Transcript from filmed lecture
November 15, 1959

Lecture M5 - Bacterial Genetics: Genetic Transduction
by J. Lederherg

In the previous lecture we discussed some aspects of the sexual mechanisms
in Escherichia coli. Although necessarily any extrapolation of the concept of semuality
to simpler microorganisms is bound to be somewhat arbitrary, we must admit that in
its main features the process of genetic recombination that was described involving the
mediation of the F factor does bear at least reminiscent resemblances to the fertiliza-
tion processes in higher organisms. The main feature that we would stress here is
the participation of intact cells as the agents which come together in the process of
recombination and also the integrity of all, or almost all, of an entire genome as the
unit of transfer. |

Only 2 few bacteria have been extensively studied for the detection of recombina-
tion processes, and, so far, Escherichia coli and its relatives among the enteric bacteria
and some filamentous bacteria, the actinomycetes, are the ouly ones in which we have
strong evidence for the existence of sexual processes. This is however more testimony
of the limited extent to which bacteria have been subjected to experimental analysis than
to the likely restriction of interesting genetic phenomena to a few forms. Further studies
have been made, some of them in an attempt to widen the range of organisms in which
sexuality could be demonstrated, and, in the course of some of these, other phenomena
have been described which are still agpects of genetic recombination, but owing to
their greater simplicity, have to be given a different category. We are concerned here
now with processes of fragmentary genetic exchange which are subsumed umdier the head-

ing of "Genetic Transductions".
One of the systems which has given rise to a large volume of experimentation
is one in which bacterlophage is intimately involved, and which concerns a reasonably
close relative of Escherichia coli, the species Salmonella typhimurium. This is the
mouse typhoid organism, and is an important agent in food poisoning disease in man.
The group Salmonella ie moderately closely related to Escherichia coli in its general
features, which is one characteristic that recommends the group for experimental physio-
logical and genetic study, since, like Escherichia coli, these bacteria can be grown in
simple media, and are subject to many of the same physiological tests which have already
been developed for E. coli.

In the course of attempts to determine whether or not strains of Salmonella
typhimurium were subject to sexual recombination along the lines described previously,
a rather erratic pattern of genetic exchange was found. It was discovered by Zinder
and myself that certain combinations of cultures when mixed together in nutrient medium
would generate new recombinant forms, somewhat along the lines previously discussed.
However, this was true only of certain combinations of strains, and it was then eventually
discovered that in contrast to the necessary participation of entire cells, which charac-
terizes the sexual process in E. coli, that, in Salmonella, from mixed cultures it was
possible to obtain cell-free filtrates that had the same activity. That is to say, froma
mixture of strains, one of them by analogy with our previous demonstration, M* T-,
and another M~ T+, it was possible to obtain what we call a filtrable agent that had
mM activity. That is to say, this cell-free filtrate, from which the cells of the original
culture had been freed by sedimenting them in the centrifuge and then filtering the cul-
ture, and was also subject to such additional precautions as heating the filtrates until

no bacteria could have survived, these cell-free preparations, nevertheless, had in
~3-

them a factor which was capable of reacting with a euiteble indicator strain to transfer
the Mt activity. And as was given in the previous design, this is demonstrated experi-
mentally by plating a mixture of the filtrate, which of course has no viable elements in
it, and of the viable cells which are methionine-less, which would be unable to grow in
minimal medium, and from that mixtare obtaining a yield of these recombinant prototrophs.

The chief feature that distinguishes this systema from sexuality ie the participa-
tion of an element, which owing to its capacity to passa through filters whick will retain
bacteria, must be smaller than the bacteria in size. And it then became necessary to
spend some effort to determine the nature of this filtrable agent. It was rather peculiar
that attempts to isolate this agent from cultures of Mt T~ alone were unsuccessful, al-
though this culture must have been involved in donating the necessary genetic element,
the M*+ factor, to the filtrate. It was eventually found that the filtrable activity itself
could be used to propagate additional M* activity. That is to say, a drop of a filtrate
once made from this mixture could be added to fresh M* cells and give rise, again, to
a new erop with great increase of M* activity. And this could be passed in series.

