4 +

CYTOCHEMISTRY

A Critical Approach

A VOLUME IN THE WILEY
BIOLOGICAL RESEARCH SERIES

This Series includes a number of small volumes
covering various topics in biology. The purpose is
to provide other scientists with an authoritative ori-
entation in the author's work as quickly as possible.
To provide this orientation, each author will be asked
to stress his own work and his own viewpoints without
loss of proper perspective of the field as a whole.

ADVISORY BOARD

Paul A. Weiss, PJi.D.
Professor of Zoology, University of Chicago

David R. Goddard, Ph.D.
Professor of Botany, University of Pennsylvania

Francis 0. Schmitt, Ph.D.

Professor and Head, Department of Biology,

The Massachusetts Institute of Technology

Previously Published

Anderson • Introgressive Hybridization
Lwoff • Problems of Morphogenesis in Ciliates

J> 4-'

CYTOCHEMISTRY

A Critical Approach

J. F. DANIELLI

Professor of Zoology

King's College

London, W.C. 2

JOHN WILEY & SONS, INC., NEW YORK
CHAPMAN & HALL, LIMITED, LONDON

Copyright, 1953

BY

John Wiley & Sons, Inc.

All Rights Reserved

This book or any part thereof must no t
be reproduced in any form without
the written permission of the publisher.

Library of Congress Catalog Card Number: 52-12420

PRINTED IN THE UNITED STATES OF AMERICA

To the Memory of
Sir Jack Drummond

A%o00S'

z ! LlBRAf?y
ACKNOWLEDGMENTS \$ 3

This book contains a slightly enlarged version of the material-
presented in a series of lectures given in the Department of
Zoology of the University of Chicago, in the spring of 1949, dur-
ing the tenure of a Rockefeller Fellowship. The main work of
preparing the manuscript was carried out at the Marine Biologi-
cal Laboratory, Woods Hole, during the tenure of a Lalor Fellow-
ship.

It is a great pleasure to me to acknowledge my debt to Pro-
fessor Paul Weiss, who invited me to Chicago; to Professor
W. L. Doyle, who made the arrangements for my lectures; to
Professor P. B. Armstrong and his colleagues, who provided
such excellent facilities at Woods Hole. To the award of
Rockefeller and Lalor Fellowships I largely owe this opportunity
to crystallise the work on cytochemistry which has increasingly
occupied my attention since 1942. A number of immediate col-
leagues have worked together as a team, of whom I am particu-
larly indebted to Dr. I. J. Lorch and Dr. L. G. Bell. From
numerous other friends I have benefited by help and advice, in-
cluding particularly Dr. Honor B. Fell, Dr. E. Kodicek, Profes-
sor D. Keilin, Dr. H. Holter, Professor T. Caspersson, Professor
B. Commoner, Dr. A. E. Mirsky, and Dr. H. Jacoby. Facilities
for carrying out the researches reported here were provided Dy
Professor J. Gray at Cambridge, by Dr. F. S. Russell at the
Marine Biological Laboratory, Plymouth, and by Professor A.
Haddow at the Chester Beatty Research Institute, London. Dur-
ing the past three years spent .at King's College my group has
benefited by grants from the Nuffield Foundation, the British
Empire Cancer Campaign, and the Rockefeller Foundation.

CONTENTS

Chapter 1. Introduction 1

Chapter 2. Fixation Procedures 16

Chapter 3. Studies on Alkaline Phosphatase . . 28

Chapter 4. The Critical Study of the Cytochemistry of

Aldehydes 78

Chapter 5. The Cytochemistry of Proteins and Nucleic

Acids 96

Chapter 6. Quantitative Studies in Cytochemistry . 126

Chapter 7. The Future Outlook in Cytochemistry . 134

i

67729

CHAPTER I

INTRODUCTION

Cytochemistry is an almost undeveloped branch of biology.
It is only comparatively recently that a vigorous attempt has
been made to solve the problems which are involved in cytochem-
ical studies. Consequently, anything in the nature of a textbook
on cytochemistry is premature. There is an insufficient body of
agreed facts for it to be possible to write a textbook which would
set forth a body of methods and knowledge which would be
agreed to without question by the great majority of workers in
this field. I wish therefore to make it clear that this book is not
intended in any way as a textbook. It is very largely a record
of experiments which I have carried out myself, or which have
been carried out by other research workers with whom I have
been closely associated. Any of the individual methods given
here may well be superseded in a few years. I do, however, hope
that a good deal of what is recorded here may be regarded as in
the nature of a blueprint for future developments in cytochem-
istry. The general endeavour which is contained in this work,
namely, the development of rigorous methods, is a key matter
without which cytochemistry is a futile study.

Cytochemistry is pre-eminently a field which calls for a team
of workers. The reason for this is that it demands a high stand-
ard of knowledge in each of the fields of biology, chemistry, and
physics. Unfortunately, many of the methods which have been
developed so far appear at first glance to be so simple that the
necessity for rigorous treatment has largely tended to be ignored.
There have, of course, been a number of outstanding instances of
compliance with the experimental criteria which are required:
this is obvious to all who know of the work of Feulgen, of Cas-
persson, or of Commoner. But the general tendency has been
to suppose that anyone who can cut a tissue section and make up
a standard solution is competent to carry out cytochemical in-
vestigations. Nothing could be further from the truth. Indeed,

1

2 INTRODUCTION

it may almost be said that a good training in histology is one
of the worst backgrounds possible for cytochemical work. In
the type of histology and cytology which is based on showing op-
timal fixation, optimal staining, and optimal methods of treat-
ment with stains, a tremendous amount depends upon the judg-
ment of the investigator as to what he wishes to stain and just
how he wishes to see it. In some sense, this type of histology
and cytology is an art. On the other hand, the object of the
cytochemist is to use fixatives and cytochemical methods in
such a way that a specific and exact treatment is applied to a
tissue, in a manner which is predetermined by the physical
properties of the specimen and a precise programme of chemi-
cal treatment. During the carrying out of this programme, the
judgment of the investigator must be suspended in the interest
of maintaining the precision of treatment. It is only after the
full processes of physical and chemical treatment have been
carried through that the investigator can allow his opinion to
operate. I feel obliged to add that the standards of cleanliness
of reagents which are frequently adequate for classical histology
are, in my experience, frequently inadequate for cytochemical
purposes. So much is this so that, whenever I hear that an in-
vestigator has failed to obtain success with a cytochemical
method, my first piece of advice is always that he should throw
away every reagent that he has been using and make up a new
set of reagents; it is surprising how frequently this elementary
step is successful.

In the second half of the nineteenth century there was a con-
siderable wave of interest in cytology and histology; this was
enabled to come to fruition by the discovery of synthetic dyes
which could be used for staining fixed material. It is quite clear
from the literature of this period that the investigators were
equally concerned to obtain information about both the physio-
logical and the chemical organisation of protoplasm. Their
success, however, was very largely limited to the morphological
and physiological side: there was an almost complete failure
on the chemical level. Amongst the reasons for this was the
fact that affinity for a given stain is only to a limited degree
determined by the detailed chemical constitution of the material
being stained. It is much more markedly determined by the
physical properties of the material. Difficulties were also en-

INTRODUCTION 3

countered with artefacts due to fixation. This matter was
brought to a head by W. B. Hardy and A. Fischer in two publi-
cations in 1899. These two investigators had a much clearer
understanding of the physical basis of cytological techniques
than had most of the previous workers in this field. They
showed that the details of the procedure of fixation, of the
nature of the fixative, and the technique of staining, have just as
profound an effect upon the final picture as has the initial
chemical composition of the material being stained. They also
showed that many of the fixatives in common use were giving
rise to structures which, in fact, did not occur in the living
cells, but were precipitation artefacts. At this time, neither the
chemistry nor the physics of the systems involved in cytology
were sufficiently understood for it to be possible to cope with
the problems demonstrated by Hardy and Fischer. Conse-
quently, by 1910 the wave of interest in the study of fixed
preparations had very largely lost its momentum so far as pio-
neering investigations were concerned. The pioneering investi-
gators instead turned to working on living cells almost exclu-
sively. There was, in consequence, a very rapid development
of experimental cytology.

Between 1910 and 1935 remarkable progress was made in the
field of experimental cytology. Some phenomena were so suc-
cessfully analyzed that the chief physico-chemical factors in-
volved could be detailed, and given an approximate mathemati-
cal treatment. However, the result of this successful investiga-
tion was that many of the working hypotheses formulated by
the experimental cytologists postulated specific cytochemical
organisations in particular parts of cells. Such postulates were
found in many widely diverse fields, e.g., ciliary and amoeboid
movement, muscular contraction, secretion, and the action of
genes. The postulated cytochemical organisations lay so fun-
damentally at the heart of the mechanisms proposed that further
progress was bound to become increasingly limited in every
field, unless further advances were possible in cytochemistry.
It is not surprising, therefore, to find that from 1935 onwards
there was a renewed wave of interest in cytochemistry. Land-
marks in this new wave of interest were the publication of
Lison's book "Histochimie Animale," and the papers of Feulgen,
Caspersson, Gomori and Takamatsu, Linderstr0m-Lang, and

4 INTRODUCTION

Holter, who blazed new trails into the wilderness of cellular
chemistry. It has, however, been somewhat unfortunate that
many of the lessons which can be learned from the work of
Hardy and Fischer and others have been overlooked in more
recent studies.

In the development of rigorous methods of investigation in
cytochemistry, it is necessary to look at each problem from
three points of view: as a chemist, as a physicist, and as a bi-
ologist. From the point of view of the physicist, the main
problems arise from diffusion and adsorption artefacts, from
estimating errors which may arise from the state of aggregation
of the substance which is being studied, from the degree of
molecular orientation of the substance, and from the scattering
of light within a specimen. No qualitative study can be re-
garded as satisfactory which does not involve the elimination
of diffusion and adsorption artefacts. Correspondingly, no quan-
titative study can be regarded as satisfactory unless the de-
gree of aggregation of molecules in the specimen, the orientation
of the molecules in the specimen, and the scattering of light
within the specimen are properly taken into consideration.

From the chemist's point of view the main problems are to
use methods which are of sufficient specificity and which shall
be quantitative and accurate to a known extent. A further prac-
tical problem which frequently arises is to find methods which
will be to a sufficient degree inert, i.e., which do not damage
the specimen. The problem of chemical specificity is a particu-
larly difficult one. The reason for this is that chemical reac-
tions are not carried out by molecules as a whole. They are
usually carried out by a very small number of atoms in a mole-
cule. They are often affected to a marked degree by a rather
larger number of neighbouring atoms; but the greater part of
the molecule may well have no effect on whether a particular
reaction occurs or not. Thus the specificity of a chemical re-
action is limited to supplying information as to whether a par-
ticular chemical group is present in a particular part of a cell.
Information as to the nature of some of the other neighbouring
groups in the same molecule may sometimes be obtained by
studying the rate at which a chemical reaction proceeds. But
it is impossible to identify the whole of any molecules, other
than the simplest, by carrying out chemical reactions. Thus, in

INTRODUCTION 5

the well-known Feulgen reaction which is commonly said to
give localization of deoxyribonucleic acid, we have in fact a
method which probably indicates a linkage between deoxy sugar
and any other group which one can split off from the glycoside
linkage at roughly the same rate as are the purines. Thus the
Feulgen reaction, in fact, gives us information which is limited
to telling us that deoxy sugar is present in the specimen and
that it is in glycosidic linkage with a substance which can be
split away by acid hydrolysis at the same rate as are certain
purines. The Feulgen reaction does not tell us whether the
substance which is split off is a purine, nor does it tell us
whether the sugar is linked through phosphate bonds to other
similar units so as to complete a nucleic acid. In a similar way,
when the ultraviolet spectrum of a specimen is studied, one
can readily detect the presence of a substance absorbing in the
same region of the spectrum as the purines and pyrimidines.
There are, of course, other substances which absorb in this
region. A notable example, recently reported by Chayen (1952),
is ascorbic acid. Thus, in a material in which ascorbic acid
may occur, it is necessary to take steps to differentiate between
ascorbic acid and purine groups. When all the necessary elimi-
nation of this type has been done, we can perhaps be certain
that in a particular part of a specimen there is purine or
pyrimidine. But it has not so far proved possible from spectro-
photometric studies to determine whether the purine or pyrimi-
dine is present as part of a nucleic acid molecule, or whether it
is present in some other form, e.g., linked directly to a protein.
It is notable in this connection that Panijel has recently found
a protein in Ascaris sperm which has the same absorption spec-
trum as nucleic acid, containing a considerable proportion of
purine, but no phosphorus.

The problems which arise from the biologist's point of view
are somewhat different in nature and are less readily defined.
They must, however, involve a constant awareness of the fact
that an animal, and a cell, cannot be dissociated from its en-
vironment. Due respect must be paid to such principles as those
of homology and analogy, and at the same time it is necessary
to maintain a more rigid guard against the acceptance of generali-
sations than is usually necessary in the fields of physics and
chemistry, owing to the fact that variables on the purely biologi-

6 INTRODUCTION

cal side are far more easily underestimated than otherwise. As
examples of some of the problems which affect one from the bi-
ological point of view, may be mentioned fixation and diet. It
is doubtful whether any method of fixation other than freeze-
drying is really adequate for cytochemical purposes. Then,
as far as diet is concerned, it will be a source of amazement
to future generations to discover in how few of the papers on
cytochemistry the diet of the animals used is at all defined.
Yet the cytochemical pattern in organs such as the liver and
intestine is astonishingly dependent upon diet. As another ex-
ample, may be mentioned the work of Dr. H. Mugard (1953),
who has recently shown that the cytoplasm of the ciliate
Ophryoglena atra which has not recently fed is apparently free
from the enzyme alkaline phosphatase: yet within a few seconds
of the formation of a food vacuole a high concentration of phos-
phatase is present in the cytoplasm. For particular studies it
will no doubt be equally important to define the age of the
animal under investigation, the time of taking the specimen,
the state of hormone activity, and other biological variables,
before a generalisation on the cytochemical level may safely
be embarked upon.

Before passing on to consider the fields in which I have played
some part in developing techniques, a number of techniques
will be considered critically, since by so doing it is possible
to see many of the major hazards which exist in the field of
cytochemistry, and which it has been my endeavour to avoid.

Centrifugal Stratification, etc.

In these techniques the common principle is, by centrifuga-
tion or other methods, to separate parts of cells which differ
from one another in their physical characteristics. The most
usual approach is to stratify a cell by centrifugation, to separate
the strata, and then examine the distribution of various sub-
stances in the different strata. In the hands of Holter and of
Shapiro, this work has provided valuable information about
the distribution of certain substances, particularly enzymes, in
large cells such as amoebae and echinoderm eggs. There are,
however, some very marked limitations to techniques of this
type. Their use is usually limited to large individual cells:

MACERATION PROCEDURES 7

cells in tissues cannot be studied readily. Then the methods of
manipulation are not devoid of action upon cell fractions which
are under study. For example, Shapiro found that when sea
urchin eggs are centrifuged, first so as to cause stratification
and then so as to cause the cell to divide into two halves — one
light and one heavy — the respiration of the two fragments so
formed is greater than that of the initial intact egg. It there-
fore follows that in an experiment of this type either there has
been a change in the physico-chemical organisation of the
enzyme systems of the cell, or else there has been synthetic ac-
tivity on the part of the cell, resulting in the formation of more
respiratory enzymes. Whichever of these changes may have
taken place, it is clear that the final condition, as revealed by
the cell fragments, cannot be taken as a close guide to what
was happening in the intact egg. In this connection, Holter re-
marks, "The only conclusions to be drawn from such experi-
ments are those based on the distributions of substances, • • •
[which] • • • permit, of course, only indirect conclusions with
regard to physiological activity." In any experiments involv-
ing the destruction of the known relationships between cellular
entities, as is the case in stratification, any deduction is dubious
unless it is established that the procedure does not lead to syn-
thesis or to destruction of chemical components. It is obligatory,
in such experiments to establish the lack of such synthesis or
degradation.

Maceration Procedures

Amongst the most common techniques employed today, par-
ticularly by biochemists seeking to make a contribution to
cytology, are techniques involving the disintegration of cells
into fragments, and fractionation of the fragments so formed.
The debris formed by maceration is commonly centrifuged at
various speeds so as to isolate fragments with different sedi-
mentation rates. It is hoped that methods of this type will
isolate granules, mitochondria, nuclei, and chromosomes, in a
condition which is closely similar to, if not identical with, the
state of those bodies in the intact cells. It would undoubtedly
be of the greatest value if it were true that cell organs could
be isolated in this way. But so far there has been an almost

8 INTRODUCTION

complete lack of proof that the bodies isolated are in the same
condition as in the intact cells.

It would, indeed, be very surprising if there were not many
and dramatic changes in the organisation and composition of
both nuclear and cytoplasmic bodies as the result of maceration
and addition of the various solutions which are used in fraction-
ation. The nucleus and the cytoplasm are both very complex
colloidal systems. Studies by Chambers and others on cell
nuclei have shown that when the nucleus is removed from the
cell by microdissection it commonly either sets into a gel or
dissolves: in either case a profound change occurs as soon as
the nucleus is removed from its normal environment. De Fon-
brune, and Lorch and Danielli have found that, although a
nucleus may readily be transferred from one cell to another
in a viable condition provided that it does not come into con-
tact with the environment of the cell, a few seconds' contact
with the environment is sufficient to destroy the viability of the
nucleus. Dr. Dounce has informed me that he has compared
the composition of nuclei isolated from macerated cells by cen-
trifugation in non-aqueous solvents with nuclei isolated by
centrifugation in aqueous solvents. The nuclei from the non-
aqueous solvents contain almost twice as much material as do
those from the aqueous solvents. From these few remarks it
is quite clear that one should anticipate profound changes in
the organisation of the nucleus when it is removed from its
normal environment, and that these changes must include dif-
fusion of substances out of the nucleus, and probably also dif-
fusion of substances into the nucleus. It should be the first
responsibility of the investigator to ascertain the extent to
which morphological changes and diffusion artefacts are in-
volved in these isolation techniques. It seems probable that at
the present time the only work of this type, i.e., on isolated
nuclei, which is reliable is that of Brachet and of Callan on
some of the properties of the whole nuclei of amphibian oocytes,
which were isolated by microdissection techniques.

It is probable that the organisation of the cytoplasmic com-
ponents is just as labile as that of the nucleus, if not more so.
Often in relatively simple protein systems, such as blood
plasma, addition of foreign materials causes profound reor-

MACERATION PROCEDURES 9

ganisation. For example, addition of salts may vary the pro-
portion of albumin and globulin found on electrophoresis of
plasma. Somewhat similar effects are found on simply adding
water to plasma, and a not immoderate addition of water causes
precipitation of part of the globulin. It is difficult to see how
experiments on the isolation of granules and other small cell
organs can be taken at their face value unless it is rigorously
demonstrated that their composition is unchanged by the isola-
tion procedure.

Often, if the composition of individual granules is identical or
closely similar to the granules in the living cells, there still re-
main many difficulties in interpretation of experiments on the
mass isolation of granules — mitochondria, nuclei, etc. The rea-
son for this is that it is practically impossible, except in a small
minority of tissues, to obtain a homogeneous concentration of
cells. In the first place, almost all tissues contain, in addition
to the typical cellular component, cells of the vascular system,
connective tissue cells, and white and red blood cells. More-
over, often the cells of the same type, e.g., hepatic cells, are
not of identical chemical composition in closely adjacent parts
of an organ. Consequently, it does not follow that a biochemi-
cal pattern observed in a particular group of granules is identi-
cal with, or even similar to, that of any individual cell type.
This can be very simply demonstrated in the case of hepatic
cells. Long-chain aldehyde (Feulgen's plasmal), ribonucleic
acid, and alkaline phosphatase will occur in the same granule
fraction obtained from liver. It would, therefore, be natural to
suppose that these three components are bound together in the
same granule and may even cooperate in carrying out certain
biochemical functions. When, however, the distribution of these
three components is studied in tissue sections, it is found that
each of them is present to a significant concentration only in
some of the hepatic cells. Some hepatic cells may be practically
free of any one of these components, some may contain one
component, some, two components, and some, three components
in significant quantities. If, therefore, all three substances
are present on a granule of the same size in the living cells,
then it appears that some of the granules contain none of the
substances, some, one of the substances, some, two, and some,

10 INTRODUCTION

three. It is thus quite misleading to suppose that all the
granules must have all three of these substances, as has been
deduced from the procedure of isolation of granules.

It should also be emphasized here that the phenomenon of
lack of biological identity of cells, even of the same cell type,
illustrates a general weakness of deductions drawn from bio-
chemical studies of extracts. As a general rule, no tissue can be
regarded as biologically homogeneous, and deductions made on
the biological level after disintegration of tissues must normally
be treated with considerable reserve. Only in the case of rela-
tively homogeneous cell preparations, such as those of yeast,
bacteria, ova and sperm, can such studies be regarded as re-
ferring to a single chemically homogeneous cell population.

In maceration experiments, it is also inevitable that activa-
tion and inactivation of enzymes will occur to a degree which
may often present formidable difficulties. How this problem
can be coped with is not at all clear. There is also the difficulty
described by Marjorie Stevenson, that one is tempted to sup-
pose that all enzymes found in a cell must have a function. Dr.
Stevenson was inclined to think this must be the case until she
found an enzyme in bacteria which could act upon chlorate. She
could not believe that this activity could be of any value to the
cell since chlorate never appears in its normal environment.
One hesitates to agree entirely with this argument: it may well
be that the enzyme which acts upon chlorate also acts upon
some other substrate which is normally found in the environ-
ment, or which from time to time occurs in the environment.
And another difficulty in the interpretation of enzyme activities
is to know whether any individual enzyme is functioning in con-
nection with a particular physiological activity which is under
observation. Indeed, it does not necessarily follow that all the
enzymes found in a cell are necessarily functioning at the same
time. It seems quite possible that some enzymes may be pres-
ent in cells at all times, but have a function to fulfill only at
exceptional periods in the life of the cell.

When all these experimental hazards and difficulties of inter-
pretation are considered, it must be clear that it will be a long
time yet before the results of maceration procedures can be
either evaluated or interpreted.

SILVER NITRATE TEST FOR ASCORBIC ACID 11

The Silver Nitrate-Acetic Acid Test for Ascorbic Acid

When tissue is treated by silver nitrate dissolved in acetic
acid, it is commonly found that deposits of silver are formed
due to reduction of the silver nitrate. It has been claimed that
reduction of silver under these conditions is a specific test for
ascorbic acid and also that the localization of the silver de-
posits is a close guide to the localization of ascorbic acid. How-
ever, owing to the diffusion factors involved, these conclusions
can hardly be true; indeed this system is an excellent example
of the complications which are bound to ensue when such chemi-
cal studies are attempted on small molecules. Molecules of low
molecular weight, such as ascorbic acid, silver nitrate, and
probably also the initial reaction product of ascorbic acid with
silver nitrate, are highly diffusible. It is thus easy to say much
in theory, but it is impossible to prove the precise localization
of ascorbic acid by this method. As the silver nitrate-acetic
acid mixture diffuses into cells a mixing zone will be established
somewhere close to the cell wall in which ascorbic acid and
silver nitrate will react. As the reaction proceeds fresh silver
nitrate and fresh ascorbic acid will be recruited into the mixing
zone until the ascorbic acid supply is exhausted. This inter-
action in the mixing zone is inevitable with two such highly
diffusible substances. Consequently, if . ascorbic acid exists in
cells as such, it can only be demonstrated in the mixing zone,
which is an artefact of fixation. In fact, the silver which is
formed by reduction in this procedure is not found in the mix-
ing zone but attached to mitochondria, etc. It is thus clear
that either one of three things must be true. The reaction may
be truly occurring at the surface of the mitochondria; but if
this is true, it is not with free ascorbic acid that the reaction
is occurring but a substance bound to the mitochondria so as to
render it non-diffusible. An alternative explanation is that the
technique does involve a reaction of silver with free ascorbic
acid but that the results of the reaction appear adsorbed on
surfaces which have a high affinity for the reduced silver and
which bear no relationship to the distribution of ascorbic acid
in the living cell. A second alternative is that diffusible ascor-
bic acid is released from the interior of the mitochondria by fix-

12 INTRODUCTION

ation, and precipitates silver in a mixing zone at the surface
of the mitochondria.

Techniques for Cytochemical Demonstration of Enzymes
These techniques usually involve carrying out an enzyme
reaction with the intrinsic enzymes of a tissue section or smear,
etc., under such conditions that one of the products of the re-
action is precipitated. Before results of this type can be evalu-
ated, it is necessary to have answers to a number of questions.
These are: (a) How much of the enzyme is destroyed before
and during the cytochemical procedure? If destruction occurs,
does it occur to an equal degree at all sites or does it occur
selectively at certain sites? (b) Is the insoluble reaction prod-
uct precipitated at the actual site of enzyme reaction, or is it
precipitated at sites which have a very high affinity for the
reaction product? (c) Is the enzyme in the specimen used for
cytochemical study in its physiological position in the material,
or has it been redistributed by diffusion processes? Studies
which do not provide answers to these questions must be dis-
carded. It is, no doubt, true that some of the methods now used
to study the distribution of enzymes do, in fact, demonstrate the
correct site of the enzyme ; it is equally true that there are other
instances in which the information provided is greatly mis-
leading.

The Feulgen Technique for Nucleic Acid

In this technique tissues are heated at 60° with normal hydro-
chloric acid for 5 minutes or more, the precise time varying with
the fixative which is used and the nature of the specimen. This
procedure splits off purine from the deoxy sugar nucleic acid,
so that, when the material is subsequently treated with reduced
fuchsin, a violet Schiff's base is formed. The procedure of hy-
drolysis, however, tends also to make nucleic acids diffusible,
probably by reducing the molecular weight, as was suggested by
Stedman and Stedman. That this was likely to be the case has
been implicit in results known for many years. For example,
Bauer showed that, when a fixative like formaldehyde is used,
or acetic alcohol, the intensity of the reaction reaches a peak
after, say, 5 or 10 minutes of hydrolysis. Further hydrolysis

THE USE OF ABSORPTION SPECTRA 13

rather rapidly diminishes the amount of nucleic acid which can
be demonstrated in a section. On the other hand, if a heavy
metal such as mercury is present in the fixative, the hydrolysis
may profitably be more prolonged, with an increase in the
intensity of reaction and no evidence of loss of material even
after 30 minutes' hydrolysis. It seems clear from these obser-
vations of Bauer that the Feulgen hydrolysis, in some way at
least, must be making it possible for nucleic acid to diffuse, and
that it may be more firmly anchored in the section by conver-
sion into the salt of a heavy metal such as mercury before the
hydrolysis is carried out. If the hydrolysis makes it possible
for nucleic acid to diffuse out of the section, then it may also be
possible for it to diffuse into the section. In consequence, there
must remain some slight doubt as to whether the details of the
picture revealed by the Feulgen technique are not to some de-
gree invalidated by diffusion artefacts. It is to be regretted
that no attempt has been made to study carefully the extent
to which diffusion does, in fact, take place during the Feulgen
procedure.

It would be of considerable value if a film could be made, by
means of ultraviolet light, of changes in the distribution of
nucleic acid in cells which are undergoing chemical fixation,
and the results so obtained compared with the results obtained
by freeze-drying, following which a study should be made of
the course of distribution of ultraviolet-absorbing material in
the same specimen, preferably by filming the cells during hy-
drolysis. It seems likely that the difficulties in such a problem
caused by prolonged exposure to ultraviolet light can be mini-
mized by using a newly developed instrument, the television mi-
croscope, which requires much less intensity of light than does
a direct photograph with a photographic film.

The Use of Absorption Spectra

Practically all cytochemical methods involve at one stage or
another the observation of the absorption spectrum of a speci-
men. It is, therefore, necessary to bear in mind certain prob-
lems which constitute limitations in the interpretation of such
studies even when the instrumental problems have been solved.

14 INTRODUCTION

These are:

1. The study of absorption spectra can provide evidence of the presence
of only certain chemical groups in a particular site: this is not in itself
conclusive evidence of the presence of any one given chemical substance.

2. When two groups are shown to exist in the same position in a cell,
it is not normally possible from studies of absorption spectrum to deter-
mine whether these groups occur in the same molecule or in different
molecules.

3. The spectrum of a molecule may be rather markedly affected by the
adjacent or neighbouring molecules. Thus Caspersson states that the ab-
sorption spectrum of nucleic acid is altered to a significant degree by the
combination of nucleic acid with basic proteins. Fox and Danielli have
shown that the absorption maximum of astacine is shifted from the yellow
into the red by adsorption at an oil-water interface. R. A. Peters has shown
that a similar reversible change occurs with astaxanthine ; when astaxan-
thine is combined with certain proteins the color is blue, and when it is
removed from this combination the colour is red. An extreme example of
this type of behaviour is found with haem, the spectrum of which varies
in a very characteristic manner according to whether it is combined with
the appropriate protein which gives, for example, a haemoglobin, a
catalase, a peroxidase, a cytochrome, or a cytochrome oxidase.

4. All quantitative studies on the absorption spectrum of different parts
of the cell need to be corrected for loss of light caused by scattering in
the specimen, and must be corrected, also, for the degree of orientation
of the absorbing molecules and their state of aggregation in the specimen.
Furthermore it must be demonstrated that the Beer-Lambert law is obeyed.

The Use of Enzymes in Cytochemical Reactions

A procedure which is rather commonly used is to allow a
purified enzyme preparation to act upon a specimen. For ex-
ample, Brachet has suggested that stains should be used to
identify the site of ribonucleic acid and that the presence of this
acid should be checked by the use of ribonuclease which should
remove all staining material. Such procedures involving en-
zymes do provide additional circumstantial evidence of the
existence of the components constituting their substrates in par-
ticular parts of cells. However, final proof can never be ob-
tained by this type of technique for the following reasons:

1. It is never possible to establish that a particular enzyme is pure. It
is, of course, possible to show that a particular enzyme such as ribonuclease
is lacking in activity, say, toward certain peptide bonds or certain forms of
linkage of sugar molecules. But it is never feasible to show that a prepara-
tion is lacking in ability to split all the various types of bonds in substances
which might be involved in anchoring a dye. Consequently, there is always

REFERENCES 15

the hazard in the use of enzyme preparations that the loss of a stained sub-
stance from a cell has been produced by some enzyme action other than
that which it was intended should occur.

2. Even if one neglects the hazard of contamination by other enzymes,
there are considerable difficulties in interpreting the removal of sub-
stances from sections by enzyme action. For example, it is possible that a
structure composed mainly of nucleic acid could be protected from the
action of the enzyme ribonuclease by quite small quantities of protein.
Even a monolayer of protein might be sufficient to prevent access of the
enzyme to its substrate. Conversely, a structure consisting almost entirely
of protein, but held together and held in the section by a fine matrix of
ribonucleic acid, might well be removed by ribonuclease although the
nucleic acid constituted an extremely small fraction of the structure as a
whole.

3. Many of the components even in a fixed preparation are not totally
indiffusible and will migrate from the fixed section if a suitable adsorbing
reagent is available. For example, Catcheside and Holmes have shown
that ribonucleic acid may be removed from chromosomes by what ap-
pears to be egg albumin, supposedly an inert protein.

REFERENCES

Bauer. 1933. Zeit. /. mikr.-anat. Forsch., 33, 143.
Brachet. 1940. C. r. soc. biol. Paris, 133, 90.

1944. Embryologie chimique (Masson, Paris).
Callan. 1952. Symp. Soc. Exp. Biol, 6 (in press).

Caspersson. 1950. Cell Growth and Cell Function (Norton, New York).
Catcheside and Holmes. 1947. Symp. Soc. Exp. Biol, 1, 225.
Chayen. 1952. Symp. Soc. Exp. Biol., VI (in press).
Danielli. 1946. Nature, 157, 755.
Dounce. 1948. Personal Communication.

Fetjlgen and Rossenbeck. 1924. Zeit. }. physiol. Chemie, 135, 203.
Fischer. 1899. Fixierung, Farbung und Bau des Protoplasmas (Gustav

Fischer, Jena).
Fox and Danielli. 1941. Biochem. J., 35, 13, 88.
Hardy. 1899. J. Physiol, 24, 158.
Holter. 1946. Nature, 158, 917.

Lison. 1936. Histochimie Animate ^(Gauthier-Villars, Paris).
Mugard. 1953. Quart. J. Micros. Sci. (in press).
Shapiro. 1939. Cold Spring Harbor Symposia, 8, 406.
Stedman and Stedman. 1943. Nature, 152, 267.

CHAPTER 2,

FIXATION PROCEDURES

Almost any tissue which is to be subjected to a cytochemical
procedure must first be fixed. The ideal result of fixation would
be to have an exact replica of the living tissue with each of the
molecules present in the original specimen in its original place,
and with no foreign molecules present. This is clearly an im-
possible ideal to achieve. In general, fixation can succeed only
in maintaining the large molecules, such as proteins, nucleic
acid, and carbohydrates, in approximately their original posi-
tions. In the case of the proteins, a considerable proportion of
the molecules are irreversibly changed. Indeed, if fixation is to
produce a specimen which will be satisfactory f6r any length
of time, the fixing agent used must effect two phenomena: it
must precipitate proteins and it must directly or indirectly
greatly restrict enzyme activity. It is not necessary that the
fixative should denature the proteins. For example, when a
heavy metal is the basis of a fixative, it precipitates the pro-
teins effectively but does not necessarily denature them. Simi-
larly, it is not necessarily the case that an enzyme should be
directly inactivated; it may be sufficient that it should be
rendered indiffusible by being captured by a matrix of another
protein. But if the enzyme systems of a cell are not, by one
method or another, prevented from displaying their normal
activity, the proteolytic and other hydrolytic enzymes rapidly
destroy the integrity of the tissue.

Fixation is commonly carried out by the action of chemical
agents alone. Occasionally freeze-drying is used, but in the
past this has been somewhat uncommon.

