Generation of HIV-1 derivatives that productively infect macaque monkey lymphoid cells

Abstract

The narrow host range of human immunodeficiency virus type 1 (HIV-1) is caused in part by innate cellular factors such as apolipoprotein B mRNA-editing enzyme-catalytic polypeptide-like 3G (APOBEC3G) and TRIM5α, which restrict virus replication in monkey cells. Variant HIV-1 molecular clones containing both a 21-nucleotide simian immunodeficiency virus (SIV) Gag CA element, corresponding to the HIV-1 cyclophilin A-binding site, and the entire SIV vif gene were constructed. Long-term passage in a cynomolgus monkey lymphoid cell line resulted in the acquisition of two nonsynonymous changes in env, which conferred improved replication properties. A proviral molecular clone, derived from infected cells and designated NL-DT5R, was used to generate virus stocks capable of establishing spreading infections in the cynomolgus monkey T cell line and CD8-depleted peripheral blood mononuclear cells from five of five pig-tailed macaques and one of three rhesus monkeys. NL-DT5R, which genetically is >93% HIV-1, provides the opportunity, not possible with currently available SIV/HIV chimeric viruses, to analyze the function of multiple HIV-1 genes in a broad range of nonhuman primate species.

The narrow host range of human immunodeficiency virus type 1 (HIV-1) has been a major impediment for developing tractable animal models for studies of viral pathogenesis and vaccine development. Because simian immunodeficiency virus (SIV) has a genomic organization similar to that of HIV-1 and some SIV strains cause disease in Asian macaques, SIV/HIV chimeric viruses (SHIVs) were generated to assess the role of some HIV-1-encoded proteins in nonhuman primates (1–3). The commonly used SHIVs contain the HIV-1 tat, rev, vpu, and env genes inserted into an SIVmac239 genetic backbone; efforts to extend the incorporated HIV-1 gene segment to include pol and gag sequences have resulted in viruses unable to replicate in monkey cells (ref. 1; unpublished data). Although SHIVs have proven useful in characterizing the immune responses to primate lentiviruses (4, 5), and specifically, the role of antibodies directed against the HIV-1 envelope glycoprotein (6, 7), the absence of the other HIV-1 structural proteins has restricted analyses of their function in vivo.

It is now appreciated that many mammalian species encode factors conferring resistance to retroviral infections. Some, such as the apolipoprotein B mRNA-editing enzyme-catalytic polypeptide-like 3G (APOBEC3) family of cytidine deaminases, modify minus strand viral DNA during reverse transcription, resulting in either its degradation or its integration into host chromosomal DNA as a hypermutated provirus (8–10). The retroviral inhibitory effect of APOBEC3G results from its packaging into progeny virions during particle assembly (11–12). The deleterious activities of APOBEC3G are countered by lentiviral Vif proteins, which prevent the encapsidation of APOBEC3G into nascent virions (13–16). The sensitivity of APOBEC3G from different animal species to the Vif proteins expressed by different viruses varies widely. For example, although HIV-1 Vif can potently suppress human APOBEC3G, it is not effective against rhesus monkey (RhM) APOBEC3G, explaining in part the restriction of HIV-1 replication in macaque cells (11).

Another recently described restriction factor, TRIM5α, targets incoming viral capsids, and it blocks retroviral replication in a species-specific manner (17–19). For example, TRIM5α from RhMs potently suppresses HIV-1 but not SIV infectivity in monkey cells (19). Although its mechanism of action is still unclear, TRIM5α restriction is thought to affect virus uncoating, thereby blocking subsequent steps in the replication cycle (20). Cyclophilin A (CypA), which binds to a proline-rich loop on the surface of the HIV-1 capsid (CA) protein, augments HIV-1 infection in human cells and inhibits its replication in monkey cells (21, 22). Recent reports suggest that by binding to HIV-1, CypA may modulate the conformation of the virion core, rendering it sensitive to TRIM5α restriction in simian cells (23, 24).