So it is now apparent that in fact we have two agents present, or at least one
agent with two manifestations. First, that agent which was capable of transferring
the M* activity, and second the agent that was capable of evoking the first agent from
the M* bacteria. Eventually it was discovered that this second agent, the provocative
agent, was a hacteriophage, and this bacteriophage was one for which the second strain
was lysogenic. That is to say, cells of the M~ strain were carrying a bacteriophage
which we have. since labeled P22, phage mmober 22, and phage 22 when grown on an
Mr strain yields particles that have the capacity to transmit M” activity to a suitable

indicator. Now it turns out that in our original system we used as an indicator the same
strain from which the P22 had been derived. However, once it was recognized that
phage 22 was the important element in provoking the transfer of genetic activity, the
routine procedure for conducting these experiments was simply to add one or a few
particles of a stock suspension of P22, add those to any of a variety of genetic types,
and then from those filtrates to obtain new preparations that would have a genetic ac-
tivity characteristic of the type of bacteria on which that phage had grown.

It would then be possible to use those filtrates having these activities on a
variety of different indicator strains, including this original M™ strain, but also in-
cluding any of a large number of other Salmonella types. And in fact, it was subse-
quently shown to be possible to transfer genetic factors almost universally to even other
species in the Salmoneila group.

Well, what does a bacteriphage now do to the cells of a suitable susceptible
strain in order to provoke the development of this activity? The answer to this ques-
tion depended on further studies of the factors that were transmitting the genes them-
selves. That is to say, although at first we knew that it required the bacteriophage to
evoke the genetic activity in these filtrates, we did not until further studies, know the
nature of the agent that was present in those filtrates. However, the simplest hypothe-
sis that one could formulate was that the bacteriophage itself had acquired the capacity
to transmit genetic factors from one strain to another. That is to say, we could make
as a preliminary proposal that the bacteriophage had acquired fragments of genetic
material from the host on which it had been grown. This would imply that in the filtrates
thet had genetic activity we should find that activity associated with the bacteriophage

liself.
This association has been verified by means of a number of tests. First the
agent shows the same quantitative susceptibility to inactivation by heat as does the
bacteriophage. That temperature and time treatment which will inactivate 90% of the
bacteriophage particles as measured by their activity as a virus, also inactivates 90%
of the transductional, that ia to say, the genetic transferring capacity of the bacterio-
phage suspension. Second, the genetic particles and the virus particles have the same
susceptibility to specific antiserum, since one can obtain, by injecting the bacterio-
phage into rabbits, antibodies which specifically block the capacity of the bacteriophage
to adsorb and attach itself to susceptible cells. Thirdly, the attachment to susceptible
cells occurs at the same rate for the genetic activity and for the virus activity of these
suspensions. This ip measured by adding known aliquots of the filtrates to an excess
of bacteria, allowing this mixture to remain together for a period of time, and then re-
moving the bacteria together with that fraction of the virus which had adsorbed to the
bacteria. Then we would conclude from filtration experiments and from sedimentation
experiments that the eize and mass of the genetic activity was approximately the same
as the bacteriophage aotivity. |

From these measures we could conclude, with a fair degree of certainty, that
the transductional activity in a filtrate was indeed associated with particles which had
essentially the same exterior properties in every respect, chemical and physical,
as the bacteriophage itself. However, owing to the very low frequency with which any
given phage particle is likely to include a particular marker, we could not be abso-
lutely sure from these experiments that the same particle which included a genetic
fragment also included a virus genome. in fact, there is a strong suspicion, which has
been much better substantiated in some other systems of transduction, that this is not
the case~ that is to say, that those particles which are carrying a genetic fragment

of the bacterium are at least defective in respect to the virus which is included in the
same particle, so that to some extent we should consider that there has been an acci-
dental replacement of a bacterial fragment in place of the virus genome which should
occupy the interior of that phage particle.