A major problem in the choice of fixative is presented by the
fact that, while a considerable amount of change must be
brought about in the specimen to render it insoluble in water
and to prevent enzyme activity, it is, nevertheless, necessary
that the group which is to be studied cytochemically shall not

16

FIXATION PROCEDURES

17

be affected to any serious degree. The choice of fixing agent has
to be made with the careful consideration of its effect upon the
group which is to be studied. It commonly happens, therefore,
that what to the cytologist has been the best type of fixative
is not available to the cytochemist. The common ingredients
of fixatives include strong oxidizing agents, such as osmic acid
and chromic acid ; strong reducing agents, such as formaldehyde
and formic acid; cross-linking substances, such as formaldehyde
and osmic acid; and acids, bases, and heavy metals which can
precipitate such substances as nucleic acids and proteins. In
addition, anhydrous fluids such as alcohol and acetic acid are
quite commonly used; they exercise their effect very largely as
denaturing agents. The anhydrous solvents cause denaturation
mainly by breaking up the lipoid adhesions which assist in
maintaining the native state in many proteins. Table I sets out

TABLE I

The action of various chemical reagents commonly employed as com-
ponents of fixatives upon chemical groups in tissue sections, -f- indicates
that a group is unaltered by a reagent. - indicates that it is unsafe to
use this component if it is desired to keep a particular group intact.

CHOH

NH2

SH

C02H

Histi-

Trypto-

Tyro-

Sugar

Alde-

Nucleic

dine

phane

sine

hyde

Acid

Formalde-

hyde

+

—

?

+

+

+

+

+

+

?

Alcohol

+

+

+

?

+

+

+

+

+

?

Acetic acid

—

—

—

?

?

?

+

?

Mercury

+

+

—

?

+

+

+

+

+

?

Osmic acid

?

?

—

?

?

?

?

Dichromate

—

?

—

+

?

__

?

_

?

Acetone

+

+

+

+

+

+

+

?

+

+

Pyridine

+

+

?

+

+

+

+

+

+

+

Picric acid

+

+

?

?

+

?

+

+

+

+

a list of fixing agents and indicates their action upon various
groups which may be of interest to the cytologist. It will be
seen that there are many of the ingredients of the better chem-
ical fixatives which commonly should be avoided. For example,
if a study is to be made of NH2 groups or of phenols such as
tyrosine, acetic acid in anhydrous solution must be avoided be-
cause of the danger of acetylation of the groups concerned. If

18 FIXATION PROCEDURES

SH groups, or sugars, are to be studied, oxidizing agents such
as osmic acid must be avoided. If amino groups are to be
studied, formaldehyde must be avoided, etc. When enzymes
are of interest, the limitations are even more severe. A sub-
stance like osmic acid must be avoided completely.

During the action of chemical fixatives, artefacts may arise
due to precipitation, diffusion, and adsorption, and also due to
activation and inactivation of enzymes. The endeavour with
regard to precipitation is not to avoid precipitation — indeed
the whole object of fixation is to secure proper precipitation —
but to see that the size of the particles precipitated is less than
the resolving power of the light used. Diffusion and adsorption
artefacts present a source of great difficulty. One method of
detecting these artefacts is to use a variety of fixing agents of
widely different physical properties. For example, if the same
structure or distribution of a substance is observed with fixatives
which are acid, neutral, basic, and anhydrous, it is rather un-
likely that the observed cytochemical pattern is an artefact.
On the other hand, the use of a wide variety of fixatives is not
an absolutely safe guarantee against the formation of artefacts.
Moreover, it frequently happens, especially when enzymes are
being studied, that the range of fixing agents which can be used
without destruction of enzymes is too restricted to permit this
method of detection of artefacts.

When the cytochemical method is applied to the analysis of
a physiological activity, e.g., enzyme action, it is very difficult
to assign a precise significance to the degree of activity which
is observed in the fixed preparation. If one knows the extent
of enzyme activity at a particular part of the preparation and
can be sure this has not been modified by the fixation procedure,
it is still not possible to estimate the degree to which this enzyme
is active under physiological conditions. Before that degree of
activity can be determined, it is necessary to have a precise
knowledge of the physio-chemical environment of the enzyme
under physiological conditions. This means knowing what other
colloids are present in the immediate environment of the enzyme,
since enzyme activation is greatly affected by the presence of
other colloidal molecules or surfaces on which the enzyme may
be adsorbed. Then it is necessary to know the hydrogen-ion
concentration, the ionic strength, the substrate concentration,

FIXATION PROCEDURES 19

the reaction-product concentration and the concentrations of
activators and inhibitors in the immediate vicinity of the
enzyme. With the exception of one or two substances, such as
cytochrome oxidase and muscle hemoglobin, it must be ad-
mitted that we are not as yet in a position to determine the exact
degree of activity of any enzyme in a particular site in the cell.

The use of chemical substances to effect fixation has been
disadvantageous in cytochemical work, from a purely chemical
point of view. It also has many disadvantages in that it is ex-
tremely difficult to detect artefacts due to diffusion and adsorp-
tion. In view of this, there can be no doubt but that the
method of choice is freeze-drying. However, it must be empha-
sized that, unless freeze-drying is carried out with rigid atten-
tion to detail, the artefacts which may arise can be even more
serious than those found with chemical fixation. The method
of freeze-drying was introduced by Altmann in 1890, and
greatly developed by Gersh. The principle involved is to so-
lidify the material by plunging a small piece into a fluid at a
very low temperature. The material is then placed in a
chamber which is evacuated, and the water is then pumped off
while the specimen is kept frozen solid. When all the water has
been removed, the specimen may be embedded in wax and sub-
jected to sectioning, etc.

The apparatus which has been used in the past for this pur-
pose has suffered from a number of deficiencies. It has been
delicate, containing a great deal of glass tubing; it has not been
easily possible to process considerable quantities of material;
mechanical refrigeration has considerably restricted the range
of drying temperatures which can be used and also is frequently
unreliable; and, not least, must be mentioned the fact that such
apparatus is expensive to build or buy. Mr. L. G. Bell and I
have therefore developed, in collaboration with W. Edwards
and Co., a freeze-drying apparatus which is sufficiently robust
and inexpensive to be suitable for routine work in almost any
laboratory. Figure 1 illustrates the principle of the drying unit.
The procedure is to fix the specimen by plunging it into isopen-
tane which has been cooled in liquid nitrogen. It is desirable
to avoid liquid air as a cooling agent, since, if for any reason
the isopentane and liquid air become mixed, a violent explosion
will ensue. It is necessary to use intermediate cooling agents

20

FIXATION PROCEDURES

such as isopentane, rather than plunge the specimen directly
into liquid air or liquid nitrogen, since, if the specimen is di-
rectly in contact with liquid air or nitrogen, gas bubbles form

-Vacuum flask

■Bath at - 78°C

Water-trapping
surface

Speciman at
controlled temperature

P,0,

Fig. 1. Illustrating principle of new-t}^pe freeze-dryer. Former freeze-
dryers pumped the water off into a P2O5 trap. In the present type the
water is trapped on the walls of the evacuated tube, which are maintained
at —78° C. until the specimen is diy. Then on warming the tube the
water is returned by the P2O5 at the tube base. During the drying process
the specimen is warmed by a thermostatically controlled heater to a de-
fined temperature, usually —40° C. When the vacuum is maintained by
a two-stage rotary pump, drying requires about 2 days. If a diffusion
pump is used, backed by a rotary pump, the drying time can be reduced
to 6-10 hours for a specimen not more than 1 mm. thick. But there is
usually no advantage in reducing the drying time to this extent.

around the specimen, due to vaporization of the liquid. As a
consequence, the specimen is insulated from the liquid air and
cools very much more slowly than when good thermal contact
is maintained between the cooling agent and the specimen.
Isopentane freezes at approximately —160°. When an optimal

FIXATION PROCEDURES 21

result is required, it is sometimes desirable to achieve a some-
what lower temperature. This can be done by using a eutectic
mixture of propane and isopentane, or propane and butane, in-
stead of isopentane.

The drying chamber which we have developed has an easily
demountable head, attached to which are the racks which carry
the specimens. The procedure which we use is to surround the
drying chamber with an acetone-solid CO2 mixture which main-
tains the drying unit at approximately —78° C. The demount-
able head is placed with the specimen racks in liquid nitrogen.
When the specimen has been frozen in the cooled pentane, the
racks are lifted out of the liquid nitrogen, the specimen placed
on the racks, and the unit placed in the drying chamber and
evacuated. From the time the specimen is lifted out of the
pentane bath to the time when it is placed in the drying chamber
at —78°, only a few seconds elapse.

The general procedure which we use is then to dry the
material for two or three days. The walls of the drying chamber
are maintained at —78°, and the drying chamber is arranged
so that the racks containing the specimens are electrically heated
to a defined temperature, variable between —60° C. and —10° C.
as required. Under these conditions the water vapor from the
specimen sublimes from the specimen to the walls of the drying
tube. This involves a very short vapor. path. Consequently,
the pumping unit has no load placed upon it. It is not necessary
for the water vapor to be pumped over into a water trap as has
commonly been the case in previous drying units. The pump
is used merely to maintain the basic low pressure, and the
water vapor sublimes from the specimen to the sides of the dry-
ing chamber. When the specimen has been sufficiently dried,
the temperature of the drying chamber is allowed to rise to
room temperature. When this occurs, the water vapor rapidly
sublimes into a phosphorous pentoxide trap. If desired, this
latter stage can be avoided by having a phosphorous pentoxide
component in the drying chamber itself. But we have encoun-
tered no difficulty with this procedure of subliming the water
from the specimen to a more highly cooled surface a short
distance away, followed by a rapid distillation of this water to
a trap, carried out at room temperature when the specimen
is already dried. The apparatus has been constructed so that

22 FIXATION PROCEDURES

it consists of metal tubing, and, as the water vapor is not be-
ing pumped into a trap at the low-temperature stage, it has
not been necessary to incorporate a diffusion pump, since a
two-stage rotary pump can do all that is required. The speci-
men temperature is defined by a thermostatic device, and the
temperature may be checked by a thermocouple which is asso-
ciated with each thermostat. As the apparatus does not in-
volve mechanical refrigeration, it can run indefinitely without
overhauling. Since there is no load thrown on the pump, it is
possible to have a number of drying tubes separately linked
to the same pump. We are at present using a system of four
separate drying tubes, each of which contains two or three trays
of specimens. Each drying tube can be opened and shut inde-
pendently of the other tubes, so that material can be processed
in considerable quantity.

Provided that the specimen is not to go into an aqueous solu-
tion or any other highly polar solvent such as glycerol or formic
acid, the specimen may be embedded, sectioned, and examined
without any further treatment. We customarily allow the
specimen to come up to room temperature in the drying chamber,
then place it in an infiltration unit containing degassed wax.
The apparatus is evacuated and the wax melted and run onto
the specimen, which may then be embedded and sectioned by
the usual procedure. Provided that the specimen is allowed to
come to room temperature, we have not encountered any diffi-
culty with condensation of water vapor from the air as the
dried specimen is transferred from the drying unit to the infil-
tration unit. It may be necessary under conditions of high
humidity to warm the specimen to above room temperature be-
fore transfer. If so, the specimen can readily be warmed to a
sufficient degree by the heating unit of the drying chamber.
Alternatively, the drying tray may be replaced by a bath con-
taining a little solidified degassed wax. When the drying is
completed, the temperature of the wax can then be raised
slightly above its melting point by the heating unit of the dry-
ing chamber; infiltration of wax then occurs without moving
the specimen.

If the material to be studied must be subjected to aqueous
or other highly polar solvents, a fixative procedure must be
used after freeze-drying. The reason for this is that freeze-dry-

FIXATION PROCEDURES 23

ing preserves most of the proteins and other substances present
in the cell in such a normal condition that they readily dissolve
in many polar substances. A fixation procedure may be ap-
plied after drying and before embedding, or it may be applied
to individual sections after embedding in wax. We are at
present experimenting with the possibility of fixatives which
can be used as vapors to act directly upon the specimen after
it has been dried. Gases such as formaldehyde, carbon sub-
oxide, hydrochloric acid, and osmic acid have potentialities for
such a procedure. Most of our work, however, has involved
material which has been embedded and sectioned after freeze-
drying. The sections are then flattened either by using a non-
aqueous solvent such as alcohol, or by flattening the section on
mercury. Then, after the wax has been removed from the sec-
tion, the latter can be subjected to the fixing action of alcohol
and other anhydrous solutions which may contain acids, bases,
heavy metals, etc., as desired. As a result of such procedures,
the specimens become sufficiently insoluble in polar solvents.

It is of interest to calculate the time relationships in freeze-
drying and in fixation by chemical agents. Studies of cell per-
meability show that when cells 2 microns in thickness are placed
in a solution of a solvent such as ethyl alcohol, the concentra-
tion of alcohol inside the cell reaches about 50 percent of that
of the outside in about 10 seconds. Thus, if individual cells
of this size were brought into direct contact with alcohol, the
concentration of alcohol would reach about 50 percent in about
10 seconds and fixation would become effective in about 10-20
seconds. But most of the material with which we deal is in
tissue sections which are much thicker than 2 microns. For a
thin slice of tissue 2 millimeters in diameter placed in alcohol,
the time required for fixation to become effective would be ap-
proximately yiOOO times longer, i.e., on the average with this
material it would require roughly 330 seconds or more before
fixation becomes really effective with a chemical fixative. When
more slowly penetrating materials, such as heavy metals, are
used as fixatives, the time for effective fixation may be con-
siderably increased.

The time taken for effective fixation to be accomplished by
the technique described above, using isopentane, can be roughly

24 FIXATION PROCEDURES

ascertained by plunging a specimen containing a thermocouple
into the pentane. It is found in this way that the temperature
in the centre of a 2-millimeter section of tissue falls below
— 50° in less than 10 seconds and falls below —170° in less
than 30 seconds. It is thus probable that a 2-millimeter piece
of material is effectively solidified in two or three seconds, com-
pared with a minimum of 330 seconds which are necessary when
a rapidly penetrating chemical fixative is used. Thus freezing
in a fluid at liquid air temperature is immensely more rapid
than any chemical agent can possibly be. If the water in the
specimen remained liquid for some time after cooling in pentane,
diffusion would nevertheless be very greatly reduced as a result
of the great increase in viscosity caused by lowering the tem-
perature. After 3 seconds the viscosity of the water in the
specimen would be more than 100 times greater than that of the
specimen at room temperature, and, after 10 seconds in the bath,
the viscosity would be of the order of 106 times that in the same
specimen at room temperature. In fact, the resistance to diffu-
sion is probably much greater than this, owing to the formation
of ice in the specimen.

However, even with this very great speed of fixation, it is
possible that diffusion artefacts may occur when the distribu-
tion of small molecules within a cell is under consideration.
Diffusion artefacts will be detectable if significant changes car
take place in the distribution of a substance over distances
roughly of the order of 0.2 micron. If a substance of the molec-
ular weight of ethyl alcohol were present in a region of a cell,
such as a mitochondrion, approximately 0.2 micron in thickness,
in the few seconds required for cooling to —50° an almost com-
plete loss of substance from the mitochondria could occur if the
effect of cooling were to make the substance suddenly as dif-
fusible as is ethyl alcohol. Significant and indeed very serious
diffusion artefacts might, therefore, arise with small molecules
and small ions, even with the rapid fixation obtained with iso-
pentane cooled with liquid nitrogen. Even with large molecules,
significant artefacts could arise. Thus, with a protein having a
molecular weight of 20,000, about 1 percent could be lost from a
region 0.2 micron thick in 1 second, and for a substance of a
molecular weight of 106 the loss could be less than 0.1 percent.
It is clear, from calculation of the rise in viscosity as the tem-

FIXATION PROCEDURES 25

perature falls in the specimen, that significant diffusion cannot
occur after liquid-air temperatures have been reached, even
with a rapidly diffusing substance such as ethyl alcohol.

It thus appears that freeze-drying, or rather freezing in cool
isopentane, produces an effective fixation in a time which is of
the order of 1 percent of the time required for the effective
action of a chemical fixative, when a specimen of about 2 milli-
meters in thickness is considered. In making these calculations,
it has been assumed that the fixative moves into the specimen
by thermal diffusion only. In some cases the rate of penetra-
tion of a chemical fixing agent may be somewhat increased or
reduced by shrinking, cracking, or swelling of the specimen.
But it is highly improbable that the time for effective fixation
by chemical substances falls much below 100 times that required
for effective fixation by cooling, even with the maximum rate of
penetration which may be achieved when an accidental flow of
fixative is procured by cracking.

In theory it is possible that serious diffusion artefacts may
develop during fixation even with the freeze-drying technique.
This is certain to be the case with small molecules and small
ions, and it is improbable with substances of high molecular
weight. Substances of low molecular weight tend to be uni-
formly distributed in cells by thermal agitation. Opposed to
the effect of thermal agitation are adsorption and other more
complicated activities of the cell. The normal distribution of
a low-molecular-weight substance represents a compromise be-
tween these tendencies. As the temperature falls the dispersing
effect of thermal agitation declines correspondingly and the
relative importance of adsorption forces increases; since the
diffusion time for small molecules is bound to be small com-
pared with the cooling time, this abnormal dominance of ab-
sorption forces must inevitably produce redistribution of small
molecules. Diffusion artefacts of this type cannot be avoided.
On the other hand, with high-molecular-weight molecules ad-
sorption is normally much greater than for small molecules, and
the diffusion time is much larger. Consequently, with freeze-
drying, diffusion artefacts are not likely to occur with molecules
of high molecular weight. But with chemical fixation the bal-
ance is thrown heavily in favour of diffusion artefacts in two
ways. First, the action time of chemical fixation is about a

26 FIXATION PROCEDURES

hundred or more times longer than with freezing fixation, so
that artefacts have much longer to develop. Secondly, chemical
fixatives are inherently rather violent reagents and inevitably
break up most of the adsorption systems of cells, and so they
temporarily render many molecules more diffusible than is
normal. Consequently, diffusion artefacts must be expected,
even with large molecules, when chemical fixation is used.

It cannot be too strongly emphasized that, although freeze-
drying presents such a great advantage over chemical fixation,
artefacts may nevertheless arise with some types of molecules
during the fixation process. Also small molecules such as in-
organic ions and sugars may diffuse to a significant degree in the
wax used for embedding. This must be a subject for future
investigation.

Until the degree of diffusion during fixation and infiltration
is ascertained, it will not be possible to tell whether the locali-
zation of small molecules is reliable, even when freeze-drying,
followed by embedding in wax, is used as the basis of technique.
Thus most of the results obtained by microincineration have in-
volved specimens infiltrated in wax. Whether the distribution
of sodium, potassium, etc., found by this technique is correct
cannot be determined until it is known what artefacts develop
during fixation, infiltration, and embedding in wax.

Dr. L. G. Bell has also found that artefacts may arise when
gels are subjected to the freeze-drying technique. Although
one knows that, when examined with a microscope, a gelatin gel
is homogeneous, a section of gelatin gel prepared by the freeze-
drying technique is not homogeneous. It appears that when
weak gels are dried and forced into wax, cracking and contrac-
tion occur so as to give rise to a somewhat fibrous material. The
fibres are, however, artefacts.

From time to time it has been questioned whether the fact
that, after a specimen has been dried, it must normally be sub-
jected to a fixing agent does not deprive one of a great deal of
the advantage which is gained from the great speed with which
diffusion is restricted by cooling at room temperatures. How-
ever, there is probably very little loss in efficiency. Once a
specimen has been effectively dehydrated, most of the molecules
of molecular weight greater than, say, 1000, are bound into a
solid unit which will break down only if the specimen is passed
into a polar solvent; e.g., the protein molecules and the nucleic

REFERENCES 27

acid molecules are held in positions from which they cannot
diffuse to a significant extent. If a fixing agent is applied to
such molecules once they are held in this way, they will be
permanently fixed in their normal position, apart from such
cracking and contraction as may occur. Artefacts due to crack-
ing and contraction are, of course, rather readily detected in a
specimen.

In conclusion, Plates I and II illustrate two of the types of
artefacts which occur in chemical fixation. Plate I, Fig. A shows
the type of distribution of glycogen which is normally obtained
in liver with chemical fixation. Figure B shows the distribution
of glycogen obtained when freeze-drying is used as the fixing
technique. There is a striking contrast between the two results,
and it is evident that, as a result of the forces set up by diffusion
processes during the action of a chemical fixative, glycogen
in a cell is forced out of its normal position. Plate II, Fig. A
shows an ultraviolet photograph of rat Walker sarcoma fixed
in Camoy's fixative (acetic acid and alcohol) which is thought
by cytologists to be one of the best fixatives for the demonstra-
tion of nucleic acid in such material. In this preparation the
contrast in the resting nuclei is quite striking, and surrounding
the metaphase plate there is a clear zone containing practically
no material absorbing ultraviolet light. Plate II, Fig. B shows
quite different results obtained by freeze-drying. Contrast in
the resting nuclei is much less than in Plate II, Fig. A and there
is no clear zone surrounding the metaphase plate. It is obvious
that in the action of a chemical fixative considerable diffusion
may occur of a substance, probably ribonucleic acid, which ab-
sorbs light in the ultraviolet. After this demonstration, it seems
hardly necessary to emphasize the desirability of using the
freeze-drying technique whenever it is available.

Plate III illustrates the quality of the results which may be
obtained by cytochemical methods subsequent to freeze-drying.

REFERENCES

Altmann. 1890. Die Elementarenorganismen und ihre Bezeihungen zu
den Zellen (von Veit, Leipzig).

Bell. 1952. International Review of Cytology, 1, 35.

Gersh. 1932. Anal. Rec, 58, 309.

Gersh, Sylven, Sjostrand, and Bell. 1952. Freezing and Drying (Insti-
tute of Biology, London).

CHAPTER ^

STUDIES ON ALKALINE PHOSPHATASE

The localization of enzymes in particular parts of cells is one
of the general problems of cytochemistry which must be largely
solved before it will be possible to make full use of the body of
knowledge of enzymes which is being built up by the biochem-
ists. Alkaline phosphatase is a particularly favorable enzyme
for experimenting on possibilities in this field, since it is rather
resistant to experimental procedures and thus makes readily
accessible study of all the various artefacts which arise in the
cytochemistry of enzymes — with, of course, the exception of de-
struction of enzyme due to the sensitivity of the enzyme itself
to experimental procedures.

Alkaline phosphatase is usually defined as an enzyme which
splits phosphate esters according to the following equation

ROPO3H2 ^ ROH + HOP03H2

with a pH optimum in the vicinity of 9.3. It is not specific
for any special group of phosphomonoesters, but, on the other
hand, it will hydrolyze only the monoesters.

The first histochemical study of alkaline phosphatase was
published by Robison in 1923. He showed that, if a thick
section of hypertrophic cartilage is immersed in a solution con-
taining an ester of phosphoric acid and calcium, a precipitate of
calcium phosphate occurs in the section at sites which may be
revealed by staining with silver. He showed that the calcium
phosphate deposits are found only in those regions in which
primary bone formation is occurring, and he was thus able to
establish a close relationship between the histological occurrence
of alkaline phosphatase in cartilage and calcification of car-
tilage.

No further advance occurred until 1939, when Gomori and
Takamatsu practically simultaneously published papers in
which they showed that the type of technique developed by

28

GENERAL TECHNICAL CONSIDERATIONS 29

Robison could be extended down to the cytological level. This
work of Gomori and of Takamatsu gave a great new impetus
to the study of the cytological distribution of enzymes. Their
work, however, suffered from the disadvantage that they had
not studied most of the major physico-chemical hazards and
problems which may arise in the cytological study of enzymes.
These problems may, as was indicated in Chapter 1, be formu-
lated as follows:

1. How much phosphatase is destroyed by the experimental procedures?
And if phosphatase is destroyed, is it destroyed selectively at certain sites
in the tissue, or is it destroyed to the same extent at all the sites at which
it occurs?

2. Does the precipitate of calcium phosphate indicate the true site of
the alkaline phosphatase, or does it merely indicate sites in the tissues
which have a high affinity for calcium phosphate?

3. Is the enzyme in a fixed section in its physiologically normal position?

Until answers can be given to these three questions, it is im-
possible to place any reliance on results obtained by this tech-
nique. This chapter will consist of a study of the methods by
which these questions may be answered, and a consideration of
some of the results which have been obtained in cases where
satisfactory answers can be provided to these questions.

General Technical Considerations

It may be noted that to obtain optimal results, i.e., results
with a minimum of fixation artefacts and a minimum of de-
struction of enzyme, the freeze-drying technique is by far the
best. Where freeze-drying is not available, one may fix in 80
percent alcohol, and then take the block after 2 hours into
absolute alcohol, in which the material must remain for at least
2 hours. After the treatment with absolute alcohol, the block
may be either taken down to water and used for cutting frozen
sections or cleared by passage through such agents as cedar-
wood oil, methyl benzoate, or benzene. It may then be infil-
trated and embedded in wax, at a temperature which should be
kept below 60° C. If the material is embedded, sections may
be cut by the usual procedures, flattened on distilled water, the
wax removed with xylol, after which the sections are taken into
absolute alcohol and then into 0.1 percent collodion, from which

30 STUDIES ON ALKALINE PHOSPHATASE

they should pass into 50 percent alcohol, then into 25 percent
alcohol and distilled water. The section should remain at least
5 minutes in the distilled water so that small amounts of pre-
formed calcium phosphate may be removed from the section.

The sections are then incubated in a mixture containing a
phosphate ester and calcium. As the enzyme splits off phos-
phate from the ester precipitation of calcium phosphate occurs
in the sections. When studying tissues of unknown phosphatase
concentration, I have found it convenient to incubate trial sec-
tions for 2 hours and for 20 hours. With the information pro-
vided by these trial sections, it is possible to determine the opti-
mal periods for which sections should be incubated to discover
the detail of distribution of enzyme in the different sites in the
specimen. It is always advisable, when a detailed study is be-
ing made after the preliminary trials, to use a series of times of
incubation, e.g., for mammalian kidney, 5, 10, 20, 40, and 80
minutes. The composition of the incubation mixture which has
been used for most of the studies reported here has been:

Milliliters

2 percent sodium veronal 20

2 percent sodium glycerophosphate 20

2 percent calcium nitrate 10

2 percent magnesium chloride 2

Water 48

100

In some of the earlier work which is included in this chapter,
magnesium was omitted. Magnesium is included to act as an
activator of the enzyme. It is not always necessary but will
often enhance enzyme activity.

After incubation, the section should be washed in 2 percent
calcium nitrate solution for a few minutes, then taken into 1
percent cobalt nitrate solution. This converts the calcium
phosphate into cobalt phosphate. The section is then washed
in distilled water to remove the excess cobalt nitrate, and then
placed in an ammonium sulfide solution, which converts the
cobalt phosphate into cobalt sulfide, which is black. The sec-
tions may then be mounted in the usual manner. To avoid loss
of material from sections due to slight acidity in the washing
medium, the calcium nitrate solution, cobalt nitrate solution,

GENERAL TECHNICAL CONSIDERATIONS 31

and distilled water should each have an addition of 0.5 milli-
liter of sodium veronal per 100 milliliters of solution.

The method just described localizes the phosphate end of the
ester. A method published by Menten, Junge, and Green in
1944 and Danielli in 1946 localizes the alcoholic end of the
molecule as an insoluble diazo dye. In this method the incuba-
tion medium contains a phenol phosphate and a diazonium hy-
droxide. Phenol phosphates do not react with diazonium hy-
droxides. But, in the presence of phosphatase, the ester is split,
giving rise to phenol, which reacts very rapidly with the dia-
zonium hydroxide. By a correct choice of phenol and dia-
zonium hydroxide, it can be arranged that the reaction product
of the phenol and diazonium hydroxide is very insoluble in
water and is therefore precipitated in the section in the vicinity
of the enzyme.

The medium which is conveniently used for these experiments
consists of a phenol phosphate solution in sodium veronal, con-
taining 0.1 gram of phenol phosphate suspended or dissolved
in 50 milliliters of 2 percent sodium veronal solution. To this
is added, just before use, 25 milliliters of diazonium hydroxide
solution containing 0.2 percent of diazonium hydroxide. The
diazonium hydroxide solutions may be prepared by the methods
given by Saunders (1936). The diazonium hydroxide may be
kept for a period of hours in the hydrochloric acid solution in
which it is prepared, or often for more prolonged periods as a
stabilized compound in solid form. Just before use it is brought
to pH 9.2-9.3 by the use of thymol blue as an indicator, and
then mixed with phenol phosphate solution and filtered, keeping
the temperature between 6 and 8° C. throughout the operation.
After nitration the slides may be placed in the medium. Many
diazonium hydroxides decompose rather rapidly under these
conditions, and it may be necessary to change the incubation
medium every 20 minutes to obtain optimal results.

Some workers have had difficulty in using this technique,
through paying insufficient attention to the adjustment of the
pH of the diazonium hydroxide solution. The diazonium hy-
droxide reacts with the indicator thymol blue, which is itself
a phenol, to give another indicator substance which has a color
change very similar to that of thymol blue, but with which the

32 STUDIES ON ALKALINE PHOSPHATASE

color change occurs only in much more strongly alkaline solu-
tions. Consequently, unless due care is used, the pH of the
diazonium hydroxide solution is not adjusted to pTL 9.2 but to
a strongly alkaline value. When the tissue sections are im-
mersed in this strongly alkaline solution, the phosphatase is
rapidly destroyed and, of course, no cytochemical reaction oc-
curs. To avoid this difficulty drops of indicator should be placed
in a white tile, and to one of these should be added a drop of
the diazonium hydroxide solution. When the diazonium hy-
droxide solution is at approximately pH 9.2-9.3, on addition of
the drop the mixture will initially turn to the appropriate blue-
grey color, which will then rapidly fade. If the solution remains
blue, the diazonium hydroxide solution is too alkaline.

Loss of Enzyme Due to Experimental Techniques

Under this general heading will be included losses due to diffu-
sion of substances out of tissue sections at various stages of the
procedure, and also loss of enzyme due to inactivation.

Effect of Fixation

The fixative originally recommended by Takamatsu and Go-
mori was 80 percent alcohol. The effect of this 80 percent alcohol
on phosphatase activity was studied on the test-tube level by
the usual biochemical procedure for the study of alkaline phos-
phatase. Pieces of rat kidney were ground up in a mortar and
the activities of aliquots determined (a) without fixation, (b)
after 2 hours' exposure to 80 percent alcohol. No significant dif-
ference could be found between the phosphatase activity of the
two samples. Further experiments were carried out (c) with 10 ^
frozen sections cut from a block of unfixed rat kidney and 10 fi
frozen sections which had been fixed for 2 hours in 80 percent
alcohol and 6 hours in absolute alcohol. The sectioned material
was ground as before and studied by the biochemical procedure.
Again no difference could be found between the enzyme content
of the fixed and unfixed material. It may therefore be concluded
that no significant loss of alkaline phosphatase occurs as a result
of fixation in alcohol. Consequently it was possible to use
the deposition of cobalt sulfide in frozen sections from alcohol-

THE CRITICAL TIME LIMITS OF OPERATIONS 33

fixed blocks as a standard for estimating roughly enzyme losses
caused by paraffin embedding, etc.

The Effect of Embedding, etc.

Since alcohol does not destroy alkaline phosphatase to a sig-
nificant degree, frozen sections from alcohol-fixed blocks may
be used to control the effect of later procedures. Sections were
cut from a paraffin block and incubated for 0, 10, 20, 40, 80, 160,
and 320 minutes in the glycerophosphate incubation mixture
given above. This series of sections was compared with a simi-
lar series prepared as frozen sections from an alcohol-fixed block
for the same periods. The comparison showed that a consider-
able loss of activity occurred during infiltration and embedding.
In a number of such experiments it was estimated that the loss
of enzyme activity may amount to 75 percent of the total. The
loss, however, did not appear to be localized but occurred to
the same extent, as far as may be distinguished by the eye, in all
sites in which the enzyme occurs. If the fixation, infiltration,
etc., were carried out under anaerobic conditions, a much greater
destruction of enzyme activity was found to occur.

The Critical Time Limits of Operations

Preliminary experiments suggested that loss of apparent
enzyme activity might be occurring in some stages of the pro-
cedure. Two groups of experiments were therefore carried
through to see to what extent this could be controlled. Three
blocks of material were fixed in 80 percent alcohol and then
taken to absolute alcohol after 2 hours. From one of these
blocks a series of frozen sections was prepared. The second
block was allowed to spend 2 hours in each of the dehydrating
and clearing agents and infiltrated with wax for 8 hours at
60° C. The third block was allowed to spend 24 hours in each
dehydrating and clearing agent and was then infiltrated for 48
hours at 60° C. When series of sections from all three blocks
were prepared by the glycerophosphate method, comparison of
the sections showed that a loss of about 50 percent of the enzyme
had occurred in both the second and third blocks (using the
frozen sections as standard) ; but there was no significant dif-
ference between the second and third blocks themselves. Thus

34 STUDIES ON ALKALINE PHOSPHATASE

extension of the time of fixation, clearing, infiltration, etc., does
not appear to increase the amount of destruction of enzyme.

In the second series of experiments, large numbers of sections
were cut from one block of paraffin-embedded rat kidney. A
number of control sections were taken through the whole glyc-
erophosphate procedure as quickly as possible, apart from a
period of 1 hour for incubation in the glycerophosphate mixture.
The experimental group of sections was taken through all steps
in the procedure, bar one, as rapidly as the control sections, but
it was exposed to this one step in the procedure for 24 hours;
i.e., an experimental section was exposed to, e.g., xylol, alcohol,
or collodion, or simply kept at 37° in air after flattening, or in
distilled water before incubation, or in calcium nitrate, cobalt
nitrate, distilled water, ammonium sulfide, or one of the alcohols
or xylols after incubation, for 24 hours. The loss of apparent
enzyme activity amounted to 100 percent in some of these steps.
Further study was made of those steps which caused loss of ap-
parent enzyme activity. The results are incorporated in Table
II which gives both the duration of each step which has been
found to be generally convenient in the laboratory and the
maximum duration of the steps which I have found caused no
loss of enzyme activity. It is probable that some of the steps
may be even more protracted than suggested in the table with-
out causing loss of enzyme activity. On the other hand, certain
steps, notably k, I, and m, of the table are very likely to cause
loss of apparent activity, so that timing of these operations
must be very carefully controlled. Although step m (washing
in tap water after ammonium sulfide) may cause loss of enzyme
activity if protracted, sections may be left for many hours in tap
water to which a little ammonium sulfide has been added with-
out loss of apparent activity.

