Transcript from filmed lecture
November 14, 1959

Lecture #43 - Bacterial Genetics: Clones
by J. Lederberg

Bacteria have become important subjects of experimental investigation for
several reasons. First, as you ali know, is their importance in the ecology of the
terrest_} rial globe, their impact on agriculture, and their relationship to human disease.
And this motivation has indeed underlain most work in bacteriology during the first 50 or
60 years of its existence. | Second, bacteria proved to be extremely convenient objects
of experimental investigation since very large populations can be manipulated with great
ease in the laboratory. And third, the simplicity of their structure, which can be illus-
trated by this schematic diagram of a cell of Escherichia coli, which measures about
1 by 4, making this among the smallest of all cells, and which contains within it a
minimum of structural detail. We see the presence within each cell of two or sometimes
four nuclei, which can be chemically defined as masses of deoxyribonucleic acid, con-
taining approximately 6 million nucleotide units. This should be compared with the con-
tent of DNA in a mouse cell which is nearly one thousand times greater or 5 billion
nucleotide units.

The slide which is now being projected will show you somewhat more realisti-
cally the appearance of a bacterial culture which has been stained so as to exhibit the
miclear material. We would like to know more than we do of the detailed morphological
mechanism by which the nuclear material is divided at each cell division, since there is
at present time a continuing controversy as to the existence of a process displaying all

it

of the features of mitosis as we see/An higher organisms. Still, there can be no doubt

whatsoever as to the fundamental chemical similarity embodied in DNA between the
genetic material of bacteria and that of higher forms, and we likewise can have no
doubt that there is a mechanism for the exact partition of the 5 million units or 6
million units of information which is contained in the DNA at each division of the bac-
terial ceil.

An important feature of experimental treatment of bacteria is the reliance on
methods of vegetative reproduction, since this, in all organisras, including those which
have an auxiliary sexual mechanism, is by far the most important means for the in-
crease of bacterial numbers. We, therefore, will recur again and again to the con-
cept of the clone. A clone is defined as a population of individuals all derived from a
single cell by vegetative reproduction. In consequence, barring some accident of
genetic mutation, or some interesting event of genetic recombination with other cells,
the clone should consist of a population of genetically equivalent individuals all of whom
have identical copies of the same information which was present in the ancestor. Be-
fore dealing with the practical exploitation of clonal growth, we should consider the
rapid rate at which bacteria multiply, and the large populations which can be obtained
from them.

in a favorable medium a typical bacterium, like Escherichia coli, can divide
vegetatively at intervals of about one half an hour. Consequently, if one plants into a
tube of suitable medium, into nutrient broth, a single cell, we will have a geometrical
increase in the number of cells along the lines indicatea by thia diagram. This doubl-
ing every half hour results in an accelerated increase in the population which can be
defined by the expression that at any given time t, measured in hours, at which there
wil: have been twice as many generations, small n, the population, large N, is given

hy the power 2 to the n, or 2 to the at. Consequently, afier a period of 30 bacterial
~3-

generations, which would require just 15 hours, a tube of clear broth into which a
single cell has been planted, will give rise to a population of 10 hillion organisms,
as indicated in this tube of turbid broth.

Now, in order to obtain single cells from which clones can be derived, which
clones are then available for further biochemical and physiological and of course
genetic study, several procedures cau be used. The most tedious, although the most
exact, is perhaps the direct isolation of a single cell by means of a micromanipulator.
But this is a fairly slow and laborious process, and in practice is used only when apecial
emphasis must be placed on the fact of the derivation of the entire clone from a single
individual which had been seen under the microscope.

In practice raany short cute are available, And in general they are based on
the fact that if one takes a suspension of bacteria in which there are large numbers of
celis one can sufficiently dilute those suspensions by diluting samples into larger volumes
of fluid, so that we will have a sample containing a few bacteria, which can then be spread
on an agar plate. ihave an example,of such an agar plate that had been spread 24 hours
age in my hand, and these are displayed further on this light box, in this example here.
We have here about 290 colonies of bacteria, each of which has been derived from a
single celi that had been planted on the surface of the nutrient agar medium, which fills
this plate. As a result of the continued vegetative multiplication of each cell, the in-
visible cell, which had been randomly put at a particular spot on the agar, now gives
rise during the course of 15, 20, or 25 generations te a visible colony. This colony
ig aclone, and, as I indicated before, unless there has been some interesting event
to intervene, it will consist of individuals all having the same genetic constitution.

