^

Biological Structure and Function
Volume I

Biological Structure
and Function

Proceedings of the First lUB I UBS International
Symposium Held in Stockholm, September 12-17. 1960

Edited by

T. W. GOODWIN

Depart lueni of Agrieu/tural Biocheiuistr

Institute of Rural Science

Peng la is, H 'ales

O. LINDBERG

The ll'enner-Gren Institute for

Experimental Biology, University

of Stockholm. Sweden

Volume I

1961

ACADEMIC PRESS • LONDON ■ NEW YORK

ACADEMIC PRESS INC. (LONDON) LTD.

17 OLD QUEEN STREET

LONDON, S.W.I

U.S. edition published by

ACADEMIC PRESS INC.

1 11 FIFTH AVENUE

NEW YORK 3, NEW YORK

Copyright © 1961, by Academic Press Inc. (London) Ltd.

Library oj Congress Catalog Card Number: 61-17329

PRINTED IN GREAT BRITAIN BY

SPOTTISWOODE, BALLANTYNE & CO. LTD.,

COLCHESTER AND LONDON

Contributors to Volume I

Per-ake Albertsson, Institute of Biocliemistry, Uppsala, Sweden.

Vincent G. Allfrey, The Rockefeller Institute, Nezv York, N.Y., U.S.A.

Erik Arrhenius, The W'enner-Gren Institute for Experimental Biologv,
University of Stockholm, Szveden.

Kare Aspberg, Research Laboratory, AB Pharmacia and the Institute of
Biochemistry, Uppsala, Szveden.

GuNNAR Blix, Medicinsk-kemiska Institutionen, Uppsala, Szveden.

H. G. BoMAN, The Institute of Biochemistrv, Uppsala, Szveden.

I. A. BoMAN, The Institute of Biochemistry, Uppsala, Szveden.

P. N. Campbell, Courtauld Institute, Middlesex Hospital, London, England.

E. S. Canellakis, Department of Pharmacohgw Yale University Medical

School, Nezv Haven, Connecticut, U.S.A.

H. Chantrenne, Laboratory of Biological Chemistry, Faculty of Sciences,
University of Brussels, Belgium.

Erwin Chargaff, Department of Biochemistry, Columbia University, Nezv
York, N.Y., U.S.A.

J. N. Davidson, Department of Biochemistry, University of Glasgozv,
Scotland.

Albert Dorfman, The LaRabida- University of Chicago Institute and the
Departments of Pediatrics and Biochemistry, University of Chicago,
Chicago, Illinois, U.S.A.

Per Flodin, Research Laboratory, AB Pharmacia and the Institute of
Biochemistry, Uppsala, Szveden.

F. Haguenau, Laboratoire de Medecine Experimentale du College de France,

Paris, France.

Mirjam Halmann, Department of Biology, Brookhaven National Labora-
tory, Upton, Long Island, A', i'., U.S.A.

Edward Herbert, Department of Biology, Massachusetts Institute of
Technology, Cambridge, Massachusetts, U.S.A.

Shlomo Hestrin, Department of Biological Chemistry, The Hebrezv
University, Jerusalem, Israel.

C. H. W. HiRS, Department of Biology, Brookhaven National Laboratory,
Upton, Long Island, N.Y., U.S.A.

Gottwalt Christian Hirsch, Zoologisches Institut, Gottingen, Germany.

VI CONTRIBUTORS TO VOLUME I

K. H. HoLLMANN, Laboratoire de Medecine Experimentale du College de
France, Paris, France.

H. HoLTER, Carhberg Laboratory, Copenhagen, Denmark.

Tore Hultin, The Wenner-Gren Institute for Experimental Biology,
University of Stockliolm, Sweden.

J. C. Kendrew, MRC Unit for Molecular Biology, Cavendish Laboratory,
Cambridge, England.

Jadwiga H. Kycia, Department of Biology, Brookliaven National Labor-
atory, Upton, Long Island, N.Y., U.S.A.

W. K. Maas, A^ra' York University School of Medicine, Nezv York, N.Y.,
U.S.A.

WiNFiELD S. Morgan, Department of Pathology, Massachusetts General
Hospital, Boston, Massachusetts, U.S.A.

H. R. Perkins, National Institute for Medical Research, Mill Hill, London,

England.
Gertrude E. Perlmann, The Rockefeller Institute, New York, N. Y.,

U.S.A.

Peter Perlmann, The Wenner-Gren Institute for Experimental Biology,
University of Stockholm, Sweden.

Keith R. Porter, The Rockefeller Institute, Nezv York, N.Y., U.S.A.

Peter Reichard, Department of Chemistry I, Karolinska Institutet, Stock-
holm, Sweden.

H. J. Rogers, National Institute for Medical Research, Mill Hill, London,
England.

Sara Schiller, The LaRabida-University of Chicago Institute and the
Departments of Pediatrics and Biochemistry, University of Chicago,
Chicago, Illinois, U.S.A.

Philip Siekevitz, The Rockefeller Institute, Nezv York, N.Y., U.S.A.

A. Tiselius, Institute of Biochemistry, Uppsala, Sweden.

Alexandra von der Decken, The Wenner-Gren Institute for Experimental
Biology, University of Stockholm, Sweden.

M. H. F. WiLKiNS, Physics Department, King's College, London, England.

Preface

In 1956 The International Union of Biological Sciences (lUBS) decided
to set up a Biochemistry Section Committee, which would be a Co-
ordinating Committee between lUBS and the International Union of
Biochemistry (lUB) and, through a Co-ordinating Committee of lUB and
the International Union of Pure and Applied Chemistry (lUPAC), would
also have contact with lUPAC. It was considered that the Committee would
be specifically concerned with chemical biology within the framework of
the Unions federated to the Councils of Scientific Unions (ICSU). The
members of the Biochemistry Section Committee are at present : R.
Brunei (Toulouse) and O. Lindberg (Stockholm) (appointed by lUBS),
A^I. Florkin (Liege) and T. W. Goodwin (Aberystwyth) (appointed by
lUB), and P. Boyer (Minneapolis) and F. Lynen (Munich) (co-opted
members). Florkin and Goodwin were elected Chairman and Secretary
respectively.

The first Committee meeting was held in 195S during the 4th Inter-
national Congress of Biochemistrv in Vienna. It had been visualized
throughout the discussions that an important function of the Committee
would be to make suggestions for various International Svmposia to both
lUBS and lUB. It was agreed that subjects would be appropriate only if
both biochemistry and the biological sciences were combining to produce
a rapidly expanding sphere of knowledge. A number of possibilities were
considered at Vienna and it was eventually decided that "Biological
Structure and Function" was most appropriate at this time. This idea
was accepted by the two International Unions and plans began to be
formulated. It was readily agreed that the most suitable centre in
Furope for such a symposium was theWenner-Gren Institute, with its w'ell-
established, international reputation in this field and, furthermore, the
project had the blessing and support of Dr. Axel Wenner-Gren himself,
who honoured the Symposium by agreeing to act as Patron of Honour
and by attending the Inaugural Session to deliver the opening address.

The lUB and lUBS have supported this Symposium financially but
the realization of the Symposium would not have been possible without
the generous aid of the Wenner-Gren Foundation, and of the various
bodies in different countries which support the attendance of scientists at
important international meetings. It was extremely satisfying to the

PREFACE

organizers to know that these official bodies considered this First lUB/
lUBS Joint Symposium worthy of support, and mention should be made
of the National Science Foundation which supported so many of our
U.S. participants; furthermore, in this connection the work done on our
behalf by Dr. Elmer Stotz, the treasurer of lUB, should not be forgotten.

The organizers hope that this Symposium will be the forerunner of a
long line of similar international symposia based on fruitful co-operation
between biochemists and biologists from all nations.

The organizers are most grateful to the Institute of Physics, University
of Stockholm, for their generosity in putting their attractive new lecture
theatre at the disposal of the Symposium.

In preparing the proceedings for the press the organizers have been
greatly helped by Miss J. T. Peel, who transcribed the recorded dis-
cussions, and by Mr. D. J. Howells, who prepared the subject index.

April, 1 96 1 T. W. Goodwin

O. LiNDBERG

Contents of Volume I

Contributors to Volume I

Preface

Contents of Volume 11

Macromolecular Structure and Function
Introduction. By A. Tiselius ....... 3

The Structure of Globular Proteins. By J. C. Kendrew . . 5

Molecular Configuration of Nucleic Acids. By M. H. F. Wilkiub . 13

Partition of Alacromolecules in .Aqueous Two-Phase Systems. Bv

Per-Ake Albertsson . . . . . . . 33

The Reactivity of Certain Finictional (jroups in Ribonuclease A
towards Substitution by i-Fluoro-2,4-dinitrobenzene. Inacti-
vation of the Fnzyme by Substitution at the Lysine Residue in
Position 41. By C. II. W. Hirs, Mirjam Halmann and Jadwiga
H. Kycia . . . . . . . . . 41

The Relation of the Secondary Structure of Pepsin to Its Biological

Activity. By Gertrude F. Perlmann ..... 59

The Problem of Nucleotide Sequence in Deoxyribonucleic Acids.

By Erwin Chargaft" ........ 67

Problems in Polynucleotide Biosynthesis. By J. N. Davidson . 95

Fnzymic Formation of Deoxyribonucleic Acid from Ribonucleo-
tides. By Peter Reichard . . . . . . .103

Studies on the Mechanism of Synthesis of Soluble Ribonucleic

Acid. By E. S. Canellakis and Edward Herbert . . .113

Microsomes and Protein Synthesis

The Endoplasmic Reticulum : Some Current Interpretations of Its
Form and Functions. By Keith R. Porter ....

Pinocytosis. By H. Holter .......

The Ergastoplasm in the Mammary Gland and Its Tumours: An
Electron Microscope Study with Special Reference to
Caspersson's and Santesson's A and B Cells. Bv F. Haguenau
and K. H. Hollmann .......

The P^xternal Secretion (jf the Pancreas as a Whole and the Com-
munication between the Endoplasmic Reticulum and the
Golgi Bodies. By Gottwalt Christian Ilirsch

127
157

169

^95

C A1 fth»

CONTENTS OI" VOLUME I

PAGE

Immunological Studies of Microsomal Structure and Function. By

Peter Perlmann and Winfield S. Morgan .... 209

Amino Acid Incorporation by Liver Microsomes and Ribonucleo-
protein Particles. By Tore Hultin, Alexandra von der Decken,
Erik Arrhenius and Winfield S. Morgan . . . .221

The Effects of Spermine on the Ribonucleoprotein Particles of

Guinea-Pig Pancreas. By Philip Siekevitz .... 239

The Correlation between Morphological Structure and the Syn-
thesis of Serum Albumin by the IMicrosome Fraction of the Rat
Liver Cell. By P. N. Campbell 255

Amino Acid Transport and Early Stages in Protein Synthesis in

Isolated Cell Nuclei. By Vincent Ci. Allfrey . . . 261

Effects of 8-Azaguanine on the Specificity of Protein Synthesis in

Bacillus cereiis. By H. Chantrenne . . . . .281

Purme and Pyrimidine Analogues and the Mucopeptide Biosyn-
thesis in Staphylococci. By H. J. Rogers and H. R. Perkins . 289

Studies on the Incorporation of Arginine into Acceptor RNA of
Escherichia coli. By H. G. Boman, I. A. Boman and W. K.
Maas .......... 297

POLYS./VCCHARIDES

Introduction. By Gunnar Blix . . . . . .311

The Growth of Saccharide Macromolecules. By Shlomo Hestrin . 315

Mucopolysaccharides of Connective Tissue. By Albert Dorfman and

Sara Schiller ......... 327

Separation of Oligosaccharides with Gel Filtration. By Per Flodin

and Kare Aspberg ........ 346

Author Index . . . . . . . . -351

Subject Index . . . . . . ... -359

Contents of Volume II

Mitochondrial Structure and Function

Effects of Thyroxine and Related Compounds on Liver Mitochondria in

Vitro. Bv Olov Lindberg, Hans Low, Thomas E. Conover, and Lars

Ernster
Components of the Energy-couphng Mechanism and Mitochondrial

Structure. By Albert L. Lehninger
Ascorbate-Induced Lysis of Isolated Mitochondria — A Phenomenon

Different from Swelling Induced by Phosphate and Other Agents.

Bv F. Edmund Hunter, Jr.
Integrated Oxidations in Isolated Mitochondria. By J. B. Chappell
Metabolic Control of Structural States of Mitochondria. By Lester Packer
Stable Structural States of Rat Heart Alitochondria. Bv F. A. Holton and

D. D. Tyler

Solubilization and Properties of the DPNH Dehydrogenase of the Respir-
atory Chain. By Thomas P. Singer

Reversal of Electron Transfer in the Respiratory Chain. By Britton
Chance

Function of Flavoenzymes in Electron Transport and Oxidative Phos-
phorvlation. By L. Ernster

Coupling of Reduced Pyridine Nucleotide Oxidation to the Terminal
Respiratory Chain. By Thomas E. Conover

Mitochondrial Lipids and Their Functions in the Respiratory Chain. By

E. R. Redfearn

The Functional Link of Succinic Dehydrogenase with the Terminal

Respiratory Chain. By Giovanni Felice Azzone
Pyridine Nucleotides in Mitochondria. By E. C. Slater, M. J. Baillie and

J. Bouman
Nucleotides and Alitochondrial Function: Influence of Adenosinetri-

phosphate on the Respiratory Chain. By Martin Klingenberg
The Role of ATPase in Oxidative Phosphorylation. By Maynard E.

Pullman, Harvey S. Penefsky and E. Racker
The Mechanism of Coenzyme Q in Heart Mitochondria. By Daniel M.

Ziegler
Reactions Involved in Oxidative Phosphorylation as Disclosed by Studies

with Antibiotics. By Henry Lardy

Structure and Function of Chloroplasts and Chromatophores
Chairman's Opening Remarks. By T. W. Goodwin
Haem Protein Content and Function in Relation to Structure and Early

Photochemical Processes in Bacterial Chromatophores. By Martin D.

Kamen

Xll CONTENTS OF VOLUME II

Observations on the Formation of the Photosynthetic Apparatus in
Rhodospirillum rubrum and Some Comments on Light-Induced Chro-
matophore Reactions. B}^ Douglas C. Pratt, Albert W. Frenkel and
Donald D. Hickman.

The Photosynthetic Macromolecules of Chlorobiuni t/iiosulfatophilum.
By J. A. Bergeron and R. C. Fuller

Some Physical and Chemical Properties of the Protochlorophyll Holo-
chrome. By James H. C. Smith

Photosynthetic Phosphorylation and the Energy Conversion Process in
Photosynthesis. By Daniel I. Arnon

The Mechanism of the Hill Reaction and Its Relationship to Photo-
phosphorylation. By Birgit Vennesland

Electron Transport and Phosphorylation in Light- Induced Phosphoryla-
tion. By Herrick Baltscheffsky

Reduction of Dinitrophenol by Chloroplasts. By J. S. C. Wessels

The Relationship between " Methaemoglobin Reducing Factor" and
" Photosynthetic Pyridine Nucleotide Reductase." By H. E. Davenport

ATP Formation by Spinach Chloroplasts. By Andre T. Jagendorf and
Joseph S. Kahn

Intact Cellular Structure and Function

Chairman's Introduction: Remarks on Control of Structure and Differen-
tiation in Cells and Cell Systems. By J. Runnstrom
The Central Problems of the Biochemistry of Cell Division. By Daniel Mazia
Studies on the Cellular Basis of Morphogenesis in the Sea Urchin. By T.

Gustafson
Cell Differentiation: a Problem in Selective Gene Activation through
Self-Produced Micro Environment Differences of Carbon Dioxide
Tension. By W. F. Loomis
RNA Synthesis in the Nucleus and RNA Transfer to the Cytoplasm in

Tetrahymena pyriformis. By D. M. Prescott
Cell Division and Protein Synthesis. By Eric Zeuthen
Structure and Function in Amoeboid Movement. By Robert D. Allen
Some Problems of Ciliary Structure and Ciliary Function. By Bjorn A.
Afzelius

Specific Membrane Transport and Its Adaptation
Chairman's Introduction. By Bernard D. Davis
Approaches to the Analysis of Specific Membrane Transport. By Peter

Mitchell
Protein Uptake by Pinocytosis in Amoebae: Studies on Ferritin and

Methylated Ferritin. By V. T. Nachmias and J. M. Marshall
Comparative Study of Membrane Permeabilitv. By E. Schoffeniels
Active Transport and Membrane Contraction-Expansion Cycles. By R. J.
Goldacre

MACROMOLECULAR STRUCTURE AND FUNCTION

Introduction

A. TiSELIUS

Biochemical Institute, The University of Uppsahr, Sweden

I think it is very interesting and significant that at this first joint
Symposium of the International Union of Biochemistry and the Inter-
national Union of Biology you have asked an ex-president of the Inter-
national Union of Pure and Applied Chemistry to be in the chair during
the first day's discussion of macromolecular structure and function. I have
ventured to interpret this as signifying that these problems, so fundamental
in biochemistry and biology, can now be discussed also in strictly chemical
terms, and it is very gratifying that the host Unions in their activities do
not feel themselves limited by barriers between different disciplines,
barriers which no longer have any meaning.

It is very gratifying that among our speakers today we have several to
whom we owe some of the most fundamental recent discoveries in this
field which have made a symposium with today's subject possible and worth
while for discussion.

Macromolecular substances in biochemistry were until not very long
ago confined to what one might call "pre-structural" chemistry. That is
to say, one had to content oneself with isolation and characterization,
perhaps involving also one or two essential structure details. This is still a
very important field of biochemistry, but it is a great advance that today
we know that an almost complete structure analysis, also in a strictly
chemical sense, can now be made on some typical and particularly im-
portant substances of this category. The methods of approach to such
structural problems have proved of such general applicability that we have
the right to be optimistic about further advances in the near future. It is
particularly important just now to observe how different modes and ways
of attacking problems of structure among biochemically important
macromolecules at last merge together so that information of difi^erent
kinds can be utilized in the final attempt to unveil a structure. This is, I
understand, what is now happening in the field of protein structure, when
X-ray crystallographic analysis data can now be combined with amino
acid sequence determinations to work out the details of the structure (e.g.
the position of the side chains). Something similar is happening in the
nucleic acid field, although the advance there has naturally been slower.

When discussing biochemical function in relation to structure on a
molecular level we are thus now gradually approaching the state in which
we can say that we share the problems and the difficulties in some of the
most advanced fields of physics and chemistry. The situation makes me

4 A. TISELIUS

recall a conversation I had with a distinguished elderly professor of
organic chemistry when I was a young man and found myself gradually
drifting from physical chemistry into biochemistry. "Be careful," he said,
"this is a dangerous subject, and you will never become a good bio-
chemist anyhow. What is the use of studying enzymes and the mechanism
of their action until the phenomena of catalysis have been analyzed and
elucidated by the physical chemists ?" I have always doubted whether he
was right in this, and today I doubt it still more. I do not propose to say
that the biochemists can get much further than the physical chemists in
providing the ultimate explanation for these and similar phenomena, but
I do believe that they have brought to light a number of observations
which must be of fundamental importance in any attempt to work out a
general theory of catalytic function in the inorganic world. Thus the
study of the functions of substances of complicated structure — such as we
find in biological materials — may contribute just as much, or even more,
to our basic knowledge in the field in general as a study confined only to
the very simplest inorganic reactions. It would appear natural that this
should be so: the phenomena of catalysis which we meet in biological
matter are among the most striking and most specific found in Nature.
Thus, quite aside from their interest per se, it must be worth while to
study them from a very general and advanced point of view.

Discussion of function on a molecular level usually involves mechanisms
of activation and the structural background of highly specific affinities.
This is of course the case when we deal with reactions between, for example,
a macromolecular enzyme and substrate molecule of smaller size. But it is
also true when we deal with those specific interactions between difi^erent
kinds of macromolecules which appear to play such an important role
in the organized chemical reactions characteristic of life. Here also,
structural aspects come into the picture and we then gradually find our-
selves discussing organization rather than intermolecular structures
without being able to define a borderline in the application of these terms.
This is also reflected in the methods used, where electron microscopy
goes hand in hand with methods of structure analysis such as X-ray
crystallography and organic chemistry have provided. I shall not dwell
upon this interesting field of "molecular biology", as this rather belongs
to the programme of the following days — especially the forthcoming dis-
cussion about structure and function of certain submicroscopic particles.
But I wish to emphasize that here again we have a field which is yielding
much new information of a very general importance — also to the chemists.
The elusiveness of the structures and functions involved in this highly
organized matter is of a kind where biologists and biochemists with their
gentle methods and somewhat greater reverence for Nature in its intact
forms are more likely to succeed than the chemists.

The Structure of Globular Proteins

J. C. Kendrevv

MRC Unit for Molecular Biology,

Cavendisli Laboratory,

Cambridge, England

Proteins have probably been more intensively studied than any other
class of molecule. Not only have they been subjected to exhaustive in-
vestigation by the classical techniques of organic chemistry, culminating
in the determination of the complete amino acid sequence of insulin (by
Sanger) and later of several other proteins ; but also a whole armoury of
physicochemical methods has been used to interpret the behaviour of
proteins in solution and, finally, the kinetics and specificity of enzyme
reactions have been investigated in great detail. All this work has had the
object of understanding the function and biosynthesis of proteins in living
organisms. Hitherto the chief obstacle in the way of applying the results
of these researches to biological problems has been our ignorance of the
structure of proteins, that is to say, of the three-dimensional arrangement
of the atoms of which they are composed. The amino acid sequence is in
efi^ect a topological description ; but in a molecule as complex as a protein
topography is much more important than topology, because it is the spatial
relations between the side-chains which determine the chemical behaviour
of the molecule, and these relations cannot be determined, except in a
fragmentary manner, by chemical methods.

We are now for the first time in a position to appreciate the general
principles of protein architecture, indeed in one or two cases to under-
stand their application in some detail. If we work upwards in the hierarchy
of organization of protein molecules we find a remarkable alternation
between the simple and the complex in structural arrangement. At the
lowest level, the polypeptide chain itself is of the utmost simplicity, having
the same backbone structure of repeated peptide groups whatever the side-
chains attached to them (with proline as the only exception). When we
examine the amino acid sequence, however — the so-called primary
structure — we find a bewildering irregularity; there are no discernible
periodicities in the sequence, which in some places seems to be entirelv
random and in others highly non-random, several identical side-chains
being grouped closely together. At the next level, the spatial configuration

6 J. C. KENDREW

of the polypeptide chain (the secondary structure), we find simphcity and
regularity once more. This regularity was first appreciated by Astbury in
his classical studies of fibrous proteins, resulting in his classification of
polypeptide chain configurations into three main types, to one of which
almost all known fibrous proteins conform. The most important of these
is the so-called a-configuration, and its structural basis was revealed by

(

<

^

0

\

0^

r V V

FRANK

Fig. I. Model of myoglobin based on the three-dimensional Fourier syn-
thesis at a resolution of 6 A, showing the general arrangement of the polypeptide
chain and the position of the haem group.

Pauling and Corey [i] when they discovered the a-helix. For some time
indirect evidence has been accumulating that the a-helix is a structural
element in the globular as well as in the fibrous proteins, but definite
proof of this has only recently been obtained, in the structure analysis of
myoglobin.

The first stage of the X-ray analysis of myoglobin [2] gave a three-
dimensional picture of the molecule at a resolution of 6 A (Fig. i), reveal-
ing the general arrangement of the polypeptide chain and of the haem

THE STRUCTURE OF GLOBULAR PROTEINS 7

group, in other words of the tertiary structure of the molecule. The
tertiary structure is highly irregular and complex, in sharp contradiction
to the simplicity of the secondary structure. More recently [3] the resolu-
tion of the analysis has been increased to 2 A (Fig. 2). Although neigh-
bouring covalentlv bonded atoms are still not resolved, it is now possible
to separate atoms which are hydrogen bonded or in Van der Waals contact,
with the result that the atomic arrangement of most of the molecule can

Fig. 2. Model of myoglobin based on the three-dimensional Fourier syn-
thesis at a resolution of 2 A. The model is seen from the same point of view as that
of Fig. I. The course of the main-chain is indicated by a white cord: side-chains
have been inserted wherever possible.

be inferred. It turns out that all the straight regions of polypeptide chain
are in the a-helical configuration ; in fact the molecule consists of eight seg-
ments of a-helix joined by irregular regions of varying length ; the helical
segments comprise 75",, of the amino acid residues, in agreement with
estimates made on the basis of optical rotation and deuterium exchange
studies. The appearance of the haem group corresponds closely with
theoretical expectation, and it can be seen that the iron atom is attached
to a neighbouring a-helix bv means of a group which is almost certainly
the imidazole ring of a histidine residue.

8 J. C. KENDREW

In the Fourier synthesis the side-chains can be seen as dense regions
emerging from the hehcal main chain at intervals corresponding to the
parameters of the a-hehx. A close examination of these regions often
makes it possible to identify a side-chain with certainty; in other cases
some ambiguity remains but the choice of side-chain can be reduced to
two or three. At the present resolution about one-third of the side-chains
can be identified with certainty, and another third with fairly high prob-
ability. It now becomes possible to correlate these X-ray results with the
preliminary data obtained by Dr. A. Edmundson who is engaged in
working out the amino acid sequence of myoglobin by chemical methods.
He has broken down the molecule into short peptides by tryptic digestion,
and has determined the composition, and in a few cases the internal
sequence, of these peptides. By comparing his results with our own it has
proved possible to place almost all the tryptic peptides along the polypep-
tide chain, and the order of peptides thus ascribed corresponds with the
order which has in a few cases been suggested by Edmundson on the basis
of a preliminary examination of the chymotryptic digest. A few discrepan-
cies remain, but although the amino acid sequence has not yet been com-
pletely determined, its main features are now beyond doubt. We are now
engaged in an attempt to increase the resolution of the X-ray results
still further, and we hope that the remaining ambiguities will then be
removed.

Large molecules are often built up of sub-units, whose spatial arrange-
ment may be called the quaternary structure. At this level of organization
we return once more to simplicity and symmetry. Thus recent X-ray
studies of haemoglobin by Perutz and his collaborators [4], resulting in a
three-dimensional Fourier synthesis with a resolution of 5.5 A, have
shown that in this protein the four sub-units are arranged in the most
symmetrical manner possible, namely at the vertices of a tetrahedron.
Another very remarkable result has also emerged, namely that each of the
four sub-units, consisting of a single polypeptide chain together with a
haem group, very closely resembles the molecule of myoglobin in tertiary
structure.

In still bigger molecules, such as the viruses, the number of sub-
units may be very large, nevertheless their arrangement is highly regular.
For example in tobacco mosaic virus there are about 2000 sub-units
arranged in the form of a helix ; in the spherical viruses the sub-units are
arranged on the surface of regular or semi-regular polyhedra.

Thus we are now beginning to get a first glance at the general nature
of protein structure at all levels of complexity. It seems certain that
during the next 4 years these preliminary glimpses will lead to a detailed
picture of the structure of proteins which will give an immense impetus
to biochemistry generally, and indeed in many respects transform it.

THE STRUfTURE OF GLOBULAR PROTEINS 9

References

Pauling, L., Corey, R. B., and Branson, H. R., Proc. iiat. Acad. Sci., Wash.

37, 205 (1951).

Bodo, G., Dintzis, H. M., Kendrew, J. C, and WyckofF, H. W., Proc. row Soc.

A 253, 70 (1959)-

Kendrew, J. C, Dickerson, R. E., Strandberg, B. E., Hart, R. G., Davies, D. R.,

Phillips, D. C, and Shore, V. C, Nature, Loud. 185, 422 (i960).

Perutz, M. F., Rossmann, M. G., Cullis, A. P., Muirhead, H., Will, G., and

North, A. C. T., Nature, Land. 185, 416 (i960).

Discussion

TisELius : Is it possible by comparing the structure, derived by your crystallo-
graphic methods, of reduced and oxidized haemoglobin or myoglobin to get any
hints about any structural changes which would accompany the combination with
oxygen ?

Kendrew : A crystal of met-myoglobin can very easily be converted into the
reduced form by diffusing into it a solution of sodium dithionite and watching the
colour change. The crystal is quite unharmed by this procedure, and its X-ray
pattern is virtually identical with that of met-myoglobin. If the same experiment is
performed with haemoglobin, the result is quite different; haemoglobin crystals
on reduction fall to pieces, and if one begins with a solution of reduced haemoglobin
and adds salt the crystals which are formed are quite different from those of met-
haemoglobin. The simplest hypothesis which would explain these results is that
during oxygenation the haemoglobin molecule changes shape, the sub-units moving
relative to one another : in myoglobin no such change could occur because there are
no sub-units. This idea is purely speculative at present, but my colleague Dr.
Perutz is now beginning a study of crystals of reduced haemoglobin with the object
of discovering exactly what differences there are between its structure and that of
met-haemoglobin.

Chan'CE: It is obvious that Dr. Kendrew "s results are important not only for
those interested in the mechanism of oxygenation but also for those interested in
the mechanism of haemoprotein action where the histidine group connected to the
iron is of special importance. I have one question which is prompted by Philip
George, as to how certain one may be that this link is histidine ; how well does
histidine fit the electron densities near the iron atom ? A second question is the
interesting electron-dense material on the other side of the water molecule, which
leads one to wonder in the reactions of the ferrimyoglobin which is really what we
are talking about, the way which you would speculate that this material will inter-
fere with ligands for the oxygen atom ?

Kendrew: At the present resolution of our Fourier synthesis of myoglobin,
namely 2 A, it is not possible to be absolutely certain of the identity of the haem-
linked side chain in myoglobin, but it is very probably histidine. As an alternative
we have tried to build a model of lysine into the electron density, but for several
reasons this solution seems very unsatisfactory.

With regard to the group on the other side of the haem group, we are in more
difficulty. We think it probable that this residue also is histidine, but there is a

lO J. C. KENDREW

definite possibility that it is glutamic acid. It is probable that we shall have to wait
for a more highly resolved Fourier synthesis before this question can be answered
definitely.

There are indeed serious problems about the attachment of large ligand groups.
We have found that /)-iodophenyl isocyanide can be diffused into a crystal of myo-
globin and that it combines at the haem group, producing the characteristic change
in spectrum. When one looks at the model of the myoglobin molecule, and notes
how closely the side chains are packed together, it is hard to understand how a
ligand as large as this can approach the iron atom without a major disturbance in
the structure. It may indeed be that some disturbance does take place, because we
find that the /)-iodophenyl /^ocyanide derivative of myoglobin crystallizes with
slightly different cell dimensions from the normal crystal, and on some occasions
assumes a totally new crystal form, which suggests that the overall shape of the
molecule may have been slightly changed. In this connexion we are contemplating
the possibility of making comparative studies of the detailed structure of myoglobin
with diflFerent ligands attached to the haem group.

Theorell: Of course, this change in shape of haemoglobin molecule on oxygen-
ation is very interesting indeed, as it has been since the observation over 30 years
ago of the strictly hyperbolic oxygenation curve of the myoglobin. Do you think
the S shape could be a consequence of the change in shape of the molecule ? Could
it explain why the introduction of the first oxygen for instance is so diflficult and
the latter ones so much easier ?

Kendrew : I agree that it is quite possible to imagine that the first oxygen
becoming attached to haemoglobin in some way alters the relative position of the
sub-units, so that subsequent oxygens can enter more readily. This idea, however,
is purely speculative at present.

Theorell: It would be very interesting to know if under low oxygen tension
the whole change occurs at the introduction of the first oxygen molecule.

Kendrew: It would be very difficult to study this question experimentally
unless somebody could discover a method of preparing crystals of haemoglobin in
the partly oxygenated state.

Hirs: In applying the method of isomorphous replacement to myoglobin you
prepared several heavy atom derivatives with reagents such as mercury diammine,
aurichloride, etc., in which you were subsequently able to locate the position of the
heavy atoms in regard to the 6 A resolution structure. Can you now tell us with
which residues these derivatives are formed ? I believe the information would be of
interest to some of us.

Kendrew : The heavy atom groups which we used for working out the structure
of myoglobin undoubtedly cause local disturbances on the side chains in their
immediate neighbourhood. These disturbances mean that it is particularly difficult
to identify just those side chains which are of most interest in the present context.
In succeeding stages of the analysis we propose to use conventional refinement
methods which do not involve the introduction of heavy atoms at all ; if these
methods are successful the problem of local disturbances will not arise and we shall
then be in a position to identify the important side chains. Our impression is that in
most cases the heavy atom groups do not combine in a strictly chemical fashion

THE STRUCTURE OF GLOBULAR PROTEINS I I

with the myoglobin molecule, but that they lodge in interstices in the lattice. This
means that if the same molecule is crystallized in a new form, with a different packing
arrangement, it reacts quite differently with heavy-atom ligands. Thus sperm whale
myoglobin crystallizes in a monoclinic form from ammonium sulphate, and in an
orthorhombic form from phosphate. In the former, p-chloromercuribenzene
sulphonate enters the lattice at one place on the molecule and indeed proved to be
one of our most useful heavy atoms. In the latter, the behaviour is entirely different ;
the group becomes attached at four different sites on the molecule and is really of
little use for analytical purposes.

Molecular Configuration of Nucleic Acids

M. H. F. WiLKINS

]\IRC Biophysics Research Unit,

Physics Department, King's College,

London, England

Need for certainty in the structure determination of DNA

Molecular theory of replication of genetic material and of mutation is
based on the structure of DNA. Since the ideas of Watson and Crick
concerning DNA are so aesthetically attractive and are now being extended
in many ways to create almost a whole subject of nucleic acid biology (e.g.
the structures of RNA's with various functions in protein synthesis are
being derived by analogy with DNA), it is important that these ideas do
not become a dogma and that alternatives are not ignored. It is also desirable
that a stage be reached where the structure of DNA can no longer be
regarded as hypothetical. It is essential therefore that the structure be
placed on a sound basis of experimental fact.

It is generallv agreed that DNA consists of two polynucleotide chains
linked together bv hvdrogen bonds between adenine and thymine and
between guanine and cytosine. It is, however, still a somewhat open
question whether the hydrogen bonding scheme in the base pairs is that
proposed by Watson and Crick or has some other form. Valuable evidence
supporting the Watson Crick scheme is supplied by the studies of enzymic
synthesis of DNA (e.g. Josse and Romberg [i]) and evidence in favour is
also given by studies of complexes of synthetic polyribonucleotides [2]. I
wish to discuss here, however. X-ray diffraction data on DNA itself and
the extent to which these data provide an exact structure for DNA and
give a unique solution.

Difficulties in the X-ray structure analysis of DNA

There are two main difficulties in studying DNA by means of X-ray
difl'raction. First, DNA, like other chain polymer molecules, does not
form single crystals. The advantage of single crystals is that they enable
diffraction to be separated in all directions in three dimensions. Second,
the resolving power of the data has until very recently been insufficient to
show individual atoms in the structure. The first difficulty has to a large
extent been overcome : the DNA molecule is highly regular and fibres

14 M. H. F. WILKINS

may be prepared which consist of aggregates of parallel microcrystals.
With care in producing the DNA (by Dr. L. D. Hamilton of the Sloan-
Kettering Institute, New York) and by taking pains with the diffraction

Fig. I. X-ray diffraction photograph of microcrystalline fibres of the Hthium
salt of DNA.

technique we have gradually improved our fibre diffraction photographs,
and such is their sharpness now that overlapping of reflections is to a large
extent avoided (Fig. i.) As a result we are able to separate a large proportion
of several hundred reflections in three dimensions and obtain a reasonably
accurate set of intensities. We have also reduced the difficulty of limited
resolving power of the data. We have recently increased the resolving

MOLECULAR CONFIGURATION OF NUCLEIC ACIDS 1 5

power by recording diffraction at angles corresponding to spacings as small
as I • I A. As mentioned later, we have not yet analyzed these new data
but expect that they will improve considerably the results described here.
In our earlier work we had in the main analyzed the diffraction data in
two dimensions and by trial had adjusted a molecular model until it was in
agreement with the diffraction data in two dimensions [3, 4]. We have
now used the three-dimensional data to check the accuracy of the model
and to find the extent to which it is a unique solution. The Fourier
synthesis method is convenient for this purpose. The syntheses which I
will describe have been obtained by Dr. D. A. Marvin (aided by a computer
programme written by Dr. O. S. Mills).

Principles of the Fourier synthesis method of structure analysis

The Fourier synthesis method [5] may in simple terms be described
as follows. The diffracted X-rays have three characteristics :

1. Direction of diffraction.

2. Amplitude (given by the measured intensities).

3. Phase.

The direction of diffraction corresponds to the spacing of the electron
density in the structure. A Fourier synthesis is produced if the various
spacings, with correct amplitudes, are added together or subtracted. The
result is that one obtains directly a picture of the structure. The phases of
the diffracted beams, roughlv speaking, tell one which amplitudes are to
be added and which subtracted. Without knowledge of the phases the
structure cannot be derived. The main difficulty with the X-ray diffraction
method is that the phases are not given directly by the X-ray photograph.
In favourable cases, phases can be derived by measuring intensities with
and without a heaw atom placed in the molecule. This method has been
used in the remarkably successful structure analysis of myoglobin [6].

Another approach is as follows. If one has already a roughly correct
idea of the structure or of an appreciable part of it, one can obtain a com-
plete and exact structure, provided that the X-ray intensity data are com-
plete and exact. The procedure is to calculate the phases from the rough
structure, in this case our molecular model, and perform a Fourier synthesis
using the experimentally determined amplitudes and the calculated phases.
The picture given by this synthesis has the following characteristics :

1. If the model is correct the picture is like the model (except that the
limited resolving power mav not enable all the detail of the model
to be seen.)

2. If the model requires modification the picture is intermediate
between the model and the correct structure. As a result one can
see how to adjust the model to make it more nearly correct.

i6

M. H. F. WILKINS

Parts of the structure not included in the model appear roughly in
the picture, i.e. approximately half-height on the electron density
contour map. For example the positions of water molecules in
crystals of vitamin 6^2 have been determined ([7] and private
communication) when phases were calculated from the vitamin
molecule alone.

Fig. 2. (Right). Fourier synthesis of section through DNA molecule. The
helix axis is vertical, approximately in the plane of the section, and on the right-
hand side of the diagram. (Left). Section through phosphate groups of two neigh-
bouring DNA molecules. A marks a peak that may correspond to a chloride ion.
Two smaller peaks lie to the right of A and probably correspond to water mole-
cules. The A' marks show positions of atoms in the molecular model.

Fourier syntheses for DNA

Fourier syntheses of sections through the structure of the lithium
salt of DNA are shown in Fig. 2. The model used for calculating the

MOLECULAR CONFIGURATION OF NUCLEIC ACIDS 1 7

phases was the final model (Model 3) described in Langridge et al. [3, 4].
The section on the right in Fig. 2 is along the helix axis of a DNA
molecule and shows the bases stacked on one another and confined to the
central part of the molecule. Parts of the deoxyribose sugar ring are shown
where the section passes through them. The section on the left is parallel
to the helix axis but removed from it and passes through various phosphate
groups. Some of the phosphate groups belong to one molecule and others

Fig. 3. Typical Fourier synthesis in plane of a base-pair. The positions of
atoms in the model are shown. + marks the position of the helix axis. The base-
pairs and deoxyribose rings show clearly in the synthesis.

to an adjacent molecule in the crystal. The resolution is insufficient to
show separately the oxygen atoms of the phosphate groups. The main
interest of this section is that in the region occupied by water and between
the DNA molecules, several peaks appear on the contour map. Possibly
several of these peaks correspond to water molecules and they do in fact
occur in stereochemically likely positions. One peak at A is higher than
the others and might be due to a chloride ion, for it is necessary that
chloride be present in the DNA for the crystalline structure to form.

M. H. F. WILKINS

Figure 3 shows a section at right angles to the helix axis and in the plane
of a base-pair. The position of the atoms in the model are marked. Since
the sequence of bases along a polynucleotide chain is not periodic, the

Fig. 4. Fourier difference synthesis through same section as Fig. 3. The
space between DNA molecules has been treated approximately as uniformly filled
with water. There is almost no indication that the separation of the glycosidic
links needs to be altered. If the Watson-Crick pairs were incorrect and the
Hoogsteen scheme correct, the region near AA should be positive and that at BB
negative. In fact the signs are the reverse.

X-ray diffraction method shows an average base. It may be seen that
regions of high electron density correspond to the positions of atoms in the
base and deoxyribose parts of the molecule. The distance between the
glycosidic links appears somewhat larger on the Fourier section than on
the model. However, the synthesis was performed without taking account
of the presence of water in the structure. When the DNA molecule is

MOLECULAR CONFIGURATION OF NUCLEIC ACIDS 1 9

treated as being immersed in water of uniform electron density [3, 4] this
discrepancy largely disappears (Fig. 4).

Preliminary study of sections through all parts of the molecule indicates
that the model requires little or no adjustment. This result is not surprising
if we assume the water molecule peaks are not spurious, for unless the

Fig. 5. Scale drawing of Watson-Crick base-pairs [10]. The glycosidic links
all make an angle of 38 with the dyacl axis and are equally distant from it.

positions of the DXA atoms were fairly accurate the synthesis would not
have shown these molecules. It is clear, however, that the limited resolving
power of the data so far used does not enable the positions of atoms in the
base-pair to be distinguished. Therefore the question arises as to whether
the atoms in the base-pair might be arranged in some other way which
would also correspond with the observed data.

20

M. H. F. WILKINS

The Watson-Crick base-pairing schemes and possible alternatives

The Watson-Crick scheme [8, 9, 10] is shown in Fig. 5, It should be
noted that the glycosidic hnks in both base-pairs are arranged symmet-
rically about a line (dyad axis) in the plane of the diagram and passing

Fig. 6. The Donohue base-pairs [10].

through the helix axis. As a result the molecule has the same appearance
(apart from the sequence of bases) when turned upside down, i.e. when
rotated 180° about the dyad axis. In both pairs the separation of the

MOLECULAR CONFIGURATION OF NUCLEIC ACIDS 21

glycosidic links is the same. Attractive features of the base-pairing are
that the hydrogen bonds are reasonable lengths [ii] and almost linear,
and that the positions and directions of the glycosidic links in all four

9 2A

Fig. 7. The Hoogsteen base-pairing scheme. The glycosidic bonds have been
placed as far apart as seems reasonable. (I am indebted to Dr. M. Spencer for this
diagram.)

nucleotides can be made exactly the same [lo]. This equivalence of
glycosidic links enables the arrangement of the deoxyribose and phosphate
atoms to be made identical in every nucleotide in the helical molecule.

As alternatives to the Watson-Crick scheme, two schemes are worth
considering. The first is the Donohue pairing [12, 10, 13] shown in Fig. 6.

22 M. H. F. WILKINS

Although the two pairs give almost the same separation of the glycosidic
bonds, the direction of the bonds cannot be made exactly the same. Some
of the hydrogen bonds deviate appreciably from linearity. However, the
stereochemical shortcomings are not so great that, considered from this
viewpoint alone, the scheme would appear unlikely to apply to DNA.
Other somewhat unfavourable features of the Donohue scheme have been
discussed [lo] e.g. considerations of the directions of atomic sequence in
the two phosphate-ester chains in the molecule. The main difTerence
between the Donohue and Watson Crick schemes lies in the symmetrical
relationship of the glycosidic links. In the Donohue scheme these bonds
are not symmetrical with respect to a dyad axis perpendicular to the helix
axis : as a result a double helix molecule with a structure of this kind will
appear different when turned upside down, i.e. rotated i8o about a line
at right angles to the helix axis. In other words a Donohue-type molecule
may be specified with respect to a direction along the helix axis while
Watson Crick molecules are symmetrical and are the same in both
directions.

The second scheme (Fig. 7) involves a hydrogen bond arrangement
found between adenine and thymine [14, 15] and derives from a suggestion
made by Professor Linus Pauling. We will refer to the scheme as the
Hoogsteen scheme. The base-pairing is symmetrical as in the Watson-
Crick scheme. The scheme requires cytosine to have a not very probable
tautomeric form in which a hydrogen atom is moved from an amino group
to a ring nitrogen atom.

Possible ways of constructing DNA models incorporating base-
pairs alternative to those of Watson and Crick

In considering molecular models of DNA involving base-pairing
schemes other than that of Watson and Crick, it is necessary that the
position of the base, sugar, and phosphate parts of the molecule be placed
in approximately the same position as in the Watson Crick type model
we have described [3, 4]. I think we can safely assume that no other
arrangement would be compatible with the X-ray data. The idea that our
model is at least in this sense unique is confirmed by comparison of the
B and C configurations of DNA. We have shown that a small change in the
configuration (mainly a displacement of the nucleotide position in the
helical structure) causes the B diffraction pattern to change into the C
pattern [16, 17].

Let us attempt to build, with Donohue pairs, a DNA model as required
with the phosphate group positions related by the dyad axis referred to
above. The glycosidic links will not be related by the dyad axis. Hence
the base pair and the sugar groups will both be placed asymmetrically in

MOLECULAR CONFIGURATION OF NUCLEIC ACIDS

23

DONOHUE

Fig. 8. (a) Donohue base-pairs arranged asymmetrically with respect to the
line joining the points of attachment of glycosidic links to the deoxyribose rings in
DNA. (b) Symmetrical arrangement of "average base-pair".

24 M. H. F. WILKINS

relation to the dyad. In each nucleotide pair the base-pair will be arranged
lop-sided. There will be two arrangements: with base-pair extending as
in Fig. 8(rt) from bottom left to top right, or from top left to bottom right.
If every base-pair is arranged in the same way, a regular helical model
will be built. If this model is turned upside down the lopsided arrange-
ment of the base-pairs will appear to have changed — the helical system
of base-pairs will have rotated relative to the phosphate groups. If the model
is viewed as a simple helix, the rotation is equivalent to a displacement in
the helix axis direction. In a fibre of DNA there will on the average be an
equal number of molecules up one way or the other. If one molecule
passes through the repeat unit in the structure, as in the A structure [i8],
each repeat unit will at random contain base-pairs in one position or the
other. This random arrangement will give rise to continuous diffraction
along the layer lines on the diffraction pattern. The A type diffraction
patterns of DNA give no trace of this continuous diffraction (Fig. 9). The
possibility that the molecule contains an appreciable degree of asymmetry,
as is required by the Donohue scheme, is therefore excluded.

There may be a possible way out of this difficulty. If the model
were not built in a regular helical fashion but if successive base-pairs in
a molecule were placed at random lopsided one way or the other, the
base-pairs would on the average be symmetrical. The shape of the resultant
average base-pair, consisting of two lopsided base-pairs lying criss-cross
(Fig. 8(6)), would be extended considerably in the plane of the base-pairs.
Such a model is unattractive because of its considerable irregularity
(probably it would not be possible to form the phosphate-ester chains)
and because the base-pairs would be stacked on one another only to a
small extent and as a result their hydrophobic surfaces would be largely
exposed.

The Hoogsteen base-pairs have the desirable feature, like those of
Watson and Crick, of symmetrically-placed glycosidic links. The distance
between the glycosidic links is smaller than in the Watson-Crick scheme
by about 2 A. (This difference could be larger — we have used a minimum
value estimated by Dr. M. Spencer.) The other main difference in the
base-pairs is that the six-membered ring in purines is placed, in the
Hoogsteen scheme, to one side of the pair of hydrogen bonds (at A' in Fig.
10), whereas in the Watson-Crick scheme this ring is at V in line with
the bonds.

A molecular model resembling DNA has been built consisting of
adenine-thymine pairs of Hoogsteen type [19]. The diffraction pattern of
the model resembles that of DNA but the agreement between calculated
and observed intensities is not so good as that obtained with Watson-
Crick pairing. However, the best way of comparing the agreement between
calculated and observed intensities for different models is to use the Fourier

MOLECULAR CONFIGURATION OF NUCLEIC ACIDS

25

/^

^^1^

Fig. 9. X-rav diffraction pattern of sodium salt of DNA in A configuration.
Clearly defined spots are seen and there is no sign of diffraction streaks along the
laver lines.

26

M. H. F. WILKINS

synthesis method as described below. Stereochemical difficulty has been
found in placing guanine-cytosine pairs in the model [19] but these difficul-
ties may not be insurmountable.

WATSON- CRICK S-S =

HOOGSTEEN S-S=9'2 A MAXIMUM

Fig. 10. Comparison of the Watson-Crick and Hoogsteen base-pairing
arrangements. The hexagonal purine ring is at A' and 1' in the two schemes.

Use of the Fourier synthesis method to assess the probability that
base-pairs in DNA are not of the Watson Crick type

We have built complete molecular models of DNA with Watson-Crick
base-pairs [3, 4]. The Fourier synthesis method enables us to assess the
probability that the base pairs are of a different type without constructing

MOLECULAR CONFIGURATION OF NUCLEIC ACIDS 27

molecular models with alternative base-pairs or calculating diffraction
from the models. If the base-pairs in the DXA model require modification,
the nature of the modification should appear when the F'ourier synthesis
is examined.

Consider first the Donohue scheme. There is no indication in the
syntheses in Figs. 3 and 4 that the average base-pair should be made
larger, in the plane of the base-pair, than the average Watson Crick pair.
Therefore it is unlikely that the Donohue scheme exists in DNA. We may
note, however, that a double-helix structure with glycosidic links arranged
as in the Donohue scheme has been established for the two-chain complex
of polyriboadenylic acid ([2] and private communication). The structure
is clearly distinguishable from that of DNA, and the sequences of atoms
in the two phosphate-ester chains run in the same direction and not in the
opposite direction as experimental evidence (other than that obtained from
X-ray diffraction) indicates is the case for DNA [i].

Examination of the Fourier syntheses also shows that the Hoogsteen
scheme almost certainly cannot exist in DNA. First, it seems most un-
likely that, as required in the Hoogsteen scheme, the distance between
glycosidic links could be 2 A less than in our model : if any alteration is
required it is that the distance should be increased slightly. Second, there
is no indication that the position of the purines in the base-pairs should be
altered to that in the Hoogsteen pairs. This is shown most clearly in the
Fourier difference synthesis (Fig. 4), Such a synthesis corresponds roughly
to the difference between the real structure and the model used in the
synthesis. If the real structure contained Hoogsteen pairs, the difference
synthesis would have positive sign at AA Fig. 4, w^here the six-membered
purine ring occurs in the Hoogsteen scheme, and negative at BB, where the
ring occurs in the average Watson-Crick pair. The observed signs are
opposite to those expected if the Hoogsteen scheme were correct. Hence,
once more, the existence of the Hoogsteen scheme appears unlikely. We
hope to confirm this conclusion by calculating a Fourier difference
synthesis for the Langridge-Rich model containing Hoogsteen adenine-
thymine pairs. If our approach is correct the synthesis will indicate that
the Hoogsteen base-pair in the model should be replaced by the Watson-
Crick pair.

It might be thought that because the X-ray data cannot resolve
spacings less than 3 A and as a result show individual atoms, the data
could not be used to distinguish between two tvpes of base-pairing in
which the positions of the atoms differed in the main bv less than 3 A.
This, however, is not so : the data can be used to distinguish between a
structure that is nearly correct and one in which the atoms are displaced
by I A or even less. In the case of well-defined groups, such as the phos-
phate group, the position may be determined to within 0-5 A. We are,

28

M. H. F. WILKINS

however, not satisfied with the present state of the X-ray analysis. The
difference synthesis should have almost zero amplitude whereas in fact
appreciable amplitudes are present. These are presumably due to errors
in X-ray intensity measurements and errors in estimating the position of

.20^ loLtfcr hr>e.

Fig. II. X-ray dillraction phcjtoKraph (with Mr. X. E. Chard) of lithium salt
of DXA showing layer lines to the 20th. The diffraction corresponds to spacings
extending down to i -7 A.

water molecules in the structure. Use of X-ray data of higher resolution
would make the errors less important and give further confirmation of the
correctness of the structure. Fortunately we have recently been able to
obtain data of higher resolution, all layer lines to the 20th (1-7 A) being
observed (Fig. 11) and also an isolated reflection on the 30th layer line

MOLECULAR CONFIGURATION OF NUCLEIC ACIDS

29

(i • I A) (Fig. 12). We intend to use these data in future Fourier analysis.
Another means of confirming the correctness of our analysis is to substitute
bromide for chloride in the structure and show that the height of the sup-
posed halide peak increases in the expected manner. We are at present
obtaining intensity data from bromide-containing DXA.

30^ /ft/cr Zinc

Fig. 12. X-ray diffraction photograph (with Mr. X. E. Chard) of lithium salt
of DXA showing i ■ i A reflection on the 30th layer line.

Notes on the molecular configuration of RNA

Improved methods of preparing RXA and the separation of various
types of RXA have helped to clarify X-ray diffraction studies of RXA.
Soluble RXA (transfer RXA with amino acids attached) can be oriented
in sheets and gives diffraction patterns which, though rather diffuse, are
clearly similar to those of DXA [20]. It is fairly certain that these RXA

3© M. H. F. WILKINS

molecules have a DNA type structure because other types of double-
helix structure formed by synthetic polyribonucleotides [2] give recogniz-
able and distinctly different patterns. Moreover the molar proportion of
adenine and uracil and of guanine and cytosine are almost equal in this
RNA. Mr. W. Fuller has modified the DNA structure and built a mole-
cular model of RNA. The hydroxyl group of the ribose is accommodated
in the structure by means of various slight distortions. The molecule is
stabilized by a hydrogen bond between the hydroxyl oxygen atom and an
atom of the adjacent phosphate group.

This type of structure appears to hold also for ribosomal RNA from
Escherichia coli [21]. It has not been found possible to orient the double-
helices in this RNA. It appears likely therefore that double helical regions
of the molecule are linked together by non-helical parts to form a cluster
of short helical regions. This is probably the structure within the ribosomes
because the diffraction pattern of ribosomes is the same as that of the RNA
mixed with protein.

It is of interest that earlier preparations of RNA, e.g. from liver, yeast,
tobacco mosaic virus, etc., always gave a double helical pattern which,
though having a general resemblance to that of DNA, showed distinct
differences [22]. It has been noted [21] that this RNA pattern resembles
that of a mixture of DNA and double-helix polyriboadenylic acid [2].
Presumably some of these RNA specimens consisted largely of ribosome
material. It may be that the earlier methods of preparation caused the
RNA molecule to be extended into a filamentous form capable of being
oriented and that parts of the molecule had approximately DNA-like base-
sequences and that other parts contained more predominantly adenine
and formed anti-parallel double-helix structures. This problem is being
investigated further. In all cases, however, the length of regular DNA-
like regions in RNA is restricted and the diffraction patterns are not well
defined. The relation of structure to function in RNA is not yet clear, it
has been found, for instance, that the original intact double-helix structure
is not required in soluble RNA for amino acid-binding activity to be
present [20].

Acknowledgments

In this brief space I will not attempt to acknowledge all those who have
contributed to this work but I wish to mention specially Drs. D. A. Marvin
and M. Spencer, Mr. W. Fuller, and the University of London Computer
Unit.

References

1. Josse, J. and Kornberg, A., Fed. Proc. 19, 305 (i960).

2. Rich A., in "A Symposium on Molecular Biology", ed. R. E. Zirkle, University
of Chicago Press (1959).

MOLECULAR CONFIGURATION OF NUCLEIC ACIDS 3 1

3. Langridge, R., Marvin, D. A., Seeds, W. E., Wilson, H. R., Hooper, C. W.,
Wilkins, M. H. F., and Hamilton, L. D.,jf. mol. Biol. 2, 38 (i960).

4. Langridge, R., Wilson, H. R., Hooper, C. W., Wilkins, M. H. F., and Hamilton,
L. D., y. mol. Biol. 2, 19 (i960).

5. Lipson, H., and Cochran, W., "The Determination of Crystal Structures",
G. Bell, London (1957).

6. Kendrew, J. C, Dickerson, R. E., Strandberg, B. E., Hart, R. G., Davies,
D. R., Phillips, D. C, and Shore, V. C, Nature, Lond. 185, 422 (i960).

7. Hodgkin, D. C, Pickworth, J., Robertson, J. H., Prosen, R. J., Sparks, R. A.,
and Trueblood, K. N., Proc. ray. Sac. A 251, 306 (1959).

8. Crick, F. H. C, and Watson, J^ D., Proc. roy. Soc. A 223, 80 (1954)-

9. Pauling, L., and Corey, R. B. Arch. Biochem. Biophys. 65, 164 (1956).

10. Spencer, M., Acta Cryst. 12, 60 (1959).

11. Fuller, W.,y. phys. Cheni. 63, 1705 (1959).

12. Donohue, J., Proc. nat. Acad. Sci., Wash. 42, 60 (1956).

13. Donohue, J., and Trueblood, K. N.,y. mol. Biol. 2, 363 (i960).

14. Hoogsteen, K., Acta Cryst. 12, 822 (1959).

15. Pauling, L., "The Nature of the Chemical Bond", 3rd Edn. Cornell Univ.
Press, Ithaca, N.Y., 504 (i960).

16. Marvin, D. A., Spencer, M., Wilkins, M. H. F., and Hamilton, L. D., Nature,
Lond. 182, 387 (1958).

17. Marvin, D. A., Spencer, M., Wilkins, M. H. F., Hamilton, L. D., y. mol. Biol.

in press (1961).

18. Langridge, R., Seeds, W. E., Wilson, H. R., Hooper, C. W., Wilkins, M. H. F.,
and Hamilton, L. D., y. biophys. biochem. Cytol. 3, 767 (1957).

19. Langridge, R., and Rich, A. Int. Ihiion of Crystallography Fifth International
Congress, abstracts 78 (i960).

20. Brown, G. L., and Zubay, G.,_7. mul. Biol. 2, 287 (i960).

21. Zubay, G., and Wilkins, M. H. ¥.,y. mol. Biol. 2, 105 (i960).

22. Rich, A., and Watson, J. D., Proc. nat. Acad. Sci., Wash. 40, 759 (i954)-

Discussion

Charg.^ff: I want to ask Dr. Wilkins a question concerning the nucleoproteins.
There is some evidence I believe that the nucleic acids in the cell are mostly present
as complicated nucleoproteins, nucleoprotamines and nucleohistones. I believe
that Dr. Wilkins has published some evidence about nucleoprotamines which
seems to indicate that the structure of DNA is maintained in these protamine
complexes. However, there are much more complicated nucleoproteins found, for
instance, in bacteria. We have described a deoxyribonucleoprotein from tubercle
bacilli which seems to contain one amino acid equivalent per two nucleotides. It is
a very peculiar and stable nucleoprotein which cannot be dissociated by strong salt
concentrations ; and I am wondering whether anything is known of the X-ray
structure of these more highly organized nucleoproteins or, for that matter, whether
anything is known about the X-ray structure of the single strand DNA as isolated
from a small virus.

Wilkins : I did not attempt in this talk to get on to the subject of nucleoproteins
which is a big and very interesting subject. There is, as Dr. Chargaff points out,
good evidence that the nucleic acids do on the whole retain their double helical

32 M. H. F. WILKINS

configuration in nucleoproteins. The diffraction diagram of nucleoprotamine is
very similar to that of DNA and it is clear that the DNA retains its configuration
when combined with protamine. Nucleohistone in the chromosomes of somatic
cells of higher organisms again gives an X-ray diffraction diagram which is charac-
teristic of the double helix of DNA; it appears that the protein is rather loosely
bound and fills the space between the DNA molecules. In the case of bacteria we
have so far only looked at one nucleoprotein which Dr. G. Zubay in our laboratory
prepared from E. coll. In that case the X-ray difTraction diagram, which we pub-
lished, shows that much of the DNA is in the double-helix form. It also shows that
the DNA is largely free from protein. We cannot be certain that some of this free
DNA might not be an artifact of the extraction procedure. Dr. Kellenberger has
evidence from electron microscope studies on E. coli that the chromosomes in these
bacteria may not consist of nucleoprotein, but might consist of DNA itself. One can
see the individual threads of the DNA molecules passing through the nucleus. In
the case of RNA the same also appears roughly to be true and the diffraction
pattern from ribosomes of E. coli is essentially the same as that from a mixture
of protein and double-helix RNA. These results have been confirmed by Dr. Klug
and his colleagues at Birkbeck College. In many nucleoproteins, and in vivo, both
the RNA and DNA have a double helical configuration and the protein is some-
how built around them and does not alter that configuration very much. There
certainly are exceptions to this rule and there is no double helix, for instance, in
tobacco mosaic virus ; there the RNA is a single chain molecule and does not have
a configuration at all resembling that of the DNA double helix ; it may be that other
exceptions will be found and that the interesting nucleoprotein from tubercle bacilli,
referred to by Dr. Chargaff has a special structure. We do know, however, that the
tubercle DNA after separation from the protein has the usual double-helix structure.

Partition of Macromolecules in
Aqueous Two-Phase Systems

Per-Ake Albertsson
Institute of Biochemistry, Uppsala, Sweden

Separation of substances by partition between two immiscible solvents
is one of the most frequently used methods in organic chemistry; in
inorganic chemistry, partition has also been widely used both for prepara-
tive and anahtical purposes. I\Iany biochemical substances, such as
polypeptides and even proteins, have also been studied by partition
methods, for example countercurrent distribution [i, 2].

One of the advantages of partition between two liquid phases is that
there is generally a greater possibility of obtaining a state of equilibrium
between two liquid phases than between a solid and a liquid phase. This
is particularlv so when macromolecules are involved. It would be of
great value if partition methods could also be applied to very high molecular
weight macromolecules of biological origin such as proteins, highly
polymerized nucleic acids, viruses, and cell particles. In recent years we
have studied this possibility in Uppsala and the present paper will describe
some experiments and a summary of the results obtained. A monograph [3]
dealing with the theoretical background and experimental details has
recently been published.

Polymer two-phase systems

In order to partition successfully biochemical macromolecules one
cannot usually use conventional two-phase systems containing an organic
solvent since this may cause denaturation. Instead a number of polymer
two-phase systems, obtained by mixing aqueous solution of two ditferent
polymers, have been used. It is a general phenomenon [4] that when
polymer solutions are mixed they give rise to two liquid phases, one phase
containing one polymer and the other phase the other polymer; both
phases have a high water content (85 99"o). Two aqueous phases in
equilibrium are thus obtained. The ditference in composition between the
phases is comparatively small; the interfacial tension, the differences
between the refractive indices and the densities of the two phases are
therefore much smaller for polymer phase systems than for conventional,

34 per-Ake albertsson

low molecular weight phase systems. Due to the presence of the polymers
in the phases, their viscosities are relatively high. The time of phase
separation is comparatively long for polymer two-phase systems; varying
from about 5 minutes to i day. Both non-ionic polymers and poly-
electrolytes may be used for the construction of a two-phase system [3].
By varying the molecular weight, the number of non-polar groups, or
the number of charged groups on the polymers, a large number of phase
systems having highly diversified properties may be obtained. The phase
systems may be complemented by adding low molecular weight sub-
stances, such as sucrose or electrolytes, to achieve a suitable environment
for biological material.

Partition of proteins, nucleic acids, and viruses

The partitioning of a particle in a two-phase system depends mainly
on its surface properties including the area and the chemical properties
of the outer layer of the particle.

The partition coefficients of some proteins and viruses have been
measured in a two-phase system of dextran and methylcellulose [3, 5] and

TABLE I

The Partition Coefficient, K, and the Surface Area of a Number of Virus

Particles and Protein Molecules in a Two-Phase System of Dextran and

Methylcellulose at 4'.*

Data from refs. [3] and [5].

Particle

K

Surface area (m/x)^

Phycoerythrin

0-9S

0-3

Haemocyanin "eighth"

0-62

0-86

Haemocyanin whole

0-25

3-5

Echo virus

0 • 2-0 • 3

1-3

Polio virus

02

2-3

Phage T3

2-3 X IO~2

8-7

Tobacco mosaic virus

(1-2) X 10-2

14-4

Phage T2

(6-10) X IO-*

25-5

Phage T4

(3-S)xio-''

25-5

Vaccinia

(4-12) X IO~^

220

the results are given in Table I together with the surface areas of the
partitioned particles, calculated from their form and dimensions as
determined with the electron microscope. As may be seen in Table I, all
particles favour the bottom phase and the partition coefficient becomes
smaller the larger the particle size. In Fig. i, the log K values are plotted
against the particle surface area; it can be seen that the points lie around a

* {K = concentration in top phase/concentration in bottom phase.)

PARTITION OF MACROMOLECULES IN AQUEOUS TWO-PHASE SYSTEMS

35

straight line indicating that in this phase system it is mainly the surface
area which determines the K value. An upper limit of the K value is
obtained which is probably due to incomplete separation of the phases.
No K value less than lO"** has been observed even for very large particles.

^O 50 220m)u^
surface area

Fig. I. Relation between the partition coefficient, K, and the surface area of
virus particles and protein molecules. The data of Table I are plotted in this figure.

The fact that the virus concentration is looo to lo coo times higher in
one phase than in the other, may be utilized for the concentration and
purification of viruses [3, 6, 7, 8, 9, 10].

tablp: II

Distribution of Some Nucleic Acid Preparations in a Na Dextran Sulphate-
Methylcellulose System

C, = concentration in top phase, C,, = concentration in bottom phase, V,,=
volume of bottom phase, a — amount of nucleic acid collected at the interface
(Lif et al. to be published; for experimental details see ref. [3]).

Nucleic acid

5,0,,,

A'.=

RNA from yeast
DNA from calf thymus (I)
DNA from calf thymus (II)
DNA from calf thymus (III;
DNA from phage Tz (I)
DNA from phage T2 (II)
DNA from phage T2 (III)

3-5
20-4

21-8
22-5
3I-I
31-5
33-0

0-67

o- 17
014
o- 12
0012
o-o8

0-02

Partition of some nucleic acid preparations has recently been studied
[11] in a dextran sulphate-methylcellulose system. Preliminary results
are shown in Table II and Fig. 2. In this phase system a major part of the

36

per-Ake albertsson

nucleic acids collected at the interface. This part was included in the
bottom phase and thus the K^ values given in Table II are not strictly
partition coefficients. As may be seen in Table II, the larger the sedimen-
tation constant the smaller the tendency for the nucleic acids to partition
in favour of the top phase.

Most proteins in the native state partition in favour of the lower phase
of the dextran methylcellulose system and the partition coefficient is

T2 DMA

Thi^mus DNA o

o S-RNA

J I 1__J I I 1 1 I L

lO

20

30

■^

20, ^

Fig. 2. Relation between the "partition coefficient", K^, and the sedimenta-
tion constant, S-.q, for a number of nucleic acid preparations. The data of Table II
are plotted in this figure.

little influenced by the electrolyte composition. Thus, if the partition is
carried out in the presence of a o-oi m buff^er and an excess (o-i m) of
NaCl, the partition coefficient is constant over a wide pH range ; see Fig. 3,
where an experiment with human serum albumin is recorded [3]. Below
pH 4-5, however, the K value increases and even becomes greater than
unity, indicating that the protein has more affinity for the top phase at
acid pH values. It is interesting to compare this result with the viscosity
and optical rotation of serum albumin as a function of pH. Thus there is

T I I I — I — I — I — I — I — r

serum albumin

PARTITION OF MACROMOLECULES IN AQUEOUS TWO-PHASE SYSTEMS

2.0

f.a

1.6

/.^

1.2

o.e

0.6

0.2

A I I \ L

37

"S (V o

tT^2>0-

2 3 ^ 5 6 7 e 9 10 n 12
pH
Fig. 3. Partition coefficient, A', of serum albumin as a function of pH in a
dextran-methylcellulose system. From ref. [3].
100

ao
7o
60
so
<o

20

15

NaCI

Fig. 4. Partition coefficient, K, of a number of proteins in a dextran-poly-
ethylene glycol system with 0005 M KH0PO4 and increasing concentrations of
NaCl ; the latter is expressed as moles NaCl added to i kg of phase system. For
experimental details see ref. [3].

38 PER-AKE ALBERTSSON

an increase in the limiting viscosity number [12, 13] and in the laevorota-
tion [14] at low pH values indicating a change in the structure of the
protein. It has also been reported [ 1 5] that there is an increase in the number
of hydrophobic groups on the protein surface at low pH values which
could explain why the protein at these pH values favours the top phase
containing more methyl groups than the bottom phase. If this explanation
is correct it means that partition provides a method for the determination
of the " hydrophobicity " of proteins and particles.

It was mentioned above that the partition in the dextran methylcel-
lulose system is only influenced to a small degree by the electrolyte com-
position. In contrast, the dextran-polyethylene glycol system shows many

Theoretical curve
Enzyme activity

tiAn

20

40 60 80
Tube number

Fig.
dextran

5. Countercurrent distribution of partly purified ceruloplasmin in a
-polyethylene glycol system. From refs. [3] and [18].

Striking salt eifects [3, 16]. Thus, a protein or particle may be transferred
almost entirely from one phase to the other by changing the salt content.
As may be seen in Fig. 4, where the partition coefficients of a number of
proteins in a dextran-polyethylene glycol system are recorded, phycoery-
thrin is in the top phase {K= ^) with o-oi m phosphate buffer pH = 6-8.
If NaCl is added up to o • i m, the protein is transferred to the bottom
phase {K=o-2)', further additions of NaCl increase the K value so that
the proteins again partitions in favour of the top phase. When the con-
centration of NaCl is increased from i to 5 m, all proteins studied transfer
to the top phase. Further information on the partition of proteins in the

PARTITION OF MACROMOLECULES IN AQUEOUS TWO-PHASE SYSTEMS 39

dextran-polyethylene glycol system may be found in refs. [3] and [16].
The dextran-polyethylene glycol system is useful in the countercurrent
distribution of proteins [17, 18]. The settling time of this system may be
as short as 5-10 minutes. A countercurrent distribution experiment [18]
with a partly purified preparation of ceruloplasmin is recorded in Fig. 5.
The enzyme activity curve is almost identical with the theoretical curve
indicating that the enzyme activity is associated with a single component.
Most experiments indicate that proteins partition according to the Nernst
partition law in the dextran-polyethylene glycol system, that is, the value
of the partition coefficient is independent of the protein concentration and
the presence of other proteins. This phase system may therefore be used
for characterization and fractionation of proteins.

References

1. Tavel, P., and Signer, R., "Advances in Protein Chemistry" II, 237 Academic
Press, New York (1956).

2. Craig, L. C, in "A Laboratory Manual of Analytical Methods of Protein
Chemistry", Vol. i, ed. P. Alexander and K. J. Block. Pergamon Press,
Oxford, 121 (i960).

3. Albertsson, P. -A., " Partition of Cell Particles and Macromolecules ". Almqvist
& Wiksell, Stockholm; John Wiley &; Sons, Inc., New York (i960).

4. Dobry, A., and Boyer-Kawenoki, F., ^. Polym. Sci. 2, 90 (1947).

5. Albertsson, P. -A., and Frick, G., Biocliim. biophys. Acta 37, 230 (i960).

6. Frick, G., and Albertsson, P. -A., Nature, Land. 183, 1070 (1959).

7. Wesslen, T., Albertsson, P. -A., and Philipson, L., Arch. Virusforsch. 9, 510

(1959)-

8. Philipson, L., Albertsson, P. -A., and Frick, G., J^irology ll, 553 (i960).

9. Norrby, E., and Albertsson, P. -A., Nature, Loud. 188, 1047 (i960).

10. Frick, G., Exp. Cell Res. (in press).

11. Lif, T., Frick, G., and Albertsson, P. -A., to be published.

12. Tanford, C, " S\Tnposium on Protein Structure", ed. A. Neuberger. Methuen
& Co. Ltd., London; John Wiley &: Sons, Inc., New York, 35 (1958).

13. Yang, J. T., and Foster, J. F.,jf. Amer. cheni. Soc. 76, 1588 (1954).

14. Sterman, M., and Foster, J. F., J. Amer. che?n. Soc. 78, 3652 (1956).

15. Foster, J. F., in Putnam, F. W., "The Plasma Proteins", Vol. i. Academic
Press, New York, i (i960).

16. Albertsson, P. -A., and Nyns, E. J., Ark. Kemi 17, 197 (1961).

17. Albertsson, P. -A., and Nyns, E. J., Nature, Loud. 184, 1465 (1959).

18. Broman, L., and Albertsson, P. -A., to be published.

Discussion

HiRs: Have there been any studies on the rate at which equilibrium has been
obtained with some of these systems ?

Albertsson: We haven't studied this systematically but it appears that just a
few inversions are enough to get an equilibrium.

Theorell: How is the separation of these materials from the polymers
achieved ?

40 PER-AKE ALBERTSSON

Albertsson: There is not one general method, of course; different methods are
used for different materials. One way, for example, is to use adsorption. In the case
of the proteins, we have adsorbed the protein on a column of calcium phosphate
and the polymers, being non-ionic, pass through and may be washed away. This is
a convenient method and a concentration of the proteins is obtained at the same
time. Another way is to transfer the substance from a polymer phase to a polymer-
free phase. If we are using the dextran-polyethylene glycol system we can, by
adding salt, transfer the protein to the top phase which contains polyethylene glycol
as the only polymer. This top phase can then be transferred to another tube and
mixed with ammonium sulphate (or potassium phosphate) when we get a new phase
system consisting of a polyethylene glycol phase in equilibrium with an ammonium
sulphate phase, this latter phase containing almost no polyethylene glycol. If one is
lucky, the protein is transferred to the ammonium sulphate phase under these
circumstances. This method works for some of the serum proteins, for example
ceruloplasmins. It is possible to adjust the volumes of the phases so that a concen-
tration of the protein is obtained at the same time.

The Reactivity of Certain Functional Groups in

Ribonuclease A towards Substitution by i-Fluoro-2,4-

dinitrobenzene. Inactivation of the Enzyme by

Substitution at the Lysine Residue in Position 41

C. H. W. HiRS, MiRjAM Halmann,* and Jadwiga H. Kycia

Department of Biology, Brookhaven National Laboratory,
Upton, L.L, NY., U.S.A.

Introduction

In this communication some of the initial results obtained in a study
of the dinitrophenylation of ribonuclease A will be presented. Publication
of the preliminary findings is warranted because some new information
has been brought to light concerning the manner in which ribonuclease A
interacts with its substrate.

While the reactivity of the functional groups in proteins has continued
to be the subject of comprehensive study, a more extensive development
of knowledge in this area of interest will have to await the elucidation of
the structures of the macromolecules involved. When it became possible
to put forward a primary structural formula for ribonuclease A [i, 2]
studies of the reactivity of certain functional groups in this protein became
of interest as a means with which to probe the secondary and tertiary
structure of the ribonuclease A molecule in solution. It was hoped that
information about the secondary and tertiary structure made available in
this way would also prove useful in delineating some of the structural
features associated with the catalytic activity of this protein.

It has become common practice to attribute variations in the reactivity
of the amino acid side-chains in proteins to their degree of accessibility
to reagents. Residues reacting more slowly than others of the same type
are considered to be "masked" or "buried". When the reactivity of such
residues returns to the anticipated level through exposure of the protein
to agents such as extremes of pH, urea in high concentrations, etc., capable
of causing the disruption of the secondary and tertiary structure, it is usual
to attribute the effect to "unmasking". While emphasizing the importance
of steric effects such views tend to draw attention away from the possibility

* Present address: Israel Institute for Biological Sciences, Nes Ziona, Israel.

42 C. H. W. HIRS, M. HALMANN, J. H. KYCIA

that in some cases large effects on reactivity may also be the consequence of
neighbouring group participation, or the influence of electrostatic inter-
actions in the vicinity of the reacting groups. A further factor of import-
ance may be the ability of the protein molecule to absorb the reagent
specifically and thereby to facilitate the reaction of a proximally located
functional group by overcoming the usual entropy factor. Finally, the
introduction of modified amino acid residues into the protein fabric may
of itself induce variations in the reactivity of the as yet unsubstituted
residues, either by direct steric interference or by alteration of the tertiary
structure. In general, therefore, unless it becomes possible to locate a
number of the functional groups with unusual reactivity by direct reference
to the primary structure, there is little chance that observations on re-
activity alone can be of value in revealing specific aspects of the secondary
and tertiary structure.

i-Fluoro-2,4-dinitrobenzene (FDNB) has a number of properties
desirable in a reagent that is to be used in studying the reactivity of func-
tional groups in a protein. It behaves in the manner of a typical reactive
alkyl halide* and is capable of reacting with unprotonated a-amino,
€-amino, and imidazole groups, as well as with thiol and phenolic hydroxyl
groups in the form of their conjugate bases. Introduction of the dini-
trophenyl group into the protein confers the characteristic spectral
properties of this chromophore to the new compound, a feature of value in
analysis, and excess reagent and dinitrophenol, formed from the reagent
by hydrolysis, are readily removed from the reaction mixture by extraction.

The groups potentially sensitive to fluorodinitrobenzene in ribonu-
clease A comprise the single a-amino group of the lysine residue at the
amino-terminal end of the peptide chain, the lo e-amino groups of the
lysine residues at positions i, 7, 31, 37, 41, 61, 66, 91, 98, and 104, the
imidazole groups of the four histidine residues at positions 16, 48, 105,
and 119, and the phenolic hydroxyl groups of the six tyrosine residues at
positions 25, 73, 76, 92, 97, and 115. Under the conditions examined thus
far, fluorodinitrobenzene does not readily form sulphonium derivatives
with the thioether function of the methionine residues, nor is reaction
with the hydroxyl functions of the serine and threonine residues in
ribonuclease A appreciable.

Analytical

In the present studies the extent of substitution on dinitrophenylation
of the enzyme has been followed by quantitative amino acid analysis with
the method of Moore et al. [3] in conjunction with an automatic analyzer

* Because of this similarity, we have elected to call substitution reactions
involving the introduction of the dinitrophenyl group alkylations, even though it
would be more accurate to use the term arvlation.

DINITROPHENYLATION OF RIBONUCLEASE A 43

of the type described by Spackman et al. [4]. It should be emphasized
that this approach is hmited in its appUcation, in the ideal case, to the
study of substitutions that lead to the formation of derivatives capable of
withstanding the customary acid hydrolysis necessary for the breakdown
of the protein to amino acids. Moreover, in using acid hydrolysis the
possibility must be recognized that the appearance of a stable derivative
on hydrolysis may in itself constitute an artifact reflecting the consequences
of an acid-catalyzed rearrangement in which migration of the dinitro-
phenyl group from the functional group of original substitution is in-
volved. When substitution takes place with the formation of derivatives
unstable to acid hydrolysis, two broad alternatives are possible: the
derivative may decompose by reversal of its formation and thereby re-
generate the amino acid, or decomposition may take place into other
products without regeneration of the amino acid. The first of these
alternatives makes it essential that the degree of substitution of the protein
be checked prior to hydrolysis by means of absorption measurements in
the ultra-violet; the second requires that the degree of substitution be
measured by difference analysis after hydrolysis.

Substitution on the amino-terminal lysine residue in ribonuclease A
may result in the formation of three products : a-dinitrophenyl-lysine,
e-dinitrophenyl-lysine, and a,e-bis-dinitrophenyl-lysine. e-Dinitrophenyl-
lysine is an a-amino acid and reacts with ninhydrin under the conditions
that obtain in the automatic amino acid analyzer to give a blue product
in approximately the same yield as typical a-amino acids. In our hands it
has proved to be stable (loss less than 5 "o after 22 hours at 110° in constant
boiling HCl) to the usual conditions of hydrolysis routinely employed as
a preliminary to quantitative amino acid analysis. a-Dinitrophenyl-lysine
contains a free e-amino group with the characteristics towards ninhydrin
of the amino group in y-aminobutyric acid. The relatively greater basicity
of the amino group in y-aminobutyric acid causes the compound to be
more strongly retained than a-aminobutyric acid upon chromatography
over columns of sulphonated polystyrene resins. Similarly, a-dinitro-
phenyl-lysine is more strongly retained by these resins than the e-isomer.
The presence of a dinitrophenyl radical in both derivatives confers on them
an enhanced affinity for the matrix of polystyrene based ion-exchangers.
Thus, when a mixture of the common protein amino acids and e-dinitro-
phenyl-lysine is chromatographed on the 15 cm. column of Amberlite
IR-120, normally used for the separation of the basic amino acids on the
automatic analyzer, the dinitrophenyl derivative emerges 35 effluent ml.
after arginine. a-Dinitrophenyl-lysine is even more strongly retained and
is not eluted satisfactorily from the column under the same conditions.
a,e-Bis-dinitrophenyl-lysine does not react with ninhydrin and is so firmly
bound by Amberlite IR-120 that only drastic conditions will serve to

44 C. H. W. HIRS, M. HALMANN, J. H. KYCIA

remove it from the resin. Both a-dinitrophenyl-lysine and a,e-bis-dinitro-
phenyl-lysine undergo decomposition during acid hydrolysis prior to
amino acid analysis to varying extents up to 30% in 22 hours, but lysine
is not regenerated in the process.

The histidine residues in ribonuclease A react with fluorodinitro-
benzene to form /w/V/rt;?o/f'-dinitrophenyl derivatives. The tautomerism of
the imidazole nucleus makes it possible for two dinitrophenyl derivatives
to be formed at each histidine residue. On hydrolysis in acid these deriva-
tives appear as the /w-dinitrophenylhistidines, unstable compounds that
break down into an as yet unidentified substance, with the properties of
an amino acid, that persists as the stable end-product of the acid hydrolysis.
Free histidine is not regenerated. The unidentified substance is eluted
more rapidly than histidine on the 15 cm. column used for the analysis
of the basic amino acids with the automatic analyzer, at a position 12
effluent ml. before lysine.

Tyrosine residues in ribonuclease A become modified at the phenolic
hydroxyl group by fluorodinitrobenzene. The resulting tyrosine ether
residues may be measured as o-dinitrophenyltyrosine subsequent to
hydrolysis. Under the conditions of acid hydrolysis used in the present
work, o-dinitrophenyltyrosine has proved to be a stable compound. It
is bound strongly by Amberlite IR-120 and is not eluted from the resin
under the conditions used in the automatic analyzer.

These observations on the dinitrophenyl derivatives of lysine, histidine,
and tyrosine have permitted the use of difference analysis to estimate the
extent of substitution of the ribonuclease A molecule on dinitrophenylation.

Completely dinitrophenylated ribonuclease A on hydrolysis yields
nine equivalents of e-dinitrophenyl-lysine and one equivalent of a,e-bis-
dinitrophenyl-lysine per molecule. At intermediate stages of substitution,
blocking of the a-amino group of lysine at the amino terminus of the peptide
chain in ribonuclease A is revealed by the inequality between ten equiva-
lents (the total lysine available) and the sum of the equivalents of unreacted
lysine and e-dinitrophenyl-lysine.

With histidine an independent check on the values obtained by dif-
ference is provided by the quantity of the unidentified degradation
product of /w-dinitrophenylhistidine formed during acid hydrolysis.
Free tyrosine, unlike histidine and lysine, undergoes destruction during
the acid hydrolysis normally used for amino acid analysis to the extent
of 5 to 8 %. Difference analysis as a measure of substitution of the six
tyrosine residues in ribonuclease A is therefore rendered uncertain to the
extent of the slight variations normally observed in the destruction of this
amino acid. For more reliable values, particularly at low degrees of sub-
stitution, it would be preferable to determine the o-dinitrophenyltyrosine
in the hydrolysates directly.

DINITROPHENYLATION OF RIBONUCLEASE A 45

Dinitrophenylation of ribonuclease A

In view of the strong affinity of ribonuclease for polyvalent anions (cf.
discussion bv Stein [5]) and the possibility that interaction with such
ions might modify the reactivity of some of the histidine and lysine side
chains, the ribonuclease A used in the present studies was subjected to a
preliminary ion-exchange step in which conversion to the acetate of the
protein was achieved.

The reaction between the protein and fluorodinitrobenzene was carried
out in aqueous solution at a constant pH, maintained with a pH-stat. A
large excess of reagent was used, and the solution was kept saturated with
reagent by vigorous stirring. As pointed out by Levy [6], under these
conditions the kinetics of the reaction with a particular functional group
become pseudo first-order in nature. The reaction may be stopped by
acidification to a pH of about 2, whereupon excess reagent and dinitro-
phenol may be removed by extraction.

Ribonuclease A can be completely dinitrophenylated. There are no
residues inaccessible to fluorodinitrobenzene, though it is evident that
wide variations in reactivity are to be found among the groups under-
going substitution. At pH 10, and at 40^, for example, approximately 30
hours are required for complete dinitrophenylation. After this time, some
0*3 residues of lysine remain unsubstituted, while free histidine and
tyrosine are absent from hydrolysates of the modified protein. After 15
hours under the same conditions, the equivalent of one lysine residue
remains unsubstituted, while reaction with histidine and tyrosine is com-
plete. It is therefore clear that a number (possibly one or two) of the
e-amino groups in the protein have decreased reactivity even at pH 10.
The propertv is probablv dependent on some feature of the primary struc-
ture, because performic acid-oxidized ribonuclease A exhibits similar
behaviour on dinitrophenylation.

It is noteworthy that completely dinitrophenylated ribonuclease A is a
soluble protein in the pH range from 2 to 10. This rather unusual property
is of value in handling the partly substituted derivatives of ribonuclease
A, for it permits the facile separation of excess fluorodinitrobenzene from
the protein.

Of more immediate interest in the present contribution are the residues
in ribonuclease A that are most reactive towards substitution by fluoro-
dinitrobenzene. At pH 8 analysis indicates that certain e-amino groups
are substituted more rapidly than any other functional groups, including
the a-amino group. This will be apparent from Table I, in which are given
the results obtained upon quantitative amino acid analysis of aliquots
removed after successive time intervals from a reaction mixture at pH
8 and at 15 . The values for all the residues in ribonuclease are shown, and

46 C. H. W. HIRS, M. HALMANN, J. H. KYCIA

in each case represent the resuhs of single determinations. The agreement
shown for the residues not affected by dinitrophenylation will serve to
give an impression of the degree of reproducibility attainable by this

approach.

TABLE I

The Dinitrophenylation of Ribonuclease A at pH 8-o, 15°

An aqueous 0-7 x lo"^ M solution of ribonuclease A (acetate) at pH 8-o and
at 15° was stirred vigorously with 20 times the quantity of i-fluoro-2,4-dinitro-
benzene required for the substitution of all the potentially available functional
groups in the protein. The initial volume was 13-6 ml. and the pH was maintained
at a value of 8-o±o- 1 with 0-2 N NaOH added with the aid of a pH-stat. Before
addition of the reagent an aliquot of the protein solution was removed, acidified
to pH I -8 with 006 N HCl, and the solution extracted 6 times with 20 volumes of
benzene. The protein was subjected to hydrolysis with constant boiling HCl in
an evacuated, sealed tube for 22 hours at iio\ after which the hydrolysate was
concentrated to remove excess acid, and the amino acid composition of the sample
determined by the method of Spackman et al. [4]. In a similar manner, aliquots of
the reaction mixture following the addition of FDNB were removed at successive
intervals of time, and the reaction in each case terminated by the addition of acid
to pH I -8. Thereafter, excess FDNB was removed by extraction, and the protein
subjected to quantitative amino acid analysis by the same procedure. The results
are expressed in terms of molar ratios of the constituent amino acids. Other aliquots
of the acidified, extracted reaction mixture were kept for determinations of
ribonuclease activity.

'1

ime of reaction

'minutes

Amino acid

0

3

6

20

40

60

90

180

240

Aspartic acid . . . .

ISO

15-3

15-1

151

150

151

149

is-i

151

Glutamic acid . . . .

120

120

120

II -9

II -9

121

II -9

I2-0

I2-0

Glycine . . . . .

3 04

313

314

301

3-07

3-00

3 -02

303

332

Alanine ....

II-Q

11 Q

II -9

II -9

II 9

II-8

I2-0

11-9

12 0

Valine ....

8-82

8-77

882

8-86

8-89

8-76

8-94

8-87

9-28

Leucine ....

IQ5

I 99

2 -OI

2 -03

2 -06

2 06

2- 00

I -96

2 -OI

Isoleucine*

2-35

2 34

2-50

2-3f>

2-57

2-53

2-74

2-42

2-53

Serinet ....

13-3

134

13-3

13s

12-9

13 -fJ

12-9

133

13-6

Threoninet

9-46

952

9-50

961

943

9-74

9-Sl

9-54

9-68

Cystine (half)t .

7-24

7-51

6-86

6-31

7 -09

5-99

6-6o

7-42

6 92

Methioninef

2 49

2-77

1-42

1-87

2-39

130

1-45

2-57

I 46

Methionine Sulphoxides

0-86

0-51

1-34

I -29

086

I -45

I -22

071

127

Proline ....

3-9

41

3-8

40

4-4

3-9

4-2

3-9

4-3

Phenylalanine

2-97

2 96

2-98

2-94

3-o8

3' 00

3 -08

2 96

3 07

Tyrosinei"

5-68

5-78

5-55

S-78

5-72

5-53

5-35

5-58

5-24

Histidine ....

4-00

4-04

3 -99

3-99

4-07

3-83

3-86

3 64

3-42

Lysine ....

10- I

100

Q-86

9-35

8-85

8 -09

7-69

6 69

6-33

Arginine ....

4 00

4 00

3-95

4-04

4-00

4 00

4 00

4-00

4 00

«-DNP-lysine .

—

00

0-2

0-53

I 04

I 40

I -97

2 06

2-54

Lysine +f-DNP-lysine

—

100

10- I

9-88

989

9 49

9-66

8-75

8-87

Unknown from histidine

—

0 0

o-o

00

00

001

008

029

042

,, ' Histidine .

—

4 04

3 -99

3 -99

4 07

384

3-94

393

3-84

♦ Incompletely liberated after hydrolysis for 22 hours,
t Correction for destruction on hydrolysis not applied.

Examination of Table I reveals that the only values altering signifi-
cantly with time are those for lysine, histidine, and tyrosine. Methionine
recovery was poor throughout, but the values for the sum of free methio-
nine and the methionine sulpho.vides remained satisfactorily constant.

DINITROPHENYLATION OF RIBONUCLEASE A 47

Had there been alkylation at the thioether sulphur of this amino acid, it is
hkely that the acid decomposition products of the alkyl derivative (cf.
Gundlach et al. [7]), particularly homoserine lactone, would have been
formed and detected. The analyses for the hydroxyamino acids are also
satisfactorily constant in the series, though it is clear that for an amino
acid like serine, of which there are fifteen residues in ribonuclease A, the
precision attainable by difference analysis would make it impossible to
detect a change of half a residue.

Comparison of the values for lysine, histidine, and tyrosine in Table I
shows that substitution on lysine is clearly the most rapid reaction. Sub-
stitution on histidine does not become detectable before the 60-minute
point, while reaction at tyrosine is not evident until the 240-minute
point (analyses of the products obtained at longer time intervals are not
included in Table I). A further breakdown of the results of the lysine
analyses is possible. The values for e-dinitrophenyl-lysine show that this
derivative appears to the extent of 1-04 equivalents in 40 minutes, at
which time the sum of the equivalents of unsubstituted lysine and e-
dinitrophenyl-lysine is equal (within the combined errors involved) to the
total available lysine. This sum becomes significantly less than 10 at 60
minutes, with e-dinitrophenyl-lysine at 1-4 equivalents, and falls to a
value of 9 at 180 minutes, at which time the value for e-dinitrophenvl-lvsine
has reached 2 • i equivalents. A value of 9 equivalents for the sum of free
lysine and e-dinitrophenyl-lysine is the value to be expected of a protein
in which the terminal a-amino group has been completely substituted.
Since determinations of a,e-bis-dinitrophenyl-lysine were not carried out,
an estimate of double substitution at the terminal lysine residue in ribonu-
clease A could not be made. Thus the values for e-dinitrophenvl-lvsine
represent minimum estimates of the extent of substitution of the e-amino
groups. Nonetheless, the results in Table I demonstrate that at least two
e-amino groups in ribonuclease A may react faster with fluorodinitro-
benzene than the a-amino group under these conditions.

