The HIV-1 derivatives were constructed on a background of infectious molecular clone NL4-3 33. NL-ScaVR, a virus containing SIVmac239 L4/5 and the entire vif gene, was constructed according to the procedure described by Kamada et al. 21. A single amino acid His (H) at the 120th position of NL-ScaVR CA was replaced with Gln (Q) by means of site-directed mutagenesis with the PCR-mediated overlap primer extension method 34, and the resultant construct was designated NL-ScaVRA1. The L6/7 of CA (HNPPIP) of NL-SVR, NL-ScaVR, or NL-DT5R was also replaced with the corresponding segments of SIVmac239 CA (RQQNPIP) by means of site-directed mutagenesis, and the resultant constructs were designated NL-SVR6/7S, NL-ScaVR6/7S, or NL-DT5R6/7S, respectively. The BssHII-ApaI fragment of NL-ScaVR, NL-SVR6/7S, or NL-ScaVR6/7S, which corresponds to matrix (MA) and CA, was transferred to env deleted NL4-3 (NL-Nhe) to generate the env (-) version of each of the constructs.

The 293 T (human kidney), LLC-MK2 (Rh kidney), and TK-ts13 (hamster kidney) adherent cell lines were cultured in Dulbecco's modified Eagle medium supplemented with 10% heat-inactivated FBS. The CD4+ CXCR4+ CM T cell line HSC-F 35 was maintained in RPMI 1640 medium containing 10% FBS. Virus stocks were prepared by transfection of 293 T cells with HIV-1 NL4-3 derivatives using the calcium phosphate co-precipitation method. Viral titers were measured with the p24 or p27 RetroTek antigen ELISA kit (ZeptoMetrix, Buffalo, NY), and viral reverse transcriptase (RT) was quantified with the Reverse Transcriptase Assay kit (Roche Applied Science, Mannheim Germany).

The HIV-1 vector expressing GFP was prepared as described previously 3637. To construct the HIV-1-WT-GFP and HIV-1-L4/5S-GFP vector, we replaced the Eco RI-Apa I fragment corresponding to MA and CA of the pMDLg/p.RRE packaging vector with those fragments from NL4-3 and NL-ScaVR, respectively. The GFP viruses were prepared from 293 T cells in a 15-cm dish by co-transfection with a combination of 24 μg of pMDLg/p.RRE derivatives, 36 μg of CS-CDF-CG-PRE (GFP encoding viral genomic plasmid), 10 μg of pMD.G (vesicular stomatitis virus glycoprotein (VSV-G) expressing plasmid), and 10 μg of pRSV-Rev (Rev expressing plasmid). Forty-eight hours after transfection, the culture supernatants were collected and used for infection.

3 × 10⁵ MT4 or HSC-F cells were infected with 20 ng of p24 of NL4-3, NL-ScaV, NL-ScaVR, NL-ScaVR6/7S, NL-DT5R, or NL-DT5R6/7S. The culture supernatants were collected periodically, and p24 levels were measured with an ELISA kit.

The culture supernatant of 293 T cells transfected with plasmids encoding HIV-1 NL4-3 derivatives was clarified by means of low speed centrifugation. Nine ml of the resultant supernatants were layered onto a 2 ml cushion of 20% sucrose (made in PBS) and centrifuged at 35,000 rpm for 2 h in a Beckman SW41 rotor. After centrifugation, the virion pellets were resuspended in PBS, and p24 antigen concentrations were measured with ELISA. SDS-polyacrylamide gel electrophoresis was applied to 120 ng of p24 of HIV-1 derivatives, and virion-associated proteins were transferred to a PVDF membrane. CA and CypA proteins were visualized with the anti-p24 antibody (Biodesign International, Saco, ME) and the anti-CypA antibody (Affinity BioReagents, Golden, CO), respectively.

HIV-1 derivatives or SIVmac particles were prepared by transfecting each of the env-deleted HIV-1 NL4-3 derivatives or SIVmac plasmids with a plasmid encoding VSV-G into 293 T cells, and culture supernatants were collected two days after transfection. One day before infection, Rh LLC-MK2 and hamster TK-ts13 were plated at a density of 5 × 10⁴ cells per well in a 24-well plate. Prior to GFP virus infection, the cells were pretreated for 2 hours with 200 ng of p24 of each of the HIV-1 or SIVmac particles pseudotyped with VSV-G. Immediately after the pre-treatment, the cells were washed and infected with the HIV-1-WT-GFP or HIV-1-L4/5S-GFP virus. Two hours after infection, the inoculated GFP viruses were washed, and the cells were cultivated in fresh media. Two days after infection, the cells were fixed by formaldehyde, and GFP expressing cells were counted with a flowcytometer. To suppress endogenous TRIM5α activity, the cells were first infected with Sendai (SeV) expressing TRIM5 lacking the SPRY domain at a multiplicity of infection of 10 plaque forming units per cell. Sixteen hours after SeV infection, the cells were treated with 200 ng of p24 of the particles and then infected with the HIV-1-L4/5S-GFP vector as described above.

