MIOcrrtftL 
between vector and the packaging system 
DNA 3' of the packaging signal is important 
in order to avoid helper virus generation from 
retrovirus packaging cell lines. 
For some purposes, virus produced tran- 
siently following transfection of packaging 
lines is suitable, and such virus can be gener- 
ated in two days at titers of from about 5 x 
10 1 to 2 x 10 s CFU/ml, depending on the 
vector and the variability inherent in DNA 
transfection. Such virus can be used to test the 
characteristics of the virus construction in 
appropriate test cells, for example, cells in 
which regulation of a gene inserted in the 
vector is expected. However, there is a greater 
chance of obtaining drug-resistant infected 
clones which contain rearranged viruses. 
We note that when a large amount of 
medium that has been exposed to either PA 12 
or Psi-2 cells is applied to cells that have 
available receptors for the virus, fusion of the 
infected cells occurs, as evidenced by the 
appearance of large numbers of multinucleate 
cells. For example, if 1 ml of medium exposed 
to Psi-2 cells for 16 h is used to infect PA 12 
cells seeded the previous day at 5 x 10 5 / 
60-mm dish, much of the cell layer fuses 
within a day. Fusion of cells can result in 
extensive cell death, but can be minimized by 
using less virus-containing medium for infec- 
tion, or by plating the cells to be infected at 
low density so that the cells are too far apart to 
interact. This effect should be considered 
when generating vector packaging clones. 
In the process of making vector packag- 
ing lines from many vectors using the strategy 
outlined in Figure 1 A, we have found that the 
titer of the virus expressed transiently follow- 
ing transfection is a good indicator of the 
usefulness of the vector. If the titer is high, we 
note that the vector packaging lines derived 
from such virus secrete unrearranged virus at 
titers exceeding 10* and will do so for many 
months in continuous culture. If the titer is 
low (less than about 10 3 CFU/ml), high-titer 
vector-producing packaging lines can be pro- 
duced, but these lines almost universally con- 
tain rearranged viruses (unpublished data). 
This is true of vectors containing one or sev- 
eral genes. For example, transfection of the 
DHFR-virus construct pLDLl into PA 12 
cells resulted in transient virus production of 
about 100 CFU/ml (4). This titer is low in 
comparison to vectors used here, and we infer 
that either expression of the mutant DHFR 
gene in this virus is not sufficient or that the 
arrangement of the virus results in poor repli- 
cation. While we were able to generate high- 
titer helper-free DHFR-virus from pLDLl 
(4), the integrated viruses in all of these clones 
are rearranged (unpublished data). A likely 
explanation for this phenomenon is that the 
virus stock used to infect the packaging cells 
contains rearranged viruses having increased 
ability to replicate or express the inserted 
gene, and these viruses are selected due to 
their increased ability to confer drug resis- 
tance to infected cells. This is the reason that 
we have used the DHFR-virus SDHT in 
experiments reported here instead of the vec- 
tors that we previously developed (4). We are 
currently trying to determine why SDHT is a 
better vector for transfer of the mutant DHFR 
gene than LDL1. 
Viral interference can be minimized by 
using inhibitors of glycosylation, in particular, 
tunicamycin (36). This effect is due to loss of 
viral envelope glycoproteins from the cell sur- 
face with the result that viral receptors 
become available for interaction with superin- 
fecting viruses of the same interference group. 
Thus an alternative to the technique described 
here for infection of retrovirus packaging 
lines, by using two packaging lines with dif- 
ferent host ranges, might involve infection of 
tunicamycin-treated packaging cells with 
virus produced by transfection of cells of the 
same packaging line. We expect that this 
alternative scheme would be more difficult to 
implement! due to the toxicity of tunicamycin 
at concentrations required for reversal of viral 
interference and to the possibility of incom- 
plete reversal of interference (36). 
A variety of retroviral vectors have been 
[366] 
Recombinant DNA Research, Volume 12 
