Milter rt «L 
rus to infect cells from different species, or 
host range, is dependent on several factors, the 
most important of which is the ability of the 
viral envelope proteins to bind cellular recep- 
tors required for entry of virions into cells. 
Entry of a virus into a cell is unaffected by 
previous infection of the cells with a helper 
virus which uses a different receptor for entry 
into the cell. In contrast, infection of cells with 
virus is severely inhibited by previous infec- 
tion of the cells with helper virus of the same 
pseudotype, or protein coat, due to occupation 
of viral receptors by endogenously synthesized 
envelope glycoprotein. Packaging cell lines 
that can produce virus having either ecotropic 
or amphotropic host range are available (2, 4, 
18, 20) and viruses having these pseudotypes 
use different receptors for entry into cells 
(29). Thus, although it should be difficult to 
infect packaging cells with virus generated by 
the same packaging cell line due to viral 
interference, it should be possible to shuttle 
retroviral vectors between packaging cell lines 
with different host range. 
Our technique for producing ecotropic or 
amphotropic helper-free retrovirus is dia- 
grammed in Figure 1, and involves only one 
cell-cloning step. Although the initial DNA 
transfection step may lead to a mixture of 
desired and rearranged viruses, isolation and 
analysis of infected cell clones in the last step 
insure isolation of a cell line producing homo- 
geneous virus of the desired genotype. We 
have used Psi-2 cells (2) as the ecotropic 
packaging line. PA 12 cells (4) were chosen 
from the available amphotropic lines (4, 18, 
20) because they produce the highest virus 
titers. Test vectors (Figure 2) included a 
Moloney murine leukemia virus-based retro- 
virus that expresses the neomycin resistance 
(Neo) gene (30), contained in the plasmid 
pN2 (provided by Dr. Eli Gilboa), and a 
spleen focus-forming virus derivative that 
expresses a mutant dihydrofolate reductase 
(DHFR) gene that confers resistance to 
methotrexate (31), contained in the plasmid 
pSDHT. These genes are transcribed using 
the transcription signals present in the viral 
LTRs. 
Table 1 shows the results of an experi- 
ment designed to test the assumptions made in 
our scheme for production of helper-free 
virus. Virus generated following transfection 
of the packaging lines was used to infect 
various cells. The day after infection, the 
medium was changed to selective medium, 
and five days later resultant drug-resistant 
colonies were counted, a measurement of the 
apparent titer of the virus stock on each cell 
type. We first checked whether virus rescued 
from the packaging cell lines had the expected 
behavior when used to infect cells already 
infected with ecotropic or amphotropic helper 
virus. The data confirmed that virus gener- 
ated from Psi-2 cells, which should have an 
ecotropic pscudotype, had greater than 300- 
fold reduced ability to infect cells already 
infected with the ecotropic virus Mo-MLV, in 
comparison to uninfected cells. Similarly, 
virus generated from PA 12 cells, which 
should have an amphotropic pseudotype, had 
at least 300-fold reduced ability to infect cells 
infected with the amphotropic virus AM- 
MLV, compared with uninfected cells. In 
contrast, infection of helper virus-infected 
cells with virus having a different pseudotype 
resulted in transfer of the selectable genes 
with an efficiency similar to that obtained 
using uninfected cells as recipients for virus 
infection. Thus viral interference occurs as 
expected in the case of infection of helper 
virus-infected cells with vectors produced 
from retrovirus packaging cell lines. 
Interference with virus infection was also 
observed with infection of the packaging lines 
(Table 1). Virus generated from Psi-2 cells 
was at least 300-fold less efficient at infecting 
Psi-2 cells in comparison to NIH 3T3 cells. 
Virus generated from PA 1 2 cells was three- to 
fivefold less efficient at infecting PA 12 cells, 
in comparison to NIH 3T3 cells. In PA 12 cells 
the block to infection by virus with the same 
pseudotype was not as severe as in the other 
cell lines tested, and this may be explained by 
[ 362 ] 
Recombinant DNA Research, Volume 12 
