. ; j . {
' September 13, 1958 I whe

Dear Joshua,

During the transition period here | must get my own work (plus some of Ki ishi'd
unpub]. obs.) published. The article will be rather long and in a way dull because it
stresses method. As you already know the results became much clearer after | dumped
glucose out of the window (that is an exaggeration because you can't open windows in

the clinical center!) The bra phar phase generation time with glyceral as
carbon source is of course much longer and 1! lost a good deal of my night sleep -
but it was worth it. 1! may have told you that galacton in cons. of | x 10 ~ mol. has

a bacteriostatic effect on the growth of 3096 (transferase-less) which is quite dramatic
(this is a glycerol instead of! gluc. as C source). After 14-18 hours a "recovery" sets
in which may be selection of mutants. Yet these suspected mutants are still Gal neg.
This analysis of this phenomenon is first on the program. Likewise, I am curious to know
the effect of galactose on 3096 on the burst size if infected with virulent lambda.
Another new observation: 4- Cfiminrae is 10-20 times higher in the presence as

in the absence of galactose in 3100, 3096, 3092 - again provided glycerol is the C

source.

This is of course not surprising at all. Va'Gal 5° (3264 and 3265) in the
constitutioa& 4-epimerase is about 50 per cent (i.e., app. same order of magnitude)
as that found in 3100 and 3096. However, add. of galactose in "Gal 3'' does not give
10-15 times increase but only 50-75 percent of an increase (which may be within the
limit of error). We have found for instance that 4eepimerase conc. is going down to 30-
4O per cent in post-log phase as compared with the titer in the middle log phase.

"Gal 9'' is of course complete devoid of heepimerase, const. as well as induc. The
kinetics of galactosidase is also intriguing. In 3096 as follows:

| suspect that in 3086 the gal kin. is abmost operating as fast as gal permeace
and hence the conc. of feee gal in the cell very low. Sufficient to induce gal kin.
but in suff. to induce galactosidase until the acc. of gal-1-P (which is staggering)
begins to inhib. the phgsphprylahi of gal. and hence permit the free gall in the cel]
to attain higher values. This is born out of the fact that in 3100 and 3092
qa laches! increases (on a cellular basés})rapidly right from the beginning.
More detailed and precise information later. - Phenol extraction of complex polyseccW ule»
(cell wall, somatic etc) showed no galactose (if any maérety-traces) which agrees with
the fact that 3099 grows happily on glycerol without gal. - | am interested in marrying off
3099 with a Shigélla. Perhaps Yarns $ deletion mutants will also be
helpful.