UNIVERSITY OE WASHINGTON 
SEATTLE, WASHINGTON 98195 
April 19, 1978 
School of Medicine 
Department of Microbiology and Immunology, SC-42 
Dr. Mary Clutter 
Rational Science Foundation 
Washington, D. C. 
Dear Mary: 
There are two kinds of recombinant DNA experiments involving 
the Agrobacterium Ti plasmid that we would like to perform 
in a highly contained plant facility. We do not anticipate 
any danger to plants from the recombinant DnA. However be- 
cause of the nature of the DM A molecules to be joined, one 
a conjugative broad host range plasmid and the other an 
oncogenic plasmid, the experiment would not be permitted 
under the present N.I.H. Guidelines. The two experiments we 
propose below can be viewed as: 1. Risk assessment, and 2. 
Benefit assessment. In the first, we propose to determine 
whether a "shrunken" Ti plasmid created by cloning can create 
a super-pathogenic bacterium. In the second, we propose to 
test whether such a "shrunken” Ti plasmid can be exploited 
as a vector to alter the nutritional quality of a crop plant. 
Experiment 1. Creation of a shrunken Ti plasmid . 
Agrobacterium Ti plasmids cause tumors in a wide range of 
dicotyledonous plants. The transformed plant cells, which 
can be propagated in vitro in the absence of the inciting 
bacterium, maintain one to several copies of a small segment 
of the Ti plasmid, i It'''is not yet proven whether the Ti plasmid 
segment (T-DNA) is covalently joined to plant DNA. The length 
of the T-DNA is variable in different tumor lines. 
We have cloned three Hind III fragments of Ti-B6-8o6 that 
encompass all or most of the T-DNA. The vector is FBK322, 
a vehicle that does not replicate in Agrobacterium so far 
as we can tell. The host bacterium is E. coli X1776. We have 
tested these clones for oncogenicity on carrot discs and the 
results were negative. In the laboratory of Jeff Schell in 
Gent, Belgium, a Ti:iRP4 cointegrate plasmid has been trans- 
ferred from Agrobacterium to a C600 strain of E. coli . While 
the cointegrate plasmid confers oncogenicity on the Agrobacterium 
host, the E. coli host does not acquire oncogenicity. Thus 
E. coli , even a robust strain, does not seem able to incite 
plant tumors, even though the entire Ti plasmid is present. 
In order to test small cloned fragments of the Ti plasmid 
for oncogenicity, it is therefore essential that they be 
transferred to A gro bacterium , the host in which the trait 
is expressed. 
CIO 5 Health Sciences Building/ Telephone: (206) 5-fl-‘>82-J 
APR 26 f978 
[Appendix A — 303] 