In this work, we have generated HIV-1 derivatives, which carry only the SIVmac239 vif gene and a short 7-aa segment from SIV gag corresponding to the HIV-1 CypA-binding loop. Molecularly cloned viruses bearing these two SIV regions are able to establish spreading infections in a cynomolgus monkey (CyM) T cell line and CD8-depleted PBMCs from pig-tailed macaques (PtMs) and RhMs. These results indicate that the incorporation of two SIV gene segments into the HIV-1 genome can effectively counter two known species-specific restriction factors that block virus replication in monkey cells. They raise the possibility of generating HIV-1 derivatives, containing all of its structural proteins and capable of infecting macaque monkeys.

Results

Construction and Characterization of HIV-1 Molecular Clones Containing CA and Vif Sequences from SIVmac239.

Three proviral DNA constructs were generated to counteract the restriction of HIV-1 replication in macaque monkey cells. In the first, the entire 214-aa Vif ORF from SIVmac239 was amplified by PCR and inserted into SmaI–XbaI-digested pNL-SX, a pNL4-3-derived vector, previously used for functional analyses of HIV-1 vif genes (25). This SIV Vif-encoding construct was designated NL-SV (Fig. 1). Because of the reported association of CypA with HIV-1 sensitivity to TRIM5α during infections of cells from Old World monkeys (21, 22), the 9-aa CypA-binding loop in NL4-3 was converted to the 7-residue SIVmac239 CA analog by site-directed mutagenesis of the pNL-SX vector carrying the HIV-1 vif gene (26). This construct was designated NL-Sca (Fig. 1). A final clone, containing both SIV elements, was generated by inserting SIV vif into NL-Sca and designated NL-ScaV (Fig. 1).

Fig. 1.

Structure of chimeric clones between HIV-1 NL4-3 and SIVmac239 in this study. Eight chimeric proviral clones shown here were generated from a pNL4-3 derived vector pNL-SX (25) as described in Materials and Methods. Each chimeric clone has the entire vif (black area) and/or partial gag (gray area; analog of HIV-1 CypA-binding loop) of SIVmac239 as shown. For insertion of vif into the clones, the SmaI–XbaI site in pNL-SX was used. The two amino acid changes in env unique to pNL-DT5 and pNL-DT5R are indicated.

Expression of the lentiviral genes present in the three newly derived cloned proviruses was assessed by immunoblot analyses of lysates prepared from transfected 293T cells. The production of Gag, Pol, Env, Vpu, and Nef proteins directed by all three constructs was comparable with that observed with the parental pNL4-3; levels of Vpr expression, however, were markedly reduced (data not shown). The latter was subsequently shown to be caused by the presence of the TCT trinucleotide, introduced into the pNL-SX vector to generate the XbaI cloning site (25). When the TCT was removed by site-specific mutagenesis, Vpr expression was restored to wild-type levels in cells transfected by all three constructs (data not shown). The “Xba site-repaired” clones were designated NL-SVR, NL-ScaR, and NL-ScaVR, respectively, as indicated in Fig. 1.

HIV-1 Constructs Bearing the SIV vif Gene Are Able to Suppress the Inhibitory Effects of Simian APOBEC3Gs.

It has been previously reported that RhM and African green monkey (AGM) APOBEC3Gs are resistant to HIV-1 vif, possibly explaining, in part, the restriction of HIV-1 replication in cells from Old World monkeys (11). To determine whether the simian APOBEC3Gs could block HIV-1 constructs carrying the SIVmac239 vif gene, VSV-G-pseudotyped viruses were generated in 293T cells in the presence of different APOBEC3Gs. For this experiment, species-specific APOBEC3G cDNAs were prepared from H9 (human), HSC-F (CyM) (27), and Vero (AGM) cells by RT-PCR and inserted into pcDNA3.1-FLAG, an expression vector containing an epitope tag, as described in Materials and Methods. Comparable levels of human, CyM, and AGM APOBEC3Gs were produced in transfected 293T cells, as monitored by immunoblotting by using anti-FLAG antibodies (Fig. 2A). The progeny virions generated in cells expressing human or the two monkey APOBEC3Gs were collected from culture supernatants at 48 h, and their infectivities were assayed in MAGI cells (Fig. 2B). Not unexpectedly, the replication of viruses (NL4-3 and NL-ScaR) bearing the HIV-1 vif gene and produced in cells expressing CyM and AGM APOBEC3Gs was potently suppressed. In contrast, the constructs (NL-SVR and NL-ScaVR) carrying the SIVmac239 vif gene were refractory to the effects of both CyM and AGM APOBEC3Gs. When the expression of the SIVmac vif gene in NL-ScaVR was abrogated by a frameshift mutation, the resulting virus (NL-ScaVR-dBgl) became sensitive to all three APOBEC3Gs, and its infectivity was markedly reduced. These results indicate that under the same experimental conditions in which simian APOBEC3Gs restrict wild-type HIV-1, the derivative clones expressing SIV Vif direct the production of virions able to replicate in MAGI cells.