Where small traces of phosphatase activity are being studied,
it is advisable to add one drop of 0.1 M disodium phosphate solu-
tion to the incubation mixture, and then filter off the precipitate
before incubation. This insures that the incubation mixture is
saturated in phosphate ion so that any trace of phosphate which
is liberated by enzyme activity is more likely to be precipitated.
As noted earlier, it is also necessary to add a little sodium
veronal to the calcium nitrate, cobalt nitrate, and distilled water
which are used after incubation; this is necessary to counteract

THE DETECTION OF DIFFUSION ARTEFACTS 35

TABLE II

Time limits of the various steps in the procedure of Takamatsu and
Gomori.

Step in Procedure
Steps before incubation

(a) Fixation

(b) Dehydration (3 changes)

(c) Clearing (3 changes of cedarwood
oil)

(d) Infiltration with wax

(e) Drying of sections on slide at room
temperature in desiccator at
37° C. in incubator

(/) Removal of wax in xylol
(g) In celloidin or alcohols
{h) In distilled water

Steps after incubation
(i) In calcium nitrate
(j) In cobalt nitrate
(k) In distilled water
(0 In ammonium sulfide
(m) In tap water
(n) In alcohols or xylols

Maximum

Duration

Found to

Cause No

Convenient

Loss of

Duration

Activity

2 hours

24 hours

6 hours

72 hours

6-24 hours

72 hours

0.5-24 hours

48 hours

2-24 hours

7 days

3 minutes

24 hours

1 minute

24 hours

5-60 minutes

24 hours

5 minutes

7 days

2 minutes

24 hours

1 minute

2 minutes

1 minute

24 hours

5-10 minutes

30 minutes

3 minutes

5 minutes

any tendency towards acidity in these solutions, which would,
of course, cause removal of precipitate from the sections.

Intact paraffin blocks will often keep 12 months or more with-
out serious loss of activity. The enzyme in blocks which have
been cut often seems to be less stable.

The Detection of Diffusion Artefacts

We may classify the artefacts which may arise by diffusion
processes into three groups.

1. Diffusion of calcium phosphate, cobalt phosphate, or cobalt sulfide.

2. Diffusion of enzyme, enzyme activator, or enzyme inhibitor after
fixation.

3. Diffusion of enzyme, enzyme activator, or enzyme inhibitor during
fixation.

36 STUDIES ON ALKALINE PHOSPHATASE

If the glycerophosphate technique is to have biological sig-
nificance, one needs to establish that the cobalt sulfide precipi-
tate produced by this technique demonstrates with precision the
site of enzyme activity, and also that the enzyme is in its
physiologically normal position. This can be best demonstrated
by separate consideration of the three groups of diffusion arte-
facts just mentioned. It should, however, be noted that as yet
no studies have been made of the concentrations of activators,
inhibitors, substrates, etc., for phosphatase which are found
in the immediate vicinity of the enzyme in the living cell. Until
such studies are made, even the demonstration of the presence
of enzyme in its physiologically normal site does not enable
us to foretell with certainty the extent to which the enzyme is in
fact active at its site.

Sites of Affinity for Calcium Phosphate

In the glycerophosphate technique the initial precipitate
which is formed is calcium phosphate. It is therefore of interest
to determine to what extent fixed material has local sites of
special affinity for such substances. The best way to test this
is to cause the liberation of phosphate in the incubation mix-
ture by the spontaneous decomposition of a labile ester, and to
observe the extent to which the liberated phosphate is precipi-
tated in sections which are devoid of enzyme activity. The
phosphatase of tissue sections may readily be inactivated by
exposure to wet steam for 10 minutes, or by being placed in
water at 90° C. for 2 minutes.

A solution containing a labile ester may be obtained simply
by adding hydrogen peroxide to the normal glycerophosphate
incubation mixture. The glycerophosphate undergoes oxidation,
and phosphate ion is slowly liberated and precipitated by the
calcium ions present in the medium. After addition of hydrogen
peroxide it is necessary to readjust the pH to 9.3, after which
inactive sections may be inserted and incubation carried out
at 37° C. Sections of kidney and of healing wounds in rat
skin were studied in this manner. The nuclei of the kidney
section were found to take up calcium phosphate with much
avidity, but the brush borders did not display any significant
affinity for calcium phosphate. In sections of healing wounds
the nuclei and the newly formed collagen fibres displayed a

DIFFUSION OF CALCIUM PHOSPHATE

37

high affinity for calcium phosphate. In these two types of sec-
tion, the distribution of phosphatase as demonstrated by the
glycerophosphate technique with sections containing active
enzyme is such that the brush-border regions of the kidney con-
tain the highest concentrations of phosphatase, newly formed
collagen the next highest, and nuclei a moderate to low concen-
tration of phosphatase. It is thus clear that some sites which
appear to contain phosphatase have a high affinity for calcium
phosphate but that there is no correlation between this affinity
and the absolute concentrations of phosphatase appearing in
the cells.

The Ability of Calcium Phosphate to Diffuse within a
Section

As noted just above, when a tissue section is heated it loses
its phosphatase activity, but retains its affinity for calcium phos-

Inactive
section

Active
section

^^

^mmzzmmz

Glass slide
Fig. 2. Illustrating arrangement of active and inactive sections on a slide

required to test for diffusion artefacts.

phate. It is therefore possible by superimposing a section con-
taining active enzyme upon a section in which the enzyme has
been destroyed to test the extent to which calcium phosphate
may diffuse in sections from one site to another. It is best to
arrange the sections as illustrated in Fig. 2 with the inert sec-
tion lying underneath the section containing active enzyme, so
as to observe the maximum effect of diffusion of calcium phos-
phate. In the first studies which were made using this tech-

38 STUDIES ON ALKALINE PHOSPHATASE

nique (in 1944), active kidney sections were placed upon in-
active sections of kidney, spleen, and healing wounds. I found
that with 12 hours' incubation there was no evidence for dif-
fusion of calcium phosphate. It is therefore clear that, when
the experiment is carried out with the conditions and timing
set out here, calcium phosphate does not diffuse during the ex-
perimental procedure to a degree that can produce a precipita-
tion which might be interpreted as indicating the presence of
enzyme in a position in which it is not intrinsically present. It
is also clear, of course, that neither the cobalt phosphate nor
the cobalt sulfide can be diffusing to such an extent as to pro-
duce an observable artefact. We must therefore conclude that
the site of precipitation of calcium phosphate is identical with
the site of enzyme activity. It is important in studies of effects
of this type that judgment should be based upon diffusion from
an active section into an inactive section of the same material.
If the underlying section is of a different material from the
active section, and evidence for a diffusion artefact should be
found, this would not constitute evidence that similar artefacts
would occur in the active section, since affinities for various sub-
stances differ from one tissue to another.

When, as may occur in some instances, the enzyme itself is
diffusible, the technique given above may not make it possible
to make an independent study of the extent to which calcium
phosphate is diffusing in a section. We may then resort to the
device of carrying out the experiment in superimposed sections
in the presence of different concentrations of calcium ion in the
incubation mixture. Since calcium ion is present in the incuba-
tion mixture in a great excess, the rate at which calcium phos-
phate can be transferred from one site on the section to another
site depends upon the solubility of phosphate ion. We have the

equation: [Ca2+]3[P043-]2 = constant

defining the solubility of phosphate ion, i.e.,

[P043~] = VConstant/[Ca2+]3

Thus we see that the rate at which an artefact can be established
due to the diffusion of calcium phosphate will depend upon the
calcium ion concentration in the incubation medium.

RELEASE OF ALCOHOLS FROM ESTERS 39

The Site of Liberation of the Alcohol Moiety of
Phosphate Esters

Although the evidence from the study of precipitated cal-
cium phosphate appears to show quite clearly that techniques
involving the precipitation of calcium phosphate indicate reli-
ably the site of alkaline phosphatase, it nevertheless seemed
desirable to reinforce this evidence by attempts to secure a pre-
cipitation of the alcohol moiety of the phosphate esters. This
would be very difficult to do with glycerophosphate. But with
a number of other esters of phosphoric acid it has been found
possible to secure the precipitation of the alcohol concerned.
The chief substrates which proved useful in this connection
are:

1. Phenolphthalein phosphate.

2. /3-Naphthol phosphate.

3. p-Nitrophenylazo-a-naphthol phosphate.

When phenolphthalein phosphate is studied, the calcium salt
is used in saturated solution in an incubation mixture contain-
ing calcium. Under these conditions both moieties of the ester
are precipitated, the phosphate moiety as calcium phosphate and
the phenolphthalein probably as a calcium salt. After incuba-
tion the sections are dried in air until the greater part of the
water has evaporated and are then exposed to ammonia vapor.
This converts the phenolphthalein into the bright-red alkaline
form. If a sufficient amount of water has been removed from
the section, phenolphthalein does not diffuse out at a significant
rate, and its distribution may be photographed. If, however, too
much water is removed by drying, ammonia does not convert
the phenolphthalein into the red form. To obtain just the right
amount of moisture in the section is a somewhat tricky pro-
cedure. When the phenolphthalein distribution has been re-
corded photographically, the distribution of calcium phosphate
may be ascertained by conversion of the calcium phosphate to
cobalt sulfide, as in the standard glycerophosphate technique.
The section may then again be photographed. It is found by
this method that the distribution of phenolphthalein is closely
similar to that of calcium phosphate, with the reservation that,
as the colour of phenolphthalein is much less intense than that
of cobalt sulfide, it is not possible to observe the presence of

40 STUDIES ON ALKALINE PHOSPHATASE

phenolphthalein in the sites which show a low activity by the
calcium phosphate method.

The distribution obtained by this method is closely similar to
that obtained by the use of glycerophosphate in the substrate.

It might be thought that phenolphthalein has a specific affin-
ity for either certain parts of tissues or for the calcium phos-
phate which is precipitated. This was shown not to be so by
exposing sections to a strong solution of phenolphthalein. There
was no specific staining of the section. Also, when to a solution
of phenolphthalein in the incubation mixture a little alkaline
sodium phosphate solution was added, phenolphthalein is not
carried down to a significant extent by the precipitate of cal-
cium phosphate which is formed. These two experiments show
quite clearly that the precipitation of phenolphthalein as a re-
sult of enzyme activity occurs owing to its insolubility and not
by adsorption at sites which have a specific affinity for phenol-
phthalein.

As mentioned previously, phenol phosphates do not react di-
rectly with diazonium hydroxides. Consequently, when a phenol
phosphate and a diazonium hydroxide are mixed in alkaline so-
lution, no reaction takes place. But, when a section is placed
in this mixture, if alkaline phosphatase is present it splits phenol
and phosphate from the ester, and the phenol rapidly reacts
with any diazonium hydroxide which is present. Thus by using
a suitable diazonium hydroxide and a suitable phenol, it should
be possible to secure the precipitation of the phenol released by
enzyme action, at the actual site of enzyme action. The most
commonly used substances for this purpose are a- or /?-naphthol
phosphate and diazotised a-naphthylamine, which react as
follows:

>OP03H2 < >OH

■» HOP03H2 + > <

enzyme

diazonium hydroxide

>N=N^ >OH

RELEASE OF ALCOHOLS FROM ESTERS 41

When mammalian kidney sections are studied by this method,
the azo dye is precipitated in the brush borders and also in the
nuclei of white cells (Table III) . It is usually not found in the
nuclei of the other cells present in the tissue. There are prob-
ably two reasons why the low concentration of phosphatase in
the other cell nuclei cannot be demonstrated by this technique.
In the first place, the azo dye is not completely insoluble in
water and it is probable that, unless the rate of production of
the azo compound at a particular site exceeds a given limit,
precipitation does not occur in the section. Secondly, the dia-
zonium hydroxide also reacts with certain components of the
tissue section themselves, producing a background stain. Very
small amounts of precipitated azo dye cannot be distinguished
from the background staining. It may also, of course, be pos-
sible that nuclear enzymes react less vigorously with /3-napthol
phosphate than with glycerophosphate. Provided that the azo
dye resulting from the interaction of the phenol and diazonium
hydroxide is sufficiently insoluble, the same result is obtained
with any mixture of phenol phosphate and diazonium hydrox-
ide. Table III summarises results which have been obtained
with a number of different bases.

TABLE III

Summary of results obtained by exposing kidney sections containing active
phosphatase to a solution at pK 9.3 containing a phenol phosphate and a
diazotised amine. The remarks refer to the localization of dye produced by
the interaction of the diazonium hydroxide with a phenol liberated by the
action of the enzyme on a phenol phosphate.

Site of Dye Precipitated with

Staining

Sodium

Calcium

of

Naphthol

Naphthol

Amine Used

Control

, Phosphate

Phosphate

a-Aminonaphthoquinone in

Deep

Not localized

Brush borders

saturated solution

p-Nitroaniline, 0.2%

Pale

Not localized

In cortex but
restricted to
brush borders

Benzidine, 0.2%

Moderate

Brush borders

Brush borders

a-Naphthylamine

Pale

Brush borders

Brush borders

/3-Naphthylamine

Pale

Brush borders

Brush borders

not

42 STUDIES ON ALKALINE PHOSPHATASE

The third substrate, p-nitrophenylazo-a-naphthol phosphate,
is itself the phosphate ester of an insoluble dye. The solution
of the phosphate ester is probably colloidal. When kidney sec-
tions are immersed in a solution of this dye phosphate at pH
9.3, the dye appears both in the brush borders and in the nuclei.
However, not all the nuclei are in fact stained in this manner.
Of two adjacent nuclei, one may stain and one may not. As
was the case when naphthol phosphate was used as a sub-
strate, this is probably due to the necessity for the rate of liber-
ation of dye in a particular site to reach a critical level before
precipitation can occur. In those nuclei which have a concen-
tration of enzyme high enough for this critical rate of libera-
tion to be reached, precipitation occurs. In those nuclei in
which the critical rate is not reached, no precipitation occurs.
The same phenomena can also be observed with this substrate
in the brush borders themselves. Some regions of the brush
borders contain a rather low level of phosphatase activity and
do not stain at all with this substrate, whereas most of the
brush-border region has a high concentration of phosphatase
and is deeply stained by the substrate.

When we consider the results of studying the precipitation of
both the phosphate moiety and the alcoholic moiety of a phos-
phate ester after liberation by enzyme action, it seems impos-
sible to escape the conclusion that in all cases, within the limits
denned by the study of diffusion artefacts, the precipitate pro-
duced is in fact at the site of enzyme activity. If this were
not so, the same region of tissues would need to have an affinity
for calcium phosphate, phenolphthalein, and various azo dyes.
Kidney tissue has no affinity for any of the azo dyes or for
phenolphthalein. Calcium phosphate has an affinity for one
of the two main sites of activity in kidney sections but, on the
other hand, has been shown to be unable to diffuse within a sec-
tion to a significant degree. It is thus certain that the site of
alkaline phosphatase in tissue sections is truly demonstrated
by these cytochemical methods.

The Effect of Thickness of the Section

All the various procedures which have been outlined above in
connection with different techniques were worked out for sec-
tions between 5 and 8 p in thickness. It is sometimes necessary

THE EFFECT OF THICKNESS OF THE SECTION 43

to use sections which are thicker or thinner than this. When
this is done, it may be necessary to consider whether any of the
times of the various treatments should be modified. Obviously
a time of washing which may be satisfactory for a 2 ^ section
may not be satisfactory for a 16 /* section. Whether this is so
depends on whether the major resistance to diffusion is encoun-
tered in diffusing out of or into the tissue section, or whether the
dead layer of water surrounding the section constitutes the
major barrier. If the resistance to diffusion is a function of the
thickness of the section, then the time required to secure a given
degree of washing should be roughly proportional to the square
root of the thickness, whereas, if the dead layer of the water
is the main factor, the time for obtaining this given degree of
washing should be independent of the thickness. To investi-
gate this point a number of series of sections were cut at 1 /*, 2 /*,
4 fi, S [x, and 16 /*. These were incubated in the glycerophos-
phate incubation mixture for 15 minutes, and then developed
with the usual cobalt technique, with the exception that different
sections of the same thickness were washed for various lengths
of time in distilled water. The times for washing varied be-
tween 0.25 minute and 16 minutes. Owing to the solubility of
cobalt phosphate in distilled water, washing for more than 2
minutes with an 8 /x section causes loss of apparent enzyme ac-
tivity from a section if carried out in distilled water. The losses
are much smaller if the washing is carried out in 0.002 percent
sodium veronal, as was the case with the results displayed in
Table IV. Under such conditions the minimum washing time
for removal of cobalt nitrate increases from about 1 minute for
1 fx sections to about 4 minutes for 16 p sections; with 1 /*
sections a significant amount of cobalt phosphate was not re-
moved from the section even after 16 minutes' washing.

It may perhaps be mentioned here that, whereas most tissues
do not take up cobalt to a significant degree from cobalt nitrate
solution, this is not the case when ova are studied. The yolk
of echinoderm ova, for example, takes up cobalt very readily:
this cobalt cannot be removed by washing in distilled water
without also removing any cobalt phosphate which is present.
So far no satisfactory solution has been found to this problem,
but it seems possible that, if a lead or copper solution is sub-

44 STUDIES ON ALKALINE PHOSPHATASE

TABLE IV

The effect of washing sections of various thicknesses for various lengths
of time. The sections were of rat kidney incubated with glycerophos-
phate by the Gomori method, after which calcium phosphate present in
the section was converted into cobalt phosphate by treatment with co-
balt nitrate. The sections were then washed in 0.002 percent sodium
veronal. + indicates that the section is washed free of cobalt nitrate
but has lost no appreciable amount of cobalt phosphate.

Section

Thickness

Washing

Time

in Minutes

in Microns

0.25 1

2

4

8

16

1

+

+

+

+

+

2

+

+

+

+

+

4

+

+

+

+

8

+

+

+

+

16

+

+

+

stituted for the cobalt solution, and the sections are washed
in dilute acetic acid or citric acid instead of in distilled water,
the yolk could be freed from stain without dissolving the pre-
cipitated phosphate.

The Physiological Sites of Phosphatase Activity-
It has been demonstrated that the various cytochemical re-
actions do indicate sites of alkaline phosphatase in the section.
It still remains to be shown that the enzyme in a fixed prepara-
tion does not diffuse to a significant degree during or after
fixation. Neumann has demonstrated that practically all the
phosphatase may be removed from some materials by prolonged
contact with 30 percent alcohol. Clearly, exposure to dilute
alcohol must be avoided if diffusion of phosphatase is not to
cause serious difficulty. So far as I am aware, Martin and
Jacoby (1949) were the first to obtain clear evidence that
phosphatase may diffuse during the course of incubation. There
are at least three possible methods of estimating the extent to
which the distribution of phosphatase is modified by diffusion.
In all these methods it is necessary to use a logarithmic series of
incubation times. The exact lengths of time of incubation must
be chosen according to the amount of phosphatase which is
present in the section.

PHYSIOLOGICAL SITE OF ACTIVITY 45

The first method is to use superimposed sections, as was done
in the study of the diffusion of calcium phosphate recorded above.
This method of using superimposed sections will, in fact, in one
operation record the diffusion, if any, of phosphatase, calcium
phosphate, cobalt phosphate, and cobalt sulfide. A detailed ex-
ample of the use of this method will be given later when the
occurrence of phosphatase in cell nuclei is discussed.

The method of using superimposed sections is quite convenient
when the tissue components which are to be studied constitute a
considerable fraction of the area of the section. But when the
tissue component constitutes only a small fraction of the sec-
tion, it is difficult to superimpose a section containing active
enzyme upon a section containing no enzyme in such a manner
as to obtain an active tissue component immediately above the
corresponding inactive component in the underlying section.
Under these circumstances, both loss of time and waste of ma-
terial may be avoided by using series of sections which have been
incubated for the usual logarithmic series of times in a number
of dilutions of the same substrate. As the substrate concentra-
tion is diminished the rate of enzyme action is also diminished,
and, consequently, the rate at which a given component attains
a given degree of staining is diminished. On the other hand, the
rate of development of diffusion artefacts due to diffusion of
phosphatase will not be seriously affected by changing the
substrate concentration. Consequently, by comparing the series
of slides which have been incubated in the different substrate
concentrations, it is possible to determine the extent to which
phosphatase is diffusing within the section.

As an alternative to various concentrations of the same sub-
strate, a given concentration of different substrates may be stud-
ied, the same logarithmic series of times of incubation being
used. If the substrates are split by the enzyme at different rates,
comparison of the series incubated in the various substrates
should again demonstrate the extent of diffusion of phosphatase.

So far I have not made use of this last method. We are at
present studying the use of phenol phosphate and paranitrophenyl
phosphate to see whether these compounds together with glycero-
phosphate constitute a satisfactory series. The rate at which
phosphate is split from these esters by ordinary alkaline hydroly-
sis varies about a hundredfold or more.

46 STUDIES ON ALKALINE PHOSPHATASE

In the various studies which I have made, artefacts due to the
diffusion of phosphatase have usually not occurred within the
time period of incubation which was necessary to obtain adequate
information about the distribution of enzyme within sections.
In the few cases where diffusion has occurred to a significant de-
gree, the methods indicated above have made it quite simple to
assess its importance, with one exception. The exceptional case
is that of tissue cultures, which will be discussed later.

Martin and Jacoby, in a study of the distribution of phos-
phatase in the small intestine, appear to have encountered very
much greater difficulty with diffusion of phosphatase: this is no
doubt due to the fact that the small intestine usually contains
a considerable concentration of diffusible phosphatase which acts
as a digestive enzyme. They appear to have had difficulty also
in interpreting their results with other tissues : this seems in part
due to incubation for an excessive length of time. But there
are also two points in their experimental procedure which may
have caused them some difficulty. In the first place, many of
their experiments involved superimposing tissue sections on
sections of different material, usually guinea-pig liver. Now it
is of no interest to know to what extent enzymes will diffuse from
one tissue into another. What is important is to know to what
extent enzymes will diffuse within a tissue and thus give rise to
artefacts. It is therefore essential that experiments with super-
imposed sections should be carried out with an active section
superimposed on an inert section of the same material. The
second fault in their procedure lay in rinsing the sections in dis-
tilled water immediately after incubation. The section at this
stage contains precipitated calcium phosphate which has a very
significant solubility in water in the absence of calcium ions, or
if the pH is shifted much below pB. 9. The effect of rinsing in
distilled water is to shift the pH towards the acid side and at
the same time to diminish the calcium ion concentration. It is
thus possible that the precipitate of calcium phosphate formed
in their sections was partially dissolved by rinsing in distilled
water when, of course, it would be freely diffusible. Subse-
quently, when they transferred the sections from distilled water
to cobalt nitrate solution, reprecipitation would occur at many
cites, and simulate diffusion artefacts.

DIFFUSION OF ENZYME DURING FIXATION 47

Diffusion of Enzyme during Fixation

We have seen that the position of an enzyme within a section
may be determined with some accuracy and that the extent
to which the enzyme diffuses within a section may also be de-
termined without difficulty. It remains to be seen whether
the enzyme present in a fixed preparation is in its physiologically
normal position.

As was indicated in the chapter on fixation, the only method
of fixation which is sufficiently rapid to exclude diffusion of an
enzyme is freeze-drying. Where this method is used, one may
be sure that diffusion of phosphatase does not occur during fix-
ation. Where freeze-drying cannot be used, the extent of dif-
fusion, and the site at which the diffused enzyme appears, can
in some cases be detected, provided that a wide variety of fixa-
tives are used. With a variety of fixatives the time avail-
able for diffusion and the adsorption conditions prevailing in
the specimen should vary considerably from fixative to fixative.

In a study of the action of chemical fixatives, a number of in-
dividual components of fixatives were tested for their direct ac-
tion upon the enzyme by placing a section from an alcohol-
fixed block in the fixative component. After 2 hours' exposure
to the fixative component the sections were washed carefully
and then incubated for the same length of time as control sec-
tions. The fixative components tested in this way were 8 per-
cent formaldehyde; saturated picric acid; saturated magnesium
sulfate; saturated ammonium sulfate; saturated calcium chlo-
ride; saturated mercuric chloride; 1 percent, 5 percent, 10 per-
cent, and 50 percent acetic acid; 1 percent trichloroacetic acid;
10 percent pyridine; acetone; chloroform; dry pyridine, alco-
holic iodine, 1 percent osmic acid; 1 percent iodine in potassium
iodide; 0.04 percent reduced thioglycollate; 0.04 percent re-
duced glutathione; and 0.04 ^percent oxidized glutathione.
Phosphatase was completely destroyed by mercury, osmic acid,
trichloroacetic acid, iodine, and concentrations of acetic acid
higher than 1 percent. There was a marked reduction in the
amount of enzyme activity caused by formaldehyde, magnesium
sulfate, ammonium sulfate, picric acid, and reduced thiols, but
the amount of reduction in activity was not sufficient to preclude
their use in fixatives.

48

STUDIES ON ALKALINE PHOSPHATASE

TABLE V

Distribution of alkaline phosphatase in sections after fixation in various
media, and after freeze-drying. ++ indicates gross shrinkage; +, some
shrinkage; 0, no change; — , some swelling; =, gross swelling.

Shrink-

Fixative

None

Freeze-drying
Ethyl alcohol 80%
Formaldehyde 4% + so-
dium chloride 1%
Acetone
Pyridine

Pyridine 70%

Pyridine 25% in alcohol
80%

Calcium chloride 1% in
pyridine 70%

Formaldehyde 4% + cal-
cium chloride 1%

Saturated picric -f- formal-
dehyde 4%

Ethyl alcohol 65% + chlo-
roform 35%

Pyridine 70% + formalde-
hyde 4%

Pyridine 20% + ethyl al-
cohol 70% + formalde-
hyde 4%

Alcohol 60% + chloroform
30% + formaldehyde

Formaldehyde 4% + ace-
tic acid 1%

Formaldehyde 4% + ace-
tic acid 5%

Ethyl alcohol 80% + ace-
tic acid 1%

Pyridine 25% + picric
(sat. soln.) 75%

Saturated aqueous ammo-
nium sulfate followed by
alcohol 80%

Saturated aqueous calcium
chloride followed by al-
cohol 80%

Saturated aqueous magne-
sium sulfate followed by
alcohol 80%

0

0

0

+

+ +
+ +

Distribution of Enzyme

Mainly in cortical tu-
bules: smeared appear-
ance
Brush borders and nuclei
Brush borders and nuclei
Brush borders and nuclei

Brush borders and nuclei
Brush borders and nuclei

Brush borders and nuclei
Brush borders and nuclei

Brush borders and nuclei

Brush borders and nuclei

Enzyme destroyed

Enzyme spread all over
proximal tubules

Very poor Enzyme spread over

whole section
Very good Brush borders and nuclei

age
0

r ixation

• • • •

0

+
0

Excellent

Fair

Good

+
+

Fair
Good

+

0

Good
Good

+

Good

+

Good

0

Fair

0

Poor

0 Poor

Brush borders and nuclei;
some spreading into
cytoplasm
Very good Brush borders and nuclei

Good
Fair

Fair
Fair

Poor

Fair

Enzyme destroyed
Brush borders and nuclei

Brush borders and nuclei
Brush borders and nuclei

Brush borders and nuclei

Brush borders and nuclei

GENERAL STATUS OF TECHNIQUES 49

Consequently, a series of fixatives were made up as indicated
in Table V. Into these fixatives were placed pieces of rat kid-
ney approximately 1 millimeter thick. The fixative was al-
lowed to act for 2 hours, after which the tissue was washed in
water, frozen sections were prepared, and the sections were
incubated for logarithmic series of times for each fixative. As
will be seen in the table, where the fixation was generally satis-
factory from a cytological point of view the enzyme in kidney
tissue was restricted to the brush-border region and to the
nuclei. This conclusion was reinvestigated later with the freeze-
drying method. On the whole, the general nature of the con-
clusions reached with chemical fixatives was confirmed. But
it was also found that with all the chemical fixatives some dif-
fusion of phosphatase into the nuclei appeared to occur. With
chemical fixation the chromocentres of the nuclei appeared to
contain much more phosphatase than in freeze-dried specimens.
Thus in the case of kidney tissue the broad outline of the re-
sults obtained with chemical fixatives is correct, but the detailed
distribution of phosphatase cannot be studied without freeze-
drying. In all the tissues which I have studied all the cytologi-
cally satisfactory fixatives give a picture substantially similar
to that obtained by freeze-drying. This, however, may not be
the case with tissues containing large amounts of freely diffus-.
ible phosphatase, such as small intestine.

The general conclusion appears to be that freeze-drying
should be employed wherever possible. Where this is impos-
sible, a wide variety of fixing agents will probably give an
overall picture which is correct in outline but misleading in
detail.

The General Status of Techniques for the Localization
of Alkaline Phosphatase
The general conclusions which may be drawn from critical
studies of alkaline phosphatase techniques differ slightly with
the various techniques. In all cases it is desirable, as indicated
above, to use methods for the detection of diffusion artefacts
based upon the technique of imposed sections, or one of the other
techniques given above. When this is done, where the initial
precipitate is calcium phosphate, fairly accurate localization
of enzyme can be determined even within a nucleus, though it

50 STUDIES ON ALKALINE PHOSPHATASE

is quite possible, for example, that the detailed distribution of
phosphatase on the chromosomes is not an accurate representa-
tion of the enzyme distribution in vivo. With the techniques
involving azo dyes, less accuracy is possible. It is quite pos-
sible to say whether the nucleus of a mammalian somatic cell
has phosphatase or not. But it is not possible with the azo-dye
techniques to determine the details of the distribution of phos-
phatase within a given nucleus of this size. Neumann, and
Martin and Jacoby, have expressed considerable pessimism
about some of these techniques. This pessimism, however, is
based partly upon results obtained with techniques which to me
appear to be defective in certain essentials, and particularly
upon failure to use adequate methods of studying diffusion
artefacts, so as to make a quantitative estimate of the extent
of diffusion.

Alkaline Phosphatases of Cell Nuclei

Studies have been made by many investigators on mass prep-
arations of cell nuclei prepared by various methods. However,
as was pointed out in the first chapter of this book, there are
two major difficulties which make it impossible at the present
time to evaluate studies on mass preparations of cell compo-
nents prepared by maceration techniques. These difficulties are
that substances diffuse in and out of nuclei fairly readily when
nuclei are suspended in aqueous solutions, and that in addition
to this the nuclei prepared from most organs cannot consist com-
pletely of nuclei from one cell type only. In view of these
difficulties, the only method which at present can provide reli-
able information as to whether there is alkaline phosphatase
in cell nuclei is the cytochemical method. A typical result is
that obtained with rat kidney, shown in Plate IV, Fig. A.
There is a certain amount of phosphatase in the nuclei of the
kidney tubule cells, particularly in the nuclei of the cells of the
proximal tubules. There is a very much greater amount of
phosphatase present in the brush border of the proximal tubule
cells. It is clearly quite possible that the phosphatase in the
nuclei may not be intrinsic phosphatase of the nuclei, but may
have reached the nuclei by diffusion from the brush-border
regions. It is, therefore, necessary to examine this material by

ALKALINE PHOSPHATES OF CELL NUCLEI 51

the superimposed-section technique described earlier in this
chapter (p. 37). Plate IV shows some of the results obtained
when kidney sections containing active enzyme were superim-
posed upon kidney sections in which the enzyme had been in-
activated. The sections were incubated for periods as follows:
5, 10, 20, 40, 80, 160, 320, 640, and 1280 minutes. Plate IV,
Fig. B, shows that the nuclei of the proximal tubules in the sec-
tion containing active enzymes are positive for phosphatase in
the case of the rat kidney after only 5 minutes' incubation.
Guinea-pig kidney in our experience has usually required 10
minutes' incubation to demonstrate the intrinsic phosphatase.
At this time no phosphatase has appeared in the inactivated
underlying section. Plate IV, Fig. C, shows the result obtained
after 320 minutes' incubation with the rat kidney. The section
containing active enzymes has produced so vigorous a reaction
for phosphatase that it is no longer possible to see all the detail
of distribution of the enzyme. The underlying section is still
devoid of phosphatase activity. Plate IV, Fig. D, shows the re-
sult obtained in the underlying inert section after 640 minutes'
incubation. The amount of phosphatase which is demonstrated
in the inert section is comparable with that found in the section
containing active enzyme after only 4 or 5 minutes' incubation.
When kidneys from different animals are studied in this way,
there is some variation in the time at which evidence of dif-
fusion of phosphatase is manifest. Sometimes in the guinea-
pig kidney phosphatase may appear in the underlying section
after only 320 minutes' incubation. The amount of phosphatase
which may have diffused into the underlying section after a
lengthy period of incubation (1280 minutes) is considerable.
With the material which we have investigated, we have never
found any considerable amount of diffusible phosphatase in rat
kidney revealed with incubations of 320 minutes or less. But
sometimes, after 1280 minutes' incubation with guinea-pig kid-
ney, there may be a very heavy reaction in the underlying sec-
tion. This reaction, of course, is entirely due to diffusion arte-
facts.

It is clear from these results that the phosphatase which is
seen in the nuclei of kidney after periods of incubation up to
160 minutes is intrinsic phosphatase. After twice this period
it is possible that part of the reaction observed is due to diffu-

52 STUDIES ON ALKALINE PHOSPHATASE

sion artefacts. But the amount of phosphatase which may have
diffused into the nuclei by this time is less than 5 percent of
the phosphatase which was intrinsically present in the nuclei.
After longer periods of incubation the amount of phosphatase
which may be present in the nuclei as a result of diffusion may,
of course, be much greater, so that the magnitude of the diffu-
sion artefact becomes increasingly serious. However, all the
detail of the distribution of phosphatase in mammalian kidney
is readily observed after only 10 or 20 minutes' incubation, so
that in this material there is no difficulty in eliminating diffu-
sion artefacts, so far as the distribution of alkaline phosphatase
is concerned.

The experiment mentioned above, however, shows quite clearly
that, where protracted incubations are carried out with mate-
rial which contains a rich source of phosphatase, the diffusion
artefact may become quite serious. Plate V, Fig. A, is a low-
power view of a liver section which was devoid of enzyme
activity, superimposed upon which was a guinea-pig kidney sec-
tion containing active enzyme. These sections were incubated
for 1280 minutes. It will be seen that the enzyme has spread
from the active section to a distance equal to more than seven
cell diameters of the kidney section. Some tissues, such as the
small intestine, when under normal physiological conditions con-
tain active diffusible phosphatase. It seems very probable that
with such tissues diffusion artefacts may be much more serious.
The work of Martin and Jacoby has, indeed, provided some evi-
dence that this is the case.

