GENETICS OF SALMONELLA
os Be See

s; School of Medicine, Son oor,
alo. ‘Alto, California, U.S.A.

 

 
GENETICS OF SALMONELLA

BY

TETSUO IINO*

Department of Microbial Genetics, National Institute of Genetics,
Misima, Shizuoka-ken, Japan

AND

JOSHUA LEDERBERG**

Department of Genetics, School of Medicine, Stanford University,
Palo Alto, California, U.S.A.

(with 2 figs.)

I. OUTLOOK

The genetics of Salmonella can be traced back to the discoveries
and descriptions of several remarkable phenomena of antigenic
variation in this genus, namely O-H variation, S-R variation,
form variation and phase variation (reviewed by KAUFFMANN,
1954).

The discovery of phage-mediated transduction in 1952 (ZINDER &
LEDERBERG, 1952) introduced the possibility of the detailed genetic
analysis of these intriguing phenomena. Antigenicities and anti-
genic variations of Salmonella were re-examined with transductional
technique, and detailed schemes on the genetic determination of
antigenicity (Section IV & V) and on the cellular regulatory system
of gene expression (Section VII) have been constructed. Further,
an evolutionary pattern of Salmonella serotypes has been postulated,
based on their genotypic constitutions (Section VI).

Transductional analysis was not confined only to the antigenic
characters, but extended also to many other bacterial characters.
In relation to the genetics of flagellar antigen, the genetic study of
motility has been undertaken with paralyzed mutants (STOCKER
et al., 1953; ENomoTo, 1962). Various nutritional mutants including
auxotrophs for amino acids or nucleic acid bases, and mutants
deficient in sugar fermentations were isolated by the application
of the selection techniques established on Escherichia coli (LEDER-

* Whose studies have been supported by a research grant from the National
Institute of Allergy and Infectious Diseases, (E-2872), U.S. Public Health Service.

** Whose studies have been supported in part by National Institutes of Health
Training Grant 26-295 and Research Grant C-4496, and by the National Science
Foundation.
112

BERG, 1950). Detailed fine structure analyses have been carried out
with these mutants (Section III). The mutants well analyzed for
their mutational sites have been used for the investigation on the
mutagenesis of base analogues (RUDNER, 196la, b; Marcon &
Muxal, 1961). A mutator gene which influences the frequency of
various auxotrophic mutations has been discovered in a strain of
S. typhi-murium (MiyaKE, 1960).

The second epoch of Salmonella genetics came with the success in
transfer of the fertility factor from E. coli to Salmonella and the
hybrid formation between them (Section III). It has led to the
establishment of the sexual recombination system in Salmonella
and has been used for the construction of the chromosome map of
Salmoneila, It also opened a way to investigate phylogenetic rela-
tions of Salmonella with other genera in Enterobacteriae.

The studies on the episomal multi-drug resistant factor R, which
confers on a bacterial cell simultaneous resistance to sulfanilamide,
streptomycin, chloramphenicol, and tetracycline, are under active
investigation, with Salmonella as well as other Enterobacteria.
(Section II). Among the chromosomal genes which control drug
resistance, a streptomycin-resistant mutant showing pleiotropism
has been studied (WATANABE, 1960). This requires thiamine and
nicotinic acid, as well as being streptomycin indifferent. Transduction-
al analysis indicated that all of these changes are caused by one
step recessive mutation of a single gene. The pleiotropism of drug-
resistant mutants was observed also on a chloramphenicol resistant
mutant (WEINER & Swanson, 1960). The mutant differs from the
sensitive parent in morphological and cultural characteristics and in
somatic antigens.

Together with the problem of drug-resistance, the problems of
pathogenicity and host specificity are important subjects that Sai-
monella genetics may contribute to epidemiology as well as basic
bacteriology. It has been known that nutritional factors influence
the virulence of the pathogen (Gowen, 1951) and that certain
auxotrophic mutations of S. typhi-murium accompany the loss of
virulence (BARON, 1953). The abundant accumulations of auxo-
trophic mutants in Satmonella provide a rich source for systematic
study in this direction. The well analyzed antigen types in this
group also favor the genetic investigation of the host-parasitic
relationships. The genetic approach has already been started
(FurNEss & Row ey, 1956). The development of this field in future
is expected.

It is not the purpose of this article to review the entire field of
Salmonella genetics comprehensively. Rather, the subjects will be
focused on a few topics, especially on the immunogenetic aspects,
which are specific to Salmonella and have not been systematically
reviewed elsewhere.
113
Il. GENETIC TRANSFER SYSTEM

Transduction

Salmonella is the prototype of phage-mediated transduction
(ZINDER & LEDERBERG, 1952). At first, the combination of the
competent bacteria was restricted to those in group A, B, and D,
because phage PLT 22, a type of Al-group phage, can attack only
those bacteria which have somatic antigen 72. Later, transduction
was demonstrated with other Salmonella phages having host ranges
different from Al-group, and the range of the competent group is
being extended (Baron et al., 1953; Sakai & Isext, 1954: EDWARDS
et al., 1955). Several techniques have been invented to use virulent
phages for transduction: namely the use of lysogenic recipient
(ZINDER, 1955), lowering of multiplicity of infection (STARLINGER,
1958) or pretreatment of the lysate with ultraviolet light (GoLp-
SCHMIDT & LANDMAN, 1962). The last method is based on the in-
formation that the plaque-forming ability of a lysate is more sensi-
tive to ultraviolet light than its transducing ability. These methods
have been successfully extended to the heterogenetic transduction
of nutritional markers between Salmonella and E. coli with the
virulent phage “chi’’ which can infect certain motile strains of
both genera (Tsuji & I1No, unpublished data).

The transductions in Salmonella so far studied are of the genera-
lized type. Each gene of the donor bacteria has approximately an
equal chance, about one to ten per 10° phage particles, to be incor-
porated into transducing phage and transferred to the recipient
bacteria. The size of genetic fragment transduced by a phage particle
at one time corresponds to a small fraction of the bacterial chromo-
some; chemically it is around 104 nucleotide pairs, less than one
hundredth of the whole chromosome length. Therefore, only very
closely linked genes are transduced simultaneously. The trans-
duced fragments may have predeterminate broken ends (OzExt,
1959).

The incorporation of a donor chromosome fragment in a recipient
cell results in the formation of a hybrid cell which is heterozygous
for a small fraction of the chromosome. Such cell is termed “‘hetero-
genote’’, and the donor chromosome fragment is called “exogenote”’,
in contrast to ‘‘endogenote’’, which is a part of the intact recipient
chromosome homologous to a given exogenote (MorsE et al., 1956).
Recombination between exogenote and endogenote occurs in a
fraction of heterogenotes, and consequently stable transductional
recombinants are produced from time to time. The rest of the trans-
ductional clones remains as persistent heterogenotes. In generalized
transduction as seen in Salmonella, the exogenotes of the trans-
ductional cells cannot multiply and the heterogenotic state is main-
tained only in a single cell in each cell generation of a transductional
114

clone until the exogenote is either lost or integrated into recipient
chromosomes. The genetic characters expressed by such hetero-
genotes are inherited linearly through cell generations. The linear
inheritance of motility appears as a trail of compact colonies on a
semi-solid medium (LEDERBERG, 1956; SrocKER, 1956a) and that
of nutritional markers as a minute colony on a minimal medium
(OzEK1, 1956). The phenomenon of linear inheritance in trans-
ductional heterogenote has been used as a main criterion of ‘‘do-
minance’ and “gene complementation’’. For example, the pro-
duction of trails on transduction from a flagellated strain to a non-
flagellated one indicated that the flagellation is dominant to non-
flagellation (STOCKER et al., 1953). The production of trails in com-
bination between two non-flagellated mutants indicates that the
flagellation alleles of these two mutants are complementary to each
other and presumably belong to different genetic functional units.

There is another related phenomenon of the hereditary change
caused by infection with bacteriophage: lysogenic conversion. In
tvansduction a phage particle carries a genetic fragment of the host
bacteria. Contrasting with it, in conversion, phage genes function as
part of the bacterial organism. Such modification occurs soon after
phage infection and lasts until lysis occurs. When a bacterium is
lysogenized by a conversion phage, the modified character is in-
herited through cell generations to the offspring as long as the lyso-
genic state persists. The conversion of somatic antigen type in

Salmonella is one of the most intensively studied examples (Section
TV).

Conjugation

As well as phage-mediated genetic transfer system, sexual mating
system has been recently established in Salmonella for the genetic
recombination analysis. The sex factor F is transferred from E. coli
to various Salmonella strains (BARON et al., 1959a; ZINDER, 1960;
MryakE, 1962) and from the latter to other Salmonella (MAKELA
et al., 1962). Both the ability to be infected by F and the degree
of fertility of male (F+-) Salmonella differ between different combi-
nations of the strains. Generally homologous strains show the
highest efficiency. The conjugation was also demonstrated between
E, coh Hf and certain strains of Salmonella (MiyAKE & DEMEREC,
1959; ZINDER, 1960; FaLtxow et al., 1962). They yield hybrids
between £. colt and Salmonella.

There are three other systems of conjugation in Salmonella:
namely F-duction (or sex-duction), colicinoduction, and R-duction.
They are mediated by different episomal factors: sex factor F,
colicinogenic factor C, and multidrug resistant factor R respectively.
These systems are common with the phage-mediated transduction
to the point that an episome is a vector of the transmission of a piece
115

of genetic fragment, and also akin to the sexual conjugation in that
both require cell contact for the transfer.

In “F-duction’’, a chromosome fragment is attached to /, re-
plicates synchronously with /, and can be transferred to competent
cells in high frequency, one per 10? to 104 cells in 1: 1 mixture of
donor and recipient cells in homologous combinations. Such an
autonomously replicating complex of F anda chromosome fragment,
designated F-prime (F'), was first obtained from £. colt and intro-
duced to Salmonella (BaRon et al., 1959a; ORsKov et al., 1961). Later,
an F’ which is combined with lactose fermentation gene (Jac) was
demonstrated in a lactose positive Salmonella isolated from nature
(Baron et al., 1959b). The F’ introduced to Salmonella can be
retransferred to the more distantly related bacteria such as Shigella
(Baron ef al., 1959b), Vzbrto (Baron & FatKxow, 1961), Serratia
(FaLKow ef a/., 1961) or Klebstella (MAKELA ef al., 1962) at lower
frequencies, as well as to other Salmonella or Escherichia. The genes
introduced to other genera by F-duction are usually not integrated
into the chromosome of recipient cell; the cell multiplies as persis-
tent heterogenote.

