OU_1 60897 >[g
OSMANTA UNIVERSITY LIBRARY
Gall No.
Author SoCAcJ-M poy
Title
This book
last marked below.
SYMPOSIA OF THE
SOCIETY FOR EXPERIMENTAL BIOLOGY
NUMBER VIII
Other Publications of the Company of Biologists.
JOURNAL OF EXPERIMENTAL BIOLOGY
THE QUARTERLY JOURNAL OF MICROSCOPIC SCIENCE
JOURNAL OF EXPERIMENTAL MORPHOLOGY
SYMPOSIA
I. NUCLEIC ACID.
II. GROWTH, DIFFERENTIATION AND MORPHO-
GENESIS.
III. SELECTIVE TOXICITY AND ANTIBIOTICS.
IV. PHYSIOLOGICAL MECHANISMS IN ANIMAL
BEHAVIOUR.
V. FIXATION OF CARBON DIOXIDE.
VI. STRUCTURAL ASPECTS OF CELL PHYSIOLOGY.
VII. EVOLUTION.
The Journal of Experimental Botany is
published by the Oxford University Press
for the Society for Experimental Biology.
SYMPOSIA OF THE
SOCIETY FOR EXPERIMENTAL BIOLOGY
NUMBER VIII
ACTIVE TRANSPORT
AND SECRETION
Published for the Company of Biologists
on behalf of the Society for Experimental Biology
CAMBRIDGE: AT THE UNIVERSITY PRESS
1954
PUBLISHED BY
THE SYNDICS OF THE CAMBRIDGE UNIVERSITY PRESS
London Office: Bentley House, N.w. i
Agents for U.S.A., Academic Press Inc.,
125 East 23rd Street, New York 10
Agents for Canada, India, and Pakistan: Macmillan
Printed in Great Britain at the University Press, Cambridge
(Brooke Crutchley, University Printer)
CONTENTS
Preface />. vii
by R. BROWN and j. F. DANIELLI
Movements of Water and Electrolytes in Invertebrates i
by j. A. RAMSAY
Vertebrate Physiology from the point of view of Active Transport 16
by HUGH DAVSON
The Concept and Definition of Active Transport 27
by THOMAS ROSENBERG \ ,
Secretion and Transport of Water 42
by j. R. ROBINSON
Osmoregulation and Ionic Regulation in Animals without Kidneys 63
by j. A. KITCHING
The Active Transport of Water under Temperature Gradients 76
by D. C. SPANNER
Water Transport in Insects 94
by J. W. L. BEAMENT
The Evidence for Active Transport of Monosaccharides across the 118
Red Cell Membrane
by PAUL G. LEFEVRE
Secretion and Transport of Non-electrolytes 136
by W. WILBRANDT
Comment on Professor Wilbrandt's and Dr LeFevre's Papers 162
by w. F. WIDDAS
Enzyme Systems of the Cell Surface involved in the Uptake of
Sugars by Yeast 165
by ASER ROTHSTEIN
Active Cation Transport in Erythrocytes 202
by MONTAGUE MAIZELS
Linkage of Sodium- and Potassium-active Transport in Human
Erythrocytes 228
by E. j. HARRIS
VI CONTENTS
The Accumulation of Amino-acids within Staphylococcal Cells
by E. F. GALE p. 242
Transport of Phosphate through an Osmotic Barrier 254
by P. MITCHELL
Anion Respiration : the Experimental Basis of a Theory of Absorp-
tion, Transport and Exudation of Electrolytes by Living Cells
and Tissues 262
by H. LUNDEGARDH
Some Aspects of Ion Transport through Membranes 297
by EDWARD J. CONWAY
Cation Absorption by Non-growing Plant Cells 325
by J. F. SUTCLIFFE
The Relationship between Metabolism and the Accumulation of
Ions by Plants 343
by R. SCOTT RUSSELL
Salt Accumulation in Plants: a Reconsideration of the Role of
Growth and Metabolism 367
by F. c. STEWARD and F. K. MILLAR
Active Transport of Inorganic Ions 407
by HANS USSING
Movements of Cations during Recovery in Nerve 423
by A. L. HODGKIN and R. D. KEYNES
The Regulation of Sodium and Potassium in Muscle Fibres 438
by H. BURR STEINBACH
Relations between Active Transport and Metabolism in some
Isolated Tissues and Mitochondria 453
by R. E. DAVIES
Active Transport through Embryonic Membranes 476
by F. W. R. BRAMBELL and W. A. HEMMINGS
Transport of Lipid through Cell Membranes 490
by A. C. FRAZER
Morphological and Molecular Aspects of Active Transport 502
by J. F. DANIELLI
PREFACE
This volume contains the papers read at a Symposium of the Society for
Experimental Biology which was held at Bangor in July 1953. It is the
eighth of an annual series of Symposium Reports. The Symposium for
1954 will be held at Leeds, on Fibrous Proteins.
In the present Symposium the first three papers are introductory in
character. These are followed by four papers on water movements and
four papers on active transport phenomena in red blood cells, yeast and
bacteria. The next group of five papers are concerned with active transport
of ions in plants, and are followed by four papers on active transport of
ions in animal cells. The remaining three papers are concerned with active
movements of proteins and fats, and with mechanisms of active transport.
The papers presented here should be considered in relation to a number
of recent reviews, notably by Conway (1953) (Biochemistry of Gastric
Secretion. Springfield: Thomas), Hodgkin (1951) (Biol. Rev. 26, 339),
Brown (1952) (Int. Rev. Cytol. i, 107), Goldacre (1952) (Int. Rev. Cytol.
i, 135) and SutclifTe (1953) (Int. Rev. Cytol. 2, 179).
The Society is deeply indebted to the British Council, the Rockefeller
Foundation and to Imperial Chemical Industries Ltd. for financial aid.
The Editors wish to thank the members of the Advisory Committee who
assisted us in preparing the Symposium programme. We also wish to
thank the Cambridge University Press for the kindness with which we were
assisted in producing this report.
R. BROWN
J. F. DANIELLI
Symposium Editors
Society for Experimental Biology
17 February 1954
MOVEMENTS OF WATER AND ELECTRO-
LYTES IN INVERTEBRATES
BY J. A. RAMSAY
Department of Zoology, University of Cambridge
I. INTRODUCTION
When I was asked to give this introductory paper it was suggested to me
that I should first and foremost present the background against which
modern developments can be seen in perspective. It would be neither
practicable nor desirable to attempt a review of all the material available ;
instead, what I shall do is to trace the development of ideas in the subject,
to consider why some lines of approach have prospered more than others,
why some questions have been answered and others left unasked, to show
how the existence of active transport mechanisms has been recognized and
to put forward some suggestions as to how they may have been evolved.
A few analyses of the body fluids of invertebrates were published during
the latter part of the nineteenth century, and work of this type continued
sporadically throughout the first two decades of the twentieth. At this
stage these investigations were not inspired by any precise theory as to the
nature of the body fluids and the ways in which their compositions were
maintained ; nevertheless, it became apparent that the body fluids of animals
were in general not unlike sea water. Owing to the relative ease with which
freezing-point measurements can be made our knowledge of the osmotic
pressure of the body fluids began to advance rapidly. Experiments were
carried out to test the effects of changes in the external medium upon the
osmotic pressure of the body fluid. On the other hand, determination of
the constituents of the body fluids was beset by technical difficulties and
advance was slower. The development of the subject can be followed in
the reviews which have appeared from time to time, notably those of Duval
(1925), Schlieper (1930, 1935), Pantin (1931), Krogh (1939) and Beadle
(1943). By the time the subject became of sufficient importance to merit
review it had also acquired a philosophy which conveniently rationalized
its ecological and physiological aspects — Claude Bernard's now famous
pronouncement 'la fixite du milieu interieur est la condition de la vie
libre'. For the subject now under discussion Claude Bernard's pronounce-
ment has the following special implication : primitive marine animals have
2 MOVEMENTS OF WATER AND
in general no means of regulating the composition of their body fluids,
and penetration into fresh water is only possible for animals which have
evolved such means. Although the reviewers I have mentioned may have
been mainly concerned with the physiological mechanisms whereby
constancy is achieved, they have very obviously accepted Claude Bernard's
proposition and incorporated it into the background of their ideas.
Let us then begin by considering a primitive marine animal having no
control over the composition of its body fluid. When such an animal is
placed in dilute sea water it swells, which may be interpreted as due to the
inward diffusion of water. After some time in dilute sea water the volume
of the animal returns to normal, which may be interpreted as the result of
the relatively slower outward diffusion of salts through the general body
surface. This, of course, is to look upon the animal as little more than a bag
containing sea water. But even a primitive animal is generally something
more than this. We have to consider that it has an alimentary canal, that
it takes in food together with some sea water, and that it voids faeces which
also have some admixture of fluid. It has an excretory organ from which
urine is eliminated. Even if the body fluid is isotonic with sea water and
the net exchange across the external surface is zero, the animal continually
gains water, partly along with its food and partly as metabolic water
produced by the oxidation of the food within the body, and loses water with
its urine and faeces. There is thus a current of water continuously main-
tained through the body, upon which other movements, such as occur
when the animal is placed in dilute sea water, are superimposed. The
recovery of normal volume in dilute sea water, which we considered a
moment ago, is not primarily due to leakage of salts and water through the
general surface; it is due to an increased flow of urine, and it is via the
excretory organ that most of the water and most of the salt leaves the body.
There are in fact three principal regions of the body through which
exchanges with the external medium can and do take place: (i) between the
body fluid and the external medium at the surface of the body, (2) between
the body fluid and the fluid in the gut, (3) between the body fluid and the
urine in the excretory organ. All of these can be the sites of active transport
mechanisms.
Next, let us consider what happens when this primitive marine animal
evolves the ability to live in fresh water. This is a question which has been
discussed at length by Beadle & Cragg (1940) and by Beadle (1943), and
they have come to the conclusion that there are two stages in the process.
I can most conveniently illustrate their thesis with examples drawn from
the Crustacea (see Fig. i). The spider crab Maia will serve to represent
the primitive marine animal. It has virtually no powers of osmotic
ELECTROLYTES IN INVERTEBRATES 3
regulation, the osmotic pressure of its blood following that of the external
medium over the whole of its viable range. The shore crab Carcinus shows
some powers of osmotic regulation, being able to maintain the osmotic
pressure of its blood above that of dilute sea water. It is commonly found
in estuaries as well as in the sea but cannot maintain itself in fresh water.
As an example of a crustacean fully adapted to fresh water we will take the
crayfish Astacus. In this case the general level of the osmotic pressure of
the blood is lower than in Maia and Carcinus, but it can be maintained at
this level even in fresh water.
3°C
Sea water A0
Fig. i. Relation between the osmotic pressure of the blood, and the osmotic pressure of
the external medium, for three crustaceans. Maia from Duval (1925), Carcinus from
Duval (1925) and Schmidt-Nielsen (1941), Astacus from Herrmann (1931).
It might well have been imagined by the early investigators that these
animals, when placed in dilute sea water, would follow what seems to us
the obvious and logical course of pumping out the water which diffuses
into them. But this does not seem to be the case. In a recent article,
Robinson (1953)* has reviewed the evidence for the active transport of
water in living systems, and as far as the aquatic invertebrate Metazoa are
concerned the evidence is as yet circumstantial. It appears that these
animals prefer to transport dissolved substances in such a way as to
compensate for the passive movements of water under osmotic gradients.
According to Beadle & Cragg, in the first stage of the evolutionary
process the animal develops the power of actively transporting salts from
* I wish to thank Dr Robinson for allowing me to see his review in typescript before
publication.
4 MOVEMENTS OF WATER AND
the external medium into the blood, to an extent which is sufficient to
maintain the osmotic pressure of the blood significantly above that of the
external medium. The excretory organ, however, does not back up the
effort of the surface membranes. In Carcinus the urine is isotonic with the
blood under all conditions, and beyond a certain point of dilution more
salt is lost via the urine than can be gained by absorption through the
surface. Ecologically this point lies for Carcinus some distance up the
estuary, but definitely short of the river. But there is a crustacean which
does in fact succeed in getting into rivers on this same inefficient physio-
logical basis, and that is the Chinese mitten crab Eriocheir. As in Carcinus,
so in Eriocheir the urine is isotonic with the blood under all conditions, but
by sheer hard work, by absorbing salts from the river water at a great rate,
this determined animal penetrates up rivers such as the Elbe for hundreds
of miles, returning to the sea only for the purposes of breeding. For
Eriocheir the ecological limit seems to lie not between brackish water and
fresh water as for Carcinus, but between hard fresh water and soft fresh
water; Eriocheir does not appear to be able to penetrate the softer waters
of the Norwegian rivers (Schmidt-Nielsen, 1941).
Then in the second stage of the evolutionary process two things happen.
First, the excretory organ becomes awakened to a proper sense of its
responsibilities and produces hypotonic urine, thus conserving the salt
content of the body ; secondly, the general level of the osmotic pressure of
the blood is lowered to about half that found in marine animals. This
reduces the strain on the active transport mechanisms at the body surface
and in the excretory organ. It also involves some readjustment of the salt
and water balance between the tissues and the blood.
It emerges from this survey that the important sites of active trans-
port lie in the general body surface — or specialized parts of it — and in the
excretory organ. There is as yet very little evidence of active transport in
the gut, at least as far as the digestive epithelium is concerned. We will
therefore now proceed to further consideration of the surface membranes
and of the excretory organs in aquatic invertebrates.
II. THE TRANSPORT OF IONS BY
SURFACE MEMBRANES
The first demonstration that an animal can maintain the osmotic pressure
of its blood by uptake of salts against a concentration gradient was given by
Nagel (1934) for Carcinus. Having confirmed that Carcinus had the power
of hypertonic regulation in brackish water and having demonstrated that
the urine was always isotonic with the blood, Nagel carried out the following
well-planned experiment. He took a number of crabs and allowed them to
ELECTROLYTES IN INVERTEBRATES 5
become adapted to a medium of considerable dilution. Some of the crabs
were killed and measurements were made of the osmotic pressure and
chloride content of their blood. The rest of the crabs were then placed in
another medium, more concentrated than the first medium but less con-
centrated than the blood of the crabs which had become adapted to the first
medium. After 24 hr. in the second medium the crabs were killed and their
blood taken for analysis. Nagel's figures (Table i) show that in the second
lot of crabs both osmotic pressure and chloride concentration of the blood
had increased. Since the body volume remained more or less constant the
increase of osmotic pressure and chloride concentration could only be
explained by the uptake of salts from the external medium against the
concentration gradient. This uptake was not affected by blocking the mouth
and Nagel assumed that it occurred at the gills.
Table i . Demonstration of active transport of chloride
by Carcinus
(From Nagel, 1934)
External medium
Blood
A°C.
Cl (mg./ml.)
A°C.
Cl (mg./ml.)
I.
0-89
8-57
•42
I2*O
•20
11-9
•23
12-3
•42
•28
13-0
"•5
Av. -31
I2'I
II.
1-18
ii'45
•So
I4-O
•52
•48
I4'3
14-2
•73
•56
15-4
I5'3
*57
14-2
•5i
14-0
Av. 1-55 14-5
Three years later, Krogh (1937 a, b) showed that fresh- water fishes and
Amphibia are able to take up chloride from extremely dilute external media,
and he later extended this work to other ions and to invertebrates (Krogh,
1938). His method was to keep the animals in a current of distilled water
until their salt reserves were depleted and then to place them in measured
volumes of dilute solutions whose final composition was determined by
analysis at the end of the experiment. He was able to demonstrate active
uptake of chloride in Astacus, in a variety of fresh- water molluscs and in the
horse leech. Active uptake of chloride from dilute solutions has also been
demonstrated by Koch (1938) for mosquito larvae, by Maluf (1939) for the
6 MOVEMENTS OF WATER AND
earthworm, by Bone & Koch (1942) for caddis larvae and by Holm- Jensen
(1948) for Daphnia. It is therefore of very widespread occurrence; but
there are some fresh-water animals in which it has been looked for but not
found, for example, in the eel and in the larva of the alder fly Siatts
(Beadle & Shaw, 1950).
One of the invertebrates which Krogh studied in particular detail was
Eriocheir, and he was able to show that there was active uptake of sodium,
potassium, chloride, bromide, cyanate and thiocyanate ; that nitrate diffused
inwards rapidly under a concentration gradient, iodide slowly and sulphate
not at all. He was also able to show that the mechanisms for uptake of
anions and cations were independent, e.g. chloride, but not ammonium,
taken up from NH4C1 and replaced by bicarbonate; sodium, but not
sulphate, taken up from Na2SO4 and replaced by ammonium. Eriocheir
does not appear to show any discrimination between sodium and potassium
or between chloride, bromide and thiocyanate when these are present in the
same solution. Astacus (Schmidt-Nielsen, 1941) will absorb sodium but
not potassium from solutions in which both are present, but does not
distinguish between chloride, bromide and thiocyanate.
It so happens that for technical reasons the best evidence for active
absorption of ions from the external medium comes from studies of fresh-
water animals, and there is no doubt that among fresh-water animals these
powers are well developed. Yet it would be wrong to suppose that they are
wholly confined to fresh- water animals.
The most recent and most accurate analyses of the body fluids of marine
invertebrates are those of Robertson (1939, 1949). Although the body
fluids of some primitive marine animals resemble sea water very closely
they are never identical with it owing to the Donnan effect which is set up
by the proteins. Robertson used the method of comparing the ionic
composition of the blood drawn from the animal with that of blood which
had been dialysed against sea water. In this way the Donnan effect is
eliminated from consideration, and it is possible to ascertain how far the
differences between internal and external media are actively maintained.
Some of Robertson's figures are reproduced in Table 2. From this table
it can be seen that no active transport of ions is needed to maintain the
composition of the sea-urchin's coelomic fluid. In the case of the lugworm
Arenicola the only ion showing a significant difference in concentration is
sulphate. But these are the exceptions. Pecten, the scallop, Loligo, the
squid, and Cancer, the edible crab, are all typical marine invertebrates
showing, like Maiay virtually no osmotic regulation; yet it appears likely
that mechanisms of active transport are at work to maintain the generally
higher concentration of potassium.
ELECTROLYTES IN INVERTEBRATES 7
From this we may infer that the active transport of ions by the surface
membranes, which is largely responsible for the hypertonic regulation of
brackish- and fresh-water animals, is not of itself a novelty of adaptive
evolution but is more probably the specialization of a mechanism which was
already in existence in their marine ancestors.
Table 2. Concentrations of various ions in body fluid as percentages
of their concentrations in dialysed body fluid
Na
K
Ca
Mg
Cl
SO4
Echinodermata
Echinus
100
102
101
IOO
IOO
101
Annelida
Arenicola
IOO
103
IOO
IOO
IOO
92
Mollusca
Pecten
100
130
102
97
IOO
96
Loligo
95
219
102
102
103
29
Crustacea
Cancer
108
1 2O
119
51
97
87
Carcinus
109
117
108
34
103
60
Carcinus figures from Webb (1940); the rest from Robertson (1939, 1949).
III. THE TRANSPORT OF IONS IN
EXCRETORY SYSTEMS
The functional unit of the vertebrate kidney is the nephron which consists
of a knot of blood vessels (glomerulus) projecting into a small coelomic
vesicle (Bowman's capsule) from which a tubule leads to the exterior. The
hydrostatic pressure of the blood in the glomerular vessels is sufficiently in
excess of the colloid osmotic pressure of the blood to cause ultrafiltration,
and the fluid in Bowman's capsule is identical with plasma except that it
contains no protein. As this fluid passes down the tubule it is modified by
the reabsorption of substances from it and the secretion of other substances
into it. As is seen from Fig. 2 a, the course of the fluid is : blood -> coelom -+
tubule -> exterior.
Now consider the excretory organs of invertebrates, some of which are
shown diagrammatically in Fig. 26, c and d. It is conceivably possible that
the antennary glands of Crustacea are homologous with the nephridia of
annelids, but it is not easy to refute the assertion that all these types of
excretory organ have been independently evolved. Yet it appears that in all
of them the course of the urine is the same, i.e. blood <-> coelom -> tubule ->
exterior. If they have this much in common, is it possible that they have
other features in common? Is it possible that, like the vertebrate nephron,
they operate on the ultrafiltration-reabsorption basis?
The suggestion that the primary process of urine formation in inverte-
brates was a process of ultrafiltration was first seriously put forward by
Picken (1936, 1937), whose main contribution to this thesis was to measure
the colloid osmotic pressure of the blood in various crustaceans and
8
MOVEMENTS OF WATER AND
molluscs and to show that it was always low in relation to the hydrostatic
pressure. Ultrafiltration was thus a possibility. Paying particular attention
to Anodon, Picken showed that the pericardial fluid was isotonic with the
blood, whereas the urine collected at the excretory pore was hypotonic and
therefore modified in composition during its passage through the tubule.
He drew off the pericardial fluid and showed that it was continuously and
fairly rapidly renewed. Further confirmation came from the work of
Florkin & Duch^teau (1948), wha found that the concentrations of calcium,
chloride and phosphate in the blood and in the pericardial fluid were
identical (Table 3).
Renal artery
- Renal artery
Dorsal blood
vessel
Tubule
Heart Pericardium
(a) Vertebrate
Tubule Tubule
Ventral blood vessel (organ of Bojanus)
^Bladder
(b) Crayfish (c) Earthworm (d) Fresh-water mussel
Fig. 2. Diagrams of the excretory organs of a vertebrate and of Astacus> Lumbricus and
Anodon to show the relation between blood system (black), coelom (stippled) and tubule.
For the others the evidence is less complete. In Astacus, Peters (1935)
succeeded in withdrawing small samples from various parts of the antennary
gland and determined the concentration of chloride (Table 4). His figures
show that within the limits of accuracy the fluids in the coelomic sac and
labyrinth are isotonic with the blood and that the urine becomes hypotonic
during its passage through the tubule. I carried out similar investigations
upon Lumbricus, measuring the freezing-point depression of blood, of
coelomic fluid and of samples collected from different parts of the tubule.
I was able to show that the coelomic fluid is isotonic with the blood and
that the fluid passing down the tubule becomes hypotonic in the region
known as the 'wide tube* (Ramsay, 1949).
Although these results are not in themselves sufficient to prove the truth
of the ultrafiltration-reabsorption theory for invertebrate excretory organs
they are at least compatible with it and can be said to raise it from the level
of mere speculation to that of a reasonable working hypothesis. But we
are still in the speculation stage in regard to the factors which have been
ELECTROLYTES IN INVERTEBRATES 9
at work in the independent evolution of this same physiological process in
so many different animals.
Table 3. Analysis of blood, pericardial fluid, Bojanus
fluid and urine of Anodon
Blood
Pericardial
fluid
(%)
Bojanus
fluid
(%)
Urine
(%>
Absolute
o/
/o
Chloride (HIM. /I.)
Calcium (mM./l.)
Inorganic phosphorus
(mM./l.)
A°C.
18-0
7-0
0-144
0-06*
IOO
IOO
IOO
IOO*
99
IOO
101
IOO*
55
75
90
60*
* Bojanus fluid ' is fluid withdrawn from the excretory organ through its internal opening
into the pericardium, while ' urine ' is fluid withdrawn from the excretory organ through
its opening to the exterior. Figures marked * from Picken (1937), the rest from Florkin &
Duchlteau (1948).
Table 4. Chloride concentration (in mM./l.) in the blood and in fluids
collected from different parts of the excretory organ o/Astacus
(From Peters (1935) as recalculated by Krogh (1939).)
Blood
Coelomic
sac
Main
labyrinth
End of
labyrinth
Tubule
Bladder
196 ±3
198 ±2
209 ±7
2I2±7
9o±6
10-6 ±0-6
In pursuing this line of thought we may next ask ourselves what happens
in animals such as the echinoderms in which there are no recognizable
excretory organs. These animals must take in water with their food and
must produce metabolic water like other animals. How do they get rid of
it? I am not aware that anyone has ever put this point to experimental test,
but it would not surprise me to be told that this water simply escapes by
seepage through the surface membranes of the body. I would be prepared
to risk a guess that if fluid were injected into an echinoderm so as to distend
its body and increase its internal hydrostatic pressure the result would be
an outward seepage of salts and water, only proteins being retained — in
fact, ultrafiltration through the thinner parts of the general body surface.
There is no serious disadvantage in this method of getting rid of water
provided that the flow is normally very small — as it is likely to be in a marine
animal whose blood is isotonic with sea water — and provided that the
animal is not concerned to maintain the composition of its blood signi-
ficantly different from that of sea water. But if the animal is maintaining
some substance X in its blood in higher concentration than in the external
medium and is actively transporting X against a concentration gradient,
10
MOVEMENTS OF WATER AND
then it is wasteful simply to allow an ultrafiltrate to be swept away from
the filtering surface by currents — for this reason, that less work is required
to get X back again from the ultrafiltrate, in which it is initially at the same
concentration as in the blood, than is required to get X from the external
medium in which it is always at lower concentration than in the blood. This
seems obvious, but it is only very recently that the point has been clearly
put, by Potts.*
In the present context the interest of this conception lies in its evolu-
tionary implications. We have seen the physiological parallels which can be
drawn between the vertebrate nephron and the excretory organs of various
invertebrates. Yet as far as we can tell they have all been evolved inde-
pendently. I have suggested that ultrafiltration may be a widespread and
primitive method of volume regulation; if this is true, then animals seem
to have been at pains to restrict ultrafiltration to certain areas of the body
and to arrange that the filtrate has to traverse some sort of tube before
leaving the body. The point which Potts has made seems to me to provide
the argument for selective advantage in this arrangement and enables us to
understand why it has been evolved independently in different phyla of the
animal kingdom.
If this is true it also implies that in marine animals having well-developed
excretory organs the urine, although it may be isotonic with the blood, is
probably not identical with the blood in composition. This is borne out
by comparison of blood and urine in Carcinus (Webb, 1940) and Cancer
(Robertson, 1939). In all these animals the urine is isotonic with the blood
under all conditions, but as Table 5 shows there must be active transport
of some ions in the excretory organ. As we saw in the case of the surface
membranes, so now do we see in the excretory organs, that the active
transport of ions, upon which depends the ability to penetrate fresh waters,
is probably widespread among animals which are exclusively marine in habit.
Table 5. Concentrations of various ions in urine as percentages
of their concentrations in blood
Na
K
Ca
Mg
Cl
SO4
Cancer (Robertson, 1939)
Carcinus (Webb, 1940)
96
95
81
78
90
94
125
390
96
98
165
224
It would appear therefore that the difference between Maia and Carcinus
is not that Carcinus can actively transport salts while Maia cannot; it is
likely that Maia, as well as Carcinus^ has powers of active transport. The
* I wish to thank Dr Potts for permission to quote his work which is as yet unpublished.
ELECTROLYTES IN INVERTEBRATES II
essential difference lies in the rate at which salts are transported relative to
the rate at which water diffuses passively in the same direction. The degree of
hypertonic regulation can be increased either by speeding up the active
transport of salts or by reducing the permeability of the surface membranes
to water, and we have perhaps paid too little attention to this second
possibility.
There is some evidence which suggests that the surfaces of fresh-water
animals are less permeable to water than the surfaces of their marine
relatives. It is not easy to present this evidence in quantitative terms of
permeability measurements because of the difficulties of measuring the
surface area of an animal. But these difficulties are not insuperable, and it
would be of great interest to know, for example, how far the success of
Eriocheir as compared with Carcinus in penetrating fresh water is due to its
powers of active transport and how far due to a decrease in the permeability
of its surface to water. It would also be interesting to know if a decreased
permeability to water is an active process in the sense that it demands a
continuous supply of energy, as suggested by Beadle (1934) for the
flatworm Gunda.
IV. OUTLOOK FOR THE FUTURE
Hitherto I have been concerned in tracing the growth of knowledge and
ideas in what may be called the general field of osmotic regulation. In
logical order, though not in chronological order, the problems were : first,
to determine the general nature of the body fluids; secondly, to show that
their composition was maintained by active transport; thirdly, to discover
the sites of active transport in the body. There are, of course, a great many
invertebrates, of which only a few have been studied, but as far as the
major phyla are concerned it is fair to claim that sufficient ground has been
covered to meet the first two points and it may be conceded that there is
some progress to report in the identification of the sites of active transport.
What is the next step to be? Are we to see the future merely as a process of
filling in the details of a design whose main outlines are already clear?
I do not think so. On the contrary, it seems to me that the next few years
will witness substantial changes in outlook.
Our present outlook is in fact still largely dominated by Claude Bernard
and his 'fixite du milieu int6rieurj. Primitive animals with no powers of
regulation are condemned to live in the sea, those with some powers of
regulation can work their way up estuaries and with the perfection of their
mechanisms can graduate to fresh water. It is a good story and by and large
it is true — but only by and large. On closer inspection the correlation
between powers of regulation and ecological distribution is not so good.
12 MOVEMENTS OF WATER AND
The animals which penetrate farther up estuaries are by no means always
those which can better maintain the constancy of the internal medium.
Cardnus has greater powers of osmo-regulation than Anodony yet Anodon
can live in fresh water while Cardnus cannot. One need look no further
than Hydra to find an animal which lives in fresh water and has no internal
medium at all in Claude Bernard's sense of the term, and of these problems
Dr Kitching is to speak later in this symposium.
The internal medium which Claude Bernard had in mind was of course
the blood. But as has often been pointed out, the internal medium in
which constancy is a prime requirement is the protoplasm of the cell.
The responsibility for maintaining constant conditions in the protoplasm
rests in the last resort upon the cell membrane. In the case of Hydra it
rests solely upon the cell membranes throughout the body. But in the
higher Metazoa living in fresh water, the task of the cell membrane can be
made easier if the medium which bathes it is not fresh water but a saline
solution whose composition is kept constant, and in so far as its task in this
respect is made easier, so we may argue that the cell will be able to apply its
resources more effectively to the main function for which it is specialized.
What the animal does, in short, is to take a part of the load which would
otherwise bear upon all the cells in its body and transfer it to those cells
which separate the blood from the external medium.
The lack of close correlation between powers of regulation and ecological
distribution need not therefore disturb us unduly. A marine animal may
be able to get into fresh water either by evolving good powers of active
transport in all the cells of its body or by evolving an internal medium which
is kept constant by active transport on the part of a few cells in the body,
those which separate the blood from the external medium. The first method
may prove successful, but the second method lays the foundation of a more
efficient physiological organization and has been adopted by Nature for all
her greater evolutionary achievements. The second method, however,
cannot be pursued to the complete exclusion of the other. Not all of the
load can be transferred to the cells of the surface membranes, for, inasmuch
as the cells of the body are not physico-chemically identical but vary from
one tissue to another, the same internal medium cannot be in equilibrium
with all of them and there is still some work to be done by their cell
membranes.
And here, I think, is where our ideas need bringing up to date. We have
been too ready to believe that once the internal medium is stabilized it is all
over bar the shouting. We have concentrated too much upon the active
transport mechanisms at the surface of the body and in the excretory
organ. We have been inclined to think of the cells of the body as being able
ELECTROLYTES IN INVERTEBRATES 13
to relax, as it were, in a medium with which they are in equilibrium. And
this in spite of the abundance of modern evidence which shows that the
cells of the body are not in equilibrium with the fluid which bathes them,
but are actively taking up some materials and are actively keeping others
out.
One of the outstanding problems in this general field is presented by the
inability, relatively speaking, of fresh-water animals to return to the sea.
Although a great many animals have become successfully adapted to fresh
water, not many have the power of passing freely from one medium to the
other. Following Beadle & Cragg, complete adaptation to fresh water
involves lowering of the osmotic pressure of the blood to a new general
level of about half that of sea water. If a fresh-water animal is placed in
sea water the osmotic pressure of its blood usually rises and the animal dies.
The few animals which can survive transference from fresh water to sea
water, such as the Salmonidae and the eel among vertebrates and the prawn
Palaemonetes (Panikkar, 1941) among invertebrates, are capable of hypo-
tonic regulation, that is, they are able to maintain the osmotic pressure of
the blood below that of the medium when they are placed in sea water.
To this the elasmobranch fishes form an interesting exception. The blood
of marine elasmobranchs is isotonic with sea water by virtue of the retention
of urea; the salt content of the blood is not widely different from that of
fresh-water animals generally. This is interesting because it suggests that
a high salt content rather than a high osmotic pressure per se is the decisive
factor.
Beadle & Cragg investigated this problem on species of Gammarus living
naturally in sea water, brackish water and fresh water and came to the
conclusion that the ability of the animal to survive changes in the external
medium was related to its ability to maintain differences in the concentra-
tions of ions between tissues and blood as well as between blood and external
medium. More recently, Camien, Sarlet, Duchateau & Florkin (1951) and
Duchjlteau, Sarlet, Camien & Florkin (1952) have shown that there is
a distinct difference between marine and fresh-water invertebrates in the
amino-acid content of their muscles. The ammo-acid content is higher
in the marine species and makes a significant contribution to osmotic
pressure. It may be that animals which have penetrated fresh water and
have reduced the amino-acid content are unable to restore it when they
are placed in sea water; osmotic withdrawal of water will then raise the
concentration of salt in the muscles to levels which are higher than those
characteristic of purely marine species and which the living cells may not
be able to endure.
14 MOVEMENTS OF WATER AND
The converse problem is also met with; there is no doubt that the tissues
of some animals are capable of working at high water content which in other
animals would be unthinkable. Anodon has succeeded in entering fresh
water not so much by its powers of active transport as by virtue of its ability
to tolerate hydration of its tissues. The osmotic pressure of the blood of
Anodon (and, so far as is known, of its tissues) is approximately that of 5 %
sea water — a quite exceptional figure — and the general wateriness of its
tissues, to which Picken drew attention, is striking.
The moral of all this is that these are problems for the cell physiologist.
In preparing this paper I was acutely aware that the problems and ideas
which have guided research in this field have had an ecological flavour,
whereas it is a common interest in processes of active transport at the
cellular level, rather than in their ecological consequences, which brings
this symposium audience together. But it seems to me that ecology cannot
give the lead much longer and that for the future we must rather look to
cell physiology for inspiration. The zoologist who seeks to interpret
ecological distribution in physiological terms will have to concern himself
more and more with the problems of cell physiology, and it is to be hoped
that those whose interest is in the fundamental problems of all living
matter will not overlook the avenues of approach to these problems which
the invertebrates provide.
REFERENCES
BEADLE, L. C. (1934). Osmotic regulation in Gunda ulvae. J. Exp. Biol. n, 382-96.
BEADLE, L. C. (1943). Osmotic regulation and the faunas of inland waters. Biol.
Rev. 18, 172-83.
BEADLE, L. C. & CRAGG, J. B. (1940). Studies on adaptation to salinity in Gam-
marus spp. I. Regulation of blood and tissues and the problem of adaptation
to fresh water. J. Exp. Biol. 17, 153-63.
BEADLE, L. C. & SHAW, J. (1950). The retention of salt and the regulation of the
non-protein nitrogen fraction in the blood of the aquatic larva, Sialis lutaria.
J. Exp. Biol. 27, 96-109.
BON£, G.-J. & KOCH, H.-J. (1942). Le r61e des tubes de Malpighi et du rectum
dans la regulation ionique chez les insectes. Ann. Soc. zool. Belg. 73, 73-87.
CAMIEN, M. N., SARLET, H., DUCHATEACJ, G. & FLORKIN, M. (1951). Non-protein
amino acids in muscle and blood of marine and fresh water Crustacea. J. Biol.
Chem. 193, 881-5.
DUCHATEAU, G., SARLET, H., CAMIEN, M. N. & FLORKIN, M. (1952). Acides
amines non-proteiniques des tissus chez les mollusques lamellibranches et
chez les vers. Comparaison des formes marines et des formes dulcicoles.
Arch. int. PhysioL 60, 124-5.
DUVAL, M, (1925). Recherches physico-chimiques et physiologiques sur le milieu
inte*rieur des animaux aquatiques. Modifications sous I'mfluence du milieu
exte*rieur, Ann. Inst. oceanogr. 2, 232-407.
FLORKIN, M. & DUCHATEAU, G. (1948). Sur I'osmor^gulation de I'anodonte.
PhysioL comp. i, 29-45,
ELECTROLYTES IN INVERTEBRATES 15
HERRMANN, F. (1931). t)ber der Wasserhaushalt des Flusskrebses. Z. vergl.
Physiol. 14, 479-524.
HOLM-JENSEN, I. (1948). Osmotic regulation in Daphnia magna under physio-
logical conditions and in the presence of heavy metals. BioL Medd., Kbh.,
20, no. ii, pp. 1-64.
KOCH, H.-J. (1938). The absorption of chloride ions by the anal papillae of Diptera
larvae. J. Exp. BioL 15, 152-60.
KROGH, A. (i 937*2). Osmotic regulation in the frog (R. esculenta) by active
absorption of chloride ions. Skand. Arch. Physiol. 76, 60-73.
KROGH, A. (1937^). Osmotic regulation in fresh-water fishes by active absorption
of chloride ions. Z. vergl. Physiol. 24, 656-66.
KROGH, A. (1938). The active absorption of ions in some fresh water animals.
Z. vergl. Physiol. 25, 335-50.
KROGH, A. (1939). Osmotic Regulations in Aquatic Animals. Cambridge University
Press.
MALUF, N. S. R. (1939). The volume- and osmo-regulative functions of the ali-
mentary tract of the earthworm (Lumbricus terrestris) and on the absorption
of chloride from fresh water by this animal. ZooL Jb. (Abt. 3), 59, 535-52.
NAGEL, H. (1934). Die Aufgaben der Exkretionsorgane und der Kiemen bei der
Osmoregulation von Carcinus maenas. Z. vergl. Physiol. 21, 468-91.
PANIKKAR, N. K. (1941). Osmoregulation in some palaemonid prawns. J. Mar.
biol. Ass. U.K. 25, 317-59.
PANTIN, C. F. A. (1931). The origin of the composition of the body fluids. Biol.
Rev. 6, 459-82.
PETERS, H. (1935). t)ber den Einfluss des Salzgehaltes in Aussenmedium auf den
Bau und die Funktion der Exkretionsorgane dekapoder Crustaceen (nach
Untersuchungen an Potamobius fluviatilis und Homarus vulgaris). Z. Morph.
Okol. Tiere, 30, 355-8 1.
PICKEN, L. E. R. (1936). The mechanism of urine formation in invertebrates.
I. The excretion mechanism in certain Arthropoda. J. Exp. BioL 13, 309-28.
PICKEN, L. E. R. (1937). The mechanism of urine formation in invertebrates.
II. The excretory mechanism in certain Mollusca. J. Exp. BioL 14, 20-34.
POTTS, W. T. W. (1954). The energetics of osmotic regulation in brackish and
freshwater animals. J. Exp. BioL (in press).
RAMSAY, J. A. (1949). The site of formation of hypotonic urine in the nephridium
of Lumbricus. J. Exp. Biol. 26, 65-75.
ROBERTSON, J. D. (1939). The inorganic composition of the body fluids of three
marine invertebrates. J. Exp. BioL 16, 387-97.
ROBERTSON, J. D. (1949). Ionic regulation in some marine invertebrates. J. Exp.
BioL 26, 182-200.
ROBINSON, Jf R. (1953). The active transport of water in living systems. BioL
Rev. 28, 158-94.
SCHLIEPER, C. (1930). Die Osmoregulation wasserlebender Tiere. Biol. Rev. 5,
309-56.
SCHLIEPER, C. (1935). Neuere Ergebnisse und Probleme aus dem Gebiet der
Osmoregulation wasserlebender Tiere. Biol. Rev. 10, 334-60.
SCHMIDT-NIELSEN, K. (1941). Aktiv ionoptagelse hos flodkrebs og strand krabbe.
• Med pdvisning av ionoptagende celler. Biol. Medd.y Kbh., 16, no. 6, pp. 1-60.
WEBB, D. A. (1940). Ionic regulation in Carcinus maenas. Proc. Roy. Soc. B, 129,
107-35.
VERTEBRATE PHYSIOLOGY FROM THE
POINT OF VIEW OF ACTIVE TRANSPORT
BY HUGH DAVSON
Medical Research Council, Department of Physiology,
University College, London
Active transport is presumably an essential feature of the vegetative activity
of all cells, so that, in this respect, we cannot expect to observe striking
differences according as we study organisms of increasing complexity,
starting from the Protozoa, say, and finishing at the mammals ; the indivi-
dual cells of all these organisms will doubtless be shown to be capable of
a high degree of active transport, and it may well be that certain highly
differentiated cells of the more complex organism, e.g. the mammalian
erythrocyte, will exhibit active transport to a less extent, and in a less
varied form, than the Protozoa. The complex organism, however, because
of its differentiation, exhibits certain structures in which active transport is
not only necessary for their vegetative activity but also — and in a very high
degree — in virtue of their specialized functions. Outstanding examples
will spring to the mind: the stomach elaborating a solution of about
0-17 N-HC1; the kidney capable of selectively removing substances from
the blood, the intestinal epithelium capable of the rapid absorption of
selected substances from the lumen of the gut; the various glands producing
characteristic secretions, and so on. Many of these specialized activities,
involving active transport, I have had occasion to review recently (Davson,
1951), and a number of them, moreover, will be subjects of specialized and
authoritative treatment in this symposium; consequently, in the present
paper, I shall confine myself to a few general aspects of active transport
taking place in specialized tissues.
Before discussing active transport — or secretory activity — it would be
interesting and instructive to consider a form of transport in which simple
physical forces appear to be adequate for the supply of energy involved in
the process. The production of the glomerular fluid in the nephron, and
of the interstitial fluid and lymph of the voluntary musculature, are
examples. The glomerular fluid is, apparently, plasma minus the plasma
proteins; the separation of this fluid requires energy to overcome the
difference of osmotic pressure between it and its parent plasma, and this
is provided by the pressure of the blood in the glomerular capillaries, i.e.
by the mechanical work of the heart. The colloid osmotic pressure in the
VERTEBRATE PHYSIOLOGY AND ACTIVE TRANSPORT 17
mammal is of the order of 30 mm. Hg, and there is little doubt that the
glomerular capillary pressure is not only adequate to effect this separation
of the plasma proteins, but also to provide the pressure-head necessary to
maintain a continuous flow against the frictional resistance of the tubules.
The separation implies, however, a membrane capable of holding back the
proteins of plasma whilst permitting a ready flow of water and the smaller
solute molecules of blood plasma; such a membrane is presumably given
by the capillary walls, the intercellular spaces being sufficiently small to
prevent — under normal conditions at any rate — the serum albumin and
globulin molecules from passing through, but sufficiently large to allow
inulin, gelatin and egg albumin to pass. The evidence in support of this
intercellular route is largely presumptive; it is argued that it is unlikely
that cellular membranes would show such a low level of discrimination as
to permit the passage of all the non-colloidal constituents of the blood at
the same rate, and, moreover, would permit substances of high molecular
weight such as inulin to pass. In the case of the muscle capillaries an
intercellular route for the flow of tissue fluid has been postulated on similar
grounds; since, in this case, there is a very definite 'leakage' of proteins,
we must assume either that the intercellular spaces are larger, or that the
glomerular membrane of Bowman's capsule acts as a second, and more
efficient, filter to ensure that only minimal amounts of protein find their
way into the tubules. The evidence that the glomerular fluid is, indeed,
nothing more than a filtrate from plasma, i.e. that no active transport
mechanisms are involved in determining the relative concentrations of
dissolved material in it and its parent fluid, is based on chemical analyses
which, because of the very small amounts of fluid available, were probably
not accurate to within less than ± 10%, although the large number of
determinations carried out, and the absence of any trend indicating active
transport mechanisms, make for a convincing body of evidence in favour
of this simple origin of the glomerular fluid (Richards, 1938). The evidence
with regard to the intercellular fluid of muscle is by no means so impressive,
in fact I only know of one analysis of the relative compositions of plasma
and this fluid, namely, that of the chloride distribution by Maurer (1938),
so that the general physiologist bases his assertion that the capillary
membrane exerts no active transport between blood and tissue fluid
largely on the belief that the phenomena of fluid exchange between inter-
cellular space and plasma are explicable on simple mechanical considera-
tions (Danielli, 1940; Landis, 1934; Pappenheimer, J. R. & Soto-Rivera,
I948).
I raise this point not with the intention of shaking belief in the general
proposition, but rather to show how inadequate such chemical evidence
l8 VERTEBRATE PHYSIOLOGY FROM THE
would be if it were desired to show that other tissue fluids, namely, the
aqueous humour and cerebrospinal fluid, were likewise formed by simple
ultrafiltration mechamisms. Thus an analysis of the main constituents of
plasma and aqueous humour, e.g. Na and Cl, carried out within the limits
of accuracy considered adequate for the study of the glomerular filtrate,
would indicate an excellent agreement between theory and experiment,
regarding the aqueous humour as a blood filtrate. Thus the ratio of the
concentrations of sodium (Na)pl /(Na)Aq was 1-03 and that for chloride
(Cl)pi./(Cl)Aq> equal to 0-955, comparing with ratios, theoretically computed
from the known base-binding power of the plasma proteins, of 1-04 and
0-96 respectively (Davson, 1939). Such a concordance was, indeed, so
convincing that for some years I was ready to believe that the aqueous
humour was, indeed, a plasma ultrafiltrate. However, the appearance of
evidence against this view made me reopen the question ; thus a deviation
of i % from the equilibrium distribution of sodium, in this case, could be
of profound significance; it could mean, for example, that the aqueous
humour contained i % more NaCl and NaHCO3 than the blood plasma,
a difference in concentration capable of maintaining a difference of osmotic
pressure of some 60 mm. Hg, a by no means insignificant contribution to
the forces driving fluid into the eye. The weak point in the work, however,
was not the accuracy of the chemical analysis, which was easily high
enough to permit the detection of a i % discrepancy, but the assessment of
the theoretical Donnan distribution of Na and Cl for a dialysate of blood
plasma, since the value of 1-04, given by Van Slyke (1926), postulates
equality of activity coefficients in the two fluids.
Clearly the best way of investigating the matter would be to dialyse
aqueous humour against plasma from the same animal, and see if there is
any migration of Na and Cl from one fluid to the other. The results for the
cat are shown in Table i, the ratios for Na and Cl being determined before
and after dialysis (Davson, Duke-Elder & Maurice, 1949). It will be seen
that there is, indeed, a migration of both Na and Cl from the aqueous
humour to the blood plasma; the true distribution ratios for a dialysate
turned out to be 1-07 and 0-97 for Na and Cl and not 1*04 and 0-96 as
computed by Van Slyke. This excess of salt in the aqueous humour is
small, from the point of view of chemical analyses, and would have been
quite undetectable by the methods used for the study of the glomerular
fluid; nevertheless, it is large enough to influence the intra-ocular pressure
and to rule out a simple filtration mechanism for the origin of the aqueous
humour. I shall be returning to the problem of the aqueous humour later;
for the moment I merely wish to emphasize the importance of, and
difficulty in, determining the existence of active transport in certain
POINT OF VIEW OF ACTIVE TRANSPORT 19
systems. For many of the contributors to this symposium this has long
ceased to be a problem, e.g. the transfer of salt by the frog's skin, the
absorption of sugars from the intestine, and so on; and the problems have
resolved themselves into determining the mechanism whereby the metabolic
energy of the cell is made available for osmotic work. In the case of the
aqueous humour and cerebrospinal fluid the problem has consisted
primarily in demonstrating the existence of active transport mechanisms
in the elaboration of these fluids.
Table i . Effect of dialysing aqueous humour against blood
plasma on the distribution ratio of Na and Cl
(Na)P,/(Na)Aq.
(Cl)PJ/(Cl)Aq
Before dialysis
After dialysis
1-042
i -068
0-945
0-971
A rather similar problem will doubtless arise with many of the more
obvious forms of active transport; for example, we shall have to differentiate
between the ' accidental' and the 'essential' in the composition of many
secretions. Thus the obvious feature of the gastric secretion is the high
concentration of hydrogen ions; the concentration of potassium is, how-
ever, about twice that of the plasma from which the secretion must ulti-
mately be derived ; the concentration of calcium is only about a tenth that
in the plasma (Gudiksen, 1943). We must ask whether these differences
are essential, in the sense that active transport mechanisms are operating
on these ions, or whether they are the result of activity directed towards
the hydrogen or chloride ion. A similar and more urgent problem, of
course, arose with muscle, nerve and the erythrocyte. The evidence
indicates that the active transport of sodium out of the muscle and nerve
fibres is adequate to account for the accumulation of potassium, because
the high internal concentration of non-permeating anions demands the
replacement of the excreted sodium. It was originally suggested that the
extrusion of sodium could explain the accumulation of potassium in the
human erythrocyte (Dean, 1941 ; Maizels, 1949), but, as I argued elsewhere
(Davson, 1951), this is to ignore the circumstance that the erythrocyte does
not have the same high concentration of non-permeating anions ; the extru-
sion of sodium would therefore only permit a limited accumulation of
potassium, and to explain the observed accumulation an active transport of
this ion must also be postulated (Harris & Maizels, 1952).
To come now to a more general aspect of secretory activity in specialized
tissue, we may note that the outstanding feature of this activity — as
contrasted with the vegetative activity in single cells — is the transport of
20 VERTEBRATE PHYSIOLOGY FROM THE
material across an organized tissue; thus the accumulation of potassium by
the erythrocyte or the extrusion of sodium by a muscle or nerve fibre, are
processes that concern only the inside of the cell and a surrounding
medium that may be considered homogeneous. Where transport across
a tissue is concerned, we are dealing with an essentially asymmetrical
system in which the medium surrounding the cells must be divided into
two specific regions — the donor region, from which the actively tranported
material is extracted, and the acceptor region, into which the actively
transported material is driven. Between the two we have the cells capable
of supplying the necessary metabolic energy. The asymmetry of this
system must, in the last analysis, reside in the asymmetrical activities of the
individual cells of the tissue, and it is worthy of note here that an important
element in this asymmetry may be the organization of the cells in a definite
layer; thus Chambers & Kempton (1933) showed that isolated cells of the
chick mesonephros showed no evidence of accumulation of phenol red,
whereas when organized in * cysts' they did so. Viewing active transport,
in these specialized tissues, as a transfer across an organized cellular
structure, we must ask next whether the substances are indeed transported
through the cells, and if so, whether they are accumulated to any extent
within them. Again, we must pay attention to the role played by the spaces
between the active cells; this role will, of course, be passive, but it is
important to know to what extent the activities of the cells are favoured or
prejudiced by the existence of regions in the tissue where diffusion may be
as rapid as that observed in aqueous solution.
The transport of the secreted substance through the active cell is highly
probable on a priori grounds — the metabolic activity of the cell depends on
enzymes that are within it and, if chemical work is to be performed on
a given molecule or ion, it seems reasonable to conclude that the molecule
or ion must penetrate the cell to participate in the energy transformation.
Nevertheless, it is worth pointing to an example of metabolic activity that
seems to be located at the surface of the cell, namely, glycolysis of the
erythrocytes of certain species ; since glycolysis seems to be the basis for the
energy available for the active transport of ions across the erythrocyte
membrane, this point is not entirely irrelevant to the discussion. Wilbrandt
(1938) showed that the permeability of the dog and rabbit erythrocytes to
glucose was so small as to preclude the possibility of its metabolism within
the cell. In the case of the rabbit erythrocyte this view seems to be borne
out by studies of the effect of fluoride. This inhibitor of glycolysis actually
causes a very rapid escape of potassium from the erythrocyte of this species ;
the effect seems to be dependent on the accumulation of intermediary
products of metabolism, since it can be prevented by adding mustard gas
POINT OF VIEW OF ACTIVE TRANSPORT 21
which inhibits glycolysis in its initial stage, and it can also be prevented by
removing from the system the necessary substrate constituents — glucose,
phosphate, calcium, magnesium and potassium ; the return of these consti-
tuents to the system causes the escape of potassium without any evidence of
a delay due to the necessity for the magnesium, glucose, etc., to penetrate
the cell. At what stage, if any, the glycolytic process becomes intracellular
is not known, but even if the entire series of chemical reactions took place
on the cell surface the energy liberated could presumably be made available
for active transport if we accept Goldacre's mechanism for this process.
In certain cases the active transport of substances through the cells of
the active tissue has been unequivocally proved; thus Chambers & Kemp-
ton, in the work previously alluded to, have demonstrated the presence of
phenol red in the epithelial cells of mesonephros cysts during the process
of accumulation; any real active transport, i.e. the transfer of phenol red
against a gradient of electrochemical potential, was only definitely proved
in the direction, cell to lumen, so that it may well be that in the donor
region the passage into the cell is a matter of simple diffusion. Again, the
secretion of HC1 by the parietal cell of the stomach may be regarded
essentially as a transfer of acid from the outside, donor, medium to the
acceptor region in the canaliculi of the parietal cell; the intermediate
accumulation of acid in the cytoplasm of the parietal cell must be ruled
out by the observations of Bradford & Davies (1950). We may thus regard
the secretory process, taking place across such specialized tissues as the
tubular epithelium of the kidney or the gastric mucosa, as the penetration
of certain substances into the active cells and their expulsion at another
region; this latter process unequivocally represents active transport, but
whether the former process, namely entry into the cells, involves any
metabolic activity will depend on an analysis of the contents of the cell and
the outside donor medium.
In this connexion we may note that many substances that are subjected
to active transport are not substances that would be expected to cross the
plasma membrane of a cell with any ease, e.g. glucose, or the sodium ion.
It seems very likely to me, however, that a specialized form of permeability,
not to be confused with active transport, will come into play in these cases.
It was Danielli who first pointed out that glycerol penetrates into certain
erythrocytes with a speed out of all proportion to what would be expected
on the basis of measurements on comparable molecules, e.g. ethylene
glycol, and he suggested the presence of active patches in the membrane
in which the activation energy necessary for penetration was very low.
About the same time I observed that, when cat erythrocytes were sus-
pended in isotonic KC1, sodium leaked out with a permeability constant
22 VERTEBRATE PHYSIOLOGY FROM THE
very much higher than that for the penetration of the smaller potassium
ion; moreover, the permeability to sodium exhibited an optimal tempera-
ture in the region of 37° C., an optimal pH in the region of 7-4, and was
-markedly inhibited by narcotics, heavy metals, soaps, etc. The permeability
to potassium, under identical conditions, was much more * orthodox',
being only mildly accelerated by narcotics and soaps, etc., and exhibiting
a continuous increase with increasing temperature without any sign of an
optimum. It was suggested (Davson & Reiner, 1942) that the permeability
to sodium was mediated by an enzyme-like grouping that lowered the
activation energy for penetration of the membrane; in other words, that the
cell membrane had become specialized to permit the rapid migration of
the sodium ion. A similar type of membrane specialization is probably
at the basis of the extremely rapid exchanges of anions observed in the
erythrocyte, a specialization that permits of the rapid acid-base exchanges
in the blood when exposed to alveolar air for the short time available.
If I have understood Le Fevre's (1952) work correctly I would suggest
that in the transport of hexoses across the erythrocyte membrane we have
another example of this catalysed or specialized permeability. The essential
feature of this permeability is that it is higher than what would be expected
of an undifferentiated lipoid membrane, but it is a permeability that is
observed with substances passing from a region of higher to one of lower
electrochemical potential, i.e. active transport mechanisms need not be
invoked. On the other hand, the specialization seems to take the form of
an enzyme-like differentiation of the cell surface, so that the permeability
is inhibited by narcotics, heavy metals, small shifts in pH, and so on. If
this type of permeability is involved in the passive transfer into the secretory
cells we may expect the active transport mechanism as a whole to be
affected by narcotics, enzyme poisons, etc., even though the actual meta-
bolic systems may not have been affected. This consideration must always
be borne in mind when considering the action of enzyme poisons on active
transport.
We have raised the question of the intercellular spaces in so far as
secretory activity is concerned; we have asked whether their presence
would be detrimental or otherwise to the process of active transport. The
obvious answer is that they would be detrimental to any transfer of material
against a gradient of electrochemical potential, in so far as they permitted
back-diffusion from the acceptor to the donor region. We may consequently
expect the intercellular spaces of a secretory epithelium to be small by
comparison with the area of the cells. Thus, in the kidney tubule, glucose
may be reabsorbed until there is no detectable concentration of this
substance in the tubular fluid ; this would suggest, either that back-diffusion
POINT OF VIEW OF ACTIVE TRANSPORT 23
is impossible — the intercellular spaces being too small to permit the mole-
cule to penetrate — or, what seems more probable, that the rate of back-
diffusion is too small to affect appreciably the concentration of glucose in
the tubule in the face of the rapid process of active transport. Further
evidence supporting the view that intercellular exchanges are not very
significant has been provided by the work of Hober (Schmengler & Hober,
1933 ; Hober, 1933) on the frog kidney using the dual perfusion technique;
according to these results, none of the sugars — glucose, galactose, fructose,
etc. — pass from the blood to the tubular fluid when perfused by way of the
renal portal vein, i.e. when presented only to the tubules. In those cases
where passive diffusion from blood to tubular fluid appears to take place,
e.g. with urea, thiourea, etc., it would seem that lipoid solubility is a
prominent factor, indicating that this back diffusion is predominantly
transcellular. Where secretory activity results in a marked difference in
osmotic pressure between the parent fluid and the secreted fluid, the
problem of back-diffusion — whether it be by way of extracellular spaces or
across the cells of the secretory tissue — raises an interesting problem.
In general we observe active transport directed towards substances that
penetrate cells slowly, e.g. ions, sugars, and amino-acids; highly lipoid-
soluble substances are generally not transported actively. This is under-
standable, since the work done in maintaining a gradient of electrochemical
potential depends directly on the permeability constant of the molecule
concerned ; to maintain, for instance, a concentration ratio of 20 between
the inside and outside of the erythrocyte, the energy requirement would
be of the order of a million calories per kg. per hour if the substance
concerned were urea, far beyond the metabolic potentialities of the erythro-
cyte or of any other cell. From energetic consideration alone, therefore,
we may expect secretory activity to be manifested towards slowly pene-
trating substances. The permeability of cells to water, is, in general,
extremely high, much higher than the permeability of the erythrocyte to
urea, for instance (Davson & Danielli, 1952; Collander, 1949), so that,
where secretory activity results in a marked difference of osmotic pressure
between donor and acceptor fluids, we must expect a modification in the
cell membranes of the secreting cells in the direction of a reduced perme-
ability to water, otherwise it would be impossible to maintain the difference
of osmotic pressure. Thus the distal tubule of the mammalian kidney may
establish a difference in concentration of the order of 1-2 M salt, equivalent
to an osmotic pressure of the order of 50 atm. ; again, the salivary secretion
is strongly hypotonic, maintaining a difference of osmotic pressure of some
7 atm. ; and in both instances the secreted fluid is separated from what is
presumably a blood-isotonic fluid by only a single layer of cells. I know of
24 VERTEBRATE PHYSIOLOGY FROM THE
no study in which the permeability to water of these cellular layers has
been examined, and it would certainly be of interest to compare, say, the
proximal and distal tubular epithelia from this aspect. The phenomenon of
the maintenance of a large difference of osmotic pressure across a secreting
tissue emphasizes once again, moreover, the limited area of the intercellular
space in this tissue.
So far we have considered the intercellular space from a purely negative
aspect; it is worth asking, however, whether active transport could con-
ceivably take place through this space, i.e. essentially over the surface of
the active cells, as opposed to through them. This, of course, is pure
speculation, but I raise the matter for what seems to me to be a good
reason, namely, the observation of secretory activity across multiple layers
of cells, as in the plant root, the frog's skin, and the ciliary epithelium. The
absorption of KNO3 by the plant root is a process of active transport
between the epidermal cells, which remove the salt from the soil or other
nutrient medium, and the stele into which it is finally exuded as root-sap.
Between this epithelium and the stele, however, there are successive layers
of cells, and we have to consider whether the salt is actively transported by
one cell-layer, excreted into the interstitial fluid in contact with the next
layer, actively transported through the next layer, and so on, until it is
finally exuded into the stele. This would appear to be a most inefficient
process, involving separate acts of transport by each successive cell-layer.
If a mode of extracellular active transport could be imagined, and at
present I refrain from drawing any picture of a hypothetical mechanism,
it might well provide a more efficient mechanism for the transport through
successive cellular layers than one based on the more conventional view of
active transport through cells.
The ciliary epithelium in the eye seems to present a similar problem,
since it is made up of two layers of what appear to be secretory cells.
Aqueous humour — essentially a solution containing the non-colloidal
plasma constituents of which the main cation is sodium — appears to be
secreted continuously from the cells of this tissue; the actual mechanism
whereby the fluid is driven out of these cells is still a matter of speculation.
In line with present thought on the active transport of sodium, we may say
that the secretory cells actively extrude this ion from their inner surfaces
into the acceptor region (the posterior chamber) ; as a result of this extrusion
anions and water follow and the remaining constituents, e.g. potassium,
sugar, urea, etc., may follow by simple diffusion from the secretory celk or
through the intercellular spaces. If this is the essential basis, once again we
must postulate a double process, involving secretory activity by the two
layers in series. If the secretory cells contained normally a high internal
POINT OF VIEW OF ACTIVE TRANSPORT 25
concentration of sodium the work done in forming the aqueous humour
would, of course, not be very high, and it would not be of much energetic
significance if the process of formation were repeated by the second cellular
layer.
One generally assumes that cells have a high internal concentration of
potassium, in which case sodium must be excreted out of the cell against
a high gradient of electrochemical potential and repeated excretion would
have much greater energetic significance. It is worth remembering, how-
ever, that certain cells actually contain sodium as their predominant cation,
e.g. the erythrocytes of the cat and dog. It would certainly be interesting
to determine the potassium content of the cells of the ciliary epithelium.
One hypothesis that I entertained for some time was that the aqueous
humour was essentially an ultrafiltrate from the blood plasma, forced
between the cells of the ciliary epithelium, and that these cells modified the
filtrate by excreting sodium into it to produce finally a hypertonic fluid as
found experimentally. If such were indeed the mechanism of formation of
the fluid, the energy requirements would be low, the route of penetration
of the various constituents being almost completely extracellular. On
injecting various substances into the blood and measuring their rate of
appearance in the aqueous humour we should not expect to find any
marked differences in rates, since the theory postulates essentially a bulk
flow through intercellular spaces. A detailed examination of this * blood-
aqueous barrier* revealed just the opposite, however, the rates of penetra-
tion of such substances as glucose, urea, sucrose, amino-acids, creatinine,
etc., being so markedly different as to indicate that these molecules must
pass a highly selective barrier — such as could be constituted by the ciliary
epithelial cells — before penetrating. It seems very likely, therefore, that
the aqueous humour is, indeed, elaborated within a cellular tissue and
extruded from this as a characteristic secretion. It would appear from the
studies of the secretory epithelia that they constitute, from the point of view
of passive permeability, exceptionally tight barriers to diffusion of sub-
stances that normally pass through cells with great difficulty; in other
words, that the intercellular spaces must be exceptionally small and by no
means comparable with the extracellular spaces of the capillary endo-
thelium.
The eye does, indeed, present an apparent exception which on further
investigation seems to * prove the rule'. Sucrose, raffinose and plasma
proteins actually do penetrate the blood-aqueous barrier; because of the
rapid drainage away of the aqueous humour, through non-selective channels,
back into the blood, the concentration of, say, serum albumin, is only a
small fraction (about i %) of the concentration in the plasma; nevertheless,
26 VERTEBRATE PHYSIOLOGY AND ACTIVE TRANSPORT
this indicates a definite leak, presumably through intercellular spaces, of
molecules of very high molecular weight. The evidence indicates, however,
that the leak does not occur through the secretory ciliary epithelium but
rather through the anterior surface of the iris over which the aqueous
humour flows on its way out through Schlemm's canal. The rates of pene-
tration of such high-molecular weight substances as have been examined,
e.g. sucrose and raffinose, suggest that the substances pass through holes
large by comparison with the size of the penetrating molecule (Davson &
Matchett, 1953). It is probably in this manner that various antibodies and
enzymes find their way into the aqueous humour. To speak teleologically,
it is in this manner that the physiological disadvantages of a highly selective
barrier, such as is constituted by a secretory epithelium, are overcome.
REFERENCES
BRADFORD, W. M. & DAVIES, R. E. (1950). Biochem.J. 46, 414.
CHAMBERS, R. & KEMPTON, R. T. (1933). jf. Cell. Comp. Physiol. 3, 133.
COLLANDER, R. (1949). Physiol. Plant. 2, 300.
DANIELLI, J. F. (1940). J. Physiol. 98, 109.
DAVSON, H. (1939). J. Physiol. 96, 194.
DAVSON, H. (1951). Textbook of General Physiology. London: Churchill.
DAVSON, H. & DANIELLI, J. F. (1952). The Permeability of Natural Membranes.
Cambridge University Press.
DAVSON, H., DUKE-ELDER, W. S. & MAURICE, D. M. (1949). J. Physiol. 109, 32.
DAVSON, H. & MATCHETT, P. A. (1953). J- Physiol. 122, n.
DAVSON, H. & REINER, J. M. (1942). J. Cell. Comp. Physiol. 20, 325.
DEAN, R. B. (1941). Symp. Soc. Exp. Biol. 3, 331.
GUDIKSEN, E. (1943). Ada physiol. scand. 5, 39.
HARRIS, E. J. & MAIZELS, M. (1952). J. Physiol. 118, 40.
H5BER, R. (1933). Pflug. Arch. ges. Physiol. 233, 181.
LANDIS, E. M. (1934). Physiol. Rev. 14, 404.
LEFEVRE (1952). J- Gen. Physiol. 35, 891.
MAIZELS, M. (1949). J. Physiol. 108, 247.
MAURER, F. W. (1938). Amer. J. Physiol. 124, 546.
PAPPENHEIMER, J. R. & SOTO-RIVERA, A. (1948). Amer. jf. Physiol. 152, 471.
RICHARDS, A. N. (1938). Proc. Roy. Soc. B, 126, 398.
SCHMENGLER, F. E. & HftfiER, R. (1933). Pflug. Arch. ges. Physiol. 233, 199.
VAN SLYKE, D. D. (1926). Factors Affecting the Distribution of Electrolytes, Water
and Gases in the Animal Body. Philadelphia.
WILBRANDT, W. (1938). Pfliig. Arch. ges. Physiol. 241, 302.
THE CONCEPT AND DEFINITION OF
ACTIVE TRANSPORT
BY THOMAS ROSENBERG
Steno Memorial Hospital and Nordisk Insulinlaboratorium,
Gentofte, Denmark
A glance at the titles of the contributions to this Symposium indicates that
the problem of * active transport' has become of considerable interest, not
only for its own sake but also in relation to general problems of intermediary
metabolism. New topics, such as nerve stimulation and recovery (Hodgkin,
1951), have been included, and certain observations (Lehninger, 1951)
indicate that very similar phenomena of transport across membranes also
occur in intracellular particles, especially mitochondria. In spite of the
importance of the problem and the large amount of excellent experimental
work which has been devoted to it recently, its detailed mechanism has
been elucidated in scarcely a single case. The reasons for our meagre know-
ledge at the present moment probably lie mainly in the difficult nature of
the problems and in the experimental inaccessibility of the systems which
have been investigated. However, one must also consider the possibility
that we have so far formulated our questions unfavourably owing to un-
suitable basic concepts. This is suggested by the fact, among others, that
at present active transport cannot be sufficiently characterized by numbers,
and there is no clear definition of its concept which might serve as a base
for a measure. Even if most workers in this field feel rather clearly what is
meant by this concept, it is not primarily a phenomenon, pre-existing in
nature, which can be recognized without being defined. There have been
numerous analogous situations in the history of science. Carnot, in 1824,
recognized rather clearly the physical meaning of the second law of thermo-
dynamics, but he could not arrive at a definition of the concept of entropy.
This was done 26 years later by Clausius, who could then give the general
formulation of this law. I think that one should not regard questions of
terminology as a mere formality. Hazy definitions and fundamentals are
not only signs of the incompleteness of our knowledge, but also often the
main obstacles to attempts to gain further theoretical and practical insight.
Fundamental definitions have varying functions in the study of a group
of allied phenomena. In the early stages they facilitate the collection of the
necessary empirical data by differentiating between superficially similar
phenomena. In later stages they assist in the formulation of models and in
28 THE CONCEPT AND DEFINITION OF ACTIVE TRANSPORT
the comparing of analogous groups of phenomena and thus in the under-
standing of the basal mechanisms. In the final stages they render possible
the immediate recognition and classification of relevant phenomena in the
study of new systems and the complete cataloguing of the whole group of
phenomena. I think that as regards the problem of active transport we are
in the second stage, and that therefore the replacement of more or less
diffuse concepts by clearer definitions is of special importance. I would
like, in connexion with these considerations, to quote a passage from Irving
Langmuir: 'The progress of modern science depends largely upon (i)
giving to words meanings as precise as possible ; (2) definition of concepts
in terms of operations; (3) development of models (mechanical or mathe-
matical) which have properties analogous to those of the phenomena which
we have observed' (Langmuir, 1929).
Early in the development of this field active transport has been con-
trasted with diffusion, which signifies the movement of a substance along a
concentration gradient by reason of the thermal movement of the molecules.
The quantitative aspects of the latter phenomenon had been worked out by
Pick (1855), who formulated the so-called diffusion laws. Although the
movement of a large number of substances across cell walls could be
expressed satisfactorily by the diffusion equations, it soon became evident
that these equations were not applicable to the movement of all substances
in the living organism. The most obvious exceptions were found among the
most important cell metabolites such as carbohydrates and amino acids,
also cations and anions, water, and a number of other substances. For such
exceptions the expression * active transport* was coined, which was designed
to convey the idea of the active participation of the cell in the movement of
the substance. It includes the concept that the cell can use part of the
energy derived from metabolism in regulating and influencing the rate and
direction of transport. The significance of such an influence can be manifold:
accumulation of certain substances within or without the cell in order to
create optimum living conditions; the preservation of substances which
are of importance to the cell or to the larger organism ; or the elimination at
increased rate of toxic substances or metabolic end-products.
If one tries to describe somewhat more closely such concepts as were
gradually combined under the general heading of 'active transport',
neglecting vitalistic considerations, one might say: by active transport is
meant the transport of substances across one or more cell membranes which
is influenced not only by the force responsible for passive diffusion, but
also by other forces which are maintained and regulated by the metabolism
of the cell.
The nature of the forces implied by such a description is unknown, as is
THE CONCEPT AND DEFINITION OF ACTIVE TRANSPORT 2Q
their connexion with metabolism, and the question arises: when can one
conclude that the transport of a substance is merely due to the force of
diffusion? Generally it is difficult to reach a conclusion, partly because
Pick's equations are not expressions of absolutely valid laws, but apply to
ideal limiting conditions, and partly to our fundamental ignorance of
conditions within the membrane. This forces us to make use of finite
differences of concentrations instead of concentration gradients and to
neglect structural properties of the membrane. For the estimation of the
errors which are thus introduced experimental data are generally lacking.
Pick's well-known first diffusion equation is
si=~D^' (0
where st is the amount of substance i which is transported per unit time
across the cross-sectional area A normal to the direction of diffusion, D{ its
diffusion coefficient and dc^dx its concentration gradient. This equation
was originally set up by analogy to heat conduction rather than derived from
first principles. Only considerably later did van Laar (1907) and Einstein
(1908) work out derivations which under certain assumptions lead to (i).
These derivations were based on the introduction of a force, the force of
diffusion, which can be identified, in modern terminology, with the negative
value of the chemical potential gradient ( — dfijdx). The term 'force of diffu-
sion ' will be used in this sense here when applied to uncharged components.
Assuming infinitely dilute solutions, expressed by d^ — RTdlnc^ and a
high resistance, one obtains an equation of the form (i); when not limiting
the case to ideal conditions, the diffusion equation has to include an additional
term with the gradient of the activity coefficient. In connexion with these
derivations it is to be noted that the application of terms like force and
resistance which are taken from analogous mechanical processes cannot be
regarded as justified a priori, but needs the confirmation of empirical and
statistical-thermodynamical methods. An extended diffusion theory, which
also considers the effect of other forces, has been presented by Onsager
(1931, 1945).
With regard to experimental evidence it is often taken as indicative of
active transport, if one of the following factors exerts an effect different from
that which it would be expected to exert on normal diffusion :
1 i ) Concentration. The rate is no longer a linear function of the difference
in concentration on both sides of the membrane ; especially at higher con-
centrations saturation phenomena may occur.
(2) Competition by chemically similar substances. This is analogous to the
just-mentioned saturation phenomena.
30 THE CONCEPT AND DEFINITION OF ACTIVE TRANSPORT
(3) Temperature. The temperature coefficient may be unusually high,
of the same order as that of enzymic reactions.
(4) Slight structural modifications of the penetrating substance. The rate
can be completely different for structurally related substances, even those
with similar molecular size or lipoid solubility, for instance, optical
isomers.
(5) Intensity of metabolism. The penetration may often be dependent
on simultaneous supply of oxygen.
(6) Effect of enzyme activators and inhibitors. In certain cases penetration
can be completely inhibited by enzyme poisons.
Some of these phenomena show that the penetration of certain substances
can be subject to regulatory mechanisms. Such mechanisms, however,
need not be equivalent to the participation of additional forces or of the
energy yielded by metabolism. One has to remember that although changes
in the rate of penetration can be due to changes in the driving forces, they
can equally well be due to changes in membrane resistance which depends
on its structure. This structure in turn might depend largely on such
factors as concentration, temperature, and occurrence of enzymic reactions
remaining normal. Perhaps the biologist is often interested primarily in
the existence of such control mechanisms and less in the question of whether
they are due to changes in the resistance or of the driving forces. Since,
however, there operate completely different mechanisms in the two cases,
it seems desirable to differentiate between them.
It should also be mentioned — as Danielli (1943) has pointed out — that
an 'abnormal* dependence of the rate of penetration on temperature and
molecular size of the penetrating substance is to be expected if the passage
of phase boundaries and microdiscontinuities are rate-determining steps.
If at such phase boundaries only a limited number of free places is available
to the diffusing substance, one might expect both saturation phenomena
and competition. Dr Wilbrandt and I (unpublished) have calculated the
case of a model with such an adsorption layer, and have found that in the
case of the human erythrocyte the dependence on concentration of the
penetration of glucose would not be inconsistent with such a structure.
For other reasons, however, a carrier mechanism was suggested in this
case. In addition, there is the often discussed possibility of a mosaic-like
structure of the membrane with transport paths of varying chemical and
structural specificity. Finally, it should be noted that even the dependence
of penetration on enzymic reactions is no definite proof for its dependence
on metabolism. It is possible, for instance, that a substance at the mem-
brane undergoes an enzymic transformation independent of cell metabolism
to an isomeric molecule, which might then be the one actually transported.
THE CONCEPT AND DEFINITION OF ACTIVE TRANSPORT 31
A method of testing for passive penetration, introduced by Ussing
(1949, 1951), depends on new principles. It involves the application of
two different isotopic forms of the diffusing substance, and has already
yielded very valuable results, especially regarding the behaviour of in-
organic ions. However, this method also has its limitations in the applica-
tion to the present problem, as has also been pointed out by Ussing. I will
not discuss the practical difficulties which arise when the method is applied
to a substance, the participation of which, in metabolism cannot be neglected,
but also for theoretical reasons it only gives an unequivocal answer to the
present question if the substance under investigation does not undergo
complex formation or any other interaction within the membrane either
with itself or with other mobile membrane constituents. In such cases the
diffusion currents of the two isotopes are not independent of each other.
Thus in the case of the simple diffusion of benzoic acid through a layer
of benzene we would find a deviation from the test equations for passive
penetration, since in benzene benzoic acid occurs mainly associated as
double molecules. An examination of the applicability of these methods
to model membranes would appear to be of great interest.
The demonstration of transport from a lower to a higher potential
(uphill transport) is a certain indication of the participation of forces other
than of diffusion, and in my opinion, at the present time such a demonstra-
tion is the only certain criterion of active transport considering our ignorance
of membrane structure and the available methods. For that reason I have,
in an earlier discussion (Rosenberg, 1948), limited the definition of active
transport to such cases. Even if such a definition appears as too narrow for
many biological purposes it has several advantages, so that in every case a
special treatment for uphill transport appears desirable. Thus one can draw
conclusions as to a general mechanism by considering the numerous non-
biological cases of transport against potential gradients. It is also possible
to express a given uphill transport in terms of the amount of substance
transported and the difference in potential. Finally, the demonstration of
uphill transport is based solely on experimental evidence without re-
quiring assumptions regarding membrane structure or mechanism.
A broader definition could be based on the above-mentioned description
of active transport and formulated in the following manner: active trans-
port is the movement of a substance which is influenced by other forces in
addition to the chemical (or analogous) potential gradient of this substance.
An advantage of this definition would be that it would roughly cover the
usual concept and permit a theoretical treatment, whereas a disadvantage
would be the difficulty of the experimental determination of whether a
given transport is active if it is not an uphill transport.
32 THE CONCEPT AND DEFINITION OF ACTIVE TRANSPORT
I would further like to add that the thermodynamical and statistical
treatments of irreversible phenomena during the last two decades have
greatly improved our understanding of the problems of the transport of
matter. Thus a number of systems which do not follow Pick's equations
san be given a satisfactory and quantitative treatment by the above-
mentioned theory of Onsager (1945), and also Bronsted's (1946) concepts
have led to good results. These same theories, however, show us the limita-
tions to which the applicability of all treatments up to now is subject. Thus,
for instance, the integration of the relevant differential equations can only
be carried out, if all conditions prevailing along the transport path are
known. In Onsager's theory, values of potential difference and flow must
be known for all quantities. Such theories are therefore of only limited
value for the quantitative treatment of membrane systems of unknown
structure with flows of unknown nature. A further limitation is that such
theories are only valid for those systems which are not far removed from
equilibrium. Transports which are induced by sudden fundamental
structural changes of short duration can apparently not yet be treated in a
satisfactory theoretical manner. Further, the possibility might be con-
sidered of whether in some case the very observation of a membrane
transport is able to induce changes in the factors affecting the transport
which cannot be neglected. Such a situation is known from atomic physics
and is there expressed by the so-called uncertainty relations.
I will not consider these limitations in what follows and discuss a few
somewhat simplified and schematic models in order to approach somewhat
more closely the problem of the nature of the additional forces and their
connexion with metabolism. Problems of the mechanism of the coupling
with metabolic processes not only play a role in the case of active transport
but also in the case of many analogous problems, e.g. muscular contraction
or the formation of energy-rich phosphate bonds. In such cases one would
often first ask the question: with which part of the total metabolism are
these processes coupled? In the case of muscular work the required energy
is evidently derived from the metabolism of carbohydrate. Such a con-
clusion does not, however, lead us to the mechanism of the coupling, for,
as is known from the work of Lundsgaard (1930), the muscle can also work
without the simultaneous utilization of carbohydrate as long as a reservoir
of energy-rich phosphate bonds is available. We can illustrate the situation
by means of a general scheme (Fig. i). Let this scheme illustrate a metabolic
reaction chain, for instance, the metabolism of glucose, which leads from
the initial compounds Gx via a number of more or less reversible reactions
to the end-products Gn. Branch chains are linked to this main chain by
coupling mechanisms, the nature of which is generally not known and
THE CONCEPT AND DEFINITION OF ACTIVE TRANSPORT 33
which are represented by cog-wheels. For as soon as the nature of such
mechanisms is known, they can be described by chemical equations or
transports. That is due to the fact that from an energetic point of view a
chemical substance can only either be transformed chemically or trans-
ported. The coupling between coenzyme oxidation and phosphorylation
of ADP is thus at the present moment still to be represented by a cog-wheel,
whereas the corresponding phosphorylation during the oxidation of
phosphoglyceraldehyde can be largely represented by chemical equations
(Warburg & Christian, 1939; Racker & Krimsky, 1952).
Assume that in such a branch chain there is a certain active transport,
e.g. A5 -> AQ. The conclusion that this transport is dependent on the
metabolism of glucose is of course correct and can be important. However,
it leads to no clue as to the nature of the link and the mechanism of the
Fig. i.
coupling. We can, however, consider this mechanism more closely by
studying the process A$ -> A^ which yields directly the energy for the active
transport. Although, of course, no specific information can be given con-
cerning this process, one can make two general statements: (i) the process
can also be described as a transport, and (2) the forces acting on it can be
represented by the gradients of a limited number of thermodynamic
potentials or homologous entities.
In this general form the above statements apply not only to such metabolic
schemes but are also valid for non-biological coupled transport processes.
For the purpose of illustration transports across cell membranes are, of
course, among the least suited examples, since it is just there that we have
no insight into detailed conditions. On the other hand, one can choose
models from the numerous non-biological instances where the movement
of a substance is not entirely due to the force of diffusion. Especially
among the separation and isolation procedures in the laboratory there are
several processes which involve the movement of a substance from a lower
34 THE CONCEPT AND DEFINITION OF ACTIVE TRANSPORT
to a higher chemical potential. This occurs, for instance, in ultracentrifuga-
tion, electrolysis, distillation of mixture, electro-osmosis, thermo-osmosis
and many others.
.- Before considering in more detail these or similar examples, let us
examine a system which is regarded as normal with respect to the usual
conditions of diffusion. Suppose a diffusion tube, placed horizontally,
contains two substances i and k and is closed at both ends by membranes
which are permeable to i and impermeable to k. The state in all parts of the
tube is defined by conditions of constant temperature, constant pressure
and by a stationary flow of i. Let the chemical potentials in two cross-
section elements I and II separated by a distance of dx be fii + dfa and /^.
In addition to this system let us consider two phases I' and II' of the same
composition as the two cross-section elements. During the reversible
transfer of one mole of i from F to IF, assuming the absence of other com-
pensatory processes, this system does work equivalent to d^ and loses an
equal amount of energy. Let us assume an equal loss of work during the
diffusion of one mole of i from I to II. The natural expression for the force
causing such a movement is then the negative value of this loss of work
divided by the distance dx.K= — =— = — — = K.. . . On these relations are
J ax dx H
based the derivations of the diffusion equations.
Let us now consider the corresponding transport in the diffusion tube
placed vertically. In this case the corresponding loss of work is no longer
dfa , for the transport of the substance involves a simultaneous transport of
mass in a gravitational field. This portion of the energy change is repre-
sented by a term Mtd(j), where Mi is the mass of one mole of i and <f> the
gravitational potential. The total decrease in energy during the corre-
sponding reversible transport is therefore dfa + Mtdip. The force acting on
the transport, again represented by the negative value of the loss of work
divided by the distance, is thus: K= — — = — -i-M^ ==!£„. + K*..
J dx dx dx H 9l
We thus have here an example of an additional force, K^ acting on the
transport of a substance. This force is, of course, capable of causing chemical
uphill transport, as happens, for example, in the ultracentrifuge.
In a similar manner chemical potential gradients can be used for lifting
mass. An example of this is well known from text-books of plant physiology.
It was originally designed by Askenasy and concerns the lifting of water in
high trees (Hulett, 1903). The model consists of a vertical glass tube which
is filled with water and is closed at the upper end with a porous gypsum
plate permeable to water vapour. On evaporation of a certain amount of
water through the gypsum plate an equivalent mass is raised from below
THE CONCEPT AND DEFINITION OF ACTIVE TRANSPORT 35
to the top. We will not concern ourselves with the stability of the system
nor with the correctness of its reproduction of the natural process, but
rather with the energetic mechanism and the force which is responsible
for the uphill transport of mass in this system. Since the water within the
tube is in equilibrium, when no evaporation occurs, there must be — as we
have just seen — a chemical potential gradient which in this case exactly
compensates the effect of the gravitational potential gradient. We thus have
in each cross-section: K = — ~-Mi~=o. The chemical potential of
dx l dx
the water and therefore also its vapour pressure is thus lower at the top
than at the bottom. The driving force for the lifting of the mass is thus the
\
I.R
,.
« i
Fig. 2.
chemical potential gradient and the mechanism is the coupled movement
of chemical matter and mass along the two potential gradients. The state-
ment that here the energy is due to the evaporation is of course not in-
correct, since the transport is due to the evaporation. It could, however,
give the erroneous impression that the heat of evaporation or the work of
evaporation are concerned in it. The value of the former, for 18 g. of water,
is approximately 10,000 cal., that of the latter about 600 cal., whereas the
work for the reversible raising of this amount of water through a tube, for
example, 25 m. in length, would be i cal. This work equals the difference
in the chemical potential at the top and at the bottom.
The result of these processes can be formulated as follows. When the
movement of chemical matter is obligatorily bound with the movement of
mass along a gradient of the gravitational potential, then there acts on this
transport an extra force in addition to the chemical potential gradient.
3-2
36 THE CONCEPT AND DEFINITION OF ACTIVE TRANSPORT
Energetically therefore a complex of chemical matter and mass is con-
ducted along the combined potential gradients.
We have based this result on the change in energy or the loss in work
during the corresponding reversible transfer. Now, the energy equation
in the form given originally by Gibbs has several terms, all of the form JdQ,
where J represents an intensity with potential character and Q a quantity.
This equation is an expression of the experience that differential energy
changes can be specified quantitatively as thermal, spatial, mechanical,
chemical, electrical, etc. In the same way, however, we are able to specify
the forces involved in the transport of matter (or in other transports).
Let us again consider two phases I' and II' with the chemical potential
Pi + fyi and fa and ^e pressures p + dp and ^. The reversible transport of
one mole of i from I' to IF again results in the loss of work rf/^, but only
on the assumption that no volume is transferred at the same time. If, how-
ever, we transfer together with i the volume vi , then the total loss of work
is given by dfa — Vidp. By applying this relation to the diffusion of a sub-
stance one can thus conclude that a pressure gradient can act as an additional
force on transport and that such an action must be effective when under
the given conditions the transport of matter and volume are dependent on
each other. The total force on the transport is then
dA_ dp, dp_
~~ ~
where vi is the volume which is being transported coupled to the transport
of one mole of i.
These considerations show not only the formal similarity but also the
difference between the action of the gradients of pressure and gravitational
potential. When the latter is involved in a transport, there is always an
additional force, K^9 which is of constant value for a given transport. That
is due to the fact that chemical substance cannot be transported without
mass and that the ratio between quantity of matter (number of moles) and
mass (number of grams) is always the same. Pressure gradients, on the
other hand, do not act on the transport of matter if this latter is not coupled
to volume transfer, and in such coupled transports the volume which is
bound to the movement of one mole of i can be variable. The same con-
ditions as in the action of pressure gradients obtain when temperature
gradients are considered (entropy movement with a variable amount of
entropy Si per mole i) as well as the gradients of other chemical potentials
d/irjdx (movement of substance r coupled to i with a variable amount nri
per mole i). On the other hand, electrical charge transport is analogous to
transport of mass if the substance in question does not during the transport
THE CONCEPT AND DEFINITION OF ACTIVE TRANSPORT 37
participate in redox reactions and so alter its charge. It is therefore possible
and also advisable for theoretical reasons to identify the force of diffusion of
ions with the negative value of the electrochemical potential gradient:
K~H == — — = — — — zi e -~- (%! e = ionic charge, and ^ the electrical poten-
tial). In the same way one can, in systems in the field of gravity, define
the force of diffusion as the negative gradient of a gravitational-chemical
potential. Such a treatment is, as mentioned, not possible for the gradients of
temperature, pressure and other chemical potentials. Since in these cases
the corresponding quantities may vary during the coupled transport, the
additional forces are also variable and depend on special conditions during
transport. For the achievement of a chemical uphill transport, for instance,
by a thermal force it is therefore especially favourable to choose a transport
path along which the amount of entropy accompanying the substance is
particularly large. The vapour phase is therefore favourable as transport
medium for chemical uphill transport between two liquid phases.
Considerations of the properties of the transport path leads us directly
to the answer of the question : Of what practical value are such seemingly
abstract relations? The most important possibility is probably that the
search for the additional forces and special mechanisms is given a certain
direction. For this, examination of the conditions for the coupling of trans-
ports within the membrane is of special importance. Generally one would
say conditions are especially favourable for coupling if an uncoupled trans-
port of the isolated quantities concerned is difficult or impossible, that is,
when the membrane is not permeable to the single quantities separately.
For instance, an effective use of temperature gradients for chemical
transports requires a relatively low thermal conductivity of the membrane.
Effective chemical uphill transports are also dependent on low back-
diffusion, i.e. on low permeability for the substance in question.
In connexion with this one can make the striking observation that in
biological systems the substances which undergo definite uphill transports
apparently all have a strongly hydrophilic character, whereas cell membranes
under conditions of normal diffusion seem to be especially permeable to
organophilic substances. Thus it would seem reasonable to search for
organophilic transport complexes formed by the hydrophilic substances in
question. This would direct attention especially on other chemical potential
gradients as additional forces, for a hydrophilic substance can only become
organophilic through chemical transformation. We deal here mainly with
the question of models in order to approach more closely the problem of
the driving forces and the mechanism of coupling. Not many models have
been described in the literature illustrating the just-mentioned coupling,
38 THE CONCEPT AND DEFINITION OF ACTIVE TRANSPORT
but the guaiacol models of Osterhout (1940) deserve interest in this
context.
Let us examine more closely an example of these models. Two identical
aqueous solutions I and II of potassium chloride and the potassium salt
of guaiacol are separated by an organic, guaiacol-containing layer. The
introduction of carbon dioxide into solution I then induces a transport of
K+ from II to I, which leads to the accumulation of potassium in I. In this
uphill transport of potassium ions the guaiacol-potassium compound un-
doubtedly acts as the penetrating transport complex. The force which
causes the transport of potassium ions therefore also contains a term with
the electrochemical potential gradient of the guaiacol ion (G~~) and can, in
complete analogy with the previously presented examples, be represented
by: K=KK+ + KG-K+= — ^~~n^-^+~^~' Here WG-K+ gives the
amount of G~ which is transported while coupled with i gram-equivalent
of K+. In a quantitative treatment of this model the expression would have
to be modified because of a possible participation of undissociated guaiacol
(GH) in the transport processes.
It is worth while to study this model a little further, for it illustrates two
other problems which are of interest in this connexion. One is the depen-
dence of the additional driving force, KG-K+y on a chemical reaction,
the second is the chemical cause for the formation of the transport complex.
On introduction of carbon dioxide into solution I the following reaction
takes place : ^ + ^ + Q- ^ GR + HCQ-
Since the guaiacol ion is consumed in this reaction there arises a concentra-
tion gradient or potential gradient of this ion. This acts as a driving
force on the potassium ion, because the guaiacol ion cannot diffuse by
itself through the organic phase, but only in combination with the potassium
ion. We can express this connexion more clearly and generally by indexing
the equation above:
CO2(I) + H2O(I) + G(II) -> GH(I) + HCO3(I).
The connexion is thus based on the fact that the reacting substances,
carbonic acid and the guaiacol ion, are separated by a layer or membrane and
cannot come in contact with each other without carrying along the potassium
ion.
The chemical problem is why the potassium compound is soluble in the
organic medium, and this question is certainly not answered sufficiently by
simple salt formation. General chemical experience shows that conditions
for the existence of alkali ions as such in organic solvents are extremely
THE CONCEPT AND DEFINITION OF ACTIVE TRANSPORT 39
unfavourable. Only such compounds can be expected to be organophilic
in which the charge is effectively shielded. In the case of ions of heavy
metals very many such organophilic compounds are known, especially the
so-called chelate complexes. Thus the copper complexes of /?-dicarbonyl
compounds like acetyl acetone and the esters of acetoacetic acid and
oxaloacetic acid are sparingly soluble in water and easily soluble in non-
polar solvents. An uphill transport of copper ions by the mediation of these
substances could thus easily be effected. Guaiacol also belongs to this class
of chelate complex formers. The formation of complexes of this kind is due
to formation of rings, usually of 5 or 6 members. Although the tendency
of alkali ions for the formation of chelate complexes is much weaker, it
nevertheless definitely exists as many findings have shown. While in the
case of the ethylene diamine tetraacetic acid (Schwarzenbach & Ackermann,
1947) the sodium complex is more stable than the potassium complex, the
reverse seems to be the case for /?-dicarbonyl compounds (Sidgwick &
Brewer, 1925). Several of the last-named complexes are easily soluble in
toluene. Also the complexes with sodium and potassium of di-yff-naphthol
sulphide, containing an 8-membered ring, have a low solubility in water
and are easily soluble in ether (Evans & Smiles, 1937). Several years ago
we carried out a series of measurements of the electrical potentials in
systems with artificial lipoid membranes to which were added such chelate
complex formers (S. O. Nielsen & Th. Rosenberg, unpublished). We noted
relatively strong potassium-binding effects of the esters of acetoacetic acid
and oxaloacetic acid while sodium was bound less strongly. Dr Wilbrandt
in Berne then studied the effect of diethyl oxaloacetate on the cation distribu-
tion in erythrocytes, but found none. One will thus have to look for more
stable complexes for such effects. Also it is likely that the solvent pro-
perties of cell membranes are not characterized sufficiently by such de-
scriptions as organophilic or lipoid.
It follows from the treatment which has been applied here that what we
observe as active transport is the transport of only one part of an unknown
transport complex, whereas the total transport of the whole complex is
never active. In observing an uphill transport one can thus conclude that
the rest of the transport complex in question is transported ' downhill * and
that this furnishes, owing to coupling, the energy for the uphill transport.
For this rest the term ' energetic carrier* was introduced in a previous
paper (Rosenberg, 1948) because of its just-mentioned function. In order
to avoid possible misunderstandings of this term, I would like to emphasize
that there is no question of suggesting a special carrier mechanism of active
transport, but merely of supplying a description which is essentially
equivalent to that of the transport being under the influence of additional
40 THE CONCEPT AND DEFINITION OF ACTIVE TRANSPORT
forces. The difference is mainly a specification of the forces which can act
on transport of matter at all, and also of the conditions under which such
an action becomes effective. On the other hand, the chemical organophilic
transport complexes may be considered as a special application of these
considerations of coupled transport.
Summarizing one can set up the following points :
(1) The concept of active transport should be clearly defined in the
interest of the elucidation of the basic mechanisms.
(2) Such a definition cannot be based merely on deviations from the
normal diffusion equations. First, one cannot presume that the diffusion
behaviour is 'normal' during the penetration through structures like cell
membranes, and secondly, we have frequently no possibility of differentiating
between changes in the diffusion resistance and changes in the driving
forces.
(3) A definition of active transport as transport under the influence of
other forces in addition to the force diffusion (defined as the negative value
of the chemical or electrochemical potential) is therefore only of limited
practical applicability because of the difficulties in demonstrating such
forces.
(4) A method introduced by Ussing, which is based on the use of
isotopes, makes possible the demonstration of passive penetration if no
specific interaction between the diffusing substance and mobile com-
ponents, including the diffusing substance itself (association), occurs in the
membrane.
(5) A chemical uphill transport, defined as transport from a lower to a
higher chemical (electrochemical) potential, is proof for the action of
additional forces, independent of the structure or of other processes in the
membrane.
(6) Just like the force of diffusion, the additional forces may be represented
by gradients of potentials or analogous thermodynamical entities. Gradients,
as, for instance, the gravitational gradient, that of pressure, of temperature
or of other chemical potentials, act as additional forces when the corre-
sponding quantities mass, volume, entropy, or other chemical components
move coupled with the substance in question.
(7) The general nature of the coupling during transport can be studied in
well-known model systems. For the combined action of several chemical
potential gradients the solubility of the transport complex and of its
separate components is of decisive importance.
THE CONCEPT AND DEFINITION OF ACTIVE TRANSPORT 41
REFERENCES
BRONSTED, J. N. (1946). K0benh. Univ. Festskr. (in Danish).
DANIELLI, J. F. (1943). In DAVSON, H. & DANIELLI, J. F. The Permeability of
Natural Membranes, p. 310. Cambridge University Press.
EINSTEIN, A. (1908). Z. Elektroch. 14, 235.
EVANS, W. J. & SMILES, S. (1937), J. Chem. Soc. p. 727.
FICK, A. (1855). Ann. Phys. Chem. 94, 59.
HODGKIN, A. L. (1951). Biol. Rev. 26, 339.
HULETT, G. A. (1903). Z. phys. Chem. 42, 353.
VAN LAAR, J. J. (1907). Lehrbuch der theor. Elektrochemie, p. 85. Leipzig.
LANGMUIR, I. (1929). J. Amer. Chem. Soc. 51, 2847.
LEHNINGER, A. L. (1951). Phosphorus Metabolism, i, 344. Baltimore.
LUNDSGAARD, E. (1930). Biochem. Z. 217, 162.
ONSAGER, L. (1931). Phys. Rev. 38, 2265.
ONSAGER, L. (1945). Ann. N.Y. Acad. Sci. 46, 241.
OSTERHOUT, W. J. V. (1940). Cold Spr. Harb. Symp. Quant. Biol. 8, 51.
RACKER, E. & KRIMSKY, I. (1952). Nature, Lond., 169, 1043.
ROSENBERG, TH. (1948). Acta chem. scand. 2, 14.
SCHWARZENBACH, G. & AcKERMANN, H. (1947). Helv. chim. acta, 30, 1798.
SIDGWICK, N. V. & BREWER, F. M. (1925). J. Chem. Soc. 127, 2379.
USSING, H. H. (1949). Acta physiol. scand. 19, 43.
USSING, H. H. (1951). Z. Elektroch. 55, 470.
WARBURG, O. & CHRISTIAN, W. (1939). Biochem. J. 303, 40.
SECRETION AND TRANSPORT OF WATER
BY]. R. ROBINSON
Department of Experimental Medicine, University of Cambridge
I. INTRODUCTION
There is now general agreement that ions are transported actively across
cell membranes, but there is no such agreement about active transport of
water. Water is certainly moved across cell membranes, but its movements
might be secondary to those of ions. If ions are transported across mem-
branes which are permeable to water, water must tend to follow because
of differences in osmotic pressure set up by the alterations of ionic concen-
tration. This is transport of water; it is even in a secondary sense active
transport of water in so far as the primary transport of ions is active. But
it is strictly passive so far as water is concerned, because the movement is
one which, given the gradients of ionic concentration, would occur spon-
taneously. Rosenberg (1948) and Ussing (1949) have stressed that trans-
port which can properly be called active differs in direction or in rate from
what could be expected on the basis of gradients of chemical or electro-
chemical potential, and that, since it is not a spontaneous process, it can
only continue so long as there is a supply of energy (e.g. from some meta-
bolic source) to maintain it.
I want to consider the possibility that active transport of water in this
stricter sense, even against osmotic gradients, occurs quite commonly in
living systems — that cells can pump water, as such, with some sort of
water-pump, and not merely secondarily by means of ion pumps. The
evidence is not so complete or compelling as one could wish, but the
possibility may have received less consideration than its deserves, and it
leads to new ways of looking at old problems. That must be the excuse for
following a rather speculative trail which future work may show to be
false. It has been suggested that it is a beneficial intellectual exercise to
practise believing impossible things, even if only before breakfast
(Carroll, 1872).
II. FRESH-WATER PROTOZOA
Protozoa living in fresh water keep their osmotic pressure higher than that
of their habitat ; they accumulate ions and they require finite concentrations
of metabolites. The concentrations within their cells have been measured
in a variety of ways, e.g. by electrical conductivity (Gelfan, 1928), by vapour
SECRETION AND TRANSPORT OF WATER 43
pressure (Picken, 1936) or by finding the lowest external concentration
which will cause the organisms to shrink. This last method has been
exploited particularly by Kitching (1934, 1936, 1938) in forms which
possess contractile vacuoles. These vacuoles seem to excrete water which
diffuses in from the environment because the osmotic pressure is higher
inside the organisms. Metabolic poisons which stop the movements of the
vacuoles cause the organisms to swell, and the external osmotic pressure
which is required to prevent swelling when the vacuoles are not functioning
can be determined.
Such experiments as these have suggested that the excesses of concentra-
tion on the inner sides of the cell membranes which form the body walls
of Protozoa are of the order of 0-01-0-05 osM/1. these correspond to
osmotic pressures of the order of 170-850 mm. of mercury. Since it
cannot be supposed that the cell membranes withstand differences of
hydrostatic pressure of this order, the stability of these organisms cannot
be explained simply by accumulation of ions and retention of metabolites
within a semipermeable membrane. There seem to be three other
explanations :
(1) The membranes are quite impermeable to water. This does not
explain the osmotic behaviour of the organisms.
(2) The membranes are permeable to solutes as well as to water. This
way round the difficulty is no more satisfactory. The contents of the
organism would exert no osmotic pressure, but neither would they remain
inside; they would escape and the organism would shrink. Active retention
of any kind would presumably cause the membrane to behave as if semi-
permeable.
(3) Water is extruded actively as fast as it diffuses in under the influence
of the osmotic gradient. This is the explanation which I prefer. When there
is a contractile vacuole the extrusion of water can actually be seen. Kitching
has shown how vacuolated organisms swell when the vacuoles are inhibited,
and how the vacuoles cease to function if the ingress of water is checked
by raising the external osmotic pressure. These organisms are not in
osmotic equilibrium with their surroundings while they are alive, but they
keep their volume constant and their contents hypertonic as a steady state
by pumping out water. The energy which this requires must be provided
by metabolism. Organisms which do not possess contractile vacuoles
presumably pump water outwards across some part of the body surface by
a mechanism which may be the same as that which transports water across
the membrane of the contractile vacuole in a vacuolated organism.
Kitching (1952) has pointed out that the emptying of the vacuole to the
exterior could be explained mechanically by a very small excess of hydro-
44 SECRETION AND TRANSPORT OF WATER
static pressure, whereas a secretory process must probably be invoked to
explain its filling. This secretion of water must be carried out across one
layer of vacuolar or cell membrane, like the secretion of sodium which
maintains a low concentration of sodium in mammalian cells bathed in
sodium-rich extracellular fluids.
III. AQUATIC METAZOA
More complex fresh-water animals have body fluids which are considerably
hypertonic, and marine teleosts keep their body fluids hypotonic to the
environment. These fluids surround most of the tissue cells and provide
them with a local environment with which they are usually supposed to be
in osmotic equilibrium. Schlieper (1930) pointed out the important
implication that the osmotic pressure of the body fluids needs to be care-
fully guarded, because it regulates the volume of the cells. We shall return
later to the question how far cells and body fluids are truly in osmotic
equilibrium. Meanwhile it is clear that, at the frontiers which separate the
body fluids from the aquatic environment, there are layers of cells whose
opposite poles are in contact with solutions which differ in osmotic pressure.
It is not necessary to discuss the various devices which reduce the area of
living tissue in the frontiers (shells and scales, for example) or which
protect it, like the slime which seems to reduce the permeability to water
of the body surface of eels (cf. Schlieper's reviews, 1930, 1935, and Krogh's
book, 1939). Kven if much of the surface is so shielded, there are osmo-
regulatory organs where the internal and external environments with their
different osmotic pressures are separated by a layer of living cells. Here
work is done to preserve the differences in osmotic pressure which are so
important for the life of the animals.
IV. TERRESTRIAL ANIMALS
Frontiers of this kind which separate fluids differing in osmotic pressure
also occur in terrestrial animals. Human saliva and sweat may have osmotic
pressures half that of the fluid on the other side of the secreting epithelium,
and greater osmotic gradients occur across the secreting epithelium which
lines the convoluted tubules of mammalian kidneys. Human urine may
be more dilute than sweat, and it may also be four times as concentrated as
the plasma (Smith, 1951). The urine of some desert mammals may attain
to higher degrees of concentration than this ; the record seems to be held by
a kangaroo rat studied by Schmidt-Nielsen, Schmidt-Nielsen & Brokaw
(1948) with a urine of which the total concentration was nearly 6 osM/1.
The kidneys and the sweat glands therefore provide good examples of
epithelia separating fluids with different osmotic pressures; moreover, the
SECRETION AND TRANSPORT OF WATER 45
epithelia lining the renal tubules and the sweat glands are only a single cell in
thickness, so that these cells must sustain huge osmotic gradients.
There may be a closer analogy than is at first apparent between Protozoa
and the cells of osmoregulatory organs. The cytoplasm in the cells com-
posing a living membrane which lies between two fluids with different
osmotic pressures cannot be truly in osmotic equilibrium with both of
them. Consequently an osmotic gradient must exist across a single layer
of cell membrane at one pole of the cells at least. The production of
anisotonic secretions has usually been regarded as a function of complete
layers of cells, the secretory work being done in the cytoplasm. When the
human kidney is producing maximal concentrated urine, the concentration
of the fluid at the luminal pole of a cell lining the distal tubule may be
I-20SM/1., compared with 0-3 osM/1. at the opposite pole, for there is
evidence that the reabsorbed fluid under these conditions is at least as
dilute as the plasma (Chambers, Melville, Hare & Hare, 1945). What then
is the osmotic pressure of the cytoplasm ? And where is the osmotic work
carried out? Water must either be actively transported into the cells at the
luminal poles, or else out of the cells at the opposite, basal, poles. In the
former case, the cytoplasm must have a lower osmotic pressure than the
body fluids in general for water to flow out spontaneously into the peri-
tubular interstitial fluid in contact with the basal ends of the cells. This
situation is not analogous to that in the Protozoa; it corresponds to a marine
protozoonwith hypotonic contents. In the latter case the cytoplasm must be
kept hypertonic to the body fluids by active extrusion of water at the basal
ends of the cells into the peritubular interstitial spaces. Reabsorption at
the luminal pole could then be passive, for water would flow into the cells
down an osmotic gradient. The urine could be concentrated up to a limit
fixed by the maximal attainable intracellular osmotic pressure; but the
limiting concentration would only be reached at low rates of urine flow
because the rate of transport across the basal ends of the cells would set
an upper limit to the rate at which water could be reabsorbed. Recent work
by Ladd (1952) discussed by Homer Smith (1952) suggests that the final
process by which the human kidneys elaborate a hypertonic urine may be
a reabsorption of water unaccompanied by solute against an osmotic
gradient (between urine and blood) at a rate of not more than 2-3 ml./min.
Hence a process analogous to that which must operate to keep the contents
of unicellular animals more concentrated than their surroundings could
account for the known behaviour of mammalian kidneys, only it would have
to operate at one pole of the cell, and not symmetrically all round its
surface. There is an obvious histological difference between the basal and
the luminal poles of the cells of the renal tubular epithelium. The basal
46 SECRETION AND TRANSPORT OF WATER
ends show striations which are due to the parallel arrangement of mito-
chondria packed closely together at right angles to the basal surface of the
cells. This would be consistent with a secretory process located at the end
of the cell remote from the lumen; it is perhaps analogous with the
grouping of mitochondria round contractile vacuoles, where a similar
extrusion of water may be occurring.
To sum up, there is no doubt that differences in osmotic pressure exist
across cell membranes in fresh-water Protozoa and in certain glands and
excretory organs of higher animals. In the former active transport of water
across a single thickness of cell membrane has to be postulated, and the
same fundamental process could account for the other examples of secretion
of water.
V. MECHANISM
Not much has been published to explain how such a process of active
transport of water might work. Mechanisms like the diffusion pump which
Franck & Mayer (1947) proposed require the cytoplasm to be divided into
compartments by semipermeable partitions. They might move water from
one end of a cell to the other end, and so pump it across a layer of epi-
thelium, but it is hard to visualize them pumping water across one thickness
of membrane either into a less concentrated environment or into a contrac-
tile vacuole. Thermo-osmosis provides a possible mechanism, which
Dr Spanner is going to discuss. The alternate expansion and contraction
of polypeptide chains combined with a cyclic alteration in their hydration
might provide a pump within a membrane (Goldacre, 1952), and there are
hints that the membrane may be a more complex organ than has been
supposed. Sjostrand (1953) has resolved some double membranes with
the electron microscope in secreting cells and around mitochondria, which
is interesting because Hartley & Davies (1952) suggested that mitochondrial
membranes perform secretory work. There seems to be an association
between double membranes and secretion, and they might contain the
secret of the water pumps. But it is not yet known how water is pumped;
an explanation may, however, be found if it is admitted that there is
something to explain.
VI. OSMOTIC EQUILIBRIUM OF BODY CELLS IN
HIGHER ANIMALS
Reasons have been advanced for believing that some of the cells in
secreting organs cannot be in equilibrium with the body fluids which
bathe them, and we must now return to the question whether the other
cells of higher animals are truly in osmotic equilibrium with these fluids.
It seems somehow natural and reasonable to regard cell membranes as
SECRETION AND TRANSPORT OF WATER 47
semipermeable, and aqueous solutions separated by such membranes as in
osmotic equilibrium. This hypothesis has been of great practical value in
clinical medicine, for it has enabled alterations in the volumes of the
various body fluids to be correlated with alterations in their composition,
and it has satisfactorily accounted for most of the observed shifts of body
water between the cells and the extracellular fluids. It explains, for
example, the paradox that a deficiency of sodium produces a more severe
form of dehydration than a pure deficiency of water (McCance, 1936;
Marriott, 1947). Some account has been given elsewhere of how this
simple and attractive view grew up (Robinson, 1953).
It was known that there was a greater total concentration of fixed base
in the water in the cells than in the extracellular fluids, but this had been
ascribed to a Donnan equilibrium, for the cells were presumed to contain
considerable amounts of protein and other polyvalent and non-diffusible
anions (cf. Newburgh's (1950) lucid account from the traditional stand-
point). J. P. Peters (1935, 1944), who did perhaps more than any other
investigator to preach the gospel of osmotic equality to the medical world,
explained the apparent excess of osmotic material inside cells, and the fact
that intracellular base sometimes varied without appropriate shifts of water,
by postulating that a variable portion of the fixed base in cells is bound in
some osmotically inactive form (Peters, 1937-8; Danowski, 1951). The
manner of binding is unknown, and the amount bound can only be
arrived at by assuming that the cells are in osmotic equilibrium, so that
this assumption can only be preserved (as a dogma) by placing it beyond the
possibility of experimental proof. The great attractiveness of this hypo-
thesis can be seen from the fact that, if it is true, the osmotic pressure of the
inaccessible intracellular fluids can be determined in living animals and
men by analysing samples of extracellular fluids, of which blood plasma is
typical and easy to obtain. But we have already seen that some of the cells
in the body cannot have the same osmotic pressure as the extracellular
fluids, and there is evidence which suggests similar inequalities in others.
Attempts to determine the osmotic pressure of cell fluids cryoscopically
have been made from time to time since Sabbatani (1901) reported a
greater depression of the freezing-point of water in the parenchymatous
organs of dogs than in their blood. This sort of discrepancy has always
been found, and seems to have been accepted at its face value by workers
who have used the method, although the results might be affected by
autolytic changes occurring in the organs post mortem. However, Sabbatani
had found that the usual difference in osmotic pressure between the blood
and the liver was abolished by poisoning with phosphorus, and Gomori &
Molnar (1932), who used the same method, found that differences in
48 SECRETION AND TRANSPORT OF WATER
osmotic pressure between a number of organs and the blood of rabbits
disappeared in the terminal stages of water intoxication. It seems that
phosphorus poisoning and water intoxication had abolished either a normal
hypertonicity of the cells, or the post-mortem autolytic changes.
VII. OSMOTIC PROPERTIES OF ISOLATED TISSUES
It has been known for some time that isolated tissues swell in solutions
which have the same osmotic pressure as the body fluids which bathed the
tissues during life. (See Robinson (1953) for some references to the earlier
literature.) Opie (1949) found with pieces of the liver and kidney of rats
that solutions of sodium chloride of about twice the osmotic pressure of
the body fluids were needed to prevent this swelling, and suggested that the
cytoplasm normally had a greater osmotic pressure than the extracellular
fluids in the body. These results might also have been due to autolysis, if
the membranes had remained impermeable to its products; but Opie (1950)
found that the cells of the liver and kidneys swelled when the animals were
poisoned with chloroform or potassium chromate, and that the behaviour
of the isolated tissues then suggested that the poisoned cells, unlike normal
ones, had been in osmotic equilibrium with their surroundings in the body.
This was also true of the livers of rats poisoned with carbon tetrachloride,
but the cells of those rats which did not die of the poisoning became hyper-
tonic again when they recovered. These effects of poisons recall Sabbatani's
(1901) old observation that phosphorus abolished the difference in freezing-
point between the liver and the blood in the dog, and suggest that Opie's
results also need not be dismissed as artefacts arising from autolysis.
VIII. OSMOTIC PROPERTIES OF SURVIVING
TISSUE SLICES
The use of sodium chloride instead of a balanced saline medium might
have contributed to the swelling of isolated tissues observed by Opie, but
I have observed the same behaviour in thin slices from the livers (Robinson,
19520) and kidneys (Robinson, 19500) of rats in a medium which repro-
duced the ionic pattern of the extracellular fluids well enough to support
respiration at a constant rate for several hours. Swelling was measured by
the percentage of water in the tissue, for experiments with inulin had
suggested that alterations in the amount of water in kidney slices reflected
changes in the volume of the cells. The important point which emerged
was that Opie's results were only confirmed when respiration was inhibited,
for instance, by cyanide or by chilling to 0-4° C. ; or when the metabolic
reaction patterns were dislocated with 2, 4-dinitrophenol (Robinson,
1950^). Chilled slices swelled unless the medium had about twice the
SECRETION AND TRANSPORT OF WATER 49
normal concentration of extracellular fluids, and the cells in slices poisoned
with cyanide at 38-5° C. roughly doubled their volume in supposedly
'isotonic' solutions. They did not swell in these same solutions when
respiring normally, and they swelled surprisingly little when respiring in
hypotonic solutions.
An essentially similar relation of water balance to respiration had been
reported by Aebi (1950 a) for liver slices from the guinea-pig, and by
Stern, Eggleston, Hems & Krebs (1949) for the liver, spleen, kidney, lung
and brain of the same animal. The latter authors found that slices from all
these tissues swelled under anaerobic conditions and concluded that some
mechanism dependent upon the supply of energy was the dominant factor
in regulating exchanges of fluid between the cells and the medium. Anoxia
and cyanide might have damaged the cells irreparably, but the swelling
which occurred when respiration was inhibited with cyanide turned out to
be fully reversible when cyanide was distilled out of the medium and the
oxygen uptake recovered (Robinson, 1950 a). This suggested that Opie was
right to believe that these cells were not in osmotic equilibrium while they
were alive; but that their volume was regulated, and the osmotic pressure
of their contents was kept above that of their surroundings, as a steady state,
by some process which used energy derived from respiration to pump
water outwards across the cell membrane. When the pumps stopped for
lack of energy the cells had to swell. The action of 2,4-dinitrophenol
suggested that energy was made available through the mediation of
adenosinetriphosphate, and approximate calculations made for the kidney
slices suggested that the amount of energy required to maintain the steady
state was roughly proportional to the measured oxygen consumption, and
that its absolute value was reasonable. Swelling in hypotonic surroundings
would be anticipated if the pumps could not speed up to cope with the
more rapid diffusion of water into the cells down a steeper osmotic gradient.
The cells in slices respiring in hypotonic solutions did swell, but they
swelled less than cells in osmotic equilibrium would have been expected
to do, and also less than cells whose respiration was suppressed. Hence the
volume of respiring cells reacted to changes in external concentration in
the direction expected for cells in osmotic equilibrium, but to a smaller
extent and for a different reason. I had made a few experiments upon
liver slices before the work of Aebi and that of Krebs's team was published,
with the idea that these would serve as controls and show the behaviour
of kidney slices to be peculiar; but when the liver slices behaved in much
the same way it began to look as though this sort of behaviour might be
more general than I had supposed. This work has been published, and it
remains to deal with a few points which merit further discussion.
5O SECRETION AND TRANSPORT OF WATER
IX. THE IMPORTANCE OF EXTRACELLULAR PROTEIN
Experiments upon tissue slices may fail in two main ways to reproduce
conditions in vivo. First, Black (1953) has revived the suggestion of
Trowell (1946) that the interstitial phase is a gel rather than a free fluid, but
there is little information available upon this point. The fact that no extra-
cellular fluid escapes from the cut surface of a dead tissue may mean only
that water has been absorbed by cells which have been deprived of oxygen
and have ceased to pump it out. Secondly, the media used in most mano-
metric work contain no protein. It is not certain how much protein the
interstitial fluids contain in the body: Drinker & Yoffey (1941) suggested
that lymph might contain half as much as the plasma; but if the function
of the lymphatics is to return to the blood stream protein which has
escaped from the capillaries, lymph may contain more protein than inter-
stitial fluid in general; how much more is uncertain. The cells of the renal
tubules are probably surrounded by a fluid which contains very little
protein, for the glomerular filtrate contains hardly any, the volume
produced each hour is about 30 times that of the whole kidney, and all but
a small percentage is reabsorbed. Hence a protein-free medium is probably
physiological for experiments upon kidney slices, although the same may
not be true of the liver, since this organ manufactures most of the protein
of the plasma, and the hepatic lymph is especially rich in protein. In any
case the colloid osmotic pressure of the extracellular fluids must be trivial
compared with their crystalloid osmotic pressure.
Parry (1936) found that the proteins of dogs' serum did not prevent the
swelling of excised portions of rats' muscle in 'isotonic' saline media.
Aebi & Meyer (1951) found that liver slices from guinea-pigs only behaved as
osmometers with respect to substances of high molecular weight when the
cell membranes had lost their semipermeable behaviour with respect to
ions (see later, p. 57). An earlier paper of Aebi (19506) had shown that
serum increased the oxygen consumption of liver slices by up to 200% by
supplying metabolites, so that its action could not be ascribed simply to
its colloidal osmotic effects. Suggestions that proteins prevent cell
membranes from becoming leaky do not help; a cell might shrink if a leaky
membrane allowed the contents to escape, but it should hardly swell.
Finally, it is improbable that the use of protein-free media vitiated the
experiments upon isolated tissue slices, because slices did not swell in
protein-free media if their respiration was satisfactory, and the swelling
which followed inhibition of respiration with cyanide was reversible in the
absence of protein.
SECRETION AND TRANSPORT OF WATER
51
X. THE MAGNITUDE OF THE INTRACELLULAR
OSMOTIC PRESSURE
The weakest point in this dynamic theory of the water exchanges of
living mammalian cells is that it has still not been demonstrated directly
that the intracellular fluids are hypertonic. Potts (1952) made some direct
03
02
z
9
E
1
i
E
01
I
I
O
©
0®
0
0
X X
VAV
o •-•••-
0
Respiration (/J./mg./hr.)
Fig. i. Estimated intracellular hypertonicity and oxygen consumption of rat kidney slices
respiring in media of different concentrations. Concentration of medium: •, 0-30 osM/1.
(* isotonic'), including experiments in presence of cyanide. 0, 0*19 osM/1. A, 0-06 osM/1.
O, 0-I20SM/1. x , 0-03 osM/1.
measurements by a micro-cryoscopic method and found that muscle cells
of Mytilis edulis and eggs of Psammechinus miliaris were in osmotic equili-
brium with their surroundings, but it remains to apply this method to cells
which might not be expected to be in equilibrium. I estimated (Robinson,
19500) that the cells of kidney slices might have an internal concentration
4-2
52 SECRETION AND TRANSPORT OF WATER
of 0-52 osM/1. when respiring in a medium of 0-30 osM/1. The corresponding
figures for slices respiring in media of other concentrations may be obtained,
assuming the membranes to behave as semipermeable, by taking the
internal concentration as m = o-52/sc, where sc is the relative volume of cell
water determined from the observed percentage of total water in the
slices (sc= i-oo in vivo, when the external concentration m0 = O"$o). Fig. i
shows (mt — m0) plotted against the oxygen consumption of slices in
0-30 OSM media with and without the addition of cyanide, and of slices in
four more dilute media without cyanide, and reveals a clear relation between
'2-0
-1-0
T
0-2 0-4
External concentration (OSM)
0-6
Fig. 2. Kflfect of concentration of external medium upon volume of cells in respiring tissue
slices (expressed as ' Relative volume of cell water ', sc). - , calculated from equation (2)
(•), observed (Robinson, 19500).
the estimated intracellular hypertonicity and the oxygen consumption of
the tissue. Moreover, approximately the same concentration difference
appeared to be maintained across the cell membrane in all these external
solutions, whose concentrations ranged from 0-03 to 0-30 osM/1.
The amount of swelling to be expected if a constant difference in
concentration was maintained across the cell membrane can be predicted
and compared with that which was observed. The condition for a constant
difference in concentration is
(i)
But since in vivo sc= i-oo when w0 =
= o-22y and
0-52
(2)
SECRETION AND TRANSPORT OF WATER 53
The smooth curve in Fig. 2 shows sc calculated from (2) ; the points are the
averages from experiments in different media. The cells did not shrink so
much in hypertonic media as was predicted, but their behaviour in dilute
solutions agreed well with equation (2). Moreover, (2) predicts a finite
volume (corresponding to ^ = 2-36) for slices in distilled water, whereas if
the cells were in osmotic equilibrium and their membranes had remained
semipermeable their volume should have become infinite. This point may
not be merely of academic interest, for human urine can approach the
concentration of distilled water quite closely. Hence the osmotic behaviour
of the slices is consistent with a constant internal hypertonicity maintained
by active transport of water.
XL CONCENTRATION OF FIXED BASE IN CELLS
In an effort to gain more direct evidence of internal hypertonicity in
respiring cells the total amount of fixed base (Na+K-fCa-f Mg) was
determined in kidney slices under various conditions (Robinson, 19526).
The concentration of fixed base was higher in the cell water than in the
medium, and the difference was about the same in a 0-17 OSM as in a
0-30 OSM medium; but it could not account for a difference in total con-
centration greater than 0-07 OSM, compared with the value of 0-22 suggested
for (mt — m0) above. The larger estimate was based on osmotic behaviour
and the smaller on determination of cations only, but it is generally sup-
posed that electrolytes are responsible for most of the osmotic pressure
exerted across cell membranes. An interesting point about this difference
between the concentration of base inside and outside the cells was that it
was abolished by chilling or by cyanide. The similarity between this and
the effect of chilling and cyanide upon respiration, rather than the magni-
tude of the difference, suggested that it was due to an active process.
Some alternative interpretations must now be considered very briefly.
The osmolar value of divalent cations is only half their chemical equivalence,
but there is no reason to suppose that the cells contained enough divalent
cations to account for the excess of total base without osmotic imbalance
(McCance & Widdowson, 1946). Intracellular cations held by the electro-
static attraction of indiffusible anions should contribute to intracellular
osmotic pressure. Conway (1945) suggested that most of these anions were
substances of low molecular weight like creatine-phosphate, carnosine and
adenosine triphosphate, which presumably also contribute to intracellular
osmotic pressure. Hodgkin (1951) concluded that little if any potassium
in the cells of excitable tissue is bound, and Klotz (1952) pointed out that
no proteins so far examined bind sodium or potassium in complexes which
should nullify their osmotic activity. Protein anions of large valency
54
SECRETION AND TRANSPORT OF WATER
making small contributions to osmotic pressure might account for what has
been called 'osmotically inactive base', though it is the accompanying
anion that is relatively inactive. The amount of base that could be held in
this way should depend upon the relation of the isoelectric point of the cell
proteins to the intracellular pH, neither of which is known. It has been
suggested that the contents of some cells are more acid than their sur-
roundings (cf. Davson, 1951), and I did some very crude experiments
which have confirmed this for the cells of kidney slices. These slices will
accumulate phenolsulphone-phthalein (phenol red), a substance which is
actively excreted by the kidney, for more of it may appear in the urine than
in the glomerular filtrate formed at the same time (Smith, 1951). Slices
from rat's kidneys incubated at 38-5° C. in the oxygenated medium used
for manometric experiments also took up :
(1) tetrabromo-w-cresol-sulphonephthalein ( = bromocresol green),
(2) dichlorophenol-sulphonephthalein ( = chlorophenol red), and
(3) dibromothymol-sulphonephthalein ( = bromothymol blue).
Table i
Indicator
Bromocresol
green
Chlorophenol
red
Bromothymol
blue
pH range
Colour of slice
Slice + acid
Slice + alkali
4-0-5-2
Blue-green
Yellow
No change
5-0-6-7
Pink
Yellow
Red
6-0-7-6
Yellow
No change
Blue-green
Table i shows the pli ranges of these indicators, the colours of slices
which had taken them up from media buffered at pH 7-4, and the final
colours produced after placing drops of N/io-HCl and N/io-NaOH upon
stained slices. The colour changes produced by acid or alkali were delayed:
nothing happened for about a minute, and then the colour suddenly
changed, so that the staining of the slices was presumably due to dye within
the cells, and the delay to the time taken for the reagent to penetrate the cell
membranes. The cytoplasm in slices respiring in a medium of pH 7-4 thus
appeared alkaline to bromocresol green, acid to bromothymol blue, and
within the range of chlorophenol red, suggesting that the intracellular pH
was around 5*5, but this must be regarded as an extremely rough estimate.
A pH of 5*5 is not far removed from the isoelectric points of many proteins
(Schmidt, 1945; West & Todd, 1951), although the isoelectric points of
intracellular proteins other than those of muscle do not seem to have been
determined. The amount of base held in cells by the electrostatic attraction
of protein anions is therefore likely to be small, and it would not be easy
to explain the swelling which occurs when respiration is inhibited by acid
SECRETION AND TRANSPORT OF WATER 55
metabolites of low molecular weight lowering the intracellular pH and
providing osmotically active partners for cations which were previously
balanced by polyvalent protein anions of low osmotic activity. A large
amount of swelling has to be explained, for the cells may double their
volume when poisoned by cyanide. Moreover, only metabolic products
to which the membrane is impermeable could account for the swelling, for
although the increase in cell volume was complete in a few minutes, it was
maintained for several hours. These considerations may make it easier to
believe that the excess of base in respiring cells indicates an excess osmotic
pressure of their cytoplasm.
XII. 'WATER PUMPS' VERSUS 'ION PUMPS'
If intracellular hypertonicity can be accepted, there seems little alternative
to postulating active transport. Moreover, the cells swell when the trans-
port system is deprived of its sources of energy. A transport system which
opposes swelling must be directed outwards; and if it leads to a higher
concentration inside the cell, it must be transporting water. The problem of
how far the phenomena could be accounted for by, say, a sodium pump,
and the swelling by water entering cells along with sodium when this pump
is stopped, can be approached experimentally by working in sodium-free
solutions. Kidney slices from rats respired satisfactorily, at least for an
hour, when the sodium chloride in the usual media was replaced by choline
chloride. Table 2 shows the percentages of water in slices in sodium-free
media of different concentrations, compared with the corresponding values
in ordinary media (Robinson, 19500), both when the slices were incubated
at 38-5° C. and when their metabolism was suppressed by chilling. When
respiration was suppressed choline solutions prevented swelling slightly
better than sodium-containing solutions, so that a small part of the swelling
might have been due to the entry of water accompanying sodium. Columns
2 and 4 demonstrate the same inverse relation between swelling and
respiration in sodium-free as in ordinary media. The differences in water
content between column 4 and column 2 cannot be ascribed to the entry
of water following sodium into the slices, because there was no sodium
outside to enter. Slices also took up water in sodium-free solutions when
their respiration at 38-5° C. was inhibited by cyanide. Hence respiration
appeared to be opposing the entry of water, rather than that primarily of
sodium.
In some experiments still in progress small concentrations of a mercurial
diuretic (thiomerin) increased the amount of water in respiring kidney
slices out of proportion to the effect upon oxygen consumption. Mercurial
diuretics are generally supposed to act by stopping active reabsorptjo&of
56 SECRETION AND TRANSPORT OF WATER
sodium by the renal tubules, so that it appeared that this swelling might be
the consequence of stopping a sodium pump. But further experiments
showed that the effect of thiomerin was about as great in sodium-free
solutions as in the ordinary media, which again points to a system trans-
porting water independently of ions.
Table 2. Percentages of water in adult rat kidney slices in ordinary
and sodium-free solutions of different total concentration
Concentra-
tion of
medium
Chilled to 0-4° C.
Respiring at 38*5° C.
OSM/1.
(i)
0'12
O'lQ
0-30
o*45
0-58
Na-free
(2)
Na present
(3)
Na-free
(4)
Na present
(5)
85-8
83-I
80-4
76-8
74'7
85-0 + 0-8
83-7±o-8
8i-6±i-3
78-4±i-5
75'8±i'8
81-9
78-5
77-5
76-6
76-5
81-9 + i-o
78-5 ±0-7
77-5 ±0-8
75'4±°'5
76-5 ±0-8
Ion pumps are presumably operating as well, and it is hoped that more
work with mercurial diuretics may help to sort out their relations to the
water pumps which also seem to exist. Aebi (1951, 19520, b) made
a detailed study of the influence of the conditions of incubation upon the
properties of liver slices from guinea-pigs, and although he did not make
use of metabolic inhibitors, his results resembled those I had observed
with slices of the liver and kidney of rats. The main difference was that an
inverse relation between the respiration and the water content of the slices
was not always present, notably when calcium was omitted from the media,
i.e. under relatively unphysiological conditions. Aebi's main conclusion
was that conditions of incubation which were unfavourable for the
retention of potassium within the cells were associated with a reduced
consumption of oxygen, with swelling by uptake of water, and with
more rapid disintegration of the cells shown by a faster loss of nitrogen
from the slices. (The effects of lack of calcium upon the loss of nitrogen
were remarkably similar to those published by Robinson (1949) on rat-
kidney slices.) Aebi (19520) suggested that the processes which controlled
the water and the ionic contents of the slices were somehow linked, and
perhaps possessed a common mechanism. Sodium entered the slices when
potassium escaped, and this would be expected to follow the stopping of
a sodium pump which normally kept the concentration of sodium in the
cells low, and so kept that of potassium high, as has been proposed in the
case of excitable tissues (Hodgkin, 1951). There is no reason on this
hypothesis why the cells should swell if the sodium pump stops, for entry
SECRETION AND TRANSPORT OF WATER 57
of sodium is balanced by loss of potassium (except for a slight swelling
which might arise from the somewhat greater osmotic effectiveness of
sodium ions compared with their equivalent of potassium; cf. Hill, 1950).
Ion pumps might, however, influence swelling indirectly in a rather
important way, if the functional semipermeability of the cell membranes
were to depend upon the dynamic separation of sodium and potassium
which they maintain. If sodium and potassium could diffuse freely across
the membranes in both directions, external sodium chloride could exert
no osmotic pressure, and its concentration would not be expected to
control the volume of the cells osmotically. This may explain Aebi &
Meyer's (1951) observation that the volume of the cells was controlled by
colloidal osmotic pressure under conditions which prevented active osmo-
regulation and allowed intracellular potassium to be exchanged for
sodium.
A sodium pump could perform two functions. It could maintain the
characteristic difference in ionic pattern between intracellular and extra-
cellular fluids, and thus also account for the apparent semipermeability of
membranes which tracer studies have shown to be really permeable. It is
doubtful how far a sodium pump could at the same time control the
amount of water in the cell. Experiments in sodium-free media suggested
that it could not account for the osmotic behaviour of the cells in surviving
slices, and that there is in addition a water pump to provide independent
control of the volume of the cells and of the tonicity of their contents. If
these two separate pumps exist they are more likely to have a common
source of energy than a common mechanism.
XIII. 'CLOUDY SWELLING'
If exchanges of water which influence the volume of the cells depend upon
the metabolism of the cells as well as upon external osmotic pressure, it
might be anticipated that shifts of water could occur in the body which
were not the consequences of changes in the composition or the concentra-
tion of the extracellular fluids. A few examples of shifts of this kind have
been produced experimentally, notably by Hamburger & Mathe (1951,
1952). They found that in poisoning with carbon monoxide or with
sublethal doses of cyanide, and in experimental acidosis and histamine
shock, water moved into the cells, with a reduction in the volume of
distribution of thiocyanate, and an increase in plasma protein concentration
and haematocrit, denoting a diminution in the volume of circulating blood.
Shifts of this kind have really been known for a long time, although they
have never been regarded in this light, for the phenomena of cloudy
swelling are recognized in the parenchymatous organs by pathologists at a
58 SECRETION AND TRANSPORT OF WATER
high proportion of post-mortem examinations. Cloudy swelling occurs
especially in association with toxic conditions, fevers and anoxia. It is one
of the mildest forms of cellular damage that can be recognized at autopsy,
and in its earlier and less severe stages it is reversible (Bell, 1913; Moon,
1951). This suggests that there is a primary functional disturbance which
precedes later modifications of structure. From the chemical standpoint the
characteristic change is an increase in the amount of water in the affected
tissues; the cells are swollen; yet they remain surrounded by body fluids
whose osmotic pressure is not diminished. They resemble the cells of
slices whose respiration is inhibited, and swelling of this kind is just what
would be expected to occur if the water pumps ceased to function properly.
Moreover, the results of Opie & Sabbatani (cf. p. 48) with poisons which
produce cloudy swelling, suggested that these poisons abolished a normal
hypertonicity of the cell contents. Cloudy swelling might be due to an
increase in the amount of osmotically active material in cells in osmotic
equilibrium; but it might also be due to the failure to maintain a normal
steady state of disequilibrium. This might also explain the watery vacuola-
tion of liver cells which Trowell (1946) found to be associated with anoxia,
and to be reversible when the cause was removed.
A further characteristic of cloudy swelling is that the mitochondria
within the cells are swollen (Turk, 1913; Anitschkow, 1914, 1923; Duthie,
1935). Mitochondria swell when the osmotic pressure of the cytoplasm is
reduced by placing cells in hypotonic solutions. Opie (1948) suggested that
cell inclusions of this kind might be regarded as intracellular osmometers.
Zollinger (1948) showed that mitochondria swelled if they were released
into isotonic saline solutions by rupturing the cell membrane, which
suggests that the osmotic pressure of the cytoplasm is normally greater
than that of these 'isotonic' solutions. This is also consistent with the
experience of Hogeboom, Schneider & Pallade (1947, 1948) that hypertonic
solutions had to be used in order to prepare isolated mitochondria from the
liver and kidney of rats without the loss of their staining reactions. Hence
the behaviour of mitochondria may indicate that the fluid within certain
cells has a greater osmotic pressure than the surrounding interstitial fluid,
although some recent work has made another interpretation possible.
Bartley & Davies (1952) found considerable concentration ratios for some
common ions between isolated mitochondria and the media in which they
were studied; and Raaflaub (1952) discovered an inverse relation between
the oxygen consumption and the swelling of isolated mitochondria in
suspensions, which is reminiscent of the behaviour of cells in tissue slices.
It is therefore possible that the behaviour of the cells might be the reflexion
of the behaviour of the mitochondria.
SECRETION AND TRANSPORT OF WATER 59
There seems to be a choice here between hypertonic mitochondria
bathed in a cytoplasm which is osmotically in equilibrium with the extra-
cellular fluid, and mitochondria in osmotic equilibrium with a cytoplasm
which is kept more concentrated than the extracellular fluid. Poisoning in
the first case would stop secretion of water outwards across the mito-
chondrial membranes, and so the mitochondria would swell, and the cells
would take in water to allow them to do so, but without increasing the
volume of cytoplasm outside the mitochondria. In the second case
poisoning would cause the cells to swell and come into osmotic equilibrium
with their surroundings ; at the same time the dilution of the cytoplasm
would allow the mitochondria to swell as osmometers. The magnitude of
the swelling which occurs when respiration of tissue slices is inhibited
favours the second alternative; but in either case, whether an osmotic
gradient is maintained across the mitochondrial membrane or across the
cell membrane, there has to be a water pump somewhere.
XIV. CONCLUSION
The existence of osmotic gradients in Protozoa and larger aquatic animals
as well as in secretory organs generally, indicates that certain cells can
transport water actively across their membranes. The cryoscopic and
osmotic properties of excised tissues, relations between the water balance of
surviving tissue slices and their metabolism, the dependence of the
concentration of intracellular base upon metabolism and the behaviour of
mitochondria inside cells and outside them suggest that this process is
not restricted to Protozoa and the cells of secretory organs, but that water
is rather generally pumped across cell membranes. Cells in dynamic
equilibrium with their extracellular fluids are less at the mercy of their
surroundings than if they were in osmotic equilibrium in the classical,
static, sense; their own metabolism allows them to take an active part in
regulating the movement of water across their membranes. There is a
certain fascination in the idea that the exchange of water between each cell
and its surroundings may be an active one. Perhaps the only immediate
value of this new way of looking at water metabolism is that it is provo-
cative, and its irritant action may stimulate future research. Whether this
is directed to showing how the water pumps work, or to filling the gaps in
the orthodox formulation and destroying the source of irritation matters
relatively little.
REFERENCES
AEBI, H. (1950 a). Helv. physiol. acta, 8, 525.
AEBI, H. (19506). Helv. physiol. acta, 8, C. 12.
AEBI, H. (1951). Experientia, 7, 346.
AEBI, H. (1952*2). Helv. physiol. acta, 10, 184.
60 SECRETION AND TRANSPORT OF WATER
AEBI, H. (19526). Biochim. biophys. Acta, 9, 443.
AEBI, H. & MEYER, A. (1951). Helv. physiol. acta, 9, 51.
ANITSCHKOW, N. (1914). Verh. dtsch. Path. Ges. 17, 103.
ANITSCHKOW, N. (1923). Arch. mikr. Anat. 97, i.
HARTLEY, W. & DAVIES, R. E. (1952). Biochem. J. 52, xx.
BELL, E. T. (1913). J. Amer. Med. Ass. 61, 455.
BLACK, D. A. K. (1953). Lancet, i, 305.
CARROLL, LEWIS (1872). Through the Looking-glass, and what Alice found there.
London: Macmillan.
CHAMBERS, G. H., MELVILLE, E. V., HARE, R. S. & HARE, K. (1945). Amer. J.
Physiol. 144, 311.
CONWAY, E. J. (1945). Biol. Rev. 20, 56.
DANOWSKI, T. S. (1951). Amer. J. Med. 10, 468.
DAVSON, H. (1951). A Textbook of General Physiology. London: Churchill.
DRINKER, C. K. & YOFFEY, J. M. (1941). Lymphatics, Lymph and Lymphoid Tissue.
Cambridge, Mass. : Harvard University Press.
DUTHIE, E. S. (1935). J. Path. Bact. 41, 311.
FRANCK, J. & MAYER, J. E. (1947). Arch. Biochem. 14, 297.
GELFAN, S. (1928). Protoplasma, 4, 192.
GOLDACRE, R. J. (1952). Int. Rev. Cytol. i, 135.
G6M5RI, P. & MOLNAR, S. (1932). Arch. exp. Path. Pharmak. 167, 459.
HAMBURGER, J. & MATH£, G. (1951). Pr. Med. 59, 265.
HAMBURGER, J. & MATHE, G. (1952). Metabolisme deVeau. Paris : Editions Medicales
Flammarion.
HILL, D. K. (1950). J. Physiol. in, 304.
HODGKIN, A. L. (1951). Biol. Rev. 26, 339.
HOGEBOOM, G. H., SCHNEIDER, W. C. & PALLADE, G. E. (1947). Proc. Soc. Exp.
Biol., N.Y., 65, 320.
HOGEBOOM, G. H., SCHNEIDER, W. C. & PALLADE, G. E. (1948). J. Biol. Chem.
172, 619.
KITCHING, J. A. (1934). J. Exp. Biol. u, 364.
KITCHING, J. A. (1936). J. Exp. Biol. 13, ii.
KITCHING, J. A. (1938). J. Exp. Biol. 15, 143.
KITCHING, J. A. (1952). Symp. Soc. Exp. Biol. no. 6, p. 145.
KLOTZ, I. M. (1952). In Trends in Physiology and Biochemistry , p. 427 (ed. E. S. G.
Barron). New York: Academic Press.
KROGH, A. (1939). Osmotic Regulation in Aquatic Animals. Cambridge University
Press.
LADD, M. (1952). J. Appl. Physiol. 4, 602.
McCANCE, R. A. (1936). Lancet, i, 643, 704, 765, 823.
McCANCE, R. A. & WIDDOWSON, E. M. (1946). The chemical composition of foods.
Spec. Rep. Ser. Med. Res. Coun., Lond., no. 235, 2nd ed.
MARRIOTT, H. L. (1947). Brit. Med. J. i, 245, 285, 328.
MOON, H. D. (1951). Amer. J. Path. 26, 1041.
NEWBURGH, L. H. (1950). Significance of the Body Fluids in Clinical Medicine.
Springfield: Thomas.
OPIE, E. L. (1948). J. Exp. Med. 87, 425.
OPIE, E. L. (1949). J. Exp. Med. 89, 185.
OPIE, E. L. (1950). J. Exp. Med. 91, 285.
PARRY, A. A. (1936). J. Cell. Comp. Physiol. 8, 277.
PETERS, J. P. (1935). Body Water. Springfield: Thomas.
PETERS, J. P. (1937-8). Harvey Lect. 33, 112.
PETERS, J. P. (1944). Physiol. Rev. 24, 491.
SECRETION AND TRANSPORT OF WATER 6l
PICKEN, L. E. R. (1936). y. Exp. Biol. 13, 387.
POTTS, W. T. W. (1952). Nature, Lond., 169, 834.
