ae - n¢ "* _' c- ue _ al rae pt et et — p ey! . we ote. ~* rol <7 — J < Te ~~. > » »* a_i =" *
: ae > = = ey pits. ; cea ons cE amo ~—oe TS Teens: rs airs at Beto aS ebro woes a
J m4 Say egret: ter = cow: i ne P45 t - es 7 epee an ett enh ey ess bites pa ohe be Feet eng te 7 A Ga
rigs 6, ei) onesies 7 “ T, wee me S fa P tee, - Soutien ve ae :
2 eS a ae ee, eee, ee $ : Ta eer er ; oat : ee :
= : ‘i 4 - a

: Age pe : . 3 ; | oe
: ae 3 = ae ees * ae var) att iy sn23 = _: ai
7 = Sc ; 7 = is i? or teat ease: Serie : ;
: E Mad ogni P 3 ter a - r a rs _ Faeae eee ot eee) ait, ° ‘
= - rarities Z

; “ye ;
SR ate pas ae, :

aot
Sa

is?
a

ane
ie
aie

1

i

on giee je

eens ree

4

‘

ene Bars
Pues
wed
ti ) ara we Pe:
p Mh 2 5)
ATES Y Hora
a * | {
pe 4 a a
te ¥
1 tele Oiae
ATy
¢

cod the 2 bs Rta SEU
Cue ais

Ss. Sheree eae
op gig as 2 Sere
ras £ 3 : me

4 wad aay “2 - —s

saan z

‘5 et
= ae,

ea
a) ee

a. R
Pant ii) Bi:

iii ie iy
Nis ye i!
Bt A! , ih }

: any
aha
nasi

Digitized by the Internet Archive
in 2009 with funding from —
Ontario Council of University Libraries

http://www. archive.org/details/biochemicaljournO/bioc

, ( 4 1
Pik } bry
Ae i | i

sit

ae a.
Bin:
ae ‘a ih my iy

e awer tl
mi

hae Haein es : 4 Ur tigrlk ube
. : r ¥ i an if Pay ot
4

Lai ; ee
: Ki :

Tih ' 4? Mu wy i 7
, . wy My At :

Te

BIOCHEMICAL
JOURNAL

CAMBRIDGE UNIVERSITY PRESS
Dondon: FETTER LANE, E.C.
Cc. F. CLAY, Manacer

€vinburah: 100, PRINCES STREET
London: H. K. LEWIS, 136, GOWER STREET, W.C.
WILLIAM WESLEY & SON, 28, ESSEX STREET, STRAND
Baris: LIBRAIRIE HACHETTE & CIE
Berlin: A. ASHER AND CO.
DLeipsig: F. A. BROCKHAUS
Chicago: THE UNIVERSITY OF CHICAGO PRESS
Bombay and Calcutta: MACMILLAN AND CO., Lp.
Toronto: J. M. DENT AND SONS, Lrp.
Tokvo: THE MARUZEN-KABUSHIKI-KAISHA

All rights reserved

TeEve

BIOCHEMICAL
JOURNAL

EDITED FOR- THE BIOCHEMICAL SOCIETY

BY
W> ME BAYLISS, FUR.S..
AND

ARTHUR BARDEN, F.R.S:

EDITORIAL COMMITTEE

Dr E. F. ARMSTRONG Dr FE. G. HOPKINS
Pror. V. H. BLACKMAN Pror. F. KEEBLE
Pror. A. J. BROWN Pror. B. MOORE
Mr J. A. GARDNER Dr W. RAMSDEN

Dr E. J. RUSSELL

Volume VII, 1913

Cambridge :

at the University Press

i913

—_—>+ —.

: 7
y

{

a

ee +
y
i |
‘a.
’
4
'
at
a 1
~~
‘ a
= 44
co

soe»

= vi cs
:
\
\
i
i

Dae

} '
i -
{
. = 7

‘aa
. Cambridge:
PRINTED BY JOHN CLAY, M.A.
AT THE UNIVERSITY PRESS _

2
7

CONTENTS

No. 1 (January)

I. The Rate of Protein Catabolism. By Epwarp Provan Carucart
and Henry Hamivron GREEN

Il. The Fate of Indolethylamine in the ada By Arruur JAMES
Ewins and Parrick PLayrarr LAipLaw ‘ 2 :

III. The Hydrolysis of Glycogen by Diastatic Envi Comparison
of Preparations of Glycogen from different sources. By Rotanp Vicror
Norris. (With four figures) ; ; , :

IV. The Metabolism of Organic Phosphorus Compounds. Their
Hydrolysis by the Action of ae By Roserr Henry ADERs
PLIMMER. ; : eh bet :

> The Hiavayee of Organic ne us Compounds by Dilute Acid
and by Dilute Alkali. By Sauter Henry ApERS PLIMMER

VI. The Nitrogenous Constituents of Lime-Juice. By Casimir Funk

VII. The Flower Pigments of Antirrhinum Majus. 1. Method of
Preparation. By Murien WHELDALE Rie it : : :

VIII. The Density and Solution Volume of some Proteins. By
Harrierre Curck and Cuartes JAMES MARTIN : ; : :

IX. A note concerning the Influence of Diets upon Growth. By
FreperRIcK GowLAND Hopkins and ALLEN NEVILLE . ; .

No. 2 (March)

X. The Condensation of Tryptophane and other Indole Derivatives
with certain Aldehydes. By Annie Homer

XI. On the Colour Reactions of certain Indole Derivatives, and their
significance with regard to the Glyoxylic Reaction. By Anniz Homer

XII. The Réle of Glycogen, Lecithides, and Fats in the Reproductive
Organs of Echinoderms. By Brensamin Moorr, Epwarp WHITLEY and
ALFRED ADAMS . : : > : : : 3 : 3

XIII. The Basic and Acidic Proteins of the sperm of Hehinus esculentus.
Direct Measurements of the Osmotic Presstire of a Protamine or Histone.
By Bexsamin Moore, Epwarp Wuiriey and ArtTHUR WEBSTER :

XIV. The Fatty Acids of the Human Brain. By Ecrrron CHAR.Es
GREY . : ; : 5 : : é : : . : :
XV. An Investigation of Phytin. By Rospert Henry Aprers PLIMMER
and Haroitp JAMES PAGE

101

116

vi CONTENTS

XVI. On the Relations of the Phenols and their Derivatives to
Proteins. A contribution to our knowledge of the Mechanism of Disin-
fection. Part II. A comparative study of the effects of Various Factors
upon the Germicidal and Protein-Precipitating Powers of the Phenols. ce
Evetyn ASHLEY COOPER

XVII. On the Relations of itis Bhotols aa eee De eee to
Proteins. A contribution to our knowledge of the Mechanism of Disin-
fection. Part III. The Chemical Action of Quinone upon Proteins. By
Evetyn ASHLEY CooPER : : : : : ; ‘

XVIII. The Rate of Fermentation by ce Yeast Cells. By
ARTHUR SLATOR . ‘

XIX. The Identity of Trimethylhistidine (Histidine. Betaine) from
Various Sources. By Grorce Barcer and Artaur JAMES Ewins

XX. A Note on the Metabolism of Nitrogenous Sugar Derivatives.
By James Artour Hewitt :

XXI. An Attempt to Estimate the Tataeiee: Rreacuee in “Milk. as
Casimir Funk :

XXII. The Enzymes of Washed are aa Dried Reet Geboioe)
J, Carboxylase. By Arraur HaArpDEN

No. 3 (May)

XXIII. The Critical Solution Point of Urine. By Witi1AmM RInGROSE
Grtston Arkins and THomas ArtHur WALLACE. (With one diagram)

XXIV. Quantitative Relations in Capillary Analysis. By Hans
Scumipt. (With three diagrams)

XXV. A Note on the Hopkins and Cole Modification of ints rere.
kiewicz Test for Protein. By Vernon Henry Mortrram . ;

XXVI. The Use of Litmus Paper as a Quantitative Indicator of
Reaction. By GrorGe STANLEY WALPOLE ; ; : :

XXVII. The Preparation from Animal Tissues of a Substance which
cures Polyneuritis in Birds induced by Diets of Polished Rice. Part I. By
Evetyn ASHLEY CooPER ;

XXVIII. On the Lipolytic Auton of bie Blood. Ry Heaters Hives
THIELE : : : : , : : : :

XXIX. On the Lipolytic Action of the Tissues. By Francis Hueco
THIELE

XXX. The Estimation of Tyrosine in Proteins by Bromination. By.

Rosert Henry ADERS PLimmMeR and EnizaBerH CowrEerR EAVES :
XXXI. The Separation of Cystine and Tyrosine. By Roperr Henry
ADERS PLIMMER : : : : : : E d
XXXII. The Factors concerned in the Solution and Precipitation of
Euglobulin. By Harrierre Cuick

PAGE

175

CONTENTS

No. 4 (July)

XXXITI. The Fat of Yeast. By ALLEN NEVILLE ; : :
XXXIV. The Influence of the Carbonates of the Rare Earths (Cerium,
Lanthanum, Yttrium) on Growth and Cell-Division in Hyacinths. By
Witiiam Hower Evans, (With Plate) : , ‘
XXXV. On the Chemical Nature of Substances from Alcoholic
Extracts of Various Foodstuffs which give a Colour Reaction with Phos-
photungstic and Phosphomolybdie Acids. (Preliminary Communication. )
By Casmme Funk and ArcnipaLp Bruce MacaLium : :
XXXVI. The Production of Acetaldehyde during the Anaerobic Fer-
mentation of Glucose by Bacillus coli communis (Escherich). By Earrron
CHARLES GREY : : d : - ; : : :
XXXVII. The Biochemical Synthesis of the Fatty Acids. By Ipa
SMEDLEY and Eva LuprzyNnskA : ; ote ; :
XXXVIII. The Condensation of Aromatic Aldehydes with Pyruvic
Acid. By Eva Luprzynska and Ipa SMEDLEY : : :
XXXIX. The Precipitation of Egg-Albumin by Ammonium ranphee:
A Contribution to the Theory of the ‘“Salting-out” of Proteins. By
Harriette Cuick and Cuarues James Martin. (With six figures)
XL. Some Observations on the Estimation of Urea. By Joun
ALEXANDER MILROY : : : : : - -
XLI. Gas-Electrode for General Use. By GrorGr Stantey WALPOLE.
(With four figures) : ; :
XLII. Some Esters of Palmitic Acid. By Marsory STEPHENSON

No. 5 (October)

XLII. Hydrolysis of Proteins with an Alcoholic solution of Hydrogen
Chloride. Part I. a CHARLES WEIZMANN and GANESH SAKHARAM
AGASHE ‘ ; .
XLIV. The Flower Pigments of Antirrhinum ee II. The pale
yellow or ivory pigment. _ Moriet WHELDALE and Harotp LLEWELLYN
BASSETT : . ; :

XLV. Observations on the use of the Folin method for the estimation of
Creatine and Creatinine. By WititiAm Henry Tompson, the late THomaAs
ArtTHuUR WALLACE and Haroip REGINALD Septimus CLOTWORTHY

XLVI. Note on the iodine content of ee ey ALEXANDER
T. CAMERON :

XLVII. The eaten of Ee baicbin with Oxygen and with
Carbon Monoxide. I. By Arcnispatp Vivian HILi

349

437

441

445

466

171

Vili CONTENTS

XLVIII. The Combinations of Haemoglobin with Oxygen and with
Carbon Monoxide. II. By JosrpH Barcrort. (With five figures)

XLIX. Separation of Proteins. Part IIT. Globulins. ee HENRY
Cospen HaAsLAM

L. The Chemistry of ae eee syphilidis and of the host’s
protecting cells. By James Eusrace Rapctyrre McDonacu and Roperr
Lauper Mackenzie WALLIS

No. 6 (December)

LI, Enzyme-action, Facts and Theory. By MHenprix Pieter
BaARENDRECHT, (With two figures)

LIT. An Investigation into the Physico-chemical Mechanien of Hae-
molysis by Specific Haemolysins. (Preliminary Communication.) By
Urrenpra Nato BrRAHMACHARI

LITT. Notes on some further Experiments on the Clotting of Caseinogen
Solutions. By Samugent BARNETT SCHRYVER

LIV. On the Cholesterol Content of the Tissues of Cats under various
Dietetic Conditions and during Inanition. By Joun AppyMAN GARDNER
and Percy Epwarp LANDER

LV. On the Oxidation of Gdapiceel and Copr ostanone. Part I. By
JoHN ADDYMAN GARDNER and WILLIAM GODDEN .

LVI. A note on a Modification of Teichmann’s Test for Blood. By
CLAUDE TREVINE SyMoNns

LVII. On the Behaviour of Amylase in the Presence of a Specific
Precipitate. By AGnes ELLEN Porter : . ; : .

LVIII. The Galactosides of the Brain. I. By Orro RosEnuEIm

LIX. The Estimation of Pyruvic Acid. By Iba Smeptey MacLean

LX. Note on Isocholesterol, Coprosterol and the Classification of the
Sterols. By CuHartes Dorke

LXI. The Hydrolysis of Glycogen - Diastatic inaeaek IL. - Phe
influence of Salts on the Rate of Hydrolysis. (Preliminary Communication. )
3y Ronanp Vicror Norris. (With one figure) :

LXII, The Enzymatic Formation of Polysaccharides by Yeast Prepara-
tions. By Arrnur Harpen and WitiiAmM Joun Youna

INDEX

THE RATE OF PROTEIN CATABOLISM«.

By EDWARD PROVAN CATHCART anp
HENRY HAMILTON GREEN.

Physiological Laboratory, Glasgow University.
Received November 9th, 1912.

The whole question of the course of the breakdown of the protein molecule
within the tissues is one of the most obscure in physiology. We have now
a fairly good idea of the form in which the protein is-absorbed from the
lumen of the intestine but its immediate fate is still unknown. Apparently
however definite evidence is now collecting [Folin, 1912] in support of the
view that no immediate synthesis, analogous to that of fat, takes place.
Unquestionably, irrespective of the form in which the protein material is
conveyed to the tissues, there is, soon after the ingestion of food, a fairly
complete disintegration of the protein molecule as evidenced by the increase
in the output of nitrogenous substances in the urine. Of course it might be
maintained, as it has been, that this material arises not from the newly
ingested material but from “effete” protoplasm broken down and discarded
when a new supply of repair material is available. If this be so then the
material excreted ought to bear some definite percentage relationship to the
normal protein of the body, for example the ratio of sulphur to nitrogen in
the urine should approximate to that of the average tissue as obtained say by
the study of the sulphur and nitrogen ratio in complete starvation.

It was thought that, by a careful study of the ratios of S: N after feeding
with specially chosen foodstuffs, light might be thrown on :—

(1) The rate of protein catabolism.
(2) The nature of the material catabolised.

Previous Work.

Little work has been carried out in this field probably because of the fact
that until recently the difficulty of carrying on a long series of sulphur analyses
was considerable. Since the introduction of the very excellent and rapid
method of S. R. Benedict this difficulty has largely disappeared.

1 The majority of the results here published were given in the form of a communication at
the Biochemical Society’s Cambridge meeting, October 1911.

Bioch, v1 1

2 E. P. CATHCART AND H. H. GREEN

x

Siven [1901] found that the sulphur output ran closely parallel to the
nitrogen output. He came to the conclusion that all the nitrogen excreted
does not come from a purely protein source.

Sherman and Hawk [1901] also found that the nitrogen and sulphur
outputs ran very closely parallel.

Von Wendt [1905] made a very thorough investigation into the output
of both nitrogen and sulphur. He found that when protein was broken
down in the body, the sulphur-rich digest products were the first to be
catabolised, i.e. that the sulphur output preceded that of the nitrogen and
that therefore the nitrogenous products which were retained in the tissues
were comparatively poor in sulphur. Von Wendt maintained that it is only
when the nitrogen and sulphur excretions are considered together that the
true picture of the total protein exchange in the body can be obtained:
individually considered they only tell whether certain decomposition products
are excreted.

Ehrstrém [1906] pointed out that, although the nitrogen and sulphur
excretions ran practically parallel, the sulphur output acted more rapidly with
any change in the intake. He suggested that the sulphur-containing amino-
acids were more readily oxidisable in the organism and that this might
account for the accelerated rate of output of the sulphur in comparison with
that of the nitrogen.

Falta [1906] also supported the view that the pretein molecule was
catabolised in a step-like fashion. He believed that certain nuclei were
more labile than others and that this lability would account for the appear-
ance of certain products before other constituent substances of the protein
molecule. He further pointed out in a very clear way that even in the case
of the nitrogen-containing moiety of the protein molecule the output of the
excess nitrogen or waste nitrogen was not immediate but that it might—
depending on the nature of the material ingested—be spread over several
days. Thus after the ingestion of caseinogen an amount of nitrogen equivalent
to that taken in was excreted in four days whereas after the administration
of egg albumin the excretion was continued to the sixth day.

Hamiiléinen and Helme [1907] as a result of their experiments with
a superimposed diet as used by Falta came to the conclusion that the output
of sulphur ran very closely parallel to the output of nitrogen. At the same
time they held that their results bore out fully the contention of von Wendt

that the sulphur-rich products arising from the catabolism of the protein

molecule are more rapidly burnt and excreted than those which are sulphur-
poor.

— << eC LLL

E. P. CATHCART AND H. H. GREEN 3

Since our work was completed, two papers by Wolf [1912] have appeared
which practically cover the same ground as our own. Wolf found that in
some instances the maximum of the output of sulphur preceded that of the
nitrogen but that in others the sulphur appeared simultaneously with or was
behind the nitrogen. He arrived at the general conclusion however that in
the majority of instances the sulphur part of the protein molecule is the site
of the first catabolic process.

MeETHODs.

In our experiments we also adopted the superimposition method of feeding
introduced by Falta, i.e. we added on a particular day to a standard diet the
special foodstuff the rate of catabolism of which we wished to establish. The
standard diet was always started four or five days previous to the test in
order to get the organism into a state of approximate equilibrium. (It may
be remarked here that in the majority of instances, particularly those on
a low protein intake, the lowest level of nitrogen was probably never reached.)
On a set day there was then added to the ordinary diet, as a rule with the
morning meal, the special foodstuff. Then on the following two, three or
four days the original diet was continued without addition of any kind. As
far as possible throughout the feeding period the intake of fluid was kept
constant. The subject H. H. G. lived an ordinary quiet laboratory life, taking
no undue exercise.

The diets which we employed were (1) a low nitrogen diet consisting of
tapioca and cream or potatoes and butter; (2) a diet containing a medium
amount of nitrogen made up of bread, butter and milk and finally (3) a diet
rich in nitrogen consisting of eggs, cheese, milk and bread.

As we were only concerned with the rate at which the nitrogen and
sulphur were excreted the output of total nitrogen and total sulphur were
alone considered. As a general rule, as the diets were creatine- and creatinine-
free, the opportunity was taken of studying the output of these substances
under the different conditions of the experiments. The total nitrogen was
estimated by the ordinary Kjeldahl method and the total sulphur by Bene-
dict’s method [1909], creatine and creatinine by Folin’s method.

In order to get as perfect a picture as possible, the outputs of total
nitrogen and of total sulphur were estimated every two hours during the day
of the experiment. As a rule, for purposes of comparison the estimations on
the preday, and sometimes on the day immediately following, were also
carried out on the two-hourly plan. The output of the total nitrogen in the
faeces was also determined.

1—2

4 E. P. CATHCART AND H. H. GREEN *

GELATIN.

Three experiments were carried out with gelatin superimposed on various
diets, (1) tapioca, (2) potato diet, (3) the egg diet.

The first experiment was carried out with the tapioca diet which had
been given for two days previously. The amount of gelatin given was that
present in 515 grams of a “table jelly,” which contained seven grams of
nitrogen and ‘503 grams of sulphur, with therefore an S:N ratio of 1:13°9.
The whole of the gelatin was taken with the breakfast meal at 9am. The
following table gives the result of the experiment.

TABLE I.

Gelatin feeding (tapioca diet). —

Total - Total
nitrogen sulphur Creatinine
Date and hour in grms. ingrms. N:S ratio in grms. Remarks
Sept. 16, 1910 9°78 “624 15°7 2°02
sy) Te KL 568 “0356 16:0 ‘162 Two-hourly collection.
11-1 “554 -0384 14:4 "168
1- 3 901 “0462 19°5 181
3- 5 697 0366 19°0 162
5- 7 618 0385 16-0 “7 FF
7-9 -680 "0549 12-4 "182
9-11 506 “0292 17°5 164
1l- 9 2°050 1366 15:0 ‘718
6°574 -4160 15°8* 1-92
Sept. 18,1910,9-11 , 490 0354 13°8 ‘170 ~=Fed day.
11- 1 *643 0505 12°7 ‘178 515grms, ‘‘table jelly’
1- 3 -902 0735 12°3 177 =7 grms. N taken at
5- 5 697 0607 115 "166 9 a.m.
5- 7 645 0658 9-8 166
7-9 ‘918 0594 15°5 187
9-11 838 0395 21-2 ‘160
1l1- 9 2-732 1389 Os 693
7°865 5237 15:0* 1:90
Sept. 19, 1910, 9-11 “462 0291 159 162
11- 1 539 "0352 15°3 ‘180
1- 3 566 0311 18°2 ‘170
3- 5 5TA *0284 20°2 168
5- 7 “454 0256 ifr 162
7-9 “798 0411 © 19°4 176
9-11 532 "0286 18-0 168
ll- 9 2°197 "1254 17°5 ‘706
67122 *B445 Tee 1:90
Sept. 20, 1910 a7 3189 18:1 1:90

* Ratio of totals.

———

E. P. CATHCART AND H. H. GREEN 5

When the two-hourly outputs of the preday and the fed day are compared
it will be noted that in the case of the total nitrogen the alteration from the
preday takes place late in the afternoon whereas with the sulphur the change
occurs a few hours after the ingestion of the gelatin. From an examination
of the S: N ratios the increased rate of the sulphur output over the nitrogen
is clearly seen. Even at the end of the first two hours the ratio of sulphur
to nitrogen has increased to 13°8 as compared with the ratio of the total
outputs of the previous day of 15:8. As will be noted the output of sulphur
accelerated over that of the nitrogen until the ratio stood at 9°8. Later on
however the rate of the output of nitrogen suddenly increased with the result
that the apparent lag was almost done away with, the ratio of S: N falling to
212. This gives an average for the total outputs of the sulphur and the
nitrogen for the day of 15:0, a result but little different from the preday.
The postday shows a fall to 17-7. As regards the total output of the super-
imposed nitrogen and sulphur there is evidence of a marked retention of both
substances. Of the ingested nitrogen only 18°6°/, is excreted and of the
sulphur 26:2°/,. On neither of the two postdays is there any evidence of
the excretion of the retained nitrogen, 5‘7 grams.

Unfortunately no estimation of the nitrogen excreted by way of the faeces
was carried out as previous experiments with other forms of superimposed
foodstuffs had shown that no marked excretion took place by that channel.
Still the retention here might have been only apparent, i.e. the bulk of the
nitrogen might have been excreted by way of the faeces. Of course at the
same time it was possible that a true retention of nitrogen lad taken place
especially as the subject was on a tapioca diet where a special demand for
protein might be presumed.

In order to test this possibility the gelatin in the following experiment
was superimposed on a diet rich in protein. The diet consisted of 10 eggs,
4 ozs. of cheese, 2 pints of milk, 8 ozs. of bread and 3 ozs. of butter. The
gelatin employed was the same as in the previous experiment and the same
amount was given. The only difference was that the subject had so much
difficulty in consuming it with his otherwise abundant morning meal that
about one third had to be left till the midday meal. Again, in order to get
nitrogen equilibrium on the diet previous to the superimposition there were
six predays of feeding (the urine on the sixth day being collected two-hourly).
As will be seen from the following table the effect of the superimposition was
not that of the previous experiment. Here there is neither a marked rise in
the output of either the nitrogen or the sulphur nor is there an increase in
the S:N ratio; indeed there is a slight decrease.

6 E. P. CATHCART AND H. H. GREEN

TABLE II.

Gelatin feeding (egg and cheese diet).

Total Total
nitrogen sulphur Creatinine
Date and hour in grms. in grms. N:S ratio in grms. Remarks
Oct. 9, 1911 16:0 1:231 13°0 — Weight 70:7 kilos.
5) Dy 5 ITO). 1:270 13°4 — sy AD) op
ayy la ly cs 16°7 1-292 13:0 1:97 TOO,
Be ae Sl 1:46 ‘1091 13-4 165 me ils a;
11-1 1:40 "0996 14°0 ‘178 Two-hourly collection.
1- 3 1°54 1036 14°8 “176
3- 5 1-48 1042 14:2 178
5- 7 1°46 1041 14:1 ‘176
7-9 1°76 *1295 13°6 ‘178
9-11 1:58 1315 12°0 165
1l- 9 6°93 5720 12°1 “755
17°61 1°349 13-0* 1:97
Oct. 13, 1911, 9-11 1°55 1179 13:2 ‘170 Weight 71:2 kilos.
11-1 1°66 1266 Ls 174. 515grms. ‘table jelly”
1- 3 Igy 1259 15°2 168 with 7 grms. N taken
3= 5 1:90 1213 15:7 “176 at. 9 a.m., } at 1 p.m.
5- 7 1°89 “1135 16°7 *178
7-9 1:92 "1122 17/Ail 183
9-11 1°63 ‘1136 16°9 179
11- 9 7°82 5793 13°5 736
20:28 1°410 14-4* 1:96
Oct. 14, 1911, 9-11 1°61 120 13°3 27/5) Weight 71°6 kilos.
11-1 1:69 113 14°9 °182
1- 3 1°62 ‘110 14°7 “79
3--5 =: 1°55 112 13:0 184
5- 7 1-52 "112 12°7 "176
7-9 1°88 143 13°1 180
9-11 1-20 33 10°9 176
1l- 9 7°60 “654 116 775
18°67 1:497 12°4* 2°03
Oct. 15, 1911 19-0 1:448 13-1 2°04
eee 18-6 1°389 13:4 1:99
Flies, 18°8 1°392 13°5 1:97 Weight 71:6 kilos.

* Ratio of totals.

If the normal output of nitrogen for the predays be taken as 17:1 gram,
the mean of the three predays preceding the superimposition, then practically
the whole of the ingested extra nitrogen was excreted at the end of the
second day after the feeding. 45:4°/, of the extra nitrogen was excreted on |
the day of feeding, 22:4°/, on the following day and 27:1°/, on the second
day after. For some inexplicable reason the output of nitrogen does not
return to its old level.

E. P. CATHCART AND H. H. GREEN 7

In this experiment the output of nitrogen in the faeces was followed.
The average output for four predays (collected and estimated each day) was
3:13 grams. The output on the day of feeding was 3:15 grams, on the first
postday 3°72 grams, and the ayerage daily output for the three following
days was 2°59 grams. Evidently then the retention of the nitrogen in the
first experiment was due to the actual retention of the nitrogen within the
tissues and not to an excretion by way of the intestine.

As regards the output of the sulphur as already noted there is a slight
fall in the ratio on the day of feeding to be followed however on the first
postday by a slight rise; on the other postdays the ratio practically returns
to the normal ratio found previous to the day of feeding.

If the normal output of sulphur be taken as the average of the four
predays, viz. 1°31 gram, then on the day of feeding 21°8°/, of the ingested
sulphur has been excreted, 35°8°/, on the first post-day, and 29°8 °/, on the
second. Thus the whole of the ingested sulphur, just as in the case of the
nitrogen, is excreted within three days but also, just as with the nitrogen,
instead of returning to its original level the output of sulphur continues
slightly above normal. It is extremely difficult to offer any adequate
explanation for this continued increase in the outputs of nitrogen and
sulphur unless it be that the ingestion of the gelatin has stimulated protein
catabolism.

Later a third experiment was tried in which there was a return to
a lower nitrogen intake. The diet on this occasion consisted of boiled rice
10 ozs. (weighed dry), potatoes 1 lb. (weighed after peeling), butter 6 ozs.,
milk 4 pint, containing about 5°5 grams nitrogen. The day of superimposition
was preceded by six days on the diet alone, the food being taken in four
meals. On the seventh day of the diet 670 grams of a “table jelly” were
superimposed. The subject found it absolutely impossible to consume this
amount of gelatin at the first meal, and accordingly half the amount was
taken at breakfast, one fourth at midday, and the other fourth with the
afternoon meal. No attempt was made in this experiment to collect the
urine two-hourly. The faeces were again examined for their output of
nitrogen and the following figures show quite conclusively that no loss took
place by this channel: average daily output on the predays 3:24 grams,
day of feeding 3°22 grams, on days immediately following 3:12 grams and
3°24 grams.

As regards the output of nitrogen in the urine it is found that there is an
immediate rise on the day of feeding and that there is also a distinct increase
in the output during the three following days. The same statement holds

8 E. P. CATHCART AND H. H. GREEN

TABLE ITI.
Gelatin feeding (rice, potatoes, butter and milk diet).

Total Total
nitrogen sulphur N:S Creatinine Weight,

Date in grms. in grms. ratio in grms. kilos. Remarks
March 19,1912 8°53 ‘658 12-9 Zs =
20 7:82 678 11°6 — 72:2
21 7:02 556 12:4 1°94 72:3
22 6°85 ~ §82 12°3 1°92 a
23 6°72 547 er) 1-92 72°0
24 6°60 564 aliigy/ 1:92 71'8
25 10:18 “883 11:4 IG — Onthe 25th670grms. ‘table
26 9:07 663 13°7 1:96 72:0 jelly’ taken( =9:12grms. N)
27 7°60 663 11°5 1:94 — 4at8.30a.m.,4at1.30p.m.
28 7:14 “582 12:3 o= — &that5p.m. Urine collected
29 6-60 560 11:8 ae 71:9 from 8.30a.m. to 8.30a.m.

good for the output of sulphur during these days. There is a curious
irregularity in the S:N ratio which cannot be accounted for. The results
were all checked by repeat analyses and found to be correct.

As regards the percentage amount of the ingested nitrogen and sulphur
which was excreted during the day of feeding and the three following days
the table below gives the result very clearly.

TABLE IV.

N Ss)

Percentage excreted
Fed day 379 59:0
First day after 25°7 18°5
Second ,, ,, 9°6 18°5
hird en 4:6 3°9
77'8 SHS)

Thus it will be noted that there is again quite a well-marked retention
of nitrogen although it does not take place in the case of the sulphur. This
retention is not so conspicuous as that which took place in the first experiment.
The difference in result may be due to the fact that in the second experiment
although the protein content is comparatively low still the diet is a more
mixed one. Both experiments demonstrate that the sulphur, as other
observers have previously found, is excreted more rapidly than the nitrogen.

Ece ALBUMIN.

Two experiments were carried out with boiled egg-white superimposed
in the first case on the tapioca diet and in the second on the potato and
butter diet.

E. P. CATHCART AND H. H. GREEN 9

In the first experiment the tapioca was given for two days previously, the
urine on the second day being collected two-hourly. On the third day 8 ozs.
of boiled egg-white containing 4°75 grams nitrogen and 594 gram sulphur,
thus with a S:N ratio of 1:8, were taken with the first meal. The collection
of the urine was continued on the tapioca diet for still another day. The
results obtained are given in the following table.

TABLE V.

Egg-albumin feeding (tapioca dict).

Total Total
nitrogen sulphur
Date and hour in grms. in grms, N:S ratio Remarks
Aug. 18, 1911 8:81 — — —
-. AUC oo 68 "0374 18:0 ‘Two-hourly collection.
11-1 “64 “0394 16
1- 3 63 “0344 19
3- 5 83 0436 19
5- 7 "78 "0428 18
7- 9 72 0507 14
9-11 38 “0266 14
11- 9 2-17 ci leu l 16
6°83 4066 abr hs
Aug. 20, 1911, 9-11 344 0266 13 225 grms. boiled egg-white
11-1 406 *0340 12 (=4:75 grms. N) superim-
1- 3 “589 0437 135 posed at 9 a.m.
3-5 *813 -0710 11°74
5- 7 *675 0580 ilicy,
7- 9 874 0827 10°6
9-11 *788 “0691 11:4
11- 9 2-070 *1870 1aleay
6°56 “O721 11°4*
Aug. 21, 1911, 9-11 399 0392 10-2
aL 382 0416 9-2
1- 3 580 “0419 14:0
3- 5 473 “0409 11:5
5- 7 468 0353 13°3
7-9 *A74 0366 13:0
9-11 672 *0425 15°8
11- 9 2-090 "1356 15-4
5°54 “4136 13°4*

* Ratio of totals.

It will be noted that the total output of nitrogen is actually less on the
fed day although there is an increase in the output of sulphur. When the
two-hourly outputs of the preday and the feeding day are compared it is
seen that the diminution in the output of the nitrogen is due to some

10 E. P. CATHCART AND H. H. GREEN

apparent retention during the early hours of the day. Although the rise
takes place on both days about the eighth hour, on the fed day instead of
a diminution following there is a marked increase, particularly at the twelfth
and fourteenth hours. During the night period however there is less nitrogen
excreted than during the night of the preday.

As regards the sulphur output on the fed day there is also less sulphur
excreted during the first four hours than during the same period on the
preday but thereafter there is a well-marked rise in the output. This
acceleration in the rate of the output of sulphur is particularly noticeable
when the S:N ratios are compared with those of the preday. The actual
maximum on the preday is 14, whereas on the fed day it rises to 10°6.

There is no rise in the output of nitrogen on the day following although
the output of sulphur still shows some increase.

In this experiment it is clear that under the conditions employed there
has been a complete retention of the 4°75 grams of nitrogen given, but, of
the ‘594 gram sulphur, 173 gram has reappeared, ie. 29 °/, of the amount
ingested.

A second experiment was carried out in which 11 ozs. of boiled egg-white
containing 6:5 grams nitrogen and ‘81 gram sulphur with a S:N ratio of
1:8 were superimposed on a potato and butter diet containing about 4 grams
of nitrogen (3 lbs. potatoes and 4 1b. of butter). In this experiment four
days of the diet were carried out and then the egg-white was taken with the
morning meal on the fifth day. The diet was continued and the urine
collected for two days more.

The results will be found in Table VI (page 11).

Here it will be noted that there is not a complete retention of the
ingested nitrogen although as in the previous experiment there is quite a
well-marked rise in the output of the sulphur. When the two-hourly
outputs of the fed day and the preday are compared it is found that the
outputs of nitrogen bear a very close resemblance to one another. In each
the maximum output takes place about the eleventh and twelfth hours.
In the case of the sulphur the output rapidly rises and continues well above
the preday throughout. This accelerated output of sulphur is also clearly
demonstrated when the S:N ratios are examined. The maximum ratio
on the preday is 10 whereas on the fed day it rises to 7°6, the total ratio for
this day compared with that of the preday being 9:8 to 12.

On the day following there is no continued rise in the output of nitrogen
and the sulphur output also falls back practically to the preday level,
although the S: N ratio is still high, 10-7, indicating a continued acceleration

a se ee el ee

—

- _—

See ss Ss rt‘=:;BPn™

*

E. P. CATHCART AND H. H. GREEN 11

in the output of the sulphur. The second day after the feeding shows
a further fall in the output of nitrogen and also in that of the sulphur.
The S: N ratio is however still high, being 10.

TABLE VI.

Egqg-albumin feeding (potatoes and butter diet).

Total Total
nitrogen sulphur Creatinine
Date and hour in grms. ingrms. N:S ratio in grms. Remarks
Sept. 3, 1911 9°8 — — -—
Sah: 45 8:3 “804 10-4 2°06
” 5, ” 76 “685 g fy fe 1:97
peOs <5) 9=LL -700 “0668 10°5 174
11-1 767 “0688 11-1 182
1- 3 526 0536 10:0 170
3- 5 743 “0597 12°4 ‘180
5- 7 *637 0626 10°2 180
7-9 “980 0653 15:0 182
9-11 ‘764 “0584 13% “186
11-9 3°100 *2482 12°5 “850
8°217 “6834 12-0* 2:10
Sept. 7, 1911, 9-11 768 “0604 12°6 ‘177 312 grms. boiled egg-
11-1 ‘788 0653 12°] ‘171 ~+=white (=6°5 grms. N)
1- 3 663 0664 10:0 "170 superimposed at 9 a.m.
3- 5 786 0974 81 ‘181
5- 7 786 1027 76 173
7-9 1:037 0933 11:0 173
9-11 781 ‘0871 9:0 169
1l- 9 3°394 3436 9-9 737
9011 “9162 9°8* 2:00
Sept. 8, 1911, 9-11 722 “0612 11°8 168
11-1 755 “0728 10°4 178
1- 3 636 0584 9:2 “176
3-5 522 “0538 Shy) 174
5- 7 593 -0608 9-7 182
7-9 *805 0755 10°6 188
9-11 *590 "0584 1071 169
11-9 2°721 "2442 ili "734
7344 “6851 LOTS 1:97
Sept. 9, 1911 6°70 “676 10:0 1:97

* Ratios of totals.

Thus of the 65 grams of nitrogen taken in (the mean of the three pre-
days’ output of nitrogen being 8:01 grams and of the sulphur for two days
‘684 gram) there has only been 1 gram or 15°3°/, excreted and that on the
day of feeding. Of the sulphur ‘232 gram or 28°6"/, was excreted on the
day of feeding and a mere trace on the first day after.

12 KE. P. CATHCART AND H. H. GREEN

These results do not agree very well with those obtained by Wolf. He
found that the sulphur output lagged behind that of the nitrogen. He had
previously observed the same curious inversion of the output rates in his
feeding experiments with uncoagulated egg albumin. This delay in the
output of the sulphur did not take place however when predigested egg
albumin was fed; in this case the sulphur output preceded that of the
nitrogen.

We had also carried out a couple of experiments with uncoagulated egg
albumin but we had used instead of the ordinary raw material from eggs
a finely powdered dry product (Egg albumin extra fine powder, Merck). In
both experiments the albumin was superimposed on a tapioca diet. As these
experiments were amongst our earliest they were not so thoroughly carried
out as those previously described. In the first, in which 6°7 grams of nitrogen
were given, there was a retention of at least 53 grams nitrogen. (The
sulphur output was not measured.) In the second experiment, in which the
same amount of nitrogen was given, there was a retention of about 60 grams.
In neither case was the urine collected on any day following the feeding.
As regards the output of sulphur in the second experiment the total amount
excreted rose above that of the preday, the S: N ratios altering from 14°3 on
the preday to 11°5 on the day of feeding. In this experiment, although it
was not very definite, there was some support for Wolf’s contention that
when uncoagulated egg albumin is fed the sulphur is at first not so rapidly
excreted as the nitrogen, but that later, when the excretion does start, it
proceeds at a faster rate than that of the nitrogen. Wolf believed that the
sulphur complex in the uncoagulated egg albumin has some power of with-
standing the onset of the ferments.

PLASMON.

Only one experiment was carried out with this substance in which it was
superimposed on a bread, butter and milk diet. After twelve predays of
feeding, 59 grams of nitrogen in the form of plasmon were superimposed
daily for eight days with the result that there was a consistent retention of
nitrogen throughout. The mean daily output of nitrogen before the plasmon
was added was 15°6 grams in the urine and faeces and despite the daily extra
intake of 5°9 grams during the plasmon period the mean daily total output
was only about 17-4 grams, giving a retention of nitrogen during the whole
plasmon period of about 33 grams nitrogen. The four days, which made up
the after-period, did not show very conspicuously the slow fall, which was to

EK. P. CATHCART AND H. H. GREEN 13

be expected from previous work, in the output of nitrogen to the original

level. It may be noted that although the subject was in very good condition

throughout this experiment there was a steady although slight gain in body

weight. The sulphur output showed nothing abnormal.

Wolf in his

plasmon experiment also found that there was no definite evidence in this
ease that the sulphur output preceded the output of nitrogen.

TABLE VII.

Veal feeding (bread and butter diet).

Total Total
nitrogen sulphur N:S Creatinine

Date and hour ingrms. ingrms. ratio ingrms.
Feb. 19, 1912 9°76 = — —
=; 2s 8°62 625 13°8 1°89
“3 PAR PSS 7:90 662 11:9 1:96
+ CP 701 616 11-4 1:96
aos 55 7:73 664 11°6 1-93
Peat. ',,, .8=10 575 0436 13:2 162
10-12 731 0586 12°5 178

12-— 2 582 0477 13-0 172

2- 4 605 0452 13°4 ‘176

4- 6 583 0460 12°7 ‘178

6- 8 732 0662 hitcal 176

8-10 680 0663 10°3 “180

10- 8 2-910 "2583 11:3 ‘779
7°40 629 TUL fee Wess)

Feb. 25, 1912, 8-10 579 0482 12-0 ‘175
10-12 847 0592 14:3 206

12- 2 991 0574 17:3 “206+

2- 4 “964 0614 15°7 196

4- 6 1131 ‘0773 14:6 “200

6- 8 1:360 0945 14:4 aS

8-10 983 0801 12°3 eo

10- 8 3577 3380 10°9 “750
10°43 8081 12S) 2°10

Feb. 26, 1912, 8-10 *882 0710 12°4 a9
10-12 929 0702 13:2 184

12- 2 “838 ‘0611 13°7 180

2- 4 ‘778 0616 12°6 “181

4- 6 “692 0583 11:9 sur

6- 8 “809 0710 11:4 180

8-10 723 0625 11-4 182

10- 8 2°932 *2658 11:0 792
8:58 7215 lO 2°05
Feb. 27, 1912 8:16 668 12-2 1:96
e286.” 55 7:87 630 12°5 1:92
99° 20, a» 7°34 “596 12°3 1:90

* Ratios of totals.

Weight,
kilos.
12°3
72°4
72:5
12°5
72°6
72°8

73°0

Remarks

26 grms. stewed veal
(=10-6 grms. N) su-
perimposed at 8.30
a.m.

+ Trace of creatine

present.

14 E. P. CATHCART AND H. H. GREEN

VEAL.

In this experiment 226 grams of veal containing 10°6 grams of nitrogen
and ‘664 gram of sulphur with a S:N ratio of 1:16 were superimposed on
a bread and butter diet which was taken for six days previous to the day of
feeding. The sixth preday urine was collected two-hourly. Following the
day of feeding there were four postdays. Table VII (p. 18) gives the
general result.

Here it will be noted that there is a well-marked rise in the excretion of |
nitrogen as the result of the feeding, an excretion which is spread over the
day of feeding and the three following days. In the case of the sulphur
there is also a rise in the output but the normal is regained at the end of
the third day after feeding. When the percentage amounts of the intake
of both nitrogen and sulphur are compared it is found that here there is no,
or perhaps only a very slight, acceleration in the output of the sulphur as is
seen from the following table.

TABLE VIII.

N SS)

Percentage output
Day of feeding 28°6 27-0
First day after 11:2 14:0
Second ,, 7°2 50
Third e- 4:4 —

Thus of the nitrogen and sulphur taken in, 56°1°/, (including 4°7°/, in the
faeces) of the nitrogen is excreted in four days, and 46°/, of the sulphur in
three days. |

This result fully substantiates the result obtained by Wolf who found
that about 45°/, of the nitrogen taken in the form of veal was re-excreted
within two days. He also found that the nitrogen and sulphur output ran
nearly parallel, although the output curve of the sulphur was a little steeper
than that of the nitrogen, indicating that the former output was slightly
more rapid. It does not however agree with the results obtained by
Hamaliinen and Helme, who did not find any marked retention of the
nitrogen and who found a much more rapid excretion of the sulphur than of
the nitrogen.

UREA.

Finally a test with urea was carried out on a comparatively low protein
diet to see if, when a nitrogenous material was used which we presumed
could not take the place of protein, retention of nitrogen took place. After

E. P. CATHCART AND H. H. GREEN 15

giving a potato and butter diet for four days twelve grams of pure urea
containing 5°6 grams of nitrogen were superimposed on the breakfast meal.
We found that over 95°/, was excreted during the day of feeding and on the
subsequent day. There was thus practically no evidence of retention. Wolf
obtained a similar result: he found an excretion of nearly 97°/, in the first
two days. This result does not favour the idea that urea can replace to
a certain extent the protein of the diet. If this were possible it would have
been expected that under the particularly favourable conditions of our
experiment, urea given on a comparatively low protein diet, a certain amount
of retention of nitrogen would have occurred. The ingestion of the urea had
no influence on the output of sulphur.

CREATININE OUTPUT.

The opportunity was taken in the course of the above experiments to
follow the output of creatinine as it was thought that the effect of the
addition of a nitrogenous substance free from creatine and creatinine to a
diet poor in protein might throw some additional light on the course of the
metabolism of creatine and creatinine. Our results show that the output of
creatinine was scarcely affected by such additions, except of course in the
case of the veal which contained creatine in small amount. This substantiates
the statement of Folin that the output of the creatinine is maintained
practically at a constant level.

DISCUSSION OF RESULTS.

From these experiments then, and from others carried out by Wolf, etc.,
it may be concluded that Falta is right in his contention that the breakdown
of protein takes place in a more or less step-like fashion. The fact that in
nearly every instance there is an attempt by the cells of the organism to
retain a certain amount of the nitrogen and to get rid at the same time of
the excess of sulphur points to the fact that the form in which the nitrogen
is retained, and apparently it is retained, although perhaps to a very
limited extent, is a special one. The evidence for the storage of protein in
the body is very scanty and has previously been discussed by one of us
[E. P. C., 1912] at some length. Briefly it may be said that the general
opinion is that a certain amount of retention can take place although the
form in which the material is retained is still a matter of dispute. Some of
the most interesting work in this connection is that of Miiller [1907] who fed

16 E. P. CATHCART AND H. H. GREEN

dogs on a definite low protein diet and then amputated a limb. Subsequently
the animal was fed on a high protein diet and killed after a long period of
feeding. An analysis of the tissues obtained from the amputated limb on the
low protein diet compared with the analysis carried out after the high protein
diet showed that a certain amount of retention had taken place. Recently
Diesselhorst [1911] repeated this work of Miiller under practically the same
conditions but making more elaborate analyses. He also found that there
was a certain gain in the amount of nitrogenous material after feeding on the
protein-rich diet. Grund [1910] also investigated the alterations in the
composition of the tissues in starvation and after feeding. He discusses this
question. of the retention of reserve protein in the cells and comes to the
conclusion that there is a certain amount of evidence in its favour, although
he at the same time maintains that if this retention does take place it
cannot play a very important part in the total metabolism. He further holds
that there is a general tendency both in periods of starvation and of feeding
for the tissues to maintain the same relative composition.

In conclusion we believe that from the study of the S:N ratios it is
clearly shown that the increase in the output of nitrogen and sulphur which
as a rule follows the ingestion of a protein meal is due to the catabolism of
the protein actually ingested and not to the displacement of. “ effete” proto-
plasm from the tissues. It will be noted for example after the superimposition
of the egg albumin with aS:N ratio of 1:8 (see Table VI) that the ratio
obtained approximates very closely to the ratio of the ingested protein.
This is particularly noticeable in the variations in the ratio in the two-hourly
collection of the, preday and the fed day. One of us [E. P. C., 1907] has
previously shown that the S:N ratio in starvation when all the nitrogen and
sulphur which is excreted in the urine must come from an endogenous source
is about 1:15. If then the extra nitrogen and sulphur excreted after the
ingestion of the egg albumin had come from the displacement of “ effete”
tissue protein it would have been expected that the S:N ratio would be
nearer 1:15 than the 1: 9°8 found.

CONCLUSIONS.

1. The sulphur-containing moiety of the protein after ingestion is, as
a rule, more rapidly catabolised and the sulphur more rapidly excreted than
the nitrogen.

2. When protein is superimposéd on a low protein diet a retention of
part of the nitrogen superimposed takes place.

E. P. CATHCART AND H. H. GREEN 17

3. The retained material is apparently stored in the tissues as a pabulum
of uniform composition.

4, The rise in the output of nitrogen and sulphur after a protein meal
is due to the catabolism of the actual material ingested.

5. The superimposition of protein has, in these experiments, no effect on
the output of creatinine.

REFERENCES.

Benedict (1909), J. Biol. Chem. 6, 363.

Cathcart (1907), Biochem. Zeitsch. 6, 109.

Catheart (1912), Physiol. of Prot. Metabolism, 58, 80.
Diesselhorst (1911), Arch. ges. Physiol. 140, 256.
Ehrstrém (1906), Skan. Arch. Physiol. 18, 281.
Falta (1906), Deut. Arch. klin. Med. 86, 517.

Folin and Denis (1912), J. Biol. Chem. 11, 87.
Grund (1910), Zeitsch. Biol. 54, 173.

Hiimilainen vy. Helme (1907), Skan. Arch. Physiol. 19, 182.
Miiller (1907), Arch. ges. Physiol. 116, 207.

Sherman and Hawk (1901), Amer. J. Physiol. 4, 25.
Siven (1901), Skan. Arch. Physiol. 10, 91.

von Wendt (1905), Skan. Arch. Physiol. 17, 211.
Wolf (1912), Biochem. Zeitsch. 40, 193.

Bicoh. yu

II. THE FATE OF INDOLETHYLAMINE IN
THE ORGANISM.

By ARTHUR JAMES EWINS anp PATRICK PLAYFAIR LAIDLAW.
From the Wellcome Physiological Research Laboratories, Herne Hull.

(Received November 10th, 1912.)

It has been shown by several observers that the naturally occurring
amino-acids are readily attacked by bacteria, and that among other products
the corresponding amines are formed [Winterstein and Kuntz, 1909;
Barger and Walpole, 1909; Ackermann, 1910; see Barger, 1911, for other
references]. It appears probable that they are produced in the intestine in
the course of the normal life of the organism and also in certain pathological
conditions. Their fate after absorption is therefore a matter of considerable
importance. The amines of this type have in most cases been submitted to
physiological investigation [Dale and Dixon, 1909; Ackermann and Kutscher,
1910; Dale and Laidlaw, 1910], and in several cases they have been shown
to be active substances. The method by which the body deals with such
bases has not been investigated so thoroughly.

Some time ago we showed [Ewins and Laidlaw, 1910] that p-hydroxy-
phenylethylamine was largely converted into p-hydroxyphenylacetic acid
and excreted as such; that the liver and the plain muscle of the body, with
the exception of the plain muscle of the lung vessels, could effect the change :
and that successive methylation of the amino-group rendered the base
increasingly resistant to destruction in the body. Another example of
a change of precisely the same type is met with in the fate of benzylamine,
which was shown by Mosso [1890] to be almost quantitatively converted into
benzoic acid and excreted as hippuric acid. We undertook the investigation
of the mode of destruction of indolethylamine (8-indole-pr-3-ethylamine) in
order to see whether it was dealt with in a similar manner, and whether its
administration would increase the kynurenic acid output in dogs.

The indolethylamine used was made according to the synthesis devised
by one of us [Ewins, 1911], which readily gives a good yield of the base.

A. J. EWINS AND P. P. LAIDLAW 19

We had already shown [Ewins and Laidlaw, 1910] that it was possible to
obtain this base by the action of putrefactive bacteria on tryptophane ; but
the method is tedious and troublesome, and the yield very poor.

We have carried out two sets of experiments in the course of our
investigation: (1) feeding experiments on dogs, and (2) the perfusion of
surviving livers of rabbits and cats’,

PERFUSION EXPERIMENTS.

The method of perfusion was similar to that already described in our paper
dealing with the fate of p-hydroxyphenylethylamine [Ewins and Laidlaw, 1910].
Small quantities of indolethylamine were perfused through surviving livers for
3—4 hours. The perfusion fluid gradually acquired the property of giving
a fine red colour, when it was mixed with one-third of its volume of strong
hydrochloric acid, with the addition of one drop of dilute ferric chloride, and
then boiled. The red colour occasionally had a faint blue component if the
hydrochloric acid was in excess or if the boiling was prolonged. A cherry
red colour also developed if a trace of sodium nitrite was substituted for the
ferric chloride in the above tests. Amy] alcohol rapidly extracted the pigment
from the solution, when a well-defined absorption band in the green was seen
with a spectroscope. The red colour is brighter and purer and the absorption
band more intense and better defined when sodium nitrite is used. When
the chromogen of the pigment appeared to be fairly abundant in the
perfusion fluid (usually about 3-4 hours) the perfusion fluid was collected,
and the liver vessels washed through with salt solution. The combined
washings and perfusion fluid were rendered faintly acid with acetic acid
and boiled. The coagulated proteins were filtered off and the filtrate
evaporated to small bulk. The concentrated perfusion fluid was made acid
to Congo-red with strong hydrochloric acid and shaken out with ether, which
readily and completely extracted the chromogen. The ethereal extracts
were combined, washed twice with water and taken to dryness. If the
residue were taken up in water and allowed to stand, a small quantity of
a crystalline acid was regularly obtained, which when recrystallised from
benzene melted at 163-164°. These crystals gave the colour reactions
mentioned above with great intensity and readily gave a deep orange-red
coloured picrate melting at 174°. These properties taken together are
characteristic of indoleacetic acid.

1 The animal experiments were performed by P. P. Laidlaw only.

20 A. J. EWINS AND P. P. LAIDLAW

Experiment. 0°25 grm. indolethylamine hydrochloride was perfused
through a rabbit’s liver for 25 hours. Perfusion fluid and washings were
worked up as described above when 50 mgm. of indoleacetic acid were
obtained. Yield 44 °/, of theoretical.

The identification was completed by comparison of colour reactions,
melting point, and mixed melting point with synthetic B-indole-pr-3-acetic
acid.

There can be no doubt then that imdolethylamine is readily converted
by the liver into indoleacetic acid, just as p-hydroxyphenylethylamine is
converted into p-hydroxyphenylacetic acid.

FEEDING EXPERIMENTS.

One half to one gram of the hydrochloride of indolethylamine is well
borne by a 7-8 kilo dog. Larger doses are inadmissible for our purpose
since vomiting and sometimes purgation ensue.

The urine of the 36 hours following the administration did not give the
colour reactions which we have always found in our perfusion experiments,
but on adding to a small quantity of urine half its volume of strong
hydrochloric acid and a trace of nitric acid and gently warming, a fine purple
colour develops, which gradually (unless the mixture 1s cooled) becomes deep
red and. then orange and ultimately yellow. The colour reminds one of
a very intense indican reaction. It is not, however, due to the development
of indigo since at no stage can a blue component be extracted by chloroform.

We found that the chromogen of this colour reaction was an acid, which
could be shaken out from the strongly acidified urine with ether or ethyl
acetate, but the complete extraction of the chromogen by either of these
solvents from acid urine was found to be almost impossible. The acid,
however, was found to be much more readily extracted if its solubility in
urine was depressed by saturation with ammonium sulphate. In a few
experiments with rabbits’ urine this preliminary saturation with ammonium
sulphate was found to be of considerable assistance, since the procedure
caused a large proportion of the hippuric acid to crystallise [cf. Roaf, 1908].

The ether or ethyl acetate extracts were carefully collected, and washed
with ammonium sulphate, and then with a small quantity of water. The acid
was then extracted from the ether by sodium carbonate solution. The alkaline
extract was acidified, saturated with ammonium sulphate and shaken out
with ether. The ether extracts were washed with water and taken to dryness.
A thick gum resulted from which nothing would crystallise. The gum was

A. J. EWINS AND P. P. LAIDLAW 21

extracted with boiling water, in which most of it dissolved, excess of picric
acid was added and the whole boiled with a small quantity of charcoal for
five minutes. On filtration and prolonged standing the filtrate gradually
deposited a deep orange-red coloured picrate. This picrate was recrystallised
from water or water with a trace of acetone several times, when it separated
in large, orange-red, rhomboidal plates melting at 145°. The erystals gave
the colour reaction with great intensity. It appeared to us to be very
probable that this acid would prove to be an indoleacetic-glycine complex,
comparable with hippuric and phenaceturic acids, but for some time we were
unable to obtain sufficient material to prove or disprove our assumption.
The yield of picrate of the acid is not good and, as we have pointed out, the
feeding experiments are limited to the administration of small quantities.

In one experiment for example, from 0°5 grm. of the hydrochloride of the
base we only obtained 0:2 grm. of the picrate or a little over 20°/, of the
theoretical maximum yield. In some other experiments the yield was rather
better but never more than 30°/,.

From a series of three experiments we obtained sufficient of the picrate
which, although not quite pure, could be analysed :

0:0986 grams picrate gave 0°1668 CO, and 0:0322 H,O
C=461 H=36
C,.H,.0;N..C,H;0,N; requires C = 46°8, H = 3:2 per cent.

The figures obtained although not accurate were sufficiently so to indicate
that the acidic chromogen was an indoleacetic-acid-glycine condensation
product, and that the substance in question has the probable constitution
denoted by the formula

7 eS oes CO. NE OH, COOH,
|
oes

For this acid we suggest the name indoleacetwrie acid, in conformity with
the nomenclature of this class of bodies to which hippuric and phenaceturic
acids belong.

It may also be mentioned that quite recently Ackermann [1912] has
shown that among the products of metabolism of nicotinic acid fed to a dog
is found a similar glycine derivative, to which he gives the name nicotinuric
acid.

Our supposition with regard to the constitution of the urinary chromogen
was further confirmed by hydrolysis of the picrate with sodium carbonate.

22 A. J. EWINS AND P. P. LAIDLAW

Experiment. 0°250 grm. of the picrate was dissolved in 80 c.c. of 5 °/, sodium
carbonate solution and hydrolysed on the water bath for one hour. On cooling
and acidifying, 0°19 grm. of indoleacetic acid picrate separated out, and from
the mother liquor 0:017 grm. more was obtained crystalline. Total 0:207 grm.
or 95°/, of the theoretical amount if the original picrate had been the
suggested indoleacetic-glycine complex.

Since the amount of acid obtainable in the experiments with indolethyl-
amine was so very small, and we have shown by perfusion experiments that
indoleacetic acid is probably an intermediate step in the production of
indoleaceturic acid from indolethylamine, we carried out a series of feeding
experiments with indoleacetic acid in the hope of obtaining larger quantities
of the desired acid. Several grams of indoleacetic acid were synthesised
by a method which will be published at a later date, and given by mouth to
a dog, when a 60°/, yield of picrate was obtained, identical in all respects
with that obtained from the feeding experiments with the base. This fact
enabled us to obtain the picrate in comparatively large quantities, since
the indoleacetic acid caused no symptoms. In this way we obtained
without trouble 3 grams of indoleaceturic acid picrate from 3 grams of indole-
acetic acid. The preparation of the free acid from the picrate could not be
carried out in the usual manner (removal of the picric acid from the acidified
solution by means of ether) owing to the well-marked acidic properties of the
indoleaceturic acid and its solubility in ether. ‘This difficulty was overcome
by the use of the base known as nitron (1.4-diphenyl-1.3.5-endanilino-
dihydrotriazole) [Busch, 1905] which precipitates picric (as well as nitric)
acid as an almost insoluble salt of the base. The experiment was carried out
as follows.

15 grams of picrate were dissolved in 150 c.c. of hot water and 1 mole-
cular proportion of nitron dissolved in hot dilute alcohol was added. The
mixture was thoroughly cooled and the precipitated picrate sucked off at the
pump. The nearly colourless filtrate was neutralised with sodium carbonate,
two or three drops of acetic acid added and the solution evaporated to dryness.
The residue was dissolved in a few (7-8) c.c. of water and made just acid to
Congo-red with dilute hydrochloric acid. An oily product separated which soon
crystallised in bunches of needle-like prisms. This was filtered off, dried in
vacuo over sulphuric acid for a short time and weighed. In this way 0°6 gram
of the crystalline acid was obtained. This was twice recrystallised from
water and then melted at 94°. The acid so obtained is very slightly soluble
in cold water (about 0°2°/, at ordinary temperature), almost insoluble in
ligroin and benzene, easily soluble in alcohol, ether and ethyl acetate. It

;
;

A. J. EWINS AND P. P. LAIDLAW 23

erystallises from water with one molecule of water of crystallisation which
is only very slowly removed by sulphuric acid in vacuo at 37°.

0:1022 gram acid (m.p. 94°) gave 02184 CO, and 0:0514 H,O
C=582 H=56
SalerG. ..0,N,"H.O C=57'6 H =5°6 per cent.
01026 gram acid (anhydrous) gave 0:2330 CO, and 0:0492 H,O
C=619 H=53
Cale. C,,H,,0,N, C= 62:0 H = 5:2 per cent.
This result fully confirms our supposition that the acid in question is indole-
aceturic acid.

It will be observed that we have never obtained anything like a quantita-
tive yield of indoleaceturic acid by giving indolethylamine by mouth. In
our best experiments about 30°/, of the base is accounted for in this manner.
It is true that this probably represents a much larger quantity of indoleacetic

acid, which must be the intermediate product; because the administration of

indoleacetic acid has never in our experience given a quantitative yield of
indoleaceturic acid. However, making some allowance for this and a little
for the imperfect methods used in the isolation of the acids from such a
complex mixture as urine, we still have not accounted for the whole of the
base. .

It appeared to us to be possible that some of the indolethylamine or
indoleacetic acid might be converted into kynurenic acid, although none of
the specimens of urine that we had examined showed obvious excess of this
metabolite. Some experiments were therefore undertaken in which kynurenic
acid estimations were made before and after administration of indolethylamine
and indoleacetic acid.

A dog was kept on a standard diet of milk and dog biscuit. On this
particular diet the kynurenic acid excretion was minimal, and traces only
could be detected in 24 hours. The kynurenic acid was estimated by Capaldi’s
method [1897].

Experiment.
Date Volume of urine Kynurenic acid Remarks

Sept. 23—24 400 c.c. Trace ==

Sept. 24—25 440 c.e. Trace =

Sept. 25-—26 600 c.c. -480 grm. On morning of 25th 1°6 grm.
tryptophane by mouth.

Sept. 30—Oct. 1 350 c.e. Nil =

Oct. 1—2 600 c.e. Trace On Oct. 1st 1-6 grm. indoleacetic
acid by mouth.

Oct. 2—4 850 c.e. Traces =

Oct. 4—5 470 c.e. _ Traces On Oct. 4th 1-0 grm. indolethy-

lamine by mouth.

24 AL'S. EWINS AND’ P. P. LAIDLAW

Neither indolethylamine nor indoleacetic acid increases the kynurenic
acid output in dogs.

It seems clear that this path of metabolism of indole-derivatives does not
aid us in accounting for the quantitative difference between the ingested
substances and those excreted.

Samples of urine were examined in various ways for other possible end
products but without success, and we suggest that the portion of the base
unaccounted for is completely broken up in the body.

OCCURRENCE OF INDOLEACETURIC ACID.

We have some evidence of the occurrence of indoleaceturic acid in normal
urine of herbivora. We have found that an acidic, ether-soluble substance,
giving the purple colour reaction, occurs in small quantity in the urine of
rabbits, but attempts at isolation of the acid or its picrate have always failed.
The small amount of the chromogen cannot be separated from the other acid,
ether-soluble substances present in all urines.

In examining urine for indican it occasionally happens that a fine purple
colour is produced by hydrochloric acid and an oxidising agent but no blue
component can be extracted by chloroform. Here, again, it is suggested that
the chromogen is indoleaceturic acid, but considerable quantities of urine
would be required to isolate the acid and such have not been at our disposal.

Herter [1908] described a case of a young girl who was suffering from
some unusual intestinal infection and from whose faeces an unusual organism
was isolated. The urine of the patient gave a marked urorosein reaction if
the urine were stale. Indoleacetic acid was isolated from this urine and
Herter identified it as the chromogen of urorosein originally described by
Nencki and Sieber in 1882. The colour, solubilities and other characteristics
of the pigment agree very well with those which Nencki and Sieber described
as characteristic for urorosein, although other observers have described other
substances as the chromogen of the same pigment. It is unfortunate that
Herter gives no description of the method of isolation of indoleacetice acid
from the urine of his patient. It is quite possible that it was originally
present as indoleaceturic acid and that this was hydrolysed to indoleacetice
acid in the process of isolation, for the complex acid is readily hydrolysed by
weak alkalies such as sodium carbonate. The fact that stale urine gave a colour,
while fresh did not, was attributed by Herter to the formation of nitrites. It is
quite possible that the indoleaceturic acid was decomposed into indoleacetic
acid and glycine, just as hippuric acid is readily split up by bacteria, and

A. J. EWINS AND P. P. LAIDLAW 25

that the nitrites, though certainly playing a part, did not furnish the whole
explanation.

If it could be demonstrated that indoleaceturic acid was a normal con-
stituent of urine it would by no means follow that its precursor in the
body was indolethylamine, for bacteria often produce indoleacetic acid from
tryptophane and proteins; in fact it is a more frequent product of bacterial
action than indolethylamine.

SUMMARY.

1. Indolethylamine is converted, by the perfused liver, into indoleacetic
acid.

2. Indoleacetic acid is excreted in combination with glycine forming
indoleaceturic acid.

3. About 30°/, of a given dose of indolethylamine is excreted as indole-
aceturic acid in dogs.

4, Neither indolethylamine nor indoleacetic acid affects the kynurenic
acid output in dogs.

REFERENCES.

Ackermann (1910), Zeitsch. physiol. Chem. 65, 504.
Ackermann (1912), Zeitsch. Biol. 59, 17.

Ackermann and Kutscher (1910), Zeitsch. Biol. 54, 387.
Barger (1911), Science Progress, 6, 221.

Barger and Walpole (1909), J. Physiol. 38, 343.

Busch (1905), Ber. 38, 859.

Capaldi (1897), Zeitsch. physiol. Chem. 23, 92.

Dale and Dixon (1909), J. Physiol. 39, 25.

Dale and Laidlaw (1910), J. Physiol. 41, 318.

Ewins (1911), J. Chem. Soc. 99, 270.

Ewins and Laidlaw (1910), J. Physiol. 41, 78.

Ewins and Laidlaw (1910), Proc. Chem. Soc. 26, No. 346.
Herter (1908), J. Biol. Chem. 4, 239, 253.

Mosso (1890), Arch. expt. Path. Pharm. 26, 267.

Roaf (1908), Biochem. J. 4, 185.

Winterstein and Kuntz (1909), Zeitsch. physiol. Chem. 59, 138.

Ill. THE HYDROLYSIS OF GLYCOGEN BY DIA-
STATIC ENZYMES. COMPARISON OF PRE-
PARATIONS OF GLYCOGEN FROM DIFFERENT
SOURCES.

By ROLAND VICTOR NORRIS, Bett Memorial Research Fellow.
From the Biochemical Department, Lister Institute.

(Received November 23rd 1912.) .

For some time considerable uncertainty has prevailed as to whether
samples of glycogen prepared from different sources are identical, the
evidence on this point being in many respects contradictory. It is agreed
that all have the same empirical formula C,H,,O;, the different results
obtained by some observers being undoubtedly due to insufficient drying of
their glycogen preparations. For Harden and Young [1902] found that the
last traces of water can only be removed by heating to 100°C. in vacuo over
phosphorus pentoxide, a procedure which had not been adopted by previous
workers. Very divergent values have however been obtained for the specific
rotation of different preparations of glycogen, the results covering a range of
about 30°. Thus Cremer [1894] found as the [a]p of yeast glycogen + 198°9°,
while Clautriau’s [1895] result was 184°5°. Harden and Young [1902] also
obtained for oyster and rabbit glycogens the value 191:2° while with yeast
glycogen the mean of several determinations was 198°3°, a result in close
agreement with that found by Cremer. It is uncertain however whether
these differences are of any real significance, the strong opalescence of
glycogen solutions making necessary the use of very dilute solutions and
thereby introducing a large experimental error.

Other points of difference have however been noticed by various observers,
though again the results are contradictory. Cremer for example states that
yeast glycogen gives a darker colouration with iodine than does oyster
glycogen, while other workers have obtained exactly opposite results. In
a similar way different preparations of glycogen yield solutions with varying
degrees of opalescence but these variations are by no means constant either
in nature or extent. At present it is impossible to explain these results and

R. V. NORRIS 27

at the same time it is uncertain how far divergences of this nature imply
any real differences between the respective glycogens.

A more valuable comparison was made by Harden and Young when they
measured the rate of hydrolysis of. different glycogen solutions by dilute
acids. The course of hydrolysis was followed by estimations of the polari-
metric and reducing powers of the solutions. By this method they were
unable to detect any differences between rabbit, yeast and oyster glycogens
and came to the conclusion that they were in all probability identical.

It seemed possible however that a study of the behaviour of various
preparations of glycogen towards diastatic ferments might prove a more
delicate means of detecting any difference in their constitution, for it has on
many occasions been pointed out what marked influences even small
alterations in the molecule may exert on the course of enzyme action. It is
the results of such an investigation which are presented in this communi-
cation. Four preparations of glycogen have been employed, these being
derived from (a) liver of dog, (b) liver of rabbit, (c) oyster and (d) yeast.
The enzyme used was an extract of pig’s pancreas, this being one of the most
convenient sources for obtaining active preparations. Before making any
direct comparison, the general conditions of action were studied, for although
a great amount of work has been carried out on the behaviour of starch
towards diastatic enzymes, the hydrolysis of glycogen has received very much
less attention. These experiments will be first described.

PREPARATION OF GLYCOGEN.

A. Dog. A large dog was fed with considerable quantities of carbo-
hydrate for forty-eight hours and then killed. The liver was rapidly
removed, minced and heated on the water bath with 60 per cent. caustic potash.
The glycogen was precipitated from this solution in the usual way [Pfliiger,
1905] and purified by repeated reprecipitation from its aqueous solution, the
first two or three precipitations being carried out with solutions rendered
faintly acid with acetic acid. Glycogen was in this way obtained in consider-
able quantity, free from nitrogen and containing only a trace of ash. In the
final stages it is frequently difficult to produce a satisfactory precipitation

with alcohol alone, the addition of a small quantity of acetone in these cases,

however, readily brings down all the glycogen, leaving a liquid which filters
well.

B. Rabbit. The method was the same as that described above, a dozen
rabbits being used for each preparation. They were previously fed on large
quantities of carrot.

28 R. V. NORRIS

C. Oyster. <A gross of oysters (1500 grms.) were worked up by Pfliiger’s
method and yielded 120 grms. purified glycogen.

D. Yeast. The glycogen was separated and purified from yeast gum by
the method recently described by Harden and Young [1912] with the
exception that yeast juice was employed as the starting material. Large
quantities of yeast juice were boiled, filtered and the filtrate precipitated
with alcohol. The precipitate was then heated on the boiling water bath with
60 per cent. caustic potash to remove protein, the solution diluted and again
precipitated with alcohol. After thorough washing with 50 per cent. alcohol
and reprecipitation, the glycogen was purified from yeast gum by saturation
of the solution with ammonium sulphate, which precipitates the glycogen but
not the gum. This process was repeated three times and the glycogen then
dialysed to remove the ammonium sulphate. The solution was then again
precipitated with alcohol and the further purification carried out as in the
final stages described above. The preparation is extremely tedious and the
yield very variable according to the history of the yeast employed.

The ash content of the four specimens was as follows. Dog glycogen,
0°38 per cent. Oyster, 0'°2 per cent. Rabbit, 0°51 per cent. Yeast, 0°87
per cent. The reaction of the two latter was faintly alkaline, while the dog
and oyster preparations both yielded slightly acid solutions. ;

PREPARATION OF ENZYME.

100 grmns. of pig’s pancreas were minced and well ground up with half
the weight of sand. Sufficient kieselguhr was added to form a fairly stiff
mass which was then pressed out in a hydraulic press. The extract obtained
in this way after filtration yielded a clear yellow-brown liquid which was
extremely active. It was usually diluted to five or ten times its volume
before use.

EXPERIMENTAL METHODS,

The course of the hydrolysis can be followed in several ways: (i) estima-
tions of the residual glycogen, (ii) determinations. of the optical activity,
and (ili) estimations of the reducing power of the solution after known
intervals of time. The first of these methods however is unsatisfactory
owing to the fact that some of the higher dextrins are also precipitated
by alcohol as well as the glycogen, for a precipitate is still obtained
in this way some time after the disappearance of the iodine reaction. The
optical activity of the solution also is due to a complex mixture of
substances, the nature of which has not been thoroughly worked out and
hence in the majority of the following experiments, the hydrolysis has been

—— oe Te ee Cae
’ al on

“<
‘

R. V. NORRIS 29

followed by estimations of the reducing power alone, though this method is
not altogether free from the same objection. he latter were carried out by
Bertrand’s method and the reducing power calculated as maltose, preliminary
experiments having shewn that under the conditions. employed very little if
any hydrolysis of the maltose into glucose took place. The reducing power
of a maltose solution of known strength was first determined and from the
results obtained, a curve constructed from which the maltose corresponding
to any found weight of copper could be read off.

The experiments were all carried out in flasks of Jena glass which were
used for no other purposes. In most cases toluene was added to the solutions
to prevent bacterial contamination.

HYDROLYSIS OF GLYCOGEN BY AN EXTRACT oF Pia’s PANCREAS.

To study the general course of the hydrolysis, 500 ce. of a 1 per cent.
solution of oyster glycogen were incubated with 2°5 ¢.c. of a pancreas extract
with the addition of 5 c.c. toluene as antiseptic. The glycogen and enzyme
solutions were each brought to the temperature of the bath (37°) before
mixing and then after known intervals of time, samples of the liquid were
removed and the following estimations made.

A. Reducing power.

B. Matter precipitable by alcohol (glycogen and higher dextrins). The
precipitation was carried out by adding two volumes of alcohol. At the same

‘time the reaction of the solution towards iodine was observed.

The results are collected in the following table.

TABLE I.
Hydrolysis of Glycogen by Pancreatic Diastase.

Wt. of ppt.
Duration thrown down by Iodine Percentage
Time of action 2 vols. alcohol Maltose reaction hydrolysis
12.30 0 4-702 grms. —_ Dark red brown 0
12.35 5 mins. — 0°223 grm. Red brown 4°45
12.45 Gs S49 is, 0°664 _,, * 13°32
1.00 30 ,, 7 et (eae 1:156 grms. = 23°2
1.15 45 ,, 2:30) Ss Val 5 Pale red brown 29°5
1.30 GoM, IESS 55 Oy (A seen Very faint 34-4
2.0 9055; PASI 5 2:003 ,, None 40°2
2.30 2 hrs. 1-20" x 2145 ,, — 43°0
3.30 oD; — 2°292 ,, — 46-0
4.30 455 0°65 grm, 2°362 ,, _ 474
5.30 Be iss — 2°434 ,, — 48°8
9.30 a Estimation 2-550 ,, -— 51°2

spoilt

30 R.. V. NORRIS

In Fig. 1 the results are shewn graphically, curve A representing the
degree of hydrolysis and curve B the residual glycogen and dextrins pre-
cipitable by two volumes of alcohol. It will be seen that the action
commences at a high initial velocity and proceeds for a very brief period in
a linear direction. By the time the iodine reaction has disappeared, the rate
has slowed down considerably and finally a long period of extremely slow
hydrolysis follows. It would appear that the rapid initial stage is chiefly
concerned with the conversion of the glycogen into dextrins with simultaneous

Percentage hydrolysis

Weight of alcohol precipitate in grms,

Duration of hydrolysis in hours

Higa

production of sugar. These dextrins, which are very resistant to further
hydrolysis, are then slowly broken down to sugar. It is obvious however
that in the above experiment this second stage would never have reached
completion and this seems to be the general rule unless very high enzyme
concentrations are employed. The iodine reaction disappeared when about
40 per cent. of the glycogen had been completely hydrolysed to maltose.
The action in short closely resembles the hydrolysis of starch but is slower
and less complete.

R. V. NORRIS 31

INFLUENCE OF TEMPERATURE.

A series of flasks each containing 50 c.c. of 1 per cent. glycogen were
incubated with 0:5 c.c. enzyme solution, at temperatures varying from 25° to
50°. Thirty minutes after the addition of the enzyme, the action was
stopped by adding 5 c.c. of N.KOH, the liquid was at once cooled and the
volume made up to 100 cc. The reducing power of 20 c.c. of this dilution
was then estimated. The results are shewn below in Table II and graphically
in Fig. 2. For the sake of comparison the results of a starch hydrolysis
with the same enzyme preparation are given also.

TABLE II.
Influence of Temperature.
A. Glycogen _ B. Starch
ia ay -

Temperature Total Percentage Total Percentage
of incubation maltose hydrolysis maltose hydrolysis
25°5° 01070 20:2 -- —
27°5° — — 071310 24°7
34° 01212 22°9 — --
37° 0°1375 25°9 071545 29°2
39°5° 0°1330 25°1 — —
42° 0°1310 24°7 0-1700 32°1
46° 01260 23°7 071895 35°7
50°5° 071190 22°5 01760 33°2

Percentage hydrolysis

25 30 35 40 "TET yo

Temperatures
Fig. 2.

32 R. V. NORRIS

Hence with the enzyme employed in these experiments, the optimum
temperature for glycogen hydrolysis was about 37°C. and this value was
unchanged by substituting different preparations of glycogen. For starch
hydrolysis on the other hand the optimum temperature found with the same
enzyme solution was 46°C. This result therefore lends some support to the
hypothesis that the two actions are brought about by two distinct enzymes.

INFLUENCE OF ENZYME CONCENTRATION.

Three dilutions of the enzyme solution were prepared and 0°5 c.c. of each
added to a series of flasks each containing 50 cc. of a 1 per cent. glycogen
solution (oyster)+1 c¢.c. toluene and previously brought up to 37°.
Altogether three flasks were used with each dilution. |

After 5, 15, and 25 minutes’ incubation one flask of each series was
removed from the bath and the action stopped by the addition of 5 ce.
N. alkali. The reducing power of each solution was then determined and
the results are shewn in Table III below.

TABLE III.

Influence of Enzyme Concentration.
Each flask contained 50 ¢.c. 1 per cent. glycogen+1c.c. toluene+0°5 c.c. enzyme.

Reduction after time given,

Dilution mgrms. Cu from 20 c.c. liquid
of enzyme = eee ee A eee
Exp. solution 5 mins. 15 mins. 25 mins.
A 1/10 7 21°5 32:0
B 1/20 3°35 10:0 —
Cc 1/40 A] 5:2 9:0

These results shew very clearly two things. (i) Under the above
conditions the rate of hydrolysis is directly proportional to the concentration
of enzyme, and (ii) the action is a linear one, the reduction after 15 minutes’
being almost exactly three times that after five minutes’ incubation and so
on. These results only obtain however when very dilute enzyme preparations
are used and during the initial part of the experiment—in other words while
there is a considerable excess of unchanged glycogen. Similar results have
been obtained by Evans [1912] in the hydrolysis of starch by saliva.

When stronger preparations of enzyme are employed the above relations
can only be observed for a very brief period of the hydrolysis, but another
point now becomes more noticeable. With dilute enzyme preparations, the
action comes to an end long before all the glycogen has been completely

ad, eh ee a

R. V. NORRIS 33

converted into maltose. This is apparent in the results quoted in Table I
where it will be noticed that after nine hours there was only a 51 per cent.
hydrolysis and the action had practically come to an end, When however
higher concentrations of enzyme are employed, the end point of the reaction
is moved very much nearer to complete hydrolysis as will be seen from the
experiments described below.

A series of flasks was incubated at 37°C. containing the following
solutions :

A. 100 cc. 1 per cent. glycogen solution+ 49 cc. H,O+01 ce.
enzyme + | c.c. toluene.

B. 100 cc. 1 per cent. glycogen solution+ 45 cc. H,O+0°5 ce.
enzyme + 1 c.c. toluene.

C. 100 cc. 1 per cent. glycogen solution+40 cc. H,O+10 ce.
enzyme + 1 c.c. toluene. 7

D. 100 cc. 1 per cent. glycogen solution+0 cc. H,O+50 ce.
enzyme + | c.c. toluene.

Percentage hydrolysis

0 1 2 3 4 5 6 7 8 9

Duration of hydrolysis in hours

Fig. 3.

‘The total volume in each case was the same, the only difference being
the concentration of enzyme. After 84 hours the action was almost over in
all four solutions and the amount of hydrolysis which had taken place was as
follows :

A, 45°6 per cent. B, 55:1 per cent. C, 60°7 per cent. D, 91°3 per cent.
The complete course of the hydrolysis in each case is shewn in Fig. 3.

Very similar results were obtained by Philoche [1908] with taka diastase
and it is not quite clear what interpretation is to be placed upon them, for, as

Bioch. vir 3

34 R. V. NORRIS

will be seen, there appears to be no direct relationship between the con-
centration of enzyme and the final degree of hydrolysis. The influence of
the products formed during the hydrolysis may be partially responsible but
judging from experiments described later, it is unlikely that this is the sole
cause. The incompleteness of the hydrolysis cannot be ascribed to the
destruction of the enzyme, for the latter after 20 hours’ incubation rapidly
attacked a second quantity of glycogen.

INFLUENCE OF GLYCOGEN CONCENTRATION.

Unless very low concentrations are used—less than 0°5 per cent. with the
strength of enzyme I have generally employed—the initial rate of hydrolysis
is practically unchanged by increasing the glycogen concentration. In the
weaker solution, however, the rate naturally begins to fall at an earlier period
owing to the glycogen being more rapidly exhausted. These points are
illustrated by the results shewn in Table IV.

TABLE IV.

Influence of Glycogen concentration.

. Reduction,
Time after megrms. Cu from 20 c.c. liquid
additionof Rae ee
enzyme A. 1°/,glycogen 8B. 2°/, glycogen

5 mins. 9:0 9:0 )

came Fae ong | Rates equal.

BLO san 45 52 Rate in A falling, B practically unchanged.
1 hr 66 89 - iain
pice 92 147 Both rates falling, but more rapidly in A,

20° 3, 108 188 Action at an end in both.

0/) hydrolysis 49-6 43-5: | a

INFLUENCE OF PRODUCTS.

As previously stated, unless high concentrations of enzyme are employed
the hydrolysis is not complete. In order to see whether this was due to the
influence of the products the following experiments were. carried out.

Influence of Maltose. In a preliminary experiment the action had stopped
when about 50 per cent. of the glycogen had been hydrolysed to maltose,
i.e. when the solution contained rather over 0°5 per cent. of maltose. This

R. V. NORRIS 35

amount of maltose was therefore added to one flask at the beginning of the
experiment and the rate of hydrolysis compared with one containing no sugar
at the beginning. The composition of the two solutions was as follows:
Flask A. 100 cc. 1 per cent: glycogen + 1 c.c, toluene +0°5 c.c. enzyme.
Flask B. 100 cc. 1 per cent. glycogen solution containing 0:5 grin.
maltose + 1 c.c. toluene + 0°5 c.c. enzyme.

TABLE V.

Influence of Maltose.

: A. Glycogen alone B. Glycogen + maltose
Time after pun Seeiene EEE SNS 5 CEE See
addition of Merms, Cu Merms. Cu Cu due to Cu due to
enzyme from 20 c.c. from 20 ¢.c, added maltose maltose formed
15 mins. 27 136 108 28
30)” ;, 46 154 zt 46
teh: 66 173 va 65
2 hrs. 80 188 - 80
Aas; 92 199 yy 91
Seay 100 207 “4 99
ZO" 55 108 212 5 104
°/, hydrolysis 49°6 47°6

It will be seen from these results that the rate of hydrolysis was
practically unchanged by the maltose added and the final degree of hydrolysis
lowered by only two per cent. The dextrins formed during the course of
the reaction, however, seemed to have some retarding influence, this being
illustrated by the following experiment.

Influence of Dextrin. 100 cc. of 2 per cent. glycogen were incubated
at 37° with 0° cc. enzyme until the solution no longer gave the iodine
reaction. The action was then stopped by heating the flask in a boiling
water-bath. The solution thus obtained was free from glycogen and in
addition to the dextrins contained about 09 per cent. of maltose. It is
referred to below as “ products of hydrolysis.” The following mixtures were
then made up and incubated at 37°.

A. 50 cc. 2 per cent. glycogen + 50 c.c. “ products of hydrolysis” + 1 c.c.
toluene + 0°5 ¢.c. enzyme.

B. 50 cc. 2 per cent. glycogen+ 50 cc. H,O+1 cc. toluene + 05 c.c.
enzyme.

Incubation was continued for six hours, the reducing power of the two
solutions being estimated from time to time. The results are collected in

Table VI.
3—2

36 R. V. NORRIS

TABLE VI.

Influence of Products of Hydrolysis.

B. Glycogen alone A. Glycogen + products
i a ee ae =
Time after Reduction, Reduction, Cu due to Cu due to
addition of mgrms. Cu from mgrms. Cu from sugar present maltose
enzyme 20 c.c. liquid 20 c.c. liquid at start formed
15 mins. 26 110 84 26
30 ,, 45 128 5 44
1 hr. 66 149 . 65
2 hrs. 81 159 3 75
a: 90 166 x 82
Ciss 97 170 sig 86
°/, hydrolysis 44°6 39°3

In this case although there was again no difference in the initial rates, after
one hour the hydrolysis in A began to lag behind that in B and after six
hours only 39:3 per cent. of the glycogen in A had been hydrolysed as against
44-6 per cent. in B. Moreover in A the action had almost stopped while in
B it was still proceeding at a fair rate. We have here then a marked
hindering action which must apparently be due to the dextrin-like substances
formed during the hydrolysis.

INFLUENCE OF ACIDITY OR ALKALINITY.

The ash almost always present in samples of glycogen consists to a large
extent of phosphates and according to the nature of these phosphates the
solution may have either an acid or alkaline reaction. Before comparing
different samples of glycogen it was therefore necessary to determine to what
extent this difference in reaction would influence the rate of hydrolysis. It
is obvious moreover that the optimum dose of acid will vary according to the
source and proportion of the enzyme used, for the latter as well as the
solution acted on may contain substances capable of neutralising some of the
acid. On the other hand the optimum hydrogen ion concentration is a
constant independent of these factors. In the latter case we obtain a
measurement of the true optimum reaction of the solution, while in the
former we have only an apparent value, the degree of accuracy of which will
depend on the amount and nature of the salts present in the solution. Thus
Sorensen [1909] working with three different preparations of invertin found
the optimal dose of sulphuric acid (apparent acidity) to be very different for
the three preparations but the optimum hydrogen ion concentration (true

R. V. NORRIS 37

acidity) was the same in each case. Hence in all the following experiments
the reaction of the solution has been determined by measurements of the
hydrogen ion concentration.

At first attempts were made to do this by means of the colorimetric
method described by Sérensen. The strong opalescence of the glycogen
solutions however rendered this method impracticable and hence the electrical
method has been used throughout. The apparatus employed was that of
Michaelis and Rona [1909] which is especially suitable since only small
quantities of solution are required. In this method, the electrodes are kept
in a still atmosphere of hydrogen and connection is made between the two
cells by means of a tape soaked in a solution of potassium chloride. By
making two determinations using different concentrations of KCl, e.g. 1:75 N.
and 35 N., the contact potential due to the salt can be calculated and
corrected for and, as equilibrium is established almost immediately, the
method is both rapid and convenient. The determinations were all made at
37°, the temperature at which the hydrolyses were carried out, the two
operations being conducted in the same thermostat. In each case the
standard solution against which the glycogen solutions were compared was
N/10 HCl and from the value of the E.M.F. thus set up and determined, the
hydrogen ion concentration was calculated in the usual way’.

The latter is denoted by the sign py, this being the logarithm of the
reciprocal value of the factor of normality. py being known, the hydroxyl
ion concentration can be calculated from the dissociation constant of water.
By graphical interpolation from a series of determinations by Lundén [1907],
the value at 37° has been taken as 2°45 x10-"=10-"%". Hence in a
N/100 solution of hydrogen ions where py=2 the hydroxyl ions will have
a concentration at 37° of 10" and so on. In the same way at absolute
neutrality the concentration of both ions will be L0~**.

DETERMINATION OF OPTIMUM REACTION.

To a series of flasks each containing 100 c.c. of a 1 per cent. glycogen
solution, were added varying amounts of N/100 NaOH or H,SO,. The total
volume of the solutions, however, was kept the same in every case. Each
flask was then in turn placed in the bath at 37° and when the solution had
acquired that temperature 0°5 c.c. of a dilute enzyme preparation was added.

1 A full account will be found in Sérensen’s paper [1909] where he also discusses the possible
sources of error in the method. :

38 R. V. NORRIS

After thorough mixing, a sample was removed and the hydrogen ion con-
centration at once determined. Thirty minutes after the addition of the
enzyme, a second sample was taken, the action stopped by the addition of
a small quantity of N. alkali and the reducing power estimated. In this way
the true reaction and the degree of hydrolysis was determined in each case.
This process was repeated with all four preparations of glycogen. In these
experiments no toluene or other antiseptic was used, Sdérensen having found
that these substances interfere somewhat in the hydrogen ion determinations,
equilibrium not being so readily established. As the duration of each
experiment, however, was only 30 minutes it is highly improbable that this
in any way influenced the results. These are collected in Table VII and also
shewn graphically in Fig. 4.

TABLE VII.

Influence of hydrogen ion concentration on the rate of hydrolysis.

C.c. of N/100 Reduction, °/, hydro-
Experi- Source of acid or E.M.F. mgrms. maltose lysis in
ment glycogen alkali added volts Pu in 20 c.c. 30 mins.
1 Dog liver 3-0 H,SO, 0°2055 4:43 7°37 3°7
2 o% 0)! 3 0°2306 4:84 33°5 - 16°8
3 53 Op5" 1, 0°2754 5°57 52°9 26°6
4 5 0 6 0:2940 5°87 59:0 29°6
5 AS 0:25 NaOH 0°3018 6:00 61:4 30°8
6 He 0:5) 255; 0°3217 6°32 60°3 30°3
7 a 1:0 oe 0°3410 6°63 54:0 27-1
8 + 3:0 . 0°3755 7:19 37°1 18°6
9 i, 6:0 % 0:4078 7:72 12-1 671
14 Oyster 05 H,SO, 0-2260 4-76 28-0 141
15 99 0 2 0-2730 5°53 40:4 20°3
16 As 0:5 NaOH 0:2870 5°76 44-2 22-2
17 - 1-0 3 0°3019 6:00 44-4 22:3
18 Xs 3:0 _ 0°3322 6°49 32°8 16°5
19 a 6:0 % 0:3980 7°56 10°7 5:4
22 Rabbit liver 6:0 H,SO, 0-2492 5°14 4:8 2-4
23 i 5-0 “ 0-2750 5°56 35°6 WS)
24 = AD 2, 0°2850 5°72 45-2 22°7
25 9 AD 5 0-3010 5:97 47-5 23-9
26 Ee 40 4, 0°3180 6°26 47:0 23°6
27 4 3:0 y 03412 6°64 42:3 21:3
28 ” 0 9 0°3818 7°30 24°9 12°5
29 Yeast 6:0 H,SO, 0:2748 5°55 25°5 12:8
30 %» BO 3 0-2940 5°87 39°5 - 19:8
31 - 3:0 5 03120 6°16 40:0 20°1
32 5: 15 a 0°3245 6°36 39°5 19°8
33 r 0 r 0-3580 6-91 27:1 13-6

R. V. NORRIS 39

The results shewn above indicate that the value obtained for the optimum
hydrogen ion concentration was practically the same with each preparation
of glycogen, that from oyster however giving a slightly lower figure than the
other three. The optimum reaction therefore for the enzyme I have used
was just on the acid side of the neutral point, the hydrogen ion concentration
corresponding to an acidity of about N/10° Alteration of the reaction in
either direction produced a considerable reduction in the rate of hydrolysis
especially in the case of the dog and oyster glycogens. Hence in comparing
the rate of hydrolysis of different glycogen preparations it is essential to see
that the reactions of the solutions are as nearly the same as possible, if the
results are to be at all reliable.

Percentage hydrolysis

4:5 5:0

Neutral Point
PH Hydrogen ion concentrations Acid <—— —~» Alkaline

Fig. 4.

COMPARISON OF GLYCOGENS.

‘1 per cent. solutions of the four kinds of glycogen were prepared and
sufficient N/100 acid or alkali added so that each solution after the addition
of the enzyme was as nearly as possible at the optimum reaction. The
amount to be added was estimated from the results obtained in the experi-
ments last described, the total volume was the same in each case. Immediately
after the addition of the enzyme, the hydrogen ion concentration was
determined to make sure that the reaction was really correct and then after
30 minutes’ hydrolysis, the action was stopped by the addition of alkali and
the reducing power estimated. Both the hydrolysis and the hydrogen ion
determination were carried out at 37° and no toluene was used for the
reason already stated. The results are shewn in Table VIII.

40 R. V. NORRIS

TABLE VIII.
Comparison of Glycogens. TI.

E. M.F. of solution Merms. maltose Percentage
Source of against N/10 HCl per 20 c.c. hydrolysis Relative
glycogen at 37° C. Pu after 30 mins. after 30 mins. rates
Dog 0:2970 volts 5:92 46°10 23°05 100
Rabbit 0:2986_,, 5°94 43°25 21-62 93°82
Oyster 0°2956_,, 5:90 41:0 20:5 88°9
Yeast 03014 ,, 599 39°5 19°7 85°7

COMPOSITION OF SOLUTIONS.

A. Dog glycogen. 100 cc. 1 per cent. glycogen+0'25 cc. N/100
NaOH +45 ec. H,O + 0°5 c.c. enzyme.

B. Oyster. 100c.c. 1 per cent. glycogen + 1:0 cc. N/100 NaOH + 40 cc.
H,0 + 0°5 cc. enzyme.

C. Rabbit. 100 cc. 1 percent. glycogen + 45 c.c. N/100 H,SO,+ 055 ec,
H,O0 + 0°5 cc. enzyme.
D. Yeast. 100 cc. 1 per cent. glycogen+425 cc. N/100 H,SO,

+0°75 cc. HO +0°5 cc. enzyme.

There is thus quite a marked difference in the rate of hydrolysis of the
four preparations of glycogen examined, the least difference being between
those from oyster and yeast. A second experiment gave very similar results

which are shewn in Table IX.
TABGE Xe

Comparison of Glycogens. II.

E.M.F. of solution Megrms. maltose

Source of against N/10 HCl per 20 c.c. Percentage Relative

glycogen at 37° C. Pu after 30 mins. hydrolysis rates
Dog 0:3046 volts 6:0 48°9 24°45 100
Rabbit 0:2950 ,, 59 46:1 23°05 94°3
Oyster 0:2901 ,, 58 42-4 21°2 S6eiee
Yeast 03112 ,, 61 40°45 20:2 83 +

* Solution rather too acid. + Solution rather too alkaline.

OPALESCENCE OF GLYCOGEN SOLUTIONS.

A 0:2 per cent. solution was made up of each of the four preparations of
glycogen and the opalescence of these compared. This was carried out in
Nessler glasses, the volume of solution being adjusted until the opalescence
appeared to be the same in each case. The comparison is not easy and the

experimental error is therefore large but the solution of rabbit glycogen was

_—

R. V. NORRIS 4]

markedly the most opalescent of the four examined, oyster and dog glycogen
being about equal and yeast less opalescent than any. The actual results
obtained were as follows:

TABLE X,

Volumes of 0:2 solutions Relative
Glycogen having equal opalescence opalescence
Rabbit 11 c.c. 4°5
Dog 30 ,, Lai
Oyster 33 Cy, 1°5
Yeast 50); 1

COLOURATION WITH IODINE.

To 10 cc. of a 0'2 per cent. solution of each of the glycogen preparations
was added 1 cc. of a 1 per cent. solution of iodine.” The volume in each
case was then made up to 50 c.c. and the colours compared in Nessler glasses
with the following results :

TABLE XI.
Volumes of solutions giving Relative strength
Glycogen equal depth of colour of colouration
Rabbit 10°5 ¢.c. 4°7
Dog OMe 2°6
Yeast Do) ss 1-5
Oyster DO 5 1

- Thus both as regards opalescence and the colouration with iodine there is
a distinct difference between the four glycogen solutions.

SUMMARY AND CONCLUSIONS.

The conditions under which glycogen is hydrolysed by a pancreatic
extract (pig) have been studied with the following results:

1. The hydrolysis at first proceeds with great rapidity, the glycogen
being quickly converted into dextrins with simultaneous production of
maltose. The further breaking down of these dextrins takes place with
extreme slowness and is as a rule incomplete.

2. The optimum temperature for glycogen hydrolysis of the enzyme
employed in these experiments was 37°. Below this there is a rapid fall
in the rate, between 37° and 50° this is less marked. With starch the
optimum temperature for the same enzyme preparation was 46°,

3. While excess of glycogen is present, the rate of change is directly
proportional to the enzyme concentration and the action is a linear one.

49 R. V. NORRIS

4, With low enzyme concentrations, the action is by no means complete ;
the degree of hydrolysis, however, rises with increasing concentration of
enzyme though not in the same ratio.

5. The concentration of the glycogen solution has little influence on the
initial rate of hydrolysis unless very low concentrations (less than 0°5 per cent.
in my experiments) are employed.

6. The action is hindered to a small extent by the products of hydrolysis,
the dextrins probably being the interfering substances.

7. The action is favoured by the presence of small traces of acid, the
optimum reaction determined with several preparations of glycogen being
a hydrogen ion concentration of 10~* (at neutrality this value is LO-** atioi oe

8. Samples of glycogen prepared from (i) dog, (1) rabbit, (111) oyster,
and (iv) yeast are hydrolysed at different rates when compared at the
optimum hydrogen ion concentration. Taking dog glycogen as 100, the
relative rates of hydrolysis found are, rabbit 94, oyster 88, and yeast 84.

9, The degree of opalescence and the colouration with iodine exhibited
by glycogen solutions vary with the source from which the glycogen has been
obtained.

10. In view of the above points of difference it is probable that the four
preparations of glycogen examined are not identical in constitution, though it
is possible these differences might be caused by variations in the colloidal
state of the four glycogen solutions. If, however, the glycogens are distinct,
the diastatic enzymes obtained from different animals should also be specific
and experiments are in progress to test this point.

My thanks are due to Professor Arthur Harden, F.R.S. and to Dr H.
MacLean for much useful advice during the course of this research.

REFERENCES.

Bertrand (1906), Bull. Soc. Chim. 35, 1285.

Clautriau (1895), ‘Etude chemique du glycogéne,’ Mem. Couronnes, Acad, Roy. Belg. 53.
Cremer (1894), Miinchen. Med. Wochensch. 41, 525.
Evans (1912), J. Physiol. 44, 191.

Ford (1904), J. Soc. Chem. Ind. 23, 414.

Harden and Young (1902), J. Chem. Soc. 81, 1225.
Harden and Young (1912), J. Chem. Soc. 101, 1928.
Lundén (1907), J. Chim. Phys. 5. 574.

Michaelis and Rona (1909), Biochem. Zeitsch. 18, 318.
Pfliiger (1905), ‘‘ Das Glykogen,’ Bonn.

Philoche (1908), J. Chim. Phys. 6, 212, 355.
Sérensen (1909-1910), Compt. rend. Carlsberg, 8, 1. rt

IV. THE METABOLISM OF ORGANIC PHOSPHORUS
COMPOUNDS. THEIR HYDROLYSIS BY THE
ACTION OF ENZYMES.

By ROBERT HENRY ADERS PLIMMER.

From the Ludwig Mond Research Laboratory for Biological Chemistry,
Institute of Physiology, University College, London.

(Recewed November 23, 1912.)

INTRODUCTION.

Our knowledge concerning the composition of the organic phosphorus
compounds occurring in both plants and animals is still very rudimentary.
Five groups of organic phosphorus compounds occur in nature :—

1. Phospholipines, of which lecithin is the principal example, and in
which the phosphorus is regarded as being in ester combination with glycerol
as glycerophosphoric acid ;

2. Phytin, in which phosphoric acid is in combination with inositol as
inositol-phosphoric acid, the compound itself being regarded as the calcium
magnesium salt ;

3. Hexosephosphate, in which a hexose is combined with phosphoric acid ;

4, Nucleic acid, in which again a carbohydrate (hexose or pentose) is
combined with phosphoric acid, and attached to the carbohydrate is a purine
base ;

5. Phosphoprotein, in which phosphoric acid is in an unknown form of
combination, presumably with one of the amino-acids composing the protein
molecule.

The phospholipines occur chiefly in animals and only to a small extent in
plants ; phytin or phytic acid has only been found in plants ; hexosephosphate
is a product formed in alcoholic fermentation by yeast ; nucleic acid is said to
be the chief constituent of the cell nuclei in both plants and animals ;
phosphoprotein is present only in the animal kingdom and forms the chief
protein foodstuff of embryo birds and suckling mammals.

Considering how little is known as to the chemical composition of these

44 R. H. A. PLIMMER

compounds it is not to be wondered at that the knowledge of their metabolism
is vague. Most investigators will agree that plants can synthesise organic
phosphorus compounds from inorganic phosphates and that they are also able
to carry out the reverse process, but how these changes are effected, or by
what enzymes they are caused, is practically unknown.

Most of the experiments on phosphorus metabolism in animals have
been feeding experiments in which animals have been fed with organic and
inorganic phosphorus compounds and the total intake and output of
phosphorus estimated, together with that of nitrogen in most cases. The
conclusions which were drawn from the general condition of the animal at
the end of the time, and from the effect of the phosphorus compounds
upon the nitrogen metabolism, were summarised by F. C. Cook in 1909.
They led to the belief that organic phosphorus compounds are directly
assimilated and are preferable to inorganic phosphates as the source of
phosphorus for animals. Against this belief the recent work of Fingerling
[1912; 1, 2] shows that the animal organism can synthesise organic
phosphorus compounds from inorganic phosphates. He fed ducks for long
periods on food containing only organic phosphorus, or only inorganic
phosphorus, as the source of phosphorus. Both sets of ducks laid approxi-
mately the same number of eggs and the lecithin and nuclein content of the
eggs was about the same in either case. He also analysed the milk of goats
which were fed with organic and inorganic phosphorus in the same way as he
fed the ducks. In both cases the composition of the milk was the same.

Gregersen [1911] has also found that the organic phosphorus compounds can’

be built up in the animal body from inorganic phosphate, thus confirming
the work of Fingerling.

Both organic and inorganic phosphorus compounds can therefore be
utilised for the formation of the organic phosphorus compounds in animals ;
the feeding experiments with organic phosphorus compounds do not show in
what form the phosphorus is assimilated, ie. whether it is absorbed in an
organic form or whether it is first hydrolysed and then resynthesised.
Phosphorus is excreted in the urine of animals almost entirely as inorganic
P.O;, whether the intake be organic or inorganic; only a small quantity
is excreted in an organic form [Mathison, 1909; Plimmer, Dick and
Lieb, 1909; and Gregersen, 1907], and the composition of this organic
phosphorus compound is unknown although it is stated to be glycerophos-
phoric acid. The phosphorus in the faeces is also mainly inorganic P,O,.

ai : a
These facts point to the hydrolysis or decomposition of organic phosphorus
compounds in the animal organism.

R. H. A. PLIMMER 45

It seemed to me that the only possible means by which the metabolism
and assimilation of the natural organic phosphorus compounds could be
studied was to ascertain if these compounds were hydrolysed by the enzymes
of the digestive tract. The action of the enzymes of the pancreas, intestine
and liver upon the various organic phosphorus compounds has therefore been
investigated and in addition the action of some plant enzymes has been
included so as to obtain a complete and comparative survey of the whole
series of organic phosphorus compounds.

I. GLYCEROPHOSPHORIC ACID.

The literature on phosphorus metabolism in plants contains no statement
as to the hydrolysis or synthesis of glycerophosphoric acid or of lecithin, and
until quite recently the knowledge of the metabolism of these substances in
animals was equally indefinite. The experiments made by Marfori [1905],
Biilow [1894], Mathison [1909], Plimmer, Dick and Lieb [1909] and others

had shown that glycerophosphoric acid was completely absorbed when

introduced into the body either by the mouth or subcutaneously, and that

the phosphorus appeared in the urine _as inorganic phosphate and not as

lycerophosphoric _aci At the time when these investigations

made’ no one had tried if the organs of the animal body were able to

were

hydrolyse glycerophosphoric acid and convert it into glycerol and phosphoric
acid. It was generally assumed that glycerophosphoric acid was either
assimilated as such, or that, since it is an ester, it was hydrolysed like the
fats by the lipase in the liver, pancreas and other organs. In 1912 Grosser
and Husler [1912] first published experiments which proved that an enzyme
(glycerophosphatase) able to hydrolyse glycerophosphorie acid was present
in certain animal tissues. They found that the kidney and intestine could
completely hydrolyse glycerophosphoric acid, that the lung hydrolysed 62 per
cent. and the liver 16 per cent. The spleen, blood and muscle did not
hydrolyse glycerophosphoric acid. -

The fact that glycerophosphatase is present in the intestine, kidney and
lung but not in the pancreas and liver suggests that the hydrolysis is not
effected by lipase since the pancreas and liver are the chief animal tissues
which contain lipase, whereas the other organs mentioned only contain small
amounts of this enzyme.

An enzyme which can hydrolyse glycerophosphoric acid is also present in
plants, especially in castor oil seeds. Like the lipase in these seeds it is only

1 The results of my investigations upon the metabolism of glycerophosphoric acid were

communicated to the Biochemical Club at their meeting in July 1911.

>

46 R. H. A. PLIMMER

liberated from a precursor by the action of dilute acids, but unlike the
lipase it is soluble in water [Connstein, Hoyer and Wartenberg, 1902;
Armstrong and Ormerod, 1906; Tanaka, 1910]. The action of the two
enzymes is thus distinct although it has been impossible to prove this
absolutely, as after repeated extraction with water the final extract always
contained some glycerophosphatase.

Extracts of yeast and of bran, which contain the enzymes hexosephos-
phatase [Harden and Young, 1908] and phytase [Suzuki and Takaishi,
1907] respectively, are also able to hydrolyse glycerophosphoric acid.

EXPERIMENTAL.

The general method for determining the action of extracts of various
tissues upon glycerophosphoric acid has been to mix a known volume of the
extract with either a known volume of a solution of glycerophosphoric acid
previously neutralised with soda, or with a solution of sodium glycerophosphate;
to withdraw a sample of known volume immediately after mixing and after
subsequent intervals of time ; and to estimate, after filtering if necessary, the
inorganic phosphate in these samples by precipitation with ammonium
magnesium citrate in the presence of ammonia, and conversion into magnesium
pyrophosphate. By the use of magnesium citrate instead of magnesia
mixture the precipitation of organic matter with the ammonium magnesium
phosphate is avoided, as was shown by Plimmer and Bayliss in 1906.
The total phosphorus content of the mixture was estimated in another
portion by Neumann’s method as modified by Plimmer and Bayliss [1906].
All results were then calculated for the same volume. In some cases a
control experiment consisting of the same volumes of water and of extract
was carried out simultaneously. Toluene was added as an antiseptic and the
mixtures were kept at 37° C. after the removal of the first sample.

The commercial preparations of glycerophosphoric acid and of sodium
glycerophosphate, which are synthetically prepared, were employed. The
synthetical compound except for its optical activity is apparently identical
with the natural compound obtained from lecithin. Both products seem to
be mixtures [Willstitter and Liidecke, 1904; Power and Tutin, 1905;
Tutin and Hann, 1906].

It was originally supposed that the glycerophosphorie acid, since it is an
ester, would be hydrolysed by the lipase of the pancreas and other tissues:
for this reason the preparations and extracts of tissues used have been
purposely made by the usual methods for obtaining an active lipase.

R. H. A, PLIMMER 47

The action of the pancreas was investigated with :

1. The residue from an aqueous extract of trypsin (Fairchild); such
residues have been shown by Dietz [1907] to be very strongly lipoclastic in
their action. The residue wag suspended in a solution of sodium glycero-
phosphate and the inorganic P.O; was estimated after filtering off and
washing the precipitate.

2. A glycerol extract of pig’s pancreas prepared by extracting the organ
for 24-36 hours with ten times its bulk of a mixture consisting of 90 parts
of glycerol and 10 parts of 1 per cent. sodium carbonate, and then straining
through muslin. In one experiment this extract was used in aqueous
solution, in another in about 50 per cent. glycerol solution. The mixture
containing the extract of sodium glycerophosphate was then made just
alkaline to phenolphthalein and maintained at 37° C. in the presence of
toluene. The samples from the glycerol solution were diluted with water
and filtered, and inorganic P.O; was estimated in the filtrate and washings.

3. Fresh pancreatic juice which is known to contain an active lipase.
The pancreatic juice in one experiment was kindly collected for me by
Prof. Starling, in another experiment by Prof. Bayliss; in the first experiment
the juice was activated by enterokinase by adding a small amount of
intestinal extract ; in the second experiment the juice was not activated.

The action of the liver was investigated with :

1, An extract of dog’s liver made by the method of Loevenhart and
Peirce [1906] for obtaining an active lipase.

2. An extract of calf’s liver prepared according to the method of
McCollum and Hart [1908] in their study of phytase in animal tissues. The
samples removed at intervals were acidified with a drop of glacial acetic acid,
and were boiled and filtered. The coagulum was thoroughly washed and
inorganic P.O; was estimated in the filtrate and washings.

The action of the small intestine was investigated by extracts made from
the mucous membrane, which was ground up with sand and extracted for
24-48 hours with water containing toluene as antiseptic ; and then strained
through lint or muslin. In the earlier experiments this extract which was
milky in appearance was used directly ; the samples, which were removed at
intervals, were acidified with acetic acid; the precipitate so formed was
filtered off and thoroughly washed and the inorganic P,O, estimated in the
clear filtrate and washings. In the later experiments the aqueous extract of
the mucous membrane was acidified with acetic acid, and filtered from the
precipitate: the clear extract so obtained was found to be active, either when
still acid, or after neutralising with sodium carbonate. These clear extracts

48 R. H. A. PLIMMER

were used in all the later experiments with the other organic phosphorus
compounds.

The action of the kidney and liver was investigated with extracts made
by grinding the organ with sand, and extracting the mass with water, to
which toluene was added, for 1-2 days. In the first experiment with the
kidney the extract was acidified with acetic acid, the filtrate neutralised with
sodium carbonate and used after again filtering. In the other experiments ~
the milky extract was used ; each sample was acidified and diluted with the
same volumes of acetic acid and water. Inorganic phosphate was estimated
in a measured volume of the clear filtrate.

A suspension of shelled castor oil seeds in dilute acetic acid was used for
the detection of a glycerophosphatase, samples being removed and the
inorganic P.O, estimated in the filtrate and washings. A suspension was also
employed when the seed was treated with acid, washed and the extracted
residue tested for glycerophosphatase.

Extracts of castor oil seeds were prepared by treating the shelled seeds
ground as finely as possible, sometimes after the fat had been previously
removed by extraction with ether, with water or with decinormal acetic acid
for 1-2 days in the proportion of 5-20 grams of seed and 100-300 c.c. of the
acetic acid. A clear filtrate was obtained on filtration. These filtrates were
used directly and in some experiments after neutralisation of the acid extract
with sodium carbonate.

Preliminary experiments to ascertain if a glycerophosphatase were present
in yeast were made with an aqueous extract of pressed yeast prepared in the
presence of toluene and by a suspension of zymin; in the latter case the
estimations of inorganic P,O; were made in the filtrate and washings from the
insoluble matter.

The later and all subsequent experiments with yeast enzymes were made
with extracts of the commercial “zymin” of Schroder and Co. of Munich.
These extracts were prepared by treating the powder with ten times its
weight of water for 24-48 hours in the presence of toluene and then filtering.
A pale brown solution resulted.

By simply covering wheat bran with five times its weight of water in the
presence of toluene a clear and active extract was always obtained on filtering.
The bran extract if kept without contact with air was of a pale brown colour,

but when it was exposed to the air and filtered it became rapidly darker in
colour,

-_——_

;

R. H. A. PLIMMER 49

The following are the analytical data:

Pancreas.

(1) Residue from 1°5 gm. |
trypsin (Fairchild) suspended

(2) 100 c¢.c. glycerol extract (3) 100 c.c. glycerol extract
_ of pig’s pancreas + 5gm.sodium | of pig’s pancreas + 5 gm. sodium

in 500 c.c. water containing | glycerophosphate in water. Mix- glycerophosphate dissolved in
5 gm. sodium glycerophos- | ture made alkaline to phenol- | 200c.c. glycerol + 200 c.c. water
phate. phthalein and volume diluted and mixture made alkaline to

with water to 500 c.é.

P,O, in gm.
At commencement 0:0121
After 31 days 0-0115

Total 0-1712

After 9 days

(4) 20 c¢.c. glycerophosphoric acid diluted |

to 500 ¢.c. with water and neutralised with
caustic soda+20 ¢.c. pancreatic juice of dog
+4-5 drops intestinal juice. Total volume
made up to 1000 e.c.

P.O, in gm.
At commence-
ment ... 0°0126 After 8days 0:0759
After 1 day ... 0°0517 a5 UD) 5, OES
» 2days 0-0569 », 138 ,, 0:0824
ip a) ae 0:0633 5 20' ,, 00891
» OF 5 0:0723 Rc ON ee Ost

Total 0:2606

Liver.

(1) A. 100 ¢.c. extract of dog’s liver by
Loevenhart and Peirce’s method + 400 c.c. H,O.
B. 100 c¢.c. same extract +400 c.c. water con-
taining 5 gm. sodium glycerophosphate.

P,O; in gm.
A B
At commencement 0:0006 0-0147
After 1 day Se 0°0013 0:0165
IeD GAYS ...; -- 0:0157
-j OME 2 -— 0°0122
Total 0-0177 0-1889

Bioch, vu

P.O, in gm.

At commencement 0:0128

Total 0°1775

After 5days... 0°0339

_phenolphthalein.

P.O, in gm.

At commencement 0:0120

00135 | After 8 days 0-0126
Total 0-1712

(5) 2°5 gm. sodium glycerophosphate in
water +50 c.c. pancreatic juice of dog. Total

| volume made up to 500 c.e.

P,O, in gm.
At commencement 0:0135
After 14 days 0:0124

Total 01661

(2) A. 125 cc. extract of calf’s liver by
McCollum and Hart’s method + 125 c.c. water.
B. 125 ¢.c. same extract+125 c.c. water con-
taining 2°5 gm. sodium glycerophosphate.

P,O; in gm.
A B Difference

At commence-
ment a.» 0°0232 0-0529 0:0297
0:0679 0:0340

Total 0:-0659 0°2460 0°1801

50 R. H. A. PLIMMER

Small Intestine.

(1) A. 100c.c. milky ex- (2) 80c.c. clear acid extract | (3) 80c.c. clear acid extract
tract of dog’s intestine + 400c.c. of cat’s intestine + 5 gm.sodium | of intestines of 2cats neutralised
water. B. 100c.c.sameextract glycerophosphate in water made | with soda + 2°5 gm. sodium gly-
+400 c.c. water containing 5 up to 500 c.c. with water. _ cerophosphate in water made up
gm. sodium glycerophosphate. | PO, iene to 250 c.c. with water.

P.O; in gm. Atcommencement 0-0256 | PO; in gm.
A B Difference After 17 hours ... 0°0313 | Atcommencement 0:0236
At commencement : Te er .. 0°0346 | After 1 day ae O08
0:0042 §=©0°0444 00402 | » Sdays ... 0:0680 | » Adays 2. sbi
After 1 day: ” 12 ” none 0:0963 | ” 9 ” eet 0:1270
0-0062 0-1506 0:1444 Total 0:1458 | Total 0:1496

After 2 days:

00061 0°1539 0°1478
After 6 days:

0:0061 0°1539 0:1478
Total:

0-:0080 0:1864 0:1784

Kidney.

(1) A.100c.c.clearacidex-| (2) A. 200 c.c. extract of (3) A. 200 ¢.c, extract of dog’s
tract of dog’s kidneys neutral- sheep’s kidneys+50c.c. water. | kidneys+50c.c. water. B. 200
ised with soda+150c.c. water. | B. 200 ¢.c. same extract +50 | c.c. same extract +50 ¢.c, water
B. 100 c.c. same neutralised c.c, water containing 2°5 gm. | containing 2°5 gm. sodium gly-

extract-+150 ¢c.c. water con- | sodium glycerophosphate. | cerophosphate.

AGE ) : oat
rack ile SCENES Uh [EE | PO; in gm. P.O; in gm.

A B A B
P.O; in gm. _ At commencement : At commencement :
A B . Difference, 0:0256 0°1300 0:0071 0:0176

At commencement : | After 1 day : ; | After 1 day:

0:0080 00090 0:0010 | 0:0321 0:2085 0:0084 01204
After 3 days: | After 4 days: After 5 days:

00075 0:0763 0:0688 | 0:0332 0°2092 00092 0°1723
After 10 days: Total: | Total :

ee RIB e pales 0-0406 02130 | 0-024 0°1750

Total :
0:0076 0:1661 071585 |
Lung.

A. 200 c.c. extract of dog’s lungs+50 c.c. water.- B. 200 c.c. extract of dog’s lungs
+90 ¢.¢c. water containing 2°5 gm. sodium glycerophosphate.

P.O, in gm.
A B Difference
At commencement 0:0069 0:0148 0:0079
After 1 day 58 0°:0110 00344 0°0234
» 2 days aP 0:0128 0:0472 0°0344
4s, a3 0°:0161 0:0730 0:0569

Total 0-0241 02282 0:2041

oO eee

Castor oil seeds.

(a) 5 gm. ground seeds
250 C.c.

10
sodium glycerophosphate.

P.O, in gm.

acetic acid containing 2°5 gm.
>

-R. H. A. PLIMMER

suspended in (b) 60 c.c. aqueous extract of ground seeds

+200 c.c. water containing 2°5 gm. sodium
glycerophosphate.

P,O, in gm.

At commencement 0-0121
At commencement 0-0089 After 12 days 0:0105
After 1 day 0°1199
», 2 days 0°1395 Total
” 3 ” 0°1538
” 8 ” 0°1852
Total 0°2308
(ec) 130 c.c. id acetic acid extract +120 c.c. (d) 100 c.c. is acetic acid extract of ground
water containing 2°5 gm. sodium glycerophos- | seeds +150c.c. water containing 2-5 gm. sodium
phate. glycerophosphate.
P.O, in gm. P,O, in gm.
At commencement 0-0217 At commencement 00356
After 1 day 0-0888 After 1 day 00920
»» 2 days 0-1027 », 2 days 0-1008
” 3 ” 0:1109 ” 7 ” 0-1179
” 6 ” 0:1204 - ” 12 ” 0:1263
Total 0°2143 Total 0:2003
50 c.c. extract contained—
At commencement 00477 gm. inorg. P20,
After 12 days 0:0504 ,, 5
Total P,O,; 0:0520 gm.
Increases are thus due entirely to hydro-
lysis of the glycerophosphate.
N eet wea | (f) Residue of seeds in| (g) Residue of seeds in expt.

(e) 200 c.c. 10

extract of ground seeds freed
from fat by extraction with
ether and kept for several
months so as to convert all
organic phosphorus into inor-
ganic phosphate + 50 c.c. water
containing 2°5 gm. sodium gly-
cerophosphate.

P.O, in gm.
At commencement 0:0410
After 1 day 0:0579
2 days 0-0711
00806
01030

”

”

Total 0-2003

' expt. (c) thoroughly washed by (e) after thorough washing was

decantation and filtration suspended in 250 c.c. of 1 per
suspended in 250 c.c. water cent. solution of sodium glycero-
containing 2°5 gm. sodium phosphate in water.

glycerophosphate. Bigs teres
P20; in gm. At commencement 0-0005
Atcommencement 0:0165 After 2 days 0:0259
After 3 days 0:0651 eee Se re 0:0338
er GS t:,: 0:0889 alti ek 00522
A PAUL ee 0-1011

Total 0°1500
Total 0°1826

52 R. H. A. PLIMMER
Yeast.
(a) 80 c.c, extract of yeast + 170 c.c. water (vb) A. 10 gm. zymin suspended in 250 c.c.
containing 2:5 gm. sodium glycerophosphate. water. B. 10gm.zymin suspended in 250 c.c.
p water containing 2°5 gm. sodium glycerophos-
P,O, in gm. phate.
At commencement 0:0396 ‘Se toa
After 1 day Ho 0:0776 : :
3 days 0:0869 A B Difference
s 0-0952 At commence-
a Se ees : ment ... 00215 0-0847 00632
Total 0-2244 After 1 day... 0°0527 01323 0:0796
», 2 days 0:0582 01424 0:0842
50 c.c. yeast extract contained— tae 0-0632 0:1539 0-0907
Did Os Vent Laon 8 Total 00850 0:2637 «00-1787
0-0930 gm. total P,O;
Increases are therefore due to hydrolysis of
glycerophosphate.

(c) A. 100 c.c. extract of zymin+150c.c. water. B. 100 ¢.c. extract of zymin +150 c.c.
water containing 2°5 gm. sodium glycerophosphate.

PO; in gm.
A B Difference
At commencement 0:0304 0:0774 0:0470
After 1 day “ise 0:0341 0:0907 0°0566
», 8 days “of 0:0379 0:1068 0:0689
Total 0:0621 0-2409 0°1788
Bran.
(1) 200 c.c. extract of bran+50 c.c. water (2) A. 125c.c.extract of bran + 125 c.c. water.
containing 2°5 gm. sodium glycerophosphate. B. 125 ¢.c. extract of bran+125 c.c. water
: containing 2°5 gm. sodium glycerophosphate.
P,O; in gm.
At commencement 0-1430 P.O; in gm.
After 1 day eee 0°1696 A B Difference
», 2days et 0-1839 At commence-
ae ee 0-1877 ment sie 0:0374 00389 0°0015
After 2 days 0:0389 00731 00342
Total 0-2904 Uf Chae 0:0389 0:0852 00463
ope Doe 0:0393 01014 0:0621
Total 0:0406 01965 0:1559

The gradual increase in the amount of inorganic P,O;, which occurred
when the extracts of intestinal mucosa, kidney, castor oil seeds, yeast and
bran were tested, is the best evidence that the hydrolysis of glycerophosphoric
acid has been effected by an enzyme present in the tissues. The enzyme is
not present in the pancreas and liver of animals. Though an increase in the
amount of inorganic phosphate was found in the experiment with pancreatic
juice activated by 4-5 drops of intestinal juice, the increase takes place very
slowly ; one-fifth is hydrolysed in 2 to 3 days, and after 36 days the hydrolysis
is less than one-half of the total P.O, in the solution. An explanation of

R. H. A. PLIMMER 53

this increase is given by the experiments with intestinal extract, which contains
the active enzyme; a small amount of the enzyme had therefore been added
when activating the pancreatic juice. The absence of the enzyme from the
extracts of pancreas and liver, which were prepared specially for lipase, and
the presence of the enzyme in the extracts of the intestine and kidney, tissues
not usually regarded as good sources of lipase, point to this enzyme being
distinct from lipase. Proof of the difference of the enzymes is given by the
experiment with castor oil seeds; lipase is insoluble in water but an aqueous
extract contains glycerophosphatase. Its presence in bran, in which lipase
has not yet been shown, is further proof that the two enzymes are not
identical. A comparison of the actual amounts of the enzymes present in
the several tissues is not possible, but it may be noticed that the amount of
glycerophosphatase in the lung is very much less than in the intestine and
kidney. The amount in yeast is also small and the amount contained in bran
is smaller still.

II. PHytvic AcIp.

Phytin, the calcium magnesium salt of inositolphosphoric acid, or phytic
acid, was first isolated by Posternak in 1903 from various seeds in which
it is present as a reserve material containing phosphorus in organic com-
bination. It also occurs in the bran of cereals [Hart and Andrews, 1903],
that of rice being particularly rich in it [Suzuki and Yoshimura, 1907}.
Phytin is thus the most important phosphorus-containing foodstuff for
herbivora, and it has been administered to man to supply organically com-
bined phosphorus. On this account more work has been done upon its
metabolism than upon that of any of the other organic phosphorus compounds.

It has been shown by Scofone [1905], Giascosa [1905], Mendel and
Underhill [1906], and also Horner [1907] that phytin, when given to animals
by the mouth, is absorbed and that the phosphorus is excreted as inorganic
phosphorus in the urine; but the absorption is never complete.

Suzuki and Takaishi [1907] have found that bran contains an enzyme,
phytase, which completely hydrolyses phytin ito inositol and phosphoric
acid; Jordan, Hart and Patten [1906] have shown that the enzymes, pepsin
and trypsin, of the animal body have no action on phytin, but according to
McCollum and Hart [1908], the liver and also the blood contain a phytin-
splitting enzyme.

The hydrolysis of glycerophosphorie acid by the enzymes of the intestinal
mucosa suggested that phytic acid might also be split up by this tissue.
Experiment, however, showed the contrary. Neither the mucous membrane

54 R. H. A. PLIMMER

of the intestine of the carnivora, dog and cat, nor that of the herbivora, sheep,
ox and rabbit, even when the latter animal was fed for some weeks on bran,
contained an enzyme which was able to hydrolyse phytic acid. Phytic acid
thus behaved very differently from glycerophosphoric acid. The difference
in the behaviour of the two compounds does not extend to the enzymes of the
liver. The hydrolysis of phytic acid by the liver observed by McCollum and
Hart was very small and it has been impossible to regard this as a hydrolysis
when compared with other results. Phytic acid is thus not hydrolysed by
the enzymes of the digestive tract of animals.

The enzymes in an extract of zymin have no action upon phytin, but it is
slowly hydrolysed by an acid extract of castor oil seeds. The only enzyme
which readily hydrolyses phytin is contained in an extract of bran.

EXPERIMENTAL.

Commercial phytin was employed as the source of phytic acid in these
experiments. This substance except for a small residue is easily soluble in
water, to which it gives a distinctly acid reaction to litmus, and from which
it is precipitated on rendering the solution alkaline with ammonia. It
contains a small quantity of inorganic phosphate. The calcium was pre-
cipitated from the aqueous solution of phytin by adding the calculated
quantity of oxalic acid, or of sodium or potassium oxalate. The solution of
the free acid, or of the sodium or potassium salt, obtained on filtermg was
used as substrate in the experiments.

Extracts of the various tissues were prepared as described under glycero-
phosphoric acid; the mixture of enzyme solution and phytate was kept at
37°C. in the presence of toluene. Samples were removed at intervals and
the inorganic P,O; estimated.

In the first experiments the estimation of inorganic phosphate was carried
out by precipitation with ammonium magnesium citrate in the presence of
ammonia, but it was soon observed that the inorganic phosphate in the
presence of phytic acid was not immediately precipitated but only came
down slowly after standing for some time, and that the filtrate frequently
gave a further precipitate on standing. The precipitate did not have the
usual appearance of ammonium magnesium phosphate and it adhered very
tenaciously to the sides of the vessel in which the precipitation took place.
On converting these precipitates by heating into the pyrophosphate the
weights obtained were also very variable. An examination into the cause of
these difficulties revealed the fact that phytic acid inhibits the precipitation

gt Sl ads.

R. H. A. PLIMMER 55

of inorganic phosphate as ammonium magnesium phosphate and that under
certain conditions phytic acid also separates as magnesium phytate from
ammoniacal solution.

Hart and Andrews [1903], in their study of the relative amounts of
Inorganic and organic phosphate in foodstuffs, precipitated the inorganic
phosphate as ammonium phosphomolybdate at 65°C. in the presence of
dilute nitric acid. This method was therefore used but the precipitation was
carried out at room temperature since it was thought that the phytic acid
might be hydrolysed by the nitric acid at 65°C. <A slight hydrolysis does
occur at this temperature (see the following paper), but it is insufficient to
vitiate the results of Hart and Andrews, or those of McCollum and Hart,
who employed this method. The precipitation of the ammonium phospho-
molybdate at the ordinary room temperature takes place slowly, but is generally
complete in 1-2 days. The completion of the precipitation is shown by the
colour of the solution; it is yellow when the reagents are added and becomes
quite colourless when the deposition of the yellow precipitate is finished. In
the presence of organic matter the solution may remain yellow, which is due
to the action of nitric acid on protein in the enzyme extracts. At the same
time protein matter is precipitated so that the estimation cannot be carried
out by titration as in Neumann’s method. When the precipitation was
complete the precipitate was filtered off and washed with water and then
dissolved in dilute caustic soda. Inorganic phosphate was then precipitated
by ammonium magnesium citrate. The details of these experiments will be
described in a later paper dealing with the analysis of phytin. Total P.O,
was estimated by Neumann’s method.

Pancreas. Liver.
130 ¢.c. pancreatic juice +130 c.c. sodium A. 125 cc. extract of dog’s liver +125 c.c.
phytate solution. | H,0. B. 125 c.c. extract of dog’s liver + 125
: | ¢.c. sodium phytate solution.
P,O, in gm. |
At commencement 0-0025 | P.O; in gm.
After 14 days... 0:0055 A B Difference
| At commence-
Total 0-0850 ment ... 0°0348 0:0408 0-0060
|Afterl day... 00351 0-:0454 0°0103
» odays... 0:0379 0:0486 0-0107
Po femeaeree 0:0376 0:0516 0-0140

Total 0-0469 0°1407 0-0938

56

Intestine.

(1) 50c.c. extract of cat’s
intestine (which hydrolysed
trypsin digest of caseinogen)
+200 ¢c.c. phytin solution.

(2)

tion,

P,O; in gm. P,O; in gm. P.O; in gm.
At commencement 0:0126 At commencement 0:0190 Atcommencement 0°0256
After 1 day 0:0145 | After 5 days 0:0179 After 1 day 0:0264
» 2 days 0:0158 ‘ », 2 days 0:0283
pa. .. 0:0180- OS Pe Slane 0:0306
Total 0:1154 Total 0°1154
(4) 200c.c. extract of ox intestine + 100c.c. (5) 150c.c. extract of sheep’s intestine +

sodium phytate solution.

P,O; in gm.
At commencement 0:0198
After 1 day 0°0219
» 9 days 0:0251
” 5 ” 0:0264
3 Oe 0:0277

Total 0:0964

(6) 100c.c. extract of rabbit’s intestine +
150c.c. sodium phytate solution.

PO; in gm.
0-0085
0:0106

At commencement
After 6 days

Total 0°0964

Castor oil seeds.

(a) 150 c.c. acetic, acid extract +100 c.c.
sodium phytate solution.

P,O, in gm.
At commencement 0:0134
After 1 day 0:0190
» 2 days 0:0230
Se ae eo 0:0291

R. H. A. PLIMMER

100 ¢.c. extract of dog’s
intestine + 150 c.c. phytin solu-

Total 0°1103

Yeast.
(1) A. 100c.c. extract of zymin+150c.c. |
water. B. 100c.c. extract of zymin +150 c.e.
sodium phytate solution,

P20; in gm.

A B Difference
At commence-
ment 0:0291 0:0362 0:0071
After 7 days... 0°0362 0°0412 0-0050
Total 0:0583 0°1509 0:0926

(3) 150 ¢.c. extract of dog’s
intestine + 100 ¢c.c. sodium phy-
tate solution.

120c.c. sodium phytate solution.

P,O, in gm.

0:0137
00139

At commencement
After 6 days _

Total 0:0912

(7) 150c.c, extract of intestine of rabbit fed
for 3 weeks on bran+100 ¢.c. sodium phytate
solution.

P.O; in gm.
At commencement 0:0188
After 2 days 0-0267
LOE 0:0299
” 8 ” 0:0301

Total 0°1065

(b) 100 c.c. acetic acid extract +150 c.c.
sodium phytate solution.

P.O; in gm.

At commencement 0°0188
After 1 day 0:0279
» 4days 0°0458
” 8 ” ove 0:0554
lees He 0:0633

Total 0°1458

(2) A. 100c¢.c. extract of zymin+150c.c.
water. B. 100 c.c. extract of zymin+ 150 c.c.
sodium phytate solution.

P,O; in gm.
A B Difference |
| At commence-
ment Pee O-Oa50 0:0373 0:0043 |
After 9days... 0°0359 0°0411 0:0052
0:0579 0-1420 0:0841

Total

R. H. A. PLIMMER

ov
Ns“

Bran.

A. 125c.c. extract +125 ¢.c. water. B. 125 ¢.c. extract +125 c¢.c. sodium phytate solution.

P.O, in gm,

A B Difference
At commencement 0:0513 0°0566 0°0053
After 1 day i 0:0531 0°1087 0°0556
» 4 days am 0°0559 071354 0:0795
SHO! 55 on 0:°0559 071354 0:°0795
Total 0:0609 0°1510 0°0901

These figures show such a marked difference in the extent of hydrolysis
by the extract of bran and by the other enzymes that it can scarcely be con-
cluded that phytase exists in any of the extracts except in that of castor oil
seeds; the hydrolysis in this case is very slow; only a very slight increase in
the amount of inorganic phosphate is noticeable with the intestinal and liver
extracts, and none with pancreatic juice and extract of zymin.

III. HEXOSEPHOSPHORIC ACID.

Harden and Young [1908] have found that hexosephosphoric acid is
formed during the fermentation of glucose and fructose by yeast or by zymase,
and that it is subsequently hydrolysed. They attribute the hydrolysis to a
special enzyme, hexosephosphatase. Harding [1912] has recently investigated
the action of enzymes upon hexosephosphate and has observed that an enzyme
present in castor oil seeds hydrolyses it, and that this enzyme is not present
in ox pancreas. These data are so closely parallel to those obtained with
glycerophosphoric acid that it was very important to test the action of the
intestine and of bran upon this compound. Extracts of both these tissues
were found to contain hexosephosphatase. The parallelism between the
action of glycerophosphatase and hexosephosphatase is thus extremely close,
and it would seem that the same phosphatase hydrolyses both these com-
pounds.

EXPERIMENTAL.

The hexosephosphoric acid employed in these two experiments was most
kindly supplied to me by Dr Young. The solution of hexosephosphoric acid
was neutralised and mixed with a neutral extract of intestinal mucous
membrane and extract of bran, both of which were prepared by the methods
described under glycerophosphoric acid. Toluene was added and the mixture
kept at 37°C. The inorganic P,O, was estimated in samples by precipitation

58 R. H. A. PLIMMER

with ammonium magnesium citrate and converted into magnesium pyro-
phosphate. Total P.O; in the solution was estimated by Neumann’s method
with another sample.

(a) Iniestinal Hetract.

100 c.c. extract +160 c.c. hexosephosphate solution.

P,O; in gm.
At commencement a ang 0:0435
After 2 days sop onic AGE 0-0815
Pe ee eee ak 50 ae 0-0819

Total 0:0860

P.O, in intestinal extract added per 50 c.c. sample=0°0146 gm.

(6) Bran Extract.

A. 100 c.c. bran extract+150 ¢c.c. water. B. 100 ¢.c. hexosephosphate solution +150 c.c.
water. C. 100 c.c. bran extract +100 c.c. hexosephosphate solution + 50 c.c. water.

A B C Difference

Bran extract | Hexosephosphate Hexosephosphate + between C

alone alone bran extract and A+B
At commencement 0-0225 0:0092 0:0352 0:0035
After 1 day xa 0:0236 0:0093 0:1006 0:0677
days op 0:0240 0:0110 071115 0-0765
ek eae a 0:0244 0:0121 0:1188 0:0823

Total 0:0279 0:1801 0:2054 —

Hydrolysis has occurred with both intestinal extract and bran extract; it—
is most marked in the case of intestinal extract where it was complete in two
days. Bran extract caused a marked hydrolysis in one day but it was not
complete in four days, in which time scarcely half of the compound was

hydrolysed.

IV. ETHYL PHOSPHORIC ACID AND DIETHYL PHOSPHORIC ACID.

The action of enzymes upon the simple esters of phosphoric acid has never
been investigated. Ethyl phosphoric acid and diethyl phosphoric acid were
therefore tested against those extracts of tissues which had been found to
hydrolyse glycerophosphoric acid, phytic acid and hexosephosphoric acid.
The parallelism in the action of the enzyme upon glycerophosphoric and
hexosephosphoric acid occurred again in their action upon ethyl phosphoric
acid; diethyl phosphoric acid was not hydrolysed by intestinal extract nor by
extract of castor oil seeds; unfortunately the quantity of diethyl phosphoric
acid available was not sufficient to test against the other extracts.

The material used in these investigations was prepared either by the

R. H. A. PLIMMER 59

action of phosphoric pentoxide or of phosphorus oxychloride upon. absolute
alcohol. The product of the reaction in each case was neutralised with excess
of baryta; barium phosphate was filtered off and the clear filtrate precipitated
by an equal volume of alcohol; barium ethyl phosphate was precipitated
in crystalline form and recrystallised from water and alcohol. This was con-
verted into the sodium salt by treatment with the calculated quantity of
sodium sulphate and the solution so obtained was used in the experiments
with the enzymes. A small quantity of barium diethyl phosphate was con-
tained in the alcoholic filtrate obtained in the preparation by means of
phosphorus pentoxide ; it crystallised out on concentration of the alcoholic
solution and was purified by recrystallisation from dilute alcohol. The solu-
tion of the sodium salt prepared by treatment of the barium salt with sodium
sulphate was employed for the investigation of the action of the enzymes.

J

EXPERIMENTAL.

The enzyme solutions were prepared in the same way as described under
glycerophosphoric acid. Inorganic phosphate was estimated by precipitation
with ammonium magnesium citrate and conversion into magnesium pyrophos-
phate. The experiments were the following :—

A. Sodium Ethyl Phosphate.

Pancreas. Intestine.

0°5 gm. trypsin (Fairchild) | (1) 125c.c. neutral intesti-| (2) 110 c.c. neutral intesti-
was dissolved in 110 c.c. deci- | nal extract +130 c.e. sodium’ nal extract+150 ¢.c. sodium
normal sodium carbonate solu- | ethyl phosphate solution. | ethyl phosphate solution.

tion and filtered. 100c.c. ex- |

tract+200c.c. sodium ethyl | P05 in gm. P.O; in gm.
phosphate solution. At commencement 0:0168 At commencement 0-0121
: : After 1 day ... 070492 | After 1 day .. 0°0457
P.O; in gm. » 2days ... 0°0572| ,, 2days .., 0-0560
At commencement 0:0052 ee ere ee 0-0628 aoe nN a 0°0626

After ll days... 0-0046 Total 0-0654 |

Total 0:0659 |

Total 0:0672

Castor oil seeds.

(1) 125 c.c. acetic acid extract of seeds+ | (2) 110 c.c. neutralised extract of castor
130 e.c. sodium ethyl phosphate solution. oil seeds+150 c.c. sodium ethyl phosphate
. solution.
: Eases PO; in gm.
mos pat At commencement 0-0075
ovegs haa : Afterlday ... 00135
we ee ie 0:0383 ees) iv i

» 2 days =f 0:0307
| Total 0:0735

Total 0:0697

60 R. H. A. PLIMMER
Yeast. Bran.
60 c.c. extract of zymin+200 c.c. sodium 60 c.c. extract of bran+200 c.c. sodium
ethyl phosphate solution. ethyl phosphate solution.
P.O, in gm. P.O; in gm.
At commencement 0-0111 At commencement 0:0148
After 1 day si 0:0283 After 1 day ise 00190
» » days Ae 0-0398 » 2 days ae, 0:0216
a es x2 0:0491 ny ee ae = 00259
031
Total 0-1052 parcixe ae eee BS

Total 0:0938

B. Sodium Diethyl Phosphate.

Intestine. Castor oil seeds.

150 ¢.c. acid extract of intestinal mucosa + 150 c.c. neutralised acid extract +100 c.c.
100 c.c. sodium diethyl phosphate solution: | sodium diethyl phosphate solution: after
after 3 days solution neutralised with a few 4 days acidified with 2 drops of glacial acetic
drops of sodium carbonate solution. acid.

PO; in gm. P.O, in gm.
At commencement 0:0250 At commencement 0:0138
After 13 days... 0:0250 After ll days ... 0-0156
Total 0:0888 Total 0:0875

Sodium ethyl phosphate is hydrolysed most readily by the extract of the
intestinal mucosa and it is hydrolysed by an acid extract of castor oil seeds ;
the hydrolysis also takes place where the extract is neutralised.

Extracts of yeast and of bran also contain an enzyme which can hydrolyse
it; in the case of bran the hydrolysis is very slow.

Sodium diethyl phosphate was not hydrolysed by either intestinal extract
or extract of castor oil seeds.

V.. HYDROXYMETHYL-PHOSPHINIC AcID (OH: CH,°: PO,: H).

The previous results of the action of enzymes upon compounds which are
esters of phosphoric acid led to an investigation of their action upon hydroxy-
methyl-phosphinic acid, a compound in which the phosphoric acid is linked to
carbon instead of through oxygen as in the other compounds.

This compound was prepared by Mr H. J. Page [1912] in this laboratory
and a quantity was kindly placed at my disposal for this purpose. No natural
organic phosphorus compounds of this constitution are known; these experi-
ments were consequently of further interest, since if hydroxymethyl-phosphinic
acid were hydrolysed by enzymes, natural organic compounds of this constitu-
tion might be expected to occur in nature.

R. H.

The results were :
(a) Action of Trypsin.
105 ¢.e. = sodium carbonate extract of 0°5

gm. trypsin (Fairchild) +150 ¢.c. sodium hy-
droxymethyl-phosphinate solution.

PO; in gm.
At commencement 0:0046
After 7 days 0:0055

Total 0°1344

(c) Action of Extract of Castor
oil seeds.

100 ¢.c. acetic acid extract of castor oil
seeds+150 c.c. sodium hydroxymethyl-phos-
phinate solution.

P,O; in gm.
At commencement 0-0009
After 4 days 0-0019

Total 0°2840

(e) Action of Extract of Zymin.

(1) 100 c.c. extract of zymin+150 c.c.
sodium hydroxymethyl-phosphinate solution.

P,O, in gm.
At commencement 0°0393
After 1 day 0:0487
» 2 days 0:0509
” 4 Pe) 0:0528
Ae eas 0:0549

Total 0°2333

A. PLIMMER 61

EXPERIMENTAL.

The solution of the free acid was exactly neutralised and mixed with the
enzyme solution which was prepared as previously described; the analysis
was carried out in the same way as with glycerophosphoric acid.

(6) Action of Extract of Intestinal
Mucosa.
100 ¢.c. neutralised extract of dog’s intesti-

nal mucosa +4150 ¢.c. sodium hydroxymethyl-
phosphinate solution,

P,O, in gm.
At commencement 0:0134
After 4 days 0:0144

Total 02422

(d) Action of Extract of Bran.

100 c.c. extract of bran+150c¢.c. sodium
hydroxymethyl-phosphinate solution.

P,O, in gm.
At commencement 0:0508
After 4 days 0°0526

Total 0-2244

(2) A. 100 c.c. extract of zymin+150c.c.
H,Oascontrol. B. 100c.c. extract of zymin +
150 ¢.c. sodium hydroxymethyl-phosphinate
solution.

P20; in gm.

A B Difference

At commence-
ment ... 0:0216 0-0227 0-0011
After 2days... 0°0296 0°0362 0:0066
bwon es. © (O0-:0310 0:0393 0-0083
chart 0:0329 00424 0°0095
Total 0:0583 0°1977 0°1394

As was expected this compound was not hydrolysed by any of the extracts
which contain glycerophosphatase even if we include the action of the extract
of zymin; in this case an increase of inorganic P.O; was observed at the
commencement but the increase was very slight later on; it is probably due
to a more rapid separation of P.O, in the enzyme solution arising from a

change in reaction when the mixture was made.

62 R. H. A. PLIMMER

VI. NUCLEIC ACID.

The action of the enzymes of animal organs upon various nucleic acids
has been studied by many investigators, but their attention has been confined
almost entirely to the separation of the purine bases from the complex mole-
cule, and little or no attention has been given to the separation of the
phosphoric acid. Milroy [1896] found that 10-16 per cent. of the total
phosphorus in a peptic digest of nuclein was inorganic phosphoric acid, and
that in a trypsin digest the amount of inorganic phosphoric acid varied
from 6-47 per cent. of the total phosphorus content. No estimation of the
amount of the phosphoric acid present in the enzyme solution was made, so
that a portion or the whole of the inorganic phosphate might have been so
introduced into his solutions. Levene and Medigreceanu [1911] in some of
their recent papers state that the enzymes of the intestinal mucosa hydrolyse
nucleic acid with the liberation of phosphoric acid; no estimation of the
amount of the inorganic phosphate was made. The formation of phosphoric
acid during the autolysis of yeast, first observed by Béchamps in 1865, was
attributed to the hydrolysis of nucleic acid; hexosephosphate and hexose-
phosphatase were unknown at that time and it is most probable that the
- phosphoric acid observed by Béchamps originated from this compound rather
than from nucleic acid.

There is thus no satisfactory proof that inorganic phosphate is hberated
by enzymes from nucleic acid though it is very probable. The following
experiments were therefore carried out to test these earlier observations and
to compare the hydrolysis of nucleic acid by the same enzymes which have
been found to liberate phosphoric acid from the other organic phosphorus
compounds.

It has not been possible to test all the various nucleic acids but the results
show that an extract of intestinal mucosa slowly separates inorganic phosphoric
acid from thymus nucleic acid and from wheat nuclein, but more rapidly and
completely from the organic phosphorus compound (inosinic acid) in meat
extract ; trypsin does not hydrolyse thymus nucleic acid, nor the compound
in meat extract. An extract of zymin showed no appreciable separation of
phosphoric acid from the thymus nucleic acid, whereas an extract of bran
hydrolysed it.

EXPERIMENTAL.

The thymus nucleic acid used in these experiments was prepared for me
by Miss R. F. Skelton by Kossel’s method [1894]. The preparation was
almost pure white in colour and contained 8 per cent. of phosphorus and

R. H. A. PLIMMER 63

17 per cent. of nitrogen. It dissolved readily in decinormal sodium carbonate
giving a solution which was faintly brown in colour.

The meat extract was a solution of lemeo, and the wheat nuclein was
also kindly prepared for me by, Miss Skelton from wheat germ by digestion
with pepsin hydrochloric acid. The insoluble residue of nuclein was dissolved
in dilute sodium carbonate, separated from starch by centrifugalisation and
then reprecipitated by acid. After washing with water it was dissolved in
decinormal sodium carbonate solution and the brown solution thus obtained
was employed.

The enzyme extracts were prepared in the same way as those used in
investigating the hydrolysis of glycerophosphoric acid; the extract was mixed
with the solution of substrate and kept in the presence of toluene at 37°C.
Samples were removed and inorganic P.O; estimated by precipitation with
ammonium magnesium citrate. Total P,O; was estimated in another sample
by Neumann’s method. : .

(a) Action of Extract of Pancreas.
(1) 2gm. trypsin (Fairchild) were dissolved in (2) 0°5 gm. trypsin (Fairchild) dissolved in

N

10 10
tered. A. 125 c.c. extract+125 ¢c.c. water. | tered. 100 c.c. extract+150 ¢.c. of a solution
B. 125¢.c. extract+125c¢.c. solution contain- _ containing 10 gm. of lemco.
ing 2°3 gm. thymus nucleic acid dissolved in

260 c.c. — sodium carbonate solution and fil- | 110 c.ce. s sodium carbonate solution and fil-

N | PO; in gm.
—~ sodium carbonate.

10 At commencement 0-1189
After 7days ... 0-1194
P,O; in gm.
in B Total 0°1331
At commencement 0:0071 0-0070
After 1 day a 0:0074 0:0087
», 11 days re 0:0074 00089

Total 0:0139 0-0951

(6) Action of Extract of Intestine.
(1) 100c.c. neutralised extract of intestinal (2) 100c.c. neutralised extract of intestinal
mucous membrane of dog+150c.c. solution | mucous membrane of dog +200 c.c. solution of
Jets : ., -. N | wheat nuclein in sodium carbonate neutralised
containing 2 gm. thymus nucleic acid in i0 | with acetic acid.

sodium carbonate. ;
P.O, in gm.

P,O, in gm.

At commencement 0-0147
At commencement 0°0152* After 1 day es 0-0211
After 1 day ae 0:0254 GE ET one 0-0200
» 2 days a; 0:0277 eae A oe 0-0228
” 4 ” eee 0°0310 “0596 :
one fa 0-0327 Total 0:0596

Total 0:0938

* 0:0117 gm. P,O; was contained in the
intestinal extract for each sample.

64 R. H. A. PLIMMER

(3) 60 c.c. neutralised extract of intestinal (4) 100 c.c. neutralised extract of intestinal
mucous membrane of a dog+11 gm. lemco | mucous membrane of cat+-10 gm. lemco dis-
dissolved in 200 c.c. water. solved in 160 c.c. water.

P,O, in gm. P.O; in gm.
At commencement 0:°1254 | At commencement 0:1409
After 1 day me 0°1458 After 1 day ‘ra 0°1668
ee eee 0:1470 35D Gaya yr ccs 0-1712
Total 0-1521 Total 0°1839

(c) Action of Extract of Zyman. (d) Action of Extract of Bran.

A. 100c.c. extract of zymin+150c.c. H,0. 100 c.c. extract of bran+150 ¢.c. solution
B. 100c¢.c. extract of zymin+150 c.c. of a | containing 2 gm. thymus nucleic acid in deci-
solution containing 2 gm. thymus nucleic acid | normal sodium carbonate.
in = sodium carbonate. P20; in gm.

: At commencement 0:0297*
E05 vou Afterlday ... 0-0432
A B Difference » 2 days a 0-0498
At commence- » 5 0-0625
ment ... 0:0216 0:0235 0:0019 oY ag os :
After 2days... 0°0296 0:0322 0:0026 Total 0:1091
pee BI) 00353 0:0043 % J :
, 8 5, ... 0°0329 0-0380 0-0051 0:0279 gm. was contained in the bran

extract for each sample.
Total 0-0583 0°1281 0:0698

Trypsin has no action upon nucleic acid; the action of intestinal extract
upon thymus nucleic acid is very slight and is still less upon wheat nuclein ;
it readily hydrolyses the organic phosphorus compound contained in lemco.
Extract of zymin has also a very slight hydrolysing action, but extract of bran
is more active. No marked hydrolysis of nucleic acid is noticeable by any of
the extracts.

VII. PHOSPHOPROTEIN.

The separation of the phosphorus of caseinogen by the action of enzymes
and by alkali was investigated in 1906 by Plimmer and Bayliss [1906], who
confirmed the earlier work of Biffi with regard to the action of trypsin.
Pepsin slowly converted the phosphorus of caseinogen into a soluble form,
but without the formation of phosphoric acid. Trypsin separated 35 per cent.
of the phosphorus as phosphoric acid, the remaining 65 per cent. being present
as a soluble organic compound. Papain also converted the phosphorus of
caseinogen into a soluble form, but no investigation was made as to whether
phosphoric acid was produced.

Complete separation of the phosphorus from caseinogen was thus not
effected by trypsin. It seemed however very probable that the enzyme of

R. H. A. PLIMMER 65

the intestinal mucosa, most probably erepsin, which according to Cohnheim
hydrolyses caseinogen, would complete the separation as phosphoric acid.
_ This was found to be the case. Papain also separates phosphoric acid from
caseinogen in both acid and alkaline media but the separation is very slow.
The vegetable proteoclastic enzyme “digestin” prepared by Yenjo in Japan
from an Aspergillus behaves like papain. Neither an extract of zymin nor
an extract of bran separates phosphoric acid from the organic phosphorus
compound remaining after the action of trypsin upon caseinogen. It might
have been expected that the extract of zymin would hydrolyse this compound
and its inactivity is probably accounted for by its method of preparation
with acetone, which coagulates the protein to which the proteoclastic enzyme
is attached, so that the latter is not extracted by water.

EXPERIMENTAL.

A trypsin digest of caseinogen was prepared by digesting 200 gm. pure
caseinogen dissolved in 2000 c.c. water containing 10 ¢.c. ammonia of sp. gr.
0:880 with 3 gm. trypsin (Fairchild) for two months at 37° in the presence
of 20 c.c. of toluene. The solution was no longer alkaline to phenolphthalein
after one month and ammonia was added till the solution became just red to
this indicator. At the end of this time the tyrosine which had separated was
filtered off. The clear brown solution contained 0:°0456 gm. total P.O, and
00209 gm. inorganic P.O, per 50 c.c. - 56 per cent. of the total P,O; was thus
present as organic P,O,;. A portion of this solution which was kept for another
four months was found to contain 0:0418 gm. total P,O; and 0:°0228 gm. in-
organic P.O, per 50 cc. A year later the amounts were 0:0430 gm. total P.O;
and 00242 gm. inorganic P,O;1.e. the solution contained 45 per cent. organic
phosphorus.

Hydrolysis thus proceeds but even after 18 months it is not complete.

The action of the extracts of the intestinal mucosa, bran and zymin was
investigated in the same way as described under glycerophosphoric acid.
The action of papain and digestin was investigated with caseinogen which
was dissolved in dilute caustic soda solution (20 gm. caseinogen + 40 c.c.
semi-normal NaOH +500 cc. water): one half of this mixture was used for
the experiment in alkaline solution and 200 c.c. of the remainder were acidified
with 25 c.c. of semi-normal sulphuric acid and employed for the experiment
in acid solution. Samples were removed at intervals and filtered for the
estimation of the inorganic P,O, and total P,O;. The action of papain was
also tested upon the above trypsin digest of caseinogen.

or

Bioch. vir

66 R. H. A. PLIMMER

The experiments were:

Action of Intestinal Extract.

(a) 125 c¢.c. extract of cat’s |
intestine + 125c.c.trypsin digest

of caseinogen. of caseinogen.

(b) 50c.c.acid extractof cat’s
intestine + 200c.c.trypsin digest

(c) 50c.c. acid extract of dog’s «
intestine + 200c.c. trypsin digest
of caseinogen.

Total 0:0431

P.O; in gm. | PO; in gm. P.O; in gm.
At commencement 0:0223 At commencement 0:0224 | At commencement 0:0244
After 3 days 0:0500 After 18 hours 0-0302 | After 17 hours 0-0329
Dae 0:0492 | wie wee Bic 0:0313 aa SOD" Be 0:0357
2 a AD 00325 | ,, 6 days 0:0367
Total 0:0579 | i edhe noaae
|
|

Action of Papain.
(a) In acid solution on ¢a-

seinogen. caseinogen.

P,O; in gm.
After 38 days 00037
” 64 ° 00052
Total 0°0203

” 64 29

Action of Digestin (Yenjo).
(a) In acid solution on caseinogen.
P.O; in gm.
After 4 months 0-0124
Total 0-0203
Action of Extract of Zymin.

A. 200 c.c. trypsin digest+ 50 c.c. extract.
B. 200 c¢.c. water+50 c.c. extract.

P,O; in gm.

A B Difference |
At commence-
ment 0-0286 0:0085 0:0201
After 2days... 0°0316 0:0131 0:0185
ie eee OL0282) 0:0139 0:0143
Total 0:°0634 0-:0279 0:0355

(b) In alkaline solution on

P,O; in gm.
After 38 days

Total 0:0349

Total 0:0406

_ (c) Ontrypsin digest. 200c.c.
trypsin digest +55 c.c. extract

of papain.
0:0075 P.O; in gm.
0:0092 At commencement 0-0166
After 1 day 0-0168
5 2 days 0-0166
yw ctie 0-0168
Total 0:0268
(b) In alkaline solution.
P,O, in gm.
After 4 months 0:0236

Total 0:0330

Action of Extract of Bran.
A. 200 ¢.c. trypsin digest+50 c.c. extract.

|B. 200 c.c. water +50 c.c. extract.

P,O, in gm.
A B Difference
At commence-
ment =) 10:0338 0-0142 0:0196
After 2days... 0°0315 0-0148 0:0167
ANG Tee 0:0327 0:0147 0:0180
Total 0°0494 0-0190 0-0304

In comparison with the other extracts that of the intestinal mucosa is the
most active in separating phosphoric acid from caseinogen. The hydrolysis
is almost complete within 24 hours; in all these experiments a small quantity
remained in an organic form. This is most probably due to impurity and
corresponds to the small quantity of organic phosphorus which always remains

_——

R. H. A. PLIMMER 67

when phosphoprotein is decomposed by alkali, as was found by Plimmer and
Scott [1908].

Papain in alkaline solution separated one quarter of the total P,O, in
35 days and this amount was slightly increased in 64 days; in acid solution
in the same time one fifth was separated as inorganic phosphate, which
amount was increased to one quarter in 64 days; further separation does
not seem to occur as papain does not hydrolyse the organic P,O, present in
the trypsin digest of caseinogen. Digestin also never separated the whole
of the phosphorus of caseinogen. Neither extract of bran nor extract of
zymin separated phosphoric acid from the organic phosphorus in a trypsin

digest.

SUMMARY AND CONCLUSIONS.

The action of enzymes upon the organic phosphorus compounds is best
summarised by a tabulation of the results:

Castor Yeast
Pancreas Liver Intestine oilseeds (Zymin) Bran

Glycerophosphorie acid fe: 0 0 + - ~ +
Hexosephosphoric _,, Ba 0 $F 4 ~ + a
Ethyl phosphoric _,, eae 0 Eee + + +
Diethyl phosphoric ,, Bak $s oe 0 0 Ss a
Phytic acid cde 0 0 0 + 0 +
Nucleic acid (Thymus) A 96 0 re + eos +? ~

= ,, (Wheat) ee “es er +?

< 5, (Meat) age 0 + a
Hydroxymethylphosphiuic acid 0 0 0 +? 0
Phosphoprotein Ae 3 + 0 0

The most active tissue in the hydrolysis of the organic phosphorus com-
pounds is the intestinal mucosa; all the organic phosphoric compounds
except phytic acid were hydrolysed by it. Diethyl phosphoric acid and
hydroxymethyl-phosphinic acid must be excluded entirely as they were not
hydrolysed by any extract. Phytic acid and phosphoprotein stand out in
marked contrast with the other compounds. Phytic acid is attacked readily
only by the enzyme in bran extract. Phosphoprotein is the only compound
which is hydrolysed by the pancreas. The constitution of phosphoprotein is
different from that of the other compounds; it is a protein and is hydrolysed
by the proteoclastic enzymes and it seems most likely that the separation of
the phosphoric acid is effected by the same enzyme which unlinks the amino-
acids from one another. The other organic phosphorus compounds examined
are esters of phosphoric acid. The inactivity of the animal tissues most rich
in lipase and the activity of an extract of castor oil seeds which does not
contain lipase shows that the lipoclastic enzymes have no action on these

5—2

68 R. H. A. PLIMMER

esters and that they are hydrolysed by a special group of enzymes. Are we
to consider that these esters of phosphoric acid are all hydrolysed by one
enzyme, or are we to consider that each compound is hydrolysed by a specific
phosphatase? If there were only one phosphatase then phytic acid should
be hydrolysed by the enzyme of the intestine, but as this is not the case we
must regard phytase as a specific enzyme. The varying extent of the
hydrolysis of the other compound by the different extracts supports the
supposition that there is a specific enzyme for each phosphoric ester. It can
scarcely, however, be admitted that the specificity is really so great. If the
chemical constitution of each of the compounds be considered it is seen that
glycerophosphoric acid and ethyl phosphoric acid are simple esters, hexose-
phosphoric acid is regarded as hexosediphosphoric acid and nucleic acid is a
complex made up of hexose or pentose mono-phosphoric acid.

One enzyme is supposed to hydrolyse the group of a-glucosides, another
enzyme the group of §8-glucosides, invertase separates fructose from sucrose
and raffinose, trypsin attacks the polypeptide linking of the amino-acids;
consequently it can scarcely be conceived that the specificity of the phos-
phatases extends to every ester and it would be better to group the phos-
phatases into several classes such as mono-phosphatases which attack mono-
esters like glycerophosphoric acid and ethyl phosphoric acid, di-phosphatases
which attack di-esters like hexosephosphoric acid, hexa-phosphatases which
attack hexa-esters like phytic acid. Glycerophosphatase will hydrolyse both
glycerophosphoric acid and ethyl phosphoric acid as well as other simple
esters. Proof of such a supposition will only be given by further work when
more simple esters have been examined and when the constitution of
hexosephosphoric acid, phytic acid, and nucleic acid have been definitely
established.

These experiments confirm the work of Fingerling and Gregersen which
shows that the animal organism can synthesise its organic phosphorus com-
pounds from inorganic phosphates. They are not in favour of the belief that
organic phosphorus compounds are essential for the well-being of animals,
a belief which has arisen from feeding experiments with inorganic and organic
phosphorus. No proper control was ever made in such experiments. When
phosphoprotein was compared with albumin as the substitute for the supply
of protein the nutritive value of albumin was reckoned as equivalent to that
of caseinogen, no attention being paid to the different composition of these
proteins in amino-acids; when phytic acid, or phytin, was compared with
inorganic phosphate the control was not made with inositol and phosphoric
acid. Further, if assimilation of organically combined phosphorus occurred,

R. H. A. PLIMMER 69

both phosphoprotein and phytic acid should be found in the blood and tissues.
Phosphoprotein occurs only as the secretion of the mammary gland in
mammals and of the yolk gland in birds; in these tissues it is synthesised.
Phytin has never been found in animals but inositol has been found in most
tissues,

The organic phosphorus compounds except phytic acid are all hydrolysed
by the intestinal mucosa and their fate in assimilation is like that of all the
foodstuffs :—the proteins are known to be completely broken down into amino-
acids before they enter the circulation; the fats are hydrolysed as they pass
through the intestinal wall; the disaccharides are broken down by enzymes
in the mucous membrane of the small intestine. The organic phosphorus
compounds are therefore almost certainly assimilated as inorganic phosphate
and the organic radicle with which the phosphorus is combined.

The non-hydrolysis of phytic acid by the intestinal mucosa was at first
attributed to the experiment being made with carnivora (dog and cat), to
which animals vegetable food is not natural, but the same experiment with
the herbivora (sheep, ox and rabbit) required some further explanation as to
the manner of assimilation of this compound. The excretion of the phosphorus
of phytic acid as inorganic phosphate, which was shown by Scofone, Giascosa,
Mendel and Underhill and also Horner, pointed to the hydrolysis of the
compound. If the absorption be not complete, as the above workers have
pointed out, unchanged phytic acid should be present in the faeces. A rabbit
was therefore fed for three weeks on food containing a large proportion of
bran, and its faeces were examined at the end of this period for organic and
inorganic phosphate. The faeces were extracted with dilute hydrochloric
acid; inorganic and total P,O; were estimated in the extract by precipita-
tion with ammonium magnesium citrate and by Neumann’s method. The
phosphorus present in the extract was almost entirely inorganic P,O,, as is
shown by the following figures :

Extract Residue
Total Total Inorganic
H,O P.O; P,0; P.O; Total P,O;

Noy. 30, 1911 14-8 percent. 1:9percent. 1-:5percent. 1:4percent. 0-2 percent.
Dec. 1 27°7 ler 16 1-4 —

oy 6 30°0 17 1°6 15 0:2

PAG 25°6 2-0 2-0 19 —

i 8 35:3 19 1:7 1-4 92

Hydrolysis of the phytic acid had occurred; this is not caused by the
acid employed in extracting the faeces, for dilute acid at the ordinary tem-
perature does not hydrolyse phytic acid (see following paper). Since the

70 R. H. A. PLIMMER

animal was fed on bran the hydrolysis must be effected by the phytase in the
foodstuff. The assimilation of phytic acid by the herbivora thus resembles
that of cellulose, which was shown by Horace Brown [1892] to be hydrolysed
by the cytase present in the vegetable food.

In carnivora and man, the assimilation of the phosphorus of phytic acid
is also in the form of inorganic phosphate. Rogoginski [1910] has shown that
in dogs the unabsorbed phytin is present as such in the faeces and that in
man the phytic acid is hydrolysed by the bacteria in the large intestine. He
also finds no beneficial effect on metabolism from the organic phosphorus
compounds. Absorption of phytic acid by animals would require rapid dialysis
through the semi-permeable membranes. The dialysis of phytic acid through
parchment membranes is slow and is only complete after several days. This
is shown by the following experiments:

26c¢.c.sodium phytateinside 30 c.c. sod. phyt. inside 70 c.c. sod. phyt. inside
50 c.c. water outside 100 ¢.c. water outside 200 c.c. water outside
P,O, in 5 c.e. P,O, in 5 e.c. P.O, in 5 c.c.
v= Ss = = SS SSS —S—— =
Inside Outside Inside Outside Inside Outside
26-9 mgm. 0 mgm. 31°5 mgm. 0 mgm. 32°5 mgm. 0 mgm.
After 1 day — 15 -- 4-4 = 2°7
» 2days — 5°3 ee 7-0 be: 5-1
Ra od — ~- _- 76 — 6°7
ee 11°5 9°3 8-0 8-0 20-0 6-6

In all probability the membrane of the intestine of animals behaves in
the same way as the semi-permeable membrane of plant seeds, which does
not allow acids to pass through, as was shown by Adrian Brown [1909].

There is thus no evidence of the absorption of phytic acid; it will be
hydrolysed either by the enzyme in the foodstutts, or by bacteria in the large
intestine, and will enter the circulation of animals as inositol and phosphoric
acid.

The value of the organic phosphorus compounds in nutrition entirely

depends on the nature of the organic matter with which the phosphoric acid
is combined.

REFERENCES.

Armstrong and Ormerod (1906), Proc. Roy. Soc. 78, B, 376.

Brown, H. T. (1892), J. Chem. Soc. 61, 352.

Brown, A. J. (1909), Proc. Roy. Soc. 81, B, 82.

Biilow (1894), Arch. ges. Physiol. 57, 89.

Cook (1909), U. S. Department of Agriculture, Bureau of Chemistry, Bulletin, No. 123.
Connstein, Hoyer and Wartenberg (1902), Ber. 35, 3988.

Dietz (1907), Zeitsch. physiol. Chem. 52, 279.

Fingerling (1912, 1), Biochem. Zeitsch. 38, 448.

Fingerling (1912, 2), Biochem. Zeitsch. 39, 239.

R. H. A. PLIMMER

Giascosa (1905), Biochem. Centr., 4, 572.

Gregersen (1907), Zeitsch. physiol. Chem. 53, 453.
Gregersen (1911), Zeitsch. physiol. Chem. 71, 49.
Grosser and Husler (1912), Biochem. Zeitsch. 39, 1.
Harden and Young (1908), Proc. Roy. Soc. 80, n, 299.
Harding (1912), Proc. Roy. Soc. 85, n, 418.

Hart and Andrews (1903), Amer. Chem. J. 30, 470.
Horner (1907), Biochem. Zeitsch. 2, 428.

Jordan, Hart and Patten (1906), Amer. J. Physiol. 16, 268.
Kossel (1894), Ber. 27, 2215.

Levene and Medigreceanu (1911), J. Biol. Chem. 9, 389.
Loevenhart and Peirce (1906), J. Biol. Chem. 2, 397.
Marfori (1905), Archivio di Fisiol. 2, 217.

Mathison (1909), Biochemical J. 4, 274.

McCollum and Hart (1908), J. Biol. Chem. 4, 497.
Mendel and Underhill (1906), Amer. J. Physiol. 17, 75.
Milroy (1896), Zeitsch. physiol. Chem. 22, 307.

Page (1912), J. Chem. Soc. 101, 423.

Power and Tutin (1905), J. Chem. Soc. 87, 249.
Plimmer and Bayliss (1906), J. Physiol. 33, 439.
Plimmer, Dick and Lieb (1909), J. Physiol. 39, 98.
Plimmer and Scott (1908), J. Chem. Soc. 93, 1699.
Posternak (1903), Compt. rend. 137, 202.

Rogozinski (1910), Chem. Zentr., m1. 1549.

Scofone (1905), Biochem. Centr., 3, 606.

Suzuki and Takaishi (1907), Bull. Coll. Agric. Tokyo, 7, 503.

Suzuki and Yoshimura (1907), Bull. Coll. Agric. Tokyo, 7, 495.

Tanaka (1910), J. Coll. Eng. Tokyo, 5, 25.
Tutin and Hann (1906), J. Chem. Soc. 89, 1749.
Willstatter and Liidecke (1904), Ber. 37, 3753.

V. THE HYDROLYSIS OF ORGANIC PHOSPHORUS
COMPOUNDS BY DILUTE ACID AND BY DILUTE
ALKALT.

By ROBERT HENRY ADERS PLIMMER.

From the Ludwig Mond Research Laboratory for Biological Chemistry,
Institute of Physiology, University College, London.

(Received November 23rd, 1912.)

Plimmer and Scott [1908] have shown that phosphoproteins may be
distinguished from other organic phosphorus compounds occurring in animal
tissues by their behaviour with dilute caustic soda. The phosphorus of the
phosphoproteins is completely separated as inorganic phosphate by one per
cent. caustic soda at 37°C. in twenty-four hours. Dilute caustic soda at
37° C. does not decompose the other natural organic phosphorus compounds
but their hydrolysis at higher temperatures has never been investigated.
Neumann’s method of preparation of nucleic acid, in which the tissues are
heated with caustic alkali, shows that nucleic acid in comparison with
phosphoprotein is much more stable; it is not definitely known if the nucleic
acid is hydrolysed by alkali. Hexosephosphoric acid is readily decomposed
by alkali. The synthetical esters of phosphoric acid with methyl, ethyl and
other alcohols are stable to alkali, as was shown by Lossen and Kohler in
189i.

In order to determine the composition of the natural organic phosphorus
compounds they are hydrolysed by heating with acid at a high temperature
and very frequently under pressure. The hydrolysis of glycerophosphoric
acid by acids has recently been investigated by Malengreau and Prigent
[1911] who find that the decomposition proceeds according to the laws of a
monomolecular reaction and that it corresponds to the hydrolysis of ethyl
phosphoric acid and other esters of phosphoric acid studied by Cavalier [1899].