I have made a few experiments to try to ascertain the nature
of the substance diffusing from an active kidney section which
leads to the apparent appearance of phosphatase activity in the
underlying inactive section. The substances most likely to be
diffusing were calcium phosphate, alkaline phosphatase itself,
or an activator for phosphatase. As was indicated on p. 38,
if the diffusing substance is calcium phosphate, the rate of ap-
pearance of the artefact should be a function of the calcium
concentration. However, when the Gomori-Takamatsu method
for alkaline phosphatase was carried out in the presence of cal-
cium concentrations 10 times greater than, or only y10, that of
the normal, no significant difference was observed in the rate of
appearance of the diffusion artefact. Had the diffusing sub-

ALKALINE PHOSPHATES OF CELL NUCLEI 53

stance been calcium phosphate, the rate of appearance of the
artefact should have varied about a hundredfold with this varia-
tion in the calcium concentration. It is, therefore, clear that
the diffusing material is not calcium phosphate.

The next test was to ascertain whether the diffusing substance
might be alkaline phosphatase itself or an activator for alkaline
phosphatase. In these experiments the underlying sections were
either kidney sections or guinea-pig liver sections in which the
phosphatase activity had been destroyed by heating, by dipping
in hot water, or by exposure to dilute acid. Superimposed upon
the inactivated section was a 50 /* section of guinea-pig kidney.
The top section was allowed to remain in contact with the
underlying section for 2 days in a sodium barbitone-calcium ni-
trate solution (corresponding to the incubation medium but
without substrate for phosphatase). After that period the
superimposed section was removed and the underlying section
incubated by itself. Plate V, Fig. B, shows the result obtained.
There is an intense reaction for phosphatase in what was
originally material lacking phosphatase. Since the active sec-
tion was removed before any incubation in the presence of sub-
strate for phosphatase was carried out, it follows that the dif-
fusing substance cannot be calcium phosphate, cobalt phosphate,
or cobalt sulfide. It must be either phosphatase itself or a
substance present in the active section which can activate phos-
phatase in an underlying section.

So far it has not been possible to discover whether the diffus-
ing substance is phosphatase or activator. If a layer of col-
lodium or cellophane is inserted between the inactive section
and the active one, the amount of diffusion is reduced to zero.
But this experiment does not decisively distinguish between the
two possibilities.

When diffusion occurs from an active section into a liver
section, the pattern of distribution of the diffused material is
considerably different from that of the intrinsic phosphatase of
the liver section. Plate V, Fig. C, shows the phosphatase which
is intrinsically present in a guinea-pig liver section. It is mainly
present in the cell nuclei. Figure B of Plate V, on the other
hand, shows that when diffusion occurs the nuclei are only one
of the sites which take up the phosphatase reaction. It thus
seems probable that, unless guinea-pig liver contains a consider-

54 STUDIES ON ALKALINE PHOSPHATASE

able amount of an apoenzyme, which is not normally active
when the phosphatase reaction is carried out but which can be
activated by an activator which can diffuse from kidney sec-
tions, the diffusing substance must be alkaline phosphatase
itself.

Investigators who are not familiar with diffusion phenomena
may perhaps not be entirely convinced by the arguments given
above that the phosphatase of the nuclei of kidney cells is intrin-
sic phosphatase. We were fortunate in obtaining quite inde-
pendent evidence that there is intrinsic phosphatase in cell nuclei
by the use of another substrate. This is

N02< >N=N< >OP03H;

which was prepared in an investigation with Dr. A. Loveless.
This substance is the phosphate of a dye, which gave a reddish-
yellow solution. Our first preparations of this material were puri-
fied rather carefully and, to our surprise, were not readily hy-
drolyzed by alkaline phosphatase. A preparation by a different
route, which was not highly purified, on the other hand, was
readily hydrolyzed by alkaline phosphatase. When kidney sec-
tions were incubated in the solution of this impure ester, there
was a heavy deposit of red dye in the brush borders and the nuclei
of the sections, Plate VI, Fig. A. The dye is precipitated because
the free dye is much less soluble in water than is the ester with
phosphoric acid. We then set out to investigate why it was that
the more highly purified material did not give a phosphatase re-
action. It was quickly found that, when a little of the products
of hydrolysis of the dye phosphate were added to purified dye
phosphate, the purified material was readily split by the enzyme.
The split products which were able to facilitate the display of
enzyme activity were, of course, phosphate ion and free phenol.
By chance, we discovered that the enzyme present in the nuclei
was more readily activated by free phenol, whereas the enzyme
present in the brush borders was more readily activated by phos-
phate. When a kidney section was equilibrated with purified
dye phosphatase to which a trace of free phenol had been added,
we were sometimes able to obtain a reaction in the nuclei only.

PHOSPHATASES OF NUCLEI OF OTHER CELLS 55

Figure B of Plate VI shows such a section. The dye material
appears only in the nuclei of the cells. As in this experiment the
brush-border enzyme was inert, it follows that the nuclear enzyme
is intrinsic in the nuclei.

In obtaining such distinction in activity between nuclear and
brush-border phosphatases, we were, of course, fortunate to a
degree which we cannot expect to see repeated with other tissues.
We have also obtained evidence, by destroying enzymes by heat,
that nuclear phosphatase is less readily destroyed than is brush-
border phosphatase of rat kidney. There thus appear to be two
differences between the alkaline phosphatases in rat-kidney nuclei
and rat-kidney brush borders. It does not, however, necessarily
follow that the enzymes in these two sites are different. It may
be that the enzymes in these two sites are adsorbed upon different
materials: this might just possibly be sufficient to account for
the difference in behaviour.

The Phosphatases of the Nuclei of Other Cells

The work mentioned above dealt only with the alkaline phos-
phatase of the brush-border cells of the proximal tubules of mam-
malian kidney. In the papers of Takamatsu and Gomori pub-
lished in 1939 it was mentioned that phosphase is rather com-
monly demonstrated in a variety of nuclei. Willmer and Krugelis
independently showed in 1942, in a study of tissue cultures,
that in mitosis the greater part of the nuclear phosphatase be-
comes concentrated on the chromosomes. From these results
it was evident that phosphatase might be playing an important
role in some of the chemical activities of the nucleus.

The tendency for phosphatase to become concentrated on the
chromosomes during cell division seems to be a fairly common
phenomenon. It may be noted, for example, in regenerating
liver cells, in tumor cells, and in various embryonic tissues.
Plate VII, Fig. A, shows phosphatase present in a normal rat
liver. The main sites of activity in the nucleolus and the chro-
mocentres are clearly demonstrated. Figure B in the same
plate shows that when a liver is regenerating, the nuclear phos-
phatase is concentrated on the chromosomes during the nuclear-
division phase. This type of behaviour, i.e., the presence of alka-
line phosphatase on chromosomes, nucleoli, and chromocentres,
is not limited to mammalian tissues. It has been found in am-

56 STUDIES ON ALKALINE PHOSPHATASE

phibia (Danielli, unpublished) and fish tissues (Lorch, 1949), in
various invertebrate tissues (de Nicola, 1949; Danielli, 1950b),
and Mugard (1953) has found that in various Protozoa phos-
phatase is a common nuclear constituent. The nuclear phos-
phatase of Protozoa is often distributed in a manner similar to
the distribution of deoxyribose nucleic acid, as demonstrated by
the Feulgen reaction, but undergoes considerable quantitative
changes during the various phases of the life history of the
Protozoa concerned.

Alkaline phosphatase is also commonly found in tumors.
Plate VII, Fig. D, shows the distribution of phosphatase in a sec-
tion of Walker rat sarcoma. Here again the distribution of alka-
line phosphatase tends to follow that of the Feulgen positive
material on the chromosomes but differs from the Feulgen posi-
tive material in being present in very high concentration in the
nucleolus.

When tumor tissues are treated with mitotic poisons, there is
again a general tendency for the Feulgen reaction and for the
distribution of phosphatase to be similar, except so far as the
nucleolus is concerned. Plate VIII shows in Fig. A the distribu-
tion of Feulgen-positive material and in Fig. B the distribution
of alkaline phosphatase, in both cases in cells in which the mitotic
poison has prevented spindle formation, so that when the nuclear
membrane broke down the metaphase chromosomes have been
distributed throughout the cell. This abnormal behaviour of the
chromosomes is frequently followed by pycnotic degeneration,
and often at this stage there is a marked parallelism between the
Feulgen reaction and the phosphatase distribution, as may be
seen from Figs. C and D on the same plate.

The marked tendency for alkaline phosphatase to follow the
•distribution of Feulgen-positive material can hardly be devoid
of significance. Although the occurrence of alkaline phosphatase
in the nucleolus is an exception to this general rule, it must be
remembered that the nucleolus is a site of high concentration of
pentose nucleic acid. Interest in this parallelism has led Danielli
and Catcheside in 1945 and Krugelis in 1945 to study the distri-
bution of alkaline phosphatase on the polytene chromosomes of
Drosophila salivary gland nuclei. When salivary glands are sec-
tioned after fixation with alcohol, it is easy to demonstrate that
the nuclei contain considerable amounts of alkaline phosphatase,

PHOSPHATASES OF NUCLEI OF OTHER CELLS 57

and that by far the greater part of this is distributed in bands
on the chromosomes. It is difficult, however, to study the detail
of the arrangement of the bands in the sectioned material since
the nuclei are so packed by the chromosomes. When the material
is prepared by the standard procedure of the cytologist, i.e.,
squashing in 45 percent acetic acid, the phosphatase activity is
completely destroyed by the acetic acid. If, on the other hand,
the nuclei are fixed by fixatives which have comparatively little
action on alkaline phosphatase, such as 80 percent alcohol or
aqueous formaldehyde, the nuclei take on a rubbery consistency
and cannot be squashed. Thus it appears that fixatives which
will give a cytologically satisfactory specimen destroy the enzyme,
and fixatives which do not destroy the enzyme do not give a sat-
isfactory cytological preparation. It was, therefore, necessary
to compromise between the irreconcilable ideals of good fixation
and retention of phosphatase activity by using various concen-
trations of acetic acid for the preparation of squashes. The best
results obtained so far were with squashes prepared in 3.4 per-
cent acetic acid. The procedure of squashing was carried out
as rapidly as possible, and immediately after squashing the
cover slips were removed under 95 percent alcohol. After the
cover slip was removed, the material was left in 95 percent
alcohol for at least 30 minutes. Under these conditions most
of the material remained on the cover slip, and most of the
material was also either cytologically poor or had its alkaline
phosphatase destroyed. In occasional specimens, however,
moderately well-fixed chromosomes were observed which had
a high phosphatase activity. An example of this is shown in
Figs. A and B of Plate IX. Catcheside and Danielli found
that in the X chromosome it appeared that the type of distri-
bution of phosphatase was identical with that of the Feulgen-
positive material. The intensity of the phosphatase reaction
in the different bands did not appear to follow the intensity
of the Feulgen reaction in the same bands.

The parallelism between the distribution of phosphatase and
of what is presumed to be deoxypentose nucleic acid in the
known sites of genes has naturally led to the suggestion that
both of these substances may be integral parts of genes. An ex-
tension of this contention has been the suggestion that a gene
may be a specific array of enzyme molecules, whose synthetic

58 STUDIES ON ALKALINE PHOSPHATASE

activity is partly defined by the nature of the enzyme mole-
cules present, and partly by the spatial orientation of the
enzymes.

It must be borne in mind that it is not always possible to
demonstrate the presence of alkaline phosphatase in nuclei.
Some of the cells of some organisms may, when examined by
the techniques given above, be lacking in phosphatase either
at some point in their life history, as is often the case with the
Protozoa, or even throughout their life history, so far as in-
formation is available at present. But it must be remembered
that none of the cytochemical methods available at present
for the demonstration of alkaline phosphatase are as sensitive
as could be desired, and that failure to demonstrate the pres-
ence of phosphatase in a particular site does not mean that
it is absent from that site, but merely that activity cannot be
detected by the cytochemical technique. Thus, as a general
rule, no alkaline phosphatase is observed by the technique of
Gomori in red blood cells. Nevertheless this enzyme may be
readily detectible in red blood cells by biochemical techniques.

Although it is true that phosphatase has not been demon-
strated in the nuclei of all tissues, it is, nevertheless, true that
its occurrence is so common, and its distribution so striking,
that, had a cytochemical method for the demonstration of al-
kaline phosphatase preceded the discovery of the Feulgen
reaction, a large part of the literature on cell nuclei would almost
certainly have been written in terms of the distribution of this
enzyme.

If, however, there is a general significance to be attached to
nuclear alkaline phosphatase of the same magnitude of impor-
tance as that to be attached to the Feulgen-positive material, it
should also be possible to demonstrate alkaline phosphatase in
the nuclei of plant cells. My first efforts to demonstrate this
were a complete failure. No doubt many other workers have at-
tempted to find alkaline phosphatase in plant-tissue nuclei by
cytochemical procedures but, as is the case with so many nega-
tive results, have not recorded the fact in the literature. How-
ever, after some time I discovered (accidentally) that unlike
the alkaline phosphatase of most animal cells, the nuclear phos-
phatase of plant tissues, or at least of those plant tissues which
I have studied, is rather rapidly destroyed by alcohol. A few

SITE OF PHOSPHATASE IN NUCLEI 59

minutes' exposure to 95 percent alcohol causes the destruction
of practically the whole of the phosphatase of the nuclei of
onion and bean root tips. By working rapidly, however, it
has been possible to squash root tips in 80 percent alcohol, re-
move the cover slip immediately under 95 percent alcohol, allow
up to 30 seconds for fixation in the 95 percent alcohol, and then,
after removal to distilled water, find some phosphatase still
present in the nuclei. For the demonstration of alkaline phos-
phatase in plant nuclei three methods have been successful: the
method of Takamatsu and Gomori, using glycerophosphate as
substrate; the use of /?-naphthol phosphate as a substrate; and
the use of the dye phosphate of Loveless and Danielli as a sub-
strate. Plate X, Fig. A, shows an onion root tip, with glycero-
phosphate as substrate; Fig. B shows the distribution of phos-
phatase in an onion root tip, with /?-naphthol phosphate as sub-
strate.

It is clear that, whereas the physical properties of the enzyme
found in plant tissues and plant-tissue nuclei may differ con-
siderably from those of the animal enzyme, this enzyme is,
nevertheless, found in plant nuclei, and that it therefore may
well have as general a significance in nuclear processes as has
deoxypentose nucleic acid.

The Details of the Distribution of Alkaline Phosphatase
in Nuclei

Considerable caution must be exercised in the interpretation
of the detailed distribution of phosphatase in cell nuclei, par-
ticularly when the nuclei are small, or as far as the finer details
are concerned. When one is concerned with large structures
such as the bands of Drosophila chromosomes, or with the dis-
tinction between chromosomes and nucleoli, the method for the
detection of diffusion artefacts given earlier in this chapter is
quite adequate. I am by no means convinced that it is pos-
sible to interpret the detailed picture of apparent phosphatase
distribution found with the individual chromosomes of normal
somatic tissues, even when full use is made of the critical meth-
ods at present available.

When the methods of analysis of the extent to which diffusion
artefacts are present cannot be used, interpretation becomes
much more hazardous. This is the case, for example, in tissue

60 STUDIES ON ALKALINE PHOSPHATASE

cultures. Figures A to D of Plate XI show the distribution of
alkaline phosphatase obtained by the method of Takamatsu
and Gomori with chick osteoblast cultures. This work on tis-
sue-culture material shows a high concentration of phosphatase
in the nuclei. In the intermitotic nuclei by far the highest con-
centration of phosphatase is found in the nucleolus: the chromo-
centres are also positive for phosphatase. In early prophase the
greater part of the chromosome is positive for phosphatase, even
before the nucleoli have disappeared. When the nuclear mem-
brane breaks down and the spindle is formed, the chromosomes
retain practically the whole of the alkaline phosphatase activity
of the nuclei, but some activity appears on the spindles and par-
ticularly on the centrosomes of the spindles. The cytoplasm also
commonly shows a certain amount of phosphatase which appears
to be present in rather large cytoplasmic granules.

The pictures observed with tissue cultures are often so clear-
cut that it is tempting to take them at their face value. But it
has not so far been possible to study the extent to which the
results are complicated by diffusion artefacts. It is clear that
the highest centres of phosphatase activity in the tissue cultures,
namely, the nucleoli of intermitotic cells and the chromosomes
of mitotic cells, have a high intrinsic phosphatase activity. But
whether the phosphatase of the chromocentres, of the spindles,
and of the cytoplasmic granules is intrinsic or is in part or wholly
due to diffusion artefacts is at present in doubt.

It would, of course, be of the greatest interest if it could be
established that there is indeed alkaline phosphatase present in
the spindles and in the centrosomes. In sectioned material it is
usually not possible to demonstrate alkaline phosphatase in these
sites, but I have occasionally found it on the spindle and cen-
trosomes in rat tumor tissues. Further investigation of this prob-
lem will be of great interest.

Possible Functions of Nuclear Phosphatase

Until recently alkaline phosphatase was thought of as an
enzyme which was apparently hydrolytic in function, producing
phosphoric acid and alcohol from a phosphate ester. Axelrod and
MeyerhofT and Green have independently demonstrated in the
last few years that phosphatases may also act as phosphokinases,
i.e., they may catalyze the transfer of a phosphate residue from

FUNCTIONS OF NUCLEAR PHOSPHATASE 61

one organic molecule to another. In considering the possible
functions of phosphatase in cell nuclei, we must bear in mind
both of these activities of alkaline phosphatase. It must also
be borne in mind that the results obtained with Drosophila sali-
vary chromosomes indicate that alkaline phosphatase is found
associated with all the genes, or at least with a very large num-
ber of quite diverse genes, so that its activity cannot be restricted
to any very specific function.

If we consider the hydrolytic function of alkaline phosphatase
there are at least three possible functions for phosphatase in
nuclei. These are:

1. A protective function. An active cell contains a number of vigorous
phosphorylating agents such as adenosine triphosphate and acetylphosphate.
Exposure of the genes to any of these substances might well be hazardous,
and it may be that the function of the alkaline phosphatase of the chromo-
somes is to protect the genes against the action of these phosphorylating
agents.

2. A final stage in synthesis of a gene product may normally be dephos-
phorylation. For example, if we are to regard the typical gene products as
proteins, we must remember that it is quite possible that protein synthesis
involves the formation of a phospho-protein, and that the action of alka-
line phosphatase in the nuclei is to split off the phosphate residue from
the protein, and thus liberate an active product from an inactive precursor.

It is of interest to reflect that the occurrence of phospho-proteins in
milk and in the yolk of eggs need not necessarily denote a special synthesis
of a phospho-protein in these instances. It may instead be an indication
of the suspension of a dephosphorylation process, which is normally car-
ried out when the function of the nucleus is to pass an active protein into
the cytoplasm. That is, by the suspension of dephosphorylation, it may
be that proteins which would otherwise participate in the chemical reac-
tions proceeding in cytoplasm are protected against such activity and thus
can be stored.

3. A third possibility is that alkaline phosphatase is concerned in some
way with nucleic acid synthesis. Such a proposal has been made by several
authors. It is, however, far from clear in what way alkaline phosphatase
might be concerned.

If we consider alkaline phosphatase as a phosphokinase, there
are at least two possibilities involved. The first of these is that
synthesis of phosphate esters may be a normal function of the
nucleus, and that the function of alkaline phosphatase is to trans-
fer phosphate groups from precursor substances to the specific
molecules which are involved in the activities of genes. An
alternative function is that alkaline phosphatase might serve as

62 STUDIES ON ALKALINE PHOSPHATASE

a phosphokinase in the conservation of phosphate bond energy.
If gene products must be dephosphorylated to release the prod-
uct of the gene in active form, a considerable economy of the
energy of the phosphate bond would be effected if, instead of hy-
drolyzing the phosphate bond, the phosphate were transferred to
some other molecule: alkaline phosphatase could act as a catalyst
for such transfers.

It should, of course, be clear from what has been said above
that we cannot at present do more than speculate about the
function of alkaline phosphatase in the cell nuclei. There must
also be considerable doubt as to whether it is at present profitable
to speculate about this matter at all. At present we can deter-
mine the cytochemical localization of very few substances in the
cell. In the nucleus we can determine pentose and deoxypen-
tose nucleic acids and alkaline phosphatase with some degree of
accuracy. But we can be reasonably confident that there are a
vast number of other substances present in the nucleus. Until
we know more about these other substances, only by the greatest
of good fortune can any of our speculations concerning the func-
tions of phosphatase and nucleic acid be correct.

Phosphatase in the Nuclei of Ciliates

Dr. H. Mugard has studied a considerable variety of ciliates,
using the alkaline phosphatase technique. The results may be
exemplified by considering the two species, Ophryoglena atra
and Opalina ranarum Stokes. With Ophryoglena no phosphatase
is found in the starved animals. Immediately after the taking
of food, phosphatase appears in the vicinity of the food vacuoles:
this occurs within a few seconds of formation of the vacuoles.
Very shortly afterwards phosphatase may also be observed in
the macronucleus, micronucleus, and throughout the cytoplasm.
Feeding is commonly followed by encystment. Following encyst-
ment the concentration of phosphatase falls in the cytoplasm,
and the macronucleus becomes negative. The micronucleus con-
tinues strongly positive. But as divisions proceed within the
cyst the amount of phosphatase in the micronucleus appears to
diminish at every division.

With Opalina there is no phosphatase to be found in the cyto-
plasm unless division is occurring. The nuclei contain a con-
siderable amount of phosphatase, particularly in the chromo-

PHOSPHATASE AND DIFFERENTIATION 63

centres, and the chromosomes found during mitosis are strongly
positive.

Dr. Mugard inclines to the view expressed by Brachet, that al-
kaline phosphatase is related to the rate of deoxy nucleic acid
turnover. It is, however, difficult to reconcile this view with the
fact that alkaline phosphatase remains constant in amount in
the sea-urchin egg until gastrulation occurs, whereas a consider-
able synthesis of deoxy nucleic acid occurs before gastrulation.

A more extended study of the relationships among nucleic acids,
proteins, and alkaline phosphatases in Protozoa is required. The
situation is becoming somewhat complicated by the fact that it
appears unlikely that there is only one alkaline phosphatase. I
have found that there is often a difference in properties between
nuclear and cytoplasm phosphatase. Ross and Ely (1951) have
recently made the very much more striking observation that there
is a high tendency for substrate specificity with some nuclei. For
example, with Habrabracon the nucleus contained no enzyme cap-
able of reacting with glycerophosphate or with adenosine phos-
phate. On the other hand, adenosine diphosphate and triphos-
phate are both rapidly attacked by nuclear enzymes. They have
also found that, in plant tissues where there is an apparent ab-
sence of glycerophosphatase (possibly due to hypersensitivity to
fixatives), there is a vigorous reaction with certain nucleotides
and not with others. So far as the evidence goes at the moment,
we must contemplate the possibility that there may be many al-
kaline phosphates, which may be individually concerned with
either nucleic acid or protein synthesis, or with both.

Alkaline Phosphatase, Protein Synthesis, and

Differentiation

Formation of Fibrous Proteins.

Fell and Danielli (1943) made a study of the action of some
chemical warfare agents upon rat skin. In the course of this
work, Miss Fell observed that at the site of a small healing skin
lesion there was a very strong phosphatase reaction. A study
was therefore made of the distribution of alkaline phosphatase
in healing wounds. We observed that phosphatase occurred to a
considerable concentration in proliferating cells, and that as new

64 STUDIES ON ALKALINE PHOSPHATASE

collagen was laid down a high concentration of phosphatase ap-
peared apparently on the collagen fibres themselves. We also
observed, as Bourne had done earlier, that phosphatase was pres-
ent in the hair follicles and thus presumably was concerned in
some way with the formation of fibrous keratin. On the other
hand, we observed no indication that skin-keratin formation
was associated with phosphatase activity. The association of
formation of two fibrous proteins with alkaline phosphatase was
suggestive, though it did not necessarily follow that there was
any causal relationship between the two substances. To obtain
further leads on this problem, we collaborated with Dr. Kodicek
in a study of skin wounds in guinea pigs, both normal and vita-
min-C-deficient animals being used. We found that there was a
close parallelism between degree of vitamin C deficiency, the
rate of formation of collagen, and the rate of appearance of alka-
line phosphatase. We were surprised to find that the optimal
rate of wound healing required a dosage of 10 milligrams of vita-
min C per animal per day. Bourne (1943) had made some
similar observations on the healing of cavities in bones. He ob-
served that the first step was the formation of protein fibres in
the cavity, and that these protein fibres were very rich in alka-
line phosphatase.

We were unable to discover a causal connection between pro-
tein synthesis and alkaline phosphatase. It seemed, however,
possible that one of the best methods of pursuing the matter
would be to investigate the formation of other fibrous proteins
such as silk. J. R. G. Bradfield has pursued this matter. Al-
though he has not been able to establish the exact site of synthe-
sis of silk proteins, he was able to show that the secreting cells of
the glands which are actively engaged in secretion of silk protein
have in one region of their cytoplasm, bordering on the lumen
of the gland concerned, a high concentration of active phospha-
tase. On the other hand, Helen Brown, in a study of the forma-
tion of the protein which constitutes the byssus thread of Mytilus,
has been unable to detect an association of the protein with alka-
line phosphatase at any stage. This latter observation recalls
the lack of association of skin keratin formation with alkaline
phosphatase. On the other hand, in this connection it is notable
that the horny teeth of Myxine and the horns of deer involved
fibrous keratin formation associated with alkaline phosphatase.

THE MECHANISM OF BONE FORMATION 65

The Mechanism of Bone Formation

As was mentioned at the beginning of this chapter, Robison
was able to establish that there was a close correlation between
sites of bone formation and of the occurrence of alkaline phos-
phatase. Before bone could be formed, the mechanisms which
had to be involved and which must operate simultaneously in-
cluded the following:

(a) The formation of a matrix (cartilage in the case of cartilage bone).

(b) The formation of extracellular relatively indiffusible alkaline phos-
phatase.

(c) Development of a local concentration of phosphate esters.

(d) A process for reorganization of the initial amorphous precipitate of
calcium phosphate which is produced under the action of factors a, b,
and c, into compact bone.

One of the key points in this process must be production of
phosphatase at the right point at the right time. It seemed im-
possible on theoretical grounds that intracellular phosphatase
could be directly concerned in producing calcification.

The relationship between intracellular and extracellular phos-
phatase and calcification has been investigated by I. J. Lorch.
The development of a number of membrane and cartilage bones
was studied in trout larvae. With both types of bones, calcifica-
tion is invariably produced by the formation of extracellular
phosphatase. During the development of the larvae there are
waves of occurrence of alkaline phosphatase in different regions.
The phosphatase builds up to a considerable concentration and
then over a few days falls to a much lower concentration. In the
case of the formation of membrane or cartilage bone, there is a
wave of development of intracellular phosphatase which is asso-
ciated with the formation of the matrix. This intracellular phos-
phatase is not associated with bone formation and may largely
disappear, to be succeeded by a. second wave of formation of
phosphatase, a great part of which is found outside the cells.
This latter wave is followed by the deposition of extracellular
calcium phosphate.

Studies were also made on the development of dogfish larvae.
Factors a, b, and c were all found to be operative in the develop-
ment of the larvae. In fact the phenomena were very similar to
those observed in trout larvae at different stages of development:

66 STUDIES ON ALKALINE PHOSPHATASE

waves of alkaline phosphatase occurred at different sites, includ-
ing a wave of intracellular phosphatase during the formation of
cartilage. But there was no calcification of the cartilage until
extracellular phosphatase appeared in the cartilage.

In the calcification of fish scales and of teeth there is a similar
development of phosphatase prior to calcification (Thorell and
Wilton, 1945; Greep et ah, 1948; Lorch, unpublished). It thus
appears that the process of calcification, where the calcium is
laid down as calcium phosphate, is invariably produced by the
formation of extracellular phosphatase. The converse of this,
however, is not necessarily true. There are probably many sites
in which extracellular phosphatase is laid down in which no calci-
fication occurs. These sites normally include healing wounds,
though occasionally calcification will occur in a healing wound.
It also appears that failure to form bone in the dogfish is not
due to any deficiency in the calcification mechanism, but is due
to a lack of the reorganization of amorphous calcium phosphatase
into organized bone. This conclusion is, of course, readily com-
patible with the view that the cartilaginous skeleton of the dog-
fish is not primitive but has evolved by loss of one of the vital
processes of bone formation. On the other hand, this state of
affairs certainly does not prove that the cartilaginous fishes are
not primitive, since the state of affairs found in their skeletons
could represent either a loss of ability to reorganize calcium phos-
phate into bone, or a system in which this capacity has never
been evolved.

It should perhaps be mentioned that it is not by any means
clear by what mechanism a sufficient concentration of phosphate
esters is produced in the calcifying regions to secure liberation
of an adequate supply of phosphate by enzyme activity. It may
be that a sufficient quantity of phosphate esters is normally pres-
ent in the intercellular fluid. Some authorities have suggested
that, since the cells in the calcifying regions are commonly rich
in glycogen, what occurs is that sugar is stored in these cells and
released as glucose phosphate or some other phosphate ester at
the time of formation of calcium phosphate. This, however, can-
not be the whole story, since, in order to form the phosphate ester,
phosphate is also required. If, therefore, the function of glyco-
gen found in calcifying regions is to provide a store of glucose for
the formation of phosphate esters, it is also necessary that the

PHOSPHATASE AND DIFFERENTIATION 67

cell should store phosphate, and that the phosphate ester should
then be released in a wave over a brief period. It seems equally-
probable, so far as the evidence at present available is concerned,
that the glycogen is not directly concerned in calcification. In-
deed, it may even be that hydrolysis of phosphate esters (derived
ultimately from the blood stream) in the cartilage raises the
concentration of glycogen precursors to such an extent as to
cause glycogen formation in cartilage cells as an indirect effect
of calcification, i.e., that the phosphate moiety of the phosphate
esters is being deposited as calcium phosphate in the cartilage
while the other moiety of the molecule is being deposited as glyco-
gen in the cartilage cells.

Alkaline Phosphatase and Differentiation

From the data which have been presented above there have
been a number of hints that alkaline phosphatase may be con-
cerned in certain steps at any rate in the process of differentia-
tion. We know so little of the mechanism of differentiation that
it is difficult to formulate a detailed suggestion as to how such
an enzyme may be participating. The difficulty is enhanced by
the fact that the same enzyme appears to operate both as a hy-
drolytic enzyme and as a phosphokinase. But, although it is
not possible at the present time to produce a detailed physico-
chemical working hypothesis for the function of alkaline phos-
phatase in connection with differentiation, there is an impressive
body of evidence that there is such a connection. Moog (1944)
has shown that in the development of the chick there are waves
of formation of alkaline phosphatase at different stages in the
development of at least most tissues. Lorch (1949) has recorded
a similar finding in the development of dogfish and trout.
Yao (1950) has made a similar observation with Drosophila
larvae. Runnstrom and his colleagues have also shown that
there is a great increase in alkaline phosphatase concentration in
echinoderm eggs immediately after gastrulation. This last ob-
servation is particularly striking in that it is not until after
gastrulation occurs that a significant quantity of new antigens
are formed in the development of the echinoderm larvae. It is
at the same stage that nucleoli become prominent in the cells of
the developing larvae: Caspersson and his colleagues have pro-
duced a good deal of evidence to show that prominence of

68 STUDIES ON ALKALINE PHOSPHATASE

nucleolar activity is associated with protein synthesis, and, as
we have noted before, nucleoli are usually one of the most phos-
phatase-rich regions of a cell. On the other hand, it is not pos-
sible to maintain that there is a direct correlation between the
cytochemically demonstrable alkaline phosphatase and protein
synthesis in all cases. Thus, though some tumours are rich in
alkaline phosphatase, other tumours which are growing equally
rapidly, such as the Rous chicken carcinoma, contain relatively
little phosphatase. We must also recall that the formation of
skin keratin and of the byssus thread of Mytilus do not appear
to be associated with alkaline phosphatase. It is, of course, pos-
sible that there are amounts of alkaline phosphatase in the cells
carrying out syntheses of these last two proteins which are suffi-
cient for the purposes of the cell but which are too small to be
demonstrated cytochemically. Alternatively, it may be that
with certain proteins the function performed by alkaline phos-
phatase is taken over by a more labile phosphatase. Whatever
the explanation may be of the fact that presence of alkaline
phosphatase cannot always be demonstrated in connection with
protein synthesis, the fact remains that there is a very close cor-
relation of occurrence of alkaline phosphatase with protein syn-
thesis and with certain stages in differentiation. As Dale has
said in connection with the correlation between liberation of
acetylcholine at motor end plates and muscular contraction, it
can hardly be that the correlation exists simply to mislead in-
vestigators ! Further work, however, is essential in this field. As
we shall see towards the end of this chapter, it is conceivable
that the alkaline phosphatase is connected, not with synthesis
of proteins, but with their activities immediately after synthesis.
It may well be that studies on Protozoa along lines similar to
those of Mugard will yield vital information.

Alkaline Phosphatase and Secretion

Verzar (1936) suggested that the process of secretion of sub-
stances across cell membranes might be connected with phos-
phorylation. Wilbrandt obtained some evidence for this by the
use of enzyme poisons to inhibit secretion. His results have been
confirmed by numerous other workers. It has also been shown

ALKALINE PHOSPHATASE AND SECRETION 69

that at many, if not all, sites of secretion of colloids, a high con-
centration of alkaline phosphatase is present. There were, how-
ever, many unsatisfactory features of this work. In the first
place, the fact that enzyme poisons which interfere with phos-
phorylating processes also prevent secretion does not necessarily
tell us very much about the actual process of secretion itself.
The general metabolism of the cell is also based on phosphoryla-
tion, and consequently failure of secretory activity as a result
of a poisoning of the phosphorylating mechanisms may simply
indicate a general breakdown in cellular activities. Secondly, it
was difficult to allot a function to alkaline phosphatase in con-
nection with secretion, since it is fundamentally a dephosphory-
lating enzyme.