RAAFLAUB, J. (1952). Helv. physiol. acta, 10, C. 22.
ROBINSON, J. R. (1949). Biochem.J. 45, 68.
ROBINSON, J. R. (19500;). Proc. Roy. Soc. B, 137, 378.
ROBINSON, J. R. (19506). Nature, Lond., 166, 989.
ROBINSON, J. R. (19520). Proc. Roy. Soc. B, 140, 135.
ROBINSON, J. R. (19526). Nature, Lond., 169, 713.
ROBINSON, J. R. (1953). Biol. Rev. 28 (in the Press).
ROSENBERG, T. (1948). Ada chem. scand. 2, 14.
SABBATANI, L. (1901). J. Physiol. Path. gen. 3, 939.
SCHLIEPER, C. (1930). Biol. Rev. 5, 309.
SCHLIEPER, C. (1935). Biol. Rev. 10, 334.
SCHMIDT, C. L. A. (1945). The Chemistry of the Amino Acids and Proteins, 2nd ed.
Springfield: Thomas.
SCHMIDT-NIELSEN, B., SCHMIDT-NlELSEN, K. & BROKAW, A. (1948). J. Cell.
Comp. Physiol. 32, 361.
SJOSTRAND, F. S. (1953). Nature, Lond., 171, 30.
SMITH, H. W. (1951). The Kidney; Structure and Function in Health and Disease.
Oxford : University Press.
SMITH, H. W. (1952). Fed. Proc. u, 701.
STERN, J. R., EGGLESTON, L. V., HEMS, R. & KREBS, H. A. (1949). Biochem.J. 44,
410.
TROWELL, O. A. (1946). J. Physiol. 105, 268.
TORK, M. (1913). Beitr. path. Anat. 56, 325.
USSING, H. H. (1949). Physiol. Rev. 29, 127.
WEST, E. S. & TODD, W. R. (1951). Textbook of Biochemistry. New York:
Macmillan.
ZOLLINGER, H. U. (1948). Experientia, 4, 312.
ADDENDUM
Some additional comments are required in the light of further discussions
and of publications seen since the foregoing went to press.
Hargitay & Kuhn (1951) showed how a small osmotic gradient main-
tained by some active process could be amplified to yield a larger difference
in osmotic pressure between the fluids entering and leaving a system
allowing counter-current diffusion, and Wirz, Hargitay & Kuhn (1951)
demonstrated a gradual increase in osmotic pressure (measured cryo-
scopically) with increasing depth below the surface of the rabbit's kidney,
as would be expected if the urine were concentrated by such a counter-
current system. More recently Wirz (1953) found, in confirmation of this
suggestion, that the blood in vessels in the renal papilla of the golden
hamster had the same freezing-point as the urine. If this is the mechanism
whereby the urine is concentrated, active transport is still required to
provide a primary osmotic gradient for the counter-current system to
amplify, but the osmotic gradient across the tubular epithelium need not
be so large as was assumed in the discussion on p. 45.
62 SECRETION AND TRANSPORT OF WATER
An important paper by Conway & McCormack (1953) described cryo-
scopic measurements made upon a number of tissues with great care to
avoid complications due to supercooling or autolysis. The results appeared
quite conclusively to refute the hypothesis that the cytoplasm in liver,
kidney or muscle is hypertonic to the surrounding extracellular fluids
during life. On the other hand, Brodsky, Rehm & Mclntosh (1953), who
clearly set out to refute the same hypothesis, and who appear to have taken
similar precautions, found that the osmotic activity of the intracellular
fluids in the liver and kidney of dogs was always greater than that of the
extracellular fluids by up to 50 % , in agreement with older work mentioned
on p. 47.
The excess concentration of total base in respiring cells, mentioned on
p. 53, was too small to account for postulated excesses of osmotic pressure,
and might have been explained by the presence of polyvalent anions in the
cells. More recently Aebi (1953) has reported concentrations of (Na-f-K)
in slices of guinea-pig liver and kidney as much as 50-100% greater than
those in the media in which the slices were incubated. These values are
probably too high to be ascribed to binding by polyvalent anions; indeed,
they are greater than have been found in tissues removed from the body,
but it may be that there is less disturbance of normal relations when a slice
is quickly lifted out of its medium than when samples of tissue are removed
from a dead or anaesthetized animal.
It is somewhat ironical that the pioneers who first suggested that some
of the fixed base in cells is present in the form of osmotically inactive
complexes thought of the amount bound in this way as small. Their
analyses were made upon dead tissues, and there was only a small excess of
base in the cells to account for. If recent chemical analyses of tissue slices
and some recent cryoscopic measurements are both correct, it seems to
follow that up to half the cations inside cells may be present in these
inactive complexes, and it ought to be possible to isolate them and to
elucidate their composition. But the last word can hardly be said until
cryoscopic measurements as careful as those of Conway & McCormack
have been made on the respiring tissue slices which appear to contain so
much base on direct analysis.
REFERENCES TO ADDENDUM
AEBI, H. (1953). Helv. physiol. acta, u, 96.
BRODSKY, W. A., REHM, W. S. & MC!NTOSH, B. J. (1953). J. Clin. Invest. 32, 556.
CONWAY, E. J. & MCCORMACK, J. I. (1953). J. Physiol. 120, i.
HARGITAY, B. & KUHN, W. (1951). Z. Elektrochem. 55, 539.
WIRZ, H. (1953). Helv. physiol. acta, n, 20.
WIRZ, H., HARGITAY, B. & KUHN, W. (1951). Helv. physiol. acta, 9, 196.
OSMOREGULATION AND IONIC REGULATION
IN ANIMALS WITHOUT KIDNEYS
BY J. A. KITCHING
Department of Zoology, University of Bristol
I. OSMOREGULATION AND IONIC REGULATION
IN COELENTERATES
In general the higher and more complicated animals excrete by means of
kidneys, and to flush these there must be an uptake of water from the out-
side. This turn-over of water imposes on the organism a clearly defined
problem in osmoregulation and ionic regulation — that of baling out water
but of conserving valuable solutes, including ions, from the urine, and of
making good from outside those which are lost. However, in small or very
thin animals, provided there is no impermeable cuticle, there seems to be no
reason why dissolved excretory nitrogenous products should not escape
into the medium by diffusion. For instance, this is probably true of
coelenterates and sponges. Even so, many coelenterates have formed an
association with symbiotic algal cells which probably benefit by the supply
of excretory solutes and help to remove them (Yonge, 1931). In any case
there is no morphological evidence nor obvious physiological need for a
continual exchange of water with the outside medium.
The coelenterates are a primitively marine group. They have only a very
few fresh-water representatives, which include Hydra and Limnocnida
(Hyman, 1940) ; Cordylophora lives in brackish as well as fresh water. Little
is known about ionic regulation in the marine coelenterates. The jelly of
Aurelia contains a slightly but definitely higher concentration of potassium
than is present in the outside medium, and this is not due merely to a
Donnan distribution, but must be the result of active transport by the cells
which surround it (Robertson, 1949). There can be little doubt but that
nervous action and muscular contraction depends on ion transport in
coelenterates as much as in higher animals, so that in the coelenterates ionic
regulation presumably came before osmoregulation. If excitation depends
on the entry of sodium as well as the loss of potassium, one might expect
that a fresh-water coelenterate would need to exercise some control over
the ionic composition of its interstitial fluid and mesogloea as well as of the
cytoplasm of its cells.
Nothing was known about osmoregulation in Hydra until the recent
work of Dr Sylvia Lilly, not yet published. I am indebted to her for
64 OSMOREGULATION AND IONIC REGULATION
permission to refer to it on this occasion. She has shown that the tissues and
cells of Hydra are rather highly permeable to water. Isolated tentacles of
H. viridis were allowed to heal up and were then treated with solutions of
sucrose. In 0-05 M-sucrose or stronger, the ectoderm, endoderm and lumen
of the tentacles shrank, and it was also possible to demonstrate shrinkage
of the ectoderm of intact Hydra.
Some information about the content of inorganic ions, and their rate of
exchange, was obtained by Dr Lilly on Pelmatohydra oligactis by the use of
radioactive isotopes. It is important to realize the nature of the information
got by experiments of this kind. The organisms are transferred from the
balanced medium with which they have been equilibrated to another which
is chemically identical but in which a small proportion of the sodium
consists of the radioactive isotope 24Na. Some 24Na enters in exchange for
23Na, until at equilibrium the ratio of 24Na ions to 23Na ions is the same
inside and outside. It is assumed that the organism cannot distinguish
between the two kinds of sodium. Thus from a knowledge of the external
concentration and radioactivity and from a measurement of the internal
radioactivity at equilibrium and the total body water, the average internal
concentration is calculated. If there is any internal unionized sodium which
nevertheless for one reason or another is able to exchange with sodium ions,
that unionized sodium will appear in the estimate obtained. On the other
hand, if any sodium is isolated either morphologically or chemically from
the 24Na ions, it will not appear in the estimate. Thus the estimate obtained
by radioactivity need not be the same as that which would be obtained by
chemical analysis. For instance, according to Abelson & Duryee (1949)
only 12% of the total sodium content of the frog's egg is readily ex-
changeable, the rest exchanging only very slowly ; and it has been suggested
that a part of the sodium content of human erythrocytes is not readily free
to exchange under experimental conditions (Solomon, 1952). It seems
likely that if there is any difference the estimate obtained from radioactivity
is nearer to the ionized sodium than is that obtained by chemical analysis,
although both kinds of information are important. However, the results
obtained for radioactivity must be viewed critically, as the correction for
decay of the isotope can lead to a considerable multiplication of the error.
To return to Pelmatohydra, this organism was found by Dr Lilly to come
substantially into equilibrium in respect of 24Na within about 12 hr. The
average concentration of exchangeable sodium was found to be remarkably
constant over a considerable range of external concentrations extending
both above and below the average level. Hydra was also found to have a
notable power of concentrating potassium, and to a small extent bromide.
These results of course do not give any indication of localization within the
IN ANIMALS WITHOUT KIDNEYS 65
organism; it is possible that sodium will be found predominantly in the
mesogloea and potassium in the cells. In order to postulate a nerve and
muscle physiology comparable with that of higher animals, it would be
necessary to suppose this to be the case and to attribute to the outermost
layer of cells in Hydra tissue the power to secrete sodium into the mesogloea
or interstitial fluid, after the manner of frog's skin. The mesogloea would
function as a primitive body fluid. Autoradiographic technique should
decide this question. It is clear that the internal osmotic pressure exceeds
the external, and that the osmotic uptake of water must be opposed actively
or in some way compensated for by an outward secretion.
II. GENERAL COMMENTS ON PROTOZOA
There can be no doubt that ionic regulation plays an important part in the
Protozoa. Many have conspicuous powers of excitation and contraction,
contraction being localized in the ciliates in myonemes in the body wall
and in some cases in a highly contractile stalk. Practically nothing is known
of ionic concentration or distribution within Protozoa. For instance, it is
not known whether the excitation of myonemes takes place at the plasma
membrane or at an internal surface. The internal osmotic pressure of
various fresh-water Protozoa is believed to exceed considerably that of the
external medium (Kitching, 1951, and earlier papers), and the conductivity
(Gelfan, 1928) suggests that this is largely due to ions. Work being carried
out by Mr L. Carter, and still in an early stage, suggests that Spirostomum
contains much more exchangeable potassium than sodium.
Although the various ions which are vital to life and activity no doubt
contribute the greatest share of the osmotic pressure of the fresh-water
Protozoa, some part must also be taken by proteins and other organic
material in solution. Very little is known of the part possibly played by the
body surface in transporting either ions or water, but it would be surprising
if it did not transport ions. The contractile vacuole is probably concerned
with baling out water which comes into the body by osmosis, and it seems
likely that it possesses the power of retaining for the organism valuable ions.
The physiology of contractile vacuoles has already been discussed at the
1951 symposium (Kitching, 1952), and I propose only to supplement this
account by reference to recent work on osmoregulation in amoebae, and to
the mechanism of vacuolar contraction.
III. OSMOREGULATION IN LABORATORY AMOEBAE
The contractile vacuoles of the various large laboratory amoebae have for a
long time been difficult to fit into the osmoregulation theory of vacuolar
activity. Adolph (1926) found that they continued their activity undi-
66 OSMOREGULATION AND IONIC REGULATION
minished even when the organism was placed in a 0-05 M solution of NaCl
or KC1. I suggested at the 1951 symposium that there might be a con-
siderable time lag before a hypertonic solution reduced the rate of vacuolar
output. Although in the long run an organism cannot continue to bale out
water if none is coming in, it can continue to do so over a short period, and
indeed the resulting shrinkage of the body may actually promote or deter-
mine the reduction in rate of vacuolar output (Kitching, 1951). Actually
I had overlooked some interesting but obscurely published work by Belda
(1942 a, by 1943) on the large multinucleate amoeba Pelomyxa carolinensis.
Belda measured the body volume of Pelomyxa by drawing the organism
into a capillary tube. He found that it shrank when placed in o-i and o-2M
solutions of non-electrolytes, but also shrank slowly and steadily, at a rate of
about 0-33% of the body volume per hour, when kept without food in its
own culture medium. After correction for the effects of starvation, it
appears that the Pelomyxa approached volume equilibrium in the hypertonic
solution in about 12 hr. In spite of the shrinkage of body volume, the con-
tractile vacuole of Pelomyxa went on evacuating water, at a progressively
slower rate, for some 80 min. after the organism has been placed in o-i M
non-electrolyte; and during this time the body volume decreased by about
7-8 % . There are differences of opinion about the internal osmotic pressure
of amoebae. Mast & Fowler (1935) found that Amoeba proteus shrank when
placed in 0-005 M non-electrolyte and regarded this as an upper limit to the
internal osmotic pressure, but Belda (1943) has criticized this conclusion
on the grounds that the shrinkage was probably due to starvation.
Belda accepted Gelfan's (1928) estimate of the conductivity of the cyto-
plasm of A. proteus , namely, o-oi M-KC1, as the best available indication.
However, Levtrup & Pigon (1951), applying the correction for starvation
to Belda's data for shrinkage in o-i M non-electrolyte made up in culture
solution of very low concentration, concluded that the body volume
decreased by exosmosis to about 80% and therefore that the internal
osmotic pressure was that of an 80 mM non-electrolyte. If non-aqueous
materials accounted for any considerable proportion of the body volume,
this estimate would have to be lowered ; but from Belda's data for shrinkage
in 0-2 M solution it appears that the correction must be small. L0vtrup &
Pigon also measured the vapour pressure of Pelomyxa raised in Pringsheim
solution of osmotic pressure about that of 7 mM non-electrolyte. Batches of
Pelomyxa were boiled or frozen in known quantities of distilled water. On
four samples they obtained closely agreeing values averaging 107 mM non-
electrolyte. They point out that this value may be too high owing to
solution of additional material from crystals, and compromise on an average
between their determination and that which they derive from Belda's work,
IN ANIMALS WITHOUT KIDNEYS 67
thus obtaining an estimate for the osmotic pressure corresponding to
94 mM non -electrolyte — very much greater than that of the external
medium.
For their study of the water relations of Pelomyxa, Lovtrup & Pigon took
advantage of the possibility of using isotopic water for the estimation of the
diffusion constant of water in the surface membrane. They also derived a
relation between the diffusion constant of Pick and the permeability con-
stant of Jacobs for water, so that from a knowledge of Pick's diffusion con-
stant and of the difference in osmotic concentration on the two sides of the
membrane it was possible to estimate the rate of osmotic entry of water into
the organism. Under conditions of steady state this should equal the rate of
vacuolar output.
The rate of penetration of isotopic water through the surface membrane
of Pelomyxa was determined by L0vtrup & Pigon by means of the Cartesian
diver. Pelomyxa previously equilibrated in a medium containing heavy
water was transferred to the Cartesian diver in a drop of medium containing
only ordinary water, or vice versa; and the changes in reduced weight were
followed.
The diffusion constants for H2O, D2O and H2O18 calculated from these
experiments were not significantly different, and averaged 2-6 x io~5 cm./
sec., from which was calculated a permeability constant of o-oi i //3///,2/atm./
min. This is of the same order as that found in osmotic experiments for
many other cells, but is at the low end of the range. Prom this permeability
constant, the surface area of the Pelomyxa, and the difference of osmotic
pressure, the rate of osmotic inflow of water was calculated to be about 2%
of the body volume per hour, or perhaps slightly more. Belda found that
the contractile vacuole evacuated 3-8% of the body volume per hour. Thus
Pelomyxa differs from other Protozoa in which water relations have been
studied only in that it is slow to react to a change in external osmotic
pressure. This may be ascribed partly to the relatively low surface area,
partly to its rather low permeability to water, and partly to a rather low
sensitivity to the changes in body volume by which the contractile vacuole
appears to be regulated.
IV. CONTROL OF RATE OF VACUOLAR OUTPUT
A study of contractile vacuoles may contribute knowledge of the control
of the secretion of water which might have wide applications in other
organisms or tissues, and therefore I have attempted to extend the sugges-
tions made about this at the 1951 symposium. On that occasion I suggested
that when the external osmotic pressure is changed, any lag in the response
of the contractile vacuole to the new rate of entry of water from the medium
5-2
68 OSMOREGULATION AND IONIC REGULATION
into the organism will cause a change in body volume. This small change
in body volume might well mediate the change in rate of vacuolar output.
It was estimated that a decrease of i|% in the body volume was
associated with complete stoppage of the contractile vacuole in the
suctorian Discophrya piriformis.
The rate of vacuolar output of the peritrich ciliate Carchesium aselli is
also much affected by temperature (Kitching, 1948 a)y and there are reasons
for supposing that temperature has an equivalent effect on the permeability
of the body surface to water; so that the body volume, which depends on a
balance of these two, remains unchanged (Kitching, 19486). A study has
now been made of the process of adjustment to a sudden change of
temperature in the suctorian Discophrya piriformis. The response of the
vacuole of this organism to a change of conditions is rather slow, so that a
change of temperature which is relatively abrupt can be imposed.
The results of a rather drastic change of temperature are shown in Fig. i.
A rise in temperature from 6 to 20° C. caused an almost immediate stoppage
of vacuolar activity, and several small kinks in the body surface gradually
filled up. 'Then vacuolar activity was resumed at a rate of output con-
siderably above the original. Thus the interpretation of the adjustment of
the organism to the new temperature is rather hopelessly complicated by an
effect on the secretory mechanism.
This effect may be exercised on the structural arrangement of proteins
or lipoproteins supposedly concerned in secretion (Kitching, 1951). For
instance, Goldacre (1952) has suggested that the folding of protein mole-
cules offers a mechanism for active transport, and Marsland (1950) has
shown that the viscosity of the cytoplasm of Arbacia eggs is increased by a
rise of temperature. Without pressing this particular interpretation too far,
we might suggest that high temperature opposes the structural change by
which secretion is brought about, and so depresses or temporarily halts
secretion, until accumulating reactants of the chemical processes involved
and increasing hydration of the cytoplasm once more force the process
forwards.
On return to a low temperature, in the experiment illustrated in Fig. i,
the rate of vacuolar output remained high for a short time, but then fell.
The body surface became wrinkled, and it was clear that there was a
decrease in volume. The shrinkage shown in Fig. i is exceptionally great.
It conforms with the hypothesis already advanced concerning the control
of vacuolar activity, but does not add anything to it. However, the experi-
ments also provide interesting information about the frequency and ultimate
diameter of the contractile vacuole, which has led me to a reconsideration
of the mechanism of systole.
IN ANIMALS WITHOUT KIDNEYS
20
15
U
10
Temperature
u20
•requency
Ultimate
diameter
$ 3
2
s 1-
Rate of output
60
Time in minutes
120
Fig. i. Effects of a sudden change of temperature on the vacuolar rhythm of Discophrya
piriformis Guilcher. This graph is plotted from data summarized by Kitching (19546) else-
where. The tentacles have been omitted from the drawings.
70 OSMOREGULATION AND IONIC REGULATION
V. MECHANISM OF SYSTOLE
Since the 1951 symposium my views on the mechanism of systole have
moved further in favour of a contractile vacuolar wall. The vacuolar wall is
visible as a clear layer, about Jy/ thick in Amoeba proteus (Mast, 1938). If
an A. lacerata is squashed the contractile vacuole may persist for some time
suspended freely in water (Hopkins, 1946). The vacuolar wall of Amoeba
is weakly birefringent, and the birefringence disappears at systole (Schmidt,
1939). These facts signify that the vacuolar wall is something a good deal
thicker than the classical permeability barrier of the cell membrane, and it
seems likely that the extra thickness is made up by a structural protein
layer. A structural layer would clearly be needed to maintain the rather
complicated shape of the permanent vacuolar apparatus in certain ciliates.
Although any further discussion of this layer is bound to be highly specula-
tive, it is natural to suggest a structure in which the long axis of the protein
molecules lies in the plane of the wall, whether or not these molecules are
contracted or partly contracted concertina-wise as in Mitchison's (1952)
model. A structural protein layer might well possess elasticity, so that the
vacuolar wall would become subject to increasing tension as the vacuole
grew, and it might also possess the power of contraction (as suggested by
Schmidt, 1939) under suitable chemical circumstances.
Body turgor is unnecessary for systole (Kitching, 1952). This is shown
again by the effects of hydrostatic pressures of 2000-3000 Ib./sq. in.
(136-204 atm.) on the suctorian Discophrya piriformis (Fig. 2). The body
surface has been thrown into creases (Kitching, 19540), but tne contractile
vacuole undergoes systole quite normally and with increased frequency.
Surface tension could easily provide the small pressure required if the
vacuolar wall were liquid, but if there is in fact some rigidity of structure in
the vacuolar wall it would probably be quite ineffective. MacLennan (1933)
has described in Ophryoscolecidae a rounding-up of the contractile vacuole
shortly before systole; then after some seconds the vacuole ' undergoes a
sudden convulsion* and discharges. There is also a rounding-up sometime
before systole in various peritrich ciliates. MacLennan has reported that
at the same time as the vacuole rounds up there is a solation of the neigh-
bouring cytoplasm, as judged by Brownian movement, and he attributes
the rounding-up to this solation. It seems very possible that the tension in
the vacuolar wall which is necessary for this rounding-up is provided by
the structural layer suggested above, and that a contraction of this is
responsible for the 'sudden convulsion' of the vacuole at the beginning
of systole. In any case it is necessary that the vacuolar wall should disinte-
grate as the vacuole contracts, and it is likely that in this process protein
IN ANIMALS WITHOUT KIDNEYS
71
4)
r •: |
2. 'i *>
o^ 1 fc>
fi 1
cv
! ? •',
1 •' '. -
M
L
*~ o
5
"o
o
s?
i
ti
• .
• 1 '
* 1 "~
>
O 3J
1
!
2'
0)
3
| '|'
', '
!
! '<
rt I,
•'. \
<4-4
a 11
O H
CJ ~
1
• i
: . i
§1
1 ii
i *
^
'i
!• i-
.'(
l
'•
1 1 - ;
'1
,
!•' •' .' •
S o|
3
E 1 1
1 ; •••• ;
c/)
S
a
<
CQ
£
VM
O
§
V)
a •:••'',
• 1
'•J '•:-. -
U
00 w
. *
•*i • • • '.
ri
i
: ••': .''•,'"
do
£
0
>3*>]
0 -|_ 2 !
i * °-i i li ii
^ « » _ _l
1 ,
*. 1 1 l 1 1 ' I. 1 1
? ^OOOL^uJL, °Wt
siauueip wumifi ( ws/^y) indino jo ;
72 OSMOREGULATION AND IONIC REGULATION
molecules will fold and become globular and so pass into the surrounding
hyaloplasm, much as postulated by Goldacre & Lorch (1950) for the tail
of amoebae. During this process tension might be developed by the con-
traction of oriented proteins, or the dissolution of the structural layer might
place the vacuolar boundary under the control of surface forces. As a
speculation, it seems to me likely that the cell will use the contractility of
proteins, which is their outstanding attribute, for the contraction which it
must carry out. The remarkably complicated action of the contractile
vacuoles of certain ciliates such as Paramecium (King, 1935) and Hapto-
phrya (MacLennan, 1944) can more readily be explained if waves of con-
traction are admitted.
In Amoeba the contractile vacuole obviously must pass into the plasmagel
before it discharges, and usually it contracts when in the tail. Here the
surrounding plasmagel is itself contracting, and any cyclic changes of the
chemical environment of the vacuole are likely to have reached a stage
favouring contraction. In organisms with stationary contractile vacuoles,
any such cyclic changes would have to take place at the vacuolar site.
However, the rounding-up of the vacuole and the solation of the surrounding
protoplasm described by MacLennan occur some 5-10 sec. before systole,
and are regularly associated with it. It is reasonable to regard them as
closely associated with some rhythmic process which sets the pace for the
vacuolar cycle. Modification of the vacuolar cycle might be caused by
modification of this rhythmic process. For instance, the characteristic over-
shoot in the depression of vacuole frequency, which occurs in Discophrya
piriformis when the temperature is lowered suddenly (Fig. i), might be
ascribed to an effect of temperature on a series of chemical reactions which
sets the vacuolar rhythm. Strong depression of one member of the series
would slow the rhythm until the reaction products of preceding members
accumulated and forced the process on again (Burton, 1939). Wells &
Ledingham (1940) have made a model with water taps, a siphon and a
kymograph to illustrate the application of this principle to rhythmic
processes.
Let us consider now the characteristics of the vacuolar cycle at constant
temperature. If the external osmotic pressure is lowered, the organism
responds by an increase in rate of vacuolar output which is associated with
an increase both in the frequency and in the ultimate diameter of the con-
tractile vacuole. In some way both frequency and ultimate diameter are
affected. Several mechanisms might be suggested: for instance, the small
increase in body volume which supposedly mediates a change in the rate
of secretion might also alter the condition of the pore plug and vacuolar
wall in such a way as to make them more sensitive to the cyclic changes by
IN ANIMALS WITHOUT KIDNEYS 73
which contraction is supposedly initiated, or the degree of extension of the
vacuolar wall itself might influence the sensitivity of the mechanism to these
cyclic changes. In either case, provided that the increase in frequency falls
short of the increase in rate of secretion, the ultimate diameter would also
be increased.
We may now return to Fig. i and consider the effect of a sharp rise in
temperature on the vacuolar frequency. There is an immediate increase in
vacuolar frequency. In fact, an overshoot would be expected, and is some-
times seen. However, the depression in secretion sets in and the frequency
also falls, being linked with the rate of secretion. Only when the secretory
activity is restored can the frequency attain the rate characteristic for the
temperature.
The effects of high pressure on vacuolar frequency (Fig. 2) are interesting
because they parallel those found many years ago on frog's heart by
Edwards & Cattell (1928). At moderate pressure (2000-3000 Ib./sq. in.)
the vacuolar frequency is markedly increased, but at 5000 Ib./sq. in. and
upwards vacuolar secretion is strongly depressed, and with it the frequency
also. Landau & Marsland (1952), in discussing the effects of hydrostatic
pressure on cultures of frog heart, have suggested a general explanation in
terms of a differential inhibition both of an enzyme-catalysed reaction and
of the denaturation of that enzyme. Although this interpretation would
give additional support to the idea of a rhythmic chemical process governing
vacuolar contraction, not enough is known to justify its application at
present.
VI. CONCLUSION
I will conclude with a return to the more general aspects of this paper. The
fresh-water coelenterates appear to control their water content by an un-
known but active mechanism and to take up various ions from the external
medium after the manner of various other fresh-water animals (Krogh,
1939). Little is known about sponges, although some fresh-water sponges
have contractile vacuoles (Jepps, 1947). The Protozoa control their water
content in many cases by contractile vacuoles, and very probably will be
found to take up ions through the body surface or part of it. It is not clear
why Protozoa should have localized the secretion of water, but it is possible
that in many cases the presence of a cuticle is concerned. This might well
interfere with the outward secretion of water by the plasma membrane,
and might also hinder the diffusion of dissolved excretory matter into the
surrounding medium, thus making necessary an organelle with the function
of a kidney. This view is proposed by MacLennan (1933) for the Ophryo-
scolecidae, into which the penetration of vital dyes was found to be much
faster during feeding; they have a thick cuticle. The cuticle plays an
74 OSMOREGULATION AND IONIC REGULATION
important part in structural differentiation in the Protozoa, and it is suggested
that its presence has rendered necessary the localization of secretion, and
possibly excretion, at an internal plasma membrane, the vacuolar surface.
REFERENCES
ABELSON, P. H. & DURYEE, W. R. (1949). Radioactive sodium permeability and
exchange in frog eggs. Biol. Bull., Woods Hole, 96, 205-17.
ADOLPH, E. F. (1926). The metabolism of water in Amoeba as measured in the
contractile vacuole. J. Exp. Zool. 44, 355-81.
BURTON, A. C. (1939). The properties of the steady state, compared to those of
equilibrium as shown in characteristic biological behaviour. J. Cell. Comp.
Physiol. 14, 327-49.
BELDA, W. H. (19420). Permeability to water in Pelomyxa carolinensis. I. Changes
in volume of Pelomyxa carolinensis in solutions of different osmotic concentra-
tion. Salesianum, 37, 68-81.
BELDA, W. H. (19426). Permeability to water in Pelomyxa carolinensis. II. The
contractile vacuoles of Pelomyxa carolinensis. Salesianum, 37, 125-34.
BELDA, W. H. (1943). Permeability to water in Pelomyxa carolinensis. III. The
permeability constant for water in Pelomyxa carolinensis. Salesianum, 38, 1 7-24.
EDWARDS, D. J. & CATTELL, McK. (1928). The stimulating action of hydrostatic
pressure on cardiac function. Amer. J. Physiol. 84, 472-84.
GELFAN, S. (1928). The electrical conductivity of protoplasm. Protoplasma, 4,
192-200.
GOLD ACRE, R. J. (1952). The action of general anaesthetics and the mechanism of
response to touch. Syrnp. Soc. Exp. Biol. 6, 128-44.
GOLDACRE, R. J. & LORCH, I. J. (1950). Folding and unfolding of protein molecules
in relation to cytoplasmic streaming in amoeboid movement and osmotic
work. Nature, Lond., 166, 447.
HOPKINS, D. L. (1946). The contractile vacuole and the adjustment to changing
concentrations in fresh water amoebae. Biol. Bull., Woods Hole, go, 158-76.
HYMAN, L. H. (1940). The Invertebrates: Protozoa through Ctenophora. New York.
JEPPS, M. W. (1947). Contribution to the study of the sponges. Proc. Roy. Soc. B,
134,408-17.
KING, R. L. (1935). The contractile vacuole of Paramecium multimicronucleata.
J. Morph. 58, 555-72.
KITCHING, J. A. ( 1 948*7). The physiology of contractile vacuoles. V. The effects of
short-term variations of temperature on a freshwater peritrich ciliate. J. Exp.
Biol. 25, 406-20.
KITCHING, J. A. (19486). The physiology of contractile vacuoles. VI. Temperature
and osmotic stress. J. Exp. Biol. 25, 421-36.
KITCHING, J. A. (1951). The physiology of contractile vacuoles. VII. Osmotic
relations in a suctorian, with special reference to the mechanism of control of
vacuolar output, y. Exp. Biol. 28, 203-14.
KITCHING, J. A. (1952). Contractile vacuoles. Symp. Soc. Exp. Biol. 6, 145-6.
KITCHING, J. A. (1954*2). The effects of high hydrostatic pressure on a suctorian.
J. Exp. Biol. 31 (in the Press).
KITCHING, J. A. (19546). The physiology of contractile vacuoles. IX. Effects of
sudden changes in temperature on the contractile vacuole of a suctorian ; with
a discussion of the mechanism of contraction. ,7. Exp. Biol. 31 (in the Press).
KITCHING, J. A. (1954 c). The physiology of contractile vacuoles. X. Effects of
high hydrostatic pressure on the contractile vacuole of a suctorian. y. Exp.
Biol. 31 (in the Press).
IN ANIMALS WITHOUT KIDNEYS 75
KROGH, A. (1939). Osmotic Regulation in Aquatic Animals. 242 pp. Cambridge
University Press.
LANDAU, J. & MARSLAND, D. (1952). Temperature-pressure studies on the cardiac
rate in tissue culture explants from the heart of the tadpole (Rana pipiens).
J. Cell. Comp. Physiol. 40, 367-81.
LOVTRUP, S. & PIGON, A. (1951). Diffusion and active transport of water in the
amoeba Chaos chaos L. C.R. Lab. Carlsberg, 28, 1-28.
MACLENNAN, R. F. (1933). The pulsatory cycle of the contractile vacuole in the
Ophryoscolecidae, ciliates from the stomach of cattle. Univ. Calif. Publ. Zool.
39, 205-50.
MACLENNAN, R. F. (1944). Tne pulsatory cycle of the contractile canal in the
ciliate Haptophrya. Trans. Amer. Micr. Soc. 63, 187-98.
MARSLAND, D. (1950). The mechanisms of cell division; temperature-pressure
experiments on the cleaving eggs of Arbacia punctulata. J. cell. comp.
Physiol. 36, 205-27.
MAST, S. O. (1938). The contractile vacuole in Amoeba proteus (Leidy). Biol. Bull.,
Woods Hole, 74, 306-13.
MAST, S. O. & FOWLER, C. (1935). Permeability of Amoeba proteus to water.
J. Cell Comp. Physiol. 6, 151-67.
MITCHISON, J. M. (1952). Cell membrane and cell division. Symp. Soc. Exp. Biol.
6, 105-27.
ROBERTSON, J. D. (1949). Ionic regulation in some marine invertebrates. J. Exp.
Biol. 26, 182-200.
SCHMIDT, W. J. (1939). Uber die Doppelbrechung des Amobenplasms. Proto-
plasma, 33, 44-9.
SOLOMON, A. K. (1952). The permeability of the human erythrocyte to sodium and
potassium. J. Gen. Physiol. 36, 57-110.
WELLS, G. P. & LEDINGHAM, I. C. (1940). Studies on the physiology of Arenicola
marina L. II. Accommodation to magnesium concentration in the isolated
extrovert. J. Exp. Biol. 17, 353-63.
YONGE, C. M. (1931). The significance of the relationship between corals and
zooxanthellae. Nature, Lond., 128, 309-10.
THE ACTIVE TRANSPORT OF WATER
UNDER TEMPERATURE GRADIENTS
BY D. C. SPANNER
Plant Physiology Department, Imperial College of
Science and Technology, London
I. INTRODUCTION
In presenting this review on a particular aspect of active transport of water
I want first of all to discuss the thermodynamics of transport processes in
a general way to try to bring out more clearly the distinction between active
and passive mechanisms, and to show just where the phenomenon discussed
in this paper fits into the general picture. The procedure I want to adopt
follows that of the relatively new Thermodynamics of Irreversible Processes
which is destined, no doubt, to play an increasingly important part in
biological theory (see, for instance, de Groot, 1951).
The meaning of ' active transport '
In a general way the meaning of * active transport' is fairly clear. Whenever
movement of matter occurs in a direction which we cannot explain as
helping towards the attainment of equilibrium we recognize a case of active
transport. Very often it is an extremely easy matter to decide whether we
have such a case or not, but sometimes it is not at all simple, and our usual
definitions land us in anomalies or fail us altogether. Consider one or two
practical examples. Electro-osmosis can be either positive or negative, i.e.
it can promote the movement of water in a direction either helping, or
hindering, the normal osmotic flow. Now in both cases there is reason to
believe that the mechanism is essentially similar; consequently if negative
electro-osmosis is regarded as active (as it must be), so must positive electro-
osmosis. In other words, an active movement can take place down a gradient
of chemical potential or activity, as well as up one. Thus the very common
definition of 'active transport' as being transport against a gradient of
activity fails us — it expresses only half of the truth. Then consider the case
to be discussed in this paper, where the temperature is not uniform. Here
no special significance attaches to activity gradients at all; in fact, the
difference in activity is only unambiguously known when the temperature
is uniform. This makes the usual criterion of active movement really
meaningless — a most unsatisfactory position. A knowledge of the real
distinction between active and passive movements is obviously desirable,
TRANSPORT OF WATER UNDER TEMPERATURE GRADIENTS 77
and I think it can be found by following the general approach introduced,
I believe, by the Norwegian physicist Onsager in 1931.
Consider a very simple type of system (Fig. i) divided into two equal
halves by some sort of membrane, and containing water and a number of
other substances. Imagine it at first to be in absolute equilibrium, so that
the temperature, pressure and the concentrations of all its constituents are
uniform everywhere. Further, let it be imagined to be entirely isolated, so
that neither matter nor energy can enter or leave it.
Suppose we proceed now to disturb its equilibrium in a number of
different ways. First, each one of the constituents can be redistributed
between the two halves so that instead of being equal in amount on both
sides it is now present in unequal amounts. Thus if glucose is a constituent
there may be a mass m J on one side of the membrane and a mass m1^ on the
II
Fig. i.
other, and the extent to which we have upset the equilibrium can be specified
by the difference m1* — m* = Amg, which of course was originally zero. For
each chemical substance present we shall have one degree of freedom in
disturbing the equilibrium, and when we have finished, in imagination,
displacing matter from one side to the other the state of our system can be
specified by the variables Am1, Aw2, A/«3, ..., Aw^, of which there will be
one for each distinct sort of substance present. We can go further than this,
however, and imagine that heat energy is made to flow from one side to the
other, setting up a temperature difference A T between them. As a matter of
fact, there will be many other ways still of disturbing the equilibrium (such
as producing an electrical potential difference between the two sides, or
promoting chemical changes), but for the purpose of illustration the matter
need be taken no further.
We now have a system not in equilibrium whose condition is completely
specified by a number of variables A/WJ, Aw2, ..., ^mk and AT", and these,
moreover, are all independent. In general, of course, our manipulations
will have produced a pressure difference AP between the sides as well, but
78 THE ACTIVE TRANSPORT OF WATER
it will not be necessary to specify it, for it is already implied. In other words,
it is not a further independent variable. We can if we so desire include it in
the specification, but then not all our variables will be independent; one of
them, say &mk, can be left out.
'Fluxes' and 'forces'
The foregoing example illustrates the point that when a system is not
in equilibrium its momentary condition can be specified by means of
a certain minimum number of variables ax, a2, ..., an, all of which are
independent. There is a certain latitude in choosing these, but their number
is always the same, however they are chosen.
Consider now how the system returns to equilibrium. The one thing that
can be said about it from the thermodynamic point of view is that its entropy
continuously increases; it manifests a continuous tendency to increase its
entropy, and when it can do so no more it comes to rest. It would seem
reasonable therefore to regard the derivative dS/da as measuring the
tendency of a to change, i.e. as a measure of the ' force ' changing it. Further,
the time-derivative da/dt can obviously be spoken of as the ' flux ' of a. To
make this clearer, however, let us refer again to the previous example. If
a = &ma measures the extent to which the glucose distribution is 'out of
balance ', then dS/da can be regarded as the force promoting transport of
glucose and da/dt as the rate of glucose transport.
Now the rate at which the entropy of the whole system changes can be
written g5 ss ds 95 ^
by the basic property of partial differential coefficients, the partial derivative
dS/dav for instance, being taken with all the other variables a2, a3, ..., an
constant. The quantity dS/dt can thus be split up into the sum of a number
of products of ' forces ' such as Xl = dS/dal and * fluxes ' such as/x = da^dt ;
furthermore, it is bound by the second law of thermodynamics to be positive.
Calling dS/dt=cry equation (i) can therefore be written in the more concise
form <r = XJ1 + X2J2 + XJ3+...+XJn, (2)
a relation which is rather analogous to the electrical one
watts = volts x amperes. (3)
Equation (2) expresses how rapidly the system is increasing its entropy in
terms of the instantaneous values of the forces and fluxes. However, these
quantities are not independent; they can be related together in a way
analogous to that in which the volts and amperes of equation (3) are related
by Ohm's law: , , ^ • A / \
J amperes = volts x conductance coefficient. (4)
UNDER TEMPERATURE GRADIENTS 79
The 'Ohm's law' relations have, however, to be much more general, for it
is a common observation that the movement of any one kind of matter
promotes to some extent movement of all the other kinds present. Thus the
flux 7i will in general depend not only on the force Xl but also on all the
other forces X2y X3 and so on. Thus in general we must write
J, = L^Xi + LuXz + LuXs+...+LlnXny (5)
and similarly for the other fluxes, the coefficients Ln, L12, ... being 'con-
ductance' coefficients whose magnitude will depend on the size and geo-
metry of the system and on the nature of the mechanism of transport across
the dividing membrane. By means of equations such as (5) (which are purely
descriptive like Ohm's law), we can express the contribution of, say, the
flux/j to the general process of increase of entropy. Calling this contribution
crj we have from (2) and (5)
Vi = XlJl = LllXl + L12XlX2 + Ll3XlX3+...+LlnXlXn, (6)
which is analogous to
watts = (conductance of circuit) x volts2. (7)
Now the interesting thing about equation (6) is that the only term on the
right which is necessarily positive is the first one ; the others may clearly be
either positive or negative. In other words, that part of the f\uxjl = dal/dt
which can be attributed to the conjugate force Xl = dS/dotl necessarily
contributes to an increase in the entropy of the system ; that part which can
be attributed to the non-conjugate forces, d*Sy?a2, dS/da^ and so on, may
contribute towards either an increase or a decrease in the entropy. This
distinction would therefore seem to be just what is wanted to mark the
division between passive and active movements; the passive component is
due to the conjugate force, the active to the non- conjugate.
As a matter of fact this division is not entirely unambiguous, on account
of the fact mentioned earlier that the variables a1? a2, a3, . . . , an can be chosen
in a variety of ways. However, the ambiguity can be overcome if the a's
are so chosen that, in the simple example given, the fluxes doc/dt represent
the rates of transfer of the individual chemical substances present together
with the rate of flow of heat. Suppose this is done in a system containing
water and one other chemical substance, say glucose. Then the passive
movement of water on the present analysis proves to be associated with the
force lvw AP + I -j~ -I Ac , where vw is the partial volume of water, AP
I \ccg/T,i> }
the pressure difference, fiw the chemical potential of water and cg the glucose
concentration per unit mass of mixture. Needless to say it represents no
departure to attribute the nature of passive agents to the pressure and
8o THE ACTIVE TRANSPORT OF WATER
concentration gradients,! though ifvw should happen to be negative instead
of positive the pressure gradient will operate in the reverse direction.
The active movement of the water likewise proves to be due to the forces
AT, the temperature differential, and z^AP-f -} Ac, where vg is
I \VC(//T,P I
the partial volume of glucose and cg is again the glucose concentration per
unit mass. This emphasizes that any movement of water caused by a gradient
in the chemical potential of another substance is necessarily an active
movement; and that a pressure gradient has an active effect as well as a
passive one, the active effect being dependent on va AP. Further, and this
will be the subject of the rest of this paper, any movement of water brought
about by a temperature difference must be regarded as an active one, the
criterion being, as mentioned earlier, whether it contributes necessarily or
not to an increase in the entropy of the system.
II. THE IMPORTANCE OF TEMPERATURE
It is usually considered that as far as makes no difference living cells are
isothermal systems. This is probably because the process of thermal
conduction over minute distances is so extremely rapid that it is difficult to
conceive of two points close together — say on either side of the plasma
membrane — possessing temperatures measurably different. It seems to me
that this view is in some respects inadequate, and elsewhere (Spanner,
1953) I have put forward the conception that where metabolism is active
any given point in the cell possesses not one temperature, but many ; in fact,
one to each different species of molecule present. What makes this view at
least a possible one is the fact that though thermal equilibration over mole-
cular distances may be an extremely rapid process, yet there are other
processes, such as chemical reaction and diffusion, which are comparable
in rapidity. Thus these processes may be able to 'take advantage* of micro-
scopic temperature fluctuations before thermal conduction has evened
them out.
An important fact in connexion with the active transport of water under
temperature gradients lies in the quite unexpectedly great effect of small
temperature differences. Consider a very simple model. Two open Petri
dishes I and II (Fig. 2) contain an aqueous solution of osmotic pressure TT.
Initially both are at the same temperature, and we can suppose them
enclosed in a larger sealed vessel. If now dish I be raised in temperature from
T to T-h AT1, the vapour pressure above it will be increased and water will
evaporate from it and condense in dish II, which will become more dilute
t These two gradients can be included together as a gradient of chemical potential.
UNDER TEMPERATURE GRADIENTS 8l
in consequence. This process will go on till the dilute and cooler solution
in dish II has the same vapour pressure as the stronger and warmer solution
in dish I, when the system will come in a sense to a standstill. There will then
be a difference in osmotic pressure ATT between the two solutions, and this
difference can broadly be said to balance the temperature difference AT.
The relation between ATT and AT can easily be shown to be (Spanner, 1952)
A77-A (8)
AT~FT' ()
where L is the latent heat of vaporization and V is the molar volume of water,
i.e. about 18-0 c.c. A simple calculation from this shows that a temperature
difference of only Y^O° C. can cause a pressure difference of over four-fifths
of an atmosphere.
Fig. 2.
A general thermodynamic relation
The simple example just given is a particular case of a general relation
governing transport under a temperature gradient. The system we con-
sidered consisted of two homogeneous parts I and II separated by a divisional
wall represented by the air space. In general this type of system can be
represented by such a diagram as Fig. 3. It is called a discontinuous system
because its properties change abruptly on crossing the membrane from one
side to the other, though within each section all properties are uniform.
Suppose that in such a system (containing for simplicity, water only)
a temperature difference AT is imposed between the two sides. In general
this will cause a movement of water from one side to the other, and provided
the walls are solid this movement will build up a pressure difference AP
which will ultimately bring the flow to a standstill. When this occurs the
system is said to be in a ' steady state', and the relation between AP and AT
will be given by the perfectly general equation
AP £*
AT~ FT' '9'
82
THE ACTIVE TRANSPORT OF WATER
where V is the volume of unit mass of water and Q* is a quantity called the
heat of transfer'. In a sense therefore equation (9) gives the equivalence
of temperature and pressure differences in promoting flow across the
membrane ; if the right-hand side is large then a very small temperature
difference can cause the same rate of flow as a much larger pressure
difference. The key to the situation obviously lies in the interpretation of
the quantity £)*, which must now be discussed.
II
Fig. 3-
III. THE 'HEAT OF TRANSFER' Q*
Imagine that the whole system represented by Fig. 3 is at a uniform
temperature, but that owing to a small pressure difference (ideally an
infinitesimal one) water is flowing across the membrane. In general, the
process by which the water traverses the membrane will act differentially
on the faster and slower molecules so that the water which crosses is not
usually a representative sample but contains either a larger, or a smaller,
proportion of ' hot' or ' cold ' molecules than the bulk. This is very obviously
the case in the example previously given, where the membrane is an air space
and the process of crossing it involves, first, evaporation and, secondly,
condensation. Thus, as the water moves across, it carries a quantity of heat
with it, and while compartment I is left cooler, compartment II grows
warmer — again a result which is very evident in the evaporation example.
Suppose now that as the water flows across the membrane heat is con-
tinuously abstracted from compartment II to maintain its temperature
constant. Then the quantity of heat so abstracted per unit mass of water
flowing is the 'heat of transfer', Q*. In the earlier example (Fig. 2) it is
obviously very nearly the same as the latent heat of evaporation, showing
that equation (8) is a particular case of equation (9).
UNDER TEMPERATURE GRADIENTS 83
Conditions for large 'heat of transfer*
It can be seen at once that the condition which must be fulfilled if there
is to be a definite heat of transfer is that the membrane should act in such
a way as to distinguish between * hot * and * cold ' water molecules. In general,
there are two ways in which this can come about, and these two incidentally
correspond to the two classical theories of cell permeability. In the first, the
membrane may act as a sieve. If the pores of this are coarse, no selective
effect is apparent; water flowing down a main does not tend to get hotter as
it goes. On the other hand, if the pores are really minute, in fact comparable
to the intermolecular distance of water, then the faster molecules penetrate
in relatively greater numbers than the slower ones, and a 'heat of transfer*
appears. This is particularly easy to show with gases and a fine-grained
porous pot. The magnitude of £)*> however, is small, and even in the most
favourable cases it never numerically exceeds \RT per mole.
In the second mechanism, the membrane acts as a potential energy
barrier. This means that it does not merely act less favourably towards
low-energy molecules (as the sieve does); it positively turns them back.
Thus in the example of Fig. 2 only those water molecules with energies
sufficient to overcome the latent heat forces can escape into the * membrane ' ;
all with lower energies are held back. From this it can at once be seen that the
* heat of transfer* can be very large in such cases; in fact it is very nearly the
same as the height of the potential energy barrier, i.e. as the * activation
energy* required to cross the membrane.
This second type of membrane may take many different forms. It may
be simply an air space with transport occurring in the state of vapour; or it
may impose chemical reactions as a condition of crossing, as when some
gases diffuse across metallic diaphragms. A very important type of mem-
brane involves what can be called 'solution in an uncongenial medium*, and
it is here that we can probably place the case of water molecules penetrating
the lipoid membrane of the protoplast. Still another type of potential
energy barrier is the electrical one. This is exemplified by a charged
membrane exposed to ions, and illustrates the fact that the barrier can be
either positive or negative, i.e. it can either oppose or facilitate crossing. In
the latter case the 'heat of transfer* may be negative, since the barrier
(which might better be called a 'ditch* in this case) acts to promote the
crossing of disproportionate numbers of few-speed molecules.
Whatever the mechanism of crossing, however, equation (9) holds,
exemplifying in this respect the typical nature of a thermodynamic result.
However, before any practical use can be made of this relation it is obviously
necessary to know the order of magnitude of Q*. Where the mechanism of
6-2
84 THE ACTIVE TRANSPORT OF WATER
'permeability' is known it may be possible to arrive at this from considera-
tion of existing and well-known data, as in the case of evaporation, where it
is equal roughly to the latent heat. As a matter of fact, as mentioned earlier,
in all cases Q* is equal very nearly to the height of the potential energy
barrier (if this is large); but as this is very rarely known, some other way of
evaluating it must be found.
IV. MEASUREMENT OF THE 'HEAT OF TRANSFER'
Fortunately, Q* can be calculated very simply from a knowledge of the
temperature dependence of the permeability. As a semi-empirical but fairly
exact relation Arrhenius showed that the equation
r) 7?
g^ln(rate) = £Tl (10)
can be used to describe the variation of rate processes with temperature, E
being the 'activation energy' of the process. Originally this equation was
applied to chemical reactions, but it can be used for transport and other
processes as well. It has been regarded till fairly recently! as only having
a partial justification in theory, but it appears (Appendix) that it can be
treated as quite exact if E, the 'activation energy', is replaced with £)*,
the heat of transfer. When this is done and the ordinary permeability // is
introduced into the equation we get the exact result
where V is the partial molar volume of water. Since V is nearly constant it
can be omitted, and assuming that Q* does not vary greatly with temperature
equation (n) becomes on integration
*
Writing 7^ « T2 = 71, jT2 — Tt = 10 and introducing the temperature coefficient
Qw of permeability we get the final result
In (1-034 Qlo) (13)
at ordinary temperatures. This result gives us a simple way of evaluating
the heat of transfer across the cell membrane.
f The theory of absolute reaction rates developed by Eyring and others since 1935 gives,
of course, a very satisfactory basis for this equation.
UNDER TEMPERATURE GRADIENTS 85
V. ACTIVE TRANSPORT OF WATER ACROSS THE
CELL MEMBRANE
Broadly speaking, the cell can be regarded as a discontinuous system of the
type described, with an internal watery phase separated from an external
one by a thin lipoid layer constituting a potential energy barrier. The height
of this barrier probably varies considerably, and sometimes the membrane
may even act as a molecular sieve. However, assuming a typical value for
the temperature coefficient of permeability it is possible to make an estimate
of the * thermomolecular pressure effect', as AP/AT in the steady state is
called. For plant cells a fairly typical value for the QIQ for water would seem
to be about 2-6 (Davson & Danielli, 1943), and using equation (13) this gives
the 'heat of transfer' as
_ 1-986x2932 , ,.
Q*= ^ IQ ^ In (1-034x2-6)
= 16,900 calories per mole.
Substituting this value in equation (9) the pressure effect is seen to be
AP 16,900x41-3
~"~
= — 132 atmospheres per degree C.,
the negative sign indicating that the pressure develops on the low-tempera-
ture side. This is an astonishingly large value; it means that a temperature
difference of only rj^° C. can cause movement of water at the same rate as
a pressure difference of well over an atmosphere. Before passing to the
question, however, of whether a thermomolecular mechanism can be of any
practical importance in the life of the organism, it is necessary to consider
two objections which raise rather serious difficulties.
Temperature gradients in the cell
The first difficulty arises from the extreme thinness of the plasma
membrane. There is a considerable amount of evidence (see, for instance,
Davson & Danielli, 1943) that the plasma membrane is normally of the
order of icr6 cm. in thickness. Now a temperature difference of yJo°C.
over this distance means a temperature gradient of io~2/io~6 = 10,000 ° C./
cm., which is hardly of the order of magnitude to be expected in living cells ;
about one-thousandth of this value would seem to be more within the
realm of possibility. However, this objection can be partially met in three
ways. The thermal conductivity of a typical oil such as olive oil is about
one-third of that of water. Further, when molecules are definitely oriented,
86 THE ACTIVE TRANSPORT OF WATER
as in crystals, the conductivity may be widely different in different directions,
a factor of even 5 or 6 being sometimes reached. Now the lipoid molecules
in the membrane are arranged more or less parallel, and it is not inconceiv-
able that this arrangement might considerably influence the heat conducti-
vity across the layer. If this is so the conductivity of the membrane might be
less than that of water by a factor as low as one-tenth or even less. This,
however, only goes a small way towards removing the objection. A second
possibility lies in the fact that over a potential energy barrier the flow of heat
is impeded as well as the flow of matter, and in a diffuse gaseous system the
resistance to heat flow can easily be augmented by such enormous factors
as io10 or more. This arises from the fact that in diffuse gases heat flow is
by diffusion of molecules, whereas in solids collision is the mechanism and
no diffusion at all need take place. Liquids come in between, and heat
conduction is partly by one means and partly by the other. To what extent
therefore thermal flow is hindered by the fact that the plasma membrane
constitutes an energy barrier is not apparent, but at least it is a factor to be
taken into account, and it may possibly be very important. Where a whole
tissue or several layers of cells is concerned of course the necessary tempera-
ture differential will be divided between all the plasma membranes, and this
will naturally reduce the gradients over them; the same is true where the cell
seems to possess multiple membranes, as reported recently by Sjostrand
(1953) for mouse pancreas and kidney cells.
Finally, it may be observed that where the temperature gradient is
excessively high the ordinary law of thermal conduction will break down,
just as the analogous laws for diffusion and electrical conduction do.
It is suggested that these three considerations, taken together, may
conceivably raise the resistance to heat flow across the plasma membrane
to several hundred, and perhaps even a thousand, times its value for a
comparable thickness of bulk aqueous phase, and this will of course
correspondingly sharpen the temperature differential over it.
The heat flow accompanying transport
The second objection is an equally serious one. It will readily be
appreciated that while transport is actually proceeding (as opposed to the
steady state in which it has ceased) very large quantities of heat energy
will be passing across the barrier, since all the molecules crossing will
be high-velocity ones. As a matter of fact, to a first approximation, the
amount of heat flowing across per unit amount of water will be the heat of
transfer, Q*. At this rate a man would have to consume a remarkable weight
of potatoes merely to keep his kidneys functioning ! In the course of an hour
between 3 and 4 1. of water are actively reabsorbed by the kidneys, and if the
UNDER TEMPERATURE GRADIENTS 87
heat of transfer was only 6900 cal./rnole (corresponding to a QIQ of 1-5) the
heat flow would be 1,340,000 cal., requiring the complete oxidation of no
less than 360 g. of glucose per hour, or 19 Ib. per day ! Fortunately, the body
functions more efficiently than this, but it looks on the surface as if any
thermomolecular mechanism is ruled out at once. The difficulty arises from
the fact that this mechanism implies that the cell membrane functions as
a heat engine, and as one working over the extremely small temperature
range AT". This at once limits the thermodynamic efficiency of the process,
as Carnot showed, to the excessively small value AT/ T, and consequently
any direct provision of heat by chemical reactions is inconceivable as a
significant contribution. However, an interesting possibility remains. If
a heat engine is very inefficient because it works over a small temperature
range, it is correspondingly highly efficient when it works in reverse as a
heat pump over the same range. In other words, the amount of heat it
transfers can be an enormous multiple of the free energy it consumes. Such
a reversed mechanism can be imagined in the cell or organ. A very simple
case is ordinary osmosis. When a Paramoecium, for instance, draws in water
osmotically from its environment this entering water will carry a large
amount of heat in with it, an amount measured by Q* for the plasma mem-
brane. There is energetically no reason why this heat should not be sufficient
to expel water by a thermomolecular process into the contractile vacuole,
whose membrane, conceivably, has a lower value of Q*. Ultimately, of
course, some source of free energy is required, and this may operate by
maintaining the required osmotic pressure difference at the points of water
entry and exit. In the case of the kidney the possibilities are somewhat
similar. Pressure filtration in the Bowman's capsule will hardly imply any
considerable flow of heat, as the membrane almost certainly acts as a fairly
coarse sieve; but in the proximal convoluted tubules where apparently
glucose is reabsorbed it is not at all impossible to conceive that the develop-
ment of osmotic activity might draw in water from the capillaries across
a high energy barrier, the heat provided being later used, perhaps in the loop
of Henle, to sustain a reversed flow into the capillaries through an energy
barrier substantially lower. Such a process, involving first a heat pump and
then a heat-engine activity, might possess quite an appreciable thermo-
dynamic efficiency. At least it appears to be a possibility deserving of
attention.
VI. CONCLUSION
The theory outlined in this paper has probably its more important
potential application to water movement, though it can also be applied
to the active transport of solutes such as sugars and ions. However, in
the case of these there seem several much more promising suggestions
88 THE ACTIVE TRANSPORT OF WATER
now being debated, and only in the case of water does it seem likely that
none of the existing mechanisms is adequate to meet the requirements.
What, in effect, the present theory does is to show that under certain
conditions a temperature differential can bring about results hitherto
associated only with a pressure difference. Thus the cells of a green alga
spending its whole life in fresh water maintain themselves in equilibrium so
far as water content is concerned only because they possess a firm wall
which is capable of sustaining a considerable internal pressure. This
possibility is normally denied to the single-celled organisms of the animal
kingdom, and they are consequently faced with the problem of actively
excreting water to counterbalance that entering by osmosis. It might, of
course, prove to be the case that they excrete actively not water, but ions or
other solutes, the water being merely drawn out after them passively by
ordinary osmosis. This can only be decided by analysis of the fluid excreted,
and it certainly remains an attractive theory, especially in the light of the
suggestions put forward a year or two ago by Goldacre (1952) on protein
contractility as a basis for osmotic work. However, living systems are
amazingly complex, and it is hardly a flight of fancy to suppose that the
ultimate explanation of the phenomena of active water movement will be
found not in one, but in a combination of several physical processes.
APPENDIX
On the thermodynamic theory
The following is a brief account of the derivation of the fundamental
equation. It follows the treatment given by de Groot (1951) and illustrates
the method of the Thermodynamics of Irreversible Processes.
Consider the simple system discussed earlier (Fig. 3), water being the
only component present and all intensive properties, such as temperature
and pressure, being uniform throughout each sub-section. Let super-
scripts I, II denote the two sections. Further, let subscripts /, e denote
increments of an extensive quantity gained internally (i.e. from the other
subsection) or externally (from the surroundings). Then it is required to
find an expression giving the rate of increase (cr) of the entropy of the
system supposing it to be held completely isolated from its surroundings.
To do this it is convenient to consider the system as merely closed, but not
isolated] and to separate the expression for its rate of entropy increase into
two parts, one representing the entropy gained from the surroundings, the
other the entropy produced internally. The latter will be the value of cr,
the quantity required. The further development of the theory will then
require the use of Onsager's Theorem.
UNDER TEMPERATURE GRADIENTS 89
If in1, m11 are the masses of water on the two sides of the membrane,
the Law of Conservation of Mass gives
mi 4. mii = constant,
or dml + dmll — o. (i)
Further, if U is the internal energy we have the analogous relation
diW + diUn^o. (2)
Finally, if the whole system is changing slowly enough each of the sides
can be regarded as an open system in internal equilibrium, and we can
apply to it the equation of Gibbs :
dU= TdS-PdV+pdm. (3)
The validity of this procedure will naturally depend on the transport
processes in operation not being of too rapid a nature; in practical cases,
however, equation (3) will hold to a very good degree of approximation.
Applying it to each side in turn we get
(4)
(5)
Rearranging these and adding,
T11}, (6)
where dU has been split up into dtU-{-dcU. The quantity in square brackets
represents the entropy gained from the surroundingsf (dPS) ; that in braces,
the entropy (dLS) generated within the system. Had the system been
isolated dcS would of course have been zero, while d^S would have been
unaffected. It is this latter quantity therefore which is relevant to the
Onsager theory.
Eliminating dm11 and dtUu from (6) by means of (i) and (2) and
dividing by dt, we get for the rate of entropy production
(T =
dt dt \T* TV) dt \T*
dm*
where the symbol AT1, for instance, stands for T"2— T\, and TlxT2=T.
Equation (7) has the form mentioned earlier ; the ' fluxes ' are Ju — —dt Ul/dt,
t The expression (dfW + ptdV1), for instance, is equal to qel, the heat absorbed from
the surroundings by subsection I. This follows from the first law, P^dV1 being the work
done by subsection I in expanding against the surroundings.
QO THE ACTIVE TRANSPORT OF WATER
the flow of energy from I to II, and Jw = —dmlldt, the flow of water. The
appropriate 'forces' are XV=—&T/T2 for the flow of energy, and
Xu = - A(^/r) for the flow of water.
The theory now relates the /'s and X's by means of rather general
equations analogous to 'Ohm's Law*. This leads to the expressions
Ju — Luu Xu -f Luw Xw, (8)
J\v = LwuXu + LWWXW, (9)
where the L's are conductance coefficients which will depend on the size
and geometry of the system and on the nature of the divisional membrane.
However, Onsager's theorem states that provided forces and fluxes are
measured in the way described the general result always holds that
LUW = LUU. (10)
This is a result of the Principle of Microscopic Reversibility, according to
which reversal of the motions of all the particles in an isolated system
would simply cause the system to retrace its former history.
Before proceeding with the theory it is convenient to replace the flow
of energy Ju with the flow of heat, JQ. In a rough way it can be seen that
the energy carried from I to II per unit mass of water will be made up of,
first, a term depending on the internal energy u per unit mass of water;
secondly, an amount of work represented by the product of the volume of
water per unit mass (v) and the pressure under which it flows; and thirdly,
a quantity of heat. The first two terms make up the heat function per unit
mass (A), since h = u + Pv. (u)
This leads to the suggestion that we define the heat flow as
JQ=JU-MW (12)
The quantity h will differ for the two sides of the system, but since the
difference is small this will be of no account. It simply underlines the fact
that the theory in any case is strictly exact only for infinitesimal departures
from equilibrium.
If we introduce JQ from (12) into (7) we find that the force Xw is
altered. This illustrates the ambiguity mentioned earlier when we were
discussing the distinction between active and passive movements; the
force Xw which is the passive agent for water movement depends on
whether we take Jn or JQ as the other flux. However, it seems logical to
take JQ instead of 7U, since not only is the flow of heat independent of the
flow of matter in a sense in which the flow Ju is not, but the choice of JQ
leads to values of the forces which are uniquely defined ; that is, they have
no unknown additive constant as h or/*, for instance, have.
UNDER TEMPERATURE GRADIENTS
Substituting for Ju from (12) into (7) the new forces become
The latter result follows from the fact that fi = h— Ts, where s is the
entropy per unit mass; also, since we are dealing with a single substance jti
is a function of P and T only. Thus, considering a fixed amount of water
we have the Gibbs equation
dG^VdP-SdT, (15)
which, on dividing by the total mass and writing A instead of d> gives
A/* = ^AP-sA7\ (16)
The 'Ohm's Law* equations can now be written in the form
AT\
AT
with LwQ = LQw. (19)
The condition for the steady state is found by writing Jw — o. This gives,
on rearrangement, ^p ^ T
Ar=-;£>r (20)
But by definition, if AT1 is put equal to zero the heat flow per unit of mass
(JQ/JU^) is the heat of transfer Q*. This gives
2*=r- <2')
•^IDW
Introducing (21) with (19) into (20) it follows that
AF_ Q*
AT" ~vT (22)
The temperature dependence of the permeability
Let y1, v11 be the rates at which molecules of water are passing across
the membrane from the two sides respectively. At equilibrium vl and yn
will be equal, and this will also be true for the steady state.
92 THE ACTIVE TRANSPORT OF WATER
Starting with the whole system in equilibrium, imagine the temperature
and pressure of section I to be raised by AT1 and AP respectively. Then the
condition that must be fulfilled if the net flow of water is to remain zero is
o. (23)
Treating v as a function of T and P this becomes
AP=O) (24)
.... AP fa \rdv f .
or, more explicitly, ^= ~dfi)P ^
_ a (In v) lit (In v)
~ 'of I i?P ' (2 >
Now it can be shown that v is proportional to the vapour pressure p.
Hence S(lnv) d(\np) v
.. v L — ___ * ___ --'- — _____ l'?'7l
?P ~ rcP ~RT (27)
by a well-known thermodynamic relation, v being the volume of unit
mass of water.
But from (26) we have
_
fi'T AT dP
or introducing (22) and (27),
fl(lni') 0* v
cT ~vTRT
( }
which is the familiar Arrhenius type of equation for the temperature
variation of a rate.
The ordinary permeability // is the net rate of flow of water under unit
pressure difference and zero temperature difference, i.e.
Using (27) it is possible to write
vv
pRT
or v = ---
UNDER TEMPERATURE GRADIENTS 93
Thus from (29) d
or since v will be nearly constant,
which is the result sought.
REFERENCES
DAVSON, H. £ DANIELLI, J. F. (1943). The Permeability of Natural Membranes.
Cambridge.
DE GROOT, S. R. (1951). The Thermodynamics of Irreversible Processes. Amsterdam.
GOLDACRE, R. J. (1952). The folding and unfolding of Protein molecules as a basis
of osmotic work. Int. Rev. Cytol. i, 135-64.
SJOSTRAND, F. S. (1953). Electron microscopy of mitochondria and cytoplasmic
double membranes. Nature, Lond.y 171, 30-2.
SPANNER, D. C. (1952). The suction potential of plant cells and some related
topics. Ann. Bot., Lond., N.S., 16, 379-407.
SPANNER, D. C. (1953). On 'active' mechanisms in biochemical processes. Physiol.
plantarum (in the Press).
WATER TRANSPORT IN INSECTS
BY J. W. L. BEAMENT
Agricultural Research Council Unit of Insect Physiology,
Department of Zoology, Cambridge
I. INTRODUCTION
Among the factors contributing to the success of the Insecta — animals of
a comparatively small order of size — the evolution of a waterproof cuticle
has been considered of prime importance. Consequently, water exchange
between the atmosphere and an insect or its egg has received considerable
attention. Other groups of animals have become established on land, either
by living in the nearly saturated atmosphere of soil or in other specialized
regions of high humidity. The higher vertebrates do not hold to this rule,
but at once the order of size becomes apparent. A medium-sized mammal
has a surface-area/volume ratio of the order 0-5 sq.cm./c.c. It can survive
comparatively high evaporation rates before it need replenish its water
supplies. A typical insect egg has by comparison a ratio of 50 sq.cm./c.c.,
and normally it has no means of replenishing its water from the environ-
ment. The active insect is only a factor of ten better off than its egg, and
a waterproof cuticle is clearly a necessity.
Wax layers
The principal device which has been evolved by the insect to resist
water loss is now well established ; it is a layer of orientated lipoid near the
surface of its cuticle, or forming one of the layers of its egg-shell (Ramsay,
1935; Wigglesworth, 1945; Beament, 1945, 1947; etc.). The degree of
efficiency which can be achieved by wax systems is exemplified by the egg
of the mite Metatetranychus ulmi (Beament, 1951). The over-wintering egg
has a diameter of only 0-14 mm., and its surface-area/volume ratio is
500 sq.cm./c.c. It has hatched successfully after being held at room
temperature in an atmosphere of 70% R.H. for a year. Other than its two
wax layers (one of which does not cover the whole shell surface) this egg
has only a keratin-like membrane and a sticky cement to protect it, the
layers together being 4/4 thick; it can survive a loss of only 5 % of its total
water content and has no means, so far as can be demonstrated, of taking
up water, even from liquid water in contact with it. If the inner wax layer
is incomplete, it dries up in 70% R.H. in a matter of minutes.
WATER TRANSPORT IN INSECTS 95
The problem of conserving water by * passive' means has thus been
overcome, and very efficiently; outside the insects proper, the wax-layer
mechanism is found in ticks (Lees, 1946; Lees & Beament, 1948), mites
(Beament, 1953), and probably in spiders (Edney, 1953), though not
apparently in myriapods (Cloudsley-Thompson, 1950); there is evidence
for believing that the cocoon of the lung fish, Protopterus (Beament, 1953),
uses a similar mechanism. Undoubtedly a parallel system occurs in the
leaf waxes of plants (Stelwaag, 1924; Fogg, 1948; Piper, Chibnall &
Williams, 1934).
At first sight, the insect has created two problems for itself by producing
this impermeable cuticle, for the very existence of the wax layer may well
prevent the absorption of water in the environmental atmosphere, or of
liquid water in contact, when such an opportunity presents itself. We must
neglect here the osmotic problems of a fresh-water insect with such
a cuticle ; at the present time no one has demonstrated the existence of wax
layers in aquatic insects. Evidence from work on water exchange through
the anal papillae of mosquitoes (Wigglesworth, 1933 a\ Beadle, 1939)
suggests that such surfaces are much more permeable than we would
expect if a wax layer were present. But one can hardly envisage a humidity
receptor, or a chemo-sensory organ, on a terrestrial insect, which could
function efficiently if covered with wax. It seems very probable that
certain surface areas of insects will be left unwaxed to allow of their
specialized functions. A general consideration of the physiology of aquatic
insects, however, must make it difficult to believe that the whole of the
cuticle is freely permeable to water, and a wax layer would surely be the
most obvious means of protection against invasion by water.
Water uptake
Because of claims for such a degree of water impermeability, a degree
of impermeability which has been repeatedly demonstrated in experiments
on isolated cuticle and artificial membranes (Beament, 1945, etc.) to be
inherent in the non-living system of the cuticle or shell, it is the more
remarkable that certain insects and eggs can take up water from the outside
atmosphere, from humidities well below those which would be in equili-
brium with the blood fluids. Lees (1946, 1948) shows that the tick, Ixodes,
can take up water down to 88% R.H., Mellanby (1932) that the mealworm
can do so from 90% R.H. and Edney (1945) claims that the flea pre-pupa
may do so from humidities as low as 50%. In all these examples, the
blood-equilibrium humidity is of the order of 99% R.H. Whether this is
a relevant figure, and how it compares with the humidity equivalent of the
epidermal cells themselves, must be considered later.
96 WATER TRANSPORT IN INSECTS
Before discussing the nature of any active water-uptake mechanism, the
physico-chemical properties of these natural membranes should be fully
understood. We must also distinguish carefully between * active' and
1 controlled* water exchange. In many insect eggs — for example, those of
Melanoplus (Slifer, 1938), Locustana (Matthee, 1951) — the shell is im-
pervious to desiccation for considerable periods of time; when eggs in
diapause are placed in liquid water, they do not take up appreciable
amounts. Once diapause is broken, however, the egg responds rapidly to
the existence of liquid on its surface, and absorbs it rapidly, increasing in
weight by as much as 15%. But Matthee has shown that water uptake
depends on the availability of oxygen, so that the process must be regarded
as ' active' at least in the sense that respiring cells are necessary. It must be
emphasized that these eggs are very waterproof up to the time of immersion
in water, even though they have come out of diapause; it is even more
striking that they apparently cannot take up water from saturated air, though
they will do so from aqueous solutions having high osmotic pressures.
Once these eggs start to take up water, the rate of transfusion is so much
greater than the apparent permeability of the shell-wax layer when the egg
is desiccated in diapause that one might be led to postulate a process in
which the cells had very remarkable powers of absorption. But if the
water-uptake process is interrupted, desiccation experiments at this time
indicate a real, though not necessarily permanent, change in the physical
permeability in both directions. This evidence could only be challenged
by experiments on inert shell membranes isolated at the relevant times.
Hence, here, an undoubtedly 'active' uptake of water is at least accom-
panied by suspected changes in the inert layers. Now the process of water
uptake might be considered as partially osmotic, though this would
continue in the absence of oxygen. But apart from the demonstrated ability
of eggs to obtain water from comparatively strong solutions, osmosis
would rapidly lead to bursting, and this is a very rare event in nature in
Locustana. The living material must be just as capable of arresting the
inflow of water, regardless of existing osmotic gradients, as it is of initiating
and maintaining the flow, and we can only conclude that an active non-
osmotic process is present. The apparent partial destruction of a wax layer
is an entirely different problem ; towards explaining this, many suggestions
could be made, but when it is followed by repair, perhaps in the presence
of water flowing inwards, we are in greater difficulties. Again, we must
know the details of the inert changes before we can assess the necessary
'vital* activities which could control the water exchange.
Water exchange in the egg of the garden chafer, Phyllopertha horticola
(Laughlin, 1953; and unpublished observations which he has kindly
WATER TRANSPORT IN INSECTS 97
permitted me to report), follows a different pattern. This egg does not go
into diapause, but in the middle of a comparatively short period of embryonic
development there occurs a space of 5 days during which the egg takes up
a considerable amount of water. The uptake is accompanied by an increased
rate of loss in dry air, ceasing at the end of the water-absorbing phase.
Throughout the life of the egg, the desiccation rate in dry air is so much
greater than that of typical insect cuticle that one would hesitate to
attribute the impermeability of the shell to a wax layer, and no such layer
has been demonstrated. To what extent the uptake phase represents an
* active' process is not known, but a contributory factor could be a ' control'
of permeability in the form of reversible changes in the shell material. The
increased desiccation rate during the uptake phase might yet be greater,
but for 'active' secretion on the part of the living material, resisting water
loss by desiccation. Suggestions have been made (Edney, 1953, etc.) that
such a process could account for the difference in water loss between dead
and living animals, and is discussed later. Here, it is much more important
to realize that these eggs can successfully complete their whole embryonic
period when floating in distilled water. One cannot neglect osmosis when
discussing water-uptake mechanisms, and one cannot easily foresee,
without the existence of a wax layer, a mechanism by which this living
material could tolerate an environment of distilled water in the early and
late periods of its existence.
When we compare this egg with that of the cricket, Gryllulus (Browning,
1953; and unpublished observations which he has kindly permitted me to
report), there is an outstanding difference. The life history of the egg shows
a middle water-uptake phase, but the apparent desiccation rate during this
period does not change materially. This might suggest, even more strongly,
the idea of an * active' mechanism which could both absorb water and
oppose loss during desiccation, though it is somewhat remarkable that the
combined increase in permeability and active process should give similar
water loss as occurs before and after the active phase. It is perhaps more
difficult to appreciate what is going on in these eggs, for while, in the
chafer, the shell shows every sign of being elastically expanded during
water uptake, there is no indication of high internal hydrostatic pressures
in the cricket egg at any time in its life — though it will develop and hatch
entirely submerged in distilled water and has never been observed to burst
due to osmosis. One can hardly suppose that the shell is completely
permeable to ions, so that, again, the living egg can in some way negate
osmotic forces.
98 WATER TRANSPORT IN INSECTS
Respiratory envelopes
Wigglesworth & Beament (1950) have shown that a large number of
insect eggs have a complete air-sponge layer round the yolk, incorporated
into the shell and allowing respiratory exchange over a large area. While
this layer does not prevent the existence of continuous solid material from
the outside to the inside of the shell, the water-path at this layer is
undoubtedly considerably restricted, and largely replaced by a probably
slower system of gas diffusion across the sponge.
Water uptake in animals during post-embryonic life
The intimacy of the epidermal cells with the cuticle, and the fact that
these cells are directly bathed in blood fluids, makes it difficult to believe
that there could be a permanent large discrepancy between the osmotic
factors of the two tissues, unless a very large amount of energy is constantly
expended to achieve this. Yet the blood is in equilibrium with 99% R.H.,
whereas in Tenebrio (Melanby, 1932), Xenopsylla (Edney, 1945) and Ixodes
(Lees, 1946, 1948) the cuticle can obtain water from much lower humidities.
There are a number of scattered phenomena which must be considered.
We are not sure that the water uptake is entirely through the cuticle, i.e.
through the externally exposed integument. It may, to an important
extent, involve the tracheal system, at the inner end of which another form
of water exchange takes place (Wigglesworth, 1933, etc.), and where, under
certain physiological conditions, water may be withdrawn and replaced in
the tracheole capillaries. But the tracheal system, or at least the inner part
of it not subjected to mechanical aeration, is understood to be filled with
water-saturated air, and there must be a very water-permeable membrane
amounting to a free water surface, present in the tracheal system to achieve
this. The main tracheal system is derived from intuckings of the integu-
ment, and although it is morphologically very different from the cuticle,
there is reason to suppose that it is secreted as the same fundamental
procession of chemical entities as compose the multiple laminae of the
cuticle : that it has a waxy waterproofing layer on it. A small but interesting
observation, which reinforces this idea, can be made when flooding the
tracheal system with aerated water. It is a well-known physico-chemical
demonstration that air bubbles appear from aerated water when placed in
a waxed beaker, but not against chemically clean glass. The appearance of
air under these circumstances in the tracheal trunks is dramatic, and more
rapid than against the outer cuticular surface. This would seem to mean
that in the tracheal lining there is a more hydrofuge surface than on the
cuticle, and, possibly, that the cement which overlies the wax layer of
WATER TRANSPORT IN INSECTS 99
typical cuticle (Wigglesworth, 1947; Way, 1950; etc.) is not present in the
tracheae.
Hence, so far as water exchange is concerned, we must regard the main
tracheal surface as being similar to cuticle, until evidence to the contrary is
presented; the tracheole ending, where fluid is visibly absorbed and
resecreted, may be a different type of active exchange with the outside
world. But the tracheal system is traditionally the main source of water
loss from the insect. Clearly the spiracular closing mechanism has been
evolved to reduce water loss to a minimum; it has been shown to act
(Wigglesworth, 1935) in response to the oxygen requirement and carbon
dioxide accumulation of the animal, and not to the saturation deficiency to
which the animal is exposed. It would thus seem most unlikely that the
tracheole — a source of great water loss — would be the main site of active
water uptake, especially when active uptake is apparently so rare a pheno-
menon, in comparison with the widespread distribution of fluid movement
in tracheoles. We must therefore look to the main cuticular and tracheal
surfaces for the site of active water exchange.
Water uptake and cuticle damage
Lees (1947) has shown that the ixodid tick is made incapable of taking
up water from surroundings at humidities lower than the equilibrium value
of its blood fluids, if the epidermis is 'wounded'. Even minute abrasion of
the epicuticle, probably only affecting the cuticular wax, the underlying
tanned epicuticle and the tips of the pore canals, is sufficient to prevent
active uptake immediately.
Now abrasion of this kind is known to elicit a typical wound-healing
activity of the underlying epidermal cells (see Wigglesworth, 1940), so
that the physiological activity of the epidermis can be said to have been
disturbed. Wigglesworth has repeatedly pointed out that the epidermis of
an insect is an entity, so that such disturbance of one part of its components
might be taken to imply that all of it will behave abnormally. However,
the tick recovers its ability to take up water against an apparent gradient of
humidity, and this recovery is accompanied by the repair of the cuticular
wax layer. Lees shows clearly that the recommencement of the water-
uptake process precedes the complete repair of the wax, but it is quite
possible that at least a monolayer of wax has been laid down over the whole
of the denuded area before uptake recommences. That water uptake can in
the later stages accompany repair would would seem to indicate that it is
the disruption of the complete wax layer, and not the repair activity, which
prevents the uptake process.
7-2
100 WATER TRANSPORT IN INSECTS
Water loss from dead insects
Parallel with these phenomena are the reports that there may be differ-
ences between the rate of water loss of some dead and living individuals
of the same species. Wigglesworth, in his extensive examination of the
cuticle in 1945, states that his figures for water loss are those for dead
insects, but providing that the spiracles are blocked, there is no difference
between water loss from dead and living animals. He includes the meal-
worm amongst his experimental material; Lees, on the other hand,
indicates the reverse phenomenon in his ticks; individuals desiccated at
a time when they could secrete water against a low humidity, lose consider-
ably less water compared with dead individuals. The desiccation rate of the
sensescent adult approaches that of the dead animal, and the aged tick loses
its ability to take up water from near saturated atmospheres. More recently,
Edney (1953), working with spiders (in which he proposes the presence of
a typical wax layer), demonstrated not only a slower rate of water loss in
the living, as opposed to dead animal, but as between cyanide-killed
spiders and those very recently killed by temperature in the course of his
temperature/evaporation experiments. He suggests that when a spider is
'dead' as judged by mechanical response, its epidermis is still alive and
may oppose evaporation. Edney does not demonstrate any active water-
uptake mechanism in these spiders, but his proposal must be considered
along with those previously raised by studying ticks and eggs — can the cell
resist desiccation? It does not necessarily follow that a process capable of
secreting water inwards by doing work will automatically also decrease the
rate of flow outwards when the direction of flow through the membrane is
reversed ; the two systems are not without their physico-chemical differences.
The mention of reversal of flow through cuticular membranes at once
introduces the much-discussed phenomenon, usually called the ' asymmetry '
of insect cuticle. The confusion which exists over this has resulted in some
curious ideas on water exchange through the cuticle, and a real under-
standing of this asymmetry is all part of the original premise: that one must
know the physico-chemical properties of the membrane before one can
assess or interpret the mechanism of the active processes which occur.
II. MEMBRANE PHENOMENA
Water uptake through a cuticle was elegantly demonstrated by Ramsay
(1935) when he showed that a droplet of water, placed on the cuticle of
a cockroach, did not evaporate, but was covered by a layer of grease,
present in a mobile state on the surface of the animal. While the grease film
substantially reduced evaporation from the droplet, the cockroach took in
WATER TRANSPORT IN INSECTS IOI
the water from the interface between it and the underlying cuticle. This
cuticle is atypical, for few insects are capable of so isolating a droplet on
their surface while they absorb it at leisure.
Ramsay apparently assumed that the cuticle underneath the drop was
denuded of its grease, but this is not so; if a cuticle is washed for a long
time in the surface of running water so that, by surface spreading, material
is swept away, most of the cuticular grease is removed (see Rideal, 1926).
But the lowermost monolayer, the one which we believe to be most orga-
nized and orientated, and which may be strongly linked to the polyphenol
tanned layer of the epicuticle, is not so removed. Indeed, while all the
evidence (Langmuir, 1925, etc., on films spread on troughs; Beament,
1945, etc., on cuticle models) indicates that this monolayer is principally
responsible for the impermeability of isolated cuticle, experiments on the
permeability of isolated cuticles make it doubtful if cold chloroform or
similar solvents can remove such monolayers, though (see p. 101) wax
solvents certainly have a considerable effect on their permeability.* Thus
the cockroach is still faced with the problem of taking up the water droplet
through the most impermeable layer of its cuticular lipoid.
Asymmetry of membranes
As long ago as 1845, Matteucci & Cima demonstrated that the rate of
flow of water through the skin of the frog and of the eel differed with the
direction of flow, when the external conditions were reversed. The pheno-
menon has been shown with the seed coats of plants (Denny, 1917), and
Hamburger's (1908) results with synthetic membranes of collodion and
chromgelatin first removed the suspicion that the phenomenon was due to
a special property of the products of living material. In arthropod cuticle,
Hurst (1941) reported of the blowfly Calliphora that ' water evaporates
through the cuticle ... of the larva more than one hundred times as rapidly
in the direction lipoid to chitin than in the reverse direction'. In his 1948
paper, Hur/st gives a graph showing apparent ratios of asymmetry of the
more reasonable order of ten to one. He, nevertheless, claims that the
order is such as to suggest an ' all-or-nothing* phenomenon, and outlines
a complicated theory of a porous valve structure in the insect cuticle. It
must be emphasized that Hurst's method of measurement of permeability
was with an osmometer tube, to which the cuticle was attached by a rubber
band; the insects were dissected under water (we are not told if they were
previously killed, so that epidermal cells may have been living). The
cuticular material was completely saturated with water at the start of the
* I am indebted to Mr M. Holdgate, for permission to refer to some experiments on the
contact angle of the cockroach cuticle which support this view.
102 WATER TRANSPORT IN INSECTS
experiment and the results on which these claims are based were apparently
for the first hour of water exchange when the apparatus was placed in a
controlled humidity.
Apart from the very great difficulty of obtaining any sort of waterproof
seal between cuticle and glass, it is the writer's experience that it is
necessary to expose any sort of permeability measuring device to a particular
humidity gradient for at least 48 hr. before a stable value is reached.
Enormous values of apparent permeability can be obtained when a wet
endocuticle, with or without attached cellular debris, is exposed to a dry
atmosphere. And since Wigglesworth (1945) and others have demonstrated
the dramatic change in permeability produced when minute cuticular
abrasion disrupts the wax layer, the only true demonstration of asymmetry
is one in which the same piece of cuticle in the identical state of preserva-
tion is successively exposed to the same humidity gradient.
Readings of permeability are thus only valid in dynamic equilibrium,
and it is obviously desirable to reverse the membrane (or gradient) a
number of times to be sure that the material has not suffered damage. It
is further necessary to run two controls; a free water surface, checking
both saturation deficiency and temperature fluctuation, and a blank
apparatus to show surface condensation and loss during weighing opera-
tions, etc. The water loss of an intact piece of cuticle is of the order of
10 mg. or less in 24 hr., and water exchange over the remainder of the
apparatus, or fluctuations in saturation deficiency may easily be of the same
order, sufficient to mask asymmetry or to double the apparent effect. It is
therefore obviously worth while to describe in some detail an apparatus
used to obtain the results quoted below.
The design of an experiment on permeability
Apparatus has been evolved from the simple membrane holder described
by Beament (1945). It consists (Fig. i) of a Pyrex water tube, having
a projecting flange at one end, and reduced to a narrow bore at the other.
A plate of electron metal is placed over the tube, with soft rubber seating
on to the glass, and an identical plate is attached below by four brass
screws, which also form supports for the whole device. The cuticle is
clamped between two electron metal rings, chosen to suit the size of the
experimental material, having pressure projections turned on both faces,
and these are themselves clamped in a brass cell with screws to provide
overall compression. All metal parts are interleaved with washers of
reinforced rubber. The unit containing the cuticle sample is thus a robust
structure, which can be mounted either way up between the plates of the
water tube ; the cuticle can be treated in various ways in its holder, while
WATER TRANSPORT IN INSECTS
103
care is taken that it undergoes no mechanical damage at all. The upper end
of the water tube is closed by a polythene tube, carrying a short length of
fine glass capillary, to allow pressure equilibration between inside and
outside, but the minimum of diffusion. The whole construction is designed
to have minimum weight (well under 100 g.) so that it can be weighed on
a chemical balance.
Sets of cuticles, cast skins, or artificial membranes were mounted in
brass cells and stored against use in a desiccator over phosphorus pent-
oxide. For each experimental determination, six membranes of the type
Fig. i. Sectional drawing of membrane holder for measuring permeabilities.
of material being investigated were placed in the holders, three with wax
outwards and three reversed. A further holder, with a cell containing a
metal disk an eighth of an inch thick in place of the membrane, formed the
control, while free evaporation was recorded by a similar apparatus con-
taining a disk of porous pot, with its circumference sealed with thick
beeswax. Water was added to a standard height in each tube, capillaries
placed over the ends, and the set of holders placed over selected humidity
solutions. They were weighed at intervals of 24 hr. until agreement of two
successive readings of water loss showed that stable conditions had been
reached. The sets of membranes were then reversed, the operation
repeated, and in some cases many reversals were carried out. It was hoped
104 WATER TRANSPORT IN INSECTS
by this scheme to eliminate sources giving rise to the appearance of
asymmetry. Results are given in Table i .
Table i. Abstract from determinations of membrane asymmetry
Rates expressed in mg./sq.cm./hr. ; inner surface against pure water; outer, 2 cm. from
phosphorus pentoxide surface in still air; temp. 20° C. ; evaporation reference, 24 mg./
sq.cm./hr. from free-water surface in membrane position; limit of experiment: asymmetry
less than 1:1*09 may not be significant.
Membrane
Treatment
Direction of flow
Rate
Rhodnius, 5th
Endo-epicuticle
0-81
whole cuticle
—
Epi-endocuticle
0-40
Rhodnius, sth
—
Initial exo-epi-
0-17
exuvia
cuticle
Average epi-exo-
3-29
cuticle
4th reversal
Exo-epicuticle
0'57
Boil in chloro-
Exo-epicuticle
form
Epi-exocuticle
i9-£
Periplaneta, late
—
Exo-epicuticle
0-42
nymphal exuvia
—
Epi-exocuticle
0-96
1 Wash in water
Exo-epicuticle
1-73
surface, mono-
Epi-exocuticle
2-06
i layer of lipoid
Boil in chloroform
Exo-epicuticle
20-4
Epi-exocuticle
20-1
Beeswax on
—
Parchment-wax
^•3
parchment
Wax-parchment
20-8
Beeswax on
—
Wing-wax
4-2
wax-free ci-
Wax-wing
cada wing
i
Rhodnius wax on
Heat to 60 C. dry
Wing-wax
0-80
wax-free cicada
before measuring
Wax-wing
3-10
wing at 20° C.
Beeswax on —
Gelatin-wax
1-97
tanned gelatin
Wax-gelatin
8-12
standard thick
membrane
Wax free
15-3
; Asymmetry
1:1-7
1:19
Not
significant
1:2-3
1:1-2
Not
significant
1:4
1:4
Discussion of results
The selected membranes are those of (i) Rhodnius ', an insect whose
cuticle is more fully understood than that of any other (Wigglesworth,
19336, 1945, 1947, etc.), (2) Periplaneta, whose cuticle is the subject of
current research into cuticular grease secretion (Beament, 19516), and
(3) 'artificial' cuticles, both of lipoid-free cicada wing covered on one side
with beeswax, and of tanned gelatin membranes (prepared as described by
Beament, 1945) similarly treated with wax.
All these membranes are asymmetric, though the ratio of permeation in
either direction, when subjected to the extreme gradient of pure water/ <
5% R.H., does not exceed five to one in any true cuticle; the ratios for
artificial systems are lower. There is no suggestion of an 'all-or-nothing'
WATER TRANSPORT IN INSECTS 105
process; rather that the highest ratio is obtained when the membrane in its
' natural ' sense has a value of impermeability of the same order as that of
an intact insect and therefore, presumably, has the most perfectly organized
lipoid layer. Hurst (1948) purports that the asymmetry ratio of blowfly
cuticle is not radically altered by chloroform treatment and concludes that
the phenomenon of asymmetry cannot be anything to do with the orientated
lipoid layer; these results would indicate exactly the reverse. A cockroach
cuticle, repeatedly washed in running water, and (see above) therefore with
only a monolayer of grease on it, still shows asymmetry of two to one in the
best example, and is significantly asymmetric in the worst case.
Comparing the order of asymmetry with the order of permeability in all
results, we might conclude that in comparatively permeable membranes
there could be pinholes. Whole cuticle preparations have often such
perforations where the ducts of dermal glands have been broken off,
whereas cast skins, though more delicate, are more likely to be intact, since
the lining of the gland is shed with the rest of the outer cuticle. Further,
in extreme cases, the swelling of artificial membranes accompanying water
uptake could be held to cause some disruption of the continuity of a super-
ficial wax layer. (This phenomenon obviously does not occur in living
systems.) But, as outlined below, any tendency for disruption of wax
through swelling will produce an effect on permeability tending to reverse
the typical asymmetry, and only enhances the values which have been
obtained here. The results with exuviae of Rhodnius need an extra word of
explanation ; the permeability in the natural sense does not seem to change
during the extensive series of reversals in the same way that values for the
reversed sense do. There can be little doubt from our knowledge of the
cuticular wax of this insect that it remains substantially unchanged and
unaffected by long exposure to water (under its cement layer), and it seems
likely that the drop in permeability in one direction only could be due to
solvation of hydrophilic material from the inner surface of the skin, during
exposure of that side to water.
III. THE PERMEABILITY OF COMPOUND MEMBRANES
The phenomenon of asymmetry in itself is probably of no great consequence
to the insect, for, so far as can be seen, no insect is capable of reversing its
cuticle at will, and therefore of making use of the phenomenon in a
particular circumstance. But an understanding of the causes of asymmetry
may give us a better understanding of the whole mechanism of transport
through cuticle, considering only the non-living system, and may therefore
give us ideas of the way in which the living material may create circum-
stances in the cuticle, such that water may flow in a particular direction.
106 WATER TRANSPORT IN INSECTS
The tanned gelatin model
Membranes of tanned gelatin (see Beament, 1945) when placed between
water and a dry atmosphere transmit water at 15 mg./sq.cm./hr. When
covered with beeswax, on one side to a thickness of 2/£, their permeabilities
are respectively 2 and 8 mg./sq.cm./hr. in the directions gelatin/wax and
wax/gelatin. (Since the tanning of gelatin is a progressive process which
may take years to complete, we should make it clear that the evidence
quoted here has all been obtained from one batch of material in a sufficiently
short period of time to ensure that the characteristics of the material have
remained constant.) If we consider any membrane of uniform constitution,
transmitting water at a steady rate Xy when placed between two different
spacial concentrations of water molecules, we can discuss the dynamic
state as follows:
(1) Across the interface between membrane and higher water concentra-
tion there will be an overall rate of flow X. This rate will be the resultant
effect of a number of forces, such as result from the hygroscopic property
of the membrane, suction forces, etc., and saturation deficiency, osmosis,
etc., in the reverse direction. For a known rate of flow we can determine
a corresponding concentration of water in the membrane surface.
(2) Similarly, there will be a set of forces at the interface between
membrane and lower water concentration, and for a flow X there will be
a corresponding water concentration in the membrane surface.
(3) At any arbitrary section of the membrane, a consideration of forces
across the section must give rise to a flow X. There must be a concentration
gradient across the membrane which may not be linear (and will not be so
unless the materials obey Pick's law). We may, however, investigate this
gradient empirically.
(i) Water uptake. A very thin membrane (from the same batch of
material as the thick membranes) was used, to ensure minimal difference of
concentration between the surface and throughout its thickness. It was
desiccated for several days, suspended by fine wire from a torsion balance,
and immersed in a beaker of distilled water at 20° C., for periods of
i min., removed, adhering water immediately blotted off with pads of
filter-paper, the balance read, and the operation repeated until the gelatin
ceased to take up any further water. Fig. 2 a shows the rate of uptake across
unit surface area, against corresponding concentrations of water. Uptake
is inversely proportional to concentration; the rate is very high until the
membrane contains some 16% by weight water of the eventual saturation
value. There is a uniform slower rate over the range 16-75%, anc^ a
rate of lowest level until saturation is reached.
WATER TRANSPORT IN INSECTS
107
(2) Water loss by evaporation. The same piece of material, fully saturated
from liquid water, was then hung from the torsion balance (through a small
hole in the lid of a desiccator) over phosphorus pentoxide. Fig. zb shows
the rate of evaporation against water concentration in the material, expressed
as a percentage of the saturation water content in liquid water. The rate of
loss, for all values of water content down to 16%, is similar to that of a free
liquid-water surface; below 16% it falls regularly with water content.
soo r
E'300
100
X
20
40 60
?o water content
80
100
Fig. 2. a. Graph showing rate of uptake of a very thin tanned gelatin membrane when in
liquid water, b. Rate of loss of water when in dry air. c. Rate of uptake of water when in
100% R.H.
Water uptake from saturated air was similarly determined. The rate/con-
centration curve is shown in Fig. 2c. The initial rate is about one-twentieth
that for corresponding concentrations of water in the gelatin when in contact
with liquid water. The gelatin comes into some form of equilibrium with the
saturated atmosphere at about 16% of liquid saturation content, but when
left in the saturated air, small random changes in weight occur over long
periods of time, presumably due to the instability of static 'saturated' air.
The gelatin, when in equilibrium with saturated air at 16% water
content, will follow Fig. 2 a when transferred to liquid water. We suggest
that the relationship between tanned gelatin and water molecules may be
108 WATER TRANSPORT IN INSECTS
discussed in terms of two factors: a hygroscopic activity, and a * suction
force*. Suction force can only be exercised against water in the liquid
phase, and while it accounts for high rates of uptake in liquid water, its
contribution to uptake from saturated air is negligible. Nor can suction
.force retard evaporation from the surface; from Fig. zb retardation of
evaporation only occurs by hygroscopic activity when water content falls
below some 16%, which is also the figure at which uptake from a saturated
atmosphere materially ceases. The state of a membrane which has a greater
water content than some 16%, placed in saturated air, is very unstable.
(3) Water distribution in membranes. The curves enable us to obtain the
water concentration in the surface of a membrane when taking up water
from the liquid phase at a known rate, and also the surface content at the
other side of the membrane, evaporating at this rate into dry air. The actual
percentage of water content of the membrane, when transmitting at a
steady rate, is obtained by rapid removal from a membrane holder, blotting,
weighing, drying and reweighing. From a consideration of two or more
identical membranes which are superimposed (the permeability measured
as one thick membrane) and then at once separated and weighed, the
linearity of the gradient can be established. Results show that the gradient
is substantially linear, within the limits of the experimental method.
Passage of water through wax
A disk of beeswax was subjected to all the experimental conditions
imposed above on the tanned gelatin disk. Within the limits of surface
condensation, there is no measurable uptake or loss. So far as a composite
membrane is concerned, we can conclude that the molecules of wax will
not introduce forces tending to move water molecules, but only produce
a resistance to water flow. It is doubtful if they will counter-attract
evaporational loss at a wax surface. But, in addition to the suggestion that
the wax of a double membrane is a high ' resistance', the restriction of flow
through the tightly packed molecules of an orientated monolayer must
mean that the wax restricts the mean free path of the water molecule to
such an extent that it is in the liquid phase when traversing the wax, and
also at the wax/gelatin (or wax/cuticle) interface.
IV. A PHYSICAL BASIS FOR ASYMMETRY
AND WATER UPTAKE
Artificial systems
When a standard tanned gelatin disk is transmitting water from pure liquid
to dry air, at 15 mg./sq.cm./hr., the inner and outer surface concentrations
(read from the curves) are respectively 80 and 10%, so that a concentration
WATER TRANSPORT IN INSECTS ICQ
difference of 4-7 % gives rise to unit flow. When waxed and transmitting
in the direction gelatin to wax, the rate is 2 mg./sq.cm./hr.; the inner
concentration of the gelatin is 97%, the drop across the gelatin component
must be 2x4*7, and in the surface against the wax, 97 — (2x4-7), i.e.
87*6%. But water at this interface is still in the liquid phase and remains
so until emitted by the wax to the dry air. Therefore there is a suction force
at the gelatin/wax boundary tending to retard the outward flow of water
molecules. Water in the wax is flowing at the observed rate due to evapora-
tional disturbance of equilibrium at the outer surface, opposed by suction
force.
In reversed conditions, a flow of 8 mg./sq.cm./hr. from the gelatin
surface by evaporation means a surface concentration of 4-5 % water. The
concentration difference between the two sides is 8 x 4-7, so that the water
concentration at the gelatin/wax interface must be 42-1%. We are still
considering water in a liquid phase, so there will be a considerable suction
force on the inner surface of the wax, while at the boundary with liquid
water there is neither suction force due to the wax nor retention force due
to the water reservoir. The rate of flow through the wax must be due to
high gelatin suction force.
Now it is very doubtful if we could talk about a * force of evaporation '
nominally measured by the rate of evaporation of a pure water surface
into a given saturation deficiency, and equilibrate this with a suction
force, measured in the same units of rate of flow across unit area. But in
this experimental example, such an equilibration,
24 — 9'5^6'5^evaporation-suction force at 87-5%
95 i suction force at 42-5% '
gives a reasonable fit with the observed asymmetry ratio. The calculations
are supported by measurement of water content of waxed membranes in
a steady state of flow; these are in close agreement with the values to be
expected from the gradient concentrations.
General principles
It would seem that any membrane, consisting of two laminae, one of
which can exercise a considerable suction force on liquid water while the
other has very little water affinity, will show the phenomenon of asym-
metric water transport as defined in this discussion (see Hartley, 1948;
Beament, 19480). But it would seem necessary to add the limitation that in
the less hydrophilic component, water must be transported in the liquid
phase, otherwise the suction forces which cause this type of asymmetry
cannot act. It follows further that the greater the impermeability of the
110 WATER TRANSPORT IN INSECTS
* resistant* layer, the greater will be the asymmetry. For, in a tanned gelatin
model, if the wax could be so impermeable as to reduce water loss to
o-i mg./sq.cm./hr. (the order of impermeability of typical insect cuticle
and shell), then, with water moving in the direction gelatin to wax, the
gelatin will be almost saturated throughout, suction force preventing water
entering the wax will be negligible, and the measured permeability will
approach the theoretical value for a very thin wax layer alone. On reversal,
the gelatin will have a very low water content, and at the surface next to
the wax will still be so unsaturated as to exercise suction forces corre-
sponding with rates of flow in the 500 mg,/sq,cm,/hr. region. There is no
opposing force on the other side of the wax. One may, without commit-
ment, compare the rates of water movement of free water placed in the
situation of the wax in both cases. This is 24:500, and the ratio, 1 121, is
very close to the recorded asymmetry of Rhodnius exuviae.
Natural systems: cuticle
Under normal circumstances — certainly at the times when it is known
that an insect can take up water — there are pore canal tips, filled with living
material immediately below the tanned outer lamina of the epicuticle. It is
not therefore necessary to consider, in the first instance, the multiple layers
of the cuticle. Whether the actual order of suction forces produced by this
tanned material is the same as that demonstrated by tanned gelatin must
await experimental evidence, but it is a heavily tanned protein and one would
expect it to behave very similarly; at least the physico-chemical behaviour
of water movement in the epicuticle must be identical to the gelatin-wax
model. Further, the material between the living pore canal and the wax
is exceedingly thin, possibly less than i// thick, so that a relatively small
concentration difference across it would maintain an appreciable flow of
water. Hence, to achieve a suction force beneath the wax and a gradient,
the tanned material next to the cells themselves need be reduced in water
content to a very small degree below liquid saturation.
Following on the explanation of asymmetry on the exuviae of Rhodnius,
it is equally apparent that the smaller ratios found with whole cuticle must
be due to the greater thickness of the overall hydrophilic component ; this
means a much greater difference in concentrations between the sides of the
hydrophilic layers, with consequent reduction in the asymmetry of the
forces acting on the wax. We have already pointed out that asymmetry in
itself is of no great biological consequence to the insect ; but some insect
egg-shells, such as those of Rhodnius when first laid, have their wax as the
innermost layer of the membrane system, and are in effect cuticles turned
inside out. It is therefore interesting to note that these eggs (Beament,
WATER TRANSPORT IN INSECTS III
19486, 1949) — and, one suspects, those of other species — add material to
the inside of the shell after oviposition, which turns the system into a
'sandwich', rather than a reversed cuticle.
Water uptake in humidities less than saturation
Here, at the inner surface of the wax, there must be a greater force acting
inwards on water molecules than those due to evaporation at the outer
surface. But the wax enables the epidermis to make use of suction force in
the tanned epicuticle. If, in the round terms of the tanned gelatin model,
a relative humidity of 90% produces an outward force proportional to
saturation deficiency, i.e. to a free water-surface evaporation-rate of
2 mg./sq.cm./hr., then from Fig. 2a it is only necessary to reduce the
water content in the underlying component by about i % to achieve a
suction force to resist outward water movement. Seeing the small order of
actual uptake rates recorded by Lees (1946) and the thinness of the tanned
cuticular component, it seems likely that a decrease in water content of this
order might well set up the necessary gradient as well. It is surely within
the ability of living cells to regulate the water content of the tanned layer
to this extent, and thus to take up water from humidities lower than that in
equilibrium with the blood fluids.
When the cuticular wax is abraided, the tick at once loses water into
air at 90% R.H. Suction force can no longer act over the abraided area;
the denuded cuticle is similar to the gelatin in Fig. 26, where, until the
water content is reduced to values of the order of 16% of liquid saturation,
there is no retardation of evaporation below the rate given by a free-water
surface. It is certainly beyond our idea of vital secretion for the epidermis
to halt desiccation by such a reduction in cuticular water. There will be
a Brown-Escombe pin-hole effect from the abraided area, and undoubtedly
the water loss from this will mask any uptake which might still be going
on over the unaffected area, since unimpeded water loss will be so much
more rapid than uptake through the considerable resistance of the intact
waxy regions.
There is no need to evoke the idea that the physiological disturbances
reported in epidermal cells, under abraided regions, have knocked out
water-uptake mechanisms. While repair activity goes on for several days,
water uptake may recommence after i day's repair. But our evidence shows
that only a monolayer of wax is necessary to make a cuticle asymmetric :
to allow use of suction force on the condensed water in the wax, and there-
fore to set up the mechanism necessary to reverse the flow through the
cuticle. If the mechanism for wax secretion in the cockroach (Beament,
1952) is commonly represented in the arthropods, and freshly secreted wax
112 WATER TRANSPORT IN INSECTS
to repair the abraided area flows over the region in a solvent, then the
laying down of a monolayer could be achieved in the first day of repair.
Further, the part played by the epidermis in uptake against saturation
deficiencies is made clearer when one considers the circumstances in which
no uptake can be demonstrated. Lees (1946, 1952), who holds that the pore
canals of the engorged tick become progressively filled with solid material,
states that the ability to secrete water is lost in engorged animals; Mr M.
Locke (in unpublished observations kindly made available to me) has
evidence suggesting that the mealworm stops taking up water when the
epidermis moves away from the cuticle prior to moulting, and starts again
soon after the moult is complete. Obviously the close application of
protoplasm to the thin tanned protein of the epicuticle is essential to the
process; the cells take up water by regulating the water content of this
layer.
Uptake from water in the liquid phase
Matthee (1951) states that in Locustana eggs, liquid water is absorbed —
even water from hypertonic solutions — but eggs in this physiological state
will not acquire water from saturated atmospheres. The shell material
(around the hydropyle in this egg, but the argument holds over the whole
surface of other eggs showing similar phenomena) can exercise suction
force against water in the liquid phase, but cannot attract water in the
vapour phase unless the concentration in the outer shell is drastically
reduced. It is obviously well beyond the vital process to achieve such
a gradient across the thick shell layers, which would mean an even lower
concentration at the inner end of the gradient. When such eggs are desic-
cated, there must be very considerable suction forces possible, acting
outwards, in the inner layers of the shell. Providing there is a wax layer
there will be a considerable resistance to be overcome by these suction
forces and if there is (as is usual) a proteinaceous layer inside the wax, then
the concentration of water here will determine suction forces opposing on
the other side of the wax. Thus in the Rhodnius egg, when there are two
waxy layers, there could very well be a low water content inside the first of
these with consequent reduction in forces tending to move water outwards.
This idea is supported by the observed dramatic drop in water loss from
the Rhodnius egg just before blastokinesis (Beament, 1949). One might
almost envisage, along these lines, that the secretion, or removal, of layers
having high suction forces inside the wax layer in an egg would have
a greater importance on regulating water exchange than the secretion and
removal of wax layers themselves.
Lees has further shown in the intact tick, in a suitable state for active
water uptake, that the rate of uptake is rather more rapid when the animal
WATER TRANSPORT IN INSECTS 113
is placed in liquid water than when it is in a saturated atmosphere. Pre-
sumably this is merely due to the greater availability of water molecules at
the outer surface. (There is no question of a suction force acting inwards
on liquid molecules near the wax.) But when abrasion has removed the
tick's ability to obtain water from humidities less than 99% R.H., it can yet
take up water at a vast rate when immersed in the liquid, for without its
wax the cuticular material can exercise suction force, and there is no wax to
provide a high resistance to flow.
Locke (1953) has shown that if, following partial desiccation, mealworms
take up water from 93% R.H., they may occasionally reach an equilibrium
weight below their starting point. Following this, a period of desiccation
for 2 hr., which does not lead to a measurable loss of weight, promotes
a further uptake of water, when the animal is then returned to 90% R.H.
This is not a matter of restoring an equilibrium water content after desicca-
tion ; the new steady level following the second uptake is higher than after
the first uptake. But after short, vigorous desiccation, the mealworm will
at first lose water from its cuticle layers; an equilibrium with the environ-
ment affecting the whole system of cuticle and cells will eventually be set
up, but the immediate effect is to produce, by disturbance, a higher suction
force against the inside of the wax, with consequent opposition to water
loss. When thus transferred to high humidity at once, a considerable uptake
rate is promoted, due to physical processes, and apparently the epidermal
cells adventitiously absorb this water until a new equilibrium is established.
This would suggest that the epidermal cells, at least in mealworms, set
up gradients, respond to changes in cuticular water content and accom-
modate themselves to humidity conditions rather slowly. The total water
content of the tanned epicuticle, and the actual amounts of water exchanged
under normal circumstances, are very small indeed, and the rates and
amounts of exchange during abrasion are so great as to be considered very
abnormal circumstances for the cells. While we have concluded that the
process of abrasion does not, in itself, knock out the vital basis for secretory
uptake, it is still possible that the epidermal cells would show a wound
reaction to the vast water exchanges produced by abrasion, regardless of
mechanical damage to protoplasmic processes.
Following along the same line of argument, an egg (such as a cricket or
chafer egg), though immersed in distilled water, and without such an
impermeable shell as would be expected if a wax layer were present, could
nevertheless prevent any inflow of water, providing it could maintain in the
innermost layer of its shell a completely liquid-saturated layer. This would
mean maintaining 100% water, as opposed to something like 99% water,
in equilibrium with the osmotic pressure of the underlying cells, though
114 WATER TRANSPORT IN INSECTS
it would still be pure water in both cases. There could, under these condi-
tions, be no water uptake, since there could be no gradient falling towards
the inner side. There would, on the other hand, be a dynamic exchange due
to random diffusion. (The process could readily be examined by putting
eggs in heavy-water solutions, to determine exchange in equilibrium,
where no high resistance to flow is expected.) If the living material
decreases the water content of the inner shell, water will flow in, as does
happen in the mid-embrionic period, and flow will cease so soon as 100%
conditions are reasserted.
Two further important points arise when considering eggs. The rate of
loss of many eggs in dry air is considerably greater than that through
a corresponding area of typical cuticle. It has been assumed on this
evidence that such shells would not have a wax layer present. But if such
shells are analogous with the physical model of reversed cuticle, transpira-
tion could be tenfold that of normal cuticle, though an equally efficient
wax be present. We must, perhaps regretfully, conclude that a desiccation
rate alone does not give a reliable criterion of the presence or absence of
a wax layer; one must also consider the effect of the other shell components.
Measurement of the permeability of the shell in both directions should,
however, give diagnostic information.
Secondly, the air-sponge respiratory system of insect eggs (Wigglesworth
& Beament, 1950) may materially affect suction force in a laminar system.
A complete air layer in between two membranes limits water transfer to
diffusion, and is known, both in models and (for example) in the dipteran
puparium, to produce a most impermeable type of membrane system. It
prevents the suction force of the dry outer membrane from affecting the
inner one. The air sponge, even if filled with saturated air, will have
a similar effect, but the proteinaceous pillars across it will act as a con-
tinuous liquid path, and so decrease its effect on transpiration.
Tracheal systems
It is early to do more than speculate on the implication of these ideas on
water exchange in tracheal systems. So far as the main tracheal trunks
are concerned, an identical exchange process to that of external cuticle is
available to the insect; in the unwaxed tracheole one has usually considered
capillary forces as the main source of activity tending to remove water from
the tracheole end cell. (Such capillary forces may be much reduced by the
presence of polar substances — octyl alcohol, for example — believed to be
a component of cockroach grease as a solvent, which must, in the cock-
roach, and may, in other insects, contaminate the surface of the tracheolar
water (Beament, 19516, 1953; see also Wigglesworth, 1953). The process
WATER TRANSPORT IN INSECTS 115
of removing water from the tracheole must involve the suction force of the
tracheole lining, and desaturation of this lining by the end cell would
readily provide the necessary mechanism. That water moves at such
apparently great speed in the tracheole is suggested to be an illusion; true,
if the water were only removed at the inner end of the tube, the meniscus
would fall at an alarming rate, when one takes into account the viscosity of
water in such a fine tube. But it is strongly suggested that there is no
difference between the end of the tracheole and the rest of its length.
Hence water is removed over the whole surface in contact with the liquid
column, and there is no actual movement of the body of liquid down the
tube. Were it possible to observe the withdrawal of water in slow motion,
one imagines the liquid column to become hollow, giving the appearance
of a falling meniscus, while there yet remained a wall of water against the
membrane absorbing it. The energy requirement of this system is an
increase in surface energy of the water, as the meniscus is expanded, and
polar material would reduce this greatly.
Wax secretion
Among the problems outlined in this paper we have mentioned the
difficulty of imagining the secretion of wax layers in the presence of a water-
flow system, as a means of arresting water uptake. The suggestions made
here minimize the need for such a hypothesis. Nevertheless, following the
report of solvents in the cockroach cuticular grease, and their maintained
secretion to offset evaporating solvent, we should consider the further
possibility of such solvents in changing the permeability of cuticular wax,
and thus providing the insect with an additional mechanism with which it
may regulate the permeability of its cuticle. Cockroach and other cuticles
and membranes have been placed in an apparatus to measure water diffusion
between two humidities (as opposed to one humidity and liquid water).
The details of this apparatus and results obtained with it will appear else-
where, but a preliminary account of information so far obtained is relevant
to this discussion. Solvent vapours have been injected into the air on both
sides of membranes, and permeabilities obtained; the membranes have
been stored under vacuum for days, and then returned to the apparatus.
It is most significant that these cuticles show no reliable change in perme-
ability in either direction when octane, decane or octyl alcohol vapours are
present on both sides, in addition to the normal humidities; there is, indeed,
a suggestion, using octyl alcohol, that permeability is decreased, but this
may be due to the formation of monolayers on the reverse side of the cuticle.
Artificially produced membranes show increased impermeability under
these circumstances; vacuum treatment to remove solvents does not alter
8-2
Il6 WATER TRANSPORT IN INSECTS
the permeability of either natural or artificial cuticles. On the other hand,
chloroform, ether, benzene and acetone vapours all lead to an increased
permeability, irreversible on storage in vacuum, and presumably due to
permanent disorientation of the wax layer.
It seems striking that the particular solvents believed to be in cockroach
grease could not affect permeability; they may even act as a piston oil to
obtain and maintain a tightly packed monolayer. Subject to further investi-
gation, then, solvents do not represent a mechanism whereby water
exchange could be controlled.
We have taken no account, in these discussions, of the cement which is
believed to cover the wax layer of certain insect cuticles. If, as some workers
believe, this material is a tanned protein, then it would materially change
the distribution of forces acting in the outer layers of the cuticle, and cause
a considerable modification of the ideas expressed here. But the behaviour
of Rhodnius cuticle and cast skin does not lead us to believe that the cement,
in this animal at least, has any great effect on observed asymmetry.
Obviously some considerable investigation into the composition of this
material is of prime importance to our further understanding of the cuticle.
Should the cement prove to be impregnated with waxy material, then this
would remove its suction activity, while in no way interfering with its
apparent usefulness as a mechanical protection to the underlying wax layer.
V. CONCLUSION
No suggestion is made that the many water-exchange problems in insects,
some of which are reviewed above, are due entirely to inert physico-
chemical processes. The evidence suggests that these processes are very
much under the active control of living cells, which act by changing the
gradients in their overlying membranes. The significance of wax layers is
obviously much greater than the simple idea that they make an insect or
egg 'waterproof and impermeable '. Epigrammatically, they have also
apparently made water exchange and active secretion a much more available
process to the terrestrial arthropod, and it is surprising that so few ex-
amples of water uptake, from vapour or liquid phase, are known to us.
I am particularly grateful to Dr T. O. Browning, and to Messrs M. W.
Holdgate, R. Laughlin and M. Locke, for permission to include details of
their unpublished work. Prof. V. B. Wigglesworth, F.R.S., Drs A. D.
Lees, J. A. Ramsay, and others of the Department of Zoology, Cambridge,
have been good enough to read and criticize the manuscript.
WATER TRANSPORT IN INSECTS
REFERENCES
BEADLE, L. C. (1939). J. Exp. Biol. 16, 346.
BEAMENT, J. W. L. (1945). J. Exp. Biol. 21, 115.
BEAMENT, J. W. L. (1947). Proc. Roy. Soc. B, 133, 407.
BEAMENT, J. W. L. (19480). Disc. Faraday Soc. 3, 221.
BEAMENT, J. W. L. (19486). Bull. Ent. Res. 39, 359.
BEAMENT, J. W. L. (1949). Bull. Ent. Res. 39, 467.
BEAMENT, J. W. L. (19510). Ann. Appl. Biol. 38, i.
BEAMENT, J. W. L. (19516). Nature, Lond., 167, 652.
BEAMENT, J. W. L. (1953). Unpublished observations.
BROWNING, T. O, (1953). J. Exp. Biol. (in the Press).
CLOUDSLEY-THOMPSON, J. (1950). Nature, Lond., 165, 692.
DENNY, S. (1917). Bot. Gaz. 63, 373.
EDNEY, E. B. (1945). Bull. Ent. Res. 35, 399.
EDNEY, E. B. (1953). J. Exp. Biol. 29, 571.
FOGG, G. E. (1948). Disc. Faraday Soc. 3, 162.
HAMBURGER, F. (1908). Biochem. Z. n, 433.
HARTLEY, G. S. (1948). Disc. Faraday Soc. 3, 223.
HOLDGATE, M. W. (1953). Unpublished observations.
HURST, H. (1941). Nature, Lond., 147, 388.
HURST, H. (1948). Disc. Faraday Soc. 3, 193.
LANGMUIR, I. (1925). J. Phys. Chem. 29, 1585.
LAUGHLIN, R. (1953). Nature, Lond. (in the Press).
LEES, A. D. (1946). Parasitology, 37, i.
LEES, A. D. (194?)- J- Exp. Biol. 23, 379.
LEES, A. D. (1948). Disc. Faraday Soc. 3, 287.
LEES, A. D. (1952). Proc. Zool. Soc. Lond. 121, 759.
LEES, A. D. & BEAMENT, J. W. L. (1948). Quart. J. Micr. Soc. 89, 291.
LOCKE, M. (1953). Unpublished observations.
MATTHEE, J. J. (1951). Sci. Bull. Dep. Agric. U.S.A. no. 316.
MELLANBY, K. (1932). Proc. Roy. Soc. B, in, 376.
PIPER, S. H., CHIBNALL, A. C. & WILLIAMS, E. F. (1934). Biochem. J. 28, 2175.
RAMSAY, J. A. (1935). J. Exp. Biol. 12, 373.
RIDEAL, E. K. (1926). Introduction to Surface Chemistry. Cambridge University
Press.
SLIFER, E. H. (1938). Quart. J. Micr. Soc. 80, 437.
STELLWAAG, F. (1924). Z. angew. Ent. 10, 163.
WAY, M. J. (1950). Quart. J. Micr. Sci. 91, 145.
WIGGLESWORTH, V. B. (1931). Proc. Roy. Soc. B, 109, 354.
WIGGLESWORTH, V. B. (19330). Quart. J. Micr. Sci. 76, 270.
WIGGLESWORTH, V. B. (19336). J. Exp. Biol. 10, i.
WIGGLESWORTH, V. B. (1935). Proc. Roy. Soc. B, 118, 397.
WIGGLESWORTH, V. B. (1938). J. Exp. Biol. 15, 235.
WIGGLESWORTH, V. B. (1940). J. Exp. Biol. 17, 180.
WIGGLESWORTH, V. B. (1945). J. Exp. Biol. 21, 97.
WIGGLESWORTH, V. B. (1947). Proc. Roy. Soc. B, 134, 163.
WIGGLESWORTH, V. B. (1953). Quart. J. Micr. Sci. (in the Press).
WIGGLESWORTH, V. B. & BEAMENT, J. W. L. (1950). Quart. J. Micr. Sci. 91,
429-52.
THE EVIDENCE FOR ACTIVE TRANSPORT
OF MONOSACCHARIDES ACROSS THE
RED CELL MEMBRANE
BY PAUL G. LEFEVRE
U.S. Atomic Energy Commission, Washington, D.C.
In the passage of the blood sugar between the human red cell and the
plasma, certain complicating peculiarities have been apparent from the
earliest investigations. Among mammalian erythrocytes, those of the
primates appear to be unique in showing an appreciable degree of perme-
ability to the hexoses. Moreover, even these cells fail to haemolyse appreci-
ably when suspended in pure isosmotic glucose solutions, so that the entrance
of the glucose appears to be limited in some manner. This is not evident,
however, in the normal distribution of the human blood sugar, which appears
to be uniform throughout the water of the cells and plasma (Kozawa, 1914;
Ege & Hansen, 1927). Also, Klinghoffer showed in 1935 that there was rapid
equilibration of glucose added in small amounts to that already present in
the blood. Ege & Hansen concluded that the totality of information on
glucose distribution in the blood was * impossible to explain ' in keeping with
the natural assumption of the sugar's free solution in the two water phases.
Klinghoffer 's investigations of the apparent paradox revealed that ready
penetration of the sugar occurred only if the glucose concentration did not
exceed about 2 %. At higher concentrations, an extracellular excess was
maintained almost indefinitely ; and this unbalance was of sufficient degree
to account for the failure of haemolysis to appear in isosmotic solutions.
Bang & 0rskov (1937) measured this divergence from simple diffusion
behaviour by showing, in a few experiments with varying glucose concen-
tration in the neighbourhood of M/2O, that the conventional red cell
'permeability constant* was approximately inversely proportional to the
glucose concentration. Guensberg (1947) greatly extended this observation,
finding that the variation of the * constant*, inversely with the glucose
concentration, is over a range of at least a thousandfold.
Such behaviour implies some limitation on the absolute rate at which the
glucose can move into the red cell; one suggestion is that the process requires
participation of some ingredient of the barrier through which the sugar
must pass to enter the cell interior. In recent years, much additional
evidence has appeared in support of this view. Over the period 1946-52,
while at the University of Vermont, I have frequently returned to this
ACTIVE TRANSPORT OF MONOSACCHARIDES IIQ
problem, and would like now to summarize the lines of evidence which
indicate that the monosaccharides, in passing through the human red cell
surface in either direction, temporarily combine with a 'carrier* molecule
which is confined to that membrane or cortex layer. This evidence is in
general along three lines :
(1) the kinetics of the sugar movements;
(2) the mutual interference with the movements in mixtures of sugars;
(3) the action of inhibitory substances.
In my own work, each of these lines was studied almost entirely by
means of a single basic method, that of 0rskov (1935). This involves photo-
metric recording of light transmittance through a very dilute suspension of
red cells, as a means of following osmotic volume changes reflecting the
movements of water across the cell surfaces. The general procedures and the
operation of the recording system have been described elsewhere (LeFevre,
1948; LeFevre & Davies, 1951). The records show, as a function of time,
the changes in direct transmittance which occur in response to various
osmotically significant alterations in the composition of the medium (which
consists of a buffered balanced salt solution to which the test substances are
added). Since the suspension volume is at least 200 times as large as the total
cell volume in these dilute suspensions, the concentrations in the medium
are not appreciably altered by the cellular events, and may be treated as
constant. Also, since the passage of water across the cell membranes, under
a diffusion gradient, is much more rapid than the movements of the sugars
with which we are concerned, it is legitimate to consider the osmotic pressure
within the cells as identical with that of the medium at all times; the rela-
tively slow volume changes recorded are then taken as a measure of the
passage of glucose across the cell surface. Excellent linearity is found
between the recorded quantity and the haematocrit or the calculated cell
volumes in saline media of varying tonicity. When the sugars are present,
small empirical corrections (LeFevre & LeFevre, 1952) must be taken into
account if precise estimation of the cell-volume changes is to be attempted;
but for any but the most critically quantitative work, direct inspection of the
records is satisfactory for general analysis of the train of events. By arrange-
ment of a suitable sequence of sudden alterations of the total osmotic pressure
and the concentration of the penetrant in the medium, one can follow not
only the entry of the substance into the cell, but also its subsequent exit.
The anomalous behaviour of glucose is apparent in the simplest series of
this sort which was first attempted, in which glucose was simply added at
various concentrations to a suspension of washed cells previously glucose-
free. The pattern of the volume changes recorded in such an experiment is
shown in Fig. i . Certain clear deviations from the predictions of simple
I2O
THE EVIDENCE FOR ACTIVE TRANSPORT OF
diffusion, as expressed in Pick's law, are immediately apparent. Approach
to the equilibrium state is decidedly the more delayed, the more glucose is
added ; in fact, the initial rate of swelling actually decreases as the concentra-
tion of sugar is increased. At still higher concentrations, the rate of swelling
0-0204 M
0-082 M
Minutes
Fig. i. Kinetics of swelling in glucose-saline mixtures. At zero time, i ml. of saline
medium, containing glucose at u times final concentration shown, was added to 10 ml.
of cell suspension (£ vol. %) in saline medium. (Medium here was only 0-6 x isotonic,
so as to render rate differences more distinct.) 38° C. Immediate deflexion at zero time is
resultant of dilution of suspension (upward deflexion) and cell-volume change (shrinkage
downward); subsequent upward deflexion records cell swelling with uptake of sugar and
water.
diminishes markedly after the first few minutes, and drops to a nearly
imperceptible rate, while the cells are still much too small for an even
distribution of glucose to have been effected (for records see LeFevre, 1948).
This latter special complication will be taken up later; the point of special
interest in the pattern shown is that the apparent uptake of the sugar is not
proportional to the gradient, but is limited to a maximum rate dictated by
some other factor necessary for the translocation of the sugar.
MONOSACCHARIDES ACROSS THE RED CELL MEMBRANE 121
Such series of tests were run with all the hexoses and pentoses readily
available: D-dextrose, D-laevulose, D-mannose, L-sorbose, D-galactose,
L-arabinose and D-xylose. Among these, a clear dichotomy was apparent:
all the aldoses (dextrose, mannose, galactose, and the two pentoses) behaved
as just described; Fig. zb shows, for instance, the behaviour of galactose.
The two ketoses, laevulose and sorbose, as in Fig. 2 a, on the other hand,
seemed to obey reasonably well the predictions of Pick's law, and there was
no reason to suppose any limiting factor other than the passive permeability
of the cell membrane and the existing gradient for the sugar.
--/
I / (a) Sorbose
0-3 H
(b) Galactose
0 5 10 15
Minutes since addition of sugar
0 5 10 15
Minutes since addition of sugar
Fig. 2. Kinetics of swelling in sugar-saline mixtures. At zero time, 2 ml. saline medium,
containing sugar at 6 times final concentration shown, was added to 10 ml. cell suspension
($ vol. %) in saline medium. 37° C. Deflexions interpreted as in Fig. i.
This dissimilarity in behaviour of the aldoses and ketoses explains the
discrepancy between the data of Kozawa (1914) and those of Wilbrandt
(1938), with respect to the comparative rates of penetration of these sugars
into the red cell. Kozawa, who worked with approximately f-isosmotic
solutions at room temperature, found (by haematocrit and direct chemical
analytic methods) the following sequence, from fastest to slowest:
arabinose, xylose > galactose, mannose, sorbose > dextrose > laevulose ;
while Wilbrandt found, with much lower concentrations of the sugars (and
at body temperature), with an optical method:
xylose, arabinose > mannose > galactose > dextrose > sorbose > laevulose.
122
THE EVIDENCE FOR ACTIVE TRANSPORT OF
The contrast in the pattern of dextrose and sorbose penetration as a
function of concentration immediately accounts for the major disagreement
between these two series. The other minor discrepancies are also attributable
to the lesser differences between the sugars in this respect, in view of the
differing concentrations at which the two investigators were working.
Sorbose
10 15
Minutes since first addition
20
Fig. 3. Unilateral inhibition of uptake between sugars. At zero time, i ml. of saline
medium, with sugar indicated at 1-8 M, was added to 10 ml. of cell suspension (} vol. %)
in medium; about 13 mm. later, at time marked, a second i ml. was added, with sugar
indicated at same concentration. Final concentration of each sugar was thus 0-15 M, or
half-isosmotic. 37° C. Deflexions interpreted as in Fig. i.
The obvious hypothesis to be derived from these observations was that
the aldoses penetrate by a process involving participation of a cell com-
ponent, while the ketoses penetrate simply by passive diffusion. This
interpretation did not survive further experimentation concerned with the
influence of the presence of one sugar on the rate of penetration of another.
If, as seemed likely, all the aldoses shared a common transport system, there
should be mutual interference with their entry when two or more aldoses
are mixed ; whereas the rate of entry of a ketose into the cells should be
unaffected by the presence of other sugars of either type. This did not prove
to be the true situation; instead, all the sugars appeared to be involved in
a common reaction, so that in mixtures of any two the rate of swelling was
MONOSACCHARIDES ACROSS THE RED CELL MEMBRANE 123
always less than would be predicted on the basis of addition of the separate
entries of each sugar. This was as true when ketoses were involved as with
the aldoses. (For records, see LeFevre & Davies, 1951.)
Further light was shed on the situation by the procedure of adding the
sugars serially rather than simultaneously, awaiting equilibration of the
cells with the first before adding the second. In such experiments, the
effect of the presence of the first sugar on the rate of entry of the second
could be readily estimated in a quasi-quantitative manner. An example is
provided in Fig. 3, which shows the characteristic situation between any
aldose and either ketose. In the mixture of sorbose and glucose, each at
o- 1 5 M, the entry of the sorbose is essentially completely prevented ; whereas,
prior addition of the sorbose had no effect on the later uptake of glucose,
other than that attributable simply to its osmotic pressure. Similar relations
were demonstrable between the two ketoses, and between any pair of the
aldoses, except that the inhibitions were not always so overwhelmingly
unilateral nor so absolute. Depending on the particular pair of sugars
involved, the effect varied all the way from no detectable influence to
apparently complete inhibition. The results of this entire series of experi-
ments are summarized in Table i, which indicates the relative effectiveness
of each of the seven monosaccharides tested against each of the others.
From this information, slightly modified by secondary factors discussed
elsewhere (LeFevre & Davies, 1951), the avidity of the several sugars in
attachment to the carrier molecule was considered to decrease in the order
in which they are listed in Table i, with the two pentoses being indistin-
guishable. The largest gaps appear to be between the aldoses and the ketoses,
and between the two ketoses. Note that these relative affinities for the carrier
do not define the relative rates of penetration, although such correlation
improves as the sugar concentration is lowered.
Table i. Mutual inhibition in uptake of sugars
In
presence of
Inhibition of uptake of
Dext.
Mann. Gal. 1 Xyl.
Arab.
Sorb.
Laev.
Dextrose
Mannose
Galactose
Xylose
Arabinose
Sorbose
Laevulose
444
o
0
444|+44'444
i + +4~ T
000
0 , O O
+
o
4 + 44
444
o
f:f
o No, or doubtful, effect.
-I- Just noticeable inhibition.
-f 4- Moderate inhibition.
-f -I- -f Very marked inhibition,
-f + + + Essentially complete block of uptake.
I24
THE EVIDENCE FOR ACTIVE TRANSPORT OF
Another procedural variation, verifying the interpretation placed on
these experiments, involved the simultaneous setting up of equal and
opposite gradients for two sugars in a mixture. This was effected by equili-
bration with one sugar in a somewhat hypotonic saline medium, and sub-
sequent addition of the second sugar together with a quantity of concentrated
saline calculated to reduce the cell volume to the point that the outward
gradient for the first sugar momentarily exactly equalled the inward
Dextrose
Sorbose
Glyc.
Sorb.
Mann.
V
^^^1 M~~_
8
8
Minutes
Fig. 4. Unilateral inhibition between sugars of movement in opposing gradient. Just
prior to zero time, 10 ml. of cell suspension (£ vol. %) in 0-7 x isotonic saline medium
had been equilibrated with either sorbose or dextrose (as labelled) at 0-262 x isosmotic.
At zero time, 2 ml. was added, containing either glycerol, mannitol, sorbose, or dextrose,
as labelled, at 1-5 x isosmotic, in saline medium at 4-0 x isotonic. This was calculated to
reduce the cell-water volume from 1-43 to 0-80 x the ' normal', and thus set up an outward
gradient for the original sugar of 0-25 isosmotic units, just equalling the inward gradient
for the second non-electrolyte. 38° C. Deflexions interpreted as in Fig. i . For significance
of records see text.
gradient for the second. Thus the immediate cell volume assumed would
correspond to the final equilibrium volume, and any intervening changes in
volume reflect the net gain or loss as the two sugars move in opposite
directions. Fig. 4 shows the behaviour of the cells in one such experiment ;
in this instance, opposing gradients for sorbose and dextrose are set up in
each of the two possible arrangements. In both cases, the equilibration of
glucose proceeded without impediment, while the sorbose movement was
reduced to a small fraction of its uninhibited rate. The accessory records
in Fig. 4 show the uncomplicated exit of the original sugar in the presence
MONOSACCHARIDES ACROSS THE RED CELL MEMBRANE 12$
of corresponding concentrations of mannitol (non-penetrating) and glycerol
(penetrating much more rapidly than the sugars). Comparison with these
control records makes it clear that in the one case the intracellular dextrose
moves outward with almost complete exclusion of the sorbose ; while, in the
other instance, the intracellular sorbose cannot escape, so that after entry of
the dextrose the cell volume remains near the maximum level attained at the
beginning of the glycerol record (the record closely resembles the mirror
image of that for glucose exit in the presence of mannitol). With either
situation, the sorbose movement is reflected by only a very slow drift of the
record back toward the final equilibrium level.
These indications of competition among all of the sugars tested led to
early abandonment of the notion that only the aldoses shared the carrier
system. Additional indication that the ketoses were similarly involved was
found in the common sensitivity to inhibitory agents (discussed in a later
section of this report); also, all showed a similar Qw of the order of 3-0
(LeFevre & Davies, 1951). The fact that the ketoses failed to show the
limitation on rate of uptake into the cells, which was observed with the
aldoses, does not in itself militate against the hypothesis that the same
carrier system is shared by both sorts of sugars. The appearance of a
pattern resembling that of passive diffusion does not imply that neces-
sarily no limiting reaction with the cell surface is involved in the movement
of the ketoses into the cells, but might reflect simply a difference in relative
velocity constants as compared with the case of the aldoses.
This is immediately apparent in considering the properties of the
simplest model of the 'carrier system' that might be proposed. In the
absence of any preliminary demonstration of the degree of complexity
that might appear in (i) the formation of the sugar-carrier complex, (2) the
movement or reorientation of the complex, or (3) the uncoupling of the
sugar from the carrier, the least involved situation was first assumed.
Steps (i) and (3) may be treated grossly in terms of only the net ingredients,
so that any enzymic participation, or rate-limiting factors arising from
step (2), are reflected only in the overall velocity constants; the following
diagrammatic presentation emerges as representing the minimal essentials:
Outside Cell surface Inside
RI &3
P+ A* --A - P^=^^===^A +P
k2 kt
Cs = conc.ofP A8 = amount of complex A-P S = amount of P
A = total amount of carrier S/V = conc. of P
A—A = amount of uncombined carrier
126 THE EVIDENCE FOR ACTIVE TRANSPORT OF
in which V is the cell-water volume, and kl9 k2, k% and &4 are the velocity
constants for the several steps as labelled ; the equilibrium constant for the
reaction at the outer surface, K19 is then equal to k2/kl9 and similarly
K2 = kz/ki for the interior reaction. Mass action law would then give the
relations
— s8-t-8v, (i)
and -^Q^-40-M.-. (2)
No explicit solution of these equations to express S in terms of t appears
to be possible ; however, with glucose and the other aldoses, the observations
noted above allow special restrictions on the system which simplify these
relations. In the early stages of the process, while S is still a negligible
factor, the rate of entry is essentially k3A81 i.e. it is proportional to the
amount of sugar-carrier complex. The finding that the process is limited
so that no increase in initial rate occurs with increased concentration
indicates therefore that the amount of this complex, A89 remains nearly
constant in the face of variation in C8 over the experimental range. (In
Fig. i, the initial slopes are approximately inversely proportional to total
osmotic pressure.) This constancy of A8 indicates that the velocity constant
£3 is the factor limiting the overall transfer rate ; that the reactions at the
outer interface must be significantly faster than at the inner interface, so
that A8 is nearly in equilibrium with the sugar in the external medium.
Furthermore, k± must be considerably larger than k2, since jK\ is evidently
small compared to the lowest C8 at which the rate clearly ceases to increase
with C8.
The sequence of the sugars with respect to their competitive prowess in
utilizing the carrier system, discussed above, is interpretable in the same
terms. It presumably reflects the order of increasing Kl9 the dissociation
constant of the sugar-carrier complexes. The range of Cs in which the
experiments of the type represented by Fig. 2 were carried out (about
0-05-0-3 M) defines a range of magnitude of K± apparently exceeding that of
the aldoses, but not that of the ketoses. More precise calculation of this
constant for the various sugars will be considered later from an entirely
different experimental approach.
Assumption that Kl is negligible compared to Cs seems then to be
justified for the aldoses in the experimental range of C89 at least in the case
of glucose, the natural blood sugar in which we have the most interest. This,
together with the conclusion that the outer reactions are near equilibrium
by reason of the lower order of velocity constants at the interior, permits
MONOSACCHARIDES ACROSS THE RED CELL MEMBRANE I2J
explicit solution of the equations. Also, since we are dealing experimentally
with osmotic volume changes, we may treat Cs and S/V more properly as
thermodynamic activities than as concentrations, and in these terms it is
impossible for K± and K2 to be unequal. With these simplifications, the
earlier equations may be reduced to
^-Al
dt~^
in which Vt is the volume of the cell water at isotonicity (C,-), and Cm is the
concentration of the non-penetrating components (salts) in the medium (all
concentrations being expressed in osmotic terms). This may be integrated
directly to give the relation of S and t\ but since we are dealing with volume
records, and since S— V(Cm + C8) — CtViy it is convenient to express the
relation in terms of V:
which may be integrated to give
C8(Cm + QT VCM CM-CJK\
= ~A~k c — \ ( [} ~ c cv^cTv v (5)
^K^m L ^m ^iyi ^m y J
in which V0 is the cell-water volume when t = o.
This equation predicts that, in any given mixture, the course of volume
changes will follow the pattern dictated by the laws of passive diffusion ;
but that, in the comparison of rates in different situations, the pattern will
be entirely different from that derived from Kick's law.
The general applicability of this system to all situations with respect to
glucose movements across the red cell membrane, in either direction, was
tested in a wide variety of experiments involving as many contrasting
situations as could be arranged. Usually, several factors were held constant
while another was varied several times; for example, the initial cell volume,
the initial glucose gradient, the initial cell glucose level, the total glucose
transferred, the glucose level of the medium, or its total osmotic pressure.
A number of examples of the results of such experiments, involving both
outward and inward movements, have been illustrated elsewhere (LeFevre
& LeFevre, 1952); space limitations here allow only one example, in Fig. 5.
The match with the predictions from equation (5), which are shown for
comparison, is evident, and was equally good for all circumstances tested,
provided the extracellular glucose concentrations (C8) were not allowed to
exceed about 70 % of isosmotic.
128
THE EVIDENCE FOR ACTIVE TRANSPORT OF
The significance of the rate equation (3) above, which gives this fit with
experiment, is much more apparent after conversion to the following forms :
(6)
(7)
or
Fig. 5. Glucose entry in two stages, with fixed final Cs. (a) At zero time, to 10 ml. of cell
suspension (\ vol. %), i ml. was added containing glucose at 1 1 x the concentration
labelled; after equilibration, at second 'zero' time, an additional i ml. was added, con-
taining glucose sufficient to bring the final concentration in each case to 0-5 x isosmotic,
as labelled. All solutions contained salt mixture at 0-6 x isotomc. 37*5° C. (b) Pattern
for same experiment, on basis of the hypothetical carrier system ; scale for deflexions is
matched approximately to fit lowest record in (a).
Equation (6) calls attention to the fact that the rate is at any instant
directly proportional to the existing gradient, at a given extracellular concen-
tration, but that with a given gradient it is inversely proportional to the
MONOSACCHARIDES ACROSS THE RED CELL MEMBRANE I2Q
extracellular concentration. The rearrangement in equation (7) combines
these variables into a single term, i.e. the ratio of the intracellular glucose
level to the extracellular; the rate of transfer is proportional to the difference
from unity in this ratio. Thus the inward rate (positive dSjdt) can never
exceed Ak3, since the ratio cannot fall below zero ; but the outward rate is
not so restricted (the negative value of dS/dt will exceed Ak3 whenever S/V
is more than twice Cs).
