Results of in vitro assays identified residues in the C-terminal domain (CTD) of human immunodeficiencyRetrovirus replication is dependent on the integration of the reverse-transcribed viral genome into a host chromosome. Subsequent to target cell entry, the double-stranded DNA substrate for integration is generated by the viral enzyme reverse transcriptase (RT) upon conversion of the genomic RNA into DNA. Acting on the attachment (att) sites at the cDNA ends, the viral DNA recombinase or integrase (IN) catalyzes two distinct endonucleolytic reactions. For the first reaction, 3Ј processing, human immunodeficiency virus type 1 (HIV-1) IN removes the dinucleotide GT from each end. This exposes a 3Ј hydroxyl moiety in preparation for the second reaction, DNA strand transfer. Upon recognition and binding to a suitable target site, IN uses the 3Ј-OHs to cut the chromosome in a staggered fashion, which at the same time joins the viral ends to the 5Ј-phosphates of the cut. Cellular enzymes are likely involved in the repair of the resultant gapped product, thus fully recombining the viral cDNA with the host (reviewed in references 16 and 38).

Retroviral replication requires the accomplishment of certain key steps in the early phase of the viral life cycle. Soon after entering a cell, the viral enzyme reverse transcriptase (RT) copies genomic RNA into linear double-stranded cDNA. The viral enzyme integrase (IN) then inserts this DNA into a host cell chromosome. In vivo, reverse transcription and integration take place in the context of large nucleoprotein complexes that are called reverse transcription complexes (RTCs) and preintegration complexes (PICs), respectively (reviewed in reference 29).Two different IN activities are required for integration. During an initial 3Ј processing reaction, IN cleaves each cDNA end adjacent to the phylogenetically conserved sequence CA. For both Moloney murine leukemia virus (MoMLV) (7,28,45) and human immunodeficiency virus type 1 (HIV-1) (12, 13, 39), 3Ј processing can occur in the cell cytoplasm. Following nuclear entry, IN transfers the processed 3Ј ends to the 5Ј phosphates of a double-stranded staggered cut in chromosomal DNA (7, 28). The final step of the integration process, which involves repairing the gaps between the unjoined viral 5Ј ends and the chromosome, can be completed by host cell enzymes (4, 54).The double lipid bilayer that envelops animal cell nuclei presents a formidable barrier for retroviral PICs. The nuclear envelope contains numerous nuclear pore complexes (NPCs) that permit the passive diffusion of relatively small macromolecules with diameters up to about 9 nm, which corresponds roughly to a 40-to 60-kDa globular protein (reviewed in reference 37). The relatively large size of retroviral PICs, estimated to be roughly the size of a ribosome for HIV-1 (25), precludes their passive transport through intact NPCs (reviewed in reference 27). Different retroviruses have apparently evolved different strategies to access the cell nucleus. HIV-1, for example, can be transported by an energy-dependent process in nondividing cells (9), suggesting that HIV-1 PICs contain specific nuclear localization sequences (NLSs) that govern their transport through intact NPCs (reviewed in references 27 and 52). Productive infection by MoMLV, in contrast, requires cells to pass through the M phase of the cell cycle (36, 44). Since animal cell nuclei disassemble during mitosis, MoMLV apparently reaches cell chromosomes in the absence of active nuclear transport by waiting for the dissolution of nuclear membranes.Recently, replication-defective MoMLV (55) and HIV-1 (3, 57) mutants were described as blocked at the nuclear import step in rapidly dividing cells. This suggested that retroviruses

Numerous factors have been implicated in the nuclear localization of retroviral preintegration complexes. Whereas sequences in human immunodeficiency virus type 1 (HIV-1) matrix, Vpr, and integrase proteins were initially reported to function specifically in nondividing cells, other recently identified sequences apparently function in dividing cells as well. One of these, the central DNA flap formed during reverse transcription, is specific to lentiviruses. It was previously reported that flap-negative (F−) HIV-1LAI was completely defective for viral spread in the MT-4 T-cell line, yet F− HIV-1 vectors were only 2- to 10-fold defective in various single-round transduction assays. To address these different findings, we analyzed the infectivity and nuclear localization phenotypes of two highly related T-cell-tropic strains, HIV-1NL4-3 and a derivative of HIV-1HXBc2 deficient for both Vpr and Nef. In stark contrast to the previous report, F− derivatives of both strains replicated efficiently in MT-4 cells. F− HIV-1NL4-3 also spread like wild-type HIV-1NL4-3 in infected Jurkat and primary T-cell cultures. In contrast, F− HIV-1HXBc2 was replication defective in primary T cells. Results of real-time quantitative PCR assays, however, indicated that F− HIV-1HXBc2 entered primary T-cell nuclei as efficiently as its wild-type counterpart. Thus, the F− HIV-1HXBc2 growth defect did not appear to correlate with defective nuclear import. Consistent with this observation, wild-type nef restored replication to F− HIV-1HXBc2 in primary T cells. Our results indicate that the central DNA flap does not play a major role in either preintegration complex nuclear import or HIV-1 replication in a variety of cell types