119 
effective colonizer by the introduction of specific genes either from E. 
coli or the organism Shigella flexneri , where the genes are known to be 
important in colonization. These experiments have failed to convert the E . 
coli K-12 to an effective colonizer. 
It also should be noted — and this experiment was actually mentioned 
earlier by Dr. Szybalski, the recent study carried out by Mark Richmond's 
laboratory. Mark Richmond is a prominent British bacteriologist. In these 
experiments that have been accepted for publication in the journal Gene , 
the feces of laboratory workers who carried out experiments with geneti- 
cally marked, tagged fe. coli K-12 strains carrying a variety of R plas- 
mids — these feces samples were monitored every two or three days over a 
two-year period. Neither this marked or identified E. coli K-12 strain, 
nor any of its plasmids, were detected in the fecal samples during this 
period of time. I would like to add that the R plasmids used in this study 
were conjugative plasmids — that is, they have all the necessary genetic 
information to promote their transfer from one cell to another. And of 
course, when we are talking about the Guidelines and cloning vectors, we 
are talking about so-called nonconjugat ive plasmids — plasmids that don't 
have, as far as we can tell, any of the so-called dozen genes that are 
required to promote the transfer of the vector with presumably its attached 
recombinant DNA — transfer of this from one cell to another. 
The detection of a plasmid transfer between E. coli cells and the 
human intestinal tract clearly has been reported as published experiments 
several years ago. There are two papers that appeared in 1975. In these 
experiments, it was clearly shown that this transfer of plasmids within the 
intestinal tract was an infrequent event, and its demonstration, to be able 
to pick that up at all required the feeding of massive doses of a donor 
strain carrying again a conjugative or self-transmissible plasmid. As I 
indicated a few moments ago, again the EK1 host-vector systems demanded by 
the Guidelines require nonconjugat ive plasmids to be used as plasmid vectors. 
The plasmids that early were used as vectors — col El and pSClOl — 
are nonconjugat ive , in that they cannot promote their own transfer from one 
cell to another; and we find now that the new generation of plasmid vec- 
tors, namely pBR313 and pBR322, that generally are being recommended for 
use in the EK2 system, not only are nonconjugat ive , but they are severely 
handicapped in their ability to be mobilized out of a cell by a self-trans- 
missible plasmid that may inadvertently establish itself as a rare event in 
an EK1 plasmid host system. 
It has been estimated, and one can argue with these kinds of probability 
calculations, but it has been estimated that the probability for transmission 
of these plasmids from the host cell is less than 10 -1 r per surviving bacter- 
ium per day in the intestinal tract of warm-blooded animals. I should like 
to add that this estimate is probably on the high side, because in a sense 
it is looking at it under best situations. It doesn't take into account 
that E. coli's metabolism is much lower in the intestinal tract than in the 
laboratory, and the higher the metabolic rate, we have found, the higher the 
[ 323 ] 