There appear to be no further data concerning the hydrolysis of the
natural organic phosphorus compounds by acids and alkalies. The statement
is usually made that they are decomposed very easily and that loss, due to
hydrolysis, occurs in their isolation from tissues when the solutions are being
concentrated. More definite information as to their hydrolysis in solution,

R. H. A. PLIMMER 73

both alone and in the presence of acid and alkali, is therefore required,
Such data are not only of use in their preparation, but may also serve for the
elucidation of their constitution. More practical knowledge concerning the
hydrolysis of phytic acid by acids was required for the estimation of inorganic
phosphoric acid in the presence of phytic acid and for the comparison of the
hydrolysis of the various compounds with the hydrolysis by enzymes (see
preceding paper). The experiments have been carried out for these purposes
“and no attempt has been made to ascertain the kinetics of the reaction.

EXPERIMENTAL.

Dilute acid and dilute alkali of concentrations varying from normal to
twice normal have been used. The hydrolysis by acids was carried out in
glass vessels, that by alkali in copper vessels. Porcelain vessels were as
unsuitable as glass owing to the solvent action of the alkali upon the glaze.
The copper vessels were also attacked by the alkali: the solution generally
contained dissolved copper as shown by the blue-green colour; the copper
compound was precipitated on neutralising with acid and a colourless filtrate
was obtained. The solution of the organic phosphorus compound was either
mixed with the acid or alkali and samples of equal volume were placed in
separate vessels which were maintained at the desired temperature, or the
entire mixture was placed in a thermostat. The vessel was then removed,
cooled to room temperature and samples taken for analysis.

The same preparations of glycerophosphoric acid, phytic acid, ethyl
phosphoric acid and the same trypsin digest were used as described in the
previous paper: hexosephosphoric acid was again kindly placed at my
disposal by Dr Young: the nucleic acid was again prepared for me by
Miss Skelton by Kossel’s method and Mr Page gave me a further quantity of
hydroxymethyl-phosphinic acid.

The analyses of phosphoric acid were made by precipitation with
ammonium magnesium citrate and conversion into magnesium pyrophosphate.
Total phosphoric acid in the solution was estimated in a separate sample by
Neumann’s method. All results were then calculated out in terms of P.O,
for the same volume.

As mentioned in the previous paper, inorganic P.O, is not precipitated in
the presence of phytic acid. Inorganic P,O, was therefore estimated by
precipitation with ammonium molybdate. The hydrolysis by acids was
carried out with nitric acid. In most experiments the phytin was dissolved
in the nitric acid; there was no necessity to remove the calcium as its
presence in the solution does not interfere with the precipitation of the

74 R. H. A. PLIMMER

ammonium phosphomolybdate. When alkali was used the cooled sample
was neutralised with nitric acid, an equal volume of 2N nitric acid, and
then 20-30 cc. of 10 per cent. ammonium molybdate solution added; the
precipitation began almost immediately and in 24 hours the solution was
quite colourless. The precipitate was then filtered off and estimated by
the Neumann method by solution in seminormal caustic soda and titration
with seminormal sulphuric acid after boiling off the ammonia. The
analytical figures obtained are evidence that inorganic P,O, can be
estimated by this method in the presence of phytic acid. Details concerning
the inhibition of the precipitation of inorganic P,O, by ammonium magnesium
citrate in presence of phytic acid and the estimation by ammonium molybdate
at room temperature will be given in a later paper dealing with the analysis
of phytin.
The details of these experiments are :

I. Hydrolysis of Glycerophosphoric Acid.

(a) By Aevd.

‘ N 4 * N 5

I By, i H,SO, at 86° C. ll. By Z H,SO, at 92° C.

20 c.c. glycerophosphorie acid diluted to 20 c.c. glycerophosphoric acid diluted to
500 ¢.c. with water. 50 ¢.c. samples+50 ¢.c. | 500 ¢.c. with water. 50 ¢c.c. samples +50 e.c.
N
ai H,SO, kept at 86° C. = H,SO, kept at 92° C.

P,O, in gm. P,O, in gm.
At commencement 0-0069 | At commencement 0:0102
After 3 hours ... 0:0092 After 1 day re 0°0547
REA Tiga a 0-0112 | 2) 2 dayse ie 00683
$9 Lo es sai 0-0391 33, (Oss as 0:0928
AR aa sist 0:0494 | 3 A S Ae 0:0998
ay GAY Sia lees 0:0756 | 55) Obes, iss 0°1578
AE eles? pee 0:0855 | lO. Sc 0°2629
spe he yp ane 00906 Total 0:2587
Total 0°2675
(b) <Autohydrolysis.

i. At 95° C. ii. At 75° C.

20 c.c. glycerophosphoric acid diluted to 25 c.c. glycerophosphoric acid diluted to
500 c.c. 50 c.c. samples kept at 95° C. 500 c.c. and kept at 75° C.

P.O; in gm. P,O, in gm.
At commencement 0:0074 At commencement 0:0157
After 2 hours... 0-0114 After 2 days... 0:0216
Sn ie Ae 0:0154 AS eS ae: 5b 0:0281
oa Olas a 0:0179 sa, ates si 0°0341
Fate ee se 0-0194 PAMB Bas sa 00404
ea ae 0:0342 Br Ofte = 0-0458
Total 0:2612 pt ape 0°0512
” 9 ey vee 0:0623

Total 0°1374

Hike
(c) By Alkali.

i. By = NaOH at 95° C.

25 e.c. glycerophosphoric acid diluted with

700 e.c. a NaOH. °
P,O, in gm.
At commencement 0:0137
After 2 days 0-0144
Bento & 0-0118
” 5 ” 0:0150
Sc mlaa ss 0°0195
ee O's ra 0:0179

Total 0-2194

PLIMMER 75

li. By a NaOH at 75° C.
6 gm. sodium glycerophosphate dissolved in
2N
550 c.c. 5 NaOH (5 ¢.c.=20°1 ee. * 1,80,)
and kept at 75° C,

P,O, in gm,
At commencement —
After 2 days 0°0011
y, Es 00023
san eee 0-0010
” 16 ” 00018
33 92 3 0:0033
” 81 ” 0:0078

Total 0°1750

Il. Hydrolysis of Ethyl Phosphoric Acid.

(a) Byacid ({ HCl) at 75°C. _ (b) By alkali € NaOH) at

(c) Autohydrolysis at 75°C.
200 e.c. ethyl phosphoric acid

300 ¢.c. sodium ethyl phos- | 75° C. solution.
2 300 ¢.c. sodi thyl phos-
phate solution + 300 c.c. = ata dnaas s fal aa P,O, in gm.
HCl. phate solution + 300 c.c. “1 | At commencement 0-0002
P,O, in gm. NaOH. After 1 day 0-0018
P Ooch om » 2 days 0-0033
At commencement 0-0001 2/5 In gm. 4 0-0064
After 8 hours 00003 | At commencement 0 ‘ 6 : 0-0094
pe day: 0-0013 After 1 day 0 aa Ware 00526
;, 2 days 0:0031 », o days 0
ic ae 0:0045 noe ae 0 | Total 0:0746
nae ae SEER 0-0065 Ae SEB We 0 |
EON 5; 0:0077 :
Dab ., 0-0091 Total 0°0926 |
ete ae 0-0115
» 11 ,, 0-0146
pee OW 35 0:0204
ee cOrs, ... 0°0240
Total 0:0469
Ill. Hydrolysis of Phytice Acid.

(a) By Aeid.
i BY 2 HNO, at 37° C.

125 ¢.c. phytic acid solution +125 c.c.
HNO,.
P.O; in gm.
At commencement 0:0127
After 1 day 00144
» 2 days 0-0142
” 5 ” 0:0134

Total 0°2156

2N

ii. By i HNO, at 37° C.

2N
1 gm. phytin dissolved in 250 c.e. a HNO,.

P.O; in gm. —
At commencement 0-0055
After 2 days 0:0062
ae Ota, 00062
8 0-0066

Total 0-:0976

76 R. H. A. PLIMMER

oN
Hf, oi = HNO, at 65°C. iv. By 4Y HINO, at 64° C.

igi. pliyin’ dissolved mwi20 ce: a0 1 gm. phytin dissolved in 260 c.c. HNO.

wee ba

130 c.c. = HNO,.
P.O, in gm. P,O; in gm.
At commencement 0:0051 At commencement 0:0046
After 1 hour 0:0056 After 15 mins. ... 0:0051
,, 2 hours 0:0057 Pers Ole as 5 0:0053
ee ee 0:0054 eel ehour) 2s 0:0053
5; Gms 06-0058 oa 2 enoOurs!... 0:0057
Total 0:0938 Total 0:0913

vy. By 2 HNO, at 64° C.

1 gm. phytin dissolved in 260 c.c. = HNO,.

vi. By > HNO, at 75° C.

1 gm. phytin dissolved in 260 c.c. = HNO.,.

P,O, in gm. P,O, in gm.
At commencement 0:0049 At commencement 0:0063
After 3 hours 00056 After 1 day wae 0:0134
eOues 0-0060 » 2 days eS, 0:0232
oy 4) 39 0:0061 pas ee me 0:0280
{DAES 0:0065 ny tet Be oe 0°0439

Total 0-1040

vii. By : HNO, at 75° C.

1 gm. phytin dissolved in 260 c.c. = HNO,. 2-5 gm. phytin dissolved in 500 c.e. a HNO,.

P,O, in gm. P.O, in gm.

At commencement 0:0056 At commencement 0:0077

After 1 day 0:0116 After 1 day “ie 0°0216

» 2 days 0°0165 ZAG yS oe 0:0349

sae tte 0-0259 59° Bi ee as 0:0571

es 0:0436 my) O55 on 0:0777

py 2) 6p Re 0:0976

Total 0:1014 ys fis 0-1040

lee a 0-1147

Hp oe B00 0°1192

pee ee Ao 0-1243

” 15 »” eee 0-1255

sy MD 5 Hr 0:1274

” 17 Pry mere 0°1312

Total 0:0926

viii. By = HNO, at 75° C.

Total 0:1471

en

mabe

lh iy A nce

R. H. A. PLIMMER 77
(6) By Alkali. (c) Autohydrolysis at 75° C.
ny 2 NaOH at 75°C. 600 ¢.c. phytic acid solution,
300 c.c. sodium phytate solution +300 c.c. P.O, in gm.
2N N At commencement 00034
i NaOH. (5 ¢.c.=9°7 c.c. DY H.SO,.) After 1 day a 0-0071
; » 2 days Bes 0-0112
P.O, in gm. oat yah a 00151
At commencement 0:0093 ey a 00191
After 1 day a 0:0089 nO ae 0°0235
ys) 2 Gays: |... 0-0091 re ihe AF ts 0:0314
din! Sa 0:0079
He re? Total 0:0527
wee. 0-0094 eh
» 16 ,, Ee 0-0106
See 55 0:0119

Total 0:1407

IV. Hydrolysis of Hexosephosphoric Acid.

ey i . (L5N (ce) Autohydrolysis at 75° C.
(a) By acid ( HCL) at75°C.| (b) By alkali > NaOH) 50 o.c. hexosephiosphoric acid

100c¢.c. hexosephosphoricacid | at 75° C. | solution + 220 c.c. water. (5 c.c.
solution neutralised with NaOH| 40c.c. hexosephosphoric acid Srqiaine: : NaOH.)

+100 e.c. a HCl. (5 c.c. =|solution + 500 c.e. = NaOH.

N N P.O; in gm.
UE :2 ce: 2 NaOH.) (9 €.c.=16°5 c.c. a H,SO,.) At commencement 00087
After 1 day a 00494
P.O; in gm. P.O; in gm. » 2days ... 00698
Atcommencement 0-0034 | At commencement 0-0136 » D8 55 00791
After 7hours ... 0-0348 | After 1 day .. 071610 » 18 ,, -» 0°1152
” i day eee 0:0443 79 2 days Ano 0-1598 Total 0°1243
» 2 days ae 0-0480 eee ae a 071641
mS). 55 aor 0:0486 E
ess. ee 0-0508 Total 0:1978
op: Ree ae 0:0508

Total 0°0527
V. Hydrolysis of Nucleic Acid.
(a) By acid G Hc!) at 75° C. () By alkali € NaOH ) at 75° C.
3°5 gm. thymus nucleic acid dissolved in 6 gm. nucleic acid dissolved in 600 c.c.

200 c.c. H,0 + 10 c.c. = NaOH; 210 c.c. + NaOH. (5 c.c. =10 c.e. = H,S0,.)
2N

= HCl then added; the precipitate went into P,O, in gm.
solution in 15 minutes at 75° C. At commencement 0
¢ After 1 day oe 0-0040
P.O; in gm. Seeedays ..: 0-0106
At commencement — ela eke ee 0-0179
After 4 hours... 00250 2 MGR: ie 0:0369
wuiidays —4.. 0-0429 ee, 0°0510
» 2days... 0-0545 Li 0-0990
hae a ae was Total 0-1128
on Fares 0-0968
ae 0°1106

Total 0°1103

78

R. H. A. PLIMMER

(c) By alkali € NaOH ) at 70° C.
4 gm. nucleic acid dissolved in 200 c.c. H,0 + 200 c.c. 2N NaOH.

Pp,O, in gm.

At commencement
After 1 day
», 2 days

4

be)

99 9
22

hey

”
be

”

Total 0:0951

0:0020
0:0108
0°0145
0:0236
0:0305
0°0447

VI. Hydrolysis of Phosphoprotein by Alkali.

i. 200 c.c. trypsin digest of
2N
caseinogen + 50 c.¢. ae NaOH.

(=1-6 per cent. NaOH.)

2N
caseinogen + 25 ¢.¢. = NaOH.
(=1 per cent. NaOH.)

ii. 175 c.c. trypsin digest of |

iii. 180c.c. trypsin digest of
N

caseinogen + 20 ¢.c. = NaOH.
(=0°8 per cent. NaOH.)

P,O, in gm, P.O; in gm. | P,O, in gm.
At commencement 0:0194 At commencement 0:0199 At commencement 0:0209
After 1 day e 0:0320 | After 1 day site 0:0327 | After 1 day of 0:0272
» 2 days Bas 0°0321 » 2 days te 0:0342 », 2 days a 0:0300
: 03% a 0:0334
NS res. 02 Bee, be Se 03

Total 0°0342 Total 0-0406
VII. Hydrolysis of Hydroxymethyl-phosphinic Acid by Acid.

: 2N
130 c.c. sodium hydroxymethyl-phosphinate solution + 130 ¢.c. ai HNO,. (5 c.c.=10°3 c.e.

= NaOH.)
P,O,; in gm.
At commencement es a5 0:0006
After 1 day Bae a6 ae 0:0004
» 2 days ae i ce 0:0006
= 7 Bare ae ig mee 0:0006
99 8 9 0:0018

Total 0°1179

Glycerophosphoric acid is slowly hydrolysed by dilute acid, complete
separation of the phosphoric acid requiring 10 days at 92°C. ‘The glycerol,
which is formed, was isolated in another experiment by removing the
sulphuric and phosphoric acids by baryta, concentrating and distilling im
vacuo: the glycerol distilled at 162°C. at 10-15 mm. pressure and a yield
of 44 per cent. was obtained. Autohydrolysis is considerably slower; one-
eighth of the glycerophosphoric acid was decomposed in 16 hours at 95°C.,
and only one-half in 9 days at 75° C. It is not hydrolysed by alkali; the slight
increase in the amount of inorganic phosphate observed after 81 days by the
action of twice normal caustic soda was due to slight evaporation through
the cork ; the total P,O; at the end of the period being 0°1750 gm., which is
a little greater than that at the beginning—0'1509 gm.

——

R. H. A. PLIMMER 79

Glycerophosphoric acid is stable to alkali like ethyl phosphoric acid ;
Lossen and K6hler’s result has been confirmed and autohydrolysis and acid
hydrolysis of ethyl phosphoric acid have been carried out at 75° C. for com-
parison with the other compounds. Autohydrolysis of ethyl phosphoric acid
was not complete in 50 days and the hydrolysis by acid seems to be slower
than in the case of glycerophosphoric acid.

Phytic acid is the most stable of the organic phosphorus compounds.
At 37°C. dilute acid effected no hydrolysis; at 64°C. a slight hydrolysis
could be detected in 24 hours ; at 75°C. normal nitric acid separated about
half of the phosphorus as inorganic phosphate in 8 days. Complete hydro-
lysis was not effected by twice normal nitric acid in 17 days at the same
temperature. Autohydrolysis at 75°C. resulted in the splitting off of about
half the phosphoric acid in 7 days.

Phytic acid is also quite stable to alkali; the slight increase in the
amount of inorganic phosphate noted after 32 days by normal alkali at 75° C.
is due to concentration of the solution by evaporation through the cork.

Hexosephosphoric acid is easily hydrolysed by both acid and alkali.
Complete hydrolysis was effected by normal hydrochloric acid in 3-5 days,
the greater part of the phosphoric acid being separated in 1 day. Auto-
hydrolysis at 75°C. was much slower than acid hydrolysis ; half the phosphoric
acid was separated in 2 days; complete hydrolysis occurred in about 18 days.
Dilute alkali produced complete hydrolysis in 1 day. In all the experiments
a small quantity of organic phosphorus remained undecomposed: it is very
probable that another organic phosphorus compound was present as impurity
in the solution.

Nucleic acid is hydrolysed by both acid and alkali. Normal hydrochloric
acid effected complete hydrolysis in 8 days at 75°C. With normal caustic
soda a slow hydrolysis was observed ; about one-third of the phosphoric acid
was split off in 8 days, about one-half in 32 days; even after 76 days
a small amount of organic phosphorus remained in solution.

The organic phosphorus compound which remains in a prolonged tryptic
digest of caseinogen is also completely hydrolysed by | per cent. caustic soda ;
there is no difference in its behaviour from that of caseinogen; the former
observation of Plimmer and Bayliss [1906] was incorrect.

Hydroxymethyl-phosphinic acid is not hydrolysed by dilute acid at 75°C.

SUMMARY AND CONCLUSION.

Ethyl phosphoric acid, glycerophosphoric acid and phytic acid are
hydrolysed by acid, but are stable to alkali. Stability to alkali is therefore
a property of the esters of phosphoric acid.

80 R. H. A. PLIMMER

Hexosephosphoric acid and phosphoprotein are so different in their
behaviour to alkali from the above three compounds that some difference in
their constitution from that of the esters must exist.

It is not known how the phosphoric acid is combined in phosphoprotein,
but it is probably united with one of the amino-acids.

Hexosephosphoric acid reduces Fehling’s solution which points to the
presence of the functional aldehyde or ketone group in the molecule.

The phosphoric acid radicles are most probably combined with two of the
hydroxyl groups leaving the reducing group free. The action of the alkali
will destroy this grouping and the whole carbohydrate molecule will be
decomposed leaving the phosphoric acid: |

et

CH,OH-CHOH.CH.CH . CH.CHOH

OP re
Ue ee oN
HOOH HOOH

Nucleic acid, since it is hydrolysed like hexosephosphoric acid by both
acid and alkali, seems to occupy an intermediate position between the stable
esters and the very unstable hexosephosphoric acid. If the purine, or
pyrimidine, base be attached to the functional aldehyde group in the same
way as the alcohols in the glucosides, the action of alkali may be to destroy

the purine base leaving the aldehyde group for decomposition of the molecule,
and phosphoric acid will remain. A formula such as

CH,—CHOH—CH.CHOH.CHOH-.CH -O—purine base

|
PO

H6 OH

would explain the slow decomposition by alkali.

REFERENCES.

Cavalier (1899), Ann. Chim. Phys. [7], 18, 449.

Lossen and Kohler (1891), Annalen, 262, 209.

Malengreau and Prigent (1911), Zeitsch. physiol. Chem. 72, 68.
Plimmer and Bayliss (1906), J. Physiol. 33, 439.

Plimmer and Scott (1908), J. Chem. Soc. 93, 1699.

VI. THE NITROGENOUS CONSTITUENTS OF
LIME-JUICE.

By CASIMIR FUNK,

Research Scholar, Lister Institute.
(Preliminary Communication.)

From the Biochemical Department, Lister Institute of
Preventive Medicine. ;

(Received November 22nd, 1912.)

In connection with the work previously published on beri-beri [Funk,
1911, 1912, 1], in which the physiological importance of certain nitrogenous
substances, belonging most probably to the pyrimidine group, was de-
monstrated, a similar fractionation of lime-juice was undertaken for the
investigation of scurvy.

From the classical work of Holst and Fréhlich [1907, 1912] on scurvy we
know that this disease has a close analogy to beri-beri and is also due to
a lack of some substance in the food. They specially advanced our know-
ledge of this disease by inducing in animals (rabbits and guinea-pigs)
a disease which possesses a great similarity in symptoms to human scurvy.
They produced the experimental scurvy by an exclusive diet of oats or
autoclaved vegetables and were able to prevent it by an addition to this
diet of fresh vegetable-juice, unboiled milk or lime-juice. Lime-juice itself
has been known for many years as an excellent remedy for scurvy.

Some time ago I was able to show [1912, 2] that lime-juice contains an
anti-neuritic substance curing polyneuritis in birds, produced by an exclusive
diet of polished rice, and a substance, present in another fraction, which is
only capable of prolonging the life of these birds. Further experiments have
shown that pyrimidine substances in general possess this life-prolonging
action and a paper on this subject will shortly appear. No anti-scorbutic
substance however was obtained from lime-juice.

I had at first intended to try each fraction, obtained during this

Bioch. vi 6

82 C. FUNK

investigation, on guinea-pigs fed on oats, but in this I found considerable
difficulties. The animals refused all food after a period of three weeks and
even an addition of freshly made potato-jnice or lime-juice was not able to
save them from starvation. The same negative result was obtained with
fractions of lime-juice which might have been expected to contain the anti-
scorbutic substance in a more concentrated form.

The only positive results were obtained with milk, 50 cc. daily of
unboiled milk in addition to the oats diet preventing loss of weight and
onset of scurvy. Even after the elimination of casemogen and the other
proteins present in milk, the latter keeps its preventive power. In doing
this extreme care must be taken not to destroy the substance, and all boiling
must be carefully avoided. The caseinogen is precipitated by addition of acetic
acid at 50° and lactalbumin by dialysed iron-solution, kaolin however carries
the substance down. The solution is then concentrated by freezing it and
taking the supernatant liquid. This subject is being further investigated.
This unstable character of the anti-scorbutic substance renders it unlikely
that the usual method of separation, with the unavoidable use of alkali, would
lead to its isolation. Although no anti-scorbutic substance was obtained
from lime-juice, several new compounds have been detected and are
described below.

Besides a substance, which apparently belongs to the terpene group and
which has only been analysed and not further investigated, as having no
direct connection with the subject, three nitrogenous substances have been
isolated. One belongs to the purine group, the second to the pyrimidine
group and the third to the choline group. The small yield obtained did
not allow however of a complete investigation, so that this paper must be
regarded as preliminary.

EXPERIMENTAL.

407 litres of commercial lime-juice were precipitated in portions of
15 litres each with 2800 grms. of neutral lead acetate. The bulky pre-
cipitate was filtered off on a large Buchner funnel and the residue pressed
out in a hydraulic press: The combined filtrates were freed from the excess
of lead, first by an addition of sulphuric acid, then by sulphuretted hydrogen,
and were then concentrated im vacuo at low temperature. During the
distillation crystals separated out, which were filtered off, and weighed
when dry 258 grms. These were recrystallised from 20°/, acetic acid and
162 grms. were obtained in needles. The substance proved to be free from
nitrogen and was recrystallised for analysis once more from water. It is

C. FUNK 83

nearly insoluble in all other solvents, except acetic acid, and is volatile
with steam; M.P. 97°-100° (uncorr.).

0°1357 grm. gave 0°3400 CO, and 0-1312 H,0.
Found 68°33 9/, C0; 10°73 %/o H.
Cale. for C,,H,,0,; 68°42°/, C; 10°52 °/, H.

This substance probably belongs to the terpene group; it possesses
a very characteristic agreeable smell when heated. The nature of this
compound, which has no direct relationship with the scurvy problem, was
not further investigated.

To the filtrate from this substance enough sulphuric acid was added to
make a 5°/, solution and then a 50°/, solution of phospho-tungstic acid
until precipitation was complete. The precipitate obtained amounted to
nearly 12 kg., and consisted chiefly of potassium salt. _

INVESTIGATION OF THE PHOSPHO-TUNGSTATE PRECIPITATE.

The precipitate was decomposed in portions of 1 kg. each with 24 kg.
baryta in a mortar and the mixture was shaken on the shaking machine
for one hour. The precipitate was then filtered off, suspended in water and
again shaken. The excess of baryta was eliminated from the combined
filtrates by the careful addition of dilute sulphuric acid. The solution
obtained was then neutralised with nitric acid, as it was strongly alkaline,
and evaporated in vacuo to about two litres. In this solution the usual
fractionation with silver nitrate was performed and three fractions were
obtained. The first fraction came down on addition of a silver nitrate
solution and contained the substances of the purine group. This precipitate
was filtered off and to the filtrate a saturated solution of silver nitrate was
added until a drop of the liquid gave with a cold baryta solution a brown
precipitate which indicates an excess of silver. A saturated aqueous baryta
solution was then added until a drop of the clear liquid gave only a small
precipitate with silver nitrate and ammonia. ‘The second precipitate consists
of pyrimidine bases and substances of the histidine type. To the filtrate
pulverised baryta was added until the whole of the silver was precipitated.
The fraction thus obtained contains, in addition to traces of pyrimidine
substances, compounds of the arginine group. The last filtrate contains
substances of the choline group. All the fractions were investigated
separately.

6—2

84 C. FUNK

INVESTIGATION OF THE FIRST SILVER FRACTION.

The bulky precipitate obtained by a simple addition of silver nitrate was
filtered off, well washed with water, suspended in water and the double silver
nitrate salts converted into the silver salts by heating them on the water-
bath with ammonia. The precipitate was filtered off and washed with water,
until the filtrate was free from nitric acid. It was then suspended in water
and decomposed with sulphuretted hydrogen. The solution on evaporation
yielded 1 grm. of substance, which however contained a little ash. It was
recrystallised from water and gave 07 grm. of crystals in the form of plates.
After drying in vacuo at 110° they became brown at 240° and melted at
282° (corr.). The substance is precipitated by mercuric acetate and partially
by gold chloride. It. differs from the known purine bodies by the great
solubility in hot water and the absence of all known reactions for these
derivatives. Heated with nitric acid it becomes yellow, and brown when
more strongly heated, no change of colouration being produced on the addition

of alkali.
0°3855 grm. loses at 110° (vacuo) 0:0190 grm. H,O

01176 ,, requires 32°7c.c. N/10 H,SO,
0°1372 ,, a 38 c.c. on 3
0:1226 ,, gave 0:1788CO, and 0:0196 H,O
Cale. for C,H,0.N; Found
(Mol. wt.=181)

C 39°78, SPT

H 3-869), 3°55

N 38°67 °/, 38°77 38°92
34 Mol. H20 4°73 °/, 4:92

INVESTIGATION OF THE SECOND SILVER FRACTION.

The second silver nitrate precipitate was filtered off, washed well with
water and decomposed with sulphuretted hydrogen. In the filtrate the last
traces of baryta were taken out with a very dilute solution of sulphuric acid
and the solution was evaporated in vacuo. As the residue did not deposit
any crystals, the solution was precipitated with mercuric sulphate. The
solution obtained by decomposition of the precipitate showed no tendency
to crystallisation, but on addition of a picric acid solution an oil separated,
which was decanted and dissolved in acetone. After evaporation a solid
powder was obtained which amounted to 0°35 grm. This was recrystallised
from a mixture of alcohol and acetone and yielded about 0'1 grm. of brown
prisms, which decomposed at 205°-210°. An amount of this fraction

fe.

C. FUNK 85

corresponding to 0°01 grm. of the picrate was decomposed and given to a
polyneuritic pigeon which showed a considerable improvement and died
only after four days.

The mercuric sulphate filtrate, freed from mereury and sulphuric acid,
was evaporated in vacuo and the residue left in a desiccator after addition
of alcohol. After a few days a colourless substance separated out which
amounted to 0°7 grm. After recrystallisation from dilute alcohol 0°25 grin,
was obtained in the form of microscopical spherolites, which after drying at
110° in vacuo, melted and decomposed at 188°—189".

0:0958 grm. subst. required 7°5c.c. N/10 H,SO,; 10-96 °/, N.
0:1034 ,, gave 0°1625 COy and 0-0636 H,O; 42:86°/, C; 6°83°/, H.
Cale. for CyH,,0,N, (Mol. wt.=250) 43:2°/,C; 7:2°/,H and 11°2°/,N.

The third silver fraction yielded but little substance. On evaporation
of the solution obtained by decomposing the precipitate no crystalline
compound separated out. Picrie acid solution however gave a precipitate
which after standing formed crystalline, yellow spherolites, which began to
decompose at 260°; yield 0:1 gr.

INVESTIGATION OF THE SILVER NITRATE FILTRATE.

Of this filtrate, which amounted to 4600 ¢c., only 220 c.c. were taken
for further investigation. To this liquid after elimination of silver and
baryta, phospho-tungstie acid was added and 420 grms. of a dry precipitate
were obtained, which was decomposed in the ordinary way. The solution
of the bases belonging to the choline group was neutralised with hydrochloric
acid and evaporated in vacuo to dryness. To get rid of the last traces of water
the residue was redissolved in alcohol and the alcohol evaporated in vacuo.
Finally an alcoholic solution was made and an alcoholic solution of sublimate
was added as long as a precipitate was formed. ‘This was recrystallised from
hot water with the addition of a little sublimate. It was found to be very
slightly soluble in water. In this way 10°5 grms. of crystals were obtained
in the form of cubes. These were reJissolved in water and decomposed
with sulphuretted hydrogen. The filtrate was evaporated in vacuo to dryness,
the residue dissolved in alcohol and evaporated again. The alcoholic solution
finally obtained was precipitated with an alcoholic platinic chloride solution,
13 grms. of a pale yellow precipitate being obtained which was recrystallised
from dilute alcohol ; 0°85 grm. was thus obtained in the form of needles, which,
after drying at 110° in vacuo, melted at 220° (uncorr.). The substance gave
the following figures on analysis, after being dried in vacuo at 110,

86 C. FUNK

0:1666 grm. gave 0:0451 grm. Pt

01622 ,, required 4:2¢.c. N/10 H,SO,

0:2173 ,, gave 0:2104 CO,; 0°0878 H,O ; 0:0589 Pt and 0-2586 AgCl.
Cale. for (C,H,,0,NHCl),PtCl, (Mol. wt. 724). Found:

C: 26°52 ; 26:40
H: 4:42; 4:48
N: 3:86 3°62
Pt: 26°94; 27°07; 27:1
Cl: 29°55 29-44

On igniting the platinum salt a peculiar smell was noticed, which was
quite different from that given by choline.

A substance of this formula, with similar properties (M.P. 219°), was

described by v. Braun [1908] as the platinichloride of methylpiperidylacetic
betaine. ‘The substance is being further investigated.

REFERENCES.

v. Braun (1908), Ber. 41, 2129.
Funk, Casimir (1911), J. Physiol. 43, 395.
— (1912, 1), J. Physiol. 45, 75.
— (1912, 2), J. State Med. 20, 341.
Holst and Frohlich (1907), J. Hygiene, 7, 634.
(1912), Zeitsch. Hyg. Infektionsk, 72, 1.

VII. THE FLOWER PIGMENTS OF AW7/RRHINUM
MAJUS. 1. METHOD OF PREPARATION.

By MURIEL WHELDALE,
Fellow of Newnham College, Cambridge.

(Received November 23rd, 1912.)

I have previously [Wheldale, 1909, 2, 3; 1910, 2; 1911] made certain
suggestions as to the nature of the chemical reactions involved in the
formation of anthocyanin. There is little doubt that anthocyanin, as
a collective term, in the same sense as protein, sugar, tannin, etc., includes
many substances having in general similar properties, but differing among
themselves as regards constitution. Evidence, from various sources, has led
me to conclude that some anthocyanins may be derived from members of
the groups of natural yellow colouring matters, known as the flavones and
xanthones.

The yellow colouring matters are largely present in the plants as
glucosides, some, or possibly all the hydroxyl groups, being replaced by sugar.
I have suggested that the reactions-involved in the formation of anthocyanin
may be represented, in very general terms, as follows :—

Glucoside + Water == Chromogen + Sugar
(Flavone or Xanthone)
« (Chromogen) + Oxygen —> Anthocyanin.

The first reaction may be regarded as controlled by one or more glucoside-
splitting enzymes and it is conceivable that specific enzymes may act on
hydroxyl groups in different positions. When certain hydroxyl groups
(position to be determined) are free from sugar, oxidation may take place at
these points, or possibly condensation, or both, with the formation of
anthocyanin. The residual hydroxyl groups in the anthocyanin molecule
would probably be replaced by sugar and hence the anthocyanins would occur
as glucosides. There is evidence that the second reaction may be brought
about by an oxidase system oe 1909, 3; 1910, 2; 1911; Keeble and
Armstrong, 1912].

88 M. WHELDALE

Some of the evidence for the above hypothesis is derived from results
obtained in cross-breeding with varieties of Antirrhinum majus [Wheldale,
1907 ; 1909, 1; 1910, 1]. The wild type of this plant has magenta flowers,
the colour due to anthocyanin. Among other varieties produced on cultiva-
tion are three without anthocyanin, i.e. ivory, yellow and white. Ivory and
yellow contain pigments very probably of the flavone class. The white has
no such pigment but carries other factors which apparently act on the yellow
pigments to form anthocyanin. None of the three varieties alone can produce
anthocyanin, but when ivory is crossed with white, a plant having magenta
flowers is produced; when yellow with white, a plant having crimson
(possibly a mixture of magenta and yellow) flowers.

With a view to testing this hypothesis I commenced a chemical investi-
gation of the pigments of Antirrhinum and in the summer of 1911 I succeeded
in finding a satisfactory method for obtaining the pigment in quantity in the
solid state. Extractions were made from yellow and ivory separately, but all
varieties containing magenta anthocyanin (ie. ivory tinged with magenta,
magenta and crimson) were extracted together. In this way all extracts
contained mixtures of pigments, as will be seen below on referring to the
varieties in detail. When attempting to purify the crude pigment later,
I found difficulty in separating some of the pigments. Hence, when I again
prepared pigment this year, the precaution was taken, in the case of certain
varieties, of extracting portions only of the corolla, containing, to the best of
my belief, either one pigment or two, which could be easily separated.

The method of preparation is as follows:—the flowers, picked off the
spikes, are boiled with water in saucepans. The pigments are readily soluble
in water and the extract is filtered through large funnels into lixiviating jars.
The pigments are then precipitated as insoluble lead salts by adding solid
crystalline lead acetate until no further precipitate appears. After standing
a few hours the greater part of the liquid is decanted from the precipitates,
which are then filtered through a large Buchner funnel attached to a filter
pump. ‘The lead salts are stirred up with 5-10°/, sulphuric acid which
decomposes the salts with formation of lead sulphate. The lead sulphate is
filtered off and the filtrate contains the pigments as glucosides in dilute
sulphuric acid solution. These solutions are boiled for several hours in two-
litre Jena flasks fitted with simple tube condensers, care being taken to avoid
concentration of the solution, since under such circumstances the pigment
may become charred. Hydrolysis of the glucosides takes place during
boiling and, on cooling and standing about 12 hours, the pigment, which is
less soluble than the glucoside, is deposited. The deposit is separated by

M. WHELDALE 89

filtering through as small a Buchner funnel as possible, a Geryk pump being
used, After washing, the pigment is dried in a desiccator over sulphuric acid.

During the current year (1912), the following varieties were extracted.
Excellent coloured figures of, the varieties are given by Baur [1911].
Figures below refer to Baur’s plate.

Ivory. (Fig. 3.) The corolla contains a pale yellow pigment (ivory) and
a patch of deeper yellow pigment (yellow) on the lower lip. Extracts from
the whole flower or from the lower lip gave an orange-yellow lead precipitate.
From the separated upper lip, a canary-yellow lead precipitate. Pigment was
prepared separately from both upper and lower lips.

Yellow. (Fig. 2.) The lips of the corolla contain yellow pigment which
is chiefly confined to the epidermal cells. The inner tissues of the lips and
the entire tube contain ivory pigment. Extracts gave an orange lead pre-
cipitate. Pigment was prepared separately from both upper and lower lips.

Ivory tinged with magenta. (Fig. 6.) Flower, ivory, with some develop-
ment of magenta anthocyanin which is chiefly confined to the epidermis.
Extracts gave a yellow-green lead precipitate. Pigment was prepared from
entire flowers only and presumably contains ivory, yellow and magenta.

Magenta. (Fig. 7.) Magenta chiefly confined to epidermis; inner tissues
contain ivory and there is a yellow patch on the lower lip. Extracts gave a
deep green lead precipitate, some pigment apparently remaining in solution
in the acetate, forming a deep green solution which is dichroic, red by
transmitted and green by reflected light. Pigment was prepared separately
from both upper and lower lips. The upper presumably contains magenta
and ivory, the lower, yellow in addition.

Crimson. (Fig. 8.) Lips of the corolla are crimson and the tube magenta.
It is at present uncertain whether the crimson colour is due to the presence
of both yellow and magenta in the cells or to a distinct crimson pigment.
Pigment was prepared from entire flowers only.

Rose doré. (Fig. 15.) Corolla contains a “ red” anthocyanin [Wheldale
1909, 2]. The inner tissues, ivory, and there is a yellow patch on the lower
lip. Extracts gave a reddish-green lead precipitate. Pigment was prepared
from entire flowers only.

Bronze. (Fig.17.) Bears the same relation to rose doré as crimson to
magenta and it is again uncertain whether the colour is due to a mixture or
to a separate pigment. Extracts gave a deep-red lead precipitate. Pigment
was prepared from entire flowers only.

Methods of purification were first carried out on pigment obtained in
1911. As mentioned above, tinged ivory, magenta and crimson flowers had

6—5

90 M. WHELDALE

then been extracted together. The crude pigment which presumably con-
tained ivory, yellow and magenta was finely powdered thoroughly dried,
placed in a Soxhlet thimble, and extracted first with warm ether in which the
ivory and yellow pigments are soluble, though not readily. The anthocyanins,
both magenta and red, are insoluble in ether.

From the ether extract two pigments were obtained by fractional erystal-
lisation from alcohol and ethyl acetate. The less soluble pigment was taken
to be ivory, the more soluble to be yellow. It is doubtful whether either
pigment was obtained in the pure state.

Ivory pigment. This is readily soluble in alcohol and acetic acid, with
more difficulty in ether and ethyl acetate, very slightly soluble in cold, more
so in hot water; insoluble in chloroform and benzene. It crystallises from
dilute alcohol in plates ; M.p. 338°. It underwent combustion with difficulty
in oxygen and did not give constant results for carbon. Tested by Zeisel’s
method for a methoxyl group, some silver iodide was obtained but only in
such quantity as to indicate impurity. An acetyl derivative was prepared by
boiling with acetic anhydride and pouring the product into sodium acetate
solution. The product was purified with difficulty by crystallisation from
ethyl acetate. The final product was pure white and crystallised in glistening
needles; M.P. 182°. The combustion results were as follows:

C=6319/, H= 4305,
63°21 4°22
62°95 431

Of the known ‘flavones, the ivory pigment bears most resemblance to
apigenin in properties and acetyl derivative. Opinion is reserved as to its
identity until further analyses have been made, especially with the ivory
pigment (free from yellow) obtained from the upper lip of the ivory variety.

Yellow pigment. Crystallises in plates from dilute alcohol but was not
obtained in the pure state; M.P. 290°-800°.

Magenta pigment. After extraction with ether for several weeks, the
magenta pigment was obtained free from yellow. It crystallises, but not
well, from a mixture of alcohol and ethyl acetate. It decomposes without
melting when heated to 340°.

Other results obtained with magenta and yellow are reserved until further
work has been done.

The crude pigment was extracted in the laboratory of the John Innes
Horticultural Institution, Merton Park, Surrey. Large numbers of plants,
the offspring of varieties originally used in breeding experiments, were grown

M. WHELDALE 91

in the garden of the Institution. The yield of pigment was small and was
not ascertained quantitatively because of the labour involved in weighing the
flowers,

The work of purification and analysis, as far as it has gone, was carried
out in the Biochemical Laboratory of the Institute of Physiology, University
College, London. I am much indebted to Dr Aders Plimmer for kind help
and suggestions.

REFERENCES.

Baur, E, (1911), Kinfiihrung in die experimentelle Vererbungslehre,” Tafel I.
Keeble, F. and Armstrong, E. F. (1912), Proc. Roy. Soc. 85 B, 214,
Wheldale, M. (1907), Proc. Roy. Soc. 79 B, 288,
—— (1909, 1), Reports to Evolution Committee of the Royal Society, Report 5, 1.
—— (1909, 2), Proc. Roy. Soc. 81 B, 44.
— (1909, 3), Proc. Cam. Phil. Soc. 15, 137.
— (1910, 1), Zeitsch. induktive Abstammungs- und Vererbungslehre, 3, 321.
—— (1910, 2), Prog. Rei. Bot. 3, 457.
—— (1911), J. Genetics, 1, 133.

VIII. THE DENSITY AND SOLUTION VOLUME
OF SOME PROTEINS.

By HARRIETTE CHICK anp CHARLES JAMES MARTIN.

From the Lister Institute.

(Received November 25th, 1912.)

In a previous communication dealing with the viscosity and other
properties of caseinogen solutions [1912], we pointed out that there was
a shrinkage in volume and corresponding increase in density when that
protein formed colloidal solution. In order to investigate the point the method
employed was to determine the density of the dry, powdered protein by
weighing in benzene or some other indifferent fluid and to compare the value
with that obtained by calculation from the observed density of a fairly con-
centrated solution. The latter value was found to be much greater than the
former.

As far as we have been able to ascertain, this is the first attempt to
determine the density of proteins with the exception of gelatin. Quincke
[1903] found the specific gravity of a specimen of the latter to be 1°368, while
Liidekings [1888] had previously given the value of 1412, and had shown
that the density calculated from that of jellies (14 to 35°/, gelatin) was as
high as 1°9.

The increase in density when proteins are dissolved in water is an instance
of the phenomenon constantly associated with colloidal solution of the lyophile
or emulsoid character. The intimate association of the protein particles with
water leads to a shrinkage in total volume, so that the increase in density
is more properly regarded as an attribute of the whole system than of the
protein alone. Analogous cases are those of gum-tragacanth and starch. In
the case of the latter Rodewald [1897] found the density in the dry state
and that calculated from that of a solution to be 1:38 and 1:49 respectively.

In our experiment with caseinogen, a sample of the pure, dry, powdered
protein was prepared by acidifying a solution in dilute sodium hydroxide,
washing and drying the precipitate, and grinding to a fine powder. A known
weight was inserted in a pyknometer bottle, covered with benzene of known

H. CHICK AND C. J. MARTIN 93

density, and all the air expelled by careful heating in vacuo. The bottle was
then filled up with benzene in a thermostat at 25°C. and weighed. The
weight obtained was compared with that of the bottle filled with benzene at
the same temperature, and the density of the powder calculated. The value
obtained was 1°318 as the mean of two determinations (see Table I). A solu-
tion of caseinogen (7°85°/,) in dilute sodium hydroxide gave a density of
1:0240, from which the density of sodium caseinogenate was calculated to be
1:42.

In order to compare this value with that obtained by direct determination,
it is necessary to apply a small correction, seeing that the solution consisted
of sodium caseinogenate. The ash present composed 4331°/, of the dry
weight and, assuming this to be sodium carbonate, 2°/, of the dry weight
would consist of sodium, and the density, for comparison with that of the
powdered caseinogen, should be reduced by that amount. After this correc-
tion has been made the caseinogen in solution is still seen to be denser than
the solid caseinogen in the proportion of about 1°39 to 1318.

In addition to caseinogen three other pure proteins have been similarly
investigated, viz. :

Crystalline serum albumin, prepared from horse-serum by the method of
Hopkins and Pinkus [1898].

Serum globulin (pseudo-globulin). Three different samples were investi-
gated, obtained from horse-serum by different methods. I and II were
prepared by repeated precipitation with half-saturated ammonium sulphate,
the euglobulin present in the serum being separated by the subsequent
dialysis ; in the case of III the latter was removed by preliminary precipita-
tion with saturated brine, after which the pseudo-globulin was precipitated
from the warm diluted filtrate by adding anhydrous sodium sulphate until
the concentration equalled 20°/, Na,SO,. The precipitate was thoroughly
washed with a solution of sodium sulphate of the same strength.

Crystalline egg-albumin, prepared from egg-white, also by the method of
Hopkins and Pinkus.

All the proteins were thoroughly dialysed in presence of toluene and filtered,
and concentrated solutions were finally obtained. Determinations were then
made of the protein-content and the density. The latter were obtained by
pyknometer readings, carried out usually at 15°C. or 20°C., the density of
water at the same temperature being taken as unity for the purpose of
calculation. It was found that the proportionality between the density of
water and that of the protein solution was, within our error of experimentation,
maintained between 15° C, and 25° C.

94 H. CHICK AND C. J. MARTIN

Dry specimens of the proteins were obtained by evaporating the solutions
to dryness and grinding the residue to a fine powder. The evaporation and
drying were carried out in vacuo at room-temperature, and were continued
until the weight remained constant, an operation requiring 2-6 weeks. It
was feared that, if the samples were dried at a high temperature, the
denaturation of the proteins might introduce a source of error. This would
not appear to be so, for, in the case of serum albumin, the densities of two
- samples dried at 100°-110° and 20° respectively, were not found to differ by
any significant amount (Table I). The densities of the various powders were
determined by weighing in benzene of known density (water at 4° C. = 1-00)
at 25°, as described above.

TABLE I.

Density of Protein in solution compared with that in the solid state.

Protein In solution Dry State
“Concentration Density Calculated Density*
of protein of the density* of of the
°/) (by weight) solution the protein protein
- Caseinogen ... as 7°85 1:0241 1°39} 1:318
Keg-albumin (eeyetalliae) ae 14:6 1:0401 1:359 1:269¢
Serum-albumin (crystalline) 22°15 1:0647 1°378 1:275§
1281+
Serum-globulinI ... ws 15:33 1:0428 1:365 1:279§
Serum-globulin II ... aE 16 35 10466 1:374 1:289f

Serum-globulin III.:. —... 11:05 1-0316 1:384 1:312t

* Density of water at 4°C.=1:-00.

+ Corrected for presence of sodium.

+ Dried at room temperature in vacuo,
§ Dried at 105°—110° C.

The results are given in Table I and show that the density of albumin
and globulin is also increased when in the state of colloidal solution, and to
a higher degree (6°8°/,) than was found for caseinogen (5°/,). Egg- and
serum-albumin have an almost identical density in the dry condition, viz.
1:269 and 1:281, and this is increased in the same proportion on entering
solution, viz. to 1°359 and 1°378 respectively. The three samples of globule
give 1:293 as mean value when dry and 1:°374 when in solution.

In Tables II and III are given the variation of the density of solitons
of serum-albumin and serum-globulin with alteration in protein concentration.
If the latter is plotted against the former, straight lines are obtained. The
curve for caseinogen, on the other hand, shows a slight convexity which

gh Te 4

H. CHICK AND C. J. MARTIN 95

is rather more than can be explained by our experimental error, the
contraction on entering solution being proportionally greater for dilute
solutions. This is shown in the third column of Table II, where the
value of the density of caseinogen calculated from that of its solution is seen
to decrease progressively from 1465 to 1412 as the concentration of the
protein is increased. In the case of the serum proteins, however, the value
remains constant (Table IIT).

TABLE II.

Density of Caseinogen (Na Caseinogenate) solutions of varying concentration.

Concentration Calculated
of caseinogen Density of density of
°/) (by weight) the solution * sodium caseinogenate
9°39 1-0283 1-412
8°33 1:0250 1:409
752 1:0232 1-424
6:05 1-0190 1-437
4:35 1:0140 1-460
2-173 1:0070 1-465
1-086 1-0033 4

* Compared with water=1-:00 at the same temperature, 15° C.
+ Solution too dilute for trustworthy calculation.

TABLE III.

Density of solutions of Horse-Serum Proteins of varying concentration.

Calculated

Concentration Density of density of

Protein 9/, (by weight) the solution * the protein
Serum-globulin a =: 15°33 1:0428 1-365
10°32 1:0290 1°374
i 6-916 1:0190 1-365
i 3-478 10096 1:370
Serum-albumin Sa ois 22°15 1-0647 1:378
a 15°16 10440 1°382
7°725 1:0220 1°381

”?

* Compared with water=1-00 at the temperature of expt.

SUMMARY.

A comparison has been instituted in case of four proteins, viz. caseinogen,
egg- and serum-albumin, and serum globulin, between the density directly
determined with dry specimens and that calculated from the specific gravity

HL CHICK AND CG. J. MARTIN

of concentrated solutions. ‘The latter is found to be 5-8 °/, in excess of the |
former, showing the extent of shrinkage in volume taking place when these |

96

proteins enter colloidal solution.

REFERENCES.

Chick and Martin (1912), Zeitsch. Chem. Ind. Kolloide, 11, 102.
Hopkins and Pinkus (1898-1899), J. Physiol., 23, 130.
Liidekings (1888), Wied. Ann. 35, 552.

Quincke (1903), Drude’s Annalen, 10, 478.

Rédewald (1897), Zeitsch. physikal. Chem. 24, 193. f

IX. A NOTE CONCERNING THE INFLUENCE OF
DIETS UPON GROWTH.

By FREDERICK GOWLAND HOPKINS anp ALLEN NEVILLE.

From the Institute for the Study of Animal Nutrition ; Department
of Agriculture, Cambridge.

(Received December 22nd, 1912.)

In a recent paper Osborne and Mendel [1912, 1] have described certain
experiments which seem to show that young animals (rats) can grow when
fed upon artificial diets consisting of “ purified” constituents alone.

These experiments would therefore indicate that the accessory factors of
uncertain nature (complex lipoids or “vitamines” or “ hormones”) which
others have believed to be necessary are not, as a matter of fact, indispensable.
Such experimental results must give pause to those who like ourselves are
engaged in an endeavour to separate, and identify more closely, the accessory
substances referred to. But they are results which contradict what is now a
considerable body of experience, and the experiments which yielded them
seem to call for repetition.

These particular experiments were in a sense merely incidental to a wide
enquiry, on the part of the authors quoted, into the nutritive efficiency of
various proteins; and those which are of significance to the matter at issue
(those, namely, which showed actual growth, and not maintenance alone)
concerned only three animals [1912, 1, pp. 356, 358]. In a later paper
[Osborne and Mendel, 1912, 2] the authors again refer incidentally to the
subject and speak of having obtained “a considerable degree of success ” by
feeding in the absence of “the hypothetical organic hormones,” etc. [1912, 2,
p- 242].

But the weight-curves given in this paper show little more than main-
tenance of the animal, without growth; while, for some reason, the three
experiments of the earlier paper which showed complete success in the
promotion of vigorous growth are not further quoted. This success the
authors attributed in the main to the fact that, in this dietary, salts were

98 F. G. HOPKINS AND A. NEVILLE

supplied in a mixture made to imitate as exactly as possible the salts of milk ;
but it is difficult to understand how animals so omnivorous as rats can
depend for normal growth upon a very exact balance of particular acid and
basic ions. In the experiments described by one of us in an earlier paper
(Hopkins, 1912] the salts administered were obtained by carefully ashing a
normal food mixture of proved efficiency. Upon an artificial diet containing
these salts the animals did not grow; but they grew at once when certain
substances were added to the diet, some of these addenda being certainly
incapable of supplying any deficiencies in the inorganic constituents of the
original diet.

We have now fed a large number of rats upon the diet employed by
Osborne and Mendel. The salt mixture as described by them was made with
the greatest care, and all their directions for the preparation of the food
mixture were exactly followed. But the protein and starch were thoroughly
extracted with alcohol, and the lactose used was several times precipitated.
from its aqueous solution by the addition of alcohol. The methods of feeding
were exactly those used in the paper by one of us already referred to, except
that the food, being more fatty and coherent, was not mixed with water.

Twenty-four rats from various sources, of weights from 50 to 60 grms.,
were placed upon the mixture. Although the consumption of food was
satisfactory, every rat, without exception, rapidly ceased to grow. In the
greater number growth ceased as early as the sixth day, in some on the
ninth, and in all before the fifteenth day. A comparatively brief period of
maintained weight was then followed by a steady decline. In the case of
eighteen of the animals the diet was administered up to the time of death,
’ which, in all but four cases, occurred before the fortieth day.

To six of the above set of rats, after the decline in their weight had begun,
2 c.c. of milk per diem were given. An immediate betterment of the general
condition of the animals followed; growth was re-established and the health
then maintained. In another experiment six rats were put upon Osborne and
Mendel’s diet, but were given milk from the first. In each case the animal
grew.

We have spoken of the food consumption (which was carefully determined
each day) as being satisfactory. It was smaller however than the consumption
of the animals upon a somewhat different artificial dietary in the experiments
described earlier by one of us. Its energy value, during the period which
preceded actual loss of weight, was just under 40 calories per 100 grms. live
weight, instead of over 50 calories. That the former value is nevertheless
well in excess of the amount required for maintenance is shown by some, as

F. G. HOPKINS AND A. NEVILLE 99

yet unpublished, experiments upon comparable animals made in the Cambridge
Physiological Laboratory by Miss Hill.

When the small ration of milk was given each day in advance of feeding
with the Osborne and Mendel, diet the food consumption remained of the
same order, and did not rise to the amount consumed by the rats in Hopkins’
earlier experiments [1912]. The resulting growth though quite definite and
steady was distinctly slower than in the experiments mentioned.

Our rats clearly behaved very differently from the three animals fed upon
a similar diet by Osborne and Mendel. The difference we are unable to
explain. Realising from previous experience how very small a remainder
of the substances which are extracted by alcohol may leave an artificial
dietary with some power of maintaining growth, and knowing that ether is a
greatly inferior solvent for them, we fed rats upon the Osborne and Mendel
mixture in which the protein (commercial casein) was extracted with ether
only (as in one of Osborne and Mendel’s experiments [1912, 1]; Curve 58,
p- 358) and the lactose not crystallised from alcohol. We were unable to
obtain growth however, though even this small difference in the diet
appreciably lengthened the period during which the animals remained in
health.

We do not in this note propose to publish the schedules of weights, ete.
They will be given elsewhere at a later date. The purpose of the present
note is to indicate that there is still reason for a continuance of the search
for special accessory substances of potent influence upon growth. It should
be pointed out that Osborne and Mendel themselves admit that such
substances may exist.

REFERENCES.

Hopkins, F. G. (1912), J. Physiol. 44, 425.
Osborne and Mendel (1912, 1), Zeitsch. physiol. Chem. 80, 307.
Osborne and Mendel (1912, 2) J. Biol. Chem. 13, 233.

7
;
is
F

X. THE CONDENSATION OF TRYPTOPHANE AND
OTHER INDOLE DERIVATIVES WITH CERTAIN
ALDEHYDES.

By ANNIE HOMER, Beit Memorial Fellow.
From the Physiological Laboratory, Cambridge.

(Received November 26th, 1912.)

The oxidation of tryptophane by methods other than that adopted by
Hopkins and Cole [1902] gives rise to coloured amorphous compounds. |
These observers, employing ferric chloride as the oxidising agent, isolated two
substances, the one having the formula C,H;ON and afterwards identified by
Ellinger [1906] as 8-indolealdehyde; the other an interesting crystalline base,
the analysis of which gave results in accordance with the formula C,,H,,N,.
This method of oxidation was used in the course of the present investigation
with a view to the isolation and identification of the base C,,H,,Ns.
Unfortunately, however, even though the conditions were varied as much as
possible, none of this substance could be obtained. Ellinger, when preparing
8-indolealdehyde by this method, had also failed to obtain Hopkins’ base.
In trying to ascertain the extent to which the ether used for extraction of
indolealdehyde had been directly or indirectly responsible for the production
of the base, Hopkins noticed that a crystalline derivative was formed when
an aqueous solution of tryptophane was in contact with ether which had
been heated locally with a glass rod. This compound however did not prove
to be the much sought after base C,,H,,N., but a new substance of an acidic
nature. At his suggestion the properties of the new substance, which may
be conveniently spoken of as the “ ether oxidation substance,’ were examined.
This work led to an investigation of the direct action of formaldehyde,
glyoxylic acid and glyoxal on tryptophane, and the action of trioxymethylene,
in the presence of a condensing agent, on indole, skatole, indolepropionic
acid, indoleacetic acid and tryptophane. It was found that the direct action
of formaldehyde and glyoxylic acid on tryptophane led to the formation of
colourless crystalline compounds ; in these cases the amino-group of trypto-
phane was the reactive group (Section A). Glyoxal reacted directly with

Bioch, vir 7

102 ; A. HOMER

tryptophane in the presence of an oxidising agent to form a coloured
derivative (Section B). For interaction with the inino-group of indole
derivatives, the presence of a condensing agent is necessary, and the resulting
products are all deeply colowred substances (Section C).

Section A.
I. Preparation of the ether oaidation substance.

(a) An aqueous solution of tryptophane was poured into a large wide-
mouthed bottle fitted with a bung, and sufficient ether was added to form a
layer a few mm. deep. In preliminary experiments the ether was heated
locally by plunging a red-hot glass rod into the bottle and keeping it there a
few minutes. Very pungent fumes were evolved, accompanied by a consider-
able rise in: temperature. The process was repeated several times, and the
bottle was well shaken and allowed to stand for a short time, when a
crystalline deposit slowly formed. This somewhat tedious process was
replaced by the following simple arrangement.

(b) The bung was pierced with four holes through which were passed a
straight wide tube open at both ends, two electrodes, and an exit tube of
narrow bore. The electrodes were connected inside the bottle by means of a
platinum spiral placed just above the surface of the ether. The wide tube
was arranged so as to have one end close to the platinum spiral. One end of
the narrow tube was flush with the bung inside the bottle, the other end was
connected with the vacuum pump and a slow current of air was drawn
through the apparatus entering by the wide tube.

Outside the bottle the electrodes were connected with a_ battery
sufficiently powerful to make the platinum spiral red-hot. The electric
current was then switched off, and the current of air was regulated so that
the heat developed during the incomplete combustion of the ether was
sufficient to keep the platinum wire glowing. As soon as the layer of ether
had disappeared, a further supply of ether was added, and the process
repeated. The bottle was then disconnected, corked and left to stand at a
temperature of 38° for several hours, during which time a crystalline deposit
slowly formed. This substance proved to be the same as that prepared
in (a).

(c) In another experiment a mixture of ether vapour and steam was
passed over red-hot glass, and the ensuing vapours were passed through a
solution of tryptophane ; under these conditions no ether oxidation substance

A. HOMER 103

was formed. Thus it is evident that, to bring about the production of this
substance, it is necessary for the tryptophane to be in contact with the
reactive principle at the moment of its formation. The pungent fumes
evolved during the slow incomplete combustion of ether were passed through
a wash bottle containing water, and on examination were found to contain
both hydrogen peroxide and formaldehyde.

(d) An attempt was made to prepare the substance by the action of
formaldehyde and hydrogen peroxide on tryptophane. A crystalline deriva-
tive was obtained, but it did not prove to be the ether oxidation substance
( post, p. 107).

In (a) and (b) the yield of crystalline product was 65 per cent. of the
original tryptophane. The substance melts and decomposes at 324°, is
colourless, crystalline, slightly soluble in water, alcohol and phenol, insoluble
in cold and hot benzene, xylene, toluene, ethyl acetate, aniline, nitrobenzene,
pyridine, amyl alcohol and ethylene dibromide. It is soluble in boiling
glacial acetic acid: with concentrated hydrochloric acid it darkens consider-
ably, and on evaporation im vacuo crystals separate out, which can be
recrystallised from ether and alcohol and melt at 285°. Several analyses of
this product were made, but failed to give concordant results. The substance
can be boiled with dilute hydrochloric and sulphuric acids (5°/,) without any
decomposition taking place. It is precipitated by the mercury sulphate
reagent, gives the glyoxylic but not the bromine tests, and does not darken
with concentrated sulphuric acid. It is soluble in alkalis, and can be
precipitated by means of acetic acid. The use of caustic alkalis is to be
avoided, owing to the pigmentation which ensues. Its sodium salt cannot be
recrystallised from water owing to hydrolysis.

The substance does not benzoylate easily, and does not form a naphthalene
sulphochloride derivative. On reduction with phosphorus and hydriodic acid
in a sealed tube at 280° it did not give products suitable for analysis. It is
unattacked by alkaline permanganate. Fuming nitric acid reacts with the
substance to yield a bright yellow nitro-compound which is soluble in water,
giving a yellow solution capable of dyeing silk. On the addition of alkali to
the yellow solution, the colour changes to deep red. Oxidation of the
substance with chromium trioxide and glacial acetic acid, and with potassium
bichromate, gave rise to brown amorphous substances. With concentrated
sulphuric acid no colour reaction is produced unless a trace of formaldehyde
be added. The purple colour produced under these conditions is identical
with that produced by the addition of a trace of glyoxylic acid to the
suspension of the substance in water, —

i

104 A. HOMER

A. The original crystalline compound prepared from solutions of
recrystallised tryptophane by using methods (a) and (0) was filtered from the
mother liquor, washed well with water, alcohol and ether, dried and analysed.

(1) Dried at the ordinary temperature in vacuo:

01600 g.; 0°3575 g. CO,, 0°0880 g. HO. C=60°94, H=6-11 °/).
0°1230 g.; 12 ¢.c. moist nitrogen at 15°C. and 768 mm. N=11-42 °%,.
C.,H,,0;N,.H,0 requires C=61-49, H=6-04, N=11-96 "/).

(2) Heated to a temperature of 155°:

0:1533 g.; 0°3730 g. COs, 0-0870 g. H,0. C=66-35, H=6-36 %J,.
0°1710 g.; 0°4180 g. COs, 0:0860 g. H,0. C=66-66, H=5-63 %/,.
0°1050 g.; 11-6 c.c. moist nitrogen at 15°C: and 768 mm. N=13-00°/,.
0:2080 g.; 21-2 c.c. moist nitrogen at 7° C. and 776 mm. N=12'90°/).
C,,H.,0,N, requires C=66°66, H=5°61, N=12°96 °/).

B. The compound was twice dissolved in ten per cent. sodium carbonate
solution and precipitated with dilute acetic acid. The constant melting
point 324° was obtained. The crystalline precipitate was washed well with
water, alcohol and ether, dried at various temperatures and analysed.

(1) Dried in vacuo at the ordinary temperature :

0-1200 g.; 0°2683 g. CO,; 0:0645 g. H,O. C=61:10, H=6-02 %/,.
01080 g.; 11:1 ¢.c. moist nitrogen at 22°C. and 767 mm. N=11'95 °%/,.
C.,H.,0;N,.H,O requires C=61:49, H=6-04, N=11°96 °/,.

(2) Dried at temperatures between 110° and 150° :

0-1187 g.; 0°2804 g. CO,; 0:0645 g. H,O. C=64:42, H=6-04 %/,.
0°1014 g.; 0-2395 g. CO,; 0°0549 g. H,O. C=64°41, H=6°05 %).
0°1252 g.; 0:2950 g. CO,; 0:0672 g. HO. C=64:3, H=6:01 %/,.
0°1914 g.; 20-9 ¢.c. moist nitrogen at 18° C. and 748 mm. N=12-51°/,.
0:1912 g.; 20 c.c, moist nitrogen at 11:5°C. and 757°2 mm. N=12°43°/,.
Co,H20;N, requires C=64:00, H=5'84, N=12-45 °/,.
(3) Heated to temperatures between 150° and 200°: .

0°1037 g. ; 0°2513 g. CO.; 0°0545 g. H.O. C=66:13, H=5'89 °/,.
Co,H2,04Ny requires C=66°66, H=5-61, N=12:96 °,.

It is thus evident that after solution of the ether oxidation substance in
alkali and subsequent precipitation by acidification, the product precipitated
has the same composition as when originally formed from the mother liquor.

It is of importance to note that the change in composition from
C.,H,O;N,.H,O (B (1)) to C,,H.,O,N, (B (2)) is of a reversible nature. That
is to say, the ether oxidation substance previously heated to 140° will readily
take up water again to give the monohydrate, But having effected the change

7
é
f
:
j

A. HOMER 105
from C,,H,O;N, (B (2)) to CH.,0O,N, (B (3)), it is not possible to get the

reverse reaction to take place. In other words, at temperatures above 150°
the ether oxidation substance loses the elements of water as the result of
some internal molecular rearrangement. The compound C,,H.,O,N; does not
undergo further change at temperatures from 200-220°, but it is decomposed
at 225°, and carbon dioxide, ammonia and skatole are evolved.

Il. The condensation of formaldehyde and tryptophane.

If a few cubic centimetres of formalin be added to a concentrated
solution of pure tryptophane, and the reaction mixture be kept at a
temperature of 38° for several hours, a crystalline derivative is deposited
from the solution.

The compound melts and decomposes at 225-240°; it is shghtly soluble
in water and alcohol, insoluble in benzene and its homologues, amyl alcohol,
chloroform, ethyl acetate and ethylene dibromide ; it is of a slightly yellow
colour which becomes intensified if the substance be kept at a temperature
of 150° for a short time, or if it be left exposed to the air at ordinary room
temperature for some days. On being boiled with water, dilute acids and
weak alkalis, it is readily hydrolysed, formaldehyde is liberated and the ether
oxidation product is formed. With concentrated sulphuric acid the substance
gives a purple colour, which is intensified if the suspension of the substance
in water be heated and cooled before adding the concentrated acid.

The condensation product was formed as follows:

(a) By the interaction of tryptophane and samples of formaldehyde
obtained (1) from Merck’s, (2) from Kahlbaum’s, and (8) from Schering’s
formalin, (4) by distilling Merck’s trioxymethylene in a current of hydrogen
or nitrogen and collecting the issuing gases in air-free water.

(6) The formaldehyde vapours evolved in (4) were passed into a
solution of tryptophane from which all air had been dispelled by a current of
hydrogen.

In each experiment a crystalline derivative melting at 235-240° was
formed,

On account of the hydrolysis which takes place with water, acids and
alkalis, the substance could not be purified by being dissolved in alkalis and
re-precipitated by acid. For purposes of analysis it was filtered from the
mother liquor, washed well with water, alcohol and ether, and dried at
different temperatures. Analyses of samples prepared at different times
and dried at the same temperature indicated constancy of composition.

106 A. HOMER

(1) Of the substance dried in vacuo :

0:1208 g.; 0°2537 g. CO,; 0:0700 g. H,O. C=57°30, H=6-49 .

0:°1135 g.; 0°2377 g. CO.; 0°0658 g. HO. C=57-'10, H=6°56 °/).

0:1340 g.; 0°2820 g. CO.; 0:°0763 g. H,O. C=57:33, H=6°38 %/).

01175 g.; 10°7 ¢.c. moist nitrogen at 16° C. and 760 mm. N=10°81 9/).

0°1512 g.; 14:4 ¢.c. moist nitrogen at 23°C. and 763-4mm. N=10°959/9.
Cy9H20.Ny.2H20 requires C=57:14, H=6°35, N=11-11%.

(2) Of the substance dried at temperatures between 110° and 130°:

01045 g.; 0°2355 g. CO,; 0-0600 g. H,O. C=61'45, H=6-43 °/,.

01135 g.; 0°2540 g. CO,; 0°0635 g. H.O. C=61:04, H=6°26 %/.

0-1265 g.; 12-7 ¢.c. moist nitrogen at 25°C. and 765mm. N=11°639/ .
C,H y205No . H,O requires C = 61°49, H = 6:04, N=11:96 9/o.

(3) The substance was heated to 160° for ten minutes, but no analyses
could be made owing to decomposition which began to take place ; there was
considerable carbonisation together with the elimination of skatole.

Analyses were made of the crystalline derivative formed by the hydrolysis
of the formaldehyde product :

(1) By water. The derivative melted at 324°, and when mixed with
some of the ether oxidation substance the melting point remained unchanged.
The substance was dried at 110°.

0:1127 g.; 0:2678 g. CO»; 0-0575 g. H,O. C=64-80, H=5-72%).
0°1055 g.; 0°2475 g. CO2; 0°0518 g. H,O. C=63°99, H=5-50 %,.

0°1064 g.; 0°2520 g. COy; 0°0572 g. HO. C=64:50, H=5:98 %/p.
0-1340 g.; 14°8 c.c. moist nitrogen at 19° C. and 768-4 mm. N=12-94°/,.

(2) By Na,CO, solution. The derivative melted at 324°.

(a) Of the substance dried in vacuo :

0°1028 g.; 0:2290 g. CO; 0°0565 g. HO. C=60-80, H=6-16 %,.

(b) Of the substance dried at 110°:

0°1000 g.; 0°2347 g. CO2; 0°0525 g. H,O. C=64:09, H=5:83 J).
0°1135 g.; 12:3 c.c, moist nitrogen at 21°C. and 768 mm. N=12-74°/,.

These results are in accordance with the assumption from the melting
point that the product of hydrolysis is the ether oxidation substance
(C,,H~0;N,.H,O and C,,H,,O0,N,).

It is thus evident that, although the product formed by the condensation
of formaldehyde with tryptophane is not the same as the ether oxidation
substance, yet there is a close connexion between these two which can be
demonstrated by the fact that on hydrolysis the former is readily converted
into the latter substance.

A. HOMER 107

IIT. The action of formaldehyde and hydrogen peroxide on tryptophane
gave rise to a crystalline compound which proved to be the formaldehyde
product II, and not the ether oxidation substance.

*

IV. The condensation of glyoxylic acid with tryptophane.

To an aqueous solution of tryptophane was added glyoxylie acid in the
proportion of one molecule of tryptophane to one molecule of glyoxylic acid.
A crystalline substance was deposited from the solution.

The substance was slightly soluble in water and alcohol, insoluble in
ether, petroleum ether, benzene and its homologues, amyl alcohol, ethyl
acetate and ethylene dibromide. It is soluble in alkalis, moderately soluble in
dilute acids, readily in concentrated acids. It is precipitated by the mercury
sulphate reagent. It gives the glyoxylic reaction and also the bromine
colour tests characteristic of tryptophane. It does not give any colour
reaction with concentrated sulphuric acid; on the addition of a trace of
formaldehyde or of glyoxylic acid solution, the glyoxylic colour is formed ;
while on the addition of a trace of hydrogen peroxide a claret colour is
obtained. |

The substance, whether separated from the mother liquors and repeatedly
washed with water, alcohol and ether, or recrystallised from hot water and
from alcohol, melted at 324°, and when mixed with some of the ether
oxidation substance the melting point was unaffected. That the derivative
under discussion was not the ether oxidation substance was shown by the
fact that when dissolved in sodium carbonate solution and the solution
acidified with acetic acid, the acid sodium salt was precipitated and not
the original condensation product. Analyses of the original acid (A), and of
the acid sodium salt (8) were made and the results obtained were in
accordance with the formulae C,,H,,O,N, for the acid and C,,H,,O,N.Na.4H,O
for the sodium salt.

A. Analyses of the original product (m.p. 324°).

(1) Of the substance dried in vacuum desiccator :
01238 g.; 0°2693 g. CO,; 0°0542 g. HO. C=59°34, H=4°89 %,.
01858 g.; 18 c.c. moist nitrogen at 21°C. and 777°3 mm. N=11:21°),.
(2) Dried at 150°:
0-1008 g.; 0°2193 g. CO,; 0°0433 g. H,O. C=59°35, H=4°82 °/).
0-2820 g.; 26-2 c.c, moist nitrogen at 20°C. and 757-2mm. N=10-70%/,.
(3) Heated to 200°:

0°1055 g.; 0°2287 g. CO.; 0°0465 g. HO. C=59-12, H=4:99.
(1), (2) and (3). Cy3Hy.0,N, requires €=59°97, H=4°66, N=10°77"/,.

108 A. HOMER

(4) Heated to temperatures between 205° and 220° for ten minutes,

carbon dioxide being evolved.
01184 g.; 0°2875 g. COz; 0:0575 g. H,0. C=66-23, H=5-44 %/,.
0°1525 g.; 16°8 c.c. moist nitrogen at 14°5° C. and 755°3 mm. N=12-94°/).
Cy4H2,04N, (ether oxidation substance) requires C=66:66, H=5°61, N=12°96 %),.

It is evident from the above analyses that at temperatures between
15° and 200° the glyoxylic condensation product has a constant composition ;
at 220° it is converted, with loss of carbon dioxide, into the ether oxidation
substance.

(5) If the glyoxylic compound be heated for a short time at a temperature
of 225-230° it decomposes, yielding carbon dioxide, ammonia and skatole
(cf. I, p. 105).

B. The acid sodium salt (see above) was twice recrystallised from water
and analysed.

0°3735 g.; 26-1 ¢c.c. moist nitrogen at 20°8° C. and 757°3 mm. N=8:08 °/).
0°1289 g.; 9 ¢.c. moist nitrogen at 21°5° C. and 761:9mm. N=8-02 %/,.
0°2048 g.; -0406 g. Na,SO,. Na=6°42 °,.

0°3700 g.; 0°0562 g. NagCO; and 0:0762 g. Na,SO,. Na=6-59 and 6-67 °/,.
0°8122 g.; 0°1542 g. Na.SO,. Na=6-15 °/,.

From these determinations the molecular weight of the substance = 355.

0:6074 g.; 0°4890 g. anhydrous residue,
. in the acid salt there are 3:84 mols. water.
0°1033 g.; 0°0528 g. H,O; 0:1612 g. COs.
Putting in the correction for the NayCO3 remainin
Oe (00150 g.) 24 + Oe stat ae
C13Hy,04N.Na.4H20 requires N=7-91, C=44-07°/,.
. H=5'37, Na=6'50 %,.
M.W. 354:3.

From these results it is obvious that the glyoxylic condensation product
and the ether oxidation substance are two distinct compounds, in spite of the
fact that the compounds when taken separately and also when mixed
together have the same melting point, viz. 324°. It is obvious that the
glyoxylic compound readily changes to the ether oxidation substance on
being heated to a temperature of 205°. (The ether oxidation substance is
not affected under the same conditions.) At 225° both are decomposed into
carbon dioxide, ammonia, skatole, ete. .

There is thus an interesting connexion between the formaldehyde
compound, the glyoxylic condensation product, and the ether oxidation
substance. The last-named substance can be produced from the formaldehyde
product by hydrolysis, and from the glyoxylic condensation product by the
application of heat.

—s ee. i

A. HOMER 109

It is probable from the following reasons that these compounds described
under I, IT and IV are formed by the simple process of condensation
between the —NH, group of tryptophane and the —CHO group of the
reacting aldehydes :

1. The ease with which the reactions take place (the presence of a
condensing agent is unessential).

2. The failure to produce naphthalene sulphochloride derivatives points
to the substitution of the hydrogen of the amino-groups.

3. The fact that the compounds so obtained are colourless points to the
condensation having taken place as a result of the activity of the amino- and
not the imino-group (post, p. 112).

4. Corresponding compounds could not be obtained with indole, skatole,
indoleacetic and indolepropionic acids, i.e. with indole derivatives in which
there is no amino-group.

For reaction to take place with these compounds, the presence of a
condensing agent is necessary, and the resulting products are intensely
coloured (post, p. 112).

Analyses of the three compounds I, II, IV are in accordance with the
following formulae :

(a) The formaldehyde product is ar Cy2H,0,N, .2H,0.
(b) The glyoxylie product... oie oe C3Hy204No.

i ie 3, above 205° ie CoyHoO4Ny.
(c) The ether oxidation substance roe Co4Hog0;Nq.

9 9 99 ” above 150° Co,H.,04Ny °

The mechanism involved in the production of these three substances may
be briefly described as follows :

I. Since the ether oxidation product is in no way an oxidation product
of tryptophane, but is closely connected with the formaldehyde condensation
product, it is suggested that in the formation of the ether oxidation substance
nascent formaldehyde is the active principle (p. 102, I c).

COOH HOOC
R.CH,.CH(NH»). COOH + 2HCHO (nascent) +R .CH,.CH.NH.CH,.O0.CH,.NH.CH.CH)R

(where R represents the indole nucleus CgHgN).

On heating this substance to a temperature of 150-200", molecular water
and not water of crystallisation is lost. After such a process the composition
of the substance corresponds to the formula C,,H.,0,N,, and can be repre-
sented graphically as

COOH

> chiata
cH. NH.CH=CH .NH.CH

R.CH, \ cH: . B

110 A. HOMER

II. The formaldehyde condensation product is formed as follows :

/ 00H

R.CH,.CH(NH,). COOH + HCHO=R.CH,.CH +H,0

\\N : CHe

III. The glyoxylic condensation product is produced in a similar manner :

COOH
VA

R.CH,.CH (NH2). COOH +COOH .CHO=R.CHp. Cae H,0

+
N:CH.COOH

It has been shown that the ether oxidation substance is readily formed
(a) from the formaldehyde condensation product by hydrolysis, (6) from the
glyoxylic condensation product by the application of heat. For the former
process (a) there are two possibilities to be considered, viz. :

(1) That during hydrolysis the molecules are linked together by means
of the addition of the elements of water, thus:

COOH COOH

/ 00H
+H,0=R.CH,.GH.NH.CH,.0.CH,.NH.CH.CH,.R

\N | CHp
(2) That during the hydrolysis tryptophane and nascent formaldehyde

2k. CH,.CH

are re-formed, and that these react as in the ether oxidation process to form
the substance C,,H,,0,N,.

(b) In the case of the production of the substance C,,H,,0,N, from the
glyoxylic product by the application of heat, it is evident that two molecules
of the latter each lose a molecule of carbon dioxide, and at the temperature
at which the reaction takes place condensation of these two molecules occurs,
accompanied by a rearrangement of the bonds into the more stable configura-
tion of the ether oxidation substance.

2R.CH,.CH(COOH).N:CH.COOH
=R.CH,.CH(COOH).NH.CH:CH.NH.CH(COOH).CH,.R+2C0,.
A study of the properties of these three compounds throws some light on
the mechanism of the Adamkiewicz reaction: a discussion of this point will
be dealt with in the following paper [Homer, 1913, p. 117, Table I].

SEcTION B.

The condensation of glyoxal with tryptophane.

If an aqueous solution of tryptophane heated to a temperature of 38° be
treated with the calculated quantity of glyoxal in the presence of hydrogen
peroxide, a crystalline chocolate-coloured deposit slowly forms.

This coloured substance is of an acidic nature and is slightly soluble in
water and in alcohol. It dissolves in alkalis, and is moderately soluble in

A. HOMER 111

dilute acids, and with the exception of acetic acid readily soluble in con-
centrated acids. It is precipitated by the mereury sulphate reagent. With
concentrated sulphuric acid it gives a brown colour which in the presence
of hydrogen peroxide changes to claret. In the presence of a trace of
formaldehyde, a brown colour is produced which is changed to the colour of
bromine on the addition of an oxidising agent.

Analyses of the substance gave the following results :

(1) Of the substance recrystallised from water and

(a) Dried at 110°:
0°1030 g.; 0°2325 g. COs; 0°0463 g. H,O. C=61-60, H=5-04 9.
0-1223 g.; 0-2780 g. CO2; 0°0560 g. H,0. C=61-99, H=5-13 %/,.
01025 g.; 02346 g. CO.; 0-0485 g. H,0. C=62-40, H=5-25 %/,.
0-2268 g.; 21-6 c.c. moist nitrogen at 24°C. and 767°5 mm. N=10-9°%,.

CogH220gNy.H,O requires C=61:90, H=4-81, N=11:1%,.

(b) Heated to 130° :
0°1006 g.; 0°2360 g. CO.; 0°0412 g. H,O. C=64:00, H=4-62 of pe
0°1080 g.; 10°7 ¢.c, moist nitrogen at 21°C. and 768 mm. N=11-51 °,.

(2) After solution in Na,CO, and precipitation by acetic acid:

01430 g.; 0°0565 g. H,O; 0°3403 g, CO,. C=64-90, H=4-43 9,
0°1060 g.; 10°2 c.c. moist nitrogen at 20°C. and 769 mm. N=11-24 de

(3) Crystallised from alcohol and heated to 110°:

01017 g.; “0441 g. H.O; 2390 g. CO,. C=64-15, H=4-86 %/,.

0°1290 g.; 12°5 c.c. moist nitrogen at 19°C. and 750°5 mm. N=11-1°/,,
(4) Crystallised from pyridine :

0°1060 g.; *0441 g. H.O; 0°2505 g. CO,. C=64:45, H=4-63 9/,.

Ib. 2, 3, a :—CogHo20gNy requires C=64°2, H=4:53, Nes od

(5) Heated to 160°:

0:1006 g.; 0-0410 g. HO; 0-2459 g. CO.. C=66-66, H=4°53 9.

0°1150 g.; 0:0460 g. H20; 0°2830 g. CO,. C=67:07, H=4:46 %,.

0°1890 g.; 19°6 c.c. moist nitrogen at 16°C. and 749°6 mm. N=12-01°),.
CogH205N, requires C=66°66, H=4:32, N=11:97 °/,.

Taking into consideration the necessity for the presence of an oxidising
agent, and also the fact that the substance produced is intensely coloured, it
is highly probable that in this reaction, besides the simple aldehyde con-
densation occurring in the reactions described in Section A, there has also
been elimination of hydrogen accompanied by complex ring formation.

Possible equation :

2C),;Hj,02N2 + 2 (CHO). + O = CogHo20gNy. HO + 2H20.

Any attempt to ascertain the constitution of this substance has been

postponed for the present.

112 A. HOMER

Secrion C.

THE PRODUCTS FORMED BY THE CONDENSATION OF ALDEHYDES
WITH THE ImINo-GRouP OF INDOLE DERIVATIVES.

I. The action of formaldehyde on indole derivatives in presence of a
condensing agent.

In connexion with the work on the condensation of formaldehyde with
tryptophane, an attempt was made to prepare corresponding derivatives with
indole, skatole, indolepropionic and indoleacetic acids, but without success.
However, it is well known that under suitable conditions, in the presence of
concentrated sulphuric acid, a solution of formaldehyde reacts with the above-
mentioned compounds to yield coloured substances. Concentrated sulphuric
acid being known to be both an oxidising and a condensing agent, the question
arose as to whether the colour effect was due to a process of simple condensa-
tion between formaldehyde and the above-mentioned substances, which have
an imino-group in common, or was in any way dependent upon the oxidising
power of the acid.

In order to test this suggestion, it was thought advisable to avoid the use
of sulphuric acid and attempts were made to form condensation products of
indole derivatives with trioxymethylene, using zine chloride and alcohol
saturated with hydrochloric acid gas as the condensing agents (with these
condensing agents an aqueous solution of formaldehyde could not be used).

II. Derivatives formed by the interaction of trioxymethylene and indole
derivatives, using zinc chloride as the condensing agent.

Indole, and indoleacetic acid reacted readily at temperatures below 100°
to produce purple-coloured products. Skatole and indolepropionic acid under
the same conditions gave brown products. Owing to scarcity of material,
the products from indoleacetic and indolepropionic acids could not be
analysed. The indole and skatole derivatives were investigated.

Tryptophane treated in the same way did not react until heated to a
much higher temperature (160°); the resulting product was of a deep brown
colour. The reaction was accompanied by a certain amount of decomposition
of tryptophane. As it was impossible to free the product of the reaction
from the results of carbonisation, no further investigation of this substance
has been made.

A. HOMER 113

Preparation of the triorymethylene derivatives of skatole and indole. One
part by weight of indole or skatole was mixed with the same weight of
trioxymethylene and ground together with one part by weight of freshly
powdered and freshly ignited) zinc chloride. The process of grinding the
reaction mixture gave rise in the former case to the formation of a purple
colour. The reaction mixture was heated on a water-bath to a temperature
of 80° for half an hour, and then transferred to an oil bath and the tempera-
ture raised to 150° for a few minutes. On cooling, the coloured mixture was
treated with hydrochloric acid solution (one part concentrated acid to one part
water) and allowed to stand for some hours. The residue was in each case
repeatedly boiled with dilute hydrochioric acid, and then with water until
free from formaldehyde and hydrochloric acid. The coloured substance was
finally washed with water, alcohol and ether, dried and analysed.

It was impossible to purify the products further on account of their
insolubility in all liquids other than concentrated sulphuric acid. Constancy
of composition was denoted by analyses of samples prepared at different
times and under slightly different conditions as regards the proportion of zine
chloride used. It is interesting to note that the indole compound is of a
purplish colour, the skatole compound is brown. They are both insoluble in
ordinary solvents, in alkalis and in acids with the exception of concentrated
sulphuric acid, in which they are appreciably soluble, and concentrated
hydrochloric acid in which they are slightly soluble. In concentrated
sulphuric acid the indole derivative gives a reddish violet colour which on
dilution becomes bluish and finally disappears owing to the decreasing
solubility of the substance as the acid becomes more dilute. The skatole
compound under the same conditions gives rise to a brown colour.

(1) Analyses of the indole derivative prepared on different occasions
were made}.

A. 0°1026 g.; 0:0475 g. H,O; 0-2702 g. CO,. C=71°85, H=5-20"/,.

B. 0°1057 g.; 0°0518 g. H,0; 0-2783 g. CO,. C=71-80, H=5-49 %/,.

A. 0°1325 g.; 9°4¢.c. moist nitrogen at 14° C. and 753°5 mm. N=8'32 9/).
0°1500 g.; 11°0c¢.c. moist nitrogen at 15°5°C. and 747°3mm. N=8'34"/).

B. 0:2074 g.; 14:2c¢.c. moist nitrogen at 13° C. and 755°3 mm. N=8:07 °/,.

Analysis of the indole derivative after being heated to 205° for fifteen

minutes gave the following result :
0-1045 g.; 0°0502 g. HO; 0:2750 g. COp. C=71'78, H=5:38 "/).
Hence the compound is stable to heat at temperatures not exceeding 205°.
Cx;H»03Ny requires C=72°41, H=5°74, N=8:04 "/,.

! These compounds are extremely difficult to burn and this probably accounts for the slightly
low values obtained.

114 A. HOMER

(2) Analyses of the skatole derivative gave the following results!:

0°1161 g. ; 0-:0658 g. H,O; 0°3120 g. COy. C=73-30, H=6°35 °/,.

01083 g.; 0°0575 g. H20; 0:2910 g. COg. C=73':30, H=5-94 9).

0°1685 g.; 9°8 c.c. moist nitrogen at 13°5° C. and 758°5 mm. N=7-13°/,.

0°1820 g.; 10-8 c.c. moist nitrogen at 11°5° C. and 765mm. N=7-15°/,.
Co3H403No requires C=73:-40, H=6:38, N=7:°42 Wi

The condensations may therefore be represented by the equations:
2C3H,N + 5HCHO= Co1;H903Ne + 2H.O
2CyoHoN + 5HCHO = Co3H9,03No+ 2H20.

It is thus obvious that in the presence of a condensing agent trioxy-
methylene reacts with the imino-group of the indole nucleus to form
coloured bodies.

Il. Derivatives formed by the interaction of triovymethylene and indole
derivatives, using a solution of hydrochloric acid gas in alcohol as the
condensing agent.

A description of the condensation products with indole, skatole and
tryptophane formed under these conditions will form the subject-matter of a
future paper. It may be stated here that they are intensely coloured
substances, and that the indole derivative is the only one of the three which
will give a purple colour with concentrated sulphuric acid.

The derivatives formed by the action of trioxymethylene on indole
derivatives in the presence of a condensing agent, whether it be concentrated
sulphuric acid, or ‘zine chloride, or an alcoholic solution of hydrochloric acid
gas, are all characterised by their insolubility in solvents other than
concentrated mineral acids.

A discussion as to the significance of activity of the imino-group of
indole derivatives in the production of the specific colour reaction known as
the Adamkiewicz reaction, will be found in the paper following this [Homer,
1913, p. 116].

During the course of the work under consideration it has been noticed
that tryptophane itself under certain conditions of oxidation gives rise in
the process of time to the formation of pigments having definite absorption
bands. The properties both chemical and physical of these pigments,
and of those formed from normal breakdown products of tryptophane, are
being investigated, and will form the subject of a future paper. At this

' These compounds are extremely difficult to burn and this probably accounts for the slightly
low values obtained.

A. HOMER 115

point it may be observed that in nearly every case these substances are
themselves brown in colour, and their solutions in organic solvents are also
brown, but the presence of acid changes the colour from brown to purple.

SUMMARY.

1. Tryptophane by virtue of its — NH, group reacts directly with
nascent formaldehyde, formaldehyde and glyoxylic acid to form colourless
erystalline compounds.

2. Glyoxal will react with tryptophane in the presence of an oxidising
agent to form an intensely coloured compound.

3. Indole derivatives, by virtue of the - NH group in the nucleus, will
react with formaldehyde and trioxymethylene in the presence of a condensing
agent to form substances of intense colour and marked insolubility in ordinary
solvents other than concentrated mineral acids.

In conclusion, the author wishes to thank Dr Hopkins for his unfailing
interest in the work and for the valuable advice given during the progress of
the investigation.

REFERENCES.
Ellinger (1906), Ber. 39, 2, 2515.

Homer (1913) Biochem. J. 7. 116.
Hopkins and Cole (1902), J. Physiol. 29, 464.

XI. ON THE COLOUR REACTIONS OF CERTAIN
INDOLE DERIVATIVES, AND THEIR SIGNIFI-
CANCE WITH REGARD TO’ THE GLYOXVIUIG
REACTION.

By ANNIE HOMER, Beit Memorial Fellow.

From the Physiological Laboratory, Cambridge, and the Institute of
Physiology, University College, London.

(Received November 26th, 1912.)

In the preceding paper [Homer, 1913] the preparation and properties of
several new indole derivatives have been described in detail.

The condensation products formed between tryptophane and nascent
formaldehyde, formaldehyde and glyoxylic acid respectively in the absence of
a condensing agent, are colourless crystalline compounds. The reaction takes
place between the —CHO group of the aldehyde and the — NH, group of
tryptophane.

In the presence of condensing agents, such as zinc chloride or an alcoholic
solution of hydrochloric acid gas, indole derivatives and pyrrole, by virtue of
the —NH group in the nucleus, combine with trioxymethylene to form
coloured substances; in the presence of concentrated sulphuric acid a
solution of formaldehyde reacts with indole and pyrrole derivatives to give
colour reactions.

In the case of the condensation of glyoxal with tryptophane, the presence
of an oxidising agent is necessary, and the resulting product, which is
coloured, no longer gives the glyoxylic reaction.

The results of an investigation of the colour reactions of the above-
mentioned substances have been embodied in the following tables, and afford
special interest with regard to the mechanism of (a) the Adamkiewicz
reaction for the detection of proteins and (b) Hopkins and Cole’s glyoxylic
reaction,

A. HOMER

SECTION LI.

117

(In the following reactions, air-free water has been used in every case.)

TABLE 1. Colour reactions produced by the action of pure concentrated sulphuric
acid on the formaldehyde, and glyoxylic acid condensation products of trypto-
phane, on the ether oxidation substance and on tryptophane.

Action of concentrated
sulphuric acid

A. On the solid:

i. In the cold
ii. On warming

Tryptophane

None

Slight darken-
ing. Finally
charring

B. Onthe solid suspended in water :

i. Directly treated with the acid

ii, Heated to 90°, cooled and
then treated with the acid

ili. In the presence of one drop
of a 1/200 solution of formalin
in water

iv. In the presence of a trace
of glyoxylic acid

v. In the presence of a trace
of ferric chloride

vi, In the presence of a trace
of hydrogen peroxide

None
None

Brown

Purple
Orange red

Orange red

Formaldehyde
product

Purple!

Colour intensi-
fied. Finally
charring

Purple

Purple colour
intensified

Purple colour
intensified

Purple
Very intense

purple
Brown

Glyoxylic
condensation
product

Yellow

Colour intensi-
fied, Finally
charring

Yellow
Yellow

Purple

Purple
Brown

Claret

Ether
oxidation
substance

None

Slight purple
tint

None
None

Purple

Purple
Green

Green

1 The term ‘‘ purple” has been used for colours that are red with a tinge of blue in them, or

vice versa.

TABLE II. The colour reactions produced by the action of concentrated sulphuric
acid on pyrrole and indole derivatives im presence of glyoxal.
Action of concentrated Indole-pro- Trypto-
sulphuric acid Pyrrole Indole Skatole pionic acid phane
In the presence of :
i. A trace of glyoxal Reddish Indian red Burnt Reddish Claret
brown Sienna brown
to which hasb een added a
trace of : Deep orange
(a) Ferric chloride __... . whichrapid- Claret
| Greenishtint Greenish tint ca lychangesto _ colour in-
(b) Hydrogen peroxide... ) rown purple mad- _ tensified
der tint
ii. An infinitesimal trace Yellow Very delicate Reddish Yellowish Russet
of glyoxal brown shade of In- brown brown witha
to which has been added a dian red tinge of red
trace of :
(a) Ferrie chloride Intense blue Green Reddish \ Orange red
changing to brown | whichrapid- Russet _
bluish pur- ly changes colour in-
ple to a purple _ tensified
(b) Hydrogen peroxide... Green Green Magenta / madder
8

Bioch, vir

118 A. HOMER

TABLE III. The colour reactions produced by the action of pure concentrated
sulphuric acid on pyrrole and various indole derivatives in the presence of

(A) formaldehyde, and (B) glyoxylic acid.
Glyoxylic Ether

Indole- conden- oxida-

Action of concentrated propio- Trypto- sation tionsub-
sulphuric acid Pyrrole Indole Skatole nicacid phane product stance
A. i. In the presence of 1 Red Purple Brown Brown Brown’ Purple Purple

drop of a solution of 1¢.c. of brown
formalin in 10 c.c. of water

ii. Inthe presence ofldrop Hed Purple Brown Brown Brown Purple Purple
of a 1/200 solution of for- brown
malin in water

to which has been added a Green Purple Purple Purple Purple Purple Purple
trace of an oxidising agent

iii. Inthepresenceofldrop Yellow- Purple Purple Purple Purple Purple Purple
of a 1/2000 solution of for- —_ ish
malin in water brown

iv. Inthepresenceofldrop Red Purple Brown Brown Brown Purple Purple
ofasolutionofformaldehyde brown ;
made by dissolving 1 part of

trioxymethylene in 1000

parts boiling water

v. In presence of 1 drop Yellow- Purple Purple Purple Purple Purple Purple
of the solution used in iv, ish
diluted ten times brown

B. In the presence of gly- Red Purple Purple Purple Purple Purple Purple
oxylic acid brown

TABLE IV. The colour reactions produced by the direct action of pure concen-
trated sulphuric acid on the trioxymethylene condensation derivatives of idole,
skatole and tryptophane.

(a) Formed in the presence of zinc chloride as condensing agent.

(b) Formed in the presence of an alcoholic solution of hydrochloric acid gas.

Action of concentrated Indole Skatole Tryptophane 4
sulphuric acid derivative derivative derivative 7

(a) On the solid Purple Brown Brown
(b) On the solid Purple Brown Brown 4a

Tables Va and Vb illustrate the masking effect due to the presence
of trioxymethylene,

A. HOMER 119

TABLE Va. Colour reactions produced by the action of pure concentrated
sulphuric acid on aqueous solutions or suspensions of indole derivatives.

Glyoxylic
“ condensation
product
‘ Indole- and ether
Action of pure concentrated propionic ‘Trypto- oxidation
sulphuric acid Indole Skatole acid phane substance
i. In the presence of 1 drop of a Purple Brown Brown Brown Purple

1/200 solution of formalin in water

to which has been added an oxi- Purple Purple Purple Purple Purple
dising agent

ii. In the presence of a trace of Purple Brown Brown Brown Purple
trioxymethylene
iii. In the presence of 1 drop of a Purple Brown Momentary Brown Purple
1/200 solution of formalin in water, appearance
an oxidising agent, and a trace of of purple
trioxymethylene, cf. i (above) colour, im-
mediately ~
masked by
brown colour

TABLE Vb. Colour reactions produced by the action of concentrated sulphuric
acid, in which has been dissolved a trace of trioxymethylene, on aqueous
solutions or suspensions of indole derivatives.

Glyoxylic
condensation
product
Action of concentrated Indole- and ether
sulphuric acid, containing propionic Trypto- oxidation
a trace of trioxymethylene Indole Skatole acid phane substance
i. Directly on the aqueous solu- Purple Brown Brown Brown Purple
tion or suspension
ii. In the presence of an oxidis- Purple Brown Momentary Brown Purple
ing agent appearance
of purple,
very soon
masked by
predominat-
ing brown
iii. In the presence of an oxidis- Purple Brown Momentary Brown Purple
ing agent and 1 drop of a 1/200 appearance
solution of formalin in water of purple,
very soon
masked by
predominat-
ing brown

Adamkiewicz [1874] demonstrated that a violet colour is produced by the
action of concentrated sulphuric acid on an acetic acid solution of protein.
Hopkins and Cole [1900; 1902] showed that the protein reacts by virtue of
its tryptophane groupings, and that pure acetic acid does not give the colour
test. As a result of their investigations they were led to the conclusion that
the reactive impurity in the acetic acid was glyoxylie acid.

Therefore, in the Adamkiewicz reaction, Hopkins and Cole regarded

s—2

120 A. HOMER

glyoxylic acid as the substance essential to the formation of the characteristic
violet colour. Rosenheim [1906] however, points out that the colour produced
by the action of formaldehyde on proteins in presence of sulphuric acid and
oxidising agents [ Voisenet 1905] is identical with that produced in the Adam-
kiewicz reaction. He criticises Hopkins and Cole’s view as to the importance
of glyoxylic acid per se; he suggests that even with glyoxylic acid an
oxidising agent is essential, and that although the latter may not have been
added, yet the glyoxylic acid itself, however carefully prepared, is always
contaminated with sufficient hydrogen peroxide for the reaction to take place.
On the other hand Dakin [1909], noticing that the colours produced by the
action of glyoxylic acid and formaldehyde. are identical, suggests that in the
case of formaldehyde, under the influence of acids, an aldol condensation and
subsequent oxidation to glyoxylic acid take place.

From Table I it is obvious that formaldehyde per se plays an important
part in the colour reaction, since of the three condensation products, the only
one which will give the colour without the addition of glyoxylic acid or
formaldehyde, is the formaidehyde condensation product. Further it has
been shown that this product is readily hydrolysed with liberation of
formaldehyde ; at the line of junction of concentrated acid and water such
hydrolysis would take place. A proof of this assumption lies in the
observation that after boiling this product with water and so ensuring
hydrolysis, the colour reaction is still more marked (Table I, B ii).

Were the essential factor glyoxylic acid per se, the glyoxylic condensation
product ought to give the colour reaction with sulphuric acid, but it will not
do so unless formaldehyde or more glyoxylic acid be added. Now glyoxylic
acid itself is decomposed by sulphuric acid with the formation of carbon
dioxide and formaldehyde’. It therefore seems reasonable to assume that
glyoxylic acid is able to take part in the colour reaction by virtue of its
decomposition into formaldehyde, and not, as Rosenheim suggests, because
of contamination with hydrogen peroxide. The fact that the colour reaction
is not given by glyoxal (Table Il) and glycollic aldehyde, neither of which
gives formaldehyde on treatment with acid, is in support of this view.

On the other hand, if formaldehyde be the essential factor, then some
explanation must be offered to account for the generally acknowledged
uncertainty of the colour reaction when that reagent is used, whereas with
glyoxylic acid there is no difficulty experienced in carrying out the test for
indole derivatives.

1 A small quantity of glyoxylic acid was heated with 5 c.c. of water and 5 c.c. of concentrated
sulphuric acid, The gases evolved were found to contain formaldehyde and carbon dioxide.

A. HOMER 121

According to Hopkins and Cole, Dakin, and Rosenheim, the presence of
an oxidising agent is necessary for the Adamkiewicz reaction to be carried
out with formaldehyde.

In the present investigation samples of pure sulphuric acid (Kahlbaum
and others) were specially procured and carefully selected so that they did
not induce a purple colour with tryptophane and a 1/500 solution of formalin
in air-free water. The addition of a trace of ferric chloride solution to the
reacting liquids induced the purple colour to appear. Samples of sulphuric
acid contaminated with a trace of an oxidising agent caused the appearance
of the purple colour with the use of a 1/100, and in some cases a 1/10 solution
of formalin in water.

A study of Table III shows that, using a 1/200 solution of formalin in
water, without the addition of an oxidising agent, indole, indolealdehyde,
indoleacetic acid, and the condensation products of tryptophane gave a purple
colour, skatole, tryptophane and indolepropionic acid gave brown colourations.
The addition of a trace of an oxidising agent to the latter was necessary for
the production of the purple colour. In this connexion the following
observations have been made:

1. In applying the formaldehyde test to tryptophane and _ skatole, if
previous to the addition of the above-mentioned formaldehyde solution a
trace of indolealdehyde be added, then the presence of the concentrated
sulphuric acid will induce the formation of the characteristic purple colour,
even though no oxidising agent has been added. This experiment shows
that for the production of the purple colour the indole derivative and not the
formaldehyde is attacked by the oxidising agent.

2. If to an aqueous solution of tryptophane one or two drops of formalin
be added, the liquid boiled, cooled, diluted and treated with concentrated
sulphuric acid, then the characteristic purple is produced, but is soon masked
by the brown colouration usually noticed.

In this experiment the presence of one of the condensation products
of tryptophane which gives a purple colour with formaldehyde has been
assured.

In order to ascertain whether this purple colour reaction was due to an
aldol condensation and oxidation of formaldehyde to glyoxylic acid [Dakin,
1909], the formalin itself was boiled, cooled, diluted to 1/200 and added
to a cold aqueous solution of tryptophane. There was no production of a
purple colour.

It will be seen from Table III, A iii and v that if the formaldehyde be
diluted to the order of 1 part to 2000 or 20,000 parts of water, then the

122 A. HOMER

purple colour is formed with skatole, tryptophane and indolepropionic acid
without the addition of an oxidising agent. This result was obtained with
Kahlbaum’s and Merck’s formalin, and also with a solution of formaldehyde
made by dissolving trioxymethylene in boiling water which had been
previously boiled for some time to expel air. The substances to be tested
were dissolved in air-free water. Even with such large dilution, there is
in the case of tryptophane a certain amount of brown colouration produced,
but not sufficient to mask the purple colour. As the concentration of
the formalin is increased, so the brown colour preponderates and masks
the purple.

These experiments, taken in conjunction with the fact that the formalde-
hyde condensation product is the only one which gives the Adamkiewicz
reaction without the addition of glyoxylic acid or formaldehyde, seem to be
conclusive as to the importance of formaldehyde per se as a factor in this
reaction. But at the same time, in performing the Adamkiewicz colour
reaction the fact that concentrated sulphuric acid causes the condensation of
formaldehyde to trioxymethylene must be taken into consideration. In the
previous paper it has been shown that trioxymethylene, in the presence
of a condensing agent, reacts with indole and indoleacetic acid to give
purple derivatives; with skatole, indolepropionic acid and tryptophane to
give brown coloured substances. In the former cases, therefore, the
secondary action with trioxymethylene will not interfere with the purple
colour formation, but in the latter cases the brown colour due to the formation
of trioxymethylene derivatives will preponderate. The correctness of this
assumption is demonstrated by the results described in Tables Va and V 0.
It will be noticed that a trace of trioxymethylene interferes with the
appearance of the purple colour which would normally be produced under
the conditions of the experiment (V aii and Vb ii). The masking effect
due to the formation of trioxymethylene by the action of concentrated
sulphuric acid on the formaldehyde may be obviated:

1. By using the formaldehyde solution excessively dilute, whereby the
formation of trioxymethylene is delayed, and the brown colouration due to
the condensation of certain indole derivatives with trioxymethylene does not
interfere with the Adamkiewicz test.

2. By the addition of a trace of an oxidising agent, whereby the
formation of an indole derivative, which will give the colour test, is induced.

3. In the case of tryptophane, by ensuring the previous formation of one
of the three condensation products described above (p. 121).

4. By the use of glyoxylic acid as the source of formaldehyde instead

A. HOMER 123

of formaldehyde itself (Hopkins and Cole’s modification of the Adamkiewicz
reaction).

The author suggests that at the line of junction of the concentrated
sulphuric acid and the aqueous layer containing a trace of glyoxylic acid and
the indole derivative, decomposition of the glyoxylic acid takes place with
the liberation of formaldehyde. The vivid colour reaction may be assigned
to one of the following causes :

1. The amount of formaldehyde produced under the conditions of the
experiment is never more than a trace, and the corresponding trioxymethylene
formation does not interfere with the colour reaction (Table III).

2. The formaldehyde is liberated in the nascent state, and is therefore
intensely reactive.

Either of these explanations can be offered to account for the vivid colour
reaction when the formaldehyde condensation product of tryptophane is acted
upon by concentrated sulphuric acid without the previous addition of
formaldehyde, glyoxylic acid or an oxidising agent. But that the nascent
condition of the formaldehyde is not essential to the colour reaction is
deduced from the fact that the addition of a 1/10 solution of formalin in
water to the ether oxidation substance or the glyoxylic condensation product
gives rise in the presence of concentrated sulphuric acid, to the formation of
a purple colour.

3. In the glyoxylic reaction with tryptophane, the colour reaction is
probably influenced by the action of the formaldehyde, liberated from the
glyoxylic acid, on the glyoxylic condensation product formed with tryp-
tophane.

Section II.

The following experiments show that the colour reaction induced by the
action of concentrated sulphuric acid on solutions of indole derivatives to
which formaldehyde and an oxidising agent have been added, is of a much
more complex nature than when formaldehyde alone is used.

In Section I (p. 121) it was demonstrated that the so-called “ glyoxylic ”
colour reactions could be obtained by the action of concentrated sulphuric
acid on solutions of tryptophane and skatole to which had been added traces
of indolealdehyde and formaldehyde. Therefore in the Adamkiewicz test the
part played by the oxidising agent is to convert the tryptophane and skatole
into some indole derivative which will react with formaldehyde to form
compounds capable of giving purple colour reactions with sulphuric acid and
not to oxidise the formaldehyde as Dakin suggests.

A study of the colour reactions of indolealdehyde was made :

124 A, HOMER

Colour reactions induced by the action of concentrated sulphuric acid
on solutions containing :

(1) Indolealdehyde.

(2) Indolealdehyde+one drop of a 1 /100
solution of formalin.

(3) Indolealdehyde+trace of trioxymethy-
lene.

(4) Indolealdehyde +a trace of an oxidising
agent.

(5) Indolealdehyde + tryptophane.

(6) Indolealdehyde + tryptophane + a trace of
formaldehyde (1/100).

(7) Indolealdehyde + tryptophane + an oxi-
dising agent.

(8) Indolealdehyde + tryptophane + a trace of
trioxymethylene.

(9) Indolealdehyde + skatole.
(10) Indolealdehyde + skatole +a trace of for-
maldehyde (1/100).

(11) Skatole+a trace of an oxidising agent +
a trace of formaldehyde (1/100).

Colour produced.

Pink. Colour becomes more intense on warm-
ing [Ellinger and Flamand, 1909].

Reddishiparnios
pomp ga | The addition of an oxidis.

ing agent causes the colour
” » | to become more blue.

Bluish magenta.

Orange red colour.

“ Glyoxylic” colour at the line of junction of

- the acid and aqueous solution. Above this
is seen an orange red colour.

Magenta.

“‘ Glyoxylic”’ colour as in (6) but it is rapidly
masked by the brown colour characteristic
of the action of trioxymethylene on trypto-
phane.

Reddish pink colour.

At the junction of the acid and solution a
reddish purple colour is obtained (as in (11)).
Above this layer the colour is reddish pink.

Reddish purple.

If the indolealdehyde in (10) be present in excess then the colour
produced is practically that produced in (9), the purple being masked.

It is obvious from these experiments that where colour reactions are
produced from tryptophane in the presence of an oxidising agent the
secondary reactions with indolealdehyde have to be considered, and in these
cases the actual shade of colour will depend upon the conditions of the
experiment.

From (6) it will be seen that the “glyoxylic” colour is produced by the
action of formaldehyde (without the addition of an oaidising agent) on the
It may be that the indole-
aldehyde reacts with the amino-group of tryptophane to form a substance
which then combines with formaldehyde as did the three condensation
products (Section I, Table I).

An inspection of reactions (4) and (7) shows that in the presence of an
oxidising agent concentrated sulphuric acid causes indolealdehyde to condense
with more indolealdehyde (4) or with tryptophane (7) to form substances
giving characteristic colour reactions which are not those of the “glyoxylic”
reaction.

mixture of tryptophane and indolealdehyde.

A. HOMER 125

Although the chemical reactions involved in the actual process of the
purple colour formation are at present unsolved yet it is clear that the
condensation of indole nuclei is an essential feature of the reaction. In the
case of those compounds which give the purple colour directly with formalde-
hyde the condensation takes place through the agency of the formaldehyde
alone. But for those substances for which, on account of interference with
the colour reaction due to the formation of trioxymethylene, an oxidising
agent is used, the reaction also involves the condensation of indolealdehyde
with the original indole derivative taken.

SUMMARY.

The formation of coloured condensation products from indole derivatives
and certain aldehydes necessitates the use of condensing agents. The
reaction takes place between the — NH group of the indole nucleus and the
— CHO group of the aldehyde.

The evidence adduced in this paper is favourable to the Adamkiewicz
reaction being primarily a formaldehyde reaction.

1. The formaldehyde condensation product of tryptophane gives the
colour test directly with concentrated sulphuric acid, without the addition of
an oxidising agent or of glyoxylic acid or of formaldehyde.

2. A trace of formaldehyde, without the addition of an oxidising agent,
gives the colour reaction with indole derivatives.

3. When formaldehyde is used in testing for indole derivatives, the
purple colour is often masked on account of further reactions taking place
between the indole substance and the trioxymethylene formed from form-
aldehyde by the condensing power of the concentrated acid.

4. In Hopkins and Cole’s modification of the test, glyoxylic acid reacts
by virtue of its decomposition into formaldehyde, and for one of the two
following reasons is the best reagent to use for the reaction:

(a) the formaldehyde is liberated in such small quantity that the
formation of trioxymethylene does not take place to any appreciable extent,

(b) the formaldehyde so liberated is in the nascent state, and therefore
more reactive.

5. In the case of tryptophane the colour test is often performed with
formaldehyde in the presence of an oxidising agent. The part played by the
latter is primarily to produce some oxidation product of tryptophane, e.g.
indolealdehyde.

The reaction under these conditions is of a complex nature as, besides the

126 A. HOMER

effects described above, (2) and (3), it involves the formation of coloured
substances by the action of the concentrated acid on:
(a) indolealdehyde and tryptophane.
(b) indolealdehyde in the presence of an oxidising agent.
(c) indolealdehyde and tryptophane in the presence of an oxidising
agent.
(d) indolealdehyde and tryptophane in the presence of formaldehyde.

REFERENCES.

Adamkiewiez (1874), Arch. ges. Physiol. 9, 156.

Dakin (1909), J. Biol. Chem. 11, 4, 289.

Ellinger and Flamand (1909), Zeitsch. physiol. Chem. 62, 276.
Homer (1913), Biochem. J. 7, 101.

Hopkins and Cole (1900), J. Physiol. 27, 418.

(1902), J. Physiol. 29, 451.

Rosenheim (1906), Biochem. J. 1. 233.

Voisenet (1905), Bull. Soc. Chim. (3), 33, 1198.

a ee oe

XII. THE ROLE OF GLYCOGEN, LECITHIDES,
AND FATS IN THE REPRODUCTIVE ORGANS
OF ECHINODERMS.

By BENJAMIN MOORE, EDWARD WHITLEY anp ALFRED ADAMS.

From the Biochemical Laboratory, University of Liverpool, and the Marine
Biological Station, Port Erin, Isle of Man.

(Received November 30th, 1912.)

In the course of a research upon the protamines and histones of the
spermatozoa of Echinus esculentus in which the reproductive organs (gonads)
were extracted with hot alcohol, we were struck by the enormous amount of
fatty matter removed by the hot alcohol. This separated out in drops and
formed a thick layer at the bottom as the alcohol cooled. This appeared to
us an unusual result to obtain from a male gland possessing only a reproduc-
tive function. It was still more surprising to find the aqueous extract made
later on, in the process of extraction of the protamines, heavily laden with
glycogen as shown by a marked pearly translucence, and confirmed by a
marked iodine reaction even after manifold dilution. In both of these
respects, the extracts of the gonads might well have been those from a liver
or hepato-pancreas in a vertebrate or higher invertebrate, so rich were they
in the usual products of metabolism. In the echinodermata, the gonads form
by far the preponderating part of the soft tissues of the animal, and contain
at least three-fourths of the organic matter of the organism. Also there exists
in these animals no structure resembling a liver, or other metabolic gland.

It was accordingly determined to make analyses of the fatty constituents
and glycogen of the gonads to elucidate their functions in this respect. For
this purpose experiments were instituted to observe the effects of seasonal
variation in April and in August—the period of sexual activity being in
April; to keep the animals in batches for some weeks in the fed and unfed
conditions respectively; and to test whether the gonads contained any
diastatic enzymes.

The exceedingly slow metabolism of these organisms during even a
lengthy period in which they were unfed, precluded us however from
observing whether the reserve stores of fat and glycogen of the gonads could

128 B. MOORE, E. WHITLEY AND A. ADAMS

be utilized for the purposes of the general metabolism of the animal as a
whole. As is shown in another research [Moore, Edie, Whitley and Dakin,
1912, p. 287] upon the metabolism of marine organisms, the amount of
organic matter required in twenty-four hours by an average-sized echinus is
only 42 milligrams, and as the gonads contain about five grams of dry
organic matter it is impossible to say whether these are acting as reservoirs
for the minimal demands of the organism during deprivation of food even for
a month or more. For example, 42 milligrams of organic matter oxidized
daily, is equivalent to 30 milligrams of moist weight, and for one month, or
30 days, this would amount to a loss in weight of only 0°9 g. This much,
however, is certain, that the gonads are very retentive of their stored-up food
supplies, such as fats and glycogen, for after even three weeks’ complete
deprivation of food, the gonads are still very rich in both oils and glycogen,
and even quantitative determinations show no constant difference between
the fed and the unfed organisms. The unfed specimens are alive and well at
the end of the period and commence at once to feed greedily when offered
suitable food such as ordinary laminaria.

It is not to be supposed that the metabolism of the echinus is so
excessively slow under normal conditions; this is only a dormant condition
which can be assumed when there is necessity for it. When the organisms
are first placed in captivity without food, there is an enormous output
of faeces for the first two days, afterwards the water in the tank is
practically free. for the rest of the period. The animals also develop an
appetite, which was shown by an attack upon the wooden bottoms of the_
tanks in which the unfed echini were stored. ‘The unfed animals supported
themselves in the stream of running water by the tube-feet just like the fed
ones and appeared quite active and well in every respect.

The echini fed in the laboratory upon laminaria and other sea-weeds, ate
quite vigorously and produced a large output of faeces daily, but at the end
of the period, the extracts from the two sets of gonads, of fed and unfed,
were very much alike in storage content. There may have been, however, an
increase in bulk of the gonads, of those which were fed, for the size of the
gonads varies considerably and they cannot be weighed at the beginning and
end of the experiment.

Another important point to remember is that there is no trace whatever
to be detected of a diastatic enzyme.

A gonad rich in glycogen may be left for some hours either at air
temperature, or heated in a bath to 35°, for the same time, and at the end
not a trace of sugar is obtainable.

B. MOORE, E. WHITLEY AND A. ADAMS 129

It may even be left for a period of two days with the same result. But,
extraction at the end shows the glycogen to be present in large amount, and
on boiling with dilute acid a copious reduction with Fehling’s solution is at
once obtained.

The diminished metabolism during fasting, compared with the rapid
uptake of food while fed, the persistence of fat and glycogen after long
starvation, and the absence of a diastatic enzyme, when all considered
together, and in conjunction with the relatively huge size of the reproductive
organs, indicate that the stored fat and glycogen are intended for the
internal metabolism of the gonads themselves and to subserve the enormously
developed reproductive functions.

The main bulk of the metabolized food would also appear to be devoted
to the nourishment and upbuilding of the reproductive organs. These
organs appear during the resting season to build up rich nutritive material
and then discharge this during the active season, when a diastatic ferment
may perhaps be temporarily developed.

It is remarkable that a few months after the active season, as in August
(the period of activity having been the end of March and earlier part of
April), the gonads are as large and full as before activity, and richly charged
with oils and glycogen. While being fed the uptake of material must be
many times the amount required for locomotion and other physiological
purposes in the unfed condition, as shown by the amount of oxygen used
daily by the unfed animal. If it may be assumed that the unfed animal
requires about as much energy for these purposes as the fed animal, as would
seem to be shown by the activity of the two being similar, then it may be
concluded that at least nine-tenths of the energy of the food taken up by the
echinus is conserved for reproductive purposes.

In regard to the carbohydrate and fatty metabolism, there is no noticeable
difference between the male and female gonads, both are rich in oils and in
glycogen.

A considerable amount of lecithides was in all cases found admixed with
the fats. The iodine value for both lecithides and oils was high, showing
a marked degree of lack of saturation. In this respect the fats of the
echinus resemble the fish-oils, or fish liver-oils.

Thus, for the oils, iodine values varying from 130 to 190 were obtained,
and for the lecithides, values of 70 to 120. These high iodine values suggest
that these fatty constituents may possess therapeutic values similar to those
of cod-liver oil in wasting diseases.

130 B. MOORE, E. WHITLEY AND A. ADAMS

METHODS OF EXTRACTION AND ANALYSIS.

In the earlier experiments carried out at Easter 1911, only the fats, oils
and lecithides as extracted by thorough treatment (a) with cold and (b) with
hot alcohol afterwards, were obtained and analysed.

The work was resumed in August and September 1911, and Easter 1912,
and then two methods chiefly were employed. In the first method the
glycogen only was estimated, by boiling out the gonads twice with water
acidulated with acetic acid, after which a third decoction showed no opalescence
and gave very little reaction with iodine. The two extracts were added
together, precipitated by the double volume of alcohol, washed with 70 per
cent. alcohol by decantation, and preserved in this for some months, after
which the precipitated impure glycogen was digested with strong caustic
potash, as in Pfliiger’s method described in the concluding part of the second
method below, followed by precipitation with alcohol, washing with alcohol,
drying and weighing. In several cases, the amount of glycogen present was
finally determined by hydrolysis to glucose, reduction by excess of Fehling’s
solution, and gravimetric determination of the CuO. When this was done it
was found that the product of the Pfliiger’s method was never pure glycogen,
but varied from 63 to 93 per cent. pure.

Both results are given alongside each other in the tables below for
comparison, The nature of the impurity was not determined, but it indicates
that this method of determining glycogen should be used with caution, and
always followed by a hydrolysis and sugar determination.

Second Method. In this, the final method used, which was found by
far the most convenient, oils, lecithides, and glycogen were all determined
in the same specimen. At the end of the preliminary treatment of the
organisms at the Biological Station, the shells were broken open by chipping
round in a sectional plane midway between mouth and anus, and the sex
determined by examining a smear from a gonad with a low power of the
microscope.

The echini were thus sorted into two heaps, the five gonads were removed
from each echinus, and two heaps made of male and female gonads re-
spectively. The weight of each quantity in the fresh condition was taken,
and they were then broken up and placed in about three times their volume
of absolute alcohol. In this condition they were transported to Liverpool
and after some months were then extracted for oils and lecithides, and for
glycogen respectively by the following method.

The alcohol was decanted and filtered off, without previous heating. A

mal

B. MOORE, E. WHITLEY AND A, ADAMS 131

fresh quantity of absolute alcohol was then added to the tissue which was
previously broken up as finely as possible, and mechanical agitation in a
bottle on a shaker was used for some hours. This second quantity of alcohol
was filtered off as before, and when mixed with the first, the mixture formed
the “cold alcoholic extract.”

The “hot alcoholic extract” was obtained similarly by two extractions
with boiling alcohol, decantation and filtration.

Each alcoholic extract was subsequently treated in precisely the same
fashion for determination (a) of salts and extractives, (b) lecithides, and
(c) fats and oils, as follows:

The alcoholic extract was evaporated till all aleohol had been removed,
and the total weight was taken. The residue was exhausted by repeated
portions of methylated ether, which portions were all united. The residue
from the ether consisted of salts, creatine, etc., and this was weighed separately.
The ethereal extract was evaporated down cautiously to a thin syrupy
consistency, acetone in excess was then added to precipitate the “ lecithide ”
fraction, this was dried and weighed and constants determined. The portion
soluble in acetone gave the product described under the heading oils and fats.
It is, of course, only claimed that this is a rough separation. The labour of
a complete separation in so many experiments would have been inordinately
great, and the figures for the constants in the two respective columns
(1) lecithides, and (2) oils and fats, show that these two fractions represent
a fairly complete separation of the lecithides from the oils. The weight,
saponification value, and iodine value, were determined in each case.

The dried tissue after removal of fats, etc., as above described, was weighed
and then digested for three hours on a steam bath with 40 per cent. solution
of caustic potash, which, as discovered by Pfliiger, leaves the glycogen intact
while it decomposes proteins. After cooling, twice the volume of absolute
alcohol was added, so precipitating the glycogen. The glycogen was filtered
off through a Gooch crucible provided with a filter of finest woven silk bolting
such as is used for flour sifting. This was found to be much preferable to
glass-wool, or asbestos, and may be recommended as a filtering medium
for ‘alkaline fluids. The precipitate is then washed with a 15 per cent.
solution of caustic potash in a 66 per cent. solution of alcohol, as recom-
mended by Pfliiger. Next it is washed through into solution in a fresh
receiver with hot water. The aqueous solution is again filtered, reduced in
volume by evaporation if necessary, and finally precipitated by twice its
volume of absolute alcohol, filtered, washed with two-thirds alcohol, dried

and weighed. Finally, the content of glycogen in the product is determined

B. MOORE, E. WHITLEY AND A. ADAMS

32

= 125 Sed ie

(ueye} ON) ——-99T

T&T

L1G OFT

STI

S0-61 ێT

9€T

G6-L6 99T

suleis
UOT}OVIYXO any{vAa
d1joyooyeR auIpoy
aje[durod 1a4y8
anpisar jo SSS

qustem Liq

ital

a €F-T
L61 €&-§
oe 09-0
006 OF-T
= 0€-T
6LT 80-6
— 99-0
106 19-7
<— O€-T
S61 &6-P
= 99-0
88T ST-Z
= OF-T
G6T §1-8
i 67-0
£06 [8.6

any[Ba Justa
moryRoy LION F

-tuodeg

se

SANTVA [IO pus 4eViT

Se

; a = LL-0
0-18 06T 08-T
= on 86-0
4-06 98T £9-0
ra ae 40-0
0-LL O8T 09-0
75 ar L&-0
1-98 9LT 09-6
‘VOZOLVNUAdS
a = TS:0
0-68 686 OL-T
SF a €&-0
0-F8 L06 OT-T
Pa — 0F-0
0-92 LOG PEG
sag a 98-0
PEL 656 01-6
‘VAO
anyvVa an[Bva qysTOM
SUTBOT wuoryeog [B40
-ruodg
coeeres —— -

San[Va opt

‘snquapnose snutyoy fo vozoznwumsady
pun nag adiy wm sng pup soy fo pun saprypnoaT fo suoyourimsajeq

ear

TI Giavi

GE-0 86-1 ri = 25%} U00.10
GL-0 L9-L 8 = _ 2oyooyy OFT
06-0 Té-T ccd a5BjU0IEg
8F-0 OL-€ =< uaF 20Y09)0 p20)
:4stoul's Gez ‘eozoyeuttodg adiy *Ay dx
OT-0 Go. T 5 ss 938) U80I10g
GL-0 69:0T ~~ ]0yoo)v .0FT
ST-0 66-T = es asRyue010g
F0-T ¥0-6 id Td J0Yyoo)v pr09
:ystoul'3Q0) “Bozojvutedg adiy “y]]‘dxq
96-0 FF-G a ania asvyUe010 F
06-1 OT-8 a ae 10Y0I]D 20 FT
90-T FI-@ = ad ane] U8010g
0S-€ OT-2 bee gy 10409)” p10)
24USIOM ysTou “3 Zee ‘VAC diy “TT “dxq
90-0 L6-T = oo asRqueo1eg
SE-0 VAI es i 10409) J0F7
0g-0 SF-T ih JU SIOM JSIOUL WO saseyuadIEg
06-% 9G.8 pd bk 0YOI]0 p)0p
4q31eM ystout “3 Ogg ‘wag adi *y ‘dxq
SOATIOBIYXO yoRayxo AATJOVIZXE PUB [VIIO}VIL
049 ‘AUT}BAID JO YSTIOA
“Sq[BG 1207,

LIGL “asvg fo squaumiadxy

B. MOORE, E. WHITLEY AND A. ADAMS 133

by hydrolysis, and gravimetric determination by precipitation in excess of
Fehling’s solution.

Experiments of Summer 1911 (August and September). The first four
experiments to be described in this series were in regard to content of
glycogen only, and the effects of lack of food upon the glycogen content of
the gonads. The gonads in all these experiments are taken at an inactive
sexual period. The ova in the females at this time of year are unripe and
small, and formed eggs are small in number relatively to other tissue; the
spermatozoa in the males are alive and moving, but small.

Thirty echini of medium and large size only were taken from Port Erin
Breakwater at low tide on August 22, all were kept in a large tank filled
with running sea-water until August 24 (about 48 hours), not being fed in
the interval. ,

Twenty of the echini were then opened and the gonads examined for
sex, twelve were found to be females and eight were males. The female and
male gonads separately were at once treated for glycogen separation as will
presently be described in Expts. V and VI. The remaining ten echini were
kept in the tank in fresh running sea-water without any food until Sept. 11,
that is for a period of twenty days in all; they were alive and well at the end
of that period, and were then opened, when gonads, and all the other organs,
presented quite a healthy and well-fed appearance: the intestines were of
course empty. On examination of these for sex, four were found to be
females and six were males. The two masses of female and male gonads
were then at once treated for extraction of glycogen in the fresh condition
with the results shown in Expts. VII and VIII.

Experiment V. The gonads from the twelve females described above
weighed 440 g. in the moist condition or an average of 36°6 g. per animal.
As subsequent determination of dry weight of a given weight of moist gonad
gave a solid content of 17:2 per cent., this weight represents 75°7 g. of dry
solid. This dry solid has salts present in the proportion of sea-water in
invertebrates so that the amount of dry organic matter is about 62°5 g.

The material was mashed up into a soft pulp, and the proteins coagulated
by dropping the pulped mass into 400 c.c. of boiling one per cent. acetic acid
in distilled water, and boiling for five minutes. The fluid was then filtered off
through fine silk gauze and pressed out. The residue of tissue was pounded
-up and extracted with 500 cc. of distilled water in two portions. ‘The last
filtrate was almost clear of glycogen, and gave only a trace of reaction with
iodine. The first filtrate in dilute acetic acid was a strongly opalescent
solution but without obvious suspension. This was again filtered through

. if
Bioch. vir :

134 B. MOORE, E. WHITLEY AND A. ADAMS

two folds of fine silk, and a small portion of the filtrate which was still
strongly opalescent gave a strong port-wine colour with iodine, and on boiling
for five minutes with 3 per cent. hydrochloric acid gave a copious reduction
with Fehling’s solution.

The main volume of the acetic acid solution and the washings were united
and precipitated with twice the volume of absolute alcohol giving an
abundant white precipitate settling easily. This was washed, after de-
cantation, with more two-thirds alcohol, and carried under alcohol to
Liverpool, where it was later passed through the Pfliiger process for
purification of the glycogen and analysed. The weight of glycogen obtained
was 7°44 ¢., but hydrolysis and a gravimetric Fehling determination showed
this to be only 79 per cent. pure glycogen giving 1°33 per cent. glycogen on
the moist weight, or the glycogen in the gonad is 9°36 per cent. of the
total organic matter.

Experiment VI. The eight sets of male gonads weighed 258 g., thus
averaging 32°25 g. per animal. Of this quantity, 200 g. were taken for
glycogen extraction, corresponding to 28:2 g. of dry organic matter. The
tissue was put through exactly the same process as used for the female
glands, and gave the same appearances and the same qualitative tests. The
filtration of the dilute acetic acid solution was more difficult, probably on
account of the presence of the minute spermatozoa, but obviously there was
much glycogen present, as shown by fine opalescence and iodine test. After
filtration through fine silk it was precipitated by two volumes of alcohol as
before, and carried off for analysis. Any spermatozoa present in the original
acetic acid precipitate would be destroyed by the subsequent boiling with
strong alkali in the Pfliiger process. The amount of glycogen obtained was
3°02 g., and this was found on hydrolysis to be 81°4 per cent. pure, corre-
sponding to 1°51 per cent. of crude and 1:23 per cent. of pure glycogen in
the moist gonad, or to 8°66 per cent. of the total organic matter.

It is thus seen that the percentage of glycogen is much the same in the
male and female gonads, corresponding to 9°36 per cent. in the female and
8°66 per cent. in the male of the dry organic matter.

Kaperiment VII. The four sets of gonads from the females which had
been kept unfed in the tank for 20 days weighed 122 g., or an average of
30°5 g. per echinus, while Expt. V shows that the twelve sets of gonads of
females killed without previous abstention from food weighed 440 g., an.
average of 36°6 g. The difference is not great and is probably quite fortuitous,
as the content of glycogen indicates no effects of the twenty-days’ fast.

The process of extraction was exactly the same as in Expts. V and VI,

B. MOORE, E. WHITLEY AND A. ADAMS 135

and yielded similar qualitative results. ‘The yield of crude glycogen was
2:09 g. but this, for some unknown reason, showed only 62 per cent. of
glycogen on hydrolysis, corresponding to 1:06 per cent. of the moist gonad, or
747 per cent. of the total organie’matter. This is a somewhat smaller figure
but the decrease is too small to mean any real demand on the gonads for
reserve food by the rest of the organism during the period of inanition, also
they are still very rich in glycogen. Moreover, the result is the reverse in
the experiment with the males now to be described.

Experiment VITT, The six sets of gonads from males kept without food
in the tank of running sea-water for 20 days weighed 237 g., thus averaging
395 g. per animal. These were extracted with dilute acetic acid and
precipitated with alcohol at Port Erin on similar lines to the previous
experiments, and the precipitated crude glycogen was afterwards worked up
by Pfliiger’s method at Liverpool, and controlled by hydrolysis and estimation
of glucose. There were obtained 4°45 g. of glycogen, shown by hydrolysis to
be 82°8 per cent. pure; this works out at 1:54 per cent. of pure glycogen in
the moist gonad, or 10°84 per cent. of the organic matter. It is here seen
that the glycogen in the unfed males is somewhat higher than in the fed males
of Expt. VI; but as the females are in the reverse direction by about
a similar amount, the proper conclusion is that the four experiments
demonstrate no appreciable disappearance of glycogen from the resting
gonads of male or female as a result of a fast of twenty days. All four
experiments concordantly show that the reproductive organs of Hchinus
esculentus are very rich in glycogen.

A second catch of 60 well-grown individuals was made from Port Erip
Breakwater on August 25 at low water with which the remaining
experiments of this series were carried out. All these were left over-night
in a large tank through which there ran an abundant supply of fresh
sea-water. Next morning a large amount of faecal matter obviously chiefly
of vegetable origin was found on the bottom of the tank. On the morning
of the day after capture (Aug. 26), 15 individuals taken indiscriminately were
opened, examined for sex, and the gonads assembled in two lots, one male,
the other female. Out of these 15 specimens, six were females, eight males,
and one was rejected as being in an abnormal condition. The six female sets
of gonads weighed 154 g., or 26°6 g. per individual; the eight male sets
weighed 235 g., or 29°4 g. per individual. The mass of gonads in each case
was sampled out into three portions which were used as follows: (1) was at
once placed directly in alcohol, for determination later of oils and_ fats
lecithides, and glycogen, (2) was at once extracted with dilute acetic acid and

g9—2

136 B. MOORE, E. WHITLEY AND A. ADAMS

precipitated with alcohol, etc. for glycogen determination, as in the preceding
four experiments, and (3) a smaller quantity was dried on a bath for deter-
mination of total solid matter in the fresh tissue of the gonad.

Out of the 154 g. of female gonad, 74°5 g. were placed in alcohol (see
Expt. IX b), 61:2 g. were extracted fresh with dilute acetic acid (see Expt. [Xa),
and 14:26 g. were taken for determination of total solids, which were found
to amount to 17:2 per cent., as the glands contain approximately 3 per cent.
of inorganic salts, on the basis that their fluids are isotonic with sea-water
this leaves 142 per cent. of dry organic matter in the female gonad.

Similarly the eight male gonads weighing 235 g. were divided into 751 g.
placed directly in absolute alcohol (see. Expt. X b), 134-4 g. extracted fresh
with dilute acetic acid (see Expt. X a), and 12°75 g. taken for determination of
total solids. The percentage of total solids found was 16°8, and after
deduction for inorganic constituents this leaves 13°8 per cent. of dry organic
matter in the tissue of the male gonad. This lies very close to the figure for
the female organ given above, and the values of 17 per cent. for total solids,
and 14 per cent. for organic solids, may be taken as an average.

The remaining 45 individuals of the second catch (Aug. 25) were used, as
follows, for experimentation in the laboratory. They were divided into three
lots of 15 individuals each, the sizes being arranged as far as possible so as to
give no advantage to any one batch!. The first batch were left unfed in the
large tank (see Expts. XI and XII). The second and third batches were
removed in order to have plenty of room and fresh running sea-water, to
a series of smaller tanks forming part of a fish-hatching apparatus not at the
time in use. The thirty echini were distributed over ten of these tanks,
three in each tank. Fifteen of these echini were fed on fresh laminaria
gathered on the shore (see Expts. XII and XIII) and they ate a surprisingly
large amount of it daily. The other fifteen, forming the third batch, received
no food of any kind during the period of the experiment (see Expts. XIV and
XV). As some index of the rate of feeding of the echini supplied with
laminaria there may be mentioned the amount of faeces found in the tanks.

The faecal masses were washed into a graduated 100 c.c. measure, allowed
to sediment, and the volume of the sediment measured. The volumes found
were as below. At this time in the experiment a few of the echini had
perished, as noted below:

Tank No. I. Two echini, 45 c.c. of faecal sediment in three days.

' As a matter of fact it frequently happens that heavier gonads are obtained from a medium-
sized than from a large echinus, so that this precaution is scarcely required so long as no
individuals below a certain size are taken in the first instance.

hapa AR

ee

.

B. MOORE, E. WHITLEY AND A, ADAMS 137

Tank No. II, Two echini, 35 ¢.c, of faecal sediment in three days.

Tank No. IV. Three echini, 100 c.c. of faecal sediment in three days.

The amount of laminaria eaten was not actually weighed, but it was
quite a surprising amount for organisms of so slow a metabolism, and was
evidently chiefly being used for storage in the gonads. The unfed batches of
these echini produced no faeces, after the first two days or thereabouts, and
at the end of the experiment it was noticed that in the tanks of the fish-
hatching apparatus, which were made of planed wood, in each case four or five
bitten out spots showed where the animals had attacked the wood of the tank
floor and scooped it out quite distinctly; no similar attack had been made
on the tanks containing those fed on laminaria. The experiment lasted from
August 25 till Sept. 13, a period of nineteen days. At the end the gonads
were removed, examined for sex, and placed at once in absolute alcohol, from
which they were taken and analysed some months later in Liverpool.

Early in this experiment a certain number of the echini in each of these
three batches succumbed. This result was in no wise due to lack of food,
for as great a percentage died in the fed batch as in the corresponding unfed
batch. Such mishaps often occur soon after the echini have been taken, for
some unknown reason, and then the survivors keep in quite good condition.
All the animals alive at the end were active and appeared quite normal and
healthy. Out of the fifteen kept in the large tank, nine survived, of which
six were males and three were females; of the fifteen fed daily in the fish-
hatching tanks, eleven survived, seven females and four males; of the fifteen
kept unfed in fish-hatching tanks, eleven also survived, six males and five
females.

Experiment IX a. Amount of glycogen in freshly extracted female gonads.
This experiment was carried out with a portion, 61:2 g., of the gonads of
females of Catch II. freshly extracted with dilute acetic acid. Details as
before, amount of glycogen 1:14 g., found on hydrolysis 87°7 per cent. pure =
1°63 per cent. of pure glycogen in fresh gland, or 11°64 per cent. of the dry
organic matter.

Experiment IX b. Determinations of (a) lecithides, (b) fats and oils, and
(c) glycogen in a portion, 74°55 g., of female gonads of Catch II, placed at
once in alcohol. Hot and cold alcoholic extracts were taken and analysed as
described, and the glycogen determined in the residue. In percentages of
the moist tissue, the following were the results obtained: glycogen, 1°53 per
cent.; lecithides, cold aleohol 0°41, hot alcohol 0°25, total = 0°66 per cent. ;
constants (cold alcohol), sapon. value 210, iodine value 78°0; (hot alcohol),
iodine value 91:1. Fats and oils, cold alcohol 0°4, hot alcohol 1:46, total =

138 B. MOORE, E. WHITLEY AND A. ADAMS

1:86; constants (cold alcohol), sapon. value 208, iodine value 180; (hot alcohol),
sapon. value 197, iodine value 174.

Percentages expressed on dry organic matter: glycogen, 10:9; lecithides,
4:6; fats and oils, 13:1.

Experiment X a. This was carried out upon a portion, 134-4 g., of the
male gonads of Catch I, freshly extracted with dilute acetic acid, ete.
Amount of glycogen obtained, 2°66 g. of 81:15 per cent. purity = 1°60 per cent.
of glycogen in fresh gland, or 11°3 per cent. of the dry organic matter.

Experiment X b. Determinations of (a) lecithides, (b) fats and oils, and
(c) glycogen in a portion, 7571 g., of male gonads of Catch II, placed at once
in alcohol. Details as in Expt. IX b. Results obtained in percentages of
moist tissue: glycogen, 1°70; lecithides, cold alcohol, 1:01, hot alcohol 0:26,
total = 1:27; constants of lecithides, (cold alcohol), sapon. value 240, iodine
value 89, (hot alcohol), sapon. value 197, iodine value 98:2 ; fats and oils, cold
alcohol 0:54, hot alcohol 1:80, total = 2°34; constants of fats and oils, (cold
alcohol), sapon. value 206, iodine value 160, (hot alcohol), sapon. value 202,
iodine value 181.

Percentages expressed on dry organic matter: glycogen, 11°90; lecithides,
8°89; fats and oils, 16°38.

Attention may again be drawn to the highly unsaturated nature of the
fats and oils as shown by the high iodine values, in both male and female
glands, and to the high percentages of reserve storage materials throughout.
There was also a large amount of organic matter soluble in 66 per cent.
alcohol, and a relatively small amount of insoluble protein and tissue debris,
but figures for these are not given, as analyses of the portions were not made.
The results were similar in this respect to Table II in which the total amount
of salts and alcohol soluble extractives are shown in a similar set of experi-
ments.

Experiments XI to XVI. The whole of the glands in each of these
experiments were placed at once, after conversion into a soft pulp, in twice
their volume of alcohol. Through a misunderstanding they were used for the
glycogen determinations after having been extracted with cold alcohol only,
so that the figures given below do not represent total amounts of lecithides,
or fats and oils. They are comparable amongst each other, however, and
show that there is no appreciable diminution of fatty substances as a result
of the nineteen-days fast.

Experiments XI and XII refer to the unfed echini in the large tanks; there
were three females, weight of gonads 73 g., average 24 g9., and six males,
weight of gonads 152 g., average, 25 g. Experiments XIII and XIV refer

—™

aes

139

€1-II
6¢-T
IF-G

6&-¢
LL-O
98-0

L
ei
-
a)
=
=
(om)
7.
=
>
62
i
=
=)
=
ca
ey
on
©
©
=
aa]

fo powad v bursnp ‘pafun puv paf “day syrup ur ‘syynsae moys sywaunsadxa oy],
pup vag adiwug wa uahool,y puv ‘sg pun sg ‘saprynoeT fo suoynurmsapaq

ues004[

pene

OST

anyjBa
aurpoy

O&%

£06

an[Ba
uoNRoy
-tn0dRg

san[va [1 pu yy

80T

1a9yvur oruesr10 Aap Jo saBujU9010 4

— 19-0 6-¢ R8-L s° AUSSI} ISIOUL JO SasBIMA0I0

Isl 0@-1 ¢.O1 10-F1 pi * “SSLT Ur syste ny
:payun vozoyeutedg “TAX ‘dxq

_— FG-8 — — 199} Bu o1uBs10 Arp Jo sadvjUe0104

— Za 1 69-2 gL Pe ONSSYy WSIOUL JO sasBIUADIIg

68T IL-@ C8.6 G81 a ‘3 ELT UL SIqSI0

:payun BAQ “AX ‘dxq

‘syuR) Sup ey-Yysy ur speuue pagum “TAX puve AX ‘sdxq

09-¢
08-0
£6-0

‘syuR} Suryoyey-Yysy ut seme ped “ATX pue [ITX ‘sdxq

19}, 8U OLaBs.10 Lp Jo sasvyued18

= aut ¢8-¢ O6-L oa AUSST] JSIOUL JO sanvjUI.IE

GIG LE-T 06-9 08-6 oe ‘3 QTL Ut sqysia Ay
poy vozoywumedg “ATX “dxg

_ FF-9 _ — 199} 8UI orURS.10 Lap Jo sasRUI0I0F

2 &6-0 6F-¢ 6-9 ‘ia AaNSsl} JSLOUL JO sasvIUAIIO

L8T 88-T 06-11 I-F1 7 B FOG UL SITIO AY

‘pay VAQ “THX “dxg

O-L — _— 9-1 — — 19}4vuI o1Uv3.10 Arp Jo sadvqyuad1eg
0-1 _ — 60-T FL-9 16-8 aNSSt} JSIOML JO SadvIUad10T
GT Sal CGE 99-1. G6-0L GG-€1 as * “B Bey Ur sqysre Ay
:payun vozoyuumedg ‘TTX “dxq
L0-L = = €1-IT — — 1a}yvu o1uBs10 Lap Jo saceyuadIeg
10-T _— =e 6¢-T G6L-21 GGT “* gnsst} JSIOUL JO sasByUIDIIg
FL-O 00T 81 91-1 06-8 LI as +B GL UL S71 AA
:payun Vag ‘TX ‘dx
*YUB] BIBL UT AoyJoFO} s[RUMIUR pezuM “]TX pue TX ‘sdxq
VU sO an[Ba aNn[va qysiem = “aga: ‘saseq JOBI}XO [OYOOTR ppoo YyTA paqovazxXe
[®10J, aulpoy uOlROy [Vjoy, snomasoaqyrae jo qYysTeA [VII9}VUI JO 9INZVN
-1aodeg ‘QUTPVAIO ‘STB [BIOT,
— rc er, oe

san[va aprqqioery

(WOT Ug Mog) fisoywsoqny ur shop 6]
‘snqueTnose snutypoy fo vorzojpusady
‘TIGL wamungy fo syuaumadag fo worsoq

II Widvi

140 B. MOORE, E. WHITLEY~ AND A. ADAMS

to echini fed during these 19 days with laminaria, but kept otherwise under
strictly similar conditions to those in the two subsequent experiments
(Expts. XV and XVI); there were seven females, weight of gonads 204 g.,
average 29 g., and four males, weight of gonads 118g, average 29 g.
Experiments XV and XVI refer to echini unfed, but otherwise similar to
those in Expts. XIII and XIV; there were five females, weight of gonads
173 g., average 34 g., and six males, weight of gonads 178 g., average 29 g.
The average total weight of gonads is less in the case of the unfed echini in
the large tank, but in those unfed, alongside the fed, in the fish-hatching
apparatus no such effect is discernible. In fact the gonads of the unfed
females are appreciably heavier than those of the fed females, while the
males are practically of equal weight in the two series. The only conclusion
valid is that even a nineteen-days period of removal of supplies is ineffectual
to cause any real starvation, or appreciably use up reserves.

It would certainly have been expected from the amount of laminaria used
by the echini that some increase of weight might have taken place in the fed
Animals, but it is possible that only a small amount of the nutritive organic
matter of the laminaria was actually utilised by the echini. Certainly the
faeces still contained a large amount of nutrition for they were swarming
with copepoda when examined microscopically. These no doubt came in
with the water supply and remained and accumulated upon the faeces of the
echini, which formed their food supply.

The analytical results of these six experiments are shown in Table II.
The figures obtained in these experiments show no clear evidence of diminu-
tion of reserves in the glands as a result of stopping the feeding for nineteen
days, for there is still abundance of glycogen and fatty materials.

Experiments of Haster 1912. The echini used in these experiments were
taken a week or two after the active sexual period which appears to have
been somewhat early in this particular Spring. At least it was difficult to
obtain any ripe eggs for fertilisation about the period of this experiment.
The catch was made on April 10 and was divided into two lots, the animals
of one of which were kept unfed from April 12 till April 21, while those of
the other lot were fed on laminaria. ‘The animals were killed at the end by
opening the shells, examined for sex, and separated into four heaps, male and
female, fed and unfed, respectively. These were immediately fixed by placing
in two volumes of alcohol and taken to Liverpool for the analyses, which were
conducted as described above ; the results are given in Table III. It will be
observed that the results do not differ materially from the others. Only
percentage figures and the constants are given in the table.

B. MOORE, E. WHITLEY AND A. ADAMS 14]

TABLE ITI.
Lecithides Oils Glycogen
ye — oo — es a
Percentage Percentage Percentage
of dry Saponi- of dry Saponi- of dry
organic ficatiom” Iodine organic fication Iodine organic
Organs extracted matter value value matter value value matter
Exp. XV. Sperm. of
fed males, wt. 205 g.
Cold alcohol 51 159°6 86°0 4:1 164°7 150-1 _
Hot alcohol 7:2 176°1 90-1 5-0 122-3 133-4 —
Total 12°3 — — 9°1 ~ — 76
Exp. XVI. Ova of fed
females, wt. 171 g.
Cold aleohol 2°87 168-1 82:8 3°15 178 134°7 —
Hot alcohol 3°92 172°8 89°7 3°92 185°1 129°5 —
Total 6°79 — -- 7°07 -— — 6°2
Exp. XVII. Sperm. of
unfed males, wt. 62 ¢. -
Cold alcohol 9-10 163°4 9495 19°43 158 148°6 <
Hot alcohol 1-12 174°3 101-0 1°12 112-1 160°6 —
Total ee hie? _ — 11°55 -- — 11'1
Exp.XVII.Ova of un-
fedfemales, wt. 160g.
Cold alcohol 4-08 164°2 88°6 6°08 168°1 142°1 is
Hot alcohol 3°76 ah A key 96°3 11°68 174:2 118°4 —
Total 7°84 _ — 17°76 — — 9-1
CONCLUSIONS.

1. Both male and female reproductive glands in echinoderms contain
large amounts of reserve metabolic products such as glycogen, fats and
lecithides.

2. These reserves are only slowly used up, if at all, when the animal is
deprived of food.

3. In a reproductive gland richly stored with glycogen, no sugar forma-
tion occurs on keeping after death, even in a period of two days.

4, The amount of food consumed is much greater than that required to
cover the daily metabolic wants of the animal, and is largely stored in the
reproductive glands during the resting period, but it has not yet been
possible to trace the conversion of this at the active reproductive season.

5. The fatty constituents of the reproductive organs of the echinoderm
are highly unsaturated, and resemble in this respect liver oils.

We desire to acknowledge much valuable assistance from Messrs. W. H.
Evans and T. A. Webster in connection with the experimental work.

REFERENCE.
Moore, Edie, Whitley and Dakin (1912), Biochem. J. 6, 255.

XIl.. THE BASIC AND ACIDIC PROTEINS GF TBE
SPERM OF ECAINUS ESCULENTUS. DIRECT
MEASUREMENTS OF THE OSMOTIC PRESS-
URE OF A PROTAMINE OR HISTONE.

By BENJAMIN MOORE, EDWARD WHITLEY aAnp
ARTHUR WEBSTER.

From the Marine Biological Station, Port Erin, Isle of Man, and the
Biochemical Department, The University, Liverpool.

(Received December 20th, 1912.)

So far as we are aware no one has hitherto made a determination of the
molecular weight in colloidal solution of any member of those interesting
classes of proteins with marked basic properties known as the protamines and
histones. These substances have been set down by workers upon them such
as Miescher and Kossel as the simplest steps in natural protein synthesis
which furnish a clue as to the constitution of the others, and it is rather
remarkable that no attempts have been made to study their degree of
complexity in solution. The present paper must only be regarded as a
tentative experiment in this direction, to be made more complete when more
material becomes available; for the chief difficulty hes in obtaining suitable
material and working it up into a supply of pure protamine or histone. The
substance which we have isolated from the ripe male gonads of Hchinus
esculentus appears to us to stand intermediately between a protamine and a
histone, for it gives feebly the Millon’s test as a reddish colouration on
boiling, with no precipitate; and yet it possesses all the properties of a
protamine. This substance dissolved as a sulphate in distilled water gave
an osmotic pressure leading to a molecular weight in solution much less than
that of an albumin or globulin, the figure being approximately 8780 when
calculated to standard temperature and pressure.

It behaved as a true colloid in being perfectly indiffusible through
parchment paper, and the osmotic pressure rose steadily to a maximum, and
remained there at a constant level. Afterwards, the sulphate of the protein

B. MOORE, E. WHITLEY AND A. WEBSTER 143

was thrown out in apparently quite unaltered form by excess of alcohol, dried
and weighed. The substance comes down quite sharply with excess of
alcohol, as a white precipitate readily soluble again in distilled water. As
this process was repeated several times before measurements were taken, it
is quite certain that if there was any admixture it must have been that of
very closely allied substances of this class,

The only other substance of this protein group from the echinodermata
ever examined was a substance isolated from the sperm of Arbacia by
Matthews [1897], and called by him arbacin. This substance is classed
amongst the histones by Matthews chiefly because it gave a positive result
with the Millon’s test. The substance with which we were working, from a
different species of echinus, in a first preparation gave a negative reply to
the Millon’s reagent, and in a second preparation gave a somewhat feeble
red colouration but no precipitate, and in absence of more complete analysis
we are consequently somewhat doubtful as to whether we ought to classify
it as a protamine or histone, so for the present we prefer to leave the
point open,

It may be of some interest to pause here to state that our original
intention in isolating this substance at Port Erin in the Spring of 1911, just
at the commencement of the active sexual period, was to observe whether
protamines and histones acted as hormones to the earlier stages of cell-
division in sexual reproduction.

The whole group of the protamines is intimately connected with the male
reproductive organs, being found in these glands almost exclusively in fishes,
and in greatest abundance just before the sperm cells are ripe for discharge.
The histones are obtained from the ripe spermatozoa of fishes, but in addition
extend up to the sperm cells of mammalia. Both protamines and histones
are strongly basic bodies, showing a markedly alkaline reaction when in
solution as free bases, and combining strongly with acids, such as sulphuric
acid, in definite molecular proportions.

Now it was shown some years ago by Moore, Roaf, and Whitley [1905]
that almost infinitesimal amounts of alkali added to sea-water increased the
rapidity of cell-divisions in fertilized echinus eggs, and that a slightly greater
amount of alkali hurried the process into marked abnormalities.

It was accordingly thought that small amounts of alkaline proteins, such
as protamines or histones, present in the sperm, ought to have prominent
functions in relationship to cell-division.

It was from this point of view that the protamine, or histone, from
echinus sperm was prepared and its action tested (a) upon unfertilized ripe

144 B. MOORE, E. WHITLEY AND A. WEBSTER

eggs and (b) upon fertilized ripe eggs, alongside controls developed in
sea-water only.

The results were uniformly negative: whether this may be due to our
being unable to realize proper experimental conditions we do not know. It
may well be, that protamine added as sulphate to sea-water is in some wise
rendered inert and incapable of penetrating the ovum, while the same
substance carried into the ovum as part of the sperm-nucleus may be most
effectual. The occurrence of these peculiar proteins in the sexual organs
would seem to indicate some relationship with a reproductive process, but we
have quite failed to realize an experimental proof.

The only positive result obtained by treatment of the fertilized eggs was
that an extract of the crude sperm powder, after treatment with alcohol as
described below, added in small amount, such as 1 in 4000, to sea-water
inhibited cell-division and induced irregularities. Sperm nuclein appeared to
possess a like action, but the protamine sulphate was quite inert.

The substance was separated from the male gonads by the well-known
process of Kossel. About sixty well-grown echini were captured upon the
breakwater at low tide on April 14th, 1911. The gonads of each individual
were examined microscopically for sex identification and a mass of about
700 g. of spermatic gonads was obtained. This material was broken up
finely and taken out in a large volume of water acidulated with acetic acid
which causes the sperm cells to agglutinate and separate off. This was
collected, extracted once with cold alcohol, and then exhausted of all fats and
lecithides by four extractions with hot alcohol, and one with ether. The ether
took out practically nothing on account of the previous exhaustive treatment
with hot alcohol. The dried residue from all this extraction, of 700 g. of
moist gland, amounted to only 21°68 g. This dried material remaining to
the end, contains the protamines or histones and the nucleins or nucleic acids
in combination together as nucleoproteins. To obtain the basic proteins
(protamines or histones) it is extracted, still following Kossel’s instructions,
with five times its volume of half per cent. sulphuric acid, shaking vigorously
at each extraction for about 15 minutes, followed by separation through fine
silk gauze and pressing out. About three extractions were found to exhaust
all protamine or histone from the material. After subsequent complete
filtration through filter paper, the sulphuric acid extracts were precipitated
with three volumes of absolute alcohol, yielding a nearly pure white
precipitate which after standing for 24 hours was decanted and filtered by a
Buchner filter. The precipitate dissolved readily in distilled water, and was

purified by dissolving and reprecipitating with aleohol four times. It was

aren

B. MOORE, E. WHITLEY AND A. WEBSTER 145

then dried after washing thoroughly with ether and weighed. The weight of
the precipitate from 700 g. of moist gland, or 21°68 g. of dried tissue residue
after alcoholic and ethereal extraction, was 1-95 g.

The residue from the extraé@tion of the tissue with dilute sulphuric acid
is exceedingly rich in nucleins or nucleic acid. It was freed from acid
and dried by extraction with alcohol and ether, and then left to extract
over night in the cold with one per cent. solution of caustic soda. When
treated with the dilute alkali it swells up into a jelly which is almost solid
in consistency and cannot be filtered in this condition, but it aggregates
somewhat on adding three times its volume of absolute alcohol. It was left
over night in this condition, but next morning was still unfilterable through
paper. It was cleared of grosser debris by filtration through the finest silk
bolting so giving a colloidal suspension free from large particles. When this
alkaline alcoholic suspension was made just acid by addition of normal
hydrochloric acid, a white flocculent precipitate was obtained which settled *
rapidly. This filtered quite readily leaving a very fine almost white pre-
cipitate on the paper. This precipitate was washed with distilled water
over night, and until the washings were practically neutral next morning.
It contained no protamine or histone as was shown by extracting it with
dilute sulphuric acid and none was obtained from it at any stage, showing
that the 1°95 g. mentioned above represented practically the total amount in
the glands. The above precipitate caused by the acid in the alkaline extract,
representing nuclein or nucleic acids, was weighed and found to be 7°45 g.
A second extraction of the tissue residue separated by the silk filtration,
yielded by the same method of alkaline extraction, followed by alcohol and
acid, about 1:12 g. of a similar nucleic product, so that the total amount of
acidic material was 7°45 + 1:12 =8°57 g. in combination with 1°95 g. of basic
protein. The final residue after extraction with both acid and alkali was a
substance which swelled up into a thick jelly in dilute alkali but did not
dissolve in it, and shrank up when treated with dilute acid and was obviously
an insoluble fibrinous form of protein. The weight after treating with
alcohol, drying and weighing was 9:10 g. This material, calculating from the
above weights, and the total weight of dried tissue extracted, is seen to form
roughly half of the organic residue remaining after extraction with dilute
acetic acid followed by alcohol and ether.

A second extraction of a different batch of male gonads made afterwards
in Liverpool gave similar results.

A small amount of the protamine (or histone) sulphate was used at Port
Erin for the investigation of effects upon cell-division, the remainder was

146 B. MOORE, E. WHITLEY AND A. WEBSTER

afterwards made up into a solution in distilled water containing approximately
two per cent. of protamine sulphate. This was introduced against distilled
water in the type of osmometer already described in previous papers [Moore
and Roaf, 1907]. The dialysing membrane was parchment paper and
complete equilibrium was established in three days’ time as shown by the
figures given below. At the end negative results were obtained on testing
for protamine on the water-side. The protamine solution on the solution
side of the membrane was quite sweet at the end of the experiment, and the
protamine was precipitable from solution by alcohol as at the beginning.
The solution had of course become somewhat diluted by expansion, and a
determination of the amount of protamine sulphate by precipitation by
excess of alcohol, filtering on to a Gooch crucible, drying by washing with
alcohol and ether and weighing gave 1:65 per cent. of protamine sulphate.
This gave a steady osmotic pressure of 34 millimetres of mercury at 15°,
which works out at 19°3 millimetres for a one per cent. solution at 0°, and
leads to a molecular weight in solution of the protamine solution of 8780.
The following is the protocol of the experiment:

Manometric
Date Time reading Temperature
1911. Nov. 21st 12 noon 388 15°
a eaend 9.30 a.m. 391 13°5°
aeeezord 9.45 a.m. 396 140°
ee Fe 5.30 p.m. 397 16°5°
» 24th 10.0 a.m. 401 15-0°
Aa 6.0 p.m. 403 16-0°
», 2oth 9.45 a.m. 406 14:5°
sot eh, 5.45 p.m. 406 S 15°5°
., 26th 10.0 a.m. 406 \& 14:5°
Wh Kas 5.30 p.m. 406 | 5 15°5°
a aTth 9.30 a.m. 406 j E 15-0°

Opened. Zero Pressure 389

Osmotic Pressure=2 (406 —389)=34 millimetres of mercury. Percentage of protamine
sulphate = 1°65.

As has clearly been shown by Moore and Roaf, and confirmed by Lilhe
and others, the molecular weight of colloids in solution varies with the
solvent, amounts of neutral salts, and acidity or alkalinity. Such reagents
affect the colloidal aggregation. Still it is of interest to compare the
complexity of solution aggregation of this basic protein, with averages for
other proteins, and colloids.

The following results have been obtained by various observers :

eee me ae

>
——_

“-

B. MOORE, E. WHITLEY AND A. WEBSTER 147

Osmotic pressure for Molecular weight
Colloid one per cent. of Colloid in solution
Gelatin (Moore and Roaf) __... “e 9°2 mm. 18,500
Egg albumin (Moore and Parker)... 13°1 13,000
Serum ( “- - Nia Ae 2°8 60,800
Alkalized Serum ( ¥ “n ip eee 17°0 10,000
Sodium Oleate ( + ) eae 10*4 16,300
” Palmitate ( Pe - ) 10-0 - 17,000
¥ Stearate ( y - alters 94 18,000
Caseinogen (Moore, Roaf and Webster) ... 129°0 1,400
Haemoglobin (Hiifner and Gausser) . 15,000 to 16,000
fe (Reid) of M- — 65,000
“s (Roaf) under varying con-
ditions from ke — 5,000 to 70,000
ce but under more natural con-
ditions and in neutral
solution about i, — 14,000 to 18,000
Histone and Protamine (present communi-
cation) nh re he. 19-3 - 8,780

The solution complexity accordingly is less than that of the coagulable
proteins, but is considerably higher than that of the acidic protein caseinogen,
which is by far the lowest of all proteins hitherto examined. The alkalized
serum and the histone le nearly at the same level.

It is interesting to compare the molecular weight in solution as given
directly by osmotic pressure with the molecular weights assigned from
analysis by Kossel and his co-workers, and with the molecular weights of the
amino-acids (hexone bases) which the protamines yield on hydrolysis. The
molecular weights given by Kossel for salmine and sturine lie between
800 and 900, so that even after allowance is made for the inorganic acid in
the sulphate there must be eight or more such large molecular groups in the
solution aggregate to yield 8,000 to 9,000 as found. The mean molecular
weight of the amino-acids forming this class of protein may be taken as 150,
and, making allowance for the sulphate in combination, this leads to the
conclusion that there are about 40 molecules of such amino-acids united to
form the complex, or colloidal solution aggregate, of the protamine, or histone,
in solution.

At present, similar measurements of the molecular complexity in solution
of nucleoproteins and nucleic acids are being carried out.

The expenses of this research have been defrayed by a grant from the
Government Grant Committee of the Royal Society.

REFERENCES.

Matthews (1897), Zeitsch. physiol. Chem. 23, 399.
Moore, Roaf and Whitley (1905), Proc. Roy. Soc. B. 77, 102.
Moore and Roaf (1907), Biochem. J. 2, 34.

XIV. THE FATTY ACIDS OF THE
HUMAN BRAIN.

By EGERTON CHARLES GREY, 1851 Hahibition Scholar.
From the Physiological Laboratory, University of Sydney.

(Received December 17th, 1912.)

In the following communication evidence is brought forward that the
solid fatty acids derived from the lipoids of the human brain do not consist
chiefly of palmitic and stearic acids as has been claimed by certain investi-
gators. The observation of Thudichum [1901] that the stearic acid of the
human brain melts not at 69° but at 52° has been confirmed, and by
fractional precipitation an acid isomeric with stearic acid has been isolated.

I have been unable to find any unoxidised fatty acids containing more
than eighteen carbon atoms in the molecule. It has also been shown that
at least 25°/, of the solid fatty acids of the human brain are hydroxy-acids of
higher molecular weight than stearic acid. From the close resemblance in
properties which the cerebral hydroxy-acids show to the fatty acids found by
Darmstaedter and Lifschiitz [1896] in lanoline, it may be of some interest
to suggest that we have here another indication of the relationship of the
nervous system to other tissues of epiblastic origin.

The cause of the high molecular weight of the brain fatty acids.

About two years ago I pointed out [1910] that the mean molecular weight
of the solid fatty acids of the human brain, as determined by titration, was
considerably higher than that of stearic acid, this fact being confirmed
by analysis of the lead soaps. I was not at that time acquainted with
Thierfelder’s paper on cerebronic acid [1905]. This acid occurs with other
hydroxy-acids of high molecular weight in the precipitate of potash soaps
difficultly soluble in alcohol, which is obtained when the saponified mixture
from the fatty acids is allowed to cool. Moreover although the potassium
salts of these hydroxy-acids are sparingly soluble in alcohol, they remain
dissolved to a considerable extent in the presence of the other soaps and

EK. C. GREY 149

subsequently separate amongst the first fractions precipitated as barium or
magnesium soaps. The existence of hydroxy-acids in the brain other than
cerebronic acid, with the possible exception of hydroxystearic acid [ Koch, 1902]
has not previously been noted. Tt is to the presence of these hydroxy-acids,
however, that the observed high molecular weights are to be attributed. In
the case of certain fractions it was observed that the apparent molecular
weight varied enormously when the acid was regained by decomposition of
the lead, barium or even the sodium soap by hydrochloric acid, the acid so
regained possessing a higher molecular weight than before. The same
phenomenon was observed in another way. When certain fractions of solid
fatty acids of high molecular weight were converted into lead soaps, and these
lead soaps subsequently exhausted with hot benzene or toluene (a process
which Frankel [1909] used to purify the lead soaps of the brain fatty acids),
a decomposition sets in with the formation of a white precipitate rich in lead,
and a brown coloured solution, which contains not lead soap, nor free fatty
acid, but a material practically devoid of acid properties. It seemed clear
that an anhydride or lactone formation had taken place in both the cases
mentioned. That this view is correct is confirmed by the following fact. If
the fatty acids are regained from the potash soaps difficultly soluble in alcohol,
taking no precaution to avoid the dehydrating effects of mineral acids, it is
found that the acid mixture so obtained is sparingly soluble in alcohol even
at the temperature of ebullition, and after repeated exhaustion with hot
spirit a portion is left which when dissolved in a mixture of chloroform and
alcohol is found by titration to be only faintly acid; but if the same material
is saponified the soap will be found completely soluble in hot alcohol and the
material regained by decomposition with acetic acid is a true acid substance
neutralising a considerable proportion of alcoholic potash. It is to the
presence of anhydride or lactone-like derivatives of the hydroxy-acids that
the observed variations in mean molecular weights as determined by titrating
with potash, are due.

The lactone formation under the conditions described occurs most
markedly, not with cerebronic acid, but with the hydroxy-acid melting
at 73°5°.

Preparation of the fatty acids from the brain substance.

The brain after removal of connective tissue and washing was heated
with 20°/, potash solution. After cooling, the solution was acidified with
hydrochloric acid and the fatty layer filtered off. This was then saponitied
with 10°/, alcoholic potash for 15-24 hours. The solution was evaporated to

Bioch. vu 10

150 K. C. GREY

dryness with previous admixture of sodium bicarbonate, and the powdered
residue extracted with anhydrous ether till the cholesterol was completely
removed. The details of this process are described elsewhere [Grey, 1912].
The fatty acids were now converted into lead soaps and treated with ether,
which was done with the aid of the centrifuge. The lead soaps insoluble in
ether yielded acids with an iodine value of 275 °/,. From 150 g. of total
fatty acids were obtained 67 g. of acids forming lead soaps insoluble in ether.
The mean molecular weight of the solid acids was 320 as determined by
titration. ‘The iodine absorption is due chiefly to admixture with a solid
unsaturated acid described later. The fatty acids from the lead soaps insoluble
in ether were now neutralised with potash. The precipitate which separated
on cooling is represented by fraction No. 1 in the accompanying table, and
consists chiefly of hydroxy-acids. The filtrate was subjected to fractional
precipitation by magnesium acetate, and the fractions 2-5 obtained. The
substance remaining in solution (fraction 6) consists chiefly of the unsaturated
acid melting at 42° to the presence of which the iodine-absorbing power of
the solid fatty acids is due, earlier fractions being practically devoid of iodine
value. The large loss of material was due to the volatilisation of the acids
by steam.

TABLE I. Showing the separation of 58 g. of solid fatty acid from
the human brain.

No. of Melting Mean mol. weight
fraction — How obtained ‘ Weight point by potash titre
No. 1 From potash soap insoluble 14°6 84 -84:5° 420
‘in cold alcohol
No. 2 From first precipitate of 6:8 68°5-69° 399
magnesium soap
No. 3 2nd ditto a Bh 6:5 55°5-56° 324
No. 4 3rd ditto na ee 54 57 —57-5° 303
No. 5 4th ditto sats se 8:0 47-2-47°6° 251
No. 6 Residue in solution Be 6:0 40 -4]° <a
Loss 10:7

The hydrowy-acids of the human brain.

These were obtained from fraction No. 1 (vide Table 1) and represent
the acids forming potash soaps sparingly soluble in alcohol. They also
separate from fractions 2 and 3 during the further purification of these.
The potash soap was brought into solution by boiling alcohol and
partially precipitated by magnesium acetate. The hydroxy-acid mixture so
obtained was further fractionated by cooling the hot alcoholic solution of the
‘acids, by the repetition of which process a fraction was obtained almost

E. C. GREY 151

insoluble in alcohol even when hot. This acid when dry was a white powder,
and melted at 100-101°. The acid is not cerebronic acid. The main portion
of the hydroxy-acids consisted of two other substances which were separated
by recrystallisation from acetone and glacial acetic acid. The portion most
soluble in acetone melted at 73°5°. The same substance was also obtained
from the earlier fractions of magnesium soaps obtained from the potassium
soaps soluble in alcohol.

In the course of these investigations the possibility of the formation of
hydroxy-acids was considered, but since the chief hydroxy-acids found were
not dihydroxy-acids, but monohydroxy-acids, they could not have been
artificial products. It is possible however that small quantities of dihydroxy-
acids may have resulted from the oxidation of unsaturated acids present.
Experiments done with coal gas yielded practically the same iodine value as
when no special precautions were taken. Moreover the isolation of acids of
great degree of unsaturation and in particular of one containing six double
bonds, indicates that oxidation during manipulation does not play any
considerable part.

Nature of the hydroxy-acids.

Three distinct hydroxy-acids have been separated, but small quantities of
others seem also to be present. The three acids isolated melted at 73°5°, 91°
and 100-101° respectively. The melting point of this last corresponds to
that of cerebronic acid, the quantity obtained was small, but judging by the
result of a single combustion analysis the substance is certainly not cerebronic
acid. On the other hand the acid melting at 91° corresponds closely in
composition with cerebronic acid as determined by many analyses. The
melting point of this substance remained constant after many recrystal-
lisations from acétone and glacial acetic acid. The analyses of these acids
are set forth in the accompanying table.

TABLE II.

