December 14, 1977 
Dr. Donald S. Fredrickson 
Director 
National Institutes of Health 
Bethesda, MD 20014 
Dear Dr. Fredrickson: 
At the suggestion of Dr. Perpich , we are writing to you to try 
to clarify the issues surrounding the problem of in_ vitro recombina- 
tion experiments in laboratory strains of Salmonella typhimurium . 
It seems to us that there are two points of importance: First, that 
there is a body of scientific knowledge of great and general interest 
which can be gained from such experiments in Salmonella which cannot 
be obtained as easily or at all using E. coli K12. Second, that the 
laboratory strains of Salmonella typhimurium in question (all deri- 
vatives of strain LT2) if handled with normal precautions do not 
pose any significant risk of causing infections in man. 
With respect to the first point, we wish to only briefly state 
that there are interesting phenomena and practical applications of 
molecular biology which depend upon Salmonella . Mutagenesis tests 
for potential carcinogens developed in the Ames laboratory use 
tester mutants whose nucleotide sequence would be of great interest. 
Application of current technology to this sequencing is made extremely 
difficult by the current interpretations of the NIH Guidelines (1-4) . 
The exciting progress in the understanding of phase variation (thus 
far only clearly described in Salmonella ) is one of the first really 
new results to arise because of recombinant DNA methods (5). In 
that case Salmonella DNA was put into E. coli . Further genetic 
analysis of this mechanism awaits, in great measure, permission to 
return these DNA molecules to their original organism. A more de- 
tailed presentation of the virtues of Salmonella research is appended. 
Next to E. coli , Salmonella is genetically the best understood free 
living organism on earth (6,7). 
In respect to pathogenicity of laboratory strains, we offer 
the following. Strain LT2 and its immediate derivatives are the 
only strains we wish to consider for in vitro recombinant DNA work. 
Virtually all Salmonella genetics has been done in these strains. 
Informally, we have polled many of the major laboratories using 
this organism. Under standard conditions of use, we know of only 
six possible laboratory infections among more than 1000 persons- 
years. Of these, one is published as having been caused by strain 
LT2 (8), but the strain in question is, in fact, a derivative of 
the more virulent strain LT7 (9) ; another turns out also not to be 
LT2 but a strain which came from a hospital isolate relatively 
recently and which differs antigenically from LT2. All of the 
remaining cases were caused by improper laboratory practice (i.e., 
resuspension of concentrated pellets by mouth pipetting) and 
followed an oral dose of more than 10^ organisms. 
[Appendix A — 129] 
