homology with human retroviral sequences. The retroviral genome can be divided 
Into two areas: the LTRs required for the high efficiency integration of the 
viral genome, and the gag , pol and env regions required for virus replication 
and virion formation after the genome has integrated. By deleting the gag , pol 
and erw regions the SAX vector can integrate but it is left without the 
capacity to replicate. The producer cell line is used to make the vector in a 
system that is theoretically helper virus free. Since the incidence of 
recombination between mutant retroviruses after coinfection is estimated to be 
0.1 to 8% (Wong and McCarter. 1973; Stephenson et al.. 1974a; Aaronson and 
Barbacid, 1980). the chance for recombination between the inserted vector and 
any helper virus may be high. Indeed, recombination between a replication 
defective virus and helper virus, restoring the replicative ability of the 
defective virus, has been shown (Goldfarb and Weinberg 1981; Goff et al., 1982). 
We assume, as suggested by Miller, that the 0.12 contamination arising in our 
S3A preparations is the result of the homology between the 5' ends of the 
packaging (-) helper virus in PA-12 and the packaging (+) vector N2 or SAX. 
Apparently, the helper virus has acquired the packaging sequence from the 
vector. 
The risks generated by recombination will be considered in detail below 
but the potential problems that could ensue from recombination include 
continued viremia with infection of the germ line, increased incidence of 
malignancy by promoter/enhancer insertion, or recombination with endogenous 
proto-oncogenes and the creation of an acutely transforming retrovirus that 
could theoretically be transmissible. By eliminating helper virus from the 
retroviral vector supernatant, the replication-competent sequences contained in 
the helper virus are not available for recombination and theoretically the risk 
of the above mentioned complications would be greatly reduced. Recombination 
with endogenous human retroviruses remains a finite possibility (see below). 
Recombinant DNA Research, Volume 12 
[501 
