Reprinted from Sctenck, November 11, 1955, Vol. 122, No. 3176, page 920,

Genetic Recombination in Bacteria

In the issue for 12 Aug. Science (p.
278) reported on the work of Wollman
and Jacob on genetic recombination in
Escherichia coli. This work concerned
the effects of mechanical disruption of
mating cultures on the patterns of segre-
gation. The authors interpreted their re-
sults in terms of fractional fertilizations
by broken gametic units, whose passage
into the fertilized bacterium is specifi-
cally polarized. The news report further
correlated this effect with virus-mediated
genetic transduction. These findings are
extremely interesting, but a report on a
single article cannot take the space to dis-
cuss diverse interpretations. This danger
is underlined by the fact that previous
news summaries [Science 118, 66 (17
July 1953)] have reported claims that
recombination in E. colt K-12 depends
on an F+ “virus” that acts as the vector;
the present article refers to “mating
pairs” and could not discuss the evolu-
tion in hypotheses. This letter is in-
tended as an extension of the remarks of
the news report that presents a more
complete picture.

Transduction, as the term was defined
when first introduced, is the transfer of
a fragment of genetic material by phage
or any other agency; fertilization implies
the union of whole genomes or nuclei.
Intermediate categories of genetic ex-
change may be found to occur naturally,
or as suggested by the current article,
may be artificially produced. For ex-
ample, fractional fertilizations (or post-
zygotic losses with the same effect) have
been indicated in frog eggs by sperm
treated with toluidine blue [R. Briggs,
J. Gen, Physiol. 35, 761 (1952)] and this
might be speculated on as one approach
to the achievement of genetic transduc-
tion in higher organisms. However, the
artificial interruption of fertilization

serves, if anything, to emphasize the com-
pleteness of the normal process and the
consequent distinction between fertiliza-
tion and transduction mechanisms.

Unfortunately, the true constitution of
the fertilized cell cannot be inferred with
certainty from data on haploid segre-
gants, owing to peculiarities of the mei-
otic mechanism in E. coli K-12. In the
cited experiments, fertilization might
have been incomplete, or it might have
been complete with later disturbances of
chromosome pairing to account for the
segregation effects. Furthermore, if fer-
tilization were fractional, the gradient of
recovery of various loci might be due, as
proposed, to the preferential orientation
of gametic chromosomes, or to depend-
ence on one locus (a centromere?) for
the completion of synapsis, crossing-over
and segregation.

These limitations of inference are
equally applicable to undisturbed mat-
ings and have provoked a diversity of
hypotheses that share the concept that
the Hfr or F+ gametes are normally in-
complete and variable in their genetic
content. However, all the experimental
data on haploid segregants are equally
consistent with a second hypothesis that
fertilization is regularly complete, but is
coincident with chromosome breakage at
specific, predetermined points on the Ft
chromosome, so that certain segments are
deleted after meiosis. Even small dele-
tions would be lethal in haploid segre-
gants and would influence the entire seg-
regation pattern.

The most direct means of analyzing
the intermediate processes of recombina-
tion and deciding between these con-
flicting viewpoints lies in the behavior of
nondisjunctional types, which have been
found both for sexual fertilization (het-
erozygous diploids) and a phage-medi-
ated transduction (heterogenotes) in E.
coli K-12. Thus the heterogenotes have

helped to clarify the nature of the frag-
ments in transduction, while the hetero-
zygotes support the second hypothesis
that in sexual recombination, the losses
are postzygotic. The evidence is that the
losses sometimes involve segments of the
F- rather than the F+ or Hfr chromo-
some, or even cross-overs between them,
and must therefore be preceded by com-
plete fertilization, synapsis, and crossing-
over. When the chromosomes are first
broken is not known, and this may occur
in the Hfr gamete, although the losses
ordinarily (perhaps not always) occur
later, after meiosis. However, the hetero-
zygotes have so far shown no evidence of
variability in the loci of breaks, for some
loci are regularly eliminated—that is,
they appear in the hemizygous state—
while others are always intact—that is,
heterozygous or, by crossing-over, homo-
zygous).

In this light, mechanical stresses might
be thought to create additional breaks
in the migratory chromosomes, to ac-
celerate the loss of previously broken
segments, or to alter their pairing ability.
The concept of mechanical breakage of
gametic chromosomes in transit is emi-
nently plausible and deserves to be tested
in other experimental materials. How-
ever, in view of the ambiguity of haploid
analysis, nondisjunctional types should
be looked for. If the hypothesis of frac-
tional fertilization is correct, interrupted
matings should lead to “diploids” that
carry variable fragments of the Hfr
genome, in contrast to the regular elimi-
nation types found from undisturbed
matings. The documentation for these
remarks may be found in another review
[J. Cellular Comp. Physiol. Suppl. 2, 75
(1955)).

JosHua LEDERBERG
Department of Genetics,
University of Wisconsin, Madison
5 October 1955