In “‘cin-duction’’, a bacterial chromosome fragment is transferred
along with the transmission of certain colicinogenic factors (OZEKI &
HowartH, 1961; SMITH & STOCKER, 1962). The sizes of chromosome
fragments transferred by F-duction or cin-duction vary in different
systems, but they are usually far longer than those in phage-
mediated transduction.

Like F’, “multi-drug resistant factor R,” originally detected
from Shigella, has a diverse range of transferability among Salmo-
nella, Escherichia, Shigella and even Serratia (reviewed by WarTa-
NABE, 1963), R-factor interferes with the genetic transfer system
by Ff, however, the joint transfer of a chromosomal fragment with
it has not been demonstrated.

Owing to the establishment of the two major recombination sys-
tems, i.e. transduction and conjugation, it is now possible to study
both micro- and macro-topology of the Salmonella chromosome.
The former is applied for the fine structure analysis of a gene or
genes closely linked to each other, but it is not adequate for the
mapping of the various genes distributed through the entire chro-
mosome. Sexual conjugation and cin-duction cover the gaps of
transduction analysis, and they have started to be used for the
mapping of Salmonella chromosome (Section III). Because of its
wide range and high efficiency of transfer of a particular genetic
fragment, F-duction promises to be a unique system for genetic
and chemical investigations of the localized chromosomal regions.
It will supplement the absence of specialized transduction in Sal-
monella.

Several consequences of possible DNA-transformation in Sal-
116

monella have been reported by some investigators, but they are still
ambiguous.

The detail on these genetic transfer processes, as well as the
nature of the mediatory factors, will not be discussed here, as they
were covered in a recently published book by Jacosp & Woo_MAN
(1961) and in two review articles by Lurta (1962) and by CLARK
et al. (1962).

III. CHROMOSOME MAP

Microtopology

In Salmonella, recombination analysis has been focused mainly
on the fine structure analysis in a gene or a short segment of its
chromosome by means of transduction. Mapping of the various
genes on a chromosome had been left behind until recently, as the
conjugation system was established on the organism.

The genetic fine structure analysis on Salmonella raised several
interesting problems on the gene structure in relation to its function.
For example, (1) clustering of the genes with phenotypically similar
effect, (2) coincidence of the order of the genes in a gene cluster with
their sequence in respect to biochemical blocks in the chain of
reaction leading to the synthesis of a certain final product, (3) hetero-
geneity among mutant sites in a genetic functional unit in respect
to mutation rate, phenotypic expression or competence to suppres-
sor. We will not discuss these problems here; they have been re-
viewed in detail by DeMEREC & Hartman (1959), and since then
work with Salmonella has mainly elaborated on the analysis of
individual genes. It may be worth noting, however, that the con-
cept of “operon” as a unit of gene expression is being extended for
the understanding of the phenomenon of gene clustering in E. coli
and several other organisms, and it is interesting to examine to
what extent this concept can be applied to the gene clusters found
in Salmonella (Section VII).

Macrotopology

Mapping of the various genes on Salmonella typhi-murium strains
LT2 and LT7 are summarized in Fig. 1 (DEMEREC & HARTMAN,
1959; FaLKow et al., 1962; Mivakg, 1962; SmitH & STOCKER, 1962,
personal comm.; MAKELA, personal comm.; I1no, unpublished data).
It is represented by a ring as revealed in E. colt K-12. The sequence
and map distance of each marker on the chromosome is roughly
the same as in £. colt (c.f. chromosome map of E. coli illustrated by
FaLkow et al., 1962). Together with the successful hybridization
between these two genera, the parallelism of their linkage maps
may reflect their phylogenetic intimacy. On detailed structure,
however, their chromosomes might differ in many respects, as
117

already suggested by preferential transduction of Salmonella mar-
kers from E. colt-Salmonella hybrids by Salmonella phage (ZINDER,
1960).

araA araB
e  JeuA
pan

 
 
 
 

  

 

met serB

( proA
thr(D-C -A-B)

  
  
  

  

F (try D-C-B-

str A-cysB)

fla(A-B-C)
* his(E-F-A-H-B-C-D-G)

adeC

Fig. 1, Linkage map of Salmonella typhi-muritum.
met: methionine, cys: cysteine, ade: adenine, pro: proline, pan: pantothenic acid,
leu: leucine, ser: serine, thr: threonine, gal: galactose, ara: arabinose, str: strepto-
mycin, O5: O-antigen 5, H,: phase-1 flagellin, H,: phase-2 flagellin, fla: flagellation,
mot: motility.
A capital letter after a gene symbol represents a cistron,

Information obtained from molecular biology endorses these
situations: the overall base composition of chromosome DNA is
similar between Escherichia and Salmonella, but DNA molecules
from each genus are imperfectly complementary as tested by their
capacity for annealing with one another (FaLKow et al., 1962).
Further comparative studies on the chromosome structures of
various Salmonella species and the related bacteria from both
genetic and molecular standpoints will throw light on the evolutio-
nary history of Salmonella.

A distinguished difference between Satmonella and Escherichia is
that the former cannot ferment lactose. In FE. colt, a gene lac, which
118

is an operon controlling lactose fermentation (Section VII; PARDEE
et al., 1959), is present near the pro gene. These two loci are three
units apart. By introduction of a chromosome fragment of E. coli
of this region to Salmonella by F-duction, the recipient Salmonella
cell acquires the ability to ferment lactose. In E. coli Hfr-Salmonella
conjugation, stable hybrids carrying Jac are obtained by transfer
of pro-lac region from the former. The recombination between pro
and lac genes is never detected among the hybrids, and the recom-
bination value between Jac and such markers as asa or gla, which
are more distal than pro from Jac, is also very low compared with
homospecific conjugation in E. coli (MryaKeE, 1962). These results
indicate that the incapability of lactose fermentation in Salmonella
is caused by deletion of Jac region on its chromosome. As well as
the Zac-locus, the P-locus, which controls the production and anti-
genic specificity of fimbriae, is present in FE. coli K-12 strain but
absent in a certain strain of S. typhi (BRINTON & Baron, 1960).
Conversely, Salmonella has a duplication of H-antigen type deter-
minants, only one of which is present in Escherichia. This will be
discussed in detail later.

IV. GENETICS OF SOMATIC ANTIGENS

O-antigen is a heat stable antigen of microcapsule on the surface
of bacterial cell wall. Its specificity is determined by sugar terminals
of lipo-polysaccharide chains. Chemistry of the antigens is described
by STavs et al. (1964) in this book. The variation of O-antigen type
has been known as ‘‘Form variation” (KAUFFMANN, 1940, 1954).

Phage conversion

The genetic determination of certain O-antigen types is an ex-
cellent example of conversion by bacteriophage. Since the conver-
sion of 3, 10 type Salmonella (E,-group) to 3, 75 type (Ey-group) by
infection of phage < was demonstrated (IseK1 & SaxKat, 1953), several
antigen types have been shown to be converted by infection of
certain bacteriophages (Table I). A clear-cut scheme on the corre-
lations between antigen type, its chemical structure and infecting
phage was constructed on the antigenic conversions among E-
group Salmonella (Fig. 2, Ucuipa et al., 1963). E,-group (3, 70)
Salmonella converts to E, type (3, 15) by infection of phage ¢!°,
The new antigen type is detected in a few minutes (UETAKE et al.,
1958). Naturally occurring E,-group Salmonella are lysogenic for
e'° or the phages relating to it. It is possible to isolate E, (3, 70)
type Salmonella from E, type through anti-15 serum selection. The
E, type cells thus obtained had lost the ¢ type phage and be
sensitive to it. The antigenic change caused by the infection of
119

Table I.

Antigenic conversion caused by phage infection in Salmonella.

 

 

conversion phage host bacteria ;changes in O-antigen) reference
: |
e, ¢15 E, group ‘3,10 3,15 | Isexr & Saxar (1953)
e384 E, group ‘3,15 >(3,15)34 UETAKE & HAGIWARA
(1960)
e153 + g34 E, group 3,10 +(3,15)34 Uetake & HaGIwaRa
| (1960)
t, P22 4, 12 type | IsEKI & KASHIWAGI
Al-grop phage group B '4,12-+1,4,12 (1957)
Salmonella |
|
phage 27 '1,4,12 type | LeMrnor et al. (1961)
| group B | 1,4,12 +1,4,12,27

 

: Salmonella

 

2° to E, Salmonella is expressed biochemically as the replacement
of the terminal sugar acetyl-galactose by galactose. The change
of E, (3, 15) to E, (3, 15, 34) by infection of ¢#4 is the addition of a
sugar, glucose, to the terminal. The addition of glucose occurs when
the terminal sugar is galactose but not acetyl-galactose. Conse-
quently, E, (3, 1/0) Salmonella does not produce antigen-34 after
infection of <4, but E, cells lysogenic for <4 produce 34 antigen
after double lysogenization with e!®. The production of antigen-1
by Al-group phage in A, B, D, group Salmonella is also an additive
reaction of glucose to a terminal of the antigenic chain by 1:6
glucosidic linkage (STOCKER et al., 1961). Thus, several O-antigen
factors were found to be determined each by a specific bacterio-
phage genome. It may control directly or indirectly the production
of an enzyme which catalyzes, modifies or inhibits the addition of
a specific sugar to the terminal of polysaccharide chains. Phage
mutants which differ in their conversion abilities have been ob-
tained (TERADA et al., 1956; UETAKE & Ucutpa, 1959).

It is interesting to note that an episome, F, also confers to the
carrier cell a new surface substance involved in male fertility,
possibly a surface polysaccharide (SNEATH & LEDERBERG, 1961),
Immunologically, the /+ antigen was demonstrated on F+ and Hfr
clones of E. coli (OrsKkov & ORsKov, 1960) and S. abony (MAKELA
et al., 1962).

