Full text of "Introduction to immunochemical specificity"

See other formats

3E3^^£

J^^^^^^^S^SI

Marine Biological Laboratory Library

Woods Hole, Mass.

Presented by

John Wiley and Sons, Inc<

July 26, 1963

3^^^^^^^^^Q^^^^S!

Introduction to
IMMUNOCHEMICAL SPECIFICITY

3 ^'^

CL^

Introduction to

Immunochemical Specificity

by WILLIAM C. BOYD

Professor of Immunochemistry
Boston University School of Medicine

INTERSCIENCE PUBLISHERS

a division of JOHN WILEY & SONS, NEW YORK • LONDON

Library of Congress Catalog Card Number 61-18288

Preface

This little book is intended to introduce the reader to our present
knowledge of immunochemical specificity (a somewhat broader
topic than the specificity of antibodies and antigens) by a discussion
of some of the important modern advances in this field and an in-
troductory account of certain earlier work that has served as a
foundation for recent progress. Aimed at the nonspecialist as well
as the specialist, the treatment is on the whole more elementary than
that of my Fundamentals of luununology, but in some ways more
detailed and up to date.

The material is based primarily on a series of lectures which I
had the privilege of giving in Moscow in the autumn of 1959. The
book is not, however, a mere retranslation of the Russian text, but a
thorough revision of the original English, with considerable additions.

Certain traces of the lecture form in which the material was origi-
nally cast still remain. Some of these may be disadvantages, but
some of them perhaps may not be. The style of lectures can and, in
my opinion, should be somewhat more informal than that of a text-
book. The lecturer is also more or less expected to do certain
things (and not to do others). He is expected, for example, to
bring his audience up to date, even to the point of presenting some
material not yet to be found in the textbooks and in some cases not
yet published. He is expected, or at least allowed, to discuss cer-
tain aspects of his own work in more detail than might seem proper
elsewhere. At the same time he is not required to cover the field

VI PREFACE

exhaustively and may he forgiven if, instead, he elects to emphasize
those portions which are of special interest to him personally. He is
also generally excused from presenting an exhaustive bibliography.

Most important of all, perhaps, lecturers are allowed to illustrate
their talks with numerous lantern slides, a feature which often greatly
increases the intelligibility of their presentation. This privilege is
reflected in the present case by a relatively high proportion of figures
and tables, which I hope will help in a similar way. In any case, the
illustrations form an integral part of the plan of the book.

Although the topic is a specialized one, little previous knowledge
of it is assumed on the part of the reader. An elementary knowledge
of organic chemistry and, for the last two chapters, a slight ac-
quaintance with the notation of partial differentiation should be
sufficient.

The author is grateful to the John Simon Guggenheim Foundation
for a fellowship that made the completion of this book possible, and
to friends and colleagues who read and criticized portions of the
manuscript.

Casa Rosada W. C. B.

May 1961

Contents

1. Antibodies I - 1

Immunity 1

Specificity 6

2. Antibodies II 12

Specificity and Chemical Structure of Antigen 12

Statistical Methods 20

Limitations of Specificity 26

Combining- Groups of Antibody 27

Formation of Antibody 29

3. Antigens 34

Definition 34

Antigenicity 34

4. Blood Groups 50

ABO Blood Groups 50

MNS Blood Groups 57

Rh Groups 58

Other Blood Groups 62

5. Plant Agglutinins (Lectins) I 64

Specificity of Proteins Other Than Antibodies 64

Plant Agglutinins 65

6. Plant Agglutinins (Lectins) II 72

Nature of Plant Agglutinins 72

Specificity of Plant Agglutinins -". 75

Role of Agglutinins in the Plant 82

Lessons from the Study of Lectins 83

7. Blood Group Antigens 85

Sources of Antigens for Study 85

Blood Group Substances A, B, H, and Le" 85

Other Human Red Cell Receptors 93

81578

viii CONTENTS

8. Salmonella Antigens 103

Endotoxins 103

The Salmonella . 105

Chemistry of the Polysaccharide Component of

Salmonella Antigens 107

Relation of Structure of Salmonella Antigens to Specificity 108
Cross-Reactions 112

9. Union of Antibody with Antigen: Thermodynamics 118

Forces Involved 1 18

Energy 124

Entropy 127

Free Energy 129

Free Energy and Equilibrium 131

10. Energy of Antibody-Antigen Reactions 134

Direct Calorimetry 134

Free Energy from Equilibrium Measurements 135

Significance of Thermodynamic Constants 140

Heat of Reaction of Isoagglutinins 145

Index 1 5 1

CHAPTER 1
Antibodies I

Immunity

Basically speaking, immunity is the increased resistance to an in-
fectious disease which often follows recovery from an initial attack.
The degree of this immunity is different with different diseases and
different patients and persists for varying periods of time. Recovery
from certain virus diseases, such as yellow fever, is followed by a
very high degree of resistance which seems to last for life in many
patients. Recovery from the common cold, on the other hand, is
followed by a very brief state of increased resistance, if indeed by
any at all. We shall also apply the term immunity to the artificially
increased resistance produced in a patient by injection or oral ad-
ministration of living virus or living microorganisms or by injection
of attenuated or dead virus or microorganisms or of antigenic prod-
ucts derived from such material. A patient whose resistance has been
heightened by such treatment is said to have been immunized, al-
though he may not be immune in the absolute sense of the word.
Animals which have been caused to produce antibodies by such ad-
ministration of antigen are also said to have been immunized, even
though they may not have obtained increased resistance to any
disease as a result.

Role of Antibodies in Imuinnity

The circulation of the immune animal often contains soluble pro-
tective proteins called antibodies, a term which is also applied to
specifically reactive proteins produced in response to any antigen,
whether it is derived from a pathogenic microorganism or not.

2 INTRODUCTION TO IMMUNOCHEMICAL SPECIFICITY

There is abundant evidence that antibodies play an important role
in an individual's resistance to many diseases. For example, the
transfer of antibody-containing blood from a convalescent patient to
a susceptible person will often make the recipient temporarily im-
mune to the disease from which the donor has just recovered (pas-
sive immunization). Transfer of antibody from an artificially im-
munized animal may be similarly efifective. A decisive change in the
sick patient's condition for the better ("crisis") many times coin-
cides with the appearance of specific antibodies in the blood. The
blood level of specific antibodies is often a fairly reliable index of
the degree of a person's immunity.

It is a characteristic feature of antibodies that they react with the
antigen which caused their production ; in fact, new proteins appear-
ing in the circulation which do not react in a detectable way with
the antigen responsible for their production in general are not called
antibodies. In a few instances antibodies have been observed which
reacted wath an antigen different from the one which caused them to
be produced and did not react visibly with their own antigen (Hooker
and Boyd, 1933; Glutton, Harington, and Yuill, 1938), but these are
exceptions.

The reaction of antibodies with their antigen can have one or more
of a number of effects : (a) Antibodies to toxins may neutralize the
toxity of the antigen, and antibodies to viruses may neutralize the
infectivity of the antigen, (b) Antibodies to soluble proteins and
other soluble antigens may precipitate their antigen (Fig. 1-1). (c)
Antibodies to microorganisms and foreign erythrocytes may cause
the antigenic cells to stick together (agglutinate) (Fig. 1-2). (d)
Antibodies to erythrocytes and certain microorganisms may cause
the antigenic cells to disintegrate. This phenomenon is called lysis,
and for its production the cooperation of certain normal components
of plasma, collectively called complement, is required, (e) Anti-
bodies to certain microorganisms, aided by complement, may cause
the death of the antigenic cells (bactericidal effect), (f) Antibody
to certain microorganisms causes the capsules of the microorganisms
to swell visibly. This phenomenon is generally referred to by its Ger-
man name Quellung. (g) Gombination of antibody with micro-
organisms and other foreign cells generally makes the invaders more
attractive to the leukocytes of the patient's circulation and thus pro-

ANTIBODIES I

Fig. 1-1. Photographs of the precipitin reaction. Reading from left to right:
negative reaction, weak positive reaction, strong positive reaction. These tests
were carried out by the interfacial technique — placing a layer of diluted antigen
over a layer of immune serum in a test tube).

■.•"** ■

I ' '■/_•' •' ^\

■ "* O- i, ""^ " ''

^ o o c*

^ °

-) '> ^ f> "p

|^ ft

Fig. 1-2. Photomicrographs of unagglutinated red blood cells (left) and
agglutinated cells (right).

4 INTRODUCTION TO IMMUNOCHEMICAL SPECIFICITY

motes phagocytosis, (h) Complement, if present, is generally taken up
when antibody and antigen combine ; thus, the occurrence of an other-
wise undetectable antibody-antigen reaction can sometimes be de-
duced from observation of complement fixation alone. The classical
Wassermann test for syphilis is based on this phenomenon, (i) The
combination of antibody and antigen may lead to the release of
histamine and other toxic substances from the tissues of the host,
as in anaphylaxis and allergy.

All these effects of antibody, except probably those of class (i),
are thought to be beneficial to the host and to aid in resistance
to infection. All of them may, under suitable conditions, be utilized
in laboratory studies.

However, although there is no doubt of the importance of antibodies
in immunity, they are by no means the whole story, and the natural,
more or less nonspecific mechanisms of resistance, such as im-
permeability of the skin and mucus membranes, and bactericidal
power of these body surfaces, the rise in body temperature which often
accompanies infection, the action of normal plasma components such
as complement and properdin, and the ingestion of invading micro-
organisms by the leukocytes (phagocytosis), are also important. In-
deed, of all the mechanisms of resistance, phagocytosis is probably
by far the most important. However, we shall here be concerned
with specific mechanisms of immunity and shall not further discuss
these other tools of resistance.

Nature of Antibodies

In view of the importance of antibodies in immunity and of their
theoretical interest as prime examples of specifically reacting bio-
logical substances, it is not surprising that many attempts have been
made to study their chemical nature. Thus far it has not been pos-
sible to ascertain by direct chemical analysis the structural basis for
the combining power and specificity of antibodies, because (a) it is
not easy to obtain large amounts of purified antibodies and (b)
protein chemistry is not far enough advanced for detailed knowledge
of the structure of any antibody molecule to be obtained.

In spite of the difficulties, some preparations of purified antibody
have been studied. There has also been analytical work on antibody-
antigen compounds, which are more readily available in a relatively

ANTIBODIES I 5

pure state. Studies of the changes in the composition of blood follow-
ing immunization have suggested that antibodies belong to the class
of serum proteins called globulins. The distinction between serum
albumin and serum globulins was originally based on solubility
characteristics in neutral salt solutions (Cohn et al., 1940; Svensson,
1941). It is now based more on the observation that in an electric
field the serum globulins move more slowly, at alkaline pH (Tiselius,
1937). In his classical paper, Tiselius (1937) pointed out that nor-
mal serum globulin showed components of at least three different
electrophoretic mobilities, and he designated them as alpha, beta, and
gamma globulins in order of decreasing mobility. Antibodies, with
some possible exceptions, belong to the gamma globulin class. It
is this group of globulins that increases most following immunization
(TiseHus and Kabat, 1939).

In man, the rabbit, and many other species, antibodies are found by
ultracentrifugal measurements to have the molecular weight charac-
teristic of serum globulins, namely about 160,000. In the horse, pig.

(a)"

l^^^^^^^l

^ggH^

^^9

1^1

■

Fig. 1-3. Models of typical antibody molecules with human serum albumin
for comparison, (a) Horse anti-pneumococcus antibody; (b) horse antitoxin;
(c) rabbit anti-ovalbumin antibody; (d) human anti-pneumococcus antibody;
(e) rabbit anti-pneumococcus antibody; (f) human gamma globulin; and (g)
human serum albumin.

6 INTRODUCTION TO IMMUNOCHEMICAL SPECIFICITY

cow, and perhaps other species, anti-pneumococcus antibodies, for
example, have a molecular weight of about 900,000. Certain anti-
bodies in human blood seem to have molecular weights somewhere
between these values (discussion in Boyd, 1956). From ultracentri-
fugal sedimentation constants, diffusion constants, and Perrin's
(1936) relation between the "frictional ratio" and axial ratio of a
prolate spheroid, the shape of protein molecules may be calculated.
Photographs of models of typical antibody molecules, with human
serum albumin for comparison, are shown in Fig. 1-3.

Actual photographs of antibody molecules, taken with the electron
microscope, reveal, as far as the still inadequate resolving power
of this instrument allows, a striking similarity to the models shown
in Fig. 1-3 (see Fig. 1-4). Also of interest in the photograph
is the apparent heterogeneity in size.

Fig. 1-4. Electron micrographs of rabbit antibody molecules. (Photograph
by Dr. C. E. Hall.)

Specificity

It is a very old observation that immunity is specific. A child who
has recovered from whooping cough is very unlikely to get this disease
again in the immediate future, but his resistance to measles is not

ANTIBODIES I 7

any greater than before. Even recovery from the superficially very
similar disease German measles does not seem to confer any im-
munity to ordinary measles, and the child generally catches all the
common childhood infections, one after the other. If he misses one
or more of them he remains susceptible to it, as he demonstrates
by promptly coming down with it when exposed later in life, possibly
from one of his own children. However, this specificity is not ab-
solute. As an example we may mention that recovery from the rela-
tively mild flea-borne typhus caused by Rickettsia mooseri is fol-
lowed by immunity to the much more serious louse-borne typhus
caused by Rickettsia prozmceki (Rivers, 1952).

Specificity of Antibodies

Just as the immunity following recovery from infections is rela-
tively specific, so is the power of antibodies to react with antigens.
Diphtheria antitoxin will neutralize diphtheria toxin and possibly
save the life of a patient with diphtheria ; it does not neutralize
tetanus toxin and is of no value in the prevention or treatment of
tetanus. In general, antibodies seem to be adapted to react just with
the antigen . which called forth their production (homologous anti-
gen). But the specificity of antibodies, like the specificity of im-
munity, is not absolute. Antibodies produced by injecting rabbits
with purified ovalbumin from the hen react also with the ovalbumin
of various other birds such as the duck. A reaction of an antibody
with a related antigen is called a cross-reaction. It is generally not
as strong as the reaction of the antibody with the homologous anti-
gen.

Today there is no doubt that the specificity of antibodies depends
on their chemical structure. But there is as yet no agreement whether
the specificity is a result of differences in amino acid composition or
in amino acid sequence, or merely of the way in which the polypeptide
chain is folded to produce the globular molecules shown in Fig. 1-3.
Pauling (1940) proposed the latter view. It is a fact that amino
acid analyses of antibodies have not yet revealed any clear-cut dif-
ferences in amino acid composition between different antibodies or
between particular antibodies and normal globulin (Boyd, 1956;
Smith et al., 1955). If the differences were mainly in amino acid
sequence and were confined to a "central, dififerential segment"

8 INTRODUCTION TO IMMUNOCHEMICAL SPECIFICITY

(Lederberg, 1959), they might be too minor to be found by the
presently available methods of analysis.

Although the brilliant researches of Sanger (1956) into the se-
quence of amino acids in the polypeptide chain have resulted in the
complete elucidation of the structure of insulin and a few other
polypeptides, such analysis has not yet been carried very far with
antibodies. However, we do know that the N-terminal sequence of
all rabbit globuHns thus far studied, including various antibodies,
seems to be (Porter, 1950; McFadden and Smith, 1955) :

Alanine- — leucine — valine — aspartic acid — glutamic acid —
In contrast to the uniformity of rabbit globulins, horse globulins,
whether antibody or normal globulin, have proved to be quite hetero-
geneous in this respect, all preparations exhibiting a wide variety of
N-terminal groups (McFadden and Smith, 1955b). The globulins
of man likewise differ among themselves in amino acid composition
(Smith et al., 1955b; Putnam, 1955).

Pozvers of Discrimination of Antibodies

It was long ago suspected that the cross-reactions of antibodies
with related antigens were due to chemical similarities between the
homologous and the cross-reacting antigen. But in the absence of
detailed information about the chemical structure of natural antigens
(a situation which has improved only slightly since the earliest
days of immunochemical work), it was not possible to state how
great the chemical similarity between two antigens had to be to
make cross-reaction possible or, to put it another way, how small a
chemical differences antibodies could detect. Karl Landsteiner (1945)
largely overcame this difficulty by the use of conjugated antigens.

It was known that chemical treatment (nitration, iodination, etc.)
of protein antigens often changed the immunochemical specificity.
Landsteiner showed that if simple chemical compounds were coupled
chemically to protein antigens it was possible to produce antibodies
which reacted specificially with the simple free compound (which
Landsteiner called a hapten). Thus it was possible to observe sero-
logical reactions which depended only on the hapten, the structure
of which was known, and not on the natural protein antigen of yet
undetermined chemical structure.

ANTIBODIES I

As an example, let us take the aromatic amine, metanilic acid,
and diazotize the amino group by treating the compound with nitrous
acid (Fig. 1-5).

HNO.. -^

JSO3H I JSO.,H

Metanilic acid- H CI Diazotized metanilic

acid

Fig. 1-5. Diazotization of metanilic acid.

The resulting diazonium salt will couple with phenols in alkaline
solution to give colored azo dyes (Fig. 1-6). It will also couple with
the phenolic group of the amino acid tyrosine, a constituent of most
proteins.

+
JSO3H H

Diazotized metanilic Phloroglucinol Azo dye

acid

Fig. 1-6. Coupling of diazotized metanilic acid with phloroglucinol to form
an azo dye.

Let us suppose we couple the diazotized metanilic acid (our hapten)
with the mixture of proteins provided by horse serum. If we repre-
sent the horse serum proteins by H and the diazotized metanilic acid
by M, we may reprsent the coupled azoprotein as HM. Injection of
this compound HM into rabbits will usually cause the production of
a number of different antibodies to the proteins of horse serum, which
we may designate collectively as anti-H, and antibodies to metanilic
acid, which we may designate as anti-M.

If, on testing the rabbit antibodies with the antigen HM we in-
jected, we get a positive serological reaction (formation of a specific
precipitate), we shall not know whether this is due to the union of

INTRODUCTION TO IMMUNOCHEMICAL SPECIFICITY

M and anti-M or the union of H and anti-H, or both. But let us
suppose we couple diazotized metanilic acid with the proteins of
chicken serum, which we may designate as C. Rabbit antibodies to
horse serum proteins do not precipitate with chicken serum proteins,
so the confusion caused by interference of the protein carrier is elimi-
nated. If we mix our immune rabbit serum with CM and obtain a
precipitate, we know it is due to the reaction of the anti-M of the
rabbit serum with the M (metanilic acid) we have coupled with the
chicken serum C (Fig. 1-7).

Rabbit immunized with complex
antigen (horse serum coupled with
hapten M) gives ontiserum "anti-M,"
which contains anti-horse antibodies
and antibodies for hopten (anti-M)

Precipitote

Precipitate

"Anti-M" Chicken No precipitate "Anti-M" Chicken-M Precipitate

Fig. 1-7. Principle of detecting antibodies to a hapten independently of anti-
bodies to the protein carrier.

References

Boyd, W. C, 1956, Fundamentals of IininiDiology, 3rd ed., Interscience,

New York.
Cohn, E. J., T. L. McMeekin, J. L. Oncley, J. AI. Newell, and W. L. Hughes.
1940, /. Am. Chem. Soc. 62, 3386.

ANTIBODIES I 11

Landsteiner, K., 1945, The Specificity of Serological Reactions, 2nd rev. ed.,

Harvard University Press, Cambirdge.
Lederberg, J., 1959, Science, 129, 1649.

McFadden, M. L., and E. L. Smith, 1955a, /. Biol. Chem. 214, 185.
McFadden, M. L., and E. L. Smith, 1955b, /. Biol. Chem. 216, 621.
Pauling, L., 1940, /. Am. Chem. Soc. 62, 2643.
Perrin, P., 1936, /. phys. radium 7, 1.
Porter, R. R., 1950, Biochem. J. 46, 473.
Putnam, F. W., 1955, Science 122. 275.
Rivers, T. M., 1952, Viral and Rickettsial Infections of Man, Lippincott.

Philadelphia.
Sanger, F., 1956, in D. E. Green (ed.), Currents in Biochemical Research

1956, Interscience, New York.
Smith, E. L., M. L. McFadden, A. Stockell, and V. Buetttner-Janusch, 1955a,

/. Biol. Chem. 214, 197.
Smith, E. L., D. M. Brown, M. L. McFadden, V. Buettner-Janusch, and

B. V. Jager, 1955b, /. Biol. Chem. 216, 601.
Svensson, H., 1941, /. Biol. Chem. 139. 805.
Tiselius, A., 1937, Biochem. J. 31, 1464.
Tiselius, A., and E. A. Kabat, 1939, /. E.vptl. Med. 69, 119.

CHAPTER 2
Antibodies II

Specificity and Chemical Structure of Antigen

We are now in a position to investigate the effect on serological
specificity of known variations, large or small, in the chemical com-
position of cross-reactive antigens. Suppose, for example, we replace
the metanilic acid which we employed in making the azo chicken
serum CM with a different aromatic amine, which we may designate
as M'. Then if we test CM' with our antiserum against horse-
metanilic acid, a positive reaction will depend on whether M' is suf-
ficiently similar to M to combine with anti-M. If M' is w-amino-
benzoic acid or w-aminoarsenic acid, for example, a positive reac-
tion may be obtained with an antiserum to horse-metanilic acid,
although the amount of precipitate will be less. This shows that the
antibodies to metanilic acid, though best adapted to the homologous
hapten, can react also with other haptens having a dift'erent acid
group in the meta position. If we make the difference greater by em-
ploying a hapten with the acid group in the para position, or use a
hapten without any acid group, little if any cross-reaction is ob-
served (Fig. 2-1). We conclude that antibodies are able to dis-
tinguish structural isomers (molecules containing the same groups
but in different positions) but can also distinguish simple groups
(acid groups in this instance) which occupy the same position.

It is a characteristic of antisera that the antibody molecules they
contain are not all alike ; they may differ in strength of reaction with
a given hapten and may differ in their specificity. This is easily shown
in the present case by allowing the anti-metanilic acid serum to react

ANTIBODIES II

NH2 NH2 NH.

Antigen
containing

JSO3H 1^ JA3O3H2 ^ JCOOH

SO-jH

Metanilic acid w-Aminoarsonic ;w-Aminobenzoic Sulfanilic Aniline
acid acid acid

Strength of ^ q

reaction ^^=^ ^ ± = u

Fig. 2-1. Reaction of antiserum for metanilic acid (Landsteiner, 1945).

with a protein coupled with one of the cross-reacting haptens vmtil
no further reaction takes place, then testing the treated antiserum with
the homologous and other related haptens. In order to avoid having
soluble complexes of antibody and antigen left in the mixture, the
anti-metanilic serum may be treated with haptens coupled to the
insoluble structures (stromata) left after lysis of red blood cells and
removal of the hemoglobin. A serum which has been thus treated
to remove all the antibody which will react with a given antigen is
said to have been absorbed with that antigen. The results of such an
experiment (Landsteiner and van der Scheer, 1936) are shown in
Fig. 2-2.

It can be seen that in each case absorption with heterologous
hapten-protein compound, to the point of reducing the reaction with
that hapten to zero, leaves considerable precipitating power for
antigens containing the homologous hapten. Generally, it also leaves
some reactivity for other heterologous haptens. Each hapten, evi-
dently, combines with that fraction of the antibody molecules for
which it has the highest affinity. The majority of the antibody mole-
cules react best with the homologous hapten, most of which is left
after heterologous absorption.

Similar results were obtained by Hooker and Boyd (1934) and
Landsteiner and van der Scheer (1940) with egg albumins of vari-
ous species, although here the exact nature of the chemical similari-
ties which led to cross-reaction was not known.

From these and similar experiments (Landsteiner and van der
Scheer (1936) drew the conclusion that "antibodies formed in re-
sponse to one antigen, although adjusted to a certain structure, are not
entirely uniform but vary in specificity to some degree." Boyd (1943)

14 INTRODUCTION TO IMMUNOCHEMICAL SPECIFICITY

Anti-metanilic acid(M) Antigens made with chicken serum and:

antiserum absorbed with

azo-stromata made with: M GAB

+ + ±

+ + + +±

(Unabsorbed) + + + ± ++ + +±

Fig. 2-2. Precipitin reactions of absorbed anti-metanilic acid antiserum with
various conjugated antigens.

suggested that we should think of the antibody molecules of an im-
mune serum as "a large family, with varying degrees of deviation from
a mean." Pauling, Pressman, and Grossberg (1944) made a similar
and more precise suggestion. In their opinion, the free binding ener-
gies of the different antibody molecules (for the determinant that
induced their formation) are distributed according to the Gauss
error function.*

In the description of the reactions of antibodies with simple sub-
stances (haptens) it was stated that to detect anti-metanilic acid
(anti-M) antibodies, for example, we make use of a protein, different
from the one used as the carrier of the hapten during immunization,
coupled with diazotized metanilic acid. It may have occurred to the
reader to ask what would happen if we mixed the anti-M serum di-
rectly with metanilic acid?

* This is the well-known "normal distribution" formula of statistics,
fix) = [l/aV(27r)]exp i-xyia')
where a, called the standard deviation, is a measure of the "dispersion" or degree
of heterogeneity of the population whose composition is summarized by the
curve. Two graphic examples of this distribution are shown in Fig. 2-3.

ANTIBODIES II

-6 -5 -4 -3 -2

Fig. 2-3. Probability ("error") distribution when standard deviation equals
1 and 2.

The answer is that with simple haptens such as metanih'c acid
no visible reaction occurs as a rule. Originally, it was thought that
the simple hapten was too small to take part in a precipitin reaction,
but we are now inclined to believe that, although size may have
something to do with it, the main deficiency of metanilic acid and other
simple haptens in this respect is that they have only one point of attach-
ment (combining group) for the antibody. Haptens containing two
or more combining groups sometimes precipitate with the anti-hapten
antibody.

Nevertheless, the anti-metanilic acid antibody has a strong af-
finity for metanilic acid, for it combines with this hapten when it is
part of a metanilic acid-protein compound and forms a specific pre-
cipitate in the usual way though it does not precipitate metanilic acid
from simple solution. The anti-M antibody may even combine with
metanilic acid itself without producing a precipitate.

Inhibition Reactions

Now it was known that an excess of antigen usually prevents the
production of a specific precipitate or greatly diminishes the amount.
Clearly it must do this by combining with the antibody, for addition
of an imrelated antigen has no such inhibitory efifect. Therefore,
Landsteiner reasoned that an excess merely of the hapten, which

INTRODUCTION TO IMMUNOCHEMICAL SPECIFICITY

logically must combine with specific antibody, ought in a similar
way to prevent or diminish precipitation of the anti-hapten antibody
with a hapten-protein compound. Experiments showed that this does
in fact happen. Thus Landsteiner invented the inhibition reaction,
which has been of enormous value in the study of immunochemical
specificity. How specific inhibition works is shown schematically in
Fig. 2-4.

Antibody

Antigen

Precipitate or ogglutinate

Antibody

Antibody- hapten
complex

No precipitate
or agglutinate

Antibody- tiopten Antigen

complex

Fig. 2-4. Principle of inhibition by a iiaptcn of serological reactions.

ANTIBODIES II 17

To inhibit completely the reaction of an anti-hapten antibody with
the conjugated hapten-antigen generally requires a good deal more of
the inhibiting hapten, in dissolved form, than that contained in the
conjugated antigen. From this we might deduce that the binding force
of the hapten alone with the antibody is less than the force which
unites the antibody and the conjugated antigen. The hapten does
combine with the antibody, however, for unrelated haptens have no
inhibitory effect, and a given hapten does not inhibit unrelated anti-
body-antigen reactions. In other words, inhibition is specific. In
Fig. 2-5 the inhibition is completely specific, that is, each hapten pre-
vents precipitation only of the homologous antibody and antigen. That
haptens combine with their specific antibody can be demonstrated by
the power of antibody to prevent a diffusible hapten from dialyzing
through a membrane otherwise permeable to it.

If closely related haptens are tested against the same antibody-

Amount of precipitate given by antisera
for corresponding antigen, in presence of

Antiserui
for

H2N

AsOsHz H2N

H2N

COOH

SO3H

H2N

+ + +

+ +

+ + +

COOH

Fig. 2-5. Inhibition of precipitin reaction with homologous and heterologous
haptens.

18 INTRODUCTION TO IMMUNOCHEMICAL SPECIFICITY

antigen system, it is usually found that the homologous hapten in-
hibits best and that other haptens inhibit more or less well, depend-
ing on the degree of their resemblance to the hapten contained in the
immunizing antigen. Thus, in Table 2-1 we see that mononitro-
strychnine inhibits the reaction of an anti-strychnine antiserum with
a conjugated strychnine-antigen as well as does strychnine itself,
whereas the related alkaloid brucine, which differs from strychnine

TABLE 2-1

Inhibition of Anti-Strychnine Sera by Various Haptens*
Hooker and Boyd, 1940

Micromoles of test subst

ance^"

Test substance

1.00

0.67

0.44

0.30

0.20

0.13

0.09

0.06

0.04

Strychnine

Mononitrostrychnine

Dinitrostrychnine

Monoaminostrychnine

Diaminostrychnine

Brucine

Morphine

» Hooker and Boyd, 1940.

*> The symbols indicate the degree of precipitation obtained when the hapten-
serum mixture was overlayered with a suitable concentration of the strychnine-
protein antigen. The symbol + indicates a positive precipitation reaction, ± a
weak reaction, "t" a faint trace, and 0 no precipitation. Absence of a symbol
means that the test was not done or could not be read because of nonspecific
precipitation.

only in possessing two methoxy groups, does not inhibit as well ; that
is, a larger amount of it is required to prevent the antibody-antigen re-
action. The unrelated alkaloid morphine does not inhibit at all. In
this case the effectiveness of different haptens was compared by
testing decreasing amounts (increasing dilutions) of the haptens
against a constant amount of antiserum, to which was later added a
suitable amount of antigen.

In general, we may expect the results obtained with haptens H, H',
H", and G, where H' is closely related chemically to H, W less

ANTIBODIES II

closely related, and G unrelated, to give results similar to those in
Table 2-1. This is shown schematically in Table 2-2.

TABLE 2-2

Precipitation Reaction of Anti-H Antibody and H Antigen,
in Presence of Hapten*

Dilution

of hapten

Hapten

1:2

1:4

1:8

1:16

1:32

1:64

+ ±

+ +

-F-H

+ +

* ++ = Strong reaction. Other symbols as in Table 2-1.

The relative effectiveness of different haptens as inhibitors can
also be shown by plotting concentrations, or maximum dilutions
which inhibit (Fig. 2-6).

I I 1 1 1

Fig. 2-6. Inhibition of precipitin reaction of anti-benzoic acid antibody by
various haptens.

20 INTRODUCTION TO IMMUNOCHEMICAL SPECIFICITY

Another way of comparing the inhibiting powers of different hap-
tens is to use constant concentrations (preferably expressed as molari-
ties) of haptens against successive dilutions of the antiserum. In
this case, again using the hypothetical haptens H, H', H", and G,
we might obtain the sort of result shown in Table 2-3.

TABLE 2-3

Precipitation Reaction of Anti-H Antibody and H Antigen,
in Presence of Hapten*

Dilution of

antiserum

Hapten

1:2

1:4

1:8

1:16

1:32

1:64

+ +

-I- +

+ +

-f-1-

+ -H

+ +

+ -f

* Symbols as in Table 2-2.

Instead of trying to find the antibody concentration which is com-
pletely inhibited by a given concentration of hapten, or the hapten
concentration which will completely inhibit a given concentration of
antibody, it is more accurate to measure the amount of precipitate
produced under the various conditions, and estimate the amount of
hapten which gives just 50 per cent inhibition.

Statistical Methods

If several such series of quantitative measurements are carried
out, it is possible to obtain a mean (average) estimate of the 50
per cent inhibiting dose and, from the standard deviation of this
mean, an estimate of its reliability. When such standard errors
are calculated they tend to be rather large, for the quantitative
precipitin technique is not as reproducible as the measurements of
inorganic quantitative analysis or physical chemistry are. For this
and a variety of other reasons, standard errors are not usually
calculated for such estimates : (a) The necessary determinations would
require too great an outlay of the experimenter's time and of an

ANTIBODIES II 21

antibody that may be in short supply, (b) The goal of such experi-
ments is not usually an estimate of the actual inhibiting dose of any
one particular hapten, but an estimate of the relative inhibiting power
of two different haptens ; in other words, a ratio. It is quickly
found that attempts to calculate the standard error of a ratio from
the standard errors of the two numbers involved leads one into
Higher Statistics.