Melting Weight Empirical
No. point taken CO, H,O C%, H %, formula
1 100-101° 0°1296 0°3395 0°1376 71:44 11°80 C);H 3403 or
CopH 440,
2 (a) 91-0° 0°1077 0°2960 0-1193 74°956 12°308
(b) a 0°1034 0-2851 — 75174 —-
(c) Py 0°1967 -- 0-2270 _- 12-530
(d) ” 0°1327 0°3660 0-1500 75°221 12-544
(e) a 02267 0°6267 02567 75°420 12-590
Meau 75°193 127493 Co,HgoOz
3 73°5° 0°1433 00-3804 0°1730 72°398 12-073
iy 0°1323 0°3526 -- 72614 ~-

' Mean 72-506 ~ 12-073 CopH yo03
10—2

152 EK. C. GREY

The composition of Thierfelder’s cerebronic acid (C.;H0;) 1s C = 75°38 ine
H = 12°56 with which substance No. 2 will be seen to be in good agreement.

Substance No. 1 corresponds to the formula C,,;H;,O, or C..H,O,. The
latter formula would be preferable since according to its melting point and
other physical properties it can hardly be a lower member of the cerebronic
acid series. In this case the acid would be a dihydroxy-acid and might have
resulted by the oxidation of a pre-existing unsaturated acid of the formula
C,,H,,.0,. Such an unsaturated acid has not yet been described as occurring
in brain lipoids.

The acid melting at 73°5° corresponds approximately to the formula
C.,HO; but it cannot be regarded as having been yet satisfactorily examined.
It is characterised by the readiness with which it may be transformed into a
non-acid lactone-like substance, as already explained. -

The unsaturated fatty acids of the human brain.

Since the discovery by Henriques and Hansen [1903] of fatty acids in the
brain more unsaturated than oleic acid, no separation of the unsaturated
acids seems to have been attempted. The following preliminary separation
is therefore described.

About two years ago the author observed that when bromine was added
to the ether solution of the unsaturated fatty acids, the temperature being
kept at 5°, a white precipitate separated which was found to contain
78-79 °/, bromine, and therefore corresponded to a fatty acid more un-
saturated than any previously described as occurring in lipoids. The amount
of such highly unsaturated acid was small, nevertheless the result will be
seen to be proof of the existence of unsaturated acids of the nature of
clupanodonic acid in the human brain. The bromine separation has also
shown the presence of the linoleic and linolenic acid series.

Separation of the unsaturated fatty acids.

An amount of 241 g. of unsaturated acids was dissolved in ether and
cooled to 5°, and bromine added until a definite excess was present; the
solution was kept in an ice chest all night, and the precipitate (0°52 g.)
separated by decantation, and washed in centrifuge tubes with ether until all
bromine compounds soluble in ether were removed. An aliquot portion of
the ether solution of bromides was evaporated as a check on the amount in
solution. From the weight of bromides obtained it is easy to determine the
percentage of bromine absorbed by the acids, which was found to be 77°5 °/o.

< 4

i. C. GREY 153

The ether was now removed from the main quantity of the soluble bromides,
the latter washed free from acetic acid by hot water, and after drying shaken
with ligroin of boiling point below 45°. A separation immediately followed,
the larger portion being soluble while another portion was apparently
insoluble in ligroin, but dissolved readily in chloroform. The quantities
found together with the percentage of bromine in the fractions is set forth
in the following table.

TABLE III. The separation of 24:1 g. of unsaturated acids.

Weightof Fatty acid Percentage of the
bromide which this Bromine acid in total un-
Fraction of fatty acids found represents _—_ per cent. saturated acids
Clupanodenie acid ay a 0°52 g. 0°13 777 0°48
Acids of linoleic and linolenic series 5°64 2°37 58°0 9°83
Oleic acid series oe ea 34°23 21°43 37°4 88°92

Properties of the individual fractions of the unsaturated fatty acids.

(1) Bromine compound insoluble in ether. 'To the acid yielding this
bromine compound I have given the name clupanodenic acid since it is still
more unsaturated than clupanodonic acid. Although the quantity used in
the analysis was small and the results must be regarded as only indicating
the existence of such highly unsaturated acids, nevertheless the substance is
quite definite and I have since prepared a gram of the bromide by precipi-
tation of another quantity of unsaturated acids.

The bromide is a white powder, which turns a faint straw colour when
heated to 120°. With further heating in a small tube it was found to
blacken at about 200° and to decompose with further application of heat
giving a sublimate and odour of stearic acid. Ignited on platinum foil the
substance blackened and melted, burning without leaving any appreciable
residue. The bromine compound is practically insoluble in ether, chloroform,
and benzene, and only slightly soluble in cold alcohol. In hot alcohol it is
more soluble and exhibits the properties of an acid. Titrated with alcoholic
potash the molecular weight was found to be over 1000. This number was
slightly low on account of the access of carbon dioxide owing to the slowness
of the titration, and the titration was moreover carried out on a very small
amount of material.

The bromine percentage was determined in two ways.

Data from bromine determinations.

(a) The substance was heated with nitric acid and silver nitrate.
0.2160 g. of bromo-acid gave 03946 AgBr. Br=77°7 °/).

154 BE. C. GREY

(b) The substance was mixed with calcium oxide and ignited as in the
method of Piria and Schiff.
0:0347 g. gave 0:0650 AgBr. Br=79-2°/,.

Calculated for dodecabromostearic acid Br= 77:9 °/,.

(2) Brominated acids soluble in ether but insoluble in ligroin. This
fraction should contain any acids of the linoleic and linolenic series, and the
percentage of bromine found is in agreement with this and points to the
mixture consisting of both tetra- and hexa-bromine compounds. It would be
premature however in view of the complexity of these lipoid fatty acids, to
calculate from the bromine figure the proportions of the bromine compounds
of either series of which this mixture may consist. During the process of
treament with ligroin it was readily seen that at least two substances were
present in the portion insoluble in that fluid for while most of the insoluble
material was of a dark colour and of a plastic adhesive nature, a small portion
was separated by repeated shaking and decantation which was of a light colour
and amorphous. This amorphous material became of a glassy transparent
nature when dried at 100° but did not melt at that temperature, whereas
the main mass of the bromine compound which was insoluble in ligroin
melted below 100°.

0°1556 g. of bromine compound heated with lime gave 0°2094 g. AgBr.

Br = 57-4°/).
Calculated for linoleic tetrabromide Br = 53°33 °/,.
- » linolenic hexabromide Br = 63°33 °/.

(3) Bromine compounds soluble in ether and also in ligroin. This fraction
is generally regarded as consisting of the dibromide of ordinary oleic acid.
But in the present case it was found that the liquid fatty acid was admixed
with a solid substance corresponding in composition to oleic acid and likewise
forming a dibromide soluble in ether and ligroin. The fatty acid has not yet
been examined in detail but it is evident that it occurs to a considerable
extent mixed with the unsaturated liquid fatty acids of the brain and is
difficult to separate from them. The acid is likewise obtained from the solid
fatty acid mixture regained from the lead soaps difficultly soluble in ether,
and is found left over in this separation, after fractional precipitation of the
saturated acids as magnesium soaps (vide fraction 6, Table I).

The total bromides soluble in both ether and ligroin gave on analysis a
percentage of bromine corresponding to oleic dibromide.

(a) 04507 g. gave 03880 AgBr. Br=36°6°/).
(b) 04507 g. gave 04050 AgBr. Br =38-2°/,.
Calculated for oleic dibromide Br = 36:2 °/,.

4g

E. C. GREY 155

Solid unsaturated fatty acid. This acid can only be briefly referred to.
Further examination will be needed definitely to establish its nature. The acid
was obtained from fraction No. 6 of the original magnesium soap separation,
and also by exhaustion of the ether-insoluble lead soaps by hot benzene. It
was purified by repeated crystallisation from acetone at a temperature of 5°,
The melting point remained practically constant at 42°. The combustion
figures show the substance to be of the nature of oleie acid.

(a) O1185 g. gave 0°1286 H,O.
(b) 0°1323 g. gave 0°1467 H,O; 0°3703 CO.,,

Calculated for

a b Mean C1sH4,0. C; & HO.
Cc — 76°34 76°34 76°60 75°66
H 12°10 12°32 12°21 12-06 12°61

The saturated fatty acids of the human brain.

The further separation of these acids obtained from fractions 2—5 of the
original magnesium soap precipitation (vide Table I) was carried out in the
usual way. The fractions were subdivided by precipitation with barium
acetate and subsequently converted into silver salts which were submitted to
combustion analysis. The fatty acids were regained from the silver salts and
the melting points determined. In many cases the acids themselves were
submitted to combustion analysis. As already pointed out, no acid of more
than eighteen carbon atoms was found. The acids isolated and identified
were ordinary stearic, palmitic and myristic acid, and in addition, in quantity
quite equal to that of ordinary stearic acid, an isomer of stearic acid melting
at 51-52°. This therefore confirms the discovery of Thudichum [1901] that
the stearic acid occurring in greatest amount in the brain is not ordinary
stearic acid melting at 69° but an isomer melting at 52° to which he gave
the name neurostearic acid.

GENERAL CONCLUSIONS,

1. At least 25 °/, of the solid fatty acids of the human brain are
hydroxy-acids, of which three have been isolated. Two of these at least are
monohydroxy-acids and cannot therefore have been produced artificially by
oxidation of unsaturated fatty acids.

2. The unsaturated acids contain besides oleic, linoleic and linolenic acids
an acid still more unsaturated to which the name clupanodenic acid has been

156 E. C. GREY

given. It combines with twelve atoms of bromine. Such a compound has
not previously been described as occurring in fats or oils.

There is present also a solid unsaturated acid melting at 42°, which is
probably an isomer of oleic acid.

3. The saturated fatty acids contain normal stearic and palmitic acids
and an isomer of stearic acid melting at 51-52°. The presence of myristic
acid has also been confirmed.

4. In the resemblance between the hydroxy-acids of the brain, and those
of lanoline, there seems to be additional evidence of the relationship of the
nervous tissue to other tissues of epiblastic origin.

In conclusion I would express my thanks especially to Dr H. G. Chapman,
at whose suggestion this work was begun, and to Professor T. P. Anderson
Stuart, in whose laboratory the investigation was carried out.

REFERENCES.

Darmstaedter and Lifschiitz (1896), Ber. 29, 1474; 2890.

Frankel, 8. (1909), Biochem. Zeitsch. 21, 337.

Grey, EH. C. (1910), Proc. Linn. Soc., New South Wales, 35, 295.

(1912), Australian Assoc. Adv. Sci., 13, 70.

Henriques and Hansen (1903), Skan. Arch. Physiol. 14, 3.

Koch, W. (1902), Zeitsch. physiol. Chem. 36, 134.

Thierfelder (1905), Zeitsch. physiol. Chem. 43, 215, 44, 366.

Thudichum (1901), Die chemische Konstitution des Gehirns der Menschen und der
Tiere, Tiibingen.

XV. AN INVESTIGATION OF PHYTIN.

By ROBERT HENRY ADERS PLIMMER anp
HAROLD JAMES PAGE.

From the Ludwig Mond Research Laboratory for Biological Chemistry,
Institute of Physiology, University College, London.

(Received January 11th, 1913.)

The chemistry of the substance phytin has attracted the attention of
workers in all parts of the world. It is generally regarded as the calcium
magnesium salt of inositol phosphoric acid (phytic acid), but the presence of
organic constituents other than inositol has been described.

OCCURRENCE.

Phytin occurs in the seeds of a large number of plants and was first
extracted from them by Palladin [1895], who distinguished it from coagulable
protein. Schulze and Winterstein [1896] prepared it by the same method and
showed that it contained phosphoric acid, calcium and magnesium. They
supposed it to be identical with the aleurone or globoid bodies described by
Pfetfer [1872]. A better method of preparation was discovered by Posternak
[1903] and he extracted the substance from various seeds. Hart and
Andrews [1903] found that the phosphorus contained in the bran of cereals
was chiefly in an organic form of combination, and Patten and Hart [1904]
isolated a substance from wheat bran which they regarded as identical with
phytin; the same substance was isolated from other cereals by Hart and
Tottingham [1909] and Anderson [1912, 4] has prepared it from cotton seed
meal. The distribution of phytin in various plants has been investigated by
Suzuki and Yoshimura [1907], Staniszkis [1909], and Geys [1910].

PREPARATION.

Palladin and Schulze and Winterstein prepared phytin by extracting the
fat-free seeds of Sinapis nigra with 10 per cent. sodium chloride solution,
boiling to remove coagulable proteins and filtering when the solution had

158 Rk. H. A. PLIMMER AND H. J. PAGE

cooled. On again boiling the solution the phytin was precipitated ; it was
filtered off, washed with boiling water and dried.

The method of preparation used by Posternak has been patented and
phytin is now a commercial article (manufactured by the Gesellschaft fiir
chemische Industrie in Basle). The preparation is as follows.

Oil-cake is extracted with dilute hydrochloric acid (0°5-1 per cent.) and
the extract is freed from protein by one of the usual methods. Sufficient
sodium acetate is added to remove free hydrochloric acid and then the
necessary amount of a soluble calcium salt is added to bring the calcium
content of the solution to a fixed amount. The phytin is completely
precipitated with copper acetate as a complex. copper, calcium and magnesium
salt. The precipitate is washed with distilled water and decomposed with
hydrogen sulphide. The filtrate from the copper sulphide is evaporated in
vacuo to a syrup, treated with alcohol and the solid mass dried and powdered.

The extract may be freed from impurities by oxidation, and organic acids
such as formic may be used instead of hydrochloric acid.

Phytic acid may be prepared by extracting acidified solutions of phytin
with a mixture of one volume of alcohol and three volumes of ether. A syrup
is obtained on evaporation which may be precipitated as copper salt as
before.

Acid salts of phytic acid may be prepared by dissolving phytin in dilute
mineral acids and precipitating with alcohol; with just sufficient acid to
dissolve the phytin the precipitate is insoluble in water, but 1f more acid be
used a precipitate of the soluble acid salt of the type CaH.X is obtained; with
still more acid the precipitate is the acid salt of the type CaH,X.

COMPOSITION OF PHYTIN.

The method of preparation of phytin from extracts of seeds and bran does
not exclude the presence of calcium and magnesium phosphate in the
preparations which have been made. .

There are many conflicting statements concerning the precipitation of
phytin by ammonium molybdate; some authors have stated that it is
precipitated, some that it is not precipitated ; some have said that it gives
a white precipitate, others a yellow precipitate. The yellow precipitate has
been attributed by many observers to impurities in the phytin. Starkenstein
[1910] claimed to have shown that phytin contained inorganic phosphate and
that it was decomposed by drying at 100°, but Jegoroff [1912] showed the
contrary. Jegoroff pointed out that there is no satisfactory method for

nl ries

R. H. A. PLIMMER AND H. J. PAGE 159

estimating inorganic phosphate in the presence of phytic acid though in some
cases that of Schulze and Castoro had been of use. Fingerling and Hecking
[1911] have also complained of the absence of a suitable method for the
estimation of inorganic phosphate in extracts of fodders containing phytic
acid.

None of the investigators mention the methods by which they have
ascertained the calcium, the magnesium and the phosphoric acid content of
phytin. Most of the methods for the estimation of calcium and magnesium
in the presence of phosphoric acid are long and involve possibilities of error.
This may account for the very different values which have been found, e.g. :

Cu Mg 1%, Na K Observer
CaMg salt is -~ _- 15°13 — — Schulze and Winterstein.
Mg salt ... at a 7°78 18°44 — — _ Winterstein.
CaMg salt fe 13°42 8°05 22°30 — — Posternak [1900] *.
CaNa salt ses 8-41 - 19°42 18°79 — “3 {1903}.
¥ He 8°16 me 19°13 1902 — 2 _
CaMgK salt (wheat) 1:13 5°80 16°38 — 2°60 Patten and Hart.
- (maize) 0-48 6°70 14°17 — 5°60 Hart and Tottingham.
a (oats) 8-60 4:70 16°70 = 1°50 2 x
oe (barley) 0°00 7:90 14°46 — 11°20 - ee
Pe (rice) 5°18 17°48 23°48 — trace Suzuki and Yoshimura.
CaMg salt
(commercial) 11°96 1°45 20°32 —_ —_ Horner.
CaMgkK salt
(from phytic acid) 13°03 4°29 19°09 — 6-42 Anderson [1912, 2].
CaMgKk salt
(wheat bran) 2°82 4:72 14°42 = a » {1912, 3}.

The carbon, hydrogen and phosphorus content of phytin and of phytic
acid as given by various workers are also very different :

Cc H 12 Observer
Phytic acid... dss 9°87 3°70 25°89 ~—-Posternak.
+ ae eo 9°97 3°66 26-00 -
Fe (wheat) ... 10°63 3°38 25°98 Patten and Hart.
-e (maize)... 10°30 _- 26°07 Hart and Tottingham.
- (oats) Bae 10°22 — 25°97 + +
PA (barley)... 10°54 = 25°88 - i
% (cotton seed) Uber Al 3°07 26°35 Anderson [1912, 4].
cs ft 10°89 3°25 27°11 5 is
a5 i 11:02 3°30 26°36 " A
Calculated for CgH240v;P, 10°08 3°36 26°05 —
CaMg salt i 9°65 2°88 15:13 Schulze and Winterstein.
CaNa salt aa ee 7°25 1°34 19-42 Posternak.
“ Ee ee 7°43 1:49 19°13 a
CaMgK salt (wheat bran) 17°30 3°63 16°38 Patten and Hart.
5 re 21°47 3°69 14°42 | Anderson [1912, 3].

* Cited in Jegorofi’s paper.

160 R. H. A. PLIMMER AND H. J. PAGE

The difficulty of burning all the carbon in the presence of phosphoric
acid has never been mentioned and the different values may be assigned to
this cause.

THE PREPARATION OF PHytTICc ACID.

Phytic acid was prepared by Posternak, who described it as a_ thick
transparent brownish syrup. Patten and Hart again prepared phytic acid
from the phytin of the bran of wheat. They pointed out the great difficulty
in obtaining it free from calcium (compare the precipitation of phytin with
copper acetate) and they adopted the following procedure. The copper
compound which is precipitated from the extract 1s decomposed by hydrogen
sulphide. The filtered solution is made alkaline with sodium hydroxide and
precipitated with barium chloride. The barium salt is washed until free
from alkali, suspended in water and decomposed with the exact amount of
sulphuric acid to combine with the barium. The solution is again pre-
cipitated with barium chloride in alkaline solution and treated as before.
The process is once more repeated and then copper acetate is added. The
copper compound which is precipitated 1s washed, suspended in water and
decomposed with hydrogen sulphide. The filtrate from the copper sulphide
is evaporated to a syrup and the residue is dried at 110°.

CONSTITUTION OF PHytTic ACID.

The evidence for the constitution at present accepted for phytic acid is
based upon the work of Winterstein and of Posternak. Winterstein [1897]
prepared a magnesium salt of phytic acid and obtained on hydrolysis with
concentrated hydrochloric acid at 130° a 60 per cent. yield of inositol.
Posternak analysed the acid and determined its molecular weight by the
freezing point method. He gave it the formula C,H,O,P, and represented it
as anhydro-oxymethylene diphosphorie acid,

Gok ant OA
\cH,--0—PO (OH).

He confirmed the observation of Winterstein that it yielded inositol on
hydrolysis (98 per cent. yield) and explained the formation of this product by
the polymerisation of the formaldehyde which is first formed.

This constitution of phytic acid was challenged by Neuberg [1908, 1] who
quoted Suzuki and Yoshimura’s result that inositol is obtained by hydrolysis
with the enzyme, phytase. Winterstein [1908] was also of opinion that the
inositol nucleus existed in phytic acid. _Neuberg and Suzuki and Yoshimura

R. H. A. PLIMMER AND H. J. PAGE 161
[1907] put forward formulae representing phytic acid as inositol hexa-
phosphoric acid :
Se ee

O O
| |
Ho-——68
| | HO),OP—O—HC—CH—O—PO(OH),
(HO),P—O—HC CH—O—P(OH); Coe bogs t (OH)2
Oo No OR eT ep an
(HO),P—O—HC 6H_0—K (OH), (HO),0P—O—HC—CH—0—PO(OH),
(Neuberg) (Suzuki and Yoshimura)

Contardi [1909] contributed further evidence to show that phytic acid is
a phosphoric acid ester of inositol: he obtained 18 g. of inositol from 100 g.
of phytin by acid hydrolysis.

Starkenstein [1910] from titrations of phytic acid supposed that phytic
acid was derived from pyrophosphoric acid and Vorbrodt [1910], who analysed

>

a “sphero” crystalline barium salt, believed that the phytic acid molecule
contained two inositol nuclei.

The synthesis of esters of phosphoric acid and inositol has been effected
by Contardi [1910] and by Anderson [1912, 1]. Contardi heated together
inositol and phosphoric acid and obtained inositol hexaphosphoric acid
together with a quantity of inositol diphosphoric acid. The inositol
hexaphosphoric acid agreed very closely in its properties with natural phytic
acid. By the same method Anderson obtained inositol tetraphosphoric acid
as the chief product of the reaction; mositol hexaphosphoric acid was not
formed. Anderson [1912, 2] has also prepared an ester of inositol and pyro-
phosphoric acid. Carré [1911] has maintained that the substances prepared
by Contardi are mixtures of inositol and phosphoric acid, but Contardi [1912]
refuted these statements.

Most observers seem to agree that phytic acid is composed of only
inositol and phosphoric acid, but the data concerning the yield of inositol
obtained on hydrolysis do not support this conclusion. Posternak alone
obtained a nearly theoretical yield; other workers have obtained yields
varying from 20 to 75 per cent. and in the majority of cases the yield has not
been given. Levene [1909] claimed to have separated phytin into two
constituents; one of them yielded inositol on hydrolysis; the other a
carbohydrate which gave pentose reactions and furfural on distillation with
hydrochloric acid. Neuberg [1909] severely criticised the claims of Levene
and showed that inositol gave small quantities of furfural when distilled with
concentrated hydrochloric acid. Considering the scanty data of the yield of

162 R. H. A. PLIMMER AND H. J. PAGE

inositol on hydrolysis, Levene’s claim that another organic constituent is
present in phytin must be considered. Anderson [1912, 3] finds that the
phytin of bran does not consist of one substance, but is a mixture.

The above summary of the work on phytin shows that there are several
lines of investigation which require attention :

(1) The presence of inorganic phosphates in phytin.

(2) The analysis of the calcium and magnesium content of phytin.

(3) The removal of calcium in the preparation of phytic acid.

(4) The yield of inositol by the hydrolysis of phytin or phytic acid.

(5) The presence of other organic compounds besides inositol in phytic
acid.

Our investigations are not complete but we believe we have sufficient
data to justify their publication, especially as our joint work has ceased for
some time, although we hope to continue it at some future date.

EXPERIMENTAL.

Commercial phytin has been employed almost entirely in these in-
vestigations, but for comparison phytin (neutral calcium phytate) was
prepared from bran. 200 g. of wheat-bran were extracted with successive
quantities of N/5 hydrochloric acid until no appreciable amount of phosphorus
was contained in the extract. The combined extracts were made alkaline
with ammonia; this caused the precipitation as neutral calcium phytate of
the whole of the calcium in the extract. The residual phytic acid in the
solution was precipitated by adding calcium chloride solution until no further
precipitation ensued. The whole was then filtered and the precipitate
thoroughly washed with water and dried. The product was dissolved in
dilute hydrochloric acid and again thrown down by adding ammonia. The
precipitate was re-dissolved in hydrochloric acid and the solution decolourised
by shaking with charcoal. After filtering, the phytin was precipitated with
ammonia and filtered off. It was washed four times by removing from the
filter, triturating with distilled water and filtering, and then dried at 100°.
Yield = 6 g. or 3 per cent.

ANALYSIS OF PHYTIN.

(1) Estimation of the Total Phosphorus. The total phosphorus content
of phytin was estimated by the method of Neumann as modified by Pliimmer
and Bayliss [1906]. Several samples were analysed. The analytical date
are all given on p. 168,

R. H. A. PLIMMER AND H. J. PAGE 163

(ii) Estimation of Inorganic Phosphorus. Inorganic phosphate in the
presence of organic phosphates can be readily estimated by precipitation with
ammonium magnesium citrate as ammonium magnesium phosphate and
conversion into magnesium pyrophosphate as was shown by Plimmer and
Bayliss in the case of caseinogen. ‘The same method can be applied to the
estimation of inorganic phosphate in the presence of glycerophosphate and
ethyl phosphate, as was shown by one of us in a previous communication, but
in the presence of phytic acid this method cannot be employed. The solution
of phytic acid obtained after removal of the calcium by oxalic acid or an
oxalate not only inhibits the precipitation of small amounts of inorganic
phosphate as ammonium magnesium phosphate but also, when larger
amounts are present, renders the precipitation slow and the precipitate
consists of ammonium magnesium phosphate and ammonium magnesium
phytate. This is shown by the following experiments.

Various amounts of a solution of sodium phosphate were added to a
solution of phytic acid prepared by adding oxalic acid to 10 g. of phytin
dissolved in water, filtering and diluting to one litre.

A. 10 ¢.c. phytic acid sol.-+1 ¢.c. NagHPO, sol. gave no ppt.
+another 1 c.c. NayHPOy, sol. gave no ppt.

+yet another 1 c.c. Na, HPO, gave ppt.=0-0692 g. P.O;.

Hence ppt. from
phytic acid

B. 10c.c. phytic acid sol. +1¢c. Na,.HPO, sol. gave ppt.=0°0313 g. P.O;. =0-0072
” » +2 c.c. ee = =0°0629.. ;, =0-0147
xe +3 ¢.c. “A “f =0°0925_ ,, =0-0202
LEO Gr > * = 0°0241
C, 10 c.c. phytic acid sol.+5 ¢.c. NazHPO, sol. gave ppt. =0-°1187 g. P20;. =0-0075
as 2 +5 €.¢. = _ =1300) 4; =0°0238
5 ¢.¢. a ie =

The estimation of inorganic phosphate in the presence of phytic acid was
carried out by Hart and Andrews [1903] by precipitation with ammonium
molybdate at 65° in the presence of very dilute nitric acid. This method
was used by McCollum and Hart in their phytase experiments. Under these
conditions there is a possibility that phytic acid is hydrolysed with the
liberation of inorganic phosphate. Experiments were therefore made to
ascertain if inorganic phosphate was completely precipitated by ammonium
molybdate in the presence of seminormal to normal nitric acid at room
temperature. The yellow solution which is first formed on adding ammonium
molybdate to a solution of sodium phosphate in the presence of dilute nitric
acid gradually deposits ammonium phosphomolybdate in a crystalline state
and the solution becomes colourless in 24 to 48 hours, On determining the

164 R. H. A. PLIMMER AND H. J. PAGE

amount of P.O, in this precipitate by the Neumann method, by dissolving it,
after filtration and washing, in semi-normal sodium hydrate, boiling off the
ammonia and titrating with semi-normal sulphuric acid it was found to be
quantitative :

1 c.c. NayHPO, sol. gave Mg.P20;=0°0241 g. P,O;.
iverc: 33 » +25 c¢.c. H20+25 c.c. 2N. oe gave Vee g. P05.
+25 c.c. Am. molybdate sol. 0:0219

29

Applying this method to the solution of phytic acid the following

amounts were obtained:
ee g. P25.
gave ~ 0:0042_ ,,
0:0044_ =O,
Total P.O; in 10 ¢.c. phytic acid sol. by Neumann’s method=0-0871 gm.

10c.c. phytic acid sol. + 25¢.c. 2N.HNO3+ 25¢.c. H2O}
+10 ¢.c. Am. molybdate sol.

Further proof that the estimation of inorganic phosphate in the presence
of phytic acid can thus be effected is given by the following experiments
which were carried out in the same way:

(a) 1e.c. NagHPO, (b) 10 c.c. phytic (c) Le.c. NagHPO, Sum of
acid sol. +10c.c. phytic (a) and (b)
acid sol.
0:0223 g. P.O; 0:0044 g. P.O; 0:0264 g. P20; 0:0267
0:0226 ,, 0:0054 , 0:0270 eee: 0:0281
0:0222 ,, 000515 as; O-0270iiees 0:0273

Hydrolysis of the phytic acid by dilute nitric acid at room temperature
therefore does not occur. Employing this method of estimation it has been
shown that hydrolysis does not occur at 37° but at 65° and at 75° phytic acid
is slowly hydrolysed [Plimmer, 1913, 2, p. 79].

If proteins be present in solution they are precipitated by phospho-
molybdic acid with the ammonium phosphomolybdate. The estimation can
be carried out by filtering the yellow precipitate, washing free from acid,
dissolving in sodium hydrate and then precipitating in the usual way with
magnesium citrate mixture and ammonia and _ finally converting into
magnesium pyrophosphate.

The above three solutions after the estimation by Neumann’s method
were filtered and the inorganic phosphate estimated as magnesium pyro-
phosphate :

(2) (d) (c) (a) + (2)

0:0239 g. P.O; 0:0042 g. P.O; 0-0281 g. P.O; 0-0281
00240, a 0-0290 ,, oe
00239, 00050 ,, 0-0289 0-0289

The estimation of inorganic phosphate in phytin can be readily effected
by this method; the presence of calcium does not interfere with the

R. H. A. PLIMMER AND H. J. PAGE 165

precipitation. A known weight of phytin is dissolved in normal nitric acid,
excess of ammonium molybdate is added and the solution is allowed to stand.
When the solution is colourless the crystalline precipitate is filtered off and
estimated by the Neumann method, or by filtering, washing free from acid,
dissolving in sodium hydrate and precipitating as ammonium magnesium
phosphate.

The results of the estimations in different samples of phytin are given on
p. 168.

(iu) Lstimation of Calcium and Magnesium. The previous investigators
have not stated the method which they employed for the estimation of these
elements in phytin. They have in all probability precipitated the calcium
as oxalate and converted it into calcium oxide, either before or after oxidation
of the organic matter, and estimated the magnesium in the filtrate as
magnesium pyrophosphate. This method involves several possibilities of
error; if phosphoric acid be not removed it may be thrown down with the
calcium oxalate and some of the magnesium may also be precipitated as
magnesium oxalate. The method of separation of these constituents in a
mixture is again a long and tedious process.

If oxalic acid or an oxalate be added to a solution of phytin a precipitate
is formed, but this precipitate does not consist of pure calcium oxalate. T his
is shown by the following experiment :

10 g. of commercial phytin were dissolved in decinormal acetic acid and
filtered from a small quantity of insoluble matter. This was dried at 110°
and analysed : it contained 13°74 per cent. of calcium (estimated as CaSQ,,
see later) and 17-4 per cent. of phosphorus (by Neumann’s method). The
filtrate was precipitated with 4:1 g. of oxalic acid dissolved in 50 c.c. of water.
The precipitate was washed and dried at 110°; it weighed 5 g. and was
analysed :

0:2188 g. heated to redness to constant weight gave 0°0956 g. residue. Calculated amount
for pure calcium oxalate is 0-0957 g.

0°3364 g. gave 0°2782 g. CaSO, (see later) and 0-0031 g. P by Neumann’s method.
0°3030 g. gave 0:0028 g. P.

Ca=24°31 per cent. (COO) .Ca requires Ca=31°25 per cent.
P = 0°92 per cent.
P = 0°93 per cent.
Another preparation of calcium oxalate from phytin contained 077 per
cent. total P and 0°27 per cent. imorganic P.
Calcium can therefore not be estimated by precipitating with oxalic acid.
Aron [1907] showed that the calcium in milk, urine, ete., could be
estimated as calcium sulphate after oxidation of the organic matter by the

Bioch. vi 11

166 R. H. A. PLIMMER AND H. J. PAGE

Neumann method. When the oxidation is completed the nitric acid is
removed as far as possible by diluting with water and again evaporating.
The sulphuric acid which remains is then diluted and four volumes of alcohol
are added. The calcium sulphate is precipitated and is filtered off, dried and
weighed.

Beyond these few estimations of calcium Aron made no other analyses,
but he states that other elements might be estimated in the filtrate.

This method therefore required verification and extension for the estima-
tion of magnesium. Analyses of mixtures of calcium chloride, magnesium
sulphate and sodium phosphate were consequently made to test the method.
The mixtures were evaporated down in a small conical flask with 10 cc. of
sulphuric acid; the acid was diluted with water and precipitated with four
to five volumes of alcohol. The calcium sulphate was filtered off on a Gooch
crucible, and dried at a red heat. The magnesium was estimated in the
filtrate as pyrophosphate after evaporating off the alcohol, oxidising the
remainder of the alcohol with nitric acid, diluting with water and making
alkaline with ammonia. Excess of phosphoric acid was present in the
mixtures to precipitate the whole of the magnesium. The results were:

Calcium Magnesium
| (ST Pa <2 = on SSS
Taken Found Taken Found
0-1736 0°2540 0°1484 01490
0°1736 02066 0:0297 0:0310
0:0347 0:0523 01484 0°1493
0°1736 0-1885 01484 0°1495
0-0264 0-0569 0:0337 0-0332 ,
0:0264 © 0:0328 0:0337 0°0332
0-0264 0:0588 0:0337 0-0329

Magnesium can thus be correctly estimated but the values for calcium are
too high. The high values are due to the presence of sodium in the solution
as is shown by experiments with mixtures of calcium chloride and sodium
phosphate and calcium chloride and phosphoric acid in various proportions :

Calcium | Calcium
ee SS SS In presence of - In presence of
Taken Found sodium phosphate | Taken Found phosphoric acid
0:0233 0-0285 5 c.c. | 0-0233 0:0233 10 c.c.
0:1038 (0) GG | 0-0233 15 ¢.¢.
0-1944 15 c.c. | 00232 20 c.c.
0-2740 20 c.c. | 0-0231 25 ¢.¢.

The presence of sodium salts in the solution thus causes a serious error in
the estimation of calcium as calcium sulphate.

If the amount of alcohol used in the precipitation be reduced and if the
amount of sodium or potassium in the solution be not greater than 0'1 g. the

—

ee ew ee

R. H. A. PLIMMER AND H. J. PAGE 167

estimation of calcium can be accurately effected, especially if not more than
one volume of alcohol be used in the precipitation :

3 volumes alcohol 2 volumes alcohol 1 volume alcohol
SSP ie ta Se ee ee —— a TE on OT ee Pa Pe ee ae eee
Calcium Amount of Na Calcium Amount of K Calcium Amount of K
found present, in g. found present, in g. found present, in g.

00231 0-0 00232 0-0 0:0231 0-0

0°0231 0-0 0°0245 0:0394 0-0228 0-0

0-0284 071228 0-0281 0-0591 0°0230 0°0197
00268 0-0614 0°0383 0°0788 00230 00394
0-0258 0:0307 0:0237 0°0197 0:0228 0-0591
0-0238 00184 0:0238 0:0158 0:0230 00788
0:0233 0:0123 0:0237 0°0118 0:0227 0:0788

The amount of calcium taken in all these experiments was 0°0233 g.

Unless the amount of potassium and sodium (if present) in phytin exceed
20 per cent. (Ol g. in 0°5 g. required for an analysis) the calcium can be
accurately estimated by precipitation with one volume of alcohol as calcium
sulphate and the magnesium can be estimated in the filtrate as magnesium
pyrophosphate. The calcium and magnesium content of various samples of
phytin have been estimated by oxidising the organic matter with 10 ce.
sulphuric acid and nitric acid, removing the nitric acid by diluting and evapor-
ating, adding 10-20 c.c. of water and then one volume of alcohol (20-60 c.c.)
to precipitate the calcium sulphate. Alcohol was removed from the filtrate
by evaporation. On oxidation with nitric acid and on making alkaline with
ammonia the magnesium was precipitated as ammonium magnesium phos-
phate and estimated as pyrophosphate. Excess of P,O; is always present in
phytin: the remainder can be estimated by adding magnesia mixture and
determining it as pyrophosphate. The sum of the figures obtained then
gives the total P.O; content. It has been found to be the same as by the
Neumann method. For the data see p. 168.

(iv) Loss of Weight on Heating and in vacuo over Sulphuric Acid.
Phytin loses weight at 110° and in vacuo over sulphuric acid. The greater
part of the loss takes place during the first heating; the weight continues
to decrease very slowly and loss of weight seems to go on indefinitely. One
specimen was heated at 110° for 54 hours. The loss in weight under both
conditions is about the same, but it takes place more slowly in vacuo.

Part of the loss in weight may be due to alcohol; one of the commercial
specimens examined gave the iodoform reaction.

The gradual decrease in weight on heating is probably due to a slow
conversion into pyrophosphate.

(v) Estimation of Carbon and Hydrogen. In the course of many com-
bustions of organic phosphorus compounds the difficulty of burning all the

11—2

168 Rk. H. A. PLIMMER AND H. J. PAGE

carbon has been noticed. Page [1912] has shown that the values for carbon
in the case of the hydroxyphosphinic acids were 1-2 per cent. below the true
value even after mixing the substance intimately with lead chromate. The
values were still lower by the Fritsch-Messinger wet method of combustion.

Previous observers have never mentioned any difficulty in ascertaining
the carbon content of phytin or other organic phosphorus compounds nor
have they mentioned any modification of the ordinary method which they
used.

One of the specimens of phytin was analysed by mixing intimately with
the finest copper oxide and it seems improbable that all the carbon was not

burnt.
Results of Analyses.

Total Phosphorus. -
Commercial phytin:

Sample 1 01120 g. : 37°8 cc. 5 NaOH. 18°68 per cent.

2 0:0682 g. : 24:7 c.c. a5 20°04 =,

3 Listavry,

4 0:0579 g. : 18°7 c.c. a USE fp

5 0:0694 g. : 24:1 cc. 5 1922 5

6 0:0614 g. : 21:2 cc. 56 Tigeiik A.

a 0:0544 g. : 18°7 cc. ae 19 7 Gm ea

8 0°0550 g. : 18°5 c.c. a SSO 2meee

Bran (1) 0°6704 g. : 15°8 cc. ‘ EBX) =

Phytin from bran 0°1326 g. : 0:0714 g. MgoP.02* 15:00 =,

* Kstimated by Mg method after ppt. of CaSO,.

; Inorganic Phosphorus.
Commercial phytin :

Sample 2 0:2664 g.: 4:3 c.c. * NaOH. 0°89 per cent.
4 0°6152 g. : 10°8 c.c. os O07 re,
5) 0°9642 g. : 17:1 cc. ne O:9S anaes
6 0°6626 g. : 11°7 cc. ¥3 OOS
7 0°5350 g.: 9:5 c.c. ae 0:98 ee.
8 0°8726 g. : 14°8 c.c. ny O9f %
Phytin from bran 0°3524 g. : 16°7 c.c. 59 202) ee
Calcium and Magnesium. Per cent.
Commercial phytin: Ca Mg
Sample 1 0°4552 g. : 0°1984 g. CaSO, : 0:0092 g. MgoP.0,. 12°82 0:44
2 0'5630 g. : 0:°234Ug. ,, : 0:0194¢. is 12:2 0°75
3 04090 g. : 01828 g. ,, +: 0:0104¢. 53 13:15 0°56
4 0°5235 g. : 0°2207g. ,, : 0°0195g. * 12°40 0°81
5 0:4245 g. : O1744'g. 5, 5 0-01 55'¢. 3 12°10 0-78
6 04224 ¢. 2 O17 75g.) 551) saOO1bi p: Bs 12°36 0-78
7 0°4052 g. : 0-1681g. ,, : 0:0151g. 4 12:20 0°81
8 0°4466 g. : 01873 g. ,, : 0:0155 g. 2 12°34 0°76
Phytin from bran 0°1326 g. : 0:0962g. ,, : 0:0000g. 33 21-127 710-0
Bran (1) 3°3430 g. : 0'°0150g. ,, =: 00786 g. 95 0:13 051
(2) 4°4868 g.: 00176 ¢g. ,, : 0-11l6¢. 5 O-1l 0°54
(8) 42028 g. : 0°0228g. ,, : 0:1058¢. a 0°15 0°55

a er

—_—.— ~~ *

R. H. A. PLIMMER AND H. J. PAGE 169

Loss of Weight (a) at 110° and (b) in vacuo over H2SO,.

Commercial phytin :

Sample 3 (a) 5°0229 g. : 0°3425 g.=6°82 per cent.
3 (b) 5-6868 g. : 0°3874 g.=6'81__,,
4 (a) 0:7376 g. 7 0°0544 g.=7°37_
5 (a) 1°7712 g. : 0°1510 g.=8'52 sé,
6 (a) 1°5502 g. : 0:1164g.=7°51 ,,
7 (a) 08472 g. : 0°0628 g.=7°41__,,
8 (a) 21870 g. : 0°1532 g.=7:00 ,,

Carbon and Hydrogen.
Commercial phytin :
Sample 2 0°3052 g. : 0°0815 g. CO, : 0°0917 g. H,O. C=7-28°/,, H=3°33%/,.
0:2392 g. : 0°0655 g. ,, : 0°0740g._,, C=7°47°/,, H=3°44°/,.

If these figures be compared with those of previous workers (p. 159)
considerable differences will be noticed. These figures agree most closely
with those of Horner for commercial phytin and of Anderson for the
preparation made from phytic acid. The total phosphorus content is
generally higher and so is the calcium content, but the magnesium content
is much lower; in all the specimens examined magnesium was less than one
per cent. Commercial phytin seems to be the calcium salt and the
magnesium may be present as impurity.

The carbon content is also lower, but the hydrogen content agrees with’
the analyses of other workers.

The commercial preparations are very constant in their phosphorus,
calcium and magnesium content, but they differ from the preparation from
wheat bran, which was the neutral calcium salt of phytic acid and hence
contained less phosphorus and more calcium; the bran preparation contained
no magnesium.

Analyses of the phosphorus, calcium and magnesium in wheat bran were
made for comparison. It will be observed that bran contains more magnesium
than calcium, which is the reverse of phytin.

PREPARATION OF PHytTiIc ACID.

Phytic acid was prepared from commercial phytin and from the phytin
from bran by the method of Patten and Hart and also by the removal of the
calcium as oxalate. Neither method sufficed to remove the calcium entirely,
and it was at first thought that some hitherto undescribed product was
present.

Preparation 1. 50g. of commercial phytin (sample 3 above) were

170 R. H. A. PLIMMER AND H. J. PAGE

dissolved in a minimum quantity of dilute hydrochloric acid and just
sufficient sodium acetate solution was added to render the solution neutral to
Congo-red paper. Copper acetate solution was then added until no further
precipitation occurred, and the copper salt filtered off by suction. Since it
was impossible to wash the precipitate thoroughly on the paper it was
removed, made into a uniform cream with distilled water and passed through
muslin to ensure the complete breaking up of all lumps. The mass was
filtered and the washing procedure repeated twice. (All subsequent wash-
ings were effected in this way.) The washed copper phytate precipitate was
suspended in water and the mixture treated with hydrogen sulphide until
saturated and set aside for twelve hours. The copper sulphide was filtered
off and washed. The filtrate and washings were exactly neutralised with
sodium hydroxide and excess of barium chloride solution added. The
precipitate of barium salt was filtered off and washed twice as described
above. It was then suspended in water and the exact quantity of sulphuric
acid necessary to combine with the whole of the barium was added, and the
barium sulphate filtered off and washed. The solution was neutralised with
sodium hydrate and the whole procedure of precipitation with barium
chloride and decomposition with sulphuric acid repeated twice. The solution
after removal of the barium for the third time was neutralised with caustic
soda and then acidified with acetic acid. Copper acetate solution was added
and the copper salt so obtained filtered off, washed and decomposed as
described above. The filtrate from the copper sulphide was evaporated to a
syrup at 100°. °

Preparation 2. A specimen of phytic acid was prepared in the same
manner from 5 g. of the phytin from bran.

Preparation 3. 100g. of commercial phytin (sample 4 above) were
dissolved as far as possible in decinormal acetic acid. The insoluble residue
(amounting to 4 g.) was filtered off and washed. To the filtrate was added
the calculated quantity of oxalic acid (37°5 g. allowing for the residue, which
contained 11°81 per cent. of calcium) dissolved in water. The calcium oxalate
was filtered off and a portion of the filtrate tested with a little more oxalic
acid to ascertain if the calcium had been completely removed ; no further
precipitate was produced. Since calcium oxalate is known to occlude
magnesium oxalate to a considerable degree it was thought that the pre-
cipitated calcium oxalate would remove at the same time the small quantities
of magnesium present. The filtrate was therefore evaporated to a small
volume im vacuo; a small amount of calcium oxalate separated out and
was filtered off. The solution was then concentrated to a syrup at 100°.

R. H. A. PLIMMER AND H. J. PAGE 171

The presence of calcium in the syrups of phytic acid.

In the case of all three preparations of phytic acid it was noticed that
after drying at 100° an appreeiable scum or skin formed on the surface and
that when small samples of each specimen were treated with alcohol a small
white floceulent precipitate was formed. Each of the preparations was
therefore dissolved in a small quantity of water and sufficient alcohol added
to precipitate completely the insoluble substance. This was filtered off,
washed thoroughly with alcohol, re-dissolved in water and re-precipitated with
alcohol. It separated as a white flocculent precipitate which rapidly became
stringy and formed a sticky mass on the bottom of the vessel ; on standing
under alcohol for some time it became quite hard and apparently crystalline.
The yields were: (1) 3 g., (2) 13 g., (8) 12 g.

On analysis when dried at 110° these precipitates were. found to contain
phosphorus and calcium, and the third preparation magnesium as well :

Prep. 1 0°2162 g. : 0:0883 g. CaSO, : Ca=12-01 per cent.
2 0:2197 g. : 0°1004g._ ,, >; Ca=13°45_ ,,
3 0:2586 g. : 0°0654 ¢._ ,, a Ci— (40) 5,
0:0357 g. Mg,P,0, : Mg= 3°12 __,,

Prep. 1 0:0527 g. : 21°10 c.c. S NaOH : P=20-79 per cent.
2 0:0227 g. : 8°95 c.c. A PS eae) ee
3 0:0304 g. : 11°80 e.c. 3 een 0) 5,

Another preparation by the oxalic acid method was found to contain
21:1 per cent. of phosphorus, 11°10 per cent. of carbon and 2°98 per cent. of
hydrogen.

The alcoholic filtrates were evaporated to syrups at 100°. They agreed
in most of their properties with the descriptions of other workers.

When dried to constant weight at 100° and in vacuo over sulphuric acid
for three months the syrup from Preparation 1 was analysed :

Prep. 1 0°3444 g. : 0°1498 g. CO, : 0:1108 g. H,O. C=11-86, H=3°57 per cent.
0:0986 g. : 38-0 c.c. S NaOH. P=21-34 per cent.

(CoHsO09P2)3 requires C=10°09 and H=3°36 per cent. P=26-07 per cent.

The carbon and hydrogen figures are higher than those required for phytic
acid whilst the phosphorus figure is much lower. It is very probable that
some alcohol is retained in the syrup and is not removed by drying; it is also
possible that some esterification took place between the alcohol and the

phytic acid.

172 R. H. A. PLIMMER AND H. J. PAGE

FORMATION OF INOSITOL BY HYDROLYSIS OF PHyTIC ACID.

The hydrolysis of phytin by dilute nitric acid at the atmospheric pressure
and at various temperatures has been investigated by one of us [Plimmer,
1913, 2]; even after seventeen days the separation of the phosphoric acid was
not quite complete.

64 ¢. of phytic acid (the syrup after removal of the crystalline calcium
compound in Preparation 3, prepared from 100g. phytin by treatment with
oxalic acid) were dissolved in 250 c.c. of twice normal sulphuric acid and
heated in a corked flask in a boiling water bath for 36 hours; about 55 per
cent. of the phosphorus was then present as inorganic phosphate ; the heating
was continued at 75° in a thermostat for 164 hours; the solution now
contained 79 per cent. of the phosphorus in an inorganic form. The
hydrolysis was proceeded with for another 118 hours after diluting with four
volumes of twice normal sulphuric acid. 84 per cent. of the total phosphorus
was now present as phosphoric acid. Since the hydrolysis proceeded so slowly
it was discontinued. The solution was evaporated to 200 c.c. and an equal
volume of alcohol was added, followed by ether until a permanent cloudiness
was produced. Inositol crystallised out when the mixture was set aside at 0°
for twenty-four hours and was removed by filtration. Two further quantities
were obtained on adding ether to the filtrate and keeping at 0°. The total
yield amounted to 45 g. The alcohol and ether were distilled from the
solution and the volume made up to 500c.c. This solution contained 10°38 g.
total phosphorus and 9:11 inorganic phosphorus, i.e. 1:27 g. organic phos-
phorus or 5-08 g. of phytic acid (25 per cent P). Allowing for the samples
removed during hydrolysis the liquid from which the inositol was isolated
contained 48 g. of phytic acid, and allowing for the organic phosphorus still
present the amount of phytic acid hydrolysed is 42 g. This quantity should
yield 10:06 g. of inositol. Since the amount actually obtained was only
4-5 g. it represents less than 50 per cent. of the theoretical.

The inositol was recrystallised from water, alcohol and ether; it melted at
220-222°. Analysis :

0°2140 g. : 0°2550 g. CO, : 0°1230 g. H,O : C=32°5 per cent., H=6°4 per cent.
CyHy,0¢ requires as ys C=40-0 per cent., H=6-7 per cent.
The figure for carbon is too low. A residue was left in the boat and
amounted to 7 per cent.; it contained particles of carbon which it was
impossible to oxidise completely after many hours in a current of oxygen,

R. H. A. PLIMMER AND H. J. PAGE 173

This combustion again shows the difficulty of burning carbon in the presence
of phosphoric acid. The ash was found to contain 52°14 per cent. of POs.

SUMMARY.

1. Inorganic phosphates in phytin can be readily estimated by
precipitation with ammonium molybdate in semi-normal nitric acid at room
temperature.

2. The calcium content of phytin cannot be estimated by precipitation
as calcium oxalate, but is easily ascertained by precipitation as calcium
sulphate.

The magnesium content of phytin can then be estimated as magnesium
pyrophosphate.

3. There is great difficulty in removing the calcium from phytin in
the preparation of phytic acid.

4. The yield of inositol obtained on the hydrolysis of phytic acid by acids
is not quantitative; we are inclined to believe that another organic constituent
is present in phytin, and hope to carry out further experiments on the
production of inositol by the hydrolysis of phytic acid at some future date.

The expenses of this work were partly defrayed by a grant from the
Government Grant Committee of the Royal Society.

REFERENCES.

Anderson (1912, 1), J. Biol. Chem. 11, 471.

—— (1912, 2), J. Biol. Chem. 12, 97.

—— (1912, 3), J. Biol. Chem. 12, 447.

(1912, 4), J. Biol. Chem. 13, 311.

Aron (1907), Biochem. Zeitsch. 4, 268.

Carré (1911), Bull. Soc. Chim. iv, 9, 195; Chem. Zentr. 1911, 1, 1196.
Collison (1912), J. Biol. Chem. 12, 65.

Contardi (1909), Atti. R. Accad. Lincei, v, 18, 1, 64; Chem. Zentr. 1909, 1, 1102.
(1910), Atti. R. Accad. Lincei, v, 19, 1, 23; Chem. Zentr. 1910, 1, 1032.
(1912), Gazz. chim. Ital. 42, 1, 408; Chem. Zentr. 1912, m1, 186.
Fingerling and Hecking (1911), Biochem. Zeitsch. 37, 452.

Geys (1910), Zeitsch. ges. Brawwesen, 33, 347-349; Chem. Zentr. 1910, u, 982.
Hart and Andrews (1903), Amer. Chem. J. 30, 470.

Hart and Tottingham (1909), J. Biol. Chem. 6, 431.

Horner (1907), Biochem. Zeitsch. 2, 428.

Jegoroff (1912), Biochem. Zeitsch. 42, 432.

Levene (1909), Biochem. Zeitsch. 16, 399.

Neuberg (1908, 1), Biochem. Zeitsch. 9, 557.

— (1908, 2), Biochem. Zeitsch. 9, 551.

—— (1909), Biochem. Zeitsch. 16, 406.

174

R. H. A. PLIMMER AND H. J. PAGE

Page (1912), J. Chem. Soc. 101, 423.

- Palladin (1895), Zeitsch. Biol. 31, 199.

Patten and Hart (1904). Amer. Chem. J. 31, 564.

Pfeffer (1872), Jahrb. wiss. Bot. 8, 465.

Plimmer (1913, 1), Biochem. J. 7, 43.

—— (1913, 2), Biochem. J. 7, 72.

Plimmer and Bayliss (1906), J. Physiol. 33, 439.

Posternak (1900), Rev. gen. Bot. 12, 5, 65.

(1903), Compt. rend. 137, 202, 337, 439; Compt. rend. Soc. Biol. 55, 1190.
Schulze and Winterstein (1896), Zeitsch. physiol. Chem. 22, 90.

—- (1903), Zeitsch. physiol. Chem. 40, 120.

Staniszkis (1909), Anzeiger Akad. Wiss. Krakau, 95; Chem. Zentr. 1909, u, 918.
Starkenstein (1910), Biochem. Zeitsch. 30, 56.

Suzuki and Yoshimura (1907), Bull. Coll. Agric. Tokyo, 7, 495.

Vorbrodt (1910), Anzeiger Akad. Wiss. Krakau, a, 414; Chem. Zentr. 1911, 1, 895.
Winterstein (1897), Ber. 30, 2299.

(1908), Zeitsch. physiol. Chem. 58, 118.

AVI. ON THE RELATIONS OF THE PHENOLS
AND THEIR DERIVATIVES TO PROTEINS.
A CONTRIBUTION TO OUR KNOWLEDGE OF
THE MECHANISM OF DISINFECTION.

PART II. A COMPARATIVE STUDY OF THE EFFECTS
OF VARIOUS FACTORS UPON THE GERMICIDAL
AND PROTEIN-PRECIPITATING POWERS OF THE
PHENOLS.

By EVELYN ASHLEY COOPER, Beit Memorial Fellow.

Lister Institute of Preventive Medicine.

(Received January 13th, 1913.)

Evidence set forth in a previous communication on this subject [Cooper,
1912, 2] suggested that the germicidal power of the phenols is due, not to a
typical chemical action upon the bacterial proteins, as appears to be the case
with formaldehyde, but to a de-emulsifying effect upon their colloidal
suspensions as evidenced by the precipitation of proteins when a certain
phenol concentration is attained.

Since H. Chick [1910] has shown that disinfection by heat is analogous
to the heat-coagulation of proteins, it seems that a similarity in mechanism
exists between the germicidal actions of heat and of the phenols.

Hardy [1899] pointed out that the heat-coagulation of proteins consists
of two processes, namely, (1) The reaction between the protein and hot water
(denaturation). (2) The subsequent separation of the altered protein in a
particulate form (agglutination). It is not at present known whether the
essential stage in the processes of disinfection by heat and the phenols is a
denaturation of the bacterial proteins or a destruction of their colloidal
suspension.

The experimental work described in this paper has been undertaken with
the object of studying more closely the apparent relationship between the
bactericidal action of the phenols and their precipitating effects on proteins.

176 KE. A. COOPER

IL THE INFLUENCE OF CHEMICAL CONSTITUTION UPON THE GERMICIDAL
AND PROTEIN-PRECIPITATING ACTION OF THE PHENOLS.

The experimental method. The minimum concentrations in water of
various phenols and certain other substances have been determined at which
a visible change in the state of aggregation of the protein occurred. Small
alterations in the magnitude of the protein-particles brought about by such
concentrations were indicated by a turbidity; larger alterations by a pro-
gressively increasing precipitation of the colloid as the concentrations of the
phenols rose.

Two proteins were used in the experiments, namely, gelatin and dialysed
ego-albumin; the former was employed in the form of small strips and the
latter as its colloidal suspension in water. The effect on the gelatin was to
produce a turbidity in the strip immersed in the solution under examination.
Throughout the experiments the weight of protein, volume of solution, and
temperature (20°) were kept constant. Observations of the effects of the
various solutions upon the proteins were made immediately after admixture
and also after standing one hour. The results tabulated below are based on
the observations made after the one hour's standing.

The experimental results. The protein-precipitating powers of the various
substances have been compared with their germicidal activities, which are
expressed in the table below in terms of the concentrations required to
disinfect B. typhosus or Staphylococcus pyogenes aureus in 15 minutes under
the standard conditions employed by Chick and Martin [1908]. The
concentrations have been selected from two previous papers on disinfection
(1) by Morgan and Cooper [1912] and (2) by the author [1912, 1], to which
papers the numbers 1 and 2 in the following table (p. 177) refer.

The following parallelisms can be deduced from these results.

1. The germicidal and proteim-precipitating powers of phenol are
decreased by the introduction of hydroxyl-groups into its benzene nucleus
and increased by the introduction cf nitro-groups and also by that of a
methyl group.

2. The monohydric phenols are superior to the monohydric alcohols and
also to acetone both as germicides and protein-precipitants.

3. The effects of substitution in the molecule of phenol upon its
germicidal and protein-precipitating powers are frequently of somewhat
similar orders of magnitude. An exact correspondence was not expected,
firstly, because it is well known that the carbolic acid coefficient of a

KE. A. COOPER 177

substance varies to some extent with the species of micro-organism employed,
and, secondly, because bacterial proteins were not used in the determination

of protein-precipitating power.

TABLE I.

Protein-precipitating power Germicidal power
Minimum conens.atwhich Concentrations killing
precipitation was visible. in 15 minutes.
(Percentages in terms (Percentages in terms
Substance of weight per volume) of weight per volume) Organism
Gelatin 0:1 g. Vol. of liquid 10 e.c.
Phenol es 2°5 0°8 [1] B. typhosus.
Trinitrophenol 0-2 Oss ply Es
(Picric acid)
Resorcinol ... 4:0 2°6 [2] se
Catechol nh 50 We [2 is
Quinol a 50 0°75 [2] . »
Pyrogallol ... 3°0 1:0 [2] a
Phloroglucinol Satd. solns. (5°/,) had >2°3 [2] 35
no visible effect :
p-Nitrophenol 0-9 0-46 [1] Staphylococcus py. aur.
m-Nitrophenol 0°6 0-27 [1] » 9
Trinitrophenol 0:2 0-14 [1] a ‘
(Picric acid)
Ethyl alcohol 50°/, and 100 °/, had 35°0 [1] ” ”
Acetone ae no visible etfect >10°6 [2] = 3
Phenol =e 2:5 OMe x ys
Egg-albumin } ¢.c. 10 °/, solution (0:05 g.). Vol. of solution containing precipitant=4°5 c.c.
Z 08 [1] B. typhosus.
neue oi re 1:0 [1] Staphylococcus py. aur.
m-Cresol __... 0-7 oe g a
0:3 [2] B. typhosus.
Acetone wr 12°0 >10°6 [2] Staphylococcus py. aur.
Methyl! alcohol 2070: — . 35°0 [1] » ”
4 35°0 [1] » ”
ao ue 32°5 [1] B. typhosus.
Propy] alcohol 12-0 14:0 [1] Staphylococcus py. aur.

II]. Tse EFFrects oF VARIOUS SUBSTANCES UPON THE GERMICIDAL
AND PROTEIN-PRECIPITATING ACTION OF THE PHENOLS.

The experimental method employed in the investigation of the precipi-
tating action of the phenols was similar to that described in Section I. The
various substances were added to the phenol solutions immediately before
admixture with the protein. The results recorded in this section are based
on observations of the condition of the proteins after one hour’s contact with

the phenol solutions.

178 E. A. COOPER

A. SoptumM CHLORIDE.

Reichel [1909] showed that the addition of sodium chloride increased
both the bactericidal action of phenol and its solubility in proteins. The
germicidal power was about doubled by 10 per cent. salt and quadrupled
by 20 per cent. salt, and the effects of salt upon the solubility in proteins
were of similar orders of magnitude. Smaller amounts of salt had little
influence.

These experiments indicated how in the presence of salt a larger amount
of phenol became available for disinfection, but did not throw any light on
the mechanism of the germicidal action of this disinfectant. It was therefore
of interest to investigate the influence of salt upon the protein-precipitating
action of phenol.

Gelatin. The minimum concentrations of phenol, which induced a visible
precipitation in the presence of varying amounts of sodium chloride, were
determined. The results are tabulated below:

TABLE II.
Percentage of salt present Minimum phenol concentrations
(weight per volume) precipitating the protein
0 ; 3%,
5 2°5
12 1-5
19 0:75

Although saturated aqueous solutions of m-cresol (1:25 per cent.) had no
visible effects upon gelatin, this protein was precipitated by 0:8 per cent.
solutions. of m-cresol in the presence of 9 per cent. salt and even by
0:4 per cent. solutions in the presence of 20 per cent. of salt.

Dialysed egg-albumin. 3°3 per cent. aqueous solutions of phenol only
induced a faint turbidity in a sample of egg-albumin, but in the presence of
6 per cent. salt an extensive precipitation occurred. 1 per cent. solutions of
phenol in water and in 6 per cent. salt had no visible effect on the albumin
suspension; a turbidity however was induced by the same phenol concen-
tration in 15 per cent. salt, and a marked precipitation in 22 per cent.
salt. solution.

The precipitating action of phenol upon proteins was thus greatly
increased by the presence of salt, but was not appreciably affected until high
salt concentrations were attained.

E. A. COOPER 179

A comparison of the results of experiments with gelatin with those
obtained by Reichel [1909] indicates that the germicidal power, the solubility
in proteins, and the precipitating action of phenol are increased to very
similar extents by the presence of salt.

The increased germicidal power thus appears to be due to an increased
de-emulsifying action of the phenol upon the bacterial proteins resulting
from the availability within the emulsoid particles of a larger amount of

phenol.

B. ALCOHOL.

Kromg and Paul [1897] showed that the germicidal action of phenol
upon spores was decreased by the presence of alcohol and that solutions of
phenol in 98 per cent. alcohol possessed no bactericidal power.

The author [1912, 2] found that the uptake of phenol by a protein was
greatly decreased by the presence of alcohol, the protein absorbing for
example eight times as much phenol from aqueous solution as from a solution
of phenol in 50 per cent. alcohol. These experiments indicated how in the
presence of alcohol less phenol became available for disinfection, but they
afforded no explanation of the nature of its germicidal action. It was
therefore of some importance to investigate the influence of alcohol upon the
protein-precipitating action of phenol.

Gelatin in the form of strips was chiefly employed in the experiments,
because it was not precipitated by immersion in alcohol of any strength.
Although this protein was precipitated by 24 per cent. aqueous phenol
solutions, there was no indication of precipitation by 5 per cent. phenol
solutions in 50 per cent. alcohol, and the extent of the precipitation by
6°25 per cent. phenol solutions in 25 per cent. alcohol was much less than
that induced by the above aqueous solutions. Furthermore, 10 per cent.
solutions of phenol in absolute alcohol had no visible effect upon gelatin.
This is of interest when correlated with Kronig and Paul’s observation
that solutions of phenol in 98 per cent. alcohol possessed no germicidal
power.

A sample of dialysed egg-albumin was immediately precipitated by
15 per cent. aqueous phenol solutions, but was not visibly affected by the
same concentration in the presence of 10 per cent. alcohol.

Although 0°5 per cent. aqueous solutions of picric acid precipitated
gelatin, similar concentrations in 30 per cent. alcohol had no visible effect
upon this protein.

180 E. A. COOPER

The inhibiting etfect of alcohol upon the germicidal power of phenol thus
appears to be due to a decrease in the de-emulsifying action of the latter
upon the bacterial proteins, resulting from a diminution in the amount of
phenol available within the emulsoid particles.

C=. .Ran

Strips of gelatin were immersed in castor-oil, containing varying amounts
of phenol. It was found that solutions of phenol in the fat as strong as
10 per cent. (by weight) had no visible precipitating action upon gelatin,
although this protein was precipitated by 24 per-cent. aqueous solutions
of phenol.

This can be correlated with the fact discovered by Koch that solutions of
carbolic acid in oil possess no germicidal power. The greatly diminished
protein-precipitating action of phenol in the presence of fat is probably due,
as in the case of alcohol, to a decreased uptake of phenol by the protein,
because, while phenol is about three times as soluble in gelatin as in water,
the partition-coefficient for phenol between fat and water is six.

D;" “ALKALI.

Kronig and Paul [1897] showed that, while 4 per cent. solutions of
phenol in water exerted a strong germicidal action upon anthrax spores,
solutions of sodium phenate equivalent to 4 per cent. phenol possessed no
germicidal power.

Morgan and Cooper [1912] showed that p-nitrophenol was about five
times as powerful a germicide as its potassium salt. The inhibiting action of
alkali upon the agglutination of proteins by heat suggested that the decreased
germicidal power might be due to the protective effect upon the bacterial
protoplasm of the alkali, produced in the hydrolysis of the alkali phenates,
against the precipitating action of the phenols simultaneously liberated.
Accordingly the effect of alkali on the protein-precipitating action of the
phenols has been studied.

Dialysed serum-albumin: Phenol. The albumin was not visibly affected
by 0°9 per cent.. solutions of phenol, but was precipitated by 1:7 per cent.
phenol. The precipitation was not only completely mhibited by N/50 to
N/150 concentrations of caustic soda, but even by a concentration as low

as N/300.

aw

K. A. COOPER 181

Gelatin: Phenol. Although gelatin was precipitated by 2°5 per cent.
aqueous phenol solutions, this protein was not visibly affected by 4 per cent.
phenol in the presence of N/25 caustic soda and was only very slightly
precipitated by the same concentration in presence of N/50 caustic soda.

Traces of organic bases, such as ethylamine and ethylenediamine, also
inhibited the precipitation of gelatin and egg-albumin by phenol.

Gelatin: p-Nitrophenol. Gelatin was precipitated by 1 per cent. aqueous
solutions of p-nitrophenol, but was not visibly affected by the same con-
centration in the presence of N/100 caustic soda, nor even by 1'4 per cent.
(N/10) solutions containing N/10 caustic soda.

Not only could alkali inhibit the precipitation of proteins by the phenols,
but when present in concentrations varying from N/1 to N/10 it could
disperse egg-albumin already precipitated by phenol.

The results indicate that very low concentrations of alkali are sufficient
to inhibit the precipitation of proteins by phenol. As alkali combines with
the phenols forming salts, which are partially hydrolysed in solution, the
inhibiting concentrations must have been even smaller than those stated
above.

In view of these results it was of great interest to investigate the
influence of alkali upon the germicidal action of phenol in greater detail, and
in particular to study the effects of very low alkali concentrations. The
experiments consisted in the estimation of the minimum concentrations of
phenol required to disinfect in a specified time under standard conditions in
the presence of varying amounts of alkali.

The following results were obtained :

TABLE III.
Organism, B. typhosus.

Concentration of Minimum concentration of phenol
caustic soda present required to disinfect in 15 mins. at 20°
0 9°5 in 1000 (N/10)

N/250 S 5
N/50 ” ”
N/25 11 in 1000 (N/8°5)
N/20 ” ”
N/15 13:5 in 1000 (N/7)

The germicidal powers of the above concentrations of alkali in the absence
of phenol were next determined and are compared in the following table
with those of solutions containing initially the same alkali concentrations
with varying amounts of phenol.

Bioch, vu 12

182 E. A. COOPER
TABLE IV.
Concentration Time required for
of NaOH complete disinfection
N/500 NaOH > 53 hours
N/250_ sy, SHO a.
N/100_,, 45 minutes
IN/50" 5 53 20 a
N/25 net < 3 FA
N/15 5 < 3 x
IN| OD a <= 3 3s
N/25 », +N/10 phenol >15 6
N/25 PN Sia) 855 15 *
NeOnh ee LENO. 4 1s ee
N/20 ,, +N/8°S .,, 15 rr
INS EAST/N >15 aN
IN| Toe eeetING | S20) 8 is >15 43
N/15 NF 3 15°
N/LO 2s EP N/12 oe: aly o
NOR SS + N/LO" .; 15 i
N/TORS Ee ND s >15 a
N/1O) °° 55 -+EN/7 ne >15 3

The above results indicate that alkali decreased the germicidal power
of phenol when present in moderately low concentration. While, however,
the precipitation of serum-albumin by phenol was inhibited by an initial
concentratior of N/300 caustic soda and that of gelatin by N/50 caustic soda,
bactericidal power was not measurably affected until the initial alkali concen-
tration was raised to N/25. ‘The precipitating action of phenol was therefore
much more sensitive to the inhibiting effect of alkali than its germicidal action.
This discrepancy is not sufficiently explained by the slight acidity of the
broth culture, nor by the toxic action of the alkali, since an analysis of the
results tabulated above indicates that not only did small concentrations of
alkali such as N/250 exert a very feeble germicidal action, but furthermore
the bactericidal action of the alkali was greatly decreased by the presence of
phenol. Owing to this depreciating effect N/10 solutions of alkali, containing
initially N/7 or N/9 concentrations of phenol, possessed a smaller germicidal
power than those which initially contained lower phenol concentrations such
as N/10 and N/12.

It is possible that a larger amount of alkali is required to preserve the
colloidal suspension from the action of the phenol in the case of the bacterial
proteins than in the cases of gelatin and albumin.

.
;
|
|

K. A. COOPER 183

EK. AcIDs.

Hailer [1910] showed that acids, arranged according to their germicidal
power and also according to their efticiencies in increasing the bactericidal
power of phenols, were not in the same order. Thus sulphuric acid exceeded
oxalic acid in bactericidal power, but was less effective in increasing the
germicidal power of phenol. Some of the experimental results described in
the previous pages suggested that the discrepancy might be due to differences
in the effects of the acids upon the protein-precipitating action of phenol.
This, however, seems not to be the case because it was found that, although
acids greatly increased the precipitating action of phenol, sulphuric acid was
effective in much lower concentration than oxalic acid.

The effects of alcohol and salt upon the precipitating action were shown
to be due to a redistribution of phenol between the water and proteins. The
greatly increased precipitating action in the presence of acids however is not
sufficiently explained in this way, as it was previously shown [Cooper, 1912, 2]
that the addition of acid only very slightly increased the solubility of phenol
in protein. It is probably due to the formation of an acid salt of the protein,
which is more readily precipitated than the protein itself.

DISCUSSION OF RESULTS.

The parallelisms deduced from the study of the effects of changes in the
chemical constitution of phenol and of the presence of certain substances
(e.g. salt, alcohol, fat) upon its germicidal and protein-precipitating powers
support the conclusion arrived at in the previous communication, namely,
that the germicidal action of phenol is associated with an alteration in the
condition of the colloidal suspensions of the bacterial proteins as evidenced
by the precipitation of such proteins as gelatin and albumin when certain
phenol concentrations are attained.

It has already been pointed out that the heat coagulation of proteins
consists of two stages.

1. A reaction between the proteins and hot water (denaturation).

2. A separation of the protein in a particulate form (agglutination).

It is therefore of interest to attempt to characterise the essential process
in disinfection by the phenols as either a denaturation of the bacterial
proteins or a de-emulsification of their colloidal solution. Chick and
Martin [1911; 1912] have shown that the process of denaturation by heat

12—2

184 EK. A. COOPER

is accelerated by acid and alkali, and retarded by salt, while the subsequent
process of agglutination is facilitated by low concentrations of acid and salt,
but inhibited by high concentrations and also by alkali.

It was shown in Section II of this paper that the bactericidal power of
phenol, its protein-precipitating action and solubility in proteins were
increased to equal extents by the addition of salt. The latter thus appears
to have no direct effect upon the germicidal and precipitating actions of
phenol, but to increase them merely by raising the phenol concentration
within the emulsoid protein-particles. It is therefore not possible by a study
of the effects of salt to correlate these properties of phenol with either the
denaturating or agglutinating action of heat upon proteins.

The fact, however, that the germicidal action of phenol is inhibited by
low concentrations of alkali, although to a less extent than its precipitating
action upon proteins, suggests that the essential process in disinfection by
phenol resembles the final phase in the heat-coagulation of proteins and
consists in the destruction of the colloidal suspensions of the bacterial
proteins.

Since the process of disinfection by hot water was found [Chick, 1910] to
be accelerated by alkali as well as by acid, it is possible that the essential
stage in this process is the denaturation of the bacterial proteins by the
action of the water and not the subsequent precipitation. A study of the
effects of salt upon the velocity of disinfection by hot water will throw light
on this problem.

SUMMARY.

1. The germicidal and protein-precipitating powers of phenol are
similarly affected by the entrance of various chemical groups into its
molecule.

(a) The introduction of hydroxyl groups decreases and that of nitro-
groups and of a methyl group increases the bactericidal and protein-pre-
cipitating powers of phenol.

(b) The monohydric phenols are superior to the alcohols both as
germicides and protein-precipitants.

2. Sodium chloride increases the germicidal and protein-precipitating
action of phenol through increasing its solubility in proteins.

3. Alcohol, on the other hand, decreases the germicidal and _ protein-
precipitating action of phenol through diminishing its solubility in proteins.
Solutions of phenol in absolute alcohol exert no germicidal action upon
spores and have no visible precipitating action on gelatin.

a

E. A. COOPER 185

4. Solutions of phenol in fat also possess no germicidal power and do
not visibly precipitate gelatin.

5. While the presence of very small amounts of alkali is sufficient to
inhibit the precipitating action of phenol upon proteins, only moderately low
alkali concentrations can effect a measurable decrease in its germicidal power.
The explanation of this discrepancy is not apparent.

6. The precipitating action of phenol is increased by the addition
of acids.

MAIN CONCLUSION.

From the evidence set forth in the previous communication and in the
present paper it is concluded that the absorption of phenols by bacteria is
merely the initial stage in the process of disinfection, and that the germicidal
action which follows is due, not to a typical chemical union of the phenols
with the bacterial protoplasm, as appears to be the case with formaldehyde,
but to a de-emulsifying action upon the colloidal suspension of some
constituent protein or proteins essential for the vitality of the organisms.

I desire to express my best thanks to Prof. Martin, F.R.S., for much
helpful criticism of this work.

REFERENCES.

Chick (1910), J. Hyg. 10, 237.
Chick and Martin (1908), J. Hyg. 8, 654.

—. (1911), J. Physiol. 43, 1.

— (1912), J. Physiol. 45, 61, 261.
Cooper (1912, 1), Brit. Med. J. 1. 1234, 1293, 1359.
(1912, 2), Biochem. J. 6, 362.

Hailer (1910), Arbeit, Kais. Gesund. 33, 500.

Hardy (1899), J. Physiol. 24, 158.

Kronig and Paul (1897), Zeitsch. Hyg. 25, 1.

Morgan and Cooper (1912), 8th Intern. Congr. App. Chem. 19, 243.
Reichel (1909), Biochem. Zeitsch, 22, 149.

XVII. ON THE RELATIONS OF THE EFHENOES
AND THEIR DERIVATIVES TO PROTEINS. A
CONTRIBUTION TO OUR KNOWLEDGE OF
THE MECHANISM OF DISINFECTION.

PART III. THE CHEMICAL ACTION OF QUINONE
UPON PROTEINS.

By EVELYN ASHLEY COOPER, Beit Memorial Fellow.
Taster Institute of Preventive Medicine.

(Received January 15th, 1913.)

Blyth and Goodban [1907] found that when pure cresylic acid was exposed
to light and air until it had become brown its germicidal power was
measurably increased. This was probably due to the formation of derivatives
of quinone through the aerial oxidation of the cresols.

Thalhimer and Palmer [1911] however could not detect any increase in
germicidal power when phenol was exposed to light and air for some time.

Morgan and Cooper [1912] showed that some of the aromatic amines
possessed a higher germicidal power when coloured by long standing than
when purified by redistillation. It is probable that the increased germicidal
power was due to the production of coloured quinone derivatives, which are
known to be formed in the oxidation of aromatic amines.

Thalhimer and Palmer [1911] showed that quinone itself was a very
efficient disinfectant and was superior to phenol, cresol, quinol, phenoquinone,
and formalin in germicidal power.

The author [1912, 1] confirmed some of the latter results and showed
further that quinone was more efficient as a germicide than the aliphatic
ketone, acetone.

In previous communications [Cooper, 1912, 2; 1913] evidence was set
forth which strongly suggested that the germicidal action of the phenols was
due, not to a typical chemical union with the bacterial proteins, but to a
de-emulsifying effect upon their colloidal suspensions. In view of the high
bactericidal power of quinone it was of great interest to investigate the

KE. A, COOPER 187

relations of this ketone to proteins, in order to compare the mechanism of its
germicidal action with that of the phenols and to arrive at a conclusion as
to the possibility of a relationship between its germicidal efficiency and
chemical reactivity.

Wiirster [1889] showed that when quinone was added to warm solutions
of various amino-acids a red coloration was developed.

Raciborski [1907] showed that quinone gave a red coloration not only
with amino-acids, but also with peptone and several proteins (egg-albumin,
serum-albumin, fibrin, globulin, legumin, and nuclein). Toluquinone reacted
similarly to quinone, and xyloquinone also gave a colour-reaction with
proteins and peptone, but not with glycine and alanine. Phenanthraquinone
and anthraquinone, on the other hand, gave no colour reactions with proteins.
The authors put forward no explanation of the above phenomena.

THE EXPERIMENTAL RESULTs.

The investigation described in this paper is divided into three parts.

Part I deals with the relations of quinone to various proteins.

Part II deals with the nature of the chemical action, which quinone was
found in Part I to exert upon proteins.

Part III deals with the relation of the chemical reactivity of quinone to
its germicidal power.

I. The relations of quinone to various proteins.
Gelatin.

When strips of gelatin were immersed in aqueous solutions of quinone
(O°l per cent.) the protein rapidly developed an intense red colour, but
retained its transparency. The reaction was irreversible, as the colour was
not removed by prolonged boiling with water or absolute alcohol. The red
colour was changed to green by the addition of alkali, but was restored by
acidification. The altered gelatin was, furthermore, completely insoluble in
hot water, and after immersion for 16 hours in quinone solutions the protein
was no longer rendered opaque (precipitated) by phenol. Immersion for
30 minutes did not inhibit the precipitation.

Gelatin after immersion in 40 per cent. formalin for 12 hours was also
insoluble in hot water and was no longer precipitated by 5 per cent. phenol,
although the original gelatin was visibly affected by 2°5 per cent. phenol
solutions.

These facts suggest that the quinone was not merely dissolved by the

188 E. A. COOPER

gelatin, but had reacted chemically with the protein. The colour-reaction
did not oceur when the gelatin was immersed in solutions of quinone in
absolute alcohol, and it only took place to a very small extent in 20 per cent.
alcohol. Solutions of quinone in toluene also gave no coloration with gelatin.

Similarly, although Witte’s peptone was coloured intensely red when
suspended or dissolved in aqueous quinone solutions, no coloration was
observed when it was suspended in an alcoholic solution of this substance.

The interpretation of these facts is probably that the quinone is dissolved
in the gelatin and proteoses before the chemical reaction, so that an efficient
external solvent for quinone (such as alcohol) by decreasing the uptake of this
substance by the colloids can inhibit the colour-reaction. (See the effect of
alcohol upon the distribution of phenol between water and proteins, Cooper
[1912, 2].)

Alizarin, like anthraquinone, gave no coloration with proteins. When
however gelatin was immersed in strong aqueous solutions of sodium alizarin-
sulphonate (Alizarin red, C,,H;O{OH),SO;Na) it assumed an intense red
colour, but, unlike gelatin treated with quinone, it was still readily soluble in
hot water and precipitated by 5 per cent. phenol, and the colour was rapidly
removed by washing with cold water.

Alizarin and its sulphonic derivative thus differed from quinone in not
reacting with proteins, and this inactivity may explain the fact that a
saturated solution (0°2 per cent.) of alizarin and a 1 per cent. solution of
alizarin-red exerted no measurable germicidal action upon Staphylococcus

py. aur.

Caseinogen.

(Merck’s Casein—prepared according to Hammarsten.)

When immersed in aqueous solutions of quinone this protein assumed
a purple colour, which was not removed by washing with water or boiling
alcohol. Unlike the original caseinogen the coloured product (after being
washed with water only) was insoluble in N/5 soda. Contact with the
alkali, however, turned it green, but the purple colour was restored by
acidification. The altered caseinogen was very slowly soluble in hot con-
centrated hydrochloric acid yielding a brown solution, and thus again
differed from the original protein, which quickly dissolved in this acid
forming a violet solution.

Egg-albumin.
As stated by Raciborski [1907] dialysed egg-albumin when mixed with

aqueous solutions of quinone soon gave an intense red coloration. The

EK. A. COOPER 189

protein was still coagulated by heat and alcohol, red coagula being formed,
which were not decolorised by prolonged washing with water and alcohol.
The protein was also precipitated from the red solution by saturated ammonium
sulphate. The precipitate was a red flocculent substance, readily soluble in
water and becoming dark brown on standing. After this colour change the
protein was frequently found to have become almost insoluble in water.
The coagula obtained from the original red solution by means of heat,
alcohol, and phenol remained permanently red.

Horse-Serum.

When aqueous solutions of quinone were added to horse-serum a red
coloration rapidly developed. ‘The serum-proteins were still coagulated by
heat and precipitated by alcohol at first reversibly and finally irreversibly.
The coagula were red and could not be decolorised by prolonged washing
with alcohol or water.

By half-saturation of the red solution with ammonium sulphate a red
flocculent precipitate was obtained, which, like the egg-albumin precipitated
by ammonium sulphate from solutions containing quinone, became dark
brown on standing. The precipitate with magnesium sulphate and the
alcohol coagulum, on the other hand, resembled the egg-albumin coagula in
remaining permanently red, but they soon became brown when immersed in
a solution containing ammonium sulphate and quinone. The cause of these
colour-changes could not be discovered.

By acidifying with acetic acid the filtrate from the precipitation with
half-saturated ammonium sulphate of the serum containing quinone another
red precipitate was obtained, corresponding to the albumin fraction from
normal serum. ‘The precipitate quickly redissolved on the addition of
water.

It was not found possible to crystallize the quinone derivates of serum
and egg-albumin by the application of the usual methods.

Witte’s Peptone.

The observation of Raciborski [1907] that solutions of Witte’s peptone
gave a red coloration with quinone was confirmed. Experiments were next
carried out with a view to the isolation of the coloured products.

Witte’s peptone has been separated into five constituents by the method
of fractional precipitation with alcohol and salt described by Haslam. The
following fractions of proteoses have been isolated in this way :

1. Insoluble in 50 per cent. aleohol and water. Crude hetero-proteose.

2. Insoluble in 50 per cent. aleohol and soluble in water.

190 EK. A. COOPER

- (1) a-proto-proteoses—precipitated by half-saturated (NH,).SO,.
(2) a-deutero-proteoses—precipitated by saturated (N H,),SO,.

3. Soluble in 50 per cent. alcohol and soluble in water.

(1) -proto-proteoses—precipitated by half-saturated (NH,),SO,.
(2) -deutero-proteoses—precipitated by saturated (NH,).SO,.

By similarly fractionating the coloured liquid obtained by mixing solutions
of quinone and Witte’s peptone together it was possible to isolate correspond-
ing precipitates all of which were highly coloured. Some of these products,
however, did not differ merely in colour from the fractionated proteoses.
Thus, while the alcoholic precipitate (a-proteoses and hetero-proteoses)
obtained from an aqueous solution of Witte’s peptone was readily soluble in
warm anhydrous m-cresol, the corresponding fraction obtained from the
aqueous solution after treatment with quinone was only slightly soluble.

Again, the a-proteoses after reacting with quinone were no longer
precipitated by formaldehyde, although they were still precipitated by
alcohol, mercuric chloride, and phosphotungstic acid. The significance of this
difference is discussed later. The fractions obtained from Witte’s peptone
after treatment with quinone corresponding to the a- and B- proteoses gave
the characteristic test for proteoses, namely a precipitate with nitric acid,
soluble on warming and reappearing on cooling.

The main conclusion to be drawn from these results is that proteins
isolated after treatment with quinone are permanently coloured and frequently
changed in solubility and precipitability, and thus appear to be chemically
altered. In the following pages the nature of this chemical change and the
possibility of its relationship to the toxic action of quinone upon bacteria are
discussed.

II. The nature of the chenical action of quinone upon proteins.

(i) The colour reactions given by quinone with simple amines.

A large number of amines readily gave colorations with quinone in
aqueous solution at ordinary temperatures.

(ii) The colour-reactions with imino-compounds.

In cold aqueous solution quinone gave red colorations with methylaniline
and di-amylamine, but not even on warming with succinimide, acetanilide,
and uric acid. Under certain conditions quinone could thus react with
substances containing the =NH group.

(ii) The inhibitory effect of formaldehyde upon the colour-reactions.

The fact that quinone gives colorations with amino- and imino- compounds

K. A. COOPER 191

suggests that the colour-reaction with proteins is due to a condensation of
the ketonic groups of the quinone with their — NH, and = NH groups.
Formaldehyde is known to react with amino- and imino- compounds by
condensation with the — NH, and =NH groups, forming methylene de-
rivatives.
NH, -CH,-: COOH + H-CHO = CH, -CH,- COOH + H,0

Bases . CH, -COOH

2NH(CH,)- CH, - COOH + H- CHO = H.C +H,0

\\N(CH,) - CH, -COOH

The colour reactions given by quinone with proteins, if due entirely to a
similar condensation, should therefore be inhibited by the previous formali-
sation of these substances.

Substance Nature of reaction
Ammonia ~ Brown coloration.
Ammonium peiehate “ce Purple a
Ethylamine me uta Violet oe
Di-amylamine ce re Rose red__,,
Guanidine (carbonate) ele Intense green coloration. (Hot solutions brown.)
Glucosamine sh ae Red coloration.
Amino-antipyrine ... sig Purple +
Tryptophane = See Red 7;
Atoxyl ... “ee ce Red _
Ree cas Pee a Red —,,__ (followed by brown precipitate).
m-Tolylene-diamine oa Violet ,, 3 re =
B-Naphthylamine .. Brown ,,
Tetra-hydro-B- naphhylamine Brown oil.
Pyridine ... : sie Dark yellow coloration passing to red.
Succinanilamine ... se Red coloration (followed by red precipitate).

It was actually found that in the case of egg-albumin, serum-proteins
proteoses, glycocoll, glycyl-l-tyrosine, methylaniline, and di-amylamine the
colour reactions with quinone were entirely inhibited by adding 40 per cent.
formalin to the compounds either previously to or simultaneously with the
addition of the quinone. Smaller amounts of formaldehyde (one to ten per
cent.) added with the quinone were found to inhibit the colour-reaction
partially.

Conversely, the a-proteoses present in Witte’s peptone after treatment
with quinone were no longer precipitated by formaldehyde, although they
were still precipitable by alcohol, phosphotungstic acid, and mercuric chloride.

In the case of gelatin, however, formalisation did not inhibit the colour-
reaction with quinone. This was also true in the case of ammonia, and here
the absence of any inhibiting effect could not be due to incomplete interaction

*

192 EK. A. COOPER

with. formaldehyde, because the product of this reaction—hexamethylene-
tetramine (urotropin), which was proved to be ammonia-free, readily gave
a red coloration with quinone.

The purified product from the interaction of formaldehyde and aniline also
gave a colour-reaction with quinone.

The inhibiting effect of formalin upon the colour-reaction given by
quinone with certain proteins and other amino- and imino- compounds, and
its inability to precipitate the compounds of quinone with a-proteoses,
although it readily precipitates the proteoses themselves, confirm the view
that it is the — NH, and = NH groups present in the proteins and their
hydrolytic products that react with the quinone. It is difficult to understand,
however, why the products of formalisation of gelatin, ammonia, and aniline
should behave exceptionally in yielding a coloration with quinone.

In order to attempt to understand the mechanism of the interaction of
quinone and proteins it is necessary here to set forth the course of the
reaction known to occur between an amine, such as aniline, and the above
ketone.

The products of the reaction depend upon the experimental conditions, as
indicated below.

1. As a result of the action of aniline upon an alcoholic solution of
quinone there are three products :

ile F 2. 3:
O O O N-C,H;
I I | I
Cc C C C
7
HC CH: CO,H;-HN-C C-NH-CsH; CoH;-HN-C C-NH-OgH; CoH;-HN-C C-NH-CoH;
ee ll | ll It al
HC CH HC CH HC CH HC SCE
Ky ae al by ait Hr
Cc C C C
| I | |
O O N - CoH; N- CoH;
Quinone. Di-anilido-quinone. Di-anilido-quinone-anil. Di-anilido-quinone-di-anil.

2. In the presence of acetic acid the chief product is di-anilido-quinone-
anil.

3. By fusing quinone with aniline and its hydrochloride the chief
product is di-anilido-quinone-di-anil.

There are thus two possible reactions between quinone and proteins and
their hydrolytic products :

1. The condensation of the ketonic oxygen atoms with the hydrogen of
the amino- and imino-groups, forming compounds of the anil-type (see
above 2, 3).

et eee

{. A. COOPER 193

2. The replacement of hydrogen attached to the benzene nucleus of
quinone by amino-acid residues through the amino- or imino-groups forming
compounds of the anilido-type (see above 1, 2, 3).

These two reactions might proceed simultaneously forming compounds
analogous to di-anilido-quinone-di-anil.

(iv) Euwperiments with quinone-dioxime.

Attempts were next made to discover to which type of chemical reaction
the red colorations given by quinone with proteins were due,

It was thought that experiments with quinone-dioxime could decide
this question, because, although substitution in the ketonic groups naturally
prevents their condensation with amino- and imino-compounds, it is known
not to inhibit the entrance of these substances through their nitrogen atoms
into other parts of the quinone nucleus.

Quinone-dioxime was prepared by the action of hydroxylamine hydro-
chloride upon quinone, the reaction being carried out in acid solution to
prevent the reduction of the quinone to quinol.

It was found that aqueous solutions of quinone-dioxime gave no colorations
with serum, gelatin, proteoses and alanine. The gelatin after immersion in
the quinone-dioxime solution was found to be still soluble in hot water and
precipitated by phenol as a white substance, and was thus not chemically
altered.

These results strongly suggest that the colour-reaction given by proteins
with quinone is due to the condensation of the — NH, or = NH groups with
the ketonic groups of the quinone, compounds similar to quinone-dianil being
produced. This conclusion is supported by the fact that no oxime could be
obtained by the treatment of the quinone-proteose compounds with hydroxyl-
amine hydrochloride.

The chemical action of quinone upon proteins thus resembles that of

formaldehyde.

(v) The relations of acetone to proteins.

Since it was found that quinone possessed a germicidal power more than
100 times as great as that of acetone [Cooper, 1912, 1] it was of interest to
compare the effects of these two ketones on proteins.

Acetone was found to differ from quinone in exerting a precipitating
action upon proteins, but while 071 per cent. solutions of quinone gave a
eolour-reaction with egg-albumin and gelatin, the albumin was not pre-
cipitated by aqueous solutions of acetone below 12 per cent. and gelatin was
not even affected by immersion in 50 per cent. and 90 per cent. solutions.

194 E. A. COOPER

The gelatin was still soluble in hot water and precipitable by phenol after
this treatment, so that there was no evidence that it was chemically altered
by the acetone. This ketone is therefore to be classed with the alcohols and
phenols as a protein-precipitant.

Ili. The relation of the chemical reactivity of quinone towards
proteins to its germicidal power.

The fact that as a germicide quinone is greatly superior to many other
para-di-substitution products of benzene is seen from the following table
[Cooper, 1912, 1 and Morgan and Cooper, 1912].

Substance Organism Carbolic coefficient
Quinone ... ak Staphylococcus py. aur. ; 10
Benzaldehyde as B. typhosus 10
Quinol ... ie - 1
p Cresol ... A. Staphylococcus py. aur. 2°4

33 3h B. typhosus 26n
Aniline... SoC Staphylococcus py. aur. 0°5

os S6c “0 B. typhosus 0:57
p-Toluidine eas e 1°25
p-Phenylene-diamine 49 Under 0:3
p-Nitrophenol ae Staphylococcus py. aur. 2°3

This itself suggests that the high germicidal power of quinone 1s
associated with its chemical reactivity.

This conclusion is supported by certain other facts.

1. Not only is quinone superior to phenol, p-cresol, quinol, p-nitrophenol
and acetone in germicidal power, but it can exert a chemical action upon
proteins in concentrations (e.g. 0'1 per cent.) much lower than those in which
the above substances can induce protein precipitation [Cooper, 1912, 2; 1913].

2. Benzaldehyde, which resembles quinone in its chemical action upon
proteins, is also approximately equal to quinone in germicidal power, its
carbolic coefficient with B. typhosus being 10.

There is thus some evidence that the mechanism of the germicidal action
of quinone consists in a chemical interaction with some constituent protein
or proteins of the bacteria essential for vitality, and not, as seems to be the
case with the phenols, in a precipitating effect upon the colloidal suspension.
The superiority of quinone as a germicide to various phenols and to acetone
is sufficiently explained by the fact that it reacts with proteins in con-
centrations much lower than those in which the phenols and acetone exert a
precipitating action.

K. A. COOPER 195

SUMMARY.

1. The observations of Wiirster and Raciborski that quinone solutions
gave a red coloration with various proteins and amino-acids have been
confirmed.

2. The proteins (egg-albumin, proteins of horse serum, gelatin, Witte’s
peptone) could be isolated in the coloured condition from the red solutions
by means of various precipitants and could not be decolorised by prolonged
washing with water or aleohol. Other physical properties of the proteins, e.g.
solubility, precipitability, were frequently changed as a result of the treatment
with quinone, as is also the case when proteins react with ‘formaldehyde.
From these results it appeared that the proteims had become chemically
altered by the quinone.

3. The colour-reaction did not occur when gelatin and proteoses were
immersed in solutions of quinone in absolute alcohol. It would appear that
the quinone was dissolved by the colloids before the chemical reaction, so
that an efficient solvent for this ketone such as alcohol, by decreasing the
uptake, could inhibit the colour-reaction.

4, The addition of sufficient formaldehyde to proteins, proteoses, amino-
acids, and imino-compounds either before or simultaneously with the addition
of the quinone completely inhibited the colour-reactions. Smaller amounts
of formalin decreased the intensity of the red colorations. Gelatin, aniline,
and ammonia however behaved exceptionally inasmuch as they still gave the
colorations with the quinone after the addition of formalin. The positive
results appeared not to be due to incomplete formalisation, since the isolated
compounds of aniline and ammonia with formalin gave colour-reactions with
quinone.

The inhibitory effect of formalin upon the colour-reactions given by
quinone with certain proteins, with proteoses and amino-acids indicates that
the latter react with quinone through their — NH, or = NH groups.

5. Proteins, proteoses, and alanine gave no colour-reaction with quinone-
dioxime, and no oxime could be prepared from the quinone-proteose
compounds. This is presumptive evidence that the constituent — NH, or
= NH groups of the proteins and their hydrolytic products condense with the
ketonic groups of the quinone. The chemical action of the latter upon
proteins thus resembles that of formaldehyde.

6. Acetone differed from quinone in acting as a protein-precipitant.

7. There is some evidence that the germicidal power of quinone is due

196 EK. A. COOPER

to its chemical action upon some constituent protein or proteins of the
bacterium essential for vitality and that the superiority of quinone as a
germicide to phenol, quinol, and acetone is explained by its reactivity
towards proteins in much lower concentration.

I desire to express my best thanks to Prof. Martin, F.R.S., for helpful

criticisms of this work.

REFERENCES.

Blyth and Goodban (1907), The Analyst, May.

Cooper (1912, 1), British Medical J. 1, 1234, 1293, 1359.

—— (1912, 2), Biochem. J. 6, 362. ;

—— (1913), Biochem. J. 7, 175.

Morgan and Cooper (1912), 8th Internat. Congr. Applied.Chem. 19, 243.
Raciborski (1907), Chem. Zentr. 1. 1595.

Thalhimer and Palmer (1911), J. Infect. Diseases, 9, 172.

Wiirster (1889), Chem. Zentr. 1. 392.

XVIII. THE RATE OF FERMENTATION BY
GROWING YEAST CELLS.

By ARTHUR SLATOR.
(Receiwwed January 16th, 1913.)

If a sugar solution which also contains the necessary food for yeast
growth is seeded with a small quantity of yeast, the yeast grows and the
sugar is fermented to alcohol and carbon dioxide. If the seeding is small
and all necessary food for yeast growth is in excess the growth during the
earlier stages of the reaction is unrestricted and follows the logarithmic law
of increase, that is the rate of increase is always proportional to the quantity
present.

At a later period retarding influences come into action, the yeast multi-
plication becomes restricted and during the final stages of the fermentation
ceases entirely.

In this communication an account is given of some measurements
made during the earlier part of the reaction where the yeast growth is
unrestricted.

If certain simple assumptions regarding fermentation are made, it can be
shown that during this period not only does yeast growth, but also the
fermentation caused by the yeast, follow the logarithmic law.

If the medium is seeded with WN cells per c.c. then the rate of growth at
any time ¢ is proportional to the number of cells present V+, where n is
the increase during the time ¢, that 1s,

w= K(N +n) ME Ms AOS Gi ars wed Ce ods (1).
where XK is the constant of growth.

On integration this becomes

1, N+n
K= Fln Ta I aa aca (2),
or eet eee) oo ees ae (2 a).

If F represents the number of grams of sugar fermented per unit of time
by each yeast cell and s the total amount fermented in time ¢, then s is
determined by the equation

f= [y+ n) F dt.
Bioch. v1 13

198 A. SLATOR

Substituting from equation 2a we have
s= [/ Nek dt

which on integration becomes

or from equation 2 a

If S is the amount of sugar fermented when the yeast grows from a very
small seeding to WV, then from equation 4

Se ree te oe eee ee (5)
Equation 3 therefore becomes
§ Siew SS ee (6)
or 2 tt a ee (6a).

It is evident from this equation that the time-fermentation curve during
this period is logarithmic and that the constant of the curve is the constant
of growth.

The validity of these equations has been tested by measuring the rate of
growth of a pure culture of a Burton yeast in lightly hopped wort of specific
gravity 1:040. It was found that K the constant of unrestricted growth of
this culture could be determined by methods of yeast counting and also by
measuring the rate of fermentation of the growing yeast cells.

Four different methods of estimating K have been worked out and
when tested with this growth of yeast were found to give almost identical
results.

In all these experiments the seeding was made with actively growing
yeast cells so that any initial retardation in the growth which would take
place if older yeast cells were used is eliminated.

METHOD 1.

The medium is seeded with a known number of yeast cells under sterile
conditions. Tubes containing the inoculated solution are kept slowly
agitated in a thermostat. At certain intervals of time the tubes are taken
out and the number of yeast cells counted. K is calculated according to
equation 2 which can be put in the form

: 1 N+n
0-434 K =" log = 5".

A. SLATOR

TABLE LI.
Temp. 20°.
t=time in hours,
N+n=number of yeast cells per c.c. at time ¢.

t . N+n 0434
0 90,100 —
17'°3 4,660,000 0-100
0 255,000 _
70 1,390,000 0°105
9°0 2,200,000 0°104
0 1,360 —
35 3,550,000 0-098

Average 0-102 T' =2°95 hrs.

199

In the above table 7 is the time for the: yeast to increase to twice the

original amount and is determined by the equation
pa 1082 _ 0-301

0-434 K 0-434 K°

METHOD 2.

Measurements of the rate of fermentation are carried out in the manner

described in previous publications. [Slator, 1906, 1908. ]

The sterile medium (75 ¢.c.) is seeded with a small seeding of yeast in a
tube of capacity about 100 c.c. The tube is connected with a manometer to
measure the rate of production of carbon dioxide. With a suitable seeding
appreciable fermentation starts in about 15 hours, the reaction proceeds
faster as the time goes on and measurements are continually taken on the

manometer at suitable intervals.
The readings on the scale can be used directly to test equation 6 a

Z wal! S+s
0-434 A = — log os

TABLE II.
Temp. = 20°.
t=time in hours from first reading S.

S+s=manometer reading in cm., at time t taking the infinitely early reading as zero.

t S+s 0:434 K

=15 (0°3)* =

0 9°7 (S) zs

0:5 10-9 0-101

0-83 11°85 0-105

1:25 12-9 0-099

1-58 13-9 0-099

1-92 15-05 0-099

Average =0°101

A second experiment gave 0°434 K=0°100. T =3-0 hrs,

* Nore. 15 hrs. before the first reading is too short a time to give exactly the infinitely
early reading and a small correction has to be made to obtain this value. As T is approxi-

mately 3 hrs. this correction is easily shown to be 1/32 of S, that is in this case 0°3 em.

13—2

200 A. SLATOR

METHOD 3.

The constant of growth can be estimated by means of equation 5
FN
K => 1S °

The yeast crop and the amount of fermentation are measured after a
convenient amount of fermentation has taken place, the initial seeding of
yeast. being small. F is determined by a separate experiment. The method
is not of much practical value as both V and F have to be determined whilst
t the time, a factor most easily measured, is eliminated.

With the culture of yeast -growing in wort at 20° it was found that 10°
yeast cells per c.c. caused a fermentation of 0°620 cm. per hour. In another
experiment with growing yeast cells it was found that when the yeast crop
was 5:2 x 10° cells per c.c. the fermentation was 14:4 cm. That is

F=0°620 x 10-*§ N=52x108, S=14°4.
It follows therefore that
0:434 kK = 0-434 FN/S = 0-097

a result in agreement with the previous experiments.

METHOD 4.

The fourth method involves measurements of the rate of fermentation
but has an advantage over method 2, as it allows measurements to be taken
in any part of the reaction even where the yeast growth is retarded and
nevertheless gives true values of A the constant of unrestricted growth.

The method is as follows: two experiments are made, the medium in each
case being seeded with small seedings in a known ratio. If the seedings are
small enough the time-fermentation curves are identical except that the one
reaction is a definite time behind the other. This time-difference is the
time for the one seeding to grow to the other, and from the two values K is
easily determined.

If & is the ratio of the two seedings and ¢, and ¢, the times at which the
two fermentations reach some definite stage in the reaction, then

0-434 K = = log R.

This method was tested in the following manner. Two flasks of wort
were seeded with actively growing yeast, one 319 times the amount of the
other. 75 c.c. of the wort with the greater seeding were placed in the tube
of the apparatus and the manometer connected. Appreciable fermentation

A. SLATOR 201

started in 15 hours and several readings were taken at definite times. The
first solution was then replaced by the second which in the meantime had
been resting in the thermostat. On the next day the times were taken
when the manometer showed the same changes in pressure as in the previous
experiment. The various differences of time were in good agreement and
the value of 0-434 4 calculated from the average came to 0°106.

TABLE III.

t,=time at which the first reaction gives the reading M on the manometer scale.
t,=time on the next day when the second reaction reaches the same point.
Temp.=20°, R=319.

M ty ty t,—t

9°4 9°20 a.m. 90 a.m. 23 hrs. 40 mins.

10°6 9°50 9°32 73 es ee: alee

11°55 10°10 9°52 ae) 53) 42" ,;

12°6 10°35 10°18 aa, 43,

13°6 10°55 10°38 Ree A

Average 23 hrs. 42 mins. =23°7 hrs.
0-434 K="8 59 0-106. T =2°85 hrs,

A comparison of the values of A determined by the different methods
shows good agreement.

TABLE IV.

Method 1 gives 0°434 K=0-102

» 2 4, 0:484 K=0-101
»> 8 4, 0°434 K=0-097
>» 4 yy 0°434 K=0°106

It is evident that the rate of growth of the yeast during this period is
regular and can be measured accurately by yeast counting or by fermentation.
It is of interest to note that the yeast crop at any time is composed mainly
of the last few generations of yeast and any dying off of the old yeast cells
would hardly affect the value of K. Further if the yeast crop were composed
of cells of different activity correct values of K would be obtained by
method 2 if the average value of the fermentative power remains constant
during the time of measurement, for equation 6a@ involves only ratios of
S and s.

The equations and experiments cover only about 2 per cent. of the
reaction corresponding to a yeast crop of 10 million cells per e.c., which is
about 8 per cent. of the final crop. Growth after this time is measurably
retarded, probably mainly by the carbon dioxide which is known to have a
considerable retarding influence on yeast growth.

202 A. SLATOR

An accurate knowledge of the yeast crop and the fermentation during
this early period is however given by these experiments and may be put in
the following form.

TABLE V.
Temp.= 20°.
F=1:2 x 10-4 grms, per sec.=4°3 x 107“! grms. per hour. 0°434 K=0°100.
Hrs. Cells per c.c. G, per c.c. fermented
— 1,000 Sx Ome
10 10,000 18x10
20 100,000 Tots) sg ke
30 1,000,000 18h Ome

40 10,000,000 Sete

These equations and methods of estimating the constant of growth can be
applied to other micro-organisms. Methods of counting may not always be
suitable, but method 4 is probably of general application. The rate of
growth of bacteria which produce acid, for instance, could be determined by
following the reaction by titration or by electrical conductivity and methods
for other growths could be devised without much difficulty.

One of the principal uses of K is to determine the time between infection
and the beginning of the chemical action brought about by the organism.
The factors which influence K are the factors which determine this time and
in extreme cases say whether a liquid is susceptible to the growth of the
micro-organism or not,

The value of A may not remain constant over the period of growth in
which one is specially interested but a row of constants is not the aim of the
investigation and a consistent though variable AH may lead to more important
results than a constant one.

Micro-organisms grow not only in liquids but also on solids and in the
liquid film covering solids and it is of importance to know whether these
equations and methods are applicable to determine rates of growth in such
cases.

Some information on the subject was obtained by measuring the rate of
growth of this culture of yeast in solid wort-gelatin (10 g. gelatin per 100 cc.
wort). The sterile solid medium was melted and seeded with about 100,000
yeast cells per c.c. 75 c.c. were poured into a bottle which was filled with
pieces of glass tubing, the whole being previously sterilised. There still
remained about 80 ¢.c. air space. The melted gelatin was cooled by rotating
the bottle under a jet of cold water so that the gelatin as it solidified became
distributed over the tubing.

The bottle was then connected to the manometer and the apparatus

A. SLATOR 203

exhausted. Measurements of the rate of fermentation were then taken and
the constant of growth calculated as in Table IT

TABLE VI.

Temp. = 15° t S+s 0°434 K
— 20 (0-1) —
0 2-05 (S) —
4 3°8 0-067
5 4-4 0-066
6 5°15 0-067
91 8°25 0-066
9-4 8°65 0-067
10°6 10°2 0-065
10°8 10-4 0-065

Average 0-066 T =4°6 hrs,

The rate of growth of a yeast colony developing in wort-gelatin from a
single cell therefore follows the logarithmic law. The experiments show that
retarding influences do not come into play up to the time the colony consists
of 200 yeast cells and probably the colony would grow regularly to a much
larger size. It is of interest to note that diffusion would play no controlling
part in determining the rate of fermentation until the colony consists of
several million yeast cells. Slator and Sand [1910] have made calculations
of the size of a yeast cell which would just ferment entirely the whole of the
sugar diffusing to it in a stationary hquid. At 30° the radius is calculated
to be 8x 10cm. The volume of such a yeast cell is 8 million times as
great as one of radius 4x 10-‘cm. Diffusion under these conditions would
not be a limiting factor in the fermentation by a yeast colony until the
colony consists of 8 million cells.

The rate of growth of this culture of yeast in wort gelatin is appreciably
higher than in wort itself, the rates being 0-066 and 0-050 respectively.

The investigation shows the possibility of measuring rates of growth when
the organism is growing on a solid medium.

REFERENCES.

Slator (1906), J. Chem. Soc. 89, 128.
—— (1908), J. Soc. Chem. Ind. 27, 653.
— and Sand (1910), J. Chem. Soc. 97, 927.

XIX.. THE IDENTITY OF TRIMETHYER ST
DINE (HISTIDINE-BETAINE) FROM VARIOUS
SOURCES.

By GEORGE BARGER anp ARTHUR JAMES EWINS.
From the Wellcome Physiological Research Laboratories, Herne Hill, SL.
(Recewed January 25th, 1913.)
In a recent paper we showed [Barger and Ewins, 1911] that ergothioneine,
a crystalline base containing sulphur, which was isolated from ergot by

Tanret [1909], almost certainly possessed the constitution denoted by the
formula 1:

I I
HN—CH NMe; yc, HN—CH NMe,
f 3
Seton ica’ Yo —> Sc. cH,-cH’ So
VA NGO va Soa
HS. C=N CO HO=N CO

Ergothioneine was thus trimethylhistidine (histidine-betaine) containing
a sulphur atom attached to a carbon atom of the glyoxaline ring. Further
we showed that on oxidation with ferric chloride the sulphur atom was
removed as in similar thiolglyoxaline derivatives [Pyman, 1911] and a new
base trimethylhistidine or histidine-betaine (II) was produced. We described
certain of its salts and pointed out the possibility of its natural occurrence.
Kutscher {1910] had previously obtained a crystalline aurichloride of a base
from commercial extract of mushrooms, and had stated that this base was
possibly trimethylhistidine, since it possessed the formula C,H,,O,N, and gave
a strong reaction with sodium p-diazobenzene sulphonate. As, however, no
details were published regarding this salt, we were unable to determine
whether our base was identical with Kutscher’s.

A short time after our paper there appeared a publication by Reuter
[1912] in which he described the isolation of histidine-betaine from Boletus
edulis, and gave a description of certain of its salts. Among others he
characterised two picrates; one, a monopicrate which was analysed, but to
which no melting point was assigned, and another, melting at 206° which
was not analysed, but which, from a picric acid determination, appeared to

G. BARGER AND A. J. EWINS 205

be a dipicrate. The picrate of histidine-betaine obtained by us from ergo-
thioniene melted at 123° and analysed well for the dipicrate, This difference
was pointed out by Reuter and for some time we were at a loss for a
satisfactory explanation. We now find that the apparent discrepancy was
due to the fact that our melting point was determined on the air-dried salt,
which contained two molecules of water of crystallisation: our analysis on the
other hand was made with an anhydrous specimen of which the melting point
was not at the time determined. Quite recently Kutscher [1912] succeeded
in synthesising the betaine in question from a-chloro-glyoxaline-propionic
acid and trimethylamine at 80°, and by means of the aurichloride established
the identity of the synthetic base with the supposed trimethylhistidine
previously isolated by him from mushrooms. On repeating Kutscher’s
synthesis we were able to isolate from the reaction product a small quantity
of the picrate of the betaine which melted at 123° as we had found in the
case of the base derived from ergothioneine. The ,two picrates were, indeed,
in all respects identical.

At this juncture we communicated with Dr C. Reuter who very kindly
supplied us with specimens of the two picrates prepared by him, together
with a full description of these salts, for which we offer him our best thanks.
We then found that the (anhydrous) dipicrate melting at 205°-206° described
by Reuter, crystallised from water with two molecules of water of crystal-
lisation, and, when air dry, melted at 123°-124°, and was then in all respects
identical with the picrate obtained by us. The water of crystallisation could
be removed by drying im vacuo over sulphuric acid, but only with some
difficulty. The picrate thus obtained always showed signs of sintering at
about 125° (doubtless owing to traces of water still adhering) but did not
melt until 205°-206° as described by Reuter. The crystalline anhydrous
salt could be readily obtained by recrystallising the picrate (dried im vacuo)
from absolute alcohol, and then melted at 213°-214°. In a private com-
munication Dr Reuter informed us that a purified specimen of his dipicrate
melted at 212°-213°. We carried out determinations of the water of
crystallisation present in our picrate (A) and that obtained by recrystallising
Reuter’s anhydrous dipicrate from water (B) with the following results.

A. 0°2061 (air dry) lost (in vacuo over H,SO,) 0:0106 H,O=5:14.
B. 0°1603 3 os sf Ps 0:0083 H,O=5:17.

C,H, ;0,N, -(C,H,0,N,),-2H,O requires H,O =5-21 per cent.
The complete identity of the two bases was established by the melting
points of the two forms of the dipicrate and of the monopicrate (m.p. 201°—
202°) which we prepared from our dipicrate according to Reuter’s direction

206 G. BARGER AND A. J. EWINS

(treating the aqueous solution of the dipicrate with one molecular proportion
of sodium hydrate). The melting points of mixtures in all cases showed
no depression.

There can be no doubt, therefore, that the trimethylhistidine (histidine-
betaine) obtained by us from ergothioneie is identical with that obtained
by Reuter from Boletus edulis and with the synthetic base obtained by
Kutscher. This result affords a further confirmation of the constitution
assigned by us, on other grounds, to ergothioneine.

For the isolation of the betaine of histidine we find that the preparation
of the dipicrate is the most convenient method; this salt dissolves in about
25 parts of boiling water and readily crystallises in thin elongated rectangular
plates with two molecules of water of crystallisation, and when air dry, melts
at 123°-124°.

In our previous paper [1911] the melting point of the aurichloride of the
betaine was given as 171°, As the quantity of material at our disposal was
at that time very small (a few centigrams only) we were unable to analyse
this salt, but now, with a further supply of material, we find that the gold
salt when pure melts at 184°, in agreement with the melting point as given
by Kutscher and by Reuter.

We also determined the rotation of the base recovered from Reuter’s
dipicrate with the following result.

Concentration of base (in aqueous solution as hydrochloride) = 0°39 per cent.

Actual rotation measured a)=+ 0°40° (mean of 6 readings) in 2°2 dm.
tube. Whence [a]p=+ 46°5°.

In conclusion we may point out that the trimethylhistidines obtained
from various sources are thus shown to be identical. This substance must be
classed with the other naturally occurring betaines from amino-acids, such as
those from glycine, proline, oxy-proline, and tryptophane (i.e. ordinary betaine,
stachydrine, betonicine and hypaphorine respectively).

REFERENCES.

Barger and Ewins (1911), J. Chem. Soc. 99, 2396.
Kutscher (1910), Zentr. Physiol. 24, 775.

(1912), Zentr. Physiol. 26, B69.

Pyman (1911), J. Chem. Soc. 99, 2172.

Reuter (1912), Zeitsch. physiol. Chem. 78, 167.
Tanret (1909), J. Pharm. Chim. [vi], 30, 145.

XX. A NOTE ON THE METABOLISM OF
NITROGENOUS SUGAR DERIVATIVES.

By JAMES ARTHUR HEWITT.

From the Physiological Laboratory, St Andrews University.

(Received February 11th, 1913.)

It is very generally accepted that carbohydrates are of fundamental
importance in the metabolism of nitrogenous foodstuffs. Loewi [1902], using
protein decomposition products which no longer gave the biuret reaction,
showed carbohydrate to be necessary for the maintenance of a nitrogen
equilibrium. Lesser [1904] subsequently demonstrated the inability of fats
to replace the carbohydrate. It has been observed further that glycosuria
following excision of the pancreas in dogs is attended by increased excretion
of nitrogen [Falta, Grote and Staehelin, 1910], and Taylor and Cathcart
[1910] found phloridzin diabetes to be followed by disturbances of nitrogen
metabolism, Finally, Falta and Gigon [1908] have concluded that carbo-
hydrate is indispensable for the utilisation of protein, and it has been
assumed that some compound is formed between the carbohydrate and the
nitrogenous material [Liithje, 1906].

In an attempt to determine the nature of this synthesis Spiro [1907]
showed that the injection of glycine and fructose together gave rise to the
appearance in the urine of a dicarboxylic acid normally absent. On the
other hand much im vitro work has been done and various definite
compounds of sugars and nitrogenous substances have been obtained. Lobry
de Bruyn [1904] prepared compounds of sugar and ammonia, Schoor!l [1900]
a urea derivative of glucose, and Morrell and Bellars [1907] have synthesised
a product from glucose and guanidine. It may be noted however that in some
cases at least the reaction does not seem to be one of simple condensation
[Irvine, 1909]. Little seems to have been done on the metabolism of
substances of this nature.

Fabian [1899] found that glucosamine is with difficulty oxidised in the
body and is in great part excreted unchanged. The extreme stability of this

208 J. A. HEWITT?

compound and its great resistance to im vivo oxidation are somewhat
unexpected; but in light of recent evidence brought forward by Irvine and
Hynd [1912] for considering glucosamine to be not a simple derivative of
glucose, but a ring compound containing a betaine grouping, Fabian’s results
are largely explained. |

For the purposes of studying the metabolism of compounds such as those
mentioned above, those referred to offer certain disadvantages; Lobry de
Bruyn’s products are somewhat indefinite, the urea derivative difficult to
obtain in quantity, and most other compounds of this nature are highly
unstable.

For the following experiments, glucose p-phenetidide, the condensation
product of glucose and p-phenetidine, was selected on account of the ease of
its preparation and its comparative stability to hydrolytic agents. At the
boiling point, however, in solution it is readily hydrolysed by 0:2 °/, hydro-
chloric acid.

The preparation of glucose p-phenetidide was carried out according to the
method devised by Irvine and Gilmour [1909].

The toxicity of the compound is also disputed [St Mostowski, 1909]
(quoted from the above-mentioned authors).

White rats were used, the faeces and urine being collected separately in
cages similar to those described by Schafer [1912]. To prevent loss of |
nitrogen from the urine in the form of ammonia, known volumes of con-
centrated sulphuric acid were generally added to the receiving vessel previous
to collection. The glucose p-phenetidide was administered either in the
food (bread and milk ad libitum), or less often by intra-peritoneal injection.
Quantities of from 0:1 to 1 g. (usually 0°5 g.) were given daily. The intake
and output of nitrogen were determined both before and during the experi-
ment, but no alteration in the nitrogen equilibrium was observed. The
presence of glucose p-phenetidide in the food caused no variation in the
amount consumed per diem which could not be accounted for by the normal
difference from day to day.

The urine of those animals receiving the compound was without exception
of a much darker colour than that of the controls or of themselves before or
after the feeding. Discontinuance of feeding at once restored the normal
light yellow colour. The colour is not caused by hydrolysis taking place in
the urine by means of the acid present, for when the urine was collected
alone the colouration persisted’.

' Oxidation of glucose p-phenetidide by means of hydrogen peroxide in neutral solution
produces a brownish black syrup of unknown composition.

J. A. HEWITT 209

Injection of 0:1 g. produced similar results, and moreover in each case a
substance displaying active reducing properties was excreted in the urine,
Greater reduction of Fehling’s solution, Barfoed’s solution and Nylander’s
reagent occurred after preyious boiling of the urine with mineral acid. In
one case following injection, a compound was obtained from the urine by
means of phenylhydrazine, which in microscopic appearance resembled the
phenylosazone of glucose, but owing to the small quantity confirmation could
not be obtained.

It was mentioned above that the toxicity of glucose p-phenetidide was
disputed, St Mostowski denying the general conclusion that it is poisonous
(we are unaware however of the quantities administered by this author).

In these experiments it is shown that in quantities up to 1 g. per diem it
produces no ill effects in rats of from 220 to 250 g. body weight. Experiments
with a view to determining the effects of larger doses were not performed
owing to scarcity of material.

Injection of p-phenetidine however is followed by marked results. The
p-phenetidine cannot be given by feeding owing to the animals refusing to
eat all food containing it.

In a cat 4.c.c. of an 8°/, solution caused a large and rapid fall of blood
pressure which returned to normal within a minute and a half and remained
steady and regular until death ensued ten minutes later. Post-mortem
examination revealed extensive intra-vascular clotting.

It seems probable that much of the glucose p-phenetidide escapes
hydrolysis or is rendered innocuous by conversion into some other substance,
from the facts that in doses of 4c.c. of an 8°/, solution or 032g. per
25 kilo body weight, p-phenetidine is highly toxic, and that glucose
p-phenetidide is harmless when given in amounts of | g. per 220 g. or 45 g.
per kilo body weight, equivalent, if complete hydrolysis took place, to about
2 g. p-phenetidine.

SUMMARY.

1. After administration of glucose p-phenetidide by mouth or by
injection, a reducing substance is excreted by the urine.

2. No effects on nitrogen metabolism are produced.

3. In amounts up to 1g. per 220 to 250g. body weight per diem
glucose p-phenetidide is not toxic.

4. Some of the glucose p-phenetidide escapes oxidation in the body or is
converted into a substance of non-poisonous nature.

210 J. A. HEWITT

5. In amounts of 0°32 g. per 2°5 kilo body weight (or 0°108 g. per kilo)
p-phenetidine is highly toxic.

The author would like to express his indebtedness to Dr Cathcart of
Glasgow University for suggesting the line of research, and also to Professor
Irvine and to Professor Herring of St Andrews University for kindly advice
and criticism.

REFERENCES.

Fabian (1899), Zeitsch. physiol. Chem. 27, 167.
Falta, Grote and Staehelin (1907), Beitr. chem. Physiol. Path. 10, 197.
and Gigon (1908), Biochem. Zeitsch. 13, 267.
Irvine (1909), Biochem. Zeitsch. 22, 357.

and Gilmour (1909), J. Chem. Soc. 95, 1150.
and Hynd (1912), J. Chem. Soc. 101, 1135.
Lesser (1904), Zeitsch. Biol. 45, 497.

Lobry de Bruyn (1904), Maandbl. Natuurw. 18, 85.
Loewi (1902), Arch. exp. Path. Pharm. 48, 803.
Liithje (1906), Arch. ges. Physiol. 113, 547.

Morrell and Bellars (1907), J. Chem. Soc. 91, 1010.
Mostowski, St (1909), Bull. Akad. Sci. Cracow, 641.
Schafer (1912), Quart. J. Exp. Physiol. 5, 204.
Schoorl (1900), Akad. Wetensch. Amsterdam, 410.
Spiro (1907), Beitr. chem. Physiol. Path. 10, 247.
Taylor and Cathcart (1910), J. Physiol. 41, 276.

XXI. AN ATTEMPT TO ESTIMATE THE
VITAMINE-FRACTION IN MILK.

By CASIMIR FUNK, Beit Memorial Research Fellow.

From the Biochemical Department, Lister. Institute.

(Received February 17th, 1913.)

The isolation of the beri-beri vitamine from milk [Funk, 1912, 1], together
with the important fact observed by Andrews [1912], that infantile beri-beri
occurs when the children are fed by mothers suffering from beri-beri, suggest
a new line of investigation into the etiology of infantile scurvy and rickets.
The problems which present themselves in connection with these investi-
gations have been already set forth [Funk, 1912, 2] and will be fully treated
in an account in the Ergebnisse der Physiologie, Vol. 13. These problems
are briefly the following. (1) What is the normal amount of vitamines
in milk of different species, including the human? (2) Is there a definite
relationship between the amount of vitamines secreted in the milk and that
ingested in the food? This point is of special interest because of the
periodical appearance of rickets in winter and disappearance in summer, a
period when the cows return to pasture. (3) What effect have boiling and
pasteurisation on the vitamine content of milk? All these points await
their solution until a method for determining such small quantities is
available. The present inquiry shows that the ordinary chemical methods
for estimating vitamines can hardly suffice and attention at present must
therefore be directed to colorimetric methods.

Nevertheless the results described below give an idea of the amount of
vitamines possibly present in milk. The whole vitamine fraction (nitrogen
precipitated by phosphotungstic acid in the alcoholic extract of the dried
milk) from 1 litre of milk amounts to 1-25 mgr. nitrogen. This would
correspond to 1-3 eg. vitamine (C,,H,,O,N,). After the elimination of the
vitamine-fraction the residual nitrogen amounts to 20-50 mgr. per litre milk.
This residual nitrogen represents in all probability allantoin; assuming this
to be correct, one litre of milk contains 0:06-0°15 grm. allantoin, a figure in
good agreement with that obtained by Ackroyd [1911] by means of a direct
method, namely 0°199 grm.

212 C. FUNK

The results further indicate that for a chemical investigation of the
vitamines of milk it will be necessary to start on a large amount of milk. At
the same time they show that the artificial protein-free milk, such as was
used by Osborne and Mendel [1912] in their experiments on growth, differs
from true protein-free milk in lacking these nitrogenous substances. It 1s
therefore quite conceivable, as the experiments of Hopkins and Neville [1913]
point out, that these substances play an important part in the process of
growth.

As the figures below show, the milk after removal of fat by centrifuging
has lost about 50°/, of vitamine and allantoin.

The milks used in these experiments were from London dairies. For
each experiment a litre was evaporated in vacuo at 30°, using a distilling
flask constructed of two parts from which therefore the residue could be
quantitatively removed. The process of drying the residue was hastened in
the last stages by an addition of alcohol. The residue was powdered in a
mortar and dried in a vacuum desiccator to constant weight. The fine
powder was then shaken with a definite amount of alcohol (400 cc.) in a
shaking machine for two hours. The extract was filtered and of the
filtrate an aliquot part was evaporated in vacuo. The residue was extracted
with water and the extract transferred to a measuring flask. The watery
extract was filtered and the filtrate found in every case to be entirely
free from proteins. A measured quantity was acidified slightly with
sulphuric acid and precipitated with 10°/, phosphotungstic acid. After
standing for 48 hours the precipitate was filtered and washed with dilute
sulphuric acid. In accordance with my previous results this phosphotungstic
acid precipitate can be regarded as the vitamine-fraction. Both in the
precipitate and filtrate the nitrogen was estimated by Kjeldahl, N/20 normal

solutions being used for titration.
Nitrogen of the vitamine _—_- Residual nitrogen

fraction in mgr. in mer.
Weight of (phosphotungstic (phosphotungstic
residue acid precipitate) acid filtrate)
1 janet milk 100 2°54 50°9
* (Control 99°5 2°46 49-6
9. { Uncentrif. 133 7 30°7
(Centrif. 98 ioe 16°8
3 { Uncentrif. 27 25 22°7
* (Centvrif. 90 1-4 12-1
yh { Uncentrif. 129 1°96 29°6
Centrif. 92 1e2 15°3
5 { Uncentrif. 132 2°4 49:2
* (Centrif. 96 1:6 27°3
6. Centrif. 89 1-06 22-4
7. Centrif. 105 1°83 44-9

C. FUNK 213

In this connection the behaviour of allantoin when boiled with Ruhemann’s
reagent (triketohydrindene hydrate) should be noted. According to Abder-
halden and Schmidt [1911] allantoin when boiled for one minute with
Ruhemann’s reagent does not give any colouration. But although this is so,
if the boiling be continued for 2-3 minutes a violet colouration appears and
this reaction might be used as a test for allantoin. Since the crude product
obtained by oxidation of uric acid by permanganate gave the reaction at
once, probably due to admixture with allanturic acid, and allantoin is known
to be destroyed by boiling, it was thought that a decomposition product of
allantoin, probably allanturic acid, was the cause of this reaction. A sample
of pure allantoin was therefore boiled with PbO, and another with water.
Both samples gave a strong allanturic acid reaction after a certain time of
boiling, the sample with PbO, giving however the stronger reaction. The
reaction seemed to disappear after prolonged boiling. The reaction may also
be obtained from hydantoin, |

It is interesting to note, that we have in allantoin an example of a
substance, which like vitamine is destroyed by boiling.

REFERENCES.

Abderhalden and Schmidt (1911), Zeitsch. physiol. Chem. 72, 87.

Ackroyd (1911), Biochem. J. 5, 400.

Andrews (1912), Philipp. J. Sc. 7, 67.

Funk, Casimir (1912, 1), J. Physiol. 45, 75.

(1912, 2), J. State Med. 20, 341.

Hopkins and Neville (1913), Biochem. J. 7, 97.

Osborne and Mendel (1912), Zeitsch. physiol. Chem. 80, 307; J. biol. Chem. 13, 243.

Bioch, vir 14

XXII THE ENZYMES OF WASHED ZYMIN AND
DRIED YEAST (LEBEDEW). I. CARBOXYLASE.

By ARTHUR HARDEN.

From the Biochemical Department, Lister Institute.

(Received February 17th, 1913.)

When zymin is thoroughly washed with water a residue is left, which no
longer has the power of fermenting glucose, but regains this power when the
washings are added to it, even when the latter have been boiled [Harden and
Young, 1910]. The dried yeast used by Lebedew behaves in a precisely
similar manner [Euler and Backstrém, 1912].

It is therefore a matter of considerable interest to ascertain how far this
treatment affects the various enzymes which are known to exist in zymin
and dried yeast. It can thus be ascertained whether these enzymes are
themselves soluble and whether if insoluble they require for their action the
presence of a soluble substance of the nature of a coenzyme.

It is also probable that some light may be thrown on the possible function
of some of these enzymes in the process of alcoholic fermentation.

Carboxylase.

Since the discovery by Neuberg and Hildesheimer [1911] of the un-
expected fact that yeast, yeast juice and zymin readily and rapidly decompose
pyruvic acid and other a-ketonic acids with evolution of carbon dioxide and
formation of an aldehyde, the opinion has been expressed in many quarters
that pyruvic acid may form a stage in the enzymatic decomposition of glucose
into alcohol and carbon dioxide [Neubauer and Fromherz, 1911; Neuberg
and Kerb, 1912; Kostytschew, 1912; Lebedew 1912].

On the other hand it is possible that carboxylase is quite independent of
the enzymes of alcoholic fermentation, its function being that of decomposing
the a-ketonic acids formed by the deaminisation of the a-amino-acids (see
Neubauer and Fromherz, 1911).

In order to examine the action of washed yeast preparations on pyruvic
acid, experiments were made not only on the free acid but also on the

it

> ew ee

A. HARDEN 215

sodium salt in presence of weak inorganic acids, by the aid of which the
acidity was diminished, whilst at the same time the whole or almost the
whole of the carbon dioxide produced was evolved.

The evolution of carbon dioxide was observed by the aid of the apparatus
previously described [Harden, Thompson and Young, 1910]. No quantitative
estimations of acetaldehyde were made, but its presence was proved by the
reaction with Schiffs reagent in all cases in which an evolution of carbon
dioxide was observed. Control experiments were at the same time carried
out with glucose and phosphate. It was thus found that the residue obtained
by washing zymin and dried yeast (Lebedew) until they could no longer
ferment glucose, was capable of decomposing pyruvic acid quite readily. The
interesting fact is thus ascertained that carboxylase does not require the
presence of a coenzyme removable by washing in order to exert its
characteristic reaction on pyruvic acid. This result does not however allow
any definite conclusion to be drawn as to the possible function of carboxylase
in alcoholic fermentation. It can only be concluded that if the decomposition
of pyruvic acid actually be a stage in the alcoholic fermentation of glucose,
the soluble coenzyme is required for some change precedent to this, so that
in its absence the production of pyruvic acid cannot be effected. The
following experiments illustrate the results obtained. The solutions through-
out were saturated with carbon dioxide at the temperature of the bath before
the commencement of incubation.

Exp. 1. A sample of Schroder’s dried yeast (nach Lebedew) was washed
3 times on the centrifuge with water and then brought into acetone and
dried. The following experiments were then made.

(a) 2 g. washed and acetoned yeast +10 ¢.c. 1 per cent. pyruvic acid
+40 cc. H,O.

(b) 2 g. washed and acetoned yeast +25 c.c. 1 per cent. pyruvic acid
+ 25 «ac. HO.

(c) 2 g. washed and acetoned yeast + 25 cc. 1 per cent. pyruvic acid
+5 cc. 03 molar Na,HPO, + 20 c.c. H,O.

(d) 2 g. washed and acetoned yeast +0 pyruvic acid +5 cc. 0°3 molar
Na,HPO, + 45 cc. H,O0 +2 g. glucose.

These were incubated at 25° and the evolution of CO, measured.

Time (a) () (c) (d)
1 hr. 50 mins. 3°5 16 18°6 19

The action of the washed yeast on glucose was extremely small (d@) and
very little fermentation of the free pyruvie acid (a and b) occurred. In
presence of sodium phosphate however a considerable evolution of CO,

216 A, HARDEN

occurred (18°6 cc.) and a strong reaction for acetaldehyde was given by the
filtrate. The small effect on free pyruvic acid is ascribed to the effect of the
acidity of the solution on the enzyme ; in presence of sodium phosphate, on
the other hand, the acidity is mainly due to NaH,PO, formed by interaction
of the pyruvic acid with the Na,HPO,.

Exp. 2. A similar experiment was made with similar results using the
same washed and acetoned yeast, but washing it again 3 times to remove
the last trace of coenzyme.

10 g. of acetoned yeast were washed and made up to 85 c.c.

(a) 25 cc. yeast suspension + 21 cc. 1 per cent. pyruvic acid +7 cc.
HO.
(b) 25 ce. yeast suspension + 21 cc. | per cent. pyruvic acid +7 cc.
0:3 molar Na,HPO, (approx. equivalent to the pyruvic acid).

(c) 25 cc. yeast suspension + 21 cc. H,O+7cc. 0°3 molar Na,HPO,
+2 g. glucose. ae ms 4 PB

55 mins. 4-1 22-0 1-2 c.c.

Here again the action on the free acid is small, whereas a considerable
evolution of CO, occurs in presence of phosphate. The amount produced
is really greater than that evolved since when pyruvic acid and sodium
phosphate are present in equivalent amount a portion of the CO, is retained
as NaHCO, in equilibrium with NaH,PQ,.

CH, - CO- COOH +Na,HPO, > CH,- CO - COONa + NaH,PO, >
CH, - CHO + NaHCO, + NaH, PO, = CH, - CHO +Na,HPO, + CO,+H,0.

Accordingly, the combined CO, was determined after incubation and, in
a separate sample, before incubation and it was thus found that 12°9 cc. of
CO, were retained and should be added to the amount observed in (b), making
a total production of 349 c.c.

Exp. 3. The effects of neutral potassium citrate and of excess of free
boric acid in presence of the sodium salt of pyruvic acid were tried. The
yeast employed was a similar preparation to that used in Exp. 2.

(a) 25 cc. yeast suspension +21 cc. 1 per cent. pyruvic acid +7 ce.
0:3 M. potassium citrate neutral to phenolphthalein.

(b) 25 cc. yeast suspension +21 ec. H,O+7 cc. 03 M. sodium
phosphate +2 g. glucose.

(c) 25 cc. yeast suspension + 21 c.c. HO + 7 cc. 0°3 MK. citrate.

(d) 4 A +21lcece. 1 per cent. pyruvic acid + 2-4c.c.
N. KHO+3 g. H;BO,.

Time (a) (b) (c) (d)
1 hr. 25 mins. 4°5 1153} 1:0 25°6

A. HARDEN 217

The sample of yeast had practically no action on glucose (b), comparatively
little on pyruvic acid in presence of citrate (a) and a large action in presence
of free boric acid (d). This result therefore is in agreement with the idea
that the action of the enzyme is greatly inhibited by acid, citric acid being
much stronger than boric acid.

Exp. 4. 10 g. dried yeast was washed 3 times and made to 100 c.c.

(a) 25 cc. yeast suspension + 25 cc. 1 per cent. pyruvic acid + 2°5 c.c,
N. KHO +3 g. boric acid + toluene.

(b) 25 cc. yeast suspension +21 cc. H,O+5 cc. 03 M. Na,HPO,
+ 2 g. glucose + toluene.

Time (a) (b)
1 hour 21-04 0:2 c.c.

Here again a good fermentation of pyruvic acid is produced by a sample
of washed yeast incapable of fermenting glucose, the total evolution from
which in 18 hours was only 0°4 c.c.

Exp. 5. 20 g. zymin (Schroder) were washed 3 times and made to
100 c.c.

(a) 25 cc. yeast suspension + 22°5 cc. HO + 2 g. glucose + 5 cc. 0°3 M.
Na,HPOQ,.

(b) is . +25 cc. 1 per cent. pyruvic acid + 2°5 ce.
N. KHO +3 ¢g. boric acid.
Time (a) (b)
43 mins. 178 0°8
18 hours 22-7 0:8

Washed zymin therefore has the same action as washed dried yeast.

SUMMARY.

Zymin and dried yeast (Lebedew) after being freed from coenzyme by
washing and thus rendered incapable of fermenting glucose, readily decompose
pyruvic acid into carbon dioxide and acetaldehyde, provided that the acidity
of the solution is kept low.

REFERENCES.

Euler and Backstrém (1912), Zeitsch. physiol. Chem. 77, 394.
Harden, Thompson and Young (1910), Biochem J. 5, 230.

and Young (1910), Zentr. Bakt. Par. 1. 26, 178.
Kostytschew (1912), Zeitsch. physiol. Chem. 79, 359.

Lebedew (1912), Ber. 45, 3256.

Neubauer and Fromherz (1911), Zeitsch. physiol. Chem. 70, 326.
Neuberg and Hildesheimer (1911), Biochem. Zeitsch. 31, 170.
— and Kerb (1912), Zeitsch. Giirungsphysiol, 1, 114.

2a. THE CRITICAL SOLUTION POINT
OF URINE.

By WILLIAM RINGROSE GELSTON ATKINS anp
THOMAS ARTHUR WALLACE.

From Trinity College, Dublin.

(Recewed January 26th, 1913.)

In order to determine the efficiency of a kidney as an excretory organ,
the most exact method is the determination of the total solid content of
the urine (neglecting colloids), as shown by the depression of its freezing-
point below that of pure water. The examination of urine and blood by this
method has proved very useful and reliable, but the time required for
carrying it out constitutes a serious drawback when only one or two
specimens are to be examined.

The following method was brought forward by one of us [Atkins, 1908]
in the hope that it might be of service in cases where great accuracy is not
essential and saving of time an advantage. Since no special apparatus 1s
required and ice is not employed the usefulness of critical solution point
determinations as a method of clinical examination is much increased
especially for work in warm climates.

The present paper gives an outline of the method of such determinations,
as the subject is a comparatively recent one. The data of the former paper
are extended and the influence of the various constituents of the urine upon
the critical solution point is traced out.

OUTLINE OF CRITICAL SOLUTION Point THEORY.

(a) Two pure substances. Consider two liquids not completely miscible
at room temperatures, such as phenol and water. If small quantities of these
two substances are shaken together a white emulsion is obtained; this on
standing separates into two layers, a lower one, consisting of phenol with
some water dissolved in it, and an upper one of water with some dissolved
phenol. If the temperature be now raised, the quantity of each dissolved in

Bioch. vir 15

220 W. R. G. ATKINS AND T. A. WALLACE

the other increases till a temperature is reached at which the two become
completely miscible; on cooling even a tenth of a degree separation again
occurs, a white emulsion is formed, and on standing the two layers separate
out as before. The temperature, at which the two layers become homo-
geneous, depends upon the proportions of water and phenol initially present,
and rises in the case of water as the amount of phenol is increased until
a mixture is reached which remains in the two phases at a higher temperature
than any other mixture of the two pure components.

The composition of this mixture is termed the “critical concentration,”
and the temperature at which the liquid becomes homogeneous is called the
“critical solution temperature.” This temperature is accordingly that at
which the two phases have the same composition and the same properties.
At the critical concentration and critical solution temperature the volumes of
the two phases, water in phenol and phenol in water, are equal.

At the critical temperature the two phases have the same surface tension,
and are therefore completely miscible; at a very slightly lower temperature
the surface tensions of the two phases are almost equal, consequently the
meniscus between them is flat. As the homogeneous mixture cools, at
a temperature very little above the critical temperature a beautiful opal-
escence appears, which is blue by reflected light and brown by transmitted.

If now a series of mixtures of water and phenol be taken and the
temperatures at which they become homogeneous be determined, and the
values thus obtained be plotted as ordinates, taking the corresponding
percentage compositions of the mixtures as abscissae, a curve is obtained, the
highest point of which is the critical solution point [Fig. 1, Schreinemakers,
1899]. This curve is somewhat flat at the top, hence the composition of the
mixture may be slightly altered to one side or the other of the critical
concentration without altermg the temperature of miscibility more than
about 01°C. The blue opalescence is still obtained, but it is not as
beautiful as at the critical solution point itself.

(b) Hffect of a third substance. The addition of a very small quantity of
a third substance to any mixture of water and phenol raises the temperature
at which the two phases become homogeneous, provided the added substance
be freely soluble in water and almost or quite insoluble in phenol, or vice
versa. If the added substance be freely soluble in both, then the temperature
at which a homogeneous mixture is obtained is lowered. Thus by adding
a small quantity of sodium oleate to a mixture of phenol and water, it is
possible to obtain a very strong solution of phenol, homogeneous at room
temperatures. Lysol, the disinfectant, is prepared in a similar way.

/

W. R. G. ATKINS AND T. A. WALLACE 221

t Maximal
Temperature t
140 =
6°47 / NaCl
nd °
Critical 130
Temperature
120°
4:91 [NaCl
(9)
110
9
100
(e)
90
» ang
Se = —% 7 dere
SO
80° é Hef
ee
YY }) 70
60-
L 50°
O
v
(oe) ~
40 FS
wv
c
= fo}
2 30
Lf
c
° Pet
20
<
10°
‘ 0-99 / NaCl.
0 = x
Oo
ozwater 207, 4o/, 60/ so/, 1007

100 /. phenol

Fig. 1.

In part after Schreinemakers. The dotted line shows the data for urine
given in Table I.

222 W. R. G. ATKINS AND T. A. WALLACE

In the case of a substance much more soluble in water than in phenol, if
curves (Fig. 1) constructed as described be plotted when one of the pure
liquids is replaced by a dilute solution of the third substance in that hquid,
the general outline of the original curve is followed, the ordinates of
temperature being everywhere greater and markedly so in what is termed
the crest region. In each of these curves there is a critical point, but it no
longer coincides with the maximum temperature as is the case when two
pure components are concerned; accordingly it must be recognised by the
beautiful opalescence and equality of volume of the layers.

The rise of critical temperature for dilute solutions, say one per cent., is
roughly proportional to the number of gram-molecules of the substance
added, with regard to its own solvent. This rise is very considerable, for an
amount of solid sufficient to depress the freezing point of water 0°5° may
raise the critical temperature of a mixture of water and phenol by as much
as 4—5°. This rise therefore which can readily be determined to 0:1° gives
about the same accuracy as a freezing point measured to 0°01" with greater
facility and simple apparatus.

APPLICATION OF THE METHOD TO URINE.

If in the preceding experiments urine, which is an aqueous solution of
certain compounds, be substituted for water, a considerable rise in the
critical solution temperature is found. This varies with the concentration of
the urine.

A specimen of normal urine was examined throughout the whole curve,
which was found to be similar to that given by a single substance when
added to a water and phenol mixture, as shown by the broken line in Fig. 1.
The method of procedure was as follows :—Weighed quantities of phenol and
urine were mixed in test tubes, and the temperatures at which they became
homogeneous were noted. A thermometer graduated in tenths from 50—100°
is suitable, but for clinical work one graduated in degrees is sufficiently
accurate. It is not so rough a measurement as it may appear to those
accustomed to Beckmann thermometers, for the temperature differences to
be measured are nearly ten times as great as those met with in freezing point
work. The data from which the broken line in Fig. 1 was constructed are
given below.

Inspection of the above table shows that the critical solution point of the
phenol-urine mixture is at 79°7°, the critical concentration being 36 per cent.
phenol. The critical solution point is by no means at the crest of the curve,

W. R. G. ATKINS AND T. A. WALLACE 223

but is recognised by the characteristic opalescence. The critical solution
temperature of the phenol-water mixture similarly determined is 66°5°. Thus

the rise produced by the solids dissolved in the urine amounts to 13°2°.

TABLE I.

Percentage Percentage Temperature of

weight, phenol weight, urine homogeneity Notes
22°42 77°58 7374 —.-
31°31 68°69 78:0 —
33°99 66°01 78°3 Opalescence.
34°28 65°72 78°7 Good opalescence.
35°93 64:07 79°7 Best opalescence,
36°48 63°52 80:0 Good opalescence.
39°78 60°22 81°2 No opalescence.
48°45 51°55 84°3
54°22 45°78 84°7 Maximum temperature = crest
60-06 39°94 82:2 point.

Having established the similarity between the curve afforded by urine
and by a single added substance it is quite sufficient to determine the
eritical solution point of any urine with phenol in the following manner. Dry
crystalline phenol is placed in a test tube to a depth of about 1 em., and

TABLE II.
Normal Urine.

No A Rk, 2
1 098° Winds 7°86
2 1:07 9°4 8°78
3 1:07 SRL 9-06
4 1:16 8°5 7:34
5 1:41 10-4 7°39
6 1:45 11°6 8:00
7 1:47 afnice} 7°68
8 1:47 125 8°50
9 1-63 11-9 7°28

10 1:67 12°3 7°36

11 crf 14:0 8:19

12 ile? 12°5 7-21

13 1°79 15:0 8°38

14 1°85 16°3 8-80

15 187 11°8 631

16 1°87 15°9 8°71

17 1:87 15*5 8 28

18 1-87 16:0 8°55

19 1:92 15°3 7:96

20 2-48 17°4 7:00

Mean 7:93

224 W. R. G. ATKINS AND T. A. WALLACE

slightly more than enough urine to cover it is added from a pipette. The
mixture is warmed and, if the critical opalescence does not appear on cooling,
urine is added in small quantities till a good opalescence is shown. In this
manner one proceeds from a too concentrated mixture of phenol to one of the
critical composition without the trouble of weighing the components. These
determinations can be very rapidly made, the time depending on how nearly
the critical concentration has been reached at the first trial.

A number of normal urines were examined by this method, their freezing
points being also obtained in the usual manner, and opposite the depression
A the rise in critical solution temperature R, has been tabulated. A third
column gives the ratio R,/A, as shown in Table IL.

It is clear from the above figures that there is a roughly constant ratio
between the two sets of numbers. The fluctuations however are considerably
beyond the range of any likely experimental error, though with highly
coloured urines it is not easy to decide as to the most characteristic
opalescence, (In practice it is permissible, within limits, to dilute such
specimens, but in obtaining the above results this was avoided.) Thus in
Nos. 15-18, which have the same freezing point, it is seen that in three cases
the critical solution points are within 0°5°, whereas the fourth is widely
divergent. As it seemed probable that such cases were to be explained by
some appreciable differences in the relative proportions of the constituents of
the urines in question, the most important of these were estimated in a
number of specimens. Furthermore, in order to study the effect of one
constituent alone, this was added in varying quantities to the analysed urine.
Table III shows the analyses of Nos. 4, 5, 6, 10, 15 of Table II, expressed as
grams per 100 ce.

TABLE SLE

b>

No. cl’ POW 80,” NH; Urea Uricacid TotalN A fh,
4 0°351 0°225 0°173 0-059 1-612 0-015 0-909 1:16 8°5 7:34
5 0°270 0-289 0:238 = 0056 1-902 0-021 1-074 1-41 10-4 7°39

6 0:721 0:227 0°138 0-049 1542 — 0-843 1°45 11°6 8:00
10 0°756 0:222 0°236 0:078 1°854 —- 1-080 NEG} LP 85} 7°36
15 0°596 0:270 0°251 0:097 2252; == 1°397 Wetsyf dL cts! 6°31

Dextrose

10a = 0453 0°1353 0-141 0°047 tes 2°02 0648 U2 1050 8°26
10b = 0°226 0-066 0-070 0°023 0:556 3°53 0°324 0°86 6°8 7°86

The analyses were made in the usual manner as described by Neubauer-Huppert, Analyse des
Harns. The first urea estimations were made by the method of Benedict and Gephart [1908]
while in the later ones Folin’s new method was employed [1912]. The latter gives slightly lower
results than the former, owing probably to the fact that in it uric acid is not hydrolysed.

W. R. G. ATKINS AND T. A. WALLACE 225

In considering what effect the various constituents will have upon the
critical solution point, it must be remembered that whereas the salts are far
more soluble in water than in phenol, urea is soluble in both. It has been
shown, moreover, that urea forms a compound with phenol; this, however, is
very largely dissociated [J. C. Philip, 1903]. Accordingly, though an increase
of urea causes an increase in the depression of freezing point, it causes a
decrease in the critical solution point temperature as shown later on.

The analyses recorded show that Nos. 4, 5, 10 are very similar in
composition, the only notable difference being the larger chloride content of
No. 10. It may be seen that the ratio R,/A is for these specimens quite
constant. In No. 6, while the chloride percentage is high, that of urea is
relatively low and R,/A has increased; whereas, in No. 15, the diminished
amount of chloride and the large rise in urea and total nitrogen combine to
lower &,/A very considerably.

To study the effect of increase in chloride, the sodium salt was added in
various proportions to No. 4, as shown in Table IV.

TABLE IV.
Chloride alone varied.
No. Cl’ A die A
4 0°351 1-158° 8°5° 7°34
4x 0°703 1°472 12°8 8°70
dy 1-055 1-797 18-0 10-02
4z a BL 2-780 29°7 10°69

Here the effect of the chloride in raising the critical solution point is well
brought out by the increased values of R,/A. Separate studies of the other
salts were not made, as, owing to their solubility relations, their effects must
be much the same as that of the chloride. Also they seem to vary less and
occur in smaller quantities.

The addition of pure urea was next studied by adding weighed quantities
to No, 5. This, of course, alters the amount of total nitrogen also.

TABLE V.
Urea and total nitrogen varied.

R,
No. Urea Total nitrogen A R, A
5 1/902 1-074 1-408° 10°4° 7°39
dx 2-446 1°328 1°527 10°2 6°68
5y 2°992 1583 1°721 9°5 5°52
zZ 4-628 2-345 2-157 78 3°62

226 W. R. G. ATKINS AND T. A. WALLACE

Thus while urea alone lowers the critical solution temperature, its effect
in urine is to lessen the rise produced by the saline bodies. Consequently a
specimen which contains more urea than is usual, will have a smaller value
for R,/A. Thus the fluctuations in this ratio, as seen in Table II, are
explained by the mutually counteracting effects of the salts and urea upon
the critical solution point.

PATHOLOGICAL URINES.

It remains to consider to what extent pathological conditions of the urine
may alter the constancy of the ratio R,/A. The simplest case is, perhaps, that
in which the urine is normal in its constituents, but dilute. Table VI shows
the results obtained by examining both normal urimes after dilution and
pathological urines.

TABLE VI.
Normal diluted and pathological dilute urines.

R,
No. : Nature of specimen A Be A
21. Normal diluted a ser bog Hee 0°33° 3:2? 9°70
22. 35 33 ae acts es aie 0:44 4:7 10°70
23. sas nee as ae 0:87 7:6 8°74
24. Patholonienl ‘ : se 0°70 7:0 10-00
25. From right kidney of A, by onlineiadeotRae ae 0°47 4-5 9°58
26. From left kidney of A, a3 pa Sy 0°73 7:0 9°58

27. From both kidneys of B, morning; nothing but water
since previous evening soe 0°63 5:5 8°73

28. From right kidney of B, by Gat herereanen at 1 an m.
Conditions as above. Albumin present ae 0-24 2:5 10°42

29. From left kidney of B, by catheterisation. Conditions
as above. Noalbumin ... a ve 0-30 3°0 10-00
Mean 9°72

These figures show that in very dilute uries the salts have a somewhat
greater effect im raising the critical solution temperature than they have
in those of average concentration. It is, however, to be noted that with
such small values of &,, an error of O0:1° affects the value of R,/A very

considerably. The similarity of R,/A in Nos. 25 and 26, and also in Nos. 28 |

and 29, is noteworthy, and gives reason for supposing that the two kidneys
were in each case excreting fluids almost identical in composition though not
in concentration. The identity in composition of the urine from the right
and left kidney is now completely established as a normal occurrence, though
the earlier work seemed to point to the opposite conclusion [Casper and

Richter’.

W. R. G. ATKINS AND T. A. WALLACE 227

A small number of other pathological urines were examined, as shown in

Table VIT.

TABLE VII.

fh,
No. Nature of specimen . A ke A
30. Chronic parenchymatous nephritis, albumin present ... 1°12 9°3 8°30
31. Specimen with albumin, which was removed by boiling 1-04 9°3 8°94
32. Alkaline, much albumin ai ee sor 1-27 11°0 8°66

In the above three instances, high values of ,/A were obtained, probably
due to a preponderance of salts in the urines. Albumin, being a colloid, can
hardly be expected to have any perceptible effect either in raising or
lowering the critical solution temperature. In the presence of the phenol in
the critical concentration (36 °/,), the albumin in any specimens met with was
not precipitated, even by boiling for five minutes.

DIABETIC URINES.

As the presence of dextrose has a great effect upon the freezing point it
was thought well to see what alteration in the critical solution temperature
its presence in urine would cause. As it is somewhat soluble in phenol, its
addition to a phenol-water mixture was first studied. Results are shown in

Table VIII.

TABLE: VIII.

Dextrose added to phenol-water system.

Dextrose, grams i hi,

per 100 ec. water A R, A
5°05 0°56° 1:7 3-04
4°04 0-42 1-0 2°38
2°02 0-21 O-4 1:90
1°01 0°10 0:3 3°00
0°50 0-05 _O-1 _ 92-00

It will be seen that the addition of dextrose causes a relatively small rise
in the critical solution temperature, the last figure being obviously due to
experimental error. As before pointed out, with low values of &, small errors
cause the ratio R,/A to fluctuate greatly. The effect of dextrose in a urine will,
therefore, be to lower the value of R,/A, for though it causes a rise in R, yet
this is not so great as is the case with the salts. To test this point further,
No. 10 was diluted by the addition of dextrose solution, so as to have the

228 W. R. G. ATKINS AND T. A. WALLACE

composition recorded in Nos. 10a and 106, in Table III. Also dextrose was
weighed out and made up to 100 cc. with No. 15, so that the resulting
solutions had very closely the composition of No. 15, but not exactly on
account of the volume occupied by the dissolved dextrose. The figures
obtained are shown in Table HT and in Table IX.

TABLE. IX.
Dextrose added to phenol-urine system.

Dextrose, grams

No. per 100 ec. urine Nig Rie &

10 — LASS 12-3° 7°36
10a 2-02 (No. 10 diluted) 1-21 10-0 8:26
10b 3°53 (No. 10 diluted) 0°86 6°8 086
15 — 1:87 11°8 6°31
15a 1:08 1-98 11:3 5°71

15) 5°40 2°43 12°3 5°06

Here it is seen that in 10a and 10} the dilution effect has more than
counterbalanced the addition of dextrose, so that R,/A is slightly raised. In
15 a and 15 b, however, it is evident that the already low value of R,/A
afforded by No. 15 is depressed still further.

In Table X are shown the values given by diabetic urines, not intention-
ally altered in any way. They had not been heated.

TABLE X.

Diabetic wrines.

Dextrose, grams fi,

No. per 100 cc. A R, A
33 1°87 igi tgl= 80° 6°84
34 5:62 1°56 10:0 6°41
35 ? 1°75 11:0 6°30
36 0-19 1:94 15°3 7°88
37 2°34 2°03 16°0 7°88
38 3°38 2°27 16°5 7°56
Mean 7:14

Here again the mean value of R,/A is lower than with normal urines,
though, owing to the variations of the other constituents of the urine, the
decrease is not proportional to the sugar concentration.

We hope to make further experiments with pathological urines, taking
into account other abnormal constituents of diabetic specimens.

—s OO

\ ell

W. R. G. ATKINS AND T. A. WALLACE 229

SUMMARY AND DISCUSSION OF RESULTS,

The examination of a considerable number of normal urines has shown that
the rise in critical solution temperature 1s roughly eight times as great as the
depression of freezing point. This holds good for urines of normal concentra-
tions, but with very dilute urines the critical solution temperature rise is
somewhat greater, consequently &,/A has a value greater than eight. This is
also the case in specimens which, for any cause, contain an unusually large
quantity of salts. On the other hand, the presence of an exceptionally
great amount of urea occasions a smaller rise of critical solution temperature ;
glucose, when present in quantity, has the same effect, but it is far less
marked. In the last two cases accordingly the ratio R,/A is somewhat
below eight.

This relation between the freezing point depression and the rise in critical
solution temperature enables one to transfer the conclusions established by
the former method to the latter. Thus A. v. Koranyi [1897, 1899 : Quoted
from Casper and Richter] gives 13° — 2°3° as the usual limits for the freezing
point of urine, from a normal kidney under ordinary conditions. Taking
R,./A=8, this range corresponds to a rise in critical solution temperature of
10°4° — 18-4".

Lindemann [1899: Quoted from Casper and Richter] gives a more
extended range, 0°9° — 2'7°, for values of A, and states that kidney patients
almost always excrete urine for which A is less than 10°. Accordingly, taking
Lindemann’s limit, 0-9", and allowing for experimental error and fluctuations
due to the composition of the sample, it seems safe to take a rise of 8° C. in
the critical solution temperature as indicating that the kidney supplying the
sample was functioning satisfactorily. It must be remembered, however,
that the converse is not always true; thus a specimen giving a rise of 7°
might really be normal, as, owing to the ingestion of a large amount of liquid,
or to the presence of much urea, the value of R, could very well be lowered.

In the latter limiting case, it would be well to make a freezing point
determination or rough urea determination by the hypobromite method, to
see if there was an unusual quantity of this substance. Ordinary values of
R, lie between 11° and 16° C.

In conclusion, we wish to thank Dr J. Timmermans, of Brussels, for

putting his extensive knowledge of critical solution point determinations

230 W. R. G. ATKINS AND T. A. WALLACE

entirely at our disposal during his visit to Trimity College. A detailed
treatment of the subject will be found in his paper [1907].

REFERENCES.

Atkins (1908), Brit. Med. J. 1, 254.

Benedict and Gephart (1908), J. Amer. Chem. Soc. 30, 1760.
Casper and Richter, Diagnosis of Kidney Disease, p. 95.
Folin (1912), J. Biol. Chem. 11, 507.

Koranyi (1897), Zeitsch. klin. Med.

(1899), Berlin ktin. Wochsch.

Lindemann (1899), Arch. klin. Med. 65.

Philip (1903), J. Chem. Soc. 83, 814.

Schreinemakers (1899), Zeitsch. physikal. Chem. 29, 597.
Timmermans (1907), Zeitsch. physikal. Chem. 58, 129.

1 See also Findlay, The Phase Rule, and Young, Stoichiometry. London, Longmans,
Green & Co.

XXIV. QUANTITATIVE RELATIONS IN
CAPILLARY ANALYSIS.

By HANS SCHMIDT.
From the Bacteriological Department, Lister Institute.

(Received March 20th, 1913.)

When a dispersed system is allowed to fall on filter- or blotting-paper
or to rise up into a strip of such a paper, one observes that the dispersed
phase remains behind the medium of dispersion, which is in most cases
water. This phenomenon, which involves an increase in the concentration
of the dispersed phase in the paper, can be used for the qualitative testing
of the nature of the substance as has been shown by the capillary analytical
experiments of Goppelsroeder [1909].

Holmgren [1908] tried to establish a formula, which gives a quantitative
relation between the extents of the dispersion medium and the dispersed
phase. He experimented with diluted acids, especially with hydrochloric
acid, by allowing the latter to drop on blotting-paper. Using Congo-red
as indicator, he was able to show how far the acid was spread over the
paper compared with the water. If a good and homogeneous blotting-paper
is used, the figure caused by the diffusion of the drop approaches a circle.

Holmgren supposed that the relation between the water-zone, the largest
radius of which may be &, and the inner circle of radius r caused by the
acid, is constant for each concentration, so that the quotient of two different
concentrations is equal to the quotient of these relations. Therefore, as
[7R?—7r*] is the surface of the water-ring, the following relation exists
according to Holmgren :

If we signify by P that concentration for which

2
fi SS
Ry? - 7?

232 H. SCHMIDT

the following equation results :

or P= 0c nee ee ee (3).

If, therefore, the concentration C and the radii of the circles produced
on blotting-paper by any drop of the acid are known, P is determined from
equation (3). Holmgren supposed P to be constant for the same paper and
independent of the concentration. He called P the paper constant and was
then able to determine any other concentration by the equation (2), but
with the limitation that this method was not to be used except for very
dilute solutions between 20 and 0:01°/,.. The results of Goppelsroeder’s
experiments, which give only the measurements of capillary height in strips
of filter-paper, can also be treated by this formula of Holmgren, if instead
of the height-measurements their squares are taken.

Skraup [1909, 1 and 2] and his co-workers were able to confirm the
applicability of this formula for capillary-analysis. On examining the amount
of free hydrochloric acid in the gastric juice by this method, I found [ Schmidt,
1913] that the results obtained agreed very closely with those obtained by
titration. Later some doubts arose in my mind as to the correctness of
Holmgren’s formula and occasioned me to undertake the following in-
vestigation.

TECHNIQUE OF EXPERIMENTS.

I experimented exclusively with hydrochloric acid, an exactly adjusted
normal solution of which was taken as the basis for the following dilutions
between 2:0 and 0:05°/,. One drop of these solutions was allowed to fall
on blotting-paper from a short distance. Much attention was paid to insure
that the drop had always the same volume. With an ordinary pipette it
was not possible to get the volumes cf the drops as exact as was desirable.
I therefore used the dropping surface of Traube’s stalagmometer, which gives
exactly equal drops, if the temperature be kept constant and the instrument
be not allowed to deviate from the vertical position. Blotting-papers are
much better fitted for capillary analysis than filter-paper, but they must
not be too thick and must have a smooth surface and a most homogeneous
structure. I obtained a series of different blotting-papers from the firm of
C. Schleicher und Schiill, Diiren i. Rheinland. The papers marked Nos. 128

toe So eae ——w i eh “

H. SCHMIDT 233

and 117 proved to be specially suitable for my purpose. The papers were
stained by floating them rapidly through an alcoholic solution of Congo-red
(0:04°/,) or methyl-orange (0'1°/,). In order to prevent the water from
evaporating or at least to reduce the evaporation to a minimum, as soon as
the drop was absorbed the paper was placed between two glass plates and
a weight laid on the plate to avoid any wrinkling. The plates were then
put on a small frame under a glass globe, the interior of which was kept
saturated with moisture by means of wetted filter-paper below. To neglect
the evaporation leads to false conclusions. The radii were measured in
millimetres by the use of transmitted or reflected light, and since the figures
of the larger spots are mostly elliptical owing to the direction of the fibres
in the structure of the paper, the main axes as well as two oblique axes
were measured, so that the given result is the average of the measurement
of 4 diameters. The number of figures examined to determine any single
point is in most cases at least 4. The measurement took place after exactly
one hour, the process having then practically reached a standstill. In
reality one may sometimes observe an increase of the radii even after
24 hours, but the amount of it is too insignificant to be noted, especially
if only one drop is taken.

Referring to dilutions of acid under 0°1°/, the measurement must take
place earlier than one hour especially when Congo-red is used, because the
blue colour disappears after a certain time, owing to chemical action.

EXPERIMENTS.

The mass of one drop given by a stalagmometer at the temperature of
18—20° is indicated by a. If the drop consisted of distilled water, its weight
was found to be 01160 gr., and if it consisted of HCl of a concentration
between 2°0 and 0:05 °/,, I found the average weight of a drop 0:116 gr., the
differences of the specific gravity in so highly diluted acids being so small
that they can be neglected without introducing an error which exceeds the
error of observation.

Therefore, putting the volume equal to the weight (in case of dilute HCl)
I give a the value of 0:116 gr. Supposing m parts of hydrochloric acid
to be present in the drop a, the concentration is given by C= - by means
of which m is easily calculated.

Under the above described conditions, a series of observations was made,
of which the following may be reproduced :

234

Indicator :

H. SCHMIDT
TABLE LI.

Paper No. 123. a=0°116 gr.

Congo-red Methyl-orange

| Sa at ae PL ce oa nl ea eh ie ae SN
Number of drops:— la 2a 3a la 2a
a ae i SN Ta teary (am a 5
Hei), am v r R r i if R ? R r
2-0 2°32 23°12 21°87) 32 29°37 = 87-87 «34:25 23°12 21°62 29°81. 26°43
1:0 1:16 23°5 «19°62 §©= 32°37 S265 88°5 —- 31'25 23°37 19°56 31°75 26
0°75 = 0°87 24 18°37 © 32°37 24°5 38°7 = 29:5 23°37 18 32 24°25
0°5 0°58 23°62 16°56 31°87 22°18 37:8 26°5 24 16°87 32°25 22°43
O-4 0-464 23°87 15°75 32°75 21°5 38-2 22°12 (?) 23°62 15°56 32°37 20°75
0°3 0°348 23°62 14°63 32°25 19°68 37:0 2375 23°62 14°75 31°62 19°56
0:2 02232) 5 23-o 7 elecol me olson Liles on .oummeOsO2 23°62, 12°25) 32-12 uae
O01 OSG 9237550 31-62 13:5 37°38 16°62 23°62 10°75 32 13°75
0°05 0:058 23-7 8°62 = 23437) —— — -—
Means of R 23-48 32-09 23-53 31°74
TABLE II.
Paper No. 117. a=0'116 gr. Indicator: Congo-red.
Number of drops: la 2a Ba
2S SSS, Sie pal La a SSS SSS
HCl °/, m R ? t ? R ip
1:0 1-16 26°5 24-25 36:0 32°41 42°25 37:5
0:75 0:87 26°85 22-40 35°75 30-1 42:5 35°5
0°5 0°58 26°68 20°62 35°43 27°75 41°62 33°18
0-4 0-464 26°87 20°5 35°43 27°5 43 33°0
0-3 0°348 25°87 18°85 35°87 25°2 42-37 30°3
0-2 0°232 26°25 16:37 36-0 22°81 42-0 26°75
0-1 0-116 26°25 13°63 35°75 18°5 42°25 22°75
0°05 0°058 26°75 = - — -= —
Means of R 26-50 35°75 : 42-98
TABLE. III.
Paper No. 123. a=0°0311 gr.
Indicator : Congo-red Methyl-orange
— SS oo
HC] °/, m R ? R 7
1-0 0-311 14°75 13°29 o --
0°5 0-155 15-1 12°5 12°5 9°3
0-4 0°124 14°65 9°4 12°8 8-9
0:3 0-0933 14:84 8°85 12-7 8°25
0-2 0:0622 15°15 7:95 12°8 7
0-182 0-0566 11838) 6-4 12-9 6°36
O01 0:0311 14:26 6:0 13-l 6
iflpis Means of Rk 14:6 ; 12°8

On the assumption that the aggregation of acid in the inner circle is

of the nature of an adsorption process, the relationship between the two

variables m and 7 must be expressible by the exponential equation

r= Bm?.

[i

H. SCHMIDT 235

It is highly probable also that the concentration of the acid or that
of the hydrogen ions is continuously decreasing from the centre to the
periphery of the spot-system in the case of any drop of a certain concentra-
tion falling on the paper. L suppose then, that r indicates that point of
hydrogen ion concentration which corresponds with the sensibility of the
indicator. But owing to the impossibility of proving experimentally the
gradual decrease of the hydrogen ion concentration in the paper between the
centre and 7, my supposition remains an hypothesis. To show how in
different concentrations the relation between m and r is expressible by
the equation r= 8m?, we use the graphic method in constructing the
m-r-curve Fig. 1.

40°1R,= 37-9
3a

2
t R, = 32:09

fe) O25 1-0 —>m 2:0
a Drteg sale

The m-r-curve rises very steeply in the beginning and then runs nearly
parallel to the m-axis, asymptotically approaching a straight line, of which
the equation is r= R. The difference (R —7) is identical with the breadth
of the water-zone and the latter is of course equal to R in case of distilled
water, when m or C=0. (R-—v) decreases with increase in the amount
of m. Theoretically, the decrease of (R—vr) is infinite but in reality m
reaches very soon such an amount, that the water-zone is no longer visible.
When the latter effect occurs with one drop, the phenomenon can still be
obtained by increasing the volume of the drop or augmenting the number
of drops, but the errors of observation are also increasing and in such a

Bioch. vu 16

236 H. SCHMIDT

degree that for concentrations higher than 2°0°/, the method of Goppelsroeder
is preferable, which, as already mentioned, involves the measurement of the
capillary height in strips of filter-paper.

That £ is in reality the constant, which it seems to be in the experimental
data, and that R depends—in the same paper—only upon the volume of the
drop and not upon the concentration, will be shown later.

By taking the logs. of the equation r= 8. m”, we get

log 7 = p. log m + log B.
This is the equation of a straight line, which is very nearly approached by
plotting the values of log r and log m, as is shown in Fig. 2.

04°03 02 O1 [0 0-9.) 0-8_; 0:7, 0-6, 0°5_, 0-4, 034

0-9

Fig. 2.

By graphic interpolation according to Freundlich’s [1907] procedure the
values of 8 and p are found to be

p= 0264
and fob NS Os
In the following table the values are calculated by the equation
(ie Ko 5) /( aca MORE ORNS Asc So (4).
TABLE IV.
Paper 123, Congo-red. 1 drop (a=0°116 gr.) r=18-75m?-64,
C=" / HCl m r calculated r observed A cale. — obs,
2:0 2°32 23°41 21°87 — 1°54
1:0 1:16 19°49 19-62 +0°13
0°75 0°87 18-08 18°37 +0°29
0°5 0-58 16°31 16°56 +015
0-4 0°464 15:12 15°75 +053
0°3 0°348 14°19 14°63 +0°44
0-2 0-232 12°74 12°37 — 0°37
0-1 0-116 10°61 10 — 0°61

0:05 0:058 8°84 8°62 — 0°22

H. SCHMIDT 237

Although the agreement is fairly good, a further calculation of 8 and p
was made by the following procedure in order to diminish the ditterences.

HCl °/, 100 x m 100 xr log m log r
2-0 232 2187 2°36549 3°33985
1-0 116 1962 2°06446 3°29248
0°75 87 1837 1°93952 3°26411
0°5 58 1656 1:77085 3°21906
0-4 4674 1575 1°64652 3°19728
0°3 34°8 1463 1°54158 3°16524
0:2 23°2 1237 1°36549 3°09237
0-1 11°6 1000 1:06446 3
0:05 58 862 0°76343 2°93551

- Mean ~=—:1°613533 — 3°167322
Deviations from means :
my r; mM, +7; (m,)"

+0°75195 + 0°172528 0°12957 0°56543

+ 0°45092 + 0°125158 0°056432 0°20333

+ 0°32598 + 0°096788 0:031551 0°051067
+ 0°15731 +0°051738 0:0081424 0024746
+ 0:032987 +0-029958 0:00098825 0:0010881
— 0:071953 — 0002082 0:000149806 0:0051772
— 0°248043 — 0:054952 0-013630 0:°061523
— 0°549073 — 0°167322 0-091871 0°30147

= 0°850103 — 0°231812 0°19705 0°72267

> : 52
parm 1) _ 07529384 =0-2733.