The determinant of O-antigen type in bacterial chromosome is
not disclosed yet, except the approximate position of the antigen-5
determinant in S. typhi-muriwm (Fig. 1). We also have no direct
120

information on the mode of gene control of the backbone structure
of O-antigen and type specific polysaccharides which are not deter-
mined by conversion phage.

gal
man
rham
glu
acetyl gal gal
pan pan
rham 15 Ne rham
acetyl-gal
man

ham
15
é
.o4

Fig. 2. Antigenic conversion by phage <15 and e* in E-group Salmonella (construct-
ed from Ucuipa et al., 1963). rham: rhamnose, man: mannose, gal: galactose,
acetyl-gal: acetyl galactose.

S-R Variation

The variant cells which had lost O-antigens manifest another
type of antigen called R-antigen. R refers to the roughness of the
colonies in contrast with the smoothness (S) of those colonies which
have O-antigens. Their type specificity has not been worked out as
extensively as the O and H antigens. The changes from S to R type,
called S-R variation, were first noted by ARKWRIGHT (1921). Since
then the phenomenon has been reported by many investigators.
S and R types are not uniquely distinct types, for there are many
intermediate types. The frequency of S-R variation has been stu-
121

died by several workers (ZELLE, 1942; PacE et al., 1951). ZELLE
found an extraordinarily high value in one unstable strain of S.
typhi-murium, amounting to one R mutation per 14 cell divisions
in an $ culture.

A suggestive result concerning the genetic basis of S-R variation
and the gene control of polysaccharide structure on cell surface was
obtained with a galactose sensitive mutant of S. ty phi-murium
(FUKAsAWA & NIKaIDo, 1961a). The mutant, called M-mutant, has
a single defect in the uridine-diphospho (UDP)-galactose-epimerase
gene and was not only unable to ferment galactose but also showed
galactose-induced bacteriolysis. On ordinary media, devoid of
galactose, they form rough colonies, and the sugar composition of
their cell walls is only glucose; while the wild type is smooth and
contains galactose, mannose, rhamnose and 3,6-didesoxyhexose in
addition to glucose. The mutant also acquired the resistance to phage
PLT22. These pleiotropic effects of the epimerase mutant are ex-
plained as follows. The polysaccharide structure at the surface of the
wild type cell is composed of three layers. The inner core consists
of polyglucose. Galactose links to the glucose and forms the second
core. The sugars of the outer layer, which is responsible for O-anti-
genicity, connect to these galactose molecules of the second core.
The accumulated chemical evidence (STAUB et al., 1964) has shown
that nucleotide-phosphate-sugars, for example uridine-diphospho-
(UDP)-galactose for galactose, are generally used as the donors of
the sugar component in polysaccharide synthesis. The defect of
UDP-galactose epimerase in the mutant blocks the conversion of
UDP-glucose to UDP-galactose. Then, the reaction of the poly-
saccharide synthesis must stop at the step of the second core
formation. The loss of outer layer results in the simultaneous loss
of O-antigenicity, including antigen-12, which is the receptor of
phage PLT 22, and the change of colonial type occurs from smooth
to rough. On the mechanism of bacteriolysis in galactose-containing
media, several possibilities were discussed in relation to the enorm-
ous accumulation of UDP-galactose (Fukasawa & NIKAIDO, 1961a).
The most plausible one is that the compound interferes with the
incorporation of other nucleotide-phosphate-sugar compounds into
the cell surface structure. These results and the accumulated in-
formation on the chemical changes which are accompanied with
the S-R variation of various bacteria (reviewed by BEALE & WILKER-
son, 1961) together with the antigen conversion by phage, lead to the
speculation that the biosynthesis of polysaccharide in the cell surface
consists of stepwise reactions: a series of genes function by bringing
about an orderly conversion of the precursor, by the addition of a
new terminal sugar together with adjacent groups.
122

V. GENETICS OF H-ANTIGENS

Flagella and flagellin as H-antigen

Flagella are long filamentous appendages, having the length” of
several times that of the bacterial cell. They originate in the cyto-
plasm and project through the cell membrane. The number and
distribution of flagella differ in different bacterial strains and cul-
tural conditions. In Salmonella, like other flagellated enterobacteria,
generally 5 to 10 flagella are observed around a cell. Their diameter,
measured on electron-micrographs, is 10 to 15 my. As can be seen
on dark-field microscope under very powerful illumination, the shape
of each flagellum is helical, with a pitch and diameter characteristic
of the strain. Flagella can be isolated easily by mechanical vibration
of the cell suspension, and purified by high-speed centrifugation
and chromatography (WEIBULL, 1950; KopBayasur et al., 1959:
Enomoto & Itno, 1962). A purified flagellum is a fibrous polymer
of a unit protein. X-ray diffraction analysis indicated that it belongs
to the keratin-myosine-epidermin-fibrinogen type. In acid solution
below pH 3.5, a flagellum dissociates into monomers called ‘‘fla-
gellin”’. Molecular weight of a Salmonella flagellin is about 38,000.
The particle size is 45 A in diameter. From the electron microscopy,
a three stranded complex helical structure has been proposed for
the arrangement of flagellin in a flagellum (KERRIDGE et al., 1962).
Amino acid composition of flagellin has been studied with several
Salmonella serotypes (AMBLER & REES, 1959; AMBLER & STOCKER,
personal comm.). They contain only 15 to 17 kinds of amino acids,
373 in number (Table II), Their N-terminal is lysine. In certain
Salmonella flagellin, an unusual amino acid ‘N-methyl lysine’”’ is
detected (AMBLER & REEs, 1959). The details and references on the
studies of general structure and function of flagella are found in a
review by STOCKER (1956b).

There is abundant evidence that flagella are responsible for the
H-antigenicity among the cell surface structures. The non-flagellated
mutants or natural isolations, called O-type, are without exception
H-agglutination negative. The reversion of these O-type strain to
flagellated type (H-type) always restores the H-agglutinability. The
parallelism between flagellation and H-agglutination is observed
when flagellated cells lose flagella by growing in a medium contain-
ing certain concentration of phenol or at the higher temperature.
Although the specificity of H-antigen is usually examined by H-
agglutination of the flagellated bacterial cells for diagnostic pur-
poses, H-antigenicity is demonstrated with flagella which have been
isolated from the cells or with purified flagellins. For the latter,
precipitation test, such as gel-diffusion technique, has been success-
fully applied (KERRIDGE et al., 1962). Both isolated flagella and
flagellins are also able to induce corresponding H-antibody by
Table II,

Amino Acid Content of Bacterial Flagellins.
moles amino acid/mole flagellin

 

 

 

 

 

S. typhi-murium .

amino acid Ew 1081, | Protews vulgaris

1.2 typet P
glycine. © 6 6 ee ee ee en 22 14
alanine 2... 1... eee ee ee ee 42 18
valine 2... ee ee 17 11
leucine. 6 6 6 ee ee ee 21 | 14
isoleucine. © 6 6 ee ee ee ee | 14 i 9
glutamic acid. . 2... ee ee ee 27 16
asparticacid . 2 2. ee ee | 44 27
serine 2 6, ee ee ee 17 | 11
threonine 2... 0. eee ee 29 14
tyrosine 6 2
phenylalanine. . 4 4
tryptophan. . 0 0
proline... 4 ; 0
cystine/2 . . 0 1
methionine. . 1 1
lysine... ee 10 8
e-N-methyl-lysine. 2... 6 6 ee ee ee 7 0
arginine cee 1 6
histidine . . 7 0
total. . 373 156
molecular weight . . . 1... ee eee 38,000 17,300

 

 

 

1, AMBLER & RzEEs, 1959 & personal comm.
2. KOBAYASHI et al., 1959.

immunization of rabbit. Their H-antigen type specificity is not
qualitatively distinguishable from that of the whole cell. Though
the possibility exists that the antigenicities of a flagellum and its
component flagellin are quantitatively different (KOBAYASHI et al.,
1959), these results strongly support the idea that a flagellin mole-
cule is the unit of H-antigens of a Salmonella cell. In other words,
H-antigen specificity is a mirror of the tertiary structure of a
flagellin molecule.

H-genes as the structural determinants of flagellin

Before starting the discussion on the genetics of H-antigen deter-
minants, it is convenient to introduce a unique phenomenon of
antigenic variation. A culture of many Salmonella strains manifests
two alternative types of H-antigen. The one is called the “‘specific
phase’’, or “phase-l’’, and another is the “non-specific phase’, or
“phase-2”’, Each Salmonella serotype has its own specific type of
phase-1 and phase-2 H-antigens manifested by flagellar protein.
124

When a mass culture of such strain is plated it dissociates into the
colonies of phase-1 and phase-2. During successive cultures of the
two types, the population of each type give rise to cells of an alter-
native phase. The phenomenon was first studied by ANDREws (1922)
and has been known as antigenic “phase variation’. The frequency
of phase variation differ in different strains ranging from 10-5 to
10-3 per bacterial division (SrocKER, 1949). The Salmonella strains
expressing phase-1 and phase-2 alternatively are called diphasic
ones.

H-antigen determinants were disclosed by transductional analysis
(LEDERBERG & Epwarps, 1953). The transduction between differ-
ent diphasic serotypes, for example from S. typhi-murium 7:7.2 to
S. abony b:enx gave the recombinants i:enx and b:7.2 but none
of the 6-1, enx: 7.2 types (an exceptional type will be described in
Section VI), This means that the antigenic specificities of phase-1
and phase-2 are determined by independent loci in each, which are
symbolized by H; and H». They are on the same bacterial chromo.
some but located far enough from each other not to be transduced
simultaneously by a phage particle (Fig. 1). Further transductional
experiments in reverse direction and between different serotypes
proved the generality of this conclusion and led to the establish-
ment of two series of multiple alleles of H; and H», The genotype
of H-antigens of S. typhi-murium is described by Hy? Ho7.2, and
that of S. abony is H,o Henz,

It must be emphasized that the antigen of each phase is a com-
plex of at least several subunits, for example e, x, and x in phase-2 of
S. abony, though they are often listed by a simplified symbol for
descriptive convenience, such as $ in phase-1 of S. abony. Many
subunits, if not all, can mutate independently from the other (I1No,
1959). In genetic recombination, however, they are transferred as a
unit by one of the H-genes.

