THE GENETICS OF PATHOGENIC ORGANISMS

Publication of the American Association
for the Advancement of Science
No. 12

PUBLICATION COMMITTEE

E. C. STAKMAN, Chairman
R. E. CoKER
L. O. KUNKEL
G. B. REED
Matco_m H. Soule
E. A. WATSON

Edited by
Forest Ray MouLTON

ot

—————— ~
Ze a yupibAL ity S SS
ia, e\,

(336226)

% aizd 40 4%
Saington, D.C:

 

        

Published for the American Association for the Advancement of Science by
Tue Science Press

1940
PROBLEMS IN THE VARIATION OF PATHOGENIC
BACTERIA

By G. B. REED

QUEEN’S UNIVERSITY, KINGSTON, ONTARIO, CANADA

From the early days of Pasteur and Koch
down to the last decade and a half, patho-
genic bacteriology was dominated by the
dogma of monomorphism. With few ex-
ceptions, the earlier studies were more con-
cerned with establishing the fixity of species
than with variation. But during the last 15
to 18 years the literature of the subject has
been flooded with evidences of variability.

It has become increasingly apparent that
pathogenic species of bacteria may pass
through thousands of generations under lab-
oratory conditions or in the animal body
without detectable change in their charac-
teristics. At the same time, it is equally
evident that most, and probably all, patho-
genic species at times undergo variation.
The fact that bacteria do vary has reper-
cussions in almost every field of the study
of infectious diseases. I should like to call
attention to problems connected with varia-
tions in antigenic composition and virulence
and to suggest some hypothetical consider-
ations concerning the causes of variations.

VARIATIONS IN ANTIGENIC COMPOSITION

From the early work, especially that of
Arkwright (1920), it has been apparent that
certain variations are generally correlated.
The most frequently observed variation,
the Smooth to Rough (S to R) colony form,
is not merely a change in the structure of
the colony; there is usually associated a
change in virulence, in antigenic composi-
tion and other characteristics, such as cell
morphology, sensitivity to bacteriophage,
and hydrophobe or hydrophile properties
of the cell. The variation is generally dis-
continuous and the variants are frequently
highly stable. In many species the § to R
change occurs readily, but reversion or R
to S occurs only rarely. The significance
of this type of variation is not apparent,

but it seems safe to assume that it represents
a relatively major change in the genetic
makeup of the organism.

The pathological importance of the change
is primarily due to variation in the anti-
genic composition which appears to be a
fundamental consideration in virulence as
well as in antibody stimulating and reacting
mechanism. The S to R variation ordi-
narily involves a loss in the heat-stable
somatic antigen that characterizes the sur-
face of the normal virulent type. This loss
may further result in the uncovering of
somatic antigens which then dominate the
variant type. The somatic antigen, prob-
‘ably most frequently concerned in this loss |
variation, we know, from the initial work
of Heidelberger and Avery, is a polysac-
charide. In chemically pure form the poly-
saccharides are haptens or partial antigens
—they react with antibodies but do not
stimulate antibody formation. In the cell,
however, they act as antigens. In the case
of the Pneumococci, the polysaccharide is
contained in the capsule. In the course of
the S to R variation the capsule and the
specific polysaccharide antigen are lost.
There remains, however, in the R organisms
an antigenic protein component common to
all types of Pneumococci. It has been

shown that a similar series of changes ac-

28

company S to R variation in several, pos-
sibly all, pathogenic species. This loss of
surface antigen means that the variant lacks
antigenic specificity and lacks the ability to
stimulate the formation of specific anti-
bodies in animals. It has not been possible
to put all antigenic differences between bac-
terial species or variants into chemical terms
or to interpret antigenic variation in this
precise manner, but the fact that certain
major changes can be so interpreted tends
to clarify the situation.
THE VARIATION OF PATHOGENIC BACTERIA ‘ 29