D(m,)? ~ 1-9365013

The value 0°273 for the exponent p gives the following values for
log 8, = log r — 0°278 log m.

log 8, and for p,
2°694072 494-4
2°728883 535°6
2°73462 542°7
2°73561 544-0
2°74778 5594
2°74438 555°1
2°711959 524°3
2-70940 512-1
2°72709 533°4

As the figures for r and m have been multiplied by 100, the values of

log 8 in the equation
a

are calculated from the following equation :
log B = log B, + 2. (0°273) — 2.
These values of log 8 are therefore the following :
16—2

238 H. SCHMIDT

log 8 B
125007 17°78
1:27488 1883
1:28062 19-08
1:28161 19:12
1-29378 19-66
1:29038 19°51
1:26559 18-43
1:25540 18-00
1:27309 18°75

Mean of 8 18°88
It results therefore for the relation between 7 and m the equation
r= 18°88. m0 10... MS Soh ee (5).

Thus substituting the calculated values for 8 and p the following equation
was obtained :
He \MoKelouni/ (mic.
which gives a far better agreement between calculation and observation, as

is shown in the following table :

TABLE V.
Paper 123. Indicator: Congo-red. One drop. a=0'116 gr. r=18°'88. m7,
C= 9/ PEC! m r calculated r observed A

2-0 2°32 23°75 21:87 — 1:88
1:0 1:16 19°66 19-62 — 0:04
0°75 0°87 18°17 18°37 +0:2

0°5 0°58 16°34 16°56 +0°22
0:4 0:464 15:11 15°75 +0°64
0:3 0:348 14°15 14°63 +0:48
0:2 0:232 12°67 12°37 —0°3

0-1 PO AG 10°48 10:0 — 0°48
0:05 0-058 8°677 8°62 — 0°05

Idem, but a=0-0311 gr.

1:0 0-311 13°72 13°29 — 0°43
0°5 0°155 11-09 12°5 +1:41
0-4 0°124 10°67 9°4 —1:27
0:3 0:0933 9-880 8°85 —1:03
0-2 00622 8845 7:95 — 0°89
0:182 0°0566 8°620 6-4 — 2°42
0-1 0-0311 7°320 6 — 1:32

The reason why the second part of the Table V does not show so good an
agreement is the inexact measurement of drops by using an ordinary
pipette instead of a stalagmometer. By increasing the number of drops, so

that instead of a, 2a or 3a were taken of the same concentration C=", the
radii of the corresponding spots must be

Ti 2) ONS:

or, generally speaking, if [na] is taken as the dropping mass, the corresponding

7
‘

radius is r/n, r being the radius for a.

H. SCHMIDT

239

Therefore, if the radii found for

a are supposed to be fixed, the radii for 2a or 3a can be calculated. The
following table gives the results:

TABLE VI.

Paper 123.

Number of drops: a

Cc = HCl

2-0
1:0
0°75
0-5
0-4
0°3
0-2
0-1
0°05

8°62

Indicator: Congo-red.

Ya

a
r, calculated

30°83
27°66
25°90
23°34
22-20
20°63
17°44
14-14

29°375
26°5
24°5
22°185
21°5
19-685
17°125
13°5

17°32

a=0'116 gr.
3a

ee a

r, calculated r, observed
37°63 34°25
33°98 31°25
31°81 29°5
28°68 26°5
27°27 22-125 (?)
25°33 23°5
21°42 20°625

16°625

The agreement between the calculated and observed values for r is still
fairly good for 2a.

But by increasing the number of drops beyond 2, the

experimental error increases, because the figure produced on paper loses
the resemblance to a circle, so that the average of r is determined with
more arbitrariness.

By using the equation (5)
P= NS Gam =

for cases where, generally speaking, (na) was taken, m becomes (nm). The

values of r, for 2m and of r, for 3m are shown in the following table :

Paper No. 123.

c=/, HCl
2-0

2m
4°64
2°32
1:74
1:16
0-928
0-696
0-464
0-232

3m
6°96
3°48
2°61
1°74
1-392
1:044
0°696
0°348

A

r, calculated
=T2=Tr/2
— 2°12
—3°91
— 3°94
— 3°68
— 3:82
— 3°53
-- 2°13
—1:47

A

rz calculated
—73=1/3

— 5:57

— 7°45

— 7:28

— 6°72

— 6°74

— 6°23

— 4°32

TABLE VII.
Indicator: Congo-red. a=0°116 gr. r,/n=18-88 (nm)??”*,
observed
log 2m r, calculated rr, observed —cale.
0°66652 28°71 29°37 +0°66
0°36549 23°75 26°5 +3°25
0-24055 21°96 24:5 + 2°54
0:06446 19°66 22-18 + 2°52
0°96755 —1 18°38 21°5 +3°12
0°84261-1 17°10 19°68 + 2°58
0°66652 —1 15°31 17:12 +1°81
0°36549 — 1 12°67 13°5 + 0°83
A observed

log 3m r, calculated 7, observed —cale.
0°84261 32°06 34°25 + 2°09
0-54158 26°53 31:25 +4°72
0-41664 24°53 29°5 +4:97
0°24055 21-96 26°5 +4:54
013364 20°53 22°12 (?) +1°59 (?)
0-01870 19°10 23°5 +44
0°84261-1 17:10 20°62 +3°52
0°54158 —-1 14°15 16°62 + 2°47

—3:17

240 H. SCHMIDT

‘The deviations of the calculated figures of r from those obtained by
observation are so considerable as to make it clear that the volume used
for capillary-analytic purposes must not exceed a certain quantity. But the
deviation of the calculated values 7,/n and 18°88m°?? must be due to the
inexactness of the constant in the equation r= 8m? and to the experimental
errors for 7;. In reality the relation

Mig 8 ania ee cin ee (6)

must exist.
THEORETICAL PART.

On detailed consideration of the processes taking place in the paper, it
becomes evident that an increase of concentration of the hydrochloric acid has,
taken place. :

If C signify the original concentration and C, the resulting concentra-
tion in the coloured spot, the relation between these concentrations may be
deduced in the following way. The paper is supposed to be previously
stained either with 0°04°/, alcoholic Congo-red solution, or with 01°/,
alcoholic methyl-orange solution, and the staining to be done in the same
manner, the paper not being allowed to adsorb much of the dye, a process
which is at room temperature only a matter of time, as has been shown by
W. M. Bayliss [1906]. Observation shows that stained paper differs slightly
in quality from unstained, a fact to which reference will be made later.

Let [a] be the constant quantity of the drop and [na] the quantity of

hydrochloric acid of the concentration C = = , dropping on a piece of blotting-

paper which is supposed to be always of the same quality. The quantity [na]
of the acid may spread over the paper, so that the acid reaches from the
centre to the distance 7/n, and the largest radius of the water-zone may
be R /n, if r and & are the corresponding values for [a]. In case of another

drop-volume, the weight being g’, the radius becomes r/2, g being
0-116 gr.

As the acid constituent of [na] is in the interstices of the paper, the
volume [na] must be equal to the volume of the paper, the air being
included, minus the specific volume of the paper-material itself in the
same space. The volume occupied by [na] is wn? and the specific volume
of the paper-material is its weight divided by its specific gravity s. The
weight of a portion of paper of square millimetre surface and thickness 6 is
indicated by p. Therefore the following equation results for [na]:

N= TRS = TNEOe aactea, eas ckeee eee (7),
s

=

H. SCHMIDT 241
or, if'n = 1, a= Rm (8-2).
(8 -2) represents a constant characteristic for the paper and may be

indicated by k, so that

and a= R?.k,

or ie Py St Retr me ent ee te or ae (8).

That is to say, R depends only upon the variation in quality of the
paper and upon the drop-volume which is taken, but it is independent of
’ the concentration C. Therefore R is to be considered as constant for all
concentrations, if the same paper and the same drop-volume are used.

The easiest and quickest way to find & is by means of the equation (8),
but by weighing the paper and determining its specific gravity and thickness
the same value for / will be obtained by the equation k= 7. (6 -2) :

In order to determine the thickness 6 of the paper, a certain number of
sheets are placed between two glass plates and by pressing them very
tightly the measure of the distance may be found, which gives when divided
by the number of sheets a rough value of 6. The best manner to determine
the specific gravities is the use of a pyknometer. The following data were

thus obtained :
Paper No. 123 6=0°17 mm.
p=0-'1408 mer.
S—3o
k=0-20923

Paper No. 117 6=0°16 mm.
p=0-'118 mer.
Sok
k=0'1651
When the value of & is found in this way, R can be calculated by the
equation (8) as is shown in the following table :

TABLE VIII.

R calculated R observed
oa == = a ‘

Paper No. a(mgr.) a 2a 3a a 2a 3a
123 116 23°537 33°28 40°77 23°48 32-09 37°9
117 116 26°49 37°48 45-90 26°5 35°75 42-28
193 13-1 12-165 a a 12°8 Methyl-orange a

14:6 Congo-red

242 H. SCHMIDT

The agreement is very close for a, but here, too, it 1s apparent that
increase of the number of drops increases the inexactness of the observations.
If the value of R, is fixed, R, and R, can be calculated as R,/2 and R, /3,
which give also a fairly good agreement, as the following figures demonstrate :

Paper No. a (mgr.)
123 116
117 116

TABLE IX.

Ry
23°5
26:5

Ro=Ry/?2 R3=Ry/3
33-13 40-70
37°36 45°89

If C=", the quantity na of HCl contains nm parts of it. These

a

nm parts of HCl have been retained by adsorption in the inner circle (7),
and if v indicates the volume of its interstices, the following equation must

express the resulting concentration C;, :

1 nm.m
C=——= .
v

=

I neglect the possibility, that the fibres may swell, supposing that the
adsorption process is much sooner finished than the swelling of the fibres.
By analogy with equation (7), v can be expressed by

v= 7rnrd — TN

p22.
s

After substituting this value for v, C, is given by

or ‘OF

eeeeeeerene

This is the equation giving the resulting concentration in the coloured
spot, and if z represents the increase of concentration, so that

and one substitutes

and

the following equation results :

CO=2+C,

H. SCHMIDT 243

If we substitute k. R? =a,
and x= 0
a ¢

we find oe ae ee eh trise aaa hata Seam T (10).

Therefore z is found to be equal to the value of the paper-constant P of
Holmgren (3).

Holmgren was of opinion that the relation between the resulting
concentration C; in the paper and the original concentration C can be

expressed by the quotient a so that

This view seems to be correct, for putting

ee ea
C; = re ay r2k?
we find R?.C.k=™m, and substituting R’k=a, the definition of C= <

results.
C.R

VT = Sanya —
When Holmgren eliminated 7? = Pi

out of the equation (3) and substi-

tuted this value for 7? in the equation (11), he got
C—O ae eae SS dca nsnuucews oveeusr (12).

Therefore Holmgren’s paper-constant corresponds with z in equation (10)
representing the increase of concentration by the adsorption, P therefore
not being a constant but a function of the concentration in question. The
following table shows the values for C, calculated by the equation (9),
P calculated by the equation (2) and C, calculated by formula (12).

TABLE X.
Paper No. 123. Congo-red. One drop (a=0°116 gr.). k=0-20923.
7am ky Seas P=0-441
C=, HCl Sank: lo Eat 7 C,=P+C C,=0°441+C

2:0 2°32 0-305 2°305 2°441
1:0 1:44 0:432 1-432 1:44
0°75 1°23 0°475 1°225 1:19
0:5 0-994 0°505 1-005 0-94
0-4 0-894 0488 0°888 0°841
0:3 0:777 0-472 0:772 0-741
0-2 0°724 0:520 0-720 0-641
0-1 0°554 0°451 0°551 0°541
0:05 0°373 0°321 0:371 07491

Mean of P=0°441.

244 H. SCHMIDT

TABLE X (cont.)

Paper No. 117. Congo-red. One drop (a=0°116 gr.). k=0°1651.

_ m0 By ini P=0-285
C=), HCl C1 ay Plo) De gas CH=P LG C,=0°285+C

1:0 1:19 0:194 1194 1-285
0:75 1-05 0-299 1-049 1:035
0:5 0°825 0-325 0-825 0-785
0:4 0-668 0-286 0:686 0°685
0-3 0:593 0-292 0°592 0°585
0-2 0-524 0°323 0523 0-485
0-1 0°377 0-277 0:377 0-385

m

We see that the values given by OS and C=P+C are almost

identical as the theory demands. But if P is supposed to be constant
according to Holmgren, and if the average value is taken for P, we find
that the values no longer agree so closely. This difference is better demon-
strated by the graphic method in constructing the C-C,-curve (Fig. 3).

Fig. 3.

It is evident that C,= P+ C represents a straight line, which cuts the
Ci-axis at a distance P from the origin and the inclination of which
must be 45°.

H. SCHMIDT 245

= 5 represents a parabola, by means of which it is possible to
demonstrate the fact that for C=0, C, must also be 0, a fact which is
not expressible by C,= P + C, P supposed to be constant.

In reality, as P varies»with the concentration, the equation C,= P + C
represents a system of parallel straight lines and the points indicating the
relation between C and C, follow the course of a parabola cutting this system.

Nevertheless, the disagreement between Holmgren’s supposition and the
observed values is so slight, that for concentrations not under 0:1 °/, his
method is for practical purposes very useful on account of its extraordinary
facility. To demonstrate the very close agreement of the calculation by
Holmgren’s formula (2) with the reality I give the following table:

TABLE XI.
Paper 123. a=0°116 er.
nw 0°273 r
P=0-441 (const.) i 18°88
r? m
j= — —=?
c="), HCl Saree. R?-? a lo
2°0 2°87 3 1:4
1:0 1-02 0-989
0°75 0-69 0°77
0°5 0°43 0°53
0-4 0°36 0-44
0°3 0°28 0°33
0:2 0-16 0-18
O01 0-097 0-084
0°05 0-068 0-048

With regard to a possible combination of the dye and the acid in the
paper, the amount of Congo-red or methyl-orange can be approximately
calculated. Bayliss [1906] has shown in regard to Congo-red, that the
staining of paper is an adsorption process. The temperature-coefficient of
the reaction-velocity is so low, that at room temperature at least 24 hours
are required for the attainment of equilibrium. I used a clear looking
alcoholic solution of Congo-red (0:04 °/,) and of methyl-orange (0°1 °/,) always
at room-temperature, and dipped the piece of paper into these solutions only
for a few seconds, then allowing the alcohol to drain off and drying the paper
at 37°. It can therefore be presumed that the amount of dye adsorbed
is very small, and if it be assumed that in so short a time only that amount
of Congo-red (or methyl-orange) can be adsorbed which was present in the
solution filling the interstices of the paper-fibres, the possible error is
certainly very small.

Of an 0°04 °/, alcoholic Congo-red solution each emm. contains 4.10~’ g. dye,

246 H. SCHMIDT

or in case of 0:1 °/, methyl-orange 0°1.10~7 g. dye. As already shown, the
volume of the interstices, which belongs to a surface of zr?, has been found
to be 7.k. According to Table I r was found =10 mm. in case of an
0-1 °/, HCl, which gives 7.4 =20°9 or 21 emm. Therefore the amount of dye
adsorbed in these 21 cmm. is in case of Congo-red 84.107’, and in case of
methyl-orange 2°1.10-7. The same consideration gives 21.1077 as the
amount of an 071 °/, HCl in the same volume of interstices.

These figures are mentioned although I was not able to prove that
any chemical action resulting in a perceptible decrease in the amount of free
acid took place between dye and acid. No difference in the measurement
of the radii of the acid-circles could be observed on varying the concentra-
tion of the indicator-solution. Holmgren, who compared the effect of 1:0
and 0:1°/, solutions of Congo-red in water on the extent of the surface
produced by allowing a drop of acid to fall on paper under similar conditions,
thought it also very improbable that the amount of dye by itself plays any
important part.

But if the effect of adsorption in unstained paper is compared with that
obtained in stained paper, a difference can be observed.

Experimenting under the conditions described in the introduction, one
observes in unstained paper that the extent of the water-zone exceeds by
25 mm. that found in paper stained by the alcoholic indicator-solutions.
This difference in the extent of the water-zone is constant, when the con-
centration of the acid employed varies, a fact which agrees with the
circumstance that R is independent of the concentration.

The extent (7) of the acid-spot is also less in stained paper than in
unstained, but according to observation this difference increases the more
the acid has been diluted, which favours the idea that a chemical action
occurs between acid and indicator.

The weight of about 500 gr., which I used to prevent the paper from
wrinkling, influences the extent of the water-zone, but the enlargement
caused by the pressure is very insignificant and does not exceed 0°1 mm.

In studying the influence of pure alcohol on the paper regarding the
adsorption of hydrochloric acid and the capillary extension of water,
experiments showed that pure alcohol has an inhibiting action on water as
well as on diluted acid.

I may suggest that the fibres of the paper when treated by alcohol,
whether pure or combined with dye, are inhibited or prevented from swelling
by water, and thus the interstices become larger, which would explain the
observation that the radii become smaller.

H. SCHMIDT 247

To apply these considerations to the method which involves the measure-
ment of the capillary height in strips of paper the different volume of liquid
must be taken into account. Skraup and his co-workers [1910] were able to
show that the amount of water raised in strips of filter-paper is different in
different parts of the strip, decreasing in a hyperbolic manner [p. 887].
Whether a similar decrease of the amount of water occurs in the case of
a drop producing a circle on the filter-paper is not proved, and I think it
very improbable, believing that this phenomenon depends upon gravitation
and therefore upon the inclination of the strip, the influence of which is
shown by Goppelsroeder [1909] in regard to the capillary height.

I found also in strips of paper that the fact whether they are untreated,
stained by alcoholic Congo-red solutions or only treated by pure alcohol,
must be taken in account. ‘Thus under the same experimental conditions
distilled water rises quicker in the unstained paper and the contrary takes
place in case of a dilute hydrochloric acid.

The following table contains measurements of capillary heights in mm.,
the time having been constant for each experiment, but I could not find
any satisfactory explanation of the phenomenon.

P ios Treated by
aper 12 Ce ee
untreated “Pure alcohol Congo-red
‘ { Water : 59 71 7a!
Stee CE Noid 50 59 59
85 105 77
Distilled water 66 78 62
76 89 69

Holmgren [1908] found that the relation between the capillary heights
of the acid and the water increases according to the concentration, and that
this relation is constant for the same concentration ; Skraup and his co-
workers [1910] found for all acids (with a few exceptions) that the stronger
the acid the higher is the degree of adsorption and vice versa.

I hope later to be able to show how the mathematical considerations
described in this paper can be applied also to Goppelsroeder’s and Skraup’s
experimental data obtained by the measurement of the capillary height.
But I am conscious of the fact that these formulae are still far from being
able to describe all the possibilities in such a complex phenomenon as the
adsorption of acids by paper.

248 H. SCHMIDT

CONCLUSIONS.

1. Diluted acids produce a ring system when dropped on blotting-paper,
the acid remaining behind the water.

2. The radius 7 of the coloured circle produced by the acid in the
paper is connected with the concentration C by an exponential equation
of the form r=£8.C”.

3. The radius R of the water-zone is independent of the concentration

and can be determined by the equation R= af, - a being the volume of

the acid drop in question and k being a constant which depends upon the
quality of the [stained] paper.

4. The adsorption causes an increase of concentration. The final
concentration is found to be a parabolic function of 7.

5. The increment of concentration varies with the initial concentration.

6. Holmgren’s calculation, which assumes that the increment of con-
centration is a constant dependent only upon the quality of the paper,
is theoretically incorrect, but it has been shown that it may be useful
for practical purposes.

REFERENCES.

Bayliss (1906), Biochem. J. 1, 175.

Freundlich (1907), Zeitsch. physikal. Chem. 57, 391.

Goppelsroeder (1909), Zeitsch. Chem. Ind. Kolloide, 4, 23 (where previous papers by the
same author are quoted).

Holmgren, J. (1908), Biochem. Zeitsch. 14, 181.

— (1909), Zeitsch. Chem. Ind. Kolloide, 4, 219.

—— (1912), Arch. Verdauungskrankheiten, 17, Erganzungsheft 57,

Schmidt, H. (1913), Deutsch. med. Wochensch. Nr. 8, 358.

Skraup (1909, 1), Sitzwngsber. K. Akad. Wiss. Wien. Math.-naturwiss. Kl. 118, Abt. ILb, 459,
559.

— (1909, 2), Monatsh. 30, 675, 773; 31, 565, 873.

——, Krause, v. Biehler (1910), Sitzwngsber. K. Akad. Wiss. Wien. Math.-naturwiss. Kl.
119, Abt. IIb, 565.

——, v. Biehler, Lang, Philippi, Priglinger (1910), Sitzungsber. K. Akad, Wiss. Wien, 119,
Abt. IIb, 873 (887).

Oke

XXV. A NOTE ON THE HOPKINS AND COLE
MODIFICATION OF THE ADAMKIEWICZ TEST
FOR PROTEIN.

By VERNON HENRY MOTTRAM.
From the Physiological Laboratory, University of Liverpool.

(Received March 19th, 1913.)

INTRODUCTION.

The following brief piece of research is the outcome of an unexpected
failure in class-work to obtain the well-known reaction with “reduced oxalic ”
acid [Hopkins and Cole, 1901, 2] protein and concentrated sulphuric acid. In
the previous experience of the writer at Cambridge and elsewhere, the test
had never failed, and a high estimate of its reliability had been formed, so
that this case, when nothing but a yellow to a yellow-brown ring was
obtained, needed investigation.

It was soon found that the failure was due to an accidental admixture of
the bench sulphuric acid with nitric acid, but during, and arising out of, the
search for the cause, the phenomena recorded below were observed.

It became obvious that three disturbing factors were involved: (a) the
presence of small quantities of impurities in the sulphuric acid, (6) the
absence of traces of impurities, and (c) physical factors influencing the rise
and fall of temperature of the reacting fluids. These have been investigated

in turn.

(1) The presence of impurities in supraminimal amounts.

The cause of the trouble recorded above was the presence in the
sulphuric acid of ‘nitrie acid,

The test, as has been pointed out by Salkowski [1888] and Hopkins and
Cole [1901, 1], and confirmed by Rosenheim [1906], is spoilt by oxidising
agents. This occurs at small concentrations, especially if, as a criterion,
rather than the formation of a ring, the colour of the resultant mixture be

250 V. H. MOTTRAM

taken’, Thus a final concentration of 1 part by volume of hydrogen peroxide
(10 vols. comm.) in 50, of 1 part by weight of ferric chloride in 1000, of 1 part
by weight of potassium chlorate in 25,000, of 1 part by volume of concentrated
nitric acid in 25,000 or of 1 part by weight of sodium nitrite in 50,000 parts
by volume prevents the formation of a violet or purple coloration.

In all cases except one (nitric acid) a yellow ring is formed, which, on
shaking the tube gently from side to side, spreads throughout the mixture.
With nitric acid there appears a purple ring above the yellow ring at the
junction of the fluids. When the test tube is shaken the purple ring moves
up the test tube, is overtaken by the yellow coloration and is destroyed. In
no case is there charring.

For laboratory practice, therefore, if, instead of a lilac, mauve or violet
coloration, a yellow, a red, a red-purple or a brown ring is formed, it is
probably due to contamination of the sulphuric acid with oxidising agents.
Commercial sulphuric acid, to judge by the samples investigated, never has
enough oxidising agents present to prevent the reaction.

(2) The influence of oaidising agents in minimal amounts

In the course of a large number of experiments it has become clear that
Rosenheim’s statement that an oxidising agent is necessary to the reaction is
justified. The observations on which this is based are as follows (for
experimental details consult Appendix, § 2):

(a) Some very pure samples of sulphuric acid (supplied by Messrs Evans,
Lescher and Webb, Messrs Baird and Tatlock, Liverpool, and Messrs Towers
and Sons, Widnes'and Liverpool) give with freshly made or with old
“reduced oxalic,” whether strong or weak, or with Benedict’s magnesium-
reduced oxalic [1909] a lilac or lilac-mauve colour instead of the typical
violet (see Appendix, § 2, Exp. LVIII etc.). But if traces of oxidising agents
are added to the “reduced oxalic” the blue-violet or violet develops at once
(Appendix, § 2). The optimal concentration is, for hydrogen peroxide
(10 vols. comm.) circa 1 part by volume in 1000 of the resulting mixture, for
ferric chloride 1 part by weight in 20,000 and for sodium nitrite 1 part in
375,000.

(b) Subsequent addition of oxidising agents after the lilac or lilac-mauve
coloration with pure sulphuric has developed converts this colour to a blue-
violet of great density and beauty (Appendix, § 1, Exps. LII—LIV)?. This

1 For experimental details see Appendix, § 1, Exps. I, II, XXX, LII—LV.

2 The paradox of the greater stability of the colour when once formed towards oxidising
reagents is simply explained as the decrease in velocity of a chemical reaction concurrent with
fall of temperature.

a nr AT

V. TH. MOTTRAM 251

does not always take place if the reacting fluids have cooled. But in this
case, warming, with previous or subsequent addition of an oxidising agent,
yields the typical blue-violet (Appendix, § 3 and § 6). Continued heating
alters the test through a series of colours typically the result of the addition
of increasingly greater quantities of oxidising agents to the “ reduced oxalic ”
(Appendix. Compare § 3, with § 2 Exp. LIX and§ 4 Exp. LXIV). Obviously
change in colour is a sign of advancing oxidation,

(ce) Commercial sulphuric acid will withstand the addition of less
oxidising agent than will pure sulphuric acid if the test is to be typical.
The reaction is not improved but spoilt by the addition of even small traces
(Appendix, § 4, Exp. XXXVI, etc.).

(d) Commercial sulphuric acid will withstand the addition of more
formaldehyde than will pure sulphuric acid (Appendix, § 4, Exp. XLI and
XLII).

(e) If the violet fluid resultant from the use of commercial sulphuric
acid be boiled, the colour changes seen in the parallel example quoted in
section (b) above are imitated with exactitude (Appendix, § 4, Exp. LXIV),
Oxidation by means of the oxidising agents present in the commercial
sulphuric acid is proceeding.

These five observations accord well with Rosenheim’s theory that oxidising
agents are necessary to the reaction. They suggest that a coloured substance
arises from the oxidation of a tryptophane-formaldehyde product formed from
the protein and that further oxidation of this blue-violet substance modifies
this colour and ultimately destroys it. We may assume that this reaction
proceeds at a velocity very much greater at high temperatures than at lower
temperatures and that, in the optimal conditions of the test, the cooling is
rapid enough to prevent the oxidation going too far.

That there is a change proceeding at a definite rate may be seen by
carrying out the test under conditions in which a high temperature is not
attained, or if attained, is rapidly reduced. For instance the initial colour
may be profoundly modified in a very wide tube, or by cooling the test tube
under the tap. A rose colour at first develops, but in the first case after the
test tube has stood some time the colour has deepened to mauve and after
a day to blue-violet. It is doubtful if this be due to the oxygen of the air,
for no observable difference was found when the test was carried out under
an atmosphere of carbon dioxide.

To the question whether a balance of formaldehyde and oxidising agents
is necessary [see Rosenheim, 1906] (another case of the two classical
antagonistic principles) my experiments give no certain answer. If these be

Bioch, vir hy

252 V. H. MOTTRAM

both in the “reduced oxalic” and in optimal equipoise, then reduction of the
strength of the “reduced oxalic” should render the equipoise more susceptible
to tilting agencies. Oxidisers should have more effect in the presence of
weak “reduced oxalic” than of strong. There is no evidence to offer that

this is the case (App. § 5, Summary of Exps. XVII to XXXYV).

(3) The influence of physical conditions.

It was thought that physical conditions, such as the size and shape of the
test tube, area of reacting fluids compared with their volume, and rapidity
with which the fluids reached their maximal temperature and subsequently
cooled would affect the results. This is indeed so but, except in the extreme
cases given above, is not of much importance. Thick test tubes give a lighter
result than thin; very narrow than normal size; and very large tubes with
inadequate amounts of fluid than normal tubes or large tubes with adequate
amounts. Cooling under the tap produces a rose colour shot with green
which subsequently, unless oxidising agents are added, does not darken much.
In three days it is lilac in colour (see App. § 6).

One reason for the different colours observed when different volumes of
the reagents are used is to be found in the dichromatism of the coloured
substance. It is lilac in thin layers or great dilutions, and bluer in thicker
layers and greater concentrations. This, of course, has its parallel in its
absorption band in the spectrum. A mixture that was lilac showed an
absorption band from » 592 to » 547, on adding a drop of ferric chloride the
band stretched from » 646 to X 511, diluting the mixture with pure sulphuric
acid brought the band to its original dimensions and the mixture to its
original colour. On warming the diluted mixture, with consequent alteration
in colour to purple, the band stretched from » 610 to A 531, and the blue and
violet were reduced in intensity. This band is identical with one described
by Hopkins and Cole [1901, 1].

Finally, the different colours observed with weaker solutions of protein are
due to the oxidation of the chromogen beyond the optimum. Weaker
solutions give with either pure or commercial sulphuric acid results that are
redder in tint than the controls. This is not dichromatism, for dilution with
sulphuric acid leaves the control many degrees bluer than the others. The
relation of this phenomenon to the others given above is clear if we assume
progressive oxidation.

V. H. MOTTRAM 253

SUMMARY AND CONCLUSIONS,

(1) The modification of the Adamkiewicz reaction introduced by Hopkins
and Cole fails in the presence of small amounts of oxidising agents.

(2) But traces of these improve the reaction when carried out with “pure”
sulphuric acid.

(3) Probably, therefore, the result with “commercial” sulphuric acid is
due to the presence of contained oxidising agents.

(4) Whether the trace of oxidising agent which gives the lilac to mauve
tint with “pure” sulphuric acid is in the reduced oxalic or in the sulphuric
acid cannot be settled. In view of Miss Homer’s work [1911] it is probably
in the sulphuric acid.

(5) The resultant coloration of the mixture depends on the extent of a
combination of a chromogen with oxygen. ‘This combination has a definite
rate of reaction which is naturally more rapid at high temperature and at
greater concentrations of the oxidising agent.

(6) It is therefore best to keep the volumes and percentage composi-
tions of the reacting fluids as constant in proportion as possible and their
absolute measure should depend on the size of the test-tube’.

(7) ‘The following method of carrying out the test yields excellent results:
1 cc. of strong “reduced oxalic” acid (Benedict’s “reduced oxalic” will do just
as well) is mixed with 1 cc. of 1°/, Witte’s peptone, and then 2 c.c. of
sulphuric acid are run down the side of the test-tube which is held as near
horizontal as convenient. The fluids are then rapidly mixed by shaking from
side to side with the test-tube vertical. The quantities given are for a test-
tube 1°5 cm. in diameter (the usual laboratory size). Subsequent addition
of one drop of 1 °/, FeCl, solution usually increases the density and the blueness
of the coloration.

The quantities given may be varied greatly and yet a good colora-
tion be obtained. 1 cc. or 5 cc. sulphuric acid to 2 cc. of the mixture yield
an unmistakable result and the quantities may be guessed and not measured.
More sulphuric acid gives a bluer result than less.

1 I can see no reason for abandoning the excellent reagent known as ‘‘reduced oxalic” acid in
favour of formaldehyde and doctored sulphuric acid. As long as the test is performed under
conditions that lead to a maximal development of heat, all samples of sulphuric acid give a
reasonable coloration. It is only when the reacting fluids are cooled or contaminated that an
imperfect result is obtained; whereas the balance of oxidising agent and formaldehyde is easily
tilted in an adverse direction, especially in the hands of students.

“Reduced oxalic” acid is a reagent easily made, particularly now that magnesium powder is a

requisite in the biochemical laboratory. It does not diminish in strength even when more than
a year old.

17—2

254 V. H. MOTTRAM

‘The details of the experiments on which the statements above made are

based are relegated to an appendix.

I have to thank Messrs Hopkins and Cole for reading through the MS. of
this paper and for suggestions they have made. Their explanation of the
phenomena recorded must be left to a future communication from them.

APPENDIX.
1. Haperiments on the destructive effect of oaidising agents.

In most of the experiments performed the production of a coloration was
taken as a criterion and not the formation of a ring. In the experiments
based on the plan of Exp. XXX (v. inf.)—the majority—all the test tubes
were prepared before the shaking of the contents took place. This naturally
does not give a maximal rise of temperature. The first test tube has to wait
a minute or so before the fluids it contains are mixed and the heat evolved
by the natural admixture at the junction of the fluids is to some extent lost.
This exaggerates the influence of oxidising agents as adjuvants of the reaction,
and decreases the influence as deterrents.

Exp. I. Equal parts stock solution of “reduced oxalic” undiluted and (circa) 1°/) acid
albumin mixed.

Mixture Pure sulphuric Oxidiser Remarks
25 ¢.c. 25 Gc. 1 drop cone, HNO; Yellow colour, faint effervescence.
25 ¢.c. 25 ¢.c. Small amt. FeCl, Normal violet, disappeared on adding
; excess FeCl...
25 ©.c. 25 ©... 1 c.c. 10 vols. HO, Light yellow coloration.

Exp. Il. Same mixture as above.

Mixture. Comm. sulph. Oxidiser Remarks
2 c.c. 2 c.c. Nil Violet ring.
2 €.¢. 2 €.¢. 2 drops 0:01 °/, HNO; Yellow ring.
PY (GG 2) ¢.c. dro pier. if Faint violet above, yellow ring below.
2 ¢.¢. 2iGsCs (1 drop water) Violet ring.
Exp. XXX. Mixture 1°/) peptone and strong ‘ reduced oxalic. ” Equal parts.
Mixture H,0z, (10 vols.) H,0 Puresu!phuric Remarks Coloration
2 c.¢. 0 drops 5 drops 74 (OO, Ring not clear. Lilac-mauve.
2 ¢.¢. 1 ae Cl Sar 2 c.c. Good ring. Clear violet.
2 ¢.¢. heer 3 5 2 c.¢. Good ring. Purple, then rose.
2 ¢.¢. Be ig an 2 ce: Vanishing ring. Light red.
A OSCE Ave I ee PaCACs Vanishing ring. Yellow.
2 c.¢. Dh ise CU) ee 2 c.¢. Vanishing ring. Straw colour.

Throughout the experiments based on this plan the same pipettes were
kept for each reagent. The drops were frequently estimated and_ their

V. H. MOTTRAM 255

average volume was 0°04 c.c. Of course in the measurement of the above
volumes, whether large or small, no great accuracy was obtained. A 1 c.c.
pipette has a high percentage error (circa 5 °/,).

>
Exp. LIT. NaNO,. Same mixture as in Exp. XXX. In the experiments following, not the
formation of a ring, but the final coloration is used as a criterion.
Mixture 0-1 °/) NaNO, H,0O Pure sulphuric Remarks
2 C.c. 0 drops 5 drops 2 ¢.c. Lilac coloration (deep blue-violet
on adding drop of NaNO).
2 ee. 1 ees r: ges 2 e.c. Dense purple coloration.
2 c.c. Pe hae 5 oie 20,0. Red coloration.
2C.C, De 55 Bish asi Orange coloration.
2°c.c. Ce LP. 2 ¢.c. Yellow coloration.
2 ¢c.¢c. OR Gtk. 2 c.c. Yellow coloration.
Exp. LIII. FeCls.
Mixture 4°/) FeCl; H,0O Pure sulphuric Remarks
2 c.c. 0 drops 5 drops 2 c.c¢. Lilac (dense blue on adding 1 drop
FeCl;).
2 c.c. i | ae Bn. PME Purple coloration.
2'C.c. Bites BY ae 2 C.c. Claret -
| 2 c.c. he dee ae 2 c.c. Red-brown ,,
2 c.¢c. Ne i 2'C.G: Brown Ae
2 c.c. 5 or) 0 o 2 c.c. ” ”
Exp. LIV. KCI10;.
Mixture 0:2 °/, KC1O, H,0 Pure sulphuric Remarks
2 ec. 0 drops 5 drops 21G.c: Lilac coloration (purple on addition
of 1 drop KCl10Os;).
2 c.c. aoe ATS 2 'C.C- Purple coloration.
2 c.c. Diana Bh. re 2 c.¢. Yellow 5
2'¢.c. Si as yin 2 c.c. a =
Exp. LV. HNOs.
HNO; 2 vols.
Mixture per 1000 H,0 Pure sulphuric Remarks
2 c.¢. 0 drops 5 drops 2 ¢.¢; Lilac (no deepening on adding
HNOs).
2 c.c. Te, eet 4D 5 2 c.c. Purple.
2 ¢.c. Di Bh a 2 c.e. Red-brown.
2 c.c. Bie ee PIE Yellow.
2 G.c. 4 ox; 1b ae 2 ce. Yellow.
2 ¢.c. Duss Ou 2 c.¢. Yellow.

In this experiment a purple ring was formed above the junction of fluids
with a yellow ring beneath. On gently shaking, the above colours were
produced.

A number of other experiments on similar lines were performed with
similar results.

256 V. H. MOTTRAM

2. Haperiments on the beneficial effect of traces of oaidising agents.
(See also Experiments immediately preceding.)

Exp. VIII. Peptone mixture.

10 9/, H,0, Pure
Mixture (10 vols.) H,O sulphuric Remarks
(a) (2 c.c. 0:0 ¢.c. 1 Okexc: PACA Ring obtained in all. On shaking,
(2 ee. 0:0 c.c. IOrerc: 2NGsCs (a) (i) and (ii) were the most
(b) (2 c.c. 0:5 c.c. 0°5 c.c. 2 €.¢. violet, (b) the deepest and (c) the
(2 ee. 0°5 c.c. 0°5 c.c. 2) 6.€. weakest. After two hours (b) was
(c) (Boe 1:0 c.e. 0:0 c.e. 2 ¢.c. nearly bleached and (c) yellow.
(2 c.e. 1:0 c.e. 0-0 c.e. 2 c.c.
Exp. IX
2) C.C. 0 drops 5 drops 2 G:C. Blue-violet coloration.
2 c.e. ibe Ares 2 Cc. Dense violet, showed up most
rapidly.
2 c.c. 2 5, Be 2 c.c. Purple.
2 c.c. Bee 2a 2 c.c. Purple.
2 c.c. Are. dr) ye 2 cc. Red-purple. :
2 C:c: ye One. 2 ¢.c. Red-purple.

A number of other experiments with different concentrations of H,O, gave
similar results. Exp. XXI is quoted. Strong reduced oxalic diluted five
times and then mixed with an equal volume of | °/, Witte’s peptone.
Exp. XXI.

ALOE Lai k0), Pure
Mixture (10 vols.) H,O sulphuric Remarks
2 cc: 0 drops 5 drops 2 ¢.c. Pale mauve.
2 c.c. sl ypere A Ny, 2 c.c. Violet-purple. Three times as dense.
2 ¢.¢. ar By Aas 2 c.¢. Purple. ,
YY) OG- Bho 5s Ble ee 2 c.c. Rose-purple.
2 ¢:c: 4s. Lee, BAG.Ce Purple-rose.
Dice Sls O-%; 2, G:C. s 95

Exp. LVIII. Mixture: 1 part freshly made reduced saturated oxalic and 1 part 1°/) Witte’s
peptone.

Mixture 0:5 °/, FeCls H,0 Pure sulphuric Remarks
2 ¢.c. 0 drops 5 drops 2 ¢.c. Lilac. (Deepened to royal blue on
subsequent addition of FeCls).
2 c.c. He eae Aree 2 c.c. Deep violet.
2 ¢.¢. Dae Bac 2 @.c. Clear violet.
2 ce. 3a De 2 c.c. Purple-violet.
2 ¢c.¢. A es Lees DiCie: Purple.
2 ¢.c. bye (var 2) (XC Red-purple.
Exp. LIX. Mixture as above.
Mixture 0°04°/, NaNO, H,0 Pure sulphuric Remarks
DyiciCs 0 drops 5 drops 2 ¢.¢. Lilac (deepening remarkably on
addition of NaNO, solution).
2 cc. 1 ss Aas. 2 @.c. Violet.
2¢e Dh. 8s Dione 2 Purple-violet (strongest at first,

bleaching later).

2 c.c BY ei Dee 2 ¢.c. Purple.
2 ce Boss Dee 2 cc. 3rown-purple.
2 ce Dee (Oieaes 2 c.e. Purple-brown.

Ss”

EE —————

V. H. MOTTRAM 957

Very many other experiments were performed, some of which are quoted
in other sections, with the same results.

3. Kaperiments on the subsequent addition of oxidising agents.

As will be seen in a number of the above experiments, addition of
oxidising agents after the lilac coloration was formed produced a deep blue
coloration. In some earlier experiments on cooling the fluids while the
mixture was being made, it was noticed that the cold mixture of sulphuric
acid reduced oxalic and peptone did not turn blue on the addition of
ferric chloride, but after the addition warming gave the intense blue
coloration. This on standing turned violet. Other tests showed that
rewarming the mixture and then adding ferric chloride produced the
same result, and that further warming, or rather boiling, made the colour
proceed through violet, purple, to red and ultimately bleached it almost
completely. Further notes will be found on similar results obtained in the
experiments on the effect of cooling the test tube while the sulphuric and
the other reagents are mixing (App. § 6).

4. Comparison between pure and commercial sulphuric acid.

Exps. XLI and XLII.

0-04 °/, form- 2 e.c, Commercial
Mixture aldehyde H,O sulphuric 2 ¢c.c. Pure sulphuric
2 @.¢. 0 drops 5 drops Violet Lilac-mauve (gives blue ring
on running in H,0,).
2 c.c. iY es. ay os Dense violet Olive green.
2 c.¢. Das Bh ee Brown-violet - ‘3
2 ¢.¢c. Bye Be is Brown-purple ss 33
2 ¢.c. A Ss 1 ar Purple-brown Es ne
2 ¢.¢. DP, O-.,, Light purple-brown = -

Exp. LXIV. Experiments on same plan as above made with different samples of acid.

= & T. commercial sulphuric haart Genta) paceeditttouch queple
. rz) 35 urple-violet é : at

me ow. a i Purple-violet | to purple-red and brown.

ite pure *s Lilae ) On heating slowly altered to purple-
Be ss We a 5 Bluish-lilae } brown obtained in XLI, 6.

Exp. LXI. When various samples of acids, commercial and pure, were used, but to 2 ¢.c. of
mixture and 4 drops of H,O in each case 1 drop of 0-°04°/, NaNO, was added, the following
results were obtained :

Sulphuric Remarks
BG. pure “2 cic: Blue-violet.
Bend ele COMM ssa 5 Purple.

T. pure, = Blue-violet.
T. comm., o- Purple.
E. L. W. pure, __,, Blue-violet.
E. L. W. comm. ,,— Purple.

258 V. H. MOTTRAM

Exps. XXXVI to XL and LX with varying strengths oxidising agents added showed that the
commercial sulphuric used had the optimal amount of oxidising agents present. Compare with
tests in §§ 2 and 3. As an example XXXVIII is quoted.

Commercial
Mixture 2°/, 10 vols. HO, H,O sulphate Coloration
2 ¢.c. 0 drops 5 drops 2 ¢.¢. Violet (densest).
2 ¢.c. Ds ZW aed 2 ¢.¢c. Purple-violet.
2 cc. Qhahe B55 2 c.¢. Purple.
PL OXG: BLES leet 2) (CG Red-purple.
2 cc. ANN 1 Ea ee, 2 c.c. Purple-red.
2 €.¢. Dress (soe 2 C.G. Red-brown.

5. Experiments on the supposed antagonism between formaldehyde and
oxidising agents and on the occurrence of the latter in reduced oxalic acid.

“weak” Hopkins’ reagent = reduced saturated oxalic diluted 5 times.

‘ ”
; normal > Py) ? ” 39 2) 24 ”?
“strong” = s & undiluted.

” > ”

The experiments were carried out as usual, six tests in each, with
increasing strengths and amounts of hydrogen peroxide; thus:

Exp. XXXII. Mixture of equal parts of strong Hopkins’ reagent and 1 °/, peptone.

Mixture 2°/, 10 vols. H,O, H,O H,S0, Coloration
(a) 2 €.€. 0 drops= 5 drops PN ONG: Lilac-mauve.
(b) 2 c.c. ikaar A 2 cc. Violet-mauve.
(c) Diccus haa 35 es 2 ¢.c. Violet | a

t.

(d) 2 c.c. ies, Deas 2 c.c. Violephes 2°
(e) 2 €.¢. A 55 1. PiCiCs Purple-violet.
(f) YONG: Sy 5; Ont 2 C:c: Purple.

Summarising the results it was found:

Densest colour

I et

10 vols. H,0, ' Strong reagent Normal reagent Weak reagent .
10 °/, (b) (violet) (b) (b)
4 (b) & (c) (purple-violet) (2) (0)
7a (c) & (d) (violet) (d) & (e) (c)
1°3 J, (a) (violet) (d) (e)
Of, (b) to (ad) (much the — (f)

same colours)

No evidence is seen of oxidising agents deflecting the optimum towards
nil concentration when weak Hopkins’ reagent is used. This should be the
case if oxidising agents are in the reduced oxalic.

————— rc ite e.hUhUCUCCOUC TT

i

V. H. MOTTRAM 259

6. Eaperiments on effect of preventing, or of accelerating,

rise to maximal temperature.

Exp. LXV.
Mixture Pure sulphuric Procedure Coloration
2 ¢.¢. 2 cc. Shaken under cold tap Pale rose with green fluorescence.
2 c.c. 2 c.¢. oe ys» Warm tap Rose.
(45° C. circ.)
2 cc. 2 c.¢. Shaken in air Lilac-mauve.
2 cc. 2 c.c. AC ,, Bunsen flame Violet-mauve, changing to brown-
violet.
Repeated first experiment left to stand... ... Mauve after three days.
» ” 5 added 1 drop 4°/, FeCl, Pale rose slowly alters to deep rose, lilac-
mauve and on gentle warming to violet.
= is “ very thoroughly cooled, Slowly darkening in about an hour to lilac-
added 1 drop 4°/, FeCl, mauve.
7. Eaperiment on dilution of peptone.
Strong Hopkins’
Peptone reagent Sulphuric Coloration.
(a) exe (1 °,) Ges 2 ¢.c. (pure) Lilac-mauve.
(b) ICsCs 55 1RCse: 2 ¢.c. (comm. ) Violet-purple.
(c) 1 ee. (0:2'°/,) HCE 2 ¢.c. (pure) Pale lilac-violet.
(d) ices!) iL Ges 2 ¢.c. (comm. ) Pale purple.
(e) = Lee. (071 %,) Wicxe 2 c.c. (pure) Very pale lilac-purple.
(f) Gon h ieze: 2 c.c. (comm.) Light rose-purple.

When (a) and (6) are diluted with pure sulphuric the result is bluer than
any parallel experiment with weaker peptone. Addition of commercial
sulphuric acid to (6) produced a redder colour than addition of pure sulphuric
acid to (b). These observations accord with the theory of progressive
oxidation of a chromiogenic substance by oxidising agents in the sulphuric
acid.

REFERENCES.

Benedict (1909), J. Biol. Chem. 6, 51.

Homer (1911), Proc. Camb. Phil. Soc. 16, 405.

Hopkins and Cole (1901, 1), Proc. Roy. Soc. 68, 21.
—- (1901, 2), Journ. Phys. 27, 418.

Rosenheim (1906), Biochem. J. 1, 233.

Salkowski (1888), Zeitsch. physiol. Chem. 12, 211.

XXVI. THE USE OF LITMUS PAPER AS
A QUANTITATIVE INDICATOR OF REACTION.

By GEORGE STANLEY WALPOLE.
From the Wellcome Physiological Research Laboratories, Herne Hill, SE.

(Received April 7th, TOTS.)

In dealing with nearly neutral solutions of feebly dissociated electrolytes,
especially when they are highly coloured, it is frequently convenient to gauge
reaction by the use of litmus paper, and so avoid either electric measurement
or the introduction of a fluid indicator into a comparatively large volume of
fluid. And it has been found that if suitable precautions are taken quite
reliable information can be obtained as to the H’ ion concentration of such
fluids in this way. The apparently contradictory results obtained in such
cases by using different commercial samples of litmus paper, immersing the
paper for a short or a long time in the fluid, using one drop on a piece of
paper or a small piece of paper in a considerable volume of fluid, and so on,
are capable of simple interpretation.

The precautions to be observed are all those which apply to the use of
litmus solution with the addition of those necessitated by the fact that it is
applied to paper. As an example of the former class the “neutral salt
effect ” may be mentioned: the “time factor” is typicak of the latter.

EXPERIMENTAL.

The following solutions were used :

Ammonia-ammonium chloride mixtures. A solution of ammonium
chloride and ammonia was prepared so as to be approximately 3 N with
respect to the former constituent and 3/32 .N with respect to the latter.
Dilutions with boiled distilled water were made x2, x4, x16, x32. According
to Fels [1904] these solutions should have the same Pj; value at 18°C. The
values obtained on examination in the hydrogen cell using a calomel
electrode were 814, 814, 813, 814 respectively, while 8:12 was that
obtained for the undiluted material.

——<—~. es

~~ a

G. 8S. WALPOLE 261

Acetic acid-sodium acetate mixtures. A solution of commercial sodium
acetate 0°61 N, containing acetic acid corresponding to 0°61/64 N was diluted
x1, x2, x8 and x16. These solutions should have the same Py, value very
closely. The values found were 6°25, 6°20, 6°27, 6°23.

Neutral phosphate solution. This was a mixture of 68°8 c.c. of Sérensen’s
N/15 Na,HPO, solution, and 131:2 cc. of his N/15 KH,PO, solution. The
Py value determined was 7:05. The figure at present accepted for absolute
neutrality is Pj, = 7:07 [Sirensen, 1909].

Solutions of serum globulin and Witte peptone were also examined, but
they presented, in addition to the phenomena observed with solutions free
from proteins and their primary disintegration products, others of a more com-
plex and different order. For this reason it was decided to deal with them more
fully in a separate communication and cite here only one or two typical cases.

Litmus solution. The reaction between litmus solution and the solution
with which it is mixed is practically instantaneous. In saline solutions it
gives an indication of the reaction of the solution subject only to a correction
for the “neutral salt effect”: for it is well known that, for solutions of the
same H’ ion concentration, indicators in the presence of larger quantities of
neutral salt give slightly different colours from those which they give when
little salt is present [Sérensen, 1912].

The litmus tincture used was itself apparently neutral, or very nearly so,
since it gave almost exactly the same tint when added in the same quantity
to equal volumes of boiled distilled water and neutral phosphate solution.

The ammonia-ammonium chloride mixtures all gave precisely the same
tint as far as could be observed with litmus solution, with the exception of
the x32 dilution, 100 c.c. of which required 0:02 c.c. N/10 NaOH to bring the
colour to the same tint as that of the others. The acetic acid-sodium acetate
mixtures, on the other hand, gave progressively pinker solutions with litmus
tincture as dilution increased. From this observation I think it is legitimate
to draw the conclusion that the neutral salt effect is observable in all the
acetic acid-acetate mixtures, and only in one case in the ammonia-ammonium

chloride mixtures. This is probably due to the fact that the P;, value of the
latter mixture is near the alkaline limit of the sensitive range of litmus.

“ Equilibrium tint.” If a small piece of litmus paper be introduced into
a comparatively large volume of any of these solutions it will, after a lapse of
a longer or shorter time, assume a definite tint which undergoes no further
change. This it is proposed to call the “equilibrium tint” of the paper used
in the solution under consideration.

262 G. S. WALPOLE

“ Reaction inertia.” A solution whose H’ ion concentration is but little
affected by the addition of H’ or OH’ ions, say in the form of hydrochloric
acid or caustic soda, can be said to have a greater “reaction inertia” than
pure water or a dilute solution of a completely dissociated electrolyte, whose
H' ion concentration would be profoundly affected by such an addition.

The above acetic acid-sodium acetate mixtures, although of the same
absolute reaction (Pj), = 6'2), differ very much in reaction inertia, as is in fact
patent from their composition. The number of c.c. of N/10 NaOH solution re-
quired to bring 100 cc. of any one solution from Pi, =62 to Py, =8'1 may be for
the purpose of this note considered as a numerical expression of this property.
A 50 cc. sample of each dilution being taken and equal quantities of litmus

tincture added to each they presented different colours but had Py = 6-2 in
each case as described above. A 50 cc. sample of an ammonia-ammonium
chloride mixture to which an equal quantity of litmus tincture was added
furnished a standard tint Py =8'l. The acetic acid-acetate samples were
then titrated with N/10 NaOH to this standard. The results are tabulated :

Acetic acid-sodium acetate mature.

Dilution Reaction inertia
1 ; 8
Ws 4
8 1
16 0°5

In a similar manner titrations of the ammonia-ammonium chloride

= its ) eis °
mixtures were undertaken from P},;=8'l to Py =62 using N/10 HCl, and
from these results a similar table may be prepared. The standard pink tint

(P}, = 62) was that of the undiluted acetic acid-acetate mixture.

Ammonia-ammonium chloride mixture.

Dilution Reaction inertia

1 90

—
lor aS)
we

ow OO bo oO

With N/15 phosphate mixtures it was found that 20 ¢.c. ofa solution Py=8'1
required 0°9 cc. N HCl to bring the reaction to Py, =62. The “reaction
inertia” for N/15 phosphate solution may then be taken as 0°9 x 5 x 10 = 45.

G. 5. WALPOLE 263

OBSERVATIONS WITH LITMUS PAPER.

Four commercial varieties of litmus paper were examined, each providing
a “red” and a “blue” paper. Three brands were glazed with the litmus
applied to one side only, and one soft and absorbent.

Of the three that were glazed, No. 1 was dyed heavily and apparently
loaded with a material of considerable “ reaction inertia”; No. 2 was not so
heavily dyed though the grain of the paper was similar; No. 3 was faintly
dyed on a perfectly smooth paper, presenting a homogeneous appearance.
The unglazed paper was labelled No. 4.

Colour tint of paper after 24 hours immersion. Small pieces of each
paper about 25 mm.? area were immersed for 24 hours in 240 ¢.c. of ammonia-
ammonium chloride solution of each dilution. At the end of that time all
the pieces of paper were of the same tint.

Only one difference could be observed. The heavily loaded paper, No. 1
(blue) was more deeply dyed in the more concentrated solutions than in
those more dilute. The adsorption equilibrium seems to shift in the
direction of greater adsorption of dye with increasing concentration of salt at
constant reaction. A special experiment ratified this conclusion. Pieces of
paper No. 1 blue after prolonged immersion in solutions Pj, = 8'1 containing
15, 7:5, 3°8, 1:9, 0°5 per cent. NaCl gave up progressively more dye to the
solution as dilution increased, though in similar solutions of Pj, = 61 no such
change could be definitely observed.

Paper No. 1 red immersed in N/15 phosphate solutions Pj} = 45, 7°1, 9:2
respectively gave up very markedly increased quantities of litmus to the
solution with increasing alkalinity. In parallel experiments the litmus found
in the alkaline solution was 5-6 times that found in the acid solution.

This, in conjunction with many similar observations in this connection,
has led me to the conclusion that in similar solutions the adsorption
equilibrium changes in the direction of diminished adsorption of litmus by
paper in solutions of increasing alkalinity.

In the acetic acid-sodium acetate solutions similarly examined small
pieces of each paper gave the same tint after 24 hours immersion, only in
one and the same dilution. Passing from one dilution to another it was
found that with increasing dilution just that change could be noticed in the
equilibrium tint which, as far as could be judged corresponded to the variation
in tint with htmus tincture due to the “neutral salt effect” described above.

Observation of rate of change of colour. In the above experiments the

264 G. 5. WALPOLE

time taken for a piece of litmus paper to acquire its equilibrium tint was
seen to vary very considerably. The same paper changed more rapidly in
a solution of greater reaction inertia than in one of less reaction inertia in all
cases, and, as one would expect, a paper took longer to reach its equilibrium
tint when the colour change from its original to its final tint was great. For
instance a red paper (No. 1 red) took several hours to come to its equilibrium
tint in the most dilute alkaline solution, while a blue paper did so in a few
minutes as it was already almost at that tint.

Observation of effect of exposure to the ar. It will have been noticed that
the two main test mixtures chosen each have a volatile constituent. Pieces
of paper that have been immersed for sufficient length of time to assume the
equilibrium tint, change that tint rapidly on exposure to the air in such
a case. The reddening of blue paper due to the volatilisation of ammonia is
naturally more rapid than the blueing of red paper due to acetic acid
evaporation.

“ Amphoteric solutions.” Dealing with approximately neutral solutions
a distinction is frequently made between a solution that is “ merely neutral ”
and one that is “amphoteric.” Milk, urine, and solutions of phosphates are
regarded as typical examples of the latter class and they are said to “turn
blue litmus paper red, and red litmus paper blue.” What is actually seen
may be better described as the “reddening of blue litmus paper and the
blueing of red litmus paper.” ‘The “merely neutral” solution is one of low
reaction inertia. A drop placed on paper wets it but only to a very small
extent alters its reaction. The colour of the paper therefore remains
unchanged. The “amphoteric” solution, on the other hand, has a high
reaction inertia. It usually contains a salt of a feebly dissociated acid or
base and, being neutral or nearly so, alters the reaction of litmus paper with
which it is brought into contact in the direction of neutrality.

It would naturally be expected that small pieces of paper, red and blue,
would, when immersed in a large volume of an amphoteric solution,
ultimately assume the same tint. During the first few hours of observation
it is seen that the blue paper is becoming progressively redder, and the red
paper bluer, but before equilibrium can be attained certain disturbing
influences may be encountered. For instance if protein be present the
colour of the paper will be interfered with. And, again, since the range of
reaction through which the litmus changes is different in the two cases, the
losses of litmus to the solution will not be the same. A blue and a red
sample of the quickly reacting paper No. 4 were seen to assume very nearly
the same tint in a large volume of milk in 24 hours.

— oe

mney

G. S. WALPOLE 265

Preliminary observations in protein solutions. In solutions of protein,
peptone and the like the use of litmus paper has, as far as my experience
goes, always indicated the solution to be more alkaline than it really is. For
instance an acid 4 per cent. solution of Witte peptone containing | per cent.
of sodium chloride, and 1 ¢.c. per cent. of normal hydrochloric acid was found
by electric measurement to have P,=65. In a tintometer arranged to
compensate for the colour of the Witte peptone [ Walpole, 1910] and using
neutral red as indicator, the Pj; value was made out to be the same as that of
a phosphate mixture P=6°8'. A neutral N/15 phosphate solution was
prepared (Pj, = 7:07) and a comparison instituted between this solution and
the acid Witte peptone solution above. It will have been noticed that
neither contains a volatile constituent.

Three papers were taken (No. 4 red, No. 2 blue and No. 3 red) and one
slip of each paper dropped into a large volume of each fluid. They were
removed at intervals, laid on white paper and examined in diffused daylight.

The unglazed, quickly-reacting paper (No. 4 red) gave with the acid
peptone solution a bluer tinge almost at once than the corresponding paper
in the phosphate solution and this relation continued indefinitely. The blue
paper No. 2 was always bluer in the acid peptone solution: the red paper
No. 3 lost most of its colour in the phosphate solution before the progressively
bluer paper in the acid Witte peptone solution could match it.

In another case a faintly alkaline 4 per cent. solution of Witte peptone of
reaction Pij;=7°3 gave with papers No. 2 (red and blue), No. 3 (red and
blue), No. 4 (red and blue) the same equilibrium tint, as nearly as could be
judged, as any of the ammonia-ammonium chloride dilutions examined,

which had P},=8'1 in each case.

EXAMPLES.

By neglecting to consider these secondary influences it is possible to
obtain some bewildering results. For instance we may take two alkaline

solutions having the same reaction, Py, = 81; A is a solution of ammonium
chloride and ammonia, 3N with respect to the former and 3/32 N with
respect to the latter; B is this solution diluted 32 times and the “reaction
inertia” of A is about 30 times that of B; C is a neutral phosphate solution ;

1 In the same apparatus using litmus tincture as indicator no real match could be obtained
but the indications were that the peptone solution was more alkaline than a phosphate mixture
Pao.

H

266 G. S. WALPOLE

and D is pure water. Red litmus paper No. 1 dipped into the solutions,
removed, and examined shows at once a bright blue colour with A, red with
B, a bluish tinge with C and red with D. Blue paper No. 4 shows blue
rapidly turning red with A, blue slowly turning not quite so red with B, red
with C, and blue with D; while blue paper No. 2 shows with A and B blue
very slowly turning red, and with C and D slight change towards red, less
marked in the latter case.

The following interpretations suggest themselves. With red paper No. 1
the concentrated solution A neutralises the acid red material of the heavily
loaded paper, and in spite of the glazing shows a pronounced blue colour at
once; with B the reaction of the absorbed fluid is determined more by the
paper than by the substances originally dissolved in the water, and the
evaporation of the free ammonia has been almost complete long before any
great effect would have been observed even if the paper had been bathed in
an excess of solution. Solution C gives the neutral tint appearing blue on a
red paper, while distilled water leaves the paper unaffected.

With blue paper No. 4 the attainment of equilibrium is very much more
rapid because the paper is not glazed, and the rate of evaporation of ammonia
is also accelerated though to nothing like the same extent. Blue paper No. 2
has a smaller “reaction inertia” than No. 1, besides it is already blue. It
differs from No. 4 in that it is glazed. The instances cited above are
exaggerated and, in the main, their interpretation obvious; but these factors
operating in a minor degree may lead to incorrect conclusions.

SUMMARY.

Except only in so far as the colour is influenced by the presence of
neutral salts, the reaction of a solution is indicated by the colour of a piece
of litmus paper which has been immersed in an excess of the fluid until no
further change can be observed. ‘The correction for neutral salt effect 1s
a small one and is of the same order as that applying to litmus solution when
used in the same way.

When the quantity of solution used is limited in amount, as when a drop
of solution is placed on the paper and the effect observed, the following
secondary influences determine the colour changes seen—

(1) Gradual attainment of reaction equilibrium between the litmus on
the paper and the solution ; more rapid with unglazed paper than with paper
that is glazed.

(2) The actual effect of the reaction of the material in the paper on the

G. S. WALPOLE 267

limited amount of solution used. This will, with the same paper, be most
marked with a solution of lowest “reaction inertia.” It is obviously fallacious
to test the reaction of tap water by watching the effect of one drop of water
on a heavily dyed paper.

(3) The “reaction inertia” of the paper. <A thick heavily loaded paper
has more effect in modifying the reaction of a drop or two of a solution than
a thin paper lightly dyed with a litmus solution of low reaction inertia.

(4) The effect of exposure to the air. The evaporation or oxidation of
some constituent of the solution affecting its reaction, will naturally produce
a corresponding effect.

(5) The liberation of indicator from the paper, which is greater in the
absence of neutral electrolytes and is more marked in alkaline than in acid
solutions. :

Determination of reaction of nearly neutral solutions of proteins and their
decomposition products by means of litmus paper is very difficult. Such cases
have not been dealt with fully here, but will form the subject of a later
communication. The general effect is to indicate that the solution is more
alkaline than it really is.

The same phenomena may be observed with litmus tincture to a much
more marked degree. For this reason it has been frequently described as
useless. in such cases.

REFERENCES.

Fels (1904), Zeitsch. Electrochem. 10, 208.
Sdrensen (1909), Biochem. Zeitsch. 21, 131.
(1912), Ergebnisse Physiol. 12, 441.
Walpole (1910), J. Physiol. 40, 27.

Bioch. vit 18

XXVII. THE PREPARATION FROM ANIMAL
TISSUES OF A SUBSTANCE WHICH CURES
POLYNEURITIS IN BIRDS INDUCED BY
DIETS OF POLISHED RICE.

PART f.
By EVELYN ASHLEY COOPER, Beit Memorial Fellow.
From the Lister Institute of Preventive Medicine.

(Received April Sth, 1913.)

Although several workers have isolated from various food-stuffs substances
capable in small amount of curing polyneuritis in pigeons and have ascribed
chemical formulae to them, attempts to prepare the active constituent in an
amount sufficient for investigating its chemical constitution and properties
have up to the present time been unsuccessful.

Funk [1911, 1913] has isolated from rice polishings, yeast, milk, and
ox-brain substances curative in oral doses of 0:02 g. and possessing similar
melting-points,

Edie, Evans, Moore, Simpson, and Webster [1912] have also prepared a
substance from yeast curative in small amount, to which they have ascribed
the formula C,H,,O;N., and Suzuki, Shimamura, and Odaki [1912] isolated
an active substance containing nitrogen from rice-polishings in the form of a
picrate, but they have not published any results of its analysis.

1. ISOLATION OF AN ANTI-NEURITIC SUBSTANCE FROM HORSE-FLESH.

In the present communication a method is described for preparing from
horse-flesh a substance capable in minute amount of curing polyneuritis in
pigeons.

The method employed was a modification of that described by Maclean
[1912, 1] for the isolation and purification of the phosphatides of the kidney.
The horse-flesh was minced, dried at 30° by means of an electric fan and
ground, and the dry powder (weight 4000 g.) extracted at 37° with absolute
alcohol on a shaking machine. The filtered extract was evaporated at 40° in

KE. A. COOPER 269

vacuo to remove the alcohol. The extract (weight 500 g.) was found to
possess marked curative properties towards pigeons affected with polyneuritis,
4 g. being sufficient to bring about complete recovery within 24 hours. The
extract was then treated with an excess of ether and the mixture allowed to
stand in the cold room for 12 hours. he ether dissolved the fats and lipoids,
but left undissolved a considerable amount of a white substance. This was
washed thoroughly with ether and then tested on neuritic pigeons. Doses of
0°3 g. were sufficient to bring about complete recovery within 12 hours. This
procedure therefore afforded a simple technique for separating a highly
curative fraction from fats and lipoids and it was much more convenient than
extracting the active substance by means of water.

The ether-soluble fraction was freed from ether by evaporating in vacuo. As
much as 12 g. of the residue possessed only slight curative properties, while 2 g.
of the lipoids separated from this fraction by means of acetone were inactive,
but 7 g. were curative within 24 hours. It was thus possible to separate
the bulk of the anti-neuritic substance from the alcoholic extract of horse-
flesh by means of ether. These results support the explanation advanced by
Maclean [1912, 2] for the curative properties of lecithin observed by several
workers, namely, that the anti-neuritic substance is not a lipoid, but is present
in ordinary lecithin as an impurity readily extractable therefrom by simple
methods.

The ether-insoluble fraction was next treated with absolute alcohol, by
which a large portion was dissolved. The insoluble residue contained both
inorganic and organic material, but possessed no curative properties. The
alcohol-soluble fraction however was strongly curative, and it was treated with
excess of ether. This caused the separation of a yellow syrup (yield 50 g.)
which was completely soluble in water and 0:2 g. of which was sufficient to
cure pigeons affected with polyneuritis. The aqueous solution of the syrup
was allowed to stand in a vacuum desiccator and a white crystalline substance
gradually separated. This proved to be carnine and possessed no curative
properties. The filtrate, which was highly curative, was next treated with
finely powdered lead acetate, until there was no more separation of a
flocculent precipitate, and the mixture was allowed to stand for 12 hours.
The precipitate was filtered off, washed with water at 35°, decomposed with
sulphuric acid, and the excess of acid removed by the careful addition of
baryta. The filtered solution possessed no curative properties. The filtrate
from the lead acetate precipitation was freed from lead by careful treatment
with dilute sulphuric acid and was left very slightly acid. It possessed
curative properties and was next treated with silver nitrate, which produced

18—2

270 E. A. COOPER

a copious yellowish-white precipitate. This was filtered off, decomposed by
hydrochloric acid and the resulting solution was nearly neutralised with soda
and was found to be curative. The filtrate from the silver nitrate precipita-
tion was also curative, but its content of active substance was completely
precipitated by silver nitrate when baryta was added. By carrying out
several animal experiments it was found that at least 3/5 of the total amount
of anti-neuritic substance present in the filtrate from the lead acetate precipi-
tation was precipitated by the addition of silver nitrate only, and about 1/4
was carried out of solution by the subsequent addition of baryta. The
remaining 3/20 was probably destroyed by the alkali The residue
obtained by the evaporation of the curative solution resulting from the
decomposition of the first silver nitrate precipitate with hydrochloric acid was
next extracted with chloroform. Only a small amount of substance was
extracted and this possessed no curative properties. The anti-neuritic
substance was therefore insoluble in chloroform. It was also found to be
insoluble in benzene and ethyl acetate. The residue insoluble in chloroform
was next extracted with absolute alcohol; a large amount was insoluble, but
the curative substance was found to have been entirely dissolved, and 0°10 g.
of the soluble fraction was sufficient to cure a pigeon affected with poly-
neuritis. This fraction was then extracted with acetone and a considerable
amount was dissolved. The insoluble residue, which contained all the
active substance, was dissolved in 50 °/, alcohol, and the sclution allowed to
stand in a desiccator for some days. A white substance gradually crystallised
out, 0:06 g. of which administered to each of two pigeons ameliorated the
symptoms of polyneuritis in a few hours and effected a complete recovery
within 24 hours. The substance dissolved with moderate readiness in water,
but was insoluble in absolute alcohol.

It was not possible to obtain a sufficient amount of the active substance
to determine its chemical composition, and the work is therefore being
repeated on a much larger scale. As, however, it was found [Cooper, 1913]
that cardiac muscle contained considerably more anti-neuritic substance than
voluntary muscle, ox-heart has been substituted for the horse-flesh and the
results so far obtained justify this change.

Some experiments have also been carried out to ascertain if the anti-
neuritic substance could be precipitated by means of ether from the fats and
lipoids derived from other animal-tissues. It was found that considerable
amounts of curative fractions could be obtained by this procedure not only
from horse-flesh and ox-heart, but also from horse-kidney, beef, ox-brain and
liver.

ss aatine lie ett mhhdeh tn cae ope mae eee a . ” a SLL RA RRAMIRETBREN ie en on. a

E. A. COOPER 271

Only a small amount of material, however, was separated by ether from
the alcoholic extract of egg-yolk, and its curative power was very feeble. The
results indicate that this fractionation affords quite a general method for
separating the anti-neuritic substance from the fats and lipoids of animal
tissues.

2. THE ABSORPTIVE CAPACITY OF ANIMAL CHARCOAL FOR THE
ANTI-NEURITIC SUBSTANCE.

Chamberlain and Vedder [1911, 1912] found that an extract of rice-
polishings containing the anti-neuritic substance after being filtered through
bone-black could no longer prevent polyneuritis in fowls and that the active
substance could only be extracted to a small extent from the charcoal by
water or alcohol. As these results are of great importance, the experiments
have been repeated. Powdered animal charcoal was extracted three times
with boiling water and then dried. A curative solution of known activity
(prepared by dissolving the ether-insoluble fraction derived -from an alcoholic
extract of horse-flesh in water) was filtered slowly through a bed of the
charcoal six times. The filtrate was still curative, but its activity was
reduced to the extent of 30 °/,. Separate portions of the charcoal were
then extracted with water and alcohol, and the extracts were found to be
highly curative. The anti-neuritic substance was thus partially absorbed by
charcoal and could be recovered from the latter by extraction with water or
alcohol. It is seen that the results do not agree entirely with those obtained
by Chamberlain and Vedder, but it is unlikely that the use of bone-black as
an absorptive agent will be of much value for the isolation of the anti-
neuritic substance.

3. THE EFFECT OF SULPHURETTED HYDROGEN UPON THE
ACTIVE SUBSTANCE.

In a preliminary fractionation of the horse-flesh extract sulphuretted
hydrogen was employed to decompose the lead and silver precipitates and to
remove the excess of the metals from solution, and during these processes a
rapid disappearance of the active substance occurred. By substituting acids
for the sulphuretted hydrogen (sulphuric acid for lead, hydrochloric acid for
silver) it was possible to conduct the operations with a much smaller loss of
active material.

These facts suggested that sulphuretted hydrogen destroyed the anti-
neuritic substance. Accordingly, the effect upon activity of passing this gas

272 EK. A. COOPER

through a curative solution for four hours was investigated, but no evidence
was obtained of any destruction of the anti-neuritic substance. The diminished
activity during the above manipulations would therefore appear to have been
due to the absorption of the curative substance by the colloidal metalhe
sulphides.

4. THe Errect oF ALKALI UPON THE ACTIVE SUBSTANCE.

Several investigators have observed that a considerable loss of anti-neuritic
substance occurs when chemical operations involving the use of alkali are
employed in the fractionations. It was thought desirable to throw some light
on this matter by quantitatively studying the effect of treatment with alkali
upon the curative power of the substance. ;

To a solution of known activity ammonia was added until its concentration
throughout the total volume of liquid reached 10 °/,. The mixture was kept
in a closed flask at ordinary temperatures for 24 hours, the ammonia then
removed by a current of air, and the curative power of the solution redeter-
mined. It was found that 50 °/, of the anti-neuritic substance had become
destroyed by contact with the alkali.

These results indicate that the isolation of the anti-neuritic substance in
large amount is only likely to be effected by employing chemical methods
that do not involve the use of alkali and sulphuretted hydrogen.

5. THe Errect oF ALKALOIDS UPON PIGEONS AFFECTED

WITH POLYNEURITIS.

Strychnine was found [Cooper, 1913] to prolong the lives of birds affected
with polyneuritis but to exert no curative action. It was accordingly of
interest to investigate the action of other alkaloids. Experiments have been
carried out with quinine, cinchonine, and cinchonidine. The substances were
dissolved in a trace of hydrochloric acid and administered to the birds
orally.

Doses of quinine ranging from 0:01 to 0:1 g. exerted a temporary curative
action. In six cases the symptoms were ameliorated for about three days and
then became once more acute, further administration of quinine having no
effect. In two other cases however the neuritic symptoms completely dis-
appeared within a few hours, and one of the birds remained free from poly-
neuritis for four days, whilst the other, receiving daily in addition to polished
rice 0'025 g. of quinine, remained healthy for 10 days and then died exhibiting

ee ee

S aula

AL

E. A. COOPER 273

symptoms merely of weakness and not of acute polyneuritis. When given in
amounts exceeding 0°1 g. however quinine had no ameliorative effect, but
actually appeared to hasten the fatal issue.

Daily doses of quinine ranging from 0°01 to 0°025 g. administered to
several pigeons fed on polished rice did not prevent or delay the development
of neuritic systems.

Cinchonine in doses varying from 0°05 to 0-1 g. like quinine effected a
marked improvement in the condition of pigeons suffering from polyneuritis,
but within 48 hours the birds again exhibited the acute neuritic symptoms
and died notwithstanding further doses of the alkaloid. Cinchonidine on the
other hand had no ameliorative effect at all.

As quinine and cinchonine are extracted from cinchona-bark by means of
acid, it was thought that their curative action might possibly be due to con-
tamination with traces of the anti-neuritic substance present in the plants.
The active substance is readily destroyed by heat, so that some light could
be thrown on this matter by observing whether the alkaloids retained their
curative properties after strong heating. It was actually found that after
being heated at 125° for six hours quinine no longer had any ameliorative
effect upon neuritic pigeons. As the alkaloid is not chemically altered by
such treatment, it would appear that its curative action is due to the presence
as an impurity of a minute amount of the anti-neuritic substance.

6. THe EFFECT OF ALCOHOL UPON BIRDS FED ON POLISHED RICE.

The administration of small doses of alcohol (0°5 c.c. three times daily) to
pigeons fed on polished rice had no measurable effect upon the period of time
elapsing before the occurrence of symptoms of polyneuritis, and thus appeared
not to influence the utilisation of the supply of anti-neuritic substance dis-
tributed in the tissues of the birds.

This is presumptive evidence that alcoholic neuritis is not caused by any
diminished capacity on the part of the organism to utilise the anti-neuritic
substance which might be expected to result from the disturbing effects of
alcohol upon the metabolism of the nervous system.

SUMMARY.

1. A fraction rich in the anti-neuritic substance can be precipitated from
the fats and lipoids (alcoholic extracts) of various animal tissues by means of
ether. ;

274 E. A. COOPER

2. A method based on this observation is described for isolating from
horse-flesh a substance small amounts of which can cure polyneuritis in
pigeons.

3. The substance is insoluble in absolute alcohol, benzene, chloroform,
ether, and ethyl acetate, but is moderately soluble in water.

4. The substance is absorbed to some extent by animal charcoal and is
readily destroyed by alkali. It is not mactivated by sulphuretted hydrogen,
but disappears in large amounts during chemical operations in which colloidal
metallic sulphides are formed.

5. Quinine and cinchonine exert a temporary curative action upon birds
affected with polyneuritis. After being heated at 125° for 6 hours, however,
quinine has no ameliorative effect, so that its curative properties would
appear to be due to its contamination with traces of the anti-neuritic sub-
stance derived from the cinchona bark.

6. The administration of small doses of alcohol to birds fed on polished
rice does not affect the period of time elapsing before the occurrence of poly-
neuritis, and thus appears not to influence the utilisation of the supply of
anti-neuritic substance distributed in the tissues of the birds. This suggests
that alcoholic neuritis does not result from a diminished capacity of the

organism to utilise the anti-neuritic substance.

In conclusion J am glad to have this opportunity of expressing my indebted-
ness to Dr Hugh MacLean for drawing attention to the curative properties of
the residues obtained by treating alcoholic extracts of voluntary muscle with
ether and for much valuable help in the course of the investigation.

REFERENCES.

Chamberlain and Vedder (1911), Philipp. J. Science, 6, 395.

(1912), Phillip. J. Science, 7, 39.

Cooper (1913), J. Hyg. 12, 436.

Edie, Evans, Moore, Simpson, and Webster (1912), Biochem. J. 6, 234.
Funk (1911), J. Physiol. 43, 395.

(1913), J. Physiol. 45, 75.

Maclean (1912, 1), Biochem. J. 6, 333.

(1912, 2), Biochem. J. 6, 355.

Suzuki, Shimamura and Odaki (1912), Biochem. Zeitsch. 43, 89.

aa

i

Rs eee ere

“he, Chae

4a phe) oie”

AXVIT. ON THE LIPOLYTIC ACTION OF
THE’ BLOOD:

By FRANCIS HUGO THIELE.

From the Research Laboratories, University College Hospital
Medical School.

(Recewed April 7th, 1913.)

Our knowledge of the ultimate processes involved in the storage of fat
and its mode of transport from the fat depots to the tissues requiring it
is very scanty. The fate of fat from the intestine to its presence in the
thoracic duct is accurately known from the experiments of Moore [1897].
From the thoracic duct it passes into the blood and is rapidly converted
into a substance no longer extractable by ether, which can pass through
cell membranes to be stored up as fat in the fat depots or to be used as
a source of energy. Of the nature of this substance our knowledge is
practically nil. It has been shown by numerous observers that the tissues
contain a “ fat-splitting ” ferment, the action of which under certain circum-
stances may also be reversible, and it was considered probable that the
blood serum contained a similar ferment. This view was first of all advanced
by Hanriot [1896], who tried the action of the blood on monobutyrin. His
experiments, however, are erroneous as has been pointed out by Arthus
[1902] because he took no precautions to ensure aseptic conditions and also
because he used monobutyrin, which ‘can be very easily hydrolysed, and
it by no means follows that ferments, breaking up such simple esters as
monobutyrin, have the same effect on neutral fats. Hagemeister [1903]
showed that the injection of fat between muscle fibres was followed by its
presence beneath the sarcolemma. He deduced that the lipolytic ferment
of the tissue fluids caused the hydrolysis of the fat, diffusion of the resulting
products occurring, followed by resynthesis in the muscle.

Fischler [1902, 1903] demonstrated the deposition of “fat” in organs
after the injection of soap into the blood, or the perfusion of surviving
organs with blood containing soap, thus lending support to the above.
Munk [quoted by Fischler, 1902, p. 149], however, has shown that soap

276 F. H. THIELE

even in small amounts exerts a toxic effect on the medullary centres.
A small amount of soap apparently does occur in the blood and according
to Schultz [1897] 28°/, of the total fatty acids in the blood are in the form
of soaps.

There is therefore no a priori objection to the possibility of the blood
containing a lipase acting on the fats in the above-mentioned way and of the
products being rapidly diffused into the cells to be resynthesised and stored
up or used as required.

Experiments, however, conducted with strict precautions as to asepsis
show that the blood serum does not contain a ferment capable of splitting
ordinary fats.

The experiments of Cohnstein and Michaelis [1894] showed that there
was a loss of ether-soluble substances on incubating mixtures of blood and
chyle and a corresponding increase in water-soluble substances, and that this
was due to a ferment in the erythrocytes. Their experiments have, however,
the drawbacks that they were not absolutely aseptic and that there were
inaccuracies in their mode of estimation.

Their results have been in part confirmed by Hamburger [1900] who
was able to demonstrate a loss of ether-soluble substance after incubating
a lipanin emulsion with blood.

Weigert [1900], however, considered that Cohnstein and Michaelis’ results
could be explained by autolysis of the lecithin. Hepner [1898] had
previously shown that the blood plasma contains cholesterol and cholesterol
esters of the fatty acids, and thought that the free cholesterol might con-
ceivably have arisen by the action of a ferment on the cholesterol ester.
Weigert therefore thought that the loss in ether extractives in Cohnstein
and Michaelis’ experiments could be accounted for by this, since the liberated
fatty acids would become converted into soaps and thus no longer be ether-
soluble. He experimented with blood plasma, blood and corpuscles alone.
He found that blood plasma on incubation suffered no loss; when blood was
incubated a loss of 18°/, of ether extract occurred ; whereas when corpuscles
alone were thus treated there was a loss of 40—50°/, of ether extract, and
since Hepner and Hoppe-Seyler [quoted by Hepner, 1898] showed that red
cells contained no cholesterol ester or neutral fat, the resulting change can
only have been due to hydrolysis of the lecithin.

In the experiments of the author the following methods of extracting fat
were adopted :

1. Drying the material with sand over a water-bath, and extracting
with ether in a Soxhlet apparatus for 48 hours.

F. H. THIELE 277

2. Mixing with anhydrous sodium sulphate as recommended by Pinkus,
finely powdering the resulting dry mass, and extracting with ether at the
room temperature for several days, with vigorous shaking at intervals. The
extraction was repeated with fresh ether till nothing further was extracted.

3. Mixing with 94°/, aleohol and keeping at 37° for several days, then
filtering and repeatedly extracting the precipitate with alcohol, and finally
boiling up. The united extracts were then evaporated and extracted with
94°/, aleohol. The original residue was also extracted with ether, and the
resulting extracts were divided into their constituents: fats, fatty acids
and lecithin.

4. Digesting with pepsin and hydrochloric acid; evaporating with sand
and extracting in the Soxhlet with ether.

Precautions to preserve absolute asepsis were taken in all these experl-
ments. The various gases were allowed to bubble through the mixtures
after having been filtered through sterile cotton wool. All specimens
showing bacterial contamination, either by direct or cultural examination,
were discarded.

I. CHANGES OCCURRING WHEN BLOOD ALONE IS INCUBATED.

Exp. 1.
100 c.c. of blood (by sand and ether extraction) yielded ae Bae 0-403 g. fat
100 ~,, S. incubated at 37° for 24 hours in closed vessel yielded... UE ToL he
HOO! s,; 3 same incubation, but with air current Ae oe 0°297 ,,
Loss in specimen in air current = 26-3 9/9.
35 without 4 = 54%).
Exp. 2.
100 ¢.c. of blood by sand and ether extraction yielded aos ate 0-366 g. fat.
1000.5 . incubated in air current for 24 hours at 37° ... xt 0°295 ,,
1000; a incubated without air at 37° for 24 hours... sek 0-299 ,,
Loss in air current=19°4 /p.
», Without ,, =18-3 %J.

Exp. 3. Extraction after drying with anhydrous sodium sulphate.

(2) 100 c.c. of blood contained 33 ie sn Ee 0-24 g, fat.
1007 _ after incubation in air current for 24 hours at 37°
contained =F se rae be 0-184 ,,
Loss = 23°3 °/,.
(b) 100 c.c. of blood contained aS ae i353 aa 0-328 g. fat.
LOOW:, - treated as in (a) contained ..,. <5 - 0°276 ,,

Loss =15°9 Jo.

For the purpose of determining to what the loss was due, the following
estimations were made. The blood was extracted with alcohol and the
residue with ether, the alcoholic extract was acidulated with dilute mineral

278 F. H. THIELE

acid evaporated and shaken out with ether. The ether was freed from
mineral acid by repeatedly shaking with water. The ethereal residues were,
united, evaporated, dried and weighed. The lecithin was determined by
precipitation with acetone, the fatty acids by titration with N/10 alcoholic
potash, and the neutral fat subsequently by boiling with a known quantity
of N/10 absolute alcoholic potash.

250 ¢.c. of blood were taken.

Unincubated blood Incubated blood
Total weight of residue 0-958 g. 0:936 g.
Lecithin 0°28 0:127
Fatty acids 0:152 ( 5:4.¢.c. N/10 ale. pot.) 0:245 ( 8:7 c.c.)
Neutral fat Orsi (G@bieey A, eee 0:310 =(11°0 e.c.)
Cholesterol 0°102 0°122

With the exception of the increase in the cholesterol figure the whole
difference is in the lecithin and fatty acids—the loss in the lecithin is
0153 g., and the increase in the fatty acids is 0'093 g. Now if lecithin
contains about 60°/, fatty acids, then 0°153 g. contains 0°092 fatty acid.

Several other experiments showed a similar change.

II. CHYLE ALONE.

Exp. 1.
50 c.c. of chyle contained (by sand and ether method) S00 0-631 g. of fat.
0) ne 3 incubated at 37° without current of air ee 0-579 ee,
100); * incubated at 37° in air current Bar as 1:064
Loss without air current= 8:2 9/9.
an in 5 Ap Sb ie

Exp. 2. Hextracted by the Sodium Sulphate-Ether method.

25 e.c. of chyle contained ae aa sie 0°317 g. of fat.
OO Mass », incubated in air current contained _... se Mois} a

Loss=7'1 /).
Exp. 3. Alcohol-ether extraction.
100 c.c. of chyle (a) Unineubated (UL) Incubated

Ethereal extract of the original extract 1°356 g. 1:243 g. (loss=8°4 °/,)

After acidulating the residue of the original extract, shaking out with
ether and removing mineral acid and adding the ethereal extracts together ;

Lecithin found 0:056 g. 0-014 g.
Fatty acids 0:0 0-034
Neutral fat 1:208 1°137

From this it would appear that there was some hydrolysis of the lecithin
and loss of a small amount of neutral fat. The fatty acid increase can be
explained by the hydrolysis of the lecithin, since 0-048 g. lecithin contains
about 0°029 g. of fatty acid.

DS ee ee

alae canada

F. H. THIELE 279

III. EXPERIMENTS ON THE INCUBATION OF A MIXTURE OF
CHYLE AND BLOoob.

(1) The sand method of extraction.

Exp. 1.
100 c.c. of a mixture of equal parts of chyle and blood were found to contain 1-281 g. of fat.
100 c.c. of same mixture incubated for 24 hours without air current were found
to contain ri eis 1200) 33
100 ¢.c. of same, incubated in an air current b for 24 Hane ponuined Boe 09155 _,,
100 e.c. of same, incubated in an air current for 48 hours contained ae 0°8025_,,
Loss was therefore :—in incubation without air current Gaol,
in 24 hours incubation with air current 28°5 °/,,
in 48 ,, PS ut y BT or Fae

Exp. 2. Performed as above, showed ;

100 c.c. of a blood chyle mixture contained ae a nt 1°41 g. of fat.
100 ¢c.c. of same incubated in air current at 37° for 24 hours mh * 0°536_—Cés,,
100 ¢.c. of same, incubated for 48 hours, same way... ae oe OTS ess
Loss after 24 hours was 61°9 °/).
ne 48 30 73°5 °f,.

(2) Sodium Sulphate Method.

100 c.c. of a mixture of blood and chyle contained Sc 1-228 g. of fat.
100 c.c. of the same mixture, incubated in a current of air at 48° i 24 howe O-6615—,,

Loss= 46°2 9/p.

These figures show that on incubating chyle and blood together there
is a large loss of ether extractive. This is not due to the method of
extraction, it is as apparent with the one method as with the other. The
loss is greater after air has been passed through the mixture, but contrary
to the observation of Cohnstein and Michaelis [1894] it also occurs in the
absence of an air current. The loss again is not the sum of the losses
that would be obtained by incubating the blood and chyle separately.

Thus in the following experiments :

100 c.c. of blood yielded es , = 0°29 g. of fat.
100 c.c. incubated for 24 hours with panne of air at 37° Ficided So Or232 05

Loss=0°058 g. or 20 °/y.

100 c.c. of chyle yielded XS ve se a Ps I OLIp Se

100 c.c. incubated as above won ae ah wa BR 103450:
Loss=0°168 g.=14°0 °/,.

Now. a mixture of 50 ¢c.c. of each contained arte ae a 0-770 g. of fat.

But incubation as before yielded Aa Ba ae Aes OATS. =:

Loss=0°298 g.=38°4 °Jpo.

280 F. H. THIELE

Now the sum of the individual losses for 50 c.c. of each = 0°117 g., which
is less than half the loss which occurred, when the two were incubated

together.
This is also shown by the following experiment :
100 c.c. of chyle were found to contain 500 Se As Be 1:268 g. of fat.
1000>5; ,, yielded after incubation ... ona sa ae Uei(cls}) 5p
Loss=0'1 g.
100 c.c. of blood yielded ine ee sie ia ve 0-314 =~,
1000s. ,, after incubation yielded... aes aoe ee 0:244 =«Ca,
Loss=0:070 g.

A mixture of 50 ¢.c. of chyle and 25 c.c. of blood were estimated to contain ... O25
After incubation a similar amount yielded ... : ach aus 0°329 i,

Loss=0°38 g.=53'8 °/),
whereas the sum of the losses for each separate incubation
=0:05 + 0:0175 =0-0675 g.

The loss is thus considerably greater than the sum of the individual losses,

(3) To see if the serum contained the substance which caused this loss
of substance extractable by ether a mixture was made of 100 cc. of chyle
and 100 c.c. of sheep’s serum.

100 c.c. of the mixture contained seg Ene nae 506 0:663 g. of fat.
1000s. at “3 after incubation in an air stream after 24 hours yielded 0‘591
Loss=0:072 g,=10°8 °/).

”

Similar experiments give like results up to 15 °/).

(4) In order to determine whether the change is due to a ferment in
the red corpuscles or other formed constituents, the following experiments
were performed with carefully washed erythrocytes. These were obtained
by centrifuging whipped blood and washing six times with sterile normal
saline in sterile centrifuge tubes.

Exp. 1.
50 c.c. of chyle and 5 e¢.c. of red cells contained * ok 0°802 g. of fat.
100 ,, aD OMe ss incubated for 24 hours in an air current
yielded 1172 ,,

Loss=0°43 g.=26°9 °/,.

In the next part of the experiment the chyle was first heated to 80° for
15 minutes:

100 ¢.c. of this with 10c.c. of the suspension of red cells incubated for 24 hours
at 37° yielded... aM 6c =e oes oe 1°107 g. of fat.

Logs for unboiled chyle was 26:9 °/,.
sone Dolled o 30°9 °/,.

i a

F. H. THIELE - 281

Exp. 2.
100 c.c. of chyle yielded iF fa ss ‘a6 “ep 1°825 g. of fat.
20 c.c. of red corpuscles yielded uae Ae 0:048 ~~,
100 c.c. of chyle +20 ¢.c. of the corpuscles incubated for 24 hours in an air
current yielded ane Aa Ser ae a 1-284" ;,
Loss was 31°4 °/,.
Exp. 3.
50 ¢.c. of chyle yielded sai ae wae wie eats 0°912 g. of fat.
20 c.c. of corpuscle suspension yielded si oe ane vf 0:048_—,,

Both were heated to 80° and then incubated at 37° for 24 hours. The total amount of ether
extract found was 0-937 g. ‘The loss was therefore practically nil.

It will thus be seen that probably the red cells, or substances centri-
fugalised down with them, have the power of causing this loss of ether
extractive.

(5) The next question investigated was whether this power was due
to the haemoglobin, to a ferment in the cells, which may be to some extent
diffusible into the plasma, or as stated above to elements precipitated by the
centrifuge with the red corpuscles.

This was investigated by obtaining crystals of haemoglobin from dog’s
blood. The blood was whipped and centrifugalised, the corpuscles were then
repeatedly washed with normal saline and then laked with ether, centri-
fugalised and washed with dilute alcohol and ammonium sulphate solution.

Exp. 1.
50 c.c. of chyle +25 ¢.c. of a watery Hb solution yielded a32 3 0°550 ether ext.
Twice this amount was incubated at 48° for 24 hours in air current and

yielded aie ane sat oe = a. 0°729 .

Loss was 0°371 g.

Exp. 2. The haemoglobin was previously treated with a saturated solution of sodium
fluoride in a saturated ammonium sulphate solution. The crystals were then repeatedly washed
with a saturated solution of ammonium sulphate solution to remove the sodium fluoride.

50 c.c. chyle together with 50 ¢.c. watery Hb solution yielded sas = 0-665 g. of fat.
The same amount after 24 hours incubation in a stream of air at 48° yielded... Oz6320

Loss was 0°033 g.=5 Jp.
Exp. 3.

100 c.c. of chyle was mixed with 50 ¢c.c. of a watery solution of Hb. The Hb
had been previously dried for several days at 50°. The mixture contained 1-268 g. of fat.
The same amount was incubated for 24 hours and yielded ih os T2027 >>>
Loss =0:056 g.=5:2 °/).

To elucidate this point, whether the loss was due to the action of the
haemoglobin, a mixture of chyle and blood was incubated and a mixed stream
of carbonic oxide and oxygen gases was passed through the fluid. The
carbonic oxide converted the haemoglobin into carboxy-haemoglobin and
the oxygen was available if necessary.

282 . ¥. A TRIE

By. previous experimentation it was determined that the “lipase”
obtained from pig’s liver was unaffected by prolonged exposure to a stream
of carbonic oxide gas as regards its action on ethyl butyrate. _

The blood used was first saturated with carbonic oxide.

Exp. 1.
50 c.c. of blood and chyle yielded... ies oe 0°490 g. of ether ext.
50 c.c. of the same mixture after incubation Pielded ais Ses 0-390 es; 55
- Loss was 0°1 g. =20°4 9p.
Exp. 2.
100 c.c. of a mixture of equal parts of blood and chyle yielded Ae 1-096 g. of ether ext.
The same amount after similar treatment yielded naa Ovo toes sh

Loss was 0°365 g.=33°3 0/,.

The experiments show that the change can still occur when the haemo-
globin is put out of action as an oxygen carrier.

In these experiments it was certain that the haemoglobin was converted
into carboxy-haemoglobin because of the colour of the mixture.

In the method of preparation of the haemoglobin crystals, protems in
the corpuscles or substances in the plasma like blood, dust particles, and
leucocyte debris may be precipitated so that any possible ferment in connec-
tion with them would be precipitated as well. The action of the carbonic
oxide and sodium fluoride leads to the supposition that there is a ferment
separate from the haemoglobin.

(6) The following experiments were performed to see if the oxygen
is really a necessity, or if the air current acts simply mechanically, thoroughly
mixing the blood and chyle. The mixtures were incubated for 24 hours
at 37°.

(a) 50 c.c. of a mixture of blood and chyle yielded ... ar sc 0°350 g. of fat.

100 c.c. of this mixture after treatment in coal gas yielded ... So 0°542 6
Loss was therefore 0°158 g. = 22°6 °/).

(b) 50 ¢.c. of blood and chyle contained . Zz ae ate 0°450 ne

100 c.c. incubated in a stream of edroped gas conan ots Ae 0°696 rp
Loss was 0°204 g.=22°5 °/,.

(c) 50 c.c. of blood and chyle mixture contained _... : aoe 0°395 ne

100 c.c. of the same mixture in a hydrogen stream paciied a ‘ie 0-608 Ss
Loss was 0:182 g.=23 °/,.

(d) 50 c.c. of chyle+ 25 ¢.c. of blood contained arp Ae 5a 0:67 3 aeeee
The same mixture incubated in air stream yielded Spe 36 0:229 eee.
The same mixture incubated in CO, stream yielded Ses ia 0°478

Loss in air=0°444=66 °/,.

” CO,=0°195 = 29 Ye
Thus it would appear that oxygen aids the action of the ferment, but
on comparing with the result in (c) of this series it is obvious that mechanical

agitation by the passage of a current of any gas also appreciably aids the
process.

;
j

‘
t.

Pell at 2 el

ee OS

F. H. THIELE ' 283

(7) The influence of protoplasmic and ferment poisons was next
examined,

Exp. 1.
(a) 100 c.c. of a mixture of chyle and blood yielded ... ae ae
100 c.c. of same mixture containing 5 °/, sodium fluoride was incubated
for 24 hours in air stream at 48° and then contained _... aie 0°657 +

0°708 g. of fat.

Loss was 0°051 g.=7 °/,.

(b) 100c.c. of the same mixture with potassium cyanide added and incubated
in the same way yielded oe nce Pe 0°641 PP

Loss was 0°067=10 °/,.
(c) 200c.c. of the same mixture was incubated in the ordinary way and yielded 0°547 oe
Loss was 0°869 = 62 °/.
(8) Some experiments were performed to determine the fate of the
individual constituents.
Exp. 1. A mixture of equal parts of blood and chyle was made. (a) 100 ¢.c. were directly

dried with sand and extracted with ether in the Soxhlet; (b) 100 ¢.c. were similarly treated after
incubation in an air current at 37° for 24 hours.

(a) Unineubated (b) Incubated
Total residue 0°732 g. 0-416 g.
Neutral fats 0°583 = (19°8 ¢.c.) » 0°327 (111 c.c. N/10 alk.)
Fatty acid 0-0 0:0
Lecithin 0-096 0-032
Cholesterol 0053 0°057

There is thus, as was expected, a loss in neutral fat. The dried residue
was then extracted with alcohol in the incubator at 37° while being con-
tinuously shaken, the alcoholic extract was evaporated and treated with
mineral acid, and shaken out with ether, the ether subsequently being
washed with water to remove the mineral acid. The ethereal extract was
then titrated with N/10 alkali in alcohol.

The unincubated required 0°8 ¢.¢, = 0°0226 g. fatty acid.

The incubated required 3'1 c.c. = ‘0874 g. fatty acid.

On subsequently boiling the extract of the incubated specimen with
a measured amount of N/10 alcoholic potash, it was found that 5°8 c.c. were
- used up =0:1709 g. of fat, which had been liberated. This fat scarcely
contained any trace of phosphorus. Hence the alcohol liberated a certain
amount of neutral fat, which had become ether-insoluble to direct ethereal
extraction.

Other experiments gave similar results as regards the amount of fatty
acids present in the ether extract of the second alcoholic extraction, and
showed that neutral fat was also present which did not come out in the
original extraction. .

Bioch. v1 19

284 F. H. THIELE

Exp. 2, A mixture of chyle and blood treated as in Exp. 1.

Non-incubated Incubated
Total residue 0°746 g.- 0:402 g.
Neutral fat 0°5938 (20:2c¢.c.N/10alk.) 0°294 (10 c.c.)
Fatty acid 0:0 0:0
Lecithin 0:99 0°029
Cholesterin 0-068 ~-0:073
Subsequent extraction as above.
Fatty acids 0:0328 g. (1:2 c.c.) 0:093 g. (3°3 c.c. alk.)
Fats 0:0 OSS") (CesT = =)

The same experiment was made with the same materials, but instead
of subsequent alcoholic extraction the finely ground material after the ether
extraction was digested with pepsin and hydrochloric acid and then dried
and extracted with ether.

Exp. 3. Thus 100 c.c. of a mixture of blood and chyle were examined.

(a) Unincubated (b) Incubated
Primary extract 0°753 g. 0°449 g. (loss=43-1 °/,)
Secondary ,, 0-028 0°32
Fatty acid in secondary extract trace 0:0789
Neutral fat nil 0°233 (7°9 N/10 alkali)

The same method as above was used. 200 cc. of a mixture of blood and
chyle were used for immediate estimation by ether extraction followed by
treatment with pepsin and extraction, and two separate lots of 200 c.c. were
incubated in an air current at 37° for 24 hours, one examined by ether-
alcohol extraction, and divided into its constituents, the other first ex-
tracted with ether and then digested with pepsin and acid and again
extracted with ether.

Exp. 4. Specimen “A” Specimen ‘‘B”
ether-alcohol digestion with
Control extraction pepsin, ete.
g. g. g.
Total residue (first extract) 1°64 0-813 0°799
(loss = 50°4 0/,) (51:8 /,)
Neutral fats... sie 1°3934 0°70 —
(47:2 c.c.N/10 alk.) (23-9 c.c. N/10 alk.)
Fatty acids... Be 0:0 0:0 —
Lecithin ae ae 0°198 0-087 --
Cholesterol ete. re 0:0486 0:021 —
Fatty acid in second extract 0:0479 0:121 0°1325
(1:7 ¢.c.N/10 alk.) (4:3¢.c. N/10 alk.) (4°7 c.c.)
Fat in second extract .... nil 0:545 0°7139
(18°5 c.c.) (24:2 c.c.)

It was also found that by extracting with alcohol on the water-bath and
then with ether, the whole of the ether-insoluble fat could be recovered.

Two lots of 100 cc. of a mixture of chyle and blood were examined
(a) unincubated, (6) incubated in the usual way.

a en

F. H. THIELE 285

(a) Unincubated (b) Incubated
Primary extract 0:712 g. 0°467 g.
Secondary ,, 0-028 0-298

These experiments further show that the neutral fats are not hydrolysed
by the blood or chyle, but that the lecithin is the only fat hydrolysed by the
ferments in the chyle or blood. This is at variance with the observation
of Hamill [1907] who states that the chyle can hydrolyse a neutral fat,
viz., olive oil. :

(9) The question then arose if this ether-insoluble fat could be isolated
and its constitution determined and especially if it were diffusible and
soluble in water. |

The method adopted was that of Cohnstein and Michaelis [1894], viz.
diffusion through parchment thimbles into distilled water. In all the
experiments it was found that only soaps, etc. diffused through, and
no substance containing the neutral fat, and that the amount of fatty acid
thus obtained corresponded to the degree of hydrolysis of the lecithins.

Experiments were also made with a lipanin emulsion, made with sterile
ascitic fluid and lipanin. A very fine emulsion was produced.

It was found that on incubation with blood and a current of air the
same results were obtained as above with chyle fat and that the fat could
be recovered again by treatment with alcohol or pepsin and hydrochloric acid.

These observations are in accordance with those of Mansfeld [1907]
who showed that the fat in lipanin after incubation with blood became
non-extractable by ether, but could be recovered by Liebermann’s method.

Hamburger [1900] was only able to obtain this result when he made his
artificial chyle with a chyloid fluid. In these experiments, as in those of
Mansfeld, ordinary ascitic fluid worked well.

CONCLUSIONS.

It would seem justifiable to draw the following conclusions :

1. Blood and chyle contain a ferment which can hydrolyse lecithin but
not neutral fat.

2. When blood and chyle fat are incubated together, the neutral fat
forms an absorption combination with the proteins and is thus rendered
non-extractable by ether. The combination can, however, be broken up by
peptic digestion, or treatment with alcohol. It is not formed during the
process of drying.

_ 3. The combination appears to be broken up by heat, since larger losses
are obtained if the incubated mixtures are not heated, and if the material be

dried at a low temperature.
19—2

286 F. H. THIELE

4. The formation of this complex appears to be due to the formed
elements of the blood, since the corpuscles or material carried down with
the corpuscles on centrifugalisation have the same effect, and serum by
itself has not.

5. The action does not appear to be due to the haemoglobin, but to
some material precipitated in the method of preparmg the haemoglobin,
since on rendering the haemoglobin inert as an oxygen carrier by carbon
monoxide, the change still occurs.

6. The change occurs best in a current of air or oxygen, and if the
mixture be thoroughly agitated.

7. The combination appears to be due to the action of a ferment
because :

(a) the haemoglobin may be rendered inert without affecting the process ;

(b) ferment poisons seriously interfere with it.

8. Similar combinations may occur when blood is mixed with a very
fine emulsion of fat. The compound, however, does not appear so stable
since it 1s practically always broken up by heating and by alcohol. .

9. The albumin-fat combination is not diffusible nor soluble in water.
After incubation an increase in diffusible material takes place, but this is
only due to the diffusion of the products of the lecithin hydrolysis, viz.,
glycerophosphoric acids, soaps, choline.

10. The chyle and corpuscles contain a ferment capable of breaking up
lecithin and liberating the fatty acids.

A summary of this paper was communicated to the Pathological Section
of the Royal Society of Medicine, Jan. 1911.

Part of the expenses of this research were defrayed by a grant from the
Government Grant Committee, Royal Society, London.

REFERENCES.

Arthus (1902), J. Physiologie, 4, 56, 455.

Cohnstein and Michaelis (1894), Pfliiger’s Arch. 55-57.
Fischler (1902), Virch. Arch. 170, 100.

(1903), Virch. Arch. 174, 34,

Hagemeister (1903), Virch. Arch. 172, 72.

Hamburger (1900), Arch. Anat. Physiol. 544.

Hamill (1907), J. Physiol. 35, 151

Hanriot (1896), Compt. Rend. Biol. 48, 123, 925, 753.
Hepner (1898), Pfliiger’s Arch. 73, 595.

Kastle and Loevenhart (1908), Lancet, July, 135.
Mansfeld (1907), Centr. Physiol. 21, 666.

Moore, B. (1897), Reports of the Johnston Lab., Liverpool.
Schultz (1897), Pfliiger’s Arch. 65, 229.

Strauss (1907), Deutsch. Med. Woch. 37.

Weigert, R. (1900), Pfliiger’s Arch. 82, 86.

Oe

1 Aoi, ae ee ee ae

Aas

——_

XXIX. ON THE LIPOLYTIC ACTION OF
THE. TISSUES:

By FRANCIS HUGO THIELE.

From the Research Laboratories, University College Hospital
Medical School.

(Received April 7th, 1913.)

The following experiments were carried out to test the lipolytic activity
of the various organs towards the neutral fats.

It is only comparatively recently that such enzymes have been demon-
strated in the tissues; their presence was inferred previously from the action _
of tissue extracts on simple esters, such as monobutyrin, ete. That these are
by no means the same, has been shown by Cohnstein [1904] for the castor-
oil seed. Arthus [1902] showed a similar difference for blood serum.

Nencki and Liidy [1887] tested the lipolytic power of tissues on amyl]
salicylate and found that the tissue extracts worked best in an alkaline
medium.

Hanriot [1896] used monobutyrin and demonstrated a lipolytic ferment
in all tissue extracts except the thyroid. The enzyme ceased to act when
a certain degree of acidity had developed and could continue when the
acidity was neutralised. He also showed that the tissue extracts could
hydrolyse the ethyl esters of formic, acetic, propionic and butyric acids.

Arthus [1902] drew attention to the errors in Hanriot’s observations and
was unable to confirm his work. Kastle and Loevenhart [1900] experimented
with ethyl butyrate and found that the ferment was closely attached to the
cells, and that liver extract was more active than the pancreatic on this ester,
but that the pancreas could hydrolyse the esters of the higher fatty acids
more readily than those of the lower.

Siebert [1900] studied the action of pancreatic extracts on egg yolk, and
further, showed that the blood serum is incapable of splitting lecithin. Sub-
sequently Umber and Brugsch [1906] experimented with tissue extracts,
using egg yolk as their emulsion of neutral fat. The various organs were
washed free from blood, then finely ground up with kieselguhr and the juice

288 F. H. THIELE

pressed out with a hydraulic press. Two series of experiments were per-
formed, one with the organs from an animal killed whilst digesting, the other
in starvation. One series of experiments was conducted under aseptic
conditions, the other under toluene.

These observers concluded that the tissue extracts can hydrolyse the fat
in egg yolk and that during digestion the pancreatic extract is the most
powerful, closely followed by that of the spleen, but that the liver and
intestinal mucosa are more powerful than the pancreas. Further that
activating bodies would appear to be present in some of the organs.

In the following series of experiments it was intended to study the
action of the various tissue extracts on:

(1) Egg yolk.

(2) Artificial emulsions of lecithin.

(83) Lecithin not in a state of emulsion.

(4) Chyle obtained from the human subject.

The tissues were ground up finely in a mortar, mixed with an equal bulk
of normal saline and pressed through a press so that the tissue extract con-
sisted of a fine suspension of the tissue in normal saline. ‘The tissues were
removed under strict aseptic precautions; pigs’ and dogs’ organs were chiefly
used. In the case of the pigs’ organs, these were removed in the slaughter-
house in gauze soaked in 2°5 °/, carbolic, the outside was removed with a sterile
knife and the tissue transferred to a sterile mortar and ground up with sterile
normal saline. The dogs’ tissues were obtained from healthy dogs, which
were anaesthetised and bled to death. The animals were subsequently
washed out with sterile normal saline. No toluene experiments are recorded,
because it was found that even adding toluene up to 2°/, did not inhibit the
growth of organisms which were capable of splitting up egg yolk very readily.
The “fat” contaming emulsions were also prepared under strictly aseptic
precautions and sterilised.

The mixtures of “fat” and tissue extract were incubated at 37° for
15-20 hours. At the end of that time-they were examined directly and
culturally for possible bacterial contamination and any showing this were
rejected. The method of estimating the effect of the tissue extracts was to
determine the total amount of “fat” present by boiling the mixture of un-
incubated material with sodium carbonate for some time on the water bath,
and then, after acidulating with sulphuric acid, extracting with ether. The
ethereal extract, after removal of the mineral acid, was then treated with
N/10 KOH to neutralise the fatty acids, and then hydrolysed with alcoholic
potash, evaporated, and taken up in water, acidulated with sulphuric acid

F. H. THIELE 289

and the liberated fatty acids extracted with ether, which was then freed from
sulphuric acid by shaking up with water. The ethereal solution of the fatty
acids was titrated against N/10 alcoholic potash and then reckoned in terms
of oleic acid; 1 ec. N/10 KOH =0:0282 g. oleic acid. The incubated mixtures
were neutralised with sodium carbonate, well heated on the water bath and
acidulated with sulphuric acid to liberate the fatty acids which were ex-
tracted with ether and purified as before. The amount of fatty acid was
thus determined by titration with N/10 alcoholic potash.

In some of the experiments it became necessary to separate oleic acid
and other high fatty acids from any lower acids, such as lactic acid, which
might be formed from the carbohydrates. This was done by having a double
series of mixtures; in the one the total fatty acids were determined in
the way mentioned above, in the second the mixture, after neutralisation
and boiling, was treated with sulphuric acid to liberate the fatty acids and
then treated with ammonium sulphate, according to the methods of Magnus
Levy, to separate the lower from the higher fatty acids.

A. Effect of tissue extracts of various organs on sterile egg yolk mixture,
containing 0°5°/, sodiwm carbonate.
10 c.c. tissue extract was used with 20 c.c. of yolk suspension. Results expressed in c.c.

decinormal alkali.
Totalfatsas Fatty acids 9/, hydro-

No. of Exp. Pigs’ organs fatty acids liberated lysed
i, Pancreas sire Foe Ape 53°2 48°0 90-0
Liver ee Ae ee 55°8 20-5 36°7
Spleen oe wes ia 53°8 20:2 37°5
Pancreas +5 c.c. Splee ae 53°4 40°0 75°0
39 > ” (the
spleen was first boiled) _.... 53°4 40°0 750
Liver+5c.c. Spleen ... ae 56'0 21°6 38-6
Pancreas +5 ¢.c. Spleen Bae 54°2 38°6 7071
Serum te an ... Notdetermined 0-0 0-0
yy) Pancreas er nef sits 53°6 30°6 57:0
10 c.c. Bile ... Bes Pee 51:8 19°08 36°7
Pancreas +5 c.c. Bile ... Sus 54°0 30°0 55-5
Pancreas + 5c.c.Bile + 5c.c. Serum 54°6 30°8 56:0
10 c.c. Serum ay ... Not determined 0:0 0-0
3 Pancreas ate ¥F <4 54:0 37°2 68-2
Liver PES cus Sas 59°2 21°8 39°5
Kidney he i ay 54°6 20-4 37°87
Muscle Fog ais “a 54-4 20°2 37°0
Spleen pe ye oe 54°4 21°6 40-0
Pancreas +5 c.c. Splee af 54°8 39°1 71:0

Muscle+5c.c. Spleen . ie 54°4 21:0 39°87

290 F. H. THIELE

Dogs’ Tissues.
5 e.c. of extract +20 c.c. of yolk suspension in each experiment. Results expressed as before.
Total fats Fatty acids °/, hydrolysed

Dog 1 | Spleen ee a ie 561 20°5 36°5
Hungry Liver zine ae sie 56:8 20°5 36:0
animal Pancreas sis AG see 55°9 22°8 40°9

Kidney bac es aa 56:3 20°2 35°8
Liver + Spleen (5 ¢.c. of each) ... 57:1 22:0 38:0
Pancreas + Spleen (5 ¢.c. of each) 56:2 37:0 65°8
Pancreas + 5 ¢.c. of Serum “ee 58-4 37°5 66°2

5 c.c. of extract +20 ¢.c. of yolk suspension in each experiment. Results expressed as before.

Dog 2 Pancreas oe eae a 54:2 41:0 75°5

Digesting Pancreas + Serum iis i 54:4 37:0 68:0
animal Serum oe soe eee ; — 0:0 0:0
Liver re ase ae 55:0 20°8 37°8

Spleen 5s ae aA 54:3 20-2 37:0

Pancreas + Spleen sth nae 54°6 44:5 81°5

Pancreas + boiled Spleen ane 54:6 41°5 77:0

5 c.e. of extract +20 ¢.c. of yolk suspension in each experiment. Results expressed as before.

Dog 3 Pancreas aa ee sis 52°8 30:0 57°0
Hungry Pancreas + Liver (5 ¢.c. of each) 53°6 34:0 63:4
animal Pancreas + Spleen (5 c.c. of each) 5a! 33°8 61:2

Spleen Fs est 596 53°9 21:2 40:0
Liver sas ane nee 53-4 22-0 401

Human Tissues.
Removed twelve hours after death. 20 c.c. of egg yolk and 10 c.c. of the tissue suspension.
Results expressed as before.

Total fats Fatty acids
as fatty acids liberated 9/9 liberated
Pancreas Ge aie aa 56°2 46-0 81'8
Pancreas and Liver (5 c.c.) i. 57°0 48:0 84:3
Pancreas and boiled Liver (5 c.c.) 57°0 49-0 85°9
Liver ... heb i Sie 56°8 21:0 37°0
Spleen nie ae asa 56:0 21°4 38:2
Liver+5c.c. Spleen... Jor 57°0 23:0 40°3
Pancreas +5 ¢c.c. Spleen... ane 56°6 48°6 85°6
Sheep’s Tissues. Quantities as before.
Liver ... es Seis ane 54°8 20°2 -36°6
Spleen dae aes ees 54°0 19°8 37°0
Liver+Spleen (5¢.c.) ... ee 55:0 20°4 37:0
Liver + boiled Spleen (5 c.c.) SEr 55°2 20°4 37:0
Pancreas sak fe Se 53°8 ~ 30:2 56:4
Pancreas+45c¢.c. Spleen... ss 54:2 32:0 58°6

From these experiments it will be seen that the tissues have the power
of breaking up the lecithin when it is exhibited to them in the form of egg
yolk emulsion.

Tn all cases the pancreas has a much more powerful action than any other
tissue extract, and there is from these experiments no distinct evidence of

= —

eo eQAILH ue ere -

On nes Cag fF

RG ine igh EID HL rm Nm

2

F. H. THIELE 291

any kinase capable of augmenting the activity of the ferments. The only
example of this effect occurs in one of the experiments with the tissues of a
dog in a fasting state, but in the other experiments with the organs of a dog,
and with those of pigs and sheep, which are always slaughtered in a fasting
condition, there is absolutely no evidence of any activating ferment in any of
the tissues. In these respects these experiments do not agree with those of
Umber and Brugsch [1906].

The above results, however, like those of Umber and Brugsch, take no
notice of the formation of the lower fatty acids, which Magnus Levy [1902]
showed were formed during the incubation of tissues from the carbohydrates
present. In order, therefore, to eliminate this error, the following experi-
ments were performed in double series; in one the total fatty acids were
estimated, in the other the lower fatty acids were estimated after separation
by ammonium sulphate, the difference between the two being due to the
amount of fatty acid liberated by the hydrolysis of the fat.

In the following series of experiments 10c.c. of tissue juice were in-
cubated with 20 cc. of yolk suspension in 0°5°/, sodium carbonate at 37° for
18 hours. The results here are in terms of N/10 NaOH as before.

Total fat
No. of Total fatty Lowerfatty Higher fatty in terms of °/, hydro-
Exp. Pigs’ acids found acids acids fatty acids lysed
1 Liver 23:0 c.c. 1-6)cre: 15°4 ¢.c. 52°0 29°6
Spleen 21:0 3:2 17°8 51°2 34°7
Kidney 18-0 4:0 14:0 51:3 27°3
Pancreas 39°0 15 37:5 51:0 73°5
2 Sheep’s
Pancreas S1-Ole.e. 2°0 c.c. 29:0 c.c. 51:0 56°8
Liver 19:0 671 12°9 51°8 27°9
Spleen 12°8 1°8 11:0 51-2 21°5
Kidney 13:2 4:0 9-2 51-4 17°9
3 Dogs’
Pancreas 38°0 c.c. 1°4 @.c. 36°6 ¢.c. 49°2 74:4
Liver 18-0 4-1 13°9 49-8 27'8
Spleen 17:0 a7 11:3 49-0 23°0
Muscle 12°0 5°2 6:8 49°3 13°7

These experiments do not very greatly alter the results of the previous
experiments, and confirm the lipolytic power of the various tissue extracts
towards egg yolk emulsion.

B. Experiments with liver extracts on lecithin, not in a state of emulsion.

Weighed quantities of lecithin were placed in sterile flasks and sterilised.
Then sterile tissue extract was added and allowed to act for varying intervals
in the cold or warm.

292 F. H. THIELE

In Exp. 1. 19 °/) was broken up. | yy 16 hours at 37°.

a 2. 2B = oe os )
e 3), SO eames “ In 14 days at 4°.
7 4, 56 Wf ”” 99 ” 28 99 oy)

The result in each case should be somewhat higher, because no account
was taken of the amount of lecithin in the liver extract added. It is
interesting to note that jecorin could also be found in all these experiments

after incubation.
C. Action of tissue extracts on chyle.

The chyle was obtained from a healthy young man suffering from
lymphatic obstruction, producing lymphangiectasis in one lower limb. The
chyle could be obtained under aseptic conditions in large quantities after a
period of blocking the fistulae in the lower limb. It contained from 1:3 to
1:6 g. of fat per cent., and from 0°064 to 0:08 g. of lecithin.

The lipolytic ferment in the chyle was destroyed by heating to 60° for an
hour, thus also sterilising the fluid.

Sodium carbonate was added to make a 0°5°/, solution.

In these experiments a double series was made in order to determine
the relative amounts of higher and lower fatty acids.

20 c.c. of extract of the finely pounded tissue were added to 50 c.c. of chyle
and incubated for 18 hours at 37°. Results expressed in cc. of N/10 NaOH.

No. of Fatty acids Lower fatty Higher fatty °/) hydro-
Exp. Pigs’ Total fats found acids acids lysed
1 Pancreas 24:0 18°8 16 17-2 71°5
Liver , 27-2 13:0 6°4 6:6 24:2
Spleen 23°1 56 2°6 3°0 13°0
Kidney 24°8 94 4:2 5:2 21:0
2 Pigs’
Pancreas 22°9 14:2 0°8 13°4 58°5
Liver 25°8 9°2 3°8 5:4 20:9
Spleen 23-2 7-4 4°8 26 11:2
Kidney 24:4 6°8 30 3°8 15°6
3 Sheep’s
Pancreas 22:0 13°2 1:0 12-2 55°4
Liver 27:1 9:2 4:0 5:2 18:0
Spleen 224 5'6 3:2 24 10°7
Kidney 25:0 6:0 2°0 4:0 16:0
4 Dogs’
Pancreas(10c.c.) 21°6 15:0 1:0 14:0 65:0
Liver 30°3 13:4 4°2 9°2 30°3
Spleen 24:7 6°5 2°8 3:7 15:0
Kidney 26°6 8-2 3:1 51 20:0

These results show that a certain amount of the fatty acids found in the

F, H. THIELE 293

incubated mixture is due to the formation of lower fatty acids. Further, that
the tissues have the power of liberating higher fatty acids when incubated
with chyle, but as chyle and the tissues contain not only neutral fat, but
lecithin as well, the question then arises as to the relative proportion
hydrolysed by the tissue ferments. Finally the figures are quite small when
compared with those obtained by using egg yolk.

To determine this question a series of experiments was carried out in
which the lecithin, the total amount of fats and the higher fatty acids formed
were determined before and after incubation.

For this purpose three sets of each mixture were made. The lecithin
was determined by extracting the material with alcohol and ether, evapo-
rating, taking up in ether, precipitating by acetone and determining the
_ phosphorus in the precipitate by Neumann’s method.

lec. N/2 NaOH = 0:014331 g. of lecithin.

Lecithin contains 66 °/, fatty acids.

Quantities as before. Results reckoned as fatty acids in terms of N /10
alkali.

Exp. 1. Pigs’ Liver Pancreas Spleen
Total fats oa fe 25°1 c.c. 21°1 c.c. 23°5 c.c.
Lecithin a os 0-261 g. 0:0902 g. 0°1504 g.
Lecithin fatty acids Bo Gere: 2-1e:c, 3°5 ¢.¢.
Fatty acids liberated 101 14:0 68
Lower fatty acids liberate 4°2 1°8 3°0
Higher . a 5°9 12°2 3°8
Lecithin left ... aes nil nil nil
Fatty acid in Lecithin... 6:1 2-1 3°5
Fatty acid found aes 5°9 12-2 3°8

Exp. 2. Dog’s Liver. Quantities as before.

Total fats Sas wide is 25°9 c.c.
Lecithin abe es Ee 0°339 g.
Lecithin fatty acids ane sae (ae
Fatty acids liberated nS ie 12-2
Lower fatty acids liberated ... Ae 4:0
Higher re ‘5 Sas ae 8-2
Lecithin left Sor nil

Fatty acids in the lecithin =7-9 c.c. N/10 NaOH.
5 », found liberated=8-2 _ ,, 3

Exp. 3. Quantities as before.

Sheep’s Liver Spleen Kidney
Total fats as fatty acids .,. 30°4 c.c. 28°9 c.c. 30°2 ¢.c.
Lecithin Sas ne 0:244 g, 0:177 g. 0°235 g.
Fatty acids in Lecithins ... b°7 G:C; 4-2 c.c. 5°5 ©.¢.
Fatty acids liberated a 11:3 8-4 78
Higher fatty acids liberated 5°5 3°9 5°3
Lower aS 3 5°8 4°5 2°5

- Lecithin left a nil nil nil

294 F. H. THIELE

Exp. 4. Spleen + Liver
Pigs’ Liver Kidney Spleen (10 ¢.c. of each):

Total fats ag nae 26°8 c.c. 25°4 ¢.c. 24:2 c.c. PAID:L OG
Lecithin ae ie 0:2178 g. 0°1576 g. 0:1089 g. 0:3152 g.
Fatty acid in lecithin 60 5:1 Ce: 3-7 Gc. AEG) 3G 7:4 c.¢.
Fatty acids liberated ae 8:3 5°2 6:5 13°7
Lower fatty acids liberated 2°8 2-1 4:0 671
Higher o 5 5°5 Bill 25 ; 76
Lecithin nen Be nil trace nil nil

In all these experiments it will be noticed, with the exception of that
with pancreatic extract, that the amount of fatty acids liberated corresponds
to the amount of fatty acids present in the lecithin, and that in all cases the
lecithin is practically completely hydrolysed.

Evidence of the inability of the lipase of the liver to attack neutral fat is
obtained from experiments on aseptic autolysis at 4°. Fat expressed as

ec. N/10 NaOH °/).

Duration of Dogs’ Sheep’s
autolysis (days) Tissues Tissues

0 16:2 21:0

7 16°6 20°4

21 15'8 21°6

42 16:2 224

91 16°8 —
168 -_ 22°8
224 — 22°4

This shows that ordinary fat is not hydrolysed, but that lecithin is
hydrolysed at 4°. These results indicate that the liver, spleen and kidney
contain a lipolytic ferment which attacks lecithin quite readily, but has no
apparent action on the simple glycerides of the higher fatty acids even when
presented in the form of the finest emulsion as in chyle, or as they exist in
the tissue itself as is seen in the autolytic experiments.

It is possible that the reason of this lies in the fact that lecithin can form
in water a colloid suspension, and so can be more readily attacked, but the
previous results show that lecithin can be readily split up even when pre-
sented to the tissue extracts in a very coarse form. The other and more
probable explanation would appear to be that when the glycerides of the
higher fatty acids are taken to the liver or other tissues they are converted
into lecithins and stored up free or in combination with the protein as the
invisible unstainable fat in the normal organ and in this form are easily
hydrolysed by the lipolytic ferment as required. This view seems the more
probable from the results of Leathes and Kennaway [1909], who showed that
the liver desaturates fats brought from the food or fat depots; this agrees
with the high iodine values of the lecithins, showing that the lecithins are

1
|

F. H. THIELE 295

formed from these desaturated fatty acids. Further, it is important to note

’

in the quantitative estimations of the “fats” in the liver, ete., that by far

the greatest amount is accounted for by the lecithins, nearly 90°/,.

D. Kaperiments to ascertain the effect of reaction on the
lipolytic power of the organs.
In each of these a double series of experiments was performed to ascertain
: the amount of higher fatty acids.

Pig’s Liver. 10 ¢.¢. of extract with 20 ¢.c. of egg yolk suspension.

1. Effect of Alkali.

Total Total
With g. Na,CO, (%/o) fatty acids Lower Higher fat °/, hydrolysed
0-05 (0:16) 14-4 - 6:0 8-4 40-8 20°7
O01 (0°32) 21-6 76 14:0 40°8 34°3
0-15 (0-5) 23-2 7-4 15°8 40°8 38°7
0:2 (0°66) 18-2 61 12°1 40°8 29°6
0-3 (1:0) 10°1 4:0 6-1 so 14°9
0-4 (1°33) 6-0 2-1 3°9 3 9-0

Same quantities as above, but 30 c.c. water added.

0-05 (0-08) 12:5 46 7-9 40-2: 19°6
0-1 (0°16) 20-0 60 14-0 Me 34°8
0-15 (0-24) 21-4 6:6 14:8 i 36°8
0-2 (0°33) 16-2 46 11°8 - 29°3
0:3 (0°5) 9°3 2-4 59 _ 14-6
0-4 (0-66) 5-0 2-0 5-0 * 7-4

: Thus it will be seen that the activity does not depend on the percentage
of alkali present, but on the total amount.

2. Effect of lactic acid. The quantities used were as before. Incubation 16 hours. The
amount of lactic acid is in c.c. of the decinormal solution.

Dog’s Liver. 10 c.c. of extract with 20 ¢.c. of egg yolk solution and all made up to 40 c.c.

Total fats Total fatty

C.c.N/10acid c¢.c.N/10alk acids found Lower Higher °/, hydrolysed
2 48-2 20-2 6:8 13-4 27°8
4 a5 23°8 79 15°9 32°9
6 rs 24:9 9°3 15°6 32-6
8 eC 23°1 12°4 10-7 22°2
10 Pa 20°8 13-0 78 16:2

Pigs Liver. Quantities as above.

2 48-2 186 81 10°5 21-8
4 4 23°8 9-7 141 29-2
6 i 24-2 10-4 9-8 20-3
8 ‘* 21-2 11:8 9-4 19°5
10 . _ -19°8 141 5-7 118

296 F. H. THIELE

SUMMARY.

From these experiments we can conclude that:
(1) The tissues possess a true lipolytic ferment.

(2) The lipolytic ferment, with the exception of the pancreatic lipolytic
ferment, can only hydrolyse phosphatides and jecorins, but not ordinary fats.
(3) The ferment is capable of acting in an alkaline or acid medium.

(4) There is no evidence of a kinase in the spleen.

Part of the expenses of this research were defrayed by a grant from the
Government Grant Committee, Royal Society, London.

REFERENCES.

Arthus (1902), J. Physiologie, 4, 56.
Cohnstein (1904), Ergebnisse der Physiologie, 3, 194.
Hanriot (1896), Compt. Rend. 123, 753, 833.

Kastle and Loevenhart (1900), Amer. Chem. J. 24, 491.
Leathes and Kennaway (1909), Lancet, 96.

Magnus Levy (1902), Beitrége, 2, 201.

Nencki and Liidy (1887), Therap. Monatsheft, 417.
Siebert (1900), Zeitsch. physiol. Chemie, 49, 50.

Slade (1903), Beitrdge, 3, 291.

Umber and Brugsch (1906), Arch. Exp. Path, 55, 164.

|
|
|
|

XXX. THE ESTIMATION OF TYROSINE IN
PROTEINS BY BROMINATION.

By ROBERT HENRY ADERS PLIMMER anp
ELIZABETH COWPER EAVES.

From the Institute of Physiology, University College, London, and the
Physiological Laboratory, University of Sheffield.

(Received April 16th, 1913.)

The tyrosine content of a protein is usually ascertained by hydrolysing
the protein with sulphuric acid, removing the acid as barium sulphate,
concentrating the solution and washings and then weighing the tyrosine
which separates out. The data so obtained by various workers are in some
cases very concordant, but it is generally considered that the real tyrosine
content of a protein is higher than these figures represent since it 1s
frequently impossible to isolate the whole of the tyrosine. Osborne and
Clapp [1906] and Osborne and Guest [1911] have particularly emphasised
this point and they consider that the tyrosine content of a protein so obtained
is only a minimal one, the true content being from one half to one per cent.
higher. The presence of cystine in the tyrosine isolated has also never been
taken into consideration. The two substances are so alike in their solubility
in water, in acids and in alkalies that their separation is a matter of con-
siderable difficulty [Plimmer, 1913]. Losses occurring in the isolation of the
tyrosine may be compensated for by admixture with cystine.

It was shown by J. H. Millar [1903] that tyrosine was readily brominated
and converted into dibromotyrosine, and that the amount of tyrosine in
a simple mixture of amino-acids could be accurately estimated by means of
this reaction.

A. J. Brown and E. T. Millar [1906] using this reaction showed that the
tyrosine was completely liberated at a very early stage in the hydrolysis of
edestin by trypsin. Their data gave the tyrosine content of edestin as
4°06 per cent., a figure which is considerably higher than that obtained
by direct isolation (2°1 per cent.). They made no estimations of the tyrosine

298 R. H. A. PLIMMER AND E. C. EAVES

content of other proteins. The higher value agrees with Osborne’s supposi-
tion that the tyrosine content of proteins as obtained by isolation is only a
minimal one and it seemed very desirable that further estimations should
be made by this method.

It was found that it was better to alter Millar’s method as it was not
sufficiently delicate for estimating small amounts of tyrosine such as are
obtained by the hydrolysis of proteins. The procedure of Brown and Millar
also involved a large deduction for the amount of bromine absorbed by the
control (protein or other decomposition products, possibly histidine) which
prevented an accurate estimation. The absorption of bromine by the
unchanged protein and its product of hydrolysis, histidine, which according
to Knoop [1908] reacts with bromine, has been eliminated by the use of
phosphotungstic acid. Another disturbing factor is the presence of trypto-
phane with the tyrosine amongst the products of hydrolysis. Though
tryptophane may be destroyed by boiling with acids, its decomposition
products still absorb bromine. The estimation of tyrosine amongst the
products of the acid hydrolysis of proteins was therefore impossible. It
remained to take advantage of the rapid and complete liberation of tyrosine
during the early stages of tryptic digestion and the slower liberation of
tryptophane, which is only complete after several days when a moderately
active trypsin solution is employed [Hopkins and Cole, 1901]. By estimating
the bromine absorption after short intervals of digestion in the filtrate from
the phosphotungstic acid precipitate, values for the tyrosine content of
proteins have been obtained which agree closely with those found by direct
isolation. |

Our work was withheld since we felt considerable diffidence as to the
exactness of our data, but since the publication of much higher figures for
the tyrosine content of proteins by Folin and Denis [1912] by a new colori-
metric method and since these figures have been criticised by Abderhalden
and Fuchs [1913] our data are of some value in support of the results of
Abderhalden and Fuchs. Folin and Denis do not seem to have taken
sufficient account of the presence of tryptophane and oxytryptophane, which
react with their reagent. If their figures express the total tryptophane and
tyrosine content, and our values the tyrosine content alone, then the
difference will give the tryptophane content of proteins which is at present
unknown and of considerable importance.

Died igs

R. H. A. PLIMMER AND E. C. EAVES 299

EXPERIMENTAL.

(1) The Estimation of Small Amounts of Pure Tyrosine
by Bromination.

J. H. Millar’s method for the estimation of tyrosine depends upon the
formation of dibromotyrosine when tyrosine is treated with nascent bromine ;
this is liberated by adding sodium bromate to an acid solution of potassium
bromide. The equations representing the reaction are :

NaBrO, + 5K Br + 6HCl = NaCl + 5KCl + 3Br, + 3H.O,
C,H,(OH).CH,.CH (NH.). COOH + 2Br,
= C,H.Br, (OH).CH,. CH (NH.). COOH + 2HBr,
from which we find that 1 g. of tyrosine absorbs 1°765 g. of bromine and
corresponds to 0°5558 g. of sodium bromate.

His procedure was to add 10-15 ce. of a 20 per cent. solution of potassium
bromide to a solution of tyrosine in hydrochloric acid and to titrate with
M/5 sodium bromate solution until it assumed a persistent deep yellow
colour. If the solution were coloured, starch and potassium iodide were
used as an indicator.

Millar found that 1°08 g. of bromine were absorbed by 1 g. of tyrosine,
a figure which is slightly higher than the theoretical value but of sufficient
accuracy to show that tyrosine can be estimated in solution by bromination ;
in one of the experiments 0°2 g. tyrosine required 3°8 cc. of the bromate
solution.

The estimation of smaller amounts of tyrosine than 0°2 g. was not
investigated by J. H. Millar. The amount of tyrosine in a protein does
not usually exceed 3 per cent. except in the case of caseimogen which
contains from 45-5 per cent. and silk-fibroin which contains about 10°5
per cent. One gram of protein would therefore usually give a solution
containing 0:01-0:04 g. tyrosine. To estimate 0°01 g. tyrosine, 0°19 ce. of
M/5 bromate solution would be required and an error of 0:05 cc. in the
titration would make an error of 25 per cent. in the estimation. It was
therefore necessary to ascertain if tyrosine could be brominated by a more
dilute solution of sodium bromate. Estimations were therefore made with
an N/5 (=M/30) solution of bromate in the same way as described by_
J. H. Millar, thus:

01911 g. tyrosine in 25 cc. hydrochloric acid solution required
21:05 ce. M/30 bromate solution ;
i.e. 1 g. tyrosine absorbs 1:75 g. bromine (theoretical 1°765).

Bioch. vit 20

300 R. H. A. PLIMMER AND E. C. EAVES

The bromination of tyrosine by the M/30 solution of sodium bromate is
slower than by the M/5 solution and it appeared that the method might
be more convenient if the bromination were effected by adding excess of
the sodium bromate solution to the tyrosine solution, allowing the reaction
to proceed for 10-15 minutes in a closed flask, adding potassium iodide and
titrating the excess of halogen with thiosulphate solution using starch as
indicator. 2 g. of tyrosine were dissolved in 500 cc. hydrochloric acid ;

50 cc. + 30 cc. M/30 NaBrO, solution: 7:1 ce. thiosulphate solution ;
22°9 cc. bromate required.

50 cc.+30cc. M/30 NaBrO, solution; 7:2 cc. thiosulphate solution ;
22°8 cc. bromate required.

i.e. 50 cc. contain 0°207 g. tyrosine.

The absorption of bromine for 1 g. of tyrosine is | $25 g., a value slightly
higher than the theoretical (1°765), but comparable with that found by
J. H. Millar (1°808). In subsequent experiments this high figure was not
obtained. It seems to have been due to the presence of a rather large
excess of bromate solution; the vapour in the flask appeared yellow and
loss occurred during the titration.

Small quantities of tyrosine can thus be accurately estimated by making
this alteration in the procedure.

(2) Estimation of Tyrosine in the Presence of Protein and its
Products of Hydrolysis.

It has been shown by A. J. Brown and E. T. Millar that proteins absorb
bromine under the conditions employed for the estimation of tyrosine by
J. H. Millar’s method, but this did not preclude the estimation of tyrosine
in the presence of unaltered protein if the amount of bromine absorbed by
the protein was deducted from the amount absorbed by the mixture of
protein and tyrosine. They showed further that there was no increase in
the absorption of bromine by gelatin during its digestion by trypsin. Gelatin
does not contain tyrosine or tryptophane, and since it was found that trypto-
phane did not absorb bromine under the same conditions it was concluded
that the increase in the absorption of bromine when edestin was digested by
trypsin was entirely due to the liberation of tyrosine. The absorption of
bromine due to the separation of tyrosine increased very rapidly and reached
a maximum in about one hour. The result gave the tyrosine content of

vw oe

eS

R. H. A. PLIMMER AND E. C. EAVES 301

edestin as 406 per cent. which differs very greatly from the value (271 per
cent.) obtained by the isolation and weighing of tyrosine. |

An examination was made for possible sources of error occurring during
the estimation of tyrosine by bromination.

Brown and Millar used M/5 sodium bromate solution and 50 ce. of a
1 per cent. solution of edestin.

0°80 cc. of the bromate solution was required for the bromination and
from this 0°42 ce. was deducted for the control. The difference of 0°38 ce.
gave the tyrosine content.

As shown above an error of 0°05 cc. in the titration, if M/5 sodium
bromate solution be used, corresponds to a difference of 25 per cent. in the
amount of tyrosine when estimating 0°01 g. of tyrosine, which is equiva-
lent to a variation of 1 per cent. in the tyrosine content of edestin. A
proportionate error must be considered when the deduction for the control
is made. The error in titration can be reduced by employing M/30 bromate
solution. An error of 0°23 cc. in the titration will now correspond to a
difference of 10 per cent. in the amount of tyrosine, which is equivalent to
a variation of 0°4 per cent. for the tyrosine content of edestin. Dilution
of the reagent will thus reduce the error but it is not in itself sufficient to
make the method an accurate one.

The amount which has to be deducted for the control is greater than the
amount used in the actual estimation; greater accuracy can therefore only
be obtained by reducing or eliminating this deduction.

The constituents of a protein which are known to absorb bromine are
tyrosine, tryptophane and histidine [Knoop, 1908]. According to Brown
and Millar tryptophane does not absorb bromine under the conditions
existing during the estimation of tyrosine so that it is most probably the
histidine which absorbs the bromine and necessitates the large deduction
for the control. Histidine can be removed by precipitation with phospho-
tungstic acid which leaves the tyrosine (and tryptophane) in solution in the
filtrate. Hence if the deduction for the control be due to the presence of
histidine, the absorption of bromine by the filtrate should be due solely to
the tyrosine contained in the enzyme preparation. A preliminary experi-
ment with gelatin showed that this was the only deduction necessary.

A 2 per cent. solution of trypsin was allowed to digest in the presence
of chloroform; at intervals 20 cc. were removed and placed in 50 ce. of
5 per cent. sulphuric acid + 20 cc. of 10 per cent. phosphotungstic acid.
To 50 ce. filtrate (= 0°333 g. trypsin) 10 cc. of sodium bromide (2 per cent.)
and 10 ce. of the sodium bromate solution (1 cc. = 0°012 g. Br) were added ;

20—2

302 R. H. A. PLIMMER AND E. C. EAVES

after 10-15 minutes 10 cc. sodium iodide (4 per cent. solution) were added
and the mixture titrated with sodium thiosulphate solution (1 cc. = 000663
g. Br) using starch as indicator’.

Time Thiosulphate required Bromine absorbed
0 15:1 ce. 0:02 g.
4 hours 11-1 0-046
6i 5 11:8 0-042

Be a Da 0-040

Taking the absorption after 6 hours, 1 g. of trypsin absorbs 0°126 g.
bromine.

A 2 per cent. solution of gelatin in 0:4 per cent. sodium carbonate
solution was digested in the presence of chloroform with 0:2 per cent. trypsin.
50 ce. samples were removed at intervals and placed in 50 ce. of 5 per cent.
sulphuric acid + 20 cc. of 10 per cent. phosphotungstic acid. 100 ce. of the
filtrate (= 0°833 g. gelatin) were treated with 10 cc. sodium bromide solution
+ 10 cc. sodium bromate solution (1 cc. = 0°012 g. bromine) and after half an
hour the mixture was titrated with sodium thiosulphate (1 cc. = 0:00663 g.
bromine) after adding sodium iodide and starch as indicator. Each sample
contains 0°833 2. trypsin.

Bromine absorbed Bromine

Thiosulphate Bromine by 0:0833 g. absorbed

Time required absorbed trypsin by gelatin
0 17-4 ce. 0-0046 -0°0025 0-0021
4 hours 16:2 0:0126 0:0104 0:0022
Cae 16°25 0-0126 0-0104 0-0022

24 ,, 15°6 — — —

The slight absorption by the gelatin (= 0:24 per cent.) is most probably
due to the presence of tyrosine; the sample (gold label) gave a distinct
reaction with Millon’s reagent.

This procedure was then applied to the estimation of the tyrosine in
caselnogen.

100 cc. samples of a 1 per cent. caseinogen solution were digested with
5 ec. of a trypsin solution for periods of 1-5 hours and were then precipi-
tated with 25 cc. phosphotungstic acid solution in hydrochloric acid; 50 ce.
of the filtrate were used for the titration:

1 In the presence of phosphotungstic acid it is better to use sodium bromide instead of
potassium bromide and sodium iodide instead of potassium iodide as potassium phosphotungstate
is precipitated when potassium salts are present and the precipitate interferes with the titration
with thiosulphate when starch is used as indicator. Auld and Mosscrop [1913] have main-
tained that starch and potassium iodide cannot be used when estimating tyrosine in digests of
protein by the Millar method. Colourless filtrates are obtained after precipitation with phospho-
tungstic acid and with the alteration in the procedure the disappearance of the blue colour when

the solution is titrated with thiosulphate is quite sharp. No difficulty has been experienced in
determining the end point under these conditions.

7
a
%

R. H. A. PLIMMER AND E.

C. EAVES

Time M/30 bromate
(hours) absorbed Difference
0 2°4 ce. —
1 3°8 14
2 4-1 17
3 4:6 2°2
4 4°6 2-2
5 4°55 2°15

3038

The maximum absorption of bromine occurred after 3 hours and then
remained constant. The percentage of tyrosine in caseinogen calculated
from the above difference is 5°08, a figure which agrees well with those
usually given (4°55).

The following duplicate experiments with another solution of caseinogen
show the reliability of the procedure’.

(1) (2)
Time Bromate Thiosulphate Bromate Bromate Thiosulphate Bromate
(hrs.) added required absorbed added required absorbed
0 5 ce. 9°65 ce. 0-2 cc. 5 cc. 9°55 ce. 0-25 ce.
0-5 s 7°45 1°3 5 1:3 1°35
1 x 5°95 2°05 a 5:9 2-05
2 “a 4:8 2°6 53 4°85 2-6
3 ss 4°35 2°85 a 4-35 2°85

The estimation of tyrosine in the phosphotungstic acid filtrate is thus
possible if no other products which absorb bromine are present. Gelatin
does not contain tryptophane and cystine and it only contains a small amount
of phenylalanine. Cystine and phenylalanine have been found not to
absorb bromine, but the behaviour of tryptophane, which according to Brown
and Millar does not react with bromine under the conditions adopted by
J. H. Millar, probably because of the presence of hydrochloric acid, required
further investigation as the method had been modified and as Dr Hopkins
had mformed us that tryptophane did react with bromine when treated
in this way. <A quantity of tryptophane was kindly supplied to us by
Dr Hopkins for this purpose and the following experiments were carried out:

00631 g. tryptophane was dissolved in hydrochloric acid and titrated
directly with sodium bromate by Millar’s method.

73 cc, bromate were required. Bromine absorbed = 0112 g.

0:0238 g. tryptophane was dissolved in hydrochloric acid: 10 ce, sodium
bromide and 10 ce. sodium bromate solution (= 1536 g. Br) were added: after

1 These experiments were carried out by one of us in conjunction with Mr 8. H. Wood and
the results were communicated to a meeting of the Physiological Society in March 1907. Con-
tinuation of the work was not then possible and no further experiments were made until 1909.
The results of these later experiments were communicated to the Biochemical Club in March 1912.

304 R. H. A. PLIMMER AND E. C. EAVES

15 minutes titrated with sodium thiosulphate solution (1 cc. = 0:00663 g. Br)
after adding sodium iodide and starch: 3:2 cc. were required. Bromine
absorbed = 0°1521 g.

0:0200 tryptophane was treated as in the previous experiment: 64 ce.
thiosulphate were required. Bromine absorbed = 01105 g.

0°57 g. tryptophane was dissolved in 1000 cc. water; 100 cc. of this
solution were used in each of the following experiments :

(1 ec. NaBrO;=0-012 g. Br and 1 ce. thiosulphate = 0:00663 g. Br.)
Titrated Thiosulphate Bromine

HCl NaBr NaBrO; after required absorbed
100 ce. 5 ce. 10 ce. 10 ce. 15 mins. 2:05 ce. 0:1064 g.

10 10 10 96 1:40 0:1107

15 10 10 s 2°30 0:1048

5 10 20 55 17:95 0-1210
10 10 20 30 mins. 17:0 0:1270

5 10 20 vs 16°5 0-1306
10 10 20 1°5 hrs. 15°05 0:1400

Another series of experiments gave similar results.

Tryptophane thus absorbs bromine under the conditions adopted by
Millar; it absorbs a greater amount of bromine under the modified con-
ditions, the absorption increasing with a larger amount of bromate and with
the time allowed for the reaction. The absorption corresponds to about
6 atoms of bromine by 1 molecule of tryptophane.

The presence of tryptophane in solution with tyrosine will thus interfere
with the estimation of tyrosine.

According to Hopkins and Cole tryptophane is destroyed by prolonged
boiling with acids.’ If its products of decomposition do not absorb bromine
the estimation of tyrosine should be possible after acid hydrolysis. Some
experiments were therefore made to see if tryptophane still absorbed bromine
after boiling with acids:

100 cc. of the above solution of tryptophane were heated for 10 hours
with excess of concentrated hydrochloric acid (50 cc.). 10 cc. NaBr + 20 ce.
NaBrO, were added and after 15 minutes titrated with thiosulphate (16°5 ce.
required). Bromine absorbed = 0:110 g.

0:058 g. tryptophane was dissolved in 100 cc. water: 10 cc. were titrated
directly: 10 cc. after boiling with concentrated hydrochloric acid and 10 ce.
after boiling with 25 per cent. sulphuric acid. The bromine absorbed was
respectively 0°0361 g., 0°0057 g. and 00106 g.

0:0998 g. tryptophane was dissolved in 50 cc. water: 10 cc. were titrated
directly and 10 ce. after boiling for 5 hours with hydrochloric acid. The
bromine absorptions were respectively 0°1017 g. and 0-074 g.

7 a teed

ae

R. H. A. PLIMMER AND E. C. EAVES 305

Absorption of bromine still occurs, but to a less extent, after boiling
tryptophane with acid, so that the estimation of tyrosine is not possible after
the hydrolysis of protein by acids.

Brown and Millar have shown that the whole of the tyrosine is very
rapidly liberated by the action of trypsin and our preliminary experiments
with trypsin and caseinogen have confirmed their results. It has been shown
by Hopkins and Cole that the liberation of tryptophane does not occur
rapidly with moderately active trypsin solutions and that its amount in
solution only reaches a maximum after several days. This difference in the
rate of liberation of the two substances may therefore permit of the estima-
tion of the tyrosine content of a protein when the bromine absorption is
measured at intervals during the digestion and when only those amounts
absorbed in the early stages, from 6-24 hours, are taken as a measure of the
tyrosine content.

(3) Absorption of Bromine during the Tryptic Digestion of Proteins.

Specimens of several animal and vegetable proteins have been procured
and examined by the method described above. The vegetable proteins were
most kindly sent to us by Prof. Osborne, the specimens in most cases being
the same as those in which he and his co-workers had determined the
tyrosine content by direct isolation and weighing.

In general, a 1 per cent. solution of the protein in 0°25 per cent. sodium
carbonate was digested in the presence of chloroform or carbon tetrachloride
with a 0-1 per cent. solution of trypsin. 50 ce. samples were removed im-
mediately and after various intervals of time and then precipitated with 100 ce.
of 10 per cent. phosphotungstic acid in 5 per cent. sulphuric acid. 100 ce.
of the filtrate were then taken for the estimation. 10 cc. sodium bromide
(2 per cent.) and 10 ce. sodium bromate solution were added ; after 15 minutes
the excess of bromine was displaced by adding 10 ce. of sodium iodide solu-
tion (4 per cent.) and the liberated iodine titrated with sodium thiosulphate
solution using starch as indicator. The estimations were generally made
after periods of 6 hours and 24 hours, since the amount of material at our
disposal did not allow of observations at more frequent intervals. These
times were chosen as the preliminary experiments with caseinogen showed
that the absorption began after about 1 hour and reached a maximum in
from 3-5 hours and that another rise sometimes occurred in about 24 hours.
The presence of tryptophane in solution was tested for by the bromine
reaction in some experiments; it was generally absent in the 6-hour period
but was faintly visible in the 24-hour period.

306 R. H. A. PLIMMER AND E. C. EAVES

The estimation with silk-fibroin was performed after hydrolysis with
20 per cent. sulphuric acid for 18 hours as trypsin has only a very shght
action upon this protein. Silk-fibroin does not contain tryptophane.

Peptone Roche, which is prepared from silk-fibroin by acid hydrolysis,
gave only a slight precipitate with phosphotungstic acid, 1t seems to contain
tyrosine or a polypeptide containing tyrosine which reacts with bromine.
Like silk-fibroin this protein contains no tryptophane.

It was impossible to estimate the tyrosine content of the alcohol soluble
proteins—the gliadins—since they are only digested with extreme slowness
by trypsin.

The following are the data:

Trypsin. (Used in experiments with caseinogen, ‘‘ peptone Roche.”)

1 per cent. in 0°25 per cent. NaygCO, solution digested in presence of carbon tetrachloride.
50 ce. samples in 100 cc. phosphotungstie acid solution. 100 ce. filtrate (=0-33 g. trypsin) for

estimation. 5 cc. sodium bromate solution (=0:0768 g. Br), ‘Titrated with thiosulphate
(1 cc. =0:00672 g. Br).

Time Thiosulphate Br
(hours) required absorbed
0 8:3'¢e: 0:0210 g.
6 6:3 0:0345
7 6°65 0:0321
24 6:7 0:0318

Caseinogen. (Hammarsten.)

12-233 g. in 487:27 cc. of 0:25 NayCOs solution + 122-33 ce. trypsin solution+2 cc. CCl.
50 cc. samples in 100 cc, phosphotungstic acid solution. . 100 cc. filtrate (=0°667 g. caseinogen
+0:0667 g. trypsin). 10 cc. sodium bromate solution (=0:1536 g. Br). Titrated with thio-
sulphate (1 ec. =0-00672 g. Br),

Thiosulphate Total Br Br absorbed Tyrosine

required, absorbed, after deducting content

Time ec: g. 0:00642 for trypsin per cent.

0 22°9 — —

5 hours 12°5 0:0696 0-0632 5°35
aueE 12°65 0-0695 0-0631 5°34
BB) Sas 10°75 00814 0:0749 6°35
SO, 7-4 0°1036 0:0972 8°23
48 ,, 79 0°1005 0-0941 7°98
4 days 7°55 0:1036 0-0972 8°23

‘© Peptone Roche.”

5735 g. in 514-2 ce. of 0°25 NagCO, solution + 57:3 ce. trypsin solution+2 ce. CCly. 50 ce.
samples in 100 ce, phosphotungstic acid solution. 100 cc. filtrate (=0°333 g. peptone + 0-0333 g.
trypsin) for estimation. 10 cc. sodium bromate (=0'1536 g. Br). ‘Titrated with thiosulphate
(1 cc. =0-00672 g. Br).

Brabsorbed after
Thiosulphate Total Br deducting 00021 Tyrosine
required, absorbed, at 0 hrs. & 0:0032 content
Time cc: g. afterwards fortrypsin per cent.
0 14°85 0:0998 0:0538 8°79
2°5 hours 14:2 0:0954 0°0582 9°32
3°75 ,, 13°5
Fads 13-1]... eee pe .
oe 13-3734 0-090 0-0636 10-23

94 ee 13-6

R. H. A. PLIMMER AND E. C. EAVES

Silk-fibroin.

4°39 g. were hydrolysed with 20 per cent. H,SO,, and the solution made up to 500 cc.
100 ce. filtrate (=0°293 g. fibroin) for estimation.
Titrated with thiosulphate (1 cc.=0-00672 g. Br).
Tyrosine content = 9°53 per cent.

samples in 100 ce, phosphotungstie acid,
5 ce. bromate solution (=0-0768 g. Br).
Thiosulphate required=4:1 ce. Br absorbed=0-0493 g.

Conglutin (Merck) containing 5-2 per cent. moisture.

5°6158 g. in 222-7 ce. NagCO, solution + 56:1 ce. trypsin solution +2 ec. CCl,.
100 ce. filtrate (=0°632 g. dry conglutin) for estimation.
Titrated with thiosulphate (1 ce. =0-00672 g. Br)

in 100 ce. phosphotungstie acid.
5 cc. bromate (=0-0768 g. Br).

Thiosulphate Total Br Br. absorbed
required, absorbed, after deducting
Time ce. g. 0-00642 for trypsin

0 It % — —

4 hours 8°55 0:0193 0-0129
eee 8°3 0-0210 0-0146
24 ,, 5°35 00415 0-0351

Excvelsin. (Own preparation.)

307

50 ce.

50 cc. samples

Tyrosine
content
per cent.

2-697 g. in 306 ce. of 0°25 per cent. NaOH + 2 cc. CHCl]; + 2 ‘‘holadin” capsules. 50 cc. samples

in 100 ce. phosphotungstie acid.

Total Br Br absorbed
absorbed, after deducting
Time g. 0-002 for holadin
0 0:00287 0:0008
5:5 hours 0-0149 0°0129
24 Fe 0-0258 0-0238
28 5 0-026 0-024
AS: oy 0:0282 0:026

Legumin (Osborne) containing 4:07 per cent. moisture.

29866 g. in 266°6 ce, Na,COs solution + 30 cc. trypsin solution +2 ce.
100 ce. filtrate (=0-03198 g. dry legumin + 00033 g. trypsin)

in 100 cc. phosphotungstie acid.

100 ce. filtrate (=0°2937 g. excelsin) for estimation.

Tyrosine
content
per cent.

2°5
4-6
4°61
50

CCl,. 50 ce. samples

for estimation. 5 cc. bromate (=0°08375 g. Br), Titrated with thiosulphate (1 cc. =0-0079 g. Br).

Thiosulphate Total Br Br absorbed Tyrosine
required, absorbed, after deducting content
Time cc. g. 0:0033 g. for trypsin per cent.
0 10°45 0-0012 — _-
6°75 hours 8°15 0-0185 0°0152 2-69
27°95; 70 0-0272 0-0239 4-23
3 days 6-0 0°0351 0:0318 5°63

Edestin (Osborne) containing 13-07 per cent. moisture.

4°6052 g. in 182-3 cc. NagCO; solution +46 ce. trypsin solution +2 ce.
100 ce. filtrate (=0°58 g. dry edestin +0-:0667 g. trypsin) for
Titrated with thiosulphate (1 cc. =0-00672 g. Br).

in 100 ee. phosphotungstie acid.

CCl,. 50 cc. samples

estimation. 5 cc. bromate (=0-0768 g. Br).
Thiosulphate Total Br Br absorbed Tyrosine
required, absorbed, after deducting content
Time ce. g. 0-00642 g. for trypsin per cent.
0 11:4 — — =
4 hours 8°35 0:0207 ~ 0°0143 1-4
Ghee 7°85 0-0241 G:0177 1:73
> ae 5-0 0-0432 00368 3°6
452-35 2°6 0:0593 0-0529 ie bry

308 R. H. A. PLIMMER AND E. C. EAVES

Vignin (Osborne) containing 7:03 per cent. moisture.

1-8982 g. in 169 ce. NagCO; solution + 18-9 ce. trypsin solution+2 ec. CCly. 50 cc. samples
in 100 ce. phosphotungstic acid. 100 ce. filtrate (=0°310 g. dry vignin + 0-033 g. trypsin) for
estimation. 5 cc. bromate (=0°08375 g. Br). Titrated with thiosulphate (1 ec. =0:0079 g. Br).

Thiosulphate Total Br Br absorbed Tyrosine
required, absorbed, after deducting content
Time ce. g. 0:0033 g. for trypsin per cent.
0 10°9 — aes mate
6°75 hours 7:95 0:0223 0:0190 3°4
Zio mss 6°35 0:0360 0:0327 5°97

Squash Seed Globulin (Osborne) containing 10-02 per cent. moisture.

1-912 g. in 170°1 ee. NayCOs solution + 19-1 ce. trypsin solution +2 ec. CCly. 50 ce. samples
in 100 ce. phosphotungstie acid. 100 cc. filtrate (=0°3 g. dry protein +0°0333 g. trypsin) for
estimation. 5 ce. bromate (=0°08375 g. Br). Thiosulphate (1 c.c.=0-0079 g. Br).

Thiosulphate Total Br Br absorbed Tyrosine
required, absorbed, after deducting content
Time Cc: g. 0:0033 g. for trypsin per cent.
0 10:8 — = =
5°75 hours 8-2 0:0205 0:0172 3°24
27 ) 65 0:0339 0:0306 5'8

Amandin (Osborne) containing 10°46 per cent. moisture.

3:3404 g. in 298°6 cc, NagCOs solution + 33°4 ce. trypsin solution+2 ce. CCly. 50 cc. samples
in 100 ce. phosphotungstie acid. 100 ce. filtrate (0°3 g. dry amandin +0:°0333 g. trypsin) for
estimation. 5 cc. bromate (=0-08375 g. Br). Thiosulphate (1 cc.=0-0079 g. Br).

Thiosulphate Total Br Br absorbed Tyrosine

required, absorbed, after deducting content

Time Oe g. 0:0033 g. for trypsin per cent.
0 10°6 — _ --
6 hours 8:2 0°:01896 0°0157 2°9
D4 6°75 0:0304 0°0271 5-1

Glycinin (Osborne) containing 9:06 per cent. moisture.

1:9166 g. in 170-5 ce. NagCO3 solution + 19-2 ce. trypsin solution +2 ec. CCly. 50 ec. samples
in 100 ce. phosphotungstic acid. 100 ce. filtrate (=0°302 g. dry glycinin + 0°0333 g. trypsin) for
estimation. 5 cc. bromate (=0°08375 g. Br). Thiosulphate (1 ec. =0-0079 g. Br).

Thiosulphate Total Br Br absorbed Tyrosine
required, absorbed, after deducting content
Time ce. g. 0:0033 g. for trypsin per cent.
0 10°4 — = =
6 hours " 9:3 0:0091 0:00058 st
28) 5; ou 0:0218 0:0185 3°47

In most cases there was an increase in the bromine absorption between
the periods of 6 hours and 24 hours. The exact time at which the increase
occurred could not be ascertained since it was impossible to make more
frequent determinations owing to the scarcity of material. An increase after
a constant period would show the point when all the tyrosine was liberated
and that at which the tryptophane became set free. There is a very close

R. H. A. PLIMMER AND E. C. EAVES 309

agreement between the figures for the tyrosine content after the 6-hour
interval and the figures obtained by isolation and weighing, as is shown by
the following table :

Percentage of Percentage
tyrosine after by

Protein 6 hours’ digestion weighing
Caseinogen 5°34 4-5
** Peptone Roche” 10°23 —
Silk-fibroin 9°53 9-10°5
Conglutin 1-31 2-1
Legumin 2°69 2-4
Edestin 1-73 271
Vignin 3-4 2-3
Squash seed globulin 3°24 3°1
Amandin 2°9 ret
Glycinin tical 19
Excelsin 2°5 ay

The correspondence in the figures in the cases of edestin, glycinin, squash
seed globulin, excelsin, legumin and silk-fibroin is very close. The value
is 1 per cent. higher for caseinogen, and the value is also higher for amandin
and vignin. The result should be slightly higher for excelsin as the material
was not dried. The amounts of amandin and glycinin available were very
small, so that much stress cannot be placed on these figures.

The method of bromination therefore appears to be of use for the
estimation of the tyrosine content of proteins if measurements of the
absorption are made at frequent intervals during a tryptic digest of the
protein, but it must be used with precautions and the figures carefully
criticised.

SUMMARY.

can be

The estimation of smal] quantities of tyrosine—0'01-004 g.
effected by J. H. Millar’s method of bromination, when a more dilute solution
of sodium bromate is used, but it is preferable to modify his procedure by
adding excess of the reagent and titrating the non-absorbed halogen with
thiosulphate solution, using potassium iodide and starch as indicator.
Tyrosine cannot be directly estimated by bromination in the presence of
protein and its decomposition products, since histidine and tryptophane
also absorb bromine. Histidine can be removed by precipitation with
phosphotungstic acid. The absorption of bromine by tryptophane is not
completely eliminated after boiling with acids so that tyrosine cannot be
estimated by this method in solutions containing the products of acid
hydrolysis of proteins which contain tryptophane. Values for the tyrosine

310 R. H. A. PLIMMER AND E. C. EAVES

content of proteins, agreeing with those obtained by isolation and weighing,
are obtained when the bromine absorption of a tryptic digest is measured

after an interval of about 6 hours.
The expenses of this research were in part defrayed by a grant to one of
us from the Government Grant Committee of the Royal Society.

REFERENCES.

Abderhalden and Fuchs (1913), Zeitsch. physiol. Chem. 83, 468.

Auld and Mosscrop (1913), J. Chem. Soc. 103, 281.

Brown, A. J. and Millar, E. T. (1906), J. Chem. Soc. 89, 145.

Folin and Denis (1912), J. Biol. Chem. 12, 245.

Hopkins and Cole (1901), J. Physiol. 27, 418.

Knoop (1908), Beitrige, 11, 356.

Millar, J. H. (1903), Trans. Guinness Research Laboratory, 1, Part 1, 40.
Osborne and Clapp (1906), Amer. J. Physiol. 17, 246.

and Guest (1911), J. Biol. Chem. 9, 388.

Plimmer (1913), Biochem. J. 7, 311.

en

ee

XXXI. THE SEPARATION OF CYSTINE AND
TYROSINE.

By ROBERT HENRY ADERS PLIMMER.

From the Ludwig Mond Research Laboratory for Biological Chemistry,
Institute of Physiology, University College, London.

(Recewed April 12th, 1913.)

Cystine and tyrosine resemble each other so closely in their solubility in
water, in alkali and in acid, that their separation is not easily effected. Both
Morner [1899, 1901] and Embden [1900] who first isolated cystine from the
products of the acid hydrolysis of scleroproteins obtained it mixed with
tyrosine. Mé6rner effected the separation of the two substances by fractional
crystallisation from ammonia; Embden by dissolving out the tyrosine with
very dilute nitric acid. Friedmann [1902] separated tyrosine from cystine
by solution in ammonia and neutralisation with acetic acid ; tyrosine
crystallised out in the neutral solution and the cystine was precipitated by
making the filtrate strongly acid with acetic acid. Cystine is usually pre-
pared from the products of acid hydrolysis of proteins by nearly neutralising
with soda and allowing to crystallise. Folin [1910] described the isolation
of cystine and tyrosine by neutralising the hydrochloric acid hydrolysis
solution to Congo-red with sodium acetate ; cystine crystallises out and on
diluting the filtrate and allowing it to stand the tyrosine separates out slowly.
Cystine thus seems to be the more insoluble in dilute mineral acids and in
strong acetic acid.

Neither Mérner’s, Embden’s, nor Friedmann’s method gives a quantitative
separation of the two substances; they only permit of the isolation of a
portion of the mixture. Folin’s method of preparation was devised for
obtaining a quantity of cystine. A method for the separation of these two
constituents of a protein is therefore required. On account of their similarity
the tyrosine isolated from the products of hydrolysis may contain cystine
unless this unit has been completely decomposed during the process of

312 R. H. A. PLIMMER

hydrolysis". For this reason the data of the actual amount of cystine in
most proteins are scanty or lacking and except in the scleroproteins and in
the protein described by Kotake and Knoop [1911] the relative amount of
cystine to tyrosine is very small. The estimations of tyrosine which have
been made by Plimmer and Miss Eaves [1918] by the method of bromina-
tion are not greatly higher than the amounts of tyrosine obtained by direct
isolation and weighing. The presence of a small amount of cystine in the
tyrosine may explain the slight differences.

Winterstein [1901] described the precipitation of cystine by phospho-
tungstic acid and Hopkins and Cole [1902] its precipitation by a solution of
mercuric sulphate in five per cent. sulphuric acid. Tyrosine is not precipi-
tated by phosphotungstic acid and its mercury compound is soluble in the
five per cent. sulphuric acid. Cystine and tyrosine have been found to differ
very greatly in their behaviour to absolute alcohol saturated with hydrochloric
acid gas. Tyrosine is readily esterified and goes into solution in the acid
alcohol ; cystine is not readily esterified and is only very slowly dissolved and
the portion which goes into solution can be precipitated by adding an equal
volume of absolute alcohol. Neither of these differences has hitherto been
used for the specific purpose of separating the two compounds and their
applicability has therefore been tested.

EXPERIMENTAL.

Tyrosine was prepared by the hydrolysis of caseinogen by trypsin. The
products which crystallised out during the digestion and on the concentration
of the filtered solution were mixed and purified by repeatedly dissolving in
dilute sulphuric acid and exactly neutralising with caustic soda, 14 g.
of pure tyrosine being thus obtained from 600 g. of caseinogen. The sub-
stance gave no precipitate with phosphotungstic acid showing the absence of
diamino-acid (diaminotrihydroxydodecanic acid) and of cystine.

Cystine was prepared from wool and hair by Folin’s method; the final
product was not perfectly white but it consisted of the typical hexagonal
plates.

1 Cystine is rapidly decomposed by boiling with alkali with the formation of hydrogen
sulphide (loosely bound sulphur) and it is also decomposed by prolonged boiling with acid. In
preparing cystine from wool or hair the best yield was obtained when the material was boiled
with concentrated hydrochloric acid for 3-4 hours; the yield was very poor when the boiling
lasted from 5-8 hours. In purifying the cystine by boiling with charcoal in acid solution loss
also occurs. If a solution of cystine in dilute hydrochloric acid be boiled for a long time it
becomes yellow or yellow-brown in colour.

R. H. A. PLIMMER 313

(1) Separation by means of Phosphotungstic Acid.

Cystine and tyrosine were mixed together in different proportions and
dissolved in 50 ce. of 5 per cent. sulphuric acid. A 30 per cent. solution of
phosphotungstic acid in 5 per cent. sulphuric acid was added so long as a
precipitate was formed. (20cc. sufficed for the precipitation of 0° g. of
cystine.) After standing for 12 hours the precipitate was filtered off and
washed repeatedly with a 2°5 per cent. solution of phosphotungstic acid in
5 per cent. sulphuric acid.

The cystine was recovered from the precipitate either by suspending it in
water or in water containing acetone as recommended by Wechsler [1911]
and adding baryta water until the solution remained permanently alkaline to
phenolphthalein, the decomposition being carried out on the water bath.
Excess of baryta was carefully avoided so as to prevent decomposition of the
cystine. The filtrate from the barium phosphotungstate was acidified with
hydrochloric acid, evaporated to a small volume on the water bath, and neu-
tralised with ammonia. The cystine crystallised out and after being left to
stand for 1-2 days was filtered off, washed, dried and weighed.

The recovery of the cystine from its phosphotungstate by decomposition
with hydrochloric acid and extraction of the reagent with ether (Winterstein)
was also attempted; the whole of the reagent was not dissolved by the ether
and the recovered cystine was contaminated with phosphotungstic acid.
Experiments were not made with amyl alcohol which Jacobs [1912] recom-
mended as a solvent for extracting phosphotungstic acid.

The tyrosine was recovered from the filtrate by adding ammonia to
remove the excess of phosphotungstic acid, filtering off the ammonium phos-
photungstate, neutralising and evaporating to a small volume. The crystals
so obtained were washed with water to remove ammonium sulphate and the
tyrosine residue was dried and weighed. This procedure’ was preferred to
the usual method of removing excess of acid with baryta which entails re-
peated extraction of a bulky precipitate of barium phosphotungstate and
sulphate with hot water and the evaporation of a large volume of liquid.

