APPENDIX: Evidence presented by Life Technologies to NIH RAC regarding the 
biosafety of SFV and derivative vector systems. 
A. Biosafety of the SFV/Helper 2 Gene Expression System 
1. RATIONALE: 
In the experiments described below we examine first the effectiveness of employing helper 
vectors as a means to prevent emergence of replication-competent virus. Using a model system 
consisting of SFV3 RNA plus Helper 1 RNA (see p. 22), we measure the frequency of 
recombinants between helper RNA and vector RNA subgenomes. A low level of replication- 
competent virus is found. 
Therefore a second biosafeguard is created, based on mutations that block the processing of coat 
proteins needed for virus infectivity. To evaluate the effectiveness of this safeguard, a model 
vector is constructed wherein mutations are introduced into the genes encoding SFV coat 
proteins in a replication-competent SFV. Our analysis of this SFV variant, SFV-SQL, reveals 
a very low level of infectious revertants. 
Finally, both biosafeguards are combined into one system: Incorporating the SQL mutation into 
Helper 1 yields the SFV/Helper 2 expression system. When this dual system is analyzed, both 
in tissue culture and in animals, no replication-competent virus is detected. 
2. RECOMBINATION BETWEEN HELPER 1 AND PSFV3-LACZ RNAS 
Use of helper-dependent packaging is an effective means to control the production of infectious 
virus. This safeguard could be subverted, however, if the helper vector and recombinant vector 
(carrying the sequences desired to be expressed) were to recombine to restore an intact viral 
genome. In fact, wild-type recombinant alphavirus genomes can be generated from double 
infections with two defective genomes (Weiss and Schlessinger, 1991). 
It was of interest, therefore, to analyze whether this also applies to the SFV in vivo packaging 
system. Recombinant virus stocks comprising SFV3-lacZ and Helper 1 were analyzed by 
indirect immunofluorescence screening of infected cells for the occurrence of E2 expression, 
using a polyclonal antibody directed against E2 protein. E2 protein would be expressed in these 
cells only if a recombination event had occurred between the recombinant and helper genomes, 
or if both RNAs had been packaged into the same virus particle. A double infection by 
recombinant and helper virus particles was minimized by using a low multiplicity of infection. 
When a total of 10 7 cells was screened, using two different stocks of SFV3-lacZ (MOI = 0.1), 
no wild-type particles were found, suggesting that recombination and/or copackaging had 
occurred in fewer than 1 in 10 6 first-generation virus particles. 
Recombinant DNA Research, Volume 18 
[445] 
