Federal Register / Vol. 45, No. 182 / Wednesday. September 17, 1980 / Notices 
61875 
followed by activated sludge treatment 
and anaerobic digestion of all sludges 
generated). This work was described in 
brief form in the Recombinant DNA 
Technical Bulletin of July 1979. 
Similar results are now available for 
the survival of E. Coli Chi 2656 (which 
carries plasmid pB322) and for E. Coli 
GF 2174 (carrying plasmid pBR325). 
Evidence also was sought in these 
studies for transfer of the plasmids to 
indegenous coliforms and, within the 
sensitivity of the methods used, such 
transfer was not demonstrated. 
Experiments favoring plasmid transfer 
in raw wastewater and in primary 
biological sludges also failed to 
demonstrate such mobilization. 
Likewise, E. Coli 2e0lC carrying a 
lambda prophage (Xcl857 nin5 plac5) 
was removed effectively by wastewater 
treatment; no free lambda phage was 
observed at any point in the treatment 
train. 
Studies have been completed 
measuring the survival of an indigenous 
£ Coli and E. Coli K-12 in these 
treatment plant models as well. The 
results of experiments utilizing such 
genetically tagged organisms confirm an 
earlier report that conventional 
domestic wastewater treatment 
processes will result in at least two log» 
reductions in concentration of 
indigenous, EKl, and EK2 organisms. 
These results are similar to those 
obtained with the indigenous 
wastewater microflora. Further 
reduction of these host-vector 
populations will depend on appropriate 
sludge treatment and effluent 
disinfection. 
c. The contractor who was performing 
tests in mice and in cultures made 
significant progress during the 
intervening year. The most notable 
result was the establishment of a 
mathematical model to study the 
exchange of plasmids between normal 
hosts and indigenous bacteria. 
Plasmid transfer in continuous flow 
(CF) cultures of defined or natural 
intestinal flora occurred with similar 
efficiency as in pure cultures in vitro. 
One can conclude that the capacity of £ 
coli K-12 strains to function as plasmid 
donors or recipients was not impaired 
by the presence of an indigenous 
microflora. £ coli Chi 1776 donated 
plasmid Rldrdl9 with somewhat lower 
efficiency than a standard £ coli K-12 
donor. 
Plasmid transfer in mice harboring a 
defined intestinal microflora appears to 
have the same degree of efficiency as in 
the CF cultures, because the rate 
constants calculated on the basis of the 
equations are similar to those obtained 
in CF cultures. However, the validity of 
this mathematical treatment for mice is 
not completely certain, because the 
mathematical model is based on the 
assumption that the bacteria are freely 
suspended, which is not likely to be the 
case in the animal. Nevertheless, the 
data suggest that plasmid transfer 
efficiency in the gut should not differ 
profoundly from that in CF cultures. 
These conclusions are based on 
similarities of transfer rate constants as 
measures of donor and recipient 
capacity. The actual rates of transfer 
are, of course, critically dependent on 
the concentrations of donors and 
recipients. The £ coli populations in 
normal mice (and in people) are so low 
(approx. 10 s per ml gut content), that 
little plasmid transfer occurred in the 
mice after the initial 20 hours of the 
experiment, i.e. after the large inoculum 
of donors had passed through the 
animals. This was true in spite of the 
fact that the resident £ coli were 
artifically implanted, highly efficient 
recipients for the highly efficient 
conjugative plasmid. It is not verjr 
common to find £ coli of such high 
recipient capacity in natural gut flora. 
Despite the fact that the capacities of 
the £ coli strains to donate or receive 
the plasmid were not seriously impaired 
in the gut, the quantitative parameters in 
the normal gut are such that little actual 
transfer occurred once the large initial 
inoculum of donors had been eliminated. 
From these data one may tentatively 
conclude that the probability for 
triparental transfer of a non-cojugative 
plasmid in the normal gut would be 
exceedingly small. The only 
circumstances in which one could 
imagine a realistic possibility for 
triparental transfer to occur, would be at 
times when abnormally high £ coli 
populations are present in the gut, as 
may be the case during fasting and other 
types of stress, in diarrhea or as a 
consequence of antibiotic therapy. 
d. The major effort of the contractor 
who was testing in both mice and 
humans in the interim has been to 
initiate triparental mating studies using 
Chi 1666 (an EKl host), the plasmid 
pBR322 and two additional mobilizing 
plasmids. 
The studies are designed to determine 
if the presence of a mobilizing plasmid 
in a bacterial host cell can cause the 
transfer of the non-conjugative pBR322 
to the indigenous microflora of either 
mice or humans. 
Two studies have been done in mice. 
In both cases germ-free mice were 
colonized with human £ coli to provide 
an array of potential recipients and then 
fed the Chi 1666 containing the three 
plasmids. In the first study 5x10* 
organisms were fed once and fecal 
samples from the mice were pooled to 
facilitate assay; Chi 1666 could only be 
isolated during the first day of the study. 
In the second mouse study, the Chi 1666 
containing the plasmids were fed daily 
for 4 days to the mice and their drinking 
water contained tetracycline. In this 
latter group the Chi 1666 survived at 
levels of lOVgram or less for the 4 days. 
In neither case, however, was the 
pBR322 mobilized to the indigenous 
flora. 
Finally, Chi 1666 containing pBR322 
and the two mobilizing plasmids was 
fed to 4 human subjects. More than 70 
fecal samples were collected during the 
study and a large number of cultures 
and subcultures (2500) performed. The 
Chi 1666 survived from 3V* to 7 days and 
that interval corresponds well with 
previous data from studies in which Chi 
1666 was fed to humans with the 
without pBR322. While the data for 
transfer of pBR322 await DNA:DNA 
colony hybridization and further 
confirmation, from the data already 
available on the two transferable 
plasmids, we can make certain 
estimates. Highly transferable pSL222-4 
was found to transfer to coliforms at a 
frequency no higher than 8x10"*. This 
value is 10"* lower than seen when the 
I tlasmid is tested under the best 
aboratory conditions. Thus pBR322, 
which is transferable by F-like R 
plasmids at 10"'-10'* would be expected 
to be transferred at an extremely low 
frequency; probably below 
experimentally detectable levels. The 
same is presumably true of potential 
mobilization by the second conjugative 
plasmid, although the data for their 
mobilization of pBR322 are not readily 
available. We would estimate that 
pBR322 would be transferred at a 
frequency at least 10M0* fold less than . 
the conjugatable plasmid, i.e., 10"*-10' 10 . 
2. Progress on the evaluation of the 
transmission of vectors from £ coli K- 
12 to other bacteria in the 
astrointestinal tract of animals and 
uman beings has been made via three 
approaches. 
a. Data reported in section 1 by two of 
the contractors relates directly to this 
•question. 
b. A meeting was convened on August 
30, 1979 at the NIH with the purpose of 
considering Falmouth Workshop 
Protocols I and II (Jour. Infect. Dis. 157, 
704-708. 1978) and some related issues. 
Dr. Stanley Falkow served as Chairman 
and sixteen participants and NIH staff 
attended. 
Protocol I addressed th ; colonization 
and transmission of plasmids from £ 
coli K-12 in the gastrointestinal tract of 
humans to other bacterial strains in the 
intestinal flora. The Working Group 
[ 218 ] 