Fig. 2.

Single-cycle replication properties of HIV-1 chimeric clones. (A) The expression of human, CyM, and AGM APOBEC3Gs containing the FLAG epitope was monitored in 293T cells by immunoblotting 48 h posttransfection. (B) VSV-G-pseudotyped viruses were prepared in 293T cells cotransfected with one of the three expression vectors for APOBEC3G (human, CyM, and AGM), and their infectivity was examined in MAGI cells. Infectivity relative to that of virus produced in the absence of APOBEC3G (null) is indicated. (C) Production of p24 Gag in human (293T), RhM (LLC-MKII), and owl monkey (OMK637) cells after infection with increasing amounts of the indicated VSV-G-pseudotyped viruses was measured on day 3 p.i. Expression levels of p24, relative to that generated by the virus sample with the highest RT activity in 293T cells, are indicated.

HIV-1 Constructs Carrying a 7-Amino Acid SIV CA Element Exhibit Increased Replication in Simian Cells.

The capacity of HIV proviruses bearing the SIV Gag analog of the HIV-1 CypA-binding loop to escape restriction in simian cells was assessed by using VSV-G-pseudotyped viruses (NL4-3, NL-ScaR, NL-SVR, and NL-ScaVR) in single-cycle replication assays by using human (293T), RhM (LLC-MKII), and owl monkey (OMK637) cell lines (Fig. 2C). Three-fold serial dilutions of each virus stock were added to the cultures, and the amounts of p24 Gag protein present in cell lysates 72 h postinfection (p.i.) was determined by ELISA. All four viruses expressed p24 with similar efficiencies in human cells. In contrast, the NL-ScaR and NL-ScaVR derivatives, both of which carry the SIV CA element, produced substantially more p24 Gag in monkey cells than the constructs (NL4-3 and NL-SVR) bearing the HIV-1 CypA-binding loop. This effect was particularly striking in owl monkey cells in which constructs carrying the SIV CA expressed 50-fold more viral protein. Not unexpectedly, the presence of SIV vif alone in NL-SVR did not result in increased p24 Gag production compared with wild-type HIV-1 in this single-cycle replication assay. The results shown in Fig. 2C indicate that the incorporation of a 21-nucleotide SIV gag gene element into HIV-1 proviral DNA is sufficient to suppress the endogenous restriction factors resident in RhM and owl monkey cells.

An HIV-1 Derivative Containing both SIV vif and the SIV CA Element Is Able to Establish Spreading Infections in a Monkey Lymphocyte Cell Line.

Although single-cycle replication experiments using pseudotyped retroviral particles like those shown in Fig. 2 B and C can furnish valuable information about virus entry, uncoating, reverse transcription, integration, and the production of viral proteins, they provide no data about the functional properties of the progeny virions that are generated. The latter information accrues from spreading multicycle infections. Toward this end, NL-ScaR, NL-SVR, and NL-ScaVR virus stocks, prepared from transfected 293T cells and normalized for equivalent amounts of reverse transcriptase (RT) activity, were used to infect human (M8166) and cynomolgus (HSC-F) cells. HSC-F is a CD4⁺CXCR4⁺CCR5⁻ CyM T cell line originally immortalized by Herpesvirus saimiri (27). Both cell types were also infected with similar amounts of the parental NL4-3 and SIVmac239, which served as controls. As shown in Fig. 3A, all of the viruses did, in fact, establish spreading infections in M8166 cells, although the three bearing the SIV vif gene (SIVmac239, NL-SVR, and NL-ScaVR) reached lower levels of peak RT activity compared with NL4-3 and NL-ScaR.