Further progress in interpreting these observations depends on the
isolation and characterization of the partly substituted protein derivatives.
Moreover, without further information about the initial steps, it would be
impossible to investigate the kinetics of subsequent substitutions in the
molecule. As will be seen some success in the direction of isolating the
initial reaction products at pH 8 has been attained. Before considering
the isolation work, the effect of dinitrophenylation on the activity of the
enzyme will be described.

Loss of enzymic activity on dinitrophenylation

Ribonuclease A (acetate) is rapidly inactivated by dinitrophenvlation
at pH 8. After an initial lag phase (also obvious from the results in Table I),

48 C. H. W. HIRS, M. HALMANN, J. H. KYCIA

probably associated mainly with the process of saturating the solution
with fluorodinitrobenzene, the inactivation follows pseudo first-order
kinetics in the range from zero to about 95% inactivation. Representative
kinetic runs at three temperatures are shown in Fig. i. Ribonuclease activity
was measured with the aid of a pH-stat by the method of Gundlach et al.
[8] using cyclic cytidylic acid as the substrate. In the initial phases of the
work measurements of depolymerase activity were also effected. The
spectrophotometric method of Kunitz [9] with purified yeast ribonucleic
acid as substrate was employed with modifications to permit the use of a
Gary recording spectrophotometer. The results obtained with this method,
though less precise than those obtained with the nucleotide substrate,
gave the same kinetic constants.

Determinations of the pseudo first-order constant at 15° gave values
ranging from o-oio to 0-020 min."^ (based on decimal logarithms). The
same sample of ribonuclease A (acetate) gave values that agreed within
10%, but different samples of ribonuclease A (acetate) gave significantly
different values for "k". A possible explanation for this variation became
apparent subsequently when it was found that the inactivation of ribonu-
clease A by dinitrophenylation was strongly inhibited by pyrophosphate
ion. A quantity of pyrophosphate equivalent to that of ribonuclease A in
an experiment of the type shown in Fig. i was sufficient to decrease the
rate of inactivation one-fifth. It is possible that the reaction would also be
inhibited by other polyvalent anions, such as phosphate and sulphate,
just as the inactivation of the enzyme by carboxymethylation on histidine
is inhibited by such ions [5]. Because of the methods of preparation different
samples of ribonuclease A (acetate) could well be contaminated to varying
extents by orthophosphate ions.

The inactivation reaction at pH 8 is also inhibited by adenylate,
cytidylate, and uridylate. Inhibition by the pyrimidine nucleotides, which
are well known competitive inhibitors of ribonuclease, has been found to
be particularly effective. Cytidylate at equal concentration to ribonuclease
in an experiment of the kind shown in Fig. i caused the rate of inactivation
to be slowed to approximately one-quarter of the rate in the absence of
the inhibitor. At a ratio of 10 equivalents of cytidylate to i of ribonuclease
A inhibition was essentially complete. Adenylate was approximately half
as effective as cytidylate at the same concentration.

While the inactivation of the enzyme by dinitrophenylation was in-
hibited by these anions, the over-all reaction with fluorodinitrobenzene,
as measured by alkali uptake in the pH-stat, was not detectably slowed
down. Thus far a detailed analysis of the kinetics of the reaction as
measured with the pH-stat has not been attempted, but it is likely that,
with improved technique and working on a larger scale, a difference of
rate in the initial stages in the presence and absence of inhibitor will be

DINITROPHENYLATION OF RIBONUCLEASE A 49

detected. At present, however, it is evident that the presence of inhibitor
serves to slow the reaction down at relatively few of the available functional
groups on the protein molecule.

Examination of the values in Table I suggested that the inactivation
reaction was primarily due to substitution at an e-amino group. Ray and
Koshland have recently developed a treatment for dealing with the results

Fig. I. Kinetics of inactivation of ribonuclease A by dinitrophenylation. An
initial lag phase in the reaction is not shown. The conditions of dinitrophenylation
were the same as those described in Table I. The points represent the average of
two determinations at different concentrations of the remaining ribonuclease
activity. The procedure of Gundlach et ol. [8] was used with cyclic cytidylic acid
as the substrate.

obtained by kinetic analysis of modification reactions with proteins [lo].
With this approach, and on the assumption that during the initial stages
of dinitrophenylation (up to 40 minutes in Table I) the drop in the total
lysine value is due exclusively to reaction of fluorodinitrobenzene at but
two e-amino groups, a pseudo first-order reaction rate constant for these
two groups of o- DIG min.~^ was calculated. Determination of the remaining
ribonuclease activity in the reaction products during the same experiment

50

C. H. W. HIRS, M. HALMANN, J. H. KYCIA

permitted the evaluation of a pseudo first-order rate constant of o-oio
min.~^ (on the basis of decimal logarithms) for the inactivation reaction.
The results are represented graphically in Fig. 2, in which the difference
in time scale for the lysine and activity values are occasioned by the dif-
ference of 2 minutes in the lag phases observed for the dinitrophenylation
and inactivation reactions. The agreement in the values is probably to
some extent fortuitous in view of the usual errors that obtain in the

10

J I L

J \ L

80

10 20 40 60

MINUTES

Fig. 2. Kinetics of inactivation of ribonuclease A by dinitrophenylation.
Conditions were the same as described in Table I. For the significance of the
lysine values, see the text.

determination of lysine and the problems involved in removing excess
reagent before undertaking the measurement of ribonuclease activity. The
values for the constant nevertheless furnish quantitative evidence that the
substitution of a single e-amino group is capable of inactivating the enzyme.
The powerful inhibition effected by the nucleotides and pyrophosphate
ion suggest that the e-amino group in question is located at either the
binding or catalytic site of the molecule.

It is likely that other groups essential to the activity of the enzyme are
among those that become substituted much more slowly at pH 8 through

DINITROPHENYLATION OF RIBONUCLEASE A 5 1

the action of fluorodinitrobenzene. Their detection by kinetic methods
would demand a greater degree of refinement in the analytical work than
has been attained thus far. On the other hand, reaction of fluorodinitro-
benzene at these more slowly reacting groups may not be inhibited as
effectively by the agents already discussed, and their detection might
therefore be facilitated by the study of the consequences of dinitrophenyl-
ation in the presence of the competitive inhibitor of the enzyme.

Isolation of inactivated mono-e-dinitrophenylaminoribonuclease A

In an experiment similar to that described in Table I, aliquots of the
reaction mixture of ribonuclease A (acetate) and fluorodinitrobenzene at
pH 8 were chromatographed on an analytical scale on columns of the
sodium form of IRC-50 equilibrated with 0-2 m sodium phosphate buffer
at pH 6-47 [11]. The chromatograms sho\\ed that the reaction mixture
rapidly becomes complex. At higher degrees of substitution, the products
present could not be eluted satisfactorily from the resin.

When the products formed from 100 mg. of ribonuclease A (acetate)
inactivated to the extent of 20'',, were chromatographed on a larger scale
the result shown in Fig. 3 was obtained. The elution diagram shows a
simultaneous plot of the absorbance, determined at 360 m^i, and of the
ninhydrin colour value developed by aliquots of the eflluent fractions. The
values for the absorbance at 360 m^ and for the ninhydrin colour value
are normalized at the maximum of the elution peak at 350 eflluent ml. As
expected, the peak at 43:; eflluent ml. was found to represent unchanged
ribonuclease A. The trailing shoulder of the peak at 350 ml. exhibits an
esentially constant ratio of the ninhydrin colour value to absorbance at
360 m/i, suggesting the presence of material of a constant degree of sub-
stitution. Aliquots of the fractions in this portion of the chromatogram
when subjected to ribonuclease activity determinations were found to be
devoid of activity. As little as o-2"o of the original specific activity could
have been detected by the procedure used. Partly overlapping the peak
at 350 ml. on the chromatogram is a smaller peak at 325 ml., exhibiting a
variable ninhydrin colour value to absorbance at 360 m/x ratio. Material
corresponding to the fractions in this peak was enzymically active, as was
the protein in the fractions corresponding to the relatively unretarded
peaks, with varying ninhydrin colour value to absorbance at 360 m^ ratios,
between 250 and 300 efiluent ml. The latter evidently represent a mixture
of different partly substituted proteins with ribonuclease activity.

A cut was made of the yellow protein represented by the trailing
shoulder of the peak at 350 ml., and the resulting solution was freed of
buffer salts by gel filtration over a column of Sephadex G-25 [12]. On
rechromatography under identical conditions of elution the inactive

52

C. H. W. HIRS, M. HALMANN, J. H. KYCIA

dinitrophenylated ribonuclease A was eluted as a symmetrical peak at the
same position. A mixture of ribonuclease A and the inactivated protein
separated in the manner to be expected from Fig. 3. When the inactive
protein was rechromatographed on IRC-50 in 0-2 m sodium phosphate
buffer at pH 6-02 [13] a single, symmetrical peak was again obtained on

EFFLUENT VOLUME ml

Fig. 3. Chromatography of the reaction mixture from ribonuclease A (acetate)
and dinitrofiuorobenzene at pH 8 and 15 after the attainment of 20°,, inactivation.
Conditions for the reaction were the same as described in Table I. The IRC-50
column measured 37 x 330 mm. and was equilibrated with o • 2 M sodium phosphate
buffer at pH 6 -47. The rate of elution was 15 ml. /hour and the effluent was collected
in 5 ml. fractions. The absorbance of the effluent fractions was measured at 360 m/i
with a Beckman DU spectrophotometer. Aliquots of the fractions were also sub-
jected to ninhydrin analysis (cf. [11]), and to ribonuclease activity determinations
by the procedure described in Fig. i.

the elution curve with a maximum at 3 -6 times the effluent volume of the
maximum observed at pH 6-47.

On quantitative amino acid analysis the protein from the 350 ml.
elution peak in Fig. 3 gave the results shown in Table II. For comparison,
the results obtained on analysis of the ribonuclease A (acetate) used in
this experiment are also included. The values for all the residues agree
satisfactorily with the exception of those for lysine. The analysis makes it

DINITROPHENYLATION OF RIBONUCLEASE A

53

clear that the inactive protein contains a single residue of e-dinitrophenyl-
lysine. The ultra-violet absorption spectrum of the inactive derivative was
typical of the spectrum to be expected of an e-dinitrophenylamino deriva-
tive of ribonuclease. The analysis in Table II permitted the evaluation of a
molar extinction coefficient at the 365 m/^ maximum of 1-51 x iC*, a
value in the range usually observed [14] for the extinction coefficient of
amino-substituted dinitrophenyl derivatives of amino acids.

Degradation of inactive mono-e-dinitrophenylribonuclease A

The kinetic analysis in Fig. 2, the analytical values of Table II, and
the chromatographic results described in the previous section, strongly

TABLE II
Anhno Acid Composition of Ribonuclease A (Acetate) and of Inactive

DiNITROPHENYLATED RiBONUCLEASE A

For procedure of analysis, see Table I. The results are expressed in terms of
molar ratios of the constituent amino acids.

Amino acid Ribonuclease A

DNP—

derivativef

Aspartic acid . . . 15 "O

14-8

Glutamic acid

I2-0

12 -o

Glycine

3-05

299

Alanine

12 0

II-8

Valine

8-62

8-95

Leucine

1-98

I 94

Isoleucine .

2-31*

2-24*

Half cystine

6-32

7-13

iNIethionine .

3-51

3-57

Serine

133

12-7

Threonine .

935

9-33

Proline

4-4

4-0

Tyrosine

5-57

5-46

Phenylalanine

3-07

292

Histidine

3-92

3 96

Arginine

4-02

4-04

Lysine

10 -o

g-i8

€-DXP-lysine

0-94

* Incompletely liberated after hydrolysis for 22 hours,
t Molar extinction coefficient at 365 m/n = 1-51 x 10*.

implied that a single inactive protein was responsible for the peak at 350
effluent ml. in Fig. 3. In order to shed further light on this question some
120 mg. of the inactive dinitrophenylated protein were prepared by

54 C. H. W. HIRS, M. HALMANN, J. H. KYCIA

repeating an experiment of the kind illustrated in Fig. 3 a sufficient number
of times. The protein was freed from contaminating chloride ion and
subsequently oxidized with performic acid under conditions identical to
those used in structural work with ribonuclease A [15]. The performic
acid-oxidized ribonuclease A derivative was subjected to quantitative
amino acid analysis, the results of which demonstrated that complete
oxidation of the disulphide bonds had occurred, that methionine had been
quantitatively converted to the sulphone, and that there had been no
destruction or alteration [15] of tyrosine during the oxidation. The oxidized
protein derivative was submitted to tryptic hydrolysis at pH 7 in the
presence of catalytic quantities of trypsin prepared by the activation of
chromatographically purified [16] chymotrypsinogen-free trypsinogen.
The conditions were similar to those already described for the tryptic
hydrolysis of performic acid-oxidized ribonuclease A [17]. The mixture
of peptides formed after 24 hours of hydrolysis was fractionated on Dowex
50-X2 columns in the sodium form by procedures very similar to those
described for the tryptic peptides from oxidized ribonuclease A [17].
Formate and acetate buffers were used to facilitate quantitative amino
acid analysis of the peptide fractions (cf. [18]).

All the peptides found in a tryptic hydrolysate of oxidized ribonuclease
A [17] were present in the hydrolysate with the exception of peptides O-
Tryp 9 and 0-Tryp 14. Identification of the peptide fractions was in each
instance confirmed by quantitative amino acid analysis. In order to make
certain that ninhydrin-negative peptides were not being overlooked, the
aliquots removed from the efiluent fractions were subjected to alkaline
hydrolysis [17] prior to ninhydrin analysis. Within the precision attain-
able by quantitative analysis of appropriate cuts from the peptide zones
on the chromatogram it was clear that the peptides in the hydrolysate
of the oxidized, inactive derivative were formed in the same yields as
they are from oxidized ribonuclease A. It was not possible to detect
any dinitrophenyl peptides present in the tryptic hydrolysate in the
effluent from the Dowex 50-X2 column. Moreover, since subsequent
extraction with o • i n NaOH failed to remove any significant quantities
of material absorbing at 360 m/x from the resin, it is possible that severe
"tailing" was responsible for the loss of the dinitrophenyl peptides. The
high affinity of dinitrophenyl derivatives of amino acids for the poly-
styrene matrix of the resin has already been described.

Peptides 0-Tryp 9 and 0-Tryp 14 are related [19] in the amino acid
sequence of ribonuclease. Peptide O-Tryp 14 is an intermediate in the
tryptic hydrolysis of oxidized ribonuclease A. On further hydrolysis it
breaks down into peptides 0-Tryp 9 and 0-Tryp 7 (Asp . Arg) by cleavage
at the carbonyl bond of the arginine residue of peptide 0-Tryp 7. The
absence of peptides 0-Tryp 9 and 0-Tryp 14 in the Dowex 50-X2 eluate

DINITROPHENYLATIOX OF RIBONUCLEASE A 55

thus reveals that modification of a lysine residue in 0-Tryp 9 must have
taken place as the only consequence of dinitrophenylation. There are two
lysine residues in peptide 0-Tryp 9. Since the peptide following 0-Tryp 9
in the amino acid sequence is 0-Tryp 5, had dinitrophenylation taken
place at the carboxyl-terminal lysine residue of peptide 0-Tryp 9, tryptic
hydrolysis at this residue would have been blocked and 0-Tryp 5 could
not have formed. Since peptide 0-Tryp 5 was present in the tryptic
hydrolysate, the dinitrophenylated lysine residue in 0-Tryp 9 must have
been the lysine residue near the amino-terminal end of this peptide.
Examination of the amino acid sequence [i] shows that this is the lysine
residue at position 41 along the chain. Experiments on the further charac-
terization of the inactive protein, and on the isolation of dinitrophenylated
peptide derivatives are now in progress.

Conclusions

These preliminarv studies have revealed that amino groups are the
most reactive functional groups in ribonuclease towards substitution by
dinitrofluorobenzene at pH 8. Two e-amino groups react faster than the
a-amino group of the lysine residue at the amino-terminal end of the pep-
tide chain. The substitution of one of the rapidly reacting e-amino groups
is accompanied bv the inactivation of the enzyme. This group is the
e-amino group of the lysine residue at position 41.

The ease of substitution of the e-amino group of the lysine residue at
position 41 is of interest in relation to the primary structure of the protein.
One of the relativelv few unique features of the amino acid sequence is
the accumulation of basic amino acid residues in the region between
residues 31 and 41. As pointed out elsewhere [i], there are five residues
capable of conferring a positive charge in these 1 1 residues, and only one
carboxyl group. A further structural feature of significance is that lysine
residue 41 is preceded by half-cystine residue 40 and followed by proline
residue 42. The combined influence of charge repulsion at pH 8 among
the cationic centres and the influence of the adjoining cystine and proline
residues would be factors adversely affecting the stability of an a-helix
formed in this part of the molecule. An additional feature of note is that
half-cystine 40 is the first half-cystine of the II-VII disulphide bond [2].
Examination of the primary structural formula will reveal that the II-VII
disulphide bond is a fulcrum about which any folding of the ribonuclease
molecule must hinge. The ease of substitution of lysine residue 41 may
therefore indeed reflect a greater degree of accessibility of this residue
because of less compact folding of the structure in this region.

The inhibition of dinitrophenylation at lysine residue 41 by cytidylate
and pyrophosphate is extremely effective even when only one equivalent

56 C. H. W. HIRS, M. HALMANN, J. H. KYCIA

of inhibitor is present per molecule of enzyme. This makes it likely that
the e-amino group of this residue is closely related to the catal}^ic function
of the protein. Substitution on lysine may, because of the size of the sub-
stituent, locally distort the tertiary structure, or prevent the formation of a
critical hydrogen bond required to maintain the configuration of the
active site during catalysis. Substitution may also, by removing a potential
positive charge, cause a collapse of an electrostatically maintained con-
figuration, or prevent electrostatic interaction with the substrate. Finally,
the introduced dinitrophenyl group may be efi^ective in preventing bound
substrate from interaction with the catalytically essential, bond-breaking
amino acids. Further work will permit a narrowing down of the possible
alternatives. In the meantime, it is worth recalling that in our earlier
work, the speculation was made [i] that the residues between positions 31
and 41 might constitute a binding site for anions. A similar idea was recently
expressed by Parks [20], who has proposed a model for the tertiary structure
and the mechanism of action of ribonuclease.

Inactivation of ribonuclease by modification of lysine residues has
been observed previously. Gundlach et al. [8] have shown that carboxy-
methylation of ribonuclease A at pH 8 is accompanied by the inactivation
of the enzyme, and that, under these conditions, the reaction is limited
to the lysine residues in the protein. Taborsky has phosphorylated ribonu-
clease A [21] with imidazole phosphate and has isolated an inactive mono-
phospho ribonuclease in which the introduced phospho group is on a
lysine residue. Carboxymethylation and phosphorylation on e-amino
groups results in a charge reversal, whereas introduction of a dinitrophenyl
group makes the lysine residue neutral. If, as seems likely, the same lysine
residue is involved in the inactivation reaction with iodoacetate, imidazole
phosphate, and fluorodinitrobenzene, it is possible that the groups intro-
duced effect inactivation by different mechanisms.

In conclusion we may briefly list the structural features in ribonuclease
A now known to be important in the catalytic action of the protein.
Studies on ribonuclease S (for summary, cf. Richards [22]) have demon-
strated that the binding of S-peptide (residues 1-20 in the original pro-
tein) is essential to the maintenance of activity in ribonuclease S. Carboxy-
methylation at pH 5 results in inactivation of ribonuclease A [8] by reaction
of the histidine residue at position 119 [23]. Limited pepsin degradation
has revealed that aspartic acid residue 121 is required for the maintenance
of activity in ribonuclease A [24]. With the implication of the lysine
residue at position 41, it is becoming increasingly clear that for the
maintenance of the configuration of the catalytic and binding sites of the
molecule, the co-operative interaction of many functional groups, located
at widely separated points in the primary structural formula, is
required.

DINITROPHENYLATION OF RIBONUCLEASE A 57

Acknowledgments

This research at Brookhaven National Laboratory has been performed
under the auspices of the United States Atomic Energy Commission. The
participation of one of us (C. H. W. H.) at the Symposium was made
possible by the combined support of the National Science Foundation
(grant No. NSF-G- 12926) and the United States Atomic Energy Com-
mission. We wish to thank Mrs. B. M. Floyd for her assistance in the
performance of the amino acid analyses.

References

1. Hits, C. H. W., Moore, S., and Stein, \V. H., J. biol. CJiem. 235, 633 (i960).

2. Spackman, D. H., Stein, W. H., and Moore, S.,y. biol. Chem. 235, 648 (i960).

3. Moore, S., Spackman, D. H., and Stein, W. H., Anolyt. Chem. 30, 1 185 (1958).

4. Spackman, D. H., Stein, W. H., and Moore, S., Analyt. Chem. 30, 1 190 (1958).

5. Stein, W. H., in Brookhaven Symposia in Biology 13, "Protein Structure and
Function", Upton, N.Y., 104 (i960).

6. Levy, A. L., Nature, Lond. 174, 126 (1954).

7. Gundlach, H. G., Moore, S., and Stein, W. H.,_7. biol. Chem. 234, 1761 (1959).

8. Gundlach, H. G., Stein, W. H., and Moore, S.,_7. biol. Chem. 234, 1754 (1959).

9. Kunitz, M.,J. biol. Cheyn. 164, 563 (1946).

10. Ray, W. J., Jr., and Koshland, D. E., Jr., in Brookhaven S\Tnposia in Biology
13, "Protein Structure and Function", Upton, N.Y., 135 (i960).

11. Hirs, C. H. W., Moore, S., and Stein, W. H.,y. biol. Chem. 200, 493 (1953).

12. Porath, J., and Flodin, P., Nature, Lond. 183, 1657 (1959).

13. Hirs, C. H. W., J', biol. Chem. 205, 93 (1953).

14. Fraenkel-Conrat, H., Harris, J. I., and Levy, A. L., in "Methods of Bio-
chemical Analysis", ed. D. Glick. New York, 2 (1955).

15. Hirs, C. H. \\.,y. biol. Chem. 219, 611 (1956).

16. Keller, P. J., Cohen, E., and Neurath, H., J. biol. Chem. 233, 457 (1958).

17. Hirs, C. H. W., Moore, S., and Stein, W. U., J biol. Chem. 219, 623 (1956).

18. Hirs, C. H. \W.,y. biol. Chem. 235, 625 (i960).

19. Hirs, C. H. W., Stein, W. H., and Moore, S.,^. biol. Chem. 221, 151 (1956).

20. Parks, J. M., in Brookhaven Symposia in Biology 13, "Protein Structure and
Function", L^pton, N.Y., 132 (i960).

21. Taborsky, G.,_7. biol. Chem. 234, 2915 (1959).

22. Richards, F. M., /// Brookhaven Symposia in Biology 13, "Protein Structure
and Function", Upton, N.Y., 115 (i960).

23. Barnard, E. .\., and Stein, D. W.,^. mol. Biol. I, 339 (i960).

24. Anfinsen, C. B.,J. biol. Chem. 221, 405 (1956).

The Relation of the Secondary Structure of Pepsin to
Its Biological Activity

Gertrude E. Perlmann

The Rockefeller Institute,
New York, U.S.A.

Two decisive advances in the understanding of protein structure are
the elucidation by Sanger of the amino acid sequence of insuhn [i] and
the more recently completed investigations on pancreatic ribonuclease by
Hirs, Moore, and Stein [2]. The biological activity of a protein, however,
depends not only on the amino acid sequence but also on the folding of
the peptide chains, their arrangement in space and how they are packed
into the protein molecule. The only technique available for obtaining
detailed and precise information about the folding is undoubtedly the
X-ray method which has been demonstrated so clearly by the work of
Kendrew and his collaborators on myoglobin [3]. In the studies on pepsin
which I shall report, more indirect methods have been used. Nevertheless,
such methods may help in defining the conformation of the peptide chain
segment necessary for the biological activity of the enzyme, as well as the
type of forces responsible for maintaining the "native" configuration.

Thus far, amino acid sequence work on pepsin has not been attempted,
which may partly be attributed to the size of the molecule. Pepsin is a
protein with a molecular weight of 35 000 and has only one peptide chain
which is cross-linked by three disulphide bonds. Examination of the
amino acid distribution, as given in Table I, reveals that the protein has
71 acidic residues (44 aspartic acids and 27 glutamic acids) and only four
basic ones (i histidine, i lysine, 2 arginines). Moreover, pepsin has a
high content of non-polar and hydroxyamino acids and 15 prolines [4]. It
is, therefore, not unlikely that such an unusual amino acid distribution
influences the secondary structure of the protein.

It is well established that in most globular proteins a certain fraction
of the amino acid residues is in the a-helical configuration. Whenever a
helical structure is present, hydrogen bonds between the oxygen atoms of
the carbonyl groups and the imino group of the peptide linkages play an
essential role in determining the folding of the polypeptide chain. However,
hydrophobic bonds may also be important in maintaining the secondary
structure.

6o

GERTRUDE E. PERLMANN

TABLE I

Amino Acid Composition of Crystalline Pepsin [4]

Nature of amino acid

No. of residues
per molecule

Per cent of
total

Acidic: (Asp,Glu) .

71

20-7

Basic: (His,Lys,Arg)

4

1-2

Non-polar: (Gly.Val, Leu, lieu

,Ala,Met)

137

40-0

Hydroxy: (Ser,Thr)

72

20-9

Aromatic : (Tyr,Try,Phe)

38

II 'I

Proline ....

IS

4-4

1/2 Cys (-S-S-) .

6

1-7

Total

343

lOO-O

That pepsin is a tightly folded molecule is indicated by its low intrinsic
viscosity. The question arose, therefore, as to whether or not reagents
could be found that would unfold the polypeptide chain and if unfolding

100

80

60-

^ 40

20

• Dilute HC
o •■ ■• + 8m urea
A ■• ■• + 3m
guanidine
hydrochloride

0 1-0 20 30 40 5-0

pH

Fig. I. pH dependence of the activity of pepsin for the hydrolysis of haemo-
globin.

occurred, how certain properties, such as viscosity, optical rotation and
the biological activity of the enzyme, would be affected.

The first two reagents which were investigated for their effect on pepsin
were urea and guanidine hydrochloride which are known to have a pro-
found influence on the biological activity of enzymes, e.g. ribonuclease [5],
chymotrypsin [6] and trypsin [7]. In contrast to the enzymes, which are

THE RELATION OF THE SECONDARY STRUCTURE OF PEPSIN

6i

readily inactivated if brought into contact with these reagents, pepsin
remains active after short exposure to urea and guanidine hydrochloride.
However, as illustrated with the aid of Fig. i, the pH of maximum hydroly-
sis, which in aqueous solutions is at pH i • 8, is shifted to an apparent pH
of 2-3 in 8-0 M urea and to pH 3-0 in 3-0 m guanidine hydrochloride.

Further investigation of the effect of urea and guanidine hydrochloride
on pepsin has revealed that the enzymic activity in the presence of these
hydrogen-bond breaking reagents depends on a variety of factors: (i)
concentration of the reagent, (2) time of exposure, (3) temperature, and
(4) pH.

Fig. 2. Dependence of pepsin activity on time of exposure to guanidine
hydrochloride at 37".

In Fig. 2, are shown the results of measurements in which the con-
centration of guanidine hydrochloride was varied but the pH of the
reaction mixture maintained at 3-4. In 3-0 m guanidine hydrochloride,
pepsin is almost as active as in buffer, whereas in 6-o M guanidine hydro-
chloride complete inactivation occurs within 30 minutes of contact with
the reagent [8]. A similar behaviour of the protein is observed in experi-
ments in which the urea concentration is varied from 2-0 M to 8-o M [9].

The second factor mentioned is temperature. If, at a constant concen-
tration of 4-0 M guanidine hydrochloride of pH 3-4, the temperature is
lowered, the rate of inactivation decreases. Thus at 37 , the first-order

rate constant, k, is 1-53 x lO"
k = 1-5 X 10^ X min."^ (8).

at

■06

iQ-* and at 25°,

62

GERTRUDE E. PERLMANN

The next point to be discussed is the effect of pH on the rate of in-
activation. In Fig. 3, is given a comparison of the effect of 8-o m urea and
4-0 M guanidine hydrochloride if pepsin is exposed to the reagent for i
hour at 37°. Three points become apparent: (i) The range of maximum
stabihty of pepsin, which in aqueous solution extends to pH 5-5, is
narrowed to pH 3-0 to 3-5 in 4-0 m guanidine hydrochloride and to pH
3-3 to 4-3 in 8-0 M urea, respectively. (2) Although denaturation of
pepsin at low pH values, i.e. i -5, has been reported by Northrop [10], the
loss of activity in the presence of hydrogen-bond breaking reagents
proceeds rapidly below pH 3 -o, and the formation of low molecular weight

_too

•

Aqi

f^rM \^

\Z\J\JO

/■ f f

A

4m

guanidine hydrochi

oride / / /

^_^

0

8m

urea

/ / /

§ 80

-

\\

/ / / '

0

>

/ / /

? 60

-

if:

1

r / /

c
I/I

g 40

-

', 1

1

c

fc 20

-

i,

/ /[ -

-=.

=T*-

^^

^«^ , JIx

100

80

-60

-40

-20

O 0)

o^

-c E
- ^

% O

0 12 3 4 5 6

pH

Fig. 3. Effect of various solvents on pepsin. Time of incubation: i hour at

37 •

peptides, resulting from the action of intact pepsin on the denatured
material, closely parallels the rate of inactivation. (3) At pH values more
alkaline than 6-o, pepsin loses its activity spontaneously without the forma-
tion of non-protein material. In the presence of guanidinium ions or urea,
this rapid loss occurs at more acid pH values. Moreover, it is of interest
to note that although the pH zone in which inactivation takes place depends
on the reagent and its concentration, the maximum rate of inactivation,
within a small pH interval is independent of the nature of the solvent. One
may, therefore, conclude that intramolecular bonds of the same nature
are sensitive to the hydroxyl ions of the medium. Thus, in the pH range
of 3-5 to 5-5 the main effect of the hydrogen-bond breaking reagents
consists in loosening the secondary structure of the pepsin molecule.

THE RELATION OF THE SECONDARY STRUCTURE OF PEPSIN 63

Consequently, some of the bonds that are necessary for maintenance of
the enzymically active configuration are broken at lower pH values than
in aqueous solution.

That the loosening of the configuration is a slight one and not similar
to the unfolding that is observed if other proteins are brought into contact
with hydrogen-bond breaking reagents is apparent from the intrinsic
viscosity given in Table II. Thus the intrinsic viscosity changes only from

TABLE II

Intrinsic Viscosity, Specific Optical Rotation, [a], and Rotatory Dispersion
Constant, A,., of Pepsin in Various Solvents at 25

Composition of solvent pH* (g/ml) ' 600 m/t 400 m/i A^

o-i M Acetate buffer . . .4-64 3-09 63-4 178 216

8 M Urea-acetate . . • S'Ss 3 "5? 64-1 179 218

3 M Guanidine hydrochloride . 3'4o 3'54 62 • i 173 217

4 M Guanidine hydrochloride . 3"4o 3'47 62-1 173 217

* Apparent pH.

3-09 to 3-54 (g/ml)^ if 8-0 M urea or 3-0 to 4-0 M guanidine hydro-
chloride is present in the reaction mixture. Moreover, as also shown in
Table II, the specific optical rotation, [a], and the rotatory dispersion con-
stant, A^., remain completely unaltered [11]. This behaviour is in contrast
to that found in the case of many proteins where the values of the specific
rotation, [a],„ of the native protein decreases by 20 to 60'' upon denatura-
tion [12]. Likewise, the dispersion constant. A,,, is lowered from the range
of 230-270 m/Li to 210 m/Li upon contact with hydrogen-bond breaking
reagents [13]. These changes usually reflect a major unfolding of the
polypeptide chain or a transition from an a-helical structure to a random
coil. Thus, the results obtained with pepsin further support the view that
no major change is brought about after short exposure to urea or guanidine
salts.

If, however, as shown in Table III, the optical rotation, [7.], and the
rotatory dispersion constant, A,,, are measured at 55", A^, increases from
216 to 230 m^ii. As illustrated in Fig. 4, the increase of the rotatory dis-
persion constant occurs above 45", but, as indicated by the dashed lines,
the loss of activity accompanies or precedes the changes of A,,. It appears,
therefore, that certain amino acid residues, e.g., proline or serine, lock the
peptide chain into a configuration which confers considerable stability
upon the protein. At higher temperatures, however, transition to a less
stable configuration takes place. It is clear that in this case the chain

64

GERTRUDE E. PERLMANN

TABLE III

Rotatory Dispersion Constant, A,., and Specific Activity of Pepsin in Various

Solvents as Function of Temperature

Relative specific
Composition of solvent pH* Temperature A^ activity per unit

nitrogen in per centf

o • I M Acetate buffer

4-34

23

2l6

100

55

227

99

8 • o M Urea-acetate

5-35

23

2X8

95

55

229

5

5 -o M Guanidine

hydrochloride

3 -So

23

217

100

55

226

4

* Apparent pH.

t The relative specific activity of a freshly prepared pepsin solution in o • i M
acetate buffer of pH 46 is taken as 100.

230

5*1: 225

^^
o o

u O

220

• 0-In acetate
buffer

- D 8-Om urea
A 5'Om guanidine

hydrochloride \i

-g---^n

100

80

60

40

20

20 30 40 50

Temperature ( C)

60

Fig. 4. Dependence of temperature of the optical rotatory dispersion constant,
A,,, and enzymic activity of pepsin in various solvents.

segment necessary for the biological activity of the enzyme has been
affected.

From the results presented here, one can conclude :

I. Although hydrogen bonds undoubtedly exist in pepsin, they are
relatively unimportant in maintaining the configuration of the protein
essential for its enzvmic action.

THE RELATION OF THE SECONDARY STRUCTURE OF PEPSIN 65

2. Pepsin does not contain large helical regions but has essentially a
"random coil" configuration. This lack of any repeating or periodically
organized structure as a helical conformation represents, however, does
not exclude that each amino acid residue is in a specific or unique location.

3. We should like to propose that in the case of pepsin hydrophobic
bonds play a considerable role in determining the secondary structure of
this enzyme.

Acknowledgment

This work was supported in part by Grant A2449 of the United States
Public Health Service.

References

1. Sanger, F., and Tuppy, H., Biochem. J. 49, 463, 481 (1951); Sanger, F., and
Thompson, E. O. P., Biocheyn. J. 53, 353 (i953); Ryle> A. P., and Sanger, F.,
Biochem. J. 60, 535 (1955); Ryle, A. P., Sanger, F., Smith, L. F., and Kitai, R.,
Biochem. J. 60, 541 (i955)-

2. Hirs, C. H. \V., Moore, S., and Stein, W. H., J. biol. Chem. 235, 633 (i960).

3. Kendrew, J. C, Davies, D. R., Phillips, D. C, and Shore, V. C, Nature,
Loud. 185, 422 (i960); Kendrew, J. C, Bodo, G., Dintzis, H. M., Parrish,
R. G., Wyckoff, H., and Phillips, D. C, Nature, Loud. 181, 662 (1958);
Bluhm, M. AI., Bodo, G., Dintzis, H. M., and Kendrew, J. C, Proc. roy. Soc.
A 246, 369 (1958).

4. Blumenfeld, O. O., and Perlmann, G. E., jf. gen. Physiol. 42, 553 (i959)-

5. Sela, M., and Anfinsen, C. B., Biochim. biophys. Acta 24, 229 (i957)-

6. Neurath, H., Rupley, J. A., and Dreyer, W. J., Arch. Biocheyn. Biophys. 65,

243 (1956).

7. Harris, J. I., Nature, Lond. 177, 471 (1956)

8. Blumenfeld, O. O., Leonis, J., and Perlmann, G. E., jf. biol. Chetn. 235, 379
(i960).

9. Perlmann, G. E., ArcJi. Biochem. Biophys. 65, 210 (1956).

10. Northrop, J. H., ^. gen. Physiol. 16, 33 (1932).

11. Perlmann, G. E., Proc. nat. Acad. Sci., Wash. 45, 915 (i959)-

12. Simpson, R. B., and Kauzmann, W., J. Amer. chem. Soc. 75, 5139 (i953)-

13. Linderstrom-Lang, K., and Schellman, J. A. Biochim. biophys. Acta 15, 203
(1955)-

The Problem of Nucleotide Sequence in
Deoxyribonucleic Acids

Erwin Chargaff
Columbia University, New York, N. Y., U.S.A.

Introduction

Now that some of the most tremendous problems in biochemistry have,
it is said, been solved successfully — such as the nature of the genetic
material, its structure and mode of replication, and even its biosynthesis —
our time, tired of so much exertion, may take a deep breath and then,
perhaps, decide that it all has to be done over again. I should not dare
contradict, for I have always been more impressed by the enormity of the
gap between claim and achievement than by the magnitude of the former.

In any event, we have set ourselves a much, more modest, but by no
means easv, task, namely, the elucidation of some of the aspects of the
primary structure of deoxyribonucleic acids, especially in regard to the
arrangement of the nucleotide constituents. In this connection, and since
this talk is given in Stockholm, I may be permitted to recall that it was
here, and also in Uppsala, that I had the first opportunity to review our
original observations on deoxyribonucleic acid. This is what I said in
1949 and what was printed a few months later [i].

"The desoxypentose nucleic acids extracted from different species
thus appear to be different substances or mixtures of closely related
substances of a composition constant for different organs of the same
species and characteristic of the species.

"The results serve to disprove the tetranucleotide hypothesis. It
is, however, noteworthy — whether this is more than accidental,
cannot yet be said — that in all desoxypentose nucleic acids examined
thus far the molar ratios of total purines to total pyrimidines, and also
of adenine to thymine and of guanine to cytosine, were not far from i."

You will recognize here, among other things, the first statement of the
well-known pairing principles.

Even the earliest observations on the nucleic acids, which showed the
existence of remarkable similarities and, at the same time, of outstanding
differences in the distribution and, therefore, the sequence of the con-
stituent monomers, the nucleotides, made it appear of great interest to

08 ERWIN CHARGAFF

learn something about the structural principles that came into play. I
shall first discuss some of the basic concepts that must be considered.

Remarks on the conceptual basis of sequence analysis

THE PROBLEM OF HOMOGENEITY

What determines the composition of a given biological polymer has
been much discussed in the recent past, at any rate in regard to proteins
and nucleic acids. Almost no attention has been paid, as has been pointed
out recently [2], to the presumably equally rigid control of the composition
of other cell-specific polymers, such as the polysaccharides or even certain
macromolecular lipids. The usual expedient, in which I wish I could
concur with greater enthusiasm [3], has been the formulation of some
sort of template which obviously makes up in versatility for what it lacks
in concreteness. This Disneyland of templates and pools and feedbacks,
presided over by never-smiling augurs, will make us the laughing stock of
later times.

What determines the size of a given biological polymer has, on the
other hand, been neither determined nor even much discussed. The cell
will in many instances contain, or at least it will be made to yield, com-
pounds belonging to the same class, but very much dilTerent in size. Pro-
teins, the best investigated group of cellular macromolecules, vary in size
over a considerable range, from something like 12 000 to many millions [4].
Deoxyribonucleic acid preparations appear, in contrast, to fall into two
principal groups of molecular weight: (a) of 6 to 8 million; (b) of 12 to 16
million [5].* Such distinctions are probably not of great heuristic value,
since the cell, which is not simply a bag containing many chemicals of
different size and quality, must impress on its components a pattern of
multiple associations which it is not possible even to define at present.

Dissimilarity in size is, however, less of a problem for the student of
the chemistry of deoxyribonucleic acids than is the possible heterogeneity
of his specimens as regards their composition and sequence characteristics.
It is, in fact, possible to divide a preparation of total deoxyribonucleic
acid into a series of fractions of graded composition (starting with a, some-
times extreme, GC-type [7] and going to a very marked AT-type) which,
however, all still exhibit the pairing regularities [8] ; but the real meaning
of such a fractionation is far from clear. It could be, as was pointed out by
us [9], that these fractionated preparations actually represent some form
of sub-units of a large aggregate, though it is not easy to describe the nature

* An interesting exception — a polydeoxyribonucleotide of mol. wt. 10 000 to
12 000 — has been described recently as a component of crystalline cytochrome
62 [6]. It is noteworthy that this small polymer of only about 40 nucleotides still
exhibits the pairing principles [i] mentioned above.

THE PROBLEM OF NUCLEOTIDE SEQUENCE IN DEOXYRIBONUCLEIC ACIDS 69

of the links that are disrupted repeatahly during the mild fractionation pro-
cedure. In this view, in its extreme form, the nucleus of a given species
could conceivably contain only one type of deoxyribonucleic acid. At the
other extreme would have to be placed the view that a given nucleic acid
preparation may comprise an entire spectrum of differently constituted
individuals so that no two nucleic acid molecules within the same nucleus
would be entirely identical [lo]. That neither of these opposite extremes can
as yet be disavowed entirely demonstrates the deep cleft that still exists
between metabiological assertions and scientific facts.

THE PROBLEM OF MACROMOLECULAR STRUCTURE

It would, of course, be good if the investigator of the structure of a
deoxyribonucleic acid knew whether he was dealing with one chain or
with two [11] or even more [12] complementary chains; but at the present,
little advanced, state of our knowledge of the nucleic acids this is not
essential. The value of anv information on the sequential characteristics
of the nucleic acid is not diminished if the complementary chains are
considered to be joined at one end or even at both ends, as may well be
the case, so as to constitute an uninterrupted sequential progression. I
am not aware of the demonstration of distinct end groups in undegraded
deoxyribonucleic acid.

The maintenance of sequence integrity of the preparations to be
examined is, on the other hand, a problem of great importance. If — quite
apart from the secondary valence forces supporting the architecture of
the secondary or tertiary structures of the deoxyribonucleic acids — there
really exist weak links in their primary structure [13], it may be essential
to avoid the rupture of the latter, since otherwise valuable features of
sequential arrangement may be lost. This applies, of course, even more to
the avoidance of chemical or enzymic degradation during the isolation.

It is quite likely that very few of the deoxyribonucleic acid preparations
described so far, and none of the models proposed to describe their struc-
ture, are representative of the native state. The actual operative entity — •
though not necessarily amenable to biological testing /// vitro — possibly is
an aggregate of very long polynucleotide chains linked to each other,
perhaps by oligopeptide bridges, and bonded to proteins in a spatially
unique configuration. If this is so, the problem of heterogeneity, men-
tioned in the preceding section, is ostensible rather than actual, having
been introduced as a necessary, but strictly non-biological, artifact of
purification. By this token, a preferred isolation method for deoxyribonu-
cleic acid would be one that avoided, as far as possible, denaturation by
chemical or physical means. Nearest to these requirements is perhaps
the procedure described on p. 326 of a previous survey [7] and applied

yo ERWIN CHARGAFF

by us to preparations from calf thymus [14] and Escherichia coli protoplasts

[15].

THE PROBLEM OF STATISTICAL SEQUENCE ANALYSIS

The possibility that no two nucleic acid molecules within the same
nucleus are entirely identical offers ' ' a prospect that would seem to condemn
us to forced statistics for life" [16]. It is this more than anything else that
has made work on the nucleotide sequence in nucleic acids appear so un-
attractive. Who would, after all, undertake to read a book that has been
passed through a grinder ? Nevertheless, being rather modest in what I
expected to gain from a perusal of the nucleic acid text, I have never been
able to share these apprehensions completely. I knew that a great deal can
be learned about an unknown language through a study of its phonemes,
their frequency, distribution density, and allophonic relationships.

If the total deoxyribonucleic acid of a given species represents a text,
it is made up of "words" — the individual molecules — that are composed
of a singularly meagre alphabet: four or five letters. But the words so
spelled out are 10 000-letter words, each of which could occur in a fan-
tastically great number of positional isomers: between lo-^oo ^nd 10^°*^*',
according to how many restrictions on neighbours are admitted [3]. The
situation facing us in examining a nucleic acid preparation comprising a
large number of isomers or homologues would, then, be comparable to
one in which all the words in a dictionary are lined up end to end in a
continuous, and essentially arrhythmic and aperiodic, sequence.

It is quite clear that the first attempt at unravelling such a clutter will
have to be based on statistics and that it must limit itself to the description
of tendencies or trends of arrangement. To give an example : running to-
gether the thirteen words making up the first sentence of King Lear I
obtain a monster word of fifty-seven letters of which twenty-one are
vowels. On this word a number of determinations can be made : (a) the
ratio of consonants to vowels ; {b) the nature of the individual consonants
and vowels; (c) the relative frequency of each constituent. If I have a
way of removing the vowels without disturbing the rest of the arrange-
ment, I shall isolate six solitary consonants, eight pairs of consonants,
three bunches of triple consonants and one cluster of five consonants in a
row. Each of these units, I would conclude, was originally flanked on
both sides by vowels. Other words would yield other combinations, with
the unambiguity of distinction increasing with the length of the consonant
clusters. In very long words composed of only two vowels and two or
three consonants, unique clusters can be expected only very rarely; but
the relative frequency of the various combinations of consonants (runs of
I, 2, 3, etc.) will be a means of unique differentiation, even though it will
not yet make it possible to reconstruct the entire text.

THE PROBLEM OF NUCLEOTIDE SEQUENCE IN DEOXYRIBONUCLEIC ACIDS 7 1

This is what we have been doing with many deoxyribonucleic acid
preparations.

Remarks on nomenclature

DEFINITION OF THE TERM " HOMOTOPE "

If one monomeric constituent can take the position of another in a
definable segment of a polymer, we propose to designate it as a homotope
(from the Greek for occupying the same place). If in the A-chain of insulin
positions 8 to lo are occupied by Ala.Ser.Val in the ox [17], but by
Ala.Gly.Val in the sheep [18], I would say that serine and glycine are
homotopic with respect to this sequence. The importance of this term
for a consideration of nucleic acid structure will become clear presently.

DEFINITIOxN OF THE TERM " PLEROMER "

If in the total composition of polymers that are characterized by
balances such as the well-known pairing principles in deoxyribonucleic
acid [i, 7], and in a more limited way in ribonucleic acids [19], one con-
stituent can ostensibly replace another in respect to these balances, we
propose to designate it as a pleromer (from the Greek for filling up the
measure). Thus, if in the deoxyribonucleic acid of wheat germ [20] the
molar quantity of guanine equals that of the sum of cytosine and 5-methyl-
cvtosine, I would sav that these two 6-amino pvrimidines are pleromeric.

SOLITARY AND BUNCHED NUCLEOTIDES

A pyrimidine nucleotide that, within the polynucleotide chain of a
nucleic acid, is contiguous only to purine nucleotides, will be referred to as
"solitary"; pyrimidine nucleotides that occur in tracts of two or more,
flanked on both sides by purine nucleotides, will be designated as
"bunched" [21]. In the diagram shown in Fig. i, solitary thymidylic acid,
for instance, appears in positions 3, 9 and 8', bunched pyrimidine nucleo-
tides are seen in positions 5 to 7.

Early attempts at sequence investigation

Since 1947, when in an early review on deoxyribonucleic acids (p. 32
of ref. [22]), I first discussed the possible importance of variations in the
nucleotide sequence for the biological specificity of a deoxyribonucleic
acid, this problem has remained among the interests of my laboratory.
Our first chemical evidence on the existence of different deoxyribonucleic
acids [23, 24] made it clear that whatever "code" was carried by the

72 ERWIN CHARGAFF

nucleic acids must be imprinted on the sequential arrangement of the
monomeric constituents and prompted us to search for ways to approach
this difficult task.

The general problem is similar to that of the structure of proteins, but
immensely more forbidding. The very great length of the chains composed
of a much smaller number of different monomers raises obstacles that
render improbable an unambiguous solution by existing methods. More-
over, the available procedures are not yet as refined as in the case of the
proteins : few specific enzymes, no generally applicable method for marking
the end groups. On the other hand, the remarkable difference in the stability

V

fjy, hi.,, .TV-,

.0.^ ^ N

)'\

;a"!ir'%;ii-<^'i4-

"..:

Fig. I . Schematic representation of a fragment of a double strand of a deoxy-
ribonucleic acid. The purines (A, adenine; G, guanine) are depicted in black, the
pyrimidines (C, cytosine; T, thymine) in white.

of the glycosidic bonds holding the purines and the pyrimidines offered a
novel possibility that we have exploited fully.

Our first attempts in the direction of utilizing a specific enzyme were
concerned with the problem of a recognizable repeating unit.

STEPWISE DEGRADATION BY DEOXYRIBONUCLEASE

When crystalline pancreatic deoxyribonuclease acts on a highly poly-
merized preparation of a deoxyribonucleic acid, it is possible to conduct
the experiment in such a manner as to collect the products detached
gradually in a dialyzable form and to separate them from the enzyme-
resistant core [7, 25]. The latter is characteristically different in composition

THE PROBLEM OF NUCLEOTIDE SEQUENCE IN DEOXYRIBONUCLEIC ACIDS 73

from the intact polymer; there is, moreover, a significant trend in the
composition of the products produced by the stepwise enzymic digestion.
No indications of any regularity in the release or the composition of
fragments were observed; no sub-units of recognizably recurrent structure
were seen : deoxyribonucleic acids apparently exhibit a largely arrhythmic
nucleotide sequence.

APURINIC ACID

This has proved a very useful compound; it is, in fact, the first inter-
mediate product in all procedures involving acid degradation [26]. When
a deoxyribonucleic acid is exposed to mildly acidic conditions (pH i-6
at 37^') all the purines are cleaved off gradually, leaving behind a polymer
representing the original polynucleotide, but composed of deoxyribo-
phosphate units, at the places of the previous purine nucleotides, and of
the pvrimidine nucleotides in unchanged ratio and at the same position
as in the native starting material. For instance, in the segment represented
in Fig. I, positions 3, 5-7, 9, i', 2', 4', 8', 10' would remain unchanged,
whereas in the remaining places the purines would now have made room
for the reactive aldehydo groups of the deoxy sugar. The presence of the
aldehydo group [27] and, therefore, of a free hydroxyl at 4' brings about a
remarkable labilization of the polymer: it is broken not only by alkali,
but even by buffers (pH 8-6) containing primary amino groups [28]. It
is likely that this susceptibility of the poly-sugar-phosphate backbone of
apurinic acid to amines is involved in the degradation reactions employing
diphenylamine under acidic conditions [29].

Studies on the arrangement of the pyrimidine nucleotides in apurinic
acid [28] were, in fact, the first that made possible an approach to the
problem of nucleotide sequence in deoxyribonucleic acid. The informa-
tion so obtained was limited, since it was only qualitative, but it showed
that a considerable portion of the pyrimidine nucleotides, and therefore
also of the purine nucleotides, was arranged in the form of tracts of several
nucleotides of one kind.

Sequence studies through differential distribution analysis

MECHANISM

The important discovery that among the fragments produced by the
acid degradation of deoxyribonucleic acids there are found the 3', 5'-
diphosphates of deoxycytidine and thymidine is due to the work of Levene,
Thannhauser and their collaborators [30-33]. A more recent re-investiga-
tion [34] by improved techniques, which confirmed the occurrence of
these diphosphates, prompted a discussion of the possible bearing of this

74 ERWIN CHARGAFF

finding on the problem of nucleic acid structure (p. 366 of ref. [7]). It was
clear that for these fragments to serve as a tool in research on nucleotide
sequence the mechanism of their release had to be better understood. It
was also indispensable to be able to study them under rigorously controlled
conditions and on a quantitative basis.

The kinetics of the liberation of the pyrimidine nucleoside diphos-
phates were first studied, in collaboration with Dr. Shapiro, on a series of
models, namely, the several deoxyribodinucleotides all having cytidylic
acid as one of their components [35]. These investigations served as the
basis of a well-controlled, difi'erential hydrolysis procedure which was
usually performed in three stages [21, 36]. The diagnostically most
valuable results are obtained in Stage I (o-i M H2SO4, 30 min., 100°),
when the release of pyrimidine nucleoside diphosphates can be taken to
reflect directly the abundance of these units as solitary pyrimidine nucleo-
tides in the polynucleotide chain. In the present survey I shall limit myself
to the findings based on this stage of the difi^erential hydrolysis procedure.
The results yielded in Stages II and III, which involve longer periods of
hydrolysis, are mainly indicative of the secondary cleavage of bunched
pyrimidine sequences; they are of great value as further means of differen-
tiation between analytically indistinguishable nucleic acids of different
cellular origin. This will be exemplified in the following section.

In its simplest form the cleavage of a polynucleotide chain and the
formation of the diphosphates of the pyrimidine nucleosides (Py) may be
regarded as a series of ^ elimination reactions [21].

A

Py

CHO

CH

CH,

CH2 0

CH

0 OH

\l
P

CH

CHOH

CH

.0— CH,

0 0-

-CH,

-O OH

\/
P

o o

CHO

I
CH,

CH

CHOH
CH,

O OH

O

o..

After the liberation of the purines the first elimination presumably occurs
at the broken line A. Since no extraneous sugar fragment is found in ester
linkage with the resulting nucleoside diphosphate, a subsequent hydrolysis
or elimination must take place at the other flanking sugar (broken line C).
It is likely that this cleavage follows a ^ aldehyde elimination (broken
line B), which has left a double bond between the second and third carbon
atom of the sugar. Similar considerations apply to the release of bunched
pyrimidine sequences. A more than tentative formulation of the reaction

THE PROBLEM OF NUCLEOTIDE SEQUENCE IN DEOXYRIBONUCLEIC ACIDS 75

mechanism will, however, have to await a better understanding of the fate
of the free deoxyribose residues liberated in the initial stage of the removal
of purines by acid.

NUCLEOTIDE ARRANGEMENT IN ANALYTICALLY INDISTINGUISHABLE DEOXY-
RIBONUCLEIC ACIDS OF DIFFERENT CELLULAR ORIGIN

It happens quite often that deoxyribonucleic acids of taxonomically
entirely different origin exhibit identity of composition as regards the

TABLE I

Differential Distribution Analysis
OF Three Otherwise Indistinguishable Deoxyribonucleic Acid Preparations*

Source: Ox Man Arbacia lixula

DNA fraction: (0-75 m) (i -o m) (2-6 m)

T C [TC] T C [TC] T C [TC]

Total

pyrimidine,

as mole "„ P 299 200 29-9 19-8 30-9 196

Solitary

pyrimidine,

as mole "o of

total constituent

in DNA 15-9 92 19-9 63 160 140

Solitary or
bunched
pyriiTiidines, as
mole % P 4-75 1-^4 I '72 5 '95 1-25. 1-71 4 '94 2-74 i-66

Total T/C,

molar ratio 1-50 1-51 i"58

Solitary T/C,

molar ratio 2-58 4 '76 i ■ So

* The figures are taken from previous papers [21, 36]. The deoxyribonucleic
acid fractions are described by the molarity of NaCl employed for the dissociation
of the histone nucleate [9]. T designates thymine or thymidylic acid, C cytosine
or deoxycytidylic acid, [TC] the two isomeric dinucleoside triphosphates com-
prising cytidine and thymidine (reported as moles of dinucleoside triphosphate
per 100 gm. atoms of nucleic acid phosphorus).

distribution of the nitrogenous constituents. (Compare the example of the
deoxyribonucleic acids of the sheep and the salmon cited previously [i6]).
In such cases, the method of differential distribution analysis discussed in

76

ERWIN CHARGAFF

a

in

U

h

u

U

a

c/2

u

3

U

h

ro

lO

t^

lo

'"'

)-H

M

„

N

vO

fO

N

C^

■^

'^

1— 1

li^

vO

o

t^

ro

h-t

't

N

O

00

N

C>

N

00

ro

N

HH

vO

r^

o

0^

N

«

"i-

00

IT)

N

'-'

Oh

'-0 <i

1) 1^

(U

-oQ

"o

s.s

s

CO ""

C8 -M

CO

. C

«

(U «

^

c 2

^

c

^ -13

•- CO

-3

1 §

1

bo

1

cs ■^

o

h

O!

-T3 0,

^1

ca

■3^
(X!

sD

ro

00

O-

O

oo

lO

o

H-l

t-»

N

00

^

N

H- 1

N

sO

r^

f^

"

►-I

N

O

_o

CS

;_(

CS

;-i

;-!

CS

"o

s

u

u

h

P

u

1

c«

CS

o

"o

h

C/2

3

^3 9

CJ

C

o

'w

"o

ca

3

C

C

iS

^

"H,

<

X

(Li

t-.

*

OJ

-G

O

THE PROBLEM OF NUCLEOTIDE SEQUENCE IN DEOXYRIBONUCLEIC ACIDS 77

this survey is destined to be of great value. It permits the examination of
detailed features of the nucleotide arrangement, far beyond the possibilities
of distinction offered by total constituent analysis. I have selected, for
inclusion in Table I, three examples of nucleic acid fractions that cannot
be distinguished analytically. There can be little doubt about the entirely
different sequential plans governing the structure of these polymers if the
values for the abundance of solitary pyrimidine units are compared.

NUCLEOTIDE ARRANGEMENT IN DIFFERENT DEOXYRIBONUCLEIC ACIDS OF

ANIMAL ORIGIN

Table II compares the distribution findings obtained with total, un-
fractionated preparations of the deoxyribonucleic acids from calf thymus,
human spleen and sperm preparations from two sea urchin species. These
data do not exhaust the information gathered by us on different animal
nucleic acids [21], but are presented here to indicate the scope of the
method.

NUCLEOTIDE ARRANGEMENT IN THE TOTAL DEOXYRIBONUCLEIC ACID OF RYE
GERM AND IN A COMPLETE SERIES OF FRACTIONS

The use of a deoxyribonucleic acid from a plant source offers several
advantages. These nucleic acids ha\ e been shown to contain a fifth nitro-
genous constituent, 5-methylcytosine, in appreciable quantity [7, 20]. In
the specimens isolated from both wheat and rye germ the concentration of
this pyrimidine amounts to 5 -9 mole % (Table III). This means that nearlv
a quarter of the cytosine has been " replaced" by methylcytosine. Since the
molar concentration of guanine equals the sum of these two pyrimidines,
one may consider them as pleromeric in the sense defined above. Minor
constituents of deoxyribonucleic acid (or, for that matter, of ribonucleic
acid) are destined to play an important role in future work on the elucida-
tion of nucleic acid structure, as they will serve as additional markers in
the array of nucleotides constituting the polymer chain. Furthermore, the
presence of a pyrimidine sharing with cytosine the property of being a
6-amino derivative and with thymine that of being a 5-methyl pyrimidine
is of the greatest interest for a better understanding of the mechanisms of
selection, incorporation, and replication that are at work.

Before the study of the nucleotide arrangement in the deoxyribonucleic
acid of rye germ could be undertaken, it was necessarv to examine the
properties of deoxymethylcytidine 3 ',5 '-diphosphate whose production is
indicative of the frequency of this pyrimidine as a solitary nucleotide [39].
At the same time the two dinucleoside triphosphates comprising 5-methyl-
cytidine and either cytidine or thymidine were also studied.

-78 ERWIN CHARGAFF

The investigation of the deoxyribonucleic acid of rye germ [37] offered
us, for the first time, the opportunity of comparing the pyrimidine distri-
bution in the total preparation with that found in a complete series of
seven fractions isolated through the fractional dissociation of the histone
nucleate. These fractions were, in their distribution and composition,
remarkably similar to those previously isolated by the fractionation of the
deoxyribonucleic acid from wheat germ [38]. The availability of such a
complete series of fractions permitted, moreover, a test of the validity of
the results given by the difi^erential distribution analysis; a weighted

TABLE III

Composition of Deoxyribonucleic Acids of Rye Germ and Wheat Germ

Rye germ*

Wheat germf

Moles/ 100-

g.-atoms P

Adenine

27-7

28-1

Guanine

22-6

21-8

Cytosine

i6-4

i6-8

Methylcytosine

59

5-9

Thymine

27-4

27-4

Molar

ratios

Adenine + thymine to guanine +

cytosine + methylcytosine

J -23

1-25

Purines to pyrimidines

I -CI

I -o

Adenine to thymine

I 01

1-03

Guanine to cytosine + methyl-

cytosine

I 01

0-96

Cytosine to methylcytosine

2-78

2-85

* Taken from a previous publication [37].
f Taken from a previous publication [38].

average of the values recorded for the individual fractions could be com-
pared with those given by the total nucleic acid preparation that had not
been subjected to fractionation and showed an excellent agreement [37].

The principal results of this study are, as regards the frequency of
solitary pyrimidine nucleotides, shown schematically in Fig. 2. In this
histogram, we give a brief description of the total pyrimidine composition
of all specimens examined — total deoxyribonucleic acid and seven fractions
— and depict the quantities of solitary nucleotide of each of the three
pyrimidines as per cent of the total constituent in the preparation. 5-
Methylcytosine, it will be seen, has a surprisingly high tendency, when
compared with the other 6-aminopyrimidine, cytosine, to occur as a
solitary unit, similar in this respect to thymine. I shall return to this point
later.

THE PROBLEM OF NUCLEOTIDE SEQUENCE IN DEOXYRIBONUCLEIC ACIDS 79

SOLITARY PY. NUCLEOTIDES IN RYE GERM DNA & FRACTIONS

BASES , MOLE % IN DNA

DIPHOSPHATES, % OF BASES
5 10 15 20 25

TOTAL RGDNA
RGDNA-0.60
RGDNA- 0.65
RGDNA- 0.70
RGDNA- 0,75
RGDNA- 0.80
RGDNA- 0.90
RGDNA- 170

27.4

16 4

5 9

26 3
17.0
6.8

27.1

16.2
6.4

28.2

16.6

6.1

27.8
16.6
5.8

28.0
16.8
6, I

28.6

16.5

5.4

29.2
15 8
4.7

//////////.v ". ///~n

////////.7:r^ZZ77A

7Z^

'/////.^//////////////A

,T

D , C

, M

Fig. 2. Pyrimidine distribution analysis of the total deoxyribonucleic acid of
rye germ and its fractions. The first column indicates the NaCl molarity at which
dissociation of the fraction from its histone salt occurred. The second column
summarizes the composition of the specimens in terms of the molar concentrations
(per 100 gm.-at. of nucleic acid P) of thymine (T), cytosine (C), and 5-methylcy-
tosine (M). The histogram shows the frequency of the solitary nucleotides, isolated
as the nucleoside 3 ',5 '-diphosphates, relative to the total concentration of the con-
stituent in the nucleic acid. Based on previously published data [37].

NUCLEOTIDE ARRANGEMENT IN A MICROBIAL DEOXYRIBONUCLEIC ACID

OF THE GC TYPE

Our survey has so far all been dealing with deoxyribonucleic acids
that belonged to the AT type. It was of interest to see whether the same
considerations could be applied to a nucleic acid of the GC type, i.e.,

8o ERWIN CHARGAFF

TABLE IV
Differential Distribution Analysis of Deoxyribonucleic Acid of BCG*

T

C

[CC]

Total pyrimidine, as mole % P

i6-7

32-3

Solitary pyrimidine, as mole % of total
constituent in DNA

13-6

I2-0

Solitary or bunched pyrimidines, as
mole % P

2-27

3-88

113

Total T/C, molar ratio

0-52

Solitary T/C, molar ratio

0-59

* The figures are taken from a previous paper [40]. [CC] designates dicytidine
triphosphate. Compare Table I for other explanations.

one in which guanine and cytosine preponderated. We first discovered
this type in tubercle bacilli [24] ; but it has since that time been found in
many micro-organisms [7, 41]. The distribution analysis of the deoxy-
ribonucleic acid of an avirulent variant (BGC) of bovine tubercle bacilli
is shown in Table IV. In the nucleic acids of the AT type, as was shown
before, the quantity of thymidine 3', 5 '-diphosphate released during
hydrolysis exceeded that of deoxycytidine 3',5'-diphosphate: this indi-
cated the occurrence of larger quantities of solitary thymine than cytosine ;
but the ratios of the solitary residues were not those that would have been
predicted statistically from the total abundance of these pyrimidines in
the polymers. In the nucleic acid of the GC type an inversion in the rates
of liberation of the nucleoside diphosphates is observed, as expected ; but
in this instance, the ratios of solitary and of total pyrimidines are not
widely different: 0-59 vs. 0-52.

NUCLEOTIDE ARRANGEMENT IN THE DEOXYRIBONUCLEIC ACID OF E. CoU
GROWN IN THE ABSENCE AND PRESENCE OF 5-BROMOURACIL

The deoxyribonucleic acid of E. coli is unusual, being composed of
nearly equimolar proportions of the four nucleotide constituents [7, 41].
No obvious distinctions appear, even when different variants and mutants
are examined, although it is not unlikely that a careful differential distribu-
tion analysis would reveal some differences. Such a study has, however,
not yet been undertaken.

We have as a first approach to the problem of the effect of a mutagenic
agent on the sequence characteristics of a deoxyribonucleic acid, investiga-
ted the nucleic acid produced by a thymine auxotroph of E. coli, grown in

THE PROBLEM OF NUCLEOTIDE SEQUENCE IN DEOXYRIBONUCLEIC ACIDS 8 1

the presence of thymine, and compared it with that formed by the same
organism when suppHed also with 5-bromouracil [42]. This halogenated
pyrimidine is known to be incorporated in what would seem to be the place
of some of the thymine of the nucleic acid [43-45]- The results of our
study were, however, quite surprising. When the results of the total
analysis of these two nucleic acid specimens are compared (Table V), it

TABLE V

Composition of Deoxyribonucleic Acid of E. coli
Grown in the Absence and Presence of 5-Bromour.^cil*

Growth Moles constituent per 100 moles total
supplements

A

G

C

T

B

T

24-8

25-0

25-5

24-8

—

T + B

25-3

24-3

25-1

16-4

9-0

Molar ratios

Pu
Py

A + T + B
G + C

A
T

A

T + B

6 -Am
6-K

0-99 0-98

T + B 0-98 1-03 1-54

* Based on previously published results [42]. The thymine auxotroph was
gro\\*n on a synthetic medium (18 hr., 37 ) supplemented, per ml., with 2 ng.
of thymine or with 2 /ng. of thymine and 10 /tg. of 5-bromouracil. Abbreviations:
A, adenine; G, guanine; C, cytosine; T, thymine; B, 5-bromouracil; Pu,
purines; Py, pyrimidines; 6-Am, 6-amino constituents (A, C); 6-K, 6-keto
constituents (G, T, B).

will indeed be seen that 5-bromouracil and thymine are pleromeric with
respect to the equality of the molar quantities of purines and pyrimidines,
of 6-amino and 6-keto nucleotides, and of 6-aminopurine (adenine) and
6-ketopyrimidines (thymine + 5-bromouracil). It could have been con-
cluded that more than one-third of the places originally occupied by thy-
mine in the polynucleotide now were filled by 5-bromouracil. The finer
structural details revealed by the differential distribution analysis showed
this, however, not to be the case (Table VI). There took place, obviously
under the influence of the halopyrimidine, a severe distortion of all
features of the nucleotide alignment that were accessible to investigation
by our method. I shall return to this problem at the end of this article.

82 ERWIN CHARGAFF

TABLE VI

Differential Distribution Analysis of E. coli DNA

Produced in the Absence and Presence of 5-Bromouracil:

Frequency of Solitary and Coupled Pyrimidine Nucleotide Units*

Growth supplements Thymine Thymine + 5 -bromouracil

B C [CC] [TC] T B C [CC] [TC]

Total pyrimidine, as

mole % P 24-8 — 25-5 16-4 9-0 25-1

Solitary pyrimidine, as
mole % of total con-
stituent in DNA i5"4 — 7-0 18-4 32-9 14-3

Solitary or bunched
pyrimidines, as
mole % P 3 '82 • — I 78 o- 37 I • 13 3 -CI 2-96 3 -60 0-58 I -38

Total solitary
pyrimidines, as

mole % P

5 60

9-57

Total T/C, molar
ratio

0-97

0-65

Solitary T/C, molar
ratio

215

0-84

Total T/B, molar
ratio

—

1-82

Solitary T/B, molar
ratio

—

I -02

Total T + B/B, molar
ratio

—

2-82

Solitary T + B/B,
molar ratio

—

2 -02

* Based on previously published results [42]. Compare Table V for the
composition of the DNA preparations and also Tables I and IV for explanations
of the designations used.

Summarizing remarks

SEQUENCE CHARACTERISTICS COMMON TO ALL DEOXYRIBONUCLEIC ACIDS

Though the first application of the procedure for the ditferential
analysis of pyrimidine distribution served to demonstrate that deoxy-
ribonucleic acids from difi^erent species, even those showing identical

THE PROBLEM OF NUCLEOTIDE SEQUENCE IN DEOXYRIBONUCLEIC ACIDS 83

base composition, were vastly different with regard to the ahgnment of
their nucleotide constituents (Table I), the more important and interesting
uses of this method bear on other aspects of the sequence problem. These
relate, in fact, to the existence of sequential patterns that all deoxyribonu-
cleic acids examined heretofore appear to have in common.

The first consequence of our studies may be the statement that in no
case can the nucleotide sequence be predicted from the knowledge of the
frequency of the individual constituents. The discarded tetranucleotide
hypothesis is a primitive example of an attempt to make such a prediction ;
I have dealt with this a long time ago [i]. This is, incidentally, quite similar
to what one would observe in a meaningful text : a certain letter may be
particularlv frequent; but this does not mean that it will occur, with
commensurate frequency, as a doublet. We shall probably find, in general,
that the symbols constituting a text are not arranged at random, if this
term is understood to refer to an alignment governed only by the relative
frequency of the elements, once the requirement of large numbers is
fulfilled. "

This brings up the problem of the non-randomness of the nucleotide
sequence in deoxyribonucleic acids which we have discussed fully in a
previous paper [21]. Whether strict randomness is conceivable in a living
system capable of orderly replication, at any rate on the level of macro-
molecular events, is a question which it would be out of place to discuss
here. The entire complex of randomness in biology would certainly deserve
a more profound examination than I am equipped to perform. In any
event, the conclusion that in no case investigated the pattern of nucleotide
distribution was that to be expected from a random arrangement of
constituents [21, 35, 37, 40] may contribute to the assurance that in
studying the nucleotide sequence in a deoxyribonucleic acid we are, indeed,
dealing with some sort of meaningful message. If the array of constituents
in a biologically functional polynucleotide has been compared to a giant
throw of dice, we must conclude that it is a throw of loaded dice.

Another consequence of our studies on deoxyribonucleic acids of
animal and plant origin [21, 37] is the conclusion that at least 70*^0 of the
pyrimidines occur as oligonucleotide tracts containing three or more
pyrimidines in a row; and a corresponding statement must, owing to the
equality relationship, apply also to the purines.

POSSIBILITIES OF CLUSTER ANALYSIS

It is quite clear that the various solitary nucleotides and the small
bunched oligonucleotides, such as doublets or triplets, will vary in fre-
quency, i.e. in relative quantity, according to the plan governing the nucleo-
tide sequence of the particular deoxyribonucleic acid; but they are most

84 ERWIN CHARGAFF

unlikely to exhibit uniqueness of quality and will, therefore, not be sufficient
to define a sequence unambiguously. Uniqueness may, however, be
expected, at least occasionally, when very large clusters can be included
in the survey. We have made a beginning with regard to clusters of
pyrimidine nucleotides [46]. When a deoxyribonucleic acid is cleaved by
acid under moderate conditions (o-i M H0SO4, 30 min., 100°) and the
fragments are fractionated on DEAE-cellulose by means of a lithium
chloride gradient, the phosphorylated pyrimidine oligonucleotides can be
separated in order of increasing chain length. Further separation within
the different size groups then is based on two-dimensional chromato-
graphy on filter paper. More than 97" o of the nucleic acid can be accounted
for. Oligonucleotides of more than five pyrimidines in a row are rare, but
occasionally perceptible. Almost all possible combinations of pyrimidine
nucleotide units are encountered, among them di-, tri-, and tetrathymidylic
acids, all carrying an additional terminal phosphate group at 3'. We are at
present engaged in extending the sensitivity of the method by the use of
radioactive labels.

These and related methods, employing specific enzymes, will doubtless
have to be refined considerably, before more definitive answers to the
sequence problem and its biological meaning become possible. It would,
for instance, be of the greatest interest to determine the nature of the
lethal sequence of a parasitic deoxyribonucleic acid that enables it to
impose itself upon, and to paralyze, as it were, the bacterial host.

WHAT IS MEANT BY "REPLACEMENT"?

The platitude that "nobody is irreplaceable" — if it ever holds true —
certainly cannot hold for the deoxvribonucleic acids, if they carry the
biological functions ascribed to them currently. Nevertheless, one often
finds the statement in the literature that an analogue of a purine or pyrimi-
dine, not normally occurring in nature, is, when offered to the cell, in-
corporated into a deoxyribo- or ribonucleic acid "in the place" of a
particular natural constituent of the nucleic acid. Such statements are, of
course, derived from the existence of the complementariness principles [i] ;
i.e. they are based on the fact that certain nucleic acid constituents are
pleromers, to use the nomenclature suggested above. Good examples of
such pleromerism can be seen in Table III with respect to cytosine and 5-
methylcytosine and in Table V with respect to thymine and 5-bromouracil ;
a more comprehensive selection of such instances will be given at the end
of this paper.

What does the assertion that X can be " replaced " in part by Y actually
signify.? (i) It means that Y is incorporated into the deoxyribonucleic
acid chain and that the cell, hence, must possess, or acquire, the enzymic

THE PROBLEM OF NUCLEOTIDE SEQUENCE IN DEOXYRIBONUCLEIC ACIDS 85

apparatus necessary for the production and utilization of the derivative
of Y that is the direct precursor of the polynucleotide chain. (2) If X and Y
share the same functional group (e.g. 6-keto or 6-amino) through which
alone the selection of the members of the newly forming chain takes place
— as foretold bv a mechanism postulating the replication of the comple-
mentarv halves of a double strand [47] — it follows that every X molecule
along the chain has an equal chance of being " replaced" by a Y molecule.
This is obviously not the case in the two systems studied by us previously
[37, 42] and illustrated in Fig. 2 and Table VI. One of these systems is
concerned with the selection of a natural analogue, a 6-aminopyrimidine,
the other with that of a fraudulent analogue, a 6-ketopyrimidine. In both
cases a discrimination, not explainable on the basis of the replication
scheme mentioned above, must have been operating. I shall discuss our
findings in the next two sections.

SELECTION OF A NATURAL ANALOGUE

When we deal with a natural analogue, such as the 5-methyl derivative
of cytosine, it is, of course, meaningless to speak of the replacement of
one by the other. Such a statement acquires meaning only in reference to
the possible mechanism of the biosynthesis of the polynucleotide chain
or to the existence of such principles of pleromerism as those mentioned
before. In other respects, in polymer chains that are not constructed
haphazardlv, each monomeric constituent must have its preordained
place; a place that it must have entered, at the time of polymerization,
through some form of a specific selection mechanism. A replication
mechanism of this nature is, indeed, offered by the scheme [47] postulating
a specific pairing of a 6-aminopurine (adenine) with a 6-ketopyrimidine
(thymine) and of a 6-ketopurine (guanine) with a 6-aminopyrimidine
(cytosine).

This postulate and the fact that 5-methylcytosine is pleromeric with
cytosine could have led to the expectation that these two pyrimidines could
replace each other at random in the deoxvribonucleic acid chain, selection
having regard only to the presence of a 6-amino group in the pvrimidine.
The amazingly constant quantity of methylcytosine incorporated into the
deoxyribonucleic acids of wheat and rye germ (Table III, and compare also
the tabulation given in a previous paper [38]) should, however, render such
a conclusion most questionable. In what manner could a guanine residue
in the preformed chain exercise a preference for pairing with cvtosine,
rather than with methylcytosine, in the newly forming counterpart chain .-
There are surely limits to the amount of foresight that we can attribute
to an ostensiblv automatic process.

Another observation, made repeatedly [S, 37, 38], has to do with the

86 ERWIN CHARGAFF

trend of the distribution of 5-methylcytosine in consecutive fractions
obtained by the fractionation of deoxyribonucleic acid from calf thymus,
wheat germ, and rye germ. In all instances, the relative abundance of
5-methylcytosine with respect to cytosine varied in such a manner that
the percentage contribution of the minor pyrimidine was higher in the
earlier fractions. In no case could one have claimed that the two 6-amino
pyrimidines replaced each other at random within the polynucleotide
chain. I have discussed, in more detail, the significance of the finding of a
disproportionate distribution of cytosine and 5-methylcytosine in view
of current replication hypotheses at a previous occasion [10].

An even more impressive diflference between the ways in which
methylcytosine and cytosine are selected for insertion into the poly-
nucleotide emerges from the study of the frequency of these pyrimidines
as solitary and bunched nucleotides [37]. As will be seen in Fig. 2, a
remarkably large proportion of methylcytosine, and a surprisingly small
proportion of cytosine, exist as solitary units. These two 6-aminopyrimi-
dines can hardly share a common pattern of nucleotide sequence. What is,
however, remarkable is that the mole fractions of 5-methylcytosine and
thymine appearing as solitary units are nearly equal in most instances. One
must conclude that, whereas there is no evidence that the two 6-amino
pyrimidines are treated indiscriminately by the selection mechanism, there
is some indication that no such distinction is exercised in regard to the two
5-methylpyrimidines. In the terms defined above, cytosine and 5-methyl-
cytosine are pleromers, but not homotopes; thymine and 5-methylcytosine
appear to be homotopic with respect to the sequence purine-pyrimidine-
purine, although they are not pleromeric.

SELECTION OF A FRAUDULENT ANALOGUE

In the instance under discussion, the incorporation of 5-bromouracil
into the deoxyribonucleic acid of E. coli, the problem of replacement can
be posed directly. In the case of a natural satellite, such as 5-methylcytosine,
the complement appears to be fixed rigidly by the cell, whereas no obvious
cellular mechanism, except the death of the cell, can be seen that would
limit the extent of incorporation of the halopyrimidine which the cell
presumably cannot but treat as an impostor. If the cellular apparatus is
inveigled into treating the fraudulent 6-ketopyrimidine as if it were
thymine, does it also insert it as such indiscriminately into the newly
forming deoxyribonucleic acid ? This does not appear to be the case. A
glance at Table VI will show that the deoxyribonucleic acids made by the
cell in the absence and in the presence of bromouracil are vastly diflFerent
with regard to their sequence characteristics. Not only is there no sign of
an equal relative utilization of thymine and bromouracil for common

THE PROBLEM OF NUCLEOTIDE SEQUENCE IN DEOXYRIBONUCLEIC ACIDS 87

sequences, such as purine-thymine-purine and purine-bromouracil-
purine, but the intrusion of the halopyrimidine seems to have brought
about so drastic a reorganization of the nucleotide pattern as to imply the
existence of two distinct populations of polymers. We may be able to fool
the cell, but it is no longer the same cell.

In the cells that were cultivated in the presence of the halopyrimidine,
the nucleic acid contains nearly twice as many solitary pyrimidine residues
as does the normal preparation; the frequency of the sequences purine-
cytosine-purine has doubled; and, relative to the total concentration of
thymine, even the runs of purine-thymine-purine have increased. In
addition — and this is quite remarkable — almost one-third of the in-
corporated bromouracil is found flanked by purines in the form of solitary
units. The abundance of the coupled unit purine-cytosine-purine has,
however, been relatively little changed.

At the same time, as Table V shows, 5-bromouracil and thymine
definitely are treated as pleromers.

THE CONCEPTS OF PLEROMERISM AND HOMOTOPY AND THE NEIGHBOUR

PROBLEM

The definitions of the suggested terms "pleromer" and "homotope"
have been given above. The compilation in Table VII, by no means
complete, gives a few examples of such pleromeric relationships in de-
oxyribonucleic acids. There can be little doubt that in all these instances
the established pairing relationships would have been destroyed, had the
minor pleromer been omitted from the computations. Only one satellite
normally occurring in nature, viz. 5-methylcytosine, could be included in
the Table. Another minor component — if it can be classified as a normal
constituent — namely, N-methyladenine, plays a baffling role. It has been
reported to occur, among other instances, in the deoxyribonucleic acid of
a thymine auxotroph of E. coli in not inconsiderable quantity [49]. Under
special conditions of growth, its concentration increases, somewhat in
measure with a decrease in the amount of thymine. It is, however, at the
moment difficult to see through what mechanism a 6-ketopyrimidine and
a 6-methylaminopurine could function as pleromers.

Pleromic relationships in ribonucleic acids are much more difficult to
establish, as the only complementariness useful for scrutiny is the equality
of the 6-amino and 6-ketonucleotides [19]. Of the three examples listed
in Table VIII only the one dealing with the incorporation of 5-ffuorouracil
(52) carries conviction. It is not impossible that, when more careful
analyses become available, 8-azaguanine and guanine will also emerge as
pleromers.

The usefulness of the concept of homotopy is, for the time being, more

ERWIN CHARGAFF

TABLE VII
Examples of Pleromeric Relationships in Deoxyribonucleic Acids

Pleromers

Molar ratios

Source

of
DNA

Analogue or

natural

satellite

(mole %)

Normal
pvrimidine
(mole %)

Comple-

mentariness

ratio*

6 Am/ Refer-
6Kt ence

Wheat 5 -Methyl- 5-9 Cytosinei6-8
germ cytosine

096

1-03

38

Rye

germ

5 -Methyl- 5-9
cytosine

Cytosine 16-4

01

I 00

37

E. coll

5-Chloro- 14-2
uracil

Thymine 7-9

•o6t

I -00!

48

E. colt

S-Bromo- 9-0
uracil

Thymine 16-4

•00

I -oi

42

E. colt

5-Iodo- 4- I
uracil

Thymine 19-6

■or

099

48

Phage
T2r +

5-Bromo- 16 • i
uracil

Thymine 15-9

•01

0-97

48

* In the case of 6-aminopyrimidines (cytosine, 5-methylcytosine), this is the
molar ratio of guanine to the sum of 6-aminopyrimidines; in the case of the
halopyrimidines, the complementariness ratio is the molar ratio of adenine to the
sum of 6-keto pyrimidines.

t Molar ratio of 6-amino nucleotides to 6-keto nucleotides.

% N-Methyladenine (1-4 mole "^^o) omitted from the calculations.

evident with respect to the amino acid sequence in proteins than to the
nucleotide arrangement in nucleic acids. More and more instances of such
homotopic relationships are being revealed as knowledge on specific
sequences in proteins increases. As concerns the deoxyribonucleic acids,
we are largely limited to the consideration of simple sequences, such as
purine-pyrimidine-purine, or of small bunches of pyrimidine nucleotides.
The development of the study of larger clusters, mentioned before, will
doubtless contribute much to our knowledge. I have emphasized above a
possible instance of homotopy, namely, the remarkable similarity in distri-
bution, as solitary nucleotides, of the two 5-methyl pyrimidines (thymine
and 5-methylcytosine) in rye germ deoxyribonucleic acid [37]. Examples
of the absence of homotopic relationships, where they would have been
expected, are more numerous: cytosine and 5-methylcytosine, e.g. in rye

THE PROBLEM OF NUCLEOTIDE SEQUENCE IN DEOXYRIBONUCLEIC ACIDS 89

TABLE VIII

Examples of Pleromeric Relationships in Ribonucleic Acids

Pleromers

Source of . ,

RNA ' 1 ' 11V Normal pyrimidine 6-Am/6-K* Reference

or natural satellite
(mole %)

nucleotide (mole ",,)

Rat liver,
soluble
RNA

Pseudouridylic 3-3
acid

Uridylic
acid

17

•4

0-95

50

E. coli,
soluble
RNA

Pseudouridylic 2 • i
acid

Uridylic
acid

15

0

o-98t

51

E. coli,
total
RNA

5-Fluoro- 9-8
uridylic
acid

Uridylic
acid

II

4

0-94

52

* Ratio of the molar sum of adenylic and cytidylic acids to that of guanylic,
uridylic and pseudo- or 5-fluorouridylic acids.

t Ribothymidylic acid (11 mole %) included among the 6-keto nucleotides.

germ nucleic acid [37] ; thymine and 5-bromouracil in the nucleic acid of
E. coli [42]. It goes without saying that, at the present state of our know-
ledge, such statements as the presence or absence of homotopy must be
taken as valid only in the most general statistical terms. Specific homotopic
replacements may, of course, have taken place even where no all-embracing
positional replacement can be affirmed.

One could conclude provisionally that in the deoxyribonucleic acids
pleromers are not necessarily homotopes, and homotopes not necessarily
pleromers. Before the law of pleromerism all bases are equal as long as
they pair otT according to functional groups. But the rules of segregation
applied to them, once they have gone through the door, still are obscure.
These rules are, I am afraid, much less simple and predictable than would
please the advocates of biological automation.

This brings me to the last topic of the present survey, namely, the
neighbour problem. It was, I believe, first stated in the following terms [10].
"The frequency of methylcytosine being linked to guanine in a dinucleo-
tide as compared with that of cytosine being so attached is twenty times as
high in calf thymus nucleic acid, and thirteen times as high in the wheat
germ preparation, than would have been expected for non-selective in-
corporation. The unexpectedly high proportion of 5-methyldeoxycytidyIic
acid that is linked to deoxyguanylic acid [53, 54] makes one, in fact,
wonder whether it is not the adjoining nucleotide rather than the one

(JO ERWIN CHARGAFF

opposite that directs the incorporation." This remarkable phenomenon —
the high tendency of methylcytosine to be next to guanine— was again
touched upon in our study on the sequence characteristics of the de-
oxyribonucleic acid of rye germ [37]. Several analogous observations also
were discussed there of which one may be mentioned. When, in this
nucleic acid, the coupled dipyrimidine units are surveyed, it is found that
methylcytosine has a greater tendency to be linked to either thymine or
cytosine than the latter has to associate with itself or with thymine ; and
there is a definite bias in favour of methylcytosine-cytosine units. In our
recent investigation of the bromouracil-containing nucleic acid of E. colt
[42] a more general formulation of the neighbour problem was attempted :
"One gains the impression that, after the selection of equal numbers of
6-aminopurines and 6-ketopyrimidines, on one hand, and of 6-keto
purines and 6-aminopyrimidines on the other, a second mechanism — an
exclusion principle, as it were — comes into play, so that certain neighbours
are tolerated and others not, or rarely."

Somewhat similar observations have also been made in the case of the
ribonucleic acids in which the analysis of neighbouring tendencies is
easier because of the facility of obtaining any given nucleoside as either
the 3' or the 5' phosphate. An examination of this type has, for instance,
been made with ribonucleic acid preparations from different cellular
portions of rat liver [55]. A particularly instructive example of the neigh-
bour principle is offered by the "soluble" ribonucleic acids of the cyto-
plasm concerned with the specific transfer of amino acids [56], in which the
terminal sequence adenine-cytosine-cytosine is required ; a sequence that
can apparently be enforced by a specific enzymic mechanism.

The manner in which a growing deoxypolynucleotide chain could
guarantee the quality of its continuation is not yet a topic encouraging
speculation.* Our present ideas about the enzymes taking part in the bio-
synthesis of specific polymers will probably have to undergo a stringent
revision before the directed synthesis of complicated sequences will
begin to be understood. For this, more than a biological form of Scotch
tape is required. Certain sequential restrictions could, incidentally, be
explained if, under circumstances, runs of more than one nucleotide were
the immediate precursors of the polymer chain. We have, for instance,
discussed the evidence, in rye germ nucleic acid, of the two dinucleotides
cytosine-guanine and methylcytosine-guanine being able to substitute
each other at random [37].

In considering the problem of the nucleotide sequence in deoxyribonu-
cleic acids we have barely turned the corner. There is a long road before
us ; and we shall not see its end.

* "Whereof one cannot speak, thereof one must be silent," Wittgenstein,
Tractatus Logico-Philosophicus.

THE PROBLEM OF NUCLEOTIDE SEQUENCE IN DEOXYRIBONUCLEIC ACIDS 9I

Acknowledgments

The experimental work forming the basis of this essay has had the
generous support of several agencies, especially, the National Institutes of
Health, United States Public Health Service ; the National Science Founda-
tion; The American Cancer Society; and the Rockefeller Foundation. I
wish to thank my colleagues whose names appear in the Bibliography ; an
especial debt of gratitude is due to Dr. H. S. Shapiro without whose work
much of what is presented here could not have been written.

References

1. Chargaff, E., Experieiitia 6, 201 (1950).

2. ChargafF, E., in "The Origin of Life on the Earth" (LU.B. Symposivim
Series, \'ol. i). Pergamon Press, London, 297 (1959).

3. ChargafF, E., in "Transactions, 4th International Congress of Biochemistry",
Vol. XIV. Pergamon Press, London, 21 (1959).

4. Edsall, J. T., in "The Proteins", Vol. I, ed. H. Xevirath and K. Bailey.
Academic Press, New York, 549 (1953).

5. Sadron, C. L., in "The Nucleic Acids", Vol. Ill, ed. V.. ChargafF and J. N.
Davidson. Academic Press, New York, i (i960).

6. Montague, M. D., and Morton, R. K., Nature, Land. 187, 916 (i960).

7. ChargafF, E., in "The Nucleic Acids", Vol. I, ed. E. ChargafF and J. N.
Davidson. Academic Press, New York, 307 (1955).

8. ChargafF, E., Crampton, C. F., and Lipshitz, R., Xature, Land. 172, 289 (1953).

9. Crampton, C. F., Lipshitz, R., and ChargafF, E.,y. biol. Chem. 211, 125 (1954).

10. ChargafF, E., /«" Fibrous Proteins and their Biological -Significance " (Symp,
Soc. Exp. Biol.) 9, 32 (1955)-

11. Watson, J. D., and Crick, F. H. C, Nature, Loud. 171, 737 (1953).

12. Cavalieri, L. F., Rosenberg, B. H., and Deutsch, J. F., Biochem. Biophys. Res.
Comm. I, 124 (1959).

13. Bernardi, G., Champagne, M., and Sadron, C, Xature. Lond. 188, 228 (i960).

14. Crampton, C. F., Lipshitz, R., and Chargaff, E.,y. biol. Chem. 206, 499 (1954).

15. Chargaff, E., Schulman, H. M., and Shapiro, H. S., Xature, Lond. 180, 851

(1957)-

16. Chargaff, E., "The Harvev Lectures", Series =52. Academic Press, New York,

57 (1958).

