MILLER AND BUTT1M0RE 
Mol. Cell. Biol. 
HPRT* virus plasmid pLPL2 (see Fig. 2). On day 3, the 
transfected cells were fed. On day 4, medium from the 
transfected cells was removed and centrifuged at 3,000 x g 
for 3 min to remove cells and debris, and samples were 
analyzed for HPRT virus by using HPRT“ rat 208F cells and 
for helper virus by using the S*L" assay. 
Cocultlvation assay for packaging function transfer. NIH 
3T3 cells nonproductively infected with N2 Neo virus (NIH 
3T3-N2 cells) were made by infecting NIH 3T3 cells with 
helper-free N2 virus from PA317-N2 cells (see Table 3) and 
selecting G-418-resistant cells. Amphotropic murine leuke- 
mia virus (AM-MLV) (see Fig. 1) infection of clonal NIH 
3T3-N2 cells resulted in production of 10* to 10 7 Neo virus 
per ml of medium exposed to the cells, showing that Neo 
virus could be efficiently rescued. Cell lines to be tested for 
production of virions containing the packaging system were 
seeded at 10* cells per 60-mm dish on day 1. On day 2, NIH 
3T3-N2 cells were seeded at 10 5 per 60-mm dish, and the test 
cells were fed. In the morning of day 3, 12 h after the test 
cells had been fed, medium from the test cells was removed 
and centrifuged at 3,000 x g for 5 min to remove cells and 
debris, and the NIH 3T3-N2 cells were infected with 1-ml 
samples in the presence of 4 p.g of polybrene per ml. The test 
cells were fed, and the NIH 3T3-N2 cells were again infected 
12 h later. Infected NIH 3T3-J42 cells were fed on day 4, and 
on day 5 they were trypsinized, divided into two samples, 
and added to 60-mm dishes containing 8 x 10 3 208F cells 
each. On day 6, the cocultivated cells were confluent and 
were trypsinized and divided 1:10. On day 9, after a total of 
4 days of cocultivation, the cells were split 1:10 into 10O-mm 
dishes in medium containing 2 mg of G-418 (ca. 50% active) 
per ml and 20 pM 6-thioguanine. Cells were fed with 
selective medium every 3 days, and colonies were stained 
and counted on day 14. NIH 3T3-N2 cells die in medium 
containing 6-thioguanine, and 208F cells die in G-418, so the 
only cells that survived were 208F cells infected with Neo 
virus rescued from the NIH 3T3-N2 cells. 
RESULTS 
Construction of mutants. Elements required for retrovirus 
replication can be divided into cis- and rro/is-acting factors. 
The iraru - acting factors include proteins encoded by the 
viral genome, which are required for encapsidation of the 
viral RNA, entry of virions into cells, reverse transcription 
of the viral genome, and integration of the DNA form of the 
virus into host DNA. The cis - acting factors include signals 
present in the viral RNA which interact with these proteins 
and other factors during virus replication. To make a retro- 
virus packaging cell line with the lowest propensity for 
generating replication-competent virus, we made alterations 
in the viral genome which should interfere with cir-acting 
elements while preserving production of rra/u-acting fac- 
tors. A cell line containing such an altered viral genome 
should not transmit this virus but would transmit other viral 
RN As containing the proper cu-acting elements, including 
retroviral vectors designed to carry foreign genes. 
We chose the replication-competent amphotropic retrovi- 
rus AM-MLV (Fig. 1) for production of mutant viruses (Fig. 
1) because of the broad host range of this virus, which 
includes mice, rats, chickens, cats, dogs, monkeys, and 
humans (8. 15, 20). The irons - acting factors encoded by this 
murine retrovirus include the gag and pol proteins, which 
are translated from unspliced viral genomic RNA, and the 
env protein, which is translated from a spliced message. 
Previous work has shown that a region between the 5' splice 
Recombinant DNA Research, Volume 12 
site and the initiator codon of the gag protein is required for 
efficient packaging of retroviral RNA into virions (13, 23). 
This packaging signal was removed from pAM-MLV to 
make pPAM. In pPAM2 the packaging signal and viral 
sequences 3' of the env protein terminator codon have been 
removed. The simian virus 40 late-region polyadenylation 
signal was added at the 3' end to provide for polyadenylation 
of viral mRNAs. The deletion at the 3' end of the virus 
removes the site for initiation of second-strand DNA syn- 
thesis and the 3' R region that is required for translocation of 
reverse transcriptase during first-strand DNA synthesis (26). 
In addition to these changes, pPAM3 has a deletion of the 5' 
end of the 5‘ long terminal repeat (LTR); thus, a proper 
integration signal cannot be made from the remaining se- 
quences and this should prevent virus integration (18). In 
pPAM4, all of the above deletions have been made and the 3' 
end of the 5' LTR and the tRNA primer-binding site have 
been removed. The viral promoter and the splice donor used 
to make the env mRNA were left intact in pPAM4. Deletion 
of the tRNA-binding site should inhibit first-strand DNA 
synthesis from the viral genome (26). 
Generation of retrovirus packaging cell lines. We intro- 
duced the DNA constructs depicted in Fig. 1 into NIH 3T3 
TK‘ cells by cotransfection using the herpes simplex virus 
thymidine kinase gene (2) as a selectable marker. TK* 
colonies were isolated and screened for production of repli- 
cation-competent virus by using the S*L" assay and for 
their ability to package a standard retrovirus vector contain- 
ing HPRT (Fig. 2). We were unable to detect helper virus 
production from any of the clones analyzed (>70 clones) by 
using the S*L" assay with a limit of sensitivity of 1 virus per 
ml. At least 50% of the clonal cell lines made with any one of 
the packaging constructs were able to package HPRT virus 
(Table 1), suggesting that rearrangement of the transfected 
DNA was not required to obtain such lines. Roughly equiv- 
alent virus titers were obtained from cell lines containing 
pPAM, pPAM2, and pPAM3; however, cell lines made with 
pPAM4 were about 10-fold less efficient in packaging HPRT 
virus (Table 1). Since virus titer is critical for many applica- 
tions of retroviral vectors, we used the best pPAM3 
transfectant, PA317, in further studies instead of the best 
pPAM4 transfectant, PA405. The deletions made in pPAM3 
should be sufficient to severely reduce packaging of RNA 
derived from this construct into virions, and even if the RNA 
is packaged, provirus formation in infected cells should be 
blocked both at the level of reverse transcription of the RNA 
to DNA and at the level of virus integration into the host 
genome. 
Comparison of virus Liters from vector-infected packaging 
ceil lines. The titers of virus produced transiently following 
transfection of PA12 or PA317 cells with HPRT virus were 
similar, indicating that PA317 cells efficiently package retro- 
virus vectors, as do PA12 cells. To further test the PA317 
line, we infected PA317 cells with a virus containing a 
dominant-acting, selectable DHFR gene (Fig. 2), assayed 
individual infected clones for production of DHFR virus, 
and compared these results with those obtained using PA12 
cells (Table 2). The DHFR virus titer produced from DHFR 
virus-infected PA317 clones was very high, up to 10 7 DHFR 
virus per ml of medium with no detectable helper virus. The 
PA317 clones on average produced slightly higher-titer 
DHFR virus than the PA12 clones, and the best clone was 
slightly better than the best PA12 clone. We concluded that 
the additional mutations present in the packaging DNA in 
PA317 cells did not adversely affect our ability to make 
high-titer-vector-producing cell lines. 
[ 369 ] 