I pointed out in 1943 that, in any case, before a simple phos-
phorylating process can effect secretion it is necessary to have
specific phosphorylating and dephosphorylating enzyme systems
present in appropriate parts of cells. For example, for the secre-
tion of glucose from the kidney the requirements are those indi-
cated in Fig. 3. Consider the upper part of the figure, which il-
lustrates the necessary siting of various enzyme systems. Since
glucose phosphate permeates cells much more slowly than does
glucose, it is necessary to have a dephosphorylating system on the
external border of the tubule cells at A, so that any glucose phos-
phate in the tubule lumen is rapidly converted into glucose and
phosphate ion. The same difference between the permeabilities
of cells to glucose and glucose phosphate also determines the
intracellular siting of enzymes if secretion is to occur. If glucose
which has diffused into the tubule cells is converted immediately
into glucose phosphate, the tendency to diffuse back into the
tubule will be greatly reduced: consequently, a phosphorylating
system is required in the cytoplasm adjacent to the lumen at B.
The converse situation is required in the cytoplasm adjacent to
the tissue spaces: i.e., to accelerate passage through the cell mem-
brane glucose phosphate must be converted to glucose. Hence
there must be a dephosphorylating system at C.

As a result of this distribution of enzyme systems, a pattern
of concentrations of glucose and glucose phosphate will be built
up as indicated in the lower part of Fig. 3. The net result of the
action of the enzyme systems is that, inside the cell, there is a
relatively high concentration of glucose against the membrane

70

STUDIES ON ALKALINE PHOSPHATASE

adjacent to the tissue spaces, and an equivalent concentration of
glucose phosphate against the membrane adjacent to the lumen
of the tubules. As glucose penetrates the cell wall much more
rapidly than glucose phosphate, there will be a net transfer of
glucose across the cell, and thus transfer of glucose from tubules

Lumen of

< money tuDuie cen

Intercellular

tubule

space

i

% ^

^ ■•£>

5 "*~

>■ ~S>

i ^

C B

i -^ a
% i

Dephosphorylation 1 Phosphorylation

5 <5

A

l/>

o

Glucose P04

•*— ■

~

Glucose

^^^^ ******~

o

^vj-^

O

n

Glucose

•>—/

^ >v

s \. ^*—*"^

-«** ^^^*~"^

Fig. 3. The upper part of the figure shows the distribution of enzyme

systems required to provide a secretory effect. The lower part of the

figure shows the concentrations that will arise in the various parts of the

system as a result of the enzyme actions.

to tissue spaces may be effected, even against a concentration
gradient.

The mechanism detailed above considers only the secretion of
glucose by kidney; in principle, of course, it can be extended to
cover the secretion of almost any type of molecule.

The techniques for the study of alkaline phosphatase have
made it possible to see to what extent this enzyme was present in
the sites required by the working hypothesis indicated above.
Gomori and Takamatsu showed that alkaline phosphatase occurs
in the brush borders of the glucose-secreting kidney tubule cells
in a variety of mammals. This has also been found to be the

ALKALINE PHOSPHATASE AND SECRETION 71

case with representative members of all the other vertebrate
classes (see Moog, 1946, for a survey of the literature). The
only case where there is alleged to be an absence of alkaline
phosphatase is with the toad fish Opsanus, one specimen of which
was studied by Wilmer. On the other hand, a number of species
of aglomerular fishes were studied by Lorch and Danielli, who
found that alkaline phosphatase was present in the border regions
of the secreting cells in all cases. It seems possible that the obser-
vation upon Opsanus is in error; post-mortem changes occur very
readily in the kidneys of fish. Whereas the detailed morphology
of the excreting organs of invertebrates differs very markedly
from that of the vertebrate kidney, the invertebrate excretory
organs always contain regions which have either been proved
to be, or are suspected to be, carrying out secretory processes
closely analogous to those occurring in the vertebrate kidney.
In all cases where these organs have been studied, these "secre-
tory" regions are found to be rich in alkaline phosphatase at sites
functionally analogous to the sites in mammalian kidney which
are rich in alkaline phosphatase. This is the case for the ne-
phridia of a terrestrial triclad (Pantin and Danielli, 1950) , for the
green, gland of the crayfish Camburus (Kugler and Birkner,
1948), for the Malpighian tubules of the tick Rhodnius, and a
number of insect larvae (Bradfield, 1950; Yao, 1950).

Apart from the excretory organs, numerous other sites of
solute secretion contain alkaline phosphatase on the free border
of the secretory cells which is exposed to the fluid from which the
solute is to be removed. This is the case with the small intes-
tine, the choroid plexus, the prostate gland, the placenta, etc. (for
a review see Bradfield, 1950) .

Thus there is an impressive body of evidence that alkaline
phosphatase is associated with the selectively active borders of
a wide variety of secretory cells. But it is still not clear whether
the phosphatase is inside the plasma membrane or on the exterior
of the cell. These two possibilities must be considered separately.

If the phosphatase is on the outside of the cells, it is in the
position in which it can carry out the dephosphorylation A of
the mechanism described on p. 69. But the situation is not
particularly satisfactory, for there has so far been no indication
of the existence of the other postulated enzyme systems at B
and C.

72 STUDIES ON ALKALINE PHOSPHATASE

On the other hand, it may be that the alkaline phosphatase is
intracellular, and, in fact, acting as a phosphokinase in relation
to secretion; if this is so, the enzyme found on the borders of
cells must have the function of transferring phosphate molecules
from a precursor such as adenosine triphosphate to substances
such as glucose as soon as the latter penetrate into the cell, thus
preventing back diffusion. Much more study is needed of the
details of organization of this system before any further conclu-
sions may be drawn.

Secretion by Contractile Proteins

In the meantime a quite different suggestion for a mechanism
of secretion has been made by R. J. Goldacre. He has suggested
that secretion is carried out by the folding and unfolding of
protein molecules. His basic concept is that, when a protein
molecule is extended or unfolded, groups are exposed which are
able to adsorb specific molecules. Then, if later the protein
molecule becomes folded, the adsorbing groups may saturate
their affinities intramolecularly, so that the adsorbed molecules
become desorbed. Impressive evidence that this may occur has
been obtained by studying the adsorption of dyes on proteins.
For example, globular serum albumin adsorbs comparatively
small amounts of dyes, whereas when spread in a monolayer it
adsorbs large amounts. Similar observations have been made on
tobacco mosaic virus and other proteins. Consequently, we may
consider it adequately demonstrated that the folding of proteins
could be one of the basic processes in secretion. Is this process
one which can be reconciled with experimental studies of se-
cretion?

Now Osterhout has pointed out that a great deal of the data on
the secretion of ions is not readily compatible with the idea that
there is a rapid movement of ions through cell membranes by
thermal diffusion. The evidence is more in favour of a one-way
mechanism in which, for example, an ion such as potassium in
the environment combines with a molecule of the cell membrane,
following which the complex so formed diffuses across the mem-
brane and discharges the ion into the interior of the cell. In
some peculiar way no back diffusion occurs. With potassium, for
example, as the external potassium concentration is raised the
rate of transport of potassium decreases relative to the external

SECRETION BY CONTRACTILE PROTEINS

73

concentration, indicating that there is a limited supply of the
transporting molecules in the cell membrane.

If we consider Goldacre's proposal in the light of Osterhout's
work and of the studies on phosphatase, it appears that we can
meet the requirements of the situation by a model such as that in-

Adsorption_
centre

-*~e

U

d

+-U

Plasma membrane

-Contractile protein

^•Phosphatase or phosphokinase
enzymic centre

A.T.P. or other
phosphate ester

^~ Enzymic centre
Adsorption centre

0or
0OPO3H2

Fig. 4. Diagram to illustrate one of the possible mechanisms for transport-
ing the substance 0 across a membrane as the result of the action of a

contractile enzyme.

dicated in Fig. 4. This model is based upon the view that the
secretory activity is performed by the folding of a contractile
protein, and that the contraction is brought about by reaction of
the protein with a substance such as A.T.P. (adenosine triphos-
phate), as is the case with actomyosin. The phosphatase thus
represents an enzymic centre through which the energy of A.T.P.
is transferred to the contractile protein. In this model, as indi-
cated on the upper part of the figure, the initial configuration
of the protein has one part of an extended chain on the cyto-

74 STUDIES ON ALKALINE PHOSPHATASE

plasmic side of the membrane, and the other part of the extended
protein chain on the environmental side of the membrane. On
the environmental side there is a specific adsorbing group which
possesses a high affinity for the substance 6. If now A.T.P. comes
into contact with the intracellular part of the protein, the protein
contracts and folds, as indicated on the lower part of the figure.
This contraction retracts the part of the protein chain containing
the complex RO from the external surface of the cell into the
internal surface. On reaching the internal surface, the folding
of the protein becomes such that the affinity of the adsorbing
group can be satisfied by internal forces in the coiled protein
molecule, so that the group 6 is discharged into the interior of
the cell.

There are a considerable number of possibilities for the details
of such a mechanism. The contraction of a protein molecule
under the action of A.T.P. is, of course, already well established
in the case of the myosin-actin complex. One of the simplest
working hypotheses would be that combination of the protein
with A.T.P. results in coiling of the protein molecule so as to
carry out the transference of 6. If, as is the case with myosin,
the protein has phosphatase activity, the A.T.P. will eventually
be split and extension of the protein will occur again so that the
original configuration of the molecule is restored, and a fresh
cycle of transfer of 0 may occur. If the protein is acting as a
simple phosphatase, inorganic phosphate will be split off. On the
other hand, it may well be that the protein will act as a phos-
phokinase, in which case it may not be 6 but the phosphate of 6
(i.e., 0OPO3H2) which is split off. It might well be that the
process of splitting off a phosphorylated 6 and the extension of
the protein are simultaneous.

A mechanism such as that just suggested would provide a very
adequate explanation of the close correlation of alkaline phos-
phatase and/or phosphokinase with sites of glucose transfer.
That A.T.P. is involved would mean that the process is indirectly
linked with respiration and/or glycolysis.

The mechanism which has been proposed above can be of very
general application. The nature of the molecule which is trans-
ported would depend almost entirely on the affinity of the adsorb-
ing group on the protein. Thus it should not be difficult for such
a system, with appropriate variations in the adsorbing group, to

FUNCTIONS OF CONTRACTILE PROTEINS 75

transport such diverse molecules as sugar, potassium, amino
acids, etc. It should also be quite possible for a modified form
of this mechanism to secure removal of a substance from a cell,
instead of secretion of a substance into a cell.

It is clear that investigation of the actual existence of the
mechanism proposed above would be very much assisted if spe-
cific poisons were available for the alkaline phosphatases and
phosphokinases, particularly those which may react with phos-
phorylating agents such as adenosine triphosphate.

Generalisation of the Contractile Protein

Hypothesis

In this chapter alkaline phosphatase has been shown to be
present in a variety of cellular sites, and to be fulfilling a variety
of functions. It has been my endeavour to indicate that many
alternative explanations must be considered in each case. But,
to conclude this chapter, I wish to propose a unifying theory,
i.e., a theory which would permit alkaline phosphatase to be con-
cerned in a wide variety of activities, and yet to be acting in
fundamentally the same way in all cases. This theory is that the
phosphatase usually represents the enzymic centre of a contrac-
tile protein. The fundamental significance of the enzyme activity
is the ability to make the chemical energy of a high-energy phos-
phate bond available for contraction of a protein.

In the preceding section of this book we have seen that this
view, in conjunction with Goldacre's theory of the performance
of osmotic work, affords a most interesting approach to the prob-
lems of secretion. We shall now consider the significance of
phosphatase in relation to three other problems: (1) chromo-
somes, chromocentres and nucleoli; (2) collagen; (3) division
spindles.

1. One of the most striking features of the behaviour of chro-
mosomes is their contraction during mitosis and meiosis. The
fact that chromosomes commonly have alkaline phosphatase, and
that during mitosis they appear to be rich in phosphatase, ap-
pears to be compatible with the view that the contraction may be
brought about by a process fundamentally similar to that respon-
sible for the contraction of actomyosin. In the intermitotic

76 STUDIES ON ALKALINE PHOSPHATASE

nucleus the richest sites of phosphatase are the chromocentres
and particularly the nucleoli: in Drosophila salivary chromo-
somes the Feulgen-positive bands contain most of the nuclear
phosphatase. It is probable that in the living cell all these re-
gions are either in a steadily contracted state, or in a state of
intermittent contraction. This latter suggestion has many points
of interest. There is no reason to believe that intermitotic
chromosomes are inert except so far as purely chemical processes
are concerned. Contractility in active regions of chromosomes
might well play a significant part in the shedding of gene prod-
ucts, or in effecting the folding of synthesized polypeptide chains
into globular proteins.

2. The association of phosphatase with newly formed collagen,
first discovered by H. B. Fell, has evoked many speculations as
to its significance. To my view, none of these speculations have
acted as a beacon by which a path for further advance could be
plotted. But, if the significance of the association is that new
collagen is a contractile protein, a new way does become clear.
One of the most dramatic phenomena in the healing of a wound
is the contraction which occurs as new collagen is laid down.
This contraction occurs at the time at which the collagen is rich
in phosphatase. In vitamin C deficiency, although much fibrous
protein is formed in a healing wound, no phosphatase is found,
and no contraction occurs. If newly formed connective-tissue
protein is normally contractile, we may indeed expect the con-
tractility to be of marked physiological importance. For ex-
ample, in organizing the structure of bone, collagen contractility
may be involved.

3. Although alkaline phosphatase has been observed to be
associated with the centrosome and spindles, it has not yet been
demonstrated by critical studies that this represents intrinsic
phosphatase. But if there is in fact phosphatase in these posi-
tions in the living cell, we should again be provided with a new
route of inquiry into the nature of spindle action.

Thus it may be seen that, although the cytochemical study of
enzymes is at present conducted on a relatively crude basis, it
bears many signs of being a skeleton key which will open many
of Nature's locks. With increasingly critical usage, we are likely
to find in such studies a far more delicate weapon than are the
common methods of biochemistry.

REFERENCES 77

REFERENCES

Axelrod. 1948. J. Biol. Chem., 172, 1.

Bradfield. 1950. Biol. Revs., 25, 113.

Bourne. 1943. Quart. J. Exp. Physiol., 82, 1.

Brown, H. Personal Communication.

Caspersson. 1950. Cell Growth and Cell Function (Norton, New York).

Danielli. 1943. In The Permeability of Natural Membranes, by Davson

and Danielli (University Press, Cambridge).
Danielli. 1946. J. Exp. Biol, 22, 110.

1950a. Cold Spring Harbor Symposia, 14, 32.

19506. XIII0 Congres international de zoologie, 205 (Protar,
Paris) .

1951. Nature (in press).
Danielli and Catcheside. 1945. Nature, 156, 294.
Fell and Danielli. 1943. Brit. J. Exp. Path., 24, 196.
Goldacre and Lorch. 1950. Nature, 166, 497.
Gomori. 1939. Proc. Soc. Exp. Biol, New York, 42, 23.
Greep et al. 1948. J. Am. Dental Assoc., 86, 467.
Kodicek, Fell, and Danielli. 1945. Brit. J. Exp. Path., 26, 367.
Krugelis. 1942. J. Cell. Comp. Physiol, 20, 374.

1945. Genetics, 80, 12.
Kugler and Birkner. 1948. Physiol. Zool, 21, 105.
Lorch. 1949. Quart. J. Micro. Sci., 90, 183, 381.
Loveless and Danielli. 1949. Quart. J. Micro. Sci., 90, 57.
Martin and Jacoby. 1949. Nature, 168, 875.
Menten, Junge, and Green. 1944. J. Biol. Chem., 158, 471.
Meyerhoff and Green. 1950. J. Biol Chem., 188, 377.
Moog. 1944. Biol. Bull, 86, 51.
1946. Biol. Rev., 21, 41.
Mugard. 1953. Quart. J. Micros. Sci. (in press).
de Nicola. 1949. Quart. J. Micro. Sci., 90, 391.
Osterhout. 1951. Biol. Bull (in press).
Pantin and Danielli. 1950. Quart. J. Micro. Sci., 41, 209.
Robison. 1923. Biochem. J., 17, 286.
Ross and Ely. 1951. Exp. Cell Research, 2, 339.
Runnstrom. Personal Communication.

Saunders. 1936. The Aromatic Diazo Compounds (Arnold, London).
Takamatsu. 1939. Trans. Path. Soc. Japan, 29, 492.
Thorell and Wilton. 1945. Acta Path. Microbiol. Scand., 22, 5.
Verzar and MacDougal. 1936. Absorption from the Intestine (Long-
mans, Green, London).
Willmer. 1942. J. Exp. Biol, 19, 11.
Wilmer. 1944. Arch. Path., 87, 227.
Yao. 1950. Quart. J. Micro. Sci., 91, 79, 89.

CHAPTER ^

THE CRITICAL STUDY OF
THE CYTOCHEMISTRY OF ALDEHYDES

Introduction

Techniques will be considered which permit the study of the
following types of aldehydes and derivatives of aldehydes:

1. Free aldehydes. Many tissues contain substances which have a very
reactive aldehyde group. This generally forms part of a fatty molecule.

2. Another common component of tissues is aldehyde existing in the
form of acetal.

3. Most sugars contain potential aldehyde groups of sugars which, how-
ever, are masked by ring formation. If at any time the ring opens, the
aldehyde group becomes reactive.

4. In addition to the aldehyde groups just mentioned, which exist in-
trinsically in tissues, many recently developed cytochemical techniques
involve formation of aldehyde groups developed as a result of oxidation
procedures.

The most successful reagent for studying these compounds is
reduced fuchsin, which was introduced by Feulgen, first for the
study of nucleic acid and later for the study of free fatty alde-
hydes. The chemical properties of the different groups of alde-
hydes mentioned above permit distinctions to be made between
them. Thus the reaction between reduced and free aldehydes
is rapid and may be complete in 15 minutes or less. The acetal
aldehydes can be split so as to release free aldehydes by a brief
treatment with 0.1 N hydrochloric acid. There is a great deal
of evidence that some of the tissue acetals can be split by the
action of mercury salts. But, as mercury may have other actions
besides that of splitting acetals, it is better to avoid mercury, and
to use hydrolysis with dilute hydrochloric acid as the diagnostic
procedure for acetals. Some glycosides may also be split by an
appropriate hydrolysis. In the case of the deoxy sugars the alde-
hyde group of a sugar liberated from the glycoside will react in
the cold with reduced fuchsin: This latter reaction is, of course,

78

INTRODUCTION 79

the basis for the Feulgen reaction for thymonucleic acid. Other
sugars will react with reduced fuchsin at higher temperatures,
but this procedure has not yet been studied in connection with
cytochemistry. There is a sharp distinction between the hy-
drolysis conditions which will liberate free aldehydes from acetal
and that required for splitting most glycosides— the acetals are
liberated by a brief hydrolysis with 0.1 N acid in the cold, whereas
most glycosides need much more vigorous treatment. The action
of 0.1 N hydrochloric acid which is sufficient to complete the
liberation of aldehyde from acetal has no significant effect in
liberating purine from thymonucleic acid. Consequently, the
two steps in hydrolysis give a clear-cut separation of these two
types of aldehyde.

It is very important with all these techniques to avoid condi-
tions under which oxidation of a tissue component may occur, for
then adventitious aldehyde groups may appear in the tissue.
Protection against this may to a large degree be secured by using
a reducing substance such as formaldehyde as a component of
the fixative, by carrying out the steps prior to exposure to re-
duced fuchsin under anaerobic conditions as far as possible, and
by avoiding the presence of oxidising substances such as mer-
cury and dichromate in the fixative. There seems to be no ad-
vantage gained by using oxidising agents other than periodate
when it is wished to oxidise groups in tissue, Periodate is a fairly
specific oxidising agent for glycols and /^-hydroxy amines. Other
oxidising agents which have been used, such as mercury and per-
manganate, are not sufficiently specific in their action to be very
valuable in cytochemistry.

In the examination of tissue sections for aldehydes we must,
of course, conform with the usual requirements of cytochemistry.
Thus it is necessary to demonstrate:

1. That the aldehyde is present in fixed tissue in its physiologically
normal position.

2. That the colour in the cytochemical reaction produced is indeed spe-
cific for aldehydes.

3. That no significant degree of destruction of aldehyde occurs during
the procedure.

4. That the colour developed in the cytochemical reaction remains in the
physiologically normal site of the aldehyde, and has not diffused to some
other part of the tissue.

80 CYTOCHEMISTRY OF ALDEHYDES

Standard Technique

In any detailed investigation it is, of course, necessary to em-
ploy a considerable number of variations in technique. But it is
convenient in making preliminary investigations to have a stand-
ard technique. The technique which will be given demonstrates
the sum of free aldehyde and acetal aldehyde. Frozen sections
must be used. The normal procedure of infiltration with wax
will remove practically all the free fatty aldehyde and fatty
acetal aldehyde. It is this which enables the Feulgen technique
for thymonucleic acid to be carried out without masking by other
aldehydes. The steps for demonstration of the sum of acetal
and free aldehyde are as follows:

1. Sections of tissue not more than 2 millimeters in thickness are fixed
in a solution containing 8 percent formaldehyde and 5 percent acetic acid.

2. After a minimum of 2 hours' fixation, frozen sections are cut.

3. The sections are washed with distilled water and then placed in cold
0.1 N hydrochloric acid for not more than 15 minutes. This liberates
aldehyde from any acetal aldehyde present.

4. The sections are washed in distilled water to remove hydrochloric

acid.

5. Sections are transferred to reduced fuchsin for 15 minutes.

6. The sections are given three washings, each of 5 minutes, in a solu-
tion containing sulfur dioxide. The sulfur dioxide solution can be pre-
pared conveniently by mixing 10 milliliters of 10 percent sodium bisulfite
solution with 10 milliliters of N hydrochloric acid, and adding distilled
water to 200 milliliters.

7. The sections are washed in distilled water and mounted in glycerol
or Canada balsam.

The fixative has been chosen because it gives a reasonable
compromise between good fixation of cytoplasm and good fixa-
tion of the nucleus in many tissues. Other tissues may need a
modified form of this fixative: for example, it may perhaps be
necessary to make the fixative up in sea water for the tissues of
marine animals. The high formaldehyde content is valuable in
protecting free aldehyde against oxidation. The whole procedure
up to step 5 can also be carried out when necessary under an-
aerobic conditions. This is most conveniently done by carrying
out the treatment with fixative, etc., with the material in a
Thunberg tube. The frozen sections can, if necessary, be cut
in a closed box in an atmosphere of nitrogen. It is best to cut
sections as soon as possible after fixation is complete. With

STUDY OF THE ACTION OF FIXATIVES 81

some material, such as rat liver, a prolonged period of fixation
does not seem to result in the production of artefacts. But as
Cain and others have emphasized, artefacts probably due to oxi-
dation arise rather readily in some tissues; I have found this
to be the case with a number of tumor tissues.

Cain has suggested that a better technique is to use very small
pieces of tissue which are placed, without previous fixation, in a
mixture containing mercury and reduced fuchsin. With the tis-
sues which I have studied, fixation under these circumstances
tends to be very poor. But in any case there is always, with tech-
niques of this type in which the colour due to a cytochemical re-
action is produced in a block of tissue, the fundamental objection
that it is impossible to study the extent to which the results are
invalidated by diffusion phenomena. It is my view that a cyto-
chemical technique which does not permit the evaluation of diffu-
sion artefacts is useless.

After a preliminary survey of a tissue made by the procedure
given above, the other points mentioned in the introduction to
this chapter may then be studied as seems appropriate.

Study of the Action of Fixatives

Where the substance which is being studied can safely be
passed through an infiltration with wax, freeze-drying is the best
method of fixation. But, for many tissue aldehydes, organic
solvents must be avoided since they dissolve fatty aldehyde ma-
terial from the tissue.

When freeze-drying cannot be employed it is necessary to use
a variety of fixatives to investigate the degree to which the dis-
tribution of aldehyde is produced as a fixation artefact, as dis-
cussed in Chapter 2. It is possible to use acid, neutral, and alka-
line fixatives, reducing fixatives, and preprecipitation with strong
salt solutions in the study of fixation artefacts. There are, of
course, limitations encountered in the use of this variety of fixa-
tives with particular types of aldehyde. Thus an acid fixative
is likely to liberate aldehyde from acetals, after which acetal
adehyde and free aldehyde are, of course, indistinguishable.

As an example of the use of this method, we may consider
studies which have been made of the fatty aldehydes of liver.
When this material is placed in fixatives containing mixtures of
formaldehyde, acetic acid, mercuric chloride, pyridine, and tri-

82 CYTOCHEMISTRY OF ALDEHYDES

chloracetic acid, either before or after preprecipitation of the
tissue with saturated calcium chloride, magnesium sulfate, or
ammonium sulfate, no variation in the distribution of aldehyde
was found except with those fixatives containing trichloracetic
acid. With the trichloracetic fixatives, after treatment with re-
duced fuchsin, some degree of nuclear staining was found which
was not encountered with the other fixatives. It seems likely that
this staining is caused by hydrolysis of the nuclear thymonucleic
acid, since trichloracetic acid is a fairly strong acid. Apart from
the nuclear staining, the aldehyde contained in the liver after
treatment with these different fixatives is always in the same
position, and it can always be removed by extraction with ace-
tone. It therefore seems justifiable to conclude that this alde-
hyde is not shifted from its physiological position by the action
of the fixative, with the reservation that strongly acid fixatives,
such as those containing trichloracetic acid, must be avoided.

Specificity of the Detection of Aldehydes

Of the naturally occurring compounds, there is no evidence
that any other than aldehydes will react with reduced fuchsin to
give a purple colour. Some unsaturated substances have been
alleged to react, but the reaction appears to be due to the pres-
ence of aldehydes or other products of oxidation, and not to the
original substance. This is the case with oleic acid. A few sub-
stances, such as those formed by atmospheric oxidation of sub-
stances containing double bonds, may perhaps be capable of
reacting with reduced fuchsin to give colours similar to those
given by aldehydes. But there is no evidence that such sub-
stances normally occur in tissues in an amount which would be
detected in a cytochemical study. However, there always exists
the possibility that some unusual substance may be encountered
or that a reaction with reduced fuchsin may occur as a result of
the presence of enzyme systems. Consequently, a number of sup-
plementary tests which can be used to check the nature of the
reacting substance are desirable. A number of reactions have
been tried for this purpose. They include the following pro-
cedures.

1. A useful distinction can be made between substances which are sol-
uble in fat solvents and those which are not. For example, it is important
in the study of fat metabolism to be able to distinguish between the long-

DESTRUCTION OF ALDEHYDE 83

chain fatty aldehydes and aldehydes which may be present in sugars, etc.
Extraction with acetone is a convenient way of distinguishing between
these two classes of substances. It is necessary, however, to carry out the
extractions on tissue sections, for with thicker pieces of tissue very pro-
longed treatment is necessary to secure complete removal of lipoidal
aldehyde.

2. Another substance which reacts with aldehydes in a fairly specific
manner is azobenzene phenylhydrazine sulfonic acid. This substance re-
acts with aldehydes to give a purple colour. The purple colour is fully de-
veloped only in rather strong acid; consequently permanent mounts can-
not be made. But the colour reaction is very useful for checking results
obtained by the reduced-fuchsin method. This reagent also reacts with
ketones, but not to give a purple colour.

3. Aldehydes readily reduce alkaline silver solutions to give a black
precipitate of metallic silver. This reaction, however, is not very sensitive
and can only be expected to be found at sites containing a high concentra-
tion of aldehyde. Ability to reduce alkaline silver solutions is not, of
course, restricted to aldehydes: many substances will carry out this re-
duction. On the other hand, any substance which is an aldehyde must
reduce alkaline silver.

4. A very useful group of reagents introduced by Bennett (1940) are the
hydrazines. Phenylhydrazine itself does not give a deeply coloured com-
pound with aldehyde groups, and indeed it is possible that it may be a
useful compound for blocking free aldehyde groups. But 2:4 dinitro
phenylhydrazine gives a fairly deep yellow colour with aldehydes. This re-
agent will also react with keto groups and is, therefore, not to be regarded
as specific for aldehydes. Camber (1949) has introduced another hydra-
zine which can be used for obtaining more intense colouration, using tech-
niques similar to those described in Chapter 5.

5. A very useful procedure is to block the aldehyde group by prior
treatment with hydroxylamine or dimedone. When treated with one of
these reagents before treatment with the reduced fuchsin, the aldehyde
group is blocked up and no purple colour develops in the reduced fuchsin.
It is probable that for many purposes dimedone is the better substance
to use as a blocking agent, for L. G. Bell has shown that during the re-
action with hydroxylamine the reagent solution may remove considerable
amounts of the protein from the tissue section. Dimedone can be used
in alcoholic solution, in which proteins do not dissolve readily.

Details of the reactions given above are provided elsewhere
(Danielli, 1949a).

Destruction of Aldehyde

It is always desirable to be sure that the aldehyde which is
being studied in an experiment is not being destroyed by a par-
ticular step in the experimental procedure. This can be done
simply by increasing the duration of each step in the procedure

84 CYTOCHEMISTRY OF ALDEHYDES

by a factor of, say, fourfold. If there is no significant difference
in the intensity of the reaction which is found, there cannot be a
significant loss of activity in any of the steps. Thus with liver
tissue it has been found that the procedure given on page 80 does
not cause any destruction of liver aldehyde.

The Detection of Diffusion Artefacts

The amount of attention which has been paid to the elimina-
tion of diffusion artefacts in the cytochemistry of aldehydes is
very small. Indeed, it may be said that most investigators have
displayed a remarkable resistance even to considering the possi-
bility that their results may be complicated by diffusion arte-
facts. Stedman and Stedman (1943) performed a very valuable
service when they drew attention to the fact that the procedure
of hydrolysis in N hydrochloric acid depolymerizes thymonucleic
acid and suggested that this depolymerization may perhaps in
some cases be sufficiently serious to make the nucleic acid dif-
fusible. Following the observations of the Stedmans, I was able
to show that, when purified thymonucleic acid is treated by the
Feulgen method, a readily soluble fraction is formed. The pro-
cedure was simply to suspend thymonucleic acid in N hydro-
chloric acid at 60° C. After hydrolysis for 10 minutes the sus-
pension was filtered, the filtrate neutralized, and reduced fuchsin
solution added. A deep purple colour immediately developed.
This coloured solution readily stains tissue sections and such prep-
arations as squashes of Drosophila salivary glands. The distri-
bution of stain in these specimens is very similar to that found
in the Feulgen reaction, except that the nucleolus is more deeply
stained than is normally found, and there also may be some
staining of the cytoplasm. It is, therefore, difficult to be certain,
without the use of diffusion studies, whether the results obtained
by the fuchsin technique for thymonucleic acid are complicated
by diffusion artefacts. This is particularly true so far as the
fine detail of distribution of thymonucleic acid is concerned. It
would not be difficult in any given instance to discover whether
a significant amount of diffusion does in fact occur, if the super-
imposed section technique, referred to in earlier chapters of this
book, were used. But I am not aware of any instance in which
this procedure has in fact been used. It is likely that any diffu-
sion artefact which does occur would be eliminated, or at least

DIFFUSION ARTEFACTS 85

greatly reduced in magnitude, if the various steps in the pro-
cedure of Feulgen are, as far as possible, carried out in the pres-
ence of a heavy metal such as mercury. The heavy metal salts
of the nucleic acids are very much less soluble than the free acids
themselves.

A somewhat similar situation has been found with the long-
chain fatty aldehydes. If a piece of liver tissue is ground with
fat solvents, such as acetone, aldehyde present in the liver goes
into solution in the acetone. If this acetone is now quickly
shaken with a large volume of water, a colloidal solution of the
aldehyde is obtained. On addition of reduced fuchsin to this
solution, a deep purple colour develops. When tissues contain-
ing no aldehyde are exposed to this solution, the purple colour is
tak^n up vigorously by the tissues. The distribution of colour
so found, however, differs remarkably from that found for alde-
hyde in most tissues, in that the nuclei tend to be deeply stained
whereas the cytoplasm is usually not very well stained. Thus,
if liver tissue is stained for the aldehyde intrinsically present,
practically all the aldehyde is found in the cytoplasm. If, how-
ever, all intrinsic aldehyde is removed from a liver section by
treatment with acetone, and this section is then stained by the
dye solution mentioned above, the staining of the section is
largely retricted to the nucleus. Thus with the fatty aldehydes
it is usually not difficult to demonstrate that the results are not
complicated by diffusion artefacts, since the sites having a high
affinity for the cytochemical reaction products differ signifi-
cantly from the sites in which the reaction products are in fact
found.

The use of superimposed sections provides a further very ade-
quate check on the diffusibility of the aldehydes and their cyto-
chemical reaction products.

Procedures for Distinguishing between Free Aldehyde, Ace-
tal Aldehyde, and Aldehyde Formed by Oxidation

When it is desired to distinguish between free aldehyde and
acetal aldehyde, it is necessary to fix with a neutral fixative such
as neutral formaldehyde solution. Table VI shows the general
plan of procedures, using material which has been fixed in a neu-
tral fixative.

8Q CYTOCHEMISTRY OF ALDEHYDES

TABLE VI

Distinction between free aldehyde, acetal aldehyde, and aldehyde lib-
erated by oxidation. The numbers indicate the order in which the steps
should be carried out to obtain a reaction for one only of the three types
of aldehyde.

Free Acetal Oxidation

Aldehyde Aldehyde Aldehyde

Cold 0.1 iVHCl .. 2 1

Oximation . . 1 2

Oxidation . . . . 3

Reduced fuchsin 13 4

To demonstrate free aldehyde only, the only procedure which
is necessary after cutting frozen sections is to expose the sections
to reduced fuchsin. This exposure should be of as short a dura-
tion as possible, and care must be taken that the fuchsin solution
is not too acid in reaction. It frequently happens that part of the
sulfur dioxide present in reduced fuchsin solution becomes oxi-
dised to sulfuric acid, thus making the solution much more acid
than it should be. It may be that, when free aldehyde alone is
to be demonstrated, it is best to use a buffered solution of re-
duced fuchsin. However, I have not investigated this point.

To demonstrate acetal aldehyde only, the first step is to
block up the free aldehyde with either hydroxylamine or dime-
done. When this has been done, aldehyde is liberated from ace-
tal by treatment with cold 0.1 N hydrochloric acid, and the alde-
hyde so liberated is coloured by treatment with reduced fuchsin.

To demonstrate aldehyde formed by oxidation procedures, it is
first desirable to eliminate both free aldehyde and acetal alde-
hyde. The procedure in such a case is first to expose to cold
0.1 N hydrochloric acid, to liberate aldehyde from acetal. Then
all the liberated and free aldehyde is blocked by treatment with
hydroxylamine or dimedone. After this, the oxidation procedure
may be carried through and then the aldehyde formed by oxida-
tion may be demonstrated by treatment with reduced fuchsin.