17. Ryle, A. P., Sanger, F., Smith, L. P"., and Kitai, R., Biochem. jf. 60, 541 (1955).

18. Brown, H., Sanger, F., and Kitai, R., Biochem. J. 60, 556 (1955).

19. Elson, D., and Chargaff, E., Biochim. biophys. Acta 17, 367 (1955).

20. Wyatt, G. R., Biochem. J. 48, 584 (1951).

21. Shapiro, H. S., and Chargaff, E., Biocliitn. biophys. Acta 26, 608 (1957).

22. Chargaff, E., Cold Spr. Harb. Symp. quant. Biol. 12, 28 (1947).

23. Chargaff, E., Vischer, E., Doniger, R., Green, C, and Misani, F., jf. biol.
Chem. 177, 405 (i949)-

24. Vischer, E., Zamenhof, S., and Chargaff, E.,jf. biol. Chem. 177, 429 (1949).

25. Hodes, M. E., and Chargaff, E., Biochim. biophys. Acta 22, 361 (1956).

26. Tamm, C, Hodes, M. E., and Chargaff, E.,y. biol. Chem. 195, 49 (1952).

27. Tamm, C, and Chargaff, E.,jf. biol. Chem. 203, 689 (1953).

28. Tamm, C, Shapiro, H. S., Lipshitz, R., and Chargaff, E.,_7. biol. Chem. 203,
673 (1953)-

92 ERWIN CHARGAFF

29. Burton, K., and Petersen, G. B., Biochem.J. 75, 17 (i960).

30. Levene, P. A., and Jacobs, W. A.,^ biol. Chem. 12, 411 (1912).

31. Thannhauser, S. J., and Ottenstein, B., Z. physiol. Chem. II4, 39 (1921).

32. Levene, P. A., J. biol. Chem. 48, 119 (1921).

33. Thannhauser, S. J., and Blanco, G., Z. physiol. Chem. 161, 116 (1926).

34. Dekker, C. A., Michelson, A. M., and Todd, A. R., J. chem. Soc. 947 (1953).

35. Shapiro, H. S., and Chargaff, E., Biochim. biophys. Acta 26, 596 (1957).

36. Shapiro, H. S., and Chargaff, E., Biochim. biophys. Acta 23, 451 (1957).

37. Shapiro, H. S., and Chargaff, E., Biochim. biophys. Acta 39, 68 (i960).

38. Lipshitz, R., and Chargaff, E., Biochim. biophys. Acta 19, 256 (1956).

39. Shapiro, H. S., and Chargaff, E., Biochim. biophys. Acta 39, 62 (i960).

40. Tsumita, T., and Chargaff, E., Biochim. biophys. Acta 29, 568 (1958).

41. Belozersky, A. N., and Spirin, A. S., in "The Nucleic Acids", Vol. Ill, ed.
E. Chargaff and J. N. Davidson. Academic Press, New York, 147 (i960).

42. Shapiro, H. S., and Chargaff, E., Nature, Loud. 188, 62 (i960).

43. Weygand, F., Wacker, A., and Dellweg, H., Z. Naturf. 7b, 19 (1952).

44. Dunn, D. B., and Smith, J. D., Nature, Lond. 174, 305 (1954).

45. Zamenhof, S, and Griboff, G., Nature, Lond. 174, 306 (1954).

46. Spencer, J. H., and Chargaff, E., Fed. Proc. 20, 353 (1961).

47. Watson, J. D., and Crick, F. H. C, Nature, Lond. 171, 964 (1953).

48. Dunn, D. B., and Smith, J. D., Biochem.J. 67, 494 (1957).

49. Dunn, D. B., and Smith, J. D., Biochem.J. 68, 627 (1958).

50. Lipshitz, R., and Chargaff, E., Biochim. biophys. Acta 42, 544 (i960).

51. Dunn, D. B., Smith, J. D., and Spahr, P. ¥.,J. mol. Biol. 2, 113 (i960).

52. Horowitz, J., and Chargaff, E., Nature, Lond. 184, 1213 (1959).

53. Smith, J. D., and Markham, R., Nature, Lond. 170, 120 (1952).

54. Sinsheimer, R. L,.,J. biol. Chem. 208, 445 (1954).

55. Shigeura, H. T., and Chargaff, E.,J. biol. Chem. 233, 197 (1958).

56. Hoagland, M. B., m "The Nucleic Acids", Vol. Ill, ed. E. Chargaff and J. N.
Davidson. Academic Press, New York, 349 (i960).

Discussion

Davis: I don't quite understand what you mean in saying that methylcytosine
has to replace thymine. Are there not solitary pyrimidine places where the methyl-
cytosine may have replaced cytosine ?

Chargaff : I have to refer you to recent papers with Dr. Shapiro in which we
discussed at great length the problem of the solitary methylcytidylic units versus
thymidylic and cytidylic units. What I had tried to point out is that you can look at
the formation of such a chain — whether it is double or single is immaterial in this
case — as being a two-stage process. First, through a sieve as it were, an equal
number of 6-amino and 6-keto compounds are selected. These guarantee the pair-
ing principles which I have mentioned before, but what goes next to each other
seems to be controlled by what I have called, for want of a better term, an exclusion
principle. For instance, you can say that in the case of bromouridylic acid in E. coli
much more occurs as solitary units between two purines than would have been
predicted. You only have to compare the total molar sums of solitary thymidylic
acid and of solitary thymidylic plus bromouridylic acids. In other words, there is a
principle operating which says that bromouracil likes to be between two purines.
This intruder has been segregated in the chain entirely differently from the way in

THE PROBLEM OF NUCLEOTIDE SEQUENCE IN DEOXYRIBONUCLEIC ACIDS 93

which the thymine is arranged; and a similar thing happens with cytosine versus
methylcytosine. In the latter case, one cannot speak of "replacement", because
you don't replace one natural compound by another, but must infer that they
occupy very different places in the chain.

Dische: These changed sequences of pyrimidines and purines which are
brought about by bromouracil appear to be very difficult to reconcile with the
function of the nucleotide sequences as a coding system. If you get a complete
redistribution of the solitary pyrimidines then you get also a complete change in the
sequence of the purines. It is difficult to see how this organism can have any
similarity to its parents if you make such a complete change in the sequence of
practically all molecules of the nucleic acids. I think this is really a most remarkable
phenomenon. The second question I should like to ask is whether the pairing is
found also in the one strand DNA ?

Chargaff : To answer your first question : on the one hand you have a hypo-
thesis and on the other you have some meagre findings. I don't know whether I
should be called upon to reconcile the experimental findings with a hypothesis. It
should be the other way around. The fact is that in normal E. coli DNA i5"o of
the thymine is solitary. This is the only 6-ketopyrimidine present in DXA. E. coli
treated with bromouracil gives i8"o of the thymine as solitary plus 33*^0 of the
bromouracil which has been incorporated as solitary unit. In other words, whereas
in the first case 1 5*^*0 of the ketopyrimidines are flanked by purines, almost 27",, are
so arranged in the second case. You are perfectly right that if there is any informa-
tion coded into the DNA, looking at the figures you would believe that we are
dealing with two entirely different organisms. I leave it to you to draw your own
conclusion. I don't want to speculate too much at this early stage. It could, of
course, be that not all the DNA is functional in a strict sense. Those fractions that
really carry information may be left relatively unattacked. In answer to your
second question, there is no obvious pairing discernible in Sinsheimer's analysis.
In any event, none of the regularities which one encounters in the other DNAs
is found in this.

Problems in Polynucleotide Biosynthesis

J. N, Davidson

Department of Biochemistry, University of
Glasgow, Scotland

One of the most exciting recent developments in biochemistry has been
the now classical work on the biosynthesis of deoxyribonucleic acid (DNA)*
by Romberg and his colleagues (for summary see ref. [i]) who showed that
extracts of Eschericliia coli contain kinases which bring about the phos-
phorylation of deoxribonucleotides to the triphosphate stage and a poly-
merase which acts on the triphosphates in the presence of a suitable DNA
primer with the production of deoxyribopolynucleotide macromolecules
built to the design of the primer molecule.

Such enzymes are also present in mammalian tissues. They have been
the subject of extensive investigations by Potter and his colleagues in
Wisconsin [2, 3, 4] and Canellakis in Yale [5, 6, 7] using regenerating rat
liver, and by our own group in Glasgow [8, 9, 10, 11, 12, 13, 14, 15] using
extracts of Ehrlich ascites carcinoma cells. Similar systems are present in
calf thymus [16] and mouse leukaemic cells [17]. The following account is
based largely on the work of a group of investigators including Dr. R. M. S.
Smellie, Dr. H. M. Keir, Dr. E. D. Gray, Dr. S. M. Weissman, Dr. J.
Richards, Dr. J. Paul and Mr. D. Bell.

The process of DNA biosynthesis can be divided into three separate
stages :

(i) The biosvnthesis of the purine and pyrimidine nucleotides and

their conversion to the deoxyribonucleoside monophosphates,
(ii) The phosphorvlation of the deoxyribonucleoside monophosphates
to the corresponding triphosphates by appropriate kinases.

* The following abbreviations are used in this paper : DNA for deoxyribonucleic
acid; dAMP, dGMP, dCMP, dUMP, UMP, OMP, IMP, and TMP for the
s'-monophosphates of deoxyadenosine, deoxyguanosine, deoxycytidine, deoxyuri-
dine, uridine, orotidine, inosine and thymidine respectively; dATP, dADP,
dGTP, dGDP, dCTP, dCDP, TTP and TDP for the s'-tri- and diphosphates
respectively of deoxyadenosine, deoxyguanosine, deoxycytidine and thymidine ;
TdR for thymidine ; ATP for adenosine-5 '-triphosphate ; DPN for diphosphopyri-
dine nucleotide; tris for 2-amino-2-hydroxymethyl propane- 1 : 3-diol; TdR kinase,
TMP kinase and TDP kinase denote respectively the enzyme systems responsible
for the formation of TMP, TDP and TTP.

96 J. N. DAVIDSON

(iii) The polymerization of the triphosphates in the presence of DNA
primer to yield new deoxyribopolynucleotide under the influence
of the polymerase enzyme.

The enzymes involved in these reactions are freely soluble in water or
bufi^er solutions and are present in extracts obtained by centrifuging dis-
rupted cells at high speeds to remove particulate material. The overall
reaction involved in the combination of the second and third steps men-
tioned above can readily be demonstrated by incubating such extracts
with labelled thymidine (TdR) and following its incorporation into DNA
(Fig. i). Thymidine has several advantages for this purpose: it is a specific
precursor of DNA and it can readily be obtained commercially, labelled
with tritium at high levels of activity. For these reasons it has been ex-
tensively employed in autoradiographic studies on DNA biosynthesis,

TdR

thymidine Jcjn<3s<z

TMP^-dUMP—UMP

TMP k/n<3s<z
TDR

TDP Ji/nasiz
TTP dATP dGTP dCTP

po/ym<zrsis<z

DNA

Fig. I. Pathway of incorporation of thymidine into DNA.

especially as tritium is a label peculiarly suitable for autoradiography (for
bibliography see ref. [18]). But thymidine has the further advantage, as
will be mentioned later, that the enzymes involved in its phosphorylation
to its triphosphate (TTP) increase and decrease according to the state of
proliferation of the tissue whereas those involved in the phosphorylation
of the deoxyribonucleoside monophosphates of adenine, guanine and
cytosine show a much smaller, if any, variation [13, 14].

Against these advantages it might be argued that thymidine is not a
normal metabolite on the anabolic pathway to TTP in vivo since the
formation of thymine derivatives proceeds at the nucleotide level by way
of UMP and dUMP to TMP (Fig. i), but this is not a serious disadvantage
since the kinase involved in the phosphorylation of thymidine to TMP
appears to run parallel with the kinases phosphorylating TMP to TTP [13].

The incorporation of pH]-TdR into DNA by the enzyme systems in
cell-free extracts of several tissues is illustrated in Table L As might be

PROBLEMS IN POLYNUCLEOTIDE BIOSYNTHESIS 97

TABLE I
The Incorporation of [^H]-TdR into DNA by Soluble Enzymes Prepared

FROM SONICALLY DISRUPTED RaBBIT TISSUES [8]

The reaction mixture contained o-i M phosphate buffer pH 8-i, 50 /imoles
glucose /ml., 2-5 /xmoles ATP/ml., 2-5 /imoles DPN/ml., 4/x moles MgCU/ml.,
500 /xg.DXA/ml., 20 /Ltmoles NaCl/ml., and 2 /iC [^H]-TdR/ml. Incubation
time : 2 hr.

„. DiS'A specific activity

^^^"^ (c.p.m.//xmole DiXA-P)

Rabbit bone marrow 24 750

Rabbit thymus 9 275

Rabbit appendix i 485

Rabbit liver 131

Rabbit kidney 71

expected, the lowest values are found for the non-proliferating tissues such
as liver and kidney and the highest values in tissues such as bone marrow
in which cell division is very active. Extracts of such rapidly prolifering
tissues may be used to illustrate the net synthesis of DNA. Table II shows

TABLE II

S-ynthesis of Polynucleotide by Polymerase Present in Extracts of Ehrlich
Ascites Tumour Cells [io]

The reaction mixture contained 100 /^imoles tris buffer pH 7-9, 0-25 /tmole
DPN, 5 ^moles MgCI.j, 25 /tg. ascites DNA and 07 ml. ascites cell extract in a
total vol. of I ml. Incubation time: 4 hr. Figures in parenthesis represent the
number of determinations.

Additions

Total DNA/tube

Nil 26-2 (4)

0-3 /^imole each of dATP, dCjTP,

dCTP and TTP 47-7 (4)

net synthesis by the polymerase present in a cell-free extract of Ehrlich
ascites carcinoma cells incubated with the four deoxyribonucleoside tri-
phosphates in the presence of primer DNA, magnesium ions and DPN.

The activities of the kinases responsible for the formation of TTP are
shown in Fig. 2. Again the highest activity is found in extracts of those
tissues showing the most active cell proliferation. The activity in liver
tissue is very low indeed but when cell proliferation is induced by partial
hepatectomy a rise in the ability of extracts of the resulting regenerating
liver tissue to form TTP from TdR is apparent within 24 hr, and is very

98 J. N. DAVIDSON

pronounced 48 hr. after operation [3, 4, 6, 7, 11, 13, 19]. The polymerase
activity is also increased. This behaviour of the thymidylate kinase system
is in striking contrast to that of the kinases phosphorylating dAMP,
dGMP and dCMP which are very little elevated in regenerating, as
compared with normal, liver [3, 14].

A detailed study of the three kinases in regenerating rat liver at various
times after partial hepatectomy shows that the appearance of the enzymes
tends to be sequential, TdR and TMP kinases appearing before TDP
kinase [13].

Fig. 2. Per cent conversion of thymidine to TTP by kinases present in
extracts of various rabbit tissues.

By 24 hr. the TdR and TMP kinases have begun to increase sharply,
reaching maximum activity after about 30 hr. The activity of TDP kinase
develops more slowly, reaching its peak at about 48 hr., which is also the
time of maximum incorporation of pH-]TdR into DNA in the overall
reaction. Subsequently all kinases decline in activity.

A similar sequential pattern of kinase appearance is also found in
cultures of the L strain of mouse subcutaneous fibroblasts [11, 13]. Such
cells in the resting phase after exhaustion of the medium show either no
kinase activity or only TdR kinase activity. On inoculation into fresh
medium a period of rapid growth occurs after a lag phase. The TdR
kinase appears early, rises sharply and remains elevated. The TMP kinase

PROBLEMS IN POLYNUCLEOTIDE BIOSYNTHESIS 99

activity rises later but before the TDP kinase, and both enzymes reach
their peak of activity during the early part of the growth phase and decline
before growth (increase in cell number) has stopped. This pattern can be
greatly modified by adding thymidine to the cultures without affecting the
growth rates. While there is little effect on the enzymes during the early
period of growth, the TdR kinase activity remains elevated for a much
longer period in the test cultures and TAIP and TDP kinases rise to a
second peak during the period when these activities in the control cultures
decline [ii, 13]. These elevations on the kinase activities can be produced
only during the growth phase and are apparently due to true enzyme
induction.

Although extracts of normal liver tissue show very little activity in
synthesizing DNA they nevertheless contain an active polymerase which
can be separated by ammonium sulphate fractionation [14]. The low

TABLE III

Effect of Liver Extract on Partly Purified Polymerase and Kinases from
Ascites Tumour Extracts [14]

Polymerase Kinases

mfx moles TTP^^ incorporated . /x/x moles ester formed
into DNA in 2 hr. per mg. in i • 5 hr. per mg.
protein protein

TMP TDP TTP

Ascites enzymes 4-7 430 830 2130
Ascites enzymes

+ liver extract 04 14 12 o

activity of whole extracts of liver suggests that they might contain factors
which interfere with polymerase or kinase action and this has indeed
been found to be the case [12, 14]. Addition of extract of normal rat
liver to partly purified polymerase and kinases for the thymidine system
prepared from extracts of Ehrlich ascites cells causes a very pronounced
drop in activity (Table III). Similarly the addition of extract of normal
liver to extract of regenerating liver greatly reduces the polymerase and
kinase activity in the latter. The activity of these enzymes in bone marrow
is likewise reduced by extracts of normal liver.

The presence of factors interfering with polymerase and kinase activity
is not confined to liver tissue. Factors inhibiting polymerase are pro-
nounced in liver, kidney and serum but not in brain or muscle. The
phosphorylation of thymidine is strongly inhibited by liver extracts but
not by extracts of other tissues so far tested while the formation of TDP
and TTP is inhibited by liver and kidney extracts, by serum and to a

lOO J. N. DAVIDSON

much less pronounced extent by muscle extracts. The polymerase used
in most of these tests has been purified from Ehrlich ascites tumour cells
but polymerases from spleen thymus and bone marrow are likewise affected.
The interfering factors are thermolabile, acid-precipitable and non-
dialysable but so far have not been identified with any known enzyme or

TABLE IV

Inhibition of Enzymes by Fractions Prepared from Extracts of Normal

Rat Liver

Inhibition of

TdR TMP ^ ,

Fraction kinase kinase Polymerase

A + + +

B ± + —

C — + +

D — — +

enzymes. They appear to be different from the enzymes which hydrolyse
the deoxyribonucleoside triphosphates with the release of inorganic
phosphate. The distribution of the interfering substances in various
tissues suggests that different factors are involved in the inhibition of the
various enzymes and this has been confirmed by fractionation of liver

AMP -■ ♦dAMP— ^dATP

Glycine 1

Glutamine — ^IMP
Asp. etc. I

GMP -.-dGMP ^dG

Aspartate Carbamyl

X phosphate
/ CTP--^CDP— ►dCDP-j^dCTP

Ureidosuccinate UTP dCMP

i f /

Orotate-^OMP— »-UMP---»dUMP — *-TMP— :^TTP

Fig. 3. Pathways involved in the biosynthesis of DNA.

extract with purification and separation from each other of the factors
inhibiting TdR kinase, TMP kinase and the polymerase [15] (Table IV).
Kinase inhibitions so far discussed have referred to the system pro-
ducing TTP. The pattern of results obtained with the kinases for dAMP,
dGMP and dCMP is quite different in that these enzymes in ascites cell
extracts are not inhibited by extracts of normal liver. The kinase system
for producing TTP would accordingly appear to be unique both in its

PROBLEMS IN POLYNUCLEOTIDE BIOSYNTHESIS lOI

tendency to vary with the state of proliferation of the tissue and in its
susceptibiUty to the action of the interfering factors present in extracts of
normal resting liver. It must therefore exercise a profound effect in the
regulation of DXA biosynthesis. Such a conclusion would be in agreement
with the well-known views of Potter [20, 21].

The overall picture of DNA formation is illustrated in Fig. 3. In
addition to feedback mechanisms controlling nucleotide biosynthesis
[20, 21] at least two main mechanisms seem to control polynucleotide
formation. One is the conversion of cytidylate to deoxycytidylate which
appears from the recent work ot Reichard, Canellakis and Canellakis [22]
to occur at the diphosphate level and to be controlled by the triphosphate
level in the cell; the second is the system involved in the formation of
TTP from TMP. It may well be that a third method of control lies in the
presence in certain tissues of inhibitors for the polymerase.

References

1. Kornberg, A., Science 131, 1503 (i960).

2. Bollum, F. J., and Potter \'. R. J. biol. Chem. 233, 478 (1958).

3. Bollum, F. J., and Potter, V. R., Cancer Res. 19, 561 (1959).

4. Bollum, F. J., Anderegg, J. \V., McElya, A. B., and Potter, \\ R., Cancer Res.
20, 138 (i960).

5. Canellakis, E. S., and Mantsavinos, R., Biochim. biophys. Acta 27, 643 (1958)

6. Mantsavinos, R., and Canellakis, E. S.,^. biol. Chem. 234, 628 (1959).

7. Canellakis, E. S., Jaffe, J. J., Mantsavinos, R., and Krakow, J. S., J. biol.
Chem. 234, 2096 (1959)-

8. Smellie, R. M. S., Keir, H. M., and Davidson, J. N., Biochim. biophys. Acta

35» 389 (1959)-

9. Keir, H. M., and Smellie, R. M. S., Biochim. biophys. Acta 35, 405 (1959).

10. Smellie, R. M. S., Gray, E. D., Keir, H. M., Richards, J., Bell, D., and
Davidson, J. N., Biochim. biophys. Acta 37, 243 (i960).

11. Weissman, S. M., Paul, J., Thomson, R. Y., Smellie, R. M. S., and Davidson,
J. X., Biochem.y. 76, iP (i960).

12. Weissman, S. AI., Gray, E. D., Thomson, R. Y., Smellie, R. M. S., and
Davidson, J. N., Biochem. J. 76, 26P (i960).

13. Weissman, S. M., Smellie, R. M. S., and Paul, J. (i960), Biochim. biophys.
Acta 45, loi.

14. Gray, E. D., Weissman, S. M., Richards, J., Bell, D., Keir, H. M., Smellie,
R. M. S., and Davidson, J. N. (i960), Biochim. biophys. Acta 45, iii.

15. Weissman, S. M., et al. (unpublished results).

16. Bollum, F. J., Fed. Proc. 18, 194 (1959).

17. Mantsavinos, R., and Canellakis, E. S., Cancer Res. 19, 1239 (1959).

18. Lima-de-Faria, A., Hereditas, Lund. 45, 632 (1959).

19. Hiatt, H. H., and Bojarski, T. B., Biochem. biophys. Res. Comm. 2, 35 (i960).

20. Potter, V. R. Univ. Mich. med. Bidl. 23, 401 (1957).

21. Potter, V. R., and Auerbach, V. H., Lab. Investigation 8, 495 (1959).

22. Reichard, P., Canellakis, Z. N., and Canellakis, E. S., Biochim. biophys. Acta
41, 558 (i960).

102 J. N. DAVIDSON

Discussion

DiscHE : You indicated in a slide that the deoxycytidine kinases did not increase
in regenerating liver but that the thymidine kinase strongly increased; is that
because the original activity of the deoxycytidine kinase in the liver was so much
higher than that of the thymidine kinase ?

Davidson : Yes, the activity of deoxycytidine kinase in the liver is much higher
than that of thymidine kinase and the same holds for the kinases for deoxyAMP,
and deoxyGMP; and this may well be the reason why they do not increase because
they are abundant already.

Reichard : I should like to ask you two questions with regard to the inhibition
of the syntheses ; one is : have you tried to fortify the system by adding an ATP-
regenerating system which might overcome this inhibition ? The second question
is : does the factor which inhibits the polymerase have any deoxyribonuclease
activity ?

Davidson: We have investigated both of these points. We haven't added an
ATP-regenerating system, we have infused ATP into the system without effect.
We have looked into the question of deoxyribonuclease activity and as far as we
can see this has nothing to do with the action of the inhibitory factor.

SiEKEViTZ : I should like to ask Prof. Davidson a couple of questions. Another
possibility of regulation of DNA synthesis derives from the possibility that DNA
itself may be a cofactor for the enzyme. Could it be that the availability of the right
kind of DNA which acts as a primer could act as a regulator of DNA synthesis ?
In the resting cell DNA could not act as a primer for the enzyme but at a certain
point something happens, it can act as a primer, and when finished it "rejuvenates"
itself until it is no longer active.

Davidson : Offhand I am not very convinced about it, but I should point out
that conditions in the resting cell are very different from those in a cell-free
system such as ours.

Reichard : If I understood Dr. Siekevitz right he implies that a specific DNA
must be present in the system in order to start off DNA synthesis. Romberg's
data would argue strongly against this, since one can use any DNA as a primer for
a polymerase from E. coli.

Siekevitz: Is that single strand or double strand DNA ?

Reichard: That's a different question again, which is not cleared up. Even in
our crude preparations, extracts of chick embryos, we get very little DNA synthesis
when we add highly polymerized double-stranded calf thymus DNA. As soon as
we heat this preparation for lo min., this treatment is supposed to split the double
strands, the DNA will act as a primer.

Siekevitz: But that might be just the point. There comes a time in the life of
the cell when single strand DNA is formed to act as a primer, and this splitting of
the RNA might be the point where regulation could occur.

Reichard : I think that is quite true and that it might be an important regu-
lating mechanism in living cells.

Davidson : I think I should perhaps mention that in the experiments which
I described we use heated DNA as our primer but, of course, this is a cell-free
system.

Enzymic Formation of Deoxyribonucleic Acid from
Ribonucleotides

Peter Reichard

Department of Chemistry /, Karolinska histitutet,
Stockholm, Sweden

The work of Romberg and his associates [i] has estabhshed that
DNA* is formed enzymically through the polymerization of deoxvribonu-
cleoside triphosphates. It is thus evident that cells which carry out a
rapid synthesis of DNA require a constant supply of these compounds,
and the biological source of the deoxyribonucleotides becomes of major
importance.

Earlier /;/ vivo experiments [2, 3] made it very likelv that deoxvribosvl
compounds in many different types of cells are formed through a direct
reduction of the corresponding ribosyl compounds. More recently it was
possible to demonstrate such conversions with enzvmes from chick embrvos
and Escherichia co/i, and it was then found that this reduction took place
at the diphosphate level, e.g. deoxycytidine diphosphate was formed from
cytidine diphosphate [4, 5].

Here I will describe experiments in which the enzymes catalyzing the
reduction of ribonucleotides are coupled with the enzymes synthesizing
DNA from deoxyribonucleotides. Thus in these experiments we studied
the over-all reaction sequence leading from ribonucleotides to the syn-
thesis of DNA. This work was carried out during a recent visit to the
Department of Pharmacology, Yale University, New Haven, U.S.A., in
collaboration with Drs. Z. N. and E. S. Canellakis.

One major purpose of our investigation was to study possible regulatory
mechanisms during in vitro DNA synthesis. The reduction of a ribotide
to the corresponding deoxyribotide represents the first step in a reaction
sequence which ultimately results in the synthesis of DNA. All the pre-
ceding enzyme reactions leading to the formation of the ribotide can be

* Abbreviations used: DX.A, deoxyribonucleic acid; CMP, CDP and CTP,
mono-, di- and triphosphates of cytidine, respectively; GMP, GTP, mono- and
triphosphates of guanosine, respectively ; ATP, adenosine triphosphate ; TTP,
th\Tnidine triphosphate ; deoxycompounds are denoted by the prefix d-.

I04 PETER REICHARD

visualized to have at least a threefold purpose: (i) The synthesis of RNA;
(2) the synthesis of coenzymes; and (3) the synthesis of DNA:

CO2, NH3, glucose, etc.

I

Ribonucleotides

RNA Coenzymes Deoxyribonucleotides — >DNA

In view of these considerations it seemed an attractive hypothesis to
consider the reductive step as a possible " pace maker " for DNA-synthesis,
and it was hoped that the construction of an /// vitro system carrying out
the synthesis of DNA from ribonucleotides might serve as a model system,
in which the influence of different factors on this hypothetical pace maker
might be studied. Ample evidence exists that such a first step in a reaction
sequence may indeed act as a rate-limiting step [6] and that it can be
influenced and regulated by the products of later reactions.

In all experiments to be described here the source of enzyme was a
high-speed supernatant fraction from a homogenate of 5-day-old chick
embryos. Both cytidine and guanosine ribonucleotides were used as sub-
strates for DNA formation.

First it was necessary to establish that our enzyme preparation could
synthesize DNA from labelled deoxynucleotides. It was found that in-
corporation of radioactivity from labelled dCMP or dGMP into DNA
required the presence of ATP, Mg + + and "primer" DNA. This in-
corporation was further stimulated by the addition of a complementary set
of the other deoxynucleoside triphosphates. These results demonstrate
that the chick embryo extract contained enzymes which catalyze the
formation of DNA from deoxyribonucleotides by a mechanism similar to
that described earlier for other systems [i, 7, 8].

Next the formation of radioactive DNA from labelled ribonucleotides
(CMP and GMP) was studied. It was found that this process again
required the addition of ATP, Mg + + and "primer" DNA. When either
^-P-labelled or tritium( = base)-labelled CMP was used as substrate, it was
found that identical amounts of isotope were incorporated into DNA
(Fig. i). This experiment demonstrates that the intact nucleotide was
used for DNA synthesis and that the incorporation of isotope did not
occur as a result of, e.g., transglycosylation.

It was then necessary to investigate in which manner the ribonucleo-
tide had been used for DNA synthesis. With techniques used earlier by
Adler et al. [9] it was possible to demonstrate that in our experiments the
isotopic ribonucleotide was first reduced to the deoxyribonucleotide, and

ENZYMK- FORMATION OF DEOXYRIBONUCLEIC ACID

105

that the labelled deoxyribonucleotide was subsequently used for DNA
synthesis. In the experiment described in Fig. 2 DNA was synthesized

Fig. I. Time curve of DX.A formation from tritium or ''-P-labelled CMP, as
measured by the incorporation of isotope into DNA [7, 8]. Incubation conditions
for one time point: o • i ^tmole of labelled CMP, i o /^^mole of ATP, 5 -o /^imoles of
MgCl,, o-i mg. of heated DNA (10 min. at 100 ) and 4-5 mg. of "enz\Tne",
final volume 0-41 ml., pH 7 -4, 37^.

CMP ►DNA

t ^ t ' ^

Micrococcal DNA-se | Pancreatic DNA-se

+ spleen diesterase ,+ venom diesterase

\-®

Md

Total counts in

CMP

'

500

dCMP

1900

7800

dAMP

3400

0

dCMP

2700

0

IMP

3900

0

Fig. 2. Enzymic degradation [9] of DNA formed from P*--CMP. 60 min.
incubation with conditions as in Fig. i.

from ^'-P-CMP and, after isolation, degraded enzymically in two different
ways :

(o) With pancreatic DNA-se + snake venom diesterase. These enzymes
split DNA as indicated by the arrow in Fig. 2 and produce 5'-deoxy-
ribonucleotides. Since 5 '-labelled CMP was the precursor in the synthesis

io6

PETER REICHARD

of DNA, the isotope stays attached to the nucleotide used as substrate for
the DNA polymerase.

(b) With micrococcal DNA-se + spleen diesterase. This results in
the formation of 3 '-nucleotides. ^^P is no longer attached to the nucleotide
used as substrate, but is transferred to the neighbour nucleotide.

The results of Fig. 2 obtained by degradation according to (a) show
that CMP was used for DNA synthesis after transformation to dCMP,
since about 95° o of the isotope resides with this deoxyribonucleotide.
After degradation according to {b) the isotope is distributed among all
four deoxyribonucleotides. This in turn demonstrates that labelled dCMP
(formed from CMP) in DNA was linked to all four deoxynucleotides in
internucleotide linkage.

Fig. 3. Stimulation of DNA formation from dCMP by an ATP-regenerating
system. Conditions as in Fig. i (ooii fxmole of dCMP substituted for CMP).
Where indicated 4 ■ 5 /xmoles of creatine phosphate + o • i mg. of creatine kinase
were added.

The results obtained so far strongly indicate that the chick embryo
enzyme preparation carried out the synthesis of DNA from ribonucleo-
tides according to the following general reaction sequence :

Ribonucleotide- — ^Deoxyribonucleotide >DNA

It was now possible to study some of the factors which might influence
this overall pathway of DNA synthesis.

The first question studied was how the process was influenced by the
ATP-level in the system. It might be expected — and has been shown
already for other similar systems — that a high level of triphosphates, as is
obtained by the addition of an ATP-regenerating system, is favourable for
the synthesis of DNA from a deoxyribonucleotide. Figure 3 demonstrates
that the addition of creatine phosphate + kinase stimulated isotope in-
corporation from labelled dCMP into DNA, and that the chick embryo
system thus conformed to expectation.

However, DNA formation from a ribonucleotide, such as CMP, was
decreased when an ATP regenerating system was added (Fig. 4). The

ENZYMIC FORMATION OF DEOXYRIBONL'CLEIC ACID

107

inhibition was located in the reductive step, as shown by the results of
Fig. 5. Our interpretation of these results is the following: the addition
of ATP + Mg ^ ^ to this crude enzyme system resulted in the phosphoryla-
tion of CAIP to CDF and CTP. As found by paper chromatography an

Fig. 4. Inhibition of DXA formation from CMP by an ATP-regenerating
system. Conditions as in Fig. i. Where indicated 4-5 /tmoles of creatine phosphate
+ o-i mg. of creatine kinase were added.

equilibrium between mono-, di- and triphosphates was rapidly established
and CDP was the predominating nucleotide. However, when the ATP-
regenerating system was added, this resulted in a very efficient phosphoryla-
tion of CMP to CTP, and little CDP was left in the system. Since CDP is

Fig. 5. Inhibition of CMP-reduction by an ATP-regenerating system.
Conditions as in Figs, i and 4. The formation of dCMP + deoxycytidine was
measured as described earlier [4].

the substrate for the reduction, the addition of the ATP-regenerating
system greatly decreased ribotide reduction. According to this interpreta-
tion our experiments indicate that in vitro the synthesis of DNA from
ribonucleotides is dependent on the maintenance of a critical level of

io8

PETER REICHARD

diphosphates and is inhibited by an excess of ATP, which increases the
level of triphosphates but decreases the diphosphate level.

As mentioned earlier, the synthesis of DNA from one labelled deoxy-
nucleotide was stimulated by the addition of the three other non-labelled

MxlO^

Fig. 6. Influence of an equimolar mixture of TTP, dATP and dGTP on
DNA formation from dCMP and CMP, respectively.

lUU

1

— 1 1

M

w^

t>

.c

c

>v

f

X TTP

<

r

• dGTP

Q

\

o dATP

^ SO

A

*\

h _

o

o^*

is

s«

•

20
MxIQS

30

Fig. 7. Comparison of different deoxynucleoside triphosphates as inhibitors of
DNA formation from CMP.

deoxynucleoside triphosphate. This is demonstrated for [^^C]-dCMP by
the upper curve of Fig. 6. With P^C]-CMP, however, isotope incorporation
into DNA was strongly inhibited by the addition of an equimolar mixture
of the triphosphates of thymidine, deoxyadenosine and deoxyguanosine

ENZYMIC FORMATION OF DEOXYRIBONUCLEIC ACID

109

(Fig. 6 lower curve). Furthermore, it was found (Fig. 7) that each single
deoxvnucleoside triphosphate inhibited DXA formation from CMP to
about the same extent.

It seemed obvious that here again the reductive step was the point of
inhibition, and we therefore studied the influence of the diflFerent deoxy-
nucleoside triphosphates on the formation of dCMP from CMP, A strong
inhibition bv dATP and dGTP was observed, TTP inhibited much less,
while dCTP showed almost no effect (Fig. 8). Addition of the purine
deoxyribonucleotides thus resulted in about a 50^0 inhibition at an initial
concentration of io~^ M. The initial concentrations of ATP and CMP
in these experiments were 2-5 x lO"'^ M and 0-4 x 10^ m, respectively.

Similar experiments were also carried out with guanine nucleotides.
It was found that DNA svnthesis from dGMP was stimulated more than

MxlO^

Fig. 8. Effects of deoxvnucleoside triphosphates on the reduction of CMP.

twofold by an equimolar mixture of dCTP, dATP and TTP. The effects
of such a mixture on the incorporation of isotope from GMP into DXA
were quite small and inconsistent. When each trip)hosphate was added
alone, different types of effects were observed (Fig. 9). Thus DNA
formation from GMP was strongly inhibited by the addition of dATP,
but was stimulated by the two pyrimidine deoxynucleoside triphosphates.
Similar divergent effects were observed when the formation of dGMP
from G^MP was investigated (Fig. 10). In this case the effect of dGTP
could also be investigated and it was found that this deoxyribonucleotide
acted as a strong inhibitor. Thus the results for the GMP^-dGMP trans-
formation showed that purine deoxyribonucleotides were inhibitors while
pyrimidine deoxyribonucleotides, stimulated the reaction.

We believe that our results show that in an in vitro system with soluble
enzvmes the reduction of both a purine and a pyrimidine ribonucleotide
was regulated to a large extent by the levels of deoxynucleotides present

no

PETER REICHARD

in the system. I should hke to stress the point that the reduction of ribo-
nucleotides was controlled not only by the product-deoxynucleotide of the

Fig. 9. Effects of deoxynucleoside triphosphates on DNA formation from
GMP. Conditions as in Fig. i with labelled GMP in place of CMP; individual
deoxynvicleoside triphosphates added as indicated on the abscissa.

Mxio^

Fig. 10. Effects of deoxynucleoside triphosphates on the reduction of GMP.
The conditions of incubation and of the assay were as described in an earlier
paper [5].

respective reactions, but also by the other deoxynucleotides ; e.g. dATP
and dGTP greatly inhibited the reduction of CMP, and the pyrimidine
deoxyribonucleotides stimulated the reduction of GMP.

ENZYMIC FORMATION OF DEOXYRIBONUCLEIC ACID III

Our results were obtained with soluble enzymes and it remains to be
shown whether thev have any bearing on the problem of DXA synthesis in
living cells. It is not vet possible to decide this point with any degree of
certainty, but it seems to be relevant that recent experiments by Klenow
[lo] and by Morris and Fischer [ii] have shown that effects similar to
those I have discussed here can be obtained with living cells. It is thus not
inconceivable that such effects are parts of an important homeostatic
mechanism for DNA synthesis.

References

1. Romberg, A., in "The Chemical Basis of Heredity", ed. W. D. McElroy and
B. Glass, Johns Hopkins Press, Baltimore, 579 (i957)-

2. Hammarsten, E., Reichard, P., and Saluste, E.,_7. biol. Chem. 183, 105 (1950).

3. Rose, I. A., and Schweigert, B. S.,J. biol. Chem. 202, 635 (1953).

4. Reichard, P., and Rutberg, L., Biochim. biophys. Acta 37, 554 (i959)-

5. Reichard, P., Biochim. biophys. Acta 41, 368 (i960).

6. Pardee, A. B., in "The Enzymes I", ed. P. D. Boyer, H. Lardy, and K.
Myrback. Academic Press Inc., New York, 681 (1959).

7. Bollum, F. ].,jf. Amer. chem. Soc. 80, 1766 (1958).

8. Mantsavinos, R., and Canellakis, E. S., j. biol. Chem. 234, 628 (1959).

9. .\dler, J., Lehman, L R., Bessman, 'SI. J., Simms, E. S., and Romberg, A.,
Proc. nat. Acad. Sci., Wash. 44, 641 (1958).

10. RIenow, H., Biochim. biophys. Acta 35, 412 (1959).

11. Morris, N. R., and Fischer, G. A., Biochim. biophys. Acta 42, 183 (i960).

Discussion

Chargaff : I was wondering whether Dr. Reichard has any explanation of the
peculiar fact that usually analogues are incorporated only into one of the two types
of nucleic acid. For instance, fluorouracil when given to E. coli will appear only as
ribofluorouridylic acid in RXA and not a trace seems to go as the deoxy compound
into DNA; similarly in the plant nucleic acids which I have mentioned in my talk
high amounts of 5-methylcytosine are found in the DNA only. I was wondering
whether you had any explanation for this.

Reichard : In the case of uracil — and I think this also applies to fluorouracil,
because from all the work which has been published now it appears that fluorouracil
behaves with enzymes as does uracil — the explanation put forward by Romberg
is that there are no kinases for deoxy-UMP; this would also be true for fluoro-
deoxy-UMP. This is a reasonable explanation, but on the other hand I am a little
worried about the fact that our own findings indicate that the reduction might take
place at the diphosphate level. I don't know whether it is really LTDP which is
reduced to deoxy-UDP although I believe so. What one might look for would be a
special phosphatase like the one which has been found in T2 infected E. coli
for deoxy-CTP, but in this case it would be specific for deoxy-UTP. That could be
an explanation.

Davidson : Is there evidence that the reduction step takes place at the diphos-
phate level in other cases than deoxy-CDP ?

112 PETER REICHARD

Reichard : In the case of G — yes, and in the case of U — no, and for A we have
no evidence.

Hess : What is the reducing system involved or have you any experience about
the type of system you would expect ? I ask because a reducing system can also act
as a controlling mechanism upon the synthesis of DXPP from XDP. Certainly, the
reducing equivalents in growing cells are available in high concentrations, i.e.
speaking for the TPNH level, which in a given control range could well be a
critical metabolite.

Reichard : I think you might be quite right but I do not know very much about
the events during the reduction. We know that it is a complicated series of reactions
and not a single reaction. We also know that we can demonstrate a stimulation of
the dialysed or ion exchange-treated enzyme preparation by TPNH. Now you
could draw the tentative conclusion that TPNH is the source of hydrogen during
the reduction, but on the other hand it might be an indirect effect. I don't know
yet but we are just starting to purify the enzymes.

Studies on the Mechanism of Synthesis of
Soluble Ribonucleic Acid*

E, S. CANELLAKIsf AND EdWARD HERBERT

Departmeni of Pharmacology, Yale University

Medical School, New Haven, Conn., U.S.A.

and

Department of Biology, Massachusetts

Institute of Technology, Cambridge, Mass. U.S.A.

Our studies have been concerned with the mechanism of synthesis of
the soluble RNA (S-RNA) present in the soluble cytoplasmic fraction of
rat liver. In this fraction of rat liver, enzymes concerned with the in-
corporation of ribonucleotides into terminal positions of S-RNA have also
been found. Within the allotted space I shall attempt to present a summary
of our studies on the synthesis of soluble-RNA, rather than present in
detail any one particular aspect of the problem.

I. Fractionation of ribonucleotide
incorporating enzyme and S-RNA

By a series of fractionation techniques outlined in Table I and Figs, i
and 2, we have been able to achieve an approximately roc 200 fold
purification of three protein and S-RNA components from rat liver.
These three fractions have been termed S-RNA-proteins a-, ^- and -y
based on the order of their elution off the hydroxylapatite columns. Each
of these fractions will incorporate ribonucleotides into the corresponding
S-RNA in terminal positions. We have been further able to separate the
protein from the S-RNA by the use of diethylaminoethylcellulose (DEAE)
(Fig. 3). Thus separated, neither the RNA nor the protein by itself will
incorporate ribonucleotides into an acid insoluble form; full activity is
restored upon recombination of the protein and the S-RNA.

* Detailed experimental evidence for the material published in this summary
may be found in the following articles which have been published or are in press.
Canellakis, E. S., and Herbert, E., Proc. nat. Acad. Set., Wash. 46, 170 (i960);
Herbert, E., and Canellakis, E. S., Biochim. b/ophys. Acta 42, 363 (i960) ; Canellakis,
E. S., and Herbert, E. (S-RNA H, IV, V), Biochim. biophys. Acta 45, 133 (iq6o),
47, 78 (1961), 47, 85 (1961).

t Senior Research Scholar of the U.S.P.H.S. wishes to thank this Service for
financial aid associated with this trip to Stockholm, Sweden.
I

114

E. S. CANELLAKIS AND EDWARD HERBERT

TABLE I
Summary of Purification Procedure

Specific

mg.

Protein

activity*

Yield

RNA

mg

. RNA

% of soluble

cytoplasm

mg

Soluble cytoplasm

5-IO

" 100"

135

100

Ammonium sulphate

(o -55-0 -86)

30-50

80

70

60

pH 5 precipitate

100-150

60

35

15

Hydroxylapatite

Ribonucleoprotein a

80-100

4

3

I

/3

750-1000

15-20

15-20

I

y

400-600

10-15

10-15

I

DEAE-cellulose

"protein "

750

50

—

100

"RNA"

—

—

8-10

0-05

* The specific activity unit is expressed as /x//moles of [^*C]-AMP incorporated
into S-RNA in the presence of i mg. protein in 6 min. at 25°.

Buffer cone

Volu

280 ml.

Fig. I. Stepwise procedure for elution of ribonucleoproteins from an hydroxy-
apatite column. The width of the steps in the chromatogram represents the volume
of eluate collected in each fraction. These volumes are as follows in the order of
potassium phosphate buffer pH 7 ■ 2 concentrations listed under the chromatogram
(0-05 M, o-o8 M, and 0-12 m); 20 ml., 10 ml., and 25 ml. Column dimensions
2 '5 cm. X 20 cm.

STUDIES ON MECHANISM OF SYNTHESIS OF SOLUBLE RIBONUCLEIC ACID II5

Buffer

Volume (ml ) Conc-
Mixer 300 OOIM

Reservoir

300 0-25 M

• Ribonucleoproteins

2-5M buffer
(HEME color)

20 30 40 50 60 70
Tube number

Fig. 2. Linear gradient procedure for elution of ribonucleoproteins from an
hydroxylapatite column with potassium phosphate buffer pH 7-2. The volume of
eluate collected per tube is i o ml. The 2 • 5 M buffer is added directly to the column
and collected in 10 ml. portions. Column dimensions 2-5 cm. x 20 cm.

2 1

1-2
06

200
100
08

3
04

0

■^

"RNA"
component

"Protein"
component

Buffer cone Eff
NaCI cone.
Volume (ml ) lO

0 OIM
0
15

T"

OOSM

0

10

007SM
0
15

II25M
0
15

02M

02M

10

J r

0 2M j

0 5M

12

Fig. 3. Stepwise elution and separation of protein and RNA component on
a DEIAE column, with potassium phosphate buffer pH 72 and sodium chloride.
Column dimensions i -5 cm. x 2 cm.

We have termed these S-RNA-protein fractions "ribonucleoproteins"
for the following reasons: (a) the protein and the S-RNA fractionate
together during an extensive purification process, (b) the protein and the

ii6

E. S. CANELLAKIS AND EDWARD HERBERT

S-RNA separate into three well-defined columns and (c) each component
contains its own RNA and ribonucleotide incorporating enzyme. We
believe that the exact type of physical relationship that exists between
these fractions will be best elucidated by detailed enzymatic studies which

S-RNA/3
n = 40

60%

H.J^

Nl

S-RNAy
n=34

237o

24%

•N.

N.

NJ
L. _in

51%

16%

IS, -4

N

14%,

6 X U

N.

\.

< 10%

n = (7AMP+ 6UMP+ I3GMP+ l3CMP-t- l\//UMP)

16AMP+5UMP + II GMP+ II CMP+li|;UMP) = n

Fig. 4. Schematic presentation of the variety of S-RNA molecules present
in S-RNA-/S and S-RNA-y. The column of figures on the left and right correspond
to the ribonucleotide content and per cent composition of S-RNA-/3 and S-RNA-y
respectively.

are at present under way. Because of the experimental simplicity involved
in isolating the ^- and y- fractions, our work has been largely limited to a
study of the properties of these two fractions.

2. Analytical data on S-RNA-^ and -7

Analytical studies on S-RNA-^ as well as S-RNA-y have shown that
they can be grossly distinguished as four molecular species differentiated
by their end-groups (Fig. 4). All four molecular species of both S-RNA-/S
and -y start with guanylic acid, that is, upon alkaline hydrolysis guanosine

STUDIES ON MECHANISM OF SYNTHESIS OF SOLUBLE RIBONUCLEIC ACID II J

diphosphate is the only diphosphate which can be found. S-RXA-^ is
rich in the molecular species which terminates in adenylic acid (yielding
adenosine upon alkaline hydrolysis) and has no detectable S-RNA species
terminating in uridylic acid. S-RXA-y is rich in the molecular species
which terminates in guanylic acid and has some of the molecular species
which terminates in uridyHc acid.

The ribonucleotide analyses of these two families of S-RNA indicate
that approximately one pseudouridylic acid molecule corresponds to one
chain (Table II). In addition, the adenylic acid content of S-RNA-^S

TABLE II
Composition of S-RNA-^ and S-RNA-y

Ribonucleosides *

Component

Adenosine

Uridine Guanosine

^ ... Guanosine
Lytidine ... , , *
diphosphate*

S-RNA-^
S-RNA

o-
o

6g
23

— 0-24

GIG 0-51

O- IG
G-I4

I 02
0-95

2'(3')-Ribonucleotides*

AMP

UMP 0-UMP GMP CMP

Total of

nucleot.

per chain*

Wt. of

chain

S-RNA-^

S-RNA-y

7-2
6-G

6

5

0 i-oi 13-0 13-2

0 I-2I II-I IIG

42

36

15 000

13 8go

* Molar composition calculated by assuming that one terminal ribonucleoside
residue corresponds to one polynucleotide chain. Chromatographic analysis made
on the alkaline hydrolysates of S-RNA-^ and of S-RNA-y.

t Including one of the nucleoside end and one guanosine diphosphate.

and of S-RNA-y approximates the sum of uridylic acid plus pseudouridylic
acid whereas the guanylic acid content approximates that of cytidylic
acid. From the end-group analysis (assuming one end-group per chain), a
molecular weight of approximately 15 000 can be derived.

3. The effect of pyrophosphate

Incubation of the enzyme with S-RNA-/3 plus -y, [^^C]-CTP, in-
organic pyrophosphate and Mg ' ^, results in an enhanced incorporation
of P^C]-CMP into the S-RNA (Table III). This effect can also be ob-
served if the enzyme and the S-RNA are preincubated with inorganic
pyrophosphate, the S-RNA then extracted and reincubated with a fresh
enzyme preparation and [^^C]-CTP in the absence of pyrophosphate. It

E. S. CANELLAKIS AND EDWARD HERBERT

TABLE III

The Effect of Pyrophosphate
ON THE Incorporation of [1'*C]-CMP into S-RNA

Treatment

['^C]-CMP incorporated
(c.p.m.)

1. None 150

2. Pyrophosphate included

in incubation mixture 600

3. S-RNA pretreated with

pyrophosphate* 610

2 -o Eoen units of S-RNA-^ and -y were used throughout this experiment. The
first tube contained [^*C]-CTP (30 m/umoles, 2 x lo^/c.p.m./jumole), S-RNA,
Mg + + 2 -o /Ltatoms, and ribonucleotide incorporating enzyme in o -08 m potassium
phosphate buffer. Final volume i o ml. The second tube contained in addition
to above, i o fimole pyrophosphate. The S-RNA used in the third tube had been
pretreated with the ribonucleotide-incorporating enzyme, Mg + + 2-0 /xatoms,
I -o /xmoles pyrophosphate in 008 M potassium phosphate buffer. Final volume
I -o ml. This was then isolated free of the ribonucleotide incorporating enzyme,
and re-incubated in an incubation mixture identical to that in the first tube.

* Inorganic pyrophosphate does not enhance the incorporation of [^■*C]-CTP
into S-RNA if during the preincubation either the S-RNA or the ribonucleotide-
incorporating enzyme is omitted.

TABLE IV

Liberation of Ribonucleoside 5'-Triphosphates into the
Acid-Soluble Fraction by the Pyrophosphorolysis of S-RNA

Radioactivity recovered in the
ribonucleoside 5'-triphosphates
(c.p.m.)

ATP CTP UTP GTP

Experiment i 7000 2600 1350 850

Experiment 2 11 60 310 330 180

Experiment 2A 6600 950 2070 1320

In Experiment i, S-RNA-/3 and -y was pyrophosphorylyzed in the presence of
[^'-P]-inorganic pyrophosphate, the ribonucleotide-incorporating enzyme, 2-o
/Liatoms Mg + + per ml. and 20 mjumoles each of non-radioactive ATP, CTP, UTP
and GTP in 008 m potassium phosphate, pH 7-2. Final volume i -o ml. In
Experiments 2 and 2A, [^-P]-S-RNA (prepared by isolating S-RNA from rat
liver which had been labelled with ^'-P in vivo) was incubated with i -o /xmole of
non-radioactive inorganic pyrophosphate, 2-0 /xatoms Mg + +, 20 m/nmoles each
of non-radioactive ATP, CTP, UTP and GTP, in the absence (Experiment 2)
and in the presence (Experiment 2A) of added non-radioactive ribonucleotide-
incorporating enzyme. The small but definite liberation of ribonucleoside s'-tri-
phosphate in Experiment 2 may be due to contamination of the S-RNA preparation
with the ribonucleotide-incorporating enzyme.

STUDIES ON MECHANISM OF SYNTHESIS OF SOLUBLE RIBONUCLEIC ACID II 9

may therefore be concluded that the inorganic pyrophosphate exerts its
effect on the S-RXA; in order that this effect be ehcited the concomitant
presence of the enzvme is required but not that of the CTP.

The effect of inorganic pyrophosphate has been elucidated as being
one of pvrophosphorolysis of the S-RXA (Table IV) because incubation
of the enzyme with S-RNA and p-P]-pyrophosphate results in the
liberation of ^-P-labelled ribonucleoside triphosphates into the medium.

800

700 X-

600

500

400

300 -

100

"C- AMP-S-RNA

C-CMP-S-RNA

1.0

0.5

2.0 3.0 4.0 5.0

Mg /I MOLES PER ML

6.0

3.0

7.0

3.5

8.0

4.0

.0 1.5 2.0 2.5

P- P^ MOLES PER ML

Fig. 5. Pvrophosphorolysis of the terminal ribonucleotides of S-RNA. Two
portions of S-RXA-/3 and -y containing 30 F.^eo units each were labelled separately
with [C^*]-AMP and with [C'^]-CMP respectively. They were then extracted free
of the nbonucleotide-incorporating enz\Tne, and Mg + + and pyrophosphate in the
indicated amounts under standard conditions. At the end of the incubation period
the residual radioactivity on each type of labelled S-RNA was determined.

If the converse experiment is performed, that is if ^'-P-labelled S-RNA
(obtained after in vivo labelling with ^-P), is incubated with non-radio-
active pyrophosphate and enzyme, again ^-P-labelled ribonucleoside tri-
phosphates are liberated into the medium.

Pyrophosphorolysis seems to be directed primarily towards the end of
the S-RNA molecule (Fig. 5). This can be shown by using S-RNA which
has been labelled in vitro with either [^^C]-ATP or with [^^CJ-CTP which
are known to be incorporated in the end of the S-RXA molecule. When

I20 E. S. CANELLAKIS AND EDWARD HERBERT

such a radioactive S-RNA preparation is added to an incubation medium
containing the ribonucleotide-incorporating enzymes, and varying con-
centrations of Mg + + and pyrophosphate, a progressive loss of radioactivity
occurs, indicating the removal of the terminal ribonucleotides in the S-RNA
molecule.

Since pyrophosphate results in the degradation of the end of the S-RNA
and in an increased incorporation of cytidylic acid, it is reasonable to expect
that new sites in the S-RNA may be exposed. If these new sites accept
cytidylic acid, then incubation of the pyrophosphorylyzed S-RNA with
p2p]-CTP (32P-P-P) followed by alkaline hydrolysis of the CM^sp-
labelled S-RNA should result in the identification of the site of attachment
of the cytidylic acid. We have performed this experiment under conditions
of low and maximal pyrophosphorolysis of S-RNA, and compared the
results with those obtained with non-pyrophosphorylyzed S-RNA. Pyro-
phosphorolysis results in the unveiling of adenylic acid as the new site of
attachment of p-P]-CMP. The distribution pattern of the ^^P obtained
with maximally pyrophosphorylyzed S-RNA (Table 5) shows the ^^P to
be equivalently distributed between 2' (3')-cytidylic acid and 2' (3')-
adenylic acid. These results make the following interpretation plausible:

P-P + AMP— CMP— CMP— AMP— S-RNA ->

AMP— S-RNA + ATP + 2CTP

32P.CTP + AMP— S-RNA ->

32P-CMP—32p.CMP— AMP— S-RNA + P-P

In other words, pyrophosphorolysis of S-RNA has exposed a terminal
adenylic acid which now accepts two cytidylic acid residues. This S-RNA
upon alkaline hydrolysis would yield equivalent amounts of P^pj.cytidylic
acid and adenylic acid. Ti is would therefore indicate adenylic acid to be
the fourth ribonucleotide in the terminal sequence. Although this inter-
pretation is in keeping with the observed facts it can in no way be con-
sidered to be the only possible interpretation. We therefore wish to suggest
this as a temporary explanation, reserving final judgment until a more
homogeneous preparation of S-RNA is available.

This interpretation of these results raises the question as to the source
of the guanosine triphosphate and of the uridine triphosphate found in the
acid-soluble fraction after pyrophosphorolysis of the S-RNA. We believe
that since guanylic acid and uridylic acid exist as terminal ribonucleotides
in our S-RNA preparations (Fig. 4), these may well be pyrophosphorylyzed
in a manner similar to that in which the terminal adenylic acid and cytidylic
acid are pyrophosphorylyzed. This is supported by the fact that pyro-
phosphorylyzed S-RNA as compared to normal S-RNA, also shows an
enhanced incorporation of uridylic acid similar to that of cytidylic acid.

STUDIES ON MECHANISM OF SYNTHESIS OF SOLUBLE RIBONUCLEIC ACID 121

4. The effect of snake venom phosphodiesterase

If S-RNA-^ plus -y is treated with snake venom phosphodiesterase
under conditions of limited degradation, a somewhat random degradation
occurs as evidenced by the fact that p-P]-CTP attaches to a variety of
terminal groups (Table V), as contrasted to its attachment to cytidylic

TABLE V

The Site of Attachment of ['-P]-CMP on Native and Pyrophosphorvlvzed

S-RNA

Radioactivity in the 2' (3')-ribonucleoside mono-
phosphates isolated after alkaline hvdrolysis of
the 32p.S-RNA

Results expressed as per cent of total c.p.m.
incorporated

AMP CMP UMP GMP

Native S-RNA

(o-o /^moles P-P,

2-0 |uatoms Mg) <5 90 S-io <s

P-P-treated S-RNA

(i -o /xmoles P-P,

20^molesMg) 15 85 <5 <s

P-P-treated S-RNA

(4-0 /^imoles P-P,

8 • o /^imoles Mg) 48 50 < 5 < 5

Phosphodiesterase

treated S-RNA 20 55 11 14

S-RNA-/S and -7 was incubated with the ribonucleotide-incorporating enzyme
under standard conditions and with the amounts of pyrophosphate and Mg + +
indicated within parentheses above. The fourth sample of S-RNA was pre-
treated with phosphodiesterase. The various S-RNA preparations were then
extracted and incubated with [^'^P]-CTP (^-P-P-P) and the ribonucleotide-in-
corporating enzNTTie under standard conditions. At the end of the incubation each
S-RNA was isolated, hydrolyzed with alkali and the radioactivity in the various
ribonucleotide fractions determined.

acid alone when native S-RNA is the substrate or to cytidylic acid and
adenylic acid when pyrophosphorylyzed S-RNA is the substrate.

The experiment summarized in Table VI was designed to establish
whether, after degradation of the S-RNA by snake venom phosphodies-
terase to such a degree that its ability to accept amino acids is lost, it is
possible to restore this ability by re-incorporating ribonucleotides into
the partly degraded S-RNA.

122

E. S. CANELLAKIS AND EDWARD HERBERT

The combined S-RNA-^ and -y was partly degraded with snake
venom phosphodiesterase. The partly degraded S-RNA was incubated
with the ribonucleotide-incorporating enzymes, in the presence of the
non-radioactive ribonucleoside triphosphates listed in Table VI. At the
end of 10 min. the amino acid-activating enzyme was added to each tube
together with radioactive threonine and the capacity of the S-RNA to
accept a radioactive threonine was measured.

TABLE VI

Reconstitution of the Capacity of S-RNA-^ and -y to Accept Threonine

Pretreatment

Incorporat

on of "C-

threonine into S-RNA

Ribonucleotide-incorpora-

c.p.m.^

m/x moles/

ting enzyme plus :

mg. RNA

mg. S-RNA

Normal S-RNA

o

606

o- 17

ATP, CTP, UTP, GTP

707

022

Diesterase

treated S-RNA

o

203

0-058

ATP

70

0-020

ATP, CTP

385

Oil

ATP, CTP, UTP

450

013

ATP, CTP, UTP, GTP

690

0197

The phosphodiesterase treated S-RNA was prepared by treatment of the
S-RNA with snake venom phosphodiesterase, and was then isolated free of the
phosphodiesterase. The pretreatment of the S-RNA was carried out as follows :
0-15 ml. of S-RNA-/3 and -y (Eogg = 40) were incubated with the equivalent amount
of ribonucleotide-incorporating enzyme (0-15 ml.), i /itmole of MgCli, 0-25
/xmole of the indicated ribonucleoside s'-triphosphate in o-o8 M potassium
phosphate buffer, pH 72; the total volume was 0-5 ml. After incubation for 15
min. at 37°, 0-5 ml. of a mixture containing 0-2 ml. of the amino acid activating-
enzyme fractions, 0-05 /xmole of [^*C] -threonine (35 x lo*^ c.p.m.//xmole), 4
/Limoles of ATP and 2 /xmoles MgClo, all in the o-o8 m potassium phosphate buffer,
were added to each tube and the incubation was continued for 10 min. more at
37' . The S-RNA was then isolated, freed of contaminant radioactivity and counted.

Degradation of the S-RNA with phosphodiesterase results in a decrease
in its capacity to accept threonine. If the degraded S-RNA was pretreated
with the ribonucleotide-incorporating enzyme and ATP, a further decrease
in its capacity to accept threonine was noted. In contrast to this, when pre-
treatment was carried out in the presence of ATP and CTP, the capacity
of the degraded S-RNA to accept threonine was greatly enhanced. Pre-
treatment with ATP, CTP and UTP further enhanced the ability of the
S-RNA to accept threonine, and finally, pretreatment of the S-RNA with

STUDIES ON MECHANISM OF SYNTHESIS OF SOLUBLE RIBONUCLEIC ACID 1 23

all four ribonucleoside 5 '-triphosphates resulted in almost complete re-
constitution of the capacity of this previously degraded S-RNA to accept
threonine (93% of the capacity of the control).

We know that at least three of the four ribonucleotides can be incor-
porated into S-RNA (Table VII), and have presented evidence that all four

TABLE VII
Effect of Phosphodiesterase on the Incorporation of ^*C-ATP, CTP, and

UTP INTO S-RNA-/3 AND -y

Radioactive

Ribonucleotide additions

ribonucleotide

-

incorporated

Radioactive

Non-radioactive

m/tmoles per
mg. S-RNA

(I)

Normal S-RNA

i*C-ATP

None

2-0

i*C-ATP

CTP, UTP, GTP

2-68

Diesterase S-RNA

"C-ATP

None

0-99

i^C-ATP

CTP, UTP, GTP

13-68

(2)

Normal S-RNA

i^C-CTP

None

2-2

i^C-CTP

ATP, UTP, GTP

2-2

Diesterase S-RNA

i^C-CTP

None

38-7

i*C-CTP

ATP, UTP, GTP

46-0

(3)

Normal S-RNA

"C-UTP

None

026

i^C-UTP

CTP, GTP, ATP

Oil

Diesterase S-RNA

"C-UTP

None

1-8

uc-UTP

CTP, (;TP, ATP

0-17

ribonucleotides are required to restore the amino acid-incorporating
capacity of the partly degraded S-RNA; we may therefore conclude that
whatever the type of reconstitution of the S-RNA that is occurring it is
not limited to the reconstitution of the terminal sequence — CMP — CMP — •
AMP but is more extensive.

These experiments represent a first crude attempt to synthesize a
defined ribonucleotide sequence in a ribonucleic acid-complex, based on
the underlying a priori assumption that the ribonucleotide sequence of
the S-RNA determines the specificity of the amino acid attachment. The
results are unduly complicated bv the fact that we probablv do not have a
single species of S-RNA, but are at present using as a substrate a mixture
of ribonucleic acid molecules, as evidenced both by the ability of the
S-RNA to accept a variety of amino acids and by the results of the end-
group analysis.

124 ^- ^- CANELLAKIS AND EDWARD HERBERT

Discussion

TisELius: May I ask did you get complete separation into S-RNA-a, -j3, -y ?

Canellakis: Yes, we do get complete separation.

TiSELius: Are the salt concentrations eluting the various components very
far apart ?

Canellakis : I think they are of the order of o -05 m to o • 2 M.

TiSELius: So you don't get much overlap between them ?

Canellakis: They come out as distinct peaks

Allfrey : I would like to ask if you have tried reconstituting this esterase-
treated RNA, not with the normal nucleoside triphosphates but with azaguanosine
phosphates or fluorouracil phosphates and see whether you still get functional
RNA?

Canellakis: No, we have not tried that.

Allfrey : It is interesting because the question is : does the specificity arise
in that end of the molecule or in the rest of it ?

Canellakis : I agree that it is interesting. I think eventually we would like to
try this experiment.

Davis : Do you have any indication of the extent to which the diesterase has
degraded the molecule, whether it has acted selectively at the ends or whether it has
also attacked the middle ?

Canellakis : On prolonged incubation we can get some attack in the middle.
We think that it is probably due to some endonuclease action in our preparation but
for a half hour or so we seem only to be getting terminal degradation.

Herbert : I should like to add that if you incubate for prolonged periods you
release 40 or 50^0 of the ribonucleotides as acid-soluble material. You can then no
longer reconstitute the amino acid-accepting capacity of our system.

von der Decken : Do you obtain the pyrophosphate exchange when you use
S-RNA-amino acid complex instead of S-RNA ?

Canellakis: I don't believe that our data permit us to answer that question.

MICROSOMES AND PROTEIN SYNTHESIS

The Endoplasmic Reticulum: Some Current
Interpretations of Its Forms and Functions

Keith R. Porter
The Rockefeller Institute, Neiv York, U.S.A.

Among the several discoveries in cell morphology which must be
credited to electron microscopy, one of the most significant is that which
has described the cvtoplasmic ground substance as extraordinarily rich in
membrane limited elements. The presence of these in such amounts and
in such complicated forms was not really suspected from light microscopy.
Their presence in the ground substance was first recognized in the electron
microscope image of cultured cells about 15 years ago, though it must be
admitted that the early images were not very convincing [i]. Subsequently,
and with increasing vigour, they have been studied in a wide variety of
animal cells and more recently in plant cells until, at the present time, they
are looked for and recognized as a standard element of cell fine structure.
The unit of structure — if such can be defined — is vesicular or tubular in
thin section profiles; i.e. a membrane-enclosed cavity or vesicle, itself
usually devoid of any internal structure. When followed through serial
sections into the depth dimension, these profiles can be seen to represent
two-dimensional images of structures which are really part ot a complex
three-dimensional system of tubules and vesicles extending into most
regions of the cytoplasm. This statement holds true for the majority if not
all cell types.

It is also now recognized and accepted as a law of fine structure that
the nuclear envelope is part of this general system. This conclusion is
supported by the structural similarity evident between the cytoplasmic
elements of the system and the nuclear envelope as well as by the im-
pressive continuity between the envelope and the cytoplasmic units [2, 3].
The envelope is essentially a large flattened or lamellar vesicle wrapped
around the nucleus. It is reasonable, therefore, to regard the nuclear
envelope and the cvtoplasmic formations as part of a single unit system of
the cell, in the same sense that we regard the mitochondria or plastids as
representing unit systems. The nuclear envelope is the most constantly
occurring part of the svstem, being found in all forms from yeast cells on
up the scale of plant and animal species. It is conceptually useful and

128 KEITH R. PORTER

possibly valid to think of the cytoplasmic part of the system as an outgrowth
of the envelope.

This is the system that has come to be known as the endoplasmic
reticulum, or for short, the ER [4]. In one structural expression or pattern
it adopts, there are many lamellar sacs or cisternae in parallel array and
this form corresponds to the cytoplasmic component, referred to for 60
years as the ergastoplasm [5]. But this is only one of many complex
organizations that are encountered and which are characteristic of different
cell types. In some other cells the system is represented in large part by a
tri-dimensional latticework of tubules. In cells of the early animal embryo
or the plant meristem, before differentiation is completed, the ER is
usually simple compared with the patterns and complexities achieved in
the completely differentiated unit. This suggests that it is designed to play
an important role in the particular activity of the differentiated cell,
whether this is secretion, contraction or impulse conduction.

These and other observations make it evident, then, that we are dealing
with a complex, finely-divided vacuolar system which ramifies and extends
to all parts of the cytosome. This effectively creates within the cytoplasm
an internal phase which is separated from the continuous matrix phase of
the cytoplasm by a membrane. These facts of morphology, as depicted by
electron microscopy and supported by phase contrast observations on
living cells, have been the basis for diverse speculations regarding the
function of the system. It has, e.g., been suggested that the system satisfies
a need in larger cells for channelled diffusion and segregation of metabolites.
Its peculiar morphological relation to the myofibrils in striated muscle has
prompted the proposal that its limiting membrane is electrically polarized
and possibly capable of transmitting intracellular impulses. The large
surfaces provided by the system have been recognized as suitable for the
support and patterned disposition of enzymes within the cytoplasm.

This information and speculation regarding the ER is now fairly well
known and probably needs no further comment. It can be found in a
number of recent reviews with ample illustration of the more important
points [5, 6, 7]. The system described has been an object of interest and
study in our laboratory for a number of years. Large blocks of what we
now know about the system have been discovered and described by
G. E. Palade and P. Siekevitz, and I am particularly grateful to them for
valuable discussions and permission to republish a few of their micrographs.

It is the purpose of this review to present a morphological background
against which the subsequent biochemical papers on this programme may
be presented. I shall focus your attention principally on functions of the
ER — functions that are suggested or defined by (a) morphological varia-
tions observed in cells with similar and different functions (comparative
cytology), and {b) modulations related to naturally occurring or experi-

THE ENDOPLASMIC RETICULUM 129

mentally induced changes in physiological function. If, as now seems
possible, these obseryations can be tied to biochemical obseryations on
fractions of cells identified as coming from this or that part of the ER, we
haye a good start on the oyer-all analysis of the role of the ER in the
intact unit. Needless to say, \ye are still yery ignorant about the reticulum.
Ho\ve^■er, some correlations of function and structure are being achieved
and with the possibilities provided by modern techniques (as yet un-
explored) the progress toward greater understanding should be rapid. I
am encouraged to believe that even the suggestive evidence provided by
observations on morphological associations of structural elements invohed
in different cell functions will be of interest to biochemists who are more
experienced in cell fractionation procedures and their potentials.

Forms and modulations of ER

The cytoplasmic part of this membranous system, as distinct from that
part comprising the nuclear envelope, appears in a wide variety of forms.
These are readily recognized as characteristic of particular cell types
whether found in one or different species of animal. It is apparent, more-
over, that cells performing certain similar functions, whether otherwise
related, show similarities in the predominant form of the reticulum in
their cytoplasms.

In some cells the system is represented in large part by only one of
these differentiations or configurations. In other instances two or more
forms may be prominently represented in a single cell and obviously
linked into a single system bv morphological continuity. There are also
cells in which a predominant form of the ER shows local specializations
or difierentiations, designed, presumably, to perform a particular function.
The system under consideration is therefore one of great variability,
possessing infinite possibilities to accommodate or control subtle func-
tional activities of the cell.

These \arious patterns or designs displayed by the system are of
course not fixed or rigid, but modulate within limits as the cell performs
its functions or responds to environmental stimuli. The pattern most
regularly encountered in microscopic examination of any cell type is
regarded or generally accepted as representing the preferred form of the
system.

The granular or rough form

The picture of the endoplasmic reticulum that is most familiar to bio-
chemists consists of arrays of parallel lines encrusted with small dense
particles (Figs, i and 4). In actual fact, of course, the lines represent
membranes which limit relatively large lamellar vesicles and it is charac-

K

I30

KEITH R. PORTER

"^^ ; -^^^^^

ff

^i

if.

J

p

%

^v^:•- ■^•

ers

j*^^

THE ENDOPLASMIC RETICULUM 131

teristic for these to lie parallel in groups or stacks of several units. Some-
times these vesicles or cisternae may possess dimensions of several microns
in the plane of the lamellae and have a thickness not greater than 50 m/i.
Such large cisternae are further characterized by having large populations
of small particles (150 to 200 A) attached to their outer surfaces. This is
the surface contiguous with the matrix phase of the cytoplasm and, indeed,
the particle may properly be regarded as a component of the cytoplasmic
matrix. In pancreas cells, such cisternae with particles occupy about three-
fourths of the volume of the cytoplasm. In liver cells they are less
prominent and occur in small arrays scattered throughout the cell (Fig. i).
In both these cells and in others as well, such arrays are observed to
coincide with the more intensely basophilic parts of the cytoplasm
(ergastoplasm) in which, histochemical tests have shown, there are high
concentrations of RNA [8, 9]. This form of the ER has a widespread
occurrence among cells and is now regarded as present in one or another
minor variation in all cells known to be engaged in the synthesis of a
protein for secretion (or export) from the cell. These facts of morphology,
together with much older observations on the behaviour of the ergasto-
plasm during phases of secretion and recovery, would convince one, even
without other evidence, that this form of the ER is somehow functional in
protein synthesis [10, 1 1]. The investigation of this topic has not, however,
stopped with microscopic cytological and histochemical studies, and much
more convincing evidence of this functional association is now available.

The microsome

This body, the subject of this session of the current symposium, was
identified and characterized first as a submicroscopic unit or particle of
the cell that could be consistentlv separated out (from a variety of animal
cells) by following a certain schedule of centrifugal fractionation [12, 13].
The microsome fraction was further characterized by Claude as being rich
in phospholipids and RNA, and as being related in some manner to the

Fig. I. This depicts the fine structure of a normal liver cell of the laboratory
rat. The nucleus {N) is at the ritjht; the marginal part of a blood sinusoid (s) is at
the left. Mitochondria are numert)us and readily identified by their characteristic
morphology. The endoplasmic reticulum is represented by two distinct forms. One,
shown in the centre of the field (er), consists of large, thin, lamellar vesicles or
cisternae with small (150 A) dense particles attached to their surfaces. In thin
sections these appear as rough, line-limited profiles, associated in groups which
coincide with the basophilic component of the light microscope image. The other
form is particle-free or smooth and is constructed of fine tubular elements ca.
50 m/^t in diameter (ers). These comprise in liver cells what appears as a loose tri-
dimensional lattice. The Golgi component is represented at G.

132 KEITH R. PORTER

chromophilic material of the liver cell cytoplasm (1943). A definitive
demonstration of the intracellular origin of the microsome v^'as not, how-
ever, provided until about 12 years later when Palade and Siekevitz [14]
published micrographs of sections through the pellet representing the

Fig. 2. Micrograph of thin section of microsome pellet from rat liver homo-
genization and fractionation. The long slender profiles are readily identified as
fragments of intracellular, particle-studded (rough) cisternae of the endoplasmic
reticulum (see Fig. 4 for comparison of rough ER in intact cell). There are also in
the field (at arrows) a few profiles of vesicles without particles (smooth or agranular).
These are thought to be derived from the smooth form of the reticulum observed
in liver cells. Mag. x 70 000. (Micrograph courtesy of G. E. Palade.)

microsome fraction. These pictures showed profiles ot particle-studded
units which by their morphology could be unmistakably identified as
fragments or pieces of the endoplasmic reticulum of the intact rat liver
cell (Fig. 2). Thus the microsome became part of a well characterized

THE ENDOPLASMIC RETICULUM 1 33

system of the cytoplasm and all the data and information that had pre-
viously been gathered on liver microsomes became applicable to the intact
svstem. The rough form of the ER could be said to incorporate amino
acids into proteins, as shown for the microsome by the pioneering experi-
ments of Hultin [15] and Borsook et al. [i6]. Then also it became valid to
ascribe to the ER glucose-6-phosphatase activity and 40 to 50",, of the
RXA in the whole tissue (li\er in this case). These and other properties
described by more refined procedures will doubtless be considered in
subsequent papers. The sole purpose here is to relate the microsome to
the ER.

As noted above, the term "rough" is applied to this form of the
reticulum because of the presence in its surface of numerous particles.
Thus the microsome is made up of at least three parts: the particles, the
membrane, and the contained space. Interest in the specific functions and
properties of these had led to several studies on them as separate entities.
With the aid of deoxycholate (DOC), the particles ha\"e been separated
from the membranes, and by electron microscopy have been shown to be
relati\elv free from contamination [14]. Examination of such preparations
have further shown these particles to contain 80 to 90",, of the RNA of
the intact or complete microsome and onlv 20",, of the protein nitrogen.
The particles are therefore appropriatelv referred to as RNP particles or
ribosomes. The other two parts of the microsome, the membrane and
content have been freed of particles with versene, likewise examined for
purity in the E.AL, and in\"estigated for other properties. As might be
expected, they possessed the other properties of the intact microsome
before the ribosomes were remo\ed, i.e. most of the protein, nearly all of
the phospholipid, and the DPNH-cytochrome c reductase activity.

Here again, with these fractions of the microsomes, isotope incorpora-
tion studies have been fruitful and, as is well known, have described the
ribosome as possessing the capacity to combine free labelled amino acids
into proteins. Thus, in the case of protein synthesis, the particle rather
than the membrane is the important structure and would seem to be
responsible for determining the character of the protein synthesized. This
conclusion finds strong support in the recent studies of Siekevitz and
Palade [17, 18] on the relation of RNP particles to the synthesis of
pancreatic enzymes.

With a specific function assigned to the ribosome, it is of some interest
to inquire into the role of the membrane and cavity of the intact system.
As previously noted, it ob^"iouslv provides a large surface for the disposi-
tion of the particles. There is, moreover, pictorial evidence that this
distribution follows certain repeating designs presumablv related to
patterns in the membrane [19]. Whether there is some functional purpose
in this is not clear. That the membranes are not essential to the svnthetic

134 KEITH R. PORTER

A?^

. •y^^'^nl

Fig. 3. Micrograph of part of an acinar cell of the guinea-pig pancreas. The
basal surface of the cell crosses the figure at the lower left ; the apical pole, with a
few elements of the Golgi (G) showing, is at the upper right. The specimen was
fixed during the recovery phase following secretion of zymogen granules.

The objects of greatest interest here are the large, dense, spherical granules
within the cisternae of the ER. Now known as intracisternal granules, these units
have been shown to possess the enzymic activities of purified zymogen granules.
It is thought that in acinar cells of guinea-pig pancreas the zymogen portions,
sequestered in the cavities of the ER, condense out as granules before being trans-
ported to the Golgi component in the apical pole for packaging as secretory
granules. The intracisternal granules are indicated at icg; the mitochondria at m.
Small (150 A) dense granules are abundant in the cytoplasm of these cells and
particularly on the surfaces of the ER. The specimen was fixed in OSO4 and stained
with lead. (Micrograph courtesy of G. E. Palade.)

activity of the ribosome is suggested by the fact that in growing cells, e.g.,
the particles seem to perform their synthetic role without membrane
support. The distinction to be made is that protein production in this latter

THE ENDOPLASMIC RETICULUM 1 35

instance is for internal consumption rather than for export. Only in other
instances where particles are associated with the membranes of the ER, is
the product prepared for secretion and the external environment of the cell.
For observations on the role of the ER in this phenomenon one must
turn to studies on intact cells engaged in protein synthesis. In an early
paper, Weiss [20] noted that the cisternae of rat pancreas cells appeared
swollen and filled with a material which he interpreted as zymogen during
the synthetic phase of the cell's activity. The real nature of the intra-
cisternal material was not, however, demonstrated. Later, Palade [6], in
some observations on guinea-pig pancreas, observed that during the
period of gland recovery after secretion, relatively large granules, having
the density of zymogen granules, appeared within the cisterna of the ER
(Fig. 3). These granules were subsequently isolated and analyzed [21, 22]
and found to have the properties of zymogen granules as normally secreted
by the cell. In this instance, then, and probably in every other case of
ergastoplasm function, the product of synthesis is, so to say, separated or
segregated from the site of synthesis (the ribosome) by the membrane
limiting the ER cisterna. One can say that the product is externalized with
respect to the continuous phase of the cytoplasmic matrix, where, if it
were left, it would presumably diffuse away among the structural proteins of
the cytoplasm.

Once sequestered within the cavities of the ER it appears to move, in
a dissolved or diffusible form, to the Oolgi region in the apical pole of the
cell, where it is apparently prepared tor secretion. Some evidence of at least
temporary continuity between elements of the ER and the Golgi com-
ponent, which would provide channels for transport, have recently been
described [23]. Thus, in addition to its role in sequestering the product of
synthesis, the ER appears to function in transporting metabolic products
from one point to another in the cell. It mav be recalled that this was
mentioned earlier as a logical function of the system.

The smooth form of the ER

In all the attention gi\en the more striking, rough form of the ER and
the ribosomes, other configurations of this intracellular system are seem-
ingly overlooked. Yet there are probably as many types of cells with
promment differentiations of the smooth ER as there are with the rough
form. In some instances, as mentioned earlier, the two are mixed in one
and the same cell and where this is the case, microsome fractions can be
shown to contain both types (Fig. 2). Thus in one of the earliest studies
of liver microsomes, note was taken of a substantial content of smooth-
surfaced vesicles derived, it was thought, from the smooth-surfaced
portions of the endoplasmic reticulum [14].

136 KEITH R. PORTER

This smooth form, hke the particle-encrusted portions of the ER
possesses a characteristic morphology. The unit of structure is tubular
(50 to 100 A in diameter), and these are arranged in varying degrees of
compaction to form complex tri-dimensional lattices. Such configurations
occur in a wide variety of cell types. In some instances their development
appears to be related to maintenance of form, e.g., in the sustentacular
cells of the frog olfactory epithelium [24]. In cells of the pigment epithelium
of the frog retina a development of the smooth ER occupies a large part
of the cytoplasm and shows no affinity for the few ribosomes present [25].
It is the predominant form in the sarcoplasm of striated muscle cells and
shows a pattern differentiation related to the sarcomeres of the myofibrils
[26]. What if any role it performs in common in these widely different cells
is a large question.

In at least one class of cells, that engaged in the synthesis of lipid
materials for export, the smooth surfaced form of the ER occurs con-
sistently. It has, e.g., been described as the dominant form in the more
mature cells of the sebaceous glands (Meibomian gland of the rat) [27].
The secretory (lipid) droplets form initially in these cells in close relation
with the tubular elements of the ER and appear later to be segregated for
secretion within the vacuoles of the Golgi complex. A similar development
of the agranular ER has been reported in cells producing steroid hormones,
for example the interstitial cells of the opossum testis [28]. Here the ER
is entirely devoid of dense particles and the cells are intensely acidophilic.
These observations have special interest and significance because Lynn
and Brown [29] in a biochemical study on similar material found evidence
that enzymes involved in the production of testosterone from progesterone
are associated with microsomes and that, though inactivated by lipases,
they are unaffected by ribonuclease at concentrations that normally inhibit
protein synthesis in liver and pancreatic homogenates. The inference is
that the enzymes active in steroid synthesis are associated with the RNA-
free microsomal membranes. This is further supported by observations on
fetal zone cells of the human adrenal in which there is a richly developed
smooth ER [29a].

These three instances suggest a correlation between the agranular ER
and contain phases of lipid metabolism, especially where a product is
synthesized for export from the cell. The exact site of synthesis with
respect to the ER has not in these cases been clearly defined. Thus one
cannot say that it is sequestered within the cavities of the ER and trans-
ported thence to the Golgi, though involvement of the Golgi component
would suggest that the sequence encountered in protein synthesis may be
followed here as well.

In this regard the morphological phenomena observed in intestinal
epithelial cells of the rat during fat absorption can be interpreted to have

THE ENDOPLASMIC RETICULUM 1 37

some significance. The story comes from recent investigations of Palay
and Karlin [30]. These authors, in following the uptake of fat by epithelial
cells of the intestinal villus of the rat, observed that small fat droplets
( ~ 50 m/.t in diameter), which appeared to traverse the cortex of the cells
in pinocytotic vesicles, reappeared in the cavities of the ER. Thev were
then transported within the channels of the system to the lateral margins
of the cell and thence, presumably through intermittently established
openings, into the inter-cellular spaces. All compartments of the ER
including the nuclear envelope were seen to contain droplets of fat within
30 min. after intragastric instillation of corn oil [31]. Besides providing
evidence of transport within the ER, these observations are descriptive of
continuity within the system. The authors express the opinion that the fat
particles which appear in the cisternae of the ER are transported directly
there by pinocytotic vesicles. It follows from this opinion that the passage
of fat through the epithelial cell is to all intents and purposes an extra-
cellular phenomena.

This interpretation of e\ents in fat absorption is introduced into this
discussion because it bears on the question of the relation of the ER to
the plasma membrane and the cell surface, and also because there is an
alternate interpretation which seems to more nearly fit the observations
and related information. According to the Palay-Karlin interpretation,
materials taken up in small bits or sips (quanta) at the cell surface are
discharged helter-skelter into the lumina of the ER. This means that a
space designed to sequester the products of synthesis mav also receive all
manner of metabolites and non-metabolites from the environment. These
concepts have grown out of observations on pinocvtic activity at the cell
surface [6] and some e\idence of occasional and probablv intermittent
continuity between membranes of the ER and the cell surface [32]. They
remain, however, as concepts and much of the evidence a\ailable on
uptake of particulate material visible with the E.AI. fails to support them
[33, 34]. As a rule, a barrier of two membranes (plasma and ER) and a
layer of cytoplasmic matrix always lies between the extracellular space and
the cavities of the ER on the uptake side. Even the Palav Karlin observa-
tions may be interpreted to fit this principle. Thev admit in their report
that definite exidence of continuity between pinocytotic vesicles and the
ER is not available and that furthermore it is difficult, through the apparent
pinocytotic actix ity, to account for the rapid fat absorption. There is good
evidence, moreover, that long-chain fatty acids proxided in dietary
triglycerides are rearranged in the fats in the chylomicrons on the other
side of the epithelium [35]. These points support an alternate interpreta-
tion that only the products of triglyceride hydrolysis pass the plasma
membrane and enter the cytoplasmic matrix ; that these are reassembled
at the membrane limiting the ER and resynthesized into fats which are

138 KEITH R. PORTER

sequestered as droplets within the cisternae. This scheme would provide
for the production of triglycerides at all levels in the ER — a possibility
suggested as well by the appearance of large quantities of fat globules in
the ER in deep parts of the cell within a short time after fat digestion
begins. The sequence of synthesis, sequestration and transport is obviously
analagous to that encountered in zymogen granule formation as depicted
in the guinea-pig pancreas.

Glycogen metabolism

The instances already cited in which the smooth or agranular ER is a
prominent component of cell fine structure do not exhaust the known
occurrences. Other reports include observations of agglomerations of
agranular vesicles and tubules in liver cells of the rat [36, 6] ; and a remark-
able tri-dimensional lattice of small tubular elements as part of the gly-
cogen-rich paraboloid found in the inner segment of the turtle cone cell
[37, 38]. Continuity between the smooth elements and adjacent rough cister-
nae describe them as parts of a single system, the endoplasmic reticulum. In
each of these instances, unlike those previously noted, the smooth ER is
specifically associated with glycogen deposits.

One of the earliest descriptions of this smooth form of the ER in liver
cells appeared in a paper by Fawcett [36]. In the liver cells of rats, fasted

Fig. 4. The image shown here represents parts of two adjacent liver cells (rat)
plus a neighbouring sinusoid (s). In the cell at the bottom of the picture, one can
identify the nucleus (A^) with its typical envelope ; mitochondria (w) ; long profiles
of cisternae comprising the granular or rough form of the ER or ergastoplasm (er) ;
and a few profiles of elements belonging to the agranular or smooth ER (ers). The
rough ER (within the rectangle) is particularly suitable for identifying the source
of the microsomes in Fig. 3.

The part of a cell at the upper left, beside the bizarre mitochondrion, contains
two elements of particular interest here. The first is very dense (following staining
with Pb(OH)2) and represents the glycogen (gl). The unit structure, which appears
as a cluster of granular subunits, usually does not exceed a maximum size of about
150 m/ti. From this, the size ranges downwards to individual subunits. This is the
liver cell component that fluctuates with physiological conditions leading to
glycogen storage or depletion. It varies directly with intensity of PAS staining and
is removed by diastase. A number of investigators have referred to it as glycogen.

These dense bodies are interspersed with less dense, line-limited elements,
which are interpreted as vesicles and tubules of the smooth ER (ers). They seem
to form an irregular lattice. While the diameters of these elements are not uniform,
it is obvious that they fall within a fairly narrow range of variation.

This is the picture of association between ER and glycogen found during the
storage phase. The specimen was taken 2 hr. after re-feeding following a 24-hr.
fast. When taken later at 4 or 6 hr., the cells show much more glycogen and the
ER profiles intermingling with the glycogen are relatively fewer. The specimen was
fixed in OsOj, embedded in Epon 812, and, after sectioning, stained with Pb(OH)2.

THE ENDOPLASMIC RETICULUM 139

and re-fed, he noted compact masses of " small vesicles and short tubules ",
especially at the periphery of the cell. A subsequent investigation dis-

covere

d that this smooth form of the ER undergoes a great hypertrophy in

W^

..^^

^■'t

••.^

m

%

^'

■«»•'

X

■if »r

A

^*

'_?-

N

liver cells of rats given the toxic azo dye, 3'-methyl-4-dimethyl aminoazo-
benzene [39]. This is a substance which, when fed to these animals, blocks
glycogen storage in affected liver cells, and eventually induces liver
tumours. In this study it was noted that these prominent developments of

140 KEITH R. PORTER

the smooth ER appeared initially in association with glycogen deposits and
that concomitantly with their development the glycogen was depleted and
did not reappear while the dye was included in the diet. It was these

/^«

^i

'4.

-T'lt '^

- • m

''« ».

W^^

5^- -•

\9

I

cm

cm

«^

^ '. v^-^

li

■V,v/-» ^^

observations that stimulated an investigation into the possible relationship
between the smooth ER and glycogen metabolism which is still continuing.
The following summary is drawn from experiments made in collaboration
with Drs. Giuseppe INIillonig and Thomas Ashford. Detailed accounts of
the results will be published elsewhere.

THE ENDOPLASMIC RETICULUM I4I

The initial obsenations were derived from a systematic examination of
liver tissue from animals fasted for 24 hrs. and for 5 days and from similar
animals after periods of re-feeding varying from i to 6 hr. Electron micro-
graphs of the liver cells of these animals reveal a constant and specific
association between the smooth ER and glycogen. The development of the
svstem is particularlv striking during the period of glycogen synthesis at,
e.g., 2 hr. after re-feeding following a fast of 24 hr. (Fig. 4). Pictures of
this kind supported the early speculation that the membranes and vesicles
might contain factors significant in glycogenesis [39]. When the stores of
glycogen have reached a maximum (e.g. 6 hr. after re-feeding), the profiles
of vesicles intermingling with the glycogen are relatively far less numerous.
Whether this represents an absolute reduction in the extent of the system
(a reversal of the absolute increase that occurs with re-feeding) or is simply
a result of dilution by the glycogen stores has not been determined.
Following a period of fast (e.g. 24 hr.), the vesicular component of the
residual glycogen complexes is again relatively prominent (Fig. 5), though
not so impressive in absolute terms as shortly after re-feeding. The system
in the fasted cells is morphologically distinctive ; many of the vesicles show
a low density content and are more closely compacted with the glycogen
units. There are, in other words, distinguishing ditferences to be noted
between the smooth ER and the fasted and re-fed cells (compare Figs. 4
and ^5). The image of the glycogen and associated ER in the cell of the
fasted animal is repeated in animals that have been given glucagon or
epinephrine. Furthermore, the effects of fasting and of glucagon and
epinephrine are repeated as well in the fine structure of rat livers perfused
under isolation for periods of 4 to 6 hr. [40].