We may now recapitulate the way in which a transduction experiment is done
in Salmonella. We must first start with a suspension of bacteria which already contain
those genes in which we are interested. And it is an important experimental fact,for
the interpretation to which we give this phenomenon, that the range of activity displayed
by a given bacteriophage filtrate is precisely the same as the genetic factors carried by
strain on which the bacteriophage is grown, and independent of the ultimate origin of
the bacteriophage. This can be shown by first growing a crop of phage 22 on a bac-
terium that might have the genotype for example M* T* X+ Y- Z-. This crop of bac-
teriophage when tested on suitable indicators one at a time, these indicators being
strains that lack M, lack T, lack Xt, lack Y*, lack Z+, on these indicators one can
show that that filtrate will have activity for the factors which were present in the donor,
in the host, and not for those factors which were absent. If this crop of phage, which I
might call the first crop, is then put onto a new bacterial strain, let's say one which is
M+ T- X+ Y* =~, we then find that that crop of phage hrs a new genetic activity, namely
it has now lost the activity for the T factor and has gained the activity for the Y factor,
precisely in correspondence with the new genetic factors presant in those bacteria used
as the host. The bacteriophage is therefore an essentially passive element in respect to

the genetic content of the fragments that it carries.
Having obtained this crop of bacteriophage from the bacterial culture that we
wish to use as donor in a recombination experiment, we then remove the excess cells
which may have survived the initial attack of the phage. And this is done most simply
by a quick sedimentation in the centrifuge, and then heating the supernates which are
almost free of bacteria to a temperature of 60° for 20 or 30 minutes, which will kill
rapidly any of the rematning bacteria in the supernate and leave the phage intact. These
supernates containing a fairly high titre of bacteriophage are then simply mixed with the
susceptible indicator bacteria to which one wishes to demonstrate the transfer of the
genetic markers in question. After a short period of incubation the bacteriophage
particles attach to these recipient bacteria, a3 we will discuss in a later lecture, the
DNA content of the bacteriophage is injected into the recipient bacterium, and, then
provided the bacteria happens to survive these particular attacks, they then have a
certain definite probability, in the present case however a rather low one, of acquiring
one of the specific markers of the strains used as a donor. This is demonstrated, as
was true in the original teat experiments, by almply plating the treated bacteria on a
selected medium which will suppress all of the original genotypes and permit the ex-
pression of bacteria which have been altered in respect to the gene that we are studying.
For example, we study the transmission of the M* factor, by mixing phage grown on an
M” strain with indicator bacteria of the M* type, and then plating that mixture on medium
deficient in methionine which will suppress all of the M minuses and allow only the M* _
to come up. Now we can symbolize this general operation, since we wish to distinguish
the trensductive phenomenon from those of sexuality, in the following way. We simply
write M*, instead of cross or X, we write dash X M™, and we will in this case have

demonstrated getting ont an M*, and, since the vector in this case is phage P22, we may
write the reaction in this way.

We now have to consider the scope of genetic exchange, since although we have
now demonstrated that the material element of transfer is something much less than the
intact cell, in fact is incorporated in an element whose DNA content is about one one
hundredth of that of the bacterium, we still have not yet said anything concerning the
genetic acon of the activity in each phage particle. This can be done by pursuit of
experiments similar to the one just described, in which we examine the coincidence of
transfer of different markers. And for example, if we conduct an experiment and, now
using the abbreviated notation, M* T+ X* where these are three factors, for examples,
methionine, threonine independence, and xylose fermentation, prepare the phage fil-
trate, put the filtrate on an indicator strain, which will be lacking the competence for
each of the three functions indicated here, and then plating these mixtures on a variety
of medium, first a medium that selects for M*, then a medium that selects for T+, and
then a medium that selects for Xt. And then we examine those progeny which are Mt,
after having purified them by picking the colonies that developed on the selected medium,
and then restreaking them a few times to reestablish definitely pure clones, and we type
them in respect te the other markers. And the typical result of such an experiment is
that those bacteria that have received the M* factor and are now manifesting it are still

_of the original recipient type in respect to the other two markers. And similarly, those
that are T* are still M™ and X~, and those that are X* are still M- and T~.

From these results we conclude that, in general, each phage particle carries
but a single marker from the donor, whose markers it is transducing. This is in con-
trast to the genetic consequences of sexual recombination, where we find large blocks
of genes being transmitted in community from the Hfr cell into the F~ cell. However,

more detailed studies have uncovered a number of specific examples where markers are
~-

