Construction ol a Petrov* us PacXagmg KA/tant 
157 
but i/-2 yielded 10 4 cfu/ml (Table 2). Mov-1 and ^-3 
produced 5 x 10 s and 10 s cfu/ml, respectively. Further- 
more, cells rendered gpt * by using supernatants from 
transfected ^-1 or i/-2 cells were producers of neither 
reverse transcriptase nor gpt* cfu and remained negative 
for both after 4 weeks of culture. In contrast, cells rendered 
gpt* using supernatants from transfected MOV-1 or i£-3 
cells were immediately producers of reverse transcriptase 
and gpt* cfu. 
Discussion 
By deleting 350 bp from a M-MuLV DNA clone, we have 
created a mutant, pMOV-^', that cannot itselt be propa- 
gated as a virus, but that still efficiently provides all of the 
obligate trans retroviral functions. Upon transfection of 
doned recombinant retroviruses, such as pMSVgpt. into 
cell lines expressing pMOV-^ - , such as i-2. high titers of 
infectious retroviral particles are produced which contain 
the recombinant genome. These particles are able to 
transmit the recombinant genome efficiently via the retro- 
virus life cyde. In contrast, the genome is not detectably 
packaged into partides and therefore helper virus is not 
produced. Thus the defect in pMOV-^ - appears to be a 
c/s-active defect and probably involves a packaging site 
on the RNA, presumably one recognized by a viral protein, 
rather than a region encoding a protein needed for pack- 
aging. Deletion of sequences in spleen necrosis virus at a 
similar position to that described here has been shown to 
result in a c/s-active defect (Watanabe and Temin, 1982). 
Murine retroviruses often encode a glycosylated form of 
their PreS®* 0 protein that is longer at the N-terminus than 
Pr65 B * B (Edwards and Fan, 1979. 1980). The pMOV-i/r 
done contains a deletion of all of the information that might 
encode such an N-terminaJ extension and yet provides ail 
of the needed helper virus functions. The 350 bp deletion 
also removes all three AUGs that are present between the 
5' LTR and the AUG used for Pr65"*° initiation, ft would 
thus appear that, at least in fibroblasts, the glycosylated 
gag precursor is not needed for efficient budding of 
infective progeny, although this has not been tested di- 
rectly. 
Dimerization and Packaging of Virion RNA 
Electron microscopic studies of 70S virion RNA have 
suggested that a site approximately 300 nudeotides from 
the 5' end of the 35S RNA molecule may be important for 
genome dimerization (Bender et af., 1978). This region is 
within the 350 bp deletion of pMOV-i/r. If dimerization is 
necessary for packaging, then the defect in pMOV-^ - may 
be a consequence of the deletion of the dimerization site. 
Studies by Cheung et ai. (1972) and Canaani et al. (1973), 
however, suggested that the packaging and dimerization 
processes are at least temporally distinct, because they 
found that freshly budded virions of Rous sarcoma virus 
contained only the monomer RNA and that dimer formation 
followed budding. Smaller mutations will have to be con- 
structeo to examine whether packaging and dimerization 
are separable processes. 
Generation of Nondefective Viruses 
When pMOV-\//” DNA is transfected into mouse cells, 
nondefective virus appears after a lag of 4 to 6 days. The 
slow appearance of nondefective virus, and the ease of 
isolating doned cell lines that harbor pMOV-f~ but do not 
generate nondefective virus, suggest that only a minority 
of transfected cells yield nondefective virus. We assume 
that recombination between some cellular sequence and 
pMOV-^“-derived DNA or RNA is responsible for the gen- 
eration of nondefective virus, but we have yet to provide 
direct evidence for this. Likely candidates for providing the 
packaging site sequence are the family of abundant, 
packagable, virus-like 30S RNAs found in mouse cells 
called VL30 RNAs (Sherwin et al.. 1978; Besmer et al.. 
1979; Scolnick et al., 1979) and the many retrovirus-like 
DNA elements found in normal mouse cells (reviewed in 
Coffin, 1982). We are currently analyzing the structure of 
the nondefective viruses as well as determining whether 
these genomes can be generated in other cell types. 
Use of i-2 
In a number of viral systems it has proved useful to design 
cell lines capable of providing certain viral functions. Such 
cell lines can then be used as hosts for growth of viruses 
defective for production of the proteins already present in 
the cells. For instance, the hr-t mutants of polyoma virus 
were isolated in this way (Benjamin, 1970), early mutants 
of SV40 can be grown in COS cells (Gluzman, 1981), and 
early mutants of adenovirus can be grown in 293 cells 
(Graham et al., 1977). The i-2 cell line described here is 
an analogous cell line for defective retroviruses. With such 
a line, stocks of natural or engineered defective retrovi- 
ruses, such as pMSVgpt and others (Mulligan, 1983), can 
be obtained totally free of detectable helper virus. This tool 
enhances the utility of retroviral vectors (Wei et al., 1981; 
Tabin et al., 1982; Mulligan, 1983) because pure stocks of 
the engineered viruses can be used to introduce genes 
into cells via the retroviral life cyde without the recipient 
cell becoming a retrovirus producer. Furthermore, genes 
doned into retroviral genomes could, in principle, be intro- 
duced into the germ line of mice without also introdudng 
helper virus, in much the same way as has been demon- 
strated for MuLV (Jaenisch, 1976, 1977; Jaenisch et al., 
1981), thus avoiding side effects inherent in having helper 
virus present. 
Experimental Procedures 
Cede and Viruses 
The NH/3T3 cel line was grown in CXXbecco's modfied Eagle mecksn 
contanng 10% cart serum. The MOV-1 cel Ime was derived by transfection 
ol pMOV-f* (actuaty pMOV-9. a gift from R. Jaervsdi) nto f*t/3T3 cels 
tdowed by dormg. Cels to be cioned were trypsrxzed. counted, and 
seeded nto 96-wel dormg trays al 0.3 cels per wel (in a voAme of 0.2 
mi per wel) and alowed lo grow for 10 to 12 days. Selection medaxn tor 
gpC was as descrtied (Mi*gan and Berg. 1961). Vrus infections were 
performed n the presence of 8 <ifl/m I potytxene tor 1-2 hr. 
Reverse Transcriptase Assays 
Reverse transenptase assays were performed as described (Gotf et al.. 
1981 ) Subconfloent cels were ted lor 12 hr before beng assayed to 
maxnxze vrus production 
[ 356 ] 
Recombinant DNA Research, Volume 12 