There are, unfortunately, several ways to interpret these observations.
The structural association of the smooth ER and the glycogen is a specific

Fig. 5. This shows parts of two adjacent rat liver cells. In this case, in contrast
to that shown in Fig. 4, the specimen was taken at the end of a 24-hr. fast period.

The cells are depleted of their glycogen (gl) except for residual units in small
marginal zones of the cytoplasm, as shown here. Following glucagon injections even
these residual units are mobilized and many cells appear devoid of any glycogen.
The space between the residual glycogen units is filled by somewhat expanded or
bloated vesicles of the smooth ER. These seem to fill the available space and crowd
close to the glycogen, possibly for functional purposes. Essentially the glycogen is
vesicle-locked. The vesicular profiles vary more in size and shape than those in
Fig. 4 and seem to have a content of lower density (possibly because something has
dissolved out). A number of these vesicles can be seen in close proximity to and
possibly blending with the cell surface or membrane (arrows).

The cell margins are contiguous and run from (cm) to (cm). It is obviously
difficult to follow the plasma membranes along this stretch, perhaps because it is
discontinuous. Mitochondria (m) are obvious; rough ER {er) is randomly scattered.
The specimen was fixed in OSO4, embedded in Epon, and stained after sectioning.

142 KEITH R. PORTER

one and naturally suggests a functional relationship. Particularly is this so
when, as in the azo dye experiments, a hypertrophy or abnormality in one
is accompanied by a diminution in the other. But whether the association
is important in glycogenesis and storage or in glycogenolysis (or both) is
not clearly evident. By way of resolving the problem one can cite certain
instances where glycogen is stored without apparent involvement of the
smooth ER. In diabetes mellitus, for example, the nuclei of liver cells show
large glycogen stores. The same have been reported in liver nuclei of young
frog tadpoles [41] and in the nuclei of cells of a chicken sarcoma [42],
where membranous elements of the ER are not normally found. Also in
chondrocytes, the cytoplasmic deposits of glycogen apparently put down
for future use of the cell, show no prominent membrane components. If
one excludes the possibility that the fixation or staining (or even the
timing of the fixation) have not been appropriate in these instances to
preserve this form of the ER, one is left with the conclusion, also supported
by recent report of Revel et al. [43], that the ER is not an obligate associate
of glycogen synthesis.

This encourages one to focus on the depletion side of the metabolism
cycle, in which the liver cell has a more unique function to perform. As is
well known, it has to export glucose on demand and keep it unavailable to
the general carbohydrate metabolism of the liver cell itself. It is important,
therefore, to have it sequestered near the site of storage or availability and
thence transported to the cell surface. If functional in this manner, the
smooth ER in the liver cell would be performing a role closely analogous
to that of the system in other types of cells described earlier in this report.

It is not difficult to find in the pictorial data, evidence to support this
concept. The morphology of the ER relative to the glycogen is identical
during the depletion phases in fasting animals and after glucagon. At such
times the association of vesicle and glycogen is more intimate. The vesicles
show a very low density suggesting that these contents are soluble. And
it is usual to find the glycogen complexes close to the cell margins with
many of the smooth surfaced vesicles up against the adjacent plasma
membrane, possibly preparatory to discharging their contents, assumed to
be glucose. Also in this regard, the association of gIucose-6-phosphatase
with microsomal membranes (presumably agranular) is not to be over-
looked. Since, however, these observations constitute the only information
we have, it is clear that the final chapters in this investigation of the smooth
ER and glycogen have yet to be written.

Endoplasmic reticulum in plant cells

Thus far in this review the observations we have considered have been
derived from animal cells. Certain generalizations have been presented

THE ENDOPLASMIC RETICULUM 1 43

relative to the structure and function of this newly defined system — again
based on animal material. Obviously if these are to be accepted as prin-
ciples of cytology, they should be equally applicable to plant cells. There-
fore a good deal of interest has been displayed in electron microscopy of
plant cells. The early attempts were less than satisfactory because of the
technical problems encountered. These early observations were, however,
of sufficient value to show that the plant cell had in common with the
animal a membrane limited system, homologous to the ER of animal cells
[44, 24]. Since that time, technical developments have provided improved
images which leave no doubt that plant cells of several kinds have in their
fine structure organelles and systems identical in many respects to those
found in the animal units [44a, 45,]. Thus the nuclear envelope shows
the characteristic two membranes and pores, and continuity at certain
points with the cytoplasmic part of the system. Cisternal units of the ER,
though randomly oriented in the meristematic cells of the root (Fig. 7), do
achieve patterns of organization in cells in more advanced stages of
difi'erentiation (Fig. 6). The RNP particles in plant cells do not show the
extreme affinities for cisternal surfaces that one notes in animal cells, but
particles are attached to some parts of the system and not to others so that
one can recognize rough and smooth divisions (Fig. 6).

The functions of the ER in the plant cell are of course not more clearly
shown by a few pictures than they were in the animal equivalent. However
the material lends itself very readily to controlled experimental study and
one can reasonably hope that a substantial amount of new information may
come from this source. Then also certain features of interest are peculiar
to plant cells, and one can hope to exploit these for clues to function not
displayed to the same advantage in animal cells. At the outset we can
reasonably assume that the little we have learned from animal material can
be transferred to the plant.

One structural association especially has held our interest and it is to
this that attention will be specificallv directed (see Porter and Machado,
[46]).

The majority of the structures representing the ER in meristematic
plant cells are large, lamellar units. When these reach out to the cell
surface they show at their margins distinctive changes in morphology.
These are reflected in the thin section by changes in the dimensions of the
profiles. Thus, instead of being long and narrow, they appear short and
often circular, which indicates that at this level the svstem is constructed
of tubular rather than lamellar units (Fig. 7). These are joined together at
this level into a reticular structure which can be more readily visualized
when included in the plane of the section (Fig. 8). Again, for the morph-
ologist this suggests a functional involvement of the ER in events taking
place at the cell surface. These events include of course some exchange of

m

«

6

Fig. 6. Micrograph of parts of adjacLiit cells in tapctal layer fourul in anthers
of Saiiitpaidia ioimtha. The cytoplasms of the fully differentiated plant cells are
richly supplied with elements of the ER. In a way peculiar to these cells, the ER
shows some of the features which characterize it in animal cells. Some of the
cisternae are lamellar and arranged in groups or clusters. These show the most
pronounced association with RNP particles. As in animal cells, these particles show
some tendency to form miniature patterns in their distribution on membrane
surfaces (arrows). Other vesicles and tubules are essentially free from particles —
and most especially those close to the cell surface. Mitochondria [m) are similar to
those in animal cells. Otherwise the figure shows proplastids {pp), a few dictyo-
somes {d), and some extracellular granules {s).

The specimen was fixed in OSO4. embedded in Epon 812, and sectioned with a
diamond knife. Contrast was improved with lead hydroxide. (Micrograph courtesy
Dr. Myron Ledbetter.)

THE ENDOPLASMIC RETICULUM

145

Fig. 7. This shows the corners of two contiguous meristematic cells of the
onion root tip. The primary wall (czv) separating the cells passes across the lower right
corner; the nucleus of one cell is at N. The wav'y lines in the cytoplasm are
actually paired lines or membranes with an intervening space (see Figs. 11 and 12
for higher magnification of same). They represent sections through large lamellar
units of the ER. Where these approach closely to the cell surface they branch and
reticulate to give in section only small circular profiles (see long surfaces indicated
by arrows). Figure 8 shows the appearance of this when section coincides with
plane of wall. Other components include mitochondria (m), proplastids (pp), and
dictyosomes (d).

The specimen, from an onion root tip, was fixed in KMn04, embedded in nietha-
crylate, and sectioned with a diamond knife.

146 KEITH R. PORTER

metabolites and the production and deposition of materials important in
wall formation. In Fig. 7, e.g., the separation of the cells is a product of a
recent division, and primary wall formation is probably in progress. The
real purpose of the association is, of course, not indicated. But it is of some
value to compare the picture at this cell surface with other instances where
wall formation is also in progress and especially where wall formation is
more clearly the major activity at the moment of fixation.

r , .

» cw ^H

>

.1*
pm

' ^ V

J

<

■ * / ^ : f

"'' '. -^

k. ^ V, 4

Fig. 8. i\Iicrugraph of section cut in plane of surface t)f onion root tip cell. The
section includes a part of the cytoplasm (bottom half) and passes out of cell
obliquely through the plasma membrane (p?n) and then into region of wall (czv).
The small densities with a ring around them represent cross sections of plasma-
desmata (/)). The component of particular interest here is the endoplasmic
reticulum and its distribution just within the plasma membrane. It obviously
consists of many tubular units interconnected to form a reticular structure which
is contiguous with the inner surface of the plasma membrane (see arrow).

One example of this latter situation is encountered in the formation of
the cell plate [47]. As is well known, this structure is characteristic of
dividing plant cells and first appears in light microscopy as a row of small
pectin vesicles arranged along the equator of the telophase spindle [48].
In the electron microscope image the vesicular phase of plate development
is preceded by another structure which has the following developmental
history.

Wjf /

/ \ . ". r

/

/ ?

•^" /

H)

Vf/

■\

erW

/

\

/,

N

Fig. 9. This shows an early stage in the development of the cell plate between
two daughter cells of onion root meristem. The telophase nuclei are at A^ and a new-
nuclear envelope has just reformed {ne). Long slender profiles of the ER, present in
the interzone, appear noticeably shorter toward the plate region {cp), and at this
level they reticulate and intermingle from the two sides to form a dense, irregular
lattice of tubules. Within this structure, which has its origins in the marginal
extensions of the ER, the early pectin vesicles of the plate first appear (see Fig. 10)
for later stage). These grow in number and size and eventually fuse to separate the
two daughter cells. (From Porter and Machado, i960.)

148 KEITH R. PORTER

During anaphase of mitosis, lamellar and tubular elements of the ER,
now divided between the two ends of the cell, invade the spindle from the
poles and sides and approach, from opposite directions, the equator of the

CW /

*y J

'Or;

v

"A

%- .^

^' I

N

\

m''

I

3 -, ^cp

\

^ i

^(

/• -,0

<?'

N

^'■~>

10

Fk;. 10. Micrograph showing di\ rIuil; clII in cortex ot onion root tip. Telophase
nuclei are indicated at A'^, cell plate at cp, and surrounding walls at cw. Numerous
slender profiles of the endoplasmic reticulum (er) are scattered throughout the
cytoplasm. At the level of the plate, and especially at its margins (arrows), the ER
profiles are much shorter and describe the existence here of a compact, irregular
lattice of tubular elements. Within this there appear first the small pectin vesicles
which by their growth and fusion achieve the separation of the daughter cells. At
the centre of the plate (cp) this has already happened, but at the margins, the cells
are still not divided.

spindle. As they reach this level, they reticulate very much as they do at
the cell surface (Fig. 9). There results from this a tri-dimensional inter-
mingling of tubular elements from the two daughter protoplasts and it is

THE ENDOPLASMIC RETICULUM

H9

within this lattice that the so-called pectin vesicles appear (Figs. lo and
ii). These increase in number and size from the centre of the plate out-
wards, and finally fuse to complete the separation of the two cells [45]. It
is not possible at this time to account for the origin of the vesicles, but it
would seem that small, restricted loci in the spindle matrix are lysed or
otherwise changed and enclosed bv a membrane — possiblv bv virtue of this

^'W^"'"!^' W

pv

1 N

(^

ph ^

-^ ■ ^

er-—

I

cw

11

Fig. 1 1. This shows, at higher magnifications, the right-hand margin of the cell
plate in Fig. 10. The region of the phragmoplast is indicated at ph. Within the
lattice of tubules, which marks the phragmoplast and the advancing margin of the
plate, the pectin vesicles {pv) first appear. To the left, in this image and toward
the centre of the plate, these have already fused. The outer surface and wall of the
cell are at cic.

change. The important thing for present purposes is to note the very
intimate association of this ER or ER derivative with the events of plate
and new wall formation. Whether the small marginal units of the ER
contribute enzyme or substrate to the biochemical events leading to pectin
and /or early cellulose formation cannot be decided from evidence available.
One may suggest, however, that the results of any participation which the
ER may take in plate and primary wall formation should reflect any
patterns or irregularities in the distribution of elements in the ER.

The intimacy shown here between the ER and the cell surface is again

pp

i^

f-

m

cw

'^12

Fu;. 12. This is identified as the external margin of an epidermal cell of the
onion root tip. The nucleus is at A^, a nuclear envelope at ne. Organelles of the
cytoplasm include mitochondria (ni), dictyosomes (d), proplastids (pp), and
endoplasmic reticulum (er). The surface of the protoplast and its relation to the
extracellular wall (nu) are the points of major interest. It appears that the wall has
been put down in layers, reflecting, possibly, phases of wall formation. The surface
of the protoplast in this instance shows no continuous plasma membrane but only
short fragments. In places (arrow), the ends of ER elements project beyond the
surface defined by these membrane fragments as though blending with the extra-
cellular material. We interpret this image as descriptive of a process of wall
formation in which a cortical layer of the cell, with fragments of the contained
organelles, is externalized by the development of a new plasma membrane behind it.

■ ^5

-er

THE f-XnOFI ASMIC RKTICII.UM I5I

J

.cw

■J %^

^^y

.:^ — er

.;■ -' ■ C

>

" s cw

\\- . -
I

■^: H^

.-v-

\

"^ ^^ X

-er

N^

I v

\

xw

\

-er

^^

V.

FiG. 13. This micrograph shows a longitudinal section through a protoxylem
cell of the onion root tip. This cell, in the final stages of its differentiation, shows
secondary ring thickenings in the wall characteristic of early xylem units. Cross
sections of these thickenings are marked civ. The point of interest here is that the
ER is patterned with respect to these thickenings ; the evident profiles are marked er.

152 KEITH R. PORTER

evident during secondary wall formation (Fig. 12). As before, the ER is
represented by small profiles at the cell surface, a reflection of the smaller
units comprising the system at this level. At a phase in active wall forma-
tion, presumably depicted in Fig. 12, a layer of the cell cortex is externalized
and the peripheral elements of the ER with it. In this process of ecdysis, as
pointed out in the figure-legend, a new plasma membrane appears to form
back of the margin. Again the morphological involvement of the ER is
taken to suggest functional involvement and one would gather from the
micrographs that parts of the ER are used up in the process [46].

If now these peripheral elements are functionally significant, let us say
in cellulose polymerization, any patterns in their distribution at the surface
should find some reflection perhaps in the thickness of the wall. One can
therefore with profit and interest explore cells where one finds ring or
spiral thickenings, as in primitive tracheids. These represent the final stage
in the difl^erentiation of these cells and they contain only a small residue of
cytoplasm. This and its properties make fixation diflicult. Nonetheless
when, after sectioning, these cells are examined, it is found that residual
elements of the ER are distributed in a pattern relative to the secondary
thickenings (Fig. 13). Earlier stages in the formation of these rings have
not been observed as yet, so the character of the ER which anticipates their
location cannot be depicted or commented upon.

The question of organization in the ER and its role, if any, in the
determination of cell form and polarity and the disposition of secondary
products of differentiation, is of course a large one and not to be solved
with a few electron micrographs. It does seem, however, that the form the
E.R adopts is determined in part at least by the molecular architecture of
the system itself and more remotely by genie information. In determining
through its morphology the spatial intracellular disposition of biochemical
reactions, it becomes an instrument for the nuclear control of these
phenomena.

References

1. Porter, K. R., Claude, A., and Fullam, E.,7. exp. Med. 81, 161 (1945).

2. Hartmann, J. F.,jf. comp. Neurol. 99, 201 (1953).

3. Watson, M. L.,^. biophys. biocheni. Cytol. i, 257 (1955).

4. Porter, K. R., J. exp. Med. 97, 727 (1953).

5. Hagunau, F., Int. Rev. Cytol. 7, 426 (1958).

6. Palade, G. E.,^ biophys. biocheni. Cytol. 2 suppl., 85 (1956).

7. Porter, K. R., in "The Cell", eds. J. Brachet and A. E. Mirsky, Vol. II.
Academic Press. 621 (1961).

8. Brachet, J., Arch. Biol. 53, 207 (1941).

9. Caspersson, T., " Cell Growth and Cell Function." New York, Norton (1950).
10. Garnier, Ch., Contribution a I'etude de la structure et du functionnement des

cellules glandulaires sereuses. Du role de Tergastoplasm dans la secretion.
Thesis, Nancy, No. 50 (1899).

THE ENDOPLASMIC RETICULUM 1 53

11. Mathews, A.,jf. Morphol. 15 suppl., 171 (1899).

12. Claude, A., Cold Spr. Harb. Symp. quant. Biol. 9, 263 (1941).

13. Claude, A., "Biological Symposium." The Jacques Cattell Press, Lancaster,
Pa. 10, 111-130 (1943)-

14. Palade, G. E., and Siekevitz, P., J. biophys. biochem. Cytol. 2, 171 (1956).

15. Hultin, T., Exp. Cell Res. 1, 376 (1950).

16. Borsook, H., Deasy, C. L., Haagen-Smit, A. J., Keighly, G., and Lowy, P. H.,
y. biol. Chem. 187, 839 (1950).

i6a. Littlefield, J. W., Keller, E.B, Gross, J., and Zamecnik, P.C. (i955)-7- Biol.
Chem. 8, 347.

17. Siekevitz, P., and Palade, G. E., J. biophys. biochem. Cytol. 7, 619 (i960).

18. Siekevitz, P., and Palade, G. E., J', biophys. biochem. Cytol. 7, 631 (i960).

19. Palade, G. E., J. biophys. biochem. Cytol. I, 59 (1955).

20. Weiss, J. ^l.,y. expl. Med. 98, 607 (i953)-

21. Siekevitz, P., and Palade, G. E.,jf. biophys. biochem. Cytol. 4, 203 (1958).

22. Siekevitz, P., and Palade, G. E., J. biophys. biochem. Cytol. 4, 309 (1958).

23\ Palade, G. E., "Symposium on Electron Microscopy." Brit. Anat. Assoc,
Arnold Press, London (i960).

24. Porter, K. R., The Harvey Lectures, Series 51, 175 (i957)- Academic Press,
New York.

25. Porter, K. R., and Yamada, E., J', biophys. biochem. Cytol. 8, 181 (i960).

26. Porter, K. R., and Palade, G. E.,y. biophys. biochem. Cytol. 3, 269 (1957).

27. Palay, S. L., in " Frontiers in Cytology," ed. S. L. Palay. Yale LTniversity Press,
New Haven, Conn. 350 (1958).

28. Christensen, A. K., and Fawcett, D. W., Anat. Rec. 136, 333 (i960).

29. Lynn, W. S., Jr., and Brown, R. H., J. biol. Cheni. 232, 1015.

29a. Ross, AL H., Pappas, G. D., Lanman, J. T., and Lind, J. (1958) J. biophys.
bioche?}i. Cytol. 4, 659.

30. Palay, S. L., and Karlin, S. ].,y. biophys. biochem. Cyt(j'. 5, 373 (1959).

31. Palay, S. L,., y. biophys. bioche?n. Cytol. 7, 391 (i960).

32. Epstein, AL A. (1957) J/', biophys. biochem. Cytol. 3, 851.

33. Hampton, J. C, Acta. Anat. 32, 262 (1958).

34. Karrer, H. E.,y. biophys. biochem. Cytol. 7, 357 (i960).

35. Peterson, M. L., The Transport of Fat in Man: A Study of Chylomicrons.
Thesis submitted for the Ph.D. degree at The Rockefeller Institute (i960).

36. Fawcett, D. 'W.,y. not. Cancer Inst. 15 suppl., 1475 (1955).

37. Yamada, E.,y. Electronmicroscopy 9, i (i960).

38. Yamada, E., and Porter, K. R.,y. biophys. biochem. Cytol., in press (1961).

39. Porter, K. R., and Bruni, C, Cancer Res. 19, 997 (1959).

40. Asford and Porter (1961).

41. Himes, AL, and Pollister, A. W., Anat. Rec. 132, 453 (1958).

42. Bingelli, M. E.,y. biophys. biochem. Cytol. 5, 143 (1959).

43. Revel, J. P., Napolitano, L., and Fawcett, D. W., J. biophys. biochem. Cytol.

8, 575 (1960)-

44. Buvat, R., and Carosso X., C R. Acad. Sci., Paris 224, 1532 (1957).
44a.Whaley, \V. G., Mollenhauer, H. H., Kephart, J. E., y. biophys. biochem.

Cytol. 5, 501 (1959)-

45. Porter, K. R., and Machado, R.,y. biophys biochem. Cytol. 7, 167 (i960).

46. Porter, K. R., and Alachado, R., European Regional Conference on Electron
Microscopy, Delft, (in press) (i960).

47. Porter, K. R., and Caulfield, J. B., " Proc. 4th Internat. Conference on
Electron Microscopy." Springer-\'erlag, Berlin, 503 (1958).

154 KEITH R. PORTER

48. Becker, W. A., Bot. Rev. 4, 446 (1938).

49. Godman, G. C, and Porter, K. R.,^ biophys. hiochem. Cytol. 8, 719 (i960).

50. Porter, K. R., and Bruni, C, Atiat. Rec. 136, 260 (i960).

Discussion

Mazia: Do you know what Dr. Palade meant when he entitled his talk "The
endosplasmic reticulum; its good fortune during the last 5 years" ?

Porter : I think he intended to show how well or fortunately the system had
lent itself to investigation.

Davis: Would the available information exclude the possibility that a major
function, perhaps more important than the synthetic function, would be that of
transport ? In lipid deposition, for example, I presume that at a time when the
animal is fed, glucose must get from the exterior to the sites of formation and
deposition, while in starvation it moves in the other direction. Is it not possible
that the endoplasmic tubes furnish means of communication, the enzymes then
being in the proper regions for the process but not necessarily part of the tubes?

Porter : I think it very unlikely that glucose or glucose-6-phosphate finds its
way into the cavities of the P3R on the uptake side of metabolism. There is, in any
case, no evidence of any continuity between pinocytotic vesicles and the ER.
Transport of lipid and of fully formed proteins is indicated and presumably the
limiting membrane functions in the sequestration of these if not in their synthesis.

Allfrey : I am no morphologist but I am brash enough to think that one can
answer this question on the penetration of substances into cells by the type of
experiment that we have been doing with frog ovocytes. If you take a frog ovocyte
and put it into a medium containing radioactive sodium and leucine and then do
an autoradiograph, you find that the substance penetrates the nucleus without an
apparent prior accumulation in the cytoplasm. We have checked this at a very early
time, and it suggests that there is a direct connection between the medium and the
membrane system around the nucleus, I would take the argument for transport
very seriously.

Porter : There are others who feel as you do about this, but where the examina-
tion of uptake has been studied with particulate materials which could be visualized
afterwards in electron micrographs, no evidence could be found for the migration
of particles from the outside into the cavities of the ER. It is conceivable that in
this particular material, which Dr. Allfrey knows better than I, there are infoldings
of the plasma membrane which reach to considerable depths into the cytoplasms
of these cells, and function in transport. It is also possible, I suppose, that the
metabolites he mentions move by unchannelled diffusion. If there is continuity
between the internal phase of the endoplasmic reticulum and the cell's exterior,
even intermittent continuity, and some uptake, it would mean that all manner of
materials of the environment would be carried to the cell's interior and even to the
surface of the nucleus. This would give the cell a substantial problem in sorting
out the useful and rejecting the unuseful. Furthermore it seems to me as rather
unlikely that a system obviously used for segregation and sequestration of products

THE ENDOPLASMIC RETICULUM 1 55

of synthesis would be designed to accept miscellaneous components of the
environment.

Jagendorf : I think there is one type of cellular synthesis which is not associated
with these membranes. Dr. Beer now at Johns Hopkins has looked at steni
collenchyma cells which continue to grow even after being secondarily thickened
and his impression is that the new cellulose is deposited very much removed from
the cell membrane.

Porter: This is a possibility, of course, and I suppose there is no reason why
cellulose precursors should not diffuse away from the cell surface before poly-
merization.

Hestrin : When you see a vesicle in the vicinity of a cellulose fibre, the possi-
bility must of course be considered that the fibre arose from the vesicle. This is not
however the only possible explanation. In the case of a bacterium Acetobacter
xylinutn, cellulose occurs, at a distance from the cells and does not appear to be in
necessary physical connection with the cells. It seems necessary to assume for this
case that cellulose was excreted in the form of a diflfusible particle and that fibre
formation occurred in the medium by a process akin to crystallization, one might
be allowed on a basis of analogy to speculate that within plant cells cellulose can
also exist in a diflfusible form.

Porter: Well, in the first instance I should have emphasized, if I failed to, that
this is strictly a morphological association which requires or asks for further
investigation. I think you may agree, however, that the association of the ER with
the cell surface is a selective one and is probably not without some significance.
W'hat it really means is, of course, not evident at this time. Now, in regard to the
other point, I recognize that cellulose polymerization or crystallization can take
place at points removed from the cell surface, but it was my impression that the
polymerization is influenced by a factor that is heat-labile and probably enzymic
and produced by the cell. One might expect it to act at the cell surface where
liberated. Am I wrong in this ?

Hestrin: No other assumption can reasonably be made.

Pinocytosis

H. HOLTER

Carhherg Laboratory, (j)peuluigen, Deiiinark

Dr. Porter's report has shown very clearly how intimately the endo-
plasmic reticulum and its morphology are related to the problems of protein
synthesis. The releyance of pinocytosis in this session is much more doubt-
ful ; but since my paper has been included here I shall try to make it as
relay ant as I can.

Perhaps I may be allowed, for the benefit of those who are not quite
familiar with the process of pinocytosis, to recapitulate very briefly its
main morphological features.

Ths term "pinocytosis", coined by Warren Lewis in 193 1 [14], was
originally intended to designate a process of active drinking by cells. In
pinocytosis, fluid is taken up discontinuously, in droplets that are engulfed
or sucked in by the cell, and the primary products of this acti\ity are fluid-
filled vacuoles or yesicles, by which a certain amount of the surrounding
medium is transferred to the interior of the cell. Pinocytosis therefore
amounts to an uptake of substances, but let me say at once that the term
"uptake" should be understood primarily in a spatial sense, as long as
the contents of the pinocytosis vesicles remain separated from the cyto-
plasm by a vacuolar membrane ; whether or not the process also amounts
to an uptake in the physiological sense, depends therefore on the later fate
of the vacuoles.

Morphologically, pinocytosis displays quite a variety in mechanism and
a very wide range of dimensions. In tissue culture cells, the classical object,
pinocytosis is brought about by the activity of membranous ruftle pseudo-
podia which, by undulating movements, form folds that enclose a certain
amount of fluid. The second variety of pinocytosis, very beautifully dis-
played by amoebae, is that of invagination (Fig. i). In this process the cell
surface forms cavities, in extreme cases long narrow tubes, from which
the pinocytosis vacuoles are pinched off^. Pinocytosis by invagination is
fairly variable with regard to the shape of the cavities formed, and extremely
variable with regard to dimensions. In amoebae the average diameter of
the primary pinocytosis vacuoles is about 1-2 /t, while in many other
objects it is only discernible by means of the electron microscope, the
diameter of the yesicles being from 100 A upwards (Fig. 2).

158 II. HOLTER

Regardless of these differences in mechanism and dimensions, the
morphologically essential feature is the same ; a certain area of the surface
membrane of the cell encloses a droplet of the surrounding medium,
separates from the surface and migrates into the cell.

Another important feature of pinocytosis is that it does not occur in
all media and all the time but is induced by the presence of certain sub-
stances in the medium. So far, the induction of pinocytosis has been
studied most extensively in amoebae [13, 9, 5, 6] with regard to proteins,
amino acids and certain salts. Typical non-inducers are, for instance, the

Fig. I. Pinocytosis channels in Amoeba proteiis, induced by i ■5"() solution of
bovine plasma albumin. (After Chapman-Andresen and Prescott [9].)

carbohydrates. Induction depends on the size of the molecule, pH and
other factors connected with electric charge. In amoebae there seems to
be a certain not very dramatic molecular specificity; but for mammalian
cells at least one case of highly specific induction of pinocytosis has been
reported, namelv the action of insulin on HeLa cells [2^] and adipose tissue

The findings regarding pinocytosis induction support a hypothesis
pronounced by Bennett [3] according to which one of the essential features
of pinocytosis consists in the adsorption of molecules or ions to the cell
surface membrane, followed by an invagination of the surface at the loaded
area, a sliding of the membrane into the invagination. The first experi-
mental support of this hypothesis came from papers by Brandt [4] and

PINOCYTOSIS 159

Schumaker [27]. Brandt induced pinocytosis in amoebae by means of a
fluorescent protein, and could directly demonstrate that a fluorescent film
is adsorbed to the surface of the amoeba and that the same film coats the
inside of the pinocytosis channels and pinocytosis vacuoles. Since then,
other examples for the surface adsorption of pinocytosis-inducers have
been found [21, 15]. With regard to the nature of the adsorbing sites
the hypothesis has been advanced by Marshall, Schumaker, and Brandt
[15] and by Chapman- Andresen [6] that they are furnished by the mucous
coat that is known to occur at the surface not only of amoebae but of many,
perhaps all, pinocytic cells. I believe that Dr. Marshall will tell you more
about this problem.

Fig. 2. Electron micrograpli ot jejcnum of 3-clay-iikl mouse (original magni-
fication X 6500) showing pinocytic vesicles of varying size originating from the
brush border. (After Clark and Wochner [9a].)

According to this well-supported hypothesis, therefore, we may assume
that the adsorption of a suitable inducer brings about a surface reaction
by which the adsorbed substance, together with some of the solvent, is
transported into the cell. This is in good agreement with the morphological
observations. The question is now: What is the physiological significance
of such a process .'' Is the physiologically predominant feature of pino-
cytosis the specific uptake of the absorbed substance, often high-molecular,
or is it the accompanying uptake of the fluid ?

This question cannot be answered in a general way on the basis of our
present knowledge. Important features of the process are certainly dif-
ferent after the induction, for instance, by salts and by proteins, and my
personal guess is that different cells in different situations will be found to
utilize the possibilities of pinocytosis for different physiological means.

l6o H. HOLTER

However this may be, it is certain that pinocytosis can be a means of
introducing certain substances, including proteins and amino acids, into
vesicles that migrate into the cell interior. The next question is then : Are
such substances available for the metabolism of the cell ? This question
is not only a figure of speech ; we must not forget that many of the sub-
stances which have been found to enter by pinocytosis, are high-molecular
compounds to which the cell membrane is assumed to be impermeable —
the same membrane, in fact, which by the mechanism of invagination
becomes the boundary of the pinocytosis vesicles. The properties of this
membrane, therefore, and its possible changes inside the cell become the
main problem in our understanding of the physiological significance of
pinocytosis.

During the migration of the vesicles from the periphery toward the
interior there occurs a rather intense dehydration which is often plainly
discernible in time-lapse movies as a shrinkage of the vacuoles. However,
just as often this process of shrinkage is counteracted or even overcom-
pensated by another frequent occurrence, namely, the fusion of the
original small vesicles to larger vacuoles, so that the actual view in the
microscope usually presents an array of vacuoles of very difl^erent sizes.
In spite of this, the actual occurrence of dehydration causing a concentra-
tion of the vacuolar contents was made rather probable by Marshall and
myself [13] by means of centrifugation experiments in which we found
that after pinocytosis the density of the pinocvtosis vacuoles, identifiable
as such by the content of a fluorescent marker, was steadily increasing.

Recently Roth [26] has published very interesting observations regard-
ing the changes occurring in pinocytosis vacuoles after ingestion. He has
shown that the vacuolar membrane displays a rather intense form of
"internal micropinocytosis" by which, as he assumes, the contents of the
primary vacuoles are distributed in the cytoplasm. Roth [26] also points
to an important fact, which perhaps had not been sufiiciently considered
by previous authors : pinocytosis, and especiallv if repeated in a second step
internally, results in an enormous increase of active internal surface available
for diffusion. Chapman-Andresen and Nilsson [8] in our laboratory
have found that this process of secondary micropinocytosis begins during
channel formation (Fig. 3).

The only morphological change of the vacuolar membrane in amoebae
that can be observed in the electron microscope is that the mucous coat
that covers the plasmalemma and after invagination forms a lining on the
inside of the pinocytosis vacuoles, disappears after some time. Miiller
and Rappay [16] have reported that the periodic acid-Schiff reaction which
is given both by the plasmalemma and the vacuolar lining, disappears
later on in the vacuole. In our laboratory [7] we studied the permeability
of pinocytic vacuoles to radioactive glucose and have obtained evidence

PINOCYTOSIS l6l

which we interpreted to indicate that the permeabihty of the vacuolar
membrane is changed after interiorization, with the effect that at least some
of the substances confined in the vacuoles can penetrate into the cytoplasm.
How this penetration occurs, and whether or not it involves some sort of
a breakdown, could not be decided by our experiments.

Thus, while there are some indications that a change in the membrane
of the pinocvtosis vacuole occurs during its migration to the interior of the

Fk;. 3. Electron micrograph of pinocvtosis channel in Amoeba proteiis,
showing micropinocytic vesicles and mitochondria close to the channel. (After
Chapman- Andresen and Nilsson [S].)

cell we can from the evidence so far discussed derive no hint as to how this
change might be brought about. I believe, however, that this question may
be answered in time, when we know more about another problem of
pinocytosis which I should now like to discuss. It concerns the relationship
between pinocytosis vesicles and certain of the cytoplasmic structures
inside the cell.

There are at present two main schools of thought regarding this
question. One connects the pinocytosis vesicles with the lysosomes, the
other with the endophisfiiic rcticith/iii, and I shall try very briefly to outline
the two views.

l62

H. HOLTER

The first important observation that ought to be mentioned in this
connection was made by Rose [25] who in a strain of HeLa cells by means
of a time-lapse film demonstrated that certain granules, which he called
" micro-kinetospheres ", actively make contact with newly ingested pino-
cytosis vacuoles, and fuse with them. He assumes that by this process
certain enzymes are injected into the vesicles, and that this is a necessary
condition for the later shrinkage and change in refraction index of the
vesicles which he observes. Rose himself suggests that the microkineto-
spheres might be identical with the lysosomes.

Alkaline
Phosphata

lII!'''^''^1ilBlFS'^'"'"''''''"'^

^^/cf

•.'oxidative •■/TTiVa' • ■ ■" • • • • . /77\

pM05P«OBYl.ATI0N,*vV^ '^^ . ' •..i?*^^^^*^^' <^ . V^^

Fig. 4. Diagrammatic representation of a cell in the proximal convolution of
the rat kidney. (After NovikofF [18].) Micropinocytosis vacuoles (p) are shown
forming at the ends of the canalicular structures (c) extending into the cell between
adjacent microvilli (mv) of the brush border. These are shown fusing into larger
vacuoles (v), which are transformed into lysosomes (L). Other abbreviations are:
A, oxidized substrate; ADP, adenosine diphosphate; AH2, reduced substrate;
Al'P, adenosine triphosphate; ATPase, adenosine triphosphatase; BM, basement
membrane; E, endothelial cell of blood capillary; GA, Golgi apparatus; N,
nucleus; and Pi, orthophosphate.

Lysosomes were detected and named by deDuve and his coworkers
[10] who have shown them to be the carriers of a whole array of
hydrolytic enzymes and who assume that they are the main instruments
of intracellular digestion. Rose's interpretation, that pre-existing lysosomes
by fusion convey digestive enzymes to the pinocytic vesicles is certainly a
very tempting one. It conflicts, however, to a certain degree with another
point of view according to which the pinocytic vesicles are transformed
into lysosomes. This point of view is held by Straus [29] who studied the
uptake of peroxidase, a protein marked by its enzymic activity, and des-

PINOCYTOSIS 163

cribed its "segregation" into granules which he called phagosomes.
Recently, [17, 18] Xovikoff has devoted himself to the question of a
possible connection between pinocytosis and lysosomes. His hypothesis
is summarized in the diagram shown in Fig. 4, which he has kindly per-
mitted me to use. According to this interpretation the lysosomes are
derived from pinocytosis vacuoles and are regarded as the end products
of their transformation inside the cell. The exact nature of this assumed
transformation is not clear, and especially not the way in which the outfit
of hydrolytic enzymes which defines the lysosomes, should be acquired.
One interesting suggestion is based on the fact that at least one of the
characteristic enzymes of lysosomes, namely acid phosphatase, has been
found by Novikoff [18] to occur in those regions of the cell membrane of
Amoeba proteiis, to which food particles attach before engulfment. This
might indicate that an essential feature of the formation of lysosomes is the
transport of surface enzymes into the cell interior by means of pinocytosis,
and this would of course tie in verv nicely with Bennett's views on mem-
brane flow from the surface into the cell.

Closely connected with the problem of a relation between pinocytosis
vacuoles and lysosomes is probably the similar claim of a connection
between pinocytosis vacuoles and mitochondria. This claim was put
forward by Gey and his coworkers [11] and has recently been revived by
Robineaux and Pinet [24] on the basis of their interference-microkine-
photographic investigations of protein uptake by macrophages. It seems to
me that the question is in reality one of being able morphologically to
differentiate between lysosomes and mitochondria, which may be difficult
without the aid of the electron microscope and cytochemical technique.

The other school of thought links the fate of the pinocytic vesicles to
the endoplasmic reticulum. This goes back to the work of Palade [19] who
claimed the continuitv of the cell surface with the membrane system of
the reticulum. I should like to present to you this view by means of
another schematic diagram, taken with the author's kind permission, from
Siekevitz's [28] Ciba lecture (Fig. 5).

The essential part of the diagram for us at the moment is that depicting
the migration of pinocvtic vesicles toward, and fusion with, the intracellular
spaces of the endoplasmic reticulum. This process is an important feature
of Bennett's [3] and Palade's [19] \iews regarding the connection between
surface membrane and the endoplasmic reticulum, but its reality was not
strictly proven and it is therefore also marked off" as hypothetical in
Siekevitz's diagram. Very recently, howe\"er, Palay [20] has described an
example of the process. He studied fat absorption by the cells of the in-
testinal wall of rats, and found that fat droplets that enter these cells by
means of pinocytic vesicles could be observed again as inclusions in the
perinuclear space. If this observation can be shown to be valid (and we

164 H. HOLTER

have learned from Dr. Porter's review that other interpretations are
possible) then it would certainly seem to indicate the existence of a
spatial continuity between pinocy tic vesicles and the endoplasmic reticulum.
In Siekevitz's diagram two alternatives are given for the fate of the
pinocytic vesicles. One is their fusion with the channels of the endoplasmic
reticulum, discussed above, the other is their disappearance into the
general cytoplasmic matrix. As we have seen, a third possibility would be
their transformation into granular cell organelles, perhaps of the lysosome
type. At present we cannot decide if only one of these possibilities is
actually realized or if several processes operate at the same time, perhaps
with shifting preponderance, according to the metabolic needs of the

Fig. 5. Stylized representation of a cell, showing a channel of the endo-
plasmic reticulum (er) and its relation to the perinuclear space and pinocytic and
secretory processes. (After Siekevitz [28].)

cell. All possibilities are highly interesting, but today the situation is not
ripe for too-detailed speculation.

There is, however, another aspect of the matter that I should like to
mention. It is well known that at least in amoeboid cells the process of
pinocytosis has its counterpart in the vacuolar discharge of material. This
is realized for instance in the filling up of the contractile vacuole [22] by
fluid-filled vesicles that fuse with the main vacuole, and also, with certain
morphological modifications, in the process of defaecation (Fig. 6). One
might say that pinocytosis works both ways, and it has also been necessary
to assume processes of inverted pinocytosis in several cases where pino-
cytosis was considered to mediate not orily the introduction of substances
into cells but also the transport of substances t/irougli cells. As depicted in

PINOCYTOSIS 165

Sieke\itz's diagram (Fig. 5) and as discussed in Dr. Porter's review, a
similar process is generally assumed to occur in the discharge of secretion
products formed in the endoplasmic reticulum. It would seem that the
endoplasmic reticulum, if it reallv can be shown to torm the link between
the ingestion and secretion pathways of the cell, might provide a mechanism
that could explain the selectivity of pinocytosis.

With regard to this problem of selecti\ity we are still very much in the
dark. It is true that the process of surface adsorption prior to pinocytosis
affords a specific enrichment of the adsorbed solute in the pinocytosis
vacuoles. But this can cover only part of the problem. So far no one has
been able to give a satisfactory answer to the question : What happens to

v-.^..

f-€5vti).'

Fig. 6. Defaecation in amoebae. Two instances of defaecation. Above: food
remnants; below: cluster of crystals. In both instances the vacuole is seen at the
right, brought close to the surface in the tip of a short pseudopod. At the left, the
objects have been expelled by rupture of the membrane and lie free in the medium.

(After Andresen and Holter [i].)

substances which are dissohed in the fluid ingested together with the
surface membrane and which are of no use, or even toxic, to the cell .'*
There must be some opportunity for a highly specific process of discrimina-
tion somewhere, but it has not seemed very satisfactory to explain any
high degree of discrimination on the basis of mere changes in the permea-
bility of the vacuolar membrane. For this it seems necessary to assume the
action of some sort of active transport mechanism in the interior of the
cell, situated either at the membranes of the pinocytosis vesicles, or
perhaps in the membrane system of the endoplasmic reticulum. Again the
experimental evidence at the present time does not justify more than very
tentative speculation on this important point.

l66 H. HOLTER

The trouble is that most of the experimental work on pinocytosis so
far has been done with amoebae, and in these organisms very little is
known about an endoplasmic reticulum, if indeed there exists a recog-
nizable equivalent of this organelle in the amoeba's constantly streaming
cytoplasm. What we need is an experimental study of pinocytosis in some
mammalian cells of well-known submicroscopic anatomy.

I should like to wind up by a short, but perhaps not wholly superfluous,
discussion of terminological problems.

There is no doubt that pinocytosis is a word that at present enjoys a
certain popular appeal, in marked contrast to the situation only a few years
ago. Once in a while one cannot help being afraid that the label "pino-
cytosis" is applied without much regard to definition in many instances in
which an actual evidence simply consists in the observation of vesicles of
unknown origin. On the other hand many new names are being created —
too many to mention here — and the revival of the Greek language has
been quite considerable in an attempt to describe more or less specialized
instances of phenomena that seem to be quite adequately covered by
the two oldest and most widespread terms, namely phagocytosis and
pinocytosis.

However, the real issue is whether or not even these two terms should
be maintained as separate concepts. Personally I have repeatedly expressed
the view [12] that there is no sharp delimitation between phagocytosis and
pinocytosis, since the main difl^erence seems not to be the mechanism of
the process, but only the nature of the ingested material. The results of
most investigations in the last years certainly seem to support the tendency
of minimizing the difference between the various forms of surface invagina-
tion.

I have tried to present to you the present state of the pinocytosis
problem, as it stands. It is a rapidly developing field, full of contradictions
and unsolved questions. But nevertheless I feel that some progress has
been made in the last years, and that some of the perspectives that have
been opened up are of great interest for all branches of cell biology and
therefore, in an indirect way perhaps, also relevant to the topics of this
symposium.

References

1. Andresen, N., and Ilolter, H., C. R. Lab. Carlsberg Ser. c/iini. 25, 107-146

(1944)-

2. Barrnett, R. J., and Ball, E. Cj., 7. biop/iys. bioc/wtii. Cytol. 8, 83-101 (i960).

3. Bennett, S.,X biophys. biochem. Cytol. 2, Part 2 Siippl. 99-103 (1956).

4. Brandt, P. W., Exp. Cell Res. 15, 300-313 (1958).

5. Chapman-Andresen, C, C. R. Lab. Carhberg. 31, 77-92 (1958).

6. Chapman-Andresen, C, "Proceedings of Symposium of Society for Cell
Biology", Paris (i960).

PINOCYTOSIS 167

7. Chapman-Andresen, C, and Holter, H., Exp. Cell. Res. Suppl. 3, 52-63 (1955).

8. Chapman-Andresen, C, and Xilsson, J. R., Exp. Cell Res. 19, 631-633 (i960).

9. Chapman-Andresen, C, and Prescott, D. M., C. R. Lab. Carhberg, Ser. cliirn.
30, 57-78 (1956).

9a. Clark, S. L., Jr., and Wochner, D., Anat. Rec. 130, 286 (1958).

10. deDuve, Chr., in " SubceUular Particles", ed. T. Hayashi. Ronald Press, New-
York, 128-159 (1959)-

1 1. Gey, G. O., Shapras, P., and Barysko, E., Ann. X.Y. Aead. Sci. 58, 1089 (1954).

12. Holter, H., in " Internation. Review of Cytology", Vol. 8, ed. G. H. Bourne
and J. F. Danielli. Academic Press, New York, 481-504 (1959).

13. Holter, H., and Marshall, J. M., Jr., C. R. Lob. Carhberg Ser chini. 29, 7-26

(1954)-

14. Lewis, W. ¥{.,yoluis Hopk. Hasp. Bull. 49, 17-28 (1931).

15. Marshall, J. M., Jr., Schumaker, V. N., and Brandt, P. W., Ann. \\Y. Acad.
Sci. 78, 515-523 (1958).

16. Miiller, M., and Rappay, G., Mag. Tud. Akad. Biol. Csoport.Kozl. 3, 81-86

(1959)-

17. NovikofF, .A.. B., /// "Developing Cell Systems and Their Control", ed. D.
Rudnick. Ronald Press, New York (i960).

18. Xovikoff, A. B., /;/ "Biology of Pyelonephritis", ed. K. Quinn and E. Kass.
Little, Brown and Co., Boston (i960).

19. Palade, G. E.,y. biaphys. biochem. Cytal. 2, Siippl. 85 (1956).

20. Palay, S. h.,y. biophys. biochem. Cytol. 7, 391-392 (i960).

21. Pappas, G. D., 10/// Int. Congr. Cell Biol. 94-95 (i960).

22. Pappas, G. D., and Brandt, P. \V., J. biophys. biochem. Cytol. 4, 485-488
(1958).

23. Paul, ].,J. biophys. biochem. Cytol. 8, 83-101 (i960).

24. Robineau.x, R., and Pinet, J., Ciba Eoundation Symposium on Cellular Aspects
of Immunity I, 5 (1959).

25. Rose, G. Cj.,y. biophys. biochem. Cytol. 3, 697-704 (1957).

26. Roth, L. E.,^. Protozoology 7, 176-185 (i960).

27. Schumaker, V. N., Exp. Cell. Res. 15, 314 (1958).

28. Siekevitz, P., Ciba Foundatioii Symposiimi on Cell Metabolism 17-49 (1959).

29. Straus, W.,y. bi(jphys. biochem. Cytol. 4, 541-550 (1958).

Discussion

DoRF.MAN : I would like to mention one thing that may have some connection,
somewhat far fetched, relating to the biosynthesis of the membrane, and that is in
the case of our study of the synthesis of polyuronic acid in bacteria. We have been
able to localize the enzyme of the final polymerization in the cells of the protoplast
membrane, which can be prepared pure. There is no RX.A. left in them and thus
this membrane is serving a function similar to the endoplasmic reticulum in the
cells that Dr. Holter suggested. I just wonder whether in this question of transport
if it isn't only the final polymerization of the macromolecules that is occurring
at this anatomically localized site because all the other enzymic activities required
to forni the viridine nucleotides are all perfectly soluble and removed. It is only the
final polymerizing reaction that is localized to these membranes.

Campbell: I would like to ask about the possibility o j-'^hether the lysosomes
are derived from the pinocytotic vesicles. I think it is shown from the work of

1 68 H. HOLTER

Coons that proteins do certainly get inside the cell intact. Now if they go in by
pinocytosis then we know also from the work of Simpson and Steinberg in the
States, that the breakdown of these proteins takes place by perhaps a reversal of
protein synthesis, that is to say not by proteolytic enzymes. As the lysosomes are
suicide squads containing proteolytic enzymes, it seems unlikely that the intact
proteins are broken down in the lysosomes. On this basis the lysosomes would not
be derived from pinocytotic vesicles.

HoLTER : I quite agree and I think it is one of the main objections to the whole
lysosome theory, but I just wanted to present the views that have been suggested.
What we really need to do is to reconcile the vacuolar membrane with the evidence
for uptake of high molecular svibstances which have been claimed to enter the
cytoplasm even without loss of their antigenic properties. To explain this we need
to get rid of the membrane barrier somehow. There are people who just assume
that the membrane of the pinocytic vesicles disappears, so that their contents are
released into the cytoplasm. But clear evidence for this is lacking. In amoebae we
have never been able to see the membrane disappear and I understand from Dr.
Porter that also the other instances in which such a membrane has been reported
to disappear are doubtful on technical grounds. So this is one of the many problems
that are still open.

Davis: I would like to mention an observation which may possibly throw light
on this problem. At Harvard Medical School Manfred Karnovsky has been study-
ing the metabolic effects of phagocytosis, and has found that the process of engulfing
particles markedly stimulates the rate of incorporation of phosphate into phos-
pholipid. This suggests that the material of the membrane is turning over.

HoLTER : Yes, I know that work, and I would also like to mention that, if I
remember correctly, the author regards the resynthesis of membrane as the limiting
factor for the duration of the phagocytic activity. Furthermore, this paper is the
only one I know of where the energetics of such processes have been considered.
It is work of that kind that we badly need also in pinocytosis investigations.

LooMis: I was wondering if the disappearance of the nuclear membrane during
mitosis casts any light on how membranes can disappear and reappear.

Holter: I don't know. Do you ?

Porter: Well, the observations available suggest that it simply fragments into
a lot of smaller vesicles without any decrease or increase in the total area of the
membrane involved.

The Ergastoplasm in the Mammary Gland
and Its Tumours:
An Electron Microscope Study with Special Reference to
Caspersson's and Santesson's A and B Cells

F. Haguenau and K. H. Hollmann

Lahoyatoire de Mcdecine Experiniciita/e dii College de France,
Paris, France

To all cytologists nowadays, the ergastoplasm is a well-defined entity,
the ultrastructure of which may be considered as characteristic ([i, 2], and
Porter this volume). Its particular importance lies in the fact that it
corresponds to the cytoplasmic nucleic acid (RNA) and is directly involved
in protein synthesis. In spite of much knowledge acquired during the last
few years the meaning of some ultrastructural aspects of ergastoplasm still
escapes our understanding. It is our purpose, by taking the example of
the mammary gland, to comment upon these aspects because though they
are mentioned and acknowledged by many they ha\e not yet sufficiently
been put forward.

At the onset we wish very briefly to go back to the question of nomen-
clature. Contrary to what might have been expected from the confused
situation which reigned at the beginning, agreement has been reached
between cytologists and, most of the time, the same terms are being used
now to design the same constituents :

Ergastoplasm: System of intracytoplasmic membranes characterized by
the presence of granules attached to its outer surface.

''Organized'' Ergastoplasm: Such a system when abundant and more or
less arranged in parallel array.

''Endoplasmic Reticulum'': Membrane component of this system (con-
tinuous with other intracytoplasmic membranous organelles in the cell
especially the Golgi apparatus).

Ribonucleoprotein granules (RNP-granules, " Ribosomes ") : The granular
component of ergastoplasm either membrane-attached or isolated in the
ground cytoplasm. The following diagram will be helpful :

170

F. HAGUENAU AND K. H. HOLLMANN

trgastoplasm
(organized)

Endoplasmic
reticulum

RNP-granules
'nbosomes'

Fig. I.

It is the relationship, if any, between one form and the other, in par-
ticular that of "organized ergastoplasm" to RNP-granules which is not
clear and that we will now consider in the mammary gland and specially
in its cancers.

I. Ergastoplasm in the normal mammary gland

Classic cytologists long ago discovered ergastoplasm in the mammary
gland and described the strongly basophilic filaments present in great
amount at the basis of the secretory cell [3, 4] and Fig. 2(0). This baso-
philia disappears under the action of ribonuclease and the reaction is
greater during the first phase of the secretory cycle. It then decreases to
appear again with reconstitution of the alveolar cells. The various changes
occurring in ergastoplasm are also easily followed in the fluorescent
microscope after acridine orange staining.

In the electron microscope, development of the ergastoplasm during
the lactation period is remarkable (Fig. 2(b)). Its order and ultrastructure
are characteristic. Ribonucleoprotein particles are regularly arranged at
the surface of the membrane system of the endoplasmic reticulum which
itself is organized in parallel lamellae (Figs. 3 and 4). This "organization"
in parallel array is one of the essential features of the ergastoplasm at that
stage. Only when ergastoplasm is found in this "organized" form does
elaboration of milk occur.

The ultrastructural details of the secretion have been described in two
recent studies in the rat [5] and in the mouse [6].

It has been shown that milk secretion in these species is made up of at
least two morphologically distinct elements probably corresponding to
protein secretion for the one and to lipid secretion for the other. The topo-
graphical relationship between lipid secretion and definite cell organelles
is still under discussion while on the contrary, formation of protein granules
appears clearly related to the Golgi complex. Indeed it is exclusively in

THE ERGASTOPLASlM IN THE MAMMARY GLAND AND ITS TUMOURS 17I

the Golgi vacuoles that these granules are found before they break out
into the granular lumen (see Fig. S).

The part plaved bv the ergastoplasm in the elaboration of these protein
granules is not clear [^, 6]. All our present knowledge, however, indicates
that its role is fundamental.

This interpretation is based on : (i) the continuity which exists between
ergastoplasm and Golgi apparatus, the whole membrane system of w^hich
constitutes the endoplasmic reticulum; (2) the enzymic studies of the

Fig. 2(a). Lactating mammary gland. Aspect of ergastoplasm in the light
microscope. .After toluidine blue staining it appears as dense basophilic masses
( -) X 1600.

Rockefeller School in particular, which have shown that in the pancreas
the same enzymes were present in the RNP-granules and in the final
secretion granule [7a].

The first fact, the continuity of the membrane svstem is established
now in a suflicient number of different tvpes of mammalian cells to be
considered as valid for all types of cells ([i], and Porter in this symposium).

The second fact is not yet biochemicallv demonstrated for glandular
organs other than pancreas but the similarity of the ultrastructural pattern
of all secretory cells studied up to now is strong indication that comparable
mechanisms are invoked in all cases.

On morphological grounds on the other hand, evidence is not easy to

172

F. HAGUENAU AND K. H. HOLLMANN

Fig. 2(b). Electron microscope aspect of a lactating mammary gland at the
peak of its activity. At right glandular lumen filled with secretion product (nunierous
small protein granules and large lipid droplets). Note the typical parallel arrange-
ment of the "organized" ergastoplasm in the cytoplasm of the glandular cells. In
contrast, the myoepithelial cell (cm), not concerned with secretion, contains no
ergastoplasm. x 7100.

THE ERGASTOPLASM IN THE MAMMARY GLAND AND ITS TUMOURS 1 73

*, ■%

-^

Fig. 3. Higher magnification of ergastoplasm in the cell of a lactating mammary
gland. "Organized" ergastoplasm is typically made of RNP-granules attached to
the surface of the endoplasmic reticulum, N = nucleus, x 51 500.

174

F. HAGUENAU AND K. H. HOLLMANN

s '^.

p

\^^ ^VTfAifl^ ^^^^

\

yC^

* /--^

Fig. 4. Another aspect of ergastoplasm in the cell of a lactating mammary
gland, to demonstrate parallel array of " organizei.!" ergastoplasm. Here fixation
with KMnOi has not preserved the RNP-granulei. x 34 000.

THE ERGASTOPLASM IN THE MAMMARY GLAND AND ITS TUMOURS 1 75

provide. It must be pointed out that even in the case of the pancreas, the
morphological demonstration of a relation between ergastoplasm and
protein granules is restricted to the one example of guinea-pig pancreas [7].

Likewise in the mammary gland, aspects showing this relationship are
exceptional. It is in one pathological case only that an example could be
found where protein granules had accumulated in the lamellar system of
the ergastoplasm (Fig. 5). When such images are found however they lead
to the belief that protein elaboration may take place in the RNP-granules
bordering the membrane of the endoplasmic reticulum and that secreted
products then proceed towards the Golgi apparatus where they will take
on their final form. Siekevitz [8, 9] has elegantly developed the matter.

When the mammary gland enters the involution phase which follows
its active secretory phase, the reduction of the alveolar trunk is spectacular
and similar to that of the virgin gland. Its ultrastructural counterpart,
likewise is striking. The rare persisting ducts or buds are lined with mono-
tonous-looking cells. The nuclei are shrunken, the cytoplasm clear, hardly
containing under-developed organelles. The ergastoplasm in particular is
now only represented by rare dispersed RNP-granules or isolated dotted
membranes. Not only is this decrease quantitative but from the special
ultrastructural point of view with which we are concerned, the ergasto-
plasm has not retained the lamellar organization which was characteristic
of active secretion and under this form it has disappeared together with
the secretion granules themselves (Fig. 6).

Though this remarkable arrangement of the ergastoplasm has always
been noted and described, its absolute relation to actual secretory activity
has not been specially emphasized. Yet it is striking. If one creates con-
ditions in which the physiological secretion is modified as under the
influence of hormones in the case of the mammary gland this organization
typical at first (Fig. 2) does not persist in the absence of sucking and is
disrupted after prolonged stimulation (Fig. 8).

2. The ergastoplasm in cancer of the mammary gland

The bond between normal secretion and presence of parallel-organized
ergastoplasm is further demonstrated if tumours are studied.

It is most striking that in over 100 cases of liiiman breast cancer studied
in the electron microscope [10, 11] none has been found where ergasto-
plasm organized in parallel arrav was present in noticeable amount.
There has been one exception to the rule onlv (Fig. 9), but in this case so
many other features were aberrant [12] that it cannot be taken int(^ account
in the present discussion. The rarity of ergastoplasm under this form is
here again linked to absence of the normal physiological secretion.
Other secretion products may be found in these cancerous cells such as

176

F. HAGUENAU AND K. H. HOLLMANN

Fig. 5. Protein granules in the cisternae of ergastoplasni in a human cancer
cell. This is a very rare finding, x 36 000.

THE ERGASTOPLASM IN THE MAMMARY GLAND AND ITS TUMOURS 1 77

V

ff

-■ ■'iLS--A

.^

¥iG. 6. Typical aspect of virgin gland cells in the mouse to show the absence
of "organized" ergastoplasm and even the rarity of RNP-granules. In the upper
part of the picture is the lumen whilst at the base of the glandular cells is a myo-
epithelial cell. X 10 100.

178

F. HAGUENAU AND K. H. HOLLMANN

Fig. 7. Low power view of a whole acinvis in a virgin mammary gland after
48 days oestrogen stimulation. "Organized" ergastoplasm is still present but lipidic
storage is conspicuous, x 7000.

THE ERGASTOPLASM IN THE MAMMARY GLAND AND ITS TUMOURS 1 79

m^ 'fc

•

f'

#.

*t

'■^

^

• ,• f

•

Fig. 8. Similar experimental conditions but at a later stage. "Organized"
ergastoplasm has been disrupted and only RNP-granules are present. Note the
typical localization of protein granules in Golgi vacuoles, x 28 250.

i8o

■*f / .»*►

'••i ^

F. HAGUENAU AND K. H. HOLLMANN

* St

.^•,-

(% ^f

' ;^ t?«

^j

.rt#^

Fig. 9. Large amount of "organized" ergastoplasm in one case of human
cancer, a rare finding. The granular mass at bottom right corresponds to a tangen-
tial section; only the surface RNP-granules have been concerned by it, while
transverse sections show both the membrane and granular constituents of ergasto-
plasm. Intermediate aspects are easily followed in this picture, x 26 000.

THE ERGASTOPLASM IN THE MAMMARY GLAND AND ITS TUMOURS l8l

>■■■ * *< f

H.

Fig. io. Human mammary carcinoma. An example of one of the most frequent
aspects. The cytoplasm is almost devoid of ergastoplasm and other organelles.
These cells correspond to the B type of Caspersson and Santesson. x 3950.

182 F. HAGUENAU AND K. H. HOLLMANN

lipid droplets for instance but these are not accompanied by the typi-
cal arrangement referred to here. On the whole, ergastoplasm in human
breast cancer is not abundant and is not present in its "organized" form
(Fig. lo).

This description, however, would lead to an erroneous representation
were it not completed by another important feature directly related to the
problem of nucleic acids in cancer cells. This concerns the A and B
types of cells described by Caspersson and Santesson [13] in the course of
their studies with ultraviolet absorption technique. These two basic types
of cells were characterized by, the A cell "a well-developed cytoplasmic
protein-forming system", the B cell, "a nucleolar apparatus showing
signs of intense function but practically none of protein or nucleic acid
formation in the cytoplasm" (Caspersson [14], pp. 142-145) (Fig. 11).

Other authors have confirmed this description though often using a
different appellation for the two types of cells [15, 16, 17, 18]. Their exis-
tence is important since a relation could be established between the type
of cells present in tumours and sensitivity to radiation [15, 17, 18, 19, 20,
21 , 22]. The consequences of this from a prognostic as well as a therapeutic
point of view would justifv greater development but would lead us into
the whole chapter of the relation between irradiation, growth, differentia-
tion, ageing and death of the cell.

What matters to us here is that in the electron microscope, two types
of cells corresponding roughly to those originally described by Caspersson
and Santesson may be readily observed [10, 11] (Fig. 12).

Though a simultaneous spectrophotometric and electron microscopic
study of the same cells has still not been achieved, parallel cytochemical
studies (methyl-green-pyronin and fluorescent acridine orange stains)
leave little place for doubt on the subject.

Though A cells are formed in all types of cancer studied in the electron
microscope so far, they are particularly noticeable in breast cancers. Not
all of these contain them nor are they present in cancers only. They can
be observed in hormone-stimulated glands or in benign tumours. In this
case their number is always small and their localization at the periphery
of a lobule. This corresponds to the typical localization noticed by
Caspersson and Santesson. In breast cancers, however, not only can their
number be greatly increased but the peripheral arrangement is no longer
respected.

In the electron microscope the morphological characteristics of A and

Fig. II. A and B cells in human mammary cancers as seen in the ultraviolet
microscope. The difference of absorption at 2570 A wavelength is very obvious
between the two types of cells, x 770 and x 550.

(By courtesy of Prof. T. O. Caspersson)

THE ERGASTOPLASM IN THE MAMMARY r.T.AXn AND IT? Tr:\IOURS 183

1 84

F. HAGUENAU AND K. H. HOLLMANN

Fig. 12. Low power view of a human cancer rich in A and B cells to show their
intrication. The B cells have a clear cytoplasm with almost no ergastoplasm, and
a voluminous spherical-shaped nucleus with prominent nucleolus. The A cells
contain a much denser nucleus with festooned limits. Ergastoplasm is abundant
and totally disorganized. In the particular example here vacuolation is marked.
This aspect possibly corresponds to necrosis, x 4300.

THE ERGASTOPLASM IN THE MAMMARY GLAND AND ITS TUMOURS 1 85

B cells correspond to Caspersson and Santesson's original description, and
refinement of criteria allows better detection, better study of their repar-
tition and lays stress on the importance of all intermediate aspects be-
tween the two A and B extreme types (Transitional T cells of Gusberg
et al. [18] Towell [22] (Fig. 14).

Neither the B (Fig. 10) nor the intermediate (Fig. 14) type of cell
which constitute the bulk of the tumour will concern us in the following
description because they have been referred to in the first part of this
paper and it has been shown that their ergastoplasm is not particularly
developed and not typically organized.

The A type of cell, on the contrary will be described in detail because
of its richness in ergastoplasm which corresponds to the high content of
cvtoplasmic RNA found with the ultraviolet absorption technique. Their
ultrastructural appearance is striking (Figs. 12, 13, 15, 16). Usually
smaller than adjacent cells they are often sharply edged, and when not
at the periphery, wedged in between the other cells. They contrast with
these by the remarkable density of both their nucleus and their cytoplasm.

The nucleus is characteristic, its chromatin is coarse and mottled, its
limits festooned. In the cytoplasm, the ergastoplasm is densely packed
and present either in its unorganized form of numerous RNP-granules or
organized in a lamellar system. This latter organization, however, is distinct
from that of the normal array mentioned above and related to physiological
secretion. No regularity is found in the repartition of the RNP-granules
along the membranes and the parallel arrangement of these is not
regular either (Figs. 13, 15). Distension of the interlamellar space is often
conspicuous and in many A types of cells is so marked that a lace-like
pattern is outlined (Fig. 16). These cells appear vacuolated and since
their nuclei are often shrunken as well, they easily evoke the picture
of a degenerating necrotic cell. It is therefore impossible, on an ultra-
structural basis to decide whether the typical A type of cell corresponds
to a young, still undifiPerentiated element as believed by most or whether
it is not simply a necrotic cell.

It is not the place here to argue about the matter. The important notion
is that two different types of cells at least have been detected with spectro-
photometric and staining techniques and are detected again in the electron
microscope. Whether thev are absolutely identical with the A and B cells
of Caspersson and Santesson and whether they represent young or aged
elements is fundamental only to explain their significance. If this could
be imderstood, it is possible that great advances would be achieved in
ultrastructural cytology of cancer because the great development of A
cells in some neoplasms represents the only ultrastructural characteristic
common to different types of cancer. Indeed there is no difference between
some A type cells observed in adenocarcinoma of the breast and those

1 86

F. HAGUENAU AND K. H. HOLLMANN

Fig. 13. Detail of an A type of cell wedged in between two "B" cells. Note
the density and irregularity of ergastoplasm pattern, x 22 500.

THE ERGASTOPLASM IN THE MAMMARY GLAND AND ITS TUMOURS 1 87

i'

''f*^

■^

rf-

Fig. 14. Two different adjacent types of cells. At bottom a "B" cell with
clear cytoplasm contrasts with a dense cell, the cytoplasm of which contains many
RNP-granules. This is not a typical "A" cell, because the nucleus is not festooned
and the nucleolus is still conspicuous but coarseness of chromatin is incipient.
This cell may be considered as intermediate between the A and B "extreme"
types. X 12 250.

F. HAGUENAU AND K. H. HOLLMANN

C*''

f ./

Fig. 15. "A" type of cells in human mammary cancer. Note their dense aspect,
the coarseness of chromatin granules, the festooned limits of the nuclei and the
abundance of peculiar ergastoplasm in the cytoplasm, x 26 000.

THE ERGASTOPLASM IN THE MAMMARY GLAND AND ITS TUMOURS l8g

*- . "-^if Tpgr, .,_2-»-i!K ^^^ ~' i fc'rf

Fig. i6. Another aspect of A cells in a human mammary carcinoma. Ik-ie
again the granularity of chromatin is evident. Ergastoplasm is restricted only to
cytoplasmic remnants due to distension of its channels. This intense vacuolization
leads to a lace-like pattern and probably corresponds as in F'ig. 12 to necrosis.
At upper left a "B" cell with normal cytoplasm shows that preservation of the
preparation cannot be responsible for the lace-like aspect, x 7200.

190 F. HAGUENAU AND K. H. HOLLMANN

observed for example in hepatoma (unpublished data) or in uterine cervix
cancer (Hinglais-Guillaud, pers. comm.).

In experimentally induced mammary tumours cells of the mouse and
of the rat ergastoplasm is again never found in its typical organized form
as will be developed now.

The mammary neoplasms of these rodents have been studied in the
electron microscope almost uniquely from the angle of their possible viral
aetiology. Most of these papers, concerned with virus detection (Bittner
agent in the mouse), mention only briefly if at all the general histological
aspect of the tumour or the characteristics of the constituting cells.

But if one concentrates on this aspect, one observes that no " organized
ergastoplasm" is to be found in these cells where no lactation occurs.
Ergastoplasm is present mostly under the form of RNP-granules, abundant
in some cells, sparse in others (Figs. 17, 18). Not only are these cells rapidly
dividing but also they are deeply involved in virus formation as evidenced
by the presence of virus particles both mature (in the acinar lumen) and
in the process of formation (at the cellular membrane) (Fig. 18).

In the rat, likewise, mammary tumours are devoid of "organized
ergastoplasm" (Fig. 19). There also, it is possible that virus formation
occurs since elements have been discovered in the electron microscope [23]
which from their morphology could be virus particles. The role they play
in the aetiology of the tumour has not yet been established.

From all these observations on the ergastoplasm of the mammary
gland cell the following conclusions may be drawn.

(i) In the normal gland, typical "organized" ergastoplasm is charac-
teristically related to physiological secretion. It is an attribute of the
normal lactating cell.

(2) In experimentally induced tumours, ergastoplasm is found mainly
under a "non-organized" form, principally RNP-granules [i, 24].

These thus appear to be related to rapid cell replication or to forms
of protein synthesis other than normal secretion, virus formation in
particular. Such aspects of ergastoplasm are also characteristic of
embryonic cells which produce protein but do not excrete it. These
have been termed "retaining cells" by Birbeck and Mercer in a recent
paper [25].

(3) In human mammary tumours, most cells also contain ergastoplasm
in the form of RNP-granules or scattered dotted membranes and this
ergastoplasm is not on the whole, particularly abundant (B and transitional
types of cells). In some less numerous cells however (A cells of Caspersson
and Santesson) ergastoplasm is remarkably dense and its pattern of a
peculiar type not related to the specific tissue of origin. The significance
of these cells is not yet known but they merit further intensive examination

THE ERGASTOPLASM IN THE MAMMARY GLAND AND ITS TUMOURS 191

■^f^^-'"*'

i

Vl

Fig. 17. Mammary tumour of the mouse. Lo\\" power view of a few cells to
illustrate the absence of "organized" ergastoplasm. Free RNP-granulcs are widely
scattered in the cytoplasm, x 21 500.

192

F. HAGUENAU AND K. H. HOLLMANN

^

•

>^J:

'^'^x

4

#

^.f

I

•7":*. .

Fig. 18. Another aspect of mouse mammary tumour in an area where virus
formation occurs (Bittner agent). Note at this higher magnification the numerous
RNP-granules and the lack of organized ergastoplasm. x 59 000.

THE ERGASTOPLASM IN THE MAMMARY GLAND AND ITS TUMOURS 1 93

^*w::A: t'i^=tJ:;:>f J5..y^^- m : : i^^

V':^.:l^i

^)i/i^,

Fig. 19. General aspect of a rat mammary tumour (G.6) to show abundance
of RXP-granules and absence of the membranes component of ergastoplasm.
X 6400.

194 F- HAGUENAU AND K. H. HOLLMANN

since they are found in human cancers of varied origin and thus correspond
to a distinctive ultrastructural pattern common to all of these.

Furthermore a relationship exists between the presence of A cells in
a tumour and its radiosensitivity. These cells thus may be of important
prognostic value.

Acknowledgments

Work referred to in this paper has been aided by Grant C-4602 of the
U.S. Public Health Service, Figure 1 1 is due to the courtesy of Prof. T. O.
Caspersson.

References

1. Haguenau, F. /// " International Review of Cytology" Vol. 7, ed. G. H. Bourne
and J. F. Danielli. Academic Press, New York, 425-483 (1958).

2. Carasso, N. and Favard, P., in " Traite de Microscopic P'lectronique" (Magnan
C. ed.) 1961 Paris, Hermann Publ.

3. Rauber, A., Schmidt's Jb. 182, 7-8 (1879) (quoted after Dabelow Hdb. Mikr.
Anat. III/3, 277-488 (1957)-

4. Limon, M., Anat. physiol. 38, 14-34 (1902).

5. Bargmann, W., and Knoop, A., Z. Zellforsch. 49, 344-388 (1959).

6. Hollmann, K. H.,^. Ultrastriict. Res. 2, 423-443. (1959).

7. Palade, G. E., and Siekevitz, P., J/', biophys. biochem. Cytol. 25, 671-690 (1956).
7a. Siekevitz, P. and Palade, G. E.,_7. biophys. biochem. Cytol. 4, 309-318 1958

8. Siekevitz, P., Ciba Sy?np. on Cell Metabolistn 17-45 (i959)-

9. Siekevitz, P., Exp. Cell. Res. siippl. 7, 90-110 (1959).

10. Haguenau, F., Bull. Cancer, 46, 177-21 1 (1959).

11. Haguenau, F., Path. Biol. 7, 989-1015 (1959).

12. Haguenau, F., C R. Acad. Sci., Paris 249, 182-184 (1959).

13. Caspersson, T. O., and Santesson, L., Acta Radiol. XLVI Siippl. 5 (1942).

14. Caspersson, T. O., "Cell Growth and Cell Function. A Cytochemical Study".
(1950) New York. W. J. Norton.

15. Barigozzi, C, and Dellepiane, G., Rass. Oncologia 22, i (1948).

16. Cusmano, L., Ttimori 21, 107-121 (1947).

17. Cusmano, L., Ibid. 23, 63-85 (1949).

18. Gusberg, S. B., Tovell, H. M. M., Long, M., and Hill, J. Ann. N.Y. Acad.
Sci. 63, II 47- II 57 (1956).

19. Cornil, L., and Stahl, A., C R. Soc. Biol. 144, 1075-1077 (1950).

20. Cornil, L., and Stahl, A., Pr. me'd. 59, 933-935 (1951).

21. Gricouroff, G., Pr. me'd. 64, 137-139 (1956).

22. Towell, A. M. M., Cancer. Res. 20, 297-306 (i960).

23. Hollmann, K. H., and Riviere, M. R., Bull. Caiuer, 46, 336-346 (1959).

24. Howatson, A. F., Ham, A. W., Cancer Res. 15, 62-69 (1955).

25. Birbeck, M. S. C, and Mercer, E. H., Nature, Lond. 189, 558-560 1961.

The External Secretion of the Pancreas as a Whole and

the Communication between the Endoplasmic Reticulum

and the Golgi Bodies*

GoTTWALT Christian Hirsch

Zoologisches Institute
Gottingen, Germany

I have only a few remarks to make here : about logical basis of methods,
about the external secretion of the pancreas as a whole, and about the
"missing link" between the endoplasmic reticulum and the work of the
Golgi bodies.

I, The logical basis of methods

The external secretion of the pancreas — as perhaps every kind of
secretion — may be compared with the production of a factory [71, loi]:
in both cases raw material comes in and products of the factory go out.
Between these two events lies the synthesis of a particular product. This
production flows through a cell in a way which may be compared to an
"assembly line", because the synthesis of the particular product does
not take place in one structure of the cell, but gradually in different
places and structural parts of the cell, and in a particular order step by
step. Naturally, this comparison between an " assembly line " and secretion
has limitations [71] : the cell has no conveyor belts or rollers. But there is
a movement of material step by step, from one cell structure to another,
carrying the slowly developing product along. There are structures which,
furthermore, approximately portray the workers on an assembly line : for
instance the ribosomes, the endoplasmic reticulum, the Golgi field. Finally
there are other cell structures which produce materials which are then
moved to the place on the assembly line where they are biochemically
necessary : for instance, the mitochondria and the cell nucleus.

It is the task of integrated structural and biochemical studies to
observe the order of occurrence in the processes of secretion. With every
new technique new observations, microscopical and biochemical are made,
which have to be integrated in the scheme of a "production line". This
integration is often difficult because the different investigators do not

* This work was supported by the Deutsche Forschungsgenieinschaft Bad
Godesberg bei Bonn, Germany.

196 GOTTWALT CHRISTIAN HIRSCH

use the same species of animals or do not follow the same points of time
in the cell production processes. And, indeed, the same species of animals
show individual differences in time in their answer to the same stimulation ;
in other words, individuals of the same species differ in their production
curves.

But in spite of all these difficulties it is possible to get a tentative scheme
of the "production line" of every species and every gland cell under the
following conditions :

Secretion is a long sequence of processes in time [64, 65]. I believe it
may be necessary to follow this sequence by putting every qualitative and
quantitative datum on a single time co-ordinate [64]. Some investigations
are ill-timed because they are made with animals, whose timing of secretion
is unknown. There are, so far as I know, only two ways of timing [70, 71].

First to investigate the ontogeny of gland cells during the development
of many well-known stages from the first "Anlage " of the cell to the stage
in which the "professional" structure and function of the cell has been
constructed [2, 11, 45, 46, 70, 121, 122, 126, 145, 179, 180].

The second way may be to investigate the secretion cycle by com-
paring the biochemical and structural information [66, 10 1] first during
the so-called "starvation time" in which the sequence of secretion pro-
cesses is stopped or is very slow — second during the "activation time"
after a certain stimulus. This "activation time" is one phase of secretion
in which the extrusion of the old secretion product takes place, following
the restitution and storage of the new products. This phase continues in the
pancreas for about i to 5 hr. [66] ; during the 1st hour the most important
processes of restitution take place [71-72]. During this phase it is, there-
fore, necessary to investigate not only one point of time, but many points
during the 1st hour and some points during the 2nd to 5th hours.

The number of methods of these investigations are numerous. I may
say that the first necessary step is to observe the living cell, either as an
isolated cell in tissue culture or preferably as one cell in living relation-
ship with other cells inside the tissue. It is possible to observe the living
tissue of the pancreas under physiological conditions during 8-10 hr.
under the light microscope [66, 67]. Only in this way can an outline of the
time of production processes be obtained.

The second way is the study of the life cycle of a cell [64-72] : you start
your investigation with a more or less uniform stage of so-called "resting
cells" during the starvation of the animal [7, 68]. Then you give the
stimulus of feeding, of drugs, of nervous or of hormonal excitation always
at the same time after the beginning of starvation (e.g. 1 or 2 days). Then
you investigate the animals at different times after the same stimulus:
e.g. 3, 5, 10, 20, 40, 60 min. and 2, 3, 4, 6, 8 hr. after stimulus. For every
point of time three animals may be used. Every animal is treated by the

THE EXTERNAL SECRETION OF THE PANCREAS AS A WHOLE 1 97

same technique, biochemical or structural. These techniques are numerous :
vou can perhaps observe the living cell [66], or study the secretion product
by fistula [68, lOO, loi], or kill many animals in the above-mentioned times
after stimulus and compare the different stages of the cell cycle [185]. The
main point seems to be that all the different techniques are applied at the
same points of time. In this manner you will really be able to integrate the
results of different techniques. Only the integration of all the different
data gives a sequence of processes and only a sequence of events gives an
idea of the life of a secretion cycle [70, 71] showing us the "assembly
line" during restitution of the product.

2. The life cycle of the exocrine pancreas cell

The life cycle may be divided into three phases : the ingestion of raw
material, the synthesis, and the extrusion of products [69 72, 140].

The ingestion [66, 10 1] consists of active transport of material from
the blood through the basal membrane (surrounding a whole acinus) and a
cell membrane [159-161, 182]. The influx of this material in some cases
seems to be changeable and higher after stimulation than during starvation
of the animal. Proteins of the blood plasma are not ingested, only amino
acids [100].

The synthesis of enzymes may be shown by the schematic Fig. i and
may be divided into the following steps :

I. The ingested amino acids [15, 27, 56, 62, 88-90, 100, 104, 105, 124,
156, 162, 191] and phosphates [19, 20, 57, 101-108, iii, 125, 154, 165,
201] in the fluid basic substance and the polynucleotides from the cell
nucleus [30, 47, 62, 1 14-1 16, 128, 170, 181] arri\e at the ribosomes (ribonu-
cleoprotein granules) [5, 12, 26, 58 61, 117, 120, 141 143, 188, 194-197,
200, 203, 206] and are absorbed [18, 173 177]. From the cell nucleolus
ribonucleic acid reaches these granules [48 55, 73, 74, 77-80, 93, 109-113,
148]. Phosphates of high energy may be formed by mitochondria as an
energy source [76, 166]. With the aid of these substances proteins are
synthesized in or on the ribosomes [76, 144].

During starvation of the animal and i hr. after feeding, the ribosomes
have the same size of --150 A diameter and the same optical density.
Their number is considerably greater, however, during the restitution
period : according to the electron-microscopic pictures of Palade I would
say about ten times greater. One may deduct from this that the ribosomes
multiplv during the ist hour after feeding. This multiplication has also
been observed during ontogenesis [2, 46]. How these granules multiply,
however, is still not clear [13, 43, 44, 178]. Forms of division have not
been seen. I have now come to the following hypothesis which I put up
for discussion: the fluid cell plasm contains free, dissolved ribonucleic

198

GOTTWALT CHRISTIAN HIRSCH

o<
O

uoi;sa5u

THE EXTERNAL SECRETION OF THE PANCREAS AS A WHOLE 1 99

acid, free amino acids, lipoproteins and synthetically active enzymes. I
can imagine that through a special enzyme mechanism these parts may be
building new ribosomes step by step. This restitution de novo may be a
synthesis under controlling factors of the cell nucleus by the amount of
ribonucleic acid, extruded by the nucleolus [24, 25, 37].

The origin of enzymes in ribosomes of the pancreas has been proved
[32, 99, 102, 103, 106, 137, 139, 156 157, 171-177, 186, 202], and in many
other cells, shown at the first Symposium of the Biophysical Society [2, 3,
4]. The ribosomes have attained, in the last 2 years, a central position in
the metabolism of many different cells. It has been shown that these
ribosomes have a different ribonucleoprotein ratio : first in different cells,
secondly — and this is most important — in the same cell at different stages
of enzyme synthesis [171-175]- The ribosomes, which appear identical
under the electron microscope, are different biochemically. Step by step
the amount of ribonucleic acid decreases, but the amount of proteins
increases.

2. What happens with these proteins synthezised in the ribosomes ?
From the investigations of Siekevitz and Palade [171 -177] and the Hokins
[77-87] it may be concluded that the proteins are released from the ribo-
somes and penetrate the wall of the endoplasmic reticulum by active trans-
port [71, 97, 129-139, 146, 149-15 1]. Inside the endoplasmic canals the
proteins may be stored for a short time as intracisternal granules.

3. The intracisternal granules are found by Palade [136-139] in the
endoplasmic reticulum of the guinea-pig only i to 3 hr. after feeding, but
not earlier or later. They have a diameter of o • 20 • 3 ^ ; they are, therefore,
to be seen in the light microscope. They appear approximately homo-
genous inside the endoplasmic reticulum; they have no outer membrane,
and differ from the zymogen granules in several respects : {a) The intracis-
ternal granules originate inside the endoplasmic canals, while the zymogen
granules come from the vacuoles of the Golgi bodies, {h) They have a
diameter of at most 0-3 /t ; the zymogen granules, however, of o • 5 to
0-6 jtx. (r) They have no membrane; the zymogen granules, however, have
one. Thus it seems that the intracisternal granules and the zymogen
granules are not identical. But it is likely that both contain digestive
enzymes.

4. Up to this stage of the secretion process I believe the sequence of
events on the assembly line has been settled to a certain degree. A further
series of arguments speaks in favour of a final condensation of the digestive
enzymes as zymogen granules in the vesicles of the Golgi bodies [66, 81,
119, 183]. Sluiter confirmed this condensation statistically using a periodic
count method by light microscope [185]. Also the experiments of Chesin
[19 21], Marshall [116], Farquhar and Wellings [42] made this conclusion
likely. The electron microscope showed a sequence of four different

200 GOTTWALT CHRISTIAN HIRSCH

vesicles and Golgi vacuoles steadily growing larger and denser. Finally the
zymogen granules are surrounded by a special membrane, containing a
large amount of packed enzymes. This work of condensing seems to be the
general function of the Golgi bodies [67, 69, 71].

5. What is the bridge between the intracisternal granules on one hand
and the condensation inside the Golgi bodies on the other ? There must be
a connection which carries the enzymes from the endoplasmic reticulum
to the Golgi field.

There are two possibilities. First, that the intracisternal granules are
dissolved in the lumen of the endoplasmic reticulum and their parts

Fig. 2. A living pancreas cell of the acinus with good blood circulation.
Pilocarpine stimulation 15.40 hr. First illustration at 16.40 is of entire cell, then
observation of a single constellation of three granules. The mitochondrion here
observed moves in a snakelike manner but was not further observed. The move-
ment of the three granules between 17.00 and 20.10 as they moved to the Golgi-
field, where they disappeared. It is probable that this type of granule is identical
with the intracisternal granules of G. E. Palade. (From G. C. Hirsch [66].)

(enzymes ?) are brought to the apex by currents in the canals. Secondly,
that the intracisternal granules wander undissolved in the canals to the
Golgi field. Both possibilities are probable in certain cases, but the dis-
solving of these granules on the cell base has until now not been seen in
the living cell. The intracisternal granules are large and can be seen also
in the light microscope. The wandering of these granules has been observed
many times (Fig. 2).

This wandering was discovered by Hirsch. In the living pancreas
after pilocarpine stimulation, small granules are produced on the cell
base in the ergastoplasm. They have a diameter of approx. 0-3 to 0-4 jjl
(Hirsch's specification of 0-03 [x is a printing error). The endoplasmic
reticulum at that time was not to be seen, but the mitochondria embedded
in the reticulum were. Because of this Hirsch made the mistake that the

THE EXTERNAL SECRETION OF THE PANCREAS AS A WHOLE 201

granules were produced on the surface of the mitochondria. With the
advent of the electron microscope it was discovered that these granules
are not produced on the surface of the mitochondria but next to these
inside the endoplasmic canals. During the ist hour after pilocarpine
stimulation the number of these granules increases. When one cuts the
optical field down to a small part of the basal plasm, one sees that during
the first 3 hr. ~ lo to 40 granules appear in this field; then the number
decreases to almost zero, owing to the fact that the particles move away.
Thev show a population growth in the beginning of 100 to 150",, per
hour.

Hirsch could show that there is a physiological connection between
these granules and the Golgi field. Figure 2 shows one of the many observa-
tions in which a single group of three granules was observed during 4 hr.
The "constellation" of these three granules "twinkled" for about an
hour moving irregularly to the Golgi-field where it disappeared by being
dissolved. The granules followed a zig-zag course, with a velocity of i /x
in 7 to 13 min. ; the smaller granules move faster than the larger ones. As
soon as a granule reaches the Golgi field it remains stationary and dis-
appears, i.e. it dissolves. The movement of the Hirsch granules from the
ergastoplasm to the cell apex was checked experimentally by three inde-
pendent investigators [39].

Are these granules of Hirsch identical with the intracisternal granules }

Size: Hirsch's granules have in the light microscope a diameter of
03 to 0-4 /x. The intracisternal granules have a diameter of 0-2 to 0-3 /x
(Fig. 2). Both have the same size, both are visible in the light microscope.

Chemical composition: Neither are lipoid drops, but contain mostly
protein, as was shown by Hirsch on his granules, and also by Siekevitz
and Palade for the intracisternal granules. Thus, I believe the two types
of granules are identical.

What happens to these intracisternal granules after their movement
into the Golgi field .' There are some possibilities : either they develop
within the Golgi field through growth directlv into zymogen granules, or
they dissolve, and the dissolved parts are condensed in the Golgi field. In
case they are directly formed into zymogen granules then the process must
be visible in the light and electron microscope. Until now, however, all
investigations under physiological conditions have revealed that the
intracisternal granules disappeared in the Golgi field. Their "future" in
the Golgi field gave birth to several hypotheses, of which only one has
been upheld : the intracisternal granules dissolve, and their components,
which are more or less finished enzymes, are taken up by the Golgi bodies,
and in some way "packed" into the developing and enlarging zymogen
granules. This condensation of the parts of the intracisternal granules thus
becomes the major function of the Golgi field.

202 GOTTWALT CHRISTIAN HTRSCH

Trial and error

The communication between the endoplasmic reticulum and the
zymogen granules and the way by which the enzymes are absorbed by the
Golgi lamellae is not entirely clear. Recently I found in the micrographs
of Palade between the endoplasmic reticulum and the Golgi lamellae many
round shaped X-bodies of about 370 A diameter. (Fig. i.) The internal
shape of these bodies is varying. It is only theoretical to say this :

1. On the end of the ER in the Golgi field, ruptures of the ER mem-
branes are to be seen in the direction to the Golgi field. Through these
ruptures the protein content of the ER seems to go out to the space between
ER and Golgi lamellae.

2. These proteins may form the X-bodies ( }) It is, however, only the
topographical relationship that gives this idea.

3. These X-bodies may be taken up by the Golgi vacuoles and packed
up to large zymogen granules. This process, however, is far from being
clear.

References

1. Afzelius, B. A., /// Wenner-Cirens Institute, Festschrift (1959).

2. Ashikawa, J. K., in "Microsomal Particles and Protein Synthesis, i." Sym-
posium Biophysical Society, ed. R. B. Roberts. Pergamon Press, London,
76 (1958).

3. Askonas, B. A., Simkin, J. L., and Work, T. S., in "4. Internat. Congress of
Biochem.", Vol. VIII. Pergamon Press, London, 181 (i960).

4. Spooner, E. T. C, and Strocker, B. A. D. (eds.), "Bacterial Anatomy".
Symposium Soc. Gen. Microbiol. Cambridge (Engl.) University Press (1956).

5. Balis, M. E., Samarth, K. D., Hamilton, Mary G., and Petermann, Mary L.,
J. bio/. Chem. 233, 11 52 (1958).

6. Bargmann, W., "Histologic und mikroskopische Anatomic des Menschen".
3. Aufl. G. Thieme, Stuttgart (1959).

7. Barnum, C. P., Jardetzky, C. D., and Halberg, F. Tex. Rep. Biol. Med. 15,

134 (1957)-

8. Bernhard, W. Cancer Res. 18, 491 (1958).

9. Bessis, M., "Die Zelle im Elekronenmikroskop." Sandoz-Monographien
(i960).

10. Block, D. P., in "Frontiers in Cytology", ed. S. L. Palay. Yale Univ. Press,
New Haven, 113 (1959).

11. Block, D. P., and Godman, G. C.,^. hiophys. biochem. Cytol. I, 531 (1955).

12. Brachet, J., "Biochemical Cytology". Academic Press, New York (1957).

13. Brachet, J., in " Mecanismes d'Autoreproduction ", ed J. A. Thomas. Masson
et Cie, Paris (1957).

14. Campbell, P. N., Greengard, O., and Kernot, B. A., Biochem. J. 74, 107 (i960).

15. Castelfranco, P., Meister, A., and Moldave, K., in "Microsomal Particles
and Protein Synthesis, i." Symp. Biophysical Society, ed. R. B. Roberts.
Pergamon Press, London, 115 (1958).

16. Chantrenne, H., in "Enzymes". Tjeenk Willink, Zwolle, 143 (1959).

THE EXTERNAL SECRETION OF THE PANCREAS AS A WHOLE 203

17. Chauveau, J., Gautier, A., and Moule, Y., C R. Acad. Sci., Paris 241, 337

(1955)-

18. Chauveau, J., Moule, Y., and Rouiller, C, Exp. Cell. Res. 13, 398 (i957)-

19. Chesin, R. B., Biokhhniya 17, 664 (1952).

20. Chesin, R. B., Petrschkaite, S. K., Tolinschis, L. E., and Paulauskaite, K. P.,
Akad. d. Wissensch. USSR Bioche?me Moskau. 22, (1957).