It might seem to the non-serologist that in the simple type
of inhibition study shown schematically in Tables 2-2 and 2-3 sta-
tistical methods could be applied and would be helpful, but this is
not generally the case either. To begin with, inhibition experiments
are ordinarily interpreted as if inhibition were an all-or-none
phenomenon. Thus from the first line of Table 2-2 we conclude
that for complete inhibition (tube 4, counting from the left) of
the amount of serum used in the experiment 1/16 of the amount
of hapten H contained in a unit volume of stock solution is sufficient.
But for all we know the amount of hapten in tube 4 may be any-
where from 1.02 to 1.98 times the minimal inhibiting dose (MID)
of H. If tube 4 contains 1.02 MID, then tube 5 in turn contains
only 0.51 MID, and the unavoidable accidental variations in experi-
mental conditions are not likely to cause tube 5 to give a nega-
tive reading, though they well might make tube 4 positive. But if
tube 4 contains 1.98 MID, which is equally possible, then tube 5
would contain 0.99 MID, and a slight variation in the conditions
of the experiment might mean that tube 5 would read negative in-
stead of positive. Thus in different experiments our estimate of the
smallest amount of hapten that will completely inhibit a given amount
of antiserum might vary from 1.0 to 2.0 to 0.5 mM. An experimenter
is likely to feel that he is wasting his time in averaging numbers like
1.0, 2.0, and 0.5, not to speak of trying to estimate a standard devia-
tion and a standard error of the resulting mean.

Also, it must be realized that, just as the results with hapten H
might vary from 1.0 to 2.0 to 0.5 mM, so the results with hapten H'
might vary from 0.25 to 0.5 to 0.125 mM. It is expected that the
variations in estimated MID's of the two haptens will generally
go in the same direction ; indeed, this is one of the reasons for run-
ning all the tests simultaneously, but it is apparent that the ratio
of the apparent MID's might vary from 16 to 1.

22 INTRODUCTION TO IMMUNOCHEMICAL SPECIFICITY

In the second place, the all-or-none interpretation of inhibition
experiments is an oversimpHfication. The experiment summarized
in the first Hne of Table 2-2 actually yields more information than
is contained in the mere statement that tubes 1 to 4 are negative and
tubes 5, 6, etc., are positive. The strength of the reaction in the
first tube to the right of the last negative tube also contributes in-
formation, for the reaction can vary from weak to strong. Taking
the simple point of view, for example, we should estimate from
Table 2-2 that hapten H' is only one-fourth as effective an inhibitor as
hapten H, for it takes four times as much to produce complete in-
hibition. But if we take account of the fact that the next tube after
complete inhibition gives a reaction of + in the case of hapten H
and only ± in the case of H', it is clear that H' is actually some-
what more than one-fourth as effective as H'. But how much more?
It is hard to put such things into numerical terms. It is possible to
invent codes for the translation of such readings into quantitative
terms, or appropriate numerical scores may be found by statistical
methods (see, for example, Fisher, 1950, pp. 289—295). In general,
however, such treatments of the results of inhibition tests have not
been found to extract enough extra information from the results
to justify the calculations involved.

Recognizing, therefore, that the results of inhibition experiments
are only semiquantitative at best, serologists who are attempting to
compare the inhibitory power of two different haptens do not
generally attempt to make quantitative estimates, but are content
to say merely that hapten H is more effective, for this particular
serum, than H' is. Some tend to rely on the old rule of thumb,
which is pretty well borne out in practice, that a difference in the
results obtained with two haptens is significant if the difference in
their inhibiting capacity differs rather consistently, from one ex-
periment to another, by two tubes (ordinarily meaning a four- fold
difference in effective concentrations). If the results do not differ
by this much, one may suspect a difference in the effectiveness of the
two haptens, without venturing a confident opinion. Even such a
difference, however, arbitrarily judged to be "non-significant," may
be of value as a guide to further experiments.

The value of statistical methods in general is of course not in
doubt. For a lonsj time certain biologists, and immunologists in

ANTIBODIES II 23

particular, often failed to avail themselves as fully as they might
have of statistical methods (Batson, 1951). The situation has
pretty well been corrected in recent years, however. A summary of
some of the current applications of statistics to immunology will
be found in my Fundamentals of Immunology (Boyd, 1956). Indeed,
it is to be feared that today there are a few biologists who feel,
as many physical anthropologists did 30 years ago, that statistics
will cure all ills. Actually, of course, the results of statistical analysis
can never be better than the data themselves. A false sense of security
stemming from a blind application of statistical methods to situations
where they are not appropriate can be as bad as the former tendency
to avoid their use.

Stereoisomerism
Having shown that antibodies can distinguish structural isomers,
Landsteiner naturally asked if they could also distinguish stereoiso-
mers and therefore next turned his attention to this problem. In
organic chemistry it is established that, whenever a carbon atom has
four different groups attached to it, there are two possible arrange-
ments of these groups which are essentially different from each other,
somewhat as the right and left hands differ from each other (Fig. 2-7).

z z

Fig. 2-7. Models of right- and left-handed molecules.

The essential difference between the two possible isomers in such
a case is correctly shown only in three dimensions. Since it is not
generally convenient to have three-dimensional models in front of
us when discussing isomerism, it is customary to represent such com-
pounds by a projection on two dimensions, as shown below.

X X

I I

W - C - Y Y - C - W

I I

z z

24 INTRODUCTION TO IMMUNOCHEMICAL SPECIFICITY

As long as we remember that such projections must not be taken out
of the plane of the paper, no incorrect conclusions will be drawn from
their use.

Organic compounds differing in the spatial arrangement of the
four different groups attached to a carbon atom have the same chemi-
cal and physical properties but differ in their effect on polarized light.
Therefore, they are said to be optical isomers. An example of such
an optically isomeric compound is aminobenzoylphenylaminoacetic
acid, which exists in two forms. One, the d form, rotates the plane
of polarized light to the right. The other, the / form, rotates it to the
left (Fig. 2-8).

H COOH

Hon/ \— conhc— / \ h.n/ \— conhc— / \

COOH H

rf-/)-aminobenzoylphenylamino- /-/)-aminobenzoylphenylamino-

acetic acid acetic acid

Fig. 2-8.

Containing an aromatic amino group, these compounds can be
diazotized, coupled to proteins, and made to function as haptens.
Since optical isomerism plays a very important role in biochemistry,
we would expect that antibodies would be able to distinguish these
two isomeric haptens. In fact, Landsteiner and van der Scheer (1928)
found that, although the undiluted antigens gave some cross-reaction,
they reacted quite specificially when diluted one to one hundred
(1:100).

In later work Landsteiner and van der Scheer (1929) showed that
D- and L-tartaric acid (Table 2-4), where two asymmetric carbon
atoms are involved, could be differentiated by the appropriate anti-
bodies and that both were distinguishable from the "internally com-
pensated" mesotartaric acid.

Next to proteins, polysaccharides are the most important natural
antigens. It was therefore logical to ask if isomers of sugars which
differ in the configuration of one or more carbon atoms and are not
necessarily optical antipodes could be distinguished by antibodies.
Goebel and Avery (1929) showed that the monosascharides D-glucose
and D-galactose, which differ only in the configuration of the fourth

ANTIBODIES II

TABLE 2-4
Serological Specificity of Stereoisomers of Tartaric Acid

Antigen from*

/-Tartaric acid

d-Tartaric acid

Mesotartaric acid

COOH

1
HOCH

HCOH

HOCH

HCOH

I
COOH

COOH

Immune serum for:
/-Tartaric acid
J-Tartaric acid
Mesotartaric acid

+ + +
0

+ + +

" Symbols indicate degree of precipitation when antisera for conjugated proteins
containing isomers on left were mixed with proteins containing isomers shown on
right. The symbol -|- + + indicates strong positive reaction ; 0, negative reaction.

carbon atom (Fig. 2-9), could be distinguished serologically. Later
work (Avery, Goebel, and Babers, 1932) even showed that the alpha
and beta anomers of glucose, when converted to the />-aminophenyl-

eCHgOH

H.OH

H,OH

H OH H OH

D- Glucose D- Galactose

Fig. 2-9.

glucosides, diazotized, and coupled to a protein, gave rise to different
antibodies (Fig. 2-10). The distinction was not as sharp as between

N = N —

CH,OH

CH2OH

Fig. 2-10. a- and /3-D-glucoside haptens.

26 INTRODUCTION TO IMMUNOCHEMICAL SPECIFICITY

glucose and galactose, which, in view of the fact that a- and y8-glucose
are spontaneously interconvertible in solution, seems reasonable.

Limitations of Specificity

It is apparent from this and similar work, that though the power
of antibodies to distinguish small chemical differences in antigens is
very considerable, this discrimination has certain limits, limits which,
in the case of the alpha and beta anomers of glucose, we have almost
reached. The number of different antibodies is certainly large, but
we are moved to ask : Is there perhaps a limit to the number of
substances which can be distinguished serologically?

We have by no means tested all possibilities, but I believe that the
answer to the above question is that there probably is a limit. In
the first place, cross-reactions are regularly found with closely re-
lated antigenic determinants (haptens), as we have just seen. In
the second place, antibodies to natural antigens are not directed
toward the molecules as a whole but toward relatively restricted por-
tions of the molecule (Chapter 3). These restricted portions of the
molecule consist of amino acid residues and combinations of amino
acid residues in the case of proteins and, in the case of carbohydrates,
of monosaccharide residues and combinations of monosaccharide resi-
dues. The number of amino acid residues occurring naturally is only
somewhat greater than twenty (Yeas, 1958). The number of mono-
saccharide residues occurring in any considerable amount in nature
is probably not much greater. The number of possible antigenic
specificities is therefore not infinite. It is accordingly not surprising
that, as more and more cross-reactions between antigens of unrelated
or remotely related origins are tried, more and more cross-reactions
are found to take place.

Thus, cross-reaction occurs between human blood group A sub-
stance and pneumococcus type 14 capsular polysaccharide (Finland
and Curnen, 1940). Anti-pneumococcus type 14 sera strongly cross-
react with a galactan isolated from cow lung (Heidelberger and
Wolfram, 1954). Pneumococcus type 2 capsular polysaccharide and
the polysaccharide from encapsulated type B Friedlander bacillus
cross-react (Avery, Heidelberger, and Goebel, 1925). Highly active
substances with specificity similar to that of the human blood group

ANTIBODIES II

substances are found in certain plants (Springer, 1958). One may
predict that the number of such serological similarities will grow
as the number of individual antigens tested for cross-reactivity
increases.

Combining Groups of Antibody

The fact that no striking chemical differences between antibodies
or between antibody and normal globulin have yet been found sug-
gests that the portion of the antibody molecule responsible for its
specific combining properties cannot be very large. This idea is sup-
ported by the evidence, to be discussed in the next chapter, that the
portion of the antigen with which an antibody combines is relatively
small, at least compared with the size of a protein molecule. From
experiments of Landsteiner and van der Scheer (1938), Campbell and
Bulman (1952) computed that the specific combining site of an anti-
body is not larger than 700 square angstrom units (700 A^).

It is believed that van der Waals forces are among the most im-
portant in the union between antibody and antigen. Since these are
very short range forces, being inversely proportional to the seventh
power of the distance, the combining groups of antibody and antigen
probably come into intimate contact to produce union as firm as that
actually observed. (The free energy change — AF is of the order of

Fig. 2-11. \^an der Waals outlines of o-, m-, and /'-azobenzenearsonates.
(From L. Pauling and H. A. Itano (eds.), 1957, Molecular Structure and Bio-
logical Specificity, American Institute of Biological Sciences, Washington, by
permission of the editors and publishers).

INTRODUCTION TO IMMUNOCHEMICAL SPECIFICITY

5 to 9 kcal. per mole. See Chapters 9 and 10.) Hooker and Boyd
(1941) suggested that the combining group of the antigen might fit
into a cavity in the antibody. Pauling made a similar suggestion.
Figure 2-11 shows Pauling's conception of antibody cavities corre-
sponding to 0-, m-, and /'-aminoarsonic acid.

To what extent such a cavity in the antibody is merely schematic
;and to what extent it is real is not yet decided. The concept has
certainly proved useful in thinking about antibody-antigen reactions.

Fig. 2-12. Schematic drawings of three possible types of cavities (deter-
minants) in antibody molecule: 1, invagination; 2, shallow trough; 3, slit
trench.

ANTIBODIES II 29

In any case, the cavity is not necessarily the deep invagination sug-
gested by Fig. 2-11 and example 1 in Fig. 2-12. The antigenic deter-
minant might alternatively be accomodated lying on its side in a
shallow trough (example 2 in Fig. 2-12), or sidewise in a sort
of slit trench (example 3 in Fig. 2-12). According to Pressman
(1957), there is evidence that all three types of antibody cavity
exist.

If the combining group of an antibody molecule is relatively small
(one such group, according to the above estimate, would amount to
about 2 per cent of the surface of an antibody molecule), we naturally
ask how many such groups an antibody has. It seems conceivable that
one group would be enough to account for the reactions of antibody.
For some time Dr. S. B. Hooker and I and some other workers in
this field maintained that on the basis of economy of hypotheses
(Occam's razor) it should be assumed that antibody was univalent.
Others assumed that antibody was multivalent. There is now con-
siderable experimental evidence indicating that neither party to this
controversy was wholly right, for the valence of antibody seems to
be two. There are certain antibodies, especially in connection with
the Rh blood groups, behaving in peculiar ways which have led to
their being described as "incomplete" or "univalent." The presently
available evidence, however, indicates that the peculiarity of their
behavior is not due to their having less than the usual number of com-
bining groups but to other features of the molecule.

Formation of Antibody

We must now ask ourselves : How does the body manage to pro-
duce relatively large amounts of globulin molecules, so precisely
adapted to combining with definite chemical groupings?

It is not easy to answer this question. A number of hypothetical
mechanisms of antibody formation have been proposed, of which we
may mention (a) the cast-off receptor theory of Ehrlich, (b) the
template theory of Haurowitz, (c) the template theory of Pauling,
(d) the "trained enzyme" theory of Burnet, and (e) the "natural
selection" theory of Jerne, which is supported by Talmage and
Lederberg.

(a) According to the theory of Ehrlich, antibodies are simply

30 INTRODUCTION TO IMMUNOCHEMICAL SPECIFICITY

natural preformed receptors of the body cell for various chemical
groupings. When the number of such chemical groupings coming in
contact with the cell is increased (antigenic stimulus), an excess of
such receptors is formed. Some are cast off into the circulation and
constitute circulating antibody. This theory was given up when it
was found that antibodies could be formed against artificial groupings
with which the organism had never come in contact in the course
of its evolution and for which it could hardly be expected to possess
preformed receptors.

(b) According to Haurowitz (1953), a template, which (as the
result of the presence of a molecule of antigen) reflects in reverse
the significant portions of the structure of the antigen held in the
expanded configuration by polar forces of a molecule of nucleic acid,
attracts to itself molecules of amino acids from which a duplicate of
itself is built up and cast off into the circulation. This theory seems to
require the persistence of small amounts of antigen throughout anti-
body formation, although this might not strictly be a necessary part
of the theory.

(c) Pauling's theory (1940) is a modification of that of Hauro-
witz and differs mainly in Pauling's supposition that preformed
normal globulin, becoming unfolded ("denatured") at the ends of the
polypeptide chain (he assumes that they have accessible to them a
number of about equally stable folded configurations), fold up (are
"renatured") on contact with a molecule of antigen and thus become
specific antibody. This theory definitely presupposes the persistence
of antigen.

(d) Burnet and Fenner (1949) suggested that enzymes involved
in the destruction of normal body constituents become adapted to
acting on similar molecules of foreign substances, are self-reproducing,
and continue to multiply after the elimination of the antigen. Anti-
bodies are supposed to be enzymatically inactive partial replicas of
these adapted enzymes. Burnet has apparently more recently changed
his views (Burnet, 1957, 1959).

(e) Jerne (1955) suggested that globulin molecules of a very
wide variety of configurations and therefore of specific reactivities are
continually being produced by the body. Some of these molecules
happen to have configurations complementary to surface groups of
some antigens ; these are the "natural antibodies." When an antigen

ANTIBODIES II 31

enters the circulation, it combines with those molecules which happen
to have the corresponding specificity. These combinations are phago-
cyted and transported to the antibody-forming cells. There the antigen
is dissociated and probably discarded, and the cell — for reasons not
specified — proceeds to make more globulin molecules like those just
introduced. The casting ofif into the circulation of these new specific
globulins constitutes the phenomenon of antibody rise.

Jerne's theory, in spite of having been proposed only recently,
has found considerable favor. Talmage (1957, 1959) considers it
essentially similar to the theory of Ehrlich but suggests that the
replicating elements are cells rather than extracellular protein. Bur-
net (1957) and Lederberg (1959) also support the theory. According
to Burnet, antigen combines with specific receptors on the surface
of lymphocytes and thereby stimulates these particular cells to settle
down and multiply in an appropriate tissue. The result of this
replication of selected cells is the production of more of the type of
globulin molecule with which the antigen combined in the first place.

Burnet and Lederberg both assume that the antibody-forming cells
are "hypermutable," i.e., that normally there are frequent changes in
the types of globulin molecules a cell is genetically capable of pro-
ducing. Thus, all possible types of gamma globulin molecules would
generally be represented in the circulation zinth the exception of those
produced by those cells that happened to combine with antigen while
they were still immature ; this is supposed to result in the elimina-
tion of such cells. This additional assumption is made to account for
the nonproduction of antibodies to antigens of the body itself and for
"acquired immunological tolerance."*

Any attempt to revive the Ehrlich theory must take account of
the objection that antibodies can be formed to antigens for which
the body can hardly be expected to have preformed natural receptors.
Talmage (1959) tries to do this by supposing that sharp specificity,
when observed, results from a mixture of globulin molecules, not all
alike, each with some degree of specificity for the antigen or hapten.
With the help of a diagram (Fig. 2-13) and by thermodynamic calcu-
lations he tries to show how the "information" and net specificity of

* When animals are injected with an antigen during fetal life, or in some
cases shortly after birth, they may be incapable of responding immunologically
to this antigen as adults (see Chapter 3).

INTRODUCTION TO IMMUNOCHEMICAL SPECIFICITY

Globulin C Globulins A, B and C

Fig. 2-13. Two-dimensional diagrams illustrating the concept that the in-
formation and net specificity of a combination of three different globulin mole-
cules may be greater than that of one globulin alone. (Redrawn from Talmage
1959.)

a combination of different globulin molecules could be greater than
those of any one type of globulin alone. Talmage suggests that the
average "monospecific" serum contains ten to 100 different kinds of
globulin molecules and points out that on such a basis the assumption
of about 5000 different possible natural globulins could account for
approximately 3 X 10^-"^ different specificities. Since this number is
larger than the number of electrons the universe is supposed to con-
tain, Talmage believes it is satisfactorily large. In fact, Haurowitz
(1956) estimated that not more than 50,000 different antibodies exist.

References

Avery, O. T., W. F. Goebel, and F. H. Babers, 1932. /. Exptl. Med. 55, 769.
Avery, O. T., M. Heidelberger, and W. F. Goebel, 1925, /. Exptl. Med. 42,
709.

ANTIBODIES II 33

Batson, H. C, I'JSl. ./. Immunol. 66, 7^7.

Boyd, W. C, l'M3, l-'iDidamciitah <if Immunology, 3r(l ed., V)SC^, Interscicnce,

New York.
Burnet, F. M., 1957, Australian J. Sci, 20, 67.
Burnet, F. M., 1959, The Clonal Selection Theory of Acquired Immunity,

Vanderbilt University Press, Nashville.
Burnet, F. M., and F. Fenner, 1949, The Production of Antibodies, 2nd ed.,

MacMillan, Melbourne.
Campbell, D. H., and N. Bulman, 1952, Fortschr. Chem. org. Naturstoffe 9,

443.
Glutton, R. F., C. R. Harrington, and M. E. Yuill, 1938, Biochem. J. 32,

1111.
Finland, M., and E. C. Curnen, 1940, /. Immunol. 38, 457.
Fisher, R. A., 1950. Statistical Methods for Research Workers. 11th cd., Oliver

and Boyd, Edinburgh.
Goebel, W. F., and O. T. Avery, 1929, /. Exptl. Med. 50, 521.
Haurowitz, F., 1953, in A. M. Pappenheimer (ed.). The Nature and

Significance of the Antibody Response, Columbia University Press,

New York.
Haurowitz, F., 1956, /. Cellular Com p. Physiol. 47, Sup pi. 1, 1.
Heidelberger, M., and M. L. Wolfram, 1954, Federation Proc. 13, 496.
Hooker, S. B., and W. C. Boyd, 1933, J. Immunol. 24. 141.
Hooker, S. B., and W. C. Boyd, 1934, /. Immunol. 26, 469.
Hooker, S. B., and W. C. Boyd, 1941, /. Immunol. 42, 419.
Jerne, N. K., 1955, Proc. Nat. Acad. Sci. U. S. 41, 849.
Landsteiner, K., and J. van der Scheer, 1928, /. Exptl. Med. 48, 315.
Landsteiner, K., and J. van der Scheer 1929. /. Exptl. Med. 50, 407.
Landsteiner, K., and J. van der Scheer 1936, /. Exptl. Med. 63, 325.
Landsteiner, K., and J. van der Scheer 1938, /. Exptl. Med. 67, 709.
Landsteiner, K. and J. van der Scheer 1940, /. Exptl. Med. 71, 445.
Pauling, L. D. Pressman, and A. L. Grossberg, 1944, /. Am. Chem. Soc. 66,

784.
Pressman, D., 1957, in L. Pauling and H. A. Itano (eds.). Molecular Struc-
ture and Biological Specificity, American Institute of Biological Sciences,

Washington, D. C.
Springer, G. F., 1958, in G. E. W. Wolstenholme and M. O'Connor (eds.),

Ciba Foundation on the Chemistry and Biology of Mueopolysaceharides, Little,

Brown, Boston.
Talmage, D. W., 1957, Ann. Rev. Med. 8, 239.
Talmage, D. W., 1959, Science 129, 1643.
Yeas, M., 1958, in Y. P. Yockey (ed.), Symposium on Liformation Theory

in Biology, Pergamon, New York.

CHAPTER 3
Antigens

Definition

We use the term antigen in at least two senses. Primarily, an
antigen is a substance which, when introduced into an animal, usually
not by way of the digestive tract, causes the production of specific
antibodies. Immunologists commonly use the term antigen also for
preparations which merely react with antibodies in vitro ; for in-
stance, the mixture of normal tissue lipids used in the Wassermann
test for syphilis is referred to as the Wassermann antigen, although
injection of it into an animal would probably not cause the produc-
tion of syphilitic antibodies.

Immunologists are also imprecise in another way in their usage
of the term antigen. The word is applied to both a purified, supposedly
molecularly homogeneous preparation, as for example crystalline
bovine serum albumin, which, when injected, will cause the pro-
duction of antibodies to this substance, and also, following tradition,
to preparations which, chemically speaking, are complex mixtures,
such as the suspensions of killed organisms which are used in the
practical production of certain types of immunity.

Antigenicity

In spite of many studies on the subject we are not yet in a position
to state positively what physical and chemical characteristics make a
substance antigenic. We can only offer certain rules of thumb: (a)
antigens are substances the molecules of which are larger than mini-

ANTIGENS 35

mum in size, and (b) they must be foreign to the circulation of the
animal in which they stimulate antibody production.

It is not possible to give a definite figure for the minimal molecular
weight which a substance must possess to be an antigen, but good
antigens generally have a molecular weight of not less than 10,000.
It is also found that by adsorbing small molecules particulate material
may become antigenic, just as conjugated antigens may be produced
by coupling simpler compounds (haptens) to proteins. Some simple
substances not conjugated to a protein cause the production of anti-
bodies, but it is believed that they act by first combining with some
of the proteins of the body.

Size alone does not seem to be enough. In addition to being
large, a molecule, to be antigenic, must possess other characteristics.
It has been suggested that a certain degree of internal complexity
may be required, and it has been found that sulfonated polystyrene
(Fig. 3-1), which is a large molecule polymer made up of a single

CH2-CH2-(-CH-CH2).-C = CH.,

SO3H SO3H SO3H

Fig. 3-1. Sulfonated polystyrene.

repeated unit, is not antigenic (Boyd, 1952). Haurowitz (1952) sug-
gested that the necessary feature is a rigid structure of the determi-
nant groups of the antigen. In support of this idea, it has found
that gelatin, a non-rigid molecule, which is ordinarily a very poor
antigen, can be made into a relatively good antigen by being coupled
with chemical groupings which would be expected to increase the
rigidity of the molecule (Hooker and Boyd, 1932; Glutton, Haring-
ton, and Yuill, 1938; Sela and Arnon, 1960). Contrary to earlier
opinion, the introduced groups do not have to be aromatic (Sela
and Arnon, 1960). Haurowitz's suggestion is also supported by
studies on the antigenicity of synthetic polypeptides. Polyglutamic
acid was found to be non-antigenic (Maurer 1957), and most of the
polymers studied by Stahmann and his colleagues (references in
Sela, 1962) were either non-antigenic or poor antigens. On the

36 INTRODUCTION TO IMMUNOCHEMICAL SPECIFICITY

other hand, Gill and Doty (1960) found a synthetic linear polymer
containing tyrosine, which would increase the rigidity of the mole-
cule, to be antigenic, and Sela (1962) found that a multichain
copolymer, in which the chains of polypeptides containing L-tyrosine
and L-glutamic acid were built on a multichain poly-DL-alanine, is
a powerful and sharply specific antigen.

It was once believed that only proteins could be antigenic, but
we now know that some carbohydrates are also good antigens. The
rigid sites in polysaccharide antigens may be the pyranose or
furanose rings (Sela, 1962). Large-molecule carbohydrates vary
in their antigenicity. Pneumococcus polysaccharides are antigenic in
man and in the mouse but not in rabbits (Dubos, 1945). Dextrans,
apparently not antigenic for rabbits, are antigenic in man (Kabat
and Berg, 1952, 1953). Purified blood group substances A and B
are fair antigens in man but not in rabbits (Morgan and van Heynin-
gen, 1944; Kabat, Baer, Day, and Knaub, 1950).

We repeat that substances must be foreign to the circulation to
be antigenic for an animal. The normal animal does not produce anti-
bodies to the protein and carbohydrate constituents of its own blood
or to the tissue components which ordinarily reach the blood. Normal
animals, however, can be induced to form antibodies to constituents
of their bodies which normally do not find their way into the cir-
culation, such as lens protein of the eye and casein, even from an
animal's own milk (Lewis, 1934).

At one time it was believed that only proteins could be antigenic.
We now know that many carbohydrates are also good antigens. Other
classes of antigens exist, but with some possible exceptions all of
them contain some protein or carbohydrate, or both.

Immunological Tolerance

For a long time it was a complete mystery why an animal did not
make antibodies for the proteins and other substances of his own cir-
culation, many of which are good antigens for an individual of a
different species. A clue has recently been found in the phenomenon
of immunological tolerance. If embryos are injected in utcro or early
in postnatal life with an antigen, not only may they not produce
any antiboch' to the antigen, but they may be rendered incapable of
responding to this antigen for the rest of their life (Burnet, 1956)

ANTIGENS Z7

although they will generally respond perfectly normally to other
antigens. This refractory state is called immunological tolerance.

To explain it and related phenomena, modern theories of antibody
formation postulate, as was pointed out in Chapter 1, that the com-
bination of antigen with immature antibody-forming cells results in
the death of these cells or at least their elimination from the body.
It seems likely that the mechanism that produces acquired immunologi-
cal tolerance, whatever it is, accounts also for the failure of the body
to produce antibodies to its own circulating antigens.

It is known that the rejection of tissue transplants from one in-
dividual to another, in contrast to the acceptance of transplants from
another part of the patient's body or from an identical twin, is an
immunological phenomenon. Acquired immunological tolerance has
been strikingly demonstrated in animals by injecting adult tissue
cells into embryos. Such injected embryos, when they are born and
grow up, may accept skin grafts, for example, from a donor of the
stock which provided the injected tissue, something they would not
do if not previously injected during fetal life. Billingham, Brent, and
Medawar (1953) believe that in such cases some of the injected cells
have survived in the recipients, thus accounting for the continued re-
ceptive state for transplants from that stock.

A utoimmunisation

Physiological mechanisms, like other machinery, can go wrong. It is
therefore not too surprising that occasional individuals are suspected
of producing antibodies to their own antigens. This process is called
autoimmunization, and naturally it is not a good thing when it occurs.
In fact, the process may be part of the etiology of a number of
hitherto mysterious diseases, most of them fortunately rare, such as
acquired hemolytic anemia, idiopathic thrombocytopenic purpura,
chronic leukopenia, periarteritis nodosa, lupus erythematosis, and
possibly other diseases.

The production of autoantibodies is made possible by a number
of abnormal factors, which may include (a) modification of one of
the patient's own antigens by combination with a drug, a bacterial
toxin, or something of the sort, so as to make it at least partially
"foreign" to his circulation, and (b) an unusual propensity of the
patient to form antibodies in general.

38 INTRODUCTION TO IMMUNOCHEMICAL SPECIFICITY

Anti(/cnic Determinants

Antibodies combine with the surface of antigenic cells or mole-
cules. Even an antigen molecule of only moderate size can combine
with several molecules of antibody. This shows that the portion of
the antigenic surface toward which the antibody is directed (anti-
genic determinants) is only part of the whole molecule. In Chapter
1 we saw that antibodies can be produced which combine specifically
with relatively small molecules such as arsanilic acid or glucose. The
cjuestion arises : How much of the surface of a antigen molecule is
actually involved in the combination with antibody?

Some information on this question has been obtained, mostly by use
of Landsteiner's inhibition technique. Landsteiner (1942) found
that hydrolysis products of silk, peptides with molecular weights about
600 to 1000, were capable of specifically inhibiting the reactions or
precipitin sera for silk. This work has recently been confirmed by
Cebra (1961), who found that tyrosine forms an important part of
the antigenic determinant in silk fibroin, but that a considerable
length of the glycyl-alanyl chain is also required for detectable
specific combination. Dodecapeptides (MW ca. 900) were the most
active of the peptides compared, giving up to 50 per cent inhibition.
Of the octopeptides tested, Gly (Glys, Alas) Tyr (MW ca. 600)
was the most effective inhibitor and probably represents a major
part of the specific antigenic determinant.

Better evidence comes from reactions with conjugated antigens con-
taining complex haptens. From quantitative studies of the inhibition of
antibodies to simple haptens Hooker and Boyd (1933, 1941) concluded
that the specificity of the antibody was influenced to some extent by
the protein tyrosine or histidine residues with which the diazotized
amines combine. This suggested that the antigenic determinant in
conjugated antigens is not quite as simple a structure as the hapten
alone (Fig. 3-2). Landsteiner (1945) studied the question by coupl-
ing to proteins haptens containing peptides made up of several
amino acids. Goebel, Avery, and Babers (1934) and Kabat (1957)
investigated antibodies directed toward determinants consisting of
several sugar molecules linked together to form an oligosaccharide.
Let us review some of these experiments briefly.

Peptide Determinants. Since some proteins are made up entirely
of amino acids, and since there is no evidence that the specificity of

ANTIGENS

Antibody

5::::; :::::::

i ^»

Y^^^Hj- ;-

:::::::, ::^

ffi

::::: P

::;::■ H^

^H i

-A \

if \

:::::::/ .

■t

r \

Antibody

lirl

Fig. 3-2. Reactions
substances.

anti-gelatin-arsanilic acid antibody with various

proteins containing a small percentage of carbohydrate is affected in
any way by the presence of the carbohydrate, we are forced to con-
clude that the specific antigenic determinants of protein antigens
consist of various combinations of amino acids. Anything we can find
out about the specificity of peptides might therefore apply also to
native proteins.

Landsteiner prepared /'-aminophenyl compounds containing pep-
tides made up of the amino acids glycine and leucine

NH2

I
HCHCOOH

CH3 NH2
I I
CHCH2CHCOOH

CH3

Leucine

Glycine

in various combinations (Fig. Z-Z). These haptens were diazotized
and coupled to protein in the usual way, using one protein for the
immunizing antigens and another for the test antigens to avoid
complicating cross-reactions due to antibody to the protein part of the
antigens.

