MEMORANDUM
BRISTOL LABORATORIES

UNIT OF BRISTOL-MYERS COMPANY

 

 

 

 

FROM J. Lederberg DATE August 17, 1955
. ¥ § -
TO J. Lein SUBJECT _ ew Antibiotics Sereening Program
CL.
Dear Joe:

Let me take up some of the issues in your last communication in order.

I am not terribly surprised that agar diffusion techniques prove to be probably
unsuitable for routine use. I think they could be made to work but only in the
hands of a professionally skilled worker. In the long run you would have to

repeat most of your experiments with liquid culture mixtures in any event so there
may be little enough point in using the agar diffusion methods. I imagine if the
need required it would not take much ingenuity to design an automatic filling machine
which would automatically deliver different ratios to input fluids upon setting a
dial or some such thing. Ina setup like yours I suppose the human hand is the more
appropriate instrument.

As far as the general sereening progr-m ia concerned, I can see very few things that
we have not already discussed and for which there already has been a decision on
one basis or another. I fully sympathize with your problem in having to make
decisions often on an expedient basis or even less than that without having very
many of the pertinent facts at all but I can't see that you have actually gone very
far off any possible path.

The only points I would bring up again would be the queries (1) Should one adhere
exelusively or primarily to Actinomycetes as the sources of new antibiotics? (2)
Whether a mtation program still could not be applied more routinely, beoause while I
think your argument that mutations leading to increased production are likely to be
on the whole infrequent, it is prebably valid that only experience would reall tell.
My suspicion would be that when you start with relatively low yields it should not

be extremely difficult to obtain mutants which give moderate performance. I would
agree that going from moderate to superior strains represents the most difficult step.
If I am not mistaken there was an entirely similar experience in the develepment of
penicillin-producing strains. I think it would be worthwhile with any culture which
seems to be producing a possibly promising or new antibiotic but in relatively small
quantity to do platings along the lines of Kelner's technique to try and look for
improved production, even simply as the bade for further preliminary analysis. Of
course, a much more exhaustive job would be done whenever you approach the question of
routine production, The matter of replating the cultures from the wild in order to
separate out the possible heterokaryons would seem to me to be much too important to
be by-passed at almost any cost and some of the occasional inconsisteneies in production
from one run to another might very well be due to this cause. I would think that as
soon as a culture seema to be producing any antiblotic in which you would have any
interest it would be worthwhile to replate it and do further studies on two or three
reisclated single colonies.
fe: J. Lein August 17, 1955 Hew Antibiotics Sereening Progran
Gonsultantship Arrangement

T wae extremely happy to hear of the sugeessful outcome of your discussions with
Bicchenistry on further emphasia on less potent broths since this is perhaps the

most obvious vulnerable point for the seareh for new materials, These are eertainly
the items which will require the most skill in order to handle and from that point

of view are the most likely to have been overlooked by your competitors, In addition
one would imagine that there will be relatively few antibiotics of very high potency
on first isolation and more of less and less potency and therefore the crean having
apparently been more or lese skimmed it is necessary to go into thinner and thinner
veins,

I would again emphasise the possibility of discovering adjuvant or aceessory anti-
biotics, 1.6., compounds which by themselves might have little antibiotic aotivity
and not enough to justify marking them alone but which could be extremely important
in maintaining the usefulness of other more striking aetivites, This may net be
expressed in the normal course of events in terms of simple synergism and I ae not
sure what the best methods would be of picking up usefal accessory agents. It might
be worthwhile, however, to consider what might be done with more or less reconstruction
experiments involving infections with challenges consisting of primarily an organisn
sensitive to a primary antibiotic containing a few organiams resistant to it to see
whether an accessory antibiotic would have the desired effect of eliminating the
last residual resistant organiams. With tuberculosis eapecially this would be an
extremely useful and pertinent result sither with etrevtomycin resistance or to a
leseer degree with isoniasid resistance,