Hand H; have been defined first as phase-1 and phase-2 antigen
type determinants. Later, certain flagellar shapes were found to be
controlled by the H-genes (I1No, 1962a). The most common mutant
type in flagellar shape of Salmonella is “curly”. Curly flagellum is
characterized by a tighter pitch, half that found in the normal
type (LEIFson & Huu, 1953). In serum agglutination test, H-anti-
gens of normal and its curly mutant flagella behave identically.
The curly character is phase specific and its expression accompanies
phase variation. For example, a curly mutant of S. typhi-murium is
curly in phase-1 while normal in phase-2. In transduction from a
normal flagella strain to such a curly phase-1 strain, transductional
clones with normal flagella were isolated. The transductional clones
thus produced showed the antigen of the donor in phase-1 and that
of the recipient in phase-2. The parallel results were obtained on
curly phase-2 mutants. The replacement of the phase-2 curly deter-
125

minant by normal one was always accompanied by the replacement
of the phase-2 antigen type by that of the donor. From these
results 1t was shown that genetic determinant of curly phase-1 is
at H; and curly phase-2 is at He.

The curly flagella are formed not only by curly mutants but also
by normal cells when flagella are synthesized in a medium contain-
ing para-fluorophenylalanine in place of phenylalanine. This fact
was demonstrated by flagellar regeneration experiment (KERRIDGE,
1959). Para-fluorophenylalanine is an amino acid analogue which
is incorporated into proteins, replacing phenylalanine (reviewed by
RicHmMonpD, 1962). When the proteins have certain biological roles,
their function may be altered or lost by the replacement. In the
synthesis of flagella, replacement of phenylalanine by the analogue
in flagellm molecules may cause alteration of their molecular con-
figuration, and consequently the change in the mode of their poly-
merization, resulting in the production of flagella with changed
curvature. The biochemical process of flagellar formation in curly
mutants might be analogous with that described above: the re-
placement of phenylalanine in flagellin by certain other amino acid
might occur by curly mutation.

Just like the curly flagellar character, one type of resistance to a
motility-specific phage has been demonstrated to be phase-specific
(MEYNELL, 1961). Motility-specific phage attacks motile, not non-
motile, bacteria. Flagellins or part of them are presumably the re-
ceptor of this phage. Among Salmonella serotypes, those which have

g... antigens are excluded from this rule. Even if motile they are
resistant to a motility-phage, cht. When g.. . antigen is introduced
into diphasic strains by transduction of #79. . ., the resulting trans-

ductional clones are resistant to cht-phage in phase-1 while sensitive
in phase-2. The antigenic phase specificity of the resistance is also
observed on a motile cfz-resistant mutant of S. typAi-murium. The
mutant is resistant in phase-2, 7.2-type, while sensitive in phase-l,
1-type (SASAKI, 1961). The determinant of the phage resistance and
1.2-type determinant, Hz, are always co-transduced.

The primacies of H; and H» as shown in the determination of the
specificity of flagella strongly support the idea that H; and H>2 are
the primary structural determinants, structure genes, of flagellin
in phase-1 and phase-2 respectively. A mutation in one of these
genes produces an altered configuration of the corresponding flagel-
lin, resulting in a change in antigen type, a modification of the
flagella shape or an alteration of the receptor site to certain bacterio-
phage.

As can be seen the excellent examples in the genetic and chemical
analysis of human hemoglobin (INGRAM, 1957), and certain bacterial
enzymes (YANOFSKY et al., 1961), it is the current view that the
structural gene carries the genetic code of amino acid sequence in
126

the form of base sequence in its DNA chain, and that the amino
acid sequence in polypeptide primarily determines the tertiary con-
figuration of the protein. An alternative mechanism was proposed
on the genetic determination of immunological specificity of insulin
(Berson & YALow, 1961), and enzymatic and immunological speci-
ficities of tyrosinase (Fox & BurNetT, 1962). That is, certain genetic
determinants of the protein specificity may control the type of
folding of polypeptide; thus they may provide the chance of produc-
ing the proteins of different specificity with the same amino acid
sequence under different genetic background (Karusu, 1960). How-
ever, there is no definite evidence so far for the second mechanism,
and the folding of a protein may be completely predetermined by
its amino acid sequence.

Though the amino acid sequence is not known yet, the compara-
tive amino acid analysis and the analysis of tryptic peptides of
Salmonella flagellin (AMBLER & STOCKER, personal comm.; ENomo-
To & lino, unpublished data) strongly suggest that the H-gene
determines the actual amino acid sequence in the polypeptide of
flagellin. The relative amount of each component amino acid
differed among the flagellins of different antigen type. Differences
are noticed even between two antigen types which represent phase-1
and phase-2 of the same strain. The finger prints of tryptic digests
of 7-flagellin and 7.2 flagellin, which represent the phase-1 and phase-
2 antigen of S. typhi-murium, respectively, differ in 15 among 35
spots (STOCKER & AMBLER, personal comm.). The difference of cer-
tain amino acid composition between flagellins of normal and curly
mutant has been suggested by preliminary experiment but is not
conclusive yet. For the clear understanding of the determination
of flagellin structure by the H-genes, we need the detailed fine
structure analysis of the H-genes, especially the localization of the
antigenic subunit determinants in H, and in parallel the analysis
of the amino acid sequence of flagellin. Some hopeful steps have
been taken in this direction. The comparative studies on the sub-
units of g... groups antigens (I1No, 1959) and on the mutants
of 7-antigens (Joys, 1961) have shown that in H; there are sequen-
tial mutational subuniis which correspond to the antigenic sub-
units. The intra-H» recombination was obtained between H 2t.2 and
He" (T1no, 1960, unpublished data),

Genetic modification of H-antigen

Besides the specificity in the relative amount of each amino acid
species, Salmonella flagellin has a remarkable chemical feature in
that some of them contain N-methyl-lysin (NML), an amino acid
that has not been previously found to occur naturally (AMBLER &
REEs, 1959), The amount of NML is about equal with that of lysine.
Among the serotypes tested, S. typhi-murium (1:17.23 ) were NML-
127

positive in both antigenic phases. In certain other serotypes, for
example S. derby (gp: -), both NML-positive and -negative strains
were found.

The analysis with transduction has shown that a gene which
determines the presence or absence of NML is linked to H; (STOCKER
et al., 1961). Unlike H; and Ho», this one gene determines the pre-
sence or absence of NML in both phase-1 and phase-2 flagellin. The
immunological study of the recombinants between NML+ and
NML- strain disclosed an interesting correlation between H-antigen
type and the amino acid: when 7.2 type NML-negative cells are
changed to NML-positive by introduction of NML*-gene, antigen
3 as well as 7 and 2 appear, and vice versa (STOCKER, personal
comm.). The correlation between NML-positive and antigen-3
positive is observed, only when the antigen type of the flagella is
1.2.(3). The H-antigen of S. abony (b-e.n.x.) is not altered by the
presence or absence of NML.

In the foregoing chapter we came to the conclusion that H,; and
Hz genes determine respectively the complete amino acid sequences
of the phase-1 and phase-2 flagellins. Consequently, the NML-gene,
which is phase non-specific and at a different locus from Hy, and
H 2, is thought to modify the product of the Hz gene, by engendering
an enzyme which methylates some lysine radicals of already formed
flagellin. The methylation of lysine, either in the form of free lysine
or S-RNA bound lysine before its incorporation into flagellin is less
likely, under our present understanding of protein synthesis.

Genetics of antigenic phases

In diphasic strains, two antigenic phases are determined at the
separate loci. Therefore phase variation cannot be construed as the
mutation of an antigen type determinant from one specific allele
to another. Instead, it must be the alternative manifestation of each
of the two antigenic specificities already inherent in the genotype.
The transduction between single phase cultures of diphasic strains
in all possible combinations showed that H; can be expressed only
when transduced into phase-1 cells regardless of the phase of the
donor, whereas H» can be expressed in any phases of the recipient
but only when donor is in phase-2. This result, together with other
supporting data (LEDERBERG & IINO, 1956), indicates that the H2
gene plays a decisive role in the expression of the antigenic phases.
The process of phase variation is thus explained as follows: “Hs
takes two different states, active and inactive. Active He is epi-
static to H; and inhibits the production of the phase-1 antigen,
while it carries out the production of phase-2 antigen. When Ho»
changes to the inactive state, corresponding to the change from
phase-2 to phase-1, the production of phase-2 antigen stops and
alternatively the production of phase-1, specified by Hj, proceeds.”
128

Occasionally serotypic variants which express three or four anti-
genic phases have been isolated from nature (EDWARDS et al., 1962).
As far as has been analyzed genetically, such tri- or tetra-phasic
strains are duplications of H; and/or Hz (LEDERBERG, 1961). They
might have been produced by either translocation or by heterozy-
gosis. The persistent heterozygotes of H; and H» are obtained
either by transduction (Spicer and Datta, 1959) or by sexual re-
combination (Hirokawa & Irno, 1961). In every case only one
antigen type among three or four is expressed at a time.

As well as Salmonella having more than two antigenic phases,
there are Salmonella serotypes and certain mutants of diphasic
strains which express only one antigenic phase, either phase-1 or
phase-2. They are called monophasic-1 or monophasic-2 type
respectively. The serotypes carrying antigen g and the related
antigens, such as S. enteritidis gm:-, are almost always monophasic-1
types, and H2 could not be transferred from any diphasic strains
to them, though the diphasic combinations of these factors are
readily synthesized by transduction of the H; to any other diphasic
strain. The genotype of monophasic-1 (g) type is represented by
H,9 H»-deficient, for no homologous locus with H» of other strains
can be detected by genetic recombination. S. typhi d: -, and some
monophasic mutants of S. paratyphi B b: -, appear to belong to the
same category.

Another group of monophasic-1 types, which have been isolated
as variants of some diphasic serotypes, especially of S. typhi-murium
and S. paratypi B, is the H» inactive type. They can become di-
phasic type by reversion or by transduction of H» from a diphasic
donor strain. In some mutants, the phase-2 antigen type recovered
by the transduction is consistently the same as that of the donor,
indicating that the inactivation of H» is caused by mutation of
a site not separable from the antigen type determinants: while in
other mutants, diphasic transductional types with concealed phase-2
antigen of the recipient appear as well as the donor phase-2 type.
Linkage analysis of such mutants disclosed a factor termed ah,
which is adjoined to the phase-2 antigen type determinant, Ho,
and controls the activity of H» (I1no, 1958a, 1962b): that is, the
mutation of ahz+ to an allele ah2- causes the inactivation of H 2
and consequently the change from diphasic type to monophasic-1
type.