The strongest precipitation was obtained with the homologous anti-
gen (Table 3-1), but cross-reactions were also obtained, generally
strongest when the terminal amino acid of the peptide in the test
antigen was the same as that in the immunizing antigen. Since these

INTRODUCTION TO IMMUNOCHEMICAL SPECIFICITY

H,NY ^CONHCH.CONHCH.COGH

/>-Aminobenzoylglycylglycine

CH3 CH3

I
CH2

H.,n/ \cONHCH2CONHCHCOOH

/'-Aminobenzoylglycylleucine
CH3 CH3

CH,

1
H.n/ \cONHCHCONHCH2COOH

p-Aminobenzoylleucylglycine
Fig. 3-3.

haptens had an amino acid — COOH at the end, Landsteiner in-
terpreted this as showing the predominant influence of the acid-
carrying group on specificity. Today we are more incHned to attribute
Landsteiner's cross-reactions to the fact that the terminal unit, acid
or not, of a composite hapten has greater influence on the specificity
than any other group does ; for, as we shall see, a similar rule applies
to haptens consisting of oligosaccharides where no acid group is
present.

TABLE 3-1
Cross-Reactions of Glycine and Leucine Haptens*

Antigen containing''

Antibody for G L

LG GGG GGL GGGG

GGGL

GGGGG

GGGGG 0 0
GGGGL 0 +

±
0

+ ±

0 -H± 0 +±
0 0 ++ 0

+ +

Landsteiner, 1945.
G = glycine, L = leucine.

ANTIGENS 41

When Landsteiner tested antibodies to larger peptides, he stiil
found that cross-reactions occurred with peptides having the same
terminal amino acids, but such cross-reactions did not always occur,
and some cross-reactions were found to be due to common amino
acids in other positions. The cross-reactions were definitely related
to similarities of constitution (Table 3-1). For instance, an antibody
for the pentapeptide GGGGG, where G stands for glycine, precipitated
— GG but not — LG antigen, where L stands for leucine, and pre-
cipitated much less — LGG than — GGG. The amount of precipitate
produced by an anti-GGGGL antiserum with various peptide-con-
taining antigens increased in the order — L, — GL, — GGL, — GGGL,
-GGGGL (Table 3-2).

TABLE 3-2
Increase in Strength of Cross-Reactions with Increase in Length of Hapten"

Antigen containing''

Antibody for

GL GGL GGGL

GGGGL

+ ±

++± +++ +++±

+ + -f- +

" Landsteiner, 1945.

'' G = glycine, L = leucine.

The strongest reactions were not always obtained with haptens
having the terminal portions identical with those of the immunizing
hapten. For instance, when Landsteiner prepared antisera against
polypeptides in which the terminal carboxyl group had been con-
verted to the amide (Fig. 3-4) he found than an antiserum for

HjX^^ ^CONHCH2CONHCH,CONH,.

Fig. 3-4. Amide of /'-aminobenzoylglycylglycine.

GGLGGAm reacted with — GGLAm and — GGGGLAm but not
with — LGGAm (Table 3-3), in spite of the fact that the terminal
three units of this last hapten are identical with the terminal three
units of the immunizing hapten. Landsteiner attributed this to a

42 INTRODUCTION TO IMMUNOCHEMICAL SPECIFICITY

TABLE 3-3
Cross-Reactions of Glycine-Leucine-Amino Polypeptides*

Antigen containing''

Antibody for

GGLGGAm

GGLAm GGGGLAm

LGGA

GGLGGAm

+ + ±

+ ± +±

» Landsteiner, 1945.

'^ G = glycine, L = leucine.

failure of the amide groups to have as strong an effect on serological
specificity as the free carboxyl groups have.

Landsteiner obtained evidence that the antibodies to such complex
peptide haptens were at least partly directed toward the whole peptide
and not merely to the component amino acids. For one thing,
varying the order of the amino acids in the peptide made a marked
change, so that — GGL, — GLG, and — LGG were serologically
different, as were — GGGGL, — GGGLGG, and — LGGGG.

Other evidence that antibody is directed toward the whole peptide
was obtained by "absorbing" an antiserum, i.e., by reacting the anti-
serum with heterologous antigens until no further precipitate formed,
and then reacting the absorbed antiserum with hapten. Suitable ab-
sorption of an antiserum for GGLGG left antibodies which reacted
with the homologous hapten but not with related haptens, except for
a slight reaction with —LGG. Tests made for comparison with diluted
antiserum showed that this change in reactivity was not due merely
to diminution in total antibody content.

A third line of evidence came from inhibition experiments. Land-
steiner found that antibodies to a given peptide were generally better
inhibited by homologous than by heterologous hapten, even when they
reacted with a heterologous antigen (Table 3-4).

From the evidence that antibodies can be directed toward the
whole of a peptide containing as many as five amino acids we may
conclude that the antigenic determinants in natural proteins may be
as large as this. Nevertheless, there seems to be a limit to the size
of the antigenic determinant to which the combining group of a
single antibody molecule can be directed, for Landsteiner and van
der Scheer (1938) found that when they used symmetrical aminoiso-

ANTIGENS 43

TABLE 3-4
Inhibition of Heterologous Reaction by Homologous Hapten*

Reaction

in presence

of hapten''

Antibody for Antigen containing

GGG

GGGGG

GGLGG

GGG GGG
GGGGG GGG
GGLGG GGG

±
+

±
±

+ +
0

" Landsteiner, 1945.

•^ G = glycine, L = leucine.

phthalyl glycine-leucine (Fig. 3-5), which they referred to as GIL,
as hapten they obtained two distinct antibodies. One reacted with
w-aminobenzoyl glycine (G) and the other with 77?-aminobenzoyl

NH,

HOOCCH2NOCI IcONHCHCOOH

CH2CH(CH3)2
GIL

Fig. 3-5.

leucine (L) (Fig. 3-6). The anti-G of stich a serum was not re-
movable with antigen containing only L, and the anti-L was not
removable with antigen containing only G. Evidently the two amino

NH2 NH2

I ICONHCHCOOH

CH2CH(CH3)2
L
Fig. 3-6.

acid residues in GIL were too far apart to be spanned simtiltaneously
by a typical antibody determinant, although there was some evidence
for the presence in the antiserum of a slight amount of a special

INTRODUCTION TO IMMUNOCHEMICAL SPECIFICITY

o = ^ ^

« ^ ^ ^

^ "JJ,

ANTIGENS

•42 >>

§1

■;2 O
CO

^ ^ ^ -^

46 INTRODUCTION TO IMMUNOCHEMICAL SPECIFICITY

antibody which might have been directed toward the whole hapten
GIL. At the same time, evidence was obtained that the anti-G and
anti-L of the antiserum produced by injecting the GIL antigen were
not quite identical with those produced by injecting G- and L-coupled
antigens.

Carbohydrate Determinants. Experiments with carbohydrate
haptens have given similar results. Goebel and co-workers (Avery,
Goebel, and Babers, 1932; Goebel, Avery, and Babers, 1934) in-
jected conjugated antigens containing monosaccharides and disac-
charides as haptens (Fig. Z-7). Coupling was done by way of an
aminophenyl group in each case. It will be seen that the cross-reac-
tions of the antisera to the disaccharide haptens occurred mainly with
the terminal sugars. Inhibition experiments showed that the /'-amino-
phenyl glycosides of the terminal sugars (monosaccharides) were
nearly as good inhibitors of the anti-disaccharide sera as the /'-amino-
phenyl glycosides of the disaccharides themselves were. This again
points to a predominant influence of the terminal group of a com-
posite hapten on the antibody produced when an antigen containing
it is injected. Nevertheless, the fact that the /'-aminophenyl glycosides
of the disaccharides were still somewhat better than the disaccharides
alone as inhibitors of their corresponding antisera suggested that the
anti-disaccharide antibodies were to some extent directed toward
the whole hapten and that a carbohydrate hapten could be larger
than a disaccharide.

Kabat (1957) was able to obtain further information on this point
by studying the antibodies produced in human beings by injections
of dextran, a large-molecule polysaccharide produced by certain
bacteria. Dextran appears to be made up entirely of glucose, pre-
dominantly connected by 1-6 linkages (Fig. 3-8). With such a
simple antigen the possible antigenic determinants are merely one
or more glucose units. Finding out how big an antigenic carbohydrate
determinant may have a specifically corresponding antibody determi-
nant is simply a matter of finding out how many glucose units an
oligosaccharide must contain to fill the combining site on the anti-
body. Kabat studied this question by measuring the relative inhibiting
power of glucose, isomaltose (two glucose units), isomaltotriose
(three glucose units), and larger polysaccharides for an anti-dextran
serum acting on dextran. The results are shown in Fig. 3-9. It is

ANTIGENS

(4)
G1(1-6)G1(1-6)G1(1-6)G1(1-6)G1

G1(1-6)G1(1-6)G1(1-6)G1(1-4)G1
(6)

Fig. 3-8. Suggested structure of dextran.

apparent that isomaltose (two glucose units) is distinctly better than
glucose as an inhibitor but that isomaltotriose is much better than
either, suggesting that the antibody determinant corresponds to an
antigenic determinant of at least three glucose units. Actually, the
data suggest that the antibodies can distinguish even isomaltohexaose
(six glucose units) from any smaller antigenic determinant, but the
difference between the pentaose and the hexaose is not great. Kabat
suggests that the hexaose is the largest group capable of entering the
cavity in the anti-dextran antibody molecule.

If the hexasaccharide is accepted as the largest group which can
totally combine with the combining site of the antibody, the contri-
bution of each glucose residue, starting with the terminal unit, to
the total free binding energy between antibody and antigen can be
computed. Such calculations are shown in Table 3-5, taken from

::::-s

GIGIGI
(Isomaltotriose)

GIGIGIGI
(Isomaltotetroose)

GIGIGIGIGI
(Isomoltopentaose)

GIGIGIGIGIGI
(Isomaltohexoose)

Fig. 3-9. Relative inhibitory power of oligosaccharides for anti-dextran serum
acting on dextran. (Redrawn from data of Kabat, 1957.)

48 INTRODUCTION TO IMMUNOCHEMICAL SPECIFICITY

TABLE 3-5

Calculated Contribution to the Free Fnerg>- of Combination of Glucoses in
Reactive Groups of Dextran"

Contribution to
Number of glucose units binding energy, %

5 98

4 95

3 90

2 60

1 (terminal unit) 39

- Kabat, 1957.

Kabat (1957). It will be seen that the terminal glucose contributes
as much as 39 per cent of the total binding energy. The first five units
together contribute 98 per cent of the binding energy, leaving only
2 per cent to be contributed by the sixth glucose residue.

These calculations give us a fairly good idea of the size of the
antigenic determinant in a typical carbohydrate antigen. We shall
see later how this information can be applied to practice.

References

Avery, O. T., W. F. Gocbel, and F. H. Babcrs, 1932, /. Exptl. Med. 55, 769.

Billingham, R. E., L. Brent, and P. B. Medawar, 1953, Nature 172, 603.

Boyd, W. C, 1952, Unpublished experiments.

Boyd, W. C, and P. Doty, 1958, Unpublished experiments.

Burnet, F. M., 1954, Proc. Roy. Soc. (London) B 146, 1.

Cebra, J .J., 1961, /. Immunol. 86, 205.

Glutton, R. F., C. R. Harington, and M. E. Yuill, 1938, Biochem. J. 32, 1111.

Dubos, R. J., 1945, The Bacterial Cell in its Relation to Problems of Viru-
lence. Immunity and Chemotherapy, Harvard University Press, Cambridge.

Gill, T. J., and p! Doty, 1960, /. Mol. Biol. 2, 65

Goebel, W. F., O. T. Avery, and F. H. Babers, 1934. /. E.vptl. Med. 60, 599.

Haurowitz, F., 1952, Biol. Rev. 27, 247.

Hooker, S. B., and W. C. Boyd. 1932, /. Immunol. 23, 465.

Hooker, S. B., and W. C. Boyd, 1941, Unpublished data.

Hooker, S. B.. and W. C. Boyd, 1933, /. Immunol. 25, 61.

Kabat, E. A. 1957, /. Cellular Comp. Physiol. 50, Suppl. 1, 79.

Kabat, E. A., H. Baer, P. I. Day, and V. Knayb, 1950, .1. E.vptl. Med. 91,
433.

ANTIGENS 49

Kabat, E. A, and D. Berg, 1953, /. Inimuiwl. 70, 514.

Kabat, E. A, and D. Berg, 1952, Ann. N. Y. Acad. Sci. 55, 471.

Landsteiner, K., 1942, /. Exptl. Med. 75, 269.

Landsteiner, K., 1945, The Specificity of Serological Reactions 2nd rev. ed..

Harvard University Press, Cambridge.
Landsteiner, K., and J. van der Scheer, 1938, /. Exptl. Med. 67, 709.
Lewis, J. N. H., 1934, /. Infectious Diseases 55, 203.
Maurer, P. H., 1957, Proc. Soc. Exptl. Biol. Med. 96, 394.
Morgan, W. T. J., and R. van Heyningen, 1944, Brit. J. Exptl. Pathol. 25, 5.
Sela, M., 1962, Paper contributed to a symposium on poly-alpha-amino acids.

To be published by the Wisconsin Univ. Press.
Sela, M., and M. Arnon, 1960, Biochem. J. 75, 91.

CHAPTER 4
Blood Groups

ABO Blood Groups

It does not take profound knowledge of science to realize that no
two human beings, with the possible exception of identical twins,
are exactly alike. There are sometimes strong resemblances in families,
and sometimes even unrelated persons look enough alike to be mis-
taken one for the other by those who do not know them well, but
close associates are very seldom deceived. Features, voice, movements,
and modes of response nearly always distinguish each human being
from all others in the world.

If we believe, and the belief hardly needs defending today, that
structural and functional differences between individuals, aside from
the effects of accidents resulting in scars or deformity and from
learned behavioral patterns, are due to underlying biochemical dif-
ferences, we should not be surprised to find that between different
individuals of the same species biochemical differences can also be
demonstrated. Surprisingly enough, this was done for the first time
at the beginning of the present century. Karl Landsteiner, then work-
ing in Vienna, discovered that not all normal human blood is alike.
Landsteiner and his pupils showed that human beings could be
classified into four groups on the basis of the reactions of their blood
with that of other normal individuals. This discovery made blood
transfusion a safe and practical procedure for the first time and had
great influence on the study of serological specificity.

Landsteiner's discovery consisted of the observation that, when
the bloods of certain individuals were mixed, the red blood cor-

BLOOD GROUPS 51

puscles adhered to each other and formed ckmips that under the
microscope looked something hke bunches of grapes. (See Fig. 1-2,
p. 3). In strong reactions all the red cells in the preparation stuck
together, leaving a clear supernatant fluid. Landsteiner showed that
this behavior could be explained by assuming that there may be two
reactive substances in the erythrocytes and two corresponding anti-
substances in the plasma which react with the erythrocyte substances.
The substances in the erythrocytes have been shown to be antigens ;
they are also called isoagglutinogens. The substances in the plasma
or serum which combine with them and thus cause agglutination have
all the properties of agglutinating antibodies, and are called isoagglu-
tinins.

It is obvious that the substance in the plasma which combines
with the erythrocytes of another individual and causes them to ag-
glutinate could hardly coexist in the same blood stream with the
corresponding agglutinogen ; for, if it did, an individual's plasma
would agglutinate his own erythrocytes. The rule, first stated by
Landsteiner and known by his name, is that those agglutinins will
be present which can coexist with the agglutinogens present in the
cells. The agglutinogens, called arbitrarily A and B, can be present
in the cell singly or together, or can both be absent. This gives us
the combinations of the four classical blood groups. (Table 4-1).

Landsteiner's discovery explained why transfusions of blood from
one individual to another had previously only occasionally been suc-
cessful. If a blood donor is selected at random, the chances of obtain-
ing one whose blood group is compatible with that of the recipient
are not good. It was only when pretransfusion blood grouping be-
came a routine that l:)lood transfusion became a safe and reliable
procedure.

TABLE 4-1
The Landsteiner Blood Groups

Blood group Antigens in cells Agglutinins in plasma

—

Anti-A and Anti-B

Anti-B

Anti-A

A and B

—

52 INTRODUCTION TO IMMUNOCHEMICAL SPECIFICITY

If a donor of the same blood group as the recipient is chosen, it
is ol)vious that the chances of a successful transfusion are good,
unless other blood factors yet to be discussed come into operation,
or unless the technique is faulty. Transfusion consists in introducing
a relatively small amount of the donor's blood, 500 ml. or less, into
the circulation of the recipient. This means that the donor's cells are
exposed to a large amount of whatever agglutinins the recipient
possesses, in full concentration. If the donor's cells contain an ag-
glutinogen capable of reacting with the recipient's agglutinin or
agglutinins, the donor's cells may be agglutinated, and a serious or
even fatal transfusion reaction may result.

It is not, however, always necessary to use a donor of exactly the
same blood group as the recipient. Introduction of a donor's ag-
glutinin which could react with the recipient's cells is often not
serious, for the agglutinin gets diluted by the recipient's plasma and
is also partly neutralized by soluble blood group substance in the re-
cipient's plasma and in his tissues. This means that, in general, trans-
fusions in the directions shown in Fig. 4-1 are possible, although
it is always preferable to use a donor of the same group as the
patient.

t
AB

Fig. 4-1. Theoretical possibilities of transfusion, based on blood groups of

donor and recipient.

Blood groups are inherited. Parents with any given combination
of blood groups may produce children of certain blood groups but
not of others, except that the mating of A X B may produce children
of any of the four groups. Inheritance is based on three allelomorphic
genes. A, B, and O, which can occur in any combination of two :
OO, AA, AO, BB, BO, or AB. The blood groups of individuals
of genotype AO are the same, so far as we can tell in the laboratory,

BLOOD GROUPS 53

as those of individuals of genotype AA. The same holds for BB
and BO. Consequently, we have to classify both AA and AO as
group A, and BB and BO as group B, giving the four Landsteiner
blood groups, as shown in Table 4-2.

TABLE 4-2

Genetically Determined Types (Genotypes) and Serologically Determined Types
(Blood Groups)

Approximate percentage of
Genotype Blood group U.S.A. population

00 O 45

aaI

AOl

bb|

BOj

AB AB 3

A 42

B 10

It was soon found that there were two kinds of A antigen. The
more common one, and in Asian populations the only one present,
reacts strongly with anti-A agglutinins and is designated as Ai. The
other, confined to Europeans and Africans and their descendents
in other parts of the world, often reacts weakly with anti-A and is
called A2. This distinction enables us to divide the population of
Europe and Africa into six blood groups instead of four, as follows :
O, Ai, A2, B, AiB, and A2B. The difference seems to be of little
importance for transfusion but is interesting to anthropologists and
students of legal medicine.

Althotigh in Table 4-1 group O erythrocytes are shown as having
no antigen, this is not strictly true. They possess antigens connected
with other blood group systems still to be discussed and also have
an antigen connected with the ABO blood group system. Human
plasma does not ordinarily contain an agglutinin for this antigen,
but the plasma of individuals of the subgroup AiB and the normal
serum of certain animals may contain one. The agglutinin can be
found in the serum of certain eels, apparently more regularly in the
European than the American eel, and may be produced by immuniz-
ing a goat with SJiiga bacilli. None of these sources is always availa-
ble, nor is the agglutinin so obtained always strong and reliable. It

54 INTRODUCTION TO IMAIUNOCHEMICAL SPECIFICITY

was therefore a considerable advance in blood grouping technique
when it was discovered that saline extracts of the seeds of certain
plants, such as Ulex europeus, which grows wild in Western and
Southern Europe and in North Africa, contain an agglutinin specific
for this antigen of the group O erythrocytes (Cazal and Lalaurie,
1952; Boyd and Shapleigh, 1954a). This plant agglutinin has ap-
parently replaced all other reagents in this application.

At first it was believed that the agglutinogen detected by this
agglutinin, whatever the source of the agglutinin, was an O antigen
which had the same relation to the O gene as the B antigen has to
the B gene. The agglutinin was therefore called anti-O. It was soon
found, however, that erythrocytes of the subgroups A2 and A2B are
also agglutinated by the agglutinin, A2 cells being affected about as
strongly as O cells. This is apparently true even when the genotype of
the Ao individual is A2A2, so that no O gene is present. It seemed im-
proper to retain the name anti-O for a reagent that detects an anti-
gen produced by both the O and Ao genes. Following the practice
of Morgan (Morgan and Watkins, 1948), the term anti-H is now
generally used for the agglutinin and the term H for the antigen it
detects.

Taking account of this and other discoveries about the blood
groups, we may revise Table 4-1 (see Table 4-3).

TABLE 4-3
Subgroups of Landsteiner Blood Groups

Blood group

Subgroup

Antigens in cells

Agglutinins in plasma

Anti-A, anti-Ai, anti-B

(A.
\A.

[A,

fAnti-B (sometimes anti-H)

IA2 + H

(Anti-B (sometimes anti-Ai)

Anti-A, and anti-Ai

fA,B

/Ai-fB

/(Sometimes anti-H)

IA2B

IA2 + B

\ (Sometimes anti-Ai)

Secretors and Nonsecretors

The antigens of the ABO blood group system are not confined to
the erythrocytes. They may occur in practically all tissues and fluids
of the body, with the probable exception of the central nervous sys-

BLOOD GROUPS 55

tern. They may occur in two forms : water soluble and lipid soluble
(i.e., soluble in lipid solvents such as alcohol-ether mixtures and
chloroform). All individuals apparently have the lipid-soluble form
in their tissues, in conformity with their blood group (Boyd and
Boyd, 1937). The water-soluble form, however, is found in only
about 85 per cent of European individuals. Such persons are called
secretors, and those in whose tissues and body fluids water-soluble
antigens corresponding to their blood group are not found are called
nonsecretors (Schlff and Sasaki, 1932). The ability to secrete the
A and B antigens in water-soluble form is inherited, being con-
trolled by a pair of genes S and s.

The saliva of all secretors, no matter what their group, contains
enough of the H antigen to make it possible to diagnose such in-
dividuals by the inhibition technique with an anti-H reagent such
as Ulex extract (Boyd and Shapleigh, 1954b). Saliva of group O
secretors is richest in H antigen, and, according to Race and Sanger
(1958), some AiB salivas may not contain enough H antigen to
make the use of Ulex extracts reliable for the diagnosis of secretors in
this subgroup.

The above sketch does not by any means give an adequate picture
of the ABO blood group system, which is one of the most complicated
known in man. A good discussion is given in the book by Race and
Sanger (1948). We may simply mention that a number of other
variants of the A antigen are known, all of them fortunately rare,
and that genes exist, also rare, which can modify the expression of
the ABO genes. The "Bombay" gene, x, when present in double
dose, XX, prevents the development of antigens B and H ; whether
it also suppresses A is not yet known. There seems to be another
gene, y, which when present in double dose, yy, modifies the de-
velopment of the A antigen in the red cells and, to a much lesser ex-
tent, in the saliva. Variants of the B antigen have also been observed.

Blood Groups of the Ancient Dead

Antigens A and B are much more stable than most protein antigens.
(The chemical nature of the A and B antigens will be discussed in
Chapter 7.) It is comparatively easy to demonstrate A and B in
dried tissue, boiled erythrocytes, or tissues which have been pre-
served in formaldehyde. These facts led Boyd and Boyd (1934,

56 INTRODUCTION TO IMMUNOCHEMICAL SPECIFICITY

1937) to attempt to demonstrate A and B in mummified human
remains. The attempt seemed to be successful. These workers tested
more than 300 specimens, mostly from Egypt in the Old World and
from Mexico and Peru in the New World. The technique, though
exacting and at times even exasperating, was simple in theory ; pul-
verized dried tissue (usually muscle) was mixed with carefully
titrated anti-A and anti-B agglutinins and the mixture tested after
a suitable interval for evidence of removal of one or more of the
agglutinins. Removal of anti-B was considered to indicate the pres-
ence in the tissue of the B antigen, removal of anti-A the presence
of A. Removal of neither suggested either that the specimen came
from an individual of group O or that any antigens originally present
had deteriorated. Removal of both anti-A and anti-B suggested group
AB or nonspecific destruction or removal of agglutinins or anti-
bodies in general.

The results obtained were on the whole in line with the present
distribution of the A and B antigens in human races, confirming
the antiquity of the ABO blood group system. (Some authors had
suggested, amazingly, that the A and B genes were of recent origin.)
The B antigen was apparently found in pre-Columbian specimens
from Mexico (Taylor and Boyd, 1943), a finding which, if ever
confirmed, might support the suggestion, made on other grounds,
that the B gene was eliminated in the aboriginal inhabitants of
America by natural selection (Boyd, 1959). This subject has been
reviewed by Smith (1960).

Origin of Isoagglutinins Anti-A and Anti-B

The presence of anti-A and anti-B in normal human plasma seems
at first glance to be an exception to the rule that antibodies to blood
group antigens do not usually appear without some history of un-
usual antigenic stimulus. The exception is a marked one, for the
occurrence of these isoagglutinins is very regular. When anti-A
or anti-B, if expected according to Landsteiner's rule, are absent,
there is usually a special explanation, as Race and Sanger point out.
It is natural to ask why these agglutinins appear with such regularity.
There have been two main theories.

According to one theory, the isoagglutinins anti-A and anti-B are a
result of the action of the ABO genes just as the ABH antigens are.

BLOOD GROUPS 57

This theory has been supported l)y Furiihata (1927) and tlie Wurni-
sers ( Fihtti-Wurmser et al., 1954). Whether this theory seems
plausible depends partly on which theory of antibody formation we
happen to believe.

Another theory suggests that anti-xA. and anti-B are immune anti-
bodies, as most other agglutinins are, being formed in response to
antigens, in food, in bacteria, and in animal parasites, which are
chemically similar to the A and B antigens of man. It is known that
a number of such related antigens exist ; in Chapter 2 the cross-
reaction of blood group A antigen and type 14 pneumococcus was
mentioned.

If the second theory of isoagglutinin formation is correct, one
wonders why natural isoagglutinins for other human blood antigens,
such as M, N, and Rh, are so seldom encountered. One possible rea-
son is the lower antigenicity of these agglutinogens ; some evidence
for this exists. Another reason might be that the human agglutino-
gens M, N, and Rh are more unusual in their structure than A and B,
so that closely related antigens in lower organisms and in food, serv-
ing as stimuli for the formation of anti-M, anti-N, and anti-Rh anti-
bodies in man, are only rarely encountered. Later on I shall mention
some recent evidence in support of this speculation.

MNS Blood Groups

Although unknown for so long, the existence of the ABO blood
groups was relatively easy to demonstrate because of the normal
presence of the isoagglutinins anti-A and anti-B. However, over a
quarter of a century went by before any other blood group system
was discovered in man. In 1927, Landsteiner and Levine demonstrated
the existence of M and N antigens (also P) by the use of absorbed
sera of rabbits injected with human erythrocytes. Because anti-M
and anti-N isoagglutinins rarely occur and because patients are not
readily immunized to these antigens, they have little importance for
transfusion ; their applications have been mainly to legal medicine
and to anthropology.

A number of variants of the M and N antigens and, in addition,
other antigens associated with the M and N factors in inheritance
have been discovered. Two of these, S and s (to be distinguished

58 INTRODUCTION TO IMAIUNOCHEMICAL SPECIFICITY

carefully from the Ss gene pair which controls the secreting phe-
nomenon in the ABO system) are fairly common nearly everywhere
and add greatly to the anthropological usefulness of the system.
Two others, called Hunter and Henshaw after the donors in whom
they were first found, are not too common in Africans and are
virtually unknown in persons of European descent.

Distribution of M and N in the Body

Boyd and Boyd (1934) were not able to demonstrate M and N
in human tissues with the technique which they had devised for
A and B. Kosyakov and Tribulev (1939, see also Kosyakov 1954)
devised a method by which M and N could be demonstrated. Their
work was confirmed by Boorman and Dodd (1943). Whereas Boyd
and Tayian (1935) could not detect M and N in boiled erythrocytes,
Kosyakov (1954) was able to do so, and also showed these antigens
to be heat stable.

Rh Groups

In 1939. Levine and Stetson reported a case of erythroblastosis
fetalis and ascribed this disease of the newborn to sensitization of
the mother to a blood antigen the fetus had inherited from the
father. It is now known that this proposed explanation of the disease
is correct and that the antigen operating in the case described was
one of those now known as Rh. Levine and Stetson, however, did
not propose any name for the new blood factor (on such a slender
hair sometimes dangles the apple of priority). It was not until
Landsteiner and Wiener (1940) discovered that serum from one of
their rabbits immunized with rhesus erythrocytes detected a new
factor in human blood that the term Rh was introduced. The new
factor still might not have attracted any more attention than had
others previously reported had not Wiener and Peters (1940) shown
that certain transfusion reactions were due to sensitization to Rh
and Levine et al. (1941) shown that Rh incompatibility between
mother and fetus could be the cause of erythroblastosis fetalis.

Reduced to the simplest terms, the way Rh incompatibility can
cause erythroblastosis fetalis is apparently this : The mother is Rh-
negative, and the fetus inherits the Rh blood factor from his father.

BLOOD GROUPS

60 INTRODUCTION TO IMMUNOCHEMICAL SPECIFICITY

Fetal antigen-containing erythrocytes, or perhaps in some cases
merely antigen from disintegrating fetal erythrocytes, passes the
placental barrier or crosses minute breaks in the placental capillaries
and starts an immune response in the mother. This results in the ap-
pearance of anti-Rh antibodies in the maternal circulation. These
diffuse through the placental barrier back into the fetal circulation
(Fig. 4-2) where they combine with the fetal erythrocytes causing
red cell destruction and, in severe cases, anemia, jaudice, edema,
and death.

It should be mentioned that erythroblastosis fetalis, although seri-
ous when it occurs, is rare. In the United States it occurs in about
one out of 400 births and rarely affects the first pregnancy, even
when Rh incompatibility exists.

The Rh blood groups are much more complicated than would the
case be if there were simply two possibilities : Rh-positive (the
erythrocytes contain the Rh antigen) or Rh-negative individuals
(the erythrocytes do not contain the Rh antigen). Our knowledge
of the Rh blood groups, which extends now to five basic Rh antigens,
is the result of the work of many able and industrious researchers.
Only a simplified outline of it can be given here.

Although it does not have priority, the notation devised for the
Rh antigens by Fisher and Race (see Race, 1944) has proven clearer
and more convenient than Wiener's notations. The five basic Rh
antigens are, according to Race, designated as C, c, E, e, and D.
Genetically C and c are alleles, and so are E and e. It is not yet cer-
tain whether this means that three closely linked loci are involved in
their mechanism of inheritance, as the British workers believed, or
a series of alleles at one locus, as Wiener continues to maintain. In
any case, the allelic behavior of C and c and E and e is an important
phenomenon for their application in legal medicine. The allele cor-
responding to D is d, but no d antigen has been demonstrated with
certainty.

The number of Rh blood groups that can l^e distinguished de-
pends on the number of anti-Rh agglutinins available. Anti-e is seldom
available and anti-c is not always easy to get for routine determina-
tions. If all five agglutinins are at hand, three different blood groups
can be distinguished with respect to the C locus, namely C-positive
c-negative, C-negative c-positive, and C-positive c-positive. (Since

BLOOD GROUPS 61

C and c hchave like alleles, the possibility of a C-negative c-negative
grouping does not exist.) Use of anti-E and anti-e further sub-
divides each of these three types into three others, and the use of
anti-D subdivides all types once more into D-positive and D-negative.
Thus, eighteen types (3 X 3 X 2 = 18) of Rh antigen can be
distinguished. If anti-e is not available, as is usually the case, he
number of Rh types becomes twelve (3 X 2 X 2 = 12) (Table 4-4).
Variants of these antigens, described by the British workers as
further alleles, have been found, and antigens which seem to be due
to other gene loci on the same chromosome are known. A rare type
of blood, classified D, which contains only the D antigen, has been
found.

TABLE 4-4
Reactions of the Twelve Rh Blood Groups Distinguished by Four Anti-Rh Sera

Reaction

with serum

Group

Anti-C

Anti-D

Anti-E

Anti-c

L cde

2. cdE

3. cDe

4. cDE

5. Cde/c

6. CdE/c

7. CDe/c

8. CDE/c

9. Cde/c

10. CdE/c

-f

11. CDe/c

12. CDE/C

-f-

The frequencies of the occurrence of Rh antigens vary widely in
different populations. The Rh-negative type (cde) is most strikingly
absent from Asian and Pacific populations and the American Indians.
It has its highest frequency in the Basques, a population in certain
regions of France and Spain, and certain inhabitants of Switzerland.
The Basques speak a non-Indo-European language and are known
to represent the remnant of an earlier European population which
was once dispersed over a much wider area, including perhaps North

62 INTRODUCTION TO IMMUNOCHEAIICAL SPECIFICITY

Africa. The D antigen is most frequent in African populations and
is so much more common there that it could almost be called an
African antigen.