Wilbrandt & Rosenberg (1950) found that the rate increased with increas-
ing concentration on the upper end of the gradient only up to a certain
maximum ; and that the rate decreased with increasing concentration on the
lower end of the gradient, but to a much greater degree than predicted by
Pick's law. It is interesting to note that this is exactly what would occur
according to the system developed above, if these experiments were per-
formed by varying the external-sugar concentration, holding fixed the cell-
sugar level. Table 2 illustrates how the carrier system would produce the
results reported by Wilbrandt & Rosenberg, if this procedure were used.
Table 2. Contrast of carrier and diffusion systems:
effect of varying sugar concentration
c.
Relative dSjdt by
Pick's law
Equation (7)
Entry: S/F=o-i,
0-2
100 — reference level
upper end of
0-3
200
133
gradient varied
0'4
300
150
0'5
400
1 60
0-6
500
167
Exit: S/V = 0-6,
o-i
-500
— 1000
lower end of
0'2
— 400
— 400
gradient varied
0-3
-300
— 200
0-4
— 200
— IOO
0-5
— IOO
-40
On the other hand, if the reverse procedure were followed, so that the cell-
sugar concentration became the experimental variable, with a fixed level in
the medium, then Pick's law and equation (7) would be indistinguishable,
and the peculiarities seen by Wilbrandt & Rosenberg should not appear.
Since these experiments have not been fully described, it is uncertain
whether they offer additional support for the scheme developed here, or
invalidate it.
The simple system definitely breaks down at sugar concentrations
.approaching isosmotic, which are of course far above the physiological
norm. The transfer of glucose slows down markedly after the first few
minutes, and may come essentially to a standstill while there is still a
E B S VIIT 9
130 THE EVIDENCE FOR ACTIVE TRANSPORT OF
considerable gradient across the cell surface. Several possible explanations
of this have been previously discussed (LeFevre & LeFevre, 1952). It was
possible to reject on experimental grounds the suggestions of loss of major
cell constituents, or of 'fixation* of the cells so as to preclude osmotic
volume changes. The most likely interpretation consistent with the facts
seems to be that the high glucose concentrations block the carrier reactions
themselves. Wilbrandt & Rosenberg (1950) have in fact taken this view in
a wider sense, claiming that the entire pattern of glucose movements
suggests a case of enzyme inhibition by an excess of substrate. Although
it does not appear from the experiments described above that this factor
is involved appreciably in the operation of the system at reasonable sugar
concentrations, it may well be the basis of its failure to function when Cs
becomes excessive. In interpreting this type of inhibition as observed with
DNAase activity, Cavalieri & Hatch (1953) point out that a molecule of
water is involved in the cleavage of the sugar-phosphate bond, and suggest
that the substrate may compete with water for a site on the enzyme. This
hypothesis could equally well be applied in the present instance.
That the complex formed in the membrane is in fact a sugar phosphate has
in no way been indicated directly by the work reviewed here ; but involve-
ment of some enzymic factor is implied. The operation of inhibitors has
been suggestive; inhibition of the uptake of glucose into red cells was
effected by very small concentrations of Hg+4~, HgJ+, or ^-chloromercuri-
benzoate (LeFevre, 1947, 1948), and by chloropicrin, bromacetophenone,
allyl mustard oil, or gold (Wilbrandt, 1950). (Iodine is also an effective
inhibitor, but only at concentrations which also lead to an obvious dis-
coloration of the haemoglobin.) The efficacy of this group of substances
suggests that some part of the transport process involves sulphydryl groups ;
if so, these groups are evidently of the not easily available type characterized
by Barren & Singer (1945), since there appears to be no inhibition at all by
Cu++, alloxan, mapharsen, iodoacetate or arsenite (LeFevre, 1948).
Use of another class of inhibitors has more recently been particularly
fruitful in the analysis of the carrier mechanism; I refer to the glucoside
phlorizin (generally considered to be rather specifically active against
phosphorylation transfer systems), and its aglucon, phloretin (/?-(/>-
hydroxyphenyl) 2, 4, 6-trihydroxypropiophenone). Either of these agents
acts as a block to the transfer of the monosaccharides across the human red
cell surface; but as Wilbrandt (1950) has shown, the simpler molecule,
phloretin, is many times more effective than its glucoside phlorizin.
Wilbrandt expressed the conviction that these agents act on the process by.
which the sugar emerges from the membrane (whether this be on the inside
or the outside) rather than on the step of entry into the membrane. He could
MONOSACCHARIDES ACROSS THE RED CELL MEMBRANE 131
show inhibition of glucose exit from the cell without any disturbance of its
entry, when (by reason of slow penetration) the agent was more concentrated
in the external medium than within the cell. Wilbrandt suggests that phos-
phorylation by hexokinase is concerned in the initial step, and dephos-
phorylation by a phosphatase in the second step, and that it is this latter step
which is sensitive to phlorizin and phloretin. In Wilbrandt's scheme, the
system does not simply consist of a reversible set of reactions, but involves
different operating units according to whether the sugar is entering or
leaving the cell. With such a system it is difficult to account for the apparent
failure of glucose ever to accumulate against a concentration gradient in
these cells; whether a reasonable fit with the observed kinetics could be
achieved with this system has not been considered.
Wilbrandt's published statements with regard to the peculiar action of
phloretin in selective inhibition of the exit process have been so far only
qualitative descriptions; a complete statement of procedure would be
helpful, as without this it is impossible to determine whether the observa-
tions actually refute the simpler interpretation of the inhibition under the
scheme offered here. My own experiments are entirely in accord with the
hypothesis that the phloretin acts on the first reaction involved, by direct
competition with the sugars for combination with the carrier molecule (or
the limiting molecule involved in the chain leading to formation of the
carrier complex). The following analysis, derived from this hypothesis, has
in fact permitted rough calculation of the dissociation constants of some of
the carrier complexes.
If the inhibitor acts at low concentrations by combining with the carrier
in the same manner as do the sugars, it must have considerably higher
affinity for the carrier (a much smaller K). Thus, when an extra ingredient
of this type is added to the former system,
dS A(k3CsIKa-k,S/V)
dt -
in which Cl and KT are respectively the concentration and equilibrium
constant for the inhibitor and K8 is the equilibrium constant for the sugar
(equal to K: or K2 of the original system). The ratio of the uninhibited rate
(R0) to the inhibited rate (RT) is then given by the relation
0 8
Thus, in a plot of this ratio against C7, in a series of tests in which only
Cz is varied, a straight line should be obtained, the slope of which is
9-2
I32
THE EVIDENCE FOR ACTIVE TRANSPORT OF
This rectilinearity is observed experimentally, as shown in Fig. 6. The
records of cell shrinkage during glucose exit, in a series of concentrations
of phloretin, under an otherwise constant set of conditions, are given in
Fig. 6 a. From such records the relative initial rates of glucose loss may be
estimated and compared as a function of the inhibitor concentration.
Fig. 6b shows the data of this same experiment, plotted in the manner
prescribed above ; a similar set of data for inhibition by Hg++ is included for
comparison, showing that with this agent the inhibition is clearly not of the
competitive type.
Minutes
M- phloretin
Fig. 6. Inhibition of glucose exit as a function of inhibitor concentration, (a) A 3 % cell
suspension was equilibrated at 38° C. for i hr. with 0-5 x isosmotic dextrose in 0-7 x iso-
tonic saline medium. Then, at zero time, 2 ml. of this was added to 10 ml. of the saline
medium, containing phloretin so as to make the final concentration of the inhibitor as
labelled in the figure, (b) The data of (a), and a similar experiment with HgCl2 in place
of the phloretin, plotted as suggested in the text.
More convincing evidence of the competitive nature of the phloretin
inhibition is obtained from consideration of the effect of varying the sugar
concentration, with a fixed inhibitor concentration. Equation (9) may be
rearranged _?/_=^//^ I\. (I0)
RQ - Rj Cf \KS J '
so that if RI(RQ-RI)~l is plotted against Cs at a fixed C7, it should yield
a straight line with -Ks as the ^-intercept and K1\Cl as the jy-intercept.
By this means, then, both Kr and Ks can be estimated. Such a graph, for
MONOSACCHARIDES ACROSS THE RED CELL MEMBRANE 133
inhibition of glucose exit by phloretin, is presented in Fig. 7 ; this experi-
ment gives a glucose K8 of 0-009 M, an<^ f°r phloretin a Kz of 4-9 x io~6M.
Thus the inhibitor's ' affinity ' for the carrier appears to be about 1800 times
that of the sugar.
The useful measurements obtained by this approach are summarized in
Table 3. The work was necessarily cut short soon after the initiation of this
phase in August 1952, and it was impossible to gather a full complement of
0-8 -
006M
Fig. 7. Inhibition of glucose exit by phloretin as a function of glucose concentration.
Procedure as with Fig. 6 a, except that glucose was added in varying amounts to the final
mixture; at each Cs, two runs were taken, one with and one without phloretin at
5-5 x io"6 M. For rationale of system of plotting data, see text.
data for estimation of Ks of each of the sugars and to check the KT for
phloretin and phlorizin using each of the sugars as test penetrant. However,
the legitimacy of the interpretation of the observed rectilinearity in the
plotted relations is attested by the rinding of reasonably similar values in the
constants with different experimental procedures. It is especially to be
noted that the dissociation constants for sugar and inhibitor are of similar
magnitude in experiments with outward movement as with inward move-
ment. Most reassuring perhaps is the finding of the same range of value
for Kj with different sugars having quite different K8.
THE EVIDENCE FOR ACTIVE TRANSPORT OF
These results lend new support to the previously postulated interpreta-
tion of the differences in behaviour between the aldoses and ketoses, and of
the pattern of competitive inhibition between the various sugars. Thus, the
value found for glucose K was appreciably smaller than the Cs range useful
in the experimental procedures; that for galactose was at the lower edge of
this range, whereas the K for the two ketoses was in excess of the upper
experimental limit for C8. The practical limitations on C8, in fact, made it
impossible to ascertain the ketoses* equilibrium constants with any satis-
factory degree of precision. In contrast to the situation with the aldoses, the
concentration of the ketoses (C8) did not affect appreciably the degree to
which a given concentration of inhibitor would act. The plot of Rj(RQ — Rf)
against C8 therefore gave for the ketoses a line of such low slope that the
location of its ^-intercept was a matter of enormous uncertainty. Neverthe-
less, the experiments with the ketoses gave a similar Kl for phloretin, and
showed equally good rectilinearity in the plot of R~l against C7. All this
is in complete accordance with the theoretical relation derived above, in
the contrast of the situation K8^>CS, with the situation 1
Table 3. Estimation of carrier-complex dissociation constants
Inhibition of
Sugar K9 Phloretin Kj* Phlorizin Krf
Dextrose entry
Dextrose entry
Dextrose exit
Dextrose exit
Dextrose exit
Dextrose exit
7'5 X I0~3 M 4-5 X I0~6 M —
10 X I0~3 M — —
9Xio~3M 4'9Xio~8M —
8 X I0~3 Mi
8xio"3M — —
7'5 X I0~3 M I'4SXIO~4M
Galactose entry
Galactose exit
5'0 X I0"a M —
4'4Xio~2M 4'8xio~9M —
Sorbose entry
Sorbose exit
1-3-2-0 Ml I'27XIO~4M
ca. 2, Mj i 4-4 x io~6 M —
Laevulose entry
ca. 2 M| 4-4 x io"6 M —
* Five values given represent complete series permitting plot as in Fig. 7; where no
value is listed, phloretin Kt was taken as 4-7 x io~8 M in calculation of Ks from plot as
in Fig. 66.
f Two values given are on basis of parallel tests with phloretin, taking K/ of latter as
4-7 x io~6 M.
t Sorbose and laevulose K cannot be satisfactorily estimated in these experiments ; see
text for discussion.
In summary, then, the several lines of attack have all fitted into the
schematic system illustrated above. This does not mean that the actual
mechanism may not be considerably more complicated, with extra steps
involving additional components, perhaps enzymic, which are not speci-
fically included in the postulated system. It does indicate, however, that
any such additional steps do not represent separate rate-limiting factors,
MONOSACCHARIDES ACROSS THE RED CELL MEMBRANE 135
so that for kinetic analysis they can be lumped together into the two overall
reactions dealt with here.
Beyond the suggestiveness of the nature of the inhibitors found to be
effective, there has not been any indication in this work of the nature of the
carrier-complex or of the probable enzymic factors involved, Phosphoryla-
tion is of course the obvious suggestion; glucose-6-phosphate does not
measurably penetrate the red cell, however, and it seems unlikely that the
complex which is supposed to be confined to the surface layer would show
this inability to enter that layer from the medium. Wilbrandt (1950) has
described a possible form of hexose-metaphosphate which should be
unionized and fairly fat-soluble, and has suggested that this could well be
the complex involved ; there does not seem to be any direct evidence of this
at the moment.
Finally, it should perhaps be emphasized that, whatever the details of
the mechanism may prove to be, there is no evidence that the red cell is
equipped with a hexose 'pump' that can provide the energy for trans-
porting sugar against a concentration gradient. The data merely indicate
that there is a temporary complex formed between the sugars and a cell-
surface component during the transfer (in a somewhat circumscribed
manner); and this is apparently the basis for the peculiar ability of the
primate erythrocyte to take up these substances.
REFERENCES
BANG, O. & 0RSKOV, S. L. (1937). J. Clin. Invest. 16, 279.
BARRON, E. S. G. & SINGER, T. P. (1945). J. Biol. Chem. 157, 221, 241.
CAVALIERI, L. F. & HATCH, B. (1953). J. Amer. Chem. Soc. 75, mo.
EGE, R. £ HANSEN, K. M. (1927). Acta med. scand. 65, 279.
GUENSBERG, E. (1947). Die Glukoseaufnahme in menschliche rote Blutkorperchen.
Inauguraldissertation, Bern, Gerber-Buchdruck, Schwarzenburg.
KLINGHOFFER, K. A. (1935). Amer. J. Physiol. in, 231.
KOZAWA, SHUZO (1914). Biochem. Z. 60, 231.
LEFEVRE, P. G. (1947). Biol. Bull., Woods Hole, 93, 224.
LEFEVRE, P. G. (1948). J. Gen. Physiol. 31, 505.
LEFEVRE, P. G. & DAVIES, R. I. (1951). J. Gen. Physiol. 34, 515.
LEFEVRE, P. G. & LEFEVRE, M. E. (1952). J. Gen. Physiol. 35, 891.
0RSKOV, S. L. (1935). Biochem. Z. 279, 241.
WILBRANDT, W. (1938). Arch. ges. Physiol. 241, 289.
WILBRANDT, W. (1950). Arch. exp. Path. Pharmak. 212, 9.
WILBRANDT, W. & ROSENBERG, T. (1950). Helv. physiol. acta, 8, €82.
SECRETION AND TRANSPORT OF
NON-ELECTROLYTES
BY W. WILBRANDT
Pharmacological Department, University of Berne, Berne
I. INTRODUCTION
Active transport of non-electrolytes has been observed mainly in con-
nexion with nutrition and excretion as well as with osmoregulation (trans-
port of water).
This report will concentrate on the transport activity of the epithelial
cells in the intestine and the kidney as well as that of red cells. Special
attention will be given to monosaccharides. There will be no discussion of
active transport of either water or inorganic cations which are covered in
other papers.
If, however, weak electrolytes, both acids and bases, were excluded in
strict accordance with the title of the paper, the list of actively transported
substances would become rather small.
This seems of some interest, because it provides a partial answer to the
first question to be asked : Which sort of substances are apt to be transported
actively?
Substances which appear to be actively transported include the common
foodstuffs, sugar, amino-acids, possibly fats and a larger series of waste
products as well as foreign substances.
Sugars and amino-acids are transported in the intestine as well as in the
mammalian kidney, due to the particular mode of action of this excretory
organ involving at first filtration so to speak of the entire ' milieu interne*
and then reabsorption of those components that are essential to the
organism.
The mode of working of the mammalian kidney has been elucidated first
by the classical work of A. N. Richards and his group with micropuncture
technique on single nephrons. Later, on the basis of the results so obtained,
extensive work on the whole kidney by various groups, including those of
H. W. Smith, Shannon, Pitts and others, has made available a large body
of information. (A comprehensive survey has recently been given by
H. W. Smith (1951).)
The function of the kidney consists of three elements : filtration in the
glomeruli, reabsorption in the tubules and secretion in the tubules.
Whereas filtration so far has been considered as passive, reabsorption may
SECRETION AND TRANSPORT OF NON-ELECTROLYTES 137
be and secretion always is an active process. Active reabsorption is assumed
if the concentration ratio urine to plasma (U/P) falls (or may under certain
conditions fall) below 1*0. Particularly this is the case with the so-called
* threshold substances' which appear in the urine only when a threshold
concentration in plasma is surpassed. They include sugars, amino-acids, urea
in the kidney of elasmobranchs, lactic acid, phosphate, sulphate, uric acid,
ascorbic acid and possibly more substances that have not been investigated
yet. With the apparent exception of uric acid, they are all substances which
serve some special purposes in the body and may be called essential
substances.
Table i a lists substances which are actively reabsorbed.
Secretion, with few exceptions, is restricted to organic electrolytes, both
weak and strong acids and strong bases (Table i b and Fig. i). They include
the group of iodinated X-ray contrast substances which have been selected
commercially, according to their ability to be concentrated in the urine,
without knowledge of the mechanism involved, and have later been among
the first cases to be recognized as secretion. Their iodine-free nuclei
behave in the same way, showing that iodine as such is not essential for the
reaction of the kidney tubule cell. There are, furthermore, several deriva-
tives of hippuric acid which for a long time has been known as a product
of ' detoxification*. Hober (1945) and later Sperber (1947) have pointed
out that an important feature of many detoxification processes is the forma-
tion of rapidly excretable compounds. Furthermore, there is penicillin whose
rapid excretion has long been one of the foremost problems in practical
penicillin therapy, and caronamide which was introduced into therapy to
Table i. Actively transported substances in the kidney
i a. Absorption i b. Secretion
Glucose Tetraethyl-ammonium
Xylose Methyl-nicotinamide
Fructose Phenol red
Galactose Penicillin
Ponceau R
/?-Hydroxy butyric acid Caronamide
Lactic acid Hippurate
Uric acid m-Amino-hippurate
Ascorbic acid />-Amino-hippurate
/>-Acetyl-amino-hippurate
(Urea) Hippuran
Glycine o-Hydroxy-hippurate
Alanine Diodrast
Glutamic acid lopax (Uroselectan)
Creatine Neo-iopax
2-Pyridine- 1 -acetic acid
Lysine Skiodan
Arginine
Histidine
Leucine
Isoleucine
138
SECRETION AND TRANSPORT OF NON-ELECTROLYTES
55
6
0=0
1
ffi
3
2
I
in
•5
-o
bo
£
o
/
£
U
O=U
a=u
CJ
/
0=0
o=u
ffi -
X
CJ
SECRETION AND TRANSPORT OF NON-ELECTROLYTES 139
compete with penicillin in the kidney. Phenol red was historically the first
substance whose secretion was experimentally established. The sulphonic
acid azo dyes studied by Hober & Briscoe-Woolley (1940) are strong acids.
Finally, methyl-nicotinamide and tetraethyl-ammonium are quaternary
ammonium bases.
In the intestine less information is available. Examples of active absorp-
tion are known mainly among the foodstuffs: sugars, amino-acids and
possibly fats. The question of secretion in analogy to the kidney has as yet
been very little investigated. It was found, however, that diodrast is
actively secreted in the small intestine (Smith, 1951).
A survey of these actively transported substances shows that those
reabsorbed are strongly hydrophilic, while those secreted quite generally
have molecules possessing both hydrophobic non-polar and hydrophilic
polar groups (some of the latter ionizing), indicating that this polar/non-
polar type of molecular structure favours secretory transport. Hober
(1940), in a study of a large number of sulphonic acid azo dyestuffs was able
to show that, in the case of the kidney, as a rule only those dyes are secreted
which are asymmetric with respect to the sulphonate group, possessing
one or two groups on one ring but none on any other.
With respect to the permeability of the cell membrane this rule indicates
that a high ability of penetration is not, as might be expected, a condition
of secretory transport. On the contrary, we find in the list practically only
one compound known to penetrate cell membranes easily, namely, urea,
whose secretory transport, however, is a specialized case in the group of
elasmobranchs. Several compounds related to hippuric acid and diodrast
have been shown not to penetrate red cells (w-amino-hippurate, />-amino-
hippurate, />-acetyl-amino-hippurate) or only slowly, not reaching equili-
brium in vivo (p-hydroxy-hippurate, o-hydroxy-hippurate, cinnamoyl
glycine, hippuran, diodrast, iopax). Amino-acids, quaternary ammonium
compounds, and in most species the sugars, also penetrate either with
extreme slowness or practically speaking not at all. This rule and its bearing
for the transport mechanism will be referred to later.
II. CHARACTERISTIC FEATURES OF ACTIVE
TRANSPORT
Before the mechanism of active transport is discussed, some remarks seem
appropriate with respect to several features more or less common to those
transports, in contradistinction to diffusion processes. From any theory
advanced for the transport mechanism an adequate explanation of these
features should be required.
140 SECRETION AND TRANSPORT OF NON-ELECTROLYTES
(i) Transport against the gradient of electrochemical potential
These transports (which have also been termed * uphill transports') are
the only ones for which the epitheton * active* is not disputable (cf.
Rosenberg, 1948). They are opposite to the thermodynamic tendency and
cannot occur spontaneously without energy-providing mechanisms.
Both kidney and intestine perform * uphill transports'. In the kidney, as
far as reabsorption is concerned, this is obvious for all substances with U/P
ratios smaller than i. They include the threshold substances as well as
fructose, galactose and others.
For substances secreted in the tubules, uphill transport is indicated if
the U/P ratio is larger than that for inulin (which is only concentrated by
reabsorption of water in the tubules). This is true for all the substances
listed in Table i b. Most of them have U/P ratios close to that of />-amino-
hippurate (which is about 5-3 times higher than that of inulin); some have
lower ratios, down to 3. If no back-diffusion occurs, these values are
minimal figures for the accumulation ratio in the secretory transport. They
hold if the transport delivers the substances into the tubular urine before
reabsorption of water begins. If the level of entrance into the tubule is
lower than the level of the beginning of water reabsorption, the accumula-
tion ratio must be higher to account for the final U/P ratio.
In the intestine, convincing evidence for 'uphill transport' has been
furnished by Barany & Sperber (1939) with respect to the absorption of
glucose. In this study elimination of the disturbing factor of water re-
absorption was accomplished by the addition of sodium sulphate, which is
poorly absorbed by the intestinal epithelium.
(2) Non-linear rate-concentration relationship
The rate of a diffusion process when following Pick's law is proportional
to the slope of the concentration gradient. In the case of diffusion from
finite to zero concentration (or in the case of the diffusion flux) it is pro-
portional to the concentration: dA\dc — constant (A== rate of transport,
c — concentration).
For active transports this is not true. With rising concentrations dAjdc
decreases. In many cases it was found finally to reach zero, indicating a
maximum rate of transport, which cannot be further increased by raising
the concentration. In kidney physiology this maximum rate has been
designated as Tm.
For the absorption from the intestine dA/dc was studied with respect to
sugars by Cori (1925), Verzar (1935), Hober & Hober (1937), Barany &
Sperber (1942), and by Vidal-Sivilla (1950), with respect to amino-acids
by Hober & Hober (1937). The results show definitely that dA/dc is not
SECRETION AND TRANSPORT OF NON-ELECTROLYTES 141
constant. The shape of the rate-concentration curve was not studied in
detail for amino-acids and could not be established very clearly for sugars
due to considerable scattering of the data. It seems, however, doubtful
whether a maximum rate is actually attained with high concentrations. In
the range studied it was not reached. It has been assumed (Verzar, 1935 ;
Donhoffer, 1935) that absorption is composed of two fractions, an active
process and passive diffusion. In this case a constant maximum rate of
absorption could not be reached because the diffusion fraction would
continue to increase with rising concentrations.
More data are available with respect to reabsorption and secretion in
the kidney. As to reabsorption, non-linearity of the rate-concentration
function is clearly shown by the existence of the threshold substances*
mentioned above. (As * threshold* critical levels of plasma concentrations
are designated, above which excretion in the urine occurs, whereas with
lower concentration reabsorption is complete.) Threshold substances in-
clude glucose, amino-acids, lactic acids, phosphate and sulphate.
Non-linearity, however, is not limited to the group of threshold sub-
stances. It has been observed in a large number of cases, both of reabsorption
and of secretion. In many of them (but by no means in all) with high
concentrations a maximum rate of transfer was reached.
Fig. 2 shows excretion and reabsorption of glycine (Pitts, 1943) in the
dog as a function of the filtered amount per minute (filtrate volume per
minute times plasma concentration). Both curves clearly show that with
high values of the filtered amount a constant maximum rate of re-
absorption is reached. This maximum rate is approached with different
rapidity for different substances, e.g. more rapidly for sulphate than for
glycine.
With respect to the form of the A\c curves and their differences reflected
in the curves of Fig. 2 some remarks seem useful.
Since the filtered volume per minute (glomerular filtration rate) varies
relatively little, the abscissa is approximately proportional to the substrate-
plasma concentration and consequently to the concentration in the
glomerular filtrate entering the tubule. Nevertheless, the curves designated
* reabsorption ' (dots) should not be taken as showing the form of the A/c
function. The amount reabsorbed per unit time, T, is an integrated value,
the sum of the amounts reabsorbed at all levels of the tubules with decreasing
values of c. It may be calculated for a given form of the A\c dependence.
In Fig. 3 curves are shown which have been calculated on the (arbitrarily
chosen) basis of the Michaelis-Menten equation
142 SECRETION AND TRANSPORT OF NON-ELECTROLYTES
SECRETION AND TRANSPORT OF NON-ELECTROLYTES
o
2
I
|
o5
"2
c
c
<u
co <-•
-2-1
II
o o
8 -o §
o o c«
" SJ
s IB-
^c
-o
144 SECRETION AND TRANSPORT OF NON-ELECTROLYTES
where A is the amount reabsorbed per unit length of the tubule and unit
time, c the concentration, Km the Michaelis-Menten constant and K a
second constant (in an enzymic reaction indicating the enzyme con-
centration). T was obtained by integration over the entire tubule, using the
relations dT=Adx (2)
and T=CF(CQ~C\ (3)
where x is the distance from the beginning of the tubule, T the amount
reabsorbed per unit time between the beginning of the tubule and the
distance x, CF the volume filtered per unit time (rate of glomerular filtration)
and CQ the concentration in the glomerular filtrate.
The integration yields the transcendental equation
mlnC°. (4)
The curves in Fig. 2 were calculated for three values of Km: o-i, i-o and
10-0 (assuming a constant value of 20 for Kx).
They resemble those of Fig. 2 in the general shape. The difference
mentioned between glucose and sulphate, under the assumptions used
here, would appear to reflect differences in Km.
It should be noted, however, that the T/CQ curves are by no means
identical with the A\C curves, as Fig. 3 shows. For instance, the concentra-
tions for 50% maximum rate are £ = o-i, i and 10 with respect to A, but
C0= 10, 10 and 15-6 for T.
Thus from TjCQ curves the function Ajc cannot be evaluated directly.
The Michaelis-Menten relation was chosen arbitrarily. It is not likely
that the true function is equally simple. It is very probable, however, that
it will contain functionally homologous factors: a factor determining the
maximal rate, homologous to K in the Michaelis expression (indicating the
enzyme concentration) and another factor determining the rapidity of
approach to the maximal rate with rising concentrations, homologous to
Km. They may be termed extensive and intensive factors.
As would be expected the values of Tm vary considerably. For the com-
parison of data obtained in different species it has become customary to
reduce them to i m.2 of body surface. So calculated the following values
of Tm in mmol./min. m.2 of body surface have been reported (taken from
Smith (1951) and changed into the units adopted here):
Glucose man 6* 1*2 Leucine dog 0-64-0-71
Glucose man $ 0-97 Sulphate dog 0-08-0-1
Glucose dog 0-67 Phosphate man 0-075
Glycine dog 1-29-2-29 Uric acid man 0-056
Valine dog 0-93 Ascorbic acid man 0-007
SECRETION AND TRANSPORT OF NON-ELECTROLYTES 145
The values of Tm appear to reflect roughly the degree of importance of
the substances as metabolites. The highest figures are those for glucose and
for glycine.
It seems useful to refer one of these values to unit surface area of the
absorbing tubule cells. The total number of nephrons in man are estimated
at two millions, the length of the proximal tubule is given as 13 mm., the
diameter of the lumen as about 20/1. This would give a total surface area
of i*6 x io4cm.2 ( = 1-6 m.2!). Tfor 1-73 m.2 body surface (instead of i m.2
as listed above) is 373 mg. /min. = 2 mmol./min. (which per day amounts to
a transported quantity of 480 g.). The amount reabsorbed per second and
cm.2 surface area thus would be 0-21 x io~8 mol. This value will be referred
to later.
For secretion likewise constant maximal rates at high concentrations or
at least non-linear relationship between rate and concentration are observed.
Diodrast Tm values determined on man range from 36 to 52-8 mg./min.
and 1-73 cm.2 body surface, the average being higher for males (49*9) than
for females (42). Tm for/>-amino-hippurate in the same units is 53 mg./min.
In terms of moles these values of Tm (0-117 mmol. for diodrast and
0-273 mmol. for^-amino-hippurate) are considerably lower than for glucose.
Phenol red shows a definite maximal rate not only in the glomerular kidney
(Tm for man 35-8 mg. = 0-103 mmol./min. per 100 ml. glomerular filtrate,
for the dog less), but also in the aglomerular kidney, where its demonstration
by Marshall & Crane as early as 1924 was one of the first observations
of this kind.
(3) Competition
If the maximal rate of transport is interpreted as due to a limited capacity
of the transport mechanism, competition among substances transported by
the same mechanism is to be expected. Several examples are known.
In the intestine Cori (1926) showed that the amount absorbed from a
mixture of two sugars was less than the sum of the amounts absorbed from
solutions of the single sugars in the same concentrations, in fact, it was only
about equal to each of these amounts.
Again in the kidney more data are available. Several examples are known
of transport substrates inhibiting the transport of others, both in reabsorp-
tion and in secretion.
With respect to reabsorption saturation of the transport mechanism by
high concentrations of glucose blocks the reabsorption of xylose com-
pletely, that of fructose incompletely (Gammeltoft & Kjerulf-Jensen, 1943)
or not at all (Levine & Huddlestun, 1947). Saturation with glycine, alanine
or glutamic acid blocks the reabsorption of creatine (Pitts, 1943, 1944).
Other pairs of substances with mutual inhibition of reabsorption are leucine
146 SECRETION AND TRANSPORT OF NON-ELECTROLYTES
and isoleucine, arginine and lysine (Beyer, Wright, Skeggs, Russo &
Shaner, 1947). The reabsorption of aminoacids, however, was not affected
by saturation with glucose.
Thus several groups of substances appear to be transported by separate
mechanisms, members of one group competing with each other, but not
with members of other groups. According to Beyer et aL (1946, 1947) and
to Pitts three groups of amino-acids may be distinguished. The first group
comprises the basic amino-acids arginine, histidine and lysine, the second
leucine and isoleucine and the third glycine, alanine, glutamic acid and
creatine. The sugars glucose and xylose form one group, possibly loosely
related to another group, containing fructose and galactose.
Such a loose connexion between groups has been interpreted on the
assumption that certain features of the transport mechanisms are shared
by two groups, others not. Gammeltoft & Kjerulf- Jensen visualized the
common feature of fructose and galactose on the one hand and of glucose
and xylose on the other as either a common phosphate donor or a common
source of energy.
In the case of tubular secretion likewise competition phenomena are
common. Tm of phenol red is lowered by diodrast as well as by hippuran,
whereas phenol red is much less effective in inhibiting the secretion of
diodrast and hippuran, indicating transport by a common mechanism, but
with differences in affinity.
The secretion of a number of organic bases, including methyl-nicotin-
amide, is not inhibited by diodrast or />-amino-hippuric acid. Thus at least
two different transport mechanisms appear to exist with respect to tubular
secretion.
(4) Enzyme inhibitors
The absorption of glucose from the intestine as well as from the kidney
tubules is depressed by phlorizine (Nakazawa, 1922; Lundsgaard, 19330;
Wertheimer, 1933; Walker & Hudson, 1937), in the case of the kidney
completely. Phlorizine, however, was also found to inhibit the tubular
excretion of phenol red in man (Chasis, Ranges, Goldring & Smith, 1938)
and in the chicken (Pitts, 1938), as well as that of diodrast in man (Chasis
et aL 1938). Cyanide inhibits the absorption of glucose (Kjerulf-Jensen &
Lundsgaard 1940). The reabsorption of sugars in the kidney was found by
Hober (1933) to be depressed by phenyl urethane. lodoacetic acid inhibits
glucose absorption both from the intestine (Wilbrandt & Laszt, 1933) and
from the tubule (Walker & Hudson, 1937), as well as secretion of phenol
red in the chicken mesonephros (Beck & Chambers, 1935).
The assumption that enzymatic reactions are somehow involved in active
transport and that depression of a transport by enzyme inhibitors may be
SECRETION AND TRANSPORT OF NON-ELECTROLYTES 147
taken as a criterion for the active nature of the transport has rather generally
been made, and, in fact, can hardly be doubted.
The question, however, arises in each case and has much been discussed
as to whether an action of an inhibitor has to be interpreted, roughly
speaking, as blocking the machine itself or the burning of the fuel, i.e.
whether it is specifically related to a reaction involved in the transport
mechanism of the substance in question or whether it inhibits energy-
supplying reactions that could be used for other transports likewise or even
for any kind of work done by the cell. Inhibitors affecting the transport
directly may be termed primary inhibitors, those acting on the energy
metabolism secondary inhibitors.
Enzyme inhibitors of high specificity are not among those listed above.
The majority would clearly be classed as secondary inhibitors. For a time
there seemed to be evidence for a primary action of phlorizine on phos-
phorylation processes involved in glucose transport. Lundsgaard (19336)
showed that the uptake of inorganic phosphate by intestinal mucosa was
inhibited by phlorizine. This, however, may have been due to the inhibition
of phosphorylase. If there is a primary inhibition by phlorizine it would
appear to be more likely an inhibition of phosphatase. But the inhibition of
various secretion processes in the kidney tubule mentioned above, as well
as the inhibition of dehydrogenases by phlorizine shown by Shapiro (1940,
1947), would be equally reconcilable with a secondary action, as this author
pointed out. One point in favour of a primary action appears to be the
effect on the glucose transport through the red cell membrane which will
be referred to later, because in this case an 'uphill' transport has not been
demonstrated and a requirement of energy-supplying reactions seems
doubtful.
Thus the particular role of enzymes in the transports considered here
has not been elucidated to a considerable extent by the study of enzyme
inhibitors so far. This situation may change, however, if effective inhibitors
of higher specificity should be found.
III. THE TRANSPORT MECHANISM
In the mechanism of transport enzymatic reactions involving the transport
substrate may be assumed which somehow affect the overall rate of diffusion.
This linking to diffusion rate has been visualized repeatedly in the form of
a combination with a carrier to form a substrate carrier complex, by which
either diffusion in the direction of the transport is enhanced or back-
diffusion is inhibited. Mainly two such mechanisms have been suggested
which may be called the cytoplasm-carrier mechanism and the membrane-
carrier mechanism.
SECRETION AND TRANSPORT OF NON-ELECTROLYTES
(i) The cytoplasm-carrier mechanism
If the substrate S after passage of the membrane by diffusion combines
with a carrier C to form a complex CS, under certain conditions the rate
if diffusion may be increased. Such a reaction was first suggested by
Hober in 1899 to account for the relatively rapid sugar absorption from the
intestine. Verzar (1931) proposed the same principle. The acceleration was
ascribed to the maintenance of a steep concentration gradient across the
membrane due to elimination of the substrate beyond the membrane by
the reaction assumed. As Danielli (1943) rightly pointed out, this accelera-
tion is only possible if CS does not (or only much slower than S) penetrate
the membrane.
Cell
Fig. 4. Schematic picture of the cytoplasm carrier mechanism. C = carrier ; S = substrate ;
M- membrane; Cy — cytoplasm. I and II: sites of the chemical reactions assumed.
Below : gradients for the substrate.
Later a second reaction releasing the substrate again at the other end of
the cell was introduced in the mechanism suggested by Shannon & Fisher
(1938) and by Kalckar (1937) and others. A schematic picture of the
mechanism is given in Fig. 4. According to this view the substrate S would
pass the first cell membrane in free form, the cytoplasm in the form of CS
and the second membrane again in the form of S. Drabkin (1948) assumed
hexokinase to be enzyme I and phosphatase to be enzyme II.
In principle this mechanism has been widely accepted and used to
explain a series of observations. The limited amount of C (whose concentra-
tion would be the extensive factor of the scheme) gives a basis for the inter-
pretation of maximum rates as well as of competition. The assumption of
different dissociation constants of the complex CS was successfully used to
SECRETION AND TRANSPORT OF NON-ELECTROLYTES 149
explain differences of Tm values for various compounds and special features
of competition, particularly the fact that competition may be asymmetric
in the sense that A competes more powerfully with B than vice versa (e.g.
phenol red and diodrast, as mentioned above). The complex constant would
be the intensive factor of the scheme.
There are, however, a number of difficulties for the acceptance of the
particular assumption that the carrier substrate complex is formed inside
the cell.
For the discussion of the potentialities in this scheme it seems useful to
introduce a value for the permeability of the membrane. Since, irrespective
of factors like lipoid solubility and porous structure of the membrane, the
rate of diffusion will always be proportional to the concentration difference
across the membrane, this may be done in the form of a * diffusion-equivalent
thickness' of the membrane, de — the thickness of a water layer through
which transport by free diffusion would occur with the same rate as it
actually does through the membrane (assuming equal concentration
difference). If the permeability constant P is expressed in cm. /sec. and the
diffusion constant D in cm.2/sec., de in cm. will be given by Z)/P.
If the dimensions of membrane and cytoplasm are chosen according to
their values of de, the concentration gradients in the steady state will be
parallel as shown in Fig. 5.
Three cases have been represented in this figure :
A, 'Downhill transport* with small (AJ and with large (A2) values of de,
B, transport between equal concentrations, and
C, 'uphill transport*.
The reactions assumed are termed I (formation of CS) and II (releasing
of S from CS).
Under the conditions of A the reactions merely serve to accelerate
diffusion as pictured by Hober and by Verzar. The acceleration ratio has a
maximum value of 2 if the concentration of S by the reaction I is kept near
zero (as assumed in Fig. 5).
Under the conditions of B and C the transport would not occur spon-
taneously without reactions I and II (acceleration ratio — oo or negative).
The presupposition for both B and C is a sufficiently high free energy of
reactions I and II. The gradients in Fig. 5 have been drawn for high values
of AF, such that the concentrations of S and CS are kept near zero by I and
II respectively. This results in maximum steepness of the gradients. If the
free energies are lower, these concentrations will rise, the gradients become
less steep and the rate of transport will decrease.
The maximum rate will be given by the maximum slope of the gradients
(as in Fig. 4), i.e. it will be determined by the substrate concentration Sl
SECRETION AND TRANSPORT OF NON-ELECTROLYTES
before passage of the first membrane, divided by de of the membrane. This
means proportionality between Sl and rate of transport for low concentra-
tions. At high concentrations the rate will be limited by the concentration
and turn-over numbers of the enzymes for I and II. It will, then, attain a
constant level in the cases B and C, but not in the case A because of con-
tinuing diffusion of free 5.
This difference between conditions A and B or C might be used to
explain the observation that a constant maximal rate of transport was
experimentally reached in general in the kidney tubules (where due to the
Cy
CS
„«.-•*
^_--
„-•*-*
Mem- Cyto- Mem-
brane plasm brane
Fig. 5. Gradients for the substrate in the cytoplasm carrier mechanism, taking the
thickness of the membrane as its diffusion equivalent value. (Dashed lines: gradients
when reactions I and II are blocked).
filtration process the conditions of B hold), but not in the intestine (where
the experimentally used concentrations in the lumen usually were con-
siderably higher than the plasma concentrations according to condition A).
It is, however, at variance with observations on the aglomerular kidney,
where the condition A prevails (for high concentrations) and nevertheless
a definite maximal rate of transport for phenol red was observed.
Being given that the rate of transport at low concentrations is limited by
the permeability of the membrane (i/</e), a maximum value for de may be
derived from the amount of glucose reabsorbed in man per unit time, T.
It was shown above that the amount transported through unit area (cm.2)
in the tubule may be approximately estimated to be 0-21 x io~8 mol./sec.
SECRETION AND TRANSPORT OF NON-ELECTROLYTES 151
This rate, according to Fig. 2, is still held at a glomerular concentration
of 20omg.%, i.e. an average tubular concentration of ioomg.% or
0-006 molar = 6x io~6 mol. /ml. Thus assuming penetration through the
membrane by diffusion of free glucose, a minimum permeability constant
P=3«5 x io~4 cm./sec. would result. (This is a minimal value, since it only
holds for zero concentration within the cell.) The diffusion constant D
being about io~10 cm.2/sec., for de a maximum value of 27 A. would be
arrived at. Since the thickness of the cell membrane is estimated to be about
1 00-200 A., this result would mean that glucose penetrates the membrane
4-8 times faster than in free diffusion which, of course, is absurd. If,
furthermore, the thickness of the brush border (about 20,000 A.) is assumed
to add to the value of de, which seems a reasonable conclusion, the dis-
crepancy becomes even 100-200 times greater.
In addition, it should be recalled here that as discussed above in general
the substances transported actively are not substances which penetrate
easily.
Inhibition of the reactions I and II would have different effects according
to the direction and value of the concentration difference. In the case A
the transport would not be blocked but slowed down, in the case B it would
be blocked, in the case C reversed. Again this would appear in harmony
with some observations, but not with others. In the intestine (condition A),
phlorizine in general only diminished the rate of absorption of sugars, in
the kidney (condition B or C) it seems to block it completely. In the double
perfusion experiments of Hober (1933), however, on the frog kidney
application of phlorizine and glucose to the tubules but not to the glomeruli
never caused glucose to appear in the urine, as it should be expected, if the
tubule cell membranes were permeable to glucose.
Thus serious difficulties arise when the implication of the cytoplasm-
carrier mechanism are considered in detail and quantitatively. This was
one of the reasons for the suggestion of the membrane-carrier mechanism
by Rosenberg & Wilbrandt (1952). In this mechanism, which shares the
useful possibilities of interpretation mentioned above with the cytoplasm-
carrier mechanism, the features just discussed offer no difficulties.
(2) The membrane-carrier mechanism
In this mechanism, which is represented in Fig. 6, the carrier substrate
complex is assumed to be formed before the passage of the first membrane,
and, in fact, to be the pre-condition of this passage, the membrane being
assumed to be practically speaking impermeable to the substrate. The
complex thus would differ from the substrate in the -ability to penetrate
the membrane, possibly due to factors like lipoid solubility.
152 SECRETION AND TRANSPORT OF NON-ELECTROLYTES
The difficulties just discussed are avoided by this interpretation.
If the membrane is impermeable to the substrate itself, blocking of the
mechanism will stop the transport completely, independent of the concentra-
tion conditions. This is in accordance not only with the observations of
Hober mentioned above, but also with data on red cells to be discussed later.
The rate of transport will be determined by the diffusion rate of the
complex rather than of the substrate. Thus the assumption of impossibly
high permeability constants will not be necessary, since the concentration
of the complex may be raised by reaction I (now at the outer surface of the
membrane) to high levels, if its free energy is sufficient.
Membrane
Enzyme
Enzyme
Substrate
Carrier
Substrate
S C
sc
.>;
Fig. 6. Scheme of the membrane carrier mechanism. The enzymes sited on the two
surfaces of the membrane catalyse the substrate carrier reactions shown in the middle and
thus build up the gradients shown below.
Finally, the maximum rate, limited by saturation of the carrier (or of the
enzyme catalysing the substrate-carrier reaction), will under all conditions
be independent of the substrate concentration, since no diffusion except
that of the complex occurs in the membrane.
A further point in which the membrane-carrier mechanism appears
superior to the cytoplasm mechanism is the efficiency of the system with
respect to back-diffusion. In the cytoplasm-carrier scheme back-diffusion
through the first membrane can be avoided, if reaction I is sufficiently rapid
and if C5, as mentioned above, does not penetrate the membrane. Back-
diffusion of S through the cytoplasm from the site of reaction II to that of
reaction I, however, appears to be inevitable. Reaction I, thus, not only will
have the task of removing S coming through the membrane in the direction
of the transport, but also S coming back from the site of reaction II. The
SECRETION AND TRANSPORT OF NON-ELECTROLYTES 153
ratio of these fractions which may be called * back-diffusion efficiency*
depends on both de and accumulation ratio. If the accumulation ratio is a,
and the ratio of the length / of the cell to the equivalent membrane thickness
ljde is 6, the back-diffusion efficiency, e, can be shown to be
Remembering that actively transported substances in general penetrate cell
membranes slowly, we may take de of a poorly penetrating substance, for
instance, glycerol in the ox erythrocyte, for which P, according to Jacobs
(1934), is 0-002 x io~5 cm./sec. Taking for D 0-812 x io~10 cm.2/sec. we
obtain a value for de of 40-6^. If the length of the cell is taken as ao//, b will
be 0*5. For an accumulation ratio of 100, which is not infrequent, the
efficiency then becomes 0-5/101-5, i.e. less than |%.
Another reason for preferring the interpretation of the membrane-carrier
mechanism, however, was the fact that with the exception of uphill trans-
port, all the features of active transport discussed above could be demon-
strated in the case of glucose transport through the red cell membrane.
These results, which were obtained in collaboration with Rosenberg
(many of them also independently by LeFevre (1947, 1948) and LeFevre &
Davies (1951)) will be discussed in the following section.
IV. THE TRANSPORT OF SUGARS ACROSS THE
RED CELL MEMBRANE
The membrane of red cells in most species has a very low permeability for
monosaccharides; in various species it is practically speaking impermeable.
The erythrocyte of man and apes, however, as was early shown by Kozawa
(1914), are highly permeable.
The transport leads to equalization of the concentration. Up to now
accumulation has not been demonstrated in a conclusive way.
In other respects, however, striking similarities were found to the trans-
ports in kidney and intestine.
Phlorizine was found to inhibit the penetration of glucose, fructose,
D-xylose and L-arabinose (LeFevre, 1947; Wilbrandt, 1947, 1953). lodo-
acetic acid, however, in contrast to its inhibition of the * uphill' absorption
in the intestine showed no effect (Wilbrandt, Guensberg & Lauener, 1947).
This would seem to indicate that phlorizine actually is an inhibitor of the
primary type, iodoacetate a secondary one.
The inhibition is also shown by phloretin, the aglucone of phlorizine, as
well as by various phosphate esters of phloretin which will be referred to
later (Wilbrandt & Rosenberg, 1950).
154 SECRETION AND TRANSPORT OF NON-ELECTROLYTES
In view of the fact that hexokinase has been named as the first enzyme in
the mechanism of intestinal and tubular absorption of glucose (Drabkin,
1948; Hele, 1950), lachrymators were tested whose hexokinase-inhibiting
action has been shown by Dixon (1948). It was found that allylisothio-
cyanate, bromacetophenone and chloropicrine block glucose transport
55
so
45
40
35
30
25
20
15
10
2
15
10
05
. Observed entrance fructose, 42°CM V 513
• SOisot.
• 30 iso'
Diffusion calculated
« 06isot
• 04isot.
02isot
0 10 20 30 40 50 60 70 mm
Observed glucose, V. 222
1-5
10
05
Enzymatic calculated
100
200
i t ,
8
Fig. 7. Observed entrance of glucose and of fructose into human red cells from varied
external concentrations as compared to the course of penetration calculated on the basis
of the E. kinetics (see p. 157).
completely. ATP, however, and magnesium never accelerated but rather
showed slight inhibition. The penetration of glycerol was not affected by
lachrymators (Wilbrandt & Rosenberg, 1951 ; Wilbrandt, 1953).
Corticosteroids which lower Tm for glucose in the human kidney were
found to decrease the rate of transport (Wilbrandt, 1953).
Other inhibitors were heavy metals (LeFevre, 1948; Wilbrandt &
Rosenberg, 1950), gold and mercury as well as chloromercuribenzoate
(LeFevre, 1948), pointing to a probable role of SH groups.
SECRETION AND TRANSPORT OF NON-ELECTROLYTES 155
No effect was observed with azide, dinitriphenol, atabrine, thyroxine,
whose inhibiting or enhancing action on other transports most likely will
be of the secondary type.
Insulin showed, if any, a slight inhibitory effect.
The simultaneous penetration of D-xylose, L-arabinose and glucose was
much slower than the sum of the individual transport rates in the same con-
centrations, indicating competition for a common transport mechanism
(Wilbrandt, 1950).
Glucose exit, V. 529
60
90
, Fructose exit, V. 530
120 min
30
60
90
120 min.
Fig. 8. Observed exit of gucose and of fructose from human red cells. S — amount of
sugar in the cell in cell units (normal cell volume x isotonicity = i).
Competition between various monosaccharides successively added has
been investigated extensively by LeFevre & Davies (1951). They showed
that the inhibition may be unilateral, e.g. the uptake of laevulose is strongly
inhibited by glucose but not vice versa. Our own results agree closely with
this observation. The authors also pointed out that certain differences in
kinetics appear to parallel the behaviour with respect to mutual inhibition.
The nature of these differences will be referred to later.
156 SECRETION AND TRANSPORT OF NON-ELECTROLYTES
The dependence of the transport rate on the sugar concentration revealed
unusual features, when studied in experiments involving both entrance
and exit of glucose. For the sake of clarity the side where glucose enters
the membrane was termed cis, that where glucose leaves the membrane
trans.
With rising cis concentrations of glucose the rate increased initially and
finally reached a constant level. This level, however, depended strongly on
the trans concentration, even (or rather particular strongly so) in very low
ranges of concentration (where the slope of a diffusion gradient would
hardly be affected). Figs. 7 and 8 show such experiments.
This surprising feature led to an attempt to calculate the kinetics of an
enzymatic membrane-carrier transport (Wilbrandt & Rosenberg, 1951).
It was assumed that a first enzyme on the cis side catalyses the substrate-
carrier reaction, that the complex diffuses across the membrane and that on
the trans side it dissociates, catalysed by a second enzyme. Using the
Michaelis-Menten equation for the rates of the enzymatic reactions and
assuming a steady state, an equation for the rate of transport was arrived at.
It can be resolved into separate terms, if instead of the rate of transport A
its reciprocal i/A is used (representing some sort of penetration resistance) :
'J^OV
A TTi I T^/ O O \ TS" * TS't (
(O (2)
n i _ Kt\
u + S,- SL1 A Kj '
M\Sl- StlKKt Si- S
(3) (4) (5)
where A = rate of transport,
Sl = cis concentration of substrate,
5H = trans concentration of substrate,
Ctot = total concentration of carrier (free and bound),
D = diffusion constant of the carrier (assumed to be equal
to that of the carrier-substrate complex),
M =Cto,A
E = concentration of enzyme (assumed to be equal cis and
trans),
K — Michaelis-Menten constant (likewise assumed equal
cis and trans),
K± and K2 = velocity constants of the enzymatic reaction and back-
reaction between substrate and carrier.
SECRETION AND TRANSPORT OF NON-ELECTROLYTES 157
Of the terms (i), (2), (3) and (5) any one may be limiting, according to the
ratios of the various constants and concentrations, as shown in the following
table:
Limiting terms and forms of penetration kinetics
Substrate. . . K>Sit Su S^K> Su Sb Sa > J
concentration
Enzymes . . . Not EI saturated Saturated
Limiting saturated En not
term K^jK^ Carrier saturated
E,,. v (i <i Saturated Z Zx E
E limiting }2 >x Notsaturated D z\ z
n,rr v 1 3 <^1 Saturated E E E
M limiting ^ ^ Notsaturated D D D
Thus five types of kinetics result which have been termed D, Z1? Z2, Z
and E and whose characteristics are shown in Fig. 9. D is the kinetics type
of free diffusion.*
It appears that under conditions under which either the enzymes or the
carrier or both are saturated the type E will frequently be found. This,
however, is the only type harmonizing with the observations reported above
for glucose.
On the other hand, the type D, i.e. a behaviour like that of free diffusion,
may also be met, particularly if the carrier is not saturated.
This agrees well with the observation that for fructose, although its
transport is also inhibited by phlorizine and thus may be assumed to be
based on the same essential mechanism as that of glucose, the striking
dependence of the rate of transport on both cis and trans concentrations is
not observed. It seems rather to follow the D type (see Figs. 7 and 8).
Accordingly the affinity of fructose to enzyme or carrier may be assumed to
be lower than that of glucose.
This difference in affinity as concluded from the kinetics, on the other
hand, is in good harmony with the observation reported above that glucose
saturation depressed fructose transport much more than vice versa.
The course of penetration for the E kinetics has been calculated both for
entrance and exit and showed close resemblance to the observed course for
glucose. In Fig. 7 calculated and observed curves for entrance are
represented.
In the calculation of equation (5) for the sake of simplicity the assumption
was made that enzymes I or II have the same Michaelis constant, or that
they are identical. This, however, is not a necessary requirement. In fact,
experiments with phloretin phosphate strongly indicate that in the red cell
the enzymes are different.
* Z2 only results if either the enzyme concentrations or the velocity constants on the
cis and the trans side are different (which was not assumed here). The general equation
for this case was given previously (Wilbrandt & Rosenberg, 1951).
158 SECRETION AND TRANSPORT OF NON-ELECTROLYTES
D
10
A-K(Si-S.i)
10
20 $,
10
0-2
30
M-ff)
40 S,
10 20 30
For legend see p. 159.
HO Si
SECRETION AND TRANSPORT OF NON-ELECTROLYTES 159
S,-0-08
Fig. 9. Types of penetration kinetics resulting from equation 5, when certain
terms become limiting (see text).
The phloretin phosphate preparations used were highly polymerized,
and a 32P labelled specimen could be shown not to penetrate the red cells
to any appreciable extent. Thus the action of this inhibitor will be restricted
to the outer surface of the membrane.
The inhibition by these phloretin esters, in striking contrast to most of
the other inhibitors tested, showed a most definite preference for the
glucose exit, as compared to the entrance. In some experiments, under
strictly parallel conditions, the entrance seems not to be inhibited at all,
whereas the exit was blocked for a long time.
l6o SECRETION AND TRANSPORT OF NON-ELECTROLYTES
Phlorizine showed qualitatively similar behaviour; inhibition of glucose
entrance, however, increased with time, most likely indicating slow penetra-
tion of phlorizine.
Several conclusions may be drawn from these results: there must be at
least two enzymes on the outer surface of the membrane, only one of which
is inhibited by phloretin phosphate, and the inhibited enzyme must be
used preferentially for outward transport, the other one for inward trans-
port. The kinetics of inhibition of the transport by various inhibitor types
(both competitive and non-competitive) has now been worked out. Marked
kinetical asymmetries may occur, even with identical enzymes cis and trans
and varying with the type of inhibitor. Thus the question whether different
enzymes cis and trans have to be assumed must be left open at present.
If some sort of a phosphorylation-dephosphorylation mechanism is in-
volved, as has been assumed for the epithelial transports, it must be the
dephosphorylation stage that is inhibited by the phlorizine group, as
suggested above.
This would appear in harmony with the results of Ellinger & Lambrechts
(1937) on various azo derivatives of phlorizine. Only those compounds
inhibited glucose reabsorption that could be shown to penetrate the tubule
cells.
Such a phosphorylation mechanism, however, cannot be as simple as
has been assumed. Glucose- 1 -phosphate, as well as glucose-6-phosphate
and fructose diphosphate, have been tested and found not to penetrate the
red cell membrane appreciably. Formation of a metaphosphate ester has
been suggested based on several theoretical considerations (Rosenberg,
1948, 1950), but no way of experimental test has yet been found.
SUMMARY
A survey is given of active transports of organic molecules through the
epithelia of intestine and kidney.
It is shown that most of the substances actively transported either are
known to or may be assumed to penetrate cell membranes only very slowly,
and that a polar-non-polar structure of the molecules appears to favour
active transport in the case of secretion.
Examples are given for the following characteristic features of these
transports :
(1) The majority of them occur or may occur against a gradient of
electrochemical potential, requiring coupling with energy supplying
reactions.
(2) Enzyme inhibitors affecting the transports may be classed as primary
inhibitors acting on reactions involved in the transport mechanism itself
SECRETION AND TRANSPORT OF NON-ELECTROLYTES l6l
and secondary inhibitors, affecting energy-supplying reactions. Few if any
primary inhibitors have become known.
(3) Competition for the transport mechanism among molecules of like
structure is frequent. In the kidney several groups of such potentially
competing molecules have become known.
(4) The rate of transport shows in general a non-linear relationship to
the concentration, in contrast to diffusion. In many cases with high con-
centrations a maximum constant level of the rate is attained.
Two carrier mechanisms that have been suggested are compared, one of
them assuming the formation of a substrate carrier complex after, the other
one before passing of the first cell membrane. The terms cytoplasm-carrier
mechanism and membrane-carrier mechanism are suggested.
Arguments against the cytoplasm-carrier mechanism are considered
which are based on consequences of the assumed free permeability of the
cell membrane for the substrate : the general character of molecules actively
transported, the facts that inhibition of the mechanism does not seem to
lead to leakage, that to account for the maximal reabsorption of glucose in
the human kidney, impossibly high permeability constants must be assumed
and that due to back-diffusion of substrate in the cytoplasm the efficiency
of the mechanism at high accumulation ratios would be very low.
Furthermore, the assumption of the membrane-carrier mechanism seems
to be favoured by the fact that most of the above-mentioned features of
active transport except the * uphill ' shift are found in the transport of sugars
across the red cell membrane; depression by enzyme inhibitors like
phlorizine and numerous others, competition among different sugars, non-
linear rate-concentration dependency with a constant maximal rate level
at high concentrations. The kinetics of a membrane-carrier mechanism
have been calculated and found to agree satisfactorily with the observed
data, particularly with respect to a striking dependence of the rate of
transport on small concentrations of glucose on the trans side of the
membrane.
REFERENCES
BARANY, E. H. & SPERBER, E. (1939). Skand. Arch. Physiol. 81, 290.
BARANY, E. H. & SPERBER, E. (1942). Ark. Zool. 34 A, no. i.
BECK, L. V. & CHAMBERS, R. (1935). J. Cell. Comp. Physiol. 6, 441.
BEYER, K. H., Russo, H. F., PATCH, E. A, & MILLER, A. K. (1946). Amer.J. Med.
Sci. 213, 246.
BEYER, K. H., WRIGHT, L. D., SKEGGS, H. R., Russo, H. F. & SHAKER, G. A. (1947).
Amer. J. Physiol. 151, 202.
CHASIS, H., RANGES, H. A., GOLDRING, W. & SMITH, H. W. (1938). J. Clin. Invest.
17, 683.
CORI, C. F. (1925). J. Biol. Chem. 66, 691.
CORI, C. F. (1926). Proc. Soc. Exp. Biol., N. Y., 24, 125.
l62 SECRETION AND TRANSPORT OF NON-ELECTROLYTES
DANIELLI, J. F. (1943). In Davson, H. & Danielli, J. F. (1943). The Permeability
of Natural Membranes. Cambridge University Press.
DIXON, M. (1948). Biochem. Soc. Symp. no. 2. Cambridge University Press.
DONH6FFER, S. (1935). Arch. exp. Path. Pharmak. 177, 689.
DRABKIN, D. L. (1948). Proc. Amer. Diabetes Ass. 8, 171.
ELLINGER, P. & LAMBRECHTS, A. (1937). C-R> S°c- BioL, Paris, 124, 261 .
GAMMELTOFT, A. & KJERULF- JENSEN, K. (1943). Actaphysiol. scand. 6, 368.
HELE, M. P. (1950). Nature, Lond., 166, 786.
HOBER, R. (1899). Pfliig. Arch. ges. Physiol. 74, 246.
HOBER, R. (1933). Pfliig. Arch. ges. Physiol. 233, 181.
H6BER, R. (1940). Cold Spr. Harb. Symp. Quant. BioL 8, 40.
HOBER, R. (1945). Physical Chemistry of Cells and Tissues. Philadelphia.
HOBER, R. & BRISCOE-WOOLLEY, M. P. (1940). J. Cell. Comp. Physiol. 15, 35.
H6BER, R. & HOBER, J. (1937). J. Cell. Comp. Physiol. 10, 401.
JACOBS, M. H. (1934). J. Cell. Comp. Physiol. 4, 161.
KALCKAR, H. M. (1937). Enzymologia, 2, 47.
KJERULF-JENSEN, K. & LUNDSGAARD, E. (1940). Hoppe-Seyl. Z. 266, 217.
KOZAWA, S. (1914). Biochem. Z. 60, 231.
LEFEVRE, P. G. (1947). BioL Bull., Woods Hole, 93, 224.
LEFEVRE, P. G. (1948). J. Gen. Physiol. 31, 505.
LEFEVRE, P. G. & DAVIES, R. I. (1951). J. Gen. Physiol. 34, 515.
LEVINE, R. & HUDDLESTUN, B. (1947). Fed. Proc. 6, 151.
LUNDSGAARD, E. (1933 a). Biochem. Z. 264, 221.
LUNDSGAARD, E, (19336). Biochem. Z. 264, 209.
MARSHALL, E. K. & CRANE, MARIAN M. (1924). Amer.J. Physiol. 70, 465.
NAKAZAWA, F. (1922). TohokuJ. Exp. Med. 3, 288.
PITTS, R. F. (1938). J. Cell. Comp. Physiol. n, 99.
PITTS, R. F. (1943). Amer.J. Physiol. 140, 156.
PITTS, R. F. (1944). Amer. J. Physiol. 140, 535.
ROSENBERG, TH. (1948). Ada chem. scand. 2, 14.
ROSENBERG, TH. (1950). Reports Steno Memorial Hosp. and Nordisk Insulin Lab. 5,
52.
ROSENBERG, TH. & WILBRANDT, W. (1952). Int. Rev. Cytol. i, 65.
SHANNON, J. A. & FISHER, S. (1938). Amer.J. Physiol. 122, 765.
SHAPIRO, B. (1940). Dissertation, Univ. Jerusalem.
SHAPIRO, B. (1947). Biochem. J. 41, 151.
SMITH, H. W. (1951). The Kidney. New York.
SPERBER, I. (1947). Proc. ijth Int. Physiol. Congr. p. 217.
VERZAR, F. (1931). Ergebn. Physiol. 32, 391.
VERZAR, F. (1935). Biochem. Z. 276, 17.
VIDAL-SIVILLA, S. (1950). Rev. esp. Fisiol. 4, 131.
WALKER, A. M. & HUDSON, C. L. (1937). Amer.J. Physiol. 118, 130.
WERTHEIMER, E. (1933). Pfliig. Arch. ges. Physiol. 233, 514.
WILBRANDT, W. (1938). Pfliig. Arch. ges. Physiol. 241, 302.
WILBRANDT, W. {i947)« Helv. physiol. acta, 5, C64.
WILBRANDT, W. (1950). Arch. exp. Path. Pharmak. 212, 9.
WILBRANDT, W. (1953). Unpublished results.
WILBRANDT, W., GUENSBERG, E. & LAUENER, H. (1947). Helv. physiol. acta, 5, C20.
WILBRANDT, W. & LASZT, L. (1933). Biochem. Z. 259, 398.
WILBRANDT, W. & ROSENBERG, TH. (1950). Helv. physiol. acta, 8, C82.
WILBRANDT, W. & ROSENBERG, TH. (1951). Helv. physiol. acta, 9, C86.
COMMENT ON PROFESSOR WILBRANDT'S
AND DR LEFEVRE'S PAPERS
W. F. Widdas, London, said:
The general findings of Prof. Wilbrandt and of Dr LeFevre have been
confirmed in experiments carried out on human erythrocytes and also on
erythrocytes of a number of mammalian species (in which the cells from the
blood of foetal and newborn animals have been shown to be permeable to
glucose and other sugars).
In interpreting the results obtained, kinetics have been used which were
based on a membrane-carrier mechanism such as Prof. Wilbrandt has
described but in which it was postulated that the rate-determining step was
the relatively slow movement of carriers and complexes between one inter-
face and the other. It was assumed that the carriers remained in the inter-
face sufficiently long to achieve, on average, adsorption equilibrium with
glucose in the respective solutions.
It follows that the net transfer of such a system would be proportional
to the difference in the fraction of carriers saturated with glucose at the two
sides. These fractions can be represented by relationships of the Michaelis-
Menten type. It can be shown that these kinetics reduce to a diffusion-type
equation when the equilibrium constant is high, but when the equilibrium
constant is low, relative to the concentrations used, the best approximation
is to take the transfer rate as proportional to the difference in the reciprocals
of the concentrations :
Transfer rate oc (7^ — 7^,
8->Sl W L
The diffusion equation applied to the red cell problem gives rise to an
expression of the form kt = \F (C V}] (i)
whereas the second approximation gives an expression of the form
fo't — CZ rp' (Ty J/YI (2)
The terms in square brackets are not identical but are of the same order,
and if one uses the diffusion equation to analyse data of swelling in glucose
solutions the theory predicts that
or log k = constant — 2 log C,
that is, a plot of log k against log C should have a slope of — 2.
164
My results and those of Wilbrandt, Guensberg & Lauener (1947) agree
well with this prediction (see Fig. i).
1-0 2-0
loglOOCs
3-0
Fig. i. Variation of apparent penetration constant (based on diffusion) with glucose
concentration. Experimental results of Wilbrandt et al. showing plot of log 100 k against
log 100 C. Points O and • refer to results obtained by a direct and indirect photoelectric
method respectively. Taking © as reference, points x represent calculated values;
the slope of the line is approximately — 2.
The kinetics have also been extended to cover competition to yield a
method of determining the equilibrium constant of the carriers reacting
with glucose (Widdas, 1953). The value reported agrees well with that
obtained by Dr LeFevre by the phloretin method.
REFERENCES
WIDDAS, W. F. (1953). 3- Physiol. 120, 23 P
WILBRANDT, W., GUENSBERG, E. & LAUENER, H. (194?)- Helv. physiol. acta, 5,
ENZYME SYSTEMS OF THE CELL SURFACE
INVOLVED IN THE UPTAKE OF SUGARS
BY YEAST*
BY ASER ROTHSTEIN
Pharmacology Division, Department of Radiation Biology, University
of Rochester School of Medicine and Dentistry, Rochester, N.Y.
I. INTRODUCTION
The hexoses constitute one of the most important sources of carbohydrate
for heterotrophic organisms such as Protozoa, bacteria and fungi, as well as
for most cells of multicellular animals. The enzymic systems by which
sugars are assimilated and degraded have been worked out in considerable
detail in cell-free systems, but a number of important questions remain
unanswered when the uptake and metabolism of sugars by intact cells is
considered — questions related to the architecture and organization of the
cell. In recent years, with the development of techniques for isolating
certain of the cellular structures such as mitochondria, chloroplasts, the
nucleus and various granules of the cytoplasm, it has been clear that certain
metabolic functions are localized in specific centres in the cell. In regard
to sugar uptake, certain questions are pertinent. Where, in the cell, are the
enzymes located which metabolize sugars? How does the sugar pass from
the environment to the enzyme centre? All substances passing into the cell
must pass through the cell membrane. What is the role of this structure in
sugar uptake? In the past, it had generally been supposed that the cell
membrane could regulate the rate of movement of glucose into the interior
of the cell, the regulatory mechanism consisting of a resistance to the
diffusion of glucose, which was defined in terms of a permeability constant.
More recently there has been increasing evidence of participation of the
membrane in an active transport of glucose into the cell, independent of the
concentration gradient, and with the necessary energy supplied by metabolic
reactions.
The permeability of cellular membranes to glucose has been the subject
of considerable study since the classical work of Overton, Hedin and others
in the late nineteenth century. There are a number of excellent reviews of
this material (Brooks & Brooks, 1941; Davson & Danielli, 1943; Hober,
* This paper is based on work performed under contract with the United States
Atomic Energy Commission at the University of Rochester Atomic Energy Project,
Rochester, N.Y.
l66 ENZYME SYSTEMS OF THE CELL SURFACE INVOLVED
1945; Heilbrunn, 1952). With the exception of the red blood cell of
primates, all have been characterized as having a relatively low permeability
to glucose and other sugars. The red blood cell of primates is a special case
involving a mechanism by which glucose is actively transported across
the membrane (LeFevre, this volume, and Wilbrandt, this volume).
Permeability studies have been largely restricted to cells which do not
actively metabolize glucose so that the osmotic equilibria could be deter-
mined. Whether the data obtained can be applied as an estimate of the
permeability of membranes of cells which actively metabolize glucose is
of course open to question. 0rskov (1945) studied the permeability of
yeast to a variety of substances by measuring volume changes in the cells
by optical or haematocrit measurements. According to his calculations, the
cells have a relatively low but definite permeability to glucose, galactose,
xylose and arabinose. However, the technique does not exclude osmotic
effects associated with metabolism of these substances, nor does it indicate
whether or not the substances are entering the cells, unaltered, by a diffusion
mechanism. Brooks (1947) has summarized attempts to estimate the per-
meability of yeast cells based on rates of metabolism of this substrate.
However, such attempts are somewhat inconclusive because of the assump-
tions that must be made in the calculations. Furthermore, the results must
be compared with data from other cells which do not metabolize glucose.
Conway ( 19500), in an interesting paper, has described the phase distri-
butions of various substances in yeast suspensions. He found that galactose
and arabinose do not distribute in the cellular water to any appreciable
extent after an hour, indicating an exceedingly low permeability of the
cellular membrane to these sugars. Using Conway's technique, it was
found that the yeast membrane is also impermeable to sorbose (Rothstein
& Meier, 1953). Since the uptake of glucose, of mannose, and of fructose
is considerable under the same experimental conditions, it must be con-
cluded that the yeast-cell membrane can discriminate between glucose,
fructose and mannose as compared with galactose, sorbose and arabinose.
Although the existing data do not unequivocally establish that the per-
meability of cell membranes is sufficiently high to account for the rate of
uptake of glucose in yeast, the specificity of the yeast membrane in terms
of its ability to pass glucose, fructose and mannose, but to reject galactose,
sorbose and arabinose, argues against a simple permeability mechanism for
the uptake of sugars.
On the other hand, evidence does suggest the existence of an active
transport system for glucose. It has already been noted that the apparently
anomalous permeability of the red blood cells of primates involves such a
mechanism (LeFevre, this volume, and Wilbrandt, this volume). The
IN THE UPTAKE OF SUGARS BY YEAST 167
movement of glucose from the lumen of the intestine across the epithelium
into the blood also involves an active transport mechanism (Hober, 1945).
In the renal tubule the glucose actually moves in the direction opposite
to that which would be dictated by the concentration gradient (Smith, 195 1).
Although the energy for these transport mechanisms is undoubtedly derived
from metabolic reactions, the exact mechanisms by which it is accomplished
are not known. The demonstration of active transport mechanisms in the
above-mentioned tissues does not necessarily permit the generalization
that other actively metabolizing tissues and cells also depend on such
mechanisms for supplies of sugar substrates.
II. AN HYPOTHESIS CONCERNING THE MECHANISM
OF UPTAKE OF GLUCOSE BY YEAST
In this section studies are presented which are concerned with sugar uptake
by yeast cells. These are not studies of permeability in the classical sense.
Glucose as it enters the cell is altered by the metabolic cycles so that osmotic
equilibrium is never attained. Therefore it is not feasible to determine con-
centration gradients, the rate of attainment of equilibrium, or the rates of
inward and outward movements of glucose. Fortunately, it has been found
that certain cations markedly influence the rate of uptake of glucose.
Furthermore, the action of these cations can be localized at the periphery
of the cell. The mechanism by which glucose proceeds through the mem-
brane can therefore be characterized in terms of the effects of these sub-
stances on rates of glucose uptake. On the basis of these studies and of
other evidence, an hypothesis concerning the mechanism of uptake of
sugars can be proposed in general terms. It will be discussed in greater
detail in the last part of this section. The hypothesis is based on the concept
that glucose is actively transported into the yeast cell. It is furthermore
suggested that this is accomplished by the presence of enzymic activity in
the peripheral layers of the cell. The enzymic activity need not be a special
mechanism for moving glucose into the cell, but may constitute the initial
phosphorylation reactions of the fermentative and respiratory pathways of
metabolism. Because of the peripheral location of the fermentative
enzymes, glucose does not have to pass into the interior of the cell, though
its metabolic products undoubtedly do so. The energy resulting from the
peripheral degradation of glucose not only accounts for the uptake of
carbohydrate by the cell, but also provides energy for the active transport
of ions such as K+, II+, Mg++, Ca++ and H2PO4~.
The hypothesis requires that enzymes be present on the surface of the
cell. Although considerable evidence has accumulated that enzymic
activity at the surface of the cell is responsible for transport of various
l68 ENZYME SYSTEMS OF THE CELL SURFACE INVOLVED
materials into the cell (Rosenberg & Wilbrandt, 1952; Danielli, 1952), the
evidence in most cases is indirect. However, a few specific enzymes have
been definitely localized on the surface of the yeast cell including invertase
(Wilkes & Palmer, 1933), maltase and lactase (Myrback & Vasseur, 1943),
trehalase (Myrback & Oertenblad, 1937), and a number of acid phosphatases
(Rothstein & Meier, 1948, 1949). In each of these cases, a single reaction
is involved. A given substrate is hydrolysed into specific products which
can often be quantitatively recovered in the medium. The specific enzymes
involved can be characterized in vivo without separation from the living
cell, in terms of the end-products of the reaction.
The surface reactions in glucose uptake cannot be so readily characterized.
The products of the initial reactions at the surface of the cell are not
recoverable in the medium. They serve as substrates for a series of reactions
leading to end-products such as CO2 and alcohol. Although the overall
rate of the entire system of reactions can be readily stated in terms either
of glucose consumption, or CO2 production, or O2 consumption, inability
to characterize the initial reactions in terms of their specific products is a
severe handicap. It has been possible, however, to characterize the surface
reactions in sugar uptake in terms of the effects of certain inhibiting and
stimulating agents, each of which can be shown to act on the cell surface.
On the basis of the results obtained, it has been possible to make informed
guesses as to the specific reactions involved in terms of the known cycles of
carbohydrate metabolism. The agents used include the following: UO2++,
Mn++, Ca++, Mg++, K+, H+, NH4+, H2PO4~.
III. URANIUM ACTS ON THE CELL SURFACE
Uranium in the form of uranyl ion has been a very useful tool with which
to explore certain properties of the surface of the yeast cell. Its usefulness
stems primarily from its ability to form complexes with a variety of bio-
chemical substances, particularly those containing phosphate or carboxyl
groups, resulting in inhibition of systems in which such groups are essential.
The metabolism of sugar by yeast is particularly sensitive to uranium
(Booy, 1940; Rothstein, Frenkel & Larrabee, 1948; Barren, Muntz &
Gasvoda, 1948).
When uranium plus glucose are added to a yeast suspension, there is no
measurable delay in the onset of the inhibiting effect, nor is there any pro-
gressive increase or decrease in the inhibiting action. Apparently equili-
brium is rapidly achieved between uranium and the cells. This was
confirmed by actual measurements of uranium binding by the cells (Roth-
stein & Larrabee, 1948). There was as much uranium binding after i min.
as after 30 min., indicating a rapid equilibration. After several hours,
IN THE UPTAKE OF SUGARS BY YEAST 169
however, an additional small increment of uranium uptake was observed
which was not accompanied by an increment in inhibition. The following
discussion is concerned with the rapid phase of uranium uptake.
The minimal rate of penetration of uranium into cells necessary to
account for the rapid phase of uptake would have to be 8-5 x io~u mol./
min./cm.2 of surface. This seems very rapid considering that the concentra-
tion gradient of uranium did not exceed i x io~5 M. The permeability
constant to account for this rate of uptake would have to be i -4 x io~7 mol./
cm.2 of surface/sec, for a concentration gradient of i mol. /I. This is about
the same magnitude as permeability of various cell membranes to water,
and considerably higher than similar constants for cations of the mono-
valent series such as potassium (Brooks, 1941). Although there is little
6 -
£ x
c_ 4
2 E 2
01 234567
Initial uranium concentration m the medium in mol /I. x 10s
Fig. i. The decrease in the uranium concentration in the medium as a function of the
initial uranium concentration and of the yeast concentration.
comparative data on rates of uptake of bivalent cations on which to base
unequivocal conclusions, the calculated minimal value for UO2++ would
seem to be inordinately high in view of its relatively high positive charge
and low mobility.
The addition of increasing concentrations of uranium to a yeast suspen-
sion results in an increased uptake of uranium approaching a maximum
value which is proportional in each case to the yeast concentration (Fig. i),
and amounts to i x io~3 mol. /I. of cells. In view of the fact that a maximum
uranium uptake is observed, and in view of the fact that the distribution
ratio between the cells and supernate is of the order of 100 to i, it is
evident that no simple distribution of uranyl ions between the medium and
the cells takes place, but rather that certain constituents of the cell are
binding the uranyl ion, and that the uranium-uptake curves represent
IJO ENZYME SYSTEMS OF THE CELL SURFACE INVOLVED
saturation of these binding sites. Among those substances which are
present in the cells which form relatively undissociated complexes with
uranyl ion are the phosphates and the carboxyl-containing compounds,
including bicarbonate, organic acids and proteins. The concentration of
orthophosphates in the cell is 1-2 x io~2 M and of bicarbonate i x lo"1.
Thus the total uranium-binding capacity of the cytoplasm is at least
i x icr1 M/l. and considerably higher if proteins and organic phosphates
are added, altogether well over 100 times the observed maximal uranium
binding. It must be concluded that uranium equilibrates with constituents
of the cell which are not in equilibrium with the cytoplasmic contents, but
which are isolated by some barrier from the total uranium-binding sub-
stances of the cytoplasm. The only part of the cell with which uranium
could combine without equilibrating with the aqueous phase of the cyto-
plasm is the membrane of the cell, and perhaps its immediately underlying
structures.
Further evidence for the peripheral action of uranium is given by the
following experiment. Cells are exposed to 2 x 10 ~5 uranium, resulting in
a 90% inhibition of glucose uptake. The addition of any uranium com-
plexing agent to the medium results in competition with the cell for the
available uranium. Consequently the binding of uranium by the cell is
reduced and the inhibition is likewise reduced. For example, the addition
of 2 x io~4 M-orthophosphate to the uranium-poisoned suspension of cells
will reduce the inhibition from 91 to 30%. In view of the fact that the
cytoplasm already contains orthophosphate in a concentration of i x io~2 M,
50 times as high as that added to the medium, the uranium must be bound
in a location in the cell which is accessible to the phosphate of the medium
rather than the phosphate of the cytoplasm. If the uranium were acting
inside the cell, the addition of relatively low concentrations of phosphate
to the medium should have little effect because much higher concentrations
of this ion are already present there. Only if the uranium were complexed
on the surface of the cell could low concentrations of phosphate achieve the
reversal of inhibition.
IV. THE CHEMICAL NATURE OF URANIUM BINDING
SITES OF THE CELL SURFACE
The binding of uranium by yeast can be characterized by a simple reversible
reaction of the form ,., y__±r IV (
The reversible nature of the binding has been established by experiments
in which uranium-complexing agents, such as phosphate or organic acids,
are added to uranium-poisoned yeast. There results a reduction in uranium
IN THE UPTAKE OF SUGARS BY YEAST
171
binding and in the inhibitory effect. Washing with very large quantities of
distilled water will also achieve the same end. Equation (i) can be tested
by putting it in the form of the mass law,
(2)
(UY)
and determining the constancy of K. Of the terms in equation (2), (U) is
the equilibrium concentration of free uranyl ion in the medium. Values for
(U) in many experiments are exceedingly low, in the range of i x io~6
to i x io~8 M. These were measured by electroplating and a-counting
techniques using natural uranium enriched with 233U which has con-
siderably higher a-activity (Rothstein, Frenkel & Larrabee, 1949). The
value for (UY), the uranium bound to the yeast cell, was determined by
subtracting (U) from the total concentration of uranium added initially.
The concentration of uranium-binding sites of the cell was estimated from
the maximal binding of uranium, assuming that each site can bind one
uranyl ion. The value for (Y) can then be calculated by subtracting (UY)
from the total number of yeast sites. Data from a typical experiment is
presented in Table i . Over a 2O-fold range of total uranium concentrations,
the calculated K from equation (2) remains relatively constant. Other
mass-law formulations based on different ratios of reactants do not give
constants. Thus equations (i) and (2) are adequate descriptions of the
binding of uranium by cell-surface sites (Rothstein et al. 1948).
Table i . Mass law constants for binding of uranium by yeast
The yeast concentration was 20 mg./ml. in each case, equivalent to a concentration of
binding sites (YT) of 2 x 10 ~5 M.
UT X 10 6 M
(£7)xio-7M
(UY)x 10 6M
(Y)X I0^6M
K X 10 7 M
0-8
0-18
078
-92
4'4
1-6
0-26
i'57
•84
3-o
2-4
0-49
2'35
77
3'7
3-2
o-75
3'i3
•69
4-1
4-0
0-86
3'9i
•61
3'5
4-8
i -08
4-69
'53
3'5
6-0
1-47
5-85
•4i
3'5
6-8
1-71
6-63
•34
3'4
8-0
2-89
7-71
•23
4-6
16-0
14-40
14-60
0'54
5'3
An attempt has been made to determine the chemical nature of the
uranium-binding sites of the cell, by comparing the properties of the yeast-
uranium complex with the properties of other uranium complexes (Roth-
stein & Meier, 1951). A large series of uranium-complexing substances
were studied. The test system contained a fixed amount of yeast, a fixed
concentration of uranium, and a variable concentration of each complexing
172
ENZYME SYSTEMS OF THE CELL SURFACE INVOLVED
agent. With no complexing agent added, 95 % of the uranium was bound
to the cells and 5 % was free in the medium. On the addition of a soluble
complexing agent, a competition was effected between the cells and the
complexing agent for the available uranium. The uranium combined with
the complexing agent remained in the supernate. Consequently, as the
concentration of complexing agent was increased, a greater percentage of
the uranium was found in the supernate. Fig. 3 shows the distribution of
uranium between cells and supernate with varying concentrations of a
number of complexing agents. In most cases a curve obtains, in accordance
with the mass law, assuming a i to i ratio of reactants. Exceptions are
0-5 1-0 1-5 20 2-5 30 3-5
Initial U-concentration in mols./l- x 10s
4-0
Fig. 2. Distribution of uranium between cells and supernate in the
presence of various concentrations of complexing agents.
orthophosphate and glycerophosphate, which form a complex with a i to 2
ratio of uranium to phosphate. The relative affinities of the various sub-
stances for uranium can be expressed in terms of the concentration of each
substance which is associated with 50% of the uranium in the medium
(C50). In the case of macromolecules, such as proteins and phosphate
polymers, it is necessary to determine the concentration of uranium-binding
sites per unit weight of material. This was accomplished with the proteins,
by an equilibrium dialysis technique, and with the phosphates, by a colori-
metric technique. For example, serum albumen has 18 sites per molecule.
Highly polymerized desoxyribonucleic acid has i site for each 4 atoms of
phosphorus.
There was a wide spectrum of stabilities among the different complexing
agents. C50's covered a 38,ooo-fold range. The weakest uranium complexors
were the polyhydroxy compounds such as fructose. Among the organic
IN THE UPTAKE OF SUGARS BY YEAST 173
acids, the monocarboxylic acids, such as acetate, formed relatively un-
stable complexes, whereas dicarboxylic acids, such as maleate, formed
more stable complexes due to chelation. Citrate, a tricarboxylic acid,
formed an even more stable complex due to multiple ring closure (Neuman,
Havill & Feldman, 1951). The most stable carboxyl complexes were
formed by the proteins. Here, accessory groups such as hydroxyl, perhaps
imidazole and phosphate, may play an important role. Probably more than
one carboxyl group is involved in binding each uranyl ion, for the number
of uranyl ions firmly bound by each protein molecule represents only about
one-fifth of the number of free carboxyl groups. Among the phosphates,
compounds with a single phosphate formed relatively unstable complexes.
As a class, compounds with multiphosphate structure, including poly-
phosphates and nucleic acids, formed the most stable complexes, with a
stability increasing with the increasing molecular weight of the material.
Of all of the classes of agents tested, only the multiphosphate compounds
gave complexes with a stability of the same order of magnitude as that of
the yeast-uranium complex. The compounds which most resembled the
cell in this respect were the metaphosphate polymers of high molecular
weight. The poly phosphates of low molecular weight, such as ATP,
showed a lower affinity for uranium by a factor of 20.
There has been a question raised concerning the validity of the mass-law
equations in heterogeneous systems, that is, in systems containing both a
solution and a solid phase (Rosenberg & Wilbrandt, 1952). Bonner,
Argersinger & Davidson (1952) and Lowen, Stoenner, Argersinger, Jr.,
Davidson & Hume (1951) studied equilibria between cations and exchange
resins. They found that the mass-law equations can be applied if the
activity factors are taken into account, and that the true equilibrium
constants can be obtained. Although the activity coefficients for the solid
phases cannot be determined in an absolute sense, they are present in the
mass-law equation in the form of a ratio which can be calculated from the
data. For example, in equation (2) the two solid phases are present in the
ratio (Y)j(UY). If each term is multiplied by its activity coefficient, then
the term becomes al(Y)la2(UY). In the case of the cation-resin systems,
the activity ratio, ajag, varies within an order of magnitude as the ratio
(Y)j(UY) is altered. However, when ( Y)j(UY) is 0-5, the ratio is approxi-
mately unity in the systems studied.
In comparative systems containing yeast, uranium, and a soluble com-
plexing agent (C), the equilibria can be expressed as
Kv= o^Y) as(UC)
K~ afUY) a4(C) '
(3)
174 ENZYME SYSTEMS OF THE CELL SURFACE INVOLVED
The comparison is made by determining the concentration of complexing
agent which will result in a ratio of ( Y)/( UY) of 0-5. Although none of the
activity coefficients is known for either the yeast groups or the complexors
studied, the activity ratios a2/a2 and a3/a4 should cancel out, as an approxi-
mation. The fact that the highly polymerized compound (molecular weight
about 2,000,000), in which the activity factors are comparable to those for
yeast, possess the same affinity for uranium as does yeast, is strong evidence
that the yeast surface sites are poly phosphates.
Can ATP be excluded as the binding agent on the surface of the cell?
The stability of the ATP-uranium complex is one-twentieth that of the
yeast-uranium complex. This difference would seem to be outside the error
attributed to activity factors. However, another factor must be considered.
Hurwitz (1953) has investigated the action of uranyl ions on the hexo-
kinase-glucose-ATP-Mg++ system. He found that the hexokinase-ATP
complex has a considerably greater affinity for uranyl ion than has free
ATP, with consequent inhibition of the enzyme activity. The finding of
Barron et al. (1948) that hexokinase was insensitive to uranyl ion was due
to the presence of a large excess of ATP which forms a chelate with uranyl
ion preventing it from acting on the enzyme. An ATP protein complex in
the surface of the cell could therefore account for the affinity which the cell
has for uranium.
Properties of the uranium complexes were also tested in regard to their
stability as a function of hydrogen-ion concentration. The experimental
technique was simple. A constant amount of yeast was mixed with a
constant quantity of uranium. A series of such suspensions was adjusted
to pH's in the range of 2- 5-4- 5. Higher pH's were avoided because above
pH 4-5 uranyl ion forms a series of complexes with hydroxyl ion. With
increasing pH there was an increased uptake of uranium complex. If the
same experiment was repeated in the presence of a fixed amount of organic
acid such as citrate, there was again with increasing pH an increased uptake
of uranium by the cells. It is known that the organic acid complexes of
uranium increase in stability as the pH is increased (Neuman et al. 1951),
therefore it must be concluded that the yeast-uranium complex increases
in stability to a greater extent than the carboxyl-uranium complex. If the
experiment is repeated with a polyphosphate (hexametaphosphate) instead
of citrate, then there is no redistribution of uranium as the pH is increased.
In other words, the increased stability of the yeast-uranium complex is
counteracted by an equal increase in the stability of the polyphosphate-
uranium complex. Thus polyphosphates are chemically similar to the yeast
surface sites not only because their uranium complex possess stabilities of
the same order, but also because they are influenced by pH in the same
IN THE UPTAKE OF SUGARS BY YEAST 175
manner. Carboxyl compounds do not resemble the cell surface on either
score.
From the evidence cited above, it seems clear that the uranium binding
of the yeast cell are multiphosphate in nature. However, on the basis of
the existing techniques, it is impossible to determine the exact nature of
the polyphosphate compound involved. It could be ATP, nucleic acid or
metaphosphate polymer.
V. QUALITATIVE EFFECTS OF URANIUM ON
METABOLISM
Low concentrations of uranium inhibit the fermentation of glucose, as
measured by CO2 production (Booy, 1940) as well as respiration of glucose,
measured by oxygen consumption (Barron et al. 1948). Similar inhibitions
are found if the disappearance of glucose is measured (Rothstein et al.
1948). Thus manometric measurements of oxygen consumption and CO2
production can adequately characterize glucose uptake in the presence of
uranium. In addition to its effect on glucose metabolism uranium also
prevents the utilization of fructose. Yeast cells do not utilize galactose
unless they have first been adapted to this substrate. Uranium blocks the
utilization of galactose by adapted cells.
Uranium, even in relatively high concentrations, blocks only 90% of
glucose uptake. Thus about 10% of glucose uptake proceeds through
uranium-insensitive pathways. All of the studies reported here are con-
cerned with the 90% of the sugar uptake which proceeds by uranium-
sensitive pathways.
The action of uranium on metabolism is restricted to an inhibition of
hexose metabolism (Rothstein et al. 1951). Thus uranium in sufficient
concentrations to block sugar metabolism has no measurable effect on the
respiration of alcohol and pyruvate, and only a small effect (less than 20 %
inhibition) on the respiration of acetate and lactate. Uranium in high
concentrations has no effect on endogenous respiration as measured by
oxygen consumption or glycogen depletion. Nor does uranium influence
the synthesis of glycogen with alcohol as a substrate.
Normally yeast does not ferment its stores of glycogen, but in the
presence of appropriate concentration of DNP, glycogen is rapidly degraded
to alcohol and CO2 (Rothstein & Berke, 1952). This DNP-induced endo-
genous fermentation is not inhibited by uranium.
In summary, then, uranium is an inhibitor of reactions specific to the
uptake of the hexoses but is without effect on all of the metabolic pathways
involved in the respiration and fermentation of other substrates, including
stored glycogen. In terms of the generally accepted scheme of metabolism
176 ENZYME SYSTEMS OF THE CELL SURFACE INVOLVED
the action of uranium must be confined specifically to those reactions
occurring at the cell surface which introduce sugars into the metabolic
machine, without any effect on the integrity of the machine itself, which is
presumably located inside the cell. The machine is therefore inaccessible
to the action of uranium even though many of its component parts are
inherently uranium-sensitive as shown by studies in vitro (Bounce & Lan,
1949).
The specific action of uranium on the initial reactions between glucose
and the cell, and the insensitivity of the rest of the metabolic cycle, offers
an explanation of the following phenomenon. If uranium is added before
or at the same time as the glucose, the onset of inhibition is immediate. If,
however, uranium is added 10 min. after the glucose, there is about a half-
hour delay in the onset of inhibition. The delay is presumably associated
with the utilization of intermediates of glucose which accumulated in the
cell before the addition of uranium. Uranium prevents further glucose
uptake, but does not interfere with the utilization of the intermediates
formed from glucose.
VI. QUANTITATIVE ASPECTS OF THE ACTION OF
URANIUM ON SUGAR METABOLISM
Two actions of uranium in yeast have been discussed : first, the formation
of a stable complex with specific sites on the surface of the cell, and second,
the inhibition of the uptake of hexoses by the cells. What is the relationship
between the uranium binding and the inhibition of metabolism? The
simplest relationship assumes that each surface site is directly associated
with glucose uptake, and also assumes that the binding of a uranium
molecule at that site prevents it from participating in glucose uptake. On
the basis of this assumption there should be a direct proportionality between
the number of sites which are combined with uranium, and the inhibition
of metabolism. Thus if one-half of the available sites are combined with
uranium, the inhibition should be 50% and with three-quarters of the sites
bound the inhibition should be 75%, etc. Such a relationship can be
expressed as (jjv\
T_(U Y) (.\
<-
where I is the inhibition, (UY) is the concentration of uranium-bound
sites, ( Y ) the concentration of unbound sites and ( YT) the total concentra-
tion of yeast sites. Equations (4) and (5) can be tested experimentally by
IN THE UPTAKE OF SUGARS BY YEAST 177
substitution in the mass-law formulation for the binding of uranium
(equation (2)). Thus / __ rx
J- (6)
Because the total uranium added (UT), is equal to free uranium ([/), plus
bound uranium (C/F),
K=[(UT)-(UY)]1-^. (7)
Again substituting equation (4),
(8)
But ( YT) is proportional to the yeast concentration. Therefore
(9)
where C is the yeast concentration in mg./ml. and k is a conversion constant
equating C and (YT). According to equation (9), at a fixed inhibition (UT)
should be proportional to C with an intercept equal to K. At 50%
inhibition, the equation can be simplified to
(UT) = o-5kC+K. (10)
In Fig. 2 the (UT) is plotted against C for 30, 50 and 70% inhibition. Data
are taken from a series of curves at eight different yeast concentrations for
inhibition of fermentation from Rothstein et al. (1948). In each case a
reasonably straight line can be drawn through the points. The lines
intersect the (UT) axis very close to the origin. For this reason the value
of K cannot be determined from the intercept with any accuracy. It is
obviously low relative to values for (t/r), less than i x io~6 (the value by
chemical determination was 3*5 x io~7, shown in Table i). On the other
hand, the value for k can be readily calculated from the slopes of the lines
to be about 7-6 x io~7 M for a yeast suspension containing i mg./yeast/ml.
suspension, i mg. of yeast contains i x io7 cells; therefore, using Avo-
gadro's number, each cell is predicted to contain about 4-6 x io7 binding
sites directly involved in fermentation. The predicted value is exceedingly
close to that found by actual experiment as shown in Fig. 4. As expected,
from the assumption made in equation (4), the inhibition is proportional to
the amount of uranium bound per cell, with essentially 100% inhibition
when 4-6 x i o7 molecules are taken up by each cell.
The 4-6 x i o7 sites per cell involved in fermentation are the same ones
whose properties were studied in regard to their mass-law behaviour
(Table i) and chemical identity (Fig. 2). However, these sites are not the
178 ENZYME SYSTEMS OF THE CELL SURFACE INVOLVED
only ones on the yeast cell capable of combining with uranium. The
presence of uranium concentration greater than that required for complete
inhibition of fermentation leads to a further uranium uptake, amounting
1-0 0-5 0-0 0-5 1-0 1-5 2-0 2-5 3-0 3-5 4-0 4-5 5-0 5-5 6-6
Log of concentration x106
Fig. 3 . The relationship beteeen yeast concentration and initial uranyl nitrate concentration
at a fixed inhibition, (i) yeast; (2) metaphosphate polymer; (3) hexametaphosphate ;
(4) desoxyribonucleic acid; (5) pyrophosphate ; (6) triphosphate; (7) ATP; (8) meta-
phosphate; (9) nucleic acid (tech.); (10) adenylic acid; (n) egg albumin; (12) serum
albumin; (13) citrate; (14) HDP; (15) orthophosphate ; (16) maleate; (17) glycero-
phosphate; (18) glucose- 1 -phosphate; (19) glucose; (20) acetate; (21) fructose.
100
80 -
60
40
20
Yeast cone,
in mg./ml.
CD 10
• 20
• 30
040
•54
Fig. 4.
012345678
Uranium uptake in molecules per cellxlO"7
The relationship beteeen uranium uptake by the cells and the
inhibition of glucose metabolism.
to at least 3 x io7 sites per cell and perhaps more. The exact extent of this
additional uranium binding cannot be readily determined for technical
reasons. The complex formed under these conditions is relatively unstable,
requiring high uranium concentrations to achieve saturation. The amount
IN THE UPTAKE OF SUGARS BY YEAST 179
bound by the cells becomes vanishingly small compared to the amount of
uranium added and falls within the limit of analytical error. The uranium-
binding data indicate the existence of at least two species of uranium-
binding loci in the cell, one more stable than the other. The * stable sites'
are involved in fermentation in a first-order relationship. The * unstable
sites' are not involved at all in fermentation.
In the case of respiration of glucose a more complicated situation exists.
A concentration of uranium which gives essentially complete inhibition of
fermentation results in only a 60% inhibition of respiration. The inhibition
of the remainder of the respiration requires very much higher uranium
concentration (Fig. 5). The uranium-binding properties of the cells are the
same under aerobic as under anaerobic conditions, therefore it must be
20 25
&
Fig. 5. The inhibition of fermentation and of respiration by uranium.
123456789 101112131415
U concentration in mols./I.
concluded that respiration involves more uranium-binding sites than does
fermentation. The inhibition curve for respiration consists of two distinct
phases. Sixty per cent of the respiration apparently passes through reactions
held in common with fermentation involving cell-surface sites capable of
forming a very stable complex with uranium. Forty per cent of the
respiration passes through a non-fermentative pathway which involves
surface sites which form a less stable complex with uranium. It seems more
than a coincidence that a similar 60:40 relationship has been shown with
urethane (Fisher & Stern, 1942). The uranium data are consistent with the
concept that two respiratory pathways are functional in the cell, one which
follows the fermentative scheme and one which does not. Perhaps the
'shunt pathway' through phosphogluconic acid is involved (Baldwin,
1952).
Differences in susceptibility to uranium have also been found in the case
of fermentation of galactose as compared to glucose. Yeast does not
normally ferment galactose, but pre-exposure of the cells to galactose plus
l8o ENZYME SYSTEMS OF THE CELL SURFACE INVOLVED
glucose results in the appearance of 'adaptive fermentation' of galactose.
The fermentation of galactose by 'adapted cells' is more sensitive, by a
factor of about 2, to inhibition by uranium than is the fermentation of
glucose by the same cells. Thus fewer surface sites are involved in galactose
metabolism than in glucose metabolism (Rothstein, Meier & Hurwitz,
It has been pointed out previously that the membrane of the yeast
cell is able to differentiate between glucose, fructose and mannose, as
compared to galactose, sorbose and arabinose, the latter sugars being able
to penetrate the cell membrane very slowly, if at all. On the basis of the
uranium studies it must be concluded that other kinds of differentiation are
also built into the membrane. First, the uptake of galactose in ' galactose-
adapted' cells is more sensitive to uranium than is the uptake of glucose.
Thus fewer surface sites are involved in the uptake of galactose than
glucose. Secondly, glucose uptake under aerobic conditions involves more
uranium-binding surface sites than does glucose uptake under anaerobic
conditions. Furthermore, glucose uptake under aerobic conditions in-
volves two kinds of sites, whereas glucose uptake under anaerobic condi-
tions involves one. The organization of that part of the cell surface
responsible for the uptake of sugars must be complex. It must contain
mechanisms possessing an inherent specificity for certain sugars. In
addition, it must contain two different systems for glucose uptake, one of
which is operative only under aerobic condition.
VII. KINETIC STUDIES
In the presence of an inhibiting concentration of uranium, the inhibited
reaction located at the cell surface determines the overall rate of metabolism.
The cell-surface reaction can therefore be characterized in terms of kinetics
and temperature effects if uranium is present. Yeast metabolism has been
shown by Hopkins & Roberts (1935) and by Gottschalk (1944) to obey the
Michaelis Menten equation
i _Km ii
F=FmS+Fm' (II)
where V is the rate of metabolism, S the substrate concentration, Vm the
maximal rate of metabolism and Km the Michaelis constant. This equation
is predicated on the existence of an enzyme substrate complex, the con-
centration of which determines the rate of the reaction. Thus
JE+ S^ES -^product. (12)
Inherent in the equation is the concept of a limited number of enzyme
sites. As the substrate concentration is increased, the rate of metabolism
IN THE UPTAKE OF SUGARS BY YEAST l8l
reaches a maximal value associated with the saturation of the enzyme sites
with substrate. Higher concentrations of substrate cannot increase the
rate of metabolism.
The Michaelis equation is tested by plotting i/V against i/S. A straight
line should obtain with an intercept equal to ijVm and a slope of Km/Vm.
In Fig. 6, taken from Hurwitz & Rothstein (1951), it can be seen that
fermentation and respiration give data consistent with the Michaelis
equation not only in the absence of uranium, but also in those experiments
200
Fig. 6. a. Kinetics of inhibition of fermentation of glucose, b. Kinetics of inhibition
of respiration of glucose.
in which an inhibiting concentration was present. It must therefore be
concluded that an interaction between glucose and the cell occurs at the
cell surface which involves a saturation phenomenon. In other words,
glucose combines in some manner with some constituent in the cell surface
which is present in limited concentration.
It is possible further to characterize the inhibition in terms of the
calculated values for Vm and Km (Baldwin, 1952). For example, in the
case of the inhibition of fermentation by uranium in Fig. 6, the calculated
value for Km is essentially the same for all three lines, 7-2 x io~3, 8-2 x io~3
and 8-3 x io~3 for control, i x io~6 M and 2 x io~6 M uranium respectively.
The Vm, however, shifts from 30 to 19 to 12 /d./mg./hr. Such behaviour
l82 ENZYME SYSTEMS OF THE CELL SURFACE INVOLVED
is typical of a non-competitive inhibition. In the case of respiration, in the
presence of uranium there is a fourfold shift in Km from 3-8 x io~3 to
1-9 x io~2, and a much smaller change in the Vm from 14 to n /^l./mg./hr.
Thus the kinetics of inhibition of respiration contains both competitive and
non-competitive elements. Although the mechanism of inhibition of
respiration is not obvious from the kinetic analysis, it is apparent that the
inhibition kinetics of fermentation are different from the inhibition kinetics
of respirations. This statement is based not only on the data of Fig. 6, but
also on other kinetic data concerned with the effect of different uranium
concentrations (Rothstein et al. 1951). The aerobic-anaerobic differences
in kinetics are undoubtedly a reflexion of the phenomenon discussed
previously. Fermentation proceeds through a reaction involving one kind
of cell-surface site, whereas respiration proceeds by reactions involving
two different kinds of cell-surface sites.
On the basis of the kinetic data certain suggestions can be made con-
cerning the role of cell-surface reactions in sugar uptake. The agreement
of the data with the Michaelis-Menten equation in the case of uranium-
poisoned yeast suggests that glucose combines with a component in the
cell membrane, the concentration of glucose-component complex deter-
mining the overall rate of metabolism. In fermentative glucose uptake,
uranium does not prevent the formation of the glucose-component complex,
but apparently prevents the breakdown of this complex into products,
doing so by combining with necessary phosphate groups. An analysis of
the respiratory glucose uptake is difficult because of the existence of two
distinct mechanisms of glucose uptake.
VIII. TEMPERATURE EFFECTS
The temperature dependence of the uranium-inhibited reactions were
characterized in terms of the Arrhenius equation
\ogV=-/il2-3RT+K, (13)
where V is the rate of metabolism, ^ the energy of activation, R the gas law
constant and T the absolute temperature. In Fig. 7, log V is plotted
against i/7\ The control data can be conveniently represented by a pair of
straight lines, the data is the presence of uranium by a single straight line.
The values for /i calculated for the two lines of the control data were 21,000
cal./mol. for the lower segment and 13,000 cal./mol. for the upper, in
essential agreement with the data of Stier (1933). The /i for the uranium
curve was 22,000 cal./mol. The lines for the control and inhibited data tend
to converge at higher temperatures. Thus elevation of the temperature
from 20 to 30° C. decreased the inhibition from 66 to 29%.
IN THE UPTAKE OF SUGARS BY YEAST 183
The value of jn of 22,000 cal./mol. is inconsistent with any mechanism
involving a free diffusion of glucose through the membrane in the aqueous
phase. It does not, however, exclude the concept of activated transport
through a lipid membrane as postulated by Danielli (1943), although it is
doubtful, in view of the low solubility of sugars in lipids, that movement
through a lipid phase occurs to any extent in sugar uptake. Such a
mechanism could not account for the selectivity of the membrane in regard
to various sugars, for the uranium effects or for the kinetics. Thus the high-
temperature coefficient is undoubtedly the characteristic of a chemical
reaction between glucose and a cell-surface constituent.
1-50
1-00
I 0-50
-0-50
33-0 33-5 34-0 34-5 35-0 35-5
Fig. 7. Temperature characteristics of respiration of glucose in the
presence and absence of uranium.
IX. THE EFFECT OF EXTRACELLULAR PH ON
THE RATE OF FERMENTATION
It has been suggested that the initial phosphorylation reactions in sugar
metabolism are mediated by enzymes located on the surface of the cell.
Such enzymes should be susceptible to the influence of extracellular pH.
Yet the reports in the literature indicate that fermentation is remarkably
independent of pH over a wide range (Euler & Heintz, 1919; Hagglund &
Augustson, 1925), a situation which is seemingly incompatible with the
hypothesis. The explanation of this contradiction lies in the fact that the
previous studies were made using potassium phosphate or potassium citrate
buffers. As will be shown, potassium counteracts the depressant effect
of hydrogen ion on fermentation, with the result that in the presence of
high potassium concentrations, the rate of fermentation is apparently
independent of pH over a wide range. However, if inert buffer systems are
used, fermentation is found to be markedly dependent on pH (Rothstein
& Demis, 19536).
Inert buffers were found for the pH range 2-0-6-0 (triethylamine(TEA)-
succinate-tartrate) and 8-0-10 (tris-hydroxymethylaminomethane or
184 ENZYME SYSTEMS OF THE CELL SURFACE INVOLVED
THAM). In the range 6- 5-7*5 no adequate buffers were found, therefore,
constant pH was maintained manually by addition, drop-wise, of TEA.
Manometric determinations of oxygen consumption and carbon dioxide
.production were feasible in the range 2*0-6*0 and 8-0- 10, with results which
agreed closely with the rates of sugar consumption. In the range of pH
6-5-7-5, metabolism was measured only in terms of glucose consumption.
No pronounced differences were observed between the effects of pH on
respiration as compared to those on fermentation. However, because there
were marked effects of pH on the end-products of fermentation, the data
presented here will concern this mode of metabolism.
50
40
30
o
£.20
6
u
10
2-0 2-5 3-0 3-5 4-0 4-5 5-0 5-5 6-0
PH
Fig. 8. The effect of pH on the rate of fermentation and on
the end products of fermentation.
The pH activity curve for anaerobic glucose uptake by living cells is a
biphasic curve, with optima at pH 5-5 and 8-5. Yeast can ferment glucose
over the pH range of 1-7-10-5, yet the average internal pH of the cytoplasm
remains constant. For example, Conway & Downey (19506) have shown
that during active fermentation in the presence of potassium with the
external pH dropping rapidly to a value below 2-0, the internal pH of the
cell becomes slightly more alkaline. Using Conway's freezing and thawing
procedure it was found that yeast fermenting in media maintained at pH's
ranging from 2-0 to 10-0 has a relatively constant internal pH in the range
6-2-6-4. Thus the rather dramatic effects of extracellular pH shown in
Fig. 9 must be characteristic of reactions occurring in the periphery of the
cell, exposed to the varying extracellular pH rather than to the constant
internal pH. The fact that the pH curve is biphasic is not entirely surprising.
IN THE UPTAKE OF SUGARS BY YEAST 185
It would be difficult to imagine any single enzyme reaction which could
show adequate activity over the wide pH range (i-7-iO'5) over which cells
show fermentative capacity. In general, enzymes have a considerably more
restricted zone of activity in regard to pH, encompassing perhaps 4-5 pH
units. On the basis that surface enzyme activity is necessary, it would seem
reasonable that the ability of yeast to take up glucose over an extended pH
range lies in the possession of two different surface enzymes, each respon-
sible for a portion of the pH range. Other evidence of the existence of two
surface enzymes has been found. At pH 8-5 mannose is respired 55 % as
rapidly as glucose, but at pH 3-5 mannose is respired 92% as rapidly as
glucose. At pH 8-5 glucose uptake is inhibited by calcium, whereas at
pH 2-0-6-0, glucose uptake is markedly increased by calcium. In fact at
pH 6 glucose uptake can be made almost entirely dependent on the
presence of calcium or other bivalent ions (see § XI).
There is some likelihood that the surface enzyme in alkaline fermentation
is the hexokinase crystallized by Kunitz & McDonald (1946) and Berger,
Slein, Colowick & Cori (1946). At least it has many similar properties.
It has a pH optimum on the alkaline side, it is only half as active toward
mannose as towards glucose, and it is inhibited by calcium.
The crystalline hexokinase could not account for fermentation in acid
solutions. It has a low activity at pH 5-0, less than 10% of that at pH 8-0.
It has no measurable activity below pH 4-0 (Hurwitz, 1953). Nevertheless,
in cell-free preparations considerable fermentative activity is found below
pH 4-0. It seems possible that a second hexokinase, with an acidic pH
optimum, is responsible for the fermentation at low pH in both cell-free
preparations and in living cells. Proof of its existence, of course, will only
be established by its isolation and characterization.
In addition to the effects of extracellular pH on the rates of fermentation,
there are some rather dramatic alterations in the end-products of fermenta-
tion. It has been known for some time that under alkaline conditions there
is an increased production of glycerol by fermenting yeast. Thus Neuberg
(Baldwin, 1952) classifies his third method of fermentation as the produc-
tion of 2 mol. of glycerol, i mol. of alcohol and i mol. of acetic acid.
Recently, Neish & Blackwood (1951) showed that as the pH is increased to
8-2, the glycerol and acetate production during fermentation increase
relative to alcohol production. Rothstein & Demis (1953 b) have done carbon
balance studies concurrently with the studies of the effects of pH on rates of
fermentation (Fig. 7). In the pH range 2-0-6-0 the end-products are largely
alcohol (35 %), carbon dioxide (20%), and glycogen (35 %) with only small
amounts of glycerol and acetate. In the pH range 8-o-io-o, the alcohol and
carbon dioxide remain about the same, but there is almost no net glycogen
l86 ENZYME SYSTEMS OF THE CELL SURFACE INVOLVED
synthesis. In place of glycogen, there is a marked increase in glycerol
production and also some increase in acetate. Sussman, Spiegelman &
Reiner (1947) had previously noted that assimilation of glucose was
decreased at alkaline pH. Wiggins, Mann, Trevelyan & Harrison (1952),
on the other hand, found no decrease in glycogen synthesis at pH 8-5. The
basis for the discrepant results does not seem obvious at the present time,
unless it is due to differences between strains of yeast.
The pH at which the change-over in end-products occurs corresponds
to the dip between the two optima in the pH curve for sugar uptake
(Fig. 7). The pH curve for glycerol production corresponds almost exactly
with the alkaline phase of the pH curve for sugar uptake. In view of the fact
that the internal pH remains constant it must be concluded that reactions
occur at the cell surface, susceptible to the external pH, which not only
determine the rate of sugar uptake, but also determine the nature of the
end-productions of fermentation. The surface mechanism for sugar uptake
is not simple in that it merely delivers glucose or a product of glucose
metabolism into the interior of the cell for conversion there into end-
products. Instead, it is complex, containing within itself capacities which
determine the nature of the end-products.
X. THE EFFECT OF MONOVALENT CATIONS ON
CELL-SURFACE REACTIONS
Potassium is present in relatively high concentrations in most cells. In the
yeast cell its concentration in the cytoplasm is of the order of o-i M.
Potassium is related to carbohydrate metabolism in yeast in at least two
ways. In the first place potassium can stimulate the rates of fermentation,
respiration and glucose consumption (Lasnitzki & Szorenyi, 1935; Farmer
& Jones, 1942; Rothstein & Enns, 1946). In the second place, during
metabolism of sugar, potassium is taken up by yeast cells (Pulver & Verzar,
1940) in a reaction which involves an exchange for hydrogen ions produced
by cell metabolism (Rothstein & Enns, 1946; Conway & O'Malley, 1946).
Conway has studied the K+-H+ exchange in some detail and has presented
a general theory of the mechanism of acid secretion (Conway, 1953).
The relationship between the two K+ phenomena has been investigated
by Rothstein & Demis (1953 a). Several possibilities seemed worthy of
consideration. Potassium might increase the rate of fermentation by
stimulating reactions occurring on the surface of the cell. Or, because
potassium is taken up by the cell in exchange for hydrogen ions, the
stimulation of metabolism might be associated either with the increased
intracellular potassium content, or with the increased acid secretion.
The stimulation of sugar uptake by potassium is dependent on two
IN THE UPTAKE OF SUGARS BY YEAST 187
factors, the potassium concentration and the hydrogen-ion concentration.
There are two zones of pH in which potassium has a substantial influence,
2-0-4-0 and 5*5-7-5, corresponding to the depressed portions of the pH-
activity curve of fermentation (Fig. 7). Because the K stimulations in the
two zones seem to be mediated by different mechanisms, they will be dis-
cussed separately.
In the lower pH range the magnitude of the effect increases as the pH is
decreased, but higher concentrations of potassium are required to invoke
2-0 30 4-0 5-0 6-0 7-0 8-0 9-0 10-0
0-2
Fig. 9. The stimulation and fermentation by potassium as influenced by pH.
the maximal effect. The rate of metabolism in the absence of potassium
decreases from a maximum at pH 5-0 to 43 % of normal at pH 2-0, but the
maximal rate of metabolism in the presence of potassium is essentially the
same at all values of pH (Fig. 8). It appears, therefore, that hydrogen ion
depresses the rate of fermentation and that the appropriate concentration of
potassium can counteract the inhibiting effect. The higher the concen-
tration of hydrogen ion, the higher must be the potassium-ion concen-
tration to achieve reversal. The interaction of potassium and hydrogen ions
apparently takes the form of a direct competition for the same loci because
there is a fixed relationship between the concentrations of the two ions.
Thus with any given hydrogen-ion concentration above ixio~5M, a
l88 ENZYME SYSTEMS OF THE CELL SURFACE INVOLVED
maximal rate of metabolism can be obtained with a ratio of K+ to H+ of
approximately 10 to i.
Under certain conditions, the effect of potassium on the rate of fermenta-
tion is considerably more dramatic. In the experiments described in the
preceding paragraph, the cells were starved with aeration for several hours
before they were used. If the cells are starved for a longer period of time,
of the order of 24 hr., they lose a considerable amount of potassium. Such
cells have a very low rate of metabolism at pH 2-0, but the addition of
0-2 M-KC1 will instantly return the rate to almost maximal value, a stimula-
tion of over 400%.
Potassium ion exerts its effect predominantly on fermentation rather
than respiration. For example, at pH 2-7 the addition of 0-02 M-KC1 to
a yeast suspension resulted in an increased rate of sugar consumption
amounting to 83% under anaerobic conditions compared to 69% under
aerobic conditions. However, the increased rate of uptake under aerobic
conditions is largely due to an increase in the rate of aerobic fermentation.
Thus, if the potassium effect is measured in terms of carbon dioxide
production and oxygen consumption, then the stimulation of fermentation
is 84%, and the stimulation of respiration only 25 %. There is an increase
in the R.Q. from 1-7 to 2-5, and aerobic fermentation is increased 146%.
It can be calculated that about 90% of the increase in the rate of aerobic
glucose consumption is due to the increased aerobic fermentation. Respira-
tion of lactate, pyruvate and alcohol is stimulated to about the same extent
as is the respiration of glucose, about 20%.
Other cations were tested for their effect on fermentation at pH 2-7. In
each case the concentration was 0-02 M. Potassium had by far the greatest
effect, 83 %, followed in order by rubidium 40%, calcium, magnesium and
manganese about 20-25%, s°dium 15-20%, ammonium, lithium and
caesium 10%. None of the other ions reduced the stimulating action of
potassium.
The stimulating action of potassium in the pH range 5-5-7-5 can be very
dramatic if the cells are starved for a few hours and then thoroughly
washed with distilled water. For example, at pH 6-0, the stimulation of
fermentation of glucose is over 100%. If the cell suspension is first
treated with a TEA-cation exchange resin (Dowex 50), the potassium
stimulation is greater than 700 % . At pH 6-0, the rate of fermentation is
normally considerably lower than at pH 5-0 or 8-5 (Fig. 7). Treatment
with resin reduces the rate at pH 6-0 to a very low level but has much less
effect on the rates at pH 5-0 and 8-5. Thus the cells are particularly
dependent on potassium at pH 6-0. The presence of potassium returns the
rate to the same level found at pH 5-0 or 8-5, without added potassium.
IN THE UPTAKE OF SUGARS BY YEAST 189
The potassium dependence of fermentation at pH 6-0 does not seem
to be the same as the potassium dependence at low pH discussed previously.
In the first place, NHj can displace K+ at pH 6-0 but not at low pH. The
stimulation of fermentation by NH^~ was first noted by Zeller (1926) and
by Smythe (1939). The effect seems to be due to NH^~ rather than un-
dissociated NH4OH. In the second place the stimulation at pH 6-0 is
associated with an increased glycerol fermentation, whereas that at pH 2-0-
4-0 leads to an increased alcoholic fermentation. Thus the stimulation at
pH 6-0 seems to be primarily of the pathway of metabolism associated
with the alkaline phase of the pH curve, with its optimum at pH 8-5.
Yeast, during metabolism of sugars, secretes acid with a resultant decrease
in the pH of the medium. In an unbuffered medium with no salts present,
the pH will drop slowly to 3-5, due to succinic acid secretion. In the
presence of 0-02 M-KC1, the pH drops rapidly to a minimum of 2-6, a con-
sequence of the exchange of K+ for H+. If the concentration of potassium
is increased the minimum pH is decreased. Thus Conway & O'Malley
(1946) found that with 0-2 M-KC1, the pH dropped to values as low as 1-7
Other ions such as rubidium, sodium and lithium also are exchanged for
H+, but at a much slower rate (Conway, 1953 ; Rothstein & Demis, 1953 a).
Is the K+ stimulation which is discussed in the previous section related
to the K+-H+ exchange? The exchange reaction is rapid at higher pH's,
but with 0-02 M-KC1 approaches zero at pH 2-6-2-7 and is actually negative
(in the reverse direction) below pH 2-6. On the other hand, the stimulation,
of metabolism with the same potassium concentration is greatest at pH 2-0.
decreasing as the pH is raised to 5-0. Thus the stimulating action of
potassium is observed, whether there is a net exchange of cellular H+ for
K+ from the medium, no net exchange, or a net exchange in the opposite
direction. The stimulation of metabolism by potassium is therefore not
directly associated either with the secretion of hydrogen ion by the cell, or
with the uptake of potassium by the cell.
Further light is thrown on the potassium effect by studies of cell-free
yeast preparation similar to that prepared by Meyerhof & Kaplan (1951).
The cells are slowly dried at room temperature, treated with acetone, then
lyophylized and pulverized (Rothstein & Demis, 1953 b). Such a dried
yeast is cell-free, and it will ferment glucose rapidly with no lag period.
The preparation was treated with cation resin previously neutralized with
TEA. Thereby all of the normal cations of the yeast are replaced by TEA.
A mixture of bivalent ions is added to the yeast preparation in a TEA-
succinate-tartrate buffer. Such a preparation will only ferment glucose if
potassium is added. It has a narrower pH-dependence curve than live
yeast. In addition to its absolute dependence on potassium, such a prepara-
IQO ENZYME SYSTEMS OF THE CELL SURFACE INVOLVED
tion shows a H+-K+ relationship similar to that of live yeast. As the pH is
reduced the rate of metabolism is reduced, but can be counteracted to some
extent by an increased potassium concentration. In such a preparation the
cell membrane is not intact. It leaks proteins and other cytoplasmic con-
stituents such as organic phosphate and potassium. There is no impediment
to movement of K+, H+ or glucose to the enzyme sites and yet the inhibiting
effect of H+ and counteracting effect of K+ can be demonstrated. Thus the
effect of K+ on metabolism can probably be attributed to its action on
fermentation enzymes rather than to its effect on acid secretion or on some
permeability property of the membrane.
The normal content of potassium in yeast is of the order of o- 1 M. In view
of the fact that concentration in the medium as low as 0-0003 M can evoke
a metabolic effect, and in view of the fact (see preceding section) that the
stimulation can occur under conditions in which there is no net uptake of
potassium, it would seem that the intracellular potassium is not involved in
the observed phenomenon. These observations were confirmed experi-
mentally by manipulating the intracellular concentration of potassium.
High-potassium yeast was prepared by pre-exposing cells to glucose plus
potassium in citrate buffer at pH 4-5. These cells took up potassium equiva-
lent to 0-05 M/l. of cells. Low-potassium yeast was prepared by starving
for 24 hr. They lost 0-03 M of potassium per litre of cells. The three types
of cells, low potassium, normal and high potassium, had cellular potassium
concentrations of 0-07, o-i and 0-15 M, a twofold range. The maximal rates
of fermentation induced by extracellular potassium (0*02 M) was the same
in each case (40-41 /^l./mg./hr.) (Rothstein & Demis, 1953 a), even though
no additional K+ was taken up during the course of the experiment. It
must be concluded therefore that the fermentation reactions influenced by
potassium are peripherally located in the cell, where they are influenced by
extracellular rather than intracellular potassium.
XL THE EFFECTS OF THE BIVALENT IONS ON
SURFACE REACTIONS
Uranyl ion inhibits sugar uptake by combining with polyphosphates on the
cell surface, substances which may be involved directly in phosphorylation
reactions. Phosphorylation reactions in general require the presence of
bivalent ions for maximal activity, particularly magnesium and sometimes
manganese. An attractive hypothesis for explaining the uranium effects
involves the displacement of magnesium and manganese ions by a com-
petitive effect from phosphorylation reactions requiring these ions.
The existence of competition between various bivalent ions and uranyl
ion was shown in two ways. In the first studies (Rothstein & Meier, 1951),
IN THE UPTAKE OF SUGARS BY YEAST 191
yeast cells were equilibrated with uranium resulting in the binding of 95 %
of this ion. On the addition of other bivalent ions, some of the uranium
was displaced from the cells by competition and appeared in the medium.
Magnesium, calcium, barium and zinc showed competitive effects. Sodium
and potassium did not. However, the cell-surface sites had a far greater
affinity for uranyl ion than for any of the other ions tested. For example,
the affinity for uranyl ion is of the order of several thousand times that for
magnesium or calcium. In the second technique, displacement of uranium
from the surface sites was not measured chemically, but instead, in terms
of a reduction in the inhibiting action of uranium. Magnesium, calcium
and manganese were tested and all were able to reduce the inhibiting action
of uranium (Hurwitz, 1953; Rothstein & Hayes, 1953). This experiment
indicates not only that uranyl ion can be displaced by the bivalent ions
tested, but also that surface sites combined with magnesium, calcium or
manganese are metabolically active. The question remains as to whether the
cell tolerates or requires that the surface sites be present in the form of a
magnesium, calcium or manganese complex for sugar uptake to proceed.
The relationship of the binding of bivalent cations by the cell surface to
the uptake of glucose has been investigated by Rothstein & Hayes (1953).
Isotope studies with 55Mn and 45Ca reveal that the cell can reversibly bind
a fixed number of cations. All of the bivalent ions tested, including Mn+ + ,
Ca++, Mg++, Ba++, Co++, Zn++, Hg++, Cu++ and UO2++ compete for the
binding sites, but of these ions, UO^"+ has by far the greatest affinity. The
equilibrium between the ions and the cells is attained very rapidly (less
than 3 min., the fastest time that could be measured) and does not alter
appreciably during the next hour. If the cells are equilibrated with 55Mn
or 35Ca and are then resuspended in solutions of non-isotopic Mn++ or
Ca++ , there is an immediate back exchange of the labelled ions, with the
attainment of the same equilibrium point, whether the experiment is
carried out with unlabelled cells in solutions containing the labelled ions
or with labelled cells suspended in solutions containing the unlabelled
ions. For example, in Table 2, in the experiments with no added phosphate
(column A), in the presence of 0-33 x io~4 M-Mn++, 66% of the 55Mn is
taken up within 3 min. and no further uptake occurs for i hr. In the
presence of 6-0 x io~4 M-Mn++ (column C), only 44 % is taken up, because
of the greater dilution of 55Mn with unlabelled Mn++ . Again the equili-
bration is complete in 3 min. Another sample of yeast first equilibrated
with 0-33 x io~4 M-Mn++ also shows 66% binding of 55Mn. After 15 min.
the Mn++ concentration is increased to 6-0 x io~4, and within a few minutes
some of the 55Mn is displaced from the cell, and the new equilibrium is
the same as that which would have resulted had the higher concentration
192
ENZYME SYSTEMS OF THE CELL SURFACE INVOLVED
of Mn++ been present initially (column B). The same equilibrium point is
achieved in either direction.
Table 2. The influence of phosphate on the uptake of 55Mn in the
presence of different concentrations of Mn++
Time (min.)
Percentage 55Mn uptake
No phosphate
Plus phosphate
A
B
C
D
E
3
15
4~
60
go
120
66
65
65
65
64
66
44
45
46
45
48
85
95
TOO
100
100
100
47
48
46
98
97
Cone, of
Mn++
added
0'33 X IO~4M
0*53 X IO~4M
until 15 min.
then in-
creased to
0-33 x io~4M
6-0 x io~4M
0'33 X IO~4M
0'35 X IO~4M
until 45 min.
then in-
creased to
6-0 x io~4M
Yeast concentration 200 mg./mlM pH 5-5, glucose o-2M, potassium chloride O-OZM,
TEA-succinate-tartrate buffer. A trace amount of 55Mn is added in each experiment,
sufficient to give 350 counts per ml. of the medium.
The characteristics of the binding of bivalent cations by the yeast cell are
similar to those described for cation exchange resins. The mass-law
derivations that have been applied to the exchange resins apply just as well
to yeast. The concentration of binding sites calculated from mass-law
measurements of 55Mn binding is approximately i x io~3 M/l. of cells,
a figure which is in good agreement with that determined by studies of
uranium binding. The dissociation constant is of the order of i x io~3 for
Mn++, Mg++ and Ca++, a figure which is only approximate because of
the long extrapolation involved. On this basis these cations form complexes
less stable than that of uranium by a factor of about 5000, a figure which
agrees with the relative stabilities calculated by direct competition.
There is considerable evidence that the binding and exchange of bivalent
ions as described in the preceding paragraphs takes place only at the
surface or periphery of the cell. For example, the existence of competition
of UO^+, Ca++, Mg++ and Mn4"1-, and the almost exact correspondence
between the number of uranium binding sites and Mn++, Ca++ and Mg++
binding sites, indicates that the binding sites for all of these cations are
identical. It has already been shown that the uranium-binding sites are
located on the surface of the cell, therefore the Mn4-4-, Ca++ and
IN THE UPTAKE OF SUGARS BY YEAST 193
binding sites must also be located on the surface of the cell. In addition,
evidence independent of the uranium competition can be cited. First,
the extremely rapid equilibration of the cell with the bivalent ions of the
medium argues for a surface phenomenon. Complete equilibration of the
contents of the cytoplasm would presumably take more than 3 min.
Secondly, the maximal binding of the bivalent ions by the cells is about
i x io~3 M/l. of cells. Since the bivalent cation content of the total
cytoplasm in the same cells is about 40 x io~3 M/l. of cells, only 2-5 % of the
bivalent cations of the cell can participate in the exchange for ions in the
medium. Thus the bulk of the bivalent cations of the cytoplasm are not in
communication with the environment. It is therefore suggested that those
ion-binding sites located on the cell surface can readily equilibrate with
the medium. This concept is similar to that put forth by Mazia (1940) and
Lansing (1942).
The absence of exchange between bivalent cations of the interior of the
cell and those of the medium has also been demonstrated by labelling the
interior compartment with 5r'Mn or 35Ca (Goodman & Rothstein, 1953).
Labelling of the cytoplasm was accomplished by setting up conditions
whereby the bivalent cations were actively transported from the medium
into the cell. Resting cells, as already indicated, do not actively take up
the bivalent cations of the medium ; they simply equilibrate in a manner
dictated by mass-law considerations. The addition of glucose and the
resultant appearance of exogenous metabolism, either aerobic or anaerobic,
does not alter the situation. However, if phosphate is added together with
the glucose, then the bivalent ions are actively transported into the cell
against the concentration gradient. In Table 2, column D, it can be seen
that in the presence of phosphate all of the measurable 55Mn has been taken
up by the cell. The Mn++ taken up in this manner is no longer in equili-
brium with the environment for it cannot be washed out, nor can it be
exchanged back when the Mn++ concentration of the medium is increased
(Table 2, column E). The amount of Mn++ that can be transported into
the cell in the presence of phosphate can be considerable, amounting to
0-02 M/l. of cells. This represents a 50 % increase in the bivalent ion content
of the cell and 20 times the maximal reversible binding of Mn+4~ in the
absence of phosphate. The detailed mechanism of the bivalent cation
transport is beyond the scope of the present discussion, but it can be
briefly stated that it depends on the active transport of the phosphate,
which carries the cation into the cell in the form of a soluble complex.
The studies of bivalent ion uptake support the conclusion that the
reversible binding in the absence of phosphate is indeed an equilibration
only with sites in the periphery of the cell and not with those in its interior.
E B S VIII 13
194 ENZYME SYSTEMS OF THE CELL SURFACE INVOLVED
Once carried into the interior of the cell, a bivalent cation is no longer
exchangeable with environmental ions. How does the binding of various
bivalent ions on the surface of the cell influence the ability of the cell to
take up glucose? The initial studies indicated that Mn++, Mg++ and Ca*"1
could each stimulate the uptake of glucose about 20-30 % . Attempts to
make the cells more dependent on environmental ions were carried out by
deionizing by prolonged washing and starving of the cells, and finally,
more successfully, by treating the cells with cation exchange resins in the
form of the triethylamine (TEA) salt. By these means it was possible to
reduce considerably the ability of the cell to take up glucose. The rate of
glucose uptake could then be returned to normal by adding back the
various ions. The predominant effects were obtained with K+ (see pre-
ceding section), but the metabolism was also remarkably dependent on
the presence of the bivalent ions, particularly at certain values of extra-
cellular pH. Sample data are given in Table 3. At pH 3-5, each of the
ions stimulates 20-25 %, but at pH 6-0 they stimulate 100 % . If the cells
are pretreated with TEA resin, the control rates at pH 6-0 can be reduced
almost to zero, and the metabolism is then completely dependent on the
presence of extracellular ions.
Table 3. Effect of bivalent ions on fermentation of glucose
pH of the medium
3'5
6-0
8'5
Control
TEA
Ca
Mg
Mn
25
32
30
30
M
H
3i
27
28
39
33
38
38
Data are in fi\, of CO2/mg./hr. Ions are 0-003 M.
The effects of bivalent ions are not simple. The dependence of glucose
uptake on these ions varies considerably with extracellular pH. At pH
values below 6-0, Mg++, Mn4"* and Ca++ all stimulate, especially at
pH 5-5-6-0. At pH 8-5, none of them stimulate, but Ca+4~, in contrast to
Mg++ and Mn++, inhibits about 50 % at 0-05 M. Regardless of complica-
tions it can be concluded that the surface sites involved in the binding of
bivalent ions are also involved in some manner in the uptake of glucose.
In fact, under certain conditions the uptake of glucose proceeds at very
low rates unless bivalent ions are present.
IN THE UPTAKE OF SUGARS BY YEAST 195
XII. PHOSPHATE UPTAKE AND PHOSPHATASES
Kamen & Spiegelman (1948) have suggested that uptake of phosphate by
micro-organisms involves an active transport system. In the case of yeast
cells, there is clear-cut evidence in this regard. Resting cells do not
exchange phosphate, lose phosphate or take up phosphate at any appreciable
rate (Hevesy, 1948). In the presence of glucose, however, there is a rapid
uptake of phosphate by the cell and incorporation into metaphosphate
compounds (Wiame, 1949). In the presence of low concentrations of azide
(Spiegelman, Kamen & Sussman 1948) or dinitrophenol (Hotchkiss, 1944),
the phosphate uptake is repressed, even though the rate of fermentation is
not diminished. If these poisons in low concentrations act by * uncoupling*
oxidative steps from phosphorylation, then it is apparent that phosphate
uptake in the intact cell is not only metabolism connected, but is directly
associated with phosphorylation mechanisms.
The uptake of phosphate during sugar metabolism does not seem to be
an inward diffusion in response to a concentration gradient associated with
a reduced internal phosphate concentration consequent to phosphorylation
processes, a mechanism aptly described by Rosenberg & Wilbrandt (1952)
as ' trapping'. In the first place there is almost no exchange of phosphate
between the cells and the medium either in the absence of glucose (Hevesy,
1948) or in the presence of glucose, during the rapid uptake of phosphate,
as shown by the constancy of the specific activity of the extracellular
phosphate using 32P (Rothstein & Meier, 1949). This indicates that the
process of phosphate uptake involves a movement only in the inward
direction, a phenomenon incompatible with the concept of an inward
movement of phosphate in response to a concentration gradient. In the
second place, orthophosphate moves into the cell against the concentration
gradient. Data taken from Rothstein & Meier (1949) and Schmidt, Hecht
& Thanhauser (1949) indicate an appreciable uptake of phosphate when the
extracellular concentration is as low as 0-0004 M> The intracellular ortho-
phosphate concentration in these experiments was of the order of o-oi-
0-02 M, or 25 times as high.
The evidence presented above suggests that phosphate is taken up by a
mechanism involving incorporation of phosphate into phosphate compounds
at the cell surface. The storage form of phosphate in the cell is metaphos-
phate (Wiame, 1949; Schmidt et al. 1949). However, in view of the high
energy content of the phosphate linkages in metaphosphates it seems un-
likely that phosphate is directly incorporated into such compounds. A more
likely explanation is the uptake of orthophosphate by phosphoglyceral-
dehyde, an essential reaction in the fermentative scheme (Baldwin, 1952).
13-2
196 ENZYME SYSTEMS OF THE CELL SURFACE INVOLVED
Phosphatases have often been thought to play a role in absorption of
sugars (Danielli, 1952). In view of the fact that phosphatases have been
localized at the surface of the yeast cell, their function has been investigated
(Rothstein & Meier, 1949). In the presence of low concentrations of molyb-
date or tungstate, the phosphatases were completely inhibited, but there
was no measurable effect either on phosphate uptake or on sugar uptake.
Thus these particular phosphatases play no role in either process.
XIII. WHAT IS THE CELL SURFACE?
In this paper, reactions are classified as 'cell-surface* reactions, if they are
directly influenced by the external environment rather than by the internal
environment of the cell, provided that the reaction is not associated with
a secreted enzyme, and provided that there is no alteration of the internal
environment that could account for the effect. The internal environment
is defined in terms of constituents which can be extracted from the cell by
such techniques as freezing and thawing, by drying, by extracting, etc.
These constituents, such as K+, H+ and orthophosphate, are presumed to
be distributed in the cytoplasmic water. On this basis the cell surface is
defined in terms of a barrier which separates the zone of influence of the
inside environment from that of the outside environment. Cell-surface
reactions occur on or outside of the barrier.
On the basis of the data on living yeast cells presented in this section, it
is not possible to make definitive statements concerning the structure of the
barrier. However, other experiments with a cell-free, fermenting system
throw some light on the problem. The fermentative enzymes of yeast are
soluble proteins. Many soluble 'zymase' preparations have been prepared
which can ferment glucose. Yet a cell-free preparation has been prepared
as described in § X, which contains over 80 % of the fermentative capacity
of the cell in an insoluble residue, even when suspended in an ionic
environment similar to that of the yeast cytoplasm (Rothstein & Demis,
1953 b). It produces alcohol, CO2 and glycogen, but has almost no ability
to respire. It is therefore suggested that in the intact cell the fermentative
enzymes are also retained in an insoluble structure. Furthermore, the
structure possesses some kind of a permeability barrier. When it ferments
glucose in the presence of inorganic phosphate, it takes up the phosphate
and incorporates it into phosphorylated intermediates of metabolism which
are retained within the structure. These do not appear in the medium
unless the structure is treated with 5 % trichloroacetic acid, in which case
they can be extracted. On the other hand, although the structure can
ferment glucose, producing phosphorylated intermediates, the latter com-
pounds when added to the medium cannot be fermented. Thus there is an
IN THE UPTAKE OF SUGARS BY YEAST 197
insoluble residue of the cell, which contains the bulk of the fermentative
activity of the cell and which has a permeability barrier to the diffusion of
phosphate esters in either direction. This barrier apparently has no re-
lationship to the general permeability barrier of the cell, which is destroyed
by the method of preparation, with the resultant leakage of most of the
soluble cytoplasmic constituents such as potassium and inorganic phos-
phate, as well as proteins. For convenience the cell-free, insoluble,
fermenting structure will be called the 'glycosome'.
The evidence previously presented concerning the actions of environ-
mental factors on glucose and phosphate uptake have been interpreted on
the basis that initial steps in fermentation are located in the periphery of
the cell. The evidence that the bulk of the fermentative activity of the cell
is contained in a structural element, the 'glycosome', suggests that this
element must also be located in the periphery of the cell, perhaps as a shell
which constitutes all or part of the cortex of the cell. Additional support
for this view is found in the similarities of the properties of the ' glycosome '
and the surface of the intact cell. Both are impermeable to phosphorylated
intermediates of metabolism. Both can take up and esterify inorganic
phosphate. Both are inhibited in regard to glucose uptake by low pH,
with a reversal of this effect by K+.
The 'glycosome' differs from the intact cell in that in addition to the
K+-H+ ion effects, its fermentation has an absolute dependence on the
presence of K+ even at the pH optimum, whereas the living cell does not.
Presumably the structure of the glycosome' is such that its outer surface,
that which is exposed to the environment, contains reactions in sugar
metabolism which are not absolutely dependent on K4", but which show
inhibition by H+ and reversal by K+. The inner surface and perhaps the
interior of the * glycosome ' contain reactions with an absolute dependence
on K+, but this does not show in the living cell because the inner surface is
always exposed to the high K+ content of the cytoplasm.
XIV. CONCLUSIONS
A number of observations have been made concerning interactions taking
place at the surface of the yeast cell, interactions which are responsible for
the uptake of sugars from the extracellular environment. The pertinent
facts can be listed as follows:
(1) The cell surface is highly selective towards various sugars.
(2) There is more than one surface mechanism by which glucose can be
taken up, one operative only under aerobic conditions.
(3) The surface reactions show properties typical of enzyme reactions
in regard to kinetics and temperature effects.
198 ENZYME SYSTEMS OF THE CELL SURFACE INVOLVED
(4) Polyphosphates and bivalent ions such as magnesium and manganese
are involved directly in sugar uptake.
(5) Orthophosphate is esterified at the cell membrane in connexion with
sugar uptake.
(6) The pH activity curve for sugar uptake is biphasic, covering a range
from less than 2-0 to greater than 10-0, but the interior pH of the cell
remains constant.
(7) Substances acting at the cell surface such as H4" and NH4+ can alter
the sugar uptake not only quantitatively, but qualitatively as well, in terms
of an alteration in the end-products of fermentation.
(8) Extracellular potassium can markedly stimulate fermentation of
living cells, whereas H+ inhibits. Similar effects have been shown in a
cell-free system in which the permeability barrier is destroyed.
(9) A cell-free preparation can ferment glucose, but not sugar phosphates,
indicating the presence of an intact structural unit impermeable to sugar
phosphates but permeable to glucose. Only on autolysis are the enzymes
for metabolizing sugar phosphates liberated.
It is obvious that a mechanism of considerable complexity is located at
the outer boundary of the cell, a mechanism with the function of absorbing
sugars. Not only is specificity for particular sugars built into this mechanism,
but in addition specific differences related to the nature of metabolism.
Thus glucose taken up under aerobic conditions is moved through the
cell surface by two mechanisms, one of which is not operative in glucose
uptake under anaerobic conditions even though the single anaerobic
mechanism has a capacity for sugar uptake which is almost twice that
accomplished by both aerobic mechanisms.
K+ and NH4+, which influence the fermentative uptake of sugars by
acting on the cell-surface mechanisms, not only alter the sugar uptake in a
quantitative manner, but exert qualitative effects on the course of metabolism
in terms of the nature of the end products. It seems, therefore, that the
surface mechanism is not simply a means for moving sugar into the interior
of the cell where it can then pass through a series of reactions controlled by
the cytoplasmic environment, but rather that the reactions at the cell
surface must proceed to the point where the determination of the end-
products occurs.
