Incorporation of integrase-LexA fusion proteins into HIV-1 and analysis of resulting virions, a strategy for achieving site-directed integration in vivo

Retroviruses are leading candidates in the development of gene therapy vectors. This is due, in part, to their ability to stably and precisely introduce a DNA copy of their genome into the chromosomes of an infected cell. A potential pitfall also exists in this reaction because integration occurs nonspecifically with regard to the chromosomal DNA. In vitro , purified fusion proteins made up of a retroviral integrase and a sequence-specific DNA-binding protein, such as the LexA repressor of E. coli, are able to direct integration toward specific sites. To determine whether these proteins can be incorporated into HIV-1 and whether the resulting virions are infectious, an in trans approach was used to deliver various integrase-LexA proteins to a virus encoding a catalytically inactive integrase gene. Integrase-LexA, integrase-LexA DNA-binding domain and N- and C-terminal integrase-LexA were fused to the viral accessory protein, Vpr. Coexpression of the Vpr-fusion proteins with an integrase-defective HIV-1 molecular clone resulted in virus production from packaging cells and efficient incorporation of the fusion proteins into virus particles. Cellular assays, measuring virus infectivity by stable expression of a reporter gene, revealed that each of the integrase-LexA fusion proteins was able to restore integration, and infectivity, to the integrase-deficient viral clone. To examine the mechanism by which the integrase-LexA proteins mediated integration, the virus-host DNA junctions of the integrated viral genomes were sequenced. In proviruses formed by the integrase-mutated virus containing the integrase-LexA protein, the characteristic hallmarks of integration were present, implying that the fusion proteins mediated insertion of the viral DNA like wild-type integrase. In addition, the aspartate 64 mutation was maintained, suggesting that a reversion of the integrase gene to wild type was not responsible for the integration events. Mixed multimers between the catalytically inactive integrase and the fusion proteins in the viruses were also determined to be formed, using complementation experiments. These results demonstrate the feasibility producing site-directed retroviral vectors using the integrase-fusion protein approach. The proteins can be efficiently incorporated into HIV-1, and mediate integration in a manner consistent with the wild-type virus.