Knowledge gained from a study of vari-
ation in the antigenicity of bacteria, during
the last dozen years, has resulted in many
modifications in methods of immunization
and diagnosis by immunological means. We
now know that in the preparation of vac-
cines for immunization against typhoid,
paratyphoid, and other infections of this
group, consideration must be given to the
complex antigenic composition of the organ-
isms and the ease with which variation in
these components occurs under laboratory
conditions. Similar problems are met with
to a greater or lesser degree in the prepa-
ration of vaccines from all types of patho-
genic bacteria. Similarly, in the prepara-
tion of antisera, as against Pnewmococci and
Meningococci, possible variation in antigenic
composition is a fundamental problem. In
such a simple procedure as the Widal and
other agglutination reactions used in the
identification of antibodies for diagnosis of
infections, due regard must be given to the
antigenic structure and possible variation of
the organisms used as antigens. Such prob-
lems appear in almost every phase of im-
munology.

It will be recalled that the French group
—Calmette et al.—obtained from an old cul-
ture of tubercle bacilli a strain, BCG, of
exceedingly low virulence, a type capable
of producing only localized lesions which are
fairly rapidly absorbed. It was shown by
Petroff (1927), Begbie (1930), Sasano and
Medlar (1931), Seiffert (1932), and our-
selves (1934) that this organism is probably
an R variant of the normal tubercle bacillus.
It has been found that the strain generally
is highly stable like most R forms. In the
hands of a few observers, sufficient variation
has, however, occurred to permit recovery of
the fully virulent or normal S form (Reed,
Orr, and Rice, 1934).

The French group and others have long
claimed that the introduction of BCG cul-
tures into infants results in a considerable
degree of immunity against ordinary tuber-
culous infection. This vitally important
problem has been seriously questioned on
statistical grounds. We have approached it
differently by examining the antigenic

structure of the BCG organisms in contrast
to normal tubercle bacilli.

In earlier papers (Rice 1931; Rice and
Reed 1932) it was shown that, if rabbits are
immunized with whole cultures of normal S
forms and other rabbits with R variants of
mammalian tubercle bacilli, a distinction
can be made in the antibody content of the
two lots of animal sera. The examination
procedure was an indirect one involving
complement fixation with antigens prepared
from S and R types of organisms. The
anti-S immune serum fixed, on the average,
twice as much complement with the S anti-
gen as was fixed with R antigen. The anti-R
serum fixed, on the average, the same
amount of complement with both § and R
antigens. These observations were inter-
preted to mean that the S organisms contain
an S-specifie and a species antigen, and that
the R organisms lack the S-specific antigen
but contain the species antigen. Corre-
spondingly, the anti-S serum contains the
two antibodies; the anti-R serum, the species
antibody, while it lacks the S-specific anti-
body.

The same examination procedure was fol-
lowed with the ordinary BCG cultures and
the S type recovered from them (Reed, Orr,
and Rice 1934). Rabbits were immunized
with the two types and the antisera tested
against antigens prepared from S and R
type mammalian tubercle bacilli. The re-
sults are summarized in Table I. It is
apparent from the column on the left that
antisera prepared from several cultures of
BCG contain antibodies which fix comple-
ment in approximately equal amounts in the
presence of S and R antigen. The antisera
prepared from 8 variants of the BCG
(which are virulent), as indicated on the
right of Table I, fixed approximately twice
as much complement in the presence of the
S antigen as in the presence of the R antigen.
The simplest explanation of the observations

‘is that the BCG organism, like other R

tubercle bacilli, lacks an antigenic compo-
nent which characterizes normal virulent
tubercle bacilli. Lacking this antigen, the
corresponding antibody is lacking in the
serum of BCG immunized animals.
30 THE GENETICS OF PATHOGENIC ORGANISMS

TABLE I

A COMPARISON oF THE REACTIVITY OF TYPICAL 8 AND R MaMMALIAN TUBERCLE Bacitius ANTIGENS WITH
THE SERUM OF RaBBirs IMMUNIZED WITH ORDINARY BCG-R ORGANISMS AND WITH DissoctaTep
S ORGANISMS aS MEASURED IN CuBIC CENTIMETERS OF COMPLEMENT FIXED SPECIFICALLY AT
Farry Pee Cent HEMOLYsIs

 

 

 