Fig. 3.

Multicycle growth potential of various chimeric viruses in human and monkey lymphocyte cell lines. Virus samples were prepared from 293T cells transfected with the indicated proviral clones, and they were inoculated into human M8166 (A) or CyM HSC-F (B) cells. HIV-1 NL4-3 and SIVmac239 served as controls. (C and D) Growth properties of viruses generated in infected HSC-F cells. Culture supernatants from NL-ScaV-infected HSC-F cells collected on days 39 and 51 p.i. (sup39 and sup51 in C) and from 293T cells transfected with a molecular clone derived from sup51 (NL-DT5 in D) were inoculated into HSC-F cells. NL4-3, SIVmac239, and NL-ScaV from transfected 293T cells served as controls. Virus replication was monitored by RT activity released into the culture supernatants.

A completely different result was obtained during infections of the CyM cell line. As expected, SIVmac239 readily established a spreading infection, which peaked on day 6 p.i.; wild-type NL4-3 produced no measurable progeny virions (Fig. 3B). Of the three NL4-3 derivatives carrying SIV sequences, only NL-ScaVR exhibited any infectivity, which became detectable on day 15 p.i. The delayed appearance of NL-ScaVR progeny is reminiscent of previously described second-site revertants of HIV-1 mutants, which acquire changes during extended tissue culture passage that confer augmented replicative properties (28, 29). Therefore, to characterize more fully the late emerging virus, new and independent infections of HSC-F cells were initiated by using both NL-ScaV- and NL-ScaVR-derived virus preparations as inocula; in both cases, the production of progeny virions was again markedly delayed (data not shown). Culture supernatants from the NL-ScaV infection were collected on days 39 and 51 p.i., normalized for RT activity, and used as inocula for infections of fresh HSC-F cells. As shown in Fig. 3C, the viruses harvested on days 39 and 51 both exhibited accelerated replication kinetics compared with the original NL-ScaV virus, suggesting that long-term passage in HSC-F had resulted in the acquisition of genetic alterations.

Molecular Cloning and Characterization of an HIV-1 Derivative Able to Cause Spreading Infections in Macaque Primary Cells.

The emergence of virus exhibiting an augmented replication phenotype prompted us to initiate the molecular cloning of cell-associated viral DNA collected from HSC-F cultures infected with the “sup 51” inoculum on day 18 p.i. Integrated proviruses were amplified from genomic DNA as two overlapping fragments by PCR, and virus stocks were prepared from 293T cells after transfection with reconstructed full-length clones. The replication properties of one of the infectious clones obtained (NL-DT5) is shown in Fig. 3D. Although production of progeny virus was delayed compared with that directed by SIVmac239, NL-DT5 still exhibited robust infection kinetics and released more particle-associated RT activity than the SIV control.

Sequencing of NL-DT5 DNA revealed that it had acquired four nucleotide changes, compared with NL-ScaV, during the 51-day passage in HSC-F cells. Two were nonsynonymous changes in env (nts 6633 and 7043), resulting in T110I (V1) and F247L (C2) substitutions in gp120. One of the other two was a synonymous change in the Pro coding sequence (nt 2300) and the other was a G to A substitution in the U3 region of the 3′ LTR. The functional significance of these changes is not presently known. Because NL-DT5 was derived from cells originally infected with NL-ScaV, the XbaI cloning site present upstream from vpr was repaired by deleting the TCT trinucleotide, as described earlier, and the resulting molecular clone was designated NL-DT5R.

Because the ultimate use of NL-DT5R would be as a virus inoculum in nonhuman primate studies, a more rigorous test of its infectivity would be replication in macaque PBMCs. In an initial experiment, unfractionated and ConA-activated PBMC from five PtM and three Indian origin RhM were infected with NL-DT5R, NL4-3, and SIVmac239 by spinoculation (30). Production of SIVmac239 progeny virions was initially detected on day 3 and peaked on day 6 p.i.; no replication of NL4-3 or NL-DT5R was observed during the 12-day course of this infection (data not shown). In contrast to these results, NL-DT5R was able to establish spreading infections in five of five PtM and one of three RhM PBMC preparations when CD8⁺ T lymphocytes were removed with magnetic beads (Fig. 4). In cells from two of the PtMs (Pt99P022 and Pt98P021), the kinetics and amounts of virus produced were similar to those seen for SIVmac239. It should be noted that NL-DT5R exhibited augmented replication in primary macaque cells compared with NL-ScaVR, the original nontissue culture-passaged construct. As expected, no replication of NL4-3 was detected in the CD8⁺ T cell-depleted primary monkey cells.