21. Chesin, R. B., in "The Origin of Life on the Earth". Moscow (1957).

22. Chevremont, M., "Notions de Cytologic et Histologic". Desoer, Liege,
(149: Microsomes) (1957).

23. Ciba Foundation Symposium: "Regulation of Cell Metabolism". J. and A.
Churchill, London (1959).

24. Clark, W. H., Jr., J. hiophys. biocheni. Cytol. 7, 345 (i960).

25. Cotte, G., Z. Zellforsch. 50, 232 (1959).

26. Crampton, C. F., and Petermann, M. L., J. biol. Chem. 234, 2642 (1959).

27. Daly, M., .\llfrey, \'. (i., and Mirsky, A. E., J. gen. Physiol. 39, 207 (1955).

28. Davidson, J. X., in "Structure of Nucleic Acids". Cambridge Univ. Press,

27 (1957)-

29. Decken, A. von der, and Hultin, T., Exp. Cell Res. 14, 88 (1958).

30. Decken, A. von der, and Hultin, T., Exp. Cell Res. 15, 254 (1958).

31. Decken, A. von der, and Hultin, T., Exp. Cell Res. 17, 188 (1959).

32. Decken, A. von der, and Hultin, T., Exp. Cell Res. 19, 591 (i960).

33. Dickman, S. R., and Morrill, CJ. A., Atner.jf. Physiol. 190, 403 (i957)-

34. Dickman, S. R., and Morrill, G. A., Ann. N. Y. Acad. Sci. 81, 585 (i959)-

35. Dickman, S. R., and Trupin, K., Arch. Birjchem. Biophys. 82, 355 (1959).

36. Dickman, S. R., Morrill, G. A., and Trupin, K. M., J. biol. Client. 235,
169 (i960).

37. Dolley, D. H., Amer.J. Anat. 35, 153 (1925).

38. Douglas, T. A., and Munro, H. N., Exp. Cell Res. 16, 14S (1959)-

39. Duthie, E. S., Proc. ray. Soc. B. 114, 20 (1933).

40. Duthie, E. S., Arch. exp. Zellforsch. 15, 352 (1934).

41. Ekholm, R., and Edlund, \ .,'}. Ultrastrncture Res. 2, 453 (1959).

42. Farquhar, M. G., and Wellings, S. R., J. biophys. biocheni. Cytol. 3, 319

(1957)-

43. Favard, P., C. R. Acad. Sci., Paris 247, 531 (1958).

44. Favard, P., C. R. Acad. Sci., Paris 248, 3344 (1959).

45. Ferreira, David, _7- Ultrastrncture Res. i, 14 (1957).

46. Ferreira, David, "A differencia^ao do condrioma aparelho de Cjolgi e Ergasto-
plasma. Estudo ao Microscopio Electronico. " Lisboa (1959).

47. F"erreira, Fernandes J., and Junqueira, L. C. U., Arch. Biocheni. Biophys. 55,

54 (1955)-

48. Ficq, A., in " Problemes d'Ultrastructures et de Fonctions Nucleaires".
Ed. J. Andre Thomas. Masson et Cie, Paris, 35 (1959).

49. Freed, J. J., Exp. Cell Res. 9, 17-34 (i955)-

50. Garfinkel, D., /// "Microsomal Particles and Protein Synthesis, i." Sym-
posium Biophvsical Societv, ed. R. B. Roberts. Pergamon Press, London,
22 (1958).

52. Goldstein, L., and Plaut, W., Proc. nat. Acad. Sci., Wash. 41, 874 (1955).

53. Goldstein, L., and Micou, ].,y. biophys. biochem. Cytol. 6, 301 (1959).

54. Goldstein, L., and Micou, J.,X biophys. biochem. Cytol. 6, i (1959).

55. Goldstein, L., and Micou, J., J. biophys. biocheni. Cytol. 6, 301 (1959).

56. Greengard, O., and Campbell, P. N. Biochem. jf. 72, 305 (1959).

57. Grycki, S. Ann. Univ. I\I. Curie-Sklodozvska, C. 8, 193 (1953).

204 GOTTWALT CHRISTIAN HIRSCH

58. Haguenau, F". Int. Rev. Cytol. 7, 425 (1958).

59. Hall, B. D., and Doty, P., in "Microsomal Particles and Protein Synthesis.
I." Symposium Biophysical Society, ed. R. B. Roberts, Pergamon Press,
London, 27 (1958).

60. Hamilton, M. G., and Petermann, M. L.,^. biol. Chern. 234, 1441 (1959).

61. Hanzon, V., Hermodsson, L. H., and Toschi, G., ^. Ultrastructure Res. 3,
216 (1959)-

62. Hecht, L. I., Stephenson, M. L., and Zamecnik, P. C, Proc. nat. Acad. Set.,
Wash. 45, 505 (1959)-

63. Herman, L., Fitzgerald, P. J., Weiss, M., and Polevoy, I. S., in " Kongress
fur EM", Vol. 2, 372 (i960).

64. Hirsch, G. C, Ronx Arch. Entw. Mech. 117, 511 (1929).

65. Hirsch, G. C, Biol. Rev. 6, 88-131 (1931)-

66. Hirsch, G. C, Z. Zellforsch. 15, 37 (1931); Roux Arch. Entw. Mech. 123, 792
(193 1); Z. Zellforsch. 14, 517 (193 1) ; Z. Zellforsch. 15, 290 (1932).

67. Hirsch, G. C, "Handb. der Allgemeinen Pathologic", Vol. H, i. Springer-
Verlag, Heidelberg, 92-213 (1955).

68. Hirsch, G. C, Junqueira, L. C. U., Rothschild, H. A., Sesso, S. R., Pfliigers
Arch. ges. Physiol. 264, 78 (1957).

69. Hirsch, G. C, Natiinvissenschaften 45, 349 (1958).

70. Hirsch, G. C, in " Handbuch der Biologie ", Vol. i. Athenaion-Verlag,
Konstanz (1959-61).

71. Hirsch, G. C, Natumissenschaften 47, 25 (i960).

72. Hirsch, G. C, in "Protoplasmatologia", Vol. 3. Springer- Verlag,Wien (1961).

73. Hoagland, M. B., Zamecnik, P. C, and Stephenson, M. L., in "Symposium
on Molecular Biology", ed. R. E. Zirkle. University of Chicago Press,
Chicago, 105 (1959)-

74. Hoagland, M. B., in "4. Intern. Congress of Biochem.", Vol. VHI, Proteins,
199 (i960).

75. Hodge, M. H., and Chapman, G. B.,^. biophys. biochem. Cytol. 4, 571 (1958).

76. Hogeboom, G. H., KufT, E. L., and Schneider, W. C, Int. Rev. Cytol. 6,

425 (1957)-

77. Hokin, L. E., Biochim. biophys. Acta 8, 225 (1952).

78. Hokin, L. E., and Hokin, M. R., Biochim. biop/iys. Acta II, 591 (i953)-

79. Hokin, L. E., and Hokin, M. R., Biochim. biophys. Acta 13, 236 (1954).

80. Hokin, L. E., and Hokin, M. R., Biochim. biophys. Acta 13, 401 (i954)-

81. Hokin, L. E., Biochim. biophys. Acta 18, 379 (i955)-

82. Hokin, L. E., and Hokin, M. R..^^. biol. Cheni. 233, 822 (1958).

83. Hokin, L. E., and Hokin, M. R., Gastroenterology 36, 368 (1959).

84. Hokin, L. E., and Hokin, M. R., J. biol. Chem. 234, 1387 (1959)-

85. Hokin, M. R., and Hokin, L. E.,J. biol. Chem. 219, 85 (1956).

86. Hokin, M. R., and Hokin, L. E.,y. biol. Chem. 233, 522 (1958).

87. Hokin, M. R., and Hokin, L. E., J^. biol. Chem. 234, 1381 (i959)-

88. Hultin, T., Exp. Cell Res. 3, 210 (i955)-

89. Hultin, T., Exp. Cell Res. li, 222 (1956).

90. Hultin, T., and Beskow, G., Exp. Cell Res. ll, 664 (1956).

91. Hultin, T., von der Decken, A., and Beskow, G.,Exp. Cell. Res. 12, 675 (i957)-

92. Hultin, T., and von der Decken, A., Exp. Cell Res. 13, 83 (i957)-

93. Hultin, T., Exp. Cell Res. 12, 290 (1957).

94. Hultin, T., and von der Decken, A., Exp. Cell Res. 15, 581 (1958)-

95. Hultin, T., and von der Decken, A., Exp. Cell Res. 16, 444 (1958)-

96. Hultin, T., and von der Decken, A., Acta chem. scand. 12, 596 (1958).

THE EXTERNAL SECRETION OF THE PANCREAS AS A WHOLE 205

97. Hultin, T., in "Wenner-Grens Institut, Festschrift" 62 (1959).

98. Ingram, V. M., in "4. Internat. Congress of Biochem." Vol. 8. Pergamon
Press, London 95 (i960).

99. Jungblut, P. W., and Turba, F., //; " 10. Coll. Physiol. Chemie". Springer-
Heidelberg, Mosbach 102 (1959).

100. Junqueira, L. C. U., Hirsch, G. C, and Rothschild, H. A., Biochetn. jf. 61,

275 (1955)-
loi. Junqueira, L. C. U., and Hirsch, G. C, Int. Rev. Cytol. 5, 323 (1956).

102. Keller, E. B., Zamecnik, P. C, and Loftfield, R. B., J^. Hisiochem. Cytochem.

2, 378 (1954)-

103. Kirsch, J. F., Siekevitz, P., and Palade, G. E., jf. biol. Chem. 235, 1419
i960).

104. Korner, A., Biochim. biophys. Acta 35, 554 (1959).

105. Kurochtina, T. P., Biochemistry, Leningr. 19, 16 (1954).

106. Laird, A. K., and Barton, A. D., Biochim. biophys. Acta 25, 56 (1957).

107. Laird, A. K., and Barton, A. D., Biochim. biophys. Acta 27, 12 (1958).

108. Lancker, J. L. van, and Holtzer, R. lu.,jf. biol. Chem. 234, 2359 (1959).

109. Leblond, C. P., Everett, N. B., and Simnions, B. Amer. J. Anat. lOl, 225

(1957)-
no. Leuchtenberger, C, and Schrader, F., Proc. nat. Acad. Sci., Wash. 38, 99
(1952).

111. Lipmann, F., iyi "4. Intern. Congress of Biochem." Vol. VIII. Pergamon
Press, London, 177 (i960).

112. Littlefield, J. W., Keller, E. B., Gross, J. and Zamecnik, P. C.,y. biol. Chem.
217, III (1955)-

113. Littlefield, J. W., and Keller, E. B.,y. biol. Chem. 224, 13 (1957).

114. Love, Robert, and Bharadwaj, T. P., Nature, Lond. 183, 1453 (1959).

115. Mahlon, B., Hoagland, M. B., Stephenson, M. L., Scott, J. F., Hecht, L. I.,
and Zamecnik, P. C.,_7. biol. Chem. 231, 241 (1958).

116. Marshall, A. Lab. Invest. 8, 460 (1959).

117. Martin, E. M., and Morton, R. K., Biochem. J. 64, 687 (1956).

118. Maurer, W. /'// " 10. Coll. Physiol. Chemie" Springer-Heidelberg, Mosbach,

(1959)-

119. Morelle, J., Cellule 37, 75 (1927).

120. Moule, Y., Rouiller, C, and Chauveau, J., J. biophys. biochem. Cytol. 7,
547 (i960).

121. Munger, B. L., Atner.jf. Anat. 103, i (1958).

122. Munger, B. L., Anat. Rec. 130, 343 (1958).

123. Munger, B. L., J. biophys. biochem. Cytol. 4, 177 (1958).

124. Nakano, E., and Monroy, A., E.xp. Cell Res. 14, 236 (1958).

125. Noback, C. R., and Montagna, W., Amer. jf. Anat. 91, 343 (1947).

126. Oota, v., and Takata, Kenzo, Physiol. Plant. 12, 518 (1959).

127. Oberling, Ch., Int. Rev. Cytol. 8, i (1959).

128. Oram, V., Acta anat. Suppl. 23, 1-114 (1955).

129. Palade, G. E., and Porter, K. R., Anat. Rec. 112, 370 (1952).

130. Palade, G. E., in " Congr. Micr. electr." London (1954).

131. Palade, G. E., and Porter, K. R.,y. exp. Med. lOO, 641 (1954).

132. Palade, G. E.,^ biophys. biochem. Cytol. I, 59 (1955).

133. Palade, G. E.,_7. biophys. biochem. Cytol. i, 567 (1955).

134. Palade, G. E., and Siekevitz, P., J. biophys. biochem. Cytol. 2, 671 (1956).

135. Palade, G. E.,y. biophys. biochem. Cytol. 2, Suppl. 85 (1956).

136. Palade, G. E.,y. biophys. biochetn. Cytol. 2, 417 (1956).

206 GOTTWALT CHRISTIAN HIRSCH

137. Palade, G. E., in "Frontiers of Cytology", ed. S. L. Palay. Yale Univ. Press,
New Haven, 283 (1958).

138. Palade, G. E., in "Microsomal Particles and Protein Synthesis, i." Sym-
posion Biophysical Societv, ed. R. B. Roberts. Pergamon Press, London,
36(1958).

139. Palade, G. E., //; "Subcellular Particles" ed. T. Hayashi. Ronald Press Co.,
New York, 64 (1959).

140. Palay, S. L., in "Frontiers of Cytology" ed. S. L. Palay. Yale Univ. Press,
New Haven, 305 (1958).

141. Petermann, M. L., Hamilton, M. G., and Mizen, N. A., Cancer Res. 14, 360

(1954)-

142. Petermann, M. L., and Hamilton, M. G., ^ biol. Chem. 224, 725 (1957).

143. Petermann, M. L., Hamilton, M. G., Balis, M. E., Samarth, K., and Pecora,
P., in "Microsomal Particles and Protein Synthesis, i." Symposium Bio-
physical Society, ed. R. B. Roberts. Pergamon Press, London, 70 (1958).

144. Peters, T., Jr.,_7. Histochem. Cytochem. 7, 224 (1959).

145. Pipan, N., Dissert. Bonn (i960).

146. Policard, A., and Bessis, M., Pr. me'd. 64, 2153 (1956).

147. Policard, A., and Baud, C. A., "Structures inframicroscopiques normales et
pathologiques des Cellules et des Tissus". Masson et Cie, Paris (1958).

148. Pollister, A. W., in " Dynamics of Growth Processes" ed. A. J. Boell. Prince-
ton University Press, 33 (1954).

149. Porter, K. R.,^. Histochem. Cytochem. 2, 346 (1954).

150. Porter, K. R., Fed. Proc. 14, 673 (1955).

151. Porter, K. R., in "The Harvey Lectures", Series 51 (1955-56). Academic
Press, New York, 175 (1957).

152. Porter, K. R., and Bruni, C, Anat. Rec. 136, 260 (i960).

153- Quastler, H., in " Information Theory in Biology", ed. H. Quastler. Univ. of
Illinois Press, Urbana, 170 (1958).

154. Rabinovitch, M., Valeri, V., Rothschild, H. A., Camara, S., Sesso, A., and
Junqueira, L. C. \J.,J. biol. Chem. 198, 815 (1952).

155. Rashevsky, N., Bull. math. Biophys. 21, 309 (1959).

156. Rendi, R., and Hultin, T., E.\p. Cell Res. 19, 253 (i960).

157. Rendi, R., and Hultin, T., Exp. Cell Res. 18, 542 (1959).

158. Rhodin, J., V^erh. dtsch. Ges. Pathol, (j. Fischer, Stuttgart, 275 (1958).

159. Robertson, J. D., Biochem. Soc. Symp. 16, 3 (1959).

160. Robertson, J. D., Anat. Rec. 136, 266 (i960).

161. Robertson, J. D., in " Kongress fiir Elektronenmikroskopie ", Vol. 2, 159
(i960).

162. Rothschild, H. A., Hirsch, G. C, and Junqueira, L. C. U., Experientia 13,

158 (1957)-

163. Runnstrom, J., in "Modern Trends in Physiology and Biochemistry".
Academic Press, New York, 47 (1952).

164. Ruska, H., Stadium gen. 12, 133 (1959).

165. Schapot, V. C, Adv. mod. Biol., Moscow 37, 255 (1954).

166. Schneider, W. C, and Hogeboom, G. H., Ann. Rev. Biochem. 25, 201 (1956).

167. Schucher, R., and Hokin, L. E.,^. biol. Chem. 210, 551 (1954).

168. Schulz, Heribert, and de Paola, D., Z. Zellforsch. 49, 125 (1958).

169. Schwarz, W., and Hofmann, M., "Vierter Internationaler Kongress fiir
Elektronenmikroskopie", 2, 369 (1958).

170. Sesso, A., and Valeri, V., Exp. Cell Res. 14, 201 (1958).

171. Siekevitz, P., and Palade, G. E.,^. biophys. biochem. Cytol. 4, 203 (1958).

THE EXTERNAL SECRETION OF THE PANCREAS AS A WHOLE 207

172. Siekevitz, P., and Palade,- G. E.,jf. biophys. biochem. Cytol. 4, 309 (1958).

173. Siekevitz, P., and Palade, G. E.,^ biophys. biochem. Cytol. 4, 557 (1958).

174. Siekevitz, P., and Palade, G. E.,y. biophys. biochem. Cytol. 5, i (1959).

175. Siekevitz, P., E.xp. Cell Res. Suppl. 7, 90 (i959)-

176. Siekevitz, P., and Palade, G.,y. biophys. biochem. Cytol. 7, 619 (i960).

177. Siekevitz, P., and Palade, G. E.,J. biophys. biochctn. Cytol. 7, 631 (i960).

178. Simkin, J. L., and Work, T. S., in "Symp. Soc. Exper. Biol". XII. Cam-
bridge (Engl.) Univ. Press, 164 (1958).

179. Sirlin, J. L., and Waddington, C. H., Exp. Cell Res. 11, 197 (1956).

180. Sirlin, J. L., and Edwards, R. G., Exp. Cell. Res. 18, 190 (i959)-

181. Sirlin, J. L., Exp. Cell Res. 19, 177 (i959)-

182. Sjostrand, F. S., and Hanzon, V., Exp. Cell Res. 7, 393 (i954)-

183. Sjostrand, F. S., and Hanzon, V., Exp. Cell Res. 7, 415 (1954).

184. Sjostrand, F. S., Int. Rev. Cyto. 5, 455 (1956).

185. Sluiter, J. W., Z. Zellforsch. 33, 187 (1944)-

186. Stephenson, M. L., Hecht, L. I., Littlefield, J. W., Loftfield, R. H., and
Zamecnik, P. C, in "Subcellular Particles", ed. T. Hav'ashi. Ronald Press
Co., Nevv^ York, 160 (1959).

187. Stich, H., in "Developmental Cytology", ed. D. Rudnick. Ronald Press
Company, New York, 105 (1959).

188. Straub, F. B., Chem. Tech. 9, 460 (i957)-

189. Strugger, S. Natunv. Rdsch. I, 7 (i960), and 2, 51 (i960).

190. Swift, Hewson, in "Symposium on the Chemical Basis of Development",
ed. W. D. McElroy, B. Glass. Johns Hopkins Univ. Press, 175 (1958).

191. Szabo, M. T., and Garzo, T., Acta physiol. hung. 12, 303 (1957).

192. Tamaki, M., and Iwashige, K., Folia, anat.jfap. 23, i (1950).

193. Tandler, C. J., Exp. Cell Res. 17, 560 (1959).

194. Tashiro, Y., Sato, A., and Furuta, Y., Cytologia 22, 136 (1957).

195. Tashiro, Y., Acta Sch. med. Univ. Kioto 34, 238 (1957).

196. Tashiro, Y., Ogura, M., Sato, A., Shinagawa, Y., Imai, Y., Hirakawa, K.,
and Hirano, S. in " International Symposium on Enzyme Chemistry, Tokyo ",
436 (1958).

197. Ts'o, Paul O. P., in "Microsomal Particles and Protein Synthesis, i."
Svmposium Biophvsical Societv, ed. R. B. Roberts. Pergamon Press, London,
156(1958).

198. Ullmann, A., and Straub, F. B., Acta physiol. hung. 8, 279 (1955).

199. Ullmann, A., Garzo, T., and Straub, F. B., Acta physiol. hung. 13, 179 (1958).

200. Vendrely, C, Vendrely, R., in "Handbuch der Histochemie ", 3, 2. Teil:
Nucleoproteide, ed. W. Graumann and Neumann, K. Crustav Fischer-
Verlag, Stuttgart, 84 (1959).

201. Vorbrodt, A., Exp. Cell Res. 15, i (1958).

202. Webster, G. C, J. biol. Chem. 229, 535 (1957); Fed. Proc. 18, 348 (1959).

203. Weiss, J. M.,y. exp. Med. 98, 607 (1953).

204. Wilson, J. W.,X Histochem. Cytochem. 2, 317 (1954).

205. Yeas, M., in " S^^Tlposium on Molecular Biology ", ed. R. E. Zirkle. University
of Chicago Press, Chicago, 115 (1959).

206. Zalockar, M., Nature, Lond. 183, 1330 (1959).

2o8 GOTTWALT CHRISTIAN HIRSCH

Discussion

Porter: I think it is appropriate to mention to the audience at this point that
Palade has been interested in this topic and has published or has in press observa-
tions which would indicate that the intracisternal granules do not have to enter the
matrix phase of the cytoplasm to get to the Golgi components, but that intermit-
tent continuity is made between the cavities of the ER and those of the Golgi.
It is probable that the granules break down into less condensed forms in making
this migration.

Immunological Studies of Microsomal Structure
and Function

Peter Perlmann and Winfield S. Morgan*

The Wenner-Gren Institute for Experimental Biology,
University of Stockhohn, Szveden

In a number of publications it has been shown that specific antibodies
can be used as tools for studying the intracellular distribution of anti-
genically distinct macromolecules of rat liver [i, 15, 20]. In this paper, the
immunological properties of microsomal fractions from various organs
will be described, and some of our recent findings will be discussed with
regard to microsomal function in protein synthesis. In the present context,
the term microsomes refers to the ribonucleic acid rich fraction, sedimen-
ted in the ultracentrifuge when a cell-free homogenate is subjected to
differential centrifugation under certain standard conditions [14, i, 10].
This definition does not involve any particular implications as to the cyto-
logical homogeneity of this fraction which consists of ribonucleoprotein
particles and different types of endoplasmic membranes. It may contain a
good number of subcellular elements which are functionally unrelated and
of different origin [13]. The enzymic or electron microscopic criteria
applied for the characterization of the microsomal fractions in this type of
work are the same as those available in other biochemical studies, e.g.
[i4> 28, 7].

Antigens in microsomes of liver and other organs of the rat

So far, rat livers have constituted the most thoroughly studied material.
The analytical reagents are antisera obtained from rabbits injected either
with total microsomes or with microsomal subfractions, with the final
supernatant after removal of the microsomes ("cell sap"), or with rat
serum proteins. Details may be found elsewhere [i, 10]. When such
antisera are reacted on agar diffusion plates according to Ouchterlony [12]
one obtains complicated patterns of antigen-antibody precipitates. Most of

* Parts of this work were carried out during a tenure of a Special Research
Fellowship granted by the National Cancer Institute, U.S. Public Health Service.
— Permanent address : Department of Pathology, Massachusetts General Hospital,
Boston, 14, Mass., U.S.A.

2IO PETER PERLMANN AND WINFIELD S. MORGAN

the antigens are both immunologically and chemically different from the
serum proteins. All subcellular fractions contain a number of these "liver"
antigens in common. These antigens may either be present physiologically
or their presence in a certain fraction may be an artefact of fractionation.
However, in addition, there also exists a considerable number of antigens
typical of various fractions. Thus, the microsomal fraction of rat liver is
characterized primarily by five to six strongly antigenic proteins which
can be extracted with sodium deoxycholate, non-ionic detergents such as
lubrol W (cetyl alcohol polyoxyethylene condensate) [ii, 5], or with
phospholipase A [i, 20]. These antigens seem to be lipoproteins. They
are difficult to characterize electrophoretically. However, it has been
possible to isolate two of these antigens in an immunologically pure form.
A lubrol extract of microsomes w'as separated by chromatography on
DEAE cellulose [23], using constant pH (o'035 m tris-buffer, pH 7*8)
and a gradient of KCl for elution. The presence of 0-2% lubrol in all
reagents was necessary for fractionation (unpublished results).

The microsomal subfraction which remains insoluble after treatment
of the microsomes with detergent (ribonucleoprotein particles) contains
some additional weak antigens. These antigens can be visualized in the
diffusion plates after extraction with chelating agents at pH 7-4 [i, 20].
Their chemistry is so far unknown.

The microsomal antigens described above are different from the anti-
gens typical of liver cell sap. Moreover, when extracts of the liver fractions
are compared with microsomal extracts or cell sap from other rat organs,
additional differences are found. This is illustrated in Fig. i. This shows
the reactions obtained when lubrol extracts of microsomes from liver,
kidney, pancreas, testis, spleen and brain were reacted with the correspon-
ding anti-microsomal sera. Each line in these photographs represents the
precipitate formed by one, or sometimes several antigens, with their
corresponding antibodies. Fusion of precipitates from neighbouring
containers suggests an immunological identity of the antigens, while the
precipitates of unrelated antigens cross. The formation of spurs in certain
cases indicates either cross reactions or an immunological heterogeneity of
the precipitates [12].

None of the antigens appearing in Fig. i is due to serum proteins,
since antibodies against these had been removed in advance by means of
absorption of the antisera with rat serum. Roughly, Fig. i demonstrates
that the microsomal extracts of liver, kidney, and pancreas are highly
organ-specific in the sense that they give the greatest number of precipitates
with their homologous antisera. The anti-liver microsomal serum contains
the largest number of antibodies reacting with antigens present in extracts
from other organs. In contrast, lubrol extracts of microsomes from testis
contain only a few components ; in common with liver microsomes spleen

Fig. I. Photograph of six agar plates showing the precipitin reactions of
lubrol extracts of microsomes from different rat organs, and the corresponding
antisera. The antigen solutions labelled with the same letters in the different plates
were aliquots of the same extracts. The extracts were diluted to contain 350-400
/xg. protein /ml. T,B,L,K,P,S: microsomal extracts of testis, brain, liver, kidney,
pancreas, spleen, respectively; oltn, akm, apm, attn, astn, abm: antisera against
microsomes of liver, kidney, pancreas, testis, spleen, brain, respectively. For
details see [19].

212 PETER PERLMANN AND WINFIELD S. MORGAN

and brain microsomes seem to be more or less devoid of lubrol-extractable
antigens specific for these organs.

A somewhat different picture emerges when the cell saps from these or-
gans are compared in similar experiments. In this case, the number of com-
mon antigens in the different organs is greater. In general, it may be stated
that the "organ specificity" of the cell sap is less than that of the micro-
somes, although organ specific antigens are present in these fractions also,

A full description of these results will be given elsewhere [19]. Ob-
viously, no immediate conclusions can be drawn from experiments such as
these, since the various homogenates are histologically heterogenous, the
cytology of some of the microsomal preparations is badly known, and the
chemistry of the antigens has not yet been studied. However, it appears
that further studies along these lines should render important information
regarding the chemistry and genetics of microsomal structures in different
types of cells.

The presence of transitory antigens in microsomal fractions

While the aim of the above-mentioned studies is to investigate the
chemical composition and relationship of subcellular structures, the
immunology of the microsomes has another important aspect. Microsomes
are believed to be important sites of protein synthesis. While the synthesis
of protein is assumed to take place in or on the ribonucleoprotein particles,
the endoplasmic membranes of the microsomal fractions are often believed
to possess a regulatory function and to be involved in transportation of
freshly formed protein out of the cells. (See [25] for references and dis-
cussion.) During recent years, there has accumulated evidence indicating
that microsomes can be characterized immunologically by the antigens
which are believed to be the transitory products of their synthetic or
transporting activities. Thus, in a number of animal species, serum albumin
can be extracted from the microsomes of the liver, where it is synthesized
[4, 21, 22]. Similarly, antibody-active proteins have been extracted from
microsomes of lymph nodes or spleen of immunized animals [2, 6]. It has
also been reported that ribosomes from Escherichia coli can be precipitated
specifically by means of an antiserum against beta galactosidase [8].

Certain results obtained in our laboratory also point in that direction.
Thus, injection of rat liver microsomes into rabbits leads to a strong pro-
duction of antibodies, not only against the "liver" antigens shown above,
but also against a great number of serum proteins of the rat [i, 20]. Antisera
against microsomes of spleen contained antibodies against rat gamma
globulin, in very high concentration. In contrast, the antisera against
microsomal fractions of the other organs shown in Fig. i contained none,
or only trace amounts, of such antibodies [19].

IMMUNOLOGICAL STUDIES OF MICROSOMAL STRUCTURE AND FUNCTION 213

Although Other explanations cannot be excluded, the antibody-
inducing potencies of the microsomes from various organs may be a
reflection of their synthetic activities. More direct indications for such
activities have recently been obtained in human patients suffering from
ulcerative colitis. In this disease, there occur autoantibodies against a
substance of the colon, and the colonic lesions found seem to be of an
immunological nature [3]. In regional colonic lymph glands of these
patients, antibody-active microsomes could be detected in a remarkably
high concentration, whereas microsomes from lymph glands of the ileal
region of the same patients were devoid of antibody activity [17].

Finally, as the last example, in the chick embryo, one can detect an
organ specific lens antigen in the microsomal fraction at a very early age,
just before lens formation is initiated [18]. Later during embryogenesis,
when lens differentiation is going on, and in the adult, this lens antigen
occurs both in the microsomal fraction and in the cell sap. Electron micro-
scopy of these microsomal fractions indicates that both ribonucleoprotein
particles and membranous elements are present.

In conclusion, serum proteins, including antibody-active globulins,
as well as the lens antigen, are most likely first synthesized in the micro-
somes. Assuming these products remain attached to any one of the
microsomal substructures for a measurable period of time, it can also be
expected that they are available for detection with immunological tools.
It has already been shown that this is true in certain cases where proteins
are temporarily bound to the endoplasmic membranes before being ex-
ported [22]. Definite proof for the immunological recognition of ribonu-
cleoprotein particles by means of antibodies against their svnthetic products
is so far lacking. It is not at all clear whether or not such transitory products
already possess an immunological specificity while being attached to the
particles. The fact that ribonucleoprotein particles of the pancreas may
carry freshly synthesized chymotrypsin suggests that this could be
possible [27].

Incorporation of isotopes into antigens of rat liver homogenates

It should be emphasized once more that the mere presence of these
transitory antigens in microsomal fractions only provides indirect evidence
for the assumption that they are products of microsomal activity. The
microsomal fractions mentioned above are all isolated from heterogenous
cellular populations. Moreover, they may contain both the ribosomal and
membranous machinery involved in protein synthesis, as well as sub-
cellular elements of a completely different functional significance. However,
questions related to the problem of microsomal activity in protein syn-
thesis can be studied by combining immunological techniques with the

214

PETER PERLMANN AND WINFIELD S. MORGAN

study of isotope incorporation. This has been done in several of the cases
referred to above, as in the case of the serum albumin [4, 21, 22], or in the
case of the beta galactosidase in E. coU [8]. In the following, this will be
exemplified briefly with some experiments from this laboratory. A detailed
discussion of the results will be published elsewhere [10]. The purpose of
these experiments was to study, qualitatively and under various experi-
mental conditions, the pattern of in vitro incorporation of isotopes into the
bulk of immunologically distinct proteins of fractionated homogenates.

Figure 2 shows the results of an experiment with rat liver slices, in-
cubated for 30 min. with ^^C amino acids. After incubation, the slices were

cA \~y

a-m

a-a

ib)

Fig. 2. (rt) Photograph, and {b) autoradiograph, of precipitin reactions of
lubrol extracts of microsomes {LU) and of a cell sap {CS), from rat liver, with
various antisera {a-m : antiserum against liver microsomes ; a-cs : antiserum against
liver cell sap; a-a: antiserum against rat serum albumin). The antigen solutions
were obtained from a fractionated homogenate of liver slices, previously incubated
for 30 min. with ['*C]-L-leucine, -L-isoleucine, and -L-valine. The radioactivity
of the TCA-precipitated protein of an aliquot of the total homogenate was 425
c.p.m./mg. protein and of an aliquot of the isolated cell sap 311 c.p.m./mg.
protein. From Morgan et al. [10].

homogenized, and a lubrol extract of the microsomes and the cell sap was
reacted with antiserum on agar plates. After completion of the immuno-
logical reactions, the agar plate was dried and covered with an X-ray film
for 3 weeks (cf. also [16, 20]).

Figure 2{a) is a photograph of the precipitin reactions. As can be seen,
the two antigen solutions were reacted with an anti-liver microsomal
serum, an anti-liver cell sap serum, and an antiserum against rat serum
albumin. Both anti-liver sera had been absorbed with rat serum and were
free of precipitating antibodies against serum proteins. The antiserum
against serum albumin contained antibodies against the albumin, and in

IMMUNOLOGICAL STUDIES OF MICROSOMAL STRUCTURE AND FUNCTION 21 5

small concentrations, against two alpha globulins and one beta globulin.
The photograph shows that the lubrol extract of the microsomes gives
four distinguishable precipitates with these antibodies. These serum
protein-like antigens are immunologically different from other microsomal
antigens, reacting with the anti-microsomal serum. Moreover, a con-
siderable amount of serum proteins had also accumulated in the cell sap.
This also contained additional antigens, some specific for cell sap, and
some in common with the microsomal extract.

From the autoradiograph of Fig. 2{b), it can be seen that an appreciable
number of the precipitates were distinctly labelled. As described else-
where [10], several lines of evidence, and control experiments such as
addition of unrelated precipitin systems to the solutions before agar
plating, or dilution with unlabelled amino acids after incubation with
i^C amino acids, all suggest that the labelling in the autoradiograph is not
due to an unspecific adsorption of radioactive material to the precipitates.
Rather, we may assume that the labelling of the precipitates indicates a
true incorporation of ^^C amino acids into the antigens. The autoradio-
graph shows that the albumin and other serum protein-like antigens
which were solubilized from the microsomes were heavily labelled, as well
as some additional microsomal antigens of unknown significance. In
contrast, in the cell sap, only the serum albumin carried enough of the
label to be easily detected. This is in spite of the fact that the concentration
of the serum proteins in this fraction probably was higher than in the
microsomal extract, as indicated by the relative positions and the appearance
of the precipitates. Other cell sap antigens were more or less inactive.

Experiments of this type make it possible to follow the distribution,
under various experimental conditions, of the radioactive label in immuno-
logically identical proteins, recovered from different cellular fractions. Thus,
the results of the experiment of Fig. 2 confirm other work reporting
the first appearance of the radioactive label in the microsomal serum
albumin followed by an increasing specific activity of this protein in the
cell sap. This is taken to indicate that freshly synthesized molecules are
transferred from the microsomes to this fraction [21, 22].

A great number of specific precipitates appearing in Fig. 2{a), par-
ticularly those formed by the antigens of the cell sap, were hardly labelled
at all. Howxver, as we do not know their chemical nature or their antibody
combining ratios, we cannot use the degree of labelling of precipitates as
the basis for comparison of the metabolism of immunologically different
antigens. Problems of this type can be approached in a different way.
Figure 3 shows a series of four autoradiographs, made from four
experiments with slices of rat liver, incubated with ^^C amino acids. The
four experiments were made with aliquots of the same preparation, and
the only difference between them was the time of incubation with the

2l6

PETER PERLMANN AND WINFIELD S. MORGAN

isotopic amino acids; for details see [lo]. The Ouchterlony reactions of all
four agar plates were practically identical and very similar to those shown
in Fig. 2(rt). Figure 3 demonstrates again that the radioactive label first
appears in the microsomal serum proteins, reacting with an antiserum
against rat serum, and in some additional microsomal antigens, reacting

a-cs

a-m

i-cs

a-m

a-cs

a-m!

a-cs

a-m

EOTA* I EDTA*^^

fi-pHS a~cs, ^-PHJ

(«)

EDT,

CS'

r

a-rs a-m?

L'

a-pHs a-cs

t")

■la-rs

(b)

4^"i

a-pH5

a-rs a-m ' a-rs

EDTA*

f >

a-pH5

a-cs

a.pH5

\

a-rs a-m

|cs'^

a-rs

(0

Fig. 3

id)

IMMUNOLOGICAL STUDIES OF MICROSOMAL STRUCTURE AND FUNCTION 21 7

with anti-microsomal serum. In the cell sap, the appearance of labelled
serum proteins is again slower. Nevertheless, after a long period of
incubation, the cell sap also contained a complete spectrum of strongly
labelled serum proteins. It should be emphasized that the concentration
of the serum proteins was practically the same in the cell saps of all four
experiments. Moreover, Fig. 3(</) shows, that even additional antigens in
the cell sap finally become considerably labelled, provided the period of
incubation with the isotope is sufficiently long. Thus, experiments of this
type suggest that the rate of isotope incorporation in the serum proteins
is faster than that of most other liver antigens reacting in our systems.

When the pattern of incorporation of ^"^C amino acids into liver antigens
is studied after incubation of mitochondria-free homogenates, or of isolated
microsomal particles [24], a number of interesting differences are observed.
Even in these systems, a reasonable labelling is obtained, both of serum
protein-like antigens, and of other antigens extracted from the microsomal
fraction. However, only trace amounts of radioactive serum proteins can
be detected in the cell sap. Instead, this fraction, and, to a still higher
extent, the ribonucleoprotein particles, contain radioactive material which
precipitates in the agar around the basins, but without reacting with anti-
bodies. The nature of this material is unknown. The significance of these
findings will be discussed elsewhere [9].

Fig. 3. Autoradiographs of precipitin reactions of lubrol extracts (Lu) and
EDTA-extracts (EDTA) of microsomes, and of the cell sap (CS), of rat liver, with
various antisera (a-m : antiserum against liver microsomes ; a-cs : antiserum against
liver cell sap; a-pH 5: antiserum against cell sap fraction precipitated at pH 5;
a-rs: antiserum against rat serum). The antigen solutions were obtained from
four separately incubated (P*C]-L-leucine, L-isoleucine, and L-valine) and sub-
sequently fractionated aliquots of a preparation of rat liver slices.

{a) 15 min. incubation with isotope. Time of exposure of autoradiograph :
3 months. The radioactivity of the TCA-precipitated protein of an aliquot of the
total homogenate was 135 c.p.m./mg. protein and that of an aliquot of the isolated
cell sap 68 c.p.m./mg.

{b) 30 min. incubation with isotope. Time of exposure of autoradiograph: 6
weeks. Radioactivity: total homogenate 402 c.p.m./mg. protein; cell sap: 280
c.p.m./mg. protein.

(f) 60 min. incubation with isotope. Time of exposure of autoradiograph: 6
weeks. Radioactivity: total homogenate 478 c.p.m./mg. protein; cell sap: 380
c.p.m./mg. protein.

{d) 120 min. incubation with isotope. Time of exposure of autoradiograph:
3 weeks. Radioactivity: total homogenate 2190 c.p.m./mg. protein, cell sap
15 10 c.p.m./mg. protein.

(Autoradiographs obtained after 6 months exposure of the plates corresponding
to the experiments of Figs. 2{a)-{c) did not show a greater number of labelled
precipitates.) The reactions obtained with the EDTA extracts and with a-pH 5
as well as other details are discussed elsewhere. From Morgan et al. [10].

2l8 PETER PERLMANN AND WINFIELD S. MORGAN

Summary

The microsomal fraction isolated from rat liver homogenates by
differential centrifugation contains a number of antigens which are
typical for this fraction. Extracts of microsomes of other organs of the rat
contain antigens which are typical for these fractions but are different
from the microsomal liver antigens. In addition to such "microsomal"
antigens, microsomes sometimes contain transitory antigens which are
only temporarily bound to microsomal particles or membranes. Examples
of such antigens are serum proteins in liver microsomes, antibody-active
proteins in microsomes of spleen or lymph nodes, and lens antigen in
microsomes of early chick embryos. The presence of these transitory
antigens in the microsomes may be an expression of their synthetic function.
This function can be studied by combining immunological techniques
with isotope incorporation. Examples are given, where the in vitro in-
corporation of ^^C amino acids into various antigens of rat liver homo-
genates was studied. After 30 min. of incubation of liver slices, serum
albumin and other serum protein-like antigens of the microsomes were
more strongly labelled than similar antigens in the cell sap. After longer
periods of incubation, strongly labelled serum proteins also appeared in
the cell sap. Under the same conditions the labelling of most other liver
antigens was weaker than that of the serum proteins. Incubation of cell-
free systems also led to the labelling of the different antigens.

Acknowledgments

This work has been supported by grants from the Swedish Cancer
Society. The technical assistance of Miss Margareta Engdahl is gratefully
acknowledged.

References

1. D'Amelio, V., and Perlmann, P., Exp. Cell Res. 19, 383 (i960).

2. Askonas, B. A., in " Mechanisms of Antibody Formation", eds. M. Holub and
L. Yaroskova, Publishing House of the Czechoslovak Academy of Sciences,
Prague, 231 (1960).

3. Broberger, O., and Perlmann, P.,jf. exp. Med. no, 657 (1959).

4. Campbell, P., Greengard, O., and Kernot, B. A., Biochem. J. 74, 107 (i960).

5. Cohn, P., and Butler, J. A. V., Biochem. J. 70, 254 (1958).

6. Feldman, M., Elson, D., and Globerson, A., Nature, Lond. 185, 317 (i960).

7. Hanzon, V., and Toschi, G., Exp. Cell Res. 16, 256 (1959).

8. Kameyama, T., and Novelli, G. D., Biochem. Biophys. Res. Comm. 2, 393 (i960).

9. Morgan, W. S., Hultin, T., and Perlmann, P. (in preparation).

10. Morgan, W. S., Perlmann, P., and Hultin, T., jf. biophys. biochem. Cytol. (in
press).

11. Ord, N. G., and Thompson, R. H. S., Biochem. J. 49, 191 (195 1).

12. Ouchterlony, O., in "Progress in Allergy", Vol. V, ed. P. Kallos. S. Karger,
Basel, I (1958).

IMMUNOLOGICAL STUDIES OF MICROSOMAL STRUCTURE AND FUNCTION 2I9

13. Palade, G. E., "First Symposium Biophysical Society". Pergamon Press,
London, 36 (1958).

14. Palade, G. E., and Siekevitz, P.,^ biophys. biochem. Cytol. 2, 171 (1956).

15. Perlmann, P., and D'Amelio, V., Nature, Loud. 181, 491 (1958).

16. Perlmann, P., and Hultin, T., Nature, Loud. 182, 1530 (1958).

17. Perlmann, P., and Broberger, O., Nature, Loud. 188, 749 (i960).

18. Perlmann, P., and De Vincentiis, M., Exp. Cell Res. (in press).

19. Perlmann, P., and Morgan, W. S., Exp. Cell Res. (in press).

20. Perlmann, P., Hultin, T., D'Amelio, V., and Morgan, W. S., Exp. Cell Res.
Suppl. 7, 279 (1959)-

21. Peters, T., ]r.,y. biol. Chem. 229, 659 (1957).

22. Peters, T., Jr.,^. Histochem. Cytochem. 7, 224 (1959).

23. Peterson, E. A., and Sober, H. A., ^. Amer. chem. Soc. 78, 751 (1956).

24. Rendi, R., and Hultin, T., Exp. Cell Res. 19, 253 (i960).

25. Siekevitz, P., Exp. Cell Res., Suppl. 7, 90 (1959).

26. Siekevitz, P., and Palade, G., J. biophys. biochefn. Cytol. 4, 203 (1958).

27. Siekevitz, P., and Palade, G., _7- biophys. biochem. Cytol. 7, 619 (i960).

28. Toschi, G., Exp. Cell Res. 16, 232 (1959).

Discussion

Siekevitz: In your autoradiographs did you see any protein which would be
labelled in the supernatant without first being labelled in the microsomes ?

Perlmann: We know that some of the typical "liver" antigens in the cell sap
will become labelled after a long period of incubation but we have not been able
to identify them since there are so many antigens in this fraction. We only know
that these labelled antigens are immunologically different from the serum proteins.
So far, we cannot exclude their presence in the microsomal extract.

Marshall : I wonder if you or anyone you know of has tried labelling similar
antigenic systems with an isotope which goes not into the proteins but into phos-
pholipids for example ? Also I should like to ask whether you have any indication
that lipid rather than protein constituents are responsible for antigenic specificity
in any of these cases ?

Perlmann: I don't think these methods have been used for the study of in-
corporation into the phospholipids or of the immunological properties of the lipid
fractions. Up to now, we have only used '*C amino acids and have followed their
incorporation into the proteins.

Amino Acid Incorporation by Liver Microsomes and
Ribonucleoprotein Particles

Tore Hultin, Alexandra von der Decken,
Erik Arrhenius and Winfield S. Morgan*

The Wenner-Gren Institute for Experimental Biology ,
University of Stockholm, Sweden

Experiments ranging over widely different groups of organisms have
shown that at least in the cytoplasm, but probably in all parts of the cell
[24, 25], the synthesis of proteins is intimately connected with a class of
submicroscopic particles (cf. [10]) remarkably rich in RNA.f By means
of isolated particles it has been possible to reconstruct, at the subcellular
level, essential parts of the protein-synthesizing mechanism, and model
experiments of various kinds have already yielded a rich supply of in-
formation about the enzymic pathways involved in the synthetic process.
However, in spite of the remarkable prosperity of the approach, the basic
function of the particles is still very vaguely understood, i.e. the reaction
by which individual, activated amino acids become linked together in a
predetermined order to polypeptide chains of specific configuration,
precursors of the final proteins. Neither has any definite answer been
given to the question of how the completed protein molecules are even-
tually released from the active sites on the particles.

From the point of view of cell function a particularly important problem
pertaining to protein metabolism is how this fundamental process is
quantitatively and qualitatively regulated in growth and differentiation.
Unfortunately, our possibilities of reaching concrete answers to this
particular problem are even more limited at present, since no model
experiments have yet been devised which unambiguously reproduce such
regulatory mechanisms in vitro.

In experiments with cell-free amino acid incorporation systems it is
often observed that the incorporation activity under optimal conditions
is approximately proportional to the amounts of RNP-particles present [5].

* Permanent address : Department of Pathology, Massachusetts General Hospital,
Boston, Mass., U.S.A.

t Abbreviations : ATP, adenosine triphosphate ; GTP, guanosine triphosphate ;
PEP, phosphoenolpyruvate ; RN.A., ribonucleic acid ; RNP-particles, ribonucleo-
protein particles.

222

TORE HULTIN et ol .

One might therefore assume that after the break-down of the gross
structural organization of the cells the activity of the particles rapidly
reaches a uniform level, independently of the previous functional state.
If this were always the case, it would seem doubtful whether the effects
of rate-regulating factors would ever become accessible to experimental
study by cell-free systems.

In the following some experimental evidence will be discussed, indi-
cating that under certain conditions the actual rate of amino acid incorpora-
tion in vitro is determined not only by the concentration of potentially active
RNP-particles. The incorporation activity may also be influenced by other
factors, including some which depend on physiological (or pathological)
conditions in the living tissues before the preparation of the homogenates.

Microsomes and RNP-particles in incorporation systems

According to the picture obtained by electron microscopy RNP-
particles may occur in the cytoplasm either in a free state, or associated
with the membranes of the endoplasmic reticulum [23, 27, 31]. Especially
in the cells of highly differentiated animal tissues which have become
specialized for protein secretion or protein accumulation, the majority of
RNP-particles may appear in a membrane-bound form. It has been
suggested that in such cells the membranes are directly or indirectly con-
cerned with the anabolic functions of the particles [11, 29].

In rat liver a high proportion of the RNP-particles are firmly attached
to the endoplasmic membranes. Membrane fragments with adhering
particles can be isolated from liver homogenates by differential centri-
fugation, accumulating in the microsome fraction. Rat liver microsomes,
as is well known, has become a classical material in the study of protein
metabolism by cell-free systems [28, 33].

It has recently been observed that the membrane components of micro-
somes are not necessarily required for the function of the particles in
amino acid incorporation [4, 17, 26]. Liver systems containing purified
enzyme preparations in combination with isolated RNP-particles have, for
instance, the same ability as systems with whole microsomes to incorporate
labelled amino acids into interior sites of protein molecules [22]. Neverthe-
less, there are certain potential differences between incorporation systems
with whole liver microsomes and such ones with isolated RNP-particles.

This fact is illustrated by the experiment shown in Fig. i. The particles
used in this experiment were prepared by treating rat liver microsomes with
a mixture of detergents (deoxycholate and Lubrol W) in a solution of
fairly high ionic strength [26]. The detached particles were then separated
from the digestion mixture by centrifugation through a density gradient
[6, 26], The experiment shows a characteristic difference between the

AMINO ACID INCORPORATION BY LIVER MICROSOMES

223

system with whole microsomes and that with isolated RNP-particles :
With the isolated particles there was a considerably longer period of active
incorporation.

1000

Period of preincubation, minutes

Fig. I. Life-span of incorporation systems with microsomes or RNP-particles
from rat liver. • RNP-particles and cell sap [26]. A Microsomes and cell sap (mito-
chondria-free homogenate) [5]. At the indicated periods of incubation (35 ) the
following components were added: o-o8 /xmole [^^CJ-L-leucine (5 75 mC/mmole),
10 /xmoles PEP, i /xmole ATP and 0-2 /xmole GTP. Final volumes i o ml. Incuba-
tion period after these additions 15 min. Proteins were purified and counted as
described previously [5].

Effects of liver poisons on the activity of microsome systems

The tendency of the membrane fragments of liver microsomes to exert
an unfavourable influence on the incorporation activity becomes con-
siderably more pronounced in experiments with liver preparations from
certain kinds of pretreated animals. The experiment illustrated by Fig. 2
[8] was carried out with liver microsomes from guinea-pigs which had
been treated for 2 hr. with the liver poison phalloidin [32] (0-5 /xg./kg.,
intravenous injection). The activity of these microsomes in amino acid in-
corporation was determined at a series of increasing concentrations by use
of a standard assay system containing excess amounts of soluble fraction
from normal guinea-pig liver. Incubation tubes with microsomes from
normal guinea-pig liver at the same concentrations were run in parallel.

224

TORE HULTIN et al.

Irrespective of whether [^"^CJ-L-leucine or [^^C]-L-lysine was used as labelled
component in the system, a considerably reduced incorporation activity
was observed with the microsomes from the phalloidin-treated animals.

That this effect on the activity of the microsomes was mediated in
some way by the membrane components was suggested by the fact that
isolated RNP-particles from the same animals showed no significant

mg of microsomal protein

Fig. 2. Incorporation activity of liver microsomes from normal and phalloidin-
treated guinea-pigs. Incorporation system as in Fig. i, with 9 mg. of soluble
proteins from normal liver. Incubation period 15 min. (35 ). After the addition
of TCA the tubes were adjusted to equal protein contents [5]. Upper curves:
normal microsomes. Lozver curves: microsomes from phalloidin-treated animals
(o'5 /^g-/kg, 2 hr.). Solid curves :[^'^C]-i.-\eucine (o-o8 /xmole, 5-75 mC/mmole).
Broken curves: ['^]C-L-lysine (0-04 /imole, 13 "5 mC/mmole).

inhibition under these conditions (Table I). Not until later, when the
phalloidin-effect had become more manifest (or under the influence of a
higher phalloidin concentration) did the isolated particles give evidence of
irreversible damage, as indicated by the incorporation data. The RNA/
protein ratio of the microsomes was not influenced by the early phalloidin-
effects described here.

In this connection it should be noted that in the early stages of the
phalloidin treatment no impairment could be demonstrated of a number

AMINO ACID INCORPORATION BY LIVER MICROSOMES 225

TABLE I

Effect of Phalloidin-Treatment in vivo on the Activity of Incorporation
Systems with Microsomes and RNP-Particles

From livers of control and phalloidin-treated guinea-pigs microsomal fractions
[5] and RXP-particles [26] were isolated in parallel. The activities of these prepara-
tions were compared in the same way as in Fig. 2.

/xg. phalloidin

Incorporation

in

per

cent of controls

per g. body weight

Microsomes

RNP-particles

0-4
0-6
I -o*

72
45
53

96
94
63

* Macroscopic effects on liver.

of mitochondrial functions, including P/0-ratio, respiratory control and
latent ATP-ase.

It may be of some interest that the inhibitory effect of phalloidin on
the incorporation of labelled amino acids into protein was not confined
only to the cell-free systems. Similar effects were readily obtained with
liver slices under conditions when there is no histological evidence of any
appreciable intracellular disorganization (Table II).

TABLE II

Effect of Phalloidin on the Incorporation of [^^C]-dl-Valine into Protein
BY GuiNEA-PiG Liver Slices

Incubation at 35° in Oo-CO., (95 : 5) gas phase. The tubes contained 80 mg.
of liver slices [12] in 2 ml. of a Krebs-Henseleit medium [18] including 0-025 ^^
glucose. Addition of 1-34 fimoles of [^*C]-DL-valine (0-14 mC/mmole) at the
beginning of incubation or after 60 min. of preincubation.

Incubation after the addition of [^*C]-DL-valine 45 min.

Concentration

Without preincubation

60 min.

preincubation

of phalloidin

c.p.m./cm^.

% Inhibition

c.p.m./cm^

% Inhibition

0

3 X IQ-^ M
lO"^ M

73
46

40

37
45

106

74
64

30

40

It is not known yet by which modification of the membrane functions
phalloidin exerts its inhibitory effects on the incorporation activity of the
microsomes. The effect seems not to be due to an induction of ATP-ase
in the membranes. As is shown by Table III about 25% of the ATP and
PEP added to the incorporation system was broken down to orthophosphate
in the course of the incubation, but no increased phosphate formation was

Q

226 TORE HULTIN et al.

TABLE III

Effect of Phalloidin-Treatment in vivo on the ATP-Breakdown in Incor-
poration Systems with Guinea-Pig Liver Microsomes

Release of orthophosphate from ATP (ii ^moles) or ATP (i /xmole) and PEP
(lo /^moles) in incorporation systems containing liver microsomes from control
and phalloidin-treated guinea-pigs. Incubation system otherwise as in Fig. 2.
ATPase activity [19] of the control liver microsomes: i 5 jumoles orthophosphate/
15 min./mg. microsomal protein.

/xmoles Pj formed in 15 min.

per mg. microsomal protein /o Inhibition
of ["C]-L-leucine

I umole ATP + , . rr.T^ incorporation

1 T^T-T^ 1 1 umoles A 1 P

10 /Limoles PEP

Control liver 2 • 6 3 • i

Phalloidin liver 1-9 20

observed in the presence of microsomes from phalloidin-treated guinea-

Effects of a somewhat similar kind have been obtained in rat liver with
the liver carcinogen, dimethylnitrosamine [12, 13], and to some extent
also with 2-aminofluorene [i]. Taken together these experiments indicate
that the incorporation in vitro of amino acids into protein by liver micro-
somes is influenced by factors acting in the membrane components of the
microsomes. Such factors may be predetermined by conditions prevailing
already in the living cells.

Evidence of stimulation effects in cell-free systems

The experiments discussed above serve to illustrate the fact that
under certain conditions the activity of cell-free incorporation systems
may be abnormally low in proportion to their content of microsomal RNA,
owing to the influence of factors induced already in vivo.

The opposite situation may also be met with in the study of cell-free
systems of this type, i.e. that the incorporation activity shows a marked
enhancement which cannot be explained on the basis of a proportional
increase in the total content of RNP-particles. The first observation of
this kind was made in sea urchin eggs in connection with their fertilization
[14]. Only a few minutes after fertilization when the total number of
RNP-particles still must have been very nearly the same as in the un-
fertilized eggs, the activity of whole egg homogenates in amino acid in-
corporation was considerably increased (Table IV). The same was true
with "mitochondria-free homogenates" (centrifuged for 8 min. at 12 000
X g). The fertilization reaction rapidly seemed to make the individual
RNP-particles more active in amino acid incorporation, in such manner

AMINO ACID INCORPORATION BY LIVER MICROSOMES

TABLE IV

227

Incorporation of P*C]-l-leucine into Protein by Cell- Free Preparations
FROM Unfertilized and Fertilized Psammechinus Eggs

Preparation of cell fractions essentially as in Fig. i. The incubation system
contained 0-4 ml. of whole homogenate or o-6 ml. of 12 000 g supernatant fraction
from equal egg samples in combination with o-6 ml. of rat liver cell sap [5], or
RNP-particles from 0-7 ml. of packed eggs in combination with 0-65 ml. of the
liver cell sap. System otherwise as in Fig. i. Incubation period (20 ) 30 min.

c

•P

m

/cm.^

Unfertilized

Fertilized

Whole homogenate
12 000 g supernatant
RNP-particles

29
28

45

71
74
58

that this effect persisted also in preparations from disrupted cells. However,
with RNP-particles isolated from the homogenates by the detergent
treatment and layering procedure briefly mentioned in a previous section
this fertilization effect was less marked.

100

80

60

E
d.
o

AG'

20

0 20 40 60

Period otpreincubation.minutas

Fig. 3. Incorporation activity of a mitochondria-free rat liver homogenate at
different periods of incubation (20"). At the periods indicated o-o8 ^tmole [^'*C]-
L-leucine, 10 /xmoles PEP, 15 /xg. pyruvate kinase, i /xmole ATP and 02 /xmole
GTP were added. Incubation period after these additions 5 min.

228

TORE HULTIN et ol.

It may be relevant in this connection to mention some characteristic
properties of cell-free incorporation systems containing mitochondria-free
homogenates from liver [7]. Contrary to what would reasonably be
expected in view of the limited life-span of these systems, the incorporation
rate is not optimal just after the beginning of the incubation, but attains
a fairly distinct maximum after some period of preincubation, the duration
of which varies with temperature (Fig. 3). This lag is not due to the time

400

E

E

Q.

200

Time in minutes

Fig. 4. Amino acid incorporation by mitochondria-free liver homogenates
from normal {a) and phalloidin-treated (h, c) guinea-pigs. Doses of phalloidin
0-4-0 -5 /xg./kg. System as in Fig. i. [^*C]-L-leucine was added at the indicated
periods of incubation, and TCA at 15 min. (35').

required for the labelled amino acids to become activated, since in the
illustrated experiment the nucleoside triphosphate-generating system was
added to the system together with the labelled amino acid at the different
periods indicated.

When mitochondria-free homogenates were prepared from phalloidin-
treated mice or guinea-pigs, the preincubation effect was less pronounced
or absent. The incorporation then proceeded at a maximal rate from the
beginning of the incubation. With animals treated with small doses of
phalloidin the paradoxical situation therefore sometimes occurred that
over a 15 min. incubation period the total incorporation became higher in
liver homogenates from the phalloidin-treated animals than in those from
the controls (Fig. 4).

AMINO ACID INCORPORATION BY LIVER MICROSOMES 229

Stimulation effects in liver systems with relationship to the action

of cortisone

As was briefly mentioned in a previous section the treatment of rats
with Hver carcinogens may lead to a reduced activity of the isolated micro-
somes in amino acid incorporation. After a single dose of 2-aminofluorene
(2-5 mmoles per kg.) this effect is usually observed most easily 4-6 hr.
after the administration.

Only a few hours later a strikingly different picture develops. As is
shown by Table V, 20 hr. after the administration of 2-aminofluorene the

TABLE V

Effect of in vivo Treatment with 2-Aminofluorene (AF) or z-Aminoaphtha-

LENE (AN) ON THE INCORPORATION OF [1^C]-L-LeUCINE INTO PrOTEIN BY CeLL-

Free Liver Systems

The carcinogens (2-5 mmoles /kg. body weight) were administered 21 hr.
prior to sacrifice. Incorporation system as in Fig. i. Incubation period (35°) 10 min.

Incorporation

in per

cent

of controls

Rats

Guinea-pigs

AF

AN

AF

Mitochondria-free homogenates

Microsomes!

RNP-particlest

490*

242

128

220
222
142

144
143

TOO

* RNA /protein ratio 120 per cent of that of normal homogenate [3].
t Primary values expressed as c. p.m. /cm-, per 100 /xg. of particle-bound RNA
added to system (cf. Fig. 2).

activity of mitochondria-free rat liver homogenates in amino acid incorpora-
tion was increased far above the normal level. At the same time there was
some increase of the RNA-content in the homogenates, but this increase
(usually less than 25%) was not nearly so great as the sometimes several-
fold rise in incorporation activity.

Owing to the lack of proportionality between the activity of these liver
homogenates in amino acid incorporation and their RNA contents it
seemed possible that the activity increase might depend on a more eflicient
amino acid activation. It turned out, however, that this was not the main
reason of the effect. Liver microsomes were prepared in parallel from
aminofluorene-treated and normal rats and their incorporation activities
were determined at different concentrations in a system containing standard
amounts of normal soluble liver fraction. The microsomal activities were
then compared at equal RNA or protein concentrations in the same way as
in the phalloidin experiments described before. As is shown in Table V

230 TORE HULTIN et ttl.

the microsomes from aminofluorene-treated animals showed a considerably
higher incorporation activity per unit amount of RNA than did normal
microsomes. With isolated RNP-particles, however, the stimulation effect
was less pronounced.

There may be some relationship between the effects of 2-aminofluorene
and its carcinogenic potency. In guinea-pigs, a species refractory to the
carcinogenic action of 2-aminofluorene, the activation after 20 hr. was
smaller than in rats (Table V). On the other hand, a significant stimulation
was repeatedly observed after 4-6 hr., i.e. at a period when rat liver
microsomes showed a definite inhibition. This species difference may
possibly be due to the more rapid disposal of aminofluorene by metabolic
detoxication in the guinea-pig liver. It may be of some interest that the
well-known bladder carcinogen 2-aminonaphthalene, which has some
potency also in liver, had about the same effects in rat liver as 2-amino-
fluorene (Table V).

TABLE VI

Effect of in vivo Treatment with 2-Aminofluorene or X-Rays on the
Glycogen Content of Rat Liver

Dose of 2-aminofluorene 25 mmoles/kg. body weight, and of X-rays (all
body irradiation, 45 min.) 2500 r. Glycogen content [16] in per cent of liver dry
weight.

Treatment of rats Glycogen content

Fed ad libitum 9 • i

Fasted 20 hr. 0"53

Fasted 45 hr. 0-43

Fasted 24 hr. Aminofluorene 4 hr. i '51

Fasted 45 hr. Aminofluorene 21 hr. 3 '26

Fasted 45 hr. X-rays 21 hr. 61

As a routine the experiments were carried out with animals which had
been kept without food for i6~20 hr. By this brief period of starvation the
glycogen content in the liver is greatly reduced, a fact of practical advantage
in the preparation of microsomes and RNP-particles. In the animals treated
with 2-aminofluorene it was observed that the content of liver glycogen
was higher than in the controls, without reaching, however, the level
characteristic of normal, fed animals (Table VI). A similarly increased
glycogen content has been observed by Kay and Entenman [15] in X-ray
irradiated rats. In view of this striking similarity it seemed pertinent to
find out whether X-ray treatment would also evoke an increased activity
of amino acid incorporation in vitro As is shown by Table VII the in-
corporation activity of mitochondria-free rat liver homogenates was in
fact increased 21 hr. after the exposure of the animals to an X-ray dose
of 2500 r.

AMINO ACID INCORPORATION BY LIVER MICROSOMES

TABLE VII

231

Effect of X-Irradiation or Prednisolone Treatment in vivo on RNA-Content
AND [i^C]-l-Leucine Incorporation Activity of Mitochondria-Free Rat

Liver Homogenates

X-rays (all body irradiation, 45 min. 2500 r) and prednisolone (50 mg/kg. body
weight, intraperitoneal injection) 21 hr. prior to decapitation. Incubation as in
Table V.

Incorporation and RNA content in
per cent of controls

X-irradiation

prednisolone

[^*C]-L-leucine incorporation
RNA content

282
no

455
120"

* Cf. [20].

Experimental evidence has been advanced by Mole [21] that the effect
of irradiation on the content of liver glycogen depends on an increased
susceptibility of the liver cells to corticoids. The following facts indicate
that the increased incorporation activity described in this section has a
similar relationship to corticoid functions: {a) The activation normally
observed 20 hr. after aminofluorene treatment was absent in adrenalec-
tomized animals (Table VIII). As a matter of fact, an inhibition was

TABLE VIII

Effect of 2-Aminofluorene-Treatment in vivo on the Incorporation of
["C]-l-Leucine into Protein by Mitochondria-Free Homogenates from
Adrenalectomized Rats

The rats were adrenalectomized 48 hr. before the incorporation experiment.
Experimental conditions as in Table V, 2-aminofluorene (AF) being administered
21 hr. prior to decapitation.

Pre-treatment in vivo

c.p.m./cm'^

Adrenalectomized controls
Adrenalectomized, AF (o'5 mmole/kg.)
Adrenalectomized, AF (10 mmole/kg.)

441

389
308

usually observed ; {b) Activation effects similar to those obtained in the rat
with 2-aminofluorene, 2-aminonaphthalene or X-irradiation were obtained
after the intraperitoneal administration of prednisolone (Table VII).
These experiments are consistent with the recent observations by Feigelson
et al. [9] that cortisone enhances the amino acid incorporation by rat liver
in vivo.

Not very much can be said at present about the working mechanisms

232

TORE HULTIN et al.

behind the activation effects described. It may be recalled from the
previous section that unexpectedly high incorporation values were often
obtained in systems containing mitochondria-free liver homogenates from
animals treated with sublethal doses of phalloidin (Fig. 4). This effect
could be related to a characteristic modification of the normal time course
of the incorporation. In liver systems from animals recovering after amino-
fluorene treatment the maximal rate of incorporation may possibly be
reached slightly earlier than usual (Fig. 5), but the quantitative importance

300

CM

E

E 200

Q.

100-

0 A 8 12 0^8

minutes of i ncubat ion

Fig. 5. Incorporation of [^*C]-L-leucine into protein by mitochondria-free
guinea-pig liver homogenates. At zero time, or after 4 or 8 min. of incubation
(35 C) o -08 /nmole of [^''C]-L-leucine, 10 /^^moles of PEP and i /xmole of ATP were
added. Incubation was interrupted 4 min. later by the addition of TCA. Open
bars: control livers. Solid bars: livers from animals given 2-aminofluorene (2-5
mmoles/kg.) 21 hr. prior to decapitation.

of this tendency must be fairly limited. As is evident from the time-course
curve shown in Fig. i, stimulation effects of the kind discussed here could
in principle be due to a prolongation of the active life-span of the micro-
somes. However, with the liver preparations from the aminofluorene-
treated animals there was no indication of any appreciably longer active
period. As is shown by Fig. 5, the increase in incorporation activity was
equally pronounced in the period with maximal incorporation rate as in
any other incubation period.

In cell-free liver systems from adult animals only a limited part of in-
corporated, labelled amino acids becomes distributed to the proteins of

AMINO ACID INCORPORATION BY LIVER MICROSOMES

233

the soluble fraction. It seemed possible, therefore, that the increased in-
corporation rates described here might be due to a facilitated release of
already labelled proteins from the anabolic sites on the RNP-particles.
The distribution of incorporated ^^C-amino acids in different subfractions
of the incubation mixture was therefore studied in experiments with
mitochondria-free homogenates from normal and aminofluorene-treated
animals. An experiment of this kind is shown in Table IX. When the

TABLE IX

Incorporation of ["C]-l-Leucine into Protein by Mitochondria-Free Rat

Liver Homogenates from Normal and Aminofluorene-Treated Animals.

Distribution of Isotope between the Proteins of Different Subfractions

OF THE Incubation System.

Experimental conditions as in Table V, 2-aminofluorene (AF, 25 mmoles/kg.
body weight) being administered 21 hr. prior to decapitation.

c.p.m

./cm^.

AF/Controls

Fraction

AF

Controls

Unfractionated

1048

449

2-33

Whole microsomes

3000

1650

1-82

Microsomes, DOC-extract

563

221

2-50

Cell sap, pH 5 precipitate

510

104

4-90

Cell sap, pH 5 supernatant

185

37

5-0

incorporation system from the aminofluorene-treated animals is compared
with the control system, it appears in fact that an especially high increase
in specific activity was obtained in the soluble fraction. A significantly
smaller increase was observed in the proteins of the microsomes.

Discussion

The experiments described above indicate that under certain con-
ditions the activity of RNP-particles in cell-free incorporation systems
may be determined by factors, established in the intact cells and still
retaining their influence in the disintegrated preparations. According to
the experimental evidence, positive effects of this kind may be observed as
well as negative. The examples which were given of negative after-effects
in cell-free systems are most easily explained on the basis of a decreased
functional stability of the microsomal membranes. In the experiments
mentioned, this effect was induced under the action of a characteristic
group of liver poisons which also includes some potent carcinogens.
The experiments clearly illustrate the general importance of the metabolic

234 TORE HULTIN et al.

state of the microsomal membranes for the functional abihties of the
attached RNP-particles.

The stimulation effects described in the previous sections cannot yet
be properly interpreted. It has been shown, particularly by experiments
with regenerating liver, that growth stimulation may be accompanied by an
increase in the total amount of RNP-particles in the cytoplasm [2, 5].
This process, by which the total capacity of the protein synthesizing
machinery rapidly increases, may, however, constitute a rather late link
in the reaction chain of growth stimulation, since, at least in regenerating
liver, it demands a fairly long induction period (12-14 hr.). There reason-
ably must be a demand for more direct regulation mechanisms, by which
the production of individual protein species can be adapted to current
needs in a more rapid and flexible way (cf. [30]). The anabolic capacity of
an average cell is probably not fully utilized under normal physiological
conditions. It seems very likely, therefore, that the early stages of growth
induction are characterized by an intensified utilization of the anabolic
units already available.

The experiments described here may have some bearing on stimulation
effects of the latter kind. They all suggest that even in cell-free incorpora-
tion systems the protein-synthesizing units are not always working at full
speed. There apparently may still be room for some influence of modifying
factors. In the experiments which were described these factors were
physiological in the sense that they could be induced at will already in the
living cells by treatments in vivo with agents such as hormones, carcinogens
and, in the case of the sea urchin egg, the fertilizing sperm. There is some
indication that the effects are related to a facilitated release of labelled
proteins from the metabolic sites on the microsome particles. In view of
this it will be of importance to attain a more detailed knowledge about the
mechanisms by which this release is mediated.

Summary

Some experimental evidence has been collected in favour of the
possibility that factors which influence the protein metabolism in the
living cells may survive the disintegration of the cell structure and continue
to influence the anabolic activity of the RNP-particles even in cell-free
amino acid incorporation systems. Both negative and positive after-effects
of this kind are discussed. The negative effects are illustrated by the action
of certain liver poisons, including potent carcinogens. A treatment in vivo
with these agents reduces the activity of isolated liver microsomes in cell-
free incorporation systems. The effect is not entirely due to a direct and
irreversible damage to the RNP-particles of the microsomes. Apparently

AMINO ACID INCORPORATION BY LIVER MICROSOMES 235

a decreased functional stability of the microsomal membranes, which
secondarily influences the activity of the attached particles, is a factor of
importance in this reduction of activity.

The positive after-effects described have been observed in experiments
with just fertilized sea-urchin eggs, and with liver directly or indirectly
influenced by corticoids. In this case the activity of the cell-free incorpora-
tion systems becomes abnormally high in proportion to the amount of
RNP-particles present. There are some indications that these effects are
related to a facilitated release of labelled proteins from the particles.

The possibility is considered that effects of this kind may reflect
simple intracellular regulation mechanisms, which to some extent may
continue to operate also when the gross structural organization of the
cells has been destroved.

References

1. Arrhenius, E., and Hultin, T., Exp. Cell Res. 22, 476 (1961).

2. Bernhard, W., and Rouiller, C, J. biophvs. biochem. CytoL, Suppl. 2, 73
(1956).

3. Ceriotti, G., J. biol. Chem. 214, 59 (1955).

4. Cohn, P., Biochim. biophvs. Acta 33, 284 (1959).

5. Decken, A. von der, and Hultin, T., E\p. Cell Res. 14, 88 (1958).

6. Decken, A. von der, and Hultin, T., E.xp. Cell Res. 15, 254 (1958).

7. Decken, A. von der, and Hultin, T., "Progress in Biophysics and Biophysical
Chemistry" (in press).

8. Decken, A. von der, Low, H., and Hultin, T., Biochem. Z. 332, 503 (i960).

9. Feigelson, P., Feigelson, M., and Greengard, O., " ist Internatl. Congr.
Endocrinology", Abstracts 823 (i960).

10. Hanzon, V., Hermodsson, L. H., and Toschi, S.,jf. Ultrastriictiire Res. 3, 216

(1959)-

11. Hirsch, G. C, Noturzvissenschofteti 47, 25 (i960).

12. Hultin, T., Arrhenius, E., Low, H., and Magee, P. N., Biochem. J. 76, 109
(i960).

13. Hultin, T., Magee, P. N., and Arrhenius, E., "4th Internatl. Congr. Biochem",
Abstracts 185 (1958).

14. Hultin, T., and Bergstrand, A., Developmental Biol. 2, 61 (i960).

15. Kay, R. E., and Entenman, C, Proc. Soc. exp. Biol. 91, 143 (1956).

16. Kemp, A., and Kits van Heijningen, A. J. M., Biochem. jf. 56, 646 (1954).

17. Korner, A., Biochim. biophvs. Acta 35, 554 (i960).

18. Krebs, H. A., and Henseleit, K., Hoppe-Seyl. Z. 210, 33 (1932).

19. Low, H., Biochim. biophvs. Acta 32, i (1959).

20. Lowe, C. U., Rand, R. N., and Venkataraman, P. R., Proc. Soc. exp. Biol. 98,
696 (1958).

21. Mole, R. H., Brit.jf. exp. Path. 37, 528 (1956).

22. Morgan, W. S., von der Decken, A., and Hultin, T., Exp. Cell Res. 20, 655
(i960).

23. Palade, G. E.,y. biophvs. biochem. Cytol. i, 59 (1955).

24. Rendi, R., Exp. Cell Res. 17, 585 (1959).

236 TORE HULTIN et al.

25. Rendi, R., Exp. Cell Res. 19, 489 (i960).

26. Rendi, R., and Hultin, T., Exp. Cell Res. 19, 253 (i960).

27. Selby, C. C.,J. biophys. biochetfi. Cytol. I, 429 (1955).

28. Siekevitz, V.,y. biol. Chem. 195, 549 (1952).

29. Siekevitz, P., Exp. Cell Res. Suppl. 7, 90 (1959).

30. Spiegelman, S., in "Symposia of the Society for Experimental Biology",
ed. F. Danielli and R. Brown. University Press, Cambridge, 286 (1948).

31. Ts'O, P. O. P., Bonner, J., and Vinograd, J., ^. biophys. biochem. Cytol. 2,
451 (1956).

32. Wieland, T., in "Festschrift Arthur Stoll", ed. P. Karrer et al. Birkhauser,
Basel, 582 (1957)-

33. Zamecnik, P. C, and Keller, E. B.,J. biol. Chem. 209, 337 (1954).

Discussion

Williams : Korner has reported effects of insulin and growth hormone on
amino acid incorporation, and I was wondering if you had confirmed these effects.

Hultin : No. I do not think that Korner's experiments say very much about
whether a high content of ribonucleoprotein particles has been induced or whether
the particles already available work more efficiently in the system, but I should
like to test that possibility.

Siekevitz: One other explanation of your results is this: In the liver cells some
RNP particles are on the membrane and some others are free, and when you mea-
sure the incorporation you measure the protein radioactivity in these mixed kind
of particles. These particles are possibly synthesizing proteins at different rates
and are perhaps also synthesizing different kinds of proteins. So that when you
think you are making these particles increase they synthetic ability, what you
might be really doing is changing the ratio from the normal state, the ratio of bound
to free particles. You, therefore, might be switching over the synthesis from one
kind of protein to another kind of protein which has a higher rate of synthesis.

Hultin : I have also been considering that kind of possibility and also the
possibility that particles may be released from the membrane and therefore work
better. I am not quite convinced that a major change in the population pattern
of the particles would be the right explanation in the case of the just fertilized sea
urchin egg.

Packer: Both you and Dr. Siekevitz are implying that what has happened is
that changes in structure are simultaneously occurring, and I was wondering
whether it would not be possible to devise an experiment, at least in an in vitro
system, where you could simultaneously follow a parameter in physical structure
and one in the function activity, such as viscosity determinations, light scattering
determinations and sedimentation velocity.

Hultin : Yes, perhaps under certain conditions what you call ergastoplasm
may form some sort of sandwich which may become released, so that particles
attached to this open sandwich may act more rapidly by becoming more easily
supplied with energy and amino acids.

Zamecnik: There is another possible site of operation of a regulator of protein
synthesis. Data of Dr. Tissieres et al. {Proc. nat. Acad. Sci., Wash. 46, 1450-1463

AMINO ACID INCORPORATION BY LIVER MICROSOMES 237

(i960)) suggest that the 70 S ribonucleoprotein particle is actively engaged in pro-
tein synthesis, but that the 30, 50 and 100 S particles are not. Since a 30 S and a
50 S particle aggregate to form the active 70 S particle, it is possible that factors
(such as Mg^+ concentration) which influence the aggregation and disaggregation
of these particles, may regulate the number of active sites available at a single mo-
ment for protein synthesis.

Packer: Detergents are one clearway by which one can dissociate microsomes;
another clear way is the type which has just been put on the board and this type
has been documented quite nicely for microsomes of pea seedlings. Recently we
have been examining suspensions of liver microsomes by adding anions and cations
and chelating agents such as ATP, and following the changes in scattered light in
the suspensions, and we find very large changes in light scattering accompany the
addition of these reagents ; so I would like to suggest that these normal chemical
substances which are present in the cell are capable of aflfecting the state of struc-
ture of particles and perhaps the reaction system you employ for measuring the
incorporation of labelled amino acids.

HuLTiN : Perhaps I should point out again that we always have a good deal of
magnesium present, and the magnesium content, as shown by Dr. Peterman, is
very important in determining the aggregation of the particles. I think that factor is
under control as we use a little more than the physiological concentration of
magnesium.

Packer : I feel that this sort of approach must be considered if you are to evalu-
ate properly the structural-functional relationships especially as the types of
changes in light scattering which I have mentioned occur very rapidly.

HuLTiN : We agree that this has something to do with structure because when
we treat the particles with detergents in a solution of high ionic strength we find
the effects fade off.

Porter: I would like to ask what you think happens to the microsomal
membrane when you detach the RNP particle from the microsomes with DOC or
any other of these reagents. Does a fragment go with the particle ?

HuLTiN : No. The particles we obtain in this way are absolutely free from all
the enzymes which are typical of the microsomes. For example, from gIucose-6-
phosphatase, from DPNH-cytochrome c reductase, from TPNH-cytochrome c
reductase, from cytochrome 56, etc. This means that they have very little impurities
from the membrane.

Campbell : Have you any idea what happens to the microsome pellet when you
incubate it, when it loses its ability to incorporate ?

HuLTiN : I am very interested in this point but unfortunately I can't say very
much about it yet.

The Effects of Spermine on the Ribonucleoprotein
Particles of Guinea-Pig Pancreas*

Philip Siekevitz
The Rockefeller Institute, NezvYork, U.S.A.

Introduction

Our laboratory has been interested in the RNPf particles of the
pancreas for the past few years [1-7]. Particularly pure preparations of
these particles can now be obtained, having a diameter of 150 A [i, 6, 7].
These particles have an RNA content of from 35 to 45",, [i, 7], a sedimen-
tation coefficient of 85 S [7] and a molar amount of Mg + + equivalent to
about one tenth that of the phosphate groups in the RNA [7]. The particles
as isolated have bound amylase, RNase and TAPase activities [6, 7].
Indeed, after the injection of radioactive leucine, the chymotrypsinogen
which can be isolated from these particles has a higher specific radio-
activity, at the early time points after injection, than the chymotrypsinogen
isolated from other parts of the cell [6]. This finding indicates that these
particles synthesize chymotrypsinogen, and possibly other secretory
enzymes of the pancreas.

When the Mg + + of the particles is removed by chelating agents such
as versene, ATP, GTP, and P-P, all of the above-named enzymes are
concomitantly removed together without about 25^0 of the total protein
and about 80",, of the RNA of the particles [7]. It was calculated [7] that
a good deal of the protein removed was made up of these bound enzymes.
The particles are still recognizable as such, but their diameter has been
increased about twofold [7]. We believe that the proteins synthesized on
the particles are there for a finite period of time and have to be removed
to complete the process of soluble protein synthesis. We were therefore
interested in finding conditions under which only the enzymes are released

* This work was made possible by grant (A- 1635) from the National Institute
of Arthritis and Metabolic Diseases, National Institutes of Health, U.S. Public
Health Service.

t Abbreviations used are: RNA, ribonucleic acid; RNA-P, ribonucleic acid
phosphate ; RNP, ribonucleoprotein ; P-P, inorganic pyrophosphate ; ATP and
GTP, adenosine and guanosine triphosphates ; tris, tris (hydroxymethyl) amino-
methane; RNase, ribonuclease ; TAPase, trypsin-activateable proteolytic activity;
TCA, trichloracetic acid.

240 PHILIP SIEKEVITZ

from the particles, without removal of the RNA from the particles, since
the total RNA of these particles has a very low rate of turnover in vivo [5].
This preliminary report notes one of these conditions. In brief, it has been
found that the polyamines, putrescine, cadaverine, and particularly sper-
mine, will effect a release of the bound enzymes from the particles. In
addition, spermine replaces the Mg + + of the particles, and in doing so,
confers a stability on the particles in the absence of Mg + +, so that very
little or no release of the RNA occurs under these conditions. The con-
sequence of these findings on the nature of the bindings among the RNA,
protein, and enzymes of the particles is discussed.

Methods

The RNP particles of the pancreas were isolated from pancreatic micro-
somes by the detergent, deoxycholate, as already described [6, 7]. The
measurements of amylase, RNase, and TAPase activities, and the con-
ditions for measurements of the release of these activities from the particles
has already been given [7], Also, the estimations of RNA, protein, and
Mg + +, and the estimations of the release of these from the particles, has
been detailed [7]. RNA-P was calculated as being 9% of the RNA, The
incubations conditions for the experiments were as follows: The surface
of the pellets of the RNP particles was washed, and the particles resus-
pended by homogenization in 30% sucrose. Particles from 250 mg. wet
weight pancreas were incubated for 30 min. at 35° in a total volume of
2-0 ml., containing the various compounds to be tested, and with the
final sucrose concentration being 15%. After incubation, the contents of
two tubes were pooled, cold distilled water was added, to make the final
sucrose concentration 5%, and the suspension was then spun for 90 min.
at 105 000 X g in a Spinco Co. refrigerated centrifuge. Enzymic activities
were tested on the original particle suspension before incubation, and on
the supernatant and sometimes on the pellet after incubation and sedimen-
tation. In other experiments, the RNA, protein, and Mg + + contents of
the original particle suspension and of the re-sedimented particles were
measured. In some cases, where duplicate assays were particularly desirable,
as in the Mg + + determinations, more of the incubated particles were
pooled for sedimentation. In the cases where radioactive spermine was
used, the spun-down pellet was washed several times with cold distilled
water to remove possible contamination by unbound spermine and then
resuspended in water. In these experiments, one-half of the suspension
was used for RNA, protein, and Mg + + determinations, while the other
half was dried down on aluminum planchets, kept in a desiccator, and
counted with a Nuclear-Chicago gas-flow counter, with a counting error
of less than 1%. In these radioactive experiments, about 10 000 c.p.m.

EFFECTS OF SPERMINE ON THE RIBONUCLEOPROTEIN PARTICLES 24I

Spermine was used in each experiment, and from 250 c.p.m. to 1000 c.p.m.
were recovered in the particles in the various experiments; this is much
more than would be expected from contamination, since the particles
were well washed after centrifugation. In the experiments in which the
release of radioactive spermine from the particles was followed, the
following conditions were used: The particles were incubated with radio-
active spermine as given above. They were then centrifuged, washed
extensively, and resuspended by homogenization in cold distilled water.
An aliquot of this suspension was then removed for immediate chemical
and radioactive analyses, while another aliquot was incubated at 35° for
30 min. with various additions in a total volume of i -o ml. After incuba-
tion, water was added, the suspension centrifuged at 105 000 x g for 90
min., and the pellet was resuspended in water and assayed as above. All
the compounds used in the experiments were brought to pH 7-0. The
radioactive spermine, specific activity approximately o • i jU,c//xmole, was
labelled with "C in the i and 4 positions of the butane part of the mole-
cule ([i,4-i^C]-i,4-di-(i',3'-diamino propyl) butane) and was obtained
through the generosity of Drs. H. Waelsch and D. Clarke.

Results

In a previous paper [7], it was reported that P-P, or even ATP or GTP,
at a cone, of 5 x 10^* m, could release from 80 to 100° 0 of the bound
amylase, RNase, and TAPase activities of the RNP particles, after incuba-
tion of the particles at 35° for 30 min. Table I shows, again as a com-
parison, the results of further experiments with P-P and also the results
of experiments with various amines and other compounds. Spermine,
even at 5 x lo"^ M, could release the amylase activity from the particles.
In experiments not shown, spermine at io~* M was equivalent to P-P at
5 X ID"'* M in releasing also the RNase and TAPase activities of the
particles. Putrescine and cadaverine, while effective, could not release the
amylase activity as well as spermine, and it would seem that the ability to
release the enzymic activity is related to the number of amine groups and
hence the charge on the molecule. As can be seen, histamine, lysine, and
ornithine had some effect, while even tris buffer had an effect, but a
marked one only at 5 x lO"^ M. It is noteworthy that pilocarpine, car-
bamylcholine, and acetylcholine have very little, if any, effect on the
amylase release, as compared to incubation in water.

The P-P release of the bound enzymic activities of the particles is
probably a consequence of the concomitant release of all the Mg + + and
most of the RNA of the particles [7]. Table II shows, again further
experiments with P-P, and with various amine compounds, on the dis-
charge of Mg + + and of RNA from the particles. However, unlike P-P,

242 PHILIP SIEKEVITZ

TABLE I

Release of Amylase Activity from RNP Particles by Spermine,
Various other Amines, and Various other Compounds

Incubation conditions are given in the text. Per cent release is that percentage
of amylase activity which was rendered soluble after spinning the incubated
particles for 90 min. at 105 000 x g. Each figure represents one experiment.

Compound
added

Cone. (M)

Per cent
release

Compound
added

Cone. (M)

Per cent
release

Spermine

5

X lO"^

83,88

Tris, pH 7-0

5 X 10-*

33

Spermine

10-*

86, 90,

Tris, pH 70

10-3

63, 4i§

82,93

Tris, pH 70

5 X 10-=*

71, 46, 86

Spermine

5

X IO-*

97. 100.
99, 100*

In. Pyrophos.

5 X lo"*

91, 80, 100,
81, 99

Spermine

5

X IO~^

I oof

Serotonin

5 X 10-*

29

Putrescine

10-^

68+

Carbamylcholine

5 X 10-*

23

Putrescine

lO"-

72

Pilocarpine

5 X lo"*

19

Cadaverine

10-='

66

Acetylcholine

5 X 10 ■*

15

Histamine

10-3

48

Liver glycogen

I mg.

8

L-Lysine

10 ^

30

Water

—

8, 9, IS

L-Ornithine

10 3

32

18, 0, 21

L-Arginine

10 3

10

* Incubated at o"" = 83.
t Plus 10-3 m Mg + + = 96.
t Plus IO-' M Mg + + = 77.
§ Incubated at o"" = 32.

spermine only discharges a small amount of the RNA of the particles, and
in fact, the ratio, RNA /protein, of the spermine-treated particles is the
same or even higher than that of the particles incubated in water. In con-
tradistinction, the P-P treated particles have an RNA /protein ratio of one-
half (Table II), or even lower [7] than, that of the control particles. How-
ever, spermine, like P-P, releases nearly all the Mg + + of the particles
into solution. Putrescine, at io~^ m, behaves similarly to spermine at 5 x
io~* M or at io~* M, but at 2 x io~^ m putrescine begins to disrupt the
particles so that some of the RNA is released.

The reason given previously for the action of the pyrophosphate
compounds was that these compounds effectively compete with the
phosphate groups of the RNA for the Mg + + complexed to these groups.
Thus, the addition of Mg + + to the incubation medium along with the
pyrophosphate compounds nullifies the effect of these compounds [7].
However, when Mg + + was added with the spermine, it did not counteract
the effect of spermine on amylase release (cf. footnote. Table I). In fact,
it was uniformly found that in the presence of both Mg + + and spermine,

EFFECTS OF SPERMINE ON THE RIBONUCLEOPROTEIN PARTICLES 243

TABLE II

Effect of Various Amines and Inorganic Pyrophosphate
ON THE Release of RNA, Protein, and Mg + + from RNP Particles

Incubation conditions are given in the text.

Exp.

No.

Compound added

Left in RNP particles after spinning incu-
bated particles for 90 min. at 105 000 x g

None

Inorg. Pyrophos. (5 x 10"* m) 140
Spermine (10 * m)
Tris, pH 7-0 (lo"^ m)

RNA

Protein

RNA/

Mg + +

(^g.)

(Mg-)

Protein

(^moles)

362

728

0-50

—

140

515

0-27

—

355

555

0-64

—

353

555

063

—

2

None

351

550

0

•64

0

•08

Spermine (5

X

10-^

' m)

260

419

0

■62

0

•00

Putrescine (5

X

lO"

^m)

179

506

0

•35

0

•00

None 915

Inorg. Pyrophos. (5 x io~* m) 456

Spermine (10"* m) 935

Putrescine (io~^ m) 920

1250

0-73

0-30

II 90

0-38

0-04

1300

072

0-05

1270

0-73

oil

4

None

142

350

0

•41

0

•04

Spermine (10

^ m)

143

287

0

■50

0

GO

Putrescine (10

|-=*m)

III

281

0

•40

0

•01

5

None

112

256

0

•44

0

•03

Spermine (10"

-* m)

102

237

0

•43

0

00

Putrescine (2

x 10-3 m)

51

237

0

•22

0

•00

None 289

Spermine (10^^ m) 272
Spermine (10* m) plus Mg + +

(io~^ m) 321

0-08

002

the RNA /protein ratios of the incubated particles were the same or in
most cases higher than those of either the untreated particles or the
particles treated with spermine alone (Table III). This might be related to
the observation that in the presence of Mg + + in the medium, spermine
did not release the Mg + + from the particles (Table II, Exp. 6).

From the results of the above experiments, it would appear that the
action of the various amines, particularly spermine, can be described as
follows. Spermine replaces the Mg + + complexed to the phosphate groups
of the RNA of the particles by forming a strong salt bond between its
charged amine groups and the charged phosphate groups of the RNA.
This salt bond is stronger than the Mg + +-phosphate complex. The

244 PHILIP SIEKEVITZ

TABLE III

Effect of Spermine and of Mg + + on Release of RNA and
Protein from RNP Particles

Incubation conditions are given in the text.

Exp.
No.