Antiserum Antigen Specific . Antiserum Antigen |  SPecifie
BCG-R Montreal (1) ...} oss BOG-8 Trades on . toc
BCG-R Montreal (2) .... s ord BOG-8 Alberta (1) -. | 8 pr
BCG-R Montreal (3) .. . a BCG-S Alberta (2) ~. ; ree
BCG-R Montreal (4) .... 8 atid BCG-8 Alberta (3) .... 8 aes
BCG-R Montreal (5) 7 sors BOG-S Ottawa (1) -.. 8 NA
BOG-E Paris .......... an 8 ose BCG-S8 Alberta (4) ~.. ‘ oon?
BCG-8 & B Alberta (5) 8 0100

 

 

 

 

 

 

It seems probable that the BCG organism
has arisen as a variant from ‘‘normal’’
tubercle bacilli and that one feature of the
variation was a loss of a specific antigenic
substance. This adds support to the doubt
raised from the statistical analysis based
upon the use of the material as an immuniz-
ing agent in man and cattle. It is in line,
too, with general findings with other species
of pathogenic bacteria, such as the Pneu-
mococci, members of the typhoid coli group,
that the 8 to R type of variation is accom-
panied by a loss of a specific antigen which
renders the variant ineffective as an immun-
jzing agent.

VARIATION IN VIRULENCE

A long recognized type of variation is
change in virulence or change in ability to
produce toxins. This change occurs in many
species under laboratory conditions, and we
know that avirulent or atoxogenic variants
may frequently be isolated from man
(Wadsworth and Sickles 1927; Okell 1929;
Gunn and Griffith 1928).. Variation in the
animal body is probably related to the inter-
action of antibodies with the surface anti-
gens of the organisms.

An attractive explanation of the rise and
fall of epidemics has been based, with some
degree of experimental support, upon these
findings. The extent to which host immu-
nity and bacterial variation contribute to
the course of epidemics has never been
determined, but there seems to be ample
evidence from both the study of epidemic
disease in man and experimental herd infec-
tions in animals that variation in the causal
bacteria is at least a factor of importance,
even though it be not a simple and obvious
relationship.

In a long series of studies of herd epi-
demics in mice, Webster and colleagues
(1930, 1933) have shown that different
strains of the same bacterium recovered at
intervals during the course of one epidemic
may possess an approximately equal viru-
lence. This they regularly found to be the
case. They, therefore, conclude that changes
in virulence play no part in the fluctuations
in mortality frequently observed in a long
continued epidemic. They have, however,
reported different degrees of virulence in
the same bacterial species recovered in dif-
ferent epidemics (1930). At the same time,
they demonstrated that in one species of
THE VARIATION OF PATHOGENIC BACTERIA 31

Pneumococeus (Webster and Clow 1933),
a strain of high virulence, as judged by
intraperitoneal and intranasal infection,
might as a result of nose-to-nose passage,
lose its intranasal virulence without affect-
ing its intraperitoneal virulence.

On the other hand, Topley, Greenwood,
and associates (1928) have considered a
second factor. They find that virulence, as
measured by direct inoculation, and the

power to induce severe epidemics by contact .

infection are not always correlated. Recent
experiments with a species of Pasteurella
(Greenwood ef al. 1936) seem to afford
proof that changes in infectivity may occur
during the epidemic spread of a bacterial
parasite. They, therefore, suggest that an
‘epidemic strain’’ is characterized by two
attributes, high virulence and high infee-
tivity. Loss of either through variation may
result in loss of ability to cause an epidemic.
It seems safe to conclude from these herd
studies that variation in the invading bac-
teria may constitute a major factor in the
spread of human infections. There is also
much supporting evidence to be found in a
study of human infections.

. THEORETICAL CONSIDERATIONS

The similarity of the variation pattern in
many species of bacteria and especially the
variation of correlated characteristics, as
just noted, have contributed largely to the
theory of cyclic change in the life history
of the bacteria. But variation by no means
always occurs simultaneously in several
characteristics. In a recently published
study (Reed 1937) of the saprophytic spe-
cies, Serratia marcescens Bizio, it was shown
that variation may oecur independently in
several characteristics.