From your previously expreseed interest in combination effects I would judge that you
have already given considerable thought in the matter of synergisms. Wow the first
proposal that I had made before with regard to accessory antibiotics would not
necessarily be reflected in any obvious synergistic effect since the primary anti-
biotie would probably be the one whieh is nost readily manifest in in vitro experiments.
Synergiema represent the seeond approach and here again, of course, there should be
many opportunities for the interaction of two antibiotics which separately sre too
weak to be very much good by themselves. I would imagine thet the point at which the
nost emphasis should be given to such accessory effeots would be where an antibiotic
has proven to be more or less active in in vitro teets, say even in the presence of
serum, but which fail when used by themselves in in yivo challenges. In such eases

it might be necessary or desirable to go right ahead with mixed therapy without the
benefit of very much knowledge of the in yitro interactions, the latter not necessarily
being of very great relevence to the therapeutic effects.

I was startled and interested by your interest in the seareh for virus-indueing agents
om the basis of the asaserine effect. It hadn't occurred to ne that this correlation
would be of direct chemotherapeutic interest but of course it is. It is doubtless true
that most of the agents that have been successfally used in anti-tumor therapy would
have this inducing effeet on lysogenic bacteria, X~rays, nitrogen mustard and asaserine.
I have been acquainted with this effect of azsaserine for some time but had forgotten

to think of asaserine as an antibiotic croduced by an Actinomycete.

I think one should look a little bit at what is behind this correlation and I think
the root of it is some fundamental common basis of mutagenic or more aceurately so-called
radiomynetic activity, that is to say, all of the agents that I have enumerated very
closely resemble X-radiation - their biological effects. To the compounds indicated
fo: J. Lein Auguat 17, 1955 New Antibiotics Sereening Program -
Coneultantship Arrangement

one can add such common agente as formaldehyde, hydrogen peroxide, and in fact I think
there are reasonably good grounds for believing that hydrogen peroxide is a key media
in many of the other radlomynetic effeotse of chemicals and of radiation in particular.

IT have some notions of my own about what the next step in this correlation ig. I have
the feeling that the one thing that is common te most of the radiomynetic agents is that
they are potentially powerful alkalating agents, that they are capable of substituting
on free amino or carboxyl groups under physiological conditions. I think even the
peroxide ean fit into the picture if one recalls that in the first place hydrogen
peroxide forme complexes with many organic compounds which have reactive hydrogen and
that these organic peroxides in turn can function as alkalating agente. In this connection
you might be interested to know that such familiar alkalating agents as acetic anhydride,
acetyl chloride, dimethyl] sulfate and ethylene oxide also have radiomynetio effeets in
other test systens, @.g., chromosome breakage and in the induced breakdown of diploids
in Escherichia coli. You will see something about this in table 10 im our paver in the
1951 Celd Spring Harbor Symposium and rather more recently in «a paver that I think you
will very much want te see ~ I believe this is a review by Levelese in either volume

1 or volume 2, probably volume 2, of ADVANCES IN CANCER RESEARCH. Levelese there has
reviewed what is common in the chemical reactivity of a variety of carcinogenic,
mutagenic and anticareinogeniec agents. Now not all of the agents indicated have heen
tested on the lysogenic system bat there are already so many parallels that I would be
willing to bet that practically any compound or agent which is mutagenic in this way
and which breaks chromosomes and breaks down dipleide of Bagherichia coli is going to

be found to be an inducing agent provided its toxic effeete:on the bacteria do not
already override any possibility of this lysogenic induetion. Therefore on this basis
the search for compounds with indueing aetivity would he equivalent te a search for
powerful, natural, mutagenic agents and I believe it has been shewn that azaserine
already falls into this category (Demerec). What may be unique about these antibiotias
that distinguishes them from such destructive chemicals ae ascetic acid anhydride and
acetyl chloride is that they are relatively stable in aqueous solution so that they
have some possibility of entering the target cell that you are aiming at.