Additions

Left in RNP particles after

spinning incubated particles for

90 min. at 105 000 x g

RNA

4 None

Spermine (10"* m)

Spermine (10^ m) plus Mg + + (io~^ m)

Protein

RNA/
Protein

I None 229 341 0-67

Spermine ( I o"* m) 209 278 0-75

Spermine (lO"* m) plus Mg + + (lo"^ m) 273 319 0-85

2

Spermine (10 * m)

Spermine (10 * m) plus Mg + + (10^ m)

188
246

272

275

0-69
0-89

3

Spermine (10"* m)

Spermine (10"* m) plus Mg + + (10"^ m)

116
156

250
271

0-47
0-58

211

315

0-67

154

282

0-55

176

287

062

0 6 12 18 24 30 36 42 48 54 60

GTP (xIO'^'m)

Fig. I . Effect of GTP concentration on the bound radioactive spermine, on the
RNA, and on the RNA/protein ratio of the RNP particles. Methods are given in
the text.

addition of Mg + + along with the spermine results in a very weak Mg + +-
spermine complex and hence has no effect on the salts formed between
the spermine and the phosphate groups. Since the spermine would appear

EFFECTS OF SPERMINE ON THE RIBONUCLEOPROTEIN PARTICLES 245

TABLE IV

Binding of Radioactive Spermine to RNP Particles, and
THE Effect of Mg + + on this Binding

Incubation conditions are given in the text.

Exp.
No.

Additions

/^moles

Spermine

spermme

bound

bound

(/tmoles)

per 10 ^moles

RNA-P

I

Spermine (5 x 10"* m)

0022

0-29

2

Spermine
Spermine

Jio"* m)

'10-* m) plus Mg + + (10-' m)

0-014*
0-031

0-31
0-74

3

Spermine
Spermine

'10* m)

,10* m) plus Mg + + (lo"' m)

o-oogt
0036

0-30
I -20

4

Spermine
Spermine

'10-* m)
lo"* m) plus Mg + + (lo"^ m)

0-034
0039

0-55
049

5

Spermine
Spermine

'10-* m)
IO-* m) plus Mg + + (10-3 m)

0-058
0-061

1-05
0-87

6

Spermine
Spermine

'10-* m)
IO-* m) plus Mg + + (10-^ m)

0049
0-049

1-49
I 09

7

Spermine
Spermine

10-* m)

10-* m) plus Mg + + (10^ m)

0-050
0014

I -12

026

* When pellet was treated with 5% TCA, then re-spun, and pellet recounted,
0002 /timoles radioactive spermine remained bound.

t When pellet was treated with 5% TCA, then re-spun, and pellet recounted,
o-ooiS ^moles radioactive spermine remained bound.

to replace the Mg + + from the particles, and since the Mg + + seems to be
one of the principal instrumentalities in stabilizing these particles (cf. ref.
in [7]), it should follow that spermine also helps to bind the RNA to the
protein of the particles, and as Tables II and III indicate, it seems to do
so.

Since we were doubly fortunate to be in New York City and knew of
certain experiments performed by Dr. Waelsch and his associates, and
were also friends of Dr. Waelsch and his associates, we were able to test
the above possibility by obtaining radioactive spermine from them. The
results are shown in Tables IV and V and Fig. i. Table IV shows, indeed,
that spermine is bound to the particles, while Table V shows that by far
the greatest part of the bound spermine cannot be washed off the particles
by further water incubation of the particles. Previously [7], we had found

246

PHILIP SIEKEVITZ

TABLE V

Effect of GTP and of Mg + + on the Release of Radioactive Spermine, and
OF Protein and RNA from RNP Particles

The RNP particles were labelled with radioactive spermine as described in
the text. These particles were then re-suspended and re-incubated at 35° for 30
min. under the conditions given in the text and with additions given below.
After re-incubation, the particles were spun at 105 000 x g for 90 min., and
radioactivity, protein, and RNA determined as described in the text.

Left in

RNP particles

after spinning re-

Additions

incubated

1 particles for 90 min. at 105

000 X g

Exp.

No.

Protein

RNA

(/Ltmoles
RNA-P)

RNA/
Protein

Spermine
1 (/xmoles)

/i.moles

spermine

per 10

/Limoles

RNA-P

I

None

—

—

—

o-oi8

—

Mg + + (10-2 m)

—

—

—

o-oiS

—

2

Control (no re-incubation)

250

0-33

0-47

0-049

1-49

None

150

0-25

0-58

0-028

I -12

Mg + + (10-2 m)

—

—

—

0-023

—

GTP (5 X 10-* m)

—

—

—

o-oi6

—

GTP (5 X io'''M)plus

Mg + + (3 X 10-' m)

—

—

—

0-022

—

3

Control (no re-incubation)

282

0-44

0-55

0-050

I 12

None

259

0-37

0-49

0-034

0-92

GTP (2 X lo-^* m)

122

004

o- 12

0-004

I 00

GTP (2 X 10^ M) plus

Mg + + (3 X 10-=' M)

188

023

0-44

0026

113

4

Control (no re-incubation)

275

0-54

0-67

0-069

1-28

None

231

0-37

0-56

0-045

I -22

GTP (2 X 10-3 m)

187

o- 14

0-27

0-014

0-97

GTP (2 X 10-3 m) plus

Mg + + (3 X 10=* m)

253

0-32

0-43

0-036

I - 12

ATP (2 X 10-3 M)

228

o-i8

0-28

0-019

1-08

ATP (2 X 10 3 M) plus

Mg + + (3 X 10-3 m)

250

0-27

0-38

0-034

1-26

that the ratio, /xmoles Mg + +/io /xmoles RNA-P, in the RNP particles
was approximately one. In Table IV it can be seen that this ratio for
spermine varied from 0-29 to i -49 in the various experiments; or, if put
in another way, from o -6 to 3 -o /xmoles primary amine /lo jumoles RNA-P,
or from 1-2 to 6-o /xmoles total amine group /lo /imoles RNA-P. It can
be noticed that these experiments of Table IV fall into three classes ; one

EFFECTS OF SPERMINE ON THE RIBONUCLEOPROTEIN PARTICLES 247

in which relatively low binding of spermine occurred but the addition of
Mg + + increased it (Exps. i, 2, 3) ; one in which an intermediate uptake of
spermine occurred and Mg + + had little effect (Exps. 4, 5); and one in
which high binding of radioactive spermine occurred but the addition of
Mg + + decreased this binding (Exps. 6, 7). The possible reasons for these
variations will be discussed below.

That the binding of the spermine is primarily due to the RNA of the
particles can be inferred from the concluding data. First, washing of the
particles with TCA releases most of the spermine counts (cf. footnote,
Table IV), indicating a non-covalent bond. Secondly, it can be seen
(Table V and F'ig. i) that the radioactive spermine which is bound to the
particles can be largely removed by treating the particles with either GTP
or ATP, and in doing so the RNA is also removed. Figure i shows that
as the bound spermine is removed from the particles, the RNA is propor-
tionally removed, whereas very little protein is lost. Hence, the ratio,
RNA /protein, follows the removal of the spermine and of the RNA, as
the GTP concentration is increased. Table V also indicates that when
Mg + + is incubated along with the GTP or ATP, the removal of spermine
by GTP or ATP is largely prevented, indicating that the added Mg + +
effectively forms a complex with the added GTP or ATP and prevents
its action in removing the spermine and the RNA. Thus it can be deduced
that the amine groups of the bound spermine were salt-bonded to the
phosphate groups of the RNA, and the addition of the pyrophosphate
compounds, GTP or ATP, effectively competes with the phosphate
groups of the RNA, for the bound spermine. When Mg + + was present,
the pyrophosphate groups were complexed with it, and hence had little
effect.

Discussion

Firstly, it should be understood that the pancreatic RNP particles do
not seem to be quite the same as the bacterial and yeast or even the liver
particles. Their size is different ; 150 A for the 85 S pancreatic particles [7],
and 200 A for the 80 S yeast particles [8] and for the 70 S Escherichia coli
particles [9, 10]. Furthermore, it appears that the E. coli particles are made
up of sub-units of 30 S and 50 .S" particles [9, 10, 11, 12] and of 100 S
dimers [9, 10, 11]. There is no evidence as yet, either morphologically or
with the analytical centrifuge, that the pancreatic particles are made up of
smaller units. There is evidence, though, that this might be the case for
the 80 S pea seedling particles [13] and the 80 S liver particles [14], but
even in these cases the sub-units seem to be 60 S and 40 S particles [13,
14]. In all the cases mentioned, except for pancreas, a decrease in the Mg + +
content of the particles results in their disruption, so that the smaller

248 PHILIP SIEKEVITZ

sub-units have the same RN A /protein ratios, indicating that the Mg^ + is
invohed in RNA to RNA binding and not in RNA to protein binding.
The reasons for this apparent discrepancy between these resuhs and ours
is not known (cf. [7]); it could be that while in the pea seedling particles
there is about i mole Mg + + for every four phosphates in the RNA [13]
and while the Mg + + requirement for the E. coli particles seems to be very
high [11], in the pancreatic particles there is only about i Mg + + for every
ten phosphates [7].

In our view, there are three components to be considered in the
pancreatic RNP particles ; the RNA and proteins themselves, and the sub-
divisions of the latter into the so-called "structural" proteins of the
particles, and into the proteins which are being synthesized by the particle
and some of which still reside there, on the "template" if you wish. The
reason for the latter part of the statement are the earlier findings [7] and
the present ones, that Mg + + chelators and replacers can release some of
the proteins of the particle without releasing the bulk of them. Even the
RNA of the particle might not be homogeneous, for we have never been
able to remove all the RNA from the protein of the particle, either by Mg + +
chelators [7] or by the removal of the spermine in the present experiments.
Thus it could be that a small proportion of the RNA is more tightly bound
to the protein, and could be metabolically different, than is the bulk of
the RNA.

What are the probable bindings between RNA and proteins in the
pancreatic particles? There is evidence that hydrogen bonding [15, 16],
electrostatic bonding [17] and Mg + + complexing (cf. literature cited in
[7]) are all somehow involved. I will just discuss the Mg + + binding and the
effects of spermine on this binding. We must look to two possibilities ; that
the Mg + + ties together various RNA chains in the RNP and thus stabilizes
the hydrogen and electrostatic bonds between the RNA and the protein
of the RNP [15], or that Mg + + links the RNA chains both to the structural
and to the newly synthesized proteins of the RNP (cf. [7]). It is difficult
to decide between these two not mutually exclusive propositions, for it is
conceivable that if Mg + + complexes adjacent RNA chains, the removal of
this Mg + + so weakens the hydrogen and electrostatic bonds between
RNA and protein that these groups become separated from each other.
It is thus clear that no one binding force seems adequate by itself to explain
the structural stability of the particles. However, it would appear that the
Mg + + complexing is of utmost importance, at least in the pancreatic
particles. Once the Mg + + is removed, and not replaced by other binding
agents, such as spermine, then any hydrogen bonds or ionic forces which
might hold the RNA to the protein are weakened, and the particle loses
identity as an RNP particle. In the case of the particles from other sources,
it would appear as mentioned, that Mg + + links together the RNA chains

EFFECTS OF SPERMINE ON THE RIBONUCLEOPROTEIN PARTICLES 249

within one particle. There is no reason to disbeHeve that the same does not
also hold for the pancreatic particles. For example, when spermine is
added to a suspension of these particles, they immediately begin to
aggregate, as if the spermine is providing more groups, in this case the
positively charged amine groups, for attachment to any negatively charged
groups in the vicinity, be they the phosphate groups in the RNA of a
single particle or the phosphate groups in the RNA of adjacent particles,
the latter explaining the aggregation.

The experiments mentioned in this paper, while clarifying some
aspects of the RNP particle structure, also add some complexities to the
picture, for which no adequate answer is available at present. It is clear,
for example, that in replacing the Mg + + of the particle, spermine dis-
charges only some of the proteins from the particles. A theory based on
this finding is as follows. The Mg + + is complexed to the phosphate groups
of adjacent RNA chains ; but it is possible that the amino groups of adenine
and guanine also contribute to the binding energy of the complex [i8].
When spermine is added, Mg + + is removed, and the salt linkages formed
by the spermine amine groups and the RNA phosphate groups are strong
enough to stabilize the particle in the absence of Mg + +, by also preserving
any hydrogen and ionic bindings between the RNA and protein. However,
some of the proteins, namely, the secretory enzymes, presumably the newly
synthesized proteins [6], are discharged from the particle. It is our con-
tention that these particular proteins are held on to the RNA {" template " .'')
by the bridging, by some of the Mg + +, between the phosphate groups of
the RNA and the amino groups of the peptide bond or the free amine
groups of the basic amino acids. The amine groups of spermine cannot
replace the co-ordination bonds of Mg + + in such a situation and the
enzymes are discharged from the particles. Since there is only one Mg + +
for every ten phosphates of the RNA, it is conceivable that all of the Mg + +
is involved in holding these newly synthesized proteins on to the particle.
For, since Mg ^ ^ has a co-ordination number of four and possibly six
[19], it is possible that some of these co-ordination bonds are involved in
only phosphate complexing, while others are involved in complex formation
between the phosphate groups and the carbonyl groups of the structural
protein or the amino groups of the synthesized proteins.

However, it appears that the binding of the RNA molecules to each
other or to the proteins is not as secure as might be assumed for a stable
configuration. It had been previously found that when all the Mg + + and
most of the RNA of the particles are removed, the remainder of the
particle has a diameter almost twice that of the original particle [7].
Even when the RNA is not removed, as in the present experiments with
spermine, the particle diameter becomes twice that of the original particle,
as found in electron micrographs by G. E. Palade. This could be explained

250 PHILIP SIEKEVITZ

by assuming that the particle is somewhat an open structure,* and the
compactness of the particle can be changed by the availability of bonding
forces, in this case, the amine groups of the spermine and the phosphate
groups of the RNA (cf. [20]). These variations in the compactness of the
particles can explain the variations noted in Table IV. Firstly, it is probable
that at pH 7 all the basic groups of the spermine are charged (cf. [21]). The
amount of spermine binding to the particles could be determined by the
availability of phosphate groups open to salt linkage. The farther apart
are the available phosphate groups, then the greater the number of
molecules of spermine which are bound, for only one amino group per
spermine can be bound. f When the phosphates are closer together, the
less spermine molecules will be bound, because all the available phosphate
groups in the vicinity can be linked to the amine groups of the same
spermine molecule. It thus might be of significance that the amount of
spermine bound per RNA goes up by factors of approximately two and
four (Table IV), as if one, two, or all four of the amine groups in spermine
are becoming involved. Thus, in the various experiments cited in Table IV,
the RNP particles as isolated may be greatly different in their compactness
of structure. When this structure is loose, more spermine can be bound
(Table IV, Exps. 5, 6, 7). In the presence of Mg + +, the intermolecular
distances might become smaller owing to the formation of Mg + +-spermine-
phosphate complexes (cf. [19]). Hence the available phosphate groups can
be linked to more than one amine groups of the spermine molecule, thus
less spermine molecules are bound in the presence of Mg + + (Table IV,
Exps. 6, 7). When the structure is tight, less spermine is bound (Table
IV, Exps. I, 2, 3). In the presence of Mg + +, again Mg + +-spermine-
phosphate complexes can be formed, but in this case these complexes can
open up a resilient structure, allowing more spermine to be bound (Table
IV, Exps. I, 2, 3). As disclosed in Table III, and has been found previously
[22] with E. coli particles, the particles have higher RNA /protein ratios
in the presence of both spermine and Mg + + than in the presence of either
alone. Also, when the particles are incubated in the presence of both
spermine and Mg + +, the spermine becomes bound, and no Mg + + or
RNA leaves the particle. Furthermore, nucleotide incorporation into RNA
by liver preparations can be stimulated by the addition together of spermine

* The mean diameter of the particles under the experimental conditions
mentioned, as found by G. E. Palade, is as follows: As isolated, 150 A; after
incubation in sucrose, 180 A; after incubation in ATP, P-P, or versene, 250-
300 A; after incubation in spermine, 400 A; after incubation in spermine plus
Mg ^ ■*■, 300 A. In some cases, when the RNA is removed from the particles, a less
dense centre appears, giving the particle a doughnut appearance [7].

t It has been estimated [20] that the distance between the terminal primary
amines in spermine is approximately 16*5 A.

EFFECTS OF SPERMINE ON THE RIBONUCLEOPROTEIN PARTICLES 25 1

and Mg + + [23]. These findings are not in disagreement with the existence
within the particles of Mg ^ +-spermine-phosphate complexes.

The finding that spermine could release the enzymes from the particles
without disrupting them prompted us to look for this polyamine in the
particles. We were encouraged by a report [24] that in E. coli and in liver
particles there exist some of these polyamines, namely putrescine, cadave-
rine, and spermidine. We have used the electrophoretic and chromato-
graphic methods mentioned [24] and have tentatively identified spermine
and perhaps spermidine in rat liver, guinea-pig liver, and guinea-pig
pancreas particles.* Contrary to the above, we could find no cadaverine
nor spermidine in the rat liver particles using the methods mentioned ; but
we also found another compound (cf. [24]) which gave a ninhydrin colour
and was positively charged, but whose mobility was different from that
of the polyamines mentioned and the basic amino acids. Thus, these
polyamines could well be the physiological agents which might effect a
release of synthesized enzymes, leaving the RNP structure intact. A
possible physiological mechanism regulating this release would thus have
to do with Mg + + availability and displacement by spermine.

However, it could be that spermine acts in these particles as a stabiliz-
ing agent, helping to hold the particle together, as it has been found to
preserve nucleic acid structure [25, 26], to preserve bacterial [27, 28]
and protoplast [27, 29] structure, as well as conferring stability on mam-
malian mitochondria [29] and nuclei [20]. Also, it has been found that
these polyamines can bind strongly to polynucleotides [30] and to phage
[21], and thus to stabilize and neutralize the DNA of the latter [21].

There have been reports that the total proteins of the RNP particles
from liver [31, 32], reticulocytes [31] and pea seedlings [31] contain
relatively large amounts of basic amino acids, but other findings suggest
that this is not the case for E. coli particles [15], nor for liver particles [33].
If this be the case, it is hard to see, from the discussion above, how these
basic amino acids contribute to the binding of the RNA to the protein,
particularly in light of the finding that even histidine does not bind
strongly to certain synthetic polynucleotides [30].

References

1. Palade, G. E., and Siekevitz, V.,J. biophys. biochem. Cytol. 2, 671 (1956).

2. Siekevitz, P., and Palade, G. E., ibid. 4, 203 (1958).

3. Siekevitz, P., and Palade, G. E., ibid. 4, 309 (1958).

4. Siekevitz, P., and Palade, G. E., ibid. 4, 557 (1958).

5. Siekevitz, P., and Palade, G. E., ibid. 5, i (1959).

* However, it is possible that in vivo the particles do not contain these poly-
amines, for their presence in the isolated particles may be a result of redistribution
during homogenization, from a soluble state to that of being bound to any available
phosphate groups of the RNA.

252 PHILIP SIEKEVITZ

6. Siekevitz, P., and Palade, G. E., ibid. 7, 619 (i960).

7. Siekevitz, P., and Palade, G. E., ibid. 7, 631 (i960).

8. Chao, F.-C, and Schachman, H. K., Arch. Biochem. Biophys. 61, 220 (1956).

9. Hall, C. E., and Slayter, H. S., J. mol. Biol. I, 329 (1959).

10. Huxley, H. E., and Zubay, G., J. mol. Biol. 2, 10 (i960).

11. Tissieres, A., Watson, J. D., Schlessinger, D., and Hollingsworth, B. R., J'.
mol. Biol. I, 221 (1959).

12. Report of Biophysics Section, Dept. Terr. Mag., Carnegie Inst. Wash. Year
Book (1959)-

13. Ts'o, P. O. P., Bonner, J., and Vinograd, J., Biochim. biophxs. Acta 30, 570
(1958).

14. Hamilton, M. G., and Petermann, M. L., J. biol. Chem. 234, 144 1 (1959).

15. Elson, D., Biochim. biophys. Acta 36, 362 (1959).

16. Hall, B. D., and Doty, v'.,J. mol. Biol. I, 1 1 1 (1959).

17. Tashiro, Y., and Inouye, A., J. Biochem., Tokyo, 46, 1625 (1959).

18. Zubay, G., Biochim. biophys. Acta 32, 233 (1959).

19. Chaberek, S., and Martell, A. E., "Organic Sequestering Agents". J. Wiley
Sons, New York (1959).

20. Anderson, N. G., and Norris, C. B., Exp. Cell Res. 19, 605 (i960).

21. Ames, B. N., and Dubin, D. T., J. biol. Chem. 235, 769 (i960).

22. Cohen, S. S., and Lichtenstein, J.,^ biol. Chem. 235, 21 12 (i960).

23. Chung, C. N., Mahler, H. R., and Enicore, M., jf. biol. Chem. 235, 1448

(i960).

24. Zillig, W., Krone, W., and Albers, M., Hoppe-Seyl. Z. 317, 131 (1959).

25. Doctor, P. B., and Herbst, E. J., Fed. Proc. 17, 212 (1958).

26. Eraser, D., and Mahler, H. R., ^. Amer. che?n. Soc. 80, 6456 (1958).

27. Mager, J., Biocliim. biophys. Acta 36, 529 (1959).

28. Herbst, E. J., Weaver, R. A., and Keister, D. L., Arch. Biochem. Biophys. 75,
171 (1958).

29. Tabor, C. W., Biochem. Biophys. Res. Comm. 2, 117 (i960).

30. Felsenfeld, G., and Huang, S., Biochim. biophys. Acta 37, 425 (1960).

31. Ts'O, P. O. P., Bonner, J., and Dintzis, H., Arch. Biochem. Biophys. 76, 225
(1958).

32. Crampton, C. F., and Petermann, M. L., J. biol. Chetti. 234, 2642 (1959).

33. Butler, J. A. V., Cohn, P., and Simson, P., Biochim. biophys. Acta 38, 386 (i960).

Discussion

HuLTiN : Did you say that if you treated these particles with spermine then they
would swell up ? If you now treat these gels with magnesium, what happens ?

Siekevitz: Yes, they do swell up. We haven't done what you suggested, but
we have incubated them with spermine + magnesium and they looked the same
way as with spermine alone.

Bergeron : Did the aggregates seen in the electron micrographs of the pellets
precede fixation or arise in the course of fixation ?

Siekevitz : This is a visible aggregation during the centrifugation. In the
absence of spermine when you centrifuge these particles they form a nice pellet
on the bottom. In all these other cases they precipitated on the side of the tubes
indicating that they were aggregating during centrifugation, and in the case of the
spermine, they aggregated before centrifugation.

EFFECTS OF SPERMINE ON THE RIBONUCLEOPROTEIN PARTICLES 253

Hunter: I was just thinking that if manganese had a greater affinity than
magnesium, that would help to get the spermine out.

SiEKEViTZ : Actually I should say that we only looked for magnesium in these
particles. There might be manganese there or even calcium. Dr. Ts'O has found
that in pea seedling particles there was five times as much magnesium as calcium.

Hunter: Even if manganese were not present physiologically, it still might be
useful for getting spermine out.

SiEKEViTZ : Yes.

The Correlation Between Morphological Structure and

the Synthesis of Serum Albumin by the Microsome

Fraction of the Rat Liver Cell

P. N. Campbell
Courtauld Institute, Middlesex Hospital, London, England

Our interest in serum albumin is based on the assumption that it is
typical of the soluble proteins synthesized by the liver. It happens to be a
rather convenient protein to study since it is homogeneous and not too
difficult to isolate in a pure form. I should like to describe some experiments
we have done in an attempt to determine the role of the various morpho-
logical structures that go to make up the microsome pellet in the synthesis
and secretion of serum albumin.

Peters [3] has shown in his work with slices of chick liver that the
microsome fraction is the most active site in the cell for the synthesis of
albumin. We have for some time been working with the isolated microsome
fraction from rat liver. We now have good evidence that such a fraction is
able to synthesize serum albumin. Initially we demonstrated the incorpora-
tion of P^C]-amino acids into the serum albumin released from the micro-
some pellet by ultrasonics. The albumin was characterized by immuno-
logical, electrophoretic and solubility criteria [i]. More recently we have
been studying the pattern of incorporation of [^^C] -leucine under in vivo
and in vitro conditions. We hope by this means to determine whether there
is de novo synthesis under in vitro conditions. So far the results obtained
are consistent with this concept.

Although the above experiments show that the microsome fraction is
an active site for albumin synthesis they do not preclude such synthesis by
other organelles such as the mitochondria. This fraction contains significant
amounts of serum albumin and may also be a site for synthesis.

As explained by Dr. Porter in his paper the predominant structure seen
in electron micrographs of the microsome fraction of rat liver is the so-
called rough endoplasmic reticulum. As Dr. Porter has shown, this consists
of vesicles bearing ribonucleoprotein (RNP) particles on the outside.

The experiments of Peters, and those on the isolated microsome
fraction, show that the albumin that is synthesized in the reticulum is at
first retained. In fact with the isolated microsome system the synthesized
albumin is not released into the soluble fraction of the incubation medium.

256 p. N. CAMPBELL

The newly synthesized albumin can be released from the membranes
either by treatment with deoxycholate (DOC) or by breaking them up with
ultrasonic vibrations. Our first idea was to determine whether the protein
released by DOC which appeared to be serum albumin by its immuno-
logical properties, had in fact all the properties of serum albumin.

We have examined extracts of the microsome pellet obtained either by
DOC extraction or ultrasonics. Immunoelectrophoresis on agar gel was
used with an antiserum prepared against a purified preparation of rat serum
albumin. We can find no evidence for the existence of precursors of
albumin with a molecular weight of less than that of the whole molecule.
This finding is of special relevance in the case of serum albumin since
Lapresle [2] and Porter [4] have both shown that degradation products
of serum albumin retain some of the immunological properties of the
native protein.

What can be learned from the nature of the binding between the
membranes and serum albumin ? As already mentioned, if the microsome
pellet is treated with ultrasonics there is a certain release of serum albumin.
If we centrifuge down the residue and treat with deoxycholate we get a
further release of albumin. The comparable amounts are approximately
two-thirds by ultrasonics and a further one-third by DOC. We are at
present carrying out studies with the electron microscope to determine the
nature of the action of ultrasonics. These studies are far from complete
but we are accepting as a working hypothesis that the ultrasonic vibrations
disrupt the membranes of the vesicles and permit the soluble contents to
be released. The albumin which is only released by DOC we assume to
be particularly closely associated with the lipid substances that go to make
up the fabric of the membranes.

In order to determine whether there was any metabolic difl^erence
between the albumin in the two fractions we injected a rat with a mixture
of p^C]-labelled amino acids and killed it after 3 min. We homogenized
the liver, isolated the microsome fraction and washed it free of the cell
sap (105 000 g supernatant). We then determined the specific activity of
the albumin in the two fractions. We found that the DOC albumin was
40% more active than the ultrasonic albumin whereas at this short time
after injection the cell sap albumin was not significantly radioactive. We
have also done similar estimations after the incubation of isolated liver
microsomes with P^C]-leucine. In this case the DOC albumin was about
four times as radioactive as the ultrasonic albumin.

We cannot at present interpret these results with certainty but they

are consistent with the idea that some stage of the synthesis of albumin

involving incorporation of amino acids takes place in the lipid membrane.

Next we must turn to the ribonucleoprotein (RNP) particles. Dr.

Siekevitz has described in his paper experiments with gumea-pig pancreas

CORRELATION BETWEEN MORPHOLOGICAL STRUCTURE 257

that suggest strongly that the RNP particles attached to the endoplasmic
reticulum are the site of synthesis of the completed molecules of chymo-
trypsinogen and other hydrolytic enzymes of the pancreas. We have natur-
ally been interested to determine whether the same holds true for the rat
liver so far as albumin synthesis is concerned.

Two aspects of this problem have occupied us. First we wanted to
know^ if the RXP particles contained any serum albumin. We have prepared
many batches of such particles after treating the microsome pellet with
DOC, extensive centrifugation at 105 000 g for 2^ hr., washing free of
DOC and further centrifugation. Such particles have the usual RNA/
protein ratio of about 0-4. We have treated the particles with a variety of
agents and compared the amounts of soluble albumin in the supernatants
with that obtained by suspension in water. We have incubated the particles
in each of the following: pyrophosphate (5 x lO"^ u), ATP (5 x iq-* m),
NaCl (0-5 m), ethylenediaminetetra-acetic acid (2^0). ribonuclease (o • 5
mg./ml.) and, finally, 70% aq, acetic acid. None of these reagents had any
detectable effect on the immunological properties of serum albumin. We
have not obtained a consistent release of serum albumin from the particles
by any of these treatments. With ribonuclease the RNA/protein ratio was
reduced to about o-i, which may be taken to indicate a considerable
degree of disruption of the particle. So far, therefore, we have no evidence
of the presence of serum albumin in the RNP particles.

The other aspect of our work has been an attemipt to obtain a synthesis
of serum albumin wdth the isolated RNP particles. We have studied the
conditions for the incorporation of amino acids into the protein of such
preparations. Like others, we find that provided the deoxycholate is
removed from the particles it is possible to obtain preparations which are
consistently active. It is not necessary to go into details, for our results
are very similar to those already described by Dr. Hultin. As with our
experiments on the synthesis of serum albumin by the isolated microsome
system, our initial aim w^as to demonstrate the incorporation of amino
acid into the purified serum albumin. The difficulty with the RNP particles
is that the cell sap is required for optimum activity and this contains large
amounts of serum albumin. Unlike the microsome preparation the particles
w^ould not be expected to retain the newly synthesized albumin so that the
latter would be diluted with the inactive albumin already present in the
cell sap. We have in fact failed to demonstrate an energy-dependent
incorporation of amino acid into the serum albumin on incubation of the
particles.

To summarize therefore we have no evidence that the RNP particles
of rat liver are the site for the synthesis of the complete albumin molecule.
At present the results are consistent with the idea that in the liver the
initial incorporation of amino acids is into the RNP particles and that an

258 p. N. CAMPBELL

albumin precursor then moves to the hpid membranes where there is a
further addition of amino acids to complete the albumin molecule. The
albumin then moves to the internal medium of the vesicles where it is
securely retained until it is passed through the cell membrane.

I am grateful for the expert assistance of Miss Barbara Kernot in all
the experiments described which were carried out in our laboratory. I am
also grateful for the encouragement of Professor F. Dickens. The work
described was made possible by a grant to the Medical School from the
British Empire Cancer Campaign,

References

1. Campbell, P. N., Greengard, O. and Kernot, B. A. Biochem.J. 74, 107 (i960).

2. Lapresle, C, Ann. Inst. Pasteur 89, 654 (1955).

3. Peters, T., ]r.,J. Histocheni. Cytochem. 7, 224 (1959)

4. Porter, R. R., Biochem.J. 66, 677 (1957).

Discussion

Porter : I was impressed by the rate at which serum albumin was synthesized
after the intravenous injection of labelled amino acids and I wondered how much
synthesis took place during homogenization and fractionation of the particles of the
microsomes.

Campbell : This seems unlikely for the conditions which exist during the
homogenization and fractionation procedure are not favourable for the incorpora-
tion of much [^^C]-amino acid. The concentration of amino acid would be very
small owing to the addition of medium, the concentration of ATP and other energy-
rich compounds would be low, and the temperature is also low. We hope that the
finger-print method may prove to be a useful technique. A mere demonstration
of radioactivity in a protein after injection should not really be taken as evidence of
complete synthesis. We think that by this method it would be possible to demon-
strate a synthesis of complete molecules.

Davis : I wonder whether the fact that you find uniform distribution of radio-
active leucine in serum albumin after only 3 min. might not be compatible with the
following alternative interpretation. The individual amino acid molecules might
line themselves up at random times, with no one site having a preferential position
compared with another in relation to time. That is, the first few amino acids might
just as well go on at one end as at the other or even in the middle. When the entire
set of slots were filled peptide bond formation would be completed and the poly-
peptide chain would zip off. Therefore during that 3 min. the radioactive leucine
might have found some templates which were almost completed but missing a few
leucines here, and other templates missing a few leucines there. All the proteins
that did get produced at that time would average out a random distribution of
radioactive leucine.

Campbell: Yes, that is a most interesting possibility. If it were true we should
not find any change in the pattern even when the animal was killed at a very short

CORRELATION BETWEEN MORPHOLOGICAL STRUCTURE 259

time after injection. The problem is of course to get enough radioactivity incor-
porated at the very short times for without this it is not possible to study the
pattern of labelling.

HoLTER : I think that we need not be afraid of the short time of 3 min. The group
at the Department of Terrestrial Magnetism of the Carnegie Institution in Wash-
ington found in E. coli incorporation times of the order of seconds, so I don't
expect there is any reason to believe that in higher organisms it should go so much
slower than in E. coli.

Campbell : There is only one snag, it is presumably easier for the amino acids
to get to the site of protein synthesis in E. coli than it is in this case.

Allfrey : I think that one answer to Dr. Davis's question comes from the work
of Dr. Friedrich-Frekse's department at Tubingen. In this department they have
shown that exposure to U.V. or X-irradiation affects protein synthesis and they have
deduced that the laying down of protein proceeds from one end of the chain ; if
it is interrupted by irradiation at any stage, then the synthesis of that particular
chain is stopped.

Porter: It is an extraordinary thing but I do not believe anyone has observed
in electron micrographs of liver cells any evidence of accumulation of material
within the extremely flat cysternae of the ER. I imagine that it could be brought
out by physiological experiments.

Williams: I wonder whether you could tell us anything about the serum
albumin in the mitochondria. Firstly, is it readily released and secondly, does it
incorporate amino acids rapidly ?

Campbell: The answer to the second question is that we don't know. As to the
serum albumin being readily released, once again Peters has done a considerable
amount of work on this with the chick liver and we have worked on the rat. Serum
albumin is released with DOC so it is certainly held in quite nicely.

Mitchell : Can you tell us more about the lipid membrane that you men-
tioned. I believe that Dr. Siekevitz in his very nice Ciba Foundation review said
that he did not think that the membrane of the reticulum had very much lipid in
it.

Siekevitz: If you isolate liver microsomes and test for phospholipid you find a
great deal, but if you isolate the pancreas microsomes you find very little, using the
same methods of extraction.

Amino Acid Transport and Early Stages in Protein
Synthesis in Isolated Cell Nuclei

Vincent G. Allfrey
The Rockefeller Institute, New York, U.S.A.

The application of tracer techniques to the study of nuclear protein
synthesis dates back to 1948, to the exciting experiments of Bergstrand,
Eliasson, Hammarsten, Norberg, Reichard and von Ubisch ; they showed
that, after the injection of [^^N]-glycine, the ^'N concentration in the pro-
teins of the liver cell nucleus was just as high as that of proteins in the
cytoplasm [i]. This great disclosure of the dynamic state of protein meta-
bolism in the interphase nucleus is now a classic observation in nuclear
biochemistry and it has become a point of departure for many subsequent
experiments on protein synthesis in the nuclei of intact organisms.

Until 1954, virtually all tracer work on nuclear protein synthesis was
based on similar experiments carried out in vivo, in which nuclei or nuclear
protein fractions w^ere isolated from the tissues of animals previously
injected with isotopic amino acids. (This procedure is still clearly the
method ot choice for studying many aspects of the physiology of the
nuclear proteins (e.g. [2, 3].)

However, in 1954 the surprising synthetic potentialities of the isolated
thymocyte nucleus came to light [4], and it became possible to study the
chemical mechanisms employed by the isolated nucleus in incorporating
free amino acids into polypeptide chains. This aspect of nuclear protein
synthesis has held our interest for the past five years, and some of the re-
actions and sites involved in amino acid incorporation are now known.
Part of this work, done in collaboration with Drs. A. E. Mirsky, J. W.
Hopkins, and J. H. Frenster, will now be described. It presents our current
views of the protein synthetic pathway in the cell nucleus. The treatment
of previously published work will omit details in order to permit the des-
cription of some newer aspects of nuclear amino acid and protein
metabolism.

Pathways in nuclear protein synthesis

AMINO ACID UPTAKE BY ISOLATED CELL NUCLEI

Suspensions of nuclei possessing a high degree of purity can be ob-
tained by differential centrifugation of homogenates of calf thymus tissue.

262

VINCENT G. ALLFREY

We have shown that if thymus nuclei are isolated under isotonic con-
ditions in a medium containing 0-25 M sucrose and 0-003 ^ CaCl2, they
are able to retain their nucleic acids [5, 6], soluble enzymes [7], and low-
molecular weight compounds such as mononucleotides [8] and free
amino acids [9].

The isolated thymocyte nucleus also retains a surprising number of its
original synthetic capabilities. These include the synthesis of ATP by an
aerobic phosphorylating mechanism [8], the incorporation of thymidine
into deoxyribonucleic acid [10], and the utilization of a variety of purine
and pyrimidine precursors for ribonucleic acid synthesis [5, 6, 11, 12] and
partial turnover [11].

Nuclear activity in amino acid incorporation experiments is especially
striking [4, 5]. The latter process is illustrated by the uptake curves shown

700

Time Course
of C'^-Amino Acid Incorporation

pH67

/Lysine-2-C

200

/ Alanine-1-C
/ ♦

100

//

/

/ /

Glycine-1-C

30 60 90
Time (min)

120

Fig. I. The time course of incorporation of [**C]-labelled amino acids into
the total proteins of isolated thymus nuclei during incubation /// vitro.

in Fig. I ; these show the time course of labelling of the nuclear proteins
during incubations in the presence of either [i-^^C]-alanine, [i-^*C]-
glycine, or [2-^'*C]-lysine.

There are two aspects of this incorporation process which mark it as
characteristically nuclear. First, amino acid uptake is sodium-ion depen-
dent. (The reasons for this are discussed below.) Secondly, amino acid
uptake does not take place if the nuclei are first treated with crystalline
pancreatic deoxyribonuclease. This evidence for the DNA-dependence of
nuclear amino acid incorporation is further supported by observations
that added DNA always restores uptake to DNAase-treated nuclei [5].
(The reason for this is now clear, since recent experiments have shown
that nuclear ATP synthesis is DNA-dependent [13] ; removal of the DNA
stops nuclear phosphorylation, and without the necessary ATP, nuclear
protein synthesis cannot proceed [5, 8]. Restoring the DNA (or substitut-

NUCLEAR PROTEIN SYNTHESIS 263

ing Other large polyanionic molecules for it) restores nuclear ATP syn-
thesis; as a result, other energy- dependent synthetic reactions in the
nucleus can resume [14].

These two facets of amino acid incorporation in the intact isolated
nucleus (sodium ion requirement and DNA-dependence) set it apart from
protein synthesis in the cytoplasm, which requires potassium, not sodium
ions [15], and is sensitive to ribonuclease [16] and not to deoxyribonuclease.
Yet a detailed study of the mechanism of protein synthesis in sub-nuclear
fractions leads to the conclusion that the protein synthetic pathway in the
nucleus is essentially the same as that found in the cytoplasm (and previously
discussed in this symposium by Drs. Zamecnik, Hultin, Siekevitz, and
Campbell).

AMINO ACID ACTIVATION

In the nucleus, as in cytoplasmic and bacterial systems, the primary
step in protein synthesis appears to be an activation reaction requiring the
participation of ATP. The reaction is mediated by activating enzymes
similar to those originally described by Hoagland [17], according to the
equation :

amino acid + ATP + activating enzyme?=^enzyme-AMP~ amino acid + PP

A convenient method of assay for this activity is based on the reverse re-
action, in which pyrophosphate is incorporated into ATP. Using ^^p.
labelled pyrophosphate and measuring the production of radioactive ATP
[18] it was possible to show that extracts of isolated thymus nuclei will
form ^-ATP, but only if amino acids are present in the incubation mixture

[19];

The ATP-pyrophosphate exchange reaction is promoted by the
addition of mixtures of amino acids or by individual amino acids tested
separately. Neutral extracts of highly purified thymus nuclei contain
activating enzymes for at least fifteen L-amino acids; the D-isomers of
these amino acids are not acted upon (Table I).

Many of the activating enzymes can be precipitated from solution by
lowering the pH of the nuclear extract to 5-2, as was observed previously
in cytoplasmic systems.

Many tests have been made to verify that the amino acid activating
activity is actually present in the cell nucleus and is not an artefact due to
the adsorption of cytoplasmic enzymes during the isolation procedure. The
most reliable evidence in this connection is that obtained from a study of
nuclei isolated in non-aqueous media by techniques which preclude any
exchange of water-soluble enzymes between nucleus and cytoplasm [20, 21].
Two types of "non-aqueous" nuclei were selected for analysis because
their purity had been established by chemical, enzymic, and immunological
tests for cytoplasmic contamination ; these were the nuclei prepared from

264 VINCENT G. ALLFREY

TABLE I

Effect of Amino Acids on ATP-Pyrophosphate Exchange by Calf Thymus

Nuclear pH 5 Fraction

Amino acids promoting Amino acids without

exchange effect in this system

L-Alanine L-Arginine*

L-Aspartic acid Glycine*

L-Cysteine L-Phenylalanine*

L-Glutamic acid All D-amino acids

L-Histidine

L-Isoleucine

L-Leucine

L-Lysine

L-Methionine

L-Proline

L-Serine

L-Threonine

L-Tryptophan

L-Tyrosine

L-Valine

* Found in low concentration in 005 m KCl supernatant before precipitation
of pH 5 fraction.

lyophilized calf thymus and chicken kidney tissues [21]. It was found that
the "non-aqueous" thymus nuclei contained slightly more activating
activity (in units per mg. dry weight) than the equivalent weight of whole
tissue. In kidney nuclei the enzyme concentration (in units per mg.) is
about half that of the whole tissue. In both instances, the high purity of
the preparations make it certain that the activating enzymes are of nuclear
origin and are not due to cytoplasmic contamination.

Amino acid activating enzymes have also been detected in calf liver
nuclei following an isolation in dense sucrose solutions by the method of
Chauveau et al. [22]. These nuclei are characterized by a high degree of
morphological purity (but other tests have shown that the adsorption of
soluble cytoplasmic enzymes can occur in heavy sucrose solutions). In a
recent communication, Webster [23] has described the properties and
partial purification of an alanine-activating enzyme from pig liver nuclei.
Considering the variety of nuclear types successfully tested for activating
activity, the occurrence of amino acid activating enzymes in the cell
nucleus promises to be a widespread phenomenon.

Ultracentrifugation experiments have shown that the activating
enzymes of the thymus nucleus tend to remain associated with the nuclear
ribosome fraction. As a result, they can be sedimented from neutral nuclear
extracts by centrifuging under conditions which bring down the ribonucleo-

NUCLEAR PROTEIN SYNTHESIS 265

protein particles of the nucleus. The association of the activating enzymes
with particles which have been proven to be intranuclear (see below) is
additional evidence for the conclusion that the nucleus itself carries out
the amino acid activation reaction.

AMINO ACID TRANSFER TO NUCLEAR RIBONUCLEIC ACID

Studies of many cytoplasmic and bacterial systems have shown that
the sequel to amino acid activation is a transfer reaction in which the

1500

RNA (00671^ chloramphenicol)

Protein (0067 M chloramphenicol)
i — n

10 20 30 40 50 60
Time (min)

Fig. 2. The time course of uptake of [i-^^C]-leucine into the protein and
" carrier" RNA of isolated thymus nuclei. The specific activity of the total nuclear
protein (c.p.m./mg.) is plotted in the upper curve. The incorporation into protein
is greatly inhibited by the presence of chloramphenicol (lower curve). The uptake
of amino acid into the RNA is the same in chloramphenicol-treated nuclei as in
"controls".

activated amino acid is coupled to a soluble ribonucleic acid, according to
the equation :

enzyme — amino acyl ~ AMP + sRNA^

amino acyl — sRNA + enzyme + AMP

[24-28]. A similar process occurs in the isolated thymocyte nucleus [19, 29].
The transfer of amino acid to RNA in the nucleus can be shown directly
by isolation of radioactive "carrier" RNA after incubating nuclei in the
presence of a radioactive amino acid and chloramphenicol [19] or puromy-
cin [29, 30]. The addition of either of these antibiotics blocks amino acid
incorporation into nuclear protein, but does not interfere with the forma-
tion of the amino acyl-RNA complex [29]. The RNA can then be prepared
by the phenol method [31, 32] without risk of contamination by radio-
active protein. The results of such an experiment are summarized in Fig. 2,
in which the specific activities of nuclear protein and nuclear "phenol"

266 VINCENT G. ALLFREY

RNA are plotted against the time of incubation in the presence of [i-^^C]-
leucine. It is clear that the addition of chloramphenicol effectively stops
amino acid uptake into protein (as first pointed out by Breitman and
Webster [33]); despite the presence of the antibiotic, leucine enters the
ribonucleic acid fraction at a very high initial rate.

Similar results have recently been obtained with puromycin [29].
Yarmolinsky and de la Haba have called attention to the fact that the
structure of puromycin closely resembles that of the proposed amino
acyl-RNA complex [30]. Considering its structure, the effectiveness of
puromycin as an inhibitor of nuclear protein synthesis can be interpreted
as due to a close competition between the antibiotic and the natural amino
acyl-RNA complex. This would then be strong presumptive evidence that
most of the incorporation of amino acids into nuclear proteins proceeds
through an amino acyl-RNA intermediate.

In the experiments described above, the formation of leucyl-RNA took
place in the intact thymus nucleus. An alternative method allows the
preparation of the amino-acid-carrier RNA complex in sub-nuclear
fractions. This procedure employs the nuclear activating enzymes (the
nuclear pH 5 fraction), ATP, RNA, and the radioactive amino acid [29,

34]-

It is of interest that much of the "transfer" or "carrier" RNA in the

thymus nucleus occurs in association with the nuclear ribosome fraction

and can be centrifuged down together with the ribosomes in the spinco.

(This differs from the usual observation that in cytoplasmic extracts, the

"carrier" RNA occurs in a relatively low-molecular weight, soluble form.)

The pH 5 fraction as usually prepared from neutral extracts of thymus

nuclei contains a few per cent of RNA which readily accepts activated

amino acids. The addition of more nuclear RNA (prepared by the phenol

method) results in a higher level of amino acid transfer.

In this connection, it should be mentioned that not all nuclear RNA
fractions can function as leucine acceptors in the transfer reaction. For
example, there is good evidence that the high molecular weight RNA of
the nucleolus does not take part in this transfer process [19, 29]. (This
RNA also fails to go into solution during the usual phenol extraction pro-
cedure.) Similar tests for amino acid transfer to DNA in the nucleus have
led to negative results [19].

The transfer reaction in sub-nuclear fractions provides a convenient
method for the preparation of leucyl-RNA. That the I'^C-uptake observed
under these conditions actually represents a true incorporation of amino
acid into RNA has been shown in several ways: (i) all of the radioactivity
can be removed from the labelled pH 5 fraction by treatment with ribonu-
clease; (2) the presence of traces of ribonuclease during the reaction
effectively prevents any uptake of isotope. (This is of interest because

NUCLEAR PROTEIN SYNTHESIS 267

amino acid uptake in the intact nucleus cannot be blocked by adding RNA-
ase to the incubation medium. This result fits in with other observations
that added RNAase does not attack the ribosomal RNA as long as it remains
within the nucleus (see below).) (3) The RNA can be prepared by the
phenol isolation method and show^n to contain bound ^^C-amino acid;
(4) all the counts are removed by alkaline digestion of the RNA or by
treatment with hot 5^0 trichloroacetic acid to remove nucleic acids; the
protein residue after these treatments is not radioactive.

PROPERTIES OF THE AMINO ACYL-RNA COMPLEX

The chemical nature of the leucyl-RNA complex has been studied
using the in vitro system of labelling the RNA associated with the nuclear
pH 5 fraction. The product of the reaction (nuclear leucyl-RNA) has
many properties which show its similarity to the amino acyl-RNA com-
plexes studied in cytoplasmic systems. It is non-dialyzable and stable in
dilute acids (e.g. 2 per cent HCIO4). On the other hand, a brief exposure
to dilute alkali (0-005 '^ NaOH) removes all the radioactivity from leucyl-
RNA, and subsequent chromatography on paper separates the counts
as free P^C] -leucine.

A short ribonuclease digestion of the nuclear leucyl-RNA complex
(adding 10 /xg. of RNAase for 15 min. at 37°) separates the radioactivity
in an acid-soluble form. It has been shown that the action of ribonuclease
releases the amino acid bound to a nucleoside [29, 34]. By following the
procedure described by Zachau et al. [35], we have isolated the amino
acyl-nucleoside by ionophoresis on filter paper. At pH 3-2 the leucyl-
nucleoside diff^ers in its mobility from any of the free nucleosides in the
RNA digest and it can be located on the paper as a radioactive, u.v.-
absorbing spot. Alternatively, the leucyl-nucleoside separates nicely in
the paper chromatographic system described by Preiss et al. [36]. In both
cases treatment of the radioactive, u. v. -absorbing spot with dilute alkali
releases the radioactivity as free ^^C-leucine and leaves the nucleoside.
Both components are readily separable by chromatography or electro-
phoresis [34].

In order to get sufficient material for identification of the nucleoside, a
large-scale preparation of leucyl-RNA was carried out, using the pH 5
fraction from over 500 g. of isolated nuclei. The product was digested wath
ribonuclease, and the amino acyl-nucleoside separated by ionophoresis.
Alkaline treatment then released the nucleoside, which, in its mobility,
Rf, and ultraviolet absorption spectrum appears to be adenosine [34].

It follows that the receptor group in nuclear "carrier" RNA is a
terminal adenylic acid, as was found earlier in the s-RNA of cytoplasmic
systems [35].

268 VINCENT G. ALLFREY

AMINO ACID INCORPORATION INTO NUCLEAR RIBOSOMES

Much of the protein synthesis in the intact nucleus occurs in its
ribonucleoprotein components [5, 37]. These are readily extractable from
the nucleus in dilute salt solutions or neutral buffers. Recent fractionation
studies of such nuclear extracts has revealed a heterogeneous population of
ribonucleoprotein particles which can be separated into classes of different
chemical composition and biosynthetic activity [37].

When the different fractions are examined under the electron micro-
scope, one observes the presence of large mmibers of dense particles of
about 100 A diameter. Particles of this size are clearly discernible in
electron micrographs of sections of intact thymus nuclei, and in tissue
sections. Similar particles have also been reported in nucleoli [38, 40],
chromosomal loops [41], in the Balbiani rings of certain chromosomes [42],
and throughout the nuclear sap in other tissues [43]. Dr. J. H. Frenster
has made some attempts at further morphological characterization of these
nuclear ribonucleoprotein particles (ribosomes), but problems of resolution
make this very difficult [37].

Tracer studies have shown that the nuclear ribonucleoprotein particles,
like those of the cytoplasm, are actively engaged in protein synthesis. Their
activity within the thymus nucleus has been followed by incubating isolated
nuclei in the presence of [^^C]-labelled amino acids and subsequently isolat-
ing the various RNP fractions [37]. The results of such experiments indicate
that the different particle fractions, separated on the basis of differing
sedimentation properties, also differ greatly in their metabolic activity [37].

There is convincing evidence that amino acid uptake takes place in
ribosomal particles within the nucleus, (i) The incorporation of amino
acids requires the presence of sodium ions in the nuclear suspension. (As
will be shown later, this requirement for sodium, rather than potassium
ions, is a sure indication of nuclear localization.) (2) Preincubation of the
nuclei with deoxyribonuclease effectively inhibits the subsequent uptake of
[^^C]-labelled amino acids into the proteins of the RNP particles. This
evidence for DNA dependency, apart from establishing nuclear localization,
also reflects the need for the nuclear phosphorylating system as a source
of ATP [13].

The synthetic activity of the nuclear ribosomes can also be studied
outside of the cell nucleus because, if they are properly supplemented, free
RNP particles prepared from thymus nuclei can incorporate [^"^C-Jamino
acids in vitro [37, 44]. The process requires, in addition to the particles,
ATP and an ATP-regenerating system, the nuclear pH 5 enzymes, and
guanosine triphosphate (GTP) [37, 44]. The time course of incorporation
of [i-^'*C]-leucine into the proteins of isolated nuclear ribosomes is
depicted in Fig. 3. It is of special interest that amino acid incorporation
can take place only in particles which were prepared under isotonic con-

NUCLEAR PROTEIN SYNTHESIS 269

ditions (upper curve) ; ribonucleoprotein particles isolated in, or exposed
to hypertonic sucrose solutions are completely inert (lower curve). This
finding is in agreement with earlier experiments on intact nuclei showing
that amino acid uptake proceeds best in isotonic sucrose solutions and is
impaired when nuclei are exposed to either hypertonic or hypotonic
conditions [44].

Leucine-l-C , ,

Isotonic nuclear

nbosomes

Ribosomes after brief
hypertonic exposure

20 40 60
Incubation time (min)

Fig. 3. The time course of incorporation of [i-^*C]-leucine into isolated
ribonucleoprotein particles prepared from thymus nuclei. The incubation medium
contained (in addition to the particles), nuclear pH 5 enzymes. ATP + an ATP-
regenerating system, and GTP. The upper curve shows the specific activity of
the proteins of the ribosomes under isotonic conditions. A brief exposure to
hypertonic sucrose solutions destroys the capacity for amino acid uptake under
these conditions (lower curve).

The incorporation of amino acids by isolated nuclear ribosomes is
inhibited if ribonuclease is added to the incubation medium. This eiTect
is of interest because it contrasts with findings on intact nuclei. It has been
shown that ribonuclease does not affect amino acid uptake into the ribo-
somes of the intact isolated nucleus [37]. The results indicate that the RNA
of the ribosome fraction is immune to attack by RNAase as long as it
remains inside the nucleus. This is one of the unique properties of the
protein synthetic pathway in the nucleus.

Some characteristically nuclear aspects of amino acid
incorporation

On the whole, the study of sub-nuclear fractions has shown that the
protein synthetic mechanism in the cell nucleus is very much like that of

270 VINCENT G. ALLFREY

the cytoplasm. It involves (i) amino acid activation, (2) transfer of the
amino acid to the terminal adenylic acid of a carrier RNA, and (3) in-
corporation into the proteins of the ribosome fraction by a reaction sequence
requiring GTP. Yet there are some striking differences between nuclear
and cytoplasmic incorporation systems. The first of these is the specific
requirement of the nucleus for sodium ions.

Why is protein synthesis in the nucleus sensitive to the presence or
absence of sodium ions ? This question has been studied in suspensions of
isolated cell nuclei by tracer methods, and the answer is now clear. Sodium
ion dependency reflects the operation of a transport system for bringing
amino acids into the cell nucleus.

NUCLEAR AMINO ACID TRANSPORT

The effect of sodium ions on the overall process of amino acid in-
corporation into nuclear proteins was discovered (largely by accident) at

200

Alanine-1-C

■ ^^ K_* Cs* and Rb*

002 0-04 006 008
Salt concentration (M)

Fig. 4. The effect of adding different monovalent cations (as chlorides) on the
incorporation of [i-"C] alanine into the proteins of intact isolated thymus nuclei.
The specific activity of the nuclear protein after 60 min. incubation is plotted
against the salt concentration of the medium.

the very beginnings of our study of uptake processes in the nucleus [5,
45]. The magnitude of the effect is strikingly demonstrated in Fig. 4,
which compares the uptake of [i-^^C]-alanine in a medium containing
sodium ions with that observed in the presence of equivalent amounts of
potassium (or other monovalent cations). The stimulatory effect of added

NUCLEAR PROTEIN SYNTHESIS 27 1

sodium is particularly large in the case of alanine uptake [46] but it is easily
demonstrated for other amino acids as well. Lithium is partly effective as
a sodium substitute, but caesium, rubidium or potassium are without
effect.

A number of quaternary ammonium ions were also tested because
Lorente de No [47] had shown them to be effective sodium substitutes in
maintaining membrane potentials in isolated nerve. However, neither
hydrazinium, guanidinium, or tetraethyl ammonium ions could promote
alanine incorporation into the proteins of sodium-deficient thymus nuclei.

The reason for the sodium-dependence of the incorporation process
was sought for but could not be found in studies of the individual reactions
in the nuclear synthetic pathway. There was no specific sodium ion require-
ment for amino acid activation, or for transfer to nuclear RNA, and a 50*^0
reduction in sodium ion concentration had no effect on amino acid in-
corporation into isolated nuclear ribosomes [44]. This led us to test the
possibility that sodium ions might influence the rate at which amino acids
enter the free amino acid "pool" within the nucleus.

This approach has not only provided an explanation of the sodium ion
effect but it has also revealed the existence of a mechanism, hitherto un-
known, for amino acid transport and accumulation within the cell nucleus.

The penetration of free amino acids into the nuclear "pool" has been
measured as follows : nuclei w'ere suspended in a buffered sucrose medium
containing sodium (or potassium ions) and a small, accurately measured
amount of the radioactive amino acid. The mixture was incubated aero-
bically, with shaking. Aliquots of the suspension were removed at zero
time and at appropriate time intervals thereafter. The nuclei were centri-
fuged down, washed with sucrose, and then extracted wdth cold 2%
HCIO4. The radioactivity in each extract was measured and, after correc-
ting for the small amount of ajiiino acid adsorbed at zero time, the counts
were plotted as a function of the time of incubation. The time course of
[i-^'*C]-alanine transport at 37° is shown in Fig. 5, which compares the
radioactivity present in the nuclear "pool" after incubation in 0-04 M
NaCl with that observed at an equivalent concentration of KCl. The
transport of the amino acid into the acid-soluble "pool" is evidently
sodium-dependent, and does not occur to any appreciable extent when
equivalent amounts of potassium replace the sodium. There is good
evidence that the radioactivity in the perchloric acid extract of the isolated
nuclei occurs as free P^C] -alanine. This was checked by chromatography
of the concentrated extract on paper and subsequent elution of the
alanine spot. A definite increase in the radioactivity of the isolated alanine
is evident after only a few minutes' incubation. Other tests have shown
that, after 30 min. at io°C., the nuclei build up an alanine concentration
that may exceed the alanine concentration of the surrounding medium [46].

272 VINCENT G. ALLFREY

It remains to be seen whether some transient derivative of alanine is formed
in the course of this accumulation process.

Competition experiments using non-radioactive D-alanine and l-
alanine before adding DL-[i-^^C]-alanine have shown that only the naturally
occurring L-isomer of the amino acid competes with radioactive alanine for
transport. It follows that the D-isomer of the amino acid is not actively
accumulated by the nucleus [29, 46]. The stereo-specificity of the reaction
suggests that amino acid transport into the nucleus involves the operation
of an enzymic mechanism.

The alanine transport reaction has a well-defined pH optimum at
neutrality. This is shown in Fig. 6. A study of the temperature-dependence

10 20 30 40 50
Incubation time (min)

Fig. 5. The specific effect of sodium ions in promoting the transport of free
[i--'*C]-alanine into the amino acid "pool" of the isolated cell nucleus. The
specific activity of the acid-soluble nuclear extracts (see text) is plotted against
the time of incubation at 37°.

of the reaction reveals the optimum in the physiological temperature range
(38°-40°) (Fig. 7). A comparison of the rates of alanine transport at 10°
and 20° (0-83 and i -6 m^moles of L-alanine per 10 min. per 20 mg. nuclei)
shows that the velocity of the reaction has doubled for a 10° rise in tem-
perature. This temperature coefficient (Ojo = 2) indicates that amino acid
transport into the nucleus involves a chemical reaction and is not simply a
result of diffusion.

It has already been pointed out that even after short periods of incuba-
tion, the [^^C]-alanine concentration within the nuclei may exceed that
of the incubation medium. This transport and accumulation of an amino
acid against a concentration gradient would be expected to require
the expenditure of energy. In this connection there is some evidence

NUCLEAR PROTEIN SYNTHESIS

273

•| 100

£

^ 0-80

0'60

E 0-40

0-20

Alanine-1-C

IO°C

6 7 8

pH of nuclear suspension

Fig. 6. The pH dependence of the nuclear alanine transport reaction. The
rate of amino acid transport (in m/^moles of L-alanine per 10 min. at 10°) is plotted
against the pH of the incubation medium.

e

^ 400

z

E

0 300

u 200

T, 100

Alanine-1-C
pH 6-65

0 10 20 30 40 50 60
Temperature ( C)

Fig. 7. The temperature dependence of the nuclear alanine transport reaction.
The rate of amino acid transport (in m/Limoles of L-alanine per 10 min. at pH 6 • 75)
is plotted against the temperature of incubation.

T

274 VINCENT G. ALLFREY

which suggests that the transport of amino acids into the nucleus is an
active process requiring the participation of ATP or some similar energy
source. For example, the addition of o-oo2 M cyanide or dinitrophenol
(both of which suppress nuclear ATP synthesis) causes some inhibition of
alanine transport. Similarly, the removal of nucleotides by selective extrac-
tion with acetate buffers [8] destroys the nuclear capacity for subsequent
amino acid accumulation [46].

Other experiments have made it clear that the primary reason for the
sodium dependence of amino acid incorporation into the proteins of intact
nuclei is simply that sodium ions are needed for the transport of amino
acids to the sites of synthesis. For example, we have found that thymus
nuclei will synthesize ^*C-labelled protein even in an all-potassium medium,
provided they are first exposed to a sodium-rich medium containing the
radioactive amino acid. (Other tests have shown that isolated nuclei cannot
retain sodium ions taken up during this preincubation.) Once transport
has occurred, the continued presence of sodium ions in the medium is not
necessary for the incorporation process [46].

Thus the sodium dependence of protein synthesis in the nucleus is due
to the operation of a specific transport mechanism. This sodium ion require-
ment becomes a useful test for establishing the intranuclear localization of
protein synthetic reactions, especially when one considers that many cyto-
plasmic systems display a specific requirement for potassium ions in amino
acid transport [48] and incorporation [15].

Other synthetic processes in isolated nuclei may or may not show a
sodium ion requirement. The range of the sodium ion requirement will
not be discussed here, except to mention that a good correlation has been
found between the ionic requirements for uptake and for transport of
several isotopic precursors of ribo- and deoxyribo-nucleic acids [46].

DNA EFFECTS ON ISOLATED NUCLEAR RIBOSOMES

A useful test for the nuclear localization of a synthetic process is a
measure of its DNA dependency. The addition of DNAase to suspensions
of intact isolated nuclei leads to a cessation of nearly all synthetic activity
[4, 5, 13]. As mentioned before, this is largely due to the fact that DNA
plays a role in mediating nuclear ATP synthesis : removal of the DNA
therefore has an indirect effect on all nuclear processes that require ATP.

However, some recent results suggest that DNA may play a more
specific role in controlling the synthesis of nuclear proteins. These results
were obtained in tracer experiments with isolated nuclear ribosomes. The
incorporation of amino acids into ribosomes outside the nucleus is not
affected by deoxyribonuclease [44] because the ATP needed for amino
acid activation is supplied by an exogenous ATP-generating system. It

NUCLEAR PROTEIN SYNTHESIS

275

was surprising therefore, to find that the addition of isolated thymus DNA
to the ribosome suspension stimulates the uptake of [i -^^C]-leucine into ribo-
somal proteins. As shown in Fig. 8, the increase in specific activity of the
protein is appreciable; DNA-supplementedribosomes are 50% more active
than the corresponding "controls". The figure also shows that the
simultaneous addition of DNAase with the DNA abolishes the stimu-
latory effect [44].

There is, as yet, no evidence for the tissue- or species-specificity of
DNA in promoting uptake. DNA preparations from calf thymus, calf

xDNA lOmg/ml.

Control

X

xDNA lOmg/ml.
DNAase 05 mg/ml

PES 10 mg/ml.

I I I I ] I I

0 20 40 60

Time (mln)

Fig. 8. Effects of added DNA, DNA + deoxyribonuclease, and polyethylene
sulphonate on [i-^*C]-leucine incorporation into the proteins of isolated nuclear
ribosomes. The specific activity of the ribosomal protein (in c.p.m./mg.) is plotted
against the time of incubation at 37°.

kidney, and wheat germ are all equally efi"ective in this test system. On the
other hand, the results cannot be explained as simply a matter of adding
large negatively-charged molecules (which could bind traces of ribonu-
clease in the preparations and so enhance their activity) because the
addition of an equal amount of polyethylene sulphonate actually inhibits
amino acid uptake by isolated nuclear ribosomes (Fig. 8).

The stimulatory efiPect of DNA takes on added interest in view of
related findings that added histones inhibit amino acid uptake by nuclear
ribosomes [44] (Fig. 9). The role of histones as repressors of nuclear
synthetic activities was observed earlier in studies of amino acid uptake in
intact nuclei (Fig. 10) [49].

The stimulation of protein synthesis in the ribosomes by DNA and
the corresponding inhibition by histones suggest that the synthetic activity
of ribosomes within the nucleus will vary depending on their proximity to

276

VINCENT G. ALLFREY

'DNA lOmg/ml.
,x Control

DNA lOmg/ml
Histone lOmg/ml

X Histone 10 mg/ml.

20 40
Time (min)

Fig. 9. Effects of added histone, DNA + histone, and DNA on [i-^^C]-
leucine incorporation into the proteins of isolated nuclear ribosomes. The specific
activity of the ribosomal protein (in cp-m./mg-) is plotted against the time of
incubation at 37°.

500

400

>r 300

Serunn albumin

L-lys[ne___

U

200-

100

CONTROL
Thymus histone 11

rotamne

Thymus histone I

0 12 3 4

Milligrams added
Fig. 10. Effects of added histone I (arginine-rich histone fraction), histone II
(lysine-rich histone fraction which is less basic than histone I), protamine, and
polylysine on [i-^*C]-alanine uptake into the proteins of isolated calf thymus
nuclei. The specific activity of the nuclear protein after 60 min. incubation at 37°
is plotted against the amount of material added to the nuclear suspension.

DNA or histone at different loci of the chromosome. This may prove to be
a physiological mechanism for controlling the rates of synthesis of different
proteins in the nucleus, and one may speculate that such control is part
of the process of cell differentiation.

NUCLEAR PROTEIN SYNTHESIS 277

Summary

A study of isolated thymus nuclei and sub-nuclear fractions shows that
the pathway of protein synthesis in the nucleus involves a sequence of
reactions similar to that observed in cytoplasmic systems.

The nucleus contains amino-acid-activating enzymes which act on at
least fifteen L-amino acids ; D-amino acids are not activated.

Following activation, the amino acid is transferred to a nuclear RNA
fraction. Neither DNA nor the high molecular weight RNA of the nucleolus
takes part in this process.

Analysis of the nuclear leucyl-RNA complex shows that the receptor
group in nuclear "carrier" RNA is a terminal adenylic acid.

Ribonucleoprotein particles in the nucleus are active sites of protein
synthesis. These particles (termed nuclear ribosomes) can be isolated and
fractionated by differential ultracentrifugation of nuclear extracts, and are
capable of independent amino acid incorporation, provided they are
isolated and tested under isotonic conditions.

Amino acid uptake into the isolated nuclear ribosomes requires the
presence of ATP, amino-acid-activating enzymes, and GTP.

The addition of DNx^ promotes amino acid uptake in ribosome sus-
pensions while added histones inhibit the process. This may prove to be
one of the physiological mechanisms for the direction and control of
nuclear protein synthesis.

The protein synthetic process in isolated nuclei responds to the addition
of sodium ions because Na + is required in a specific amino acid transport
mechanism. This transport of amino acids to the sites of nuclear protein
synthesis is apparently an enzymic reaction that is specific for the L-form
of the amino acid.

References

1. Bergstrand, A., Eliasson, N. A., Hammarsten, E., Norberg, B., Reichard, P.,
and von Ubisch, H., Cold Spr. Harb. Synip. quant. Biol. 13, 22 (1948).

2. Daly, M. M., Allfrey, V. G., and Mirsky, A. E.J. gen. Physiol. 36, 173 (1952).

3. Allfrey, V. G., Daly, M. M., and Mirsky, A. E.,y.gen. Physiol. 38, 415 (i955)-

4. Allfrey, V. G., Proc. nat. Acad. Sci., Wash. 40, 881 (1954).

5. Allfrey, V. G., Mirsky, A. E., and Osawa, S., J. gen. Physiol. 40, 451 (1957).

6. Allfrey, V. G., and Mirsky, A. E., Proc. nat. Acad. Sci., Wash. 43, 821 (1957).

7. Stern, H., and Mirsky, A. E.,y. gen. Physiol. 37, 177 (1953).

8. Osawa, S., Allfrey, V. G., and Mirsky, A. E.,J. gen. Physiol. 40, 491 (1957).

9. Allfrey, V. G., and Mirsky, A. E., Proc. nat. Acad. Sci., IVash. (in press).

10. Friedkin, M., and Wood, H., J. biol. Chem. 220, 639 (1956).

1 1 . Allfrey, V. G., and Mirsky, A. E., Proc. nat. Acad. Sci., Wash. 45, 1325 (1959).

12. Breitman, T., and Webster, G. C, Nature, Lond. 184, 637 (1959).

13. Allfrey, V. G., and Mirsky, A. E., Proc. nat. Acad. Sci., Wash. 43, 589 (1957).

14. Allfrey, V. G., and Mirsky, A. E., Proc. nat. Acad. Sci., Wash. 44, 981 (1958).

278 VINCENT G. ALLFREY

15. Sachs, H.,y. biol. Chem., 233, 643 (1958).

16. AUfrey, V. G., Daly, M. M., and Mirsky, A. E.,jf. gen. Physiol. 37, 157 (1953).

17. Hoagland, M. B., Biochim. biophys. Acta 16, 288 (1955).

18. Crane, R. K., and Lipmann, ¥.,y. biol. Chem. 201, 235 (1953).

19. Hopkins, J. W., Proc. nat. Acad. Sci., Wash. 45, 1461 (1959).

20. Behrens, M., Hoppe-Seyl. Z. 209, 59 (1932).

21. Allfrey, V. G., Stern, H., Mirsky, A. E., and Saetren, H.,_7. gen. Physiol. 35,
529 (1952).

22. Chauveau, J., Moule, Y., and Rouiller, C, Exp. Cell Res. li, 317 (1956).

23. Webster, G. C, Biochem. Biophys. Res. Co?nmun. 2, 56 (i960).

24. Hoagland, M. B., Keller, E. B., and Zamecnik, P. C,J. biol. Chem. 2l8, 345
(1956).

25. Holley, R. W., J. Amer. chem. Soc. 79, 658 (1957).

26. Ogata, K., and Nohara, H., Biochim. biophys. Acta 25, 660 (1957).

27. Berg, P., and Ofengand, E. J., Proc. nat. Acad. Sci., Wash. 44, 78 (1958).

28. Weiss, S., Acs, G., and Lipmann, F., Proc. nat. Acad. Sci., Wash. 44, 189
(1958).

29. Allfrey, V. G., Hopkins, J. W., Frenster, J. H., and Mirsky, A. E., Ann. N.Y.
Acad. Sci. 88, 722 (i960).

30. Yarmolinsky, M. D., and de la Haba, G., Proc. nat. Acad. Sci., Wash. 4$^ 1721

(1959)-

31. Schramm, G., and Kerejkarto, B., Z. Naturf. 76, 589 (1952).

32. Kirby, K. S., Biochem. J. 64, 405 (1956).

33. Breitman, "P., and Webster, G. C, Biochim. biophys. Acta 27, 409 (1958).

34. Hopkins, J. W., Allfrey, V. G., and Mirsky, A. E., Biochim. biophys. Acta ^j^
194 (1961).

35. Zachau, H. G., Acs, G., and Lipmann, F., Proc. nat. Acad. Sci., Wash. 44,
885 (1958).

36. Preiss, J., Berg, P., Ofengand, E. J., Bergmann, F. H., and Dieckmann, M.,
Proc. nat. Acad. Sci., Wash. 45, 319 (1959).

37. Frenster, J. H., Allfrey, V. G., and Mirsky, A. E., Proc. nat. Acad. Sci., Wash.
46, 432 (i960).

38. Porter, K. R.,y. Histochem. Cytochem. 2, 346 (1954).

39. Bernhard, W., Bauer, A., Gropp, A., Hagenau, F., and Oberling, C, Exp.

Cell Res. 9, ^^(1955)-

40. LaFontaine, J. Cj.,y. biophys. biochem. Cytol. 4, 229 (1958).

41. Gall, J. G.,y. biophys. biochem. Cytol. 2 (Suppl.), 393 (1956).

42. Beerman, W., and Bahr, G. F., Exp. Cell Res. 6, 195 (1954).

43. Callan, H. G., in " Symposium on Fine Structure of Cells (Leiden) ", LU.B.S.
publ. B21, 89 (1956).

44. Frenster, J. H. Allfrey, V. G., and Mirsky, A. E., Biochim. biophys. Acta 47,
130 (1961).

45. Allfrey, V. G., Mirsky, A. E., and Osawa, S., Nature, Lond. 176, 1042 (1955).

46. Allfrey, V. G., Mirsky, A. E., Hopkins, J. W., and Naora, H. Proc. nat. Acad.
Sci., Wash, (in press).

47. Lorente de No, R.,^. cell. comp. Physiol. 33 (Suppl.), (1949).

48. Christensen, H. N., and Riggs, T. R.,^. biol. Chem. 194, 57 (1952).

49. Allfrey, V. G., and Mirsky, A. E.. Trans. N. Y. Acad. Sci. Ser. H, 21, 3 (1958).

NUCLEAR PROTEIN SYNTHESIS 279

Discussion

Canellakis : Have you tried adding ribonucleic acid instead of deoxyribonu-
cleic acid to your isolated ribosomes ?

Allfrey: Yes. In isolated ribosomes the addition of commercially prepared
yeast RNA failed to stimulate the uptake of [i-'*C]-leucine.

Canellakis : Was it a high molecular weight polymer ?

Allfrey: Yes. That is, it was non-dialyzable.

Arnon : When you speak of ATP synthesis by the nucleus do you imply that
the nucleus is self-sufficient with respect to its ATP requirement for protein syn-
thesis or does it still depend on ATP formation by mitochondria ?

Allfrey : I would say that for amino acid incorporation in isolated thymus
nuclei it is not necessary to add mitochondria or some other ATP-generating system.
Amino acid uptake into proteins by isolated nuclei is not as active as in whole
cells (perhaps only one-third as active) but the process can continue actively for
8-9 hr. in the absence of mitochondria. Moreover, mitochondrial contamination is
not a problem because we inhibit mitochondrial phosphorylation by adding Ca-+
ions. The nuclei are still able to carry out all of these synthetic reactions.

Arnon : What is the substrate for ATP formation ? Is oxygen being consumed ?

Allfrey: Yes, oxygen is consumed. The QCOg) usually lies between i and 2.

Arnon : But what is the chemical substrate ?

Allfrey: We don't know how the coupling of oxidation and ATP formation
takes place. We are working on that, and know that glucose is metabolized in these
nuclei.

ScHOFFENiELS : Did you find competition between the l and the d forms of the
same amino acid in the process of transfer from the solution to the nuclei ?

Allfrey: There is no competition between the l and the d forms. It is only the
L form of the non-radioactive amino acid which will compete with the radioactive
amino acid ; the d form does not.

Schoffeniels : If you add both, will not the d form be inhibitory ?

Allfrey: It has no action at all at the concentrations we have tested.

Herbert : I wonder if you have characterized the RNA, which I presume is
much like soluble RNA, further than to establish that the amino acid complexes
with adenylic acid, and whether you have characterized the activating enzymes to
any extent. It has been reported that activating enzymes isolated from nuclei are
different from the cytoplasmic enzymes with respect to certain amino acids.

Allfrey : We have tried both specificity reactions, testing the nuclear amino
acid-activating enzymes and cytoplasmic enzymes from guinea-pig liver, and
ribonucleic acids from diflferent types of nuclei. Now, the answer you get depends
on the amino acid you use. If you use leucine, both Dr. Webster and I agree there
is no sign of specificity, i.e. the nuclear enzyme will transfer to cytoplasmic RNA
or the RNA of other nuclear types, and the cytoplasmic enzyme will transfer to
nuclear RNA. But in Dr. Webster's experiments with alanine-activating enzymes
there is evidence of specificity between liver nuclear enzymes and cytoplasmic
enzymes. We haven't checked this with thymus nuclei. We have now started to
characterize the amino acid " carrier "-RNA of the thymus nucleus. We do not
call it " soluble "-RN A because in our system it occurs in association with the

28o VINCENT G. ALLFREY

nuclear ribosomes. This may be because we flood the system with Ca'^^ ions during
the course of isolating the nuclei.

Bergeron : Would it be correct to say that DNA-ase has free access by diffusion
into a compartment into which alanine does not diffuse freely but has to be moved
by active transport ?

Allfrey : I think that is right. It is not only true for DNA-ase, which has a
molecular weight of 6i coo, and for RNA-ase with a molecular weight of about
13 000, but also for some of the polyelectrolytes which we have used with molecular
weights up to 200 000. However, an alternative explanation may be offered to the
compartment hypothesis. Perhaps the apparent presence of alanine in a free "pool"
is an instance of specific adsorption rather than membrane-limited compart-
mentalization.

Bergeron : Might it not be true that you have different pools in the one system
you are dealing with, which you call the nucleus?

Allfrey : I think it's true. Certainly the effect of sodium in stimulating amino
acid entry into the "pool" differs greatly for different amino acids.

Porter : Did you imply that the radioactive sodium is concentrated intra-
nuclearly or in the perinuclear space ? This arrangement of channels bringing it to
the perinuclear space would account for the increased density in the nuclear area
in a radioautograph of a whole cell.

Allfrey : The figure shown was a radioautograph of a section of the frog
oocyte. The high grain density over the nuclear area indicates the presence of
^^Na within the nucleus. This result has been checked by direct isolation of these
nuclei. Dr. H. Naora made a set of very small chopsticks out of wood and devised
techniques for removing a single nucleus from frozen oocytes. Thus we were able
to count the nuclei separately from the cytoplasm.

Porter: It seemed to me that sodium ought really to be concentrated in the
perinuclear space and that the migration of the sodium is through the pores in the
nuclear envelope from the cytoplasm matrix.

Allfrey : Well the sodium is freely diffusible in our system of isolated thymus
nuclei.

Mitchell: Could you tell us what evidence there is that the alanine which is
accumulated is free inside the nucleus ? This is a difficult problem, and one that
many people have encountered in trying to prove that a given process was active
transport. How do you get the alanine out to show that it was free ?

Allfrey : To put it as briefly as possible, we extract with cold trichloroacetic or
preferably cold 2% perchloric acid. This leaves the proteins behind. Then we take
the perchloric acid extract which is concentrated and chromatographed. We find
a single spot corresponding in Rf to free alanine. Now this does not mean some
other intermediate wasn't formed in the process, but we have not yet found evi-
dence for any other radioactive product.

Effects of 8-Azaguanine on the Specificity of Protein
Synthesis in Bacillus cereus

H, Chantrenne

Laboratory of Biological Chemistry, Faculty of Sciences,
University of Brussels, Belgium

If 8-azaguanine is added to a growing population of Bacillus cereus,
the rate of increase of the optical density of the bacterial suspension is
reduced by about 50% [i, 2, 3]. Such an observation is usually expressed
in the following way: "azaguanine reduces the rate of growth of the
bacteria". This statement can be misleading; it depends what is meant
by "growth". A closer study of the effects of 8-azaguanine revealed the
following facts [3, 4, 5]. The analogue inhibits the increase of bacterial dry
weight by 40 to 50°,,- Protein synthesis is almost completelv suppressed,
whereas the formation of cell wall material, including the peptides it
contains, is practically untouched. The synthesis of ribonucleic acid can
be stimulated, unaffected, or slightly depressed depending on experimental
conditions, especially aeration. Xo effects are observed on the synthesis of
DNA when low concentrations (4 /xg./ml.) of azaguanine are used; a
partial inhibition slowly establishes in the presence of higher concentrations
(40 /xg./ml.) of the drug.

The formation of various essential cell constituents is thus very
differently affected by azaguanine, with the obvious result that the gross
composition of the bacteria is progressively changed. On the other hand,
the number of viable bacteria, i.e. the number of bacteria which are able
to form a colony when plated on a normal medium, ceases to increase
within 20 min. of the addition of azaguanine; it then stays constant for
about 2 hr. before dropping sharply. Therefore, the changes in chemical
composition of the bacteria which are caused by 8-azaguanine are not
immediately lethal. Recovery of the bacteria from the toxic effects of the
analogue is a slow and progressive process.

The inhibition of protein synthesis in B. cereus is fully expressed
within about 10 min. of the addition of azaguanine (36 /xg./ml.). If guano-
sine (135 /xg./ml.) is added together with azaguanine, it completely protects
the bacteria against the analogue. If added not later than 30 min. after
azaguanine, guanosine rapidly restores the synthesis of proteins. But this

282

H. CHANTRENNE

restoration becomes more and more sluggish as guanosine is added later
and later. Moreover, qualititative changes are observed among the protein
produced during restoration.

Figure i{a) shows how the synthesis of protein material — as measured
by phenylalanine incorporation — is restored when guanosine is added 30,
45 and 60 min. after azaguanine. In the same experiment, the synthesis of

750

500

250

3,000

•

^ /

•

/

2,000

/

•

/ .

1,000

0

1 . . 1 , , 1

30

45

180
(min)

(a)

360

180
(mm)

(b)

360

Fig. I. Restoration of protein and penicillinase synthesis [7]. To five growing
suspensions of B. cereus (penicillinase-constitutive mutant) containing 008 ^C
of L-['*C]-phenylalanine per ml., 36 /xg. of 8-azaguanine were added at time o.
Guanosine (135 /xg./ml.) was added respectively o, 30, 45 and 60 min. after
azaguanine. No guanosine was added to the last suspension (00).

(a) : '""C in protein material (c.p.m. per ml. suspension).

(6) : penicillinase (units of activity per ml. suspension).

constitutive penicillinase — as measured by its enzymic activity — was
determined; the results are shown in Fig. i(b). Comparison of figures i{a)
and i{b) indicates that when guanosine is added 45 or 60 min. after
azaguanine, penicillinase production is still completely obliterated at a
time when the synthesis of protein material was already restored to a
considerable extent [7].

The damage caused by 8-azaguanine to the formation of the enzyme
penicillinase is therefore more severe or more difficult to repair than the
damage caused to the synthesis of protein material as a whole. Moreover,
it becomes worse and worse as the time of action of azaguanine increases.
All the enzymes, however, do not suffer as much as penicillinase. For

EFFECTS OF 8-AZAGUANINE ON THE SPECIFICITY OF PROTEIN SYNTHESIS 283

instance, the synthesis of catalase is restored almost at the same time as
the average protein material ; for some time during restoration the rate of
catalase formation per weight of newly formed protein material is even
greater than in normal bacteria (Fig. 2).

Clearly, besides inhibiting protein synthesis in general, the purine
analogue must jam some mechanism upon which the specificity of protein
synthesis depends.

Cells or bacteria which are unable to incorporate azaguanine into
ribonucleic acids and nucleotidic compounds are insensitive to the analogue

750
^1 500

c

V

i 250

■1,500'

V 1.000

500

Fig. 2. Restoration of the synthesis of protein material, catalase and penicil-
linase [7]. Two growing suspensions of B. cereiis (penicillinase-constitutive mutant)
in a casein hydrolysate were shaken in a water bath at 30 . At time o, each sus-
pension received o*o8 ^tC L-[^^C]-phenylalanine and 36 ^ig. 8-azaguanine per ml.
Guanosine (135 /^tg./ml.) was added respectively at the same time as 8-azaguanine
(control, clear points) and 45 min. later (black points).

O, • : ^■'C in protein materia! per ml. suspension.

A, A : Penicillinase activity (units per ml. suspension).

V, ▼ : catalase activity, per ml. suspension.

[8, 9]. The inhibitory action of azaguanine therefore is probably due to
some harmful synthesis of nucleotidic compounds containing the analogue
instead of guanine. Since guanosine triphosphate is required for the passage
of the amino acids from transfer RNA to the nascent polypeptides on the
ribosomes [10, 11, 12, 13], an obvious possibility is that azaguanosine
triphosphate might interfere at this stage. The degree of inhibition of
growth is indeed more closely correlated with the concentration of azagua-
nine in acid soluble compounds than with the total amount of analogue in
the nucleic acids [i]. On the other hand, in Tetrahymena geleii which
requires exogenous uracil to make RNA, the degree of inhibition of
growth by azaguanine depends on the level of uracil in the medium [14]

284 H. CHANTRENNE

and it was observed in our laboratory (Allinckx, unpublished) that azagua-
nine inhibits amino acid incorporation into the proteins of this protozoan
only when uracil is provided. This suggests that the harmful synthesis in
which azaguanine is involved depends on uracil. One of the lesions caused
by azaguanine thus probably concerns some RNA fraction undergoing a
relatively rapid metabolism.

The possibility that 8-azaguanine might change the acceptor specificity
of soluble transfer RNA for certain amino acids was investigated recently
by Osawa (personal communication) : in soluble RNA as isolated by the
phenol method from azaguanine-inhibited B. ceretis, about 25*^0 of the
guanine molecules were replaced by the analogue. In spite of this, no
difference was observed between this abnormal soluble RNA and the
normal compound in their ability to bind P^C]-leucine or in the total
amount of ^^C bound from a mixture of radioactive amino acids in the
presence of a normal extract of bacteria.

According to these data, azaguanine does not seem to interfere with
the activation of the amino acids, at least if the classical scheme of protein
synthesis involving Hoagland's activation enzymes is valid for B. cereus.
The possibility that the "incorporation enzymes" of Beljanski and Ochoa
[15] is involved should also be investigated, for these enzymes require the
participation of the four nucleoside triphosphates [16] and azaguanine
might possibly interfere at this point.

If the activation steps are not disturbed by azaguanine, and if the
acceptor specificity of transfer RNA is not changed, then azaguanine
probably interferes with the passage of the amino acids from transfer RNA
to the ribosomes or with their condensation into polypeptides in the
genetically controlled structure. These are the most obscure steps of
protein synthesis, and azaguanine might be a useful tool for the study of
these steps. An important observation in this respect is that azaguanine,
which is incorporated into both soluble and ribosomal RNA, disturbs
the normal balance between these two groups of nucleic acids. The ratio of
sedimentable to non-sedimentable RNA decreases considerably as azagua-
nine is being incorporated (Otaka, in press). In our laboratory, current
research by J. Lahou indicates that azaguanine drastically reduces uracil
incorporation into ribosomal RNA, while non-sedimentable RNA accumu-
lates. The addition of guanosine, which restores protein synthesis, causes
an increase of sedimentable RNA, and part of this originates from non-
sedimentable RNA which was formed in the presence of azaguanine.

There is certainly a block at this stage. Unfortunately, it has not been
possible to establish yet whether the inhibition of formation of ribosomal
RNA is the cause or the consequence of the inhibition of protein synthesis.
It is indeed conceivable that ribonucleic acids which would normally be
bound into ribosomes remain in the non-sedimentable fraction for lack

EFFECTS OF 8-AZAGUANINE ON THE SPECIFICITY OF PROTEIN SYNTHESIS 285

of protein with which they could combine. But this certainly does not
completely explain what happens. When guanosine is added to intoxicated
bacteria, a reorganization of RNA takes place : both non-sedimentable and
ribosomal RNA lose the largest part of the azaguanine that they had in-
corporated, whereas all the pyrimidine residues that had been taken up
during azaguanine action are retained as polynucleotides; part of the
pyrimidines which were incorporated into soluble RNA are now found in
the ribosomal sediment. It would seem that the presence of azaguanine
in the ribonucleic acids prevents some normal interaction between soluble
and ribosomal RNA which is important for protein synthesis.

In a discussion on the structure and function of transfer RNA [17] it
was pointed out that the individual transfer ribonucleic acids must be
recognized by the activation enzymes, and that in turn the transfer RNA
must recognize the different sites on the template. In this perspective, it
seems that transfer RNA's which have incorporated a considerable amount
of azaguanine are still recognized by the activation enzymes ; on the other
hand, it is possible that they cannot recognize the template any more.

The differential effects of azaguanine on the synthesis of individual
enzymes indicates an action of the analogue on RNA fractions which carry
genetic information (messenger RNA or template RNA). To account for
the differential susceptibility of various enzymes, one can imagine that
certain templates are blocked, whereas others are not. This would result
in the production of an abnormal assortment of proteins. It is also con-
ceivable that the templates are all modified and that they produce abnormal
proteins, the abnormalities being of such a nature that certain enzymes are
more adversely affected than others. Incorporation of 2-thiouracil or of 5-
fluorouracil into RNA of Escherichia coli indeed causes the production of
proteins which differ from the normal ones in their serological properties,
enzymic activity and amino acid composition [18, 19, 20].

These possibilities are presently being investigated in our laboratory
in the case of catalase and penicillinase synthesis in B. cereus ; the work is
not advanced enough for results to be reported at present. A few data on
catalase may be quoted. The catalase which forms during restoration of
protein synthesis at a time when the production of active penicillinase is
not yet restored (see Fig. 2) was compared to the enzyme of normal
bacteria. No difference in sensitivity to azide or hydroxylamine inhibition
was detected; the concentration of inhibitor required for 50% inhibition
was the same for both the normal enzyme and the enzyme made during
restoration by guanosine. However, the relative activities of the two samples
of enzyme at different temperatures were markedly different, as shown in
Fig. 3. These activities were measured on crude bacterial extracts prepared
by sonication, and indirect effects cannot be completely excluded at present.
As the protein content of the extracts differed, an excess of serum albumin

286

H. CHANTRENNE

was added to the bacterial extracts in certain experiments; this did not
change the resuhs appreciably.

These results suggest that the protein moiety of catalase made during
the restoration period was abnormal, although the enzyme was active.
Considerable change in the protein moiety of catalase can probably be
brought about without changing much the catalytic properties of the haem;
to the contrary, the active centre of penicillinase must be a region of the
polypeptide, the catalytic properties of which are directly affected by

Fig. 3. Changes in catalase properties. Extracts were prepared by sonication
from normal bacteria (•) and from bacteria which had been incubated with azagua-
nine for 45 min. before adding guanosine. These bacteria were collected i2omin.
after the addition of guanosine (O)- The catalase activity of the extracts was
determined according to von Euler and Josephson [21] at four different tem-
peratures: o", 15", 30 and 45". In the second experiment, an excess bovine serum
albumin was added to both extracts. The results are expressed as per cent of the
activity measured at o".

malformations of the chain. This may account for the greater sensitivity
of penicillinase to the damages caused by azaguanine.

Our results suggest that 8-azaguanine, like 2-thiouracil and 5-
fluorouracil, can affect the structure of the proteins produced by the
bacteria which have incorporated the analogue into their nucleic acids. On
the other hand, azaguanine exerts a general inhibition of protein synthesis,
most probably by disturbing the normal interactions between some soluble
and ribosomal ribonucleic acids.

References

Mandel, H. G.,J. biol. Chetn. 225, 137 (1957).
Chantrenne, H., Rec. Trav. chim. Pays-Bas 77, 586 (1958).
Chantrenne, H., and Devreux, S., Exp. Cell Res. Suppl. 6, 152 (1958).
Chantrenne, H., and Devreux, S., Nature, Lond. 181, 1737 (1958).
Chantrenne, H., and Devreux, S., Biochini. biuphys. Acta 39, 486 (i960).

EFFECTS OF 8-AZAGUANINE ON THE SPECIFICITY OF PROTEIN SYNTHESIS 287

6. Chantrenne, H., Biochetn. Pharmacol, i, 233 (1959).

7. Chantrenne, H., and Devreux, S., Biochim. biophys. Acta 31, 239 (i960).

8. Matthews, R. W., in " Ciba Foundation Symposium on the Chemistry and
Biology of Purines", Churchill, London, 270 (1957).