In this species, as is generally the case,
colony structure, cell form, and specific
somatic agglutinogen are always associated :
S colonies are composed of short rods and
coceoid cells containing specific somatic
antigen, while the R colonies are composed
of long rods and filaments without the
specific somatic antigen. Where variation
occurs from § to R or in the reverse direc-
tion, the associated characteristics vary syn-

chronously. This may arise from an insep-
arable linkage of the inheritance factors or
these three characteristics may result from
the transmission of a single gene. What is
measured as somatic antigen may, for in-
stance, determine the cell form and cell form
may determine the colony topography.

Several other characteristics exhibited in-
dependent variation. Capsulated cells with
the resulting mucoid structure of colonies
occur in both S and R forms. Variation
may proceed in the direction of either gain
or loss in capsulation, and this variation may
occur irrespective of variation in S or R
colony form. It is evident, therefore, that
the factor for the inheritance of mucoidness
or its absence is independent of the factor
or factors for colony structure. Again,
variation in pigmentation from red to
orange-red or white, or from orange-red to
white, occurs regardless of variation in
colony structure or capsulation of the cells.
Color factors must, therefore, be inherited
independently of factors for colony struc-
ture or for capsulation.

Independent variation in three sets of
characteristics does not appear to be con-
sistent with any theory of cyclic change in
the organisms. It seems more likely that,
in these cases at least, the inheritance of
individual characteristics has been subject
to some irregularity. It is more in line with
current genetical opinion to assume that this
is the result of change in individual genes
or ‘‘gene mutation.’’ A study of Siaphylo-
cocct now in progress adds additional evi-
dence in support of this hypothesis.

There are, however, certain well known
variations in pathogenic bacteria which pos-
sibly permit a simplification of the gene
hypothesis, a simplification which may be
applicable to all inheritance and variation
in the bacteria, and perhaps in a wider field.

Since the work of Neufeld in 1909, we
have been familiar with the existence of
antigenically different types of Pneumococct.
Three types and a heterogeneous group were
first distinguished, and quite recently the
unclassified group has been differentiated
into some 27 additional types. There is
ample evidence that the antigenic difference
32 THE GENETICS OF PATHOGENIC ORGANISMS ©

TABLE II

CHARACTERISTICS OF THE TYPE-SPECIFIC POLYSACCHARIDES OF PuEvMOCcOCcCI
(After Avery and Heidelberger)

 

 

 

 

 

: ‘ . : Dilution
tical Per cent Substances obtained so:
Type Op . : . giving specific
rotation nitrogen on hydrolysis precipitation
To resecssee 300° 5 Galacturic acid and amino-sugar de-

THVAEEV OR aeesecnnnncnenentnnensnnnestnsnnnntneensttn 1: 6,000,000
rr Ww. 74° 0 Glucose nsetssnnsstnsnnretereennenenennnttnt 1: 5,000,000
TI ....... — 33° 0 Glucose and. glucoronic acid .—-—-.. 1: 6,000,000

 

 

between at least the original three types
depends upon chemical differences in the
polysaccharide, which is the principal com-
ponent of the capsule surrounding the
organism, as indicated in Table II.

As just noted in’ another connection,
variation frequently occurs in the strain of
this species in which the capsule and type-
specific polysaccharide is lost. Such
variants, therefore, do not belong to a
specific type; they have lost their type char-
acteristic. It was shown by Griffith (1928)
and recently confirmed particularly by Daw-
son and Sia (1931) that, if a non-capsulated,
non-type-specific variant arising from type I
is grown in a menstruum including type I
polysaccharide, reversion to a capsulated
type-specific form may occur. And, very
significantly, the new form will be not type I
from which the non-capsulated strain arose
but type II, the type of the polysaccharide
present in the menstruum in which the
variation occurred. Once this new type IT
has arisen, it continues to breed as a char-
acteristic type II; that is, the newly gained
ability to synthesize the type II polysac-
charide is passed on from generation to
generation. ,

The fact that the type-specific polysac-
charide is taken up from the environment
and continues to be formed in subsequent
generations suggests that it may be a self-
propagating substance like Stanley’s virus
protein. If so, the type-specific inheritance
may be regarded simply as a cutting off, in
the fission division, of a portion of the sub-
stance sufficient to permit the propagation
of more similar substance in the developing
daughter cell. A loss variation is then

simply a cell division in which the poly-
saccharide fails to divide. If this is true of
specific polysaccharide, possibly there are in
the cell other self-propagating substances.
Such a simple hypothetical mechanism
probably makes it unnecessary to postulate
genes, although it is conceivable that the
genes are just such substances.

