THE UNIVERSITY OF WISCONSIN
COLLEGE OF AGRICULTURE

Madison 6

DEPARTMENT OF GENETICS

December 25, 1952

Dear Bruce:

Hail, wassail stout fellow, joyeux noel, etc. We miss you too: it
would have been some party last night could we all have gotten togebhsr
over our cups. Spike (Spicer) helps remind us of Merrie England, but 'ere
too long you'll have him back too.

I'm sorry not to have ansered you sooner about your draft-- I find it dif-
ficult to say anything about it at this stage, except that it seems to cover
all the facts. I'd rather not have gone at this, and would rather see your
final draft, as it is only thete that various points of emphasis and fact
can be seen to be criticized. Nort’ hasn$t said anything to me about his
views on the papwr: I expect you'll hear straight from him. I'll try +o
dig out what suggestions I can at this stage, as you've prodded Esther for
them. First, thanks for the numerous cultures received on the 23d. I was
most interested in SL-13 -14 (Are these supposed to be identical). I wondered
whether there weren't some inconsistency between your transduction of a para
A, and its supposed incompetence to be adsorbed. After looking at its very
low efficiency of transinduction ( by PLT22/2 and 223/609 etc), I see there
may be no inconsistency, but I haven't tried to measure phage adsorption yet.
Fla I think may be better than F, although more cumbersome. F may already
have built up some connotations re compatbility. Fla ban be pronounced; Flg
cannot. Hy is probably easier than Hap for typographical reasons, but this
is no strong argument. I would agree on Fla, for 543, and will accede to your
other definitions henceforth, except that I think it unwise to save Flay for
SL-15. If it does turn out to be isolocal with Fla,, that will leave a gap
at the very beginning of the series which is awkward in completing checker-
boards, ete. I would classify SL-13 as provisionally Fla, on the basis of
the common linkage to Hj, and give it a distinctive number later if necessary
In fact, it would be psychologically advantageous not to give adjacent nume-
rals to linked loci, as this often tenda to evoke umix a speciously simple
image of structure. I would assign Flay to some locus definitely different
from Fla,. When it comes to numbering loci, I suggest the burden ts to prove
difference, though the shoe's on the other foot when it comes to assertions
of allelism. Slo 's ok, but I think a more elegant prefix would be desirable.
Have you any classical scholars in your pocket? Re phage-typing, fine. Unbe-
knownst to me, Edwards had been doing the same on the strains sent him, with
consistent results. On basis of rhamnose fermentation alone, the parent of
534 has to be #157. #157 came from N25, which I kmow have, and is a typical
"java" strain. The change b--1,2 must have been a sporadic mutation, but I
am making some further tests on the genetic homologies of its H,>. A major
difficulty is a j phase (233?) which crops up now and then, and for which I
have not yet gotten serum for further genetic analysis. Dave should have this
in hand soon, however. I applaud your plans to study the tracks. I don't
understand your remark on gp-1,2. Have I miscopied something? The experiment
was dublin 0 --x typhimurium, not the contrary. All that needs be postulated
is the substitution of H,8P for Hi. I haven't tried Ho* --x dublin.

The adaptation of PLT22 to paraB is still unsettled--just haven't gotten
round to a careful titration of "228" grown back on typhimurium, though I've
made the lysates.

If I can get round to it, I'll add a briefing on some recent experiments
on lysogenization. I'd better get to your draft first:
Ibe, II

Title: I like 'em shorter, viz. "Genetic analysis of Salmohella flagella", or
genetic determination, control, e.+.

Ia. Norton and I will probably set to review and summarize the evidence for
FA@ phage. Our recent lysogenization expts, seem to clinch this, and there's
more yet. You will have to make a lucid recap., but I would suggest not
including other experimental details in the first draft. They can be added

if there seems to be room.

b,g. Check. Table would be all right; not improper to group them, anticipe-
ting later work.

Ile. Definitely. Itéé reputation sheuld be restored.

How abeut some discussion of physical properties, structure of gelatin agar?
(BeG. temperature respense).

III check. Incidence of lysogenicity may depend on adaptation of phage to
host. Eeg,. with LT-2 and PLT22, it is very high; with PLT22 ani 666, very low,
IVa. table here: alwage aasier to refer back than forward. Gallinarim not
known to be transinduced for any character.

Vcheck. VI cy. VII che Hirsch's 0 mutant might be very good for this in
view of its high rate of phase-variation.