In contrast to monophasic-1 types, H, deficient type has not
been detected among monophasic-2 strains either from nature or
as laboratory mutants. All the monophasic-2 mutants so far examin-
ed were found to be the mutants in a region, termed ahy, adjoining
to H; (Itno, 1961a). The function of ah; parallels that of ake in
relation to H2. The mutation of af;- causes the inactivation of Ay,
but it does not concern the activity of H». In such mutants Ho
129

changes its activity as in diphasic strains. Consequently, in phase-1
both Hy and Hy» are inactive and the production of flagella is
entirely stopped, while in phase-2 normal flagella are produced: the
mutants undergo oscillatory changes between flagellated phase-2
and non-flagellated phase-1.

The last group of monophasic types include those which actually
carry both H; and Hy» genes but have a greatly reduced rate of
variation. A representative of this group is S. abortus-egui (Ep-
warps & Bruner, 1939). S. abortus-equi isolated from nature is
generally in phase-2, enx-type. The phase-2 culture of this serotype
is very stable and the alternative phase, a-type, is occasionally
isolated after anti-ewx serum selection. The phase-1 clone thus
obtained is also stable: reversion to phase-2 can be detected only
with strenuous selection with anti-a serum. Though there is no
detailed quantitative measurement of phase variation on these
serotypes, it is estimated to be less than 107 per bacterial division:
this value is less than one thousandth of the frequency of phase
variation in ordinary diphasic strains. Thus S. abortus-egui carries
both Hz and H» and undergoes phase variation, but the frequency
is so low that a culture behave as if it is either (usually) a mono-
phasic-1 or (rarely) monophasic-2 type. The transductions between
such stable phase strain and a diphasic strain demonstrated that a
major factor which regulates the stability of antigenic phases is
transduced simultaneously with H»2 at the frequency of 30% (IrNo,
1961b). The factor is designated vA. An allele vhe- is in S. abortus-
equt and stabilizes H2 in its existing state, whether inactive or active,
and produces monophasic-1 or monophasic-2 types respectively. The
ah,~ monophasic-2 and the vhg~ monophasic-2 type are identified by
plating the culture on semi-solid media; the former dissociates non-
flagellated phase-1 cells while the latter is stable in flagellated phase-
2. The distinction between ah2- monophasic-1 type and vh2- mono-
phasic-1 type is possible only after the stability of the phase of their
phase-2 derivatives is examined: the phase-2 culture originated
from the former by reversion of ah;~ to ah;+ is diphasic, whereas
those from the latter is monophasic-2. The genetic change from
vh2- to vhet or the opposite direction has not been reported since.
A similar type of monophasic behavior as that of S. abortus-equi
has been reported in S. paratyphi A (BRUNER & Epwarps, 1941];
Epwarps et al., 1950). The monophasic property of this serotype
may be caused by stabilization of H2 as in S. abortus-equi. Whether
the genetic factor which stabilizes the antigenic phase in S. para-
typhi A is identical to that of S. abortus-equi or not awaits further
genetic analysis.

An extreme type of mutation in flagellar antigenic phase is that
from flagellated to non-flagellated (O) type. The change has been
called O-H variation. The mapping of flagellation genes has been
130

carried out in parallel with transduction (Imno, 1958b and un-
published data) and with cin-duction (SmirH & SrocKER, 1962
and unpublished data). The results are consistent with each other,
and two chromosomal regions responsible for the flagellation both
in phase-1 and phase-2 are assigned. The one is adjacent to H; and
composed from at least three cistrons, fla-A, fla-B and fla-C. The
other is located between try and gal loci, and the mutants in this
region so far examined belong to one cistron. Fla~ mutation of any
one of these genes causes the loss of flagella.

Among the Salmonella serotypes, S. gallinarum and S. pullorum
in group-D do not produce flagella and are H-negative. The spon-
taneous mutation to H-type of these serotypes has not been observ-
ed. Though the detailed genetic analysis of these serotypes has not
yet been carried out, the preliminary transduction experiment
suggests that they carry H19™ gene, but non-flagellated because of
the deletion or multi-site mutation of a Fla region (LEDERBERG &
Epwarps, 1953).

The genotypes of several representative H-antigen types are
summarized in Table III.

VI. EVOLUTIONARY ASPECTS OF SEROTYPES

As it can be presumed from the earlier discussion on the genetic
determination of O-antigen, the differentiation of the O-antigen
types may proceed by the loss, gain or modification of the ability
to add a specific sugar to the polysaccharide skeleton of the micro-
capsule. This may be attained stepwise by either phage conversion
or mutation of antigen type determinants. Actually, the differences
of O-antigen types between E,, E,, E, and Ey groups are explained
by the presence or absence of symbiotic conversion phage(s) (Fig. 2).
The differences of O-antigen type in group B are also explained by
two conversion phages for antigens 7, and 72, and a mutation in a
particular chromosomal gene for antigen 5. As we have only scarce
information on the chromosomal genes controlling the O-antigen
production, it is not certain whether the differences in O-antigens
between different groups are solely explained by accumulation of
stepwise mutations or not.

The H-antigen situation is somewhat more complicated. If you
look through the Kauffmann-White scheme (KAUFFMANN, 1964)
you may notice that many H antigen types appear only in phase-l
while others in phase-2. However, there are several antigen types,
like e, /, and w, which appear both in phase-1 and phase-2, indicating
that the antigen type is not specific to antigen phase. Indeed, there
are occasions that the same type of antigenic subunit appears in
both phase-1 and phase-2 of the same strain as, in S. salinatis deh,
denz;;. The same antigen Jw is determined by H2 in S. wien b: lw
 

 

 

 

 

 

a ‘yueunuiayep oseyd oruasyue :4yd ‘oyeys oTqRysun tA "JULINUL JPISIQINW Jo UOTZDII[EP :O ‘aday uasique possaidxoun :( )
re
TT
0 0 ti ol o w3 + :(u13) oiseydryynu wmnavUYyvs *S
|
+ 0 | é > OQ 0 uid + 73 ]-o1seydouow “assa *S
| |
| |
+ + — xa | + e +: xua : (2) Z-oiseydouowr anba-snpaogn *S
an
-k — i CT + / ke + (o-[) :e [-oseydououwr F 1yddywoand *S
+ A, +f : xue + | q + xua :q orseydip Au0gv “Ss
+ é + ZL — I + (vit gorseydouour 99TTAAS Jue {NUT
, : wnranm-1y4gaq 'S
+ é + | ZT + 4 I — ST 2()— [-oseydouow ISOLMS Jue PNW
| : wninu-tydhy Ss
++ A + ol + ! 1 +! GL! aseydip mntunu-144dq Ss
| | |
r
e - Bud BUA ery Sue | TY lye : |
- ve fad | fg foe! === addy uastyue aseyd s1uastque | adAjzor19s
adAjouo3 : |

 

 

“sodAjO19S VTTOUOW[S aaTpejUasaidas ay} UT S}JUVUTUTIOJAp UssT}UL-F] SY} Jo sodAjousy

“TH 9192®,L
132

while by Hz in S. day-es-salaam lw: enzys (EDwarps et al., 1955).
In some instances the antigenic phases classified by antigen type
do not agree with the phases defined by genetic analysis. The phase-1
and phase-2 of S. worthington have been registered as Jw and z
respectively; in transduction, /w is found to be determined by Hz
and z by H; (Epwarps et al., 1955). There is an isolation of S.
paratypht B java, which expresses b and 7.2 alternatively but at a
lower frequency than the ordinary phase variation. Not only the 8,
but also the 7.2 factor of this strain was found to be determined by
an allele of H; (LEDERBERG, 1961). The strain also carries a nuli-
allele of Hz, allowing the synthesis of such types as(H7?) H21.2Hyenz,

Regardless of the seeming confusion in the correlation between
antigen type and antigenic phases, there is actually a regular strict
separation of H; and H» homologies in all strains above. Assign-
ments of alleles to Hz versus H» can be made unambiguously by
genetic analysis, and the further behavior of each strain is perfectly
consistent. They still precisely fit the rule of phase variation:
Frequent state change of H» and epistasis of Hz to Hy.

An unique anomaly in the homology of H-genes was, however,
found on one occasion of transduction from S, abony b: enx to a
monophasic-1 type of S. typhi-murium-: 1.2 (lino, 1961c). The
recombinant expresses non-flagellated phase and b-phase, b:-, alter-
nately, instead of b- 7.2 or -: enx. By further transduction or rever-
sion, 7:6 type was obtained from the recombinant. H» of the strain
behaves as an allelic locus of phase-2 antigen type determinant,
H»> rather than the original H;%, For example, the transduction
from the recombinant to @:7.5 type gave recombinants a:6 and
2:7.5. Phase variation in these strains occurs as frequently as the
original recipient strain. Thus the anomalous recombinant is presum-
ed to have originated by an exceptional recombination: H2 locus is
replaced by H; of the donor in the transduction, indicating residual
structural homology between Hz and H». The very rare chance of
the event and the exclusive affinity of the translocated H»? to
other H» genes indicate that the barriers of the synapsis between
1 and H» are not so much their own structural differences but
may also be the differences of the other genes involved in a trans-
duction fragment together with H; or Ho.

The genetic homology between H, and Ho revealed from the
studies of unequal recombination leads us to the speculation on the
phylogenic relations of phase-1 and phase-2. We may infer that one
of the H genes originated by duplication of the other. Then, which
of H, and Hz is the original locus? Like many other questions on
evolution, one may never get the conclusive evidence on it. How-
ever, there are several indications which favor the idea that Ho
arose by duplication of H;. For instance, there are numbers of
monophasic-1 types deficient in H», but no H; deficient types have
133

been detected among monophasic-2 types. Clusters of the genes
which control the sequence of a reaction process have been demon-
strated for many nutritional characters. A similar gene cluster of
flagellar formation is present at the H; region. On the contrary,
the genetic factors known to be closely linked to H» are only the
H» activity controler, az and its stability controller, vi. The
interaction between H2 and vh2 is analogous to the variegated type
position effect observed by translocation of certain genes to hetero-
chromatic region (reviewed by Lewis, 1950). Furthermore, in E. coli
which is monophasic, only one H gene has been found and linked
to hes, to which H,; of Salmonella is also linked (Fig. 1), Though
allelism test of E. coli H and Salmonella H has not been led to a
definite conclusion, it suggests that the H locus of E. colt is allelic to
fT; of Salmonella.