Stability of Rh Antigens
The chemical nature of the Rh antigens is still completely unknown
(see, however, Chapter 7). The antigens appear to be less stable than
A, B, M, and N; at least Kosyakov (1954) reported that they are
destroyed by boiling for 10 minutes.

Other Blood Groups

Once laboratories were set up to examine routinely human sera
which showed atypical agglutination reactions, the discovery of
other human blood groups followed rapidly. It is doubtful if all
have yet been reported. Those already demonstrated have generally
been named after the donor in whose blood the new antigen or anti-
body was first identified. They have names such as Lutheran, Lewis,
Duffy, Kell, and Kidd. A blood factor which may possibly be different
from any of these has been found with the aid of a plant agglutinin
from peanuts (see Chapter 5).

References

Boyd, W. C, 1959, /. Med. Educ. 34, 398.

Boyd, W. C, and L. G. Boyd, 1934, /. Immunol. 26, 489.

Boyd, W. C, and L. G. Boyd, 1937, /. Immunol. 32, 307.

Boyd, W. C, and E. Shapleigh, 1954a, /. Lab. Clin. Med. 44, 235.

Boyd, W. C, and E. Shapleigh, 1954b, Blood 9, 1195.

Boyd, W. C, and E. Tayian, 1935, /. Immunol. 29, 511.

Boyd, W. C, D. M. Green, D. M. Fujinaga, J. S. Drabik, and E. Waszczenko-

Zacharczenko, 1959, Vox Sanguinis 4, 456.
Boorman, K. E., and B. E. Dodd, 1943, /. Pathol. Bacterial. 55, 329.
Cazal, P., and M. Lalurie, 1952, Acta Haematol. 8, 73.
Filitti-Wurmser, S., et al., 1954, Ann. Eugenics 18, 183.
Furuhata, T., 1927, Japan Med. World 7, 197.
Kosyakov, P. N., 1954, Antigennye Veshchestva Organismu i ikh Znachenie v

Biologii i Meditsine (Antigenic Substances of the Body and Their Significance

in Biology and Medicine), Medgiz, Moscow.
Kosyakov, P. N., and G. P. Tribulev, 1939, J. Immunol. 37, 283.
Landsteiner, K., and A. S. Wiener, 1940, Proc. Sac. E.rptl. Biol. Med. 43, 223.

BLOOD GROUPS 63

Levine, P., and R. E. Stetson, 1939, /. Am. Med. Assoc. 113, 126.

Levine, P., P. Vogel, E. M. Katzin, and L. Burnham, 1941, Am. J. Obstcf.

Gynocol. 42, 925.
Morgan, W. T. J., and W. M. Watkins, 1948, Brit. J. E.vptl. Pathol. 29, 159.
Race, R. R., 1944, Nature 153, 771.

Race, R. R., and R. Sanger, 1958, Blood Groups in Man, Blackwell, Oxford.
Schiff, F., and H. Sasaki, 1932, Klin. IVochschr. 11, 1426.
Smith, M., 1960, Science 131, 699.
Taylor, W. W., and W. C. Boyd, 1943, Year Book Am. Phil. Soc, p. 178,

Year Books, New York.
Wiener, A. S., and H. R. Peters, 1940, Ami. Infernal. Med. 13. 2306.

CHAPTER 5
Plant Agglutinins (Lectins) I

Specificity of Proteins Other Than Antibodies

Antibodies are not the only large molecules with specific biological
activity. Enzymes (also proteins) and hormones, many of which are
proteins, also exhibit this phenomenon. Enzymes, outstanding ex-
amples in this respect, show various degrees of specificity. Some en-
zymes, such as barley maltase and succinic acid dehydrogenase, are
very specific, catalyzing one reaction and only one reaction. Other
enzymes are specific for a particular chemical grouping in their sub-
strate. Enzymes catalyzing reactions in which optically active sub-
stances such as sugars or amino acids are involved frequently act
primarily or exclusively on one of the enantiomorphs. Even enzymes
such as trypsin attack only certain linkages in their substrates. En-
zymes can be inhibited by an excess of one of the products of the
reaction they catalyze, a behavior reminiscent of the specific inhibition
of antibody-antigen reactions by haptens.

Although enzymes resemble antibodies in many ways, there are
striking differences. An enzyme combines with its substrate and then
catalyzes a chemical reaction in which the substrate is involved. The
result is often complete destruction of the substrate. Antiliodies, on
the other hand, have no known catalytic activity and do not them-
selves cause chemical changes in the antigens with which they com-
bine.

Other proteins exhibit specific comljining power. Serum albumin
has the power of binding certain dyes and a number of other natural
and synthetic substances (Klotz, Walker, and Pivan, 1946; Karush,
1950).

PLANT AGGLUTININS (LECTINS) I 65

Grabar (1947) suggests that the power of plasma proteins to
combine with various substances explains one of their important
roles — that of carrier (transportcitr). According to his view, the
lipid carrying role of beta globulins is analogous to the function of
antibodies in their union with antigens.

Plant Agglutinins

Possession of proteins capable of such firm and relatively specific
combination with other substances is not confined to the higher ani-
mals. It has long been known that extracts of certain plant seeds will
bring about the agglutination of animal erythrocyte suspensions to
which they are added. In fact, the agglutinative action of extracts of
the castor bean, Ricinus communis (Table 5-1), was discovered be-

TABLE 5-1

Agglutination of Animal Erythrocytes by Ricin*

Extract of Rich

nus communis, diluted

Red cells from

1:64

1:128

1:256

1:512

1:1024

1:2048

1 :4096

1:8192

Rabbit

1.5

Cat

1.5

0.5

Man

1.5

0.5

Chicken

1.5

Rat

^ Numbers indicate degree of agglutination; 4 indicates complete agglutination,
0.5, weak agglutination, 0, no agglutination.

fore agglutinins for erythrocytes were demonstrated in the blood
of animals (Lau, 1901), and was described soon after bacterial ag-
glutination. Extracts of other plant seeds, such as Abrus precatorius
and certain other Leguminosae, were shown to have similar action.
Because these plant agglutinins act on the red cells of several
animal species they were called nonspecific by many workers. Yet
Landsteiner (1945) observed that the substances are not without a
certain degree of specificity. He illustrated this fact with a little table
in his book on the specificity of serological reactions (Table 5-2). It

66 INTRODUCTION TO IMMUNOCHEMICAL SPECIFICITY

TABLE 5-2
Titers of Different Plant Agglutinins for the Red Cells of Different Species*

Titer for blood of:

Agglutinin from:

Rabbit

Horse

Pigeon

Beans

Lentils

A brus precatorius

Castor beans

125
160

128
4

2000

256

512

■^ After Landsteiner, 1945.

will be seen that some of the seed extracts show specificity, since
they agglutinate the erythrocytes of one species more strongly than
those of another and in one case do not agglutinate blood of a
certain species at all.

Blood Group-Specific Plant Agglutinins
One day toward the end of 1945, looking at this table (Table 5-2)
in the new edition of Landsteiner's book, I was seized with the
idea that if such seed extracts could show species specificity, they
might even show individual specificity ; that is, they might possibly
affect the red cells of some individuals of a species and not affect
those of others of the same species. I asked one of my assistants to
go out and buy dried lima beans. Why I said lima beans instead of
pea beans or kidney beans I shall never know. But if we had bought
practically any other kind of bean we should not have discovered any-
thing new.

The lima beans were ground and extracted with salt solution.
The extract agglutinated erythrocytes of some human individuals
very strongly, btit those of others only weakly if at all. It was im-
mediately evident that the differences were correlated with blood
groups (Table 5-3). The agglutinin from lima beans is almost com-
pletely specific for the blood group A antigen.

This discovery was made so easily that I was disarmed. The whole
process of thinking of the experiment, obtaining the materials, and
testing the idea were the events of perhaps two hours. So it is not
surprising, perhaps, that I failed to realize the importance of the
observation and did not immediately follow it up. I was also then in

E\\^

+ + + +

WCB

+ + + +

PLANT AGGLUTININS (LECTINS) I 67

TABLE 5-3
Test of Lima Bean Extract (December 10, 1945)

Reaction of extract with cells of group:

A BO

LH ± BD 0

BR 0

SJ 0

ON 0

BA 0

CTS 0

the process of writing the second edition of my book on immunology.
I did inckide a short and rather obscure reference to the observation
in the new edition (Boyd, 1947).

After about two years, I returned to the study of the plant ag-
glutinins. In 1949, I published a report on 262 varieties of plants be-
longing to sixty-three families (Boyd and Reguera, 1949). Of
these plants, 191 showed no agglutinating activity. Some agglutinated
human erythrocytes of all blood groups. Extracts of certain varieties
of Phascolus limensis and Ph. lunatus agglutinated strongly only
blood of groups A and AB. One species only, Vitis aestivalis, gave a
weak reaction only with B, but I have not been able to reproduce
this result w^ith later material.

Meanwhile, in 1948, Renkonen had published a paper dealing with
independent studies on fifty-seven species belonging to twenty-eight
genera. Among the blood group-specific plants he studied were Vicia
cracca, specific for A, and Laburnum alpiniim, Cytisiis scssijolius,
and Lotus tctragonolobus, specific for H.

A number of laboratories are now engaged in the study of these
interesting substances. Reviews have been published by Krtipe
(1956), Makela (1957), and Bird (1959). Seeds of a number of
plants are reported to contain anti-A; that of Dolichos biflorus re-
acts so much more strongly with Ai than with Ao as to be virtually
specific for Aj (Bird 1951). An anti-N has been found in Vicia
graminea (Ottsooser and Silberschmidt, 1953) and, more recently,
in Bauhinia purpurea (Makela, 1957; Boyd, Everhart, and McMaster,
1958). An anti-M is on the market.

68 INTRODUCTION TO IMMUNOCHEMICAL SPECIFICITY

I shall discuss the nature of the plant agglutinins in Chapter 6.
In the meantime may I anticipate by saying that although it is proper
to refer to them as agglutinins, as has been done for more than half
a century, it would be begging the question to refer to them as anti-
bodies. In fact, I do not believe that this is what they are. I have
therefore suggested referring to these proteins as lectins, from the
Latin legere, t» pick out or choose (Boyd and Shapleigh, 1954a), in
order to call attention to their specificity without implying any as-
sumptions concerning their origin. I suggested that the term be
applied also to those normal antibodies supposedly not the result
of antigenic stimulus. But, if Jerne's "natural selection" theory of
antibody formation proves to be correct (Chapter 2), it may turn
out that the difference between "normal" and "immune" antibodies
is not as great as has been thought. In that case the term lectin may
come to be restricted to antibody-like plant proteins. In fact, there
already seems to be a tendency to use the word in this way.

No good anti-B lectin is routinely available. Extracts of Sophora
japonica agglutinate blood of group B more strongly than that of
group A but react too strongly with A to be satisfactory as a labora-
tory anti-B reagent (Kriipe, 1953). Euonymus eiiropeiis extracts
contain anti-B and anti-H (Schmidt, 1954). Marasmius oreades,
which sometimes furnishes a satisfactory, though weak, anti-B, is
a small mushroom not commercially available (Elo, Estola, and
Malmstrom, 1951). The best anti-B is said to be that from Bandeiraea
simpUcijoUa (Makela and Makela, 1956), although the samples of
this plant I have tested have been disappointing.

Because of the absence of an anti-B, lectins are not used routinely
in the determination of the ABO blood groups, in spite of the fact
that satisfactory anti-A is available from several plants. However,
the anti-Ai of Dolichos biflorus and the anti-H of Ulex europeiis
make an ideal combination of reagents for the routine determination
of the subgroups of A and AB (Boyd and Shapleigh, 1954c), as
shown in Table 5-4. Anti-A lectins, especially from lima beans, have
had a number of applications in special experiments where a large
amount of anti-A agglutinin is needed (At wood and Scheinberg,
1958). Testing for H substance in saliva, by inhibition of the anti-H
of Ulex (Boyd and Shapleigh, 1954b), has become the preferred
method of diagnosing secretors and nonsecretors (Table 5-5).

PLANT AGGLUTININS (LFXTINS) I 69

TABLK 5-4
Determination of Subgroups of A and AB with Lectins^'

Reaction with

extract of:

Subgroup

Dolichos biflorus

Ulex europeus

Ai
A2
AiB
A2B

+ + + +

+ + +

0
+ + +

^ Boyd and Shapleigh, 1954c.

Lectins have been used by Morgan and Watkins (1956) to show
that the blood group antigen of group AB individuals is not a
mixture of A and B substances, but a unique molecule containing
both A and B specificities. The anti-N of Vicia graminca is actually
better than the absorbed anti-N prepared from immune rabbit serum
and v^ould doubtless be used routinely if more of the tiny seeds
of this South American plant were available. This lectin has already
proven valuable in clarifying the MN system in chimpanzees (Levine
et al., 1955). The anti-N of Bauhinia is not quite as good, but may
nevertheless come into routine use because the seeds are availalile in
many parts of the world (Boyd, Everhart, and McMaster, 1958).

TABLE 5-5
Inhibition of Anti-H (Ulex) by SaHvas of Secretors and Nonsecretors*

action of Ule
Sa

X lectin with
liva, diluted

group 0 cells

Blood groups

1:2

1:4

1:8

1:16

1:32

A (secretor)
A (nonsecretor)
B (secretor)
0 (secretor)
0 (nonsecretor)

+ + + +
0
0

++++

+ + + +
0
0

++++

+ + + +

0
+ + + +

+ + + +

0
+ + + +

+ + + +

+ +

+ + + +

Boyd and Shapleigh, 1954b.

70 INTRODUCTION TO IMMUNOCHEMICAL SPECIFICITY

Another example of the practical use of lectins was discovered by
Levine, Celano, Lange, and Berliner (1957). Using the anti-N of
Vicia graminea, they observed that horse erythrocytes contain the
N antigen, a fact that absorbed rabbit sera had failed to reveal because
species-specific agglutinins had remained in the absorbed sera. The
finding of N in the erythrocytes of the horse led Levine and his
colleagues to look for natural anti-M in horse serum. Having found
it, they were led to predict that the horse should be a good producer
of immune anti-M. This prediction was verified, and thus a new and
abundant source of immune anti-M was discovered.

Until recently all plant agglutinins tested fell into two categories :
either they agglutinated the red cells of all human beings and were
thus considered of little interest because they were "nonspecific," or
they reacted with one of the already known agglutinogens of human
blood. (It is a striking fact that in spite of the thousands of species
of plants already tested, no lectin reacting specifically with any of
the Rh antigens has been found. I shall suggest a possible explana-
tion for this in Chapter 7.) Recently, Boyd et al. (1959) found in
extracts of ordinary peanuts (AracJiis hypogaea) an agglutinin which
seems to be an exception to this rule. It agglutinates some human
erythrocytes and does not agglutinate others. But the agglutinogen
detected, if a distinct agglutinogen is involved, seems to be different
from A, B, H, M, N, P, S, s, C, D, E, c, e, V, Fy^ Fy^ K, k,
Le^, Le", Lu"*, Lu*", Jk", Jk'', or Js. Unfortunately the new agglutinin is
very weak and works only with erythrocytes suspended in serum
albumin.

References

Atwood, K. C, and S. L. Scheinberg, 1958, /. Cellular Comt^. Physiol. 52,

Suppl. 1, 97.
Bird, G. W. G., 1951, Current Sei. (India) 20, 298.
Bird, G. W. G., 1959, Brit. Med. Bull. 15, 165.
Boyd, W. C, D. L. Everhart, and M. H. McMaster, 1958, /. Immunol. 81,

414.
Boyd, W. C, D. M. Green, D. M. Fujinaga, J. S. Drabik, and E. Waszczenko-

Zacharczenko, 1959, ['o.r Sanguinis 4, 456.
Boyd, W. C, and R. M. Reguera, 1949, /. Immunol. 62, 22,i.
Boyd, W. C, and E. Shapleigh, 1954a, /. Immunol. 73, 226.
Boyd, W. C, and E. Shapleigh, 1954b, Blood 9, 1195.

PLANT AGGLUTININS (LECTINS) I 71

Boyd, W. C., and E. Shapleigh, 1954c, /. Lab. Clin. Med. 44, 235

Elo, J., E. Estola, and N. Malmstroni, 1951, Ann. Med. E.rptl. et Biol. Fcnniae

(Helsinki) 29, 297.
Grabar, P., 1947, Les Globulins du Scrum Sanguin, Editions Desoer, Liege.
Karush, F., 1950, /. Am. Chcm. Soc. 72, 2714.
Klotz, I. M., F. M. Walker, and R. B. Pivan, 1946, /. Am. Chem. Soc. 68,

1486.
Krupe, M., 1953, Z. Hyg. Injckiionskrankh. 318, 167.
Kriipe, M., 1956, Blutgruppcnspczifische Pfldnclichc Eizi'cis::kdrpcr (Phytag-

glufininc), Ferdinand Enke Verlag, Stuttgart.
Landsteiner, K., 1945, The Specificity of Serological Reactions, 2nd rev. cd..

Harvard University Press, Cambridge.
Lau. C, 1901, tJber vcgetabilische Blutagglutinine, Inaugural-Dissertation,

Rostock.
Levine, P., F. Ottensooser, M. J. Celano, and W. Pollitzer, 1955, Am. J. Phys.

Arthropol. 13, 29.
Levine, P., M. J. Celano, S. Langc, and \'. Berliner. 1957, I'o.v Sangitinis 2.

433.
Makela, O., 1957, Studies in Hemagglutini)is of Legumiuosae Seed. W'eilin and

Goos, Helsinki.
Makela, O., and P. Makela, 1956, Ann. Med. E.vpfl. et Biol. Peuniae (Helsinki)

34, 402.
Morgan, W. T. J., and W. M. Watkins, 1956, Nature 177, 521.
Ottensooser, F., and K. Silberschmidt, 1953, Nature 172, 914.
Renkonen, K.'O., 1948, Ann. Med. E.rptl. et Biol. Fcnniae (Helsinki) 26, 66.
Schmidt, G., 1954, Z. Immunitdtsforsch 111, 432.

CHAPTER 6
Plant Agglutinins (Lectins) II

Nature of Plant Agglutinins

There is no reason to suppose that the blood group-specific ag-
glutinins for which I proposed the name lectins are essentially dif-
ferent from the "nonspecific" agglutinins which have been known
so long. Ricin, from Ricinus communis, has been studied more
thoroughly than any other of the plant agglutinins. Although a toxin
as well as a hemagglutinin, it is, on the whole, a typical lectin. The
chemistry and immunology of ricin were studied and reviewed fairly
recently by Kabat, Heidelberger, and Bezer (1947) and by Kriipe
(1956). It is a globulin (i.e., soluble in salt solutions but not in dis-
tilled water), with an isoelectric point of 5.4-5.5 and a molecular
weight of about 80,000. It can be crystallized ; the crystalline protein
is highly toxic and strongly hemagglutinative.

The most complete immunochemical study on a specific lectin
carried out so far seems to be that of Boyd, Shapleigh, and McMaster
(1955). These authors found this anti-A lectin (from lima beans) to
be globulin also. They were not able to obtain it in crystalline form
but studied concentrated and partially purified preparations of which
about 36 per cent were specifically reactive with A antigen. The
partially purified material was electrophoretically heterogeneous but
nearly homogeneous in the ultracentrifuge. The observed sedimenta-
tion constant suggested a molecular weight of about 80,000.

Boyd and co-workers observed that the lectins, in addition to
their ability specifically to agglutinate erythrocytes of the appropriate
blood groups, precipitate specifically with the purified antigens (Boyd

PLANT AGGLUTININS (LECTINS) II

72>

and Shapleigh, 1954a; Bird, 1959). Boyd, Shapleigh, and McMaster
made a quantitative study of the precipitin reaction of their partially
purified lima bean anti-A and purified A antigenic substance from
hog stomach. The general course of the precipitin reaction was found
to be very similar to that of the precipitation of A substance by hu-
man anti-A antibody (Fig. 6-1).

5 10 15 20 25 30 35 40 45 50

Micrograms A substance added
Fig. 6-1. Protein nitrogen specifically precipitated from anti-A lima bean
lectin solution (open circles) by hog A substance, compared with nitrogen
precipitated from human anti-A serum (solid circles).

A characteristic of the lectins, and one that strikingly differentiates
them from immune antibodies, is their homogeneity. I do not mean
physical homogeneity, for in most cases the lectins have not been
purified sufficiently for us to know whether they are electrophoretically
and ultracentrifugally homogeneous.* I mean they are homogeneous
in their affinity for red cell antigens.

* By eluting proteins of different mobilities from paper electrophoresis paper,
Ensgraber, Kriipe and Ensgraber-Hattingen (1960) obtained evidence sug-
gesting that the agglutinins of twelve species studied by them were not
homogeneous electrophoretically. In two cases they were able to obtain fractions
of the total agglutinin present that did seem to be electrophoretically (and
ultracentrifugally) homogeneous.

74 INTRODUCTION TO IMMUNOCHEMICAL SPECIFICITY

When an immune serum or a "normal" isoagglutinin is found to
agglutinate more than one type of cell, absorption with one of these
types will generally remove the antibodies which affect it, leaving the
antibodies which react with other types of cells. The method of
preparation of a reagent for Ai, used before the introduction of
Dolichos lectin for this purpose (see Table 5-4 in Chapter 5) demon-
strates the behavior of "normal" isogglutinin. The serum of an in-
dividual of group B agglutinates erythrocytes of both subgroups Ai
and A^. Absorption with A2, however, removes the antibody which
reacts with A2 cells, leaving an anti-Ai agglutinin. Absorption of
an antibody to hen ovalbumin with duck ovalbumin removes the duck-
reactive antibody, leaving the anti-hen antibodies (some of which
will react also with ovalbumins of other avian species).

If it is attempted to repeat such an experiment with a plant ag-
glutinin instead of an antiserum from an animal, the results are
generally different. Absorption with one type of cell of a lectin that
agglutinates two different types of cells nearly always removes both
types of antibodies. If, for example, we try to make the lima bean

TABLE 6-1

Effect of Absorption on Agglutinating Activity of 10 Per Cent Solution of Lima
Bean (Sieva) Proteins''

Extract,

diluted

Test cells

1:2

1:4

1:8

1:16

1:32

1:64

1:128

1:256

1:512

1:1024

Before absorption

After 3 absorptions

with B cells"

* Boyd and Shapleigh, 1954d.

•^ U = undiluted.

" After absorption, 0.05 ml. of the indicated dilution of the protein solution was
mixed with 0.05 ml. of a 1% suspension of erythrocytes of the indicated group
and the mixture centrifuged, shaken, and read microscopically.

PLANT AGGLUTININS (LECTINS) II 75

lectin more specific by treating it with B cells to remove the anti-B,
we find we can do so only by absorbing an amount of the protein
sufficient to reduce the lectin concentration to such a point that it
will no longer agglutinate 1j cells (Table 6-1). This is seen from
the fact that, while the absorption was successful in lowering the
titer of the lectin against B cells about two stages, it also lowered
the titer against Ai cells two stages (Boyd and Shapleigh, 1954d).

Another indication of the homogeneity of lectins is the observation
of Morgan and Watkins (1953) : the agglutinin of SopJiora japonica
can be absorbed by either Ai or B cells, but the agglutinin, on being
eluted from either type of cell, still agglutinates A and B cells as
before, showing that no separation into molecules of different speci-
ficities has been effected.

This homogeneity of the lectins possibly sheds light on what their
function may be in the plant, and also, because of its contrast with
antibodies, has a bearing on theories of antibody formation.

Specificity of Plant Agglutinins

We have already seen that there are degrees of specificity. The
specificity of an antibody may be described as lozv if it reacts with
a large number of antigens, especially if they are closely related
chemically. Specificity is said to be sharp if an antibody reacts only
with a small group of chemically closely related antigens. An antibody
may be said to have absolute specificity if it reacts with only one
antigen.

When blood group-specific plant agglutinins were first discovered,
it was natural to suppose, since their reaction with the blood group
antigens was thought to be merely a chemical accident, that their
specificity was less sharp than that of the normal isoagglutinins. The
contrary has proved to be true, at least in some cases.

The specificity of some lectins is very sharp. The anti-A of human
group B plasma, for example, reacts nearly as well with Ao as with
Ai cells. The lectin of Doliclws biflorus, on the other hand, has a
much greater affinity for Ai than for A2 (over 500 times as great),
so that it is virtually specific for Ai. Human anti-A reacts also with
the Forssman antigen, the J substance present in the blood of some
cattle, the R antigen present in some sheep, and hog A substance.

76 INTRODUCTION TO lAIMUNOCHEMICAL SPECIFICITY

whereas the Dohchos lectin is not specilic fur any of these, hnt
detects a previously undescrihed heterogenetic factor present in
the erythrocytes of sheep, goat, horse, dog, and pig (Bird, 1959).

The anti-A of lima beans is somewhat less specific ; its affinity
for A2 is higher, and, when concentrated, it weakly agglutinates
B cells also. The anti-H lectins are still less specific, for, when
sufficiently concentrated, they may agglutinate Ai and B cells. This,
however, might be because human erythrocytes of these groups con-
tain some H antigen.

In addition to these relatively specific lectins, others are known
which seem to react with more than one receptor on the red cell,
such as A and B, or B and H. Finally, we come to those which
agglutinate human cells of all groups. Even these, however, may have
their own specificity, as I shall mention later.

As we shall see in the next chapter, human isoagglutinins react
not only with purified A and B substances but also with certain frag-
ments into which these antigens can be split, as by hydrolysis. In
the case of fragments too small to precipitate, specific activity has to
be demonstrated by the inhibition technique. In view of what we
know about antigenic determinants, we should expect a limit to this
fragmentation process ; i.e., if the A blood group antigen, for example,
is split into portions that are too small, the fragments will no longer
show A specificity or, in other words, will no longer inhibit the re-
action of anti-A agglutinin with A antigen. It was found the smallest
fragment of the A antigen which specifically inhibited the reaction
of human anti-A antibody with the A antigen is a disaccharide con-
taining A^-acetylgalactosamine as a terminal group. (The structure
of this disaccharide will be discussed in Chapter 7.) Strikingly
enough, anti-A agglutinins from plants were inhibited not only by
this disaccharide but by A^-acetylgalactosamine itself (Morgan and
Watkins, 1959). From this observation we could draw either of
two opposed conclusions. We could say that plant anti-A reagents
are less specific than human anti-A, since they are inhibited by a
simpler substance, or we could say they are more specific, since
they cross-react less with other portions of the A antigen.

In the case of anti-H agglutinins the plant reagents have not proven
to be any less specific than those of animal origin. The anti-H of eel
serum is inhibited by L-fucose, and L-fucose inhibits the anti-H of

PLANT AGGLUTININS (LECTINS) II

Lotus tctragouolobits and Ulex curopeus. L-Fucose does not inhibit
the anti-H of Cyfisits sessilifolius or of Laburnum alpinum, but salicin,
a gkicoside of D-glucose and sahgenin, does (Bird, 1959). If this
means that the lectins of Cytisus and Laburnum are directed toward
a part of the H antigen different from that recognized by the animal
anti-H reagents, it might suggest that the specificity of the lectins
is greater, not less than that of the animal agglutinins.

The anti-H of Lotus tctragonolohus is inhibited also by 2-deoxy-
L-fucose, L-galactose, 6-deoxy-L-talose, D-arabinose, and vV-acetyl-
glucosamine. Morgan and Watkins (1953) pointed out that, except
for the last, all the inhibiting sugars, when written in the pyranose
form, have the same configuration at carbon atoms 3 and 4 (Fig. 6-2) .
In all of them the hydroxyl groups are on the same side of the pyran
ring and pointing down.

Kriipe (1956) noticed that the sugars which inhibited the anti-
(A-|-B) agglutinin of Sopliora japonica (A^-acetyl-D-galactosamine,

H.OH

OH
6-Deoxy-L-talose

H.OH

^6h

H.GH

D-Digitoxose
Fig. 6-2. Haworth formulas of sugars inhibiting anti-H of Lotus tctra-
gonolohus. Arrows point to carbons 3 and 4, which liave the same configuration
in all these substances. (Redrawn from Morgan and Watkins, 1953).

INTRODUCTION TO IMAIUNOCHEMICAL SPECIFICITY

D-galactose, lactose, melibiose, L-arabinose, and D-fucose) also all had
the same configuration at carbons 3 and 4. Here, too, the hydroxyl
groups at carbons 3 and 4 are at the same side of the ring, but are
pointing up, which is just the opposite of that found in the sugars
inhibiting Lotus (Fig. 6-3).

CH2OH

H,OH

^ W OH

D-Galoctose

CH2OH

H.OH
3H H/'

^^H NHCOCH3

A/-Acefyl- D-galactosamine

H,OH

Fig. 6-3. Hawortli formulas of sugars inhibiting the anti-(A-|-B) of Sophora
japonica. (Redrawn from Kriipe, 1956).

Alakela (1957), who made a much more extensive study of plant
agglutinins and their inhibition reactions, suggested that monosac-
charides fall into four classes with respect to their specific inhibiting
activity for plant agglutinins and that this is based on their configura-
tion at carbons 3 and 4 (Fig. 6-4).

PLANT AGGLUTININS (LECTINS) II 79

The assignment of the aldohexoses and aldopentoses to Makela's
four groups and the steric similarities of these sugars in each classi-
fication are shown in Fig. 6-5. The relation between the pentoses and
hexoses shown in books on organic chemistry is based on possible
synthetic pathways in the laboratory and does not always show the
actual spatial relations of the ring structures.

Kriipe observed that the agglutinin of Ricinus communis was in-
hibited by sugars which fall into Makela's group 2, and the agglutinin
of Pisum sativum by sugars of group 3. Apparently these two "non-
specific" agglutinins do show a certain degree of specificity. In his

-0.

H0X3

•0.

Fig. 6-4. Clas.sification of pyranosc forms of sugars into four groups on the
basis of the configuration of carbons 3 and 4 (Makelii, 1957).

more extensive study Makela found many other leguminous seeds
having agglutinins which fell into one of these two classes. Other
legumes did not fall into either class. In addition to seeds with ag-
glutinins inhibited by sugars of groups 2 and 3, Makela's tables give
examples of seeds not inhibited by any sugars tested but inhibited
by blood group substances and some seeds that are not inhibited by
any of the substances tried.

Among plants the seeds of which contain agglutinins inhibited
by sugars of group 2 are Bandeiraca simplicifolia, various species of
Bauhinia, Sophora japonica, various species of Crotalaria, various
species of Cytisus, various species of Caragana, Wisteria cJiinensis ,
Coronilla varia, various species of Erythrina, and Glycine soja. In-
hibited by sugars of group 3 are Parkia filicoidea, Lathyrus latifolitis.

80 INTRODUCTION TO IMMUNOCHEMICAL SPECIFICITY

Lens culinaris, Pisuiit sativum, and various species of Vicia (Table
6-2).

Fig. 6-5. Steric relations of aldopentoses and aldohexoscs (Boyd, 1960).

PLANT AGGLUTININS (LECTINS) II 81

TABLE 6-2
Lectins^ Inhibited by Sugars of Group 2 and 3.

Group 2 Group 3

Bandeiraca simpUcijoUa Parkia filicoidca

Bauhinia spp. Lathynis latifoliiis

Sophora japonica Lens culinaris

Crotolaria spp. Pisuin sativiDii

Cytisus spp. Vicia spp.

Caragana spp.
Wisteria ehinensis
Coronilla varia
Erythrina spp.
Glycine soja

* Extracted from plant seeds of the species listed.

It seems clear that these plant agglutinins are not nonspecific, but
react with a definite chemical structure in the red cell, probably one
having as terminal group a sugar of group 2 or 3, as the case may
be. It happens that the erythrocytes of all human beings contain both
these particular receptors ; so therefore no individual differences are
found in the reactions of these lectins with human erythrocytes.
Further study with more complicated carbohydrates will enable us
to make a better guess at the detailed structure of these receptors.
As to the receptors with which other plant agglutinins combine we
have as yet no clue.

In a systematic study of the inhibition of two "non-specific" lectins,
that of Ricinus communis and that of Bauhinia purpurea (dialyzed
free of the group 2 sugars making it A/'-specific), Boyd and
Waszczenko-Zacharczenko (1961) found considerable similarities,
but some differences. Bauhinia lectin was inhibited by sugars of
Makela's group 3, but Ricinus was not. It was concluded that the
receptors in the human erythrocyte with which these two lectins
combine, though similar, are not identical. Both lectins were inhil)ited
by "unnatural" sugars of group 4.

