UNIVERSITY OF CALIFORNIA

DEPARTMENT OF PHYSIOLOGY
SCHOOL OF MEDICINE
BERKELEY 4, CALIFORNIA

September 21, 1955

Dr. Joshua Lederberg
University of Wisconsin
Department of Genetics
Madison 6, Wisconsin

Dear Dr. Lederberg:

Your letter of July 28th, with information on the Oak Ridge Symposium
volume, was forwarded to me in Europe. ‘Thanks very mich indeed. I was able
to locate this publication in Geneva and to read a good part of it then with
great interest.

As a consequence of writing my paper on the origin of sexuality (which
is now in the hands of an editor, who has not informed me of its fate), I have
been inspired to speculate on a number of interesting problems. One thing that
intrigues me is the question as to whether reciprocal "crossing over" occurs ;
between adjacent nuclear bodies in @‘mltinucleate (usually quadrinucleate) cell :f
swm@mees Escherichia coli, when it undergoes ordinary asexual division in the
“haploid” state. So far I have found no answer to this question either from
the literature or from discussion with friends and colleagues (Adelberg, Garber,
Szybalski). Yet it seems quite likely that you have information that could
dispose of the question quite readily.

 

On the off-chance that you have not investigated the problem, I thought
to write you, describing a theoretical model that I have derived as a possible
means of testing such a phenomenon. As a non-bacteriologist, I do not have the
experience and knowledge to suggest mach more than certain principles and I
realize that an actual test depends upon the availability of suitable genetic
systems, which may or may not exist.

The basic requirement for such a test would, as I see it, be that one
mst be able to distinguish the four types of descendants that would be derived
from a single heterokaryotic E. coli cell containing four nuclei of the type a
AB, Ab, 8B, and ab. The only way that I can conceive of deriving such a cell] is “etecfavie
would be to have (induce) a double mrtation followed by a "cross-over" one nuclear (“>
generation after this mtational event. A scheme of doing this may be proposed
as follows:

i. One mst have a double auxotrophic mtant ab, requiring substances
a@and 6. Certain necessary, or desirable, properties of AB, Ab, aB,
and ab organisms are:

&) Wild type (AB) mst produce enough diffusible @ and B to
permit both types of single auxotrophs (Ab and aB) and
(preferably to a lesser extent) double auxotrophs (ab) to Co
grow more slowly in the presence of AB cells on agar plates,¢f minim! ivitiom,

b) It would be highly desirable if some easy way of distinguish-
ing the four genotypes (by color; colonial morphology;
Dr. Joshua Lederberg -2- September 21, 1955

2.

36

h,

reactions to some indicator medium, or media; etc.) were
to exist.

The mitation rate of a +A and b - B probably would have to be

sufficiently high under normal conditions that double mtations

within a of ee cell would occur at, say, about the level of

one in 10° or 107 cells. (I realize that this is very high, but

I understand at least one "mitabile" strain of E. coli exists.)

Or, to put it a bit differently, the rate of le=-back

mutations after heavy irradiation (say on 10° survivors out oo,
of 10” cells) should be about one in 10° or 10°; cells. Bee ONG

One would then grow ab cells on an xe celle/n medium (contain-

ing a and 6B) to a density of say 1x10° cells ful. »,irrediate
heavily so that a small percentage survives (1x10"), and then wis,
dilute appropriately, and plate out on minimal medium so as to

give between (?)100-500 surviving cells per plate (if all could
grow).

Presumably only AB cells would grow, plus mutant cells in close
enough proximity to receive enough substances a and/or B.
Possibly also Ab cells could grow in company with aB if they
produced the reciprocally needed factors.

The colonies should then be studied for sectoring, which would
be evident by irregular colonial shape or (better) by more
easily detectable morphological properties. Those colonies
with four types of sectors representing AB, Ab, aB, and ab could
presumably come only from a cross-over (in accordance with

accompanying diagram).
I should be mech interested in your comments on the foregoing scheme.

There may be fallacies not apparent to me. Also some other (more practical)
way of distinguishing four such genetic types deriving from a single cell my
well oceur to you.

Ene Losure

ECDs: jd

Very sincerely yours,

Ellsworth C. Dougherty
dovotc metahon
oA

\

- 4. “Ctserng C66
F, mere AA Nf)
‘ah [ON Ce wile’ hy -

eae oe

 

 

Pes ble cedlecua (

mer pheleg 4j

Sther types ok amved cadens mrghh be -

. . ~ diveche ravale tics feu
g ah snaich might reset Fron 6 unt
so
okt

yucle: atte, devote me heartiery wll cel!

‘ > (pz Hw ¢ Ev
onsrage

aks . ve eke searcygeanen
: hk omaght be expected fo otc Free eG SY
A \e AG Whre 4 vot
fo |! wb - = cpa rite | motatee
q ZV Xng \ ,
4B

c 7 vue be : Laom © WX; le cel!
of jebicnenidag ima

+=