Further information on Research Plan:
Studies on the Mechanism of Transformation in B. subtilis

This material is submitted in response to a request to furnish
the finer brushstrokes that will paint in the details
we. SO that the [study section] may find themselves better able to

evaluate the finalized picture."

In ny covering letter, I offered some of the reasons that make
this a difficult task, if the research plan is expected to be a
reliable chart of future developments. Attachment (1) illustrates
the role of unexpected findings in changing the detailed plan on a
short time scale, and knowing that this is the main fruit of our
effort makes the construction of very elaborate plans seem an
exercise in futility. In addition, we face the special problems of
satisfying foreseeable regulatory restrictions that complicate
efforts to examine some of the most telling scientific issues.

The field of genetic transformation also has a long history.
Fortunately, Microbiology-1976 has appeared since the original
proposal was submitted. It includes extensive reviews by Frank
Young, David Dubnau and others, that cover most of the issues of
concern for the present status of the mechanism of transformation.
I hope it is legitimate for me to refer to that source rather than
repeat a great deal of detail that is already in print.

Dubnau summarizes the stages of transformation as
1) binding
2) double~strand fragmentation of DNA (on the cell surface)
3) single~strand formation
4) interaction with recipient DNA
5) postsynaptic events (resolution of the heteroduplex DNA)
expression, etc.

As mentioned earlier, it was my plan to reexamine each of the
classical experiments that led to this model (some of which were
done here: .Bodmer & Ganesan °64-66) with the use of HOMOGENEOUS
populations of genetically (and isotopically) marked DNA such as is
now accessible , through the amplification of inserted DNA
segments in plasmid vectors. Dubnau furnishes an apt quotation to
illustrate the motive: "Our knowledge of the impact of DNA
structure and chemistry on binding is rudimentary." There are many
confusing and conflicting claims about the influence of
single~strandedness, nicks and other structural modifications of
the DNA on binding, and thereafter in the whole process.

Even such a simple issue as the role of DNA size on the
efficiency of transformation cannot be readily rationalized; though
it has long been known that larger molecules (about 10 nepgadaltons)
are most effective. Even here, lacking homogeneous samples, it is
difficult to separate issues of size from freedom from damage, and
freedom from competition from DNA lacking the specific markers.
(2

Since DNA bound to the cell can be sheared off, it would also
be desirable to have honogenous material to look for the
specificity of the binding site on the DONA: is it preferentially
one ends; are there preferred sites interstittally, and so on.

The basic theme of all these experiments is simply to trace
the fate of marked donor DNA through the 5 stages summarized above.
The use of honogenous (amplified DNA) with well defined termini
offers several advantages: high specific biological activity;
predetermination of the relevant input segment, €.8-, SO that it
can be recovered later by re~segmenting the target DNA with the
same restriction enzymes; definition and control of the other input
DNA species present. Current experiments use fragmented DNA of
variable size, extent and terminations, and generally highly mixed
with fragments of the entire source genome, With such heterogeneous
DNA, some tests are qualitatively unfeasible; others, e.g. precise
measures of the timing and variability of fragmentation, complex
formation, etc.,: could not be made with very high precision and
interesting anomalies would then be obscured.

For example, sone of our recent work (Harris-Warwick et al.,
1975) used partially purified EcoRI~sepments. These showed graded
efficiency of marker transfer that we speculated came fron
‘edge~effects’, namely that markers near the boundary of the
segment were more likely to be eroded at an early stage of
transformation than those at the center. This is a simple, perhaps
even intuitively obvious expectation. Nowever, there were still
some uncomfortable discrepancies between this view and the linkage’
statistics. Most desired would be a series of amplified segments,
embracing the same general region, but tailored at the ends for a
direct measure of the effect of the geography of a marker on its
acceptance in the chromosome. There should be no difficulty in
finding, and we have some strong hints already of mutations in the
restriction-target site that will give us segnents of different.
sizes as material for such experiments. Conversely, we can seek
deletions near the edges that will bring critical markers closer to
an edge.

In older work, many people may have noticed anomalies in the
frequency of co~transfer (linkage) of a pair of markers as a
Function of DNA concentration [sic] and of size. These might be
explained by a peculiar kinetics of competition between the larger
(2-marker~embracing) fragments and smaller ones, both fron the same
region and from other parts of the chromosome. In the past, we
could not reliably test our primitive speculations put forward to
explain such anomalies; we certainly were in no position to nake
observations that would critically challenge the consensual picture.