Precisely why so many radiomynetic agents should in fact prove to have anti-tuaor
activity ia not at all clear any more than what is certain why I~rays do have sone
differential effect on tumor tissue, presumably thie is all bound up with the question
of rates of growth and the like which distinguish tumor tissue from other cells. There
may be also problems of differential penetration which will be one of the few points

of aeceas to differential tumor chemotherapy. I think the question of sereening for
anti-tumor agents on this basis then perhaps reduces to what is the most effective

method of looking for agents with radiomynetic aetivity. WM I were working in « research
laboratory primarily, I would think that the breakdown of diploids would be one of the
pest systems for examining antibiotics from this point of view and you would get results
along the lines of table 10 that I referred to in the Cold Spring Harbor Syeposiun. I
regret to say that these diploids are so hard te handle that there is really no poesibility
of adapting them to routine use in thie way. However, inatead of diploids of Escherichia
eoli it might be possible to work up techniques based on the breakdown of diploid
heterocycles in other organisms, especially yeast. Here again, there may be many
technical problems in working out the nature of the effects of the killing agent, but

the handling of the yeast culture itself would be very much less of a problem. I think
for the moment that this approach, this use of a test for genetic effects, is something
fo: J. Lein August 17, 1955 Hew Antibiotics Sereening Program -
Consultantship Arrangement

that ought to be thought about in the long run but is a leng ways from being anywhere
near ready for routine application, Any teet of this Kind is bound to mean much nore
attention to the antibiotic effeete of your reagent than you are likely to want to give
in a routine preliminary screening program. These are thinga that you might de on an
intermediate level before going on to test your material, let's say on an Ascites tumor
situation.

fhat leaves now the question of whet other systems would be most appropriate for

looking for this kind of activity. Another system that would be poasible would be one
which involves teeting for active mutagenie potency and some of Demerec's fundamentally
rather sloppy but etill workable assay systems, for exemple, the induced reversion from
SD to SR again might be applicable in a research laboratory atmoephere but I wonld
ehudder to think of having to set up such a program on a routine basis. I think that

the possibility of lyscgenic induction as a screening technique does deserve cleser
coneideration. I think it is possibly the one which is most likely to be readily useable
which ia the least likely to be misinterpreted in the details of ite results but again

it 1s going to require a lot of attention in trying to set it up on a routine basis. Now
you are quite right thet almost the only way of leoking for effects of thia kind is

going to be to leok fer the preduetion of phage rather than for the greas destruction

of the lysogenic bacterium. One might hope that one could get a differential on that
basis and that is the reason that I gave you the strains but you would be beund to miss
many reagents as you already know that you would if you relied on an observable difference
in the survival of LP+ as compared to LPS bacteria. You are quite right too in

supposing that the normal untreated lysogenic strain is going te produce phage without
the intervention of some inducing agent. However, there ought to be and there is as a
rule a very sharp distinction between the amount of phage which is produced spontaneously
and the amount which is produced under the influence of a euitable indueing agent. How
one will have to pay some attention to detail on this because in order to get effective
induction it is necessary that the cells being treated, until the time to which the
material is added, be in optimal physiclegical condition and you are also going to have
to work out some more or less @mi-quantitative way of meaeuring the phage that is produced
in the vresence of an antibiotic, That means of course that you are going to hava to

do these tests for this particular purpose to begin with in a liquid medium.

I think the quantitative assay does not nesessarily present too many problems because
one can fairly easily set up a leepful dilution type of technique which will get

around that question, that is to say, you might use a routine along the following lines,
grow your bacteria in broth, use them at leg phase, that is to say, add the antibiotisa
to them and make sure they are being adequately aernted during the time that the anti-
biotie is working, say some 2, 3 or 4 hours later take a loopful of the LP+ bacteria and
antibiotie mixture and dilute that in a ee of water and then take a leepful of that and
spread it on a etrevtomycin-resistant indicator.