In some cases the addition of an inhibiting sugar to a non-specific
plant agglutinin does not suppress all activity but leaves the prepara-
tion able to agglutinate cells of certain blood groups, thus revealing a
new specificity. Makela (1957) found that the agglutinin oi Bandeiraca

82 INTRODUCTION TO IMMUNOCHEMICAL SPECIFICITY

simplicifolia became B-specific when tested against cells suspended
in 2 per cent glucose. A suitable concentration of galactose makes the
anti-(A+B) agglutinin of Calpiirina aurea A-specific (Bird, 1959).
Boyd, Everhart, and McMaster (1958) found that some preparations
of Baiihinia purpurea were nonspecific (really specific for sugars be-
longing to group 2), but could be made N-specific by the addition of
D-galactose or other sugars of group 2. Most nonspecific plant ag-
glutinins, however, do not develop a new specificity when treated with
inhibiting sugars in this way. The three lectins just mentioned would
seem to be exceptional in this respect.

Role of Agglutinins in the Plant

We do not really know the role in the plant of the proteins which
we recognize by their ability to agglutinate certain types of erythro-
cytes. In speculating about this role we may follow several lines of
thought.

One possible approach, but in my opinion a naive one, is to as-
sume that because the lectins behave like antibodies they are real
antibodies. There are several arguments against this assumption :
(a) Although the literature on plant immunity is enormous, it has
not been demonstrated that plants manufacture antibodies, (b) There
is no evidence that the plants have ever been exposed to the blood
antigens with which their lectins react. It is extremely unlikely, for
example, that Vicia graminea has ever come in contact with the blood
group N antigen, (c) Lectins may occur in some varieties of a
species and be absent in others. This difference persists even when
the varieties are grown in identical environments ; experiments carried
out in Puerto Rico have indicated that the difference is hereditary
(Schertz, Jurgelsky, and Boyd, 1960).

Another point of view assumes that the configuration which enables
the plant proteins to combine specificially with certain blood group
antigens is merely an accidental feature of their structure and that
the proteins are present in the seed merely as storage material, or
for some similar purpose.

A third point of view, which I favor, holds that it is no accident
that the lectins are adapted to combine specifically with certain car-
bohydrates but that their function in the plant is to combine with.

PLANT AGGLUTININS (LECTINS) II 83

transport, and perhaps immobilize in the seed one or more of the
carbohydrates with which they have the power to combine. Kriipe
(1956) first suggested the possible role of the lectins as "Kohlen-
hydratfixierer" ("carbohydrate catchers"), and Boyd, Everhart, and
McMaster (1958) also thought that lectins might so function in the
plant.

Lessons from the Stutly of Lectins

The study of plant agglutinins promises to throw new light on the
specificity of the blood group antigens and on the nature and number
of carbohydrate groupings which are present on the surface of the
erythrocyte. I shall discuss this in the following chapter. Study of
inhibition reactions of the lectins has already thrown considerable
light on the structure of the ABH antigens. The difference in specifi-
city between lectins and human and animal agglutinins, whether we
care to regard this difference as evidence that lectins are less specific
or more specific, makes lectins particularly suitable for a study by
the inhibition reaction of the cell receptors. (See Chapter 7.)

The greater homogeneity of the lectins with regard to specificity
presents an interesting contrast with antibodies and might suggest
that a protein molecule which has a certain specificity as a part of its
role in metabolism is likely to be more uniform in this respect than a
gamma globulin mixture which has acquired a certain specificity by
some process of natural selection. This line of thinking may support
the antibody-formation theory of Jerne and subsequent modifications
thereof.

References

Bird, G. W. G., 1959, Brit. Med. Bull. 15, 165.

Boyd, W. C., 1960, /. Immunol. 84, 231.

Boyd, W. C, D. L.Everhart, and M. H. McMaster, 1958, /. Immunol. 81, 414.

Boyd, W. C, and E. Shapleigh, 1954a, Science 119, 419.

Boyd, W. C., E. Shapleigh, and M. H. McMaster, 1955, Arch. Biochcm. Biophys.

55, 226.
Boyd, W. C., and E. Shapleigh, 1954d, /. Immunol. 73, 226.

r^)Oyd, W. C, and E. Waszczenko-Zacharczenko, 1961, Transfusion, 1, 223.

ICnsgraber, A., M. Krupe, and R. Ensgraber-Hattingen, 1960, Z. Immunitats-
forsch. 120, 340.

84 INTRODUCTION TO IMMUNOCHEMICAL SPECIFICITY

Kabat, E. A., M. Heidelbergcr, and S. Hc/at, 1947, J. Biul. Clinii. 168. 629.
Kriipe, M., 1956, Bhitgruppcnspccifischc Pfliiiialiclic Envcisckdrpcr (Fhytny-

glutin'me), Ferdinand Enke Vcrlag, Stuttgart.
Makela, O., 1957, Studies in Hcinagglutinins oj Lcginiiiiiosac Scrcis, Wcilin and

Goos, Helsinki.
Morgan, W. T. J., and W. M. Watkins, 1953, Brit. J. Exptl. Pathol. 34, 94.
Morgan, W. T. J., and W. M. Watkins, 1959, Brit. Med. Bull. 15, 109.
Schertz, K. F., W. Jurgelsky, and W. C. Boyd, 1960, Proc. Nat. Acad. Sci. 46,

529.
Schiff, F., and L. Adelsberger, 1924, Z. Imiiiunitiitsforsch. 40, 335.

CHAPTER 7
Blood Group Antigens

Sources of Antigens for Study

The human erythrocyte is a compHcated structure, the blood group
antigens apparently making up only a small part of its mass. It is
not surprising, therefore, that attempts to determine the structure of
the blood group antigens by analyzing material isolated from erythro-
cytes have never given information of much value. Not only is the
starting material complex and the desired antigens only a small por-
tion of it, but the antigens seem to be bound in some way to the
lipids, and possibly to the proteins, which are present on the surface
of the red cell, making purification extremely difficult (Morgan and
Watkins, 1959) . If it were not for the much more abundant occurrence
of the blood group substances, in water-soluble form, in the saliva,
gastric juice, ovarian cyst fluid, and meconium of secretors, and the
occurrence of closely related antigens in hog and horse stomach, little
would be known today of the chemistry of blood group substances.

A number of methods of isolating and purifying blood group sub-
stances from the sources just mentioned have been described. The
extraction with cold 90 per cent phenol, employed by Morgan and
King (1943), used more than any other method, eliminates most of
the accompanying nonspecific protein and other impurities. High-
speed centrifugation and further fractionation from water and other
solvents results in further purification.

Blood Group Substances A, B, H, and Le"*.

As a result of such methods, four blood group substances have been
obtained in amounts sufficient for chemical study : A, B, H, and Le*.

86 INTRODUCTION TO IMMUNOCHEMICAL SPECIFICITY

The Le^ antigen (one of the antigens of the Lewis blood group sys-
tem) has been studied nearly as thoroughly as the ABH antigens be-
cause it also occurs in water-soluble form in body fluids.

Analytical Results

The results of chemical analyses of the ABH and Le^ blood group
antigens have been disappointing : they have not revealed any chemi-
cal differences that can be correlated with dififerences in blood group
activity. At a glance, the four antigens seem very much alike. They
each contain the same two sugar components, L-fucose and D-galactose,
and the same amino sugar components, D-glucosamine and D-galac-
tosamine. They also contain the same eleven amino acids (Kabat,
1956; Morgan and Watkins, 1959). The role of the amino acids is
not clear, for the specificity of the antigens seems to be determined
mainly by the carbohydrate portions. Morgan believes, however, that
the blood group antigens are not merely a loose combination of a
macromolecular polysaccharide with protein but consist of carbo-
hydrate chains and peptide units bound together by primary chemical
bonds.

Typical analytical values for preparations of the specific substances
are shown in Table 7-1. The observed dififerences are within the
range of variation found with different preparations of the same anti-
gen (Morgan and Watkins, 1959).

From such data we are forced to conclude that the specific serologi-
cal dififerences between the A, B, and H antigens are due not to
dififerences in over-all composition but to variations in the arrange-

TABLE 7-1

Typical Analytic Values for Preparations of Human Blood Group
Antigenic Substances"

Substance N, % Fucose, % Acetyl, % Hexosaniine, % Reduction, %

5.4

9.0

5.3

8.6

Le"

5.0

9.9

5.6

7.0

5.6

—

Morgan and Watkins, 1959.

BLOOD GROUP ANTIGENS 87

ment of the component parts. In fact, recent evidence suggests that
only certain parts of the complex polysaccharide molecules are re-
sponsible for the specific serological properties.

Of all the sugars present in the H blood group substance, for ex-
ample, only L-fucose (Fig. 7-1) specifically inhibited the agglutinat-

Fig. 7-1.

ing action of an anti-H from eel serum. Similar results were obtained
with an anti-H of plant origin, of the seeds of Lotus tetragonolobus.
It was also found that anti-H from either of these two sources was
inhibited more strongly by a-methyl-L-fucopyranoside than by the
yt?-furanoside or by fucose alone. These results suggested that l-
fucose is an important part of the H substance molecule ; by analogy
with Landsteiner's findings with composite haptens (p. 40), L-fucose
is probably the terminal group of the specific part. The fact that the
a-methylfucopyranoside inhibited better than the /8-methylpyranoside
suggested that the fucose was connected by an alpha linkage to the
next residue of the reactive portion of the H molecule.

The first information concerning the role of a particular sugar in
the specificity of the A substance was obtained by tests on anti-A
reagents of plant origin (Morgan and Watkins, 1959). Anti-A lectins
were specifically inhibited by A^-acetylgalactosamine (Fig. 7-2).
Most of the human anti-A reagents tested were not inhibited by this
amino sugar but were inhibited by the disaccharide 0-a-A^-acetyl-D-

H NHCOCH3

A/- Acetyl - D-galacfosamine

Fig. 7-2.

88 INTRODUCTION TO IMMUNOCHEMICAL SPFXIFICITY

galactosylaminoyl-(1^3)-D-galactose (Fig. 7-Z). This suggests that
this disaccharide must be very similar to, or possibly identical
with, the terminal disaccharide portion of the specific part of the
human A substance. (Morgan and Watkins, 1959).

'°k^

VOH

NHCOCH3

H,OH

o-cf-ZV- Acetyl- D-galactosylaminoyI- (I -^3) - D -galactose
Fig. 7-Z.

Kabat and co-workers (1956), also using the inhibition technique,
found evidence bearing on the structure of the specific part of the B
antigen. Of the monosaccharides present in the molecule, D-galactose
was the best inhibitor of anti-B antibodies, but the galactose-containing
disaccharide melibiose, the trisaccharide raffinose, and the tetrasac-
charide stachyose (Fig. 7-4) inhibited even better than galactose
alone (Fig. 7-5). This would have been expected if the specific part
of the B antigen consisted of a terminal nonreducing galactose unit
joined by an alpha linkage to another sugar unit. That the linkage
is alpha is pretty well shown by the fact that a-methylgalactoside
inhibits better than galactose, but the /;?-galactoside inhibits not as
well (Fig. 7-5).

Kabat was also able to draw some conclusions about the sugar
residue next to galactose in the specific side chain of the B antigen.
It could not be glucose, for glucose is not a part of the B molecule.
It was not likely to be another galactose, for, if it were, stachyose,
which contains a terminal galactose bound by a l-»6 alpha linkage to
another galactose, would be a better inhibitor than melibiose or
rafBnose, where the sugar next to galactose is glucose. But stachyose
is no better an inhibitor than melibiose or raffinose. According to
Kabat, A'^-acetylglucosamine is the only remaining possibility for the
next-to-terminal sugar in the specific side chain of B antigen.

BLOOD GROUP ANTIGENS

"°kr

N°"

Sfachyose
Fig. 7-4.

I I 1 II

Galoctose Melibiose Roffinose Stochyose

ji- Methylgalactoside a- Methylgolactoside

Fig. 7-5. Relative inhibiting power for anti-B of various sugars and glycosides
(redrawn from Kabat, 1956).

INTRODUCTION TO IMMUNOCHEMICAL SPECIFICITY

Some of this information may eventually see practical application,
for Kabat suggests that the introduction of a number of melibiosyl
residues into a polysaccharide would endow it with substantial blood
group B activity.

How many other sugars are present in the active side chains of
the A and B molecules is not known, but on the basis of his determi-
nations of the size of the reactive group of dextran (p. 48) Kabat
believes that the total is of the order of six. This would mean that
in the light of our present knowledge the specific portions of the
A and B molecules would resemble the structures shown in Fig.
7-6, where x stands for a number of the order of four.

H NHCOCH3 H OH

A Substance

CHgOH

1"'

"°A

/°i^

— 0
\

lSS!i_

HO^

^OH

1~"

NHC

OCH3

B Substance

X = approx. 4
Fig. 7-6. Suggested structures of reactive groups of blood group A
substances.

and B

Additional and independent evidence for the part played by
L-fucose, A^-acetylgalactosamine, and D-galactose in H, A, and B
specificity, respectively, was obtained by Watkins and Morgan (1955)
from the results of enzyme inhibition by these sugars. It is known
that an enzyme can be inhibited by an excess of one of the products
of its action on its substrate. An enzyme preparation was available
from Trichomonas foehts which destroyed the substrates consisting

BLOOD GROUP ANTIGENS

of A, B, and H substances. As expected, the destructive action of
the enzyme preparation on A substance was inhibited by A^-acetyl-
galactosamine, the action on B substance by galactose, and the action
on H substance by fucose.

There is at present no clue to the identity of the monosaccharide
unit next to fucose in the specific part of the H substance. Our best
picture of its structure is shown in Fig. 7-7 , where x stands for a
number of the order of five.

H Substance

X = approx. 5
Fig. 1-1 . Suggested structure of blood group H substance.

Watkins and Morgan (1957) found that the destruction of the
serological activity of the Le"* antigen by the Trichomonas enzymes was
inhibited by L-fucose, which suggested a role for this sugar in the
specificity of the Le"* antigen. However, the agglutination of Le(a+)

Q^ ,. NHCOCH3

-[i - D - A/- AcetylglucosaminoyI ■

' ^ \r

H\OH h/h

H OH

/9-D-GalQCtosyl -

Fig. 7-8. Suggested structure of terminal portion of blood group Le^ sub-
stance (Morgan and Watkins, 1959).

INTRODUCTION TO IMMUNOCHEMICAL SPECIFICITY

red cells by human or rabbit anti-Lewis sera was not detectably in-
hibited by L-fucose or by any other components of the blood group
substances. Certain oligosaccharides containing fucose did inhibit,
however, which showed that a-L-fucopyranosyl groupings were in-
volved in Le' specificity (Morgan and Watkins, 1959). From a
study of the inhibitory activity of various oligosaccharides, mostly
isolated by Kuhn and his colleagues from human milk (Kuhn, 1957),
Morgan and Watkins suggest that the terminal portion of the specific
part of the Le' substance is a trisaccharide of the structure shown
in Fig. 7-8.

Action of Genes
The way in which the ABO, secretor, and Lewis genes cooperate
to produce the various blood group substances found in the body
fluids of persons of different genotypes is still to be worked out.
Watkins and Morgan (1959) have proposed the following scheme
as a first approximation. Three independent gene systems, L' and V,
S' and s', and the ABO genes, are supposedly involved. In various

TABLE 7-2
Possible Genetic Pathways for the Production of Blood Group Substances. I'

Sequence and products of

gene action

Secretor type

Pre
suh

cursor l' gene

Precursor substance (not Nonsecretor ABH

stance (inactive)

acted on by S', A, or B genes) Nonsecretor Le''

L' gene

(+ a-fucosyl units)

Le=

sub

s' gene

_^ Le'^ substance (not acted on Nonsecretor ABH

Stance (inactive)

by A or B genes)

Secretor Le"

O gene

H substance +

S' gene

( + a-fucosyl units)

-^ Unconverted Le''
+ Leb

(inactive)

H substance Le''

B gene

B substance -(-

-^ Unconverted H

and Le" + Le^

•ABH

secretor

-|- Unconverted Le*

(+ a-galactosyl units)

A substance —

■^ Unconverted H
and Le» -|- Le^

(+ a-AT-acetyl
galactosaminoyl uints)

Watkins and Morgan, 1959.

BLOOD GROUP ANTIGENS

lAiii.i-; 7-,i

Possible (lenetic I'athways for the Production of Blood Group Substances. II"

Precursor substan

+ Inconvertcd I.(

m'"c

mene

(inactive)

i ^

[ ^

Asubst.

1! subst.

II s

l.c'

+ 11+ I.C-

+ 11 + u-

-e'

subst.

+ l,c''

+ I.e''

I.e''

Sccrctor

AliHsccre

tor

ABH nonsecreto

T>pe

Le» secretf
Le"" secret

■■

Le''

.onsecretor

• After W

itkins

and Morgan

I9S9.

recnrsor

lis

bstancc

substance

A.B.O

genes

Rcne

gene

I i ^ i

Rir Asubsl. Bsubsl.
+11 +11

Le"" nonsecrctor

ways they act to modify the precursor substance, a mucopolysac-
charide which is believed to be identical with the material found in
the secretions of the individuals who secrete neither A, B, or H nor
Le^ or Le** substances. The U gene acts to add a-fucosyl units to
this precursor substance, and the S' gene adds still more. The B
gene adds a-galactosyl units, and the A gene adds a-galactosaminoyl
units (see Table 7-2).

The scheme of Table 7-2 is inadequate in some respects, and
Watkins and Morgan suggest replacing it by the more complicated
system shown in Table 7-3.

Other Human Red Cell Receptors

Bauhinia Receptor
In addition to the red cell receptors characteristic of the various
blood groups, there are a number of receptors, some common to all

94 INTRODUCTION TO IMMUNOCHEMICAL SPECIFICITY

human erythrocytes, about which a certain amount of information
has been gained by a study of the inhibition of lectins by carbo-
hydrates. Let us first consider a receptor detected by extracts of
Bauhinia purpurea (Boyd, Everhart, and McMaster, 1958). As al-
ready mentioned, extracts of the seeds of this plant can be specific for
the N antigen; if they are not N-specific, they can be made so by
adding galactose or disaccharide containing galactose. The anti-N
specificity of Bauhinia extracts depends on the presence of sufficient
amounts of one or more sugars, probably galactose or galactose

TABLE 7-4
Inhibition of Dialyzed Bauhinia Extract by Carbohydrates*

Cells

Sugar,

diluted

Sugars

1:2

1:4

1:8

1:16

1:32

1:64

1:128

1:256

Melibiose

Raffinose

Lactose

L-Arabinose

D-Galactosamine HCl

(neutralized)

D-Galactose

Stachyose

1-,^

^ Equal amounts of the carbohydrate solution, lectin, and cell suspension were
mixed. The symbol U means that the carbohydrate solution (O.IM) was used
undiluted. The numbers signify strength of agglutination, 4 being the strongest
(all the erythrocytes stuck together in one large clump). Negative reactions are
recorded as 0.

BLOOD GROUP ANTIGENS 95

derivatives. Removal of these sugars by dialysis makes Bauhinia
purpurea extracts nonspecific; that is, they then agglutinate human
blood irrespective of blood group. The effect of certain sugars on
such nonspecific Bauhinia extracts is shown in Table 7-4.

It vv^ill be seen that the nonspecific activity of Bauhinia extracts
is inhibited by sugars of Makela's group 2. It may even be that the
red cell receptor detected by nonspecific Bauhinia extracts is the
same as that detected by other plant agglutinins which are inhibited
by group 2 sugars. No adequate comparison has yet been made.*
From the inhibiting power of the disaccharide, trisaccharide, and
tetrasaccharide containing galactose, it can be assumed that the
Bauhinia receptor is an oligosaccharide containing at least one more
unit beyond galactose. It probably contains several monosaccharide
units, although, if it does, one can conclude that the next-to-terminal
unit is not galactose, as in stachyose, for this sugar is not a very good
inhibitor here. Although the Bauhinia receptor has some features
in common with the B receptor, it is obviously not identical with
it since it occurs in all human erythrocytes. As a matter of fact,
B substance does not react with the Bauhinia agglutinin.

Peanut Receptor
Another receptor has been detected with the aid of extracts of
ordinary peanuts. This plant agglutinin is also inhibited by sugars
of group 2 (Table 7-5). That galactose is the most effective monosac-
charide inhibitor suggests that galactose is the terminal group of
this receptor also. The receptor consists of more than one sugar
unit, however, since two disaccharides, trehalose and lactose (the
former not even containing galactose), inhibit better than galactose.
Two other disaccharides containing only glucose (maltose and cel-
lobiose) also inhibit well. The inhibitory power of the three all-
glucose disaccharides and the fact that lactose contains glucose sug-
gest that glucose may be the next-to-terminal group in the peanut
receptor. This is supported by the observation that glucose itself
has some inhibiting power. The effectiveness of trehalose suggests

* In my laboratory we have recently made a detailed comparison of two
lectins that are inhibited by group 2 sugars (those from BauJiiiiia purpurea
and Ricinus communis) , and obtained evidence that the two receptors are not
identical (Boyd and Waszczenko-Zacharczenko, 1961).

INTRODUCTION TO IMMUNOCHEMICAL SPECIFICITY

TABLE 7-5
Inhibition of Peanut Lectin (Anti-Gy) by Sugars^
Sugar, diluted

Undil.

1:2

1:4

1:8

1:16

1:32

1:64

1:128

Saline

(control)

+ +

D-Galactose

+ +

L-Arabinose

+ +

D-Arabinose

+ +

D-Glucose

+ +

D-Mannose

+ +

Cellobiose

+ +

Maltose

+ +

Trehalose

+ +

Melibiose

+ +

Lactose

* + = Agglutination of test

cells, 0 =

= no aggl

utination.

that the peanut receptor contains glucose as the next-to-terminal
unit, possibly linked to the terminal galactose by a l^-l link (the
linkage in trehalose). The peanut agglutinin, in spite of a presumed
galactose terminal unit, is not inhibited by B substance and in that
respect resembles the Bauhinia agglutinin. The two receptors are
different, however, since the peanut receptor, unlike the Bauhinia re-
ceptor, is not found on all human erythrocytes. Another sign of
difference is that the peanut agglutinin is inhibited by sugars (cello-
biose, trehalose, melezitose, and maltose) which do not inhibit the
Bauhinia agglutinin.

There thus seem to be at least three receptors containing galactose
as a terminal unit, one of them present on all human red cells, the
others only on those of certain individuals. Their structure in the
light of our present scanty information is shown in Fig. 7-9.

It has already been mentioned that plant agglutinins inhibited by
sugars of Maleka's group 3 react with a red cell receptor present
on all human red cells. It would seem likely that the terminal unit
in this receptor is a sugar of group 3. Since Makela found mannose
the best inhibitor for such agglutinins, one could hazard the guess that

BLOOD GROUP ANTIGENS

CHaOh

H0\°1.

H
B Substance

n . —

NHCOCH

CH20H

H OH

Bauhinia receptor

CH2OH

Peanut receptor (Gy)
Fig. 7-9. Suggested structure of the terminal portions of three receptors of
the human red cell which contain galactose as a terminal monosaccharide unit
(Boyd, 1960).

the terminal unit is mannose. We cannot go further than this on
the information now available.

The RH Receptors
In my laboratory we recently applied the specific inhibition tech-
nique to a study of the Rh blood group receptors. Hackel, Smolker, and
Fenske (1958) reported that anti-Rh sera are inhibited specifically
by a number of ribonucleic acid derivatives, which suggested that the
Rh antigens are at least partly ribonucleotide in nature. We found
that human anti-D serum is also inhibited, weakly it is true, but
apparently specifically, by the "unnatural" sugars of Makela's group
4, including L-mannose, L-glucose, and D-gulose (Table 7-6). It is
not inhibited by the "natural" enantiomers of these sugars or by
any other sugars we have tried.

98 INTRODUCTION TO IMMUNOCHEMICAL SPECIFICITY

TABLE 7-6

Inhibition of Anti-D by 0.2M Solutions of Various Sugars'*

Serum,

diluted

Substance added

Undiluted

1:2

1:4

1:8

1:16

1:32

Saline (control)

3.5

D-Glucose(3)

3.5

1.5

L-Glucose (4)

2.5

D-Mannose (3)

3.5

2.5

L-Mannose (4)

2.5

D-Gulose (4)

» Numbers indicate strength of agglutination, from 4 = complete agglutination,
to 0 = no agglutination. The numbers in parentheses after the names of sugars
indicate the group of the sugar in Makela's (1957) classification (see Figs. 6-4
and 6-5).

The results suggest that the D receptor may contain a sugar of
group 4 as terminal unit. They are supported by the observation
that streptomycin, a natural glycoside of A''-methyl-L-glucosamine,
and rutinose [6-0-(/3-L-rhamnosyl)-D-glucose] also inhibit (Table
7-7). Streptomycin does not inhibit much better than L-mannose or

TABLE 7-7
Inhibition of Anti-D by Glycosides

Serum,

diluted

Substance

Undil.

1:2

1:4

1:8

1:16

1:32

Saline

Rutinose (4)
Streptomycin

(4)

4
4
0

4
0
0

3
0
0

2
0
0

0
0
0

L-glucose. This is not surprising considering that the next-to-terminal
unit is 5-deoxy-3-formyl-L-lyxose, which one would not expect to
find in red cells (Fig. 7-10), though such preconceived notions may
be dangerous. Rutinose, however, on a molar basis (the solution
available was only 0.14 as strong as the other sugar solutions studied),
inhibits better than streptomycin or L-mannose, which might suggest

BLOOD GROUP ANTIGENS

0 (Streptose)

(Streptidine)

OH H

(A/- methyl -t-glucosomine)

Streptomycin
Fig. 7-10.

that the next-to-terminal unit in the D receptor is D-glucose or a
similar sugar. The likelihood that the terminal unit of rutinose,
L-rhamnose (Fig. 7-11), is the terminal unit of the D receptor
is diminished by the observation that rhamnose itself does not inhibit.

H,OH

It is hardly necessary to mention that a knowledge of the chemical
structure of the D antigen could have considerable practical value.
It might enable us to make good anti-D agglutinins by immunizing
animals, which is now impossible. Injections of a nontoxic oligosac-
charide with high D activity into pregnant women might possibly
neutralize the anti-D of the maternal and fetal circulations and pre-
vent erythroblastosis in the infant.

The inhibition behavior of anti-C and anti-E seems to be more
complicated (Table 7-8). L-Glucose has some inhibitory effect on

100

INTRODUCTION TO IMMUNOCHEMICAL SPECIFICITY

TABLE 7-8
Inhibition of Anti-Rh Sera bv D- and L-Glucose

Sugar

Serum,

diluted

Serum

Undil.

1:2

1:4

1:8

1:16

1:32

Anti-D

Saline

3.5

D-Glucose

3.5

1.5

L-Glucose

2.5

Anti-C

Saline

3.5

1.5

D-Glucose

3.5

1.5

L-Glucose

Anti-E

Saline

1.5

D-Glucose

3.5

0.5

L-Glucose

1.5

both agglutinins, but anti-E is also inhibited by D-glucose. Other
sugars of group 4 do not seem to inhibit anti-C ; therefore I have no
confidence yet that the terminal unit is a stigar of this group (Table
7-9). It seems possible that a sugar of group 3 is the terminal unit

TABLE 7-9
Inhibition of Anti-C

Substance

Undil.

Serum, diluted

1:2

1:4

1:16

1:32

Saline

D-Glucose (3)
L-GIucose (4)
D- 1 dose (4)
Rutinose (4)
Streptomycin (4)

3.5
0
4
3.5

2.5

3.5

2.5

of the E receptor, considering the effectiveness of sugars of this group
in inhibiting anti-E (Table 7-10). The inhibition by L-allose, a
group 2 sugar, is unexpected and is not paralleled by inhibition by
other group sugars.

If it should prove that the D receptor, and possibly the C and E

BLOOD GROUP ANTIGENS 101

TABU

<: 7-10

Inhibition

of Anti-

Serum,

diluted

Substance

Undil.

1:2

1:4

1:8

1:16

1:32

Saline

3.5

1.5

D-Mannose (3)

D-Gliicose (3)

0.5

L-Mannose (4)

dtz

0.5

L-Glucose (4)

1.5

D-Gulose (4)

1.5

Rutinose (4)

Streptomycin (4)

3.5

1.5

L-AUose (2)

D-Galactose (2)

3.5

L-Arabinose (2)

3.5

receptors as well, contains a sugar of group 4, it may surprise some
people, for sugars of this group have not previously been found in
human tissues. We may still expect many surprises regarding the
natural occurrence of sugars. In the field of protein chemistry it is
commonly assumed that only one enantiomer of each amino acid occurs
in nature, and yet Oncley (1959) has pointed out reasons for doubting
this. The finding of a derivative of L-glucose in streptomycin has
already shown that some of these "unnatural" sugars occur in nature.
In the next chapter it will be seen that such sugars play a role in
the antigens of Salmonella and certain parasites.

Since the above was written, Dodd, Bigley, and Geyer (1960),
starting from the observation that the receptor-destroying mumps
virus liberates from human erythrocytes a specific anti-D inhibitor,
found that A^-acetyl-neuraminic acid and other compounds related
to sialic acid inhibited anti-D but not anti-C or anti-E, thus pro-
viding a cltie to the chemical differences between D and the other
Rh antigens. In my laboratory we found that colominic acid, thought
to be a polymer of A^-acetyl neuraminic acid, also inhibits anti-D
specifically, and suggested it might even have clinical application
in preventing erythroblastosis fetalis (Boyd and Reeves, 1961).

We have also found other substances, including some amino acids,
not closely related to any of the substances discussed above, to have

102 INTRODUCTION TO IMMUNOCHEMICAL SPECIFICITY

considerable inhibitory power. None of these compounds inhibited
anti-A, anti-B, anti-H, anti-M, or anti-N. It is clearly premature
to present any detailed picture of the structure of any of the Rh
antigens, but if progress continues at the present rate something
should be known in a few years at the latest.

References

Boyd, W. C, 1960, /. Immunol, 85, 221.

Boyd, W. C, D. L. Everhart, and M. H. McMaster, 1958, /. Immunol. 81, 414.

Boyd, W. C, and E. Reeves, 1961, Nature 190, 1123.

Boyd, W. C, and E. Waszczenko-Zacharczenko, 1961, Transfusion, 1, 223.

Dodd, M. C. N. J. Bigley, and V. B. Geyer, 1960, Science 132, 1398.

Hackel, E., R. E. Smolker, and S. A. Fenske, 1958, Vox Sanguinis 3, 402.

Kabat, E. A., 1956, Blood Group Substances, Academic Press, New York.

Kuhn, R., 1957, Angew. Chem. 60, 23.

Morgan, W. T. J., and H. K. King, 1943, Biochem. /., 37, 640.

Morgan, W. T. J., and W. M. Watkins, 1959, Brit. Med. Bull. 15, 109.

Oncley, J. L., 1959, Rev. Mod. Physics 31, 30.

Watkins, W. M., and W. T. J. Morgan, 1955, Nature 175, 676.

Watkins, W. M., and W. T. J. Morgan, 1957, Nature 180, 1038.

Watkins, W. M., and W. T. J. Morgan, 1959, Vox Sanguinis 4, 97.

CHAPTER 8
Salmonella Antigens

Progress in bacteriology was greatly aided by the development ot
methods of staining microorganisms to facilitate their microscopic
observation. The staining methods were developed empirically. Al-
though their mechanism is still obscure, two of them are now known
to detect fundamental and significant differences in the cellular
structure of bacteria. These two reactions are the Gram stain and
the acid-fast stain. On the basis of their behavior toward the reagents
used in the two reactions, all bacterial species may be classified into
three broad groups : Gram-positive, Gram-negative, and acid-fast.
There also exist intermediate forms with poorly defined staining
reactions (Dubos, 1952).

Endotoxins

Gram-negative bacteria are characterized by the fact that, when
they are dyed with a basic triphenylmethane dye such as gentian
violet, the color can be removed by washing with alcohol. Gram-
negative bacteria have a number of other features in common. One
of the most interesting of these features to the immunologist is their
content of endotoxins. These characteristic substances, not released
to any great extent into the culture medium as the organism grows
(in contrast to exotoxins such as diphtheria toxin), can be obtained
by lysing the bacteria or by extracting them with trichloracetic acid,
diethylene glycol, etc. They are toxic in animals in very small amounts
(of the order of 0.001 microgram per kilogram of body weight), pro-
ducing fever, leukopenia, etc.