The amounts of cystine and tyrosine taken and recovered were as

follows :
Taken Recovered
aaa —— ea Vz SS =F
Cystine Tyrosine Cystine Tyrosine
0°5 g. 0°5 g. 0°35 g. 0°35 g.
0: 1-0 0°47 0-6
0-5 1:0 0°45 0°55
1:0 0-5 0-80 0-25
1:0 0-5 0°82 0-4
05 = 0:27 ae
0°5 = 0:3 a

314 R. H. A. PLIMMER

Neither the cystine nor the tyrosine was completely recovered. The
cystine is precipitated practically completely by the phosphotungstic acid ;
the loss seems to take place in its recovery from the phosphotungstic acid
precipitate ; the solution must be made slightly alkaline to ensure complete
decomposition of the phosphotungstate and some of the cystine is most
probably also decomposed by the alkali. Some decomposition may also occur
during the evaporation of the acid solution. Except in the last experiment
the cystine always contained tyrosine and this accounts for the loss of
tyrosine. The cystine was only obtained free from tyrosine in the last
experiment in which the precipitate was washed some twenty times by
removing it from the filter, stirring up with the washing reagent and again
filtering until the washings showed no reaction with Millon’s reagent: over
80 per cent. of the tyrosine was then recovered.

2 g. of the cystine recovered from the earlier experiments were found to
contain 0°6 g. of tyrosine which was isolated by means of alcohol saturated
with hydrochloric acid (as described below).

(2) Separation by means of Mercuric Sulphate.

Mixtures of cystine and tyrosine in various proportions were made and
dissolved in 5*per cent. sulphuric acid and treated with mercuric sulphate
dissolved in 5 per cent. sulphuric acid (Hopkins and Cole's tryptophane reagent)
until no further precipitate occurred. The precipitate was filtered off after
standing for 12 hours and washed repeatedly with 5 per cent. sulphuric acid
by removing from the filter, stirring up with the acid and again filtering
until the washings gave no reaction with Millon’s reagent.

The cystine was recovered from the precipitate by suspending in water
and decomposing with hydrogen sulphide. The filtrate from the mercuric
sulphide was evaporated on the water bath to a small volume and then
neutralised with ammonia. The cystine crystallised out and was filtered off,
washed, dried and weighed. The acid filtrate containing the tyrosine was
evaporated on the water bath to about 400 cc. and filtered from the mercuric
sulphate which had separated out. A slight excess of ammonia was added
and after again filtering the solution was evaporated almost to dryness. The
crystals so obtained were filtered off and washed with water until free from
ammonium sulphate and the residue of tyrosine was dried and weighed.
This procedure was preferred to the removal of the sulphuric acid with baryta
which would have necessitated the repeated extraction of the insoluble barium
sulphate with boilinggwater.

4

R. H. A. PLIMMER 315

As with the previous method the amounts of cystine and tyrosine re-

covered were far from quantitative as is shown by the following figures:

Taken Recovered

Cystine Tyrosine Cystine Tyrosine
0°5 g. — 0°30 g. —
0:5 — 0-31 —
0-5 0-5 g. 0-26 0°45 ¢g.
O°5 0-5 0°28 0°47
0-5 1:0 O31 0°75
0-5 1-0 0°15 0-96

The loss of tyrosine was apparently less than that of cystine but the
tyrosine was very impure and contained a brown pigment arising from the
decomposition of cystine. On further investigation the precipitation of
eystine by mercuric sulphate in 5 per cent. sulphuric acid was found to be
incomplete, as was shown by an estimation of the nitrogen in an experiment
with cystine alone:

1 g. of cystine was dissolved in 100 ce. of 5 per cent. sulphuric acid; 20 ce.
were found to contain 0°0224 g. N by Kjeldahl’s method. The remaining
80 cc. (=0:0896 g. N) were precipitated with 21 cc. mercuric sulphate
solution. 70 ce. of the filtrate contained 0-0172 g. N.

Hence the amount precipitated was 00724 g. or 81 per cent. of the cystine.
The amount of cystine recovered from the precipitate by the procedure
described above was 0°3 g. instead of 0°6 g. Loss occurs not only in the
precipitation but also in evaporating the solution before neutralising with
ammonia. Cystine is much more unstable to acid than one is led to expect
from the description of its isolation.

(3) Absolute alcohol saturated with hydrogen chloride.

Whilst preparing tyrosine ethyl ester from some tyrosine it was observed
that complete solution of the material could not be effected and an ex-
amination of the insoluble residue showed it to be cystine; 5 g. of the
material yielded 0:05 ¢. cystine and 4 g. yielded 0:2 g. cystine

This difference in the behaviour of cystine and tyrosine suggested a
simple method for effecting their separation.

Preliminary experiments were made with pure cystine and pure tyrosine ;
05 g. cystine was covered with 20 cc. absolute alcohol saturated with
hydrogen chloride, warmed on the water bath and allowed to stand for
12 hours. The undissolved substance was filtered off, washed with absolute.
alcohol, dried and weighed. Yield=039 g. A white precipitate was pro-
duced when the wash alcohol came into contact with the filtrate. This

Bioch, vit 21

316 R. H. A. PLIMMER

precipitate was filtered off, washed, dried and weighed. Yield =0-11 g.
On dissolving a test portion of each of these quantities in ammonia and
allowing to crystallise the typical hexagonal plates characteristic of cystine
were formed,

0:5 g. tyrosine was covered with absolute alcohol saturated with hydrogen
chloride and warmed on the water bath. Complete solution readily occurred
and the tyrosine was converted into its ethyl ester. On adding water and
neutralising with sodium carbonate no tyrosine was precipitated, but on
acidifying and boiling for 4-5 hours the ester was hydrolysed and on again
neutralising with sodium carbonate tyrosine was precipitated. It was
filtered off, washed, dried and weighed. Yield = 0-45 g. <A further 0°05 g.
was obtained on acidifying the filtrate, boiling, and neutralising once more.

A mixture of 0°5 g. of each was then treated in the same way with 30 ce.
absolute alcohol saturated with hydrogen chloride. 0°35 g. was insoluble
and the alcohol used for washing the residue precipitated an additional 0:1 g.
The filtrate after hydrolysis of the tyrosine ester yielded 0°47 g. tyrosine.

In these experiments the cystine was recovered in two fractions, but if an
equal volume of absolute alcohol be added to the absolute alcohol saturated
with hydrogen chloride before filtering the whole of the cystine can be ob-
tained in one operation. The cystine which goes into solution seems to be
cystine hydrochloride, for it dissolves in water. The insoluble portion seems
to consist mainly of cystine but a portion of it dissolves if it be washed with
water, and is apparently cystine hydrochloride as the cystine is precipitated
in hexagonal plates on neutralising with ammonia. ‘There is no evidence
that the cystine is:converted into its ester. According to Friedmann cystine
is not easily converted into its ethyl ester but its methyl ester is more easily
obtained (Fischer and Suzuki).

The separation of cystine and tyrosine by this method has been tested by
the following experiments :

Taken recovered
a a stad [avi GE Wirz ae > ae
Cystine Tyrosine Cystine Tyrosine

0°5 g, 0-0 g. 0:47 g. 0-0 g.
0°5 0-0 0:50 0:0
0:0 0°5 0:0 05
0-0 0°5 0:0 0°49
0°5 0°5 0°45 0°47
0°5 1:0 0:49 0-96
1:0 0:5 1-00 0-50

The mixtures were treated with absolute alcohol saturated with hydrogen
chloride and warmed on the water bath. An equal volume of alcohol was
added and the insoluble cystine filtered off, washed, dried and weighed. The

R. H. A. PLIMMER 317

filtrate was diluted with 2 volumes of water and boiled for 8 hours, water being
added when necessary. ‘Tyrosine was precipitated on neutralising; it was
filtered off, washed, dried and weighed.

The usefulness of this method is illustrated by the first experiment
in which presumably pure tyrosine prepared from wool by Folin’s method
had been used. The presence of cystine was not observed by microscopic
examination and the cystine present was found to be unevenly distributed.

The cystine recovered from the phosphotungstate precipitate above men-
tioned contained tyrosine; 2 g. contained 1:4 g. cystine and 0:6 g. tyrosine.

A mixture weighing 3 g. was found to contain 19 g. cystine; the
tyrosine was unfortunately lost. This mixture actually contained 2 g. cystine
and 1 g. tyrosine.

SUMMARY.

1. Cystine and tyrosine can be separated from one another by precipi-
tation with phosphotungstie acid. The precipitation of cystine is almost
complete, but loss occurs in its recovery from the precipitate. Almost the
whole of the tyrosine can be recovered from the filtrate and washings.

2. Cystine and tyrosine can be separated from one another by precipi-
tation with mercuric sulphate in five per cent. sulphuric acid. The cystine
is not completely precipitated and the tyrosine which is recovered is impure.

3. Cystine and tyrosine can be completely and quantitatively separated
by means of absolute alcohol saturated with hydrogen chloride. The
tyrosine is rapidly converted into tyrosine ester and goes into solution.
It can be recovered by boiling the solution when diluted with water for eight
hours and then neutralising with ammonia. Almost the whole of the cystine
is insoluble; the portion which goes into solution (perhaps cystine hydro-
chloride) is precipitated by adding an equal volume of absolute alcohol. The
eystine is not converted into its ethyl ester since on dissolving the insoluble
portion in dilute hydrochloric acid and neutralising with ammonia the cystine
is precipitated in the typical hexagonal plates.

REFERENCES.

Embden (1900), Zeitsch. physiol. Chem. 32, 94.
Folin (1910), J. Biol. Chem. 8, 9.

Friedmann (1902), Beitrdge, 3, 1.

Hopkins and Cole (1902), J. Physiol. 27, 418.
Jacobs (1912), J. Biol. Chem. 12, 429.

Kotake and Knoop (1911), Zeitsch. physiol. Chem. 75, 488.
Morner (1899), Zeitsch. physiol. Chem. 28, 595.
(1901), Zeitsch. physiol. Chem. 34, 207.
Plimmer and Eaves (1913), Biochem. J. 7, 297.
Wechsler (1911), Zeitsch. physiol. Chem. 73, 188.
Winterstein (1901), Zeitsch. physiol. Chem. 34, 153.

XXXII. THE | FACTORS “CONCERNED? siti
THE SOLUTION AND PRECIPITATION OF
EUGLOBULIN.

By HARRIETTE CHICK.

From the Lister Institute.

(Received April 15th, 1913.)

CONTENTS.

PAGE
I. Introduction . 5 : : 5 . : : : : : 318
II. ‘‘Solution” of euglobulin by acids and alkalies : : : 319
Ill. “Solution” of euglobulin by neutral salts : 322

IV. Effect of neutral salts upon acid and alkaline dispersions of
euglobulin ; ; F j : : : f ‘ : 332
V. Summary i. : ; : ; : : : : 3 ; 338

I. INTRODUCTION.

There is no reason for regarding euglobulin, the material precipitated
from serum by dialysis or by dilution and acidification, as a chemical entity.
Michaelis and Rona (1910, 2] are of opinion that it consists of that portion
of the total proteins which owes its solution to the dispersing power of the
electric charge upon its particles. It is true that the procedure adopted
for separating euglobulin is one which renders it iso-electrie with the
solution; but the conditions determining its dispersion in serum are not
so simple as they suggest. Euglobulin, as shown by Hardy [1905], is dis-
solved by neutral salts, e.g., sodium chloride, in a concentration of 1/10th
normal, to form a colloidal solution in which the particles are electrically
neutral, a result which I have been able to confirm (see p. 329, below).
Serum contains enough salt to produce this effect and unless it is at the
same time diluted, the euglobulin it contains cannot be completely pre-
cipitated by the addition of the amount of acid necessary to render it
iso-electric. The dispersion of euglobulin in serum is due to two entirely
distinct causes: (1) Electric charge on the particles owing to the alkalinity
of the fluid; (2) Formation of a “soluble” compound with the salt present,

H. CHICK 319

concerning the mechanism of which there are a variety of theories, which
are discussed in detail below.

By the process of “heat-denaturation” proteins acquire many of the
properties peculiar to euglobulin, as regards the factors conditioning their
solution (dispersion) and agglutination (see Hardy [1900], Michaelis and
Mostynski [1910] and Chick and Martin [1912]). The results of the last-
named, however, afford no support for the opinion expressed by Starke
[1900] and Moll [1904] that “albumin” is converted into “globulin” by
heating. Abderhalden [1903, 1904] has found evidence of difference in
chemical composition between the two sets of proteins which is not obli-
terated after the former has been heated. Further, the analogy mentioned
above is by no means complete; the union of euglobulin with neutral salts
to form a dispersion that is without drift in an electric field has no parallel

in the case of denaturated serum-proteins.

II. SoLutTion oF EUGLOBULIN BY ACIDS AND ALKALIES, AND
THE ISO-ELECTRIC POINT.

Solution of euglobulin by means of acid or alkali was shown by Hardy
[1905] to be associated with the possession by the dispersed particles of
an electric charge which was respectively positive or negative in sign.

Michaelis and Rona (1910, 2) found the iso-electric point of euglobulin
to be at a concentration of hydrogen-ions equal to 36 x 10~ normal and
to coincide with the point of optimum flocculation for this protein.

The iso-electric point has been re-determined in the present instance
and the result of Michaelis and Rona has been confirmed.

The euglobulin was prepared as follows: horse serum was diluted ten
times with distilled water and the globulin was precipitated by acidifying
with acetic acid (about 3-4 cc. N acetic acid per litre according to the
original reaction of the serum). The precipitate was allowed to settle, was
centrifuged off, and purified by dissolving in a minimal amount of standard
sodium hydroxide solution (according to the amount of the precipitate) and
reprecipitating with hydrochloric acid, the precipitate again being separated
by centrifuging. This operation was repeated once or twice and the pre-
cipitate finally washed with distilled water. A fairly concentrated suspension
was made in distilled water from which, by dilution, the material was
prepared which was used for the various experiments’. The particles of

1 In some experiments, for example those in Tables IV and V, the euglobulin was prepared
from horse-plasma (oxalate) by the method described by Mellanby [p. 339, 1905]. The purifica-
tion in this case was just as above and no differences were detected between samples prepared
by the two methods.

320 H. CHICK

suspensions prepared as above invariably had a slight negative charge and
the addition of a little acid was necessary in order to render the particles
iso-electric with the solution.

In Table I is shown the degree of dispersion of a sample of euglobulin
(0032 °/, solution) corresponding with various concentrations of acid and
alkali. The charge carried by the particles was at the same time determined

TABLE I.
Influence of reaction (hydrogen-ion concentration) upon the dispersion of
euglobulin, and the electric charge carried by the protein particles ;
influence of sodiwm sulphate.

Concentration of protein=0°032 °/).

Ce. Ce.
N/100 = =N/100
HCl NaOH
(orequi- (orequi- Concen-
valent) valent) Concen- tration of Sign of
added added trationof hydrogen- electric
in total in total salt, in ions, in charge
Exp. volume volume Salt terms of terms of on the Degree of
No. of10cc. of10cc. added normality normality particles agelutination
il — 0°2 = = 0°36 x 10-7 ~ Faintly opalescent soln.
2 — 0-1 — — 0; Sie ~ Opalescent solution.
S — 0:0 = — 32 4, = Agglutinated.
4 “Al _ — — (AS. - Agglutinated later.
5 oa — — — 11405: + Opalescent solution.
6 “2 = — = 1390 ,, as Faintly opalescent soln.
7 9) — —_ os 5080 ,, Clear solution.
8 2 — Na,SO, 0:03 0:90 es - Opalescent solution!.
9 5 — #5 0:05 15900. - Opalescent solution!.

! Agelutinated on standing.

by observing their behaviour in an electrical field, using the microscopic
method previously employed by Martin and the author [1912, p. 285] in in-
vestigating the electrical properties of denaturated proteins. In the 5th
column is given the concentration of hydrogen-ions in the various solutions,
and the point of optimum agglutination is seen to be at a concentration
equal to 32x 1077 normal. This figure is in good agreement with that
found by Michaelis and Rona.

Confirmation of these values was incidentally obtained in the course of
experiments made to elucidate other points. For example, the range of
agglutination of another euglobulin suspension, containing 0:6 °/, protein
was determined, after first dissolving in a minimum amount of dilute sodium
hydroxide solution, by adding dilute hydrochloric acid to a series of tubes
until precipitation occurred and finally dispersion was again obtained. Dis-
persion corresponded with a concentration of hydrogen-ions equal to 2°2 x 107

H. CHICK 32]

normal on the alkaline side and 62:1 x 10-7 normal on the acid side. The
solutions, of course, all contained a trace of salt.

In case of a third sample of euglobulin, see Table IT (containing 0-016 °/,
protein) the limits of agglutination were found to lie between concentrations
of hydrogen-ions equal to 5 x 10~? and 213 x 10 normal and _ precipitation
to be rapid at a concentration of 18 x 10-7 normal.

TABLE II.

Influence of reaction (hydrogen-ion concentration) upon the dispersion of
euglobulin and upon the electric charge carried by the protein particles;
influence of sodium sulphate.

Concentration of protein=0-016 °/,.

Ce. Ce.
N/100 =N/100
HCl NaOH
Concen- (orequi- (or equi- Concen-
tration valent) valent) tration of
of salt added added hydrogen- Sign of
added, in in total in total ions, in electric
Exp. Salt terms of volume volume terms of charge on Degree of
No. added normality of10ce. of 10ce. normality particles agglutination
1 = ae 0°3 — == + Clear solution.
a = = 0-2 — 1280 x 10-7 + 3 5
3 = = Ol = Dilidw + Agglutination partial.
1 — -- 0:00 136i, - Agglutination complete.
5 — = = oal) =: - Agglutination partial.
6 = = — 0-2 —_ - Dispersed.
v Na,SO, 02 0-5 -- 1010 x 10-7 - Agelutination partial,
less good than No. 8.
8 “3 03 3 _ LS Ties - Agglutination best, but
not quite complete.
9 a “04 .: — Gps = Agglutination less good
than No. 8.
10 As 05 2 = = 0 Dispersed,
11 4 ‘O07 y — — ; 0 Dispersed.

The experiments as to the amount and sign of the charge carried by the
euglobulin particles, detailed in Tables I and II, are not calculated to define
the iso-electric point with as great accuracy as those of Michaelis and Rona.
As a rule the particles, while agglutinating rapidly, retained the negative
charge which they originally held and only showed a positive charge when
dispersion by acid was already well begun. Speaking generally, however,
the point where the charge on the particles changed its sign coincided with
the agglutination-zone. On the whole, there is a tendency to retain the
negative charge rather than the positive. <A distinctly positive charge was
first detected on the globulin particles at a concentration of hydrogen-ions

322 H. CHICK

equal to about 200 x 10-7 normal (see Exp. 3, Table II, and Exp. 2,
Table VII).

The agglutination and dispersion of euglobulin by acids and alkalies, in
confirmation of the results of Michaelis and Rona, is seen to be primarily
dependent upon hydrogen-ion concentration, and in this respect a close
analogy is presented with heat-denaturated proteins [Michaelis and Rona,
1910, 1; Sdrensen and Jiirgensen, 1911; Chick and Martin, 1912]. In both
cases flocculation takes place when the protein particles are iso-electric with
the solution, and, in cases where the reaction of the solution is more acid or
alkaline than the iso-electric point, dispersion of the protein is due to the

possession of a positive and negative electric charge respectively.

III. So.tution oF EUGLOBULIN BY NEUTRAL SALTS.

There has been some difference of opinion as to the mechanism involved
in the solution of euglobulin by electrolytes. Hardy [1905] came to the
conclusion that salt-solutions of globulin are without drift in an electric field
and the particles of the system must be regarded as electrically neutral. He
considered [1905, p. 325] “salt-globulin” to be the result of a molecular
union of globulin and salt! by a process analogous to the formation of amino-

acid-salt compounds’, thus :

Na Cl
N=H, N=H,
R +NaCl=R
~ 20 ee A

A similar theory had already been put forward by Pauli [1899]. |
Mellanby [1905], on the other hand, has expressed the opinion that the ;
solution of globulin by electrolytes is the result of activity of the constituent
ions. He found that the efficiency of a salt in “dissolving” euglobulin
depended on the valency of its ions, and, moreover, that the amount of
globulin dissolved by a given concentration of any salt was proportional to
the initial concentration of the protein. This curious fact was also noticed ‘
by Hardy [1905, p. 310].
A third view is that of Schryver [1910]. He considers (1) that, when
precipitated, euglobulin is in a dehydrated condition, the amphoteric mole-

Peas i

cules having united by twos in salt formation with loss of the elements of

1 Osborne and Harris [1905] also interpreted the solution of edestin by certain electrolytes as
due to the formation of such compounds.
2 These appear to have been prepared in the crystalline form by Pfeiffer and Modelski [1912].

ee

H. CHICK 323

water and (2) that, when dissolved by electrolytes, the molecules remain
separate in their hydrated condition, the aggregation in pairs being prevented
by adsorption of the electrolytes on the surface of the molecules, thus

COOH NH,OH COO—-NH,
fe a
R + R = H,0+R R
~ rid
NH,OH COOH NH,OH COOH
~ = ~- el ~~ = cr —
Salt-dissolved globulin Precipitated (dehydrated)
(hydrated) globulin

The insolubility of the dehydrated euglobulin Schryver attributes to the
increased size of the molecule’. There seems to be, however, better experi-
mental evidence in support of the more usual conception that the euglobulin
is precipitated in such circumstances as will allow of the aggregation of the
protein particles under the influence of surface tension, that is to say, in
the absence of other causes tending to keep them dispersed. Such causes
are the presence of acid or alkali, in which case the euglobulin particles have
been shown to be charged electrically, or the presence of electrolytes in
whatever way the latter may act.

Schryver’s view is in accord with that of Hardy in so far as the solution
of euglobulin by electrolytes is attributed by both to the agency of the salt
as a whole rather than to the activity of the constituent ions. His adsorp-
tion hypothesis, however, affords no explanation of the comparatively high
concentration necessary for solution in case of many salts.

An additional theory, for which some experimental support is here given,
is that dispersion of euglobulin by salts may be due to a specific adsorption
by the protein particles of one—the more potent—ion of the electrolyte
employed, and the acquisition of an electric charge by this means. Such
an hypothesis would explain the influence of valency, demonstrated by
Mellanby, and be consistent with what is known of the action of neutral
salts in dispersing denaturated serum proteins [Chick and Martin, 1912,
p- 281], with which euglobulin is in many ways analogous.

?

In the experiments detailed below, the “solution” of euglobulin by
neutral salts was studied, using a series of electrolytes of more varied
character, as regards the valency of the constituent ions, than those employed
by the above observers. By this means some facts have been discovered

which go to show that while, in some instances, solution of euglobulin by
1 The conclusion of Dabrowski [1912] that in solutions of crystalline egg-albumin containing

ammonium sulphate, the colloidal particles are of smaller size than in the case of the dialysed
protein, is of interest in this connection.

324 H. CHICK

neutral salts appears to be due to some action of the whole salt, in other
cases it is undoubtedly associated with an acquirement of electric charge by
the protein particles, the sign of which is determined by the more potent

ion of the electrolyte employed.
1. Experiments to determine whether salt-globulin ws electrically charged.

A series of experiments with a dilute euglobulin suspension (0:016 °/,
protein) and a variety of salts is set forth im Table HI. The object was
to compare quantitatively the solvent and dispersing power of different 1ons
and to decide whether solution was accompanied by the deposition of an
electric charge upon the protein particles. The latter, in this instance, was
determined by the microscopic method, using a dark ground and side
illumination. The fluids were placed in a specially constructed cell and the
behaviour of the particles in an electric field was observed. In using this
method with fairly concentrated salt solutions (0°01 normal and upwards)
there is some risk of disturbance by gas-bubbles given off at the electrodes,
as the result of electrolysis. This has to be very carefully watched. As
a rule the difficulty is minimised if the electrodes are previously platinised
and the circuit closed only for the short time necessary to make the observa-
tion. The electromotive force employed was about 9 volts.

The results of this series of experiments show the relative solvent power
of six different salts, viz. the chloride, sulphate and citrate of sodium, the
chlorides of barium and calcium and lanthanum nitrate.

It is very clear that the influence of valency in preventing agglutination
of euglobulin is far from negligible, multivalent ions being much the more
powerful in order of increasing valency. For example, the concentration of
lanthanum nitrate and sodium citrate required for the purpose is respectively
1/30th and 1/60th of the necessary concentration of sodium chloride
(5th column, Table IIT).

The electrical condition of the dispersed particles is set out in column 4,
and it will be seen that, in four cases, a charge was demonstrated of similar
sign to that carried by the more potent ion of the electrolyte employed. In
the case of calcium! and barium salts, on the other hand, no charge could be
detected on the protein particles which remained visible.

The character of the dispersion obtained was also different in the case
of different salts, and could be grouped according to two types, which I have
tentatively called the “electrical” and the “molecular” type of solution.

1 This was also found to be the case when denaturated serum-proteins were dispersed by
calcium salts [Chick and Martin, 1912, Table XIV}.

ee An?

ey

are

Salt
added

NaCl

Na,SO,

Na,Cit

CaCl,

BaCl,

La (NO,),

H. CHICK 325

TABLE III.

Solution of euglobulin by electrolytes.

Concentration of protein =0:016 °/,.

x x x Complete agglutination, filtrate protein-free.
x x Less complete agglutination, filtrate contained trace protein.
x Partial agglutination.
x — Almost complete dispersion.
- Complete dispersion, clear solution.

Sign of Concentration of salt, in
Concentration of the electric terms of normality, necessary
salt, in terms Degree of charge upon (a) toprevent (b) tocomplete
of normality agglutination the particles agglutination dispersion
0-01 SCE 0-03 —
0-02 Kes .
0:03 sux
0-04 Sa
0-045 xx _*
0-05 x =
0-0005 SOC 0-02 0°05
0-001 aS es
0-01 KEKE
0:02 ox _*
0-03 x
0-04 x — --*
0:05 ~
0:0005 x : 0-0005 0-003
0-001 :
0-002 x —
0-003 ~
0-01 Sa SSH 0-02 —
0°02 ;
0:03
0-04
0:05 x 0
0-001 bs > * 0-005 0-05
0:005 See
0-008 Sue
0-01 Scere 0
0:02 x 0
0:03 x -
0:05 - 0
0-001 xx + 0-001 0-02
0-003 x
0-005 505
0-006 x — +
0-008 x
0-01 x
0:02 -

** Aoclutinated particles.

326 H. CHICK

That with a lanthanum salt or a citrate was exactly analogous to dispersion
by acids and alkalies, the globulin suspension, with progressive increase in
concentration of the salt, passed through various gradings of opalescent
solutions until a clear transparent fluid was finally obtained. At the same
time, under the microscope, using a dark ground illumination, innumerable
particles could be seen which became smaller and smaller, acquiring at the
same time an unmistakably positive or negative charge. Finally the whole
field became a uniform grey in colour, until at length no particles could be
distinguished.

In the case of the other salts used, there was usually no grading of
opalescence, but, as the concentration of salt was increased, a decreasing
amount of the original suspension remained undissolved in an otherwise clear
fluid. Under the microscope there was a black and white effect, fewer and
fewer agglutinated masses remaining visible on the black ground. This type
of solution obtained with sodium sulphate and the chlorides of sodium,
calcium and barium. In the case of the two latter salts, no charge was
discovered on the visible particles; with sodium chloride and sulphate a
negative charge was observed on the agglutinated particles which remained
visible. It must however be remembered that the sign of the charge in both
cases was negative, Le. similar to that carried by the original globulin
suspension.

The microscopic method is clearly unsuitable for studying the electrical
condition of salt-globulin in such cases as the above, where “solution” is
associated with a degree of dispersion which renders the dispersed particles
immediately invisible. A series of corresponding experiments were accord-
ingly made, in which the “U” tube method was employed, the object being
to find out whether actual kataphoresis of the “ salt-globulin” did or did not
take place in an electric field.

This was the method used by Hardy [1905] whose conclusion that “ salt-
globulin” was without electric charge has already been referred to. In his
experiments, however, as was pointed out by Michaelis [1909] the protein
solution may not have been adequately protected from the influence of the
acid and alkali produced at the electrodes by electrolysis of the salt present.

In order to avoid all complications due to electrolysis five U tubes were
arranged in series in the present instance. The “salt globulin” occupied the
centre tube, which was connected by a three-way tap on either side for
convenience in filling. The other four tubes, two on either side, contained
a solution of the same electrolyte in the same concentration as that used
to disperse the euglobulin. All five tubes were fastened by brass clips to

I

a= — ~

H. CHICK 327

a stand, so constructed that their position could be altered until the heignt
of liquid in all the tubes was accurately adjusted to a standard level. Only
after this was accomplished were the taps opened, and the euglobulin
solution placed in contact with the other tubes. Control experiments
showed that under these circumstances there was no transference of liquid
from one tube to another. The electrodes were placed in the further arms
of the two end tubes and the globulin solution was thus securely protected
from the influence of any acid or alkali produced there by electrolysis.
Litmus and phenol-phthalein were added to the tubes containing the positive
and negative electrodes respectively and in all cases the experiment was
discontinued long before there was any danger of acid or alkali reaching the
centre tube. The resistance of this arrangement was very great, and,
although the electrodes (small strips of platinum foil) were connected with
the lighting cireuit (200 volts) the current which passed amounted to only
00001 to 0:°005 amperes according to the nature and concentration of the
electrolyte employed. At the close of the experiment, which usually lasted
from 5 to 10 hours, the contents of the nearer arms of the U tubes adjacent
to the globulin tube were tested for the presence of protein by addition of
Ksbach’s reagent or otherwise.

The results are set out in Table IV below, and, in general, confirm the
results of Table LII. In the case of calcium chloride, barium chloride and

TABLE IV.

Electrical properties of euglobulin, dispersed by various electrolytes.
(U-tube method.)

Protein content=0:16 °/).

Concentra- Dura- Signof Sign
tion, in tionof electric of electric
terms Electric experi- charge charge, observed
Exp. of nor- field, Current, ment, upon Appearance of by micro-
No. Salt mality volts. ampéres hours protein solution scopic method
1 Na,SO, 0-05 200 0:0025 85 0 Clear solution con- =
taining some par- (agglutinated
ticles undissolved. particles)
2 CaCl, 0-05 so 0-002 55 0 Almost clear solu- 0

tion, containing
some particles.

3 BaCl, 0-04 5 0-005 50 0 Almost clear solu- 0
tion, faint opales-
cence.
4 La(NO,), 0-005 os 0:0001 55 + Opalescent solution +
5 Na,Cit 0-003 i 0-001 50 - Opalescent solution
6 > 0-05 ae 0-0015 9°75 0 Clear solution con-

taining someagglu-
tinated particles.

328 H. (CHICK

sodium sulphate in concentrations of 0:05, 0:04 and 0:05 normal respectively,
no migration of the dissolved protein was demonstrated. With the last
salt, however, the negative charge carried by the undissolved particles was
again shown by the settling which took place less rapidly in the “ positive a
than in the “negative” arm of the centre U tube, showing that in the
latter case the action of gravity was reinforced by motion of the particles
towards the positive pole. ‘

Euglobulin dispersed by a small concentration of lanthanum nitrate
(=0:005 normal) displayed a marked positive charge and in the case of
sodium citrate? a negative charge was shown under similar conditions
(concentration = 0:003 normal). If the concentration of sodium citrate
was increased to 0:05 normal, however, the dispersed euglobulin showed no
migration at all.

The case of sodium citrate is a very interesting one, as the character
of the dispersion obtained by employing these two concentrations was also
quite different, conforming respectively to the “electrical” and “molecular”
types of solution described above (p. 324). With the dilute salt the dis-
persion is to an opalescent solution and is accompanied by the acquisition
of a negative charge; in the stronger solution of citrate, the protein
is uncharged and the liquid shows a small precipitate suspended in a
clear fluid.

The change from the one type of solution to the other can be seen if
a series of solutions be made up containing equal amounts of euglobulin
and a concentration of sodium citrate ranging from 0001 to 0°05 normal.
Dispersion is already well marked at 0:001 normal and continues through
various grades of diminishing opalescence to a concentration of about 0:0038
normal, At a concentration of 0°005 normal the opalescence is greater and
at 0°01 normal, a distinct precipitate can be seen. In higher concentrations
a gradual clearance takes place, but the solution is now of a different
character, less and less of the precipitate remaining undissolved in an other-
wise clear solution. In those experiments the protein content was 0°16 °/).

Another interesting experiment is the following: The U-tube apparatus
was arranged so that in the centre tube was a dispersion of euglobulin in
0:003 N sodium citrate and in the side tubes 0°05 N sodium chloride. As
the current passed a precipitation took place in the negative arm of the
centre tube, the zone of which continually progressed towards the positive
arm. This was followed by a zone of clear solution, also moving in the same

1 These citrate solutions were carefully prepared to be quite neutral and the hydrogen ion
concentration was approximately 10~7 normal.

buen

H. CHICK 329

direction. The interface between the advancing chlorine ions and the citrate
ions will not be sharply defined, as the former travel more rapidly and tend
to over-run and intermix with the latter. A possible explanation is therefore
that as both citrate and chlorine ions move towards the positive pole, the
latter replace the former im the solution around the protein which is also
moving towards the positive electrode, but at a slower rate than either,
In low concentration chlorine ions are unable to disperse globulin and
precipitation occurs, to be followed again by solution when the concentration

of the chlorine ions is sufficiently increased.

The conclusion to be drawn from these experiments is that solution of
euglobulin by neutral salts, in case of the more ordinary electrolytes, is due
to the formation of some unionised and uncharged compound of salt and
globulin—whether by a molecular union (Hardy) or as the result of adsorp-
tion (Schryver) there is not, at present, enough experimental evidence to
decide. At the same time the phenomenon of dispersion by salts, at any
rate in its beginning, is due to an electric charge being deposited on the
particles of the euglobulin suspension by the agency of the ions of the
electrolyte. The influence of salts upon dispersions of euglobulin by acids
and alkalies, dealt with in the next section, is in support of this view.

In the case of the commoner salts, containing only mono- or divalent
ions, the electric charges brought into play are not powerful enough to
disperse the euglobulin until the concentration is increased to a point where
the second (“molecular”) type of solution takes place. With such electro-
lytes the first or “electrical” type of solution as a rule is negligible.
It was, however, detected in the case of sodium chloride and sodium sulphate.
With sodium citrate, on the other hand, the one type was seen to give place
to the other as the concentration of salt was progressively increased.

When euglobulin is denaturated by heat, it loses its characteristic
property of forming electrically neutral solutions with electrolytes. For
example, after heating a dispersion of euglobulin in 0:05 normal sodium
sulphate, in which no kataphoresis of the protein could be demonstrated,
the protein particles were found to migrate to the positive pole in an electric
field. At the same time an alteration took place in the appearance of the
solution—the degree of dispersion was diminished and a thick opalescence
was developed. With dispersions in sodium chloride (0-1 normal) and
calcium chloride (0°15 normal) almost complete agglutination took place

on heating.

330 H. CHICK

2.- Alteration in electrical conductivity during solution of euglobulin
by electrolytes.

It is clear that if a molecular union or an adsorption compound is formed
during solution of euglobulin by electrolytes, there must be some diminution
in electrical conductivity. Hardy [1905, p. 307] states that a loss of con-
ductivity takes place equal to 1-4°/, to 2°4°/, in case of solution by magnesium
sulphate and sodium chloride respectively.

The results of a special set of experiments, made for the purpose, confirm
the result of Hardy.

The same sample of euglobulin was used as for the previous set of
experiments, and the mixtures when prepared contained 0°6°/, protein. The
conductivity was determined :

(1) of the “salt-globulin” solution,

(2) of an equal concentration euglobulin suspension in distilled water.

In the case of (1) and (2) the solutions were allowed to settle or were
centrifuged and the conductivity measured in the supernatant liquid.

(3) of an equal concentration salt solution in distilled water.

‘The conductivity of the distilled water used was found to be negligible
in comparison; hence direct comparison was made between (1) and the sum
of (2) and (3). All determinations were made at 18°.

The results are set out in Table V and show a loss of conductivity in
every case investigated. The salts used were the chlorides of barium and
sodium, sodium citrate and lanthanum nitrate. In the case of what I have
termed the “electrical” type of dispersion the diminution in conductivity
was proportionally greater than with the molecular type; in the former case
it occurred with iso-electric euglobulin to the extent of 13:6 °/, and 11°6°/, in
case of dispersion by weak (0:004 normal) lanthanum nitrate and sodium
citrate respectively, see Exp. 1. With more concentrated salt, 0°05 normal,
the loss varied from 1—4°/, in the case of the four salts employed.

3. Relation of the amount dissolved by a salt to the total
amount of euglobulin present.

Notice has already been made of the observation of Mellanby [1905,
p. 342] confirmed by Hardy [1905, p. 310] that the amount of euglobulin
dissolved by a given concentration of salt is, within certain limits, approxi-
mately proportional to the concentration of protein in the original suspension.
In the case of the salt used by Mellanby (sodium chloride in concentration

Aeron in Se

H. CHICK 331

TABLE V.

Change in electrical conductivity on solution of euglobulin by electrolytes.

Concentration of protein=0-6 °/,, temperature 18° C.

Concen-
tration
of eu-
Concen- globulin
tration Conductivity, in reciprocal ohms Per- in solu-
Condition of salt, ——_——_ ~ centage tion as
of euglo- in terms Salt “Salt-glo- “ Salt- lossin ‘Salt-glo-
Exp. bulin sus- of nor- solu- Euglobulin bulin” globulin” conduc- bulin,”
No. pension Salt mality tion suspension (found) (calculated) tivity oF)
1 Approxi- NaCl 0:05 0°1389 0:00050 ~=—-0"1378 0-1394 1 tea ly —
mately BaCl, 0:05 = 0°1368 0°00126 =01368 0°1381 0-99 —
iso- Na,Cit 0-004 0:005489 0:00046 0:005331 0:005949 11-6 —
electric ~ 0°05 0°05289 0:00046 0°05236 0:°05335 1-90 —
La(NO.), 0°004 0°006785 0:000345 0:006277 0-007130 13:6 —
ne 0:05 0-06708 ue 0°06536 0°006742 3:14 —
2 Alkaline Na;Cit 0:004 0:005562 0:000569 0-005578 0:006131 9-93 0:28
st 0°05 0°05357 x 0:05337 =0°05414 1:45 0°41
La(NO,), 0:004 0:006845 5 0006482 0-007414 14-4 0°34
# 0:05 0:06681 Bs 0:06511 0°06738 3°5 0°51
3 Acid Na,Cit 0:004 0:-005541 0°002145 0:006907 0-007686 11-3 0-07
- 0:05 0°05359 35 0°05415 0°05573 2°93 0-42
La(NO.), 0°004 0-006850 5 0:008264 0:008995 8-85 0°45
= 0:05 0:06794 5 0:06717 0:07008 4°33 0°53

from 0:04 to 0:09 normal) solution is of the “molecular” type. On the
assumption that this solution is the result of a molecular union between
salt and protein (Hardy) it is difficult to explain the existence of the above
relationship. On the other hand, if Schryver’s view of the salt-solution of
euglobulin be accepted in one respect, that is to say if we consider the
second or “molecular” type of globulin to be the result of “adsorption ”
of the salt as a whole, the amount of globulin dispersed by a given concentra-
tion of salt might be approximately proportional to the extent of adsorbing
surface i.e. to the concentration of euglobulin’.

It was also possible, however, that solution-rate might be conditioned
by the size of the euglobulin particles (of which there would be every variety
in such a suspension as that used by Mellanby) and that in the time of
experiment, final equilibrium had not been maintained but only the
particles of small size and comparatively large surface had been successfully
attacked.

1 The experimental support for this view is based upon the fact that the globulin-dissolving
capacity and surface-tension (against air) of series of solutions of similar salts, were found to be
inversely related to one another, which is the result that should obtain, in accordance with the
Willard Gibbs hypothesis. The application of surface tension measurements of air against
solution to the case of the protein against solution is not, however, without risk of error.

Bioch. vir 22

332 H. CHICK

~ Some experiments were made to examine this theory, using equal volumes
of a 3°/, suspension of globulin and 1/10th normal sodium chloride solution.
It was found, however, that almost perfect equilibrium was attained
within a few minutes after mixing. Experiments were also made, using
in the one case finely divided euglobulin and in the other a similar suspension
of globulin previously aggregated by freezing; the amount of protein dis-
solved in the second case by 0°05 normal sodium chloride, under similar
conditions, was about 20°/, less than in the former. A phenomenon of this
magnitude is, in itself, inadequate to explain the relationship observed by
Mellanby and Hardy; it may, however, be a contributing factor.

The facts would appear to be best met by some such conception as the
following: Solution of euglobulin by such electrolytes as sodium chloride
or sodium sulphate, is due to the formation of a “soluble” compound of the
globulin and the salt, which is electrically neutral, and to prevent the
dissociation of which a large excess of the salt is necessary. Under such
circumstances the final equilibrium, i.e. the relative proportion of “dissolved”
and “undissolved” globulin, will largely depend upon the concentration of
the salt employed and show an approximate constancy if the latter is
maintained constant. In other words, the amount of dissolved euglobulin
will be roughly proportional to the original concentration (Mellanby).

IV. THE EFFECT OF NEUTRAL SALTS UPON ACID AND ALKALINE
DISPERSIONS OF EKUGLOBULIN.

The influence of salts in causing precipitation of euglobulin previously
dissolved in acid and alkali is analogous to the corresponding action with
denaturated serum protems.

Hardy [1905, p. 317] has drawn attention to the fact that the pre-
cipitating action! of a salt upon “acid or alkali-globulin” is due to one
only of its ions, viz.: that which carries a charge opposite in sign to that
carried by the protein, and that the higher the valency of this ion, the
greater is the power of the salt.

These results have been confirmed by me and I have also been able to
show that, if the concentration of salt is further increased, dispersion will
again occur, the particles now taking a charge whose sign is determined by
that of the more potent ion of the electrolyte employed. Thus the addition
of an appropriate electrolyte to an alkaline or acid solution of euglobulin

1 Mellanby [1905] has dealt with the precipitation of globulin from its solution in electrolytes.
This is a different phenomenon and a high concentration of salt is required (salting out).

H. CHICK 333

may, after first precipitating the protein, cause the particles to disperse
again with a charge opposite in sign to that originally carried.

For example, an alkaline dispersion of euglobulin containing 0016 °/,
protein, with the particles megatively charged, was precipitated by the
addition of lanthanum nitrate to a concentration of 0°001 normal; at a
concentration of 0-008 to 0:01 normal, dispersion again took place, the
particles now being positively charged. In a similar experiment with sodium
chloride the precipitation took place at a concentration of salt equal to 0°02
normal; at 0:05 normal the globulin was again dispersed, this time bearing
a negative charge.

These results present a close analogy with what takes place in case of
denaturated serum-proteins, where the degree of dispersion in acid and
alkaline solution is also greatly influenced by the presence of neutral salts,
and in two ways. In the first place the reaction of protein-containing
solutions, whether denaturated or not, becomes altered on addition of neutral
salts [Chick and Martin, 1911, p. 21; 1912, p. 280], being shifted in the
direction of the neutral point. In acid solution, the concentration of hydrogen
ions 1s lowered and the concentration of hydroxy] ions lessened if the solution
be alkaline; the effect is related to the valency of the anion and kation
respectively of the electrolytes in the two cases. In the second place, the
electric charge carried by the protein particles is modified and may be
lessened or even changed in sign by the addition of electrolytes if of opposite
sense to that carried by the more potent ion of the electrolyte added.

The effect of the above-mentioned salts in modifying the reaction? of solu-
tions containing protein has also been demonstrated in the case of euglobulin
(see Tables I, VI and VII). It was therefore necessary to determine in how
far the precipitating effect of salts was due to this effect. It is evident
that solutions either too acid or too alkaline for precipitation of euglobulin
might be adjusted to the iso-electric reaction by the addition of an appropriate
electrolyte, and indeed this frequently occurred. For example (Experiment
12, Table VI), addition of sodium citrate to a concentration of 0:002 N,
caused precipitation of an acid dispersion of euglobulin (0:016 °/, protein) at
the same time reducing the concentration of hydrogen ions to a point very
near the iso-electric point for this protein. In some cases the change of
reaction extended to the other side of the iso-electric point, e.g. Table VI,
Experiment 13; in this case the observed change of sign in electric charge
taking place simultaneously with dispersion by sodium citrate, could be
explained on the ground of change of reaction alone. By comparison of

1 The phenomenon in absence of protein is perceptible but negligible,

334 H. CHICK

Experiment 7 with 12 and 13, Table VI, this change in reaction is seen
to be increased proportionally with the degree of valency possessed by the
ions of the salt employed.

TABLE, VAL.

The effect of sodium sulphate and citrate upon an acid dispersion
‘of euglobulin (0-016 °/,).

Ce. N/100 HCl

Concentra- (or equiva- Concentration Sign of
tionofsalt lent)addedin of hydrogen electric charge
Exp. Salt intermsof totalvolume ions,interms carried by Degree of
No. added normality of 10 cc. of normality . the particles agglutination
1 — — 0-0 = Partial agglutination.
2 — 0:075 - _ Agglutinated completely.
io oo 0-1 - 5 Ls
4 — 0°2 LOS 20 (S105 1054) + Dispersed, clear solution.
5) 0:3 af ” ” ”
6 — 0-7 ar » ” ”
7 \Na sO, 0-03 0°5 10-4: (357 x 107’) ~ Agelutinated.
8 0-04 x — Partial agglutination.
9 0°05 * - Dispersed.
10 0:07 ” 0 ”
11 Na,Cit 0-001 0-7 ib Dispersed.
12 0-002 rs LORS 25 (SOA Oey) - Almost complete aggluti-
13 0-003 = ORO OFZ 104) : Dispersed. [nation.
14 0-004 > - Dispersed, faintly opal-

escent solution.

In many cases, however, acid and alkaline solutions of euglobulin, with
their positively and negatively charged particles respectively, were first
precipitated and afterwards dispersed, the particles bearing an electric
charge of changed sign in solutions whose reaction still remained more acid
and more alkaline respectively, than the iso-electric point’. A good example
of this is seen in Table I, Experiment 9, where addition of 0°05 N sodium
sulphate to an acid dispersion of euglobulin (0°:032°/, protein) caused the
particles to be dispersed and to carry a negative charge in a solution where
the hydrogen ion concentration, equal to 1590 x 10~7 normal, was far on the
acid side of the iso-electric point. Experiment 7 of Table II is another
instance of the same phenomenon}.

With an alkaline suspension of euglobulin a corresponding series of
results was obtained with lanthanum nitrate. The results are set out in

! In higher concentration of sodium sulphate, the dispersed globulin appears to be electri-
cally neutral, see Exp. 10, Table VI.

H. CHICK 335

Table VII. A preliminary set of experiments (1 to 8), in absence of electro-
lytes, showed a positive charge to be acquired by the protein particles at
a concentration of hydrogen ions equal to 204 x 10-7 normal. In the alka-
line suspension used, the concentration of hydrogen ions was equal to
00013 x 10-7 normal. In presence of 0:006 normal lanthanum nitrate the

TABLE VII.

Effect of lanthanum nitrate upon an alkaline dispersion of euglobulin.
Protein=0-016 °/,.
Ce.N/100 Ce. N/100

HCl NaOH
(or equi- (or equi- Concen- Sign of
Concen- valent) valent) tration of electric
tration of addedin addedin hydrogen- charge
salt, in a total a total ions, in carried
Exp. Salt terms of | volume volume terms of by the Degree of
No. added normality of 10ce, of 10cce. normality particles agglutination
1 ae. as 0-2 — Dispersed, clear solution.
2 ae Ol eG ORC? + Dispersed, faintly opal-
( (204 x 10-7) escent solution.
3 et 0:05 = Ieee - Agglutinated.
|{ (1:05 x 10-7)
4 = 0 = = Dispersed, opalescent sol.
5 oS a 0°5 29 ” ”
6 a = 0-1 Dispersed, clear solution,
u —* =a 0-2 ” ” ”
8 = — é 0°5 ( 103 oe) ” ”
{(0:0013 x 10-7)
9 La(NOs)3; 0:0005 — 0:5 Dispersed, opalescent sol.
10 0-001 = re Agglutinated.
11 0-002 — 33 + Agglutinated almost com-
pletely.
12 0-004 os 5 af 5 Ag *.
13 0:005 = 5G + Agglutination not com-
plete.
14 0-006 — fe ee ee + » 9 9
| (0-24 x 10-7)
15 0:008 -- 5 Dispersed partly.
16 0-01 — + » 9 ;
17 0-02 a ” be) ” ”

alkalinity was reduced almost to the neutral point. At the same time the
charge on the particles was found to be positive in a solution whose reaction
was on the alkaline side of the iso-electric point, and where, in the absence
of any electrolyte, the particles would be negatively charged.

In these instances the effect of electrolytes recalls the exactly similar set
of phenomena obtaining in the case of denaturated serum-proteins mentioned

22—3

336 H. CHICK

above, and also the analogous influence of salts which has been observed
in case of inorganic colloidal solutions [Burton, 1909]. In certain of the
latter there has been demonstrated a selective adsorption of the ion bearing
a charge opposite in sign to that of the charged colloidal particles [Linder
and Picton, 1895; Whitney and Ober, 1902; Freundlich, 1910] and this
explanation may be extended to the case of proteins.

In this connection, certain of the conductivity determinations in Table V,
i.e., those of Experiments 2 and 3, are of special interest. These experiments
were made with suspensions of euglobulin which were slightly on the acid
and alkaline side, respectively, of the iso-electric point. The effect of adding
electrolytes containing trivalent positive (La(NO,),) and negative (Na,Cit)
ions was carefully studied. The loss of conductivity was found to be greater
for sodium citrate if the globulin suspension were originally acid (i.e. charge
on the protein particles positive) and greater for lanthanum nitrate if the
globulin suspension were alkaline. At the same time, the degree of dis-
persion was measured by the content of protein in the supernatant fluid
after centrifuging (see last column), and was found to be less. In all cases
equivalent solutions of the salts were compared and the effects were more
marked in the experiments with less concentrated salt (0-004 normal), in
which case dispersion has been shown to be of the “ electrical” type.

Analogy with euglobulin presented by caseinogen and other proteins.

Caseinogen has been found to show close analogy with euglobulin as
regards the effects of electrolytes upon its solutions in either acid or alkali.

Michaelis and Rona [1910, 2] drew attention to the fact that among the
naturally occurring proteins whose solution was accompanied by the acquisi-
tion of electric properties, in addition to globulin, were caseinogen, gliadin,
and edestin, and they determined the iso-electric point in each case. The
result was especially interesting in the case of caseinogen, where the
particles were found to be iso-electric with the solution, and to be precipitated
when the concentration of hydrogen ions was 1°8 x 10~ normal, a degree
of acidity far beyond that determined for euglobulin or for the heat-
denaturated proteins of serum.

The following experiments, set out in Table VIII, show the analogy with
euglobulin to be closely maintained in respect also of the action of electrolytes.
For example, in an opalescent solution (0°05 °/,) in weak hydrochloric acid,
in which the protein particles were yet visible under the microscope using
a high power and dark ground illumination, the caseinogen was found

;

= — en x Fae pet er oe eee wate, PR EPR tit wit ery. -

H. CHICK 337

to be positively electrified. This solution was readily precipitated by a
minute concentration of sodium sulphate (equal to 0°0005 normal) or sodium
citrate (0:00005 normal); with increased concentration of either salt (0°015
and 0001 normal, respectiyely) the protein particles were again dispersed
and found to carry a negative charge in both cases, see Table VIII, Ex-
periment A.

TABLE VIII.

Influence of electrolytes wpon agglutination of caseinogen
(Merck’s pure casein).

(A) Dispersed with a little HCl. (B) Dispersed with a little NaOH.
x x x Complete agglutination. x Partial agglutination.
x x Almost complete agglutination. x — Almost complete dispersion.

— Complete dispersion.

Concentra- Sign of
Concentra- tion of salt, electric charge
tion of Salt in terms of Degree of carried by
Exp. No, protein, °/, added normality agglutination the particles

(A) 0°05 0 0-00 - +
+ Na,SO, 0:0005 x xX Xx
0-001 x x
0-005 x xX
0-01 <<
0-015 = =
0-02 =

es Na,Cit 0:00001 x xX
0°00005 x x x
0-0001 x xX
0-0005 x
0-001 = =
0:002 =

(B) 0-03 a 0-00 * as
a CaCly 0-01 =

0:02 x =
0°05 x

0-09 x 0
0-10 x
0°15 x

0-2 x - 0
0:5 -

Re La(NO,). 0:00002 -
000005 x =
0:00008 x
0-0001 Gio
0-0002 x x x
0-0005 x -
0-001 _ +

338 H. CHICK

Experiment (B) with an alkaline dispersion of caseinogen 0:03 °/,, showed
an exactly analogous set of phenomena with lanthanum nitrate.

With calcium salts the action appears to be different in character. In
the first place no complete precipitation takes place; the size of the particles
is increased, but no complete agglutination occurs although the solution
becomes turbid. The particles, when dispersed again by increased con-
centration of the salt (0°2 to 0°5 normal) do not appear to carry any electric
charge. In this respect also the analogy with both denaturated serum-
proteins and with euglobulin is maintained.

V. SUMMARY. -

1. The iso-electric point for euglobulin has been re-determined and
found to coincide with the point of most rapid agglutination, viz.: at a
hydrogen ion concentration of about 3 x 10~ normal, a figure which agrees
with that obtained by Michaelis and Rona [1910, 2}.

2. The solution or dispersion of euglobulin by electrolytes is shown to
be much influenced by the nature (especially as regards valency) of the
constituent ions and to be of two general types :

(a) “electrical” type of solution in which the euglobulin dispersion
is accompanied by the acquisition of an electric charge by the protein
particles, the sign of which is similar to that of the more potent ion of the
electrolyte employed.

(b) “molecular” type of solution, in which the dissolved euglobulin is
electrically neutral.

In type (a) the dispersion is considered to result from a specific adsorp-
tion of the ion possessing the higher valency, in (b) from a molecular union
with (Hardy) or adsorption of (Schryver) the salt as a whole. Both types
of solution are accompanied by loss of electrical conductivity in the liquid.

The “electrical” type of solution is well seen in case of dispersion by
such salts as sodium citrate and lanthanum nitrate in low concentration ;
in case of the more ordinary salts, containing mono- or divalent ions only,
the electric forces concerned are not powerful enough to disperse globulin
until the concentration is raised to a point where “ molecular” solution takes
place. In the case of sodium citrate, the “electrical” type of solution was
found to change to the “molecular” type as the concentration of the salt was
increased.

3. Euglobulin, when denaturated by heat, no longer possesses the
property of forming the “molecular” type of solution with electrolytes. On

de aon

H. CHICK 339

heating the latter, in some cases the degree of dispersion is merely diminished,
and the protein particles acquire an electric charge, whose sign is determined
by the more potent ion of the electrolyte employed ; in other cases agglutina-
tion takes place.

4. The reaction of acid and alkaline solutions of euglobulin is greatly
influenced by the addition of electrolytes, the hydrogen and hydroxyl ion
concentration being reduced respectively. In case of the former the effect is
much increased with rising valency of the anion and in alkaline solution the
result is determined by the valency of the kation.

5. The influence of electrolytes in causing precipitation of globulin
dissolved in acid and alkali may, in some instances, be adequately explained
by the alteration in reaction, described under 4; in this way solutions too
acid or too alkaline for agglutination of the globulin may be adjusted to the
iso-electric point by the addition of an appropriate electrolyte.

Precipitation by electrolytes may, however, also take place in solutions
whose reaction is still far removed from that of the iso-electric point. In
these instances it is attributed to neutralisation of the electric charge
originally carried by the protein particles by means of a specific adsorption
of the oppositely charged ion of the electrolyte; the effect is related to
valency.

6. In the properties regarding solution and precipitation detailed under
1, 4 and 5, euglobulin, in common with caseinogen, and the vegetable
globulins presents a very interesting analogy with heat-denaturated proteins.
Kuglobulin differs from heat-denaturated protein in its capacity to form
solutions with electrolytes in which the protein particles are electrically
neutral.

In conclusion, I wish to express my indebtedness to Prof. C. J. Martin
for much helpful advice and criticism.

REFERENCES.

Abderhalden (1903), Zeitsch. physiol. Chem. 37, 49
(1904), Zeitsch. physiol. Chem. 44, 17.

Burton (1909), Phil. Mag. (6) 17, 583.

Chick and Martin (1911), J. Physiol. 43, 1.

(1912), J. Physiol. 45, 261.

Dabrowski (1912), Bull. Akad. Sci. Cracow A, June, 485.
Freundlich (1910), Zeitsch. physikal. Chem. 73, 385.
Hardy (1900), Proc. Roy. Soc. 66, 110.

(1905), J. Physiol. 33, 251.

Linder and Picton (1895), J. Chem. Soc. 67, 63.

en edt ‘ie St Spee a
Poe bad ic ee

340 | H. CHICK

Mellanby (1905), J. Physiol. 43, 338.

Michaelis (1909), Biochem. Zeitsch. 19, 181.

and Mostynski (1910), Biochem, Zeitsch. 24, 79.

—— and Rona (1910, 1), Biochem. Zeitsch. 27, 38.

— -— (1910, 2), Biochem. Zeitsch. 28, 193.

Moll (1904), Beitriige, 4, 563.

‘Osborne and Harris (1905), Amer. J. Physiol. 14, 151.
Pauli (1899), Pfliiger’s Archiv. 78, 315.

Pfeiffer and Modelski (1912), Zeitsch. physiol. Chem. 81, 329.
Schryver (1910), Proc. Roy. Soc. B, 83, 96.

Sérensen and Jurgensen (1911), Biochem. Zeitsch. 31, 397.
Starke (1900), Zeitsch. Biol. 40, 494.

Whitney and Ober (1902), Zeitsch. physikal. Chem. 39, 630.

AAU THE PAT OF YEAST:

By ALLEN NEVILLE.
Department of Agriculture, University of Cambridge.

(Received April 24, 1913.)

Probably owing to the small quantity of material usually available, only
one previous examination of yeast fat, that by Hinsberg and Roos [1903],
appears to have been carried out, and, as this investigation was inconclusive
on several points, it appeared to be worth while to go over the ground again.
In an examination of yeast products for the “vitamines” or “hormones”
connected with the phenomenon of animal growth, fairly large quantities of
the crude fat of yeast became available and on this the following investiga-
tion was carried out. The results confirm the statement of the above
mentioned workers that the principal saturated fatty acid present is a
pentadecoic acid of empirical formula C,;H;,0,, but in addition the author
separated a small quantity of arachidic acid, C.,H,O.. There might be noted
here the diversity of opinion in the literature as to the existence of a
pentadecoic acid in the natural fats and the widely different values given
to the physical constants of this acid by those authors who claim to have
isolated it. The values given for the melting point vary from 70° to 52°,
while the same constant for the methyl ester varies from 68° to 38°.
Probably some of the acids described were impure as it is unlikely that
a sufficient number of branched chain acids occur in the natural fats to
account for the varying values given. The highest melting point obtained
by the author for this acid was 59° while Hinsberg and Roos give 57°. An
important point to be noticed is that for the synthetic normal acid of this
composition, synthesised by Krafft [1879] and also by Le Sueur [1905], the
melting point given is in both cases lower than the one found for the yeast
acid, being stated as 51° by the former and 53° by the latter worker. The
melting points of the simpler derivatives also disagree with those of the
corresponding derivatives of the synthetic acid. The work was started in
the belief that the substance was probably a mixture, but all attempts to
separate it into more than one substance were unavailing and the only

Bioch, vir 23

342 A. NEVILLE

conclusion possible was that the substance was really a single acid of the
composition C,;H;,O, and, at the same time, that it was not the normal
pentadecoic acid.

In dealing with the unsaturated acids the author experienced considerable
difficulty in directly isolating any pure product, the simpler derivatives being
difficult to deal with and indicating no satisfactory means of separation.
By oxidation with potassium permanganate, however, very satisfactory
specimens of the corresponding di- and tetra-hydroxy acids were obtained
and these showed the presence in the original fat of the acids C,,H;,O,,
C,sH;,0, and C,,H;,0,. Hinsberg and Roos obtained results which they inter-
preted as showing the presence of the acids C,H»O, and C,,H,,O, though
they themselves query these formulae.

The phytosterol obtained melted at 145-147° and was apparently identical
with that obtained by the other workers from one sample of beer yeast. The
compound melting at 159° obtaimed by them from the majority of yeast
samples appeared not to be present in the sample under examination. |

PRELIMINARY EXAMINATION.

The crude fat was hydrolysed with alcoholic potash and after the alcohol
had been distilled off the crude dry soap was extracted with ether. From
the ethereal solution the yeast cholesterol was obtained. The crude soap
was then redissolved in water and after neutralisation of excess of alkali,
the fatty acids. were precipitated in the form of their lead salts and the
usual rough separation of the salts of the saturated and unsaturated acids
made by extraction with ether. The free acids were then liberated from
their lead salts by hydrochloric acid and treated as described below.

THE SATURATED Farry ACIDs.

When the dark brown solid mass, obtained from the ether-insoluble lead
salts, was distilled under a pressure of 100 mm. by far the larger part passed
over between 250° and 260°. This was redistilled and came over at about
215° under 15 mm. pressure. In the first distillation the process was not
pressed to an extreme extent and the residue left in the flask was purified
by crystallisation only. The portion which had been twice distilled was
crystallised several times from methyl alcohol and then melted at 54-55°.
An iodine absorption determination gave however an iodine value of 29'5,
and, as this was only very slowly altered by crystallisation, the acid was dis-
solved in dilute potash solution and treated with 15 per cent. permanganate

A. NEVILLE 343

solution so as to convert the impurities consisting of unsaturated acids into
the corresponding hydroxy-acids. The solution having been treated with
sulphur dioxide, the precipitated acids were filtered off and treated with
a small quantity of ether. Filtration from the hydroxy-acids and crystallisa-
tion of the substance obtained from the ethereal solution gave a white
crystalline body, melting at 59° and showing no iodine absorption when
treated with Hiibl’s solution.

Pentadecoic Acid.

On analysis the substance, whose preparation has just been described,
gave the following result :
0°1686 g.; 0°4611 CO,; 0°1878 H,O.
Found 74°59 °/,C; 12°38 °/,) H.
Cale. for C,;H 3,0, 74°38 9/,C; 12°39 9/, H.
This substance was evidently the same as that separated by Hinsberg and
Roos and described by them as being definitely pentadecoic acid.

A determination of the molecular weight by the freezing point method
showed that 04310 g. substance dissolved in 242174 g. glacial acetic acid
gave a depression of the freezing point of 0°295°; whence molecular weight
= 235. Calculated for C,,;H,,O, = 242. On titration with alcoholic potash
18344 g. acid took 14°9 ec. N/2 KOH for neutralisation.

KOH required = 22-74 °/, weight of acid.
Cale. for C,;H,,0,=23°14 °/, weight of acid.

The silver salt, prepared by precipitating a neutral solution of the
potassium salt with silver nitrate, was obtained as a white amorphous
precipitate and gave the following result on ignition :

0-4288 g. salt; 0°1308 g. metallic silver. Molecular weight of the acid=247.
Cale. for Cy5H3902= 242.
The above estimations point to the substance being pentadecoic acid or a
mixture with practically the same composition. If not a single compound the
substance was most likely to be a mixture of two of the three acids, myristic,
palmitic and stearic acid. Referring to the determinations by Heintz of the
melting points of mixtures of any two of these acids it was found that with
a melting point of 59° the substance might be a mixture of (a) 15°/, myristic
with 85°/, palmitic acid, (b) 40°/, myristic with 60 °/, stearic acid, (c) 45°/,
palmitic with 55 °/, stearic acid. Assuming the substance to be one of these
mixtures, weighed quantities of pure myristic, palmitic or stearic acids were
added and the melting point curves plotted. In no case did the curve agree
with that for any pair of the above three acids. The melting point for a
23—2

344 A. NEVILLE

mixture of all three of the above acids lies below that actually determined.
Other methods for the separation of a possible mixture in the substance were
tried but without result. Thus the fractional precipitation of the magnesium
salts and the continual crystallisation of the lithium salts from alcohol were
without effect. To obtain further evidence of the homogeneous nature of
the substance several derivatives were prepared and compared with the
corresponding compounds of the better known fatty acids.

The methyl ester, prepared by passing gaseous hydrochloric acid into a
mixture of the acid and methyl alcohol, was obtained as a white crystalline
substance which could be crystallised from a small quantity of methyl alcohol
and melted at 26°, which is considerably lower than the melting point for
either methyl palmitate or methyl stearate.

The anilide was prepared by boiling the acid for some hours with excess
of aniline. It crystallised from alcohol in pearly scales and melted sharply
at 86-87°.

Myristic anilide melts at 84°, palmitic anilide at 90°5° and stearic anilide
at 94°.

Analysis ; 0-2117 g.; 10°3 cc. moist N at 16° and 772 mm. N=4°48 °/,.

Cale. for CoH; ON N=4°41 %,.

The amide was prepared by heating together equal quantities of
phosphorus pentachloride and the acid in chloroform. The solvent and
phosphorus oxychloride were distilled off under slightly reduced pressure
and the residue poured into strong ammonia. It crystallised in small flakes
from 50 °/, alcohol and melted at 94-95°.

The amide of the synthetic pentadecoic acid, prepared by Le Sueur,
melted at 102°5°, while palmitic amide melts at 107°.

0°1443 g.; 7:0 cc. moist N at 15° and 760 mm. N=5°71 °/).
Cale, for C1;H3,0N N=5'81 %,.
The whole of the evidence points to the substance being a definite compound
of the composition C,,H;,O,, but at the same time being not identical with
the normal synthetic acid of this composition.

Arachidic Acid.

It has been mentioned that a small quantity of material remained behind
in the flask when the pentadecoic acid was distilled under reduced pressure.
This small quantity of material was several times crystallised from alcohol
and was finally obtained as a white crystalline body which melted sharply

ait shal.

1 ern, Reyer -

A. NEVILLE 345

On analysis it gave the following results :
0°1212 g.; 0°3440 g. CO.; 0°1385 g. H,O.
O=77-40 °,, H=12-70 %,.
Cale. for CopHyo02 C=76°92 °/,, H=12-82 %/).
0°4585 g. required 14:2 cc. N/10 KOH for neutralisation.
KOH required =17°34 °/, of weight of acid.
Cale. for CopHyoO2 17°95 °/, of weight of acid.

The substance was optically inactive, did not absorb iodine from Hiibl’s
solution and did not react with acetic anhydride. The substance was almost
certainly therefore arachidiec acid.

THe UNSATURATED ACIDS.

The unsaturated fatty acids, liberated from the ether-soluble lead salts,
formed a dark brown oily mass which behaved as a “semi-drying” oil when
exposed in films to the air. Attempts were at once made to purify this mass
by distillation under reduced pressure but without much success, as de-
composition of some portion took place and the distillation was rendered
useless. Distillation in steam was also of little avail the quantity passing
over being very small. In an attempt to obtain some separation by means
of salts, it was noticed that the barium salts, a sticky mass containing a small
quantity of water, were soluble to some extent in benzene, toluene and
ether, but almost insoluble in cold alcohol while with hot alcohol the mass
yielded an apparently homogeneous glue with the solvent. These observa-
tions together with an iodine value of 138 for that portion of the acids whose
barium salt was soluble in benzene, point to the presence in the mass of
some acids of the linoleic series and the presence of some proportion of
these acids would account for the “semi-drying” properties of the oil.

Attempts at direct separation of the mixture having failed, the whole
“unsaturated” fraction was neutralised with caustic potash, dissolved in
a large volume of water and an equal volume of 1:5 per cent. potassium
permanganate solution run into it. After standing for ten minutes sulphurous
acid solution was added and the precipitated acids filtered off. The filtrate
was found not to contain any appreciable quantity of hydroxy-acids, thus
showing the absence in the original mixture of any appreciable quantity of
acids of the linoleic series.

The acids precipitated by sulphur dioxide were washed with small
quantities of ether to remove unoxidised material and afterwards treated
with large quantities of the same solvent. The ether-soluble portion was
crystallised from alcohol and the purified dihydroxy-acid is described below.
The portion insoluble in ether was crystallised from large quantities of water
and was identified in the manner hereafter described as a tetrahydroxy-acid.

346 A. NEVILLE

The Dihydroxy-acid.

This substance was obtained as a well crystallised white powder, insoluble
in water, slightly soluble in ether but soluble in hot alcohol. It melts at
124-125°.

Analysis : 0°1568 g.; 0°3894 g. CO.; 0°1566 g. HO.

C=67°73 J,, H=11:09 %/,.
Cale. for Cy7H3202(0H)2 C=67°55 hs nb oine
0°6453 g. required 21:0 cc. N/10 KOH for neutralisation.
KOH=18-22 °/, of weight of acid.
Cale. for Cy;H320.(0H),, KOH=18-54 °/, of weight of acid.
The acetyl derivative gave the following result :

0:8212 g. gave acetic acid equivalent to 53-7 cc. N/10 acetic acid.
Acetic acid=40°45 °/, weight of substance taken.
Calc. for Cy;H2404 39°73 %/,.

The above figures point to the substance being a dihydroxy-acid of the
composition C,,H;,0, and the melting point and general properties agree
exactly with a substance of this composition described by Fahrion [1893].
This substance was shown by Ljubarsky [1898] to be probably a mixture
of dihydroxypalmitic acid and dihydroxystearic acid, which mixture the
latter author found to behave in many respects as a chemical individual.
In view of the fact that an unsaturated acid of the oleic series with the
composition C,;H;,0, has hitherto not been separated with certainty from
natural fats, it seems more probable that the substance separated was the
mixture which Ljubarsky claims it to be, than a single chemical individual,
although no separation was possible by repeated crystallisation from alcohol.

If this is so the original fat must have contained two unsaturated fatty
acids of the oleic series with compositions C,,H,,O0, and C,,H;,O, respectively.

Tetrahydroxy-acid.

The oxidation product insoluble in ether but cerystallismg from large
quantities of water was analysed with the following results:

0°1465 g.; 0°3330 g. CO.; 0°1380 g. HO.
C=61599/0/ Hi 0-460
Calc. for C1gH3202(0H), C=62:07 Uh H=10°34 hi
1°0878 g. requires 32°3 ec. N/10 KOH for neutralisation.
KOH=16°'63 °/, of acid.
Cale. for CjgH3202(0H), KOH=16:18 °/,.
04321 g. when converted into acetyl compound and hydrolysed gave acetic acid
equivalent to 50:5 ec. N/10 acetic acid.
Acetic acid obtained = 70°12 °/, of weight of acid taken.
Cale. for CygH3202(OH), 68-96 °/).

_—

A. NEVILLE 347

The substance is a white crystalline body of silky lustre, melting at
156°, soluble in hot alcohol and glacial acetic acid but insoluble in nearly
all other solvents. The general properties of the acid, together with the
analyses given above, point to the substance being a tetrahydroxy-acid of
the composition C,,H,(OH),. COOH. The best known acid of this composi-
tion is tetrahydroxystearic acid (sativic acid), formed by the oxidation of
linoleic acid, but this melts at 173-174°. Apparently, however, there are
several isomeric acids of this composition, for other observers have reported
acids melting at 152° and 165° and the acid separated may be identical
with one of these. At any rate, the isolation of this substance points
definitely to the existence in yeast fat of an unsaturated acid of the linoleic
series with the empirical composition C,,H;,O,, that is, an acid certainly
isomeric, but perhaps not identical with ordinary linoleic acid,

YEAST CHOLESTEROL,

The ethereal extract of the crude potassium salts obtained from the first
hydrolysis of the yeast fat was worked up in the usual way for the yeast
cholesterol. The product obtained, after being crystallised several times
from alcohol, melted at 145-147°, and this melting point was not altered by
repeated crystallisation. It had [a]p=—75°54° and gave similar colour
reactions to the majority of the phytosterols. It was apparently identical
with the cholesterol obtained from one sample of yeast by Hinsberg and
Roos. The cholesterol, melting at 159°, which these authors most frequently
obfained from yeast was not isolated in spite of a careful search being
made for it.

SUMMARY.

A comparison of the results obtained by Hinsberg and Roos and by the
author is shown in columns I and II respectively of the following table.

I II
Saturated acid Cy;H4 0s. Saturated acid Cj;H 02.
M.P. 56°. MER. 59°.
Not identified in the mixture. Saturated acid CopHyyO2. M.P. 77°.
Arachidic Acid.

Unsaturated acid Cy,H» 02. Not identified in the mixture.

Unsaturated acid C,gH3,09. Unsaturated acids CygH3)04 and C,gH340o.
Presence shown by their oxidation
products.

Not identified in the mixture. Unsaturated acid C,;gsH3:02, shown by its

oxidation product.

Yeast cholesterols melting at 159 and 145- Yeast cholesterol melting at 145-147°.

148°.

pd anata

me _ na y > ae a i —_
x - ties . ~~

=a Ss vl ase ]
348 ‘A. NEVILLE " ‘

The thanks of the author are due to Prof. T. B. Wood and Dr F. ce
Hopkins, F.RS., for kindly placing the material at his disposal and for the
interest they have taken in the work throughout.

REFERENCES.

Fahrion (1893), J. Soc. Chem. Ind. 936.

Hinsberg and Roos (1903), Zeit. physiol. Chem. 38, 1.
Krafft (1879), Ber. 12, 1671.

Le Sueur (1905), J. Chem. Soc. 87, 1898.

Ljubarsky (1898), J. pr. Chem. [2] 57, 19.

XXXIV. THE INFLUENCE OF THE CARBON-
ATES OF THE RARE EARTHS (CERIUM,
LANTHANUM, YTTRIUM) ON GROWTH AND
CELL-DIVISION IN HYACINTHS.

By WILLIAM HOWEL EVANS.
From the Biochemical Department, University of Liverpool.

(Received May 3rd, 1913.)

The effects of various organic and inorganic bodies on growth and cell-
division in both plants and animals have been studied by several researchers
in this laboratory during the past few years. Moore, Roaf, and Whitley
[1905] investigated the action of acids and alkalis on the development of
the fertilized eggs of the sea-urchin, Echinus esculentus. Their inquiry was
directed to the effects of variation in hydrogen and hydroxyl ion concentra-
tion in the medium (sea-water) in which the eggs were allowed to develop,
and they found that within certain narrow limits of concentration, alkalis
favoured cell-division, whereas acids were invariably fatal. Small additions of
acids inhibited cell-division and growth, and at a concentration of 0°001 molar
practically all cell-division was stopped. A slight increase in alkalinity
favoured development, at the same time producing irregularities in cell-
division. Beyond the optimum concentration of alkali, exceedingly irregular
division, resulting in particular in the production of multi-nucleated cells,
was observed. At a concentration of 0°0015 M. of caustic alkali, however,
cell-division was completely arrested.

Whitley [1905] observed similar effects on the eggs of plaice, and also
noted the action of the indicators phenolphthaléin and dimethylaminoazo-
benzene on the eggs both of plaice and echinus, He found that while
phenolphthaléin was deadly to the eggs of echinus, it was harmless to those
of the plaice. On the other hand, the azobenzene derivative quickly killed
the latter, and appeared, if anything, to have a favourable effect on the
development of the former.

Moore, Knowles, and Roaf [1908] extended the observations to plants,
using for their experiments the common hyacinth. They obtained similar

350 W. H. EVANS

results in regard to the influence of acids and alkalis, and they also showed
that the cation had a specific effect, potassium being much more stimulating
than sodium to both rootlets and foliage leaves. The phosphatic ion also
had a special effect on the flower, causing an increase in size at optimum
concentration, and at higher concentrations an irregular inflorescence, with
packed florets on a dwarfed stalk.

Histologically they observed depression of nuclear division with acids,
and thickening of the cell walls; with alkalis, increased nuclear division,
changes in chromosomes and irregular figures, while the cell outlines
became obscured.

Coppin [1912] studied the effect of allantoin and other purine derivatives
on the growth of hyacinths, and also salts of organic acids such as sodium
huminate, sodium malate, and sodium oxalate.

These latter substances, as well as sodium urate, had a stimulating effect
on the growth of the hyacinths, but allantoin and the other purine substances
inhibited both growth and cell-division.

Working on somewhat different lines, Ransom [1912] observed the action
of caffeine upon the germination of seeds. He used a large number of
different seeds, and his method consisted in soaking the seeds for a short
time in his caffeine solutions before sowing them in the usual manner. He
found that caffeine, even in a very dilute solution, had a powerful effect
in retarding germination and growth; while a concentration of 1 per cent.
in many cases completely inhibited germination.

In the present inquiry, salts of the rare earths were used to test whether
these produced any physiological effects on growth and cell-division. For
this purpose the carbonates of cerium, lanthanum, and yttrium were selected.

Preparation of carbonates.

The carbonates were prepared, according to Moissan, in the following
ways:

(1) Cerium. A solution of cerium nitrate was treated with ammonium
carbonate, and the precipitate of cerium carbonate filtered off, and thoroughly
washed.

(2) Lanthanum and Yttrium. The hydroxides of the metals were taken,
suspended in water, and thoroughly saturated with carbon dioxide over a
period of several hours. The precipitates were then filtered off.

W. H. EVANS 351

Preparation of the carbonate solutions.

Owing to the very slight solubilities of the carbonates of cerium, lan-
thanum and yttrium, some difficulty was experienced in preparing solutions.
Finally two or three grams of each of the respective carbonates were
suspended in about two litres of Liverpool tap-water’, and carbon dioxide
was passed into the bottles for one or two hours, thus ensuring complete
saturation. After being allowed to stand overnight, the undissolved residues
were filtered off.

These clear filtrates were the actual solutions used in the experiments.
Their concentrations were estimated by taking a measured volume of each,
evaporating to dryness, and weighing the residue. The following figures
were obtained:

Cerium cs ae fs 0°007 °/,
Lanthanum ... Ae oes 0:01 °/,
Yttrium Sat Bs aA 0:017 °/,

In all probability they contained a mixture of the respective carbonates
and bicarbonates.

Effect on the growth of hyacinths.

The plants used were a common variety of hyacinth. Healthy bulbs of
as nearly uniform size as possible were selected. They were placed in
hyacinth glasses which had been blackened on the outside with black lacquer
to prevent action of light upon the rootlets. The hyacinth glasses held about
450 cc. and the solutions mentioned were filled in until they just wetted
the bulbs. In addition, controls were grown under precisely similar con-
ditions in Liverpool tap-water. Water was added to the glasses from time
to time to make up for the loss due to evaporation. A few of the ends
of the growing rootlets were cut off on the twenty-fifth day for the purpose
of studying the effects of the solutions on cell-division and nuclear changes.
These were all immediately fixed in Flemming’s solution, cut in paratiin,
and stained for nuclear figures by Heidenhain’s iron-alum, haematoxylin
method.

On the twenty-fourth day, a measured volume of the fluid was removed
from each of the different solutions, filtered, and evaporated to dryness,
organic matter being removed by ashing with ammonium nitrate. The
residues were then weighed. Results:

1 Liverpool tap-water is very pure surface water and practically free from inorganic salts.

352 W. H. EVANS

24th day Original
Cerium 0:0068 °/, 0:007 °/,
Lanthanum 0:0076 °/5 0:01 °/,
Yttrium 0:0066 °/, 0:017 °/,

Thus in the case of the cerium carbonates there had been little or no
absorption ; but in the other two cases, particularly in that of yttrium, there
had been marked absorption.

In addition measurements and observations were made at intervals of
the growth and conditions of the plants, the pomts noted being the length
of the green leaves, length and condition of the rootlets, and the condition
of the florets and the flower spike. These observations and measurements
will be found set out in the accompanying Table (p. 353).

While the results obtained were not in all cases concordant, a few definite
points may be made. In the first place, all the bulbs in the experimental
solutions reached maturity before the controls in water.

Secondly, the plants in the cerium carbonate solution were the first to
attain maturity, though their development was not so marked as that of
those in the lanthanum carbonate, which followed next in point of time.

The effect of the yttrium solution was somewhat anomalous, since while
all the plants matured a few days before the controls, the rootlets were
dwarfed, and looked yellow and unhealthy from the beginning.

The lanthanum ion seems to have a special effect on the flower stalks,
resulting in very tall plants. No irregularities manifested themselves in the
inflorescences.

HISTOLOGICAL INVESTIGATION OF GROWING ROOT-TIPS UNDER THE
INFLUENCE OF THE ABOVE REAGENTS.

The varying effects produced by the different metallic ions are best seen
from the accompanying photomicrographs (Plate I). Speaking generally
it may be stated that the cerium and lanthanum ions have a decidedly
stimulating effect on cell-division in the rootlets, while that of yttrium has
a deleterious effect.

The following are brief notes of the histological examination of slides
prepared from each of the twelve plants as mentioned in the Table. They
were stained by Heidenhain’s iron-alum, haematoxylin method.

1. Ceriwm carbonate. (Nos. 2 and 3.) (See photomicrograph 2.)

In both these preparations a noticeable feature is the beautiful regularity

of the arrangement of the cells, the sections showing a large number of

PatoqTA G-G1 G-L OT G Or g G-L oy G
uado sqyar0[ iy
ST Geb, cT gL S-61 GL él 9 S-TL 9 Or 9 ¥
uado sjato, 7
91 al ST OT ST OT G-G1 OT G-GL OT OT OT L
uado sya 0] J
aad = 8 € c-L § G-L § GL § 9 GG G
uado s]910[ J
= Fa GT G S-6I G £41 G $-6L G GL G G
uado sqyar0],iT :
= = = a = = S-GL 9 OT ¥ Goh 14 GG
uado szolo]
=. > a Ta C-61 g G-61 g OT § 8 ¥ i
uedo sya10[ 7
= = = = = ST 8 G-6T 8 OT GL L
uado
$}010Ff ITV
= 7 = = == a LT ST LT ST OT ct G-6T
uado
$10.10} [TV
= = a = = aa G-LT ST ST ST al 61 GL
uado
8}0108 TV
= ae aj a a a =F = L I L i! I!
poraygIAA 99301 yoorT
quatadopasep ON G:F T it
SOABATT SJOOY SOAVATY SJOOY SOAVOTT SJOOY SaavaTyT sJOOY SeavaT soo saavay sjzooy ATUOs}ooY
— FS or oOr'’rowrr'vTY’ '“'- OOOO hop prez

Sep wags ‘Fre Avp UF, ‘82 Gad Aep qI0L ‘F3“Gaq ep astg ‘eT ‘qeq Ap MLS ‘IT “Gea APP UPS ‘BGT —“g -uee

Ud UL saanay pun szoo4 fo JuaUMaimsvayy

It

OL

1D

G

19VVA
s[oaquog

"/y L10-0 9y8U0q
“ABQ UUNTAGI A

*/o 10-0 ayBU0G
-1BQ UNURYUBT

°/) LO0-0 238U0q
-1eQ WnITaD

IGT “09 poqreyg

354 W. H. EVANS

rather small, closely packed, square cells, containing many dividing nuclei.
The cell-walls are clearly defined, and the cytoplasm is granular and stains
well. In many cases the nuclei are elongated, and show one or two nucleolus-
like chromatin dots surrounded by a clear space.

2. Lanthanum carbonate. (Nos. 4, 5 and 6.) (See photomicrograph 3.)

The same regularity of the arrangement of the cells is seen in these
sections, but it is not quite as marked as in the cerium preparations. Many
dividing nuclei are seen, and the nucleolus-like dots above referred to are
also conspicuous. The cytoplasm is faintly granular, and the cell walls are
less sharply defined than in Nos. 2 and 3 (cerium).

3. Yttrium carbonate. (Nos. 7, 8 and 9.) (See photomicrograph 4.)

In this series hardly any dividing cells are to be seen. The arrangement
of the cells is irregular. The nuclei are very deeply stained, and irregular
in size and shape. The cell walls are not distinct, and the cytoplasm is
scanty, ill-defined, and very faintly staining. The whole appearance is that
of an irregular mass of cells, with scattered deeply-stained nuclei, and
presents a very different picture from the compact and regular arrangement
shown in the cerium and lanthanum series.

4. Controls—tap-water. (Nos. 10,11 and 12.) (See photomicrograph 1.)

In these sections no great amount of cell-division is noticeable. The cells
are fairly regularly arranged, and the nuclei are deeply stained. They are
for the most part in the resting condition, and many exhibit the darkly
staining dots resembling nucleoli, each surrounded by a clear space.

oa

CONCLUSIONS.

1. Marked effects are produced upon the dividing cells of hyacinth
rootlets by the addition of the carbonates of cerium, lanthanum and yttrium
to the medium. The concentration of these substances necessary to produce
physiological effects is very small.

2. The cations produce diverse effects ; lanthanum especially, and cerium
being favourable to growth and cell-division, while yttrium is unfavourable.

3. The lanthanum ion has a special effect on the flower stalk, causing an
increase in length.

©
E z 2° es, es
al
a é °% " » & : e o>
ibe &s 9 2¢ « %

oe aut §@2 0° = & © ¥

vet chet an ean,

. e é . 3° segs Pd
” er 6 “@2,%, a

NO
=.
aS
har
. 3
EI
“2
at
+

Zs
#.
oy
1

THE BIOCHEMICAL JOURNAL, VOL. VII.
OF
_ =
oe
4%
a)
cz
aa
ye
Pad
<e! o
Fig

460 diameters.

70
‘ 0

‘OO

Cerium carbonate

x 460 diameters.

tapwater.

Control

160 diameters,

‘O17

Yttrium carbonate

< 460 diameters.

Lanthanum carbonate 0:01 9/,.

AS

W. H. EVANS 355

4. The cytological effects are best seen in the accompanying photographs.
| In the cerium and lanthanum preparations there is a marked increase in cell-
division, accompanied by a beautiful regularity in the arrangement of the
cells. With yttrium there is a diminution in cell-division, and the cells are
irregularly arranged.

I should like to take this opportunity of thanking Professor Benjamin
Moore for suggesting the line of research, and for his kindly help and
criticism throughout the work.

REFERENCES.

Coppin (1912), Biochem. J. 6, 416.

Moore, Knowles and Roaf (1908), Biochem. J. 3, 279.

Moore, Roaf and Whitley (1905), Proc. Roy. Soc. B. 77, 103.
Ransom (1912), Biochem. J. 6, 151.

Whitley (1905), Proc. Roy. Soc. B. 77, 137.

"
o
>

XXXV. ON THE CHEMICAL NATURE OF SUB-
STANCES FROM ALCOHOLIC EXTRACTS OF
VARIOUS FOODSTUFFS WHICH GIVE A
COLOUR REACTION WITH PHOSPHOTUNG-
STIC AND PHOSPHOMOLYBDIC ACIDS. (PRE-
LIMINARY COMMUNICATION.)

By CASIMIR FUNK, Beit Memorial Research Fellow, AND
ARCHIBALD BRUCE MACALLUM, Beit Memorial Research Fellow.

From the Biochemical Department, Lister Institute.

(Recewed May 22nd, 1913.)

Two colour reactions have recently been described, one with phospho-
tungstic acid for uric acid {Folin and Macallum, 1912], the other with
phosphotungstic and phosphomolybdic acids for polyphenols [Folin and Denis,
1912], which they recommend for the determination of these substances.
One of us previously observed during the fractionation of yeast and rice-
polishings [Funk, 1913] that the vitamine-fraction constantly gave a blue
coloration with both reagents, that the reaction gradually disappeared during
the further purification of the products and at the same time that the
curative power for polyneuritis in birds slightly diminished. As these
reacting substances might be of some importance for the process of curing,
we have tested a large number of compounds with both reagents, in order to
obtain some information as to the chemical nature of these substances, and
we have found that a certain number of them which occur in nature show
the colour reactions.

With the uric acid reagent certain purine bodies and some tyrosine
derivatives give a very strong reaction. The alcoholic extracts of different
foodstuffs, especially those which are known to be rich in vitamines, give in
general both reactions very markedly. The substances giving the reaction
seem to pass into the phosphotungstic acid filtrate and we are proceeding to
isolate them.

C. FUNK AND A. B. MACALLUM 357

The reactions with purine and pyrimidine derivatives.
Substance Uric acid reagent Phenol reagent

Xanthine tay os 3 - ++
Hypoxanthine ee ax ae ~ +
Paraxanthine fhe a “ep - <
Guanine oes wT: sa - ++
Alloxantin be ae ok ++4 foto
Alloxan a ar or ; _ =
Thymine oF int ois = a
Theophylline sed Bi: rae = z
3-Methyl-uric acid ..- or 28 + ots
ie ” ” bos we ae trace ++
Hydantoin or vd ar = =
Uracil . : vat ay = =
Guanidine carbonate + ae _ =
Hydantoic acid... Hee sec = =
Allantoin as sey ae = =
Adenine nas or a = =
Uridine a as ae - =
Guanosine Ae at See = trace
Adenosine bie e “¢ = =
Cytidine nitrate ... er i = =
Yeast nucleic acid ... sud a trace trace
Thymonucleie acid Hic ae = =

Tyrosine derivatives.

l-Tyrosine Sat Sine Sa = ast ce
Nitro-tyrosine ae ete ce = =
3-4-Dihydroxyphenylalanine ... AS Sr ae se rita cts
2-Aminotyrosine ... is Aa ae Gr SE eich ke
3-Aminotyrosine ... Ba 560 a5 Se ar aS a
1-Tyrosine anhydride ae fet = =?
Glycyl-l-tyrosine ... ae fe trace a

We have also investigated a number of amino-acids, polypeptides and
diketopiperazines, not including tryptophane, oxy-tryptophane and oxy-proline,
all of which were entirely negative to both reagents.

Foodstuffs.
Ceridin (alcoholic extract of yeast) aes trace +
Zymin ... Sob trace +
Alcoholic extract of rice- pola aah + aia c
Subs. Cy,Hj )09N; from vitamine-fraction of
yeast = =
Subst. Co9H»309N; io the same fenotiod - +
Nicotinic acid from yeast and rice es - =
CogH a0 Ny from vitamine-fraction of rice ~ -
Alcoholic extract from caseinogen (crude) + 4
Alcoholic extract of milk see he -—- +
Whey powder from milk ; ute + +
'Filtrate from milk precipitated by acid... trace +
Alcoholic extract of whey powder eae trace +
Cod liver oil and the aqueous extract aes + ++
Alcoholic and aqueous extract of ae
chicken sarcoma st = trace et

bo
_

Bioch, vi

358 C. FUNK AND A. B. MACALLUM

_ We see from the table that the reactions are very specific for purine de-
rivatives and polyphenols and they may therefore serve as a guide as to what
groups the substances giving the reactions may belong. As the reactions
are very sensitive, it seems to us that they might be used to ascertain the
purity of phosphatides and other substances like caseinogen which are
prepared from foodstuffs, and for which up to the present we possess no
standard of purity. This test has already been found very useful in the
investigation of the vitamine-fraction.

The table for purine derivatives shows that a substitution of one hydrogen
atom in the purine ring lessens or destroys the power to give the uric acid
reaction. In the case of the phenol reagent this is also brought about when
two hydrogen atoms are substituted. The colour reactions with tyrosine
derivatives-and alloxantin were remarkably stable as compared with the
others.

REFERENCES.
Folin and Macallum junr. (1912), J. Biol. Chem. 11, 265; 13, 363.

Folin and Denis (1912), J. Biol. Chem. 12, 239.
Casimir Funk (1913), Brit. Med. J. April 12th; J. Physiol. 46, 173.

XXXVI. THE PRODUCTION OF ACETALDEHYDE
DURING THE ANAEROBIC FERMENTATION
OF GLUCOSE BY BACILLUS COLI COMMUNIS
(ESCHERICH).

By EGERTON CHARLES GREY, 1851 Hahibition Scholar.
From the Biochemical Laboratory of the LInster Institute.
(Received June Sth, 1913.)

Acetaldehyde has long been known as a product of alcoholic fermentation
by yeast [see Roeser, 1893]. It has likewise been found to be formed by the
leaves of higher plants when their metabolism is restricted to anaerobic
conditions [Kostytschew, 1913].

In view of this wide distribution, it was to be expected that acetaldehyde
would be a product of bacterial fermentation. This is the case.

The quantity of acetaldehyde found by Roeser in various wine musts after
a fermentation lasting from five to fifteen days, varied from 20-200 milligrams
per litre of the fermented fluid. In the case of B. coli communis acting
on glucose, one litre of fluid containing originally 15 grams of the sugar
yielded in 15 days when half the sugar had been fermented 2°34 mg. of
acetaldehyde, while, at the same time, 1375 grams of alcohol had been
produced. Since in the case of the true alcoholic fermentation quoted above,
the quantity of alcohol formed was certainly twenty times that produced in
the experiment with B. coli communis, it will be seen that the ratio of acetal-
dehyde to alcohol is of the same order in the two cases.

It is not at present settled, in the case of yeast fermentation, whether
acetaldehyde is a primary product or whether it results from the oxidation of
preformed alcohol, the following result is therefore significant in showing that
the production of acetaldehyde by B. coli communis is related to the forma-
tion of alcohol, carbon dioxide, and acetic acid, rather than to the other
products.

The experiment consisted in comparing the production of acetaldehyde
from normal B. coli communis (Escherich) with that from a strain derived
from the normal organism by selection with sodium chloracetate [see Penfold,
1911, and Harden and Penfold, 1912]. This derived strain had completely

24—2

360 E. C. GREY

lost the power of producing gas from carbohydrates. That it was closely
related to the original organism was demonstrated by testing the agglutinating
power of both the original and the derived organism towards a rabbit serum
obtained after inoculating with the former. Both organisms agglutinated
completely with the serum up to a dilution of one part of the serum in 25,600
of normal saline solution.

REACTIONS EMPLOYED FOR THE DETECTION AND DETERMINATION
OF ACETALDEHYDE.

(1) Rimini’s reaction [1898].

A few drops of diethylamine are added to a similar quantity of a solution
of sodium nitroprusside. A blue colour appears which rapidly fades (2-3
minutes). If this solution be now diluted with 1-2 ¢.c. of water a pale yellow
or yellow-green solution is obtained but if diluted with 1-2 cc. of a solution
containing acetaldehyde a blue colour results characteristic of this aldehyde.
Though specific for acetaldehyde the reagent is not so delicate as that of
Schiff, and the coloration is moreover less permanent.

(2) Schiff’s reagent. This reagent was used for the determination of
acetaldehyde.

It is important to make sure that no alkali is present in the fluid to be
tested by Schiff’s reagent since even the alkalinity of tap water produces a
definite coloration. The maximum intensity of colour, by simple addition of
alkali, was obtained with 0:1 c.c. normal KHO, added to 50 cc. distilled water
containing 0°25 c.c. Schiff’s reagent. A greater concentration of alkali destroys
the colour, which is regained on the addition of acid. The solution before
examination was therefore always tested with phenolphthalein paper.

Acetone is said to give a coloration with Schiff’s reagent, but this effect
which is not produced for many hours cannot possibly be confused with the
reaction for acetaldehyde in which case the coloration reaches a maximum in
about twenty minutes and has very considerably faded in less than an hour,
It is important to emphasise this fact, since Mendel [1911] states that he
found acetone amongst the products of the action of B. coli communis and
certain other organisms on glucose.

DETAILS OF THE DETERMINATION OF ACETALDEHYDE.

(1) About 750 cc. of the fluid containing the liquid and solid products
of fermentation (acetaldehyde, alcohol, and calcium salts of volatile and non-
volatile acids etc.) was treated with oxalic acid in excess of that necessary for

er

E. C. GREY 361

the complete precipitation of the calcium. When the precipitate of calcium
oxalate had settled, as much as possible of the supernatant liquor was
removed with the aid of a siphon, and distilled. The first 400-500 c.c. of the
distillate was made alkaline with barium hydroxide and redistilled, using a
fractionating column, until 100 e.c. had been collected. In both distillations
the receivers were well cooled by means of ice.

50 c.c. of this solution were used for the determination of the acetaldehyde,
the method of procedure being an adaptation of that of Ryffel [1909].

(2) A standard solution of formaldehyde was prepared as described by
Ryftel. The strength of this solution was determined by comparing the colour
produced from it on the addition of Schiff’s reagent, with that produced on
adding the same quantity of the reagent to a solution of acetaldehyde pre-
pared by the distillation of a known weight of lactic acid.

002494 gram lactic acid was converted into acetaldehyde. The intensity
of colour produced on adding 0°5 c.c. Schiff’s reagent to 100 c.c. of this acetal-
dehyde solution was 19 times that of the standard formaldehyde colour as

measured by a Duboseq tintometer, 100 cc. of the standard formaldehyde

0 = = 001312 gram lactic acid.

Applying the ratio determined by Ryffel empirically for these conditions,
viz.: 04 mg. formaldehyde = 3-435 mg. lactic acid = 1:765 mg. acetalde-
hyde, it follows that 100 cc. of standard formaldehyde solution contains

solution was therefore equivalent to

O-4 x 0°01312 _ 1-528 x 1-765 _
—_ = 1528 mg. formaldehyde, and is equivalent to ae 6°471mg.
acetaldehyde.

(3) 50 cc. of the solution of formaldehyde described in paragraph (2)
was diluted fivefold and ae with the distillate described paragraph (1),
with the result:

standard 1
€ oO .
The ratio of 5 x distillate 1-33

From this it was calculated that the total amount of acetaldehyde in the
original litre of fermented fluid was 2°34 mg.

DISCUSSION OF RESULTS.

In order to demonstrate the relationship between the production of acet
aldehyde and that of alcohol, carbon dioxide, hydrogen and acetic acid, the
analyses of the products formed by the normal and the artificially selected
strain are compared below, the results being calculated to 100 grams of the
sugar.

362 E. C. GREY

Products from glucose
A.

t Strain artificially i
Normal selected by the
B. coli communis chloracetate method
per cent. per cent.
Hydrogen 0:42 Nil
Carbon dioxide 16-9 Nil
Alcohol 18-1 5:3
Acetic acid 18°5 10°8
Formic acid oer ilaleal
Lactic acid 36°8 68:0
Suecinic acid 0:7 0°8
Acetaldehyde 31:345 mg. Nil
(Glueose decomposed 7:°465 gram Terie:
| Acetaldehyde found 2°34 mg. Nil

It will be seen that the artificially selected organism has (as in the case
of the organism examined by Harden and Penfold), besides having lost the
power of producing gas, also produced less alcohol and acetic acid. Coinci-
dentally with these changes there has been an abolition or great reduction of
the yield of acetaldehyde.

In the light of these results we may consider the question of the
origin of the acetaldehyde.

If this substance were a secondary product formed by the oxidation of
alcohol through the activity of oxidases concerned in the cell-growth as was
suggested by Roeser for the case of yeast, then since there is a large excess
of alcohol in both cases, the amount of acetaldehyde formed in the two
experiments should be approximately the same, or at least, of the same order.
Moreover in view of the fact that not only was air excluded from the fermen-
tation flasks, but that in the experiment in which acetaldehyde was produced,
350 c.c. of hydrogen had been evolved, such an oxidation would seem unlikely
to have occurred. The strongly reducing conditions of the experiment may
be well seen from the experiment in which Harden attempted to employ
asparagine as a source of nitrogen. This substance, he found, was completely
reduced to ammonium succinate with corresponding diminution in the
hydrogen evolved.

It may be suggested therefore that the production of acetaldehyde and
part of the alcohol occur simultaneously. This would agree with the view
held by Kostytschew [1912] that acetaldehyde is an intermediate product in
alcoholic fermentation and would moreover suggest that part of the alcohol
produced by the action of B. coli communis on glucose, passes through at
least one of the same stages as that produced by the zymase of yeast.

From a consideration of the structure of glucose and mannitol, and the

E. C. GREY 363

fact that B. coli communis produces about twice as much alcohol from the
latter as from the former, Harden [1901] suggested that the terminal group
—CH(OH).CH,OH which occurs twice in mannitol though only once in
glucose, is related to the production of alcohol.

Mannitol CH,OH. CHOH- (CHOH), - CHOH - CH,OH.

Glucose CH,OH . CHOH - (CHOH),- CHOH - CHO.

If then the final group — CHOH.CH,OH conditions the production of
alcohol, the corresponding group in glucose - CHOH .CHO might stand in
the same relation to acetaldehyde.

Clearly since an accumulation of this product would be harmful to the
organism natural selection would have evolved the organism capable of trans-
forming the aldehyde by reduction into ethyl alcohol, or oxidation to acetic
acid.

Further results will be shortly forthcoming with regard to the production
of acetaldehyde from other substances allied to glucose.

SUMMARY.

(1) Acetaldehyde has been detected as a product of the action of B. coli
communis on glucose, under anaerobic conditions.

(2) By artificial selection of B. coli communis by means of growth on
sodium chloracetate, strains of the original organism have been obtained
which produce either a greatly diminished amount of acetaldehyde or none
at all.

(3) It has been found that the production of acetaldehyde is related to
the formation of alcohol, carbon dioxide and hydrogen rather than to the
other products. This has been ascertained by a comparison of the products
formed by normal B. coli communis with those from an artificially selected
strain produced by growth on agar containing sodium chloracetate.

(4) It is therefore suggested that acetaldehyde is a primary and not
a secondary product of fermentation, and also that the process of alcohol
formation by B. coli communis is in part analogous to the alcoholic fermenta-
tion set up by the zymase of yeast and to processes which occur in the leaves
of higher plants.

REFERENCES.

Harden (1901), J. Chem. Soc. 79, 610.

and Penfold (1912), Proc. Roy. Soc. B. 85, 415.
Kostytschew (1912), Zeitsch. physiol. Chem, 79, 130.
(1913), Zeitsch. physiol. Chem. 83, 105.

Mendel (1911), Centr. Bakt. Par. 11. 29, 290.
Penfold (1911), Proc. Roy. Soc. Med. 97.

Rimini (1898), Chem. Zentr. 2, 277.

Roeser (1893), Ann. Inst. Pasteur. 7, 41.

Ryffel (1909), J. Physiol. 39. Proc. v.

XXXVIL THE BIOCHEMICAL SYNTHESIS
OFTHE FATEY ACIDS.

By IDA SMEDLEY (Beit Memorial Fellow) AND
EVA LUBRZYNSKA.

Biochemical Department, Lister Institute.

(Received June 10th, 1913.)

The methods by which fat is formed in the living organism remain at
present completely unknown to us. Even the chemical reactions by which
fatty acids are built up, comparatively simple though these must be, have
hitherto found no satisfactory explanation. The one fact which appears to
be completely established, by a large mass of expérimental evidence, is that
the carbohydrate of the food may be converted into fat inside the living
organism, although neither the place where this change takes place nor
the method by which it is accomplished is known with any degree of
certainty.

The evidence as to the formation of fat from protein is less convincing,
but it is at any rate possible that a conversion of protein into fat may also
take place.

But little may be learnt from an attempt to correlate the composition
of the fat stored up in the organism with the nature of the food supplied.
There is abundant evidence that fatty acids taken in the food may be merely
stored up unchanged in the body. The glycerides of palmitic, stearic and
oleic acids are the constituents which most generally occur. Acids belonging
to more highly unsaturated series than oleic acid have been demonstrated
but these are more probably connected with further changes in the building
up of fat into complex molecules, possibly of the nature of lecithin, than with
the synthesis of the fatty acids themselves.

Two of the most prolific factories of fat are perhaps to be found (1) in
plants, in such nuts as that of the cocoa-nut tree (Cocos Nucifera), where
an abundant transformation of carbohydrate into fat must take place, and
(2) in animals in the active mammary gland.

In both these instances, where a comparatively rapid conversion of

I. SMEDLEY AND E. LUBRZYNSKA 365

carbohydrate into fat is probably taking place the resultant fats are charac-
terised by the presence of considerable quantities of the lower fatty acids.
In cocoa-nut oil, the acids containing the even numbers of carbon atoms from
six to eighteen, in butter from four to twenty, have been described. In these
acids the carbon atoms are linked in straight chains and there is no evidence
that any acid with a branched structure exists.

The question now arises whether the normal fatty acids present in butter
are products of synthesis or of degradation. Knoop (1904] and Dakin [1908,
1909] have shown that the fatty acids are broken down by oxidation of the
8-carbon atom; all the lower fatty acids present in butter may therefore be
derived by oxidation from the arachidic or stearic acids present.

Some evidence on this point may be obtained from agricultural experi-
ments; the problem has been directly investigated in an attempt to
determine the reason of the variations which occur in the proportion of
volatile fatty acids present in butter fat. Swaving (1906) carried out
feeding experiments in the North of Holland to determine the cause of
the low percentage of volatile soluble acids. Van der Zande and Siegfeld
showed that a diet rich in carbohydrate, e.g. turnips, increased the pro-
portion of the lower fatty acids and more recently Siegfeld [1907] and
Amberger [1907] have shown that the increase is more especially in the
insoluble volatile acids (i.e. caprylic, capric and lauric). Amberger, in a
series of experiments carried out on the same set of cows showed that
whereas food rich in protein such as malt germs diminishes the proportion
of lower fatty acids, food rich in carbohydrate such as turnips increases
this proportion. If the percentage of the lower fatty acids increases
with the amount of the carbohydrate in the food, it would appear more
probable that they exist as intermediate synthetic products on their way
to the higher fatty acids, than as degradation products. Such evidence as
exists is therefore in favour of a synthesis in which all fatty acids containing
even numbers of carbon atoms from four to twenty linked together in
straight chains are formed from carbohydrate in some way through the
agency of the mammary gland.

Previous Hypotheses as to the Nature of the Reactions by which Fatty
Acids are formed from Carbohydrate in the Animal Organism.

Emil Fischer suggested that stearic and oleic acids are formed by the
condensation of three hexose molecules or of six triose (glycerose) mole-
cules in such a way that a straight chain containing 18 carbon atoms

366 I. SMEDLEY AND E. LUBRZYNSKA

is formed. From this, by further processes of oxidation and reduction,
stearic and oleic acids are formed. Palmitic acid with its chain of 16 carbon
atoms would be compounded from two pentose and one hexose molecules.
Glucose, gluconic and glucuronic acids are suggested as the precursors of
the pentose molecules. In favour of this hypothesis it is difficult to find
any evidence either of a chemical or biological nature. Against it the
following considerations. may be urged:

(a) No laboratory method is known by which two hexose molecules
may be made to condense in such a way that a straight chain of twelve
carbon atoms is produced.

(b) Pentoses are known to exist in the organism in combination in the
nucleoproteins but there is no indication that the pentoses are in any way
connected with the formation of fat or with normal carbohydrate metabolism.

(c) If it be granted that the fatty acids of butter are products formed
synthetically from carbohydrate, the hypothesis presents insuperable diff-
culties. No combination of hexose and pentose molecules will produce
myristic acid (C,,H,,0,) by direct addition, yet this acid occurs commonly
in fats, e.g. lard, butter, cod-liver oil and many vegetable fats. The existence
of an intermediate tetrose sugar would have to be assumed as a normal

constituent.
3 hexose molecules give stearic acid.
2 pentose and one hexose molecules ,, palmitic acid.
2 pentose and one tetrose r. », Mmyristic acid.
2 hexose molecules », laurie acid.
2 pentose ,, », Caprice acid.
2 tetrase - », caprylic acid.

There is no evidence of the formation of a tetrose molecule in the
organism and it is exceedingly unlikely that a regular series of fatty acids
should be formed in this way.

(d) In the case of stearic acid the reduction of seventeen hydroxyl
groups would be assumed.

This hypothesis has therefore no evidence in its favour and involves
reactions which are not analogous with any of those known to us. It does
not therefore furnish us with a satisfactory explanation of the problem under
consideration.

The second hypothesis, which is perhaps the more generally accepted, is
that the fatty acids are built up by repeated condensations of a compound
containing two carbon atoms. This was first suggested by Nencki and
afterwards developed by Magnus-Levy [1902], Leathes and others who regarded
acetaldehyde as the substance from which by a series of aldol condensations

4

I. SMEDLEY AND E. LUBRZYNSKA 367

the straight chains containing even numbers of carbon atoms were formed.
The reactions involved would be represented by the following equations :
(1) CH,.CHO+CH,.CHO =CH,.CHOH.CH,.CHO
(2) CH,.CHOH.CH,.CHO=CH,.CH,.CH,.COOH+H,O
10:

If the aldol on the other hand were reduced to butyl aldehyde, it would
be available again to take part in a similar condensation :
CH,.CHOH.CH,.CHO+H, =CH,.CH,.CH,.CHO+H,0
CH, .CH,.CH,. CHO +CH,.CHO=CH,.CH,.CH,.CHOH.CH,.CHO.
a normal six-carbon-atom chain being thus produced.
In favour of this hypothesis, it may be urged:
(a) It does account for the production only of those fatty acids con-
taining even numbers of carbon atoms, since only multiples of two will exist.
(b) Hoppe-Seyler showed that by the action of caustic alkali on lactic
acid at from 200°-300°, acetic, butyric and caproic acids were formed.
Pasteur had previously shown that butyric and caproic acids were formed
by the bacterial fermentation of sugar. Acetic aldehyde may be obtained
from lactic acid and may therefore be a degradation product of sugar.
On the other hand it is open to the following criticisms:
(a) Lieben [1883, 1901] and his pupils have shown that when the
higher aldehydes condense with acetaldehyde under the influence of dilute
alkalies, the resulting aldehydes possess a branched and not an open-chain

structure:
CH; . CH,. CH, . CHO + CHO . CH, —» CH;.CH,.CH.CHO

|
CHOH
|

CH3.

It has since been shown that both aldol and crotonaldehyde will undergo
auto-condensation with the formation of a normal eight-carbon-atom chain
[Raper, 1907; Smedley, 1911]; but the difficulty of adding on acetic alde-
hyde to a higher aldehyde so as to build up chains increasing by the addition
of two carbon atoms has not been surmounted. One must therefore assume
that the condensation of aldehydes in the body does not take place in the
same manner as it does when brought about by the action of condensing
agents in the laboratory.

(6) No free aldehydes other than the sugars have been detected in the
body. If present in quantity they would probably be injurious to the life
of the cell. Parnas [1910] has shown that an enzyme is present in the liver
by which free aldehydes are at once removed.

368 I. SMEDLEY AND E. LUBRZYNSKA

(c). There is no biological evidence that acetaldehyde is formed as an
intermediate substance in the body metabolism.

The aldol condensation does not therefore furnish us with a satisfactory
analogy for the method by which the fatty acids are built up.

A survey of the general methods of producing fatty acids in the laboratory
shows that the most satisfactory method by which fatty acids may be built
up by increments of two carbon atoms is by means of Reformatski’s reaction
in which aldehydes are condensed with bromoacetic ester in the presence
of zinc;

OH
R.CHO+BrCH,. COOKt+ Zn + H.O=R.CHOH.CH,. COOHKt+ Zn,
r

As however neither zinc nor bromoacetic ester occurs in the body, this does
not furnish us with any helpful analogy for biochemical synthesis.

The Degradation-products of Carbohydrates: their suitability as Units
in the biochemical synthesis of Fatty Acids.

But little is known as to the manner in which carbohydrate breaks down
within the body. It has been repeatedly established that when a solution
of glucose in Ringer’s fluid is perfused through the isolated heart sugar
disappears [Locke and Rosenheim, 1907 ; MacLean apd Smedley, 1913]. This
is the only instance in which it has been established beyond the region of
controversy that sugar disappears when subjected to the action of an isolated
organ. But even here the decomposition products of the sugar molecules
are unknown. The controversy as to whether glycogen is a storage product
or a stage in the normal metabolism of sugar throws little light on the
problem under consideration. The discussion as to whether glucose is the
source of the lactic acid in the animal organism has more bearing on the
subject of fat formation. Embden has shown that the transfusion of blood
rich in sugar through a glycogen-free liver resulted in the abundant forma-
tion of lactic acid: blood poor in sugar similarly transfused gave rise to lactic
acid in inconsiderable amount. The formation of lactic acid from carbo-
hydrate is also indicated by the experiments of Mandel and Lusk on phlorizin
diabetes. It seems probable that both carbohydrate and protein may give
rise to Jactic acid in the body. The occurrence of lactic acid as a possible
cleavage product of carbohydrate suggests that the breaking down of sugar
takes place in such a way as to give rise to compounds containing three
carbon atoms. Lactic acid itself is not a very reactive substance nor does

—_

I. SMEDLEY AND E. LUBRZYNSKA 369

it appear a hopeful starting material for the synthesis of fatty acids. It is
however closely related to pyruvic acid.

CH;.CHOH .COOH CH;.CO.COOH
Lactic Acid Pyruvic Acid

There is also evidence that pyruvic acid itself is probably of considerable
importance in animal metabolism. It has been demonstrated that a close
connection exists in the organism between the a-amino- and the a-keto-acids.
Embden and Schmitz [1910, 1912] have shown that if a solution of ammonium
pyruvate be perfused through a liver, alanine is formed. Fellner [1912]
further showed that if a liver rich in glycogen be perfused with blood con-
taining ammonia, alanine is formed, and froma consideration of Embden’s
experiments pyruvic acid is indicated as the intermediate substance,
Neubauer and Knoop and Kertess [1911] have also suggested the formation
of alanine from pyruvic acid in the body.

Knoop [1910] and Knoop and Kertess [1911] have shown that if y-phenyl-
a-amino-butyric acid be fed to a dog, a considerable proportion of the acid
appears in the urine as the acetyl derivative; the same phenomenon was
observed by Neubauer and Warburg [1910] in their perfusion experiments.
There is some reason to believe that the acetylating agent may be pyruvic
acid, since de Jong [1900, 1904] showed that ammonium carbonate and
pyruvic acid react with formation of acetyl-alanine. It seems therefore
probable that pyruvic acid may be an intermediate substance formed in the
body from carbohydrate.

Pyruvic acid is a reactive substance, readily losing carbon dioxide under
the influence of oxidising agents and forming acetic acid. A study therefore
of its chemical properties and of its powers of condensation seemed of especial
interest.

The condensation of Pyruvic Acid with Fatty Aldehydes and the
oxidation of the products formed.

It had already been shown that if anhydrous hydrochloric acid be passed
into a mixture of benzaldehyde and pyruvic acid, cinnamoyl-formic acid results
[Erlenmeyer, 1901];

C.H;.CHO+CH;.CO.COOH=C,H;.CH:CH.CO.COOH+H;0.
Later both benzaldehyde and cinnamyl aldehyde were condensed with
pyruvic acid by adding a small amount of 10°/, caustic soda to the mixture
[ Erlenmeyer, 1903}.

In order to make use as far as possible only of reagents which may be
considered to bring about reactions somewhat similar to those brought about

370 I. SMEDLEY AND E. LUBRZYNSKA

by enzymes within the body, the condensation of the fatty aldehydes with
pyruvic acid was attempted in very dilute alkaline solution. The intermediate
unsaturated a-keto acid which was expected to result was not isolated, but
the product was at once oxidised by silver oxide in alkaline solution or by
hydrogen peroxide in neutral solution.

The behaviour of crotonaldehyde was first investigated.

EXPERIMENTAL.

Condensation of crotonaldehyde with pyruvic acid and oxidation of the
product formed.

5 grams pyruvic acid, 5 grams crotonaldehyde, 75 cc. n. NaOH and
I litre of water were added together and left for three days at the room
temperature, the solution being approximately 1/50 normal. The liquid
became deep yellow but no insoluble oil separated as in the condensation
of crotonaldehyde alone. The solution was neutralised by the addition of
125 cc. n. H,SO, and steam distilled to remove any free aldehyde.

Oxidation of Reaction-product.

Silver oxide was precipitated from 30 grams silver nitrate and added
to the solution of the condensation product of crotonaldehyde with pyruvic
acid after it had been steam distilled. 200 cc. of a 1/3 normal solution of
baryta were gradually added and the whole allowed to stand over night.
Next morning the silver oxide had been largely converted to silver. The
precipitate was filtered off, washed and concentrated under reduced pressure
until 50 ce. remained. In order to convert any hydroxy-acid that might
conceivably be present to the corresponding unsaturated acid, 10 grams of
baryta were added and the mixture boiled for 30 minutes.

Excess of sulphuric acid was then added and the whole steam-distilled.
1500 ce. of the steam distillate required 28°5 cc. normal potash for neutral-
isation. The neutral distillate was evaporated almost to dryness and to the
potassium salt so obtained, 10°/, H.SO, was added. Crystals separated which
melted at 132° after once recrystallising from dilute alcohol. The melting
point was unchanged on mixing with a specimen of sorbic acid prepared
by the condensation of crotonaldehyde and bromoacetic ester, and hydro-
lysis of the ester formed. The crystals were therefore satisfactorily identified
as sorbic acid.

In subsequent experiments the oxidation of the neutral condensation
product was carried out by means of hydrogen peroxide. An amount of
hydrogen peroxide exactly equivalent to the pyruvic acid originally taken

I. SMEDLEY AND E. LUBRZYNSKA 371

was used and the neutral mixture of condensation product and peroxide left
to stand over night at the ordinary temperature; the product was concen-
trated under reduced pressure and steam distilled as before, and the final
product consisted of a mixture of acetic and sorbic acids, the yield being
somewhat improved by this means. From 5 grams of crotonaldehyde 0°5 g.
sorbic acid was thus obtained.

The reaction must therefore have proceeded as follows:

CH;.CH:CH.CHO+CH;.CO.COOH =CH,.CH:CH.CH:CH.CO.COOH
CH;.CH:CH.CH:CH.CO.COOH+H,.0.,.=CH;.CH:CH.CH:CH.COOH+H,0+CO,
Sorbie Acid.

The condensation of Butyl Aldehyde with Pyruvic Acid.

10 grams of butyl aldehyde, 10 grams of pyruvic acid and 150 ce. normal
potash were shaken up with 2 litres of water and at the end of 12 days the
mixture was neutralised and concentrated under diminished pressure. To
the concentrated residue silver oxide from 43 grams of silver nitrate and
200 cc. n/3 baryta were added. After standing over night, the silver
precipitate was filtered off and the filtrate concentrated to 250 cc., 50 grams
of baryta added and boiled for 30 minutes. The whole was then acidified
with dilute sulphuric acid and distilled in steam. 1500 ce. of distillate were
neutralised by 59°2 cc. normal potash and evaporated to dryness. The
residue was acidified and extracted with ether. After evaporating off. the
ether, the residue was distilled under a pressure of 20 mm. About 3 grams
boiling from 130°-140° were obtained. The liquid rapidly decolourised
bromine water and gave on analysis the numbers required for the compound
C.H,,0;.

0°1409 g.; 0°3262 g. COs; 0°1134 g. HO.
C 63-149), H8-94%,.
Cale, for CgH 02 C 63-15 %/, H 8-77 %,.
In subsequent experiments, the product similarly prepared appeared to
consist of a mixture of octylenic acid (probably obtained by the self-con-
densation of the butyl aldehyde) and of hexylenic acid obtained from butyl
aldehyde and pyruvic acid. The difficulty of separating these in a small
quantity of a liquid mixture is considerable.

In another experiment where hydrogen peroxide was used as the
oxidising agent as described under crotonaldehyde, the product obtained
distilled under reduced pressure (15 to 20 mm.) from 120°-128° and gave on
analysis the following numbers.

0°1220 g.; 0°2858 g. COy: 0-1074 g. H.0.
C 63-85 °/,: H 9-75 %).

372 I. SMEDLEY AND E. LUBRZYNSKA

In investigating the condensation of iso-valeraldehyde and oenanthol with
pyruvic acid, chiefly the products of condensation of the aldehydes with
themselves were isolated. From the condensation product of iso-valer-
aldehyde and pyruvic acid, a small amount of the barium salt of an acid
was obtained, the percentage of barium in which agreed with that required
for the barium salt of the corresponding unsaturated keto-acid.

Condensation of these higher fatty aldehydes with pyruvic acid under
varying conditions is now being investigated. The condensation of croton
and butyl aldehydes with pyruvic acid and the oxidation of the product
formed with hydrogen peroxide in neutral solution furnishes a method by
which an unsaturated fatty acid may be built up containing two more carbon
atoms than the aldehyde from which it is derived. These condensations
have also been investigated under similar conditions in the aromatic series
[Smedley and Lubrzynska, 1913].

CONCLUSIONS.

The hypothesis now brought forward [Smedley, 1912] suggests that
pyruvic acid, formed in the body as a decomposition product of carbo-
hydrate, is the starting-point for the synthesis of the fatty acids. The stages
which are assumed to occur are represented by the following equations:

(1) CH3;.CO.COOH=CH;.CHO+CO,.
Pyruvic Acid>Acetaldehyde.
(2) CH;.CHO+CH;.CO.COOH = CH;.CHOH.CH,.CO.COOH
CH3;.CH:CH.CO.COOH + H,0.
Pentylenic a-keto-acid.
(3a) CH3.CH:CH.CO.COOH+0= CH;.CH:CH.COOH+CO,.

(4a) CH,.CH:CH.COOH+2H = CH;.CH,.CH,.COOH.
Butyric Acid.
(3b) CH;3;.CH:CH.CO.COOH = CH;.CH:CH.CHO+CO,.

Pentylenic a-keto acid.
(4b) CH3.CH:CH.CHO+CH;.CO.COOH=CH;.CH:CH.CHOH.CH,.CO.COOH.

and by reactions similar to 3a and 4a

CH; . CH,.CH,.CH,.CH,.COOH.
Caproic acid.

The evidence supporting this hypothesis may be briefly summarised as
follows.

1. Pyruvic acid is probably a degradation product of carbohydrate in
the body.

The perfusion experiments of Embden, Knoop and Neubauer show that
pyruvic acid is converted into alanine through the agency of the liver cells

teh er TT

_es =v 2

I. SMEDLEY AND E. LUBRZYNSKA 373

and that a close connection exists between the a-amino- and a-keto-acids.
Pyruvic acid may probably be an intermediate stage in the transformation
from glycogen to alanine (Kellner).

There is some reason to believe that in the acetylation of certain amino-
acids which has been observed both in perfusion and in feeding experiments,
pyruvic acid is the acetylating agent.

The close connection between alanine and pyruvic acid suggests that the
alanine group of the protem molecule may furnish an additional source of
the pyruvic acid available for the synthesis of fatty acids.

2. The decomposition of pyruvic acid into acetaldehyde and carbonic
acid, which constitutes the first stage of this process, has been shown by
Neuberg to be readily brought about by an enzyme present in yeast, termed
“carboxylase.”

3. The present hypothesis postulates that free acetaldehyde is not
liberated but that the decomposition of the keto-acid is in some way
regulated by the pyruvic acid with which the “nascent” aldehyde combines.

The condensation of fatty aldehydes with pyruvic acid has now been
shown to take place in the laboratory under the influence of dilute alkalis at
ordinary temperature.

4. Oxidation of the a-keto acid according to the equation

R.CO.COOH+0=R.COOH + CO,

may be brought about in the laboratory by hydrogen peroxide at the
ordinary temperature in neutral solution (p. 370).

5. The reduction of the unsaturated acid is the final stage; there is
abundant evidence that reduction can take place in the body although very
little is known as to the mechanism by which it is accomplished.

6. The a-keto-acid, synthesised as above, may be split into CO, and
aldehyde, and a further condensation with pyruvic acid may then be effected.
An acid with two more carbon atoms than the original aldehyde would thus
be synthesised.

As yet no a-keto-acids have been detected within the body: it may be
that they occur only within the cell and that reduction or oxidation always
accompanies their liberation. The above hypothesis accounts for the forma-
tion of a series of straight chain acids beginning with four carbon atoms and
increasing by increments of two carbon atoms: it involves only reactions
which are analogous with those which are known to occur in the laboratory
and there is reasonable evidence for believing that the starting material,
pyruvic acid, can be formed from carbohydrate in the body.

re)
Vt

Bioch, vu

374 I. SMEDLEY AND E. LUBRZYNSKA

REFERENCES.

Amberger (1907), Zeitsch. Nahr. Genussm. 13, 516.

Dakin (1908), J. Biol. Chem. 4, 63, 419; 5, 173.

(1909), J. Biol. Chem. 5, 303, 409; 6, 203, 221, 235.
Embden and Schmitz (1910), Biochem. Zeitsch, 29, 423.
(1912), Biochem. Zeitsch. 38, 392.

Erlenmeyer (1901), Ber. 34, 1817.

(1903), Ber. 36, 2527.

Fellner (1912), Biochem. Zeitsch. 38, 414.

De Jong (1900), Rec. Trav. Chim, 19, 259.

—— (1904), Rec. Trav. Chim, 23, 131.

Knoop (1904), ‘‘Der Abbau aromatischer Fettsiuren im Tierkérper.”
-—— (1910), Zeitsch. physiol. Chem. 67, 489.

Knoop and Kertess (1911), Zeitsch. physiol. Chem. 71, 252.
Lieben (1883), Monatsh. 4, 10.

(1901), Monatsh. 22, 289.

Locke and Rosenheim (1907), J. Physiol. 36, 205.

MacLean and Smedley (1913), J. Physiol. 45, 462.
Magnus-Levy (1902), Beitrdge, 2, 261.

Mandel and Lusk (1906), Amer. J. Physiol. 16, 129.
Neubauer and Warburg (1910), Zeitsch. physiol. Chem. 70, 1.
Neuberg and Hildesheimer (1911), Biochem. Zeitsch. 31, 170.
Parnas (1910), Biochem. Zeitsch. 28, 274.

Raper (1907), J. Chem. Soc. 91, 1831.

Siegfeld (1907), Zeitsch. Nahr. Genussm. 18, 616.

Smedley (1911), J. Chem. Soc. 99, 1627.

(1912), J. Physiol. 45, Proc. xxv.

Lubrzynska and Smedley (1913), Biochem. J. 7, 375.
Swaving (1906), Zeitsch. Nahr. Genussm. 9, 505.

XXXVI. THE CONDENSATION OF AROMATIC
ALDEHYDES WITH PYRUVIC ACID.

By EVA LUBRZYNSKA anp IDA SMEDLEY (Beit Memorial
Research Fellow).

From the Biochemical Department, Lister Institute.

(Received June 18th, 1913.)

In a preliminary communication published by one of us [Smedley, 1912],
and in the foregoing paper [Smedley and Lubrzynska, 1913], the condensation
of certain fatty aldehydes with pyruvic acid in very dilute alkaline solution
has been described.

In order to obtain a better knowledge of the reaction and hence possibly
to overcome some difficulties with which we found ourselves confronted in
our experiments in the aliphatic series we have investigated the behaviour of
certain aromatic aldehydes under similar conditions. A number of the
By-unsaturated a-keto-acids have been prepared and converted by oxidation
with hydrogen peroxide into the corresponding unsaturated acids containing
one carbon atom less than the original keto-acid.

Claisen and Claparéde [1882] studied the action of anhydrous hydrochloric
acid on a mixture of benzaldehyde and pyruvic acid and obtained from the
mixture cinnamoyl formic acid. This reaction was further studied by Erlen-
meyer, junr. [1899, 1901], who showed that in addition to the above product
y-phenyl-8-benzylidene-a-ketobutyrolactone is formed :

CH; .CHO+ CH; .CO.COOH= C,H; .CH :CH.CO.COOH+ H,0.
2C4H; . CHO + CH,.CO. COOH =, H;.CH : C.CO.CO.0.CH. C,H; +2H,0.
nd

——

The condensations of benzaldehyde, cinnamic aldehyde, piperonal and
anisic aldehyde with pyruvic acid have been studied in dilute alkaline
solution. The reactions proceed in all cases readily at the laboratory
temperature. The amount of potash used was generally such that the
strength of the solution was from 1/40 to 1/10 normal, and the mixture was
allowed to stand at the temperature of the laboratory for periods varying
from 2 to 7 days. As the reaction proceeds the aldehyde disappears and the

25—2

—

376 E. LUBRZYNSKA AND I. SMEDLEY

liquid gradually becomes yellow. When the reaction is completed, the keto-
acid is thrown down on acidifying the liquid, and after purification, is
oxidised to the corresponding @8-unsaturated acid.

Yellow crystalline products are formed by the condensation of anisic
aldehyde and piperonal respectively with pyruvic acid, the formulae of which
are shown on analysis to agree with those required for the -unsaturated
a-keto-acids :

R.CHO+CH,.CO.COOH=R.CH: CH.CO.COOH+H,0.
R.CH : CH.CO.COOH+H,0,=R.CH : CH. COOH +CO,+H,0.

After we had completed the above series of experiments, we found
that we had overlooked two papers by Erlenmeyer, junr. [1903, 1904], in
which he had described the condensation of benzaldehyde and cinnamic
aldehyde respectively with pyruvic acid in the presence of strong caustic soda
(10 °/,). The products which we obtained on condensing cinnamic aldehyde
with pyruvic acid were similar to those described by Erlenmeyer. He
isolated two forms, an orange-red, melting at 75° and gradually converted in
vacuo into a yellow form melting at 107°. Our products melted at 73° and
104° (uncorrected) respectively.

From anisic aldehyde and pyruvic acid yellow crystals melting at 130°
were obtained. Piperonal under similar conditions gives beautiful yellow
needles which at a temperature of about 70° undergo a remarkable transition
into a deep orange-red form and finally melt at 163°. On keeping the red
form so obtained in an evacuated desiccator or in a well-stoppered bottle it
was still unchanged after a month had elapsed. If however it was exposed
to the air it was gradually reconverted into the yellow variety.

On oxidation with the calculated quantity of hydrogen peroxide in neutral
solution at the ordinary temperature the keto-acids split off carbon dioxide
and form unsaturated acids:

R.CH : CH.CO.COOH+H,0,=R.CH : CH.COOH +CO,+ H,0.

Silver oxide which was successfully used to oxidise the corresponding
keto-acids in the fatty series [Smedley, 1912], was found to be without action
in the aromatic series.

EXPERIMENTAL.

Condensation of piperonal with pyruvic acid. 4g. of pyruvic acid and
6 g. of piperonal were shaken up with 1 litre of water and 140 cc. N. KOH
added; the mixture was allowed to stand at the ordinary temperature for
a week, during which it became yellow. The small amount of unchanged

E. LUBRZYNSKA AND I. SMEDLEY 377

piperonal was filtered off and the solution neutralised with 36 ec. N. H,SO,.
The neutral solution was extracted with ether several times to free it from
traces of unchanged aldehyde. Air was drawn through to remove the last
traces of ether and the solution acidified. A yellow solid was precipitated
and allowed to settle: it was then filtered off and well washed with cold water.
After several recrystallisations from dilute alcohol it was obtained in
beautiful yellow needles which turned to deep orange-red at about 70° and
melted at 163°.

01062 g.; 0:2336 g. CO, ; 0:0368 g. H,O.

C 59-98 9/3 H 385 >.

Cale. for CHs0;, C 60-00 °/,; H 3°63 °/,.

Condensation of anisic aldehyde with pyruvic acid. 5 g. anisic aldehyde
and 3°5 g. pyruvic acid were shaken up with 1 litre of water and 70 ce. N.
KOH added, the mixture being allowed to stand for 4 days during which a
yellow colour developed. It was then neutralised with 25:05 ec. N. H,SO,
and the mixture treated exactly as in the case of the condensation of
piperonal with pyruvic acid.

The yellow needles finally obtained melted at 130° and no change in
colour was observed either on standing or on heating.

0°1100 g.; 0°2585 g. COg; 0:0500 g. H.O.
C 64-09 °/,; H 5:00 %/,.
Cale. for C,,H 904, C 64:07 °/,; H 4°85 °/,.
The specimen, part of which gave the above results on analysis, was
allowed to remain in a closed test-tube for two weeks and then analysed :
0-0972 g. ; 0-2330 g. CO,; 0-0407 g. H,0.
C 65-32 /,; H 4-63 °/,.
A week later it was again analysed :

0°1208 g.; 0°2927 g. CO,; 0°0562 g. H,0.
C 65-98 /,; H 5°16 %/,.

The acid was then reerystallised and again analysed :

0°1244 g.; 0-2972 g. CO,; 0-0564 g. H,0.

C 65-11 /,; H 5-06 °/,.
The preparation was repeated and the freshly prepared product analysed :

0-1169 g.; 0:2743 g. CO,; 0°0518 g. H,0.

C 6410 °/,; H 4-91 9/,.
The acid therefore on standing undergoes a slight decomposition, the

nature of which was not ascertained.

Condensation of cinnamic aldehyde with pyruvic acid. 5 g. cinnamic
aldehyde and 4 g. pyruvic acid were shaken up with 1 litre of water and
80 cc. of normal potash. The mixture was allowed to stand for 4 days at the

378 E. LUBRZYNSKA AND I. SMEDLEY

ordinary temperature, neutralised and treated as in the experiments already
described with piperonal and anisic aldehyde.

The condensation product was obtained in two forms; red crystals melting
at 73°, which pass gradually into a yellow substance melting at 104°. In one
experiment red crystals melting at 73° were exposed in an evacuated desic-
cator for 48 hours: they had then become yellow and melted at 95°. After
another 24 hours the melting point was 101° and after recrystallisation from
dilute alcohol it rose to 103°. Analysis showed that the composition of the
two forms after drying in a vacuum desiccator was identical.

01274 g.; 0°3344 g. CO.; 0:0600 g. H,0.

C 71°58 °/,; H5-17%,.

Cale. for CyHyo03, C 71°28 °/,; H 4:95 "/,.

Oaidation of diovymethylene-benzylidene-pyruvic acid :
CH,0,CgH3CH : CH. CO. COOH. |

The keto-acid was dissolved in alcohol, neutralised with potash and the
theoretical quantity of hydrogen peroxide then added. It is important that
no excess of the peroxide be present or the yield will be found to be
diminished. The oxidation product was heated for an hour on a water-bath
to complete the reaction and then acidified. A very pale yellow acid was

precipitated, which after one recrystallisation was almost white and melted
at 242°.
0:1074 g.; 02462 g. CO,; 0°0423 g. H,0.
C 62-47 J,; H 4:37 °).
0-1148 g.; 0:2690 g. COz; 0:0458 g. HO.
C 62°51 °/); H 4:32 %.
Cale. for CjgHs04, C 62°50 °/,; H 4:16 %/,.

Oxidation of cinnamylidene-pyruvic acid :
C,H, .CH: CH .CH :CH.CO. COOH.
The oxidation was carried out in a manner similar to that just described for
the piperonal product. The acid obtained after recrystallisation from benzene
and from alcohol melted at 165° and was therefore identical with the cinna-
mylidene-acetic acid prepared by Perkin from malonic acid and cinnamic

aldehyde, melting at 165°.

0:0844 g.; 0°2335 g. CO,; 0:0442 g. H,O.
in C 75:47 °/); H 5:80 %J,.
After again recrystallising :
0-1093 g.; 0:3052 g. COy; 0-0578 g. H,0.
C 76°12 %/,; H 5-85 °/,.
Cale. for C};Hy)02, C 75-86 %/,; H 5:74 %/,.

Oxidation of methoay-benzylidene-pyruvic acid :

CH,0O . CoH. CH : CH . CO. COOH.

E. LUBRZYNSKA AND I. SMEDLEY 379

This acid was dissolved in alcohol, carefully neutralised with potash and
hydrogen peroxide added. In this case it was found that the best yield was
obtained on adding 1°5 times the theoretical quantity of the peroxide. After
several recrystallisations, the acid melted at 172°. Analysis gave the
following result :
01020 g.; 0°2517 g. CO.; 0°0530 g. HO.
C 67°23 Jy; H 5°70 Jp.
Cale. for CyopH 903, 67°41 %/o; H 5°62 %.

Condensation of benzaldehyde with pyruvic acid. The condensation was
carried out as in the cases already described, but the intermediate keto-acid
was not isolated. The product was at once oxidised in neutral solution with
hydrogen peroxide.

It was found in this case of the greatest importance to have the solution
exactly neutral and to avoid any excess of the peroxide.

On acidification, cinnamic acid was precipitated. It was identified by
taking the melting-point of a specimen mixed with pure Kahlbaum’s
cinnamic acid.

The yields in the above experiments varied from 50-70 °/, of the
theoretical.

REFERENCES.

Claisen and Claparéde (1832), Ber. 14, 2472.
Erlenmeyer, junr. (1899), Ber. 32, 1450.

—— (1901), Ber. 34, 817.

—— (1903), Ber. 36, 2527.

—— (1904), Ber. 37, 1318.

Smedley (1912), J. Physiol. 45, Proc. xxv.

and Lubrzynska (1913), Biochem. J. 7, 364.

XXXIX. THE PRECIPITATION OF EGG-ALBUMIN
BY AMMONIUM SULPHATE. ACONTRIBUTION
TO THE THEORY OF THE “SALTING-OU rE]
OF PROTEINS.

By HARRIETTE CHICK anp CHARLES JAMES MARTIN.
From the Lister Institute.

(Received June 11th, 1913.)

INTRODUCTION.

The first systematic and quantitative researches upon the precipitation of
proteins by salts in large amounts were made by Hofmeister and his pupils.

In the experiments of Kauder [1886] and Lewith [1888] with serum
proteins, and Hofmeister [1888] with egg-protein, comparison was made of
the precipitating power of a large series of electrolytes, as a result of which
the latter were arranged in what is known as the Hofmeister series. Most of
the experiments were made with sodium salts, among which sulphate,
phosphate, acetate, citrate, tartrate, chromate, chloride, nitrate and chlorate
formed a series arranged in descending power of precipitation ; of the kations
examined lithium was the most effective, and sodium, potassium, ammonium
and magnesium came afterwards in order of decreasing efficiency.

Hofmeister [1889] came to the conclusion that the precipitation was
caused by the electrolytes depriving the protein of the amount of water
necessary to keep it in solution, and was confirmed in this view by the
results of some experiments showing the influence of various salts in modi-
fying the imbibition of water by gelatin [1891]; work in this direction was
later extended by Pauli [1898] and Pauli and Rona [1902].

Spiro [1904] demonstrated that the precipitation from their solutions of
caseinogen and gelatin by sodium sulphate was analogous to the “salting out”
of alcohol recently studied in detail by de Bruyn [1900]. In neither case is
the phenomenon one of simple precipitation, since, owing to the appropriation
of water by the salt, separation into two phases occurs. Hach phase con-
tains all the constituents of the system and any alteration in one of

Pte...

H. CHICK AND C. J. MARTIN 381

the three constituents leads to readjustment of the composition and
relative volumes of the two phases. Spiro also pointed out that, since in the
case of alcohol the effect of electrolytes is not attributable to the consti-
tuent ions, any influence of the latter in the salting out of proteins must be
regarded as a subsidiary phenomenon.

Spiro’s conception explains to some extent the divergent results obtained
in the precipitation of proteins by the addition of neutral salts, when the
whole conditions are not maintained constant.

The series of observations which we are about to record concern the
“salting out” by ammonium sulphate of pure recrystallised egg-albumin.
Our observations show that in this case also we have to deal with the
separation of the original system (itself not homogeneous) into two distinct
phases, and that the influence upon the volume of these phases of concentra-
tion of protein, salt and water in the system is, as Spiro found for caseinogen
and gelatin, analogous to what occurs in alcohol, salt and water mixtures. In
addition, however, we find that the charge carried by the protein particles is
an important factor in the final equilibrium.

The results of the experiments will first be set forth and the proposed
explanation discussed later.

PRECIPITATION OF PURE EGG-ALBUMIN BY AMMONIUM SULPHATE.

Material. The material employed was egg-albumin crystallised from egg-
white in presence of ammonium sulphate according to the method of Hopkins
and Pinkus [1898]. The albumin was recrystallised once or twice, separated
from the mother liquor by pressing between filter paper, and finally dissolved
in distilled water. A concentrated stock solution was thus obtained, the
composition of which, as regards (1) protein, (2) ammonium sulphate,
(3) water was accurately ascertained by analysis, and which, when diluted to
a suitable degree, served for most of the following experiments.

Since the salt employed for “salting out” in these experiments was also
ammonium sulphate, the small concentration of the latter always present in
the original albumin solution presented no complication ; an allowance was
made for this amount in the calculations. In those cases where an electrolyte-
free solution was required, the albumin solution was previously dialysed.

Egg-albumin prepared in this way we believe to be as homogeneous
a protein as it is possible to obtain. Hopkins [1899-1900] came to the
conclusion that egg-albumin, crystallised from faintly acid ammonium
sulphate solutions by the above method, was a pure substance. The

382 H. CHICK AND C. J. MARTIN

rotatory power remained absolutely constant after repeated recrystallisa-
tions (p. 312) and the proportion of carbon, hydrogen, nitrogen and sulphur,
as well as the ash, remained constant. His experiments were made with four
different samples after three or four recrystallisations.

1. Influence of concentration of salt upon the amount of
protein precipitated.

Mellanby [1907] made a quantitative study of the influence of concentra-
tion of ammonium sulphate on the precipitation of the proteins from horse
serum, but, as far as we are aware, no experiments have as yet been made with
a pure protein. In the present instance two sets of experiments were made, in
both of which the concentration of protein was about 1 °/,. In the first the
concentration of protein in the whole system was left constant = 1:11 °/, by
weight, and the ratio salt to water was varied (Table I and Fig. 1). In the
second set the ratio protein to water was kept constant and the precipitation
studied by varying the amount of salt present; the concentration of protein
in the whole system varied from 1:0 °/, to 0°93°/, (by weight). (Table II and
Fig. 2.)

TABLET:
Precipitation of pure egg-albumin with ammonium sulphate ;

influence of concentration of salt.

Temperature, 20°.
Protein constant=1-11 °/) by weight of total system.
Ratio salt/water varying.

G. albumin G. salt G. albumin

Albumin Water Salt in 100 g. in 100 g. in 100g.
g. g. g. total svstem total system filtrate
1:00 69-00 20:00 111 22-22 1-089
1:00 68-00 21-00 a 23°33 0-711
1-00 67:00 22-00 a3 24°44 0°302
1:00 66°00 23°00 5 25°55 0-104

1-00 65°00 24-00 5s 26°66 0:0315
0-50 32°35 12°15 Bp 27-00 trace
0°50 32-00 12°50 by. 27°77 trace

The method of experiment was as follows. Mixtures were prepared by
weighing into stoppered bottles the required amount of water and protein,
and the necessary amount of ammonium sulphate was then added in large
crystals and gently shaken. This prevented over-saturation with ammonium
sulphate in the neighbourhood of the crystals, the large size of which

H. CHICK AND C. J. MARTIN 383

prevented a too rapid solution. The bottles were placed in a thermostat at
20°, for from 1 to 2 hours, the contents filtered and the protein estimated in
the filtrate by weighing the coagulum formed on heating.

G. albumin in 100g. filtrate.

22 23 24 25 26 27 28 39

G. (NH,4),SO, in 100 g. total system.

Fig. 1. Influence of concentration of the salt upon precipitation of egg-albumin by (NH4).SO,
at 20°, see Table 1; protein constant=1-11 °/, (by weight) of total system, ratio salt to
water varying.

TABLE IL.

Precipitation of pure egg-albumin with ammonium sulphate ;
influence of concentration of salt.

Temperature, 20°.
Ratio protein/water constant =1-3/100.

G. salt to G. albumin
Albumin Water Salt 1:3 g. protein in 100 g.
g. g. g. and 100g. H,0 filtrate
1:30 100 29 29 0-998
1:3 100 29°7 29°7 no precipitation.
1:30 100 30 30 0°938
1°95 150 46°5 31 0°733
if33 100 32 32 0-487
2°6 200 66 33 0:273
2°6 200 67 33°5 0°232
2°6 200 70-01 35°0 0-105

2-6 200 76 38-0 0-022

G. albumin in 100 g, filtrate.

384 H. CHICK AND C. J. MARTIN

29 30 31 32 33 34 35 36 37 38
G. (NH,),SO, to 13g. protein and 100 g. H,O.

Fig. 2. Influence of concentration of the salt upon the precipitation of egg-albumin by
(NH4)2SO,4 at 20°, see Table II; ratio protein to water constant=1-3 to 100, amount of
salt present varying.

The curves in Figs. 1 and 2 are both of the same general form, in both
cases the percentages of albumin in the filtrate are plotted as ordinates and
the abscissae are respectively percentage by weight of ammonium sulphate in
the whole system and grams of ammonium sulphate present, which explains
the fact that the curve is steeper in the former instance. The point of
commencing precipitation which is very sharply marked is at about the same
concentration of salt in both cases, viz. 22°2 and 22:7 °/, ammonium sulphate
respectively ; the curve then descends steeply and approaches the base line
asymptotically.

2. Influence of concentration of protein.

Kauder [1886] showed that serum albumin was more readily precipitated
by ammonium sulphate if in more concentrated solution and determined the
diminishing concentration of ammonium sulphate necessary to cause com-
mencing precipitation in a series of solutions of increasing protein concentra-
tion. Hofmeister in 1888 published the results of similar experiments, using
egg-white and potassium acetate and ammonium sulphate. Similar evidence
has since been brought forward by other workers, e.g. Mellanby [1907], but
in no case was a pure protein employed.

We have made an experiment with pure egg-albumin, estimating the

H. CHICK AND C. J. MARTIN 385

protein precipitated by a constant concentration of ammonium sulphate (ratio
salt to water constant) when the amount of protein was varied. The results
are given in Table III and graphically set forth in Fig. 3, where the propor-
tion of the protein separated is plotted as ordinate against the concentration
(percentage by weight of whole system) of protein in the original mixture as
abscissa. Not only is more protein separated from the more concentrated

TABLE III.

Precipitation of pure egg-albumin with ammonium sulphate ;
influence of concentration of protein.

Ratio salt/water constant =31/100.
Concentration of protein varying.

G.albu- G.albumin G. G. albumin
min in to 31g. albumin precipitated Protein
Albumin Water Salt 100g.total salt and inl00g. from 100g. ppted.
g. g. g. mixture 100g.H,O filtrate total mixture 90

1:90 56°92 17°65 2-481 Bret 1:130 1°351 54:4
1-90 36°92 11-44 3°775 5:14 1°115 2-660 70°4
3°79 50°85 15°76 5°383 7°45 1159 4-224 78°5
7:59 53°69 16°64 9°738 14°13 0-935 8-803 90-4
4°74 17°31 5°36 17-306 27°4 0:72 167534 95°5

#

= 100

=

is) elo)

Be

SE 80

% 2

2 70

S

3

5 60

—

60 5 E 10 15 20

G. albumin in 100 g. original mixture.

Fig. 3. Influence of concentration of protein upon the precipitation of egg-albumin by
(NHy4).SO, at 20° C., see Table III; ratio salt to water constant=31 to 100, amount of
protein varying.

solution at a given concentration of salt, but a greater proportion is
precipitated (see last column, Table III). The concentration of protein
in the filtrate is not constant, but varies from 1°13 °/, to 0°77 °/, (col. 6,
Table III), as the initial protein concentration is varied from 2°5 to 17-3 °/,
(col. 4). This suggests that the precipitation is a phase separation, analogous
to the case of “salting out” of alcohol with ammonium sulphate. As will be

386 H. CHICK AND C. J. MARTIN

seen later, when the results are given of determinations of the protein-salt-
water content of filtrate and precipitate respectively, this was proved to be
the case.

3. Influence of hydrogen ion concentration.

It is common experience, e.g. with serum, that the addition of a little
acid enhances the amount of protein precipitated by the same concentration
of ammonium sulphate, and that proteins not precipitated by saturation with
sodium chloride are thrown down on acidification of the solution.

Mellanby [1907] called attention to the increased amount of precipitation
of horse serum by neutral salts after addition of various acids, and gave some
quantitative data, using sodium chloride. In the present investigation the
influence of acidity was directly measured in a series of mixtures in which the
concentration of protein and ammonium sulphate was maintained constant,
and so chosen that precipitation had just begun in the control solution. The
reaction of the solution was adjusted to various degrees of hydrogen ion
concentration by the addition of small quantities of standard sulphuric
acid.

Determinations of hydrogen ion concentration were made with the type
of concentration cell described by Michaelis and Rona [1909], the contact
fluid between the two cells being saturated potassium chloride solution.
It is possible that the determinations of H+ concentration are not very
accurate owing to the high concentration of ammonium sulphate present ;
those in any one series are, however, perfectly comparable.

TABLE IV.

Precipitation of pure egg-albumin with ammonium sulphate ;
influence of hydrogen won concentration.
Experiment I at 18°.

G. ammonium sulphate in 100 ce. original mixture=30°4.
G. protein in 100 ce. original mixture =0°575.

G. protein present Hydrogen ion concentration
in 100 ce. in filtrate, in
filtrate terms of normality
0-572 10-527 ( 54 x 10-7)
0°391 WOE (ail se 5.)
0:097 NO=23* (PBL Se! 5.)
0-062 NOR (CRISS 5, })
0:048 Dee (RSS 55)
0:035 IOs" (AROS 35)

H. CHICK AND C. J. MARTIN 387

G. albumin in 100 cc. filtrate.

als) =5 -4
Hydrogen ion concentration, exponents.

Fig. 4. Influence of hydrogen ion concentration upon the precipitation of egg-albumin by
(NHy)2SO,4 at 18° (see Table IV).

Concentration of (NH4)2SO, in whole system, constant =30-4 grams per 100 cc.
” ” Protein ” ” ” =0°575 » >

The concentration of protein varied from 06 to 0°9°/, and that of
ammonium sulphate from 304 to 28°6°/, (by volume) in the three different
series of experiments set forth in Tables IV to VI. The mixtures were placed
for two hours at 18° in order that equilibrium might be attained; they were
then filtered and the concentration of protein and of hydrogen ions in the
filtrate was determined.

The influence of hydrogen ion concentration was found to be very
marked although the range through which it operates is not extensive.
In one case (Exp. I, Table IV, protein concentration = 0°57 °/,, ammonium
sulphate = 30°4°/,) an increase in acidity from 54 x 10~ normal (control) to
780 x 10-7 normal was enough to cause precipitation of nearly all the protein.
In Exp. II, Table V (0°86 °/, protein and 30:03 °/, ammonium sulphate) and
Exp. III, Table VI (0:9 °/, protein and 28°6 °/, ammonium sulphate) the
effective range of hydrogen ion concentration was from 2°9 to 360 x 10~
normal, and from 12 to 203 x 10~7 normal respectively.

388 H. CHICK AND C. J. MARTIN

TABLE V.

Precipitation of pure egg-albumin with ammonium sulphate ;
influence of hydrogen ton concentration.
Experiment II at 18°.

G. ammonium sulphate in 100 g. original mixture=30°03.
G. protein in 100 g. original mixture=0°856.

No. of ce. No. of ce.
N/10 H,SO, N/10 NH,;OH
(or equivalent) (or equivalent) G. protein Hydrogen ion concentra-
added in total added in total in 100 ce. tion (filtrate), in

volume of 27 cc. volume of 27 ce. filtrate terms of normality
— 1:0 0:864 LOpE 47952 °88<1 0e2)
— 0°5 0-809 HO moot T( GOS 2s a)
— 0:3 O95 tame TORE (iGO) Se og J)
— _ 0-514 10 me 039-8 nex )
0°5 —- 0-056 10; 4404 Soe; )
1:0 — 0:01 (about) j= SOEs} SS 55.

The last trace of protein present, however, does not appear to be precipi-
tated by alteration of reaction alone, a slight trace in Exp. III being still left
in solution at a hydrogen ion concentration of about 1/100 normal or
73000 x 1077 normal.

The range of reaction where hydrogen ion concentration has its great
effect is presumably just on the acid side of the iso-electric point (see below
p- 392). From Exp. III it is seen that for ammonium sulphate change in
hydrogen ion concentration at and on the alkaline side of the neutral point is
without much influence (viz. from Ht =12 x 10~ normal to 0°008 x 10
normal), see Table VI.

ee ee

~~

TABLE VI.
Precipitation of pure egg-albumin with ammonium sulphate ;
influence of reaction, hydrogen von concentration.

Experiment III at 18°.
G. (NH,4)280, in 100 g. original mixture=28°6.

G. protein A, re “ =0-910. 24

Ce. N/10H,S0, Ce. N/10 (NH4)OH ) ,
(or equivalent) (or equivalent) G. protein Hydrogen ion concentra- a
added in total added in total in 100 ce. tion (filtrate) in §
volume of 25:4 cc. volume of 25-4 cc. filtrate terms of normality 2 |
— 3°6 No precipitation 10~*-"1 (0-008 x 10-7) 7

— 2-0 0-899 1077-3 (0-59 ~ 95 t

- 1:0 0-908 10-656 (2°7 fat) 4

= = 0-899 10-68 (12-0, ) 5

0°25 — 0°533 Ome 2235 ae) .:

0°5 — 0-174 TWO-CAR 55), ‘

1-0 — 0-063 10 4-59)(203 eee) . |

5:0 — trace 10-*-4 (35000 ,, ) ‘

10-0 == slight trace 10-2-14 (73000.,, )

H. CHICK AND C. J. MARTIN 389

G. albumin in 100 cee. filtrate.
a

=o —S =F —6 =) —4 18) =

Hydrogen ion concentration, exponents.
Fig. 5. Influence of hydrogen ion concentration upon the precipitation of egg-albumin by

(NHy)2SO, at 18° (see Table VI).

Concentration of (NH4)2SO, in whole system, constant=28°6 grams in 100 ec.
” ’” protein ” ” ” aes 0-910 ” ”
H* concentration expressed as exponents.

4. Influence of temperature.

Lewith [1888] in case of ox-serum proteins, showed that rise in tempera-
ture assisted the precipitation by ammonium sulphate and Hofmeister [1888]
made the same observation with egg-white and various salts. Spiro [1904]
stated the same to be true of crystalline serum albumin and ammonium
sulphate.

We have confirmed the above observations for serum proteins! and pure
egg-albumin if the reaction be alkaline, but in faintly acid solution (10°
normal) we have found the reverse to be true above 9°.

In Table VII and Fig. 6 are set forth the results of an experiment with
0°85 °/, protein and 28 °/, ammonium sulphate. <A series of exactly similar
solutions were placed for from 1 to 2 hours in a thermostat at the required
temperature, after which they were rapidly filtered and the protein estimated
in the filtrate. From 0° to 9° the temperature coefficient of precipitation
was positive, above 9° it remained negative to 50°% This can be readily

1 Mellanby [1907, p. 294], on the other hand, states that the temperature coeflicient of serum-
protein precipitation with (NH,),SO, is negative between the temperatures of 0° and 40° C. and
so small as to be unimportant.

* No denaturation occurred.

Bioch. vir 26

390 H. CHICK AND C. J. MARTIN

demonstrated if ammonium sulphate be added to an egg-albumin solution at
9° to the point of opalescence, and just short of the formation of a definite
precipitate. If, then, the mixture be divided into three portions, of which
one is placed at 0° and a second at 20°, a definite precipitation will occur at
both temperatures, whereas the portion maintained at 9° will remain merely
opalescent.

TABLE VII.

Influence of temperature upon the precipitation of pure crystalline
egg-albumin with ammonium sulphate (28 °/,).

Protein content =0°85 9/9.
Hydrogen ion concentration about 10~° normal.

Concentration of

Temperature protein in filtrate, °/,
50 0-828
36 - 0°703
20 0°364
9 0:268
0 0:318

G. albumin in 100 ce. filtrate.

10) 10 20 30 40 50

Temperature, degrees centigrade.

Fig. 6. Influence of temperature upon the precipitation of egg-albumin by (NHy4)2SOq, see
Table VII.

Concentration of protein, salt, water and hydrogen ions in whole system constant.
INTERPRETATION OF RESULTS.

From the experiments detailed above, it is clear that the amount of
protein precipitated from solution (volume of the protein-rich phase) is

H. CHICK AND C. J. MARTIN 391

dependent upon the amount of salt present, concentration of protein, concen-
tration of hydrogen ions and temperature.

Spiro [1904] showed that, in the “salting out” of caseinogen and gelatin by
sodium sulphate, the ratio ofsalt to water in the more protein-rich phase was
less than in the watery phase

a similar relation to that found by de Bruyn
[1900] in alcohol separation.

The same is true in the case of egg-albumin. An albumin solution was
precipitated by ammonium sulphate and a complete analysis made of the
original solution, the precipitate and the filtrate. In order to free the
precipitate from adherent mother-liquor, it was not simply dried between
filter paper, as done by Spiro, but placed between filter paper, surrounded
by kieselguhr and submitted to a pressure of about 3 tons to the square inch.
This pressure was, however, more than enough to squeeze out adherent
mother-liquor, and actually removed some of the imbibed water of the
protein-rich phase, for the concentration of salt in the liquid expressed, in
the one case where it was collected and analysed, was only about half of that
in the watery phase. The analyses given in Table VIII, where the results of
these experiments are set out in detail, do not therefore express the com-
position of the two phases which were in equilibrium. ‘The experiments prove,
however, that the protein had appropriated some of the water for, notwith-
standing this squeezing out of weak salt solution, the ratio salt to water
in the compressed cake was considerably less than in the watery phase.
See Table VIII.

This means that egg-albumin, like caseinogen and gelatin, dissolves or
imbibes water just as alcohol does, and a mixture of protein, salt, water can
be made to separate into two phases, either by the further addition of salt or
of protein in the same way as a mixture of alcohol, salt, water separates on the
addition of either salt or alcohol. In either case it amounts to increasing the
concentration of salt in the watery phase.

To explain the influence of hydrogen ion concentration in the precipita-
tion of egg-albumin, we must suppose that the electrical condition of the
colloidal particles is a factor which modifies the ease with which they
aggregate under the influence of the salt present. Hofmeister [1889] was
influenced by a similar idea when he made parallel experiments with
colloidal ferric hydroxide and egg-white, and Posternak [1901, 1 and 2], when
investigating the precipitation of vegetable globulins, found that the
efficiency of various electrolytes was influenced by the reaction of the
suspension and presumably by the charge originally carried by the protein
particles. .

26—2

392 H. CHICK AND. C. J. MARTIN

TABLE VIIL

Precipitation of pure egg-albumin by ammonium sulphate ;
composition of filtrate, pressed precipitate and press- liquor.

Composition °/, In original In pressed In In press
Exp. by weight of mixture precipitate filtrate liquor
I Egg-albumin AGE 9°56 64°65 0-84 -
(NH,),SO;... oe 23°35 29-83 26-67 —
H,O af Ses 67:08 6°39 72°49 —
Salt present in 100 parts 25°82 17°6 26°9 =
salt and water
Il Egg-albumin ie 12-48 63°69 0°59 —-
(NH,),SO, ... fat 22-60 29-91 27-04 =
HO BSS ie 64:92 8°47 72°37 “=
Salt present in 100 parts 25°82 22-1 27-2 a
salt and water
Il Egg-albumin Pec 9°47 73:57 1°58 0
(NED) SO, = sie 24-09 22°04 27°66 15:2
H,O es ee 66°42 6°39 70°76 84:8
Salt present in 100 parts 26°62 22°47 28-1 15:2

salt and water

The particles of proteins in acid or alkaline solutions carry respectively a
positive or negative electric charge. Only at the iso-electric point, which is
slightly to the acid side of the true neutral point in all cases hitherto
investigated, does the charge disappear.

Salts like ammonium sulphate and sodium sulphate will cause separation
into two phases, one protein-rich and the other protein-poor, whatever the
charge upon the protein aggregates of the colloidal solution, if enough of the
salt be added. Such phase separation is very materially assisted if the
particles are positively charged, ie. in solutions more acid than the iso-
electric point. In the case of sodium chloride, indeed, acidification is
necessary with many proteins.

From the analogy of the “salting out” of alcohols, phenols, etc. one must
suppose that, by the gradual withdrawal of water from the protein aggregate
by the salt, a critical dispersion point is reached when the surface tension at
the interfaces causes the particles to run together. Supposing the particles
are not iso-electric with the continuous phase, either from the original
solution being acid or alkaline, or from the preferential adsorption of one of
the ions of the electrolyte, the possession of charge will lower surface tension,
so that, with charged particles, a higher concentration of salt will be required
to arrive at the critical dispersion point. With negatively charged particles,
in an alkaline solution, the charge cannot be neutralised by the more potent

H. CHICK AND C. J. MARTIN 393

ion SO,’ which, if adsorbed, would still further increase the negative charge.
We find that once the solution is more alkaline than the neutral point the
amount of alkali added makes no difference! to the amount of salt required
for precipitation, see Exp. ILI, Table VI and curve Fig. 5. The moment the
reaction is made more acid than the iso-electric point and the protein
particles carry a positive charge, this will at once be neutralised by the
adsorption of SO,” ions. As these are in such high concentration they are
apparently able to counteract in this way the maximum positive charge
imposed upon the particles by the addition of acid (see Exp. III, Table VI,
H* = 0-007 normal).

We venture to put forward the above interpretation from the analogous
action of SO,” upon protein particles under conditions which permit of the
demonstration of the existence of charge. In acid solution, protein particles,
carrying a positive charge, have been shown to be sensitive to the anion of
any electrolyte they may encounter. (Hardy [1900] for heated egg-white ;
Chick and Martin [1912] for denaturated serum proteins and egg-white ;
Chick [1913] for euglobulin and caseinogen.) Arguing from analogy we may
suppose that in the case of egg-albumin the SO,” ion of ammonium sulphate
will also be more readily adsorbed and any charge on the protein
neutralised if the solution be originally on the acid side of the iso-electric
point, 1.e. if the protein particles carry a positive charge.

The statements detailed above have not actually been substantiated in
case of pure egg-albumin, nor has the iso-electric point been determined.
This has, however, been done in case of serum-albumin [Michaelis and
Mostynski, 1910; Michaelis and Rona, 1910], and these two proteins other-
wise display close similarity as regards the conditions of their solution or
precipitation. We have not been able to put our interpretation to the
direct test because it is impracticable to determine the charge carried
by the particles of egg-albumin in the presence of excess of ammonium
sulphate, nor under these conditions were we able to study the influence of
addition of ions of varying valency. To test our hypothesis we therefore had

1 In the case of NagSO,, which as regards the influence of acid behaves in an analogous
manner to (NH,).SO,, excess of alkali (NaOH) favours precipitation, but, compared with the acid,
a high concentration is required. In accordance with the explanation of the influence of reaction
set forth above we presume that, while in acid solution the positively charged protein particles
attract the SO, ion, in alkaline solution the Na ion of the sodium salt is preferentially absorbed.
The electric charge on the particles is neutralised in both cases, but more readily in the first,
owing to the greater potency of the SO, ion. With (NH,),SO, no effect in solutions made
alkaline with ammonia can be demonstrated, owing, presumably, to the low ionisation of
(NH,)OH, especially in the presence of: excess of (NH).,SO,.

394 H. CHICK AND C. J. MARTIN

recourse to the expedient of withdrawing the water from the protein-water-
combination by the addition of alcohol to the point where a surface tension is
just manifest at the interfaces. In other words, the alcohol was added until
the solution became opalescent but short of commencing precipitation, which
could then be brought about by addition of a small amount of various
electrolytes. Two sets of experiments were made; in one set the original
protein solution was acid, and in the second set alkaline, and it was found, as
was expected, that an electrolyte was efficient in causing precipitation of the
protein in order of increasing valency of its anion in the first case and of its
kation in the second. All solutions contained the same concentration of
alcohol.

The results of a series of experiments with serum proteins are given in
Table IX. In solution A (acid) salts containing divalent (ammonium sul-
phate) and trivalent (sodium citrate) anions caused precipitation in respective
concentrations of 0:00055 and 0:00036 molar, whereas, in case of a monovalent
anion (magnesium nitrate) about ten times the concentration was required.
The valency of the kation was not without effect, but worked in the opposite
direction ; a small concentration of lanthanum nitrate (00007 molar) cleared
up the original opalescence of the solution. For the same reason magnesium
sulphate proved to be a much less efficient precipitant than ammonium
sulphate.

An exactly converse set of results was obtained when the protein solution
was originally made alkaline (B. Table IX). In this case lanthanum nitrate
was the most powerful precipitant of all the salts tried and the sulphate of
magnesium was much more effective than that of ammonium, At the same
time a small concentration of sodium citrate (0°0007 molar) caused the solution
to become clear. ;

With alkaline solutions higher concentration of lanthanum nitrate
(0:0036 molar) prevented the formation of a precipitate, no doubt owing to
the acquisition of a positive charge in excess of that needed to neutralise the
negative one originally possessed. In acid solution an analogous result
followed addition of sodium citrate to a concentration of 0:0007 molar.

Some experiments made with pure egg-albumin are set out in Table X.
The original solution was acidified and after addition of sufficient alcohol to
cause turbidity, determination was made of the concentration of a series of
sodium salts necessary to cause a precipitate to form. The influence of
increasing valency of the anion is very marked, sodium citrate and sulphate
being respectively about 800 and 25 times as powerful in this respect as
sodium chloride.

eg ery Ge

H. CHICK AND C. J. MARTIN

TABLE

3S

395

Precipitation of serum proteins by various electrolytes in presence

of alcohol (almost to precipitation) at 0°.

A. In acid solution.

Salt
Na,Cit

(NH4)2S04

MgSO,

Mg(NO3)2

La(NOs)3

Cone.
(molar)

000007
0-00036
0:00071
0:0036

0-00011
0°00055
0-0055
0-055

0°00055
0-0011
0-0055
0°055

0-0011
0°0055

0-00071
0:0036

B. In alkaline solution.

La(NO3)3

Mg(NOs)2

(NH4)2S04

Na 3Cit

0-000071
0:00036
0:00071
0-0036

0-00011
0:00055
0:0055
0:055

0-0011
0:0055
0:0011
0:0055

0:00071
0°0036

* Clearer than control solution containing alcohol only.

Protein content=about 0-1 °/,.

Cone.
(normal)

0:00022
0-0011
0°0022
0-011

0:00022
0-0011
0-011
0-11

0-0011
0-0022
0-011
0-11

0°0022
0-011

00022
0-011

0-00022
0-0011
0-0022
0-011

0-00022
0-0011
0-011
0-11

0:0022
0-011

0:0022
0-011
0°0022
0-011

Degree of
precipitation

Cone. (molar)
required for com-
plete precipitation

0-00036

0:00055

0°0011

0:0055

0-00036

0°00055

0:0011

No precipitation

— =clear solution, clearer than the control, containing alcohol only.
— + =opalescent solution.
+ =partial precipitation.
+ 4+ =complete precipitation, filtrate protein-free or containing trace only.

396 H. CHICK AND C. J. MARTIN

TABLE X.

Precipitation of pure egg-albumin, in acid solution, by sodium salts,
in presence of alcohol almost to precipitation ; influence of anions.

Protein content=0°7 /p.
Temperature, 0°.

Concentration required Concentration necessary
to cause precipitation to commence precipitation

in presence of alcohol in absence of alcohol

Salt (Molar) (Molar)
Citrate (neutral) 0-00013 1°37
Phosphate 0-00016 2°50
Tartrate 0-003 1-60*
Sulphate 0:004 1:69
Acetate 0-09 =
Chloride 0:10 4°3*
Chlorate 0°35 ; —
Chromate 0:4* 1:71*

* Did not precipitate.

In the same Table are given the results of a series of experiments in
which precipitation took place in absence of alcohol and it will be seen that no
such influence of valency can be detected here. The relation of the various
electrolytes, with the exception of sodium citrate, is that expressed by the
Hofmeister series [1888] and doubtless conditioned by their relative water-
drawing capacity. When the protein is on the point of being precipitated
after the necessary water-withdrawal has taken place, the influence of the
anion or kation of the electrolyte present can complete precipitation by
neutralising a charge. This occurs for example when protein almost pre-
cipitated by ammonium suphate is made slightly acid, or when to protein
almost thrown out by alcohol a trace of an appropriate electrolyte is added.
In the former case, as long as the solution remains alkaline, the effect of
presence of ammonium sulphate with its divalent anion will rather be to
increase the negative charge on the protein.

An additional piece of evidence in support of the view that the
charge carried by the protein particles is an important factor in deter-
mining the concentration of salt requisite to occasion separation into two
distinct phases, was obtained from studying the “salting out” of egg-
albumin with calcium chloride. In this case the positive ion of the electro-
lyte is prepotent and we should expect it to act more efficiently if the
solution containing the protein is made more alkaline than the iso-electric

H. CHICK AND C. J. MARTIN 397

point. This proved to be the case’. It was found that in a solution contain-
ing 0°8 °/, protein and 37°7 °/, CaCl, an opalescence was developed on standing
for about 2 hours at 18° when the hydrogen ion concentration was about
23 x 10-7 normal. In presence of a small concentration of Ca(OH),, under
otherwise similar conditions, the hydrogen ion concentration fell to about
0:05 x 107 normal. In this case almost all the protein was precipitated and
only a trace remained in the filtrate.

Precipitation by calcium chloride differs from that by ammonium sulphate
in the fact that the process is irreversible, and the precipitate is formed
slowly and becomes insoluble. We were not able to make satisfactory
experiments with other electrolytes of the same character, e.g. the nitrates and
chlorides of magnesium and barium, because with egg-albumin precipitation
is only very partial even in alkaline solution, but in these cases also, we were
able to satisfy ourselves that alkalinity of the solution assisted the separation
of a precipitate (protein-rich phase).

SUMMARY.

1. The precipitation of egg-albumin by ammonium sulphate is, as Spiro
demonstrated to be the case with sodium sulphate, and casemogen and gelatin,
due to the separation of the system into a protein-rich phase and a watery
phase, and to a certain extent is analogous with the salting-out of alcohol.