REFERENCES CITED

ARKWRIGHT, J. A. 1920, Variation in Bacteria in
Relation to Agglutination by Salts and by Spe-
cific Sera. J. Path. & Bact., 23: 358.

Bresiz, BR. 8. 1930. Microbic Dissociation, with
Special Reference to Certain Acid-fast Bacilli.
Edinburgh Med. J., 37: 187.

Dawson, M. H. and 814, B. H. P. 1931. In vitro
Transformation of Pneumococcal Types. I. A
Technique for Inducing Transformation of Pneu-
mococcal Types in vitro. J. Exp. Med., 54: 681.

Grumnwoop, M., Hut, A. B., Torey, W. W. C. and
Wrison, J. 1936. Experimental Epidemiology.
Med. Res. Council Sp. Rept., no. 209.

Grirmrs, F. 1928. Significance of Pneumococcal
Types. J. Hyg., 27: 113.

Gunn, W. and GervirH, F. 1928. Bacteriolog-
ical and Clinical Study of 100 Cases of Scarlet
Fever. J. Hyg., 28: 250.

OxeLt, CO, C. Relationship of Virulent to Avirulent
Diphtheria Bacilli. J. Hyg., 29: 309.

Prrgorr, 8. A. 1927. Microbe Dissociation: The
Tubercle Bacillus. Proc. Soc. Exp. Biol. ¢ Med.,
24: 632, 956.

Rexp, G. B., Onr, J. H. and Rice, C. E. 1934.
Studies in Variability of Tubercle Bacilli; BCG.
Canad. J. Res., 11: 362.

Rgep, G. B. 1937. Independent Variation of Sev-
eral Characteristics in §. marcescens. J. Bact.,
34: 255.

Rice, C. E. 1931. Studies in the Variability of
Tubercle Bacilli. IV. Antigenic Properties of
§ and R Cultures. Canad. J. Res., 5: 375.

Rice, C. E. and Rezp, G. B. 1932. Studies in the
Variability of Tubercle Bacilli. VIL. Antigenic
Activity of ‘8’? and ‘‘R’’ Cultures as Measured
by Complement-fixation. J. Immunol., 23: 385.
THE VARIATION OF PATHOGENIC BACTERIA

Sasano, K. T, and Mepuar, E. M. 1931. Studies
on the Calmette-Guérin Strain of the Tubercle
Bacillus (BCG). I. The Effect in vitro of En-
vironment on the Virulence of BCG. Amer. Rev.
Tuberc., 23: 215.

SEIrFERT. 1932. Dissoziationsversuche mit BCG.
Klin. Wochenschr., 11: 181.

TopLey, W. W. C., GREENWooD, M., WILsoN, J. and
NEwsouip, E. M. 1928. Epidemic Poteney of
Strains of Bact. aertryeke of Varying Virulence;
Report to Medical Research Council. J. Hyg.
27: 396.

WapswortH, A. B. and Sickues, G. M. 1927. A
Study of Pneumococci Isolated from Horses Un-
dergoing Pneumococeus Immunization. J. Exp.
Med., 45: 787.

Wesster, L. T. 1930. The Role of Microbie
Virulence, Dosage and Host Resistance in Deter-
mining the Spread of Bacterial Infections among

Mice. III. B. enteritidis Infection. J. Ezp..
Med., 52: 931.
Wesster, L. T. and Crow, A. D. 1933. Intra-

nasal Virulence of Pneumococei for Mice. J.
Exp. Med., 58: 465.

33