X. Here is the first place my emphasis in thinking differs from yours. It
would be premature te conclude that the linkages have any phyelological
significance: could be coincidence. Alsc, argument is historically reversed.
The linked transduction was stumbled on before itcwas looked for. Perhaps

the closest analogy in EB. coli is V,P-¥,7-V,8 system (OSH '51). Very close
linkuge of "pseudoalleles" is excep tional—- your draft seems to give the 5
opposite impression. Try lwoffing 548? pell line 6 f.2 Which 2 instance:
"in the present instances" 2 Oriticlem ef 543 as possibly a "monophasic abom,"
--what dees it mean? What would you call tgymmexsexxx sundiege --x 543: monoph
asic chester, sandiego, stpaul...? Are you asking the question, is 545 recent
ly or simply derived from a typical b:12 or bionx strain? K. can't tell you,
but I'd be amused to hear his response. Most people will recognize what you
mean 4f you write paratyphi B, not java. Edwards mentioned that it reacted
with various paraB phages, but no simple pattern.

Ratio of bti is variable with FA from different stocks, and is rather poorly
defined as later wwarms may show diffenent ratios fron earlier. I would not

go into this, but simply put ca. 5% for tem -—-x 543. Might mention that

Gal- marker of 666 rules out contam. to explain i's, as doee their monophasic
behavior.

Detailed report (mirror absorptiens) on crucial strains would be defirable.

Can also mention general confirmation by Edwards, but phrase this carefully
(perhape, "mtmttar cultures obtained in similar expts. by JL have been veri-
fied by FRE) to avoid implication that we felt independent verification of

their findings was necessary. xtuxxpxsifckexxx

But you really can't hold me to a serious discussion of the draft, which is
superb. Will you give me an oppertunity to see the final papert I promise to
be most prompt.

Back to lysogenization expts: (I must have included some remarks with my
shipment of cultures -Have you gotten these yet? Also, have just sent out
some reprinte ef Z&L, and that review of which most of the punctuation is
yours. Supplies are fairly tight, but suggestions on who might profitably

use addn'l reprints are welcome). In my last letter, I mentioned a prelimin.
expt. This has been repeated with a more satisfactory control: Gal<dtla¢ was
added in small numbers to Gal-Fla- plus FA(Gal+). Papillae were picked and
scored as Fla+/~ and lysegenic/sensitive, after purification. The transductions
(Fla- Gal+) were 18 lysogenic | 3 sensitive; the insertions (Fla+ Gal+) came
out 3 lysogenic : 43 sensitive. Esther has done an equally decisive experimen
with lambda. In both cases, the multiplicity of phage was high, and it was
difficult at first to give detailed explanation, except in the general terms
that lysogenization is directly connected with trameduction. The former has
yet to be well understood, but the following reasoning may be useful. The non-
lysogenic sutvivors in these experiments can hardly be regarded as cells that
have never heen infected. Instead, they are likely to be part of the progeny of
infected celle, and, often, sibs to celle that have either lysed or become lyso-
genic. In this connection, 1t is worth noting that Esther found that the trans-
ductiona were not obviously mixed lambdat and s, whereas most of the controls
that showed any lambda were. What we are doing, then, is to fix on that part of
the progenies which have had the best opportunity to become stably lysogenic by
separating the transductions from the parent cells, and this accounts for the
higher incidence of lysogenicity. In E. coli, this incidence was 100% in the
transductions; about 2/3 in the reat of the population (a good deal of this is
overestimated by reinfection, of course). I had been plarming a similar experiment
at lower multiplicities, and using antiserum to prevent reinfection, so that the
limiting factor would be supposedly the amount ef phage, but until the lysogeniza-
tion aspect is better understood, this may not be meaningful. In addition, even
undiluted serum ia not entirely effective in preventing cross-infection from
bacterial infective centers. The point of these experiments is, of course, to try
to show that Fa=phage not only with respect to the skins, but also the sontents.

Another approach is now possible with the help of a “lytic variant" of PLT22,
22V, noticed casually on LT-2. 22V lyses almost completely (survivors mostly roughs,
not lysogenic, so far), but LT2(22) de resistant (Cf. C and C' of Burnet and Lueh
1936). Thus, if 10° PLT22 ie added to 10 LT-2, followed 10 minutes later by excess
22V one gets nearly 10° survivors. The plating of the PLT22-infected LI-2 by itself
gives an expected proportion ef contaminated, rather than lysogenic colonies, in
agreement with Esther and my previous expts. with lambda. So I do not think, wfor-
tunately, that resistance to 22V requires the ultimate stable lysogenic state. (I
note your wention cf a etudent doing something similar-- we shall have to arrange
to avoid unnecessary overlap.) This protection experiment is something I longed to
do with lambda some time ago, especially whea I thought of the “4ransformatione”
as a sort of distorted lysogenization, but I never could find a t:utant lambda or
other phage with the hecessary properties. Bominski had something similar too, of
course. 22V should make it possible to connect FA with particles with protective
ability, which will in turn produce lysogenic survivors. If most of the transduc-
tions are preserved (after infection at low multiplicity with PL1T22) tichmonmmia
while most of the rest of the population 1a degroyed by by 22V, this would again
correlate phaze particles as actual carriers of FA. (In this, I an contending with
the counterhynothesis that FA is phage max skins that have incorporated bact#ial
fragments instead of phage nuclei.) As 22V itself has a trace, though rather definite,
tranducing activity, teated on lysogenic receptor strains, some elementary precau-
tions are needed for the experiment, and these have dedayed 1t momentarily. 22V seems
to be temperate for SW666.