103

104

INTRODUCTION TO IMMUNOCHEMICAL SPECIFICITY

The first important step in the study of endotoxins was the de-
velopment by Boivin (Boivin and Mesrobeanii, 1933) of a method
of extracting them with trichloracetic acid. Boivin reported that the
endotoxins of a number of gram-negative bacteria consisted mainly
of polysaccharide and lipid.

Modern knowledge of the chemistry of endotoxins derives mainly
from the work of Morgan (1937, 1940, 1941, 1942) and Goebel
(1945), who showed that endotoxins are complexes containing phos-
phorylated polysaccharide and protein. Further work by these and
other workers on the degradation products of the endotoxins revealed
that they have the make-up shown in Table 8-1 (Westphal and
Liideritz, 1954).

TABLE 8-1
The Endotoxin Complex of the Cell Wall of Grain-Xegative Bacteria''

Lipopolysaccharide-Protein-Lipid-Coniplex

Lipopolysaccharide
(iindegraded poly
saccharide)

Phosphorylated
polysaccharide

Conjugated protein

Lipid A

Lipid B

(easily split off)

Protein

» Westphal and Liideritz, 1954.

It appears that the lipid A component is responsible for many of
the toxic effects of these complex substances ( Schmidt, Eichenberger,
and Westphal, 1958; Westphal, 1960). All the preparations of this
component examined from various enterobacteria seem to be similar
or perhaps identical (Westphal, 1960), containing about 20 per cent
D-glucosamine, 7-8 per cent phosphoric ester, 50 per cent long-chain
fatty acid (including hydroxy-fatty acids), and a peptide side chain
containing serine and dicarboxy amino acids.

SALMONELLA ANTIGENS 105

Although the Hpid A component of the endotoxins of the gram-
negative bacteria is essential for many of the endotoxic manifesta-
tions and can act as a potent adjuvant in the production of antibodies
(Westphal, 1960), the portion which determines the specificity of
protective antibodies is the polysaccharide component. Such anti-
polysaccharide antibodies do not protect the organism producing them
against the pyrogenic effects of endotoxin if it is experimentally in-
jected, but they do account for the species-specific immunity which
generally follows recovery from an infection with one of the micro-
organisms. Consequently it is the anti-polysaccharide antibodies
which are of greatest interest to immunologists. Considerable progress
has recently been made in the study of the chemical basis for the
immunological differences which are observed, especially in the group
of gram-negative bacteria known as the Salmonella. Before we can
discuss them we must pause to recall a few salient facts about this
important group of microorganisms.

The Salmonella

The Salmonella are Gram-negative, non-spore-forming, motile
bacteria which are generally pathogenic for both man and animal.
S. typJwsa, causative agent of typhoid fever, .S". paratyphi A, and
possibly ^. sendai, cause disease only in man.

The Salmonella are mostly flagellated. The flagella as well as the
body of the organism contain antigens. The flagellar antigens are
called H antigens, the somatic antigens O antigens. The letters origi-
nated with German writers who observed that colonies of the motile
(i.e., flagellated) Salmonella on agar medium were surrounded by a
"Hauch" (breath or emanation), while colonies of the nonmotile
organisms were "Ohne Hauch" (without emanation).

The H antigens are of two kinds : those shared by a number of
species or types, and those peculiar to a particular species or type,
or shared by only a few species or types. Many of the species or
types are diphasic ; that is, at one stage of a culture the specific
flagellar antigens may occur (specific phase), whereas at another
the group antigens may be present (group phase). Any given cul-
ture of such an organism may consist entirely of one or the other of
the phases or may contain both. A bacillus in one phase usually keeps

106 INTRODUCTION TO IMMUNOCHEMICAL SPECIFICITY

the same phase for a number of generations, but is always capable
of giving rise to the other phase. As a matter of fact, the antigens of
either phase may occur in various types, although the specific antigens
are generally restricted to a smaller number of types (Dubos, 1945).
These complicated antigenic properties of the Salmonella can be a
source of confusion unless they are understood.

TABLE 8-2
Somatic and Flagellar Antigens in Certain Common Salmonella*

H Ant

igens

Phase I

Phase 2

Group

Species

0 Antigens''

(specific)

(group)

S. paratyphosa

(I), II, XII

S. schottmuelleri

(I), IV, (V),

XII

1,2

S. typhim.uriinn

(I), IV, (V),

XII

1,2

S. hirschfeldii

VI, VII,

1,5

S. choleraesuis

VI, VII

1,5

S. oranienburg

VI, VII

m,t

—

S. niontevideo

VI, VII

g,m,s

—

S. newport

VI, VIII

e,h

1,2

S. typhosa

IX, XII,

—

S. enteritidis

(I), IX, XII

g,m

—

S. gallinarum

I, IX, XII

—

S. pullorum

I, IX, XII

—

S. anatum

III, X

e,h

1,6

* Modified from a table in Bacterial and Mycotic Infections of Man, edited by
R. J. Dubos, 1952. Courtesy of Dr. Dubos, Dr. H. R. Morgan, J. B. Lippincott
Co., and the National Foundation for Infantile Paralysis.

•^ Parentheses indicate that the antigen is not invariably present.

It was formerly the practice to designate the somatic (O) antigens
by Roman numerals, as shown in Table 8-2, but following a decision
made in 1953 at the Sixth International Congress of Bacteriology in
Rome, the workers who have recently contributed so much to our
knowledge of the chemical structure of these antigens, employ
Arabic numerals. I shall follow this usage. It was former practice
to denote the species flagellar antigens by small Roman letters, and
the group flagellar antigens by Arabic numerals. Thus Salmonella
nczvport possesses O antigens VI and VIII, species H antigens e and

SALMONELLA ANTIGENS 107

h, and group H antigens 1 and 2. From here on, however, we shall
be speaking of 5". ncivport as possessing O antigens 6 and 8. The
flagellar antigens will not come into the picture, because I do not
intend to discuss them further.

The antigenic structure of the Salmonella has been studied in
great detail by Kauffman (1937) and White (1926) ; the classifica-
tion of these authors, based on the somatic and flagellar antigens,
is in common use. In general, the species of Salmonella are divided
into groups on the basis of similarity with respect to the O antigens,
and the species within a group are often differentiated according to
differences between their H antigens (Kauffmann, 1950, 1951). The
species have been arranged in groups designated A, B, C, etc., ac-
cording to similarities in the content of O antigens. All this, it should
be remembered, was done purely on the basis of serological evidence.

Chemistry of the Polysaccharide Component
of Salmonella Antigens

On hydrolysis, the Salmonella polysaccharides split into monosac-
charides and phosphoric acid. Chromatographic study of the sugars
shows that they represent a fairly complicated mixture ; a single
polysaccharide may consist of six to seven different sugars, in-
cluding hexosamines (glucosamine and galactosamine), heptoses,
hexoses, pentoses, and deoxy sugars (Davies, 1955; Mikulaszek
et al., 1956; Davies, 1960). The dideoxy sugars move faster on
chromatograms than the other sugars do, and their discovery, based
on this property, by Staub (1952) and Westphal (1952) was a new
fact of great immunochemical interest. They play a very important
role in the structure of the Salmonella antigens because :

(a) Brief acid hydrolysis of the Salmonella lipoidpolysaccharides
always splits off these dideoxy sugars before significant amounts of
other sugars are released. This shows that the deoxy sugars are
terminal and acid labile in the branched polysaccharide structure.
It is known (see above, pp. 39-40) that the terminal groups play
a predominant role in hapten specificity.

(b) When pathogenic "smooth" Salmonella forms change to the
nonpathogenic "rough" forms, the fast chromatographic sugar com-
ponents in hydrolysates of the antigens are missing, although the
endotoxic lipoid A component is still present.

108

INTRODUCTION TO IMMUNOCHEMICAL SPECIFICITY

These dideoxy sugars are all 3,6-dideoxyhex()se.s, and five have so
far been identified in natural antigens (Table 8-3).

TABLE 8-3
Naturally Occurring 3,6-Dideoxyhexoses^

Name

First found in

References

Abequose

Endotoxin of S. abortus eqiii

Westphal, Liideritz, Fromme,
and Joseph (1953).

Tyvelose

Endotoxin of S. typhosa

Pon and Staub (1952), West-
phal, Fromme, and Joseph
(1953).

Ascarylose

Glycolipid of eggs of Para-

Fouquey, Polonsky, and Lederer

scans cquorum

(1957).

Paratose

Endotoxin of S. paratyphi

Davies, Fromme, Liideritz,
Staub, and Westphal (1958).

Colitose

Endotoxin of Eschenchia

Liideritz, Staub, Stirm, and

coli 0 111

Westphal (1958).

=* Westphal,

1960.

The structures of these dideoxyhexoses are shown in Fig. 8-1.
It will be noted that two of them, colitose and ascarylose, have the
configuration of the "unnatural" L-series of hexoses, which are sus-
pected of playing a role in the structure of the human Rh antigens
(see Chapter 7). This does not necessarily mean that any serological
similarities between the Salmonella antigens and the Rh blood group
antigens are to be expected, although this is a point which so far as
I know has not been tested. But it does tend to confirm our suspicion
that the "unnatural" sugars are more widely distributed in nature
than was expected. What their relative abundance will turn out to
be is another question.

Relation of Structure of Salmonella
Antigens to Specificity

Comparison of the results of chromatographic analyses of Sal-
monella antigens with their position in the Kaufifmann-White classi-
fication (Staub, Tinelli. Liideritz, and Westphal, 1959; Staub, I960:
Westphal, Liideritz, Staub, and Tinelli, 1959) showed that the

SALMONELLA ANTIGENS

109

An,

)iH,OH

Colitose

(3,6-D

'ideoxy-

-L-galoctose)

CHj

S—

-V

)JH,OH

— r

Porafose

(3,6- Dideoxy

-D-g

lucose)

CH,

H0>

HO/

JH.OH

Tyve

lose

(3,6- Dideoxy-

•D- monnose)

Abequose
( 3,6- Dideoxy- D- galactose)

H,OH

Ascorylose
(3,6- Dideoxy- L- mannose)

Fig. 8-L Five naturally occurring 3,6-dideoxyhexoses.

terminal dideoxy sugars did play an important antigenic role, as ex-
pected. Each Salmonella species produces only one such sugar, and
this sugar is characteristic of the group. A, B, etc., into which the
species falls in the Kaufifmann- White scheme. Paratose is characteris-
tic of species in group A, for example, and colitose of group O (Table
8-4).

It has been further shown (Staub, Tinelli, Liideritz, and Westphal,
1959) that different dideoxyhexoses function as terminal groups
of various antigenic factors of the Kauffmann- White scheme, abequose
being the terminal unit of antigen 4 of group B, tyvelose of antigen 9
of group D, and colitose of antigen 35 of group O.

There seems to be no evidence that more than one of these 3,6-
dideoxyhexoses occurs in the endotoxin of any one species of
Salmonella. When a 3,6-dideoxyhexose does occur it always occupies
the terminal position in a side chain of the antigenic determinant of

110

INTRODUCTION TO IMMUNOCHEMICAL SPECIFICITY

TABLE 8-4

Carbohydrate Structural LInits of Specific O Antigens (Endotoxins) of Salmonella

Groups A, B, D, and O'^

(Heptoses and aminosugars not inchided)

6-Deoxy-

3,6-Dideoxy-

Kauffmann-

Hexoses

1 . i

hexoses

xoses

White

CJ '~ "

Group species

antigens

C2J

A 5. paratyphi

1,2,12

B S. schoUmuellen

1,4,5,12

B 5. typhimurium

1,4,5,12

B S. abortus equi

4,12

B S. budapest

1,4,12

B S. Stanley

4,5,12

B S. salinatus

4,12

D 5. typhosa

9,12

D 5. enteriditis

1,9,12

D 5. gallinarum

1,9,12

D S. dar-es-salaam

1,9,12

0 S. adelaide

0 6'. monschaui

" Westphal, 1960.

the carbohydrate antigen. This does not mean that other sugars can-
not be terminal, for glucose and rhamnose can occur in this position.

As would have been expected on the basis of what we have learned
in earlier chapters of this book, the most informative way of studying
the terminal sugar of these antigens was found to be by the inhibition
reaction. Staub, Westphal, and colleagues (Staub and Tinelli, 1957;
Staub et al., 1959) took advantage of the fact that degree of in-
hibition can be measured quantitatively if the reaction inhibited is
the precipitation of a soluble antigen by a precipitating antibody,
see above, p. 20 ; they made use of soluble antigens obtained by
acetic acid lysis of the microorganisms and purification by Freeman's
method (1942) of the product (Table 8-5).

In this table PsTy stands for the polysaccharide extracted from
5". typhosa, PsTyB for the polysaccharide from S. schottmuelleri
(formerly paratyphoid B), and PsTyox for the carbohydrate of

SALMONELLA ANTIGENS

111

TABLE 8-5

Specific Inhibition of Precipitation of Salmonella Antigens by
Anti-5. typhosa Antiserum*

Horse

anti-typhoid i

;erum

Rabbit anti-typhoid

serum

Inhibitor

reacted with

reacted w

ith

PsTyb

PsTyo.«

PsTyBd

PsTyb

PsPtRd

(12,9)"

(9)

(12)

(12,9)

(12)

Glucose

Galactose

Mannose

Rhamnose

Tyvelose

* Staub et al., 1959. Numbers indicate per cent inhibition.

^ Polysaccharide extracted from S. typhosa.

" PsTy oxidized with periodic acid.

<^ Polysaccharide from 5. schottmuelleri (formerl}- paratyphoid B) = S. para-
typhi B.

« Somatic antigens 9 and 12 of the KaulTman — White scheme. The italic number
indicates the antigen which characterizes group D, the group that includes
5. typhosa.

S. typhosa after treatment with periodic acid. The reason for inckid-
ing such oxidized antigens in the studies is that periodic acid destroys
substances possessing two adjacent hydroxyl groups, such as terminal
ghicose or galactose. Terminal 3,6-dideoxyhexoses, however, do not
possess such a combination of hydroxyls and are not attacked.

From the results obtained with the horse anti-typhoid serum shown
in Table 8-5, Staub et al. (1959) concluded that tyvelose is the
terminal sugar of antigen 9 and rhamnose that of antigen 12.

It will be seen from Table 8-5 that the results obtained with the
rabbit serum were quite different from those of the horse serum.
The precipitation of the polysaccharide of S. typhosa (PsTy) by
horse anti-typhoid was inhibited significantly only by rhamnose and
tyvelose, whereas these sugars inhibited precipitation of the same
antigen by rabbit anti-typhoid very poorly. Glucose was much more
active with rabbit serum. The difference in inhibition of precipitation
of the polyoside of 6^. scJwttiniiellcri (PsPtB) was even greater. It
was therefore concluded that antigen 12, common to S. typhosa and

112 INTRODUCTION TO IMMUNOCHEAIICAL SPECIFICITY

^. schottinuclleri, contains a side chain terminating in glucose as
well as one terminating in rhamnose.

Similar studies carried out by Staub et al. on antisera to .S". sclioff-
muelleri (containing antigens 4, 5, and 12) showed that abequose in-
hibited the precipitation of PsPtB and especially of PsPtBox- This
showed that abequose is the terminal unit of either antigen 4 or 5.
Since abequose and antigen 4 are found in all Salmonella of group B,
but antigen 5 is lacking in some members of this group, Staub et al.
concluded that abequose plays no role in antigen 5. This was con-
firmed by the observation that the precipitation of an extract of
S. typhimurimn, which contains no 5 antigen, is inhibited by abequose
and by the finding that when all the antibody precipitable by an ex-
tract of this 5". typJiiniuriitui was removed from the anti-PsPtB
serum, the action of the serum on PsPtB was not inhibited by the
abequose.

Staub et al. (1959) suggest that the dideoxyhexoses may play an
especially important role in the specificity of the Salmonella antigens,
not only because they are always terminal, but because the two hydro-
phobic CH2-groups they contain are able to approach much closer
to the corresponding surface of the antibody than the hydrophilic
CHOH-groups, thus strengthening the van der Waals forces between
the antigenic determinants and the antibody (see Chapter 9).

Cross-Reactions

As a result of extensive studies similar to those just outlined, Staub
et al. concluded that although distinct Salmonella antigens generally
have different terminal sugars, this is not always the case. For in-
stance, abequose occurs at the extremity of both antigens 4 and 8,
and glucose at the extremity of antigens 1 and 12. It seems reasonable
to conclude that in such cases the next-to-terminal sugar is difTerent,
or attached in a different way. In order to test this idea, the authors
carried out quantitative cross-reactions with a number of polysac-
charides. Some of their results are shown in Table 8-6.

From the precipitation observed with the galactomannans of gum
ghatti, lucerne, and clover, Staub concluded that the Salmonella an-
tigen 4 has structural similarities with these polysaccharides ; for.
whenever the antibodies to antigen 4 were removed, precipitation of

SALMONELLA ANTIGENS

113

TABLE 8-6

Cross-Reactions of Horse Serums for 5. schottmuelleri and S. typhosa
with Certain Polysaccharides"

Precipitating

serum s

ibsorbed w

ith:

— PsPtB

PsPtBox PsTy

PsTm>'

Antibody re-

(•^,5,12) —

(12)

(^,5)

(5)

maining for

Polysaccharide

antigens

1. Anti-5.

schottmuelleri serum

Galactomannan

Gum ghatti

270 19

—

Lucerne

255 28

—

243

Clover

200 25

—

Dextran

67 —

—

2. Anti-5. typhosa

Antibody re-

P,12 9

maining for

antigens

Dextran

108 7

* Staub et al., 1959. Numbers indicate micrograms of precipitate nitrogen.
^ Polysaccharide from S. typhimuriuni. Other antigens abbreviated as in
Table 8-5.

the galactomannans was reduced virtually to zero. Antibody to anti-
gen 12, on the other hand, seems not to cross- react with these galacto-
mannans, as is shown by the fact that removal of anti-12 by absorp-
tion with polysaccharide of 5". typhosa does not mtich afifect the pre-
cipitation of the serum with lucerne.

The antibodies precipitable by dextran are seen to be, at least in
part, anti-12 antibodies, since removal of anti-12 by treatment with
PsTy considerably reduces the amount of precipitation with dextran.
This is shown even more clearly by the fact that removal of the
anti-12 from the anti-^'. typhosa serum eliminates, for all practical
purposes, precipitation with dextran.

From these results Staub and her co-workers concluded that anti-
gen 12 contains glucose units linked as they are in dextran. They
felt they could not decide whether these glucoses were in the side
chain which terminates in rhamnose or part of a chain terminating

114 INTRODUCTION TO IMMUNOCHEMICAL SPECIFICITY

in glucose. They further conckided that antigen 4 contains groupings
similar to those present in the galactomannans. These polysaccharides
contain long chains of mannose linked 1^4, with occasional side
chains consisting of galactose linked 1^6, as shown in the following
scheme :

galactose galactose

(1-6)

— (1-^4) mannose (1— »4) — mannose — (1-^4) mannose (l->4) —
It is evident that the precipitability of anti-4 antibodies by these
galactomannans is due to their specificity for a terminal galactose, a
galactose-mannose group, or a chain of mannose linked 1— >4. The
last possibility is eliminated by the fact that oxidized paratyphoid
polysaccharide still precipitates this antibody, for mannose linked
1^'4 would be destroyed by periodic acid oxidation. The probability
that the cross-reaction is due to a terminal galactose was diminished
by the failure of Heidelberger and Cordoba (1956) to obtain cross-
reactions with other polysaccharides containing terminal galactoses.
Also, periodic acid oxidation would destroy a terminal galactose, yet
the oxidized polysaccharide is still able to absorb out the anti-4
antibodies. One is, therefore, led to conclude that the grouping com-
mon to antigen 4 and the galactomannans is the galactose-mannose
grouping. But, since antigen 4 terminates in a nonoxidizable sugar
and the only such sugar present is abequose, the terminal portion
of antigen 4 may be :

abequose — galactose — mannose

Staub et al. (1959) were able to detect a weak cross-reaction be-
tween S. schottmuelleri and 5. nezvport owing to the terminal
abequose which forms part of antigen 4 in the former and part of
antigen 8 in the latter. This cross-reaction took place with horse
serum only, which suggested that the horse produces antibodies spe-
cific for the terminal sugar more readily than the rabbit does.

In later work Staub et al. established the terminal sequence of
sugars in antigens 1 and 12 as :

o-glucose — galactose — mannose — rhamnose
The linkages between the glucose and galactose are different in the

SALMONELLA ANTIGENS

lis

Abequose

Poratose Rhamnose

Glucose

GROUP A (I, 2, 12)
)Tyvelose

Rhamnose

Glucose

GROUP B ((I), 4, 5, 12)

GROUP D (9, 12)

GROUP Cg (6, 8)

GROUP P (35) E. coli

Fig. 8-2. Scheme showing our present knowledge of the role of known sugars
in the specificity of some Kauffmann- White antigens (Staub, 1959). Ellipses
indicate bacteria, projecting chains O antigens.

two antigens, probably 1-^6 in antigen 1 and 1— >4 in antigen 12.
(Staub, 1960; Stocker, Staub, Tinelli, and Kopacka, 1960; Tinelli
and Staub, 1960).

A summary of the conclusions of Westphal, Staub, et al. about the

116 INTRODUCTION TO IMMUNOCHEMICAL SPECIFICITY

antigenic structure of several Salmonella species in terms of chemi-
cal structure of the Kaufifmann- White classification is shown in Fig.
8-2. To anybody familiar with the (largely unavoidable) vagueness
of serological methods of classifying bacteria, the concreteness of
the new results will seem like a ray of light in a dark room. We may
confidently anticipate that this ray will grow brighter until the whole
intricate structure is illuminated.

References

Boivin, A., and L. Mesrobeanu, 1933, Cowpt. rend. soc. hiol. 112, 76, 611, 1009;

113,490; 114, 307 ff.
Boivin, A., and L. Mesrobeanu, 1935, Rev. immunol. 1, 553.
Boivin, A., and L. Mesrobeanu, 1936, Rev. immunol. 2, 113.
Boivin, A., and L. Mesrobeanu, 1937, Rev. immunol. 2, 113, 3, 319.
Davies, D. A. L., 1955, Biochem. J. 59, 696.

Davies, D. A. L., 1960, Advances in Carbohydrate Chem., in press.
Davies, D. A. L., I. Fromme, O. Liideritz, A. M. Staub, and O. Westphal,

1958, Nature 181, 822.
Dubos, R. J., 1952, Bacterial and Mycotic Infections of Man, 2nd ed., Lippincott,

Philadelphia.
Dubos, R. J., 1945, The Bacterial Cell in Its Relation to Problems of Virulence,

Immunity and Chemotherapy, Harvard University Press, Cambridge.
Fouquey, C. J. Polonsky, and E. Leder, 1957, Bull. Soc. Chim. Biol. 39, 101.
Freeman, G. G., 1942, Biochem. J. 36, 340.
Goebel, W. F., 1945, /. Exptl. Med. 81, 315.
Heidelberger, M., and F. Cordoba, 1956, J. E.vptl. Med. 104, 375.
Kaufifmann, F., 1950, The Diagnosis of Salmonella Types, C. C. Thomas,

Springfield.
Kaufifmann, F., 1951, Enterobacteriaccae , Ejnar Munksgaard, Copenhagen.
Kaufifmann, F., 1937, Z. Hyg. Ifektionskrankh. 120, 177.
Liideritz, O., A. M. Staub, S. Stirm, and O. Westphal, 1958, Biochem. Z. 330,

193.
Mikulaszek, E., et al., 1956, Ann. inst. Pasteur 91, 40.
Morgan, W. T. J., 1937, Biochem. J. 31, 2003.
Morgan, W. T. J., and S. M. Partridge, 1940, Biochem. J. 34, 169.
Morgan, W. T. J., and S. M. Partride, 1942, Brit. J. E.vptl. Pathol. 23, 151.
Morgan, W. T. J., and S. M. Partridge, 1941, Biochem. J. 35, 1140.
Pon, G., and A. M. Staub, 1952, Bull. soc. chim. biol. 34, 1132.
Schmidt, G., E. Eichenberger, and O. Westphal, 1958, Experientia 14, 289.
Staub, A. M., 1960, Ann. inst. Pasteur 98, 814.
Staub, A. M., 1960, Ann. inst. Pastetir, 98, 814.

Staub, A. M., and R. Tinelli, 1957, Bull. soc. chim. biol. 39 (Suppl. 1), 65.
Staub, A. M., R. Tinelli, O. Liideritz and O. Westphal, 1959, Ann. inst. Pasteur

96. 303.

SALMONELLA ANTIGENS 117

Stockcr, B., A. M. Staub, R. Tinelli, and B. Kopacka, 1960, Ann. insf. Pasteur

98, 505.
Tinelli, R., and A. M. Staub, 1960, Bull. soc. chini. biol.. 42, 583, 601.
Westphal, O., 1960, Angnv. Chcm. 72 (Dec.)
Westphal, O., 1952, Angczv. Chcm. 64, 314.
Westphal, O., and O. Liideritz, 1954, Angezv. Chcm. 66, 407.
Westphal, O., O. Liideritz, I. Fromme, and N. Joseph, 1953, Angczv. Chcm. 65,

555.
Westphal, O., O. Luderitz, A. M. Staub, and R. Tinelli, 1959, Zcntr. Baktcriol.

I. Orig. 174, 307 ff.
White, P. B., 1956, Further Studies t>l the Salmouella Group. Great Britain

Med. Res. Council Special Rep. Series No. 103, 1951, London.

CHAPTER 9

Union of Antibody with Antigen :
Thermodynamics

The exact mechanisms by which antibodies produce their effects
have not been cleared up in all cases, but that the first step is com-
bination of the antibody and antigen is not in dispute. It is therefore
of interest to inquire into the forces involved and the firmness of
the union. A proper treatment of these points will require the intro-
duction of a few elementary thermodynamic notions.

Forces Involved

Landsteiner (1936) pointed out that the covalent bond (e.g.,
the bond holding the two carbons together in ethane, H3C — CH3)
does not generally form fast enough and is not reversible enough
to be a plausible explanation of antibody-antigen reaction and that
some compounds which can react with antibodies cannot form covalent
bonds. Similar arguments probably rule out the coordinate link or
semipolar double bond.

We are left with three possibilities : coulomb forces, van der Waals
forces, and hydrogen bonding. Coulomb forces are those causing
positive and negative charges to attract each other. All antibodies
and many antigens are proteins, and it is pertinent to remark that
prominent among the charged groups in protein molecules are the
positive free e-amino groups ■ — NH3+ of lysine and the negatively
charged free carboxy groups — COO~ of the dicarboxylic amino
acids such as aspartic acid. A separated, fixed, pair of positive and
negative charges constitutes a dipole. It is easy to see how dipoles

118

UNION OF ANTIBODY WITH ANTIGEN

119

+ -+ -+ - +

Schemes of dipole association

First step

Second step

Third step

Attraction of o dipole by an ion
Fig. 9-1. Schemes showing dipole-dipole association and attraction of a
dipole by an ion.

may attract other dipoles as a result of coulomb forces, or attract
positive or negative ions (Fig. 9-1).

Van der Waals forces constitute the most general intermolecular
attraction and may operate between any two molecules. They depend
not upon permanent but upon instantaneous dipole moments. A mole-
cule which has no permanent dipole moment, for example methane
(CH4), may have at a certain instant an instantaneous dipole mo-
ment when the center of charge of the rapidly moving negative elec-
trons surrounding the carbon nucleus lies to one side of the center
of charge of the positive nucleus. This instantaneous dipole moment
produces an instantaneous electric field which may influence another
molecule in the immediate neighborhood. As a result the electrons
of the second molecule move relative to their nucleus in such a way
as to produce a force of attraction for the first molecule.

Van der Waals forces decrease very rapidly with distance, being
inversely proportional to the seventh power of the distance, and are

120 INTRODUCTION TO IMMUNOCHEMICAL SPECIFICITY

consequently negligible when two molecules are separated by any ap-
preciable distance. They are quite strong between molecules that
can bring parts of their "surfaces" into close contact.

Hydrogen bonds (also relatively short-range j consist essentially
of a hydrogen atom which is attracted simultaneously to two dififerent
atoms. For example, the two oxygens in salicylaldehyde (Fig. 9-2)

.0.

-c=o

Salicylaldehyde
Fig. 9-2.

are connected by a hydrogen bond. Many of the unusual properties
of water are due to hydrogen bonding. It is believed that hydrogen
bonds play an important part in maintaining the characteristic con-
figurations of protein molecules.

The role of coulomb forces in holding antibody and antigen to-
gether was removed from the realm of pure hypothesis by the ex-
periments of Singer (1957). This worker pointed out that if a nega-
tively charged group is involved in an antibody-antigen bond, it is
possible to calculate the effect of pH on antibody-antigen combination.
The assumption is made that, if the negative group is in the antigen,
the antibody contains a corresponding positively charged group, and
vice versa. For our present purposes it is immaterial which molecule
contains the negative group. Singer and Campbell (Singer, 1957)
suggested that, if there is one negative group characterized by an
intrinsic hydrogen ion association constant K^ and if we neglect the
nonspecific repulsion between antibody (Ab) and antigen (AG)
molecules, the following relation should hold in the acid region :

log (l/K-l/Ko) =\og(Ku/Ko) - pH (1)

where K is the apparent equilibrium constant at a given pH for the
reaction

Ab+Ag;eAbAg
and A'o is the value of K at neutral pH where both the positive and

UNION OF ANTIBODY WITH ANTIGEN

121

the negative group are fully ionized. A similar relation would apply
in the alkaline region. If two negative and two positive groups were
critically involved in each Ab-Ag bond, the expected relation would
now contain a (pH)- and a 2(pH) term.

Singer tested this relation by ultracentrifugal observations on Ab-
Ag mixtures at different pH. Typical results are shown in Fig. 9-3,

Fig. 9-3. Ultracentrifugal diagrams of mixtures of bovine serum albumin
and its antibody at various pH. Sedimentation is proceeding in the direction of
the arrow. Ag stands for antigen, Ab for antibody, a for an antibody-antigen
complex thought to be AgoAb, b represents a complex thought to be AgsAb-,
and 5 is gamma globulin. At pH less than 4.5, progressively larger amounts of
free gamma globulin (antibody) appear, while the amounts of the complexes
diminish (Singer, 1957).

which shows the sedimentation diagrams of mixtures of bovine
serum albumin and rabbit anti-bovine serum albumin. As the pH
falls, more and more free gamma globulin (antibody) appears in the
mixture while the amount of the antibody-antigen complexes de-
creases. The changes are clearly a function of pH* and were fomid
to be entirely reversible.

Enough results at different pH were obtained to show that the
linear relation predicted by the equation holds quite well (Fig. 9-4).
This was found to be true for both systems studied, namely, oval-
bumin-antiovalbumin and bovine serum albumin-antibovine-serum-

* Habeeb et al. (1959), however, conclude from chemical modification studies
that "the removal of the positive charge on the same amino groups of Ab by
an increase of the pH of the solution, instead of by acetylation, might have
the same effect on the Ab molecule and its capacity to precipitate witli a
large Ag molecule. The generally-observed dissociation of Ag-Ab bonds in
alkaline solution might therefore be attributable to such a deformation of the
Ab molecule, rather than ... to titration of specific critical groups within the
Ab sites."

122

INTRODUCTION TO IMMUNOCHEMICAL SPECIFICITY

albumin. The constant A'h had in both cases a vahie of about 105,
which is consistent with the idea that a carboxyl group, — COO~,
is critically involved in the antibody-antigen bond in these systems,
and must be ionized for maximum bond strength. Singer concluded
that the attraction of this group for its complementary positive group
accounts for about half of the strength of the antibody-antigen bond
in these cases. The remainder is presumably due to some or all of
the other forces mentioned above.

-3.0

' 1

1 1

-3.4

ffV

-3.8

O^v

-4.2

1 1

3.2

3.6

Fig. 9-4. Effect of pH on antibody-antigen equilibrium in the bovine serum
albumin system, plotted according to equation (1). The slope of the line is
-1.2 (Singer, 1957).

In the case of the attraction of antibody to /'-(/''azophenylazo)-
benzene arsenate, Nisonofif and Pressman (1957) found that the
negatively charged - — COO~ group contributed over 4.8 kcal./mole
to the binding energy, again indicating the presence of a positive
charge in the combining group of the antibody. The uncharged
/j-phenylazo group contributed 2.3 kcal./mole.