What is the nature of the interaction between the cell surface and glucose?
It is obvious that no simple diffusion mechanism either through pores or
through a lipid layer is compatible with the described properties such as
the specificity, the kinetics, the temperature coefficient. The transport of
sugars in the form of a complex with its liberation from the complex on the
inside of the membrane might account for the specificity, kinetics and
IN THE UPTAKE OF SUGARS BY YEAST 199
emperature coefficient, but could not readily account for aerobic-anaerobic
differences, for the predetermination at the surface of the end products of
metabolism, nor for the K+-H+ ion effects, which can be reproduced in
a cell-free system in which the permeability barrier is destroyed.
The hypothesis described in the introduction to this chapter to the effect
that sugar uptake involves metabolic interactions at the outer surface of
the cell is consistent with the available data. It is suggested that the
phosphorylation reactions of the fermentative schema are located in a
structural element of the cell which lies directly underneath the perme-
ability barrier (plasma membrane). This structural element may constitute
the peripheral gelatinous zone of the cell generally called the cortex or
ectoplasm. The integrity of the cortex remains intact in cell-free prepara-
tions even though the permeability barrier is destroyed as indicated by the
failure of such a preparation to utilize sugar phosphates, even though it can
rapidly ferment glucose. The fermentation by cell-free preparations also
has a restricted pH range and a complete dependence on the presence of
K+; whereas that by live cells has a very wide range and no absolute
dependence on potassium. Thus in a living cell the cortex must be shielded
from the external environment. It is suggested that glucose is phosphory-
lated at the permeability barrier by ATP and hexokinase, that the sugar
phosphate resulting from the reaction proceeds through the glycolytic
reactions within the structure of the cortex, the ATP being regenerated by
the coupled oxidation at the phosphoglyceraldehyde dehydrogenase reaction.
This reaction also serves to pick up orthophosphate and could account for
its uptake from the medium. It is assumed that sugar phosphates cannot
penetrate beyond the permeability barrier, that orthophosphate can
penetrate into the cortex and that glucose can move into the cortex only by
undergoing phosphorylation. One reaction in glucose fermentation which
is directly exposed to extracellular pH is the first one in the chain, the
hexokinase reaction. It is suggested that in order to encompass the pH
range 1-7 to n-o, there must be two hexokinases with different pH optima
explaining the biphasic curve for the effect of pH on fermentation, the
different actions of Ca++ at high and low pH, and the relative rates of
mannose metabolism. The effect of uranium is on the first reaction of sugar
uptake by chelating with ATP and thus preventing the phosphorylation.
The uranium does not penetrate into the cortex as shown by its failure to
inhibit other metabolic reactions and by the fact that the uranium complex
with the cell is directly susceptible to the effects of external pH. However,
uranium also inhibits a second reaction, one which participates in respira-
tion but not in fermentation. The effects of K+ and NH4+ on fermentation
seem to be due to their ability to counteract the inhibiting effects of
2OO ENZYME SYSTEMS OF THE CELL SURFACE INVOLVED
external H+ on fermentative enzymes. A similar effect can be shown with
a cell-free fermenting system in which the permeability barrier is destroyed,
and also on purified yeast hexokinase in the case of K+ (unpublished data).
Another fermentative reaction susceptible to K+ and NH4+ is phospho-
hexokinase (Muntz, 1947).
This hypothesis is put forward on a tentative basis, with the realization
that the existing data, although consistent with the general nature of the
suggested mechanism, do not allow more than educated guesses concerning
the details.
REFERENCES
BALDWIN, E. (1952). Dynamic Aspects of Biochemisty , 2nd ed. chap. 15. Cambridge
University Press.
BARRON, E. S. G., MUNTZ, J. A. & GASVODA, B. (1948). y. Gen. Physiol. 32, 163.
BERGER, L., SLEIN, M. W., COLOWICK, S. P. & CORI, C. F. (1946). J. Gen. Physiol.
29, 379-
BONNER, O. D., ARGERSINGER, W. J. Jr. & DAVIDSON, A. W. (1952). J. Amer.
Chem. Soc. 74, 1044.
BOOY, H. L. (1940). Rec. Trav. bot. neerl. 37, i.
BROOKS, S. C. (1947). Advanc. Enzymol. 7, i.
BROOKS, S. C. & BROOKS, M. M. (1941). The Permeability of Living Cells, 19.
Protoplasma Monographien.
CONWAY, E. J. (1953). The Biochemistry of Gastric Acid Secretion. Springfield,
111.: Bannerstone House.
CONWAY, E. J. & DOWNEY, M. (19500). Biochem.J. 47, 347.
CONWAY, E. J. & DOWNEY, M. (19506). Biochem.J. 47, 355.
CONWAY, E. J. & O'MALLEY, E. (1946). Biochem.J. 40, 59.
DANIELLI, J. F. (1943). In DAVSON & DANIELLI (below).
DANIELLI, J. F. (1952). Symp. Soc. Exp. Biol. 7, i.
DAVSON, H. & DANIELLI, J. F. (1943). The Permeability of Natural Membranes.
Cambridge University Press ; New York : Macmillan Co.
DOUNCE, A. L. & TIEN Ho LAN (1949). Pharmacology and Toxicology of Uranium
Compounds, i, part 2, ed. Voegtlin and Hodge. McGraw-Hill Book Co., Inc.
EULER, H. V. & HEINTZ, S. (1919). Hoppe-Seyl. Z. 108, 165.
FARMER, S. N. & JONES, D. A. (1942). Nature, Lond., 150, 768.
FISHER, K. C. & STERN, J. R. (1942). J. Cell. Comp. Physiol. 19, 109.
GOODMAN, J. & ROTHSTEIN, A. (1953). In preparation.
GOTTSCHALK, A. (1944). Aust. J. Exp. Biol. Med. Sci. 22, 291.
HAGGLUND, E. & AUGUSTSON, A. M. (1925). Biochem. Z. 155, 334.
HEILBRUNN, L. V. (1952). An Outline of General Physiology. Philadelphia and
London: W. B. Saunders Co.
HEVESY, G. (1948). Radioactive Indicators. New York: Interscience Publishers,
Inc.
H6BER, R. (1945). Physical Chemistry of Cells and Tissues. Toronto (Philad.):
The Blakiston Co.
HOPKINS, R. H. & ROBERTS, R. H. (1935). Biochem.J. 29, 919.
HOTCHKISS, R. D. (1944). Advanc. Enzymol. 4, 153.
HURWITZ, L. (1953). Ph.D. Thesis, Univ. of Rochester, N.Y.
HURWITZ, L. & ROTHSTEIN, A. (1951). J. Cell. Comp. Physiol. 38, 437.
KAMEN, M. D. & SPIEGELMAN, S. (1948). Cold Spr. Harb. Symp. Quant. Biol. 8,
IN THE UPTAKE OF SUGARS BY YEAST 2OI
KUNITZ, M. & MCDONALD, M. R. (1946). y. Gen. Physiol. 29, 393.
LANSING, A. I. (1942). Biol. Bull, Woods Hole, 82, 385.
LASNITZKI, A. & SZORENYI, E. (1935). Biochem.J. 29, 580.
LOWEN, W. K., STOENNER, R. W., ARGERSINGER, W. J. Jr., DAVIDSON, A. W. &
HUME, D. N. (1951). y. Amer. Chem. Soc. 73, 2666.
MAZIA, D. (1940). Cold. Spr. Harb. Symp. Quant. Biol. 8, 195.
MEYERHOF, O. & KAPLAN, A. (1951). Arch. Biochem. 33, 282.
MUNTZ, J. A. (i947). J- Biol. Chem. 171, 653.
MYRBACK, K. & OERTENBLAD, B. (1937). Z. Biochem. 291, 61.
MYRBACK, K. & VASSEUR, E. (1943). Hoppe-Seyl. Z. 277, 171.
NEISH, A. C. & BLACKWOOD, A. C. (1951). Canad.J. Technol. 29, 123.
NEUMAN, W. J., HAVILL, J. R. & FELDMAN, I. (1951). J. Amer. Chem. Soc. 73, 3593.
0RSKOV, S. L. (1945). Acta path, microbiol. scand. 22, 523.
PULVER, R. & VERZAR, F. (1940). Nature, Lond., 145, 823.
ROSENBERG, TH. & WILBRANDT, W. (1952). Enzymatic processes in cell membrane
penetration. Int. Rev. Cytol. i, 65, ed. G. H. Bourne and J. F. Danielli.
New York : Academic Press Inc.
ROTHSTEIN, A. & BERKE, H. (1952). Arch. Biochem. Biophys. 36, 195.
ROTHSTEIN, A. & DEMIS, C. (1953 a). Arch. Biochem. Biophys. 44, 18.
ROTHSTEIN, A. & DEMIS, C. (19536). In preparation.
ROTHSTEIN, A. & ENNS, L. (1946). J. Cell. Comp. Physiol. 28, 231.
ROTHSTEIN, A., FRENKEL, A. & LARRABEE, C. (1948). J. Cell. Comp. Physiol. 32,
261.
ROTHSTEIN, A. & FRENKEL, A. & LARRABEE, C. (19490). U.S. A.E.C. Declassified
Report AECD 2815.
ROTHSTEIN, A. & FRENKEL, A. & LARRABEE, C. (19496). U.R. Report 73.
ROTHSTEIN, A. & HAYES, A. (1953). In preparation.
ROTHSTEIN, A. & LARRABEE, C. (1948). y. Cell. Comp. Physiol. 32, 247.
ROTHSTEIN, A. & MEIER, R. C. (1948). y. Cell. Comp. Physiol. 32, 77.
ROTHSTEIN, A. & MEIER, R. C. (1949). y. Cell. Comp. Physiol. 34, 97.
ROTHSTEIN, A. & MEIER, R. C. (1951). y. Cell. Comp. Physiol. 38, 245.
ROTHSTEIN, A. & MEIER, R. C. (1953). In preparation.
ROTHSTEIN, A., MEIER, R. C. & HURWITZ, L. (1951). y. Cell. Comp. Physiol. 37, 57.
SCHMIDT, G., HECHT, L. & THANHAUSER, S. J. (1949). y. Biol. Chem. 178, 733.
SMITH, H. (1951). The Kidney. Structure and Function in Health and Disease.
Oxford University Press.
SMYTHE, C. V. (1939). Enzymologia, 6, 9.
SPIEGELMAN, S., KAMEN, M. D. & SUSSMAN, M. (1948). Arch. Biochem. 18, 409.
STIER, T. J. B. (1933). J. Gen. Physiol. 16, 815.
SUSSMAN, M., SPIEGELMAN, S. & REINER, J. M. (1947). J. Cell. Comp. Physiol. 29,
149.
WIAME, J. M. (1949). y> Biol. Chem. 178, 919.
WIGGINS, E. H., MANN, P. F. E., TREVELYAN, W. E. & HARRISON, J. S. (1952).
Biochim biophys. Acta, 8, 537.
WILKES, B. G. & PALMER, E. T. (1933)- J. Gen. Physiol. 16, 233.
ZELJ.ER, H. (1926). Biochem. Z. 175, M5-
ACTIVE CATION TRANSPORT
IN ERYTHROCYTES
BY MONTAGUE MAIZELS*
Department of Clinical Pathology, University College Hospital
I. INTRODUCTION
While the muscle cell has been regarded as typical of a living dynamic unit,
the mammalian erythrocyte was for a long time considered to be a senescent
structure specialized for oxygen carriage and subject to simple physical
laws uncomplicated by vital processes. It is easy to understand how this
view arose. The human erythrocyte, though rich in potassium and poor in
sodium, survives for many days in a circulating plasma rich in Na and poor
in K. Clearly, such a situation could only persist if the cell were imper-
meable to cations or if the paradoxical distribution were maintained by an
active transport against the concentration gradients. It seemed improbable,
however, that such a vital process could be attributed to a non-nucleated
cell whose respiration was minimal, and so it was generally assumed that
the anomalous distribution of cations arose as an active process in the
immature nucleated respiring cell and was perpetuated in maturity by the
cell membrane becoming impermeable to cations. In support of this view
was the well-known observation that in short-term experiments mammalian
erythrocytes in solutions of varying tonicities behave as osmometers. In
short, then, the mature erythrocyte was regarded as a dead cell, permeable
to simple anions, but impermeable to cations and to the cell and plasma
proteins. On this basis the experiments of Henderson, Van Slyke and
Gamble were founded, and though the fundamental assumptions were
wrong, their observations conducted under controlled conditions are still
valid and fruitful.
Between 1930 and 1940 several observers described the passive penetra-
tion of erythrocytes by Na (Jeanneney, Servantie & Ringenbach, 1939;
Maizels & Whittaker, 1940) and also the passive escape of K (Duliere, 1931 ;
Drew, Esdall & Scudder, 1939; Downman, Oliver & Young, 1940) and the
observations, if acceptable, would imply as a corollary the existence of an
active process to compensate for the passive cation movements. As, how-
ever, the experiments were carried out in vitro under highly artificial con-
ditions, they were not held to invalidate the view that erythrocytes in vivo,
or when freshly shed, were impermeable to cations. But in 1936, Henriques
* In receipt of a grant for technical assistance from the Nuffield Foundation.
ACTIVE CATION TRANSPORT IN ERYTHROCYTES 203
and 0rskov showed that when lead was injected into rabbits, the erythro-
cytes lost K and then gradually recovered the lost K as the effects of the
injection passed off. The experiment is perhaps not quite conclusive since
it is uncertain how much of the rise of cell K is due to actual uptake during
recovery and how much is contributed by new-formed cells or by cells
derived from reservoirs possibly inaccessible to lead. More conclusive were
the experiments of Cohn & Cohn (1939) and of Maizels & Paterson (1940).
The former showed that 24Na injected into dogs rapidly exchanged with
Na in the erythrocytes, while the latter authors, without using tracers,
demonstrated permeability and active cation transport in the case of human
erythrocytes in the following way. Group O blood was cold-stored until
cell Na had risen to about 60 m.equiv./l., and 1000 ml. were then transfused
to a group A recipient, causing an immediate rise of Na in the circulating
cells from 14 to 32 m.equiv.; 6 hr. later, however, cell Na had fallen to
1 6 m.equiv./l., and since differential agglutination showed the donor cells
to be still surviving, it followed that Na had left the donor cells against the
concentration gradient. Moreover, since loss of Na was unaccompanied by
cell shrinkage it was presumed that there had been a compensatory uptake
of K — also against the gradient. The permeability of rabbit cells to 42K was
demonstrated in vivo by Hevesy & Hahn (1941) and to 24Na by Mullins,
Fenn, Noonan & Haege (1941) and also by Hahn & Hevesy (1942). Mean-
while, an impetus had been given to these studies by Steinbach (1940),
who showed that excised frog's muscle lost K to K-free Ringer solution,
regaining the K against a steep concentration gradient when only a small
amount of K was added to the external medium. A year later (1941)
Harris and also Danowski showed that, in stored blood, erythrocytes
which had in the cold lost K in accordance with the concentration gradient
regained K in vitro if incubated at 37° C. with glucose. Maizels (1949)
confirmed these observations and also showed that Na gained during cold-
storage was excreted during incubation.
Cation transport in erythrocytes has thus been established for just
12 years, and its metabolic basis and energetics subjected to much study.
But in spite of the apparent simplicity of the mammalian erythrocyte, the
nature and cause of cation movements across the cell wall remain quite
obscure.
II. METABOLISM AND TRANSPORT
Both Harris and Danowski showed that glucose was essential to cation
transport and that fluoride inhibited the active movements. Maizels (1951)
showed that active transport was inhibited or abolished by fluoride and
monoiodoacetate in high dilution, but not by poisons acting chiefly on
the respiratory cycle, such as cyanide (10 mM), dinitrophenol (i ITIM) and
204 ACTIVE CATION TRANSPORT IN ERYTHROCYTES
malonate (10 HIM): Table i shows that fluoroacetate (10 mM) is also without
effect; mepacrine and arsenite increase the permeability of erythrocytes
but are probably without direct effect on transport. It follows from what
has been said, that cation transport in human erythrocytes is based on
energy derived from glycolysis and not from respiration: this is to be ex-
pected from the studies of Harrop & Barron (1928) and of Dische (1937)
which identify red-cell metabolism with glycolysis. It is probable that
transport in other mammalian erythrocytes is also based on glycolysis ; this
is so with rabbits and according to McKee, Ormsbee, Anfinsen, Geiman
& Ball (1946) is also the case with monkey cells.
Energizing substrates. Glucose and mannose energize cation transport in
high or low concentrations, while fructose is effective in high concentration
(500 mg./ioo ml.) but not in low (100 mg./ioo ml.) (Maizels, 1951). The
observations are surprising but accord with those of Meyerhof & Geliazkowa
(1947) on the glycolysis of sugars by brain and sarcoma slices: here glucose
and mannose are rapidly glycolysed in high or low concentrations, galactose
is little affected, while fructose is rapidly glycolysed in 2% solution and
only slowly in 0-2% solution. Meyerhof & Geliazkowa (1947) attribute
this to differing affinities of the sugars for hexokinase. Galactose has little
or no ability to energize active cation transport: neither has lactate nor
pyruvate.
The findings in the case of pyruvate may seem to contradict those of
Wilbrandt (1940); but this is not so and the subject requires further con-
sideration. Wilbrandt, using massive amounts of fluoride (20-40 m.M/1.),
found that human erythrocytes shrink and become resistant to haemolysis
by hypotonic solutions. The effect, fully developed in an hour or two, was
presumably due to a rapid loss of K from the cells without corresponding
gain of Na. It was inhibited by pyruvate — possibly through energy
derived from the conversion of pyruvate to lactate. In Maizels's (1951)
experiments much lower concentrations of fluoride were used (1-5-10 mM)
in the presence of glucose: here some cell swelling was usually observed,
though the writer remarks that this increase 'was sometimes less than
would have been expected from the failure of active transport*. In any
case it is clear that transport and glycolysis in human cells are inhibited
by an amount of fluoride which is far less than that required to elicit the
Wilbrandt effect, and this inhibition is not significantly affected by
pyruvate. This finding is illustrated in Table i which also shows that,
although inhibition of transport by fluoride is little if at all affected by
pyruvate, the haemolysis which is constantly present in systems incubated
with large or small amounts of fluoride is inhibited by the addition of
pyruvate and this is observed consistently; lactate has no such effect.
ACTIVE CATION TRANSPORT IN ERYTHROCYTES
2O5
The findings suggest the following conclusions: (i) Small amounts of
fluoride (5 mM) produce maximal inhibition of glycolysis and transport
while having little effect on the permeability of the cell membrane and
such energy as may be liberated by the reduction of pyruvate cannot be
used for cation transport (which may well depend solely on energy-rich
phosphate bonds). The inhibition of transport results in a slow increase of
cell Na and a gradual loss of K extending over many hours. (2) High
concentrations of fluoride do not further increase inhibition of glycolysis
Table i. Cation transport in human erythrocytes : effects of fluoroacetate
and of fluoride and pyruvate
(Blood stored at 4° C. for 7 days and then incubated for 18 hr. External concentrations,
K 10, Na 150 m.equiv./l. glucose 6 mM.)
Cells
No.
Hr.
at
37° C.
Additions mM/1. cell
suspension
Lysis
(%)
V
pH at
20° C.
Contents
(m.equiv./l.
cells)*
Concentra-
tions
(m.equiv./l.
cell water)
K
Na
K
Na
i a
0
(Unincubated)
o
i°5
7-24
66
54
88
72
b
18
None
0
99
7-20
9i
24
132
35
c
18
None
o
101
7-n
89
24
126
34
d
18
Fluoroacetate 10
0
101
7-14
88
26
124
36
^a
0
(Unincubated)
0
103
7-02
52
61
7i
84
b
18
None
0
IOO
6-93
74
34
1 06
49
c
18
NaF4
I'2
106
7-12
46
70
61
93
d
18
NaF 4, pyruvate 28
0-4
104
7-20
47
68
63
92
e
18
NaF 4, lactate 28
i-o
104
7-22
45
70
61
94
F=cell volume as a percentage of the original cell volume.
* Contents corrected for changes in volume by reference to the original cell volume.
or transport, but have an additional direct effect on membrane permeability,
so that most of cell K is lost in an hour or two and is not immediately
replaced by entering Na, whose rate of penetration is much less affected
by the presence of fluoride. (3) The haemolysis found in the presence of
low concentrations of fluoride and the marked increase in permeability to
potassium which accompanies high fluoride concentrations are both
mitigated by the presence of pyruvate, acting either in some simple
physical way or by its reduction to lactate.
Nucleated erythrocytes
This account is based on unpublished work by the author.
Cation distribution in the erythrocytes of the chicken is very similar to
that in man, but active transport in these nucleated cells may be energized
206 ACTIVE CATION TRANSPORT IN ERYTHROCYTES
either by lactate, pyruvate or glucose and when the two former substrates
are added, fluoride and iodoacetate cause only a small decrease in active
transport, any inhibition probably being non-specific and exerted on the
respiratory cycle: for by contrast respiratory poisons like cyanide, carbon
monoxide and dinitrophenol, and even simple deprivation of oxygen inhibit
transport strongly, notwithstanding that glycolysis is active in the presence
of these poisons.
Transport is not greater in the presence of glucose than it is with lactate,
nor is the poisoning effect of cyanide enhanced by the addition of moderate
amounts of fluoride or iodoacetate: both these observations show that
active transport in chicken cells derives little or no immediate energy from
glycolysis.
Cation transport in chicken erythrocytes thus has a respiratory basis, and
in this resembles transport in brain and retina, where Terner, Eggleston &
Krebs (1950) have found the process to be energized by glucose, lactate or
pyruvate, but to require also the presence of glutamate. Thus, if potassium-
depleted retina be incubated in a medium containing glutamate (5 mM/1.)
and glucose, active transport of glutamate occurs, the tissue level rising
from the normal of 6 IBM to 20 mM/kg. ; at the same time there is a move-
ment of an equimolar amount of K into the tissues: in the absence of
glutamate little or no cation transport occurs. In the chicken and man,
however, erythrocytes contain less than 0-5 mM/1. glutamate, which if con-
cerned at all in transport must act in some obscure cyclical fashion and
certainly transport is active even in the complete absence of added
glutamate.
Cation transport in the red cells of the grass-snake is qualitatively
similar to that in the chicken, being based on respiration and not directly
on glycolysis. The case of the tortoise is complicated. At 25° C. some
degree of cation transport may be manifest, but at 37° C. there is a marked
rise of cell Na and a moderate fall of K (both movements with the concen-
tration gradients) and so the cells swell and may even haemolyse (Table 2).
The effects are due to the use of a standard calcium-free medium for
suspending the cells, and the addition of 3-5 mM Ca (probably less would
suffice) greatly decreases loss of K and gain of Na. It is probable that Ca
acts by decreasing cation permeability rather than by enhancing active
transport, for were the action mainly on transport, the effects of calcium
lack would be more marked at 23 than at 37° C. So, too, if cyanide is
added to a suspension of tortoise erythrocytes at 25° C. cations move
passively with the concentration gradient, but if calcium be present as well
the passive penetration is prevented and cell K and Na remain practically
unaltered (Table 2, nos. ia, d and e): here, too, it must be presumed that
ACTIVE CATION TRANSPORT IN ERYTHROCYTES
207
in the cyanide-poisoned systems containing Ca, loss of respiratory activity
and cation transport is offset by decreased permeability to cations. The
effects of calcium on tortoise erythrocytes corresponds with its action on
the erythrocytes of the snapping turtle (Lyman, 1945): here the cells swell
and haemolyse in an artificial medium containing less than 1-7 mM Ca.
Lyman found the phenomenon to be peculiar to the snapping turtle and
absent in the golden-striped, box and marine turtles and also in the
diamond-backed terrapin. In the case of carp's blood, Black & Irving
(1938) have shown that the addition of oxalate promotes haemolysis, while
Table 2. Cation transport in tortoise erythrocytes: effects of temperature,
calcium and cyanide
(Blood stored at 4° C. for 7 days and then incubated. External concentrations, K 10,
Na 160 m.equiv./l. glucose n mM.)
No.
Incubation
Additions
mM/1. cell
suspension
Erythrocytes
PH
at
20° C.
V
Haemo-
lysis
Contents
(m.equiv./l.
cells)*
Concentra-
tions
(m.equiv./l.
cell water)
Temp.
(°C.)
Time
(hr.)
K
Na
K
Na
25
22
*5
37
25
120
16
la
b
c
d
e
f
8
23
23
23
23
37
37
o
3
3
3
3
3
3
(Unincubated)
No addition
CaCla 5
NaCN2
NaCN 2,
CaCla 5
No addition
CaCl2 5
6-90
6-78
6-76
6-65
6-66
6-67
6-72
105
103
IOI
112
103
I78
IOO
0
o
o
0
o
Marked
0
IO2
IO4
106
94
98
53
105
19
16
u
30
18
176
ii
136
142
150
"5
134
36
IS*
F = cell volume as a percentage of the original cell volume.
* Contents corrected for changes in volume by reference to the original cell volume.
fluoride has a similar effect on dogfish blood (Ferguson, Horvath &
Pappenheimer, 1938), and also on the tautog (Tautoga onitis], sea robin
(Prionotus carolinus] and squeteague (Cyonoscion regale) (Hamdi & Fergu-
son, 1940): Ferguson et aL suggested that removal of ionized magnesium
was the cause of haemolysis in the bloods of these various fishes, but it is
probable that Ca is the effective ion; certainly in the case of the tortoise Mg
cannot replace Ca as an inhibitor of haemolysis. In my own experiments,
cation transport in the erythrocytes of the frog, chicken and man were not
affected significantly by the presence or absence of Ca: in the case of the
tortoise, however, Ca is, as has been seen, an essential factor; it cannot be
replaced by Mg, Cd, Ni or most other metals, while by contrast, Ba, Sr
and (oddly) Co are as effective as Ca in preventing cell swelling and
haemolysis.
208 ACTIVE CATION TRANSPORT IN ERYTHROCYTES
So far then, we have encountered four types of cation transport in animal
cells : one type in mammalian erythrocytes based on anaerobic glycolysis,
a second type in chicken and snake erythrocytes based on respiratory
activity and unaffected by Ca, a third type seen in the African tortoise (no
other variety was investigated), the snapping turtle and a variety of teleosts
and elasmobranchs is also probably based on respiration, but requires the
presence of calcium to control permeability; the fourth type of transport,
also aerobic, is seen in brain, retina and other tissues, and requires the
presence not only of oxidizable substrate, but also of glutamate. With
regard to the third type of transport, it is possible that a survey of erythro-
cyte susceptibility to Ca-lack might reveal interesting relationships between
different species.
Sites of cation transport in erythrocytes
In non-nucleated erythrocytes the mechanism for cation transport is
presumably located in the cell membrane because this is the sole site for
those phosphorylations on which transport depends. Since no phosphoryla-
tion occurs at the outer cell face, but only dephosphorylation (Clarkson &
Maizels, 1952) it may be assumed that K transport from without inwards
must be preceded by a physical penetration of the cell surface before the
cation carriers concerned can become effective. In the case of chicken
erythrocytes, it has been seen that transport is based on respiration. But
the cells do not stain with janus green and appear to contain no mitochondria,
though it is possible that the respiratory apparatus is diffusely disposed in
the cell membrane; in any case the actual transport must obviously take
place across the cell membrane, and it is here that cation carriage and
transport must be activated even in those cells where the energizing
enzyme systems can be localized to respiring mitochondria within the cell.
Energy for transport
Raker, Taylor, Weller & Hastings (1950) found that 1-5 mM glucose were
metabolized by i 1. cells in i hr. when the pH of the suspending medium
was 7-5 ; at pH 7-0 the figure was 0-93. Maizels (1951) gives 1-5 mM at cell
pH 7-4 (7-6 external), calculated to be more than ten times greater than is
needed to account for the observed cation transport. It is probable that the
transport depends on the presence of energy-rich phosphate bonds and
there is no evidence in favour of Solomon's (1952) suggestion that the
energy derives from the reduction of pyruvate to lactate in the presence of
coenzyme i . This view is based on a misconception of the Wilbrandt effect
which, as has been shown, is primarily concerned with the effects of
fluoride and pyruvate on cation permeability and not on transport.
ACTIVE CATION TRANSPORT IN ERYTHROCYTES 209
III. THE MOVEMENTS OF CATIONS IN HUMAN BLOOD
Dean (1941) first stressed the idea of a cation pump for muscle. He wrote:
* If potassium and sodium are mixed inside the fiber as free ions, then the
pump that builds up the internal concentration of potassium must be
pumping potassium in or sodium out or both/ He discusses especially
the Na pump, but continues later: 'it makes little difference whether
potassium or sodium is pumped.' In the case of muscle, Krogh (1946)
suggests the existence of a Na pump, but of the erythrocyte he says: 'No
evidence is available to show whether Na or K or only one of these ions
is actively transported, but a K transport appears most likely.' Elsewhere,
however, it was remarked (Maizels, 1949) that the physical results of a large
amount of non-penetrating anion (haemoglobin and organic phosphate)
within the erythrocytes were such as to tend to swelling and rupture — a
tendency which would be accentuated by the presence of an inwardly
acting K pump and which could not be adequately opposed by any purely
physical process acting on the sodium ion. He further pointed out that an
outwardly acting sodium pump, on the other hand, would suffice to over-
come the Donnan effect of non-penetrating cell anion and might even
determine such a backflow of K that at equilibrium the cation distribution
characteristic of the human erythrocyte was attained. Solomon (1952)
criticizes this view on the grounds that at equilibrium Na leaving the cell
equals Na entering and so no free energy is left for K transfer. The observa-
tions, however, are irrelevant, for the cold-stored blood systems considered
by Maizels were not at equilibrium during the early stages of incubation
and with cell Na actually falling at the rate of 2-3 m.equiv./l. per hour
against a steep concentration gradient, a backflow of K is possible. How-
ever, in systems at equilibrium Solomon ignores completely the Donnan
effect of non-penetrating cell anion which, with Na 'fixed' by transport at
a constant low level, must tend to attract K and water instead of Na and
water.
But while Solomon's criticism is not valid, the simple theory of Na
transport with compensatory movement of K is unacceptable on other
grounds (Davson, 1951 ; Harris & Maizels, 1952). Thus, if the distribution
of K were secondary to that of Na and governed by potentials arising from
the presence of non-penetrating cell anion, [K]cell/[K]piasma should equal
[H]cell/[H]piasma and [Cl]piasma/[Cl]cell. Harris & Maizels (1952) have in
fact found good agreement between the hydrogen and chloride ion ratios,
the actual values depending on the pH and not on the metabolic state of
the cells. But the potassium ratio is very different, for in fresh human
blood [Cl]piasma/[Cl]cells=I'4, while [K]cell/[K]piasma==3°- Tne con-
E B S VIII 14
210
ACTIVE CATION TRANSPORT IN ERYTHROCYTES
Table 3. Cation transport in human erythrocytes
(Exp. A: blood stored 7 days at 4° C. and then incubated in solutions containing glucose 8 mM, KC1
lom.equiv./l. and varying proportions of NaCl and LiCl. Exp. B: fresh blood incubated in NaCl-LiCl mixtures.)
Cells
External medium
Exp.
Hr. at
37° C.
V
pHat
20° C.
Contents
(m.equiv./l.)*
Concentrations
(m.equiv./l.
water)
Concentrations
(m.equiv./l. water)
[Na],/[Na],
K Na K
Na
K
Na
Li
Ai
0
101-5
7-11
69-5 j 48 5
97'5
68
10
142
0
0-48
4
98
7-10
7I-5 39
106 58
10
142
0
0-41
lit
98
7-26
75'5
32
112
47
10
142
0
0'34
24
97
7-02
82
24
122
36
10
142
0
0-25
27
100
7'iS
81-5
24
117
34'5
10
142
0
0-24
A2
0
101 5
7-ii
69'5
48-5
97'5
68
10
53
89
1-28
4
96
7-11
68
30-5
103-5
46-5
IO
53
89
0-88
"I
97
7-28
72
i9'5
108
29
10
53
89
0'55
24
98
7-06
72
1 1-2
1 06
16-5
10 53
89
0-31
27
98
7-09
7i
10-5
104-5 i5-5
IO
53
89
0-29
30
99
7-10
70
10-3
102 15
10 53
89
0-28
Bi
0
100
7-40
101
12-0
144 17-2
10
145
0
OT2
6
96
7'27
99'5
n-6
151 17-5
10
145
0
0'12
24
97
7-21
102
II-2
152 16-7
10 145
0
0-12
30
100
7-20
105 io'6
151 15-2
10 145
0
OTI
47
101-5
7-28
103-5 12-6 145 17-9
IO
H5
o
0'12
B2
0
IOO
7-40
101 12-0
144 17-2
10
70
75
0'24
6
96
7-27
96-5 1 7-7
146 n-6
10
70
75
0-17
24
101
7-19
90-5
TS
128 10-6
10
70
75
0-15
30
101
7-18
89-5
5'6
126 7-9
IO
70
75
o-ii
47
103-5 7-26
88-5
6-6
120 (/O
10
70
75
0-13
Exp. Ai, &out = o-ii hr."1; kin =0-027 hr.'1. Exp. A2, fcout^o-13 hr."1, Al
F^cell volume as a percentage of the original cell volume.
* Contents corrected for changes in volume by reference to the original cell volume. Note: cell contents and
concentrations corrected for the Na content of intercellular fluid.
Table 4. Cation movements in the cells of human blood stored at 4° C. for
i week with LiCl 135 and Na 15 m.equiv.ll. water and then incubated in
mixtures of KCl and LiCl containing glucose 6mMfor 18 hr.
(Control blood (Exp. 2) stored at 4° C. in NaCl 150 m.equiv./l. water and incubated with KCl and NaCl.)
Cells
Incubation medium
No.
Cold
stored
in
Hr.
at
37° C.
Contents
pH at (m.equiv./l. cells)*
0 /-I
Concentrations Concentrations
(m.cquiv./l. water) (m.equiv./l. water)
20 C.
K
Na
Li
K Na Li K
Na Li
i a
LiCl
0
6-96 73
8
33
IOO 11 46 —
_ I _
b
LiCl
18 7-16 64
2-6
S2
94 4 70 ' o
o | 160
c
LiCl 1 8
7-16 63
2-4 52
93 3*5 , 76 10
o 150
d
LiCl 18 j 7-07 69
2-2 45
102 3 66 , 22
0
138
e
LiCl 1 8
7-03 , 81
2'4 ' 37
112 3'5 52 83
0
77
f
LiCl
18
7-06 , 90
2'5
28
129 3-5 40 117
0
43
8 LiCl
18
7-01 92
2'3
24
U2 j 3'5 j 35 l6°
0
o
20
NaCl 10 7-02 52
61
0
7I i 84 | o — —
—
b NaCl
1 8 6-93 74
34 o
105 ; 49 | o 10
150
o
Contents corrected for changes in volume by reference to the original cell volume.
ACTIVE CATION TRANSPORT IN ERYTHROCYTES 211
elusion is thus enforced that cation distribution in blood involves a trans-
porting mechanism for K in addition to that for Na. There is no evidence,
however, that the two mechanisms are independent, and it is more likely
that they are closely integrated: thus, in sodium-depleted erythrocytes
where further net output of Na is limited, little uptake of K is manifest
(Table 4); so, too, when cells are incubated in media of exceptionally low
K content, significant uptake of K is no longer possible, and at the same
time output of Na ceases to be apparent. These matters are discussed later :
in the meantime it is interesting to note that though a simple Na trans-
porting device acting alone might suffice to give blood its characteristic
distribution of Na and K, the cells in the absence of a simultaneous active
transport of K would be so acid and so lacking in chloride and bicarbonate
ions as to function imperfectly as units for buffering and oxygen carriage
(Harris & Maizels, 1952).
Sodium
In what follows, the word transport' will be used for movements
against gradients of concentration and potential, * transfer* will be used for
movements in any direction and ' diffusion* for purely passive movements.
Transfer constants may be obtained most accurately by means of tracers,
or in the case of Na the * chemical method' of Harris & Maizels (1951,
1952), which requires no radio-sodium, may be used. In the latter type of
experiment, blood is first cold-stored and then incubated with glucose so
that the decline in cell Na concentration may be estimated : the volume of
the external phase is always very large, so that for practical purposes its
cation concentration may be regarded as constant throughout the experi-
ment. In either case (but especially in the * chemical method*) error is
introduced by decreased permeability which occurs on incubating erythro-
cytes (Sheppard, Martin & Beyl, 1951; Harris & Prankerd, 1953), and
by failure of metabolism resulting from dephosphorylation. As a result, the
curve for log 24[Na] against time is only linear for a few hours, tending
to ' flatten* thereafter and to give the impression that only part of cell Na is
freely exchangeable. According to data obtained in vitro at 37° C. by
Solomon (1952) about 3 m.equiv. cell Na are unexchangeable or exchanged
very slowly. Solomon contrasts this with his own findings at lower tem-
peratures and with the results in vivo of tracer experiments communicated
to him personally by Edelman, James & Moore, where in either case cell Na
appeared to be fully exchangeable. Solomon concludes : ' the present results
showing a more slowly exchangeable Na fraction in vitro must be accepted
with reserve and are certainly not indicative of the true state in nature.*
Using both the chemical and tracer methods for human erythrocytes in
14-2
212 ACTIVE CATION TRANSPORT IN ERYTHROCYTES
phosphate-Nad media at 37° C., Harris & Maizels (195 1) obtained kout (the
outward transfer constant) as 0*25 and kin (the inward transfer constant) as
0-023 hr."1. Solomon (1952), using a NaHCO3-NaCl medium, obtained
-kout 0-6 hr."1 and kin 0-022 hr."1. Solomon's higher figures were due in
part to his allowance for the slowly exchangeable Na fraction in the cells,
but also to the superiority of his medium, for with a medium similar to
Solomon's figures of 0-3 hr."1 have been obtained for &0ut. Even so, results
for the transfer constants of Na obtained with the phosphate-NaCl medium
have a relative, if not an absolute, value. Certain other findings will now
be reviewed.
The ratio of cell and external sodium concentrations. The investigation is
complicated by the necessity for adding substances like lithium chloride or
sucrose when lowering the external sodium concentration, if hypotonicity
is to be avoided. Flynn & Maizels (1949) cold-stored blood for several days
so that cell Na was high, and then measured the fall of cell Na on incubating
the cells in relatively large volumes of solutions containing glucose and KC1
(10 m.equiv./l.), together with NaCl and LiCl in varying proportions: they
found that after about 24 hr. the ratio of cell to external sodium concentra-
tions tended to be constant. Harris & Maizels (1951) prepared time curves
for Na output on incubation and confirmed by extrapolation to t— oo, that
[Na]In/[Na]0ut was approximately constant as was k0utlkin. The individual
values of the constants may vary, though in fact in actively transporting
bloods kout and kin are both little affected by variations in external Na. The
findings conflict with those of Solomon (1952) who reports that lithium has
no effect on K transfer, but may decrease Na transfer by 30% or more. It
may be noted, however, that the absolute changes effected by Li in Solomon's
transfer figures are in fact quite small: /eout alters from 0-319 to 0-274 hr."1
and kln from 0-00874 to 0-0102 hr."1. The original experiments on the Na
output of stored cells have therefore been repeated (Table 3): it will be
seen that the transfer constants of Na are little affected by the presence of
Li, whose entry into the cell nevertheless decreases by about 10 m.equiv./l. ,
the final level attained by K in the cells. So, too, when external Na is
decreased and tonicity is maintained by use of K, [Na]J[Na]e still tends to
constancy (Table 5).
External potassium concentration and sodium transfer. Variations of
external K between 14 and 4 m.equiv./l. (Solomon, 1952) and even down
to 2 m.equiv. (Harris & Maizels, 1951) have little effect on the transfer
constants for Na, but with [K]f below i m.equiv. both constants and
especially &out fall, the Na efflux being depressed, and it may well be, as
Flynn & Maizels (1949) suggest, that in the theoretical (but hitherto
unattained) K-free system, Na efflux would cease altogether.
ACTIVE CATION TRANSPORT IN ERYTHROCYTES
213
pH and cell sodium. Between pH 7-2 and 7-6 (cells) output of Na from
stored cells is maintained at a steady level (Flynn & Maizels, 1949). Below
pH 7-1 the manifest output falls, presumably because of decreased efflux,
while above pH 7-7 manifest output is also decreased, probably as a result
of increased influx: the value of the latter observations is uncertain because
of the haemolysis which becomes apparent as cell pH exceeds 7-7.
Lithium
Data from Table 4 (where K and Na were estimated chemically and Li
with a flame photometer) suggests that in sodium-poor cell suspensions at
37° C. &in and &out for Li are, over a period of 18 hr., about 0-016 hr.-1.
Similarly, when cells depleted of Na by cold-storage in LiCl solutions
were incubated in glucose-containing media rich in Na and Li, but poor
in K, the inward transfer constant determined from direct measurement of
Li was found to be about 0-022 hr."1 (Table 5). Without attributing any
great degree of accuracy to these figures, they do suggest that the passive
transfer rate for Li is of the same order as those of Na or K. The data also
show that there is no active efflux of Li comparable to that occurring with
Na. As in the case of Na, however, the rate of passive influx is not constant
throughout the incubation period, but decreases with time. Thus in
Table 3, the concentrations of Na + K in Exp. Bi (where total base con-
centration is fairly constant throughout) compared with the concen-
trations of Na + K in Exp. 62, suggest that in the latter the concen-
trations of Li at o, 24 and 47 hr. are about o, 23 and 32 m.equiv./l. cell
water respectively. Hence, the transfer constant for influx in the first
24 hr. is about 0-017 hr."1, while between 24 and 47 hr. it is about
0-007 hr."1. Other experiments of this type show similar reductions of Li
influx with time.
Table 5
Cell Li (m.equiv./l.)
External concentration
Transfer constant
for Li influx
at 37° C.
At o hr.
At 1 8 hr.
K
Na
Li
27
28
29
26
36
43
10
10
10
120
95
70
25
50
75
0-024
0'020
Potassium
According to Dean, Noonan, Haege & Fenn (1940) about 1-4% of cell
K exchanges per hour; Raker et al. (1950) give 1-6% at 37° C. and
Sheppard & Martin (1950) 1-8% at 38° C.
External sodium and cell potassium. When external Na is lowered, tonicity
being maintained by LiCl, uptake of K by cells incubated after cold-
214 ACTIVE CATION TRANSPORT IN ERYTHROCYTES
storage is decreased: this is shown in Table 3. In five similar experiments
reduction of external Na from 140 to 50 m.equiv. (with increase of external
Li from o to 80 m.equiv.) decreased the average gain of K by 62%, while
it increased the loss of Na by 43 % ; similar figures are reported by Ponder
If it be assumed that cell K in Table 3 is approaching equilibrium in
24-30 hr., then the data for stored cells suggests that when external Na is
lowered from 142 to 53 m.equiv./l. K influx falls by about 15 % ; in the case
of the fresh cells in Exp. B, decrease of [Na]e from 145 to 70 m.equiv. lowers
the equilibrium value of [K]z- (at 30 hr.) from 151 to 126 m.equiv./l. cell
water; assuming that the rate constant for K efflux remains unaltered or
changes similarly in the KCl-NaCl and KCl-NaCl-LiCl systems during
incubation, then halving [Na]e reduces K influx by about 16%.
Cell sodium and cell potassium. Flynn & Maizels (1949) remarked that if
conditions during cold-storage (e.g. in a LiCl medium) were such as to
cause cell Na to fall to a low level, little further active output could occur
during the subsequent incubation, and under these circumstances there is
practically no active uptake of K. Thus in Table 4 of the present paper,
where Na output during incubation is only about 5 m.equiv./l. cells, uptake of
K does not exceed this figure so long as [K]e lies between o and 22 m.equiv.
and it is not until [K]e approaches 80 m.equiv. that definite increase in
[K]^ occurs. It is thus clear that when Na efflux falls, K influx is also
decreased. The findings contrast with those of the control experiment
(Table 4, no. 2), where cells from the same blood cold-stored and incubated
in Na-rich media show very active movements both of Na and K. It should
be noted that decrease in K transport can only be demonstrated when both
erythrocytes and suspending medium are depleted of Na before incubation:
if the medium be Na-poor and the cells Na-rich, output of Na and uptake
of K will still occur, and this has led Ponder (1950) to state that K transport
remains active in ' LiCl and CsCl systems ' . In short, Ponder's experimental
conditions correspond to those already described in the preceding section.
It has previously been noted that Na transport depends on an adequate
amount of K in the external medium and the present argument shows the
correlation of K transport with Na efflux: thus the integration of the
various cation movements is close.
Concentration of cell potassium and the magnitude of the potassium influx.
Under normal conditions K influx amounts to about 1-6 m.equiv./l. cells
hr.-1 at 37° C. (Raker et al. 1950; Sheppard & Martin, 1950; Solomon,
1952), but when stored blood with K-depleted cells is incubated with
glucose the influx in active preparations may exceed 2 or even 3 m.equiv.
hr."1 (for examples see Flynn & Maizels, 1949 ; Ponder, 1950). Such a high
ACTIVE CATION TRANSPORT IN ERYTHROCYTES 215
rate is only maintained for a few hours, the influx tending to fall as cell K
rises and perhaps also as a result of metabolic failure. Hence, in our experi-
ments over a period of 18 hr. incubation, increase of cell K was about
1-2 m.equiv. hr."1, averaging 1*4 for sixteen stored bloods with a mean
cell K content of 62 m.equiv./l. at the end of cold-storage and i-o for
thirteen bloods whose mean cell K was 75 m.equiv. ; it thus appears that
the lower cell K is at the beginning, the greater the rise during incubation.
External potassium and cell potassium. In the metabolizing erythrocyte
increase in the external K concentration, [K]e, might affect cell K by
increasing adsorption, or else by displacing cell Na, or by increasing the
passive entry of K into the cell, or by affecting the active K influx. The
effect of K adsorbed is very slight. Cells suspended in simple KC1 solution
(0-175 M) gain about 6 m.equiv./l. K almost immediately, subsequent gain
being much slower (Maizels, 1935): assuming that this quick gain is due to
adsorption and allowing for K in the intercellular fluid, this gives a probable
figure for K adsorbed of about 2 m.equiv. Hence, with external K at
75 m.equiv./l. this immediate rise, attributed to adsorption, should be very
small: experimentally, the net value was found to lie between 0-5 and
i m.equiv./l. cells.
Displacement of cell Na is more significant. Since, as we have seen,
[NaJJfNa]^ is constant, rise of [K]e from 5 to 75 m.equiv., with fall of [Na],,
from 140 to 70 m.equiv./l. will about halve the concentration and also the
content of cell Na (Tables 3 and 6) ; assuming that K influx is linked with
and equal to about half of Na efflux, increase of cell K directly due to
displacement of 4-8 m.equiv. Na, will be about 2-4 m.equiv./l. cells in
24 hr.
Passive influx is negligible under physiological conditions with plasma K
set at the low level of 5 m.equiv./l.; so, too, passive efflux of Na may be
ignored when cell Na is only about 12 or 15 m.equiv./l. But in artificial
systems with [K]e or [Na]^ raised to a high level, passive influx of K or
efflux of Na become significant. This aspect is usually ignored, with the
implication that the whole of K entry into and Na exit from the meta-
bolizing cell is active.
It is true that there is no direct evidence of passive K penetration into
normal erythrocytes, for though cells depleted of K during cold-storage
and then incubated in glucose-free K-rich media permit the passive entry
of K at a rate comparable to passive K efflux from the metabolizing cells
(0-01-0-02 hr."1), this does not prove that there is a passive entry of K
under physiological conditions. But since in the actively metabolizing cell
Li enters and leaves at about the same rate, while passive penetration of
Na is also free, it seems likely that under these conditions passive influx of
2l6
ACTIVE CATION TRANSPORT IN ERYTHROCYTES
K also occurs; indeed, it is unlikely that a membrane passively permeated
by an ion in one direction, would be impermeable in the reverse direction.
Hence, it may be assumed that the rate constant for passive influx of K(R)
^quals that for passive K efflux (£"), provided that allowance is made for the
Donnan asymmetry. The passive influx may then be calculated from
Harris's (1953) equation, R =f2 x E x [K]e/[K]t-, where [K]e and [K]t. are
Table 6. Effect of external potassium concentration on the potassium
content and concentration of human erythrocytes
(Fresh blood incubated in mixtures of KC1 and NaCl containing glucose.)
Cells
External
medium.
Concentra~
Exp.
Buffer
Hr.
at
37° C.
V
pH
at
Contents
(m.equiv./l.)*
Concentrations
(m.equiv./l.
water)
tions
(m.equiv./l.
water)
20° C.
K
Na
K
Na
K
Na
i
Phosphate
0
100
7'3i
96
15
138
21
5
150
0
IOO
7'35
97
14
140
20
75
80
24
96
7'37
95
H
144
21
2
i53
24
97
7-27
97
ii
H5
16
5
150
24
98
7'25
102-5
9'5
151
H
25
130
24
IOI
7-23
112
6'i
157
8-5
75
80
2
NaHCO2
o
IOO
7-36
98
n-9
140
17
5
150
0
IOO
7-40
99
11-3
142
16-2
75
80
24
97
7'49
98
12-8
146
19
5
150
24
102
7'57
116
6-9
1 60
9'5
75
80
3
NaHCO2
o
IOO
7-40
IOI
n-i
145
i5'8
5
150
0
IOO
7'44
101-5
9-8
146
14-0
76
80
24
98
7'45
101-5
9'5
149
14-0
5
150
24
104
7*45
117-5
5'7
159
77
75
80
F=cell volume as a percentage of the original cell volume.
* Contents corrected for changes in volume by reference to the original cell volume.
Note: cell contents and concentrations corrected for Na and K contents of the inter-
cellular fluid.
the respective external and internal concentrations and/ is the asymmetry
factor of Harris & Maizels (1952) which equals [Cl]e/[Cl]< or about 1-14 at
the usual experimental pH. Thus, with [K]e at 5 m.equiv./l. it may be
calculated that the active moiety of influx is 1-54 and the passive 0-06 per 1.
cells hr."1; giving the observed total influx of r6 m.equiv., at which level
the K content of normal cells remains constant with a rate for outward
transfer of 0-016 hr.-1. But with [K]e at 75 m.equiv. K influx should rise
from 1-52 + 0-08 m.equiv. to 1-52+1-1 or 2-62 m.equiv. hr.-1, and cell K
should rise by about 22 % during the 24 hr. incubation. Observed increases
are shown in Table 6, and after making the appropriate deductions for K
adsorbed and K displacing Na, the total net influx in 24 hr. is about
ACTIVE CATION TRANSPORT IN ERYTHROCYTES
217
13 m.equiv., corresponding to a reduction in the active component of influx
of about 25 %, when [K]e rises to 75 m.equiv./l. and [Na]e falls from 145 to
75 m.equiv. : it has already been seen that if [Na]e is similarly reduced by
the addition of LiCl (instead of KC1), active influx of K suffers a similar
but smaller reduction of about 15 %.
These findings may now be compared with those of other workers.
Flynn & Maizels (1949) found that when stored cells were incubated in
media containing glucose the rise in cell K was rather greater with [K]e at
25 than at 5 m.equiv./l.; owing to increase in cell volume, however, the
increase in cell K concentration was relatively less: this is also evident in
Table 6 of the present paper. Davidsen & Kjerulf- Jensen (1950) also
report increased K influx with rise of [K]e. Solomon (1952), on the other
hand, working with a range of [K]e between 4-5 and 16-75 m.equiv./l.
states that (total) K influx is unaffected by external K concentration, though
his results show a reduction in total influx in three of five experiments ; in
one case rise of [K]e from 9-35 to 11-35 m.equiv. lowered the transfer rate
for K by 22 % : it should perhaps be noted that most investigators of the
effects of [K]e on influx have limited the external concentration of K to
between 5 and 20 m.equiv., which could in any case cause little rise in total
influx — certainly not more than 15%. Indeed, the only evidence that K
influx is unaffected by very large rises in [K]e, is limited to a single record
in the paper by Raker et al. (1950). The table below is modified from their
table 6 and the figures in columns i and 2 are from Raker's and his co-
workers' own data:
Table 7
[K],
m.equiv./l.
mM K exchanging per 1.
cells hr."1 at 37° C.
(total influx)
Passive
influx
Active
influx
4*47
38-5
74'4
1-67
i'33
i-58
0-07
0-62
1-19
i -60
0-71
0-39
The data evoke the following comments : comparison of the figures for
total influx with those for [K]e suggests that the experimental error must
be high, for increase of [K]e from 4*47 to 38-5 m.equiv. appears to
decrease total K influx by 20% and active K influx by 50%, while increase
of [K]e from 4*47 to 74-4 m.equiv./l. seems to depress total influx by only
6%, while lowering the true active influx by no less than 75%. In the
absence of further supporting data and in view of the findings in this paper
and in that of Davidsen and Kjerulf- Jensen it would seem reasonable to
conclude that with marked rise in the external K concentration there is
some small depression of true active K influx, and as a result total K influx
2l8 ACTIVE CATION TRANSPORT IN ERYTHROCYTES
is a little less than would be expected from the data for passive K permea-
tion. This matter is considered again later.
pH and potassium transfer. According to Raker et al. (1950) change of
external pH between 7 and 7-7 leaves K influx unaffected, but as the par-
ticular system beginning at pH 7-7 ended at pH 7-2, the significance of the
range indicated is doubtful, and a range of 7-7*5 is more probable. Flynn
& Maizels (1949) found net uptake of K during the incubation of stored
cells to vary rather little for a range of cell pH between 7 and 7-6, though
there was a definite maximum at pH 7-3 (pH readings at 20° C.); Bonder's
(1950) figures are similar.
Temperature coefficients
According to Solomon (1952) the apparent energy of activation for K
transfer is about 12,300 calories/mole, and he derives the following values
from the respective data of Raker et al. (1950), Sheppard & Martin (1950)
and Ponder (1950); 14,500, 15,800 and 16,200. In the case of Na transfer
across the erythrocyte membrane, Solomon gives 20,000 for the activation
energy of influx and 15,000 for efflux, from which it follows that cell Na
concentration must be less at 25° C. than at 37° C. Actually, the reverse
is the case, and in a recent experiment with stored blood [Na]^ fell from 65
to 33 m.equiv./l. cell water after 24 hr. incubation at 37° C. and only to
50 m.equiv. at 25° C.; it follows that the activation energy for Na efflux
must in fact be greater than for influx. According to Harris (1953), the
activation energy for the passive fluxes of Na and K equal 14,000 cal./mole,
while the value for the active fluxes of Na and K is nearly twice as great.
Ponder (1950) and also Solomon (1952) remark that the energy of activation
for K transport and for glycolysis is similar and they imply that the corre-
spondence arises from the interrelation of transport and glycolysis. This
may well be true for active cation transport, but the high energy of activa-
tion for passive movements does not necessarily suggest that cations are here
also combined with chemical carriers, and the high energy of activation
may well be a physical result of the difficulty with which cations penetrate
the cell wall barrier (see Danielli & Davson, 1934). The matter is of impor-
tance because there is still a tendency to ascribe any reaction with a high
activation energy to chemical processes. Thus, Solomon (1952) states that
'the high temperature coefficient observed by Gourley & Gemmill (1950)
for phosphate transfer is typical of a metabolically linked process rather
than simple diffusion' — a view which Gourley & Gemmill themselves
advance. In the case of erythrocytes suspended in pure isotonic phosphate
solutions the activation energy calculated from Maizels's (1932) data is
16,000 — a figure which holds both at pH 8 and 5-4. This independence of
ACTIVE CATION TRANSPORT IN ERYTHROCYTES 2IQ
pH suggests that neither phosphorylation nor dephosphorylation are major
factors in the penetration of phosphate from pure phosphate media, and
this view is supported by the speed of entry at pH 5-4 which with an
external phosphate concentration of 130 mM amounts to 21 mM/1. cells per
min. at 37° C. With erythrocytes in 6-9% glucose entry may exceed
30 mM glucose per min., a rate which would seem to exclude a chemical
basis for glucose transfer; the energy of activation was 18,000; Masing's
(1914) figure was 14,000. In the case of ox cells, the activation energy for
penetration of polyhydric alcohols may rise as high as 23,000 (Jacobs,
Glassman & Parpart, 1935). Hence, it is necessary to be cautious when
ascribing a chemical basis to a phenomenon on account of its high energy
of activation.
Transfer constants of various species
Most of these observations were made between 1939 and 1942, and in
view of complicating factors such as slowly exchanging Na (Sheppard
et al. 1951) and K (Hevesy & Hahn, 1941) their significance is uncertain.
However, data collected from various sources by Sheppard et al. shows that
in the cow and sheep, as in man, Na penetrates the erythrocyte membrane
more readily than K. So too, in the case of dog erythrocytes, Krogh (1946),
using the data of Cohn & Cohn (1939) and of Hahn & Hevesy (1942),
showed that here also Na penetrates more readily than K.
In the case of chicken erythrocytes, the transfer constants of Na at 25° C.
(measured by the chemical method of Harris & Maizels, 1951) are as
follows: at pH 6-8, &0ut = o-45; atpH 7*2-7-75, 0-7 and at pH 7-7, 0-6 hr."1:
£0iit for human cells in a similar medium (NaCl solution buffered with
phosphate) was 0-06 at pH 7-3 and 25° C. £ln for chicken cells at 25° C.
varied between 0-03 and 0-05 (four experiments). Taking an average figure
for chicken erythrocyte volume and area of I3O//3 and i8o/^2, the perme-
ability constants are found to be for efflux 4-9 x io~5 and for influx
2-0 x io~6/cm. hr."1.
IV. GENERAL CONSIDERATION OF CATION TRANSPORT
IN HUMAN ERYTHROCYTES
The cell wall barrier
It has been seen that in the metabolizing human erythrocyte the active
and passive fluxes for Na exceed the corresponding K fluxes; the same
holds for the erythrocyte of the dog and for other, if not all, mammalian
erythrocytes. So, too, in the case of the non-metabolizing human cell
(kept at 4° C. or incubated in the absence of glucose) Na penetrates more
rapidly than K, and the naturally K-rich cells suspended in a Na-rich
22O ACTIVE CATION TRANSPORT IN ERYTHROCYTES
medium swell. This last observation is complementary to that of Davson &
Reiner (1942) on the cat: cat erythrocytes are naturally poor in K and rich
in Na and if suspended in KC1 solutions show an excess of Na lost over K
gained. So, too, when tortoise cells are suspended in a calcium-free NaCl
solution, gain of Na is thrice as fast as the simultaneous loss of K.
Since of the two hydrated ions Na is the larger, it is likely that Na and K
penetrate unhydrated, presumably through a non-watery lipoid phase and
possibly combined with lipoid soluble carriers. This possibility has been
advanced by Davson & Reiner (1942) and by Solomon (1952). It is
interesting to recall that according to Teorell (1952) erythrocyte ghosts offer
considerable resistance to the passage of cations, a resistance which is much
decreased by the addition of oleate; nevertheless, even in the absence of
oleate, K appears to penetrate more quickly than Na. Teorell considers that
the cation penetration of ghosts is compatible with passage in a watery
solution through a positively charged membrane, enhancement of passage
by oleate being due to a decrease of positive charge. It has already been
seen that in certain circumstances K may cross the membrane of the intact
human erythrocyte more rapidly than Na; this occurs in the presence of
lead or of high concentrations of fluoride, and possibly these poisons effect
a phase reversal similar to that which may occur in the membrane of the
erythrocyte after haemolysis.
Cation carriage
Before proceeding with the discussion of this subject, it will be well to
recall the phenomena characteristic of transfer in the human erythrocyte.
(i) The rate of passive transfer of Na is greater than that of K, while
active transfer of Na also exceeds that of K. (2) During cold-storage, cell
Na, total base, water and volume all increase, while during subsequent
incubation with glucose all these changes are reversed and cell Na, total
base, water and volume all decrease. The greater the increase in base at
4° C., the greater is the decrease at 37° C. (3) In cell suspensions, the ratio
of the Na concentrations in cells and suspending medium is constant and
independent of the external Na concentration (provided that [K]e exceeds
a certain minimum critical value. (4) When external K is reduced below
this critical value (1-2 m.equiv./l.) manifest output of Na and uptake of K
fail. (5) As external K is raised above the critical value, further increase in
Na efflux and K influx becomes relatively small and since with high [K]e
some cell swelling occurs, the concentration of cell K is even less affected.
(6) When, as a result of cold-storage in a Na-free medium, cell Na falls to
a very low level, output of Na during subsequent incubation is necessarily
limited and in these circumstances uptake of K is much reduced. (7) In
ACTIVE CATION TRANSPORT IN ERYTHROCYTES 221
cells cold-stored in a Na-rich medium, Na is high and K low; when such
cells are incubated Na efflux and K influx both increase and may be twice
as great as in cells whose composition is normal. Of these observations,
Krogh (1946) especially has emphasized the first; the last was derived by
Harris & Maizels (1952) from data presented by Flynn & Maizels (1949)
and Ponder (1950). The remaining observations were made by Flynn &
Maizels (1949) as a result of direct chemical analysis and those relating to
Na were confirmed by Harris & Maizels (1951) using tracers.
It may be said at once that the theory recently applied to frog's muscle
by Ling (1952) is not applicable to cation transport in erythrocytes. Ling
supposes that in the presence of certain energy-rich phosphate compounds,
the ability of myosin to adsorb ions is much increased, K then being
preferentially adsorbed for purely physical reasons. There is thus a transfer
of K from one phase, the extracellular medium, to a second phase, the
muscle cell which is considered to present a network of interfaces. Outward
transport of Na is not a feature of Ling's theory of cation transport in
muscle. Such a device, however, is not applicable to the erythrocyte,
where Na as well as K is actively transported and where K transport occurs
from one inert phase — the plasma, across a second phase — the cell mem-
brane, where cation movements are energized, to a third phase — the cell
interior where catabolic but no anabolic activities occur. Moreover, there
is in the erythrocyte no protein which could fill the role taken in muscle
by myosin ; stromatin, the only possible candidate constitutes but i % of
the dried weight of the human erythrocyte. Hence, most workers in this
field adhere to the view that Na and K are transported in complex com-
bination with a carrier, presumably lipoid or lipoid soluble. Thus, if the
complex XC of the carrier X with the cation C is broken down at only one
face of the cell membrane, the concentration gradient for XC will carry
the complex to that face, irrespective of the concentration of C on either
side of the membrane and a means is afforded for the transport of C.
Flynn & Maizels (1949) originally suggested that in the maintenance of
constancy of composition and volume of the erythrocyte, Na transport was
the dominant factor, basing this view on the observation that in media
whose composition approximates to that of plasma the manifest entry
of K into the metabolizing cell never exceeds, and is often less than,
the corresponding exit of Na ; that when Na efflux is limited by experimental
conditions K influx also falls ; and that any cell containing non-penetrating
anions such as protein and organic phosphate and unprotected by an external
or internal resistant structure, must inevitably rupture unless a Na-excre-
tion mechanism exists: the need of a device for the active uptake of K is not
inherent. However, in view of the fact that active K influx as well as active
222 ACTIVE CATION TRANSPORT IN ERYTHROCYTES
Na efflux must occur, Harris & Maizels (1952), while retaining the idea of
the dominance of Na transport, suggested that ' inward transport of K is
" geared " to outward Na transport by the use of a common carrier'. If this
t>e so, the gear ratio must equal 1/2 since about 1-6 m.equiv. K/l. hr."1 are
transported into the cells for about 3-2 m.equiv. Na transported out.
The apparent reciprocity between Na and K transport is further shown
by the observations of Flynn & Maizels (1949) that lack of K in a medium
containing suspended cells decreased the output of Na by the cells they
attributed this effect to an increased passive influx of Na resulting from an
insufficiency of K in the external medium to satisfy the physical require-
ments of the system : it was also suggested that the output of Na ' is in some
way potentiated by the presence of K in the plasma and that in the
theoretical but hitherto unattained K-free medium, Na efflux would cease
altogether'. So, too, Hodgkin & Keynes (1953) remark that lack of K in
a medium containing axons of Sepia reduces the efflux of Na and suggest
that 'there may be a coupling between K influx and Na outflux'.
Harris & Maizels (1952), in their further discussion of transport in
erythrocytes, continue : ' If the turn-over rate of K carrying groups were
sufficiently low, this rate and not the external K concentration would
become the rate-determining factor. Moreover, if the same groups carried
Na out of the erythrocytes, the occurrence of a high Na efflux by bringing
more groups to the exterior, could facilitate the increased K influx. ' This
theory is compatible with many of the known facts of cation transport.
Thus, during cold-storage the Donnan asymmetry and the greater speed
of passive Na movements compared with those of K would increase cell
base and volume, while subsequent incubation in glucose containing media
would restore active transport and with a K/Na ratio of \ would cause Na
loss to exceed K gain and so lead to shrinking of the cell, the change being
opposed to some extent by the Donnan asymmetry. Moreover, increased
Na efflux from the Na-rich stored cell, by bringing more K-carrying
groups to the exterior, would hasten K entry and cause the rate of influx
to rise above that characteristic of the normal cell. So, too, if a cell is
depleted of Na during cold-storage, efflux during incubation will be small,
fewer carriers will be brought to the external face of the cell membrane
and K influx will fail. On the other hand, one must presume that with cells
in a K-free medium, return of carriers from the outer to the inner face of
the cell membrane is restricted, and this may be one factor in the decreased
Na efflux from cells suspended in media with very low K concentrations.
Again, if the turn-over rate of K carriers were sufficiently slow, the
failure of K influx to alter appreciably when the external K concentration
([K]c) was raised from 5 to 15 m.equiv./l. would be explained. But such
ACTIVE CATION TRANSPORT IN ERYTHROCYTES 223
an explanation could hardly apply to systems where [K]e had been raised
to 75 m.equiv./l., for with active influx unaltered, passive influx now
becomes significant and should raise the total influx to about 2'6 m.equiv./l.
cells hr."1. E. J. Harris (this symposium) explains the observed depression
of the active component of K influx when [K]e is raised, by an ingenious use
of the theory of linked carriage. He observes that when [K]e is raised from
5 to 75 m.equiv./l. [Na]e is necessarily halved so that at equilibrium Na
influx and hence efflux must also be halved. Applying Harris's explanation
to existing data leads to the conclusion that the active component of
K influx must fall from i *6 to 0*8 m.equiv./l. cells hr."1, which with a passive
component of K influx at i-i gives a total influx of 1-9 m.equiv. hr.-1,
sufficient to raise cell K by about 6 m.equiv. in 24 hr. The suggestion that
active K influx decreases, while passive and total K influx increase with
rise of [K]e throws doubt on the experiments of Raker et al. (1950),
Sheppard et al. (1951) and of Solomon (1952) which seem to show that
total K influx is independent of [K]e. On the other hand, it does not
wholely agree with the findings in the present paper where cells in a
medium containing 75 m.equiv./l. K show a net increase of K content
(corrected for extra K adsorbed in the K-rich medium and for uptake of K
associated with a falling cell Na) of about 13 m.equiv./l. cells: this corre-
sponds to a depression of the active component of influx by about 25 and
not by 50%; though there is an increase in total influx of 40%. When
[Na]e is halved by the addition of Li instead of K, the active component of
influx falls by only 15%, and since [Na]^ is halved at the same time, the
K/Na gear ratio must rise by about 50% when half of [Na]e is replaced by
K and by about 75% when Li is the replacing cation.
It is possible that the small increase in K influx with rise of [K]e recorded
by E. J. Harris arises from his assumption that changes in cell volume are
negligible. Experimentally (Table 6), if 100 ml. cells are transferred from
a medium containing 2 to one with 75 m.equiv. K/l. they gain 5 ml.
water and this will correspond to an additional gain of about 7 m.equiv. K
of which Harris takes no account. His figures corrected for changes in cell
volume correspond to a gain not of 6 but of 13 m.equiv. in 24 hr. — a figure
which agrees closely with the corrected figures derived from Table 5 : this,
as has been seen, corresponds to a depression in true active influx not of 50
but of 25 %.
Thus, the theory of linked Na-K carriage, though explaining most of
the phenomena associated with cation transport in erythrocytes, does not
agree with the observed effects on K influx of halving the value of [Na]e,
either by adding K of Li to the suspending medium and an alternative
theory is set out below.
224 ACTIVE CATION TRANSPORT IN ERYTHROCYTES
Evidence has already been put forward to show that Na and K traverse
a lipoid phase of the cell membrane and it has been presumed that active
transport occurs by means of carriers. However, no simple theory suffices
to explain all the relevant facts of transport. Thus the existence of the
complex JHMa and YK respectively broken down at the outer and inner
faces of the cell membrane would not explain why K influx rises when
Na efflux increases, nor why active K influx falls when [K]e is raised.
Hence, a more complex theory is advanced involving several assumptions,
(i) It is suggested that the cell membrane contains two sorts of separated
transmitting zones, both lipoid: the passive zones, through which Na
and K move with the gradients, the active zones through which movements
occur only by means of carriers, Na and K being transported in opposite
directions against gradients of concentration and potential. (2) The amount
of K adsorbed at the external phase boundary of the active zone increases
rapidly as the external concentration rises from o to a certain critical low
level (1-2 m.equiv./l.) above which level further increase in the external
concentration of K is accompanied by little further increase in the amount
of K adsorbed. (3) Over the rest of the external phase boundary and over
the whole of the internal phase boundary of the cell membrane K adsorbed
is more closely related to the concentration in the adjacent aqueous phase:
a similar relation is thought to hold for Na over the whole of both faces
of the cell membrane. (4) There are in the active zones common carriers
for Na and K, their number being in excess of Na and K to be carried.
(5) Na is liberated from the Na-carrier complex by specific enzymes
related to the external phase boundary, while K is liberated from the
K-carrier complex by specific enzymes related to the internal phase
boundary. (6) The Na and K carriers at the respective transporting
surfaces are in equilibrium with Na and K adsorbed from the adjacent
watery phases, much as a citrate-calcium complex or a protein calcium
complex in solution is in equilibrium with ionized calcium. In short, the
active zones are envisaged as isolated lipoid cylinders containing no cations
but Na and K complexes: they are thought to be bounded by active
surfaces where the liberation of cations from the corresponding complex
at the lipoid side of the interface is governed by the amount of the same
cation which is adsorbed at the watery side of the interface, the latter
opposing to some extent the energizing effect of the underlying metabolic
process. The number of assumptions is somewhat large, but none are
unlikely. Thus, the second, involving strong adsorption of K at low
external concentrations, with little increase as [K]e rises above 2 m.equiv./L,
is supported by known facts: it will be recalled that with fresh cells, [K]e
at 2 m.equiv. loads half as many carriers as does Na with an internal
ACTIVE CATION TRANSPORT IN ERYTHROCYTES 225
concentration of 15 m.equiv., so that the association of K with the external
phase boundary seems to be about four times as great as that of Na at the
internal phase boundary; moreover, in stored cells uptake of K remains
maximal with [K]e at 2 m.equiv., although [Na]^ may equal 80 m.equiv.
and under these conditions K at the external face must be 20 times as
effective as Na at the internal face, in its ability to load carriers. If then,
this theory be accepted, it becomes possible to explain many of the
phenomena previously listed as characteristic of transport in the human
erythrocyte. The greater speed of passive Na transfer over passive K
transfer accounts for increase in cell Na, total base and volume during
cold-storage, while reversal of these changes during incubation with glucose
arises from the rate of Na transport exceeding that of K transport. Again,
the conditions envisaged are compatible with constancy of the ratio
[Na] J[Na]e. So, too, as [K]e falls below the critical level, K adsorbed will
fall sharply, the loading of K carriers will fail and K influx will fall : K de-
adsorbed from the external face of the cell membrane will be replaced by
Na, whose increased adsorption there will hinder the breakdown of out-
going Na-carrier complexes with a corresponding decrease of Na efflux.
As [K]e rises above the critical level, it is assumed that K adsorbed at the
external face of the active zones becomes maximal; thereafter K trans-
ported across the active zones remains constant, but K penetrating through
the passive zones will increase with rise of [K]e and the resulting rise of
[K]; will hinder the breakdown of K carriers at the inner phase boundary
of the active zones and so cause the observed decrease in active influx
which accompanies a rise in [K]e. The fall in K influx which accompanies
values of [Na]^ so low as to preclude much Na efflux during incubation is
more resistant of explanation: possibly, however, the Li which enters
passively from the Li-rich Na-poor external medium, displacing cell Na,
may be less strongly adsorbed at the inner phase boundary than Na, in
which case there would be increased adsorption of the competing K within
the cell and this according to the theory would lead to a decrease in active
K influx. Finally, during cold-storage [Na]^ rises and [K]^ falls; during
the subsequent incubation, the initially high [Na]^ will increase the rate
at which Na carriers are loaded, while the low level of [K]^ will facilitate
the breakdown of K carrier complexes. It follows that the increases of Na
efflux and K influx observed when stored blood is incubated, are not
linked but parallel events.
If all this be correct, then the apparent linkage between Na and K
transport really arises from the fact that the osmotic and electrical require-
ments of cells and plasma are such that Na and K in the cell phase and
also in the plasma phase are necessarily complementary. Thus, when cells
E B S VIII 15
226 ACTIVE CATION TRANSPORT IN ERYTHROCYTES
are incubated in a medium rich in K, external Na must be reduced if
isotonicity is to be maintained, while during cold-storage the initially high
levels of cell K and plasma Na lead to a fall in cell K and to a comple-
mentary rise in cell Na.
REFERENCES
BLACK, E. C. & IRVING, L. (1938). J. Cell Comp. Physiol. 12, 255-
CLARKSON, E. M. & MAIZELS, M. (1952). J. Physiol. 116, 112.
COHN, W. E. & COHN, E. T. (1939). Proc. Soc. Exp. BioL, N.Y., 41, 445.
DANOWSKI, T. S. (1941). J. BioL Chem. 139, 693.
DAVIDSEN, H. G. & KJERULF-JENSEN, K. (1950). Proc. Soc. Exp. BioL, N.Y., 74,
477-
DANIELLI, J. F. & DAVSON, H. (1934). J- CelL Comp. Physiol. 5, 495.
DAVSON, H. (1951). A Text-Book of General Physiology. London: J. and A.
Churchill Ltd.
DAVSON, H. & REINER, J. M. (1942). J. CelL Comp. Physiol. 20, 325.
DEAN, R. B. (1941). BioL Symp. 3, 331.
DEAN, R. B., NOONAN, T. R., HAEGE, L. & FENN, W. O. (1940). J. Gen. Physiol.
*4» 353-
DISCHE, Z. (1937). Enzymologia, i, 288.
DOWNMAN, C. B. B., OLIVER, J. O. & YOUNG, I. M. (1940). Brit. Med. J. i, 559.
DREW, C. R., ESDALL, K. & SCUDDER, J. (1939). J. Lab. Clin. Med. 25, 240.
DuLifcRE, W. L. (1931). C.R. Soc. BioL, Paris, 107, 261.
FERGUSON, J. K. W., HORVATH, S. M. & PAPPENHEIMER, J. R. (1938). BioL Bull.,
Woods Hole, 75, 381.
FLYNN, F. & MAIZELS, M. (1949). J. Physiol. no, 301.
GOURLEY, D. R. H. & GEMMILL, C. L. (1950). J. Cell. Comp. Physiol. 35, 341.
HAHN, L. & HEVESY, G. (1942). Ada physiol. scand. 3, 193.
HAMDI, T. N. & FERGUSON, J. K. W. (1940). Proc. Soc. exp. BioL, Paris, 44, 427.
HARRIS, E. J. (1953). Biochem.J. 54, xivP.
HARRIS, E. J. & MAIZELS, M. (1951). J. Physiol. 113, 506.
HARRIS, E. J. & MAIZELS, M. (1952). J. Physiol. 118, 40.
HARRIS, F. J. & PRANKARD, T. A. J. (1953). J. Physiol. xai, 47°-
HARRIS, J. E. (1941). J. BioL Chem. 141, 579.
HARROP, G. A. & BARRON, E. S. G. (1928). J. Exp. Med. 48, 207.
HENRIQUES, V. & ORSKOV, S. L. (1936). Skand. Arch. Physiol. 74, 78.
HEVESY, G. & HAHN, L. (1941). K. dansk. vidensk. Selsk. (BioL Medd.), 16, i.
HODGKIN, A. L. & KEYNES, R. D. (1953). J. Physiol. 120, 15^.
JACOBS, M. H., GLASSMAN, H. N. & PARPART, A. K. (1935)- J- Cell. Comp. Physiol.
7> 197-
JEANNENEY, G., SERVANTIE, L. & RINGENBACH, G. (1939). C.R. Soc. BioL, Paris,
130, 472.
KROGH, A. (1946). Proc. Roy. Soc. B, 133, 140.
LING, G. N. (1952). Phosphorus Metabolism, Vol. n. Baltimore: Johns Hopkins
Press.
LYMAN, R. A. (1945). J. Cell. Comp. Physiol. 25, 65.
McKEE, R. W., ORMSBEE, R. A., ANFINSEN, C. B., GEIMAN, Q. M. & BALL, E. C.
(1946). y. Exp. Med. 84, 569.
MAIZELS, M. (1932). y. Physiol. 77, 22P.
MAIZELS, M. (1935). Biochem.J. 29, 1970.
MAIZELS, M. (1949). J. Physiol. 108, 247.
MAIZELS, M. (1951). J. Physiol. 112, 59.
ACTIVE CATION TRANSPORT IN ERYTHROCYTES 22?
MAIZELS, M. & PATERSON, J. H. (1940). Lancet, 2, 417.
MAIZELS, M. & WHITTAKER, N. (1940). Lancet, i, 590.
MASING, E. (1914). Pflug. Arch. ges. PhysioL 156, 401.
MULLINS, L. J., FENN, W. O., NOONAN, T. R. & HAEGE, L. (1941). Amer. J.
PhysioL 135, 93.
MEYERHOF, O. & GELIAZKOWA, N. (1947). Arch. Biochem. 12, 405.
PONDER, E. (1950). J. Gen. PhysioL 33, 745.
RAKER, J. W., TAYLOR, I. M., WELLER, J. M. & HASTINGS, A. B. (1950). J. Gen.
PhysioL 33, 691.
SHEPPARD, C. W. & MARTIN, W. R. (1950). J. Gen. PhysioL 33, 703.
SHEPPARD, C. W., MARTIN, W. R. & BEYL, G. (1951). J. Gen. PhysioL 34, 411.
SOLOMON, A. E. (1952). J. Gen. PhysioL 36, 57.
STEINBACH, H. B. (1940). J. BioL Chem. 133, 695.
TEORELL, T. (1952). J. Gen. PhysioL 35, 669.
TERNER, C., EGGLESTON, L. V. & KREBS, H. A. (1950). Biochem. J. 47, 139.
WILBRANDT, W. (1940). Pfliig. Arch. ges. PhysioL 243, 519.
15-2
LINKAGE OF SODIUM- AND POTASSIUM-
ACTIVE TRANSPORT IN HUMAN
ERYTHROCYTES
BvE. J.HARRIS
Biophysics Department, University College, London
I. INTRODUCTION
The application of tracers has enabled a considerable amount of information
to be amassed about the rates of penetration of various ions into living cells.
Amongst these the human erythrocyte has been extensively studied, and
the purpose of this contribution is to show that by a simple assumption the
kinetics of movement of the alkali ions can be interpreted coherently and
quantitatively for a number of different experimental conditions. That some
similar assumption may be applicable to certain other cells, in particular to
the erythrocytes of other species, is possible, but experimental evidence is
not nearly so profuse and a stringent test is difficult to make.
The passive penetration of a membrane will normally depend upon the
size and charge of the penetrating particle, and on the structure and charge
of the membrane. In the following it is supposed that, for passive penetra-
tion taking place as a result of thermal agitation, the respective ions have
different probabilities of crossing the membrane per collision made with it ;
and it is assumed that, in addition, an active process potentiated by metabolic
reactions causes an outward flux of Na ions and an inward flux of K ions
so linked that the inward active K flux is numerically related to the outward
active Na flux. This might happen if the K carriers were formed chemically
from Na carriers when the latter reach the outer surface of the cell.
The qualitative facts which lead to this hypothesis will be briefly men-
tioned, and then a step-by-step examination of the experimental data will
be made in the light of it. The data have been drawn either from recent
literature or from experiments made recently by the writer in collaboration
with M. Maizels and T. A. J. Prankerd.
Under a variety of conditions the sum of Na and K in the cell undergoes
rather little change. If fresh cells are stored in the cold, or if metabolism is
stopped, an exchange of internal K for external Na proceeds, together with
some net gain of Na, so that total base per cell increases. The inference
drawn from this is that the rate of passive movement of Na is greater than
that of K, and, as the respective concentration gradients tending to lead to
K loss and Na gain are not very different, one may further conclude that the
TRANSPORT IN HUMAN ERYTHROCYTES 22Q
rate constant determining passive Na movement exceeds that determining
passive K movement.
When cold-stored cells, which have high Na and low K content, are
incubated they lose Na and gain K, but not equivalent to the Na lost, so
some diminution of volume and total base also takes place. Thus the rate of
active expulsion of Na against an electrochemical potential gradient is
somewhat more than the rate of active accumulation of K. That the active
fluxes of Na and K are not independent but are linked in some way is
suggested by the following : (a) K influx is high into cells which (on account
of high Na content) are expelling a high flux of Na; this can be deduced
from results given by Maizels (1951) and Ponder (1950), though these
authors do not specifically draw this conclusion ; (b) sufficient reduction of
external K reduces Na efflux (Harris & Maizels, 1952) and does not conflict
with the fact; (c) the K influx, provided a certain minimal K concentration
is present, does not vary with external concentration (Raker, Taylor,
Weller & Hastings, 1950; Sheppard & Martin, 1950; Solomon, 1952). This
suggests that the ingoing K is drawn from an adsorbed layer which is
saturated when Ke ^ c. 2 mmol./l. That the coupling is not a simple conse-
quence of the Na extrusion setting up a sufficient potential difference to
bring about a purely physical attraction of the K is shown by examination
of the distribution of other penetrating ions, e.g. Cl or H (Sheppard, 1951 ;
Davson, 1951; Harris & Maizels, 1952).
II. MOVEMENT OF Na
The movement of Na into the cell does not require a supply of energy
because it takes place down an electrochemical potential gradient, to which
the small electrical contribution is determined by pH. The rate of entry of
Na seems to be proportional to its concentration between 150 and
75 m.equiv./l., if the external osmotic pressure is maintained by substitution
of KC1, choline chloride, or sugar. When still more NaCl is replaced by
sugar the cells lose Cl~ and gain OH~ (Davson, 1939), and this makes
invalid a comparison of rates of movement of ions because pH affects the
rate constants. However, over the range mentioned, and in tracer experi-
ments in which total Na remains constant with only an exchange of 24Na
for 23Na, one can set Na influx =fkl\Nae] into unit volume of cell water,
in which k± is the first-order rate constant which would hold if no electrical
asymmetry were present, [Nae] is the external Na concentration, and/2 is
equal to [Cle]/[Cli] = [Hi]/[He] and depends upon pH (the Cl ratio as
function of pH is plotted in Harris & Maizels, 1952). The factor / allows
for the effect of the electrical asymmetry (see Appendix).
Movement of Na from the cell is assumed to take place by both passive
230 LINKAGE OF SODIUM- AND POTASSIUM-ACTIVE
and active processes. The passive process, analogous to the Na entry,
accounts for an efflux A1[Nai]//from unit volume of cell water, the factor/
appearing here in the denominator (cf. Appendix, p. 241). In addition, an
active process, with a rate constant k'2, operates to expel a flux ^[NaJ. In
each case [Nai] is expressed as concentration in cell water. The total Na
efflux is given by
The rate of change of internal Na, allowing for the possible variation of cell
volume (cf. Harris & Maizels, 1952) is
[Nad, (i)
where VQ is cell volume at t=o and V is cell volume. Provided V remains
sufficiently constant the equation can be integrated, and a useful form,
holding for [Nae] constant and initial [Nat] = [NaJ (o), is
k
using kz = -~ + k'% for brevity. This equation was used to find k% by Harris &
Maizels, who did not, however, separate k2 into the active and passive
components.
Energy requirements of Na efflux
At a given temperature the energy required for Na extrusion depends
upon the efflux and on the logarithm of the ratio of the external to internal
Na concentration, with allowance for the electrical asymmetry. In addition,
if one accepts the linkage between Na and K transport, a further term
appears involving the logarithm of the ratio of internal to external K
concentration. This last term is multiplied by a factor depending upon the
ratio of the number of K's carried in to Na's carried out. When one K is
carried in for each two Na's carried out, as seems to be the case (as shown
later), the expression
gives the Na and K movement energy requirement in unit time of unit
volume of cell fluid. The value of the expression is greater when [NaJ is
high than when it is low. Therefore the energy demand of the Na extrusion
process is more when the cells have a high Na content than when they have
expelled so much that a steady, low, Na level has been reached. An example
of three states during incubation of cold-stored cells is given in Table i .
TRANSPORT IN HUMAN ERYTHROCYTES
231
Similar figures are obtained even if the link between K and Na transport
differs from that assumed in the calculation.
Table i. Relative rates of energy consumption and values of energy required
per Na ion at different stages during incubation of cold-stored cells
External Na 150 mmol./I. External K = 4mmol./I. /=i-2 (pH 7'i).
[Na,]
m.equiv./l.
[KJ
m.equiv./l.
Relative Na
efflux ions/unit
time
Energy /Na ion
(for 2 Na out,
coupled to i K in)
Relative rate of
energy consumption
100
SO
10
50
IOO
140
lOO&z
50*2
IO&2
i'77kT
2-SikT
4>6okT
ijyktkTjunit time
i40'5k2kT/umt time
46&a£77unit time
& = Boltzmann's constant.
The fact that the extrusion process, although needing less energy, requires
it at an increasingly high electrochemical potential evidently sets the
lowest limit attained by the concentration of internal ionized Na. The
extrusion process becomes decreasingly efficient as the energy require-
ment per mol Na expelled approaches the free energy of the potentiating
reaction, and this will set a limit down to which the ratio [NaJ/[Nae] can
be driven. This fact could explain the observation of Harris & Maizels
(1951) that, although the transfer constant &2 varies widely, the ratio
[Nai]/[Nae] attained by the cells at 37° is comparatively constant, and even
remains so if part of the external Na is replaced by Li.
It is perhaps desirable to repeat the evidence that Na efflux is first order
with respect to [Nai], for another possibility would be that the flux should
correspond to a constant energy requirement. Table 2 summarizes an
experiment taken from Harris & Maizels (1952) showing that efflux is equal
to the product of a constant and [Nai]. In addition, there is the fact that
similar values for kz are found both by tracer methods (constant Nat) and
the chemical method (variable Nai).
Table 2. To show that Na efflux varies as Nai
Time
(hr.)
Interval
(hr.)
[NaJ
Mean Na,
(m.equiv./l.
water)
Mean net
loss
(m.equiv./hr.
x 1. water)
Mean efflux
(m.equiv./hr.
xl. water)
k*
hr.-1
0
6
ii
24
6
5
13
48-4
38-5
34
61-8
43'5
36-2
4*4
2'0
0'35
9'5
7-0
5'45
0-15
0-16
0-15
The mean efflux is found by adding the influx (5-1 /fequiv./ml. cell water x hr.) to the
mean net loss rate. The value for influx is found by equating the 24 hr. value of efflux
(kz 0-15 hr."1, Nai = [34]) to tne influx. Values, obtained by use of tracer Na, of the influx
at 37° are similar.
232
LINKAGE OF SODIUM- AND POTASSIUM-ACTIVE
The effect of temperature on Na transfer
In the discussion of the effect of temperature upon the rates of ion
transfer it will be necessary to distinguish carefully between the flux and
the rate constant. This is because the internal concentrations can vary; for
example, [Nai] increases and [Ki] decreases as the temperature is reduced.
At the steady state of internal Na and K the Na efflux must equal Na influx,
i.e./AjfNae] = /e2[Nai], If k2 diminishes more rapidly than k± as temperature
is reduced, then Nai will rise to a higher value until a new balance is
struck. Therefore the values of steady state flux (in or out) will vary equally
with temperature, although the rate constants may vary differently. When
the cells have not attained a steady state, as, for example, will be the case if
they have recently been cooled, the effect of temperature on the flux will
approximate to the effect it is having upon the rate constant. The well-
known fact that cooling the cells causes [Nai] to rise must mean that the
temperature-dependence of k2 is greater than that of kly or more specifically
the active rate constant k'2 diminishes more rapidly than does the passive
rate constant k± when the cells are cooled (for /is constant at a given pH).
The activation energy applying to the passive rate constant k± has been
found to be about 14,000 cal./mole in the first hour of treatment at the
lower temperature (Harris & Prankerd, unpublished).
Effect of pH on the Na rate constants
Provided the potential difference (E) across the cell membrane is small,
as is the case in the human erythrocyte, the effect is to increase the inward
rate constant by the factor/, and diminish the outward passive one by the
same factor, where f2 = exp(EF/RT), which is assumed to be equal to the
ratio [Cle]/[Cli] (see Appendix, p. 241). As the potential difference across the
cell membrane is determined by pH (Harris & Maizels, 1952) the passive
fluxes will vary with pH. In addition, if the fixed charge in the membrane
Table 3. The rate constants (fk^ determining Na influx at various pH. The
pH-dependent factor f has been evaluated from the ratio Cle/Cli— /2 taken
from Harris & Maizels (1952). It appears that ^ itself also varies with pH
pH
Inward rate
constant /&!
hr."1
Ratio /&!
acid/alk.
f2
Ratio /
acid/alk.
Ratio &x
acid/alk.
Source of
rate
constants
7*53
7-06
0-0226)
0*0162 f
0-72
•35)
•12)
0-91
0-79
Solomon
7-36
7-12
0-0194)
0-0159)
0-82
•35)
•i6F
0-93
0-88
Solomon
7'4
6-8
0-013)
o-oio)
0-77
•29)
•02)
0-89
0-87
Harris &
Maizels
TRANSPORT IN HUMAN ERYTHROCYTES 233
(such as that discussed by Teorell, 1951) is pH-dependent, the value of k^
itself will alter. Table 3 shows that the ratio of the quantity fk± at two
different pH's is not equal to the ratio of the respective/'s, that is to say, k±
is also varying with pH. The change of k^ happens to approximate to the
change of/, but the figures are rather scattered.
The active rate constant k'2 is much more dependent upon pH, and seems
to pass through a maximum at about pH 7-4 (Harris & Maizels (1951) make
this statement about &2, in which the greatest source of variation lies in the
active component k'2, so it is valid to shift their conclusion to k'2).
III. MOVEMENT OF K
Experiment indicates that the active K influx is numerically related to the
active Na efflux provided Ke ^ c. 2 mmol./l. In addition, by study of systems
lacking metabolite, it can be shown that passive movements in both direc-
tions take place, and it seems reasonable that eventually the ratio [Ki]/[Ke]
in such a system will become equal to the ratio [Cle]/[Cli] =/2. Then, as for
Na passive movement, there will be passive K influx =/A[Ke] and
efflux = h[Ki]/f, where h is the first-order rate constant holding in absence
of a potential difference. The active K influx will be written as r^[Nai],
where r is the number of K ions carried in per Na ion carried out. This active
component does not include the external K concentration as a factor,
because it seems that there must be ample time for the inward-bound
carrier to acquire a K ion irrespective of the K concentration provided the
latter is more than about 2 mmol./l. This is accordingly an empirical
formulation, but it may be supported, for example, by the demonstration
that the cells carry adsorbed K, for which there is already some evidence
(Maizels, 1935), and which is finding further support in recent experiments
(Harris & Maizels, unpublished).
The differential equation describing movement of internal K can then
be written dVK\
ai. (3)
When it is sufficiently accurate to take the cell volume as a constant the
equation can be integrated, and for constant Ke, [Nai] = Nai(o) and
[Ki] = [Ki] (o) at t = o, one obtains
(4)
x I i-exp I - — I |4- — j-
[A[Nae]-*t[Nai](o)][i-exp(-V)]
234 LINKAGE OF SODIUM- AND POTASSIUM-ACTIVE
The equation is interesting because it contains two exponential terms of
different time constant. One of these has the rate constant characteristic of
movement of Na, and the other the rate constant characteristic of passive
outward movement of K. The two terms are operative when [Nai] is varying.
An example of this system has been provided by Ponder (1951), who
examined the time course of the loss of K by cells suspended in NaCl
solution. To this solution cells at first rapidly lose K, so after a short time
the external solution will contain sufficient K for the equation to apply.
Eventually (20 hr.) metabolite supply fails and a rapid K loss again occurs
as if Na carriers, for want of the requisite chemical transformation, are
operating on internal K. Ponder was able to express his curve as the sum of
two exponentials having exponents (at 37°) 0-27* and o-ont (t in hours),
which compare with values of Na transfer constant (k2) 0-25-0-35 hr."1 and
K exchange constant 0-016 hr."1 found in isotope experiments. When
[Nai] remains steady the equation loses the more rapidly varying term, and
in observations of K exchange using isotopic K the quantity h/f determines
the rate. Substitution of the two steady-state conditions
|[K1] = r^[Nai]+/A[Ke] for K, (5)
and AtNaJ-fe + aijtNai] for Na, (6)
allows the equation for constant [Ke] of constant specific radioactivity to
be reduced to *Ki/Ki = [i — exp( — ht/f)], where *Ki/Ki expresses the
specific radioactivity of the internal K taking that of the external K as unity.
No evidence of coupling between Na and K transport can remain under
these conditions. If, however, [Nai] diminishes in cells exposed to isotopic
K (denoted by *K) there will be some uptake of *K with the rapid rate
constant of the Na-active movement, in addition to the slow exchange,
and this might invalidate comparison of the K exchange rates under
different conditions. Similarly, cells containing *K will exchange it slowly
for ordinary K from a non-radioactive solution, but will lose it (in exchange
for Na) rapidly if conditions do not favour constancy of internal K level.
IV. NUMERICAL VALUES OF Na AND K FLUXES
At 37-38° C. the Na flux into and out of the cells has been given as 4-74
(Solomon, 1952) and 4-7 (Sheppard, Martin & Beyl, 1951) in //equiv./ml.
cell water per hr. Under similar conditions and in the same units K flux
is 2-51 (Raker etal. 1950), 2-57 (Solomon, 1952), 2-52 (Sheppard & Martin,
1950, with assumed Ki=i4O /^equiv./ml. cell water). To express the
figures per ml. cell water a value of cell water equal to 65 % of the cell
volume has been used.
TRANSPORT IN HUMAN ERYTHROCYTES 235
To find r, the ratio of the active K influx to the active Na efflux it is
necessary to separate active and passive components of the respective
fluxes. For Na: as in the steady state at 37° [NaiJ/fNa^] comes down to
Ak \
-j + k'2\ =0-07, and with /2= 1-28, fkl comes to
about o-o8&2> i-e. the active Na efflux is 0*92 x total Na efflux, or about
4*3 /^equiv./ml. cell water per hr. For K ions the rate constant determining
efflux h/f is 0-0165 hr."1 (mean of Raker et al. and Solomon). This figure is
multiplied by/2(i -28) to obtain/A and by Ke to obtain the passive component
of the K influx. For Ke = 4 mmol./l. this comes to o-o8/^equiv./ml. cell
fluid. Subtracting this from the mean of the values for total K influx one
has: active K influx at 37° = 2-45 /^equiv./ml. fluid per hr. Then
r = 2-45/4-3 =0-56.
An alternative evaluation can be made using the two steady-state equations,
for since *
Kt=/2Kc+^r (active Na flux),
with active Na flux 0-92 x total Na flux, and h/f the exchange rate constant
f°r K K, = /«Ke + °'9^ (total Na flux)
(K exchange constant)'
Putting in usual values for KI, Ke and the flux and K exchange rate constant
the value of r comes to 0-51 at 37°.
It is more difficult to evaluate r at other temperatures because infor-
mation is less precise, and particularly it is doubtful how constant the
internal levels of Na and K were in the experiments. In a recent set of
isotope exchange experiments at 27-5°, for example, it was found that *K
uptake and *Na loss was more rapid initially than after 3 hr. The value of
r found was about 0-5.
It is interesting and suggestive to compare k± with /z, the respective
passive transfer constants (to which the permeability constants are pro-
portional) for Na and K in absence of an electrical asymmetry. To find k±
it is necessary to divide the inward rate constant/^ by/(i-i3). Using
figures from the work of Solomon (1952) or Sheppard et al. (1951), &i is
about 0*034/1-13 =0-030 hr."1 at 37°. (It is to be noted that Solomon's
inward rate constant 0-022 hr."1 appears to refer to unit volume of cells,
whereas in the present paper all figures refer to unit volume of cell water,
taking this as 0-65 of the cell volume.) The rate constant of K exchange, h/f,
is 0-0165 hr."1 at 37°, so h is 0-0187 hr."1. Therefore one has
Passive permeability to K __o-oi87_ ,
Passive permeability to Na 0-030
236 LINKAGE OF SODIUM- AND POTASSIUM-ACTIVE
At low temperatures, when the active processes are nearly stopped, the rate
constants for Na and K movement are known. That for K entry at 4° is
0-0023 hr."1 (Sheppard & Martin, 1950) or 0-0016 hr."1 (Raker et al. 1950),
whereas that for Na entry lies between 0*0035 anc^ 0*0042 hr."1 (Harris &
Maizels, 1952). If it be assumed that no active K influx is present at this
low temperature the passive permeability ratio (K/Na) lies between 0-5
and 0-7, so it is probably the same as at 37°.
The fact that the passive permeability to Na ions is greater than that to
K ions is strong evidence that neither ion penetrates the cell in the normally
hydrated state, for Na ions in solution are larger than K ions, as shown by
their lower diffusion constant. It seems likely that the mechanism of both
passive and active processes is similar, involving attachment of ions to
groups in the membrane which carry the ions across. The only difference
between passive and active movement would then be that the former
involves directed flow of K-carrying groups inward, and Na-carrying
groups outward, whereas the latter is purely random, as in Ussing's (1949)
exchange diffusion process.
The effect of Ke on the exchange of K
All investigators who have examined the effect of variation of Kc on the
rate of K turn-over at 37° agree that the rate is little affected over a wide
range of concentrations, which are obtained by replacing Na by K in the
medium. In particular, Raker et al. (1950) used Ke between 2 and
74 mequiv./l.
This result has in fact been incorporated into the formulation by making
active K influx independent of [Ke], but it transpires that the sum active K
influx 4- passive K influx is nearly independent of Ke, which is not directly
obvious. One has
active K flux = r*JNai = r*1/Nae- (from (6)),
passive K flux=/AKe.
So, total K influx =y(r*1Nae + AKe)-*1 Nat.
As rk^ is c. 0-016 hr."1 and h is 0-0187 hr."1, from the figures already
mentioned, the factors multiplying Nae and Ke are nearly equal, so the first
term is nearly constant so long as the sum (Ke -f Nae) is constant, and the
second, negative, term is in any case small, about 5 % of the positive term
at most. Thus K influx will not depend very appreciably upon Ke. Calcula-
tion indicates that from 74 m.equiv./l. the influx would be 7 % greater than
from 4 m.equiv./L, which is not very easy to detect experimentally,
particularly as cell volume may change.
TRANSPORT IN HUMAN ERYTHROCYTES 237
The effect of pH on K transfer
When internal K remains steady the rate of exchange of K is determined
by h/f. /varies with pH, but only from about 1-13 at pH 7-4 to 1*0 at pH 6-7.
Probably h varies in the same direction as/, for pH is likely to have the same
influence upon h as it has upon kv Therefore it is unlikely that a distinct
influence of pH upon the K exchange will be observable, and this is in fact
in agreement with the observations of both Raker et al. and Solomon. On
the other hand, if K-depleted cells are incubated the uptake of K in exchange
for cell Na involves the active Na rate constant k'2. Then the time course of
K net gain will depend upon pH in the same way as the rate constant deter-
mining Na extrusion. Flynn & Maizels (1949) and Ponder (1950) both find
a pH optimum near 7-4 for net K gain (and pH net loss), agreeing with the
optimum for the rate of Na extrusion (Harris & Maizels, 1951).
The effect of temperature on the rate of K transfer
The activation energy of the K transfer process has been measured by
Raker et al. (1950), Sheppard & Martin (1950) and Solomon (1952). Figures
for both inward and outward rate constants in the range 40-25° are between
13,000 and 15,000 cal./mole. A difficulty which arises in evaluating energy
of activation applying to the outward rate constant is that even a very slow
rate of net loss of K from the cells may considerably increase the value of the
efflux rate constant above that deduced from exchange measurements
(entry of *K), in which it is assumed that cell K concentration remains
constant. Error also arises when cell volume is not constant and as in fact
below 37° the K concentration falls (cf. Raker et al., table 3), although the
content remains steady (at 24*4°) because the cells also swell, the transfer
constants applying to K movement at this, and lower temperatures should
be evaluated graphically, bringing cell volume into the equation. An experi-
ment made here on K entry at 37-5°, 27-5° and 18° showed that the rate
constant applying to K entry at the two lower temperatures was diminishing
for several hours after cooling the cells. The activation energy calculated
after 3 hr. was nearly twice that calculated after i hr. The subject evidently
requires further study, but it seems correct to state that the activation
energy applying to the K efflux is close to that found for Na influx
(about 13,000 cal./mole), which is to be expected if they are both similar
processes. The activation energy applying to K influx is certainly higher
than this when observations are made at lower temperatures (Sheppard &
Martin, 1950), and it seems likely that it will be found that the active K
influx has an energy of activation of some 25,000 cal./mole over the whole
temperature range wrhen account is taken of disturbing factors. Attention
238 LINKAGE OF SODIUM- AND POTASSIUM-ACTIVE
must be drawn to the fact that cells which are not in a steady state at the
temperature in question will have time variable K fluxes, and the activation
energy deduced will apply to variable proportions of active and passive
processes.
It is interesting that the activation energy of the glycolytic reaction is
very close to that of the passive-ion movement. This might indicate that the
passive movement of some ion across the cell membrane controls glycolysis.
In addition, it is possible to envisage an active transfer mechanism whose
rate would depend upon the product of the concentration of a labile inter-
mediate and the concentration of those ions having 13,000 cal./mole activa-
tion energy. The overall activation energy of the active process might then
come to the sum of the energies applying respectively to passive penetration
and to the rate of glycolysis.
The exchangeability of cell Na
Both Solomon (1952) and Sheppard et al. in two papers (1951) found
that one-third to one-half of the cell sodium did not appear to participate
in the exchange process in vitro. Analyses made recently (Harris & Maizels,
unpublished) also indicate that cells suspended in Na-free solutions do not
lose all their Na. The tracer experiments quoted, and others made here,
indicate a variable behaviour of the cell Na, as if some external factor,
operative in vitro, brings about the partial immobilization of part of the Na.
In saline suspensions the rate constant for Na efflux diminishes after some
hours storage at 37° (Harris & Prankerd, 1953). Use of plasma-saline
mixtures appears to favour consistent results, but even whole bloods to
which are added a very little radioactive preparation frequently show an
incomplete equilibration of their Na after 24 hr. incubation. One possible
explanation for this apparently incomplete exchange (though not for the
slowing of the rate constant) would be provided if a part of the Na was so
readily exchanged that the shortest wash, as usually applied to remove
radioactive extracellular fluid, brought about a substitution of ordinary Na
for the isotopically labelled Na. As mentioned when discussing K exchange
there is evidence for Na and K being adsorbed, and it is found that unwashed
samples of cells drawn from a suspension just after addition of tracer Na
carry several times more radioactivity than that expected for the extra-
cellular fluid alone.
The effect of fluoride
In low concentrations (2-10 m.equiv./l.) fluoride inhibits glycolysis, and
after 1-2 hr., during which the cells use up reserves of metabolite, the active
Na extrusion ceases (Maizels, 1951 ; Harris & Maizels, 1951). Na movement
is then governed by the passive rate content. In high concentrations of
TRANSPORT IN HUMAN ERYTHROCYTES 239
fluoride (60 m.equiv./l.) Davson (1941) found that a drastic change in
behaviour was caused in the rabbit erythrocyte. K was lost without a
corresponding gain of Na, so the cells shrank. An experiment has indicated
a similar behaviour on the part of human cells, which under these conditions
have a Na efflux several times that of the usual passive efflux, and a high
permeability to K, so it seems that a change in the membrane structure has
been induced.
Other alkali ions
Some isolated observations of the effect of other alkali ions on Na and K
movements have been made (Ponder, 1950; Flynn & Maizels, 1949;
Solomon, 1952). Using tracer K in a mixture containing Rb, Solomon made
two experiments. These suggest that, within experimental error, K influx
is reduced in the ratio Ke/(Ke + Rbe). The figures are
Ke/(Ke + Rbe)
K flux in presence Rb
K flux in absence Rb
0-58
0-61
0-52
0-66
This result would follow if the inward-going carriers made little distinction
between the two ions. Li and Cs apparently are not taken in by the K
carriers, for they did not reduce K influx directly; an indirect effect of Li
is mentioned later.
Substitution of Li, K or Rb for two-thirds of the external Na reduces
Na influx, Li having rather more effect than the other ions. The total influx
of Li + Na is about 4*0 m.equiv./l. cell fluid per hr. as compared to the usual
Na influx of 4-7 m.equiv./l. cell water per hr., which suggests that the
passive rate constant applying to Li entry is lower than that applying to Na
entry.
Cells exposed to mixtures of Li and Na retain their usual Na efflux rate
constant (&2) (Harris & Maizels, 1951), and ultimately the ratio Nai/Nae
becomes equal to that found in the high Na solution. The reduction of
[Nai] which occurs in the Li mixture will lead to Na efflux diminishing in
proportion, and on the hypothesis advanced here this will lead in turn to
a reduced influx of K, so [Ki] will fall. That this reduction of K influx and
KI does occur has been observed both by Ponder (1950) and Flynn &
Maizels (1949).
Other erythrocytes
Figures for the rates of turn-over of Na and K in cells of some other
species have been given by Sheppard et al. (1951). These cells differ from
those of man because they have low [KJ and high [Nai] ; thus they are not
240
LINKAGE OF SODIUM- AND POTASSIUM-ACTIVE
likely to require a very efficient K transporting system. A rough evaluation
of the ratios 'r' and h/kl9 i.e. the respective active and passive permeability
ratios for K ions compared with Na ions has been made :
Animal
[NaJ
/^equiv./ml.
cell water
[KJ
/^equiv./ml.
cell water
hlki
r
Dog
Sheep
Cow
1 68
156
102
8
15
37
o-i
o-i
0-3
O-02
O'OI
0-3
It is remarkable that again the rate constants determining K passive move-
ment are lower than those determining Na passive movement, and the active
process carries fewer K ions than Na ions.
V. CONCLUSION
In conclusion, the results obtained for human red cells indicate that a single
mechanism brings about active Na extrusion and active K accumulation.
The K must be drawn from a reservoir which is filled to a level independent
of the external concentration when the latter exceeds 1-2 m.equiv./l. In
addition to the active process, there are passive fluxes applying to each ion
operating in each direction. The rate constant determining passive sodium
movement is greater than that determining passive potassium movement.
Recent observations by Steinbach (1952) on muscle and Hodgkin &
Keynes on nerve (1953) suggest that in these tissues also there is a linkage
between the active movement of sodium and the accumulation of potassium,
so it is possible that a common chemical process operates in a number of
different biological materials. The elucidation of the chemistry of this
process appears to be one of the most important issues in the field of so-called
permeability studies.
REFERENCES
DAVSON, H. (1939). Biochem.J. 33, 389-401.
DAVSON, H. (1941). J. Cell. Comp. Physiol. 18, 173-85.
DAVSON, H. (1951). Textbook of General Physiology. London: Churchill.
FLYNN, F. & MAIZELS, M. (1949). J. Physiol. no, 301-18.
HARRIS, E. J. & MAIZELS, M. (1951). J. Physiol. 113, 506-24.
HARRIS, E. J. & MAIZELS, M. (1952). J. Physiol. 118, 40-53.
HARRIS, E. J. & PRANKERD, T. A. J. (1953). J. Physiol. (in the Press).
HODGKIN, A. L. & KEYNES, R. D. (1953). J* Physiol. 120, 46 P.
MAIZELS, M. (1935). Biochem.J. 24, 1920-82.
MAIZELS, M. (1951). J. Physiol. 112, 59-83.
PONDER, E. (1950)- J- Gen. Physiol. 33, 745~57-
PONDER, E. (1951). J. Gen. Physiol. 34, 359-72.
RAKER, J. W., TAYLOR, I. M., WELLER, J. M. & HASTINGS, A. B. (1950). J. Gen.
Physiol. 33, 691-702.
SHEPPARD, C. W. (1951). Science •, 114, 85-91.
TRANSPORT IN HUMAN ERYTHROCYTES 241
SHEPPARD, C. W. & MARTIN, W. R. (1950). J. Gen. Physiol. 33, 703-22.
SHEPPARD, C. W., MARTIN, W. R. & BEYL, G. (1951). J. Gen. Physiol. 34, 411-29.
SHEPPARD, C. W. & BEYL, G. (1951). J. Gen. Physiol. 34, 691-704.
SOLOMON, A. K. (1952). J. Gen. Physiol. 36, 57-110.
STEINBACH, H. B. (1952). Proc. Nat. Acad. Sci.y Wash., 38, 451.
TEORELL, T. (1951). Z. Elektrochem. 55, 460-9.
USSING, H. H. (1949). Physiol. Rev. 29, 127-55.
APPENDIX
The effect of a small potential difference upon inward
and outward rate constants
By assuming a constant field across the membrane, Goldman (1943) and
later authors have derived integral forms of the diffusion equation which can
be written for a univalent positive ion:
. fl _ ptfF (outside concentration)
influx - ~~ '
ffl ^eV exp ( — eV/kT) (inside concentration)
efflux = /^ i-exp(-«F/*T) '
P= permeability constant, proportional to the transfer constant, in absence
of the electrical field, k = Boltzmann's constant, V the magnitude of the
potential difference which is across the membrane (inside of cell negative),
e = the electronic charge, T the absolute temperature.
When eV is small compared with kT, as is the case in the human
erythrocyte as judged by the chloride ratio, the exponentials approximate to
i-eV/kT+l>(eV/kT)*.
Substituting this one obtains
. fl P (outside concentration)
influx = -j— \tfjkT '
efflux = (l _eV/kT+ Hf
= P(inside concentration) x (i —%eV/kT), nearly.
Putting i — \eV\kT-f, it is seen that the operative inward permeability
constant in presence of the field is/P, and the outward one is P/f.
Also, for a positive ion (e.g. H) which is not subject to active transport,
when the concentrations are steady influx = efflux and (inside concentra-
tion) =/2 (outside concentration). For a negative ion, such as Cl, the
equations are reversed and Cle/Clt =/2.
REFERENCE
GOLDMAN, D. E. (1943). J. Gen. Physiol. 27, 37.
E B S VIII l6
THE ACCUMULATION OF AMINO-ACIDS
WITHIN STAPHYLOCOCCAL CELLS
BY E. F. GALE
Medical Research Council Unit for Chemical Microbiology,
Department of Biochemistry, University of Cambridge
I. INTRODUCTION
Our attention was first drawn to problems of active transport in bacterial
cells when, in the course of studies on the assimilation of amino-acids by
bacteria (Gale, 1953), we found that certain Gram-positive bacteria con-
tained high concentrations of free amino-acids inside the cells (Gale,
1947; Taylor, 1947), and the problem arose, among many others, of the
mechanism whereby these amino-acids were retained in such high
concentrations.
It is now known that many cells of microbial, plant and animal origin
contain free amino-acids. Christensen, Riggs, Fischer & Palatine (19520, b)
have followed their earlier studies on the accumulation of glycine by various
animal cells by a detailed investigation of amino-acid accumulation by
mouse ascites tumour cells. These cells affect a marked concentration of
many amino-acids, including such as a-y-diaminobutyric acid which do
not occur naturally, and it seems that some amino-acids displace potassium
ions within the cells. In the case of diaminobutyric acid the potassium
displacement is almost complete and results in an intense concentration of
the amino-acid across the cell wall. The effect of metabolic activity of the
cells on the amino-acid concentration has not been studied in detail and,
although no oxidizable substrates are added in addition to the amino-acids,
it is probable that the tumour cells accomplish an endogenous respiration
during the concentrative action, since the addition of inhibitors such as
cyanide, dinitrophenol and arsenate markedly decreases the ability to
concentrate glycine (Christensen & Riggs, 1952). The accumulation of
glycine is accompanied by a swelling of the cells, so it would appear that
the amino-acid is osmotically active within the cells. Our own studies have
concerned the accumulation of amino-acids by bacteria and yeasts, and,
for the purposes of this symposium, I intend to restrict my discussion
mainly to one organism, Staphylococcus aureus. Taylor (1947) studied the
ability of a range of bacteria to accumulate amino-acids, and Staph. aureus
strains proved to have the highest concentrating activity towards glutamic
acid. Although glutamic acid may not be the amino-acid which undergoes
ACCUMULATION OF AMINO-ACIDS WITHIN CELLS 243
the highest concentration across the staphylococcal cell- wall (see Table i),
its ease of estimation by the specific decarboxylase method (Gale, 1945)
has resulted in a greater amount of knowledge accumulating concerning its
transport than for any other amino-acid. The investigations have been
reviewed elsewhere (Gale, 1949^, 1953), and certain aspects of them have
been discussed in a previous symposium of this Society (Gale, 1948).
The concentration of the free amino-acids within staphyloccocal cells
varies with the conditions holding at the time of harvesting of the cells
from a growth medium. Factors which affect the concentration include the
concentration of free amino-acids in the external medium, the phase of
growth when harvesting occurs, the availability of sources of energy such
as fermentable carbohydrates, the availability of other amino-acids, the
rate of protein synthesis within the cells, and the presence of phosphate and
other ions in the external medium. By a judicious selection of conditions
it is possible to obtain suspensions of organisms which possess very little
free amino-acid within the cells; these 'deficient' cells can then be used to
investigate the conditions under which specific amino-acids will accumu-
late within the cells.
II. ACCUMULATION OF LYSINE AND
GLUTAMIC ACID
If such deficient cells are suspended in a solution of lysine, free lysine
begins to accumulate rapidly within the cells, and the accumulation
continues until the internal concentration is some 4-60 times greater than
that in the external solution, the concentration gradient becoming larger as
the external concentration decreases. This accumulation is not affected by
metabolic inhibitors such as cyanide or dinitrophenol, takes place rapidly
at 2° C., and has a temperature coefficient of 1-4 which is within experi-
mental error of the value for free diffusion. The process appears to be one
of diffusion across the cell-wall followed by a distribution between internal
and external media essentially similar in properties to a Donnan distribution
(Gale, 1947, 1953; Najjar & Gale, 1950; McQuillen, 19500).
If deficient cells are suspended in a solution of glutamic acid, even at
37° C., there is no accumulation of glutamic acid within the cells as long as
precautions are taken to exclude metabolic sources of energy. If an energy
substrate such as glucose is added to the incubation mixture, then rapid
accumulation of free glutamic acid takes place within the cells, and the
internal concentration may rise to as much as 400 times that in the external
medium. The accumulation is abolished by any inhibitor which prevents
the metabolism of glucose, does not take place at 2° C. or less, and has
a temperature coefficient of 2-7-2*8. If the rate of internal accumulation is
16-2
244 THE ACCUMULATION OF AMINO-ACIDS
determined at various external concentrations, that for lysine is found to
vary in an approximately linear manner with external concentration, while
that for glutamic acid is independent of external concentration except for
low values of the latter (Gale, 1947). Further, if the amount of amino-acid
appearing within the cell is compared with that disappearing from the
external medium, it is found that the transport of lysine is quantitative,
whereas there is frequently a marked over-all loss during the transport of
glutamic acid.
If cells are loaded with glutamic acid by incubation in the presence of
glucose, then washed and resuspended in water or saline, there is a very
slow loss of glutamic acid from inside the cell to the external medium. The
addition of glucose stops this loss, but uncoupling agents such as sodium
azide or dinitrophenol do not accelerate it.
The accumulation of glutamic acid by Staph. aureus is accompanied by
an increase in the potassium content of the cells. The amount of potassium
taken up by the cells corresponds to not more than i atom per molecule of
glutamic acid accumulated (Davies, Folkes, Gale & Bigger, 1953). If the
experiments are carried out in a medium freed from potassium as far as
possible, the accumulation of glutamic acid is decreased but not abolished,
the increase in amino-acid within the cells being partially balanced by an
increase in sodium ions. The ability to accumulate glutamic acid is fully
restored by the addition of potassium but not magnesium or ammonium
ions.
The accumulation of glutamic acid thus takes place only when metabolic
processes are occurring and, as a first approximation, would appear to take
place as a result of active transport across the cell-wall structures.
If the accumulation process involves a metabolic link between glutamic
acid and glucose breakdown, it should be possible to inhibit the enzyme
or enzymes concerned in the link and so prevent glutamic acid accumula-
tion without inhibiting glucose fermentation or respiration. In the course
of our studies we have found six substances having such inhibitory actions:
sodium azide, 2.4-dinitrophenol, crystal violet (Gale, 1951), 8-hydroxy-
quinoline (Gale, 19496) and> under somewhat different circumstances,
penicillin (Gale & Taylor, 1947) and bacitracin (Paine, 1951). Sodium
azide and dinitrophenol have been shown to uncouple oxidative phos-
phorylation processes in yeast and mitochondria, and the general properties
of their inhibition of glutamic acid accumulation may mean that similar
coupled reactions are involved here. 8-Hydroxyquinoline appears to act
by inactivation of a metal (manganese or magnesium) necessary for the
accumulation process (Gale, 19496). Penicillin and bacitracin prevent
glutamic acid accumulation without affecting glucose metabolism if the
WITHIN STAPHYLOCOCCAL CELLS 245
cells are allowed to grow in the presence of either antibiotic for 30-90 min.
before harvesting. It is probable that the effect in this case is a secondary
one representing a loss of function following upon some more direct
inhibition.
III. ACCUMULATION OF OTHER AMINO-ACIDS
Chromatographic examination of extracts from staphylococcal cells shows
that many amino-acids, other than glutamic acid and lysine, exist in the
free state within the cells. It would be of interest to know which of these
accumulate as the result of active processes. Enzymic methods of estima-
tion which can clearly distinguish between amino-acids inside and outside
cells are available for only a few amino-acids (Gale, 1945; Krebs, 1950).
Such methods show that the accumulation of glutamic acid, aspartic acid
and histidine is increased by the presence of glucose. Recently, C14-
labelled amino-acids have become available, and it has been possible to
investigate the distribution of radioactivity across the cell-wall when
Staph. aureus is incubated with specific amino-acids with and without
glucose. The organism can acquire endogenous stores of energy under
some conditions of growth, and some accumulation of, for example,
glutamic acid will occasionally occur in the absence of added glucose as
a result of the utilization of such stores. However, this utilization can be
abolished by the use of 2.4-dinitrophenol, and the method adopted has
been to investigate the accumulation of radioactivity in the presence of
glucose with and without an uncoupling concentration of 2.4-dinitrophenol
(DNP). The measurements give the radioactivity of the external and
internal media, and it does not follow that the activity of the cell contents
is necessarily due to the amino-acid used alone. However, the ability of
the staphylococcus to break down amino-acids is very limited (Hills, 1940),
and the ratios obtained give an indication of the importance of active
processes in the assimilation of the amino-acids studied. Table i summarizes
the results obtained.
The concentration ratio for internal and external radioactivity was
determined in each case and column (c) of Table i shows the increase in
this ratio when DNP was omitted. In the case of glutamic acid, which has
been exhaustively investigated by the enzymic method, the ratio obtained
in the absence of DNP was 324, which agrees well with the value of 400
obtained by the enzymic method for low external concentrations of glutamic
acid. In the presence of DNP the ratio falls to 2-8, and it is clear that active
processes, which can be abolished by DNP, result in a very great increase
in the accumulation of glutamic acid. In the case of lysine, the isotope
experiments indicate a marked accumulation in the presence of DNP, and
246 THE ACCUMULATION OF AMINO-ACIDS
this accumulation increases about 3 times when DNP is omitted. It would
appear that, although lysine accumulates in the absence of active processes,
the accumulation increases when energy is made available. The amino-
acids which resemble lysine, in that a marked accumulation occurs in the
presence of DNP, are arginine, alanine and glycine. Glycine gives the
highest concentration ratio under these conditions, and its accumulation
would appear to be decreased by glucose metabolism. It is possible that the
accumulation of glycine in Staph. aureus resembles that in mouse ascites
Table i . Active processes involved in the accumulation of
free ammo-acids within Staphylococcus aureus
Washed suspensions of cells were incubated with C14-labelled amino-acids, glucose,
buffered salt solution with and without 0*01 M 2'4-dinitrophenol. After 60 min. at 37° C.,t
the cells were centrifuged down and the free amino-acids liberated, after exhaustive
washing, by treatment with cetyl-trimethylammonium bromide. The radioactivities of
the internal fluid and of the external medium were determined and expressed as a ratio.
Concentration
. Internal
ratio - :
Increase in concen-
External
Amino-acid
tration ratio due to
active processes
(a)
(b)
Glucose
Glucose + DNP
present
present
c = a/b
Proline
1670
9*7
171 i
Glutamic acid
324
2-8
104
Phenylalanine
16
0-9
17-2
Aspartic acid
307
4-24
7-2
*Methionine/valine
26-2
3-96
6-6
Threonine
13-8
7-0
1-97
Tyrosine
8-65
5-36
0-62
Lysine
196
63
3-12
Arginine
96'5
76
1-26
Alanine
19-7
19*6
I'OI
Glycine
8i-5
92-8
0-89
* The labelled amino-acids were obtained biosynthetically and fractionated chromato-
graphically; valine and methionine were combined in one fraction.
tumour cells and takes place by displacement of potassium. The amino-
acids which do not accumulate markedly in the presence of DNP and
whose concentration is increased by active processes are proline, glutamic
acid, phenylalanine, aspartic acid and methionine/valine. The internal
concentration in all cases is calculated on the assumption that the whole
volume of the cell is available for accumulation purposes ; the true * water-
space' of the cell must be considerably less than this, and consequently the
concentration gradients given are smaller than those actually occurring
across the cell-wall. It would seem that some degree of accumulation of all
the amino-acids must occur even in the absence of active processes.
WITHIN STAPHYLOCOCCAL CELLS 247
IV. ACTIVE TRANSPORT OF GLUTAMIC ACID
The purpose of this symposium is to discuss the nature of the active
transport process, and I intend to devote the rest of this contribution to
a consideration of the active accumulation of glutamic acid in the staphylo-
coccus. There would appear to be three possible explanations for the
experimental observations :
(1) That the cell- wall structures include a membrane impermeable to
glutamic acid and that, during the breakdown of glucose, some linked
metabolism occurs which results in the production of a derivative of
glutamic acid which can diffuse through this membrane and is reconverted
to glutamic acid within the cell.
(2) That the cell-wall structures are impermeable to glutamic acid but
contain a substance which is capable, during glucose metabolism, of
combining with glutamic acid and acting as a carrier across the impermeable
barrier.
(3) That the cell-wall structures are freely permeable to glutamic acid,
but that, when glucose is metabolized within the cell, a derivative of
glutamic acid is formed which is unable to diffuse out of the cell and
consequently accumulates. This derivative would be estimated as free
glutamic acid in the experimental procedures so far used.
A fundamental difference between hypotheses (i) and (3) is that, in the
former, coupled metabolism must take place near the surface of the cell,
outside the permeability barrier, whereas, in the latter, the metabolism
occurs within the cell. If the reaction is on the surface of the cell it should
be possible to reproduce the reaction by supplying breakdown products of
glucose. We immediately think of the possibility of generating energy-rich
bonds probably involving phosphate; this would give a system analogous
to that postulated by Gourley (1952) for the transport of phosphate in the
erythrocyte. However, it has not been possible to promote glutamic acid
transport by the addition of hexose- or triosephosphate, adenosinetri-
phosphate, or metaphosphates. Negative results of this nature can always
be explained away by the further postulate that the site of the coupled
metabolism is not actually on the surface of the cell but lies beneath the
surface while being above the barrier to glutamic acid penetration.
Cytologists will not necessarily quarrel with a postulate which endows the
bacterial cell with a series of outer coats.
Serious consideration has been given to the possibility that glutamine
might be the form in which glutamic acid passes across the cell-wall
(Mitchell, 1949). The synthesis of glutamine from glutamic acid requires
the intervention of adenosinetriphosphate (Elliott, 1951), and an attractive
248 THE ACCUMULATION OF AMINO-ACIDS
hypothesis can be built around the idea that coupled phosphorylation near
the surface of the cell results in synthesis of glutamine, which diffuses
through the cell-wall and undergoes irreversible hydrolysis to glutamic
acid within the cell. However, addition of glutamine to suspensions of
staphylococci does not result in the accumulation of glutamic acid within
the cells unless glucose is also added. Further, the addition of a-amino-y-
methylsulphinylbutyric acid does not affect the accumulation of glutamic
acid within Staph. aureus, although it acts as an inhibitor of the system
converting glutamic acid to glutamine (Waelsch, Owades, Miller & Borek,
1946; Elliott & Gale, 1948) and affects the assimilation processes for
glutamic acid in Strep, faecalis (Gale, 1949 a).
If a metabolic modification of glutamic acid occurs prior to its passage
into the cell, it should be possible to obtain some indication of the nature of
this modification by a study of the accumulation of glutamic acid within
cells presented with derivatives of that amino-acid. A number of glutamic
acid derivatives have been tested in this sense both in the presence and
absence of glucose. Of the substances tested, only two have been found
which will give rise to glutamic acid inside the cell when incubation takes
place in the absence of glucose: 7V-phosphoryl-glutamic acid and the
diethyl ester of glutamic acid. 7V-phosphoryl-glutamic acid is relatively
unstable and decomposes to release glutamic acid, but addition of the
substance to washed suspensions of Staph. aureus is followed by the
appearance of free glutamic acid within the cells at a rate which varies
widely but has been as high as 75 % of that obtained in the control incubated
with glucose and glutamic acid. The temperature coefficient for the over-
all process is 2-4, and it is possible that phosphoryl-glutamic acid may be
acting as an energy source rather than entering the cell unchanged. In the
presence of glucose, the internal accumulation of glutamic acid occurs at
the same rate whether the external source is glutamic acid or the N-
phosphoryl derivative. The diethyl ester of glutamic acid gives rise to free
glutamic acid within the cells at a slow rate with a low temperature
coefficient (1*8). The esterase activity of these cells is low, and the slow rate
of accumulation of free glutamic acid within the cells is probably due more
to a slow rate of de-esterification than to a slow penetration. No evidence
could be obtained for esterification of glutamic acid by the cells.
A number of simple peptides containing glutamic acid have also been
tested (see Table 2), but in no case has free glutamic acid appeared within
cells incubated in the absence of glucose. In the presence of glucose, some
peptides, e.g. a-glutamyl-valine or a-glutamyl-leucine, give rise to
glutamic acid within the cells more rapidly than does glutamic acid itself
as external source. a-Glutamyl-glutamic acid or a-glutaminyl-glutamic
WITHIN STAPHYLOCOCCAL CELLS 249
acid give rise to free glutamic acid within the cells but at a slower rate than
glutamic acid itself. There is a suggestion that the effectiveness of the
external source may be related to its lipoid solubility, but further investiga-
tions are needed with a wider range of peptide structures.
Table 2. Glutamic acid derivatives giving rise to free
glutamic acid within Staphylococcus aureus
Comparative
rate
Temperature
coefficient
20-30° C.
A. Free glutamic acid obtained only in
presence of glucose
a-L-Glutamyl-L-leucine
146
2'33
a-L-Glutamyl-L-valine
140
2-25
L-Glutamic acid
100
2'8
a-L-Glutamyl-glycine
88
—
a-L-Glutaminyl-L-glutamic acid
69
2'5
L-Glutamine
60
2-4
a-L-GIutamyl-L-glutamic acid
45
2'0
Glycyl-L-glutamic acid
22
y-L-Glutamyl-L-valine
20
y-L-Glutamyl-L-leucine
16
Glutathione
15
2'5
y-L-Glutamyl-L-glutamic acid
13
2'4
y-L-Glutamyl-L-aspartic acid
5
—
y-L-Glutamyl-glycine
3
—
B. Free glutamic acid obtained in absence of
glucose
JV-phosphoryl-L-glutamic acid
40-75
2*45
Diethyl-L-glutamic ester
10-35
1-8
Comparative rate gives the rate of appearance of free glutamic acid within the cells
incubated at 37° C. compared with the rate when glutamic acid and glucose form the
external source.
Rothstein & Meier (1948), using isotopic methods, demonstrated the
presence of phosphatases at the surface of yeast cells, while Barron, Muntz
& Gasvoda (1948) showed that glucose oxidation by yeast could be inhi-
bited by uranium ions and the inhibition released by addition of phosphate.
Since cell-free enzyme preparations were not affected in the same way, the
suggestion was put forward that the uranium was acting by combination
with proteins on the cell surface which were responsible for the oxidative
processes. McQuillen (1950^) has shown that the staphylococcal cell will
adsorb uranium ions, and we have further found that saturation of the cell
surface with uranium will prevent the accumulation of glutamic acid within
the cell. Again, the inhibition can be reversed by washing the cells in
phosphate buffer. These findings are consistent with the hypothesis that
factors on the surface of the cell play a role in the over-all accumulation
process.
250 THE ACCUMULATION OF AMINO-ACIDS
Free penetration: active accumulation
We can now turn our attention to the third hypothesis which postulates
that glutamic acid penetrates the cell-wall structures and that a non-
diffusible derivative is formed inside the cell. First we must consider the
grounds on which the amino-acid within the cell is called 'free glutamic
acid'. In our studies at Cambridge we have estimated glutamic acid by the
specific decarboxylase method (Gale, 1945). Glutamic decarboxylase will
attack the L-isomer of glutamic acid only, and requires that both carboxyl-
groups and the amino group are unsubstitut'ed (Gale, 1946). The enzyme
will not attack any peptide of glutamic acid. Glutamine and JV-phosphoryl-
glutamic acid are attacked only after the substituting groups have been
removed ; the crude preparations of decarboxylase contain enzymes which
will carry out these removals. The material which is liberated from within
the staphylococcal cell is attacked by the enzyme at exactly the same rate
as glutamic acid. Since the enzyme cannot penetrate the intact cell, it is
only possible to test the action of the enzyme on the intracellular material
after its release from the cells, and the necessity for rupturing or altering
the cell structure introduces doubt concerning the identity of the material
investigated and the material as it exists in the untreated living cell.
Methods that have been used to release the internal material are
(1) subjecting the cells to 100° C. for 10-15 min.;
(2) shaking the cells with small glass beads at 50 vibrations/sec, for
15-20 min. in the cold;
(3) subjecting the cells to supersonic vibration (20,000 cyc./sec.) for
5-10 min. in an ice-cooled vessel;
(4) treating the cells with acetone-ether mixtures or with 5 % trichlor-
acetic acid;
(5) treating the cells with bactericidal concentrations of tyrocidin or
detergent substances such as cetyl-trimethyl-ammonium bromide;
(6) crushing the cells under pressure at -15° C. in the Hughes press
(Hughes, 1951).
In all cases material is obtained which contains glutamic acid when tested
with the specific decarboxylase or by chromatographic procedures. Never-
theless, it is clear that many of these treatments are drastic and might well
cause decomposition of labile derivatives either as a result of heating or of
the liberation of catabolic enzymes. Even the material obtained from
method 6 must be thawed before it can be tested. Although no evidence has
been obtained that the internal glutamic acid is other than free glutamic
acid there can be no confidence that some alteration in its nature has not
occurred at some stage in the procedure.
WITHIN STAPHYLOCOCCAL CELLS 251
Cowie, Roberts & Roberts (1949), Roberts, Roberts & Cowie (1949) and
Roberts & Roberts (1950) investigated the somewhat similar situation that
arises when potassium accumulates within Escherichia coli during glucose
metabolism. The cells do not accumulate potassium in the absence of
glucose breakdown, and the process has the appearance of involving active
transport. By the use of isotopically labelled metals, these authors showed
that free diffusion of both Na and K ions occurs across the bacterial cell-
wall, and that rapid equilibration occurs between the external medium
and the 'water space' within the cells. From the sodium distribution they
calculated that the water space occupied approximately 70 % of the cell
volume. When the cells were actively metabolizing glucose, the uptake of
potassium increased and the accumulated potassium appeared to be held
in a non-diffusible form within the cells. It was suggested that potassium
combined with breakdown products of glucose to produce the non-
diffusible material. The accumulation was therefore not due to active
transport of potassium across an impermeable barrier but to the formation
of a non-diffusible metabolic product within the cell, free diffusion of
potassium ions still occurring between the internal and external media.
The situation therefore corresponds to hypothesis 3.
The experimental findings concerning glutamic acid accumulations are
superficially very similar to those found for potassium accumulation in
Esch. coliy so Britten (1952, 1953) undertook an investigation with
Staphylococcus aureus using 14C-labelled glutamic acid. In the first place
he found free exchange between glutamic acid of the 'water space' in the
cells and of the external medium whether metabolism of glucose was
occurring or not. Over and above this exchange, however, he found that,
when active metabolism was occurring, glutamic acid became stored
within the cells in a form which was less readily exchangeable with the
external medium. If the temperature was dropped to 4° C., exchange
between external glutamic acid and 'stored glutamic acid' fell to insigni-
ficant values. It was further found that, if the cells were incubated with
glutamic acid and then removed from solution, the amount of 'stored
glutamic acid' continued to increase at the expense of the glutamic acid
in the water-space of the cells. The picture that develops from these
investigations is that glutamic acid passes freely into the water space of the
cells and, when glucose metabolism occurs, enters into some metabolic
process which results in the formation of a non-diffusible form, this non-
diffusible material constituting the 'free glutamic acid' of the previous
studies. Britten was again unable to distinguish between the 'stored
glutamic acid' after liberation from the cell and free glutamic acid, but,
unless the interpretation of this type of experiment is wrong, there would
252 THE ACCUMULATION OF AMINO-ACIDS
appear to be no escape from the conclusion that the internal glutamic acid
cannot be free glutamic acid. It may be some highly labile derivative or
possibly it is combined in some labile manner with the cell proteins.
We know little or nothing of the interior organization of the bacterial cell.
A botanist presented with a situation similar to that discussed here might
suggest that free diffusion of potassium and glutamic acid occurs into the
internal medium of the cell and that, when glucose is available, active
transport into, and accumulation within, a vacuole takes place. I do not
know whether vacuoles exist inside the staphylococcus. Britten calculates
that the water-space of the staphylococcus occupies 35-45% of the cell
volume, so there may be room for a small vacuole in the rest of the cell.
If there is, then our problem is merely removed from the door-step to the
serving-hatch but, for the present, there would seem little point in postu-
lating a hole in a bacterium to explain a hole in a postulate.
REFERENCES
BARRON, E. S. G., MUNTZ, J. A. & GASVODA, B. (1948). J. Gen. Physiol. 32, 163.
BRITTEN, R. (1952). Yearb. Carneg. Instn, p. 92.
BRITTEN, R. (1953). In preparation.
CHRISTENSEN, H. N. & RIGGS, T. R. (1952). J. Biol. Chem. 194, 57.
CHRISTENSEN, H. N., RIGGS, T. R., FISCHER, H. & PALATINE, I. M. (19520).
y.Biol. Chem. 198, i.
CHRISTENSEN, H. N., RIGGS, T. R., FISCHER, H. & PALATINE, I. M. (1952^).
J. Biol. Chem. 198, 17.
COWIE, D. B., ROBERTS, R. B. & ROBERTS, I. Z. (1949). J. Cell. Comp. Physiol. 34,
243-
DAVIES, R., FOLKES, J. P., GALE, E. F. & BIGGER, L. C. (1953). Biochem. J.
54, 430.
ELLIOTT, W. H. (1951). Biochem. J. 49, 106.
ELLIOTT, W. H. & GALE, E. F. (1948). Nature, Lond., 161, 129.
GALE, E. F. (1945). Biochem. jf. 39, 46.
GALE, E. F. (1946). Advanc. Enzymol. 6, i.
GALE, E. F. (1947). J. Gen. Microbiol. i, 53.
GALE, E. F. (1948). Symp. Soc. Exp. Biol. 3, 233.
GALE, E. F. (1949^). Johns Hopk. Hosp. Bull. 83, 119.
GALE, E. F. (19496). y. Gen. Microbiol. 3, 369.
GALE, E. F. (1951). Biochem. y. 48, 286.
GALE, E. F. (1953). Advanc. Protein Chem. 8 (in the Press).
GALE, E. F. & TAYLOR, E. S. (1947). J. Gen. Microbiol. i, 314.
GOURLEY, D. R. H. (1952). Arch. Biochem. Biophys. 40, i, 13.
HILLS, G. M. (1940). Biochem. y. 34, 1057.
HUGHES, D. E. (1951). Brit. y. Exp. Path. 32, 97.
KREBS, H. A. (1950). Biochem. y. 47, 605.
McQuiLLEN, K. (19500). Thesis: Electrophoresis of [Bacteria. Cambridge
University Library.
McQuiLLEN, K. (19506). Biochim. biophys. Acta, 6, 66.
MITCHELL, P. D. (1949). Symposium. Nature of the Bacterial Surface.
NAJJAR, V. A. & GALE, E. F. (1950). Biochem. J. 46, 91.
WITHIN STAPHYLOCOCCAL CELLS 253
PAINE, T. F. Jnr. (1951). J- Bact. 61, 259.
ROBERTS, R. B. & ROBERTS, I. Z. (1950). J. Cell Comp. Physiol. 36, 15.
ROBERTS, R. B., ROBERTS, I. Z. & COWIE, D. B. (1949). J. Cell Comp. Physiol.
34, 249-
ROTHSTEIN, A. & MEIER, R. (1948). J. Cell. Comp. Physiol. 32, 261.
TAYLOR, E. S. (1947). J. Gen. Microbiol. i, 86.
WAELSCH, H., OWADES, P., MILLER, H. K. & BOREK, E. (1946). J. Biol. Chem.
166, 273.
TRANSPORT OF PHOSPHATE THROUGH AN
OSMOTIC BARRIER
BY P. MITCHELL
Department of Biochemistry, University of Cambridge
Much of the present-day knowledge of biochemistry is centred about the
part played by phosphate in coupling thermodynamically natural with
thermodynamically unnatural processes in living organisms. Nevertheless,
very little is known of the mechanism of transfer of phosphate molecules
through the membranes which form the connecting links between biological
media. The view, expressed nicely by Rosenberg (1948), that by virtue of
specific permeability properties, the natural membranes act as connecting
links between particular components of the phases which they separate has
its counterpart in the view of the enzymes as couplers of reactions which
can proceed only on or in the enzyme molecules. Rosenberg's treatment
shows, in fact, that the energetics of the reactions in two phases connected
by a membrane can be described in the same terms as 'homogeneous'
enzyme-linked reactions ; the important implication being that the efficiency
(or reversibility) of transport reactions is determined by the specificity
of membrane permeability, exactly as the efficiency of coupled enzyme
reactions is determined by the enzyme-substrate and enzyme-carrier
specificities. This merging of the terms of description has appropriately
coincided with the realization that the permeability properties of mem-
branes to the substrates which they transport may be dominated by enzymic
specificities. In complex biochemical systems, such as those carrying out
oxidative phosphorylation (e.g. Slater & Cleland, 1953), the osmotic and
enzymic specificities appear to be equally important and may be practically
synonymous.
Although the study of membrane-transport phenomena in combination
with active metabolism has been responsible for great advances in the out-
look upon permeability problems, it is unfortunate that the experimental
systems in which active transport can be studied should, in general, contain
so many variables. This would seem to be particularly true in the case of
phosphate transport, since the phosphate molecules may be involved at
several points in the train of energy-yielding reactions required to drive the
active transport. There is no a priori reason to suppose, however, that the
transport reaction may not be studied when isolated from the energy-
yielding reactions. Ussing (1947) has proposed to account for the passive
TRANSPORT OF PHOSPHATE THROUGH OSMOTIC BARRIER 255
one-to-one exchange of cations across a membrane by a process, referred
to as exchange- diffusion, in which the ions are supposed to saturate a
carrier, in combination with which they pass across the membrane which
they cannot traverse in the free state. In 1950, Roberts & Roberts, using
32P, found that a one-to-one exchange of inorganic phosphate molecules
occurred across the cell membrane of resting Bacterium coli. They main-
tained, however, that the so-called inorganic phosphate of Bact. coli
(extracted in dilute trichloroacetic acid) was not free, but was adsorbed on
a fixed number of acid-labile sites within the cells, the membrane of which
they considered to be freely permeable to phosphate ions. The one-to-one
phosphate exchange was therefore described by the self-explanatory term
exchange-adsorption. Roberts & Roberts ( 1 950) were not alone in suggesting
that the so-called internal inorganic phosphate might be fixed in an adsorp-
tion complex within the cells. Similar suggestions were, for instance, made
by Kamen & Spiegelman (1948) working on baker's yeast, Green, Atchley,
Nordman & Tepley (1949) working on the cyclophorase system of
mamalian tissues, and Harman (1950) working on mammalian heart-
muscle mitochondria.