Some more data with UV: By very long expesures (20 ming., our sterilamp at 50 om)
lytic activity of FaCe.g. SW618) can be reduced over six decades, while FA is dimin/
ished less than 1. This permite traneductions (Gal+) and plaques to be counted on
the same plate. The former are not lysogenic. As I may have mentioned, the intent
of these experiments wae to dissociate Fla from H; in the linked transduction. It
appears doubtful, however, that UV 1s reaching the genetic material at all. In addi-
tion, the proportion of b:i in SW618--x 666 varies as between the early and the late
swarms, and cannot therefore be accurately measured. I do not understand thie very
well, the purified isolates have about the same motility afterwards. Either there is
a difference in the time of initiation, or the initial differences in rates (polygenic
linked modifiers?) are levelled by subsequent selection before the isolates can be
purified. One expt. with X-ray was not very promising: 200,000 r left about 10%
phage and 30% (+) FA. Theseg doses are too large to make any thorough investigation
feasible.
Larry has been doing some UV-induction (lwoff) experiments for me: LT2, LT22
do not look very promising, but the SW543 line, infected with 22B, seems to be
working very well. Lwoffates titrate to about 1010, have the same pattern of
Gal+, Fla+ transducing activity as lytically grown phage.

On phase genetics, I have been temporarily stopped, waiting for some new cul-
tures from Edwards which may be sufficiently stable. For good technical reasons,
I need diphasics (like abony) of which H, is neither i nor 6, and of which Ho

does not react with anti-1,2.... If you yourself happen to be acquainted with

any such serotypes of which you have the knowledge that they are relatively stable
in flagellar phase, I should appreciate them. (Perhaps Joan Taylormight--would you
ask her? I've sent Edwards the specifications, suggesting bispebjerg, abortus—bovis
(aomptent to absorb?), durban or other abony's.)I could use some of my own
recombinants, and perhaps will, but this might not be regarded as cricket. One gets
into some perplexities wondering how a "heterophasic" transduction will work out

on my scheme.

I have not yet succeeded in getting 1 from 1:1,2 —x 666, but have perhaps not
tried enough. It is possible that gx H, 1+ ds capable of functioning in the mono-
phasic residual genotype of SW666. This could be tested by looking at the competence
of ,the (Hy te ~-x 666)i —x abony, where it would no longer function. Alternatively,
Hy, +* may undergo sufficiently rapid variation to Hy in the 666 residue to allow

the i phase to be detected, especially if Fla+ H,1* 1s imuotile, and therefore
strongly selected against.

On a visit to Boston last month, I had an excellent time with Dieses, and took
some of his material back with me. I think I can now grow lifcrms from Proteus
without tco much trouble, but quite large inocula are still required at each
transfer. He has promised to send me some material of the same scrt from Salmo~
neila for some genetie experimenta (e.g. Fla+ L —x Fla~ Bact.) These L&s are
entirely irrevertible, which is some slight advahtage. It seems to be necessary
to preserves the integrity of microcolonies cf the L's to permit further growth:
they either must make a surface film for themselves on bpoth, or have physical
support, either agar at just the right concentration, or cotton fibers, or as
seems to be working a film of collodionm. Microscopically, they ave wierd; I do
not yet have any evidence of my own that they are genetically derived from the
original bacteria, and am keeping an open mind on the whole business. The small
size and growth habit almost suggests that the genetic equivalent of a bacterial
cell is the L-colony, but this is a speculative fancy.

Word of your magnificent talk also reached here via a letter to Spicer.

Have you heard of Novick&Szilard's recent work on antdmitagens?~~you must have

en route home, It would be fun (forvyou?) to try these on the Salmonella mutations
(phase variation, Flat..) In fact, there's still no evidence of any mutagenic
erfect on these factors, is there?

Esther and I have been discussing your phage-chromosome speculation, Most of this
is already deeply engrained (as speculation) around here, but why does the site
have to be a chromosome ehd? What kinds of experiments would bear on this?

A riverdice,

Qt

\pu. f
i. LA