Based on these speculations, we may trace the evolutional path-
way of antigenic phases as follows: The original type is primary
monophasic-1 type having H, but not H». Duplication and trans-
location of H; happened in the monophasic-1 type. The synaptic
homology between H; and duplicated Hz was blocked by the
disparity of the loci closely linked to each of Hz. The translocated
H is now identified as a new locus, H2, different from the original H;.
The H2-gene was unstabilized by an influence of the adjoining
chromosomal region. Then, the structural differentiation might have
occurred between H, and H» and now the strain is recognized as a
diphasic strain. Subsequently, a variety of further types has evolved
through mutations and deletions in the diphasic system (of EDWARDS
& Brun, 1936), affecting the H; and H» genes themselves, their
controllers ah; and ahe, the variation factor vhs and the fla genes,
respectively.

Next we will look into the combination of H-antigen and O-
antigen of a serotype. The various H-antigen types are almost
equally distributed to each group and it is difficult to find any
correlations in the combination of O- and H-antigens. It might be
a possibility that same types of mutations of H (or O) antigen have
occurred in parallel in two independent strains having different
O (or #) antigens. The more important mechanism might be,
however, the formation of the new combination through genetic
recombination. Several facts have been known to support this idea:
Namely wide distribution of a specific genotype such as H 19H 9~--x
in various O-groups, 7m vivo occurrence of phage mediated trans-
duction and conversion on antigenic characters (VELAUDAPILLAI,
1960) and the isolation of Salmonella-Escherichia hybrid (BARON
et al., 1959b).

VII. REGULATION OF GENE ACTION
The mechanism of regulation of gene expression is a highlight of
134

modern genetics which aims to bridge between heredity and differ-
entiation. The progress of biochemical and molecular genetics in
the past decade succeeded in the establishment of the fundamental
scheme of protein synthesis. The primary structure of a protein is
determined by the nucleotide sequence of a chromosome region
called structural gene. Messenger-RNA, which is complementary
to the structural gene, is synthesized copying the later as the
primer. The messenger-RNA (m-RNA) moves from nucleus to
cytoplasm and fixes on a ribosome system. Transfer-RNA (t-RNA),
each molecule of which carries an activated amino acid, comes to
the ribosome and arranges itself along the m-RNA. A t-RNA has
both amino acid specificity and pairing specificity to a coding unit of
m-RNA. The activated amino acids, which arranged together with
t-RNAs on the m-RNA, combine with polypeptide linkage side by
side and leave from the ribosome system. The sequential polypeptide
formation and stripping from the ribosome allow the synthesized
polypeptide to fold in proper order and to form a protein with a
specific tertiary structure. The mechanism of genetic regulation is
being discussed in connection with these biochemical steps of gene
expression (FRISCH, 1961).

A prevailing concept on the regulation of gene function is the
regulator-operator theory formalized by Jacop & Monon (1961).
There is a unit segment of chromosome called operon. An operon is
composed from an operator-gene and one or more functionally
related structural genes linked to it. The operator-gene switches
on and off the expression of the structural gene(s) directly, without
mediation of cytoplasmic factor. Its function is presumably the
initiation of m-RNA synthesis on the template of the following
segment of DNA. The function of the operator-gene is managed by
the product, termed repressor, of another gene, regulator-gene. The
attachment of the repressor to the operator-gene prevents m-RNA
initiation in that region and consequently inactivates the structural
genes in an operon. The regulator gene may or may not be linked to
the operon. In enzyme synthesis, the ‘‘inducer’’ is explained by its
combination with the repressor, to inhibit the latter from attaching
to the operator-gene. The repressor has been presumed to be non-
protein substance; for example RNA. The possibility still remains,
however, that certain kinds of protein might be a repressor (Hort-
ucHI & Novi, 1961), e.g. histone, which combines with chromo-
somal DNA and suppresses the synthesis of chromosomal RNA
(Huanc & Bonner, 1962).

In Salmonella, a cluster of genes controlling the related function
has been found on various auxotrophic mutants and fermentation
mutants (DEMEREC & Hartman, 1959). In two groups among them,
results have been reported which seem to suggest that the cluster
fit to the concept of operon. They are histidine mutants (AMEs et al.,
135

1960) and galactose mutants (FuKasAwa & Nikarpo, 1961b). In
each of these groups, a type of mutants show pleiotropic effect to
block the function of all structure genes in the cluster (LEDERBERG,
E.M., 1961). The mutants are not complemented by any of the
structural gene mutants. Their mutant sites map at one extremity
of the cluster. They were explained as the mutants in the operator
gene.

The genetic system controlling the flagellar formation provides an
interesting model of gene interaction. In the flagellar forming sys-
tem, the genes H; and H2 are assigned as the structural genes. The
expression of the H-genes are regulated by several other genes:
namely ah,, ah, vh, and flas, as well as the interactions between
H, and Ho (section V). The ah-genes are phase specific regulato1s
An allele ah+ switches on the production of flagellin controlled by
adjoining structural gene, ah; for H; and ahz for H» respectively.
A mutant allele, ah,~, or ah2- switches off the reaction in the corres-
ponding phase. AA+ is dominant to ah-, but the function of ah
appears only in cis-position with the adjacent H (I1No, 1962b).
It means that H; (or H2) and ah; (or ahe) behave as two component
parts of a genetic functional unit. In chemical terms, an ah-H
complex may be a unit of transcription of genetic code carried by
a chromosomal DNA. It must determine the production and struc-
ture of a m-RNA. The ah-region in the ak-H complex may be a
terminal segment which codes a part of flagellin polypeptide not
responsible for the antigenic specificity but important for the folding
of flagellin polypeptide; or it may be a region where the synthesis of
the complementary m-RNA starts; or it may correspond to a termi-
nal of m-RNA where polypeptidation of the orderly arranged amino
acid starts. Another possibility is that ah carries the code of a region
of m-RNA which is not responsible for the amino acid sequence
in flagellin but for the specific affinity of the m-RNA to ribosome.
The comparative analysis of the genetic fine structure of ah-H
region and the amino acid sequence in flagellin will give the final
conclusion on these aspects in the future.

You may notice that the ah-H system is in many respects ana-
logous to the operon. The ah; (or az) may correspond to the opera-
tor of H; (or #2), and ah;- and ah2~ to the operator negative mutants
(O°), which is inactive regardless the presence or absence of re-
pressor. Then, is the phase variation explained as the repression
and derepression of the ah2-H» operon? If it is the case, we may
assign ahg* as the repressible operator (O+); in phase-1, O+ is re-
pressed and Hg is inactive, while in phase 2, O+ is unrepressed and
His active. The phase controller in diphasic strains is functionally
indistinguishable from ahz except for the high frequency of state
change in the former. Moreover, the recombination of the phase
controller has not been demonstrated with neither He nor aho.
136

Therefore, it is an acceptable hypothesis that both the phase con-
troller and the ah2* gene are identical. So far we have not detected
any agent which works as exogenous inducer or repressor of phase
variation. Consequently, it is premature to speculate on the nature
of the endogenous factors. We can infer a phase-2 specific product
which accumulates in the phase-2 cell and at a certain level of con-
centration represses the function of ah, perhaps the same agent
that maintains the H» activity. The possibility that either flagella
or flagellin in phase-2 is the repressor is excluded because of the
fact that the state change of H»2 occurs even in the mutant which
has lost the ability to synthesize flagellin (Ivo, unpublished data).
It is also unlikely that vk» is equivalent to the regulator gene which
produces a repressor because in vf2~ cell both phase-1 and phase-2
are stable.

“Formation and dissociation of the ah2-repressor complex’’ is an
attractive hypothesis to explain the state change in phase variation
but we must keep in mind that it is not the sole possible explanation.
It must be recalled that the H2 locus had been originated by trans-
location of H; gene (Section VI). Thus the translocation may result
in a certain structural anomaly at the junction of the chromo-
some and the translocated segment of the H» region. The vhe
factor would then represent such a structural anomaly which
causes the instability of the adjacent region where az is located;
consequently ahz frequently changes its activity independently
of any repressor substance. Both the constancy of the frequency
and the randomness of the occurrence of the phase variation
in different ages of a single phase colony under various cultural
conditions (STOCKER, 1949, IINo, unpublished data) favors this
hypothesis rather than the hypothesis of self-reinforced repressor
production in phase-2, at least for the H, effect.

On the expression of the H-genes, the interaction between the
two homologous genes, H; and H», superimposes on the ah-H
system. In di-, tri or tetra-phase strains of various origin, we have
noticed that only flagella of one phase are produced at a time by a
cell. It is the general rule on all of these strains that H2 is epistatic
to H, regardless of the antigen type. How is the epistasis of a gene,
namely H»2, to another homologous one, H;, established? Several
working hypotheses may be proposed for this question. It might be
the interaction at the chromosomal level; for example, the ordered
transcription of the genom in m-RNAsynthesis favors the prefer-
ential production of phase-2 m-RNA, or either phase-2 m-RNA or
(less likely) phase-2 flagellin acts as a cross-repressor of H, pre-
sumably by complexing at ah;. Still other possibilities remain that
the functional competition occurs at a localized site in cytoplasm,
between phase-1 and phase-2 m-RNAs on ribosomes or between
phase-l and phase-2 flagellins at the polymerization step on a
137

flagella-forming apparatus. For the present, we can not exclude the
possibility of any one of them.

