August 4, 1946.

Dr. P.R. Edwards,

Dept. Animal Pathology,
Kentucky Agr. Exp. Sta.,
Lexington, Ky.

Dear Sir:

As I suggestea when I wrote to you and Dr. Bruner a year ago, in a request
for reprints, the phenomenon of phase variation in Salmonella has long seemed
to me to be one of the most interesting in genetics. On the hasis of research
now in press by Dr. E.L. Tatum and myself, it may now be feasible to attack the
problems of inheritance in coliform bacteria by Mendelian methods. Since it
will be some time Before this work is in print, I will take the liberty later
in this letér of summarizging that work. This letter is addressed to you and
Dr. Bruner, on the basis of my reading of the Salmomplla literature, in the
hope that we can initiate a collaborative project on Salmonella genetics, or
if you prefer, that we can .timlate you to apply some of the procedures that
we have developea in E. coli.

Briefly, we have found in E. coli evidence, that we find difficult to refute,
for the recombination of characters in mixed populations. The characters that
were used were a) a series of X-ray and ultra-violet induced mutations for growth
factor requirements and b) spontaneous mutations for resistance to spe@ffic
phages. Biochemical mutants grown together give rise to wild type cells; in
addition, new combinations af nutritional requirements and phage resistance are
found. There is sufficient data at this time that we are fairly confident
of the existence of 4 recombination process, but before the work is published
in extenso we should like to accumulate a good deal more. Recombination would
seem to imply a method of sexual reproduction, on which we have no evidence.
Under cultural conditions so far used the process is rare, about 1 cell in a
miliion being a recombination type, and we have had to use selective media to
demonstrate them. So far only one coli strain has been successfully used; another:
which was non-motile being unproductive. The recombinations occurred between
mutants of the same strain.

Nhile these conclusions are in the broadest sense highly tentative,if
is my feeling thit they are weal enough established to explore the possibilities
in other organisms, The tabulation of nature lly occurring antigenic types in
Salmonella has always suggested to me a recombination process, whereby different
antigens can be associated with each other in the very many divers permutations
characteristic of the numerous 'species,' It seems to me that it would be very
worthwhile to attempt the production of new types by the method sketched above.
This woukd involve inducing (by radiation) nutritional mitations in different
Salmonella strains, grovth in mixed culture, plating on minimal media to select
the nutritional recombinations, and examination of these 'wild type' cultures
for antivenic variations. ‘hile we have considerale experience in obtaining
and handling nutritional mutants, we are not at all adept or equipped to do
antisen analyses, ,for which reason we hope that we can arrange to work together.
To illustrate what is proposed consider the flagellar antigens of
S. typhi-murium and 5, abortus bovis, which 4n my table are characterised
as i,.1,2,3 and be. enx. “e would obtain biochemical mitants in both
these strains, let us say biotin-less «nd leucineless, respective ly.They
would be grown in mixed culture, and then plated to isolate colonies of
nutritional recombinangs, requiring neither biotin, nor leucine. Such
colonies (which occur in &, coli, and we hope may occur in Salmonella)
would te sent by us to you for examination of antigenic status. Cne might
look for the occurrence of flageliar antigenic formul:e 1..e,n,x such as is
found in bonamiensis or Deeds 2rd in abony « ,

“y cursory study suggests that ‘these ‘strains might be the most suitable
as a start. Such an andlysis is required befcre one can begin to consider
phase variation, If you are disposed to work cooperatively with us, I would
Sugeest that you send us cultures of the orgenians mentivuned, in both phases.
It would also be helpful to have a small amount of monospecific sera so
that we might do some preliminary screening of what we sent back to you (if
anything/ arbse ). I hope that the cultures mentioned are only mildly patho-
genic if at all; we cre not a medical laboratory. If you should have any
other suggestions as to hich strains might best be used, please voice them,

It is possible that you would prefer to work on this problem in your
own laboratory. If so please let us know, and we shall send you our publications
as seon as they are available. The barrier of distance may be of some import-
ance; if there is anyone of special competence in the ~almonella serology
in this area, with whom you feel thks problem might be more profitably discussed
please let me know,

Lastly,. I should be grateful to you and Dr. Bruner if you would send
me reprints of your papers for my files. J. Bact 37:365, 1938; 42:467, 1941,
and Froc. Soc, 41:223 1939 have been recived, for which, thanks.

Very sincerely yours,

Joshua Lederberg.