We have found that EMB agar without sugar but with peptone is one of the better tyves

of media to use for the detection of plaques of lambda. They show up rather more
sharply I think than they do on straight nutrient agar - that is a straight surface
plate technique that you can use. Now in order to use this method you will have to

have a streptomycin-reaistant mteant of a lambda-sensitive atrain. I ses no reason at
all that the W148§ strain could not be directly adapted — I'm not sure off hand whether
or not we have a W1485 streptomycin-resistant stant of our own, If we do, I will
gladly send it to you. Another stock that you could use would be strain C or some stocks
fo: J. Lein August 17, 1955 Yew Antibioties Sereening Program ~
Consultantship Arrangenent

related to it. There may be some advantages to thie. I will arrange to send you a few
different leaabda-sensitive atrains that you can leck at yourself and see which one of
them would be the beet indiaator for your purposes. I think you may have to prepare
your own etreptomycin-resistant mutants of each of these but that is routine and
shouldn't hold you up fer very long. I just don't think that we have very wany of
them. At any rate the ona that I will send you next will include strain ¢ and another
one that we may already have that ia streptomycin-resistant which is sensitive to
lambda.

Dees this kind of a protocol seem too difficult to be worthwhile fer your purposes? I
ado not see many ways in which it can be simplified and you are going to have to work
out «2 few of the details with known agents like azaserine or nitrogen mustard or
ultraviolet light in order to see exactly how it is wing to go.

One alternatively could use differential quantitative measurements of killing of
lysogenic as against sensitive bacteria but on the whole I think that would be less
reliable than would be the production of phage. The reason would be that if even

only 15% of the bacteria are lysed that represents a very large output of phage as
compared to the normal situation whereaa that redueotion in viable titer would not mean
a great deal at all.

De you think I have given you enough detail on the probable method? This is something
that I am sure you have pretty well thought out for yourself. I think rather than try
to gpell the whole thing out for you I'll give it to you ae a form and then if you
want to try it you san come up with thie or that question and we'll de our best to
help you out. I think you conld get started right now with your W1485 strain and I
will see about digging up some of these others, In the meantime I will give the matter
some more thought myself and see if I can imagine the way of getting at the question.
On the whole though, I think the use of a streptomycin-resistant indicative strain on a
etreptomycin medium would be the best approash to leeking for the croduction of lambda.
I think tee that there is little question that the lambda system is perhaps the beat
for doing this tyne of work. There are many other indusible bacteria; you could
probably pick up any of them. I think there is snough generality in the effect of
mutagenic agents on different organisma that there would not be a great deal of point
in spreading your effort over too wide a wariety ef responding organisms. This is
rather a different situation, I think, from the usual modes of antibiotie agtivity.

fhis itself may be something of a hint because I think it would be reasonable then
to suppose that most any antibiotic with a narrow speotrum of activity among the
bacteria is not going to be a likely candidate for this sort of work whereas those
antibiotics that have seme effect on a wariety of organiems are those which ara most
likely to work.

That just now led me to another suggestion. Another teehnique of lesking for such
differential effects may in fact be at least as good as any of these, and that is
taking advantage of the rather remarkable differential insensitivity to radiation of
strain B and strain B/r. Wow as you probably know this differential is not unique
to UV light and to X-ray# but to some unknewn mechanism appears to apply also to
other kinds of radiomynetic agents so I think that one might give some attention te
the very simple possibility of comparing strain B and B/r,and any antibiotic which
differentiates seriously between them would be a likely candidate as a radiomynetic
agent. I don't know off hand whether asaserine has been tested directly -- I would
To: J. Lein August 17, 1955 New Antibiotics Sereening Program
Consultantship Arrangement

ve willing to predict that there would be a significant differential between the two.
Wo one knows what the basic mechanism ia for this difference in sensitivity. This

in turn has led te another thought that might be quite apropo to this kind ef analysis,
and that is a comparison of the survivorship of hapleid as against diploid yeast.