In contrast to the above regulation systems, the regulations by
the fla-genes are non-specific with respect to antigenic phase. The
fla-genes are in functional units distinct from both of H; and He
genes. The dominance of flat to fla~ excludes the possibility that
fia” produces repressor substance of the H-genes by itself. We may
assume three other mechanisms on the function of the fla-genes.
(1) Fla+ supports the polymerisation of flagellin monomers. (2)
Flagellin molecules are synthesized on specialized ‘‘flagellosomes”’
and fla* controls the production of such specific ribosomes. (3) Flat
produces internal inducer of flagellin. The test on the production
of immunologically cross-reacting-material (CRM) of flagellin with
twenty-five fa~ mutants of S. typhi-murium and S. abortus-equi
showed that, with one exception, they do not produce flagellin
CRM at all (Irvo & Enomoro, 1952). Presumably, they belong to
the category (2) or (3). The remaining one, obtained from a strain
of S. abortus-equi, produces a flagellin antigenically indistinguish-
able from the parental flagellin (Ivo & Haruna, 1960); the mu-
tant can produce flagellin molecules but cannot polymerize them
into a flagellum, hypothesis (1). The mutant gene is in a region
different from either any other /la-genes or the H-genes.

As for hypothesis (2), the presence of specialized flagellar-forming
apparatus in the flagellated Salmonella cells has been predicted from
the flagellar regeneration experiment (KERRIDGE, 1960) and the
deflagellation experiment with phenol (I~vo & MrTant, 1962). The
CRM* fa~ mutant may be an excellent material for the studies
on the genetic control of such a flagellar forming apparatus. Hypo-
theses of type (3) are always available when other evidence is not.

The central problem of development in higher organisms is the
quasi-stable switching of gene activity — for example the transition
from fetal to adult hemoglobin synthesis, the differentiation of
isozymes, or the formation of specialized nerve-endings in the brain.
Phase variation in Salmonella is a prototype of endogenous switching,
in a simple organism amenable to genetic analysis. (Mainly lacking
is a direct criterion of the chemical state of DNA from H>»-active
and -inactive states.) It has already justified continued study by
the specialist for its significance in the epidermology and evolution
of Salmonella. Even more important it embodies the contemporary
tradition of bacteriological research: Its unification with general
biochemistry, genetics and cell biology.

 
REFERENCES

AMBLER, R. P. & Rees, M. W., 1959. e-N-Methyl-lysine in bacterial flagella
protein. Nature, Lond. 184, 56—857.

Ames, B.N., Garry, B. & HERZENBERG, L. A., 1960. The genetic control
of the enzymes of histidine biosynthesis in Salmonella typhimurium.
J. gen. Microbiol., 22, 369—378,

Anpbrews, F. W., 1922. Studies in group-agglutination. I. The Salmonella
group and its antigenic structure. /. Path. Bact. 25, 515—521,

Baron, L.5., 1953. Genetic transfer by means of Vi phage lysates. Cold
Spring Harbor Symp. quant. Biol. 18, 271—272.

Baron, L.S., Sprpman, W.M. & Carey, W. F. 1959a. Hybridization of
Salmonella species by mating with E. coli. Science. 130, 566—567.

Baron, L.S., Carry, W. F. & Spitman, W.M. 1959b. Characteristics of a
high frequency of recombination (Hfr) strain of Salmonella typhosa
compatible with Salmonella, Shigella and Escherichia species. Proc.
Nat. Acad. Sci. U.S, 45, 1752—1757.

Baron, L. S. & Fatkxow, S., 1961. Genetic transfer of episomes from Salmo-
nella typhosa to Vibrio cholerae. Rec. Genet. Soc. Amer. 30, 59.

Baron, L.S., Format, S. B. & Spr~man, W., 1953. Uses of Vi phage lysates
in genetic transfer. Proc. Soc. exp. Biol. Med. 83, 292—-295.

Beate, G.H. & Wirxinson, J. F., 1961. Antigenic variation in unicellular
organisms. Ann. Rev. Microbiol. 15, 263—296.

Berson, S. A. & Yatow, R.S., 1961. Immunochemical distinction between
insulins with identical amino-acid sequences. Nature, Lond. 19, 1392—
7393.

Brinton, C. C. & Baron, L. S., 1960. Transfer of piliation from Escherichia
coli to Salmonella typhosa by genetic recombination. Biochim. biophys.
Acta 42, 298—37].

Bruner, D. W. & Epwarps, P. R., 1941. The demonstration of non-specific
components in Salmonella paratyphi A by induced variation. J. Bact.
42, 467—478.

Crark, A. J. & ADELBERG, E. A., 1962. Bacterial Conjugation. Ann. Rev.
Microbiol. 16, 289—379.

DemeEREc, M. & Hartman, P. E., 1959. Complex loci in microorganisms. Aun.
Rev, Microbiol. 13, 377406.

Epwarps, P. R. & Bruner, D. W., 1939. The demonstration of phase varia-
tion in Salmonella abortus-equi. J. Bact. 58, 63-—72.

Epwarps, P. R., Barnes, L. A. & Bascock, M. C., 1950. The natural occur-
rence of phase 2 of Salmonella paratyphi A. J. Bact. 59, 135—136.

Epwarps, P. R., Davis, B. R. & Cuerry, W. B., 1955. Transfer of antigens
by phage lysates with particular reference to the /. w antigens of
Salmonella, J. Bact. 70, 279—284,

Epwarps, P. R., SAKAZAKI, R. & Kato, I., 1962. Natural occurrence of four
reversible flagella phases in cultures of Salmonella Mikawashima. J.
Bact. 84, 99-103.

Enomoto, M., 1962. Grouping of paralyzed mutants in Salmonella, Ann.
Rept. Nat. Inst. Genetics (Japan). 13, 75.

Enomoto, M. & Irno, T., 1962. Purification, chromatography and electro-
phoresis of Salmonella flagellin. Ann. Rept. Nat. Inst. Genetics. (Japan).
12.

FaLkow, S., Marmur, J., Carey, W. F., Sprrman, W.M, & Baron, L.S.,
1961. Episomic transfer between Salmonella typhosa and Serratia
marcescens. Genetics 46, 703—706.

Fackow, S., Rownp, R. & Baron, L. S., 1962. Genetic homology between
Escherichia coli-K12 and Salmonella. J. Bact. 34, 1300—1372.
139

Fox, A. 5. & Burnett, J. B., 1961. Tyrosinases of diverse thermostabilities
and their interconversion in Neurospora crassa. Biochem. biophys.
Acta 61, 108—120.

Friscu, L. (ed.), 1961. Cellular regulatory mechanisms. Cold Spring Harbor
Symp. quant, Biol. 26, pp. 408.

Fukasawa, T. & Nixarpo, H., 196la. Galactose-sensitive mutants of Sal-
monelia II. Bacteriolysis induced by galactose. Biochim. biophys. Acta.
48, 470—483,

Fukasawa, T. & Nikarpo, H., 1961b. Galactose mutants of Salmonella
typhimurium. Genetics. 46, 1295—1303.

Furness, G. & Row.ey, D., 1956. Transduction of virulence within the
species Salm. typhimurium. J. gen. Microbiol. 15, 140-—145,

GoLDsHMIpT, E. P. & Lanpman, O. E., 1962. Transduction by ultraviolet
irradiated virulent bacteriophage. J. Bact. 83, 690—691.

GoweEn, J. W., 1951. Genetics and disease resistance. Genetics in the 20th
Century, (L. C. Dunn ed.). 407-—429.

Hirokawa, H. & Iino, T., 1961. H-antigen of heterozygous hybrids be-
tween Salmonella abony and Salmonella typhimurium. Ann. Rept. Nat.
Inst. Genetics (Japan), 12, 81—82.

Horiucui, T. & Novick, A., 1961. A thermolabile repression system. Cold
Spring Harbor Symp. quant. Biol. 26, 247-248.

Huanc, R.C. & Bonner, J., 1962. Histone, a suppressor of chromosomal
RNA synthesis. Proc. Nat. Acad. Sci., U.S. 48, 7277—1222.

Irno, T., 1958a. Immunogenetics of Salmonelia. Thesis. University of Wisconsin.

Ito, T., 1958b. Cistron test of motility genes in Salmonella. Ann. Rept. Nat.
Inst. Genetics (Japan). 9, 96.

Iino, T., 1959. Subunits of #7 gene in Salmonella. Ann. Rept. Nat. Inst. Gene-
tics (Japan) 10, 777—173.

tino, T., 1960. Transductions between curly flagellar mutants in Salmonella.
Ann. Rept. Nat. Inst. Genetics (Japan) 11, 73—74.

Ito, T., 196la. Genetic analysis of O-H variation in Salmonella. Japan. f.
Genet. 36, 268—275.

Iino, T., 1961b. A stabilizer of antigenic phase in Salmonella abortus-equi.
Genetics 46, 1465—1 469.

Iino, T., 1961c. Anomalous homology of flagellar phases in Salmonella.
Genetics 46, 14711474,

Ino, T., 1962a. Curly flagellar mutants in Salmonella. J. gen. Microbiol. 27,
767—175,.

Iino, T., 1962b. Phase specific regulation of the flagellin genes (H, and H,)
in Salmonella. Ann. Rept. Nat. Inst. Genetics (Japan) 13, 72—73.

lino, T. & ENomoto, M., 1962. Further studies on the non-flagellated mu-
tants of Salmonella. Ann. Rept. Nat. Inst. Genetics (Japan) 13, 73—74.

lino, T. & Haruna, I., 1960. A non-flagellated mutant which produces
flagellar protein in Salmonella. Ann. Rept. Nat. Inst. Genetics (Japan)
11, 74.

Iino, T. & Mirani, M., 1962. Effect of phenol on the flagellation of Salmo-
nella cell. Ann. Rept. Nat. Inst. Genetics (Japan) 13, 74,
Isex1, S. & Kasurwaci, K., 1957. Induction of somatic antigen 1 by bacterio-
phage in Salmonella B group. Proc. Japan Acad. 33, 481—A85.
IsexI, S.& Sakat, T., 1953. Artificial transformation of O antigens in
Salmonella E group. II. Antigen-transforming factor in bacilli of sub-
group E,. Proc. fap. Acad. 29, 127—731.

Incram, V.M., 1957. Gene mutaton in human hemoglobin: the chemical
differences between normal and sickle-cell hemoglobin. Nature, Lond.
180, 326—328.

Jacos, F. & Monon, J., 1961. On the regulation of gene activity. Cold Spring
Harbor Symp. quant. Biol. 26, 793-—277,
140

jJacos, F. & Woirtman, E. L., 1961. Sexuality and the genetics of bacteria.
Academic Press. New York & London. pp. 374.