2. The first effect of concentrated salt is to withdraw water from the
protein aggregates. A surface tension is in consequence developed at the
interfaces, which causes the protein particles to aggregate, thus dividing the
system into two distinct phases (precipitate and filtrate).

3. All three constituents of the system, viz. protein, water and salt, are
present in each phase ; the proportion, however, is different. The precipitate
(protein-rich phase) contains relatively little water and salt, and the filtrate
(watery phase) relatively little protein. Under appropriate conditions practi-
cally all the protein may be precipitated, only a trace remaining in the
filtrate.

The two phases are in equilibrium, and any alteration in the amount of
any one of their three constituents is followed by a change both in their
composition and volume. Thus increase in concentration of salt or of protein

1 Precipitation by calcium chloride is also assisted in cases where the solution is markedly
acid, a comparatively large concentration of acid (HCl) being needed to demonstrate the
phenomenon. The explanation is exactly analogous to that offered for the case of sodium

sulphate, see footnote, p. 393, but in the reverse sense, much more acid being required to produce
this effect than is the case with alkali.

398 H. CHICK AND C. J. MARTIN

is followed by a corresponding increase in the protein-rich phase (pre-
cipitate). Owing to rigidity of the latter readjustment is, however, slowly
accomplished.

4. The relative volume of the two phases (amount of precipitate) is
altered by varying the temperature.

5. The exact proportion of salt, protein and water at which phase
separation (precipitation) occurs and the relative volume of the two phases
(amount of the precipitate) 1s very sensitive to hydrogen ion concentration in
the neighbourhood of the iso-electric point. In the case of (NH,),SO,, when
the hydrogen ion concentration varied from 10~* to 10~* normal, the amount
of the precipitate increased from a negligible amount to a maximum.

6. With CaCl, in which the kation is prepotent, a similar effect was
observed, but in the opposite direction.

7. An interpretation of the dominating influence of hydrogen ion con-
centration over this range is put forward and some experiments in support of
it adduced.

The principles discussed in this paper must be borne in mind whenever
salting-out is made use of in the fractionation of protein solutions and the
purification of isolated fractions by re-solution and reprecipitation. Con-
clusions as to the homogeneity of proteins isolated by salt precipitation must
also be reconsidered in the light of these results.

REFERENCES.

de Bruyn (1900), Zeitsch. physikal Chem. 32, 63.
Chick (1913), Biochem. J. 7, 318.

and Martin (1912), J. Physiol. 45, 261.
Hardy (1900), Proc. Roy. Soc. 66, 110.

Hofmeister (1888), Arch. exp. Path. Pharm. 24, 247.
—— (1889), Arch. exp. Path. Pharm. 25, 1.

—— (1891), Arch. exp. Path. Pharm. 28, 210.
Hopkins (1899-1900), J. Physiol. 25, 306.

and Pinkus (1898), J. Physiol. 23, 130.
Kauder (1886), Arch. exp. Path. Pharm. 20, 411.
Lewith (1888), Arch. exp. Path. Pharm. 24, 1.
Mellanby (1907), J. Physiol. 36, 288.

Michaelis and Mostynski (1910), Biochem. Zeitsch. 24, 79.
and Rona (1909), Biochem. Zeitsch. 18, 318.
— (1910), Biochem. Zeitsch. 27, 38.

Pauli (1898), Pfliiger’s Archiv, 71, 333.

and Rona (1902), Beitréige, 2, 1.

Posternak (1901, 1), Ann. Inst. Past. 15, 85.

(1901, 2), Ann. Inst. Past. 15, 169.

Spiro (1904), Beitrédge, 4, 300.

XL. SOME OBSERVATIONS ON THE
ESTIMATION OF UREA.

By JOHN ALEXANDER MILROY.

From the Physiological Laboratory, the Queen's University of Belfast.

(Received June 11th, 1913.)

The chief aim of the following investigation was to find out a simpler
method for the estimation of urea than those at present in use. The hypo-
bromite method, although rapid and simple, has been shown to be unreliable
[Morner, 1903]. All the more accurate methods depend like that of Folin
[1901, 1902] on the conversion of urea into an ammonium salt, and the
estimation of the ammonia by distillation. These methods necessarily involve
a preliminary estimation of the preformed ammonia of the urine. No useful
purpose would be served by an enumeration of the chief modifications which
the methods have undergone. Reference may be made to the exhaustive
account given in the revised edition of Neubauer and Huppert’s Analyse des
Harns [1910].

The distillation of the preformed ammonia and that derived from. the
hydrolysis of urea occupies in each case about an hour. A considerable
reduction of time and simplification of the method would be secured, if the
distillation were replaced in both instances by titration in the presence of
formaldehyde. The main objections to the latter method are that other
amine derivatives, notably amino- and diamino-acids affect the result of
the titration [Schiff, 1899, 1901, 1902], and that the best method for the
determination of the neutral point, from which one starts, is still unsettled.
Ronchese [1907] and Malfatti [1908] employ phenolphthalein throughout
as indicator in their method for the estimation of ammonia in the urine
by formol-titration, while Sérensen [1907, 1910] and Henriques [1909] use
azolitmin paper as the indicator in neutralising the solution and phenol-
phthalein for the titration of the acids set free by the addition of neutral
formol in their method for estimating amino-acids. De Jager [1909, 1;
1910, 1] follows the same procedure as Malfatti after adding excess of
neutral potassium oxalate or of potassium sulphate. Folin [1903] had

400 J. A. MILROY

previously made similar use of potassium oxalate in determining the acidity
of the urine. The fact that solutions of chemically pure ammonium salts
react acid to phenolphthalein has long been recognised. In consequence,
the results of formol-titrations with this as the sole indicator are too low.
One objection to the method of overcoming this ‘difficulty adopted by
Sérensen and Henriques is the inconvenience of the frequent testing with
azolitmin paper required in the course of the titration. The addition of
litmus to the solution renders the titration difficult and uncertain owing to
the conflicting colour changes of the two indicators. In its place, I have
used methyl-red [Rupp and Loose, 1908 ; Tizard, 1910; Howard and Pope,
1911] as the indicator of the neutral point and phenolphthalein for the
determination of the amount of acid set free by the addition of neutral
formol. Methyl-red has the advantage of being an exceedingly sensitive
indicator for the titration of weak bases such as ammonia, while it is
comparatively insensitive to weak acids. On the other hand, phenolphthalein
is a good indicator for weak acids, but is not a reliable indicator for weak
bases. The colour changes of the two indicators do not interfere with one
another, and they may therefore be used together without any difficulty
arising. Methyl-orange may be used instead of methyl-red, but it does not
give such a sharp endpoint.

1. PRELIMINARY OBSERVATIONS ON THE TITRATION OF PURE
AMMONIUM SALTS.

The following experiments were carried out in order to ascertain the
degree of accuracy of the method as compared with formol-titrations of
ammonium salts in the presence of phenolphthalein alone. Five drops of
a 0°5°/, alcoholic solution of phenolphthalein were added to 10 ce. of a
decinormal solution of pure ammonium chloride, which had been previously
rendered faintly acid by the addition of dilute hydrochloric acid. The
solution was then neutralised by the addition of sufficient decinormal caustic
potash just to render the solution faintly red. Excess (about 5 cc.) of neutral
formol (50 cc. 40 °/, formaldehyde, 1 cc. of 0°5°/, alcoholic phenolphthalein,
and sufficient decinormal caustic potash to render the solution faintly pink)
was then added, and the addition of decinormal alkali continued until the
solution just acquired a faint permanent red colour. The addition of 9°55 ce.
of decinormal alkali was found to be necessary to reach this point instead
of the true value 10 cc. Two to three drops of a 0:2°/) alcoholic solution
of methyl-red were added to another 10 cc. of a decinormal solution of

ee

J. A. MILROY 401

ammonium chloride, and then sufficient decinormal caustic potash just to
convert the red colour of the solution to a yellow. On the addition of
neutral formol, the solution again acquired a red colour owing to the
hydrochloric acid, set free by the interaction of the formaldehyde with
the ammonium chloride, acting upon the methyl-red. On the further
gradual addition of a measured quantity of decinormal caustic potash, the
red colour fades and is replaced by a yellow. The addition of alkali was
continued until a faint permanent red colour, due to the action of the alkali
on the phenolphthalein, just appeared. 9°95-10 cc. were required to reach
this point. Repeated control estimations of the same character with solu-
tions of pure ammonium sulphate as well as of ammonium chloride gave
similar results. It will be noted that, when the two indicators are used, the
solution during titration passes through three changes of colour. It is first
red owing to the action of free mineral acid on the methyl-red, then orange
to yellow when the solution contains only free organic acid, and finally red
when just alkaline owing to the action of slight excess of free alkali on
the phenolphthalein. The organic acid is formic acid present in the formal-
dehyde solution and resulting from the oxidation of part of the formaldehyde.
Thus, prior to the addition of the neutral formol, the solution contains only
ammonium chloride; while after its addition the fluid contains hexamethy-
lenetetramine, free hydrochloric and formic acids, the latter two being in
varying proportions according to the amount of formic acid present in the
original solution of formaldehyde. In the presence of traces of free formic
acid, methyl-red has nearly the same tint as in neutral solution, and, there-
fore, when all the free mineral acid has been neutralised, the colour of the
methyl-red changes to faint orange and later to yellow, although the solution
is found still to be distinctly acid to phenolphthalein.

2. APPLICATION OF THE FOREGOING RESULTS TO THE ESTIMATION
OF PURE UREA IN AQUEOUS SOLUTION.

Folin’s method of hydrolysis [1901, 1902] is not applicable in this case,
since the presence of a large amount of magnesium chloride interferes with
the subsequent titration. A slight modification of Benedict and Gephart’s
method of hydrolysis [1909] was consequently adopted. The modification
consisted merely in using more dilute acid for the hydrolysis of the urea.
When a considerable excess of acid is used, the large amount of neutral salt
formed in the course of neutralisation interferes with the accuracy of the
determination of the endpoint [Michaelis and Rona, 1909; Andersen, 1911].

402 J. A. MILROY

The use of a smaller amount of acid also lessens the risk of the decomposition
of other nitrogenous substances, when the method is applied to urine. After
using this method of hydrolysis for some time, I found that it had been
previously employed by Henriques and Gammeltoft [1911].

At first the hydrolysis was carried out in sealed glass tubes heated in
a glycerol bath to 160° for four hours. T'wo cc. of a 2:006°/, solution of
pure urea and 2 cc. of normal sulphuric acid were placed in a glass tube,
which was then sealed and heated in a glycerol bath. After being allowed
to cool, the sealed tube was opened and the contents carefully transferred
to a flask, the broken tube being thoroughly washed with distilled water.
The fluid was then neutralised in the presence of methyl-red as indicator,
and the ammonia determined by means of formol-titration using phenol-
phthalein as the indicator. The following are the results given in tabular
form. One ce. of decinormal alkali is equivalent to 0-003 g. of urea,

Ce. of Ce. of N. acid Ce. of N/10 alkali G: of Theoretical
2-°006°/, urea _ for hydrolysis (formol titr.) urea found amount of urea
2 2 H.SO,4 13°3-13°35 0-0399-0:04 0:0401
2 2°5 HCl 13°4 0:0402 0:0401

Since the use of sealed tubes is somewhat inconvenient, Benedict and
Gephart’s method [1909] of heating the solution in test-tubes covered with
leadfoil and placed in an autoclave kept at a temperature of 155° for 1°5 hours
was afterwards employed. A large number of estimations can readily be
carried out simultaneously by means of this method. The temperature in
the autoclave was varied in the first experiments in order to find the most
suitable temperature for hydrolysis.

23101 g. of pure urea were placed in a 200 ce. flask, dissolved in water,
100 cc. of normal sulphuric acid were added, and the whole made up to
200 cc.; 5 cc. portions of the acid solution were then hydrolysed. The
following table gives the results.

Ce. of N/10 alkali G. of Theoretical
Temperature for formol-titr. urea found g. of urea
144-145° 18°8 0:0564 0:0577
(70 minutes)
152-153° (1) 19°15 0:05745 95
(30 minutes (2) 19-2 0°0576 5p
longer) (3) 19°15 0:05745

Since the application of the foregoing results to the estimation of the
urea in the urine involves a preliminary formol-titration of preformed
ammonia and other amine derivatives, it was essential to test the method
by applying it to the titration of amino-acids as well as to mixtures of
amino-acids with ammonium salts.

een

J. A. MILROY 403

3. TITRATION OF AMINO-ACIDS ALONE AND WITH AMMONIUM SALTs.

Glycine being easily prepared in a pure state, was selected as a type
for the tests. Since one hag to deal with a mineral acid solution in the
final formol titration, it was obviously essential to carry out the estimation
under similar conditions, and also to compare the results with those obtained
when phenolphthalein was used as the sole indicator. The following are
the results given in tabular form. The glycine solution was 0°71 °/,.

These results indicate that, when the two indicators are used, the
estimations of amino-acids and ammonium salts are more accurate than
when phenolphthalein is the sole indicator. The fact that formol-titration
yields too low results in the case of mixtures of amino-acids and ammonium
salts was first noted by de Jager [1909, 2]. Henriques and Sorensen [1910]
have suggested an explanation of this anomaly. More recently de Jager
[1910, 2] has shown that the addition of urea enables one to obtain correct
results. The latter fact is illustrated by the results given in (4) of the
table.

Ce. N/10 alkali Formol-titr. Theoretical
Solutions used (for neutralisation) (found) result
(1) 10 ce. glycine 2°55 (indicator methyl-red) 9-4 9°46
2-55 cc. N/10 HCl
(2) 10 cc. glycine 2°85-2-9 (indicator phenol- 9-25 PP
2°75 ce, N/10 HCl phthalein)
(3) 10 cc. glycine 2 ce. (methyl-red) 18°65 19°46

10 ce. N/10 NH,Cl
2 ec. N/10 HCl

(4) 10 cc. glycine 2 ce. (methyl-red) 19-4 aA
10 ee. N/10 NH,Cl
2 ce. N/10 HCl
0-5 g. urea

4. ESTIMATION OF UREA IN URINE.
(1) Without a preliminary purification of the urea.

Since the presence of phosphates renders the neutral point indefinite
for both indicators, and since the neutral points for the two indicators
are markedly different in solutions of phosphates, it was necessary to pre-
cipitate the phosphates with baryta mixture. The addition of an unduly
large excess of barium chloride is to be avoided, as its presence interferes
slightly with the accuracy of the titration; but a small excess is immaterial.
When sulphuric acid is used for the hydrolysis of the urea the excess of
barium present in the filtrate is precipitated as barium sulphate, and it

404 J. A. MILROY

was mainly for this reason that normal sulphuric acid rather than hydro-
chloric acid was used for the majority of the hydrolyses. The results of the
method of formol-titration were controlled by comparison with the values
for ammonia obtained by distillation with magnesia and with barium
hydrate.

100 cc. of normal urme were mixed with 10 cc. of a binormal solution
of barium chloride and sufficient barium hydrate to make the mixture
distinctly alkaline, and finally made up with water to 200 cc. After the
fluid had been thoroughly shaken, it was allowed to stand either for a few
minutes or preferably in the presence of toluene for twelve hours. Henriques’
method [1909] of freeing the urine from phosphates, sulphates and carbonates
may be used as an alternative to the foregoing. The solution was then
filtered. 25 ce. of the filtrate required 18°5 cc. of decinormal hydrochloric
acid to render it neutral to methyl-red. The endpoint is not quite sharp.
Excess of neutral formol (about 5 cc.) was added, and the titration with
decinormal alkali completed. The average of two titrations was 5:2 cc. of
decinormal alkali. This result gives the approximate amount of nitrogen
in the form of preformed ammonia and amino-acids in 12°5 cc. of urine.

A number of 10 cc. portions of the filtrate were then mixed with 7-8 ce.
of normal sulphuric acid, and placed in test-tubes in the autoclave, which
was kept at a temperature of 155° for 15 hours. Although the fluid had
darkened considerably in colour as a result of the heating with acid, the
colour on dilution was not so deep as to interfere appreciably with the
subsequent formol-titration. The ammonia in some portions was then set
free by the addition of magnesia or barium hydrate, distilled off, and
absorbed by excess of decinormal sulphuric acid (50-100 cc.) placed in the
receiver. In others the ammonia and amino-acids were determined by
formol-titration. Prior to carrying out the formol-titration it 1s advisable
although not essential to filter off the barium sulphate. A considerable
amount of pigment adheres to this and is removed with it. The test-tube
and precipitate were then thoroughly washed, and the collected filtrate and
washings treated by formol-titration. The following table (I) gives the
results obtained. The deduction referred to in the table corresponds with
the amount of substances in the original urine capable of titration by the
formol method, consisting mainly of preformed ammonia and amino-acids,

An examination of the table indicates that the results of the formol-
titration are invariably higher than those obtained by distillation; but
agree well with one another. The difference is to be explained by the
presence of non-volatile amino-groups set free by the hydrolysis with acid,

Se eee eee ;™

J. A. MILROY 405

which raise the results obtained by formol-titration ; but obviously cannot
affect the values for ammonia obtained by distillation. The hydrolysis with
normal acid splits up hippuric acid and any polypeptides which may be
present, and thus renders the results of formol-titration higher than those
obtained by distillation. The preliminary formol-titration, on which the
deduction is based, gives the amount of nitrogen in the form of preformed
ammonia and amino-acids: while, after hydrolysis, the final formol-titration
gives the preformed ammonia and amino-acids, together with the amino-
acids set free by hydrolysis, and the ammonia resulting from the decomposition
of the urea. The difference between these two values represents the ammonia
derived from urea plus the nitrogen of any amino-acids derived from hydro-
lysis of polypeptides and of hippuric acid. On the other hand, the method of
distillation gives only the preformed ammonia and that formed by hydrolysis
of urea. In calculating the results given in the table, the same deduction
was made from the values obtained by distillation as from those by formol-
titration. Strictly speaking, a smaller deduction should have been made
from the latter. An independent determination of the preformed ammonia
would be required to ascertain the true value of this correction. This point
will be dealt with in a later part of this paper.

TABLE LI.
(The quantity of urine was 5 cc. for each estimation.)
Method Ce. of N/10 alkali Deduction /) of urea °/) of urea-N.
Formol-titration (1) 32°15 2-1 ce, 1°803 0°8414
N/10 alkali

ef (2) 32:1 2-1 1-800 0-840

53 (3) 32:4 2-1 1°818 0°848
Distillation with (1) 32°05 2-1 1-797 0°8386
barium hydrate

(2) 32 2-1 1-794 0°837

A (3) 31-9 2-1 1-788 0-834
Distillation with MgO (1) 31:7 2-1 1776 0°829

=f 8 (2) 31:75 2-1 Herr fs) 0-830

The difference between the result obtained by distillation and that by
formol-titration gives an approximate value for the non-volatile amino-
derivatives. A more accurate estimate might be made by taking advantage
of the fact ascertained by de Jager [1910, 2] that the addition of urea is
necessary in order to secure the true result for the sum of amino-acids and
ammonia, This addition is unnecessary in the first formol-titration, since
sufficient urea is present in the urine; but after hydrolysis the addition

Bioch. vi 27

406 " J. A. MILROY

of pure urea to the neutralised solution would be necessary before carrying
out the second formol-titration. I have not considered the correction of
sufficient importance in the case of the normal urines examined to warrant
its introduction into the table; but in the event of a marked difference
being found between the values obtained by distillation and formol-titration
indicating excess of amino-acids, it would be advisable to make the
correction.

The next table (II) gives the results of a similar series of estimations in
a sample of another urine. Normal hydrochloric acid was used in this
instance instead of sulphuric.

The same features will be noted as in the previous table. When barium
hydrate is used as the fixed alkali, the values for ammonia are usually
higher than when magnesia is employed. In order to obtain constant
results, the distillation has to be more prolonged in the case of magnesia
than with barium hydrate. The distillation usually occupied 1-2 hours
in both cases. The differences between the results of formol-titration and
of distillation is greater than in the other urine.

TABLE IZ.

(The quantity of urine used for each estimation was 5 cc.)

Method Ce. of N/10 alkali Deduction 0/9 of urea %/ of urea-N.

Formol-titration (1) 36°65 2°5 2-049 09562

5 (2) 36°6 zs 2-046 0°9548

rf (3) 36°6 B 2-046 0:9548
Distillation with
barium hydrate (1) 35°45 2°5 ICT 0-9226

a (2) 35°65 5 1-989 0-9282

3 (3) 35°75 59 1-995 0:931
Distillation with MgO (1) 35 2°5 1:95 0-910

» ” (2) 35°5 = 1:98 0-924

(2) Estimation of urea in urine after a preliminary extraction
of the urea with a mixture of alcohol and ether.

Folin’s [1901] and Benedict and Gephart’s methods [1909], as well as all
other methods depending on the conversion of urea into the ammonium salt
of the acid used for hydrolysis, give inaccurate results when carbohydrates
are present in the urme. Mdérner [1903] has shown that this is due to the
fact that the humin-substances formed by the action of the acid on the

sugar contain firmly bound nitrogen, when ammonium salts are present or
formed during the hydrolysis.

J. A. MILROY 407

Morner adopted the following method in order to obtain a solution
of urea free from carbohydrates. The phosphates were precipitated from
5 cc. of urine by the addition of 1°5-2 g. finely powdered barium hydrate,
and the fluid was then extracted with 100 ce. of a mixture of two volumes
of aleohol and one of ether. After being thoroughly mixed, the solution
was allowed to stand for 24 hours, then filtered, and the precipitate washed
with the mixture of alcohol and ether, the filtrate and washings being
collected in the flask afterwards to be employed in the distillation. The
alcohol and ether were then distilled off, and the residue after being freed
from ammonia by concentration at 50° after the addition of magnesia, was
hydrolysed by Folin’s method [1901]. I carried out the first stages of
the method in the same way; but instead of freeing the aqueous solution
from ammonia by the addition of magnesia, I determined the amount of
amino-nitrogen by means of formol-titration. The method of hydrolysis
adopted was the modification of Benedict and Gephart’s method [1909]
previously described. In some cases normal hydrochloric acid was used for
the hydrolysis, in others normal sulphuric acid. The following table
summarises the results.

TABLE III.

Results in ee.

Method of N/10 NH, Deduction 9], urea 9/, urea-N,
1. Direct formol- (a) 29°55 0°85 1-722 0°8036
titration (b) 29-7 1-731 0-8078
(c) 29°65 fe 1:728 0-8064
2. Distillation with (a) 29-1 0°85 1695 0-791
Ba(OH), (b) 29°15 5 1-698 0-7924
3. Distillation with (a) 28°45 0°85 1-656 0°73
MgO (b) 28° 5 1°659 0-774
(c) 28-1 7 1-635 0-763

The following are the chief features revealed by an examination of the

table. The comparatively low

value of the deduction is to be explained

by the fact that the method of preparing the aqueous solution of urea
involves the loss of the greater part of the preformed ammonia. ‘The values
for urea obtained by the formol method are, as in the previous estimations,
higher than those obtained by distillation. Amino-derivatives evidently
pass into the alcohol and ether mixture. Hippuric acid is probably the most
important of these. The glycine derived from its hydrolysis must raise
the results of the formol-titration, while it obviously cannot affect the values
obtained by distillation. The results obtained by distillation with magnesia
are considerably lower than those with barium hydrate.

27—2

408 J. A. MILROY

5. .FINAL CONTROL OF THE METHOD BY AN INDEPENDENT DETERMINATION
OF THE PREFORMED AMMONIA.

The estimations hitherto given suffer from the defect of not having been
controlled by an independent determination of the preformed ammonia.
Reference was previously made to this fact. In a final series of experiments,
I have estimated the total nitrogen by Kjeldahl’s method, the nitrogen of
the preformed ammonia by the method of Kriiger, Reich, and Schittenhelm
[1903], the ammonia and urea nitrogen after hydrolysis by distillation,
and the nitrogen of the amino-acids by Henriques’ method [1909]. The
results for 100 cc. of urine are stated in Table IV in terms of decinormal
alkali and in Table V as percentages.

TABLE IV.
Results for 100 ce. urine expressed in terms of decinormal nitrogen.
Urea + NH; + amino-
Urea and NH3-N. acid nitrogen NH; + amino-N. Ammonia-N.
Total N. (distillation) (formol-titration) (in original urine) (distillation)
587 529 550 33°6 23°9
(Sorensen)
585 528 544 35:2 24°2
(methyl-red)
583 524 546 — —
585 (mean) 527 (mean) BAT (mean) os 24-05 (mean)
TABLE V.
Urea-N.
Total N. Urea-N. (formol-titr.) Amino-N. Ammonia-N.
585 503 513 9°6 24°05
(Sdrensen) (Sérensen)
— — 511°4 11°2 —
(methyl-red) (methyl-red)
—0819% 0-704°%, 0-718 J, (S.) 0-013 "/, (8.) 0:0337 °/,
0-716 °/, (M.) 0:0157 °/, (M.)

The value given in column 2 of Table V is arrived at by deducting
the amount of preformed ammonia as determined by the method of Kriiger,
Reich, and Schittenhelm from the total amount of ammonia obtained by
distillation after complete hydrolysis of the urea. This may be regarded as
the correct value for urea arrived at by a standard method. The values for
the nitrogen of urea given in the third column of Table V are obtained by
deducting the results of column 4, Table IV, from those of column 8, Table IV,
using in the first instance the results of Henriques and Sérensen’s method
and in the second case those obtained by the use of methyl-red and phenol-
phthalein as the two indicators. The values for the nitrogen of the
amino-acids given in column 4 of Table V are arrived at by subtracting the

——— ———@<@<@<_—— Se

J. A. MILROY 409

results of the method of Kriiger, Reich and Schittenhelm given in column 5
of Table IV from those of column 4, Table LV, using in the one case Henriques’
method, and in the second formol-titration with methyl-red as the first indicator.

Sorensen [1907, 1910] used fifth-normal alkali for the titration of the
amino-acids. Notwithstanding the fact that this strength is to be preferred
for reasons stated by that author, I was obliged to continue the use of
decinormal alkali as it had been used throughout the foregoing investigation.
These final results indicate that the error in the method of estimating
urea in this urine by formol-titration is a positive one amounting to approxi-
mately 1:7 °/,.

The chief advantages of the formol-method of estimating urea are that
it is not only more rapidly carried out than the methods involving distilla-
tion, but also entails less arrangement of apparatus. The fact that the method
gives at the same time an approximate estimate of the preformed ammonia
and amino-acids is also an advantage, which compensates to some extent for
the less degree of accuracy attainable. These facts suggest that it might
find useful application for the examination of urines in clinical laboratories
as a substitute for the hypobromite method. Another possible application
of the method might be to the study of the hydrolysis of urea by enzymes

and micro-organisms.
REFERENCES.

Andersen (1911), Skand. Arch. Physiol. 25, 104.

Benedict and Gephart (1909), J. Amer. Chem. Soc. 30, 1760.
De Jager (1909, 1), Zeitsch. physiol. Chem. 62, 333.

—— (1909, 2), Zeitsch. physiol. Chem. 64, 337.

—— (1910, 1), Zeitsch. physiol. Chem. 67, 1.

(1910, 2), Zeitsch. physiol. Chem. 67, 107.

Folin (1901), Zeitsch. physiol. Chem. 32, 504.

(1902), Zeitsch. physiol. Chem. 36, 333.

(1903), Amer. J. Physiol. 9, 263.

Henriques (1909), Zeitsch. physiol. Chem. 60, 1.

—— and Gammeltoft (1911), Skand. Arch. physiol. 25, 153.
and Sérensen (1910), Zeitsch. physiol. Chem. 64, 120.
Howard and Pope (1911), J. Chem. Soc. 99, 1333.

Kriiger, Reich, and Schittenhelm (1903), Zeitsch. physiol. Chem. 39, 73, 165.
Malfatti (1908), Zeitsch. anal. Chem. 47, 273.

Michaelis and Rona (1909), Biochem. Zeitsch. 23, 61.
Moérner (1903), Skand. Arch. Physiol. 14, 321.

Neubauer and Huppert (1910), Edition 11, pp. 560-577.
Ronchese (1907), J. Pharm. Chim. (6), 25, 611.

Rupp and Loose (1908), Ber. 41, 3905.

Schiff (1899), Annalen, 310, 25.

(1901), Annalen, 319, 59, 287.

(1902), Annalen, 325, 348.

Sérensen (1907), Biochem. Zeitsch. 7, 45.

(1910), Biochem. Zeitsch. 25, 1.

Tizard (1910), J. Chem. Soc. 97, 2485.

XLI. GAS-ELECTRODE FOR GENERAL USE.
By GEORGE STANLEY WALPOLE.
From the Wellcome Physiological Research Laboratories, Herne Hill, S.E
(Received June 11th, 1913.)

The standard forms of hydrogen-electrode, such as those used by
Dolezalek [1899], Wilsmore [1900] and Bjerrum [1910] and, for fluids
containing carbon dioxide, Hasselbalch [1910], must remain indispensable
when physical measurements of the highest order of accuracy are under-
taken.

With these, after the expenditure of considerable time and care, and
with continued vigilance over every part of the electrical apparatus, con-
secutive results concordant to about 0:05 millivolt can sometimes be obtained.
But, for these results to have any meaning beyond the nearest millivolt or
two, if is essential that the diffusion potential difference of the cell should
be accurately known in every determination: and as this is recognised, by
Cumming and others whose investigations continue to throw light on this
difficult subject, to be almost impossible, it follows that, as far as absolute
H’ ion determinations are concerned, attention should be focussed more
particularly on diffusion potential errors (in view of their greater magnitude)
than on those found at the electrode itself.

In general laboratory practice, therefore, especially when dealing with
protein-containing materials, which may or may not be free from carbon
dioxide, the most suitable electrode is not one which, when coupled with
an ideal electrical apparatus, will give results of this high order of accuracy,
regardless of the expenditure of time and material. Rather it is one which,
without exaggerating appreciably the diffusion potential error, will give,
under ordinary working conditions and in a few minutes, results correct to
1 millivolt or so on one or two ce. of fluid without loss or contamination.

Anticipating a long series of H’ ion determinations on protein solutions,
some of them containing carbon dioxide, I endeavoured to devise such an
electrode. The arrangement subsequently adopted and the experiments
made to discover its imperfections are described below. Only one form of

© TS ea ee =

G. S. WALPOLE 411

apparatus has been used for all cases. When dealing with a solution
containing carbon dioxide or other dissolved gases a slight modification of
technique is all that is necessary.

APPARATUS.

The modifications of existing apparatus which I have employed consist of
(1) the electrode vessel,
(2) the filling apparatus,
(3) the support in the constant temperature bath.

The electrode vessel (ABD, Fig. 1) is somewhat more complicated than the
simple V shape generally employed, but this is amply compensated for by the
simplicity of the other parts of the apparatus and the many conveniences
accompanying its use. I have used throughout the platinum point advocated
by Michaelis, just making contact with the surface of the fluid. The platinum
point is mounted at the end of a glass tube A. The protruding end is
blackened in the usual manner. The other end makes contact with a small
globule of mercury placed inside the tube.

The lower end of A may be ground to fit the outer tube B or else a joint
may be made by means of a rubber stopper as in the diagram. Attached to
B at the side is a very fine bore capillary tube carrying a tap D, and at the
lower end is a second capillary tube not quite so fine, of about 1 mm. bore.
At the lower end of this may be fitted a very small glass stopper, but it is
not necessary and has certain disadvantages.

The filling apparatus used with all these electrode vessels is depicted
together with other apparatus in Fig. 2. It consists of a three-way piece F
connecting a 5 ce all-glass syringe G, well lubricated with vaseline, a short
piece of fine-bore stout rubber tubing H and a glass tube carrying a glass
stopcock which is connected with the hydrogen supply.

As support for the electrode vessels in the constant temperature bath,
while taking potentiometer readings, I have used a glass trough KX (Fig. 1)
in which the electrode vessels stand side by side in a suitable connecting
solution. For single electrode vessels a large test tube serves excellently.
The other half-electrode employed, a tube from which leads into the glass
trough, is also immersed in the water-bath.

I have used principally the calomel-saturated potassium chloride half-
electrode recently described by Michaelis and Davidhoftf [1912] and a standard
calomel N/10 KCl half-electrode. The connecting solutions were saturated
potassium chloride solution, saturated ammonium nitrate [Cumming, 1907],

412 G. S. WALPOLE

and solutions of potassium chloride 1°75 N and 3°5 N respectively, employed
for the extrapolation method of Bjerrum.

In some experiments on diffusion potential two gas electrodes of the
vertical type were stood side by side in the trough containing connecting

solution.

EXPERIMENTS WITH SOLUTIONS FREE FROM CARBON DIOXIDE
AND OTHER DISSOLVED GASES,

Technique. The electrode vessel is first of all connected by the capillary
side tube D (Fig. 1) to the rubber tube of the filling apparatus (Shown at GFH,
Fig. 2), and hydrogen passed for a few seconds, both glass taps being open.
Then, by alternately drawing out the piston with the tap on the electrode
vessel shut, and pushing it home with the tap open, the last traces of air are
expelled from the dead space of the syringe and the T-piece. The lower end
of the electrode vessel is now brought under the surface of a sample of the
fluid to be examined and the glass tap on the T-piece shut. By pulling out
the plunger of the syringe the fluid is drawn up until its surface is just at
the point where the platinum point is sealed to the glass. Slight movements
of the plunger in and out now cause the liquid by rising and falling to wash
the platinum point well without wetting the tube A (Fig. 1) supporting it.
Finally the height of the liquid is adjusted till the point just touches the
surface, the glass tap D on the side capillary closed, and the vessel containing
the remainder of the sample of fluid taken away.

The electrode vessel is now disconnected from the filling apparatus, and the
lower capillary wiped dry with a piece of clean filter-paper and placed in the
glass trough in a constant temperature bath. There is sufficient connecting
solution in the trough to cover the electrode vessel above the rubber stopper.
The two copper wires from the potentiometer are now led, one into the
calomel half-electrode, and the other into the tube A and the reading is

taken from which the value of Py (the negative exponent of the H’ ion
concentration) may be calculated directly. .

In order to be certain that the platinum electrode is working properly
and that equilibrium is established instantaneously, a few minutes at least
may be allowed to elapse before a second reading is taken. Meanwhile the
filling apparatus can be used to fill another electrode vessel with another
sample, and when this is introduced into the trough, the copper wire can be
changed over from the first electrode vessel and a reading of that taken.
While the equilibrium of the second electrode is being checked, the first can

the filling apparatus

G. 8S. WALPOLE

413
be refilled. It is removed, dipped into a large beaker of distilled water, and

dried externally with a clean towel or filter paper and connected again to

“i GF
Gi
es

“ELA —~|

Fig. 1

Opening the tap D (Fig. 1) and pushing home the syringe, the electrode
vessel is emptied, and the tap on the filling apparatus opened for a few
seconds to fill it with fresh .hydrogen

The next sample having been brought
underneath the electrode vessel, the tap on the three-way piece is closed

414 G. S. WALPOLE

liquid drawn up by the syringe, expelled so as to get rid of traces of the
previous sample, and then the vessel filled with a fresh sample as before.

If an electrode is properly blacked, and in good working order, it should
give the equilibrium value immediately and scarcely change afterwards—two
millivolts at most—during the next ten or twenty minutes. Diffusion’ of
hydrogen into and through the rubber stopper causes slight inconvenience
if the electrode has not been used for a day or two. The liquid rises in the
vessel and the platinum becomes immersed beyond the point. Equilibrium
under these conditions would not be reached for a long time. To rectify
this, the glass tube A can be slid upwards through the rubber stopper. It
is advisable when setting the electrode aside, to push the glass tube holding
the electrode well through the cork, fill with hydrogen, suck up distilled
water till the platinum is immersed and after disconnection from the filling
apparatus leave the vessel immersed in distilled water.

Washing out of electrode vessel. One rinsing only between two samples
has been found sufficient. Several experiments were performed to demonstrate
this. One set of figures is given below.

Temp. 15°. Bar. 750 mm. Half-electrode sat. KCl-calomel. Connecting fluid sat. KCl.

Calculated Py value for 0-0993 NHCI (dissociation 91:6 °/)=1:041. Calculated Py value
0-1 N NaOH (14:14 - 1-075) =13-065. Sample used was known not to be absolutely free from CO,.

Solution Volume Time of Potential pta™— 0251
used used Remarks observation tg Hy O0507
0:0993 N HCl 2 ce. Apparatus at commence- 8.45 0-3110 1-039
ment clean and dry 8.50 0°3110 1-039
9.10 0-3113 1:045
3 4 ec. © Used 2 ce. for rinsing out 1.45 0:3113 1:045
last sample 2.00 0°3115 1:048
2.15 0-3120 1:057
2.45 0-3120 1:057
0-1 N NaOH 4 ee. Used 2 cc. for rinsing out 2.50 1-0020 13°015
last sample 2.55 1:0022 13°019
3.15 1:0022 13-019
3.50 1:0030 13-032
0:0393 HCl 4 ce. Used 2 ec. for rinsing out 4.19 0°3115 1:048
last sample
Sy cire. 20 ee. Rinsed out many times and 4.45 0°3105 1:031
filled side capillary, see 4.55 0°3110 1-039
page 425. Passed bubble 5.00 0°3110 1-039
for 5 mins. next morning
8.5 0°3108 1:036

Temperature. Although most of these determinations have been made at
18° exactly, it is unnecessary to pay too great attention to this point as long
as the temperature of the calomel-saturated KCl half-electrode and the gas

Pi Po EP

G. S. WALPOLE 415

electrode are the same in any one determination. Calculation from the
constants given by Michaelis for this combination demonstrates that the

error in measurement of the value of Py per 1° is 0:007 for a neutral solution
the H’ ion concentration of which has no temperature coefficient. The error
with a N/10 KCl-calomel electrode is about three times this. Approximate
determinations with the saturated KCl half-electrode may quite well be
made on the open bench without a constant temperature bath.

When a N/10 KCl-calomel half-electrode or a N/10 HCl-hydrogen half-
electrode are used in combination with the gas electrode, more careful
temperature adjustment is necessary. As a direct result of the diminutive
size of the electrode vessel, and the small quantity of fluid used therein, the
time taken to arrive at temperature equilibrium is very short.

Quantity of material used. The size of electrode vessel found most
convenient for general use has the following principal dimensions; length of
tube B (Fig. 1) 75 mm., maximum external diameter 10°5 mm., length of lower
capillary 25 mm. The capacity when full is 15 ce., so that a determination
can be performed easily on 2 cc. of the material if the electrode vessel be
clean and dry to start with. When the electrode vessel is not cleaned and
dried between two determinations, 4 cc. are required—2 ce. to rinse out the
vessel and 2 cc. from which the electrode vessel is filled.

If only smaller quantities of material are available, a smaller electrode
vessel may be used. Quite good results were obtained by one holding
03 ce. This is not necessary, however, for the quantity of material available
may be drawn up into a vessel of the most convenient size and then saturated
potassium chloride drawn up afterwards until contact is made with the
platinum point. Mixing of the two fluids only takes place very slowly by
diffusion unless the density of the fluid examined approximates to that of
saturated KCl. In those cases when it is known that the H’ ion concentra-
tion does not change by such treatment the sample may be diluted with
pure water [cf. Michaelis and Davidhoff, 1912]. This is the substitution of
an indirect method for a direct one and so an additional possibility of error
is introduced.

The extent of the contamination of the fluid during a determination is
very small since the boundary surface between it and the saturated potassium
chloride solution is only that of a section of the capillary tube. When a very
small glass stopper is fitted at the bottom of the capillary, possibility of
contamination is still further removed. It is usually found that enough
potassium chloride diffuses round the stopper to diminish the resistance
sufficiently for a reading, not always very sharp, to be taken.

416 G. S. WALPOLE

In many cases, therefore, where a small quantity of liquid only is available,
it is possible to determine its reaction electrometrically and recover it with
very little contamination, dilution, or loss, for further experiments. Usually
a moderate economy of material is advantageous and, on this and other
grounds, the use of the vertical electrode filled by suction has been found
by me much mofe convenient than the V-type for general work. . Frothing
fluids do not exhibit their disagreeable characteristic, as hydrogen is not
bubbled through them, but they are forced up by external atmospheric
pressure. There is no need, therefore, to make an indefinite contact with
the froth. The abolition of tapes or wool threads soaked in potassium
chloride solution was found a great convenience.

Diffusion Potential. In the electrolytic cells of which one half-electrode
is hydrogen in contact with an aqueous solution there are really four
differences of potential involved. There are the two electrode potentials
proper and two differences of potential where the electrode solutions come
into contact with the connecting solution. The latter-differences of potential
are for the most part dependent upon the natures of the electrode solutions
and of the connecting solution, but the time the solutions have been in
contact is a concomitant factor. Cumming and Gilchrist [1913] have
recently investigated this quantitatively and find that this “time change”
is more marked with capillaries than open tubes. They, therefore, reeom-
mend that capillaries should be avoided in the construction of an electrolytic
cell; so also should membranes, cotton wool plugs and, presumably, tapes.

In the vertical electrode vessel filled by suction, the surface between the
solution examined and the connecting fluid is in a capillary or at the end of
one. It may readily be made to occur at the wider part of the tube (p. 415,
paragraph 4). I have made a number of careful measurements with this
cell to determine how far variations in potential difference can be traced to
boundary changes, and to what extent these alter when the surface between
the electrode solution and the connecting solution is in the broader part of
the tube instead of at the end of the capillary.

Using N HCl and N/10 HCl as electrode solutions connected to a
saturated KCl-calomel half-electrode by saturated KCl solution, I could
not detect with certainty any difference in the time change over half an
hour in each case—first with boundary at end of capillary, second with
boundary in broader part of the tube. Over longer periods, differences in
the “time changes” could probably be detected, but do not rightly enter
into consideration in the ordinary use of the electrode.

I have sometimes used saturated ammonium nitrate [Cumming, 1907]

G. 5S: WALPOLE 417

instead of saturated potassium chloride as connecting fluid when examining
acid solutions. With the N/10 KCl-calomel half-clectrode N and N/10 HCl
gave the potentials 0°3390 and 03974 respectively ; calculated values 0°3407,
03977. With the saturated KCl-calomel electrode, however, the results
were about 18 millivolts too low, the contact potential between saturated
potassium chloride solution and saturated ammonium nitrate being of that
order.

Comparison between the results obtained with the vertical and the
V-electrode vessels,

Both electrode vessels were fitted with rubber stoppers through which passed the glass tube
in the end of which the blackened platinum point was mounted. When necessary this was
adjusted so as just to touch the surface of the liquid. Temp. 18°. Connecting solution saturated
KCl solution,

Potential readings

SE
V-elec- Vertical
trode electrode Previous

Date Solution Half-electrode vessel vessel results
18 April 0-01 Na,CO, Aq. Calomel. Saturated KCl 0°8770 0-8780 —
11 April Sodium Citrate 5 “3 3 0-6240 0°6235 0°6238
(Sorensen) (Sérensen)
6 May Standard Acetate 3 <5 5 0°5165 0°5170 0°5175
solution (Michaelis) (Michaelis)
7 May 0-1N HCl FP s A 0°3122 0°3125 --
8 May 0-1N HCl Fs or 0°3122 0°3125 —
15 May N HCl % O-LN KCl 0°3435 0°3425 os
15 May N HCl =3 Saturated KCl 0°2558 0°2560 —

ELECTROMETRIC TITRATION.

In acidimetry and its various applications the addition of a standard
alkaline solution is continued until the titrated fluid has a certain H’ con-
centration previously agreed upon. This is called the “end point.” Generally
the “titre” or amount of standard alkali which must be added to a certain
amount of the fluid to be titrated in order to arrive at the “end point :
is all the information required. The observer, with a proper understanding
of the nature of the reaction, knows the correct “end point” for the object
he has in view, and chooses an indicator which will tell him by its colour
change when the “end point” is reached.

For instance, in the titration of an acetic acid solution it is desired to
add standard NaOH solution until all the acetic acid is converied to sodium
acetate. The hydrogen-ion concentration of sodium acetate solution is about
Pj;=7°5 to 95 so that phenolphthalein or another indicator exhibiting its
colour change over this range of reaction is employed.

418 G. 8S. WALPOLE

Again, in the Kjeldhal titration N/10 NaOH is added to a very
dilute solution of sulphuric acid containing some ammonium sulphate.
The “end point” is reached when all the sulphuric acid is converted to
ammonium sulphate and sodium sulphate. Ammonium sulphate is faintly
acid so that a trace of soda in excess will bring the mixture nearly neutral.
Hence an indicator is chosen which is very sensitive at the neutral point—
methyl orange, alizarin, rosolic acid have all been recommended. Such
simple titration processes depend for their accuracy and reliability on a sudden
change taking place in the reaction of the titrated fluid at some point during
the addition of the alkaline solution, and the correct*choice of an indicator
which demonstrates this change. When dealing with very feebly dissociated
acids and bases, of which proteins and some of their products of hydrolysis
are excellent examples, such sudden changes of reaction do not occur. The
titration process becomes more difficult. Moreover at any time it gives no
information as to the rate of change of concentration of H’ ions taking place
progressively step by step as the alkaline solution is added. The measurement
of this is important, and its neglect is frequently the outcome of the fact
that it is a laborious process. To do it colorimetrically would be tedious and,
in small quantities of fluid, impossible. Electrometrically with an electrode
vessel of the standard type and a tape it would be impracticable, since loss
and contamination of the fluid by potassium chloride would inevitably occur.
With a vertical electrode vessel filled by suction these measurements of
change of H’ ion concentration during the titration process have been found
quite practicable. There is no loss, practically no contamination by potassium
chloride, and each determination of H’ ion concentration during the titration
process takes about two minutes.

Apparatus. Fig. 2 is a scale drawing of the apparatus used for titrating
10 cc. quantities. A small beaker contains the fluid to be titrated and
dipping into it is a small electrode vessel (B) of capacity 3 cc. and a capillary
tube filled with saturated KCl solution, connected to the saturated KCl-
calomel electrode. Passage of liquid bodily along the capillary is prevented
by the insertion of a long tightly packed plug of cotton wool, and if necessary
a clip on the rubber tube connecting it to the half-electrode. The con-
ductivity of the potassium chloride is so good that readings may still be
taken with the clip on and sometimes, if the tap on the half-electrode be not
vaselined, with this tap shut. A standard burette is clamped above the
beaker. Two wires from the potentiometer lead to the half element and the
electrode vessel respectively.

Technique. The electrode vessel is filled with hydrogen, and, while the

Hydrogen

AH

an

VI

i

Fig. 2:

420 G. S. WALPOLE

gas is still passing, its lower end brought below the surface of the measured
quantity of liquid to be titrated which has been introduced into the beaker.
The tap on the filling apparatus is turned off and by withdrawing the
plunger, liquid is sucked up until contact is made. A reading of the
potential indicates the reaction of the original liquid. The plunger is pushed
home, a few drops of the titrating solution in the burette are run in, and
by imparting a rotary motion to the beaker its contents are mixed until
uniform throughout, and the reaction determined as before. In practice it
is found best to draw the liquid up into the side tube and push it out several
times in order to rinse the platinum point well between each reading,

From the results a curve may be plotted showing the relationship
between the quantity of fluid added and the reaction. If any doubt exists
as to the accuracy of the results obtained in this way, one or two points on
the curve can be checked by single additions of N/10 alkali followed by
determinations in a larger electrode vessel.

Eaample 1. 10 ec. of N/10 glycocoll solution containing in 1 litre
7505 g. glycocoll and 5°85 g. sodium chloride [Sdrensen, 1909, 1] were
titrated with N/10 HCl electrometrically. Two separate titrations A and B
were performed and subsequently the amount of potassium chloride acci-
dentally introduced was found to be equivalent to 3:3 cc. of decinormal
solution in one case and 3:0 cc. in the other, though no clip was used and
the tap was open. The 10 determinations in titration B took 13 minutes;
the 21 determinations in titration A took 43 minutes, though no particular

attempt was made at speed.

Electrometric titrations of 10 cc. glycocoll solution with N/10 HCI.

Connecting fluid—saturated KCl solution. Half-electrode—saturated KCl-calomel. No
correction for diffusion potential.

Vol. N/10 HCl K.M.F. in H.M.F. in Calculated from
added titration 4 titration B Sorensen’s results
1 ce. 0°446, 0°445, 0°445 0-443 0:446
2 0-426, 0°426, 0°426 0°425 0°426
3 0°413, 0°414, 0-414 0:412 0:413
4 0:404, 0°404 0-402 0-403
5 0°394, 0°394 0°393 0-394
6 0:387, 0°387 0°386 0°386
if 0-381 0°379 0°379
8 0°374 0:°373 0°373
9 0°369 0°367 0°367
10 0°365, 0°365, 0°365 0-362 0°362

Eaample 2. Advantage was taken of the opportunity now afforded of
following the changes of reaction during a Sérensen formaldehyde titration.

tier!

G. 8S. WALPOLE 421

A 4°/, solution of Witte peptone was employed and the titration performed
with every care. First colorimetrically, compensating for the colour of the
peptone solution by a special tintometer, and then electrometrically.

1, SORENSEN TITRATION OF 4°/, WITTE PEPTONE SOLUTION
COLORIMETRICALLY,

Stage (1). The neutralisation of the peptone solution. Neutral red solution
was used as indicator, and as a standard of neutrality a mixture of N/1L5
phosphate solutions (84°4 ce. acid + 65°6 cc. alkaline). 10 ce. of this solution
—its reaction does not change on dilution [Szili, 1904J—to which 1 ce. of
neutral red solution was added, were placed in a cylindrical vessel with a flat
bottom D. A similar vessel surmounting this contained 10 ce. of Witte
peptone B. Two corresponding vessels, A and C, contained distilled water,
and 10 ce. Witte peptone solution plus 1 ce. neutral red solution respectively.
Looking down AC the tint appeared redder than that seen looking down BD,
and successive additions of N/10 NaOH solution were made until 0°5 cc. had
been added, and the two columns were seen to match. It was considered
that the 10 cc. of Witte peptone solution were neutralised by the addition
of 0°5 cc. N/10 NaOH. Unfortunately the colorimetric result is misleading

here. The Pj, value is 7°66 and not 7:07. The errors in Pj, found by
Soérensen [1909, 2] in similar determinations with 2 per cent. Witte peptone
solution containing 0°'l N NaCl and correcting for the colour of the solution
by 2 drops of Bismarck brown and 2 drops of helianthin II were 0°18 and
0:12 in the same direction in two cases. Rosolic acid gave a similar result.

Stage (2). Titration in presence of formaldehyde. The problem is to
discover how much 071 N NaOH must be added to a mixture of 10 cc. of
Witte peptone solution and 10 cc. of neutral 40 per cent. formaldehyde
solution in order that the resulting mixture should have the reaction
P= 8°68 (7 borate + 3 HCl); what further addition must be made to bring
Pi, value to 891 (8 borate +2 HCl); and what still further addition will
bring the value to P}, = 9:09 (9 borate + 1 HCl).

Using the same apparatus the solutions were arranged as shown on p. 422.

Formaldehyde bleaches slightly the colour of Witte peptone so that its
colour must be compensated for by a solution similarly bleached. 1 ce.
saturated sodium chloride solution is added to each formaldehyde-peptone
mixture to inhibit the formation of polymethylimino-compounds. The same
quantity was added to the 20 ce. of Sérensen mixture in D in order that the
“neutral salt effect” on the phenolphthalein should be of the same order in

Bioch, vu 28

422 G. S. WALPOLE

A. B.
10 ce. Witte peptone solution. 10 ce. Witte peptone solution.
10 ce. formaldehyde containing 1:0 cc. of 10 ce. formaldehyde solution.
0-1 °/y) phenolphthalein in 50 °/9 alcohol. 1 ce. saturated NaCl solution.

1 cc. saturated NaCl solution.
(4:8 cc. 0-1 N NaOH were added to this tube
to make a match.)

C. D.

20 cc. distilled water. 20 ce. Sdrensen mixture.
(7 borate +3 HCl.)
1-0 ce. of 0:1 °%/9 phenolphthalein in 50 °/)
alcoliol.
1 cv. saturated NaCl solution.

A and D. The amount of 0:1 N NaOH added to“A before a match resulted
was 48 cc. D was now removed and replaced by D’ containing 20 ce.
(8 borate +2 HCl) mixture, 1-0 cc. of 0-1 °/, phenolphthalein in 50 °/, alcohol,
and 1 ce. saturated sodium chloride solution. A further 0°3 cc. of 0°l N NaOH
was required in A to make a match. D’ was now replaced by D” containing 20 ce.
(9 borate + 1 HCl) mixture, 1:0 cc. of 071 °/, phenolphthalei in 50 °/, alcohol,
1 ce. saturated NaCl. A further 0°3 cc. of 0:11 N NaOH was required to
make a match. Hence by a colorimetric method it has been found that the
following mixtures have the corresponding H’ ion concentrations given in
the table:

Pa Composition of mixture

7:07 10 cc. Witte peptone solution + 0°5 ce. 0:1 N NaOH

8°68 ae Py * +10 ce. neutral 40 °/, formaldehyde sol. + 4-8 0°1N NaOH
8-91 * xp > + Py) “ 6 +5°1 5

9-09 Of 5 5 + 5 Ap a +5°4 3

By a simple artifice the formation of polymethylimino-compounds in the
above titrations may be prevented without the addition of sodium chloride
solution. If, before adding the formaldehyde to the peptone solution some
0'1 N NaOH be added, then no cloudiness or precipitate forms. Repeating
the above titrations in this manner and adding the solutions in the order
named, the following results were obtained.

Pu
8:68 10 cc, Witte peptone soln.+4:0 cc. 0°:1N NaOH +10 ce. neutral 40 °/, formaldehyde +
0°80 cc. 0-1 N NaOH

8°91 os a ie Bee 55 +10 cc. neutral 40 °/, formaldehyde +
1:10 cc. 0'1N NaOH
9-09 » ” » SF gn zs +10 ce. neutral 40 °/) formaldehyde +

1:40 ec. 0-1 N NaOH
In Fig. 3, the relation between the compositions of these mixtures and
their Pj, values is represented diagrammatically.
The points obtained using phenolphthalein are marked +. The neutral
red point is marked +f.

G. S. WALPOLE 423

Millivolts against
Sat. KCl half-electrode

Negative Exponent

78 er) Pi

-76

*74

Sy.

“70

64

formaldehyde addition

"62

drop of Pit value due to

‘60

) Ws hs a2 3 4 5 6
ce. N/10 NaOH added to 10 cc. Witte peptone.

Fig. 3. Sorensen formaldehyde titration of 4 °/, Witte peptone performed electrometrically.

Note.—At C 10 cc. neutral formalin was added.
© are points obtained in the titration process—electrode containing 0°3 cc.
O are points obtained with large electrodes in separate determinations.
+ are points obtained colorimetrically using phenolphthalein.
++ is point obtained colorimetrically using neutral red.

28—2

424 G. S. WALPOLE

2. SORENSEN TITRATION OF 4°/, WITTE PEPTONE SOLUTION
ELECTROMETRICALLY.

10 ce. of peptone solution were taken and the same small electrode vessel
as before. Connecting fluid and fluid in calomel half-electrode were saturated
KCl solution; room temperature about 17°. As the alkaline formaldehyde
solution becomes less alkaline on prolonged exposure to the air it 1s necessary
to work with reasonable rapidity. After the addition of 1°5 ce. of 0-1 N NaOH
solution the potential reading was found to be 0°7350; 10 ce. of neutral
formaldehyde were then added and the potential immediately fell to 0°5615.
Further additions of decinormal alkali were made

and corresponding potential
readings taken—until after adding 4°5 ce. in all, the change of potential was so
great, on account of the addition of the last 0°5 cc., that the ordinary rinsing
out of the apparatus between two determinations was evidently not sufficient
for this case, and the value 0°6800 is probably too low, the point being off
the curve.

EXPERIMENTS WITH SOLUTIONS CONTAINING DISSOLVED CARBON DIOXIDE.

The difficulties of H’ determinations of fluids containing carbon dioxide
have been so thoroughly dealt with by Michaelis and Rona [1909], Hassel-
balch [1910], and Michaelis and Davidhoff [1912] that their repetition here
is unnecessary. It is sufficient to remark that before a final unchanging |
potential reading can be obtained the hydrogen atmosphere, the fluid, and
the platinum electrode must all be in equilibrium, no further gas exchange
taking place between them. This refers to hydrogen, carbon dioxide, and
oxygen. If the hydrogen be not at 760 mm. a correction must be applied
to the formula expressing the relation between the potential of the half
element and the hydrogen ion concentration of the fluid.

Generally speaking, electrode vessels may be divided into two classes,
those in which hydrogen is passed in turn through a small portion of the
fluid examined, and then through a portion of the same fluid in the cell
itself until equilibrium is attained, and those in which one portion of
hydrogen only is used. The methods are referred to respectively as those
of the “moving” and “still” hydrogen atmosphere.

In electrode vessels in which a “still” atmosphere is employed it may be
considered that there are three degrees of accuracy with which the H
concentration of a fluid containing carbon dioxide may be determined.

Firstly, the bubble of hydrogen may be brought to the surface of the
fluid examined, the blackened platinum point saturated with hydrogen

G. S. WALPOLE 425

adjusted so that it is just in contact with this surface, and the potential
measured. The form of electrode used may be the V-electrode or the
vertical electrode described. The potential will not represent accurately
the reaction of the fluid. Owing to the diffusion of carbon dioxide from
the fluid into the hydrogen the surface layers will rapidly become more
alkaline than the solution was originally. The potential reading will there-
fore be high, falling slowly as equilibrium is established and only reaching
a constant value some hours afterwards when that end has been attained.

Secondly, the small bubble of hydrogen may be passed backwards and
forwards through the fluid in the cell a few hundred times before a reading
is taken. By this means equilibrium is established, and it will be found that
the value is constant and nearly correct. The electrode used may be a
V-electrode manipulated in accordance with the instructions given by
Michaelis [1912], or the vertical type filled by suction, or the Hasselbalch
electrode. It will be seen that, though equilibrium is established between
the hydrogen atmosphere and the fluid, the fluid has given up some carbon
dioxide to the hydrogen and has therefore become more alkaline than it was
originally and it is not the H’ ion concentration of the original solution that
has been measured.

This brings us to the third step where after equilibrium is reached the
hydrogen bubble is retained, but the fluid in the cell replaced by a fresh
volume with which the hydrogen is again brought into equilibrium. This
process may of course be repeated until a definite final potential reading
is observed which will then represent the true reaction of the fluid containing
carbon dioxide.

The only electrode vessel described which permits of this is the Hassel-
balch electrode vessel. Michaelis used it to control the results obtained by
him using the V-electrode and passing the bubble up and down many times
to obtain equilibrium—the second case above.

Since finding that the same thing can be done quite simply with the
same electrode vessel as that described for gas-free solutions, I have made
a number of determinations of the reactions of carbonate solutions by its
means and have obtained consistent results.

Technique. If the electrode vessel be tipped sideways when filling, the
liquid drawn up may be made to enter the side capillary leaving a bubble
of hydrogen. After closing the tap the apparatus may be taken in the hand
and by a slight movement at the wrist the bubble made to pass from one
end of the vessel to the other as many times as are necessary to obtain

equilibrium,

426 G. S. WALPOLE

Bringing the vessel into a vertical position again the lower end is dipped
below the surface of the fluid in the beaker, and the glass tube moved up
or down through the rubber stopper until the platinum point just touches
the surface of the column of fluid standing up in the electrode vessel. This
may now be wiped dry externally, placed in the trough, and a reading of the
potential taken. The value obtained corresponds exactly to that obtained by
the V-tube used in the manner advocated by Michaelis. When small
quantities only of carbon dioxide are present, this value will be very nearly
correct. It may be checked by attaching a rubber tube to the tube at D
(Fig. 1), introducing the lower end of the vessel in a sloping position into the
beaker again, aspirating at D, and opening the tap gently. In this way fresh
solution is drawn into B in the place of the old solution which passes out
through D without disturbing the hydrogen bubble. This fresh quantity of
solution is now brought into equilibrium with the bubble of hydrogen as
before. The process may be repeated indefinitely and the result obtained
is of the third order of accuracy—previously only obtained by the Hasselbalch
electrode. The pattern of electrode vessel having a ground glass joint cannot
be used for these operations as the height of the platinum point in the vessel
is not then adjustable. It is essential that the platinum point shall only
just touch the surface of the fluid when a reading is to be taken, otherwise

equilibrium between the electrode and the solution, instead of taking only >

a minute or two, may take hours. In these cases, where the investigated
liquid contains carbon dioxide, with experiments lasting over a number of
hours, the rubber jot has a further disadvantage over and above that
already mentioned. Carbon dioxide, like hydrogen, permeates rubber, and
though the rubber joint is immersed in potassium chloride solution, transpira-
tion of gases through the rubber is not prevented, and this slow transpiration
is detrimental to accurate work when experiments last several hours.
Carbonate solutions. In order to check the results obtained when using
the vertical electrode filled by suction for fluids containing carbon dioxide,
I have determined the hydrogen ion concentration of mixtures of a sodium
carbonate solution and dilute hydrochloric acid. For each determination
12°5 cc. of 0°2 N sodium carbonate were taken and diluted to nearly 100 ce.
Then a measured quantity of 0-1 N hydrochloric acid was added and the
volume made to 100 cc. exactly. No barometric correction has been applied
to the determinations for diminished hydrogen pressure due to the carbon
dioxide present. Neither have any steps been taken beyond the use of
saturated potassium chloride as connecting solution to correct for diffusion
potential. The results are plotted on a curve (Fig. 4). Abscissae represent

G. S. WALPOLE 427

ce. of 01 N HCl taken: while ordinates are proportionate to the Pj values
less a constant. It will be seen that a mixture of 12°5 ec. of 0°2 N sodium
carbonate + 12°5 ce. 0:1 N HCl diluted to 100 cc. corresponds to a solution
of 0°0125 molecular NaHCO, which is also 0:0125 N with respect to
sodium chloride. At this point the potential is 0°7200 against the calomel
saturated-KCl half-electrode corresponding to Pj; = 813. At this reaction
phenolphthalein gives a pale pink colour, thus confirming the propriety of
the analytical device of titrating carbonates in the presence of caustic
alkali using phenolphthalein and methyl orange.

Potentiometer readings
against Calomel
Sat. KCl electrode
in volts

+
H values
11g

10

re
fo.)

© QTc
as = O\S
c eqs
[= ESS 26 \5
Ss =s2s= oo ln
— =

= ron Pal
= 0.20 SL
oh NON )
A ro]
(@) 5 VOY 1255 5 90 95

ce. 0:1 N HCl added to 12°5 ce. 0-2 N Na,COsz solution.

Fig. 4.

428 G. S. WALPOLE

Mixtures of 12°5 cc. 0:2 N Na,CO, and varying quantities of 0-1 N HCl.
diluted to 250 ce.

Connecting fluid sat. KCl. Half-electrode calomel-sat. KCl.

Vol. of pe Ds eel
0-1N HCl Potential x #00577

0 0:883 10:95

5 0-841 10°22
10 0-798 9:48
it 0-782 9-20
12 0°755 8°73
12°3 0-738 8°44
12°5 0°720 8:13
12°7 0-701 7:80
13 0-684 7°50
14 0°652 - 6°95
15 0:638 6-71
18 0618 6:36
20 0:598 6:01
23 0570 5:53
24 0°558 5 +32

REFERENCES.

Bjerrum (1910), Zeitsch. physikal. Chent. 73, 724.
Cumming (1907), Trans. Faraday Soe. 11, 1.

— and Gilchrist (1913), Trans. Faraday Society.
Dolezalek (1899), Zeitsch. Elektrochem. 5, 533.
Hasselbalch (1910), Biochem. Zeitsch. 30, 317.
Michaelis and Rona (1909), Biochem. Zeitsch. 18, 317.
‘and Davidhoff (1912), Biochem. Zeitsch. 46, 131.
Sorensen (1909, 1), Biochem. Zeitsch, 21, 167.

(1909, 2), Biochem. Zeitsch. 21, 245.

Szili (1904), Zeitsch. Hlektrochem. 113.

Wilsmore (1900), Zeitsch. physikal. Chem. 35, 291.

XLII. SOME ESTERS OF PALMITIC ACID.
By MARJORY STEPHENSON.

From the Ludwig Mond Research Laboratory for Biological Chemistry,
Institute of Physiology, University College, London.

(Recewed June 14th, 1913.)

The physiological importance and commercial value of the fats early
attracted chemists to the study of their properties and composition. The
interest which they then aroused has been fully maintained with the result
that the present extensive literature on the compounds of the higher fatty
acids deals almost entirely with the esters of glycerol to the exclusion of
those of other polyhydric alcohols.

The chemistry of the fats may be said to have had its birth in the work
of Chevreul [1823] to whose researches in the first quarter of the nineteenth
century we owe the proof that the fats occurring in the animal body are
triglycerides of higher fatty acids chiefly of stearic, palmitic and _ oleic.
About forty years later Berthelot [1860] first succeeded in synthesising the
mono-, di-, and tri-glycerides of many fatty acids and in identifying the
tri-olein, tri-palmitin and tri-stearin so obtained with naturally occurring
animal fats.

The method of synthesis adopted by Berthelot consisted in heating
together glycerol and the fatty acid in a sealed tube to 200°-250°. This direct
method of synthesis with various slight modifications has been largely used
by later workers and still enjoys a wider application than the yields which
are so obtained appear to warrant. Modifications of this method are chiefly
directed towards eliminating the water formed during the reaction and thus
enabling the process to be carried out at a lower temperature and also
increasing the yield. A modification of this kind was introduced by Scheij
[1899] and extended by Belucci and Manzetti [1911]. These workers found
that the yields obtained by Berthelot’s method could be greatly increased by
carrying out the reaction under diminished pressure; thus Belucci and
Manzetti synthesised triolein by heating together glycerol and oleic acid to
260° under a pressure of 20 mm., and state that in this way a yield of 95-98 °/,

430 M. STEPHENSON

was obtained; they afterwards found that the reaction could be carried out
equally well under atmospheric pressure if a current of carbon dioxide were
passed through the flask. Griin [1905] carried out the synthesis of digly-
cerides by warming the fatty acid and glycerol in concentrated sulphuric
acid; thus a-a-distearin was obtained from glycerol and stearic acid, glycerol-
disulphonic acid being formed as an intermediate product. None but
primary alcohol groups are esterified in this reaction. Bloor [1910, 1912]
applied this method to the condensation of mannitol with stearic and lauric
acids; he identified his products as mannitol distearate and mannitol dilaurate
respectively. .

The halogen derivatives of glycerol have also been used in the prepara-
tion of glycerides. Partheil and Von Velsen [1900] prepared trilaurin and
tripalmitin by heating the silver salts of these acids and tribromohydrin to
140° to 150° in the presence of xylene. A similar method was used by Guth
[1903]; he prepared monostearin by heating together sodium stearate and
monochlorohydrin, also a-distearin from a-dichlorohydrin, and §-distearin
from -dibromohydrin; tristearin was obtained from tribromohydrin and
excess of sodium stearate. The yields obtained by these last two methods
are not given.

Since the esters of the lower fatty acids are so readily prepared by the
action of the acid chloride on the alcohol it seemed probable that this method
might also be applied to the synthesis of esters of the higher fatty acids, not
only with glycerol but also with other polyhydric alcohols. This method has
been used recently by Griin and Schreyer [1912] who prepared a-myristo-
a-chlorohydrin by the action of myristyl chloride on a-monochlorohydrin, and
8-myristo-a-a-dichlorohydrin from myristyl chloride on a-a-dichlorohydrin.
This is the only example of the use of this method which the writer has been
able to find, though the wide use of the method in other fields makes 1t seem
possible that some cases have been overlooked.

In order to test the efficiency of the above method for the synthesis of
fatty bodies, palmityl chloride has been exclusively used and was prepared by
the method described by Marie [1896] for melissyl and cerotyl chlorides. Its
action was tried on ethylene glycol, glycerol, mannitol and glucose. In the
case of glycerol esterification was effected by heating the two substances
together to 120°. The action however was slow and this method was
abandoned in favour of allowing the substances to react in the presence of
pyridine with chloroform as a solvent. The action then took place in the cold

and was completed by gently warming on the water bath.

lye ag

LA Lp CALA ie Ne Bi

M. STEPHENSON 431

EXPERIMENTAL.

Preparation of palmityl chloride. 45 g. of palmitic acid were mixed with
40 g. of phosphorus pentachloride (calculated amount 38 g.) in a distillation
tlask and the action started by warming. When the action was complete the
excess of phosphorus pentachloride and the phosphorus oxychloride were
distilled in vacuo on a boiling water bath; when these substances were
removed the temperature was raised; the palmityl chloride distilled over as a
colourless oil at 198°-200° under a pressure of 15 mm. Yield obtained 37 g.
C09.)

Preparation of glycol dipalmitate. 1 g. of dry glycol and 5 g. of pyridine
were added to 12'8 g. of palmityl chloride (4 g. in excess of that required for
complete esterification); about 50 cc. of dry chloroform were then added to
dissolve the solid compound of pyridine and palmityl chloride formed; the
contents of the flask turned a deep orange colour. The flask was closed and
allowed to stand for 24 hours, it was then gently warmed on the water bath
and well shaken; the chloroform was evaporated off and dilute sulphuric acid
added to liberate the pyridine from its combination with palmity] chloride;
the insoluble ester and palmitic acid (from excess of reagent) were filtered
off and washed free from sulphuric acid and pyridine on a hot water filter;
they were then dissolved in hot absolute alcohol and the free palmitic acid
neutralised with alcoholic potash; the solution was'evaporated to dryness and
the residue transferred to a Soxhlet apparatus and extracted with ether. The
ester crystallised from the ethereal solution on cooling in clusters of fine
needles.

Yield before recrystailisation 7°5 g. (94 °/,). The substance so obtained
was brownish yellow in colour and was soluble in hot and cold alcohol and
also in chloroform, ether, ethyl acetate and the usual solvents for fats. It
was recrystallised from a mixture of chloroform and alcohol since it was found
that by this method the crystals obtained were white whereas when other
solvents were used the colouring matter was thrown down with the crystals.
The crystals so obtained were deposited in rosettes of fine needles ; after four
recrystallisations the substance had a constant m.p. of 65° (corrected). It
was dried to constant weight in vacuo over sulphuric acid; it still, however,
contained traces of moisture, which were only finally removed by heating in
a toluene bath to 106°. The substance was identified by the results of
elementary analysis and by the percentage of palmitic acid estimated by
Hehner’s method. As these values calculated for glycol monopalmitate,

432 M. STEPHENSON

glycol dipalmitate and for palmitic acid approach somewhat nearly to each
other, these figures are appended for purposes of comparison.

Jo Palmitic

%%) Hydrogen %y Carbon acid
Calculated for glycol monopalmitate 12-00 72:00 85:3
% 5, glycol dipalmitate 12°27 75°84 95°2

35 ;, palmitic acid 12°5 75°01 100

Analysis :
(1) 0°2195 g.; 0:2406 g. H,O; 0:°6108 g. COg.
(2) 0:2598 g.; 0:2913 g.H,O; 0:7227 g. COp.
(3) 1°8598 g.; 1°7750 g. palmitic acid.

%y Palmitic

°) Hydrogen 9/9 Carbon acid

Analysis (1) 12°44 75:9 --
A (2) 12°45 75°84 —

» (3) = = 95°5

The compound may therefore be identified as glycol dipalmitate corre-
sponding to the formula

CH,00CC,;H.

CH,00CC,;H31.

Glycerol tripalmitate (tripalmitin). 1:2 g. of glycerol and 9 g. of pyridine
were added to 15'8 g. of palmityl chloride (4°7 g. in excess of that required
for complete esterification) ; 30 cc. of chloroform were added. The procedure
followed was the same as that described for the glycol compound. In this
case the ester obtained from the ether extract was brown in colour and much
difficulty was experienced in obtaining white crystals. After two recrystal-
lisations from chloroform and alcohol followed by four from ether the substance,
though melting sharply at 62°, was still biscuit coloured. It was then
dissolved in wet ether and boiled for six hours with charcoal. The substance
which crystallised from the ether after filtration was quite white and was
deposited in fine needles; m.p. 62° (corrected). It was sparingly soluble in
hot alcohol and soluble in ether, chloroform and all the usual solvents for fats.
Traces of charcoal clung to the substance with great obstinacy and could not
be removed by the filtration of the ethereal solution. They were finally
removed by filtering the substance itself in the steam oven through a very
small filter. The substance was dried at 110°, and identified by the results
of elementary analysis; the corresponding values calculated for di- and tri-

palmitin and for palmitic acid are appended for comparison.

M. STEPHENSON 433

%/) Hydrogen %/y Carbon
Calculated for dipalmitin 11°90 74:06
o3 ,, tripalmitin 11°91 75°95
a », palmitic acid 12°5 75°01
Analysis : (1) 0°2055 g.; 0°2222 g. H.O; 0°5728 g. COo.
(2) 0°1945%.; 0°2155 g. HO; 0°5405 g. COo.
%9 Hydrogen 9/) Carbon
Analysis (1) 12°01 76°00
53 (2) 12°30 75°78

The substance was therefore identified as tripalmitin.

Owing to the difficulty experienced in purification sufficient substance
was not available for a determination of palmitic acid.

Yield before recrystallisation 5:5 g. (50 °/,).

Yield of pure substance 2 g.

Manmitol hexapalmitate. 1:5 g. of mannitol and 8 g. of pyridine were
added to 19:2 g. of palmityl chloride (4°8 g. in excess of that required for
complete esterification). The procedure followed was identical with that
described for glycol. The total yield of ester obtained after extraction in the
Soxhlet apparatus and before further purification was 11°2 g. (78°/,). The
substance obtained was almost insoluble in cold alcohol, and very sparingly
soluble in hot alcohol; it was easily soluble in ether, light petroleum, chloro-
form and ethy! acetate. It crystallised readily from a mixture of chloroform
and alcohol or from ethyl acetate in rosettes of fine needles; m.p. 64°5
(corrected). It was dried at 106°.

The substance was identified by the results of elementary analysis and by
the percentage of palmitic acid estimated by Hehner’s method; these values
calculated for mannitol hexapalmitate, mannitol pentapalmitate and for

palmitic acid are also appended.
%/9 Palmitic

°l) Hydrogen °/) Carbon acid
Calculated for mannitol hexapalmitate 12-05 76-03 95-4
S »» Mannitol pentapalmitate 11-90 75°26 93°37
mm », palmitic acid 12°5 75°01 100
Analysis : (1) 0°2245 g.; 0-2430 g. H,O; 0°6270 g. CO,.

(2) 0°2135 g.; 0°2335 g. H,O; 0°5940 g. CO,.
(3) 1°7067 g.; 1°6220 g. of palmitic acid.
/) Palmitic

%ly) Hydrogen 9/9 Carbon acid
Analysis (L) 12°03 76°19 =
” (2) 12°14 75°90 —_—

» _—«(3) — ss 95°06

The substance was therefore identified as mannitol hexapalmitate corre-
sponding to the formula
C,sH3,CO . 0. CH, . (CH . OOCC,;H31)4. CHa. O . OCCysHs1 .

434 M. STEPHENSON

~ Glucose pentapalmitate. 1 g. of glucose and 9 g. of pyridine were added
to 83g. of palmityl chloride (2°4 g. in excess of that required for complete
esterification). The procedure followed was the same as that previously
described; as in the case of the glycerol compound decolorisation with
charcoal and subsequent filtration of the melted ester were necessary. The
total yield obtained after extraction in the Soxhlet apparatus was about 5 g.
(83 °/,). The total yield of purified material was 2°8 g.

The substance obtained was very slightly soluble in cold alcohol and only
sparingly soluble in hot alcohol; it dissolved readily in all the solvents
commonly used for fats. It was recrystallised from ethyl acetate and melted
at 62° (corrected). The substance was readily hydrolysed by alcoholic potash
as was shown by the rapid darkening of the solution due to the action of the
alkali on glucose, coupled with the fact that the substance was soon completely
dissolved. The action of the glucose on the alkali prevented a saponification
value from being taken. An attempt was made to estimate the palmitic
acid by Hehner’s method; a known weight of the substance was saponified
with alcoholic potash, the alcohol evaporated off and the soap dissolved in
water. The solution was very dark owing to the presence of the resins
already mentioned, and it was found on acidifying and attempting to filter off
the palmitic acid that the presence of these substances caused the palmitic
acid to pass through the filter paper. An attempt to effect the complete
acid hydrolysis and afterwards to estimate the glucose was next made; 0°397 g.
of the ester was placed in a flask with 50 ce. of absolute alcohol in which,
even on boiling, it only partially dissolved, the greater part sinking as oily
drops; 25 ce. of 20 °/, hydrochloric acid were added and the flask boiled
under a reflux condenser for 84 hours. At the end of this time the hydrolysis
was still incomplete as was shown by the presence of small drops of insoluble
ester in the boiling alcohol. When the contents of the flask were filtered to
remove the unchanged ester and the filtrate neutralised the presence of
glucose was shown by the reduction of Fehling’s solution. Owing to the
impossibility of removing the unchanged ester quantitatively the glucose
could not be estimated. It was therefore necessary to rely solely on the
combustion results for the identification of the substance.

0°2233 g.; 0°2410 g. H.0; 0°6163 g. COp.

0/9 Hydrogen 0/9 Carbon
Analysis 11-99 75°26
Calculated for glucose pentapalmitate 11°82 75°32

“A BA », tetrapalmitate 11°66 74:20

M. STEPHENSON 435

The substance was therefore identified as glucose pentapalmitate corre-
sponding to the formula

CH, . OOCC,;H3; .(CH.O.OC. C,;H3;),. CHO.

This research was carried on during the tenure of a maintenance grant
from Newnham College, Cambridge; the expenses of the research were
defrayed by a grant from the Government Grant Committee of the Royal
Society. To both of these bodies I owe my thanks. My thanks are also
due to Dr R. H. A. Plimmer, in whose laboratory the work was done, for
his kind interest and valuable advice.

REFERENCES.

Belucci and Manzetti (1911), Atti R. Accad. Lincei [v] 20, i, 125 and 235.
Berthelot (1860), Chimie Organique fondée sur la Synthese, Vol. u.
Bloor (1910), J. Biol. Chem. 7, 427.

(1912), J. Biol. Chem. 11, 141 and 421.

Chevreul (1815-1823), Recherches sur les corps gras d’origine animale.
Griin (1905), Ber. 38, 2284.

and Schreyer (1912), Ber. 45, 3420.

Giith (1903), Zeitsch. Biologie, 44, 78.

Marie (1896), Bull. Soc. Chim. 15, 503.

Partheil and Von Velsen (1906), Arch. Pharm. 238, 267.

Scheij (1899), Rec. Trav. Chim. 18, 169.

ZL =HYDROLYSIS OF PROTEINS WITH ‘AN
m_COHOLIC “SOLUTION” OF “HYDROGEN
CHLORIDE.

PART I.
By CHARLES WEIZMANN anp GANESH SAKHARAM AGASHE.
(Received June 30th, 1913.)

In the usual method of isolating the mono-amino-acids resulting from
protein hydrolysis, aqueous hydrochloric acid is used as the hydrolysing
agent, and subsequently a saturated alcoholic solution of hydrogen chloride
is used for the purpose of esterification. An attempt was made to shorten
this double process by using a saturated alcoholic solution of hydrogen
chloride from the beginning to serve both as a hydrolysing and an esterifying
agent.

Further, a saturated alcoholic solution of hydrogen chloride is undoubtedly
a milder reagent than a saturated aqueous solution of the same, first, because
the solubility of hydrogen chloride is smaller in alcohol than in water, and
secondly, because it cannot be heated to as high a temperature as the aqueous
solution ; so it was expected that the process of hydrolysis would not be quite
complete, but might stop at an earlier stage than that of the amino-acids.

While the research was in progress, Pribram [1911] and Abderhalden and
Hanslian [1912] published their investigations, in which the action of this
reagent on proteins was studied. But their object being quite different from
that of the present investigation, they have not pursued the subject further.

In the present investigation, experiments have so far been made on
caseinogen from cow’s milk, and on silk-fibroin. In each case it was found
that the protein was only partially attacked. The corresponding mono-
amino-acids were isolated by the usual methods, and converted into copper
salts, which were analysed. It must be admitted that the evidence of these
analyses is not absolutely conclusive; but the authors believe that it is
sufficient to show the identity of the various substances, which are known
to be present in the different fractions by the more thorough investigations
made by using the ordinary method.

Bioch. vir 29

438 C. WEIZMANN AND G. S. AGASHE

Although the proportion of the protein attacked could not always be
exactly determined, as the reaction liquid could not be filtered free from
the solid floating in it, it appeared certain that a very large proportion had
been attacked. The yield of the amino-acids, however, was comparatively
poor. This makes it extremely probable that a large proportion of products,
more complex than the simple amino-acids, had been formed. ‘The isolation
of individual substances. from this complex mixture is an extremely difficult
task ; and experiments are still being carried on for that purpose.

EXPERIMENTAL.

Casetnogen.

200 grams of caseinogen were hydrolysed by boiling for several hours
with 2 litres of alcohol, saturated with hydrogen chloride. There was some
unchanged solid which could not be filtered. The mixture was, therefore,
evaporated in vacuo, and the residue decomposed with sodium hydroxide,
and extracted with ether in the usual way. A very strong smell of ammonia
was evident throughout this process.

The ethereal extract was then fractionated under 14 mm. pressure, and
the following fractions collected. The temperatures are those of the bath.

I -60°, II 60-100°, III 100-140°, IV 140-170".
They were further worked up in the usual way.

Fraction I: This gave 0:4 gram of alanine, which was confirmed by
analysing the copper salt.

0:0500 grm. gave 00162 grm. CuO. Cu = 25:92 per cent.
Calculated for alanine, Cu = 26°51 per cent.

Fraction IL: This gave 3°5 grams of a mixture of leucine (also isoleucine),
valine, and proline. They were isolated in the usual way, and confirmed by
analysing the copper salts.

Leucine (and isoleucine) :

0:1225 grm. gave 0:0293 grm. CuO. Cu =19°13 per cent.
Calculated for leucine, Cu = 19°63 per cent.

Valine :

0:1098 grm. gave 0:0300 grm. CuO. Cu = 21°85 per cent.
Calculated for valine, Cu = 21°48 per cent.

ee ee

*

CART USS 9 \ mere a OE

a
=!

C. WEIZMANN AND G. 8S. AGASHE 439

Fraction III: This gave about 1°5 gram of a mixture of leucine, ete., and
x little glutamic acid. The latter was converted into the copper salt, which
was analysed,

00724 grm. gave 0:0280 grm. CuO. Cu= 30°93 per cent.
Calculated for glutamic acid, Cu = 30°45 per cent.

Proline: About 1 gram of this was obtained from the last two fractions.
It was confirmed by analysing the copper salt, which was first made anhydrous.

0°0572 grm. gave 0:0154 grm, CuO. Cu = 21°53 per cent.
Calculated for proline, Cu = 21°77 per cent.

Fraction IV: Phenylalanine was isolated from this in the usual way, and
confirmed by estimating the chlorine in the hydrochloride, of which about
0'4 gram was obtained.

01105 grm. gave 0:0778 grm. AgCl. Cl=17°38 per cent.
Calculated for phenylalanine hydrochloride, Cl = 17°61 per cent.

The rest of the fraction gave 0°5 gram of a mixture of solids.

The dark: brown residue in the distilling flask was extracted with hot
water. On evaporating the aqueous solution, about 10 grams of a yellowish
white powder were obtained, which was most probably a mixture of diketo-
piperazines.

The residue in the extraction flask was not worked up as usual for
re-esterification, but is being worked up with the object of isolating the
more complex products of hydrolysis.

Silk-fibroin.

50 grams of silk-fibroin were boiled for several hours with about a litre
of alcohol, saturated with hydrogen chloride. The fibre was entirely broken
down, but a very fine solid was still floating in the brown liquid, which, in
this case, could be easily filtered off. The fine solid weighed about 14 grams
on drying.

The liquid was further dealt with exactly as in the last case; here, too,
the smell of ammonia was very evident during the process of extraction with
ether. The ethereal extract was distilled in the usual way, the temperature
of the bath being only taken up to 100°. About 5:5 grams of a mixture of
glycine and alanine were obtained. They were not separated quantitatively ;

but the picrate of glycine was prepared according to Levene’s method, and
melted at 190°.

29—2

440 C. WEIZMANN AND G. 8. AGASHE

_No erystalline solid could be obtained from the residue in the distilling
flask, which, however, gave very beautifully Deniges’ test for tyrosine, as
modified by Mérner [1902].

The residue from the extraction is being worked up for ie einen of
complex hydrolytic products, as in the last case.

Another interesting experiment was made with silk-fibroin. About a
gram of it was heated with about 10 cc. of absolute alcohol saturated with
hydrogen chloride, in a sealed tube for about seven hours at 110-120°. The
whole of the fibroin was found to have gone into solution without any

residue.

REFERENCES.

Abderhalden and Hanslian (1912), Zeitsch. physiol. Chem. 77, 285.
Moérner (1902), Zeitsch. physiol. Chem. 37, 86.
Pribram (1911), Zeitsch. physiol. Chem. 71, 472.

XLIV. THE FLOWER PIGMENTS OF AWNTIR-
RHINUM MAUS. 2. THE PALE “YELLOW
OR IVORY PIGMENT.

By MURIEL WHELDALE (Fellow of Newnham College, Cambridge) AND
HAROLD LLEWELLYN BASSETT (Trinity Hall, Cambridge).

From the Laboratory of the John Innes Horticultural Institution
and the Balfour Laboratory, Cambridge.

(Received July 22nd, 1913.)

In a previous communication [ Wheldale, 1913], the method of preparation
of the crude pigment has been described. The varieties of A. Majus originally
used for Mendelian experiments were the following :

White

Yellow

Ivory

Yellow tinged with bronze
Ivory tinged with rose doré
Bronze

Rose doré

Yellow tinged with crimson
Ivory tinged with magenta
Crimson

Magenta.

Ivory is dominant to yellow and since white may carry any factor but the
one producing yellow, the result of crossing any variety with white depends
on the gametic constitution of the white individual used. Of the varieties
forming anthocyanin, the Mendelian relationship may be represented in the
scheme on p. 442, Dominance is denoted by the direction of the arrows;
thus rose doré is dominant to bronze, magenta to rose doré, etc.

It has been previously suggested that ivory contains a chromogen of the
nature of a flavone, from which the red and purple anthocyanins are formed
by stages of oxidation or polymerisation or both. Also that the pigment

442 M. WHELDALE AND H. L. BASSETT

of the yellow variety and of the yellow patch on the palate of all varieties
(except white) is due to a second, more deeply coloured flavone. Microscopic
examination and microchemical tests show that anthocyanin and yellow
pigments are mostly limited to the epidermis of the corolla, while the inner
tissues contain the ivory chromogen. It is obvious therefore that all crude
extracts of entire flowers will contain two or more pigments.

alae Crimson

Yellow tinged Crimson —

- Bronze
Rose dore

Ivory Yellow tinged Bronze

tinged Rose doré

Although no analyses have yet been made of the anthocyanins from
crimson and bronze, yet purification of the crude pigment indicates that these
colours are merely due to mixture of magenta with yellow, and red with
yellow respectively and not to specifically different pigments.

The constituent pigments of the varieties may be thus expressed :

Yellow (ivory, yellow).
Ivory, lower lip (ivory, yellow).
,, upper lip (ivory).

Yellow tinged bronze, bronze, ivory tinged rose doré, rose doré (yellow,
ivory and red).

Yellow tinged crimson, crimson, ivory tinged magenta (yellow, ivory,
magenta).

Magenta, lower lip (yellow, ivory, magenta).

si upper lip (ivory, magenta).

The crude pigment from yellow, ivory (upper lip), ivory tinged magenta,
bronze, rose doré, crimson and magenta (both upper and lower lips separately)
was extracted with ether for several months.

The yellow ether extracts were in each case crystallised from alcohol, the
first deposits, consisting mainly of the ivory pigment, separated off and the

M. WHELDALE AND H. L. BASSETT 443

melting points determined. All the products melted between 336° and 340°.
In five cases the acetyl derivative was prepared by boiling with acetic
anhydride as described in the previous paper and crystallising several times
from ethyl acetate. The final products were white substances crystallising
in needles and all melting at 181-182”.

The five acetyl derivatives were submitted to combustion, giving the
following results :

Acetyl derivations of ivory pigment from

C. lal
(Eh yellaver nn atcsene ep ach oceans GS'10 4 /o 28568 O49,
(2) avory; upper lip... 62°84 /,...... 3°99 °/,
(3): TEGECE, camnicescwsud aud totnadaer 013 1 il Ae 4°14 °/,
(4). ROME ern cca ncaa aaneneet CIO Yea ce asd 92
(5) ivory, tinged magenta...... 63:23°),...... 3°88 °/,

which agree closely with combustion results of the acetyl derivative previously
obtained from ivory extracted from magneta [Wheldale, 1913]:
Acety] derivative of ivory pigment from magenta:

w H.
(ayeenen 63:19 /,......4°30
Giese 63-21 ",......4°22
Geers 62-95 /,......4¢31

It was there suggested that the ivory pigment might be apigenin,

of which the acetyl derivative should theoretically give:
C. las
63°64 °/, 4-04°/).

An acetyl derivative of apigenin was obtained by Kostanecki [ Rupe, 1909].
Perkin [1897], on the other hand, was unable to acetylate apigenin but
prepared a benzoyl derivative by the Schotten-Baumann process. The
derivative thus prepared was slightly soluble in alcohol, more readily in
benzene, from which it crystallised in colourless needles melting at 210-212°.

A benzoyl derivative having the same properties and melting point as
the above was prepared from the ivory pigment by the following method,
The pigment was dissolved in about five times its weight of pyridine, cooled
in ice and benzoyl chloride added. On addition of dilute sulphuric acid, a
yellow oil separated out which was collected and spread on a porous plate.
It solidified after a time and was extracted with ether which removed
benzoic acid and some yellow products leaving a white wax-like substance.
The latter was crystallised from benzene. Melting point 210-212’.

444 M. WHELDALE AND H. L. BASSETT

Analysis results :
Benzoyl derivative of ivory pigment :
C. balk
73°82 °/, 4-02 °/,.
Theoretical result for benzoyl derivative of apigenin :
C. Jeb
74°23 °/, 3°78 4.
The properties of apigenin given by Aberhalden [1911] are as follows:
Formula, C,;H,,O;.
O
Ho —OH

CH

M. Pt. 347°. Crystallises in yellowish-white platelets, readily soluble in
alcohol, slightly in ether and hot water, Becomes bright yellow in alkali
solution. The yellow solution in concentrated sulphuric acid has at first a
greenish, later a bluish fluorescence. The alcoholic solution gives a blackish-
brown coloration with ferric chloride and brown-red with ferrous sulphate.

The ivory pigment corresponds entirely with this description. From
the acetyl derivative, the pure pigment was obtained by hydrolysis with
alcoholic soda. On neutralisation with acid, the pigment came down as a
yellowish-white precipitate and after crystallisation from dilute alcohol,
melted at 347°.

We therefore conclude that the ivory pigment is apigenin, and that it is
present in each of the main classes of varieties of Antirrhinum with the
exception of the white. In the plant, apigenin exists undoubtedly as a
glucoside, though the kind of sugar and the number of molecules attached
still remain to be ascertained.

It appears possible that the deeper yellow pigment may prove to be
a flavone, similar in constitution to apigenin, but deeper in colour owing to
the presence of an additional hydroxyl group.

REFERENCES,

Abderhalden, EK. (1911), Biochemisches Handlexikon, 6, 51.
Perkin, A. G. (1897), J. Chem. Soc., 71, 801.

Rupe, H. (1909), Die Chemie der naturlichen Farbstoffe, 2, 63.
Wheldale, M. (1913), Biochem. J. 7, 87.

XLV. OBSERVATFONS ON THE USE OF THE
FOLIN METHOD FOR THE ESTIMATION OF
CREATINE AND CREATININE.

By WILLIAM HENRY THOMPSON,
THE LATE THOMAS ARTHUR WALLACE! ann
HAROLD REGINALD SEPTIMUS CLOTWORTHY.

From the School of Physiology; Trinity College, Dublin.
(Received July 22nd, 1913.)

During a continuous use, for over three years, of the Folin colorimetric
method of estimating creatine and creatinine we have made observations by
way of testing its accuracy and limitations which we think may be useful to
others if published. Some of these have meanwhile been pointed out by
workers in other laboratories, but we include them for the sake of

corroboration.

PROCEDURE.

The colorimeter we mostly used was that of Jobin—a modification
of the Du Boscq instrument. We also compared its readings with those
of two similar instruments made by Pellin? and with one of the Kagenaar
pattern as described by Hoogenhuyze and Verploegh [1905]. We found that
consistent and reliable readings were obtainable with all, after a little
experience in their use had been acquired.

For making the standard bichromate solution we used Kahlbaum’s purest
bichromate of potassium certified to contain 99°97 °/, of the pure salt. This
was fused, then cooled in a desiccator and a semi-normal solution made from
it containing 25°54 g. in the litre.

The readings of the colorimeter were made by one of us (W. H. T.), but

1 Immediately before sending this paper for publication we learned, to our great sorrow, of
the death of our co-worker, Thomas Arthur Wallace, at Agra in India, whither he had gone,
a short while before, to take up a post in the Missionary College there. W.H.T., H. R.S. C.

2 One of these was very kindly lent to us by Dr Ritchie of the Royal College of Physicians
Laboratories, Edinburgh, to whom we express our best thanks.

446 W. H. THOMPSON, T. A. WALLACE AND

were checked in all important cases by comparative readings made mostly by
a second person in the laboratory familiar with the use of the instrument’.

To avoid unconscious bias, the solutions in all the critical readings were
made up by a second person and given to the reader under a number which
gave no indication of which solution it was, this being disclosed only when
the series was finished. We found this a very useful safeguard.

Great care was taken to obtain equal illumination of the two halves
of the colorimeter. One important cause of inequality was found to be
uneven blackening of the interior of the metal box in which the glass
parallelopipeds are contained. The possibility of shifting of the zero on
one or other side of the instrument was constantly examined and allowed for.

Another routine precaution, taken at the beginning of every set of
readings and repeated as a rule after every four readings (if the series
exceeded six in number), was to read the standard solution against itself.
This we regard as of the utmost importance.

Few will believe, without actual experience, how much the reading of
the standard by an individual may vary and how necessary is this control.
The procedure also furnishes a ready test of fatigue, whether retinal or
cerebral. When fresh, there is no difficulty in obtaining similar readings
on the two sides using the standard solution in both tubes. When fatigued
it was found that the readings were too high or too low, oftenest the latter and
frequently to the extent of 0°5 mm.

The degree of illumination unquestionably affects the result. This has
been pointed out by several workers. On dull days the readings tended to
be lower than on bright days. Our instrument was placed opposite a large
window, with northern aspect, and the readings made, for the most part, in
the middle hours of the day. The practice of reading the standard against
itself reduces the error for variation in the illumination.

CONTROL OF THE STANDARD SOLUTION.

Folin [1904], in fixing his bichromate standard, tested it against pure
creatinine prepared from urine, also against creatinine solutions made by the
conversion of creatine, and likewise against the double picrate of creatinine
and potassium obtained from urine.

Of subsequent workers, there appear to be only three who controlled

1 We are grateful to Professor Noel Paton and Dr Rosenheim for assisting us in this way on
more than one occasion.
* This last was overlooked by one of us (W. H. T.) in a recent note on the subject [1913].

= iy A Dae, eh

he EIN EL LAD GO Ee

>

H. R. 8. CLOTWORTHY 447

their standard solutions by the first of these procedures, namely: Hoogen-
huyze and Verploegh [1905], Koch [1905] and Klercker [1907].

The creatinine used by Hoogenhuyze and Verploegh was prepared from
urine and gave almost theoretical readings (8°14 mm. instead of 81 mm.).
That of Koch was obtained from Merck and required 11°46 mgms. to give the
standard reading of 8°1 mm. in the colorimeter, this being normally given
by 10 mgms. of creatinine treated in the way described by Folin. Klercker
used a solution of creatinine prepared from urine. The content of this was
determined by estimating the total nitrogen and found to be 96°6°/, of
the theoretical, assuming all the nitrogen to be present as creatinine.

We endeavoured to control our standards by this method, but early
attempts to prepare pure creatinine did not prove satisfactory. Later we
prepared a small sample of pure creatinine free from creatine. With this
we obtained theoretical readings. The difficulty we encountered of pre-
paring pure creatinine is admitted by Folin. This has been largely reduced
by his newer methods.

Of these latter, that of heating pure crystalline creatine in the autoclave
at a pressure of 4°5 kilos per sq. em. for 3 hours was tried by one of us. The
yield was not so high as claimed for it by Folin, nor the creatinine so pure
as anticipated, but the method proved valuable in giving an easy means
of preparing pure creatinine picrate which we used later as a control.

Most workers have tested their bichromate standard solution against solu-
tions of creatinine prepared by the conversion of pure creatine. ‘This was
the method followed by Dorner [1907], Gottlieb and Stangassinger [1907],
Hoogenhuyze and Verploegh [1908], Weber [1908], Benedict and Myers [1907].
We gave it an exhaustive trial. For this purpose a large number of
samples of pure creatine, obtained from different kinds of flesh (horse, rabbit,
dog), were prepared from time to time, and purified by recrystallising six to
eight times till absolutely pure. Several Dumas analyses of these were
made, the results of which are here shown.

TABLE I. Showing results of Dumas analyses of samples of creatine.

Creatinetaken N. evolved Temp. Pressure N. per cent. Theory
1 0°1056 29°4 c.c. 16° C, 740 mm. SPP | 32°06
Y 02220 61°8 18 740°8 32°03 =
3 071625 44°0 ity 755°5 31°92 a
4 02020 54°8 16°5 756 B222 =
5 0-1260 35°4 22 753°4 32°10 Pe
6 0°2160 58°0 18°8 754-6 32°30 *

(Two of these estimates, Nos. 5 and 6, were kindly made for us by
W. Caldwell, M.A., Senior Demonstrator of Physiology in this school, who

448 W. H. THOMPSON, T. A. WALLACE AND

also checked readings for us at various times. For this assistance we express
to him our best thanks.)

There can be no doubt from the above analyses that our preparations
of creatine were absolutely pure. This is borne out also by the fact that the
crystals in every case, after being fully air dried at room temperature, lost
the calculated amount of water of crystallisation when further dried in the
oven at 110°C. In this latter drying—notwithstanding statements to the
contrary—creatine does not decompose and we found it unsafe to trust for
complete drying to a temperature lower than this. In weighing the
anhydrous creatine, which is very hygroscopic, great care was taken to guard
against the absorption of water.

In making the test solutions of creatinine we usually took the quantity
of creatine requisite to give when fully converted a 0°1°/, solution of creatinine,
that is to say, 0116 grm. of anhydrous creatine or 0:132 erm. of crystalline
creatine for 100 cc. In a few cases, solutions of anhydrous creatine of other
strengths were used.

From a large number of estimates made with these solutions we found,
as is shown in detail in Tables XIII and XIV, that the average conversion by
the water bath method amounts to 96°5°/, of the theoretical yield—varying
from 93°1°/, to 100:2°/,. The readings on the colorimeter scale corresponding
to these were 8°39 mm. for the average, 87 mm. for the lower, and 8°08 mm.
for the higher, instead of 8°1 mm. for the theoretical yield.

On more carefully examining the results published by others we found
that our experiences were not singular. Folin gives no data on this point in
his first article, but speaks of the difticulty of complete conversion of creatine
in the article in Hammarsten’s Festschrift [1906]. Dorner [1907] with 0°1°/,
solutions of creatine obtained results varying from 85 to 100°/, of the
theoretical yield. Gottlieb and Stangassinger [1907] 83:19 to 99°98 (with
4°56°/, HCl) and 91:55 °/, to 92°98°/, (with 2°28°/, HCl). Jatté’s [1906] best
result with 2—2°5°/, HCl was 94°3°/,. It did not seem to us therefore that
this method was sufficiently accurate to be used as a control for the standard
bichromate solution and another test was substituted by one of us the
results of which have been published [Thompson 1913]. It was found that
solutions of creatinine picrate and also of the double picrate of creatinine and
potassium, when taken in quantities equivalent to 0:1 grm. creatinine and
treated in the usual way, fulfilled the required conditions, giving readings
of 81 mm. in the Du Boscq colorimeter. This result settled some doubts
which had previously existed in our minds with regard to the accuracy of
the Folin standard.

H. R. S. CLOTWORTHY 449

CONDITIONS WHICH AFFECT THE READINGS.

Our object was not to test all the possible conditions which might affect
the results. We confined ourselves to those which came under our notice in
the application of the method to our work.

These were (1) in the estimation of creatine, the best amount of acid (HCl)
to use for the conversion into creatinine; (2) the optimum time and
temperature for the development of the colour in the estimation of creatinine
and creatine; (3) the influence of the quantity of alkali added for the same
purpose ; (4) the relative values of readings at different parts of the scale ;
(5) the influence of urinary pigment on the estimation of creatine in urine ;
(6) the estimation of creatine in the presence of dextrose and the use of
phosphoric acid ; (7) the recovery of creatine from diabetic urine; and (8) the
reliability of the autoclave method.

1. Estimation of Creatine. The amount of acid to be used in the conversion.

For the conversion of creatine into creatinine, Folin in 1904 recommended
5 cc. of normal HCl to be added to 10 cc. of a dilute solution of creatine.
Later [1906] he increased this amount to 10 c.c., a quantity which has been
adopted by most subsequent workers. A few however, namely Dorner {1907 ],
‘Hoogenhuyze and Verploegh [1905], have used double that quantity, namely
20 cc. of N. HCL.

We have tested all three quantities and are convinced that the best
results, both with the water bath and autoclave methods, are got by using
10 cc. N. HCl with 10 cc. of weak creatine solution. The conversion with
5 cc. is liable to be less complete, particularly if glucose be present.

From several observations the following two consecutive sets of readings

are taken:
N. HCl Method Reading N. HCl Method Reading
20 c.c. water bath 8°59 mm. 20 c.c. autoclave 8°52 mm.
20 autoclave 8°69 10 Bi 8°26
10 Pe 8°46 5 95 8°32
5 ae 8°6

The readings with 20 cc. and 5 cc. are uniformly higher, that is, the
yield is uniformly less, than with 10 c.c. It would seem as if a destruction
of creatinine took place when 20 c.c. of acid were used.

Except at the very outset we have always used in these investigations
10 cc. of N. HCl for the conversion of creatine.

450 W. H. THOMPSON, T. A. WALLACE AND

2. The Time and Temperature for Development of the Colour.

Folin originally gave a time limit of 5 to 10 mins. for the development
of the Jaffé reaction. Later [1906] he restricts the time to 5 mins. and this
period has been adopted by most workers. Some however have departed
from it, thus Dorner [1907] gave 5-15 mins., Benedict and Myers [1907]
only 35 mins., Mellanby [1908] 5 mins. with some latitude, Mendel and ~
Rose [1911] 10 mins.

Comparatively few observers however have recorded the temperature
which they adopted. Amongst these are Hoogenhuyze and Verploegh [1905],
who diluted with water at 15° C., Dorner [1907], who cooled the solution after
adding the picric acid and Klercker [1907], who used water for dilution at
different seasons of the year of about the same temperature. Mellanby [1908]
also states that it 1s necessary for the temperature to be constant but does
not give the degree.

We found with temperatures of 10° C. to 15°C. that the maximum colour
was not developed under eight minutes. Between 15°C. and 17°C. the
time required was seven minutes, above 17°C. and below 20°C. five minutes
is sufficient.

The results are shown in the following table:

TABLE II. Creatinine estimation: influence of tume and temperature on
the development of the Jaffé reaction.

Readings Readings Method of

Time 5 mins. Time 7 mins. Temp. conversion
8°75 mm. 8°52 mm. 10 Water bath 5 hours,
9°30 8°85 10 Autoclave HCl 25 mins.
8°55 8°45 15 Water bath.
8°76 8°50 15 Autocl. HCl 25 mins.
8°80 8:70 15 Autoel. H,PO, 30 mins.
8°30 8:11 15 Water bath HCl.
8°48 8°35 iy Water bath HCl 3 hours.
8°29 8°30 20 Water bath HCl 3 hours,

Nearly all our observations were made at a temperature lying between
15° C.-17°C. The flasks in which the reaction was carried out were weighted
with leaden rings and kept in a basin of water at this temperature.

3. Effect of the amount of alkali used in the Folin method.

We found that the quantity of alkali added to develop the colour of the
Jaffé reaction distinctly affected the results. For creatinine Folin recommended
5 cc. of 10°/, NaOH solution; for creatine, the same amount over and above

H. R. S. CLOTWORTHY 451

that necessary to neutralise the acid used in the conversion. These quantities
have been adhered to generally by others with few exceptions. Thus Benedict
and Myers [1907] used 10 c.c. for 3°5 mins.; Mellanby [1908] concluded that
no great accuracy was necessary; he got the same results with 10 cc. and
3c.c. His observations were chiefly concerned with meat extracts. Grindley
and Woods [1906], also working with meat extracts, used 5 cc. to 10 cc.
10°/, NaOH. Hehner [1907] soon afterwards stated, in a short publication
giving no data, that an excessive quantity of alkali diminished the colour in
the case of meat extracts, and that to obtain the best results the amount of
alkali added must be quite small. He also considered that more picric acid
than the quantity prescribed by Folin should be used when dealing with these
extracts. In reply to this Emmett and Grindley [1907] repeated the work of
Grindley and Woods, finding that in the determination of preformed creatinine
the use of a small or large amount of alkali made almost no difference, but
that slightly better results were obtained with 10 c.c. or 15 ¢.c. than with 5 c.e.
For transformed creatine 10 cc. and 15 c.c. gave the same results, and both
were better than 5 cc. It is not clear, however, that the acid used in the
conversion was neutralised before adding the picric acid solution, nor was
their method strictly comparable to that followed by others—the quantities of
creatinine containing solution were much larger.

Cook [1909] also found in estimating the creatine content of meat extracts,
that 5 cc. of alkali did not give the maximum colour, while 10 c.c. and 15 c.c.
gave identical results. With solutions of creatinine he found that 5 c.c. and
10 c.c. gave similar results. Accordingly, he recommended the use of 10 c.c.
all round. For a considerable time we followed this recommendation without
question, but later discovered that 10 c.c. of 10°/, NaOH destroyed some of
the creatinine and gave too low results. This applies chiefly to pure solutions
of creatinine but also, though to a less extent, to urine. We have not tested
the point in the case of meat extracts. Our results are shown in the
following readings.

TABLE III. Influence of the quantity of alkali on the colour developed.

Readings with 5 c.c. Readings with 10 e.c.
NaOH of 10°/, strength NaOH of 10°/, strength
(a) Creatinine in aqueous solution
8-29 8°75
8°25 8-70
8°58 8°75
8°78 9°05
8:20 8:58

Mean 8:12. Mean 8:77

452 W. H. THOMPSON, T. A. WALLACE AND
TABLE III (continued)

(b) Creatinine in urine

9°54 9°66
9°05 9°45
9°55 9°75
6°30 6-50
fico 79
8:25 8°27
6°26 6°41
bia 5°85
Mean Terie Mean Besley

It will be seen that with pure solutions of creatinine, the average reading of
the series, when 5 cc. of alkali were used, was 842 mm. (=0:0966°/, of
creatinine): when 10 c.c. ofalkali were used the average reading was 0°35 mm.
higher, that is 8°77 (=0:0924°/, of creatinine). There was also a difference
with urine though less marked. The mean of the series of readings with
5 ce. of alkali is 7°77 mm. (=0°1042°/,), that with 10 cc. alkali is 7°97 mm.
(=0:1016°/, creatinine). So far as these solutions are concerned we are
unable to confirm Cook: on the contrary, we find that better results are
given when Folin’s original directions are strictly followed.

4. Relative values of readings at different parts of the scale.

It has been generally accepted that within a lower limit of 5 mm. and an
upper one of 12 mm., as stated by Folin, the values of the readings are
proportional. We have not found this to be so, nor is it generally true in
colorimetry that the dilution ratio of a solution is inversely proportional to
the height of the column of coloured solution. Folin natu ally refers to this,
and explains the exceptional result in his method as follows. He concludes
that dilution produces a diminution of the total colour of the solution—not
an increase such as would arise if the colour were due to a red ion—but that
the diminution is hidden by the increased relative depth of tint which occurs
with increase of the height of column seen through. It has, however, since
been shown by Chapman [1909] that the colour in the Jaffé reaction is due
not alone to the formation of picramic acid, as had long been held, but to a
mixture of picramic acid and a still redder reduction product of picric acid,
namely diaminonitrophenol. It is doubtful therefore if Folin’s explanation
will hold good, nor are the facts quite in accordance with his view.

On this point we made a large number of observations, comparing in each
case the reading given at a certain dilution with that given by the same
solution diluted to one and a quarter, one and a half, or double the volume.

H. R. 8S. CLOTWORTHY 453

In the following table we give some of our results confining these to a
comparison of readings with dilutions to 250 c.c. and 500 e.c., but making
the selection so as to show the results with different solvents. These latter
were water, dog’s urine, normal human urine, and diabetic human urine.

TABLE IV. Relative value of readings at different parts of

colorimeter. scale.

Readings Readings

Solvent Dilution 10-250 = Dilution 10-500 Ratio
1. Water (HCl) 63° mm. 11°2 mm. Tela)
2. ”% 6°5 11°36 AREY i?
3. ” 6°55 11°60 1 ra Ie
4, as 5°50 10-10 Lt 80
5 7 5°45 9°80 Daa ey if
6. 3 6-15 11-00 1: 1:79 |.Mean=1: 1-76
is a 6:28 11-00 eee
8. > 6:15 10°90 Se i
9. e 6-15 10-70 1:1:74
10. te 6°30 10-90 tte Near
ll. : 6°24 11°15 ivpaa heresy)
12. Dog’s urine (HCl) 4:80 8-90 A sc1-6a))
13. - - 4°43 8°50 oe
14. ” 93 5°90 11°30 hean ios!
15. - zs 4°85 9°30 1: 1-92 | Mean=1: 1:91
ae 4°63 9-00 1: 1-94 |
17. , 4 4:60 8-96 1: 1-95 |
18. aC a 4°87 9°24 e907)
19. Human urine (HCl) 6°30 12-20 1g et
20. ss ns 4°30 8:10 11°88 }
a. : 5-55 10-30 MAVaieT oa ee
22. BA s3 5°50 10°67 Hes dO4
23. Diabetic urine 5°22 10:10 1 ae] 3
24. re is 6°60 12-80 1:1-94¢ iter oe
cS) 6-60 12-00 Levey fe st St
26. a a 6°25 11°30 1:1°81)

On examining the list it will be seen that the higher reading is never

double the lower and that the ratio differs considerably when urine is
compared with water as the solvent, also that there is a difference, though

On the other hand,
The mean ratio

not marked, between normal and diabetic human urine.
the dilution ratio is identical for human and dog’s urine.
instead of being 1 to 2, is, for solutions in water 1 to 1°76, in dog’s urine and
human urine 1 to 1°91, in diabetic urine 1 to 1°87. These results were borne
out by dilutions to other degrees than one to two.

Our general conclusions are that for accurate work readings can only be
regarded as strictly proportional if they lie between a lower limit of 7 mm.
and an upper one of 9 mm. If separated by a wider interval a correction

factor has to be applied, and this varies for different solvents, but for all
Bioch. vir 30

454 W. H. THOMPSON, T. A. WALLACE AND

practical physiological purposes may be based on the mean of all the above
ratios for a difference of column of 5 mm., viz. 1:1'8 instead of 1:2.
In general therefore the upper readings of the scale give relatively too
high results, the lower readings on the contrary too low results, as compared
with the normal reading of 8-1 mm.

5. The influence of urinary pigment: recovery of creatine from
normal urine.

Several observers have called attention to the influence of urinary pigment.
Weber [1908] found that the darkening which occurs on boiling urine with HCl
lowers the reading and increases the estimation of creatine by fully 5°).
Benedict [1912] also found a similar effect and recommended the use of
granulated zine in the boiling to remedy it. Rose [1912] did not find this
procedure efficacious with diabetic urines and recommended the use of phos-
phoric acid instead of hydrochloric. Dreibholz [1908], with diabetic urine,
found a difficulty in matching the colour with the standard bichromate
solution after boiling the urine with hydrochloric acid, but failed to find a
useful remedy. With normal urine when the pigment was removed by
filtration the result was not affected.

Early in these investigations it was clear to us that in the case of dog’s
urine the darkening on boiling with hydrochloric acid lowered the colorimeter
readings. We therefore proceeded to investigate the point, and did so in
two ways, both of which were applied to normal human and dog’s urine.

In the first method the ordinary creatine readings of a series of urines
were compared with those given by similar samples of the same urines after
boiling with an equal quantity of normal HCl. In the boiling we used not
alone the water bath for three hours, but also the autoclave at 117°-120° C.
for two different periods, namely, 15 mins. and 25-30 mins. ‘The results
are shown in the following tables:

TABLE V. Showing colorimeter readings of human urine before and
after boiling with normal HCl.

Water bath Autoclave Autoclave
Unboiled 3 hours 15 mins. 25-30 mins.
1 78 7°33 7d 776
2 6°5 6°3 6°5 6°51
3 8-1 7-9 8:15 81
4 7°24 eilte 716 7°26
5 76 7:3 = 75
6 6°25 6°16 = 6-1
a 751 7°54 — 7:5

Mean 7°29 PTS eT Fg 7-22

H. R. 8S. CLOTWORTHY 455

The effect on human urine is therefore very slight. Taking the mean
yield of the unboiled urines as 100°/,, that of the water bath series amounts
to 102:25°/,, and of the autoclave, 25-80 mins., to 101°/,. The yield of
those boiled in the autoclave for 15 mins., when compared with the same
group of urines unboiled, works out at 101‘1°/,.

TABLE VI. Showing similar readings to the above for dog’s urine.

Water bath Autoclave Autoclave
Unboiled 3 hours 15 mins. 25-30 mins.

ik 6-2 57 6°35 6:06

2. 74 6-08 6°46 6°45

3. y 6-4 6°84 6°73

4. 7°74 73 7°54 7°47

5. 6°62 6°2 — 6°67

6. 8-5 7°87 -- 79

Ws 9°3 8-03 — 8-13
Mean 7°55 6°8 6°8 7°06

The results for dog’s urine, taking the mean of the unboiled series as
100°/,, work out as 111°/, for urine treated by the water bath method, and
106°9°/, for the same boiled in the autoclave for 25-30 mins. The mean yield
of the smaller autoclave series boiled for 15 mins. gives 1046 °/, when compared
with that of the same group unboiled. It is interesting to note that in both
series the effect on the pigment is more pronounced in the water bath than
in the autoclave method.

The second method consisted in comparing the recovery of creatine when
dissolved in urine with that of the same substance dissolved in water. The
procedure adopted in testing the recovery of creatine from urine was as follows.
Normal urines were taken and known quantities of the pure samples of
creatine we used throughout this research were dissolved in them—for the
most part 0°116 g. in 100 c.c. (=0°1°/, of creatinine).

Determinations were then made, (1) of the preformed creatinine in the
urines unboiled, (2) of the readings of the same urines without any addition
of creatine but treated by the water bath and autoclave methods, (3) of the
total creatinine in the urines after creatine was added, these being treated as
in (2). Two sets of subtractions were then made; (a) the creatinine of the
unboiled urine was subtracted from the total creatinine of the urine to which
creatine was added; (b) the creatinine of the boiled urine was subtracted
from that of the same urine plus creatine.

The values for the first set of deductions should give too high results for
recovered creatine, at all events in the dog’s urine, if the darkening of the
pigment affect the recovery. Those of the second set, presumably the

30—2

456 W. H. THOMPSON, T. A. WALLACE AND

true creatine values, should correspond to the results of recovery of creatine
from solution in water provided no other disturbing factor, than the effect of
the pigment, entered into the reaction.

Observations were made both with human urine and with that of the
dog. The following tables show the results:

TABLE VII. Showing the recovery of creatine from dog’s urine:
expressed as volume percentages of the urine taken.

Creatine recovered Creatine recovered
a, expressed as creatinine expressed as creatinine

Creatine added water bath 3 hrs. autoclave 25 mins.
expressed as ZB SS ss = eS eae

creatinine (a) (b) (a) (0)
He 0:05 0:0627 0-0598 0:0575 0:0567
2. 0:05 0:0594 0:0475 0:0613 0:0533
3. 0:10 0:1003 0-1003 0:0916 0:0910
4, 0-10 0-1029 0:0935 0:0939 0-0892
De 0:10 0-0392 0-0809 0:0925 0:0930
6. 0:10 0°1297 0-1221 0:1022 0:0950
Uk 0:05 0-0598 0:0461 0:0605 0:0478
Totals 0°5500 0°6040 - 0°5502 0°5595 0°5260

= O92 8,0 )/ O04), = L017 of, 9 Ay

The total yield of the first set of values by the water bath method,
column (a), works out as 109°8°/, of the amount added. That obtained by
the second method of deducting, column (0), gives a yield of 100°/,. The
corresponding values furnished by the autoclave method (25 wins. at
117-120°C.) are 101°7°/, and 95:4°/, respectively. Those therefore in
column (a) show an increase due to pigment of 10°/, by the water bath
method and of 6-7°/, by the autoclave method, figures which correspond
very closely with the effects of darkening of the pigment of dog’s urine
seen in Table VI. Moreover, the mean of the recovery by the two methods,
water bath and autoclave, obtained by the second method of deduction is
97°7°/, which is very close to the average recovery from solution in water
mentioned in the earlier part of this paper, namely, 96°5 °/,.

The results of recovery from human urine are seen in Table VIII.

On examining these figures it will be seen that the results of column (a)
in the water bath method, which include the pigment effect, are just over
2°/, higher than those of column (6). Similarly, those of the autoclave
method show a difference of 1:2°/, due to the pigment, and these differences
correspond almost exactly with those given earlier for the effect of the
pigment in human urine alone. The mean result for the two methods
(water bath and autoclave), taking column (b) in each case, which excludes

Ay

* ov ped eee > Pee

H. R. 8S. CLOTWORTHY 457

the pigment effect, is 98°26°/,, that is slightly higher than the general
average for aqueous solutions. It may be concluded therefore that the
darkening of the pigment is the only disturbing factor in normal urine in
the recovery of creatine, and'that its augmenting effect in round numbers
is 2°/, for human urine and 10°/, for dog’s urine. The effect, moreover, may
be practically eliminated by the procedure of boiling the urine for the normal
period (in any given research) and deducting the results from those given
by the urine of subsequent periods treated in identically the same way—
that is assuming the pigment has remained unaltered throughout.

TABLE VIII. Showing the recovery of creatine from human urine:
erpressed as volume percentages of the wrine taken.

Creatine recovered Creatine recovered
; expressed as creatinine expressed as creatinine

Creatine added water bath 3 hrs. autoclave 25 mins.

expressed as = A ~w

creatinine (a) (b) (a) (b)
1 0:05 0:0521 0:0496 0:0563 00563
2 0°10 0:0980 0-0970 0-0970 0:0974
3. 0-10 0:0980 0-0937 0:0958 0:0944
4 0°10 071111 0-1092 0:1104 0-1072
5 0°10 0-0881 0:0885 0:0922 0-0920
Totals 0°45 04473 0:4380 0:4517 0:4473

=99°4 9/, = 97°33 9/p = 100-4 Jp = 99-2 9/,

6. Estimation of creatine in the presence of destrose.

Different statements had been made with regard to the influence of
dextrose on the Jaffé reaction in its application to the estimation of creatinine
and creatine. Klercker [1907] found that on long standing, when glucose
was present, the colour became deeper; Hoogenhuyze and Verploegh [1905]
that glucose produces no effect in the time necessary for the Folin estima-
tion; Dreibholz [1908] that the caramel produced on boiling diabetic urine
with HCl adversely affected the estimation of creatine; Taylor [1911] that
glucose introduces no effect on the estimation of creatine unless present to
above 5°/,; while Rose [1912] strongly supports the view of Dreibholz.

We felt it necessary from the outset of our investigations to ascertain for
our own satisfaction which statement was to be accepted. Our first observa-
tions dealt with aqueous solutions of creatine and dextrose, later we extended
them to diabetic urine in which creatine was dissolved.

458 W. H. THOMPSON, T. A. WALLACE AND

TABLE IX. Estimation of creatine in presence of dextrose.

Water bath Autoclave Autoclave Autoclave Theoretical amt.
3 hours 25 mins. 30 mins. 30 mins. Dextrose expressed as
HCl HCl HPO, 3 /o H,PO, 2 °%/o present creatinine
il, 0-1012 0-1020 — — 5% 0-1000
2, 0-1004 0:1002 — — 10 %Jo Bs
3h 0:0937 0-0936 0-0886 = 10 %J/o r
4, 0-0910 0 0907 0:0814 — 10 %Jo Pa
os 0-0895 0-0900 0:0815 0-0900 10 %J 0-0900
6. 0:0975 0:0968 0:0917 0-0960 10 Jo 0:1000
Uf 0-0324 0-0366 0:0323 0:0326 10 %Jo 0:0342
8. 0:0960 0:0964 0:0850 0-0889 10 %Jo 0:1000
Av. 9/o of
ae ee 96-6 96-6 86-7 925 0-1000

The foregoing table gives the first series of observations. It includes also
a comparison of the results of the water bath method with those of the
autoclave. Further, the efficacy of phosphoric acid in strengths of 3°/, and
2°/, for the recovery of creatine is compared with that of normal hydro-
chloric acid. ,

On examining the table it will be seen that with HCl—both water bath
and autoclave methods—the average percentage of recovery (96°6°/,) is
practically the same as when dextrose is not present in the solution. This,
as previously mentioned, was 96°5°/,. Our conclusion, therefore, is that
dextrose per se has very little, if any, influence on the results within the
time necessary for estimating creatine.

In the above observations we used 10 cc. of normal HCl for 10 cc.
of creatine solution, but in some of our early work the quantity of acid
added was only 5c. In these we found, when the quantity of sugar was
5°/, or over, that the creatine results were much too low. This has also

been the experience of M. Ross Taylor [1911]. It is essential therefore to _

use the larger quantity of acid.

It has been stated by Rose [1912] that the disturbing influence of
dextrose is due to the action of hydrochloric acid in causing a transforma-
tion of some of the sugar into caramel. It occurred to him therefore to
employ phosphoric acid in the conversion of creatine when dextrose is present,
since it does not caramelise sugar. Rose claims to have obtained better results
in this way than with hydrochloric acid. We tested the effects of the two
acids, using solutions of creatine both in the presence and absence of sugar.
In these tests varying strengths of phosphoric acid were used, 5°/), 4°/o, 3°/o
and 2°/,.. The results are shown in Table IX. In all cases the best results were
obtained by us with 2°/,, by Rose, however, with 3°/, phosphoric acid, the

> -

H. R. S. CLOTWORTHY 459

volume of acid added in both sets of observations being 20 c.c. The results
with 5 °/, and 4°/, phosphoric acid were much inferior—these strengths causing
a considerable destruction of creatinine.

We then felt it necessary to make a comparison of these same acids,
using creatine in aqueous solution. The following table gives the results:

TABLE X. Estimation of creatine in aqueous solution: comparison of
hydrochloric and phosphoric acid in transforming creatine into creatinine.

Autoclave Autoclave Theoretical amt.
Water bath 30 mins. 30 mins. expressed as
N. HCl H3PO, 3 °/, H;PO, 2 °, creatinine

1c 0°0942 0:0910 0:0940 0-1000

2. 0:0935 0-0885 0:0920 =f

3 0:0931 0-0915 0:0921 7

4 0:0928 0:0910 0-0931 -

5. 0:0924 — 0:0925 5

6. 0:0951 00932 0:0944 Pr

7 0:0976 0:0931 0:0976 ee

8. 0:0972 0:0944 0:0980 =z

9. 0-0960 00935 0:0942 43
Mean recovery °/, 94°6 81:8 94-6 ~

On examining the table it will be seen that the results with normal
HCl and 2°/, H,PO, closely correspond, while those with 3°/, H,PO, are
considerably lower. The mean result with HCl is 946°/, of the theoretical
yield; with 2°/, phosphoric acid the same; with 3°/, phosphoric acid 81:8°/,.
It still remained to test the recovery of creatine from diabetic urine—the
results are shown in the following section.

7. Recovery of creatine from diabetic wrine.

In these observations we made use of a series of diabetic urines obtained
from time to time from clinical hospitals in Dublin through the kindness of
members of the staff. We applied the same four methods for recovery
of creatine from diabetic urine which were used when the substance was
added to aqueous solutions of dextrose (see Table IX). The results are
necessarily somewhat more complicated than with simple solutions, since the
preformed creatinine and creatine had to be determined separately.

The full data are given in Table XI Pure creatine was added to eight
samples of diabetic urine obtained from different patients. Three of the
samples contained both acetone and aceto-acetic acid, four were free from
both these bodies, and one contained a trace of aceto-acetic acid but no
acetone.

W. H. THOMPSON, T. A. WALLACE AND

460

,8-88= %o9-F6= Yots-e6= So ¥-96=

6FL0-0 8610-0 980-0 F080-0 6FZ0-0 9610-0
oS F6= 10 0-98= lo 8-26= fy 8-96=
2¥60-0 0980-0 8260-0 8960-0 0000-0 0100-0
1 F-98= Wy tl8= Yo e-16= %y 6-88=
£980-0 2180-0 €160-0 6880-0 . 1920-0 €1Z0-0
7 9-26= %oT-00T= %oe-GOT= %/o¢-901=
0°26-0 TOOT-0 €0T-0 C90T-0 F¥00-0 FOT0-0
"lo F-96= Yo T-86= Y%oe-26= 5 9-F6=
£960-0 1860-0 €260-0 9760-0 0F00-0 €100-0
‘lo G-T8= "10 9-98= % 6-G8=
— 2180-0 0998-0 6680-0 = 1200-0
%TII= Yo s.= a
= 1190-0 020-0 0220-0 = OTE0-0
7 8-F76= %S6= %9-F6=
= CF60-0 060-0 8760-0 == 020-0
0G %9 € "SUIUI GZ ‘TM ‘SUITI OG ‘SUIUT YE
tog'H foa’H *[oojnB IOH “[o0jnB , “[90jne
(p) (9) 10H () %8°Od*H %oe’Od*H
Siesta aay he ae ee)

Aq punoj sv pateaooas aUlZBaIH

6920-0

6600-0

TFF0-0

6610-0

6800-0

6160-0

TLE0-0

9760-0

“SUIU Gz
*Jooyne

IOH
9

8960-0

1600-0

C¥F0-0

8610-0

€€00-0

6760-0

IT&0-0

G¢s0-0

porjeut

‘aM
10H
(>)

ae
Aq punoj sv outlan Jo aulyvedo poulIojatg

€¥80-0

000T-0

000T-0

0O0T-0

000T-0

000T-0

OOOT-0

000T-0

aulutyvaI0

8B

“poppe

aul} BaD

FLE0-0

SL¥0-0

0960-0

[190-0

L9%0-0

¢9€0-0

6860-0

0LG0-0

‘yuoo rad
aUIUT}RAIO
pesseidxe poawlqojarg

‘sauLin oyaqnip 02 pappy uaym aurynato fo hwaaoda. ay, bumoysy

IX WIV

plo® 91J99B-03998 10 9104908

ON ‘%o LT es0rjyxeq °*g
ploe oleoe
-0}9908 JO 90¥13 ‘at0}00B

ON °°) 8 B8013X0q *L
“OB “4908-0908 PU 9U0}008
WO “/oL-L esoryxeg “9
Pl08 91}90B-01998 10 9u0yo0R
ON ‘°/) 8. asomyxaq ‘G
Pl0B 01}908-04908 10 904908
ON “7/o 9.F e80I}xeq “fF
‘OB *J90B-0J90"8 PUB dTIOJONR
WIO_ *"/)L-6 W04xEq “¢
“0B “J90B-0J90¥ PUB BU0J00B
wog °°/) OT esouxeqg °%
plow 91}90¥-03908 10 oMOJe0B
ON ‘% ZG eS01yxeq ‘T

poureyuo» sult

H. R. S. CLOTWORTHY 461

On examining the results it will be seen that the percentage recoveries
with the different methods were as follows: (1) water bath HCl, 72°/, to
106°5°/,, the mean of the series being 91°8°/,; (2) autoclave HCl, 72°,
to 105°3°/,, the mean being91:2°/,; (3) autoclave H,PO, 3°/,, from 61 °/, to
100°/,, the mean being 87:1°/,; (4) autoclave H,PO, 2°/,, from 86°4°/, to
96°4°/,, the mean being 91°76 °/,.

The results where HCl has been used are about 5°/, inferior to the
corresponding recoveries from aqueous and dextrose solutions. Those with
2°/, H,PO, are less variable but the mean (taken however from a smaller
series) is identical with that of HCl. With 3°/, H,PO, the results are less
favourable. We are forced therefore to conclude that while no special ad-
vantage is gained by using phosphoric acid, there is a liability to destruction
of creatine by this reagent even with a strength of 3°/,. It will be remem-
bered that the same effect was found in the recovery of creatine from
aqueous solutions both in the presence and absence of glucose.

The remarkable feature, however, of the results is the variability in the
degree of recovery from different diabetic urimes. This was not seen in
the recovery from normal urine nor from aqueous or dextrose solutions, and
points to a disturbing factor other than pigment or sugar. It is also
remarkable that the lowest results were obtained from urines containing
acetone and aceto-acetic acid. We have not studied the effects of either
of these bodies, but they have been called attention to by Jatfé (acetone),
Folin (acetone, aceto-acetic acid, and aceto-acetic ester), Klercker (acetone),
Hoogenhuyze and Verploegh (acetone), Krause [1910] (acetone and aceto-
acetic acid), Wolf and Osterberg [1911] (aceto-acetic ester), Rose [1912]
(aceto-acetic acid), Greenwald [1913] (acetone and aceto-acetic acid). Krause
and Greenwald find the disturbing effect of aceto-acetic acid more serious
than that of acetone. Its presence increased the values for preformed
creatinine, and thereby lowered those of creatine. This is in harmony
with our observations, but we think the matter deserves a more thorough
investigation, starting with solutions of pure creatine to which dextrose,
acetone, and aceto-acetic acid are added in succession.

Since the foregoing was written Graham and Poulton [1913] have
published a note in which they show that aceto-acetic acid and its sodium
salt diminish the colour of the Jaffé reaction and thus lower the estimate of
creatinine. Aceto-acetic ester, used by Wolf and Osterberg, also by Folin,
has no appreciable effect.

462 W. H. THOMPSON, T. A. WALLACE AND

8. Comparison of the autoclave and water bath methods.

It may seem superfluous to have undertaken this part of the work, but at
the period when we began it, the reliability of the autoclave method was not
universally accepted. We therefore decided to test it for ourselves, and also
to determine the optimum time in the autoclave for the transformation of
pure creatine into creatinine at the temperature (117°C. to 120°C.)
recommended by Benedict and Myers.

Two series of observations were made. In the first the results of heating
on the water bath for three hours were compared with those of the autoclave,
heated for 15 mins., 25 mins., and 35 mins. respectively. They are given in

Table XIII:

TABLE XIII. Creatine estimation: comparison of water bath
and autoclave methods.

Theoretical amt.

Water bath Autoclave ~ Autoclave Autoclave expressed as

No. 3 hrs. 15 mins. 25 mins. 35 mins. creatinine

1h 0:1002 0-0994 0°1004 0:0997 0-1000

2. 0:0981 0:0944 0-0982 — +3

ae 0:0976 0:0935 0:0982 : 0:0942 .

4. 0:0997 0:0942 0:0995 — 33

5. 0-1000 0-0947 0-1000 0-1006 E

6. 0:0944 0-0880 0 0962 0:0953 x
Total 0-5900 05642 05925 03898 0-6000

=98:35%, =94-03 %, = 98°78 9/, =97°45 °/,

It will be seen that the results of heating in the autoclave for 15 mins.
are too low, while those for 25 mins. are identical, or almost so, with the
water bath results. Consequently we are of opinion that the optimum time
in the autoclave is 25 mins. at a temperature of 117°C. The results of
heating for 35 mins. are slightly lower than those for 25 mins., indicating
that the optimum period has been exceeded.

The second series gives a number of comparisons between the water
bath results and those of the autoclave heated for one period only, namely,
25 mins.

On examining the mean results of the foregoing a remarkably close
correspondence—indeed identity—will be seen, though the degree of con-
version 95°7°/, is a little lower than the average as given at the beginning
of this paper, namely 96°5°/,, which was obtained by taking the gross mean
of the determinations in Tables XIII and XIV. When we compare this

hyd Pay & 4 ee,

H. R. 8. CLOTWORTHY 463

with the gross mean of the autoclave determinations (25 mins.) the corre-
spondence also works out very close, namely 96°6°/, as against 96°5 Wire
Observations were also made for periods of one hour and two hours in the
autoclave. In both cases there was a considerable destruction of creatinine,
the results at one hour giving a yield of 92'4°/, and at two hours of 88°6 Pe

TABLE XIV. Creutine estimation: comparison of water bath
and autoclave methods.

Water bath Autoclave Theoretical amount
No. 3 hours 25 mins. expressed as creatinine
{fe 0:0975 0-0969 0°1000
8. 0:0964 0-0961 f
9. 0:0964 0°0961 +
10. 0-0994 0:0999 7
Nile 0:0994 0:0993 -
12. 0:0942 0:0948 ae
13. 0°0935 0:0932 -
14. 0:0931 0:0927 33
15. 0:0931 0:0925 Pa
16. 0°0944 0-0971 %
LT: 0:0959 0:0953 45
18. 0:0953 0:0942 ; Pie
19. 0-0960 0:0958 fe
20. 0:0955 0:0959 re
Total 1°3401 1:3398 14000
=95°7 /, =4b-7 "/,

It may be stated that we found no difficulty in obtaining consistent and
concordant results with the autoclave, provided that care was taken to see
that the time and temperature were kept constant and that the autoclave was
not opened till the pressure had fallen to zero. We allowed at least ten
minutes for this and then opened the valve slowly. Having set the auto-
clave for 117°C. it is desirable if possible not to use it for other purposes
necessitating different temperatures, during the time it is in use for estima-
ting creatine. It must be borne in mind that the velocity of reaction is
greatly accelerated as compared with that at 100° C. and therefore slight
differences of time and temperature produce much greater effects.

We may perhaps add also in regard to the water bath that our ordinary
rule was to immerse the flask in the water and use a reflux condenser.
Latterly we have returned to the use of the funnel and watch glass for
condensation and find it more convenient, but the flask should be immersed
in the water. If placed on the top of the water bath the results are less
consistent. Dr Rosenheim informed one of us (W. H. T.) that he had had
the same experience. .

464 W. H. THOMPSON, T. A. WALLACE AND

SUMMARY OF RESULTS.

1. For control of the bichromate standard solution used in the Folin
method of estimating creatinine and creatine we did not find the degree of
conversion of creatine into creatinine by boiling with normal HCl sufficiently
constant. We recommend for this purpose creatinine picrate as described
by Thompson.

2. For the estimation of creatine in weak solutions the best results were
got by using an equal quantity of N. HCl and boiling either on the water
bath for three hours or in the autoclave for 25 minutes at 117°C.

3. For the development of the colour in the Folin method the optimum
time and temperature in our hands proved to be 7 mins. at 15°-17°C.

4. The addition of too much alkali reduces the colour. For aqueous
solutions and urine the best results are given by the quantity recommended
by Folin, namely 5 cc. of 10°/, NaOH over and above what may be necessary
to neutralise the solution.

5. The range of proportional readings on the colorimeter scale should
not go lower than 7 mm. nor higher than 9mm. _ If it is necessary to compare
readings separated by a wider interval a correction factor should be used as
explained in the body of this paper.

6. The darkening of the pigments of urine which occurs on boiling with
normal HC] adds to the estimate of creatine-creatinine contained init. With
human urine the increase is slight (1-24 °/,), with dog’s urine as much as 10°/).

7. The presence of dextrose to the extent of 10°/, did not affect the
estimation of creatine. Phosphoric acid was not found to be superior to
hydrochloric in estimating creatine when dextrose is present.

8. The recovery of creatine from diabetic urine gave results lower by
5°/, than from aqueous or sugar solutions. Phosphoric acid did not prove
more useful for this purpose than hydrochloric. The smaller recovery is
probably due to the effect of aceto-acetic acid.

9. ‘The autoclave method gave results identical with those of the water
bath. The optimum time for urine and weak solutions of creatine was
25 mins. at 117°-120°C.

REFERENCES.

Benedict, F. G. and Myers, V. C. (1907), Amer. J. Physiol. 18, 397.
Benedict, S. R. (1912), J. Biol. Chem. 12, 73.

Chapman, A. C. (1909), Analyst, 34, 475.

Cook, F. C. (1909), J. Amer. Chem. Soc. 31, 673.

H. R. 8. CLOTWORTHY 46

Dorner, G. (1907), Zeitsch. physiol. Chem. 52, 225.

Dreibholz, W. (1908), Inaug. Dissert. (Greifswald).

Emmett, H. D. and Grindley, H. S. (1907), J. Biol. Chem. 3, 491.
Folin, O. (1904), Zeitsch. physiol. Chem. 41, 223.

(1906), Hammarsten’s Festschrift, ‘The Chemistry and Biochemistry of Kreatin and
Kreatinin.”

Gottlieb, H. and Stangassinger, R. (1907), Zeitsch. physiol. Chem. 52, 1.
Graham, G. and Poulton E. P. (1913), J. Physiol. 46, Proceedings, xliv.
Greenwald, I. (1913), J. Biol. Chem. 14, 87.

Grindley, H. S. and Woods, H. S. (1906), J. Biol. Chem. 2, 309.
Hehner, O. (1907), Pharm. J. 78, 683.

Hoogenhuyze, C. J. van and Verploegh, H. (1905), Zeitsch. physiol. Chem. 46, 415.
— —— (1908), Zeitsch. physiol. Chem. 57, 161.

Jaffé, M. (1906), Zeitsch. physiol. Chem. 48, 430.

Klercker, K. O. (1907), Biochem. Zeitsch. 3, 45.

Koch, W. (1905), Amer. J. Physiol. 15, 15.

Krause, R. A. (1910), Quart. J. Exp. Physiol. 3, 289.

Mellanby, E. (1908), J. Physiol. 36, 447.

Mendel, L. B. and Rose, W. C. (1911), J. Biol. Chem. 10, 213.

Rose, W. C. (1912), J. Biol. Chem. 12, 73.

Taylor, M. Ross (1911), Biochem. J. 5, 362.

Thompson, W. H. (1913), J. Physiol. 46, Proceedings, i.

Weber, S. (1908), Arch. expt. Path. Pharm. 58, 93.

Wolf, Ch. G. L. and Osterberg, E. (1911), Amer. J. of Physiol. 28, 71.

or

XLVI... NOTE. ON ‘THE IODINE. CONTENT (OF
FISH-THYROIDS.

By ALEXANDER T. CAMERON.

From the Department of Physiology and Physiological Chemistry,
University of Manitoba.

(Recewved Aug. Sth, 1913.)

The presence of iodine in the tissues of, and oils from, various fishes has
been known for a considerable time. It was found in herrings by Jonas
[1838], in the oil from the liver of Raia clavata by Girardin [1842], and in
crabs by Rieger [1853]. Barral [1877] states that iodine is always present
in the oils from fish-livers. Stanford [1883] analysed a large number of fish
oils, and invariably found iodine present. His analyses seem to have been
controlled carefully, although the amounts quoted are minute. Thus:

Cod liver oil (various samples) averaged 0:000322 per cent. iodine (0:000138—0-000434),

Dry cod fish contained 0:000829 __,, .;
Seotch herring, salted 55 0:00065 “is Ae
Scotch herring brine e 0:00012 x5 iF
Seal oil <8 0:00005 ar =

Cod liver oil dragées (liver residues from
which oil had been removed) contained 0:05637

Bourcet [1899], using analytical methods similar to those of Baumann,
found iodine present in a large number of different species of fish, in amounts
somewhat larger than those just quoted, ranging from 00007 °/, (Leuciscus
cephalus) to 0:024°/, (Merlangus carbonarius). Raia clavata contained
0:002°/,. It is uncertain from his published data whether Bourcet’s figures
refer to fresh or dried fish.

Baumann and Goldmann [1896] fed dried cod to a dog as a probable
source of iodine, and confirmed the presence of the element indirectly by
noting that such feeding increased the iodine content of the thyroid.

So far as I am aware, fish thyroids have not hitherto been tested for
iodine, although in them, as in mammals, its presence is to be expected, and
in larger amount than in other tissue. Some indirect evidence on the normal
presence of iodine in the thyroids of fishes is afforded by the results of

hea pree.

A. T. CAMERON 467

Marine and Lenhart [1910], who found that the so-called thyroid carcinoma
of brook-trout is severe endemic goitre, which can be cured by the
administration of iodine as in mammals. This would seem to indicate that
these thyroids function in the same way as matimalian thyroids, and hence
should contain the same constituents.

While it is still to be proved that iodine is an indispensable constituent
of the thyroid, there can be no doubt that it is almost invariably present.
Even in those cases where absolutely negative results have been recorded
(compare for example Roos [1899], who found that the thyroids of many
carnivorous animals contained no trace of iodine), the method employed—
that of Baumann and Roos [1895]—1s inconclusive for minimal amounts, since
it frequently yields too low figures, indicating that minute quantities might
entirely escape detection. (See in this connection Seidell [1911].) The
actual amount of iodine found varies considerably, and seems to depend
largely, if not entirely, on the diet. Roos’ results [1899] show that the
thyroids of herbivorous animals, whose food contains small quantities of
iodine, are richer in iodine than those of carnivorous animals, whose food
contains little of the element. For further literature in this connection see
Swale Vincent [1912].

There is direct experimental evidence that when iodine compounds are
fed the iodine content of the thyroid is markedly increased. Thus Baumann
[1896] found that the thyroids of dogs previously fed for some weeks
on lean meat (iodine free) contained little or no iodine. That of a dog
which had received sheep’s thyroids (9 grams) 14 days before death con-
tained 0°34 °/, (dried gland). Roos [1899] obtained a very similar result.
While a large number of analyses showed that the usual iodine content in
the dried thyroid of the dog averaged about 0°1°/,, that of a dog to which
potassium iodide had been administered for some time contained 0°35 °/,.
Baumann and Goldmann [1896] removed the left thyroid of a dog. Its
weight, fresh, was 1°92 grams, and the iodine content 0°06 milligram. After
the wound had healed, the dog was fed for 14 days on 14 pounds of dried
cod (see above). The right thyroid was then removed. It weighed 2°5 grams
fresh, 0°8 gram dry, and contained 2°9 milligrams iodine (0°36 °/,).

It might therefore be expected, taking into consideration the constant
presence of iodine in sea-water, that the thyroids of salt-water fishes would
contain a maximum amount of iodine. If the conclusions of Gautier [1899]
are correct, and the iodine in sea-water is largely present in organised matter,
it should be the more capable of assimilation and transmission.

The results given below afford evidence that unusually large amounts are

468 A. T. CAMERON

constantly present in some species, and hence lend support to the view that
the iodine content of the thyroid is a function of the iodine in the diet.

Through the kindness of the authorities we have been furnished by the
Marine Biological Association, Plymouth, with samples of glands of two
species of elasmobranchs, Raza clavata and Scylliwm canicula. In all, six
samples were received, each consisting of a large number of glands. These
had been obtained during the early spring of this year, and stored in absolute
alcohol during collection. The alcohol was subsequently evaporated, and the
residue added to the glands. The weights of fresh glands given below are
therefore only approximate.

The samples were dried in a steam-oven until constant weight was
attained, and were analysed by Hunter’s method [1910]. This has now been
tested by a number of investigators (see for example Seidell [1911]), and is
undoubtedly very accurate when appreciable quantities of iodine are present.
(In testing for minimal quantities of iodine in other work I have not found
this method completely satisfactory, and my results so far seem to support
Kendall’s conclusions [1912].)

In all, four blank determinations were carried out (one with each series
of analyses) using powdered fibrin as organic material; each gave an
absolutely negative result. A test of the accuracy of the method, with a
known quantity of potassium iodide, gave the following result:

Amount of iodine taken 0:000306 gram.
9 », found 0:000310 ,,
This is of a lower order of magnitude than the amounts shown below, and
in these the error is probably less. The data for the thyroids are seen in the
following table :

Weight (grams) Amount Amount Percentage
a= = = taken of iodine of iodine
Species Sample Fresh Dry (gram) (gram) (dry substance)

Raia (1 & 2) (10°5) 2°541 0-503 0:002208 0-439
clavata 0°503 0:002198 0:437
(Mean 0:438)
(3) (6:4) 1:053 0°513 0-001678 0°327
(4) (?) 1:294 0:544 0:001542 0:283
Scyllium (5) (4°5) 0°987 0°516 0:003712 0°719
canicula (6) (?) 4°393 0:504 0:005831 1157
0-524 0-006099 1-164

(Mean 1-160)

All the thyroids in sample 5 except two were from male specimens and
all those in sample 6 from female specimens.

A. T. CAMERON 469

The results may be contrasted with the maximum iodine content hitherto
observed in different mammals. In all cases the figures indicate percentage
iodine in the dried gland.

Maximum iodine content in thyrords.
Fish thyroids 1:160 per cent.

Dog “f 0-692 ~=C«, recorded by Marine and Lenhart [1909, 2].
Human ,, 0:588 si, be 5, Seidell [1911].

Stag BS 0°54 a3 or » Blum [1899].

Pig ss Orb31 SC, as » Seidell and Fenger [1913] (fat-free).
Sheep ,, 0°53 = sx », Baumann and Roos [1895, 1].

Beef - 0-477 =; a », Marine and Lenhart [1909, 1].
Goat ,, 028 ,, “ ,» Blum [1899].

The average iodine content can probably be taken as from 30 to 50°/,
of these figures.

While the thyroids of Raia clavata only contain a quantity of iodine
of the same order as these maxima, both samples of Scyllium canicula
contain more, and one much more iodine than any thyroid previously reported
on. This is the more remarkable since this figure is the average for a large
number of glands, while several of the other maxima (including the next
three highest figures) are for single individuals. The theory that the iodine
content of the diet plays a considerable, if not the whole, role in determining
that of the gland therefore receives strong support from these figures.

The different results from the two sexes in Scylliwm are possibly accidental.
No previous marked variation seems to have been noted.

I hope to examine other species shortly, and to contrast thyroid with
other tissue in what seems most probably the optimum condition for maximal
iodine content.

I wish to thank Professor Swale Vincent for his kind interest in this
work.

The research of which this forms part is being carried out in connection
with the Ductless Glands Committee of the British Association for the
Advancement of Science, and the expenses are being in great part defrayed
by grants from the British Association, and (to Professor Vincent) from the
Government Grant Committee of the Royal Society.

REFERENCES.

Barral (1877), Compt. rend. 84, 308.

Baumann (1896), Zeitsch. physiol. Chem. 22, 1.

and Goldmann (1896), Miinch. med. Wochensch. 43, 1153.
— _ and Roos (1895, 1), Zeitsch. physiol, Chem. 21, 481.

Bioch, v1 31

470 A. T. CAMERON

Baumann and Roos (1895, 2), Zeitsch. physiol. Chem. 21, 489.
- Blum (1899), Arch. ges. Physiol. 77, 70.

Bourcet (1899), Compt. rend, 128, 1120.

Gautier (1899), Compt. rend. 128, 1069; 129, 9.

Girardin (1842), Compt. rend. 14, 618.

Hunter (1910), J. Biol. Chem. 7, 321.

Jonas (1838), Annalen, 26, 346.

Kendall (1912), J. Amer. Chem. Soc. 34, 904.

Marine and Lenhart (1909, 1), Arch. Internal Med. 3, 66.

—— and Lenhart (1909, 2), Arch. Internal Med. 4, 440.
and Lenhart (1910), J. Exp. Med. 12, 311.
Rieger (1853), Jahresber. Chemie, 329.
Roos (1899), Zeitsch. physiol. Chem. 28, 40.
Seidell (1911), J. Biol. Chem. 10, 95.

—— and Fenger (1913), J. Biol. Chem. 13, 517.
Stanford (1883), Chem. News, 48, 233.
Vincent (1912), Internal Secretion and the Ductless Glands, p. 314 (Arnold, London).

XLVII. THE COMBINATIONS OF HAEMOGLOBIN
WITH OXYGEN AND WITH CARBON MON-
OXIDE. ok

By ARCHIBALD VIVIAN HILL (Fellow of Trinity College, Cambridge).
From the Physiological Laboratory, Cambridge.

(Received August 16th, 1913.)

In a previous paper Barcroft and Hill [1910] gave evidence to show that
in dialysed solution haemoglobin has its smallest possible molecular weight,
contains, namely, one atom of iron. In a later communication Hill [1910]
suggested that the dissociation curves of oxyhaemoglobin in the presence of
salts and carbon dioxide can be -calculated from the hypothesis that these
bodies tend to aggregate the large haemoglobin molecules into larger
molecules, which then combine with oxygen according to the equation

Hb, + nO, = Hb,(O2)n-

If y be the percentage saturation of the haemoglobin with O,, and « the
tension of the latter in the solution, it can be shown that this hypothesis
leads to an equation for the dissociation curve of the type

JES eer rae (1),
where X is the equilibrium constant and n is a whole number > 1.

This equation seems to suit all known dissociation curves of oxyhaemo-
globin with a very high degree of accuracy, as numerous published and
unpublished experiments of Barcroft [1913] and others (see e.g. Douglas,
Haldane, J. S. and Haldane, J. B. 8. [1912]) will show. In point of fact
does not turn out to be a whole number, but this is due simply to the fact
that aggregation is not into one particular type of molecule, but rather into
a whole series of different molecules: so that equation (1) is a rough mathe-
matical expression for the sum of several similar quantities with m equal to
1, 2, 3, 4 and possibly higher integers.

The basis of the suggestion lies fundamentally in the idea that the
31—2

472 A. V. HILL

oxygenated haemoglobin molecules consist almost entirely of the fully
saturated types

HbO,, Hb,(O,)., Hb;(O,)s, etc.,
rather than of the partially saturated types Hb;O,Flb,@;, iby. ete; sor
the observed dissociation curves in the presence of salts exhibit a double
curvature (i.e. are of an S-shape), differing therein from the dissociation curve
of dialysed haemoglobin, which latter is a rectangular hyperbola always
concave to the horizontal axis. The former dissociation curves start from
the origin almost—if not quite—horizontally, bend upward at first with
concavity facing the y-axis, pass through a point of inflexion, and then bend
inward again with concavity towards the #-axis. This main property of the
curves is simply represented physically by the idea that the aggregated
haemoglobin molecule, taken e.g. to be Hb,, is oxidised into the form
Hb,(O.), by combination with two oxygen molecules simultaneously, and
that the unsaturated molecule Hb,O, either does not exist at all or exists in
negligibly small quantities. For in the former case the dissociation curve

Kx?

is of the form y= 100 ere the tangent to which at the origin is y=0, Le.

the horizontal axis: and this approximates to the fact, observed experi-
mentally. While if the unsaturated Hb,O, exists in appreciable quantities

the dissociation curve involves an equation of the type y = B YT _ 4 ete., where
1+Ka

B is some constant less than 100, the tangent to which curve at the origin
is y= BKza, a line going out at a slope as in the dissociation curve of dialysed
haemoglobin, and not as observed when salts are present.

Now it seems at first sight almost unreasonable to suppose that Hb,
would combine only with 20,, i.e. with two molecules at once, and would not
oxidise partially at first by combination with the single molecule O,: that in
fact the oxidation goes on according to the equation

Hb, + 20, == Hb,(0,),
and not according to the equations

Hb, + O; == Hb,0;
and then Hb,0, + 0; Hb; (O); }

One could see no analogy for an equation of the former type when the latter
scheme was so obviously possible. That the latter scheme does not immediately
satisfy the facts is obvious from a short calculation, which however I believe
will lead us to the clue to the whole matter. In equations (A) above, if
K and K' be the equilibrium constants of the two reactions respectively, and

oe

Py

OM; HILDE, 473

if the concentration of Hb,O, be uw, of Hb,(O,), be v, of Hb. be w, and if « be
the tension of oxygen and y the °/, saturation we find from the laws of

mass-action
u v ,
—=K and —=—K’.
wr uc

Also since (w+2v) molecules of Hb are saturated with O,, and there are
(2u 4+ 2v+2w) molecules of Hb altogether,
y=100, u+2v

2u+2v+2w?

: Ka +2K' Ka?
al ) S = .
which becomes y = 100 5 oxi

when we put u=Kwe and v=K'eu= KK'e*w.
The tangent at the origin to this curve is
2y = 100Ka,

a straight line which is not the horizontal axis; thus the scheme represented
by equations (A) does not give a dissociation curve of the observed kind—
coming in horizontally to the origin—unless K is exceedingly small.

If however K is exceedingly small the difficulty entirely disappears, and

the equation becomes

which is of the type required to fit the dissociation curve, with tangent at
the origin y=0. (It should be noted that if K is small K’ must be large, so
that KK’ is finite.)

Can we therefore assume that K is small, that in fact oxygen combines
far more readily with Hb,O, than it does with Hb,? In other words, is it
a justifiable assumption that the partially saturated body Hb,O, is very
unstable, and exists in almost inappreciable quantities: that in fact Hb,O,
once formed almost immediately combines with another O, to form the
saturated compound Hb,(O.),? In support of this contention! one may argue
that there is no transitional spectrum corresponding to unsaturated oxides,
and therefore that, if there is any truth in the aggregation theory, the
unsaturated oxides are present only in negligible quantities. The theory
will obviously lead directly to a dissociation curve of the observed type, to
which by adjusting the value of KK’ it can be made to conform with some
degree of accuracy. So far then as we are concerned with dissociation curves

of oxyhaemoglobin, the equations to these can be deduced at once from the

1 First made to Mr Barcroft by Mr W. H. Mills of the Chemical Laboratory, Cambridge.

474 Ae We EEE

aggregation hypothesis with the conception of gradual combination with
oxygen according to the scheme

Hb, a5 O, ae Hb,0,,
Hb,0; + Opa Hb,,(05),,
Hb,(Oz)» +0,.== Hb,,(O,)., etc.,

provided always that we may assume that O, combines much more readily
with Hb, 0, to form Hb,(O,),, than with Hb, to form Hb,O,: in other words
that the partially saturated molecule is very unstable, is difficult to form and
easy to combine further with oxygen. For all such combinations lead to a
dissociation curve with double curvature, which reaches the origin §hori-
zontally.

A plausible suggestion as to the reason why O, should combine much more readily with
Hb,O, than with Hb, may be put forward here. The haemoglobin molecules are probably
aggregated by surface effects, by the neutralisation e.g. of the electric charges on their surfaces.
It may be however that the free chemical ‘‘ bonds” of Hb, with which reduced haemoglobin
attaches itself to oxygen, are in addition combined with one another in the aggregated molecule,
so that if reduced haemoglobin is Hb= then the aggregated molecule is not merely Hb, Hb
(two molecules just sticking together by surface forces) but rather Hb=Hb. In that case
Hb,0., the half saturated molecule, has to be formed by the breaking loose of these bonds, and
will be =Hb, Hb=O,. This body one would naturally expect to be very unstable, as is
Hb= in the presence of oxygen. It would be very difficult for O, to form this compound,
because it has to break down the Hb=Hb molecule into =Hb, Hb= in order to do it: and the
compound =Hb, Hb=O, once formed, presenting as it does two unsaturated bonds, would
immediately seize on another O, molecule becoming O,=Hb, Hb=O,. Whether this turns out
to be the case or not, it presents a very clear physical conception as to why Hb,O, should exist
in only very minute concentrations.

So far then the aggregation hypothesis can be made to fit the observed
facts very closely, and to have a reasonable physical explanation. Some
doubt has been thrown on the particular form of it advanced in my earlier
paper, by Douglas, Haldane and Haldane [1912]. These authors determined
experimentally the CO-haemoglobin curve in the presence of a partial
pressure of oxygen, and found it to be a rectangular hyperbola, According
to the rough scheme, dealing with single molecules,

HbO, + CO = HbCO + O,,
(and supposing the haemoglobin to be all combined with either CO or O,)
the equation to the CO-dissociation curve in the presence of a constant
partial pressure of O, should be that of a rectangular hyperbola. But
dealing with aggregated molecules, say Hb,, the formula

Hb,(O,), + 2CO = Hb,(CO), + 20,

leads, by the laws of mass action, to an equation ““, = K, if # and 2’ be the

”")

yal?

Buy .

A. V. HILL 475

partial pressures of O, and CO, and y and y’ be the °/, saturations with O,
and CO. If all the haemoglobin is combined y + y’ = 100, so that

oe eee

100 — y’ ee’

which represents a CQO-dissociation curve of the type described above,
containing a point of inflexion, and coming in horizontally to the origin.
Thus this, and as a matter of fact all similar conceptions, of the balanced
action between CO, O, and haemoglobin, led to the same type of equation,
which gives a curve in no way like a rectangular hyperbola: so that the
experiment of Douglas and Haldane led one to doubt the validity of the
whole hypothesis of aggregation, in spite of the accuracy with which it fitted
some of the facts. Why is it that the CO-O,-haemoglobin dissociation curves, in
the presence of a constant tension of O,, are unaffected by salts and CO,, when
the O,-dissociation curves and the CO-dissociation curves are so very largely
affected? This very puzzling question can, as a matter of fact, be quite simply
answered by the aid of the conception sketched above.

Let us assume for simplicity that the haemoglobin exists as double Hb,
molecules, and that according to the following scheme it combines gradually
and by steps with CO and O,. The Hb, combines first with one O, or with
one CO, the new molecule Hb,O, or Hb,CO combines further with O,, or CO,

to form one of the three compounds
Hb,(O,)., Hb,(O,)(CO), or Hb,(CO),.

We then have the equations (in which the small letters represent concen-
trations and the k’s represent equilibrium constants)
Hb, os O; a DAOs ky ’
Hibs 0, + 0, == Hb(0)); k, |
Hh.0, $00 = Hb,0, (CO) k, |
Ht, + CO = Hb,CO al eee (A).
Hb, CO + cos nes Hb, (CO)s ey

Hb,CO + 0, =® Hb(COKO,) ky

From these, by the laws of mass action, we find
E zy E, = = Ki
II. aE 4!) iL eae
IIT. Y=k,, "=k.

476 A. V. HILL

From equations I and IT, multiplying up, we find

TV, Begs hele, a = ky'ke’,

xy ayy)

and from equations I and III
VS rr =A ari =k,
so that h,k, = k,'k,’ = k say.
Let & be the amount of haemoglobin per cc. saturated with O,: & the
amount saturated with CO. Then :
E=k(ut+w)ty,
E=h(w'+w)t+v.
Using the values of u, v and w given in equations I, IV and V we find
E-bay tea) ee @)
and &=4(k/a'y t+ kav'y) + kykya’y
Ya Shi the) + hy hye!
Be Ge 4 (ky + ka’) + kykox
Thus apparently the °/, saturations of the Hb with O, and CO are not
in the direct ratio of the latter’s tensions, as 1s necessary if the dissociation

curve is to be a rectangular hyperbola. We come however to the hypo-
thesis outlined above, viz. that the unsaturated Hb,O, is very unstable and
tends to break down either into Hb, or Hb,O,. In this case we may assume
that k, is very small, and k, very large, their product remaining finite.
Similarly we have assumed that k,' is very small and k,' is very large. We
therefore find, neglecting k, and ky,

, ,

b kar + ky’ koa!
TE a essresestensneennnneen (C).

This, again, is not a direct proportion between &'/& and a'/z, as we should
expect from Douglas and Haldane’s observations [1912, see especially pp. 278,
290]. Looking however at the chemical equations (A) above, if we were
to assume that CO has a very much higher affinity. for the partially saturated
= Hb, Hb=0, (ie. Hb,O,) than has O,, ie. that nearly all the Hb,O,
combines with CO rather than with O,: and if we further assume that the
partially saturated = Hb, Hb =CO (ie. Hb,CO) has a much higher affinity
for the CO than it has for the O,, i.e. that nearly all the Hb,CO combines
with CO rather than with O,, then the whole difficulty ‘is immediately solved.
For’ in the first case we find that kx is very much greater than k,«, and

1 It should be noted that the assumption is not simply that Hb,O, has a greater affinity
for a given pressure of CO than it has for the same pressure of Oy: this, of course, we should
expect. It is that the affinity of Hb.O, for CO exceeds the affinity of Hb,O, for O, more than
the affinity of Hb for CO exceeds the affinity of Hb for O,: that the CO-affinity is even greater,
relatively to the O,-affinity, in the aggregated molecule than it is in the simple.

A. V. HILL 477

in the second that /,’a’ is much greater than k,’#. These assumptions are
in the highest degree reasonable, and making them, equation (C) above
, /

immediately gives the required result. Putting k=k/k, in the numerator

and k=k,k, in the denominator (as shown above) we find
Yat hy bhy'e-+ hye’
&” a ky 4kga'+kox
and assuming, as above, that /;’a can be neglected in comparison with k,‘a’,

and that k,a can be neglected in comparison with ka’, we find

That is to say, the CO- and O,-saturations of the haemoglobin are
proportional to. their partial pressures, and if the haemoglobin be fully
saturated with CO and O,, Le. putting €+€ equal to a constant (say 100)
we find, if 2h,'ky'/(kiks) = K, :

liege
100-2 7 XX.

This is a rectangular hyperbola between &’, the °/, saturation with CO,
and a’, the partial pressure of the CO. Thus on quite simple assumptions
the theory is found to agree perfectly with the experimental observations
of Douglas and Haldane.

Now Douglas and Haldane have brought to light some other experi-
mental facts which seem at first sight to be difficult of explanation. “They
found [1912, p. 279] that the CO-O;-haemoglobin dissociation curve,
in the presence of a fixed tension of CO and a variable tension of O,, was
a rectangular hyperbola for large tensions of oxygen but bent back and
fell again as the O,-tension became very small and gradually vanished.
Now, on our theory as advanced above,-a very pretty explanation of this
fact can be found, which may elicit some facts of further physiological
interest. The assumption made above is practically that CO combines much
more readily with Hb,O, in the presence of O,, than it does with Hb, in the
presence of the same O,. When the O,-pressure is reduced beyond a certain
limit very little Hb,O, is formed, and therefore the CO is forced to combine
with Hb, direct—which it does much less readily, and therefore the curve
falls again. An analytical expression of this fact is given below. In order to
find the dissociation curve in question we have to find the relation between &,
the CO °/,-saturation, and z, the O,-tension. We know that the total amount
of haemoglobin present can be represented by 100, so that, neglecting the

small amounts of the unsaturated compounds present,
2E + 2E’ + 2y = 100.

478 AL OVS FEE:

Putting into this equation the values of € and & in terms of y given in
equations (B) above, we find

Shy me 100
Y on 4 (kya + hy’x’) + hero’ + kykox® + ky/k,/x!2 ”

so that, from equation -
f= 100i "Kgl! +4 kyla! +4 kee’)
cv (kya + ky’x') + kaw’ + kykgu? + ky/ky'x’?*

This relation between * the °/,-saturation with CO, and a, the O,-tension,
is too complicated for simple calculation and comparison with experimental
results. It can however be shown that it may possess the general qualities
of the experimental curve found by Douglas and Haldane [1912, p. 279].
It rises from the horizontal axis in much the same way as their curve, with
diminishing O,-tension, and then reaches a maximum and falls again
slightly as the O,-tension becomes very small and finally vanishes. This
can be simply tested by finding the value of 0&/dx at «= 0, ie. the slope
of the curve at the point of zero O.-pressure. If 0£/dx is positive, 1.e. if the
curve slopes upwards at first, & first increases and then of course later
decreases as x increases. If 0£'/dxv is negative at « =0, then the curve slopes
down from the beginning. .

We should expect therefore, from Douglas and Haldane’s results, that
the value of 0£'/dx at «=0 should be > 0.

A simple calculation will show that

1 (5) _ ake'+ t 3 kky'x arf 2+ kk, ke, gr’ 3 — k phy! Tig'a!? — 3 ky key'a! — 2h ky’ ky’ a’ 3 — kky’a’?
100 z=0 (1 + kya! + 2hy’key'a’ 2)?

Let us assume that «' is very large, ie. that there is a large tension
of CO. Then (d£’/Ax),—9 is obviously negative, because of the preponderating
negative value of the term — 2kk,’k,'a’®.
curve will not exhibit a fall of CO °/,-saturation as the O,-tension is diminished
to zero, if the CO-tension is high. The assistance provided to the combina-
tion of CO by the presence of a little unsaturated Hb,O, is of no importance
if the CO-tension is very high. Let us assume that 2 is very small. Then
the value of 73, (G&'/0x),—9 is

Hence Douglas and Haldane’s

3 kav’ -—4 kyky'2’

(tho)?

neglecting squares or higher powers of a’. This quantity is therefore positive
ifk>hk,k. Nowk=kk,. Hence we must have k,>k/. This we should
expect to be the case. From the assumption already made, &,' is
very much greater than k,’, CO combines much more readily with the
partially saturated Hb,CO than with the completely unsaturated Hb,: one

AUVs HILL 479

would expect it therefore to combine much more readily with the partially
saturated Hb,O, than with the completely unsaturated Hb,. In that case
k, would be as much greater than k,’, as ki,’ is: and on the latter inequality
the whole of the argument of this paper is based. Without therefore intro-
ducing any new hypothesis we have shown that, for a small constant CO-
tension, the CO-O,-dissociation curve in the presence of an increasing tension
of O, rises at first, reaches a maximum and falls slowly to zero, as in Douglas
and Haldane’s recorded observations.

In all the above considerations Hb, has been taken as the general type
of the molecular cluster. Any other type of molecule, Hb,, Hb,, ..., can be
considered of course in exactly the same way: the mathematical treatment
would be more complicated, although exactly the same in type.

Douglas, Haldane and Haldane in their paper [1912, p. 296] put forward
an ingenious theory to account for the several dissociation curves observed,
both with O, and CO: and they gave an equation for a dissociation curve
which fits the experimental points fairly closely, but not so closely as does
the equation deduced from the hypothesis of this paper. In order to arrive
at this equation they made various assumptions as to the exact values of
the dissociation constants of the various reactions, for which, as they admit
[1912, p. 301], there are no real a priori reasons: and there are other
improbabilities in their assumptions, at least as great as in those made above
in this paper. For example the idea that “the aggregated molecules do
not give up or take up O, or CO without first splitting up into simple
molecules” seems difficult to admit: and it is not easy to understand, if
oxyhaemoglobin molecules are continually aggregating and reduced molecules
are continually aggregating, how the oxyhaemoglobin and: the reduced
haemoglobin molecules do not aggregate together to form unsaturated
clumps. It is true however that their theory does agree to a considerable
extent with observed facts, even though they have made many rather difficult
assumptions in the development of it: so that, even though the theory
sketched in this paper seems to me to be more probable, I do not feel that one
can as yet decide definitely between the two. It is noticeable however (as
Barcroft shows) that the equation to the dissociation curve given here, and in
the previous paper, does definitely suit the experimental facts better than
the equation deduced from the hypothesis of Douglas, Haldane and Haldane.
That a modification however of their assumptions would make their equation
fit the experimental facts better than at present, seems to me very probable :
so that a judgment between their theory and mine can scarcely be given on
the facts hitherto discussed. At present however the theory and assumptions

480 A. V. HILL

given here seem to me to be simpler and easier, as well as to fit the ex-
perimental data better, than the hypothesis advanced by Haldane, Douglas
and Haldane.

SUMMARY AND CONCLUSIONS.

The O,- and CO-dissociation curves of haemoglobin are known to
differ, according as salts and CO, are present or are not. A theory to
explain this has already been advanced, viz. that the simple molecules of
haemoglobin are aggregated into molecular clusters. This theory explained
many of the facts very exactly, but certain objections had been raised to it
in particular by Douglas, Haldane and Haldane, especially in relation to the
CO-O,-dissociation curves.

It is shown here that the theory is capable of including all the known
facts in relation to CO- and O,-dissociation curves, if one is allowed to make
certain simple assumptions as to the order of magnitude of the equilibrium
constants in the several reactions involved. These are:

Gi) that the half saturated molecules Hb,O, and Hb,CO are very
unstable, and change at once into either Hb, or Hb,(O,),, Hb,(CO), or
Hb,(CO)(O,), ,

(ii) that the half saturated molecules Hb,O, and Hb,CO combine
much more readily with CO than with O,.

The first of these assumptions can be explained as due to the fact that
Hb, is Hb = Hb, while Hb,O, is = Hb, Hb = O,, with two unsaturated bonds,
which tends to combine at once with more O, to form O, = Hb, Hb = O,.

If these assumptions are justified, we may deduce that since CO combines
much more readily with Hb,O, than with Hb,, haemoglobin will take up
more CO at a given tension, if a little O, is present than if O, is completely
absent.