Many of the papers on the cytochemical demonstration of alde-
hydes are extremely difficult to interpret, owing to the failure of
the authors to distinguish clearly between free aldehyde, acetal
aldehyde, and aldehyde formed by oxidation, both adventitious
and intentional. Part of the difficulty has been, as Feulgen and

PROCEDURES FOR ALDEHYDES 87

Voigt found, that acetal aldehyde in tissues is rather readily
broken down to free aldehyde by mercury solutions. It is prob-
able that mercury and other heavy metals can sometimes acceler-
ate oxidation processes, so that when mercury is used an arte-
fact is introduced, owing to oxidation of substances in tissues
which would not otherwise give an aldehyde reaction. This diffi-
culty is not encountered if fixatives and reagents devoid of mer-
cury are used, as in the procedure described above.

The situation has also been somewhat complicated by the claim
of Dempsey, Bunting, and Wislocki that a-hydroxy-£-keto-
steroids may also be oxidised by cold aqueous mercuric solution
with the formation of aldehydes. This is very unlikely to occur.
Boscott, Mandl, Danielli, and Shoppee showed that when pure
deoxycorticosterone is treated with aqueous mercuric chloride
there is no indication of oxidation of the steroid. Deoxycorticos-
terone is a typical a, /^-hydroxy ketone. We may, therefore, say
with confidence that typical a,/?-hydroxy steroids will not be oxi-
dized in the manner postulated by Dempsey et al. There re-
mains, of course, the possibility that there may be some very
atypical hydroxyketones which would react with cold aqueous
mercuric chloride. However, if the procedure given above (pp.
85 and 86) is carried through there is no danger of complication
from such atypical compounds. It must, however, be empha-
sized that the claim that hydroxyketosteroids have been demon-
strated in tissues such as the adrenal gland, by the use of mer-
cury solutions as an oxidising agent, must be discounted until
the presence of atypical substances is in fact demonstrated by
orthodox chemical means.

It should perhaps again be emphasized that one of the most
important control experiments in the study of aldehydes is to
carry a sample of tissue through the various steps of fixation
etc., under anaerobic conditions,, to eliminate the formation of
aldehydes, or substances which may react like aldehydes, by at-
mospheric oxidation of tissue components. Cain, following the
observations of Gomori and other workers, has quite correctly
drawn attention to the danger of encountering artefacts due to
atmospheric oxidation. But the procedures suggested by Cain
do not offer any satisfactory means of eliminating both this
and other artefacts. The procedure recommended above on the

88 CYTOCHEMISTRY OF ALDEHYDES

other hand should enable all the various artefacts to be dis-
tinguished clearly.

Aldehydes in Fat Transport and Metabolism

Introducton

Feulgen and Voigt (1924) showed that aldehydes of fatty
type were to be found in many biological materials. My interest
in these substances was aroused by the fact that the Rous chicken
sarcoma virus is said to contain a considerable quantity of acetal
aldehyde. Thus there appears to be a possibility that Rous virus
could be localized in cells by analysis of the distribution of alde-
hyde. A preliminary examination of this possibility was made
together with Dr. L. M. J. Rinaldini. We found that the situa-
tion was complicated by the fact that the Rous tumour is char-
acteristically heavily infiltrated with fat, and that there appeared
to be an association between fat droplets and sites of high alde-
hyde activity, particularly in the necrotic areas. As will be seen
later, there can be no doubt that aldehydes are intimately con-
cerned with fat metabolism, so that the occurrence of long-chain
aldehydes in the Rous tumour can hardly be taken as diagnostic
for the localization of virus. It may be, of course, that sites con-
taining the aldehyde are also rich in virus, and that one of the
major consequences of infection of cells with the Rous virus is
a disturbance of fat metabolism.

Comparatively little is known of the steps involved in the syn-
thesis and degradation of the higher fatty acids, such as palmitic
or stearic acid. It is known that the naturally occurring fatty
acids almost invariably contain an even number of carbon
atoms and that, when the fatty acids are degraded, they usually
lose carbon two atoms at a time. A suggestion which has, there-
fore, been considered in connection with fat synthesis is that it
proceeds by the condensation of aldehydes, with acetaldehyde as
the main condensing agent, in an aldol reaction:

R-CH2.CHO + CH3-CHO -> R-CH2-CHOH-CH2-CHO

The biochemists have not found it possible to find acetalde-
hyde performing this function in tissues. However, it is quite
certain that some 2-carbon compound, perhaps acetyl phosphate,

FAT METABOLISM IN THE LIVER 89

or a free radical related to acetyl phosphate, is a main inter-
mediate in the synthesis and degradation of fats. It is, therefore,
of considerable interest in connection with fat metabolism as a
whole to consider the function which may be performed by long-
chain aldehydes. These substances have been unduly neglected
by the biochemists, possibly because they so readily undergo
polymerization and oxidation when treated by the common
methods for examination of fats.

It was decided to study this problem in the liver of rats and
mice. A concurrent study was also made of aldehyde in rela-
tion to absorption of fat from the digestive tract.

The material used in these experiments was fixed either in neu-
tral formaldehyde or in 8 percent formaldehyde -f- 5 percent
acetic acid.

Absorption of Fat from the Digestive Tract

Mice were fed with either emulsified olive oil or emulsified
oleic acid ; they were killed at intervals up to 7 hours after feed-
ing; and various sections of the small intestine were examined.
The fat droplets found in the intestinal epithelial cells were usu-
ally completely lacking in aldehyde, especially when a consider-
able amount of fat had been fed to the animals. When mice were
fed on a diet containing a significant but small amount of fat, the
occasional droplets of fat which are found in the cells often con-
tain a considerable amount of aldehyde, and there may be a con-
siderable amount of aldehyde distributed diffusely in the cyto-
plasm of the epithelial cells. But, when there is a considerable
flow of fat droplets through the cells, no aldehyde can be observed.
It thus seems clear that long-chain aldehydes are not concerned
with either the absorption of fat or fatty acid into the cells, or
with the transport of fat or fatty acids through these cells.

Fat Metabolism in the Liver "

The results obtained depended somewhat on the diet. When
the diet contained fat, protein, and carbohydrates, considerable
areas of the liver, as seen in tissue sections, contained a diffuse
distribution of aldehyde in the cytoplasm. There appeared to be
no free aldehyde in the nucleus. When the cells contained fat
droplets, each droplet appeared to be surrounded by a spherical
shell containing a high concentration of aldehyde. There were

90 CYTOCHEMISTRY OF ALDEHYDES

also large areas of the liver sections which appeared to be free
from aldehyde. These areas were also lacking in fat droplets.

When the animals were placed on a diet rich in fat, there was
an increase in the number of fat droplets to be seen in the
hepatic cells, and a concomitant increase in the amount of alde-
hyde found in the liver. All the droplets were surrounded by a
spherical shell of aldehyde (e.g., Plate XII, Fig. A).

Some animals were also starved up to 6 days, although allowed
free access to water. It is known that under such conditions fat
passes from the fat depots of the body to the liver, where it is
broken down into carbohydrates. It was found that under these
circumstances the liver becomes heavily infiltrated with fat, and
gives an intense aldehyde reaction, with a high concentration of
aldehyde as a spherical shell surrounding the fat droplets. This
result and the result recorded in the preceding paragraph make
it clear that long-chain aldehyde is closely connected with the
conversion of fat to carbohydrate.

Another group of animals was fed on a diet containing protein
and a high concentration of carbohydrate, but no fat. It is
known from isotope experiments that under these conditions no
fat is lost from the fat depots, and that fat is synthesized in the
liver from carbohydrate. When animals were killed at various
times after being placed on this carbohydrate-rich diet, it was
found that the initial response was a sharp decline in the amount
of aldehyde in the liver, together with the loss of fat drops from
the liver. After 2 or 3 days on this diet, a considerable amount
of aldehyde appeared diffusely distributed in the cytoplasm. Fat
droplets reappeared and were closely associated as a general rule
with a spherical shell of aldehyde. Some fat droplets, however,
did not contain aldehyde. On this diet aldehyde was also fre-
quently found in the bile canaliculi. It was concluded that it
was probable that there is also a close connection between long-
chain aldehydes and fat synthesis, although the evidence for this
is not quite so strong as is the case for the relationship between
fat degradation and long-chain aldehydes.

All the experiments reported above, both those on the liver and
those on the intestine, were performed on material which had
been fixed for little more than 2 hours. The intensity of the alde-
hyde reaction was not influenced by conducting fixation, etc.,
under anaerobic conditions.

NATURE OF FAT SYNTHESIS 91

The Nature of Fat Synthesis and Degradation

The following group of facts needs to be considered in rela-
tion to fat synthesis and degradation:

1. Long-chain aldehydes are intimately connected with fat metabolism
in the liver.

2. The fat droplets found in tumours are also commonly associated with

long-chain aldehyde.

3. When fat appears in tissue cultures the droplets are commonly sur-
rounded by a shell of aldehyde.

4. Biochemical evidence shows that changes in fat are mainly mediated
by loss or addition, not of CO2, but of 2-carbon fragments.

5. Whereas free fatty acid is not attacked by liver extracts, the acyl
phosphates are readily attacked by such extracts (Lehninger, 1945).

There are a number of questions the answers to which are far
from clear at the present time in connection with the function
of the aldehyde. Thus we do not know whether the aldehyde
involved occurs as a true intermediate, as an end-product or as a
by-product of processes occurring in the liver. If it is a true
intermediate, we must envisage a series of reactions in which
aldehydes, or compounds such as acetals which readily give rise
to aldehydes, are involved in synthesis and degradation of at least
the higher fatty acids (Danielli, 1950b), and occur as part of the
process of ^-oxidation.

It must, of course, be pointed out that we do not know whether
the aldehyde which is involved in fat metabolism occurs as free
aldehyde, as aldehyde acetal, or as some other compound of alde-
hydes. One possibility is that the true intermediate is aldehyde
phosphate such as R-CHOH-OP03H2. This last compound
would be highly reactive and unstable in the presence of water,
and so it would be difficult to discover in tissues. If it occurred
as a true intermediate, one would probably find by cytochemical
processes either free aldehyde derived from its decomposition, or,
as we shall see later, an acetals It is clear that there is room
for a considerable amount of work, particularly on the enzymes
which might be associated with reactions involving aldehydes.
At the present no methods have been worked out for the cyto-
chemical localization of such enzymes. But it is interesting to
note that Worden (1943) has found that there is a very high con-
centration of aldehyde oxidase present in adsorbed form on the
surface of the fat droplets of milk.

92 CYTOCHEMISTRY OF ALDEHYDES

The Significance of Aldehyde Acetal

Feulgen and his colleagues showed that, whereas some free al-
dehyde (which they originally called plasmal) exists in tissues, a
large part of the naturally occurring aldehyde occurs as the
acetal phosphate of glycerol (which they called plasmalogen), as
shown in Formula III. The acetal phosphate III has been shown
to be a common constituent of tissues. Its mode of formation is
not at present clear. If an aldehyde phosphate occurs in tissues,
it might well combine spontaneously with glycerol as indicated
in I, to give the compound II. This latter compound would prob-
ably pass spontaneously into III by wandering of the phosphate
radical, as occurs in the equilibrium between a- and ^-glycero-
phosphates. Alternatively, it may be that the compound III is
derived by condensation of free aldehyde and glycerophosphate
to give the acetal phosphate. Various hypotheses could be put
forward for the function of the acetal compound III. It might
be concerned in fatty-acid metabolism, in the formation of tri-
glycerides or in the formation of phospholipins. But there is
little evidence available on these points as yet.

R R

I I

CH CH

/ \ / \

OH OPO3H2 O OP03H2

+
HO OH

OH
CH2— CH— CH2OH2

CH2— CH— CH2OH

1 11

R

CH
/ \

o 0

CH2 CH— CH2OP03H2

in

The Study of Glycols in Tissues

Groups such as — CHOH— CHOH— do not react directly with
any of the reagents for aldehydes. Recently, however, it has been
shown that periodate reacts in a fairly specific manner with these

THE STUDY OF GLYCOLS IN TISSUES 93

groups, to form aldehyde groups which may then be detected
with the aldehyde reagents. Hotchkiss has shown that treat-
ment with periodate may be a very valuable procedure in the
study of carbohydrates in tissues, provided that the carbohy-
drates are insoluble in water (see, for example, Plate XII, Fig.
B) . This technique has at present been far less extensively ex-
ploited than is desirable. There is a possibility that it could be
used to demonstrate the presence of pentose nucleic acid. Nor-
mally a nucleic acid does not contain an a,/?-glycol grouping.
After hydrolysis of pentose nucleic acid, as in the Feulgen pro-
cedure, an extra hydroxy group is liberated on Ci by the re-
moval of the purine moiety from the nucleic acid. Consequently
if there is also a free hydroxyl group on C2, the compound now
contains an a,£-glycol group which should be susceptible to oxi-
dation by periodate. In the case of the thymonucleic acid C2
does not bear an hydroxyl group, unlike pentose nucleic acid. It
is not at present known, however, whether the hydroxy group at
C2 is normally phosphorylated in pentose nucleic acid. If it is
not, after a Feulgen-type hydrolysis, pentose nucleic acid should
contain a glycol grouping which can react with periodate.

An attempt has been made to study the distribution of a sub-
stance with the above structure in tissues, using the following
procedures :

1. The tissue is embedded in wax, one result of which is to remove any
fatty aldehyde which may be present.

2. It is necessary to remove any preformed glycol grouping. This is
quite readily done by treatment with periodate which transforms the gly-
col groups into aldehyde groupings. These groups are later (in Step 4)
blocked with hydroxylamine or dimedone.

3. Sections are submitted to a Feulgen hydrolysis in the presence of
mercury. The hydrolysis splits off purine. The mercury prevents
nucleic acid from going into solution.

4. As a result of the Feulgen hydrolysis reactive aldehyde groups are
liberated on the deoxy sugars but not on the normal sugars. This reactive
aldehyde, together with that formed in Step 2, is blocked by treatment
with hydroxylamine or dimedone.

5. The sections are then treated with periodate to oxidize any a,/3-glycols
liberated in Step 3 to aldehyde groups.

6. The section is exposed to reduced fuchsin to colour the sites of alde-
hyde groups.

When this procedure is carried out it has been found that with
a number of tissues, e.g., pancreas, a colour appears in those sites

94 CYTOCHEMISTRY OF ALDEHYDES

in the cytoplasm and nucleus which are believed to contain pen-
tose nucleic acid. This investigation, however, has been made
in only a very preliminary way, and it by no means follows that
the colour which is developed under those conditions is necessarily
due to pentose nucleic acid. All that can be said is that the
procedure which is given has demonstrated the presence of a sub-
stance, which is normally masked in tissues, but which is brought
into a form which will react with periodate by a Feulgen-type
hydrolysis.

Procedures of the type which have just been outlined will per-
mit the detection of a considerable variety of substances. Treat-
ment of a tissue with periodate followed by reduced fuchsin will
demonstrate the presence of a,£-glycol groups such as those of
glycogen. By using a Feulgen hydrolysis instead of periodate
oxidation, deoxynucleic acid may be demonstrated. An oxida-
tion carried out after hydrolysis will demonstrate certain types
of masked glycols such as those found in readily hydrolyzed
glycosides. It is also possible that a good deal of evidence can be
collected about the orientation of the hydroxyl groups in these
glycols. For example, the cis glycols can probably be prevented
from reacting with periodate by formation of the acetonyl com-
pounds, or, as Seymour Cohen has suggested to me, perhaps by
the use of borate in the periodate mixture. Trans hydroxy groups
do not readily form acetonyl compounds, nor do they have the
property, possessed by cds compounds, of forming a complex with
borate. Neither the acetonyl compound nor the complex with
borate is likely to react with periodate.

By using procedures of this type it should be possible to obtain
very useful information about many of the natural polysac-
charides, including the polysaccharides present in the mucopro-
teins and intracellular matrixes.

REFERENCES

Bennet. 1940. .Am. J. Anat., 67, 151.

Boscot, Mandl, Danielli, and Shoppee. 1948. Nature, 162, 572.

Cain. 1949. Quart. J. Micro. Sci., 90, 411.

Camber. 1949. Nature, 163, 285.

Danielli. 1949a. Quart. J. Micro. Sci., 90, 67.

19496. Quart. J. Micro. Sci., 90, 309.

1950. Quart. J. Micro. Sci., 91, 215.

REFERENCES 95

Dempsey et al. 1943. Endocrin., 33, 387.

1944. Endocrin., 35, 409.
Feulgen and Rossenbeck. 1924. Zeit. Physiol. Chem., 135, 203.
Feulgen and Voigt. 1924. Pflugers Arch., 206, 389.
Hotchkiss. 1948. Arch. Biochem., 16, 131.
Lehninger. 1945. J. Biol. Chem., 161, 415.
Stedman and Stedman. 1943. Nature, 157, 740.
Worden. 1943. Nature, 152, 505.

CHAPTER $

THE CYTOCHEMISTRY OF PROTEINS
AND NUCLEIC ACIDS

Introduction

A common characteristic of the proteins and the nucleic acids
is that they both normally occur as macromolecules. It is, of
course, desirable if possible to characterize macromolecules by
their properties as a whole. But this is usually impossible, par-
ticularly in cytochemistry, and it is usually only by studying the
reactivity of certain chemical groups present in the macromole-
cules that these substances are identified. Consequently, there
is much similarity in the methods which are employed in the
study of the various proteins and nucleic acids, and in the diffi-
culties which are encountered in these studies. A convenient
classification of the groups present in proteins and nucleic acids
which are available for study from the cytochemical point of
view is:

1. Carbohydrate groups. These may be found both in nucleic acids and
in proteins. Part of the cytochemistry of these groups has been discussed
in Chapter 4.

2. Phosphate groups. Until recently phosphate groups have mainly
been studied in relation to nucleic acid. It is known, however, that in
addition to the phospho-proteins of yolk and of milk, other phospho-
proteins occur in nature. Myosin, for example, is probably a phosphoryl-
ated protein.

3. Resonating ring structures. These are groups such as purines, pyrimi-
dines, tryptophane, tyrosine, and haem. These groups have the capacity
for absorbing light vigorously. Most of the naturally occurring resonating
ring groups of proteins and nucleic acids absorb in the ultraviolet region,
but a few, such as haem, also absorb light in the visible spectrum.

4. Peptide linkages. These are found in proteins only, so far as is known.

5. Reactive side-groups. These include such groups as NH2, CO2H, SH,
OH, tyrosine, histidine, tryptophane, etc.

6. Enzyme activity. One method of characterizing a protein molecule
as a whole is by its enzyme activity. The study of alkaline phosphatase
discussed in Chapter 3 is, of course, an example of this.

96

INTRINSIC ABSORPTION SPECTRA 97

7. Antigens. It is probable that individual proteins can conveniently
be studied by their reaction with coloured antibodies.

The number of methods available for study of the nucleic acids
and the proteins is considerable, although as yet far from suffi-
cient for obtaining all the information which is required about
their cytochemistry. The methods include:

1. Observation of the intrinsic absorption spectrum.

2. Observations made after the use of chromogenic reagents which
react with particular groups.

3. Determination of the localization of enzyme activity.

4. Observation of the reactions of tissues with colored antibodies.

5. Staining with dyes.

6. Study of the action of enzymes such as proteolytic enzymes and
nucleases, on tissue sections and squashes, etc.

In this chapter we shall consider only the methods involved
in 1, 2, and 4. The methods concerned in 3 have already been
considered in Chapter 3. The methods involved in 5 and 6 are,
in my view, not sufficiently reliable except for preliminary studies.
The reasons for this have been given elsewhere (Danielli, 1946),
and there seems to be no reason for revising the opinions given
in this earlier publication.

The Analysis of Intrinsic Absorption Spectra

As has been shown by the accomplished work of Caspersson
and his school, a remarkable amount of information about pro-
teins and nucleic acids can be obtained even when one is limited
to studying the ultraviolet absorption spectra. When reflecting
microscopes become more readily available it should be possible
to use, with more or less equal facility, absorption bands ranging
from the infrared to the far ultraviolet. At present most studies
have been limited to the regions between 2500 A.U. and 3000
A.U., which is the region in which the pyrimidine ring and some
of the groups of proteins have strong absorption bands. It is un-
likely that the extreme ultraviolet beyond 2400 will be a very
profitable region for study, since so many substances have strong
absorption bands in this region. There is also a considerable lim-
itation on the usefulness of moving into the infrared regions, due
to the decline in resolving power as one moves into the longer
wavelength regions of the spectrum. This limitation is more
severe than is generally realized. Thus, although the limit of

98 CYTOCHEMISTRY OF PROTEINS

resolution is given by A/2, it is necessary to have a region at least
3 A. in diameter before quantitative studies of its absorption spec-
trum can be made. Thus, if one is employing light with a wave-
length of 4 fi7 whilst the resolution limit is 2 p, the size of area
which is needed for studying the absorption spectrum for light of
wavelength 4 ju is at least 12 /x. Thus studies in this region of
the infrared would only be useful with very large cells.

The characteristic absorbing unit of the nucleic acids is the
pyrimidine ring, found in both the pyrimidines and in the purines.
In nucleic acids these rings produce a very strong absorption of
light in the region of 2600 A.U. The exact position of the absorp-
tion band of individual purines and pyrimidines varies, according
to the side-groups which are present on the ring structure. Thus
a considerable change in position of the absorption bands may
occur when an NH2 group is replaced by an OH group. Until
now observations on the nucleic acids have been limited to those
which can be made without modification of the purine and pyrim-
idine rings. But it may quite well be possible to modify the
absorption spectrum by carrying out reactions with these ring
structures.

In the proteins the main absorbing components are trypto-
phane and tyrosine, which absorb strongly in the region of 2800
A.U. The absorption of light by tyrosine is sensitive to pH
changes. The remaining amino acids absorb light rather strongly
in the region of 2400 A.U.

Generally speaking, it is probably true that one can determine
the amount of the nucleic acid and of protein by making observa-
tions at a number of different wavelengths in the ultraviolet. It
must, however, be remembered that the absorption of light is a
property not of a molecule as a whole, but of some of the smaller
constituent groups. It is not impossible for some of the purines
and pyrimidines occurring in nucleic acids to occur as constituent
groups of proteins, and for some of the amino acids to occur as
constituent groups of nucleic acids. Panijel has found a protein
in Ascaris sperm which contains purine but no phosphorus. Thus
although this protein has an absorption spectrum similar to
nucleo-protein since it lacks phosphorus it cannot contain nucleic
acid.

Very few studies have been made in the visible region of the
spectrum, since most proteins do not absorb light to a significant

USE OF CHROMOGENIC REAGENTS

99

degree in the visible. An exception to this is the very interesting
study which Thorell (1947) has made of the formation of hemo-
globin during the development of red blood cells.

A typical example of the work of Caspersson is illustrated by
Fig. 5.

1.0

2200 2400 2600 2800

A.U.

3000

3200

Fig. 5. Extinction coefficients for a point on a chromosome (upper curve)

and in the cytoplasm (lower curve) of a Tradescantia pollen mother cell.

The absorption at 2800 A.U. is mainly caused by protein, and at 2600 A.U.,

mainly by nucleic acid. After Caspersson, 1950, page 66.

General Principles Involved in the Use of
Chromogenic Reagents

There are a wide variety of agents which may be used in order
to produce an absorption of light by proteins or nucleic acids in
various parts of the spectrum. Up to the present these agents
have been almost entirely restricted to the development of ab-
sorption bands in the visible region of the spectrum. It seems
likely that in the future attempts will be made to use reagents

100 CYTOCHEMISTRY OF PROTEINS

producing absorption in the ultraviolet and infrared regions.
But with all regions of the spectrum, the same general problems
must be taken into consideration. The general requirements of a
chromogenic reagent are:

1. The reagent should not give a generalized staining, i.e., it shall be
inert to most components of a tissue section and also it should be colour-
less itself.

2. The reagent should be readily washed out of sections by inert
solvents.

3. The reagent should have a known specificity, i.e., it must be known
just what groups it will attack.

4. The product of the reaction between the agent and the protein or
nucleic acid shall be indiffusible.

5. The reagent shall produce its effect under mild conditions. For ex-
ample, it is essential that none of the steps involved shall dissolve com-
ponents of the tissue section.

6. It is desirable that there should be several independent methods for
colouring the same group, so that the results obtained by these different
methods may be compared.

Many of the points concerned have been discussed in more
detail elsewhere (Danielli, 1946-1949) .

Practically all the chemical reagents which are available will
in fact attack more than one group of a protein or nucleic acid.
Consequently, it is frequently necessary to use two reagents, one
of which will be a non-chromogenic reagent used to block one of
the groups which would otherwise be coloured by the other reagent.
In fact, it can safely be said that the development of non-chro-
mogenic blocking agents is a key point in the cytochemistry of
the proteins.

The complication which has not as yet received consideration
is the accessibility of some of the chemical groups of proteins and
nucleic acids to the reagent being used. It is well known from
the work of Anson and Mirsky and others that tyrosine and SH
groups of proteins may be masked in a native protein, and only
become accessible to chemical reagents after denaturation. Simi-
lar observations have been made on histidine, and no doubt all
groups of proteins are to some extent inaccessible to chemical
reagents. Sometimes, as is the case with histidine, a group which
is accessible to a reagent in a native protein may become inacces-
sible in a denatured one. This question of masked groups pre-
sents a very difficult problem for cytochemistry, particularly in

DIAZONIUM HYDROXIDES 101

view of the fact that a considerable part of the protein in a fixed
tissue section is in the denatured state. I shall not dwell on this
point further at the present time. But it should perhaps be men-
tioned that this problem does not really become important in
most cases until quantitative studies are being made. At present
there is so much doubt as to how far microspectrophotometric
observations are quantitative that it does not seem worth while
to go into the problems arising from masking and unmasking of
chemical groupings.

Diazonitjm Hydroxides as Chromogenic Reagents

Diazonium hydroxides will react with histidine, tryptophane,
and possibly nucleic acid to give azo dyes. There are, of course,
many other possible components of tissues with which diazonium
hydroxides will react, including most phenols. The most con-
venient diazonium hydroxides have so far proved to be tetrazot-
ised benzidine and dianisidine. Frequently it is found that, when
one of these compounds is allowed to react with a tissue section,
a sufficient intensity of colour is developed for cytochemical pur-
poses. Under these conditions the reagent is linked onto the pro-
tein or nucleic acid by only one of its two active groups. If, how-
ever, the colour is not sufficiently intense, it may be intensified
by washing the section to free it from excess of reagent, and
then placing the section in a solution of phenol or amine which
then couples onto the free end of the reagent. For example,*

Pr< >OH

Pr< >OH

N=N< >— < >N=N-OH

1

Pr< >OH

N=N< >-< >-N=N< >NH<

* Here and elsewhere, Pr denotes protein.

102 CYTOCHEMISTRY OF PROTEINS

A typical procedure would be as follows:

1. A piece of tissue is submitted to freeze-drying and sectioned after
embedding in wax.

2. After removal of the wax and passing through alcohol the sections
are placed for 10 minutes in a solution of 0.2 percent tetrazotised benzidine
at about 4° C, in sodium veronal buffer.

3. The sections are then washed for 1 minute in each of three changes
of 0.1 TV hydrochloric acid. It is essential that this washing should be
very thorough, and the solutions must be changed frequently to secure
this.

4. The sections are then placed for 15 minutes in a suspension of 0.1
gram of a- or /3-naphthol suspended in sodium bicarbonate or sodium
veronal solution.

5. The sections are washed in distilled water and mounted in balsam.

Since the diazonium hydroxides can react with a number of
tissue components, it is usually necessary to use blocking agents
to prevent the chromogenic reagent from combining with all the
groups which are open to attack. Among suitable reagents for
blocking purposes are dinitrofluorobenzene, performic acid, and
benzoyl chloride. Dinitrofluorobenzene will block all phenols
and also histidine ; performic acid destroys tryptophane ; and ben-
zoyl chloride destroys histidine, tryptophane, and tyrosine. Some
of the possible permutations and combinations of blocking agents
are shown in Table VII.

TABLE VII

The ability of certain amino acid residues to react with diazonium hy-
droxides after treatment with various reagents.

2 : 4 Dinitro

Fluoroben- Performic Benzoyl
Reagent None zene Acid Chloride

Histidine -+- ? + —

Tryptophane + + — —

Tyrosine + — + —

The use of diazonium hydroxides for the study of nucleic acids
was first suggested by J. S. Mitchell. He suggested that a tissue
section should first be subjected to benzoylation and then ex-
posed to a diazonium hydroxide. The reactive amino acids would
be destroyed by benzoylation, and only the nucleic acids should
be free to react. Mitchell's conditions of benzoylation were un-
satisfactory. He used benzoyl chloride in sodium hydroxide so-

NITRO COMPOUNDS 103

lution: under these conditions the reaction with benzoyl chloride
is patchy, and also the sodium hydroxide extracts much material
from the sections. I have found that a much more reliable tech-
nique is to use a 5 percent solution of benzoyl chloride in dry
pyridine, for about 12 hours.

There is also some doubt as to whether nucleic acid can, in fact,
react with diazonium hydroxides under the circumstances sug-
gested by Mitchell. It is certainly true, however, that after ben-
zoylation, treatment with a diazonium hyroxide produces a dis-
tribution of colour which is very similar to the distribution of
nucleic acid revealed by other processes. Further inquiry into
the chemistry of this reaction is clearly necessary.

Nitro Compounds as Chromogenic Reagents

There are many nitro compounds in use in organic chemistry
as analytical reagents. Many of these will react with nucleic
acids and proteins. Some of these substances give a reaction
product which is coloured. Thus, according to Sanger, dinitro-
fluorobenzene will react with tyrosine, histidine, NH2 groups,
and SH groups. Of these, the reaction products with histidine,
NH2 groups, and SH groups are coloured. Usually, however, the
intensity of colour is too low to be satisfactory for cytochemical
problems. The intensity of the colour may be increased to a sat-
isfactory degree by reduction of the nitro group of the reagent
to an amino group, diazotisation of the resulting amino group,
followed by linkage to a phenol or amine. This procedure pro-
duces an azo dye, which is linked by covalent bonds to the
group which is being studied. A typical reaction is the following:

N09 N02

Pr-NH2 -> Pr-NH<^ >N02 -> Pr-NH<" ^NH2

i
N02
<- Pr-NH/~~^)>N=N-OH
N02
Pr-NH<^ ^>N=N<^ "\NH9

104

CYTOCHEMISTRY OF PROTEINS

The details of the procedure necessary for using some of these
nitro compounds have been worked out together with L. G. Bell
and A. Loveless. The basic procedure is as follows:

1. Sections are placed for 1 hour in a suspension of 0.5 milliliter of
dinitrofluorobenzene and 1 gram of sodium bicarbonate in 100 milliliters
of 70 percent alcohol. Under these conditions linkage of the dinitrofluoro-
benzene to tyrosine, histidine, NH2 groups, and SH groups may occur.

2. The sections are washed in three changes of 70 percent alcohol, 5
minutes in each change, to remove uncombined dinitrofluorobenzene.

3. The sections are then treated for 2 minutes with a solution of
chromous chloride in 0.1 N hydrochloric acid. This reduces the nitro
groups of the dinitrofluorobenzene which is linked to various groups in
the section (i.e., R-N02 -» R-NH2).

TABLE VIII

Some nitro reagents, and groups for the localization of which these
reagents may be particularly useful.

N02

N02

N02

NH,

NOs

N02

N02

>N02

>NCO

>CH2Br

>AsO

Tyro-
sine

+

+

>NH-COCH2I

>0-CO-CH2I

>COCl

SH

+

+

NIL

+

+

CO2H

CHOH

+

+

+

+

+

+

ALDEHYDES AS CHROMOGENIC REAGENTS 105

4. Sections are washed in two changes of 0.1 N hydrochloric acid for 3
or 4 minutes, and transferred to

5. A nitrous acid solution obtained by dissolving 0.5 gram of sodium
nitrite in 100 milliliters 0.1 N hydrochloric acid kept at 0° C. The sec-
tions are kept in this solution for 5 minutes to diazotise the amino groups.

6. The sections are transferred to an alkaline solution of a phenol or a
solution of an aromatic amine, e.g., H acid or a-naphthol or /3-naphthyla-
mine. After 15 minutes in this solution coupling with the diazotised sec-
tions should be complete.

7. The sections are washed in tap water and mounted in balsam.

Table VIII gives a list of some of the nitro reagents which are
usually available and indicates the groups of proteins for the
study of which they are particularly suitable.

Aldehydes as Chromogenc Reagents

A third family of compounds useful in the cytochemistry of
proteins are those containing two reactive groups, one of which
is an aldehyde group and the other of which may be any of a

TABLE IX

Aldehyde reagents under investigation for the cytochemistry of
and — NH2 groups.

-SH

CHO

CHO

CHO

o

NH,

CHO<

>N02

CHO^ >COCH2Br

CHO

>0-CO-CH2I

+

+

+

CHO

o

H • CO • CH2I

SH

+

+

+

106 CYTOCHEMISTRY OF PROTEINS

considerable variety of groups. Table IX shows the formulae
for a number of these reagents. These are being investigated
in collaboration with A. Loveless and L. G. Bell. The reagents
may be used in two ways. Thus the CHO group may be allowed
to react with the NH2 groups of a protein. After this reaction
has gone to completion, the second reactive group (e.g., N02 or
NH2) may be subjected to the procedure discussed in the previ-
ous section, to produce a colour. In other cases, the second re-
active group is allowed to react with the protein, the CHO group
being kept intact. The section is then exposed to reduced fuchsin,
which forms a purple colour with the aldehyde group. The fol-
lowing is an example of this type of reaction:

Pr-NH2 + CHO<^ ^>NH2 -> Pr-NH-CHOH^ ^>NH<

i

Pr • NH • CHOH^ \N=N ■ OH

Pr-NH-CHOH< >N=N< >NH;

Blocking Agents

As was pointed out above, chromogenic reagents are usually
able to react with several of the groups found in a tissue section.
Consequently, before the results obtained with a chromogenic
reagent can be analyzed, it is necessary to make studies on sec-
tions in which some of the reactive groups have been blocked with
non-chromogenic reagents. Table X shows a list of such reagents
and of the groups which they can conveniently be used to block.

