NOTiCfS 
33059 
fiuous. because an exchange experiments 
have undoubtedly been performed already 
in nature. 
In every instance of artificial recombina- 
tion. consideration must be given to the pos- 
sibility that foreign DNA may be translated 
into protein (expressed', and also to the pos- 
sibility that normally repressed genes of the 
host may be expressed and thus change, 'un- 
desirably. the characteristics of the cell. It 
is assumed that the more similar the DNAs 
of donor and host, the greater the probabil- 
ity of expression of foreign DNA. or of pos- 
sible derepression of host genes. In those 
cases where the donor exchanges DNA with 
E. coli in nature, it is unlikely that recom- 
bination experiments will create new genet- 
ic combinations. When prokaryote donors 
not known to exchange DNA with E. coli in 
nature are used, however, there is a greater 
potential for new genetic combinations to be 
formed and be expressed. Therefore, it is re- 
quired that experiments involving prokaryo- 
tic DNA from a donor that is not known to 
exchange DNA. with E. coli in nature be car- 
ried out at s higher level of containment. 
Recombination using prokaryotic DNA from 
an organism known to be highly pathogenic 
is prohibited. 
There are only limited data available con- 
cerning the expression of DNA. from higher 
forms of life (eukaryotes) in E. coli (or any 
father prokaryote). Therefore, the contain- 
ment prescriptions for experiments insert- 
ing eukaryotic DNA into prokaryotes are 
based on risks having quite uncertain prob- 
abilities. 
On the assumption that a prokaryote host 
_Uht translate eukaryotic DMA, it is fur- 
ther presumed that the product of that for- 
eign gene would be most harmful to man if 
it were an enzyme, hormone, or other pro- 
tein that was similar homologous) to pro- 
teins already produced by or active in man. 
An example is a bacterium that could pro- 
duce insulin. Such a "rogue” bacterium 
could be of benefit if contained, a nuisance 
or possibly dangerous if capable of surviving 
in nature. This is one reason that the 
higher the phylogenetic order of the eukar- 
yote. the higher the recommended contain- 
ment. at least until the efficiency of expres- 
sion of DNA from higher eukaryotes in pro- 
karyotes can be determined. 
There is a second, more concrete reason 
for scaling containment upward as the eu- 
karyote host becomes similar to mam. This 
is the concern that viruses capable of propa- 
gating in human tissue, and possibly causing 
diseases, can contaminate DNA. replicate in 
prokaryote hosts and infect the experimen- 
talist. Such risks are greatest when total 
DNA from donor tissue is used in "shotgun” 
recombinant experiments: it diminishes to 
much lower levels when pure cloned DNA is 
used. 
The structure of the classification 
for permissible experiments is based, 
therefore, on assumptions governing 
potential risk. It should be emphasized 
again that although recombinant DNA 
experiments have now been performed 
for over five years in hundreds of labo- 
ratories throughout the world with 
hundreds of thousands of different re- 
combinant DNA molecules produced, 
no case of hazard has been demon- 
strated. 
Part III of the guidelines assigns to 
each specified class of experiments a 
level of physical containment and a 
level of biological containment at 
which the experiment shall be pier- 
formed. As noted before, there is 
10.000- to 100.000-fold protection in 
going from a class I or n biological 
safety cabinet to a class III biological 
safe r y cabinet (i.e., from P3 to P4). 
Similarly, in going from PI to P3 there 
may be a 10.000- to 100.000-fold in- 
crease in safety. For biological con- 
tainment, there is the criterion for 
HV2 systems that "escape of the re- 
combinant DNA either via survival of 
the organisms or via transmission of 
recombinant DNA to other organisms 
should be less than 1/1C S under speci- 
fied conditions.” However, that crite- 
rion is not relative to the HV1 host- 
vector systems but absolute: thus, this 
might be a characteristic found for 
some host-vectors in the HV1 system, 
but it is mandated for all EV2 sys- 
tems. This level was chosen, it was 
pointed out. because it represents a 
practical limit which one can measure 
experimentally. 