Another way in which one can conveniently separate cells in order to produce
-4-

clones is to use the well known, and very complicated instrument, the innoculating loop.
One merely takes a sample from the broth culture, I'm omitting the sterile precautions
which are necessary in practice, but a little hard to do at the lecture table, and simply
spreads out the content of the loop onto a fresh agar plate, and then eventually there

will be a point on the plate where single ceils are deposited on the agar at some dis~
tance from one ancther. This gives rise to a plate that looks like this, and I've used

the loop to streak out two different samples of different kinds of bacteria, and those

are indicated by the different colors of the colonies on this half of the plate, on one hand,
and the other half of the plate on the other.

This plate, that I've just shown, a'so illustrates how we must in practice rely
on the average behavior of a large number of organisms for most of our genetic work.
There are very few markers which can be accurately discerned in individuals cells.

Now there are some kinds of experiments where one will rely on a cell marker, for
example motility, which is based on the presence of flagella -- the organs of locomotion,
but as a rule there simply is not enough material in a single cell to give us a convenient
way of studying its phenotype, and we therefore rely on the fact of the genetic homo--
geneity of a clone, and the ability to recognize the physiological and biochemical b2-
havior of a large number of identical organisms as a way of typing the original bacterium
from which the clone had derived.

This plate, in fact, illustrates a physiological difference between a wild-type
and a mutant strain of Escherchia coli which has been of some practical importance in
the development of this subject. The medium on which these bacteria are plated is an
eosin methylene blue agar medium containing the sugar lactose. And the dark colonies

that appear at the top half of this plate represent organisms that are capable of fer-
menting lactose, and thereby generating the color that you see in the colony as a result
of the acid produced from the fermentation. The bacteria that have been streaked on
the lower part of this plate are a lactose-negative mutant, that had been derived from
the original wild-type culture by exposing their cells to ultraviolet radiation. The
comparison between the wild-type and the mutant strain has been carried to some de-
gree of precision, and has in fact given us some insight into the genetic control of the
forrnation of this enzyme responsible for lactose fermentation, beta galactosidase.

Before we touch further on physiological mutants of bacteria, it would be im~-
portant to verify the fact that these are spontaneous, or ultraviolet, or X-ray induced
reutations, in the same sense as applied to mutations of higher organisms. Since bac-
teria are so intimately exposed to their immediate chemical environment, one might
inquire whether in these instances, for example, the production of lactose-positive re-
verse mutants from the lactose-negative strain, or in other cases, where one could in-
sert some question of the inductive effect of the environment on the bacteria, whether
these chemical environments have resulted in a direct way in the modification of the
hereditary constitution of the cells.

This question which may appear remote in the example I have just indicated,
has been a very pressing one with respect to mutations of another form, rautations to re-
sistance to antiblotics. It has been known, almost since the introduction of these agents
in chemotherapy, that populations of bacteria which had been exposed to an antibiotic,
for example t.b. organisms expoged to streptomycin during the course of therapy of
tuberculosis, might on some occasions give rise to progeny which were resistant to the
drug in question. At first sight the mechanism would appear to be quite simple, and has

been so proposed by many workers. Namely, that some chemical reaction had taken place,
=f =

as between the streptomycin and the tuberc_le organisms to result in a change in heredi-
tary constitution of these bacteria, and, consequently, the development of drug resis-
tance. And this might seem to be the aimplest statement of what, in gross, happens
when one conducts this experiment cither in the treated patient, or in the laboratory,
when one adds a given amount of antibiotic to a large culture of bacteria.

However, the geneticist, accustomed as he is to the recognition of changes in
populations on the basis of natural selection of spontaneously occurring mutants, has
persistently raised an alternative hypothesis, namely that mutations to drug resistance,
or wutations of other adaptive types, have in fact no relationship to the environment in
which they occur, except that,once having occurred,these environments select those or-
ganisms that now have an adaptive value therein.

It has been a matter of some difficulty to establish in any precise way a criterion
for deciding between these two interpretations since in most instances it would be neces-
sary, in order to demonstrate the presence of one drug-resistant organiam among 4
population of 10 billions of cells, to expose that population to the drug in question. How
then can we determine whether the mutants were present in the population before the drug
was added to it, that is to say, whether the mutants were pre-adaptive, or whether they
had arisen as a result of a direct chemical interaction with the chemical -- whether
they are post~adaptive ?

