NOTICES 
33163 
tainment levels for recombinant DNA 
experiments involving eukaryotic viral 
DNA inserts. However, if the virus 
itself must be handled at higher levels 
of physical containment it seems pru- 
dent at the present time to use the 
more stringent containment condi- 
tions. It was emphasized that contain- 
ment practices must include adequate 
training and the use of high quality 
microbiological technique. 
EXPERIMENTS WITH EUKARYOTIC HOST- 
VECTORS 
1. Vertebrate host-vector systems in 
which viral DNA vectors are used to 
propagate other DNA segments. 
The workshop participants endorsed 
the portion of the NIH guidelines that 
describes the features required for an 
animal virus to be used as a cloning 
vector (i.e., the first portion of section 
III. B. 3. a. of the draft version, revised 
NIH guidelines for recombinant DNA 
research). The following is our recom- 
mended version of this section: 
Because this work will be done almost ex- 
clusively in tissue culture cells, which have 
no capacity for propagation outside the lab- 
oratory, the primary focus for containment 
is the vector; It should be pointed out that 
the risk of laboratory acquired infection as 
a consequence of tissue culture manipula- 
tions is very low. Given good microbiological 
practices, the most likely mode of escape of 
recombinant DNAs from a physically con- 
tained laboratory is carriage by an infected 
human: thus the vector with an inserted 
DNA segment should have little or no abili- 
ty to replicate or spread in humans. Fur- 
ther, a recombinant virus should not inad- 
vertently pose a threat to any species. 
For use as a vector in a vertebrate host 
cell system, an animal viral DNA molecule 
ideally should display the following proper- 
ties: 
(a) It should not consist of the whole 
genome of any agent that Is Infectious for 
humans or that replicates to a significant 
extent In human cells in tissue culture. 
(b) It should be derived from a virus 
whose epidemiological behavior and biologi- 
cal properties are well understood. 
(c) Its functional anatomy should be 
known— that Is, there should be a clear idea 
of the location within the molecule of: 
(1) The sites at which DNA synthesis origi- 
nates and terminates: 
(ii) The sites that are cleaved by restric- 
tion endonucleases; 
(iil) The template regions for the major 
gene products. 
(d) It should be defective when carrying 
an inserted DNA segment: that is, propaga- 
tion of the recombinant DNA as a virus 
must be dependent upon the presence of a 
complementing helper genome. In almost 
all cases this condition would be achieved 
automatically by the manipulations used to 
construct and propagate the recombinants. 
In addition, the amount of DNA encapsulat- 
ed in the particles of most animal viruses is 
defined within fairly close limits. The inser- 
tion of sizeable foreign DNA sequences, 
therefore, generally demands a compensa- 
tory deletion of viral sequences. It may be 
possible to introduce very short insertions 
(50-100 base pairs) without rendering the 
viral vector defective. In such a situation, 
the requirement that the viral vector be de- 
fective is not necessary. 
If possible the helper virus genome 
should: 
(1) Be integrated into the genome of a 
stable line of host cells (a situation that 
would effectively limit the growth of the 
vector recombinant to such cell lines): or 
(ii) Consist of a defective genome, or an 
appropriate conditional lethal mutant virus, 
making vector and helper dependent upon 
each other for propagation. 
However, neither of these stipulations is a 
requirement. 
The group discussed at length the 
possibility that use of eukaryotic vec- 
tors under these conditions could pose 
a threat to the community or environ- 
ment, but could not envisage a plausi- 
ble set of circumstances whereby a 
risk to the community could develop. 
Given the extent of the genetic dele- 
tion required to satisfy the conditions 
stated above and in table 2, and the 
constraints that encapsidation places 
on the size of an inserted gene seg- 
ment, the workshop participants saw 
no way that the experiments could 
generate a competent virus containing 
new genetic information sufficient to 
code for a functional gene product. 
The group tried to conceive of ways by 
which a recombinantional event be- 
tween the defective recombinant 
genome and the helper genome or the 
genome of an indigenous virus in an 
exposed laboratory worker might gen- 
erate nondefective viral recombinants, 
but as mentioned, was unable to iden- 
tify any. 
The possibility that a defective re- 
combinant genome might be main- 
tained in nature through complemen- 
tation by a helper virus was considered 
too unlikely to be a cause for concern, 
given the absence of any precedent in 
animal virology. The group considered 
the adeno-associated viruses, which 
are defective parvoviruses maintained 
in nature, to be a unique case in that 
they are known to integrate into host 
cells and to have at least 33 different 
potential helpers (human adenovir- 
uses) available. 
Having considered the use of the 
genomes of different DNA animal vir- 
uses to propagate DNA sequences 
from different sources, the group pro- 
posed the following recommendations: 
TABLE 2 
Viral Vector ENA 
polyana virus — all or part 
SV40 — unconditionally 
defective by deletion of 
all or part of the sequences . 
of any of the genes 
adenoviruses 2 and 5 — 
incapacitated by deletion of 
at least two capsid genes 
marine adenovirus strain 
FL — all or part + 
Herpes simplex virus — 
incapacitated by a large 
deletion 
Precaution Level Depending 
Source of Foreign 
on 
prokaryotic 
eukaryotic 
viral 
L or M 
L 
etc 
L or M 
M 
etc 
L or M 
L 
etc 
L or M 
L 
CbC 
etc 
etc 
etc 
FEDERAL REGISTER, VOL. 43, NO. 14&— FRIDAY, JULY 28, 1978 
[ 124 ] 