Use of E. coli K-l 2 
A number of comments were made 
concerning the use of E. coli host- 
vector systems. It was observed that 
because E. coli K-l 2 is currently a 
“poor" pathogen doesn’t mean that it 
might not be converted to a “good” 
pathogen with the addition of one or 
two genes; the enfeebled nature of E. 
coii K-12 “is presumably the conse- 
quence of mutation(s) introduced 
during its laboratory passage.” but 
that perhaps different strains of K-12 
with different histories may not all be 
similarly enfeebled. 
Further, it was claimed that the fail- 
ure to convert K-12 to a pathogen by 
the use of certain plasmids or Salmc- 
netta genes is not definitive; to be de- 
finitive, we must have the detailed 
nature of the mutaiions in K-12 
"which prevent the expression of 
pathogenicity.” Also, it was noted that 
there is no way to assess the absolute 
risk associated with these experi- 
ments, and that it is important to 
assess the potential harm net only to 
man but to plants, anima ls, and the 
enviroment. 
Another commentator urged that 
this section be supplemented with the 
evidence from the Falmouth Confer- 
ence to show that the potential risk is 
minimal. A commentator cited the po- 
tential risk on the basis that “virtually 
any highly conserved physiologically 
active eukaryotic protein * * * or frag- 
ment thereof could be highly toxic 
when introduced out of context by a 
bacterium which received the appro- 
priate gene in a recombination experi- 
ment.” This criticism of the E. coli K- 
12 system does not detract from the 
scientific knowledge over the past two 
years of the great safety of this 
system. This evidence is presented in 
detail in the Environmental Impact 
Assessment. I agree that different 
strains cf K-12 with different histories 
may not all be similarly enfeebled and 
that failure to convert K-12 to a path- 
ogen to date does not prove it can 
never happen. However, the safety of 
E. coli K-12 has been clearly shown, 
and there is no need to limit or specify 
particular strains for EK1. After 30 
years of work with many different 
strains, there is still no known patho- 
genic E. coli K-12 strain. Thus, there 
is presumptive evidence that all K-12 
strains are safe. They are well suited 
for laboratory experiments because 
they take up DNA easily, but their cell 
wall makes them unsuited to compete 
in nature with wild-type E. colt 
On the basis of the Falmouth Con- 
ference (which is discussed further in 
the Environmental Impact Assess- 
ment), the conclusion can he drawn 
that it is essentially impossible for E. 
coli K-12 to be transformed by recom- 
binant DNA into a wild-type, patho- 
genic E. celt An E. coli K-12 contain- 
ing toxic genes through recombination 
could theoretically present a risk to a 
laboratory worker who accidentally in- 
gested it; but it would only be to that 
laboratory worker. There is evidence 
to show that harmful genes will have a 
very low probability of being tra'is- 
ferred from E. coli to another organ- 
ism. The plasmids used at the HV2 
level are engineered so that they nei- 
ther self-transfer nor transfer when 
another plasmid induces conjugation. 
Thus, the high degree of safety of this 
system is clear and explains why it is 
preferable to any other host-vector 
system at present. 
General Classification 
Disagreement was expressed over 
whether the PRG-RAC was too strin- 
gent or too lax. Those arguing the 
former position maintain that the 
guidelines should be relaxed even fur- 
ther because all the experimental evi- 
dence gathered and analyzed in the 
past 2 years indicates that the initial 
fears concerning the potential hazards 
were extremely exaggerated: more- 
over, the benefits to be derived from 
the research are great. Aiso, it is 
pointed out that recombinant DNA ex- 
periments not allowed under the cur- 
rent NTH guidelines are proceeding 
with the approval of responsible na- 
tional committees in a number of Eu- 
ropean countries. Those opposing this 
view argue that there is a lack of ex- 
perimental data for a sound evaluation 
of the potential risks, and the fact 
that a recombinant DNA experiment 
is permitted in Europe is irrelevant to 
the establishment of standards in the 
United States. 
Recombinant DNA Experiments In- 
volving Viral DNA 
FEDERAL REGISTER, VOL. 43, NO. 146— FRIDAY, JULY 28, 1978 
[ 20 ] 