REFERENCES.

Barcroft (1913), Biochem. J. 7. 481.

Barcroft and Hill (1910), J. Physiol. 39, 411.

Douglas, Haldane, J. 8. and Haldane, J. B. S. (1912), J. Physiol. 44, 296.
Hill, A. V. (1910), J. Physiol. 40, Proceedings, iv.

Pie ere

XLVIII. THE COMBINATIONS OF HAEMOGLOBIN
WITH OXYGEN AND WITH CARBON MON-
OXIDE. . II.

By JOSEPH BARCROFT.
From the Physiological Laboratory, Cambridge.
(Received Aug. 16th, 1913.)

In the preceding paper Hill [1913] has expanded his formula
y/100 = ne

in which y= the percentage saturation of the haemoglobin with oxygen, « the
oxygen pressure, AK the equilibrium constant of the reaction and n the
average number of molecules of haemoglobin in each aggregate. The formula
was originally put forward [Hill, 1910] to cover the dissociation curves of
haemoglobin in pure aqueous solution and in certain saline solutions. With
the introduction of certain assumptions it may now be applied (1) to the
affinity of blood for oxygen in the absence of carbonic acid, (2) to the affinity
of blood for carbon monoxide in the absence of oxygen, (3) the partition
of haemoglobin between oxygen and carbon monoxide in the presence of the
two gases and either with or without acids such as CO, in the system. The
validity of the theory depends not only upon the soundness of the reasoning
on which it rests, but upon the accuracy with which it fits the vast number
of experimental data by which it may be tested.

It may here be pointed out that the formula contains but two constants
n and k: the experimental test therefore is a much more crucial one than in
the case of the more adaptable formula of Douglas, Haldane and Haldane
* [1912], which contains three constants.

Since the formula was first published by Hill, I have devoted a great
deal of attention to the accuracy with which it fits the facts. The
correspondence is so striking that I propose to record it in the present paper.

I will treat of the evidence as regards

(1) the reaction Hb +O, = HbO,,
where the haemoglobin is dissolved in solutions of various salts,

(2) Hb+O, = HbO, with reference to blood.

482 J. BARCROFT

HAEMOGLOBIN IN SALINE SOLUTIONS.

In Hill’s preliminary communication he tabulated the comparison of the
curves calculated from the formula and the freehand curves drawn by Camis
and myself [1910] for solutions of haemoglobin in water, 0°7°/, NaCl and
0:9 °/, KCl.

The figures lost much of their force from want of a fuller presentation.
It was not evident from a casual study of these data that all the curves they
represented for which the value of n was more than unity were in essence S-
shaped. It therefore seemed that there was an essential difference between
these curves and those which were known to exist for blood, the latter being
S-shaped curves.

Blood.

Blood offers a far more searching test than haemoglobin solution because
the curves are much more spread out. There is no case on record in which
a great number of determinations, say more than about a dozen, have been
carried out on the blood of any one person at one time. Fortunately the
blood of the persons who have been most thoroughly investigated seems the
same at one time as at another!; a single curve may be taken as representing
the blood of one individual, for instance Douglas, Zuntz or myself, irrespective
of the time at which the determinations were made.

Thirty-eight determinations of points on Douglas’ normal dissociation
curve at his existing alveolar CO, pressure of 40-41 mm. have been made.
The series is especially valuable because the determinations have been made
by somewhat different methods, by different persons, and on different scales.
Rather more than half were made by me, and for all these the differential
method of blood gas analysis was used, dilute ammonia unaided by saponin being
employed in the blood gas bottles and in most cases 0°1 ¢.c. of blood being used
for each analysis. Sixteen of the points shown were determined by Haldane
and Douglas; they did not use the differential apparatus, but Brodie’s adapta-
tion of the old Barcroft-Haldane apparatus. They used 1 ¢.c. of blood for their
analyses. Five determinations were made with dilute sodium carbonate as the
fluid for taking the blood whilst the rest were made with ammonia. They
used saponin in addition to the alkali. Moreover, Haldane and Douglas
treated the gas in their tonometer somewhat differently from the way in which
I did. Their method involved them in much greater corrections for the
changes during analysis in gases held physically in solution. In spite of this
their method was probably somewhat more accurate than mine.

1 Barcroft [1911], Douglas, Haldane and Haldane [1912, p. 283].

J. BARCROFT 483

In treating of the data I have adopted two methods. The first is to take
all the data in order, to divide them into three groups as nearly as possible
equal, to state (1) the difference between the observed and the calculated
percentage saturation in the case of each observation, (2) the greatest
difference in each direction and the mean difference in the case of each
group and (3) the same in the case of the whole series. The following is
the result :

Pressure °/)sat. Dev. from Pressure °/)sat. Dev. from Pressure °/ sat. Dev. from
obs. obs. curve obs. obs. curve obs. obs. curve
9 6 +1 36 63 4+0°5 59 84 —]
1 6 +3 36 67 +45 61 83 -—3
17°5 26 +3 37 63 -1 61 84 —2
18 24 +1 38 65 -l 70 89 -0°5
20 27 -0°5 39 71 +3 77 92 +05
20 30 +2°5 39 67 -—l 85 92 - V3
23 38 +3 41 71 +1 85 94 +0°5
24 37 -l 42 75 44 90 92 —2°5
26 39 —3 42 72 +1 95 99 +4
29 51 +2 49 77 —1°5 97 94 —1°5
30 50 -l 49 79 +0°5 96 96 +0°5
34 57 —2 53 78 —3 104 96 0
59 81°5 —3°5 104 sO +1
Greatest error Mean
Total of 38 ee 0-14 Jy
— °O

The second plan of treating the results is to divide according to the
methods by which they were obtained. Taking the 16 results obtained by
Haldane and Douglas, eight are above the line and eight below it. The
greatest individual discrepancies are + 4°/, and —3°/,. The mean of the
sixteen is 0°19°/,. Of their sixteen results one differs from the calculated
number by 4°/,, three by 3°/,, three by 15-2 °/,, eight by 0°5-1°/, and one by
under 0°5°/,: whilst of the whole series of thirty-eight, there is one of
4°5°/,, three of 3°5—4°/,, nine of 2°5-3°/,, six of 1°5-2°/,, eighteen of 0-5-1 °/,
and one of under 0°5°/,. The general run of the figures is so nearly the
same in each of the two series, as to prove that the method of experiment
does not affect the comparison at issue.

Nine points have been determined on the Zuntz blood at 40 mm. CO,
pressure. I could not find that it differed from that of Douglas, it is
therefore referable to the same curve. Of these nine, three fall on the
curve (within 0°5°/, of it), four are below it and two are above it.

Oxygen pressure 0 12 30 37 37 43 61 79 S7
Observed °/) saturation 0 4 50 61 Gb 72 86:5 92 95
Divergence from curve 0 -5 -1 —4 —3°5 0 +°5 +1 0

Extreme divergence = Mean divergence -—1°3 7g.

5°

484 J. BARCROFT

10 20 30 40 50 60 70 80 90 {00° “te

Fig. 1. Ordinate=percentage saturation with oxygen ; abscissa =oxygen pressure.
x Blood of Douglas. m Blood of Zuntz. + Blood of Haldane.

Referable to this curve also is the blood of Haldane which I have taken
from the figure in the Journal of Physiology, vol. XLIV, page 283, as faithfully
as the scale of the figure permits. ‘The points are as follows:

Oxygen pressure 26 20:3 20°5 20°55 29 381 37 37 47 48 52°5 62 62 82 99 105
Observed %/y sat. 19 28°5 28 23 49 52 63 66 81 80 83 87 895 93 98 96

Div. from curve 0 -0°5 0 -5 0 +05 -1 +2 44 4241 O +425 41425 0
Extreme divergence Be . Mean divergence 0°6 9/9.

The results of the comparisons between the curve drawn from the
equation, and the actual determinations which I have cited may be summarised
as follows. (1) Of the 63 determinations none differs by more than 5°/,
from the calculated value, (2) the “average error” reckoning all departures
from the calculated value as positive quantities is 1°6°/,, (3) the mean of all
the errors reckoning those above the line. as positive and those below it as
negative is 0°02 °/,.

ee

rib

J. BARCROFT 485

Tt is known that the blood of different persons differs. I have compared
the points which have been determined on my own blood with the curve,
the constants of which are n=2°5, K =0:000292. Only about 20 points are
available for the comparison,

Dividing them into two groups they are as follows :

Pressure of oxygen 0 15 ii 13 22 28 36 36 37 38°5

%y sat. observed -1 2 4 12 37 55 69°5 71 68 72°5

Div. from caled. value —-1 41:5 0 -4 -2 0 0 +15 --2°5 +40°5
Extreme divergence a ie Mean —0°6 %J .

Pressure of oxygen 38°5 43 50 #50 50 50) 57 65 65 65 89

"ly sat. observed 74°55 795 83°5 83°5 84°5 85 89 89 90 90°5 95

Div. from caled. value +2°5 +1 -0°5 -05 +05 +1 0 -8 -2 -I'5 0
Extreme divergence | cei Mean — 0°23 9%.

Taking the two series together the results may be summarised as follows:
(1) no single determination differs by more than 4°/, from the calculated
value, (2) the “average error,” calculating all errors as positive, is 1:2°/,,
(3) the mean divergence, calculating points above the line as positive and
those below as negative, is 0°36 °/).

100

80

60

40

20

10) 20 40 60 80 100
Fig. 2. Zuntz! dissociation curve, 35 mm. CO,; ordinate=percentage saturation
with oxygen, abscissa=oxygen pressure in mm.
Of the cases of human blood which have been studied at all exhaustively
by me, the least satisfactory from the present point of view is that of Zuntz
at his existing alveolar CO, pressure of 34-35 mm. of CO,; here, however,

1 The data and pressure given on p. 62, Journal of Physiol. Vol. x11, are slightly erroneous
owing to a wrong correction for the zero of the gas burette.

Bioch. yu oe

486 J. BARCROFT

the chief cause of discrepancy is a point done in duplicate (pressure 19 mm.
percentage saturation 41 and 40°/,). The data are as follows. The curve
here is very steep, and a slight error in the pressure measurement would
account for it.

Pressure of oxygen 155 19 19 23 27 31 36 36 49 50 50 60 60°5 60°5 72

0/) sat. observed 21 40 40 50 58 63 75 68 83 86 83 85 92 88 90
Div. from caled. value —1 +6 +7 +443 0 +4 -3 -1 41-2 -5 +2 -2 -3

Of these fifteen points one is on the line, seven above and seven below it,
the average error counting all errors as positive is 3°/,, whilst the mean
counting errors above the line, as positive, and those below it, as negative, 1s
0°66 °/,.

So far as I know there are no other cases in which a sufficient number
of points has been determined and these spread sufficiently uniformly over
the curve to test the truth or otherwise of the theory.

At this point a few words may be said by way of comparison between the
curve obtained from Hill’s formula and that obtained from the formula
of Douglas, Haldane and Haldane. Over a great portion of the curves
obtained from the two formulae they agree so closely as to make an
experimental test between the one and the other almost impossible.

At pressures below 20 mm. the two curves differ and it may be well to
compare this portion of them with the observed points. This task is made
easier because at very low oxygen pressures the difference between the blood
of different persons is very slight and therefore the determinations—all too
few—for human blood may be taken together for the purpose. This may be
done the more freely by me because such error as there is in the application
of results obtained from my blood to the curves of Douglas’ blood which are
those depicted, would favour the formula of Douglas, Haldane and Haldane.

Below 15 mm. there appear to be eleven determinations on human blood,
of these the three lowest, at 0 and 15 mm. offer no test of the relative
merits of the two curves, yet I have included them because they are all
important to the argument as showing that the percentage saturations as
observed start from the zero. Had there been some such error as incomplete
laking of the blood or wrong correction for the gases in physical solution
these points would not have fallen on the base line.

Below are the data of the eleven points.

Pressure of oxygen ... oe
Percentage saturation observe

Divergence from Haldane’s curve
Divergence from Hill's curve ...

0 “15; 6 7 8 9. al aa
-1 2 |-1. 4 42 6° 33°)° 4p
-1 +41/-6 -3 -7 -35 +1 -10 -4 -35
-1 42)-25 41 -25 415455 -5 +1 +1

ooc co

OT ee oa

J. BARCROFT 487

Averaging the points between 5 and 15 mm., (1) the extreme departures
from Haldane’s formula are +1 and —10 and from Hill’s +5°5 and —5,
(2) reckoning the “average error” (counting all the errors as positive) it is
with regard to Haldane’s formula 4°8°/, and with regard to Hill’s 2°5%/,,
(3) reckoning errors above the line as positive and below it as negative, the
mean divergence from Haldane’s curve would be 4:5°/, and from Hill’s curve
1:2°/,. It seems clear from the figures and more clear from Fig. 3, that Hill’s
formula more closely represents the facts than does that of Haldane. Eight
points is however a small number to take for such a purpose and while it is

K: r n

1+ Ka"
many more points would have to be determined before the argument could

evident that the equation y/100= has the better of the comparison,

be regarded as settled on this count.

30

25

20

15

10

Fig. 3. Lower portion of dissociation curves of human blood. The upper curve is drawn
from Douglas, Haldane and Haldane’s formula, the lower curve from Hill’s formula.
Ordinate= percentage saturation with oxygen; abscissa=oxygen pressure in mm.

It may be taken as proved that either by accident or otherwise the
curves for human blood at normal alveolar CO, pressure closely follow Hill’s
equation, which is founded upon a physical conception. The natural test
is to see whether curves obtained under widely different conditions also
agree with the formula.

32—2

488 J. BARCROFT

The simplest way of varying the curves is to alter the concentration of the
carbonic acid. In this way a whole series of curves may be produced as was
shown by Bohr, Hasselbalch and Krogh [1904]. Poulton and I [1913] investi-
gated the question of whether three curves of the series other than the normal
curve at 40 mm. pressure could be represented by Hill’s formula. The answer
was striking for not only were the curves capable of expression by the general

Tego
formula y/100 =, ae

constants (A) varied, n remaining at a constant value throughout. Hence it

but over the whole series only one of the two

is possible from a graph relating the concentration of carbonic acid to the
value of K to obtain the dissociation curves for concentrations of carbonic
acid intermediate between those at which the determinations were actually
made. As this has proved useful on many occasions I append the diagram.

ACee SSOP eRe eae

Bree
ae

Speceeeeaee
J

FC
SAE

ae

—_—

Fig. 4. Ordinate=CO, pressure in mm. Abscissa=K.

It was shown by Orbeli and me [1911] that carbonic acid had no specific
effect in changing the affinity of haemoglobin for oxygen, but that it shared
this property with lactic acid: Mathison subsequently showed that other
organic and inorganic acids had a like effect.

|
t
:
:
®
t
:

J. BARCROFT

489

The following points were obtained by the addition of lactic acid (0°2°/,)

to blood from which the carbonic acid had been as completely as possible

removed by shaking.

Pressure of oxygen Bee $e 10 20 20 30 30 40
Percentage saturation observed ... 11 39 40 64 68 82
Divergence from calculated results +1 =] 0 0 +4 +2

n=2°5, K=0-000363

50

TRY
me
Ty
H/o
YZ

take
| Paes | | IAM

ari Sales e
O°

Z

*24 °32

| WV
HECeW

ie i Se ee eS

-40

50
86

Fig. 5. Equilibrium curves of CO haemoglobin drawn from Hill’s formula. The points are the
observations of Douglas, Haldane and Haldane. Ordinate=percentage of haemoglobin

with CO; abscissa=CO pressure in mm.

In addition to this, curves might be cited in which both lactic and

carbonic acid were added to blood, such curves for example as those given by

me in the record of my work in Teneriffe and by Orbeli and myself on the

effects of low oxygen pressures on animals. The points on curves all follow
the equation with the same degree of exactitude as those cited in this paper,

but the points determined in each case were too few to make any detailed

discussion of them worth while.

490 J. BARCROFT

“Most of what has already been said as regards oxyhaemoglobin curves
might be repeated with’ regard to CO-haemoglobin. Fortunately a very
complete series of data on this subject has recently been published by
Douglas, Haldane and Haldane. In Fig. 5, these data are shown and
along with them a system of curves drawn from Hill’s formula. The scale
on which the diagram is plotted is such that the centre point marked A
of the curve corresponding to 40 mm. CO, would be on the same place in
the figure as the centre point of the oxyhaemoglobin curve corresponding to
the same CO, pressure. That being so it will be found that the whole
figure is superposable upon the same figure for oxyhaemoglobin. Not only
are the points then faithful to Hill’s formula, but on a different scale the
relation of K to the CO, pressure is the same in the case of oxy- as of CO-
haemoglobin.

There is one more point to which I should draw attention, and which is
involved in the fidelity of the observed points to the theoretical lines. Did
the lines in Fig. 5 actually conform to Haldane’s formula, there would have
been the same slight discrepancy between the lines and the points at the
bottom of the diagram in the case of CO-haemoglobin as in the case of oxy-
haemoglobin and to just the same extent. When this difference reappears
in another series such as the CO-series in which the points have been
arrived at by wholly different methods (these were obtained by the method
of carmine titration) the weight of evidence in favour of their having a real
significance is greatly increased.

CONCLUSION.

1. The available data for the dissociation curves of blood agree very
closely with the theoretical curves deduced from the following physical
conceptions :

(a) ‘That the reaction between haemoglobin and oxygen is a reversible
chemical change Hb, + nO, == Hb,Onn.

(b) That » is the average number of molecules aggregated together,
the value of n depending upon the nature and concentration of the electro-
lytes in the solution.

(c) That the effect of acids is to change the equilibrium constant of the
reaction without sensibly altering the degree of aggregation of the molecules.

(d) That the above reaction does not involve the breakdown or
reformation of the aggregates.

(e) That unsaturated oxides are unstable and break up into
haemoglobin and saturated oxides.

J. BARCROFT 491

2. The available data with regard to the reaction of CO and oxygen
support an entirely similar conception of eeamenote

3. So far as the curves deduced from the formula y/100 = (Hill)

ae n
ean be distinguished from those yielded by the formula

=e (Douglas, Haldane and Haldane),
(1l-y)(l+ay)

the experimental evidence leans towards the former.

REFERENCES.

Barcroft (1911), J. Physiol. 42, 47.

— and Camis (1910), J. Physiol. 39, 118.

— and Orbeli (1911), J. Physiol. 41, 355.

— and Poulton (1913), J. Physiol. 46, Proceedings vi.

Bohr, Hasselbalch and Krogh (1904), Skand. arch. Physiol. 16, 402.
Douglas, ©. J., Haldane, J. S. and Haldane, J. B. (1912), J. Physiol. 44, 275.
Hill, A. V. (1910), J. Physiol. 40, Proceedings iv.

(1913), Biochem. J. 7, 471.

Mathison (1911), J. Physiol. 43, 347.

XLIX. SEPARATION, OF PROTEMS:
PART III. GLOBUEINS’.
By HENRY COBDEN HASLAM.

(From the Pathological Laboratory, Cambridge.)
(Received July 27th, 1913.)

Up till the middle of last century it was generally held that the protein
of blood serum and other kindred fluids was a single homogeneous substance
—albumin. The word globulin, apparently first used by Berzelius, was applied
by him to two substances, the protein part of haemoglobin (called by him
haemato-globulin) and the protem of the lens, and came thus to be associated
with cell-protein. That the serum protein could be split up, or that serum
contained more than one protein was first suspected on account of some
experiments of Liebig, Zimmerman and others, who found that great dilution
of blood-serum produced a precipitate, especially after neutralisation with
acetic acid. It was, however, Panum [1851] of Copenhagen who first
discovered serum-globulin, and who developed the now classical method of
preparation by first diluting the serum with water, and then adding acetic
acid. Panum showed that the substance occurs constantly in human blood
both in health and in disease, and called it serum casein from its resemblance
to milk casein, Shortly after Zimmerman [1854] published the carbonic acid
gas method of preparation. A few years later Alexander Schmidt [1862]
in the course of his extensive observations on serum and kindred fluids showed
in all cases that, side by side with the more soluble albumin, there was always
the more insoluble globulin. In pursuance of his views on blood clotting he
called the globulin in serum fibrino-plastic substance, though as often as not
in his writings he calls it globulin.

Considerable activity followed on the publication of Schmidt’s views:
several ways of producing protein precipitates im serum were discovered and
discussed ; and there was some little confusion as to whether there was more
than one substance, and as to nomenclature, till Heynsius [1869, 1876] gave

1 Part I Haslam [1905], Part II Haslam, [1907].

H. C. HASLAM 493

good reasons for believing that, whether precipitated by simple dilution,
carbonic acid gas, dilute acids, or saturation with sodium chloride there was
only one substance—para-globulin as it was then called—a conclusion that,
except in regard to sodium chloride, still holds good.

In 1878 Hammarsten [1878] published his great attempt at the purifica-
tion and estimation of this substance. He introduced magnesium sulphate
which produced a much larger precipitate than the reagents hitherto employed.
In 1883 Burckhardt [1883], in repeating Hammarsten’s experiment showed
that, in addition to the globulin which was insoluble in water, a water-soluble
substance was contained in the magnesium sulphate precipitate. Thus the
probability of the existence of a third protein in serum was shown, though
Hammarsten maintained that the magnesium sulphate precipitate was a
single substance. Burekhardt’s observation was, however, confirmed by
Marcus [1899].

Meanwhile Hofmeister had conceived the idea of dividing serum and
other liquids into fractions by means of taking precipitates at ditferent
degrees of concentration of one and the same salt solution, and thus isolating
the different proteins. The first attempt was carried out by Kander [1886]
when he precipitated globulin by half-saturation with ammonium sulphate,
and albumin by complete saturation of the filtrate, the globulin roughly
corresponding to Hammarsten’s. Later on this system was elaborated by
Pick [1902], Fuld and Spiro [1900], and others, both in the case of albumoses
and serum. It was at first found that the globulin brought down at half-
saturation could be split into two portions. The first was found to resemble
the original water-insoluble substance of Panum, Schmidt, Heynsius and
others, and was accordingly named eu-globulin ; the second was more soluble
and was called pseudo-globulin. Ina further research Porges and Spiro [1903]
thought there were three distinct fractions, both by ammonium, sodium, and
magnesium sulphate; and Reiss [1904] also in Hofmeister’s laboratory,
decided on three—eu-globulin, pseudo-globulins I and I. Freund and Joachim
[1902] on the other hand, taking the two fractions eu- and pseudo-globulin,
showed that each contained a water-soluble portion, and took the view that
there were four globulins. In regard to this system of fractionation I showed
by direct experiment in Part I [1905] of this series that no single precipita-
tion, both in the case of albumoses and serum proteins, whether by acid, salts
of heavy metals, salting out or alcohol, ever produced a complete separation :
that in the precipitate the substance of the filtrate could always be demon-
strated, often up to 20 or 30°/,; and similarly, mutatis mutandis, with the
filtrate. And Wiener [1911]; working on quantitative lines, showed that

494 H. C. HASLAM

reliable estimations of globulin could not be made by single precipitations.
J. Mellanby [1907] from a determination of the percentage of protein at
gradually increasing concentration of aleohol drew the conclusion that there
were three different proteins in serum. Fractional methods, then, seemed to
have left the subject in a more confused state than that in which they found
it. I believe, however, and hope to show in these pages, that the method
is inherently sound, and must for the present be regarded as one of the most
important, and in many cases the only, method, we have of separating proteins.

Fractional precipitation.

The fundamental observation on which the method rests is that when
a precipitant is added gradually to a protein fluid, and the resulting precipi-
tate collected in successive portions, it is found that one portion differs from
another. It is inferred from this that a separating process has been set up.
And the proof of the correctness of this inference is found in the fact that if
the process is continued sufficiently, substances that are undoubtedly distinct
from each other can be obtained. Now in the case of the precipitant being
a salt such as ammonium sulphate each increment of the salt in the protein
fluid is followed by an increment of the precipitate. It is possible that at
some concentrations the precipitate falls rather more thickly than at others:
and indications may thus be afforded as to how a separation may be attempted.
But it is entirely fallacious, as I have already pointed out, to suppose that
one protein is wholly, or even nearly wholly precipitated before the next
begins to come down. If there are two or more proteins their precipitation
commences almost if not quite simultaneously, and they continue to come
down together till the end of the precipitation, though at any one time
different quantities of each might be coming down. Mellanby has de-
monstrated this point in the case of serum and ammonium and magnesium
sulphate, having made quantitative experiments and plotted curves therefrom.
He finds the precipitate falls very uniformly and that at no point is there
a cessation of precipitation. His conclusion, however, that no splitting up of
the serum can be brought about by these means, is erroneous. If no
separation were brought about by the salting out, the protein in any one
fraction would resemble that in every other. But this is obviously not the
case. The chief cause of confusion hitherto has been of the opposite kind ;
that there has been no way of determining in the case of differing fractions,
whether each fraction connotes a separate substance or not.

Let us suppose that there are two substance A and B; that A in a pure

H. C. HASLAM 495

state is precipitable more easily and is all precipitated at half’ saturation ;
while B, in a pure state, does not commence to come down till the half-
saturation point is reached, and only comes down completely on full saturation.
Suppose now that three successive precipitates are taken at one-third, two-
thirds, and full saturation. The portions would be composed as follows :

Fraction 1. n, parts A, ms parts B.
2. ny parts A, my parts B,
3. ng parts A, m, parts B.

where ny >Ng>nz and my>mMy>Mz.

Each fraction would thus show some difference from every other. Let us
suppose that each fraction is redissolved and again precipitated at the same
concentration. The three fractions would then differ from each other more
markedly. The first would contain a greater proportion of A ; the last would
contain a greater proportion of B; while the middle one would remain about
the same, much the greater part coming down between the precipitation
limits. Generally speaking the separating power of the salt is not sufficient
to cause any great change at any one precipitation. Thus far we might
consider that we were dealing with three separate substances.

There are two principle methods by which to determine whether a
fraction represents a substance or not.

1. Constancy of quantity under repetitions of the process. Fractions 1
and 3 would show this constancy after a time. Fraction 2 consisting of
a mixture, would not, but would gradually disappear.

2. Subdivision of a fraction to find whether it is consistent or not.
Fraction 2 on subdivision would show that it consisted of substances properly
belonging to fractions 1 and 3. At half-saturation, in short, 1t can always be
divided into a precipitate containing a higher proportion of A than B, and a
filtrate containing more B than A.

Having decided on the number of substances into which the parent body
can be split, there next arises the question as to how far each can be separated
from the other. In the case of A, after a certain number of precipitations,
the filtrates contain a constant quantity of organic nitrogen: that is to say,
the separating process has entirely ceased. The separation, therefore, 1s
presumably complete, but I will return to this point later. In regard to B,
when we arrive at the point that half-saturation produces no precipitate, we
cannot infer that the separation is complete. I have shown in previous

1 As in previous papers I follow the usual convention of describing as ‘“‘half” saturated a
_ solution made by mixing equal volumes of saturated salt and protein solution. In reality such
a solution is less than half saturated.

496 H. C. HASLAM

papers that, if a further small addition of salt be made, the resulting
precipitate will consist largely of A; that is to say, a quantity of A remains
dissolved in the B fraction. Means must be adopted, therefore, to continue
the separation. Hitherto, the procedure has been to take a small fraction,
dissolve it, and reprecipitate at the same concentration. The precipitate
consisting mostly of A, is withdrawn, while the filtrate is returned to the
main solution. This may be continued until no further trace of A can be
found in the fractions. The separating process, then, comes to an end in this
case also. But although the separating process has come to an end we cannot
assume that the substances are completely separated, because in analogous
separations in fractional distillation, crystallisation and precipitation it may
happen that a certain amount of the substance being got rid of remains with
the substance being purified, the two together, in this instance, acting as one
substance towards the separating agent. Proof positive, therefore, can only
be obtained by means of independent reactions. In the case of most proteins
these, at present, are few in number: it is obvious that some separation must
precede the discovery of typical reactions based on constitutional differences.
In one of the separations to be described, we have such a reaction, and it goes
far, I think, in demonstrating the validity of fractional precipitation in the
case of serum proteins.

In regard to the question of the chemical individuality of the products of
such separation, it can only be said that it is convenient to regard them as
individuals until they are shown to be capable of further subdivision, Each
can be tested with all the means at our command.

First Separation.

In the case of serum (ox-serum was used in all these experiments) my first
procedure was to divide it by means of half-saturation with ammonium
sulphate. The serum was diluted some four times with water, and an equal
volume of the saturated salt solution was added. The resulting precipitate
was collected, redissolved in water, and the process repeated until the
separation, as shown by the sulphuric acid decomposition test!, was complete.
The separation was an easy one at this concentration; three or four
precipitations got rid of all but a small quantity of albumin, and the

1 8-10 c.c. of the filtrate are mixed with an equal volume of concentrated sulphuric acid in a
test-tube and warmed to boiling point. As the protein in the filtrate diminishes the tint gets
lighter and when two successive filtrates give the same tint the amount of protein in the
filtrates has become constant and the separation is at an end.

ee oe ee

H. C. HASLAM 497

separation was completed in some six precipitations. The globulin fraction
so obtained was almost entirely soluble in salt and water.

I then sought to divide it further by ammonium sulphate. Precipitation
at one-third saturation (one volume of saturated solution being added to two
volumes of the globulin solution) was carried out, and in this way the protein
was divided into two fairly equal portions, the greater portion, perhaps, being
that precipitated. The precipitate was re-dissolved and re-precipitated ; and
the process was repeated until the filtrate showed only small quantities of
protein. During the progress of the precipitations, a larger and Jarger
portion of the precipitate became insoluble in salt and water, and the
experiment, on this account was not continued to a point of constancy in the
usual way, since the insoluble matter might have held to itself soluble
protein and so have vitiated the result. It served to indicate, however, that
the globulin of half-saturation could be further divided into a water-insoluble
part, precipitable at one-third saturation, and a more soluble portion not so
precipitable. The latter was then treated by the process of fractional
precipitation to remove any water-insoluble globulin that had remained
dissolved in it. Powdered ammonium sulphate was added gradually and
dissolved by stirring until a small precipitate, say some 15-20 °/, of the
amount of protein present appeared. This was filtered off, redissolved in
water, and saturated salt solution was added till the solution was at “one-
third” saturation. This caused precipitation of a good proportion of the
fraction. Examination showed it to consist largely of water-insoluble
globulin. The filrate from the precipitation was returned to the main liquid,
and the whole process was repeated some five times, the last fraction showing
only quite a small amount of globulin. A further small amount of globulin
was removed by dialysis.

The resulting substance was easily and completely soluble in water, and
could be precipitated at half-saturation with ammonium sulphate. It was
tested for albumin. A portion was dissolved in 50 ¢.c. water and precipitated
by the addition of 50 c.c. saturated salt solution. This precipitate was
redissolved and re-precipitated in the same way at the same volume!

90 c.c. of Ist filtrate gave by Kjeldahl 1-0 c.c. N/10 NH;,
GOiciens,,, ands ae. Pe a O:9b7e.cn. ,, 3

The small amount of protein in the filtrate, being constant in « juantity, is
presumably pseudo-globulin.

It was mostly precipitable by saturation with sodium chloride and nearly
completely by magnesium sulphate. With ammonium sulphate most of it

* For details of this method see Haslam [1905].

498 H. C. HASLAM

(say five-sixths) was precipitated by the addition of only three-quarters of a
volume of saturated solution (43 °/, saturation). Half-saturation precipitated
all but traces. This, then, was the pseudo-globulin of Hofmeister, the
existence of which was indicated by the experiments of Burckhardt and
Marcus. Its solubilities are intermediate between those of the water-
insoluble globulin and albumin, .

I next proceeded to ascertain whether this substance could be split up
further by ammonium sulphate. Freund and Joachim found that the
pseudo-globulin of Hofmeister was partly soluble, partly insoluble in water :
but the insoluble portion was probably merely globulin left dissolved, as we
have seen, in the pseudo-globulin. It was possible, however, that it could be
split into two soluble bodies after Spiro and Reiss. A portion was divided
into two roughly equal parts by ammonium sulphate in the following way.
A dilute solution (about 0°25 °/,) was made. A small test portion was com-
pletely saturated with salt and placed ina cylindrical beaker. Saturated salt
solution was then added to the main liquid until a small amount, double the
volume of the test portion, placed in an exactly similar cylindrical beaker,
showed on looking downwards the same degree of density of precipitate.
I then assumed that about half the protein was precipitated. This, after
being allowed to stand, was filtered off and the filtrate saturated to obtain
the remainder. These two fractions closely resembled each other in their
solubilities. Each was largely precipitable on the addition of three-quarters
its volume of saturated salt solution, and the portions left over in solution
appeared in no way different from those precipitated. Nor did further treat-
ment on fractional lines show that the substance could be divided.

Since we had a water-soluble globulin commencing to fall at 33 °/, satura-
tion, and nearly finishing at 43°/,, that is to say when free or nearly free
from other proteins, it might be supposed that a substance existed that
commenced to fall at 43°/, and was mostly precipitated at 50°/, saturation.
It might be supposed that if such a substance existed it might be nearly or
entirely lost in the series of precipitations at 50°/, in which albumin is got rid
of. To test this point, I began afresh with some serum and endeavoured by
suitable fractionations to find such a body; but entirely without success.

Eaperiment. 300 c.c. ox-serum were taken and diluted to 1000 cc. and
500 c.c. saturated salt solution were added to get rid of some of the globulin.
To the filtrate were added some 700 c.c. (over half-saturation) further of
saturated salt solution, so that most of the searched for substance should be
precipitated while as much albumin as possible should be left in the filtrate.
The precipitate was dissolved again in 1000 cc. water and the above

H. C. HASLAM 499

precipitations were repeated. The protein resulting was then dissolved in
900 c.c. water and 500 c.c. of saturated salt solution were added so that not
only globulin but pseudo-globulin might be largely precipitated also. To the
filtrate was added salt solution again to make rather more than half-satura-
tion. I proceeded in this way to eliminate gradually portions of globulin and
pseudo-globulin, on the one side, and albumin, on the other; always being
careful to work well outside the precipitation limits of the body for which I
was looking, so that as little as possible might be lost. No such substance,
however, could be found. The process was continued until a quantity too
small to work with remained, and up to the last, pseudo-globulin on the one
hand and albumin on the other could be demonstrated. This experiment was
repeated on another sample of ox-serum.

I concluded, therefore, that no substance exists in ox-serum precipitable
between 43 and 50 °/, salt concentration, at any rate in quantities comparable
to globulin and albumin.

Water-insoluble globulin.

As it was not possible to carry to completion the separation of globulin
by means of ammonium sulphate, owing to its becoming increasingly insoluble
in the salt solution, I tried precipitations with acetic acid. Some water-
insoluble globulin that had been prepared from ox-serum by a_ few
precipitations at one-third saturation and then diluting with water, was
shaken well with water, thrown on the filter, and washed with water. It was
then suspended in water, and a few drops of ammonia were added. The
globulin completely dissolved. The volume of the solution was 400 c.c.
About 0°65 g. of globulin was recovered at the end of the experiment, so that
if we allow for some 20°/, of other proteins and loss during the experiment,
there would be some 0°8 g. protein. Thus the strength of the solution was
about 0:2 %/,.

The globulin was then precipitated by the addition of dilute acetic acid.
The precipitation began when the solution was about neutral, and was rapidly
completed on the addition of further acid, the solution being finally faintly
acid. The mixture was allowed to stand over-night. The precipitate was
then filtered off, suspended in water, and redissolved by the addition of
ammonia. Normal solutions of ammonia and acetic acid were used, the
quantities required being added out of burettes. Before precipitation the
volume of the solution was always made to 400 cc. 16 c.c. alkali were used
to dissolve: the precipitate was usually complete on the addition of some.

500 H. C. HASLAM

13 to 14 cc. of acid, but 16 were added. On one occasion a small excess was
added, but no re-solution of the precipitate occurred. It was noted that the
globulin took more time to dissolve as the experiment advanced.

In the second filtrate the presence of water-soluble protein was easily
demonstrated. The fourth and fifth filtrate each showed a faint cloud on
boiling, and a faint precipitate on saturation with salt. In each, however, by
saturating the whole with salt, and collecting the small quantity of protein
obtained, water-soluble protein could be demonstrated. The eighth and
tenth filtrates were examined quantitatively for organic nitrogen by the
method described in previous papers.

Kighth filtrate volume 345 c.c. gave 2°4 mg. N.
Tenth 33 5 343) CC eas ee:

Thus in precipitating the water-insoluble globulin by. acetic acid a point
can be arrived at in which the organic nitrogen in the filtrate is constant
and the soluble pseudo-globulin presumably completely eliminated. Reckon-
ing that globulin contains 16°/, N, the amount in these filtrates would be
0:0073 g. or 0:004°/,. It may be noted that the separation of globulin from
pseudo-globulin is considerably more difficult than the separation of pseudo-
globulin from albumin.

Physical Changes caused by Separation.

The substance thus obtained is distinctly less soluble than that prepared
more rapidly by a ‘smaller number of precipitations. Some of the essential
peculiarities of globulin are, however, preserved. The addition of alkali
causes it to swell up into a jelly-like consistence. It is, on the other hand,
only slightly soluble in salt solutions. But the question of the alteration of
the physical properties of proteins on precipitation is a difficult one.

Globulin prepared by a single precipitation with acetic acid from diluted
serum is a comparatively soluble substance. It also has mixed with it
considerable quantities of pseudo-globulin and albumin. The more it is
precipitated, the more insoluble does it become, and it is very generally
believed that this is due to a physical alteration. But it is not necessary to
‘postulate physical alteration, at any rate, to the degree generally done.
Globulin can be held in solution by the other proteins of serum. This can
be shown by taking a portion of serum-protein, by precipitating with
ammonium sulphate at half-saturation twice or thrice and dialysing until all
the salt is removed. A certain amount of globulin will be precipitated ; but
no matter how long the clear solution from which the precipitate may be

oo

H. C. HASLAM 501

removed is allowed to dialyse, water-insoluble protein in large quantity can be
demonstrated therein. Or, more simply, serum may be dialysed indefinitely
and after filtering off the resulting precipitate, large quantities of globulin
can be found in the filtrate. Again, during the course of fractional experi-
ments, globulin is commonly found in solutions half-saturated and more than
half-saturated with ammonium sulphate, there being excess of other protein
present. Indeed, it is this property that makes the whole difficulty of the
separation. It is to get rid of the globulin that remains dissolved in the
pseudo-globulin that the many precipitations at high salt concentration have
to be resorted to. It is this property that makes it impossible by any one
precipitation by any agent to obtain more than a proportion of the globulin
present. Pseudo-globulin and albumin are, in fact, the chief solvents of
globulin in serum. But in this case it is clear that in any series of precipita-
tions for its preparation and purification, where it loses at each precipitation
a portion of the other protein, it loses a solvent. The more the other proteins
are withdrawn from it, the more insoluble must the preparation become.
Quite apart, therefore, from any changes that may go on in the globulin
as a result of the physical action of the reagent used, we see that any prepara-
tion must become more insoluble as it becomes purer. But there is little
doubt that some of the various reagents used may also cause insolubility.
This is evident from the different behaviour of globulin when precipitated by
acetic acid or ammonium sulphate. With ammonium sulphate, as we have
seen, globulin gradually becomes so insoluble as to resemble coagulated
protein. It may be noted that this change does not occur uniformly through-
out the globulin, but a small portion of such insoluble matter appears, and
continues to increase as the precipitations proceed. We may therefore argue
that, apart from its separating power, ammonium sulphate has an action on
globulin tending to render it insoluble. With acetic acid such action is
extremely small ; after many precipitations only quite a small proportion of
this very insoluble matter appears.

But although it is quite easy to prepare solutions in which globulin is
held in solution by the other proteins, it is apparently difficult to reproduce
this condition when the various proteins have been separated from each
other. At any rate, it cannot be produced by the simple addition of globulin
to a solution of albumin or pseudo-globulin. I have no quantitative experi-
ments on this point, but it is quite clear that only small quantities, if any, of
completely separated globulin dissolve directly in watery solutions of
separated albumin or pseudo-globulin. So far as my experiments have gone,
it seems to be the case that, after a small degree of separation, say two or

Bioch,. vu 33

502 H. C. HASLAM

three precipitations, the component parts can be restored to the status quo ;
but that the farther the separation is pushed, the more difficult does the
restoration become. This would seem to show that each act of separation
causes some secondary molecular change in the parts separated.

In this connection I may recall the remarkable fact to which Hardy
draws attention: that while serum itself can be readily filtered through a
porous pot, globulin, even in the early stages of preparation, cannot. Hardy
also takes occasion to point out that neither alkali nor salt is capable of
producing so high a grade of solution as that of globulin as it exists in serum,
and that a further dissolving agent must be present. To say that that agent
is the other serum protein is not very different from saying that there is some
sort of combination among them; a conclusion to which Hardy is led on
other grounds. In the first paper of this series I brought forward some
evidence that this seems also to be the case with the albumoses.

Allowing, however, for the greatest amount of change in the process of
separation, it can hardly be supposed that any radical change of constitution
occurs: the changes are extra- rather than intra-molecular. So that from the
point of view of chemical analysis, the matter is not of great importance.

Possibility of Substances between Globulin and Pseudo-Globulin.

In the series of precipitations at one-third saturation, it was found that,
even after seven or eight precipitations at sufficient dilution, the product
obtained was largely soluble in water. After dialysis a good proportion,
perhaps 50 °/,, remained in solution. The clear solution certainly contained
some pseudo-globulin, and might well have consisted entirely of this and
globulin. We have seen that pseudo-globulin cannot be further split up,
but there was a possibility that the bulk of the soluble matter consisted of
a third substance, more soluble than globulin, but precipitable at one-third
precipitation with the globulin. To ascertain whether such a substance
existed, experiments were conducted on the following lines. Diluted serum
is precipitated some eight or nine times at one-third saturation: this gets rid
of all albumin, and a considerable quantity of pseudo-globulin. Globulin is
then removed by fractional precipitation, care being taken to remove only
insoluble matter, all doubtful portions being returned to the main solution,
so that the quantity of substance sought for should not be diminished in
this way.

This leaves us with a more soluble substance again, and from this, more
pseudo-globulin is removed by precipitation at one-third saturation.

Thus alternately portions of globulin and pseudo-globulin are removed

H. C. HASLAM 503

from the protein under examination. This can be continued until the
substance sought for is found, or there is no more substance to work with.

Two separate experiments were carried out, and on several occasions
portions of protein were obtained which were precipitable at “ one-third”
saturation, and soluble in water—further examination, however, always
showed them to consist of globulin and pseudo-globulin. I select, the follow-
ing for quotation.

500 c.c. ox-serum were diluted to 2,000 cc., and a precipitate formed by
the addition of 2,000 c.c. saturated salt solution. The precipitate was
dissolved in 2,000 c.c. and a further precipitate formed by the addition this
time of 1,000 c.c. saturated salt solution (one-third saturation). This latter
precipitation was repeated eight times when the filtrates were found to
contain only quite small quantities of protein. During the series of precipi-
tations large quantities of insoluble matter appeared, and were removed from
time to time. The final precipitate contained a fair proportion of protein that
was soluble in water. From this I sought to remove any water-insoluble
globulin that it might contain by means of the system of fractional precipi-
tation before described. The protein was dissolved in 200 c.c. water (the first
series of precipitations had considerably reduced it in bulk), and some 60 c.c.
saturated solution were added. The small precipitate that fell was found to
consist largely of water-insoluble globulin, but mixed with it was a certain
amount of water-soluble substance, and this was returned to the main
solution. To this a further quantity of salt was added and another fraction
obtained, which was treated in the same way. Some six fractions were
taken and in the last two only quite small quantities of water-insoluble matter
were found. The experiment was accordingly brought to an end by nearly
complete saturation of the liquid, and the resulting protein was examined.
On being dissolved in 80 c.c. of water it was found that on addition of 40 c.c.
saturated salt solution (one-third saturation), only some two-thirds of it were
precipitated. The remainder came down on half-saturation. It was clear,
therefore, that pseudo-globulin was present to a considerable extent. A
fresh series of precipitations at one-third saturation was therefore undertaken,
some eight in number. The resulting substance of which there was only a
small quantity, was again found to contain both water-insoluble and water-
soluble parts. Owing to the withdrawal of a large quantity of pseudo-
globulin, the insoluble globulin was easily demonstrated again. The water-
soluble part, after further precipitations of a similar character to those
already described, was shown to contain both water-insoluble globulin and
pseudo-globulin.

33—2

504 H. C. HASLAM

It was thus seen that no body precipitable at one-third saturation and at
the same time water-soluble, could be obtained that remained constant in
quantity, and could not be resolved into globulin and pseudo-globulin. Such
fractions, therefore, must be looked upon as mixtures. Or if we look upon
the serum proteins as one molecular structure in the serum, such fractions
must be regarded as portions of that molecular structure which have so far
resisted the disintegrating action of the salt; but which, by further action can
be resolved into globulin and pseudo-globulin, And it may be further
remarked that by precipitation at suitable concentrations fractions of material
or portions of the original molecular aggregate may be obtained having any
required solubility, and exhibiting a certain appearance of constancy. This
is much more the case with mixtures of globulin and pseudo-globulin than
with those containing albumin. The two former are much more closely
connected at any rate in regard to their behaviour to salt, than are pseudo-
globulin and albumin. As we have already seen, albumin can be removed
comparatively readily.

I thus arrived at the general conclusion that there are two, and only two
different proteins that are precipitable at half-saturation with ammonium
sulphate; the historic water-insoluble globulin and the protein soluble in
water corresponding to Hofmeister’s pseudo-globulin. In addition to the
precipitation differences between these two bodies, there are also certain
differences in appearance. The soluble substance comes down in finer
particles and has not the flocculence of the globulin precipitate. It is
perhaps rather whiter and does not become discoloured so readily as globulin.

The Separation and Purification of Pseudo-Globulin.

We are now in a position to discuss the separation and purification of
pseudo-globulin in greater detail, with a view of determining the limits of
fractional methods, and the most convenient way of applying those methods to
this particular case.

To obtain crude pseudo-globulin fcr further experiment, diluted ox-serum
was taken; globulin and albumin got rid of at one-third and half-saturation
respectively. After some two to three repetitions, a sample of pseudo-globulin
can be obtained, giving no precipitate at one-third saturation, and leaving only
traces of protein in the filtrate at half saturation. I will now describe the
method of fractional precipitation I have hitherto employed in greater detail.

The crude protein, in this case pseudo-globulin, is dissolved in water and
saturated salt solution is added until a precipitate begins to fall. This

|
|

H. C. HASLAM 505

is usually just after one-third saturation. In the early stages of the experi-
ment a good precipitate is obtained at “one-third” saturation of the fraction.
After the first few removals of globulin, however, this is no longer the case.
The concentration of protein in the solution at this point may be 2—0°5°%/,.
The readiest way of determining this is by a Kjeldahl determination of the
total organic nitrogen, calculating the protein as containing 16°/,N. The
precipitate formed may be some 20 to 25°/, of the whole quantity of the
‘protein present. This is determined by comparing the opacity of the pre-
cipitate formed with that in a small quantity of the solution in which the
protein has been completely precipitated by the addition of solid salt as has
been described before. In using this method it must be remembered that
the precipitate continues to fall gradually for some hours after the addition
of a quantity of salt, so that only a rough determination can be made at the
time of adding salt. The rate of fall of the precipitate appears to depend, to
some extent, on the concentration of the protein : the diluter the solution the
slower the fall. After the formation of the precipitate the solution is allowed
to stand 24 hours. If the precipitate has sunk to the bottom, most of the
liquid can be decanted. The remainder is filtered and the precipitate is
redissolved. Some 35-30°/, of this is then precipitated at about the same
concentration as in the previous experiment.

This precipitate being some 9-6 °/, of the total protein present, is removed,
and the filtrate is added to the main solution. The amounts withdrawn are
gradually lessened as the experiment is repeated and may be tested for the
protein it is desired to eliminate, in this case dialysis being used to detect
globulin.

The process can be repeated indefinitely. When, however, the separation
proceeds slowly, that is, when at each precipitation the proportions of the
substances to be separated do not differ much in the precipitate and filtrate,
it is expensive both of material and time. If the fractions are large the
material is soon used up, and if small the process becomes very slow.

The amount of the fraction should correspond to some extent with the
amount of protein present which it is desired to eliminate. As will be seen
from the experiment quoted, some six or eight repetitions of this process will
reduce the insoluble globulin in pseudo-globulin from 8-10°/, to 2-3°/,.
The volume of the solution is increased at the end of each pair of precipita-
tions. When it becomes inconveniently large it can be reduced by salting
out the protein completely from a portion, dissolving the protein so obtained
in a minimum of water, and returning it to the main solution.

In employing this method in the separation of the albumoses in some

506 H. C. HASLAM

cases I had no means of determining the progress of the separation, beyond
that of noting the increasing solubility of the protein being purified and
obtaining from successive fractions diminishing quantities of the more
insoluble protein. Where hetero-albumose was being salted out there was
the more independent test of precipitation by dialysis. This is the case in
the present instance. This test, however, is only a solubility one, and cannot
be relied on in the same way as some independent reaction. In the present
case we have this latter: since, while globulin contains phosphorus, pseudo-
globulin does not contain any. Globulin only contains about 0'1°/, P, so that
in spite of the possibility of detecting very minute quantities of phosphorus
the test is limited in its usefulness. It is sufficient, however, to show the
general course of the separation, and to enable us to place a value on the
solubility tests.

Hardy [1905] first pointed out that globulin contained phosphorus and
that other fractions of serum-protein containing less globulin contained less
phosphorus. The relationship of phosphorus to globulin is not quite simple
as will be pointed out more fully later on. It is sufficient for present
purposes to note that salting out operations and dialysis apparently leave the
phosphorus content of globulin unchanged; while samples of pseudo-globulin
can be obtained which, by the most careful tests, give no phosphorus.

For detecting and roughly estimating minute quantities of phosphorus
the ammonium phosphomolybdate method of Neumann was used; the
decomposition of the protein being effected by Bayliss and Plimmer’s
modification to avoid using more sulphuric acid than is necessary. Am-
monium phosphomolybdate contains only some 1°6°/, P, and as very small
quantities can be precipitated and detected, it forms, under proper conditions,
a very delicate test. Further, owing to the way in which minute quantities
of the precipitate fall, it is possible, simply by inspection, to estimate the
amounts with rough accuracy.

Taking known quantities of a solution of sodium phosphate I have found
that 0°005 mg. P can certainly be detected, sometimes 0°0025 mg. P. And if
the conditions as to total volume of sclution, quantities of reagents present,
and heat used to effect precipitation, are maintained equally, the differences
between such quantities as 0:005, 0:0075, 0:01, 0015, and 0:02 mg. P, can be
readily appreciated; and, by comparison, a rough estimate can be formed of
the trace of phosphorus under consideration.

Good results were obtained with solutions differing slightly from
Neumann’s:

H. C. HASLAM 507

30 ¢.c, Water.

30 c.c. 50 °/) ammonium nitrate solution,

30 c.c. 3 °/) ammonium molybdate solution.
Heat to 85° C,

Experiments were also made to determine how far Neumann’s method
could be used to estimate small quantities of phosphorus. Using a deci-
normal in place of a seminormal standard solution, I found that the method
did not lose in accuracy until the quantities were below 0°2 mg. P, when the
estimations became too high. Below 0:1 mg. P (which corresponds to 0°9 c.c.
decinormal solution) the values were quite unreliable :

(1) 0°52 mg. P gave 4°8 c.c. N/10 solution...0°539 mg. P

(2) 0-26. ,, me. ye Pe OB.
(3) 0-26. ,, et eee ee Os C07),
(4) 0-208 ,, a ae Pei OOS. 4,
(5) 0-208 ,, ee ea 028” ,,
(6) 0-104 ,, Pia \ Wee. 0-181 ;,
(7) 0-104 ,, aie 3, a O-1r
(8) 0-104 ,, Shh ol es Pao 07196,
(9) 0-078 ,, a ae oe ,..0°099" ,,
(10) 0-078 ,, 5S Qe ht eee) 0-088> ,;
(11) 0-052 ,, » 0-65 ,, mee. 0-0675. ,,
(12) 0-052 ,, gah 5 eta Sone .0-088"

For the experiments about to be described some three different samples
of ox-serum (each of about one to two litres) were mixed. In order to
determine the amount of phosphorus in the globulin of this mixture of
serums, some globulin was separated and purified, first by salt precipitation,
and then by acetic acid in the way previously described.

200 c.c. of 7th filtrate gave 1°2 c.c. N/10 NH;.
200: -,,, Sih” 5. yf Atl ”

The globulin was washed on the filter paper until no more acid was shown
in the filtrates and then washed with alcohol and ether. It was dried at
110° to constant weight.

10965 g. SB gave 1:2055 mg. P=0°11 9/p) (by Neumann).
0-732 ,, , 0785 0-107 Ye.
Mean =0°1085 °/o.

Therefore, 1 mg. P indicates 0°921 g. globulin.

I. Crude pseudo-globulin, giving no precipitate at one-third saturation
and containing traces only of albumin. ‘To determine the amount of globulin
contained in it phosphorus was estimated by Neumann’s method.

2°172 g. gave 0°19 mg. P corresponding to 8:4 °/, globulin.

The fractional process described above was then carried out six times in
succession ; large fractions were taken, some 28°/, in the first instance and

508 H. C. HASLAM

35-40 °/, in the second. There was a marked increase in the solubility of
each successive fraction, until, in the last fraction removed, I was unable, even
after four sub-fractionations, to find any protein insoluble at one-third
saturation. On the other hand, the fraction gave a distinct, though small,
precipitate on dialysis. The experiment was terminated by salting out all
the protein. To get rid of the small amount of albumin it was redissolved
in water and thrice precipitated at half-saturation.

200 c.c. of 2nd filtrate gave 2°4 c.c. N/10 NH.

200%. .,) aord “5 Sie Side Be ”
Thus albumin was eliminated.

On dialysis of a portion a good cloud appeared. The whole amount of
protein was precipitated by the addition of alcohol in the cold (below 15° as
advised by Mellanby) and washed with alcohol and ether. A portion was
dried at 110° to constant weight and the phosphorus estimated :

(1) 0:4115 g. gave 0-0125 mg. P by inspection...2°8 °/) globulin.

(2) 0-442 4, O015 ,, 23 oS as
If we take a mean and allow 20°/, error, the amount of globulin in the sample
is from 2°3-3°5 °/). .

The great contrast between the solubility in salt and the precipitability
on dialysis was surprising and will be referred to later. This inconsistency
was no doubt the basis of the observation by Freund and Joachim, who found
that the fraction which gave no precipitate of globulin at one-third saturation
gave a precipitate on dialysis so large that it was considered another
substance.

II. Crude pseudo-globulin as in I. This experiment was on a much
larger scale, and the process was repeated 36 times. The fractions were
smaller than in I, being at first 20°/,, but dropping later to 12-15°/,. The
dilution varied from 1—0°5 °/, in the earlier stages to 0°2—0'1 °/, in the later.

The salt solubility of succeeding fractions was noted, and a general
agreement with the results in I was found.

After the ninth and twentieth fractions, the whole protein was salted out
and dried. On each occasion it was found on re-solution that the solubility
had decreased, and further that fractions taken immediately after re-solution
were less soluble. The tenth fraction had the same solubility as the fifth, and
the twenty-first as the fifteenth.

All the sub-fractions were tested by dialysis. It was sought by this means
to follow the progress of the separation. The quantity of protein precipitated
showed in general a decline. The decline was not a regular one, however.
The first nine tests showed a steady decrease of substance. The tenth (after

H. C. HASLAM 509

drying) showed an increase, and from this to the seventeenth a slower and
more irregular decline was noted. The twenty-fifth fraction showed no pre-
cipitate on dialysis, but the sub-fraction gave a cloud. The thirtieth fraction
and sub-fraction remained clear. The thirty-first sub-fraction, however,
gave a good cloud and the fraction a faint cloud. The thirty-fourth fraction
and sub-fraction were again clear and also the thirty-fifth.

In general it was noted that when the fractions taken were small, the
separation seemed to make very little progress. Observations were made
as to the composition of the precipitates on dialysis, since in some cases they
appeared out of proportion to the amounts of globulin that could be present.
It was found that a considerable amount of pseudo-globulin was precipitated
with the globulin, as the experiment became more advanced. The precipitate
caused by the dialysis of serum or of the mixture of proteins obtained by one
or two precipitations - by “salting out” consists mostly of globulin, a certain
proportion of pseudo-globulin and a small amount of albumin. The earlier
fractions on dialysis gave such precipitates. In the later fractions, after the
twelfth, it was found that on dialysis a large quantity of pseudo-globulin was
precipitated with the globuli—in some cases an excess of pseudo-globulin.
In one case where some 6-7°/, of the protein present was estimated to be
globulin, the precipitate was some 15-20 °/, of the protein present. Allowing,
therefore, for 2°/, globulin remaining in solution, the precipitate must have
contained some 70-80°/, pseudo-globulin, to 20-30°/, globulin. Without
attaching too much importance to these estimates, it is clear that considerably
more pseudo-globulin than globulin was in this precipitate.

Precipitates of this kind remain insoluble, or nearly insoluble in water.
They may be largely dissolved by the addition of salt, or better, a trace of alkali.
In the latter case, if the solution be neutralised, no reprecipitation occurs,
and further, if the solution be then dialysed there may be only quite a small
proportion of the protein precipitated, or even none at all. This latter result
could be attributed to traces of acid or alkali being left over after imperfect
neutralisation. This was shown to be the case in many fractions containing
small quantities of globulin. A portion dissolved in water and dialysed gave
a small precipitate. Another portion which had been dissolved in dilute
ammonia and then neutralised gave no precipitate on dialysis: nor did
subsequent additions of traces of acid or alkali, nor prolonged dialysis, have
any effect whatever in producing a precipitate.

The size and formation of the precipitate on dialysis appeared also to be
connected with the rate at which the salt was withdrawn from the solution.
In the majority of the experiments the precipitate reached its maximum in

510 H. C. HASLAM

24 hours; although there was then further salt in the solution, further
dialysis produced no precipitate, even though continued many days. If,
however, the solution was diluted four or five times with distilled water, a
further precipitation was sometimes obtained.

It was also frequently noted that if pseudo-globulin containing a small
proportion of globulin were dried, the precipitate on subsequent dialysis
was increased. Both globulin and pseudo-globulin are rendered more
insoluble by drying: by repeated drying both may be rendered entirely
insoluble. So that the physical condition of pseudo-globulin would seem to
be an important factor in inducing its precipitation with globulin on dialysis,
and may have been one of the causes of the increase in the amount carried
down with globulin already referred to. |

It was also noted that, as the experiment advanced, a tendency to
mechanical coagulation developed. This was chiefly noted in filtering, after
a fraction had been precipitated. The protein in the solution then was just
on the point of being precipitated. Unless the filtration was conducted
slowly, small flakes and strings of protein appeared, which tended to adhere
together at the end of the filter.

Towards the end of the experiment, also, the fractions were more affected
by drying than in the early stages. _ Thus there is distinct evidence that the
prolonged contact with salt induces slow physical changes in pseudo-globulin.

Only one intermediate observation of the amount of phosphorus was
made. After nine fractions a portion was freed from salt, and precipitated
by alcohol and dried.

0°347 g. gave 0°01 mg. P by inspection...... 2°6 °/, globulin.

This shows a rather slower rate of progress than in I.

Of the final substance 0°561 g. was tested for phosphorus with a negative
result.

The physical changes noted during the experiment were :

1. A gradual increase of solubility in salt solutions which, after a time,
appears to cease. In no case did the increase go so far that pseudo-globulin
became soluble in half-saturated, or nearly half-saturated solutions (that is at
ordinary dilutions, 5-0°2 °/,).

2. The substance gradually became more liable to pass into the insoluble
or coagulated state.

In regard to the dialysis test it was found :

1. That the amount of pseudo-globulin carried down with globulin on
dialysis increases as the experiment proceeds; the pseudo-globulin, after a
time, being in excess of the globulin.

;
’

H. C. HASLAM 511

2. That if the substance be dried before dialysis a considerably larger
precipitate results.

3. That the amount and constitution of the precipitate is also dependent
on various chemical conditions,

In regard to the phosphorus estimations, assuming that 0°005 mg. P is the
smallest amount that can be detected, the final result would mean that the
pseudo-globulin might contain a quantity less than 0:00087 °/, P, but not more.
This, together with the rather better result in III, where nearly a gram gave
a negative result, is practically conclusive that pseudo-globulin does not
contain phosphorus. We can, therefore, attribute the small and diminishing
amounts of phosphorus found in the experiments to the presence of globulin,
and so far as we can estimate the phosphorus we can estimate the globulin.

Experiments I and II, therefore, show :

1. That by the method described all but some 2-3°/, of the globulin can
be eliminated in some six or eight fractionations.

2. That the larger the fractions taken, the more rapid the separation.
So that the method becomes too lengthy if used economically as regards
material.

3. That both solubility tests as guides to the progress of the separation
are fallacious.

As to the progress of the separation in II after the ninth precipitation,
it is clear from the negative result in the final test for phosphorus, that most
of the remaining 2°6°/, of globulin is eliminated. It would seem probable
that the elimination is a very gradual one.

In the albumose separation I relied on solubility tests, especially the salt
solubility one. In no case, however, did the indications point to such a rapid
conclusion of the separation as in I: nor were such large divergences as those
between I and II met with. Although, therefore, there may have been some
small alterations in solubilities which would tend to falsify results, it is
hardly possible that it could have been so extensive as in the present case.
In all respects the albumoses showed themselves more stable under treat-
ment than the serum protein.

Finally, I tried a different method with a view of obtaining greater
rapidity. In place of taking only a small proportion of protein at each
precipitation, which is only efficient when a comparatively large proportion
of globulin is contained therein, I tried the following plan which is based on
the principle of dividing the protein into two equal parts at each precipi-
tation.

III. Some 10 g. of crude pseudo-globulin which had been fractioned

512 H. C. HASLAM

four- times on the method described above, and contained about 4°/, of
globulin, were dissolved in 2,000 cc. water. By the addition of solid salt
about half the amount of protein as estimated in the way above described
of comparing opacities was precipitated. The precipitate P_, was dissolved
in 2,000 c.c., and again by solid salt equally divided into P_, and F,. The
filtrate F,, was further divided equally as far as possible into P, and F,, by
the further addition of salt. At this stage, there were four divisions each
containing about a quarter of the original protein. The process can be best
explained by means of a diagram.

O
oe iaara ae
Py Fy
T a ] Is “I al

IPs Fo=Po Fi
SSS ea b Sal
Ios IPI Hej Ieee F453

ie te at ip ae al d
Pig Fo= Po Pyo=F 49 Fy4
ieee Rese | are
Pa, Bay F.3= 1g F445

F, and P, were mixed being considered nearly of the same grade in
purity; the precipitate being dissolved in a minimum of water and added to
the filtrate. The mixture was then precipitated by the addition of solid salt
as before and divided into P_, and F,,.

At each precipitation the globulin is divided into two portions, most
going into the precipitate ; so that the amount of globulin in the filtrates on
the right F,,, ie, and F',, is constantly diminishing. The process which
involved a considerable number of precipitations was continued until a
sufficiency of material of grades 3, 4, and 5 was obtained. It was found that
an exactly equal division of the protein in any given experiment could not
be obtained without an undue amount of time being spent. The method
used was that already described of comparing opacities of precipitates.
Owing, however, to the gradual fall of the precipitate which is the more
accentuated the more dilute the solution, the precipitate cannot properly be
estimated until twenty-four hours have elapsed. To save time, therefore, I
endeavoured ‘to allow for the extra fall on standing, by precipitating only
some 35 or 40°/, as estimated shortly after adding the salt. This led to
several inexactitudes and probably accounts, in part, for the slight inequality
of the results.

The phosphorus estimations were all done at the same time, and com-
pared with each other as well as with known amounts:

H. C. HASLAM 513

Parent substance 0°335 g. gave 0°015 mg. P...4°1 °/) globulin. (By inspection)

Third grade 0-511 re 0:0075 ,, ...1°3 %o ve PF
Fourth grade OSB67) 5, O007b" 5, 3.0°79%5 5 “
Fifth grade 0°883 4, negative, i.e. contains less than 0°5 °/) globulin.

The figures, allowing for the roughness of method of estimation, are
sufficient to show that this method is more suited to eliminate small
quantities than the other. It also has the advantage of economising time
and material, as the precipitations, though numerous, can be carried on
simultaneously, after the first few; and all material in various grades can be
labelled and kept ready for use at any time.

These estimations of phosphorus, restricted as they have been owing to
the small amount present, are, I think, sufficient to establish the main point
that the separation of these two proteins can be completed, or nearly
completed, by means of fractional precipitations.

In regard to the practical point of the preparation of pseudo-globulin
containing, say, not more than 1°/, globulin, I do not think any very
extensive procedure will be necessary. I found that in _ precipitating
pseudo-globulin containing some 2—3°/, of globulin with alcohol in the cold,
that a small fraction could be obtaimed which was comparatively rich in
globulin. I have not concluded my experiments on this point, but it is not
unlikely that after a few fractions with salt have got rid of all but some 3 °/,
or so of the globulin, that a large part of the latter can then be eliminated by
a few fractions with alcohol.

The Preservation of Globulin and Pseudo-Globulin.

I have already pointed out that both globulin and pseudo-globulin on
being dried, at any rate with salt, become more insoluble. To keep them
without drying I have found that, by the addition of a little ether to their
solutions in dilute saline, they will keep for considerable periods in stoppered
bottles. The solutions should be well shaken with the ether.

I have samples which I have kept over a year which are quite clear and
show no tendency to precipitate.

Relation of Phosphorus to Globulin.

In the preceding section I have assumed that globulin contained
phosphorus as an integral part of itself. This was because throughout such
reactions as there described the globulin always retains phosphorus. It is
difficult to believe that if the phosphorus could be split off by salt, ammonia,

514 H. C. HASLAM

or acetic acid, that any except a trace would have remained with the globulin
after the prolonged operations entailed by its purification. However, to
obtain further assurance I estimated the phosphorus in a sample of crude
globulin prepared from serum by four precipitations of salting out at
one-third saturation, followed by two by acetic acid, after solution in
ammonia.

1-4725 g. globulin gave 1:462 mg. P...0-099 °/o.

This is distinctly less than in the case of the more purified globulin although
it had suffered less than half the number of precipitations.

I have already quoted (page 507) the phosphorus determinations in one
sample of purified globulin. Another sample from ox-serum was analysed :

0:5495 g. globulin; 0°575 mg. P...0°105 Jo,

a very similar result.

Treatment of dried globulin with alcohol or ether gives a yellow, rather
fatty extract, which contains phosphorus. This extract is only separated
from globulin with some difficulty. I have tried various means; treatment
of the finely powdered globulin in a Soxhlet apparatus with ether ; prolonged
shaking with alcohol and ether; and boiling with alcohol, after Plimmer.
The latter was the most effective; an extract could generally be obtained by
its means from samples which had been previously treated in other ways.
I found, however, that though most of the extract could be obtained after
some eight to ten hours boiling, that further small quantities could be got
by more prolonged boiling. In no case could I get anything like a phosphorus-
free globulin. I quote two experiments in which the globulin, well powdered,
was boiled with successive portions of alcohol for periods of seven hours,
until no more extract was obtained.

(1) 0:5265 g. globulin gave ext. cont. 0°3096 mg. P; 0:056 °/y. Boiled 21 hrs.
(2) 0°68625 g._,, é vx (O3RUT 5,02 0-048. Sa ee
Thus about half the original phosphorus is got rid of in the extract. In
previous experiments in which prolonged shaking (200 to 300 hrs.) was
adopted, not more than 30 or 35°/, of the phosphorus was eliminated. I also
shook solutions of crude globulin and diluted serum directly with ether.
In the former case the results were negative, though in one or two experi-
ments minute traces of an extract similar to that looked for were obtained.
With serum, no trace of the extract looked for was found, but an orange
pigment was extracted without much difficulty. From pure or nearly pure
pseudo-globulin no extract could be obtained by any of these means, though

H. C. HASLAM 515

small quantities were found in samples containing 3 or 4°/, of globulin, as
would be expected. The amount of extract obtained from globulin measured
by weight was some 8-10°/, of the globulin treated.

In view of the difficulty of completely extracting solids by liquids and
especially the very slow action of the extracting liquids in this instance,
it is tempting to suppose that the whole of the phosphorus in globulin
belongs to some lecithin-like body or bodies which would appear closely
connected with the globulin, but not part of the molecule. These would
then amount to no less than 15-20°/, by weight of the globulin. On the
other hand the present facts do not warrant us in accounting for more than

half the phosphorus in this way.

SUMMARY.

1. There are two proteins of ox-serum insoluble in _half-saturated
solutions of ammonium sulphate, saturated magnesium sulphate, or sodium
chloride; the historic water-insoluble globulin, and the water-soluble body
pseudo-globulin.

2. These two bodies cannot be split up further by means of fractionation
with salt and water.

3. Intermediate fractions are shown to be mixtures of these two bodies.

4. Globulin contains, or is closely associated with phosphorus, rather
more than 0'l mg. P°/, being found. About half this belongs to a fatty,
lecithin-like body which amounts to some 8—10°/, of the globulin freed from
pseudo-globulin. Apparently no part of this body is detached from globulin
through prolonged treatment with acids, alkalies, or salts.

5. Pseudo-globulin does not contain phosphorus.

6. Repeated precipitations of globulin at constant volume finally give
filtrates in which the amount of protein is constant. Globulin, presumably,
can thus be freed from pseudo-globulin.

7. Pseudo-globulin can be freed in a similar way from albumin, the
separation being considerably easier.

8. Pseudo-globulin can also be freed, or nearly freed, from globulin by
suitable methods of fractional precipitation.

9. The presence of phosphorus has given us an independent test for
following the progress of the latter separation. The result has been to
establish the validity of the general method of fractional precipitation in this
case, to give important indications as to the most suitable method to use,
and to estimate the value of the solubility tests.

516

H. ©. HASLAM

REFERENCES.

Burckhardt (1883), Archiv f. exp. Path. u. Pharm. 16, 322.
Freund u. Joachim (1902), Zeitsch. physiol. Chem. 36, 407.

Fuld u. Spiro (1900), Zeitsch. physiol. Chem. 31, 132.
Hammarsten (1878), Pfliiger’s Arch. 17, 413.

Hardy [1905], J. Physiol. 33, 330.

Haslam (1905), J. Physiol. 32, 267.

—— (1907), J. Physiol. 36, 164.

Heynsius (1869), Pfliiger’s Arch. 2, 1.

——— (1876), Pfliiger’s Arch. 12, 549.

Kander (1886), Arch. exp. Path. u. Pharm. 20, 411.
Marcus (1899), Zeitsch. physiol. Chem. 28, 559.
Mellanby, J. (1907), J. Physiol. 36, 288.

Panum (1851), Virchow Arch. 3, PAR

(1852), Virchow Arch. 4, 17 and 419.

Pick (1902), Beitrdge 1, 351.

Porges u. Spiro (1903), Beitrage 3, 277.

Reiss (1904), Beitriége 4, 160.

Schmidt, A. (1862), Miéiller’s Arch. 429.

Wiener (1911), Zeitsch. physiol. Chem. 74, 29.
Zimmerman (1854), Miiller’s Arch. 377.

L. THE CHEMISTRY OF THE LEUCOCYTOZOON
SYPHILIDIS AND OF THE HOST’S PROTECT-
ING CELLS.

By JAMES EUSTACE RADCLYFFE McDONAGH anp
ROBERT LAUDER MACKENZIE WALLIS, Gillson Research
Scholar in Pathology, Society of Apothecaries.

(Received June 28th, 1913.)

INTRODUCTION.

The life-history of the Leucocytozoon Syphilidis has been fully described
by McDonagh [1913, 1, 2]. The present communication represents an
attempt to elucidate the micro-chemistry of the organism in question.

The first consideration will be given to the physical and chemical
properties of the staining reagents used, then to the characters of the
syphilitic bodies as evidenced by staining in vivo; the same by the staining
of fixed specimens, in which the protoplasmic and nuclear portions will be
dealt with separately. Reference will then be made to some _ physical
characters of the various cells, and attention given to some special chemical
features of the syphilitic bodies and of the host’s protecting cells. Finally,
the physical and chemical characters of the lecithin-globulin complex will be
mentioned.

THE PHYSICAL AND CHEMICAL PROPERTIES OF THE STAINING REAGENTS USED.

In all micro-chemical investigations it is important fully to consider the
chemical and physical properties of the staining reagents before drawing any
conclusions as to the chemical nature of the structures under consideration.
The dyes as a group are almost entirely colloidal in nature, that is to say,
they do not readily diffuse through animal membranes. Their colloidal
nature is also emphasised by their high molecular weight and absence
of crystalline forms. In solution these dyes form typical colloidal suspen-
sions, the size of the particles varying in different cases; consequently we
notice that some stains appear quite homogeneous, whilst others show the

Bioch. vu 34

518 J. E. R. McDONAGH AND R. L. M. WALLIS

well-marked characters of suspensions. Further, many dyes are removed by
filtration and may be separated by this means, as, for example, methylene
blue occurring in the urine after subcutaneous injection.

Owing to their colloidal nature the small particles of a dye in suspension
become electrically charged, and we may for convenience divide up dyes into
negatively and positively charged colloids. The negatively charged colloidal
particles of a dye will, therefore, show electrical migration to the anode,
whilst positively charged colloids will travel towards the kathode. It follows
from this also that on mixing a positive dye with a negative dye the
charges will be neutralised and the resulting colloidal mixture will then
exist in an uncharged condition; wide the eosin-methylene blue mixture,
the former being negatively charged whilst the latter is positively charged.

But the one great feature of all dyes is that they exhibit the phenomenon
of adsorption (Bayliss). By adsorption we mean a “combination” between
two substances which is not strictly in the nature of a chemical union, that
is to say, in which there is no direct proportionality between the concentration
of the solution and the amount adsorbed. There is some kind of physico-
chemical affinity between the bodies adsorbed, and those which take them
up, but this affinity is more in the nature of a mechanical than a true
chemical affinity. To take an example, if a series of solutions of Congo red
of varying concentrations be taken, and the same amount of filter paper be
placed in each, a part of the dye is taken up, but in relatively larger
proportions in the more dilute solutions of the dye.

There are, however, instances where the combination of the dye with the
material acted upon is in the nature of a true chemical combination, but
these are exceptions rather than the rule, vide the staining of nuclei with
rongalit white. ‘

A third and most important factor comes into play in the processes
of histological preparations, namely, the presence of electrolytes. The dyes
as a whole are very sensitive to electrolytes in regard to their adsorption
properties and the effect appears to be proportional to the degree of their
colloidal nature. Electro-negative dyes, like Congo red, are increased in
adsorptive power by the addition of kations, such as lithium, potassium,
sodium, ammonium, magnesium and calcium. On the other hand, anions
such as hydroxyl, acetate, chloride, oxalate, sulphate and phosphate
depress adsorption. With electro-positive dyes, such as toluidine blue, the
reverse holds good, namely anions facilitate and kations depress. The effect
of electrolytes, however, is much more marked in the case of kations than it
is with anions. The salts of the heavy metals which form positively charged

J. E. R. McDONAGH AND R. L. M. WALLIS 519

colloidal hydroxides, for example, iron, have a very powerful effect on the
adsorption of electro-negative dyes, and in every case the ion promoting
adsorption of the dye is carried down with it.

If we turn now to histological preparations, perhaps we may in the light
of these observations be able to interpret the changes induced in the act
of staining. In the living cell the substance to be stained has a negative
charge, as the reaction of the tissues is always on the alkaline side of
neutrality, and we know that protein solutions in an alkaline medium are
always negatively charged ; it follows, therefore, that living cells will take up
basic dyes and that electrolytes are not essential for the process. When the
cells die the electrolytes attached to the protein constituents of the cell are
split off, with the result that the cells now readily take up acid dyes.
Further, we know that the fixation of a dye is facilitated by heat, and this
fact has been made use of in Altmann’s method of acid-fuchsin staining.
Mayer has also shown that the Nissl bodies of nerve cells have their affinity
for basic dyes destroyed by previous treatment with neutral salts; an
observation which further emphasises the importance of electrolytes in the
staining act.

Let us now turn our attention to the observations upon the syphilitic
parasite and see whether the results obtained can be interpreted in the light
of these views of staining. In the first place, we will consider the two
principal stains used, viz. pyronin and methyl green. Both these dyes are
positively charged colloids and in consequence will be facilitated in their
staining by anions, particularly hydroxyl and sulphanion. The pyronin will,
therefore, act as a basic dye and this explains why it is precipitated by
nigrosine and also diamine green, but not by diazine green. Making use of the
chemical characters of the dyes used and reviewing all the results obtained
in this light, it is obvious that much more useful and valuable knowledge
of the micro-chemistry of the cells under investigation can be ascertained.
The specific staining properties of the syphilitic parasite will be again
referred to and all we would point out here is the importance of regarding
pyronin as a positively charged colloid, and influenced by anions.

THE CHARACTERS OF THE SYPHILITIC BODIES WHEN STAINED “IN VIVO.”

The simplest and best way to use a reagent for vital staining is to spread

a solution on to a slide free from fat and alkali and then allow it to dry in

the air. The film should be made just before it is required and not kept

several days. The material to be examined is placed upon a cover slip and
34—2

520 J. KE. R. McDONAGH AND R. L. M. WALLIS

the latter so adjusted to the slide that no air remains in between and so
that as little as possible of the fluid to be examined exudes at the sides.
Examination is possible until the fluid dries, which takes several hours.
Ringing the cover slip with wax greatly facilitates the process. The slides
used for hanging drop preparations are useless, but a warm stage may some-
times be advantageously employed.

We have used all the stains which have been from time to time
advocated, but have not obtained the results which one might be supposed
to get from reading the literature about them; therefore, it might be as well
to state that neutral red, neutral violet, Bismarck brown, auramine, diazine
green, malachite green, tropaeolin 00, and Congo red do not give good
vital staining. Owing to the use which is now being made of the azo-dyes
in determining the functional activity of certain cells in the body, we
attempted several for our method of staining im vivo, but by no means
with success, because they have such feeble staining properties and are
general protoplasmic stains without possessing affinity for certain structures ;
moreover, they do not possess metachromatic properties. The dyes which
gave the best results were aqueous solutions of borax-methylene blue and
polychrome-methylene blue; brilliant crystal blue, Nile-blue sulphate and
alcoholic solutions of toluidine blue, thionine and azure II. The disadvantage
of the alcoholic solutions is that they must be diluted well before use as
crystals so readily form, and this is especially the case with thionine. Azure II
stains deeply, but unfortunately has no metachromatic properties. The
intracellular stages are perhaps better depicted by the alcoholic stains, but
taking everything together the results are not so good as when the aqueous
solutions of either borax-methylene blue or brilliant crystal blue are used :

and, of the last two, the former is superior. The metachromatic properties:

of borax-methylene blue are dependent upon the basic sodium biborate, which
acts upon methylene blue by producing methylene violet which, as a basic
dye, shows affinity for acid substances, and methylene red which is an acid
dye and so shows affinity for basic substances.

We tried various other bases with methylene blue with the hope
of obtaining greater metachromatic action by substituting a base of higher
valency for the borax, for which purpose we employed colloidal aluminium
hydroxide. The methylene blue remained unaltered, no doubt owing to the
fact that the aluminium hydroxide did not contain sufficient free hydroxy]
ions and was in itself an unstable colloid and therefore had no action upon
the positively charged methylene blue.

Borax-methylene blue, when freshly prepared, has practically no

J. EK. R. MCDONAGH AND R. L. M. WALLIS 521

metachromatic action, but this property increases the longer the stain is kept,
until finally the methylene red becomes the stronger dye and stains the cells
just as the methylene violet does. Borax-methylene blue appears to be at
its best when it has been kept for a year or two. As the methylene red
scarcely comes into play in the fresh solutions, no harm is done by adding
O'1 g. eosin to 100 cc. borax-methylene blue, as a true chemical compound
results. The eosin picks out the granules in the polymorphonuclear
leucocytes, also stains brilliantly the eosinophile granules, but not one of the
stages of the syphilitic organism. When we had learnt that the syphilitic
organisms contained lecithin in the form of a lecithin-globulin complex,
and being aware of the affinity that this complex shows towards dextrose,
it struck us that it might be possible to increase the staining properties
of the organism by adding dextrose to borax-methylene blue. Although
the dextrose did not carry the colloidal dye particles to the cells in
question, it nevertheless was taken up by every cell which contained the
lecithin-globulin complex, since the protoplasm of such cells swelled and
absolutely refused to stain, but, owing to the swelling, were as easily
discernible as if they had stained, consequently the plasma cells and syphilitic
bodies could be well studied, as their nuclei were not prevented from
staining. Witnessing the act of impregnation in a dextrose-borax-methylene
blue specimen, we were able to compare it with what we had previously
seen. The nucleus of the female appeared to float about, surrounded as it
was by the clear ring of unstained protoplasm; when first seen one end
of the spirochaeta had already entered, as one extremity was fixed to the
nucleus; the spirochaeta was also thickened (due to the dextrose), The one
end of the spirochaeta remained fixed to the female nucleus and, in spite
of knocking against all the other cells in its progress, it remained attached
to the same spot, but did not enter any further. Suddenly the female cell
stopped and the spirochaeta pallida vanished inside, but 55 minutes had
elapsed before this happened. No further change was noticed in the female
cell, as it had not stained very well, but about four minutes later it became
very active and discharged a clear non-staining polar body which seemed to
be emitted with some force. A few seconds later another clear polar body
was extruded and then the female cell came to a standstill again. It is
possible that the dextrose made these polar bodies appear bigger than they
really were, as each was certainly 2-3 w in diameter.

From what has been stated it will be easily seen that a description of the
syphilitic organism in vivo from its reactions will entirely depend upon the
characters of the borax-methylene blue which is used, and as impregnated

in

522 J. E R. McDONAGH AND R. L. M. WALLIS

female cells do not stain with eosin or the methylene red of freshly prepared
borax-methylene blue, the increase in the basicity resulting from impregnation
cannot be very great. It is far more probable that no change in reaction
occurs and that the reason for staining with methylene red, a fact which we
have frequently observed, is due to an increase in the reducing action, as will
be shown later.

The sporozoites may remain for some time unstained or stain immediately
a dense violet. The intracellular phases stain late and the early ones show
an affinity for the methylene violet moiety, whilst the late ones, viz. the
coils, take up the methylene red. The females before fertilization remain
unstained except their chromatic network and blepharoplasts which stain
immediately with methylene violet. The spirochaeta pallida stains pink
and when it has impregnated a female cell and the whole cell has come to a
sudden standstill, a pink diffuse stain comes over the cell like a mantle.
The sporozoites while in the spore cysts show a greater affinity for methylene
red than methylene violet, and some spore cysts are seen which stain
distinctly metachromatically.

In staining in vivo with a negatively charged colloid it follows that
alkaline basic dyes will react best, and this has been found to be the case;
and the reason why neutral red, neutral violet, Bismarck brown, auramine,
diazine green, malachite green, tropaeolin 00, and Congo red were found not
to give good results is simply the fact that they are negatively charged
dyes and therefore cannot stain cells which contain colloids in solution with a
negative charge existing in a medium on the alkaline side of neutrality. It
follows that good staining can only be obtained by using dyes with a positive
charge, hence the reason why borax-methylene blue and polychrome-
methylene blue serve so admirably, as it is only under such conditions that
adsorption can come into play.

Summary. Basic stains are the most suitable for in vivo work, and of
these borax-methylene blue is the best. Owing to the presence of a lecithin-
globulin envelope the syphilitic bodies can be made to stand out clearer by
adding dextrose to the stain. The varied affinity shown by the different
bodies on the one hand for methylene violet and on the other hand for
methylene red is due to the prevalence of a substance which has strong
reducing properties (lecithin-globulin) and not to a change in the reaction.

J. E. R. MCDONAGH AND R. L. M. WALLIS 5

bo
2D

THE CHARACTERS OF THE SYPHILITIC BODIES WHEN STAINED IN
FIXED SPECIMENS.

Both Pappenheim and Martin Heidenhain explained the specific action
of methyl green for chromatin as being due to the breaking down of the
weak basic salt by the strong nucleic acid radicle; on the other hand the
nucleus did not stain with pyronin which they regarded as a stronger basic salt.
As a matter of fact methyl green, being a triamino-stain, is by far a stronger
base than the diamino-stain pyronin, the basicity of which is also diminished
by its O-ring formation. Furthermore, acetic acid increases methyl green
staining and if acids combine with the free amino-group of the salt, acetic
acid would have done this before the nucleic acid got a chance. Therefore,
the explanation of its action, which had held sway for some years, cannot be
the correct one.

A stain which had been largely used by Unna, namely rongalit white,
was found to resemble methyl green in many respects and was known only
to stain the oxygen positions of the tissues.

Rongalit white is the leuco-base of methylene blue which is prepared
with sodium sulphite and formalin. It is therefore a colourless and basic
mixture and the methylene blue is only brought out as a dye in the presence
of oxygen. As rongalit white stains the nuclear part of the cell, Unna
concluded that methyl green also picked out the oxygen foci of the tissue
and was, therefore, a reduction-sensitive stain.

By a series of experiments Unna showed that methyl green was far more
sensitive to reducing substances than methylene blue and that malachite
green came in between, but that such reducing agents as grape sugar and
hydroxylamine were without effect, 1.e. did not decolourise methyl green.
Another difference between methyl green and methylene blue and malachite
green was that the leuco-bases of the last two could be reconverted into their
coloured bases by the addition of hydrogen peroxide, but not so methyl
green.

Until Unna enunciated his theory of staining by oxidation and reduction
we were under the impression that staining depended upon reaction ; in
other words, that acid substances stained with basic dyes and were therefore
termed basophilic, and basic substances stained with acid dyes and were
therefore termed acidophilic. No doubt in part this conception is correct,
but there is also no doubt that it was carried very much too far. After all,
fixed protoplasm is an amphoteric substance, i.e. it can act as a base or an
acid; for instance, the protoplasm of plasma cells stains well with acid

524 J. E. R. MCDONAGH AND R. L. M. WALLIS

fuchsin, which is an acid dye, or with pyronin, which is a basic dye. Although
it stains with both, it does so better with the latter than the former; there-
fore under ordinary circumstances, it can be stated that protoplasm, using the
word in a very general sense, prefers to act as an acid.

From McDonagh’s description of the Leucocytozoon Syphilidis and the
coloured plates which illustrate his article [1913, 2], it will be seen that by
using Pappenheim’s stain (a mixture of pyronin and methyl green) the
syphilitic bodies stain with pyronin, but differ from all other cells in that
the nucleus also apparently stains with pyronin and with a much deeper red
than the rest of the cell. Working on the reaction hypothesis, or on the
electrolytic theory, one must assume then that the protoplasm and especially
that of the nucleus of the syphilitic organism is strongly basophilic and
negatively charged, but we have already shown it to be partly acidophilic
from our in vivo examinations, when we pointed out that certain phases
showed an affinity for methylene red. We have then a paradox and a
solution to the problem can only be found if we adopt Unna’s theory of
oxidation and reduction.

Methylene violet, like methyl green, is a reduction-sensitive dye, although
not to the same degree, and the reason why certain phases stain with
methylene red is not because the protoplasm is acidophilic, but because it
has reducing properties, and as methyl green is far more sensitive than
methylene violet, every phase stains with pyronin, while only in those in which
the reducing action is greatest is the affinity of methylene violet for nucleic
acid overcome, with the result that the reducing substance stains with
methylene red. In fixed specimens this reducing substance does not stain
very readily with acid dyes, should a basic dye be present as well, because
if pyronin is supplanted by acid fuchsin, most of the nuclei of the syphilitic
organisms stain with methyl green; therefore, this characteristic pyronino-
phile substance of the syphilitic organisms is a strong reducing agent, is
basophilic, and, accordingly, negatively charged according to the electrolytic
theory ; but the action of its electric charge is overshadowed by its reducing
action. Therefore, we have another extremely important factor coming into
play in the act of staining.

Seeing how sensitive a stain methyl green is, it at once appears obvious
what caution must be taken in choosing the most suitable fixing reagent,
and that any fixing reagent which robs the nucleus of its oxygen will
naturally prevent staining with methyl green, and will also alter the action
of the medium. Fixing tissue for 24 hours in a 1°/, solution of platinum
chloride increases the capacity of the nuclei for methyl green, but it is an

J. E. R. MCDONAGH AND R. L. M. WALLIS 525

expensive solution and does not give such good results as some other fixing
reagents. Mercuric chloride has the disadvantage that the sections may
stain unevenly, since it only increases the capacity for methyl green staining
in the situation where it remains and diminishes it in those situations where
it is reduced by the tissue. If mercuric chloride is employed in an alcoholic
solution as a fixing reagent, quite good sections may be obtained with
Pappenheim’s stain, and the effect is enhanced if a little acetic acid is added
to the mixture. Chrome salts render staining with pyronin and methyl
green impossible. Osmic acid alone or in conjunction with other acids
diminishes the receptivity of protoplasm to most dyes.

Formalin, owing to the formic acid which it so frequently contains as
an impurity, not only diminishes the staining properties of protoplasm, but
markedly reduces the power of the nuclei to stain with methyl green. In-
cluding other fixing reagents which are seldom used and are not available for
obtaining satisfactory sections with Pappenheim’s stain, we are left only with
alcohol, and from several experiments we have undertaken we are firmly
of the opinion that alcohol is far and away the finest fixing reagent we
possess at present, as it allows staining with most of the stains in general
use, it fixes by coagulation, and forms no chemical compound with the cells
and is, therefore, extremely well adapted for the purpose of micro-chemical
research when fresh sections are not employed.

We either employ absolute alcohol or 50°/, alcohol; the former causes
shrinking of the intercellular tissue but not so much of the individual cells
and its main advantage is that coagulation is immediate. The moment the
tissue is removed from the body it is placed at once into absolute alcohol and
allowed to rest on wool—whenever alcohol is used a pad of wool should rest
on the bottom of the bottle so as to allow the alcohol to remain the same
strength throughout while the water extracted from the tissue sinks through
the wool; this simple device means also a great saving of alcohol. The
tissue remains for 12 hours in absolute alcohol and can then be changed into
two further lots of absolute alcohol for 12 hours each or be put back to 50°/,
and gradually taken up in the usual way. Cedar wood oil is used for clearing,
as it does not harden the tissues to the same extent as xylene, and before it
is put into wax two changes of xylene are used for 1-2 hours each, as wax
penetrates better after the tissue has been through xylene. The wax used

is a mixture of:

Paraffin (melting point 60° C.) sa 84 parts.
Stearin sas on — eas 1 part.
Wax ake fess Res es 4 part.

526 J. EK. R. MCDONAGH AND R. L. M. WALLIS

The melting point of the prepared article is 53°C. and the tissue is left in
this, changed three times, for about six hours altogether. The whole of the
secret of getting good paraffin sections is to be absolutely certain that the
tissue is fully dehydrated.

The tissues can also be fixed in 50°/, alcohol, allowed to remain in the
solution for 24 hours and then for 24 hours each in 70°/, and 90°/, and
12 hours each in three changes of absolute alcohol, and so on as before.
There is not so much shrinkage of the tissue, and most excellent Pappenheim-
stained sections may be obtained; but, owing to the solubility of certain
substances in 50°/, alcohol, this method cannot take the place of absolute
alcohol in micro-chemical research. .

For some tests of minor importance celloidin sections are preferable, but
for general purposes we much prefer paraffin, as thinner sections can be
obtained ; they can be more easily fixed to the cover slips, and one does not
have to go through the troublesome procedure of removing the celloidin,
which is necessary when aniline dyes are being used, owing to the intense
avidity which celloidin has for many of them.

If Pappenheim’s stain is to be used we proceed as follows: Roughly two
parts of a saturated aqueous solution of pyronin are mixed with one part of a
saturated aqueous solution of methyl green immediately before use, or until
the mixture assumes a red-purple colour. In this stain the sections may be
left from 5 minutes indefinitely as overstaining is impossible. After being
in the stain the sections are transferred to a freshly prepared distilled water
solution of resorcinol, which is just used for washing off the stain, i.e. about a
minute, and then they are put into a freshly prepared absolute alcoholic
solution of resorcinol and kept in until all the supertiuous stain has come away.

These resorcinol solutions are absolutely essential, as they act as mordants
and we regard mordanting after staining as superior to Unna’s method,
which consists in adding carbolic acid to the stain, which enables the stain to
be prepared and always ready for use. The stain is sold under the name
carbol pyronin methyl green. The great disadvantage of the ready prepared
stain is that the pyronin comes out too quickly in the dehydrating process,
while if resorcinol is used this is not the case. The amount of resorcinol
crystals used in the first watch glass is about 0°3 g. and in the absolute alcohol
watch glass, just double the quantity. From the resorcinol the sections go
through three changes of absolute alcohol, two of xylene, and are then mounted
in balsam.

As ethyl alcohol may abstract the pyronin stain from the sections, clearing
fluid may be used to take its place, such as chloroform, lavender oil or

J. E. RL. MCDONAGH AND R. L. M. WALLIS 527

bergamot oil. Clove oil should never be employed owing to its powerful
reducing action.

Xylene fortunately acts indifferently, but Canada balsam in time, owing
to its reducing action and being an acid, destroys the staining effect; dammar
dissolved in xylene would form a better medium for preserving the section.
For a year or two, or even longer, sections stained in the above method and
mounted in Canada balsam show a much sharper contrast of colour; the
pyronin stands out clearer, the orange colour of the mast cells is more distinct,
but the methyl green staining is somewhat weaker, and as time proceeds it is
the methyl green stain which first disappears. Some sections prepared seven
years ago are as good as and in many respects, owing to increase of sharpness,
better now than then.

In a Pappenheim-stained section the protoplasm of the groundwork and
connective tissue cells stains a rose pink and has a finely granular appearance,
whereas the protoplasm of the plasma cells stains a clear red. All nuclei
stain green, the nucleoli a brilliant red, the mast cell granules orange and all
bacterial and protozoal bodies red. The protoplasm of the syphilitic bodies
stains a rose pink to red, and the nuclei a deeper red. The difference in the
rose pink to red of the protoplasm being most marked in the female cells is
dependent upon whether they have been impregnated or not, as the fertilised
female cells and zygotes always stain deeper. At first sight one might
conclude that the supposed nuclear part of the syphilitic organism stains
deeply with pyronin, because it contains no nucleic acid; but we have
endeavoured to prove that such a surmise is incorrect.

We added acetic acid to the alcohol in which the tissues were fixed in
the proportion of 1 ce. glacial acetic acid to 75 ce. either absolute or 50°/,
alcohol, with the hope that if any nucleic acid was present the acetic acid
would precipitate it. When the sections were stained we found that the
general pyronin staining had not been interfered with, that the methyl green
staining was strongly intensified, and that some of the nuclei of the syphilitic
bodies stained a brilliant green, proving at once that they contained nucleic
acid. The addition of acetic acid in the alcohol used for fixing, for ordinary
staining with pyronin and methyl green gives better results than when
alcohol is used alone, owing to the fact that apart from the nucleic acid
being precipitated, the swelling action of the acetic acid on the cells is
counterbalanced by the shrinking action of the alcohol and vice versa. For
special staining, as when the demonstration of micro-organisms is required,
alcohol alone is the one and only fixing reagent.

As methyl green is one of the ingredients of Ehrlich’s triacid stain and

528 J. E. R. McDONAGH AND R. L. M. WALLIS

the red dye, acid fuchsin, is an acid dye in contradistinction to pyronin which
is basic, we stained some sections with this mixture, with the result that the
protoplasm of the syphilitic bodies stained red, while the nuclei stained
green, another proof that the nuclei contain nucleic acid. We can, therefore,
assume that there is some substance either in or over the nucleus of the
syphilitic parasites which is a strong reducing agent, as it prevents the
methyl green from getting at the nucleus and that it prefers basic to acid dyes.

Our next step was to try to determine the reducing action of this
substance and also its degree of solubility in various reagents.

Reducing action. (a) Sections were stained for 1-2 minutes in a freshly
prepared 1°/, solution of potassium permanganate, then washed in water,
decolourised in oxalic acid if overstained, dehydrated and mounted in balsam.
All protoplasm has a reducing action and consequently stains brown with
potassium permanganate while the nuclei remain unstained. The protoplasm
of the syphilitic bodies has a greater reducing action than ordinary grano-
plasma, but the nuclei do not stain and can be counterstained with methyl
green; therefore, the reducing body has not a very marked action on
potassium permanganate.

(b) Sections were placed for 5 minutes in a mixture of equal parts of a
1°/, solution of ferric chloride and a 1°/, solution of potassium ferricyanide.
It is imperative that these two solutions shall only be mixed immediately
before use, as 1f allowed to stand a blue precipitate is slowly formed. The
reducing action of the granoplasma converts the ferricyanide into ferrocyanide
with the result that where the reduction is greatest a beautiful Berlin blue
colour is formed in the presence of the ferric chloride. Ordinary granoplasma
has a weak reducing action and so stains green; the granoplasma of plasma
cells and the syphilitic bodies have a stronger reducing action and so stain
darker, while the nuclei of the ordinary cells do not stain, except some of the
“nuclei” of the syphilitic bodies, which stain a faint Berlin blue colour.
Red blood corpuscles also give the Berlin blue reaction and so do the
aminoplasma cells, both to a more marked degree than the nuclei of the
syphilitic bodies. Ordinary nucleoli have likewise a reducing action on
ferric ferricyanide.

There is something quite peculiar in the appearance of the syphilitic
bodies stained with potassium permanganate and ferric ferricyanide, because
in the former, when counterstained with methyl green, the nucleus appears
smaller than it really is; there is less to take the methyl green. In both it
appears irregular, and scattered here and there about the nuclei are small
non-staining transparent areas.

1
2
?
4
=
>

>

J. KE. RL. McDONAGH AND R. L. M. WALLIS 529

We will deviate somewhat from our course here and dwell upon the
aminoplasma cells.

The aminoplasma cell is a form of plasma cell which Unna has called,
from his examinations thereof in fixed specimens, Hyaline plasma cell.
The term “ Hyaline” rather suggests some relationship to cartilage, although
it is very largely used for substances of which the observer has no knowledge.
Now, hyaline cartilage is a_ strongly basophilic substance owing to its
chondroitin-sulphuric acid radicle and therefore possesses great affinity for
methyl green. Unna’s hyaline plasma cells are, on the other hand, acidophilic
and contain no acid radicle; furthermore, they have very strong reducing
properties and so cannot stain with methyl green and, as we shall show
presently that this reducing action is due to tyrosine, we consider that the
best name for them is aminoplasma cells.

The cell is frequently to be met with in syphilitic material, but it is also
to be found in any very chronic inflammatory lesion, viz. Rhinoscleroma and
Uleus Molle Serpiginosum. In in vivo specimens an aminoplasma cell is
apt to be mistaken for a zygote, owing to the affinity which both have
for methylene red, but the distinction becomes clear when it is borne in mind
that the former may vary in size from 7-14 ~ or more in diameter, that it
may have no nucleus, that the nucleus stains homogeneously with methylene
violet, that it may be situated in the centre of the cell or at the periphery
and that it sometimes possesses the power of motion and may be extruded
and finally excluded from the cell altogether. In the aminoplasma cells also
are generally to be seen dots, masses or strands situated anywhere and
irregularly scattered about the cell, which have no connection with the
nucleus, although they stain deeply with methylene violet.

In fixed specimens the appearance of these cells is very different and
instead of being round, homogeneous cells, they are often irregular in shape
and divided up into irregular sized loculi or balls of protoplasm, many of
which become loose and scattered about in the tissue. These balls stain with
safranin, and acid fuchsin, and give a Berlin blue reaction with ferric ferri-
cyanide. They do not stain well with pyronin, but in some specimens strands
of protoplasm are to be noticed in between the loculi which do stain with
pyronin. The strands are no doubt the same as the dots, masses and strands
which were described in the im vivo method as showing an affinity for
methylene violet.

These ballooned plasma cells have in some cases lost their nuclei, whilst
in others the nucleus is lengthened out and fits one apex of the cell like a
cap does the head, and not infrequently sends stringlike processes down over

530 J. E. R. MCDONAGH AND R. L. M. WALLIS

the cell protoplasm. These cells are no doubt degenerated cells, because the
protoplasm gives amino-acid reactions and in the most degenerated cells the
nucleus gives the histone reaction and fails to stain with methyl green.

From what has been said it will at once be seen that the syphilitic bodies
bear points of resemblance to the aminoplasma cells, but that they differ in
the very striking point that the most reducing part of the syphilitic body
stains deeply with pyronin, whilst the most reducing part of the aminoplasma
cell stains faintly with pyronin, and, as we have shown that the reducing
substance of the former is basophilic, it at once appears obvious that that
of the latter is more acidophilic.

The Berlin blue formation is a fixed chemical process between tissue and
reagent, since, although the colour can be caused to vanish with alkalies, it
immediately returns on the addition of an acid, and, as the feeble reduction
areas are more quickly decolourised than the firm Berlin blue areas, weak
alkalies may be used to decolourise the former and the protoplasm of the
cells can then be counterstained with an aniline dye.

(c) The third reaction we tried was with tetranitrochrysophanic acid
(C,;H,O,(NO,),). It is a crystalline product obtained from chrysarobin which
is dissolved in acetic acid and treated with nitric acid. The reagent is
insoluble in water, and owing to the reducing power of ethyl and methyl
alcohol it has to be dissolved and kept in chloroform or xylene.

After staining for 10 minutes the sections are returned to chloroform, put
through three changes and through xylene, and then mounted in balsam.
Weak reducing agents stain a pale rose red, strong reducing agents stain red.
The protoplasm stains pale rose red, nuclei remain unstained, syphilitic
bodies stain a deeper red, but the contrast is not so clear as in (a) and (6).

In tissues there are four chief classes of bodies :

1. Proteins. 2. Carbohydrates. 3. Fats:
4. Cholesterol, lecithin and allied lipoids.

All four groups possess reducing properties in varying degrees, but the
second we can rule out in the present discussion as will be shown later.
Therefore, the reducing substance must be a protein, a derivate of a protein,
a fat, or a lipoid. Speaking generally, pure proteins, fats (olein excepted)
and lipoids are not strong reducers, but derivates of proteins are, especially
the amino-acids, and here is appended a list of amino-acids with their
action on potassium permanganate and the ferric ferricyanide mixture
(after Unna).

J. E. R. MCDONAGH AND R. L. M. WALLIS 531

Amino-acids KMn0O, Iron mixture

Asparagine is 3

Alanine 1 Ame -

Phenylalanine

Leucine i % +

Glutaminie acid bi +

Glycokoll i one + +
Cystine... ef ae + +
Tyrosine and Tryptophane +++ ++ 4

l++ i

The amino-acids par excellence which give the Berlin blue reaction are
tyrosine and tryptophane, and to prove that these plasma cells contained
tyrosine we stained some sections in Millon’s reagent and succeeded in
getting the recognised reaction. Millon’s reaction was, on the other hand,
very feebly marked in the syphilitic bodies; therefore, the reducing substance
of the syphilitic bodies is not dependent upon tyrosine for its property.

So far as amino-acids are concerned we can say that the syphilitic bodies
contain none free, Feeling that the protein molecule existed as such, we
first directed our attention to the albumoses and repeated all Unna’s ex-
periments.

(a) The protoplasmic portion of the syphalitic organism.

The arbitrary division of albumoses is, in our opinion, not quite justifiable,
and to say that granoplasma is a very special albumose, as Unna attempts
to do, can scarcely be correct, since the granoplasma of connective tissue cells
behaves differently from that of plasma cells; the greatest difference is also to
be noticed in the individual plasma cells themselves depending upon their
stage of development and degeneration, and lastly that of organisms and
protozoa is as different again, and all behave differently according to the
method of fixation.

Unna states that granoplasma is a deuteroalbumose and not a primary
albumose. Although it differs from the former owing to its greater in-
solubility and in this respect resembles an acroalbumose which belongs to
the latter group, the assumption is made that granoplasma is a deutero-
albumose, which has probably been formed from an acroalbumose.

To make this very complicated part of our work as clear as possible, it
may be said that the most soluble granoplasma is that met with in the
connective tissue cells and groundwork, then comes the granoplasma of some
plasma cells, then of other plasma cells, then of the embryonic lymphocytes
and nucleoli, and finally of the syphilitic bodies; but here we must halt for
a moment, as in the syphilitic bodies we are dealing with two distinct
proteins, one rather granular, which stains lighter with pyronin, the other

532 J. E. R. MCDONAGH AND R. L. M. WALLIS

highly refractile, which stains deeply with pyronin; the former is the
groundwork or granoplasma of the cell and resembles ordinary granoplasma ;
the latter may cover the whole cell or only the nucleus and is extremely
resistant to reagents, and therefore does not resemble ordinary granoplasma ;
this is the protein which we will call the pyroninophile substance.

As the granoplasma of the syphilitic bodies resembles ordinary grano-
plasma, the protein of .the syphilitic bodies will be referred to as the
pyroninophile substance, since it is the chemistry of this substance that we
wish to unravel.

Unless otherwise stated, the following experiments were undertaken with
sections which had been fixed in 50 °/, alcohol and which were placed in the
different reagents for 12 hours at room temperature and then stained with
pyronin and methyl green.

1. In distilled water granoplasma begins to dissolve, the action is very
much quicker at 37°, but the syphilitic bodies remain unaltered. If kept in
for several days, the avidity for pyronin disappears and the nucleic acid is
left behind to stain with methyl green. One may say that the protein of the
syphilitic bodies is insoluble in water; this rules out protoalbumose and
deuteroalbumose. If normal saline is substituted for distilled water, the
action is much the same and the syphilitic bodies are still insoluble; by
this test acroalbumose is excluded.

2. 30°/, alcohol behaves like normal saline and dissolves a greater
portion of the granoplasma, but has no action on the syphilitic bodies.
In 60°,, 70°/,, 80°/o, 96°/,, and absolute alcohol the granoplasma remains
mostly intact, depending upon the concentration, as no granoplasma is soluble
in absolute alcohol. In no percentage of alcohol are the parasites dissolved ;
this excludes at once peptone and protoalbumose, also deuteroalbumose a,
which is soluble in 70 °/, alcohol and all the other deuteroalbumoses which
are soluble in 80 °/, alcohol and upwards, except thiodeuteroalbumose and
glycodeuteroalbumose 8 II, both of which are insoluble in absolute alcohol.

3. In a 10 °/, solution of metaphosphoric acid granoplasma and_ the
syphilitic bodies are insoluble, owing no doubt to the precipitation of all
proteins by the acid; this is likewise the case with 1 °/, phosphomolybdic
acid, alone, or with 1°/, hydrochloric, 1°/, phosphotungstic acid, picric acid,
and weak solutions of the mineral acids. These tests rule out protoalbumose,
which is soluble in picric acid and dilute mineral acids, and Kiihne’s
heteroalbumose which is also soluble in dilute mineral acids. It is very
difficult to work with the above acids owing to the fact that they all prevent
staining with methyl green, and everything stains a diffuse red with pyronin.

>
i
i

J. E. R. MCDONAGH AND R. L. M. WALLIS 533

4. Granoplasma and the proteins of the syphilitic bodies are insoluble in
all strengths of acetic acid, and the reason why the nuclei stain green, giving
the first impression that the pyroninophile substance over or in it has been
dissolved, is only due to the marked precipitating action of acetic acid on
nucleic acid.

5. Granoplasma is very soluble in a 1 or 2°/, solution of boric acid but
insoluble in a 5°/, solution; the syphilitic bodies, on the other hand, retain
their affinity for pyronin. Such a pretty instructive picture is obtained by
leaving a section in 1 °/, boric acid for 12-20 hours at ordinary temperature
and then staining in the usual way with pyronin and methyl green, that
more than a passing mention is desirable. The granoplasma of all the cells
has dissolved, the nucleoli have vanished, those embryo lymphocytes which
stain red and might be confounded with certain phases of the syphilitic
organism now all stain a brilliant green, and the only bodies which stand out
a brilliant red colour are the syphilitic parasites. So in this very simple
method we have as fine a differential stain as the Zieh] Niellsen for tubercle
bacilli.

6. In 1°/, potassium ferrocyanide not only does the ordinary granoplasma
disappear, but the protein of the syphilitic bodies does also, with the result
that only the nuclei stain with methyl green. If acetic acid is added to the
potassium ferrocyanide the granoplasma and protein of the syphilitic bodies
remain unaltered and stain in the ordinary way.

7. In a 2°/, solution of copper sulphate the granoplasma has gone,
nuclei remain, nucleoli have disappeared, also the granoplasma of the
aminoplasma cells, the groundwork protoplasm of the syphilitic female bodies
has vanished, but the pyroninophile substance over the nucleus remains
intact and stains with pyronin, and most of the spore cysts stain red.

8. In 1°/, caustic potash nuclei and all have dissolved.

9. In mercuric chloride and alcohol there is no change.

10. The syphilitic bodies remain unchanged after treatment with a
1-10 °/, solution of lead acetate; nucleoli are likewise not dissolved in this
reagent.

From these experiments it is clear that the pyroninophile protein of the
syphilitic parasites is not ordinary granoplasma, it is not an albumin,
albumose or peptone; this leaves us with only globulin. So when the
insolubility of the protein under question is considered, we think we are
justified in saying that it is a globulin, or, as we shall see later, a globulin
complex. If the sections have been fixed in absolute alcohol which prevents
the extraction of salts and acts as.a very powerful coagulant, many of the

Bioch. vir 35

534 J. EK. R. MCDONAGH AND R. L. M. WALLIS

above-mentioned substances fail to make any alteration, boric acid, for
instance, being innocuous, and the syphilitic protein does not dissolve in
potassium ferrocyanide. To produce the same results sections must be left
in the reagents for several days.

50 °/, alcohol can extract electrolytes from the cells, hence they become
less negatively or positively charged and the charge may be still further
diminished by reagents, and as electrolytes are essential for the staining of
fixed specimens, their complete abstraction will result in the pyronin not
staining. Hence, it may be quite wrong to say that this or that protein
dissolves in this or that solution, as it may be only its electrolytes which are
removed; therefore, as previously stated, the arbitrary division of the cell
proteins is not justifiable.

Absolute alcohol extracts few, if any, of the electrolytes, hence staining 1s
not interfered with. As the syphilitic protein when fixed in 50 °/, alcohol
resists reagents so remarkably, it can from what has just been said be
assumed that the salts or electrolytes are firmly bound up with the protein.
This would not be the case unless the protein was itself also bound up in a
highly organised and stable complex, so we have the first clue in this simple
observation that the pyroninophile substance of the syphilitic parasite is a
protein (globulin) complex.

(b) The nuclear portion of the syphilitic organism.

After treatment with acetic acid before staining with pyronin and methyl
green, or employing Ehrlich’s triacid mixture, in order to get the nuclei of
the syphilitic bodies to stain with methyl green, on careful examination
marked differences can be discerned between the parasitic nuclei and those of
other cells. In the former the methyl green stain gives a purer green colour,
the stain is more brillant and it is evenly distributed throughout the nucleus,
or in other words, is homogeneous. If small lymphocytes be now examined,
or the nuclei of plasma cells, and contrasted with the above, it will be noticed
that the green is darker and has a mixture of blue or black; it is duller and
moreover distributed into dots and strands, which are the chromatin bodies
and filaments. If attention is now paid to the dividing cells and the embryo
lymphocytes it will be noticed that the green resembles that met with in the
syphilitic bodies and that the stain is again homogeneous. The only phase
of the syphilitic organism which at all resembles in colour the lymphocytes
or nuclei of the plasma cells is the spore cyst, or rather the sporozoites which

|
|

J. E. R. MCDONAGH AND R. L. M. WALLIS 535

the cyst contains. This very simple observation is yet another very im-
portant argument in favour of McDonagh’s view, since the nuclei of the
developing syphilitic bodies resemble those of the developing lymphocytes
and the dividing plasma cells, while the sporozoites which have no further
need to develop and are in the resting stage resemble the mature lymphocytes
and resting plasma cells. Degenerated nuclei behave quite differently ; the
division into chromatic filaments becomes more marked, the methyl green
stains them bluish, then not all and at this stage the slender chromatic
filaments stain red with the pyronin, or a diffuse red stain of the remaining
protein may result; therefore, the syphilitic bodies are not degenerated
nuclei.

The chemical substance of the nucleus is a compound of a protein and
nucleic acid and is therefore called a nucleoprotein. This nucleoprotein on
hydrolysis breaks down into protein and nuclein; the nuclein into protein
and nucleic acid; the nucleic acid into purine bases, viz. guanine, adenine,
xanthine and hypoxanthine, pyrimidine bases, viz. thymine, cytosine and
uracil, pentoses and phosphoric acid.

Methyl green only stains nuclein and nucleic acid, whilst the protein
stains with pyronin; therefore the reason why degenerated nuclei stain in
some cases with pyronin, is that the nucleic acid has become further split up,
while the protein remains behind.

(c) Action of reagents on nucler.

We next tried a series of experiments by leaving sections which had been
fixed in 50°/, alcohol for 20-24 hours at room temperature in several reagents
to see if we could get any different reactions with the nucleic acid of the
syphilitic bodies, and that of ordinary cells. The sections were stained with
pyronin and methyl green, and every experiment which was undertaken was
also repeated with sections of the soft roe of a herring (Clupea harengus).

1. Leaving a section of roe for 20 hours in a concentrated solution of
ammonia results in a partial disappearance of the nucleic acid, but the cells
still stain with methyl green and retain their form; the chief difference from
the normal is the appearance of a diffuse mass of protein which stains red
with pyronin and is no doubt a mixture of histone and protamine.

In the syphilitic section the nucleoprotein has also been broken up; in
some nuclei the nucleic acid has disappeared altogether, in other nuclei there
are masses of it left as the red field is dotted here and there with some blue
masses (methyl green), The granoplasma of the cells has dissolved. The

35—2

536 J. E. R. McDONAGH AND R. L. M. WALLIS

nuclei of the syphilitic bodies have partly gone and the granoplasma has
completely disappeared and also the pyroninophile substance.

Nucleoli have mostly gone, and the aminoplasma cells completely. The
best maintained of the syphilitic bodies are the sporozoites in the spore cysts
which still stain quite intensely with methyl green and are, on the whole,
even less damaged than the nuclei in the herring’s roe. Therefore, thie
sporozoites are not only extremely rich in nucleic acid but also extraordinarily
resistant to chemical reagents.

2. After leaving sections of roe in saturated sodium chloride solution all
the nucleic acid disappears and all that is seen is a diffuse mass of histone
which stains well with pyronin.

In the syphilitic sections the granoplasma of the cells is well preserved,
if anything the pyronin staining of the protoplasm of the plasma cells is
increased. The nuclei on the other hand are very much altered; they stain
homogeneously a slate-grey colour; the chromatic bodies and filaments are
not to be seen, but the nucleic acid is less disturbed than in that of the fishes’
roe. The nucleus of the syphilitic parasite does not stain with methyl green,
not because the nucleic acid is dissolved but because the brilhant refractile
pyroninophile substance has been precipitated and therefore has had its
properties intensified.

The nucleoli are well preserved and the aminoplasma cells are intact.

3. A section of roe which has been treated with a 1 in 3 solution of
magnesium sulphate has lost all its nucleic acid, no cell outline is even
discernible and all that remains is the precipitated histone, which stains
especially brilliantly with pyronin.

In the syphilitic sections the granoplasma has partly dissolved, the nuclei
stand out, stain a brilliant green and are intensely refractile, looking like
pieces of green glass. The great difference between the nuclei found in the
roe and in the syphilitic section can possibly be explained by the fact that
in the latter the nucleic acid has not been extracted before the precipitation
of the histone, with the result that the nucleoprotein is maintained, as is the
case in sections which have remained in potassium ferrocyanide. All nucleoli
have vanished. The nuclei of the syphilitic phases stain a brilliant green and
are much better preserved than the nuclei of other cells; all the pyroninophile
substance has completely disappeared.

4. The only difference noticed in a section of roe which has been in a
01 °/, solution of calcium chloride is that a trace of histone has been ab-
stracted from the nuclei.

In the syphilitic sections the staining is not quite as good as under

|
|
|
}

J. E. R. MCDONAGH AND R. L. M. WALLIS 537

ordinary circumstances. The pyroninophile substance is not destroyed, but
the sporoblasts and sporozoites show a greater affinity for methyl green than
usual; therefore a trace of the body is soluble in calcium chloride. These
sections show up some points;which we have frequently observed in other
sections, and which have more than once raised the question as to whether
we were dealing with spore cysts or the degenerated nuclei of plasma cells.

In some spore cysts which stain with methyl green, bright red bodies are
to be seen resembling nucleoli. There may be one or more, varying in size,
and frequently the largest red mass lies on and looks as if it was part of the
biggest spore body. In this latter observation lies the solution. This large
spore body is a part of a sporoblast which has to divide still further to form
sporozoites and, as mentioned above, the oldest sporozoites have generally
lost their pyroninophile constituent, which would not be the case with
undeveloped sporozoites. The pyroninophile substance is no doubt separated
off and breaks up into fragments in the groundwork of the spore cyst.

When the nucleus of a plasma cell degenerates it does not break up into
a mass of circular bodies, but into a ring of bodies, some of which are circular
and others oval; and there is nothing in the centre except a few strands
which stain red with pyronin.

5. In 2°5°/, sodium chloride all sections show the points brought out by
the preceding reagent in a slightly more pronounced degree.

6. 20°/, ammonium sulphate solution has the effect of so breaking down
the nucleoprotein of the nuclei found in a section of roe that no distinct
nuclei are seen, only a diffuse red purple colouration of the nucleic acid being
left upon a deeply red diffuse groundwork of histone.

In syphilitic sections destruction is not nearly so marked, so far as the
toute ensemble is concerned; the pyronin stain seems to have been increased
probably owing to some abstraction of histone from the nuclei, as the staining
capacity of the nuclei is very much weakened. The pyroninophile substance is
maintained.

If specimens fixed with absolute alcohol are used, quite different results
are again obtained. Nuclei are not affected by concentrated ammonia, but
the pyroninophile substance dissolves. In 20 °/, NaCl nucleoprotein has been
split up and part of the nucleic acid dissolved, but the pyroninophile substance
is unaltered. If the syphilitic bodies are closely examined, a clear white halo
surrounds the nucleus and tiny areas of white are to be seen in the nucleus.
The refractility of the protoplasm of the plasma cells is markedly diminished
while that of the syphilitic bodies appears to be increased, which has the
effect of strongly differentiating them from the other cells.

538 J. E. R. McDONAGH AND R. L. M. WALLIS

-%. In 06°/, lithium carbonate the granoplasma of cells has mostly
disappeared, the nuclei stain homogeneously and of a bluish colour; nucleoli
are well maintained, and the syphilitic bodies remain practically unaltered.
So here again is a fine differential method.

(d) Action of sera on cells.

We thought it possible that something might be learnt by treating
sections with different sera, so the following experiments were undertaken.

Several cc. of blood were withdrawn from a non-syphilitic, a case of early
secondary syphilis, and a case of late tertiary syphilis. Great care was taken
to have the serum absolutely free from haemoglobin, and when put into the
watch glass, a covering of pure toluene was used to prevent any bacterial
action. Unless the latter precaution is taken bacteria multiply in the sera
and exert a pronounced hydrolytic action on the sections so that even after
20 hours the individual cells are only just discernible. All sera have the
same action as normal saline. If specimens fixed with absolute alcohol are
treated in the same way, there is no change and in neither case can any
difference be detected between the action of the three sera.

Whether the pyroninophile protein of the syphilitic bodies is part and
parcel of the nucleus or only its sheath is at first sight difficult to ascertain,
but we incline to the latter view for the following reasons.

Morphologically it looks more like a cover, this being especially noticeable
in potassium permanganate specimens which have been counterstained with
methyl green, as the nucleus stains faintly and appears hazy and gives
exactly the impression of being covered with a veil. It is soluble in
potassium ferrocyanide and when sections are left for 12 hours in a 1°,
solution the nuclei stain better than ever and appear clearer with methyl
green, which would scarcely be the case if it was part of the nucleus.

Moreover, it gives a faint Berlin blue reaction, which no nucleus is known to

do. Further, by staining with dextrose-borax-methylene blue, during im-
pregnation, there appears to be a marked differentiation of ecto- and endoplasm
and even the spirochaetae are swollen. If it is not part of the nucleus, in
what combination with protein does the nuclein and nucleic acid of the
syphilitic bodies exist? That a protein does exist is quite clear, since the
nuclei of the syphilitic parasites can be made to stain with acid dyes, viz.
diazine green, which mixes very well with pyronin; and moreover, will stain
well with pyronin when the nucleic acid has been separated off, when it is
found to behave like ordinary granoplasma. In this respect the protein

ory Ase ee eiees S @ S

J. KE. R. MCDONAGH AND R. L. M. WALLIS 539

radicle of the syphilitic nucleoprotein does not differ from that of ordinary
cells.

The protein of nucleoprotein is always regarded as quite a special protein,
and said to possess strong basic properties and to be extremely rich in hexone
bases, viz. lysine, arginine and histidine.

Nucleoprotein is a complex in which the properties of the protein can be
very materially altered by the prosthetic group, since we are by no means
aware what all the other constituents are and therefore cannot tell when
they have been removed ; we can never be sure that we are dealing with the
protein only.

We have already shown that the division of the proteins of the cell
protoplasm is purely arbitrary and the separation of the nucleoproteins is
likewise no doubt artificial, since when the protein of the nucleoprotein is
separated out it gives the same micro-chemical tests and behaves in the same
way to stains as does the cell granoplasma; therefore the syphilitic nucleus
does not differ in gross details from the nuclei of its host’s cells.

For lysine and arginine we have as yet been unable to find a specific
microchemical test, as immersing sections first into a freshly prepared solution
of diazobenzene-sulphonie acid and then washing with a 1°/, solution of
NaOH, did not give the characteristic pink colour which is seen in the
microchemical test for arginine. Reversing the solutions by using the alkali
first made no difference. Whichever method was employed a beautiful stable
orange colour, which stained both the protoplasm and the nuclei, resulted.

The well-known test for histidine, namely, bromine water and acetic acid,
also gave negative results, both before and after hydrolysing with a 0°5°/,
solution of hydrochloric acid. Oddly enough a section which had been treated
with bromine water and acetic acid quite failed to give the orange colour
with diazobenzene-sulphonic acid and sodium hydroxide, which looks as
if either the histone or the hexone bases had formed a halogen compound,
which of course they are well known to do, and the bromo-histidine compound,
for instance, does not give the characteristic reactions of the base.

Failing to get the typical hexone reactions we did not expect nor did we
succeed in getting a positive imidazole reaction by treating sections with
ammonia and silver nitrate. We also tried, but once more with negative
results, to get a positive biuret reaction which is obtained with histone; the
failure was due to the destruction of the tissues by the strong soda solution,
which is unfortunately necessary for the reaction.

All nuclei are stated to contain phosphorus, which is intimately bound up
with the nucleic acid radicle of the nucleoprotein and disappears when the

540 J. E. RX MCDONAGH AND R. L. M. WALLIS

nucleic acid is hydrolysed, and as it is the nucleic acid radicle which shows the
affinity for methyl green it is possible that this acid plays a part in the selective
action of nuclei for methyl green. The nuclei of the syphilitic parasites, when
the pyroninophile membrane covering them is removed or prevented from
staining, stain not only brilliantly with methyl green but also show a greater
resistance to hydrolytic agents than do the nuclei of ordinary cells; hence it
might be expected that the parasitic nuclei were especially rich in phosphorus.

To see whether this surmise was correct, the following experiments were
undertaken :

Phosphorus. Both fresh sections and specimens, fixed with absolute
alcohol, the latter giving quite as good results as the former, were placed in
a mixture of molybdic acid, ammonia and nitric acid, and kept therein at
37° for 10 minutes to 48 hours. One is supposed to regard as inorganic
phosphorus that which makes its appearance in the first 10 minutes, but
when sections are examined so early only negative results are obtained.
After 24 hours good staining effects can be obtained, but the staining is
sharper if the sections are left in the mixture for even another day.

The sections are washed well in distilled water and are then placed in a
2°/, solution of phenylhydrazine hydrochloride, taken direct through alcohol,
and mounted in balsam.

The presence of phosphorus is indicated by a green colour, and when the
sections are examined it is found that both the protoplasm and the nuclei
of nearly all cells are stained. The staining is deepest in the syphilitic
bodies and then in the plasma cells; there appears to be no great difference
in the staining properties of the nucleus compared with the cell protoplasm,
which indicates that the phosphorus in a cell is not restricted to the nucleus.

As nuclei also contain iron the following tests were undertaken :

Iron. A. Inorganic.
B. Organic.

(A) Inorganic. Specimens fixed with absolute alcohol were after
removal of wax transferred direct from absolute alcohol into a freshly prepared
solution of equal parts of 0°5°/, hydrochloric acid and 1°/, potassium
ferrocyanide, and allowed to remain therein for one hour. By this means the
cells-did not give the Berlin blue reaction, nor did they even stain green.

Instead of the hydrochloric acid potassium ferrocyanide mixture, a 0°5°/,
haematoxylin solution was employed, which also failed to prove the presence
of inorganic iron.

(B) Organic. Sections prepared as above were placed in a mixture

J. E. R. MCDONAGH AND R. L. M. WALLIS 541

of sulphuric acid (4 vols.) and absolute aleohol (100 vols.) and left therein at
37° for 24-48 hours.

After 24 hours the sections were washed in absolute alcohol and some
were transferred for half an hour into the hydrochloric acid potassium
ferrocyanide mixture, whilst others were stained in 0°5 °/, aqueous solution
of haematoxylin. From both solutions the sections were taken through
alcohol, ete. and mounted in balsam. When the former were examined it
was seen that the protoplasm of the cells remained unstained while the nuclei
stained green, the arrangement of the chromatin remaining unaltered.
Other nuclei stained a light Berlin blue and the colour was homogeneous ;
examined for action on polarised light they were found to react slightly, while
one of the other cells showed a trace of reaction. These cells were no doubt
the syphilitic parasites; the nuclei of which contained more organic iron
than those of the leucocytes and connective tissue cells.

The same difference in degree of staining was also noticeable in the
haematoxylin specimens.

Summary. The protoplasm of the syphilitic bodies is strongly pyronino-
phile, which proves it to possess reducing properties. The reducing action is
not so strong as that of the aminoplasma cells and therefore is not due to an
amino-acid; and moreover it prefers basic to acid dyes which still further dis-
tinguishes it.

The protoplasm is very resistant to reagents and in all respects resembles
a globulin. The nucleus in its behaviour to dyes and reagents most closely
simulates the nuclei of dividing cells.

Hence, neither can the syphilitic bodies be taken for cell degenerations

nor nuclear degenerations.

FURTHER POINTS CONCERNING THE REDUCING ACTION OF CELLS,
WITH SPECIAL REFERENCE TO THEIR PHYSICAL CHARACTERS.

None of the proteins we have mentioned have sufficient reducing power
to form Berlin blue from the ferric ferricyanide reagent; therefore, the
reducing action of some of the syphilitic bodies, red blood corpuscles, ete.,
must be due to a carbohydrate, a fat or a lipoid.

Carbohydrates. Our endeavours to find a carbohydrate in the syphilbtic
bodies failed and the tests used are only of interest from the negative
information derived therefrom.

Keeping sections for some time in a hot solution of Fehling’s reagent gave
no yellow precipitate. Leaving sections for 48 hours at 37° in a solution

542 J. E. R. McDONAGH AND R. L. M. WALLIS

of 0'5°/, potassium hydroxide in 90°/, alcohol and then immersing them into
a 2°5°/, solution of dimethylparaminobenzaldehyde in 1°/, HCl gave no
carmine reaction, which indicates the absence of glucosamine.

Treating sections with a 15°/, alcoholic solution of a-naphthol and then
examining them in sulphuric acid for the furfurol reaction, owing to the
destruction of the tissue, gave no information.

The principal organic tests and group reactions either necessitate the use
of strong acids or alkalies; with the former the tissue is as often as not
dissolved in toto; with the latter it is almost certain to leave the cover slp,
swell and become practically unmanageable. Ammonia is the least offender
in this way, but unfortunately it cannot as a rule take the place of the
potassium or sodium hydroxides. |

We tried also leaving fresh and fixed sections in a saturated solution of
copper acetate at 40° for 24 hours and then washing them in a strong
solution of sodium carbonate. A general reduction of the copper occurred in
the fresh sections, but the individual cells could not be studied. In the fixed
films the only cells which reduced were those of the rete malpighii. Therefore,
although we can say that the tissue cells do contain sugar, we cannot differ-
entiate them individually.

We therefore resorted to a physical test.

By the use of Nicol’s prisms it is seen that the syphilitic cells when
properly focussed appear as bright stars against a black background. The
phenomenon is better marked in specimens fixed with absolute alcohol than
when 50°/, alcohol has been used for the same purpose, and can be beautifully
demonstrated in sections stained with pyronin and methyl green. It is most
marked over the nuclear area, is greater in zygotes than in female gametocytes,
is very well marked in the trophozoite and male gametocyte and much less
evident in sporoblasts and sporozoites. In short, it is greatest where the
pyronin staining is deepest; therefore, it is the pyroninophile substance that
is concerned. The only cell constituents which exhibit this property are
cholesterol, sugar, and lipoids.

Cholesterol gives a crystalline, and not a star-like appearance as seen in
the syphilitic parasites, and moreover cholesterol is not increased in the serum
of syphilitic patients. Sugar can be excluded, because if it were present in
sufficient quantities to show the phenomenon, it would give the micro-chemical
tests; moreover sugar is soluble in water which the pyroninophile substance
is not. Further, the active substance of the parasitic cells must therefore be
lecithin, or rather its fatty acid constituent. To bring still greater evidence
to bear on this assumption we compared the colloidal particles in a case of

J. KE. R. McDONAGH AND R. L. M. WALLIS 543

pseudochylous fluid with dextrose and in a case without dextrose, with the
result that the former were more active than the latter, the activity of which
was comparable in degree with that of the syphilitic cells. The pseudochylous
fluid with dextrose gave a marked Berlin blue reaction, which was not the
case with the other. The fluid containing no dextrose had a reducing action
equal to that of the syphilitic cells; therefore it appears justifiable to state
that the syphilitic bodies contain no dextrose.

Leaving sections for hours and even days in ether, absolute alcohol, and
absolute alcohol and ether mixed, the cells still retain this property and
therefore the lecithin cannot exist alone. The pyroninophile substance
contains protein, and as a lecithin-protein complex is known to exist and the
protein is a globulin which we have also shown the pyroninophile substance to
be, we can assume that this substance is lecithin-globulin.

Fatty acids. Now what evidence have we for assuming that it contains
a fatty acid? We stained sections with iodine and with Nile blue sulphate,
aceording to the method of Lorrain Smith. The syphilitic bodies stained
with iodine and also very deeply with a saturated solution of Nile blue
sulphate. Red blood corpuscles also stained with iodine and all nuclei and
nucleoli stained with Nile blue sulphate. The only granoplasma that stained
with Nile blue sulphate was that of the syphilitic bodies and some of the
plasma cells. The inclusion of a fatty acid in the lecithin-globulin complex
is still further supported by the invariable occurrence of a saturated fatty
acid (stearic acid) in the lecithin-globulin complex of pseudochylous fluids,
and the visibility of the fluid between crossed Nicol prisms is due to the fatty
acid and not to the lecithin itself.

The reducing action of red blood corpuscles, syphilitic bodies, protoplasm
of some plasma cells and nucleoli is in all probability due to the lecithin-
globulin complex and the change that takes place in impregnated female
cells is an increase of this body which, owing to its reducing action, prevents
the cell from staining with methyl green, and, owing to its acid action,
prevents the cell from showing a marked affinity for negatively charged dyes, as
carbol fuchsin, ete. Since the complex is not so marked in unimpregnated
female cells and the protein is less acid, it will stain with carbol fuchsin
in the carbol fuchsin-carbol iodine green method. Owing to the staining
reactions, the fatty acid must be a saturated one, and as neither the syphilitic
bodies nor the colloidal particles of pseudochylous fluid stain with osmic acid
or Sudan ITI, the fatty acid is clearly not oleic acid.

As the syphilitic parasite was shown to contain lecithin we thought
it necessary to repeat the more important work which had been done in

544 J. E. R. McDONAGH AND R. L. M. WALLIS

connection with the lecithin in nerve tissue, so the following tests were
carried out.

1. The tissue was fixed in Miiller’s fluid for 10 days, frozen and sections
cut. The sections were transferred into 1°/, osmic acid for 24 hours at 37°
and then placed in the followimg mixture :—

Pyrogallic acid i uae parts 15
Sod. sulphite 6 wae eee
Sod. nitrate ... pe oe s (AY)
Water oe Bi sc 300

and differentiated in 0°1°/, potassium permanganate which reoxidises the
osmium which has not combined. The brown colour of the permanganate
can be removed if desired with 1 °/, oxalic acid.

Considering how insoluble the syphilitic lecithin. appeared to be in
alcohol, we tried alcohol fixed specimens and paraffin sections, but instead of
floating the cut sections out on to warm water, we employed a hot 7°/,
solution of potassium bichromate and obtained quite as good results. The
syphilitic bodies had a greater reducing action upon the osmium than other
cells, staining a deep olive green, the only other structures which resembled
them were nucleoli and the aminoplasma cells.

2. Tissues were fixed for four days in 10°/, formalin and were then
placed for the same length of time in Weigert’s chrome alum copper acetate
mixture in the incubator at 37°. Some sections were cut in the frozen state,
others taken through paraffin. The cut sections were put into sulphuric acid
alcohol (1:500 H,SO, in 50°/, alcohol), and then stained for 10 minutes in
1°/, osmic acid ; they were then well washed and treated with 5°/, pyrogallic
acid solution, differentiated in 0-1 °/, potassium permanganate taken through
sulphurous acid and the fresh sections were mounted in liquid paraffin ; while
the paraffin sections went through the usual stages to balsam. The results
tallied with those obtained from No. 1.

3. Alcohol fixed specimens, the paraffin sections of which had been
floated out on to potassium bichromate, were stained for 24 hours at room
temperature in Weigert’s haematoxylin.

The syphilitic bodies stain intensely and in this respect resemble the
nucleoli. The chromatin of nuclei also stains well, but the colour is not so
fast.

4. Paraffin sections of tissue which had been floated out on potassium
bichromate were stained for 24 hours at 37° in Kultschitzky’s haematoxylin,
washed and decolourised in a 0°25 °/, solution of sodium carbonate until the
sections retained a light blue colour. The sections were then washed,

|
|
,

OP Ae RIDE wl

J. KE. R. McDONAGH AND R. L. M. WALLIS 5A5

transferred to a 0°1°/, solution of potassium permanganate, then into
sulphurous acid, washed in a 01°, solution of lithium carbonate, taken
through alcohol and mounted in balsam.

The sections gave the same results as were obtained from method No. 3.

All sections were compared with and controlled by sections from a case of
neurofibromatosis, with the result that the reducing action of the syphilitic
bodies was practically equal to that of the medullated nerve fibres, and the
capacity for stainig with haematoxylin resembled that of the axis cylinders,
the deep colour of which is no doubt due to the presence of a lecithin-protein
complex.

Summary. The reducing action of the pyroninophile substance of the
syphilitic bodies is not due to cholesterol or a carbohydrate. Although it is
active optically, it is not so markedly so as the two substances just mentioned.

The activity is due to a fatty acid and resembles that of the particles met
with in the fluid from a case of pseudochylous ascites. The particles of such
a fluid consist of lecithin-globulin with a saturated fatty acid (stearic acid) in
its radicle. This physical phenomenon is a most useful means of picking out
the syphilitic bodies in a section.

OXIDISING ENZYMES.

Our next efforts were to ascertain whether the presence of an oxydase or
a peroxydase could be detected in the syphilitic bodies, since some observers
have stated that the bodies which have been described as parasitic are really
mast cells. For this purpose we fixed some tissue in formalin embedded in
gelatin, and cut sections after freezing the material with carbon dioxide.

In testing for oxydases, the following methods were employed :—

1. Sections were placed in a 1°/, alcoholic solution of benzidine, but failed
to stain, proving that no oxydase could be demonstrated; therefore, the
syphilitic bodies do not harbour a tyrosinase, a ferment which at first sight
might be expected to be present, owing to the increased pigmentation which
is somewhat characteristic of syphilitic lesions.

2. Sections were placed in a 1°/, alkaline solution of a-naphthol
(prepared by heating 1 grm. anaphthol in 100 ce. dist. water and then
adding just sufficient KOH to bring the a-naphthol into solution) for a few
minutes and then transferred to a 1°/, sol. of tetramethylparaphenylene-
diamine hydrochloride (violamine) washed in distilled water and mounted in
glycerol or liquid paraffin. This method also failed to show the presence of
an oxydase.

546 J. E. R. McDONAGH AND R. L. M. WALLIS

The following methods used for demonstrating the presence of a per-
oxydase met with more success :—

1. Sections were placed in a mixture of equal parts of a 1°/, alcoholic
solution of benzidine and 3°/, hydrogen peroxide.

Such sections can be taken through alcohol, etc. and mounted in balsam,
or of course mounted in the ordinary way. A very strong peroxydase reaction
was found in the red blood corpuscles and mast cell granules. The syphilitic
bodies remained unstained.

It is owing to the presence of peroxydases that the mast cells stain orange
with pyronin.

2. Equally good results can be obtained by leaving sections in a mixture
of paraphenylenediamine tartrate (ursoltartrate) and H,0,.

The presence of a peroxide in tissues cannot be demonstrated, as no blue
colour results on leaving sections in a mixture of benzidine and extract of
horseradish. The above affords sufficient evidence against the view that the
syphilitic bodies are mast cells.

PHYSICAL AND CHEMICAL PROPERTIES OF THE
LECITHIN-GLOBULIN COMPLEX.

The lecithin-globulin complex when in solution gives rise to a definite
opalescence, and exists as a colloidal suspension which can be removed by
filtration through a Chamberland candle. It appears microscopically in the
form of very fine refractile granules which do not stain with osmic acid, or
Sudan III. These fine granules no doubt owe their refractility to the
associated lipoid lecithin. The particles exist in an alkaline medium and
possess a negative charge, and in consequence we find that the lecithin-
globulin complex is readily precipitated by kations, particularly the divalent
kations, Mg, Ba, and Ca.

The lecithin-globulin compound is readily precipitated by acetic acid even
in the cold, also by alcohol; half saturation or full saturation with ammonium
sulphate, and removal of the salts by dialysis results in separation of the
lecithin-globulin complex. Treatment of the complex with ether has no
effect and previous addition of alkali such as caustic potash does not
appreciably alter the solubility of this body. In all respects this lecithin-
globulin behaves exactly like the pseudo-globulin fraction isolated from serum.
One third saturation with (NH,).SO, precipitates the euglobulin fraction
from serum and this fraction is soluble in a 0°6 °/, solution of sodium chloride,

J. KE. R. MCDONAGH AND R. L. M. WALLIS 547

whereas the pseudo-globulin remains insoluble. The pseudo-globulin fraction
of the serum thus behaves in every way like the lecithin-globulin compound.
The solubility of the globulin present in serum or in the body cells will
therefore depend upon the amount of lipoid in association with the globulin,
and the former will influence the optical properties, the electrical charge on
the colloidal particles in suspension, and also the relationship of globulin to
electrolytes. In connection with these observations, it may be noted how
important both constituents of the complex are in the maintenance of life.
During starvation, for example, the blood contains a larger amount of
globulin, and after excessive bleeding the first constituent of the blood to
return to its normal amount is the globulin fraction.

With regard to the nature of the lpoid present in association with the
globulin, it is usually found to be lecithin. This lecithin is generally of the
type described as a monoaminophosphatide, yielding choline on hydrolysis, and
fatty acids of the stearyl group. The lecithin is insoluble in water, and is so
firmly united to the globulin as to remain undissolved when treated with
ether. The production of lecithin is probably determined by the processes of
degeneration of cellular material, which takes place in diseases in which
effusions may result, viz. tuberculous infections, malignant disease, and
syphilis. In the production of milky effusions it seems highly probable that
the destruction of lecithin-containing cell elements takes place in the blood
itself, and that the lecithin so formed diffuses through the peritoneal
membrane into the serous cavities. This would explain the sudden changes
from a clear effusion to a milky one noted by some observers, or the reverse
condition where a milky is later replaced by a clear transparent fluid.

The resistance to putrefaction exhibited by all fluids containing this
complex is very striking in view of the fact that lecithin readily undergoes
auto-oxidation when free. The complex seems to confer increased stability
upon both constituents, but at the same time the lecithin is capable of fully
exerting all its special functions, particularly its influence on the neutralisa-
tion of toxines and bacterial growth. The chemical configuration of the
lecithin molecule possibly accounts for this property, since it contains a
large number of hydroxyl groups capable of uniting with such bodies as
ferments, proteins, sugar, and other lipoids. The power lecithin possesses of
resisting putrefaction suggests a possible function of this body in the produc-
tion of immunity.

From what has been stated it will be noticed how close is the resemblance
between the pyroninophile substance of the syphilitic parasite and the
colloidal particles of the pseudochylous fluid, both substances being no doubt

548 J. E. R. McDONAGH AND R. L. M. WALLIS

identical and the relationship becomes the closer when the staining properties
of both are considered, and for brevity’s sake we need only state that both
the colloidal particles and the syphilitic parasites are Gram negative.

REFERENCES.

McDonagh, J. E. R. (1913, 1), Proc. Roy. Soc, Med. (Path. Sect.), 6, 85.
(1913, 2), Dermatolog. Wochensch. 56, 413.
Unna, P. G. (1913), Biochemie der Haut (Verlag von Gustav Fischer).

NOTE:

CHICK & MARTIN. The Precipitation of egg-albumin by ammonium sulphate. This Journal,
vir, 1913, p. 380. In Table VIII, p. 392, 4th column, the values given for the percentage by
weight in the Pressed Precipitate of (NH4)2SO4 and water respectively should be interchanged
and the Table should read thus:

Composition, °/, In pressed

Exp. by weight of precipitate
H,O 29-83
II (NH4)2S04 8:47
H,0 29°91
Ill (NH4).S04 6°39

H,0 22°04

LI. ENZYME-ACTION, FACTS AND THEORY.
By HENDRIK PIETER BARENDRECHT.

From the Laboratory of the Netherland Yeast- and Spirit-Manufactory,
Delft, Holland.

(Received August 30th, 1913.)

The researches of the last few years on the kinetics of enzyme action
have brought more confusion than clearness into this field. Even about the
action of the best studied enzyme, invertase, there is contradiction and
uncertainty; it seems as yet not quite established if, or under what con-
ditions, the hydrolysis of cane-sugar by invertase follows the simple law
of mass-action.

As to the nature of enzyme action there is a general inclination to
suppose, that some chemical combination of enzyme and substrate plays a
leading part in the process and that the general properties of the colloids,
to which enzymes seem to belong, will supply sufficient explanation of their
remarkable activities.

The aim of this paper is to clear up some of the contradictory statements
in the literature about the kinetics of the most simple enzyme actions, and to
show at the same time that there is another theory, which is more in
accordance with the peculiarly characteristic behaviour of enzymes.

The hydrolysis of cane-sugar by invertase has been chosen by many
investigators as the best process on which to study the velocity of enzyme
action in relation with the different factors, which are here of importance.
Both enzyme and hydrolyte are readily accessible, the action is simple, starts
with the well-known substance saccharose and leads to the also fairly well-
known substances glucose and levulose.

Chemically speaking, this hydrolysis is a monomolecular reaction. It
was therefore a confirmation of what chemists expected, when O’Sullivan
and Tompson [1890] found in an elaborate investigation, that the inversion
of saccharose by invertase from yeast proceeded in accordance with the
law of mass-action, i.e. if y represents the inverted fraction at a time f,

1 1
k=- log [ay Was nearly a constant.

Bioch. yu 36

550 H. P. BARENDRECHT

- Duclaux [1898] however, without doubting the correctness of the experi-
mental results of O’Sullivan and Tompson, pointed out, that there was still
a remarkable difference between the action of acids and invertase (and of
enzymes generally).

In differently concentrated solutions of saccharose the same amount of
acid gives the same velocity-constant k, thus for instance a 5, 10 or 15 ah
saccharose solution will be inverted by a 1 »/, acid solution in the same time
to the same extent, eg. half way. But for invertase Duclaux had found
as early as 1883, that the same quantity of enzyme, acting in different
concentrations of cane-sugar, inverted in the same time not the same fractions
but the same absolute amounts. ; |

Henri [1901] then took up again the experimental study of invertase and
found, that not only was there the difference between the actions of enzymes
and acids, established by Duclaux, but that & did not even remain constant
during one and the same action of invertase on cane-sugar; on the contrary
steadily increased.

Now it cannot be denied, that Henri’s experiments were open to criticism
owing to his not taking sufficiently into account the mutarotation of the
freshly formed glucose in his polarimetric estimations.

It was therefore a valuable contribution on the experimental side of the

subject, when A. Brown [1902] also investigated the kinetics of the invertase-
action and found the same regular increase of the constant k= logs
for sufficiently concentrated saccharose solutions as Henri. On p. 375 of his
article Brown says: “Conditions of experiment similar to those used by
C. O'Sullivan and Tompson were employed, but the invertase used was
prepared in a different manner from the enzyme with which these authors
experimented.”

To this last point we will return further on.

In order to remove all doubt, possibly also expressed by others as to the
correctness of Brown's estimations, I wrote to him on this subject and was
assured, that he “was well aware of O’Sullivan’s experiences with regard to
mutarotation and took exactly the same precautions to avoid errors in
connection with it, which O'Sullivan adopted.”

As I myself [1904] found the same deviation from the simple law of
mass-action as Brown, without using a polariscope at all, but estimating the
hydrolysis of cane-sugar by the accurate gravimetric method of Kjeldahl with
Fehling’s solution, Hudson’s [1908] criticism of his predecessors (excepting
Henri) must be rejected. How it was, that Hudson could again find validity

H. P. BARENDRECHT 551

of the simple formula of mass-action in his experiments with mvertase was
explained by the work of Sérensen [1909].

In his elaborate article on: “The measurement and importance of the
concentration of hydrion in enzyme-action” this author shows, that low
concentrations of hydrion in the invertase-solution give the well-known
increasing k. Going beyond this concentration one can work with an
invertase-solution, where the reaction velocity follows the formula

1 1
k=; log

and in still more distinctly acid solutions / diminishes even more and more
rapidly during the reaction. That this irregularity in the reaction-velocity
is due to a gradually decaying of the enzyme by the action of the hydrion
was perhaps not sufficiently emphasised in this article, though it is clearly
expressed on p. 146 for the series with the highest hydrion concentration,
where the reaction did not even come to a finish.

Indeed it follows from Sérensen’s observations and deductions, that 1t is
dangerous to work with enzymes at their optimum concentration of hydrion
as well as at the optimum temperature of their activity. Both factors in
that condition give splendidly rapid enzyme action, but at the same time are
gradually destroying the enzyme itself.

It is obvious, that, where the quantity of the active catalyst is diminishing
during a reaction, the velocity of this reaction will diminish at the same
- rate and may thus seem to fulfil the simple formula of mass-action. With
low hydrion concentration, that is without adding acid or using extracts
that have become acid by deterioration, one always gets for invertase (provided
the saccharose solution is sufficiently concentrated) the same increase in the

1 1 Aes
constant k, calculated as b =o lens , or a real constant *, when it is

calculated according to the empirical formula of Henri or to the theoretical
formula, which I developed in 1904.

Sérensen drew attention to the fact, that O’Sullivan and Tompson
brought their solutions by the addition of sulphuric acid to about the
maximum of activity of the invertase, while Hudson worked with solutions
containing acetic acid. Now we can add to this, that Brown “employed in
(his) experiments an extract of invertase, prepared from dried yeast by
digestion with water,” and “without the complicating addition of sulphuric
acid.” In my own experiments I used an extract of fresh yeast, mixed
with kieselguhr and dried rapidly at a low temperature. No acid at all
was used,

36—2

552 H. P. BARENDRECHT

When examining the results of O’Sullivan and Tompson as closely as
these authors themselves did, one sees clearly, that the acidity they employed
had not reduced the real velocities quite to an extent sufficient for the law
of simple mass-action to be perfectly imitated. On p. 846 they say: “But
if we look more closely at the diagram, we see that the deviations from the
theoretical line, though small in themselves, are constant in the three
experiments; in fact they correspond in a very striking manner.” And
a little further: “These figures all show a continuous increase in activity
up to a certain point, about 80 per cent. of inversion, and after that a large
and increasing decrease.” That is exactly what the gradual destruction
of the invertase must produce. |

Sdrensen goes however too far, when he generalises this influence of
hydrion concentration to suggest a sufficient explanation for the marked

falling off of the constant k = flog ps in the experiments of Armstrong

with lactase. My own experiments as well as the theory of enzyme action,
which I proposed some years ago, lead to a different explanation and to a
real reaction-constant in this case also.

As the application of this theory may now be extended in some points
beyond what was published before, it may be of interest to survey once
more a part of this much entangled field of enzyme actions with the aid of
this theory.

Any hypothesis of enzyme action must be founded on the principal
characteristics of this kind of catalysis.

It seems more and more established that these are:

1. That the same quantities of enzyme, brought into differently, but
sufficiently, concentrated solutions of substrate, produce in the beginning
the same action in the same time.

2. That catalysis by enzymes has a marked specific character’.

Starting from these general facts, I ventured the hypothesis [1904] that
enzyme action spreads like radiation from an enzyme particle as centre, and
that an enzyme particle contains the same molecule, which is liberated or
acted upon by this enzyme, in some active state.

What kind of radiation this may be cannot yet be decided. It is however
very remarkable, that an imitation of specific enzyme actions has been
produced by Rosenthaler [1908] by electromagnetic oscillations. By placing

1 This specificity is known to enzyme workers to be probably not so absolute as is often taken
for granted. According to the following theory a limited specificity must also be looked for.

|

2 OA hae a Gs

H. P. BARENDRECHT 553

his substrate solutions inside a solenoid, through which a frequently inter-
rupted direct current or an alternating current was passed, a diastase-like
action on starch was observed for a frequency of 440-480 per second, and
a hydrolysis of proteins for one of 320-360.

In order to explain the principal facts it is necessary to make the natural
assumption, that this radiation may be absorbed by the substrate itself, by
the substances liberated from it by the enzyme action, or by any other
foreign substance added to the solution. The possible absorption by the
water itself must be practically the same in every case and can therefore be
left out of account.

It is evident that the sphere, within which this radiation-like effect
around an enzyme particle may be expected to be active, must be very small ;
otherwise enzyme actions would be instantaneous.

In the beginning the spheres of action in differently, but sufficiently
‘concentrated solutions of substrate will be inversely proportional to that
concentration ; therefore in the first time-unit the same quantity of enzyme
will produce the same amount of effect in these solutions, for practically all
radiation is absorbed by the substrate only. *

In more dilute solutions not all the radiation will reach a substrate
particle before it is too much weakened by spreading. This explains the
fact that in moderately dilute solutions the constancy of the amount of
initial action has disappeared. In sufficiently dilute solutions of substrate
we shall practically obtain the condition, that on the average no two particles
of substrate are on the same radius of the radiation sphere, so that all these
particles will be acted upon at the same time and the amount of action will
be directly proportional to the concentration of substrate ; a fact observed by
myself [1904] as well as by many other investigators.

The advantage of a reasonable working hypothesis proves here again to
be its giving an indication, where to look for regularities and where these are
not to be expected.

This hypothesis can also be quantitatively developed and leads, as I have
shown, to a theoretical formula for the reaction-velocity in sufficiently
concentrated solutions of substrate. We shall call the initial concentrations
of the substrate a, the same after a time ¢ of reaction x (all expressed for
instance in grammes per 100 c.c.), and the amount of action in the first
time-unit by a given quantity of enzyme in 100 cc. of these sufticiently
concentrated solutions m. According to theory as well as to experiment m is
independent of a.

If further, n is the ratio between the absorption power for the enzyme

554 H. P. BARENDRECHT

radiation of the reaction-products and that of the original substrate, the
general differential equation for the rate of reaction is:
—dxz= He ea dt.

As however catalysis by enzymes is generally recognised to be as a matter
of fact a balanced action, this formula for the rate of reaction can only be
applicable, where there is evidence that the reverse action, the synthesis,
does not come into play. It will be shown further on, that this happens
to be the case as.a rule in the experiments with two of the best studied and
relatively simple enzymes, invertase and lactase.

Bearing in mind this restriction, we obtain by substitution of <—* =¥

and integration, the general formula for the rate of enzyme action in
sufficiently concentrated solutions of substrate :

log 7 + +=" Q-43dy = ™ 0-434
expressed in decimal logarithms.

The two constants in these experiments can be determined in different
ways.

They may be calculated from the observed action at two different time-
intervals, or m can be directly found by estimating the change during the
first time-unit and n from the same estimation in similar conditions, but
in a solution, in which had also been dissolved a known amount b per 100 c.c.
of the reaction-products.

In the first estimation the initial velocity will be

dy _™m
dt y=0 Ee a ?
in the second, from

dx ag

ian he L+n(a—@) +nb dt,
dy Thee. A
we get (3) Se Se
it /y=
ate Din Te lin 2

a

Evidently the value of n will depend only on the nature of enzyme and
substrate, and must therefore be the same in the experiments of different
observers.

For invertase from ordinary yeast the estimations of all the investigators,
who avoided the errors mentioned in the first part of this article, agree with
the value n =4, and I found the same also by direct estimation of initial
velocities. ’

H. P. BARENDRECHT 555

As (a—«) grammes saccharose give by inversion }4$ grammes invert-
sugar, the rate of inversion of cane-sugar in sufficiently concentrated solutions
by unaltered invertase is thus generally represented by the equation :

log; + 0393y = Kt.

The enzyme lactase is not as readily accessible as invertase, therefore its
action is not so often studied. Ordinary brewers- or distillers-yeast is inactive
in regard to milk-sugar. It is necessary to employ the special yeast, which
is known to be the principal active agent of “kephir.”

In order to extract the enzyme lactase as fresh and as unaltered as
possible from the living cell, which produces it, I prepared a sufficient
quantity of a pure culture of “Saccharomyces kephir” and worked only with
the fresh extract from this yeast, which had been mixed with kieselguhr and
dried in vacuum at a low temperature [1906].

In estimating the velocity of inversion of an 8°/, milk-sugar solution at
30° by this lactase, I found a series of figures, to which it will now be shown
that the general equation:

log + 41=" 0-434y =™ 0-434¢
=i! | n na
can also be applied with the same success.

Only, as explained above, the value of n may be expected to be different
from that for invertase.

Direct observations of the very large retarding influence of glucose and
of galactose showed that n is here much greater than in the case of
invertase.

Calculation from two estimations of the hydrolysis of lactose at different
time-intervals gives about n=6. Thus for lactose the velocity of inversion
is represented by

log ~~, — 0362y = ht.

For 8°/, milk-sugar at 30° I found, estimating the hydrolysis again by
the gravimetric Fehling-Kjeldahl method:

1 1 ;

t (minutes) y

5 0°23 —(- 0-0061
10 0°38 0-0070
15 0:47 0-0070
20 0°57 0-0080
25 0°61 0:0075
40 0°72 0-0073
66 0-82 0-0068

90 0-90 0:0075

556 H. P. BARENDRECHT

Armstrong [1904], though also emphasising the importance of working
rapidly and using only fresh extracts, did not it seems give care enough
to these precautions. Kephir-grains, an extract of which he used, are a
mixture of kephir-yeast cells with lactic acid bacteria and other residues,
all together a mixture of somewhat unknown history and origin. The
extract will contain some acid and in fact was not very active, compared
with that obtainable from fresh pure lactose yeast. The experiments lasted
several hours and even days, so that it was necessary to add toluene. The
risk of decaying of the enzyme during these long experiments at 35°-37°
was therefore not sufficiently excluded.

Still the first series of his angers (p. 506) calculated with our
lactose equation, gives

il 1
2 hams i ‘es log Tae p= 7 (oe p= EF -0: 302y )
1 0:221 0-1085 0-0285
2 0-312 0:0812 0:0247
3 0°384 0:0713 0:0244
4 0°458 0:0665 0:0250
5 0-515 0-0629 0-0256
6 0-566 ~ 0-0604 0:0263
10 0-69 0-0509 0:0259
LG 0-842 0:0471 0:0290
23 0°924 0:0461 0:033
29 0-953 0:0457 0-034
38 0:98 0:0447 0-035

The last of these estimations are less trustworthy, owing to exposure of
enzyme and of milk-sugar to the combined action of the slight acidity and
the temperature of 35°.

The other series of Armstrong’s article lasted still longer, the inversions
were far from finished at the end, and proceeded only very slowly during the
last 24 hours. Evidently the enzyme was here to a large extent destroyed.

There are further regularities in the working of these relatively simple
enzymes, invertase and lactase, which only become apparent when the
investigator takes sufficient care to work only with enzymes, extracted with
caution directly from the fresh living cells.

As I have shown before, the experiments not only agree with the
conclusion, drawn from our hypothesis, that generally all neutral substances
as well as the inversion products must retard the enzyme action, but striking
quantitative regularities also exist.

For unaltered! invertase from ordinary yeast the facts are these :

1 By drying the yeast at a high temperature or treatment with alcohol I got an invertase, the
action of which was more retarded by levulose than by glucose.

H. P. BARENDRECHT 557

If we represent the retarding power (or the absorption coefficient ac-
cording to our conception) of glucose by n=4, then the retarding powers
of levulose and of invert-sugar are found to be exactly n=4 also. For
the retarding powers of galactose and mannose however the experiment
gives the double value n=1. Nearly all neutral substances give retarda-
tion of the same order as the sugars, e.g. dulcitol, mannitol, urea, salicin and
other glucosides.

It is very remarkable, that milk-sugar alone does not appreciably retard
the action of invertase. In connection with this may be mentioned the
small absorption power that milk-sugar was also shown to have for lactase
action, compared with that of all other sugars. ‘The figure for the inversion
couple from milk-sugar was found to be n=6. Galactose retards twice as
much as glucose, so for galactose the absorption power of lactose action is
n= 8, that of glucose n=4. For levulose, a hexose which is here not con-
tained in the substrate, experiment gives also the sum of the absorption
powers of the inversion products of milk-sugar, that is n = 12.

In connection with the recent work of Tanret [1905] and of Hudson
[1908] I wish to draw attention here to another point about the action of
invertase and maltase.

If enzymes, as seems now more and more probable, are able to decompose
a substance as well as to actuate the reverse process, then invertase, maltase
and lactase should produce an equilibrium between bihexose and hexoses.
In the case of maltase Croft Hill [1898] seemed to have established some
synthesis of maltose from glucose, but it was proved afterwards, that the
bihexose formed with difficulty was, at least for the most part, another
substance. As for invertase and lactase, the evidence of synthesis of cane-
sugar and milk-sugar is even slighter.

The facts, that a fairly concentrated solution of cane-sugar is practically
completely hydrolysable by means of invertase and that after all no appreci-
able enzymatic synthesis of maltose from glucose is observed, suggest the
conclusion, that the hexoses, split off by these enzymes, undergo a secondary
change, by which the reverse enzyme action is prohibited.

Through the investigations of the mutarotation of sugars, especially by
Tanret and by Hudson, much clearness has been brought into our knowledge
of the isomerism of the most important members of this group. It was proved
by Tanret, that glucose, dissolved in water is at first totally in the a-form
({a]o= 110°) and then gradually changes into a mixture in equilibrium
((a]p = 52°5°) of about two-thirds a-glucose and one-third §-glucose
({a]p = 19°).

558 H. P. BARENDRECHT

As two-thirds of glucose in solution are finally in this a-modification,
then, if the glucose, split off by enzyme action from cane-sugar or maltose
should be originally in this same form, as is generally assumed, it is difficult
to see why the reverse action of the enzyme should not produce a considerable
amount of maltose.

Also the hydrolysis of cane-sugar by invertase should then not be
expected to come to a finish. |

The rate of mutarotation of this a-glucose, i.e. the velocity with which
it comes to this equilibrium, was carefully measured by Hudson, who made
a similar investigation of the mutarotation of levulose. Combining these
investigations with his experimentally quite correct estimation of the
velocity of inversion of cane-sugar by invertase, Hudson came to the con-
clusion, that it is really a-glucose, which is originally formed by enzyme
action in this case.

As the mean value of 18 experiments he found however for the specific
rotation of this “fresh” glucose 120°7°, while this figure for a-glucose
is 110°. This difference is so considerable, that Hudson’s work also seems
to point to the non-identity of enzyme-made fresh glucose with a-glucose.

This result is in complete agreement with the experimental evidence
I got in 1904 of an unstable form of glucose, which seems to be the primary
product of enzyme action and which gradually is converted to ordinary
glucose in solution, that is, as we know now, to a mixture of a-glucose and
§-glucose, according to the new nomenclature of Tanret.

It was expected, if the inversion of maltose was accomplished by a very
active maltase solution’ rapidly enough for the velocity of inversion to be
great, compared with the velocity of conversion of the freshly formed
glucose, that after the real equilibrium point maltosez@glucose had been
reached, the rate of inversion would appear to be approximately constant.

The curve, which must result from these two simultaneous reactions, may
be demonstrated with the aid of Fig. 1.

After the few minutes that are needed for the principal reaction to
come to the real equilibrium point P, the relatively slow secondary change
of unstable to stable glucose has practically not yet started. Being a mono-
molecular reaction, this rate of change of glucose will be at any time
proportional to the concentration of unchanged unstable glucose.

As soon however as the concentration of unstable glucose has been

1 In my article of 1904 I omitted to mention, that these very active maltase extracts of the
dried yeast-kieselguhr were obtained by using water with a small amount of sodium hydroxide.
Extraction with water only gave a maltase solution of smaller activity.

eaten «i

H. P. BARENDRECHT 559

diminished by this change, the equilibrium between maltose and glucose
is disturbed and part of the maltose is split again. As we have arranged
the conditions so that this inversion of maltose is much quicker than the
transformation of glucose, the concentration of unstable glucose and
therefore the rate of its change will be maintained constant (in first
approximation).

Y=4

Gene

Fig. 1.

Thus in Fig. 1 we can represent, also in first approximation, the forma-
tion of stable glucose by the dotted straight line PQ, and the amount of
unstable glucose by the horizontal dotted line PR. At any time, e.g. after
A minutes, the equilibrium between maltose and unstable glucose will be
immediately restored by the quick enzyme action. As however the con-
centration of maltose at the time A is somewhat smaller than that at
the time S, the concentration of unstable glucose, required to balance
by synthesis the hydrolysis of maltose, is also somewhat smaller than
SP or AB. .

The unstable glucose which has disappeared therefore need not be
totally replaced by newly inverted maltose and the result will be, that at
the time A this inversion will not have proceeded exactly to D, but to C.
The estimations of the inversion of maltose will thus give a curve PM,
slightly bending downwards from the straight line PQ.

Fig. 2, which illustrates the type of result obtained in my experiments
of 1904, may serve also here to demonstrate the concordance with the
theoretical type of Fig. 1. i

560 H. P. BARENDRECHT

0-4

0-3

0-2

0-1

10 20 30 40 50. 60 70 80 90
minutes

Fig. 2. A, 5°, maltose. B, 10 °/) maltose.

All the estimations of the hydrolysis of maltose were executed with
Fehling’s solution by the gravimetric method of Kjeldahl [1895] for mixtures
of sugars.

These experiments with different concentrations of maltose showed
further, that the positions of the equilibrium points, Le. the points of
flexure, where the curves become approximately straight lines, were con-
nected together as these figures are in a balanced action between a
monomolecular and a bimolecular reaction.

Following up the conception, that an enzyme particle extends its catalytic
action in a sphere as regards both the reverse action and the hydrolysis of
maltose, and that the size of this sphere in sufficiently concentrated solutions
must be inversely proportional to the sum of the concentrations of maltose
and glucose (the concentration of glucose multiplied by the absorption
coefficient n), the complete formula for the velocity of action of maltase
on maltose in conditions, where the secondary change of glucose is not yet
perceptible, is

—dz=m ee was a dt.

x+n(a-2)

E . . , . ae ud . . “zy. .
After substituting again —~ = y, the equation, relating the equilibrium
points (where — 2 =()) with the initial maltose concentration, 1s:
1 = — gay? = 0 2... ee -nen- oer eee eee (£).

For a solution of 10 grammes maltose in 100 ce. for instance, the
experiment gave about y=0'15, from which followed q = 4.

The other experiments with concentrations of 7, 5 (see Fig. 2), 3 and 1 g.
for 100 c.c. agreed then with the figures, calculated from equation (Z).

H. P. BARENDRECHT 561

a y
10 0°15
7 0-17
5 0°20
3 0°25
1 0°39

The synthesis of bihexoses in the living cell and the indications of
synthesis in vitro, however slight they may be, suggest the conclusion, that
the transformation of enzyme-made glucose to stable glucose is in fact also
a balanced action, in which the equilibrium is far on the side of stable
glucose.

In enzyme actions like this one, where the real equilibrium is so soon
attained and the reverse action is thus vigorously hindering the hydrolysis,
it is clear, that addition beforehand of the product of the reaction may
retard more than that of any other compound of the same kind. The
conversion of the unstable enzyme-made glucose is retarded considerably
by the added glucose and with that the reverse action can assert itself.

For maltose I found a much larger retardation by glucose than by
galactose and levulose.

Hence there is in fact no discrepancy between the peculiar retardation
effects of foreign substances on the action of invertase and lactase on one
side and the general rule, applied with success by ter Meulen [1905], on
the other, that the enzymatic hydrolysis of a glucoside is most retarded
by the same sugar, which is split off from the glucoside. The diagrams,
representing the rates of hydrolysis of many of these glucosides, have the
same character as that of maltose-inversion. Here also the reverse action
would soon stop the further splitting of the glucoside, if the sugar, which
is formed, were not liable to a secondary change, which change is effectively
retarded by addition of a large amount of the stable form of the same

sugar.
REFERENCES.

Armstrong (1904), Proc. Roy. Soc. 73, 500.
Barendrecht (1904), Zeitsch. physikal. Chem. 49, 456.
(1906), Zeitsch. physikal. Chem. 54, 367.

Brown (1902), J. Chem. Soc. 81, 373.

Croft Hill (1898), J. Chem. Soc. 73, 634.

Duelaux (1898), Ann. Inst. Past. 12, 96.

Henri (1901), Compt. rend, 133, 891.

Hudson (1908), J. Amer. Chem. Soc. 30, 1564.
Kjeldahl (1895), Compt. rend. Carlsberg, 4, 329.
O’Sullivan and Tompson (1890), J. Chem. Soc. 57, 865.
Rosenthaler (1908), Sitzungsber. K. Preuss. Akad. Wiss. 20.
Sérensen (1909), Compt. rend. Carlsberg, 8, 120.
Tanret (1905), Bull. Soc. Chim. (3), 33, 337.

ter Meulen (1905), Rec. Trav. Chim. 24, 444.

LII. AN INVESTIGATION INTO THE PHYSICO-
CHEMICAL MECHANISM OF HAEMOLYSIS
BY SPECIFIC HAEMOLYSINS. (Preliminary
Communication.)

By UPENDRA NATH BRAHMACHARL
From the Campbell Medical School, Calcutta.
(Received October 2nd, 1913.)

The mechanism of haemolysis by specific haemolysins has not been much
investigated from the physico-chemical standpoint. It is supposed that
the haemolysin affects the permeability of the envelope. Baumgarten has
noticed that the first stage of haemolysis, as produced by a biological poison,
is the swelling of the corpuscles, just as occurs when the serum is made
hypotonic. Bang [1910] considers that haemolysis by means of Cobra
poison is due to change in the composition of the lipoid membrane of the
erythrocyte rendering it more permeable to salts. Through this membrane
the extracellular salts and water soak gradually, until in the end the corpuscle
ruptures. The behaviour of the nucleated erythrocytes of amphibians towards
specific haemolysins tends to shew that their permeability is alone affected
during the process [Landau 1903].

So far as I am aware, no work has been done on the behaviour of
erythrocytes loaded with amboceptor towards distilled water before the
complement has been allowed to act upon them. The present paper gives
the results of investigations on the determination of the resisting power of
amboceptor-loaded human erythrocytes to haemolysis when under the influence
of distilled water. The haemolytic antisera were obtained from fowls and
rabbits that had been immunised by means of injections of washed human
corpuscles.

It was originally expected that amboceptor-loaded corpuscles would be
less resistant than normal ones in their behaviour towards distilled water.
Contrary to what was expected, I noticed a remarkable increase of the
resisting power of the erythrocytes.

The method adopted for determining the resisting power of the erythro-
cytes is that previously described by me [1909, 1911] for the determination

———

————

U. N. BRAHMACHARI 563

of the specific resistance of erythrocytes to haemolysis. A series of experiments
was performed and these are described as follows :—

In the first series, the erythrocytes were suspended in 0°85°/, NaCl and
allowed to haemolyse with two parts of distilled water. The mixture was
allowed to stand for ten minutes and the dissolved haemoglobin was estimated
in the manner previously described. This gave the resistance of the normal
erythrocytes to haemolysis.

