Federal Register / Vol. 46, No. 233 / Friday, December 4, 1981 / Notices 
59387 
could not transfer readily to other hosts 
(nonconjugative plasmids) were 
specified by the Guidelines. The 1976 
Guidelines did not specifically 
acknowledge the additional conclusion 
that such exchange implies: that for at 
least some pairs of organisms, 
recombinant DNA experiments may 
simply imitate nature in moving blocks 
of genes from one species to another. 
A series of promiscuous plasmids that 
move freely among severl species of 
bacteria, have been found in a large 
series of gram negative organisms, 
ranging from E. coli through RhizobJum 
and Pseudomonas (Alexander & Jollick, 
1977; Haas & Holloway, 1978; Baron et 
al., 1968). These plasmids frequently 
carry, in addition to genes for their own 
maintenance and transmission, 
antibiotic resistance genes and 
occasionally chromosomally derived 
genes (Olsen & Gonzalez, 1974; 
Holloway, 1978). Evidence of in vivo 
transfer of these plasmids from virtually 
any gram-negative species to any other 
gram-negative species has been found 
(Ingram et al., 1974; Smith, 1969). In 
laboratory experiments, chromosomal 
genes from a variety of organisms can 
be transferred by these plasmids into 
other species (chromosome 
mobilization). In some cases, such 
chromosomal genes can be stably 
recombined into the recipient 
chromosome. In other cases, the 
mobilized genes can be foimd stably 
associated with the plasmid, and, 
therefore, can be transferred at high 
efficiency to many organisms 
(Holloway, 1978). 
Although the precise nature of this 
mobilization is not know, it seems to be 
somewhat plasmid specific. At least 
some plasmids can mobilize the DNA of 
many diverse organisms (Haas & 
Holloway, 1978), and many genes from 
any given organism can be mobilized by 
these plasmids. Givei\the ubiquity of 
such plasmids, the ample opportunity for 
exchange in the environment these 
organisms share, and the very long time 
periods available, it seems reasonable 
to assume that, for most gram-negative 
bacteria, any given gene has in fact been 
introduced at some time into any given 
recipient. This probability was explicitly 
recognized in the 1978 revision of the 
Guidelines by the inclusion of a list of 
“exchangers" (Federal Register, July 28, 
1978). Recombinant DNA experiments 
between two exchangers are exempt 
from the Guidelines on the rationale that 
such a pair should have been exposed to 
each others DNA, and, therefore, no 
unique combination should be produced 
via recombinant DNA technology. If this 
principle is accepted in the broadest 
sense, essentially any cloning among the 
enterobacteriaceae, both pathogenic and 
non-pathogenic, would be exempted 
from the Guidelines. 
Many gram positive bacteria take up 
DNA from the environment (Low & 
Porter, 1978). When the DNA taken up 
has similar sequences to the host 
bacteria DNA, the bacteria are capable 
of recombining that DNA into their own 
chromosome. Some of these organisms 
have been shown to be transformed 
(changed genetically) by DNA released 
by neighboring bacteria in the soil 
(Graham & Istock, 1978; Burke & Le, 
1980). Therefore, such organisms 
exchange genetic information with other 
organisms by transformation; some pairs 
of microorganisms have been added to 
the exchanger list on this basis. 
Similarly, bacteria may frequently 
encounter mammalian DNA, either from 
decaying matter or from intestinal cells 
or ingested food in the case of gut 
bacteria. Mammalian cells in the 
intestine should frequently be exposed 
to DNA released from resident bacteria. 
It is difficult in these cases to 
realistically estimate uptake and 
persistence, although mechanisms for 
“illegitimate recombination" do exist 
both in bacteria and mammalian cells 
(Cold Spring Harbor Symposium, 1981; 
Kleckner, 1981). Large pieces of non- 
homologous DNA can be incorporated 
into the mammalian genome (Scangos & 
Ruddle, 1981). 
There have been reports of the 
existence of mammalian hormone-like 
proteins elaborated by bacteria (Koide & 
Maruo, 1981; LeRoith et al., 1981); this 
may reflect uptake and integration of 
mammalian DNA by bacteria sometime 
in the past. 
In at least one case, a system for 
genetic transfer from bacteria to plants 
exists. The plant pathogen 
Agrobacterium fume/ac/ens transfers a 
specific part of its extrachromosomal 
DNA, implicated in plant tumor 
formation, stably to the infected plant 
cells where the DNA is expressed and 
stably maintained (Chilton et al., 1977). 
The transferred piece is derived from a 
large, transferable plasmid which has 
the capacity to enter, by conjugation, 
many other prokaryotes. It is not known 
whether other such mechanisms exist, 
but even this isolated example provides 
a mechanism for moving bacterial and 
perhaps other types of information into 
many kinds of plants. 
Among animals and plants, viruses 
seem to be the most likely mechanism 
for pick up of chromosomal genes and 
transfer of genetic information from one 
host to another. In the last few years, it 
has become clear that many 
chromosomal sequences can be found in 
association with viral DNA (Bishop, 
1981; Weinberg, 1980). 
c. Counterarguments. 
1. While exchange can be documented 
in the laboratory, most genes will not be 
transferred in vivo in stable association 
with plasmids; only the most 
homologous regions will recombine into 
the recipient chromosome. Species that 
do not occupy the same ecological niche 
are unlikely to have the opportunity for 
natural exchange of genetic information. 
Introduction of DNA via transformation 
will rarely lead to stable diploid 
formation or replacement of the original 
DNA with new sequences. The use of 
the recombinant DNA technique may 
increase these processes dramatically 
by providing a mechanism for 
maintaining new sequences in the 
absence of homology. 
ii. The vectors used for recombinant 
DNA experiments may be engineered to 
exist in many copies in the celL or other 
mechanisms may be used to increase 
expression of genes beyond that found 
in nature. Such optimization of 
expression could lead to an organism 
unique in its ability to produce an 
excess of a particular product. 
2. Dissemination. 
a. Stability of Recombinant DNA. 
Much of the discussion of the possible 
hazards associated with recombinant 
DNA in E coli K-12 has centered on the 
inability of this organism to establish 
itself in the environment or disseminate 
recombinant DNA to other organisms. 
We have summarized some of these 
data in the Appendix, and further 
discussion of the data can be found in 
the report of a meeting in Falmouth, 
Massachusetts, where the 
epidemiological consequences of cloning 
in E. coli K-12 were discussed (Gorbach, 
1978). 
Some of the arguments made with 
respect to E. coli K-12 can be 
generalized to other systems as well. 
Laboratory strains, which most 
frequently will be used as hosts for 
recombinant DNA, may lose some of the 
characteristics which permit growth 
outside the laboratory. Many vectors, in 
being redesigned as useful vectors for 
recombinant DNA experiments, will lose 
the capacity for self-transmission. 
However, in a discussion of whether 
essentially any recombinant DNA 
containing organism is likely to be 
disseminated, we must assume that at 
least in some cases the host will be one 
not far removed from the wild 
environment, and that, in some cases, 
self-transmissible vectors will be used 
to carry the recombinant DNA, 
Therefore, one must ask if there are 
[ 276 ] 