Fig. 4.

Growth potential of the chimeric viruses in CD8-depleted PBMCs from PtM and RhM. Virus samples were prepared from 293T cells transfected with the proviral clones indicated at the bottom, and they were infected into CD8-depleted PBMCs by spinoculation (30). Virus replication was monitored daily by RT production in the culture supernatants. HIV-1 NL4-3 and SIVmac239 served as controls. Animal identifications are indicated at the top of each panel.

Discussion

Our results are consistent with and extend numerous previously published single-cycle virus replication experiments that have reported species-specific APOBEC3G and TRIM5α restriction of HIV-1 in monkey cells. In our work, the establishment of spreading HIV-1 infections in simian cells represents an important step in significantly increasing the host range of HIV-1. It was conferred by inserting a 21-nucleotide SIV Gag CA element and the entire SIV vif gene into the genetic backbone of the pNL4-3 HIV-1 molecular clone, plus four additional nucleotide changes acquired during long-term passage in a CyM lymphoid cell line. The proportion of HIV-1 sequences in the molecularly cloned NL-DT5R derivative obtained (93%) is substantially greater than that present in currently available CXCR4 (X4) using SHIVs (28–30%). It may be possible to increase the HIV-1 content of NL-DT5R further by mutating the DRMR amino acid residues at positions 14–17 of HIV-1 Vif to SEMQ, which is similar to the analogous region of SIV Vif. Such a change in the vif gene has recently been reported to confer replication competence to HIV-1 constructs in the presence of RhM APOBEC3G (31). Construction of other NL-DT5R variants bearing CCR5 using env genes from a variety of HIV-1 clades is also a future goal of these studies.

In contrast to commonly used X4 SHIVs, which carry SIV gag and pol genes, the NL-DT5R variant provides the opportunity to assess nonnucleoside RT inhibitors and a full spectrum of protease inhibitors, which specifically target HIV-1-, not SIV-, encoded enzymes. HIV-1 variants like NL-DT5R may also permit analyses of the cellular responses directed against HIV-1 Gag proteins that are associated with immunologic control and escape, not possible with currently available X4-tropic SHIVs.

Although the host range of the HIV-1 NL-DT5R derivative has expanded to include a monkey lymphoid cell line and CD8-depleted RhM and PtM PBMC, it still replicates less efficiently than SIV in simian cells. This observation undoubtedly reflects the high proportion (93%) of HIV-1 sequences present in the final construct, which have evolved for optimal replicative potential in human, not monkey cells. Additional changes will be required to achieve more robust infectivity for simian cells, which has already occurred to a limited extent with the acquisition of nucleotide substitutions after in vitro passaging. In addition to expected alterations in viral structural proteins, long-term passaging of NL-DT5R in monkeys may introduce changes in other HIV-1 sequences affecting analogous but subtly different SIV nonstructural proteins and cis-acting elements involved in processes such as transcriptional regulation and T cell activation pathways in monkey cells. Such alterations are likely to occur in the HIV-1 Nef protein, which is significantly smaller (205–210 aa long) than SIV Nef (260–265 aa long), and the HIV-1 LTR, which can be distinguished from its SIV analog by encoding: (i) different numbers/types of binding sites for transcriptional regulatory proteins and (ii) a single, not a double, stem-loop-bulge TAR element present at the 5′ termini of all viral transcripts (32). Although the replication properties of HIV-1 NL-DT5R in inoculated monkeys are presently unknown, they are very likely to be less robust than existing X4 SHIVs for the reasons noted above. Nonetheless, it is anticipated that extensive in vivo passaging of NL-DT5R will greatly augment its infectivity in macaque cells. In this regard, improved replicative and disease-inducing properties attended serial animal-to-animal transfers of first-generation nonpathogenic SHIVs, and they were associated with extensive sequence changes affecting multiple viral genes (33).