In the second series, the erythrocytes were mixed with two parts of
a 10°/, dilution of complement-free antihuman fowl’s haemolytic serum with
0°85 °/, NaCl and kept in the incubator for one to three hours and subsequently
treated in the same way as above with distilled water after having been
thoroughly washed in 0°85°/, NaCl, and the resistance of the amboceptor-
loaded erythrocytes determined in the same way as above.

In the third series, the erythrocytes were treated with complement-free
antihuman rabbit's haemolytic serum and their resistance to haemolysis
determined.

As a control test, a few experiments were made to determine the
resistance of human erythrocytes after having been treated with normal
fowl’s serum in the same way as above.

In this way the resistance of the erythrocytes under different conditions
was determined and the following tables worked out. The resistance of the
erythrocytes is expressed in terms of the relative haemoglobin-value of
the resistant erythrocytes which is the ratio between the amount of haemo-
globin in the resistant corpuscles and the amount in the total suspension of
erythrocytes used.

Tables shewing the resistance of erythrocytes to haemolysis under
different conditions :— .

TABLE IL
Resistance Resistance of
of normal erythrocytes loaded
erythrocytes with amboceptor
to haemolysis obtained from fowl
1 0°345 0-500
2 0°379 0-456
3 0-242 0°407
4 0°333 0-480
5 0°346 0452
6 0-269 0°531
7 0-333 0-466
8 0°350 0444
9 0°301 0°421
10 0-288 0°459

ul 0-288 0:368

564 U. N. BRAHMACHARI

TABLE IL.
Resistance Resistance of
of normal erythrocytes loaded
erythrocytes with amboceptor
to haemolysis obtained from rabbit
1 0°195 0-600
2 0-188 0:378
3 0:296 0:592
+ 0-240 0-480
5 0°132 0-220
TABLE AM.
Resistance Resistance of
of normal erythrocytes treated
erythrocytes with normal
to haemolysis fowl’s serum
1 0-288 - 0243
2 0-288 0:298
3 0-282 0-259
4 0-240 0-282

It will be seen from the above tables that in every case the addition of
the amboceptor increased the specific resistance of erythrocytes, while the
addition of normal fowl’s serum did not bring about any difference in their
resisting power.

The problem that we are now to solve is how it is that the erythrocytes
become more resistant when they become fixed to the amboceptor but lose
all their resisting power as soon as the complement acts and what is the
biological significance of this remarkable phenomenon. An explanation may
be offered, which however is hypothetical. In the first process of the action
of haemolysin during which the amboceptor combines with the erythrocytes,
we may assume that this combination is purely of the nature of adsorption
and not true chemical combination. In this process the dimensions of the
pores between the molecules of the outer wall of the erythrocytes become
small due to the amboceptor molecules filling the pores of the original
erythrocytes. As diffusion depends, specially, upon the dimensions of the
pores of the membrane through which it takes place, less haemoglobin passes
out through the red corpuscles loaded with amboceptor than the unloaded
ones when they are treated with distilled water. In the case of the
colloidal complex of amboceptor and erythrocyte molecules, the bodies
brought into close contact with each other do not react with one another
in the chemical sense. Chemical reaction which, according to Bayliss
[1912], is the third and last stage of the heterogeneous reactions of colloidal
complexes, takes place between the molecules of the amboceptor and of the

U. N. BRAHMACHARI 565

erythrocytes only through the agency of the complement, which probably
acts more or less like a ferment. When this takes place, there is a con-
densation of the molecules of the colloidal complex in each erythrocyte and
as a result of this, the dimensions of the pores of the membranes of the red
corpuscles increase to such an extent that haemoglobin diffuses out of the
corpuscles and haemolysis results.

This theory will easily explain the inhibitory influence exerted by
hypertonic saline solutions on the action of specific haemolysins as has
been observed by Sutherland and McCay [1911] and others. Such saline
solutions would tend to reduce the size of the erythrocytes by exosmosis
of water from them and as a result of this the molecules of the erythrocytes
come closer to each other and haemolysis is prevented. In other words the
widening of the pores of the erythrocytes brought about by the action of the
complement on the amboceptor-loaded erythrocytes is counteracted by the
presence of the hypertonic saline.

To test the accuracy of the theory the following experiments were
performed :—

A suspension of sheep’s corpuscles in normal saline was mixed with
amboceptor and complement and kept in the incubator. The process of
haemolysis was stopped in from five to ten minutes before it was complete,
the corpuscles quickly centrifuged and the supernatant fluid replaced by
N/2 NaCl, after thoroughly washing the corpuscles several times with the
same. A portion of the suspension of the corpuscles in N/2 NaCl was again
centrifuged and the N/2 solution was pipetted off and in its place an excess
of 0:85°/, NaCl solution was substituted—a marked haemolysis resulted.
Thus we have the following results :—

(1) Amboceptor-loaded corpuscles partially acted upon by complement,
thoroughly washed in N/2 NaCl—no haemolysis.

(2) Amboceptor-loaded corpuscles partially acted upon by comple-
ment, thoroughly washed in N/2 NaCl and then the N/2 NaCl replaced by
0°85 °/, NaCl—distinct haemolysis.

In other words amboceptor-loaded corpuscies which have been partially
acted upon by the complement are haemolysed when brought in contact
with normal saline. The dimensions of the pores of such corpuscles when
suspended in N/2 NaCl are much less than when suspended in normal
saline, i.e. there is marked widening of these pores when they are suspended
in normal saline and haemoglobin diffuses out. Evidently, therefore, the
pores of the amboceptor-loaded corpuscles that have been acted upon partially
by the complement are much wider than those of the normal ones. The

Bioch, vu 37

566 U. N. BRAHMACHARI

complement must have therefore brought about a condensation of the
molecules of the amboceptor-loaded corpuscles, as explained before, thereby
leading to widening of their pores to such an extent as to allow the haemo-
globin molecules to pass out of them, even when they are suspended in
normal saline.

An attempt was made to determine the volume of the corpuscles before
and after combination with amboceptor. After rejecting different methods,
the following method was devised for determining the size of the corpuscles.
Human and sheep’s corpuscles were used for the purpose.

The corpuscles were first suspended in 0°85°/, NaCl and the suspension
was sucked into two large-bore haematocrit tubes of exactly the same
diameter and the corpuscles in both were centrifugalised at a high speed
for about ten minutes, after which their volume was noted. The clear
supernatant fluid from one of the tubes was now pipetted off and a
dilution of haemolytic amboceptor in 0°85°/, NaCl (one part of amboceptor
and nine parts of 0°85°/, NaCl) introduced in its place. The amboceptor
was then thoroughly mixed with the corpuscles and the mixture allowed
to stand for ten minutes. The other tube which was used as a control was
then shaken up and the corpuscles again mixed with the supernatant fluid.
The two tubes were again centrifugalised for ten minutes and the volume
of the corpuscles noted. If the volume of the corpuscles in the control tube
was found to be exactly equal to what was noted in it in the first experiment,
then there was no error due to any change in the speed of the centrifuging
tube. The volume of the amboceptor-loaded corpuscles was then noted.
If the volume’ of the corpuscles in the control tube was different in the
second experiment from what was noted in the first, then the volume of
the amboceptor-loaded corpuscles was accordingly changed. In this way,
absolutely accurate results were obtained.

Thus, if V be the volume of the corpuscles in the control tube after the
first centrifuging and V’ after the second, then V— VV’ is the difference due
to changes in the speed of the centrifuging machine.

If now v be the volume of the corpuscles before being loaded with
amboceptor, and v’ after being loaded with amboceptor, it is evident that
y—v is not actually the change in the volume of the corpuscles. The
exact change is, evidently,

(v—v’)+ ark V—V’").

In this way, the exact change in the volume of the corpuscles due to
combination with amboceptor was found and the following results obtained.

U. N. BRAHMACHARI 567

Human corpuscles.

Volume before action Volume after action
of amboceptor of amboceptor
(1) 2-75 cm, (1) 2°75 cm.
(2) 2°75 (2) 2°75
(8) 2°50 (8) 2°50
(4) 1:50 (4) 1:50
(5) 1:45 (5) 1:50

Sheep's corpuscles.

Volume before action Volume after action
of amboceptor of amboceptor
(1) 3°30 cm. (1) 3°30 cm.
(2) 5:38 (2) 5:20
(3) 4°50 (3) 4:20
(4) 6°23 (4) 6°22

It will be seen from the above tables that there was no reduction in the
volume of the corpuscles after the action of the amboceptor and therefore
the increase in their resistance to haemolysis was not due to any diminution
in their volume after the action of the amboceptor.

I am deeply indebted to Lt.-Col. Sutherland, I.M.S., of the Calcutta
Medical College, for kindly providing me with the haemolytic antisera and
giving me every facility to work in his laboratory.

REFERENCES.

Bang (1910), Biochem. Zeitsch. 23, 463.

Bayliss (1912), Proc. Roy. Soc. B, 84, 81.

Brahmachari (1909), Biochem. J. 4, 59, 280.

(1911), Biochem. J. 6, 291. (See also Studies in Haemolysis, 2nd Ed.)
Landau (1903), Ann. Inst. Past. 17, 52.

Sutherland and McCay (1911), Biochem. J. 5, 1.

37—2

LIII. NOTES ON SOME FURTHER EXPERI-
MENTS ON THE CLOTTING OF CASEINOGEN

SOLUTIONS.
By SAMUEL BARNETT SCHRYVER.
From the Research Institute of the Cancer Hospital.

(Received October 10th, 1913.)

Some months ago, I communicated to the Royal Society the results of
a series of experiments on the clotting of casemogen solutions [1913]. It
was there shown that caseinogen, when freshly precipitated from diluted
skimmed-milk under certain specified precautions, is an unstable substance
which on treatment with water at 37°, or with 0:1°/, acetic acid at room
temperature, is readily converted into another product or series of products
(metacaseinogens), which differ from the original substance in that they are
very much less soluble in lime water, and consist therefore apparently of
smaller aggregates of lower molecular weight. Whena half-saturated solution
of lime water is saturated with the natural caseinogen by rotating the mixture
for several hours in a thermostat at 20°, an acid milky solution is obtained
after centrifugalisation and filtration, which readily clots without the addition
of a soluble calcium salt in the presence of rennin; it will also clot in the
absence of rennin when gently warmed (to 25°) in the presence of soluble
calcium salts when the latter exist in the mixture within certain definite
limits of concentration. In the presence of milk-serum, peptones, amino-acids
ete., a clot can be produced from the above-mentioned calcium caseinogenate
solution only in the presence of both soluble calcium salts and rennin.

In the paper quoted it was shown, furthermore, that the casein of the clot
produced from calcium caseinogenate by rennin alone is about half as soluble
in lime water as the natural caseinogen, and appears to possess, therefore,
approximately half the molecular weight, and corresponds in properties with
the metacaseinogens produced from natural caseinogen by the action of water.

Similar conclusions as to the molecular weight of casein, and its
relations to caseinogen have been arrived at recently by Lucius van Slyke and
A. W. Bosworth [1913], who have attacked the problem from a quite different

S. B. SCHRYVER 569

standpoint, and by other experimental methods, and in a still more recent
paper by Bosworth [1913] a method is described for the production of a clot
from caseinogen solutions by rennin alone, which does not differ materially
from the method described by me’.

Bosworth also came to the same conclusion as I did, that the production
of casein from caseinogen is not an ordinary proteoclastic reaction, although
it may be the result of a hydrolytic cleavage. I have ventured to suggest
that the two substances bear a relationship to one another similar to that of
metaphosphates to pyrophosphates. Bosworth and I are also in accord
with reference to the identity, or at any rate the approximate identity, of the
empirical chemical compositions of caseinogen and casein.

Accurate stcechiometrical relationships between casein and caseinogen
were not to be attained by the method of experiment adopted by me,
owing to the existence of an adsorption equilibrium when lime water is
treated with excess of various caseinogen and casein preparations, a point
which has been the subject of numerous experiments, which are briefly
mentioned in my previous paper [1913, p. 466].

The casein prepared from the clot differs in one important respect from
metacaseinogen produced by the action of water at 37° or by 0°1°/, acetic
acid at room temperature from caseinogen, for whereas the latter, on solution
in excess of caustic alkali and reprecipitation from this solution with acetic
acid under certain defined conditions is reconverted into a caseinogen of
relatively high solubility in lime water, which readily yields clots on treatment
either with calcium chloride or rennin, casein obtained by means of rennin
cannot be converted into caseinogen under these conditions; on dissolving in
alkali and reprecipitation, a product of only about half the solubility of
caseinogen in half-saturated lime water is obtained. The solution in lime
water furthermore does not clot on addition of rennin, and yields, on mixture
with solutions of calcium salts within certain limits of concentration, an
immediate precipitate. This precipitate is readily distinguishable from the
clot produced under similar circumstances from caseinogenate solutions,
which only forms when the temperature is raised above that of the tempera-
ture of solution.

Further experiments have been directed towards ascertaining the reason
of this difference between the casein of the clot and the metacaseinogen

1 Throughout my papers, I have consistently referred to the preparations which have not been
clotted as caseinogen, whereas the preparations obtained from the clot are described as casein.
In this respect, I have kept to the English method of nomenclature. Van Slyke and Bosworth
adopt the method of nomenclature usual on the continent, describing the products as casein and

paracasein, respectively.

570 S. B. SCHRYVER

produced by the action of water alone on caseinogen. These experiments are
not yet complete, but I have been tempted to communicate them in their
present form, owing to the fact that they have been subjected to a temporary
interruption due to my change of laboratories.

CLOTTING OF CASEINOGEN SOLUTIONS BY MEANS OF PANCREATIN.

Investigations have been carried out with the object of comparing the
process of clot production by means of other ferments. The results obtained
with pancreatin indicate that its action can be readily distinguished from
that of pepsin (or rennin). When half-saturated lime water is saturated with
natural caseinogen, the solution thus obtained in the manner described
[Schryver, 1913, p. 462] readily clots on the addition of small amounts of
pepsin solution (1°/,). No soluble calcium salt need be added. A clot under
similar conditions cannot be produced by a pancreatin solution, although
the latter will clot milk. A soluble calcium salt appears, therefore, to be
necessary when pancreatin is employed for clot-production [compare
Mellanby, 1912]. A clot was also prepared from an artificial milk by
pancreatin in the following way: 5 c.c. of calcium caseinogenate solution
were mixed with an equal bulk of milk-serum containing calcium chloride in
the concentration of N/25. Under these conditions, as has already been shown,
no clot is produced. If to this mixture are added a few drops of 1°/)
pancreatin solution, and the whole is incubated, a clot is produced similar
in appearance to that produced by pepsin under like conditions.

The clot produced from artificial milk has the same properties as that
produced from the natural milk, but differs in certain important respects from
the clot produced by pepsin (or rennin) preparations.

The general method employed for examining the casein from the clots was
as follows. The serum was poured off from the clot, which was then thrown
on calico, and squeezed. It was then ground in a mortar with 0:1°/, acetic
acid, washed with water, dilute alcohol in increasingly graded strengths,
absolute alcohol and ether, and then air-dried. The fine powder thus produced
was then heated for 5 minutes with absolute alcohol to destroy the enzyme,
the hot alcohol was filtered off by means of a Buchner funnel and the powder
then washed with alcohol and ether, and again air-dried. The powder was
then transferred to a mortar, ground with water into a paste, and N sodium
hydroxide solution was added with continual grinding, until nearly all
dissolved, and the solution was alkaline to phenolphthalein. It was then
diluted and filtered through paper-pulp on to a Buchner funnel. Dilute

S. B. SCHRYVER 571

acetic acid was then carefully added with constant stirring until precipitation
was complete, the supernatant liquid was rapidly poured off, and the
precipitate washed by decantation or on a filter with ice-cold water, alcohol
in graded strengths up to absolute alcohol, and finally with ether, and it was
then air-dried.

Sometimes the above process was modified in various details, e.g. the clot
after washing with water was first dissolved in alkali and reprecipitated, and
the product after drying in the usual manner treated with hot alcohol. In all
cases, however, a similar product was obtained. When the clots produced by
pepsin and pancreatin were compared, the same procedure was always
employed. The clots produced by pepsin alone from calcium caseinogenate
solution, or by the combined action of the ferment and calcium chloride and
ferment in the presence of milk-serum, yielded caseins with the same
properties.

Now whereas the original caseinogen preparations had a solubility in
half-saturated lime water of about 353, and after solution in alkalis and
reprecipitation a solubility of about 30, the solubility of the casein prepara-
tion from the pepsin clot was about 14-17 and that of the pancreatin clot
about 6-8. Unlike metacaseinogen, therefore, caseins are not converted into
the more soluble products by solution in alkali and reprecipitation with acids,
even when all precautions for preventing subsequent change have been taken.
Furthermore, a saturated solution of caseinogen in half-saturated lime water
is opaque like milk, whereas that of pepsin-casein is opalescent, and that of
pancreatin-casein is generally water-clear. The two latter yield immediate
precipitates and not clots on treatment with calcium chloride and do not clot
on treatment with rennin. There appears, therefore, to be a difference in
the molecular weights of the caseins produced by pepsin and pancreatin, both
of which are considerably lower than that of caseinogen, that of pepsin-casein
being about one-half and that of pancreatin-casein about one-quarter of that
of caseinogen. Again, attention must be called to the fact that exact
steechiometrical relationships are not to be expected from the method of
experiment employed.

THE DIFFERENCES BETWEEN METACASEINOGENS AND CASEINS.

As already repeatedly stated, the metacaseinogens differ from the caseins
in that the latter are not reconverted into caseinogen on solution in alkali

1 Le. 5 c.c. of the solution after Kjeldahlisation required 35 c.c. N/10 sulphuric acid to
neutralise the ammonia produced (see former paper).

572 S. B. SCHRYVER

and reprecipitation with acids, whereas the former are. In spite of the large
number of experiments which have been carried out, I have not so far
succeeded in producing from a casein a solution which clots on treatment
with rennin.

Investigations have, therefore, been carried out with the object of
ascertaining the reason of the differences of behaviour between these
substances, which in other respects have similar properties.

Up to the present, no entirely satisfactory proof as to the actual nature of
the difference has been found, but certain facts have been recently discovered
which afford some indication. S

It seems not improbable that the ferment forms a combination with its
substrate, from which it cannot afterwards be separated, and that the
properties of the latter are then so altered that it can no longer form the big
aggregates of the character of caseinogen. |

Experiments were therefore carried out with the object of ascertaining
the properties of the substance produced, when ferments were allowed to act
on caseinogen at low and high temperatures, 1.e., under conditions under
which the proteoclastic action of the ferment is excluded.

Pepsin was allowed to act on calcium caseinogenate for 15 minutes at 0°,
the conditions of experiment being otherwise the same as those employed in
clot production. No clot formed under these conditions. The free caseinogen
was then precipitated with ice-cold dilute acetic acid and a little sodium
chloride?, and rapidly washed several times with ice-cold water, graded
strengths of alcohol, absolute alcohol and ether. These operations were
carried out as rapidly as possible. The powder after air drying was then
heated for 5 minutes with boiling alcohol, dissolved in excess of alkali, from
the solution in which the casein(ogen) was obtained in the usual way.
Although the product thus obtained possessed about the same solubility in
half-saturated lime water as ordinary caseinogen which had been dissolved in
alkali and reprecipitated, the solution obtained from caseinogen which had
been treated with the pepsin differed from that got from normal caseinogen
(treated in a similar way but without pepsin), in that the former did not clot
in the presence of rennin, and yielded an immediate precipitate, and not
a clot, when treated with calcium chloride solution of the requisite con-
centration.

The mere precipitation therefore, of caseinogen in the presence of a ferment
under conditions which exclude enzyme action, causes appreciable alterations
in its properties. When one considers the important part that surface

1 The precipitate does not readily separate in ice-cold solutions without this addition.

8S. B. SCHRYVER 573

actions play in such processes, this result is not altogether surprising. The
above result was obtained several times. A similar experiment was also
earried out in which pancreatin was used instead of pepsin, the conditions
being the same as those described above. The product in this case had
a solubility of only about 8 in half-saturated lime water, and was identical
in properties with the casein obtained from a pancreatin clot. It may
be remarked here that pancreatin produces the clot in the highly acid
solution.

A large number of experiments were also carried out with the object of
ascertaining the action of the heated ferment on clot formation. The results
obtained up till now are not, however, quite conclusive, and this is due to the
varying properties of the different pepsin preparations. Rohonyi [1913] has
recently succeeded in preparing definite compounds of proteins with proteo-
clastic ferments, and has shown that such compounds can be obtained when
ferment solutions are employed, the proteoclastic activity of which has been
destroyed by heat? If the conception is true that the caseins differ from
metacaseinogens in the fact that the former are combinations with the ferments,
then, in the light of Rohonyi’s recent work, it is conceivable that caseins
should be formed by the action of heated enzymes. In carrying out the
experiments devised for testing this theory precisely the same difficulties have
been encountered as those described by Rohonyi, due to the differences of
characters of the various enzyme preparations. This investigator has shown
that it is possible to prepare compounds of pepsin and proteins by the
employment of solutions of the ferment which have been heated for
15 minutes at 100° provided that no precipitate is formed during the heating.
Preparations having this property appear to be rare. I have myself not
succeeded in obtaining such. I have, however, obtained a preparation which
after heating to 80° for some time was still capable of actively clotting
calcium caseinogenate solutions. Prolonged heating of the ferment solution
in quartz tubes at this temperature produced a precipitate, and the solution
gradually became inactive.

It is very probable that experiments of this nature will only succeed when
a ferment preparation relatively free from extraneous proteins is obtained.
Further work in this direction is contemplated.

S. B. SCHRYVER

or
“NI
ne

THE NATURE OF “NATURAL CASEINOGEN.”

L. van Slyke and Bosworth have called attention to the fact that various
commercial preparations of caseinogen contain large quantities of acid calcium
salt, due apparently to the precipitation of the product with insufficient
quantities of acid. A criticism might therefore be directed against the
“natural caseinogen ” described by me, that such consisted of a mixture of free
acid and acid calcium salt. I therefore subjected my preparation to ex-
amination with a view of ascertaining its nature, and comparing it with
a calcium-free preparation prepared according to the method of van Slyke
and Bosworth. The latter was found to have a low solubility in half-saturated
calcium hydroxide, as it had stood some time in contact with hydrochloric
acid during the process of filtration. After solution in alkali, and _re-
precipitation with acid, with the precautions already repeatedly described, it
had a solubility of about 26, ie., about the solubility of my own caseinogen
preparations which have been “ purified ” by solution in alkali and reprecipita-
tion by acid}.

The solubilities of my “natural caseinogen” and of the caseinogen
preparations prepared according to van Slyke and Bosworth (both the
preparation directly precipitated with hydrochloric acid and that redissolved
in alkali and reprecipitated by acetic acid with precautions for preventing
change) in warm 10°/, sodium chloride were determined. Van Slyke and
Bosworth show that under these conditions the calcium salts readily dissolve,
whereas the free acids do not. The preparations were mixed with 10 times
their weight of the saline solution and maintained at 37° for about 1 hour
with continual shaking. The mixtures were afterwards centrifuged, and the
nitrogen in the clear filtered supernatant liquid was determined by Kjeldahl’s
process.

The saline extract from all preparations contained about the same small
amount of nitrogen, viz. 2 or 3 mgrms. in 5 ¢.c. Only minute quantities of
the preparations were soluble in saline, and my “natural caseinogen” is
therefore a free acid and not a calcium salt. Asa control, an acid calcium
salt was prepared by precipitation of sodium caseinogenate solution by
calcium chloride [Schryver, 1913, p. 469]. The precipitate was washed with
dilute alcohol till chlorine free, and obtained dry after alcohol and ether
washing in the usual way. On treating this with 10 times its weight of 10°/,
sodium chloride at 37°, it rapidly dissolved, leaving only a very small residue,
and yielding a thick syrupy solution.

1 Preparations thus ‘‘ purified” have never as high a solubility in lime water as ‘‘ natural
caseinogen.”

a ie 2 a RY a

S. B. SCHRYVER 575

CONCLUSIONS.

1. Casein produced from a pepsin (or rennin) clot differs from meta-
caseinogen, a product produced by the action of water at 37° on caseinogen,
in that it cannot be converted by solution in alkali and reprecipitation into
a more soluble product which dissolves in calcium hydroxide to yield clottable
solutions. The statement made in my earlier paper that casein produced
by pepsin is an aggregate of about half the size of that of caseinogen is
confirmed.

2. The action of pancreatin in clot-production differs from that of pepsin,
in that it produces clots only in the presence of soluble calcium salts. It
produces these clots, however, under conditions under which calcium salts
alone fail, e.g.,in the presence of milk-serum. The casein from pancreatin
clots also differs from that of pepsin clots. The former have only about half
the solubility in half-saturated lime water of the latter. Such lime water
solutions are usually water-clear.

3. Certain experiments are quoted, which indicate that the difference
between metacaseinogens and caseins is due to the fact that the latter are
combinations of the enzyme with the protem. The evidence produced both
by myself and independently by van Slyke and Bosworth and by Bosworth
alone, indicate that the action of the ferment is not an ordinary proteoclastic
action. It is conceivable that if the enzyme contains both haptophoric and
zymophoric groups, the former only take part in clot formation.

4. Evidence is brought to show that the “natural caseinogen,” the
preparation of which is described in my earlier paper, is not a calcium salt.

REFERENCES.

Bosworth (1913), J. Biol. Chem. 16, 231.

Mellanby (1912), J. Physiol. 45, 345.

Rohonyi (1913), Biochem. Zeitsch. 53, 179.

Schryver (1913), Proc. Roy. Soc. B. 86, 460.

Van Slyke, L. and Bosworth, A. W. (1913), J. Biol. Chem. 14, 203, 207, 211, 227 and 231.

LIV. .ON * THE . CHOLESTEROL (CONTENT) OF
THE TISSUES OF CATS UNDER VARIOUS
DIETETIC CONDITIONS AND DURING _IN-
ANITION.

By JOHN ADDYMAN GARDNER anp
PERCY EDWARD LANDER.

From the Physiological Laboratory, University of London, South Kensington.
(Recewed October 11th, 1913.)

Since the discovery of cholesterol by Conradi in 1775, and its analysis by
Chevreul in 1815, it has been found to be very widely distributed in the
animal, and in its isomeric forms in the vegetable, kingdom, and it is now
generally recognised as an integral constituent of all cells in the animal
body [Dorée, 1909]. The constant presence of such a substance clearly
suggests its importance from a vital standpoint, and necessitates its recognition
as a primary constituent of all protoplasm. Though a considerable amount
of work has been done on the subject, however, we have as yet little definite
knowledge of its physiological functions.

As a result of experiments on herbivorous animals published by one of
the present writers and various colleagues and the researches of Pribram
[1906], Kosumoto [1908, 1 and 2], Harley and Barratt [1903] the working
hypothesis was put forward that in these animals cholesterol is a substance
which is strictly conserved in the animal organism; that when the destruction
of the red blood corpuscles and possibly other cells takes place in the liver,
their cholesterol is excreted in the bile, and that the cholesterol of the bile
is reabsorbed in the intestine along with the bile salts and finds its way into
the blood stream to be used in cell anabolism [Dorée, Ellis, Fraser and
Gardner, 1908-1912]. In the case of carnivorous animals there would not
appear to be the same necessity for such economy, as cholesterol is a normal
constituent of their food, whereas the food of herbivora contains no cholesterol,
but instead the isomeric phytosterols, and these only in minute quantity.
The absorption of the food cholesterol would, however, be limited by the
reduction of cholesterol to coprosterol by the bacteria of the gut. In the

ee ee eee ee

J. A. GARDNER AND P. E. LANDER 577

human subject cholesterol is excreted in the faeces entirely in the form of
coprosterol under ordinary conditions [Bondzynski, 1896; Bondzynski and
v. Humnicki, 1896], though it has been shown by Miiller [1900] that
a prolonged milk diet leads eventually to the excretion of cholesterol as
such. In the case of dogs and cats the change into coprosterol is generally
only partial, though under some conditions it may be complete, and takes
place only on meat diets [Dorée and Gardner, 1908, 2]. Feeding experiments
[Dorée and Gardner, 1908, 2; Kosumoto, 1908, 2; Ellis and Gardner, 1909, 2]
on dogs and cats have shown that the total cholesterol found in the faeces,
either as cholesterol or coprosterol can be more than accounted for by the
cholesterol taken in with the food, provided that the animal remains in
health and constant in weight. This is also probably the case in man
[Ellis and Gardner, 1912, 3]. Feeding cats, however, on artificial diets free
from cholesterol led to inconclusive results from the point of view of
deciding whether in the case of carnivora the cholesterol of the bile is
normally reabsorbed along with the bile salts in the intestine [Ellis and
Gardner, 1909, 2]. Earlier experiments made by Dorée and Gardner [1909]
to trace a connexion between the percentage of cholesterol in the blood of
dogs and the cholesterol content of the food taken, also led to inconclusive
results, partly owing to the defects in the methods then available for
estimating cholesterol in the tissues. Recent improvements in the methods
of estimating cholesterol—notably the digitonine process of Windaus [1910]
—rendered it desirable to repeat and extend the experiments referred to.
In the present paper we give an account of analyses of blood and the tissues
of cats fed on diets containing different amounts of cholesterol and also of cats
kept in a state of imanition.

It is much more difficult in the case of carnivorous animals to maintain
constant conditions during experiment than in the case of herbivorous animals
such as the rabbit. In the latter animals a natural standard diet, free from
cholesterol and similar bodies, to which measured portions of cholesterol
could be added was available; and as these animals are practically continuous
feeders the bile flow and consequently the cholesterol content of the blood
due to this source would remain practically constant. In the case of cats
we had to make use of different foods, though the difference was not so
marked as in the earlier experiments. Little is known concerning the
influence of food on the secretion of the bile, but from experiments that
have been made by various observers, there is good reason to suppose that
the nature of the diet would not be without influence [vide Goodman, 1907].
Furthermore the cat and the dog are discontinuous feeders and the flow

578 J. A. GARDNER AND P. E. LANDER

of bile into the intestine would be intermittent. Under these circum-
stances the portion of the floating as distinguished from the constituent
cholesterol in the blood and the tissues due to the reabsorption of the
cholesterol of the bile, would not necessarily be strictly comparable in
different cases. What would be the limits of such variation, if any, we have
no data at present upon which to form an opinion, but a variation of the
kind suggested might wholly or partially mask any variation due to the
cholesterol absorbed from the food, which at best would not be great in
absolute magnitude. We did not therefore expect to obtain results comparable
in precision with those obtained in the case of herbivora, but we attempted
to render the conditions as comparable as possible by careful selection of
diets, and by always killing the animals during a period of active digestion.

DETAILS OF THE DIETS USED.

Diet A, cholesterol free. We had some difficulty in finding a cholesterol-
free diet which cats would take freely, but ultimately the followmg proved
suitable. The whites of four eggs were taken raw and mixed with an equal
weight of mashed boiled potatoes. The well-mixed material was flavoured
with a small quantity of Liebig’s extract of meat, and heated until the white
of egg coagulated. The animals appeared to relish this diet, not the
slightest trouble being experienced in getting them to eat it.

Diet B. Diet A, with the addition to each daily ration of a weighed
amount of free cholesterol.

Diet C. Lean cooked beef.

DETAILS OF THE ANIMALS USED.

Diet A, cholesterol free.

Cat No.1. The animal was fed for eight days and consumed 1003 grams
white of egg and 1008 g. potatoes, the total weight when cooked being 1932 g.
During this period its weight remained quite constant—2‘7 kilos. It was
killed one hour after the last meal. The stomach was full, and all the organs
appeared normal.

Cat No. 2. This cat ate the same amount of food as No. 1, and during
the eight days of the diet period its weight was as follows :—2°67, 2°58, 2°6,
2°55, 2°5, 2°5 and 2°55 kilos. It was killed two hours after the last meal.
The stomach was full, and all the organs seemed normal.

Cat No. 3. This cat was pregnant, and during 11 days consumed

J. A. GARDNER AND P. E. LANDER 579

728 g. of egg white, 728 g. of potatoes, weighing when cooked 1396 g.
It lost weight gradually during the diet period—8, 2°7, 2°7, 25, 2:4, 2°3
and 2°3 kilos. It was killed one hour after the last meal.

Cat No. 4. This animal ate in the course of 11 days 1138 g. egg white
and 1210 g. of potatoes, weighing when cooked 2188 g. It remained
approximately constant in weight—3-2, 3:2, 3°2, 3:1, 3:0, 3:1, 3:1, 3:1 kilos. It
was killed three hours after the last meal. The stomach was half full, and
the organs seemed normal.

Diet B.

Cat No. 5. During 11 days this cat consumed 693 g. of egg white and
693 g. of potatoes, weighing when cooked 1256 g., and 2°75 g. of cholesterol,
i.e. about twice the amount it would have got from the same weight of meat.
It weighed 2°3 kilos. at the beginning of the experiment and 2°2 kilos. at the
end. It was killed one hour after the last meal, and all its organs were
quite normal.

Cat No. 6. This cat consumed during 11 days 728 g. of egg white
and 728 g. of potatoes, weighing when cooked 1303 g., and 2°75 g. of
cholesterol. It weighed 2°4 kilos. at the beginning and gradually decreased
in weight to 2:1 kilos. It was killed 1°5 hour after the last meal. Its
organs seemed normal.

Diet C.

Cat No. 7. This animal was fed for ten days on lean cooked beef as
free as possible from fat and consumed 1563 g. of meat weighing in the raw
state 2165 g. This would correspond to an intake of about 17 g. of
cholesterol. The weights taken from time to time were 3°37, 3:3, 3°4, 3°33,
3°42, 3°45 and finally 3:43 kilos. It was killed four hours after the last meal.
The stomach was full and all the organs normal.

Cat No. 8. This cat consumed during 11 days 1910 g. of cooked
lean beef, weighing in raw state 2433 g. This would correspond to an
intake of about 1-9 g. of cholesterol during the period. The animal remained
constant in weight—3'3 kilos. It was killed four hours after the last meal.
The stomach was full and the organs normal.

ANIMALS IN A STATE OF INANITION.

In these experiments well-fed animals were kept without food for seven
days, but were allowed as much water as they wished. At the end of this
period they all appeared to be in good health. No faeces were passed.

580 J. A. GARDNER AND P. BE. LANDER

Cat No. 9. The weights taken daily, except on Sunday, were 3:1, 2°9,
2-8, 2°6, 2:55 and 2°5 kilos. The loss of weight was therefore 19°4°/,. After
the animal was killed the stomach was found to be completely empty.
There was a little fluid in the small intestine, and some faecal matter in
the rectum.

Cat No. 10. The weights recorded were 3:1, 2:95, 2:8, 2°75, 2°7,26 and
25 kilos., the loss in weight being as before 19°4°/,. The stomach was quite
empty; a small amount of fluid was found in the small intestine and some
faeces in the rectum.

Cat No. 11. This animal similarly treated only lost 10°/, of its body
weight, the weights recorded from time to time being 2'8, 2°7, 2°6, 2°5 and
finally 2°5 kilos. The stomach was empty, but there was some fluid in the
intestine and faeces in the rectum. The weight of the contents was 92 g.,
and this was found to contain 0:24 g. of cholesterol.

Immediately the animals were killed, they were cut up and the various
tissues submitted with the least possible delay to the treatment described
below. The following were examined—blood, liver, muscle taken from
various parts of the body, kidney, suprarenal glands, heart, brain, lung and
contents of the gall-bladder.

METHOD OF EXTRACTION AND ESTIMATION OF CHOLESTEROL.

The tissues were finely minced and ground up with sand and enough plaster
of Paris to make the whole mass fairly dry and friable. It was allowed to
stand until quite dry, powdered and extracted in a Soxhlet’s apparatus with
ether for about 14 days. The ethereal extract was made up to known
volume, and divided into two portions, or in some cases into suitable aliquot
portions. In one portion the free cholesterol was estimated and in the other
the total free and ester cholesterol after saponification in ethereal solution
by means of sodium ethylate. The cholesterol was estimated by the modifi-
cation of Windaus’ digitonine process described by Fraser and Gardner
[1910]. The ester cholesterol was thus determined by difference, and it was
assumed that no hydrolysis of the esters took place in the time occupied in
dissecting out the tissue, mincing up and drying with plaster of Paris.
Though we think this assumption is probably justifiable in the case of
most of the tissues, it is perhaps open to some question in the case of liver.
This tissue is known to contain ferments, which hydrolyse cholesterol esters,
but the hydrolysis does not appear to take place rapidly even under most
favourable conditions. Thus a sample of liver which on treatment as

a

=

y
a Co

J. A. GARDNER AND P. E. LANDER 581

described above was found to contain 0°28 °/, of total cholesterol and 0°13 °/,
of free cholesterol and therefore 0°15°/, of ester cholesterol, was still found
to contain 0:05°/, of ester after incubating for four days at 37° in saline
suspension in the presence of a little toluene, though another sample allowed
to putrify in saline suspension at 37° yielded no ester. We endeavoured
however to reduce any such error to a minimum by carrying out the
operations as rapidly as possible.

DISCUSSION OF ANALYTICAL RESULTS.

Blood. The results of the analysis of the blood of the eleven animals are
collected together in Table I.

TABLE I.
Total free
Weight of and com- Cholesterol
blood taken bined cholest- Free cholest- in form of
No. of for analysis, erol in g. per erol in g. per esters per
cat Diet in grams 160 g. of blood 100 g. of blood 100 g. of blood
1 A 68-2 0-044 0-044 Nil
(non chol.)
2 A 87:9 0-060 0:046 0-014
3 A 84-1 0-058 0:056 0-002
4 A 83-2 0°044 0-040 0-004
Mean of 4 cats 0052 0:047 0005
5 B 74:2 07102 0098 0-004
(cholest.)
6 B 59°8 0-178 0:049 0-129
Mean of 2 cats 0:140 0'074 o:'066
{i C 126:2 0-083 0-066 0:017
(meat)
Sr C 74:2 0:078 0°054 0-024
Mean of 2 cats foot: # | 0'060 o'021
9 Inanition 65°8 0°163 0-074 0-089
10 Inanition 76°5 0:196 0-069 07127
11 Inanition 75°6 0-151 0:126 0-025
Mean of 3 cats o-:170 o:090 o:'080

On comparing the figures in column four of Table I it will be seen that
the total cholesterol of the blood increases with the cholesterol taken in
with the food. In the case of the animals in a state of inanition, which
were living on their own tissue, the increase is very marked. This increase is
in all cases chiefly due to an increased quantity of ester cholesterol, though the
free cholesterol itself also increases. The high ester content is very marked
in the case of animals in a state of inanition. This may very likely have

Bioch. vi . 38

582 J. A. GARDNER AND P. E. LANDER

something to do with the transference of the fat from the depots during
starvation, and in this connexion it may be noted that in the case of cats
9 and 10 which lost 19°/, of their body weight their blood contains
0:089 and 0:127°/, of ester cholesterol, while that of cat 11, which only
lost 10°/, of body weight, is 0°025. In the case of cats 2 and 3, both
of which lost weight during the diet period, the total cholesterol of the
blood is somewhat higher than in the case of cats 1 and 4 which had the
same diet but remained constant in weight. The same thing will be noticed
in the case of cat 6 compared with cat 5 on the same diet.

The variations observed are similar to those found in the case of
rabbits fed on diets containing varying amounts of cholesterol, and the
figures are of much the same order as will be seen from Table I in which
only average values are given.

TABLE II.
Number of Total free
animals of and
which the 2 combined Free Ester
figures are cholesterol cholesterol cholesterol
the average Animal Diet per cent. per cent, per cent.
4 Cat Cholesterol-free A BS 0°052 0:047 0:005
4 Rabbit! Bran extracted by ether ... 0:058 0-044 0-014
2 Cat B (A+ cholesterol) mae 0-140 0:074 0-066
4 Rabbit! Extracted bran + cholesterol 0-081 0:059 0:022
2 Cat C (lean beef) a 500 0-081 0-060 0-021
2 Rabbit! Ordinary bran ... a 0:082 0:069 0:013
2 Rabbit! Cabbage leaves and stalk ... 0:068 0:063 0-005
3 Cat | Inanition a 000 0-170 0-090 0-080
2 Rabbit! Inanition - : 0:154 0:113 0-040

1 [Hillis and Gardner, 1912, 2.]

Liver. In Table IIl we give the total cholesterol, free and combined,
actually found in the livers of the animals, also the weights of cholesterol
per cent. of liver and the weights of liver cholesterol per kilo. of body weight.
On comparing the figures for animals fed on diet A (cholesterol free) it will
be seen that both the total percentages of cholesterol in the liver and the
total liver cholesterol per kilo. of body weight are fairly uniform, though
not so constant as was found to be the case in rabbits fed on cholesterol-free
diet. On comparing these figures with those of cats 5 and 6 which were
fed for a period on the same diet, but with the addition of free cholesterol,
a marked increase in the total cholesterol is obvious. This increase is partly
due to free cholesterol, but mainly to ester cholesterol. On the other
hand in the case of cats 7 and 8, fed on lean meat, 1e. a diet containing

J. A. GARDNER AND P. E. LANDER 583
TABLE III.
Wet. of Wt. of Wt. of Wt. of
animal total cho- free cho- ester cho- Wt. of cholesterol Wt. of liver cholesterol
when Wt.of lesterol lesterol lesterol per cent. of liver per kilo. of body wt.
No. of killed, liver, in liver, inliver, in liver, -—— A> ——- —~.
cat Diet inkilos. ing. in g. in g. in g. Total Free Ester Total Free Ester
1 A 2°70 59°8 0-128 0-092 0°036 0-214 0-154 0-060 0-047 0:°034 0-013
2 A 2°55 54°6 0-136 0-041 0-095 0249 0-076 0:173 0:°053 0-016 0:037
3 A 2°30 63-0 0-102 0:102 Nil 0:162 0°162 Nil 0044 0°044 = Nil
4 A 3°10 70°8 0-192 0:048 0-144 0-254 0:067 0-187 0:062 0°015 0-°047
Mean values 0O'220 0'115 0'105 0'052 0'027 0'025
B 2°20 48-4 0-194 0-089 0-105 0-401 07185 0°216 0:088 0°041 0:047
6 B 2°10 43°7 0-228 0-049 0°179 0:466 0-117 0°349 0-108 0:023 0-085
Mean values 0'434 0'151 O°283 0: 098 0'032 0'066
7 Cc 3°43 77°5 0-243 0-221 0-022 0°316 0°285 0:031 0-071 0-064 0-007
8 C 3°30 107°7 0-362 0-362 Nil 0°346 0°346 = Nil 07109 0-109 Nil
Mean values 0'331 0'316 0'015 O'O090 0'087 0'003
9 Inanition 2°50 50°3 0-345 0-063 0-282 0-550 0°125 0°425 0°138 0:025 0°113
10 Inanition 2°50 49°8 0°295 0-069 0-226 0°637 0°143 0°494 0°118 0:027 0-091
11 Inanition 2°50 48°2 0-245 0:059 0-186 0°509 07122 0-387 0:098 0:024 0-074

Mean values

a considerable proportion of cholesterol but no added carbohydrate, there
is a similar increase in total cholesterol, but the increase is almost entirely
free cholesterol, the ester being nil in one case and considerably below
the average of diet A in the other. In the case of cats 9, 10 and 11, in
which the animals were living on their own tissues, we find a similar storing
up of the cholesterol in the liver. It is stored almost entirely in the form
of esters and there is only a slight increase if any of free cholesterol. This
may also have something to do with the transference of the fat from the
depots to the liver during starvation. In the case of cat 11, which lost
only 10°/, of body weight, this storimg up is not quite so great; in this cat
the percentage in the blood was also somewhat lower than in the case of
cats 9 and 10. Those changes are generally similar to those observed by Ellis
and Gardner [1912, 1] in the case of rabbits during starvation. Assuming
that the loss in weight of the cats during inanition is due to muscle tissue,
then, if the cycle postulated for herbivora be true for carnivora, cats 9 and 10
both of which lost 600 g. of body weight should have accumulated each about
0:5 g. of cholesterol in the blood, liver and the contents of the intestine.
Taking the average figures for diet A as representing the normal cholesterol
content of the blood and liver under conditions in which body weight is kept
constant, but no cholesterol is absorbed with the food, 0°42 g. of cholesterol
was accumulated in the blood and liver of cat 9 and 0-41 in cat 10. The

38—2

0'565 0°'130 0'435 O'118 0'025 0'093

584 J. A. GARDNER AND P. E. LANDER

contents of their intestines were not analysed, but in cat 11 about 0:2 g.

was found.
Suprarenal Glands. This organ was only examined in five of the eleven

cats, and the results are given in Table IV.

TABLE IV.
Total cho- Free
Weight _ lesterol, free cholesterol J, of cion fy Oxi
No. of of organ, and combined, actually total free ester
cat Diet in g. actually found found cholest. cholest. cholest.
3 A 0°6569 0:0116 0:0060 Hore 0:91 0°86
4 A 0°5886 0:0155 0:0034 2°63 0°58 2°05
5 B 0°6310 0:0373 0:0373 5°92 5°92 Nil
6 B 0:4408 0:0228 0:0070 5:13 1°59 3°58
11 Inanition 0:°5208 0:0048 00035 0°92 0:67 0:25

Here the addition of cholesterol to the diet produces a marked increase
in the total cholesterol content of the organ, and the change appears to run
parallel to that in the blood. In cat 5, for instance, the increase in the blood
is all free cholesterol, the actual amount of ester cholesterol being slightly
below the average of the cats on diet A. In the suprarenal the increase is
also all free cholesterol, the ester being nil or very small. In cat 6 on the
other hand, where the increase is almost entirely in ester cholesterol, the
ester cholesterol of the blood is very high, while the free cholesterol is
almost exactly equal to the average on diet A. In inanition however this
parallelism no longer appears to hold, for while the cholesterol accumulates
in the blood it decreases almost to the vanishing point in the suprarenals.
More experiments will however be required before this interesting point can
be settled. The inanition result appears, however, to be in agreement with
the observations of Landau [1913]. In this connexion it may also be noted
that Albrecht and Weltmann [1911] and Hueck [1911] have shown that in
diseases such as Carcinoma, Tuberculosis, etc., both the free and ester
cholesterin of the suprarenals sink and sometimes are absent altogether.

Muscle. The muscle analysed was a mixed sample cut from various parts
of the body. The results are given in Table V.

TABLE V.
No. of : Percentage Percentage Percentage
cat Diet total cholesterol free cholesterol ester cholesterol
1 A 0:056 0:031 0-025
2 A 0-047 0-032 0:015
7 C 0-087 0-083 0:004
8 C 0-081 0-073 0-008
9 Inanition 0:059 0-053 0-006
10 Inanition 0:065 0:051 0:014

J. A. GARDNER AND P. E. LANDER 585

It will be noticed that on diet A, which contained no cholesterol, but was
rich in carbohydrate, there is a marked increase in the ester cholesterol
compared with the cats on meat diet C. The free cholesterol is however
much lower. In the case of rabbits, Ellis and Gardner were unable to trace
any influence of food cholesteré! on the cholesterol content of muscle.

Heart. The results are given in Table VI.

TABLE VI.

No. of Wt. of Total cholest. Free cholest. Ester cholest.
cat Diet organ, in g. per cent. per cent. per cent.
1 A 79 0-080 0:073 0:007
2 A 8°8 0-068 0068 Nil
7 C 16°3 0:073 0-073 Nil
8 Cc 12°5 0-072 0:063 0-009
9 Inanition 10°5 0-071 0-063 0-008
10 Inanition 9°4 0-091 0:039 0-052

The cholesterol of this organ is very constant, and nearly all present in
the free state. The percentage is much the same as that given for heart
muscle in the ox, 0:066 to 0:071 by Ellis and Gardner [1908].

Kidney. The results on this organ are given in Table VII.

TABLE VII.
Wt. of

animal Wt. of cholesterol Wt.ofkidney cholesterol

when Wt.of Totalcho- Freecho- Estercho- per cent. of kidney per kilo. of body wt.
No. of killed, kidney, lesterol, lesterol, lesterol, —— SS SS
eat Diet inkilos. in g. in g. in g. in g. Total Free Ester Total Free Ester

1 A 2°7 21°6 0-064 0-064 Nil 0:298 0:296 0:002 0-024 0-024 Nil

2 A 2°6 17°8 0-056 0-056 Nil 0°326 0°326 = =6Nil 0022 0:022 Nil
7 C 3°43 27°5 0-073 0-069 0-004 0:265 0:252 0-013 0:021 0-020 0-001
8 C 33 24°8 0-077 0-070 0:007 0-313 0:284 0:029 0:023 0-021 0:002

Mean values O'301 0:'290 O'O11

9 Inanition 2°5 20°8 0-070 0°053 0:017 0°335 0-212 0:123 0:028 0-021 0-007

10 Inanition 2°5 23°8 0-066 0°054 0-012 0:275 0°226 0:049 0:026 0:022 0-004

The diet as in the case of rabbits appears to have no influence on the
kidney cholesterol. The average values for normal animals are not very
dissimilar from the mean of those given by Windaus [1910] for normal
human kidneys, viz. 0°24 free and 0-02 ester per cent. In the case of the
starved animals the ester cholesterol shows a marked increase, which Ellis
and Gardner also showed to be the case in the rabbit. These high ester
values recall the high ester value found by Windaus in human pathological
kidneys.

Lung. The results are given in Table VIII.

586 J. A. GARDNER AND P. E. LANDER
TABLE VIII.

Wt. of Percentages of cholesterol
No. of organ taken for ee =~

cat Diet analysis, in g. Total Free Ester

1 A 17:1 0:416 0-295 0-121

2 A 156 0°361 0:296 0-065

7 C 17°5 0:435 0°247 0-188

8 C 11:4 0°453 0-270 0:183

9 Tnanition 13°4 0-445 0°413 0:032
10 Inanition 16°5 0°376 0°376 Nil

The cholesterol content, as one would expect, appears to have nothing to
do with diet. The values are of much the same order as those given for
rabbits by Ellis and Gardner, viz. average of six animals, free 0'442 and
ester 0053. The ester cholesterol, as in the case of rabbits, is very variable.

Brain. The brains of cats 1 and 2 on diet A, 7 and 8 on diet C and
9 and 10 in a state of inanition, were examined with the following results :—
3°05, 2°65, 2°94, 2°91, 2°33 and 2°06 per cent. In no case was any ester found.
The average is 2°66. For rabbits, Ellis and Gardner give as an average of
eight experiments 2°32, with a variation of 2°02—2°88, and in no case was
there any evidence of the presence of esters.

Bile. Attempts were made to estimate the cholesterol content of the
bile of the cats under different diets and during starvation, but the contents
of the gall-bladder were so small that it was difficult to estimate the chol-
esterol with any degree of accuracy. We give however the results for what
they are worth in the following Table IX.

TABLE IX.
; Percentage of cholesterol
No. of Wt. of bile in 2 (ee a
cat Diet _—_gall-bladder, ing. Total Free Ester
1 A 0:96 0°56 0°42 0-14
2 A 1°48 0:28 0°28 Nil
7 C 0°61 0°36 0°34 0°02
8 C 1:30 0°61 0°61 Nil
9 Inanition 2°53 0°25 0:03 0°22
10 Inanition 1°62 3°01 3°01 Nil

The results of the experiments on diets A and C do not indicate any
connexion between either the actual amount of cholesterol in the bile or the
percentage content and the diet. In the case of cat 10 there was a relatively
high percentage of cholesterol in the bile, and this appeared to be all or
mainly in the free state. This is in agreement with the observations of
Ellis and Gardner on starving rabbits. On the other hand in the case of
cat 9, also in a state of starvation, both the actual amount found and the

percentage are of much the same order as in the fed animals, but it was
nearly all found in the ester condition.

J. A. GARDNER AND P. E. LANDER 587

GENERAL CONCLUSIONS.

The results obtained in these experiments are strikingly similar to those
obtained by Ellis and Gardner on the rabbit, and are in agreement with the
hypothesis advanced at the beginning of this paper, viz. that cholesterol
is a constituent constantly present in all cells, and that when these cells are
broken down in the life process the cholesterol is not excreted as a waste
product, but is utilised in the formation of new cells. A function of the
liver is to break down dead cells, e.g. blood corpuscles, and eliminate their
cholesterol in the bile. After the bile has been passed into the intestine in
the process of digestion, the cholesterol is reabsorbed, possibly in the form
of esters, along with the bile salts and is carried in the blood stream to the
various centres and tissues for reincorporation into the constitution of new
cells. Waste of cholesterol is made up from that taken in with the food.

We take this opportunity of thanking the Government Grant Committee
of the Royal Society for help in carrying out this work.

REFERENCES.

Albrecht and Weltmann (1911), Wiener med. Wochensch. No. 14.
Bondzynski (1896), Ber. 29, 276.
and v. Humnicki (1896), Zeitsch. physiol. Chem. 22, 396.
Dorée (1909), Biochem. J. 4, 72.
Dorée and Gardner (1908, 1), Proc. Roy. Soc. B, 80, 212.
—— —— (1908, 2), Proc. Roy. Soc. B, 80, 227.
—— (1909), Proc. Roy. Soc. B, 81, 109.
Ellis and Gardner (1908), J. Physiol. 38, Proc. i.
—— —— (1909, 1), Proc. Roy. Soc. B, 81, 129.
—— —— (1909, 2), Proc. Roy. Soc. B, 81, 505.
B,
B,

—— —— (1912, 1), Proc. Roy. Soc. 84, 461.
—— —— (1912, 2), Proc. Roy. Soc. 85, 385.
—— — (1912, 3), Proce. Roy. Soc. B, 86, 13.
Fraser and Gardner (1909), Proc. Roy. Soc. B, 81, 230.
(1910), Proc. Roy. Soc. B, 82, 559.
Goodman (1907), Beitrdge, 9, 91.

Grigaut (1913), Le Cycle de la Cholestérinémie.
Harley and Barratt (1903), J. Physiol. 29, 341.
Hueck (1911), Miinchener med. Wochensch. 2588.
Kosumoto (1908, 1), Biochem. Zeitsch. 13, 354.
(1908, 2), Biochem. Zeitsch. 14, 416.
Landau (1913), Deut. med. Wochensch. 12, 1.
Miiller (1900), Zeitsch. physiol. Chem. 29, 129.
Pribram (1906), Biochem. Zeitsch. 1, 414.
Wacker (1912), Zeitsch. physiol. Chem. 80, 381.
Windaus (1909), Ber. 42, 238.

—— (1910), Zeitsch. physiol. Chem. 65, 110.

LV. ON THE OXIDATION OF COPROSTEROL
AND COPROSTANONE.

PAR Tai
By JOHN ADDYMAN GARDNER anp WILLIAM GODDEN.
From the Physiological Laboratory, University of London, South Kensington.
(Received October 31st, 1913.)

Dorée and Gardner [1908] showed that coprosterol, C.,H,O, was readily
oxidised on treatment with the theoretical quantity of chromic acid with
the production of the corresponding ketone, coprostanone, C;H,0:- Bhis
coprostanone crystallised in glistening leaves which, under the microscope,
appeared to consist of thin plates, generally square, with one or all of the
corners slightly truncated. It melted at 62-63° to a clear liquid. It yielded
an amorphous semicarbazone, melting at 192°, and an amorphous oxime,

melting at 71°. On treatment with phenylhydrazine it behaved in an ~

abnormal manner and yielded a crystalline derivative, melting at 192°, which
was subsequently shown by Dorée [1909] to be coprosteryl-carbazole, formed
from the coprostanone-phenylhydrazone with elimination of ammonia. As
the yield of pure coprostanone in the above preparation was only 60°/,, it
appeared desirable to submit the reaction to further investigation.

Oxidation of coprosterol by chromic acid. 'To make sure of oxidising the
whole of the coprosterol, a considerable excess of chromic acid was used—
three times the amount taken by Dorée and Gardner.

10 g. coprosterol dissolved in 130 ce. of glacial acetic acid, and heated
to 70°, were treated in the course of 1 hour with a solution of 8:5 g. chromic
anhydride in a little dilute acetic acid, the temperature being maintained by
the heat of the reaction. After adding the chromic acid, the liquid was
heated on the water bath for one hour, and then diluted with water. The
bulk of the acetic acid was next neutralised with sodium carbonate, and the
still strongly acid liquor was thoroughly extracted with ether. The ethereal
solution was now thoroughly extracted with dilute caustic soda, and on
evaporating the ether nearly pure coprostanone was obtained in a yield in all
experiments of about 70 per cent.

. J. A. GARDNER AND W. GODDEN 589

The caustic soda extracts on acidification with hydrochloric acid gave
a white crystalline precipitate of an acid mixed with an oily substance. The
oil was got rid of by washing with petroleum ether, in which the crystals
were insoluble. During the extraction of the ethereal solution of the
oxidation product with caustic soda a small quantity of an insoluble basic
chromium salt also separated, and this on treatment with hot hydrochloric
acid yielded a further quantity of the crystalline acid. This acid is soluble
in ether, but only difficultly soluble in benzene and in chloroform. It is
sparingly soluble in hot acetone and somewhat more readily in boiling ethyl
acetate or alcohol. It can be readily purified by recrystallisation from a
mixture of ether and acetone or from ethyl acetate. It usually separates as
a fine powder, and under the microscope appears in the form of well-defined
square rods. It melts at 247° to a clear liquid, without any decomposition.
For analysis the substance was dried in vacuo. It proved rather difficult to

burn.
(1) 0°2054 g.; 05604 CO,; 0-1934 H,O.
(2) 02433 g.; 0°6628 CO,; 0-2358 H,O.
Found (1) C=74:41; (2) 74:30. Cale. for C,,H,,0,, C=74:94; C,,H,,0,, C=74°59.
H=10°46; 10-76. H=10-26; H=10°67.

In other combustions in which the carbon came low, the hydrogen was
10°38, 10°23 and 10°63.

Sodium salt. On titration with caustic soda in alcoholic solution
0°6304 g. required 14°43 cc. N/5 NaOH. With another specimen 03134 g.
required 7:24 ce. N/5 NaOH. Mol. wt. found (1) 436°8, (2) 432°8. Cale. for
C.,HyO,, 434.

Ammonium salt. The acid dissolved in strong ammonia giving a soapy
solution. The bulk of the free ammonia was boiled oft, and on cooling the
liquid set to a perfectly clear stiff jelly. The jelly dried up in vacuo, with
gradual evolution of ammonia, to a white amorphous mass, which dissolved
in hot water giving a slightly opalescent soapy solution. For analysis the
substance, dried finally at 100°, was decomposed by excess of standard acid,
filtered and the acid titrated back with soda. 0°3157 g. was decomposed with
25 ce. N/5 sulphuric acid and the excess of acid’ required 42°5 cc. soda.
25 ec. N/5 acid=49'2 soda; percentage of N=3:19. Cale. for C.,H,O,N =3:1°/,.

Silver salt. This was prepared by precipitating a neutral solution of the
sodium salt with silver nitrate. The white precipitate was well washed with
water, in which it was quite insoluble, and dried at 100°.

(1) 0°298 gave on ignition 0-0998 g. of silver.
(2) 0-1467__,, oh i, CO es

Mol. wt. of acid found (1) 430°6. Cale. for C,,H,,0,, 434.
(2) 433-8.

590 J. A. GARDNER AND W. GODDEN :

From the analysis of these salts and the combustions of the acid itself, the
acid appears to be a dibasic acid of the formula CG, H,07

The oil formed along with this acid and dissolved away by petroleum
ether as described above, would not crystallise on long standing. It was an
acid and formed an insoluble barium salt on precipitating the solution in
ammonia with barium chloride. This substance has not yet been further
investigated.

Oxidation of coprostanone by chromic acid. In order to ascertain whether
the acid C,,H,0, was a bye-product in the preparation of coprostanone, or
formed by the further oxidation of the coprostanone by the excess of chromic
acid, one gram of coprostanone was dissolved in glacial acetic acid and heated
to 70° with 0°35 gram of chromic anhydride dissolved in dilute acetic acid
until all the red colour had gone. It was then heated to boiling, cooled and
poured into water. The turbid liquid was extracted with ether, and on
evaporating the ether a green oil containing chromium was obtained. ‘This
was freed from chromium by warming with hydrochloric acid on the water
bath. The liquid was then extracted with ether, the ethereal solution
evaporated and the brown oil left taken up with petroleum ether. The
petroleum was allowed to evaporate spontaneously and left a brown oil, which
on long standing in a desiccator deposited some crystalline matter. The oil
was dissolved away by petroleum ether, and the crystalline matter recrystal-
lised from hot ethyl acetate. It melted at 247° and was identical with the
acid described above. ‘The yield was however small. The oil dissolved by
the petroleum was acid in character and formed an insoluble barium salt.

Owidation of coprostanone by ammonium persulphate. As the yield of
acid C,,H,O, was small by both methods of preparation, we thought it
desirable to try other oxidising agents and ultimately selected ammonium
persulphate. This reagent is known to act on cyclic ketones [Baeyer and
Villiger, 1899] breaking the ring and changing them to hydroxy-acids or their
lactones; and we hoped in this way to break the ring of the coprostanone at
the CO group and obtain a hydroxy-acid which might on further oxidation
give the dibasic acid C.,H,,O,. 10 g. coprostanone were dissolved in 375 ce.
of glacial acetic acid, and to the solution were added 10 g. ammonium
persulphate, dissolved in the least possible amount of water and diluted with
an equal volume of glacial acetic acid. The mixture was then heated on a
water bath and another 10 g. of powdered ammonium persulphate gradually
added. The heating was continued for four hours, during which gas was
slowly evolved. This gas was mainly oxygen from the persulphate, but
contained some carbon dioxide. After standing several days the liquid was

4
$

J. A. GARDNER AND W. GODDEN 591

diluted with an equal volume of water and thoroughly extracted with ether.
The ethereal solution was repeatedly shaken with water to get rid of the bulk
of the dissolved acetic acid, and then extracted with 10 °/, caustic soda solution
several times, until a sample of the alkaline extract no longer gave a
precipitate on acidification. The soda solution on shaking with the ethereal
extract turned dark brownish red in colour. The various caustic soda
extracts were mixed and on standing overnight bulky gelatinous clots of a
colourless sodium salt separated out (A). These were somewhat difficult to
filter, but it was found possible to syphon off the greater portion of the
mother liquor. The mother liquors and filtrates, which were now pale yellow,
on acidification with hydrochloric acid gave an amorphous precipitate of an
acid substance (B). The ethereal solution, freed from acid matter, gave on
evaporation a small quantity of a neutral substance (C).

Examination of (A). The clots were readily soluble in hot water giving
a clear soapy solution, from which they separated on cooling, but more readily
if the solution was made strongly alkaline. On acidifying the aqueous
solution with hydrochloric acid a crystalline precipitate was thrown down
which, in the preparation described, amounted to about 65 °/, of the weight
of the coprostanone taken. In different preparations, however, the relative
amounts of (A), (B) and (C) varied considerably with the conditions of
reaction.

A small quantity of the clotted matter was filtered on the pump, washed
with cold water to get rid of excess of alkali and further purified by means
of alcohol. It was obtained as a white powder perfectly soluble in hot water.
For analysis it was dissolved in water and precipitated by a measured
quantity of standard acid. The precipitate was filtered, well washed and the
filtrates titrated with alkali.

0-2488 g. was found to contain sodium equivalent to 6:1 ce. N/10 caustic
soda ; percentage sodium = 5°64. Cale. for C,,H,,O,Na = 5°21.

The crystalline precipitate obtained on acidifying the aqueous solution of
this sodium salt can be separated into two substances by fractional crystalli-
sation from ethyl acetate or better from alcohol. The less soluble substance,
present in small quantity, crystallised in thin plates and after several
erystallisations melted at 183-184°. The more soluble substance, con-
stituting the bulk of the material, crystallised in minute needles and
melted at 157-158".

Substance melting at 157-158°. This body was easily soluble in ethyl
acetate, acetone and benzene, but somewhat less so in alcohol. It was
sparingly soluble in petroleum ether, but more soluble in the fraction of

592 J. A. GARDNER AND W. GODDEN

petroleum boiling at 156-195°. On combustion the fallowing results were
obtained :—

(1) 0°1549; 0-4568 CO,; 0-1595 H,O.

(2) 0-1571; 0:4625 ,, ; 0-1617 ,,

(3) 0°1640; 04834 ,, ; 01680 ,,

(4) 0-2557; 0°7503 ,, ; 0°2627.,,

I Il Til LIN Cale. for C.,H,,0,
C 80°42 80°29 80°40 80-02 C 80°52
H 11-44 11°43 11°38 11°41 H 11:50

Combustion does not however throw much light on the number of carbon
atoms in the molecule, as decrease of the molecular weight by CH, makes
very little difference in the percentage composition thus—

C H
Co,H1gOo Zs a 80°52 11°50
CogH 4402 eae ee: 80°34 11°42
Co5Hy20e viele eee 80°14 iW oi
Co4H 4902 aes wee 719-93 11:19

The molecular weight was ‘therefore determined by depression of the
freezing point of benzene :—

(1) 0-2975 g. in 20 cc. benzene of sp. gr. 0°883, depression of freezing point 3°115° to 2°915°.
(2) 05993 g. in 15 ce. benzene of sp. gr. 0-883, depression of freezing point 3:115° to 2°565°.

Molecular weight found I 412°7. Cale. for Co7Hyg02, 402°37.
II 403:1.

The substance did not go into solution on boiling for a short time with
an aqueous solution of caustic soda, nor on shaking a solution in ether
with either 10, 20 or 30°/, caustic soda, though as originally obtained in the
acetic acid solution it did so quite readily. It apparently contained no OH
group, as it was quite unaffected by prolonged boiling with either acetic
anhydride or acetyl chloride, and on pouring into water and extracting with
ether it was recovered unchanged in both cases. It was probably a very
stable lactone of a hydroxy-acid of the formula C.,H,O;.

Examination of (C). The neutral substance left in the ether after
extraction with caustic soda, can be readily purified by recrystallisation from
alcohol, or ethyl acetate, or petroleum ether. It melted at 183-184° and was
identical with the substance of the same melting point obtained along with
the lactone melting at 157-158°. This was proved by a mixed melting point.
It crystallised in long thin plates, or glistening leaves, which under the
microscope appeared to be rectangular plates, some square, some long. It
proved difficult to burn and on combustion gave the following results :—

J. A. GARDNER AND W. GODDEN 593

(1) 0-1197; 0°3514 CO,; 01235 H,O.
(2) 0°1720; 0°5052 ,, ; 0-1672
(3) 0-1455; 04264 ,, ; —
(4) 0°1580; 0-4615 ,, ; 01594 ,,

I I Ill IV
C 80-06 80-11 79°92 79°66
H 11°46 10°80 — 11°21

0°8365 g. in 15 ce. benzene of sp. gr. 0°883, depression of freezing point
076°. Molecular weight found, 407; calc. for C,,H,O,, 402°37. The sub-
stance was not affected by shaking in ethereal solution with either 10, 20 or
30 °/, caustic soda solution.

The analyses threw no light on the relationship of the substance to the
lactone melting at 157-158°. The molecular weight determination perhaps
suggests that it is a stable lactone isomeric with the 157-158" body. It may
however be a lactone of lower molecular weight produced by further oxidation,
or possibly some derivative of a ketonic or diketonic nature. ‘To test this 0°95 g.
of the lactone (M.P. 156°) was dissolved in glacial acetic acid and 5°7 cc. of 10°/,
chromic acid solution added. The mixture was kept at 70° for half an hour,
and then the temperature was raised to the boiling point for a few minutes.
The liquid was then poured into cold water and the precipitated matter
filtered off. The filtrate was extracted with ether and a further quantity of
solid obtained, which was added to the precipitate. This precipitate was
crystallised from alcohol from which it separated in needles and_ plates,
melting at 169-171°. This weighed 0°38 g. or 40°/, of the weight of the
original substance. After several recrystallisations it melted at 183-184°.
This was shown by a mixed melting point determination to be identical with
the substance melting at 183-184° obtained from (C). The lactone (MP.
157-158°) appears therefore to yield this neutral substance melting at
183-184° on oxidation. The latter body itself on similar treatment did
not appear to be further oxidised, and was recovered unchanged. It did
not appear to have ketonic properties for the following reasons: (1) sodium
amalgam had no effect on the solution of the substance in moist ether in
presence of sodium bicarbonate ; (2) in attempts to prepare a semicarbazone
in the usual manner, the substance was recovered unchanged ; (3) on heating
1 mol. of the substance in alcoholic solution with 6 mols. of hydroxylamine
hydrochloride and the equivalent amount of dry sodium acetate for 20 hours,
the substance was recovered quantitatively and unchanged. On repeating
experiment (3) however using caustic soda instead of sodium acetate, needle
shaped crystals were obtained on evaporating the alcohol. These melted at
268-274". This was not however an oxime as it contained only a trace of

594 J. A. GARDNER AND W. GODDEN

nitrogen—under 1°/,. The same or a very similar substance, melting at 264—
286°, was also obtained on substituting phenylhydrazine hydrochloride for
the hydroxylamine in the above experiment. This contained about 0°9°/,
of nitrogen. We had not sufficient material for the further investigation
of these substances. It would seem probable however from these results
that the substance of M.P. 183—184° is a stable lactone either isomeric with
or containing less carbon than the lactone of M.p 157-158".

Examination of substance (B). The acid precipitated on acidifying the
alkaline solution was a light brown amorphous resin-like substance, readily
soluble in alkalis giving a soapy solution. It was easily soluble in the
common organic solvents, but would not crystallise. From petroleum ether
it dried up to a brittle resin. It was dissolved in ammonia and reprecipitated
by hydrochloric acid several times, and well washed. After drying it melted
at 88° to astiff paste and decomposed at 114°. The yield was only small.

Barium salts. The resinous acid was dissolved in ammonia, the excess of
alkali exactly neutralised by hydrochloric acid and the solution precipitated
with baryta water. The precipitate, after being well washed with water,
dried to a light brown powder. This was readily soluble in chloroform, and
when the solvent was allowed to evaporate spontaneously separated on the
sides of the beaker in light brown brittle flakes. On analysis—

(1) 0-1367; 0-0327 BaSQ,.
(2) 02173; 0-0494_,,
(3) 03541; 0°8468 CO,; 0-3047 H,0.

Ba C H

I 14°11 —

II 13°50 - os

Ill a= 65°22 9°56

Cale. for Cy7H4;0,Ba,/, 13°68 64°53 9°03
Cale. for C54H9;0;Ba 13°83 65°25 9°64

The filtrate obtained after precipitating with baryta in the preparation of
the above salt was acidified with hydrochloric acid and the small amount
of amorphous resin precipitated was filtered and well washed. It was then
dissolved in hot baryta water and the excess of baryta precipitated by carbon
dioxide. The milky liquid was well boiled and the barium carbonate
filtered off. The filtrate was evaporated to small bulk and left in a desiccator.
Crystals gradually separated, which under the microscope appeared to be
small plates. There was only sufficient for a barium estimation after drying
at 100°.

0:1424; 0:0590 BaSO,; Ba found=24°39; Cale. for Co7HyyO,Ba=24:12; C;,H930;Bag= 24°35.

J. A. GARDNER AND W. GODDEN 595

We hope shortly to be in a position to give a further account of the
reactions of these various substances.

We take this opportunity of expressing our thanks to the Government
Grant Committee of the Royal Society for assistance in carrying out this

work,

REFERENCES.

Baeyer and Villiger (1899), Ber. 32, 3625.
Dorée and Gardner (1908), J. Chem. Soc. 93, 1629.
Dorée (1909), J. Chem. Soc. 95, 653.

LVI. A NOTE ON A MODIFICATION OF
TEICHMANN’S TEST FOR BLOOD.

By CLAUDE TREVINE SYMONS.
From the Government Analyst's Laboratory, Colombo, Ceylon.
(Received November 4th, 1913.)

For the identification of blood as such, several tests are in common use.
But tropical conditions, such as are found in Ceylon, in many cases render
some of these tests of little value. Hence I do not hesitate to put forward
this modification of an old test, as I have found it to be very satisfactory.

The routine methods which may be adopted in cases of suspected blood
stains are as follows:

I. The benzidine test. Benzidine is freshly dissolved in a small quantity
of glacial acetic acid and hydrogen peroxide is added. This will give an
immediate blue colour if it comes in contact with any blood. As, however,
the immediate blue colour is produced by some other substances, the test
merely serves to weed out a large number of stains which superficially
resemble blood stains.

II. The spectrum test. In ordinary routine work it has been found
that in suitable cases good results are obtained by soaking out the stain
with ammonium hydrate and then adding ammonium sulphide to obtain the
haemochromogen spectrum. But with very small stains there is often in-
sufficient material to get the spectrum without special microscopical apparatus.
In addition, when the stains have been long exposed to the tropical sun, they
are apparently rendered so insoluble that the pigment is broken up before
a solution can be obtained for the spectroscope. Further, in cases where
the stain has been rusted on to iron and is not in large quantity, it has been
found to be practically impossible to obtain a solution which will yield
a spectrum.

Ill. The corpuscle test. This has been found to be of very doubtful
value and very often impracticable in the case of old stains.

IV. TYeichmann’s haemin crystal test. This at first sight appears to
provide a very convenient and reliable method of making a complete

(ANALG re GA

x

C. T. SYMONS 597

identification of blood as such. In routine practice in the tropics, however,
I understand that it has been largely discontinued, in consequence of the
hitherto insuperable difficulties, which arise under certain conditions. As
in the case of the spectrum test, old stains which have been exposed for
a long time to air and sun in the tropics, and stains which are rusted on
to iron, usually give negative results, even in cases where it is known that
the stains were caused by blood.

Therefore I considered it of interest to examine the literature on the
subject, and to evolve if possible some modifications which would give better
results. The routine method in vogue had been to use glacial acetic acid
with a trace of sodium chloride. Following the recommendations of various
writers, I used, in place of sodium chloride, at first potassium iodide, and
later sodium iodide with the glacial acetic acid. Good results were obtained
with fairly fresh blood, but the results in the above mentioned obstinate
cases were still unsatisfactory. So an attempt was made to substitute some
other acid for the acetic acid.

Sulphuric, formic, valeric, butyric, and several other acids and combina-
tions of acids and salts were tried without much success. Eventually strong
lactic acid (sp. gr. 1:21 Kahlbaum) was tried with sodium iodide. This acid
is mentioned in the literature as having been tried by Teichmann himself.

At first a very small fragment of the stain was placed in a drop of a solution
of sodium iodide (approx. 5°/,) in water, on a glass slide. This was carefully
dried on an asbestos mat over a Bunsen flame. The cover glass was then
placed over it and the lactic acid allowed to run under the cover glass, so
as to cover the stain. The slide was then warmed on the asbestos mat and
examined for crystals. Excellent results were obtained.

Later, solid sodium iodide, freshly prepared, was dissolved in the lactic
acid to about 1°/,. This solution becomes brown on keeping, but the
change does not appear to impair its action, at any rate for some months.
The fragment of stain is covered with the solution, and the cover glass is
placed in position, and the whole slowly warmed as above to such a tempera-
ture that the solution is just about to boil under the cover glass. In most
cases five minutes treatment is sufficient to produce crystals. If they
have not then appeared, the heating must be continued; more solution
must be allowed to run under the cover glass if necessary, but this is not
often the case, as the lactic acid does not evaporate as rapidly as acetic
acid. The resulting crystals differ very much in size in different cases, being
usually rather small, but dark and perfectly characteristic.

By this method I have not failed to obtain crystals, in any case which

Bioch. vu 39

598 C. T. SYMONS

I have tried up to the present. Stains on filter paper which had been kept
for three years exposed to the air, and were then exposed to direct sunlight
for some hours, gave excellent results. An old stain, rusted on to a large
iron knife, also gave perfectly clear results. These two failed to give any
results with the other tests, except with benzidine.

One great advantage of the method is that only the very smallest
fragment of stain is necessary. In fact it is essential for the production of
good results that only a small amount of blood be taken. But it has also
been found rather essential that the blood should not be very dilute. The
smallest scraping from a solid article, or a shred of stained material about
a millimetre long, is quite sufficient.

In using fresh blood it is advisable to dry the blood first.

SUMMARY.

The benzidine test for blood stains is merely a preliminary test as it 1s
not characteristic.

The spectrum and corpuscle tests fail in many cases, as does the ordinary
method of using Teichmann’s test. The use of a mixture of sodium iodide
and lactic acid is recommended for the latter, in place of sodium chloride and
acetic acid.

LVII. ON THE BEHAVIOUR OF AMYLASE IN
iie PRESENCE OF A SPECIFIC PRECIPITATE:

By AGNES ELLEN PORTER,

Lister Research Scholar, Lister Institute.

(Received Nov. 10th, 1913.)

The complement-binding phenomenon which takes place in the presence
of antigen and antibody has become of great practical importance in the
recognition of disease. The difficulties of the test are great, owing to the
fact that five different, chemically impure, and little known substances must
be employed: antigen, antiserum, complement, amboceptor, and red corpuscles.
The last three components are very troublesome to obtain, and it is not
surprising that many efforts should be made to minimise this trouble, and
to lessen the difficulties which are necessarily dependent on such a
complicated process. The most attractive simplification would be one in
which all these three components of the haemolytic system, viz. complement,
amboceptor, and red corpuscles, are entirely done away with. This would
be possible if some substance could be found which would become fixed, like
complement, by the antigen-antibody mixture, but whose absence could be
tested for more easily. x

Like complement, animal ferments seem to be very rapidly absorbed by
coagulated colloids and even by organic precipitates of different kinds, The
ferment shows some selective action in this tendency. Pepsin absorbed into
egg-white [Dauwe, 1905, p. 426] or pepsin, ptyalin, etc. into collodion
[Porter, 1910, p. 382] can be recovered in some degree by a solution of the
particular substance which they digest. As ferments can only be recognised
through their effects, this possibility of partial recovery presents the chief
objection to the use of ferments as substitutes for complement.

Most ferments are powerfully influenced by the presence of serum. — Pro-
teolytic ferments are inhibited by serum, probably because their attraction
to the serum proteins, even in a dissolved condition, is already so great. On
the other.-hand other ferments, such as starch-splitting ferments and lipases,
are markedly accelerated by the presence of serum.

39—2

600 A. E: PORTER

Hailer [1908, p. 280] first made the attempt of absorbing a ferment
by means of a specific precipitate. He chose rennet, with sheep’s serum
as antigen, and the serum of a rabbit immunised against sheep’s protein as
antibody. This choice was perhaps unfortunate, as not only rennet but also
antirennet have been described in serum [Fuld and Spiro, 1900, p. 141]. Also,
as I have pointed out, rennet suffers very little from dilution, when the
dilution has been freshly made. Rennet has indeed acted perfectly in my
hands [1911, p. 394] at a dilution of 1/340,000. It can therefore be almost
entirely absorbed without any evidence of the fact. On this account I thought
it advisable to attempt the absorption with another ferment. Pepsin and
trypsin are inactivated by serum; I therefore chose amylase. Szumowski
[1898, p. 162] has proved that amylase is absorbed by fibrin. It is much
less easily absorbed by serum proteins in their natural state, as serum is
itself amylolytic. The advantages of the ferment were therefore as follows :

(1) It is present in serum, so that, were the absorption successful, no
further addition of ferment to the antiserum would be required.

(2) It is apparently not absorbed by serum proteins in their normal
state, ie. before the union of antigen and antibody.

(8) It loses rapidly on dilution, so that even a small loss could be
measured,

METHOD.

Ferment, guinea-pig-serum, saliva, and taka-diastase.

Antibody, serum of rabbits immunised against egg-white or against
horse-serum.

Antigen, egg-white or horse-serum.

The ferment, in as small bulk as possible, 1.e. 0°05 cc. saliva, or 0°15 ce.
diluted taka-diastase, was added to 0°15 cc. immune or normal serum, and
0:15 cc. of the various dilutions of egg-white or horse-serum, the mixture
being left at room temperature overnight. Next morning 3 cc. of a 1 °/
soluble starch solution were added, and the tubes were placed at 37° for
5 to 20 minutes, or in the case of guinea-pig-serum for a half to one and
a half hours. The mixtures were then tested for sugar by means of Fehling’s
solution, and for dextrins by means of iodine. On account of the protein
present, use was made of the dialysed iron method, which had the double
advantage of removing both protein and ferment. When the mixtures were
taken from the incubator, they were made up to a bulk of 5 or 6 cc. with

physiological saline solution, and 0°5 ce. saturated salt solution, and 0:5 ce.

A. E. PORTER 601

B.P. Liquor Ferri Dialysatus were added to each and the resulting pre-
cipitate immediately filtered off. A measured amount of the perfectly clear
and inactive filtrate was then tested quantitatively for sugar. The colour
given by the iodine test in the filtrate was clear and permanent.

The difference between the effect of immune and normal serum on
amylase in the presence of antigen was very slight, but on account of the
accuracy of the method and the number of times the experiment has been
repeated, if may be taken as the expression of a genuine though very partial
absorption.

Absorption experiment.

A. Immune serum. Rabbit immunised against egg-white.

(To each tube were added 0:05 ce. saliva, 0-15 ce. anti-egg-serum, 0°15 cc. diluted egg-white,
method as above. )

Dilution of Time at 37° Ce. of Iodine solution
egg-white with starch Fehling sol. 0°5 ee,
1/10 10 minutes 0:2 Dark mauve.
1/50 aC 0°225
1/100 0°225 e
1/250 Ar 0°25 oa
1/500 “5 0°25 Light mauve.
NaCl 35 0:3 Pale pink.
1/10 20 minutes 0°25 Dark mauve.
1/50 = 0°25 Medium mauve.
1/100 0°25 x a
1/250 0-275 Medium light manve.
1/500 0°3 Light mauve.
NaCl 0°35 Very pale.

B. Normal rabbit-serum.

(To each tube were added 0:05 ce. saliva, 0°15 ec. normal serum, 0°15 ce. diluted egg-white.
method as above.)

1/10 to 1/500 10 minutes 0°3 All pale.
NaCl si 0°3 Pale.
1/10 to 1/500 20 minutes 0°375
NaCl ne 0°375

be]

When serum is used as antigen a curious phenomenon may be noticed.
In this case all three ingredients, antigen, antibody, and saliva, contain
amylase from three separate species, and this gives rise to an acceleration
which is decidedly beyond the sum of the three separate activities. Although
the antiserum was heated at 56° to reduce its amylolytic power, it was
still able to exert an accelerating influence. In the following experiment,
where the antigen is horse-serum, acceleration is to be seen and must be
allowed for. In spite of it, absorption can be observed if the “immune ”
and “normal” columns are compared.

602 ! A. E. PORTER

A. Rabbit-serum, immune to horse-serum.
(To each tube were added 0-05 cc. saliva, 0°15 ce. anti-horse-serum, 0-15 cc. diluted hoe

serum, method as above.)

Dilution of Time at 37° Ces of Colouration with
horse-serum with starch Fehling sol. iodine solution
1/10 10 minutes 0-5 Red mauve.
1/50 oe 0°45 Pe
1/100 = 0-4 3
1/250 Be 0-4 Same, but deeper.
1/500 Py 0°35 Mauve.
NaCl 3 0°35 ~
1/10 20 minutes 05 Pale pink.
1/50 : 0-5 -
1/100 5. 0-45 os
1/250 os 0:4 | Mauve.
1/500 5S 0°35 =f
NaCl = 0°35

B. Normal rabbit-serum.
(To each tube were added 0:05 cc. saliva, 0°15 cc. normal rabbit-serum, 0°15 cc. diluted

horse-serum, method as above.)
”

1/10 10 minutes 0°55 Colourless.
1/50 Bcd 0°55 rs
1/100 35 0°5 3
1/250 nD 0-4 Pale.
1/500 55 0°35 Mauve.
NaCl 35 0°35 35

1/10 20 minutes 0-6 Colourless.
1/50 5 0°6 +
1/100 - 0°55 bs
1/250 4 0:45 “s
1/500 ss 0-4 Pale.

NaCl © % 0°35 Mauve.

Taka-diastase displayed no tendency whatsoever to become absorbed by
a specific precipitate. This is interesting, as I have noticed before [1910,
p. 386] that while ptyalin was inactivated by collodion membranes in a day,
taka-diastase was hardly affected in a month’s time:

BEHAVIOUR OF SERUM AMYLASE ON PRECIPITATION WITH
CARBON DIOXIDE.

Fresh guinea-pig-serum, diluted to 1/10 in water, was saturated with
carbon dioxide, and the globulins separated and redissolved in saline solution,
after centrifuging. The upper fluid taken from the centrifuge had its
isotonicity restored by sodium chloride, and both portions were tested for
amylase. The ferment was found to be almost unaffected by the precipi-
tation, practically the whole ferment remaining free in the upper fluid.

A. EK. PORTER 603

Ce, of Colouration
Time 14 hours Fehling sol. with iodine
1 ce. serum 1/10 +1 ec. NaCl 0°85 °/, 0°5 Brown.
1 ce, precipitate (1/10) +1 ec. NaCl 0-85 °/, 0°05 Dull, dark.
lec. upper fluid +1 ec. NaCl 0:85 °/, 0-4 Brown.
1 ce. serum 1/10 +1 ce. precipitate 0°475 3
1 ce. serum 1/10+1 ce. upper fluid 0°65 Colourless.
1 ce. precipitate +1 ce. upper fluid 0-425 Brown.

Although these experiments have been attended with little success, I am
venturing to record them as of some small theoretic interest, and partly to
recommend a good method for testing amylolytic action in the presence

of serum.

REFERENCES.

Dauwe (1905), Beitriige, 6, 426.

Fuld and Spiro (1900), Zeitsch. Physiol. 31, 132.
Hailer (1908), Arbeit. kaiserl. Gesund.-Amt. 29, 277.
Porter (1910), Quart. J. exp. Physiol. 3, 375.
(1911), J. Physiol. 42, 389.

Szumowski (1898), Arch. Physiol. 5° série, 160.

LViliy “FHE GALACTOSIDES OF “THE GRAIN at
By OTTO ROSENHEIM.
From the Physiological Laboratory, King’s College, London.
(Received Nov. 10th, 1913.)

Nearly forty years ago Thudichum [1874] recognised the existence in
the brain of substances analogous in their constitution to the vegetable
glucosides. He distinguished two principal representatives of this class
which he called phrenosin and kerasin. Phrenosin being relatively easy
to prepare in sufficient amount for further study, Thudichum succeeded in
elucidating its constitution by an investigation of its hydrolytic cleavage
products. The carbohydrate obtained on hydrolysis of phrenosin was
originally called “cerebrose” by its discoverer, but was identified later
with galactose [Thierfelder, 1890; Brown and Morris, 1890]. The group
name “cerebroside,” in analogy to “glucoside,” was introduced by Thudichum
on the assumption that cerebrose was the typical carbohydrate of the brain,
but later on he himself adopted the name cerebro-galactoside. In view of
the fact that the occurrence of these substances is not limited to the brain’,

ce

the use of the general name “galactoside” seems to be preferable to the

c

special term “ cerebroside.”

The work of subsequent investigators was limited, until quite recently,
to the preparation by modified methods of the substances first obtained by
Thudichum, without contributing to the knowledge of their cleavage products.
Unfortunately a great confusion in the nomenclature of these substances
has resulted from the failure of earlier workers to co-ordinate their results
with each other and with those of Thudichum. It seems that too much
importance was attached to minor differences in the figures of elementary
analysis and to variations in the melting points of these substances of high
molecular weight ; it was mainly on this basis that new names like pseudo-
cerebrin, cerebrin, homocerebrin, cerebron were introduced for substances,
which from the method of their preparation were nothing else than
Thudichum’s original phrenosin and kerasin, or mixtures of the two.

' Similar substances have since been obtained from pus (Kossel and Freitag), from blood
(Bang and Forssmann), from adrenals (Rosenheim and Tebb), and also from certain mushrooms
(Zellner), ete.

ar

O. ROSENHEIM 605

In the course of a prolonged investigation of these substances, carried
out during the last six years, the author has been able to confirm Thudichum’s
main observations and has worked out a new method for the preparation
of the galactosides from brain. The complete separation of the galactoside
mixture into its two principal constituents, phrenosin and kerasin, was made
possible for the first time by the help of a new physical test, depending on
the behaviour of these substances under the polarising microscope, which
will be described later.

The investigation of the optical activity of the two substances brought
out an interesting difference between them. Whilst phrenosin, in confirma-
tion of the previous observations of Thierfelder and Kitagawa [1906], proved
to be dextro-rotatory, it was found that kerasin, when completely separated
from all dextro-rotatory phrenosin, possessed a laevo-rotation. Levene and
Jacobs [1912, 2] and subsequently Thierfelder [1913] have since described
as inactive a “kerasin fraction” prepared by different methods and evidently
not completely freed from phrenosin.

The two products were further subjected to hydrolysis and all the
cleavage products were identified. As far as phrenosin is concerned, the
results agreed with Thudichum’s observations which had already been
confirmed in the main outlines by Thierfelder’s work [1904, 1905].

The complete hydrolysis of kerasin had hitherto not been carried out
and its relationship to phrenosin had remained unexplained. In its
elementary composition it agrees closely with phrenosin and yet it differs
widely from it in appearance, solubility and other physical constants.
Thudichum suggested that the essential difference between it and phrenosin
might be due to a difference in the fatty acid radicle. Neither the fatty
acid nor the carbohydrate nor the base had been isolated or identified by
Thudichum. Levene and Jacobs [1912, 2] recently advanced the hypothesis
that the two substances represented optical isomers.

My own results show that Thudichum’s suggestion was correct and that
kerasin contains a different fatty acid from phrenosin, On hydrolysis of
kerasin, purified as far as possible, I obtained as the only fatty acid one of
the composition C.,H,.O., which I was able to identify with lignocerie acid
[1913]. The carbohydrate was found identical with d-galactose and the base
with sphingosine, as was to be expected from Thudichum’s preliminary work.

Phrenosin, on the other hand, gives rise to the optically active hydroxy-
acid C,;H,,O;, as well as to d-galactose and sphingosine’.

1 Thierfelder [1904] was the first to show that the fatty acid obtained by Thudichum from
phrenosin represented a hydroxy-acid. Levene and Jacobs [1912, 1] satisfactorily explained the

606 O. ROSENHEIM

In connection with this work the question was investigated in which
form the galactosides occur in the brain, and it is proposed to deal with this
fundamental question first, leaving the description of the methods used for
the separation of phrenosin and kerasin, as well as for their purification and
hydrolysis, for a subsequent communication.

«tae GALACTOSIDES EXIST IN THE BRAIN IN THE FREE STATE.

Previous investigations leave the question undecided whether the galacto-
sides occur preformed in the brain or whether they are split off from more
complex substances during the process of preparation. Whilst some recent
authors incline to the view that part at least exist preformed, others are
of the opinion that, as only small quantities are obtainable directly from
the brain without drastic methods, they owe their origin to the partial
decomposition of a preformed complex substance. No experimental proof
for either view has been brought forward.

It is evident that this question cannot be decided by the use of those
methods hitherto employed for the preparation of the galactosides, in which
either baryta is used for the removal of the phosphatides, or in which
solvents like alcohol are used at their boiling temperature. In both cases
a decomposition of a preformed combination is possible. It seemed feasible,
however, to obtain conclusive evidence on this point by the use of an inert
solvent which dissolves galactosides in the cold, and in which phosphatides
and sulphatides are insoluble. Such a solvent was found by Rosenheim and
Tebb [1910] in pyridine.

In applying the pyridine method directly to brain, it is preferable
previously to remove water and cholesterol by means of cold acetone
[ Rosenheim, 1906], and the unsaturated phosphatides, lecithin and kephalin,
by means of ether or petroleum ether. As the same method was used —
subsequently for the isolation of the galactosides, the procedure adopted
may be described here in detail.

differences in melting point of Thudichum’s and Thierfelder’s acid by showing that the acid
existed in two isomeric modifications, of which the optically ‘inactive one melts at 82-85°
(Thudichum’s ‘‘neurostearic’’ acid), whilst the dextro-rotatory form melts at 106-108° (Thier-
felder’s ‘‘ cerebronic”’ acid), From a ‘‘kerasin fraction” Thierfelder [1913] has recently obtained
an acid of the composition Cy,H4s0.; Levene [1913] has since identified a similar acid as
lignoceric acid, which he isolated on hydrolysis of a ‘‘cerebrin fraction.” Both Thierfelder’s
and Levene’s fractions were optically inactive and contained therefore probably dextro-rotatory
phrenosin admixed with the laevo-rotatory kerasin. Thierfelder, indeed, obtained 6 °/, of
cerebronic acid, the typical acid of phrenosin, on hydrolysis of his ‘‘kerasin fraction.” The
carbohydrate had not been identified by either of these workers.

O. ROSENHEIM 607

10 kg. of finely minced ox brain were suspended in 10 litres of acetone
and allowed to stand with frequent stirring for 24 hours at room temperature,
The watery acetone extract was decanted and the brain pulp strained through
several layers of fine muslin. At least six subsequent extractions with
sufticient acetone to cover the tissue were made, until the last extract on
evaporation yielded only an inappreciable amount of cholesterol. The total
quantity of crude cholesterol obtained from these extracts varies between
240-260 g., i.e. about 2°5 °/, of the fresh brain.

The tissue was next spread in a thin layer on large glass plates, gently
warmed from below, and freed from acetone by means of an air current
(electric fan).

The dry, somewhat waxy powder was now subjected to extraction with
cold petroleum ether, Five to six extractions were usually found sufficient
to remove the unsaturated phosphatides, which may be obtained from the
extracts by the usual methods. After the removal of the petroleum ether
by an air current in the way described above, the tissue was passed through
an Excelsior mill and was thus obtained as a fine cream-coloured powder,
ready for the pyridine treatment. 10 kg. of fresh brain yield on the average
1300 g. (from 1200-1400 g.) of this powder, which seems to keep indefinitely
in this condition.

For the preparation of the galactosides it was found convenient to work
up 500 g. of this powder at a time. This quantity was covered with 1500 ce.
of pyridine (B.P. 115°) and after having been warmed to 45° by being kept
for about twenty minutes in a water-bath at 50°, it was rapidly cooled to
room temperature. Although the galactosides are readily soluble in cold
pyridine, the initial warming is necessary in order to allow the solvent to
penetrate the tissue. Filtration proceeds easily by means of a large Buchner
funnel. From their pyridine solution the galactosides were obtained by
pouring it into 3-4 volumes of acetone. A bulky white precipitate is
formed, from which the supernatant fluid can be easily decanted off after
standing some time. The mother liquor deposits still further on cooling on ice.
In the earlier experiments this deposit was filtered off separately, but as the
amount was found to be very small, the pyridine-acetone mixture was cooled
down directly to 0° in the later experiments.

The filtration and washing of the mixed galactosides under pressure
proceeds very slowly and it is therefore advisable to filter through a plain
filter. After thorough washing with acetone, the prec:pitate is suspended in
acetone and only then filtered under pressure. After being dried in vacuo,

1 The pyridine solution may, of course; be previously concentrated by distillation in vacuo.

608 O. ROSENHEIM

the powder is extracted with ether in a Soxhlet, in order to remove the last
traces of ether-soluble phosphatides.

The crude galactosides are thus obtained as a slightly yellowish powder,
the average yield amounting after two extractions to 205 g. from 10 kg. of
moist brain, i.e. 2 per cent.

It is of interest to compare this experimental yield with the theoretical.
Unfortunately there is at present a lack of a reliable quantitative method for
the estimation of galactosides and the data available in the literature as to
the percentage of galactosides in normal brain are correspondingly scanty.
Probably the method worked out recently by Lorrain Smith and Mair [1913],
in which the galactosides are weighed as such, gives truer results than the
older methods in which the reducing power of hydrolysed brain extracts
served as a basis of calculation. Lorrain Smith and Mair state their results
in percentages of the chloroform extract of dried brain. Taking the mean
water content of brain as 78°/,, I have calculated from their results the
percentages of galactosides as 7°3°/, in dry and 1°6 °/, in fresh normal human
brain. This result agrees well with the experimental yield of 9°8 °/, in dry
and 2°/, in fresh brain, as stated above, if we consider that these figure refer
to the crude product.

After having been recrystallised twice from 15 volumes of an alcohol-
chloroform mixture (1:2) the substance was obtained as a white powder,

which contained 1:68°/, of nitrogen and only a very small percentage of
phosphorus (0:08 °/,).

0:3000 g.; 2°36 ec. N/10 KOH by Neumann’s method.

0°4893 g.; 5°86 cc. N/10 KOH by Kjeldahl’s method.
The product represents, as will be shown in a subsequent communication,
a mixture of at least two substances. For the purpose of the present investi-
gation it was considered sufficient to show that it consists essentially of

galactosides, and the substance was therefore subjected to hydrolysis without
any further purification.

Method of estimation of the products of hydrolysis.

Hydrolysis was carried out in methyl alcohol solution with sulphuric acid,
and the cleavage products estimated in a way similar to that described by
Thierfelder [1905] for “cerebron.” Galactose was estimated polarimetrically

and the fatty acids (as esters) and bases (as sulphates) were collected as
carefully as possible and weighed.

1 g. of the substance was dissolved in 50 cc. methyl alcohol containing

QO. ROSENHEIM 609

5 ce. concentrated sulphuric acid and boiled under a reflux condenser on a
water bath for six hours. After standing overnight, white glittering scales
of the esters (and free fatty acids) had crystallised out. The flask was kept in
the ice-chest for some hours and the crystals were filtered off, washed with
cold methyl alcohol, dried in vacuo and weighed. To the filtrate water was
added, and the clear solution boiled for some time in order to hydrolyse the
methyl galactoside formed during hydrolysis. The alcohol was evaporated
on the water bath. During the concentration of the watery solution, oily
droplets of the sulphates of the bases appeared, which solidified on cooling.
The deposit was filtered, taken up in boiling alcohol and the solution
evaporated in vacuo. The residue consisting of the sulphates of sphingosine
(and dimethylsphingosine) was weighed after having been dried to constant
weight in vacuo. The final filtrate was made up to 100 cc. and examined
polarimetrically.

Previous experiments had shown us that the results of galactose estima-
tions by the polarimeter agreed with those made by Kjeldahl’s gravimetric
method.

Experiment I. 1g. substance gave 0°303 g. esters and 0620 g. sulphates
of bases. Galactose solution: actual rotation measured = + 0°33° (mean of
six readings) in 2 dm. tube. Whence galactose = 20°39 °/, [see Landolt, 1892,
p. 452].

Experiment II. 1 g. substance of a ditterent preparation gave 0°280 g.
esters, Galactose solution : actual rotation measured = + 0°34° in 2 dm. tube.
Whence galactose = 21:01 °/).

d-Galactose was identified in the final solution by means of methylphenyl-
hydrazine. The free mineral acid was neutralised with solid sodium acetate
and methylphenylhydrazine was added. After standing at room temperature
for some hours, crystals of the hydrazone settled down. They were filtered
off and recrystallised from alcohol. The white crystals melted sharply at 191°.
The melting point of galactose methylphenylhydrazone is given by Neuberg
[1907] as 191° [see also Frinkel, 1910]. From another portion of the
hydrolysate, galactose was prepared as such by the usual methods and
identified as mucic acid by oxidation with nitric acid. The dry ammonium
salt of the mucic acid thus prepared gave on heating a strong pyrrole reaction.

The results are given in the following table (page 610), which shows also
for comparison the figures obtained by a similar method by Thierfelder from
the hydrolysis of more or less purified galactosides,

It will be seen from these results that the cleavage products obtained on
hydrolysis of the galactoside mixture show a close resemblance in their nature,
as well as in their quantitative distribution, with those obtainable from the

610 0. ROSENHEIM

Galactoside mixture prepared
by the pyridine method

(a Ee eS Phrenosin! Kerasin
I II (cerebron) fraction
Nitrogen °/, 1:68 — 1°76 1-61
Galactose °/, 20°39 21°01 19-88 19°35
Esters+ fatty acid °/, 30°3 28°0 38°7 29°6
Bases (as sulphates) 62:0 o- 50°3 d7°2

1 The figures given represent the mean values calculated by me from Thierfeldev’s figures.

galactosides isolated by a different method. This result therefore seems to
justify the conclusion that the product obtainable from brain by extraction
with cold pyridine, consists practically entirely of galactosides. In their
preparation any possible decomposition of a preformed complex substance
has been carefully avoided by limiting the time of extraction to a minimum
and by employing an inert solvent at a low temperature. As, further, the
amount obtainable experimentally agrees very well with the theoretical
yield, we must assume that» the whole of the galactosides exist in the brain
in the preformed state.

The expenses of this research have been in part defrayed from a grant
from the Government Grant Committee of the Royal Society.

SUMMARY.

(1) A new method for the preparation of galactosides from brain by
means of pyridine is described.

(2) Evidence is brought forward to show that the galactosides exist in
the brain entirely in the preformed condition.

REFERENCES.

Brown, Horace T. and Morris, G. H. (1890), J. Chem. Soc. 57, 57.
Friinkel, S. (1910), Biochem. Zeitsch. 26, 41.
Landolt, H. (1892), Das Optische Drehungsvermigen, Braunschweig.
Levene, P. A. and Jacobs, W. A. (1912, 1), J. Biol. Chem. 12, 381.
(1912, 2), J. Biol. Chem. 12, 389.
(1913), J. Biol. Chem. 15, 359.
Neuberg, C. (1907), Biochem. Zeitsch. 3, 531.
Rosenheim, O. and Tebb, M. C. (1910), J. Physiol. 40, Proceedings, i
—— (1906), J. Physiol. 34, 104.
—— (1913), Trans. Internat. Congress of Med. Sect. 1. 626.
Smith, J. Lorrain and Mair, W. (1913), J. Path. Bact. 17, 609.
Thierfelder, H. (1890), Zeitsch. physiol. Chem. 14, 209.
— (1904), Zeitsch. physiol. Chem. 43, 21.
—— (1905), Zeitsch. physiol. Chem, 44, 366.
and Kitagawa, F. (1906), Zeitsch. physiol. Chem. 49, 288.
—— (1913), Zeitsch. physiol. Chem. 85, 35.
Thudichum, L. J. W. (1874), Reports of the Med. Off. of the Privy Council and Local
Gut. Board, New Series, No. III.
(1882), J. pr. Chem. 25, 19.
—— (1901), Chem. Konstitution d. Gehirns d. Mensch. u. d. Tiere, Tiibingen.

LIX. THE ESTIMATION OF PYRUVIC ACID.
By IDA SMEDLEY MacLEAN, Beit Memorial Research Fellow.

From the Biochemical Department, Lister Institute of
Preventive Medicine.

(Received Nov. 11th, 1913.)

Pyruvic acid forms with phenylhydrazine a hydrazone by means of which,
as Emil Fischer pointed out, one part of this acid in one thousand parts
of water may be detected. The yield of hydrazone however is not quanti-
tative and the attempts made by some authors to estimate pyruvic acid by
weighing the phenylhydrazone precipitated have not proved satisfactory.
Subsequently nitrophenylhydrazine was used as the precipitating reagent
and by this means Neuberg and Karezag [1911] recovered 92°/, of pyruvic
acid from a 1°/, solution, as the nitrophenylhydrazone. In estimating the
amount of pyruvic acid in solutions containing 0°1°/, and of still lower
concentrations this method is however of very little value, since the error
introduced by the appreciable solubility of the hydrazone becomes of in-
creasing importance with increasing dilution. An investigation into the
action of the tissues on dilute solutions of pyruvic acid which I had under-
taken had temporarily to be abandoned since the apparent removal of the
acid observed might have been explained by an increase in the amount of
hydrazone held in solution, and even when controls of corresponding dilution
were used the results were unsatisfactory.

The action of asymmetrical diphenylhydrazine was investigated and was
found to present similar difficulties. .

If the experimental errors found above were due to the solubility of the
hydrazone, the determination of the amount of phenylhydrazine removed
from the solution in combination as hydrazone should give satisfactory results.

Estimation of Phenylhydrazine.

Fischer [1878] showed that phenylhydrazine was oxidised by cold dilute
Fehling’s solution with evolution of nitrogen; benzene and aniline were

612 I. 8S. MacLEAN

formed and cuprous oxide precipitated. Strache and Kitt [1892] estimated

the volume of nitrogen liberated and showed that if boiling solutions were

used, no aniline was formed and the whole of the nitrogen was liberated in

the free state; under these conditions six molecules of cupric oxide were

necessary to oxidise two molecules of phenylhydrazine, a mixture of benzene

and phenol being obtained. ‘The reaction may be represented as follows:
2C,H;. HN. NH,+ 80 =C,H,+ C,H,OH + 2N, + 2H.0.

The benzene formed during the reaction exerts an appreciable influence
on the vapour tension, a difficulty which Strache overcame by saturating the
gas both with benzene and with water vapour and introducing the necessary
corrections. Strache [1891, 1892] estimated ketones and aldehydes by allowing
warm solutions of the carbonyl compound and phenylhydrazine to react and
then measuring the excess of phenylhydrazine in the solution by determining
the volume of nitrogen evolved when oxidised by boiling Fehling’s solution.

The method is not very convenient and it would be preferable to estimate
the cuprous oxide formed. By the above method, however, in working with
tissue-extracts containing pyruvic acid, any sugar present would react with
the boiling Fehling’s solution. If, however, the cuprous oxide formed when
the excess of phenylhydrazine reacts with Fehling’s solution at air temperature
be estimated, this difficulty can be obviated.

Experiments were therefore made in order to determine whether the
amount of cuprous oxide precipitated by a certain weight of phenylhydrazine
was constant.

The Fehling’s solution was made up as in Bertrand’s method for esti-
mating glucose.

Solution I. Copper Sulphate crystals 40 grams per litre.
Solution II. Rochelle Salt 200 grams |

: é yer litre.
Caustic Soda 150 ~— ,, { |

5 ce. of a solution of phenylhydrazine containing 3'5236 g. in 100 cc. of
50°/, acetic acid were diluted to 100 ce. with water, and allowed to stand
at the ordinary temperature for 30 minutes: in nine experiments, 20 ce. of
each of the Fehling’s solutions, I and II made up as above, were added
to 10 ce. of the diluted phenylhydrazine solution and the mixture allowed
to stand at the ordinary temperature for times varying from half an hour
to four hours and a half. The cuprous oxide formed was then filtered
through a Gooch crucible, dissolved in ferric sulphate solution as in Bertrand’s
method for the estimation of glucose and the ferrous sulphate produced

titrated with deci-normal permanganate solution.

:

\

I. S. MacLEAN 613

Results.
Ce. N/10 KMnO,
eqtuvalent to
Time Cu,O formed
30 minutes 6°0
60 ra 6°05
90 ne 6-0
150 + 6°05
150 PE 6°0
180 ey 6-0
210 ¥s 6°0
240 i) 5°95
270 Be 595

The reaction between phenylhydrazine and Fehling’s solution appears
therefore to reach a definite stage of equilibrium within half an hour at the
ordinary temperature, after which no further oxidation proceeds.

In three experiments, the following values were obtained :

1 ce. of N/10 KMnO, was equivalent to (1) 0°002962 gr. phenylhydrazine,
(2) 0002962
(3) 0°002935
As the mean of these experiments therefore
1 cc. N/10 KMnO, is equivalent to 0:00295 g. C,H;NHJ|NH,.

It may be mentioned that in filtering the cuprous oxide from the cold
Fehling’s solution, it is advisable to filter only under a slight difference of
pressure, as there is a tendency for the cuprous oxide to pass through the
asbestos. If this happens the filtrate should be filtered through a clean
Gooch crucible and the two results added together.

”»

»

Ystimation of Pyruvic Acid.

A specimen of Kahlbaum’s pyruvic acid was distilled under diminished
pressure and the fraction boiling at 77°-78° under a pressure of 15-20 mm.
used for the estimation. The method adopted was as follows:

A solution of pyruvic acid was made up containing 1°5228 g. per 100 ce.
Quantities of from 2 to 10 cc. of this solution were diluted to about 80 cc.,
5 ce. of a solution of phenylhydrazine, approximately 4°/,, added, and the
mixture made up to 100 cc. and allowed to stand half an hour at the
ordinary temperature ; 5 cc. of the hydrazine solution were diluted to 100 ce,
and allowed to stand for the same time. After half an hour the pyruvic
hydrazone which had separated was filtered off and 10 cc. of each filtrate

Bioch. vir 40

614 I. S. MacLEAN

added to 40 cc. of Fehling’s solution. Ten ce. of the control phenyl-
hydrazine solution were similarly treated. The cuprous oxide was estimated
as above described.

Thus:
10 cc. phenylhydrazine solution require 6:00 N/10 KMnO,
6-00 99 9
10 cc. hydrazone filtrate require 600 Dar vas -
2°75 99 29

3°25 ce, N/10 KMn0, are equivalent to 3°25 x 0:00295 g. phenylhydrazine.

C,H;HN . NH» + CH;. CO. COOH+CH;. C (NyHCgH;) . COOH.
108 88
88 x 0°00295 x 3°25

108
diluted pyruvie solution contained 0-00780 g. pyruvic acid, and the original solution contained
1°556 9).

10 ce. of the diluted solution contained by weight 0-007614 g. and the original solution
1-5228 97).

3°25 ec. N/10 KMn0O, are equivalent to g. pyruvic acid, and 10 cc. of the

The following table shows some of the results obtained :

Ten ce. of a solution of pyruvic acid contained :

(b) By above method

(a) By weight of estimation Error
2°68 milligrams 2°57 milligrams - 0°11 mer.
eyilg) 2°15 —1:04
3°57 2°51 — 1:06
5°36 5-14 — 0:22
6°28 6°57 +0°29
7:14 7:46 + 0°32
fol” 7°76 +0°15
8-04 Corn — 0°33
9°42 9-80 +0°38
9°58 9:20 +0°38
10°71 11°22 +051
10°71 10°34 — 0°37
SO 12°31 + 0°34
12°77 12°67 — 0:10
13°40 13°15 — 0°25

In carrying out the above estimation it is important that the phenyl-
hydrazine solution shall be freshly made up and if it is at all discoloured that
the phenylhydrazine shall be freshly distilled. |

Influence of Glucose.

Under the conditions above described the presence of glucose does not
appear to interfere with the estimation of pyruvic acid. In one experiment,

I. S. MacLEAN 615

5 ce. of a solution of phenylhydrazine acetate, 10 cc. of a solution of pyruvic
acid and 10 ce. of a 1°/, glucose solution were made up to 100 ce. and the
pyruvic acid estimated as above, and compared with a solution similarly made
up but from which the glucose was omitted.

Residual hydrazine (without glucose) required 4°80 cc. N/10 KMnO,.
ap 3 (with 2s) 4°85

The method therefore gives satisfactory results in estimating solutions
of concentrations above 0:03°/,. Below this concentration probably more
accurate results would be obtained by using a more dilute solution of
permanganate.

The advantages may be summarised as follows:

(1) It is easily carried out; the whole estimation can be done in little
more than an hour.

(2) It gives a greater degree of accuracy than the unsatisfactory method
of gravimetric estimation at present in use.

(3) The presence of glucose does not interfere with the estimation.

(4) The method promises to be of general value for the estimation of
carbonyl compounds and also for measuring the rate of interaction of these
compounds with phenylhydrazine.

REFERENCES.

Fischer, E. (1878), Annalen, 190, 101.

Neuberg and Karezag (1911), Biochem. Zeitsch. 36, 63.
Strache (1891; 1892), Monatsh. 12, 524; 13, 299.
— and Kitt (1892), Monatsh. 13, 316.

40—2

LX. “NOTE: ON ISOCHOLESTEROL COERG-
STEROL AND THE CLASSIFICATION OF THE

STEROLS.
By CHARLES DOREE.

(Received Nov. 9th, 1913.)

Isocholesterol was discovered by Schulze [1873] in the fat extracted
from raw sheep’s wool with ether. The fat is derived from the secretion
of the fatty glands surrounding the hair follicle in the skin. For many years
the individuality, and even the existence of isocholesterol, were questioned,
as many workers, including the present writer, were unable to obtain it.
In reply to criticisms of Darmstidter and Lifschiitz [1898], Schulze [1898]
repeated his original experiments and subsequently Moreschi [1910] obtained
the substance from wool fat and fully confirmed the observations of Schulze.
In 1910 additional interest was given to the question by the discovery of
isocholesterol in a plant product, the so-called South African rubber, which
consists of the coagulated latex of various species of Euphorbiaciae. The
latex contains about 6 per cent. of rubber and 70 per cent, of resm. The
isocholesterol was found as a constituent of the latter (excretory) product
[Cohen, 1908]. A careful comparison of the substance with a specimen of
the original isocholesterol supplied by Schulze was made by Cohen, and in
my opinion no doubt can be entertained as to the identity of the two
products and consequently of the recognition of isocholesterol as a definite
member of the sterol group.

Coprosterol occurs normally in the faeces of men on ordinary diets, and
in those of animals, in place of cholesterol, when diets rich in cholesterol,
e.g. raw brain, are given [Dorée and Gardner, 1908]. It is no doubt produced
from the cholesterol of the food by changes taking place in the intestine.
Konig and Schluckebier [1908] have stated also that it appears in the
excrement of animals which have been kept on diets rich in phytosterol such
as peas, maize, coco-nut cake, etc. [Cp., however, Kusumoto, 1908; Dorée
and Gardner, 1909. ]

The recognition of isocholesterol as a definite member of the sterol group

leads me again to point out that while isocholesterol and coprosterol, in their

C. DORER 617

mode of occurrence and properties, ¢losely resemble one another, they differ
in these respects markedly from all the other sterols [ Dorée, 1909, 1]. Hitherto
the sterols have been classified roughly as zoo- or phyto-sterols according
as they were obtained from animal or vegetable sources respectively. But
in order to emphasise the speéial properties and relationships of isocholesterol
and coprosterol it is now suggested that a third class should be formed to
include them and any other similar substances that may be discovered, and
that the terms zoo- and phyto-sterols should be limited to a somewhat more
exact definition. Seeing that coprosterol and isocholesterol undoubtedly
stand in a close relationship to cholesterol and phytosterol, and that they
appear to form a connecting link between these characteristic constituents
of animal and vegetable protoplasm, they may be classified as “natural
derived sterols” or metasterols.

The metasterols occur in the excretions and secretions of animals (in-
testinal canal, glands of the skin) and of plants (resin). They are never
found as constituents of animal or vegetable protoplasm. The connection
between them and the other sterols from a biochemical standpoint is shown
in the following scheme.

CHOLESTEROL PHYTOSTEROL

L a
(inte, i <— FZ aot

Ne) ~~ _ Coprosteron—

The metasterols are monatomic alcohols of high molecular weight. Unlike
the other sterols they (a) are saturated towards bromine and hydrogen
[Moreschi, 1910], (b) do not crystallise from alcohol in plates, (c) give the
colour tests of Liebermann and Salkowski in a modified way, (d) have practi-
cally no anti-toxic action towards haemolytic poisons [Hausman, 1905],
(e) are dextro-rotatory.

It is probable that spongosterol [Henze, 1908], which has no anti-toxic
power and apparently is a saturated alcohol, will fall into this class. The
position of stigmasterol, C,,H,.O [Windaus and Hauth, 1906], and brassica-
sterol, C,,H,O [Windaus and Welsch, 1909], found, together with large
proportions of ordinary phytosterol in Calabar beans and rape seed re-
spectively, is uncertain.

With the formation of this class it is now proposed to limit the terms
zoo- and phyto-sterol to sterols which are found as tissue constituents of

618 Cc. DOREE

animals and plants respectively. Cohen [1908], in the hght of his discovery
of isocholesterol as a plant product, has stated that the distinction between
zoo- and phyto-sterols can no longer be maintained. But with the proper
recognition of isocholesterol as a derivative produced from zoo- or phyto-
sterols by metabolic changes, there is now every reason for maintaining it.
For in spite of a great number of researches (Hauth [1907] quotes some
50 papers and many more have since appeared), no substance resembling
cholesterol has ever been obtained from vegetable protoplasm or one re-
sembling phytosterol from animal protoplasm; and this in spite of the fact
that herbivorous animals take in phytosterol with their food and the
carnivorous plants ingest a cholesterol. If phytosterol taken with the food
is used by the animal in building up its tissues, it must first be converted
into cholesterol. Furthermore a zoosterol is an invariable constituent of all
animal organs and tissues so far examined from Chordata to Coelenterata
[Dorée, 1909, 1; Welsch, 1909], and a phytosterol seems similarly to be
contained in the tissues of plants of all orders. So fundamental is this
distinction considered to be that it has become an important criterion in
chemical technology for deciding the animal or vegetable origin of various
natural substances. Thus Lewkowitsch [1913] defines these products to be
of animal or vegetable origin respectively according to the presence of a z00-
or phyto-sterol in the unsaponifiable portion. [Cf also the ‘phytosterol
acetate test.’ |

The zoo- and phyto-sterols will therefore include such sterols as are found
entering into the composition of animal and vegetable protoplasm respectively.
Speaking only of those whose properties have been carefully examined (for
zoosterols see Dorée [1909, 1] and Welsch [1909]; for phytosterols see Hauth
[1907]) the zoo- and phyto-sterols may be held to include “a number of
secondary, monatomic alcohols, chiefly C.,H,O, containing one ethylene
linking in the molecule. They crystallise from alcohol in plates. Their
benzoates melt at about 145° and show the phenomena of liquid crystals.
They are all laevo-rotatory and have a powerful anti-haemolytic action towards
saponine, ete.” Their constant presence in the tissues is in part due to the
necessity for the protection of the cells from such toxic substances: their
occurrence in the fluids and secretions shows that they are continually being
metabolised and, it is believed, conserved by the organism [Dorée and
Gardner, 1909].

A summary of the properties of the best known sterols and their
derivatives is given in the following table:

|
|
.

ae 4 4

Ag ee

‘ch Saeraes

ee

. DOREE 619

(a) Zoosterols : Acetate Benzoate Dibromide
M.p. an m.p. m.p. m.p. Occurrence

Cholesterol 147° — 37° 114° 145° 123° Universally distributed in
the animal kingdom,

Bombicesterol 148° — 35° 129° 146° ‘fi b te In Bombyx mori [Menozzi
and Moreschi, 1908}.

Clionasterol 138° — 37° 133° 143° 114° In Cliona celata [Dorée,
1909, 1].

(b) Phytosterols :

Phytosterol 137° — 34°! 127° 146° 98° Universally distributed in
the Phanerogams.

(c) Metasterols :

Isocholesterol OS pes + 60° 134° 194° None Wool fat; resin of En-
phorbiaciae,
Coprosterol 100° + 24° 88° 122° None Excrement of animals.

1 In ether solution ; others in chloroform.

Of the chemical relationship existing between cholesterol, phytosterol and
the metasterols little is known. On a milk diet, during which bacterial
change in the intestine is reduced to a minimum, cholesterol [ Miiller, 1900]
and phytosterol [Windaus, 1908] are not converted to coprosterol. Normally
in the case of men, and on diets rich in sterols in the case of animals, copro-
sterol is produced. Analyses, to which however no great weight can be
assigned, seemed to indicate that the coprosterol molecule contained two
atoms of hydrogen more than that of cholesterol, and it has been thought
that coprosterol was dihydrocholesterol produced by bacterial reduction.
Dihydrocholesterol artificially produced [Willstitter and Mayer, 1908] is
not however identical with coprosterol, although very similar in its properties
as will be seen from the following table :

Acetate Benzoate Ketone ..

M.p. [a]p m.p. m.p. m.p. Obseryer
Dihydrocholesterol 142° + 29° 111° 155° 128° Willstitter.
Coprosterol 100° +24° 88° 122° 63° Dorée.
Isocholesterol 137° + 60° 134° 194° ? Cohen.

Observations on a large number of cholesterol and phytosterol derivatives
have shown that any modification of the ethylene linking causes a change in
the optical rotatory power from negative to positive and at the same time
the antihaemolytic function is abolished or reduced to a minimum [Hausman,
1905]. The presumption therefore is strong that it is the side chain
containing the double linking in cholesterol that is modified to produce
coprosterol. The change however is more than one of simple reduction.
The only other explanation that has been offered, based upon the behaviour
of a peculiar carbazole derivative of coprostanone [Dorée, 1909, 2], is that the

620 CG. DOREE

bacterial action might simultaneously bring about reduction and rearrange-
ment of the unsaturated side chain, producing two methyl groupings, thus:

CH

NcH.CH,.CH : CH, + H, DeH.cH :

“a : i We CH,
Cholesterol Coprosterol

Attempts have been made to solve the problem by studying the oxidation
of coprosterol and cholesterol in the hope of obtaining, among the products,
a derivative common to both. Coprosterol is difficult to obtain in any
quantity, and considering the similarity between coprosterol and isochole-
sterol and the fact that the raw material from which isocholesterol is obtained
is a commercial product, I decided two years ago to work up a large quantity
of wool fat in the hope of preparing sufficient isocholesterol to enable a
thorough examination of that substance to be made. - Quantities of 2 to 3
kilos of wool fat from various sources were saponified under pressure by the
method of Lewkowitsch and the mass extracted with ether in the usual way.
The unsaponifiable residue was then fractionated in alcohol, acetone, methyl
alcohol, etc. None of the fractions appeared to contain isocholesterol and
they were therefore severally benzoylated. A number of crystalline benzoates
were obtained, but none melting higher than 145°. A quantity of wool fat
treated by the method of Darmstiidter and Lifschtitz also gave a negative
result. An explanation of these results may be found in the observation
of Cohen [1908] that isocholesterol undergoes a change of the nature of —
autoxidation on keeping. The somewhat drastic treatments to which
commercial wool fat is subjected may therefore have destroyed the isochole-
sterol. The best method of obtaining it would seem to be to extract fresh
clipped raw wool with ether, or to employ the resin of South African rubber.

The expenses of these experiments were covered by a grant from the
Government Grant Committee of the Royal Society for which I desire to
express my thanks.

REFERENCES.

Cohen (1908), Arch. Pharm. 246, 518, 592.

Darmstadter and Lifschiitz (1898), Ber. 31, 98, 1126.

Dorée (1909, 1), Biochem. J. 4, 73.

(1909, 2), J. Chem. Soc. 95, 655.

—— and Gardner (1908), Proc. Roy. Soc. B. 80, 230.

— —— (1909), Proc. Roy. Soc. B, 81, 109.

Hausman (1905), Beitriige, 6, 567.

Hauth (1907), Zur Kenntnis der Phytosterine, Diss. Freiburg.
Henze (1908), Zeitsch. physiol. Chem. 55, 427.

C. DOREE 621

Kénig and Schluckebier (1908), Zeitsch. Nahr. Genussm. 15, 654.

Kusumoto (1908), Biochem. Zeitsch, 142, 407.

Lewkowitsch (1913), The Chemical Technology and Analysis of the Oils, Fats and
Waxes, 3rd Ed. Macmillan and Co.

Menozzi and Moreschi (1908), Atti R. Accad. Lincei A Fie Ab PT

Moreschi (1910), Atti R. Acead. Lincei [v], 19, ii, 53-57.

Miiller (1900), Zeitsch. physiol. Chem. 29, 129.

Schulze (1873), J. pr. Chem. 7, 169.

—— (1898), Ber. 31, 1200.

Welsch (1909), Vorkommen der Sterine in Tier. u, Pflanzenreich, Diss, Freiburg.

Willstitter and Mayer (1908), Ber. 41, 2199.

Windaus (1908), Arch. Pharm. 246, 121.

—— (1911), Die Sterine, Biochemisches Handlexikon (Abderhalden), 3, 268,

—— and Hauth (1906), Ber. 39, 4378.

—— and Welsch (1909), Ber. 42, 612.

LXI. THE HYDROLYSIS. OF .GLYCOGEN BY
DIASTATIC ENZYMES. Il’ THE INFLUENCE
OF SALTS: ON THE RATE OF HYDROLYSIS:
(Preliminary Communication.)

By ROLAND VICTOR NORRIS, Beit Memorial Research Fellow.

From the Biochemical Laboratory, the Lister Institute.

(Received Nov. 17th, 1913.)

In a recent communication [Norris, 1912] it was pointed out that samples
of glycogen prepared from different animals were hydrolysed at different
rates by pancreatic amylase. In view, however, of the fact that these
preparations contained varying amounts of salts it seemed desirable to
examine how far this might influence the rate of hydrolysis, and a quantita-
tive study has therefore been made of the action of neutral salts on diastatic
action. The results quickly showed that the experiments referred to were
in no way invalidated by the varying salt content, the latter in every case
being sufficient to produce the maximum degree of hydrolysis and not great
enough to cause any inhibition.

The experiments have however been continued in the hope of finding
some explanation of the manner in which salts exert their influence.

It has been known for some time that a dialysed amylase solution when
added to a starch solution free from salts produces but little hydrolysis, and
that on adding certain salts in small quantities the activity of the enzyme
is restored. In spite of a considerable amount of work on this point, the
explanation of this fact is not by any means clear. It is agreed that the
most active salts are those of the halogen acids. With regard to sulphates,
however, very divergent results have been obtained. Cole [1906, 1], for example,
states that sulphates accelerate the action while Griitzner [1902] maintains
that magnesium and sodium sulphate are “specific poisons” for the ferment.
In this connection however it must be pointed out that in one at least of
Cole's experiments the addition of sulphate produced no acceleration although
the concentration of the salt was similar to that used in previous experiments

in which an acceleration had been obtained. These and other divergencies

_

ran bape PEA nan

R. V. NORRIS 623

may perhaps be partly explained by one of the following reasons. Firstly,
the source and method of preparation of the enzyme has differed with nearly
every worker. Hence while some investigators have employed solutions
containing only traces of proteins, in other cases these have been present in
considerable quantity. Again, in the few cases where the action has been
followed quantitatively, the diastatic activity has been estimated either by
Roberts’ achromic point method [Roberts, 1891] or by the method of
Wohlgemuth [1908, 1], and it has been shown by Evans [1912, 1] that
neither of these is satisfactory. Finally, in some cases, the enzyme and
starch solution employed have been by no means free from salts, the controls
all showing a marked degree of hydrolysis, that is to say the observed effect
was really due to a mixture of salts and not alone to the particular salt under
investigation.

In the following experiments glycogen has been employed instead of
starch, while the enzyme has consisted of an extract of pigs’ pancreas.

Preparation of glycogen. This was obtained from dogs’ liver by Pfliiger’s
method. The crude glycogen was purified by repeated precipitation of its
solution by alcohol and was finally dialysed for a week, the last three days’
dialysis being against running distilled water.

Preparation of enzyme. This consisted of a Buchner extract of pigs’
pancreas which was dialysed for three days, in the course of which a certain
amount of protein usually separated out. The dialysed extract was then
filtered till perfectly clear and diluted from ten to twenty times with distilled
water,

Experimental methods. The following may be described as typical of the
method employed.

A 1 per cent. or 2 per cent. solution of glycogen was as a rule used and
to this was added a suitable concentration of the salt under investigation.
The mixture was then brought to a temperature of 37° in a thermostat and
the enzyme added. After 15 minutes and 30 minutes hydrolysis, 20 ce. of
the mixture were removed and the sugar immediately estimated by Bertrand’s
method, The sugar solution was added directly to the alkaline copper mixture
and hence the hydrolytic action stopped instantaneously. The strength of
enzyme used was adjusted so that the readings taken fell on the linear portion
of the hydrolysis curve. [See Evans, 1912, 2 and Norris, 1912.] The salts
employed were nearly all Kahlbaum’s “for analysis with certificate of
guarantee,”

624 . R. V. NORRIS

Effect of dialysis on glycogen hydrolysis.

As in the case of starch the result of dialysis of both enzyme and
glycogen resulted in almost complete inactivation. Similarly the hydrolytic
power was again restored by the addition of certain salts.

Influence of salts.

Under this heading neutral salts only are considered. The rate of
hydrolysis is of course greatly influenced by any change in the reaction
of the medium, but this point has been dealt with in a previous communica-
tion [ Norris, 1912].

Hydrogen ion determinations made on glycogen solutions containing
varying concentrations of sodium chloride showed that the reaction of the
medium was not altered by the presence of this salt» in the concentrations
employed, hence the accelerating effect of sodium chloride is not due to this
cause.

Sodium chloride.

A series of mixtures was made up each containing 1 per cent. glycogen
and a concentration of sodium chloride ranging from zero to 0:003 N. These
were in turn incubated with 1 cc. of a dilute enzyme preparation and
hydrolysis allowed to proceed for 15 minutes. The sugar in 20 cc. of each
solution was then estimated.

The results are shown in Fig. 1 where the abscissae represent the cc. of
0:1 N NaCl in 100 ce, of the mixture and the ordinates the cc. of KMnO, used
in the sugar estimation.

3-0

2:0

Sugar estimation
cc. KMnO, used for 20 ce. liquid.

ie) 0-5 1:0 1-5 2:0 2:5 3-0
cc. N/10 NaCl in 100 ce. reaction mixture.

Big ies

52 eet

|

AP Sa te

R. V. NORRIS 625

The results show that there is a rapid increase in the hydrolysis with
increasing NaCl content until the latter reaches a concentration of about
0002 N (0012 per cent.). This value is in close agreement with Cole’s
results with starch and ptyalin.

On further addition of salt there is at first no change in rate, but with
high concentration a slight retardation may be produced.

An investigation was next undertaken to decide whether the value of
this optimum concentration (0°002 N) would be changed by alteration in the
concentration of either glycogen or enzyme. On this point results have
been somewhat contradictory but it seems probable that the glycogen
concentration has but little influence on the amount of salt required to
produce the maximum rate of hydrolysis. The optimum salt content for
2 per cent. glycogen has usually been about 0°002 N, that is to say the
same as for 1 per cent. glycogen.

On increasing the amount of enzyme, however, a higher concentration of
NaCl is usually required. This is in agreement with the results of Cole
[1906, 1] and Starkenstein [1910]. The latter, working with starch, states
that the amount of NaCl necessary to give the maximum rate of hydrolysis
varies directly with the concentration of enzyme and has even based on this
a method for the estimation of diastase in animal organs [Starkenstein,
1912]. The results of my experiments do not point to such a simple
relationship, though as already stated more NaCl is usually required with
increased concentration of enzyme. It must be remembered that it is usually
in the enzyme that impurities such as proteins etc. will be found and the
latter may begin to play an important part when the amount of enzyme
present is not small.

The following table shows the results of one experiment in which the
concentration of enzyme was twice that usually employed.

TABLE L

Determination of NaCl optimum with 1 per cent. glycogen and
high enzyme concentration.

Percentage hydrolysis

Concentration SS SS ~
Experiment of NaCl 15 mins. 30 mins.
A 0 55 9°20
B 0:002 N 15*1 26°5
C 0-004 N 16°5 28-0
D 0-006 N 18°3 29°8
E 0-01 N 18°3 29°5

626 R. V. NORRIS

In this case therefore by doubling the concentration of enzyme the
optimum salt concentration was raised from 0:002 N to 0-006 N, that is to
say three times.

In other experiments however very much lower results were obtained
and there seems to be some unknown factor concerned. The point is still
under investigation.

Comparison of sodium chloride with other salts.

Chlorides of different metals.

Table II shows the accelerating effect of the chlorides of sodium, -
potassium, calcium, barium and magnesium.

These were added in insufficient quantity to produce the optimum rate
of hydrolysis, so that any variation in their accelerating power could be
detected.

TABLE II.

Accelerating power of various chlorides. 1 per cent. glycogen.
Concentration of salt =0:0005 N.

Percentage hydrolysis
=

RE Ee a re eS
Salt 15 mins. 30 mins.
NaCl 10°83 18°76
KCl 10°61 18:8
CaCl, 10-70 _—
BaCl, 10°61 17°6
MgCl, 10°61 18°52
Control 0 0°68 1:14

All the above salts are therefore of equal accelerating power, that is to
say the kation exerts practically no influence on the reaction. This is in
agreement with the results of Starkenstein [1910]. Cole [1906, 1] considered
that the anion accelerated while the kation depressed the action. If the
latter were correct, however, one would expect the chlorides containing a
divalent kation to be less active than sodium or potassium chloride and as
shown this is not the case. The view that the anion is the more important
factor is further strengthened by a comparison of the accelerating power of
chlorides, bromides and iodides. The results given in Table III confirm
those of previous investigators, namely, that the acceleration decreases in the
order given, the drop from bromide to iodide being much greater than that
trom chloride to bromide.

Comparison of KCl, KBr and KI.

R. V. NORRIS

TABLE III.

Percentage hydrolysis

aa Ce OO ---— +--+

Salt 15 mins. 30 mins.
KCl 10°61 18°80
KBr 8°31 15°60 ~
KI 2°65 3-70
Control 0°68 1°40

(Concentration of salt = 00005 N.)

Influence of sulphates.

The influence of three sulphates has been investigated but it has been
found that they have no accelerating power at all; these results are therefore
in agreement with those of Wohlgemuth [1908, 2] but in opposition to those
of Cole [1906, 1]. The latter however found that sulphates were much less
effective than the halogen compounds.
as one would expect if the anion were the factor concerned, that the divalent
anion would be extremely potent. The situation is complicated, however, as
pointed out by Cole, by the fact that sodium sulphate in moderate dilutions
chiefly dissociates into Na+ and NaSO, in which case the anion is
monovalent.

On the other hand it has not been found that sulphates have any de-
pressing action [cf. Griitzner, 1902], nor do they hinder the acceleration
produced by NaCl ete.

These results are of some interest,

TABLE IV.
Influence of sulphates. 1 per cent. glycogen.

Percentage hydrolysis

Concentration - —_—-- a
Salt of salt 15 mins. 30 mins

if 0 Control 1:60 5°04
Na,SO, 0-002 N 1°65 5:0
Na,SO, 0-01 N 1°55 50
MgSO, 0-001 N 1°60 —

2. MgSO, Each

+ 0-001 N 9°10 16°04

NaCl
NaCl 0-001 N 9-20 16°10

For the sake of comparison the results obtained with different salts have

been collected in Table V.

628 R. V. NORRIS
TABLE V.

‘Comparison of various salts. 1 per cent. glycogen containing 0:0005 WN salt.

Percentage hydrolysis

oe —— aa

Salt 15 mins. 30 mins. Remarks
0 (Control) 0°68 1:14
NaCl 10°83 18°76
KCl 10°61 18-8 | Halogen salts.
KBr 8°31 15°6 Both ions monovalent.
KI 2°5 3:7
CaCl, 10°70 ee Halogen salts with divalent
BaCl, 10°61 17°6 kati

1 . e ation.

MgCl, 10°61 18-5 )
KNO, 2°75 5:08 Both ions monovalent.
La(NOs) 2°75 4:94 Trivalent kation.
Na,SO, 0°55 1:0 Divalent anion.
K,S0, 0-64 1-20 . .
MgSO, 0-70 — Both ions divalent.

It will be seen from the results tabulated above that the only salts of
those tried which have a powerful accelerating action are those of the
halogen acids, although nitrates have a small influence. It is also clear that
the anion is much more concerned in the reaction than is the kation,
Further it is probable that the action of the salts is chiefly confined to the
enzyme, for it has been shown by Cole [1906, 2] that in the action of invertase,
where the substrate is not a colloidal solution, salts have again a powerful
influence. In this case, however, the action is reversed, that is to say, the
hydrolysis is retarded by chlorides.

If the function of the anion be simply to alter the charge on the enzyme,
it is difficult to. understand why sulphates have no accelerating effect, for
sulphates discharge a ferric hydroxide solution much more readily than
chlorides. The fact that the number of salts producing an acceleration is so
restricted, however, points to some other explanation.

In the meantime it seems desirable to examine separately the effect of
salts on the glycogen and enzyme respectively from the point of view of
adsorption and charge, and experiments on these lines are in progress but are
not as yet sufficiently advanced to furnish an explanation. The action of a
further series of salts is also being investigated.

SUMMARY.

(1) A dialysed glycogenase (pancreatic) solution has practically no
hydrolysing action when added to a dialysed glycogen solution.

(2) The activity of the enzyme is restored by the addition of small
quantities of certain salts.

R. V. NORRIS 629

(8) The most powerful of these are the salts of the halogen acids, the
activity diminishing in the order chlorides, bromides, iodides; nitrates have
also a slight accelerating action.

(4) Sulphates do not restore the activity of a dialysed enzyme solution,
neither are they inhibitors [ef. Griitzner, 1902].

(5) ‘The concentration of salt required to produce a maximum degree
of hydrolysis rises with increasing enzyme concentration but appears to be
independent of the glycogen concentration within the limits tried.

(6) The anion is probably the part of the salt concerned in the accelera-
tion, the nature of the kation (valency) having no influence,

REFERENCES.

Cole (1906, 1), J. Physiol. 30, 202.

—— (1906, 2), J. Physiol. 30, 281.

Evans (1912, 1), J. Physiol. 44, 220.

(1912, 2), J. Physiol. 44, 191.

Griitzner (1902), Pfliiger’s Archiv, 91, 195.
Norris (1912), Biochem. J. 7, 26.

Roberts (1891), Digestion and Diet, London, p. 68.
Starkenstein (1910), Biochem. Zeitsch. 24, 210.
—— (1912), Biochem. Zeitsch. 47, 300.
Wohlgemuth (1908, 1), Biochem. Zeitsch. 9, 1.
— (1908, 2), Biochem. Zeitsch. 9, 10.

Bioch, vu 41

LXII. THE ENZYMATIC FORMATION OF POLY-
SACCHARIDES BY YEAST PREPARATIONS.

By ARTHUR HARDEN anp WILLIAM JOHN YOUNG.

Biochemical Department, Lister Institute.

(Received Nov. 17th, 1913.)

Two reasons led to the institution of the followig experiments. In an
earlier paper [1904] the authors showed that the carbon dioxide evolved
in the alcoholic fermentation of sugars by yeast juice was not equivalent to
the sugar which disappeared from the solution, and ascribed this fact
to the production of a hydrolysable compound of low reducing power. It
was subsequently found that in alcoholic fermentation by yeast preparations
a certain amount of hexosephosphate is formed, which has a lower reducing
power than the sugar (about 75 per cent.) from which it is formed and would
therefore in part account for the phenomenon.

In the second place it appears to follow from the authors’ equations of
fermentation [1908] that in the normal fermentation both of fructose and
glucose half the sugar passes through the form of hexosephosphate, which is
then hydrolysed. Since these two hexoses appear to yield the same hexose-
phosphate it would be expected that as the fermentation proceeded fructose
and glucose alike would be partially converted into the same product of
hydrolysis, and the rotations of their solutions should therefore tend “to
approximate to each other. The exact nature of the substance produced
along with phosphoric acid by the hydrolysis of hexosephosphate in yeast
juice is not definitely known, but all the evidence points towards its being
fructose. Hence we should expect in all fermentations of glucose by
yeast or yeast preparations a progressive conversion of glucose into fructose
[compare Slator, 1911]. This however has not been observed and experi-
ments were therefore made on the subject, especially with the object of
ascertaining whether the product of hydrolysis of the hexosephosphate
underwent any secondary change, such as condensation to a polysaccharide.

The present experiments show that both from glucose and_ fructose

one or more dextrorotatory polysaccharides are produced during alcoholic

Ss

A. HARDEN AND W. J. YOUNG 631

fermentation by yeast preparations. The previous conclusion of the authors
is therefore confirmed, but it is not yet settled whether the polysaccharide
formation takes place at the expense of the glucose and fructose themselves
or occurs indirectly as the result of the action of some enzyme on the
product of hydrolysis of the hexosephosphate. Further investigations on
this point are in progress.

It is well known that living yeast forms glycogen when brought into
excess of sugar solution [see Pavy and Bywaters, 1907], and the behaviour
of yeast preparations therefore indicates that the enzymes involved in this
synthesis are probably, at least to some extent, still present and active.
The isolation of a substance having the qualitative reactions of glycogen is
a further confirmation of the observation of Cremer [1899] who found that
in yeast juice free from glycogen a substance was slowly formed in the

presence of sugar which gave the characteristic glycogen reactions.

EXPERIMENTAL.

Experiment 1, Three lots of 100 cc. maceration juice (from Schroder’s
dried Miinchener yeast) were incubated at 25° with toluene until they had
attained the temperature of the bath. To Nos. 1 and 2 were then added
25 ce. of a 40 per cent. solution of glucose and the evolution of carbon
dioxide observed. At the same time No. 3 was boiled and cooled, and 25 ce.
of the same glucose solution added.

In Nos. 1 and 2 a maximum rate of 216 cc. per 2 minutes was slowly
attained, which then rapidly diminished until in 52 minutes a constant rate
of 2°8 cc. per 2 minutes was reached. The initial high rate was due to the
presence of free phosphate in the maceration juice, which was converted
into hexosephosphate. During this period the total gas evolved was 293°6 cc.
at room temperature and pressure. After 52 minutes No. 2 was boiled,
whilst the fermentation in No. 1-was allowed to proceed for 17 hours
38 minutes, during which time 1458 ce. of CO, had been evolved. No. 1
was then boiled. The contents of all three flasks were then filtered, and
the free phosphate estimated in an aliquot portion of each. The amount
of glucose in each was determined by precipitating the proteins in aliquot
portions with Patein’s mércuric nitrate solution and estimating the glucose
by means of Pavy’s method.

The treatment with mercuric nitrate precipitates the hexosephosphate,
so that in order to determine the amount of sugar used up allowance must
be made for the quantity bound up in the form of hexosephosphate. This is

632 A. HARDEN AND W. J. YOUNG

readily done since it has given rise to the CO, equivalent to the free
phosphate present at the beginning, and can therefore be determined by
subtracting the carbonic acid corresponding with the constant rate of
fermentation from the total actually evolved, in the manner frequently
described before.

Free phosphate in No. 3=1110 g., in No. 2=0:180 g. and in No. 1
= 0127 g. Mg.P.O,, showing that the same quantity of sugar is still bound
up as hexosephosphate in No. | as in No, 2.

CO, evolved up to the time when No. 2 was boiled (after

52 mins. fermentation) = Sci Fe 293°6 cc.
Rate=2°8x26 .... Pap aa ae TPS! 55
Equivalent of phosphate ae a4 eats 220°8 ce. or 201 ce. at

N.T.P., ie. 0-394 g.

This corresponds therefore to 0°79 g. glucose.

The amount of glucose converted into hexosephosphate may also be de-
termined from the phosphate combined during the experiment. Phosphate
bound up in No. 2=1:110 —0:130 = 0980 g. Mg,P.0,; equivalent therefore
to 0°980 x 489 = 0°795 g. glucose.

The tables show the amount of glucose which cannot be accounted for as
CO, and alcohol or as hexosephosphate, the glucose originally present being

obtained from No. 3 by analysis.

Flask (2).
CO, evolved =293°6 cc. or 267 cc. at N.T.P.=0°53 g. equivalent to 1-06 g. glucose.

Glucose bound up as hexosephosphate se bee 0:79 g. i"

Total accounted for a Lae ae a 1 1°85 nah Gr
Original glucose 9°98 g.
Final glucose 8°15 ¢.
Actual loss 1°83 g. :

No disappearance of glucose was observed.

Flask (1).
CO, evolved = 1458 cc. or 1339 cc. at N.T.P. or 2°634 g. equivalent to 5-27 g. glucose.

Glucose bound up as hexosephosphate et, ae 0:79 g. <
Total accounted for 348 re ate ate 6:06 g. RA
Original glucose 9°98 g.
Final glucose 0-91 g.
Actual loss 9-07 g.

Thus 9:07 — 6-06=3-01 g. of glucose have disappeared.

The ratio between the reducing power and the optical rotation was
determined in each mixture after the treatment with Patein’s solution i

|

A. HARDEN AND W. J. YOUNG 633

order to see if any active substance other than sugar were present. For
the sake of convenience the rotation observed in a 400 mm. tube is compared
with the reducing power determined by Pavy’s method expressed as grams of
glucose in 100 ce. .
: : : Rotation in 400 mm. tube
re glucose this ratio —~—___— oe = POG
With pure gtucos ° Reduction (g. glucose per 100 cc.) + 2°05,

whilst with pure fructose it is — 4°03.

These ratios were found to be
(1) +o35¢=+1791.
(2) +7 = 42-06.
2°84:
(3) +23 =4 215.

It is thus seen that in No. 1 some substance is present which has a much
greater dextrorotatory power than has glucose, whereas in No. 2 all the
rotatory power may be accounted for by the quantity of glucose present.

Experiment 2. A similar experiment was carried out with fructose
(Kahlbaum), the following mixtures being employed:

(1) 100 cc. maceration juice + 25 cc. 40 per cent. fructose + toluene.

(2) 100 ce. » +25 ce. 40 per cent. “3 2
(8) 100 ce. if 4, +25 ce. water + toluene.
(4) 100 cc. oF ¥) + 25 cc. ” »

The juice in each case was kept in the bath until the temperature was
attained, and the fructose and water then added. Nos. 1 and 3 were boiled
immediately, whilst the others were incubated and the fermentation observed.
In No. 2, a high phosphate rate of 53 cc. per 5 minutes was rapidly reached
which then decreased as usual to a constant rate of 7°5 cc. per 5 minutes; at
the end of 17 hours both Nos. 2 and 4 were boiled. During the first 70 minutes
No. 2 had given off 352°5 ec. of carbon dioxide, the amount due to the
phosphate thus being 352°5 — 14 x 75 = 3525 — 105 = 247°5 ec.; the total
evolved in the 17 hours was 1179°5 cc. at 762°8 mm, and 15°. No. 4 showed
no fermentation at all.

A portion of each of the four filtered mixtures was treated with Patein’s
solution as in the last experiment, and the reducing power and rotations
determined in an aliquot portion of the filtrate. The figures given are
calculated for the total volume of the juice.

1. Original 2. After 3. Original 4, Juice incu-
mixture 17 hrs. juice alone bated alone
Rotation in 400 mm. tube — 30°81° +4°65° —0-061° — 0:066°
Total sugar as g. glucose 9731 1-10 0 0

41—3

634 A. HARDEN AND W. J. YOUNG

- Loss of sugar = (1) — (2) = 821 g.

CO, evolved = 1179°5 ec. at 762°8 mm. and 15° = 2:08 g.

CO, corresponding to fructose bound up as hexosephosphate = 247°5 cc. at
762°8 mm. and 15°=0°41 g. .

Total sugar accounted for as CO, and hexosephosphate therefore

=2x 249= 498 g,

Sugar disappeared = 8:21 — 4:98 = 3:28 g.

The ratios of rotation to reduction expressed as before were also de-
termined in the solution after treatment with Patein’s solution and were
found to be (1) — 414, (2) +.5°28; pure fructose = — 4°03.

It is thus seen that in this experiment a substance having a high dextro-
rotation was formed, so that although all the fructose was not used up, the
mixture had changed in rotation from laevorotatory to dextrorotatory, whilst
a much larger proportion of the original fructose had disappeared than could
be attributed to the fermentation.

In the solutions before treatment with Patein’s solution the free phos-
phate was estimated and was found to be:

(1) 1-208 g. Mg,P,0,.

(2) 0174s. -.,,
(yea ls210 oF) a.
(4) 12082. .,,

From this it is seen that the phosphate and hence an equivalent portion
of the sugar was still bound up as hexosephosphate at the end of the
experiment (No. 2). These numbers serve as before as a check on the
quantity of sugar which has been converted to hexosephosphate, viz. that
amount corresponding to

1:203 — 0174 = 1029 g. Mg.P,0, or 1:029 x 483 fructose = 0°834.
Fructose calculated from equivalent of CO, as above = 0°41 x 2 =0°82 g.

The mixtures before treatment with Patein’s solution were tested with
iodine solution; Nos. 1, 3, and 4 gave no colouration, whereas No. 2 gave a
deep reddish brown colouration.

A portion of No. 2 treated with three volumes of alcohol gave a white
precipitate, which was redissolved in water and again precipitated with
alcohol. This last precipitate gave an opalescent solution in water which
was precipitated by saturation with ammonium sulphate and gave a red
colouration with iodine. The only difference which could be seen from the
behaviour of glycogen was that it gave a somewhat different red colour
with iodine.

te a at Foe 4

A. HARDEN. AND W. J. YOUNG 635

The other solutions Nos. 1, 3, and 4 gave slight precipitates with alcohol,
the aqueous solutions of which gave however no colouration with iodine,

It is thus seen that during the fermentation of fructose by maceration
juice a dextrorotatory, glycogen-like substance is formed,

These results appear to us as already indicated to throw some light on
the cause of the difference which exists between sugar fermented and carbon
dioxide evolved, not only in the case of yeast preparations but also in that
of living yeast. Euler and his colleagues in recent papers have argued from
the existence of this difference between the amount of sugar actually removed
by living yeast from a glucose solution and the amount equivalent to the CO,
evolved, which he terms A—C, that the hexose requires to undergo some
change which renders it directly fermentable and that the difference A—C
represents the amount which is in this intermediate condition, [Euler and
Johannson, 1912; Euler and Berggren, 1912.] There seems however to be no
good reason to suppose that Euler and Johannson’s A — C cannot be accounted
for by the well-known formation of glycogen which has been shown by Pavy
and Bywaters [1907] to be of the order of magnitude required.

In Euler and Berggren’s experiments on the effect of yeast extract in
increasing both rate of fermentation and A—C [1912], no counts of yeast
cells before and after the experiments were made. As the earliest observa-
tions were made after an hour at 15°-18° and the experiments in some cases
extended to over six hours (1 g. pressed yeast in 25 cc. of solution), the
possibility of yeast growth must not be overlooked. This is still more
probable in the cases in which only 0:25 g. of pressed yeast was taken and
tested with yeast extract itself, various precipitates from yeast extract, and
with sodium nucleinate or ammonium formate [1912, pp. 216, 217; Euler
and Cassel, 1913], in a total volume of 40 cc.

An experiment made on this point showed that under similar conditions
of concentration growth readily occurs at 25°. The yeast was added as
5 ce. of a suspension of 5 g. yeast in 100 cc. H,O, ie. 0°25 g. yeast.

(1) and (2) 5 ce. yeast suspension + 20 cc. of 20 per cent. glucose
+15 cc. H,O.

(3) and (4) 5 cc. yeast suspension + 20 cc. of 20 per cent. glucose
+15 ce. yeast extract.

In 345 mins. the evolutions were respectively 61, 59°6, 1463, 150 ce,
At the close of this time the numbers of cells present per cc, were 68'°7 x 10°,
68°5 x 108, 105°25 x 10°, 98°8 x 10°.

Asparagine acts in a precisely similar manner, 0°25 g. added to 0°5 g,
yeast in 30 cc. sugar solution increasing the evolution in 2 hrs, at 25° from
73'6 to 89°4 cc,

636. A. HARDEN AND W. J. YOUNG.

_It therefore seems that the experiments in which Euler has shown the
accelerating effect of yeast extract, sodium nucleinate, etc., on the action of
living yeast require revision from this point of view.

The method of testing for a co-enzyme by the action of solutions on living
yeast is moreover open to the criticism that the yeast cell is, if at all, only
imperfectly permeable to the co-enzyme so that negative results would be of
little value.

SUMMARY.

During the alcoholic fermentation of glucose and fructose by Lebedetf’s
maceration extract of dried yeast, dextrorotatory polysaccharides are pro-
duced, and it is to the formation of these that the difference between the
sugar removed and that equivalent to the carbon dioxide evolved is principally
to be attributed.

REFERENCES.

Cremer, M. (1899), Ber. 32, 2062.

Euler and Berggren (1912), Zeitsch. Gdrungsphysiol. 1, 203.
— and Cassel (1913), Zeitsch. physiol. Chem. 86, 122.
and Johannson (1912), Zeitsch. physiol. Chem. 76, 347.
Harden and Young (1904), Ber. 37, 1052.

—— (1908), Proc. Roy. Soc. B, 80, 299.

Pavy and Bywaters (1907), J. Physiol. 36, 149.

Slator, A. (1911), J. Inst. Brewing, 17, 147.

NOTE:
HASLAM. Separation of proteins, Part 111, Globulins. This Journal 1913, 7, 492.
In section 4 of Summary, p. 515, for 0:1 mg. P °/, read 0:1 Py.

a,

INDEX

Acetaldehyde, production of, during the an-
aerobic fermentation of glucose by B. coli
communis (Grey) 359

Adamkiewicz test for protein, Hopkins & Cole
modification of (Mottram) 249

Apams, A. see Moorr, B.

AGasHE, G. S. see WetzMann, C.

Aldehydes, aromatic, condensation of with
Pyruvie acid (Lubrzynska and Smedley)
375

Amylase, behaviour of in presence of a specific
precipitate (Porter) 599

Analysis, capillary, quantitative relations in
(Schmidt) 231

Antirrhinum Majus, flower pigments of (Whel-
dale and Bassett) 87, 441

Atkins, W. R. G. and Wauuace, T. A. Critical
solution point of Urine 219

Banrcrorr, J. Combinations of Haemoglobin
with Oxygen and with Carbon Monoxide
481

Barenprecu?t, H. P. Enzyme action, facts
and theory 549

Barger, G. and Ewrns, A. J. Identity of
Trimethylhistidine (Histidine-Betaine) from
various sources 204

Bassett, H. L. see WHEtLDALE, M.

Birds, preparation from animal tissues of
substance which cures polyneuritis in
(Cooper) 268

Blood, lipolytic action of the (Thiele) 275

Blood, modification of Teichmann’s test for
(Symons) 596

Braumacuari, U. N. Investigation into the
physico-chemical mechanism of haemolysis
by specific haemolysins 562

Brain, fatty acids of human (Grey) 148

Brain, galactosides of the (Rosenheim) 604

Cameron, A. T. Iodine content of fish-
thyroids 466

Capillary analysis, quantitative relations in
(Schmidt) 231

Carbon monoxide, combinations of haemo-

globin with (Hill, Bareroft) 471, 481

Carboxylase (Harden) 214

Caseinogen, clotting of solutions of (Schryver)
568

Carucart, BE. P. and Green, H. H. Rate of
protein catabolism 1

Cell-division, in hyacinths, influence of rare

_ earths on (Evans) 349

Chemistry of the Leucocytozoon Syphilidis
(McDonagh and Wallis) 517

Cuicx, H. Factors concerned in solution and
precipitation of Euglobulin 318

Cnuick, H. and Martin, C. J. Density and
solution volume of some proteins 92
—— Precipitation of egg-albumin by
Ammonium Sulphate 380, 548 j

Cholesterol, content of tissues of cats (Gardner
and Lander) 576

Classification of the sterols (Dorée) 616

Clotting of caseinogen solutions (Schryver) 568

Crorwortny, H. R. S. see Tompson, W. H.

Colour reaction with phosphotungstic and
phosphomolybdic acids, chemical nature of
substances from alcoholic extracts of
various foodstuffs which give a (Funk and
Macallum) 356

Colour reactions of certain indole derivatives
(Homer) 116

Coorrr, E. A. Preparation from animal
tissues of substance which cures polyneu-
ritis in birds 268

Relations of the Phenols and their deriva-
tives to proteins 175, 186

Coprostanone, oxidation of
Godden) 588

Coprosterol, oxidation of (Gardner and Godden)
588

Coprosterol (Dorée) 616

Creatine and Creatinine, use of Folin method
for estimation of (Thompson, Wallace, and
Clotworthy) 445

Cystine and Tyrosine, separation of (Plimmer)
311

(Gardner and

Density, of some proteins (Chick and Martin)
92

Diastatic enzymes, influence of salts on rate of
hydrolysis of glycogen by (Norris) 622

Diets, influence of, upon growth (Hopkins and
Neville) 97

DoréE, C. Isocholesterol, coprosterol and the
classification of the sterols 616

Eaves, E. C. see Puimmer, R. H. A.

Echinoderms, réle of Glycogen, Lecithides,
and Fats in the reproductive organs of
(Moore, Whitley and Adams) 127

Echinus esculentus, basic and acidic proteins of
sperm of (Moore, Whitley and Webster) 142

Egg-albumin, precipitation of by Ammonium
Sulphate (Chick and Martin) 380, 548

Enzyme action (Barendrecht) 549

Enzymes of washed zymin and dried yeast
(Harden) 214

Euglobulin, factors concerned in solution and
precipitation of (Chick) 318

Evans, W. H. Influence of carbonates of rare
earths (Cerium, Lanthanum, Yttrium) on
growth and cell-division in hyacinths 349

638

Ewins, A. J. see Barcrr, G.
and Lamzaw, P. P. Fate of Indol-
ethylamine in the organism 18

Fat, of yeast (Neville) 341

Fats, role of in the reproductive organs of
Echinoderms (Moore, Whitley and Adams)
127

Fatty acids, biochemical synthesis of (Smedley
and Lubrzynska) 364

Fatty acids of human brain (Grey) 148

Fermentation of Glucose by B. coli communis,
production of Acetaldehyde during the (Grey)
359

Fermentation, rate of, by growing yeast cells
(Slator) 197 :

Folin method, use of for estimation of Creatine
and Creatinine (Thompson, Wallace and
Clotworthy) 445

Foodstuffs, chemical nature of substances from,
which give colour reactions with phospho-
tungstic and phosphomolybdic acids (Funk
and Macallum) 356

Funk, ©. Estimation of vitamine-fraction in
milk 211

—— Nitrogenous constituents of lime-juice
81

and Macauuum, A. B. On the chemical

nature of substances from alcoholic extracts

of various foodstuffs which give a colour
reaction with phosphotungstic and phos-

phomolybdie acids 356

Galactosides, of the brain (Rosenheim) 604

GARDNER J. A. and GoppEN, W. Oxidation of
coprosterol and coprostanone 588

and Lanner, P. E. Cholesterol content
of the tissues of cats under various dietetic
conditions and during inanition 576

Gas-electrode for general use (Walpole) 410

Germicidal properties of phenols (Cooper) 175

Globulins, separation of (Haslam) 492, 636

Glycogen, hydrolysis of, by diastatic enzymes
(Norris) 26

Glycogen, influence of salts on rate of hydro-
lysis of by diastatic enzymes (Norris) 622

Glycogen, role of in the reproductive organs
of Echinoderms (Moore, Whitley. and
Adams) 127

Glyoxylic reaction, significance of colour re-
action of certain indole derivatives with
regard to (Homer) 116

GoppEN, W. see GARDNER, J. A.

Green, H. H. see Carucart, E. P.

Grey, E. C. Fatty acids of human brain
148

— Production of acetaldehyde during the
anaerobic fermentation of Glucose by
Bacillus coli communis 359

Growth, influence of diets upon (Hopkins and
Neville) 97

Growth, of hyacinths, influence of rare earths
on (Evans) 349

Haemoglobin, combinations of, with Oxygen
and with Carbon monoxide (Hill, Barcroft)
471, 481

INDEX

Haemolysis, physico-chemical mechanism of,
by specific haemolysins (Brahmachari) 562

Harpen, A. Enzymes of washed zymin and
dried yeast. Carboxylase 214

— and Youne, W. J. Enzymatic formation
of polysaccharides by yeast preparations
630

Hasna, H. C, Separation of Proteins. Glo-
bulins 492, 636

Hewitt, J. A. Metabolism of nitrogenous
sugar derivatives 207

Hint, A. V. Combinations of Haemoglobin
with Oxygen and with Carbon monoxide
471

Histidine-Betaine, identity of, from various
sources (Barger and Ewins) 204

Histone, osmotic pressure of a, direct measure-
ment of (Moore, Whitley and Webster) 142

Homer, A. Colour reactions of certain indole
derivatives 116

Condensation of Tryptophane and other
indole derivatives with certain aldehydes
101

Hopkins and Cole modification of Adamkiewicz
test for Protein (Mottram) 249

Hopxrins, F. G. and Nrvinur, A. Influence
of diets upon growth 97

Hyacinths, influence of carbonates of rare
earths on growth and cell-division in
(Evans) 349

Hydrogen Chloride, hydrolysis of proteins
with alcoholic solution of (Weizmann and
Agashe) 437

Indole derivatives, colour reactions of certain
(Homer) 116

Indole derivatives, condensation of with certain
aldehydes (Homer) 101

Indolethylamine, fate of, in the organism
(Ewins and Laidlaw) 18

Iodine content of fish-thyroids (Cameron) 466

Isocholesterol (Dorée) 616

Larpnaw, P. P. see Ewrins, A. J.

Lanper, P, E. see Garpner, J. A.

Lecithides, role of in the reproductive organs
of Echinoderms (Moore, Whitley and
Adams) 127

Leucocytozoon Syphilidis, chemistry of (Mc-
Donagh and Wallis) 517

Lime-juice, nitrogenous constituents of (Funk)
81

Lipolytic action of the blood (Thiele) 275

Lipolytic action of the tissues (Thiele) 287

Litmus paper, use of as a quantitative in-
dicator of reaction (Walpole) 260

Luprzynska, E. and Smepury, I. Condensa-
tion of aromatic aldehydes with pyruvic
acid 375

Lusrzynska, E. see Smepiey, I.

Macauium, A. B. see Funes, C.

McDonaceu, J. E. R. and Wautis, R. L. M.
Chemistry of the Leucocytozoon Syphilidis
517

MacLnaan, I. Smepiey.
acid 611

Estimation of pyruvic

INDEX

Martin, C. J. see Cuick, H.

Metabolism of nitrogenous sugar derivatives
(Hewitt) 207

Metabolism of organic phosphorus compounds
(Plimmer) 43

Milk, estimation of vitamine-fraction in (Funk)
211

Minroy, J. A. Some observations on the esti-
mation of Urea 399

Moore, B., Wurriry, E. and Apams, A. Role
of Glycogen, Lecithides, and Fats in re-
productive organs of Echinoderms 127

—— — and Wessrrr, A. Basic and acidic
proteins of sperm of Hehinus esculentus.
Osmotic pressure of a protamine or histone
142

Morrram, V. H. Hopkins and Cole modifi-
cation of Adamkiewicz test for protein
249

Nevinite, A. The fat of yeast 341

— see Horkins, F. G.

Nitrogenous constituents of lime-juice (Funk)
81

Nitrogenous sugar derivatives, metabolism of
(Hewitt) 207

Norris, R. V. Hydrolysis of Glycogen by
diastatic enzymes 26

—— Influence of salts on the rate of hydro-
lysis of glycogen by diastatic enzymes
622

Osmotic pressure of a protamine or histone,
direct measurement of (Moore, Whitley and
Webster) 142

Oxygen, combinations of haemoglobin with
(Hill) 471

Oxygen, combinations of haemoglobin with
(Barcroft) 481

Pace, H. J. see Puimmer, R. H. A.

Palmitic acid, some esters of (Stephenson)
429

Phenols, effect of various factors upon germi-
cidal and protein-precipitating powers of
(Cooper) 175

Phenols and their derivatives, relations of, to
proteins (Cooper) 175, 186

Phosphorus compounds organic, hydrolysis of
by dilute acid and by dilute alkali (Plimmer)
72

Phosphorus compounds organic, metabolism
of (Plimmer) 43

Phosphotungstic and phosphomolybdie acids,
chemical nature of substances from alcoholic
extracts of various foodstuffs which give a
colour reaction with (Funk and Macallum)
356

Phytin, investigation of (Plimmer and Page)
157

Pigments, flower, of Antirrhinum
(Wheldale and Bassett) 87, 441

Purmmer, R. H. A. Hydrolysis of organic
phosphorus compounds by dilute acid and
by dilute alkali 72

— Metabolism of organic phosphorus com:
pounds 43

Majus

639

Puen, R. H. A. Separation of Tyrosine and
Cystine 311

and Eaves, BE. C. Estimation of Tyro-

sine in proteins by bromination 297

and Pacer, H, J. Investigation of Phytin
157

Polyneuritis in birds, preparation from animal
tissues of substance which cures (Cooper)
268

Polysaccharides, enzymatic formation of by
yeast preparations (Harden and Young)
630

Porvrer, A. E. Behaviour of Amylase in pre-
sence of a specific precipitate 599

Precipitation of egg-albumin by Ammonium
Sulphate (Chick and Martin) 380, 548

Protamine, osmotic pressure of a, direct
measurement of (Moore, Whitley and
Webster) 142

Protein catabolism, rate of (Cathcart and
Green) 1

Hopkins and Cole modification of
Adamkiewicz test for (Mottram) 249

Proteins, basic and acidic, of sperm of Echinus
esculentus (Moore, Whitley and Webster)
142

chemical action of quinone upon (Cooper)
186

— density and solution volume of some
(Chick and Martin) 92

estimation of Tyrosine in, by bromina-
tion (Plimmer and Eaves) 297

— hydrolysis of, with alcoholic solution
of Hydrogen Chloride (Weizmann and
Agashe) 437

precipitating power of phenols for (Cooper)
175

— relations of Phenols and their derivatives
to (Cooper) 175, 186

salting-out of (Chick and Martin)

separation of (Haslam) 492, 636

Pyruvie acid, condensation of aromatic alde-
hydes with (Lubrzynska and Smedley)
375

Pyruvie acid, estimation of (MacLean, I. Smed-
ley) 611

380

Quinone, chemical action of upon proteins
(Cooper) 186

Rare earths, influence of carbonates of, on
growth and cell-division in hyacinths
(Evans) 349

Reaction, use of litmus paper as a quantitative
indicator of (Walpole) 260

RosenuHEIM, O. Galactosides of the brain 604

Salting-out of proteins (Chick and Martin)
380

Salts, influence of, on rate of hydrolysis of
glycogen by diastatic enzymes (Norris)
622

Scumpt, H. Quantitative relations in capil-
lary analysis 231

Scuryver, 8S. B. Notes on some further
experiments on the clotting of caseinogen
solutions 568

640

Staror, A. Rate of fermentation by growing
yeast cells 197

Smepury, I. and Lusrzynsxa, EH. Biochemical
synthesis of fatty acids 364

see Luprzynska, E.

Solution-volume, of some proteins (Chick and
Martin) 92
SrePHENSON, M.
acid 429

Sterols, classification of (Dorée) 616

Sugar derivatives, nitrogenous, metabolism of
(Hewitt) 207

Symons, C. T. Modification of Teichmann’s
test for blood 596

Synthesis of fatty acids, biochemical (Smed-
ley and Lubrzynska) 3864

Some esters of Palmitic

Teichmann’s test for blood, modification of
(Symons) 596

Tutene, F. H. Lipolytic action of the blood
275

Lipolytic action of the tissues 287

Tompson, W. H., Wautuacr, T. A. and Crot-
wortHy, H. R. 8S. On the use of the Folin
method for estimation of Creatine and
Creatinine 445

Thyroids, fish, iodine content of (Cameron)
466

Tissues, lipolytic action of (Thiele) 287

Trimethylhistidine (Histidine-Betaine) from
various sources, identity of (Barger and
Ewins) 204

Tryptophane, condensation of with certain
aldehydes (Homer) 101

Tyrosine, and Cystine, separation of (Plimmer)
311

Tyrosine, estimation of in proteins by bromina-
tion (Plimmer and Hayes) 297

INDEX

Urea, estimation of (Milroy) 399

Urine, critical solution point of (Atkins and
Wallace) 219

in milk

Vitamine-fraction, estimation of

(Funk) 211

Wautack, T. A. sce Arxins, W. R. G. and
Tuompson, W. H.

Watts, R. L. M. see McDonacu, J. HK. R.

Wanpoutr, G. 8. Gas-electrode for general
use 410

—— Use of litmus paper as a quantitative
indicator of reaction 260

Wessrer, A. see Moore, B.

Wuizmann, C. and Acasuu, G. 8. Hydro-
lysis of proteins with alcoholic solution
of Hydrogen Chloride 437

Wuevpatr, M. Flower pigments of Antirrhi-
num Majus. Method of preparation 87

and Basserr, H. L. Flower pigments of
Antirrhinum Majus. (2) The pale yellow
or ivory pigment 441

Wuitiry, EK. see Moors, B.

Yeast cells, growing, rate of fermentation by
(Slator) 197

Yeast, dried and washed, enzymes of (Harden)
214

Yeast, enzymatic formation of polysaccharides
by preparation of (Harden and Young)
630

Yeast, the fat of (Neville) 341

Youne, W. J. see Harpen, A.

Zymin, washed, enzymes of (Harden) 214

CAMBRIDGE: PRINTED BY JOHN CLAY, M.A, AT THE UNIVERSITY PRESS.

yh U
. My

a
P im

‘Dh

:
0 ee A
‘ery, %

A Soha a
1 ‘eh. wl + F ' ‘ janny +
‘3 i ‘ z. ;

My \
} |
j h
" ae
-
i ) ~_
i
'
\
’
f
1 .
; 5
¢ <4 é
| ¥

QP The Biochemical journal
501

B47

Val

cop.4

Biological

& Medical

Serials

PLEASE DO NOT REMOVE
CARDS OR SLIPS FROM THIS POCKET

UNIVERSITY OF TORONTO LIBRARY

gor
Pe

Pei F I Eae.

Say Soe
Ss

aaet

ae

A.
RIT p=

Paso,

igre:
nes

s nets ote
Pars once Mi at arses rege bes ~
; Sap ere A ig este
ex

peer so Gee
oe =

: eee han gt nn r =
Seren ep coe cir ap ieer ea eae
Sra 4 Gans ee ae