Since there are quite remarkable differences in the susceptibility of strains of yeast
of different ploidy to radiomynetic agents, particularly UV and X-rays, I would assume
without knowing it to be a fact that the same is likely to be true with other agents,

De you happen to know whether asaserine is active on yeast? If you tell me that it is
et all I will be glad to dig up some haploid and diploid yeast strains for you and you
may want to see for yourself if there is a differential in killing aetivity. One

point about almost all of these tests is that they are going neeessarily to invelve some
sort of semi-quantitative procedure in liquid medium because none of the effects are
going to be all or no inhibition or bactericidal effect such as you have been aceustomed
to seeing with ordinary antibiotic activity.

Let me know if any one ef these approsehes seems interesting enough to you that you want
to pursue it further and I will do my best to dig up the material. I can, I think, send
you stocks of B and B/r or very closely related derivatives of them. I haven't

retested these myself for their differential response to irradiation. You cen get, of
course, the basic data on them from Dr, Bwelyn Witkin at Cold Soring Harbor if you want.
T am not sure that hey haven't been working with this particular problem for quite some
time and I am really not at all sure whether they could heir you a great deal further

on their specific properties, One of these methods certainly should be suitable for what
you are ultimately aiming at here.

To turn to ancthr question, I certainly did make a bad guess as to what you intended to
use these coli strains for. I am a little surprised that you presented the poseibility
of biological protection to cover antibiotic @mporession of enterie formula simply as a
promotional stunt and I really wonder if this is what you think of it. It seeme to ne
that you are touching here on some of the most subtle.and nost important preblems of host-
parasite relationships, and I would certainly encourses you to go en in this direetion
and would lend every effort I could to help, vrovided that it was clearly understood that
there are some very serious problems involved here that are going to take a good deal of
experimental werk before they can have any hope of use. Against that should be measured,
I think, the great imcortanee that this approach would have. I don't know why you should
think that this is so absolutely silly. Inthe first place, behind this notion is a
considerable tradition of bacteriological therapy which had as one ef its poorer
expressions, for example, the fad of drinking Lactobacillus ris milk some while azo.
In the German literature of the 20's and 30'a there is «2 good deal to be said about
so-called "eoli therapy", the fundamental reasoning of whieh was quite similar to what
you have in mind as a cover for antibiotie suppression of other organisms although of
course the practical situation wes a little different. In facet, a couple of years ago

I talked to Dr. Finland at Massachusetts General Hospital about juet this possibility --
it's one that had ceourred to him toe. This was just at the time when he was first
describing these very serious Staphylococcal enteritides that follewed upon Aureomycia
administration and of course the first point that he would bring up was just how dangerous
an experiment it would be to introduce any new organiem into an unknown situation.

Deepite those reservations, they're ones which you have expressed yourself of course, it
seers to me that there is some vary important worthwhile work that needs to be done here
and that an important adjuvant to antibiotic therapy of the gut is going to be replacement

with suitable healthy so-called flora when they questiom however whether there is such
fo: J. Lein August 17, 1955 Wew Antibioties Sereening Program -
Consultantehip Arrangement

a thing as a healthy flera but it may not turn out that each individual will have his
own range of microfloral adaptatione. Thies is something that will require a great
deal of experiment. Frankly I don't think that your approach is geing to work; I
think that using a purineless mutant is not going to be the answer. If I am not
miataken it has been pretty well shown by the wofk of Burrows and Bacon that the reason
that purineless mutants of the Salmonellas are relatively non-pathogenic is precisely
peeause they failed to grow sufficiently in the habitat into which they are introduced
unless artificial growth faetor supplements are administered to the animals as well.
New if your're going to give purine along with purineless bacteria I think you might
as well give up that partiamlar bleck in the first place. I think too that net many
physicians are going to be convinced by a purely apriori argument that purineless
organisms under any circumstances are really likely to be lees virulent than the
normal. There are some very serious questions as to what one should consider as the
nevmal flora and perhaps as a firet approaeh in experimental animis I weuald use for
administration and prevention of overriding infections thet flera which does reourr

