MILLER AND BUTTIMORE 
Mol. Cell. Biol. 
I SELECT CELLS IN 
G-4H PLUS G-TMIOGUANINE 
FIG. 3. Scheme for assay of p»cir« r n l iy»tem transfer. 
with rat cells allowed transfer to occur directly between cells 
and therefore increased the sensitivity of the assay. 
Using this cocultivation assay, we found that medium 
horn Psi-2 cells consistently induced low-level Neo virus 
release from NIH 3T3-N2 cells (Table 5). Medium from 
PA12 cells or PA12 cells producing DHFR virus also induced 
Neo virus release but at even lower levels. We did not detect 
colony production using medium from PA317 cells or DHFR 
virus-producing PA317 cells. We ruled out endogenous 
mouse viruses as being responsible for Neo virus rescue in 
these experiments because these viral genomes were present 
in all of the packaging lines, whereas not all induced Neo 
virus production. Retroviruses can induce cell fusion, but 
colony formation as a result of cell fusion was ruled out 
because fusion products between the HPRT” rat and G-418- 
resistant mouse cells would still be sensitive to 6- 
thioguanine, since the mouse cells were HPRT*. In addition, 
we ruled out colony formation as a result of cell fusion 
because virus from the packaging cell lines should induce 
fusion at similar rates, and some lines did not induce colony 
formation. We concluded that deletion of the packaging 
signal was not sufficient to prevent transfer of the packaging 
function. However, additional alterations in the packaging 
system in PA317 cells reduced packaging function transfer to 
undetectable levels. 
DISCUSSION 
We made a systematic series of deletions in a replication- 
competent retrovirus in an attempt to make a nontransmis- 
sible virus which still provides rranr-acting factors required 
for packaging and transmission of retroviral vectors. These 
deleted constructs were introduced into cells by cotransfec- 
tion with a selectable marker to make various retrovirus 
packaging cell lines. Packaging cells made using the most 
heavily deleted construct, pPAM4, did not produce virus 
vectors at high titer. Packaging cell lines made using an 
intermediate construct. pPAM3, did produce virus vectors at 
high titer (10* to 10 7 CFU/ml). The packaging construct 
pPAM3 was made by deletion of the retroviral packaging 
signal, all of the 3' LTR, part of the 3' LTR, and the site for 
second-strand DNA synthesis. Thus, RNA transcribed from 
pPAM3 should have been inefficiently packaged into virions, 
and reverse transcription and integration of any resultant 
virus should have been impossible. Indeed, we were unable 
to detect packaging function transfer from PA317 cells' 
containing pPAM3. 
Others have attempted to make packaging lines analogous 
to PA317 cells. A packaging cell line that contains a virus 
with deletions of the origin of second-strand DNA synthesis 
as well as the packaging signal has been previously described 
(24), but vector titers from this line (10 3 CFU/ml) are low. A 
packaging system with a limited host range that is based on 
avian reticulocndotheliosis virus has been described before 
(27). The gag and pol proteins on the one hand and the env 
protein on the other are synthesized from separate DNA 
constructs. The packaging signal was also removed from 
both constructs. However, helper virus production or pack- 
aging system transfer from this line was not extensively 
analyzed. 
We did not observe helper virus production from cells 
transfected with any of the packaging constructs. This 
includes the construct pPAM, in which only the retroviral 
packaging signal was removed. This is similar to results 
obtained by other groups using constructs similar to pPAM 
(3, 24). In experiments leading to construction of the Psi-2 
packaging cell line (13), helper virus production from the 
packaging signal-deleted helper virus used was very fre- 
quent; however, a later report suggested that this was 
probably due to the presence of a small amount of wild-type 
helper virus plasmid in the construct plasmid stock, as other 
preparations of the construct did not lead to helper virus 
production (12). In addition, we have never observed helper 
virus production from PA12 cells carried in the laboratory 
for many months, as measured by S*L“ assay of the cells 
and their derivatives containing various retroviral vectors. 
TABLE J. Detection of Neo virus rescue from cells following 
infection with medium exposed to packaging cell lines 
Test cells 
No. of cokxucs induced 
Exp* 1 
Exp* 2 
NIH 3T3 (TK-) 
0 
0 
Psi-2 
18 
10 
PA 12 
3 
2 
PA12-SDHT 
0 
1 
PA317 
0 
0 
PA317-SDHT 
0 
0 
Recombinant DNA Research, Volume 12 
[ 373 ] 