An exchange of inorganic phosphate across the cell surface, apparently
similar to that reported for Bact. coli, was observed in a Staphylococcus by
Mitchell & Moyle (1953). The rate of the exchange in resting cells suspended
in a balanced salt medium containing 10 mM-phosphate at neutral pH was
i-4/miole/g. cell dry wt./min. at 25° C. The reciprocity of the exchange
was extremely strict when precautions were taken to minimize residual
respiration (Mitchell, 1953 0), the net flux being much less than i % of the
exchange flux. When, however, the cells were allowed to respire in the
presence of glucose, for which they possess powerful oxidative systems, a
net flux of phosphate was directed inward at the rate of i-4//mole/g./min.
as before, and the outflux fell virtually to zero (Mitchell & Moyle, 1953).
We therefore considered it probable that the same transport mechanism
was employed in both the uptake and the exchange reactions. This being
the case, it occurred to us that a study of the latter reaction might yield in-
formation about the transfer process because of the elimination of the large
number of variables connected with the energy-yielding reactions. It was
necessary, however, to postpone any study of the exchange reaction per se
until it could be decided whether the uptake of inorganic phosphate repre-
sented a true active transport of phosphate across an osmotic barrier
impermeable to phosphate ions, or whether it represented an adsorption
process like that considered by Roberts & Roberts (1950).
Two kinds of experimental approach towards a decision between the
alternative exchange hypotheses were adopted. The first was to determine
256 TRANSPORT OF PHOSPHATE THROUGH
whether or not an osmotic barrier impermeable to phosphate ions exists
near the cell surface. This could be done by determining the degree of
dilution of a known addition of phosphate to a thick cell suspension and
estimating the part of the volume of the suspension accessible to the
externally added phosphate. If a phosphate-impermeable barrier were to
exist near the external surface of the staphylococci, the effective cell volume
should be approximately equal to the volume of a close-packed centrifuged
pad less 26 % , the interspace volume for close-packed spheres (Conway &
Downey, 1950). If, on the other hand, there were no osmotic barrier for
phosphate, or if the barrier were broken by reagents such as trichloroacetic
acid or butanol, the effective cell volume should be approximately equal to
the apparent specific volume of the materials of the cell, namely, c. 0-8 ml./g.
The effective phosphate-impermeable volume of normal staphylococci in
o- 1 M-NaCl was found to be 2-42 ± 0-05 ml./g. cell dry wt., in tolerably good
agreement with the figure 2-67 + 0-01 ml./g. for the close-packed cell
volume less 26%. After treatment with dilute trichloroacetic acid or n-
butanol, the phosphate-impermeable volume decreased to 0-76 ± 0-07 ml./g.
cell dry wt., which, on correction for loss of diffusible constituents, gave
0*93 ± 0*09 ml./g. (Mitchell, 1953 a). These observations were in accord
with the view, based upon chemical analyses of the components of the cell
envelope, that the osmotic barrier of staphylococci is a lipoprotein layer
c. 1 5m// thick which is supported by an external cell wall c. 25 m/i thick
(Mitchell & Moyle, 1951). There is therefore little doubt that an osmotic
barrier impermeable to phosphate ions exists near the surface of the cells.
This would imply either that the resting exchange of phosphate must occur
by exchange-diffusion across the barrier, or that the so-called inorganic
phosphate must be adsorbed outside the barrier. We can calculate that the
effective internal cell volume was 1-66 ± 0-07 ml./g. cell dry wt. and that the
normal inorganic phosphate content (c. i5o/miole/g.) would give a con-
centration of c. o-i M-phosphate in the internal medium. The concentration
of phosphate that would be required if it were all present outside the barrier
seems almost to preclude this possibility. However, a more direct con-
firmation of the occurrence of exchange diffusion was sought by studying
the sensitivity of the exchange reaction to inhibitors.
If the exchange-adsorption hypothesis were correct, the so-called in-
organic phosphate of the cell would be held by combining groups corre-
sponding in number to the acid-soluble inorganic phosphate molecules.
The inhibition of the exchange of phosphate on a combining group would
not be expected to occur unless the inhibitor molecule were close to the
combining group. Hence, unless action at a distance were postulated, the
number of molecules of inhibitor required to retard the exchange reaction
AN OSMOTIC BARRIER 257
should be of the same order as the number of molecules of acid-soluble
inorganic phosphate. If, on the other hand, the exchange- diffusion
hypothesis were correct, the exchange would be mediated by a relatively
small number of carrier groups in the osmotic barrier. In this case, an
inhibitor might retard the exchange reaction when the number of molecules
present was similar to the number of carrier groups and therefore small in
comparison to the number of molecules of acid-soluble inorganic phosphate.
Out of a large number of inhibitors, the heavy metals and their derivatives
were found to be the most potent group, headed by phenyl-Hg+. The
relationship between the degree of inhibition of the exchange reaction and
the concentration of phenyl-Hg+ was represented by K=Mnl(ioo — ri),
M being the amount of phenyl-Hg+, n the percentage depression of the
rate of phosphate exchange and K a constant. This indicated a reaction of
the type M+X=MXy X representing the sites which, when combined
with inhibitor (as MX), cause inactivation of a corresponding proportion
of the units controlling the exchange of phosphate. The value of K was
2-2//mole/g., and since the cells contained 147/miole acid-soluble inorganic
phosphate per g., the value of X per 100 molecules of acid-soluble phosphate
or hypothetical adsorption sites was only 3 (Mitchell, 1953 a).
There remained little doubt that the exchange of phosphate across the
osmotic barrier in resting cells is due to some kind of exchange-diffusion
system, and that the uptake of phosphate during respiration is due to a
coupling of the exchange-diffusion system with other reactions. A more
detailed study of the exchange system in resting cells was therefore pursued
(Mitchell, 19536).
The effectiveness of inhibitors upon the exchange reaction was studied
to determine the specificity of the reaction for phosphate and to obtain
information about the mechanism of inhibition which might be relevant to
the mechanism of transport of the phosphate molecules across the osmotic
barrier. Out of a total of some forty anions which were tested, some such
as chlorate and borate acted as inhibitors, but only arsenate was able to
substitute effectively for phosphate in the exchange reaction. A remarkable
feature of the behaviour of arsenate, however, was that although it entered
into the exchange system somewhat less readily than phosphate, it did so
with the same efficiency as phosphate: In other words, the same one-to-one
exchange is observed between arsenate and phosphate as between phosphate
and phosphate. It is therefore possible to cause active transport of arsenate
in exchange for phosphate. By washing the cells in i mM-arsenate, the
internal inorganic phosphate, initially at a concentration of c. 100 mM,
was found to pass out in exchange for arsenate which, towards the end of
the exchange process, moved from an external concentration of i mM to an
E B S VIII 17
258 TRANSPORT OF PHOSPHATE THROUGH
internal concentration approaching 100 mM. The exchange of arsenate and
phosphate in this system is analogous to the cation exchange for which
Ussing has postulated exchange-diffusion; but it should be appreciated
that the tightness of the coupling of the arsenate-phosphate exchange is
of a much higher order than that reported (Ussing, 1949) for the cation
exchange.
Of the inhibitory anions, 2 : 4-dinitrophenate was one of the most in-
teresting. At a concentration of i mM at pH 7 it caused c. 50% depression
of the rate of phosphate exchange, approximately the same as the depression
of endergonic processes dependent upon glucose fermentation (Gale,
1951). Aureomycinate inhibited phosphate turn-over to about the
same extent as 2: 4-dinitrophenate. It seems possible that the exchange
reaction upon which 2 : 4-dinitrophenate and aureomycinate act in the
Staphylococcus may have its counterpart in the oxidative phosphorylation
systems of tissue particles which are so effectively decoupled by these
reagents (Loomis & Lipmann, 1948; Loomis, 1950).
Following up the inhibitory action of the heavy metal cations, it was
found that inhibition was caused by substances such as iodine or bromine
at pH 5-5, or by chloroacetophenone or JV-ethylmaleimide at pH 7, but not
by iodoacetate — indicating that the exchange- diffusion reaction involves
thiol groups of low reactivity. The inhibition of the exchange reaction by
the decoupling agents and by the reagents reacting with thiols suggested
that, although the exchange of phosphate across the osmotic barrier did not
in itself involve an over-all free-energy change, a small residual metabolism
might be necessary to allow the exchange reaction to occur, as has been
suggested for the cation exchange in tissue particles (Davies, 1954). It was
found, however, that under conditions in which residual metabolism was
reduced to a level where it could not be detected manometrically, the ex-
change of phosphate continued at the normal rate. The rate of exchange
was therefore presumed to be determined by thermal movements such as
those considered by Ussing (1949) to operate the carrier groups in the
exchange-diffusion systems, and not by the rate of triggering by some
associated energy-consuming metabolic reaction. This being the case, it
was thought that kinetic studies might be amenable of interpretation
(Mitchell, 1953 i).
At an external phosphate concentration of i mM, the velocity of the
exchange reaction exhibited a sharp maximum at pH 7. This maximum,
however, was a function of the external phosphate concentration, the
optimum rate increasing and the optimum pH moving towards 9 as the
external phosphate concentration was increased towards 100 mM. The
dependence of the rate of exchange upon the external phosphate concentra-
AN OSMOTIC BARRIER 259
tion at constant pH and ionic strength followed closely the classical
Michaelis & Menten (1913) law, giving a value of 1-6 mM for the Michaelis
constant (Km) at pH 6-8 and ionic strength o-i. When expressed in terms
of total phosphate concentration the Km value was a function of pH, but
when expressed in terms of the concentration of H2PO^ ion, it was estimated
to have the value 0-8 ± o-i mM independent of pH over a range from pH 5-5
to 8-5.
It seems probable, therefore, that the H2PO^ ion and not the HPO^~ ion
enters the exchange reaction. The rate of the exchange reaction appears to
be a linear function of the degree of saturation of an externally accessible
reactant with H2PO^~ ions in exactly the same way as the rate of an enzyme
reaction is a linear function of the degree of saturation of the enzyme with
its substrate. This does not necessarily mean that the externally accessible
reactant is an enzyme, for in the exchange-diffusion system of Ussing,
provided the movement of the carrier compound from one side of the
membrane to the other were the rate-limiting step, exactly the same type
of kinetics would be expected. However, the phosphate-exchange system
differs from Ussing's in a most important respect. It has been pointed out
by Ussing (1949) that when the carrier compound of his exchange-diffusion
system is not saturated with its substrate on one side of the membrane or
the other, considerable leakage of the substrate across the membrane should
occur. According to Ussing's exchange- diffusion model, the Km (0-8 mM)
of the phosphate-exchange reaction would represent the dissociation con-
stant of the carrier for the H2PO^ ion ; yet, even when the external H2PO^~
concentration was lowered to O'i mM a strictly reciprocal exchange of
phosphate still occurred. The phosphate-exchange system must therefore
differ from Ussing's model in a way which makes it more strictly coupled.
This might be accomplished by a number of obvious mechanisms. Probably
the simplest and most attractive hypothesis, however, is as follows: Phos-
phorylated carriers are present in the osmotic barrier the phosphate groups
of which, due to thermal movements of the carriers, come into contact with
the media on either side where an enzyme-catalysed exchange may occur
between the phosphate groups of the carriers and phosphate ions. The
essential difference between Ussing's system and this one is that in the
latter the free energy of formation of the carrier-ion complex (or compound)
is assumed to be conserved during the enzyme-catalysed exchange of the
ion, the carrier-ion complex thus not being in equilibrium with its dissocia-
tion (or hydrolysis) products on either side of the barrier as is the case in
Ussing's system.
It was found that the value of Km (expressed as a concentration of
H2PO^) was constant to within ± 10% over a temperature range from
17-2
2&O TRANSPORT OF PHOSPHATE THROUGH
5 to 25° C. Thus, the heat of formation of the external phosphate complex
must be small. On the other hand, the temperature coefficient of the rate
of the reaction at constant external phosphate concentration was greater than
10, and gave a constant Arrhenius energy of 38,000 cal./mole over the
temperature range from 5 to 20° C. The heat of activation of the exchange
process (equal to the Arrhenius energy — RT) is therefore a little over
37,000 cal./mole. This heat of activation represents the increase in total
heat required for the thermal movements of the carriers which effect the
translocation of the phosphate groups across the osmotic barrier in the
model system proposed above.
It is hardly necessary to point out that the value of the heat of activation
of the exchange reaction is much higher than that usually observed in
diffusion processes. In fact, it is unusual for reactions with such high heats
of activation to proceed at a measurable rate at room temperature. This,
however, is equivalent to the statement that the free energy of activation of
the exchange reaction (upon which the absolute reaction rate depends) is
probably much smaller than the heat of activation, the larger part of the
heat of activation being due to an entropy change as in reactions such as
the denaturation of proteins (Glasstone, Laidler & Eyring, 1941). The
thermodynamic data are thus in accord with the view that the osmotic
barrier, across which the carrier effects the translocation of phosphate and
of which the carrier forms an integral part, is a well-organized internally
bonded structure. The translocation reaction may be regarded as a kind
of reversible denaturation of the osmotic barrier.
We have no definite information at present as to the chemical nature of
the substance which carries the phosphate across the osmotic barrier of
the Staphylococcus. Gourley (1952) has suggested that adenosine triphos-
phate (ATP) may be the intermediate which carries phosphate across the
membrane of red blood corpuscles; for, during active uptake of isotopic
inorganic phosphate the isotope enters the internal ATP fraction faster than
the internal inorganic phosphate fraction. The phosphate-exchange system
proposed for the Staphylococcus might operate with ATP as the carrier and
phosphokinases as the enzymes, although efforts to locate an externally
situated phosphokinase have failed. If this were the case, however, the
strongly lipophobic properties of ATP would almost certainly have to be
masked by some large molecule such as a protein of low water solubility,
in company with which the ATP would have to pass across the osmotic
barrier by a translational or rotational movement of the large molecule.
Although such a system is not more complicated than we might anticipate
for the organization of biological transfer reactions, it includes rather more
speculation than the present experimental evidence will safely bear. The
AN OSMOTIC BARRIER 261
fact must not be overlooked that Gourley's results might be explained
adequately if the intermediate of the carrier system were to pass its phos-
phate direct to adenosine diphosphate inside the cells during active
phosphate uptake.
REFERENCES
CONWAY, E. J. £ DOWNEY, M. (1950). An outer metabolic region of the yeast cell.
Biochem. J. 47, 347.
DAVIES, R. E. (1954). His paper at this Symposium.
GALE, E. F. (1951). The assimilation of amino acids by bacteria. 10. Action of
inhibitors on the accumulation of free glutamic acid in Staphylococcus aureus
and Streptococcus faecalis. Biochem. J. 48, 286.
GLASSTONE, S., LAIDLER, K. J. £ EYRING, H. (1941). The theory of Rate Processes.
New York: McGraw-Hill.
GOURLEY, D. R. H. (1952). The role of adenosine triphosphate in the transport of
phosphate in the human erythrocyte. Arch. Biochem. 40, i.
GREEN, D. E., ATCHLEY, W. A., NORDMAN, J. & TEPLEY, L. J. (1949). Studies on
the cyclophorase system. XII. Incorporation of 32P. Arch. Biochem. 24, 359.
HARMAN, J. W. (1950). Studies on mitochondria. II. The structure of mito-
chondria in relation to enzymic activity. Exp. Cell. Res. I, 394.
KAMEN, M. D. & SPIEGELMAN, S. (1948). Studies on the phosphate metabolism
of some unicellular organisms. Cold Spr. Harb. Symp. Quant, biol. 13, 151.
LOOMIS, W. F. & LIPMANN, F. (1948). Reversible inhibition of the coupling
between phosphate and oxidation. J. Biol. Chem. 173, 807.
LOOMIS, W. F. (1950). On the mechanism of action of aureomycin. Science, in,
474-
MICHAELIS, L. & MENTEN, M. L. (1913). Die Kinetik der Invertinwirkung.
Biochem. Z. 49, 333.
MITCHELL, P. & MOYLE, J. (1951). The glycerophospho-protein complex envelope
of Micrococcus pyogenes. J. Gen. Microbiol. 5, 981.
MITCHELL, P. & MOYLE, J. (1953). Paths of phosphate transfer in Micrococcus
pyogenes: Phosphate turnover in nucleic acids and other fractions. J. Gen.
Microbiol. 9, 257.
MITCHELL, P. (1953 a). Transport of phosphate across the surface of Micrococcus
pyogenes: Nature of the cell 'inorganic phosphate'. J. Gen. Microbiol. 9,
273-
MITCHELL, P. (19536). Transport of phosphate across the osmotic barrier of
Micrococcus pyogenes: Specificity and kinetics. J. Gen. Microbiol. (in the
Press).
ROBERTS, R. B. & ROBERTS, I. S. (1950). Potassium metabolism in Escherichia coli.
III. Interrelationship of potassium and phosphorus metabolism. J. Cell.
Comp. Physiol. 36, 15.
ROSENBERG, T. (1948). On accumulation and active transport in biological systems.
Acta chem. scand. 2, 14.
SLATER, E. C. & CLELAND, K. W. (1953). The effect of tonicity of the medium on the
respiratory and phosphorylative activity of heart-muscle sarcosomes. Biochem. J.
S3, 557-
USSING, H. H. (1947). Interpretation of the exchange of radio-sodium in isolated
muscle. Nature, Lond., 160, 262.
USSING, H. H. (1949). Transport of ions across cellular membranes. Physiol.
Rev. 29, 127.
ANION RESPIRATION
THE EXPERIMENTAL BASIS OF A THEORY OF
ABSORPTION, TRANSPORT AND EXUDATION OF
ELECTROLYTES BY LIVING CELLS AND TISSUES
BY H. LUNDEGARDH
Penningby, Sweden
I. INTRODUCTION
The processes of absorption, accumulation and transport of electrolytes
are of universal importance in all living organisms. They appear more
prominently in plants because (i) plant cells accumulate large quantities of
salts from dilute media, and (2) the long-line transport of salts, known as
sap movement, is closely related to this cellular power of active salt trans-
port. ' Active ' here means requiring supply of energy, active transport thus
covering both the case of accumulation against a concentration gradient
and also polar movement of ions.
The question as to the possible mechanism of active transport of cations
and anions has aroused much speculation. The problem may be looked
upon from different angles, and it is probable that active transport of ions
may be realized in different ways. At present, only the theory of anion
respiration is supported by experimental material of a magnitude which
permits the building up of a serviceable theory. My laboratory has devoted
more than twenty years of extensive experimental work to lay the ground
for the theory of anion respiration, and since 1941 brilliant contributions
have emanated from the laboratory of R. N. Robertson. Important
studies were also made in the laboratory of D. R. Hoagland. Owing
to limited space I will restrict myself to a survey of the main facts and
conclusions.
Roots of cereals, primarily spring wheat, have served as the experimental
material in my own laboratory. Roots are very suitable objects, because
they act as the salt pumps of the whole plant and thus reveal both sides of
the problem, i.e. both absorption and exudation of ions. In some other
laboratories, e.g. that of Robertson, slices of storage tissue (carrots, etc.)
served as the main experimental material. It is of considerable interest to
note that the chief results from such different tissues agree fairly well.
ANION RESPIRATION 263
II. THE PRINCIPLE OF ION CARRIERS
Several earlier workers (e.g. Osterhout, Hoagland) observed that the cations
and anions of a salt were frequently absorbed in non-equivalent quantities.
It is a well-known fact that much more nitrate than calcium is absorbed
from calcium nitrate, whereas from potassium nitrate the quantities of
absorbed cations and anions are more equal. It has been shown that the
balance may be moved in one direction or the other by pre-treating the
objects with salts; roots charged with calcium absorb very little of this
cation, whereas the absorption of nitrate is only little influenced (Lunde-
gardh, Burstrom & Rennerfelt, 1932; Lundegardh, 1937). These observa-
tions support the idea of a separate entrance of the cations and the anions
into the cell.
Outgrown plant cells have a large sap space (vacuole) surrounded by a
thin layer of protoplasm, the peripheral part of which is the protoplasmic
membrane. This construction of the cell permits a sharp definition of active
accumulation; it is the absorption of c0M+A~ from the medium and its
accumulation as c±M+A~ in the cell sap, c: being larger than CQ and M+A~
symbolizing a neutral salt. The energy (E) required for the necessary
osmotic work is expressed by the formula E = RTln C^CQ. For a gradient
cilco=looi not uncommon in roots, E amounts to c. 2700 cal./mol. at
20° C.
Most workers in the field of salt accumulation accept the idea of ion
carriers, or large organic ions in the protoplasm, symbolized as R+ and R~,
endowed with more or less specific attraction to salt ions. The idea of ion
carriers goes back to Osterhout and other earlier investigators. It has been
further developed by the work in my laboratory. The entrance of salt ions
into the cell is figured as the formation of compounds M+R~ and R+A~.
The accumulation of the free salt M+A~ is figured as the reversal process,
implying the dissociation of the carrier compounds somewhere inside the
protoplasmic membrane. The complete process of salt uptake follows the
scheme :
a b
Absorption
Translocation
R+A- + M+R-
Accumulation
The supply of accumulation energy may be at a and/or b. It is, further-
more, clear that only one of the ions of the salt MA requires a supply of
accumulation energy. Provided anions are actively accumulated the acid
HA will spontaneously decompose the compound MR, resulting in free
salt MA and regenerated carrier HR able to combine with a new cation
264 ANION RESPIRATION
from the medium. It is not even necessary that the carriers R+ and R~ are
rotating as indicated in the above scheme, because repeated ion exchange
(see below) will serve the same purpose.
The outlined mechanism very probably exists in roots. Some of the
investigations performed in my laboratory have been devoted to the
elucidation of the question of cation carriers R~ and the spontaneous
reaction R- + M+ -> MR.
In the first decade of this century, when I started work on salt uptake in
plant roots, much attention was paid to the interaction of salts and proto-
plasm along the lines of 'salt permeability', implying the passive diffusion
of salts into the living cells. These studies and speculations followed upon
early work in colloidal chemistry, later followed by the modern concept of
macromolecules and their physical chemistry. The protoplasmic membrane
is to-day pictured as a more or less flexible network of large molecules,
forming a mosaic pattern and composed of a variety of chemical com-
pounds. Some of these are certainly non-dissociated, others are more or
less ionized, forming electrically charged islands in the electro-neutral
ground substance of the protoplasmic membrane. In the surface of the
plant roots the ratio of R~ to R+ groups appears to be 100 or more. Varying
conditions of dissociation may, however, severely change the balance.
In an aqueous medium the carriers are balanced by small movable ions
of water or neutral salts. As shown below, some of the carriers, if not all,
are more or less specific to the attracted small ions of opposite charge, thus
inducing a selective absorption. The bonds between carrier and movable
ion are similar to those characteristic of 'adsorption' and obey the law of
mass action (see Lundegardh, 1941). The phenomenon of ion exchange is
a manifestation of these principles.
Ion exchange is the main path of non-metabolic salt absorption by
protoplasm as was demonstrated in extensive experiments (see below).
Simple diffusion of salts plays a more subordinate role, as is shown by the
absence of diffusion potentials (Lundegardh, 1938) and by direct experi-
ments with strong solutions with respiration inhibited (Hoagland &
Broyer, 1942). Judging from experiments on the exchange of cations the
total 'exchange capacity' (sometimes called 'free space') is of such a
magnitude that it probably includes most of the protoplasm. Hope &
Robertson (1953) claim that the concentration of the R~~ groups may be as
high as io~2 mol. Measurements of the surface potential give values of
io~2-io~3 mol. as shown in my work on these questions.
ANION RESPIRATION 265
III. THE SURFACE POTENTIAL OF THE PROTOPLASM
The dissociation of the protoplasm creates an electrokinetic potential
difference between the surface layer and the medium. In the case of
single cells, e.g. bacteria or erythrocytes, it can be measured electro-
phoretically. In the case of whole organs, such as roots, it may be measured
as electromotive power. In my experiments during 1938-41 the root
potential was measured by means of a cathode-ray oscillograph, permitting
the recording of very rapid changes. The surface potential of root hairs can
be measured electrophoretically (Lundegardh, 1941) if they are observed
by microscopy in a field of alternate current (2 V. and 20 cycles/sec.). The
amplitude of the oscillations in the field is a measure of the magnitude
of the charge. It was shown that the amplitude is approximately propor-
tional to the pH, if diluted HC1 is used as medium, the vibrations ceasing
at pH 3 in the case of wheat roots (see Lundegardh, 1941, fig. 4). Whole
roots submerged in a dilute nutrient solution to a depth of 20-30 mm.
show a negative electrokinetic potential of c. 60 mV. That the site of the
charge is the surface of the living roots was demonstrated from the ex-
tremely rapid time course of potential changes induced by a change of the
ionic composition of the medium.
By observation of the decrease in potential on increasing the H-ion con-
centration of the medium it is possible to plot a titration curve of the
protoplasmic membrane and from this to calculate its apparent dissociation
constant (see Lundegardh, 1941, fig. 12). The surface of the living epidermis
cells of the wheat roots behaves as a comparatively strong acid, the pK
being 1-2. No carboxylic acids attain this high dissociation, only sub-
stituted phosphoric or sulphonic acids. The latter are absent from the cell
surface. It was accordingly concluded that phosphate groups in organic
linkage are probably serving as carriers of cations in the root surface. This
conclusion is to some extent supported by a later discovery of nucleotides
among the normal exudates from the living root surface (Lundegardh &
Stenlid, 1944). Also observations on bacteria support the idea of nucleotides
present in the surface of the cell.
According to our conception of the mosaic pattern of the surface of the
protoplasm the electrokinetic potential is an expression of the dominating
acid dissociation, and it behaves essentially as a Donnan potential. If the
medium is changed from distilled water or a very diluted mineral acid to
solutions of neutral salts, the negative potential is lowered towards zero at
increasing concentration of the salt and finally changes to positive. This
fact is explained as follows. In the absence of metallic cations in the medium
the R- groups are balanced by H+. The root surface behaves as an H-ion
266 ANION RESPIRATION
electrode, the measured potential being an expression of log H^"/H^(", if
HjJ" represents the cH of the medium and Hf the cH of the root surface.
If a neutral salt is added the M-ions exchange with H^~ to an extent deter-
mined by the activity and a factor a, representing the specific adsorption
qualities of the cation in question (see Lundegardh, 1941). The potential
is accordingly lowered, the drop being approximately proportional to the
logarithm of the salt concentration (see Lundegardh, 1940, fig. 5; 1941,
%• IS)-
During prolonged exposure to a salt solution these exchange processes
in the surface, which are so excellently reflected in the electrokinetic
potentials, are followed by similar processes in the bulk of the protoplasm
resulting in a chemically measurable cation exchange between a root and
the surrounding medium (see Lundegardh et al. 1932; Lundegardh, 1945,
1 950 a). These observations teach us that cation carriers are also present
in considerable quantities in the bulk of the protoplasm. A useful tool for
measuring the quantities of cations absorbed by the root and the quantities
of other cations simultaneously given off is an application of the method of
quantitative spectrum analysis which I introduced (1929). By means of
this technique it was shown (Lundegardh et al. 1932) that in the alkali
series K is most intensively absorbed, followed by Cs and Rb. Of the highly
hydrated Li less is absorbed, and of Na, the hydration of which is much
higher than that of K, only small quantities are absorbed. The simul-
taneously exuded quantities of cations reflect the power of ion exchange.
Charging of the roots with Ca considerably increases the exchangeable
quantities of this ion. As expected, H+-ions have a high exchange power.
From electrically recorded salt absorption experiments (Lundegardh,
1949*:) it was calculated that a solution of 0-0005 mol. HC1 exchanges
quantities of K from the roots which considerably exceed the quantities of
this ion actively absorbed from a o-ooi mol. solution of KC1. The pre-
dominating power of cation exchange is reminiscent of soil colloids, with
the difference that living protoplasm absorbs cations more intensively
than soil colloids do. The dominating attraction of cations, however, does
not exclude a certain exchange of anions, as may be observed from potential
experiments with salts of different anions and the same cation, or from the
reversal of charge in strong salt solutions. The normally low surface con-
centration of the anion carriers R+ (see above) hampers the analytical
determination of the anion exchange, but there is little doubt that also in
respect of anions exchange processes are at work. In several papers I have
pictured the exudation of salts from the protoplasm into the vacuoles as
a passive exchange process. Some investigators, e.g. Hoagland, Arisz
(1952) and others, believe in an active excretion through the tonoplast.
ANION RESPIRATION 267
Their arguments are not convincing, and a fact directly speaking against
an active excretion into the vacuoles is the rapid outflow from the tissue
at times of salt starvation (see Lundegardh, 1945, fig. 2). The observed
facts support the idea of an exchange equilibrium between protoplasm and
vacuole, the latter serving as a transitory store of salts. There is no conflict
between this idea and the frequently observed slow rate of diffusion through
the tonoplast.
IV. NON-METABOLIC ABSORPTION AND EXUDATION OF
IONS. POSSIBLE MECHANISMS OF SALT ACCUMULATION
OTHER THAN ANION RESPIRATION
Rotation of the carrier molecules at the cell surface serving as ion acceptors
will bring the attached ions into the protoplasm, and they are then taken
over by other carriers circulating in the protoplasm (fig. i). One may speak
Fig. i . Scheme of the absorption of cations by carrier molecules and their transport
inwards by means of rotation of the carriers.
Fig. 2. Scheme of rapid transport of ions along 'adsorption tracks'.
of 'adsorption tracks' or 'surface gliding of ions' if the ions are inter-
mittently pushed and pulled along structures (see the scheme in Fig. 2).
It was experimentally shown (Lundegardh, 1945, 1950) that this movement
is non-metabolic, i.e. not linked to respiration. In the case of roots, metallic
268 ANION RESPIRATION
cations will dominate over nutrient anions owing to the predominating
acidic qualities of the cell surface, but certain quantities of salts will never-
theless always be available for passive transport by means of the trans-
piration stream or for metabolic purposes.
In the case of metabolic taking over of ions from their carriers to other
chemical systems, one may speak of active salt absorption at b in scheme (i).
But the common metabolic activity of the cell offers more direct mechanisms
of accumulation. Two of these may be mentioned here, namely, variations
in the dissociation of the carriers R, and variations in the acid/base balance.
The quantitity of carried ions (MR and RA) corresponds to the quantity
of H+ and OH~ held in exchangeable state by the large carrier ions. The
dissociation is regulated by the pH of the surroundings, and a lower pH
at the cell surface in combination with a higher pH in the cell sap, as
realized in many plant cells, undoubtedly favours an import of metallic
cations. But such conditions favour either cations or anions, never both.
If, however, the dissociation constant (k) of the carrier is changed, con-
ditions may be fulfilled for a real accumulation mechanism, according to
the following scheme (k1>k0):
Absorption Metabolic Accumulation
(2)
• *
Transformation
Transformation
A change in the acid dissociation constant of R~ groups is probably
induced by certain ions, e.g. Ca, Mn and BO3, because these ions maintain
a high negative potential of the root surface whereas K-ions act in the
reversed direction (Lundegardh, 1940). An increase of the concentration
of R~~ in the surface will, however, give the same result. Changes in the
dissociation constant of large molecules, induced by small stoichiometric
changes, are well known. In the chain of reactions participating in aerobic
respiration, changes of the dissociation constant occur at many points,
especially in connexion with the phosphate metabolism.
Variations in the acid/base balance have been studied from many angles.
Of special interest are metabolic regulations of the internal pH of root cells
(Ulrich, 1941; Burstrom, 1942). At storage of organic acids, e.g. malic
acid, in the cells, metallic cations are released from the protoplasm in
exchange for H-ions. More cations are concomitantly absorbed from
the medium. If some of the acid is consumed in the tricarboxylic cycle the
n
Absorption
M+
^- r\
^ /OA\
^* f\A
t
*- M (f)A\
— ^- ri n
Accumulation
>• (CM)
t
Production
>• n (UA)
Consumption
ANION RESPIRATION 269
remaining metallic cations will release adsorbed anions from the protoplasm
and more anions will then be absorbed from the medium. As freely
movable anions are not so easily available, owing to the dominating acidoids
in the protoplasm (see above), supernumerary metallic cations may also be
exuded from the root surface. For maintaining a continuous accumulation
of salts a metabolic mechanism must be thought to produce acids in one
region or layer of the cell and to consume acids at another place according
to the following scheme (OA = organic acid) :
(3)
Some authors believe that such a scheme is realized in connexion with
the activity of the cytochrome system (Helder, 1952; Vervelde, 1952), but
this is a mere speculation. In cells deprived of a cytochrome system
organized for polar transport (see below) mechanisms operating by means
of potential gradients controlled by dissociation and/or continuously
working cycles of the acid/base balance may possibly exist. But such cases
will have to be experimentally demonstrated. A specific absorption of
anions may be expected if they are caught by a special carrier also serving
as a mechanism of accumulation. It is commonly believed, though never
strictly proved, that phosphate is absorbed by the aid of phosphorylation.
Phosphorylation at the surface of the cell and dephosphorylation inside of
it will of course serve as a suitable mechanism for accummulation of
phosphate. The fact that the absorption of sugar is probably linked to a
special mechanism working in the cell surface (cf. Lundegardh & Burstrom,
1944) supports this idea, because a reversible phosphorylation is probably
acting here too.
From the point of view of specific ion carriers attention is called to the
coenzymatic function of a number of cations, Ca, K, Mn and Mg. The
selective absorption of cations has been known for a long time. In experi-
ments on the absorption of the members of the alkali series (Lundegardh
et al. 1932) it was observed that K was absorbed more rapidly than would
be expected from its physico-chemical properties (hydration, dimensions,
etc.). From equimolar solutions (0-0025 m.) of chlorides the relation K/Na
amounted to 24. Li is considerably more hydrated than Na, but the
relation K/Li nevertheless amounted to only 1-8. The slow entrance of Na
270 ANION RESPIRATION
is obviously caused by a low percentage of carriers suitable for this ion.
The selectivity is frequently a group character. It was observed that K
and Rb behave very similarly (Burstrom, 1937). The recent results of
Epstein (1952) only corroborate the earlier research work.
V. THE SAP MOVEMENT
Salts which have penetrated the epidermis layer are comparatively rapidly
distributed through the root tissue. This internal transport is partly inde-
pendent of respiration and glycolysis (Lundegardh, 1950). It is probably
facilitated by the absence of potential differences between the internal
surfaces of the cells and by exchange processes. The rapidity of ion exchange
processes was experimentally demonstrated (see above). Owing to the
mobility of salts stored in the root cortex these are continuously emptied into
the open central vessel of the roots (Fig. 3). Under anaerobiosis the
exudation continues, but with decreasing speed and with a certain
preference for cations (Lundegardh, 1945, 1950).
The non-metabolic character of these internal movements of salts is
emphasized by the low temperature coefficient, the QIQ amounting to only
1-5-1-6, whereas the Qw of active accumulation amounts to 2-25-2-37
(Lundegardh, 1950, p. 146). Wanner (1948) found a QIQ of 1-4 for the
absorption of K-ions from the medium, but £)10 = 2 to 2-5 for the absorption
of the nitrate ion. These results point to the existence of separate mechanisms
for the accumulation of cations and anions through the epidermis, but the
translocation inside the epidermis may be in part non-metabolic as shown
by a lower Q1Q for both ions. My own results show that salts may be
passively transported through the root tissue by means of non-metabolic
processes (probably a diffusion enhanced by ion exchange along * adsorption
tracks'). Only active transport yielding accumulation against the concen-
tration gradients requires the assistance of metabolic processes.
The concentration of the exuded sap is normally higher than that of the
nutrient solution (Lundegardh, 1943, 1945, 1950*)- In wheat it usually
amounts to 15-25 mmol. KNO3 (the dominating salt) per litre, but may
rise to 37 mmol. x I.-1. This concentration is, however, lower than the
mean concentration of KNO3 in the root cells, or 40-80 mmol. x I.-1.
From the point of view of exudation into the vessels the process conse-
quently does not involve osmotic work. As shown in a discussion of the
kinetics of salt exudation (Lundegardh, 1950, p. 104), even the exudation
of much higher concentrations need not necessarily involve active work.
This conclusion may hold good also for the highest exudation value ever
observed in my experiments, viz. c. 120 mmol. x I."1 from roots treated
with o-oi mol. NaF at pH 4-4. An indispensable condition for all this non-
ANION RESPIRATION 271
metabolic translocation of salts is of course that a certain level of concentra-
tion has been created by preceding active work (see below).
In principle there is no large difference between the exudation of salts
from the vascular epithelium into the vessels and the exudation of K from
the epidermis into a diluted solution of HC1, mentioned above. Also
quantitatively the two cases are very similar. The influx and outflow of
cations from the epidermis is essentially an ion-exchange process, due to
existing gradients of the concentration (activity) of the involved ions. The
same scheme may also be applied to the internal exudation from the
Epidermis
Fig. 3. Diagram of the main tissue layers in a grass root. From the surface to the centre:
(i) epidermis (single layer of cells without intercellulars); (2) cortex (several layers of large
cells with large intercellulars); (3) endodermis (single layer of cells without intercellulars);
(4) vascular epithelium (tissue of narrow cells surrounding the vessels); (5) central large
vessel.
vascular epithelium. The fact that the outflow from the epidermis is mainly
restricted to cations, whereas normally neutral salts are exuded from the
vascular epithelium, points to a less acid reaction of the exuding surface of
the latter, favouring a more equivalent activity of cations and anions
according to the Donnan principle. The outflow of salts may be pictured
as a 'canalized leakage* (see Lundegardh, 19506, p. 112), viz. a per-
meability restricted to salt ions, escaping from the ionized points of the
cell surface.
Organic substances are not exuded by the vascular epithelium of wheat
seedlings. The outside of the epidermis behaves differently also in respect
of exudation of organic substances which are regularly exuded in varying
quantities (Lundegardh, 1932; Lundegardh & Stenlid, 1944; Stenlid,
272 ANION RESPIRATION
1950). The fall in the concentration in the tissue caused by the internal
exudation into the vessels may be followed by intermittent analysis. It is
reflected in a falling concentration in the exudate. Because water is exuded
into the vessels not only owing to the osmotic conditions of salt exudation
but also owing to metabolic processes, e.g. the glycolytical disappearance of
sugar (see Lundegardh, 19496, 19506), interesting variations of the con-
centration of the sap may be induced under the influence of inhibitors
affecting these processes, but these questions are outside the scope of the
present survey.
VI. DEFINITION OF THE ANION RESPIRATION
Cations and anions may be absorbed, transported and exuded on the basis
of what we have called non-metabolic processes, but a continuous exudation
of salts in the ascending sap requires a continuous renewal of the level of
concentration in the exuding cells, a condition not realized with inhibited
respiration. Under anaerobic conditions salts exude into the vessels for
several hours, but to a slowly sinking degree, and no salts are absorbed
through the epidermis to restore the sinking osmotic gradient. Only an
active anion respiration is capable of filling the salt magazines of the root
cells.
The discovery of quantitative relations between the active absorption of
salts and respiration was published in 1933 by Lundegardh & Burstrom.
Independently, Steward (1932) had shown that aerobic respiration (O2
consumption) is a necessary condition for salt accumulation in slices of
storage tissue. Later Steward, together with Hoagland and other in-
vestigators, insisted upon the thesis that there are no quantitative relations
between aerobic respiration and salt accumulation. Lundegardh &
Burstrom (1935) and Lundegardh (1937 and later papers) were able to
distinguish two groups of aerobic respiration, the ground respiration and
the anion respiration. The former is not inhibited by o-ooi mol. cyanide
and shows no direct relation to salt accumulation. The latter is inhibited
by very low concentrations of cyanide, as also by NaN3 and CO, and its
intensity is closely related to the intensity of salt absorption.
The quantitative relation between absorbed salt and the cyanide-
sensitive respiration is primarily an effect of the anions. Already the
pioneer experiments showed a clear proportionality between absorbed
anions and one fraction of the respiration. No clear proportionality was
traced in respect of the simultaneously absorbed cations. In later in-
vestigations (Lundegardh, 1937) with roots pretreated with Ca in parallel
experiments with Ca(NO3)2 and KNO3 in equinormal concentrations
found approximately the same absorption of nitrate and the same anion
ANION RESPIRATION 273
respiration, whereas no cations were absorbed from the Ca(NO3)2 but
quantities of K from the KNO3 . Other experiments with bicarbonates of
K, Ca and Ba showed no anion respiration, but a considerable uptake of
cations. It is important to note that all these experiments were performed
with carefully treated 2-3 weeks old intact seedlings, ' desalted* by pretreat-
ment with distilled water under illumination in photothermostats. The
experimental vessels were specially constructed to maintain perfect
aeration of the intact plants ('circulation vessels'). Each experiment lasted
from 10 to 20 hr. In later experiments also detached root systems were
submerged in the solution and shorter experimental times were used,
usually i hr. The absorption and exudation of cations and anions was
closely followed by spectrographic and chemical analysis of solutions and
in certain cases also of the roots. In some later experiments the electrical
conductivity was also used as a measure of the concentration of the
medium. It is important, however, to check the conductivity measurements
by chemical analysis (Lundegardh, 1949^).
VII. THE ANION RESPIRATION AS AN ION CATALYSIS
The idea of a respiratory system, operating with active Fe as the source of
accumulation energy, was developed at an early date (Lundegardh &
Burstrom, 1935 ; Lundegardh, 1935, 1937). It was later shown that another
heavy metal, Mn, had a considerable influence on the ground respiration
and on nitrate assimilation (Lundegardh, 1939; Burstrom, 1939). The iron
catalysis was later assumed to be identical with the cytochrome-cytochrome
oxidase system. This idea was accepted and further developed by R. N.
Robertson (1941 and later papers). Robertson devoted extensive attention
to salt respiration in slices of storage tissue, primarily of carrot. He was
able to demonstrate the existence of an anion (or salt) respiration and a
ground respiration, both quite similar to the corresponding processes in
roots. Robertson observed the initial non-metabolic phase of salt absorp-
tion, the rise and the fall of the anion respiration with the salt content of
the medium, its high sensitivity to cyanide, and the restricted passive
leakage of anions from the cells.
The theory of anion respiration in its present shape was outlined by
introducing the hypothesis that the active transport of anions is causally
linked to the transference of electrons in the cytochrome system (Lunde-
gardh, 1945). The leading idea is the postulate that variations in the charge
of two reacting cytochromes, Ct1 and CY2, induced by the transference of
one electron according to the formula
(4)
l8
274 ANION RESPIRATION
are reflected in the distribution of the movable anions in the surroundings,
one anion being attracted by C*2.Fe3+ at the same moment it is uncoupled
from CVFe2+.
In an intact cytochrome-cytochrome oxidase system the electrons are
moved from a dehydrogenase system (DH ; as a rule succinic dehydrogenase)
through the potential ladder of the cytochromes to oxygen. As shown in
the following scheme :
H (from DH)
O2 — * -« c -* b -< 4
H+ (5)
Anion ^ ^~~ ** Anion
anions will concomitantly be moved in the opposite direction. The final
acceptor of the electrons is the O2. The final acceptor of the anions are
cations circulating in the surroundings of b. Protons are produced when
the hydrogen of the reduced dehydrogenase is oxidized by cytochrome b
(or some intermediate factor; see below and Lundegardh, 1952, 1953).
If the cytochrome system is part of a membrane structure (Lundegardh,
1950), with the oxidase facing the medium and cytochrome b facing the
place of accumulation, conditions will be fulfilled for an active absorption
of salts from the medium into the cell.
VIII. THE QUOTIENT Q an./O2
Robertson (1941) called attention to the fact that four electrons are re-
quired for the complete reduction of one molecule O2 according to the
formula
2 = 2H20. (6)
If one electron is exchanged for one monovalent anion the quotient
an./O2 or (mol. absorbed anions/mol. consumed oxygen) will at best attain
the value 4. Robertson & Wilkins (1948) found in experiments with slices
of carrot that in strong salt solutions Q an./O2 rose to approximately 4 but
never higher.
The fact that in roots Q an./O2 seldom attains higher values than 1-2
(Lundegardh, 1949) nad now *° ^e investigated. I had previously observed
(Lundegardh, 1937) that roots held in distilled water show a cyanide-
sensitive fraction of the aerobic respiration amounting to some 50-75 %
of the anion respiration in salt solutions. This * distilled- water respiration'
(d.w.-respiration) was now more closely analysed (Lundegardh, 1949^).
It was shown that it decreases with increasing periods of washing of the
roots in distilled water (see Table i). A considerably more rapid decrease
near to zero was rapidly obtained in 0-0005 mol. HC1, whereas a similar
ANION RESPIRATION
275
period in 0-005 mol. KHCO3 raised the d.w.-respiration above the original
value. The * idling ' of the anion respiration was explained as follows :
What is measured in a tissue is not only the metabolism of the surface
layer in direct contact with the medium but primarily the bulk of the
tissue, the single cells of which communicate only with surrounding cells.
Table i . Decrease of the cyanide-sensitive respiration by
washing in distilled water or dilute acids
Medium
Distilled water during
0'000<> M-
HC1
0-005 M-
KHCO8
Time
i day
2 days
4 days
5 days
i day
i day
Cyanide-sensitive
respiration
(relative)
100
56
36
27
ii
131
The large quantities of salts accumulated in the single cells of the cortex
have traversed a number of cell walls on their way from the medium. The
solution in the cellulose walls and in that part of the protoplasm which is
responsible for the exchange capacity is the medium from which the
cytochrome system of the internal cells pumps its anions. Assuming a
steady state Accumulated anions
Anions xcCt ' (7'
in which cCt is the effective capacity of accumulation of the cytochrome
system, and K the ' coefficient of accumulation', we arrive at the conclusion
that there is an intracellular solution of salts in a steady-state balance with
the accumulated salts in the sap space of the single cells. This intracellular
salt solution is of course a very small quantity, because even the total
volume of protoplasm is very much lower than the total volume of vacuoles,
but it will be able to furnish a rapid stream of water, as in transpiration,
with certain amounts of salts (see Lundegardh, 1945, 1950). Its main task,
however, is to furnish the internal cells with material for salt accumulation.
Equation (7) teaches us that this movable non-accumulated fraction of
anions exists also in roots held in distilled water and that cCt must be at
work unless the level of accumulation sinks. Salts may under these cir-
cumstances be transported from regions of lower cCt value to regions of
higher power of accumulation. It was shown experimentally that internal
salt transport is accelerated by the anion respiration (see above on sap
movement). The slowly decreasing d.w.-respiration reflects the sinking of
the salt level in the tissue (Table i).
In addition to inorganic salts, anions of organic acids participate in the
d.w.-respiration. This is shown by the rapid decline of organic acids at low
18-2
276 ANION RESPIRATION
pH (HC1) and the corresponding rise in alkaline solutions (KHCO3). An
analysis of the content of malic acid in the tissue (unpublished results)
corroborates this conclusion. The d.w.-respiration may consequently be
interpreted as an anion respiration by means of * native anions ', representing
internally transported salts plus organic acid, and to include also true
'idling', viz. the re-accumulation of anions passively leaking out from the
cells (cf. Lundegardh, 1937). After transference of the roots to a salt
solution anions absorbed from the medium are added to the native anions
thereby raising the total anion respiration to the top level which is attained
if the supply of anions fully covers what is needed for a cytochrome system
operating at maximal speed. It was calculated that absorbed anions parti-
cipate with c. 25-30 % , circulating inorganic anions with about the same, and
anions of organic acids with about 40-50 % of the total capacity of the accu-
mulation mechanism in roots. A participation of organic acids (succinate)
in the anion respiration is also assumed by Turner & Hanly (1949).
This analysis of the d.w.-respiration opens a possibility of explaining
the differences of Q an./O2 if different anions are compared. From
equinormal solutions nitrate, chloride and sulphate were absorbed with
sinking g-values (Lundegardh & Burstrom, 1933). Also the cations have
an influence, easily movable cations as a rule raising, slowly movable cations
lowering the Q-values (Lundegardh, 1937). All these observations may be
explained from a competition between absorbed anions and native anions.
At low availability of absorbed anions, caused by low concentration or low
ionic mobility, the native anions share a larger part of the total assembly
of transported anions.
From the viewpoint of quantitative relations it is now important to
know that Q an./O2 maintains its value if the capacity of the cytochrome
system is gradually slowed down by rising concentrations of cyanide, as
shown in experimental series with nitrate and chloride (Lundegardh, 1949 b).
IX. THE SALT ACCUMULATION AS A STEADY STATE
In addition to equation (7) the salt situation of the root cells may be figured
as a steady state between an uphill process, the anion respiration, and a
number of downhill processes collected under the name of passive leakage
or exudation (Lundegardh, 1948). The passive exudation includes diffusion,
ion exchange, reversed adsorption, destruction of carriers, etc.
The steady state of single cells
Intracellular
salt solution
_ Cytochrome system
CQ.M+A-
^ ^^-m
Passive exudation
ANION RESPIRATION
277
The relation q/£0 or the degree of accumulation is determined by the
dimensions of the cytochrome system in the single cells and may ac-
cordingly be different in different layers of the root (see Fig. 6). Owing
to its high osmotic pressure the epidermis is probably provided with a very
active cytochrome system. The endodermis may possibly also develop a
high level of accumulation, as a sluice of salts to the vascular epithelium.
Very little is known, however, of these physiological-anatomical differentia-
tions. The bulk of salts is certainly stored in the cortex. See the following
scheme :
The steady state of the root tissue
Epidermis
Cortex, stele
Exuded sap (9)
X. SPECTROSCOPIC IDENTIFICATION OF THE
CYTOCHROMES
Before 1950 the identification of the anion respiration mechanism with the
cytochrome-cytochrome oxidase system was hypothetically built on the
extremely high sensitivity of anion respiration and absorption to cyanide.
Similar and more specific effects were found with other inhibitors of active
iron, e.g. aa-dipyridyl (Stenlid, 1950). A good criterion of the presence
of cytochrome oxidase is the inversion of the CO inhibition by light.
Independent investigations with wheat roots (Sutter, 1950) and slices of
carrot (Robertson & Wilkins, 1948) now showed a synchronous inhibition
of the anion respiration and the absorption of chloride anions under the
influence of a mixture of 95 % CO and 5 % O2 and a similar synchronous
recovery of both processes after illumination with strong white light.
Final evidence of the causal relation between the cytochrome system and
the anion respiration mechanism was yielded in my recent studies of the
absorption spectrum of the living roots (Lundegardh, 19510, ft, 1952,
1953 a-c). A special photoelectric spectrophotometer was constructed and
built for this purpose, enabling the automatic recording of the absorption
spectrum of a thick bundle (mostly 15 mm.) of roots under varying con-
ditions (solutions of different oxygen pressure and salt content, of different
inhibitors, at different temperature, etc.).
278 ANION RESPIRATION
The cytochromes are fairly dominant in the absorption spectrum of the
wheat roots. Of other coloured substances only carotenes show a prominent
band at c. 482 m/i, whereas peroxidases (band at 404 m/*) and flavoproteins
(band at c. 455 m/^) interfere only little in vigorously growing roots. It is
important, however, to select such roots, because at certain periods of the
year, e.g. October to February, the seeds frequently suffer from a reduced
activity of the cytochrome system, combined with a reduced salt absorption.
Under these circumstances other coloured substances, e.g. flavoproteins,
are more prominent and the roots also contain quantities of non-specified
haemin substances, possibly serving as precursors of cytochromes.
For identification of the cytochromes the oxidation-reduction spectrum,
viz. the extinction of reduced cytochrome minus the extinction of oxidized
cytochrome (see Lundegardh, 1951, 1952), proved very useful. The
difference spectra of the roots were compared with pure cytochrome c and
with the spectra of cytochrome oxidase (here also called cytochrome a) and
cytochrome b published by biochemists. Several hundred spectrograms
were recorded from vigorously growing wheat roots (only the lower c. 60 mm.
were used; the roots were divided in sections of 20 mm. length; some 300
such root pieces were tightly packed in a 15 mm. quartz tube provided
with inlet and outlet for the solutions).
The wheat roots have a complete cytochrome system. The prosthetic
groups are identical with or very similar to those of the cytochrome oxidase
and the cytochromes c and b known from animal preparations, yeast, etc.
The presence of cytochrome oxidase was also demonstrated from the
oxidation of added reduced cytochrome c by living or homogenized roots.
The kinetics of the cytochrome system of wheat roots is very similar to that
of the heart muscle, both in respect of the turn-over number of the cyto-
chrome oxidase, the state of oxidation of the operating system and its
linkage to succinic dehydrogenase (dk).
Because the molar extinction coefficients of the different cytochromes
calculated per mol. Fe are fairly alike, the approximate percentage of the
single enzymes may be calculated from the absorption spectrum. Vigorously
growing wheat roots contain on an average o*84/£mol. cytochrome oxidase
(a value corroborated by determinations of the oxidation of added cyto-
chrome c), i'56/emol. cytochrome c, and 2-46 /imol. cytochrome b per kg.
fresh weight. This corresponds to 21*6 mg. xkg."1 fresh weight, or
237 mg, x kg."1 dry weight in respect of cytochrome c, values corresponding
to quantities found in animal tissue. Cytochrome b is, however, more
dominant than in most animal tissues.
The cytochromes are approximately uniformly distributed in the lower
60-100 mm. of the roots, a fact coinciding with the earlier established
ANION RESPIRATION 279
uniform distribution of the power of salt accumulation (Lundegardh,
19496). A cytochrome system is also present in the green leaves and the
coleoptiles of wheat. Roots of rye and maize behave similarly.
XI. CO-VARIATION OF THE ABSORPTION OF ANIONS AND
THE STATE OF OXIDATION OF THE CYTOCHROMES
If the presence of movable anions is facilitating the electron transference
through the cytochrome system to the oxygen an increasing anion respira-
tion will be accompanied by an increase of the quotient oxidized cytochrome/
reduced cytochrome. A co-variation of the active absorption of anions and
the value of this quotient must strongly support the hypothesis of a direct
causal linkage between electron transference and anion transport.
A close study of the time course of the reoxidation of the completely
reduced cytochrome system of the roots (Lundegardh, 19536) confirmed
the conclusion that the electrons are transferred in the following sequence :
. ^ 1 ri 2,345^ , \
succmate-^a/z— b-^c—^ a—* O2. (10)
The dehydrogenase dh is believed to have a band at 570-575 m/4 when
reduced (Lundegardh, 1952, 1953). At the start of reoxidation the cyto-
chrome oxidase is immediately oxidized by the oxygen, cytochrome c is
then oxidized by the oxidase, and so on. A steady state is finally attained,
in which the degree of reduction of each member corresponds to the
prevailing conditions of the whole system and the potential states of the
single members. Under conditions of optimal anion respiration (0-05 mol.
aerated salt solution) and at 18-20° C. the cytochromes are predominantly
oxidized (60-84 %, cytochrome b leading). This means that the first link
in the chain, viz. the splitting up of one hydrogen atom uncoupled from
succinate in the reaction
Succinate^fumarate + 2H+ -f 2£, ( 1 1 )
acts as a brake, resulting in a comparatively low state of reduction of cyto-
chrome b and the subsequent links of the chain.
The influence of neutral salts on the oxidation-reduction balance of the
cytochrome system may be studied by recording the changes in the absorp-
tion spectrum occurring at a change of the medium from distilled water to a
salt solution, provided the roots were previously desalted. A few seconds
after the contact with the salt the reduction peaks (in the oxidation-
reduction spectrum) of the cytochromes are visibly lowered (Fig. 4). The
observed time-course of the increasing oxidation coincides with the time-
course of reoxidation from anaerobic to aerobic in distilled water (Fig. 5),
and it also coincides with the time-course of absorption of anions, as
280 ANION RESPIRATION
observed in parallel experiments with the same material and at the same
temperature.
In a series of experiments with the addition of 0-05 mol. KCI the oxida-
/ oxidized cytochrome \ .
ion I = quotient . — -^-v — xiool rose from 48% in aerated
tion
+ 0-04
.6 CO-6KCI
a-CO
-0<H
Fig. 4. Oxidation-reduction spectrum of the cytochrome system of wheat roots. Black
curve: difference spectrum between roots in 0*05 mol. KC1 and roots in aqua, showing the
decrease of the reduced bands of the cytochromes a (cytochrome oxidase) and b. Dotted
curve : reduction spectrum in carbon monoxide.
50
100
O 2 -0-05 mol. KCI I -aqua dest.
X I -aerobic KCI I -anaerobic KCI
J beco
/ |
ids ~ "*
I t 1 1
60
120
180
240
300
Fig. 5. Time course of oxidation, measured as decrease of the a-band of cytochrome b,
at a shift from anaerobic to aerobic (dotted curve), and at a shift from aqua dest. to
0-05 mol. KCI. The two processes take an approximately identical course.
distilled water to 69 % in the areated salt solution for cytochrome oxidase,
and from 6 1 to 84% for cytochrome b. The higher oxidation of the latter
reflects its lower oxidation potential state; a comparatively small rise of the
oxidation level of a cytochrome of high potential will secure a comparatively
large oxidation of a cytochrome of lower potential. The response of the
ANION RESPIRATION 281
cytochrome system to salts and other conditions changing the state of
oxidation may be observed on the Soret bands in the violet (y-bands) or on
the bands in the green (a-bands), the former showing a 3-5 times higher
extinction in the oxidation-reduction spectrum than the latter. In the case
of cytochrome c, the a-band at 550 m/£ is to be preferred, because the
appearance of changes of the y-band at 418 m/£ (oxidation-reduction) is
frequently delayed, probably owing to molecular complex reactions
(Lundegardh, 1953 d). The values of the response of cytochrome c to salts
(Lundegardh, 1953 a, table 3) are for this reason too low. The a-band at
550 m/£ shows a very marked response to changing conditions of oxidation
(see Lundegardh, 1953 £, fig. 2).
Considering the fact that the response of the roots to salts comprises
only about one-quarter of the total anion respiration (see above on d.w.-
respiration) the observed changes of the oxidation-reduction state of the
cytochromes are fairly large. In one experiment with a 10 mm. thick bundle
of desalted roots (see Lundegardh, 1952, p. 491) the height of the Soret
band of the cytochrome oxidase was 0-040 ( = log /0//i). The addition of
0-02 mol. KC1 lowered the value to 0-021. The final addition of o-ooi mol.
HCN raised the value to 0-066. The effect of the salt thus amounted to
30 % of the amplitude of oxidation-reduction between the state of oxidation
in aerated distilled water and complete reduction in HCN. A simple cal-
culation teaches us that the cytochrome system would probably be com-
pletely reduced if the ' native anions' could also be removed.
The idling caused by the native anions (organic acids plus the circulating
fraction of previously stored salts) is reflected in the state of oxidation-
reduction of the cytochrome system if the state of approximately 100%
oxidation obtained after treatment with malonate or fluoride (and not the
situation in distilled water) is chosen as zero value (see Lundegardh, 1953 d).
The values of the oxidation in distilled water just mentioned (48-61%)
were calculated from the absolute zero value of reduction (100 % oxidation).
In the case mentioned the increase in oxidation of cytochrome oxidase
caused by salt amounted to 69 — 48 = 21%. The corresponding value of
cytochrome b was 84 — 61 = 23 % increase of the oxidation. These values are
in good agreement with the conclusion drawn from absorption experi-
ments, that the active salt absorption of wheat roots on an average corre-
sponds to 25 % of the total electron transference.
XII. THE ANIONS AS COENZYMES
A further analysis of the effect of salts on the cytochrome system reveals
two partial effects or phases :
(i) The exchange of electrons and anions, an electrochemical balancing
282 ANION RESPIRATION
process controlling the velocity of turn-over of the enzymes according to
the scheme :
CYOXt . anion + e. (12)
Phase i represents the fundamentalcoenzymatic effect of anions.
(2) The coenzymatic effect may be developed into a physiological
process, the active transport and accumulation of salts, if the structural
arrangement of the cytochromes turns over the coenzymatic attraction of anions
into a polar stream from the medium to the interior of the cell, or from one
point of a cell to another point, from which a rapid backflow of salt is
prevented by a structural barrier.
This analysis of the salt effect explains why cytochrome systems may,
in many cases, e.g. in animal cells, operate without any salt accumulation
to speak of. These cells apparently lack an organization suitable for salt
accumulation. It is, on the other hand, experimentally possible to separate
the phases i and 2 in objects showing both.
In homogenized root tissue the aerobic respiration slowly decreases,
probably owing to an uncoupling of the succinic dehydrogenase, and the
power of salt accumulation is lost. The cytochrome oxidase and probably
also the other cytochromes, however, maintain their activity several hours
after homogenization. In a series of experiments (Lundegardh, 1953 £)
reduced cytochrome c was added to homogenate and the activity of the
cytochrome oxidase determined from the disappearance of the band at
550 m/4. Controls were made with addition of cyanide. The activity of the
oxidase is largely dependent on the presence or absence of salt ions. If
desalted roots are homogenized in distilled water the activity of the
cytochrome oxidase was increased 60% after addition of o-i mol. KC1 or
KNO3. Similar results were yielded by roots from nutrient solutions if
they were homogenized together with anion-absorbing resin. Also in this
case the addition of a surplus of salts considerably accelerated the activity
of the cytochrome oxidase. It was calculated that the abridged chain
c-+ a-+ O2 is about as active as in intact roots.
XIII. THE DIFFUSION BARRIER AND CYTOCHROME b
Uncoupling of phase 2 may also be brought about by treatment of the roots
with dinitrophenol (DNP) or fluoride. Experiments by Stenlid (1950) and
Robertson & Wilkins (1948) have shown that DNP in certain concentrations
stops the salt absorption in spite of the fact that the cyanide sensitive
respiration continues.
Stenlid showed that DNP in some way severely disturbs the cytochrome
system, causing a qualitative change in the respiratory processes. Similar
ANION RESPIRATION 283
effects were observed under the influence of methylene blue. My own
experiments with DNP have corroborated the observations of Robertson
and Stenlid and show a rapid decline of the active chloride absorption in
concentrations of DNP higher than io~6 mol. (at pH c. 5 ; at lower con-
centrations the normal anion respiration was slightly stimulated), whereas
the cyanide-sensitive respiration was even stronger than in the controls up
to a concentration of c. 3 x io~5 mol. DNP. The simultaneously observed
absorption spectrum of the roots showed a quite normal behaviour of the
cytochromes a and cy whereas cytochrome b was obviously more or less put
out of action.
Fluoride is a more effective and at the same time more indulgent
inhibitor of cytochrome b. At a pH promoting the absorption of molecular
NaF (c. 3*3-3-4; see Lundegardh, 1949 a) cytochrome b remains completely
oxidized whereas a and c continue their enzymatic activity in connexion
with a second dehydrogenase system, the succinic dehydrogenase being
put out of action because it cannot transfer electrons to cytochrome b if
fluoride is present. The second dehydrogenase system is possibly linked to
cozymase and a flavoprotein. At inhibited activity of cytochrome b this
second dehydrogenase system conducts a part of the end-oxidation via
cytochrome c and cytochrome oxidase. With still active succinic dehydro-
genase the second system is probably more exclusively served by a second
oxidase system, possibly flavoprotein and/or a non-ferrous metalloprotein
(see Fig. 6).
These recent results (see Lundegardh, 1953 a) define the question as to
the organization of the cytochrome system as a body promoting a polar
stream of anions. The theory of anion respiration as it has been elaborated
in my later work (1945-53) postulates the cytochromes as carriers of the
anions and attributes the accumulation work to the liberation of the anions
at cytochrome b (i or 2 in scheme (10)), this place of release of the anions
from their carriers obviously being structurally arranged to prevent a rapid
back-flow of the anions to the point of absorption (at the cytochrome a or
5 in scheme (10)). This theory introduces a minimum of additional hypo-
theses and is a logical development of the co-enzymatic action of movable
anions.
It was stated that phase i , or the co-enzymatic action of movable anions
in the surroundings of the cytochrome system, is probably of universal
importance. Few living cells are, however, completely deprived of salts.
Moreover, movable anions of organic acids are produced in the stages of
aerobic respiration preceding the end-oxidation. The experiments with
desalted roots and roots held under conditions lowering the production of
organic acids (Lundegardh, 1949 a), like desalted slices of storage tissue
284 ANION RESPIRATION
(Robertson and his group), clearly demonstrate the existence and physio-
logical importance of the co-enzymatic function of anions.
If phase i means an exchange between anions and electrons the move-
ment of the former may be characterized as an * active transfer', but in the
absence of structural barriers this active transfer will be unable to accom-
plish a real salt accumulation, because the released anions will find their
way back to the starting point by means of normal diffusion. The inter-
mittent function of the cytochromes as anion carriers (in the moments of
Carbohydrate
/ •
fM + adsorption layer
Aqua
Salt
Fig. 6. Diagrammatic representation of the aerobic respiration in wheat roots and its
linkage to salt accumulation. The encircled figures indicate points sensitive to inhibitors:
(i) the coenzymatic effect of anions (A~)-y (2) malonate and fluoride, inhibiting succinic
dehydrogenase (deh.!); (3) inhibitors of cozymase; (4) urethane, inhibiting the oxidation
of cytochrome 6; (5) DNP inhibiting phosphorylation and the reduction of b; (6) fluoride,
inhibiting the production of organic acids; (7) cyanide, azide, CO, etc., inhibiting the
oxidation of cytochrome oxidase.
oxidation) may possibly result in an attraction sphere of anions around
the system, a circumstance possibly conveying certain biochemical
consequences.
Accumulating evidence supports the assumption that mitochondria are
the site of the complete cytochrome system of animal cells and possibly
of certain plant cells too. Most animal cells do not accumulate salts to any
considerable extent in spite of the dominance of the cytochrome system as
respiratory mechanism. The mitochondria do not seem to have any pro-
nounced function as salt accumulators. They may have a membrane (see
Farrant, Robertson & Wilkins, 1953), but are apparently not built for the
ANION RESPIRATION 285
development of a high osmotic pressure. But they may, of course, be
figured serving as carriers of an adsorbed layer of salt ions, thus trans-
porting salts from one point of the cell to another by means of protoplasmic
streaming or a more independent mobility of the mitochondria themselves.
The assumption of mitochondria as carriers of salts between the surface
of the plant cell to the sap space would possibly meet difficulties in ex-
plaining the observed quantitative relations between the electron activity
and the transported anions. At the present state of knowledge it seems
Electron from DH.Hi
\
Anion
accumulation
\
Anion Electron
absorption to oxygen
Fig. 7. Scheme of the coenzymatic action of anions and their transport from the point of
absorption to the point of accumulation. The large circles symbolize the apoenzymes
(c is actually smaller than a and 6), the black semicircles the coenzymes.
more appropriate to assume a structural linkage of the cytochrome system
to the protoplasmic membrane in cells endowed with the power of an
effective salt accumulation. This assumption is, moreover, supported by
the real presence of cytochrome oxidase in the surface of the roots (Lunde-
gardh, 1952, 1953). The kinetics of the cytochrome system (Lundegardh,
1953^, c) leaves little room for ideas about a local separation of the cyto-
chromes, e.g. the oxidase placed in the surface and cytochrome b at the
vacuole membrane. The cytochromes are very probably joined to a struc-
tural body in which, however, the single enzymes have free thermal
mobility (cf. Fig. 7). Under these circumstances the loci of absorption and
accumulation of salts lie within a distance of only c. 150-200 A. (see below),
286
ANION RESPIRATION
or on both sides of a membrane. The original place of salt accumulation
will then be the protoplasm. As previously mentioned, a free-exchange
equilibrium probably exists between the protoplasm and the cell sap and
there is little experimental support for assuming an active secretion of salts
through the vacuole membrane.
A simplified scheme of the transference of electrons and anions is shown
in Fig. 7 and in the following sequence of reactions:
Absorption i. Cytochrome oxidase (Fe?+) + anion
chrome oxidase (Fe2+)+.anion-h(iO2)".
2. Cytochrome oxidase (Fe2+ )+ . anion 4- cytochrome
c (Fe2+) ^cytochrome oxidase (Fe2+) -f cyto-
chrome c (Fe2+)+. anion.
3 . Cytochrome c (Fe2+)+ . anion + cytochrome b (Fe2+)
^cytochrome c (Fe2+) + cytochrome b (Fe2+)+.
anion.
4. Cytochrome b (Fe2+) H. anion + 1 DH.H2^ cyto-
Accumulation chrome b (Fe2+) -f £ DH ~h H+ -I- anion.
The molecular construction of the haemin groups is still incompletely
known (see Lemberg & Legge, 1949). It may be sufficient to symbolize the
loss or gain of one electron as a change of the net charge of the molecule.
The cleavage of H into H+ -f e marks the end-stage of accumulation. The
postulated diffusion barrier must be situated somewhere between stage
(i)~(a) and the rest of the electron ladder because it was experimentally
shown that the active accumulation stops if cytochrome b only is inactivated
(see above). The barrier may simply be regarded as retarding diffusion of
anions through the membrane structure. The activating effect exerted by
the cytochrome system will then move the anions preferably in a centripetal
direction, the stages (5) and (i}-(z) of the scheme (10) respectively
providing the energy for the entrance and exit of the anions.
According to reaction (4) the accumulated anions are combined with an
equivalent quantity of hydrogen ions, A~ and H+ together forming a strong
acid, but this acid immediately reacts with the cation carriers M+.7?~,
as a result of which preferably neutral salts will be accumulated. As an
equivalent quantity of protons are consumed in the reaction
(i3a)
(5)
the accumulation of a neutral salt does not change the cH balance.
ANION RESPIRATION 287
XIV. ACTIVE TRANSPORT OF CATIONS
If the anion respiration is considered as an electrochemical phenomenon,
implying the transference of electricity from one point to another, the whole
process may also be pictured as an electrophoresis between the positive
pole of the system, represented by cytochrome i, and the negative pole,
represented by cytochrome oxidase. The streaming of anions may from this
viewpoint be interpreted as a regular electrophoresis between the poles of
an electric battery (Figs. 6, 8). The electron transference between the single
cytochromes corresponds to the internal transference of electrons pro-
ceeding in a battery. This scheme opens certain new aspects as to an active
transport of metallic cations.
Electrophoresis comprises a streaming of an equivalent quantity of
cations in a direction opposite to that of the anions, in our case from the
region of the dehydrogenase system to the surface boundary. This reversed
stream would imply a continuous loss of metallic cations if a corresponding
quantity of protons were not produced at the positive pole and consumed
at the negative pole (reaction (5)). The net result will then be a one-sided
transport of anions (Fig. 8 A). If, however, the protons produced at
the positive pole are partly consumed in an excretion of the free acid
HA — a process implying an acidification at the point of accumulation —
metallic cations will be caught by the centrifugal stream of positive
electricity (Fig. 8 C), and we have a respiratory mechanism excreting
cations.
An excess of metallic cations (K, Ca, Na, etc.) may exist after periods of
abundant salt absorption concomitant with internal consumption of anions,
especially nitrate of which c. 50% is normally proteinized in the wheat
roots (Burstrom, 1939). An excess of metallic cations may also result from
an abundant production and storage in the cell sap of organic acids followed
by a period of their metabolic consumption. As shown by Ulrich (1942)
and Burstrom (1943) absorbed cations and produced organic acids —
primarily malic acid — are important factors in maintaining a constant
internal pH. During metabolic consumption of organic acids the super-
numerary cations may be returned to the medium partly by non-metabolic
exchange, partly by means of active excretion in the anion respiration
mechanism (Fig. 8 B). Theoretically a situation may be figured in which no
anions are actively absorbed from the medium, only organic acids meta-
bolically produced. Given a previous abundant storage of metallic cations
in the cell the anion respiration will at first sight behave exclusively as a
mechanism for the excretion of cations. Anions acting as co-enzymes are
in this case OH~, HCOj, and organic acids. The participation of the
288
ANION RESPIRATION
cytochrome system in an active transport of cations has been very little
studied, but the problem certainly deserves more attention.
The electrochemical properties of the complete cytochrome system are
well fitted for a separation of anions and cations of a salt on the two sides
of a membrane barrier. Examples of this property are the excretion of
chloride from the gastric mucosae, and the strong acidification of the cell
sap in certain plants.
DH
D
H
1
J
I
/
-1
^
t
~*^
acc
H'
«
M A
ace.
H
— +1
HAc ace.
/
H
-* h
A\
71
e
r
/
6
I
r i
+
A
r
M
H
h
H
h
t
E
i/>
.= Ac
c
vt
c
o
o
c
o
ll
O
Q.
<J
rt C
E
0
E
c
0
| / i
|
E
Q.
s
u
JC *-»
°fr
O
C £
O u
i_
1
(U
CO
E
5~ / <u
1 *-
1 u
0
s
O
C
0
x
X.
rt
ft_ X
t*
i
/ ^
*"
Q.
c
0
U
C
0
E
8 S
-o o
U
•_>
Z
/ c
/ -
u
C
0
E
u
O
•o
<3
^
< i
/ rt
^
^
U
—
u
r U
t
e
iH2C
)<-
i02
A
- /*
s<
^2
l<
^2
/
\~ A
1f
.
J
1
j
i
M.
OH
|H
20
<<—
>
Oh
'
\.\-
J
[
NOH JH2O
Fig. 8. Diagram illustrating the electrophoretic activity of the enzyme body of the
cytochrome system and its activity as a mechanism for absorbing anions and excreting
cations. See the text.
XV. THE LOCALIZATION OF THE CYTOCHROME
SYSTEM IN THE CELL
The probable localization of the cytochrome system of wheat roots at
the cell surface is supported by the observation of an oxidation of added
reduced cytochrome c by living roots (Lundegardh, 1953). Wheat roots
in good health contain 2'$6/imol. cytochrome b x kg."1 (fresh weight), or
2-5 x io~3/4mol. xml."1. At a diameter of 0-5 mm., i ml. tightly packed
roots have a total surface of 63 cm.2 or 63 x io16 A.2. The total surface of all
cortex cells is approximately 8 x io19 A.2. At a number of 6 x io23 molecules
in i mol., i ml. root tissue thus holds 15 x io14 mol. of cytochrome b, viz.
ANION RESPIRATION 289
only one molecule on a surface of c. 5 x io4 A.2, the mean distance between
two molecules amounting to c. 220 A. The molecular weight is known only
for cytochrome c (13,000) and is calculated to be 75,000-80,000 for cy to-
chrome oxidase (Warburg, 1946). Assuming the molecular weight of
cytochrome b to equal that of the oxidase its diameter, at close packing of
the atoms, would amount to 35-40 A. At a mean distance of 220 A. cyto-
chrome b would then be conveniently placed in a single surface layer (of
the dimensions of membranes of mitochondria and other structures, see
Sjostrand, 1953). It was earlier calculated (Lundegardh, 1940) that the
mean distance of the cation carriers in the root surface is 160-170 A., a
value of similar magnitude to the above. Owing to the lower concentration,
a surface layer of cytochrome oxidase would, however, show a more
spacious distribution, resulting in a surface concentration amounting to
only about one- third of the concentration of the cation carriers. Because the
molecules of the cytochrome oxidase according to the theory are serving
as anion carriers (R+)y the calculated figures are well in accord with the
actually shown predominance of the R~ groups (see above). From what is
known about the kinetics of the cytochrome system (Lundegardh, 19536)
the three cytochromes are probably comparatively tightly packed in groups
of i a, 2C, and 3-46. Even the distribution of such groups in one single
layer in the surface of the cell would leave ample space for other con-
stituents of a complex membrane of mosaic pattern, possibly also including
other enzyme systems.
XVI. DISCUSSION
Considering the outstanding importance of an active transport of salts and
other solutes into cells and between cells of a tissue or organism, the
number of investigators devoted to these problems is still very limited.
And the single investigators mostly reveal a keen ambition to present new
theoretical schemes. It is a remarkable fact, to which I have previously
alluded (Lundegardh, 1940, 1949^, p. 324), that workers in the field of
' diffusion permeability ' ignore the results on salt accumulation in spite of
the fact that the existence of an ionic activity of the protoplasmic membrane
certainly interferes with the properties of passive permeability on inter-
action between sugar absorption and ions (see Lundegardh, 1940; Lunde-
gardh & Burstrom, 1944 and other papers). It is, furthermore, known that
passive diffusion is not the simple line of communication between a cell
and its surroundings once believed, because a number of common non-
electrolytes, e.g. sugar, asparagine, etc., are actively transported and still
obscure manifestations of 'active transport are involved in transport
mechanisms in sieve tubes or along cell surfaces in parenchymatic tissue.
E B S VIII 19
290 ANION RESPIRATION
Investigations on anion respiration and related problems inevitably
include biochemical work. The discovery of the co-enzymatic effect of
anions on the cytochrome system has, however, evoked no rejoinder from
the side of the biochemists. Chance (1952, 1953) overlooks this effect in his
recent attempts to understand the reaction kinetics of the cytochrome
system. The research work on the cytochrome system of wheat roots,
presented here, illustrates the usefulness of suitably chosen and treated
living material as objects for biochemical work. It has been shown that the
living cells are surprisingly permeable to all biochemical inhibitors, if
attention is paid to the conditions of dissociation (Lundegardh, 19490;
Stenlid, 1950). A co-enzymatic effect of anions was recently observed in
the photosynthesis of isolated chloroplasts (Gorham & Glendenning, 1952).
The attempts of these authors to explain the results from obscure effects of
ions on * colloidal properties' are, however, futile. Cytochrome is present
in the chloroplasts (Hill, 1951) and a co-enzymatic effect of anions is thus
feasible, but anions may of course also operate as balancing agents in other
processes of electron transference than redox reactions of cytochromes.
Certain authors refuse to discuss the theory of anion respiration because
they think it is not * universally accepted' (Preston, 1948, p. 130), and
recently Overstreet & Jacobson (1952) failed to recognize what the theory
really contributes, owing to an incomplete knowledge of various papers in
which it was experimentally elaborated. The critical attitude of Steward
(1935 and later papers) is unique, because he simply denies the existence of
any evidence of a quantitative relation between an anion or salt respiration
and the absorption of ions. In fact, Robertson and his group in a series of
papers clearly demonstrated the existence of an anion (or salt) respiration
in the same material that Steward used.
Besides this negative attitude, Steward also presents his own theory on
salt accumulation, namely, an assumed linkage to growth and protein
metabolism (Steward & Preston, 1941). I have previously shown (1945,
p. 31) that the extremely active cytochrome system is obviously linked to
various synthetic processes, the synthesis of proteins among others. The
active transport of salts is by no means the main purpose of the cytochrome
system, but merely an accessory process consuming a very small fraction of
the converted energy (see p. 281). We know at present that the main part of
the energy of respiration is stored in high-energy phosphate bonds of which
the protoplasm disposes for special purposes. The active accumulation of
salts thus runs simultaneously with the uphill side of the complicated
steady-state situation characteristic of living cells.
In the case of wheat roots it is known that active growth is restricted to a
tip zone of a few millimetres length. The nitrate assimilation, supporting
ANION RESPIRATION 29!
the synthesis of proteins in the root (Burstrom, 1943; Lundegardh, 1945,
195 1 a), is extended over the lower 20-30 mm. of the root. In this zone
only, c. 50 % of the absorbed nitrate is consumed and utilized for synthetic
work. In the zones above 30 mm. up to 100 mm. from the tip, 100% of
the absorbed nitrate is exuded in the sap stream (Lundegardh, 1951 a).
The cytochrome system and the power of active salt absorption is, however,
uniformly distributed over the whole length of c. 100 mm. (Lundegardh,
19496, 1952, 1953^)- Also experiments on the effect of a number of in-
hibitors on growth and salt absorption illustrate the lack of parallelism
120 -
110 -
100
90
80
70
60
50
40
30
20
10
0
Bleeding (30 mm.)
* • * ** s
„.-'** Cl absorption (1 hr.) ^ 9
V
10 9 10-8 10 7 10 6 10-s
Indol acetic acid
10"
10 3mol.
Fig. 9. Diagrammatic representation of experimental series on the effect of indole acetic
acid on growth, bleeding, respiration, and Cl absorption.
between these processes. The same conclusion is drawn from experiments
with indole acetate (IA), the results of which are plotted in Fig. 9. Owing
to a complex effect of the hormone and the H-ions (see Lundegardh, 1949^)
the rapid growth reaction (30 min.) varies along a reversed optimum curve,
the increasing retardation dominating up to a concentration of c. io~5 mol.
I A and a pH (caused by the partial dissociation) of c. 4-9, followed by an
increasing stimulation in the concentrations io~4 to io~3 mol. I A and pH
values 4-4 to 3-9. The respiration and the absorption of added chloride
(unpublished experiments) give curves of a different pattern, the chloride
absorption remaining intact up to io~6 mol. I A (at a growth inhibition down
to 30% ) and after a rapidly passing stimulation at io~5 IA sinking to 45 %
ANION RESPIRATION
in io~4 and 5% in io~3 mol. IA. The total respiration decreases slowly,
followed by the curve of bleeding (gross volume of the exuded sap).
Separate determinations of the respiration show disturbances both in Q
an./O2 and the ground respiration.
There is consequently ample experimental evidence of a full activity of
the cytochrome system in cells of inhibited growth and of minimal protein
synthesis. The cells 100 mm. from the tip of rapidly growing grass roots
are of course comparatively * young', and it is possible or even probable that
in older parts both the cytochrome system and the power of active salt
absorption disappear.
Humphries (1951, 1952), conducted experiments with roots of barley
and pea from which he concluded that there is 'no evidence of a salt
respiration'. As Humphries made no experiments with cyanide he
apparently knows very little of anion respiration in his objects. Also his
discussions reveal lack of knowledge of the vast experimental material on
which the theory of anion respiration is founded. Humphries advances
speculations as to the sugar level of the cells as a promoter of salt accumula-
tion. The hypothesis that sugar 'may be the parent substance for the forma-
tion of a chemical compound capable of combining with ions ' has no real
meaning because sugar participates in a multitude of biochemical processes
and is 'the parent substance' of nearly everything in the protoplasm. In
my experiments the role of sugar as the fuel for the anion respiration was
quantitatively demonstrated. At low sugar level the anion respiration may
be retarded. It then accelerates after sugar has been supplied. Unpub-
lished experiments show that sugar supplied simultaneously with chloride
stimulates the absorption of anions, whereas roots observed after a pre-
ceding period of feeding with sugar reveal a slight retardation of the anion
absorption. It was previously observed that feeding with sugar raises the
level of acidity in the cell sap. As shown above, organic acids compete with
inorganic anions in the anion respiration. Sugar is absorbed into the roots
by means of an active process (Lundegardh & Burstrom, 1944), probably
respiratory phosphorylation. In this activated state the sugar is probably
more accessible for the anion respiration mechanism.
Humphries has observed that salts may be absorbed without any
appreciable change in the total respiration. This observation is neither new
nor is it surprising. The analysis of the d.w.-respiration teaches us that
roots can show a considerable anion respiration if sufficient 'native anions'
are present. If such roots are transferred to a salt solution the newly
imported anions slowly take over the co-enzymatic function of a corre-
sponding number of native anions. During this competition process the
intensity of the anion respiration is but little changed. But it is continuously
ANION RESPIRATION 293
working, as may be seen from the inhibition by cyanide or from a direct
observation of the cytochrome system. Only if the objects are washed in
aerated distilled water of a pH somewhat below 6-5 during a sufficiently
long period do the native anions disappear to a degree permitting accurate
measurements of the quantitative relations between absorbed anions and
the anion respiration. One cannot expect that a living tissue will expose
the results of one single physiological phenomenon unless a number of
interfering reactions are slowed down to a minimum. In long-term experi-
ments with intact plants the passive transport of salts due to respiration has
to be considered. As previously mentioned, the exchange capacity (or ' free
space') of the cells provides the plant with an instrument of passive trans-
location of salts, the extension of which, however, can be determined only
from careful studies of the simultaneous active anion respiration. Such an
analysis has been omitted in the studies of Humphries and in the more
recent work of Hylmo (1953).
The investigations and discussions of Robertson & Wilkins (1948) as to
the effect of DNP has awakened doubts on the validity of the theory of
anion respiration also among other writers. Overstreet & Jacobson (1952,
p. 202) venture that 'it is rather difficult to fit the DNP effect into the
Lundegardh hypothesis, as presently postulated, without some further
assumptions or extensive modification*. It was shown in the preceding
pages that no 'extensive modifications' are needed. The key position of
cytochrome b plus dh elucidates the problem.
The observations by Stenlid, Robertson & Wilkins and Lundegardh that
DNP also promotes an exuberant exudation of organic substances from the
tissues points to a severe disturbance of the structural qualities of the
protoplasmic membrane. It has been assumed that the succinic dehydro-
genase is involved in the high-energy phosphate (=~ph) metabolism
(Schlenk, 1951). A tentative scheme is the following :
Succinate + 2Fe3+ + ADP + PO4^Fumarate + 2Fe2+ + 2H+ + ATP. (14)
The balance will be moved to the right in the presence of predominantly
oxidized dehydrogenase, a situation realized in a respiring cell given a
sufficient supply of salts. At very high concentration of fumarate and
predominantly reduced cytochrome b the equilibrium will move to the left
side, a situation observed in living roots after addition of fumarate + HCN
(Lundegardh, 1953). According to the formula AF= -wFAE1, where
AF=free energy, F = Faraday or 23,000 cal., w = the number of electrons
and Z? = the potential gap (Kaplan, 1951, p. 64), one ~ph will be syn-
thesized at £" = 0-3 V. if two electrons are participating in the equilibrium.
These facts are concordant with the calculated relatively high oxidation-
294 ANION RESPIRATION
reduction potential state of cytochrome b (Lundegardh, 19530). It is a
common belief that ATP is an important guard of the structural integrity
of the protoplasm, and the key position of cytochrome b in the anion
respiration may also be considered from this viewpoint. Fluoride inhibits
the succinic dehydrogenase and enzymes participating in the production
of organic acids, but not phosphorylation. The similar effects of DNP and
fluoride on the anion respiration (Lundegardh, 1952, 1953) points to the
inhibition of the electron transference from dehydrogenase to cytochrome b
as the immediate cause of inhibited accumulation of anions.
As to the suggestion of Robertson & Wilkins, that the ~ ph metabolism
might be directly involved in accumulation work too, this possibility was
already discussed in connexion with other possible mechanisms of salt
accumulation. Because values of Q an./O2 above 4 were not observed in
plant material there is so far no need for any other mechanism than
the anion respiration, if a cytochrome system is at work. Speculations as to
variations in potential gradients, viz. the acid/base balance (Helder, 1952;
Vervelde, 1952), or in the dissociation of ion carriers (p. 264), must be
supported by quantitative experimental investigations before the matter can
be taken under consideration. Helder(i952, p. 421) ventures as an argument
against the carrier function of the cytochromes that the ratio of the absorbed
quantities of different anions ' is quite different from that of the composition
of the external solution*. This is a poor argument because (i) the cyto-
chromes, as other ion carriers, may of course show selective qualities, and
(2) there are also other anion carriers in the surface of the cell, with which
the cytochromes may exchange anions. If the latter act selectively the total
absorption will be selective.
A few words may finally be said about electrical currents in salt-absorbing
organs. Roots produce an electrical current continuously if a circuit is
closed between the tip and the base (Lundegardh, 1940). If the current
is accelerated by the addition of a 2 V. battery the anion respiration and the
salt absorption are also accelerated. If the extra current flows in the
opposite direction salt absorption and respiration are retarded. These
results are quite in accord with the assumption that the stream of electrons
through the cytochrome system is the source of electricity. But they are
not quite conclusive because electricity will be produced in connexion
with every active transport of ions. An operating cytochrome system
augments the electron density in the organ, hence lowers its ohmic
resistance. Unpublished measurements show an increase of the con-
ductivity at full activity of the cytochrome system and a decrease after the
addition of cyanide. But also the production of organic acids has to be
considered here. Measurements of bioelectric potentials and bioelectric
ANION RESPIRATION 295
production of electricity may reveal important physico-chemical properties
of the cells, as shown in the discussion of root potentials. Reliable con-
clusions as to the mechanism of salt accumulation are, however, attained
only from a biochemical and biophysical analysis of the phenomenon
in vitro and in vivo.
REFERENCES
ARISZ, W. H. (1952). Ann. Rev. Plant Physiol. 3, 109.
BURSTROM, H. (1937). Medd. CentAnst. Forsoksv. Jordb., Stockh., no. 475.
BURSTROM, H. (1939). Planta, 30, 129.
BURSTROM, H. (1943). LantbrHogsk. Ann. n, i.
BURSTROM, H. (1945). LantbrHogsk. Ann. 13, i.
BURSTROM, H. (1949). Mineral Nutrition of Plants, p. 251. Madison Symposium.
CHANCE, BRITTON (1952). Nature, Lond., 169, 215; Ann. Rev. Biochem. 21, 687.
EPSTEIN, E. (1953). Nature, Lond., 171, 83.
FARRANT, J. L., ROBERTSON, R. N. & WILKINS, M. J. (1953). Nature, Lond., 171,
401.
GORHAM, P. R. & GLENDENNING, K. A. (1952). Arch. Biochem. Biophys. 37, 199.
HOPE, A. B. & ROBERTSON, R. N. (1953). Aust.J. Sci. 15, 197.
HELDER, R. J. (1952). Ada hot. neerl. i, 361.
HILL, R. (1951). New Phytol. 50, 98.
HOAGLAND, D. R. & BROYER, T. C. (1942). J- Gen. Physiol. 25, 865.
HUMPHRIES, E. C. (1952). J. Exp. Bot. 3, 291.
HYLMO, B. (1953). Physiol. Plant. 6, 333.
KAPLAN, N. O. (1951). The Enzymes (Sumner, Myrback), 2, part i, p. 55.
LEMBERG, R. & LEGGE, J. W. (1949). Hematin Compounds and Bile Pigments.
Now York.
LUNDEGARDH, H. (1929-34). Die quantitative Spektralanalyse der Elemente, i and
2. Jena.
LUNDEGARDH, H. (1935). Naturwissenschaften, 23, 313.
LUNDEGARDH, H. (1937). Biochem. Z. 290, 104.
LUNDEGARDH, H. (1938). Biochem. Z. 298, 51.
LUNDEGARDH, H. (1939). Planta, 29, 419.
LUNDEGARDH, H. (1940). LantbrHogsk. Ann. 8, 233.
LUNDEGARDH, H. (1941). Protoplasma, 35, 548.
LUNDEGARDH, H. (1943). Ark. Bot. 31 A, no. 2.
LUNDEGARDH, H. (1945). Ark. Bot. 32 A, no. 12.
LUNDEGARDH, H. (1948). Disc. Faraday Soc. no. 3, p. 139.
LUNDEGARDH, H. (19490, b). LantbrHogsk. Ann. 16, 339, 372.
LUNDEGARDH, H. (1949^). Physiol. Plant. 2, 388.
LUNDEGARDH, H. (1949^). Ark. Bot. i, 289.
LUNDEGARDH, H. (1950). Physiol. Plant. 3, 103.
LUNDEGARDH, H. (19510, b). Ark. Kemi, 3, 69, 469.
LUNDEGARDH, H. (1952). Nature, Lond., 169, 1088.
LUNDEGARDH, H. (19530). Ark. Kemi, 5, 97.
LUNDEGARDH, H. (19536, c, d). Nature, Lond., 171, 477, 521 ; 172, 303.
LUNDEGARDH, H. & BURSTROM, H. (1933). Biochem. Z., 1933, 261, 235.
LUNDEGARDH, H. & BURSTROM, H. (1935). Biochem. Z. 1935, 277, 223.
LUNDEGARDH, H. & BURSTROM, H. (1944). LantbrHogsk. Ann. 12, 51.
LUNDEGARDH, H., BURSTROM, H. & RENNERFELT, E. (1932). Svensk hot. tidskr.
26, 271.
LUNDEGARDH, H. & STENLID, G. (1944). Ark. Bot. 31 A, no. 10.
296 ANION RESPIRATION
OVERSTREET, R. & JACOBSON, L. (1952). Ann. Rev. Plant Physiol. 3, 189.
PRESTON, R. D. (1948). Disc. Faraday Soc. no. 3, p. 130.
ROBERTSON, R. N. (1941). Aust.J. Exp. Biol. Med. Sd. 19, 265.
ROBERTSON, R. N. (1951). Ann. Rev. PL Physiol. 2, i.
ROBERTSON, R. N. & TURNER, J. S. (1945). Aust. J. Exp. Biol. Med. Sci. 23, 63.
ROBERTSON, R. N. & WILKINS, M. J. (1948). Aust.J. Sci. Res. ser. B, i, 17.
ROBERTSON, R. N., WILKINS, M. J. & WEEKS, D. C. (1951). Aust. J. Sci. Res.
ser. B, 4, 248.
SCHLENK, F. (1951). The Enzymes (Sumner, Myrback), 2, part i, p. 316.
SCOTT, G. T. (1944). J. Cell Comp. Physiol. 23, 47.
SJOSTRAND, F. S. (1953). Nature, Lond., 171, 30.
SLATER, E. C. (1950). Nature, Lond., 166, 982.
STENLID, G. (1948). Physiol. Plant, i, 185.
STENLID, G. (1947). LantbrHogsk. Ann. 14, 301.
STENLID, G. (1949). Physiol. Plant. 2, 350.
STENLID, G. (1950). Physiol. Plant. 3, 197.
STEWARD, F. C. (1935). Ann. Rev. Biochem. 4, 519.
STEWARD, F. C. & PRESTON, C. (1941). Plant Physiol. 16, 85.
SUTTER, E. (1950). Experientiay 6, 264.
TURNER, J. S. & HANLY, V. F. (1949). New Phytol. 48, 149.
ULRICH, A. (1942). Amer.J. Bot. 29, 220.
VERVELDE, G. J. (1952). Zoutophoping door Plantewortels. Wageningen.
WARBURG, O. (1946). Naturwissenschaften> 33, 92.
SOME ASPECTS OF ION TRANSPORT
THROUGH MEMBRANES
BY EDWARD J. CONWAY
Department of Biochemistry, University College, Dublin
I. INTRODUCTION
In the following article, work relating to active transport of sodium and
potassium ions in yeast and some new evidence relating to the localization
of sodium in muscle from the author's laboratory are mainly considered,
together with some recent work from other laboratories relevant to the
'redox-pump' theory of active transport (Conway, 1951, 1952, 1953). At
the outset some questions of a general kind arise, including the nature of the
immediate energy source in transport of ions to a higher electrochemical
potential. Under this latter heading no general review is attempted, but
the ' redox-pump ' theory mainly considered (which is dealt with much more
fully elsewhere; Conway, 1953), as well as the irreversible energy change in
the transport of free ions through membranes.
Active transport across a membrane connotes movement of the solute or
ion across the membrane dependent on the activity or energy change of
another system. Passive transport is equivalent either to free diffusion or
to * exchange diffusion J where the energy of the net movement comes from
the same system, as where urea diffuses from a higher to a lower concen-
tration across a membrane. Here the system is urea on both sides of the
membrane.
Active transport so defined may in turn be considered as equivalent to
functional transport, and the following may be considered to exemplify
different kinds of functional transport or transference.
(a) The solute is removed from free solution by a carrier, the resulting
complex then traversing the membrane and yielding the carried solute on
the other side by enzyme action. In this the carrier itself crosses the
membrane in quantitative or equivalent relation with the carried solute.
Lipoid solubility of the carrier complex may be assumed.
(b) This is a process similar to (a) but with cyclical restoration of the
carrier representing what are usually regarded as the most typical cases of
active transport. Here also lipoid solubility would appear advantageous or
necessary. Cyclical activity may be assumed associated with redox changes
either directly or indirectly, the latter through conversion of the electron
energy into that of phosphate bonds.
298 SOME ASPECTS OF ION TRANSPORT
(c) An ion is brought across the membrane at a certain rate, by a
potential difference or gradient produced by the carrier transport and
release of an ion of opposite charge. Here the ion itself is passive and the
process not usually regarded as one of active transport, and distinguished
therefrom ; yet in a steady state energy has to be continuously expended upon
the maintenance of the potential difference which pulls across the free ions.
Such transport, however, may be considered as secondary to * carrier
transport'. (In this context one may consider the question whether an iron
filing jumping up to a magnet is * actively' transported; and if it jumps
through a viscous fluid, the question is even more pertinent.)
Clearly, apart from verbalisms, the significant feature is the expenditure
of energy by another system in the process, and the energy requirement for
the passage of such * passive J ions is considered later.
(d) Functional transport may also be exemplified when a current of
fluid is set up through the pores of a membrane as by electro-endosmosis,
carrying with it various solutes, and if such are selected, as by molecular
diameter and pore size, a certain specificity can be obtained in the solute
carriage.
Such a flow of water might also be produced by the special secretion or
active transport of one type of solute, with subsequent osmotic flow of water,
or again could be produced by the increase of individual molecules or ions
through enzyme action on one side of a membrane.
(e) Another form of ion passage, which at least arises for consideration,
occurs when the chemical potential of a salt, such as potassium phosphate,
is lowered on one side of a membrane by metabolic activity incorporating
the phosphate ions in organic esters, and a consequent movement of both
potassium and phosphate ions to restore the balance (Boyle & Conway,
1941), assuming provisionally that phosphate ions pass as free ions through
the membrane. A characteristic of the movement here is that a new equi-
librium is reached after the withdrawal of the phosphate ions, and to main-
tain such equilibrium no additional metabolic activity directed thereto is
essential (apart from the slow extrusion of sodium ions).
The transport in such examples (a)-(e) is effected by special carriers, or
by the provision of an electrochemical gradient, or a flow of water or by
special metabolic activity. All of these processes depend on activity of some
kind and the expenditure of energy by another system, and it may be assumed
that all are of functional significance. One may then speak of active or
functional transport as providing the widest heading, divided in turn into
carrier transport and free transport.
Carrier transport may bring an ion to a higher or to a lower electro-
chemical potential (vide Rosenberg, 1948). If to a lower then the functional
THROUGH MEMBRANES 299
significance of the carrier transport would appear to be the facilitation of
membrane passage. Free transport occurs always to a lower electrochemical
potential.
In (c) and (e) above the passage across the membrane is necessarily to a
lower electrochemical potential, but not necessarily so in (d). The reader may
also be referred to an interesting discussion of this question from a somewhat
different point of view by Linderholm (1952).
II. THE ENERGY SOURCE IN TRANSPORT TO A HIGHER
ELECTROCHEMICAL POTENTIAL
When carrier transport of an ion occurs to a higher electrochemical potential,
then immediately on release of the ion there must be a simultaneous
transference of free energy.
In other words, on the breakdown of the complex, part at least of its
chemical energy must appear in the increased electrochemical potential of
the transferred ions. Such energy may be directly derived from a redox
system transferring electrons, or by the indirect use of such energy as by
phosphate bonds.
The use of phosphate bond energy
One way in which bond energy may be used in active transport has been
suggested by Danielli (1952), who pictures such energy utilized, as it is in
muscular contraction, for the contraction of protein chains on which ions
have been loosely combined. Another type of theory may be illustrated by
that of Nielsen & Rosenberg (195 1) for the secretion of hydrogen ions by the
gastric oxyntic cells. (This is referred to again at the end of this article.)
Davies & Krebs (1951) also discuss the operation of ATP in the control of
the interaction of the ferri-ferrocytochrome and flavine systems.
The 'redox-pump* theory for the active transport of inorganic cations. This
has been described elsewhere (Conway, 1951, 1952), and the theory is
treated in much greater detail in another publication (Conway, 1953).
The following is a brief account.
M«±M+e (i)
represents a metal respiratory enzyme in the reduced and oxidized condition.
j?m, the potential of the system, is independent of the pH.
The equation ^^ ^ ta + H ^ JQ + H+ + * (2)
represents a system, the potential of which is dependent on the hydrogen-ion
concentration. The potential of system (i) is written E"m, and of system (2),
Ect with both systems at a pH, say, of 6-0. It is assumed for convenience in
300 SOME ASPECTS OF ION TRANSPORT
writing that the reductant concentration is the same as the oxidant in both
systems. If they are joined by metal electrodes and liquid bridge, and dn
equivalent electrons allowed to pass from system M to system Ct then one
may write for the free-energy change
dn¥(Em - Ecta) + dnRTln (H)a = dn x a constant. (3)
If the hydrogen-ion concentration of the system is now raised until the
potential of system (2) is equal to that of system (i), then, as before,
dn(Em - EC(b) + dnRTln (H)b = dnxa constant, (4)
and, as the first term is zero,
dnRTln(H)b = a constant,
then on subtracting equations (3) and (4), then
F(Em-Ecla) = RTln(H)b/(H)a. (5)
This may be interpreted to mean that the whole of the externally available
free energy associated with the passage of one equivalent of electrons from
system (i) to (2) at a H level of (H)a can be converted into the energy required
to raise one equivalent of hydrogen ions from the level (H)a to (H)6. The
use of such a principle for secreting H ions in high concentration requires
a certain organization of enzymes with respect to the cell membrane. This
has been treated at length elsewhere as the ' redox theory ' for the secretion
of H ions (Conway, 1952; vide also Davies, 1951). The evidence for such
a theory both for the oxyntic cell of the gastric mucosa and for the yeast cell
is very strong.
The secretion of inorganic cations by such a redox-pump system, that
is, by direct use of the electron energy, may be treated in an analogous
manner.
Thus if one considers the system
(Red.- J9+) *± Ox + e + fi+, (6)
in which (Red.~5+) represents an adsorption complex of an inorganic
cation with the negatively charged reductant of a respiratory enzyme, M',
then, as above, one may deduce that the relation
= a constant. (7)
(Here, also, both redox systems may be assumed for convenience to have
their oxidant and reductant concentrations equal.) Thus at a low con-
centration of 'B* (Em — Em,^ may be assumed to be relatively high, but
if ' B 9 is progressively increased, then a concentration (B)b is reached when
(Em-Em'b) approaches zero.
THROUGH MEMBRANES 3<DI
Equation (7) may be interpreted to mean that all the electron energy
available from the transference of one equivalent of electrons from one
system to the other at a * B ' level of (B)a can be transformed into the osmotic
energy required to raise one equivalent of B from level (B)a to (B)b. An
analogous equation may be written for the active transport of anions,
concerning which in root hairs much valuable work has been done by
Lundegardh (1940, 1947, 1948, 1949).
Application of the principle summarized in equations (5) and (7). This has
been outlined in detail elsewhere (Conway, 1951, 1952, 1953). In the case
of the secretion of H ions, metabolic hydrogens are considered to be trans-
ferred by way of flavine enzymes to a metal catalyst in the membrane which
receives the H atoms, splitting off H+ ions and retaining the electrons,
which are then transferred to oxygen in the case of the oxyntic cell. When
the electrons combine with oxygen, H ions are taken up in this last stage
from within the cell, in equivalent relation to those produced outside.
With the transport of inorganic cations, in the first stage, the metal
catalyst receives H atoms, splits off H ions and retains electrons, the nega-
tively charged catalyst then combining with the inorganic cation in a
complex, which passes into the cell membrane. Here, with a final transport
of electrons to oxygen within the cell, there is an uptake of H ions, which is
equivalent in amount to the H ions liberated therein at the first stage, so that
the cell remains neutral.
The question of specificity of carriage
When K+ ions are carried into the yeast cell or Na+ ions carried out
considerable specificity of carriage is shown. It would appear then that the
complex of inorganic cation and reduced respiratory enzyme does not
result simply from long-range electrostatic forces.
The following suggestions may be made concerning this point. It may
be pointed out that considerable physical differences may occur between
certain salts of K and Na. One such salt of an acid may be practically
insoluble and the other by comparison very soluble. Also, in the case of
certain salts of polyphosphates (Von Wazer & Campanella, 1950), specificity
is shown with respect to the degree of ionization. Hodgkin (1951) cites the
experiments of Schwarzenbach, Kampitsch & Steiner (1945, 1946) showing
that certain organic compounds had a weak affinity for Na but not for K,
and evidence from Lamm & Malmgren's experiments (1940) that the
polymers of metaphosphoric acid have a special affinity for sodium. It may
also be considered that, in general, enzyme action begins with formation of
a compound of enzyme with substrate, and often the intimate nature of such
a compound is very problematical. In the present context we may speak of
3O2 SOME ASPECTS OF ION TRANSPORT
a Na enzyme or a K enzyme. In any case, if there are more than electrostatic
forces, it would seem that an ionic bond is operative, the question being
how such a bond is affected by the redox cycle. The following general possi-
bility may be considered. In the oxidized state of the molecule one may have
some such arrangement as
/M+— e /M—
-Xj —>-*<(
yX~ + B+ \XB)
where an attractive force is exercised between a charged atom which suffers
a valency change in the redox cycle and a negatively charged atom in
another group on the molecule. This force may be sufficient, taking the
group of molecules as a whole, to displace Na or K ions wholly or in part
from attaching at Xy in the manner of an undissociated salt. When the
valency change occurs in the M atoms and they gain electrons, the attractive
force with X disappears, and Na ions are taken on. In this case the inorganic
ion attachments are not directly to the M atoms. (Also it may be noted that
such an atom changing its valency in a redox cycle may be nitrogen.)
Kinetic limitations
It has been considered that the full conversion of the electron energy to
osmotic work occurs when the donating system and the acceptor systems
have the same potential. In such a theoretical case the transfer of electrons
would not in fact occur, and if there is a slight difference only it would occur
presumably at a relatively slow rate. In such a case, if the metabolic system
feeding the electrons or H atoms into the mechanism proceeds more or less
unchanged, the donating system will become more and more reduced and
the potential difference between it and the accepting system increased.
III. THE SIGNIFICANCE OF TRANSPORT POTENTIALS
Here a distinction may be drawn between assumed and true transport
potentials. The assumed transport potential (Ussing & Zerahn, 1951;
Teorell, 1952 ; Linderholm, 1952) has arisen in connexion with the develop-
ment of flux equations. Thus the ratio of the flux of an ion across a membrane
in direction 1-2 to that in the 2-1 direction may be written
-nt'K', (8)
where n£ ' and nf ' are the moles of a cation species transferred, c£' and cf '
the concentrations on each side, £ being the Planck symbol ( = eEFIRT).
If active transport is associated with the ion movements, then this ratio does
THROUGH MEMBRANES 303
not hold. To equalize the ratios, cf could be multiplied by a number a.
This equalizing number may then be presented as ^ or as eEtF/RT, in which
the potential Et is assumed to be, or called, a transport potential.
Objections to the use of this concept of a transport potential are that,
depending on the mobility of the free ions through the membrane, a whole
series of figures can be obtained for Et, even though the active transport
mechanism remains the same in kind and in energy output, and even if the
number of active carrier molecules remains the same. Also if Et were
regarded as a hypothetical potential which could move the free ions (carried
in active transport at the same rate), the energy required for this is quite
different from that involved in their actual transport.
These points become clearer when one considers energy requirements
in relation to fluxes and transport of ions.
However, there is one value of £, which may be regarded as truly signi-
ficant when, for instance, as in transport of Na ions through the frog skin,
the concentration of NaCl on both sides of the membrane is the same
(ci ' = cz' m equation (10)), and a counter-potential is applied across the
membrane sufficient to prevent the active transport of Na ions, then £ = ^
(Ussing & Zerahn, 1951). Here the applied potential gives the maximum
potential under the conditions against which the Na+ ions can be actively
transported.
Apart from this it is shown below that a true transport potential far free
ions is a significant figure. It may be defined as the extra electrical potential
required to bring a free ion across a membrane at a given rate. Such a
transport potential in turn arises from the carrier system ferrying bound ions
of opposite charge.
IV. MINIMAL ENERGY CHANGES INVOLVED IN
ACTIVE TRANSPORT
In the following treatment, it will be considered for simplicity that Na+
ions are being actively transported, that Cl~ ions are being transferred in
a secondary way by the resulting potential difference, and, further, that the
actively transported Na+ ions do not diffuse back again through the
membrane or, in other words, that the membrane is impermeable to free
Na+ ions.
The classical differential equation (Planck, 1890; Nernst, 1888, 1889)
relating the flux of an ion to the electrical and concentration gradients may
be written in the form
dn+' __ Ac+' IRTdc+'
~dt ~ ~ /7Vn \ c+' dx
304 SOME ASPECTS OF ION TRANSPORT
(Here n+' is the net number of moles of a cation species in unit time from
side i to 2, c+> is the concentration of a cation species on side i, and dc+ /dx
the gradient of concentration of this ion across the membrane along a line
normal to the surface; dE/dx is the potential gradient along the same line,
A is the surface area and / the frictional resistance per unit of average
velocity for the single ion along the normal line. N0 being the Avogadro
number, the expression i///V0 can be replaced by w', the absolute mobility
of the cation, or its mean velocity under unit force along the line of force. It
may be noted that the Pick diffusion coefficient D = RT/fN0 = RTu'.)
Another treatment is possible apart from this Nernst-Planck differential
equation, deriving originally from Danielli (1943) as given in Davson &
Danielli's book, The Permeability of Natural Membranes, and by Davson
(1951) in his General Physiology. In this there is applied to membrane
kinetics the activation concepts already in use for chemical kinetics and is
especially applicable to non-aqueous membranes. An important elaboration
of this activated-state theory has been recently published by Zwolinski,
Eyring & Reese (1949). The present treatment, which is dealt with in more
detail elsewhere (Conway, 1953), follows the classical Nernst-Planck
concepts.
The integration of equation (9), as Teorell (1952) points out, was given
by Behn (1897) as
„+'- u'[ARTt fo-'i) h^/
~ U
(Here c2 and cl are the total concentrations on each side of the membrane
and 3 the thickness of the membrane. A steady state is assumed.) This is
the solution with a homogeneous and non-ionic membrane. Teorell (1952)
has given the more complete solution, for an ionized membrane, but for the
present treatment the simpler condition is sufficient.
The expression between brackets is common to the various diffusing
ions, and this expression multiplied by cf may be taken as the flux of the
ion in one direction, and multiplied by c£' E, it may be taken as the flux in the
opposite direction, the ratio of the two partial fluxes being
When the total electrolyte concentration on each side of the membrane
approaches equality, equation (10) expressed for the partial flux from phase i
to phase 2 changes to u'AtRT Ing ..
+— +
or
THROUGH MEMBRANES 305
which is the same as the equation of Levi & Ussing (1948) when i///V0 is
substituted for u' '. (A very recent article by Ussing (1952) dealing more fully
with flux equations was received by the author when this manuscript was
going to press.)
Considering now the net flux of the cation (here Na) across the membrane,
this is also the net flux of the salt. The minimal energy expenditure in the
process is made up of the reversible work done plus an irreversible loss of
free energy.
The reversible work done. This may be expressed as
-AG^ztfrin^/q. (14)
(Here it is assumed that NaCl or similar salt is the only electrolyte on each
side of the membrane.) The sum of the work against the potential difference
for the Na and Cl across the membrane is zero.
The irreversible work. This corresponds to the loss of free energy in
diffusion, which in turn may be regarded as irreversible work done against
the frictional resistance to the net diffusion of molecules or ions. For
a neutral solute diffusing across a membrane, this may be expressed as
A^> ™ o nAtRTD, ., C, , ,
- AG = TAS = — - fo - c2) In ± . (15)
o c2
This irreversible change of free energy when Na ions are actively carried is
made up of such components as the diffusion to a higher chemical (but
lower electrochemical) potential of the Cl ions ; secondly, to the loss of free
energy in the back-diffusion of Na ions which were actively carried against
the electrochemical gradient; it would also be increased to some extent by
appreciable water fluxes through the membrane, though this latter may here
be neglected.
Loss of free energy in the diffusion of the Cl ions
In the present context this is considered to occur to a higher chemical
potential, or RT\n c2/cl is a positive quantity, but to a lower electrochemical,
or (RTlnc2/cl — EF) is negative. This latter expression is zero when there
is no net flow of the Cl ions ; then
E = RT/Flnc2/cv (16)
When a flow of Cl ions occurs the potential must exceed this figure.
The amount may be obtained from Behn's equation above, where £
is replaced by i/£ and u' by v', the latter being the mobility of the
Cl ions.
306 SOME ASPECTS OF ION TRANSPORT
From this equation it follows that
^ RT , c2
Here n~ in a steady state is also the total NaCl transported and the amount
of Na actively carried.
Transport potentials for the free ions
It is obvious that the second member on the right of equation (17) is the
surplus potential carrying across Cl~ ions. One may write
n-Slnc2/c1 . „
=
where Et is the transport potential. It will be seen that this is proportional
to the active transport of Na ions (which in a steady state is the same as
Cl~ transport) also to the thickness of the membrane, and inversely as the
mobility of the Cl~ ions.
From equation (18),
where n is the net transport of NaCl and n = n+ = n~.
Another source of free-energy loss occurs if there is a back-duffusion of
the Na+ ions, since for any effective carriage of Na+ ions against the electro-
chemical gradient, this back-diffusion represents a surplus carriage, and so
much extra energy is required for a given net transport. Such energy loss is
clearly the flux of Na+ ions from 2 to i, multiplied by (RTlnc^^ + EF) or
Collecting these three energy quantities one obtains
AG -2nRT\nc Ic + n'8lnc^ Atu'^-c^RTlnc^/c^
-AGrTotal-2»^^n^l + ^7(~^j+ 8\^fc) •
(20)
Approach of the transport system to the maximum efficiency
It will be seen from equation (20) that the efficiency of the transport of
NaCl increases as the mobility of the free Na+ ions through the membrane
decreases towards zero, and as the mobility of the free Cl~ ions increases.
It is very probable that no appreciable amount of free energy is lost in the
cyclical movements or displacements of the carrier itself in the membrane.
For it we may assume no overall diffusion gradients, and that the complex
as a whole is neutral.
THROUGH MEMBRANES 307
Where the redox pump operates, efficiency will also depend on how much
of the energy of the potential jump of the electrons is transferred to osmotic
work on switching their passage through the membrane system. As this
energy transference may reach close to 1 00%, it will appear that efficiency
of transport may be very appreciably increased where the anion as well
as the cation is actively carried.
V. ACTIVE TRANSPORT OF Na+ AND K+ IONS IN YEAST:
TWO DISTINCT CARRIERS INVOLVED
Active K absorption during fermentation
An account has been given of the active uptake of K+ ions by fermenting
yeast (Conway & O'Malley, 1946). When one part of yeast ferments with
one part of 5 % glucose unbuffered and KC1 is present to the extent of 100
or 2OomM/L, then K+ ions are rapidly taken up in quantity from the
suspending fluid and H+ ions returned in practically equivalent amount,
and with previous oxygenation for many hours the pH is of the order of
i -6-1 »7, but if the suspending fluid is reduced to lower and lower volumes
the pH approaches 1-5. In the fermenting suspension cyanide 2 or 4 mM/1.
has no appreciable effect, but azide in similar strength practically abolishes
the effect but increases the amount of alcohol produced. (It may inhibit the
rate of alcohol production.)
Simultaneously, with the excretion of acid, there is an equivalent, or near
equivalent, production of alkali inside the cells. The process is very specific
for K+ ions as against Na+ ions, and it takes some 25 times the concentration
of Na+ as compared with K+ to have equal uptake. Rb is taken up at about
one-half the rate of K+ ions, and both Cs and Li behave like Na.
At the end of the fermentation the K+ ions taken up are slowly returned
to the suspending fluid.
If no KC1 is present in the 5 % glucose, acid is still excreted in somewhat
reduced quantity, and considerably so if the yeast has been oxygenated for
hours beforehand. The H+ ions here are associated with succinate ions as
succinic acid. Most of such succinate ions were already present in the yeast
cells before the fermentation, but if a large amount of suspending fluid is
used, and there is prolonged fermentation, new formation of succinate ions
(or succinic acid within the cells) may be shown to occur throughout.
(This continued formation of succinic acid is in agreement with the work
of Kleinzeller, 1941 .) The excretion of H+ ions by the cells into the suspend-
ing fluid and their appearance as succinic acid is carried out by the same
mechanism as when H+ ions appear in exchange for K ions. Though the
most striking lowering of the external pH is produced by the H+ and K+
308 SOME ASPECTS OF ION TRANSPORT
exchange, the secretion of free succinic acid is here analogous to the secretion
of hydrochloric acid by the oxyntic cells.
The redox theory. The active secretion of H+ ions here as for the gastric
oxyntic cells is explained by the redox theory (Conway & Brady, 1948;
Conway, 1951, 1952). A similar theory was advanced shortly afterwards by
Crane & Davies (1948). A full account is given in a recent publication
(Conway, 1952); vide also the article of Davies, 1951).
Briefly, in this theory it is considered that metabolic H atoms are trans-
ferred via a flavine enzyme to a metalic respiratory enzyme in the membrane.
This retains the electrons and sets free H+ ions. The appearance of the free
H+ ions result in a potential (a transport potential as described in the
previous section) which can pull across accompanying anions, or, alter-
natively, such are in turn actively excreted. In the case of yeast with K+
and H+ exchange, K+ ions are brought across in an adsorption complex of
the type discussed above, and as the electrons are passed internally K ions
are liberated within the cell.
Post-fermentative permeability of yeast cells to K ions
If resting yeast cells are suspended and shaken anaerobically, the entrance
of K+ ions is extremely slow (Conway & Moore, 1950). After some hours
labelled K+ externally has mixed with the internal K+ ions to only a small
percentage or less.
If such yeast be shaken in air, labelled K+ ions enter quite readily, and
after i hr. the mixing may proceed to some 50 or 40 %.
This entrance is almost entirely abolished by cyanide, which contrasts
with the effect of this inhibitor during active fermentation, but like the
effect in fermentation azide almost altogether abolishes it.
It is considered that the same K-carrier is here operative as during
fermentation, but cannot transfer electrons to an organic acceptor as during
active fermentation, but finally to oxygen.
The active excretion of Na+ ions by yeast
As mentioned above, provided Na+ ions are in high proportion and K+
ions absent, or in very low concentration, Na+ can be absorbed in
fermentation.
When unbuffered N/io-NaCl exists outside the cells during fermentation
without any KC1 very little Na ions enter; where, however, M/5 -sodium
citrate is present outside the cells and the yeast is suspended in 20 times its
volume of this solution containing also 5 % glucose, very appreciable
amounts enter after about i hr. fermentation. When the yeast is then washed
THROUGH MEMBRANES 309
twice with 20 times its volume of water the cells retain as much as 60 m.equiv.
Na+ ions.
The Na+ ions may be assumed to be brought into the cells on the K-
carrier, and even a small amount of K+ ions outside will prevent the Na+
uptake. If a series of such fermentations in M/5-sodium citrate is carried
out, it is possible to exchange practically the entire amount of K ions in the
yeast cell for Na ions (Conway & Moore, 1952).
The special Na-carrier within the cells
Once the Na ions are introduced in quantity they are actively excreted
outwards by a special carrier for Na ions (Table i), which is as much, or
even more, specific for Na as the K-carrier for K ions. The existence of this
special carrier is shown by the following facts.
Table i . Changes of Na and K content of Nz-yeast preparations (given as
mmol./kg. of centrifuged yeast) on suspension in N/io-KCl and in water
Time of
suspension
(hr.)
InN/io-KCl
In water
Na change
K change
Na change
K change
0
I
3i
19
— 20-6
-36-8
-40-9
-47-8
+ 25-3
-f39'8
+ 41-2
+ 41-8
-7'9
— 14*0
-19-8
-23'4
-0-3
-0-8
+ 1-0
+ 3'7
Average Na and K content of yeast at zero time :
Na content = 64-9 mmol./kg. centrifuged yeast.
K content = 84-7 mmol./kg. centrifuged yeast.
(a) Specificity. Using labelled Na+ and K+ in suitable experiments with
the Na-yeast, it has appeared that labelled K+ in a Na-yeast prepared as
above (with about 60 m.equiv. Na and 70 m.equiv. K/l.) is not carried across
the membrane in appreciable amounts over some hours whereas the Na
ions are freely carried, the external fluid containing 100 mmol. NaCl and
100 mmol. KC1/1.
(b) Effect of inhibitors. Whereas azide inhibits the K-carrier system, and
practically abolishes the carriage of K+ when present in 2 mM/1. this has no
effect on the Na-carrier system, as shown when Na is being excreted from
a Na-yeast into water. Also if Na is being excreted at a much more rapid
rate into 100 mM/1. KC1, azide reduces the Na+ transport to the same rate
as into water only.
On the other hand, cyanide (2 mM/1. strong) and anoxia strongly inhibit
both the K+ and Na+ carriers in Na-yeast (but do not inhibit these carriers
in active fermentation). Table 2 shows the effect of oxygen lack.
3IO SOME ASPECTS OF ION TRANSPORT
Table 2. Changes of Na and K content of yeast suspended for 90 mm. (at
18° C.) in tap water and in M/io-KCl, and the effect on such changes of
oxygen lack. (The washed Na-yeast was suspended in 20 times its
volume of fluid)
Conditions
Na change mmol./kg.
K change/kg, centrifuged
yeast
In tap water
In M/io-KCl
In tap water
In M/io-KCl
Oxygen present
Oxygen absent
-i3'7
-3'8
-38-0
-8-6
-0-8
— 1-7
33'3
-0-3
The Na-yeast immediately after suspending had 60 mmol. Na/kg. and 78 mmol. K/kg.
of centrifuged yeast.
Interaction of the Na+ and K+ carriers
(a) If the Na-yeast is suspended in 20 times its volume of o-i M-NaCl no
net excretion occurs over 24 hr. At the same time there is very appreciable
movements of Na+ ions across the membrane, which we may assume to be
carried to and fro.
The inclusion of Na azide in the suspending fluid entirely inhibits the
entrance of Na+ ions, and as a result there is a net excretion of Na+ ions.
The interpretation here is that, without the azide, Na ions without any
competition from K ions are being actively carried back into the cell by the
K-carrier at the same rate as they are being brought outwards by the specific
Na-carrier. Azide inhibits the K-carrier and so there is a net excretion.
(b) When the Na-yeast is suspended in o-i M-NaCl a small inclusion of
KC1 (0-005 M) in the fluid causes a marked excretion of Na+ ions. The rate
can be doubled or more by increasing the KC1 further to about 0-025 M,
but it is not further affected by increase of external KC1.
K+ ions are now taken up in approximately equivalent exchange for the
Na+ ions excreted.
The interpretation here is not that K ions displace Na+ ions on the Na-
carrier, since the large concentration of K+ ions inside the cells does not
affect the Na+ carriage, no appreciable amounts of K+ ions being actively
extruded ; but rather that electrons on the Na-carrier are transferred to the
K-carrier and K+ ions carried inwards in consequence, the electrons being
then transferred to the oxygen system and the K+ ions liberated.
(c) When the Na-yeast excretes Na+ ions into water, or into o- 1 M-NaCl
plus Na azide (0-002 M) or into o-iM-NaCl plus o-i KC1 plus 0-002 M-Na
azide, the rate of excretion in all three cases is the same. However, when
cyanide is included (2-5 mM/1.) the excretion drops towards zero.
The interpretation here is that the Na-carrier is not affected by azide,
but can transfer its electrons to the oxygen system when the K-carrier is cut
THROUGH MEMBRANES 311
out but then at a much slower rate than when this is operative in the transfer
of K+ ions. Cyanide, however, blocks the path of these electrons to oxygen
whether directly transferred by the Na-carrier alone or via the K-carrier.
While the * redox-pump ' theory affords an explanation for the various
facts in connexion with the active transport of K ions and excretion of Na
ions, the following special supporting evidence may be given.
(a) The effect of water soluble redox dyes on the K and H ion exchange
during fermentation
During fermentation in the presence of N/io-KCl, with one part of
centrifuged yeast to 20 volumes of 5 % glucose containing M/io-potassium
succinate buffer, a steady pH value of approximately 4-5 was maintained
over some hours. Conducting the fermentation anaerobically (Conway &
Kernan, 1953) in a stream of nitrogen and including various redox dyes
in strength M/ 10,000 with characteristic potentials ranging from 290 to
— i6omV, it was found that the potential is registered by a platinum
electrode against the saturated calomel electrode (and calculated as against
the normal hydrogen electrode) altered in correspondence with the
characteristic potential of the redox dye.
Without any such redox dye the pH was about 180 mV. When it was
raised beyond this value there was an increased K- and H-ion exchange.
When it was reduced below this value the exchange was reduced, and
could be altogether abolished at potentials at or somewhat below 100 mV.
Plotting the potentials with the dyes included against the ratio of the
H-ion secretion to the control value (without any dye) gave a practically
linear relation, with some scatter for different dyes.
When the H ions secretion was in this way abolished, there was almost
no effect on the CO2 or alcohol production.
(b) The relation of the oxygen consumption to the Na output by Na-yeast
excretion at its maximum at room temperature
From the purely energetic standpoint when Na-yeast excretes actively
into solutions containing N/io-NaCl and sufficient KC1 to reach maximum
output, the work done on such excretion is only a small percentage of the
total energy output. When, however, the number of Na ions excreted for
one molecule oxygen uptake is examined (the suspension being i part of
Na-yeast to 100 of suspending fluid), the ratio is near to 4 for the first
50 min. In short, under such conditions, the obvious interpretation is
that all or nearly all the electrons which finally reach oxygen pass through
the transport system involving the Na and K carriers (Conway, Ryan &
Carton, 1954).
312
SOME ASPECTS OF ION TRANSPORT
(c) The azide effect
Azide (2 HIM) inhibits the K-carrier, but not the Na-carrier. Azide in
such concentration completely abolishes new formation of organically
bound phosphate, even during active fermentation. From this latter action
(Conway, Carton & McGovern, 1953) the transference of the oxidative
energy to energy-rich phosphate bonds ceases and the practically exact
agreement of the active transport of Na ions under such conditions and
into water without any azide shows that the direct usage of electron energy
rather than indirectly through phosphate bond energy interprets the
results.
160 r
Fig. i. Curves of extrusion of Na (as mg./kg. yeast) from a Na-yeast suspended in 20
times its volume of o-i KC1 and 5 % alcohol (Conway & Moore, 1952) which had about
half its K replaced by Na ( = 60 mmol./kg.). Control curve (without hormone). Curve
with DOCA. Curve with i7-hydroxycorticosterone. Control curve (middle) without
hormone. Curve with DOCA. Curve with i7-hydroxycorticosterone.
The effect of adrenal cortical hormones on active transport
of ions in yeast
The effect on the excretion of Na from Na-yeast of various cortical
hormones and other steroids included in the suspending fluid was examined
(Conway & Hingerty, 1953). Ethanol to the extent of 5 % was incorporated
in the fluid to assist the steroid solubility. Of the steroids examined, only
DOCA, cortisone and compound F showed any effect. DOCA always
definitely inhibited, cortisone to a much smaller extent, and compound F
had a small but definite stimulating effect (Fig. i). The amounts of these
steroids present per litre suspending fluid were approximately 0-002, o-o
and o-oo mM respectively.
THROUGH MEMBRANES
313
VI. THE ACTIVE EXCRETION OF Na IONS
FROM MUSCLE
To observe this in mammalian muscle in vivo (Conway & Hingerty, 1946),
rats were used with skeletal muscle containing relatively large amounts of
Na (upwards of 50 m.equiv./kg.) as a result of feeding for a month on a
K-free diet in accordance with the experiment of Heppel (1939). Such rats
were then put on a diet with high K content, and the skeletal muscle actively
excreted its contained Na slowly. Fig. 2 illustrates the results obtained. In
about 12 days the muscle Na had decreased to normal level, and about
3 days were required for the average half-period of excretion. This may be
considered the first unequivocal demonstration of active Na extrusion in
quantity from muscle, and under conditions in which a considerable
amount of Na had entered the muscle fibres in vivo.
32-
30-
z24"
22-
20-
18-
Muscle Na*
11
10
9
8 E
7 I
c
**
5
4
1 234567
Days on high K diet
Fig. 2. Mean curve (six rats used) of Na content of skeletal muscle, after a period of
K-free diet and restoration to a high K diet. The dotted line gives the normal level.
Curves for plasma K and muscle K for the same rats are also given.
The question o/Na excretion from the isolated frog sartorius
Steinbach (1940) investigated the question of the active extrusion of Na
from the isolated frog sartorius when much sodium had entered the
muscle fibres after 24 hr. immersion in K-free Ringer fluid. Such immer-
sions were carried out using two sartorii from each frog. One of these was
then re-immersed in Ringer fluid containing K ions (10 m.equiv./l.) and it
was concluded that excretion of Na ions was demonstrated.
Such experiments have been already discussed (Conway, 1945, 1946).
Steinbach has recently published experiments (1951) in which the con-
ditions were better controlled and with 12 pairs of muscle he
314 SOME ASPECTS OF ION TRANSPORT
obtained an average of 48 ± 2-7 m.equiv./kg. after the first immersion
and 39 ± 1*7 m.equiv./kg. after the second immersion or a difference of
9 ±3-2 m.equiv./kg. While this made the excretion of Na very probable
we have been unable to obtain quite the same results on repeating his
procedure, due, no doubt, to some undetermined difference in technique.
Very recently (Desmedt, 1953) considerable excretion of Na has been
demonstrated using single isolated sartorii immersed in Ringer fluid con-
taining K ions. The muscles had gained much Na by previous immersion
in K-free Ringer fluid. It is, however, important to note that the Ringer
fluid used differed from that of Steinbach in so far as it represented the
average inorganic composition of the frog's plasma (Boyle & Conway,
1941).
The localization of Na in muscle
It is in general assumed that when the interspace Na is subtracted from
the total muscle Na, the remainder exists in the fibres and is evenly distri-
buted through their substance. The following evidence shows that such is
very probably not the case, and that most of the fibre Na is either localized
in the sarcolemma, or alternatively concentrated in a small group of fibres
(about 10% of the whole) with the Na content of the typical fibre, and
behind the sarcolemma, only of the order of 2 or 3 mM/kg. total muscle.
Such a conclusion is based on the following evidence (Conway & Carey,
(a) The curve of entrance of labelled K into the sartorius in the cold. When
the isolated sartorius is immersed in Ringer fluid containing, say,
iomMKCl/1., and a small amount of labelled K is introduced, this enters
at first very rapidly to a level representing about 0*65 of the whole muscle.
After this there occurs a very slow further entrance. This is linear up to an
hour or more, and extrapolating the line to cut the ordinate gives the 0-65
fraction as given above. The half-period of this first zone is about 8 min. at
or near to o° C. (Fig. 3).
The picture is clearer if sulphate-Ringer instead of Cl is used, the Cl ions
being entirely replaced by sulphate.
This first zone appeared in each of a large number of experiments
conducted with all suitable precautions. It was also evident when the
muscles were immersed in frog heparinized blood and a little labelled K+
introduced, the muscles being cut at no point, and having a little of the
pelvic bone attached; also their surface was not dried in any way before
immersion.
It will be seen that the space of 0*65 is far higher than the interspace
between the fibres, which is approximately 0-13 (Conway, Kane & O'Reilly,
1941) so that there must be some fibre region into which labelled K in the
THROUGH MEMBRANES
315
cold enters relatively very swiftly. Such a region may be the sarcolemma, or
alternatively some group of fibres.
This first region of entrance, even with heparinized frog blood and
muscle surface not dried was first described at the Physiological Congress
in Copenhagen (1950). Very recently a similar initial swift entrance of
42K into the fibres of the frog sartorius was described by Harris (1953)
and attributed to entrance into injured fibres.
(b) The effect of increasing the external K concentration on the level of this
first entrance. If K ions can enter a fibre region so quickly it seems likely that
Na ions would enter it more rapidly than into the general fibre space, and
in turn that such a region may well contain a high proportion of Na ions.
1-8
1-6
Sg 0-8
* E
S 0-6
x
0-2
20
40 60
Minutes
80
100 120
Fig. 3. Curves of labelled K entrance (given as ratios count of i c.c. muscle: count of i c.c.
external fluid) into isolated sartorii at o° C. Curve A: mean curve for six experiments
with sartorii immersed in sulphate Ringer (all Cl replaced by SO4) and K in external
fluid = 10 m.equiv./l. Curve B: entrance of 42K into the companion muscles used in A,
immersed in chloride Ringer with 10 m.equiv. K/l. Curve C: mean entrance of 42K into
sartorii immersed in heparinized blood.
This may be tested by increasing the K content outside the muscle. If
the region in question contains approximately the same non-diffusible anion
content as the general fibres, then it should contain, if all the diffusible
inorganic cations were K and Na, about 94 mM/kg. of total muscle. On
increasing the K content outside up to high values the level of the first
region of the curve of entrance of 42K should appear to decline to about
one-ninth the value when the external K was 10 m.equiv./l. The observed
decline, however, is not nearly so marked, being about one-third at most.
Such is readily explicable if the region in question contains a relatively large
amount of Na ions with low levels of K in the external fluid.
(c) The effect on the first region of immersing the muscles in 3*2% glucose
containing 10 m.equiv. labelled ^.-sulphate. The level of the first region
316 SOME ASPECTS OF ION TRANSPORT
(average of three experiments) under these conditions was found to be
0-99 as compared with 0-52 for the sulphate-Ringer (the fibre interspace
being subtracted). The increased ratio is in accordance with expectation
if the first region contained a high proportion of Na+ ions, which would then
exchange entirely for K+ ions and give an increased level for the first part of
the curve of entrance of 42K.
A calculation of the volume of the fibre region in question may be made
from such data. Thus if a be its relative volume and 94-6 m.equiv./kg.
represent the total of Na plus K in the fibres (Boyle & Conway, 1941), then
a is given by ax 94-6
10
- = 0-99,
so that a is approximately 0*10 or 10 % of the total fibre space.
25 r
Normal level of muscle sodium
10 20 30 40 50 60 80 90 100 110 120
Minutes
Fig. 4. Curves of Na and Cl losses from frogs' sartorii immersed in 3-2 % glucose at room
temperature. Centre curve (•) : Na losses into 3*2 % glucose. Lower curve ((•)) : Cl losses
into 3*2% glucose. Upper curve (x): Na losses into 3'2% glucose containing 100
m.equiv. KC1/1. Dotted line gives mean level of Cl and HCO3 in normal muscle.
From more detailed studies it would appear that in vivo about 70 to
80 m.equiv./kg. of Na may be assumed to exist in this space and about 20-
30 m.equiv. K. Thus it would account for about 7-8 of the 10 m.equiv. of
Na in the muscle fibres. The K ratio across the membranes of such a region
would be about 8-12, and could be expected to give a membrane potential
of about 56mV.
This Na region is considered from such evidence to be not only very
freely permeable to K ions but also to Na ions, and that under normal
conditions some active extrusion of Na allows a K increase in the region
beyond the plasma level, but the major fraction of the non-diffusible anions
are balanced by Na ions.
THROUGH MEMBRANES 317
(d) Sodium exit from the sartorius into 3-2% glucose. When sartorii are
immersed in 3-2% glucose the Cl, which is almost all in the interspace,
comes out rapidly. The interspace Na comes out as fast as the Cl, but the
fibre Na much more slowly (Fig. 4). Most of the fibre Na is lost when only
a small fraction of the K has emerged. This may be explained by active
extrusion of Na ions from the whole fibres. At the same time the difference
is in large measure dependent on the localization of the fibre Na. It is
apparent from the evidence given above that the fibre Na comes out for the
most part from a special fibre region with a much freer K and Na perme-
ability than the remaining region of the fibres.
If this latter conclusion be true then if 100 mmol./l. of KC1 is included
in the glucose, the Na+ ions should come out far more rapidly than into the
free glucose owing to free K and Na exchange. This in fact happens as
shown in Fig. 4, giving the exit of fibre Na when 100 mM-KCl/1. is included
in the glucose solution. Some, however, of the fibre Na remains for an
indefinite period or comes out very slowly. This Na fraction may be
regarded as the Na behind the sarcolemma or within the substance of the
typical fibres.
Is the readily interchangeable Na present in the sarcolemma or
localized in about 10% of the muscle fibres?
The available evidence as a whole appears to favour its presence in a group
of fibres rather than in the sarcolemma. The following may be noted.
(a) When sartorii are immersed in K-free Ringer overnight in the cold,
and much Na has entered the fibres, such Na appears to come out just as
readily into K2SO4, for example, as the normal fibre Na. Further, neither
cyanide, iodoacetate nor azide when each is present as a mM/1. has any
appreciable effect on the rate of emergence (Conway & Carey, 1954). Here
the simpler interpretation is that individual fibres on immersion in K-free
Ringer change and allow much Na to enter.
(b) The fact that when nearly half the muscle K has been displaced by
Na, due to a K-free diet (Heppel, 1939) extending over some weeks, then
24Na exchanges very rapidly with all this muscle Na. But it takes about
3 days for half of it to be extruded on return to a high K diet (Fig. 4) (Conway
& Hingerty, 1946). With such diet the plasma K passes above the normal
plasma value after i day. Exchange diffusion (Ussing, 1949) has been
proposed to explain this result, but it is more simply interpreted by the
alteration in permeability of a group of fibres. With this change Na ions
come to move in and out freely and the fibres may then be relatively
impermeable to Cl ; for although much Na has entered the fibres in quantity,
there is but little increase in the muscle Cl.
318 SOME ASPECTS OF ION TRANSPORT
(c) At the same time Nastuk & Hodgkin (1950) have examined the
potential across the membrane of a number of individual fibres and obtained
an average value of about 88 mV. at 18° C. and 84 mV. at 7° C. The
potentials at room temperature 'usually lie between 80 and 95 mV. Resting
potentials as low as 60 mV. were occasionally observed but are regarded as
due to faulty impalement.* These fibres if included 'would not have altered
the average values by more than i or 2mV.' If we were to take the 2mV
here and an average potential for the accepted fibres as 84, then 8 % of the
whole fibres giving 60 mV. would only alter the 84 to 82 mV., and 8% is
close to the 10 % considered above and which can only be considered a very
approximate figure.
Table 3. Relative net entrance rates of ions into muscles compared with the
relative theoretical diffusion constants through water
Cation series *Z)' for single ions, with K value = 100
KC1
100
K
100
RbCl
38
Rb
103
CsCl
8
Cs
104
NaCl
0
Na
67
LiCl
o
Li
52
CaCl2
o
Ca
40
MgCl2
0
Mg
35
Anion series
KC1
100
Cl
100
KBr
63
Br
105
KNO3
17
NO,
96
K phosphate
4
H2P04
50
KOOC.CH3
3
HPO4
39
KHCO3
i
CH3COO
54
K2S04
o
SO4
53
The values of D were determined from the formula RT/zF2. The entrance rate of
K phosphate is for nearly equal mixtures of K2HPO4 and KH2PO4 . (Table from Conway,
1 947 ; entrance rates of salts from Conway & Moore, 1 946.)
At the same time no very definite conclusion can be drawn as to the
region of localization. It may also be pointed out that for experiments
where exact figures are advanced for the Na ion rates across the muscle-
fibre membrane (e.g. Harris & Burn, 1949), the result is complicated by the
localization of the fibre Na in two regions, from one of which the Na will
move rapidly, and from the other very slowly. An incorrect comparison
could thus be drawn between the rate of passage of free K and Na ions
across the typical muscle-fibre membrane.
Passage of K ions across muscle and nerve-fibre membranes
Whereas in yeast the passage of K ions into the cell is dependent on an
active transport mechanism the passive entrance of the free ions being
relatively very slow, the entrance of K and Cl ions into muscle was shown to
THROUGH MEMBRANES
319
be a passive process (Boyle & Conway, 1941), and similarly for liver cells
(Conway, 1944) and the cells of the proximal convoluted tubules of the
frog's kidney (Conway, FitzGerald & MacDougald, 1946). Shanes (1946)
showed such permeability to apply also to nerve fibres, working with
spider-crab nerves; and this was also found applicable to nerve fibres by
Hodgkin (1947) and by Hodgkin & Huxley (1947).
Immersion of frogs' sartorii in the cold in Ringer solution containing
various amounts of KC1 and NaCl produced equilibrium values of K, Cl,
water content and of resting potentials which were predictable from
a consideration of simultaneous Donnan, electrical and osmotic equilibria.
It appeared that for membranes of the soft tissues of the body in general
a certain type of equilibrium applied which may be described as follows :
(a) Below a certain size level solute particles will penetrate the cell
membrane freely, independently of their lipoid solubility.
(b) Small cations and anions can in general both pass the cell membrane
with varying degrees of freedom.
(c) The critical size in general for rapid passage of cations is at the K
level (hydrated ion) or between it and that of the Na ion, and the critical
size for the anions is at or near the dimensions of the Cl anion (Table 4).