in organiame which heve been chronically treated with drugs like Aurtomyein, but which
animals have remained healthy. ! am not proposing giving Staphyloeocel, Candida,
Proteuses etc. The organisms I think should be repurified and studied to cet some sort
ef exvectation of their later behavior when they are reinoculated. $0 I don't think
that this is a silly idea, Joe, not at all, but I do think it's not something thet one
ean afford te go into lightly. [I have every hope thet you have the facilities to be
able to do a proper atudy of it since I don't think that anyone else is doing very much
along theese lines. Of course one place one would hone to have gotten information on
this queation would be from the germ—free animal studies of Renier's group at Notre
Dame. I have had oceasion to leek over that work again lately; the performance so far
hae been very little beyond the technicsl accomplishment of maintaining animals in
germ-free condition. They give hinte of experiments that they are going te do on rure
culture inoculating germfree animals but in fact very little is done about it. A
person who might be able te give you some information on these effeets would be
Mortimer Star (at the Department of Bacteriology at the University of California)
Davis - he was interested although he was primarily a plant vcathogenie bacteriologist.
He was interested for awhile in the growth estimating effeats of antibiotics and felt
sure that their modification of intestinal flora waa an important part of it. At any
rate he did notiee that after the administration of streptomyein to turkeys there was
an eventual reourrence of streptomycin-resistant coliforms and that when this finally
happened his growth-stimulating effeetsa ceased. I think he may have had a note or two
about this in the Bactearielogical Proceedings but I'm not sure.

You may be interested in a case that happened here in our department. fhe baby girl
of one of our assistant professors came home from the hospital with what eventually
proved to be a pure gulture of Serratia marcescens in her gut. No other organism
could be isolated at first for «a peried of about two weeke. After she began to ze on
a diet of non-sterile light foods she eventually picked up some coliform as well.
She is still excreting this Serratia. I'll be glad te send you a eulture of this
organism if it would have some interest for you as an organiem that could readily be
followed in artificial infection experiments -long these lines. There is at least
the justification that in one instance it had no pathogenic effects whatsoever. I
think you may find some resistance among pediatricians to the administration of a
chromogenic organiam. I know that this particular family was alarmed because of the
red diapers for a period of some time, This question is one that I'm really quite
interested in myself, Joe, and wish there were an opportunity to do mere research on.
To: J. Lein. August 17, 1955 New Antibiotics Screening Program ~
Consultantship Arrangement

There ought to be better facilities for germfree researeh and maybe something
eventually is going to be done about it at institutions other than Notre Dama.

That about ties up these main points. The only thing that I want to add now, Joe,

is a note that you may or may not have seen in an issue of WATURE, July 16, 1955,

Pe 121, There is = paver by Simonti and Simonti on genetic recombination in Strevtonyces.
They worked with a strain of Streptomyces coelicolor, and they have about the same
category of evidence as was represented in the firet paper in NATURE by Tatum and myself
on FE, goli. Obviously as a result of thie Gallen, Bradley and I are going to have to
reevaluate our own position in thie work and we're going to have to diseuss this
somewhat with you. The main thing that I would have been concerned abovt in freely
discuseing our work wag that I did mot want to get any information from you which had

to be held confidential and I did not want to have a purely one-sided discussion on
work of this kind. I will just leave it at that, Joe, and if you can see any reasonable
common grounds that we can meet on to talk about the genetics of Actinomycetes I'1) be
glad to reopen the question. Hewever, as long as we were direotly occupied with this
work ourselves it would do us, I think, a good deal of harm to be influenced by
information which was not publialy available. Esther and I are taking about 10 days

off on a trip to Colorade and for that reason there may be some delay in your receiving
some of the eultures that I promised although I left instructions behind for then. I
will be back just before the end of the month.

I've just learned that we are going to have our laboratory remodeled sometine during

the fall semester and if that is the case there may be a very suitable cecasion for me

to take off and pay you ancther visit in Syracuse some time in October or November or
early December if that would suit your plans as well. Let's leave that open for the time
being but I thought I would just mention it to you. Probably by the time this has been
assimilated we will have gotten back from Madison and I will weit to hear from you

again.

Yours sincerely,

Joshua Lederberg
/J1e