Some of the reagents given in Table X are able potentially to
react with a considerable number of the chemical groups of a
protein. But the actual range of their blocking activity can be
restricted by controlling the pH at which the blocking reaction
is carried out, and controlling the time of exposure to the
blocking agent. This is the case, for example, with phenyl iso-
cyanate and diazonium hydroxides. Some of the blocking re-
agents have the useful property of having a reversible action.
This is the case with Hg, ClCH2-0-CH3, PhSCOCl, ClCH2-0-
CH=CH2, PhAsO, PhHgOH. With these reversible reagents it

BLOCKING AGENTS

107

TABLE X

Reagents which may be used for blocking various groups in protein
molecules. Reagents which may readily be removed are marked with
an asterisk.

Tryp-
to-
phane

Tyro-
sine

Histi-
dine

SH

s— s

NH2

C02H

CHOH

N02

F<^ ^>N02

+

?

+

+

HN02
HCO-OOH

+

+

+

+

/~ ^>COCl

+

+

+

+

+

+

Hg*
ICH2-CONH2

+
+

?

<^ ^>COCH2Cl

+

Glyoxal
Ketene

( + )

+

+
+

/ Nnco

+

+

+

ClCH2-0-CH3*

+

/ Ns-coci*

+

-

Naphthoquinone

+

+

Methylnaphthoquinone
ClCH2-0-CH=CH2*

+

+

/ \AsO *

+

/ ")>HgOH *

+

N02<^ ^)>N=NOH

+

+

+

is often possible to block a group, carry out a cytochemical pro-
cedure, and then remove the blocking agent by a mild treatment
with dilute acid, dilute alkali, dithioglycerol, or a heavy metal.
After removal of the blocking agent a further cytochemical pro-
cedure may be carried out with the group which has been liber-
ated. Thus, for example, NH2 groups may be protected by
treatment with PhSCOCl against a reagent for tyrosine. Then

108 CYTOCHEMISTRY OF PROTEINS

treatment with a solution of heavy metal will liberate the NH2
group which may then be treated with a second cytochemical
reagent. A more detailed discussion of these reagents has been
given elsewhere (Danielli, 1949).

Methods of Increasing Sensitivity

Where tissue components are present in small amounts, the
methods given above may not be sufficiently sensitive. So far we
have considered three methods for increasing sensitivity.

1. Probably the method which is potentially able to give the
greatest increase in sensitivity is to modify procedures of the
type given in preceding sections so as to obtain a fluorescent com-
pound. The specimens may then be studied in the fluorescent
microscope. A considerable gain in sensitivity should be achieved
even by visual inspection. If photographic means are used there
should be, in theory, a very great increase in sensitivity indeed.
On the other hand, the intrinsic fluorescence of the specimen may
pose some difficulties. These, however, may perhaps be met by
using selective quenching agents and by observing the fluores-
cence only in selected regions of the spectrum. Imperfections in
the optical equipment will, of course, also set a limit to the sen-
sitivity as determined by photographic methods. We have not
as yet paid much attention to the use of fluorescent compounds,
but it does hold much promise.

2. A technique which we have found of considerable value
involves only a slight modification in many of the procedures
set out above. In these procedures colour has been obtained by
forming an azo dye in the section. If, in the final stage of the
formation of this dye, an amine is used which can be diazotised
after coupling, it is in theory possible to repeat the procedure,
linking on yet another amine, after diazotisation. Theoretically,
this process may be repeated an indefinite number of times. In
practice, however, although there is a very considerable increase
in the intensity of coloration produced by repeating this step a
small number of times, the increase is not directly proportional
to the number of couplings, the increment diminishing as the
number of couplings is increased. Consequently, we have found,
with an amine such as a-naphthylamine, that eight couplings are
the maximum which it is profitable to undertake.

THE DETECTION OF DIFFUSION ARTEFACTS 109

3. Another procedure which we have under investigation is to
link a sugar onto the compound which is being coloured. To se-
cure this coupling it is necessary to have a sugar or related com-
pound which has been linked with a phenol or an amine in such
a way that the resultant compound is competent to react with a
diazonium hydroxide. Thus the actual mode of operation is to
produce a diazonium hydroxide in the section, possibly a diazo-
tised azo dye as suggested in method 2, and then to couple on the
sugar compound just mentioned. When the coupling is com-
pleted the sugar may be oxidised with periodic acid. This will
give aldehyde groups where previously there were a,/3-glycol
groups. The aldehyde groups are then treated with reduced
fuchsin, which, of course, gives a vigorous colouration. We are
at present experimenting in this way with a derivative of inulin.
The use of such a polysaccharide is, however, complicated by
steric factors, and to what extent the use of such compounds will
prove satisfactory cannot at present be predicted.

The Detection of Diffusion Artefacts

All the methods for the study of proteins and nucleic acids
which have been discussed have involved the formation of dyes
in the sections which are attached by stable covalent bondings to
the compound which is to be identified. Consequently, there are
no problems arising from the actual diffusion of the dye itself,
or from the adsorption of the dye by other compounds in the
section. All the methods are so designed that each individual re-
agent is a small colourless compound, the excess of which can be
washed out of the tissue section with comparative ease. Conse-
quently, there should be little trouble arising from unwanted
colour produced by residual amounts of reagents. The most seri-
ous type of diffusion artefact which is likely to arise under these
circumstances is that due to diffusion of the protein or nucleic
acid element itself. There is a very serious possibility that such
diffusion may occur despite all that can be done in the way of
fixation. To detect artefacts of this nature, three methods are
in use.

1. To test for the loss of appreciable amounts of components
by diffusion, about 0.2 gram of sections is suspended in the vari-
ous reagents. Each step in the procedure is then carried out on

110 CYTOCHEMISTRY OF PROTEINS

this considerable quantity of suspended sections, after which the
sections are centrifuged down and the fluid phase then tested
for protein and nucleic acid. In this way the gross loss of mate-
rial from sections by diffusion may be detected.

2. Where the components of the tissue which are being stained
can be obtained in a soluble condition, a very useful method is to
carry out the cytochemical reaction with this solution of the
components. Then a tissue section is exposed to the coloured
soluble component to see whether it adsorbs upon the section.
This reveals the sites of affinity within the section for the coloured
component. In cases where the distribution of sites of affinity
coincides with distribution of sites of the compound as revealed
by a cytochemical method, further enquiry is clearly indicated.

3. Perhaps the most useful method of all is the use of superim-
posed sections, somewhat along the lines used in the case of
alkaline phosphatase. Let us consider as an example the dis-
tribution of tyrosine in a section. A section is placed on a slide
and treated with a blocking agent so that no residual tyrosine
groups remain in it. Then upon this is placed the section which
is to be examined, and the cytochemical reaction is carried out
with the superimposed sections. If any diffusion artefact occurs,
it should be revealed by the appearance of colour for tyrosine in
the underlying blocked section.

Methods for the Electron Microscope

In 1948 Mr. Walker and Miss Hanson, of Professor Randall's
Biophysics Research Unit at King's College, London, asked me
whether it would be possible to develop methods of the types
discussed above for use with the electron microscope. The physi-
cal problem involved is markedly different from that encountered
with the light microscope. In the light microscope perception of
an object is based upon non-transmission of light as a result
either of scattering or absorption, or both. If scattering is re-
duced to a minimum, and the component which it is wished to ob-
serve has an absorption spectrum markedly different from that
of any other component present, quite accurate estimations may
be made even of small absorptions of light. For example, an
absorption of 5 percent of the light passing through an object
may be made with an accuracy of better than 10 percent. On

METHODS FOR THE ELECTRON MICROSCOPE 111

the other hand, with electron microscopes, at least as at present
conceived, perception is based entirely upon the scattering of
electrons. Scattering is a function of the atomic number of the
atoms in the material upon which the electron beam impinges.
In biological objects, the great bulk of the scattering material is
carbon, oxygen, and nitrogen, which differ relatively little in
scattering power. Consequently, the image obtained in an elec-
tron microscope depends almost entirely upon the mass of mate-
rial present in different parts of the specimen, and there is little
prospect of perceiving chemical differences between different
parts of a specimen. Furthermore, the application of reagents
to the specimen will facilitate the perception of chemical differ-
ences only so far as the reagent adds selectively to the effective
electron-scattering power of individual chemical components.

Thus the sensitivity of methods for use in the electron micro-
scope is limited by the sensitivity of the instrument to differences
in scattering power. With the techniques available at the mo-
ment, it appears that, to obtain reproducible results, the scatter-
ing power of a component must be altered by at least 10 percent.
To increase the scattering power by this amount is difficult,
compared with the small changes in chemical composition which
will give a significant change in the transmission of visible light.
Consequently, one cannot expect, without marked improvement
in the instruments available, to obtain methods of the same de-
gree of sensitivity as those available for the light microscope.
On the other hand, the great advantage in resolving power pos-
sessed by the electron microscope justifies the development of
even crude cytochemical methods. In this direction we have
had some success. The work has been carried out in the main
by Mr. Stuart-Webb and Mr. Bell of my department, and Dr.
Bovet and Dr. Lamb of Professor Randall's department.

After some search for a suitable test object, it was found that
thin gelatin films could be obtained which were sufficiently uni-
form in thickness for preliminary investigations. With these thin
films the following procedure was worked out, in which the scat-
tering power of components which will react with dinitrofluoro-
benzene is markedly increased, partly by combination with a
large organic group, and partly by subsequent linkage with a
heavy metal.

112 CYTOCHEMISTRY OF PROTEINS

1. Reaction with dinitrofluorobenzene.

2. Reduction of the nitro groups to NH2 with chromous sulfate.

3. Diazotisation of the NH2 groups.

4. Combination with K acid or dithiol.

5. Combination with silver or lead.

N02 1 N02

PrH -> Pr— /~ ^\N02 -> Pr/~ ~\nK2

NO

2

Pr< >N=N-OH

NOo SH SH

SO3H SO3H

Pr<T ^N=N<^ y Pr<^ ^>N=N

ch3 x/V

OHNH2

i t

Pb Ag Ag

S03 S03

NOo S S

Pi/ \N=N<^ y Pr<^

CH3

OHNH2

This technique was applied to spermatozoa. The mean opac-
ity of the middle piece of bull spermatozoa, before treatment, was
0.825. After the treatment via K acid and silver the opacity-
was increased to 1.21, and after the treatment via dithiol and
lead the scattering power was 1.20. From these increases it was
calculated that between 8 percent and 14 percent of the middle
piece consisted of material able to react with dinitrofluoroben-
zene.

Potentially, most of the methods outlined above could be modi-
fied for use in the electron microscope. But there are many diffi-
culties to be overcome, from the points of view of technique of
staining, of interpretation, and of accuracy of estimating scatter-
ing power.

PROTEIN SYNTHESIS 113

The Systems Concerned in Protein Synthesis

At the date of writing the main contribution, from the cyto-
chemical point of view, to our knowledge of the systems con-
cerned in protein synthesis has come from the laboratory of
T. Caspersson in Stockholm. These contributions have been
published in a series of papers commencing in 1932, and have
been summarised in a recent monograph entitled Cell Growth
and Cell Function (1950) . Probably the most important of these
contributions has been the observation of the constant associa-
tion of nucleic acids with protein formation. In his monograph
Caspersson recapitulates the generalization made by Caspersson
and Schultz, based on the study of nuclei, bacteria, viruses, etc.,
"that all self-reproducing protein molecules need the coopera-
tion of nucleic acids for reproduction" (p. 100). Caspersson
also states that "the nucleic acids take part in the processes
leading to the synthesis of cellular proteins, not only in the
nucleus but also in the cytoplasm" (p. 140) . These views have
illumined many previous observations and have inspired a re-
markable volume of work by other investigators in a wide variety
of fields.

In pursuing the further analysis of the association of protein
synthesis and nucleic acids, there are three important questions
which will be considered here. These are: .

1. Do nucleic acids participate chemically in protein synthesis? Alter-
natively,

2. Are nucleic acids essential chemical or physical components of the
environment in which protein synthesis occurs?

3. What are the sites of protein synthesis?

To obtain a satisfactory answer to question 1 is by no means
simple. For example, if the nucleic acids had catalytic functions,
it is unlikely that chemical changes in the nucleic acid during
protein synthesis could be detected even by the use of isotopes.
On the other hand, if the phosphate groups of nucleic acids par-
ticipated in protein synthesis, it is likely that a phosphorus turn-
over could be detected by isotopic studies.

Question 2 is also by no means simple, for there are many con-
ceivable functions of nucleic acids which fall within the frame-

114 CYTOCHEMISTRY OF PROTEINS

work of this question. Two examples of these may be considered
as illustrations. One much publicised theory has been that the
nucleic acids serve as templates or moulds, upon which the pro-
tein is formed. So far as this theory implies that the surface of
a nucleic acid molecule can determine the sequence of amino
acids in a polypeptide chain, it seems to me quite naive, from a
physico-chemical point of view. To the extent to which the tem-
plate function is limited to determining the manner in which a
polypeptide chain is folded into a unique configuration there is
more to be said for this theory. A quite different alternative
theory is that the nucleic acids constitute a necessary part of the
environment in the same way that water does. Water is an
essential for protein synthesis, acting as a solvent, as part of the
protein structure, and possibly also as a catalyst and even as a
direct participant in the synthetic process; nucleic acid could,
so far as present knowledge goes, have a similarly essential role
as an environmental factor.

The third question engenders a different set of enquiries.
The simplest approach to this problem is to suppose that the
places in cells which have the highest concentrations of protein
are the sites of synthesis. The intrinsic fallacy in accepting such
a hypothesis without proof may readily be seen if we consider
the somewhat broader scene of protein production in a country-
side: the highest concentrations of protein are to be found in the
slaughterhouses, cold-storage centres, and granaries, but the
actual sites of synthesis are the fields, pastures, and woodlands,
in which protein may be said, both literally and figuratively, to
be thin on the ground. Consequently, although it is of much
interest to discover and record the sites of occurrence of proteins
in cells, we must refrain from identifying the richer sites with
sites of synthesis until proof of this identity is obtained.

With these considerations in mind, we may now proceed to
examine the problem in more detail.

Caspersson's Cytological Systems

First consider the observations connected with chromosome
reproduction. It was shown by studies on chromosomes, particu-
larly those of grasshopper spermatocytes, that the chromosomes
are richer in nucleic acid during nuclear division than at any
other time. This was interpreted as indicating "a connection

CASPERSSON'S CYTOLOGICAL SYSTEMS 115

between the duplication of the genes and the presence of nucleic
acid."

Taking the salivary gland nucleus of Drosophila as typical of
an intermitotic nucleus, Caspersson advances the view that the
euchromatic * regions of a chromosome produce proteins which
are rich in tryptophane and tyrosine, whereas the heterochro-
matic regions produce proteins which are rich in diamino acids.

Then, turning to reproduction of cytoplasmic protein, Caspers-
son believes that the process commences by formation of nucleo-
lar protein within a chromocentre. Thus the chromocentre be-
comes distended and the Feulgen-positive material appears as a
ring, halo, or crescent contiguous with the newly formed nucleo-
lus. The chromocentre in this condition is commonly called
the "nucleolus-associated chromatin." There is a very great
volume of evidence to indicate that the nucleolus is intimately
connected with formation of cytoplasmic protein. So far as is
known there are few, if any, exceptions to the rule that cells
which are rapidly forming cytoplasmic protein have prominent
nucleoli. During the formation of ova in the ovary, in which
very large quantities of cytoplasmic protein are formed, one or
more nucleoli are usually by far the most striking feature of the
nucleus. In developing eggs, nucleoli are far from prominent
during the early cleavage phases, which appear to be primarily
concerned with securing distributions of the. original yolk mate-
rial and the establishment of the early differentiation pattern.
But after this new antigens, i.e., new proteins, appear in the
embryo; during this phase the nucleoli become prominent.

According to Caspersson, the nucleolar material is rich in di-
amino acids. Although the nucleolus itself is lacking in deoxy
nucleic acid, it may absorb light strongly in the region of 2600
A.U. and is then supposed to be rich in pentose nucleic acid. The
content of pentose nucleic acid is,^ however, very variable.

It was also found that within the nucleus there is a gradient
in concentration of the protein rich in diamino-acid from nucleo-
lus to the nuclear membrane. In the cytoplasm there is a high
concentration of pentose nucleic acid and of protein adjacent to

* Euchromatic regions have a low content of deoxy nucleic acid between
mitoses, whereas heterochromatic regions retain much deoxy nucleic acid
during the intermitotic phase.

116 CYTOCHEMISTRY OF PROTEINS

the nuclear membrane, but the concentration diminishes farther
away from the nuclear membrane.

From these observations Caspersson concluded that "the
nucleic acids take part in the synthesis of cellular proteins, not
only in the nucleus but also in the cytoplasm."

Caspersson's Chemical System

Corresponding to the pattern of behaviour and composition of
the various cell organs just outlined, a scheme of chemical or-
ganisation is postulated. The scheme may be summarised under
four headings:

1. Nucleic acids participate in an essential manner in the formation of
both nuclear and cytoplasmic proteins.

2. In the reproduction of the gene protein, the participating nucleic acids
are of the deoxypentose type.

3. In the formation of the cytoplasmic proteins, the participating nucleic
acids are of the pentose type.

4. The nucleolus-associated chromatin (or heterochromatin) governs the
level of nucleic acid formation, and consequently also determines the rate
of protein production.

These four postulates all derive ultimately from the experi-
mentally observed correlation between protein synthesis and the
existence of a high concentration of nucleic acids within the same
cell. This correlation is quite adequately established. But, on
the other hand, there are other ingredients in these four postu-
lates which are not so well established and need more critical
examination. To conduct this examination, and to obtain more
precise information about the manner in which nucleic acids
control protein production, we can profitably consider a number
of studies made by other than cytochemical methods. But be-
fore doing so it is desirable to consider briefly the nature of the
systems in which the nucleic acid-protein synthesis has been se-
curely established.

Systems Displaying Protein Synthesis and High Nucleic
Acid Concentrations
A very interesting system to study from this point of view is
the growing oocyte and its attendant nurse cells, and the subse-
quent development of the fertilised ovum. In the formation of
the oocyte a great increase in cytoplasmic protein occurs with

SYSTEMS DISPLAYING PROTEIN SYNTHESIS 117

very little reproduction of genes. During this period the deoxy
nucleic acids of the oocyte nucleus are scanty in amount, rela-
tive to the size of the cell, whereas pentose nucleic acids are
plentiful in the cytoplasm both of the oocyte and of the nurse
cells, as has been shown by Brachet. When the oocyte has
reached its maximum size, the cytoplasmic nucleic acid may di-
minish. After fertilisation there ensues a period of rapid cell
division, and therefore of gene synthesis, with little change in
cytoplasmic protein: during this period the total nuclear deoxy-
nucleic acid increases rapidly. When the individual organs are
forming, the cytoplasm and nuclei are both rich in nucleic acid
during the phase of multiplication of cells and of cell growth.
But once the organ is established the cytoplasmic nucleic acid
may fall sharply and is high only in those cells which produce
large amounts of protein, and in these cells only at the time of
protein production. All these observations are consonant with
Caspersson's views.

So far as investigation has gone, observations on plant growth
fit into the same pattern. In the growing root tip, the cells which
are multiplying rapidly are much richer in deoxynucleic acids
than are the other cells. The cells in which cytoplasmic protein
is being formed rapidly are rich in pentose nucleic acids, whereas
those cells which are not producing protein may be almost devoid
of the pentose acids. In yeasts and bacteria the situation is the
same: high protein production is associated with a high concen-
tration of pentose nucleic acids, whereas resting phases contain
little pentose nucleic acid.

Rapidly growing tumours have cells whose cytoplasm is rich in
pentose nucleic acids; under the action of growth-restricting
treatment, the cytoplasmic nucleic acid falls rapidly.

When the different cells composing a protein-secreting organ
are examined, the same pattern emerges. Those cells which are
vigorously secreting protein are rich in nucleic acid, and the re-
mainder are not. Thus the exocrine cells of the pancreas are rich,
the endocrine cells are poor. In the gastric mucosa the HC1-
secreting cells are poor in nucleic acids, and the protein-secret-
ing cells are rich. Amongst the richest cells of all in pentose
nucleic acid are the motor neurons studied by Hyden (1947).
When vigorously stimulated, nerve cells may lose two-thirds of
their protein : in a subsequent resting phase this protein is rapidly

118 CYTOCHEMISTRY OF PROTEINS

reconstituted. It is thus not surprising to find the neurons so
rich in nucleic acid. If, however, the nerve axon is severed there
is a marked decline in cytoplasmic nucleic acid, followed by a
recovery as regeneration proceeds.

Thus the evidence shows that there is a widespread association
between protein production and high nucleic acid contents of
cells.

Biochemical Studies of Protein Synthesis

E. S. G. Barron has studied protein synthesis in bone marrow,
using isotopic phosphorus to detect phosphorus turnover in the
nucleic acids. In this system cytoplasmic protein synthesis is
very rapid indeed. But there is no detectable turnover of nucleic
acid phosphorus.

The reproduction of bacteriophage is another fascinating sys-
tem. Type T phage growing in E. coli contain a large amount of
deoxynucleic acids: they may thus be regarded as providing a
model experiment in nuclear-protein formation of the type con-
cerned in chromosome reproduction. The facts have been sum-
marized by Seymour Cohen (1949).

In the growth phase, normal E. coli produce about three times
as much pentose nucleic acid as deoxy nucleic acid. When the
bacteria are infected with phage T2r+, nucleic acid formation
appears to stop abruptly. Protein synthesis, on the other hand,
proceeds rapidly at a linear rate. Then, at approximately 7
minutes after infection, deoxy nucleic acid formation recom-
mences, at a linear rate which is about four times the normal rate
of formation. There is no significant formation of pentose nucleic
acid. Thus protein formation substantially precedes formation
of new deoxy nucleic acid. It was shown by Hershey, Kalmunson,
and Bronfenbrenner that T2r+ phage does not react with anti-
sera to E. coli; consequently, the protein which is formed is not
bacterial protein but a new antigen. The time relationships of
formation of new protein and nucleic acid after infection are
shown in Fig. 6.

From the data obtained with phage it is clear that after infec-
tion substantial protein synthesis occurs without the formation
of new nucleic acid. It could be concluded that the formation of
the specific protein of phage does not involve participation of

BIOCHEMICAL STUDIES

119

nucleic acid. But this would hardly be justified, since the infect-
ing phage particles may carry sufficient phage nucleic acid into
the bacteria to promote synthesis of a considerable amount of
phage protein. Furthermore, there is as yet no evidence that
bacterial nucleic acid present before infection does not partici-
pate in phage-protein synthesis.

4

1

T T

1 /

_3 °

(Z

>

—

—

1§

3$X

% 2

o

V

—

Increments

D 1—

/ 1

1

10 15

Min. after infection

20

25

Fig. 6. Synthesis of deoxy nucleic acid and of protein in E. coli after

infection by phage.

Further valuable information about the course of phage forma-
tion in bacteria has been obtained by the use of radiations.
Latarget (1948) studied the effect of X- irradiation of infected
cells. For up to 7 minutes after infection the results were char-
acteristic of a system in which it is necessary to secure only one
hit per original phage particle to prevent reproduction of that
particle. In the period from 9 to 13 minutes after infection it is
necessary to record more than one hit per infecting particle to de-
stroy the phage. From 13 minutes after infection until the
bursting of the bacteria the actual number of phage particles
present does not appear to increase, but the resistance of the
particles to irradiation does increase markedly. The rise in re-
sistance may be due to an increase in size of the new virus par-
ticles.

120 CYTOCHEMISTRY OF PROTEINS

Thus from the combination of biochemical and radiation stud-
ies it appears that new protein formation markedly precedes
new nucleic acid formation. But the formation of new genetically
active particles is not complete until new nucleic acid is formed.

Further information may be derived from studies of bacteria
which have multiple infection, i.e., more than one phage particle
per cell. Lea and Salaman (1946) had concluded that a large
phage particle has a zone sensitive to X-rays which is composed
of at least 14 distinct units. Luria and Latarget (1947) and
Luria (1947), using ultraviolet irradiation, obtained results com-
patible with the conclusions drawn from the use of X-rays, and
showed that there are 30-50 dissociable loci in a phage particle,
all of which are necessary for successful reproduction of phage.
In cases of multiple infection, reproduction was still possible so
long as at least one of each type of locus remained in the cell,
i.e., new phage could be produced as a result of some form of col-
laboration between different loci of separate phage particles in
the same cell, although none of the phage particles taken singly
in a cell could reproduce.

If it were permissible to generalise from the results on phage to
protein production in general, one would be tempted to ask
whether protein production is in fact determined by nucleic acid.
In the case of phage the evidence, though not definitive, provides
a hint that protein production may be independent of nucleic
acids, or alternatively that the nucleic acids may be responsible
merely for the organisation of synthesized protein, or for some
final stage in protein formation other than synthesis of peptide
bonds. It also seems certain that nucleic acid phosphorus is not
necessarily labile during protein synthesis. But to compare the
evidence obtained with phage with cytochemical studies is in any
event premature, because of a number of technical weaknesses in
the evidence available from cytochemical methods. These weak-
nesses include:

(a) The fact that little work has been done on the cytochemi-
cal changes of individual cells through a cycle of growth and divi-
sion. Most of the results available have been obtained on fixed
material, the cells of which are in different stages of growth. The
exact placing of individual cells of a fixed preparation in a se-
quence corresponding to their growth stages is necessarily some-
what a matter of opinion, and to that extent unsatisfactory.

BIOCHEMICAL STUDIES 121

(b) Most of the work on fixed cells has involved the use of
fixatives such as acetic alcohol. As indicated in Chapter 2, such
fixation may result in serious diffusion artefacts, and it is hardly
profitable to consider results on tissues which have been prepared
other than by freeze-drying.

(c) In the interpretation of cytochemical results the data
which are available show, at best, only that at certain times in
the life of a cell there are high concentrations of proteins and
other compounds at certain sites in the cell. There is no infor-
mation available as to where these compounds come from, where
they are synthesized, and often there is little information as to
what their ultimate fate is within the cell. Caspersson has built
up a fairly elaborate system based on the assumption that all the
substances concerned move down their concentration gradients.
This, however, is a quite unsatisfactory hypothesis, for at least
two reasons. First, it is just as characteristic of living cells that
substances should move up, rather than down, concentration
gradients. Secondly, there can be no question of the substances
diffusing down the concentration gradients in the cell: if thermal
diffusion were concerned, all the substances in a cell would be
uniformly distributed over cells in a few seconds. Thus it is
clear that, since no such uniform distribution exists with proteins
and nucleic acids, these substances are not free to diffuse, and
consequently one cannot tell whether concentration gradients in
these substances have any relevance to their future distribution
in a cell.

We are thus bound to conclude that, although much recent
work, especially that of Caspersson and his school and the stud-
ies of phage, is most stimulating and provides much valuable
evidence, the fact remains that the sites of synthesis of protein
within the cell are as yet unknown, and that the exact relation-
ship of nucleic acids to protein synthesis is also unknown. To
emphasize this point I wish to conclude this chapter by bringing
forward two quite independent alternative theories of the rela-
tionship of nucleic acid to protein synthesis, neither of which in-
volves direct participation of the nucleic acids in the formation
of the chemical bonds of peptide chains.

That there is a marked degree of species specificity in the bio-
logical action of nucleic acids was clearly demonstrated by the
experiments of Avery, McLeod, and McCarty, in which it was

122 CYTOCHEMISTRY OF PROTEINS

shown that type changes in bacteria were mediated only by the
action of the nucleic acids of that species. Further evidence was
obtained by Mazia, who showed that nucleic acids block the de-
velopment of fertilised sea-urchin eggs and that the action is
markedly dependent upon the species from which the nucleic
acid is obtained. Horstadius, Lorch, and Danielli also injected
nucleic acids into sea-urchin eggs, and obtained activation which
was to a considerable degree dependent upon the source of the
nucleic acid. It is therefore clear that any theory of the role of
nucleic acid must provide for a considerable degree of species
specificity. This, however, can be done in a variety of ways.
Two such mechanisms which will be discussed here are (1) that
nucleic acids are folding agents and (2) that nucleic acids are
trapping agents.

The first of these theories is based upon the fact that the
natural proteins of cells consist not merely of a specific peptide
chain but of a peptide chain which is folded into a specific con-
figuration. Treatment of natural proteins with so-called de-
naturing agents, which cause a loss of this specific configuration,
causes a loss of biological activity and even a marked loss of
ability to react with antibodies to the original natural protein.
Thus the unique configuration of a natural protein is funda-
mental to its normal action. But these unique configurations
cannot be produced without some highly specific mechanism.
May it not be that the steric factors in this mechanism are pro-
vided by nucleic acids? Astbury has pointed out that the dis-
tance between nucleotides in nucleic acids closely corresponds to
the distance between peptides in a polypeptide chain, so that a
precise adlineation between protein and nucleic acid is possible.
If now the mechanical properties of the nucleic acid are such as
to facilitate a particular manner of folding, we can picture a
process in which a polypeptide adsorbs on a nucleic acid mole-
cule, is folded into a unique configuration by folding of the
nucleic acid, and is then shed as a folded globular protein. This,
of course, leads to the question, what folds the nucleic acid?
The reply to this may well reside in a possibility considered in
Chapter 3, namely, that the chromosomes, in their active regions,
are constantly folding and unfolding, through the action of phos-
phate esters upon a phosphatase-contractile protein unit in the
chromosome. Thus the complete picture of protein production

BIOCHEMICAL STUDIES 123

could involve first synthesis of a polypeptide chain— a process
not involving nucleic acid— then the adsorption of the polypep-
tide onto the nucleic acid of a chromosome, followed by contrac-
tion of the chromosome region concerned. The contraction would
fold the polypeptide-nucleic acid complex into a configuration
partly determined by the nucleic acid, and the folded protein
would then be shed from the chromosome. If this were so, it
would require that, in many cases at least, the genetically active
parts of chromosomes would be nucleic acids. Such a view stands

Chromosome

Histone nucleate

Protein + ribose nucleic acid + histone

Protein nucleate

Fig. 7. To illustrate a simple system for protein synthesis, in which pentose
nucleic acid is acting as a trapping agent.

in sharp contrast to the theory which regards genes as arrays of
enzymes. But it is easy to reconcile such a theory with the
species-specific actions of nucleic acids, and with the fact that
viruses (which many regard as analogous to genes) appear to
have no intrinsic enzymes but do have intrinsic nucleic acids.

The second theory is that nucleic acid is a trapping agent. If,
by one means or another, an equilibrium is produced in a cell
between a protein and its precursors, then if the protein combines
with a trapping agent, which virtually removes it from the sys-
tem, further protein production must ensue. It is possible that
nucleic acids, through their well-known capacity to form com-
plexes with proteins, act as trapping agents. In such a case, dur-
ing active protein production, it would be necessary to produce
nucleic acid in proportion to the amount of protein produced.
This leads to a scheme such as that illustrated in Fig. 7 (Danielli,
1949). In this scheme a gene is conceived as a site of protein
synthesis and degradation, which is shifted in favour of synthesis
by the trapping action of pentose nucleic acid. The equilibrium

124 CYTOCHEMISTRY OF PROTEINS

between protein and pentose nucleic acid is itself controlled by
the concentration of histone: since histones are more basic
than most proteins, histone will compete for nucleic acid more
successfully than most proteins. Hence an increase in histone
will cut down protein synthesis, and an increase in pentose nucleic
acid will enhance protein synthesis.

The work of Caspersson and Brachet and their colleagues ap-
pears to be compatible with this scheme. It is interesting to ob-
serve that Stedman and Stedman (1943) found that the concen-
tration of histone is lowest in rapidly growing tissues such as
tumours and embryos, as is required by this theory. Correspond-
ingly, in red cells, which have a very high histone content, growth
has practically ceased.

If this scheme were correct one would predict that, since pro-
tamine is a more basic protein than histone, it should be a more
efficient competitor for nucleic acid than is histone, and so should
be a strong inhibitor of protein synthesis. That is, synthesis
should be very low in cells containing a high protamine content;
this is certainly the case with spermatozoa.

We thus have two distinctly possible roles for nucleic acids in
protein synthesis, neither of which involves direct participation
in formation of polypeptide chains. These two theories seem to
me to be equally possible, and the fact that this is so makes it
plain that much more information is required. But it must be
noted that it is not impossible that both theories may be cor-
rect, i.e., that nucleic acids may act both as trapping and as fold-
ing agents. One possibility would be that deoxynucleic acid acts
mainly as a folding agent, and pentose nucleic acid mainly as a
trapping agent.

REFERENCES

Avery, McLeod, and McCarty. 1944. J. Exp. Med., 79, 137.
Brachet. 1940. C. r. soc. biol, 133, 90.
Barron. 1949. Personal Communication.
Caspersson. 1947. Symp. Soc. Exp. Biol, 1, 127.

1950. Cell Growth and Cell Function (Norton, New York).
Caspersson and Schultz. 1938. Nature, 122, 294.

1939. Arch. exp. Zellforsch., 22, 650.
Cohen. 1949. Bact. Rev., 13, 1.
Danielli. 1946a. Nature, 157, 755.

19466. J. Exp. Biol, 22, 110.

1949. Cold Spring Harbor Symposia, 14, 32.

REFERENCES 125

Hershey, Kalmunson, and Bronfenbrenner. 1943. J. Immunol., 46,

267.
Horstadius, Lorch, and Danielli (unpublished).
Hyden. 1947. Symp. Soc. Exp. Biol., 1, 150.
Latarget. 1948. /. Gen. Physiol, 31, 529.
Lea and Salaman. 1946. Proc. Roy. Soc, B, 133, 434.
Luria. 1947. Proc. Nat. Acad. Sci. Wash., 33, 253.
Luria and Latarget. 1947. J. Bad., 53, 149.
Mazia. 1949. Personal Communication.
Mitchell. 1942. Brit. J. Exp. Path., 23, 296.
Panigele. 1948. Personal Communication.
Sanger. 1945. Biochem. J., 39, 507.
Stedman and Stedman. 1943. Nature, 152, 556.
Thorell. 1947. Cold Spring Harbor Symposia, 12.