transduced together. From these results we conclude that there is a possibility of
linked traneduction and from. a comparison of tranghidotion by a bacteriophage Pi,
which is active in Escherichia coli, with the results of sexual recombination in
Escherichia coli, we conclude that the average genetic content of each bacteriophage
particle is approximately 1% of the total genotype of the donor strain. This is at least
roughly concordant with the proportion of DNA which can be contained in a phage
particle, with the DNA in a bacterial nucleus. As it happens, the linkage that ean
be demonstrated by co~transduction gives us a very powerful tool for studying the
close linkage of biochemical markers which are physiologically related. Demerec and
his co-workers at Cold Spring Harbor have extensively studied the relationships between
markers involved in related steps of a biosynthetic sequence, for example, the synthe-
ais of tryptophan which proceeds via anthranilic acid and indol as precursors of trypto-
phan. They have found that not only are those distinct mutants which are involved in
various ways in introducing defects into the formation of one particular enzyme closely
linked, in a cluster at a particular point on the bacterial chromosome, but also those
mutants which are involved in the sequential steps in the biosynthesis of tryptophan are
also closely inked to one another. And the same pattern has been found with respect
to the biosynthetic patterns of histidine as well as of other substances. In fact, we may
take it that as a general rule, to which there are however some exceptions, there will
be this correspondence between biosynthetic and genetic association, to the point where
we oan assign a definite segment of the chromosome to a cluster of genes concerned with
related functions.

Since this situation appears not to pertain to some higher organien»and does

not apply in general for example even to Neurospora, we are not in position to assess the
-10-

exact significance of this correlation. However, it has been suggested that the adap-
tive value of this linkage of physiologically related genes must be in the regulatory
functions which should function so as to turn an entire set of enzymes on and off in
accord with the physiological requirements of the organism in different situations.

The study of transduction by bacteriophage has been greatly helped by the
appearance of another type of transduction in which the frequency with which markers
are carried is several orders of magnitude higher than it is in Salmonella. We find
that the villain of the piece occurs again; we are dealing here once more with the
bacteriophage, lambda. Now you may recall that we had evidence from bacterial
crosses in which the markers lysogenicity and Gel were in coupling and both segre-
gating, that these markers are closely linked on the bacterial chromosome. A
second testimonial to this linkage is the fact that when bacteriophage lambda is har-
vested by the ultraviolet induction of the prophage in a galactose-~positive lysogenic
cell, that that lambda is found to have transducing activity for the Gal genes to which
it is linked. Unlike the Salmonella situation where we find that any marker of the en-
tire genome is susceptible of transudction by phage P22, so iar we have not found any
markers other than Gal, the ones which are clogely linked to the prophage, to be cap~
able of being transduced by this particular element.

The rate of transduction by lambde, originally grown from typical haploid
lysogenic cells, is again quite amall, and of the order of one per hundred thousand
of the phage particles per marker. However, the progeny of the transduction from
galactose-positive lysogenic to a galactose-negative strain had been found to exhibit
an unusual structure, which is similar to heterozygosis but which involves only the
Gal and perhaps the lambda segment. Because of the restricted scope of this dupli-

cation of genetic material we have called these unusual types heterogenotes. And the
~11-

simplest way to display the genetic behavior of these heterogenotes is to suppose that
in addition to the unchanged chromosome of the recipient type,which we might indicate
as having a Gal~ marker and some kind of prophage site, that there is an additional frag-
ment which remains outside the chromosome although perhaps attached to it by some stil!
not very well specified mode of attachment, and this fragment has the gene that had been
carried over in the lambda segment. Now we know that the heterogenotes are capable of
undergoing further changes equivalent to reduction as we see it in diploid forms, which
permit the exchange of the galactose factor into the galactose mimus form.

Of some importance is the fact that the lambda isolated from a heterogenote has
a very high activity, to the point where nearly every lambda particle will transmit Gal
activity. This is indicated by a simple experiment which I've diagrammed here, if
one cross-brushes a suspension of lambda, (with) containing a relatively small number
of particles, that lambda having come from a galactose positive heterogenote, against a
suspension of galactose-negative bacteria, one can almost count the lambda particles by
the number of galactose-positive colonies that appear in the zone of intersection. This
is displayed further in the accompanying alide.

We have accordingly discussed the major mechaniams of genetic recombination
as displayed in bacteric: heterozygosis and the sexual mechanisms of E. coli K12, a
more complex form which is preceded by the concurrence of melei in common cytoplasm
in filamentous fungi, previous lectures have discussed the direct role of DNA in the pneumo-
coccus transformation, and we have just referred in the present lecture to the role of

bacteriophege as a vector.