Page 36 of Attachment E 
Section VI 
COGENE-INITIATED RISK ASSESSMENT 
EXPERIMENTS: 
(1) Recommendations from March 5, 1977 
Richmond-Falkow-Anderson Meeting 
(2) Recommendations from March 30, 1978 
Richmond-Falkow-Anderson Meeting 
(3) Recommendations of the Working Group 
on Risk-Assessment. — April 3, 1978 
OUTLINE PROTOCOL OF TESTS FOR 
COLONIZATION AND TRANSMISSION 
OF E. COLI AND E. COL/ CONTAINING PLASMIDS 
The following outline was submitted to COGENE by Professor E.S. 
Anderson, S. Falkow and M.H. Richmond . . . 
“ Even though a certain amount of work on the survival of E. coli K-12 
in man has already been published, we feel that there is a case for 
carrying out a wider study. In view of the particular concern of COGENE, 
we feel that experiments should be of two broad types. First, studies in 
which volunteers swallow marked E. coli lines carrying appropriate 
bacterial plasmids and the excretion of the strains and of the plasmids 
is then determined; and secondly, studies in which workers involved in 
“genetic engineering ” are monitored to see whether they ever excrete 
detectable numbers of the bacteria and plasmids with which they are 
working. 
As far as the feeding experiments are concerned, we would broadly 
envisage two types of experiment. In the first, a nalidixic acid resistant 
E. coli K-12 line, carrying a well defined self-transmissible plasmid with 
an unusual marker pattern, such as trimethoprin + kanamycin 
resistance, would be used, and the excretion of other E. coli strains 
which had picked up other plasmids would be monitored. The object of 
this experiment, therefore, is to test: 
a) the survival of a fed E. coli K-12 line, 
b) the frequency of transfer of a self-transmissible plasmid from 
that strain to the resident Gram-negative flora, 
c) the frequency of transfer of other R-plasmids to the fed 
strain. 
It was envisaged that about 20 volunteers would be used in each test 
centre, and it was thought an advantage if the strains were fed in milk to 
some of the volunteers, and in nutrient broth to others. The first 
procedure would be likely to give a “worst case" response, while the 
latter would be more characteristic of a laboratory accident. 
The second type of feeding experiment would be broadly of the same 
type as the first, except that the E. coli K-12 strain would be one carrying 
a non-self-transmissible plasmid with appropriate markers for 
- 30 - 
[ 415 ] 
