Page 25 of Attachment E 
stein, based on a model outlined by Dr. Ron Davis. This model in- 
volved tests of lambda carrying yeast shotgun fragments, serially 
passaged through coli. The object of these tests was to arrive at 
some rational way of dealing with populations of clones which would 
include collectively, the entire gene content of a particular organism. 
It seemed most reasonable to start with yeast because its genome is 
much smaller than most eukaryotes. Further tests could include 
more complicated eukaryotes. The test would involve taking an entire 
shotgun from a cloning experiment and to passage it through animals 
(either normal or germ-free mice tor example). After colonization was 
established, the bacteria would be reisolated, and checked to see 
whether any clones had survived better than others. The “winners ", 
those present in highest number, would be assumed to be those 
which were best adapted for colonization. These could then be 
compared by various criteria with wild type E. coli. The results with 
cloning of yeast genes in coli on the lambda chromosomes have 
shown that in every case in which good survivors were selected and 
checked by comparison with wild-type lambda, they grew less well 
than the wild-type strain. 
The question then arose, what does one do with clones that have 
been purified from shotgun experiments? What sort of test could be 
required of such clones? The participants in the conference were 
unable to offer any concrete test in the particular situation other than 
using the same sort of tests that had been outlined for the shotgun. It 
is clear, however, that some way to cope with these kinds of clones 
must be developed, since most Guidelines place shotgun ex- 
periments in a higher risk category than those using purified genes 
simply to compensate for the lack of knowledge concerning possibly 
harmful gene elements within the population. Some way of putting a 
value on that of the “lack of knowledge’’ component has to be 
developed if guidelines are to evolve rationally. 
Another topic of discussion was centered on experiments in- 
volving animal virus DNAs. Here the critical question was whether or 
not the cloning procedure would in any way create a path of entry for 
animal viruses that would bypass the normal defense mechanisms of 
the host. If so, then containment higher than that required for the 
animal virus itself would be warranted. However, it was considered 
equally possible that putting animal virus DNA into E. coli would 
really represent an attenuation of viral virulence. If it turns out that 
attenuation is the rule, then cloning may represent a preferred way to 
work with animal virus pathogens. It was considered absolutely 
essential that risk assessment experiments to test the effects of 
cloning animal virus genes be undertaken. There was unanimous 
agreement that the Martin and Rowe type of experiment (that is, 
cloning of polyoma DNA in E. coli and exposure of mice to the 
recombinant clones in various ways) constitutues the most sensitive 
and least hazardous way to test this question. Drs. Martin and Rowe 
have in fact done some preliminary non-cloning experiments in an- 
ticipation that they would soon have the facilities to do these kinds of 
- 19 - 
[ 404 ] 