There are a number of reasons for believing that the van der Waals
forces are among the most important of the non-coulomb forces. One
of the arguments supporting this belief derives from the fact that the

UNION OF ANTIBODY WITH ANTIGEN 123

strength of the antibody-antigen bond is greatly decreased if the
hapten or antigen combining group is sHghtly changed in shape.
This is shown by work with haptens of known chemical constitution,
such as the experiments discussed in Chapter 1, and by measurements
of the bond strength for groups of related haptens, discussed below.
The strong influence of shape suggests that close contact between the
various atoms of the combining group of the antibody and the atoms
of the haptens or antigenic combining group is necessary for a
strong antibody-antigen bond. Such closeness of contact accords well
with the suggestion, made by Hooker and Boyd (1941), Pauling and
Pressman (1945) (Fig. 2-12), and Karush (1956), that the com-
bining group of the antibody may in fact be a cavity into which the
hapten or antigen combining group fits snugly. Close fit would make
the van der Waals forces strong, and any change in the hapten or
antigen combining group that lessened that fit would markedly weaken
the strength of the bond, which is precisely what we observe.

Although in the two systems studied by Singer the non-coulomb
forces (which, if the argument in the preceding paragraph is valid,
may be second in importance) were thought to account for only
about half the strength of the antibody-antigen bond, there are
cases where the non-coulomb forces presumably account for the
entire bond. strength. These cases apply to antigens which do not con-
tain positively or negatively charged groups in their specifically re-
active portions. Good examples of such antigens are provided by the
blood group substances (Chapter 7). Here, no positive or negative
groups are present, at least not in the portions responsible for the
antigenic specificity. Yet the blood group antigens combine firmly and
typically not only with antibody but with the blood group-specific
plant proteins I have called lectins (Chapter 6). These reactions have
been studied quantitatively (Kabat, 1956; Boyd, Shapleigh, and
McMaster, 1955). Karush (1958) believes that the forces between
antibody and carbohydrate antigens are mainly hydrogen bonds.

Wurmser and Filitti-Wurmser (1950) suggest that the combining
energy of the isohemagglutinins with their receptors on the human
erythrocyte is equivalent to that of about four hydrogen bonds or
twenty van der Waals bonds. Before we can discuss such quantitative
estimates further it will ])e necessary to go into some of the concepts
of thermodvnamics.

124 INTRODUCTION TO IMMUNOCHEMICAL SPECIFICITY

Energy

We shall need to discuss only the first two laws of thermody-
namics. The first law is well known and today needs only to be
stated to be believed. It is simply that energy can neither be created
nor be destroyed. It is understood that we are not thinking of changes
involving changes in atomic nuclei ; if we were, we should have to
formulate the law more broadly.

From the first law of thermodynamics it follows that no perpetual
motion machine of the "first type," i.e., one getting all or part of its
energy from nowhere, can ever be constructed. The total energy of a
completely isolated system, therefore, remains constant (If the sys-
tem is not isolated its total energy may change from time to time.)
We designate this total energy, which may be made up of heat ( which
Count Rumford proved to be a form of energy ) or of mechanical
or chemical energy and at times of other forms, as E. The science
of thermodynamics grew out of a study of the process by which heat
may be converted by suitable machines partly into work. If we let
Q stand for the heat content of the system and W for the work done,
we may write the simple equation

AE = ^Q - AW (2)

which states that the increase in the total energy of the system,
a£, equals the heat taken up, AQ, minus the work done, AlV. This
is a statement of the first law of thermodynamics in symbols.

If we consider an extremely small change in the system and ignore
certain questions of mathematical rigor, we may replace the finite
dift'erences AE, AQ, and AlV by the differentials dE, dQ, and dW,
and write

dE = dQ - dW (3)

The meaning of this equation is not as obvious as the beginner
might think. It looks as if the equation means that, if you measure
the infinitesimal increase in the total energy of a system, you can
show experimentally that it equals the experimentally determined in-
finitesimal absorption of heat minus the experimentally determined
amount of work done. But this is not the meaning at all, for we have
no "energy meter" with which we can measure the total energy of
a system, or even the change in total etiergy. The onl}' way we have

UNION OF ANTIBODY WITH ANTIGEN 125

of getting dE is by measuring ciQ and dlJ^ and taking the difference.
It looks as if equation (3) is a trivial tautology.

This is not the case, however, because there is an essential dif-
ference between dE on one hand and dQ and dW on the other
(Klotz, 1950). For dE is an exact differential, and dQ and dlV are
not. The meaning of the mathematical term exact differential is dis-
cussed in textbooks of the calculus. Here we need only recall that,
if a differential dX is exact, the values of A' at two different points,
.Yi and Xo, depend solely on the initial and final values of the in-
dependent variables of which X is a function, whereas, if dX is inexact,
the values of X depend upon the particular route we take from Xi
to Xo. In physics, if the pressure P and volume V of steam in an en-
gine are fixed, the values of the other variables such as the tem-
perature T are thereby determined. Since the values of P and V
determine the state of the system, P and V are called the independent
variables. We could have chosen other sets of two, such as P and
T or V and T, but in the study of heat engines, where thermodynamics
originated, the set P, V is particularly useful.

We find that specifying P and V does not uniquely determine
Q or W , for the amount of heat a system may take up can vary in
spite of this, and it is well known that the portion of the heat a
machine converts into work depends on the efficiency of the machine.
Consequently, dQ and dW are inexact differentials and final values
of Q and VV depend not merely on the final values of P and V, but
on the route we choose in getting from the state Pi, V\ to Fo. f^2-
Two possible routes are shown schematically in Fig. 9-5.

On the other hand, the value of E is completely determined by P
and V, and no matter what route we take from Pi, ]\ to P^. V2,
the final value of E, E^, will he the same. Consequently, if we go from
point 1 to point 2, then back to point 1 (this we call a reversible
cyclic process), AP must equal zero, while AQ and AfF will in
general be different from zero. All this is a mere restatement of the
first law of thermodynamics, but it is of the greatest importance.

A thermodynamic quantity which depends only upon the values
of the independent variables is called a thermodynamic junction.
Thus, the total energy E is such a function. Knowing that P is a
thermodynamic function, we can write other expressions which are
also thermodynamic functions. For example, if we write

126

INTRODUCTION TO IMMUNOCHEMICAL SPECIFICITY

Fig. 9-5. Two possible reversible routes from state 1 to state 2.

E + PV

(4)

it is obvious that // is a thermodynamic function, for we have seen
that E depends solely on the values of P and V, and the product
PV depends only on these variables. Therefore dH is an exact dif-
ferential. The quantity H is called the total heat, or enthalpy.
Enthalpy is important because, when the pressure remains constant,
which it does during most chemical reactions, the change in enthalpy
is equal to the heat absorbed or heat given off, as follows :

±(AH)p = ±(A(2)p (5)

The subscript P indicates that the variable P (pressure) remains
constant. The sign convention, positive for heat absorbed and nega-
tive for heat given off, was more natural in the study of heat en-
gines than it is in chemistry, but is now firmly established.

A// is interesting because in many cases a large negative \H
for a chemical reaction goes along with a strong tendency for the re-
action to go spontaneously. Indeed it was long believed that —(\H)p
was the proper measure of the spontaneity of a reaction. It was
eventually found that not — (AH)p but the change in another thermo-
dynamic function, the free energy, is the proper measure of the

UNION OF ANTIBODY WITFI ANTIGEN 127

spontaneity of a chemical reaction.* The more spontaneous a reac-
tion, the stronger the chemical bonds formed. But before we can dis-
cuss free energy we must introduce the second law of thermody-
namics.

Entropy

Going back to equation (3), we may rewrite it as follows:

dQ = (IE -\- dW (6)

If the pressure on a system remains constant, any work done is the
product of the change in volume times the pressure, thus

dQ = dE -\- PdV (7j

Since we know that £ is a function of P and V, we have, by an ele-
mentary and purely formal application of the calculus

dE = {dE/dV) dV + (dE/dP) dP (8)

where g indicates partial differentiation. Substituting in equation (6),
we obtain

dQ = (dE/dV + P) dV + (dE/dP) dP (9)

Since dQ is not an exact differential, equation (9) cannot be inte-
grated as it stands. It is shown in the calculus (e.g., Osgood 1925)
that, whenever you have an equation of the form

dQ = XdV -\- YdP (IC)

where X and Y are functions of the variables P and V, there is al-
ways an integrating factor B = j(P,V), in fact a number of such

*A reaction may be spontaneous and, nevertheless, not take place at any-
appreciable rate of speed under ordinary conditions. For example, the reaction

2H2 -1- O2 ^ 2H..0

has a large negative \H and is also spontaneous by the free energy criterion
(see below). Nevertheless, mixtures of gaseous hydrogen and oxygen can be
stored indefinitely at ordinary temperatures and pressures. The reaction is
spontaneous, however, as is clear from what happens when an electric spark
is passed through the mixture.

128 INTRODUCTION TO IMMUNOCHEMICAL SPECIFICITY

functions of P and /', such that when both sides of equation ( 10 ) are
multipHed by one of them, the product B dQ becomes an exact
differential. In the present case such a function is easily found. The
simplest one is \/T. Multiplying by \/T, we obtain

dQ/T = [{dE/dV + P)/T]dV + [{dE/dP)/T]dP (11)

That dQ/T is an exact differential is proved in thermodynamics
(Klotz, 1950) by showing that dQ/T is an exact differential (a)
for an ideal gas carried through a certain sequence of reversible
changes called a Carnot cycle, (b) for any substance carried through
a Carnot cycle, and (c) for any substance carried through any re-
versible cycle.

The sequence of changes which constitute a Carnot cycle is so
simple and symmetrical that it is easy to show that, for such a cycle
and by virtue of part (c) of the above-mentioned proof for any
reversible cycle,

W/Q2 = {T, - T,)/To (12)

where W is the work done by the system during the cycle, Qo is
the heat taken in at the higher temperature To, and Ti is the lower
temperature. The fraction W/Qo is called the efficiency of the cycle.
In thermodynamics it is further proved that (a) the efficiency of
a real substance carried through a Carnot c}-cle cannot be greater
than that of an ideal gas and cannot be less, and (b) the efficiency
of any substance carried through any reversible cycle is the same as
that of an ideal gas carried through a Carnot cycle. The fraction
IV/Q2 is therefore the maximum theoretical efficiency of any heat
engine which takes in heat Qo at temperature To and returns part
of the heat to the surroundings at temperature T]. The efficiency of
an actual engine will be less than this : it is impossible for the
efficiency of any engine, actual or theoretical, to be more.

Since dQ/T is an exact differential, it can be integrated. As a
result of this integration we shall obtain a function of the independent
variables P and V. This is a new thermodvnamic function, and we
can give it a name. The name of the new function is entropy. It is
represented by the symbol S. and we write

dS = dQ/T (13)

UNION OF ANTIBODY WITH ANTIGEN 129

The discovery that the integral of dQ/T* defines a new thermo-
dynamic function constitutes also a discovery of the second law of
thermodynamics. It is probahly the best way of introducing the con-
cept of entropy, which is not, like the concepts of temperature, pres-
sure, heat content, etc., an obvious generalization of ideas already
more or less familiar to the non-scientist but a subtle and powerful
new concept. The best attempt to explain the concept in words is to
say that it is a measure of the disorder of a system, or of the extent
of the loss of availability of energy.

The second law of thermodynamics can be stated in words in a
number of other ways, though none of them adequately suggests the
significance and applicability of the principle. For example, we may
say that no heat engine can produce work by taking a quantity of heat
from the environment at a certain temperature and returning the
unused heat to the environment at the same temperature. Such an
engine would be a perpetual motion machine of the "second type,"
and the second law of thermodynamics asserts that no such machine
can ever be constructed.

The significant thing about the second law for chemists is that it
provides a valid measure of the tendency of a process to take place,
when the change in entropy

(the subscripts meaning volume and energy are constant) is large
and positive, the process will tend to take place spontaneously, and
this tendency is greater the larger a5"f,£;.

Free Energy

Although fine for the processes that take place in heat engines,
as a measure of the spontaneity of a chemical reaction ^.Sv.e has its
drawbacks. In chemical reactions, the volume of the system often
does not remain the same and the energy practically never does.
Pressure and temperature are usually constant but volume and
entropy vary. Consequently, we want a new thermodynamic function.

* The dQ in the definition of entropy must be the heat absorbed in a re-
versible process and is sometimes written explicitly rfQrev.

130 INTRODUCTION TO IMMUNOCHEMICAL SPECIFICITY

let us say F, such that Ft,p = j(V,S). We can get this rather simply
by the definition

F = H - TS (14)

It is easy to show that F is a function of V and 5" when P and T
are constant. So is TS, obviously : when T is constant it is a function
of 5" alone. From our original definition of H we have

H = E -\- PV (15)

When P is constant, PF is a function of V only. We saw above that
£ is a function of P and V only ; consequently, when P is constant,
£ is a function merely of V. Therefore,

F = H - TS = E + PV - TS (16)

is a function of V and S. Consequently, the thermodynamic function
F defined by this expression is a function of V and S. The new func-
tion is called the Gibbs free energy.

When T and P are constant, we have from equation (16)

A£p,r = A£ + PAF - rA5 (17)

Now, from equation (3) above, we have Aii = AQ — Al>F. If we
ignore complications such as osmotic effects, the only work the system
does is mechanical, MV = P AV, and AE = AQ — P AV. Substitut-
ing this into equation (17), w^e obtain

AFp,T = AQ - PAV -\- PAV - TAS (18)

From the definition of entropy, AQ = T AS ior a. reversible process,
we find that for a reversible process, or at equilibrium,

AFp,T = O (19)

If the pressure does vary but the temperature continues to remain
constant, we have from equation (16)

dF = dE + P dV + J' dP - T dS

Again, dE = dQ - P dV = T dS - P dV, and we obtain

dPr = VdP

For a perfect gas we have PV = nRT, or V = nRT/P, so that

UNION OF ANTIBODY WITH ANTIGEN 131

dF= - {nRTdP)/P

Integrating, we obtain

F^ - f., = _AF = nRT\n{Pi/P2) (20)

Spontaneously, a perfect gas can only expand ; it cannot spontaneously
contract. In other words, the pressure can only decrease. From this
we see that in a spontaneous reaction AF will be negative. The larger
the negative value of AF, the greater the tendency of the process to
go.

Strictly, equation (20) applies only to a perfect gas. But it also
applies without serious error to many real gases. If we replace Po
and Pi by the thermodynamic activities, which for the dilute solu-
tions used in immunochemistry do not differ appreciably from the
molar concentrations, we may apply this equation to antibody and
antigen solutions.

Free Energy and Equilibrium

We now proceed to derive an important relation between AF and
the equilibrium constant of a chemical reaction. Let us suppose we
have a reaction between two perfect gases A and B, to give two
other perfect gases, C and D. Then if we represent the numbers
of moles involved by lower case letters, a, h, c, and d. the initial
pressures as Pa and Pb, and the final pressures as Pc and Pd> we
have to write

aA{PA) + bB{PB) -^ cC(Pc) + dD(P,>) + AF (21)

where AF represents the change in free energy which accompanies
the reaction. In order to compare free energy changes, and therefore
tendencies of reactions to take place, we need free energy changes
where the starting and stopping points are always the same ; in
other words, all reactants must start at a standard state and finish
up in a standard state. In the case of gases the standard state is
atmospheric pressure. In the case of dissolved substances, which we
mostly deal with in immunochemistry, the standard state is unit
activity.

We can find the free energy change, called the standard free
energy change and represented by AF°, which would result if the

132 INTRODUCTION TO IMMUNOCHEMICAL SPECIFICITY

reacting gases shown in equation (21) started at atmospheric pres-
sure and the products ended up at atmospheric pressure. We simply
systematically add to equation (21) a series of equations, each one
of which carries one of the gases from the standard pressure to the
partial pressure Pa, Pb, etc., actually observed, adding also each
time the free energy change which such a change in pressure entails.
For instance, the first equation we add is

aA{PA = 1) -^ aA{PA = Pa), AF = oRT\n{PA/\) (22)

After performing all these additions we combine the logarithmic
terms and obtain

AF° = AF+ RT\n[(PAy(PB)W{Pcy{PDy]

AF° = AF - RT\n[{Pcy{PDY/{PAY{PBy] (23)

If the amounts a, b, etc., and the pressures Pa, Pb, etc., are those
found at equilibrium, the free energy change AF in the reaction
shown in equation (21) is zero, and the term AF drops out. And
since we see that the expression whose natural logarithm appears
in (23) is in that case simply the equilibrium constant K, equation
(23) reduces to

AF° = - RT In K (24)

which is the relation we were seeking. Again we see that when
there is a strong tendency for the reaction as written to go to the
right (K is large), AF° will be large and negative.

The equilibrium constant of a reaction is a measure of the extent
to which a reaction goes to completion. The standard free energy
change, which can be calculated from it, is thus a proper measure
of the strength of the chemical bonds that are formed, and broken,
during the reaction. Whenever the equilibrium constant of a reaction
can be measured, we can calculate the standard free energy change.
If we know AF°, we can calculate the entropy change A^"", if
AH° is known from calorimetric measurements, by using equation
(14) in the form AF° = AH° — T A5"°. A//° has been measured
directly for only a few immunochemical reactions. When it cannot
be measured it can often be calculated from van't Hoff's equation

UNION OF ANTIBODY WITH ANTIGEN 133

^(In K)/dT = AH°/RT' (25)

If we assume A//° is independent of T in the range of temperatures
studied, we can integrate equation (25) to obtain

\n(K,/Kr) = - {AHyR)[l/T, + l/T^] (26)

which makes it possible to estimate AH° if observations on the
equihbrium constant are available at two different temperatures.
In fact this is the commonest way of obtaining AH°.

References

Boyd, W. C, M. Shapleigh, and McMaster, 1955, Arch. Biochem. Biophys. 55,

226.
Habeeb, H. F. S. A. et al., 1959, Biochim. ct Biophys. Acta 34, 439.
Hooker, S. B., and W. C. Boyd, 1941, /. Immunol. 42, 419.
Kabat, E. A., 1956, Blood Group Substances, Academic Press, New York.
Karush, F., 1956, /. Am. Chem. Soc. 78, 5519.
Karush, F., 1958, Trans. N. Y. Acad. Sci. 20, 581.
Klotz, I. M., 1950, Chemical Thermodynamics, Prentice-Hall, Englewood Cliff's,

N. J.
Landsteiner, L., 1936, The Specificity of Serological Reactions, C. C. Thomas,

Springfield.
Nisonoff, A., and D. Pressman, 1957, /. Am. Chem. Soc. 79, 1616.
Pauling, L., and D. Pressman, 1945, /. Am. Chem. Soc. 67, 1003.
Osgood, W. F., 1925, Advanced Calculus, Macmillan, New York.
Singer, S. J., 1957, /. Cellular Comp. Physiol. 50, Suppl. 1, 51.
Wurmser R., and S. Filitti-Wurmser, 1950, Biochim. et Biophys. Acta 4, 238.

CHAPTER 10
Energy of Antibody-Antigen Reactions

Direct Calorimetry

In the early days of immunochemistry, methods were not available
for measuring the amounts of free antigen or antibody, or both, re-
maining after antibody and antigen have reacted. Therefore calcula-
tion of the free energy change from direct measurements of the equi-
librium constant was not possible. The earlier estimates of the strength
of the antibody-antigen bond were based on attempts to measure the
heat of reaction AH. It will be seen from equation (14) in the previous
chapter, which we can rewrite as follows,

AF° = AH° - TAS° T = const. (1)

that if the entropy change were zero, AH° would equal AF° , and
such a measurement would be an adequate measure of the strength
of the antibody-antigen bond. In fact, we may regard equation (14)
as a statement that in order to make AH° a reliable index of the
tendency of the reaction to take place, we have to correct it by al-
lowing for the entropy change AS° . A positive entropy change will
make a negative AF° still more negative, a negative entropy change
will make it less negative or even positive. Somewhat unexpectedly,
recent work suggests that in serological reactions AS° is. in fact, not
large, though usually not zero.

Nevertheless, not too much has been learned about antibody-
antigen reactions by direct calorimetry. The first attempt, by Bayne-
Jones (1925), gave results that we now know were nearly a million
times the correct value. Two later determinations, the first by

134

ENERGY OF ANTIBODY-ANTIGEN REACTIONS 135

Kistiakowsky and his group (Boyd et al.^ 1941), and the second by
Steiner and Kitzinger (1956), gave —40 and —6 kcal per mole of
antibody, respectively. I doubt if this difiference v^as due to experi-
mental error ; more likely it should be traced to differences in features
between the two very different antibody-antigen systems used.
Steiner and Kitzinger 's result agrees better with values of AH°
calculated indirectly for other serological reactions, as we shall see
below.

Free Energy from Equilibrium Measurements

Various methods have been used to measure the equilibrium be-
tween free antibody and antigen and their compounds, or between
antibody and hapten and their compounds, including (i) equilibrium
dialysis, (ii) direct analyses of precipitates and supernatants, (iii)
electrophoretic and ultracentrifugal observations, and (iv) light scat-
tering. Details of the experimental procedures will have to be found
in the references cited. Here we may say merely that all are methods
of determining or calculating the concentrations at equilibrium of free
antibody, free antigen, free rapten, or compounds thereof. From such
measurements the equilibrium constant K and AF° can be calculated.
If measurements can be made at more than one temperature, AH°
and AS° can also be estimated.

Of the above methods, (i) and (iv) are applicable only to simple
antibody-hapten systems, the former only to univalent hapten sys-
tems. Method (iii) can be applied to systems in which the antibody
is reacting with a protein, but application of the method may in some
cases disturb somewhat the very equilibrium which it is desired to
measure. Method (iv) does not disturb the equilibrium.

In applying method (iii), allowance must be made for the fact
that antibody is divalent, at least usually, and protein antigens are
multivalent (Epstein, Doty, and Boyd, 1956). Therefore, if we
measure the equilibrium in which each antibody is combined with as
many antigen molecules as possible (two in the case of divalent anti-
body), in the presence of free antigen and the compound AG, where
A represents antibody and G represents antigen, our equilibrium
constant corresponds to

(AG2)/(G) (AG) = K (2)

136 INTRODUCTION TO IMMUNOCHEMICAL SPECIFICITY

What we are interested in, however, is the strength of a single
antibody-antigen bond. The compound Ad:- contains two such bonds,
and each mole of free antigen G contains %' moles of free combining
sites, where v is the valence of the antigen. Consequently, we have to
obtain the value of K\ where K' is the equilibrium constant corre-
sponding to the equilibrium

(antibody-antigen bonds)/

(free antigen sites) (free antibody sites) = K' (3)

by waiting

2(AG2)A(G) (AG) = K'
or

K' = {2/v)K (4)

Therefore, the standard free energy of a single antibody-antigen bond

A7^°, equals -RTXnK' = -RTlnK - RT\n{2/v),or

AF°, = -RTlnK + RT\niv/2) (5)

The exact value of the correction will depend on the valence of the
antigen and the exact nature of the reaction the equilibrium state of
which is being studied.

As an illustration, let us consider the results of Baker et al. (1956)
on the reaction of anti-benzenearsonic acid antibodies with ben-
zenearsonic acid-azo-bovine serum albumin (bovine serum albumin
coupled with diazotized arsanilic acid ) . The reaction studied by these
workers was

A' ; + AG ^ AC-
and their bovine serum albumin contained thirteen introduced
benzenearsonic acid azo groups per molecule. They calculated a AF°
of — 5.2 kcal. per mole. From the above this is equivalent to a bond
free energy change AFi° of

-5.2 -\- 7?rin(13/2)
or

-5.2 + 1.1 = -4.1 kcal. /bond
Contrary to expectations, this value is less (i.e., more positive) than

ENERGY OF ANTIBODY-ANTIGEN REACTIONS 137

the value of —7.4 kcal. per bond found by Epstein, Doty, and Boyd
(1956) for the reaction of anti-benzenearsonic acid antibodies with
the divalent hapten T (terephthalanilide-/',/''-diarsonic acid) (Fig.
10-1).

H203As<' p>NHOC<r ^CONH<^ ^AsOaH,

Fig. 10-1. Divalent hapten used by Epstein, Doty, and Boyd (1956).

It would have been expected that the benzenearsonic acid groups
in the coupled bovine serum albumin, being coupled through the azo
linkage with tyrosine and histidine residues just as in the coupled
protein used for immunization, would correspond to the combining
sites of the antibody better than the amide-coupled benzenearsonic
acid groups of the hapten T. Epstein, Doty, and Boyd suggested
that the decreased bond strength was due to some unfavorable feature
in the orientation of the groups in the coupled protein.

In dealing with multivalent antigens which may combine simul-
taneously with a number of molecules of antibody, the mathematical
problems of formulating the reaction become formidable unless we
introduce simplifying assumptions. The simplest assumption is that
the free energy of combination of an antibody molecule with a com-
bining site of the antigen is the same for all such sites and is not
afifected by the number of antibody molecules which have already com-
bined with the antigen. With the aid of this assumption, which can
hardly be strictly true but which is certainly adequate as a first ap-
proximation, we can easily solve the problem, as shown by Linder-
str0m-Lang (1924), von Muralt (1930), Eowler (1936), Wyman
(1943), and Klotz (1946). If we let the association constant for
the formation of a single antibody-antigen bond be K, and the num-
ber of combining sites on the antigen molecule (or cell) be ;//, we
find the ratio r of antibody molecules combined with an antigen mole-
cule (or cell) to be

r = mKiA)/l\ + A'(A)] (6)

where (A) is the concentration of free antibody.

A summary of the principal thermodynamic studies on the antibody-
antigen or antibody-hapten reaction is given in Table 10-1.

138

INTRODUCTION TO IMMUNOCHEMICAL SPECIFICITY

-a

03 ~

°9

II •= oa"

1^ T! ""

1 00 00

00 Ov

z 1 ^

^ -H -H

O ro O O t^

4^ -H ^ooiq

— oo-Hc^t^oodoo

<» -H (>J (M ™ 1

^ -M u

- rt g

§ H -^

OJ O fcuO

r^ (M

"n 00 00 vo ■^ t^ ON

£ '-' —

:§ -H -H

1 .^ 1

O (TJ O O 1^

r-j O (>i d -; !>.■ O^ ro'

-S fe :s

^ 1 1

II .= 0)

1 ' '

1 1 1 1 1 1 1

^it

CQ ^ §

- tn en

.5 '" <«

O o) r^4

rv) tN

00 o o"

1 -H -H

d d

2^ ^ ..,

^ lO lO vO

1 CVJ lO -+ 00 ^) CN in

^ _0 T3

O *^ lO lO On

"^ 1 1 1 1

. liS 00 t^ NO w 1^ vd

2 1 1 1 1 1 1 1

5 *i cn

1 ' ' ' '

.H I- "o

> OJ 3

^1!

,-^ ,— - ^^

II -v a

^, r^ ^ O -H

c -^

c:j- c- 2^ C- c.

ssssssss

.s" "^"^ "S

^_^

S o S

-o

^ s -

"2 3^

S tj >.

■§

a = -T3

•5-

^c^l

= c c

(!!)
(!!!)
(!!!)

fn r;^ •=■ .T o .o o .-^i^

11 « rt

< S "1

< U Q Q CQ

cq Q CQ fc U W U

rt .5 c5 "cS

e oj

^' ^ ^ !

-M 3

, Hey
, HSA
. BSA
, 0

, Car boxy

-a

'lis

o -a

-H CN PO '* lO

vd i-^ 00 o\ d -^ CN qJ

T-H .rt .^H

ENERGY OF ANTIBODY-ANTIGEN REACTIONS

139

TABLE 10-la
Reactions for which Data Are Presented in Table 10-1

A. nA + G^ A„G

B. A + A„-iG ;=i A„G

C. A + 2G ^ AG2

D. AG + G ;^ AG2

E. A + 2H ^ AHo

F. Haptenic group + antibody site ^ hapten-antibody bond

A = antibody, G = antigen, H = hapten.

AsOjHj

N = N —
COOH

SO3H

N^N-

CH,

CONHC6H4ASO3H2

CONHCcH.AsO^H,

./^

CHNHCO<

coo ^

>N(CH3)2

N(CH,U

Fig. 10-2. Structures of haptens referred to in Table 10-1.

140 INTRODUCTION TO IMMUNOCHEMICAL SPECIFICITY

References for Table 10-1 :

1. Baker, M. C, D. H. Campbell, S. I. Epstein, and S. J. Singer, 1956,
/. Am. Chem. Soc. 78, 312.

2. Boyd, W. C, J. B. Conn, D. C. Grcgs', G. B. Kistiakowsky, and R. M.
Roberts, 1941, /. Biol. Chem. 139, 787.

3. Carsten, M. E., and H. N. Eisen, 1955, /. Am. Chem. Soc. 77, 1273.

4. Epstein, S. I., P. Doty, and W. C. Boyd, 1956, /. Am. Chem. Soc. 78,
3306.

5. Haurowitz, P., C. F. Crampton, and R. Sowinski, 1951, Federation Proc.
10, 560.

6. Karush, P., 1950, /. Am. Chem. Soc. 72, 2705.

7. Karush, P., 1956, /. Am. Chem. Soc. 78, 5519.

8. Karush, P., 1957, /. Am. Chem. Soc. 79, 3380.

9. Singer, S. J., and D. H. Campbell, 1955, /. Am. Chem. Soc. 77, 3499.

10. Singer, S. J., and D. H. Campbell. 1955, /. Am. Chem. Soc. 11, 4851.

11. Smith, E. L., et al., 1952, /. Biol. Chem. 199, 789.

12. Steiner, R. P., C. Kitzinger, and T. H. Benzinger, 1956, Research Rept.
Vaval Med. Research Inst. 14, 73.

Significance of Thermodynamic Constants

The figures in Table 10-1 present some unexpected features. Most
surprising, perhaps, is that AF° is generally not large ; —9 kcal per
mole seems to be about an upper limit. This is not a large value for
standard free energy changes. The free energy of formation of water,
for example, is — 54.65 kcal. per mole (for two hydrogen-oxygen
bonds) ; that of carbon monoxide is —33.0 kcal. per mole (for one
carbon-oxygen bond). On the other hand, it can be seen from Fig.
10-3 that the free energy changes involved in the formation of the
antibody-antigen bond are sufficient to cause the reaction to go sub-
stantially to completion if the reagents are concentrated. (This figure
shows the relation between the equilibrium constant K and the free
energy change. Also shown is the per cent of product B at equilibrium
in a hypothetical reaction A ^ B.)

Not only are the values of AF° small by physical and chemical
standards, but the values for the different reactions are surprisingly
alike, suggesting that no antibody-antigen reaction is likely to have
a large free energy change. If antibody is formed through contact
with a molecule or portion of a molecule of antigen or with some

ENERGY OF ANTIBODY-ANTIGEN REACTIONS

141

Per cent of B

40 50 60

1 1 1

-1

^— -

— -

-2
-3

—

-4

1 1

1 1 1 1 1 1

1 1

1 1 1

3 4 5

10 20

0.05 0.1 0.2 03 0.4 0.5 1.0 2.0

Equilibrium constant {K)
Fig. 10-3. Relation between standard free energy change (AF°) of a reaction
A-^B and the equilibrium constant. Also shown is relation between AF" and
the per cent composition of the equilibrium mixture with respect to B. (Slightly
modified from H. B. Bull, 1951, Physical Biochemistry, 2nd ed., Wiley, New
York, by permission).

intracellular template which causes part of the new molecule to have
a configuration complementary to the antigenic determinant, a small
value for AF° is understandable. An antibody molecule that possessed
too strong an affinity for the fixed antigen molecule or intracellular
template would have difficulty leaving its place of formation and
getting into the circulation, as has been pointed out by Pauling
(1940) and Singer (1957).

Another unexpected feature of Table 10-1 is that the values of
A^°, with two exceptions, are positive instead of negative. When
antibody molecules combine with a molecule of antigen, their freedom
of motion is restricted, and this loss of freedom constitutes a loss of
"configurational entropy." Therefore, one would expect antibody-
antigen reactions to be accompanied by a decrease in entropy. The
positive values reported therefore demand explanation.

It has generally been supposed that the positive values for A5°

142 INTRODUCTION TO IMMUNOCHEMICAL SPECIFICITY

are due to the fact that in most cases there is mutual neutralization
of positive and negative charges (see p. 120), with resulting loss
of attraction for water molecules. Restoring freedom of motion to
water molecules previously bound to the antigen or antibody surface
causes an increase in entropy, and this might be more than enough
to compensate for the loss of entropy due to the decreased mobility
of the antibody molecules. For instance, Epstein, Doty, and Boyd
(1956) calculated that in the reaction studied by them the release
of about twenty-four water molecules accounted for the observed
A6"°. In line with this argument, Karush (1958) found a negative
entropy change of about nine units for the reaction of antibody with
his lactose-hapten "lac," where there is no charge to be neutralized.

