8 
It should be noted here that the above five examples of 
how a gene might be considered "harmful" were given little, if 
any, attention in eitfier the Guidelines or the impact statement. 
It would seem that in so expanding the defiiiition of "harmful 
gene, "the probability of untoward effects occurring might oe 
somev/hat greater tVian that assumed by the RecomJainant DNA 
Advisory Committee or the authors of the impact statement. 
(b) The decision to go ahead with the approval of E. coli 
as the predominant host for recombinant DNA experiments on 
bacteria, ignores the fact that E. coli , in addition to being a 
conuTion and possibly beneficial denizen of the human colon, is 
also a human pathogen . Although it is correctly asserted that 
E. coli K-12 does not normally colonize the human colon, it 
is incorrect to infer that this property renders it harmless 
to humans. The primary vectors for inserting "foreign" DNA 
into strains of E. coli K-12 are plasmid DNA's. The following 
points, most of which received little of no discussion in either 
the Guidelines or tlie impact statement would argue that the use of 
E. coli as a recipient for recombinant DNA is inherently riskier 
than either the NIII guidelines or the impact statement v/ould lead 
us to believe: 
(1) E. coli K-12 can colonize the bowels of a small fraction 
of individuals, usually already debilitated by some other pathological 
condition. In sucli individuals, recombinant DNA containing 
Appendix K — 62 