In short, for free entrance of cations or anions, it is approximately 8 A.
diameter. Thus, while K, Rb and Cs ions can enter the cell at appreciable
rates over short periods, Na and Li ions are virtually excluded; and while
Cl, Br and NO3 ions enter freely, HCO3 and CH3COO ions diffuse very
slowly and SO4 ions are practically excluded. That the ion size is not the sole
determinant of the entrance rate is shown by the fact that with muscle Cs
enters much less rapidly than K. (The same would appear also for nerve,
though the difference is less marked (Hodgkin, 1947).
Table 5 shows the relative rate of ions into muscle fibres with the relative
theoretical diffusion constants through water.
Table 4. Ion sizes and hydration
Ionic radii
Hydration
Hydrated
Non-hydrated
(mol. H2O)
Rb
,
3-6 1 0-49
0'5
Cs | 3-6
1-65
0-2
K
3'8
i'33
3'8
Na
5'6
0-98
8-0
Li
7'3
0-78
12-6
Ba
8-8
i'43
i3'5
Ca
9-6
i -06
17-6
Sr
9-6
1-27
14-6
Mg
10-8
0-78
22'2
Table from Conway (1947); data compiled from various sources. Ionic radii in
Angstroms.
32O SOME ASPECTS OF ION TRANSPORT
From various reviews by other workers, e.g. Hodgkin (1951), Keynes
(1951), Ussing (1949) and Katz (1952), it will be seen that the general nature
of the above views have been accepted.
With respect to muscle-fibre sodium, Krogh (1946), relying on the isotope
studies of Heppel and others, concluded that Na ions entered the muscle
fibre as fast as, if not faster than, K ions. This was shown to be erroneous
from energy considerations (Conway, 1946), and from the energy require-
ments alone Na could enter only at a fraction of the rate of K ions, there
being a constant active extrusion of the entering Na in the steady state.
Maizels (1951) has represented us in this connexion as holding that Na
could not be actively excreted because of energy requirements. The point
made against Krogh's views was, however, that Na could not be actively
excreted (and a steady state reached) if Na ions enter as fast as K ions
per unit concentration. From the energy calculations they could only enter
much slower.
If one judges the true rate of passage of Na ions and their active extrusion
from the muscle fibres in vivo then the rat experiments (Conway &
Hingerty, 1946) commented upon above show a half-period of some days
for extrusion of Na which had entered muscle in quantity. Hodgkin &
Huxley (1947) find that free Na ions move across the fibre membrane of the
squid axon at only about one-fortieth the rate of K ions.
For isolated sartorii muscle Harris & Burn, using isotopes, give the rate
for K ions as about seven times that of Na ions, but as discussed above the
result for Na ions will give in effect only the rate from localized regions of
Na accumulation. The true rate for the typical muscle fibre is probably
much slower than that given by Harris and Burn (1949).
VII. THE ACTIVE TRANSPORT OF Na IONS ACROSS
THE FROG'S SKIN
The study of the electrical phenomena of the frog's skin and its active
transport of ions has had a long history proceeding from Du Bois Reymond
(1848), through Galeotti (1904) and many later observers. In recent years
much valuable work has been done by Ussing and co-workers (vide Ussing,
1949, 1952). In a recent important paper by Linderholm (1952) from
Teorell's laboratory, a theoretical introduction of interest is given, also
a historical review. Linderholm confirms the findings of Ussing & Zerahn
(1951) and adds much new material. He interprets his results with respect
to the active transport of Na ions in terms of the * redox pump ' (Conway,
When a counter potential is applied across the mounted frog skin to
reduce the potential difference across the skin to zero a current flows in the
THROUGH MEMBRANES 321
controlling circuit which is entirely due to actively carried Na ions (as
shown by Ussing & Zerahn (1951) and confirmed by Linderholm (1952)),
no net exchange of Cl ions occurring.
The oxygen consumed and Na ions transferred in these short-circuited
skins is 4-8/1!. O2cm."2hr.~1, and 238/^/*F. cm.~2sec.~1 respectively. The
ratio of Na ions carried to oxygen molecules consumed is therefore 3-4. The
theoretical ratio, if all the electrons received by the oxygen were diverted to
Na carriage, is 4-0. The Na ions would appear to be carried by the epithelial
cells of the skin which have a high oxygen consumption (Erdman & Schmerl,
1926). Some oxygen is no doubt consumed by other cells of the skin, but it
is quite probable that the epithelial consumption accounts for nearly the
full amount.
Linderholm also discusses other conditions, in which the ratio would
appear also in harmony with the theory. Such ratios may be compared with
those found in the active extrusion of Na by yeast dealt with above. He
points out too that the theory of the 'redox pump* readily explains the
current produced by the skin, and the dependence of the electrical potential
on the oxygen supply as well as the inhibition of the potential by several
respiratory inhibitors.
VIII. SOME RECENT VIEWS CONCERNING THE
NATURE OF GASTRIC ACID SECRETIONS
Work on this question up to 1950 was reviewed by the author (Conway,
1952), and while this was in proof a review on the same subject by Davies
(1951) appeared.
Nielsen & Rosenberg (1951) have advanced a theory, involving the
transfer of energy by ATP in the formation of a monometaphosphate ester
of a sterol, and the change of this by enzyme action into a derivative of
orthophosphoric acid with H-ion increase in the membrane (which is anion-
but not cation-permeable). Cl ions are exchanged for the organic phosphate
ions across the membrane which in turn are split by a phosphatase, and the
original sterol or sterol-like substance reformed. Discussion must here be
limited to pointing out that the existence of the required monometaphosphate
ester and the special enzyme have not been demonstrated. Also the theory
was advanced chiefly to account for the very high HC1/O2 ratios recorded
by Davies which have been recently criticized by Davenport (Davenport,
1952; Davenport & Chavre, 1952).
Another kind of theory has been advanced by Hogben (1951). In this
the H ions of the gastric juice derive from the reaction
322 SOME ASPECTS OF ION TRANSPORT
in the canaliculus, HCOjj" being exchanged for Cl~ ions actively transported
across the membrane, leaving H+ and Cl~ ions. The assumptions in this
theory are too considerable to make it plausible. Thus it is assumed that Cl""
on a carrier is displaced by HCCKf in a concentration of the order of io~9 M,
or that the carrier is about 10 million times more specific for HCO3 than
Cl ions in the membrane or on the edge of canaliculus, while inside the
substance of the cell the carrier is relatively specific for Cl ions, since the
efficiency of the whole process depends on this specificity. Further it is
assumed that the moment HCOg" is in this way attached to the carrier its
energy level is raised by about 10,000 cal. of free energy per ion equivalent,
and no indication is made as to how this happens. It is also assumed that
a back-flow of HCOjf into the canaliculus occurs, or that the membrane is
permeable to free HCO^~ ions, in which case a very high diffusion potential
(about 350-400 mV.) could be expected.
Also, the frog's mucosa contains at least three different kinds of cell
lining the boundary between the serosal and mucosal sides, and to produce
an overall electrochemical potential difference of zero across the membrane
does not necessarily imply that this holds for the oxyntic cells; and the
evidence for the active carriage of Cl ions (which has never otherwise been
shown for animal tissue cells) should be examined in the light of this
objection. Finally, it may be remarked that carrier transport of Cl ions
would be quite in harmony with the 'redox theory' (vide Conway, 1952),
in which it is not assumed that H ions are transported but rather metabolic
hydrogens. These are transferred in a similar way to normal metabolism
except that a localization of certain phases occurs in the membrane.
SUMMARY
1. Mechanisms for the immediate energy source in active transport,
chiefly the 'redox pump', have been discussed.
2. The energy required in active transport, involving reversible and
irreversible energy changes, have been expressed in equational form.
3. The significance of * transport potentials' has been examined.
4. Recent developments in the nature of Na- and K-active transport
in yeast have been described. There are two distinct carriers, one specific
for Na, which removes Na from the cell, and one specific for K, which
introduces K into the cell. Azide (2 mM/1.) does not inhibit the Na-
carrier but cyanide and oxygen lack practically entirely inhibit both.
The cyanide inhibition does not occur if the cells are rapidly fermenting
glucose.
5. DOCA (and to a lesser extent cortisone) in 0-025 mM/l- concentration
markedly inhibits the extrusion of Na ions from Na-yeast immersed in a
THROUGH MEMBRANES 323
KC1 solution, but it does so indirectly by its effect on the K-carrier.
Compound F stimulates the Na excretion or the K uptake.
6. Evidence is advanced for the view that the fibre Na in skeletal muscle
is either localized in the sarcolemma, or exists in a special group of fibres.
7. Criticisms have been made of recent theories of the nature of HC1
secretion.
REFERENCES
BEHN, V. (1897). Ann. Phys. Chem., N.F., 62, 54.
BOYLE, P. J. & CONWAY, E. J. (1941). J. Physiol. 100, i.
CON WAY, E. J. (1944). Unpublished observations.
CONWAY, E. J. (1945). Biol. Rev. 20, 56.
CONWAY, E. J. (1946). Nature, Lond., 157, 715.
CONWAY, E. J. (1947). Irish J. Med. Set. 262, 263.
CONWAY, E. J. (1951). Science, 113, 270.
CONWAY, E. J. (1952). The Biochemistry of Gastric Acid Secretion. Springfield,
U.S.A. : Thomas.
CONWAY, E. J. (1953). Int. Rev. Cytol. 2, 419.
CONWAY, E. J. & BRADY, T. (1948). Nature, Lond., 162, 456.
CONWAY, E. J. & CAREY, M. (1954). Biochem. J.
CONWAY, E. J., CARTON, E. & McGovERN, H. (1954). In preparation.
CONWAY, E. J., FITZGERALD, O. & MACDOUGALD, T. C. (1946). J. Gen. Physiol.
29, 305-
CONWAY, E. J. & HINGERTY, D. (1946). Biochem. J. 40, 561.
CONWAY, E. J. & HINGERTY, D. (1953). Biochem. J. 55.
CONWAY, E. J. & KANE, F. (1942). Unpublished observation.
CONWAY, E. J., KANE, F. & O'REILLY, H. (1941). J. Physiol. 99, 401.
CONWAY, E. J. & KERNAN, R. (1954). In preparation.
CONWAY, E. J. & MOORE, P. (1946). Unpublished observations.
CONWAY, E. J. & MOORE, P. T. (1950). Biochem. J. 47, iii.
CONWAY, E. J. & MOORE, P. T. (1952). International Congress of Biochem.:
Abstracts.
CONWAY, E. J. & O'MALLEY, E. (1946). Biochem. J. 40, 59.
CONWAY, E. J., RYAN, H. & CARTON, E. (1954). Biochem. J.
CRANE, E. E. & DAVIES, R. E. (1948). Biochem. J. 43, xlii, xliii.
DANIELLI, J. F. (1952). Symp. Soc. Exp. Biol. no. vi, p. i. Cambridge University
Press.
DAVENPORT, H. W. (1939). Amer.J. Physiol. 97, 32.
DAVENPORT, H. W. (1952). Fed. Proc. 2, 715.
DAVENPORT, H. W. & CHAVRE, V. J. (1952). Amer.J. Physiol. 171, i.
DAVIES, R. E. (1951). Biol. Rev. 26, 87.
DAVIES, R. E. & KREBS, H. A. (1951). Symp. Biochem. Soc. no. 8, p. 77.
DAVIES, R. E. £ ROUGHTON, F. J. W. (1948). Biochem. J. 42, 618.
DAVSON, H. (1951). Textbook of General Physiology. London.
DAVSON, H. & DANIELLI, J. F. (1943). The Permeability of Natural Membranes.
Cambridge University Press.
DESMEDT, J. E. (1953). J. Physiol. 121, 191.
Du Bois REYMOND, E. (1848). Untersuchungen iiber Tierische Elektricitat. Berlin.
ERDMAN, R. & SCHMERL, E. (1926). Arch. exp. Zellforsch. 2, 280.
GALEOTTI, G. (1904). Z. phys. Chem. 49, 542.
HARRIS, E. J. (1953)- J- Physiol. 120, 246-53.
HARRIS, E. J. & BURN, G. P, (1949). Trans. Faraday Soc. 45, 508.
324 ASPECTS OF ION TRANSPORT THROUGH MEMBRANES
HEPPEL, L. A. (1939). Amer.J. Physiol. 127, 385.
HODGKIN, A. L. (1947). J. Physiol. 106, 305.
HODGKIN, A. L. (1949). Arch. Sci. physiol. 3, 151.
HODGKIN, A. L. (1951). Biol. Rev. 26, 339.
HODGKIN, A. L. & HUXLEY, A. F. (1947). J. Physiol. 106, 341.
HODGKIN, A. L. & KATZ, B. (1949). J. Physiol. 108, 37.
HOGBEN, C. A. M. (1951). Proc. Nat. Acad. Sci., Wash., 38, 13.
KATZ, B. (1952). Symp. Soc. Exp. Biol. no. vi, p. 16.
KEYNES, R. D. (1951). The Role of Electrolytes in Excitable Tissues. Published by
the Institute Universitade.
KLEINZELLER, A. (1941). Biochem.J. 35, 495.
KROGH, A. (1946). Proc. Roy. Soc. B, 133, 140.
LAMM, O. & MALMGREN, H. (1940). Z. anorg. Chem. 245, 103.
LEVI, H. & USSING, H. H. (1948). Acta physiol. scand. 16, 232.
LINDERHOLM, H. (1952). Acta physiol. scand. 27.
LUNDEGARDH, H. (1940). LantbrHogsk. Ann. 8, 233.
LUNDEGARDH, H. (1947). Amer. Rev. Biochem. 16, 503.
LUNDEGARDH, H. (1948). Disc. Faraday Soc. B, 139.
LUNDEGARDH, H. (1949). LantbrHogsk. Ann. 16, 339, 372.
MAIZELS, M. (1951). J. Physiol. 112, 59.
NASTUK, W. L. & HODGKIN, A. L. (1950). J. Cell. Cornp. Physiol. 35, 39.
NERNST, W. (1888). Z. phys. Chem. 2, 613.
NERNST, W. (1889). Z. phys. Chem. 4, 129.
NIELSEN, S. O. & ROSENBERG, T. (1951). C.R. Trav. Lab. Carlsberg, 27, no. 19.
PLANCK, M. (1890). Ann. Phys. Chem., N.F., 39, 161 ; 40, 561.
ROSENBERG, T. (1948). Acta chem. scand. 2.
SCHWARZENBACH, G., KAMPiTSCH, E. & STEiNER, R. (1945). Helv. chim. actay 28,
828.
SCHWARZENBACH, G., KAMPITSCH, E. £ STEINER, R. (1946). Helv. chim. acta, 29,
364-
SHANES, A. M. (1946). J. Cell. Comp. Physiol. 17, 57.
STEINBACH, H. B. (1940). J. Biol. Chem. 133, 695.
STEINBACH, H. B. (1950). Amer.J. Physiol. 163, 236.
STEINBACH, H. B. (1951). Amer.J. Physiol. 167, 284.
TEORELL, T. Z. (1951). Elektrochemie, 55, 460.
TEORELL, T. Z. (1952). Progr. Biophys. 18.
USSING, H. H. (1949). Physiol. Rev. 29, 127.
USSING, H. H. (1952). Advanc. Enzymol. 13, 21.
USSING, H. H. & ZERAHN, K. (1951). Acta physiol. scand. 23, no.
VON WAZER & CAMPANELLA (1950). J. Amer. Chem. Soc. 72, 655.
ZWOLINSKI, B. J., EYRING, H. £ REESE, C. E. (1949). J. Phys. Chem. 53, 1426.
CATION ABSORPTION BY NON-GROWING
PLANT CELLS
BY J. F. SUTCLIFFE
Department of Botany, King's College, London
I. INTRODUCTION
The uptake of inorganic solutes by plants clearly involves a complex of
interrelated processes, any one of which may, in suitable circumstances,
limit the overall absorptive capacity of the organism. Amongst these
controlling factors are the rate of utilization and the translocation of ions
from one part of the plant to another. Both of them are intimately related
to growth, and the influence which growth exerts on the course of mineral-
salts absorption may be attributed at least in part to these associated
processes. Helder (1951) has shown that the uptake of nitrate by maize
plants is dependent upon its incorporation in the organic constituents of
the cells. Numerous researches with animal tissues (Kamen & Spiegelman,
1948; Sacks, 1948) have indicated that the mechanism of phosphate
absorption cannot be divorced from the functional importance of phos-
phorus in metabolism, and the same is probably also true for plants.
In order to distinguish utilization from a more fundamental absorption
mechanism, much attention has been paid to the uptake of such ions as
potassium and chloride, which are not appreciably metabolized, but
accumulate within plant cells. Rubidium and bromide ions have often
been favoured as indicators (Steward, Prevot & Harrison, 1942), since they
do not usually occur in plants at all. There is no evidence that the absorp-
tion mechanism for these relatively metabolically inert ions is in any way
fundamentally different from that which is involved in the case of nitrate or
phosphate. On the contrary, it is probable that the uptake of inorganic
solutes always involves a combination with organic cell constituents. But
whereas with nitrate, for example, these become immediately concerned
in metabolic processes leading to the synthesis of protein, in the case of
accumulated ions, the complex is subsequently broken down to release free
ions into the cell vacuoles. In the present study, attention is confined to
the absorption of two cations, potassium and sodium, which accumulate
readily within the tissues under investigation.
In order to reduce the effect, on absorption, of the transport of materials
away from the absorbing region, frequent attempts have been made to
examine ion-uptake, using less complex systems than those which are
326 CATION ABSORPTION BY NON-GROWING PLANT CELLS
presented by the intact angiosperm. Some investigations have been made
upon the coenocytic algae such as Valonia, Halicystis and Nitella spp.,
where the problem of translocation is not involved (Hoagland, Davies &
Hibbard, 1928; Brooks, 1937; Jacques, 1938). Mainly because of the
difficulty of culturing them in the laboratory, these algae, however, are
inconvenient material for most absorption studies, and many research
workers have favoured the use of excised roots as experimental objects
(Lundegardh & Burstrom, 1933; Hoagland & Broyer, 1936; Humphries,
1950). Excised roots are easy to grow in culture under controlled condi-
tions ; they are the natural absorbing organ of the plant, and when isolated
they absorb ions rapidly without the modifying influence of transport into
the shoot.
That the movement of ions away from the region of absorption may still
exert a complicating effect with this material is, however, indicated by the
suggestion of Lundegardh (1949) that mineral salts may be moved across
the cortex of excised wheat roots into the conducting elements of the stele,
and from there, through the cut surface, back into the culture medium.
Investigations with excised roots are further complicated by the inevitable
presence in the organ of many different types of cell with diverse absorptive
capacities. Prevot & Steward (1936) found that there is a pronounced
longitudinal gradient of accumulation along intact roots, and differences of
this kind make it difficult to interpret results based on a net uptake of
solutes at the cell level. This difficulty is enhanced when the relationship
between respiration and absorption is being studied, since the respiratory
activity of different cells in such a complex is also not the same.
Another important type of material which has been extensively used in
ion-absorption studies consists of tissue slices of various storage organs
(Nathanson, 1904; Stiles & Jorgensen, 1915; Steward, 1937; Robertson,
1941). An advantage of these objects is the greater uniformity of cells
comprising the tissue, but the intensive investigations of Steward and his
collaborators, with disks of potato-tuber tissue, have shown that even this
material is not without certain complicating features. Freshly cut slices of
the tissue are not immediately capable of accumulating ions metabolically,
but a capacity to do so develops when the material is suspended for several
days in an aerated solution of mineral salts. During this treatment there is
an increase in the rate of respiration, protein synthesis begins, and the cells
at the surface of the block may show a tendency to divide, forming a layer
of callus (Steward, Berry, Preston & Ramamurti, 1943). A study of the
absorption of ions by disks of different thickness led to the conclusion that
only the cells at the surface of the slices were involved in metabolic
absorption. Further, Steward & Preston (1940, 1941) demonstrated a close
CATION ABSORPTION BY NON-GROWING PLANT CELLS 327
relationship between the rates of ion absorption and protein synthesis
under various conditions.
It would appear, however, that a general application of the hypothesis,
that salt absorption and protein synthesis are directly related, cannot be
upheld. Ulrich (1941) has observed that there were no significant changes
in the amounts of amino- or amido-nitrogen during the accumulation of
ions by excised barley roots. Furthermore, it is by no means certain that,
in normal circumstances, when protein synthesis is proceeding in roots,
the regions of most active absorption and synthesis exactly coincide
(Kramer & Wiebe, 1952).
6-0 r
1-5
Fig. i. Increase in length ( — 0 — ) and K absorbed ( — Q — ) by maize-root sections
placed in 2 % sucrose + 0-0 1 M-KC1 at 25° C. during 72 hr.
Nevertheless, it is evident that the growth of cells is of profound signifi-
cance in relation to absorption, and it is therefore of interest to examine the
extent to which other features associated with growing cells, such as in-
creasing surface area or volume, are important controlling factors in
situations where protein synthesis does not appear to be intimately in-
volved. Jacques (1939) concluded that there is little connexion between
surface area and the rate of uptake of ions by Halicystis, but Burstrom
(1942), on the other hand, has shown that there is a close similarity between
the changes in cell length and the amounts of osmotically active materials
in actively growing epidermal cells of wheat roots.
328 CATION ABSORPTION BY NON-GROWING PLANT CELLS
I (Sutcliffe, 19526) studied the course of potassium uptake by extending
root segments of maize, and found that the rate of accumulation remained
constant during the growth phase, whilst the surface area of the tissue
increased by more than 200% (Fig. i). A similar observation has been
made by Brown & Cartwright (1953). There is no significant increase in the
amount of protein during the growth of these fragments, and it may be
concluded that the rate of absorption of ions in this case is related to the
bulk of the protoplasm, rather than to the surface which it presents to the
external medium. After growth of the segments had ceased, accumulation
gradually stopped, and here the finite volume of the mature tissue may be
the limiting factor.
II. THE DEVELOPMENT OF AN ABSORPTIVE CAPACITY IN
NON-GROWING CELLS OF STORAGE TISSUE
The study of solute accumulation by actively growing cells, even in the
simplest experimental situation, presents such a complicated picture that it
seemed advisable to get more information about the course of absorption
with non-growing cells. Successful investigations of this kind have already
been made with mature animal cells, such as erythrocytes (Solomon, 1952).
Cells of red beet root proved to be a satisfactory plant material for this
study. The tissue was cut into small disks, 0-75 cm. in diameter, and 0-5 or
0-75 mm. in thickness. After cutting, the disks were observed to swell in
water during 24-48 hr., by about 10% of their original volume, and then
there was no significant change for many days. Cell counts showed that
there were no cell divisions during this time, and the amount of protein
synthesis was small. On these grounds it was concluded that the material
consisted of mature cells.
Experiments showed that with disks that were thinner than i mm.
absorption was directly proportional to the total number of cells present,
and hence it can be justifiably claimed that all the cells of the tissue,
irrespective of their position in the block, are involved in accumulation.
By using small disks, it is possible to select uniform groups of cells in the
material, avoiding the anomalous phloem rings occurring in beet, and data
thus obtained may be interpreted at the level of individual cells.
Beet tissue resembles that of potato, inasmuch as the absorption of ions
by freshly cut disks is restricted, and metabolic accumulation may be
stimulated by washing the material in aerated distilled water or a mineral
salts solution (Fig. 2). During this treatment there is a gradual increase in
the rate of respiration (Bennet-Clark & Bexon, 1943; Stiles & Dent, 1947)
resembling the changes that were observed by Steward & Preston (1940)
with potato. In the case of beet, however, it is likely that the changing rate
CATION ABSORPTION BY NON-GROWING PLANT CELLS 329
of respiration is associated with increasing protoplasmic activity, involving
little overall synthesis of protein.
Lundegardh (1940) and Robertson & Turner (1945) have claimed that,
whilst there may be no quantitative relationship between the total respira-
tion of a tissue and ion absorption, accumulation of solutes is closely
related to a cyanide-sensitive respiratory component, which is stimulated
by the presence of mineral salts. An attempt was therefore made to observe
changes in this aspect of respiration during the development of absorptive
capacity in beet disks. It was found that although a solution of O-O2M-KC1
0-1 s r
0-10 -
o
U
0-05 -
Fig. 2. Internal K concentration of beet disks during aeration in distilled water (- - x - -)
and C-02M-KC1 ( ), after a preliminary period in distilled water of a few hours (A),
4 days (Q) and 8 days (®) at 25° C.
had no influence on the rate of respiration of freshly cut tissue, oxygen
absorption was stimulated by salt after the material had been washed in
distilled water for several days. The level of the salt-induced respiration
reached a maximum after about 7 days in water at 25° C., which was also
the time required for the development of maximum absorptive capacity
under the same conditions. Finally, it was shown that the salt-stimulated
component of respiration was cyanide-sensitive.
These observations support the contention of Lundegardh that ion
absorption and cytochrome-mediated respiration are closely related. But
the evidence is not conclusive that cytochrome acts directly as the carrier
for anions, in the way that is suggested by this worker. It is possible that
330 CATION ABSORPTION BY NON-GROWING PLANT CELLS
many, if not all, of the energy-requiring processes in most plant cells are
linked to the cytochrome terminal oxidase system, and salt absorption may
not be especially favoured in this respect (Lemberg & Legge, 1949).
Studies of the effect of various respiratory inhibitors on growth, for
example, by Hackett & Schneiderman (1952), have shown the importance
of cytochrome-mediated respiration in this connexion.
The intensification of metabolic activity during the washing of tissue
slices may in part be attributed to increased oxygen tension, lowered
carbon dioxide concentration, and the higher temperature to which the
cells are exposed after cutting. It is likely that, through the influence of
these factors, new active elements of the respiration machinery are
synthesized, and cytochrome components are probably particularly in-
volved. Parallel changes in the activity of cytochrome upon aeration of
yeast cells were observed by Chin (1950).
Another possible cause of the increased rate of respiration of washed
tissue is the removal of a metabolic inhibitor from the material. Skelding
& Rees (1952) have demonstrated the presence of an inhibitor of ion
absorption in an extract from freshly cut slices of beet, and they claim
that this is gradually removed from the cells during washing, partly by
diffusion into the bathing medium, and partly by metabolic degradation.
Skelding & Rees found that the inhibitor does not affect respiration, and
they have postulated therefore that its influence on ion absorption may be
exerted through a physical effect on the permeability of the cell proto-
plasts.
Using the same technique as that of Skelding & Rees, I obtained a
diffusate from beet tissue which inhibited the absorption of potassium by
washed disks, but, at the same time, strongly stimulated respiration. A part
of the respiratory effect may be attributed to the presence of various organic
acids in the extract. Bennet-Clark & Bexon (1943) caused large stimulations
of respiration by applying expressed sap from beet tissue to the outside
of intact cells, and they concluded that part of the effect was due to the
presence of malic and citric acids. It is unlikely, however, that either of
these substances can be identified as the absorption inhibitor, since Machlis
(1944) has shown that the influence of organic acids on ion accumulation
by barley roots is rather one of stimulation than the reverse.
Until it is possible to separate the inhibitor from other metabolically
active substances in the beet extract, it cannot be established with certainty
whether, or not, it also affects respiration. When this point becomes clear it
will be possible to understand, perhaps, how the inhibitor functions in
preventing accumulation, and its relationship to the increased metabolic
activity following the washing of disks.
CATION ABSORPTION BY NON-GROWING PLANT CELLS 331
The diffusion of various substances from freshly cut tissue which stimu-
late respiration may account for the phenomenon of 'wound respiration*
which has often been observed with storage tissues, as well as with other
plant materials (Steward, 1933; Robertson, Turner & Wilkins, 1947). The
absence of this temporary stimulation of respiration, imposed on the
gradual increase in metabolic activity, in our experiments, in those of
Bennet-Clark & Bexon, and of Steward & Preston (1940) may perhaps be
related to a more rigorous washing procedure in these cases, which pre-
vented the accumulation of respiratory active substances in the medium.
III. THE EFFECTS OF PRETREATMENT ON THE
ABSORPTIVE CAPACITY OF DISKS
There is now a considerable amount of evidence that an adsorptive phase is
closely associated with the overall accumulation mechanism for cations
(Lundegardh, 1946; Overstreet & Jacobson, 1946). An investigation was
therefore made of the effect of washing on the adsorptive capacity of beet
disks, in an attempt to correlate it with the metabolic changes and the
increasing rate of accumulation that have been described above. The
adsorptive capacity of the material was determined by taking batches
of disks which had been washed in distilled water for various periods of
time, and placing them in a solution of O-O2M-KC1 for 4 hr. at 7° C. At
the end of the experimental period, an analysis of the medium was made,
and the non-metabolic uptake of potassium was determined. The data
obtained (column A of Table i) show that, when the disks were transferred,
immediately after cutting, to the experimental conditions, there was a loss
of ions from the tissue to the external medium. After washing the material
for 2 days in aerated distilled water, however, there was an immediate
uptake of salt on transference to KC1. The amount of this non-metabolic
absorption was increased with a longer period of pretreatment up to 6-8
days.
A comparison of the amounts of potassium adsorbed and the amounts
lost from the tissue during washing (column C of Table i) shows that,
although a part of the adsorbed potassium may be replacing that which is
leached from the cells, this cannot be the only factor involved since the
adsorptive capacity of the disks continues to increase after the material
has been washed in water for several days, when the total potassium
content is no longer decreasing. Ions which are taken up by the disks
may be located in intercellular spaces, cell walls, protoplasts, or vacuoles.
Since the volume of the intercellular spaces and cell walls probably does
not increase after about 2 days, it may be concluded that the increased
adsorptive capacity of the cells is associated with an increased affinity of
332 CATION ABSORPTION BY NON-GROWING PLANT CELLS
the protoplasm for ions, or else to a greater fixation of potassium in the
vacuoles. It is likely that non-metabolic adsorption occurs in the proto-
plasts rather than in cell vacuoles, since Brooks (1937) and Hoagland &
Broyer (1942) with Nitella have observed that the movement of ions
into the protoplasm is rapid, whilst subsequent entry into the vacuole
takes place slowly and, under anaerobic conditions, perhaps not at all.
Table i. Amounts of potassium adsorbed in 4 hr. at 7° C. (A), amounts of
potassium subsequently exchanged in 6 hr. at 7° C. (B), and the total
amounts of potassium in beet tissue. (C), after washing in aerated distilled
water at 25° C. for various periods of time
Days of
washing
o
2
4
6
8
Amounts of K (/^g./g. fresh wt.)
A
— 50 ±20
i85±38
4i5±39
6Q9±5i
73i ±46
B
645 ± 43
58?±38
455 ±29
405 ±32
4i3±30
C
23i5±i75
19501163
1895 ± 122
I9O5 ±148
1875 ±143
Further evidence in this connexion has been obtained by studying the
non-metabolic exchange of potassium by beet disks with 42K. The tissue,
after being washed for various lengths of time, was allowed to adsorb
potassium for 4 hr. at 7° C. as described above. It was then transferred to
a O-O2M-KC1 solution containing the isotope for 6 hr. at 7° C., after which
the radioactivity of the medium was determined, and the amount of
potassium exchanged with the material was calculated. The results of this
experiment, shown in column B of Table i, indicate that the amount of
potassium exchangeable in 6 hr. under the conditions of the investigation
decreased during the washing of the tissue. After about 4 days of pre-
treatment on transferring the material to KC1 only a part of the potassium
adsorbed, exchanged readily, and about %oo/ig. out of more than joo/ig.
adsorbed per g. fresh weight of well-washed tissue, did not exchange.
The exchange data indicate that the adsorbed ions exist in at least two
different states outside the vacuoles. Those which are easily exchanged may
be present in the intercellular spaces, cell walls, or loosely bound in the
protoplasm, whilst the rest are probably located entirely in the protoplasts.
Jacobson & Overstreet (1947), from their study of isotopic exchange in
roots, were led to the conclusion that cations may be bound into plant
cells with varying degrees of non-exchangeability. They believe that the
binding of cations in the form of chelated complexes with proteins, amino-
acids, or organic acids, may be of particular importance in relation to
accumulation. Ion bonds formed in these combinations are relatively
CATION ABSORPTION BY NON-GROWING PLANT CELLS 333
strong, and ions so held are exchanged only with difficulty. The adsorbed
potassium may become increasingly bound in such forms following the
pretreatment of disks, and if these combinations are important in relation
to the accumulation mechanism as a whole, then the increased capacity
of the tissue to absorb after washing may be accounted for.
The relatively large amount of the adsorbed potassium, which becomes
bound in the non-exchangeable form in washed tissue, suggests that the
adsorptive centres involved are located throughout the protoplasm, rather
than only at the outer surface of the cells. Suggestions that protoplasts as
a whole contain negatively charged immobile anions, to which cations may
be fixed, have already been made by Blinks (1940) and by Robertson
(1951). It is not possible to make a calculation of the * apparent free-space'
into which ions move non-metabolically from the present experiments as,
for example, Hope & Stevens (1952) have done for bean roots, since the
concentration of ions which is attained in this region cannot be deter-
mined. Nevertheless, it seems unlikely that the surface of the protoplast
is the sole region of adsorption as Lundegardh (1940) suggested.
It appears to be more likely from the present data that there is a barrier
to the free diffusion of ions, and to exchange processes in the vicinity of the
tonoplast, separating the protoplast from the vacuole of a plant cell.
Arisz (1945) has concluded, from his studies of ion absorption by leaves,
that the tonoplast is the region of a cell across which active accumulation
occurs. There is some evidence from isotopic investigations with algal
cells (Hoagland, 1944) that the concentration of ions in the vacuoles may be
greater than in the protoplasm, and this again supports the idea that ions
do not move readily by diffusion across tonoplasts. This hypothesis is also
in agreement with the observation of Bennet-Clark & Bexon (1943) that
metabolically important substances, such as organic acids, may be re-
latively inactive when they are confined to cell vacuoles.
During the washing of freshly cut disks there is clearly an activation of
the metabolic mechanism by means of which ions are able to traverse this
physical barrier against an activity gradient. This is associated with an
increased capacity of the cells to adsorb ions strongly, and it may be
postulated that the formation of complexes with particular cell constituents
forms an essential intermediate link in the movement of solutes from an
external medium into vacuoles. Metabolic energy is probably involved in
the preliminary synthesis of the carrier molecules which occurs during the
washing treatment. In this way the importance of aeration and temper-
ature in determining the rate at which the absorptive capacity of the tissue
is developed, and the effect of respiratory inhibitors in retarding this
process can be explained.
334 CATION ABSORPTION BY NON-GROWING PLANT CELLS
The fact that potassium appears to combine with the carrier at low
temperature in the present experiments indicates that metabolic energy is
not involved at this stage. But it may be required during a series of re-
actions leading to the subsequent breakdown of the complex, and the
release of free ions into the vacuoles. The spatial separation of ion ad-
sorption on to the carrier, and their release, which is responsible for actual
transport may be accomplished by an aggregation of the carrier substances
in microscopic particles, such as mitochondria, which undergo random
movements in the protoplasm (Robertson, 1951), from a region where the
ions are taken up to one where the breakdown of the complex occurs; or
the contraction of protein molecules, as has been suggested by Goldacre
(1952), may be involved.
IV. THE COURSE OF ION ABSORPTION WITH
WELL-WASHED TISSUE
When beet disks have been washed in aerated distilled water at 25° C. for
about 7 days, a maximum absorptive capacity is attained. If, at the end of
this time, the material is transferred to a solution of mineral salts, ab-
sorption occurs at a rate which is dependent upon various external factors,
such as the nature of the ions involved, and their concentration ; the presence
of other ions, or various substances which may influence metabolism;
temperature, and aeration. The effects of these factors on absorption are
mostly well understood, and will not be further discussed here.
There are, however, in addition, various internal influences which are
effective in controlling the rate of absorption. One of these is clearly the
capacity of the accumulation machinery, already considered above as a
limiting factor during washing, and another is the internal concentration of
ions. A reduction in the rate of uptake of solutes, when the mineral-salts
content of the material is high, has been observed by a number of workers,
including Hoagland & Broyer (1936) with barley roots, Jacques (1938) with
Valonia, and Alberda (1948) working with whole maize plants. The same
effect with beet disks is shown in Fig. 2.
Although the phenomenon is well established, the mechanism of it is not
yet clearly understood. Broyer (1951) wrote: 'If roots (however) have
accumulated inorganic solutes in the past, under favourable environmental
conditions, approaching upper limits imposed by their hereditary potenti-
alities, they may be restricted under such circumstances from further
accumulation. Such high-salt roots are close to their dynamic equilibrium
relative to inorganic solutes/
The establishment of an equilibrium of the type visualized by Broyer is
one of the characteristics of ion accumulation by non-growing cells.
CATION ABSORPTION BY NON-GROWING PLANT CELLS 335
Growing tissues are capable of absorbing ions indefinitely, and this may be
due in part to the increasing volume of the material which prevents the
internal concentration from attaining its limiting value. Utilization and
translocation are also obviously important, in this connexion, in situations
where they are operative ; and protein synthesis, in so far as it results in the
production of new units of the absorption machinery, must also have
a modifying influence.
In studying the effect of internal concentration on the uptake of ions by
non-growing cells, various alternative hypotheses may be considered. It is
possible, for example, that the net accumulation of solutes is the resultant
of an absorption and a leakage process, which are mutually opposed
(Krogh, 1946). Assuming that uptake is metabolically controlled, and that
leakage occurs passively by diffusion along a concentration gradient, as
more ions are accumulated the rate of outward movement will increase. If
uptake remains constant, then the rate of accumulation will be reduced,
and an equilibrium will be established when the opposing processes are
equal. Evidence that the absorption mechanism itself may be relatively
unaffected by the increasing ion content of the material is provided by the
observation (Sutcliffe, 19520) that the level of the cyanide-sensitive respira-
tion in a cell which has become saturated with potassium is as high as that
of one which is still absorbing ions rapidly. If the energy of salt respiration
is, in fact, involved in accumulation, then it is clear that this energy is still
available after net uptake has ceased, and it may be postulated that ions are
still moving into the tissue, but are being balanced by a contrary move-
ment in the opposite direction.
There are, however, a number of reasons why this attractive hypothesis
must be discarded in the present case. It has been observed, for example,
that variation of the temperature between 15 and 30° C. does not alter the
equilibrium position with beet disks, although it does of course profoundly
affect the time which is required for the establishment of it. Even though
Danielli (1952) has explained that diffusive processes through cell mem-
branes may have a high Qw, it seems unlikely that a metabolic absorptive
process and a passive leakage will be influenced simultaneously to the same
extent by temperature. Yet this must be the situation if the equilibrium is
established in the way outlined above.
Moreover, if there is an appreciable passive leakage of inorganic ions from
intact cells and tissues, it ought to be possible to detect this by observing
changes in the composition of the external medium, when the material is
placed under conditions which prevent metabolic reabsorption. Actually,
Hoagland & Broyer (1942) found that the rate of outward diffusion of ions
from excised barley roots, when oxygen is withheld, was very low. A similar
336 CATION ABSORPTION BY NON-GROWING PLANT CELLS
observation was made with beet disks when the accumulation mechanism
was inhibited by KCN (Sutcliffe, 1952 a).
Although the evidence appears to be conclusive that there is no signifi-
cant passive leakage of ions by diffusion from healthy plant cells, the
possibility cannot be excluded that solutes are being transported outwards
from vacuoles by a process which depends either directly or indirectly on
respiratory energy, and cannot therefore be detected when metabolism is
inhibited. That root cells, in some circumstances, will allow the transport
of previously accumulated ions away into the shoot was demonstrated by
Steward et al. (1942), and the active movement of cations in both directions
through the membranes of erythrocytes is well established (Solomon,
1952).
This possibility was examined, with beet cells under the present
experimental conditions, by observing the rate of exudation from disks at
25° C. of previously accumulated 42K. Table 2 shows the changes which
were observed in the radioactivity of the external medium, when disks,
which had been equilibrated with O-O2M-KC1 containing 42K, were trans-
ferred either to distilled water or to an inactive KC1 solution at 25° C.
during 6 and 24 hr. Corrections have been made for the decay of the
isotope during the experimental period.
Table 2. Radioactivity (counts per min.) of the external medium, after 6 and
24 hr. when disks, which had previously been allowed to accumulate 42K,
were transferred to either distilled water or O-O2M-KC1, at 25° C.
Time
(hr.)
Total initial activity
of the material
Activity of the medium
H20
0-02M-KC1
o
6
24
2460+ 115
32±4
27±3
105 ± 16
ii9± n
The data show that a negligible amount of the previously accumulated
42K left the tissue, when the disks were suspended in distilled water, even
during 24 hr., thus confirming a similar observation of Jenny & Overstreet
(1939) with excised barley roots. A greater increase in radioactivity was de-
tected when KC1 was present in the external medium, indicating that a
certain amount of exchange occurred, but the extent of this was clearly
limited since the amount exchanged after 24 hr. was not significantly greater
than after 6 hr. In order that isotopic equilibrium between the tissue and
the medium should be established, it may be calculated that about 43 % of
the 42K originally present in the disks should have passed into the external
solution. The fact that only about 5 % was exchanged indicates that, if there
CATION ABSORPTION BY NON-GROWING PLANT CELLS 337
was any metabolic transport of ions from the material, it was proceeding
extremely slowly. It is probable that all the 42K which appeared in the
medium was derived from the cell protoplasts, and that vacuolar ions were
not significantly involved.
This conclusion was confirmed by a study of the exchange of 42K when
disks, which had been allowed to equilibrate with inactive KC1 solution,
were transferred to a medium containing the isotope. The experiment
showed that between 400 and 500 /^g. of potassium, out of a total of more
than 5Ooo//g./g. fresh weight of the material, was readily exchanged within
6 hr. at either a high or low temperature. The amount exchanged in this
experiment was thus about the same as that which exchanged in the washed
tissue capable of absorbing ions rapidly, although the total potassium
content of the disks in the present case was more than twice as great
(Table i). It is evident therefore that cell protoplasts may contain only
a limited amount of easily exchanged potassium, and this is independent of
the total cation content of the material.
So far attention has been confined to the absorption of potassium, but
the study has also been extended to include sodium. These two ions
together comprise about 90% of the total inorganic cations in the material
which we have examined, and of the two there is a considerable excess of
potassium. In one case, which may be regarded as typical, there were
present in the disks, immediately after cutting, about 2-25 mg. of potassium,
and 0-55 mg. of sodium/g. fresh weight of tissue. When washed beet tissue
was placed in a mixture of KC1 and NaCl containing equal quantities
of each, the two cations were absorbed at approximately equal rates, and
appeared to compete with one another on more or less equal terms, for the
use of the same accumulation mechanism. Absorption eventually ceased
when the total concentration of both ions together was about the same as
that which was attained when absorption occurred from a solution of
either KC1 or NaCl alone (Table 3).
When disks, which had been allowed to equilibrate with o*O2M-NaCl,
were transferred to a solution of O-O2M-KC1, a limited uptake of potassium
Table 3. Potassium and sodium contents (mg./g. fresh weight) of disks allowed
to equilibrate with (a) O-O2M-KC1, (b) o*O2M-NaCl, (c) o-oiM-KCl
-f o-oiM-NaCl, at 25° C. for 7 days
Medium
0-02M-KC1
o-02M-NaCl
o-oi M-KCl + o-oi M-NaCl
Cation content
K.
5'75 ±0-38
i'53±o-o8
4-15—0-27
Na
0'53 ±0-02
5-13+0-32
2'38±o-i8
K-Na
6-28
6-60
6-53
338 CATION ABSORPTION BY NON-GROWING PLANT CELLS
was observed within 24 hr., which was accompanied by a corresponding
output of sodium. This may be interpreted as representing the replacement
of some sodium in cell protoplasts by potassium. Subsequently there was
no significant change in either the potassium, or the sodium content of the
disks, although the experiment was prolonged for a further 7 days (Fig. 3).
When the disks were treated first with KC1 and then placed in NaCl, the
Disks transferred
Fig. 3. Amounts of K (0), Na (Q) and Na + K (A) in beet disks placed in O-O2M-KC1
( ) or o-o2M-NaCl ( ) for 9 days at 25° C., and then transferred to the alternative
medium for a further 8 days at 25° C.
effects observed were substantially the same. In this case, however,
following the preliminary exchange, the sodium content of the tissue con-
tinued to increase throughout the experimental period. This was probably
due to the fact that the material was not completely equilibrated at the time
of transfer, since the total cation content of the disks also increased. The
experiment as a whole confirms the view that most of the cations in beet
cells, whether potassium or sodium is considered, do not exchange readily
over relatively long periods of time, and that the movement of ions between
the external medium and cell vacuoles is probably very slow.
CATION ABSORPTION BY NON-GROWING PLANT CELLS 339
It is also evident from the above experiment, that in the presence of
a high internal concentration of either potassium or sodium, the absorption
of the alternative ion is retarded. This is contrary to the observations of
Humphries (1950, 1951, 1952) that the absorption of potassium by excised
root systems is related to the amount of this ion which is present in the
tissue, and is independent of any other. Humphries grew potassium-
deficient plants by replacing this ion in the culture medium by sodium,
and found that the excised roots of these plants subsequently absorbed
more potassium than the controls. This discrepancy may be resolved, if
excised roots, unlike beet disks, selectively absorb potassium rather than
sodium. It would be interesting to know the sodium content of these
potassium-deficient plants to establish this point.
Since neither a passive leakage nor a metabolic leakage of ions from cells
can adequately account for the equilibrium which is established with beet
tissue, an alternative hypothesis may be proposed. Hoagland & Broyer
(1936) suggested that the accumulation of ions in high-salt barley roots
might eventually cease, because of the progressive saturation of proto-
plasmic constituents. This theory has been developed by Overstreet,
Jacobson & Handley (1952) as a part of their theory of the mechanism of
ion absorption in roots. They have proposed that uptake depends on the
presence in the protoplasm of metabolically produced carrier substances,
designated HR, which combine with the ions, e.g. M+, and then subse-
quently break down again according to the reversible reaction
In this reaction, one of the rate-determining factors is the concentration
of the complex MR. It is suggested that the level of MR may be higher in
high-salt than in low-salt tissues, so that a smaller amount of HR is available
to combine with more M+ from the external medium, and the rate of
absorption is reduced.
In extending this hypothesis somewhat, it may be proposed that the
effect of a high internal salt concentration is to prevent the breakdown of
the ion-carrier complex at the inner surface of the tonoplast of the plant
cell, and this may perhaps be associated with the increased activity
gradient across which the ions are being moved. It is possible that when
ions have been adsorbed on the carrier, this complex is metabolized to a
high-energy form, which can then be decomposed to release the ions, as
long as the free energy of the whole system is thereby reduced. When
equilibrium is attained in the cell, owing to the increased free energy of
potassium ions in the vacuole, this condition may be no longer satisfied, so
that the ion and its carrier remain in the combined form. No carrier is
340 CATION ABSORPTION BY NON-GROWING PLANT CELLS
therefore available to accept ions from the medium, and accumulation
stops.
In this study, therefore, the failure both of freshly cut disks and of salt-
saturated tissue to accumulate ions has been considered, and in each case it
is possible that this may be related to the lack of available carrier molecules.
With freshly cut tissue, the carrier is either absent from the protoplasts,
or else it is inactivated by an unknown mechanism, whilst with cells
which have already absorbed ions to their maximum capacity, the carrier
is probably present entirely in the combined form. In either of the two
situations, absorption is prevented.
V. THE INTERRELATIONSHIP BETWEEN ABSORPTION
OF ANIONS AND CATIONS
So far it has been assumed that a metabolic mechanism is directly involved
in the absorption of cations, but in the accumulation of neutral salts the
uptake of one ion cannot be divorced from that of its associate. The active
absorption of cations automatically creates an electrical gradient along
which a passive movement of anions may occur, and vice versa. With many
types of animal cell it appears that cations are metabolically transported,
whilst the anions follow passively. On the other hand, Lundegardh (1940)
has claimed for plants that anions are actively absorbed, and cations move
in along the electrical gradient so created.
The evidence presented above shows that the protoplast of a plant cell
presents a very considerable barrier to the free diffusion or exchange of
cations, indicating that the presence of an electrical gradient alone may
not be sufficient to account for the absorption of K and Na at the rates
which have been observed. It seems probable therefore that a specific
accumulation mechanism for cations exists in these cells. The inquiry may
then be made whether in fact the anions are absorbed passively and the
scheme of Lundegardh be reversed. Experiments with 22Br along the lines
of those outlined above for 42K have shown, however, that there is a similar
lack of exchangeability between vacuolar anions and the medium, so that
for the present it must be concluded in agreement with Hoagland &
Steward (1939) that both anions and cations are accumulated directly by
a metabolically controlled mechanism. Whether it is the same or a different
carrier system which is involved in the two cases must remain the subject
of further investigation.
VI. SUMMARY
The complexity of the experimental situation with growing cells and
tissues is indicated, and the importance of attempting to study ion-
absorption in non-growing cells is emphasized. It is proposed that the
CATION ABSORPTION BY NON-GROWING PLANT CELLS 341
development of a capacity to absorb solutes in beet disks may be related to
the metabolic synthesis, or activation of carrier substances in the tissue.
Evidence is presented that combination between the ion and its carrier
occurs in the protoplasm by a non-metabolic mechanism, and that active
transport takes place across the tonoplasts. It is probable that the physical
resistance of these membranes to diffusive processes is very high, and that
exchanges of ions between cell vacuoles and the external medium are
prevented. The establishment of an equilibrium, when accumulation
ceases in non-growing cells, may be attributed to a failure of the ion-carrier
complex to break down when the activity gradient across which the ions
are moving becomes too great.
REFERENCES
ALBERDA, T. (1948). Rec. Trav. hot. merl. 41, 541.
ARISZ, W. H. (1945). Proc. K. Akad. Wet. Amst. 48, 420.
BENNET-CLARK, T. A. & BEXON, D. (1943). New Phytol. 42, 65.
BLINKS, L. R. (1940). Cold Spr. Harb. Symp. Quant. Biol. 8, 204.
BROOKS, S. C. (1937). Trans. Faraday Soc. 33, 1002.
BROWN, R. £ CARTWRIGHT, P. M. (1953). J. Exp. Bot. 4, 197.
BROYER, T. C. (1951). In Mineral Nutrition of Plants, p. 217, ed. Truog, E.
University of Wisconsin Press.
BURSTROM, H. (1942). Ann. Landw. Hochsch. Schzv. 10, i.
CHIN, C. H. (1950). Nature, Lond., 165, 926.
DANIELLI, J. F. (1952). Symp. Soc. Exp. Biol. 6, i.
GOLDACRE, R. J. (1952). Int. Rev. Cytol. i, 135.
HACKETT, D. P. & SCHNEIDERMAN, H. A. (1952). Abstr. Meeting Amer. Soc.
Plant Physiol. Ithaca.
HELDER, R. J. (1951). Proc. K. Akad. Wet. Amst. ser. C, 54, 275.
HOAGLAND, D. R. (1944). Lectures on the Inorganic Nutrition of Plants, p. 52.
Waltham, Mass. : Chron. Bot. Co.
HOAGLAND, D. R. & BROYER, T. C. (1936). Plant Physiol. n, 471.
HOAGLAND, D. R. & BROYER, T. C. (1942). J. Gen. Physiol. 25, 865.
HOAGLAND, D. R., DAVIES, A. R. & HIBBARD, P. L. (1928). Plant Physiol. 3, 473.
HOAGLAND, D. R. & STEWARD, F. C. (1939). Nature, Lond., 143, 1031.
HOPE, A. B. & STEVENS, P. G. (1952). Aust.J. Sci. Res. B, 5, 335.
HUMPHRIES, E. C. (1950). J. Exp. Bot. i, 282.
HUMPHRIES, E. C. (1951). J. Exp. Bot. 2, 344.
HUMPHRIES, E. C. (1952). J. Exp. Bot. 3, 291.
JACOBSON, L. & OVERSTREET, R. (1947). Amer. J. Bot. 34, 415.
JACQUES, A. G. (1938). J. Gen. Physiol. 22, 147.
JACQUES, A. G. (1939). J. Gen. Physiol. 22, 757.
JENNY, H. & OVERSTREET, R. (1939). Soil Sci. 47, 257.
KAMEN, M. D. & SPIEGELMAN, S. (1948). Cold Spr. Harb. Symp. Quant. Biol.
13, 151-
KRAMER, P. J. & WIEBE, H. H. (1952). Plant Physiol. 27, 661.
KROGH, A. (1946). Proc. Roy. Soc. B, 133, 140.
LEMBERG, R. & LEGGE, J. W. (1949). Hematin Compounds and Bile Pigments.
New York: Interscience Publishers.
LUNDEGARDH, H. (1940). LantbrHogsk. Ann. 8, 234.
342 CATION ABSORPTION BY NON-GROWING PLANT CELLS
LUNDEGARDH, H. (1946). Nature, Lond., 157, 575.
LUNDEGARDH, H. (1949). LantbrHogsk. Ann. 16, 339.
LUNDEGARDH, H. & BURSTROM, H. (1933). Biochem. Z. 261, 235.
MACHLIS, L. (1944). Amer.J. Bot. 31, 183.
NATHANSON, A. (1904). Jb. wiss. Bot. 39, 607.
OVERSTREET, R. & JACOBSON, L. (1946). Amer.J. Bot. 33, 107.
OVERSTREET, R., JACOBSON, L. & HANDLEY, R. (1952). Plant Physiol. 27, 583.
PREVOT, P. & STEWARD, F. C. (1936). Plant Physiol. n, 509.
ROBERTSON, R. N. (1941). Aust.J. Exp. Biol. Med. Set. 19, 265.
ROBERTSON, R. N. (1951). Ann. Rev. PI. Physiol. 2, i.
ROBERTSON, R. N. & TURNER, J. S. (1945)- Aust. J. Exp. Biol. Med. Sci. 23, 63.
ROBERTSON, R. N., TURNER, J. S. & WILKINS, M. J. (1947). Aust. J. Exp. Biol.
Med. Sci. 25, i.
SACKS, J. (1948). Cold Spr. Harb. Symp. Quant. Biol 13, 180.
SKELDING, A. D. & REES, W. J. (1952). Ann. Bot., Lond., N.S., 16, 513.
SOLOMON, A. K. (1952). J. Gen. Physiol. 36, 57.
STEWARD, F. C. (1933). Protoplasma, 18, 208.
STEWARD, F. C. (1937). Trans. Faraday Soc. 33, 1006.
STEWARD, F. C., BERRY, W. J., PRESTON, C. & RAMAMURTI, T. K. (1943). Ann.
Bot., Lond., N.S., 7, 221.
STEWARD, F. C. & PRESTON, C. (1940). Plant Physiol. 15, 23.
STEWARD, F. C. & PRESTON, C. (1941). Plant Physiol. 16, 481.
STEWARD, F. C., PREVOT, P. & HARRISON, J. A. (1942). Plant Physiol. 17, 411.
STILES, W. & DENT, K. W. (1947). Ann. Bot., Lond., N.S., u, i.
STILES, W. & JORGENSEN, I. (1915). Ann. Bot., Lond,, 29, 6n.
SUTCLIFFE, J. F. (19520). J. Exp. Bot. 3, 5Q.
SUTCLIFFE, J. F. (19526). Unpublished data.
ULRICH, A. (1941). Amer.J. Bot. 28, «>26.
THE RELATIONSHIP BETWEEN METABOLISM
AND THE ACCUMULATION
OF IONS BY PLANTS
BY R. SCOTT RUSSELL
Department of Agriculture, University of Oxford
I. INTRODUCTION
A discussion of the probable nature of the mechanism whereby ions are
accumulated in plant tissues must turn largely on two fundamental
questions which are still matters of controversy, namely :
(1) Is the cytochrome-cytochrome oxidase system the only mechanism
able to make energy available for the active accumulation of electrolytes in
plant tissues or do other terminal oxidases share this characteristic?
(2) Is the accumulation of ions directly mediated by the electron transfer
in respiration or does a product of respiration function as an ion-carrier?
The investigation of the former subject to be described in this paper was
undertaken jointly with Dr W. O. James of the Department of Botany,
Oxford University. The remaining experiments were carried out in the
Department of Agriculture, Oxford University, in collaboration with
Dr R. P. Martin, Miss Joyce Ayland and others.
It is a pleasure to acknowledge my indebtedness to those who have
collaborated in this work and also to Dr J. L. Harley and Dr J. F. Sutcliffe
who have both made unpublished data available to me.
II. THE ABILITY OF TERMINAL OXIDASE SYSTEMS OTHER
THAN CYTOCHROME TO MEDIATE ACTIVE
ACCUMULATION
Lundegardh (1945, 1950) and Robertson & Wilkins (1948) have concluded
that the active accumulation of ions is dependent upon respiration mediated
by cytochrome oxidase. Their results suggest that in wheat roots and carrot
slices other terminal oxidases may be without effect in this respect.
However, as cytochrome appears to be a principal path of respiration in the
tissues they investigated, it would seem rash to conclude that the same result
would be obtained in tissues possessing different terminal oxidase systems.
Thus it was decided to examine the relationship between respiration and
salt uptake in material in which a large fraction of the normal respiration is
mediated by systems other than cytochrome. The roots of young barley
plants were suitable for this investigation, since, at certain stages of their
344
THE RELATIONSHIP BETWEEN METABOLISM AND
development, ascorbic acid oxidase is the principal terminal oxidase (James,
1953). Grain of the variety Spratt Archer was germinated on moist cotton
gauze supported above distilled water at room temperature. Apical 10 mm.
segments of root were detached for experimental treatment after varying
periods of growth. Respiration was determined by the standard Warburg
procedure, and the absorption of bromide and rubidium was measured
by tracer methods, both the root apices and the external solution being
assayed. The duration of each experiment was 3 hr. Diethyldithiocarb-
amate, which chelates copper (Albert & Gledhill, 1947), was employed
as a respiratory inhibitor. James & Garton (1952) have shown that at
the concentration of 2 x io~4M this substance inhibits the ascorbic acid
100 r
J50
0L
Age of roots (days)
Fig. i. Effects of 2 x io~4M-diethyldithiocarbamate in the apical 10 mm. of barley roots.
Curve — respiration. Histograms: unshaded — absorption of rubidium ; shaded — absorption
of bromide.
oxidase of barley to an extent of over 85 %; by contrast the inhibition of
cytochrome oxidase is less than 10%. It was found (James, unpublished)
that the extent to which the respiration of barley root apices is affected
by 2 x io~4M-diethyldithiocarbamate varies with the age of the tissues
(Fig. i). Three days after germination this inhibitor reduced respiration
by approximately 35%; thereafter the effect increased until in the apices
of roots 7 days old respiration was inhibited to over 60%. Concurrently
with the increasing sensitivity of the tissues to diethyldithiocarbamate,
the extent of light-reversible inhibition by carbon monoxide fell. The
respiration of recently germinated embryos was inhibited to the extent of
80% by carbon monoxide, whereas when roots had reached the age of
7 days virtually no inhibition occurred. The complexity of the pattern of
respiration in this tissue is further indicated by that fact that, in root
THE ACCUMULATION OF IONS BY PLANTS
345
apices exceeding 7 days in age, the effect of diethyldithiocarbamate
declined.
A general discussion of the balance of terminal oxidase aystems is beyond
the scope of the present inquiry, and consideration will be here confined to
the relationship between respiration and salt uptake in roots 5 and 7 days
old. The effect of 2 x io~4M-diethyldithiocarbamate on the absorption of
bromide and rubidium from 0-002 M-solutions is shown by the histograms
superimposed on Fig. i . It is apparent that the effects of this inhibitor on
respiration and on salt absorption were similar. The inhibition of the
absorption of both ions rose from between 50 and 60% at the fifth day to
over 70% two days later. Since the extent to which 2x io~4M-diethyl-
dithiocarbamate inhibits salt uptake greatly exceeds the extent to which it
inhibits cytochrome oxidase, the observed effect on salt uptake cannot be
attributed to the inhibition of the cytochrome system. Further evidence in
this direction is provided by results of James's experiments with carbon
monoxide which indicate that cytochrome oxidase is responsible for
a negligible part, if any, of the respiration of roots 7 days old. A clear
relationship between the inhibition of ascorbic acid oxidase and the in-
hibition of salt uptake is instead indicated.
Table i. The effect of diethyldithiocarbamate on the absorption of
rubidium and bromide by barley plants at the second leaf stage
Percentage
Ion
Concentration
of diethyldithio-
reduction of
absorption
5 % fiducial
limits
carbamate (molar)
induced by
inhibitor
Rb, io~3M
2 X I0~4
69-9
±2-9
Br, 8 x 10 "3M
2 X I0~4
67-2
+ 0-8
10 3
73'7
±i'8
Additional evidence that respiration through the ascorbic acid oxidase
system can provide energy which is utilized in salt absorption was obtained
in experiments in which intact barley plants at the second leaf stage were
treated with 2X io~4M-diethyldithiocarbamate (Table i). The absorption
of both rubidium and bromide was inhibited to between 65 and 70 %. When
the concentration of diethyldithiocarbamate was increased to io~3M, salt
absorption was decreased to a small but significant extent. James &
Garton (1952) have shown that this concentration completely inhibits
ascorbic acid oxidase and reduces the activity of cytochrome oxidase by
upwards of 25 %. It is possible therefore that a small fraction of the salt
absorption of the plants was mediated by cytochrome. Some further
evidence of the relationship between ascorbic acid oxidase and salt
346 THE RELATIONSHIP BETWEEN METABOLISM AND
absorption is provided by an experiment to which reference is made later
(Fig. 4), in which xo^M-diethyldithiocarbamate inhibited the absorption
of phosphate by barley plants to the extent of over 30 % when the external
concentration of phosphate was 10 p.p.m.
Having shown that respiration mediated by ascorbic acid oxidase can
bring about the active accumulation of salts, it is natural to inquire if this
capacity is common to all terminal oxidases, or whether metallo-enzymes
alone possess it. In this connexion interest attaches to the recent investi-
gations of Harley and his associates (1953) on the nutrient absorption and
metabolism of beech mycorrhizal roots. In fresh excised mycorrhizas
(Table 2) it was found that cyanide, azide, fluoride iodoacetate and arsenate
had no inhibiting effect on oxygen uptake, though they inhibited phosphate
uptake to a marked extent. Malonate at the concentration used had no
significant effect on respiration or absorption. The respiration rate was
unchanged by the addition of salts to the external medium. Tissues which
had been stored in distilled water gave contrasting results; a marked 'salt
respiration* was induced by 64 x io~3M-KH2PO4 (Table 3). Cyanide,
surprisingly, also stimulated respiration, and when cyanide and phosphate
were applied together a still greater increase was observed. Phosphate
absorption was reduced by cyanide to the extent of approximately 50%,
a comparable degree of inhibition to that in fresh roots.
Table 2. Effect of inhibitors on the absorption of oxygen and phosphate
by fresh, excised beech mycorrhizas at pH 5-5
Harley, McCready & Brierley (1953).
' % of control
02
P
Cyanide io~3M !
1 06
49
Azide 2 x io~6M
96
50
Iodoacetate io~3M i
100
48
Fluoride
4 x io~2M
97
35
Malonate
5 x io~2M
90
100
Arsenate
10 3M
99
39
Table 3. Effect of 5 x io~3 M-KCN and 64 x io~3 M-KH2PO4 on the
oxygen uptake of beech mycorrhizas after storage in distilled water
(Data from Harley, McCready & Brierley, 1953.)
Treatment
% of c
control
Exp. (i)
Exp. (ii)
KCN
KH2P04
KCN + KH2P04
164
170
192
190
179
241
THE ACCUMULATION OF IONS BY PLANTS 347
No detailed interpretation of these puzzling results appears at present
possible; it can only be concluded that there is no simple relationship
between the activity of metallo-terminal oxidases and the accumulation
of phosphate in this tissue. This being so it is probable that respiration
through a flavo-protein system was associated with salt uptake.
III. THE LINK BETWEEN RESPIRATION
AND SALT ACCUMULATION
In view of the recent reviews by Broyer (1951) and Overstreet & Jacobsen
(1952) it is unnecessary to discuss in general terms the theories which have
been advanced to explain the active accumulation of salts. It is widely
agreed that ionic exchange processes play an important part in the initial
entry of ions into the cytoplasm. But the conflict of opinion is sharp with
regard to the subsequent steps in the transfer of ions across the cytoplasm
to the vacuole or the vascular stele. The 'anion respiration' concept
developed by Lundegardh and by Robertson and his associates is well
known. The distinctive features of this postulate are that the movement of
anions across the cytoplasm is directly mediated by the electron transfer
through the cytochrome-cytochrome oxidase system, while cations move
by diffusion. The alternative mechanism most favoured in recent years
postulates a product of respiration or ' carrier ' whereby ions are placed
under restraint and are moved against an ionic gradient (Wohl & James,
1942; Jacobsen, Overstreet, King & Handley, 1950).
While the results so far presented are incompatible with Lundegardh's
concept in its present form, they would not conflict with a theory postu-
lating that the electron transfer mediated by all terminal oxidases can
promote the accumulation of salts in the manner attributed by him to
cytochrome. Such an interpretation would, however, be unsupported by
positive evidence, and, furthermore, would be difficult to reconcile with
the more recent extension of Lundegardh's theory (1952) in which it has
been necessary to visualize the three cytochromes, a, b and c, as parti-
cipating in different steps of the mechanism.
Clearly an entirely different experimental approach is necessary if one or
other of the postulated mechanisms is to be established and the other
rejected. It seemed that some progress in this direction would be made if it
could be shown that roots can store a capacity to transfer ions across the
cytoplasm by virtue of prior respiration ; the concept of ' anion respiration '
denies this possibility, but if a carrier mechanism is operative it would be
expected that at any instant healthy tissues will possess some capacity to
accumulate ions by virtue of the carrier within them formed by prior
respiration which has not yet mediated the accumulation of ions.
348 METABOLISM AND ACCUMULATION OF IONS BY PLANTS
Since such an effect has not been observed hitherto, it is reasonable to
infer that, if it does occur, the quantity of ions which can be accumulated
in this way is insignificant by comparison with that absorbed under normal
experimental conditions. An investigation of this question can therefore
be expected to be profitable only if very low concentrations of salt are
employed in the external solution; tracer methods make this relatively
easy, and the principal difficulty to be overcome is to determine whether
the ions entering the plant have in fact been subject to metabolic accumula-
tion. Exchange or adsorption effects on the cytoplasm would be expected
to account for a considerable entry under such conditions, and the im-
portance of determining whether or not the absorbed ions have been held
in this manner is obvious.
This requirement can be largely satisfied by using intact plants. A
number of investigators have concluded that the movement of ions across
the symplast to the stele is a process analogous to the movement of ions
across the cytoplasm of a single cell to its vacuole (Wiersum, 1947; Lunde-
gardh, 1950; Arisz, 1951). Thus if experimental periods are so short that
the nutrient status of the plant is virtually unaffected by the ions absorbed,
changes in the content of shoots may be regarded as reflecting similar
changes in the extent to which ions are transferred across the cytoplasm.
The principal drawback of such studies is the labour they entail, and the
considerable magnitude of experimental errors. Simpler material, such as
detached roots or tissue slices, cannot, however, provide equally un-
equivocal evidence of active transport.
An investigation of the absorption of phosphate by young barley plants
provided the opportunity for examining this question. Plants were grown
to the second leaf stage in a solution containing a balanced supply of
nutrients other than phosphate. The phosphate present in the seed was
sufficient to prevent visual symptoms of deficiency, and the resultant plants
showed a considerable capacity for phosphate accumulation. Before dis-
cussing the evidence which these investigations provide on the mechanism
of nutrient accumulation, it is necessary to summarize data, some of which
has been presented elsewhere (Russell & Martin, 1953; Russell, Martin &
Bishop, 19530) regarding the effect of the amount of phosphate entering
plants on its distribution between roots and shoots, and the relationship
between the external concentration and the rate of absorption of phosphate.
When plants were treated for 24 hr. with concentrations of labelled
phosphate ranging from io~8 to io~4M-H2PO4 (i.e. 0-000316 to 3-16
p. p.m. P), the fraction of the absorbed phosphate which was found in the
shoots increased progressively as the external concentration of phosphate
was increased (Fig. 2, curve A). Still higher concentrations caused the
90
k \
, 1 I
o/
/o
70
50
30
10
1 ill UU
0-01
0-1 1-0 10 iooVd
Mg. phosphate absorbed per plant (log scale) \ .'
Fig. 2. The absorption, distribution and subsequent loss of phosphate by young barley
plants treated with different concentrations of labelled phosphate for 24 hr. Concentration
of labelled phosphate applied (p. p.m. P). a, 31-6; b, 10-0; c, 3-16; </, i-oo; e, 0-316;
/, o-ioo; #, 0-0316; h, o-oioo; i, 0-00316; /, o-ooi ; kt 0*000316. Curve A, percentage of
absorbed phosphate in shoots. Curve B, relative absorption. Curve C, percentage of
absorbed phosphate lost by plants transferred to phosphate free solutions for 7 days.
Exponential mean external concentration
Fig. 3. Relationship between absorption and exponential mean external concentration
for the same experiments as Fig. 2. (Co-ordinates not to scale, «ee text.)
350 THE RELATIONSHIP BETWEEN METABOLISM AND
proportion of phosphate in the shoots to decrease ; between experiments the
concentration which induced this effect varied. The initial phosphate
content of the plants was in the order of 0-15 mg., and it is apparent from
Fig. 2 that over the greater part of the concentration range, the phosphate
content of the plants was changed to a negligible degree during the course
of the experiments. Thus differences in the distribution of the absorbed
phosphate cannot be attributed to effects on the rate of growth of the plants
caused by the nutrient absorbed during the experimental periods. More-
over, the interaction of phosphate with other ions simultaneously absorbed
cannot explain the changing pattern of its distribution, since the same
general effect was observed when phosphate was supplied alone, or in the
presence of other nutrients.
An examination of the relationship between the external concentration of
phosphate and the rate of absorption is complicated by a number of circum-
stances resultant on the low concentrations of phosphate employed.
Evidence to be presented elsewhere (Russell, Martin & Bishop, 1954)
shows, however, that the general nature of the effect of external concentra-
tion on the rate of absorption can be validly assessed. The wide range of
concentrations employed make it impossible to represent the relationship
between absorption and the exponential mean concentration in a graph of
manageable size, using a linear scale. As in the present discussion interest
centres on changes in the slope of the curve, it is sufficient to consider
a diagrammatic representation (Fig. 3) in which the vertical and horizontal
scales have been varied concurrently, so that the lines joining adjacent
points are of equal length, but their slope is correctly shown. A sigmoid
relationship between absorption and the external concentration is apparent.
The region of maximum slope occurred when the external concentration
was between o-i and 0-03 p. p.m. P. The steady decrease in the angle of
the curve below this region indicates that the amount of phosphate
absorbed decreased more rapidly than the external supply of phosphate
when the external concentration was lowered. The same relationship is
shown by calculated values of 'relative absorption' in which the phosphate
absorbed is expressed as a percentage of the amount available to the plants ;
the values show a well-marked maximum for plants supplied with
o-i p. p.m. P (Fig. 2, curve B).
This relationship is strikingly at variance with the proportionality
between external concentration and absorption which normally obtains in
dilute solutions. A possible explanation is suggested by other observations.
When plants which had been treated with varying concentrations of labelled
phosphate for 24 hr. were transferred to phosphate-free media, the pro-
portion of the phosphate which had been absorbed in the previous period
THE ACCUMULATION OF IONS BY PLANTS 351
which was lost from the plants to the external solution was greatest in the
plants which had been supplied with the most dilute solutions (Fig. 2,
curve C). It is apparent from Fig. 2 that plants supplied with less than
OT p. p.m. P showed a marked increase in phosphate loss, while the pro-
portion of the absorbed phosphate found in the shoots and relative
absorption both declined. This suggests that smaller fractions of the
absorbed phosphate which were found in the shoots of plants supplied with
low concentrations of phosphate, and their reduced efficiency of absorp-
tion, were due to phosphate being retained by some mechanism in the
roots which subsequently released it to the outer medium. Phosphate
retained in this manner had clearly not equilibrated with the phosphate
already present in the plants ; this was shown by the fact that no significant
loss of phosphate occurred from plants which were grown to this stage
in the absence of external sources of that nutrient.
In seeking an interpretation of these effects, consideration was first given
to whether the observed retention of phosphate in roots was due to a meta-
bolic process or to a physical mechanism such as adsorption on inert
surfaces. If retention were due to adsorption on inert surfaces, the extent
of depletion of the solution, and therefore the rate of relative absorption,
would be expected to be greatest when the lowest concentrations of
phosphate were provided. This was not so. The loss of labelled phosphate
when plants were transferred to phosphate-free media is also difficult to
reconcile with the view that the phosphate had been retained by a simple
physical process. Further evidence regarding the nature of the mechanism
of phosphate retention was obtained by pretreating plants for 24 hr. with
unlabelled phosphate of concentration 10 p. p.m. In this manner their
phosphate content was increased by approximately 65%. Pretreatment
had little effect on the absorption of labelled phosphate from solutions
containing o-ooi or 10 p. p.m. P during the subsequent 24 hr., nor was the
distribution of phosphate between roots and shoots significantly affected in
the plants supplied with 10 p.p.m. P. When, however, the external con-
centration of phosphate was o-ooi p.p.m. P, the content of shoots was
increased over thirty times by the pretreatment (Table 4). An effect of the
same type, though of lesser magnitude, was shown when plants of con-
siderably greater age were similarly treated.
The foregoing results are most readily interpreted as indicating that
metabolic processes in the root retain phosphate against upward movement,
and that the proportion of the entering phosphate which is retained in this
way is greatest when the phosphate content of the roots is low. Thus it
appears that the distribution of phosphate between the roots and shoots of
barley plants reflects the interaction of two metabolic processes, namely,
352
THE RELATIONSHIP BETWEEN METABOLISM AND
that responsible for retention in the roots and that responsible for the
transference of phosphate to the stele.
Table 4. Effect of pretreatment with phosphate (lop.p.m. P) for 24 hr. on
the absorption and distribution of labelled phosphate in young barley
plants during the subsequent 24 hr.
When a logarithmic transformation was necessary for statistical analysis, the trans-
formed values are shown in italics.
Initial concentration
of solution (p. p.m. P)
/ig. P per plant
%of
absorbed
phosphate
in shoots
Pre-
treatment
Treatment
Root
Shoot
Total
Nil
10
S.D.
Nil
IO
S.D.
10
10
o-ooi
O'OOI
39'i
42-0
n.s.
0-0647
0-0649
n.s.
5i-9
52-5
n.s.
0-0004
(0-604}
0-0149
(2'i73)
(0-163}
55'0
91-8
94*5
n.s.
0-0651
0-0798
n.s.
55-8
55'5
n.s.
0-63
(0799)
18-80
(2-274)
(0-152)
Phosphate absorbed
during pretreatment
43'4
98-4
56-0
Experiments with respiratory inhibitors enable some conclusions to be
drawn with regard to these two mechanisms. Typical results for experiments
lasting 24 hr. are summarized in Fig. 4. At the highest concentration of
phosphate (i.e. 10 p. p.m. P), the application of a range of concentrations of
sodium azide, 2, 4-dinitrophenol and diethyldithiocarbamate markedly
reduced absorption, the content of the shoots being depressed to a greater
extent than that of the roots. When, however, the initial concentration of
the external solution was o-ooi p. p.m. P, the phosphate content of the
shoots of plants at the second leaf stage was increased by certain concentra-
tions of the inhibitors; io~~3M-sodium azide, io~4M- diethyldithiocarba-
mate and io~5M-2, 4-dinitrophenol were most effective in this regard. The
phosphate content of the roots of plants which showed this effect was in
some experiments increased though to a lesser extent than that of the shoots ;
in other experiments it was reduced but to a smaller extent than in plants
supplied with 10 p.p.m. P. Higher concentrations of the inhibitors
reduced the content of both roots and shoots.
As in all studies conducted in a greenhouse in which the vagaries of
climate are not fully controlled, there was considerable variation
between different experiments. It was shown in experiments with io~3M-
sodium azide, the inhibitor most extensively used, that the nature of the
effect of the inhibitor on the phosphate content of the shoots of plants
THE ACCUMULATION OF IONS BY PLANTS
353
supplied with o-ooi p.p.m. P varied with plant age (Fig. 5). When
plants of which the first leaf was developing were used, the inhibitor
reduced the phosphate content of shoots, though the magnitude of the
10 p.p.m. P
0-001 p.p.m. P
+ 600
+ 400
o
-5+200
•100
Shoots
TTL
!!
_u
-50
-100L
Roots
T-i
10-5
10-4
10-3
10-2
10-s
10°
10-1
Molar concentration of inhibitors
Fig. 4. Changes in phosphate content of roots and shoots of young barley plants induced
by inhibitors in experiments lasting 24 hr. Left : Concentration of phosphate in external
solution 10 p.p.m. P. Right: Concentration of phosphate in external solution o-ooi p.p.m. P.
Solid lines: sodium azide. Broken lines: diethyldithiocarbamate. Dot-dash lines:
2,4-dinitrophenol. Values expressed as percentages of controls. Significant differences
from control treatments are shown and significant effects are indicated by solid circles.
effect was less than in plants supplied with 10 p.p.m. P. When progressively
older plants were similarly treated, increasing stimulation of phosphate
transfer to the shoots were shown. This result indicates that the effect of
354
THE RELATIONSHIP BETWEEN METABOLISM AND
the inhibitor in increasing the content of shoots was not a necessary con-
sequence of a small amount of phosphate entering the plant. A comparison
of the curves for the percentage change in shoot content induced by
+ 50r
4-1000
+ 800
+ 600
+ 400
+ 200
1 2 3 41 2
Youngest leaf fully emerged
Fig. 5. Effect of sodium azide on the absorption of phosphate by barley plants of different
ages from solutions containing o-ooi p.p.m. P, in experiments lasting 24 hr. A and B,
changes in phosphate content of shoots and roots respectively. Solid line: io~3M sodium
azide. Broken line: io~5M sodium azide. Significant effects are indicated by solid circles.
C, mean dry weight of control plants.
io~3M-azide (Fig. 5 A) and for plant dry weight (Fig. 5C) shows an
obvious similarity. The relationship between these quantities was ex-
amined on the pooled data of this and two other experiments, and a highly
significant linear regression of plant dry weight on the percentage increase
THE ACCUMULATION OF IONS BY PLANTS 355
in shoot content induced by io~3M-azide was demonstrated, the equation
"einS y = (0-102 ±0-021) x -21-4.
Since the plants received no phosphate before the experimental treatment,
their phosphate status varied inversely with their weight. On the basis of
the results obtained by pretreating plants with phosphate (Table 2), it is to
be expected that plants of low phosphate status will show the greatest
metabolic retention of phosphate in roots. Thus the observed relationship
between plant weight and the effect of the inhibitor is regarded as indi-
cating that the extent to which shoot content is increased by the inhibitor
depends on the extent to which phosphate entering the plant would be
metabolically retained in the roots if no inhibitor were present. It is
therefore considered that the primary effect of the inhibitor was in all
Outer medium
Process (C)
Root
Process (A)
(Formation of complex
j with a product of
I respiration ,
Process (B)
Metabolic retention
Vascular stele
Fig. 6
probability the inhibition of phosphate retention, and that the stimulation
of phosphate transfer to the shoots was an indirect effect due to a larger
fraction of the entering phosphate being available for transference to the
stele.
The results of these experiments can be explained in terms of the three
interrelated processes shown in Fig. 6. Each process, in all probability
embraces a number of stages: Process A: the formation at or near the cell
surface of a complex between the entering ion and a product of respiration ;
this process is envisaged as resulting in the ultimate release of the ion into
the stele, unless it is diverted into Process B, the mechanism of metabolic
retention in the cytoplasm of the root. Process C results in the ultimate
release into the outer medium of a fraction of ions which have previously
been utilized in process B. It is postulated that the root possesses a
certain capacity to accomplish process A in consequence of the necessary
metabolic product having been produced by prior respiration. Thus the
inhibition of respiration will bring about a comparable inhibition of
23-2
356 THE RELATIONSHIP BETWEEN METABOLISM AND
process A only if the amount of respiratory intermediate is negligible by
comparison with the number of ions available for absorption. The extent
to which phosphate is utilized in process B is, however, regarded as
dependent on the requirements of concurrent metabolism. It is postu-
lated further that process B has a greater affinity for phosphate than
process A. Thus, if the amount of phosphate entering the plant is reduced
so that it is insufficient to saturate both mechanisms, the proportion of the
phosphate diverted to process B will be increased. This will in turn increase
the extent of process C relative to process A. In this way the reduction in
the efficiency of absorption which occurred when the external concentra-
tion was reduced below o-i p. p.m. P can be explained.
The results of the experiments with inhibitors are also compatible with
the proposed mechanism. When the external concentration of phosphate
was high (10 p. p.m. P) absorption and translocation to the shoot was
greatly reduced by inhibitors, as would be expected if the rate of process A
was closely related to the rate of simultaneous respiration. When the
external concentration of phosphate was low (o-ooi p. p.m. P) and the
extent of metabolic retention in the plants not treated with inhibitors was
also low, absorption and translocation were reduced by inhibitors to a much
smaller extent; this is attributed to the fact that the product of prior
respiration was now able to mediate the absorption of a significant pro-
portion of the ions available to the plants. When, however, the external
concentration of phosphate was low (o-ooi p. p.m. P) and the extent of
metabolic retention in the plants not treated with inhibitors was high,
upward movement to the shoot was increased by inhibitors, and in some
cases the total amount of phosphate absorbed was increased also. The
increased shoot content is interpreted as indicating that the extent of
inhibition of process B sufficiently exceeded that of process A to cause a
greater rate of release of ions into the vascular stele. The increase in total
plant content induced by inhibitors in some experiments suggests that
phosphate loss through process C was reduced to a greater extent than
primary absorption (process A).
The speculative nature of this interpretation is obvious. The results,
however, appear irreconcilable with any theory which denies the ability of
roots to * store* the capacity to transfer ions across the cytoplasm. The
concentrations of sodium azide and diethyldithiocarbamate which in-
creased the movement of phosphate to the shoots of plants supplied with
o-ooi p. p.m. P have been shown to inhibit respiration; thus the concept
that the electron transfer in respiration directly mediates the accumulation
of ions is unacceptable. The formation of a product of respiration which
serves as a carrier must therefore be envisaged. It is of interest that
THE ACCUMULATION OF IONS BY PLANTS 357
Sutcliffe reached a similar conclusion in his study of the absorption of
cations by slices of beet, which have been described earlier in this
Symposium.
Because of the special role of phosphate in metabolism, results obtained
with that ion must be treated with caution in making generalizations re-
garding absorption. However, the view that information on the relation-
ship between absorption and respiration gained by the use of phosphate
may be of general application is encouraged by the fact that respiratory
inhibitors appear to affect the absorption of different ions from relatively
concentrated solutions to apparently the same extent. In a series of
experiments in which io~3M-diethyldithiocarbamate was applied to barley
root tips, the percentage inhibition of the absorption of phosphate and of
bromide from 0-002 M solutions, together with the 5 % fiducial limits were:
P: 67-3 ± 10-4% inhibition,
Br: 73-3 ± 9-7% inhibition.
IV. THE INTERACTION BETWEEN EXCHANGE PROCESSES
AND THE CARRIER MECHANISM
Before considering the inference which can be drawn concerning the
probable nature of the mechanism whereby ions are accumulated, it is
desirable to consider the part played by ionic exchange in this process.
The importance of exchange reactions has been recognized since the earlier
investigations of Briggs (1930) and Brooks (1939) and the simultaneous
occurrence of ionic exchange reactions and of the movement of ions by
the postulated carrier mechanism must be envisaged. Knowledge of the
relationship between these two processes is of obvious importance if
information concerning the functioning of the carrier mechanism is to
be derived from observations of salt uptake by plant tissues. If the two
processes are relatively independent one of the other, each may exert
a separate influence on the rate of influx of ions into tissues; thus changes
in salt uptake induced by experimental treatments may be due to effects
on either or both mechanisms. If, on the other hand, the two processes
are closely related, as would be the case if all exchangeable ions in the
cytoplasm were located on the carrier, changes in the rate of association
of ions with the carrier would be indicated by changes in the rate of
entry of ions into the tissue. The study of the carrier mechanism would
thus be greatly simplified. The results of many investigations suggest that
the rates of the two processes vary independently (e.g. Brooks, 1939;
Broyer, 1950). Recently, however, Jacobsen et al. (1950) have proposed
a carrier mechanism which implies that the process of ionic entry by iso-
358 THE RELATIONSHIP BETWEEN METABOLISM AND
topic exchange is identical with the association of the ion with the carrier.
They have postulated that binding substances (HR and R'OH) are pro-
duced by metabolism and that entering ions (M+ and A~) react with them
at or near the surface of the protoplasm in the following manner :
The complexes MR and R'A are considered to transfer the ions to the
vacuole in which they are released by the chemical alteration of the com-
plexes. On the basis of this concept, Epstein & Hagen (1952) and Epstein
(1952) have discussed the mode of binding of ions in the postulated carrier.
They observed the effects of varying external concentrations of potassium
and sodium on the rate of absorption of rubidium by detached barley
roots, and after a kinetic analysis of their results arrived at the conclusion
that potassium competes for the same sites as rubidium in the postulated
carrier, while sodium does not. Applying the same procedure to anions,
Epstein (1953) concluded that chloride, but not nitrate, is bound at the same
sites as bromide. These interpretations rest on the assumption that the effect
of one ion on the rate of entry into the cell of another ion of the same sign is
due solely to competition for sites in the complex which effects the trans-
ference of ions across the cell. Stated otherwise, this means that changes in
the concentration of one ion will affect the rate of entry of another ion of the
same sign into the cytoplasm, and its rate of transference across the cyto-
plasm, in the same manner. If these assumptions are valid, the methods
of Epstein & Hagen would provide an important approach to the study
of the carrier mechanism. Since, however, they did not examine the fate
of the ions absorbed by the tissues which they studied, their interpre-
tation lacks proof. Information on this question could be obtained by
applying combinations of ions similar to those used by Epstein & Hagen to
a system in which it was possible to make separate observations of total
absorption and of the metabolic transport of ions across the cytoplasm.
For the reasons stated earlier, barley seedlings were considered suitable for
the purpose.
It was found that when barley plants at the second leaf stage were treated
with combinations of rubidium and potassium, both the absorption of
rubidium and its distribution between roots and shoots were markedly
affected. Absorption was reduced in a manner comparable to that reported
by Epstein & Hagen. The extent and nature of the effect on the distribution
of rubidium between roots and shoots, varied dependent on the prior
treatment of the plants. Their potassium status was a major factor in this
regard.
THE ACCUMULATION OF IONS BY PLANTS
359
For the experiment illustrated in Fig. 7 and Table 5 plants were raised
to the second leaf stage in a potassium-free nutrient solution of the
following composition in milli-equivalents per litre: Mg, 3; Ca, 9; Na, 2;
SO4, 3; H2PO4, i; NO3, 10. Half the plants were then pretreated for
24 hr. with 0-5 x io~3M-potassium, the remainder being maintained in
a potassium free medium. io~3 or io~2M-labelled rubidium was then
provided for 4 hr. in the presence and absence of 2 x io~2M-potassium.
The absorption of rubidium and its distribution between roots and shoots
Absorption of rubidium (mg /plant) Percentage of rubidium in shoots
r-0-6
L-0-5
-0-4
-0-3
-0-2
-0-1
Pretreatment
0-5
Nil
0-5 mriK
Nil
Fig. 7. The absorption and distribution of rubidium in young barley plants to which
potassium was provided prior to, or simultaneously with, rubidium. For statistical
analysis see Table 5.
was significantly affected by all single-factor effects and first-order inter-
actions, with the exception that absorption was not affected by the inter-
action between pretreatment and simultaneous treatment with potassium.
From the viewpoint of the present discussion, the effect of potassium on
the distribution of rubidium is of particular relevance, and it alone will be
discussed. Whereas pretreatment with potassium reduced the fraction of
rubidium absorbed in the subsequent period which was moved to the
shoots, the reverse effect was induced by potassium in plants which had not
been pretreated with that ion when rubidium was supplied at the higher
level. Somewhat contrasting results were obtained when the same con-
centrations of potassium and rubidium were applied to plants which had
360 THE RELATIONSHIP BETWEEN METABOLISM AND
been raised in io-3M-Ca(NO3)2 and io~3M-CaSO4. Their rate of growth
was greatly retarded as compared with plants grown in the potassium
free nutrient solutions. Under these circumstances the presence of
potassium significantly increased the proportion of the absorbed rubidium
found in the roots at both levels of rubidium (Table 6).
Table 5. The absorption and distribution of rubidium in young barley plants
to which potassium was provided prior to, or simultaneously with,
rubidium
Before the experiment the plants were grown in a potassium-free nutrient solution.
0-48 hr.*
Pretreatment with potassium
(M x io~3)
o-5
Nil
72-76 hr.
Concentration of rubidium
(MX io~8)
I 10
I IO
Rubidium absorbed mg. /plant
K absent
K present (2 x io~2M)
i'33 2'33
0-32 2-35
5-20 7-29
1-20 6-15
% of absorbed rubidium in
shoots
K absent
K present (2 x io~2M)
43-87 44-04
36-90 40-79
55-69 63-66
52-33 66-82
* From 48 to 72 hr. all plants were in distilled water.
Statistical analysis
Varianc<
i ratios
Effect
Absorption!
% of absorbed
of rubidium (mg.)
rubidium in shoots
Pretreatment with K (Pt)
Simultaneous treatment with K (S)
Rubidium (Rb)
667-21
278-29
611-04
275-1
5'6
36-7
PtxS
2-17
5'5
PtxRb
8-44
17-6
SxRb
219-00
6-3
5 % point
4'
54
i % point
8-
68
o- 1 % point
16-
59
t Data transformed to logarithmic basis for analyses.
No detailed interpretation of the interaction between rubidium and
potassium with respect to the absorption and distribution of the former ion
can yet be advanced. It is, however, apparent that the two ions interacted
independently with respect to entry into roots and to transference across
the cytoplasm to the stele. This situation is most simply interpreted by the
postulate that the ions interact differently in exchange processes following
their initial entry into the root, and in the subsequent process when they
are transferred across the cytoplasm. Clearly, then, no conclusions with
THE ACCUMULATION OF IONS BY PLANTS 361
regard to the competition of ions for sites on the carrier can be reached
on the basis of observation of changes in the absorption of one ion induced
by varying the concentration of other components of the external solution.
Table 6. Effect of 2 x io~2M potassium on the absorption of
rubidium from solutions by young barley plants
Duration of treatment (hr.) . . .
4
20
Concentration of rubidium in
T
10
i f\
external solution (M x io~3)
Rubidium absorbed (mg. /plant):
K absent
0-050
0-079
0-234
0-320
0-385
K present
0-0078
0-071
O-O3I
0-126
0-233
S.D.
o-
01
O-O27
% of absorbed rubidium present
in shoots:
K absent
19-5
27-1
64-0
67-3
68-0
K present
32-1
38-1
68-8
72-4
72-8
S.D. 5-
25
1-6
V. CONCLUSIONS
In the absence of any direct evidence as to the nature of the postulated
carrier in active accumulation, no detailed interpretation can be possible of
the mechanism whereby ions are transferred from the outer medium
against an apparent concentration gradient to the vacuole or vascular stele.
The evidence here presented indicates, however, that two distinct phases
must be envisaged, namely, the entry of ions into the cytoplasm by
exchange and the subsequent reaction involving the carrier already
postulated. Discussion will here be confined to these two processes, but it is
obvious that they do not embrace all reactions in which the entering ion
may take part; the utilization of ions in metabolic processes near their site
of entry may, as has been shown in the studies with phosphate, have a
profound effect on the rate of absorption. The examination of the nature of
these processes is beyond the scope of the present discussion, though the
work of Kamen & Spiegelmann (1948) encourages the view that the
phosphate metabolically retained in roots was involved in esterification
processes.
Exchange mechanisms have frequently been described as passive or
physical (e.g. Broyer, 1951) to distinguish them from the subsequent active
step of accumulation. This terminology seems misleading, since ionic
exchange or Donnan equilibria depend on metabolically produced sub-
stances equally with the active mechanism of accumulation. The apparent
independence of respiration frequently shown by the passive process clearly
indicates not independence of metabolism, but that metabolism is main-
362 THE RELATIONSHIP BETWEEN METABOLISM AND
taining the process at a relatively steady state. This is readily demonstrated
by the effect of gross changes in metabolism : sufficient concentrations of
inhibitors destroy the exchange capacity of cells. Conversely, as Eddy &
Hinshelwood (195 1) have shown, the induction of a state of rapid glycolysis
in resting bacteria causes the rapid entry of cations, apparently by an
exchange process. Such effects are clearly of an entirely different category
from those which lead to the transference of ions to cell vacuoles or to the
vascular stele of roots. The total exchange capacity of the cytoplasm and
probably also the affinity of different sites for ions must thus be expected
to be variable, depending on metabolic activity. The variable effects of
potassium on the transference of rubidium to shoots here described would
appear to reflect such differences.
Outer medium i Cytoplasm
Entry of ions
by exchange for
hydrogen ions
Movement of
ions to vacuole
or stele
u_
Barrier
zone
-MX-
X1
Vacuole or stele
Fig. 8. Postulated mechanism for the metabolic accumulation of cations. (A similar
mechanism for anions is envisaged.) M = entering cation. X= carrier produced by meta-
bolism. Xl = product of breakdown of carrier-ion complex.
A generalized scheme embracing the two processes, whereby it is
suggested ions are metabolically accumulated, is proposed with more than
a little trepidation in Fig. 8. For reasons stated earlier it is assumed that
the influx of ions into the vacuole of single cells and into the vascular
tissues of roots depends on broadly similar processes. However, it is to be
expected that more detailed studies will reveal differences between the two
mechanisms.
The outer layers of the cytoplasm are regarded as being relatively
permeable to ions by exchange or diffusion. This conclusion is indicated
not only by the marked isotopic exchange of cations which Broyer &
Overstreet (1951) and Sutcliffe have demonstrated, but also by the ease
with which phosphate which has been retained in metabolic processes in
the roots subsequently diffuses into the outer medium. Both because of
the marked concentration gradient which exists between the vacuole of
cells and the outer medium and because of the apparently low exchange-
THE ACCUMULATION OF IONS BY PLANTS 363
ability of ions which have penetrated sufficiently deeply into the cell
(cf. Sutcliffe), the inner layer of the cytoplasm, presumably the tonoplast,
must be regarded as a * barrier' to the free movement of ions by exchange
or diffusion. The transference of ions across this zone is considered to be
effected by the carrier mechanism.
It is visualized that respiration leads to the production of a substance (X)
which is freely diffusible throughout the cytoplasm, and which places ions
under restraint on the outer side of the ' barrier', forming a complex which
diffuses across the 'barrier' and breaks down to liberate the ions on the
inner side. This scheme differs in no great respect from that put forward by
Wohl & James (1942). Rosenberg (1948) has discussed the thermodynamic
aspects of mechanisms of this type, and their existence has been suggested
by Steward & Street (1947), and Jacobson et al. (1950). The view that the
carrier may be amphoteric is encouraged by the fact that the present work
and that of Sutcliffe indicates the existence of comparable mechanisms for
cations and anions respectively.
While no direct information with regard to the nature of the carrier is
available, the correlation between protein synthesis and salt accumula-
tion which occurs in many tissues has led to the suggestion by Steward &
Street that ions combine with nitrogenous compounds and are released
when these compounds break down. Some evidence compatible with this
view has been obtained in the present investigations; over periods of 3 hr.
or longer, the absorption of phosphate by barley plants has been increased
by the presence of nitrate (Table 7). It appears also that the absorption of
rubidium is increased to a smaller extent in experimental periods of 24 hr.
Table 7. Effect of 3 x io~4M nitrate on the absorption of phosphate by
barley plants at the second leaf stage from a solution containing
3xio-8M-H2PO4
Duration of
experiment
(hr.)
Nitrate
present
Nitrate
absent
S D.
6
24
i-63
2-OI
2-42
0-96)
1-42
1-52 )
0-27
An alternative interpretation of these results is, however, that the increase
in respiration rate occasioned by nitrogen metabolism and not the produc-
tion of nitrogenous compounds is effective in increasing absorption.
The energy relations of the carrier mechanism may now be considered.
Since the earlier observations of Steward (1933) it has been apparent that
the accumulation of ions is dependent on the release of energy by respira-
364 THE RELATIONSHIP BETWEEN METABOLISM AND
tion. That two metallo-terminal oxidases, cytochrome and ascorbic acid
oxidases, can effect the necessary electron transfer is now established and it
appears from the work of Harley that flavoproteins may have the same
capacity. A question of obvious interest is whether the formation of the
carrier is the only step in the accumulation process which is quantitatively
dependent on respiration. Some evidence that this is the case is provided
by the fact that the carrier produced by prior respiration can apparently
mediate the transfer of phosphate when respiration is subsequently
inhibited. Results presented by Sutcliffe also support this suggestion.
He considers it probable that metabolic energy is not involved in the
combination of the ion with the carrier because this reaction can appar-
ently occur at low temperatures. Moreover, the view that the breakdown
of the carrier-ion complex is likewise independent of respiration is com-
patible with his observations of the effect of temperature changes between
15 and 30° C. on the absorption of cations. Increasing temperature
within this range markedly accelerated the rate of absorption of ions,
but the concentration in the tissues when equilibrium had been attained
with the external solution was unaffected. The increased rate of trans-
port when temperature was raised can be attributed to the effect of the
increased rate of respiration on the rate of formation of the carrier. If,
however, the association of the ion with the carrier and the breakdown of
the resultant complex are not directly dependent on respiration, the
equilibrium concentration in the cell will be determined by other factors;
accumulation will cease when the slow outward diffusion of ions from the
vacuole into the cytoplasm equals their rate of entry into the cytoplasm
from the outer medium. This will result in the concentration of MX being
constant throughout the system and its inward diffusion will therefore
cease. Since both the diffusion of ions into the vacuole and the initial
process of entry are expected to be little affected by temperature, the
equilibrating concentration will be unaffected by metabolism. If, on the
other hand, the breakdown of MX at the interior did depend on a glycolytic
process, its rate would increase with rising temperature, and the equi-
librium concentration in the tissues would be consequently affected. This
accords with the conclusion of Lundegardh (1950) that the release of ions
into the stele is not linked with glycolysis.
The type of mechanism here discussed is unsupported by direct evidence.
No published data appear, however, to be irreconcilable with it. The close
correlation between the activity of cytochrome and salt absorption which
has been demonstrated in great detail by LundegSrdh (1945, 1950, 1952,
I953a> ft) cannot be regarded as proof of the theory of anion respiration.
Such results can equally be interpreted as indicating that, in the tissues
THE ACCUMULATION OF IONS BY PLANTS 365
investigated, a carrier is produced by respiration through the cytochrome
system, and that the rate of utilization of the carrier is an important factor
determining the rate of its production. Similarly, the quantitative relation-
ship between respiration and salt uptake demonstrated by Robertson &
Wilkins (1948) can be interpreted as indicating that the transfer of one
electron is required for the production of the carrier necessary to bring
about the accumulation of one univalent ion of each sign. The di fficulties
Robertson, Wilkins & Weeks (1951) and Lundegardh (1952) have ex-
perienced in reconciling the fact that 2, 4-dinitrophenol may inhibit salt
uptake without inhibiting respiration can be explained by the postulate
that this substance inhibits the formation of the carrier by interference
with energy transfer.
VI. SUMMARY
Evidence is presented which indicates that respiration through the ascorbic
acid oxidase system can mediate salt absorption in the roots of young barley
plants. An examination of the effects of respiratory inhibitors on the
absorption of phosphate by barley plants from dilute solutions leads to the
conclusion that tissues can store the capacity to accumulate ions. It is
therefore considered that the accumulation is affected by a carrier substance
produced by respiration. The relationship between the transfer of ions
across the cytoplasm and ionic exchange process is considered. The view is
advanced that the accumulation of ions must be interpreted in terms of both
exchange reactions and of the association of ions with a metabolically
produced carrier.
REFERENCES
ALBERT, A. & GLEDHILL, W. S. (1947). Biochem.J. 43, 636.
ARISZ, W. H., HELDER, R. J. & VAN NIE, R. (1951). jf. Exp. Bot. 2, 257.
BRIGGS, G. C. (1930). Proc. Roy. Soc. B, 107, 248.
BROOKS, S. C. (1939). J. Cell. Comp. Physiol. 14, 383.
BROYER, T. C. (1950). Plant. Physiol. 25, 367.
BROYER,T. C. (1951). Miner al Nutrition of Plants, zd. E. Truog, p. 187. Wisconsin.
EDDY, A. A. & HINSHELWOOD, C. (1951). Proc. Roy. Soc. B, 138, 237.
EPSTEIN, E. (1952). Proc. Fourth Annual Oak Ridge Summer Symposium, p. 418.
Oak Ridge.
EPSTEIN, E. (1953). Nature, Lond., 171, 83.
EPSTEIN, E. & HAGEN, C. E. (1952). Plant Physiol. 27, 457.
HARLEY, J. L., MCCREADY, C. C. £ BRIERLEY, J. K. (1953). Personal com-
munication.
JACOBSON, L., OVERSTREET, R., KING, H. M. & HANDLEY, R. (1950). Plant
Physiol. 25, 639.
JAMES, W. O. (1953). Proc. Roy. Soc. B, 141, 289.
JAMES, W. O. & GARTON, N. (1952). J. Exp. Bot. 3, 310.
KAMEN, M. D. & SPIEGELMANN, S. Cold Spr. Harb. Symp. Quant. Biol. 13, 151.
LUNDEGARDH, H. (1945). Ark. Bot. A 32, i.
LUNDEGARDH, H. (1950). Physiol. Plant. 3, 103.
366 METABOLISM AND ACCUMULATION OF IONS BY PLANTS
LUNDEGARDH, H. (1952). Nature, Lond.y 169, 1088.
LUNDEGARDH, H. (19530). Nature, Lond., 171, 477.
LUNDEGARDH, H. (19536). Nature, Lond., 171, 521.
MARTIN, R. P. & RUSSELL, R. SCOTT (1950). J. Exp. Bot. i, 141.
OVERSTREET, R. & DEAN, L. A. (1951). Mineral Nutrition of Plants, ed. E. Truog,
p. 79. Wisconsin.
OVERSTREET, R. & JACOBSON, L. (1952). Ann. Rev. PL Physiol. 3, 189.
ROBERTSON, R. N. & WILKINS, M. J. (1948). Aust.J. Set. Res. B, i, 17.
ROBERTSON, R. N., WILKINS, M. J. & WEEKS, D. C. (1951). Aust.J. Sci. Res. B,
4, 248.
ROSENBERG, T. (1948). Ada chem. scand. 2, 14.
RUSSELL, R. SCOTT & MARTIN, R. P. (1950). J. Exp. Bot. i, 133.
RUSSELL, R. SCOTT & MARTIN, R. P. (1953). J. Exp. Bot. 4, 108.
RUSSELL, R. SCOTT, MARTIN, R. P. & BISHOP, O. N. (19530). J- ExP- Bot- 4» J36.
RUSSELL, R. SCOTT, MARTIN, R. P. & BISHOP, O. N. (19536). In the Press.
STEWARD, F. C. (1933). Protoplasma, 18, 208.
STEWARD, F. C. & STREET, H. E. (1947). Ann. Rev. Biochem. 16, 471.
WIERSUM, L. K. (1947). Rec. Trav. hot. neerl. 41, i,
WOHL, K. & JAMES, W. O. (1942). New Phytol. 41, 230.
SALT ACCUMULATION IN PLANTS:
A RECONSIDERATION OF THE ROLE
OF GROWTH AND METABOLISM
A. SALT ACCUMULATION AS A CELLULAR PHENOMENON
BY F. C. STEWARD AND F. K. MILLAR
B. SALT ACCUMULATION IN THE PLANT BODY
BY F. C. STEWARD*
Botany Department, Cornell University, Ithaca, New York
The problem of salt accumulation in plants necessarily begins with the
process in single cells. This phase of the problem will be considered in
Part A of this review.
With respect to a vascular plant the problem involves other considera-
tions which lead to an understanding of what may be called the ' internal
nutrition ' of the organism and to the recognition in the vascular plant body
of 'centres of growth and salt accumulation*. While such centres clearly
involve the method by which individual cells accumulate solutes they also
involve other interrelationships, and the problem thus impinges upon the
larger ones of growth, development and morphology of the entire organism.
An attempt will be made to analyse these questions in Part B of this review.
No attempt will be made to deal comprehensively with the extensive
literature in this field ; this will permit the presentation of some new data,
and more particularly will allow space to develop certain speculative ideas.
A. SALT ACCUMULATION AS A CELLULAR
PHENOMENON
I. INTRODUCTION
A principal feature of the Nitella experiments was the exchange of entering
bromide for issuing chloride. It was tacitly assumed that this was a true
exchange of bromide for chloride. However, in the experiments a large
population of Nitella cells was used and the possibility, in fact the prob-
ability, existed that the bromide entered the younger and still growing,
expanding cells, while the chloride issued from older and more senescent
* Using data from experiments with S. M. Caplin, J. A. Harrison, F. K. Millar,
R. Overstreet, B. M. Pollock, and A. G. Steward.
368 SALT ACCUMULATION IN PLANTS: A RECONSIDERATION
cells. An oft-quoted experiment showed that a given population of Nitella
cells only attained a steady state with respect to the concentration of
bromide and chloride in the cells after the lapse of a relatively long period,
namely, 40 days. It is inconceivable that any biological process involving
a large population of cells and requiring the lapse of so long a time should
proceed independently of the concomitant changes due to growth and
development. Therefore, even in the Nitella experiments, there was already
some evidence that the actual development of the cells in question might
be involved, in a determining way, in the intake of salt into their cell sap.
It was the experiments on storage tissue, particularly those on thin slices
of potato tissue made in the first instance in Hoagland's laboratory
(Steward, 19320), that turned our attention and also that of Hoagland to
the importance of respiration as the metabolic process that mediated the
energy required by salt accumulation.
On previous occasions the work on salt accumulation, largely using cut
disks of plant storage tissues as experimental material, has been summarized
from the following points of view :
(a) From the standpoint of the role of respiration and metabolism in salt
accumulation (Steward, 1935, 1937), recognizing that the respiratory and
metabolic activity of the cell is involved because it furnishes the ultimate
source of energy for the process, and that the relationship is more indirect
than if the entry of ions was determined by the exit of a specified amount
of carbon dioxide.
(b) From the standpoint of the status of the cell or organ for further
growth and development, recognizing that their potentialities for growth,
by division or cell enlargement, profoundly modify the way that meta-
bolism may be used in salt accumulation (Steward, 1935) and that their
previous nutritional status, i.e. whether they are of high or low salt content,
will profoundly alter the amount of salt that may be absorbed (Hoagland &
Broyer, 1936).
(c) From the standpoint of the varied metabolic processes that may be
observed in potato cells that are accumulating salts (Steward & Street,
1947), recognizing that in order to understand this system one needs to
know how these various processes (respiration, protein synthesis, oxidase
activity, starch^sugar equilibria, etc.) are influenced by the principal
variables that determine salt accumulation (the nature and concentration
of the salt, oxygen tension, temperature, surface/volume relations of the
tissue, the time drift after cutting the tissue, etc.).
From these various summaries and the papers to which they refer, the
following main ideas and conclusions were obtained prior to the work now
to be presented :
OF THE ROLE OF GROWTH AND METABOLISM 369
(i) In thin disks of potato tuber, intake of cation and anion from very dilute
solutions may proceed in approximately equivalent amounts at rates which may
be linear with time for long periods. The rates of ion intake are determined by
oxygen tension, i.e. pO2 in the gas stream (Steward, 1933) and temperature
(Steward, Berry, Preston & Ramamurti, 1943).
(ii) The relative absorption (accumulation ratio) of bromide was greater from
more dilute solutions. In very dilute stirred solutions, furnished in large volume,
the accumulation ratio could be very high (Steward, 19320). In the range
0-045, °'°°45> 0-00045 M-KBr a tenfold increase of external concentration
doubled the internal concentration and the effect on the respiration during the
period was small (Steward, 1933).
(iii) The effects of oxygen pressure on salt intake and respiration indicated
that salt intake varied along with a large component of the aerobic respiration
that was only oxygen saturated in solutions in equilibrium with air, and this
component was probably mediated by the polyphenol oxidase system, rather
than by a cytochrome system. The pO2 at which oxygen saturation occurred
was different for tuber tissue and for excised roots (Steward, Berry & Broyer,
(iv) The relation of salt intake to respiration did not only concern a part of
the respiration due solely to the presence of the salt, but it varied with respiration
and metabolism which occurred even in the absence of the salt and was deter-
mined by other variables (/>O2; temperature, time, proximity to the surface of
the disk, etc.). The relation of respiration to salt intake was not regarded as
a simple and direct one, nor was it the direct consequence of carbon dioxide
production per se.
(v) Salt accumulation and the metabolism with which it is associated proceed
at a greater intensity in the surface layers of cut disks, and it is in these cells
that visible signs of growth and vital activity may be seen, especially in disks
that are exposed to moist air.
(vi) To accumulate bromide within, potato disks require other properties
than a high rate of respiration. These are the ability of the cells to grow and
divide and to synthesize protein, properties which are eliminated after long
storage of tubers at low temperature (Steward et al. 1943). In the case of
artichoke tuber (Steward & Berry, 1934) the intake of bromide is linked with
properties which change with the lapse of time after cutting, absorption being
higher at first and declining thereafter.
(vii) The variations in the metabolism brought about by the nature and
concentration of the external salt all tend to show that at least a major part of
the aerobic respiration of the tissue slice is mediated in ways in which protein
synthesis and respiration vary together. Parallel effects due to salts on respira-
tion and protein synthesis were observed, and these were ascribed primarily to
the cations, with contrasting K : Ca effects, but were modified by the anions.
Throughout, the effects of salts and oxygen on respiration and on the use of
soluble nitrogen compounds in protein synthesis run parallel (Steward &
Preston, 1940, 19410).
(viii) External pH's of 7-0 promoted respiration and protein synthesis, but
sharply contrasted effects were observed at constant pH due to the use of
CO2/HCO^ or phosphate buffers. Whereas increased phosphate concentration
E B S VIII 24
370 SALT ACCUMULATION IN PLANTS: A RECONSIDERATION
increased respiration and protein synthesis, increased concentrations of the
CO2-bicarbonate buffer at pH 7-0 decreased it and could eliminate bromide
accumulation altogether (Steward & Preston, 1941^).
(ix) In thin slices of storage tissue many vital processes are linked together.
Increased respiration, the activity of the oxidase system, synthesis of protein
from soluble nitrogen compounds and salt accumulation all tended to run
parallel as affected by external conditions. Protoplasmic streaming (Steward,
19320) was also a visible indication that while the tissue was doing osmotic
work in accumulating salt it was also capable of mechanical work. But the
property that above all expresses the ability of the cells to accumulate salts
seemed to be their ability to synthesize protein — if this were lacking much
metabolic acitivity and respiratory carbon dioxide seemed to be of no avail.
(x) The effects of rapid aeration were now seen to be twofold. First it
furnishes oxygen at the appropriate pressure for respiration and salt accumu-
lation. Secondly, it sweeps out carbon dioxide as fast as it is formed. In the
latter way carboxylation reactions are limited and decarboxylations are fostered
(Steward & Street, 1947). Thus the high rate of respiration that obtains under
these conditions is maintained largely, if not solely, by use of the carbon residues
from deaminated amino-acids — residues that can be fed in to the Krebs cycle
in lieu of the breakdown products from sugar. Meanwhile the nitrogen groups
thus transferred form protein, using a carbon framework derived from sugar.
It is this * nitrogen cycle* that is fostered by rapid aeration and is inhibited by
carbon dioxide. In * handing on' the nitrogen for protein synthesis, glutamine
and glutamic acid play the key role. Any treatment which retards, or stops, this
synthesis of protein, retards or eliminates the accumulation process (cf. effects
of CO2/HCOjJ~, [Ca++], prior storage at low temperature).
(xi) In mediating the salt intake the role of protein synthesis was conceived
to involve a carrier molecule. If this molecule (say a phosphorylated glutamine)
could bind (or hold) anions and cations, transport them across the cytoplasm,
then, if it were condensed to protein, it might leave the ions in a situation in
which they might be readily accumulated in the vacuole (Steward & Street, 1947,
p. 496; cf. Franck & Meyer, 1947).
(xii) The scheme (Steward, 1935) which attempted to relate the activity of
ion-absorbing systems to their growth and development reached the following
conclusion. If it were possible to deal with cells that were actively dividing and
multiplying, as in a meristem, the concentrations of the solutions they contained
would be greater than at any other time, although the amounts per cell would be
limited by the small volume of the aqueous phases in strictly dividing cells.
This seemingly pardonable extrapolation of the data then available was based
on such facts as that approaching the root apex the concentration of ions absorbed
in the total water of the cells steadily increased. Also bromide is accumulated
in potato cells as they reapproach the conditions in which cell division may occur
and, if the cells are treated in such a way that the power of division is eliminated,
the power of accumulation disappears also. Slices of potato tissue in which
protein breakdown occurs even lose their solutes to an external solution.
In this state of affairs a new approach to the problem was required.
Before describing this, reference should be made to the work of Sutcliffe
OF THE ROLE OF GROWTH AND METABOLISM 371
(1952) since this followed the general pattern of the work summarized
above.
After exposing beetroot disks to three changes of distilled water at hourly
intervals they were treated with daily changes for periods up to 8 days. In
consequence the K content of the tissue declined to half its original value. On
subsequent transfer to relatively strong solutions (up to 0-04 M-KC1) K was
reabsorbed in such a manner as to suggest
(i) The initial rate of uptake was a function of the salt and solute deficit
created by the long washing.
(ii) The final plateau of K content attained after many days was relatively
unaffected by either the concentration of the KC1 or the initial deficit due to the
long washing but was in fact the maximum that these cells could contain.
(iii) This intake of K, clearly induced by the long treatment, was cyanide
inhibited and the tissue acquired a more conspicuous salt-induced, cyanide
sensitive, component of respiration the longer it was pretreated with distilled
water. There was, however, no simple relation between K absorbed and salt-
induced respiration. (The respiration of beet tissue in water was surprisingly
sensitive to cyanide.)
The fundamental feature of these experiments, however, still evades us. What
were- the metabolic consequences of the long washing with distilled water
(cf. Steward & Preston, for the potato tuber, 1940)? What happened to all the
other cations and the various organic electrolytes and non-electrolytes (cf.
Steward, 19320)? Is it conceivable that, without obvious growth, the tissue
would increase its total K content threefold over its initial value unless :
(a) The pretreatment depleted the cells of salts and solutes.
(b) Placing the tissue in such a strong solution (0-04 M-KC1) permitted it to
embark upon a vicarious intake of salt, unaccompanied by growth, till the cells
readjust to this applied concentration. This is an effect comparable to that
induced in Hoagland's 'low salt' roots though here it is due to a superimposed
high concentration of salt.
(c) With the elapse of time the metabolic characteristics of the cells changed
in response to aeration. It is still possible, lacking evidence, that these changes
were in the direction of growth (e.g. protein synthesis) even though they did not
proceed to actual cell division or proliferation.
The objective of the new approach was to find the point of contact in the
metabolism of growing cells between the diverse processes of water and
salt intake on the one hand and respiration and protein synthesis on the
other. Two main requirements seemed to be apparent.
First, biochemical techniques were required by which to fractionate the
nitrogen compounds in the hope that the intermediary metabolism involved
in the protein synthesis might be traced. The technique that had emerged
and was adaptable to this end was the qualitative and quantitative technique
of paper chromatography (Steward & Thompson, 1950; Thompson,
Zacharius & Steward, 1951; Thompson & Steward, 1951). In this way,
24-2
372 SALT ACCUMULATION IN PLANTS: A RECONSIDERATION
the nitrogen compounds of the potato-tuber tissue have been described and
some investigations have been made of the changes which occur under
conditions conducive to salt intake (Steward & Thompson, 1950).
The second requirement was quite different, for one needed a system
in which cells could be placed at will in either of two contrasted states,
namely, the proliferating-dividing condition or the non-dividing condition.
The search for such a tissue system prompted a re-examination of tissue-
culture techniques. Again the detour proved rewarding beyond expecta-
tions, but, from the standpoint of this review, the desired system was found
in the use of standard explants of secondary phloem from the carrot root.
Under the conditions described (Caplin & Steward, 1948, 1949) such
explants will either remain under sterile conditions with only sluggish
expansion for a long period of time, or, if supplied with nutrients, vitamins
and the growth factors that are contained in coconut milk, they will embark
on a most remarkable period of active growth by cell division.* In the
coconut-milk factor, or factors, lay the secret of the transition from the
relatively inactive non-growing state to the actively proliferating, growing
condition. Here then was the type of system desired for a study of the kind
of salt absorption which is characteristic of the cells in these two states
and which would enable one to recognize the metabolic features which, in
these cells, are linked with salt accumulation. The mechanical devices by
which these ends were achieved have been described and may, therefore,
be merely referred to here (Steward, Caplin & Millar, 1952).
II. ABSORPTION OF 137Cs BY GROWING AND NON-
GROWING TISSUE CULTURES
Here a confession is necessary. The first expectation was that the actively
growing, proliferating carrot-tissue cultures, which increase in fresh weight
some eighty times in 20 days, would, if supplied with an ion that could be
absorbed, absorb much more actively than their relatively non-growing
analogues whose weight, though maintained, was only increased slightly
in comparison. By the use of the radioactive isotope 137Cs it was possible
to investigate this question quite accurately, and this contribution is
especially associated with Miss Millar (1953). The unexpected and dramatic
result was as follows : per unit weight the relatively non-dividing cells of
the carrot explant absorbed more caesium than the dividing cells of the
growing cultures (see Table i). This result was too striking to be ignored.
* Using maceration techniques that he has applied to the interpretation of the growth
of roots, Dr R. Brown, working with one of us (F. C.S.), has examined these carrot
cultures with the following results. The initial tissue explant weighing 2'6 g. contained
about 25,000 cells. After about 13 days the number was over i x io6 if the explant was in
the medium containing coconut milk under the standard conditions referred to above.
OF THE ROLE OF GROWTH AND METABOLISM 373
Table i . Fresh weight and absorption of 137Cs by proliferating and
non-proliferating carrot explants in carrier-free solution
Days after inoculation ...
o
2
4
6
10
H
*Growing cultures, mg.
4-0
5'7
7'3
ii'5
31-8
87-0
fresh weight
Counts/sec. /mg. fresh
—
1-14
2-17
2-85
2'55
2-14
weight
'(•Non-growing cultures,
4-0
5*i
5'9
6-3
7'2
8-4
nig. fresh weight
Counts/sec. /mg. fresh
—
2-37
4-87
8-17
17-0
30-4
weight
* Rapidly proliferating cultures in basal medium + coconut milk.
t Cultures which only very sluggishly expand m basal medium only without coconut
milk.
The 137Cs could be obtained and used in carrier-free solution (io~8 mol./
1.), and by addition of non-radioactive caesium (at io~3 mol./l.) one could
vary the total caesium concentration over a very wide range ( x io5), keeping
the concentration of radioactive caesium constant. Again it was found that
the relationship of these two types of absorbing system, the growing and
the non-growing, to the total concentration of caesium were quite different.
The growing-dividing cells behaved again in a quite unexpected fashion
(Table 2).
Table 2. 137Cs absorption ratio* as a function of time for proliferating and
non-proliferating carrot-tissue cultures with and without the addition of
inert carrier caesium
Days after inoculation
4
7 io
14
^Growing cultures with carrier caesium
7-30
9-14 : 14-1
I3H
absorption ratio
Growing cultures without carrier caesium,
9'43
10-7
12-9
H'5
absorption ratio
JNon- growing cultures with carrier caesium
27-2
53'i
77'7
79'5
Non-gi owing cultures without carrier
375
649
2050
2130
caesium
* Absorption ratio is counts/sec. /g. fresh weight of tissue divided by counts/sec./c.c.
of external solution.
f Rapidly proliferating cultures in basal medium -f coconut milk,
j Cultures lacking coconut milk.
Table 2 gives the data for uptake of 137Cs in terms of its absorption ratio,
i.e. the concentration of 137Cs in the total water of the tissue divided by
the final concentration of 137Cs in the external solution. From the data in
Table 2 the surprising result was that the presence of a relatively large
excess of caesium ions over the 137Cs had very little, if any, significant
effect on the absorption of the 137Cs in the case of the growing-proliferating
cultures.
374 SALT ACCUMULATION IN PLANTS: A RECONSIDERATION
The results could only be interpreted on the very definite conclusion
that the caesium absorbed by growing-proliferating tissue cultures bore
a linear relationship to the external concentration, that is, to the first power of
the external concentration. Because that external concentration was varied
over such a wide range ( x io5) this result was most surprising.
In contrast the relatively non-growing tissue cultures, that is, those
lacking the coconut milk growth factor(s), and merely placed in a calcium
chloride solution, behaved in a more conventional fashion because the
relative uptake of 137Cs was greater from the more dilute solution. (In
other words, the accumulation ratio increased with dilution of caesium in
the familiar fashion.)
These results lead to the following conclusion. In the angiosperm plant
body the complete cycle of events from cell division, through the formation
of the vacuole and the complete enlargement to the mature cell, is charac-
terized by salt absorption which proceeds at two distinct stages by two
distinct types of mechanism.
The first of these (stage I) is characterized by, or is emphasized in, the
cell in the state in which it is capable of continuous and active cell division,
while it remains small ; in this state the cell absorbs its ions in such a way
as to suggest that they are bound on certain sites produced continuously
and in the process of this growth. The relationships to concentration clearly
suggest that this binding is a stoichiometrical one, the suggestion being that
positive ions like caesium are bound to negative, or acidic, locations. Such
a process would be expected to reside specifically in the cytoplasm where
the synthesis occurs.*
The second type of absorption mechanism (stage II), however, is quite
different, and it is characteristic, not of the dividing cell, but of the cell in
which division has slowed down and expansion of the vacuole is the chief
event. This type of absorption, now regarded as a process of secretion into
the vacuole, is characterized by the familiar accumulation mechanism which
causes the relative absorption to be greatest from the most dilute external
solution. Both of these absorption processes are related to time and
aeration in ways that necessitate that the metabolism of the cells determines
the intake of the ion.
However, we can now see that the extrapolation in the earlier scheme
of 1935 to the state of active cell division was not wholly justified. At the
point where growth ceases to be predominantly by cell division and becomes
predominantly by expansion, an abrupt change occurs in the nature of the
* Even in the ' bound ' state the ion may be ' accumulated * in the sense that the amount
present, expressed as a concentration in the total water of the cell, may be much greater
than in the external solution. Ratios of 8 to io were observed in these experiments.
OF THE ROLE OF GROWTH AND METABOLISM 375
absorption process. It is difficult to see how this important conclusion
could have been reached without the particular advantages presented by
the tissue-culture system to which reference has been made. The obvious
suggestion is that with the formation, or rather the enlargement, of vacuoles
the activity of the tonoplast intervenes, in the way that de Vries originally
visualized (1885), to cause secretion internally into the vacuole.
One may now ask what are the other metabolic characteristics of the cells
in these two contrasted states.
The dramatic metabolic effect produced by the coconut-milk growth
factor on the carrot-tissue cultures is the marked stimulation to protein
synthesis which accompanies the cell division. Cells of the resting carrot,
like those of potato tuber, contain a relatively large percentage (>5O%)
of their total nitrogen in the form of soluble nitrogen compounds.* In the
cells cultured on coconut-milk media this is not so, for the bulk (70%) of
the nitrogen is now present in the protein form (Steward et al. 1952).
Much more striking, however, is the fact that the proportion of the different
substances in the alcohol-soluble nitrogen fraction is quite different in the
growing culture and in the relatively non-growing culture. This is particu-
larly true of the amide, asparagine, which is conspicuous in the non-
growing cells and is either absent, or is very much reduced, in the growing
cells.
The same sort of result has been demonstrated by the use of small
explants of potato tissue which, however, require the intervention not only
of coconut milk but of 2, 4-D, or an analogous substance, in dilute solution
to make them grow (Caplin & Steward, 1951). One can, therefore, relate
the first kind of ion absorption which occurs in growing-dividing cells and
which is proportional to the first power of the external concentration of the
absorbed ion, to cells in a particularly active state of protein synthesis and
cell division and, therefore, of multiplication of their self-duplicating units.
Accompanying the transfer from the resting to the dividing state the
composition and metabolism of the tissue changes profoundly. Even the
protein that is synthesized in the dividing cells is recognizably different,
for it may be shown by quantitative paper chromatography that it contains
hydroxyproline, unlike the protein of resting cells, and contains other
amino-acids in somewhat different proportions. Moreover, it seems
permissible to regard the binding sites which take up the ion in question
as being formed continuously as the new protein is synthesized. So long
as the coconut-milk factor is present the carrot cells are arrested in their
* The data on nitrogen metabolism of tissue cultures were communicated by Steward
& Thompson to the American Society of Plant Physiologists at Minnesota, September
1951, and they will appear in thej. Exp. Bot.
376 SALT ACCUMULATION IN PLANTS: A RECONSIDERATION
differentiation in the sense that they continue to divide but they do not
fully enlarge and vacuolate, nor do they pass to the state of maturity in
which cell divisions would not occur.
Therefore, what the carrot-coconut milk system does is to enable one
to separate the two steps (stages I and II) of the absorption process giving
emphasis, at will, to the kind of absorption which is characteristic of the
growing-dividing plant cells or to that which is characteristic of the cells
whose further growth is mainly by enlargement, rather than by division.
Metabolic inhibitors have been used to characterize even further the
differences between cells in these two contrasted states (Steward & Shantz,
1951). The effect of various inhibitors upon the growth of carrot cells in
coconut-milk media is shown in Table 3. The striking thing is that the
growth of the carrot in presence of coconut milk is so insensitive to cyanide.
From what is known of the effect of cyanide upon respiration this would
seem to imply that much of the respiration proceeds over pathways which
do not directly involve cytochrome, or other enzyme systems catalysed by
heavy metals.
Table 3. Effect of enzyme inhibitors on growth (mg. fresh weight per culture)
of carrot-tissue explants* during 21 days in medium + coconut milk
Mean fresh weight (mg.) at
specified
Inhibitor
cone, of
inhibitor
IO 3M IO'4M ! IO~GM
Fluoride
132 141
161
Cyanide
147 134
117
Dimtrophenol
10
ii
79
Dimtrocresol
4
5
13
* Weight of initial explants = 4*0 mg. Weight of explants 21 days in basal medium
only = 7-0 mg. Weight of explants 21 days in basal medium + coconut milk— 137 mg.
The effects of a long (7 days) exposure to cyanide on both the non-
growing cultures in calcium chloride solution and the growing cultures in
basal medium -f- coconut milk are shown by its effect upon 137Cs uptake.
The data are shown in Table 4.
These data show a relatively greater sensitivity to cyanide in the case of
the ion intake typical of the non-growing cultures than of that typical of
the growing ones.
The effects of cyanide upon the respiration of the tissue in the two states
have not yet been fully worked out. It is clear that the tissue which grows
in media containing coconut milk, whose growth and ion intake are alike
relatively insensitive to cyanide, produces carbon dioxide by pathways
that are largely insensitive to cyanide. The tissue of the experiment in
Table 4 which had been treated for 7 days with potassium cyanide when
OF THE ROLE OF GROWTH AND METABOLISM 377
removed to a Warburg apparatus, produced carbon dioxide at the rate of
0-337/^1. Og/mg./hr. while still in the presence of io~4M-KCN. Comparable
tissue not treated with cyanide respired at the rate of 0-372/4!. O2/mg./hr.
Therefore, the respiration of the growing tissue is only suppressed by
to the extent of about 10%.
Table 4. Uptake of 137Cs by carrot tissue explants as
affected by lo^M-CN during 7 days
Type of culture
Medium
137Cs uptake per day
(counts/sec. /mg.)
% Inhibition
by CN
CN-treated Control
Non-growing
Growing
CaCl2
Basal medium -f-
coconut milk
r1
0-363 | 5-36
0-0225 i 0-0259
i
93-2
i3-i
The respiration of normal carrot tissue as freshly excised from the
carrot root is well known to be cyanide-sensitive (Marsh & Goddard, 1939).
Our own data show that even after 2 days under the culture conditions
described the respiration was inhibited to the extent of 40% by io~4M-
KCN. Therefore, the ion intake which occurs in the non-dividing tissue is
associated with a respiratory metabolism which is markedly cyanide-
sensitive, particularly in freshly explanted tissue. Whereas the growing cell
in the presence of coconut milk is relatively less sensitive to cyanide, as
compared with the non-proliferating tissue in the absence of coconut milk,
the converse is true in the case of those inhibitors that affect phosphoryla-
tion. Such enzyme inhibitors as the nitrocresols markedly inhibit the
growth of the proliferating cultures, and their effect on ion intake will now
be examined (Table 5).
The effects of treatment with nitrophenolic inhibitors were tested on
cultures which grew in full nutrient -f coconut milk for 21 days but which
received the inhibitor at two stages, 0-2 and 7-9 days. The effect of the
Table 5. Effect of dinitrocresol on growth and subsequent 137Cs uptake by
carrot-tissue cultures grown 21 days in presence of coconut milk
All relative data for the inhibitor treated cultures expressed as a % of the untreated.
Cone, of
inhibitor
Period of contact
with inhibitor Period of contact with inhibitor
o— 2 days
% inhibition
7-9 days
% inhibition
Rel. F.W.
Rel.
137Cs
uptake
F.W. 137Cs
Rel. F.W.
Rel.
137Cs
uptake
F.W.
137Cs
to-4
io-5
28-9
93'5
70
106
70 1 30
81-1
34'7
90-2
96
65
378 SALT ACCUMULATION IN PLANTS: A RECONSIDERATION
inhibitor is recorded in terms of the total growth (mg. fresh weight) and
the 137Cs uptake which occurred subsequent to the treatment in comparison
with the uninhibited controls. The data show that marked inhibition
occurred as a result of contact with lO^M-dinitrocresol but that io~5M was
much less effective. They also show that the inhibitor applied at 7-9 days
reduced both the fresh weight and the 137Cs uptake proportionally much
more than when it was applied to the tissue at 0-2 days. (There is a lag
period of approximately 4 days after inoculation, during which little or
no external growth of carrot cultures in coconut milk occurs.) The relative
insensitivity of the carrot cultures exposed to cyanide in the early period
(0-2 days) is therefore to be associated with conditions in the tissue before
it has fully responded to the coconut-milk stimulus.
Therefore, cyanide inhibition of metabolism and ion intake is typical of
the freshly explanted tissue and of the non-proliferating tissue, i.e. tissue
not treated with coconut milk. On the other hand, nitrophenolic inhibition
of metabolism and of ion intake is typical of the cultured tissue in media
that contain coconut milk. Furthermore, the effect of the nitrophenols on
the tissue in coconut milk is most marked after time has elapsed and the
tissue has fully responded to the coconut-milk treatment.
Both of these salt-absorbing mechanisms should be regarded as depen-
dent upon oxidative metabolism. However, the first stage (stage I) of the
absorption process, i.e. the one in which ionic binding is mainly in question,
is closely linked to protein synthesis and to multiplication of cells and of
their self-duplicating units. Although this mechanism is relatively cyanide-
insensitive, phosphorylation seems to be necessary for its maintenance,
since it is more sensitive to nitrophenols.
In the second type (stage II) of the salt-absorbing mechanism, where any
concomitant protein synthesis is unaccompanied by cell proliferation,
cyanide inactivates the mechanism. Here the prime event is the attainment
of relatively high concentrations in the total water of preformed cells, and
this involves the removal of the ions from their initial binding sites and their
secretion into the vacuolar fluid.
III. ABSORPTION OF 137Cs BY ARTICHOKE TISSUE:
GROWING AND NON-GROWING
Proof of these two types of absorption process by reference to another
tissue system was clearly desirable. Use was therefore made again of tissue
from the Jerusalem artichoke tuber. In passing from its initial high rate of
respiration through a time drift (when respiration falls to an eventually low
level) cut disks of artichoke-tuber tissue pass from an active absorbing
condition to a less active condition (Steward & Berry, 1934).
OF THE ROLE OF GROWTH AND METABOLISM 379
If coconut milk is added to artichoke tissues which, with the lapse of
time and through exposure to solution, had become adjusted to a low rate
of respiration (order of 0-08 mg. CO2/g./hr. in contrast to 0-25 mg./g./hr.
for fresh-cut disks), the cells (like those of the carrot) will return to the
actively dividing state and will grow, and their rate of metabolism is
increased. (Prior to our knowledge of this effect of coconut milk all
attempts to raise the respiration of artichoke disks, which had passed through
their time drift in distilled water, to their initial high level and to maintain
this by the use of such added metabolites as phosphate, nitrate, sugar,
amides, etc., had failed (Ramamurti, 1938). The effect of the coconut milk
is, therefore, a highly specific one.)
One might have expected that the coconut milk and its attendant growth
would cause the caesium uptake per unit weight of these cells markedly to
rise. On the contrary, the tissue now takes up less caesium per unit weight
per unit time than the controls which lacked the coconut-milk growth factor
during the period of absorption. The data of Fig. i show that these effects
attributable to coconut milk may be brought about at will in the artichoke
tissue. This and other results confirm, for the case of the artichoke tissue,
that the new ideas on salt intake which have been described apply also to
these cells as well as to those of explanted carrot-root phloem.
The tissue lacking coconut milk absorbs 137Cs in the manner of the
growing but non-dividing cells (Text-fig, ic), its ion intake decreases
during the time drift with the respiration in the manner previously referred
to for bromide. However, by addition of coconut milk at points far along
the time drift, e.g. 8 days, the cells change their metabolism, grow and,
thereafter, absorb less, rather than more, 137Cs per unit of fresh weight
(Fig. i c). It should be noted that on passing from the non-dividing to the
dividing state the tissue produces much more carbon dioxide but absorbs
much less 137Cs (cf. Text-fig, i B, C).
The comment may well be made that these experiments have not simpli-
fied the problem, they have only complicated it. Instead of one mechanism
of salt intake and accumulation we now have two. And each of these
mechanisms bears its characteristic relationship to the metabolism that is
involved, particularly to protein synthesis. Also it may rightly be claimed
that it is hardly a simplification to regard an explanation of salt intake as
carrying with it the need also to explain something about the perhaps still
more mysterious process of protein synthesis.
Therefore, a further digression must be made, but this will be justified
because it leads to a possible picture of the events that may underlie the
kind of result that has been described.
B. Carbon dioxide production per hour
C 137Cs absorption per day
16
18
20
8 10 12 14
Time (days after inoculation)
Text-fig, i . Effects of time and coconut milk on fresh weight, carbon dioxide production
and 137Cs absorption of artichoke-tuber tissue cultures,
medium -f coconut milk; O O cultures in calcium chloride;
ferred at 8 days ( f ) from calcium chloride to basal medium -f- coconut milk.
cultures in basal
cultures trans-
THE ROLE OF GROWTH AND METABOLISM 381
IV. PROTEIN SYNTHESIS AND ION BINDING:
A TEMPLATE HYPOTHESIS
The task now to be attempted is as follows: Some picture should be formed
of how protein synthesis may occur at the expense of the soluble nitrogen
compounds. This picture should visualize the central role which glutamic
acid may play as the starting-point from which a variety of amino-acids
may be derived by transamination. It should also give prominence to the
role which the amide glutamine seems to play, for it is related to protein
synthesis in ways that suggest that it is peculiarly able to donate nitrogen
to the protein synthesizing surface, so that one may visualize the soluble
nitrogen as being canalized through the form of glutamine. We should also
be able to visualize how, by the aid of respiration, both protein synthesis
and ion intake are enabled to proceed so that the nitrogen for synthesis,
the energy for synthesis and salt accumulation and the ions to be accumu-
lated seem to be presented to the active surfaces simultaneously and in
a compatible form.
The possibility exists that the relation of protein synthesis to ion accumu-
lation may concern the entropy changes. Formation of highly ordered large
molecules from disordered amino-acid moieties will cause the kind of
entropy change that might give a positive change of free energy if the
accompanying heat change is small.
Bounce (1952) has outlined the template hypothesis in a detailed manner.
He has visualized the chemical reactions by which a nucleic acid surface
could reduplicate itself, and he has also visualized how it could permit the
amino-acids to be combined with the template and finally removed again,
or * peeled off ', in such a way that when they are removed protein is formed.
First Bounce visualizes that the nucleic acid surface is activated by
phosphorylation. By amino-transphorases the amino-acids are induced to
combine with the nucleic acid surface. It is thought that when the protein
is removed the adjacent amino-acids are induced to form the a-peptide
links that exist in the protein. This is shown in Text-fig. 2.
Hanes, Hird & Isherwood (1950, 1952) and Hanes, Connell & Bixon
(1952), however, have elaborated another idea. Glutamic acid and
ammonia, with energy supplied from ATP, itself regenerated by respira-
tion, can form glutamine (Speck, 1947; Elliott, 1948). Glutamine can be
regarded as a hydrogen peptide in which the energy of respiration produces
the amide linkage. Hanes has drawn attention to the fact that the single
most important step in protein synthesis may well be the formation of the
y-glutamyl bond which can be built into a series of y-glutamyl dipeptides.
This process requires the intervention of systems which incorporate
382 SALT ACCUMULATION IN PLANTS: A RECONSIDERATION
glutathione.* Enzymes exist in both plants and animals by which the
glutamyl moiety of glutathione may be split off and combined with a variety
of amino-acids to give y-glutamyl dipeptides. The residue cysteinyl-glycine
may then be available to regenerate glutathione, by using again the energy
of respiration and by a process not dissimilar to that which has been
described for glutamine. So the cycle could go around again.
The importance of this mechanism is that it visualizes a means by which
a variety of y-glutamyl dipeptides may be formed, and to do this one needs
catalytic amounts of glutathione and the means, through respiration and
phosphate band energy, to regenerate the glutathione from the cysteinyl-
glycine residue and glutamic acid. However, the apparent dilemma is that
the variety of peptides so formed do not incorporate the a-amino-peptide
linkage which is actually necessary for the formation of protein. (The
dilemma may of course be overcome by assuming that once energy is
incorporated into these y-dipeptide linkages the appropriate a-dipeptide
link may be formed by some intramolecular rearrangement.)
Here, however, one can bring the ideas of Bounce and of Hanes together.
If it is assumed that the amino-acids were presented to the template,
consisting of a pentose nucleic acid surface, not in their free state, but only
in the form of these y-glutamyl peptides then some interesting conse-
quences could follow. First, one could dispense with the need to phos-
phorylate the nucleic acid surface, because the energy inherent in the
y-glutamyl bond would suffice to enable the amino-acid to combine with
the nucleic acid surface. The glutamyl residue would then become available
for resynthesis to glutathione and so permit the reversible cycle to continue.
By this means all the amino-acids could be presented in the form of the
same linkage, the y-glutamyl linkage, for which the nucleic acid surface
could be regarded as a dipeptidase or transferase.f In fact, according to
Brinkley (1952), the hydrolysis of peptides by nucleic acids has already
been demonstrated, even though they are virtually free from a protein
moiety. Since some synthetic resins are also now known to hydrolyse
protein (Underwood & Deatherage, 19520, b) this does not appear as
improbable as it might otherwise have seemed to be. The remaining process
of removing the amino-acid from its combination with the nucleic acid
surface and, in the process, the formation of a-peptide links remains
exactly as under the Dounce hypothesis.
* Brinkley (1952) visualizes an activated substrate such as ' y-glutamyl-coenzyme-A '
as the actual donor of the glutamyl residue, so that glutamine and glutathione are reserves
of this radical.
f It is true that the transfer in question, would be of the 'arnine transfer' kind
(Hanes et al. 1950), which though possible, is not yet demonstrable with isolated enzyme
systems.
OF THE ROLE OF GROWTH AND METABOLISM
383
\ z
kzX /
I
O
it
OH I
\ /
: (j u z u 2
\/\/\/\/xii:
N
tX)
384 SALT ACCUMULATION IN PLANTS: A RECONSIDERATION
Thus, one can now regard the role of glutamine and glutamic acid in
protein synthesis in the following way. Glutamic acid is the acid which,
more than any other, enables the amino groups to be handed on to other
keto-acids through transamination, to generate, in the variety required, all
the amino-acids necessary for protein synthesis. This would leave a-keto-
glutaric acid, which would either be drawn into the Krebs cycle and
respired away or regenerated to glutamic acid by nitrogen drawn from
soluble nitrogen compounds that are not immediately available for protein
synthesis. By the aid of respiratory energy, glutamic acid could form
glutamine and by similar processes glutathione could also be produced.