One of the first approaches to this problem, that was at ali convincing in its
implications, was presented by Luria end Delbruck in 1943, the basis of their now well
known fluctuation test. If we consider the clonal nature of bacterial growth we must
keep in mind that when a mutation should occur during the course of development of

a clone that the number of mutant cells which will be present in the final population will
depend not only on the mutation rate alone, on the frequency with which mutational events
occur, but also on the extent to which individual mutants will have had an opportunity to
multiply during the further growth of the clone. For example, a mutation which occurs
late in the development of the clone, let us say at the last division before we examine

it, will be represented by one mutant individual per mutational event. Much more in-
frequently, hecause there will have been many fewer cells at that time, we might have
some mutations occurring say at the first or the second division. And these mutations
will now have a very large mutant progeny, since the mutant character will be propa-
gated to ali of the descendants of those few celle that were present at that time. Conse-
quently, the statistics of distribution of mutant cells in bacterial populations would be
expected to be highly skewed, that ia to say, there should be a very few cultures which
have a very large nuraber of mutants far in excess of the expectation from any normal
distribution of the mutant individuals.

On the other hand, if the mutation were the resuit of an interaction of the final
celis in the clone with the chemical used to demonstrate the mutants, we might then
expect something much more nearly approaching a random or normal distribution.

The experimental finding which Luria and Delbruck found, and proposed as
evidence that the mutations for resistance to viruses in bacteria and to drugs was based
on spontaneous mutation, was that there was this highly skewed distribution shat there
were, so to speak, afew, and an unexpectedly large number, however, jackpots of
cultures in which a very much larger number of mutants was present, than would have
been predicted from the random occurrence of mutants in the final generations.

However, this argument is essentially a statistical one, and rigorous as it

may be, it has not been understood as fully as might arguments of a more direct nature.
One of these is based on a new technique for handling the large numbers of clones which
are necessary in analyzing problems of this kind. You will have recognized that we
might have made a direct test of the pre-adaptive occurrence of resistant mutants in a
very simple but tedious way. We might make a plating of a clone, so that there would
be at our disposal some hundreds of thousands, or millions, or even billions if necessary,
of isolated colonies. And we might then test those colonies in a way that I've indicated on
this plate on the light box; pamely, each colony might be put on a medium which lacks the
drug, and then a sample of it stroked over to a line in which a certain amount of the drug
had been added -- in this case the actual demonstration concerns streptomycin and cer~
tain mutants of E. coli which are resistant to streptomycin. If a sufficiently large num-
ber of randomly plated bacteria were used, and their clones tested against streptomycin,
if these mutations can occur in the absence of the drug, one might in fact expect, without
having subjected the bacteria to any prior exposure to the drug, that an occasional colony
would be resistant. However, we now know that in order to do this experiment it would
necessitate of the order of one hundred million, or one billion tests, and this ie far be-
yom! the range of any practical experiment that can be done by such a random procedure.
Fortunately, we have now at our disposal another method which makes it pos-
sible to handle large numbers of clones efficiently for this purpose. This is called replica
plating, and it relies on the possibility of transferring the growth from one plate to an-
other by meane of a sheet of velvet which is placed on a hoop, as I have just done. If
we start with an agar plate which has already had distributed on it a collection of colon-
ies, whose types we are interested in determining, we can press the plate down on the
velvet, leaving on the velvet surface an impression of the growth that was on the plate,
and then use the velvet as a master to pliant a corresponding pattern of grows on one

or a aeries of additional plates, as I am now doing. I have on the light box an example
of this operation that was done about 24 hours ago, which indicates the identical pat-
terns of growth which one can obtain, having started from this master, and here are
two of the copies, and the similar distribution of colonies as between the two plates
shows the extent to which one can obtain a faithful replica of the original pattern on one
or several secondary plates,

The application to the present problem is the possibility of replica-plating a
pattern growth from one master fo a plate not containing the drug in question, to the
standard medium, and then to additional plates which do contain the drug. On the plates
containing say streptomycin, only those bacteria will be able to form colonies which are
resistant to the drug. Consequently, by this method in a single procedure it is possible
to teat not just one colony for resistance to an antibiotic, but to test as many colonies
as can be crowded onto a single plate. And this means that say a thousand clones can
be tested in one operation.

In fact, even this is not sufficient for an efficient test of the pre-adaptation
hypothesis. Since the frequency of occurrence of drug-resistant mutants in many cases
is very much lower even than one in a thousand, and it may be one in a million or one
in 2 billion.

Even sc we can then use replica plating by starting with an original plate on
which had been smeared, say one million or one billion organisms, which will give
rise to small colonies go closely spaced together that there appears to be continuous
growth. In that plate, however, there will be imbedded a few small clones, not dis-
cernible to the eye, ‘which should, on the pre-adaptation hypothesis, be resistant to
the drug. The position of these clones will be registered when replica plates are made

from the original to subsidiary plates containing the drug. It is then possible to go back
~10~

to just the corresponding site on the original place corresponding to the resistant
colony and from that site to obtain an innoculum which will be greatly enriched in the
proportion of resistant as“: compared to sensitive cells. Because the colonies are
so closely crowded together, the first return to the original plate will not give a pure
culture of resistant organisms. However, it will have a higher proportion of them
since one can go to the place where the resistant organisms are found, and ignore
parts of the original plate which do not have these resistant mutants.