Joys, T. M., 1961. Mutation in flagellar antigen 1 of Salmonella typhimurium.
Thesis, London University.

Karusu, F., 1960. Role of disulfide pairing in the biosynthesis of antibody.
Science 132, 7494.

Kaurrmann, F., 1940. Zur Serologie des I-antigens in der Salmonella Gruppe.
Acta path. microbiol. scand. 17, 135—144.,

KAUFEMANN, F,, 1954. Eniterobacteriaceae. 2nd ed. E. Munksgroard, Copen-
hagen.

KAUFFMANN, F., 1964. Das Kauffmann-White Scheme, in “The World Problem
of Salmonellosis”. Dr. W. Junk-Publishers, The Hague. p. 27—66.

KarriDGE, D., 1959. The effect of amino acids analogues on the synthesis
of bacterial flagella. Biochim. biophys. Acta 31, 579—581.

KERRIDGE, D., 1960. The effect of inhibitors on the formation of flagella by
Salmonella typhimurium. J. gen. Microbiol. 33, 519-—538.

KErRRIDGE, D., Horne, R. W. & Gravert, A. M., 1962. Structural compo-
nents of flagella from Salmonella typhimurium. J. Mol. Biol. 4,
227-238.

Kopayasul, T., Rinker, J. N. & Korrrer, H., 1959. Purification and che-
mical properties of flagellin. Arch. Biochem. Biophys. 84, 342—J361.

LEDERBERG, J., 1950. Isolation and characterization of biochemical mutants
of bacteria. Methods in med. Res. 3, 5—22.

LEDERBERG, J., 1956. Linear inheritance in transductional clones. Genetics
41, 845—871,

LEDERBERG, J., 1961. A duplication of the H, (flagellar antigen) locus in
Salmonella. Genetics 46, 1475—148].

LEDERBERG, J. & Epwarps, P. R., 1953. Serotypic recombination in Sal-
monella. J. Immunol. 71, 232—240.

LEDERBERG, J. & Itno, T., 1956. Phase variation in Salmonella. Genetics 41,
744—757.

Lerrson, E. & Hucu, R., 1953. Variation in shape and arrangement of
bacterial flagella. 7. Bact. 65, 263277.

Le Minor, L., Le Minor, S. & Nicotze, P., 1961. Conversion de cultures de
Salmonella schwarzengrund et Salmonella bredeney, dépourvues de l’an-
tigéne 27, en cultures 27 positives par la lysogénisation. Ann. Inst.
Pasteuy 101, 577/—589.

Lewis, E. B., 1950. The phenomenon of position effect. Advances in Genetics
3, 73—115.

Luria, S. E., 1962. Bacteriophage genes and bacterial functions. Science
136, 685—692.

Marco.in, P. & Muxar, F. H., 1961. The pattern of mutagen-induced back
mutations in Salmonella typhimurium. Z. Vererbungst., 92, 330—335.

MAKELA, P.H., LEDERBERG, J. & LEDERBERG, E.M., 1962. Patterns of
sexual recombination in enteric bacteria. Genetics 47, 1427-1439.

Marmor, J., Fatxow, S. & ManxveL, M., 1963. New approaches to bacterial
taxonomy. Ann. Rev. Microbiol. 17, 329-—372.

MEYNELL, E, W., 1961. A phage, ox, which attacks motile bacteria. J. gen.
Microbiol, 25, 253—290.

MrvakeE, T., 1960. Mutator factor in Salmonella typhimurium. Genetics 45,
17-14.

Miyake, T., 1962. Exchange of genetic material between Salmonella typhi-
murvium and Escherichia coli K-12. Genetics 47, 1043-1052.

Miyake, T. & DEMEREC, M., 1959. Salmonella-Escherichia hybrids. Nature,
Lond. 183, 7586.

Morsz, M.L., LEDERBERG, E. M. & LEDERBERG, J., 1956. Transductional
heterogenotes in Escherichia coli, Genetics 41, 758—779.

 
141

OrsKov, S. & Oxrskov, E., 1960. An antigen termed ft occurring in Ft £,
coli strains. Acta path. microbiol. scand. 48, 37—46.

Mxskov, F., OrsKxov, I. & Kaurrmann, F., 1961. The fertility of Salmonella
strains determined in mating experiments with Escherichia strains.
Acta path. microbiol. scand. 51, 291—296.

OzeExt, H., 1956. Abortive transduction in purine-requiring mutants of
Salmonella typhimurium. Genetic Studies with Bacteria. Carnegie
Inst. Wash. Publ. 612, 97—106.

OzeEKI, H., 1959. Chromosome fragments participating in transduction in
Salmonella typhimurium. Genetics 44, 454—470.

Ozex1, H. & Howarth, S., 1961. Colicine factors as fertility factors in bacteria.
Nature, Lond., 190, 986—989.

PacE, L. A., GRoopLow, R. J. & Brawn, K., 1951. The effects of threonine
of population changes and virulence of Salmonella typhimurium.
J. Bact. 62, 639647,

Parpeg, A. B., Jacos, F. & Monon, J., 1959. The genetic control and cyto-
plasmic expression of “inducibility” in the synthesis of f-galactosidase
by £. colt. J. Mol. Biol. 1, 165—178.

Ricumonp, M. H., 1962. The effect of amino acid analogues on growth and
protein synthesis in microorganisms. Bact. Rev. 26, 398—420.
Rupyer, R., 1961a. Mutation as an error in base pairing. I. The mutagenicity
of base analogues and their incorporation into the DNA of Salmonella

typhimurium. Z. Vererbungsl. 92, 336—360.

Rupner, R., 1961b. Mutation as an error in base pairing. II. Kinetics of
5-bromodeoxyuridine and 2-aminopurine-induced mutagenesis. Z.
Vererbungsl. 92, 367—379.

SakKal, T. & ISeEK1, S., 1954. Transduction of flagella antigen in Salmonella E.
Group. Geumma J. med. Sci. 3, 195—199.

Sasaki, I, 1961. Chi-phage resistance of the Salmonella serotypes having
g-antigen. Ann. Rept. Nat. Inst. Genetics. (Japan) 12, 82—83.
SmiTH, S. M. & Stocker, B. A. D., 1962. Colicinogeny and recombination.

Brit. med. Bull. 18, 46—47.

SNEATH, P. H. A. & LEDERBERG, J., 1961. Inhibition by periolate mating in
Escherichia colt K. 12. Proc. Nat. Acad. Sci., U.S. 47, 86.

Spicer, C. C. & Datta, N., 1959. Reversion of transduced antigenic
characters in Salmonella typhimurium. J. gen. Microbiol. 20,
736—1 43.

STARLINGER, P., 1958. Uber einen Defect des transduzierenden Salmonella-
Phagen. 2. Naturforsch. 136, 489-—493.

Staus, A.M. & Raynaup, M., 1964. Connaissances actuelles sur la nature
chimique des antigénes présents dans les Salmonella, in “The World
Problem of Salmonellosis’. Dr, W. Junk Publishers, The Hague,
p. 743—770,

Stocker, B. A. D., 1949. Measurement of rate of mutation of flageller anti-
genic phase in Salmonella typhimurium. J. Hyg. 47, 308—473.

Stocker, B. A. D., 1956a. Abortive transduction of motility in Salmonella,
a non-replicated gene transmitted through many generations to single
descendant. J. gen. Microbiol. 15, 575—398,

STOCKER, B. A. D., 1956b. Bacterial flagella: morphology, constitution and
inheritance. Symp. Soc. gen. Microbiol. 6, 19-—40.

Stockrr, B. A. D., Staus, A. M., TINELLT, R. & Kopacka, B., 1960. Etude
immunochimique sur les Salmonella. V1. Etude de Vantigéne I présent
sur deux Salmonella B et Ey. Ann. Inst. Pasteur 98, 505—523.

Stocker, A. D., ZInpeR, N.D. & LEDERBERG, J., 1953. Transduction of
flagellar characters in Salmonella. J. gen. Microbiol. 9, 410—433.

TerapA, M., Tomu, T. & Kurosaka, K., 1956. A doubly lysogenic strain of
S. typhimurium. Virus 6, 274—287.
142

UETAKE, H., Luria, S. E. & Burrous, J. W., 1958. Conversion of somatic
antigens in Salmonella by phage infection leading to lysis or lysogeny.
Virology 5, 68—97,

Ucuipa, T., Ropsins, P. W. & Luria, 5. E., 1963. Biochemistry (in press).

VetTAKE, H. & Haciwara, $., 1960. Somatic antigen 15 as a precursor of
antigen 34 in Salmonella. Natuve, Lond., 186, 26/262.

Usrtake, H. & Ucuipa, T., 1959. Mutants of Salmonella phage 15 with ab-
normal conversion properties. Virology 9, 495—505.

VELAUDAPILLAI, T., 1960. Transduction in vivo. Z. f. Hyg. 146, 470—480.

WaTANABE, T., 1960. Transductional studies of thiamine and _ nicotinic
acid requiring streptomycin resistant mutants of Salmonella typhi-
murium. J. gen. Microbiol, 22, 102—172.

WATANABE, T., 1963, Infective heredity of multiple drug resistance in bacte-
ria. Bact. Rev. 27, 87-775.

WEIBULL, C., 1950. Investigations on bacterial flagella. Acta chem. scand. 4,
268—276,.

WEINER, L. M. & Swanson, R. E., 1960. Chloramphenicol-resistant strains
of Salmonella typhosa. J. Bact. 79, 863—868.

YANOFSKY, C., HELINSKI, D. R. & Marine, B. D., 1961. The effects of muta-
tion on the composition and properties of the A protein of Escherichia
coli tryptophan synthetase. Cold Spring Harbor Symp. quant. Biol. 26,
17—24.

ZELLE, M., 1942. Genetic constitution of host and pathogen in mouse typhoid.
J. infect. Dis. 71, 137—152.

ZINDER, N. D., 1955. Bacterial transduction J. Cell, comp. Physiol. 45 (Suppl.
2), 23—49.

zinpen N.D., 1960. Hybrids of Escherichia and Salmonella, Science 131,
813—815.

ZINDER, N. D. & LEDERBERG, J., 1952. Genetic exchange in Salmonella. J.
Bact. 64, 679—699.