It may illustrate what we are looking for to recall how the
advent of Hfr strains of E. coli, and the careful kinetic studies
of E. coli conjugation by Jacob & Wollman were instrumental in
uncovering what was really going in in E. coli crossing. Kinetics
of that level of precision has not really been possible until now
with inhomogenous transforming DNA.
The basic experimental methods are already well~established,
in other Laboratories and in our own. Besides the conventional
biochemical~genetic marker mapping methods, DNA used in uptake and
integration studies would be double labelled (a la Bodmer &
Ganesan, for example) to permit pycnographic separation of mixed
complexes of donor and recipient DNA. Now we have the additional
advantage of restriction-enzyme~segmentation of the target PNA,
fractionation by gel-electrophoresis, and heteroduplex analysis by
electronmicrography, or with c=RNA staining (a la Sothern ). {The
latter is particularly pertinent in dealing with conplexes
involving heterologous segments, as we have studied in R. subtilis/
globigii crosses; and which arise in confrontations of reconbinant
plasmids with the recipient DNA. }

With homogeneous and amplified DNA we also expect to be in a
far better position to study the specificity of competition at
every stage of the process ~~ being able to make well formed
mixtures of known inputs. Likewise, the search for and
categorization of mutants that influence the efficiency of
transformation, either with regard to the donor DNA or the
recipient, will be greatly facilitated. A number of
recombination~-deficient mutants are known, mostly concerned with
uptake or polymerase~defects; much less is known especially in B.
subtilis about variations in the donor DNA arising by mutation.

There remains the task of getting appropriate material for
such studies. Our reported finding (soon going to press) about the
expression of some B. subtilis genes in E. coli opened the door to
the use of £. coli plasmids as vectors, and was our initial plan.
However, for many reasons -~ including biohazard considerations
especially in the current regulatory climate ~~ we would prefer a
system that allowed us to amplify subtilis DNA in B. subtilis
cells. We realize there may soon be many alternative possibilities
for this; but are relieved that there is one very likely avenue by
exploiting the growth of Staphylococcus plasmids in RB. subtilis.
These plasmids are technically [e.g. size, number of restriction
sites, selectible markers] superior to any other that we know of
right now in Bacillus. Yes, we have heard “rumors” about E. coli
plasmids also being transferable to subtilis, but would be in some
dilemma if we had to stake this proposal on unpublished information
from other laboratories.

We are just now in the process of trying to fulfill the
promise that the staph plasmids offer as cloning vehicles, which
entails the usual experiments of ligating sticky~-ended marked
segments to restriction-enzyme~segmented plasmid DNA. It is sone
advantage that we already have phi~3-t segments of known efficacy
for the thy+ marker, and can also try the subtilis segments that
have been cloned and found to effect transformation both in E. coli
and and in subtilts.
In addition, we are just starting some experiments on the
transmissibility of staph plasmids to B. subtilis in mixed culture,
viz., by some conjugal mechanism. This could be an operational
convenience; perhaps more important today, it would also throw some
light on the promiscuity of bacterial gene exchange under
conditions even less intrusive than the extraction of DNA for
transfection,

These experiments were motivated by Davies” speculation that
antibiotic-resistance factors in common pathogens ultimately
‘originated in soil actinomycetes which are the source of
antibiotics, and which have evolved adaptations to cope with them
in nature. If this can be substantiated, we are likely to find many
other instances of ’promiscuous” gene exchange as a natural
process. Indeed, several other workers are energetically seeking
other wide exchanges of plasmids with B. subtilis. It is quite
likely that the present staph system will not be the ultimately
ideal one, and we are prepared to shift gear in the light of
developing new information.

Finally, we have in mind that the staph plasmids appear to be
unstable in B. subtilis -~ an important biological and
methodological issue. We will inevitably be preoccupied with what
is behind stability. What will doubtless happen is that,
incidentally to other work, changes of stability pattern will be
noted: and if this occurs in an experimentaily advantageous
context, we will try to understand what genetic and physiological
factors influence the fitness of the exotic plasmid. In particular,
insertions of additional segments may have paradoxical effects:
from previous experience, larger plasmids will have some trouble
propagating. But in some cases the augmentation of the plasmid
with segments of the host genome may facilitate the durability of
the plasmid. The fact that, so far as we know, there is no
detectable homology of DNA sequences in the plasmid and in the
chromosome is actually highly advantageous for further analysis.
(We are having some trouble in tracking phi-3~-t on this account.)