Using this as an innoculum it is possible to either repeat this enrichment
procedure, or finally to do a plating at such a dilution that there will be discrete colon-
ies and from such discrete colonies to detect the one which is resistant to the drug by
making a replica plate. The point that should be stressed is that in these operations
of transferring colonies from a master plate to the velvet, and then from the velvet
to a subsidiary plate, when we eave the master plate we do not expose any of its cells
directly to streptomycin. We have taken only a sample of the cells from each colony
in order to test the status of that colony or that site on the agar. Consequently, when,
as has proven to be the case, it is experimentally possibile in many instances to isolate
drug resistant clones of bacteria by the repetitive use of the replica plating technique,
we can be assured that the drug, say streptomycin, or penicillin or chloramphenicol,
has not played a direct part in the production of the resistant mutants. These resis-
tant mutants although present in very small numbers must have been present in the
original clones before they were exposed to the drug.

Since we can so easily develop clones containing the 10 billions of organisms
which I was indicatipg previously, and since it is such a simple operation to take the

entire contents of such a culture and to plate into agar containing a drug like strepto-
-11-

raycin, we consequently have a very powerful method indeed for detecting the presence
of even very small numbers of mutant cells in very large populations. We can, by this
method, measure the smallest spontaneous mutation rate thet has yet been possible to
be measured, and this is in fact the routation of Escherichia coli from its norma! status
of sensitivity to streptomycin to the status of streptomycin-resistance. This is a
change which occurs approximately just once for one billion cell divisions. And we
can visualize the difficulties that we would encounter if we set ourselves to the task of
detecting, much less than enumerating the frequency, of this kind of mutant in popula-
tions of any other organism that we can use in the laboratory.

Having thus shown the spontaneous character of those mutants which arise and
can be selected by the presence of deleterious agents, or similarly the spontaneous
character of mutations that lead to improved adaptive quality of the organism in other
stressful environments, for example the placing of nutritional mutants in media which
lack the specific nutrients required by a given strain, we are then in a position to use
this technique as a routine method for the specific isolation of those genotypes which
may occur in 4 bacterial population and whose occurrence may give us important infor-
mation as to genetic processes taking place in those populations.

In the next lecture we are going to discuss the application of such techniques
for the discovery of genetic recombination. For the moment I think we cat merely
point to the fact that we now have available a method of precisely counting the incidence
of mutants which may arise in bacterial cultures that had been exposed to specific treat-
ments, particularly designed to induce the occurrence of various mutations. Histori-
cally, the most important of these is of course X-irradiation. And just as it does with

every other organism we know, X-rays will cause an increased frequency of the pro-
~12~

duction of various kinds of mutants in bacterial populations. I would like very much to
stress the various kinds of mutants, since neither with X-rays nor in fact with any other
agent now available to us have we succeeded in producing any single unique type of mu-
tant at will. This is presumably ultimately a reflection of the chemical structure of the
hereditary material as represented in DNA. The fact that DNA has relatively simple
units, the nucleotides of adenine, guanine, cytosine, and thymine, should tell us that
we will face great difficulties in trying to obtain an agent which will specifically induce
the mutation of any particular gene. Any gene in which we might be interested is cer-
tain to be composed of large numbers of these elementary units, differing from other
genes only in their arrangement. In order fo obtain a specific mutagen we should then
require an agent not which merely separates its actions on thymine as compared to ade-
nine, or on guanine as compared to cytosine, but which is able to recognize specific
assemblages of these elementary nucleotides into larger units.

Such an agent, in other words, would have to have all of the necessary chemi-
cal discrimination which we now attribute to a gene. And we may reasonably expect
that such an agent will have the same chemical complexity, and perhaps even structure,
that the gene itself displays. This order of complexity is beyond our present powers
of chemical synthesis, although it may perhaps eventually be within our grasp. Al-
though specific routagenesis is still for future investigation, this does not mean that
we conopletely lack important information on the chemical environment concerned with
mutation. For example, Novick and Szilard have found that purines, for example,
caffeine, adenine, guanine, will greatly increase the mutation rate as observed in bac-

teria, and furthermore that the corresponding riboside derivatives will greatly de-
-13-
crease the mutation rates. These observations, and many others, have led us to the

conclusion that spontaneous mutation, sporadic though it be with respect to the identity

of ite targets, is in many respects an incident of the normal metabolism of the cell.

Thank you.