Materials and Methods

Construction of HIV-1 Proviral Clones.

A pNL4-3-derived (34) vector, previously used for functional analyses of HIV-1 vif genes and designated pNL-SX (25), was the genetic backbone for the constructs shown in Fig. 1. In pNL-Sca, the 9-aa CypA-binding region of pNL-SX/NLVif (25) was replaced with the corresponding 7-residue segment from SIVmac239 CA by using the QuikChange site-directed mutagenesis kit (Stratagene, La Jolla, CA). To construct pNL-SV, the entire SIVmac vif sequence was amplified by PCR by using pMA239 (1) as template with forward TCCCCCGGGATGGAGGAGGAAAAGAGGTGG and reverse GCTCTAGATCATGCCAGTATTCCCAAGACC primers containing embedded SmaI and XbaI sites, respectively. The reactions were heated at 95°C for 5 min for 1 cycle; 95°C for 1 min, 51°C for 1 min, and 72°C for 1.5 min for 10 cycles; 95°C for 1 min, 60°C for 1 min, and 72°C for 1.5 min for 25 cycles; and 72°C for 10 min for 1 cycle. The amplified product was subcloned into the SmaI–XbaI site of pNL-SX. To construct pNL-ScaV, the SmaI–XbaI fragment from pNL-SV was cloned into the equivalent sites of pNL-Sca. A negative-control clone, designated pNL-ScaV dBgl, contained a frameshift mutation at the BglII site of SIVmac239 vif in the pNL-ScaV.

Cell Culture.

The 293T (human embryonic kidney), LLC-MKII (RhM kidney), Vero (AGM kidney), and OMK637 (owl monkey kidney) adherent cell lines were cultured in Eagle's MEM supplemented with 10% heat-inactivated FBS. A CD4⁺CXCR4⁺CCR5⁻ CyM T cell line, HSC-F (27), was maintained in RPMI medium 1640 containing 10% FBS. RhM PBMCs were prepared and cultured as described previously (35). For PtM PBMC, a mixture of 95% Ficoll–Paque Plus (GE Healthcare, Piscataway, NJ) and 5% Dulbecco's modified PBS was used as a separation medium. To remove CD8⁺ T cells from PBMCs, cells were stained with phycoerythrin (PE)-conjugated antihuman CD8 antibody (clone SK1; BD Bioscience, San Jose, CA), and then they were incubated with magnetic beads conjugated with anti-PE antibody (anti-PE MicroBeads; Miltenyi Biotec, Auburn, CA). Unstained cells were collected as the pass-through of a depletion column (LD Column, Miltenyi Biotec) according to the manufacturer's instructions.

Transfection, Infection, and RT Assays.

Virus stocks were prepared by transfecting 293T cells with cloned HIV-1 NL4-3 derivatives by using either calcium phosphate coprecipitation (36) or Lipofectamine Plus (Invitrogen, Carlsbad, CA); 48 h later, culture supernatants were collected and stored at −80°C until use. Virion-associated RT activity was measured as described previously (28). HSC-F cells (1 × 10⁷) were infected with equivalent amounts (1 × 10⁷ RT units) of different virus preparations, and then they were monitored for RT activity in the culture supernatants. Macaque PBMCs (5 × 10⁶) were infected with similar amounts (1 × 10⁷ RT units) of the indicated viruses by spinoculation (30) for 1 h, and they were maintained for 12 days. The tissue culture medium was replaced daily.

Cloning of APOBEC3G Genes.

Species-specific APOBEC3G cDNA was amplified from H9 (human), HSC-F (CyM), and Vero (AGM) cells by RT-PCR (described in Supporting Methods, which is published as supporting information on the PNAS web site) and cloned into pcDNA3.1-FLAG, an expression vector containing the FLAG tag sequence in pcDNA3.1 (Invitrogen). The expression levels of the three APOBEC3Gs in transfected 293T cells were monitored by immunoblotting using the anti-FLAG antibody.

Single-Cycle Replication Assays.