CHAPTER

QUANTITATIVE STUDIES
IN CYTOCHEMISTRY

My own contribution to quantitative studies in cytochemistry
has been limited to pointing out some of the theoretical difficul-
ties in this field. The requirements of a system which will permit
quantitative studies are far from generally appreciated, and the
purpose of this chapter is to set out briefly the more general re-
quirements, so that the limitations of individual studies may be
more readily assessed. Since it is not practicable to give a de-
tailed account of all possible requirements in all possible situa-
tions, I shall limit the discussion to rather general problems, and
it should not be supposed that a study which complies with the
points to be set out is necessarily adequate. On the other hand,
any study which does not provide adequately for the solution of
all the problems which will be set out here is certainly inade-
quate. The problems which will be considered are the following:

1. Errors due to scattering and refraction.

2. Adherence to the Beer-Lambert law.

3. Errors due to states of aggregation.

4. Errors due to fixation and diffusion.

5. Errors due to inadequate chemical procedures.

6. Damage caused to specimen by illumination.

Errors Due to Scattering and Refraction

In almost all quantitative cytochemical studies, the experi-
mental procedure is to estimate the proportion of light which is
transmitted by the specimen at a variety of wavelengths. Since
cells are too small to permit such studies without magnification,
the experimental procedure involves examination of the distribu-
tion of light in a microscope image of the object, by means of
either a photographic method or some form of photocell. The
distribution of light in the image may be influenced by the

126

3>--*a

»

A*- :/"*>#

>

*%

-

*v*

f**«*J

>

■ *x

. t

*

j

1 :?

'T -i

>.

.JR

i

1 4;

-■4
: ■ §

* **
J,-

fc

f

... I

p

*«

A,

i *

• * #**

-

*

*

1 - . 4 **

>

¥0

% PJJI
1

4 ,; •

•1k ;; •

i

'

B

Plate I. Comparison of freeze-drying and chemical fixation. Both figures
show glycogen in rat liver as demonstrated by the reaction with periodate
(see Chapter 4) . Figure A shows the effect of fixing with 80 percent alcohol.
Figure B is a frozen-dried preparation. Figure A shows a polarised distri-
bution of glycogen, as though the cells had been centrifuged. Figure B
shows the physiological condition of the tissue. Preparations by L. G. Bell.

Plate II. Comparison of freeze-drying and chemical fixation. Both figures
are ultraviolet photographs of Walker rat sarcoma, taken by R. King;
2570 A.U.; 2-mm. quartz objective. Figure A shows fixation by Carnoy's
method, and Fig. B by freeze-drying. Note that with chemical fixation
there is a halo surrounding the metaphase plate, and the intermitotic nuclei

show much contrast, whereas with freeze-drying the cytoplasm absorbs
ultraviolet light more or less uniformly right up to the metaphase plate,
and there is comparatively little contrast in the intermitotic nuclei. Figure
B resembles ultraviolet photographs of undamaged living cells, and Fig. A
resembles ultraviolet photographs of damaged cells.

Plate III. Two examples of rat-kidney sections prepared by the freeze-
drying method, and then treated by cytochemical techniques. Figure A
shows a section treated by the tetrazodianisidine technique, followed by
K acid (see Chapter 5). Note the excellent preservation of the mito-
chondria and the nuclei, and the fine detail of the free borders of the
secretory cells demonstrated in the central tubule (L. G. Bell). Figure B
showrs a similar section treated by the glycerophosphate technique for alka-
line phosphatase; incubation time, 20 minutes. All the structures in Fig. B
appear by virtue of their phosphatase content only. Note that phosphatase,
though most abundant in the secretory borders of tubule cells, is evident
also in all cell borders of the tubule cells. Note also the uniform distri-
bution of phosphatase in the nuclei apart from the nucleoli, which are
richer in phosphatase than the remainder of the nucleus.

Plate IV. Rat-kidney tubule sections (fixative 80 percent alcohol), dem-
onstrating the Gomori-Takamatsu technique for alkaline phosphatase.
Figure A, incubated 20 minutes. Compare this figure with Plate III, Fig. B.
By comparison the nuclei of the alcohol-fixed section are grossly precipi-
tated, and during fixation phosphatase has diffused from the cell borders
into the cytoplasm. Figure B, incubated 5 minutes: the main sites of
phosphatase are already apparent. Figure C is a section containing phos-
phatase superimposed on a section in which phosphatase has been destroyed.
Incubation time, 320 minutes. The active section has been grossly over-
incubated and appears almost opaque. The microscope was focussed on
the underlying section, which is devoid of phosphatase activity, but which
shows a faint image by diffraction. The absence of any apparent phos-
phatase activity in the underlying section shows that, during the incubation
period of 320 minutes, no significant amount of phosphatase or of calcium
phosphate has diffused into the underlying section, so that there is clearly
no significant diffusion artefact. Figure D is the same as Fig. C, but incu-
bated 640 minutes. By this time some of the nuclei of the underlying
inert section appear black, showing that either calcium phosphate or phos-
phatase has diffused into the underlying section. The intensity of staining
of the nuclei of the inert section has been deliberately exaggerated in the
photograph to provide contrast with the unstained background seen by

diffraction.

A

*

P

fll

•

Hj

m

•b

IP"

* a

*

u* '

i

i

it

*•

4f

0
$ 4

w

#

*»..

1

^f

*-

1S&

* »

*

IF

#

m

!

c

Plate V. Experiments on the diffusion of phosphatase during use of the
Gomori-Takamatsu technique. Figure A, guinea-pig kidney superimposed
on inert section of guinea-pig liver, incubated 1280 minutes. Note that the
staining artefact on the inert section extends seven or more cell diameters
from the opaque active section. Figure B, photograph of liver section.
This was an inert section, upon which an active kidney section was super-
imposed, in buffer, for 2 days. Then the active section was removed, and
the previously inert section incubated for 1280 minutes, with the result
shown in the figure. All the blackening shown represents phosphatase
activity. Since the active section was removed prior to incubation, the
factor which had diffused from the active to the inert section must be
either phosphatase or phosphatase activator. Figure C, intrinsic phos-
phatase of guinea-pig liver: incubation time 1280 minutes. The phospha-
tase intrinsic to the liver sections of Figs. A and B was destroyed. Com-
parison of the three figures shows that the artefact phosphatase has a
different distribution from the intrinsic phosphatase.

.' «,

If

> *

. ... . „•; • *«. . '••«%*;• -v >.\;4%*a^

♦ « • * km * ** * „ • ■ * » » *» * **

»*, •» Tf • * * • * ' «• ' » • . •

• i • • » • .», t. . •

• •.*•- - > • <.* V . »• • «».♦ *

. .* . •* &*• * »*. J*f" % JC if • • ■•• *

v ♦» «♦■ . * ., v » ** •• • *

* .'

• _. •

• , ♦ »

Jm ♦.••#**» «f

>•«

WW 4

:v 0*.w.

/fj)- *• o:

\ H B ft .A > ♦». /■ ^ S V U A H V |.

ML ^^L. . f i, 5 \ I ft i *x 4 r&

-•J-fl-'tv',.!.^^* i^y*?j

\ * r

B

Plate VI. Rat-kidney phosphatase by the method of Loveless and Danielli.
Figure A shows a section incubated with substrate and p-nitrophenylazo-
4-naphthol. This mixture contains no activator for brush-border phospha-
tase, so that only nuclear phosphatase is demonstrated. Figure B, same
as Fig. A, but incubation carried out with activators for both brush -border

and nuclear phosphatases.

*&•

A

B

Plate VII. Figure A, phosphatase of normal rat liver: note high concen-
tration in nucleoli. Figure B, phosphatase of regenerating liver: note
nuclear phosphatase concentrated on chromosomes (shown as equatorial
plate). Figure C, Feulgen of Walker rat sarcoma: fixative Carnoy. Figure

D, phosphatase of Walker rat sarcoma: note high concentration of phos-
phatase in nucleoli, mitotic chromosomes, and also on spindle * ixative
for phosphatase preparations, 80 percent alcohol; incubation time, 120

minutes.

#

•

gn

* •

D

Plate VIII. Walker rat sarcoma 4 days after treatment by nitrogen mus-
tard. Figures A and C, Feulgen; fixative, Carnoy. Figures B and D,
alkaline phosphatase by Gomori-Takamatsu method; fixative, 80 percent
alcohol; incubation time, 120 minutes. Note parallelism between distri-
bution of phosphatase and deoxy nucleic acid. Figures A and B show cells
in which spindle formation has failed, so that metaphase chromosomes are
scattered throughout the cytoplasm. Figures B and D show pycnotic

degeneration.

B

Plate IX. Alkaline phosphatase of Drosophila salivary chromosomes by

Gomori-Takamatsu method. Fixative, 3.8 percent acetic acid, followed

by 95 percent alcohol; incubation time, 1280 minutes.

Plate X. Alkaline phosphatase of onion root tips. Fixative, 80 percent
alcohol. Figure A, with glycerophosphate as substrate.

Plate X. Alkaline phosphatase of onion root tips. Fixative, 80 percent
alcohol. Figure B, with /3-naphtholphosphate as substrate.

•

*

B

Plate XI. Alkaline phosphatase of chick osteoblast cultures. Cultures
by H. B. Fell. Fixative 80 percent alcohol; incubation time, 1280 minutes;
substrate, glycerophosphate. Figure A, prophase and intermitotic nuclei:
note high phosphatase concentration in nucleoli, still visible in prophase,

and high concentration on prophase chromosomes. Figure B, note meta-
phase plate and anaphase. Figure C, note anaphase. Figure D, note meta-
phase plates. There appears to be phosphatase on the spindles and centro-

somes.

Plate XII. Figure A, aldehyde in a single liver cell. The large, clear
spherical body is the nucleus, and the two dark rings are zones of high
aldehyde content surrounding two fat droplets. There is also diffuse alde-
hyde in the cytoplasm. The surrounding cells are devoid of aldehyde.
Figure B, carbohydrate in rat intestinal mucosa, revealed by periodate
oxidation. Note high concentrations in the absorbing borders and in the
goblet cells, and a much lower concentration in the Golgi zone of the

absorbing cells.

ERRORS DUE TO SCATTERING 127

scattering and refraction of light in the object, as well as by the
absorption of photons by individual constituents of the object.
Since the analysis of the distribution of chemical substances in
the specimen depends entirely on the amount of absorption of
light, it is necessary to reduce the effects of scattering and refrac-
tion to a minimum, and, if these effects are still significant, to
make measurements of the errors due to them in order to correct
apparent absorption values. Caspersson has set out these prob-
lems in some detail. His chief conclusions are as follows:

1. The object must be clearly resolved. A satisfactory absorption curve
cannot be obtained if the diameter of the object is less than three times
the wavelength of the light used.

2. It is not sufficient to measure the absorption of light at one wavelength
only, because the absorption bands of the various components of tissues
overlap to a considerable degree.

3. The optical system must fulfill Abbe's sine condition. Otherwise the
distribution of light in the image does not correspond to that in the
object.

4. Every point in the object must be illuminated by incoherent light,
as, for example, by using Kohler's method of illumination.

5. A correction must be applied for the amount of light lost by scatter-
ing. This is best done by measuring the amount of light actually scattered.
A much more common procedure is to measure the loss of light in the
specimen at a wavelength at which the specimen is believed to have
practically no power of absorbing light. The light loss at this wavelength
is then assumed to be caused by scattering. It is then commonly assumed
that the scattering is varying inversely as the fourth power of the wave-
length, and the light loss by scattering at other wavelengths is calculated
on this basis. This method is by no means satisfactory, since, whereas the
loss of light varies as 1/Xn, where X is the wavelength, the value of n is a
function of particle size in the specimen and varies from 2 to 4 for particles
between 10 m/A and 10 fi in diameter. Since this is a range of particle
sizes to be expected in cells, the correction is obviously difficult to make
from observations at one wavelength. From this point of view there is
much to be said for working at as large a wavelength as is compatible with
adequate resolving power.

6. To minimise scattering and refraction, the specimen must be mounted
in a medium of as nearly as possible the same refractive index as the speci-
men. Caspersson suggests that the ratio of the two refractive indexes
should not exceed 1.1.

7. If the points mentioned above are adequately taken into considera-
tion, including the limit to the ratio of the refractive indexes of speci-
men and mounting medium, then Caspersson calculates that the minimum
permissible numerical aperture which can be used is 0.85.

128 QUANTITATIVE STUDIES

Adherence to the Beer-Lambert Law

When an adequate optical system has been used, and losses
of light by scattering and refraction minimised, the difference of
the intensities of the light incident upon the specimen and that
transmitted is due to absorption of photons. The question then
arises: what is the relationship between the amount of light
absorbed and the number of absorbing molecules. It is com-
monly assumed that this relationship is given by the Beer-Lam-
bert law, i.e.:

E = log I0/I = ecd

where E = extinction, I0 = intensity of light entering specimen,
I = intensity of light leaving specimen, e = extinction coefficient,
c = concentration of absorbing substance, and d = thickness of
absorbing layer. This postulated linearity between the extinc-
tion and concentration is by no means an unbreakable rule, and,
unfortunately, in biological systems, as Caspersson has noted, it
is difficult to obtain adequate evidence of the extent to which the
rule is valid. The conditions prevailing in fixed specimens are
difficult to evaluate from the point of view of the relationship be-
tween E and c. Errors due to fluorescence may also be of im-
portance: these can be detected by use of a fluorescence micro-
scope.

Commoner has made valuable contributions to this field. In
an elegant paper he showed that the Beer-Lambert law was valid
for the pigments in Coleus hair cells. And in an interesting paper
written with Lipkin he pointed out that, where the specimen
under examination contains optically anisotropic molecules,
marked deviations from the Beer-Lambert law occur if the aniso-
tropic molecules are markedly oriented. In the case of an army
of molecules which are completely oriented, with unpolarised
light the extinction is proportional to the number of molecules
only at relatively low extinction values (less than 0.15). As the
concentration rises E deviates increasingly from the linear rela-
tionship, and as E approaches 0.3 the extinction becomes inde-
pendent of the number of absorbing molecules: the proportion of
light absorbed may never exceed 50 percent. With such a sys-
tem, containing oriented anisotropic molecules, it is possible to

ERRORS DUE TO STATES OF AGGREGATION 129

obtain reliable results only if the specimen is examined with

polarised light.

Although it is probable that serious errors due to orientation
of anisotropic molecules are rare, the danger of error is none
the less very real. It has been shown by Schmidt (1937), Frey-
Wyssling (1948), Caspersson (1950), and others that nucleic acid
molecules are markedly anisotropic and quite commonly oriented.

Errors due to States of Aggregation

These may be classified into two groups: errors due to the
aggregation of individual molecules, and errors due to aggregates
of size near the limit of resolution.

Even when the group which is absorbing light is known, it is
not always easy to define its absorption band in a specimen. The
position of an absorption band may often be readily altered
by the formation of complexes with other substances in the speci-
men. Well-known examples of this are the change in colour of
astaxanthin from red to blue when it is adsorbed on certain pro-
teins, and the wide variations in the spectrum of haem when it is
adsorbed upon different proteins. These changes in spectrum
are not restricted to the visible spectrum. Caspersson has found
a marked difference in the ultraviolet between the positions of the
maximum of absorption by pure nucleic acid and the maximum
found with nucleic acid in cells. These effects must be expected
whether the absorbing molecule is an intrinsic part of the cell
or an added dye; they are likely to be affected by the mode of

fixation.

The second effect may be most clearly appreciated if we con-
sider a chequerboard, as in Fig. 8. The square as a whole is the
area whose absorption is under consideration. The hatched
areas contain a high concentration of absorbing substance; the
clear areas have no absorbing material. Consider the case in
which the concentration of absorbing material in the hatched
areas is so high that, if a large area had the same composition
as the hatched areas, practically no light would be transmitted.
Then the actual absorption of light by a chequered area depends
upon the relationship between the distance d and the wavelength

of light.

If d » a, the chequers will be clearly resolved and the area will
transmit 50 percent of the incident light. If A » d, the chequers

130

QUANTITATIVE STUDIES

will not be resolved, and the transmission of light will be prac-
tically nil. If d is of the order of A/3 or A/4, the percentage of
transmission is unpredictable. The great danger therefore lies
in this region, since the chequers cannot be resolved when d is of
this order of A/3 or A/4. Thus, in a specimen in which the ab-
sorbing material is aggregated in this way, errors of hundreds of

Fig. 8. To illustrate the absorption of light by a chequerboard distribution
of an absorbing substance. The areas containing the absorbing substance

will be resolvable if d > X/2.

percent could arise in calculating the concentration of absorbing
substances from extinctions.

Probably the most effective way of surmounting this latter
difficulty is to study a specimen at several different regions of
the spectrum.

Errors due to Fixation and Diffusion

Displacement of substances in the cell by diffusion may occur
at the time of fixation or at a later stage of treatment. As was
indicated in Chapter 2, the only reliable method of fixation for
tissues is freeze-drying. Diffusion, however, may occur even
within the medium used for mounting a section, or during the
various stages of a cytochemical reaction. A method of study

ERRORS DUE TO FIXATION AND DIFFUSION 131

which has general validity for the detection of diffusion arte-
facts is to use superimposed sections, with the section to be stud-
ied superimposed upon a similar section which is devoid of the
substance under investigation. This technique has been dis-
cussed previously in this volume, and in two earlier papers
(Danielli, 1946, 1949). So far, however, no studies have been
made of diffusion during paraffin infiltration and embedding.
Although diffusion in paraffin is unlikely to be significant when
such substances as nucleic acid are considered, with many of the
substances for which the microincineration technique has been
used the errors may be serious.

It should be feasible to study diffusion in wax, by means of
the superimposed-section technique.

As a result of fixation marked increases usually occur in the
refractive-index differences between different parts of cells, and
as a result diffraction and light scattering become more pro-
nounced. Caspersson suggests that this may be minimised by
either of two methods. The first is that commonly practiced in
microscopic work — namely, mounting the specimen in a medium
of the same refractive index as the part of the specimen to be
examined. A second most ingenious device, probably applicable
only to frozen-dried material, is to make use of the very slow
diffusion of proteins, etc., which occurs in specimens mounted in
glycerol. When a frozen-dried specimen is mounted in glycerol
the boundaries of the various cell organs are sharp, and the re-
sultant abrupt discontinuities in refractive index cause marked
losses of light. But after some time in glycerol a small degree
of diffusion reduces the sharpness of the boundaries, so that
there are continuous rather than discontinuous changes in re-
fractive index, and light losses are correspondingly reduced.

Errors due to Inadequate Chemical Procedures

The first requirements of a -cytochemical reaction are that the
end-products should be known and either non-diffusible or read-
ily examined by techniques for detecting diffusion. Attention
has already been directed to these problems in preceding parts of
this book. For quantitative studies a further chemical problem
becomes prominent, for it is necessary to know to what extent
the cytochemical reaction is quantitative. The requisite infor-
mation is seldom available at the present time.

132 QUANTITATIVE STUDIES

This latter problem is by no means simple. Suppose, for ex-
ample, that we consider a reaction such as that of dinitrofluoro-
benzene with tyrosine. With free tyrosine the reaction is rapid
and quantitative, but with a protein steric effects may be suffi-
cient to mask some reactive groups from the reagent. When
denaturation occurs, some masked groups may be exposed to
the reagent, and other hitherto exposed groups may be masked.
Thus the results obtained may be a function of the degree of
denaturation. Material which has been frozen-dried is prob-
ably little denatured but, on the other hand, can rarely be used
without further denaturation. And the greater the denaturation,
the greater is the deviation of masked groups from that prevail-
ing in the intact cell. Consequently, before commencing a study
of proteins it is necessary to decide first whether what is wanted
is a measure of the total concentration of a particular amino acid,
or whether it is the concentration of these groups which is steri-
cally unmasked in the native protein. And, when this decision
has been made, it still remains to design an effective way of
getting the required information.

Damage Caused by Illumination

When work is being done in the visible region of the spectrum,
damage caused to the specimen by the incident illumination is
unlikely to be serious, but as one moves into the ultraviolet the
hazard steadily becomes greater. First, it becomes impossible to
make continuous studies upon living cells, and then as the energy
of the incident radiation increases damage to the specimen on
the chemical level may become evident. This hazard can prob-
ably be adequately allowed for by measuring changes in extinc-
tion with time.

With studies in the electron microscope damage again may be
caused by the electron beam. The magnitude of this damage
on the chemical level has yet to be explored.

REFERENCES

Caspersson. 1947. Symp. Soc. Exp. Biol, 1, 127.

1950. Cell Growth and Cell Function (Norton, New York).

Commoner. 1948. Ann. Mo. Bot. Gard., 85, 239.

1949. Science, 110, 31.

REFERENCES 133

Commoner and Lipkin. 1949. Science, 110, 41.
Danielli. 1946. Nature, 167, 755.

1947. Symp. Soc. Exp. Biol, 1, 101.

1949. Cold Spring Harbor Symposia, 14, 32.
Frey-Wyssling. 1948. Submicroscopic Morphology o] Protoplasm (Else-
vier, Amsterdam). .
Schmidt. 1937. £>ie Doppelbrechung von Karyoplasma, etc. (Berlin).

CHAPTER 7

THE FUTURE OUTLOOK
IN CYTOCHEMISTRY

Of the many fields of cytology, cytochemistry has probably
been that which has grown most rapidly and been the scene of
the most intense activity in the past ten years. The next ten or
twenty years must inevitably see the consolidation of this field,
the refinement of many methods, and the invention of many new
methods. These activities seem likely to fall into three groups:

1. The study of qualitative methods.

2. The study of quantitative methods.

3. The application of cytochemistry to the study of the fundamental
problems of cell biology.

Under the first of these headings we must envisage the exten-
sion of studies to other regions of the spectrum, including the far
ultraviolet and the infrared, the development of new microscopic
instruments, particularly reflecting microscopes and perhaps also
the television microscope now being studied by A. K. Parpart.
The refinement of electron-microscope methods is much to be
hoped for. Methods involving the use of radioactive isotopes
may become prominent. The investigation of fluorescence meth-
ods and the use of specific quenching agents can be anticipated.
Great prominence will probably be given to methods for the
localization of enzymes and the study of substrate specificity as a
means of revealing the diversity of enzymes in cells. And the
method recently published by Coons and his colleagues, of using
tagged antibodies, if sufficiently sensitive may be of the greatest
value in studying a wide variety of proteins.

In the study of quantitative methods probably the main tasks
are the defining of the limits of accuracy which can be obtained,
from both the chemical and the physical viewpoints. The de-
velopment of more sensitive photocells, for example, is likely to
be valuable, provided that the extent to which cytochemical re-

134

THE FUTURE OUTLOOK 135

actions are quantitative is adequately explored, and that more
attention is paid to losses of light by scattering, etc.

The biological problems which can be studied by cytochemical
methods are already numerous, and beyond doubt the number
will increase rapidly. A very profitable field to survey will be
that of finding the relationship between purified enzymes and
intracellular enzymes. To a considerable degree the biologist
must look upon purified enzymes as artefacts, and, as indicated
in the chapter on phosphatase, to proceed from knowledge of
purified enzymes to knowledge of their biological function may
be a long and difficult journey. It is certainly a journey which
can be shortened by cytochemical studies. Valuable examina-
tions will also be made of the extent to which the biochemical
theories of metabolic pathways are valid. At present almost
all the evidence on the nature of metabolic pathways has been
built up by in vitro studies of enzyme systems and by a limited
number of genetical studies. Some degree of confirmation of
these theories is being obtained by the use of isotopes, but it is
doubtful if isolation of isotope-tagged compounds from living
tissues will finally solve many problems. The chemist who can-
not think of a variety of alternative relationships between the
members of a given set of tagged intermediates is unusually lack-
ing in ingenuity. Cytochemical studies, by revealing the distri-
bution of enzymes and substrates in the different phases of cell
activity, can contribute greatly to this field.

The biochemistry of tissues is at present a very rough-and-
ready affair. The histologist knows that all tissues are com-
pounded of a wide variety of cells, all of which must be indis-
criminately ground up together in the formation of "acetone-ex-
tracted powders," homogenates, breis, and microsome suspensions,
etc. The cytochemist knows not only that these various cells
are chemically different but also that even cells of the same type
may, in different parts of an organ, be widely different. For ex-
ample, the hepatic cells close to the central vein in a hepatic
unit may differ widely from the more peripheral cells in their
content of nucleic acid, phosphatase, fat, and glycogen, and so
they may be in markedly different physico-chemical states. The
combination of the present undiscriminating approach of the bio-
chemist and the more perceptive but undeveloped approach of the
cytologist should be richly rewarding.

136 THE FUTURE OUTLOOK

The situation is the same in comparative biochemistry. At
present comparative biochemistry is founded on the results ob-
tained on a small number of so-called typical animals, plants,
yeasts, and bacteria. The sole basis of the claim that these or-
ganisms are typical rests upon the facts that all of them may,
by rather simple methods, be obtained in great quantities and
that many of them have been prominent in the preliminary medi-
cal biology courses and their equivalents. The great majority of
living organisms cannot be obtained in sufficient quantity for
biochemical studies without immoderate and unjustified expense.
But cytochemistry has the potentiality of throwing almost the
whole of the living world open to study by the chemist.

Medicine and agriculture cannot fail to benefit from the funda-
mental studies just outlined, but in addition to this there is prom-
ise for the direct sudy of pathological conditions by cytochemical
methods. For example, studies of gene action, of the evolution
of genes and viruses, and of differentiation hold great promise
for the elucidation of the nature of cancer and the development
of new drugs for the treatment of cancer. It has already been
shown that various tumours are particularly rich in certain
enzymes. This in itself has been effective in promoting the study
of new types of drugs. The drugs which are in use at present
lack selectivity so that, although they have a potent action on
tumour cells, their use is restricted by their high general toxicity.
At King's College we are now investigating the use of drugs
which are comparatively inert, but which, upon coming into con-
tact with an enzyme present in high concentration in a tumour
cell, liberate a derivative of very high activity. Thus we hope
that, by designing a drug to fit the enzymic pattern of a tumour,
we may attain a much higher degree of specificity of drug action.

Equally, we can expect assistance in diagnosis. A biopsy speci-
men can seldom provide sufficient material for a variety of bio-
chemical tests, but it is sufficient for dozens of cytochemical tests.

Thus the outlook in cytochemistry is full of promise. This
promise, however, will only be fulfilled if cytochemistry is used
as a rigorous technique by an adequately trained staff.

REFERENCE

Coons. 1952. Symp. Soc. Exp. Biol., 6 (in press).

INDEX

(For author references, see pages 15, 27, 77, 94, 124, 132, 136)

Absorption spectra, see Light
Acetal, see Aldehydes
Activators, 36, 52-54, Plates V, VI
Actomyosin, 73-75
Adenosine triphosphate, 73-75
Aldehydes, 9, 78-95, Plate XII

acetal significance, 92

artefacts, 84, 87

as chromogenic reagents, 105, 106

as protein-sensitivity reagent, 109

destruction, 83

fat transport and metabolism, 88-
91

fixation techniques and reagents,
78-89, 93, 94

glycol and, 92-94

supplementary tests, 82
Alkaline phosphatase, 6, 9, 28-77

activator, 52-54, Plates V, VI

artefacts, 29, 35-38, 42, 45, 46, 49,
52, 53, 59

azo compounds and dyes, 41, 42,
50

bone formation and, 65

calcium phosphate, 36-40, 42, 44,
46, 49, 53, 66, 67, Plate IV

cartilage and, 66, 67

collagen and, 64, 75, 76

definition, 28

dephosphorylation function, 61,
62, 68-71

diazonium hydroxides, 40, 41

diffusion, 44-49, 53, 54

excretory organs and, 71

Feulgen reaction, 56-58

Alkaline phosphatase, fixation tech-
niques and reagents, 29^44,

47-49, 56-58
freeze-drying, 29, 47-49
genes and, 57
glucose, 69, 70, 74
glycogen, 66, 67, Plate I
glycerophosphate technique, 36,

39, 42, 44, 59, Plates III, X,

XI
healing wounds, 63, 64
heat destruction, 55
hydrolysis, 54
hydrolytic functions, 60, 61
incubation time, 30, 33-35, 51-53
kidney and, 32, 34, 36, 38, 41, 42,

44, 49, 50-55, 69-71, Plates

III, IV, V, VI
mitosis, 55, 56, 60, 75, 76, Plate XI
/3-naphthol phosphate and, 39, 40,

59, Plate X
p-nitrophenylazo-a-naphthol

phosphate, 39, 42
nucleic acids and, 56, 57, 61, 63,

Plate VIII
nucleus and, 49-56, 58-63
phenolthalein phosphate, 39, 40,

42
phosphate esters, 39-42
alcohol moiety, 39, 42
phosphorylation, 60, 61, 68, 69, 70,

72, 74
physiological activity sites, 44
protein synthesis, 63, 64, 68, 75
research history, 28
secretion and, 68-73
section thickness, 42, 43

137

138

INDEX

Alkaline phosphatase, section thick-
ness, washing time, 43, 44

substrates, 39-42, 45, 59, Plates
VI, X, XI

superimposed-sections technique,

37, 38, 45, 49, 51-53, Plates
IV, V, VIII, IX

tissue differentiation and, 67, 68
Artefacts, 3, 4, 13, 18, 19, 25-29, 35-

38, 42, 45, 46, 49, 52, 53, 59,
60, 81, 84, 87, 109, 110, 121,
131, 135, Plates IV, V

Ascorbic acid, 5, 11, 64

Azo dyes, 41, 42, 50, 101, 108, 109

Bacteria, 117, 118-120, 122
Bacteriophage, 118-120
Beer-Lambert law, 128
Biology, cytochemical viewpoint, 4,

5, 6, 134, 135, 136
Blocking (non-chromogenic)

agents, 100, 106-108
Bones, 65, 118

Calcium phosphate, 36-40, 42, 44,

46, 49, 53, 66, 67, Plate IV
Carbohydrates, 89, 90, Plate XII
Cartilage, 66, 67
Chemistry, cytochemical viewpoint,

4

Chromocentre, 55, 75, 76, 115
Chromogenic reagents, 97, 99-106
Chromsomes, 15, 50, 55-57, 59, 61,

75, 99, 114, Plates VII, VIII,

IX, XI
Collagen, 64, 75, 76
Cytochemistry history, 2-4
Cytoplasm, 8, 60, 62, 63, 89, 90, 94,

99, 113, Plates II, IV, VIII,

XII

Dephosphorylation, 61, 62, 68-71
Diazonium hydroxides, 40, 41, 101—

103, 109
Digestive tract, 52, 89, 117, Plate

XII

Enzymes, 6, 7, 10, 12, 14, 16, 18, 19,
28, 29, 57, 58, 96, 134-136
See also Alkaline phosphatase
Excretory organs, 71

Fats, aldehydes and, 88-91, Plate

XII
Feulgen reaction, technique, and

characteristics, 5, 12, 13, 56-

58, 76, 78, 84, 85, 94, 115,

Plates VII, VIII
Fixation and fixatives (reagents),

2, 3, 6, 16-27, 29-44, 47-49,

56-58, 130, Plates I, II, III
Freeze-drying, 6, 13, 16, 19-27, 29,

47-49, 81, 121, 130, 131, Plates

I, II, III

Genes, 57, 61, 115
Glucose, 69, 70, 74
Glycogen, 66, 67, Plate I
Glycols, 92-94

Glycerophosphate, 36, 39, 42, 44, 59,
92, Plates III, X, XI

Hydrolysis, 12, 13, 64

Infrared, see Light
Inhibitors, 36

Keratin, 64

Kidney, 32, 34, 36, 38, 41, 42, 44, 49,

50-55, 69-71, Plates III, IV,

V, VI

Light, 4, 5, 13, 14, 97-108, 126-130,

132, 134
Liver, 9, 46, 52, 53, 55, 81, 85, 89-91,

Plates I, V, VII, XII

Maceration, 7, 50

Mitochondria, 9, 11, 24, Plate III
Mitosis, 55, 56, 60, 75, 76, Plate XI
Molecules, 4, 11, 16, 26, 70

Neurons, 117

Nitro compounds, 103-105
Nucleic acids, 5, 9, 12-15, 17, 56, 57,
63, 96-124, Plate VIII

INDEX

139

Nucleic acids, absorption spectra,
97-108
artefacts, 109-110
azo dye method, 108, 109
bacteriophage and, 118-120
chromogenic reagents, 97, 99-106
aldehydes, 105, 106
diazonium hydroxides, 101-103
nitro compounds, 103-105
classification of groups, 96
electron microscope methods,

110-112
fluorescence methods, 108
relation to protein synthesis, 113,
114, 116-124
chemical system, 115
cytological systems, 114, 115
study methods, 97
sugar linkage method, 109
Nucleolus, 55, 56, 59, 60, 67, 68, 75,

76, 115, Plates III, VII, XI
Nucleus, 8, 9, 49-56, 58-63, 94, 115,
Plates II, III, IV, VIII, XII

Phosphate esters, 39-42, 67
Phosphorylation, 60, 61, 68, 69, 70,

72, 74
Physics, cytochemical viewpoint, 4
Plants, 58, 59, 63, 99, 117, Plate X
Proteins, 8, 15-17, 63, 64, 72, 73, 75,
96-124
adsorption of dyes, 72
artefacts, 109, 110
azo dye method, 108
bacteriophage and, 118-120
chromogenic reagents, 97, 99, 106
aldehydes, 105, 106
diazonium hydroxides, 101-103
nitro compounds, 103-105
classification of groups, 96
electron microscope methods,

110-112
fluorescence methods, 108
non-chromogenic blocking agents,

100, 106-108
relation to nucleic acids, 113, 114,
116-124

Proteins, relation to nucleic acids,
chemical system, 115
cytological systems, 114, 115
study methods, 97
sugar linkage method, 109
synthesis, 63, 64, 68, 75, 113,
114, 116-124

Qualitative methods, future, 134
Quantitative studies, 126-133
errors due to, Beer-Lambert law,
128
diffusion, 130
fixation, 130
inadequate chemical procedure,

131
scattering and refraction, 126,

127
states of aggregation, 129
future, 134

Reagents, see also Fixation

chromogenic, 97, 99-106
Ribonuclease, 14

Salivary glands, 56, 57, 76, 84, Plate

IX
Sarcoma, see Tumours
Scales, fish, 66
Silk, 64

Spermatozoa, 112
Spleen, 38
Substrates, 36, 39-42, 45, 59, Plates

VI, X, XI
Superimposed-sections technique,

37, 38, 45, 49, 51-53, 85, 131,

Plates IV, V, VIII, IX

Teeth, 66

Tissue culture, 55, 59, 60
Tumours, 55, 56, 60, 68, 88, 91, 117,
Plates II, VII, VIII

Ultraviolet, see Light

Vitamin C, 64

Wounds, 36, 38, 63, 64

Yeasts, 117