Through the sequence already described glutamyl peptides could be
formed in almost any variety, so that glutamine would become the agent
through which, in the necessary variety and proportions, the various
amino-acids could be presented to the protein synthesizing surface and,
in so doing, the peptide would bring up, not merely the amino-acid, but
also the energy in the y-glutamyl bond necessary to combine it with the
nucleic acid template.
To relate these ideas to salt accumulation one needs to make still one
more hypothesis. If instead of regarding glutamine itself, possibly
phosphorylated, as the carrier molecule of ions (cf. Steward & Street,
1947), transporting them from the external surface to the place where
protein synthesis occurs, one could now regard the y-glutamyl-peptides
as being the carriers, some interesting consequences follow.
Such peptides would have, or could have, free acidic groups that would
bind cations and free amino groups which could bind anions. In the form
of the complex the ions could be transported across the cytoplasm. If one
regarded the seat of synthesis as in the vicinity of the tonoplast, then the
ions with the amino-acids would be bound to the nucleic acid surface, as
to an ion-binding resin when by the dipeptidase action, the amino-acid
residue was fixed to the nucleic acid surface with the liberation of
glutamine.
So long as new nucleic acid surfaces were being created, i.e. so long as
self-duplicating surfaces are being multiplied, and so long as the amino-
acid could be fixed and remain bound in this way, then no free accumulation
in the vacuole would be expected to occur. This is visualized as the
condition which obtains in the growing, dividing cell as exemplified by the
cells stimulated to divide by coconut milk (stage I). Probably similar
circumstances also obtain in meristematic cells proper.
However, when further synthesis of nucleic acid ceases, and if protein
synthesis is to continue, the bound amino-acids should be * peeled off* in
the form of free protein. In doing so and at one and the same time, the
OF THE ROLE OF GROWTH AND METABOLISM 385
a-peptide links are to be formed and the sites hitherto available to bind
anion and cation disappear. If one visualizes this event as occurring at, or
near, the tonoplast surface it seems a not improbable outcome that the
ions thus freed could be moved into the vacuole along a very short, but
very steep, diffusion path or they could be actually released on the side
toward the vacuole. Such a hypothesis admittedly involves much specula-
tion. However, the hypothesis has the attraction of bringing together so
much otherwise unrelated evidence that it is here presented in outline in
the form of the following charts (Text-figs. 3, 4). Knowledge of the intra-
cellular location of pentose nucleic acid is still limited. Caspersson (1950)
stresses that pentose nucleic acid congregates at the nuclear membrane
surface and here protein synthesis occurs but from the work of Commoner
(1950) on plant cells it is not impossible that pentose nucleic acid may
accumulate at cytoplasmic surfaces.
These ideas bear obvious resemblance to views held on the events that
occur with contraction and expansion in muscle. In the resting muscle,
with its expanded fibres K+ and PO|~ are conceived to be bound in non-
dialysable form, at negative and positive sites respectively. After the muscle
contracts, energy being required to bring this about, K+ and POl~ are
liberated. Thus at the protein surface ions are alternately bound and
released concomitantly with expansion and contraction (Haurowitz, 1950,
cf. p. 164).
Some other evidence lies behind this train of thought. Work on the
effects of mineral deficiency in relation to nitrogen metabolism has
emphasized the striking consequences of growing plants without sulphur.
Under these circumstances protein synthesis is arrested, soluble nitrogen
compounds accumulate and (in Mentha) among these soluble nitrogen
compounds the two which attain very great concentrations in the cells are
glutamine and arginine. This is consistent with a mechanism in which
sulphur, as glutathione, is required to catalyse protein synthesis and,
lacking sulphur, the mechanism becomes blocked at the point where some
of the principal products that accumulate (glutamic acid ^glutamine)
would otherwise give rise to the variety of amino-acids required to give
protein.
However, for ion intake the crucial evidence would be the recognition
that 137Cs could be bound in some organic carrier molecule that could be
detected. Following a suggestion from the work of Bolton (1950) and
Roberts et al. in a private communication, chromatography of anhydrous
methanol extracts, obtained from lyophilized tissue cultures grown in
media containing coconut milk and 137Cs, showed that the caesium exists in
at least two states which differ in their mobility in methanol on paper.
E B S VIII 25
386 SALT ACCUMULATION IN PLANTS: A RECONSIDERATION
Spraying the methanol chromatogram with ninhydrin produced the pattern
shown at Text-fig. 5. The more mobile 137Cs component was closely
superimposed by the pattern of ninhydrin activity in segment 2.
Glutamic + NH3
Glutamic + cysteinyl-
glycine
Glutamine + cysteinyl-
glycine
ATP
ATP
Glutamine
Glutamyl-cysteinyl-glycine
Glutathione
Glutamyl-cysteinyl-glycine -f amino-acid (Rlt
Glutamyl-*! Cysteinyl-glycine
Glutamyl-/?2
Glutamyl-/^
Glutamyl-7?n
(y-Glutamyl peptides)
Pentose-nucleic acid
Template ©
^e
+ Glutamic "^ glutamine
+ NH3 + ATP
ct°
OH
ct°
I OH
HONH,
HONHa
I cation I
CHa binding CHa
CHa COOTL CHa COOH
C— NH— CH i— NH-CH
PNA
Ions free at the tem-
plate surface
Peptide removed with for-
mation of a-amino-acid
linkages
NH
*** Anion binding sites
y-Glutamyl-peptides may present to synthesizing surface:
(i) Amino-acids
(ji) Energy in amide linkage
(iii) Associated ions :
Anions bound to basic amino-acids
Cations bound to more acidic amino-acids
Text-fig. 3.
Similar segments i and 2 cut from the chromatogram were extracted with
water and rechromatographed on a 5 x 5 in. paper using phenol : collidine-
lutidine. Other segments from positions well above and below segments
i and 2 were also treated in this way. These two directional chromatograms
OF THE ROLE OF GROWTH AND METABOLISM
C O
WCQ
25 2
388 SALT ACCUMULATION IN PLANTS: A RECONSIDERATION
were then sprayed with ninhydrin and also treated with the starch-
chlor-iodide method of Rydon & Smith (1952).
Although some traces of ammo-acids were present in the original paper
(washed only with methanol), the following results were quite clear.
Anhydrous
methanol
1
Segment 1
Sample
origin *
Weak
ninhydrin *
reaction
Strong w
ninhydrin
reaction
Segment 2
Text-fig. 5. One-dimensional chromatogram in anhydrous methanol cf a methanol extract
of lyophilized carrot-tissue culture, i, ninhydrin reactive compounds; 2, 137Cs.
Segment 1
r
Segment 2
~\ r
~i
| xFront
3
Phenol _J L_ X Front
-o
«3
J3
C
70
Phenol _'
Text-fig. 6. Identification of ninhydrin reacting compounds that move in methanol
alongside 137Cs. Phenol : collidine-lutidine chromatogram of eluates from segments of
methanol chromatogram.
When the two 137Cs containing segments (i and 2) of the methanol
chromatograms are extracted and the extracts chromatographed in aqueous
solvents both give one and the same 137Cs spot (Text-fig. 6). Therefore
there is the possibility that 137Cs existed in the dry methanol extract as an
organic complex, which broke down in the extraction and the more acid
aqueous phenol. In this case one product of the decomposition could be
aspartic acid from segment 2, for this substance appeared on the papers of
Text-fig. 6 in amounts far beyond any possible error, and it now moved
OF THE ROLE OF GROWTH AND METABOLISM 389
independently of the 137Cs. Segments lower than 2 also showed that there
were more mobile amino compounds in the methanol extract. The
compound B detected by the starch-chlor-iodide test was not associated
with the 137Cs because it also appeared in the segment below no. 2.
As additional controls 137Cs was chromatographed in methanol in
presence of other salts (KC1, KNO3, K2SO4) and a mixture of amino-acids.
137Cs did not give two sharply defined spots (cf. Text-fig. 5) in any of these
cases, and it ran more slowly than, and independently of, all the amino-
acids and ninhydrin reacting materials.*
Therefore, the probability exists that 137Cs occurred in the methanol
extract of lyophilized carrot-tissue culture as a complex, decomposable by
water, with a ninhydrin reactive material of which one product at least
could be aspartic acid.
These observations are mentioned because they suggest lines along
which further investigations of possible carriers of ions undergoing salt
accumulation may be sought but not to claim that the present evidence
alone is more than suggestive.
V. RESPIRATION AND SALT ACCUMULATION:
THEORETICAL CONSIDERATIONS
If is not proposed to review all the recent papers that bear upon the relation-
ship of respiration to salt accumulation. The following observations, how-
ever, are meant to relate the work here described to the interesting work of
Robertson and of Lundegardh to which it is so closely related.
The work of Robertson (1951 and references there cited) has been done
with carrot-root tissue, mainly secondary xylem. After long washing and
in nutrient-free solution the capacity of the cells for further growth
and anabolism (protein synthesis) will be limited. The tissue as used by
Robertson would seem to be in a somewhat comparable state to the
carrot-phloem tissue which lacks the growth-promoting factors of coconut
milk, i.e. it has a limited ability to grow, but this will be predominantly, if
not exclusively, by enlargement of preformed cells. Early in the time drift
the respiration is high and comparable with that of potato disks; later in
the time drift it falls to a lower value. In this respect the behaviour of the
carrot root resembles that of artichoke tuber (Steward & Berry, 1934).
According to Robertson the respiration early in the time drift, i.e. at the
initial high rate, is markedly cyanide-sensitive; later in the time drift it is
less so.
* Adding 137CsCl to a methanol extract of lyophilized tissue gave the same chromato-
graphic results as if the tissue had absorbed the CsCl. Therefore, the Cs-containing
complex can be formed in an anhydrous extract of tissue.
390 SALT ACCUMULATION IN PLANTS: A RECONSIDERATION
Robertson has used the carrot tissue in this particular state, i.e. after it
has adjusted to a steady but low level of metabolic activity, as the means to
investigate the relation of respiration to salt intake. Certain features of this
technique require to be noticed.
(1) The single salt solutions to which the tissue is exposed are relatively
strong (up to 0-05 M) — these contrast with the very dilute solutions from
which 'accumulation' often occurs.
(2) The determination of ion intake is almost wholly by the non-specific
conductivity method.
(3) Where the papers, e.g. Robertson, Wilkins & Weeks (1951), Text-
figs. 2, 3, refer to * accumulation ' of ions the data clearly relate to the amount
of ion (or rather salt) absorbed from the external solution. Since the internal
concentrations of the ions in question were not usually determined and are
often considerable (as, for example, potassium, calcium, chloride) they
would modify greatly the estimate of * accumulation ' (and therefore the
work done), using the term 'accumulation' in its conventional sense.
With these points in mind the following features of Robertson's investi-
gations on carrot tissue may be noted:
(1) Having adjusted to a low rate of respiration the tissue responds to the
presence of a relatively high concentration (0-05 M) of salt by an increase in
respiration which is steep at first. In the case of potassium salts the high
level reached tends to be maintained; in the case of calcium salts it more
quickly declines.
(2) After the initial increase the salt-induced respiration and the absorp-
tion of salt proceed concomitantly. (One cannot really tell from this alone
that 'salt respiration' is causally related to the salt absorption, for it may
be possible to devise situations in which the absorption of ions occurs with-
out recourse to measurably increased respiration, as, for example, in the
case of 137Cs in very dilute solution.)
(3) The salt-induced respiration is cyanide- and carbon monoxide-
sensitive, so these inhibitors retard both respiration and salt uptake by the
tissue in this stage.
(4) The effects of dinitrophenols (loc. cit. pp. 258 et seq.) are, however,
in this respect somewhat anomalous because they increase respiration
without concomitant increase in salt uptake and may even, on the contrary,
cause it to decrease.
(5) Using the tissue in the condition described, Robertson compared the
oxygen uptake to the salt absorbed in the following way. For every
molecule of oxygen absorbed, 4 atoms of hydrogen would ultimately be
transferred, over pathways mediated by the appropriate dehydrogenase
and cytochrome oxidase as terminal oxidase. Each electron transfer would
OF THE ROLE OF GROWTH AND METABOLISM 391
ultimately result in a hydrogen ion. The supply of hydrogen ions so
produced furnishes the absorbing power for the cations to be absorbed. The
basic scheme is that of Lundegirdh (1945), which visualizes the salt-
induced respiration that is involved as proceeding over an iron-catalysed,
cyanide-sensitive, respiratory system.
The data here reported on carrot and artichoke tissue — extending the
body of data on potato tissue and roots already published — present on this
view the following problems.
In dilute solutions (io~3 for KBr down to io~9 for 137Cs) the direct effects
of the presence in the external solutions of the ions to be absorbed are either
small or not measurable, as in the case of very dilute solutions of 137Cs. In
these cases, however, great accumulation (based on actual test of the tissue
and the external solution) does occur, and it is determined by time and
oxygen concentration in such a manner that it obviously proceeds pan passu
with the aerobic respiration and metabolism of the tissue as a whole — not
merely with a component of it due only to the presence of the salt.
In the case of both carrot and artichoke tissue there are clearly two
distinct relationships to respiration that are to be considered. In nutrient
solutions plus coconut milk, respiration is greatly stimulated and absorp-
tion of an indicator ion (137Cs) proceeds in time concomitantly with growth
and respiration, but the internal concentrations and the degree of accumu-
lation attained, despite the increased aerobic respiration, are reduced in
comparison with the non-dividing cells. Per contra, as the cells develop
and pass out of the dividing state, or by withdrawal or deprivation of the
coconut milk, though their respiratory intensity may decline, their attained
salt concentrations may increase. This is another though dramatic example of
the kind already noted (Steward et al 1943) which suggests that the nature
of the oxidative pathways by which the carbon dioxide emerges has a
profound effect on whether it is, or is not, linked to the process of ion
intake and accumulation. Previously stress has been laid on the fact that,
to be effective in promoting salt accumulation, the carbon dioxide, and the
oxidation by which it is produced, needs to be linked to protein synthesis.
The data here recorded also stress that there is a profound difference
between dividing, proliferating cells and cells whose growth is mainly, if
not solely, by enlargement of their vacuoles.
It is, therefore, not possible to frame a hypothesis covering all the facts
of ion accumulation based upon the direct intervention of oxygen uptake
or carbon dioxide output per se in producing salt accumulation. The
ultimate explanation requires a much more intimate understanding of the
reactions that lead up to the final emergence of carbon dioxide and the
transfer of hydrogen to molecular oxygen of the air, and of the way in
392 SALT ACCUMULATION IN PLANTS! A RECONSIDERATION
which these reactions are used by the dividing and by the non-dividing
cell.
Thinking in this general field has been influenced by the dramatic
advances made since Lipmann (1941) outlined energy transfer through
phosphorylation and, notably, by the knowledge that mechanical work in
muscle can be visualized to flow from the energy actually donated to
shortening muscle fibres as the muscle protein splits off phosphate from
ATP.
In plant cells the kind of salt-absorbing system is clearly determined by
the manner of growth of the cells in question (i.e. whether predominantly
by division and self-duplication or by enlargement). For carrot and
artichoke tissue, particularly, this is experimentally controllable by the use
of the growth factors in coconut milk. However, within each metabolic
pattern as thus determined there must be specific points, or reactions, at
which the energy for the salt accumulation is furnished in a milieu in which
the details of molecular architecture are conducive to it. But clearly there
are in the overall mechanism two distinct steps, or types of process
(stages I and II), and they each have a metabolic basis and a relation to
a distinct and definite phase of growth in the cell.
Stage I (promoted by the coconut-milk factor) is typical of the cell in
active division and in active multiplication of its self- duplicating parts. In
this case the metabolism and growth are characterized by cyanide-insensi-
tivity but great sensitivity to nitrophenols. Here the relation of ion intake
to metabolism seems to be that the overall respiration promotes the
synthesis of new ion-binding sites and in this sense respiration and salt
accumulation are linked. If the nitrophenol acts by uncoupling phosphory-
lation (Loomis & Lipmann, 1948) without which growth and protein
synthesis do not proceed, this is of more consequence to the mechanism of
ion intake than its effect on respiration alone, for it is of no avail that
carbon dioxide is produced, if the energy cannot be donated through
phosphorylation to perform useful work.
When the cell ceases to divide and expands its vacuole, accumulation of
ions in free solution (stage II) begins. During the process the activity of
the growth factors that previously determined division has either expired,
or is suppressed. The sensitivity of the system to cyanide is now greater,
and at some point in the kind of metabolism that is associated with cell
enlargement there must be a stage at which, through definite molecular
arrangements in which energy is transferred, the secretion of ions into the
vacuoles is negotiated. If Robertson's carrot tissue is to be regarded as
predominantly in stage II it is suggestive that an increased respiration,
caused by nitrophenols, did not, as expected under the Robertson-
OF THE ROLE OF GROWTH AND METABOLISM 393
Lundegardh view, inevitably result in an increased ion intake. It is,
however, possible to harmonize this otherwise perplexing result with many
others in which increased respiration alone, that is, respiration which is not
harnessed to growth and protein synthesis, fails to promote salt intake.
The nitrophenol may act by dislocating the mechanism of energy coupling
through phosphorylation, for it is only indirectly that respiration is in this
way made effective in ion intake.
In the meristem and its derivative tissues all these events (stages I and II
and their associated metabolic reactions) proceed in rapid and orderly
sequence as part of the pattern of growth and differentiation. However, in
evaluating the different materials used for experiment, whether these are
excised roots, aerated potato discs, carrot explants with or without coconut
milk, etc., it is important to understand their special relations to this overall
pattern and to realize that no single explanation can possibly cover the
behaviour of all.
It still seems, however, that the relationship that is most necessary to the
understanding of ion accumulation is its relation to the processes of growth
and protein synthesis. The speculations that have been advanced recognize
that all of these essentially endogonic reactions are coupled ultimately
with exogonic reactions of respiration. But, as shown by the work on
tissue cultures, the details of metabolic coupling and the very nature of
the ion-accumulation process that occurs are different in cells that are
growing by division and in cells that are growing only by enlargement.
B. SALT ACCUMULATION IN THE PLANT BODY
I. CENTRES OF GROWTH AND ACCUMULATION
Since the primary processes of salt accumulation are characteristic of cells
capable of growth and division, one may survey the plant body and
define certain centres of growth and salt accumulation. From the evidence
on tissue cultures one may now recognize that these centres of growth and
salt accumulation will behave differently according as their chief character-
istic is growth by cell division or by cell enlargement.
In the angiosperm plant body attention should, therefore, be focused
upon the special problems that the following centres of growth and salt
accumulation present.
The root. A gradation of salt accumulation along the axis of unbranched
roots has been recognized and correlated with similar gradients of meta-
bolic activity (Prevot & Steward, 1936; Steward, Prevot & Harrison, 1942;
Machlis, 1944). While it is true that these gradients dealt with segments of
394 SALT ACCUMULATION IN PLANTS: A RECONSIDERATION
root, in which those nearest the tip contained a larger number of smaller
cells,* it is still true that per unit of water the concentrations of absorbed ion
followed a well-defined pattern along the root axis which is also recogniz-
ably correlated with the gradient along the axis of respiratory intensity per
unit water content. The behaviour of the root may now be re-examined by
reference to some experiments made with radioactive isotopes. These
experiments again raise the question of the gradient of salt accumulation
along the axis, the contrasted behaviour of the root apex, or meristem, and
of the regions of cell elongation and differentiation, and also the problem
of the mechanism of secretion into the stele.
The shoot apex. As the shoot apex produces leaf elements in orderly
sequence they each pass successively through their developmental sequence.
This sequence includes a brief phase in which cell divisions predominate
followed by the * Sach's grand period of growth', in which vacuolation and
extension predominate as growth in size rises to a maximum rate and
subside as the organ reaches maturity.
Entry of salt into the leaf, however, demands, first, access to salt via the
vascular system and then the ability to accumulate it, whether this is done
by virtue of the properties of growing-dividing or of growing-extending
cells. Also reference should be made to the much neglected fact that while
the leaves present on the axis at any one time on a herbaceous angiosperm
shoot represent a developmental series, they are also to be regarded as more
closely connected within vertical ranks or units, i.e. orthostichies. Within
each orthostichy the possibility exists that the stimuli and interrelations
which determine salt accumulation can operate in a more intimate fashion
among leaves which are more directly connected by vertical vascular strands
than throughout the plant body as a whole.
Also the problem of access to salt, the means whereby the solutes
are drawn off laterally from the axis into the lateral organs, assumes a
different aspect according as one considers it in terms of the organi-
zation of the herbaceous dicotyledon or of the monocotyledonous plant
body.
Perennial woody dicotyledons also present their special problems. Here
attention may be focused upon the role of the cambium as a centre of
growth and salt accumulation in the axis and as an active agent in longi-
tudinal and lateral movement of salt. The role of resting and active buds
and the effect of the periodicity in their development requires to be
evaluated in terms of their ability to absorb and accumulate salt.
* For a dicotyledon root Robinson & Brown (1952) state that the number of cells per
segment increases up to 2 mm. from the tip, the volume per cell up to about 12 mm. from
the tip.
OF THE ROLE OF GROWTH AND METABOLISM 395
The following summary will make brief reference to investigations in
this field made upon the following plants and organs:
(1) the uptake of 137Cs by detached and attached roots of Narcissus
with special reference to the longitudinal gradation of accumulation along
single roots (making reference to experiments with R. Overstreet and
S. M. Caplin, and F. K. Millar);
(2) the intake of bromide by the shoot of Cucurbita (with A. G.
Steward);
(3) the intake of 137Cs by the shoot of Narcissus (with S. M. Caplin);
(4) the intake of bromide by the shoot of Populus with special reference
to the role of the cambium and intake by the growing buds (with J. A.
Harrison) ;
(5) direct absorption at the cambial surface of Tilia and of Acer (with
B. M. Pollock);
(6) the intake of 137Cs by the buds of Acer (with B. M. Pollock and
F. K. Millar).
Absorption of 137Cs by Narcissus roots. Overstreet & Jacobson (1946)
used carrier-free radioactive isotopes and determined the absorption of
these along the axis of single roots. Working at low temperatures (order
of 2° C.), where metabolism would be at a minimum, they attributed the
intake that they encountered to a non-metabolic absorption, or binding,
which they regarded as the prelude to the metabolically determined
accumulation to follow. Thus, even in non-dividing cells, Overstreet
identified what would seem to be a transient phase in ion intake which is
comparable (in the sense that it depends on ion binding) to that which
persists when cells remain in the permanently dividing state, as in the
tissue cultures already referred to. However, work with Narcissus roots,
selected because they do not branch and because they grow well in water,
produced some unexpected results. Experiments made in collaboration
with Drs Overstreet and Caplin (see Steward, 1948) revealed that the
gradients of 137Cs absorption along the axis of single roots of Narcissus
were very variable. In some roots the highest concentrations were obtained
near, but just behind, the tip; in others the maximum absorption occurred
many millimetres, even up to 2 cm., from the tip. Apparently a large random
sample of roots will reproduce the smooth longitudinal gradation of ac-
cumulation that Prevot & Steward (1936) described, but this is only
statistically true, and individual roots may deviate widely from this ' ideal *
behaviour. The range of differences encountered with isolated roots taken
from the same bulb at the same time may be seen in Text-fig. 7.
It is evident that even when the maximum absorption occurs near the
apex, it still occurs a few millimetres behind the root tip. Also, super-
396 SALT ACCUMULATION IN PLANTS: A RECONSIDERATION
imposed upon the otherwise smooth basipetal gradation of accumulation,
there are secondary maxima which may occur at almost any point, even
too far back from the tip for their association with growing or dividing
cells to be readily plausible. The radio-autographs of longitudinal half-
sections of Narcissus roots give direct, visible evidence of this phenomenon
(PL i, fig. i). This effect was so surprising when it was first encountered
that it suggested the possibility that the concentrations of absorbed ion in
10 20 30 40 50 60 70 80 90100 10 20 30 40 50 60 70 80 90100
Distance from apex in mm
Text-fig. 7. Relationship between absorption of 137Cs and distance of root segment from
apex. All root attached. Absorption period 17 hr. Temp. 19° C. in dark.
the cells of the root are not static but that ' waves ' of absorption may pass
along a root. In this way the point of maximum concentration may migrate
along the axis of the root, though it naturally occurs most frequently and
exists for the longest time interval near the tip.
Experiments have been made by Miss F. K. Millar (1953) in the attempt
to settle this point by a technique which allowed her to trace out the distri-
bution of 137Cs along an attached root and then to follow the changes that
occur with time. The full data and technique cannot be given here, but it
must suffice to say that evidence was obtained that the point of maximum
accumulation, first located at or near the root apex, could migrate along
the axis of an attached root, and its rate of backward movement was of the
OF THE ROLE OF GROWTH AND METABOLISM 397
order of 2-3 mm./48 hr. (Text-fig. 8). In these experiments, for reasons
dictated by the technique, the leaves were removed from the bulb and the
roots were in nutrient-free solutions of carrier caesium. Therefore these
changes, i.e. the redistribution of the accumulation of previously absorbed
137Cs with time, refer to the maturation of the preformed tissues of the root.
Growth in length by the formation of new cells did not occur appreciably
during these experiments.
100-
80-
E 60-
g 40-
S 20-
£ 0-
£ 100-
V 80-
§s 6°-
£ 2 40-
a 3 2°-
-o S 0-
8jioo-
L^80-
2g 60-
£ g 40-
!o" 20-
z o-
S. 100-
:> 80-
1 60-
3, 40
20-
-*-*•
0123456789 1011 12
Number of root segments from the tip
(1 segment = 0-5 mm.)
Text-fig. 8. Linear distribution of absorbed 137Cs in roots of Narcissus] the effect of time
after absorption. A. Root excised before absorption. — O — distribution of 137Cs at end of
absorption period ; x distribution of 137Cs 20^5 hr. after absorption period. B. Root
attached. — O — distribution of 137Cs 6-5 hr. after absorption period ; x distribution
of 137Cs 27 hr. after absorption period;— -A distribution of 137Cs 54-5 hr. after absorp-
tion period. C. Root attached. — O — distribution of 137Cs 3 hr. after absorption period ;
x distribution of 137Cs 51 hr. after absorption period. D. Root attached.
— O — distribution of 137Cs at end of absorption period ; x distribution of 137Cs
48 hr. after absorption period; A distribution of 137Cs 69 hr. after absorption
period.
It is clear, therefore, that the root responds to stimuli which regulate its
salt accumulation but which are as yet only incompletely known. The cells
of the root meristem should resemble the * growing-dividing' cells of the
carrot-tissue cultures and thus owe their salt intake to stage I of the
accumulation process, in which stoichiometrical ion-binding may pre-
dominate, but in which the highest concentrations are not necessarily
398 SALT ACCUMULATION IN PLANTS: A RECONSIDERATION
produced. The drop in 137Cs content in the apical segment is consistent
with this (Text-fig. 8). The * growing-extending* cells, in which divisions
are less frequent, would be expected to attain higher concentrations ac-
cumulating their salts by stage II of the absorption process and secreting
them into their expanding, aqueous vacuoles. Cells at the height of their
growth by extension, just behind the meristem, should thus attain the
highest concentrations, and indeed they often do.* But even so, there are
maxima of absorption superimposed upon this ' ideal * behaviour and which
exist so far from the tip as to suggest that other effects intervene to deter-
mine the absorption that may occur in these regions of the root. These
'other effects' are as yet unexplained except that they are probably part of
a periodic or * wave-like* movement that locates the point of maximum
accumulation in the maturing cells of the root at different distances from
the root apex at different times.
The more strongly absorbing attached shoot in the light has an un-
paralleled ability to deplete an attached root of its absorbed ions (Hoagland
& Broyer, 1936). This depletion of the root by the shoot seems to occur
more readily from near the tip (Steward et al. 1942). The problem of ion
secretion into the stele and of their removal to the shoot seems still to be
a complete mystery, except that it involves metabolism and respiratory
energy; for in this respect the root is subject to regulatory control by the
shoot, and the basis of this is still unknown.
The relative accumulation in leaf, stem and root. Text-figs. 9 and 10 show
the concentration and total amounts of ions absorbed in the different
regions of the plants named. The technique here is to plot the concentra-
tion (quantity per unit fresh weight or unit weight of water) as ordinate
and the fresh weight of the sampled region as the abscissa, so that the
height of each histogram is a measure of concentration and the area of the
resultant rectangle is a measure of the total amount in the sampled region.
In Cucurbita and for bromide the relative order of ion accumulation is
root < stem < leaf, and in the stem the internodes higher on the axis tended
to attain the higher concentrations (Text-fig. 9). In Narcissus the order
obviously is root > crown > leaves (Text-fig. 10).
We are clearly not able, as yet, to specify what determines the ability of
one organ to deplete another, nor can we venture to explain the differences
* Brown has rendered a service by showing that various properties reach their maximum
value per cell coincidentally, or nearly so, with the attainment of maximum cell size,
notably protein-N content and the activity of certain enzymes (Robinson & Brown, 1952).
There can be little question also that the interval in which the cell embarks upon and
reaches its maximum intake of salt spans its own growth. It is still true, however, that the
' intensity * of the salt absorption is best measured by referring it, as a concentration, to the
quantity of water in which it occurs.
OF THE ROLE OF GROWTH AND METABOLISM
399
Series A
MS- 0-50- Leaves as one sample
6
Apex
bromide/2.0'40
fresh 0- 30-
Scaleoffr«hwt. Wt'°;20;
|
ives
4
5
Le,
3
1 2
'~|6' Roots o-10-
| 0-2QJ
1 I 2 |
1 Inter
UAj
_Stem_as_pne sar
3
nodes
3pi~
4
5
6
0-30-H
0-40-
0-50-
M«- 0-50-
bromide/g.o-40-
freshO-30-
Scale of fresh wt. wt-0-20-
I 1 0-10-
110 r'
r--
Lea\«
fple'
5
-i
i
i
i
1 x
1 &>
o.
6l< 1
fa's o
L<
3
\€tt
aves
4
2
1
1 »• Roots nin
1
2 I 3
Internodes
ne sample
4
5
6
J O'TU'
Sterr
^ as o
0-20J
0-30-(
0-40H
0-50-
Text-fig. 9. Distribution of bromide in Cucurbita pepo. Plants of series A received KBr
via their roots during periods in which the shoots were in the light; plants of series B
received KBr via their roots but only during the alternating periods in which their shoots
were in the dark. Light and dark periods were adjusted to 12 hr. each.
Initial T activity of external solution
Text-fig. 10. Absorption and distribution of 187Cs in Narcissus. SS, sheath leaves in
acropetal succession; L, expanded leaf with sheathing base; stem, axis of lateral bud
bearing S 1—3 and L i— 6; crown, flattened main axis of bulb.
400 SALT ACCUMULATION IN PLANTS: A RECONSIDERATION
between the behaviour of Cucurbita toward bromide and Narcissus toward
137Cs in this respect.
However, it is possible to analyse the accumulation in the leaf in relation
to its development.
Absorption and accumulation in the leaves of herbaceous plants. Cucurbita,
which grows sympodially, is the selected example of a dicotyledon and
Narcissus the selected example of a monocotyledon.
The Cucurbita experiments were performed in such a way that the
shoots of one series (A) received potassium bromide via the roots at the
time that their shoots were in the light; the other series (B) received
potassium bromide via the roots while their shoots were in the dark.*
Relative bromide concentration of leaves in
an acropetal succession
Series B
Text-fig, ii. Relative bromide concentration of leaves in an acropetal succession. Entry
of bromide into the leaf of Cucurbita pepo. (Relative concentration in each leaf equals the
concentration in the leaf divided by the average concentration in the leafy shoot as a
whole.)
Text-fig, n, series B, shows that in leaves of the first phyllotactic
series of a Cucurbita seedling the intake of bromide by the leaves reflects
their own growth and development since the leaves which are expanding
most have the highest concentration of bromide \ the curve of bromide
absorption against leaf number on the axis clearly reflects the stage of
each leaf relative to 'Sach's grand period of growth* for that leaf.
However, superimposed upon this pattern are the following effects. When,
as in series A, the leaves are in the light and simultaneously the roots have
access to salts, the apex grows, absorbs salt (potassium bromide) direct
from the roots and also depletes the lower leaves of previously absorbed
* The periods of light and dark were the same duration (12 hr.) and the plants were
transferred as necessary from culture solution with bromide to culture solution without.
The experiment involved several such cycles.
OF THE ROLE OF GROWTH AND METABOLISM
4OI
salts so that they become able to take in more of the indicator ion (bromide).
These facts become intelligible when it is realized that the leaves i, 6,
n, etc., of Cucurbita constitute a vertical series, or orthostichy, and these
leaves are in direct connexion via one of the five cauline bundles (Text-
fig. 12).
Thus the leaves in one orthostichy constitute a closely integrated
nutritional system. As each new leaf is cut off from the apex and is added
to the series it passes through its own sequence of development and of
primary salt accumulation. This salt intake proceeds slowly at first, when
Vascular system
===== Cauline bundles
(O\ Common bundles
Leaf traces
Nodal anatomy
Leaves
1 and 6
/01
,01
• II IIIIVV V
Text-fig. 12. The vascular pattern of Cucurbita pepo, showing direct connexion of leaves
i and 6 via cauline bundles. Each leaf receives three leaf traces which fuse into a network
at the nodal plate. Of these three, one springs directly from the cauline bundle.
cell division predominates, gathering momentum and reaching high
concentrations at the height of its own expansion. Thereafter, further
intake of salt declines as expansion subsides unless, in response to the
competition from growing leaves above, the leaf in question is depleted of
total salts and thus acquires a vicarious ability to absorb again from the
root, when the supplies of absorbed salt are restored.
Before the bromide could be accumulated in the leaves of Cucurbita, it
had to be drawn off laterally and enter the vascular system of the leaf.
This only occurred extensively in the light under the conditions in which
the leaf itself grew. The mechanism which enables salts to enter the
vascular supply to the leaf is best postponed until the conditions in a tree
are described. However, in dicotyledons the vascular cambium, by its
E B S VIII 26
402 SALT ACCUMULATION IN PLANTS: A RECONSIDERATION
very potentiality for growth, commands attention as a region of potential
salt accumulation in the stem and as the means by which salts are directed
by the ebb and flow of its activity to the lateral buds.
Attention in the monocotyledon should be focused upon the intercalary
growing region at the base of each leaf; for this is not only the means by
which these linear leaves grow, but it can also be a centre of salt accumu-
lation (at the expense of the solutes in the stele) from which salt is accumu-
lated and diverted to the leaf. Experiments in which 137Cs was allowed to
enter the shoot of Narcissus substantiate this view, for 137Cs accumulated
strongly in the basal meristematic region of both the leaves and sheathing
scales of the bud (PI. i, fig. za, b). In these figures the sheathing leaf bases
are shown * opened out* and as mounted for the radio-autograph.
In the shoot of Narcissus the concentration of absorbed 137Cs in leaves
increased markedly the nearer they approach to the apex of the shoot
(Text-fig. 10). However, even the smallest segment tested does not permit
one to distinguish between the meristem proper and the rapidly elongating,
differentiating cells to which it gives rise.
Centres of salt accumulation in a tree. In the shoot of deciduous woody
dicotyledonous trees, each lateral bud, as a potential growing region, is
a potential region of salt accumulation which, granted its ability to gain
access to salts, will recapitulate the potentialities of the terminal growing
point. Buds, however, arise in isolation in the axils of leaves, and their
vascular supply does not immediately integrate with that of the main axis.
In trees it is a common condition that the bud rapidly becomes dormant in
the year of its formation ; in its dormant condition absorption of salts is at
a minimum. In the ensuing spring, with the onset of the familiar but
mysterious process of ' bud-break ', a renewed activity in the vascular
cambium emanates from the bud and spreads basipetally downward. In
the outcome, the effective vascular supply to the bud integrates with the
vascular tissues in the stem, which develop, not in the year of the formation
of the bud, but in the year of its growth. The significance of these facts in
the internal nutrition of trees has been largely overlooked; but their
pertinence was revealed by an investigation into the uptake of bromide by
a population of poplar trees (Populus nigrd) throughout one annual cycle of
growth in England (Harrison, 1938).
By appropriate experimental devices (not here reported) it was possible
to show that in the dormant condition there was a slow upward, probably
diffusive, spread of solute in the old wood. In the spring, however, the
active cambium becomes a growing region, active in the accumulation of
solutes, which are drawn laterally from the old wood within (Text-fig. 13).
Not until the growth and differentiation of vascular elements develops
OF THE ROLE OF GROWTH AND METABOLISM
403
o
su-sii
5 * -s .a
u
llllli
26-2
404 SALT ACCUMULATION IN PLANTS: A RECONSIDERATION
backward from the bud and passes downward in the stem, in the current
year of growth, does any appreciable entry of salt into the bud occur; but
thereafter it continues smoothly throughout the period when the bud grows
and develops. The data show that the leaves of Populus could reach an
appreciable size (4-4 g.) before intake of bromide from the old wood became
effective, but thereafter intake and growth in size went hand in hand. The
regression equation of growth (x = g. fresh weight) on total bromide
content (j> = mg.equiv.) is given by ^ = 0-023^ — o-ioi.
Thus, the vascular cambium in a tree is to be recognized as a prominent
region of growth and of salt accumulation. A primary function of the
cambial region is to absorb salts from the dilute solution in the old wood
within and transfer them in the developing vascular traces to growing buds
immediately above.
The knowledge that the cambium of woody dicotyledons acts as an
absorbing and accumulating region suggested the possibility that a radio-
active isotope could be applied directly to the exposed external surface of
the cambium at the time of its activity in the spring. When this is done,
by building small cups around the cambium exposed by removing a window
of bark, the absorbed isotopes move preferentially to the bud immediately
above and on the same side of the shoot as its point of application (Steward,
1948)-
Following upon this technique, experiments have also been made to
indicate the kind of absorption mechanism by which the young expanding
leaves of maple absorb 137Cs (Pollock, 1950; Millar, 1953). The following
points are particularly relevant to this discussion.
(1) Entry into the very young leaf of 137Cs supplied via the cambium
surface is almost completely confined to leaves in the light. This was true
even though the leaf was obviously expanding and so might not have been
expected to absorb by the method typical of actively dividing cells.
(2) Entry of 137Cs into the young leaf is affected by the presence of
added carrier in ways which are consistent with the view that here also, as
in the rapidly dividing tissue cultures, the 137Cs is being stoichiometrically
bound to cation-binding sites, probably on nucleic acid surfaces. Also,
such attempts as have been made to apply both 137Cs and respiratory
inhibitors to the cambium surface suggest that, in its response to inhibitors
(comparative insensitivity to cyanide), the tissue of the young leaves of the
buds of maple resembles the behaviour of the growing-dividing cells
rather than of the growing-extending but non-dividing cells.
These observations would, however, become intelligible if new binding
sites (as at a nucleic acid template surface for protein synthesis) were being
multiplied in these leaves in the light.
OF THE ROLE OF GROWTH AND METABOLISM 405
The role of the root apex, the shoot apex, the developing leaf buds, the
vascular cambium and the intercalary meristems of monocotyledons as
centres of growth and salt accumulation in the plant body is, therefore,
demonstrable. All the special problems that each presents can obviously
not be dealt with here, or even on present knowledge. The information that
has accrued from the investigation of cells at different stages of their
development tells something of the diverse ways in which cells appear to
use their metabolism to absorb and accumulate their salts. Each region,
however, can hardly operate in isolation, for shoots deplete roots, leaves
in the same orthostichy interact with each other and the active cambium
accumulates ions from the dilute xylem fluid within and supplies them, via
the current year's growth, to the buds above. Leaves in one orthostichy
constitute a more closely knit nutritional unit than the leaves of the whole
shoot. The fact that this complex pattern is controlled and integrated is
evident; the method by which it is accomplished is, however, totally
unknown. Though we now have some idea how a given cell absorbs its
solutes from dilute solution in the first place, we have no idea of the nature
of the stimulus that prompts that cell to part with those solutes so that
they may be directed to even more strongly accumulating cells elsewhere
in the plant body.
REFERENCES
BOLTON, E. T. (1950). Fed. Proc. 9, 153.
BRINKLEY, F. (1952). Exp. Cell Res. Suppl. 2, 145.
CAPLIN, S. M. & STEWARD, F. C. (1948). Science, 108, 655.
CAPLIN, S. M. & STEWARD, F. C. (1949). Nature, Lond., 163, 920.
CAPLIN, S. M. & STEWARD, F. C. (1951). Science, 113, 518.
CASPERSSON, T. O. (1950). Cell Growth and Cell Function. New York.
COMMONER, B. (1950). Disc. Faraday Soc. 9, 449.
DE VRIES, H. (1885). Jb. zviss. Bot. 16, 465.
DOUNCE, A. L. (1952). Enzymologia, 15, 251.
ELLIOTT, W. H. (1948). Nature, Lond., 161, 128.
FRANCK, J. & MAYER, J. E. (1947). Arch. Biochem. 14, 297.
HANES, C. S., CONNELL, G. E. & DIXON, G. H. (1952). Phosphorus Metabolism,
2, 95. Baltimore.
HANES, C. S., HIRD, F. J. R. & ISHERWOOD, F. A. (1950). Nature, Lond., 166, 288.
HANES, C. S., HIRD, F. J. R. & ISHERWOOD, F. A. (1952). Biochem. J. 51, 25.
HARRISON, J. A. (1938). Ph.D. Thesis, University of London.
HAUROWITZ, F. (1950). Chemistry and Biology of Proteins, pp. 164-6. New York.
HOAGLAND, D. R. & BROYER, T. C. (1936). Plant Physiol. n, 471.
HOAGLAND, D. R., HIBBARD, P. L. & DAVIS, A. R. (1926). J. Gen. Physiol. 10, 121.
LOOMIS, W. F. & LIPMANN, F. (1948). J. Biol. Chem. 173, 807.
LUNDEGARDH, H. (1945). Ark. Bot. A, 32, i.
MACHLIS, L. (1944). Amer.J. Bot. 31, 281.
MARSH, P. B. & GODDARD, D. R. (1939). Amer.J. Bot. 26, 724.
MILLAR, F. K. (1953). Ph.D. Thesis, University of Rochester, Rochester, N.Y.
OVERSTREET, R. & JACOBSON, L. (1946). Amer.J. Bot. 33, 107.
406 SALT ACCUMULATION IN PLANTS
POLLOCK, B. M. (1950). Ph.D. Thesis, University of Rochester.
PREVOT, P. & STEWARD, F. C. (1936). Plant Physiol. n, 509.
RAMAMURTI, T. K. (1938). M.Sc. Thesis, University of London.
ROBERTSON, R. N. (1951). Ann. Rev. PL Physiol. 2, i.
ROBERTSON, R. N., WILKINS, M. J. & WEEKS, D. C. (1951). Aust.J. Sci. Res. 4, 248.
ROBINSON, E. & BROWN, R. (1952). J. Exp. Bot. 3, 356.
RYDON, H. N. & SMITH, P. W. G. (1952). Nature, Lond., 169, 922.
SPECK, J. F. (1947)- J- Biol. Chem. 168, 403.
STEWARD, F. C. (1932 a). Protoplasma, 15, 29.
STEWARD, F. C. (19326). Protoplasma, 16, 497.
STEWARD, F. C. (1933). Protoplasma, 18, 208.
STEWARD, F. C. (1935). Ann. Rev. Biochem. 4, 519.
STEWARD, F. C. (1937). Trans. Faraday Soc. 33, 1006.
STEWARD, F. C. (1948). Brookhaven Conference Report BNL-C-4, Biological
Applications of Nuclear Physics, p. 96.
STEWARD, F. C. & BERRY, W. E. (1934). J. Exp. Biol. n, 103.
STEWARD, F. C., BERRY, W. E. & BROYER, T. C. (1936). Ann. Bot., Lond., 50, 345.
STEWARD, F. C., BERRY, W. E., PRESTON, C. & RAMAMURTI, T. K. (1943). Ann.
Bot., Lond., N.S., 7, 22.
STEWARD, F. C., CAPLIN, S. M. & MILLAR, F. K. (1952). Ann. Bot., N.S., 16, 57.
STEWARD, F. C. & PRESTON, C. (1940). Plant Physiol. 15, 23.
STEWARD, F. C. & PRESTON, C. (1941(2). Plant Physiol. 16, 85.
STEWARD, F. C. & PRESTON, C. (19416). Plant Physiol. 16, 481.
STEWARD, F. C., PREVOT, P. & HARRISON, J. A. (1942). Plant Physiol. 17, 411.
STEWARD, F. C. & SHANTZ, E. M. (1951). Gordon Research Conferences, AAAS
Conference Vitamins and Metabolism, p. 75.
STEWARD, F. C., STOUT, P. R. & PRESTON, C. (1940). Plant Physiol. 15, 409.
STEWARD, F. C. & STREET, H. E. (1946). Plant Physiol. 31, 155.
STEWARD, F. C. & STREET, H. E. (1947). Ann. Rev. Biochem. 16, 471.
STEWARD, F. C. £ THOMPSON, J. F. (1950). Ann. Rev. PL Physiol. i, 233.
SUTCLIFFE, J. F. (1952). J- Exp. Bot. 3, 59.
THOMPSON, J. F. & STEWARD, F. C. (1951). Plant Physiol. 26, 421.
THOMPSON, J. F., ZACHARIUS, R. N. & STEWARD, F. C. (1951). Plant Physiol. 26,
375-
UNDERWOOD, G. E. & DEATHERAGE, F. E. (19520). Food Res. 17, 425.
UNDERWOOD, G. E. & DEATHERAGE, F. E. (19526). Science, 115, 95.
EXPLANATION OF PLATE
Fig. i. Radio-autograph of longitudinal half-sections of Narcissus roots containing 137Cs.
Exposure 10 min. on X-ray film. Six roots of Narcissus from a bulb grown in tap water.
Fig. 2. (a) Radio-autograph of dissected Narcissus shoot. Exposure 30 days on X-ray
film. Younger leaves accumulated relatively more caesium than older ones. The sheath
portion of a leaf accumulated almost all the caesium taken up by the leaf. The sheath
portion opposite the blade accumulated relatively more than the sheath portion below the
blade, (b) Narcissus shoot dissected and mounted for radio-autograph. Lower row : sheath
leaves, the outermost one at right. Upper row: expanded leaves ; oldest expanded leaf on
right, youngest leaf enclosing the stem growing-point on extreme left.
PLATE 1
Fig. i.
Fig. 2 (a). Fig. 2(6).
For explanation see p. 406
ACTIVE TRANSPORT OF INORGANIC IONS
BY HANS H. USSING
Zoophysiological Laboratory, Department of Biological Isotope Research,
University of Copenhagen
The last two decades have seen a greatly increased interest in the pheno-
menon of active transport of inorganic ions. It has been accepted by
most workers in the field that the peculiar ionic distribution between living
cells and their surroundings as well as the secretion by glandular tissues
must involve the specific transport one way or the other of certain ion
species. Transport which is energized by the living cells is termed active
transport. This definition is not entirely unambiguous, however, and it
might be appropriate at the outset to discuss briefly how the term * active
transport* should properly be applied. We have to admit that, despite the
considerable effort put into the study of the behaviour of the inorganic ions,
there is still very much lacking in our understanding of the mechanisms by
which they are handled in the organism. Apparently the best thing we can
do now is to systematize our ignorance, or, in other words, to sort out the
features of the behaviour of the ions which can be explained by known
physico-chemical forces; what remains then is pooled under the heading
* active transport'. Consequently, as our understanding of the processes
has improved, there has been a tendency to restrict the use of the term
' active transport'. Let us, for example, consider the distribution of cations
between the muscle fibre and its surroundings. The concentration of
potassium in the fibre is perhaps 40 times higher than in the medium. At
first sight this requires some active transport mechanism pushing potassium
into the fibre. As pointed out originally by Dean (1941) the high potassium
concentration, however, is probably the result of the functioning of a trans-
port mechanism, a 'sodium pump', which pushes sodium ions out of the
fibre, thereby creating the potential difference across the fibre membrane
which in turn forces the potassium ions to enter in excess over the concen-
tration outside. Obviously, both the high potassium concentration and the
low sodium concentration in the fibre are the results of the functioning of
the sodium pump, but only the movement of sodium should be termed
'active transport'. The sodium ions move from a low concentration and
a negative potential to a high concentration and a positive potential, a
process which requires work on the part of the cell. The potassium ion, on
the other hand, seems to be practically in Donnan equilibrium across the
408 ACTIVE TRANSPORT OF INORGANIC IONS
fibre membrane (Boyle & Conway, 1941). Thus we speak of active trans-
port when the ion is transferred against an electrochemical potential
gradient, and one might indeed limit the use of the term to such cases
(cf. Rosenberg, 1948; Ussing, 19496). There is, however, no reason to
believe that the mechanism which is able to transport a certain ion species
against an electrochemical potential gradient should be unable to assist in
transporting the same ion species Mown hiir under a different set of
experimental conditions. Logically the latter phenomenon should also be
considered active transport. The only question is then whether it is experi-
mentally possible to demonstrate that an active transport process is
'assisting' the spontaneous diffusion of an ion.
Let us consider first the behaviour of a non-electrolyte, for instance,
glucose. This substance is supposed to diffuse through a living membrane
which separates an * outside solution' from an 'inside solution'. The
glucose concentration in the outside solution is maintained constant at coy
whereas that of the inside solution is maintained constant at ci (c0 > c{).
When the steady state has developed we have by Pick's law that the total
amount of glucose passing from the inside to the outside solution per unit
time is Moui = k.ci. Similarly, the influx is M^ = k.cQ. We thus have
Min/Mout = r0/£:1. (i)
In other words, if only diffusion is operative, the flux ratio has to be equal
to the concentration ratio for the diffusing substance. In principle the flux
ratio can easily be measured using two isotopically labelled types of glucose
to measure influx and efflux, respectively. One may, for instance, add
glucose labelled with heavy carbon, 13C, to the inside medium and calculate
the efflux from its rate of passage into the outside medium. Simul-
taneously, the influx could be measured by aid of 14C-labelled glucose
added to the outside medium.
It is easily seen that the flux ratio (Mln/Mont) cannot be larger than the
concentration ratio (cjc^ if the process is one of simple diffusion. If the
flux ratio is nevertheless found to be larger than the concentration ratio,
we would speak of active transport.
For electrically charged particles Pick's law does not apply. In spite of
this it can be shown that an equation formally identical with (i) is valid
for ions that diffuse passively through a membrane. The equation may be
written -
M. IIM — J_°C_v pzFEIRT (~\
inliviout~ f ^ e > \2/
Jici
where /0 and f± are the mean activity coefficients, c0 and ci the concentra-
tions of the ion in question in the outside and inside solutions, respectively,
ACTIVE TRANSPORT OF INORGANIC IONS 409
and E is the potential difference between the two solutions (cf. Ussing,
19490, Teorell, 1949).
As soon as the steady state has been established, this equation applies
for any ion species passing by simple diffusion only. The equation holds
not only for a homogeneous membrane, but for any number of super-
imposed layers, or for a membrane where the properties vary continuously
all the way through.
If the parameters occurring in equation (2) are available for measure-
ment, we are in a position to decide whether or not the passage of the ion
through the membrane is due to active transport.
Until now we have been considering a membrane which constitutes
a non-aqueous phase. If we are dealing with a pore membrane, the situation
is complicated by the fact that there may be a net flow of solvent through
the pores. If this is the case, the flowing solvent will exert a 'drag' upon
all diffusing particles. Such a drag would speed up the ion flux in the
direction of solvent flow while opposing the flux in the opposite direction.
A discrepancy brought about by such a drag between the flux ratio found
and that calculated according to equation (2) should not be considered
evidence of active transport. This is obvious if the flow of solvent is due
to a hydrostatic or osmotic pressure difference across the membrane. But
even if the solvent flow is due to processes in the membrane itself, it
would be preferable to speak of active water transport rather than active
ion transport, as long as the deviations from equation (2) were due to the
drag only. The drag force is non-specific, just as the electric potential
force is. The problem, whether or not the living membranes, and the cell
membranes in particular, have pores, thus attains considerable interest.
As is well known Collander (1937) considered his experiments with the
diffusion of non-electrolytes into Characean cells as indicating a lipoid-
pore structure of the cell membrane. This interpretation has been criti-
cized by Davson & Danielli (1943), who proved theoretically that similar
results might be obtained with a continuous membrane in which the
molecules dissolved. On the other hand, it was pointed out by the present
author (1952) that in case pores are present in a cell membrane, its water
permeability, as measured by the rate of osmosis, would come out higher
than its permeability as measured by the rate of diffusion of isotopic water.
This is a simple consequence of the fact that flow through pores is depen-
dent upon the pore size (compare Poiseulle's law) as well as upon the total
number of pores. Diffusion, on the other hand, depends solely on the total
area available to diffusion. Experiments in our laboratory by Zeuthen &
Prescott (1953) did indeed show that, according to the criterion just
mentioned, several types of egg-cell membranes must have pores. Nothing
410 ACTIVE TRANSPORT OF INORGANIC IONS
can be said as yet about whether pores are also present in the ordinary
cells of the animal body, but experiments comparing the osmotic and
diffusion permeabilities of the amphibian skin to water do indicate the
presence of pores in this structure (Koefoed-Johnsen & Ussing, 1953),
and the experiments performed by Visscher, Fetcher, Carr, Gregor,
Bushey & Barker (1944) on the rate of passage of heavy water through the
intestinal wall may be interpreted as indicating the presence of pores in
the intestinal mucosa.
In organs where there is a considerable net flow of solvent across
a membrane we should be prepared to meet with deviations from equa-
tion (2) brought about by the solvent drag. The solvent drag is pro-
portional to the linear rate of flow of solvent through the pores through
which the ion in question diffuses, and inversely proportional to the
diffusion coefficient of the ion in water. It is relatively easy to evaluate the
drag effect as long as it is known that water and ions follow the same path
through the membrane; if, however, the membrane presents a whole
spectrum of pore sizes of which only some are accessible to ions, the
experimental approach becomes more involved. But although the flow of
solvent is able to increase the rate of passage of a solute, the ratio between
solute and solvent will be lower in the solution leaving the membrane than
in that entering. Simple drag therefore cannot be used to concentrate the
solute. Only if, as in Visscher's well-known fluid circuit hypothesis
(Ingraham, Peters & Visscher, 1938), the pure solvent is returned by an
active process, will the net effect of the solute drag be a concentration of
the solute. Whether or not a mechanism of this type is operative in certain
cases, it is hard to believe that it is of general importance. It is, for instance,
difficult to see how a stream of water could be specific enough to carry
along sodium ions without affecting potassium or chloride ions. Ruling out,
then, the fluid circuit hypothesis as a general mechanism of active trans-
port, we are left with the hypotheses that are based on chemical reactions
between the ion which is being transported and the cell constituents. This
interaction is usually visualized as the chemical binding to the carrier at
one boundary and the release at the other boundary of the ion from the
complex after some chemical alteration of the carrier. The realization that
even the monovalent ions form organic complexes has somewhat lessened
the objections to the carrier theory. Nevertheless, it is still based on the
method of elimination, and the chemical isolation of some carrier molecule
is highly desirable to complete the case.
Summing up then, for membranes where the net water transfer rate is
insignificant or nil, it seems justifiable to speak of active transport (a) if the
transfer takes place from a lower to a higher electrochemical activity and
ACTIVE TRANSPORT OF INORGANIC IONS 411
(b) if the transfer takes place in such a way that the flux ratio found
(Mfo > Mout) is larger than that calculated from equation (2). The number
of living systems where data necessary for an analysis according to equa-
tion (2) have been obtained is still rather small. Until a few years ago the
electric potential difference across most cell membranes could not be
measured precisely. Recent progress in the construction and use of micro-
electrodes has, however, made the measurement of intracellular potentials
technically possible. The mean activity coefficient of the cell interior is,
on the other hand, still a matter of dispute. It is likely, and the work on
isolated nerve by Keynes (1951) has indeed strengthened this belief, that
the activity coefficient to be used for potassium in the cell interior can be
put equal to that of the bathing fluid. We are on safer ground, however, in
cases where dilute solutions of inorganic ions can be used as bathing
solutions on both sides of the membrane under study. Particularly simple
conditions ensue if we can use identical solutions on both sides of the
membrane. Membranes useful for such an approach are the isolated frog
skin and the isolated frog gastric mucosa. The experiments which I am
going to discuss presently were performed on the isolated, surviving frog
skin.
This organ is rather remarkable. In 1857 Du Bois-Reymond observed
that it maintains a potential difference between its inside and outside. Ever
since, frog skin has been one of the favourite objects of electrophysiologists
and students of permeability problems. About twenty years ago another
surprising property of frog skin attracted the interest of physiologists. Huf
(1935) found that the isolated surviving frog skin, when in contact with
Ringer's on both sides, performs an active transport of sodium chloride
from the outside solution to that bathing the inside. Shortly afterwards
Krogh (1937) observed that frogs in need of salt are able to take up sodium
chloride from the surrounding medium, even if the latter is as dilute as
io~5M with respect to sodium chloride. Even more surprising perhaps was
the finding (Krogh, 1938) that the mechanism is specific to sodium.
Neither potassium nor calcium were taken up at all. Among the anions
Br~ and HCO^~ were able to substitute for chloride.
If we consider the isolated skin with Ringer's on both sides, it is readily
seen that the transfer of sodium must be due to active transport. Taking a
skin which maintains a potential difference of 60 mV. between the inside
and the outside (the former being positive relative to the latter) it follows
from equation (2) that the sodium efflux ought to be ten times the influx,
if the sodium diffused passively. Experiments showed, however, that the
reverse is more nearly true. The influx is always higher, and sometimes
more than ten times higher, than the efflux. Based on these observations
412 ACTIVE TRANSPORT OF INORGANIC IONS
the present author some years ago (Ussing, 1948) advanced the hypothesis
that not only is the sodium ion actively transported, but this transport is
the source of the electric potential across the skin. Obviously this hypothesis
goes beyond the simple recognition of the active transport of sodium. It
requires further that sodium is transferred in such a way that there is an
equivalent transfer of positive electric charges. (A mechanism by which
a sodium ion from the inside solution is exchanged against another positive
ion from the outside solution does not fulfil this requirement.) Further-
more, it requires that no other active transport process or ion-forming
metabolic processes are going on by which charges are transferred across
the skin.
Just as it appears from simple inspection that the sodium ion is being
actively transported, it is seen that the chloride ion might diffuse passively
and that its net transfer through the skin is due to the electric potential
difference. But even if the potential difference is such that it renders a
passive chloride transfer thermodynamically possible, still the transfer is
not necessarily truly passive. Active processes might well be aiding or
resisting the diffusion of chloride. But if one measures the influx and
efflux of chloride with radioactive isotopes and compares the flux ratio
found with that calculated from the potential difference and the chloride
concentrations (cf. equation (2)), it turns out that the agreement is quite
satisfactory. Fig. i shows the chloride flux ratio found plotted against that
calculated from a series of experiments. The efflux was determined with the
radioactive 36C1, whereas the influx was determined as the sum of the
efflux and the net flux (the latter value as determined by chemical analysis).
The outside medium was i/io Ringer's, whereas the inside medium was
Ringer's solution. The experiments also showed that, generally, a high
potential difference was associated with a low chloride permeability and
vice versa. This result is not unexpected, since the flow of chloride ions
constitutes an electric current which short-circuits the skin potential more
or less completely, depending on whether the resistance to chloride is low
or high.
So far the observations are in agreement with the hypothesis that the
potential difference is created by the active transport of sodium. However,
even if the diffusion of chloride influences the skin potential only in so far
as it represents a short-circuit of the electromotive force, the behaviour of
the less abundant ions, notably those formed in the cell metabolism, might
still contribute significantly to the potential difference observed. In
particular we have to consider those metabolic ions which, like H+ and
HCO3~, are considered important in some hypotheses advanced to explain
the skin potential.
ACTIVE TRANSPORT OF INORGANIC IONS
413
Judged superficially, the odds against the electric asymmetry of the frog
skin being the result of one transport process only seemed quite high.
Nevertheless, it can be shown that, normally, the active transport of sodium
ions is the sole process responsible.
Just like any electric battery the surviving frog skin with its maintained
electromotive force can be short-circuited. This was demonstrated by
Francis (1933), who found that the partially short-circuited frog skin would
give off electric current for many hours continuously. Later, Stapp (1941)
and Lund & Stapp (1947) improved the technique, using electrodes of low
resistance to bring about a nearly total short-circuit of the skin. These
12345678
N.n/Nout found
Fig. i. Calculated versus found flux ratios (Mln/Mout) for chloride ions in the
isolated surviving frog skin.
workers did not correlate the current created with ionic movements. From
the foregoing it would appear, however, that such a comparison could give
important information.
It is clear that no net transfer of passive ions can take place if the skin is
short-circuited, so that the potential drop over it is nil, and if, further, the
bathing solutions on the two sides are identical. Ions which are subject to
active transport will, on the other hand, flow faster one way than the other,
and thus contribute to the total current flowing through the short-circuit.
An experimental apparatus was therefore constructed by Dr Zerahn and
myself (Ussing, 1950; Ussing & Zerahn, 1951), making possible the
simultaneous determination of short-circuit current and ionic fluxes. The
transport rate of sodium across the skin is so low that the determination by
chemical analysis of the current/active-transport relationship would meet
ACTIVE TRANSPORT OF INORGANIC IONS
with great difficulties. The tracer method, on the other hand, makes
possible the determination of the transport rate with accuracy. The influx
of sodium can be determined with 22Na and the efflux with 24Na (cf . Levi &
Ussing, 1949), and this procedure is now being regularly used. It turned
out, however, that the efflux of sodium was only a small fraction of the
influx, so that it suffices in most cases to apply a suitable correction to the
influx in order to obtain the net sodium transport.
V W
rAAAAAi
Fig. 2. Diagram of apparatus used for determining Na flux and short-circuit current.
C, celluloid chamber containing, on each side of the skin, 40 ml. Ringer; S, skin; a, inlets
for air; A, A', agar-Rmger bridges connecting outside and inside solutions, respectively,
with calomel electrodes; B, B' ', agar-Rmger bridges used for applying outside e.m.f. ;
Z), battery; W, potential divider; M, microammeter ; P, tube potentiometer.
The apparatus used is shown in Fig. 2. The skin, »S, is placed between
two celluloid chambers, C, containing Ringer's solution. The potential
differences across the skin is read on a potentiometer, P, which is con-
nected through calomel electrodes to two agar- Ringer bridges, A and A',
opening close to the skin. Another pair of agar-Ringer bridges, B and B',
opening at a distance sufficient to give a homogeneous electric field at the
level of the skin, are connected through silver/silver-chloride electrodes
with a microammeter, M, and a battery, D. The current in this circuit is
now adjusted by aid of a variable resistance, W, so that the potential drop
across the skin is zero. It is obvious that this accomplishes a total short-
ACTIVE TRANSPORT OF INORGANIC IONS 415
circuit of the skin. The current generated is read on the microammeter.
Table i shows the results of some of our first experiments. Influx and
efflux were not determined simultaneously, but in parallel experiments.
The figures are arranged so that those from influx experiments are to the
left and those from the efflux experiments to the right. Both the flux values
and the current are expressed as millicoulombs/cm.2/hr.
Table i . Sodium flux and total current values obtained in i hr. periods on
totally shorted normal frog skin. Group A comprises results from five
influx experiments, group B results from six efflux experiments
A (influx)
B (efflux)
| millicoulomb cm. ~2hr.~1
millicoulomb cm.~2hr.~1
Date
Date
Na Current
Na
Current
26. iv
102
99
28. iv
9'7
130
93
99
10-5
139
27. iv
177
174
2. V
5'3
in
176
162
9-1
1 08
124
123
13-0
1 08
13-6
112
3-v
64
64
63
55
II. V
6-0
136
57
49
5'
124
4. v
248
253
8.vi
14-7
13-2
92
100
260
224
205
205
22. ix
2-6
164
2-4
118
23. ix
139
133
118
112
4.x
0-8
102
It is noticed that the efflux is always much smaller than the current,
whereas the influx is identical with, or, in some cases, a little higher than,
the current. On an average from a considerable number of runs, the influx
is 5 % higher than the current, whereas the efflux is very nearly 5 % of
the current. Thus, the net sodium flux is exactly equal to the short-circuit
current. Consequently we arrive at the conclusion that the total current
which can be drawn from the short-circuited frog skin comes from active
sodium transport.
This holds true even if the skin is treated with agents known to affect
the skin potential. Thus, 5 % carbon dioxide in the air (or oxygen) used
for mixing the solutions depresses the current to zero. At the same time
the sodium influx drops to a low value which, incidentally, is about equal
to the efflux. This means that even the active sodium transport is stopped
by 5 % carbon dioxide. This inhibition is fully reversible as can be seen
from the table.
416 ACTIVE TRANSPORT OF INORGANIC IONS
Another agent with a striking effect on the skin potential is neuro-
hypophyseal extract which, according to Fuhrman & Ussing (1951), brings
about an increase in the skin potential. The table shows that it increases
the skin current as well as the influx and efflux of sodium. The net sodium
flux, however, remains equal to the current.
We shall revert later to the effect of adrenaline, which is rather remark-
able and constitutes the only exception so far observed to the rule of
equality between current and active sodium transport. This rule has been
found to hold in the presence of a long series of biologically active
substances.
One may ask the question whether the mechanism responsible for the
active sodium transport is specific to this ion or whether we are dealing
with a more or less unspecific cation transporting system. As long as the
bathing solutions are ordinary Ringer's, sodium is likely to dominate the
picture compared to, say, potassium, simply due to its much higher
concentration. Experiments which have recently been performed by Dr
Zerahn and myself indicate, however, that the transport mechanism prefers
sodium to potassium to a remarkable extent. Table 2 shows some of the
results. Instead of ordinary Ringer's, bathing solutions were used where,
expressed on a molar basis, 35 % of the total monovalent cation was
potassium and the remaining 65 % sodium. Influx and efflux of sodium
were determined simultaneously with 22Na and 24Na, respectively. It is
seen that even under these conditions the total current is accounted for by
the net sodium flux, indicating that potassium contributes insignificantly,
or not at all, to the short-circuit current.
Table 2. Showing that the current generated by the short-circuited frog skin
is carried by sodium ions, even when one-third of the sodium in the bathing
solutions is replaced by potassium. Area of skin , 7-1 cm.2.
(K/Na)xioo
in solutions
Influx
(/^equiv.
Na/hr.)
Efflux
(/tequiv.
Na/hr.)
Na
(//equiv./
hr.)
Na
current
Total
current
I
35-o
13-2
0-58
I2'6
339
289
35'°
9'4
0-81
8-6
230
236
II
35'°
5'8o
0-38
5-42
H5
157
35'°
6-13
0*35
578
155
154
35*0
0-41
136
143
III
35-0
4*05
i '04
3-01
81
85
35'°
i -06
4-24
114
116
35'0
5'35
0-94
4-41
118
121
Experiments now in progress (Zerahn, unpublished) indicate that
calcium and magnesium penetrate very slowly indeed even when present
ACTIVE TRANSPORT OF INORGANIC IONS
417
in high concentrations in the bathing solutions. The choline ion does not
penetrate at all (Kirschner, in preparation).
There is, however, one ion that can to some extent substitute for sodium,
namely, the lithium ion. This is clearly borne out by work which Dr
Zerahn has been doing during the last year. Table 3 gives a few examples
of the substitution of lithium for sodium. As bathing solutions, mixtures
of ordinary Ringer's and lithium Ringer's were used. The molar ratio
between lithium and sodium for the mixture used in each experiment is
indicated in column 2. Influx and efflux of sodium were determined with
22Na and 24Na, respectively. It is seen that the net sodium current is in all
cases smaller than the total short-circuit current. The part of the current
not accounted for as sodium current is given in column 8 as 'lithium
current'. The last column shows that this 'lithium current* comprises
about the same fraction of the current as lithium does of the total mono-
valent cation of the bathing solutions. Furthermore, in specially designed
experiments Zerahn was able to demonstrate by chemical analysis that the
frog skin can transport lithium against a concentration gradient. Despite
this fact lithium cannot totally substitute for sodium. In pure lithium
Ringer's the skin deteriorates rapidly.
Table 3. Showing that if part of the sodium in the solutions bathing the
short-circuited frog skin is replaced by lithium, the latter ion carries
a corresponding part of the current. Area of skin, 7-1 cm.2.
(Li/Na)
X IOO
in
Influx
(/fequiv.
Efflux
(/^equiv.
ANa
(/iequiv./
ANa
Total
current
'Li
current'
'Li
current '
o/ Of
solutions
Na/hr.)
Na/hr.)
hr.)
(/<amp.)
(/rnnp.)
total
I
21-2
8-5
0-7
7'7
206
284
78
27-4
2IP2
7-0
0-6
6-4
171
249
78
3i'3
II
33-3
6-0
0-4
5-6
150
194
44
227
33-3
3'7
0'5
3'2
86
*3l
45
34'3
III
52-0
2-81
0-40
2-41
65
107
42
39-2
52-0
2-17
0-69
1-48
40
56-0
IV
80-5
0-62
0-25
i'37
37
104
67
64*0
80-5
0-85
0-37
0-48
13
44
71-0
The active sodium transport of the frog skin seems to be strictly
dependent upon the oxidative metabolism. Thus it is stopped by oxygen
lack and by cyanide poisoning. This does not mean, however, that there
is a simple relationship between the rate of sodium transport and the
metabolic rate. This is clearly borne out by the fact that 5% carbon
dioxide, which inhibits the sodium transport entirely, depresses the oxygen
consumption by only about 25 %.
418 ACTIVE TRANSPORT OF INORGANIC IONS
Dr Fuhrman (1952) in our laboratory tested the effect of a number of
drugs upon the sodium transport and the oxygen consumption of the frog
skin. Some of the results are shown in Table 4. It is seen that, out of six
drugs which inhibit sodium transport, three stimulate, two do not in-
fluence, and one inhibits the oxygen consumption. The fact that dinitro-
phenol inhibits the active sodium transport so strongly is perhaps an
indication that ATP plays a role in the functioning of the * sodium pump'.
Table 4. Effect of some drugs on sodium transport and oxygen
consumption of short-circuited frog skin (Fuhrman)
i = inhibition ; s — stimulation.
Concentration
drug* (M/l.)
Effect on
Na transport
Effect on
O2 consumption
Dinitrophenol
5 x io~5
i
s
^-Nitrophenol
2 X I0~4
i
s
Sulphanilamide
2XIO-2
i
None
/>-Toluene sulphonamide
2X I0~2
i
s
Prontosil red
I X IO~2
i
i
Quinone
I X I0~5
i
None
-AWM/VHh
vwwvwvwww
* Concentration of inhibitor necessary to give 25-75 % inhibition of sodium transport.
Although the short-circuit current is a measure of the rate of active
sodium transport across the skin it does not give any indication of the
electric work performed by the living
cells, which depends not only on the D.. ^Na
amount of sodium transferred, but also
on the frictional resistance which has
to be overcome during the passage of
the ions through the skin. The tracer
experiment does, however, provide the
data necessary to calculate the electro-
motive force of the sodium transporting
mechanism as well as its internal
resistance. This is most easily seen
if we consider an equivalent circuit
representing the frog skin (see Fig. 3).
£"Na is the electromotive force of the
sodium transport mechanism; J?Na is the internal resistance of the * sodium-
battery', or in other words, the reciprocal of &Na, that part of the total
d.c. conductivity of the skin which is due to the sodium ion; R^r is the
resistance of the shunt brought about by all the passive ions present in
the skin. When in our experimental apparatus we adjust the potential
Shunt
Fig. 3. Equivalent circuit representing
the short-circuited frog skin. £"Na, electro-
motive force of the sodium transporting
mechanism; RZI, resistance to the Na
current; R$&, resistance to passive ions.
The lead marked * Shunt ' represents the
net effect of the applied e.m.f.
ACTIVE TRANSPORT OF INORGANIC IONS 419
difference between the two sides of the skin to zero, the effect is that of
connecting the points A and B with a shunt of infinitely high conductivity.
The fraction of the current passing that shunt will therefore be infinitely
larger than that passing through #L/, which therefore becomes virtually
zero. The amount of current drawn from the skin under these conditions
depends only upon £Na and R^aJ whereas the value of 7?E/ is immaterial.
E"Na can be estimated by three independent methods. Perhaps the most
obvious method is to apply a counter electromotive force and adjust it so
that the sodium influx and efflux become equal. This method leads to
values around no mV.
If J?Na is to be determined according to the second method, the potential
is maintained at zero while the sodium concentration of the outside solution
is lowered until influx and efflux become equal. The electromotive force
of the * sodium pump ' pushing sodium inward then is equal to the dif-
fusion force tending to force sodium outward. The best procedure is to
replace the sodium chloride of the outside solution stepwise with the
chloride of a non-penetrating monovalent ion, for example choline. In
that case the ionic strengths of the inside and outside solutions remain the
same and one can probably disregard the activity coefficients and put the
diffusion force acting on the sodium ion equal to
where c± and c0 are the sodium concentrations in the inside and outside
solutions, respectively.
The third method for determining ENa> depends on the assumption
that the electromotive force of the ' sodium pump ' affects the flux ratio
of the sodium ion in exactly the same way as an applied electromotive force
would affect the flux ratio of a passive ion. Thus, if we find the flux ratio
10/1 for a shorted skin we conclude that the £"Na responsible must be
58 mV. In general, we have that with identical solutions on both sides
and zero potential difference
Min
Since the sodium current strength is (Min — Mout), we obtain by Ohm's law
RT M^
~F mMout
*N* Mhl-Moufc-
It is seen that the tracer experiments provide the data for the estimation
of both the electromotive force and the internal resistance of the ' sodium
27-2
420
ACTIVE TRANSPORT OF INORGANIC IONS
pump*. Linderholm (1952), working in Teorell's laboratory, has shown
that the sodium conductivity (i/R^*) plus the chloride conductivity
(i//?ci) equals the d.c. conductivity of the frog skin. In the equivalent
circuit (Fig. 3) we can thus put R^j equal to Rcl which can be obtained
from experiments with radioactive chloride. Thus drugs can affect the
potential difference across the frog skin by affecting one or more of the
three parameters £"Na, 7?Na and jRcl. Cu++ may be mentioned as an
agent that increases the skin potential by increasing RC{ without affecting
£"Na and ^?Na. Most agents affecting the skin potential do so, however, by
affecting the latter two parameters or one of them. Table 5 (Kirschner,
1953) shows two examples of this. Tetraethylpyrophosphate, a potent
inhibitor of cholinesterase, when added to the inside medium, depresses
-^Na very markedly while increasing R^A. Other inhibitors of cholin-
esterase, like eserine, have similar effects. Atropine has exactly the
opposite effect, increasing the electromotive force of the ' sodium pump '
while decreasing its internal resistance. These findings suggest, but do not
prove, the participation of acetylcholine in the sodium transport mechanism.
Anyway, it is tempting to see the active sodium transport of the frog skin
as a manifestation of a general property of animal cells, which in the nerve
fibre and muscle fibre serves to extrude sodium and keep up the membrane
potential, but which in the frog skin, the kidney tubule and the intestinal
mucosa serves the purpose of salt transport.
Table 5. Effect of tetraethylpyrophosphate (TEPP) and air opine on active
sodium transport of short-circuited skin of Rana esculenta
Influx and efflux: /*M./hr./7'O7 cm.2
Current: //amp./7'O7 cm.2
C, control period.
Ey experimental period.
E*
R Flux (/tM/hr.)
/
(mV.)
(ohms/cm.2)
In
Out
(/lamp.)
I. 4x10 3M-TEPP
(inside) C
35
4060
3'24
0-84
61
E
7
(16500) 0-57
0-46
3
II. 4xio-3M-TEPP
(inside) C
45
3030
4'93
0-84
104
E
2
(7000) 0-90
0-82
2
I. I X IO~2M-
Atropine
(outside) C
35
7000 i '86
0-48
35
E
67
2300 8-66
0-60
205
II. 9 x io~3M-
Aropine
(outside) C
41
3800 3-62
0-78
76
E
66
I57O I2'l6
1-04
297
ACTIVE TRANSPORT OF INORGANIC IONS 421
This should not lead us to believe, however, that electric potentials
across living membranes are always the result of active sodium transport.
Even the frog skin may under certain conditions perform active transport
of at least one more ion, namely, chloride. When the skin is stimulated by
adrenaline, an extra source of electromotive force is aroused which turns
out to be an active outward transport of chloride ions (Koefoed-Johnsen,
Ussing & Zerahn, 1952). This transport seems to be performed by the
skin glands which start secreting under the influence of adrenaline.
Recently Jorgensen (unpublished) has observed that live frogs are able to
perform an active transport of chloride inward when in need of this ion.
Active transport of chloride ions is also at work in the isolated frog
gastric mucosa. Hogben (1951) has shown that the total electric current
generated by the short-circuited gastric mucosa comes from active trans-
port of chloride ions. Table 6 shows one of his experiments. Identical
solutions were used on both sides of the mucosa and the potential was
short-circuited according to the principles outlined above. Influx (secre-
tion to nutrient side) was determined with 38C1, whereas efflux was
measured with 36C1. The flux values are expressed as micro-equivalents/
cm.2/hr. It is seen that the net flux of chloride is 3-82. This figure equals,
within the accuracy of the methods used, the sum of the electric current
drawn from and the hydrochloric acid secreted by the mucosa. This is not
the place to discuss the role of active chloride transport in the formation of
the gastric juice, but it is evident that no explanation of the function of the
gastric mucosa is complete until it takes into account the active chloride
transport.
Table 6. Chloride transfer across the short-circuited gastric
mucosa of the frog
38CI N-S 6-80 Current 3-06
38C1 S-N 2-98 Hion 0-71
3-82 3'77
Fluxes and current expressed as /^equiv./cm.2/hr. N, nutrient side; Sy secretion side.
REFERENCES
BOYLE, P. J. & CONWAY, E. J. (1941). J. Physiol. 100, i.
COLLANDER, R. (1937). Trans. Far day Soc. 33, 985.
DAVSON, H. £ DANIELLI, J. F. (1943). The Permeability of Natural Membranes.
Cambridge University Press.
DEAN, R. B. (1941). Symp. Soc. Exp. Biol. 3, 331.
FRANCIS, W. L. (1933). Nature, Lond.y 131, 805.
FUHRMAN, F. A. (1952). Amer.J. Physiol. 171, 266.
FUHRMAN, F. A. & USSING, H. H. (1951). J. Cell Comp. Physiol. 38, 109.
HOGBEN, C. A. M. (1951). Proc. Nat. Acad. Set., Wash., 37, 393.
422 ACTIVE TRANSPORT OF INORGANIC IONS
HUF, E. (1935)- PflVg- Arch- 8es- Physiol. 235, 655.
INGRAHAM, R. C., PETERS, H. C. & VISSCHER, M. B. (1938). J. Physiol. Chem. 42,
141.
KEYNES, R. D. (1951). J. Physiol. 114, 119.
KIRSCHNER, L. (1953). Nature, Lond. 172, 348.
KoEFOED-JoHNSEN, V., Lfivi, H. & UssiNG, H. H. (1952). Ada physiol. scand. 25,
150-
KOEFOED-JOHNSEN, V. & USSING, H. H. (1953). Ada physiol. scand. 28, 60.
KOEFOED-JOHNSEN, V., USSING, H. H. & ZfiRAHN, K. (1952). Acta physiol. scand.
*7, 38.
KROGH, A. (1937). Skand. Arch. Physiol. 76, 60.
KROGH, A. (1938). Z. vergl. Physiol. 25, 335.
LEVI, H. & USSING, H. H. (1949). Nature, Lond., 164, 928.
LINDERHOLM, H. (1952). Acta physiol. scand. 27, Suppl. 97.
LUND, E. J. & STAFF, P. (1947). In Bioelectric Fields and Growth, by Lund, E. J.
ROSENBERG, T. (1948). Acta chem. scand. 2, 14.
STAPP, P. (1941). Proc. Soc. Exp. BioL, N.Y., 46, 382.
TEORELL, T. (1949). Arch. Sci. Physiol. 3, 205.
USSING, H. H. (1948). Cold Spr. Harb. Symp. Quant. BioL 13, 193.
USSING, H. H. (1949*2). Acta physiol. scand. 19, 43.
USSING, H. H. (19496). Physiol. Rev. 29, 127,
USSING, H. H. (1950). Abstr. Comm. XVIII Int. Physiol. Congr. Copenhagen.
USSING, H. H. (1952). Advanc. Enzymol. 13, 21.
USSING, H. H. & ZERAHN, K. (1951). Acta physiol. scand. 23, no.
VISSCHER, M. B., FETCHER, E. S., CARR, C. W., GREGOR, H. P., BUSHEY, M. &
BARKER, D. E. (1944). Amer.J. Physiol. 142, 550.
ZEUTHEN, E. & PRESCOTT, D. (1953). Acta physiol. scand. 28, 77.
MOVEMENTS OF CATIONS DURING
RECOVERY IN NERVE
BY A. L. HODGKIN AND R. D. KEYNES
Physiological Laboratory, University of Cambridge
I. INTRODUCTION
There are good reasons for believing that the conduction of impulses in
excitable tissues is intimately connected with movements of sodium and
potassium ions. In the giant axons of squid and cuttlefish the sequence of
events is thought to be as follows. When the membrane is depolarized,
either by application of a cathode or by activity in a neighbouring region of
the fibre, it becomes highly and specifically permeable to Na+ ions. As the
sodium concentration is much higher outside than inside, Na+ ions rush
inwards, at first further reducing the membrane potential, and finally
reversing it by some 50 mV. Near the peak of the spike, the permeability
to sodium is reduced, while that to potassium is considerably increased.
The potassium concentration gradient is directed outwards, so that there
follows a rapid outward passage of K+ ions, which only ceases when the
membrane potential has been restored to its original value. The impulse
having passed, the fibre is left with slightly more sodium inside it — and
less potassium — than it had before. Evidence of this has been provided by
tracer studies (Rothenberg, 1950; Grundfest & Nachmansohn, 1950;
Keynes, 19510, b) and analyses of stimulated axons (Keynes & Lewis,
1951), while the permeability changes have also been investigated in detail
by electrical recording methods (Hodgkin & Katz, 1949; Hodgkin, 1951;
Hodgkin, Huxley & Katz, 1952).
If the nerve is to continue to conduct impulses over long periods of time,
it must possess a mechanism for pumping out sodium and for reabsorbing
potassium. Evidence for potassium reabsorption is provided by the
experiments of Shanes (1951), but little is known about the fate of the
sodium which enters nerve and muscle fibres during electrical activity. In
contrast to the conduction mechanism, where ions move down pre-
existing concentration gradients, the recovery process necessitates the
performance of secretory work, since the ions are transported from weak to
stronger solutions. We have recently been using radioactive tracer tech-
niques to study this active transport of sodium and potassium across nerve
and muscle membranes, and shall present here some of our preliminary
findings.
424 MOVEMENTS OF CATIONS DURING RECOVERY IN NERVE
II. THE NATURE OF THE SODIUM EFFLUX IN
NERVE AND MUSCLE
If giant cephalopod axons or frog muscles are loaded with 24Na by stimula-
tion or prolonged soaking in radioactive solutions, and are then washed in
a rapid stream of inactive artificial sea water or Ringer, their radioactivity is
found to decrease in a roughly exponential manner (Levi & Ussing, 1948;
Harris & Burn, 1949; Keynes, 195 ib). There is, therefore, a continual
movement of sodium outwards through the cell membrane. This sodium
efflux can conveniently be investigated experimentally, but we must first
inquire whether it can legitimately be identified with the operation of
a secretory mechanism, or whether it should more properly be regarded
merely as a passive diffusion along the concentration and potential gradients.
One test is to apply the equation derived by Ussing (1949 a) and Teorell
(1949) relating the ratio of the inward and outward ionic fluxes to the
electrochemical activities of the ions on either side of the membrane. It
has been shown that for the independent diffusion of free ions,
_ Inwards _ Jo y o ~EFIRT ( T ^
M ~~ f r ' v '
m out wards /I °1
where the M's represent the fluxes of a given ion, /0 and /j are its activity
coefficients, C0 and Q the external and internal concentrations, and E is the
potential difference between the external solution and the axoplasm. There
is no a priori reason for assuming that sodium ions do cross the membrane
independently, but electrical studies suggest that equation (i) may be
valid, at least approximately, for the rapid sodium movements which
occur during a nervous impulse (Hodgkin & Huxley, 1952). The first
measurements of the sodium fluxes in aoo// Sepia axons (Keynes, 195 ib)
gave an influx of 61 and an efflux of 31 p.mol./cm.2/sec., but in most of our
recent experiments the fluxes have been nearly equal, averaging about
40 p.mol./cm.2/sec. for internal sodium concentrations generally between
40 and 100 mmol./l. axoplasm. These axons had all been stimulated for
about 10 min. at 50 impulses/sec, before making any measurements, in
order to ensure that the recovery process was working under approxi-
mately standard conditions. On the whole, they were probably in some-
what better condition than the axons used in the original work, being
longer and dissected by improved methods, and this is likely to explain
their lower sodium influx. In three axons similarly recovering from
stimulation, the resting potential determined with internal microelectrodes
was 60-70 mV. ; Weidmann (1951) found 62 mV. in some unstimulated
axons. Assuming equal activity coefficients in the axoplasm and in sea
MOVEMENTS OF CATIONS DURING RECOVERY IN NERVE 425
water, and taking C0 as 485 mM, the theoretical flux ratio calculated from
equation (i) lies between 200: i and 50: i, as compared with the observed
ratio close to unity. It follows that although from 0-5 to 2% of the out-
ward sodium flux can reasonably be ascribed to a purely passive diffusion
process, some other explanation must be found for the remainder. For
frog muscles the conclusion is similar, since the figures given by Hodgkin
(1951) lead to a theoretical ratio of over 150/1.
If the amount of energy required to extrude sodium at the observed rate
were greater than the resting metabolism of the tissue could possibly
provide, it would be justifiable to conclude that only part of the sodium
efflux could be an active secretion. Levi & Ussing (1948) calculated that
about 30% of the resting metabolism of a frog sartorius muscle would be
needed for sodium extrusion — a figure which seemed to demand a rather
high efficiency for the sodium pump. However, there is some indication
that their estimate may have been too large, since some more recent
experiments in which the extrusion of 24Na from pairs of frog sartorii was
measured in parallel with their oxygen consumption gave an average energy
requirement of only 10% (Keynes & Marshall, 1954). Furthermore, in
Sepia axons the energy used for sodium extrusion would not constitute an
immoderate proportion of the resting metabolism. It can be calculated
that sodium is moved outwards at a rate of roughly 3 x io~5 mol./g.
axoplasm/hr., against a total electrochemical potential difference of the
order of ii5mV. This would need about 0-08 cal./g. axoplasm/hr. of
secretory work, which represents some 10% of the resting oxygen con-
sumption. (Cardot, Faure & Arvanitaki (1950) found that isolated Septa
axons consumed i-6cu.mm. O2/mg. dry weight/hr. when soaked in sea
water.) Thus in neither tissue do energy considerations rule out the
possibility that the whole of the sodium efflux may represent an active
secretion.*8
An alternative to active secretion of sodium has been suggested by
Ussing (19496), who pointed out that sodium might conceivably be
exchanged across the membrane without the performance of any osmotic
work, if its inward and outward movements were suitably linked. Such
a mechanism would not, of course, bring about any net transfer of sodium.
We cannot be certain that this idea is wholly inapplicable to Sepia axons,
but observations of their behaviour in sodium-free solutions show that it
will not provide a complete explanation of the observed sodium efflux.
When axons containing 24Na are transferred from inactive artificial sea
* An interesting way of expressing the relationship between total oxygen consumption
and sodium movement is to give the number of ions transferred for each molecule of
oxygen consumed (cf. Davies, this Symposium). The figures used here indicate a ratio of
four Na+ ions per O2 in both Sepia axons and frog muscle.
426 MOVEMENTS OF CATIONS DURING RECOVERY IN NERVE
water to a solution in which the sodium is completely replaced by either
choline or dextrose, the immediate effect is an increase in sodium efflux
of about 30 %. This seems inconsistent with an obligatory coupling between
influx and efflux of the type postulated by Ussing.
Another reason for thinking that the sodium efflux observed with tracers
does represent a secretory extrusion is that fluxes of the same order can be
calculated from the net movements of sodium which have been observed in
frog muscle. Following Steinbach (1951, 1952), Desmedt (1953) measured
the rate at which frog muscles can pump out sodium when they are taken
from a potassium-deficient medium in which their intracellular sodium
concentration has become abnormally high, to a potassium-rich solution in
which they extrude some of the sodium against the concentration gradient.
From the curves given by Desmedt it can be calculated that the maximum
rate of net outward sodium movement is about 20 p.mol./cm.2/sec., which
is not very different from the sodium efflux found by tracer methods
(Keynes, 1954).
It is difficult to decide how much importance should be attached to the
suggestion that part of the observed sodium efflux arises from an exchange
of sodium bound in some way on the outer side of the membrane, and does
not represent a passage of ions through the membrane (Harris, 1950). In
squid axons most of the 24Na appears to be intracellular, since it can be
extruded with the axoplasm. In muscle, the size of the overshoot of the
action potential under various conditions is consistent with the assumption
that all the sodium associated with the fibre space is intracellular (Desmedt,
1953). We shall therefore take it that there is not enough externally bound
sodium to cause material disturbance to our arguments.
III. METHODS OF MEASURING IONIC FLUXES
Two types of technique have been used to measure the ionic fluxes in giant
cephalopod axons. One consists in measuring the radioactivity of the axon
itself, immersed in a rapidly flowing inactive medium, using the apparatus
described by Keynes (19516). Influxes are obtained from the amount of
radioactivity found to enter the axoplasm during a short soak in a labelled
solution, while effluxes are calculated from the rate at which the count
decreases over a relatively long period in the inactive artificial sea water.
This method serves well for determining influxes, but is not really suitable
for observing small changes in the ionic effluxes. For this purpose it can
easily be shown that the reliability of the results is greatly improved by
measuring the radioactivity which appears in the external medium,
instead of that which remains inside the axon, the gain in accuracy being
most marked when the efflux is smallest. Our second technique is illus-
MOVEMENTS OF CATIONS DURING RECOVERY IN NERVE 427
trated in Figs, i and 2, which show two forms of the apparatus used for
slightly different purposes. The axons were first loaded with radioactivity
by stimulation in 24Na or by soaking in 42K artificial sea water, and were
Wide U-tube to
equalize levels
in side puddles
Fluid in open
side puddles
continually
changed
Fluid withdrawn by
motor-driven syringe
at a rate of 05 ml. /mm.
Fig. i. Apparatus (not to scale) for collecting radioactive ions emerging from giant
cephalopod axons. The overall length of the capillary was 30 mm. For 500 fi squid axons
the internal diameter of the tubing was i mm. ; for 200 ^ Sepia axons it was 600 /*.
Side flow
\ 002ml./mm
Side flow
002ml /mm
Outflow _
05ml./mm. ."T 10-36 V.
Fig. 2. Apparatus (not to scale) for collecting 24Na extruded from Sepia axons during
application of a polarizing current. The central bulb was 6 mm. across, and the length of
capillary tubing on either side was 20 mm. The internal diameter of the capillary was
375 /i. A and D were silver tubes, chlorided on the inside. B was a large chlorided silver
wire, and C a small one.
then pulled through a short length of precision-bore capillary tubing. For
experiments with inhibitors (Fig. i) fluid was withdrawn from a single
side-arm, so that all the radioactive ions extruded from the central 30 mm.
of the axon were collected; end-effects were minimized by continually
changing the fluid in the open side puddles. The radioactivity of 5 ml.
428 MOVEMENTS OF CATIONS DURING RECOVERY IN NERVE
samples was measured in a conventional liquid counter. For other experi-
ments, in which membrane potential was one of the variables, radioactive
ions were only collected from a central 6 mm. bulb across which artificial
sea water flowed (Fig. 2). 24Na emerging from the part of the axon in the
capillary was prevented from reaching the bulb by ensuring that about
10% of the washing solution flowed sideways into the sealed side puddles,
from which fluid was sucked very slowly by a pair of motor-driven syringes.
The membrane potential of the 6 mm. central stretch of the axon could be
raised fairly uniformly by applying a polarizing voltage between the bulb
and the two side puddles. Again the radioactivity of 5 ml. samples collected
from the bulb was determined with a liquid counter.
On every occasion when the absolute size of the sodium or potassium
efflux was to be determined, it was essential to know the total concentra-
tions of sodium and potassium in the axoplasm. At the end of each success-
ful experiment on a Sepia axon, a known length of axon was cut out and
dried on a quartz thread, for subsequent determination of sodium and
potassium by activation analysis (Keynes & Lewis, 1951); concentrations
were calculated from the length of axon excised and the mean axon
diameter. In squid experiments, a sample of axoplasm was extruded on to
a quartz hook, weighed with a torsion balance, dried, and stored in a
quartz tube for analysis.
IV. THE EFFECT OF METABOLIC INHIBITORS ON
THE SODIUM FLUXES
Fig. 3 shows the results of an experiment on the sodium efflux from a Sepia
axon, using the method of Fig. i. In normal artificial sea water the rate of
appearance of 24Na in the external medium (expressed in counts/min./min.)
declines exponentially, because the intracellular 24Na is being diluted all
the time with inactive sodium; nevertheless, the absolute size of the sodium
efflux is probably nearly constant. When 0*2 mM-dinitrophenol (DNP) is
added to the washing fluid, the efflux decreases gradually to about one-
twentieth of its initial value. The effect is largely reversed by washing the
DNP away, when the efflux soon recovers to a level comparable with that
before inhibition. The older technique gives an equivalent result, the 24Na
content of a poisoned axon remaining so nearly constant over a period of an
hour or so that it is hard to obtain a reliable figure for the very small
residual sodium efflux. Two other metabolic inhibitors, cyanide in concen-
trations of i and 10 mM and 3 mM-azide, have been found to have a very
similar effect; a single experiment with iomM cyanide, performed some
time ago with the original method, showed no effect (Keynes, 1951^),
presumably because the inhibitor was not applied for long enough.
MOVEMENTS OF CATIONS DURING RECOVERY IN NERVE 429
The action of DNP on the sodium efflux persists under all conditions so
far tested. Thus the efflux remains at a very low value when choline is
substituted for sodium in the external medium, and even when the axon is
transferred to an isotonic dextrose solution containing no salts other than
the 0-2 mM of DNP. The variation in sodium influx during treatment with
DNP has also been examined. As the figures in Table i show, there is
some reduction in influx in a fully poisoned axon, but not by much more
than half.
100 rr
100
150
Minutes
200
250
Fig. 3. Sodium efflux from a Sepia axon during treatment with o-2mM-dinitrophenol.
At beginning and end of experiment axon was in normal artificial sea water. Abscissa:
time after end of stimulation in radioactive solution. Ordinate : rate at which 24Na leaves
axon. Vertical lines are ±2XS.E. (From Hodgkin & Keynes, 1953 a.)
This interruption of the sodium efflux in cephalopod axons is by no
means the only example of the action of metabolic inhibitors on the active
secretion of ions. Poisons like cyanide and DNP have been shown to
block ionic transport in a variety of animal tissues, such as gastric mucosa
(Davies, 1951), kidney slices (Mudge, 1951), frog skin (Fuhrman, 1952)
and chicken erythrocytes (Maizels, 1954), as well as in plants (Robertson,
Wilkins, & Weeks, 1951). However, in frog muscle inhibitors do not seem
to have a very pronounced effect on the sodium efflux, although it is
difficult to be certain that they do not cause some reduction. Neither
0*2 mM-DNP, nor a combination of 3 mM cyanide with 0-5 mM iodoacetate,
applied to sartorius muscles for over 3 hr. at 20° C., caused any large
reduction in the rate of loss of 24Na in inactive Ringer (Keynes & Marshall,
430 MOVEMENTS OF CATIONS DURING RECOVERY IN NERVE
V. THE EXCITABILITY OF POISONED AXONS
The marked reduction in the sodium efflux of a poisoned axon is not
accompanied by loss of excitability. On one occasion a Sepia axon 180/4 in
diameter, treated with DNP, conducted impulses for 70 min. at a frequency
of 5<D/sec. — a performance comparable with that expected of a normal
fibre. This is not surprising, because the net rate of gain of sodium at
50 impulses/sec, is fairly high, and it makes little difference to the internal
sodium concentration whether the pump is working or not. Experiments
with external electrodes in Sepia axons and with long internal electrodes in
mV.
+ 50
- 50
-100,
-O—hO-
I
-0-0 1
•0-02 -
50
100 150
Minutes
200
Fig. 4. Effect of dinitrophenol on membrane potentials in a squid axon. The potential
of an internal electrode relative to that of the external solution was determined at rest (2),
at crest of spike (i), and at maximum of positive phase (3). At beginning and end of
experiment axon was in normal artificial sea water.
squid axons have shown that the action potential and resting potential are
almost unaltered by DNP. Similar but less complete results have been
obtained with azide and cyanide. A typical experiment with 0-2 mM-DNP
on a squid axon is illustrated in Fig. 4. The only noticeable effects of the
inhibitor were (i) a slight acceleration in the rate of decline of the spike,
such as might occur if sodium were slowly accumulating inside the axon,
and (2) an initial rise in the resting potential of about 2 mV., followed by
a very slow decline. After about i hr. in DNP the resting potential and
action potential were almost exactly equal to the means of the values before
and after treatment. In these experiments the axons were surrounded by
a relatively large volume of sea water, and it is possible that accumulation of
potassium in the external fluid might have caused inexcitability had the
MOVEMENTS OF CATIONS DURING RECOVERY IN NERVE 431
volume been much smaller (see the work of Shanes & Hopkins (1948) and
Shanes (1951) on the effect of anoxia on crustacean and squid nerve).
It is not unreasonable to find that inhibitors do not interfere with the
excitability of giant axons, because there is already other evidence that the
immediate source of energy for the transmission of the nerve impulse is
not metabolic, but is derived from the movement of ions down the con-
centration gradients. In order to test this point further, we have used 24Na
to measure the rapid sodium movements during stimulation of poisoned
fibres. At the beginning of one experiment, the result of 4 min. stimulation
(at 50 impulses/sec.) of a squid axon in artificial sea water made up with
24Na was to cause an extra sodium entry of 11-5 p.mol./cm.2/impulse; the
resting influx was 50 p.mol./cm.2/sec. The axon was then treated with
o«2 mM-DNP for an hour, reducing the sodium efflux almost to zero as
usual. Stimulation in a 24Na solution containing 0-2 mM-DNP now caused
a sodium entry of 10-7 p.mol./cm.2/impulse (assuming the resting sodium
influx to have been unaltered by DNP ; if the influx were actually halved,
the correct result would be slightly greater — iri p.mol./cm,2/impulse).
When the axon had been allowed to recover in normal artificial sea water
for an hour, a final period of stimulation gave an entry of 1 1-9 p.mol./cm.2/
impulse. In another experiment, the extra outward movement of sodium
was measured during electrical activity, using the technique shown in
Fig. i. In normal artificial sea water at the start and finish, stimulation
gave rise to outward sodium movements estimated as 5*9 and 5-4 p.mol./
cm.2/impulse. In 0*2 mM-DNP, with the resting efflux reduced to about
one-tenth, stimulation still resulted in an outward sodium movement of
4-6 p.mol./cm.2/impulse.
At least in cephalopod axons, it thus seems that the mechanism re-
sponsible for the conduction of impulses can be dissociated from that which
restores the ionic concentration differences after activity. This conclusion is
not necessarily valid for other kinds of excitable tissue. Interpretation of
the results of treating frog muscle (Ling & Gerard, 1949), frog nerve
(Lorente de No, 1947) and crustacean nerve (Shanes & Hopkins, 1948)
with inhibitors, or of depriving them of oxygen, is complicated by the
possibility that some of the observed effects may arise secondarily from
changes in internal ionic concentrations together with an accumulation of
potassium outside the fibres. However, there are definite indications that
in some cases metabolic poisons may have a direct influence on the
membrane potential.
432 MOVEMENTS OF CATIONS DURING RECOVERY IN NERVE
VI. THE EFFECT OF EXTERNAL POTASSIUM
ON THE SODIUM EFFLUX
Steinbach (1940, 1951, 1952) has shown that when frog muscles are
soaked in potassium-free Ringer's solution they lose potassium and gain
sodium, and that if they are then transferred to potassium-rich Ringer they
are able to extrude some of their surplus internal sodium. Part of the
explanation for this behaviour appears to be that the sodium efflux is
appreciably decreased in potassium-free Ringer, and is increased well
above normal in potassium-rich Ringer. The changes are most easily
observed in a very small muscle like the extensor longus dig. IV; in the
sartorius they tend to be obscured by diffusion effects (Keynes, 1954).
* K-free -H
5=^20
10
50 100 M 150
Minutes
Fig. 5. Effect of a potassium-free solution on sodium efflux from a Sepia axon. At the
beginning and end of the experiment the axon was in normal artificial sea water, in which
potassium concentration was 10-35 mM. Abscissa: time after end of stimulation in
radioactive solution. Ordinate: rate at which 24Na leaves axon. (From Hodgkin &
Keynes, 19536.)
A similar reduction of sodium efflux in a potassium-free medium has been
noted in human erythrocytes by Harris & Maizels (1951). As Fig. 5
shows, Sepia axons behave in the same way. In potassium-free artificial
sea water the sodium efflux is reversibly reduced to about one-third of its
normal value, the effect apparently being immediate, in contrast to the
delayed action of inhibitors. High potassium concentrations cause an in-
crease in sodium efflux, but not a very large one, the efflux only being
raised to 30% above normal when the external potassium concentration is
50 mM.
It seemed of some interest to inquire into the reason for this effect. In
the first place, we have found that it cannot be due to a decreased sodium
permeability in both directions, since the sodium influx is not significantly
different from normal in potassium-free artificial sea water (see figures in
Table i). Another explanation might be that there is a rise in resting
potential when the external potassium is removed, and that this slows
MOVEMENTS OF CATIONS DURING RECOVERY IN NERVE 433
sodium extrusion by increasing the electrical potential gradient against
which a positively charged ion has to be ejected. The effect of potassium
concentration on the resting potential in Sepia axons has been determined
with internal microelectrodes, and the rise in potassium-free sea water
turns out to be 5-10 mV. Other types of nerve fibre give changes of the
same order (see Hodgkin, 1951).
In order to find out whether an increase of 10 mV. in membrane potential
could reduce the sodium efflux enough to account for the effect of re-
moving external potassium, we used the apparatus shown in Fig. 2. This
10
7
5
c
rj
£
3
z
<u
^_
ex
2
o
c
X
£
E
"^
1
UJ
c
0
07
0-5
03-
02-
17mV.
K-free
50
100 150
Minutes
200
Fig. 6. Effect of anodal polarization and a potassium-free solution on sodium efflux from
a Sepia axon. Abscissa: time in minutes. Ordinate: rate at which 24Na leaves axon.
Vertical lines are ±2 XS.E. The figures of 17 and 28 mV. are based on pairs of measure-
ments giving 15 and 19 mV. in one case, and 36 and 20 mV. in the other. The apparent
lag in the effect of the potassium-free solution is explained by the time taken to change
solutions, the apparatus not being washed out between samples as it was in the experiments
of Figs. 3 and 5.
was so designed that the efflux from a 6 mm. length of Sepia axon could be
measured while the membrane potential was varied. Before starting an
experiment, current was applied between A and By and the electrotonic
potential it produced was measured between C and Z). A similar measure-
ment was made at the end of each group of determinations, in order to
allow for changes in membrane resistance. From these readings it was
possible to estimate the average change in membrane potential produced
when current was applied both to A and D as in Fig. 2. The results of
a typical experiment are shown in Fig. 6. It will be seen that with polarising
currents which gave mean potential increases of 17 and 28 mV., there was
no significant alteration in the rate at which sodium left the axon, although
E B S VIII 28
434 MOVEMENTS OF CATIONS DURING RECOVERY IN NERVE
a potassium-free solution caused a substantial reduction. Similar results
were obtained in other experiments, the average change in efflux being
0-99 + 0-04 (s.E. of mean) for an increase in membrane potential averaging
1 8 mV. This indicates that the effect of removing external potassium
cannot be due to its action in increasing membrane potential.
VII. COUPLING BETWEEN POTASSIUM INFLUX AND
SODIUM EFFLUX
Another possibility is that there may be some more specific form of linkage
between potassium influx and sodium efflux, of the type proposed for
erythrocytes by Harris & Maizels (1952). Thus one might imagine a cyclical
mechanism in which a potassium carrier (X ) moved inwards in association
with potassium, was converted by metabolism into a sodium carrier ( Y ) on
the inside of the membrane, and returned to the outside in association with
sodium. A further conversion of Y into X on the outside of the membrane
would complete the cycle, and sodium would move outwards on one limb
while potassium moved inwards on the other. A system of this kind would
be inhibited both by removal of external potassium and by interfering with
the metabolic activity of the cell. Support for this type of hypothesis is
provided by the action of DNP and cyanide on the potassium influx of
Sepia fibres recovering from a bout of stimulation. As Table i shows,
when these inhibitors were applied in concentrations sufficient to cause
a drastic reduction in sodium efflux, there was a reversible reduction in
potassium influx to about one seventh of the normal value. This was not
due to a reduction in the potassium permeability of the membrane, since
the potassium efflux was, if anything, somewhat increased by 0-2 mM-DNP
or cyanide. Nor was it due to a change in membrane potential, since we
have shown that this is virtually unaltered by DNP. The most reasonable
explanation is that, in addition to moving passively through the membrane,
potassium ions may also be drawn into the cell by a metabolic process
coupled to one which simultaneously extrudes sodium.
The idea of a coupling between potassium influx and sodium efflux is
attractive because it would explain both the action of inhibitors on potas-
sium influx and the effect of external potassium on sodium efflux. It also
seems consistent with the observations that inhibition of the sodium pump
has little effect on the resting potential, and that alterations of membrane
potential do not change the sodium efflux. Thus if sodium extrusion were
coupled to potassium absorption, the secretory process would transfer no
charge across the membrane, so that it might be expected neither to affect
the membrane potential directly, nor to be altered by changes in membrane
potential. On the other hand, the hypothesis raises a number of difficulties
MOVEMENTS OF CATIONS DURING RECOVERY IN NERVE 435
which have not yet been resolved. In the first place it is clear that the link
between potassium influx and sodium efflux is not at all rigid, since the
latter is found to continue at about its normal rate (or somewhat above it)
when all ions are removed from the external medium and tonicity is main-
tained with dextrose. Under these conditions sodium is no longer moving
against a concentration gradient, so that the situation is hardly comparable
with the normal one. Nevertheless, it seems clear that the sodium efflux
into sugar solutions does not involve a passive movement, since it is still
blocked by DNP. We have checked that there is no detectable pH change
when sodium is extruded into a small volume of isotonic dextrose, so that
an exchange of Na+ and H+ ions can be eliminated. The only remaining
alternative is that under certain conditions sodium may move out of the
axon with some unidentified anion.
Table i . The effects of various solutions on the
sodium and potassium fluxes in Sepia axons
JMorrnal
Flux ratios in abnormal media
fluxes
(p.mol./
Dinitro-
phenol
Cyanide
Azide
[K] =
[K] =
(cm.2/sec.)
3 mM
o mM
50 mM
o-i mM
O'2 mM
i mM
2 mM
10 mM
Na influx
35
0'5
—
—
—
0-9
—
Na efflux
40
O'l
0-05
0-06
—
<o-o8
<o-o8
0-3
i'3
K influx
20
—
0-13
—
0-3
—
—
0
6
K efflux
30
c. 1-5
—
C. I 'I
—
—
i
5
Ail these figures were obtained with axons which had been stimulated for about 10 min. at
50 impulses/sec, before making any measurements. The first column gives the average fluxes in
a normal medium (artificial sea water), i p.mol. = i /i/imol. = 10 ~12 mole. The other columns
show the ratios of the fluxes in abnormal media to the geometric mean of the normal fluxes
determined before and after treatment with the abnormal solution. Some of the figures are
based on few experiments, and are subject to revision.
Another difficulty raised by our experiments is that the evidence for
a secretory potassium influx destroys the apparent agreement between the
observed flux ratios for potassium and those calculated for independent
transport from equation (i). In the axons used in the present experiments,
the potassium influx averaged just under 20 p.mol. /cm. 2/sec., and was
usually between one-third and two-thirds of the efflux. As has been
argued previously (Keynes, 19516; Hodgkin, 1951), this ratio is approxi-
mately what one would expect if K+ ions were free to diffuse independently
across the membrane. It is now clear that the agreement must be a coin-
cidence, since a large fraction of the influx depends on metabolism, and the
flux ratio falls to 0-05 in poisoned axons without appreciable changes in
resting potential or potassium concentration. It also follows that even in
28-2
436 MOVEMENTS OF CATIONS DURING RECOVERY IN NERVE
a poisoned axon the movements of potassium ions across the membrane
are not independent in the sense required for equation (i) to hold.
The evidence for an active uptake of potassium in fibres recovering from
stimulation has been emphasized because it was an unexpected result.
However, it should not be thought that potassium can only move into nerve
fibres through a secretory channel. If axons are depolarized by raising the
external potassium concentration to 50 mM, the membrane becomes much
more permeable to potassium, and permits large fluxes of potassium to pass
in both directions. Under these conditions we have found that DNP
apparently has very little effect, since the influx and efflux in poisoned
fibres are close to the average values for unpoisoned fibres in 50 mM-
potassium (see Table i ). This is satisfactory, because there is other evidence,
both from electrical studies (Hodgkin & Huxley, 1952) and from tracer
experiments (Hodgkin & Huxley, 1953), that depolarization causes a
maintained increase in the potassium conductance of the nerve membrane,
and that this change has an important role in the conduction of impulses.
It therefore seems likely that potassium can cross the membrane by two
parallel routes — a secretory channel and a passive permeability channel.
But there is clearly not yet enough evidence to apportion the normal
fluxes between these two pathways.
REFERENCES
CARDOT, H., FAURE, S. & ARVANITAKI, A. (1950). .7. Physiol. Path. gen. 42, 849.
DAVIES, R. E. (1951). Biol. Rev. 26, 87.
DESMEDT, J. (1953). J. Physiol. 121, 191.
FUHRMAN, F. A. (1952). Amer. J. Physiol. 171, 266.
GRUNDFEST, H. & NACHMANSOHN, D. (1950). Fed. Proc. 9, 53.
HARRIS, E. J. (1950). Trans. Faraday Soc. 46, 872.
HARRIS, E. J. & BURN, G. P. (1949). Trans. Faraday Soc. 45, 508.
HARRIS, E. J. & MAIZELS, M. (1951). J. Physiol. 113, 506.
HARRIS, E. J. & MAIZELS, M. (1952). J. Physiol. 118, 40.
HODGKIN, A. L. (1951). Biol. Rev. 26, 339.
HODGKIN, A. L. & HUXLEY, A. F. (1952). J. Physiol. 116, 449.
HODGKIN, A. L. & HUXLEY, A. F. (1953). J. Physiol. 121, 403.
HODGKIN, A. L., HUXLEY, A. F. & KATZ, B. (1952). J. Physiol. 116, 424.
HODGKIN, A. L. & KATZ, B, (1949). J. Physiol. 108, 37.
HODGKIN, A. L. & KEYNES, R. D. (1953 a). J. Physiol. 120, 45 P.
HODGKIN, A. L. & KEYNES, R. D. (19536). J* Physiol. 120, 46 P.
KEYNES, R. D. (1951 a). J. Physiol. 113, 99.
KEYNES, R. D. (19516). J. Physiol. 114, 119.
KEYNES, R. D. (1951^). The Role of Electrolytes in Excitable Tissues. Rio de
Janeiro : Institute de Biofisica.
KEYNES, R. D. (1954). In preparation.
KEYNES, R. D. & LEWIS, P. R. (1951). J. Physiol. 114, 151.
KEYNES, R. D. & MARSHALL, G. W. (1954). In preparation.
LEVI, H. & USSING, H. H. (1948). Acta physiol. scand. 16, 232.
MOVEMENTS OF CATIONS DURING RECOVERY IN NERVE 437
LING, G. & GERARD, R. W. (1949). J. Cell. Comp. Physiol 34, 413.
LORENTE DE No, R. (1947). A Study of Nerve Physiology, i and 2, in Studies from
the Rockefeller Institute for Medical Research, 131 and 132. New York.
MAIZELS, M. (1954). This Symposium.
MUDGE, G. H. (1951). Amer.J. Physiol. 167, 206.
ROBERTSON, R. N., WILKINS, M. J. & WEEKS, D. C. (1951). Aust. J. Sci. Res. B,
4, 248.
ROTHENBERG, M. A. (1950). Biochim. biophys. Acta, 4, 96.
SHANES, A. M. (1951). J. Gen. Physiol. 34, 795.
SHANES, A. M. & HOPKINS, H. S. (1948). J. Neurophysiol. n, 331.
STEINBACH, H. B. (1940). J. Biol. Chem. 133, 695.
STEINBACH, H. B. (1951). Amer.J. Physiol. 167, 284.
STEINBACH, H. B. (1952). Proc. Nat. Acad. Sci., Wash., 38, 451.
TEORELL, T. (1949). Arch. Sci. physiol. 3, 205.
USSING, H. H. (i949a). Acta physiol. scand. 19, 43.
USSING, H. H. (19496). Physiol. Rev. 29, 127.
WEIDMANN, S. (1951). J. Physiol. 114, 372.
THE REGULATION OF SODIUM AND
POTASSIUM IN MUSCLE FIBRES
BY H. BURR STEINBACH
Department of Zoology, University of Minnesota, Minneapolis
I. INTRODUCTION
One of the first manifestations of life must have been a differentiation of
living forms from the environment. In basic chemical make-up there is
a remarkable uniformity found in all types of cells that have been investi-
gated. This uniformity appears to extend even to such stable, diffusible
substances as the inorganic ion-forming metals, sodium and potassium.
The ratio of internal to external concentration is always higher for potassium
than for sodium; usually the absolute internal concentration of potassium
is higher than that of sodium. Since these elements are rather similar with
respect to their physical and chemical properties, and since all normal active
cells are known to be permeable to them, it follows that mechanisms for
selective elimination or uptake of the two must have developed at an early
period in the evolution of life.
Leaving, for the moment, single cells, and turning to higher multi-
cellular forms, for example, the frog or the crab, an equally remarkable
ability to regulate the internal body fluids with respect to ions is found.
Oddly enough, the regulation of the body fluids with respect to sodium and
potassium is the precise opposite of that of the individual cells. As is well
known, many salt-water Crustacea can tolerate considerable dilution of
sea water in which they live with only minor variations in the salt content
of the body fluids. As the external medium is diluted, sodium and chloride
of the blood drops slightly to a new level which is then maintained or
regulated at a constant value in the face of further external dilution (cf.
Krogh, 1939). On the other hand, if potassium of the body fluid is measured
it is found to vary nearly directly with the external ionic strength. If tissues,
such as muscle and nerve, are assayed in parallel experiments, potassium
tissue concentration is found to be closely regulated; sodium tissue con-
centration varies as the external medium within wide limits (Steinbach,
unpublished).
Thus the cells of higher animals maintain a relatively constant internal
potassium concentration, yet these same cells, organized as tissues of the
skin, alimentary tract and excretory tract, serve as devices for regulating
the sodium content of the body fluid. It is perhaps noteworthy that, while
REGULATION OF SODIUM AND POTASSIUM IN MUSCLE FIBRES 439
individual plant cells appear to show the same sort of chemical differentiation
with respect to sodium and potassium as animal cells do, the whole organisms
have not utilized their ion-transporting devices to give the highly regulated
body fluids (high sodium) shown by their more mobile animal relatives.
During normal resting existence, cells typically maintain a high-potas-
sium, low-sodium composition of the intracellular fluids. Whole organisms,
among the higher animals, also maintain a high-sodium, low-potassium
body fluid. In general, there are two common categories of conditions
under which both cells and whole animals tend to lose their abilities for
maintaining characteristic ionic distributions: (a) any violent alteration
tending to lead to the death of the living unit (shock, extreme temperature
changes, etc.); (b) transitory responses to stimulation. Living units, during
a response to a stimulus or during 'death changes', lose potassium from
the cells and gain sodium, and the body fluids correspondingly tend to
equilibrate with the external environment.
Thus, not only is the ionic differentiation a characteristic of the living
organism, but it is completely dependent upon the metabolic processes
that contribute to the state we call * living'. It is towards an explanation
of this delicate criterion of life that this paper is directed.
II. THE CASE OF THE ISOLATED FROG SARTORIUS
MUSCLE
Of all the cell types that have been studied, probably the isolated striated
muscle of the frog is the best known, barring, of course, the ubiquitous
erythrocyte with all its special complications. The frog sartorius appears to
be a good choice of material in general, since the distribution and move-
ment of sodium and potassium in it appears to be quite comparable to
similar phenomena in a variety of muscles from many different types of
organism, vertebrate and invertebrate (cf. Steinbach, 1947). In addition,
it is almost indestructible; even biochemists and physiologists can excise
it with a minimum of damage, and a great many physical facts are known
about its resting and contracted state.
Over a period of many years, a variety of explanations have been offered
for the ability of muscle fibres to maintain a high-potassium, low-sodium
condition. These explanations fall into three categories.
(1) It is assumed that muscle cells are either impermeable to all cations
or to the special cation, sodium, or to cations of hydrated volume similar to
sodium or greater.
(2) It is assumed that some constituents, presumably organic, within the
muscle fibre can effectively bind a considerable fraction of the potassium
in preference to sodium.
44° THE REGULATION OF SODIUM AND
(3) It is assumed that the high internal potassium concentration is
maintained by an electrochemical gradient, sodium entering the fibre but
being transported outward by a sodium extrusion system.
The first two of these assumptions will be examined briefly, the third in
more detail.
(i) Permeability v. Impermeability
Since the work of Fenn (cf. Fenn, 1936, for references) it has been known
that sartorii, isolated into the usual Ringer's fluid (potassium concentration
c. 0*002 M) tend to lose potassium and gain sodium. Except for a short
interval immediately after isolation, only slight changes in concentration
of inorganic anions such as chloride were observed, hence it was concluded
that sodium entered the muscle fibres in exchange for potassium. Earlier,
Meigs (Meigs & Atwood, 1916) had shown by many good chemical analyses
that muscles treated with solutions containing great excesses of potassium
would swell markedly, and that this swelling was reversible under certain
conditions involving a reversible gain and loss of both potassium and
chloride.
Thus, for a long time, the evidence has been very good that muscle
fibres are permeable, in the strict sense of the word, to sodium, potassium
and chloride. Recent work with isotopic tracers (reviewed elsewhere in
this volume) has served to demonstrate this point with great clarity and to
give us some rather precise numbers to work with, with respect to relative
rates of penetration of the substances in question.
Permeabilities expressed in dimensions comparable to diffusion rates
are of prime importance for processes that are highly time-dependent.
Reversible changes during excitation, occuring in milliseconds, might well
be controlled by relative rates of penetration of specific ions. Even the
slower changes during growth and differentiation might depend in their
details upon permeability rates. However, the steady-state changes which
are reflected in the normal maintenance of the internal sodium and
potassium concentration of muscle fibres are probably best regarded as
relatively time-independent and hence permeabilities, as rates, will be
invoked only incidentally in the present discussion.
Teorell (1949) has used a special term of some interest. In describing
cases where, in the gross distribution of a substance, cells behave as though
they were impermeable (i.e. the substance does not seem to enter to cause
net changes in concentration), he speaks of * false impermeability'. The
frog sartorius would thus show * false impermeability' to sodium, the
erythrocyte would show ' false impermeability* to cations in general. In such
instances we may be aware of a false impermeability, but in calculations of
various electrochemical equilibria it is frequently useful to assume true
POTASSIUM IN MUSCLE FIBRES 441
impermeability. However, the usefulness of this assumption should not
blind us to the fact that the assumption is wrong in its fundamental
details.
(2) Ion binding
In order to explain the distribution of sodium and potassium in the frog
muscle, ion binding is almost as attractive an assumption as is selective
impermeability. It is, however, very difficult to discuss on a precise basis
because of the lack of good evidence. Certain it is that whole muscles
behave as though they contained, under certain conditions, 'quotas' of
potassium which can leave the tissues with varying degrees of ease. A num-
ber of abstracts and short papers have been published claiming that myosin
(the major protein constituent of muscle) and haemoglobin (occupying that
same quantitative position in the erythrocyte) can selectively bind appre-
ciable potassium in preference to sodium and thus contribute to the general
ionic distribution (cf. Stone & Shapiro, 1948; Steinbach, 1950). However,
studies in which pure haemoglobin (Battley & Klotz, 1951) and muscle
particulate matter (Steinbach, 1950) were analysed gave no evidence for
a selective binding amounting to more than a small percentage at most. In
the case of the muscle tissue, any selective binding was in the direction of
excess sodium rather than potassium. Recently, good evidence has been
brought forth for a non-exchangeable fraction of potassium in liver
mitochondria held under aerobic conditions. However, even here, only
a very small fraction of the total base-binding capacity of the mitochondria
is involved (cf. discussion of paper by Mudge, 1953).
That there must be some selective combination of cell constituents with
both sodium and potassium is indicated by the activating and inhibiting
effects of these ions on certain enzyme systems. However, here again, the
amounts involved are very small as compared with the total numbers of
ions present in the protoplasmic system.
Lastly, it is worth reiterating that all studies of osmotic concentration
and electrical conductivity of the internal protoplasm of muscle fibres are
consistent with the idea that a major fraction of the internal sodium and
potassium is present as freely dissociable ions, exerting osmotic pressure
and carrying electric current as in normal solutions.
Ion binding then must be discounted as a factor leading directly to exces-
sive analytical concentrations of potassium in protoplasm in preference to
sodium. However, the small bindings that have been indicated may very
well be of crucial importance in active transport mechanisms.
442 THE REGULATION OF SODIUM AND
(3) Active transport of ions
Since neither impermeability nor ion binding can be demonstrated to
play dominant roles in cells and protoplasm, it is only natural that recent
investigations have been interpreted from the point of view of active
transport of ions. Such mechanisms can give rise to * physiological imper-
meabilities* but still be reversible and dynamic. For our purposes, active
transport may be defined as any process leading to the movement, into or
out of a fixed volume, of the substance in question against its electro-
chemical gradient. This is the simplest case to consider, since the complica-
tions introduced into systems with bulk movement of fluids are considerable.
With specific reference to the ions, sodium and potassium, not only can a
constant volume of the muscle fibre usually be assumed but also a constant
number of base-binding groups per unit of protoplasmic volume. This
assumption, however, should always be regarded with suspicion, since it
does not hold for some muscle types nor need it always hold for the same
muscle under different conditions (cf. Steinbach, 1947).
The electrochemical gradient for a charged unit such as an ion is repre-
sented by the sum of the diffusion gradient and the electrical gradient. The
details of operation of these gradients have been worked over admirably by
others reporting in this symposium. At electrochemical equilibrium the
electrical potential difference between two phases is a function of the ratio
of concentrations of any given mobile charged unit in the phases multiplied
by the appropriate constant and sign. This, of course, holds only for an
equilibrium state. The relationship between potential difference and con-
centration ratio during attainment of equilibrium is not so simple, nor has
it been well worked out in a way useful to the biologist.
As has been known for many years, the potential difference between the
inside and the outside of the muscle fibre is very close to that calculated
from the ratio of potassium concentrations. Hence muscle fibres (and cells
in general) have been referred to frequently as potassium electrodes. In a
sense, then, there is nothing to explain about the potassium distribution;
it is just as it should be, granted that there is some device for producing the
electrical potential difference. With the knowledge that sodium, of similar
charge to potassium, can penetrate into and out of muscle fibres and yet be
maintained in low internal concentration, distinctly not in electrochemical
equilibrium with external solutions, emphasis has shifted to sodium trans-
port systems as devices leading to the potassium equilibrium.
While evidence for the active extrusion of sodium from the interior of
frog muscle fibres has been available for some time, recent work has been
concerned with two types of phenomena: (i) the demonstration that the
POTASSIUM IN MUSCLE FIBRES 443
exchange of sodium and potassium between inside and outside of the fibre
is completely reversible (Steinbach, 1950), and (2) the demonstration with
isotopic tracers that influx and efflux of sodium are distinctly different
(Levi & Ussing, 1948). In addition, it has recently been shown that no
potassium accumulation is noted when internal sodium of sodium-enriched
muscles is replaced by choline (Steinbach, 1952). Choline appears not to
be involved in the transport mechanisms, therefore it is not extruded and
hence there is no space (charges) made available in which potassium can be
concentrated rapidly in excess into the interior of the fibre according to the
electrochemical gradient. In general, the evidence available at present is
entirely consistent with the idea that the major system for regulating the
sodium and potassium balance of the frog sartorius fibre is a sodium
extrusion system. It should be noted, of course, that a contribution by a
potassium transport system is by no means ruled out. It merely is not
necessary to explain results at hand.
Assuming the validity of the foregoing analysis, some new evidence will
be presented relating to the general characteristics of the sodium transport
system. These results, which will be reported in more detail elsewhere,
have all been obtained by methods previously described (cf. Steinbach,
Sodium-rich, potassium-poor sartorius muscles were prepared by soaking
pairs of muscles overnight at ice-box temperatures (c. 2° C.) in potassium-
free Ringer's. Rate of loss of potassium and rate of gain of sodium is by no
means constant for different sets of muscles, but the rates are more nearly
comparable between members of a pair. Therefore most results are
expressed in terms of differences between members of pairs of muscles.
In general, one muscle of a pair, after equilibration with the potassium-
free salt solution, was analysed as an initial control, the other member of
the pair being treated with potassium-containing solution as the experi-
mental muscle. It will be assumed throughout the presentation of the
data that sodium extrusion is the primary process. However, it must be
remembered that, operationally, rate of sodium extrusion is identical with
rate of potassium uptake.
The rate of sodium extrusion
By exposing experimental sodium-rich muscles to potassium Ringer's
(0*01 potassium, slightly above the minimum maintenance concentration)
for different periods of time the data obtained in Fig. i were obtained. It is
clear that the major fraction of the excess sodium has been extruded during
the first 30 min. of recovery at room temperature. Since about 15 min. is
necessary for equilibration with the extracellular spaces (Levi & Ussing,
444 THE REGULATION OF SODIUM AND
1948), it is obvious that sodium extrusion is a very rapid process under
these conditions. Without attempting any precise corrections for diffusion
into extracellular spaces, a rate of 40 mmol./kg./hr. may be set as a reasonable
rate.
ssoh
JO
X
40 -
E
o
£ 20
_
o
10
" i
XK
3 4
Hours recovery
Fig. i
330
E * X
• X
-10
•Na
XK
50 60 70 80 90
Initial Na mmol./kg. muscle
Fig. 2
The effect of internal sodium concentration on the rate of sodium extrusion
Taking advantage of the normal variability of rate of uptake of sodium
during the initial equilibration period, the data of Fig. i may be replotted
to show the relationship between internal sodium concentration and rate
of extrusion of sodium. For this plot it is assumed that the sodium extru-
sion during the first 30 min. of recovery can be used as an index of initial
POTASSIUM IN MUSCLE FIBRES 445
rate. Again, no attempt has been made to make these figures more precise
by correcting for extracellular space diffusion.
Fig. 2 shows that, when muscle sodium is above about 60 mmol./kg.,
the rate of sodium extrusion is independent of the internal concentration.
At lower internal sodium concentrations, the initial rate of extrusion is less.
This, of course, is the behaviour that would be expected of a saturable
carrier mechanism of sodium extrusion.
The work previously cited on substituting choline for sodium in internal
and external fluids (Steinbach, 1952) also gives evidence that the rate of
sodium extrusion is independent of external sodium concentration at all
levels from isotonic to one-tenth isotonic.
The effect of external potassium concentration on sodium extrusion
For each sodium ion that is extruded to the outer medium without an
anion, some other cation must enter to maintain electro-neutrality. Nor-
mally this would appear to be accomplished by entrance of potassium,
presumably by diffusion along an electrochemical gradient. If this is true,
then the rate of inward diffusion of potassium might well be a limiting
factor for outward sodium extrusion. To test this, initial rates of sodium
extrusion were measured when sodium-rich muscles were immersed in
recovery solutions containing different concentrations of potassium. The
results, presented in Fig. 3, show that when external potassium concentra-
tion is about 0-02 M the rate of sodium extrusion is optimal. Lower
external potassium concentrations result in lowered rates of sodium
extrusion.
Variable external potassium concentrations have many different effects
on the frog sartorius (Sandow & Kahn, 1952), so the interpretation of these
experiments is by no means easy. However, the provisional conclusion
may be drawn that inward leakage of potassium may be a limiting factor
for outward sodium extrusion.
The temperature coefficient of sodium extrusion
Table i summarizes data relating to the effects of temperature on the
rate of sodium extrusion. The data were obtained for only two temperatures.
A Qw value calculated for the range between these two points is above 3-0.
III. THE MECHANISM OF SODIUM TRANSPORT
The results of experiments outlined above are all consistent with the
following conclusions:
(i) Sodium is moved at a rapid rate outward against an electrochemical
gradient. Compensation for positive sodium charges moved outward is
446
THE REGULATION OF SODIUM AND
70 r
60
50
40
430
o
E
*r
-o 20
10
30 min. 90 min.
o Na o
X K +
10
20
K cone, medium
Fig. 3
30
40
Table i . Average changes in sodium and potassium of muscle fibres during
recovery at 4° C. and at 22° C. Eleven pairs of muscles for each tempera-
ture, sodium and potassium analyses on each muscle.
For calculating fibre concentrations as mmol./kg. fibre, an interspace value of 25 % is
assumed. Statistical treatment of data according to Simpson & Roe, Quantitative Zoology.
Change in concentration
Na
K
22° C.
4°C.
22° C. 4° C.
Average change*
Difference 22°~4
d1<r'd
V
-13
2-3
— II
2-8
3-9
4-15
— 2
I'S
15 3
3*4 i '4
12
37
3-2
3'4
* Average change: concentration in fibres at end of 30 min. recovery minus concen-
tration of control member of pair at end of extraction period and at zero recovery time.
POTASSIUM IN MUSCLE FIBRES 447
effected by the inward diffusion of potassium along an electrochemical
gradient.
(2) The rate of sodium transport outward may be so fast that the rate
of inward diffusion of potassium becomes a limiting factor.
(3) With high-sodium concentrations in the fibre substance, rate of
extrusion of sodium is independent of internal sodium concentration.
With low internal sodium concentrations, extrusion outward is a function
of internal concentration.
(4) The rate of outward extrusion of sodium is independent of external
sodium concentration.
(5) Sodium extrusion is highly temperature-sensitive with a QIQ above 3-0.
All of these conclusions suggest that sodium transport is effected by
some specific chemical carrier device, presumably located near the external
limiting boundary and present in an effective concentration which can
become limiting at high internal sodium concentrations. Furthermore,
combination with sodium to give net outward transport appears to take
place at the internal surface, not the external. This might merely mean
that there are more active combining groups directed inwardly than there
are directed outwardly.
Direct chemical evidence on the nature of specific sodium transport
mechanisms is not at hand. Various vehicles, such as non-aqueous organic
radicals, phosphate combinations and lashing protein tails, have been
proposed from time to time. In frog muscle homogenates an excess,
amounting to a small percentage of the total sodium, is found to be bound
to the sedimentable particulate matter (Steinbach, 1952). At the moment,
there is no special reason for implicating any one type of compound. It
should be emphasized, however, that the carrier compound need not be
concerned exclusively with sodium or any other single ion. All that is
required is a discrimination such that the outward movement of ions by
the carrier pathways should have a higher sodium/potassium ratio than
the inward leakage, presumably by diffusion through pores.
The requirements may be illustrated by a few simple considerations.
The electrical resistance of the fibre membrane is known to be high; a
figure of 1000 ohms/cm.2 may be assumed. If all of the ions carrying current
through this resistance are regarded as moving through a water-filled hole,
then that hole need occupy an area of the order of only one-millionth of
the total surface. Therefore there is a great expanse of surface available for
specific chemical transport systems. Calculations show that even assuming
improbably slow turn-over values for carrier molecules, maximum rates of
transport could be accommodated without crowding of carrier molecules
at the surface.
448 THE REGULATION OF SODIUM AND
If we assume that the external solution has a sodium/potassium ratio of
10 and that diffusion rates within the membrane are in the ratio of 50 for
potassium to i for sodium, then, with zero concentration of both ions
inside originally, the first ions to penetrate would give a sodium/potassium
ratio of 0-2, or an accumulation of potassium. However, as diffusion went
on with no increase in volume, the internal ionic ratio would approach that
of the external fluid, unless there were some outwardly directed carrier
system that had an effective association constant for its sodium compound
that was at least five times greater than the association constant for the
corresponding potassium compound. Furthermore, some mechanism
must exist for the creation (or rejuvenation) of this carrier substance within
the fibre and its destruction externally. Our present information does not
allow us to attempt to name specific compounds. The presence of esterases
on the outside of yeast and other cells might implicate esters as part of the
carrier system, and the polyphosphates are known to be effective binders of
alkali and alkaline-earth metals.
IV. THE PHYSIOLOGICAL ACTION OF INTRACELLULAR
SODIUM AND POTASSIUM OF MUSCLE
There have been many investigations reported in the literature on the
effects of alteration of the sodium and potassium concentrations of the
environmental fluids on muscle action. In general, two types of effect
may be distinguished : the one a rapid effect, occurring soon after applica-
tion of the agent presumably to be ascribed to alterations of external
limiting layers and the other a long-delayed effect, occurring after half an
hour or so of treatment and which may be related to intracellular changes
in the bulk concentration of the ions of the protoplasm. A recent paper by
Sandow (Sandow & Kahn, 1952) reviews some of this literature and
presents new information.
Recently we have carried out a few preliminary investigations of tension
development by sodium-enriched muscles and muscles after they have
recovered for a short period in the high-potassium medium. Isolated
sartorii, treated as usual with a preliminary soaking in potassium-free
solution, were tested with short tetanic stimuli from an induction coil.
Tensions were registered with a simple torsion-wire myograph. Since
muscles immersed in the usual high-potassium recovery fluid (containing
o-oi M potassium) are non-irritable, all muscles were tested for tension
development only after equilibration for half an hour in the potassium-free
medium. Seventeen muscles, soaked 24 hr. in potassium-free solution,
gave an average tension of 94 g./g. of muscle. The corresponding control
muscles soaked for the same period in high-potassium solution gave an
POTASSIUM IN MUSCLE FIBRES 449
average tension of 256 g./g. of muscle. During this period of soaking other
sets of muscles lost approximately one-third of their potassium with a
corresponding gain of sodium. If the individual results are plotted there
is an indication of an almost linear relationship between the highest tension
observed at a given internal potassium concentration and the internal con-
centration of potassium. Marked deviations toward abnormally low
tensions are observed, however, leading to the conclusion that tension is
related to internal potassium concentration but also to other unknown
factors. In another series of twelve pairs of muscles, all muscles were
soaked overnight in the potassium-free medium. One member of each pair
was then allowed to ' recover* for i hr. in the high-potassium medium, then
washed in potassium-free medium to develop irritability, and tensions
developed were compared with the controls which had been carried through
similar manipulations in the potassium-free medium. The 'recovered'
muscles gave an average tension of 150 g./g. of muscle, the unrecovered
controls, 98 g./g. Again, the variations were great so the results must be
regarded as preliminary. However, they are entirely consistent with other
observations on whole animals which show that muscles in animals main-
tained on a low potassium diet, and which develop low-potassium, high-
sodium muscles, are weaker muscles than normal controls (Heppel, 1939).
Preliminary measurements of membrane potential have also been made,
using the Ling & Gerard (1949) technique of membrane puncture with
microelectrodes. We confirmed their results completely with normal
muscles in the usual Ringer's fluid (0-0025 M potassium), the average
membrane potential being 93 mV. Muscles soaked overnight in the high-
potassium fluid (o-oi M potassium) and measured immediately after removal
to room temperature in the same medium gave an average potential of
54 mV., decreasing to 50 mV. in half an hour. Similar muscles, soaked
overnight in the potassium-free medium and removed to the high-potassium
medium and room temperature for measurement gave an average membrane
potential of 46 mV. immediately after exposure to the high-potassium and
48 mV. after half an hour recovery. The change during the recovery period
is not statistically significant, but the fact that the control muscles showed
a consistent decrease in membrane potential during the same period
probably indicates some significant effect of the extrusion of sodium and
uptake of potassium that was occurring in the potassium-depleted tissues.
It should be noted, however, that the loss of potassium during the soaking
period must have amounted to 30-50% of the initial potassium of the
muscle. Hence, if the membrane potential measured really reflected the
ratio of potassium, internal to external, it should have decreased by from
40 to 70 % . Since the decrease amounted to less than 20 % it seems obvious
E B S VIII 29
450 THE REGULATION OF SODIUM AND
that the measured membrane potential depends upon other factors than
the potassium gradient. It should be recalled that Tobias (1950) showed
that waterlogged muscles after all the potassium had been removed by
distilled water still showed respectable membrane potentials.
V. EFFECTS OF SODIUM AND POTASSIUM ON
MUSCLE CONSTITUENTS
While it is true that ionic strength has a marked effect on the activity of the
acto-myosin system, no clear-cut distinction can be made between sodium
and potassium (cf. Szent Gyorgi, 1951; Mommaerts, 1950). So far as the
actual contractile machinery is concerned, there appears to be no property
which might be expected to be altered greatly by change of the intracellular
concentrations of the two ions. However, if we turn our attention to the
enzyme systems, most of which might be expected to be concerned with
the flow of energy to the contractile system (recovery or charging-up), there
are many examples of specific effects of potassium and sodium. In general,
potassium is stimulatory, sodium is inhibitory, although the two need not
be antagonistic (cf. Utter, 1951; von Korff, 1953). From the variety of
information available about ion effects on enzymes, it is difficult to make
much physiological sense. Most ion effects on enzyme systems have been
reported on extracts, and it is known that the effects of a given agent on a
soluble enzyme may be quite different from its effects on the same enzyme
attached to the structural elements of protoplasm (cf. Steinbach, 1949).
However, based mostly on studies of soluble systems, some generalizations
may be hazarded. In the first place, in those instances where potassium
activates an enzyme system (mostly phosphate transfer systems), the
activation is maximal over a broad range covering the concentration of
potassium normally found in muscle (i.e. near o-i M). No sharp ' cut-off'
points are reported, although there are a few systems which will not func-
tion at all in the complete absence of potassium. If potassium exerts
similar effects on enzymes in cells then it seems rather improbable that
potassium, as an activating ion, has much of a regulatory function. From
all of our available knowledge, it would seem probable that the normal
concentration of potassium is always higher than is necessary for very
adequate activation of the enzyme systems concerned with glycolysis and
phosphate transfer. Therefore potassium within the cell is probably to be
regarded as almost indifferent in the sense that there is so much of it
around that the interior never need worry about a lack. So far as potassium
is concerned, activation is probably always normally maximal.
With sodium, however, in those instances where there is a pronounced
sodium inhibition, the case is otherwise. For example, the important
POTASSIUM IN MUSCLE FIBRES 451
acetate-activating system is 50% inhibited by 2 x io~2 M-sodium (von
Korff, 1953). This concentration is seldom found in the interior protoplasm
of the frog muscle fibre but might well be present in the outer layers con-
cerned with sodium transport. Any displacement of the sodium from these
layers by an indifferent ion such as potassium would then not only release
an inhibition rapidly but would also have its effect in a highly reversible
fashion due to the interaction of sodium and potassium. Since the acetate
activation system is already strongly implicated in surface phenomena
concerned with excitation, the pronounced sodium inhibition becomes of
considerable interest. The surface of the resting muscle fibre, complete
with sodium transport mechanism concentrating sodium from the interior
and moving it outward to a strong sodium environment, probably does
represent a 'high-sodium* area of the cell, where the acetate-activating
enzymes are held inhibited. This inhibition might be presumed to be
released immediately following stimulation due to the formation of a new
mixing zone in the outer boundary where the sodium concentration would
be lower than normal and the potassium concentration would be higher.
In addition to effects of sodium and potassium on enzyme systems, other
effects due to structural alterations might be expected. Sodium, a compact
ion of high charge density might be expected to promote complex formation
much more than potassium which is larger with lower charge density
(cf. Steinbach, 1952). Differences of mobility of the two ions have been
strongly emphasized in the past. However, so far as the physiological effect
is concerned, it would seem probable that the effect of an ion depends more
upon what it does when it gets to its site of activity than upon the speed
with which it gets there.
REFERENCES
BATTLEY, E. H. & KLOTZ, I. M. (1951). Biol. Bull, Woods Hole, 101, 215.