The one large negative entropy change in study 1 of Table 10-1 is
harder to explain. However, it should be remembered that, in the
first place, it is based on a value of AF° which was merely assumed
and, in the second place, hemocyanin is a rather special antigen in a
number of ways, being much larger and more multivalent than most
antigens and constituting an associating and dissociating system.
Steiner and Kitzinger (1956) suggested that a change in the state of
association of the hemocyanin might account for the large enthalpy
change observed and for the large negative entropy change calculated
from this value.

A third feature of the results of Table 10-1 is that the enthalpy
(heat content) changes are small, with the exception, again, of that
found in study 1. Aside from this perhaps atypical value, the largest
enthalpy change in the table is the — 9.7 kcal. per mole calculated by
Karush (1958) for the reaction of antibody with the "lac" hapten.
This is definitely on the small side when compared with the AH°
of —94.03 kcal. per mole for the reaction of hydrogen and oxygen to
form water, or the — 26.4 kcal. per mole for the reaction of carbon
and oxygen to form carbon monoxide. It is also of interest that, in
all cases where AH° is not zero, or so close to zero that its exact
magnitude is not known, it is negative, i.e., the reaction is exothermic.

The enthalpy changes of all the antibody-antigen or antibody-hapten
reactions studied, with the exception of that in study 1, are too small
to account for the firmness of the bond and the fact that the reaction
goes to substantial completion. Obviously, in many, perhaps most,
cases the major portion of the driving force of the reaction AF° is

ENERGY OF ANTIBODY-ANTIGEN REACTIONS 143

contributed by the term T ^S° and is thus due to the positive entropy
change (equation (25).

In spite of the relative weakness of the antibody-antigen or anti-
body-hapten bond, antibodies display very sharp specificity, as we
have already seen. For instance, when Karush compared the reac-
tions of anti-"lac" antibody with lactose with the reaction of the
same antibody with cellobiose, he found a value for Ai^" of — 5.52
kcal. per mole for lactose and only — 1.96 kcal. per mole for celloboise,
although the only difference between the two sugars is the arrange-
ment of the hydrogen and hydroxyl groups on carbon number 4
of the terminal hexose unit (Fig. 10-4). This again accords with
the notion that the hapten fits quite precisely into a portion of the
antibody.

The importance of close fit of antibody to hapten is also shown
by the work of Nisonoff and Pressman (1957) who found that
substitution of an iodine atom ortho to the carboxy group of the
benzoate ion decreased the antibody-hapten combining energy by
2.4 kcal. per mole. Substitution of an iodine in the meta position
decreased the binding energy by about 0.7 kcal. per mole.

It has been known for some time that the antibody molecules in
any given antiserum are heterogenous (references in Boyd 1956).
This heterogeneity manifests itself, among other ways, by differences
in their specific affinity (Karush, 1958; Epstein, Doty, and Boyd,
1956; Nisonoff and Pressman, 1958). This means that the AF°

CHjOH H

Lactose

144

INTRODUCTION TO IMMUNOCHEMICAL SPECIFICITY

for the antibodies of an antiserum will be different for the different
antibodies ; combination of antigen or hapten will be firmer with some
than with others. Heterogeneity of antibody is responsible, for ex-
ample, for the fact that the relation between r/c and r, where r is the
average number of hapten molecules bound per antibody molecule
and c is the concentration of free hapten, is not a linear one (see Fig.
10-5). It was suggested by Pauling, Pressman, and Grossberg
(1944) that the standard free energy of combination of the various
antibody molecules may follow the distribution of the normal error
function (Gaussian distribution) (see Fig. 2-4). This suggestion has
been worked out in detail by Karush (Karush and Sonenberg,
1949; Karush, 1956).

If Kq is the average binding constant and a a measure of the

( \J

1 1 1
7.1° C.

1 1

—

25° C. \
\ \

\ \

••

°\

1 1 1

1 ^^1

Fig. 10-5. Binding results at 25 °C. and 7.1 °C, for the reaction between D-Ip
hapten and purified anti-D-Ip antibody (Karush, 1957, 1958). The points are
experimental and the curves theoretical.

ENERGY OF ANTIBODY-ANTIGEN REACTIONS 145

range of values of K, then the "normaHzed" Gaussian function
(PauHng, Campbell, and Grossberg, 1944) is

[1/V(7r)cr]exp[ - \niK/Ko)/aY

The fraction of total combining sites, n, which have a specified
binding constant K will be, for an infinitesimally small area (in, ex-
pressed as follows :

dn/n = [1/ V (7r)a] exp[ - \n(K/K,/a]- d \n{K/K,)

From this Karush and Sonenberg (1949) found (the derivation is
given by Klotz, 1953) that the fraction of antibody sites occupied,
7'/n, where n is the number of combining sites per antibody (found
by Karush to be two in confirmation of much earlier work), is in
terms of the concentration c of free hapten, as follows :

r/n = 1 - [ 1/V (tt)] I {[1 - exp( - a^-)]/[i + ^V exp(a'(7)]l da

where a is [ln(A7A'o)] A. Karush (1957, 1958) found that if
for his D-Ip anti-D-I,, system he took the heterogeneity index o- of
antibody to be 2.3, the above equation enabled him to account satis-
factorily for the experimentally formed relation between r/c and r.
(See Fig. 10-5, where the circles are the experimental points and the
curves are theoretical.)

Heat of Reaction of Isoagglutinins

The thermodynamic constants for the reaction of the human iso-
hemagglutinins have been estimated by Wurmser and Filitti-Wurmser
(Filitti-Wurmser, Jacquot-Armand, and Aubel-Lesure, and Wurmser,
1954; Wurmser and Filitti-Wurmser, 1957), who have devoted a
great deal of penetrating thought and experimental skill to the prob-
lem. The methods used are somewhat different from those involved
in the studies just discussed and deserve a little space to themselves.

Wurmser and co-workers showed that the combination of iso-
agglutinins with human erythrocytes is reversible, so that equilibrium
considerations apply. We can use equation (25), which gives us a
relation between the equilibrium constant K and the concentration of
free antibody at equilibrium. Equation (6 ) contains two unknown
constants : the number of combining sites on a red cell, in, and the

146 INTRODUCTION TO IMMUNOCHEMICAL SPECIFICITY

association constant K. Precise values of m are not yet available,
but Wurmser and Filitti-Wurmser devised methods of calculation
which did not require a knowledge of m.

If we invert both sides of equation (6), we obtain

\/r = \/m + \/mK(A) (7)

This means that if we plot the reciprocal of the number of moles
of agglutinin combined with a mole of red cells against the reciprocal
of the concentration of free agglutinin, we should get a straight line
with slope 1/mK. If we make such determinations at two different
temperatures, the ratio of the two slopes (l/mK2)/{l/mKi) gives
us the ratio of the association constants at these two temperatures,
Kx/K2. From this we may calculate AH° from van't Hoff's equation
(25) (p. 133).

The amount of isoagglutinin remaining free in equilibrated mix-
tures of erythrocytes and serum cannot be estimated with sufficient
accuracy by the method of serial dilutions generally used to estimate
the strength of an agglutinating serum, and the quantitative methods
of Heidelberger and his school are not sensitive enough. But the
Wurmsers hit upon the device of expressing the agglutinin con-
tent of their sera in terms of the maximum number of red cells
they agglutinate, and of determining the free agglutinin in the super-
natant of erythrocyte-agglutinin mixtures in the same way. This
enabled them merely by cell counting to obtain the data for determin-
ing the requisite slopes and ratios of slopes described in the last para-
graph (Fig. 10-6).

The values of AH° calculated by these methods are shown in Table
10-2. It will be seen that these values of AH° are in several cases
larger than the rather small values calculated by other workers for
other antibody-antigen and antibody-hapten systems. The most sur-
prising feature of Table 10-2, however, is the marked differences in
the anti-B isoagglutinin values obtained from the blood of persons of
different blood group and even of different genotype. This has been
confirmed by the examination of the serum of 36 AiO individuals, six
of group AiAi, and eight of group OO. The anti-B in the serum of
any given individual seems always to be homogeneous. This homog-
eneity is in marked contrast to the heterogeneity found for immune
antibodies (p. 14) and, if confirmed, might go far toward supporting

ENERGY OF ANTIBODY-ANTIGEN REACTIONS

147

2.0

0 0.5 1.0 1.5

1 //V4 (reciprocol of concentration of free onti-B) xiO^
Fig. 10-6. Relation between reciprocals of fraction of anti-B agglutinin com-
bined with erythrocytes and concentration of free anti-B agglutinin, at 37 °C.
and 25 °C., showing linear relationship and different slopes at the two tempera-
tures (Filitti-Wurmser et al., 1954).

TABLE 10-2
Heat of Combination of Isohemagglutinins with Erythrocytes"

Isoagglutinin

Agglutinogen

Genotype of donor

Ai7°,kcal./mole

Anti-B

AiO

-16 ± 2

Anti-B

AiAi

-6.5 ± 1.1

Anti-B

A2O

-9

Anti-B

-1.7 ± 0.4

Anti-A

-10 ± 3

Anti-Ai

-33 ±2.5

" Wurmser and Filitti-Wurmser, 1957.

the views of workers such as Furiihata (1927) who postulated that
the isoaggkitinins anti-A and anti-B were as much a product of the
blood group genes as the agglutinogens A and B were.*

''On the whole the less probable view, see Chapter 4, p. 57.

148 INTRODUCTION TO IMMUNOCHEMICAL SPECIFICITY

By estimating the molecular weights of the isoagglntinins and by
estimating m by determining the amount of protein nitrogen taken
up by erythrocytes from agglutinating sera, Wurmser and Filitti-
Wurmser were able also to obtain approximate values for the free
energy and entropy changes for these anti-B-B reactions. The ap-
proximate molecular weights obtained are shown in Table 10-3.

TABLE 10-3
Approximate Molecular Weights of Human Anti-B Isoagglutinins*

Genotype of donor

M'^

AiO
A,Ai
GO

~500,000
-200,000
-125,000

" Wurmser and Filitti-W'urmser, 1957.
^ Molecular weight.

It will be seen that the different kinds of anti-B, according to
Wurmser, also differ in molecular weight.

The calculated free energy and entropy changes are shown in
Table 10-4. It will be seen that these results suggest that the binding
energies AF° are not very different for the three kinds of anti-B, but
that the differences in AH° correspond to significant differences in
A^'". Wurmser and Filitti- Wurmser concluded that the specific com-
bining groups of the three different kinds of anti-B are not very
different and suggested that the increase in entropy which results
when the anti-B of group O serum combines with B erythrocytes may
be connected with a perturbation of the entire protein molecule, pos-

TABLE 10-4

Free Energy and Entropy Changes for Binding of Anti-B Isoagglntinins
by B Erythrocytes''

Genotype of donor AF°,kcal./mole A5°,e.u.

GO -9.2 -1-24

AiAi -9.5 +9.7

A,0 -9.8 -20

» Wurmser and Filitti-Wurmser, 1957.

ENERGY OF ANTIBODY-ANTIGEN REACTIONS 149

sibly some sort of reversible denaturation which results in greater
disorder and consequent absorption of heat. These effects might mask
the evohition of heat and the decrease in entropy which are caused by
the local reaction of the specific combining group with the B receptor
on the cell.

According to Wurmser and Filitti-Wurmser, these findings in-
dicate that the isoagglutinins in man are produced not by a process
of immunization, as antibodies are in general, but directly under
control of the blood group genes. This view, if correct, would sharply
distinguish these "natural agglutinins" from those produced by im-
munizing animals, or presumably even from those produced by in-
jection of A and B blood group substances into human volunteers.

Kabat (1956) does not believe such a distinction exists, and does
not believe there are such substances as "natural agglutinins." He
has also criticized the calculations of Wurmser and Filitti-Wurmser
in detail, but to me none of his criticisms seem conclusive ; indeed,
some seem quite beside the point. (Cf. Filitti-Wurmser, Jacquot-
Armand, and Wurmser, 1960.) A decision as to the validity of the
Wurmser and Filitti-Wurmser conclusions will have to await con-
firmation or disproof of their work in another laboratory.

References

Baker, M. C, D. H. Campbell, S. I. Epstein, and S. J. Singer, 1956, /. Am.

Chcm. Soc. 78, 312.
Bayne- Jones, S., 1925, /. Iiiiiiiuiiol. 10, 663.

Boyd, W. C, 1956, Fundamentals of Immunology. Interscicnce, New York.
Boyd, W. C, J. B. Conn, D. C. Gregg, G. B. Kistiakowsky, and R. M. Roberts,

1941, /. Biol. Chem. 139, 787.
PIpstein, S. I., P. Doty, and W. C. Boyd, 1956, /. Am. Chcm. Soc. 78, 3306.
Filitti-Wurmser, S., Y. Jacquot-Armand, G. Aubcl-Lesure, and R. Wurmser,

1954, Ann. Eugenics 18, 183.
Filitti-Wurmser, S., Y. Jacquot-Armand, and R. Wurmser, 1960, Rev. hematol.

15, 201.
Fowler, R. H., 1936, Statistical Mechanics, Cambridge University Press,

Cambridge.
Furuhata, T., 1927, Japan Med. World 7, 197.

Kabat, E. A., 1956, Blood Group Substances. Academic Press, New York.
Karush, F., 1956, /. Am. Chem. Soc. 78, 5519.
Karush, F., 1957, /. Am. Chcm. Soc. 79, 3380.
Karush, F., 1958, Trans. N. Y. Acad. Sci. 20, 581.
Karush, F., and M. Sonenberg, 1949, /. Am. Chem. Soc. 71, 1369.

150 INTRODUCTION TO IMMUNOCHEMICAL SPECIFICITY

Klotz, I. M., 1953, Protein Interactions, in H. Neurath and K. Bailey (eds.),

The Proteins, vol. 1, part B, p. 727. Academic Press, New York.
Linderstr0m-Lang, K, 1924, Compt. rend. trav. lab. Carhbcrg. Ser. cliim. 15,

no. 7.
von Muralt, A., 1930, /. Am. Chem. Soc. 52, 3518.
Nisonoff, A., and D. Pressman, 1957, /. Am. Chem. Soc. 79, 1616.
Pauling, L., 1940, /. Am. Chem. Soc. 62, 2643.
Pauling, L., D. Pressman, and A. L. Grossberg, 1944, /. Am. Chem. Soc. 66,

784.
Singer, S. J., 1957, /. Cellular Comp. Physiol. 50, Supl. 1, 51.
Steiner, R. P., C. Kitzinger, and T. H. Benzinger, 1956, Research Kept. Naval

Med. Research Inst. 14, 73.
Wurmser, R., and S. Filitti-Wurmser, 1957, Progress in Biophysics 7, 88.
Wyman, J., 1944, in Cohn, E. J. and J. T. Edsall (eds.), Proteins, Aniinoacids

and Peptides, Reinhold, New York, p. 451.

Index

A antigen, subdivisions of, 53, 54
A substance, formula, 90

(hog), 75
Abequose, 108, 110, 114, 115

formula, 109
ABH blood group antigens, 85

analytical results, 86
ABO blood groups, 51, 55

complications of, 55
Abriis prccatorins, 65, 66
Absorption, of lectins, 74

of serum, defined, 13
iV- Acetylgalactosamine (A^-Acetyl-D-

galactosamine), 76, 77, 78, 87
7V-Acetyl-D-galactosamine, formula, 78
A/'-Acetyl-D-glucosamine, 77
A^'-Acetylneuraminic acid, 101
"Acquired immunological tolerance"
(immunological tolerance), 31, 2)6,
2,7
Agglutination, heat of, 145 fif.

photographs of 3
Aldohexoses, steric relationships, 80
Aldopentoses, steric relationships, 80
D-Allose, formula, 80
L-Allose, formula, 80
D-Altrose, formula, 80
L-Altrose, formula, 80

Amino acids, inhibition by, 102
Analyses of blood group antigens,
Anemia, acquired hemolytic, 27
Anti-A-lectins, 66, 67, 76

inhibition of, 87
Anti-Ai lectin, 67, 75
Anti-(A+B) lectin, 77
Anti-B lectin, 68
Antibodies, formation of, 29 ff.

molecular weight, 5

nature of, 4 fif.

reactions of, 2

role in immunity, 1, 2

specificity of, 7

valency of, 29
Antibody-antigen

136
Antibody-antigen

involved, 118 fif.
Antibody molecules, cavities of, 2i

models of, 5

photographs of, 6
Antigenic determinants, 38, 42, 47
Antigenicity, 34
Antigens, chemical alteration of, 8

defined, 34

molecular weight of, 35
Anti-Gy lectin, inhibition of, 96
Anti-H agglutinins, 54, 67, 76
Anti-H lectins, 67

bond, strength of,
combination, forces

151

152

INDEX

Antihapten antibodies, detection of, 10
Anti-Lewis agglutinins, 92
Anti-M from horses, 70
Anti-M lectin, 67
Anti-N lectin, 67, 69
D-Arabinose, 11, 96

formula, 80
L-Arabinose, 78, 96

formula, 80
Arachis hypogaca, 70
Ascarylose, 108

formula, 109
ATWOOD and SCHEINBERG, 68
Autoantibodies, 27
Autoimmunization, 2>7
AVERY, GOEBEL, and BABERS,

45
AVERY, HEIDELBERGER, and
GOEBEL, 26

B substance, formula, 90

BAKER, 136

Bandeiraca siwplicifolia. 68, 79, 81, 82

Basques, 61

Bauhinia purpurea. 81, 82, 94, 95

Bauhinia receptor, 93

Beans, 66

BIRD, 67, 82

Blood group antigens, 85 ff.

sources of, 85
Blood groups, inheritance of, 52
BOIVIN, 104
"Bombay" gene, 55
BOORMAN and DODD, 58
BOYD, 2, 7, 23, 35, 56, 80, 97, 135
BOYD and BOYD, 55, 58
BOYD, EVERHART, and McMAS-

TER, 67, 69, 94
BOYD and REEVES, 101
BOYD and REGUERA, 67
BOYD and SHAPLEIGH, 55, 68, 75
BOYD, SHAPLEIGH, and McMAS-

TER, 72, 123

BOYD and WASZCZENKO-

ZACHARCZENKO, 81, 95
BURNET, 29, 36
BURNET and FENNER, 30

Calorimetry, 134

Calpurnia aurea, 82

CAMPBELL, 27, 120

CAMPBELL and BULMAN, 27

Caragana spp., 79, 81

Carbohydrate haptens, 46

Carbohydrates, anttigenicity of, 36

Carnot cycle, 128

Cavities in antibody molecule, 28

CEBRA, 38

Cellobiose, 95, 96

Cellobioside, B, 44

Chimpanzees, 69

Clover gum, 113

GLUTTON, and HARINGTON, and

YUILL, 2, 35
Cold, common, 1
Colitose, 108, 110

formula, 109
Colominic acid, 101
Complement, 2
CoroniUa varia, 79, 81
Coulomb forces, 118
Covalent bond, 118
Cross-reactions, 112
Crotalaria spp., 79, 81
Cytisus sessijolius, 67
Cytisus spp., 79, 81

D antigen, structure of, 98, 99
DAVIES, 107

5-Deoxy-3-formyI-L-lyxose, 98
2-Deoxy-L-fucose, 77
Deoxy sugars, 77, 107, 108
6-Deoxy-L-talose, 77

INDEX

153

Dctcnninants, antigenic, 38

carbohydrate, 46

peptide, 38
Dextran, 36, 113

antibodies to, 46

composition of, 46
Dicarboxylic amino acids, 118
3,6-Dideoxyhexoscs, 108, 109
Diethylene glycol, 103
D-Digitoxose, Tl
Diphasic bacteria, 105
Dipole, attraction by an ion, 119
Dipole association, 119
DODD, BIGLEY, and GEYER, 101
DOTY, 36

Dolichos bifloriis, 67, 68, 75
DUBOS, 36, 106
Duffy blood groups, 62

Eel, 53, 75

EHRLICH, 29

ELO, ESTOLA, and MALM-
STROM, 68

Endotoxins, 103
composition of, 104

Energy, 124

ENSGRABER, KRi)PE, and ENS-
GRABER-HATTINGEN, 73

Enthalpy (H), 126

Enthalpy changes in serological reac-
tions, 138

Entropy, 127 ff.
defined, 128

Entropy change in serological reac-
tions, 138

Enzymes, inhibition of, 91
specificity of, 64

Erythrina spp., 79, 81

EPSTEIN, DOTY, and BOYD, 137,
143

Equilibrium constant, relation to free
energy change, 131-133

Error function, 14

Erythroblastosis fetaiis, etiology of, 58,

59
Escherichia coli, 115
Euonymus eiiropcns. 68
Exact differential, 125

FILITTI-WURMSER. JACQUOT-
ARMAND, and AUBEL-
LESURE, 145
FINLAND and CURNEN, 26
First law of thermodynamics, 124
FISHER, 22
Forssman antigen, 75
FOWLER, 137
Free energy, 129 ff.

defined, 130

and equilibrium, 131

from equilibrium measurements, 135
Free energy changes in serological re-
actions, 138
D-Fucose, 78
L-Fucose, 76, 86, 91
FURUHATA, 57, 147

Galactomannans, 113
D-Galactose, 78, 89, 95, 96, 110, 111,
114

antibodies to, 25

formula, 80
L-Galactose, 77

formula, 80
Galactoside, )3, 44
Gamma globulin, 5, 121
GAUSS, 14
Genes, action of, 92, 93
Gentiobioside, jS, 45
"GIL," 43

Glucose, antibodies to, 25
D-Glucose, 95, 96, 98, 110, 111, 115

formula, 80

154

INDEX

L-Glucose, 97, 98

formula, 80
Glucoside, a, 44

/3, 44
Glycine, 40
Glycine soja, 79, 81
Goat, 53

GOEBEL, 46, 104
GOEBEL, AVERY, and BABERS,

38,45
GRABAR, 65

Gram-negative bacteria, definition, 103
Gram stain, 103
D-Gulose, 97, 98

formula, 80
L-Gulose, formula, 80
Gum ghatti, 113

H (enthalpy), 126

H antigens (of Salmonella), 105, 106

H blood factor, 54

H substance, formula, 91

HABEEB, 121

HACKEL, SMOLKER, and

FENSKE, 97
Hapten, defined, 8
Haptens, carbohydrate, 46

synthetic, in thermodynamic studies,

139
HAUROWITZ, 29, 30, 35
Heat of combination, of isoagglutinins

with erythrocytes, 147
Henshaw antigen, 58
HEIDELBERGER and WOLFRAM,

26
Hemocyanin, 142
Hexosamine, 86
Homogeneity of lectins, 73, 74
HOOKER, 2, 29
HOOKER and BOYD, 13, 28, 35, 38,

123
Hunter antigen, 58
Hydrogen bonding, 118

Hydrogen bonds, 120, 123

D-Idose, formula, 80
L-Idose, formula, 80
Immunity, defined, 1
Immunological tolerance, 36
"Incomplete" antibodies, 29
Information and specificity, 32
Inhibition, 19, 20, 47, 98. 100, 101

of anti-C, 100

of anti-D, 98, 100, 101

of antidextran serum, 47

of anti-E, 100, 101

of enzymes, 91

by haptens, 19, 20

principle of, 16

quantitative, 20, 110
Inhibition reactions, 15, 20, 110
Invagination, in antibody molecule, 29
Isoagglutination, enthalpy changes, en-
tropy changes, free energy changes
in, 148
Isoagglutinins, heat of combination, 147

molecular weights of, 148

origin of, 56

J substance of cattle blood, 75
JERNE, 29, 30, 31,83

KABAT, 5, 36, 38, 47, 48, 86, 88, 89,

90, 123, 149
KARUSH, 64, 123, 143, 144, 145
KAUFFMAN, 107, 108, 109, 115
Kell blood groups, 62
Kidd blood groups, 62
KISKIAKOWSKY, 135
KLOTZ, 125, 128, 137, 145
KLOTZ, WALKER, and PIVAN, 64

INDEX

155

KOSYAKOV, 58, 62
KRUPE, 67, 68, 11, 78, 79

Laburnum alpinum, 61

Lactose, 78, 95, 96

Lactoside, jS, 45

LANDSTEINER, 8, 13, 15, 38, 39,

40,41,65, 118
Landsteiner blood groups, 51
LANDSTEINER and LEVINE, 57
LANDSTEINER and VAN DER

SCHEER, 13, 24, 27
LANDSTEINER and WIENER, 58
Lafhynis latifoliiis, 19, 81
LAU, 65

Lea antigen, 85, 86
Lea substance, formula, 91
Lectin, defined, 68
Lectins, 64 ff.
discovery of, 66
inhibition of, 76
lessons from study of, 83
role in plant, 82
specificity of, 75
LEDERBERG, 8, 29, 31
Lens culinaris, 66, 80, 81
Lentils (see Lens culinaris)
Leucine, 40
Leukopenia, 31
LEVINE, 69
LEVINE, CELANO, LANGE, and

BERLINER, 70
LEVINE and STETSON, 58
LEWIS, 36
Lewis blood groups, 62
Lima bean lectin, 66, 72, 76

properties of, 72
Lima beans, 66, 68
LINDERSTR0M-LANG, 137
Lipopolysaccharides, 107
Lotus tctragonolobus, 67, 11, 81
Lucerne gum, 113

Lupus erythematosis, 31

Lutheran blood groups, 62

Lymphocytes, 31

Lysine, 118

Lysis, 2

D-Lyxose, formula, 80

L-Lyxose, formula, 80

MAKELA, 67, 78, 81,96

Makela's classification of sugars, 78,

79,98
MAKELA and MAKELA, 68
Maltose, 95, 96
Maltoside, /3, 45
D-Mannose, 96, 98, 110, 111, 114

formula, 80
L-Mannose, 97, 98

formula, 80
Marasmius orcadcs, 68
MAURER, 35
MEDAWAR, 31
Melibiose, 78, 96

formula, 89
a-Methyl-galactoside, 89
;8-Methyl-galactoside, 89
MILULASZEK, 107
MNS blood groups, 57 flF.
Molecular weights of antibodies, 5, 148
MORGAN, 36, 104
MORGAN and KING. 85
MORGAN and WATKINS, 54, 69^

75, 16, 11, 86, 87, 88, 92
Mummies, 56
Mushroom, 68

N in horse erythrocytes, 70
"Natural agglutinins," 149
NISONOFF and PRESSMAN, 122,
143

156

INDEX

North Africa, 61, 62

O antigens (of Salmonella), 105
O blood group, 51, 53
OCCAM, 29
ONCLEY, 101

OTTENSOOSER and SILBER
SCHMIDT, 67

P antigen, 57
Paratose, 108, 110, 115

formula, 109
Paratyphoid B (see SalmoncUa schott-

mucllcri)
Parkia filicoidca, 79, 81
PAULING, 7, 28, 29. 30, 141
PAULING and ITANO, 27
PAULING and PRESSMAN. 123
PAULING, PRESSMAN, and

GROSSBERG, 14, 144
"Peanut receptor" (Gy), 70, 95

formula, 97
Peptides, antibodies to, 39 ff.
Periarteritis nodosa, Z7
Periodic acid. 111
Phascolus limcnsis, var. macrocarpus

(see Lima beans)
Pisiim sativum, 80, 81
Plant agglutinins, 65 ff.

specificity of, 65, 66
Pneumococcus, 57
Polymers, 35, 36
Precipitation of Salmonella antigens,

111
Precipitations, photographs of, 3
Precipitin reaction, with A substance,
73

of lectins, 73
PRESSMAN, 29

Purpura, idiopathic thrombocytopenic,
37

Ouclluiig, 4

R antigen (sheep), 75
RACE and SANGER, 55
Raffinose, formula, 89
RENKONEN, 67
Rh, discovery of, 58

in various populations, 61
Rh antigens, stability of, 62
Rh blood groups, 58 ff.
Rh inheritance, 60 ff.
Rh nomenclatures, 60
Rh receptors, 97 ff.
L-Rhamnose, 99, 110, 111, 115
Ribonucleic acid derivatives, 97
D-Ribose, formula, 80
L-Ribose, formula, 80
Ricin, agglutination by, 65

properties of, 72
Rici)ii(s connniiiiis, 65, 66, 72, 79, 81,

95
Rickettsia, 7

"Rough" forms (Salmonella), 107
Rutinose, 98

formula, 99

S (entropy), 128
S (secretor gene), 58
S antigen, 57
Saliva, antigens in, 55
Salmonella, described, 105
Salmonella abortus cqui, 110
Salmonella adelaide, 110, 115
Salmonella anatum, 106
Salmonella antigens, 103 ff.

structure of, 112-116
Salmonella budapest, 110
Salmonella choleraesuis, 106
Salmonella dar-es-salaam, 110
Salmonella cnteriditis, 106, 110

INDEX

157

SalnioncUa galluiariuii , 106, 110

Salmonella hirschjcldii, 106

Salmonella monschaui, 110

Salmonella montevideo, 106

Salmonella ncivport, 106, 107

Salmonella oranienhurg, 106

Salmonella paratyl>hi A, 105, 115

Salmonella paratyphosa. 106, 110

Salmonella salinatus, 110

Salmonella pulloniin, 106

Salmonella schottmucllcri, 106, 110,
112, 113, 115

Salmonella scndai, 105, 106

Salmonella Stanley, 110

Salmonella typhimuriunt, 106, 110, 112

Salmonella typhosa, 105, 106, 110, 111,
115

SANGER, 8, 55

SCHIFF and SASAKI, 55

SCHMIDT, 68

SCHMIDT, EICHENBERGER, and
WESTPHAL, 104

Second law of thermodynamics, 129

Secretors, 55
diagnosis of 55, 69

SELA, 35, 2,6

Serological reactions, thermodynamic
values of, 138

Serum albumin, 64

Shiga bacillus, 53

Silk, 38

SINGER, 120, 122, 141

SMITH, 7, 8, 56

"Smooth" forms (Salmonella), 107

Sophora japonica, 68, 77, 79, 81

Specificity, effect of chemical composi-
tion on, 12
of enzymes, 64
limitations of, 26
of plant agglutinins, 75

Spontaneous reactions, 127

SPRINGER, 27

Stachyose, formula, 89

STAHMANN, 35

Statistical methods, 20 fif.

STAUB, 107, 108, 110, 113, 115
STAUB and TINELLI, 110
STAUB, TINELLI, LtJDERITZ,

and WESTPHAL, 108, 109
STEINER and KITZINGER, 135,

142
formula, 99
Stereoisomerism, 23 ff.
Steric relationship, of aldopentoscs

and aldohexoses, 80
Streptomycin, 98
Strychnine, antibodies to, 18
Subgroups of ABO blood groups, 54
SVENSSON, 5
Switzerland, 61

TALMAGE, 29, 31, 32

D-Talose, formula, 80

L-Talose, formula, 80

Tartaric acid, antibodies to, 24, 24

TAYLOR and BOYD, 56

Template theory of antibody formation,

30
Terminal unit of antigenic groups, 40
Thermodynamic function, defined, 125
Thermodynamics, 124 fif.

first law, 124

second law, 129
TINELLI and STAUB, 115
TISELIUS, 5
Transfusion, 52
Trehalose, 95, 96
Trichloracetic acid, 103
Trichomonas enzymes, 91
Trypsin, 64
Tyvelose, 108, 110, 111

formula, 109

Ulex europeiis, 54, 58
Ulex extract, for diagnosis of secre-
tors, 69

158

INDEX

"Univalent" antibodies, 29
Ultracentrifugal observations, on anti-
body-antigen mixtures, 121
"Unnatural" sugars, 97, 101

Valency of antibodies, 29

Van der Waals forces, 21, 118. 119,

123
Van't Hoff's equation, 133, 146
Vicia cracca, 67
Vicia graminea, 69
Vicia spp., 81
VON MURALT, 137

WESTPHAL, 104, 105, 107, 110
WESTPHAL and LUDERITZ, 104
WESTPHAL, LUDERITZ, STAUB,

and TINELLI, 108
WHITE, 107, 115
WIENER and PETERS, 58
Wisteria chincnsis, 79, 81
WURMSER, 57, 145
WURMSER and FILITTI-WURM-

SER, 123, 145, 147, 148, 149
WYMAN, 137

D-Xylose, formula,
L-Xylose, formula.

WATKINS and MORGAN, 91
Watkins and Morgan's scheme for
gene action, 92, 93

YCAS, 26

Internet Archive Audio

Featured

Top

Images

Featured

Top

Software

Featured

Top

Books

Featured

Top

Video

Featured

Top

Mobile Apps

Browser Extensions

Archive-It Subscription

Save Page Now

Full text of "Introduction to immunochemical specificity"

See other formats