9. Brockman, R. W., Hutchinson, D. J., and Skipper, H. E., Fed. Proc. 17, 195
(1958).

10. Hoagland, M. B., Zamecnik, P. C, and Stephenson, M. L., Biochim. biophys.
Acta 24, 215 (1957)-

11. Hoagland, M. B., Stephenson, M. L., Scott, J. F., Hecht, L. I., and Zamecnik,
P. C.,y. bid. Ghent. 231, 241 (1958).

12. Webster, G. C, Arch. Biochem. Biophys. 85, 159 (1959).

13. Lamborg, M. R., and Zamecnik, P. C, Biochim. biophys. Acta 42, 206 (i960).

14. Heinrich, M. R., Dewey, V. C, Parks, R. E., and Kidder, G. W.J. biol. Chem.

I97> 199 (1952).

15. Beljanski, M., and Ochoa, S., Proc. tiat. Acad. Sci., Wash. 44, 494, 1 157 (1958).

16. Beljanski, M., Biochim. biophys. Acta 41, 104 (i960).

17. Zamecnik, P. C. (communicated to the Symposium).

18. Hamers, R., and Hamers-Casterman, C, Biochim. biophys. Acta 33, 269 (1959).

19. Horowitz, J., and Chargaff, E., Nature, Loud. 184, 1213 (1959).

20. Gros, F., in " Dynamik des Eiweisses, 10. Colloquium Gesellschaft fiir
Physiologischen Chemie ". Springer, Berlin, 82 (i960).

21. Von Euler, H., and Josephson, K., Ann. 455, i (1927).

Discussion

Campbell : If I understand it right the incorporation of your azaguanine into
the S-RNA makes your system rather different from that of Dr. F. Gros of the
Institut Pasteur in which there was very little incorporation of fluorouracil into his
S-RNA ?

Chantrenne: These were very short experiments, a matter of 10 or 20 sec, if
I remember correctly.

Campbell: But later on the label moved to 70S and not S-RNA ?

Chantrenne: Yes. In the case of azaguanine, there are two effects. The first is
an inhibition of protein synthesis ; it might be due to an action of the analogue
on GTP or on S-RNA which prevents this from interacting normally with the
ribosomes. The second effect bears on the specificity of protein synthesis, and in
this respect it resembles very much that observed by Gros with fluorouracil; it
might concern the genetic messenger or the ribosomes.

Chargaff: Fluorouracil goes very largely into the S-RNA; it actually replaces
more uracil of the S-RNA than it does of the ribosomes.

Herbert: In Dr. Gros's experiments fluorouracil inhibited DNA synthesis
and he had to add thymine to restore it.

Chantrenne: Yes, that is right.

Herbert : Does azaguanine affect DNA synthesis ?

Chantrenne: With low concentrations of azaguanine, DNA synthesis is not
inhibited. To get an inhibition, you must use rather large concentrations and even
so the inhibition establishes only after about one or one and a half hours. The
incorporation of azaguanine into DNA is very low, but it might possibly be sufficient
to cause changes in specificity.

Purine and Pyrimidine Analogues and the Mucopeptide
Biosynthesis in Staphylococci

H. J, Rogers and H. R. Perkins

National Institute for Medical Research,
Mill Hill, London, England

Both Gram positive and Gram negative micro-organisms are sur-
rounded by a rigid layer or cell wall. The outer surface of this material is
distinguished with some difficulty from capsular material; its inner
surface is applied closely to the "membrane" which, among many other
functions, controls the passage of smaller metabolites into and out of the
cell. If this rigid outer layer of insoluble material, the so-called cell wall,
be removed, the remainder of the cell will only remain whole, as a spherical
protoplast, providing high concentrations of material which cannot readily
penetrate the "membrane" are present in the external medium. If this
condition is not met the cell bursts because the internal concentration of
solutes in the cell produces an osmotic pressure equivalent to about twenty
atmospheres for cocci and about ten atmospheres for some rod forms, too
great a force for the weak membrane, unsupported by the cell wall, to
withstand. A large proportion of this insoluble cell wall material, in Gram
positive micro-organisms, consists of a limited number of amino acids,
usually two amino sugars and in some species a few hexoses. Such com-
plexes, we have suggested, should be called mucopeptides [i]. In the
staphylococci, these mucopeptides contain only alanine, lysine, glycine
and glutamic acid together with glucosamine and muramic acid (3-O-
carboxyethylglucosamine) ; most of each of the amino sugars exists as the
N-acetyl derivative ; some of the alanine and probably all the glutamic acid
are in the d configuration [2, 3]. Apart from mucopeptide the wall material
of staphylococci also contains 40-60",, of a teichoic acid [4, 5] consisting
of a polymer of ribitolphosphate with N-acetylglucosamine and D-alanine
attached. In some strains it has been suggested that this material may
account for up to 70-80*^ ,j of the weight of the isolated wall material [6].

The biosynthesis of mucopeptide is inhibited by penicillin, bacitracin
and oxamycin [7, 8, 9] but not by chloramphenicol [10, 8]. It seems prob-
able that the inhibition of this process represents a primary site of action
of the former group of antibiotics. During inhibition caused by penicillin

290 H. J. ROGERS AND H. R. PERKINS

a small nucleotide-linked mucopeptide accumulates which consists of
uridine diphosphate linked to N-acetylmuramic acid, alanine, glutamic
acid and lysine in the molar ratios of 1:1:3:1:1 [11]. Fragments of the
mucopeptide, namely uridine diphosphate-N-acetylmuramic acid and
uridine diphosphate-N-acetylmuramic acid linked to one molecule of
alanine, also occur. On the basis of the analytical similarity of the first of
these mucopeptides (excluding the nucleotide) with the cell wall muco-
peptide of their strain of staphylococcus, Strominger and Park [12]
suggested that the nucleotide-mucopeptide was a precursor of cell wall.

The cell wall mucopeptide from our strain of organism [8, 13] prepared
by the Cummins and Harris [14] method differed in quantitative composi-
tion from that examined by Park and Strominger [12] which had been
prepared by the Salton and Home [14a] method. Such differences promp-
ted further investigation of the relationship between uridine nucleotide
linked compounds and cell wall mucopeptide synthesis. Also the com-
pounds isolated by Park [11] contained neither N-acetylglucosamine nor
glycine both of which are prominent components of the cell wall material
in all strains of staphylococci so far examined.

A possible general way of examining the role of uridine compounds in
biosynthesis appeared to be to supply the cell with a pyrimidine analogue
which behaved sufficiently like the natural compound to be incorporated
into the small molecular weight compounds without these being able to
polymerize to form the macromolecules. The uracil analogue which
appeared most suitable for such a study was 5-fluorouracil. It had been
shown by Heidelberger and his colleagues [15, 16] that this uracil analogue
was converted by mouse tissues to the mono- di- and triphosphate nucleo-
tides and probably to fluorouridine-diphosphoglucose. Also on structural
grounds fluorouracil might be expected to behave more like uracil than
other analogues such as 5-bromo- and 5-iodouracil or 6-azauracil.

The effect of 5-fluorouracil upon the synthesis of mucopeptides was
therefore studied and its action compared with that of other purine and
pyrimidine analogues. A full account of most of this work has already
appeared [20].

The effect of the compounds upon the biosynthesis of cell wall muco-
peptide was examined by measuring the rate of incorporation of [i-^^C]-
alanine, glutamic acid and glycine or [G-^*C]-lysine into washed cells of
Staphylococcus aureus strain 524/SC incubated in the presence of chloram-
phenicol under the conditions of Mandelstam and Rogers [8]. After incuba-
tion, the measurements of biosynthesis were made using material, prepared
from the cells, which was insoluble in hot trichloroacetic acid. It was
shown that incorporation into such material reflected cell wall mucopep-
tide synthesis by preparing the cell walls according to the method of
Cummins and Harris [14]. Of the following compounds only 5-fluorouraciI

PURINE AND PYRIMIDINE ANALOGUES 29 1

inhibited cell wall mucopeptide synthesis: 5-bromouracil, 5-fluorouracil,
6-azauracil, 6-azathymine, 8-azaguanine and its riboside, 2 : 6-diamino-
purine, and 6-mercaptopurine. At a concentration of o-i jumole/ml. of
5-fluorouracil the biosynthesis of cell wall mucopeptide was inhibited
30-60%. The other compounds were tested in concentrations up to i
/xmole/ml. without effect upon biosynthesis. A separate section will be
devoted to the examination of the effect of certain substituted benzimida-
zoles which have also been examined.

The accumulation within the cells of compounds soluble in trichloroace-
tic acid and which contain N-acetylhexosamine in a bound form, was
examined by the technique previously used by Strominger [18] for penicil-
lin-treated cells. It was found that a considerable accumulation of such
compounds occurred. For example, in the presence of 0-4 jumole of 5-
fluorouracil/ml. the cells accumulated 600 /xg. of N-acetylhexosamine/ 100
mg. of bacterial dry wt. during i hr. incubation, whilst the control cells
incubated without 5-fluorouracil had accumulated only 50 60 /xg. of bound
N-acetylhexosamine. The amino sugar concentration is expressed in terms
of N-acetylglucosamine. The amount present in the cells increased accor-
ding to the concentration of fluorouracil added to the incubation medium
up to the highest concentration of inhibitor examined (o •4/zmole/ml.). This
accumulation could be prevented equally by the simultaneous presence
of either uridine or uracil; thymidine was about a third as effective as
these two compounds on a molar basis.

The nature of compounds containing N-acetylhexosamine

Washed cells of Staphylococcus aureus 524/SC were incubated at 35°
in the solution used previously which contained the four cell wall amino
acids, glucose, phosphate, and chloramphenicol. A concentration of 0-4
/imole/ml. of [2-^^C]-5-fluorouracil was included. Incubation was con-
tinued for I hr. at 35° with aeration. At the end of this time the cells
were removed by centrifugation and washed with cold o • i m sodium-
potassium phosphate buffer at pH 7-0. They were then extracted with
cold 10% trichloroacetic acid [18] as before. The extract was freed from
trichloroacetic acid by ether extraction, the residual ether removed by
aeration and the extract then chromatographed on a column of Dowex-i
(X2, chloride form, 50-100 mesh). The eluents used were those described
by Strominger [18]. The emerging samples were examined for bound
N-acetylhexosamine by the usual technique [18] and for radioactivity. Two
peaks containing both N-acetylhexosamine and [2-^^C]-5-fluorouracil were
found corresponding in position to the bound N-acetylhexosamine peaks
obtained when extracts from penicillin-treated cells were examined in a
similar manner. Both these peaks appeared likely to contain more than

292 H. J. ROGERS AND H. R. PERKINS

one substance. Consequently the materials from each peak after concentra-
tion by adsorption on to, and elution from charcoal [19], were re-chromato-
graphed on columns of diethylaminoethyl-cellulose. The columns were
developed with a gradient of ammonium acetate at pH 6-5. Material from
the first peak from Dowex-i yielded only one substance containing both
N-acetylhexosamine and [2-^^C]-5-fluorouracil when examined on the sub-
stituted cellulose column whilst the second yielded two such peaks, one
of which was still heterogeneous. This latter peak was chromatographed
once more on diethylaminoethyl cellulose, the column being developed
this time with an increasing concentration of ammonium acetate at pH 5-0.
Table I shows the analysis of the compounds isolated. It will be seen
that I A, 2C1 and 2C2 are the fluorouridine analogues of the compounds
previously isolated by Park [11] from penicillin-treated staphylococci.
Huorouridine-diphospho-N-acetylglucosamine (2B) was also isolated: no
other compounds containing N-acetylhexosamine could be recognized.

TABLE I

Analysis of the Compounds Extracted from Staphylococcus aureus Strain 524

WITH Trichloroacetic Acid after the Cells had been Incubated with 04

/amole of [2-^'*C]-5-Fluorouracil

(The molar proportions are related to the concentration of [2-^*C]-5-fluoro-
uracil which is taken as i 00.)

Molar proportions

■^-Fluoro- Phos- Glutamic ^ • ai • Muramic Glucosa-

•' ., , . , Lysme Alanine . ,

uracil phorus acid • acid mine

I A I • 00

2B I 00

2C1 I 00

2C0 I 00

099

1-05

283

I ■ 12

0

0

0

0

0

0-97

0

0

0

I 0

0

0

0

1-3

I -o

0

* = Not estimated.

It will be noted that none of the compounds contained glycine. In
earlier experiments in which lower concentrations (o-i /xmole/ml.) of 5-
fluorouracil had been used, traces of glycine were found in hydrolysates
from the materials in both peaks which emerged from the Dowex-i
columns [17]. Only very small amounts of a definite compound, however,
were found containing glycine, glutamic acid, alanine, and muramic acid.
This compound seemed no longer to be present when 5-fluorouracil was
present at the higher concentration in the incubation medium.

From the above observations it seems reasonable to suppose that at
least part of the cell wall mucopeptide in our strains of staphylococcus is

PURINE AND PYRIMIDINE ANALOGUES 293

synthesized via uridine diphospho-N-acetylglucosamine and a uridine
diphospho-mucopeptide similar to that isolated by Park [ii]. Equally it
seemed possible that glycine which did not accumulate in any form associa-
ted with 5-fluorouridine might be incorporated into the wall mucopeptide
via a co-enzyme other than uridine. If such a precursor also contained
other amino acids besides glycine, the differences in molecular proportion
between our cell wall mucopeptides and those of other strains [13] might
be explained. Earlier work, already described, showed that some analogues
of thymine, guanine, and adenine did not inhibit mucopeptide synthesis or
lead to the accvmiulation of compounds containing bound N-acetylhexosa-
mine which were soluble in trichloroacetic acid.

TABLE II

The Effect of Substituted Benzimidazoles on the Incorporation of [i-"C]-
Glycine and [G-^*C] Lysine into Material Insoluble in Hot Trichloroacetic

Acid

The organisms were incubated under conditions [8, 10] for cell wall synthesis
such that little or no protein synthesis took place.

Compound

Concentration

"0 Inhibition of in<

;orporation

(/tmoles/ml.)

[ I -i*C] -glycine

[G-

-i«C]-lysine

5 : 6 dimethyl-

0-4

12

24

benzimidazole

I 0

15

32

5 : 6 dichloro-1-

(jS-ribofuranosyl)-

05

24

32

benzimidazole

I 0

62

63

5 : 6 dichloro-

0-5

12

15

benzimidazole

I 0

21

31

2 -o

64

59

Another group of analogues, the substituted benzimidazoles, have been
shown to inhibit [8-^^C]-adenosine incorporation into ribonucleic acid
of chorioallantoic membrane [21, 22] and to inhibit incorporation of
amino acids into trichloroacetic acid insoluble material from staphy-
lococci [23, 24, 25]. The effect of these substances upon cell wall
formation under our conditions was examined. On the hypothesis that
the small amount of material referred to above as having been isolated
from cells treated with o-i /^mole/ml. of fluorouracil, which contained
glutamic acid, glycine, and alanine but not lysine, might be a cell-wall
precursor, the effects of the substituted benzimidazoles on the incorpora-
tion of glycine and lysine were compared. It was found (see Table II) that
two of the compounds inhibited incorporation of both amino acids when
tested in the same system as used before, but that both amino acids were

294 H. J. ROGERS AND H. R. PERKINS

equally affected. The third, 5, 6-dimethylbenzimidazole, was only weakly
inhibitory but rather more active in inhibiting lysine incorporation than
that of glycine. No definite accumulation of glycine-containing compounds
was found. Full analysis of the mechanism of the inhibition has not yet
been attempted but it does not seem to be possible to inhibit differentially
the two possible pathways of synthesis in an obvious manner by means of
the substituted benzimidazoles. It may be, of course, that even if two such
pathways exist, it is impossible to obtain a polymer principally made from
only one set of precursors. If this were so, then one would not expect to
be able to inhibit the incorporation of amino acids differentially, although
earlier work with penicillin [8] held out promise of such possibilities.

References

1. Perkins, H. R. and Rogers, H. J., Biochem.J. 72, 647 (1959).

2. Strominger, J. L., and Threnn, R. H., Biochim. biophys. Acta 33, 280 (1959).

3. Salton, M. R. J., Nature, Loud. 180, 338 (1957).

4. Armstrong, J. J., Baddiley, J., Buchanan, J. G., Carss, B., and Greenberg,
G. R.,y. chem. Soc. 4344 (1958).

5. Baddiley, J., Proc. chem. Soc. 177 (1959).

6. Hancock, R., Biochim. biophys. Acta 37, 42 (i960).

7. Park, J. T., Biochem.J. 70, 2P (1958).

8. Mandelstam, J., and Rogers, H. J., Biochem.J. 72, 654 (1959).

9. Strominger, J. L., Ito, E., and Threnn, R. H., J. Amer. chefn. Soc. 82, 998
(i960).

10. Mandelstam, J., and Rogers, H. J., Nature, Loud. l8l 956 (1958).

11. Park, J. T.,J. biol. Chem. 194, 877, 885, 897 (1952).

12. Strominger, J. L., and Park, J. T. Science 125, 99 (i957)-

13. Rogers, H. J., and Perkins, H. R., Nature, Loud. 184, 520 (i959)-

14. Cummins, C. S., and Harris, H., J', gen. Microbiol. 14, 583 (1956).

14a. Salton, M. R. J., and Home, R. W., Biochim. biophys. Acta 7, 177 (1951).

15. Chaudhuri, N. K., Montag, B. J., and Heidelberger C, Cancer Res. 18, 318
(1956).

16. Harbers, E., Chaudhuri, N. K., and Heidelberger C.,J. biol. Chem. 234, 1255

(1959)-

17. Rogers, H. J., and Perkins H. R., Biochem.J. 74, 6P (i960).

18. Strominger, J. L,.,J. biol. Chem. 224, 509 (1957).

19. Cabib, E., Leloir, L. P., and Cardini, C. E.,J. biol. Chem. 203, 1055 (i9S5)-

20. Rogers, H. J., and Perkins H. R., Biochem.J. 77, 449 (i960).

21. Tamm, I., Science 126, 1235 (1957).

22. Tamm, I. "8th Symposium of the Society for General Microbiology",
ed. S. T. Cowan and E. Rowatt. Cambridge University Press, 178 (1958).

23. Gale, E. F., and Folkes, J. P., Biochem.J. 64, 4P (1956).

24. Gale, E. F., and Folkes, J. P., Biochejn.J. 67, 507 (1957).

25. Gale, E. F., "8th Symposium of the Society for General Microbiology", ed.
S. T. Cowan and E. Rowatt. Cambridge University Press, 212 (1958).

PURINE AND PYRIMIDINE ANALOGUES 295

Discussion

Reichard : Do you believe that the fluoronucleotides are less effective as donors
of the peptides, and connected with this question is the second question didn't you
find any uracil in your fluorouracil nucleotides ?

Rogers : In answer to the first question we have taken the cells in which we have
accumulated the pyrimidine mucopeptide and have tried in all ways to obtain
utilization of them. We get no utilization of the small mucopeptide even when the
cells are incubated in the presence of a large excess of normal uridine. Our hypo-
thesis is that the fluorine-substituted compounds inhibit the polymer formation by
competitive inhibition of the final steps of condensation. We haven't any very good
evidence, but one can specifically prevent the accumulation of these materials by
adding uridine simultaneously with purines. The second question concerns uracil
again ; in preliminary experiments we have used half as much fluorouracil as those
I have shown today and under these conditions we got approximately equal amounts
of the uracil compound and the fluorouracil compound accumulating. When we
doubled the amount of fluorouracil, the normal uracil compounds completely
disappeared.

Studies on the Incorporation of Arginine into Acceptor
RNA of Escherichia coli*

H. G. BOMAN, I. A. BOMANf

The Rockefeller Institute, New York, U.S.A.

and

W. K. Maas|

New York University College of Medicine,
New York, U.S.A.

Introduction

The activation of amino acids and their transfer to acceptor RNA
(soluble RNA, sRNA) has been extensively investigated in several labora-
tories (for a review see Zamecnik [i]). Although it is generally assumed that
these reactions are involved in the biosynthesis of proteins, relatively
little experimental evidence from whole cell studies is available on the
exact nature of their participation [2]. To characterize further the role of
these reactions in cellular metabolism we began a study of the incorpora-
tion of arginine into acceptor RNA. It has been shown that in E. colt
the intracellular concentration of this amino acid controls the formation
of enzymes involved in its own biosynthesis [3]. When the concentration
of arginine is high, such as in cultures growing in the presence of exo-
genously provided arginine, the formation of these enzymes is repressed ;
when it is low, as in an arginine-requiring mutant growing with limiting
arginine in a chemostat, the level of enzymes rises to 500 times that of the
repressed culture. In view of recent studies which have implicated RNA
in the regulation of protein synthesis [4, 5] we felt that the arginine system,
in which it is possible to vary the rate of formation of certain specific
enzymes, would be suitable to elucidate a possible regulatory function of
RNA.

The present paper is chiefly a description of a system for the in vttro
incorporation of arginine into the acceptor RNA. Certain compounds

* This work was supported by grant no RG-6048 from the U.S. Public Health
Service.

f Present address : Institute of Biochemistry, Uppsala, Szceden.
% Senior Research Fellow, U.S. Public Health Service.

29S H. G. BOMAN, I. A. BOMAN

Structurally related to arginine have been investigated and their effect
on the rate of formation of arginine-RNA will be described. In addition
cells obtained from different states of repression will be compared in
regard to the arginine-activating enzyme and the arginine-RNA. In
particular this comparison involves a mutant in which arginine no longer
represses enzyme formation.

Methods and Materials

The strains used were E. coli Hfr 30S0 (" wild type ") and Hfr 30S0A5,
a canavanine-resistant mutant [6] in which arginine can no longer repress
the synthesis of some enzymes involved in its own biosynthesis (in the
following this mutant is designated arg.R").

Growth of the cells was carried out aerobically in two types of minimal
media, A and S-2, [7, 8] both supplemented with monosodium glutamate
(1-67 g./l.) and glucose (5 g./l.). The cells were harvested near the end
of the logarithmic growth phase, collected by centrifugation in a Sharpies
centrifuge, and stored as a frozen paste.

The preparation of the acceptor RNAcan be summarized as follows [9] :
The procedure was designed to avoid large volumes and ultracentrifuga-
tions and is essentially a combination of steps taken from the methods of
Crestfield ei ah [10] and Kirby [11] to which a lanthanum precipitation
[12] has been added. It consists of a heating step which breaks up the cells,
a deproteinization with phenol, an alcohol precipitation to remove phenol
and low molecular weight material, an extraction of the precipitate with
I M sodium chloride to separate the acceptor RNA from the high molecular
weight RNA, and a lanthanum precipitation to separate the acceptor RNA
from polysaccharides.

The assay of the arginine-RNA* was carried out in a volume of o • 5
ml. at pH 7-0. The concentration of the components were o-i m tris-
maleate buffer, 0-002 M ATP, 0-004 ^ Mg + + (added as MgO), 0-002 m
glutathione, 0-002 M fructose diphosphate, 0-28 mM L-[^^C]-arginine
with 5-93 X 10*' c.p.m./|umole, 15-20 jtxg. arginine-activating enzyme in
0-5 ml., and 5-25 optical units of RNA in 0-5 ml. The reaction mixture
was incubated at 37° for 15 min. Mixing of reagents and termination of the
reaction by the addition of 2 ml. of o-oi m lanthanum nitrate in 0-5 M
perchloric acid were done in an ice bath. After the reaction was stopped,
o ■ 2 ml. of 5% commercial yeast RNA was added as carrier. The precipitate
was washed twice with 2 ml. of cold 5",, trichloroacetic acid containing

* The following abbreviations are used: RNA, ribonucleic acid; tris, tris-
(hydroxymethyl)aminomethane ; ATP, adenosine triphosphate; c.p.m., counts
per min. ; optical unit (o.u.), the amount of RNA which in i ml. gives an extinction
of I at 260 m/x; U.S. P., United States Pharmacopoeia; DEAE, diethylaminoethyl.

STUDIES ON THE INCORPORATION OF ARGININE 299

1% casein hydrolysate, and once with 2 ml. of cold ethanol-ether (2 :i).
It was dissolved by adding o • 3 ml. of o ■ 2 m triethylamine and heating to
about 60° for 2 min. transferred to planchets, dried and counted with a
windowless gasflow counter. A sample in which RNA was omitted was used
as a blank except if otherwise stated. Correction for self absorption and
geometry was read from a standard curve.

The large scale labelling of RNA with [^"^CJ-arginine was carried out
with the same reaction mixture as that used for the assay. The reaction
was terminated by addition of an equal volume of 88% phenol in water
and shaking for 2 min. After cooling to 4°, the phases were separated by
centrifugation. The water layer (at the top) was removed and alcohol was
added to it to a final concentration of 20% (to increase the solubility of the
remaining phenol). The arginine-RNA was then precipitated with lan-
thanum nitrate [8], redissolved in a small volume of 0-2 M potassium
ethylene diamine tetraacetic acid, pH 7, and finally dialysed for 15 20 hr.
with several changes of water.

The protein and the nucleic acid contents were determined from the
extinctions at 280 and 260 m/x [13].

Chemicals used: Nitroarginine was synthesized according to Kossel
and Kennaway [14]. The melting-point was 248^ d. and it was found to be
homogeneous by paper electrophoresis at pH 6 and by chromatography in
isopropanol : ammonia : water (7:1 : 2) (R/= o • 56). The strong u.v. absorp-
tion of nitroarginine makes it convenient to observe it on paper by quench-
ing of fluorescence in the same way as nucleotides. Sulphaguanidine
(U.S. P.) was purchased from Amend Drug and Chem. Co., New York,
New York, L-canavanine sulphate from Nutritional Biochemical Corp.,
Cleveland 28, Ohio; streptomycin sulphate (U.S. P.) from Ely Lilly and
Co., Indianapolis, Indiana; L-[^'*C]-arginine with 5*93 x 10^ c.p.m./m-
jumole from Nuclear Chicago Corp., Des Plaines, Illinois; L-[^'*C]-arginine
with 88-8 X 10^ c.p.m./m/Limole and [^*C]-alga protein hydrolysate with
1-4 X 10'' c.p.m./|U,g. were obtained from Volks Radiochemical Co.,
Chicago 40, Illinois.

Results

PARTIAL PURIFICATION OF THE ARGININE-ACTIVATING ENZYME

The frozen bacterial paste (about 11 g.) was placed in a temperature-
insulated mortar. Liquid nitrogen was poured into it until the paste was
well covered. As temperature equilibrium was approached the paste
became brittle and was then ground to a fine powder by gently tapping
with a pestle for 5-10 min. The liquid nitrogen was allowed to evaporate
and the frozen powder was spread over the bottom surface of a plastic
container. It was melted by dipping the container into water at 30° for
2-3 min., and cooled to 0°. It was suspended with stirring in about 4 ml.

300 H. G. BOMAN, I. A. BOMAN

of 0-02 M MgClo in 0-02 M tris-HCl, pH 7-3 until a viscous but homo-
geneous mixture was obtained. This was centrifuged at 100 000 g for i hr.
The pellet (about one-third of the volume) was discarded and the clear
supernatant was dialyzed at 4° for 12-15 hr. against 0-04 m tris-HCl,
pH 7-4. The resulting solution (Fraction I) contains about 9-9 mg. of
protein and 0-9 mg. of nucleic acid per ml. It was used as a general
source of all activating enzymes as well as for the further purification of
the arginine-activating enzyme.

For this further purification the extract was chromatographed on DEAE
cellulose [15]. A column containing a 100 ml. of absorbent was equilibrated
with 0-04 M tris-HCl, pH 7-4, and was then loaded with a volume of
Fraction I corresponding to 30 mg. of protein. The chromatogram was
developed by stepwise elution, using the following amounts of tris-HCl
bufl^ers, pH 7-4: («) 65 ml. of 0-23 m; (b) 45 ml. of 0-38 M, and (r) 100 ml.
of I M. Step (a) elutes a protein fraction with none or very little arginine-
activating enzyme ; step (b) contains most of this activity ; step (c) elutes a
fraction consisting mainly of nucleic acid. It was observed that the first use
of a batch of DEAE cellulose caused a considerable loss of enzyme activity.
Repeated use of the same column gave, however, satisfactory results.

The fractions containing the arginine-activating enzyme were pooled
and concentrated about twenty times using negative pressure dialysis with
I cm. wide dialysis tubing. The concentrated solution (and also Fraction I)
were stored at — 15° with 30-40% ethylene glycol as antifreeze because it
was found that repeated freezing and thawing results in marked loss of
activity. A typical concentrated solution of the arginine-activating enzyme,
including the ethylene glycol, had a protein content of 2 -9 mg. per ml. and
a nucleic acid content of 002 mg. per ml.

LEVEL OF ACTIVATING ENZYMES UNDER DIFFERENT CONDITIONS OF GROWTH

As mentioned previously the addition of arginine to the growth medium
represses the formation of enzymes in its own biosynthesis. It was therefore
of interest to examine the effect of arginine on the formation of its activating
enzyme. For comparison a similar test was carried out for leucine, although
a corresponding repression by leucine has not yet been investigated.

Two cultures of strain 3oSq were grown, one in medium A + l-
arginine (200 /ig./ml.), the other in medium A + L-leucine (200 yug./ml.).
In the latter, L-isoleucine (100 /xg./ml.) and L-valine (100 /xg./ml.) were
added to counteract the slight inhibition of growth produced by L-leucine.

Fraction I was prepared from the harvested cells as described in the
preceding section and designated I^ for the arginine-grown and I^ for the
leucine-grown cells. The levels of the activating enzymes were determined
by following the formation of arginine- and leucine-RNA. In all tests the

STUDIES ON THE INCORPORATION OF ARGININE

301

same preparation of acceptor RNA was used. The initial reaction volume
was I -5 ml., and included 100 fx\. of Fraction I. Aliquots of 0-5 ml. were
removed at different times and assayed as described under Methods. The
results are shown in Fig. i where the formation of arginine-RNA is rep-
resented by circles (open for Fraction I^ and filled for Fraction Ij) and
the formation of leucine-RNA is denoted by triangles (open for Fraction 1^
and filled for Fraction I^). The figure includes an experiment with a
twenty-times diluted sample of Fraction I^ (open squares) to show that
this reaction like that with leucine procedes linearly with time. As Fig. i

II

300

n

1 1 ■ 1 —

/^^^'"^^"^""^'^'^

y^ ^xy^\

/ / / y'

// ^^ ^

too

/ / / ^

1 / /^

/ /

// / ^

// / /

// / '^

11 / /

I ^/^ ^^^^

1 ^ _^— — -^ — '

iZ-'-'^ — ' ^

30

f1
I I

Time in minufes

Fig. I. Rate of formation of arginine-RNA (filled and unfilled circles) and
leucine-RNA (filled and unfilled triangles). The enzyme sources were Fraction I^^
obtained from cells grown in the presence of arginine (open circles and triangles),
and Fraction I^ from cells grown in the presence of leucine (filled circles and
triangles). The open squares represent the formation of arginine-RNA with a
twenty-times diluted sample of Fraction I.^.

shows, the addition of either amino acid to the growth medium is without
effect on the formation of either activating enzyme.

EFFECT OF GUANIDINO DERIVATIVES ON THE ARGININE-ACTIV.'VTING ENZYME

Certain guanidino derivatives were tested for their effect on the rate
of formation of arginine-RNA. Purified arginine-activating enzyme (about
3 /xg. per assay) was used in all experiments with arginine, and Fraction I
was used in the experiments with the amino acid mixture (the alga protein
hydrolysate). Table I shows that canavanine and streptomycin were found
to be inhibitory whereas nitroarginine was slightly stimulatory and sul-
phaguanidine without a definite effect.

302 H. G. BOMAN, I. A. BOMAN

TABLE I

Effect of Guanidino Derivatives on the Arginine-Activating Enzyme

Incorporation of

Addition

Activity

c.p.m.

per cent

All amino acids
Arginine

Canavanine
Canavanine

6140
5090
2830
1320

100

83
100

47

All amino acids

Arginine
>>

Streptomycin
Streptomycin

1550
mo
3680
2270

100
72

100
62

Arginine

Sulphaguanidine
Nitroarginine

3390
3480
3860

100
103
114

The concentration of the compounds listed under "Addition" were 3 mM, and
the [^*C]-arginine concentration was o • 28 mM in all experiments. The concentration
of the [^*C]-alga protein hydrolysate was 3-4 /Ltg./ml. in the experiment with
streptomycin and 8-4 /xg./ml. in the one with canavanine. Except for the use of an
excess of RNA from 3oSq, the assay was carried out as described under Materials
and Methods. The activity in c.p.m. is given without correction for self-absorption
and geometry.

fOO

\

so

\\

\ \

\ \

\ \

foO

\ \

\ \

\\

\ ^v

^0

\ \^

s \.

N. N.

\ \.

v^ \

■^ \^

^0

1 1 1 1 1

2 < 6 e

mM Canavanine

Fig. 2. Activity of the arginine-activating enzyme as a function of canavanine
concentration. The concentration of [^^C]-arginine was o -28 mM. Enzyme purified
from strain 30S0 represented by open circles; from strain 30S0A5 by filled circles.

STUDIES ON THE INCORPORATION OF ARGININE

303

Since the arg.R~ strain 30SQA5 was originally isolated as a canavanine-
resistant mutant [5], a comparison was made of the canavanine inhibition
of the arginine-activating enzymes isolated from 308^ and from 30S0A5.
About 3 • 3 fjig. of purified activating enzyme was used in each assay and
the canavanine concentration was varied over a twentyfold range. The
results from these experiments (see Fig. 2) show no difference in the
sensitivity between the enzymes from the two strains.

CHARACTERIZATION OF THE ARGININE-RNA

A comparison was made between the acceptor RNA from the wild
type and the arg.R~ mutant, using 15 fig. of the same preparation of
purified arginine-activating enzyme in each assay. Figure 3 shows the RNA-

10 15 20 25

Optical units s-RNA

Fig. 3. RNA dependence for the formation of arginine-RNA. Acceptor RNA
from strain 30S0 (wild type) represented by open circles, from strain 30S0A5 (the
arg . R-mutant) denoted by filled circles.

dependence for the formation of arginine-RNA. The slopes of the lines
correspond to an uptake of i arginine per 1900 nucleotides for the wild
type and i arginine per 1200 nucleotides for the arg.R" mutant. Figure 4
shows the enzyme dependence of the formation of arginine-RNA for
a pair of RNA preparations other than those used for the experiments of
Fig. 3. At all levels of enzyme, the amount of arginine-RNA formed in
1 5 min. is larger in the arg . R~ mutant than in the wild type. To test whether
or not the difference observed in the two previous experiments was specific
for arginine, the incorporation of arginine was compared to that of the
amino acid mixture (which contains about 6% arginine). Table II shows

304

H. G. BOMAN, I. A. BOMAN
TABLE II

Amino Acid Incorporation into Acceptor RNA

Incorporation of

Type of RNA

Specific activity of RNA

c.p.m./o.u.

per cent

Arginine

All amino acids

Wild type (30S0)
Mutant (30S0A5)
Wild type (30S0)
Mutant (30S0A5)

176
282
568
675

100
160
100
119

The concentration of the [^*C]-alga protein hydrolysate was 3-4 /xg./ml. ;
other conditions of the assay as described under Materials and Methods. The
specific activity in c.p.m./o.u. is given without correction for self adsorption and
geometry.

0 10 20 30

//g Arginine activating enzyme

Fig. 4. Enzyme dependence for the formation of arginine-RNA. Purified
arginine-activating enzyme from strain 3oS(, (the wild type) was used with 24-0
optical units acceptor RNA from the same strain (open circles) and with 23 • 8
optical units acceptor RNA from the arg.R" mutant 3oS|,A5 (filled circles). The
blanks used in this experiment were the c.p.m. obtained by extrapolation to zero
jj-g. of enzyme.

that the incorporation of the amino acid mixture is 19% higher into the
mutant RNA than into the wild type RNx\, whereas the incorporation of
arginine into the mutant RNA is 60*',, higher. However, it should be
pointed out, that in all experiments in Tables I and II, the incorporation
of the amino acid mixture is proportional to its concentration, while the

STUDIES ON THE INCORPORATION OF ARGININE

305

formation of arginine-RNA is independent of arginine concentrations
above o • i mM.

As a qualitative test of the product formed, a sample of acceptor RNA
from 30S0A5 was labelled with the more active of the ["C]-arginine
preparations (see Methods and Materials). The labelled RNA obtained
had a specific activity of 5730 c.p.m./o.u. It was subjected to mild alkaline
treatment (0-05 M triethylamine for 3 min. at 60°) and analyzed by paper
chromatography. Only the arginine spot could be detected, despite the
fact that the arginine used for the incorporation experiment showed four
spots of impurities.

The hydrolysis of arginine-RNA at pH 8-o was studied at three
temperatures. Samples of p^C]-arginine-RNA with 160 c.p.m./o.u. were
incubated at different temperatures; aliquots were removed at various

Fig. 5.

20 30 40 50 60
Time (min)
Rate of hydrolysis of arginine-RNA at pH 8 -o and different temperatures.

times and added to i ml. of cold o-oi M lanthanum nitrate in 0-5 M per-
chloric acid. The precipitate collected after centrifugation was counted.
Figure 5 shows these data as a percentage of the counts of the zero time
precipitate.

Discussion

THE ARGININE-ACTIVATING ENZYME

The chromatography of the arginine-activating enzyme described here
removes nucleic acid rather well but gives on a protein basis only about a
threefold increase in specific activity. However, the amount of protein
required to label i mg. of RNA in 15 min. is only 15 fig. (see Fig. 4). The
corresponding figure for a seven-times purified leucine-activating enzyme
calculated from Fig. 3 in ref. [16] was found to be around 4-7 mg. A
similar figure has been reported also for an isoleucine-activating protein

306 H. G. BOMAN, I. A. BOMAN

fraction [17]. There may thus be considerable differences in turnover
number between activating enzymes.

The experiments in Fig. i show a comparison of the incorporation of
arginine and leucine when the enzyme fractions had been obtained from
cells grown in the presence either of arginine or of leucine. The results
show that the addition of the amino acids to the media had no effect on
the level of the corresponding activating enzyme and that the repression
mechanism found for the enzymes in the biosynthesis of arginine [18]
does not operate on the arginine-activating enzyme. Further support for
this conclusion is provided by the results of several experiments in which
no appreciable difference was found between the quantities of arginine-
activating enzyme obtained from the wild type and the arg . R" mutant.

/;/ vivo experiments have earlier shown that canavanine inhibits growth
by interfering with the utilization of arginine in protein synthesis [19].
The present finding of the inhibition of the arginine-activating enzyme
suggests this step as the site of action of canavanine. Since the enzyme
from the mutant is as sensitive as that from the wild type, the mechanism
of the resistance in the mutant does not involve an alteration of this enzyme
(cf. ref. 20 and 21).

The experiments in Table I show that streptomycin inhibits the
formation of arginine-RNA and that this inhibition is greater for arginine
than for a mixture of all amino acids. Streptomycin is known to complex
with nucleic acids [22] and it is difficult to exclude that this presumably
non-specific effect contributes to the inhibition we have recorded here,
although no precipitation occurred in the presence of the buffer used in
our experiments. The increased sensitivity of the arginine incorporation
compared to that of the mixture of all amino acids (including arginine),
though rather slight, indicates some degree of specificity for the inhibition
of the formation of arginine-RNA. Some further support for a relation
between the utilization of arginine and the action of streptomycin is
suggested by the fact that Gorini [23] has obtained a mutant which
requires either arginine or streptomycin. It has also recently been found
that the genes for streptomycin resistance, the arginine repressor and some
of the arginine synthesizing enzymes are closely linked [6].

Davis and his co-workers [24] have recently shown that streptomycin
damages the cell membrane but they have concluded that this action though
necessary, may not be sufficient to account for the bactericidal effect. It
may well be that both the specific action of streptomycin in the incorpora-
tion of arginine into acceptor RNA and its non-specific complex formation
with nucleic acid contribute to the bactericidal action.

In vivo experiments with nitroarginine showed that it neither affects
growth, nor does it replace arginine for the growth of an arginine-requiring
mutant. The i^/^ increase in the rate of formation of arginine-RNA

STUDIES ON THE INCORPORATION OF ARGININE 307

observed in the presence of this compound is obscure, ahhough an in-
hibitor normally present in the cell and antagonized by nitroarginine could
account for an effect like this.

When comparing the four guanidino derivatives tested in Table I it
can be seen that those in which the electronegative character of the guani-
dino group has been decreased (nitroarginine and sulphaguanidine), have
no inhibitory action on the arginine-activating enzyme. This is in agree-
ment with the observation that the enzyme is relatively acidic and eluted
from the DEAE cellulose column with 038 m tris-HCl pH 7-4 whereas
the threonine-activating enzyme from calf liver was eluted by 0-12 m
tris-HCl pH 7-8 during otherwise similar conditions [25].

Sharon and Lipman [26] have studied the influence of analogues on the
trvptophan-activating enzyme and have compared their in vitro results
with in vivo studies on the growth of E. coli [27, 28]. It was found that
5-methyltryptophan and 6-methyltryptophan, inhibited both growth and
the tryptophan-activating enzyme. At a concentration ratio of analogue :
tryptophan of 200 :i the inhibition was 70" o for 5-methyltryptophan and
42% for 6-methyltryptophan in the in vitro experiments. In comparison
we found at a concentration ratio of 10:1 (analogue: arginine) 47^0
inhibition for canavanine and 38" ^ inhibition for streptomycin. The
somewhat greater effect of the arginine analogues lends additional sup-
port to the suggestion that the in vivo action of these inhibitors are on
the level of the activating enzymes.

THE ARGININE-RNA

The alkaline liability of the formed arginine-RNA as well as the other
characteristics of the reaction are consistent with the general notion of an
amino acid-acyl-RNA compound of the type previously described for
other amino acids [29, 30].

The difference in the amount of arginine- RX A between the wild type
and the arg.R~ mutant (see Figs. 3 and 4) has been observed in two
independent pairs of RNA preparations. At present we do not know
whether this difference is only a quantitative one or also a qualitative one.
The physiological significance of the difference in relation to the mechanism
of the arginine repression is also not clear. However, the observed
difference supports our original notion that repression is linked with RNA
metabolism and encourages us to continue studies along these lines.

308 H. G. BOMAN, I. A. BOMAN

Acknowledgments

We should like to express our thanks to Dr. F. Lipmann for much
help and many stimulating discussions.

References

1. Zamecnik, P. C, "The Harvey Lectures" Series 54 (1958-9), Academic
Press, New York (i960).

2. Lacks, S., and Gros, F.,^. mol. Biol, i, 301 (1959).

3. Gorini, L., and Maas, W. K., Biochim. biophys. Acta 25, 208 (1957).

4. Maaloe, O., in "Microbial Genetics". Cambridge University Press, 272
(i960).

5. Neidhardt, F. C, and Magasanik, B., Biochim. biophys. Acta 42, 99 (i960).

6. Borgois, S., Lavallee, R., Maas, W. K., and Wiame, J., in preparation.

7. Davis, B. D., and Mingioli, E. S., J. Bact. 60, 17 (1950).

8. Hager, L. P., Ph.D. Thesis, Univ. of Illinois (1954).

9. Method developed in the laboratory of Dr. F. Lipmann.

10. Crestfield, A. M., Smith, K. C, and Allen, F. W., J. biol. Chem. 216, 185

(1955)-

11. Kirby, K. S., Biochem.J. 64, 405 (1956).

12. Hammarsten, E., Hammarsten, G., and Theorell, T., Acta med. scand. 68, 219
(1928).

13. Warburg, O., and Christian, W., Biochem. Z. 310, 384 (1941).

14. Kossel, A., and Kennaway, E. L., Hoppe-Seyl Z. 72, 486 (191 1).

15. Peterson, E. A., and Sober, H. E.,X Afiier. chern. Soc. 78, 751 (1956).

16. Allen, E. H., Glassman, E., Cordes, E., and Schweet, R. S.,jf. biol. Chern. 235,
1068 (i960).

17. Hele, P., and Finch, L. R., Bioclieni. jf. 75, 352 (i960).

18. Gorini, L., and Maas, W. K., in "The Chemical Basis of Development",
ed. McElroy and Glass. Johns Hopkins Press 489 (1958).

19. Schwartz, J., and Maas, W. K.,^. Bact. 79, 794 (i960).

20. Davis, B. D., and Maas, W. K., Proc. nat. Acad. Set., Wash. 41, 775 (1952).

21. Hotchkiss, R. D., and Evans, A. H., Cold Spr. Harb. Symp. quant. Biol. 23,
85 (1958).

22. Cohen, S. ^.,y. biol. Chem. 168, 511 (1947).

23. Gorini, L., personal communication.

24. Anand, N., Davis, B. D., and Armitage, A. K., Nature, Lond. 185, 23 (i960).

25. Acs, G., Hartmann, G., Boman, H. G., and Lipmann, F., Fed. Proc. 18, 178

(1959)-

26. Sharon, N., and Lipmann, F., Arch. Biochem. Biophys. 69, 219 (1957).

27. Halvorson, H., Spiegelman, S., and Hinman, R. L., Arch. Biochem. Biophys.

55> 512 (1955)-

28. Pardee, A. B., Shore, V. C, and Prestidge, L. S., Biochim. biophys. Acta 21,
406 (1956).

29. Zachau, H. G., Acs, G., and Lipmann, F., Proc. nat. Acad. Sci., Wash. 44,
885 (1958).

30. Preiss, J., Berg, P., Ofengand, E. Y., Bergmann, F. H., and Dieckmann, M.,
Proc. nat. Acad. Sci., Wash. 45, 319 (1959).

I

POLYSACCHARIDES

Introduction

GUNNAR BlIX

Medicinsk-kemiska Institutionen, Uppsala, Sweden

In the present symposium problems related to polysaccharides occupy
a relatively modest place, being dealt with only at this afternoon's session.
This is perhaps a reasonable limitation in view of the general aim of the
symposium. But polysaccharides undoubtedly play a very important role
in "biological structures and functions" throughout the animal and plant
kingdoms down to the bacteria. Since the lectures we shall listen to during
the next few hours seem to be dealing mainly with metabolic and methodo-
logical aspects, it will perhaps be appropriate as an introduction to touch
upon the arrangement and functions of the polysaccharides in the tissues.
Because I have little personal experience of the conditions in plants I shall
confine myself to some words about polysaccharides in animal tissues. In
some regards they may have a wider application.

Bone, cartilage, skin, tendons, blood vessels, etc., all contain con-
siderable amounts of acid mucopolysaccharides. As you know it has turned
out during later years that there are many kinds of these substances:
hyaluronic acid, chondroitin sulphuric acids of different types, kerato-
sulphuric acid and others. Within each species minor modifications
probably occur, at least as regards degree of polymerization. Certain
general statements may be made about the arrangement of the mucopoly-
saccharides in the structural pattern of the connective tissues, and also
about their physiological functions. They belong to the amorphous ground
substance, they are present in the form of a sol or a gel in the pores of a
collagenous network and thev are loosely or more firmly associated with
proteins. It also seems safe to say that they are in part responsible for the
physical, and not least the mechanical properties of the connective tissues.
But when we come to questions such as exact ultrastructural or molecular
arrangements and precise mode of functions we are still very much in the
dark.

The physiological function of a biological substance is in the first
instance dependent on its chemical and physical properties. But these
properties may in difl^erent biological structures or environments have
very different physiological significance or consequences.

I should like to illustrate this point with an example. Hyaluronic acid,

312 GUNNAR BLIX

wherever it occurs, is a high-molecular, poly-anionic, highly hydrated,
essentially unbranched polysaccharide, the molecules of which seem to be
present in living tissues as randomly kinked coils. As might be expected,
aqueous solutions of this substance are highly viscous. The viscous
character of the joint fluids is due to hyaluronic acid, which occurs dis-
solved in the fluid together with some serum proteins. In this environment
the properties of hyaluronic acid make it suitable to serve as a lubricant,
protecting the cartilaginous surfaces in the joints against mechanical
damage. This may be regarded as the main function of the hyaluronic
acid of the joint fluid.

In the skin, subcutaneous, and other connective tissues the hyaluronic
acid has quite other functions. The characteristic mechanical properties
of these tissues, their rigidity, degree of compressibility, resistance to
injected fluids and so on are no doubt due to the particular arrangement
in which a sol or gel of hyaluronic acid (and of some other mucopoly-
saccharides) is included into the fine three-dimensional network of
collagen fibres. Variation in pore-size, in fibre-width, fibre-length and
orientation and in concentration, polymerization and ionization of the
polysaccharides may of course influence and modify the mechanical and
other properties of the tissues, but these fields are largely unexplored.

Karl Meyer and his collaborators have made the interesting observation
that the ratio between hyaluronate and chondroitin sulphate and the
ratio between different chondroitin sulphates vary with age. In pig's skin
the ratio between chondroitin sulphate B and hyaluronate was found to
be 1 : 5 in the new-born but a good i : i in the adult. In new-born infants
the rib cartilage contains a high concentration of chondroitin-4-sulphate.
This concentration decreases with increasing age, whereas the concentra-
tion of keratosulphate increases from the first year to adulthood and then
remains constant. Furthermore chondroitin-4-sulphate is practically all
replaced by the 6-sulphate. The physiological significance of these changes
is unknown, but since they concern chemical substances with different
properties they must undoubtedly influence the mechanical and other
properties of the tissues. The mechanism by which these chemical changes
are initiated may be something essential in the process of ageing.

There is evidently a long way to go before arriving at a clear under-
standing of the precise physiological significance of these and other
structural polysaccharides. The work still to be done must comprise
investigations of their primary and conformation structure, as well as
studies on their interactions with fibre and globular proteins. It should
also include a close inquiry into the physicochemical properties of the
pure substances and of the complexes formed between them and proteins.
It should of course involve direct studies on living tissues with optical
and other methods. Investigation on artificial models may also be valuable.

INTRODUCTION 313

The so-called structural polysaccharides should by no means be
regarded as metabolically inert material. Already the relatively rapid
metabolic turn-over of some of the acid mucopolysaccharides may be
taken as evidence of that. It has been suggested that the architecture of the
collagen bundles is determined by the type of polysaccharide formed by
the cells in various types of connective tissues, and that the polysaccharide
in some way or the other influences fibre formation. In embryonic develop-
ment and wound healing the fibre formation is preceded by or con-
comitant with the formation of mucopolysaccharides. Of these, hyaluronic
acid is usually the first produced, followed by sulphated forms. Some kind
of template function of the polysaccharides bearing upon secondary or
tertiary structure of connective tissue proteins might perhaps be con-
ceivable.

In any event the problems pertaining to formation and metabolic
turn-over of the polysaccharides must of course be of great interest in
connection with biological structure and function. These problems have
been very much to the fore during latter years and Prof. Dorfman and Prof.
Hestrin have, as you all know, made outstanding contributions in these
fields. We have learned that biosynthesis is not, as was earlier assumed, a
simple reversal of the process of hydrolysis. The enzymic polysaccharidic
syntheses work with donor substances, a common feature of which is that
they consist of a sugar substituted on the anomeric carbon atom. The
splitting of the bond between sugar and substituent supplies the energy
required for the polymerization. The availability of suitable donors and
appropriate enzymes (and perhaps other catalysts) will be the primary
factors directing and guiding the formation of the polysaccharides. This
field is as fascinating as it is complicated.

We await with great interest what Prof. Hestrin and Prof. Dorfman
have to tell us about their work.

As with all other sciences progress in the polysaccharide field is very
much dependent on the finding of new tools, principles, and methods of
investigations. The separation of biological substances in unchanged state
from the complex mixture in which they occur in living tissues, is one of
the basic methodological problems of biochemistry. Dr. Flodin will give
us a report of a new and promising device for the fractionation and
separation of biological substances, which seems promising not least for
the carbohydrate field, and I should not be surprised if it spread like an
epidemic in biochemical laboratories.

The Growth of Saccharide Macromolecules

Shlomo Hestrin

Department of Biological Chemistry, The Hebrew University,
Jerusalem, Israel

In the context of a symposium on " Biological structure and function",
it is surely appropriate to consider processes of polysaccharide synthesis
which directly underlie striking morphological change in living organisms.
Levan and dextran synthesis from sucrose and bacterial cellulose synthesis
afford instructive examples of reactions in this class. The syntheses of
these polymers proceed often to levels of product concentration at which
a polymer-rich aqueous phase separates out in the surround of the cells.
The ability for extracellular polymer synthesis appears to be so firmly
established in the genetic make-up of widely different species, that one
cannot but wonder whether some important biological function is not
fulfilled by this property. Experiments reported in this symposium by P. A.
Albertsson [i] are of interest in this connection. He has discovered that
when phases arise in a mixed aqueous solution of appropriate species of
macromolecules, particles dispersed in the system are often partitioned
between the phases in a highly selective manner, and that particle size
and shape affect this distribution very markedly. When phases separate
out in a living system in the wake of polysaccharide synthesis, a similar
specific partition pattern must occur. Thus a means may be afforded in
the evolutionary process whereby potent macromolecular agents could be
selectively concentrated or excluded from any separated phase. Such a
phenomenon might have particular importance in the coacervate systems
to which a prominent role has been assigned in a recent hypothesis
concerning the nature of the primaeval milieu in which biological structure
originated [2].

Size of the radical transferred from donor to the growing polymer
chain in syntheses of levan and dextran

Reactions catalyzed by a range of carbohydrases are now^ known to
consist in the transfer of a glycosyl rather of a glvcosido group from the
donor to an oxygen atom in the acceptor. If a change in substrate structure
at an atom position close to that at which the bond breakage is to occur

3l6 SHLOMO HESTRIN

can be expected to exert a more profound effect on the reaction rate than
a similar change effected at a relatively remote atom position, it might be
possible on the basis of enzyme specificity studies in a family of substrate
analogues to infer the position of the bond at which the enzymic cleavage
occurs [3]. However, an attempt to derive this inference from a considera-
tion of the substrate range of the polymer-synthesizing carbohydrases —
levansucrase and dextransucrase — encounters the difficulty that in both
these systems minor changes in glycose structure at atom positions on
both sides of glycosidic oxygen in sucrose result in a complete suppression
of the reactivity [4]. The question therefore arose whether these poly-
merizing enzymes are glycosylases, as are the common hydrolases, or
whether they are glycosidases and thus perhaps different in a salient aspect
of structure from the common hydrolase group. An investigation which
made it possible to select between these concepts was carried out in
collaboration with Dr. Frank Eisenberg during a visit to the National
Institute of Health at Bethesda.

Sucrose was synthesized enzymically by levansucrase-catalyzed trans-
fer of fructose from raffinose (melibiosyl fructoside) to [i-^^0]-glucose.
Sucrose formed was converted chemically first into its octoacetate and
then, by transacetylation to methanol, was reconstituted in crystalline
form. If a fructosido radical is transferred by the action of levansucrase,
synthesized sucrose should have been devoid of any ^^O. If fructosyl rather
than fructosido was the group transferred from raffinose to the added
glucose acceptor, the atom excess of ^*^0 in the synthesized sucrose was
expected, on the basis of the atom excess of ^^O in the glucose used, to be
o • 65 %. Experimentally the atom excess of ^^O in the recovered crystallized
sucrose was found to be o-6\%, in close agreement with the calculated
value. Hence it could be concluded that levansucrase, like the common
hydrolyzing /S-fructofuranosidase, is a glycosylase and not a glycosidase.

P^O]-Sucrose which had been synthesized in the above manner was
then incubated with dextransucrase of Leiiconostoc mesenteroides. Dextran
formed in this system and isolated from aqueous solution by repeated
precipitation with ethanol proved to be devoid of ^^O. Hence we are able
to draw the further conclusion that dextransucrase, like levansucrase, is a
glycosylase and not a glycosidase.

It follows that levansucrase and dextransucrase attack sucrose each on
a different side of the oxygen bridge. Should we be inclined, accordingly,
to write the structure of sucrose with a Lipman wiggle ( ^ ) to indicate the
site of the "high-energy-bond", we would be at a loss to decide the side
of the oxygen bridge in which the symbol could properly be placed.

The results with dextransucrase further afford conclusive proof that
the only site in enzymically synthesized sucrose at which ^^O occurred
was the glycosidic oxygen, P^O]- Sucrose synthesized as described by the

THE GROWTH OF SACCHARIDE MACROMOLECULES 317

action of levansucrase is easily available and can be expected to be a useful
aid for the elucidation of the intimate mechanisms of reactions in which
an interglycosidic oxygen bridge is concerned.

Fig. I. An early reaction time in synthesis of cellulose by Acetobacter xylimini.
Cells were incubated in a droplet of glucose solution on a collodion film for 2 min.
Product was freed from non-polymeric solutes in the extracellular phase by flotation
on water. A typical intercellular space is shown. There is an abundance of a poly-
meric material in the form of granules, rods and branched processes extending
from slimelike regions. At zero time of reaction the space was optically empty.
Pt shadow-cast (shadow ratio i :5). Magnification x 50 000. (Electronmicrograph
prepared by Dr. D. Danon and Air. I. Ohad.)

Growth of cellulose fibre in an extracellular phase

Synthesis of cellulose in Acetobacter provides an example of a poly-
merization system in which, as in the case of the production of levan and
dextran by cells, the accumulation of polymeric product occurs in the
extracellular phase. In the case of cellulose production, unfortunately,
even the general nature of the reaction mechanism remains obscure.

Greathouse [5] has reported that an ATP-fortified homogenate of
A. xylinum cells readily synthesizes cellulose. Stacey [6] has reported, on
the other hand, that attempts to duplicate this result have failed. Dr.

3l8 SHLOMO HESTRIN

Z. Gromet-Elhanan in our laboratory likewise attempted to repeat the
result described by Greathouse, and used for this purpose the same strain
in conditions which resembled as closely as possible those used by the
earlier workers. We have been unable in our laboratory to obtain any
significant synthesis of cellulose in this homogenate system.

Fig. 2. Cellulose fibres formed from glucose in a dilute suspension of cells of
Acetobacter xyliuiim. Fine filaments (" ultrastrands "), whose thickness as estimated
by measurement of shadow length is about 15 A, are seen both in solitary dis-
positions and in intertwisted bundles ("composite fibre"). Along the length of
some of the latter, aggregations of rods and granules ("amorphous formations")
are seen. Photograph was taken at a reaction time of 20 min. Intercellular space
at zero time was free from polymer. Specimen is mounted on a collodion film and
was freed from non-polymeric solutes by filtration over agar (Kellenberger's
technique). Pt shadow-cast (shadow ratio, i :6). Ahignification x 50000. (Elec-
tronmicrograph prepared by Dr. D. Danon and Mr. I. Ohad.)

The work of Glaser [7] has demonstrated that UDPG labelled in the
glucose moiety incorporates ^^C into cellulose in the presence of a sub-
cellular particle prepared from A. xylinmn. Mr. I. Ohad in our laboratory
has successfully repeated this experiment. However, it should be noted
that the cellulose-synthesizing activity manifested by this particle prepara-
tion is very poor. It has, moreover, not as yet been shown that the observed

THE GROWTH OF SACCHARIDE IMACROMOLECULES 319

incorporation reaction involves a polyrepetitive process rather than the
transfer of one or only a few glucose residues to the acceptor site. Hence,
one cannot be confident that this system is indeed one of complete cellulose
synthesis by a cell-free agent. Nor can it as yet be asserted with any con-
fidence that the major donor system involved in cellulose synthesis by this
bacteria is indeed UDPG itself rather than an analogue thereof.

Fig. 3. Morphological elements in cellulose formed from glucose in a dilute
suspension of cells of Acetobacter xylhnitn. In addition to the elements listed in
the legend of Fig. 2, this field reveals characteristic arrangements ("mats") of
polymer in parallel rows of granules or rods apparently patterned by an enclosed
framework of ultrastrands. Technique as in Fig. 2, except for the use of a longer
reaction time (about 2 hr.). Magnification x 25 000. (Electronmicrograph prepared
by Dr. D. Danon and Mr. I. Ohad.)

A production of electron-microscopically demonstrable cellulosic
fibre has been observed by Colvin [8] in A. xyliniim homogenates in-
cubated with glucose in presence of adenosine triphosphate. Colvin has
also reported that an ethanol extract made from a suspension of cells
incubated with glucose contains a solute which on transfer to water assumes
a fibrillar form, is alkali-insoluble, and affords glucose on acid hydrolysis.
When such an ethanol extract was heated and then supplemented with an
ultrafiltered Acetobacter preparation (aqueous extract of a dried suspension

320 SHLOMO HESTRIN

of cells in a glucose solution), the yield of fibre, as judged on the basis of
fibril incidence in an electron-microscope field, was significantly increased.
However, in view of the complexity of design and the sparsity of quantita-
tive information in these experiments, the chemical interpretation of these
results remains still uncertain.

In a recent note, Klungsoyr [9] revealed the existence in A. xylimim of
a disproportionating enzyme system which can catalyze a transfer of glu-
cosyl units from soluble cellodextrins into an insoluble cellulose fraction.
This may indeed prove an important clue towards the understanding of
the intermediary mechanism in cellulose synthesis.

Although hexose phosphate (a- and ^-glucose- 1 -phosphate andUDPG)
supplied to the cell exogenously does not yield cellulose, the assumption
that hexose phosphate is an intermediary of cellulose production can be
strongly supported. In this connection it may be noted that these cells
readily form cellulose both from hexonic acids (gluconate, 2-ketogluconate,

TABLE I

Distribution of Radioactivity in Cellulose
Formed from Specifically- Labelled Fructose

Substrate

Total radioactivity in

cellulose monomer

("o of that in

fructose)

Distribution of radioactivity in

cellulose monomer (total in

monomer = 100)

Ci

C2 C3 C4 Cs

C6

[ I -i*C] -fructose
[2-^*C]-fructose
[6-"C]-fructose

37 to 42

76

103

17
6

63
o

1 1
o

5-ketogluconate) and from other compounds (glucose, fructose) which
could readily be converted by an Acetobacter cell into hexose phosphate.
Furthermore, the view that hexose phosphate is an intermediate is also
supported by the observation that radioactivity recovered in cellulose
formed from specifically-labelled glucoses presents a distribution pattern
quite diff^erent from that in the original substance but which is similar
in unique features to the pattern which would arise if the intermediate
on the pathway between exogenous hexose and formed cellulose is hexose
phosphate in pentose cycle [10]. An analysis of radioactive carbon distri-
bution in the cellulose afforded from specifically-labelled fructose was
carried out in our laboratory by Dr. Gromet-Elhanan with results
shown in Table I. These findings have given additional support to the
view that the cellulose arises from hexose phosphate in a pentose cycle. It

THE GROWTH OF SACCHARIDE MACROMOLECULES

321

should be noted, however, that a detailed examination of the results does
reveal a quantitative deviation in some of the data from values that would
be predicted on the basis of the conventional scheme of the pentose cycle.
This implies, as indeed can readily be assumed, that hexose phosphate in
the metabolic pools of this cell is probably involved also in additional and
perhaps still undefined metabolic transformations.

Fig. 4. Cellulose pellicle formed from glucose in a relatively concentrated
suspension of cells of Acetobacter xylimmi. Cells are enmeshed in a cellulosic
film which consists of fibres running through relatively amorphous regions of
pol>Tner. Ultrastrands are resolvable in the fibre regions. Granular matter con-
stitutes the amorphous phase. Specimen, which was freed from non-polymeric
solutes by flotation in water, is mounted directly on a copper grid. Pt shadow-cast
(shadow ratio i :5). Magnification x 25000. (Electronmicrograph prepared by
Dr. D. Danon and Mr. I. Ohad.)

The Acetobacter enzyme-system which converts hexose phosphate into
cellulose is known to be anchored to the cell. Since the formed cellulose
fibre is observed in the extracellular medium as a free entity rather than
as a physical appendage of the organism, we may assume that the mor-
phological precursor of the fibre is a diffusible cellulose form — probably
a lone cellulose molecule — which escapes from the cell into the medium
wherein it finally enters into a crystalline fibrous habitat.

322 SHLOMO HESTRIN

If the diffusible entity which enters into the extracellular phase and
there serves as a precursor of the cellulose fibre is indeed itself chemically
identical with the cellulose in fibre, it could be anticipated that when cells
are incubated with glucose the point of time at which fibre appears might
be preceded by an interval during which a relatively large fraction of
cellulose molecules in the extracellular phase is still in a relatively dis-
organized state. Miihlethaler [ii] has shown that, when A. xylinum grows
in a complex medium, cellulose fibres arise in the vicinity of the cell within
an amorphous "slime". Studies recently conducted in our laboratory by
Mr. I. Ohad in collaboration with Dr. D. Danon at the Weizmann Institute
have indicated that when washed cells of this organism are suspended in
radioactive glucose solution the synthesis of cellulose proceeds linearly in
time without any observable induction phase, but that deposition of
fibre becomes apparent only after an initial interval during which electron-
microscopically discernible material is accumulated in the extracellular
phase within relatively amorphous formations which consist of an alkali-
insoluble, radioactive macromolecular compound — presumably cellulose
itself. Within a few minutes after contact of the cells with glucose, almost
all cellulose in the extracellular phase assumed a well-defined crystalline
habitat. Ribbons and ropes of cellulose arose by side-to-side aggregation
and intertwisting of an element which presented a remarkably constant
morphology — the cellulose " ultrastrand " The thickness of this element
was estimated by means of measurements of the length of the shadow which
it casts. On this basis, the thickness was shown to be in the range 15 A.
Width estimates were relatively more diflicult to arrive at in view of the
distortion imposed by metal-shadowing. Allowing about 80 A for con-
tribution made by metal to apparent width (cf. Hall [12]), the net width
of the ultrastrand can be estimated to have been about two or three times
the thickness. This would imply that the basic morphological element in
bacterial cellulose fibre is a bundle which comprises about twelve glucose
chains in its cross-section.

Celluloses from widely different sources including preparations of
bacterial origin, have been generally supposed, on the basis of electron-
microscope studies, to consist of fibrils ^ 60 A in diameter [13]. However,
correction was not made in any of these earlier studies for the effect of
metal on the apparent fibril width, nor was the thickness estimated on the
basis of measurements of length of shadow. In view of the present findings
on bacterial cellulose, a re-examination of the value which has been assigned
to the thickness dimension of the cellulose fibril in celluloses of different
sources may be desirable.

THE GROWTH OF SACCHARIDE MACROMOLECULES 323

References

1. Albertsson, P. A., these proceedings, Vol. i, p. 33.

2. Oparin, A. I., in "The Origin of Life ", First International lUBS Symposium.
Pergamon Press, London, 428 (1959).

3. Koshland, D. E., Jr., in "The Enzymes", Vol. i, ed. P. Boyer et ol. Academic
Press, New York, 305 (1959).

4. Hestrin, S., Feingold, D. S., and Avigad, G., Biochem.J. 64, 340 (1956).

5. Greathouse, G. A., J. Amer. chem. Soc. 79, 4503 (1957).

6. Stacey, M., in "Soc. Exp. Biol. Symposium XII". Cambridge University
Press, Cambridge, 185 (1958).

7. Glaser, L.,7- hioL Chem. 232, 627 (1958).

8. Colvin, J., Arch. Biochem. Biophys. 7, 294 (1957); Colvin, J., Nature, Loud.
183, 1 135 (1959)-

9. Klungsoyr, S., Nature, Lotui. 185, 104 (i960)

10. Schramm, AL, Gromet, Z., and Hestrin, S., Nature, Loud. 179, 28 (1957).

11. Miihlethaler, K., Biochim. biophys. Acta 3, 527 (1949).

12. Hall, C. E., J. biophys. biochem. Cytol. 2, 625 (i960).

13. Cf. review by Frey-Wyssling, A. in "Symposium on BiocoUoids ", _7. cell,
comp. Physiol. 49, (Supplement i), 63 (1957).

Discussion

Rogers : I wonder if Dr. Hestrin could tell us a little more about his comparison
of the rate of synthesis of cellulose by the UPDG system and his system, because
these comparisons are a little difficult to make especially when you are isolating
particles from bacteria with rather tough cell walls. It is a little difficult to know
what proportions of "particles" you have got out of the organisms, or how damaged
or undamaged the preparation is. The second point I am not quite clear about,
although I think you may have explained it already, is why you used fructose in
these experiments, because if the UDPG system were functioning then presumably
the fructose must get in, be reconstituted to the appropriate glucose-phosphate, the
UDPG be made and this transferred back to the cellulose and during the process
I would have thought there was an equal chance that endogenous supplies of glucose
might be used in preference to fructose, in any case the penetration of cells by
fructose is sometimes rather difficult.

Hestrin : As in the work reported by Glaser, we were only able to recover about
I °o or less of the cellulose-synthesizing activity of the cells in the equivalent
amount of particle. Even on a weight per weight basis, the activity of the particle
was less than that of the cell.

The advantage of using fructose as a substrate in the synthesis of radioactive
cellulose relates to the circumstance that fructose does not appear to be subject to a
direct oxidation, whereas glucose tends to be oxidized to gluconate and thence to
ketogluconate isomers each of which in turn can give rise to cellulose. The existence
of many alternate pathways by which glucose can form cellulose complicates the
calculation of the distribution which '^C supplied as glucose may be expected to
assume in the synthesized cellulose. Using fructose, however, we have no glu-
conate, we have no ketogluconate, but we still have of course, all the intermediates

324 SHLOMO HESTRIN

of the pentose cycle to contend with. There existed a fairly close agreement between
a pattern predicted on the basis of a schematic pentose cycle and the actual findings.

DoRFMAN : The first part of your paper might be of interest in this question
of specificity of the aglucone or glucone of glucosidase. Dr. Julio Ludowieg has
recently studied this problem with Dr. Vennesland and myself using ^*0 with
the hyaluronidases. In the case of the streptococcal hyaluronidase, as some of you
know, the product is an unsaturated compound. Presumable this is an elimination
reaction and, as one might have guessed, no incorporation of ^*0 from the medium
occurs during this hydrolysis. The cleavage is thus of the ether rather than the
glycoside link. With testicular hyaluronidase, which is a conventional glucosidase,
"O incorporation apparently occurs, the cleavage is thus apparently of the glyco-
side bond. Testicular hyaluronidase does not act on chondroitin sulphuric acid
B which contains L-iduronic acid instead of D-glucuronic acid but it does act on
both hyaluronic acid and chondroitin sulphuric acid A, one of which has glucos-
amine and the other has galactosamine. Thus the specificity does not reside in the
glycone portion of the molecule.

Hestrin : Such cases encourage the consideration of the possibility that the
interaction of enzyme and sugar does not exclusively involve one side of the sugar.
Conceivably the substrate has to be fitted into a pit on the protein surface, or per-
haps, as Wallenfels has conjectured, the protein winds itself around the substrate.

Mitchell: I would like to ask Dr. Hestrin a question stemming from his last
remarks. It seems as though the specificity of some of these enzymes implies that
there may be a hole into which the precursor goes and out of which the polymer
has to be extruded. Now, could I preface my question by saying that about five
years ago. Dr. Moyle and I discovered that in certain micrococci there was an
autolytic system that seemed to be capable of cutting a ribbon out of the spherical
cell wall so that it would fall readily into two hemispherical parts ; and this system
seemed to be mechanically attached to the wall, for it centrifuged with the cell
wall fraction of disintegrated cells and could not readily be washed oflF. We made
the suggestion that this system might be a synthetic system, and we visualized the
synthesis as though the enzymes were like the clasps on a set of zip fasteners fixed
in the plasma membrane and the zips, representing the unpolymerized cell wall
precursors, we imagined as being pushed through from the inside to become
zipped together, forming a wall on the outside. Now, could I ask in that context,
whether you know where the particles come from in your particulate preparations ?
Is it at all likely that they are originally part of the plasma membrane, and if so,
do you think perhaps a precursor is being pushed out which is not actually visible
yet as a fibre ?

Hestrin: The particles could be fragments of the cell wall; we don't know.
If they are fragments of the cell wall we could be more confident that the observed
incorporation activity truly represented a polymerization process leading to fibre
production. If the particles are of an endocellular origin, it is difficult to see how
they could have operated in vivo to give rise to extracellular deposition of fibre.
Perhaps, in that case, the observed incorporation of radioactivity was only a mani-
festation of an oligorepetitive rather than of a polyrepetitive process of trans-
glycosylation.

If we assume that the polyrepetitive step in synthesis occurs at the cell membrane

THE GROWTH OF SACCHARIDE MACROMOLECULES 325

and results in the liberation of a large cellulose molecule, we are attracted to suppose
that such molecules diffuse into the medium and there crystallize to yield fibrils.

Mitchell: Could I try to clarify my question a little ? In these organisms, one
would presume that the cell wall, the outer stiff region, is a molecular sieve with
holes in it that would be quite big enough to let through, say, sucrose molecules
or an individual fibre, one sugar molecule in width, which is going to form a
cellulose fibril. But, the plasma membrane, which is underneath the cell wall and
is impermeable to glucose phosphates will also certainly be impermeable to glucose
and fructose. We have never done permeability determinations on the organism
that Dr. Hestrin has mainly been speaking about, but with many others, even
when glucose is rapidly metabolized, the membrane is nevertheless quite imper-
meable to free glucose in the normal sense. You therefore have the problem of
how the precursor comes through the plasma membrane and why it polymerizes
outside and not inside the protoplast. We would like to imagine that the trans-
location and the polymerization of the precursor is controlled by a concerted cata-
lytic process.

Porter : As you may have guessed the other day I am fairly naive in this area
of wall formation but I have been impressed by the general facility with which cells
seem to shed their outer layer and then form a plasma membrane under the
shed cortex. I don't suppose this happens in the micro-organisms but I did want
to ask you if you thought the proposals made the other day that elements of the
ER might carry the catalysts for cellulose polymerization were at all feasible ?

Hestrin : There does not appear to be any protoplasmic connection between
cellulose in the medium and a cell. One might speculate that the cellulose mole-
cules diffuse from the cell and crystallize in the medium at a distance from the cell.
As to the polymerization step leading to the cellulose molecule, we still do not know
whether it occurs in the cell wall, on the outer surface, or within the cell.

Mucopolysaccharides of Connective Tissue*

Albert Dorfman and Sara Schiller

The LaRahida-Lniversity of Chicago Institute and the Departments of
Pediatrics and Biochemistry, University of Chicago, Chicago, III., U.S.A.

Polysaccharides, like other macromolecules, are generally classified on
the basis of structure and composition. However, rapidly expanding
research has made feasible an examination of the relationship between
biological function and chemical structure. Consideration of the physio-
logical role of polysaccharides suggests a distinction between energy-
yielding and structural polysaccharides. In general, the former appear to
be branched glucose polymers containing a glycoside linkages. Most
typical of this group are starch and glycogen. In contrast, the structural
polysaccharides are a more complex group of substances containing diverse
monosaccharides. Many appear to be /S-linked polymers. Examples of
these are cellulose, chitin, and the acid mucopolysaccharides of mammalian
tissues. Other polysaccharides, such as the hemi-celluloses, pectins, and
bacterial capsular polysaccharides, may be regarded as structural, but
chemical correlations have not been clearly established, particularly with
regard to steric configuration of the glycoside bonds. It seems likelv that
with continuing investigation of chemical-biological correlations, a better
understanding of the physiology of these substances may be reached.

Limitation of space does not permit an extensive review of the numerous
structural polysaccharides. This presentation will be confined rather to a
discussion of the acid mucopolysaccharides of mammalian connective
tissues. These compounds are linear polyanions which contain alternating
units of N-acetylated hexosamine and uronic acid (with one exception) ;
and in some cases, sulphate. They exist in the ground substance, a complex
mixture in which the formed elements of connective tissues are imbedded.

Figure i illustrates a portion of the chain of chondroitin sulphuric
acid-A, present in high concentration in mammalian cartilage. Alternating
glucuronic acid and N-acetylgalactosamine units are linked glvcosidically
to form linear molecules. The galactosaminidic bond is 1^*4 while the

* Original investigations reported in this communication were supported by
grants from The National Heart Institute of the United States Public Health
Service (No. H 311), The National Foundation, and The Chicago Heart Associa-
tion,

328 ALBERT DORFMAN AND SARA SCHILLER

glucuronidic bond is i— *3, a pattern which is followed in the three chon-
droitin sulphuric acids and hyaluronic acid [i]. The sulphate in chondroitin
sulphuric acid-A is esterified at C-4 of the acetylhexosamine residue
[2, 3]. The studies of Mathews and Lozaityte [4] indicate that in cartilage

i

Fig. I. The structure of the disaccharide unit of chondroitin sulphuric acid-A.

chondroitin sulphuric acid-A exists as coiled linear chains of molecular
weight 50 000 attached to a protein core forming a macromolecule with a
minimum molecular weight of 4 000 000. The state of other mucopoly-
saccharides is less well known although there is evidence that hyaluronic

TABLE I

Mucopolysaccharides of Connective Tissues

Amino sugar

Uronic acid

Sulphate

Hyaluronic acid

N-acetylglucosamine

Glucuronic acid

—

Chondroitin sulphuric

N-acetylgalactosamine

Glucuronic acid

+

acid-A

Chondroitin sulphuric

N-acetylgalactosamine

Iduronic acid

+

acid-B (/j-heparin)

Chondroitin sulphuric

N-acetylgalactosamine

Glucuronic acid

+

acid-C

Chondroitin

N-acetylgalactosamine

Glucuronic acid

—

Keratosulphate

N-acetylglucosamine

(Galactose)

+

Heparin

Glucosamine

Glucuronic acid

+

(N-sulphated)

(?)

Heparin mono-

Glucosamine

(?)

+

sulphuric acid

acid may exist in chains of considerably greater molecular weight [5]. The
presence of anionic groups along the chain imparts the capacity to bind
small cations in a manner similar to other polyanionic macromolecules
such as resins.

The known acid mucopolysaccharides, together with their component
sugars, are listed in Table I. Hyaluronic acid consists of alternating units
of N-acetylglucosamine and glucuronic acid and contains no sulphate.
Although the extent and nature of the linkage between hyaluronic acid

MUCOPOLYSACCHARIDES OF CONNECTIVE TISSUE 329

and protein is not yet established, it appears unlikely that a covalent bond
exists [6]. Hyaluronic acid occurs in a large number of mammalian tissues.
The presence of hyaluronic acid in the capsule of Group A streptococci
is unique. Polysaccharides of similar structure are found in capsules of
other microorganisms, but in no other authenticated case is the capsular
substance identical with a mammalian polysaccharide. The biological
function of hyaluronic acid is not clearly delineated; it is neither anti-
thrombic nor anticoagulant. Hyaluronic acid appears together with other
mucopolysaccharides in a number of connective tissues, but seems to be
characteristic of tissues with high water content. It is the principal poly-
saccharide of vitreous humour, synovial fluid, Wharton's jelly of umbilical
cord, cock's comb, and sex skin of certain primates. As will be demon-
strated below, thyroid deficiency with its concomitant oedema is charac-
terized by an increase in hyaluronic acid. While this association is highly
suggestive, its significance is not entirely clear. The specific association of
tissue hydration with hyaluronic acid rather than with the sulphated
polysaccharides is not readily explained on the basis of osmotic considera-
tions. Reinits [7], in a study of the rate of accumulation of hyaluronic acid
and water in the sex skin of certain primates found that hydration precedes
the peak of hyaluronic acid accumulation. He suggests that hyaluronic
acid serves to induce an accumulation of protein in the extracellular space
by hindering the access of protein to the lymph capillaries, with the con-
sequent increase in osmotic pressure due to the proteins. Reinits 's experi-
ments may be open to criticism on the basis of the inadequacy of the
methods for estimation of hyaluronic acid. Fessler [8] has emphasized the
role of hyaluronic acid in forming a system of relative non-compres-
sibility as a result of the interaction of a viscous solution with intermingled
collagen fibres. However, this study establishes no specificity for hyaluronic
acid in contradistinction to other acid mucopolysaccharides. That hyaluronic
acid does afford a mechanical barrier is attested by the action of hyaluroni-
dase in promoting the spread of particulates.

The association of hyaluronic acid with tissue hydration may be
attributed to the fact that hyaluronic acid is bound less avidly with protein
than are the sulphated polysaccharides. Insufficient data are available for
localizing individual polysaccharides in tissues which contain mixtures. It
is possible that the sulphated polysaccharides are more closely associated
with structural elements while hyaluronic acid is present in higher con-
centration in the amorphous gel between these elements. Mathews [9]
has demonstrated a strong afiinity of the chondroitin sulphuric acid
protein complex for collagen.

Chondroitin sulphuric acids-A and -C are present in largest con-
centration in cartilage and appear to be responsible to a considerable
degree for the unique physical characteristics of this tissue. The possible

330 ALBERT DORFMAN AND SARA SCHILLER

role of chondroitin sulphuric acids in calcification has been considered
but no mechanism has been clarified [lo]. Chondroitin sulphuric acid-A
diflFers from chondroitin sulphuric acid-C only in that the sulphate group
is esterified at carbon 4 of the galactosamine rather than at position 6.
The relationship of the sulphate position to the calcification process
merits further consideration. It is of interest to note that chondroitin
sulphuric acid-C predominates in elasmobranch cartilage which is not
converted into bone [11]. The interrelationships of chondroitin sulphuric
acids and collagen are not entirely clarified. Jackson [12] found that
treatment of tendon with hyaluronidase increased the solubility of collagen.
Chondroitin sulphuric acid-B is distinguished from chondroitin
sulphuric acids-A and -C by the presence of L-iduronic acid instead of
D-glucuronic acid [13, 14]. The sulphate group occupies the same position
as is characteristic of chondroitin sulphuric acid-A [2]. Figure 2 shows the

H

y^^ \h

OH H

OH OH\| 1/ OH

H OH

D-glucuronic acid L-iduronic acid

Fig. 2. The structures of D-glucuronic acid and L-iduronic acid.

structure of L-iduronic acid compared with that of D-glucuronic acid. The
two uronic acids are epimers difiFering only with respect to the stereiso-
merism at C-5. L-idose has not been identified in natural materials, but
L-iditol has been found in mountain ash berry by de Bertrand [15] in 1905.
Chondroitin sulphuric acid-B (/3-heparin) was isolated by Marbet and
Winterstein [16] from a commercial preparation of heparin and because
of its anticoagulant properties was believed to be an isomer of heparin.
It was found to be more potent as an antithrombic substance than as a
whole blood anticoagulant. Grossman and Dorfman [17] found that the
antithrombic activity of chondroitin sulphuric acid-B varied with the
concentration of thrombin. At low thrombin concentrations it is more
active than heparin, but at high thrombin concentrations it is inactive.
Like heparin, chondroitin sulphuric acid-B requires for activity a plasma
cofactor, the natural plasma antithrombin. In the absence of plasma it is
completely inactive. On whole blood the anticoagulant activity is only 5%
that of heparin. The physiological importance of the antithrombic activity
of chondroitin sulphuric acid-B is obscure. The relatively high concentra-
tion in certain tissues may permit a significant homeostatic role in the
regulation of fibrinogen-fibrin conversion within tissues.

MUCOPOLYSACCHARIDES OF CONNECTIVE TISSUE 33 1

The contrast between the anticoagulant properties of chondroitin
sulphuric acids-A and -B suggests that the substitution of L-iduronic acid
for D-glucuronic acid alters biological properties. It is possible that anti-
thrombic activity is related to some molecular characteristics which also
confers an increased capacity to bind certain ions [i8]. Hoffman et al.
[13] suggested that chondroitin sulphuric acid-B is associated with coarser
collagen bundles. This polysaccharide has been found in a variety of sites.
It was identified in gastric mucosa by Smith and Gallop [19] and has been
found in rabbit skin by Schiller et al. [20]. Of particular interest is the fact
that chondroitin sulphuric acid-B and heparin monosulphuric acid are
excreted in the urine and found in the tissues of patients afflicted with the
Hurler syndrome [21]. The enzymic defect in this heritable disorder of
connective tissue is unknown.

Davidson and Meyer [22] isolated from cornea a fraction composed of
D-glucuronic acid and N-acetylgalactosamine and less than one mole of
sulphate per disaccharide repeating unit. It was suggested that this com-
pound, chondroitin, is a metabolic intermediate in the biosynthesis of
chondroitin sulphuric acids-A and -C. Suzuki and Strominger [23] have
recently demonstrated the transfer of radioactive sulphate from PAPS * to
a chemically desulphated chondroitin sulphate by an enzyme isolated from
chicken oviducts, suggesting that sulphation of chondroitin sulphate
occurs at the macromolecular stage. This pathway cannot be considered as
established since the amount of sulphate incorporated was extremely low,
and biosynthesis of chondroitin has not yet been achieved. The physio-
logical and metabolic role of chondroitin must yet be considered uncertain.

Unlike other mucopolvsaccharides, keratosulphate lacks a uronic acid
component. Instead, a galactose group is substituted. Hirano et al. [24]
have indicated that the galactosyl bond is 1^4 and the glucosaminidic
bond is 1^3 with sulphate substituted on carbon 6 of the amino sugar.
Keratosulphate has been isolated from the nucleus pulposus [25]. Shetlar
and Masters [26] and Kuhn and Leppelmann [27] demonstrated an
increase in the ratio of glucosamine to galaciosamine in cartilage with
advancing age. This observation has found explanation in the demon-
stration by Kaplan and Meyer [28], that rib cartilage of the human adult
contains a large proportion of keratosulphate in contrast to the absence of
this substance in rib cartilage of newborns. Even larger amounts of kerato-
sulphate were reported in patients with the Marfan syndrome, although
it is dubious whether the difference was significant. Hallen [29] similarly
found that the relative concentration of glucosamine in nucleus pulposus
increased with advancing age. This was interpreted as an increase of
keratosulphate /chondroitin sulphate ratio. Davidson and Woodhall [30]

* '3-phosphoadenosine 5'-phosphosulphate.

332 ALBERT DORFMAN AND SARA SCHILLER

made similar observations. The difference between the physical-chemical
properties of keratosulphate and chondroitinsulphuric acids-A and -C
may be responsible for changes in the physical properties of cartilage and
nucleus pulposus observed with increased age. These facts have obvious
implications with respect to the pathogenesis of collapse of intervertebral
discs.

Heparin monosulphuric acid was isolated by Jorpes and Gardell [31]
in 1948 from a commercial preparation of heparin. Like heparin, it was
found to contain glucosamine and to demonstrate a positive optical rotation.
In contrast to heparin, N-acetyl groups and approximately one mole of
sulphate per disaccharide were found. Subsequently, Linker et al. [32]
isolated a similar compound from the liver of a patient with amyloidosis
and from aorta. They named the substance heparitin sulphate. Large
amounts of a similar compound have been isolated from the urine and
tissues of patients with the Hurler syndrome by Brown [33], Dorfman,
and Lorincz [21], and Meyer et al. [34]. Hen oviducts [35] and a mast cell
tumour [36] have also been shown to contain a similar substance.

Heparin monosulphuric acid is characterized by a high colour yield in
the Dische carbazole reaction for uronic acid, a positive optical rotation,
less than one mole of acetyl per disaccharide unit, and variable sulphate
content. In a recent study, Cifonelli and Dorfman [37] have shown that a
number of compounds can be separated from a crude preparation, obtained
as a by-product of heparin by the Upjohn Company. These materials all
exhibited a positive optical rotation and contained both N- and 0-sulphate
as well as N-acetyl. The total of N-acetyl and N-sulphate was approxi-
mately I. The N-acetyl content of most fractions approximated 0-7
mole per disaccharide unit. Preparations which manifested high N-acetyl
values contained virtually no N-sulphate. Preliminary studies of the
structure of these compounds suggest that they are not linked through the
3 position of the hexosamine as are the chondroitin sulphates and hyalu-
ronic acid, but are probably linked through the 6 position of hexosamine
[38]. These substances have little, if any, antithrombic or anticoagulant
properties. Their origin and metabolic importance have not been clarified
but chemical properties suggest a relationship to heparin. They may
represent intermediates in the biosynthesis of heparin.

Heparin has been studied intensively from a biological point of view
but its chemistry is yet poorly understood. It contains more than 2 sulphate
groups per disaccharide unit and demonstrates a positive optical rotation.
Unlike other acid mucopolysaccharides it has an N-sulphate group but no
N-acetyl group. In addition, one to two sulphate groups per disaccharide
repeating unit are present in other parts of the molecule.

The striking anticoagulant properties of heparin are well known.
However, its role in the homeostasis of blood coagulation is not clear.

MUCOPOLYSACCHARIDES OF CONNECTIVE TISSUE 333

Heparin has not been isolated from blood but there is no doubt that it is
present in mast cells. The possible physiological role of heparin assumed
increased importance with the discovery that injection of heparin induced
the release of a lipase responsible for " clearing" of blood [39]. The relation
of these observations to fat metabolism and arteriosclerosis is obvious.

The foregoing discussion reviews briefly the status of knowledge
regarding chemical structure and biological function of the acid muco-
polysaccharides of connective tissues. Alterations in tissue functions may
be a reflection of changes in metabolism of these substances. Clarification
of the metabolism of acid mucopolysaccharides required first a delineation
of the pathways of biosynthesis. These have been reviewed extensively
elsewhere [40].

Briefly, Alarkovitz et a/. [41] have shown that the acid mucopoly-
saccharide, hyaluronic acid, is formed from uridine diphosphoglucuronic
acid and uridine diphospho-N-acetylglucosamine by an enzyme isolated
from a strain of Group A streptococci. Although attempts to solubilize
this enzyme have not been successful, it has been possible to show that the
enzyme resides on the protoplast membrane [41]. Of interest, is the recent
report by Smith et al. [43] that Type III pneumococcal capsular poly-
saccharide is synthesized from uridinediphosphoglucuronic acid and
uridinediphosphoglucose.

Early studies on the metabolism of the acid mucopolysaccharides were
concerned with rates of turnover in mammalian connective tissue. Skin
was used as a source of connective tissue since it contains acid mucopoly-
saccharides in sufficient quantity for isolation and degradation [20]. The
rate of turnover of hyaluronic acid and a chondroitin sulphuric acid fraction
was determined in both rats and rabbits by utilizing [^^C]-glucose and
-acetate as well as ^'SO^, as precursors [44]. Glucose is a precursor of the
uronic acid and hexosamine moieties of the mucopolysaccharides ; acetate
is a precursor of the acetyl groups; and ^^SOj" is incorporated as ester
sulphate. These experiments showed that hyaluronic acid, with a half-life
of approximately 2 • 5 days, w^as metabolized at a rate comparable to other
metabolically active substances [45]. Chondroitin sulphuric acid was
metabolized more slowly. The rate of turnover was found to be similar
when measured with acetate, glucose, or sulphate, indicating complete
turnover of the entire molecule. Sulphate has been used by many investiga-
tors to label acid polysaccharides in diverse tissues.

Having determined these parameters of normal metabolism it was
possible to utilize similar techniques for a study of the efl^ects of hormones.
Cortisone and hydrocortisone were found to decrease the rate of turnover
of both the sulphated and non-sulphated mucopolysaccharides of skin [46].
This inhibition is time-dependent. Although the mechanism of action of
the adrenal hormones is unknown, this metabolic effect may be responsible

334 ALBERT DORFMAN AND SARA SCHILLER

for the delayed wound healings and demineralization of bone typical of
Cushing's disease.

Investigation of the metabolism of acid mucopolysaccharides in alloxan
diabetes indicated a decreased capacity to metabolize mucopolysaccharides
and a restoration of the defect toward normal by the administration of
insulin [47]. The rate of turnover of polysaccharides of partly fasted
animals, which served as controls for the weight loss of the diabetic rats,
was not different from that found in normal rat skin.