CONWAY, E. J. (1945). Biol. Rev. 20, 56.
FENN, W. O. (1936). Physiol. Rev. 16, 450.
HEPPEL, L. A. (1939). Amer.J. Physiol. 128, 440.
VON KORFF. Personal communication.
KROGH, A. (1939). Osmotic Regulation of Aquatic Animals. Cambridge University
Press.
LEVI, H. & USSING, H. (1948). Ada physiol. scand. 16, 232.
LING, G. & GERARD, R. W. (1949). J. Cell Comp. Physiol. 34, 383.
MEIGS, E. B. & ATWOOD, W. G. (1916). Amer.J. Physiol. 40, 30.
MOMMAERTS, W. F. (1950). Muscular Contraction; A Topic in Molecular Physio-
logy. New York : Interscience Press.
MUDGE, G. H. (1953). Trans. 4th Macy Conf. Renal Function.
SANDOW, A. & KAHN, A. J. (1952). J. Cell. Comp. Physiol. 40, 89.
STEINBACH, H. B. (1947). Ann. N.Y. Acad. Sci. 47, 849.
STEINBACH, H. B. (1949). Arch. Biochem. 22, 328.
STEINBACH, H. B. (1950). Amer.J. Physiol. 163, 236.
29-2
452 REGULATION OF SODIUM AND POTASSIUM IN MUSCLE FIBRES
STEINBACH, H. B. (1952). Proc. Nat. Acad. Sci., Wash., 38, 451.
STONE, D. & SHAPIRO, S. (1948). Amer.J. Physiol. 155, 141.
SZENT GYORGI, ALBERT (1951). Chemistry of Muscular Contraction, 2nd ed. revised
and enlarged. New York : Academic Press.
TEORELL, T. (1949). Ann. Rev. Physiol. n, 545.
TOBIAS (1950). Injury and membrane potentials in frog muscle after depleting
potassium and producing other changes by soaking in potassium free salt
solution or distilled water. J. Cell. Comp. Physiol. 36, 1-13.
UTTER (1951). (Personal communication.)
RELATIONS BETWEEN ACTIVE TRANSPORT
AND METABOLISM IN SOME ISOLATED
TISSUES AND MITOCHONDRIA
BY R. E. DAVIES
Medical Research Council Unit for Research in Cell Metabolism,
Department of Biochemistry, The University of Sheffield
I. INTRODUCTION
Most of the early work on what we now call active transport consisted of
investigations with the whole animal and was a study of secretion and
absorption. This has given information about the nature of the secretions
produced by organs such as the stomach, pancreas and intestines, and
about the way in which the activities of these organs are controlled in the
whole animal. For more than fifty years it has been possible to study
secretory activity in perfused organs, but these approaches are not suit-
able for finding out what goes on inside the cells of these organs, and this
is only possible when investigations are made with much simpler systems
under controlled conditions and with quantitative measurements. Within
the last few years this has been accomplished and experiments have been
carried out in many laboratories on isolated slices or sheets of tissue which
have been handled by normal biochemical techniques. These investi-
gations have thrown light on some details of the mechanisms of secretion
and accumulation and have made possible comparisons between the
processes of metabolism, which supply energy, and of active transport,
which require energy supplies. This has led to the discovery that the whole
cell is not the simplest unit which is able to maintain active transport. It is
now known that the mitochondria, which are the structures responsible
for virtually all the respiratory metabolism of cells, are able both to secrete
and to accumulate a variety of inorganic and organic cations and anions.
This last conclusion comes from a line of work which we have followed in
Sheffield for eight years, and has involved investigations on isolated gastric
mucosa, on brain and kidney slices and finally on kidney mitochondria.
II. ION MOVEMENTS IN BRAIN SLICES
The problem of hydrochloric acid production has been reviewed recently
and will not be discussed directly here, but instead some of the work at
Sheffield on the problems of ion transport in brain and kidney will be
considered.
454 ACTIVE TRANSPORT AND METABOLISM IN SOME
Stern, Eggleston, Hems & Krebs (1949) showed that many isolated tissues
could accumulate L-glutamic acid during aerobic metabolism and found
that brain cortex slices were most effective. Although in the presence of
glucose these slices could maintain an internal concentration of at least
about 20 times that in the medium when this latter was low (approx.
i mM), the transport of L-glutamate into the tissue stopped when the
difference between the internal and external concentrations became about
20 mM. This applied over the range from approx. i to 20 mM-L-glutamate
final concentration in the external fluid, and this ability to accumulate
glutamate against a concentration gradient was inhibited in the absence of
oxygen or the presence of 2:4-dinitrophenol.
Further work led to the discovery by Terner, Eggleston & Krebs (1950)
that L-glutamate can play an important role in potassium accumulation in
brain cortex slices; slices of guinea-pig and rabbit-brain cortex incubated
for i hr. in a saline medium lost half their potassium ions, but this could
be prevented if both glucose and L-glutamate were added to the medium.
Once again this effect could be inhibited by 2.'4-dinitrophenol and failed to
occur anaerobically.
Ox retina is an easier tissue to handle, and with it they showed that
sodium movements were approximately equal and opposite to potassium
movements, and that during the recovery phase glutamate and potassium
were taken up in approximately equivalent amounts. There is a discrepancy
here that has not yet been cleared up. If two positive ions exchange and
glutamate enters, then, since glutamate is on balance a negative ion, some
other ion as yet unknown must move to balance the electric changes.
The recovery phase was characteristic of ox retina which had been
brought to the laboratory in ice-cold saline and was found to have a rela-
tively low potassium and high sodium content.
However, a recovery phase was also found to occur in the brain slices
which were found to lose rapidly 40 % of their potassium content within
a few minutes of the start of incubation at 40° C. and then to re-accumulate
this lost potassium during the next half-hour if the conditions were those
already described.
This showed the dynamic nature of the processes involved in producing
the steady level of potassium in the slices ; so the next step was to find out if
this steady level of potassium was itself a reflexion of to-and-fro movements
of this ion. This was likely because work with intact animals had shown
qualitatively that the potassium ions of animal tissues and body fluids
continually interchange (Joseph, Cohn & Greenberg, 1939; Hahn,
Hevesy & Rebbe, 1939; Fenn, Noonan, Mullins & Haege, 1941-2;
Noonan, Fenn & Haege, 1941).
ISOLATED TISSUES AND MITOCHONDRIA 455
Experiments with 42K showed conclusively that the potassium was
rapidly exchanged even though there were no net changes in concentration
(Krebs, Eggleston & Terner, 1951). The average turn-over rates for brain
were between 3-5 and 4'0%/min. and for retina 7 and io%/min. These
rates are very high compared with the turn-over rate of potassium in red
blood cells which is 0-03 %/min. (Raker, Taylor, Weller & Hastings, 1950;
Sheppard & Martin, 1950; Solomon, 1950). Thus in brain potassium
exchanges on average about 120 times, and in retina about 250 times more
rapidly than in human red cells.
The experiments were not designed to study the rate of the net accumu-
lation, but for brain this must have had a QK (uptake) of at least —13
expressed as /d./mg. dry wt./hr. (where 22-4 jul. = i ^mole). This can be
compared with the steady-state exchange with a QK (exchange) of ±32.
The rate of oxygen uptake expressed in the same units was Qo2= — 18.
On the assumption that this turn-over was caused by an active uptake and a
passive leakage Krebs et al. (1951) calculated that this turn-over rate would
require at least 2-5% of the energy available from metabolism in brain.
Since it was not easy to test this assumption in brain tissue because of
the swelling that occurs during incubation, Mr R. Whittam and I turned
to kidney to investigate the problem of the turn-over of sodium with this
tissue.
Now work with frog muscle and giant axons of invertebrates, which have
a negative intracellular electric potential, had made it probable that a sodium
pump operates to maintain the observed ionic concentration differences
in these tissues (Dean, 1941; Krogh, 1946; Hodgkin, 1951), and that
potassium movements are largely or entirely passive with no change in
electrochemical potential. Thus the work done in moving this ion inside the
cell against a concentration gradient would equal the work gained by moving
the positive charges to the negative inside of the cell (cf. Ussing, 1952).
Although the electric potentials inside brain or kidney cells have never
been measured, it was important to investigate sodium movements and
turn-over in case a sodium pump operates in these tissues.
The swelling of brain slices in saline solutions, which has been mentioned
already, was discovered by Elliott (1946), and Stern et al. (1949) showed
that the slices swelled about 25 % aerobically but 50% anaerobically ; thus
respiration is an important factor in controlling the fluid uptake of the
tissues. They also found that kidney slices maintained their original wet
weight during incubation, and this makes them suitable for investigations
on sodium turn-over which would be difficult with tissues which swell,
because of the problem of allowing for the sodium in the fluid imbibed by
the swollen tissue.
456 ACTIVE TRANSPORT AND METABOLISM IN SOME
III. ION MOVEMENTS IN KIDNEY SLICES
Net changes of ions
It was first necessary to find suitable experimental conditions and to
measure the sodium, potassium and water content of the guinea-pig kidney
cortex slices under a variety of conditions. It was known from the work of
Krebs et al. (1951) that when no substrate was added about one-third of
the tissue potassium was lost on aerobic incubation, and about two-thirds
. Na-HK*
-X Na4
0 5 10 20 30 40
Time (mm.)
Fig. i. Changes in the sodium and potassium content of guinea-pig kidney cortex slices
on incubation in vitro at 37° C. in bicarbonate saline containing 155 mM-Na1, 5 mM-K+
and 10 mM-a-ketoglutarate ; gas 5 % CO2 in 95 % O2 .
on anaerobic incubation for 30 min. at 40° C. in a bicarbonate saline
containing 5 mM-potassium ions and gassed with 5 % carbon dioxide in
oxygen. Addition of a-ketoglutarate prevented this loss, L-glutamate and
L-aspartate reduced it, but pyruvate, succinate, citrate, fumarate and
glucose had no effect.
The initial experiments showed that kidney cortex behaved in a very
similar way to brain cortex. The tissue was sliced with a dry razor and on
aerobic incubation at 37° C. it was found that there was an immediate loss
of potassium and gain of sodium during the first 2 or 3 min. (Fig. i).
ISOLATED TISSUES AND MITOCHONDRIA
457
These changes were largely reversed during the next 20-30 min., after
which a steady state was maintained for a further 40 or 50 min. After
still longer periods the slices began to leak out potassium once again and
the sodium content increased.
Table i. The relation between the changes in sodium and potassium and the
net accumulation of oc-ketoglutarate during aerobic incubation of guinea-
pig kidney cortex slices for 40 min. at 37° C. ( ± standard error)
All concentrations given as /onoles/g. tissue
± 1*9
No. of exps.
19
H
33
13
AK+
ANaf
/. A(Na+4K+)
A a-ketoglutarate = -f- 12-0 ± 0-58
.'. . -: . - — _ =4- 2- 1 4 0-17 —
Aa-ketoglutarate ~
a-ketoglutarate metabolized in 40 min. = 53'3±5'46 15
Table 2. Effect of a-ketoglutarate concentration on the potassium
content of guinea-pig kidney cortex slices
Slices (150 mg.) incubated at 37° C. for 40 min. in 2 ml. bicarbonate saline gassed with
5%C02in02.
Initial a-keto-
glutarate concen-
tration in medium
(mM)
Amount of potassium in tissue (/^moles/g. tissue)
Before
After
Change (%)
0
2
4
6
8
10
78-8
78-8
78-8
78-8
78-8
78-8
63M
65-8
73'7
78-3
78-5
79'5
— 20
-16
- 6-5
- 0-6
- 0-4
+ 0-9
The gains and losses of sodium and potassium were approximately equal
and opposite, but, whereas the lost potassium was completely recovered
under these conditions, some of the sodium was not extruded. Measure-
ments of the pH of the medium with a glass electrode showed that no
changes at all could be detected, so some other ion appeared to be involved
since the excess of positive ions accumulated had to be accounted for
somehow.
Measurements were therefore made of the movement and metabolism of
the substrate a-ketoglutarate with results given in Fig. i and Tables i
and 2. It was first confirmed that concentrations of a-ketoglutarate lower
than 6 mM were less effective in maintaining the potassium concentration
(see also Krebs et al. 1951), and found that there was an accumulation of
a-ketoglutarate inside the slices. Although the fresh slices had a very low
458 ACTIVE TRANSPORT AND METABOLISM IN SOME
content of this anion (0-15 mM) this increased during incubation and could
reach at least 2-5 times the concentration in the medium. This is comparable
with the situation in brain where the specific substrate required for
potassium re-accumulation, the L-glutamate anion, is also accumulated
against a concentration gradient. If, as is likely in both brain and kidney,
the insides of the cells are negative relative to the outsides, then these
accumulations of L-glutamate and a-ketoglutarate must be accomplished
not only against concentration gradients but against electric potential
gradients. This is a process for anions quite similar to the extrusion of the
cation, sodium, in nerve and muscles which takes place against an electro-
chemical gradient.
In these experiments on kidney, in contrast to those on brain, there was
no ionic discrepancy, since the increase in sodium plus potassium, which
was almost all an increase in sodium, was exactly twice the increase in
a-ketoglutarate ions. This means they were electrically equivalent, since
a-ketoglutarate is a dibasic acid and virtually completely ionized at the
experimental pH values. Another difference between the two tissues is
that in brain both glucose and L-glutamate are required to recover the
potassium, whilst in kidney a-ketoglutarate is completely effective on its
own and the addition of glucose makes no difference.
It is of interest that four times as much a-ketoglutarate is metabolized
and disappears from the system as is accumulated and found inside the
tissue slices (Table i). This makes it probable that the actual accumulation
is much greater than the net accumulation because all oxidation of a-keto-
glutarate occurs inside the cells on the mitochondria. Further work, which
is in progress, is therefore needed on the fate of this substrate before the
situation can be clarified.
When these preliminary experiments had been completed Mudge (195 1 a)
published experiments on rabbit-kidney cortex slices with rather different
experimental conditions. The kidney cortex was wet sliced at 2° C. and the
slices were then leached in 0-15 M-sodium chloride for 2-3 hr. at room
temperature and kept for a further 40 min. at 2° C. in the absence of
substrate, phosphate buffer, calcium or potassium. They were then incu-
bated at 25° C. in a medium containing all these substances but with
a potassium content of 10 HIM, which is twice that in a physiological saline
solution.
Under these conditions Mudge (19510) found that after the period of
leaching and cooling the slices had lost potassium and gained sodium. He
showed that during the subsequent incubation these changes could be
largely reversed and stated that * Changes in tissue Na are the reciprocal
of K'. However, his results all showed an overall increase of sodium in
ISOLATED TISSUES AND MITOCHONDRIA 459
the tissue even when the lost potassium had been recovered. This increase
was probably associated with an accumulation of the substrate, but this is
uncertain because no measurements of substrate changes were recorded.
Mudge (1951 a) found that several substrates were effective in supporting
ion movements, and these included pyruvate, succinate, citrate and
fumarate which Krebs et al. (1951) had found to be ineffective. The
explanation for this is not clear, but it may reside in the use by Mudge
(19510) of a relatively low experimental temperature and a high potassium
content in the medium, both of which may help to make it easier for the
slices to move these ions, and thus allow other substrates, besides a-keto-
glutarate, to support the active transport.
This suggestion that the conditions are less critical at 25° C. than at
40° C. is supported by our finding that inhibitors of carbonic anhydrase
lower the potassium content of kidney slices at 40° C. (Davies & Galston,
1952) but not at 25° C. (see also Mudge, 19516).
The initial rates of ion movement found by Mudge (1951 a) (calculated
from Figs. 2 and 3) were a QK (accumulation) of — 12 and a £)Na (extrusion)
of + 1 6. However, if the rate of sodium extrusion is taken from the tangent
of the smooth curve rather than from the first two points, the QN& becomes
+ 33. These results were with o-oi M-acetate as substrate at 25° C. and
the QOz was only —3-8.
The net rates found in our experiments are given in Table 3. It should
be noted that the initial rates during the recovery phase are likely to be too
low because under our conditions there was a smooth transition from the
period of leakage to the period of recovery. The results are therefore
minimal values.
Table 3. Rate of respiration and maximum net rates of change of
sodium and potassium in guinea-pig kidney cortex slices
Incubated aerobically at 37° C. in bicarbonate saline containing 10 mM-a-ketoglutarate
cf. Fig. i). All Q values are in /vl./mg. dry wt./hr., where 22-4/^1. = i /tmole.
Ion change
Kh leakage out
Kf accumulation
0
+67
-25
Naf leakage in j —93
Na+ extrusion
Oxygen uptake (20-60 mm.)
Before considering the dynamic exchanges during the maintenance of
the steady-state conditions there are a few other points relevant to the net
changes observed. These concern the importance of aerobic metabolism,
the effects of temperature and the presence of a water pump.
460
ACTIVE TRANSPORT AND METABOLISM IN SOME
125
100
75
50
25
-Tissue transferred to
medium at 0°
15 30 60 _ 90 120
Time (mm )
86
84
82 5
80
78
76
Fig. 2. Effect of 2 x 10 4M-2 : 4-dinitrophenol on the sodium and potassium content of
guinea-pig kidney cortex slices incubated aerobically at 37° C. with TO mM-a-keto-
glutarate as substrate.
140 -
120 -
100
40
20
DNP added at start
Na
r*
DNP added Control
at 35mm.
10
15
20 25 30
Time (mm.)
35
40
45
50
Fig. 3. Changes in sodium, potassium and water content after transferring guinea-pig
kidney cortex slices, which had been pre-incubated aerobically at 37° C. for 35 min., to
the same medium at o° C.
ISOLATED TISSUES AND MITOCHONDRIA 461
The accumulation of potassium and a-ketoglutarate, and the extrusion
of sodium all depend on active aerobic metabolism and did not occur
either anaerobically at 37 or o° C. when there is no respiration, or aerobi-
cally at o° C. when respiration is very low. At o° C. a-ketoglutarate
addition had only a slight effect on the Q02, and this substrate was not
accumulated. This is in contrast to 37° C., where a-ketoglutarate was
accumulated and its addition increased the QOz by 69%.
These effects of temperature and lack of oxygen, which were accom-
pdnied by a swelling of the tissue slices, were reversible (Table 4) (see also
Mudge, 1951 a] Robinson, 19500; Aebi, 1952). As in the case of brain
this swelling was far greater than could be accounted for by the difference
between the water content of the hydration spheres of the sodium ions,
which increased, and that of the potassium ions, which decreased, in the
absence of active aerobic metabolism, so some other mechanism must be
invoked for this 'water pump'.
Respiration alone is not sufficient to maintain these manifestations of
active transport because 2:4-dinitrophenol could stop them (Fig. 2) with-
out affecting the initial rates of leakage or inhibiting the rate of oxygen
uptake, which was in fact somewhat increased (Mudge, 19516; Robinson,
19506). These results make it probable that both ion transport and the
ability to maintain the normal water content of the kidney slices are
mediated through high-energy phosphate compounds.
However, it seems likely that these two mechanisms are not very closely
linked together, for the following reasons. When the slices were incubated
at o° C. swelling was complete within 10 min., but the changes in sodium
and potassium were still occurring slowly after 2 hr. (Fig. 3). Under these
conditions there was normally almost no gradient of sodium plus potassium
between the tissue water and the medium (Table 4), but if the slices were
transferred to medium oxygenated at 37° C., water was pumped out of the
slices and at the same time they developed a gradient of sodium plus
potassium. Thus the total content of these ions increased whilst the water
content decreased. These results favour the view that kidney slices actively
metabolizing in oxygen are not in osmotic equilibrium but maintain
hypertonic internal compartments by the expenditure of metabolic energy.
Steady-state turn-over of ions
As in the case of brain it was important to find out whether the steady-
state concentration differences of ions were maintained by the cells
because they were impermeable to these cations, or whether there were
to-and-fro exchanges, and if so what were the relations between these
exchanges and the energy supply.
462
ACTIVE TRANSPORT AND METABOLISM IN SOME
K
s
c c
o c
§
<X>
R
K 7
p ±'
•Ki ° W
1 c°S
1 1I
^i O
1 ^
J^ g Z3
•2 ^1
5 |.S
^ |i
y- §2
I -°s
. ?? l-H f \
a,
s"
.a
"s
^«
jy
3
E2
Is
gg
ii
o
I
CO
B S
4j CO
T B
b
+1
ON
b
+1 ON
O
b
-fl
<<f
O
b
+1
O
b
+1
O
b
+1
00
b ,
rf O
N CO
J!
CO
11
'l-t TJ
2^? §•
ISOLATED TISSUES AND MITOCHONDRIA 463
It was first necessary to find out if all the sodium and potassium in the
kidney slices were exchangeable with sodium and potassium in the
medium, so slices were incubated in physiological saline solutions con-
taining spectroscopically pure 24Na+ or 42K+ and were removed after
various times. The results are shown in Figs. 4 and 5. They show that,
within the experimental errors of 2 or 3 %, the whole of these ions exchanged
rapidly, with sodium exchanging faster than potassium. The rates were
slower under strictly anaerobic conditions, but the turn-overs were still
complete. This appeared to be in contrast to the findings of Mudge (1952),
that at 25° C. under anaerobic conditions only 60% of the potassium in
rabbit kidney cortex slices had exchanged after 4 hr. Experiments were
therefore carried out at o° C. with guinea-pig kidney cortex slices and these
showed that even after 10 hr. about 20%, aerobically, and 40%, anaero-
bically, of the tissue potassium had failed to exchange despite rapid initial
rates of exchange. Thus at o° C., in addition to the small amounts of
extracellular potassium, there were two forms of tissue potassium which
exchanged at fast and slow rates (Table 5).
The turn-over rates of both sodium and potassium given in Table 5 were
calculated from the experimentally determined curves showing the uptake
of the radioactive isotopes with time. In these experiments the slices had
been pre-incubated and were in steady states with only slight changes in
the concentrations of these two ions. In contrast to potassium the rapidly
exchanging extracellular sodium could not be neglected, and the results
given refer to the intracellular sodium. Since the concentration of sodium
and potassium varied so much with the conditions (see Table 4), the rates
per gram of tissue may not be a fair basis for comparison. The rates are
therefore given as /^moles turned over per 100 /^moles of intracellular
cation in the tissue.
The turn-over rates given in Table 5 for the aerobic slices are the
fastest so far found for sodium and potassium in any mammalian tissue
(cf. Davies & Galston, 1951), and it is remarkable that in kidney an amount
equal to the whole intracellular potassium exchanges in less than 7 min.,
i.e. more than 200 times a day. Table 5 also shows that, although the
contents of sodium and potassium depended on respiration (Table 4), the
turn-overs of these ions are largely independent of respiration, the rates in
anaerobic conditions being about 75 % of those in the presence of oxygen.
The remaining quarter of each turn-over rate is associated with aerobic
metabolism, and for sodium, and the slow fraction of the potassium, shows
large changes with temperature. The direct effect of temperature on the
slowly exchanging part of the tissue potassium was very large, the increase
aerobically being iO4-fold and anaerobically i84-fold.
464
ACTIVE TRANSPORT AND METABOLISM IN SOME
100
80
S 60'
40
20
Aerobic
50 100
Time (min.)
150
Fig. 4. Exchange of 24Na+ in isolated guinea-pig kidney cortex slices at 37° C. Changes in
sodium content were occurring throughout the experiment, but from 30 to 50 min. the
aerobic slices contained 112 mmoles Na+/l. tissue water, and the anaerobic slices contained
135 mmoles Na+/l. tissue water.
Aerobic
150 200
Time (min.)
Fig. 5. Exchange of 42Kf in isolated guinea-pig kidney cortex slices at 37° C. Changes in
potassium content were occurring throughout the experiment, but from 30 to 50 min. the
aerobic slices contained 96 mmoles K+/l. tissue water, and the anaerobic slices contained
35 mmoles Kf/l. tissue water.
ISOLATED TISSUES AND MITOCHONDRIA 465
These results show the relations between the aerobic metabolism of
isolated guinea-pig kidney cortex slices, the amount of water in the tissue
and the content and turn-over of both sodium and potassium. It follows
that even so-called resting cells must do work continuously to maintain
their internal ionic environments. However, the amount of this work
cannot be calculated straightforwardly from measurements of the turn-
over rates of the ions in aerobic conditions without reference to the situation
in, for example, anaerobic conditions (cf. Krebs et al. 1951). It thus
becomes important to know the location in the cells of the enzyme systems
which do the work which maintains this active transport.
Table 5. Steady-state turn-over rates of sodium and potassium in
guinea-pig kidney cortex slices
(All turn-over rates for sodium refer to the slowly exchanging fraction.)
Increase of
turn-over
on
Gas
phase
Temp.
Turn-over rate
(/tmoles of
cation/min./
Ratio of
turn-over
rates
37° C./o° C.
rate due to
02 at 37° C.
divided by
the increase
of turn-over
rate due to
O2 at o° C.
^
o,
37
16-21
1-81
o
8*0 1
r > 2O*
N2
37
12' I
o
8-9)
I"35
C+
os
37
o
Fast 1 8- 1
0-88
1 0-82
j
Slow 0*155
104
f
N2 37
o
"'5 I
Fast 12-71
0-91
—
Slow 0-0622
184
47
Zeo-Karb 225 37
II'O
(10-30 mesh) o
7-1
K salt in
o-oiM-KCl
* Calculated after making allowance for the maximum possible error in the measure-
ment of the turn-over rates.
IV. ACTIVE TRANSPORT IN MITOCHONDRIA
The experiments just described refer to the activity of intact cells in
kidney cortex slices and were done in collaboration with Dr A. W. Galston
and Mr R. Whittam. The work to be presented now shows that these
phenomena are largely reflexions of more fundamental events at the sub-
cellular level, and was done in collaboration with Mr W. Bartley. We now
know that for some activities the elementary secretory units are not the
cells but the mitochondria.
466 ACTIVE TRANSPORT AND METABOLISM IN SOME
During the last few years much work has been published on the chemical
activities of various fractions of the cell, and this has shown that the mito-
chondria are responsible for virtually all the respiration and oxidative phos-
phorylation that goes on in the cell (see reviews by Green, 1951 ; Schneider,
1953). In view of the connexion between active transport and oxidative
phosphorylation it therefore seemed reasonable to suppose that mitochon-
dria were, energetically, closely associated with secretory activity in the cell.
Active transport by cells is nearly always in one particular direction, and this
means that cells must have special structures which cause the cells to direct
secretions to one particular wall, for example, of the cell. In secreting
kidney cells the mitochondria are arranged longitudinally from the cell
wall bordering the lumen of the tubules, so it seemed worth while investi-
gating the possibility of a still closer association and hence to look for
secretory activity in isolated actively metabolizing mitochondria. This
possibility was supported by the well-known fact that the osmotic strength
and composition of the suspending medium is of critical importance for
maintenance of the size, shape and activity of isolated mitochondria. It
was also known that potassium was necessary for maximum oxygen uptake
in liver mitochondria (Pressman & Lardy, 1952).
Harman (1950) had investigated the distribution of potassium between
mitochondria and the suspending solution but did not find any evidence
for an accumulation. However, the particles he had isolated had been
kept for more than 20 min. at o° C., and this would have drastically reduced
their rate of metabolism and hence any active transport that depended upon
a continuous supply of energy. This was confirmed and the technique
developed of separating the actively metabolizing mitochondria from the
incubation medium as quickly as possible at 20° C. in the high-speed head
of an International Centrifuge which took only 24 sec. to pack down the
mitochondria in potassium chloride solutions. Under these conditions it
was possible to observe a rather labile ability of the particles to maintain
concentration gradients which was dependent on the experimental condi-
tions. Table 6 shows that well-oxygenated, actively metabolizing kidney
cortex mitochondria can actively transport a wide variety of organic and
inorganic cations and anions (Bartley & Davies, 1952). In all the cases
shown there were no, or only small, gradients in the initial material, and
the extra ions were accumulated during metabolism. At very low external
concentrations of sodium (6 x io~4M) the freshly prepared material could
maintain a ratio of 26. This may have been due to some * bound* sodium in
the mitochondria. Similarly, ratios for magnesium of up to 4-5 were
observed, but there were no clear effects of metabolism on the concen-
tration gradients or the absolute content of this ion. It is possible that the
ISOLATED TISSUES AND MITOCHONDRIA 467
gradients for magnesium are maintained by a type of 'binding' process
similar to that described recently for calcium by Slater & Cleland (1953).
The results given in Table 6 are for the centrifuged material without
allowance for the extraparticulate fluid, so the ratios for the cations must
be minimum values. The results for the substrate anions are also un-
corrected for the amounts used up during the isolation of the particles.
This explains the low ratio found for oxaloacetate, and shows that pyruvate,
fumarate and a-ketoglutarate were concentrated in the mitochondria.
Since both cations and anions can be accumulated and this accumulation
depends on metabolism, any explanation along the lines of a Donnan
equilibrium seems most unlikely. It is remarkable that an important
function of the cell, the ability to pump water, can also be carried out by
mitochondria, since the water content of these particles increased in the
absence of metabolism and decreased during active metabolism. Similar
results with liver mitochondria have been found by Macfarlane & Spencer
Table 6. Ratios of internal f external concentrations found for meta-
bolizing skeep kidney cortex in mitochondria at 20° C.
Substance
Ratio
Concentration in
medium after
separation (M)
<H+'
Na4
K+
Orthophosphate
Adenosme polyphosphates
Pyruvate
Fumarate
Oxaloacetate
a- Ketoglutarate
Citrate
2'5
i'5
2-4
6-0
0-7
i'i
8-0
O'i
I'D
0-8
1-6 x jo 7
2'6 X IO~2
9-0 x io"2
i'9X io~4
4-3 x 10 4
3-5 xio-3
2-8 x io"4
1-8 x io"3
6-3 x io~2
1-4 x io *
Water content of metabolizing mitochondria = 80 %.
Water content of non-metabolizing mitochondria = 91 %.
It was important to find out whether these ionic differences were due to
static accretions of bound ions or whether they were manifestations of
dynamic activities, and if so what were the relations between the rates of
ion exchange and metabolism. Once again this problem could be investi-
gated with the help of radioactive isotopes.
The whole of the sodium and potassium was exchangeable with extremely
high turn-over rates. These rates were so high as to be immeasurable with
our techniques at 20° C., even in the absence of substrate. Fig. 6 gives
the results obtained at 20 and at o° C. with sodium and potassium. This
shows the effect of added substrate at o° C. on the sodium turn-over. It
30-2
468 ACTIVE TRANSPORT AND METABOLISM IN SOME
was found that the increase in £>o2 following the addition of fumarate was
only 0-5, but the increase in QN& in the first minute was 2000. Thus the
uptake of one molecule of oxygen was causing the uptake and output of
4000 sodium ions ; with a-ketoglutarate this figure was at least 6400 sodium
ions. These results appear to rule out many simple ideas on the mechanism
of sodium transport. It is possible that the extra oxygen uptake in the
presence of substrate is used by the mitochondria to make available carrier
molecules which can themselves transport the ions to and fro at very great
rates. It is noteworthy that at o° C. in the absence of substrate the sodium
does not exchange uniformly (Fig. 6). There is a fraction which turns over
100
80
*>
| 60
<
j
So
20
1 2
Time (mm )
(0)
20
40 60 80
Time (min.)
100
Fig. 6. Exchange of (a) 24Na+ and (b) 42K+ in sheep kidney cortex mitochondria.
(a) x with no substrate at 20° C. (100 % exchange by the first measurement also occurred
with 2'$ mM-a-ketoglutarate or 2-5 mM-fumarate at 20° C.); • with 2-5 mM-a-keto-
glutarate at o° C. ; O with 2-5 mN-fumarate at o° C. ; • with no substrate at o° C.
(b) • with no substrate at o° C. (100% exchange by the first measurement also occurred
with no substrate at 20° C. and with 12* 5 -mM-a-ketoglutarate at o or 20° C.).
very slowly, but its turn-over rate can be greatly increased by increasing
the rate of metabolism of the mitochondria. Thus sodium turn-over is
closely linked to mitochondrial metabolism.
In the case of potassium the turn-over rate was immeasurably fast under
all conditions. It was not possible to show any effect of an increase of
metabolism on the rate of potassium exchange, and this means that if the
exchange did depend on metabolism then at least 10,000 potassium ions
were exchanged for every oxygen molecule used by the mitochondria.
These steady-state exchange rates are so high that they rule out any
explanation based on an active uptake and a passive leakage. Two reasons
will clarify this. One is that no simple reaction could form a stoichio-
chemical link between several thousand sodium ions and one oxygen
ISOLATED TISSUES AND MITOCHONDRIA 469
molecule. The other is a difficulty concerning the energetics. Even on the
basis of the formation of three high-energy phosphate bonds for each
oxygen atom this would only make available 15 cal./mole of sodium ions
transported, and this could only produce a concentration gradient of 1-03
even if it were used with 100% efficiency. Much greater concentration
gradients are actually maintained (Bartley & Davies, 1952; Macfarlane &
Spencer, 1954), so this possibility is excluded.
This means that in these rapid steady-state exchanges there must be
energetic coupling of the accumulation and the leakage, and this could
occur in two types of way. One way is a balance between two forms of
energy, so that, say, a gain of concentration energy goes with a loss of
electrical energy as in the exchanges taking place at equilibrium between
ion-exchange resins and salt solutions. This is only possible when net
transport involves no change in electrochemical potential, and this is
unlikely to be the case in mitochondria even if they behaved like a mixture
of a cation and an anion exchange resin which could accumulate and
exchange both cations and anions. The second way is possible when net
transport involves changes in electrochemical potential and must occur in
mitochondria. This is the * exchange diffusion', postulated by Ussing
(1949), which requires a fully saturated ion-carrier moving to and fro by
thermal motion between the two compartments. Thus when one ion loses
energy, another gains it. This allows an exchange of ions to proceed
without net transport or an energy supply.
This idea is easier to visualize if one thinks of the Eiffel Tower during
peak periods. The top lifts are always full and are linked, so that when one
moves up another moves down to balance it. If the lifts were frictionless
there would be no net work done. There is, in fact, no net transport of
people, but, nevertheless, there is mixing and exchange of those at the
top and those at the bottom, and any one individual has had his potential
energy increased and again decreased.
The possibility of ' exchange diffusion ' must always be borne in mind in
interpreting steady-state exchange rates measured with radioactive isotopes.
The net uptakes of cations and anions against electrochemical gradients
must of course require energy, so an investigation was made into phosphate
metabolism which is a link between the energy and the transport.
The results so far obtained throw light on the mechanism by which
phosphate is transported by mitochondria and were obtained using radio-
active phosphate and radioactive adenosine triphosphate (ATP). In the
first experiment a washed sheep kidney mitochondrial suspension in
potassium chloride was added to a medium containing pyruvate, bicarbon-
ate, magnesium, chloride, ATP and phosphate buffer, containing labelled
470 ACTIVE TRANSPORT AND METABOLISM IN SOME
phosphate. This system was able to maintain its organic phosphate content
and rapidly incorporate radioactivity into the ATP, so that isotopic
equilibrium with the phosphate in the medium was 50 %, reached in about
3-4 min. when oxygenated at 20° C. Fig. 7 shows the results obtained
when the mitochondria and medium were separated at different intervals
of time. In this experiment there was a large accumulation of phosphate
within the mitochondria which reached a maximum after 4 min. However,
I2r
10
200
I
I
I
I
12345
Time (min.)
2345
Time (mm.)
Fig. 7. Changes in the concentration (a) and specific activity (b) of the orthophosphate in
sheep kidney cortex mitochondria and in the medium during incubation at 20° C. with
32PO4 in the medium at the start. For further details see text.
the fall in the specific activity of the phosphate in the mitochondrial
fraction after the first half-minute in the face of a net increase in the
amount of phosphate within the mitochondria and a constant specific
activity outside showed that the phosphate accumulating within the mito-
chondria did not come directly from the external orthophosphate. The
only possible source of this non- radioactive phosphate was the ATP in the
medium which, on breaking down, was acting as a carrier of phosphate
from medium to mitochondria.
It is important to realize that the specific activity calculated for the
mitochondrial phosphate includes the activity of the ' extramitochondrial1
ISOLATED TISSUES AND MITOCHONDRIA
471
phosphate in the spaces between the particles. The inulin space of the
packed-down suspension was measured and gave the high value of 60%,
so these particles may have a system of intercommunicating passages
rather like a sponge or loofah. When allowance was made for this, the
specific activity of the intramitochondrial phosphate was calculated to be
only 50 % of the activity of the medium at the first measurement, so some
12
10
z
E
300 -
Mitochondria
Supernatant
1 2 3
Time (mm.)
(a)
Mitochondria
Supernatant
1234
Time (min.)
Fig. 8 Changes in the concentration (a) and specific activity (b) of the orthophosphate in
sheep kidney cortex mitochondria and in the medium during incubation at 20° C. with
[0 y — 32P,] ATP in the medium at the start. For further details see text.
external phosphate must have got in on its own. It is clear, however, from
these results that the large increase of non-radioactive intramitochondrial
phosphate can only have originated from the added ATP.
This conclusion was confirmed by repeating the experiment with ATP
labelled in the two terminal phosphate groups (Fig. 8). These results show
that in contrast to the previous experiment the specific activity of the
mitochondrial phosphate was higher and increased more rapidly than that
in the medium. This could only occur if the mitochondrial phosphate came
472 ACTIVE TRANSPORT AND METABOLISM IN SOME
from the radioactive ATP in the medium. Thus ATP is identified as the
carrier molecule for accumulating phosphate from medium to mitochondria
(Hartley & Davies, 1954).
These observations with mitochondria are in accordance with similar
results for whole cells, since it is now known that the absorption of inorganic
phosphate is often a process requiring phosphorylation to form an organic
phosphate, rather than a simple exchange of ions across cell membranes.
This has been shown for heart muscle by Sacks (1948), for body cells in
general by Popj&k (1950), for red cells by Gourley (1952) and for yeast
cells by Nickerson & Mullins (1948) and Spiegelman, Kamen & Sussman
(1948). It now seems probable that all these cases reflect the powers of
active transport possessed by subcellular particles, and it is predictable
that many other types of active transport shown by intact cells will be
explicable in terms of the activity of enzyme assemblages within the cells,
and in particular of those complex assemblages of enzymes the mito-
chondria. The multicompartmental structure of these particles has already
been demonstrated convincingly by Sjostrand (1953), and this gives
a physical basis for their secretory activity which has to be postulated to
explain these experimental results.
V. THE PROBLEM OF THE EFFICIENCY OF
ACTIVE TRANSPORT
There have been so many theories put forward recently to account for
active transport that there is an urgent need for experiments to test them,
rather than for more theorizing. As a contribution to this, Table 7 gives
the collected results of the experiments which have been discussed here
for the relations between metabolism, net transport and turn-over in those
cases where the information is available. These results show that, though
the rates of net movement of sodium and potassium against concentration
gradients can be accounted for by the possibilities given in Table 7, (c) or
(d\ but in many cases not by (a) or (6), the steady-state rates of ion turn-
over must be caused by other mechanisms. This is confirmed by the
occurrence of large rates of ion turn-over in anaerobic conditions where
net transport against gradients does not occur. Many more such com-
parisons are required, and a start has already been made for the problem of
hydrochloric acid and electricity production in gastric mucosa (Crane,
Davies & Longmuir, 1948; Gray, Hokin & Rehm, 1948; Teorell, 1949;
Crane & Davies, 1951; Rehm, 1950; Davenport, 1952; Davenport &
Chavre, 1951, 1952), of salt uptake in plants (Robertson & Wilkins,
19480:, b\ Lundegardh, 1949) and of electricity production in frog skin
(Francis, 1933; Stapp, 1941; Lund & Stapp, 1947).
ISOLATED TISSUES AND MITOCHONDRIA
473
8" «
^ .••*»
^ S •1^U'S»
S« "f*! C 1) 4-J
o A
s s °
0 |
g s N
t^ Mioooooo
1 *&
r^oo NMO^tOMM^t- ro
d co N 0 v<
^^ ^ & 8
OM •^•CO^ONMM rooo "^ O
O O co M
•O Tj- lO
V 2
t^ N
A A
6°<
A
C £'
*g 2 *oo
00
00 CO vO
« o'obx
00 CO CO O
oo Y* 9 9 oo
4^ ,JH fv^ rj ---
M N
N 0 *
H M
0? & ^ >
III 1
1 1
1
s S"
lih3 1&
i § § h?
* > g § bt.C
18-^4?
<U - * <N-< ?
o i i coor^ooMfooP
^ 1 1 a \O rl-NiOMO
1 + +1 +1+1 +1+1
<??; 1 1
r i
II co »o
+1+1
> C r<n 5L
Q O.S°^
C | M
^ ^-V^-% x-^ ^-V
+~^ -^ "
-s^-v
«*M §4J
P^P +rt 4^ fed fc3 4~N+^+c3 +~"+rt + ^ ^
^"J> ««+ 03 W+ + +
Q} £ H^ ^ ^^ ££^^£^£^*'
6£ ^^^
' ^ K/ K/1 V
_, ^ hM hM HM
*J ^ ^>
l-< c "^
CO OJ \O N CO M f>. iOvO O co^O
MCO MM O^'-'ONNOOiOMO
°?^ 0 0<
10 O^ O O <
D 0 0 O <t
3 O O to rj-
1 s g
1 +1 + 1 1 ++ 1 +I+I+I+I+I
| | xO IONX
3 00 >0 CO 10
11 M" to co -f 10
« ff +1 +1+1 +1+1 +1+1
cti 0
O c "S
GO
A
A A
|cj
o o r^ o
-f N co
oo o o o o r-
CI CO
OJ 1)
*j O)
« i 03 0 ^
03 ! c -< 4-»
S
o>
4-1
-c
!~s sS
fj ,_ 03 in ij
B 2 o 2
3 <u J2 aj 2 i>
0^ 8 ^^
o § r
= bd D S
3 hM T*. O
i- < i60
j z p
s « M^;
03 . 03
-
1
1J
4-1
03
%
^ t« 03
co z o> a;
{50 D C o OJD O
'as j}^ 'a'S
03 ^ X 03 X
<U C « 1) <U <U
5 03 ^ t S £
3^-0 3 0
O^tf U0U
Rat-liver
mltochondri
Sheep-kidney
mltochondri
w
• i
o o
CO
^ £ q^
^
O
fl, O to
o IO
0
c
H
S s~*' o^i M
r^ M
o
^ H >»— '
JS i-, ^s*^
5 ^^
s^ &S S^
03 0 oT S
03 c/)
*
a3 10 ^ 10 •« ^
»-j !> ^J O^ ^ 03
*C§C£Q
^ S "5 £
«5 a «.2
^c^) PQQ
i(S
474 ACTIVE TRANSPORT AND METABOLISM IN SOME
There has been much other work on these and other topics, but usually
there is some gap in the data so that quantitative comparisons cannot be
made. I wish to conclude by appealing to all workers in this field to try
and obtain results in two forms if possible, the stoichiochemical and
energetic, i.e. the number of ions moved actively in relation to the number
of molecules of oxygen used or substrate metabolized, and also, if this can
be estimated, the amount of energy needed to drive the active transport
relative to the amount of energy made available from metabolism. When
sufficient information of this type has become available it should be possible
to clear the field of useless theories and make a major step forward in the
elucidation of the details of the mechanisms of ion transport.
REFERENCES
AEBI, H. (1952). Helv. physiol. acta, 10, 184.
HARTLEY, W. & DAVIES, R. E. (1952). Biochem. J. 52, xx.
BARTLEY, W. & DAVIES, R. E. (1954). Biochem. J. 57, 37.
CRANE, E. E. & DAVIES, R. E. (1951). Biochem. J. 49, 169.
CRANE, E. E., DAVIES, R. E. & LONGMUIR, N. M. (1948). Biochem. J. 43, 321.
DAVENPORT, H. W. (1952). Fed. Proc. n, 715.
DAVENPORT, H. W. & CHAVRE, V. J. (1951). Amer.J. Physiol. 166, 456.
DAVENPORT, H. W. & CHAVRE, V. J. (1952). Amer. J. Physiol. 171, i.
DAVIES, R. E. & GALSTON, A. W. (1951). Nature, Lond., 168, 700.
DAVIES, R. E. & GALSTON, A. W. (1952). 2nd Int. Congr. Biochem. Abstr. Comrnun.
p. 142.
DEAN, R. B. (1941). Symp. Soc. Exp. Biol. 3, 331.
ELLIOTT, K. A. C. (1946). Proc. Soc. Exp. Biol., N.Y., 63, 234.
FENN, W. O., NOONAN, T. R., MULLINS, L. J. & HAEGE, L. (1941-2). Amer. J.
Physiol. 135, 149.
FRANCIS, W. L. (i933)- Nature, Lond., 131, 805.
GOURLEY, D. R. H. (1952). Arch. Biochem. 40, i.
GRAY, M., HOKIN, L. E. £ REHM, W. S. (1948). Amer.J. Physiol. 155, 440.
GREEN, D. E. (1951). Biol. Rev. 26, 410.
HAHN, L. A., HEVESY, G. C. & REBBE, O. H. (1939). Biochem. J. 33, 1549.
HARMAN, J. W. (1950). Exp. Cell. Res. i, 394.
HODGKIN, A. L. (1951). Biol. Rev. 26, 339.
JOSEPH, M., COHN, W. E. & GREENBERG, D. M. (1939). J. Biol. Chem. 128, 673.
KREBS, H. A., EGGLESTON, L. V. & TERNER, C. (1951). Biochem. J. 48, 530.
KROGH, A. (1946). Proc. Roy. Soc. B, 133, 140.
LUND, E. J. & STAPP, P. (1947). In Bioelectric Fields and Growth. Lund: E. J.
Austin; Texas: University of Texas Press.
LUNDEGARDH, H. (1949). LantbrHogsk. Ann. 16, 372.
MACFARLANE, M. G. & SPENCER, A. G. (1953). Biochem. J. 54, 569.
MUDGE, G. H. (1951 a). Amer.J. Physiol. 165, 113.
MUDGE, G. H. (19516). Amer.J. Physiol. 167, 206.
MUDGE, G. H. (1952). Fed. Proc. n, 109.
NICKERSON, W. J. & MULLINS, L. J. (1948). Nature, Lond., 161, 939.
NOONAN, T. R., FENN, W. O. & HAEGE, L. (1941). Amer.J. Physiol. 132, 472, 612.
PopjAx, G. (1950). Nature, Lond., 166, 184.
PRESSMAN, B. C. & LARDY, H. A. (1952). J. Biol. Chem. 197, 547.
ISOLATED TISSUES AND MITOCHONDRIA 475
RAKER, J. W,, TAYLOR, I. M., WELLER, J. M. & HASTINGS, A. B. (1950). J. Gen.
Physiol. 33, 691.
REHM, W. S. (1950). Gastroenterology, 14, 401.
ROBERTSON, R. N. & WILKINS, M. J. (19480). Nature, Lond., 161, 101.
ROBERTSON, R. N. & WILKINS, M. J. (19486). Aust. J. Sd. Res. B, i, 17.
ROBINSON, J. R. (19500). Proc. Roy. Soc. B, 137, 378.
ROBINSON, J. R, (19506). Nature, Lond., 166, 989.
SACKS, J. (1948). Cold Spr. Harb. Symp. Quant. Biol. 13, 180.
SCHNEIDER, W. C. (1953). J- Histochem. Cytochem. i, 212.
SHEPPARD, C. W. & MARTIN, W. R. (1950). J. Gen. Physiol. 33, 703.
SLATER, E. C. & CLELAND, K. W. (1953). Biochem. J. 54, xxii.
SJOSTRAND, F. S, (1953). Nature, Lond., 171, 30.
SOLOMON, A. K. (1950). iSth Int. Congr. Physiol. Abstr. Commun. p. 460.
SPIEGELMAN, S., KAMEN, M. D. & SUSSMAN, M. (1948). Arch. Biochem. 18, 409.
STAPP, P. (1941). Proc. Soc. Exp. Biol., N.Y., 46, 382.
STERN, J. R,, EGGLESTON, L. V., HEMS, R. & KREBS, H. A. (1949). Biochem. J.
44, 410.
TEORELL, T. (1949). Experientia, 5, 409.
TERNER, C., EGGLESTON, L. V. & KREBS, H. A. (1950). Biochem. J. 47, 139.
USSING, H. H. (1949). Physiol. Rev. 29, 127.
USSING, H. H. (1952). Advanc. Enzymol. 13, 21.
WHITTAM, R. & DAVIES, R. E. (1954). Biochem. J. 56, 445.
ACTIVE TRANSPORT THROUGH
EMBRYONIC MEMBRANES
BY F. W. R. BRAMBELL AND W. A. HEMMINGS
Department of Zoology, University College of North Wales, Bangor.
I. INTRODUCTION
Young birds and mammals acquire passive immunity from their mothers.
At the time of hatching or birth they become exposed to the risks of
infection, but it is not for some weeks, or even months, later that they can
be shown to have developed the power of producing circulating antibodies
to antigens which they have themselves encountered. During this interval
maternal antibodies are present in the circulation and are partly, if not
wholly, responsible for the young animal's powers of resistance to infection.
This passive immunity of maternal origin gradually wanes as the young
animal develops its own active immunity. At first the titres of antibody
of the blood of the young animal may be as high as, or even higher than,
those of the blood of the mother.
It is evident, therefore, that the process of transmission of passive
immunity from mother to young must be efficient. However, it is known
to be accomplished in very different ways in different species. In the fowl
the maternal antibodies are secreted in the yolk of the egg while still in the
ovary and subsequently are gradually absorbed by the yolk-sac of the
developing chick. In the horse, cow, sheep, goat and pig they are present
at high titres in the colostrum, or first milk, and at very much lower titres
in the subsequent milk. The young of these species are devoid of antibodies
at birth, or virtually so, but absorb them through the gut wall with great
rapidity from the first feed. Thereafter, the capacity to absorb antibodies
from the gut contents declines rapidly and virtually disappears by 24 hr.
after birth. In the dog it is probable that some immunity is acquired from
the colostrum and milk after birth, the puppy being able to absorb anti-
bodies from the gut contents until 10-12 days old. In rats and mice it
seems certain that some immunity can be acquired before birth but that
the greater part is acquired after birth, the young being able to absorb
antibodies from the gut contents until they are weaned. In man at least
the major part of the transference of passive immunity takes place in the
uterus but there is weighty evidence of some additional transference by
way of the mammary secretions after birth. In rabbits and guinea-pigs
transmission takes place before birth, and there is no conclusive evidence
ACTIVE TRANSPORT THROUGH EMBRYONIC MEMBRANES 477
of any transfer thereafter by way of the colostrum and the milk. In the
rabbit the maternal antibodies are secreted into the uterine lumen during
pregnancy and are absorbed by the foetal yolk-sac splanchnopleur. The
allanto-chorionic placenta does not appear to play any part in the trans-
mission. Since we have reviewed recently (Brambell, Hemmings & Hen-
derson, 1952) the literature and the experimental evidence for this route of
transmission in rabbits we do not propose to dwell on it further here, but
to accept this conclusion as established. Unfortunately, at present our
knowledge is confined almost entirely to these few species, amongst which
there is much diversity.
It is our purpose in this paper to discuss the problem of active transport
of proteins in the light of our work on the transference of passive immunity
in rabbits.
II. ANTIBODIES AS MARKERS OF
SERUM GLOBULIN
Antibodies occur in the globulin fraction of the serum proteins, principally
in the y-globulin. The molecules of y-globulin which possess antibody
activity are specifically adapted to unite with the antigen which has stimu-
lated their production. It is believed that this activity is due to a surface
configuration of some of the globulin molecules which renders one or more
regions on the surface of each molecule capable of forming a bond with the
prosthetic group of the antigen. Apart from this capacity to react with
the specific antigen, the antibody is chemically and physically indistin-
guishable from other globulin molecules of the same serum fraction, and
presumably performs the same functions.
Hence, if an immune serum is injected into a non-immune animal, the
antibody can be used as a marker of the injected globulin. Appropriate
immunological reactions, in vivo or in vitro, provide tests for the presence
of antibodies at very high dilutions ; those for antitoxins being particularly
useful because they are not only amongst the most sensitive but they are
also quantitative. Since antibody activity depends on surface configuration
of the molecules it provides a safe means of recognition ; any considerable
metabolic alteration of the molecule is liable to destroy its immunological
activity. The technique is limited in that, ipso facto, antibodies can be used
as markers of that particular fraction alone of the serum proteins in which
they occur, usually the y-globulin. Immune sera prepared in other species
have the advantage that while the antibodies contained in them can be
titrated directly, the serum proteins, themselves species-specific antigens,
can be recognized as such by precipitin reactions with their homologous
antisera. This furnishes a complementary means of establishing the in-
tegrity of the foreign protein at least as delicate as that provided by the
478 ACTIVE TRANSPORT THROUGH EMBRYONIC MEMBRANES
titration of the contained antibody. Precipitin ring tests are highly sensitive
but recognize a serum fraction which is not necessarily coextensive with the
antibody. Hence the two tests cannot readily be related quantitatively but
have been used to reinforce each other.
Our investigations on the passage of proteins into the foetus have been
confined to rabbits. We used antibodies as markers, hence the information
obtained was limited to the globulin fraction. The animals employed were
not immune, and hence did not make the antibody themselves.
III. THE ARRANGEMENT OF THE
FOETAL MEMBRANES
The experiments, the results of which concern us here, were performed on
rabbits 24 days pregnant, the immune serum employed being injected into
the uterine lumen. The animals were killed 24 hr. after injection, and
samples of the foetal fluids and of the maternal blood were then collected
and tested for their antibody content. The technique has been described
elsewhere (Brambell, Hemmings, Henderson, Parry & Rowlands, 1949)
and further detail is unnecessary here.
The arrangement of the embryonic membranes of the rabbit is shown in
Fig. i. It can be seen that the yolk-sac splanchnopleur and the chorion are
exposed to the contents of the uterine lumen. The yolk-sac is open after
mid-pregnancy, the lower or bilaminar hemisphere having disappeared,
and the upper or splanchnic hemisphere being inverted, with its endodermal
surface exposed to the lumen but not attached to the uterine wall. This
membrane is highly vascular, containing the network of blood-vessels and
blood-islands, bounded by the sinus terminalis, which constitutes the area
yasculosa. These vessels are connected with the umbilicus by the vitelline
arteries and veins in the yolk-sac stalk. The mesenchyme containing the
vessels is covered on the side next the uterine lumen by a thick columnar
epithelium of endoderm cells resting on a fine basement membrane. The
surface of these endoderm cells exposed to the uterine lumen is provided
with a typical brush border (Morris, 1950). On the side away from the
uterine lumen, the vascular mesenchyme is bounded by a thin squamous
mesothelium, which forms the lining of the exocoel. Between the sinus
terminalis and the margin of the placenta there is a broad zone of chorion,
also exposed to the uterine lumen. This is a thinner membrane which is
almost non-vascular, being traversed only by a few small vessels which
connect the sinus terminalis with the placental circulation. It is covered
on the side next to the uterine lumen by a thin trophoblastic epithelium
and it is lined, on the side next the exocoel, by the same squamous meso-
thelium as covers the inner surface of the yolk-sac splanchnopleur.
ACTIVE TRANSPORT THROUGH EMBRYONIC MEMBRANES 479
It is apparent that the direct approach to the foetal vitelline circulation
from the uterine lumen involves traversing the yolk-sac endoderm first,
then the basement membrane and mesenchyme surrounding the vessels
and finally the endothelial walls of the vessels. Passage of substances from
the uterine lumen into the exocoel would involve traversing either the
endoderm, the basement membrane, the vascular mesenchyme and the
mesothelium of the yolk-sac splanchnopleur, or alternatively the tropho-
blast, mesenchyme and mesothelium of the chorion. The amnion is
Fig. i . Diagrammatic transverse section of the uterus and foetal membranes of a rabbit
in late pregnancy. The foetus is omitted. Foetal ectoderm shown as continuous lines,
foetal endoderm as cross-hatched, foetal mesoderm as broken lines. A, amnion; AL,
allantois; AS, allantoic stalk; C, chonon; E, exoceol; L, uterine lumen; O, remnant of
bilaminar omphalopleur; P, allanto-chorionic placenta; S, yolk-sac splanchnopleur;
Uy umbilicus ; YS, yolk-sac stalk.
separated from the uterine lumen by these membranes and by the fluid-
filled cavity of the exocoel. The amnion is a very thin membrane, covered
on the outside by the mesothelium lining the exocoel and lined on the inside
by a thin squamous epithelium of ectoderm cells. By the 25th day of
pregnancy the amount of exocoelomic fluid remaining, though variable,
is usually small, so that often it was impossible to obtain a sufficient
sample for testing. However, the considerable number of samples that
were obtained did not vary greatly from the amniotic fluid in antibody
480 ACTIVE TRANSPORT THROUGH EMBRYONIC MEMBRANES
content. The latter could always be obtained in sufficient quantities at this
stage, and it became the standard practice to collect the amniotic fluid and
to neglect the exocoelomic fluid.
To study the possible entry of antibodies from the uterine lumen into
the foetal vitelline circulation by the indirect route through the chorion
and the exocoelomic surface of the yolk-sac splanchnopleur, thereby
avoiding passage through the endoderm, immune serum was injected directly
into the exocoel in a further series of experiments (Brambell, Hemmings,
Hemmings, Henderson & Rowlands, 1951). Thus, in these, the outer
surface of the amnion and the inner surface of the yolk-sac splanchnopleur
were directly exposed to the immune serum.
It is worthy of emphasis that the arrangement of the foetal membranes in
the rabbit is such as to render them peculiarly accessible to experiments
involving exposure to solutions, since these can be injected so readily into
the uterine lumen or even into the cavities enclosed by the membranes.
It should be observed that passage through vascularized membranes, such
as the yolk-sac splanchnopleur, presents a special problem. Passage could
occur directly through such membranes, from one side to the other, as
with non-vascular membranes, or passage could occur from one or other,
or from both, sides into the blood vessels.
These are important considerations in relation to the results which are
to be considered.
IV. EXPERIMENTAL RESULTS
Antibodies enter rapidly into the foetal circulation from immune rabbit
serum injected into the uterine lumen. This also is the route by which
antibodies enter from the maternal circulation. Provided that the quantity
of immune serum injected is 0^5 ml. or over per embryo, the titre of
foetal blood is related remarkably closely to the titre of the immune serum
employed. This relationship holds whether the antibodies are bacterio-
agglutinins, haemolysins, or antitoxins. The results of a series of experi-
ments with agglutinins and haemolysins are shown in Table i as a correla-
tion table of titre of foetal serum to titre of immune serum injected. It is
evident that there is close correlation.
The same data are presented in Table 2 as distributions according to the
ratio of the titre of the foetal serum to the titre of the immune serum
injected; the data for agglutinins and haemolysins being shown separately.
It is evident that, from the data available, there is no significant difference
in behaviour between agglutinins and haemolysins, that in both the modal
ratio is 1/16 and that the individual variation of foetuses is confined to two
dilutions either side of the mode.
ACTIVE TRANSPORT THROUGH EMBRYONIC MEMBRANES 481
The data for diphtheria antitoxic sera are given as the ratio of the con-
centration of the foetal serum to the concentration of the immune serum
injected in Table 6, column 2. Whereas the data presented in Tables i and 2
are based on the serum titres of individual foetuses, those in Table 6 are for
the pooled sera of all the foetuses in a litter, individual variation being
largely masked by pooling. It will be observed that the ratio conforms
closely to the modal ratios for agglutinins and haemolysins.
Table i . Correlation of litre of foetal serum to litre of immune serum
injected
(Extracted from Brambell et al. 1949, 1950.)
,—,. - Titre of foetal serum
Titre of
serum injected
1,10 1/20 1/40
1/80 ! 1/160
1/20,480
—
! i
1/10,240
— i —
—
1/5,120
—
2
1/2,560 , —
2
4
I I
1/1,280
6
6
I
1/640 i
2
4
5
1/320
5
ii
i
1/320 1/640
Table 2. Titres of foetal sera expressed as ratios of the titres of the
immune sera employed
(Extracted from Brambell et al. 1949, 1950.)
Titre of foetal serum
1/64
i/32
1/16
1/8
No. of foetuses observed
jlutinins
Haemolysins i
i
9
13
5
45
6 i
5 !
J6 |
I
|
Total
The obvious route by which antibodies in the uterine lumen enter the
foetal circulation in the area vasculosa is through the endoderm of the yolk-
sac splanchnopleur. There is an alternative route through the chorion into
the exocoel and thence through the mesothelium of the exocoelomic surface
of the yolk-sac splanchnopleur. This would avoid traversing the endoderm.
To distinguish between these, immune serum was injected directly into the
exocoels of alternate embryos. Almost invariably some leaked back through
the puncture into the uterine lumen. Hence the outer or endodermal
surface of the yolk-sac splanchnopleur of all the embryos, both experi-
mental and control, was exposed to the immune serum. In the experi-
i. n ^ vin 31
482 ACTIVE TRANSPORT THROUGH EMBRYONIC MEMBRANES
mental embryos the inner or exocoelomic surface also was exposed. The
immune serum injected was diluted by the exocoelomic fluid. The results
are given as the ratio of the titre of the foetal serum to the titre of the
immune serum injected in Table 3. This ratio could not be calculated for
those foetal sera which were negative at the lowest dilution of i/io, but
these were too few to invalidate the results. It can be seen from the close
correspondence between the distributions of experimental and control
foetuses that entry of antibodies into the foetal circulation must be mainly,
if not entirely, through the outer surface of the yolk-sac and hence through
the endoderm.
Table 3. Titres of foetal sera expressed as ratios of the litres of immune
sera injected into the exocoels
(Extracted from Brambell et al. 1951.)
Titre of foetal serum
Experimental
foetuses
Control
foetuses
Titre of serum injected
Negative
3
4
1/128
i
2
1/64
10
10
1/32
i?
IO
1/16
9
3
1/8
3
3
i/4
i
Total
43
33
Comparison of the entry of antibodies prepared in cattle or horses with
the entry of those prepared in rabbits shows that absorption into the foetal
circulation is a selective process, the bovine or equine antibodies being
transported in much lower degree than the rabbit antibodies. Antibrucella
serum prepared in rabbits was injected into the lumen of one uterine horn
and antibrucella serum prepared in cattle into the lumen of the other
uterine horn in each of five rabbits pregnant 24 days. The titres of the
foetal sera, expressed as ratios of the titres of the sera injected are shown
in Table 4. In most cases the bovine antibodies could not be detected in
the foetal serum even at the lowest dilutions, although the immune bovine
sera employed were of higher titres than the immune rabbit sera. In those
cases where bovine antibodies were detected in the foetal sera they were at
much lower relative titres than the rabbit antibodies.
Similar experiments with antidiphtheria serum prepared in horses, the
test for which is more sensitive, showed that the antitoxin invariably
entered the foetal circulation but at concentrations of the same order as
those of bovine agglutinins. It appeared possible that these results might
be due to properties of the serum other than those of the antibody molecules
ACTIVE TRANSPORT THROUGH EMBRYONIC MEMBRANES 483
themselves: for example, the foreign sera might be toxic or might block
pores in the membrane, thus rendering it impermeable. On these grounds
mixtures of immune rabbit serum with either immune bovine serum or
immune equine serum were injected into the uterine lumen. The anti-
bodies prepared in rabbits entered the foetal circulation as readily from such
mixed sera as from pure rabbit serum, as is shown by the data in Table 5.
The antibodies prepared in cattle or horses did not enter any more readily
from such mixed sera than from pure sera.
Table 4. Comparison of entry of antibrucella agglutinins prepared in
rabbits and in cattle
(Extracted from Brambell et al. 1950.)
Titre of foetal serum
Foetuses exposed
to rabbit anti-
bodies
Foetuses exposed
to bovine anti-
bodies
Titre of serum injected
1/256
—
3
1/128
—
3
1/64
—
1/32
i
—
1/16
4
—
1/8
i
—
i/4
5
—
Total: Positive
ii
6
Negative
—
16
Table 5. Entry of agglutinins or haemolysins prepared in rabbits from
mixtures of rabbit and equine or bovine sera compared with that from
pure rabbit sera
(Extracted from Brambell et al. 1950.)
Titre of
foetal serum
Pure serum
Mixed serum
Titre of serum injected
1/64
i
6
1/32
8
6
1/16
13
14
1/8
5
14
i/4
5
i
It was possible, consequently, to compare entry of antitoxins from
mixtures of immune sera prepared in rabbits, cattle and horses, so long as
each antiserum was prepared against a different antigen. It was necessary
to pool the sera from all the foetuses in each litter to obtain quantities
sufficient for all three tests. The results are shown in Table 6. It is evident
that upwards of fifty times more antitoxin prepared in rabbits enters the
foetal circulation than antitoxins prepared either in cattle or horses.
31-2
484 ACTIVE TRANSPORT THROUGH EMBRYONIC MEMBRANES
It may be concluded that the transport of antibodies from the uterine
lumen into the foetal circulation is a selective process depending on some
character of the antibody molecules peculiar to the species of animal in
which they have been produced. It seems improbable that the molecular
weight of the antibody is the determining factor. As is well known, these
tend to fall into one of two size groups ; those with molecular weight about
180,000, and those with molecular weight about 900,000. The rabbit
agglutinins and the equine antitoxins belong to the group of smaller
molecules, whereas the bovine agglutinins and the rabbit haemolysins
belong to the group of larger molecules, a distribution that cuts across that
of entry into the foetal circulation.
Table 6. Entry of antitoxins prepared in rabbits, cattle and horses into the
foetal circulation expressed as ratios of the concentrations of the mixture
of immune sera employed
(Extracted from Brambell et al. 1952.)
No. of
experiment
„ . concentration of foetal serum
Ratio: . - — j
concentration of serum injected
Rabbit
antitoxin
Bovine
antitoxin
Equine
antitoxin
569
613
578
615
c. 0-047
0-059
0-077
0-040
0-0002
O-OOI2
< 0'0002
0-0008
0-0013
0-0009 to 0-0005
< 0-0008
Results, as yet unpublished (Brambell, Hemmings and Oakley), obtained
with antitoxins prepared in other species, show that the entry of these is
intermediate between those for rabbit antitoxins on the one hand and
bovine or equine antitoxins on the other hand.
There is evidence of another kind that the foetal yolk-sac splanchnopleur
in rabbits is selective as between the different maternal serum protein
fractions. Electrophoretic analyses of maternal and foetal sera in rabbits
(Brambell, Hemmings, Henderson & Kekwick, 1953) have shown that the
proportions of the electrophoretic fractions differ widely, the ratios of the
foetal/maternal concentrations being for albumin c. 1/3, for a-globulin 2/3,
for /?-globulin 3/2, and for y-globulin c. 2/5. This suggests a differential
selection of these protein fractions by the foetal membrane, which we hope
to investigate by using radioactive isotopes as markers.
Estimations of the antitoxin content of the tissues of the yolk-sac splanch-
nopleur after exposure to a mixture of antitoxic sera should provide a
useful clue to the mechanism of selection. Such determinations of the
antitoxic content of the foetal membranes were made on material from the
ACTIVE TRANSPORT THROUGH EMBRYONIC MEMBRANES 485
experiments referred to in Table 6 and from two other animals, exposed
for shorter periods of 3 and 6 hr. respectively. The results are given in
Table 7.
Table 7. Relative concentrations of antitoxins in other foetal fluids and
in the tissues of the membrane
(Extracted from Brambell et a/. 1952.)
No
5<>9
613
578
6l5
' 902 1
903
Duration of
'
exposure (hr.) ...
24
22
24
23
i 6 i
3
Rabbit/bovine :
1
Washings
0-97
I '5 I
2-83
0'9Q
i-5o 1
1-23
Splanchnopleur
<2'26
2-72
2' 1C)
I'59
1 1-09 !
1-27
Amniotic fluid
2-24
•30
I-78
I-48
; 1-56 !
Stomach contents
•30
I-87
3-66
—
Rabbit/equine :
'
Washings
—
•07
2-04
0-86
1*34
i'i5
Splanchnopleur
—
•86
1-62
1-29
I -00
I'OO
Amniotic fluid
—
•25
I'34
I'I2
1-12
—
Stomach contents
—
•14
I-58
I-84
—
—
It might be supposed that if the bovine and equine antitoxins were
prevented from entering the membrane, then the relative concentrations
in the tissues would be similar to those in the foetal circulation, the rabbit
antitoxin preponderating greatly. If all were admitted equally freely by
the endoderm, and exclusion of the heterologous antitoxins was effected
by the vascular endothelium, the only other continuous cellular layer
separating the circulation from the uterine lumen, then a large excess of
bovine and equine antitoxins might be built up in the tissues around the
vessels. Neither of these expectations was realized. The ratios observed
varied from unity to 2*72. Such determinations are, however, liable to error
from two sources, first from antitoxin adhering to the outer surface of the
membrane (although this was well washed in saline), and secondly, from
residual blood remaining in the vessels. Although they were drained, they
were not transfused, and it can be calculated that 4% of blood in the tissues
would be sufficient to account for the small but significant excess of rabbit
antitoxin observed in the tissues. These results are not conclusive, and it
would be unwise to place too much reliance upon them.
Antibodies enter the amniotic fluid from the uterine lumen also. The
absolute concentrations attained in the amniotic fluid may not be of much
significance, since the volume of fluid in the amnion is very variable, but
the relative concentrations of antibodies prepared in different species are
interesting. The ratios for antitoxins, in the four experiments already
referred to in Table 6, are given in Table 7. It will be apparent that
486 ACTIVE TRANSPORT THROUGH EMBRYONIC MEMBRANES
whereas rabbit antitoxin enters the amniotic fluid much less freely than the
foetal circulation, the reverse is true of bovine and equine antitoxins.
Consequently, the relative entry of rabbit/bovine and rabbit/equine anti-
toxins into the amniotic fluid only slightly, though significantly, exceeds
unity. This suggests that entry takes place directly through the membranes
from the uterine lumen and not indirectly by way of the foetal circulation.
The nearly uniform entry of rabbit, bovine and equine antitoxins into the
amniotic fluid does not justify the assumption that it is due to an active
process of selective absorption and secretion on the part of the cells, such as
is required to account for the entry of rabbit antitoxins into the circulation.
It may be of an entirely different character not involving cellular activity.
Since the concentrations attained within the amnion after 24 hr. exposure
are of the order of less than i/ioo of those without, it is evident that an
effective barrier to free passage does exist. The gradual passage through
this barrier, in slightly different amounts for each kind of antitoxin, could
be accounted for by purely physical forces. It may be that antibodies * seep '
into the amniotic cavity between the cells of the multicellular membrane,
perhaps even through local faults that must occur in such membranes.
Since the entry of bovine and equine antibodies into the foetal circulation
is of the same order as into the amniotic fluid, it is possible that it may be
due to * seepage' between the cells.
Antibodies are found also in the stomach contents of foetuses which have
been exposed to immune serum. These appear to be derived from the
amniotic fluid, but often are more concentrated in the stomach than in the
amnion. The ratios are given in Table 7.
Finally, attempts were made to recover the residues of the serum injected
from the uterine lumina at the conclusions of the experiments. Too little
fluid remained to collect, so saline washings were taken. The results are
included in Table 7. The value for no. 569 was obtained from saline trans-
fusion of the intact uterus. The other values refer to saline in which the
excised foetal yolk-sac splanchnopleurs were washed. It is probable that
minute quantities of foetal blood remaining in these membranes may have
contaminated the latter group and may account for the tendency for the
ratios to be above unity.
V. DISCUSSION
The mammalian foetus provides an excellent approach to the problem of
the relation of the plasma protein reserve to the tissue protein. In some
species maternal plasma protein reaches the foetus in the form of antibody.
Possibly maternal plasma protein, of which the antibody is a conveniently
recognizable fraction, may play normally an important part in foetal meta-
ACTIVE TRANSPORT THROUGH EMBRYONIC MEMBRANES 487
holism. Certainly, in some cases, the antibody may cause serious damage
to the foetus.
Technically it is not easy to determine the quantity of protein transported
across unit area of the maternal/foetal barrier in unit time, and, so far, we
have not attempted to do so. What has been measured is the concentration
of antibody built up in the foetal serum, a quantity not necessarily related
to specific rate of transport in any simple fashion. It would be premature
to attempt a coherent hypothesis of molecular exchange across this barrier
on the evidence available, but some discussion of the problems involved
appears justified. The reasons for regarding the appearance of the specific
antibody in the foetal circulation as evidence of the transport of the
molecules through the foetal membranes substantially unaltered, have been
discussed already.
The embryonic membranes are multicellular and complex. It would be
quite misleading to treat them as though they were simple membranes of
uniform structure. First, being multicellular, transport through them could
occur either between or through the component cells. Individual cells must
die, and may leave, temporarily, actual gaps in the membrane. Openings
may form due to changes of shape and movements of the cells. The inter-
cellular substance between the cells may be permeable to protein. It is
difficult to conceive a mechanism adequate to explain the large entry of
protein through such a membrane, which is capable at the same time of
retaining the foetal fluids, unless it be by active transport by the cells. The
cells need not all be in the same physiological state ; the sum of their cell
membranes cannot have the uniformity of the cell membrane of a single
cell at any given moment. Active transport of large molecules by a cell
could take place by ingestion in droplets of fluid and their transport in
vacuoles or by their attachment to and passage through the cell membrane
into the organized substance of the protoplasm. It is worth noting that
whereas the former process could be selective only in an all-or-none fashion
of accepting certain fluids and rejecting others, the latter process could be
selective also in the sense of accepting certain molecular specie* while
rejecting others presented to it simultaneously in a common substu a.
It was found that bovine and equine antibodies occur in the foetal sera
at nearly the same concentrations, relative to each other, as in the mixed
sera injected. They are both at much lower concentration relative to rabbit
antibodies. This led to the suggestion that the difference might be due to
qualitatively different methods of entry into the foetal circulation (Bram-
bell et al. 1952). It was suggested that the entry of the foreign antibodies
might be due to a process of slow 'seepage' between the cells of the multi-
cellular membrane. This would not necessarily involve an active process
ACTIVE TRANSPORT THROUGH EMBRYONIC MEMBRANES
of absorption and secretion on the part of the cells such as the rapid entry
of rabbit antibodies appeared to imply. However, more recent work, still
in progress (Brambell, Hemmings & Oakley, unpublished) shows that
antibodies prepared in certain other species appear in the foetal circulation
in concentrations intermediate between the low bovine and equine values
and the high rabbit ones. In this respect it appears now that antibodies
prepared from a range of species form a graded series rather than two clearly
defined groups.
The selective effect could be produced in any one of three ways, or in
any combination of them. First, by differential destruction of the anti-
bodies before reaching the barrier between the uterine lumen and the
foetal circulation. Secondly, by differential destruction after passing the
barrier. Thirdly, by selection at the barrier. The first two hypotheses
involve differential destruction of antibodies, such as occurs when foreign
proteins are administered parenterally to the adult organism. In the non-
immune animal the removal is gradual while the antigenic properties of
the foreign antibody globulin result in its rapid removal through serological
reactions in the immune adult. No satisfactory evidence has been obtained
so far of such destruction of the foreign antibodies, either in the uterine
lumen or in the foetus during the course of the experiments. Some con-
trary evidence has been forthcoming. Material recovered from the uterine
lumen, either by perfusion or by washing the foetal membranes at the con-
clusion of the experiments, was found to have ratios of rabbit to bovine or
equine antibodies not greatly exceeding those in the mixed serum injected.
Moreover, the ratios in the tissues of the splanchnopleur, in the amniotic
fluid and in the stomach contents were similar (Table 7). If differential
destruction in the uterine lumen, before reaching the barrier, accounted
for the ratio in the foetal circulation, the ratios in the other foetal fluids
and tissues should be similar to that in the foetal serum. This is strong
evidence that destruction in the uterine lumen alone cannot account for the
results.
Destruction of the foreign protein after passing the barrier would need
to be exceedingly rapid to account for the ratios observed. Serological
reactions scarcely can be invoked since the mammalian foetus does not
appear to be capable of producing such antibodies itself, so far as is known
at present. No clear evidence of destruction has been obtained. Bovine and
equine antibodies are found in the foetal circulation for at least 96 hr. and
the concentrations after 48 hr. are not much lower than after 24 hr. from
the time of injection into the uterine lumen. Although this could be
accounted for by continued uptake rather than persistence of the foreign
antibodies, it does not provide any positive evidence of rapid destruction.
ACTIVE TRANSPORT THROUGH EMBRYONIC MEMBRANES 489
Although the hypothesis of differential destruction after passing the barrier
cannot be excluded, that of selection at the barrier appears preferable at
present.
Wherever selection occurs, the means by which it is brought about is a
problem of no less interest. A selective transport favouring certain mole-
cular species at the expense of others in the same substrate would appear
to be a very strong indication of a process involving at some point an inti-
mate relation with the organized structure of the protoplasm. It is not
possible to separate the corresponding fractions from mixed sera by any
known physical means, and immunological methods must be employed to
effect such separation in vitro. Hence it may be suspected that the
mechanism of separation in vivo is immunological in character, using the
term in a broad sense, and that it involves something in the nature of the
union of antibody with antigen or of enzyme with substrate. It must be
remembered that the rabbits employed had not been immunized previously,
and that the 24 hr. duration of the exposure to the foreign proteins during
the experiments was too short to permit of the development in the mother
of any known kind of immunity.
REFERENCES
BRAMBELL, F. W. R., HEMMINGS, W. A., HENDERSON, M., PARRY, H. J. &
ROWLANDS, W. T. (1949). Proc. Roy. Soc. B, 136, 131-44.
BRAMBELL, F. W. R., HEMMINGS, W. A., HENDERSON, M. & ROWLANDS, W. T.
(1950). Proc. Roy. Soc. B, 137, 239-52.
BRAMBELL, F. W. R., HEMMINGS, G. P., HEMMINGS, W. A., HENDERSON, M. &
ROWLANDS, W. T. (1951). Proc. Roy. Soc. B, 138, 188-95.
BRAMBELL, F. W. R., HEMMINGS, W. A., HENDERSON, M. & KEKWICK, R. A.
(!953)- Proc. Roy. Soc. B, 141, 300-14.
BRAMBELL, F. W. R., HEMMINGS, W. A., HENDERSON, M. & OAKLEY, C. L. (1952).
Proc. Roy. Soc. B. 139, 567-75.
BRAMBELL, F. W. R., HEMMINGS, W. A. & HENDERSON, M. (1952). Antibodies and
Embryos. London: University of London, The Athlone Press.
MORRIS, B. (1950). Quart. J. Micr. Sci. 91, 237-49.
TRANSPORT OF LIPID THROUGH CELL
MEMBRANES
BY A. C. FRAZER
Department of Pharmacology, University of Birmingham
I. INTRODUCTION
The metabolism of lipids in the animal body involves their absorption from
the intestinal lumen, distribution in the body, uptake by and release from
adipose tissue and utilization for energy purposes in the liver and other
cells. The passage of lipids into the intestinal cell, the adipose tissue cell
and the liver cell and the transport of lipid material in the blood between
the intestine, the fat stores and the liver are facts that are generally accepted
by physiologists. The object of this paper is to consider in detail the evidence
for the transport of lipids into these three groups of cells in the living body,
to examine the type of lipid involved in this transfer and the factors that
appear to influence it in each case.
Types of lipid molecules
The term ' lipid' is applied to a group of substances with certain common
lipoid characteristics that are frequently associated together in biological
systems. The most important lipids are :
(i) Fatty acids
(a) Saturated series. This comprises acetic acid up to fatty acids con-
taining 30 carbons or more. Natural fatty acids usually contain an even
number of carbons; those with 10 carbons or less are volatile in steam and
are liquid at mammalian body temperature. The commonest natural
saturated fatty acids are lauric, myristic, palmitic and stearic acids.
(b) Unsaturated series. This includes the most common of all natural
fatty acids, oleic acid ; this is liquid at mammalian body temperature. There
are also several polyethenoid acids. Those containing the grouping
— CH : CH . CH2 . CH : CH — are dietary essentials since this configuration
cannot be synthesized in the body.
(ii) Glycerides
Fatty acids occur as mono-, di- and tri-esters with glycerol, the latter
being the commonest form in the body.
TRANSPORT OF LIPID THROUGH CELL MEMBRANES 491
(iii) Phospholipids and cerebrosides
Lecithins and cephalins are glyceride esters containing fatty acids,
phosphoric acid and a base. Plasmalogens are similar but contain fatty
aldehydes instead of fatty acids. Carbohydrate containing phospholipids
and a non-glyceride phospholipid, sphingomyelin, have also been isolated.
Cerebrosides contain fatty acid, carbohydrate and base, and are closely
related to the sphingomyelins. The most important phospholipid for con-
sideration in connexion with lipid transport would seem to be lecithin.
(iv) Sterols
Many sterols occur in the body and have various pharmacological actions.
So far as lipid transport is concerned, cholesterol appears to be the most
important.
II. THE INTESTINAL CELL
(i) Structure of the cell
The absorbing unit in the small intestine is a columnar cell with a thick
striated outer border next to the intestinal lumen. This outer border has
been shown by Baker (1942) to be penetrated by fine canals about 0*3/4 in
diameter. Electron microscope studies (Granger & Baker, 1950) have
revealed a fine fibrilliform structure of the outer border — the fibrils running
in the same axis as these canals. The canals can be shown to be filled with
lipid-staining material only during the absorption of fat (Baker, 1951).
Between the outer border of the cell and the nucleus is the Golgi organ ;
this part of the cell appears to be rich in phospholipid. The cell contains
a number of enzymes including alkaline phosphatase.
(ii) Evidence for transport of lipids through the intestinal cell
Numerous observers have noted that glycerides disappear from the
intestinal lumen and simultaneously lipid-staimng material can be demon-
strated inside the intestinal cells while the chyle becomes laden with fat.
This situation persists just as long as there is lipid being absorbed from the
intestinal lumen. Although the intestinal cell may synthesize glycerides,
visible lipid cannot be demonstrated in the intestinal cells or the chyle
during carbohydrate or protein absorption. Labelled glyceride molecules
have been traced through the cell into the chyle from the intestinal lumen.
Lecithin can be readily traced into the intestinal cells, and labelled chole-
sterol was found to pass from the intestinal lumen with fat into the chyle
(Biggs, Friedman & Byers, 1951). It has been claimed that plant sterols are
not absorbed (Schoenheimer, 1931), but the evidence does not preclude
their destruction in the intestinal cell.
492 TRANSPORT OF LIPID THROUGH CELL MEMBRANES
There seems little doubt that lipids pass through the intestinal cell
during absorption. The cell is an active metabolic unit capable of modifying,
or even synthesizing, many types of lipid molecule, so that comparison of
the lipid material found in the intestinal lumen with that recovered from
the chyle is not a reliable indication of selective permeability of the
intestinal membranes.
(iii) Forms of lipid transported
For nearly half a century it has been thought that lipids had to be
reduced to some water-soluble form before absorption could occur (Verzar
& McDougall, 1936; Bloor, 1943). In the case of glycerides this was said
to be achieved by complete hydrolysis and solubilization of the liberated
fatty acids by the hydrotropic action of bile salts. On the basis of considerable
experimental evidence to the contrary, this concept of complete hydrolysis
has been challenged (Frazer, 1938, 1946, 1952 a). That hydrolysis is partial
and not complete has recently been conclusively proved by studies using
labelled materials (Favarger & Collet, 1949; Favarger, Collet & Cherbuliez,
1951; Karnovsky & Gidez, 1951; Reiser, Bryson, Carr & Kuiken, 1952;
Borgstrom, 1952).
Partial hydrolysis results in the presentation of a water-insoluble as well
as a water-soluble lipid fraction to the intestinal cell for transport. This
water-insoluble portion is present in the intestinal lumen as a finely dis-
persed oil-in-water emulsion of glycerides containing other lipid-soluble
molecules. It can be shown that fine emulsification is an essential step in
the absorption of the glyceride fraction (Frazer, Schulman & Stewart, 1944;
Daniel, Frazer, French & Sammons, 1951; Frazer, 1952^). Phospholipid
is not hydrolysed in the intestinal lumen in rats or human subjects (Frazer,
Sagrott & Sammons, 1949). Cholesterol appears to be absorbed in the
glyceride fraction. The water-soluble lipid fraction presumably enters the
cell in ionic or molecular form.
(iv) Factors concerned in lipid transport through the intestinal cell
It would appear that the first essential factor determining whether the
lipid molecule will pass in the water-soluble or water-insoluble fraction is
its partition coefficient between oil and water. Thus, the lipid molecules
are 'partitioned' into two groups in the intestinal lumen. Normally the
water-soluble component consists essentially of the shorter chain, and
perhaps some of the more unsaturated, fatty acids. Long chain, and more
saturated, fatty acids tend to remain in the oil phase. This partition of
absorbed fatty material originally predicted from relatively indirect
evidence (Frazer, 1948 a) has now been conclusively proved using 14C-
TRANSPORT OF LIPID THROUGH CELL MEMBRANES 493
labelled fatty acids (Bloom, Chaikoff, Reinhardt, Entenman & Dauben,
1950; Chaikoff, Bloom, Stevens, Reinhardt & Dauben, 1951; Bloom,
Chaikoff & Reinhardt, 1951 ; Kiyasu, Bloom & Chaikoff, 1953). The water-
soluble fatty acids are mainly absorbed by diffusion, but there is not
sufficient evidence to show whether any selective transfer of these molecules
occurs. The absorption of tributyrin and fatty acids is not affected by
double adrenalectomy in rats (Bavetta & Deuel, 1942; Frazer, 19486).
Particulate absorption of the water-insoluble glycerides occurs mainly
in the upper intestine. The particles must be less than o-$/i in diameter
and negatively charged for absorption to occur (Frazer et al. 1944).
Particulate absorption is entirely prevented by faulty intraluminar
emulsification, but in pancreatic enzyme deficiency it can be re-established
by the intraduodenal administration of finely emulsified fat (Frazer, 1952^).
Glyceride absorption is depressed by double adrenalectomy in rats (Verzar
& Laszt, 1934, 1935; Barnes Miller & Burr, 1941), but facilitated by the
simultaneous administration of choline, glycerophosphate or lecithin
(Frazer, 1951). Particulate absorption is depressed in human subjects in
the sprue syndrome possibly due to excessive mucus secretion and decreased
intestinal motility. Pumping of the villi and water flow may be important
factors (Frazer, 1952^).
It would therefore appear that the most important form of lipid trans-
ported through the intestinal cell is glyceride in particulate form. These
glyceride particles may contain other lipid-soluble molecules such as
cholesterol. Those lipids that can be removed from the oil into the water
phase may pass into the cell by diffusion.
III. TRANSPORT OF LIPID IN BLOOD AND LYMPH
(i) Structure of plasma
The resting blood plasma contains 400-500 mg. of lipid per 100 ml. of
plasma, which is mainly in the form of phospholipid and cholesterol. The
phospholipid is largely lecithin containing palmitic, oleic and linoleic acids,
while 60% of the cholesterol is esterified with more unsaturated fatty acids.
The level of cholesterol is affected by age (Keys, Mickelsen, Miller, Hayes
& Todd, 1950), by the dietary level of fat (Keys, 1952) and by endocrine
factors.
These lipids are associated with protein to form macro-molecules. Two
plasma lipoproteins have been characterized — an a-lipoprotein that is about
300 A. long and 50 A. wide and has a molecular weight of about 200,000;
and a /?-lipoprotein that is spherical with a diameter of about 185 A.
and molecular weight about 1,300,000. Some 75% of the lipid in fasting
494 TRANSPORT OF LIPID THROUGH CELL MEMBRANES
plasma can be accounted for as /?-lipoprotein (Gurd, Oncley, Edsall &
Cohn, 1949).
After fat feeding and under certain other conditions there is a marked
increase in blood lipids. This normally takes the form of an increase of the
glyceride fraction which is commonly accompanied by a concomitant
smaller increase of phospholipid and cholesterol. The glyceride is in par-
ticulate form and causes a marked turbidity of the plasma. This emulsion
of glycerides remains stable in the plasma for long periods. Stability
appears to depend upon lecithin since destruction of the lecithin by
D-lecithinase causes clumping and creaming of the chylomicrons (Frazer,
Elkes, Sammons, Govan & Cooke, 1945). The phospholipid appears to
link the glyceride particles to globulin so that they display certain reactions
characteristic of this protein fraction (Elkes, Frazer & Stewart, 1939).
(ii) Evidence for lipid transport in the blood and types of lipid involved
Lipid transport in the blood is proved if lipid can be shown to be added
to the blood at one point and removed at another. This is the case with
glycerides which are injected into the blood stream from the thoracic duct
and subsequently removed into adipose tissue or the liver and other cells.
Labelled glyceride has been traced from the intestinal lumen to these
different destinations. Turn-over of plasma phospholipids has been
demonstrated, and it was suggested that glycerides might be converted
into lecithin in the liver in preparation for transport to, and metabolism by,
the extrahepatic tissues. Recent work shows that this is not the case and
that the liver is the main site of both formation and removal of plasma
lecithin (Entenman, Chaikoff & Zilversmit, 1946). Furthermore, extra-
hepatic tissues can readily use glycerides or fatty acids (Geyer, Cunningham
& Pendergast, 1950; Goldman, Chaikoff, Reinhardt, Entenman & Dauben,
1950; Geyer & Cunningham, 1950), and can synthesize phospholipids in
situ if they are required. Similarly, there is little reason to believe that
cholesterol plays a major part in the transport of other lipids in the blood.
(iii) Factors affecting lipid transport
Little is known about factors controlling lipoprotein structure and
stability. There is some evidence that abnormal lipoproteins can occur
under certain circumstances. An abnormal and unstable lipoprotein
fraction has been demonstrated in ageing subjects and may be responsible
for deposition of lipid in the blood vessel walls (Gofman, Jones, Lindgren,
Lyon, Elliott & Strisower, 1950). Whether lipoprotein structure varies
significantly under normal physiological conditions, other than with ageing,
remains to be seen.
TRANSPORT OF LIPID THROUGH CELL MEMBRANES 495
With regard to paniculate glyceride, the importance of the integrity of
the stabilizing lecithin has already been noted. It has also been shown that
heparin can abolish the abnormal turbidity of alimentary hyperlipaemia.
This was said to be due to finer dispersion of the glyceride (Hahn, 1943;
Weld, 1944, 1946), but recent work indicates that heparin may be more
concerned with the removal of fat from the blood stream than dispersion
of the lipid in the plasma (Brown, 1952). While the question must be left
open at present, there is no doubt that the effect of heparins and anti-
heparins must be taken into account in any consideration of lipid transport
in the blood.
(iv) Transfer to and from the blood and lymph
Negatively charged particles appear to pass more readily from the tissue
spaces into lymphatics than into blood capillaries. The reasons for this
apparent selectivity are not clear. In the small intestine the negatively
charged particles of glyceride enter the lacteal rather than the capillaries
and so pass almost exclusively into the chyle. The particulate fat is trans-
ferred from the lymph into the blood stream by 'intravenous injection* via
the thoracic duct. Water-soluble lipids, such as short-chain fatty acids, may
pass into the capillaries and be conveyed in the portal blood.
The particulate glycerides pass out of the blood to cells where they may
be stored or metabolized. How this outward passage of particulate fat
occurs is still a mystery. Certain factors appear to control this removal —
the possible effect of heparins and antiheparins has already been mentioned.
In occasional subjects the rate of removal is slowed so that progressive
accumulation of particulate glyceride in the blood follows normal fat intake
and the blood fat may exceed 10%. The level of blood glycerides rapidly
decreases if fat is excluded from the diet (Holt, Aylward & Timbres, 1939;
Lawrence, 1946; Stanley & Thannhauser, 1949).
It may, therefore, be concluded that the only certain evidence of lipid
transport in the blood at the present time is that concerned with the con-
veyance of glycerides in particulate form from the intestine to the adipose
tissue and from these fat stores to the liver and elsewhere. The other lipid
components of plasma appear to be essentially structural and turn-over of
lecithin and cholesterol mainly reflects metabolic activity in the liver.
(i) Structure IV" ADIPOSE TISSUE CELLS
Adipose tissue cells are macrophages that are particularly concerned with
the assimilation of fat from, and subsequent release into, the blood and
tissue spaces. The detailed mechanism of this uptake and mobilization of
496 TRANSPORT OF LIPID THROUGH CELL MEMBRANES
fat is unknown. Adipose tissue cells can also synthesize fats and glycogen
(Tuerkischer & Wertheimer, 1942). Special collections of adipose tissue —
so-called ' brown fat' — occur in hibernating animals and seem to have more
extensive functions (Wendt, 1937; Eger, 1938). When fully developed,
adipose tissue cells appear as large nucleated cells in which the whole
cytoplasm appears to be replaced by triglyceride fat.
(ii) Evidence of lipid transport into and out of adipose tissue cells
As already mentioned, adipose tissue cells can synthesize glycerides from
non-lipid material. However, labelled glycerides have been traced from
the intestine into the fat depots and thence back into the blood stream to
the liver. The half-life of the fat depots in the rat was found to be 6-8 days
(Schoenheimer & Rittenberg, 1936). When hyperlipaemic blood circulated
through the capillaries, there was a marked reduction in glyceride content as
shown by simultaneous arterial and venous analyses. There would seem to
be no doubt, therefore, that lipids can pass into the adipose tissue cell, be
retained in the cell for storage purposes and mobilized again into the blood
stream ultimately to be used by the tissues.
(iii) Form of lipid transported
The lipid in the adipose tissue cells consists of triglyceride esters of long-
chain fatty acids. There is a close similarity between the depot fats and
dietary fat provided that there is a reasonably high level of fat in the diet
(Hilditch, 1947). The selection of the more saturated fats for deposition in
the fat depots can be explained on the basis of the Partition Hypothesis of
fat absorption — synthesized fats also tend to belong to the saturated series.
Hilditch & Stainsby (1935) have suggested that biohydrogenation may
occur. In any case, the lipid is presented to the intestinal cell as fine par-
ticles of triglyceride. There is no evidence at present to indicate whether
these particles of fat undergo any preparatory changes before entry into the
adipose tissue cell. Although esterases occur in most tissues, there is no
evidence of a significant degree of true lipase activity in adipose tissue —
the introduction of a potent lipase into such tissues causes classical 'fat
necrosis'.
The adipose tissue cells are continually releasing glycerides into the
circulation, so that a fairly rapid turn-over of the fat occurs. The mobiliza-
tion of the depot fat causes a visible hyperlipaemia consisting of a marked
increase of particulate glycerides. By what means, and in what form, this
lipid passes from the adipose tissue cells is not known.
TRANSPORT OF LIPID THROUGH CELL MEMBRANES 497
(iv) Factors concerned in deposition and mobilization of fat from adipose
tissue cells
The main factor that determines the accumulation of fat in fat depots is
the relationship of supply to demand. Thus appetite greater than that
required to meet the energy demand results in surplus intake of carbo-
hydrate and fat, both of which may give rise to increased deposition of
glyceride in the adipose tissue cells. As already mentioned, heparins and
antiheparins may play some unknown part in the mechanism of removal of
paniculate fat from the blood stream. The major cause of obesity is probably
excessive synthesis of fat from non-lipid sources ; this synthesis is stimulated
by insulin (Renold, Marble and Fawcett, 1950; Scott & Engel, 1950).
Various factors are alleged to enhance the mobilization of fat from the
adipose tissue cells. An endogenous hyperlipaemia occurs in starvation
but to what extent this is an unmasking of normal turn-over which is
obscured by alimentary hyperlipaemia is not known. A more definite fat
mobilization seems to follow the administration of lipogenic hormones of
the anterior part of the pituitary gland (Stetten & Salcedo, 1945 ; Campbell
& Lucas, 1951). Adrenalectomy in rats is said to alter the mobilization of
fat from adipose tissue (Stoerk & Porter, 1950). Nervous lesions (Wert-
heimer, 1926) also influence the rate of mobilization from the affected area.
Fenton & Carr (1951) demonstrated genetic factors — yellow mice becoming
obese while their non-yellow litter-mates did not, although consuming the
same synthetic diet.
r. o. , V. LIVER CELLS
(i) structure
The liver cell is enclosed in a membrane that comes into intimate contact
with the blood stream. Lipids enter into the structure of the liver cell: the
mitochondria may contain more than 25 % of lipid, mostly phospholipid
containing arachidonic acid. These highly unsaturated fatty acids do not
appear to undergo oxidative degradation and may be protected by their
close association with protein (Bensley, 1937; Kennedy & Lehninger, 1949).
The integrity of the liver cell is important for phospholipid and cholesterol
synthesis (Ada, 1944; Hevesy, 1945; Bloch, Borek & Rittenberg, 1946).
(ii) Evidence for lipid transport into and out of the liver cell and types of lipid
involved
It can be readily demonstrated that intact or isolated liver preparations
utilize labelled fatty acids or glycerides. This metabolism of lipids seems to
occur inside the cell and, if it is interfered with, intracellular accumulation
of lipid may occur. It seems clear that lipid can readily pass from the out-
E B S VIII 32
498 TRANSPORT OF LIPID THROUGH CELL MEMBRANES
side to the inside of the liver cell. Does lipid also pass in the other direction ?
Some lipids, such as phospholipid and cholesterol, must pass from inside
the liver cell into the blood since the turn-over of these molecules in the
plasma is a reflexion of turn-over in the liver. Glyceride, on the other
hand, does not normally pass out from the liver cell again but is utilized.
It was at one time thought that long-chain saturated fatty acids were de-
saturated in the liver and that the unsaturated fatty acids formed were
transported to the extrahepatic tissues for combustion. It is now clear,
however, that the most unsaturated fatty acids in the liver are neither
oxidized nor transported away from the liver, that desaturation is not a
necessary preliminary to utilization of fats by the liver cells or extrahepatic
tissues, and that saturated fats can be utilized by the extrahepatic tissues in
the absence of the liver. In choline deficiency fats accumulate in the liver ;
this has been attributed to faulty phospholipid formation (McArthur,
Lucas & Best, 1947), and it has been suggested that this prevents the trans-
port of the fat away from the liver cells. Hartroft (1951) has demonstrated
the accumulation and removal of fat in the choline deficient liver in a series
of dramatic histological preparations. While his interpretation may be
correct under the abnormal conditions of gross fat accumulation observed
in choline deficiency, it can be shown that transport of lecithin away from
the liver is not an effective form of expenditure of liver glycerides. Choline
deficiency does not impair the ability of the glycerides to enter the liver
cells, but it probably interferes in some way with the utilization of glycerides
within the cell. Thus the available evidence supports the view that gly-
cerides can be transported into, but not out of, the liver cells except as
degradation products.
(iii) Factors affecting liver lipid
The essential factor affecting the glyceride in the liver cells is the rela-
tionship of supply and demand. If large quantities of lipid come to the
liver, as in starvation or after a heavy fatty meal, accumulation of glyceride
may occur. Alternatively, interference with the rate of utilization of
glycerides, as in poisoning or choline deficiency, may also cause a signifi-
cant increase in intracellular and extracellular glyceride. The origin of this
lipid from the diet or fat depots can be checked by the use of labelled fats.
Endocrine factors control the synthesis of lipids by the liver, insulin
favouring fatty acid and cortisone cholesterol synthesis. Thus, marked
changes in liver lipids that are not concerned in any way with lipid transport
can occur.
During recent years much attention has been paid to phospholipids and
cholesterol as possible forms of transportable lipid. Emphasis has been
TRANSPORT OF LIPID THROUGH CELL MEMBRANES 499
largely placed on formation of some water-soluble and diffusible compound
or complex. It is becoming increasingly apparent that this is not the case,
and that the most important lipid component in fat transport during
intestinal absorption, or in the blood, or for storage in adipose tissue, or for
metabolic use by the liver and extrahepatic tissues, is glyceride fat in
particulate form. Phospholipid plays an important part in cell structure ; it
is essential to the stability of particulate glyceride in the blood and may
represent a vital step in the metabolism of glycerides in the liver cell, but
it does not appear to be significantly used as an intermediate in lipid
transport. Cholesterol also has structural functions, and it can be slowly
transported to the tissues, but it is not an important factor in fatty acid
transport. The passage of finely dispersed, negatively charged fat particles
through membranes raises new problems, many of which cannot be solved
until certain basic information becomes available.
VI. SUMMARY
1. Since hydrolysis of glycerides is incomplete in the intestinal lumen,
lipids are 'partitioned' before absorption into water-soluble and water-
insoluble fractions.
2. The water-insoluble glyceride fraction enters and leaves the
intestinal cell in particulate form.
3. Particulate fat passes preferentially into the lacteals while the water-
soluble lipid may enter the blood capillaries direct.
4. The main form of lipid transported in the blood stream is particulate
triglyceride. Phospholipids and cholesterols are only removed extremely
slowly by the extrahepatic tissues. The turn-over of lecithin and chole-
sterol in the plasma is a reflexion of turn-over in the liver.
5. Particulate glyceride is removed from the blood stream into adipose
tissue. Heparins and antiheparins may play an important part in the
removal of particulate fat from the blood.
6. Fat is mobilized from the fat depots and appears in the blood as
triglyceride in particulate form. The mechanism of mobilization is obscure.
7. The liver takes up particulate glycerides and metabolizes them.
There is no adequate evidence of fatty acids or glycerides passing out from
the liver cells. Phospholipids and cholesterol synthesized in the liver cells
can pass into the blood stream, but the liver is the main site for their
removal from the blood.
32-2
500 TRANSPORT OF LIPID THROUGH CELL MEMBRANES
REFERENCES
ADA, G. L. (1944). Biochem.J. 45, 422.
BAKER, J. R. (1942). Quart. J. Micr. Set. 84, 73.
BAKER, J. R. (1951). Quart.?. Micr. Sci. 92, 79.
BARNES, R. H., MILLER, E. S. & BURR, O. (1941). J. Biol Chem. 140, 241.
BAVETTA, L. A. & DEUEL, H. J. Jr. (1942). Amer.J. Physiol. 136, 712.
BENSLEY, R. R. (1937). Anat. Rec. 69, 341.
BIGGS, M. W., FRIEDMAN, M. & BYERS, S. O. (1951). Proc. Soc. Exp. Biol., N.Y.,
78, 641.
BLOCH, K., BOREK, E. & RITTENBERG, D. (1946). J. Biol. Chem. 162, 441.
BLOOM, B., CHAIKOFF, I. L., REINHARDT, W. O., ENTENMAN, C. & DAUBEN, W. G.
(1950). 3- Biol. Chem. 184, i.
BLOOM, B., CHAIKOFF, I. L. & REINHARDT, W. O. (1951). Amer. J. Physiol. 166,
45i.
BLOOR, W. R. (1943). Biochemistry of the Fatty Acids. New York: Reinhold
Publishing Corporation.
BORGSTROM, B. (1952). Acta physiol. scand. 25, 140.
BROWN, W. D. (1952). Quart. J. Exp. Physiol. 37, 75.
CAMPBELL, J. & LUCAS, C. C. (1951). Biochem.J. 48, 241.
CHAIKOFF, I. L., BLOOM, B., STEVENS, B., REINHARDT, W. O. & DAUBEN, W. G.
(1951). J. Biol. Chem. 190, 431.
DANIEL, J. W., FRAZER, A. C., FRENCH, J. M. & SAMMONS, H. G. (1951). J.
Physiol. 114, 26 P.
EGER, W. (1938). Klin. Wschr. 17, 1033.
ELKES, J. J., FRAZER, A. C. & STEWART, H. C. (1939). J> Physiol. 95, 68.
ENTENMAN, C., CHAIKOFF, I. L. & ZILVERSMIT, D. B. (1946). J. Biol. Chem. 166,
15-
FAVARGER, P. & COLLET, R. A. (1949). Helv. Physiol. Pharmacol. Acta, C, 8, 15.
FAVARGER, P., COLLET, R. A. & CHERBULIEZ, E. (1951). Helv. chim. acta, 34, 1641.
FENTON, P. F. & CARR, C. J. (1951). J. Nutr. 45, 225.
FRAZER, A. C. (1938). Analyst, 63, 308.
FRAZER, A. C. (1946). Physiol. Rev. 26, 103.
FRAZER, A. C. (19480). J. R. Soc. Arts, 96, 582.
FRAZER, A. C. (19486). Arch. Sci. Physiol. 2, 15.
FRAZER, A. C. (1951). Parsons' Modern Trends in Paediatrics. London: Butter-
worth.
FRAZER, A. C. (19520). Biochem. Soc. Symp. 9.
FRAZER, A. C. (19526). Avery Jones' Modern Trends in Gastro-enterology . London :
Butterworth.
FRAZER, A. C. (1952 c). Proc. European Congr. Gastroent. 3, 60.
FRAZER, A. C. (1952^). Trans. R. Soc. Trop. Med. Hyg. 46, 576.
FRAZER, A. C., ELKES, J. J., SAMMONS, H. G., GOVAN, A. D. T. & COOKE, W. T.
(1945). Lancet, i, 457.
FRAZER, A. C., SAGROTT. P. E. & SAMMONS, H. G. (1949). Proc. Int. Congr.
Biochem. i, 596.
FRAZER, A. C., SCHULMAN, J. H. & STEWART, H. C. (1944). J- Physiol. 103, 306.
GEYER, R. P. & CUNNINGHAM, M. (1950). J. Biol. Chem. 184, 641.
GEYER, R. P., CUNNINGHAM, M. & PENDERGAST, J. (1950). J. Biol. Chem. 185, 461.
GOFMAN, J. W., JONES, H. B., LINDGREN, F. T., LYON, T. P., ELLIOTT, H. A. &
STRISOWER, B. (1950). Circulation, 2, 161.
GOLDMAN, D. S., CHAIKOFF, I. L., REINHARDT, W. O., ENTENMAN, C. & DAUBEN,
W. G. (1950). J. Biol. Chem. 184, 719.
TRANSPORT OF LIPID THROUGH CELL MEMBRANES 501
GRANGER, B. & BAKER, R. F. (1950). Anat. Rec. 107, 423.
GURD, F. R. N., ONCLEY, J, L., EDSALL, J. T. & COHN, E. J. (1949). Disc. Faraday
Soc. 6, 70.
HAHN, P. F. (1943). Science, 98, 19.
HARTROFT, W. S. (1951). Ciba Foundation Symp. on Liver Disease. London:
Churchill.
HEVESY, G. C. (1945). Nature, Lond., 156, 534.
HILDITCH, T. P. (1947). The Chemical Constitution of Natural Fats. London:
Chapman and Hall.
HILDITCH, T. P. & STAINSBY, W. J. (1935). Biochem.J. 29, 90.
HOLT, L. E., AYLWARD, F. X. & TIMBRES, H. G. (1939). Johns Hopk. Hosp. Bull.
64, 279.
KARNOVSKY, M. L. & GIDEZ, L. I. (1951). Fed. Proc. 10, 205.
KENNEDY, E. P. & LEHNINGER, A. L. (1949). Fed. Proc. 8, 213.
KEYS, A. (1952). Proc. Int. Dietetic Congr. i.
KEYS, A., MICKELSEN, O., MILLER, E. v. O., HAYES, E. R. & TODD, R. L. (1950).
y. Clin. Invest. 29, 1347.
KIYASU, J. Y., BLOOM, B. & CHAIKOFF, I. L. (1953). J. Biol. Chem. 199, 415.
LAWRENCE, R. D. (1946). Lancet, i, 724, 733.
MCARTHUR, C. S., LUCAS, C. C. & BEST, C. H. (1947). Biochem.J. 41, 613.
REISER, R., BRYSON, M. J., CARR, M. J. & KUIKEN, K. A. (1952). J. Biol. Chem.
194, 131-
RENOLD, A. E., MARBLE, A. & FAWCETT, D. W. (1950). Endocrinology, 46, 55.
SCHOENHEIMER, R. (1931). Science, 74, 579.
SCHOENHEIMER, R. & RlTTENBERG, D. (1936). J. Biol. Chem. 121, 249.
SCOTT, J. L. & ENGEL. F. L. (1950). Endocrinology, 46, 574, 582.
STETTEN, D. & SALCEDO, J. (1945). J. Biol. Chem. 156, 27.
STANLEY, M. M. & THANNHAUSER, S. J. (1949). J. Lab. Clin. Med. 34, 1634.
STOERK, H. C. & PORTER, C. C. (1950). Proc. Soc. Exp. Biol., N.Y., 74, 65.
TUERKISCHER, E. & WERTHEIMER, E. (1942). y. Physiol. IIO, 385.
VERZAR, F. & LASZT, L. (1934). Biochem. Z. 270, 35.
VERZAR, F. & LASZT, L. (1935). Biochem. Z. 276, 396.
VERZAR, F. & McDouGALL, E. J. (1936). Absorption from the Intestine. London:
Longmans Green and Co.
WELD, C. B. (1944). Canad. Med. Ass. J. 51, 578.
WELD, C. B. (1946). Canad. Med. Ass. J. 54, 71.
WENDT, C. F. (1937). Z. phys. Chem. 249, 4.
WERTHEIMER, E. (1926). Pflug. Arch. ges. Physiol. 213, 262.
32-3
MORPHOLOGICAL AND MOLECULAR
ASPECTS OF ACTIVE TRANSPORT
BY J. F. DANIELLI
Department of Zoology, King's College, London, W.C. 2
I. INTRODUCTION
In the majority of papers contributed to this Symposium the morphological
and molecular aspects of active transport and secretion have been some-
what neglected, although the necessity for paying attention to these aspects
was emphasized in the introductory papers of Ramsay and of Davson, and
in the paper by Frazer. To some extent contributors to the discussion,
particularly V. B. Wigglesworth, have emphasized the importance of
paying more attention to the morphology of cells and membranes involved
in active transport. But for the most part the papers have accepted as their
starting point that active transport occurs, and have concentrated on studies
of kinetics of transfer, upon the linkage of transport with metabolism and
upon the facts revealed, for example, by the use of competing molecules and
of enzyme poisons. This paper will concentrate on the two aspects, morpho-
logical and molecular, which have been relatively neglected.
II. ANALYSIS OF ACTIVE TRANSFER AT DIFFERENT
MORPHOLOGICAL LEVELS
There are at least six levels at which it is profitable to consider active
transfers. The first level is that of the simplest known cellular membranes,
such as those of red blood cells, where morphological complications are
minimal. . At this level we encounter processes ranging from simple diffusion
under thermal agitation, through facilitated diffusion, up to the full
mechanism of active transfer. Thus with human red cells ethyl alcohol
appears to pass through the membrane by simple diffusion, glycerol and
glucose by facilitated diffusion and simple diffusion, sodium and potassium
by active transport and simple diffusion. Electron-microscope studies have
not so far given any indication that the red cell membrane is at all com-
plicated morphologically. At present simple diffusion processes have a
satisfactory quantitative theory, theories are being developed for facilitated
diffusion, and active transport is largely lacking a theory.
The second level is that of the cell membranes which are specialized for
the performance of active transport, such as the membranes of intestinal
epithelial cells, kidney tubule cells and motor end-plates. These specialized
MOLECULAR ASPECTS OF ACTIVE TRANSPORT 503
structures are often very rich indeed in particular enzymes, e.g. alkaline
phosphatase at the surface of intestinal epithelium and proximal tubule
cells, choline esterase at the surface of motor end-plate cells. In A. C.
Frazer's paper in this volume, particular attention is directed to the prob-
ability that the structure of the free border of the intestinal epithelium is
critical in fat transport. However, in general we have at present very little
knowledge of the real functions of these specialized cell membranes. It
has been suggested that the high concentrations of enzymes in these
specialized membranes are concerned in active transport by serving as the
centres through which chemical energy is transformed into mechanical
energy, just as adenosine triphosphatase acts as the centre through which
the chemical energy of adenosine triphosphate is transformed into the
mechanical energy for contraction of muscle actomyosin (Danielli, 1951,
1952, 1953, 1954). Recently Sjostrand & Rhodin (1953) have shown by
electron microscopy that the apical zone of kidney proximal tubule cells
is a system of fine ducts, about 600 A. in diameter, thus resembling the
apical zone of intestinal epithelial cells, but having ducts which are smaller
in diameter. It seems possible that an exploration of the function of enzymes
in relation to this type of structure would be richly rewarding.
The third level is that of mitochondria. The work of Davies and his
colleagues (reported in this Symposium) shows that mitochondria have a
remarkable ability to transport ions across their membranes. Sjostrand
(1953) and Sjostrand & Rhodin (1953) have revealed to a remarkable degree
the complex membrane structure of mitochondria, but it has not so far
been possible to relate the physiological function of mitochondria to their
structure and enzyme organization, so far as the field of active transport is
concerned. Sjostrand & Rhodin have also made many suggestive observa-
tions on the relationship between mitochondria and cellular membranes.
Unfortunately, much of this work has been on osmium tetroxide fixed
material rather than on frozen-dried material, so it is uncertain how far
these observations have physiological significance. Some incidental
observations which I have made on the mitochondria of frozen-dried
sections suggest that mitochondria may at times protrude into the lumen
of a tubule or become part of the cell membrane. At present we do not
know, for example, how mitochondria participate in ionic regulation in a
mammalian kidney. It would be very much easier to understand how this
occurred if we knew that mitochondria project into the lumen of a tubule or,
as Wigglesworth suggests, absorb material from the lumen by being applied
to the base of the apical ducts.
The fourth level of approach is to vacuoles. What is the significance,
from the point of view of active transfer, of those vacuoles which include
MORPHOLOGICAL AND MOLECULAR
among their properties the ability to concentrate neutral red? Are the
osmotic and regulative properties of contractile vacuoles to be ascribed
solely to their associated mitochondria, or have the vacuoles intrinsic
regulative activity?
The fifth level of approach is to pinocytosis,* the phenomenon of
absorption of environmental fluid as discrete droplets by undulating mem-
branes, first observed by W. H. Lewis. Anyone who has seen a film showing
the flood of vacuoles passing across a cell as a result of pinocytosis will
readily realize that any theory of kinetics of penetration through cell
membranes must break down utterly if applied to a cell which can display
pinocytosis.
The sixth approach is to discover the complete physiological significance
of extracellular processes, such as lymph flows, capillary flows and pumping
activities such as those displayed by intestinal villi.
A complete understanding of active transport in an organ will only be
possible when we can give an integrated account of the functions of all the
processes concerned with active transport which occur at these six morpho-
logically distinct levels. In making the analysis which will permit such an
integrated account we must recognize both the hierarchy of morphological
units discussed above and a hierarchy of processes which may occur as part
of the functioning of each morphological unit. This hierarchy of processes
includes (a) simple diffusion, (b) facilitated diffusion and (c) active transport.
Simple diffusion, i.e. movement of molecules under the driving force of
thermal agitation, unrestricted by steric factors and unlimited by mole-
cular structure, can be handled quantitatively. Facilitated diffusion cannot
yet be treated quantitatively, and has been recognized as a process occurring
at a limited fraction of the cell surface (at so-called active patches), only
since quantitative treatment of simple diffusion through cell membranes
was possible (Danielli, 1943). Facilitated diffusion occurs under the driving
force of thermal agitation, and the equilibrium reached is the same as that
achieved by simple diffusion. But facilitated diffusion at any one site,
unlike simple diffusion, is restricted both by structural and steric factors so
that only a small number of molecular species are concerned. For example,
a hexose penetrates cell membranes by simple diffusion about io3-io4 times
more slowly than does ethanol. But at a site of facilitated diffusion specific
for certain hexoses, these hexoses will penetrate at a rate comparable to
ethanol, i.e. io3-io4 times faster than by simple diffusion. For the moment
we may contrast facilitated diffusion with active transport by defining
active transport processes as those in which molecules are transferred across
membranes by the use of an energy supply other than, or additional to,
* At the Symposium a film of pinocytosis, lent by A. Hughes, was shown at this point.
ASPECTS OF ACTIVE TRANSPORT 505
thermal agitation. But it is wisest to regard these definitions as of transient
value, for, on the one hand, as our knowledge increases it may become
necessary to subdivide what we now call 'active transport', and, on the
other hand, as we shall see later, some conceivable types of facilitated
diffusion may require energy for maintaining structural units, including
carriers or expanded lattices, although the actual movement of a molecule
from one side of a membrane to the other may proceed under thermal
agitation only. Thus there may not be a sharp division between all forms of
facilitated diffusion, and all forms of active transfer, so far as energy re-
quirements are concerned.
It seems to me better to adopt this analytical approach to active transport,
and recognize that some of our definitions are of transient value, rather than
to adopt the method, which Rosenberg has advocated in this Symposium,
of attempting to find a rigorous all-inclusive definition of active transport,
for the latter method is handicapped by emphasis on energy differences
between initial and final states, to the exclusion of the processes intervening
between these states. For example, a cellular membrane may well transfer
sodium chloride solution from one side to the other of the membrane, say
by a vacuolar process. Thermodynamically, the initial and final states on
the two sides of the membrane are indistinguishable, and thus by Rosenberg's
definition the process is not active transfer. But since vacuolar processes
require energy, to a biologist this transfer is active. Thermodynamically it
would be recognized as active transport if the free energy of the whole
system (including the membrane) was assessed. But this, in the majority
of cases, would be impossibly difficult, since it is usually impracticable to
measure experimentally what part of the free-energy change occurring in
a cellular membrane is associated with active transfer rather than with other
cellular processes.
III. THE RELATIONSHIPS OF ENZYMES AND
ENZYME POISONS TO TRANSFER
That certain substances which are enzyme poisons inhibit transfer processes
is certain; how these substances act is a matter of conjecture. Outstanding
examples of such action include the inhibition of glucose transfer in
intestinal epithelial cells by iodoacetate (Wilbrandt & Laszt, 1933); of
glucose transfer in kidney proximal tubules by phloridzin (Nakazawa, 1922) ;
of facilitated diffusion of glycerol in red cells by copper (Jacobs & Corson,
1934), by SH reagents (LeFevre, 1948); of facilitated diffusion of glucose in
red cells by phloretin phosphate (Wilbrandt & Rosenberg, 1950), by
dinitrofluorobenzene and by diazonium hydroxides (Bowyer, 1953); of
active transfer of sodium and potassium in red cells by eserine (Holland
506 MORPHOLOGICAL AND MOLECULAR
& Grieg, 1950) and by diazonium hydroxides (Bowyer, 1953). But it does
not follow that these substances necessarily act by poisoning enzymes.
Enzymes, by definition, are cellular catalysts, i.e. enzymes are substances
found in (or synthesized by) cells, which accelerate chemical reactions but
do not modify chemical equilibrium conditions. Since not all cellular
processes are chemical, it is improbable that all accelerated biological
processes are normally accelerated by enzymes. Precise evidence is re-
quired before we can decide whether a process is enzymic or not. To illus-
trate this necessity we may consider generalized cell permeability: at one
time, because permeations had temperature coefficients of the same order
of magnitude as those found for chemical reactions, it was supposed that all
permeations involved chemical reactions (and must be enzymic). This rather
widely held theory was based almost entirely on this one point of resem-
blance between permeations and chemical reactions. But a more detailed
analysis showed that this resemblance was misleading ; any process in which
a resistance, which may be represented by an activation energy, is overcome
directly by thermal agitation will be exponentially related to temperature.
Chemical reactions are one such process; permeations of cell membranes
by simple diffusion are another. Hence these processes show similar
temperature coefficients (Danielli, 1934). This example makes explicit the
danger of concluding that two processes are identical because they have a
few quantitative or qualitative similarities. Facilitated diffusion and
enzymic processes have several such similarities: they are inhibited by
similar substances; structural and steric properties are of importance in
both cases; the processes have similar temperature coefficients. But such
resemblances are no proof of identity. In 1939 I was able to show, by
analysis of the kinetic studies of Jacobs, that glycerol penetrates the red
cells of certain species by two processes, one of which was simple diffusion
and the other involved a selective activity of a small fraction (less than i %)
of the surface area of the cells. These active patches in many ways resembled
enzymes, and this resemblance was strengthened by Davson's (1942)
studies. But there is still no proof available that enzymes are in fact in-
volved. Proof could perhaps be obtained by inhibiting a facilitated diffusion
process with an irreversible reagent such as dinitrofluorobenzene, followed
by a demonstration of a strict parallelism between enzyme inhibition and
inhibition of facilitated diffusion, backed up by demonstration that enzyme
isolated in purified condition from inhibited cells was in fact combined with
the inhibitor. At present we lack evidence of this quality for a single case.
On the other hand, if we make a short list of enzymic processes which
may be involved indirectly in facilitated diffusion and active transfer, we at
once see that many enzymes may be concerned and that many substances
ASPECTS OF ACTIVE TRANSPORT 507
may poison them, so stopping transfer, and yet none of them may be in-
volved in the direct process of transfer. Enzymic processes of interest in
active transport include :
(a) Enzymes providing energy for transfer , e.g. all enzymes concerned in
adenosine triphosphate synthesis are of interest if ATP is the basic energy
source. Phosphokinase inhibitors, or substances like dinitrophenol which
break the linkage between phosphate uptake and oxidation, will inhibit
transfers based on ATP, without necessity for a direct action on the process
of transfer.
(b) Enzymes with a trapping function. Once a substance has entered a
cell its accumulation may be assisted by conversion to a derivative, e.g.
glucose and inorganic phosphate may be accumulated as a less diffusible
glucose phosphate. If glucose is the penetrating species, the conversion
may be effected by ATP, catalysed by a hexokinase. If phosphate is the
penetrating species, it may be trapped by formation starch or glycogen
catalysed by a phosphorylase, or taken up in an enzymically catalysed
phosphorylative oxidation. In both cases the enzyme has a trapping function
and accumulation will be inhibited if the enzyme is inhibited.
(c) Enzymes with a maintenance function. It is possible that the active
regions of a cell membrane are unstable and require * servicing ' by enzymes.
For example, enzymic phosphorylation of an unstable carrier molecule
may be required, or the active form of the membrane may be an expanded
lattice structure which collapses into a more stable structure and requires
energy for re-expansion. The servicing enzymes in such cases would not
participate directly in transfer, and there would probably be no stoichio-
metric relationship between maintenance energy and number of molecules
transferred.
(d) Enzymes forming or emptying vacuoles. In addition to any enzymes
which are concerned with the transfer of substances into vacuoles, there
may be other enzymes concerned in the movement and discharge of
vacuoles, actuating contractile proteins, as is found with the ATPase-
actomyosin complex. Inhibition of such enzymes will also inhibit over-all
active transfer.
(e) Enzymes participating directly in active transfer. The contractile
protein theory of Goldacre (1952), as developed by Danielli (1951, 1952,
1953, 1954), is concerned with one type of mechanism whereby an enzymic
centre is an integral part of a true active transfer mechanism. It is only in
cases where an enzyme poison acts upon mechanisms of this type that we
learn anything of the actual mechanism of transfer across a membrane.
It seems likely that enzymes which are concerned with actuating or
servicing the direct mechanisms of transfer will vary from cell to cell, and
508 MORPHOLOGICAL AND MOLECULAR
also with the type of molecule to be transferred, for such variation will
permit transfers of different molecules to occur independently (Danielli,
1954). It is of interest in this connexion that the evidence at present
available suggests that phosphatases are concerned in the transfer of glucose
across the membranes of red cells, intestinal epithelial cells and kidney
proximal tubule cells, whereas choline esterase is concerned with transfer
of sodium and potassium across red cell membranes and motor end-plate
membranes.
IV. THE IMMEDIATE MOLECULAR MECHANISM OF
FACILITATED DIFFUSION AND ACTIVE TRANSFER
The plasma membrane of most cells appears to consist basically of a lipoid
layer about 5oA. in thickness (roughly bimolecular) with protein layers on
either side. Movement of molecules across such membranes is limited
mainly by hydrogen-bond formation between water and the molecules
concerned, and by the effective viscous and structural resistance of the
membrane. In the main, facilitated diffusion is concerned with the ac-
celerated transfer of molecules which penetrate only slowly by simple
diffusion because of extensive hydrogen-bond formation, and active trans-
port is concerned primarily with accelerated transfer and establishment of
concentration gradients of the same molecular species. Both facilitated
diffusion and active transport must occur within the limits laid down by the
size and structure of the membrane, and the central problem is probably
that of overcoming the hydrogen-bond limitation, for there is little evidence
so far that reducing the viscous and structural resistances of the membrane
are ever of importance except in so far as these factors also are involved in
the hydrogen-bond problem. There appear at present to be two alternative
approaches to this central problem: the first from the point of view of
macromolecular processes of the membrane and the second from the point
of view of the molecular process acting upon the transferred molecule.
Macromolecular membrane processes
(a) The simplest process which might have validity is the diffusing
shuttle or carrier molecule, to which Osterhaut has given much attention
(Fig. i a). The essentials of such a molecule are that it should be relatively
insoluble in water and readily dissolved in lipoid, should not itself form
many hydrogen bonds with water, should readily and reversibly form a
complex with the molecule to be transferred which also will not form many
hydrogen bonds with water, and further that both the free carrier and the
complex should diffuse readily in the lipoid layer. If these criteria were
satisfied, the carrier molecules, while executing the random movement of
ASPECTS OF ACTIVE TRANSPORT
509
thermal agitation, would effectively shuttle between the two sides of the
membrane. Such mechanisms will accelerate the attainment of equilibrium,
and so may be classified as facilitated diffusion. Wilbrandt & Rosenberg
(1952) have discussed such a system in relation to glucose transport.
(6) A second mechanism is that of the propelled shuttle (Fig. i b). In
this mechanism an enzyme centre such as phosphatase or choline esterase
la
Membrane
—Co-
Membrane
Adsorption carrier i
4
*
Protein
— Enzyme centre
Diffusing shuttle or carrier Propelled shuttle or carrier
Membrane
Membrane
1c
Adsorption centre
Adsorption centre
Rotating
carrier
Rotating
segment
Rotating molecular carrier Rotating molecular segment carrier
Fig. i.
supplies energy to actuate a contractile protein. The contractile protein has
an adsorption centre which will be on one side of the membrane when the
protein is contracted and on the other side when the protein is extended.
The properties required of the adsorption centre are similar to those re-
quired for the carrier molecule in mechanism (a). According to the details
of the contraction and expansion, such a process could merely accelerate
the attainment of equilibrium, or cause the building-up of concentration
gradients. This mechanism has been discussed in more detail elsewhere
(Danielli, 1952, 1953, 1954).
5IO MORPHOLOGICAL AND MOLECULAR
(c) A third mechanism is that of a rotating molecular carrier (Fig. i c).
A molecule of diameter equal to the thickness of the plasma membrane will
rotate very rapidly in a medium of the viscosity of water, and even in a
medium t)f the apparent viscosity of the plasma membrane (about io5 times
that of water) the rate of rotation will be high. Some special properties
would be required to retain the molecule in the membrane, such as in-
ability to form hydrogen bonds with water. And it would need to have a
special adsorption centre with the same properties as the adsorption centre
in mechanism (6), or the carrier in mechanism (a). With such properties a
rotating carrier would provide a mechanism for facilitated diffusion.
(d) The concept of a rotating carrier molecule presents certain mechanical
and stability problems which would perhaps be more readily resolved if the
rotating unit were not a whole molecule, but a molecular segment (Fig. i d).
A molecular segment such as that shown in the figure would be able to
rotate about an axis composed of two covalent bonds, such as carbon-
carbon, carbon-oxygen, or carbon-nitrogen. The energy required for
rotating about such bonds is of the order of 3000 cal., so that the average
energy required to rotate a segment would be 6000 cal., and the proportion
of such segments rotating at any time would be proportional to e~QOGQIRT.
The order of magnitude of the average rate of rotation would be given by
kT=Iw2y where / is the moment of inertia and w the angular velocity of
the segment. An adsorption centre would be required as in mechanism (c).
It is not quite inconceivable that the rotating units in mechanisms (c)
and (d) could be energized by enzymic centres, so that the heat of reaction
at the enzyme centre gave additional rotational energy to the unit.
(e) A fifth quite different mechanism can be based upon the idea of
expanded lattices. If the structure of the lattice of say a protein in the
membrane were expanded, molecules of the correct size and structure would
be able to penetrate into and through the lattice, and thus through the
membrane. This would constitute facilitated diffusion. The energy for this
expansion might be provided by enzyme action. F. Bowyer has pointed
out to me that if such a lattice expanded and collapsed asymmetrically, it
would constitute a pump imparting one-way transfer to those species able
to enter the expanded lattice.
The five types of macromolecular process envisaged above, with their
variants, provide possible mechanisms, the energy requirements of which
range from zero in some cases, through a ' servicing' energy requirement
in other cases to further cases in which the energy requirement would be
stoichiometrically related to the rate of transfer. These mechanisms have in
common the necessity for an adsorption centre (or lattice) which is specific
for a limited range of molecules. The forces available for adsorption with
ASPECTS OF ACTIVE TRANSPORT 511
most molecules of biological interest appear at first sight to be limited to
Van der Waals forces, electrostatic forces and hydrogen-bond forces
(Danielli, 1944), which are not specific in themselves. Consequently
specificity must reside in the distribution of these forces in space. However,
exactly the same physical problem arises in the determination of the
specificity of enzymes. It therefore seems quite possible, for example, that
a substance which will inhibit the hydrolysis and synthesis of glucose
phosphate by phosphatase will also interact with the specific adsorption
centres of transfer mechanisms for glucose or phosphate, and so inhibit
these also. Thus at this stage of the analysis it appears that inhibition of a
specific transfer process by a specific enzyme poison is not good evidence
that the enzyme for which the poison is specific is part of the intimate
mechanism of transfer.
The molecular process acting upon the transferred molecule
Turning now to consider the molecules which are transferred, our
attention is immediately focused upon the mechanisms whereby hydrogen
bonding between these molecules and water may be broken. Specific
adsorption upon a carrier unit will be almost useless in all the mechanisms
described above (except possibly the expanded lattice, mechanism (e))
unless the number of hydrogen bonds is reduced to a maximum of two.
In the case of glucose, for example, which forms five hydrogen bonds with
water, the activation energy for permeation into cells by simple diffusion is
of the order of 15,000-25,000 cal. With an activation energy as high as this
permeation is very slow. When glucose is penetrating into human or rabbit
red cells by facilitated diffusion, the activation energy is reduced to about
4000 cal., so that four out of the five bonds have been broken by some
process other than thermal agitation, and penetration occurs at about the
same rate as is found for methyl alcohol, which also forms only one hydrogen
bond with water.
Some insight into possible mechanisms may be gained by considering
methods of denaturing proteins. The structure of proteins is largely main-
tained by hydrogen bonds ; these bonds are broken by adding protons (as
hydrogen ions) or by adding alternative hydrogen-bonding agents (e.g.
urea). It is reasonable to suppose that these two methods will also break
hydrogen bonds with water. Thus we conclude that for maximum efficiency
the specific adsorption centres in carrier molecules should provide hydrogen-
bonding groups and should be proton conductors so that protons may be
readily available. Proteins, nucleic acids and probably polysaccharides will
be able to act as hydrogen-bond formers, and in so far as these compounds
are themselves hydrogen-bonded, or rich in groups such as COOH, they
512 MORPHOLOGICAL AND MOLECULAR
will be proton conductors. On the other hand, the hydrocarbon moiety of
lipoids would be quite unable to form hydrogen bonds or conduct protons.
Whilst the ideal adsorption centre is a hydrogen-bond forming proton
conductor, its selectivity will be determined by its organization in space,
and here the distribution of hydrogen-bond forming groups may be of
particular importance. We now know that native proteins are often not
homogeneous in their structure, but consist of lamellae, each of which is
composed of one or more polypeptide chains. In each lamella the poly-
peptide chains are oriented so that the polar side-chains are in one surface
and the non-polar side-chains are in the opposite surface. In a soluble
protein these lamellae are grouped in pairs, with their non-polar surfaces
together, and with polar surfaces either exposed to water or facing a polar
surface of another pair of lamellae. Where polar surfaces of lamellae are in
contact with one another they may adhere strongly, as is the case, for
example, in haemoglobin, where the pairs of lamellae may be dissociated
by urea. Since such surfaces adhere so strongly it is possible that units
based on protein lamellae may extend right through the membrane, as
indicated in Fig. 2 a. With the arrangement of lamellae of Fig. 2 a a
hydrogen-bonding channel or slit will exist right through the membrane
(which will otherwise be lipoid in character). Such channels may contain
much water, as is the case with native proteins in crystals. But in so far as
such channels exist they must constitute a small fraction of the cell surface,
and must be too tightly bonded to permit entry of the great majority even
of small molecules. If this were not so we should not find the quantitative
relationship between oil-water partition coefficients and rates of penetration,
which has been established for many years. For example, if these hydrogen-
bonding channels were equivalent to aqueous channels, methyl alcohol and
glycerol should have the same activation energy of permeation and should
penetrate at almost the same rate (according to the equation PM* = constant).
In fact, the activation energies for permeation of these molecules are
normally about 4000 and 12,000 cal. respectively, and methyl alcohol
permeates at least 100 times faster than does glycerol. Slow permeation
for glycerol is often coexistent with rapid facilitated diffusion of other mole-
cules such as amino-acids and sugars. On the other hand, (Hollander has
shown that in some cells, although the great majority of molecules obey
the quantitative relationships expected for molecules diffusing through a
lipoid layer, the smallest molecules, such as formamide and methyl alcohol,
may enter cells more rapidly than is calculated, which he explains in terms
of minute pores in the membrane. Similarly, Jacobs (1952) and Ussing
(p. 409) have found anomalies in the rate of penetration of water which they
explain in terms of pores. It is possible that the hydrogen-bonding channels
ASPECTS OF ACTIVE TRANSPORT
SI3
indicated in Fig. 2 a would permit the rather indiscriminate passage of such
very small molecules, but would be quite impermeable to other molecules
unless these molecules had a specific structure.
Returning to Fig. 2cy we see that a polypeptide chain or lamella, forming
part of a hydrogen-bonding channel, may extend on to the outer and inner
surfaces of the lipoid part of the membrane, and thus could constitute the
active unit in a propelled shuttle or carrier (mechanism (6), Fig. 1 6). Thus
a system such as Fig. 2c, if not energized, would permit selective facilitated
diffusion through its hydrogen-bonding channels, and if energized would
provide active transport and a means of building up concentration gradients.
Hydrogen bond —
forming (probably
hydrated) channels .
Membrane
Lipoid
mm*
WWW
Non-polar side-chains
\ /
Polar side-chains
26
Lipoid
mm
Lipoid
2a
•*- A
2c
Fig. 2.
In his introductory paper to this Symposium, Davson has suggested that
the remarkable permeability of red cells to anions is a case of facilitated
diffusion; it is limited to very small anions such as Cl~. It is tempting to
suggest that this is due to the incorporation of haemoglobin lamellae in the
membrane, thus providing positively charged hydrogen-bonding channels
which will facilitate the diffusion of small anions.
Other characteristics of proton-conducting hydrogen-bonding systems
The synthesis and hydrolysis of esters, acetals, glycosides and peptides
are catalysed by protons (supplied as hydrogen ions). It might therefore
be suspected that a proton-conducting hydrogen-bonding system would be
514 MORPHOLOGICAL AND MOLECULAR
catalytic. This consideration reminded me that S. J. Holt had drawn my
attention to the catalytic effect of ion exchange resins. Cation-exchanging
resins are of necessity proton conductors, and although their surfaces, un-
like those of proteins, cannot be expected to be highly specific towards
reactants, some catalytic effect might nevertheless be expected. Papers by
Sussman (1946) and by Underwood & Deatherage (1952) show that cation-
exchanging resins are in fact catalytic for hydrolysis and synthesis of a wide
variety of compounds, including hydrolysis of proteins. A paper by
Thomas & Davies (1947) shows that resins are more effective catalysts than
can be explained by their hydrogen-ion content. By comparison with
hydrogen ions in bulk solution, the hydrogen ions of resins are more
effective by the following factors :
i-y-fold for hydrolysis of methyl acetate,
2'3-fold for hydrolysis of ethyl acetate,
9'8-fold for hydrolysis of butyl acetate.
The order of increasing efficiency is the same as the order of increasing
adsorption of the ester. Consequently it appears that in addition to the
catalytic effect of the protons of the resin, the resin surface is contributing
a second catalytic influence.
From these facts it is apparent that if we specify that the active membrane
unit in facilitated diffusion or active transport is a stereochemically specific
proton-conducting hydrogen-bonding system, we are probably also
specifying that this system shall have catalytic activity. Indeed, it may well
be that all enzymes of the great family which catalyse synthesis and
hydrolysis of esters, glycosides, peptides, etc., owe their activity to being
proton-conducting hydrogen-bonding systems whose stereochemistry
determines which reactions they may catalyse. The active membrane units
in facilitated diffusion and active transport may be members of this family
in the same sense that the haemoglobins are members of the family of
enzymes which owe their activities to the properties of the ion-porphyrin
complex. If the active membrane units are catalytic, fresh possibilities
arise for their selectivity; they may select molecules from the environment
by reversible exchange reactions. For example, an amino-acid may be
selected from the environment by an exhange reaction with a peptide chain
of the active unit, thus :
polypeptide - CO . NH . CUR . CO . NH - polypeptide + NH2 . CUR1 . COOH,
polypeptide - CO . NH . CKR1 . COOH + NH2CRR . CO . NH - polypeptide.
And after transfer to the far side of the membrane an exchange reaction in
the reverse direction would release the amino-acid. A similar process could
be involved in transport of sugars, etc.
ASPECTS OF ACTIVE TRANSPORT 515
It may well be that polynucleotides and polysaccharides can be integral
parts of such processes, in addition to proteins. Some polynucleotides have
recently been found to have peptidase activity, and the structure of nucleic
acids is now believed to be maintained by hydrogen bonding. Neuberg &
Roberts (1949) and Lansing & Rosenthal (1952) have obtained evidence
suggesting that polynucleotides are concerned in active transport.
V. CONCLUSION
It is apparent that physico-chemical considerations provide a wealth of
possibilities which need further consideration ; many of these arise from
Goldacre's theory of the function of contractile proteins. It is possible also
that studies of active transfer have brought us to the point where we must
reconsider the cellular function of enzymes. Cellular enzymes may be not
merely catalysts which facilitate certain reactions but polyfunctional centres
whose function in one context may be chemical, in another mechanical,
perhaps sometimes both. So far as active transport is concerned it seems
clear that future work on any biological system involves :
(1) Analysis of the number of morphological units involved, as outlined
in § II of this paper.
(2) Analysis of the number of distinct processes involved in each
morphological unit, as outlined in §§ II and III.
(3) Analysis of the molecular mechanism of each process, as outlined
in § IV.
REFERENCES
BOWYER, F. (1953). Personal communication.
DANIELLI, J. F. (1934). J. Cell. Comp. PhysioL 5, 495.
DANIELLI, J. F. (1943). In Davson & Danielli, Permeability of Natural Membranes,
Cambridge University Press.
DANIELLI, J. F. (1944). J. Exp. Biol. 20, 167.
DANIELLI, J. F. (1951). Nature, Lond., 168, 464.
DANIELLI, J. F. (1952). Symp. Soc. Exp. Biol. 6, i.
DANIELLI, J. F. (1953). Cytochemistry, John Wiley.
DANIELLI, J. F. (1954). Proc. Roy. Soc. B (in the Press).
DAVSON, H. (1942). J. Cell. Comp. PhysioL 20, 325.
GOLDACRE, R. J. (1952). Int. Rev. Cytol. i, 135.
HOLLAND, W. C. & GRIEG, M. E. (1950). Arch. Biochem. 26, 151.
JACOBS, M. H. (1952). In Modern Trends in Physiology. New York. Academic
Press.
JACOBS, M. H. & CORSON, A. (1934). Biol. Bull., Woods Hole, 67, 325.
LANSING, A. I. & ROSENTHAL, T. B. (1952). J. Cell. Comp. PhysioL 40, 337.
LEFEVRE, P. G. (1948). J. Gen. PhysioL 31, 505.
NAKAZAWA, F. (1922). Jap. J. Exp. Med. 3, 288.
NEUBERG, C. & ROBERTS, I. S. (1949). Arch. Biochem. 20, 185.
SJOSTRAND, F. S. (1953). Nature, Lond., 171, 30.
516 MOLECULAR ASPECTS OF ACTIVE TRANSPORT
SJOSTRAND, F. S. & RHODIN, J. (1953). Exp. Cell Res. 4, 426.
SUSSMAN, S, (1946). Industr. Engng Chem. 38, 1228.
THOMAS, G. C. & DAVIES, C. W. (1947). Nature, Land., 159, 372.
UNDERWOOD, G. E. & DEATHERAGE, F. E. (1952). Science, 115, 95.
WILBRANDT, W. & LASzx, L. (1933). Biochem. Z. 259, 398,
WILBRANDT, W. & ROSENBERG, T. (1950), Helv. physiol. acta, 8, C82.
WILBRANDT, W. & ROSENBERG, T. (1952). Int. Rev. Cytol. i, 65.