NOTICES 
33123 
There are those, however, who view 
the ubiquity of E. coli in vertebrates 
as an argument against the use of even 
the K-12 strain as a host for foreign 
DNA. Concern has been expressed, for 
instance, that the bowels of persons on 
anitbiotics, ill persons, human infants, 
or members of other species may be 
susceptible to colonization. ( 7) More- 
over, if E. coli K-12 can survive in a 
person or animal, it might confer ge- 
netic characteristics to hardier bacte- 
rial inhabitants. 
Scientists engaged or interested in 
recombinant DNA research have ad- 
dressed these concerns in three ways - 
first, through guidelines specifying 
practices and conditions of contain- 
ment for experiments classified ac- 
cording to the presumed hazard; 
second, through attempts to develop 
safer hosts and vectors; and third, 
through risk-assessment studies. The 
results of a workshop held in June 
1977 at Falmouth, Mass., to evaluate 
the potential risk of recombinant DNA 
experimentation with E. coli K-12 will 
be discussed below. 
The Evidence That E. Coli K-12 Is 
Nonpathogenic. The laboratory var- 
ients of K-12 permitted in recombin- 
ant DNA experiments have never been 
reported to cause disease, even in labo- 
ratory workers. K-12 has been grown 
in large quantities -up to hundreds of 
liters containing as many as a trillion 
bacteria. These cultures have been 
pioduced in countless laboratories the 
world over, and under containment 
conditions lower than the minimal 
ones in the NIH, guidelines. K-12 has 
none of the properties generally asso- 
ciated with pathogenic bacteria.(S-75) 
It does not— 
» survive and multiply readily in 
natural environments. 
• spread from animal to animal or 
plant to plant, 
• multiply readily on body surfaces 
or intestines and lungs, 
e penetrate animal cells or spread 
through animal bodies, 
• produce a toxin or otherwise alter 
other living things to cause disease, or 
• resist normal body defense mecha- 
nisms. 
Even after as many as 10 billion K- 
12 organisms have been ingested, their 
multiplication in normal humans is 
only transient, and after a time none 
can be recovered. (10, 11, 12, 14) Thus, 
K-12 does not establish itself as a per- 
manent resident of human beings. On 
the other hand, K-12 can reside under 
abnormal conditions, as during antibi- 
otic therapy. (.16) 
A micro-organism, in order to cause 
disease, must have the genetic capabil- 
ity to do so, as well as the ability to es- 
tablish itself in the body. It is difficult 
to conceive how- K-12, itself nonpatho- 
genic, could become pathogenic as a 
result of genetic manipulation. Highly 
attenuated, it is known only to inhabi- 
tat the biological laboratory. 
Even when genetic determinants of 
pathogenicity in other E. coli strains 
were introduced into K -12, no instance 
of capacity to induce diarrheal disease 
or urinary tract infection could be de- 
tected. (14, 15) The workers conclude 
that the inadvertent transformation 
of K-12 into a highly pathogenic form 
by the introduction of a single frag- 
ment of foreign DNA is highly unlike- 
ly. (.15) Indeed, the number of charac- 
teristics that a microbe must have in 
order to cause disease is believed to be 
great, not to mention additional char- 
acteristics needed to produce an epi- 
demic. (7 7) 
Transfer of Foreign DNA from E. coli 
K-12. While it would appear impossi- 
ble to render E. coli K-12 pathogenic 
by the introduction of foreign DNA, 
there is still to be considered whether 
the inserted fragment could be trans- 
mitted to another bacterium with 
which the K-12 comes in contact, in- 
cluding other strains of E. coli. Such a 
transmission might convert the recipi- 
ent into a pathogen or render a patho- 
gen more viable. The case of plasmid 
vectors is considered first. 
Plasmids are intracellular particles 
composed of DNA and not dependent 
on chromosomes for their replication. 
Hence, they can be used as vectors, or 
vehicles, for transporting foreign DNA 
into the bacterial host, where they 
multiply and propagate the genes they 
bear. Certain plasmids (called “conju- 
gative”) are inherently able to migrate 
from one bacterial cell to another. 
These are prohibited for nearly all re- 
combinant DNA experiments. Only 
plasmids not capable or barely capable 
of spontaneous intercellular migration 
("nonconjugative”) may be used. 
The nonconjugative plasmid's ability 
to migrate is augmented if the cell 
harboring it is invaded by a conjuga- 
tive plasmid, which may confer this 
property. Then even the non- conjuga- 
tive plasmid may become a potential 
DNA-bearing invader. It has been cal- 
culated, however, that the chance of 
this occurring with certain K-12 plas- 
mid systems is less than 1 in 10 16 (10 
quadrillion) K-12’s surviving per day 
in the intestine of warm-blooded ani- 
mals. (13, 14) The probability is even 
lower in sewers, sewage treatment 
plants, and waterways. It should be 
noted that since most of the estimates 
of probability are based on data ob- 
tained under laboratory conditions, 
animal ahd human feeding studies are 
needed to verify the predictions.US) 
Consideration must also be given to 
the question of transfer of foreign 
DNA from the initial K-12 host to 
other bacteria by means of bacterio- 
phage vectors. Bacteriophages are vir- 
uses that infect only bacteria. They 
could escape the laboratory either as 
mature infectious particles or in bacte- 
rial hosts in which the phage DNA is 
carried as a plasmid or within the 
DNA of the cell. 
The survival of phage DNA when re- 
leased as infectious particles depends 
on their stability in nature, their in- 
fectivity, and the probability of effec- 
tive encounters with naturally occur- 
ring E. coli. The bacteriophage used in 
recombinant DNA experiments is 
known as lambda. It is considered very 
unlikely to survive and to infect resi- 
dent E. coli in animals and humans, 
being highly sensitive to stomach acid, 
reluctant to infect smooth E. coli cells 
(the type normally found in the gut), 
and susceptible to drying, as would 
occur if it escaped into the air. More- 
over, E. coli vulnerable to lambda is 
uncommon in nature. Infective lambda 
ingested in large amounts (10 -, or 100 
billion, particles) could not be detected 
in human feces.(79) 
Establishement of lambda as a resi- 
dent of the E. coli host cell’s DNA is a 
well-known example of natural recom- 
bination, In certain cases, it is a fre- 
quent event, as likely to occur as not. 
Hence, this mode of escape would be 
the preponderant laboratory hazard. 
However, most variants of lambda 
used (or under consideration for use) 
in recombinant DNA experiments 
have a much reduced ability to become 
so incorporated. (2d- 22) Here the prob- 
ability drops to 10 5 or 10 G — 1 in 
100,000 or 1,000,000.(22-25) 
The estimates for containment in 
the use of bacteriophage host-vectors, 
while not exact, are sufficient to 
assure that the probability of transfer- 
ring a foreign DNA fragment from the 
original K-12 host to other bacteria is 
remote. 
Ability of E. Coli K-12 To Survive 
and Spread in Nature. Thus far, the 
suitability of K-12 for recombinant 
DNA experiments has been considered 
in relation to its ability to do harm 
FEDERAL REGISTER, VOL. 43, NO. 146— FRIDAY, JULY 28, 1978 
[ 84 ] 