The influence of hypophysectomy [48], growth hormone, and thy-
roxine have also been investigated but these studies are difficult to interpret
since changes in pool size occurred during the course of the experiments.
Turnover studies depend upon the maintenance of a steady state or
constant pool size ; differences in rates of synthesis are difficult to define
if a reasonably constant pool size does not obtain.

The information derived from turnover experiments suggested that
marked variations in the concentrations of mucopolysaccharides might
occur. It therefore became important to devise methods for the quantita-
tive determination of acid mucopolysaccharides in tissues.

By the application of relatively simple procedures it is now possible to
perform such analyses. The method, which is published elsewhere in
detail [49], depends upon the solubilization of tissues with papain, followed
by further deproteinization with trypsin and trichloroacetic acid. The
partly purified polysaccharide preparation is separated on the basis of the
differential solubility of complexes formed with cetyl pyridinium chloride
(CPC). In the presence of 0-03 5-0 -040 m NaCl all the polysaccharides are
quantitatively precipitated by CPC providing Celite is added. When the
resultant precipitate is extracted with varying concentrations of NaCl
sharp separation of three fractions is obtained. The CPC-hyaluronic acid
complex is solubilized by 0-4 M NaCl in o- 1% CPC, while the complexes
of chondroitin-sulphuric acids-A, -B, -C, and heparin monosulphuric
acids are solubilized by i -2 M NaCl in o-i^',, CPC. Least soluble is the
complex of heparin which is solubilized in 2 • i m NaCl. Further separation
of these fractions can be achieved by subsequent chromatography on
Dowex 1x2, chloride.

Table II illustrates a typical separation of a preparation of rat skin.
The small amount of colour found in tubes i and 2 is due to interference
by protein with the carbazole reaction and does not indicate significant
loss of mucopolysaccharide. In the case of rat skin the fraction obtained
with 1-2 M NaCl contains chondroitin sulphuric acids, but in other
tissues heparin monosulphuric acid may be obtained in this fraction.

The validity of this method has now been established by a variety of
prodedures. Table III indicates the recovery of polysaccharides added to
rat skin in duplicate experiments. Excellent recovery and reproducibility

MUCOPOLYSACCHARIDES OF CONNECTIVE TISSUE

TABLE II

The Separation of Acid Mucopolysaccharides from
AN Extract of Rat Skin*

335

Flask No.

Solvent

Volume

Uronic acid

1 Supernatant

2 003 M XaCl in oiO„ CPC

3 0-4 M NaCl in o-if-o CPC

4

5

6

7
8

9 I • 2 M NaCl in o • I "^'o CPC

10
II
12
13

14 2-1 M NaCl

IS
16

ml.

\J-g-

9-5

317 (brown)

15-0

95 (yellow)

412

30-0

1620

29-8

1 09 1

30-0

612

29-8

122

30-0

24

30-0

0

3469

30-5

885

30-0

297

30-3

127

30-0

30

30-5

0

1339

30-0

716

30-0

38

29-8

0

754

* The uronic acid content was 5 ■ 89 mg. in a volume of 9 ■ 5 ml. To this were
added 0-38 ml. of i ivi NaCl and 0-32 ml. of a 10*^0 solution of CPC, bringing the
total volume to 10 -2 ml. The mixture was incubated for i hr. at 37". Approximately
200 mg. of Celite were added prior to centrifugation as described in the text, o • 2
to I -o ml. aliquots were removed from each flask for uronic acid determinations.
The recovery was 5-97 mg. uronic acid or ioi°o-

was achieved. The procedure has now been successfully adapted to analysis
of human skin, human aorta, human lung, and basement membranes of
dog and human kidney.

Table IV illustrates a series of analyses on rat skin [50]. A surprisingly
large amount of heparin w^as found. Variation of polysaccharide content
with age was noted. Particularly striking was the decrease in concentration
of heparin beyond 44 days. Little or no heparin was found in hog, guinea-
pig, rabbit, foetal calf, camel or human skin. The physiological significance

336 ALBERT DORFMAN AND SARA SCHILLER

TABLE III

Recovery of Acid Mucopolysaccharides added to Skin

Recovery of total mucopolysaccharides Recovery of
Amount added muco-

Substance
added

Hyaluronic
acid

Chondro-
itinsul-
phuric
acid-B

Heparin

o 4 M NaCl I • 2 M NaCl 2 • i M NaCl polysaccharides

mg.^

2-53

0-95

0-71

mg. uronic acid per 5 g.
acetone-drv skin

2 -96

I -40

0-57

0/
/o

5-08

1-36

0-45

93

4-42

I -05

0-47

77

3-II

2-40

0-62

102

3-20

2-33

0-62

99

2-88

1-44

I • 12

87

2-72

I-I3

116

90

* As uronic acid.

of this reduction in polysaccharide content, particularly with respect to
hyaluronic acid and heparin, is not entirely clear. Previous claims [51]
that with advancing age there is a shift from hyaluronic acid to chondroitin
sulphuric acid are not substantiated by these data although a relative shift

TABLE IV
Effect of Age on Concentration of Mucopolysaccharides in Rat Skin

Age
(days)

Hvaluronic acid

Chondroitin
sulphuric acid

Heparin

21*
23*
44*
57*
74*

217*

406

473
561
893
951

fj.g. uronic acid/g. dry rat skin

979 385

894 406

1232 337

439 218

540 197

578 273

320 140

540 200

480 257

380 135

538 256

421
562
388
100

92
1 1 1

70
1 10

68

35
100

* These data were obtained from pools of eight to ten rat skins for each age
group. In the older age groups, the results represent values from individual
animals.

MUCOPOLYSACCHARIDES OF CONNECTIVE TISSUE 337

does occur. As noted, however, hyaluronic acid does appear to be correlated
with the higher water content of the skin of young rats [52]. Further study
is necessarv to appreciate fully the effect of this shift on physical structure
of ground substance and the consequent influence on the metabolism of
cells imbedded therein.

This method of analysis has been used to study further the effect of
insulin on the svnthesis of acid mucopolysaccharides. The turnover
studies discussed previously indicated that alloxan diabetes effects a
decrease in turnover rate of both hyaluronic acid and the chondroitin
sulphuric acid fraction. Analysis of skin from these animals confirmed a
suspected decrease in concentration of these substances. Table V indicates

TABLE V

Effect of Insulin on Mucopolysaccharide
Content of Rat Skin

Type of

Age

Body
weight

Distribution

of mucopolysaccharides of skin

rat

HA

CSA

H

eparin

days

S-

H'S-

U.A.

per g. dry

skin

Normal

74

340

720

319

158

Restricted

food

74

200

730

282

159

Diabetic

74

185

448

234

196

Normal

81

317

624

274

lOI

Restricted

food

81

189

701

260

156

Diabetic

81

182

371

190

205

Diabetic

+ insulin

81

273

506

22s

175

the results of these experiments. Noteworthy is the marked decrease in
hyaluronic acid concentration and a less striking decrease of chondroitin
sulphuric acid in contrast to heparin. The latter compound is actually
increased in the diabetic animals. This may be due to a relative decrease in
total ground substance. As in the turnover experiments, these effects were
reversed by insulin. It seems possible that this defect of mucopolysac-
charide metabolism in diabetes mav plav a role in the delayed wound
healing, decreased resistance to infection and accelerated degenerative
changes which occur in diabetes melitus.

Of particular interest has been the study of the effects of thyroid
hormone on mucopolysaccharide metabolism. For this purpose propyl-
thiouracil was employed to inhibit synthesis of thyroid hormones in rats.
Table VI illustrates the results of such experiments. In contrast to previous

338

ALBERT DORFMAN AND SARA SCHILLER

TABLE VI

Effect of Thyroid Hormone on Concentration of
Mucopolysaccharides in Rat Skin

Age

Weight

Thyroid
weight

Distribution

of

mucop

olysaccharides

HA

CSA

Heparin

days

av. in g.

av. in g.

MS-

U.A.

per g.

dry

skin

Normal
Normal

88
86

334
333

191
19-4

785
704

290
264

158
147

PTU
PTU

PTU

78
86
88

194
190
192

77-5
90-7

964

852

976

1024

182
220
204

93

115

87

PTU + T4
PTU + T4
PTU + T4

78
86
88

254
279

273

13-7
25 I

i8-5

778
748
809

249

234
271

117

94

137

experiments a striking differential effect was observed. Concomitant with
a decrease in the concentration of chondroitin sulphuric acid, there was
marked increase in hyaluronic acid. Administration of thyroxine reversed
these effects toward normal. The biochemical or endocrinological
mechanisms are not clear. Nevertheless, these findings may afford an
explanation for the mechanism of myxoedema.

Summary

The information presented in this paper indicates that the connective
tissues represent not only a mechanical support for parenchymal cells, but
a controlled and controlling environment. The acid mucopolysaccharides
are a family of compounds with a chemical unity. Nevertheless, the
chemical variations within the group are mirrored by differences in
biological activity. Various tissues exhibit variations in polysaccharide
composition. Furthermore, it is apparent that under physiological and
pathological influences, this composition may be altered both qualitatively
and quantitatively. Such variation results in a change in the milieu in
immediate contact with parenchymal cells.

Acknowledgment

The authors are grateful to Dr. Martin B. Mathews for many valuable
discussions of ideas presented in this paper.

MUCOPOLYSACCHARIDES OF CONNECTIVE TISSUE 339

References

1. Mever, K., "The Harvey Lectures", Series 51 (1955-56), Academic Press, New
York, 88 (1957) •

2. Mathews, M. B., Nature, Loud. 181, 421 (1958).

3. Hoffman, P., Linker, A., and Meyer, K., Biochini. biophys. Acta 30, 184 (1958).

4. Mathews, M. B., and Lozaityte, L, Arch. Biochem. Biophys. 74, 158 (1958).

5. Laurent, T. C, Ark. Kemi ll, 487 (1957).

6. Ogston, A. G., and Sherman, T. F., Biochem. J. 72, 301 (1959).

7. Reinits, K. G., Biochem. J. 74, 27 (i960).

8. Fessler, J. H., Biochem. J. 76, 124 (i960).

9. Mathews, M. B., Circulation 14, 972 (1956).

10. Sobel, A. E., Burger, M., Samachson, J., and Slovik, N., "Resumes Commns.
3rd Intern. Congr. Biochim., Brussels", 48 (1955).

11. Nakanishi, K., Takahashi, N., and Egami, F., Bull. chem. Soc, Japan 29,

434 (1956).

12. Jackson, D. S., Biochem. J. 54, 638 (1953).

13. Hoffman, P., Linker, A., and Meyer, K., Biochim. biophys. Acta 69, 435 (1957).

14. Cifonelli, J. A., Ludowieg, J., and Dorfman, A.,^. biol. Chem. 233, 541 (1958).

15. de Bertrand, G., Bull. Soc. chim. Fr. 33, 264 (1905).

16. Marbet, R., and Winterstein, A., Helv. chim. Acta 34, 231 1 (1951).

17. Grossman, B. J., and Dorfman, A., Pediatrics 20, 506 (1957).

18. Mathews, M. B., Biochim. biophys. Acta 37, 288 (i960).

19. Smith, H., and Gallop, R. C, Biochem. J. 53, 666 (1953).

20. Schiller, S., Mathews, M. B., Jefferson, H., Ludowieg, J., and Dorfman, A.,
J. biol. Chem. 211, 717 (1954).

21. Dorfman, A., and Lorincz, A. E., Proc. nat. Acad. Sci., Wash. 43, 443 (1957).

22. Davidson, E. A., and Meyer, K.,y. biol. Chem. 211, 605 (1954).

23. Suzuki, S., and Strominger, J. L.,^. biol. Chem. 235, 274 (i960).

24. Hirano, S., Hoffman, P., and Meyer, K., Fed. Proc. 19, 146 (i960).

25. Gardell, S., Acta chem. scand. 9, 1035 (1955).

26. Shetlar, M. R., and Masters, Y. F., Proc. Soc. exp. Biol. 90, 31 (1955).

27. Kuhn, R., and Leppelmann, J., Leibigs Ann. Chem. 6ii, 254 (1958).

28. Kaplan, D., and Meyer, K., Nature, Lond. 183, 1267 (1959).

29. Hallen, A., Acta chem. scand. 12, 1869 (1958).

30. Davidson, E. A., and Woodhall, \i.,y. biol. Chem. 234, 2951 (1958).

31. Jorpes, E. A., and Gardell, ?).,y. biol. Chem. 176, 267 (1948).

32. Linker, A., Hoffman, P., Sampson, P., and Meyer, K., Biochifn. biophys.
Acta 29, 443 (1958).

33. Brown, D. H., Proc. nat. Acad. Sci., Wash. 43, 783 (1957).

34. Meyer, K., Grumbach, M., Linker, A., and Hoffman, P., Proc. Soc. exp.
Biol. Med. 91, 275 (1958).

35. Schiller, S., Biochim. biophys. Acta 32, 315 (1959).

36. Roden, L., and Dorfman, A., Acta chem. scand. 13, 2121 (i960).

37. Cifonelli, J. A., and Dorfman, A.,^. biol. Chem. 235, 3283 (i960).

38. Cifonelli, J. A., and Dorfman, A., Fed. Proc. 18, 204 (1959).

39. Korn, E. D.,jf. biol. Che?n. 215, i (1955).

40. Dorfman, A., Markovitz, A., and Cifonelli, J. A., Fed. Proc. 17, 1093 (1958).

41. Markovitz, A., Cifonelli, J. A., and Dorfman, A., Jf. biol. Chem. 234, 2343

(1959).

42. Markovitz, A., and Dorfman, A., Fed. Proc. 19, 146 (i960).

34° ALBERT DORFMAN AND SARA SCHILLER

43. Smith, E. B., Mills, G. T., Bernheimer, H. P., and Austrian, R.,.7- ^ioL Chem.
235» 1876 (i960).

44. Schiller, S., Mathews, M. B., Goldfaber, L., Ludowieg, J. and Dorfman, A.,
y. biol. Chem. 212, 531 (1955).

45. Schiller, S., Mathews, M. B., Cifonelli, J. A., and Dorfman, \.,jf. biol. Chem.
218, 139 (1956).

46. Schiller, S., and Dorfman, A., Endocrinology 60, 376 (1957).

47. Schiller, S., and Dorfman, A.,_7. biol. Chem. 227, 625 (1957).

48. Schiller, S., and Dorfman, A., Fed. Proc. 16, 242 (1957).

49. Schiller, S., Slover, G. A. and Dorfman, A., J. biol. Chem. 236, 983 (1961).

50. Schiller, S., and Dorfman, A., Nature, Loud. 185, 11 1 (i960).

51. Loewi, G., and Meyer, K., Biochim. biophys. Acta 27, 453 (1958).

52. Lowrey, G., Anat. Rec. 7, 143 (1913).

Discussion

DiscHE : I think that the findings of Dr. Mathews in your laboratory about this
combination between the chrondroitin sulphate and proteins are of particular
interest because there appears here a possibility to introduce the factor of specificity
into the field of hexuronic acid-containing polysaccharides. The specificity cannot
be related in these compounds solely to the composition of the carbohydrate but
where proteins enter also into the structure and there is a multiplicity of poly-
saccharide molecules on every protein, there is also a possibility of variations, which
can be associated with changes in specificity ; this is a very interesting development
in this respect. I should like to make one remark on changes with age. I think we
must be very cautious in establishing any such correlations. Dr. Karl Meyer who
found such changes in the ratio between chrondroitin sulphate and keratin sulphate
told me that this finding is not valid for the rat.

Rogers: Just two points, one of which might be of possible interest to Dr.
Dorfman, he just touched on the question of the /3-linkage. Recently we have been
examining cell wall preparations of Bacillus subtilis for a different reason altogether
and we had cause to make a trichloroacetic extract of the walls and found there,
rather to our surprise, a polysaccharide which we isolated in a homogeneous state.
It had exactly the same composition as chondroitin but had positive rotation instead
of a negative rotation and was hydrolyzed very readily with acid. We could not
separate a disaccharide from it after mild acid hydrolysis as you can with chon-
droitin. Finally, following very beautiful work of Barker and his colleagues, we
examined the infra-red spectrum at low wave numbers and this showed the
expected band at about 850 cm~^ which these authors found to be very typical of
a-linkages and indeed this band was not present in chondroitin, thus it rather looks
as though in this organism there is an a-linked substance which so far as we can see
at the moment has the same composition as chondroitin; but we have no evidence
that it behaves as a linear polyelectrolyte like chondroitin ; it may be a highly
branched chain compound. We have also looked at several species of micro-
organisms during the last few years for the distribution of hexosamine-containing
compounds and found that over 95% of materials which contain amino sugars
appear to be outside the permeability membrane. We took the soluble protein con-
stituents and we could never find more than about 5% of the total hexosamine of

MUCOPOLYSACCHARIDES OF CONNECTIVE TISSUE 34I

the cells there and we could always account for the other 95 "o either in the capsular
material or in the cell wall. I wonder if you would like to speculate as to why
hexosamine-containing substances seem to occur principally outside the per-
meability membrane.

Dorfman: With regard to Dr. Dische's comments I am glad that he did
mention this point because there wasn't time to expand on it. As a matter of fact
we are at present reinvestigating the question of antigenicity of these substances
with the protein complexes which are quite different compounds from alkali-
degraded material. We have been able to show in work that I didn't cover (Gross,
J. I., Mathews, M. B., and Dorfman, A., jf. biol. Chem. 235, 2889, (i960)) that
in vivo the entire protein-polysaccharide complex appears to turn over as a unit.
We have measured the metabolism of protein and carbohydrate separately and
simultaneously in the same animals and have shown that the rates of turnover are
identical for the protein part and the carbohydrate part. This raises some important
implications on carbohydrate biosynthesis. With regard to Dr. Rogers' comment, I
was careful not to make an absolute correlation between a- and ^-linkages because
I don't think that it would hold up. I think that there are other exceptions. ^-Link-
ages arise more readily from linear compounds.

As far as my speculating as to why hexosamine compounds are outside the mem-
brane, I suppose I should quote a very famous Swedish teacher I had, Prof.
Anton J. Carlson who, whenever asked a question of that kind, said he wasn't
there when they made it. I don't know what it is about the chemical properties of
hexosamine which is desirable for Nature to incorporate it into structural polysac-
charides. Chitin for instance in the insects and fungi, is also a ^-linked glucosamine.

Mitchell: I wonder if I might speculate a little on why these compounds
appear outside, because I am afraid that my remarks earlier about the zip fastener
idea were not very clear and were not really understood. Consider a piece of proto-
plast membrane as shown in the diagram. It is creating a growing cell wall on the
outside and one believes that the precursors are inside, in the cytoplasm and one
knows that the cell wall is outside the plasma membrane. The most interesting
question is why is the wall made outside and not inside the plasma membrane ? This
is a directional matter: why should there be a vector component of the chemical
process, seen as a whole ? The simplest way of explaining this would be to say that
an enzyme {A) (Fig. i), rather like the ones which Dr. Hestrin is imagining with a
sort of hole through it, is accepting the precursors on the inner surface of the
plasma membrane and polymerizing them as they pass through to the outside — the
polymerized chain becoming extruded as it forms. Such as enzyme, catalyzing a
vectorial metabolic process would, of course, have to be specifically located to some
substratum in the plasma membrane complex by means of primary or residual
bonds.

As the cell wall is fairly porous, the externally secreted polysaccharides, such as
the capsular polysaccharides and the cellulose of Acetobacter xylinum, could be
produced and positioned by a similar mechanism. An enzyme polymerizing an
extracellular polymer is depicted at B, and the fine chains of the polysaccharide are
supposed to be thin enough to go through the holes in the substance of the wall, so
that again we would have a vector component of the metabolic process playing a
morphogenetic role. This conception is of interest to those trying to relate structure

342 ALBERT DORFMAN AND SARA SCHILLER

to function, I think, because it suggests a connection between morphogenesis and
the unique asymmetry of protein molecules.

Dorfman: I like this very much and there isn't time to expand on it. Last year
we proposed a mechanism for polysaccharide synthesis which envisaged a mechan-
ism of this kind. One of the difficulties is that we find no low-molecular weight
intermediates. We have to think of an enzyme which somehow or other forms
hyaluronic acid on the enzyme while the chain keeps getting longer until it is large
enough and then is detached from the enzyme.

Hestrin : In our laboratory Schramm and Zelinger have found that /3-glucose-i-
phosphate can be condensed in the presence of maltose phosphorylase with
glucosamine to afford in good yield the ot-glucosyl-i,4-glucosamine. They have
also obtained similarly the acetylglucosamine derivative. These materials are thus

Cytoplasm

Plasma membrane

Medium

Cell wall polymerizing enzyme ""Xl-s:*^^^

^"v aJ

Cell wall precursors — ^^=*'^

Extracellular polysaccharide precursors _ ^^

NO

Extracellular polysaccharide ' ^■^=t^-'

polymerizing enzyme

Extracellular
polysaccharide

Fig. I.

now readily available and may be useful for studies designed to elucidate the effect
of an a-linkage on the properties of glucosamine-containing polymers.

A general comment concerning the interpretation of data obtained in enzyme
solution on a substrate of a saccharide synthesis might perhaps be appropriate.
There is a tendency to take the view that a demonstration of synthesis in an en-
zyme solution suffices to demonstrate that the substrate in question is also the
physiological substrate. However, especially if the reaction observed in a solution is
sluggish, an observed reaction may be only an artifact. Some of the uridine
diphosphoglycoses, for example, can still conceivably merely be analogues of
unidentified physiological substrates. Before accepting a particular polymer syn-
thesis reaction as having been demonstrated, we should insist on a realization of the
reaction under conditions in which there occurs a net increase of the mass of the

MUCOPOLYSACCHARIDES OF CONNECTIVE TISSUE 343

polymer, with attendant increase either in the number of the polymer molecules
present and/or large increase in their molecular weight.

DORFMAN : I should point out that we have shown net synthesis of hyaluronic
acid.

Marshall: Dr. Dorfman's work is tremendously stimulating to some of us
concerned with very different problems in cell physiology. May I add to Dr.
Porter's comment a remark based on experience with amoebae ? We have come to
view the amoeba surface as a combined structure, with an inner membrane and
an outer mucoid coat. Both components must be thought of in dynamic terms,
because in amoebae we observe new formation and expansion of surface without
the appearance of successive layers, such as Dr. Porter has just described in
chondrocytes. Rather there appears to occur some sort of interstitial formation of
new membrane and new coat material. I don't know whether the amoeba surface is
unusually labile, or unusually rapid in turnover compared to that of other cells,
but it suggests an even more dynamic way to conceive of the passage of substances
in or out of a cell, without there being at any time a stable structure in which holes
must appear.

Separation of Oligosaccharides with Gel Filtration

Per Flodin and KAre Aspberg

Research Laboratory, AB Pharmacia and the Institute of Biochemistry,

Uppsala, Sweden

In a series of oligosaccharides the individual members differ very
little in properties and the preparation of them requires highly selective
methods. Dickey and Wolfrom [i] succeeded in separating the acetylated
cellodextrins up to the cellohexaose chromatographically in magnesium
and calcium silicate columns. The acetates were adsorbed from a chloro-
form solution and the chromatogram was developed with a benzene-
ethanol mixture. The most commonly used procedure is to adsorb the
oligosaccharide mixture on charcoal from a water solution and then elute
them with a gradient of increasing ethanol concentration [2]. With minor
modifications the method has been applied to cellodextrins, maltodextrins
and isomaltodextrins. The selectivity is large enough to give a complete
separation up to the hexaoses. The higher members are generally obtained
as fairly pure substances containing the neighbouring oligosaccharides as
contaminants.

In an ion exchange resin sucrose and glucose move at different rates
when conditions are chosen so that no adsorption takes place. Similarly,
in swollen starch the retardation is smaller the higher the degree of
polymerization [3].

In columns packed with particular gels formed by cross-linking
dextran, large molecules were observed to emerge first from the column
while small molecules were retarded [4]. The same phenomenon was
observed when low molecular weight dextrans were fractionated [5]. It
should be mentioned that all operations were made in the same solvent.
The behaviour is in contrast to what occurs in the adsorption chromato-
graphic techniques, where adsorption is stronger the higher the molecular
weight. It is consistent with a molecular sieve mechanism according to
which a larger part of the gel particle is available the smaller the molecular
size. The object of this communication is to show that the method, which
has been named gel filtration, may be used to separate cellodextrins.

Experimental

The dextran gel used was Sephadex G-25 (Pharmacia, Uppsala,
Sweden) 200-400 mesh dry sieved. It was swollen in water and sedi-

346 PER FLODIN AND RARE ASPBERG

mented, and the fines remaining in the supernate was removed by decanta-
tion. It was then packed into a glass column with the dimensions 4-5 by
150 cm. The packing procedure was that described by Flodin [6]. When
all the material (300 g.) had packed, a filter paper was put on top of the
bed. The height of the bed was 126 cm. and its volume 2000 ml.

By passing a zone of indian ink through the bed the void space was
determined and found to be 515 ml. and simultaneously a check of the
packing was obtained.

The oligosaccharide mixture was obtained by acetolysis of cellulose
followed by hydrolysis as described by Whittaker [2]. The product was
extracted with water and the solution obtained contained about 10%
oligosaccharides. 20 ml. were applied to the column and eluted with water

4,000 -

?, 3,000

§ 2,000

1,000

600

Fig. I. Elution curve for a cellodextrin mixture. Numbers above the peak
indicate the degree of polymerization.

at a rate of 30 ml. per hour. The efiluent was collected in 10 ml. fractions
and the concentrations were measured in a Rayleigh interferometer. The
resulting curve is shown in Fig. i. The numbers above the peaks indicate
the degree of polymerization.

The material in each peak was identified by paper chromatography.
The cellohexaose, cellopentaose, cellotetraose and cellotriose peaks from
three almost identical experiments were pooled, dried by lyophilization
and crystallized twice from water-ethanol solvents. The molecular weight
was determined with the sodium borohydride-anthrone method of Peat
et al. [7] (Table I).

The specific optical rotations were difficult to measure with sufficient
accuracy for the higher cellodextrines because of the limited amounts
available and their low solubility. The results obtained (Table I) are of
the same order of magnitude as those given in the literature.

SEPARATION OF OLIGOSACCHARIDES WITH GEL FILTRATION 347

TABLE I

Oligosacchan

de

Molecular

weight

[-]t

Theor.

Found

Cellotriose
Cellotetraose
Cellopentaose
Cellohexaose

504
667
829
991

524
704
852
920

22-4

17-0

II-4

9-5

In collaboration with Dr. Bengt Nygard an X-ray diffraction analysis
of the cellodextrins was made in a Guinier camera with Cu K^ radiation,
and an exposure time of i hr. The films obtained were analyzed photo-
metrically and the values for sin'^^ calculated. In Table II the number of

TABLE II

Oligo-

Number of
reflections

Sin^^

saccharide

I

II

III

Cellobiose

Cellotriose

Cellotetraose

Cellopentaose

Cellohexaose

38
28
21

15
10

0-0308
0-0298
0-0296
0-0296
0-0298

0-0367
0-0361
0-0363
0-0367
0-0367

0 - 1 044
0- 1065
0- 1069
0-1075
o- 1082

definitely identified reflections and the sin-^ values are given for the three
strongest and most significant ones. The reflexions obtained from cello-
hexaose were diffuse indicating heterogeneity. Likewise the molecular
weight was lower than the theoretical value, whereas the tendency for the
other oligosaccharides was towards somewhat too high values.

Discussion

Preliminary experiments to separate cellodextrins were made with a
50-100 mesh sieve fraction of Sephadex. A high flow rate could be used
and an experiment in a 2500 ml. column was made in 7 hr. The analysis
of the eflluent fractions gave a smooth curve. Paper chromatography of the
material in the fractions showed, however, that the fractions contained
only two and sometimes three components.

The results were so encouraging that it was decided to try more
efficient columns in order to obtain a complete separation of the oligosac-
charides. Earlier investigations had shown that the best way to increase
the efficiency was to use smaller dextran gel particles, and consequently

348 PER FLODIN AND RARE ASPBERG

this approach was chosen. A sieve fraction passing through 200 mesh
proved satisfactory.

As seen from Fig. i the volume between the peaks is only between 50
and 100 ml. compared with a total bed volume of 2000 ml. Thus, it was
necessary to pack the column very carefully, but once packed it could be
used for a long time. For example, in three experiments made with the
same cellodextrin material the elution patterns were practically identical.
There was a slight variation in the positions of the maxima but the form
of the peaks and the resolution was the same. As seen in Fig. i the zones
overlapped slightly indicating that a complete separation was not obtained.
The amount of contamination in the peaks was so small that it was elimina-
ted in the crystallizations. This was evidently not the case for the cello-
hexaose which showed inhomogeneity by some of the criteria used.

The volume of the sample must not be larger than l\{Kj) — -/v^,)
where F, is the volume of water inside the gel grains and Kj^ and Kj) are
the distribution coefficients of the solutes to be separated. For the oligosac-
charides in the column used the sample volume must be less than 80 ml.
and in practice considerably less. On the other hand the process itself is
insensitive to the solute concentration if the viscosity is low. Thus, an
optimal amount is separated if the concentration is high and the volume
about one-fourth of the calculated value. To isolate a large amount of one
of the oligosaccharides a two-step procedure may be preferable in which a
preliminary rapid separation is made in a short column packed with, say,
the 50-100 mesh sieve fraction. A further fractionation is then made in a
high-efficiency column.

The outlined procedure is in principle applicable to any separation
within a series of homologues in which the molecular size difference is at
least as large as for the oligosaccharides. With ionized solutes, however,
it is necessary to take the effect of the charges into consideration. This
means that separations often have to be made in the presence of strong
electrolytes. High-efficiency separations may also be made in higher
molecular weight ranges. Dextrans on molecular weights up to about
40 000 have been fractionated with excellent results in dextran gels with
different degrees of cross-linking.

References

1. Dickey, E. E., and Wolfrom, M. h.,jf. Aniey. c/ietn. Soc. 71, 825 (1949).

2. Whittaker, D. R., Arch. Biochefti. Biophys. 53, 439 (1954).

3. Lathe, G. H., and Ruthven, C. R. J., Biochem. J. 62, 665 (1956).

4. Porath, J., and Flodin, P., Nature, Loud. 183, 1657 (1959).

5. Flodin, P., and Granath, K., /;/ "Symposium iiber Makromolekyle", Wies-
baden, Oct. 1959.

6. Flodin, P.,X Chromatography 5, 103 (1961).

7. Peat, S., Whelan, W. J., and Roberts, J. G.,y. chem. Soc. 2258 (1956).

SEPARATION OF OLIGOSACCHARIDES WITH GEL FILTRATION 349

Discussion

Hestrin : Would it be possible with the help of this method to separate linear
from branched oligosaccharides when they have the same molecular weight ?

Flodin : It might be possible. There is, however, no experimental evidence as
yet.

Smith: Does this maximum molecular weight hold for proteins also ?

Flodix : Roughly. The size is probably the determining factor in the separa-
tions and therefore the molecular weight, though more convenient to use, is an
inadequate measure.

Smith: If you have a molecule which is the size of, say, 100 A could you do
anything with that ?

Flodin : Molecules larger than about 50 000 in molecular weight can be
separated from smaller ones on the loosest gels. Even haemoglobin, however, is
somewhat retarded in relation to gamma globulin and serum albumin.

AUTHOR INDEX

Numbers in brackets are reference numbers and are included to assist in locating
references in which the authors' names are not mentioned in the text. Numbers in
italics indicate the page on which the reference is listed.

Acs, G., 265 [28], 267 [35], 27S, 298 [9],
307 [25, 29], J08

Adler, J., 104, 105, iii

Afzelius, B. A., 20J

Albers, M., 251 [24], 2§2

Albertsson, P. A., 33 [3], 34[3, sl. 35 [3, 6,
7,8,9, II], 36[3],37[3],38[3, 16, 18],
39[3, 16, 17, 18], 39, 315, 3-3

Allen, E. H., 305 [16], 308

Allen, F. W., 298 [10], 308

Allfrey, V. G., 197 [27], 203, 261 [2, 3, 4],
262[5, 6, 8, 9, II, 13], 263[i4, 16, 21],
264[2i], 265 [29], 266 [29, 34], 267 [29,
34], 268 [5, 13, 37, 44], 269 [44], 270 [5,
45], 271 [44, 46], 272 [29, 46], 274[4, 5,
8, 13, 44, 46], 275 [44, 49], 2yj, 2y8

Ames, B. N., 25o[2i], 252

Anand, N., 306 [24], 308

Anderegg, J. W., 95 W, 98 [4], ^01

Anderson, N. G., 250 [20], 2^2

Andreson, N., 166

Anfinsen, C. B., 56 [24], 3/, 6o[5], 65

Armitage, A. K., 306 [24], 30S

Armstrong, J. J., 289 [4], 204

Arrhenius, E., 225 [12], 226[i, 12, 13], -J5

Asford, 141 [40], 133

Ashikawa, J. K., i96[2], i97[2], i99[2],
202

Askonas, B. A., I99[3], 202, 2i2[2], 218

Auerbach, V. H., loi [21], loi

Austrian, R., 333 [43], 340

Avigad, G., 3 16 [4], J-'J

B

Baddilev, J., 289 [4, 5], 2g4

Bahr, G. F., 268 [42], 2j8

Balis, M. E., 197 [5, 143], 202, 206

Ball, E. G., i58[2], 166

Bargmann, W., i7o[5], 171 [5], rg4, 202

Barigozzi, C, 182, 194

Barnard, E. A., 56 [23], 57

Barnum, C. P., 196 [7], 202

Barrnett, R. J., 158 [2], 166

Barton, A. D., 197 [106, 107], 199 [106], 203

Bar\^sko, E., 163 [11], 16-

Baud, C. A., 206

Bauer, A., 278

Becker, W. A., 146 [48], 153

Beerman, W., 268 [42], 2j8

Behrens, M., 263 [20], 278

Beljanski, M., 284, 287

Bell, D., 95[io, 14], 96[i4], 97[io], 98[i4].

99 [14], 10 1
Belozersky, A. N., 80 [41], 92
Bennett, S., 158 [3], 163, 166
Berg, P., 265 [27], 267 [36], 278, 307 [30],

308
Bergmann, F. H., 267 [36], 278, 307 [30],

308
Bergstrand, A., 226 [14], 233, 261, 277
Bernardi, G., 69 [13], 9/
Bernhard, W., 202, 234 [2], 235, 278
Bernheimer, H. P., 333 [43], 340
Beskow, G., 197 [90], 204
Bessis, M., 199 [146], 202, 206
Bessman, M. J., i04[9], 105 [9], iii
Bharadwaj, T. P., 197 [114], 205
Bingelli, M. F., 142 [42], I53
Birbeck, M. S. C, 190, 194
Blanco, G., 73 [33], 92
Block, D. P., 196 [11], 202
Bluhm, M. M., 59 [3], 65
Blumenfeld, O. O., 59 [4], 60 [4], 61 [8], 65
Bodo, G., 6[2], 9, 59 [3], 65
Bojarski, T. B., 98I19], ^01
Bollum, F. J., 95 [2, 3, 4, 16], 98 [3, 4], loi,

104M, 105 [7], m
Boman, H. G., 298 [9], 307 [25], 308
Bonner, J., 222 [31], 235, 236, 247 [13],

248[i3], 251 [31], 252
Borgois, S., 298 [6], 306 [6], 308
Borsook, H., 133, 153
Boyer-Kawenoki, F., 33 [4], 39
Brachet, J., 131 [8], 152, I97[i2, 13], 202
Brandt, P. W., 158U], I59[i5], i64[22],

166, 167
Branson, H. R., 6[i], 9
Breitman, T., 262 [12], 266, 277, 278
Broberger, O., 213 [3, 17], 218, 2ig
Brockman, R. W., 283 [9], 287
Broman, L., 38 [18], 39 [18], 39
Brown, D. H., 332, 339
Brown, G. L., 27(20], 29 [20], 30(20], 31
Brown, H., 71 [18], 91
Bruni, C, 139E39]. 141 [39], ^53, 206
Buchanan, J. G., 289(4], -94
Burger, M., 330(10], 339
Burton, K., 73 (29], 92

352 AUTHOR INDEX

Butler, J. A. V., 2io[5], 218, 251 [33], 252
Buvat, R., 143 [44a], 153

Cabib, E., 292 [19], -'9-^

Callan, H. G., 268 [43], 278

Camara, S., 197 [154], 206

Campbell, P., i97[56], 202, 203, 2i2[4],

214U], 218, 255 [i], 258
Canellakis, E. S., 95, 98 [6, 7], loi, loi,

i04[8], 105 [8], III
Canellakis, Z. N., loi, loi
Carasso, N., 169 [2], IQ4
Cardini, C. E., 292 [19], 2g4
Carosso, N., 143 [44a], 153
Carss, B., 289(4], 297
Caspersson, T., 131 [9], 152, 182, 194
Castelfranco, P., 197 [15], 202
Caulfield, J. B., 146 [47], 153
Cavalieri, L. P., 69 [12], 91
Ceriotti, G., 229 [3], 235
Chaberek, S., 249 [19], 250 [19], 252
Champagne, M., 69 [13], 'ji
Chantrenne, H., 202, 281 [2, 3, 4, 5], 282

[7], 286, 287
Chao, F.-C, 247 [8], 252
Chapman, G. B., 204
Chapman-Andresen, C, 158 [9, 5, 6], 159,

160, i6b, 167
Chargaff, E., 67 [i], 68 [i, 2, 3, 7, 8, 9],

69[7. 10], 7o[3. 14- 15. if>i. 71 [i. 7,
19, 21, 22, 23, 24], 72[7, 25], 73 [26,
27, 28],74[7, 21, 35, 36], 75 [9, 21, 36],
75[i6], 76[2i]. 77[7. 21, 39], 78[37.
38], 79 [37], 80 [7, 24, 40]. 81 [42], 82
[42], 83 [i, 21, 35, 37, 40], 84 [1,46],
85[8,37,38,42],86[io,37],87[i9,S2],
88 [37,38, 42], 89[io, 37, 42, 50, 52],
90 [37, 42, 55], gi, ()2, 285 [19], 287

Chaudhuri, N. K., 290 [15, 16], 294

Chauveau, J., 197 [18, 120], 203, 205, 264,
278

Chesin, R. B., 197 [19, 20], 199, 203

Chevremont, M., 203

Christensen, A. K., 136 [28], 153

Christensen, H. N., 274[48], 278

Christian, W., 299 [13], 308

Chung, C. N., 251 [23], 252

CifonelH, J. A., 33o[i4], 332, 333 [40, 41,
45], 339, 340

Clark, S. L., Jr., 167

Clark, W. H., Jr., 199 [24], 203

Claude, A., i27[i], 131 [12, 13], 152, 153

Cochran, W., 15 [5], 3T

Cohen, E., 54[i6], 57

Cohen, S. S., 250 [22], 232, 306 [22], 308

Cohn, P., 2io[5], 218, 222 [4], 235, 251
[33], 252

Colvm, J., 319, 323

Cordes, E., 305 [16], 308

Corey, R. B., 6, 9, 20 [9], 31

Cornil, L., i82[i9, 20], IQ4

Cotte, G., 199 [25], 203

Craig, L. C, 33 [2], 39

Crane, R. K., 263 [18], 278

Crampton, C. P., 68[8, 9], 7o[i4], 75 [9],

85 [8], gi, i97[26], 203, 251 [32], 252
Crestfield, A. M., 298, 308
Crick, F. H. C, 20 [8], jr, 69 [11], 85 [47],

gi, 92
Cullis, A. P., 8 [4], 9
Cummins, C. S., 290, 294
Cusmano, L., 182 [16, 17], ig4

D

Daly, M. M., i97[27], 203 261 [2, 3],

263J16], 277, 278
D'Amelio, V., 209 [i, 15, 20], 2io[i, 20],

212 [i, 20], 2i4[2o], 218, 2ig
Davidson, E. A., 331, 339
Davidson, J. N., 95 [8, 10, 11, 12, 14], 96

[14], 97[8, 10], 98[ii, 14], 99[ii, 12,

14), loi, 203
Davies, D. R., 7b], 9, 15 [6], Ji, 59[3], 65
Davis, B. D., 298 [7], 306(20, 24], 308
Deane, H. W., 194
Deasy, C. L., 133 [16], 153
de Bertrand, G., 330, 339
de Duve, Chr., 162, 167

Dekker, C.A., 73[34], 9-'

de la Haba, G., 265(30], 266, 278

Dellepiane, G., 182 [15], 194

Delhveg, H., 8i[43], 9-'

de Paola, D., 206

Deutsch, J. P., 69(12], 91

De Vicentiis, M., 213(22], 219

Devreux, S., 281 (3, 4, 5], 282(7], 286, 287

Dewey, V. C, 283(14], 287

Dickerson, R. E., 7(3], 9, 15 ((6], 31

Dickey, E. E., 345, 348

Dickmann, S. R., 203

Dieckmann, M., 267(36], 278, 307(30],

308
Dmtzis, H., 6(2], 9, 59 [3], ^5, 251 (31], 252
Dobry, A., 33(4], 39
Doctor, P. B., 251 (25], 252
Dolley, D. H., 199(37], 203
Doniger, R., 71 (23], 91
Donohue, J., 21, 31
Dorfman, A., 330(14, 17], 331 (20, 21], 332,

333 [20, 41, 44, 45, 46], 334 [47, 48,

49], 335 [50], 339, 340
Doty, P., 197 [59], 204, 248(16], 252
Douglas, T. A., 203
Drever, W. J., 60(6], 65
Dubin, D. T., 250(21], 252
Dunn, D. B., 81(44], 87(49], 88(48], 89

[51], 92
Duthie, E. S., 201 [39], 203

Edlund, v., 203
Edsall, J.T., 68(4], 91
Edwards, R. G., 196(180], 207
Egami, P., 330(11], 339
Ekholm, R., 203
Eliasson, N. A., 261, 277

AUTHOR INDEX

353

Elson, D., 71 [19], 87[i9], 97, 212 [6], 218,

248 [15], 252
Enicore, M., 251 [23], 2^2
Entenman, C, 230, 2J5
Epstein, 137 [32]. ^53
Evans, A. H., 306 [21], 308
Everett, N. B., 197 [109], 20^

Farquhar, AI. G., 199, 203
Favard, P., 197 [43. 44]. ^94, -'03
Fawcett, D. W., 136 [28], 138 [36], 142 [43],

153
Feigelson, AI., 231, 233
Feigelson, P., 231, 255
Feingold, D. S., 316 W. 3^3
Feldman, AI., 212 [6], 21S
Felsenfeld, G., 251 [30], 2^2
Ferreira, David, i96[45, 46], i97[46], 203
Ferreira, Fernandes J , 197 [47], 203
Fessler, J. H., 329, 339
Ficq, A., 197 [48], 203
Finch, L. R., 306 [17], 30S
Fischer, G. A., iii, iii
Fitzgerald, P. J., 204

Flodin, P., 51 [12], 57, 345 U, SJ. 346, 34S
Folkes, J. P., 293 [23, 24], 2g4
Foster, J. F., 38 [13. 14. 15], 39
Fraenkel-Conrat, H., 53 [14], 57
Eraser, D., 251 [26], 252
Freed, J. J., 197 [49], ^03
Frenster, J. H., 265 [29], 266 [29], 267 [29,

37], 268[44], 269(44], 271 [29, 44],

274 [44], 275 [44], -'7-5
Frey-Wyssling, A., 322 [13], 323
Frick, G., 34[5], 35 [5, 7, 8, 10, 11], 39
Friedkin, AI., 262 [10], 277
Fullam, E., 127 [i], 152
Fuller, W., 21 [11], 31
Furuta, Y., 197 [194], 20/

G

Gale, E. F., 293(23, 24, 25], 294

Gall, J. G., 268 [41], 278

Gallop, R. C, 331, 339

Garden, S., 331 [25]. 332, 339

Garfinkel, D., 197 [50], 203

Gamier, Ch., 131 [10], 152

Garzo, T., i97[i9i], 20/

Gautier, A., 203

Gey, G. O., 163, 16/

Glaser, L., 318, 323

Glassman, E., 305 [16], 308

Globerson, A., 212 [6], 218

Godman, G. C, 153, i96[ii], 202

Goldfaber, L., 333 [44], 340

Goldstein, L., 197 [52, 53, 54, 55], 203

Gorini, L., 297[3], 3o6[i8,'23], 308

Granath, K., 345 [5], 34^

Gray, E. D., 95 [10, 12, 14], 96 [14], 97 [10],

98[i4], 99[i2, 14], JOi
Greathouse, G. A., 317, 323

Green, C, 71 [23], gi
Greenberg, G. R., 289 [4], 294
Greengard, O., 197 [56], 202, 203, 212 [4],
2i4[4], 218, 231 [9], 235, 255[i], 258
Greenstein, J. P., 194
Griboff, G., 81 [45], 92
Gricouroff, G., 182 [21], 194
Gromet, Z., 32o[io], 323
Gropp, A., 2y8

Gros, F., 285 [20], 28j, 297 [2], 308
Gross, J., 197 [112], 205
Grossman, B. J., 330, 339
Grumbach, AI., 332 [34], 339
Gr\cki, S., 197 [57], -'03
Gundlach, H. G., 47, 48, 49, 56, ^y
Gusberg, S. B., 182 [18], 185, 194

H

Haagen-Smit, A. J., 133 [16], 133
Hager, L. P., 298 [8], 299 [8], 308
Haguenau, F., 128 [5], 152, 169 [i], 171 [i],

i75[io, II, 12], i82[io, 11], i9o[i],

^94, 197 [58], 204
Halberg, F., 196 [7], 202
Hall, B. D., 197(59], -'04, 248 [16], 252
Hall, C. E., 247 [9], 252, 322, 3-'3
Hallen, A., 331, 339
Halvorson, H., 307 [27], 308
Ham, A. W., igo[23], 194
Hamers, R., 285 [18], 28y
Hamers-Casterman, C, 28s [18], 28y
Hamilton, L. D., 15 [3, 4],^i7[3, 4], i9[3,

4], 22 [3, 4, 16, 17]. 24[i8],26[3,4], J7
Hamilton, M. G., 197 [5, 60, 141-143], 202,

204, 206, 247 [14], 252
Hammarsten, E., 103 [2], iii, 261, 2J7,

298 [12], 308
Hammarsten, G., 298(12], 308
Hampton, J. C, i37[33], ^53
Hancock, R., 289 [6], 294
Hanzon, V., i97[6i, 182, 183], 204, 20/,

209 [7], -'^8, 221 [10], 233
Harbers, E., 290 [16], 294
Harris, H., 290, 294
Harris, J. I., 53 [14], 37, 60 [7], 65
Hart, R. G., 7[3], 0, i5[6], jx
Hartmann, G., 307 [25], 308
Hartmann, J. F., 127(2], 132
Hecht, L. I., 197 [62, 115], 199 [186], 204,

205, 207, 283 [10, 11], 257
Heidelberger, C, 290, 294
Heinrich, AI. R., 283 [14], 287
Hele, P., 306 [17], 308
Henseleit, K., 225 [18], 233
Herbst, E. J., 251 [25, 28], 232
Herman, L., 204

Hermodsson, L. H., 197 [61], 204 233
Hestrin, S., 3i6[4], 32o[io], 323
Hiatt, H. H., 98 [19], loi
Hill, J., 182 [18], 185 [18], 194
Himes, AI., 142 [41], 133
Hinman, R. L., 307 [27], 308
Hirakawa, K., 197 [196], 207

354 AUTHOR INDEX

Hirano, S., I97[i96], 20-] , 331, jjg

Hirs, C. H. W., 41 [i], 51 [11], 52[ii, 13],

54[i5, 17, 18, 19], 55[i], 56[i], 57,

59. 65
Hirsch, G. C, 195 [71, loi], 196 [64-72,

loi], 197 [66, 68-72, 100, loi, 162],

i98[72], i99[66, 71], 20o[67, 69, 71,

72], 204^ 205, 206
Hoagland, M. B., 90 [56], Q-', 197 [73, 74,

lis], 204, 203, 263, 265 [24], -V<S', 283

[10, 11], 28y
Hodes, M. E., 72 [2s], 73 [26], gr
Hodge, M. H., 204
Hodgkin, D. C, 16 [7], J/
Hoffman, P., 328 [3], 33o[i3], 33i[i3, 24],

332 [32], 33Q
Hofmann, M., 206

Hogeboom, G. H., i97[76, 166], 204, 206
Hokin, L. E., i97[77-8o], 1 99 [77-87], -O-/
Hokin, M. R., i97[78-8o], i99[78-87],

204
Hollcy, R. W., 265 [25], 278
Hollingsvvorth, B. R., 247 [11], 248 [i i], -',5-'
Hollmann, K. H., i7o[6], 171 [6], i9o[6],

J 94
Holter, H., i58[i3], i6o[i3, 7], i66[i2],

166, i6y
Holtzer, R. L., 197 [108], 203
Hoogsteen, K., 22 [14], 3 J
Hooper, C. W., 15 [3, 4], i7[3, 4], I9[3, 4],

22[3, 4], 24[i8], 26[3, 4]. J/
Hopkins, J. W., 263 [19], 265 [19, 29], 266

[19, 29, 34], 267 [29, 34], 271 [46], 272

[29, 46], 274 [46], 2y.S
Home, R. W., 290 [14a], 207
Horowitz, J., 87[s2], 89[s2], 9-', 28s [19],

287
Hotchkiss, R. D., 306 [21], JOcV
Howatson, A. F"., 190 [23], i()4
Huang, S., 251 [30], 252
Hultin, T., 133, 153, 197 [30, 88-90, 93,

156], 199 [32, 97, 156, 157], -'oj, 204,

205, 206, 209[lO, 20], 210 [20], 212
[20], 2 14 [10, 16, 20], 215 [10], 216 [10],
217 [10, 24], 218 [9], 218, 2iq, 222 [5,
6, 22, 26], 223 [5, 8], 225 [5, 12, 26],
226[l, 12, 13, 14], 227 [5], 228[7], 234

[5], -'J5
Hutchinson, D. J., 283 [9], 287
Huxley, H. E., 247 [10], 252

I

Imai, Y., 197 [196], 207
Ingram, V. M., 203
Inouye, A., 248 [17], 232

ItO, E., 289 [9], 2Q4

Iwashige, K., 207

J

Jackson, D. S., 330, 339
Jacobs, W. A., 73 [30], 92
Jaffe, J. J., 95 [7], 98 [7], lOT
Jardetzky, C. D., 196 [7], 202

Jefferson, H., 331 [20], 333 [20], 339

Jorpes, E. A^, 332, 339

Josephson, K., 286, 287

Josse, J., 13, 27[i], 30

Jungblut, P. W., 199 [99], 205

Junqueira, L. C. U., 195 [loi], 196 [68],

i97[47, 68, 100, loi, 154, 162], 20J,

204, 203, 20b

Kameyama, T., 212 [8], 214 [8], 218

Kaplan, D., 331, 339

Karlin, S. J., 137, 153

Karrer, H. E., 137 [34], i53

Kauzmann, W., 63 [12], 65

Kay, R. E., 230 [15], 233

Keighly, G., 133 [16], 133

Keir, H. M., 95 [8, 9, 10, 14], 96 [14], 97 [8,

10], 98[i4], 99[i4]- ^0/
Keister, D. L., 251 [28], 232
Keller, Y.. B., i97[io2, 112, 113], I99[i02],

20s, 222 [33], 23b, 265 [24], 278
Keller, P. J., 54[i6], 57
Kemp, A., 23o[i6], 233
Kendrew, J. C.,6[2], 7 [3], 9, i5[6], J-r, 59,

Kennaway, E. L., 299, 308

Kephart, J. E., 143 [44], ^53

Kerejkarto, B., 265 [31], 278

Kernot, B. A., 202, 212 [4], 2 14 [4], 218,

255[i], -5A'
Kidder, G. W., 283 [14], 287
Kirby, K. S., 265 [32], 278, 298, 308
Kirsch, J. F., I97[i03], I99[i03], -05
Kitai, R., 59[i], 71 [17, 18], 91 65,
Kits van Heijningen, A. J. M., 23o[i6], 233
Klenow, H., 1 1 1, ///
Klungsoyr, S., 320, 3-3
Knoop, A., i7o[5], 171 [5], i94
Korn, E. D., 333 [39], 339
Kornberg, A., 13, 27[i], Jo, 95[i], loi,

103, i04[i, 9], 105 [9], ^11
Korner, A., 197 [104], 203, 222 [17], 23s
Koshland, D. E., Jr., 49[io], 57, 3i6[3],

323
Kossel, A., 299, 308
Krakow, J. S., 9s [7], 98 [7], ^oi
Krebs, H. A., 225 [18], 235
Krone, W., 251 [24], 232
Kuff, E. L., 1 97 [76], 204
Kuhn, R., 331, 339
Kunitz, M., 48, S7
Kurochtina, T. P., 197 [105], 203

Lacks, S., 297 [2], 308

LaFontaine, J. G., 268 [40], 278

Laird, A. K., 197(106, 107], i99[io6], ^05

Lamborg, M. R., 283 [13], 287

Langridge,R., 15 [3, 4], 17, I9[3, 4], 22[3,

4], 24 [18, 19], 26[3, 4, 19], 3t
Lapresle, C, 2s6, 238
Lathe, G. H., 345(3], 34^

AUTHOR INDEX

Laurent, T. C, 328 [s], 339

Lavallee, R., 298 [6], 306 [6], 308

Leblond, C. P., 197 [109], J05

Lehman, L R., i04[9], 105 [9], ///

Leloir, L. F., 292 [19], ■^Q4

Leonis, J., 61 [8], 63

Leppelmann, J., 331, 33Q

Leuchtenberger, C, 197 [no], 203

Levene, P. A., 73(30, 32], gj

Levy, A. L., 45, 53 [14], 57

Lewis, W. H., 167

Lichtenstein, J., 2 so [22], -'5_'

Lif, T., 35, jg

Lima-de-Faria, A., 96 [18], loi

Limon, M., 170(4], IQ4

Linderstrom-Lang, K., 63 [13], 65

Linker, A., 328(3], 33o[i3], 33i[i3], 332,

339
Lipmann, F., i97[iii], 205, 263[i8], 265

[28], 267 [35], 278, 298 [9], 307 [25, 26,

29], 308
Lipshitz, R., 68 [8, 9], 70 [14]. 73 [28], 75 M.

78 [38], 85 [8, 38], 88 [38], 89 [50], 91,

9-'
Lipson, H., 15 [5], 31
Littlefield, J. W., i97[ii2, 113, 186], 203,

207
Low, H., 223 [8], 225 [12], 226[i2, 19], 233
Loewi, G., 336 [51], 340
Loftfield, R. B.,"i97[i02], i99[io2, 186],

205, 207
Long, M., 182 [18], 185 [18], 194
Lorente de No, R., 271 [47], 278
Lorincz, A. E., 331 [21], 332, 339
Love, R., 197 [114], 203
Lowe, C. U., 231 [20], 233
Lovvrey, G., 337 [52], 3 40
Lowy, P. H., 133 [16], 133
Lozaityte, I., 328, 339
Ludowieg, J., 33o[i4], 331 [20], 333[20,

44]. 339, 340

M

Maaloe, O., 297 [4], 308

Maas, W. K., 297 [3], 298 [6], 306 [6, 18, 19,

20], 308
McElya, A. B., 95 [4], 98 [4], lur
Machado, R., 143 [45, 46], 149 [45], 133
Magasanik, B., 297 [5], 303 [5]. 3^8
Magee, P. N., 225 [12], 226 [12, 13], 233
Mager, J., 251 [27], 232
Mahler, H. R., 251 [23, 26], -',5-'
Mahlon, B., i97[ii5], 203
Mandel, H. G., 281 [i], 283 [i], 286
Mandelstam, J., 289 [8, 10], 290 [8], 293

[8, 10], 294 [8], 2Q4
Mantsavinos, R., 95 [5, 6, 7, 17], 98 [6, 7],

707, 104 [8], 105 [8], III

Marbet, R., 330, 339
Markham, R., 89(53], 9-
Markovitz, A., 333 [4°]. 339
Marshall, A., 197 [116], 199, 203
Marshall, J. M., Jr., i59[i5], 160, 767
Martell, A. E., 249[i9], 25o[i9], 232

355

Martin, E. M., 197 [117], 203

Marvin, D. A., 15 [3], i7[3], i9[3], 22[3,
16, 17], 26 [3], J7

Masters, Y. F., 331, 339

Mathews, A., 131 [ii], 133

Mathews, M. B., 328 [2], 329, 330 [2],
331 [18, 20], 333 [20, 44, 45], 339, 340

Matthews, R. W., 283 [8], 287

Maurer, W., 203

Meister, A., 197 [15], 202

Mercer, E. H., 190, 194

Meyer, K., 328[i, 3], 33o[i3], 331 [13, 22,
24, 28], 332 [32, 34]. 336 [51], 339, 340

Michelson, A. M., 73 [34], 92

Micou, J., 197 [53. 54, 55], -'OJ

Mills, G.T., 333 [43], J^o

Mingioli, E. S., 298 [7], 308

Mirsky, A. E., 197(27], 203, 261 [2, 3], 262
[5, 6, 7, 8, 9, II, 13], 263[i4, 16, 21],
264[2i], 265 [29], 266[29, 34], 267[29,
34], 268 [5, 13, 37, 44], 269 [44], 270
[5. 45]. 271 [44, 46], 272 [29, 46], 274
[4, 5, 8, 13, 44, 46], 275 [44, 49], 277,
278

Misani, F., 71 [23], 9/

Mizen, N. A., i97[i4i], 206

Moldave, K., 197 [15], 202

Mole, R. H., 231, 233

Mollenhauer, H. H., 143 [44], T33

Monroy, A., 197 [124], 203

Montag, B. J., 290 [15], 294

Montagna, W., 197 [125], 203

Montague, M. D., 68 [6], 97

Moore, S., 41 [i, 2], 42, 43 W. 46 [4]. 47
[7], 48 [8], 49 [8], 51 [i I]. 52 [i I],
54[i7, 19]. 55[i. 2], 56[i, 8], 37, 59,
65

Morelle, J., 199(119], -'05

Morgan, W. S., 209[io, 20], 2io[2o], 21 1
[19], 2i2[2o], 2i4[io, 20], 2i5[io],
2i6[io], 2i7[io], 2i8[9], 218, 2ig,
222 [22], 233

Morrill, G. A., 203

Morris, N. R., in, 777

Morton, R. K., 68[6], 97, i97[ii7], 203

Moule, Y., i97[i8, 120], 203, 20^, zG^-lzz],
278

Miihlethaler, K., 322, 323

Muirhead, H., 8 [4], 9

Muller, M., 160, 767

Munger, B. L., i96[i2i, 122], 203

Munro, H. N., 203

N

Nakanishi, K., 330 [11], 339

Nakano, E., 197 [124], 203

Napolitano, L., 142 [43], 75J

Naora, H., 271 [46], 272 [46], 274(46], 278

Neidhardt, F. C, 297 [5], 303(5], 308

Neurath, H., 54(16], 37, 6o[6],'65

Nilsson, J. R., 160, 767

Noback, C. R., 197(125], 205

Nohara, H., 265(26], 275

Norberg, B., 261, 277

356 AUTHOR INDEX

Norrby, E., 35 [9]- 39

Norris, C. B., 250 [20], J5J

North, A. C. T.", 8 [4], 9

Northrop, J. H., 62, 65

Novelli, G. D., 2i2[8], 2i4[8], 218

Novikoff, A. B., 163, i6y

Nyns, E. J., 38[i6], 39[i6, 17], 39

O

Oberling, C, 203, 278

Ochoa, S., 284, 287

Ofengand, E. J., 265 [27], 267 [36], 2yS,

307 [30], 308
Ogata, K., 265 [26], 2/8
Ogston, A. G., 329 [6], 339
Ogura, M., i97[i96], 2oy
Oota, Y., 196 [126], 205
Oparin, A. I., 315 W- J-i
Oram, V., 197 [128], 205
Ord, N. G., 2io[ii], 218
Osawa, S., 262[8], 268[5], 27o[5, 45],

274 [5, 8], 277, 278
Ottenstein, B., 73 [31], 9-'
Ouchterlony, O., 2io[i2], 218

Palade, G. ¥.., i28[6], 132, 133 [14, 17, 18,
19]. 135. 136 [26], 137 [6], 1 38 [6], J5-\
153, 163, 167, i7i[7], 175 [7], 194,
i97[io3, 173, 174, 176, 177], i99[i03,
129-139], 171-174, 176, 177, 20^^,206,

207, 209[l3, 14], 213 [27], 218, 2H),

222 [23], 2J5, 239 [i, 2, 3, 4, 5, 6, 7],

24o[5], 241 [7], 242 [7], 245 [7], 248 [7],

249 [6, 7], 250(7], 2^1, 252
Palay, S. L., I36[27], 137, 153, 163, 167,

1 97 [140], 206
Pappas, G. D., i59[2i], i64[22], 167
Pardee, A. B., io4[6], iir, 307(28], 308
Park, J. T., 289, 290(1 1, 12], 292, 293, 2(j4
Parks, J. M., 56, 57
Parks, R. E., 283(14], 287
Parrish, R. G., 59(3], 65
Paul, J., 95[ii. 13]. 96[i3]. 98[ii, 13].

99[ii, 13]. loi, i58[23], 167
Paulauskaite, K. P., 197(20], 199 [20], 203
Pauling, L., 6, 9, 20(9], 22 [15], 31
Peat, S., 346, 34S
Pecora, P., 197 [143], 206
Perkins, H. R., 289[i], 29o[i3, 17], 292

[20], 293 [20], 2()4

Perlmann, G. E., 59[4], 6o[4], 61 [8, 9],

63 [11], 65
Perlmann, P., 209[i, 15, 20], 2io[i, 20],

2ii[i9],2i2[i,i9, 20], 213 [3, 17, 18],

214(10, 16, 20], 2i5[io], 2i6[io],

217(10], 2i8[9], 218, 21Q
Perutz, M. F., 8, 9
Petermann, M. L., 197 [5, 26, 60, 141-143],

202, 203, 206, 247(14], 251 [32], 252

Peters, T., Jr., I97[i44], 206, 2i2[2i, 22],
213(22], 2 14 [2 1, 22], 215 [21, 22], 2ig,
255, ^58

Petersen, G. B., 73(29], 92

Peterson, E. A., 2io[23], 2ig, 30o[i5], 308

Peterson, M. L., 137 [35], 153

Petrschkaite, S. K., 197 [20], 199(20], 203

Philipson, L., 3 s [7, 8], J9

Phillips, D. C, 7 [3], 9, 15 [6], 31, 59 [3], 65

Pickworth, J., 16 [7], 31

Pinet, J., 163, 167

Pipan, N., 196 [145], 206

Plaut, W., 197(52], 203

Polevoy, I. S., 204

Policard, A., 199(146], 206

Pollister, A. W., 142(41], 133, 197(148],
206

Porath, J., 51 (12], 57, 345(4]. 34^

Porter, K. R., 127(1], 128(4, 7], 136(24,
25, 26], 138(38], 139(39], 141 [39, 40],
143(24, 45, 46], 146(47]. 149(45], -fi-',
153, rg4, 199(131, 1 49-1 51], 205, 206,
268(38], 278

Porter, R. R., 256, 258

Potter, V. R., 95, 98(3, 4], loi, loi

Preiss, J., 267, 278, 307(30], 308

Prescott, D. M., 167

Prestidge, L. S., 307(28], 308

Prosen, R. J., 16(7], 31

Putnam, F. W., 38(15], 3g

Quastler, H., 206

Q

R

Rabinovitch, M., 197(154], 206
Rand, R. N., 231 [20], 235
Rappay, G., 160, 167
Rashevsky, N., 206
Rauber, A., 170(3], i<)4

Ray, W. J., Jr., 49(10], 57

Reichard, P., loi, 101 , 103 (2, 4, 5], 107(4],

1 10(5], ///, 261, 277
Reinits, K. (j., 329, JJ9
Rendi, R., 197(156], 199(156, 157], 206,

217(24], 2/9, 221(24, 25], 222(26],

225 (26], 2.75, 236
Revel, J. P., 142, 153
Rhodin, J., 206
Rich, A., 13(2], 24(19], 26(19], 30(2, 22],

JO, 31
Richards, F. M., 56, 57
Richards, J., 95(10, 14], 96(14], 97(10],

98(14], 99(14]. ^07
Riggs, T. R., 274(48]. 278
Riviere, M. R., 194
Roberts, J. G., 346(7], 348
Robertson, J. D., 197(159-161], 206
Robertson, J. H., 16(7], 31
Robineaux, R., 163, 167
Roden, L., 332(36], 339
Rogers, H. J., 289(1, 8, 10], 290(8, 13, 17],

292(20], 293(8, 10, 20], 294(8], 2g4

Rose, G. G., 162, 167

Rose, I. A., 103 [3], III

Rosenberg, B. H., 6g[i2], qi

Ross, 153

Rossmann, M. G. 8 [4], 9

Roth, L. E., 160, i6-j

Rothschild, H. A., i96[68], i97[68, 100,

154, 162], 204, 205, 206
Rouiller, C, i97[i8, 120], 203, 205, 235 [2],

235, 262 [22], 278
Runnstrom, J., 206
Rupley, J. A., 60 [6], 65
Ruska, H., 206

Rutberg, L., 103 [4], 107 [4], m
Ruthven, C. R. J., 345 b], 34^
Ryle, A. P., 59[i], 65, 71 [17], gi

S

Sachs, H., 263 [is], 274[is], 2j8

Sadron, C. L., 68[5]- 69[i3], 9i

Saetren, H., 263 [21], 264[2i], 2jS

Salton, M. R. J., 289[3], 29o[i4a], 21/4

Saluste, E., 103 [2], ///

Samachson, J., 330 [10], ijy

Samarth, K. D., 197 [5, 143], 202, 206

Sanger, F., 59, 65, 71 [17, 18], 91

Santesson, L., 182, ig4

Sampson, P., 332 [32], 339

Sato, A., 197 [194, 196], 2oy

Schachman, H. K., 247 [8], 232

Schapot, V. C, 197 [165], 206

Schellman, J. A., 63 [13], 65

Schiller, S., 331, 332 [35], 333 [20, 44, 45,
46], 334 [47, 48, 49]- 335 [50], 339, 34o

Schlessinger, D., 247 [11], 248 [11], 232

Schneider, W. C, 197 [76, 166], 204, 206

Schrader, F., 197 [i 10], 20^

Schramm, G., 265 [31], -'7<S', 320 [10], 323

Schucher, R., 206

Schulman, H. M., 70 [15], 91

Schulz, H., 206

Schumaker, V. N., 159, i6y

Schwartz, J., 306 [19], 308

Schwarz, W., 206

Schweet, R. S., 305 [16], 308

Schweigert, B. S., 103 [3], m

Scott, J. F., 197 [115], 203, 283 [10, 11], 28-

Seeds, W. E., 15 [3], i7[3l, i9[3], 22[3],
24[i8], 26 [3], JJ

Sela, M., 60 [5], 65

Selby, C. C, 222 [27I, 235

Sesso, S. R., i96[68], i97[is4, 170], 204,
206

Shapiro, H. S., 7o[i5], 71 [21], 73[28],
74[2i, 35, 36], 75[2i, 36], 76[2i],
77[2i, 39], 78[37]. 79[37]. 81 [42],
82 [21, 42], 83 [21, 35, 37], 85 [37, 42],

86 [37], 88 [37, 42], 89 [37, 42], 9o[37,

42], 9-r, 02
Shapras, P., 163 [11], 16/
Sharon, X., 307 [26], 308
Sherman, T. F., 329 [6], 339
Shetlar, M. R., 331, JJQ
Shigeura, H. T., 90 [55], 92

AUTHOR INDEX 357

Shinagawa, Y., 197 [197], 20J

Shore, V. C, 7 [3], 9, 15 [6], 3i, 59 b]. 65,

307 [28], 308
Siekevitz, P., 132, 133, 135 [21, 22], 133,

163, 167, 171 [7], 175, 194, I97[i03,

173-177]. I99[i03, 134. 171-177].
205, 206, 207, 209 [14], 212 [25], 213
[27], 21Q, 222(28, 29], 236, 239 [i, 2, 3,
4. 5. 6, 7], 24o[5], 241 [7], 242[7],
245 [7]. 248 [7], 249 [6, 7], 250 [7],
-'5^, 252

Signer, R., 33 [i], 39

Simkin, J. L., 197 [178], 199 [3], 202, 207

Simmons, B., 197 [109], 205

Simms, E. S., I04[9], 105 [9], iii

Simpson, R. B., 63 [12], 65

Simson, P., 251 [33], 252

Sinsheimer, R. L., 89(54], 9-

Sirlin, J. L., I96[i79, 180], i97[i8i], 207

Sjostrand, F. S., 197 [182], 199 [183], 207

Skipper, H. E., 283(9], 287

Slayter, H. S., 247(9], 252

Slovik, N., 330 [10], 339

Sluiter, J. W., 197 [185], 199, 207

Smellie, R. M. S., 95(8, 9, 10, 11, 12, 13,
14], 96[i3. 14]. 97[8, 10], 98[ii, 13.
14], 99[ii, 12, 13, 14], ^'^^'^

Smith, E. B., 333, J70

Smith, H., 331, 339

Smith, J. D., 81 [44], 87(49], 88(48], 89(51,

53]. 92
Smith, K. C, 298(10], 308
Smith, L. F., 59(1]. 65, 71(17]. 91
Sobel, A. E., 330(10], 339
Sober, H. A., 210(23], -^9
Sober, H. E., 300(15], 308
Spackman, D. H., 41(2], 42(3], 43, 46,

55(2], 57
Spahr, P. F., 92
Sparks, R. A., 16(7], j/
Spencer, J. H., 84(46], ()2
Spencer, M., 19(10], 2o[io], 21(10], 22

(10, 16, 17], 31
Spiegelman, S., 234(30], Jjj, 307(27], 3o8
Spirin, A. S., 80(41], 92
Spooner, E. T. C, 199(4], -'^'~
Stacey, M., 317, 323
Stahl, A., 182(19, 20], 194
Stein, D. W., 56(23], 57
Stein, W. H., 41(1, 2], 42(3], 43(4], 45,

46(4], 47(7]. 48(5. 8], 49(8], 51(11].

52(11], 54(17. 19]. 55(1, 2], 56(1. 8],

57, 59. 65
Stephenson, M. L., 197(18, 62, 115], 199

[186], 204, 203, 207, 283(10, 11], 287
Sterman, AI., 38(14], JO
Stern, H., 262(7], 263(21], 264(21], 277,

278,
Stich, H., 207

Strandberg, B. E., 7(3], 9, 15(6], jr
Straub, F. B., 197(188], 207
Straus, W., 162, 167
Strocker, B. A. D., 199(4]. 202
Strominger, J. L., 289(2, 9], 290, 291, 294,

331. 339

358 AUTHOR INDEX

Strugger, S., 2oj

Suzuki, S., 331, JJ9

Swift, H., 20-]

Szabd, M. T., ig7[i9i], 2o-j

W

Tabor, C. W., 251 [29], 2^2

Taborsky, G., 56, ^"j

Takahashi, N., 330 [11], JJ9

Takata, Kenzo, 196 [126], 20^

Tamaki, M., 2o-j

Tamm, C, 73 [26, 27, 28], <)i

Tamm, I., 293 [21, 22], 2()^

Tandler, C. J., 20-]

Tanford, C, 38 [12], jg

Tashiro, Y., I97[i94-i96], -'07, 248[i7],

-'5-'
Tavel, P., 33[i], J9
Thannhauser, S. J., 73 [31, 33], 9-
Theorell, T., 298 [12], 30^
Thompson, E. O. P., 59 [i], 65
Thompson, R. H. S., 2io[ii], 218
Thomson, R. Y., 95 [11, 12], 98 [11], 99

[11, 12], TOl

Threnn, R. H., 289 [2, 9!, 29^
Tissieres, A., 247 [11], 248 [11], 2=^2

Todd, A. R., 73 [34], 9-'

Tolinschis, L. E., 197 [20], 199(20], 203

Toschi, G., 197 [61], 204, 209(7, 28], -'/<S',

-'•^9, ^35
Tovell, H. M. M., 182 [18], 185 [18], i<)4
Towell, A. M. M., i82[22], i8s[22], i<)4
Trueblood, K. N., i6[7], 21 [13], j/
Trupin, K., 203
Ts'O, P. O. P., i97[iy7], -"7, 247 [13],

222 [31], 23b, 248[i3], 251 [31], 2S2
Tsumita, T., 8o[4o], 83 [40], 9-
Tuppy, H., 59 [i]. 65
Turba, F., 199 [99], 20^

Wacker, A., 81 [43], 92

Waddington, C. H., i96[i79], 2oj

Warburg, O., 299 [13], 308

Watson, J. D., 20, 30(22], 31, 69[ii],

85 [47]- 9J, 9-^ 247 [11], 248 [11], 2S2
Watson, M. L., 127(3], ^5-
Weaver, R. A., 251 [28], 2^2
Webster, G. C, i99[202], 20-], 262[i2],

264, 266, 2-j-j, 2j8, 283 [12], 287
Weiss, J. M., 135, 153, 197 [203], 20/
Weiss, M., 204
Weiss, S., 265 [28], 278
Weissman, S. M., 95 [11, 12, 13, 14, 15],

96 [13, 14], 98 [11, 13, 14], 99[ii, 12,

13, 14], ioo[i5], loi
W'eUings, S. R., 199, 203
Wesslen, T., 35 [7], 39
Weygand, F., 81 [43], 9-
Whalev, W. G., 143 [44], 153
Whelan, W. J., 346 [7], 34^
Whittaker, D. R., 345(2], 346, 348
Wiame, J., 298 [6], 306 [6], 308
Wieland, T., 223 [32], 2j6
Wilkins, M. H. P., 15 [3, 4], i7[3. 4],

i9[3, 4], 22 [3, 4, 16, 17], 24[i8], 26

[3. 4.1 3o[2i], J/
Will, G., 8 [4], 9
Wilson, H. R., i5[3, 4], i7[3, 4], i9[3, 4],

22 [3, 4] 24[i8], 26[3, 4], jr
Wilson, J. W., 2oy
Winterstein, A., 330, 33')
Wochner, D., 159, 16 j
Wolfrom, M. L', 345, J^cS'
Wood, H., 262 [10], 2JJ
Woodhall, B., 331, JJ9
Work, T. S., 197 [178], 199 [3], -02, 207
Wvatt, G. R., 71 [20], 77 [20], gi
Wyckoff, H. W., 6[2l, 9, 59(3], 65

Ullmann, A., 207

U

V

Valeri, V., 197 [154, 170], 206
van Lancker, J. L., 197 [108], 20^
Vendrely, C, 197 [200], 207
Vendrely, R., 197 [200], 207
Venkataraman, P. R., 231 [20], 23^
Vinograd, J., 222[3i], 236, 247[i3], 248

[13], -\5-'
Vischer, E., 71 [23, 24], 80 [24], 9/
von der Decken, A., 197 [30], 199 [32], 203,

204, 222 [5, 6, 22], 223(5, 8], 225(5],

227(5], 228 [7], 234(2], 2'j5
von Euler, H., 286, 287
von Ubisch, H., 261, 277
Vorbrodt, A., 197(201], 207

Yamada, E., i36(2s], 138(38], 153
Yang, J. T., 38(13], J'V
Yarmolinsky, M. D., 265(30], 266, 278
Yeas, M., 207

Zachau, H. G., 267, 278, 307(29], 308

Zalockar, M., 197(206], 207

Zamecnik, P. C, 197(18, 62, 102, 112, 115],
199(103, 186], 204, 205, 207, 222(33],
236, 265(24], 27S, 283(10, II, 14],
285(17], 287, 297, 308

Zamenhof, S., 71(24], 80(24], 81(45], 9^,

9-'
Zillig, W., 251, 252
Zubay, G., 27(20], 29(21], 30(20, 21], 31,

247(10], 249(18], 232

SUBJECT INDEX

A

Acceptor-RNA (see s-RXA)

Acefobacter xylinum, cellulose synthesis

by, 317-322

N-Acetylhexosamine, compounds in muco-
peptide, 290-293

Adenosine monophosphate (AMP), incor-
poration into s-RNA, 114, 119

Adipose tissue, pinocytosis in, 158

Alanine, incorporation into mucopeptide,
290-292
incorporation into proteins of thymus
nuclei, 270-274, 276

Albumin (serum), formation in micro-
somes in rat liver, 255-258

Alloxan diabetes, effect on mucopolysac-
charide metabolism, 344, 337

Amino acid activation, in cell nuclei,
263-265

Amino acid incorporation, eflFect of histones
on, 275-276
into ribosomes of cell nuclei, 268-269

Amino acid transfer, to RNA in cell nuclei,
265-267

Amino acid transport to nucleus, 270-274

Amino acid uptake, by isolated cell nuclei,
261-263

2-Aminofluorene, effect on liver glycogen,
230-231
effect on protein synthesis, 226, 229—

233
2-Aminonaphthalene, effect on protein

synthesis, 229-231
Amoeba, defaecation in, 165

pinocytosis in, 157-161, 163-166
Amylase, release from RNP particles,

241-242
Antigens, incorporation of isotopes into,
213-218
in microsomes, 209-218
transitory antigens of microsomes,
212-213. 218
Apurinic acid, 73
Arginine, incorporation into acceptor-

RNA, 297-307
Arginine-activating enzyme, 305-307

effect of guanidino deri\atives on, 301-

3°3 . .

levels during different growth conditions,

300-301
purification from Escherichia 299-300
Arginine-RNA, 300-301, 306-307
assay of, 298-299
properties of, 303—305
8-Azaguanine, elTect on protein svnthesis,
281-286

B

Bacillus ceretis, DNA synthesis in, 281

protein synthesis in, 281-286
Benzimidazole compounds, effects on cell

wall mucopeptide synthesis, 293-294
"Bunched" nucleotides, definition of, 71

Calf thymus, DNA biosynthesis in, 95
DNA from, isolation of, 69-70
nucleotide arrangement in, 77, 89

Caspersson's and Santesson's cells, "A",
182-190, 194
"B", 181-188, 189-190

Catalase, svnthesis in Bacillus cereus,
283-286

Cellodextrin (Sephadex), in gel filtration,
345-348

Cellulose, extracellular synthesis of,
317-322

Cellulose "Ultrastrand", 322

Ceruloplasmin, partition of, 37-39

Chicken embryos, enzymes for DNA syn-
thesis from, 103-110

Chondroitin sulphuric acids (A, B and C),
327-322

Chymotrypsin, eflFect of urea and guanidine
hydrochloride on, 60

Cluster analysis, 83-84

Coenzymes, synthesis of, 104

Cortisone, and amino acid incorporation
in liver systems, 229-233
effect on turnover of skin mucopoly-
saccharides, 333-334

Counter-current distribution, 33, 38, 39

Cytidine monophosphate (C]\IP), incor-
poration into DNA, 104-109
incorporation into s-RNA, 11 7-1 18, 120

Cytidine triphosphate (CTP), incorpora-
tion into s-RNA, 11 8-1 19

D

Deoxy cytidine monophosphate (deoxy-
CMP), incorporation into DNA, 104,
106, 108

Deoxyguanosine monophosphate (deoxy-
GMP), incorporation into DNA, 104,
109

Deoxyribonuclease (DNA-ase), micro-
coccal, 105—106
pancreatic, 72-73, 105

Deoxyribonucleic acid (DNA), acid de-
gradation of, 73-75

360 SUBJECT INDEX

Deoxyribonucleic acid (DNA) — continued
and nuclear amino acid incorporation,

262-263
biosynthesis of, 95-101, 281
primer molecule in, 95, 97
cluster analysis of, 83-84
degradation by deoxyribonuclease, 72-73
differential distribution analysis of, 73-82
Donohue pairs for, 22-24, 27
effect, on isolated nuclear ribosomes,

274-276
enzymic synthesis, from ribonucleotides,
103-111

primer molecule in, 104

with enzymes from chick embryos,

103-1 10
with enzymes from Eschcricliia coli,
103
enzymic degradation of, 105-106
Fourier synthesis for, 16-19, 26-29
homogeneity of samples, 68-69
homotopy in, 87-89
Hoogsten base-pairs for, 24-27
nucleotide arrangement in,

from bo\ine tubercle bacilli, 79-80
from calf thymus, 77, 89
from E. coli, 80-82, 90
from man, 75-77
from ox, 75-77
from rye germ, 77-79, 90
from sea urchins, 75-77
from wheat germ, 77-78, 89
nucleotide sequence of, 67-90
"parasitic", 84
pleromerism in, 84-89
"replacement" in, 84-85
Watson-Crick scheme for, 20-22, 24,

26-27
X-ray analysis of, 13-15
Dextran, biosynthesis of, 315-317
Dextransucrase, 316
Diesterase, from snake venom, 105

from spleen, 105-106
Differential distribution analysis, of DNA,

73-S2
Dimethylnitrosamine, effect on protein

synthesis, 226, 233
Donohue base-pairs, 20-24, 27

E

Ehrlich ascites carcinoma cells, DNA

synthesis in, 95, 97, 99, 100
^-Elimination reactions, 74
Endoplasmic reticulum, and pinocytosis,
161, 163-166
and the Golgi component, 135-136, 169,

171, 175, 195, 198-202
definition of, 169
forms and functions of, 127-152

granular or rough form, 1 29-1 31
smooth form, 135-138
glucose-6-phosphatase activity of, 133—

142
in plant cells, 142 152

Epinephrine, 141
Ergastoplasm, definitition of, 169
function and form of 128, 131, 135
in mammary gland and its tumours,
170-194
in pancreatic exocrine cell, 198, 200-201
"organized" ergastoplasm, definition of,
169
Escherichia coli, acceptor-RNA of, 297-307
DNA biosynthesis in, 95
DNA from, isolation from protoplasts,
69-70

nucleotide arrangement of, 80-82, 90
enzymes for DNA synthesis from, 103
ribosomal RNA of, structure of, 30
precipitation of, 212

Fat, absorption of, 136-138, 163
P^luorodinitrobenzene, 42-56
5-Fluorouracil, effect on biosynthesis of

mucopeptides, 290-293
FIuorouridine-diphospho-N-acetylglucos-

amine, in mucopeptides, 292
Fourier synthesis, 6-8
principles of, 15—16 ■
for DNA, 16-19, 26-29
Fructose, incorporation into cellulose,

320-321

G

Gel filtration, for oligosaccharide separa-
tion, 345-348

Glucagon, 141-142

Glucose, cellulose synthesis from, 318-322

Glucose-6-phosphatase, 133, 142

Glucose- 1 -phosphate, in cellulose synthe-
sis, 320-321

Glutamic acid, incorporation into muco-
peptide, 290-292

Glycine, incorporation into mucopeptide,
290-294

Glycogen, effect of 2-aminofluorene on
liver content, 230-231
effect of X-rays on liver content, 230-231

Glycogen metabolism, and the endoplasmic
reticulum, 138-142

Golgi component, and the endoplasmic
reticulum, 135-136, 169, 171, 175, 195,
198-202

Guanosine, and effects of azaguanine on
protein synthesis, 281-286

Guanosine monophosphate (GMP), incor-
poration into DNA, 104, 109-110

Guinea pig, pancreatic RNP particles of,
effect of spermine on, 240-251

H

Haemoglobin, X-ray studies of, 8
He La cells, pinocytosis in, 158, 162
a-Helix, 6-8, 59, 63

SUBJECT INDEX

361

Heparin, 330, 332-338
Heparin monosulphuric acid, 331-332
Histidine, dinitrophenyl derivatives of, 44
Histones, effect on amino acid uptake by

nuclear ribosomes, 275-276
Homotope, definition of, 71
Homotopy, in DNA, 86-89
Hoogsten scheme, 21-22, 24, 26, 27
Hyaluronic acid, 328-329, 332-333
Hydrocortisone, effect on turnover of skin

mucopolysaccharide, 333-334

I

Insulin, effect on concentration of acid
mucopolysaccharides in connective
tissue, 337

Isoleucine, incorporation into liver antigens,
214-281

Keratosulphate, 331-332

L

Lens antigen, 213

Leucine, incorporation into li\er antigens,

214-218
incorporation into protein, 223-224,

226-229, 231-233, 255-256, 265-268,

275-276
Leucine-activating enzyme, 300—301
Levan, biosynthesis of, 315-317
Levansucrase, 316-317
Liver, structure of RNA from, 30
Lysine, dinitrophenyl derivatives of, 43-44,

. 47 . .

incorporation into mucopeptides,

290-294
incorporation into protein, 224, 262
Lysosomes, and pinocytosis, 1 61-163

M

Mammary gland, normal, ergastoplasm of,
170-175
cancerous, ergastoplasm of, 175-194
3-Methyl-4-dimethylaminoazobenzene,

effect on endoplasmic reticulum,
139-140, 142
effect on glycogen metabolism, 139-140,
'4-2
Micro-kinetospheres, 162
Micropinocytosis, 160-162
Microsomes,

form and function of, 1 31-135
immunological studies of, 209-218
in incorporation systems, 222-235
serum albumin synthesis by, correlation
with structure, 255-258
Mitochondria, and pinocytosis, 163
Mono-dinitrophenylaminoribonuclease A,
degradation of, 53-55
isolation of, 51-53

Mouse, leukaemic cells, DNA synthesis in,

95
subcutaneous fibroblasts, DNA synthesis
in, 98-99
Alucopeptides,

biosynthesis of, 289-294

effect of benzimidazoles on, 293-294
effect of purine and pyrimidine analo-
gues on, 290-291
Alucopolysaccharides,
acid, 327-338

effect of age on concentration of,

335-337
effect of insulin on concentration of,

337
effect of thyroid on concentration of,

3377338
quantitative analysis of, 334-336
turnover of, 333

effect of hormones on, 333-334
Myoglobin, amino acid sequence in, 8
X-ray analysis of, 6-7, 15

N

Nernst partition law, 39
Nuclei, protein synthesis in, 261-277
Nucleic acids (see RNA and DNA), of
cancer cells, 182, 185
partition of, 34-36

O

Oligosaccharides, separation by gel filtra-
tion, 345-348

Pancreas,

exocrine cell,

enzyme synthesis in, 197-201.
RNP particles of,

effect of spermine on, 251
structure of, 247-25 i
Penicillinase, svnthesis in Bacillus cereiis,

282-286
Pepsin, amino acid composition of, 59-60
enzymic activity,

effect of guanidine hydrochloride on,

60-64
effect of urea on, 60-64
relation of secondary structure to,
59-65
Phagocytosis, 166
Phagosomes, 163
Phalloidin, effect on microsome systems,

223-226, 228-229, 232
Phosphodiesterase, from snake venom,

121-122
Phycoerythrin, partition of, 38
Pilocarpine, effect on pancreas cells,
200-201

362

SUBJECT INDEX

Pinocytosis, 157-166

definition of, i 57

inverted pinocytosis, 164-165
Pinocytotic vesicles, and fat absorption,

137.. 163

and mitochondria, 163
Pleromer, definition of, 71
Pleromerism, in DNA, 84-89
Polymer two-phase systems, 33-39
Polymerase, DNA, 96-101, 106
Proteins, partition of, 34-39

structure and size of, 68, 72

R

Rabbit tissues, DNA synthesis in, 96-98
Rat, adrenalectomized, 231

intestinal fat absorption, 136-138
liver, antifjens of, 213—218
DNA synthesis in, 98-100
fractionation of s-RNA from, 1 13-1 16
"Retaining" cells, 190
Ribonuclease, amino acid composition of,

53
dinitrophenylation of, 42-56

effect on enzymic activity 47-51
histidine residues in, 44-45
lysine residues in, 43, 45, 47
release from RNP particles, 241
serine residues in, 47
tryrosine residues in, 44-45
Ribonucleic acid (RNA),

incorporation of azaguanine into, 281-286
molecular configuration of, 29-30
nucleotide sequence in, from rat liver, 90
pleromeric relationships in, 87, 89
replacement in, 84-85
ribosomal, structure of, 30
soluble (acceptor-RNA, s-RNA),

eflFect of phosphodiesterase on, 1 21-122

fractionation of, 11 3-1 16

fraction of ribonucleotide-incorporat-

ing enzyme, 11 3-1 16
incorporation of AMP into, 114, 119
incorporation of arginine into, 297-307
incorporation of CMP into, 117-118,

121
incorporation of threonine into 122-

123
pyrophosphorolysis of, 1 17-120
synthesis of, 1 13-123
synthesis of, 104, 262
Ribonucleoproteins (see Ribosomes, and
RNP particles),
from rat liver,

fractionation of, 113-116
S-RNA-/3 and -y, analytical data on,
1 1 6-1 17

effect of phosphodiesterase on,
121-122
Ribonucleoprotein particles (Ribosomes,
RNP particles),
amino acid incorporation into, 268-269
and amino acid acti\ation, 264-265
and serum albumin synthesis, 256-258

binding of spermine to, 244-247

definition of, 169

in cancer of mammary gland, 176-177,
179-180, 185, 187, 190-193

in incorporation systems, 222-235

in normal mammary gland, 170-175

in pancreatic exocrine cell, 197—199

in plant cells, 143-144

pancreatic, effect of spermine on, 240-251
structure of, 247-251

release of amylase from, 241-242

release of protein from, effect of Mg++
on, 243-244

effect of spermine on, 243-244

release of RNA from, effect of Mg++ on,
242-244

effect of spermine on, 241—244

release of spermine from, effect of guano-
sine triphosphate on, 246-247
effect of Mg++ on, 246-247
Ribosomes (see RNP particles)

S

Sea urchin eggs, amino acid-activating

enzymes in, 226—227
"Solitary" nucleotides, definition of, 71
Soluble ribonucleic acid (s-RNA, see RNA)
Spermine, effect on ribonucleoprotein

particles, 239-251
Staf>/iylococciis aureus, mucopeptide syn-
thesis in, 289-294

Threonine, incorporation into soluble
RNA, 122-123

Thymidine, incorporation into DNA,
96-99, 262

Thymidylate kinase system in DNA syn-
thesis, 96-101

Thyroid hormone, effect on mucopoly-
saccharide metabolism, 334, 337-338

Tobacco mosaic virus, arrangement of
structural sub-units, 8
structure of RNA from, 30

Transitional "T" cells, 185, 190

Trypsin, effect of urea and guanidine hydro-
chloride on, 60

Tyrosine, dinitrophenyl derivati\es of, 44

U

Uridine diphosphate glucose (UDPG), in
cellulose synthesis, 318-320

Uridylic acid (UMP), incorporation into
soluble RNA, 120

Valine, incorporation into liver antigens,
214-218
incorporation into protein, 225

SUBJECT INDEX 363

Viruses, and aetiology of tumours, 190, 192 effect on liver glycogen, 230-231

arrangemept of subunits of, 8 effect on liver RNA content, 23 1

partition of, 34-35

Vitamin Bj,, X-rav analvsis of, 16

Y

W

Yeasts, structure of RNA from, 30

Watson-Crick scheme, 19-22, 24, 26-27

X Z

X-rays, effect on amino acid incorporation, Zymogen granules, in pancreatic exocrine
230-231 cell, 198-202