The effects of the species-specific APOBEC3Gs on virus replication were evaluated by using VSV-G-pseudotyped HIV-1 stocks, prepared from 293T cells cotransfected with (i) individual env-deficient NL4-3 clones (NL4-3, NL-SVR, NL-ScaR, NL-ScaVR, and NL-ScaVR dBgl); (ii) pCMV-G (37), a VSV-G protein expression vector; and (iii) an individual species-specific APOBEC3G expression vector at a ratio of 8:1:1. The infectivity of the resultant viruses was determined by MAGI assay as described previously (38). To assess the effect of gag gene substitutions during single-cycle replication in cells from different primate species, VSV-G-pseudotyped viruses were prepared from 293T cells cotransfected with (i) individual env-deficient NL4-3 clones [NL4-3, NL-SVR, NL-ScaR, and NL-ScaVR] and (ii) pCMV-G, at a ratio of 9:1. Virus released into the medium and normalized for RT activity was added directly or as 3-fold serial dilutions to 293T, LLC-MKII, and OMK637 cells, plated at a density of 5 × 10⁴ cells per well in 24-well plates on the day before infection. On day 3 p.i., cells were lysed with CHAPS-based lysis buffer (28), and the amounts of intracellular p24 were determined by using the RETROtek p24 ELISA kit (ZeptoMetrix, Buffalo, NY). The total amount of protein in each sample was determined in parallel with the DC protein assay kit (Bio-Rad, Hercules, CA) to normalize for different cell-harvesting efficiencies.

Generation of pNL-DT5.

HSC-F cells were infected NL-ScaV virus prepared from transfected 293T cells as described above. Half of the culture medium (5 ml) was replaced every 3 days, and the harvested supernatants were stored at −80°C. Fresh HSC-F cells (1 × 10⁷) were added on days 27, 36, and 45 p.i., and the culture was maintained until 51 days p.i. The supernatants collected on days of 39 and 51 p.i. were filtered through a 0.45-μm filter and used to initiate a second round of infection (5 × 10⁶ RT units of viruses added to 1 × 10⁷ HSC-F cells). On day 18 p.i., cells infected with the day 51 supernatant (sup 51) were collected (Fig. 3C), and the integrated provirus was amplified from genomic DNA as two overlapping fragments by DNA PCR. The 5′ fragment extended from 5′ LTR to the Vpr-coding region, whereas the 3′ fragment spanned the Vif-coding region to the 3′ LTR. The 5′ fragment was amplified with the NL1–24Aat-5′ (AGTCAGACGTCTGGAAGGGCTAATTTGGTCCCAAA at nucleotide positions 1–24 in NL4-3) and NL5832–5855Bam-3′ (ATCGCGGATCCTCTAGTCTAGGATCTACTGGCTCC at 5832–5855) primer pairs, whereas the 3′ fragment was amplified with the NL5596–5619Xba-5′ (GCTAGTCTAGAAGCCATACAATGAATGGACACTAG at 5596–5619) and NL9686–9709Sph-3′ (ACATGGCATGCTGCTAGAGATTTTCCACACTGACT at positions 9686–9709) primer pairs. The reactions were heated at 95°C for 5 min for 1 cycle; 95°C for 0.5 min, 51°C for 0.5 min, and 72°C for 6 min for 10 cycles; 95°C for 0.5 min, 60°C for 0.5 min, and 72°C for 6 min for 25 cycles; and 72°C for 10 min for 1 cycle. The amplified 5′ and 3′ viral DNA segments were digested with AatII–EcoRI and EcoRI–SphI, respectively, and they were then cloned together into pUC19 digested with AatII–SphI. The resultant proviral clone was designated pNL-DT5.

Abbreviations

AGM

African green monkey

APOBEC3G

apolipoprotein B mRNA-editing enzyme-catalytic polypeptide-like 3G

CA

capsid

CyM

cynomolgus monkey

CypA

cyclophilin A

PBMC

peripheral blood mononuclear cell

p.i.

postinfection

PtM

pig-tailed monkey

RhM

rhesus monkey

RT

reverse transcriptase

SHIV

SIV/HIV chimeric virus

SIV

simian immunodeficiency virus

SIVmac

simian immunodeficiency virus isolated from rhesus macaques

VSV-G

vesicular stomatitis virus type G.