CM PBMCs were suspended in RPMI medium 1640 supplemented with 10% (vol/vol) FBS, and the CD8+ cells were removed with a magnetic bead system (Miltenyi Biotec, Auburn, CA) and stimulated for 1 day with 1 μg/ml of PHA-L (Sigma, St. Louis. MO). For prolonged stimulation, CD8-depleted CM PBMCs were first stimulated with 1 μg/ml of PHA-L for 2 days and then with human IL2 100 U/ml for 2 more days. 3 × 10⁵ cells were then inoculated with 200 ng of p24 of NL-DT5R, NL-DT5R6/7S or with 200 ng of p27 of SIVmac239 and incubated at 37°C in a medium containing 100 U/ml of human IL2. The culture supernatants were collected periodically, and the levels of p24 or p27 were measured with an antigen capture assay (Advanced BioScience Laboratories, Kensington, MD)

Several proviral DNA constructs have been generated to counteract the restriction of HIV-1 replication in CM T cell line HSC-F 38 (Fig. 1). We first generated NL-SVR and NL-ScaVR according to the procedure described by Kamada et al. 21. NL-ScaVR, a virus with SIVmac239 L4/5 CA and vif, could replicate slowly in HSC-F and replicated well in MT4 as previously reported (Fig. 2A). We recently discovered that the 120th amino acid of CA affected the sensitivity of HIV-2 to CM TRIM5α 32. We, therefore, introduced an additional amino acid substitution, His to Gln, at this position in NL-ScaVR. The resultant virus was designated NL-ScaVRA1; but this virus unexpectedly showed less efficient replication than did the parental NL-ScaVR in both MT4 and HSC-F cells (Fig. 2A), probably due to a reduced viral fitness created by this mutation. We, therefore, replaced the entire L6/7 CA of NL-ScaVR (HNPPIP) with the corresponding loop from SIVmac239 (RQQNPIP), and the resultant virus was designated NL-ScaVR6/7S. The amount of RT per 1 ng of CA of NL-ScaVR (0.083 ng) was comparable to that of NL-ScaVR6/7S (0.081 ng), indicating that the replacement of L6/7 in HIV-1 with the corresponding loop of SIVmac did not affect the reactivity of CA antigen. Although NL-ScaVR6/7S grew slightly slower in MT4 cells, it could replicate more efficiently in HSC-F cells than the parental NL-ScaVR could (Fig. 2A). Similar results were obtained when we inoculated 20 ng of RT equivalent of NL-ScaVR or NL-ScaVR6/7S into HSC-F cells and measured the periodic RT production in culture supernatants (data not shown). These findings demonstrated that L6/7 CA of SIVmac improved the replication in CM cells of an HIV-1 derivative that already contained a SIVmac L4/5 and vif. We then generated NL-SVR6/7S, in which the L4/5 sequence was from HIV-1, but the L6/7 and vif came from SIVmac. NL-SVR6/7S showed better replication than NL-ScaVR6/7S in MT4 cells, but lost its replicative capability in HSC-F cells (Fig. 2B). NL-SVR, a virus with SIVmac vif, could replicate in MT4, but failed to do so in HSC-F (Fig. 2B). These results indicated that both L4/5 and L6/7 of SIVmac are required for efficient replication in HSC-F.

We then introduced SIVmac L6/7 into NL-DT5R, a molecularly cloned virus with two nonsynonymous changes in the env gene gained during long-term passages of NL-ScaVR in HSC-F cells 21. The resultant virus was designated NL-DT5R6/7S. Although the peak titer of NL-DT5R6/7S was almost the same as that of NL-DT5R, NL-DT5R6/7S could replicate faster in HSC-F than the parental NL-DT5R (Fig. 2C). This finding confirmed that SIVmac L6/7 CA sequence improved the replication in CM cells of HIV-1 derivatives that contained SIVmac L4/5 and vif. The finding suggested that HIV-1 L6/7 and L4/5 CA sequences are important for intrinsic restriction in CM cells.

Several studies have demonstrated that CypA augments HIV-1 infection in human cells 39, but inhibits its replication in OWM cells 181920. CypA was packaged in HIV-1 but not in SIVmac virus particles. To determine whether the replacement of HIV-1 L6/7 with that of SIVmac affects CypA binding of HIV-1 CA, we performed Western blot analysis of viral particles from HIV-1 derivatives. As shown in Fig. 3 (upper panel), CypA proteins were clearly detected in the NL-SVR particles (lane 1) but not in those of NL-ScaVR (lane 3), thus confirming that the L4/5 sequence of HIV-1 but not of SIVmac is required for CypA incorporation into viral particles. CypA proteins were detected in NL-SVR6/7S (lane 2) but not in NL-ScaVR6/7S (lane 4), indicating that the additional replacement of HIV-1 L6/7 with that of SIVmac had little effect on CypA incorporation. This finding suggests that the effect of L6/7 replacement on viral growth was independent from CypA binding of HIV-1 CA. When we used anti-p24 antibody (Fig. 3, lower panel), p55 Gag precursors and p24 proteins were clearly detected. There were no differences in the amount of p24 or the ratio of p24 to p55 among the four HIV-1 derivatives, indicating that the HIV-1 Gag precursor proteins with SIVmac L4/5 and L6/7 were processed normally by the viral protease.

It is known that the intrinsic restriction factors working against HIV-1 in CM and Rh cells can be saturated by inoculation of a high dose of HIV-1 particles 19404142. To determine whether alteration in the CA of HIV-1 would affect its ability to saturate restriction factors, Rh LLC-MK2 cells were pre-treated with equal amounts of VSV-G pseudotyped HIV-1 particles that were with or without SIVmac L4/5 and/or L6/7 CA to saturate intrinsic restriction factor(s). The pre-treated cells were then infected with GFP-expressing HIV-1 carrying SIVmac L4/5 CA (HIV-1-L4/5S-GFP), since we wanted to exclude any effects of CypA on the GFP expressing virus in LLC-MK2 cells. The susceptibility of particle-treated cells to virus infection was determined by the percentage of GFP-positive cells. The cells treated with the wild type (WT) particles showed greatly enhanced susceptibility to HIV-1 infection compared with non-treated cells (Fig. 4A, left), demonstrating that the intrinsic restriction factor(s) in LLC-MK2 cells were saturated by a high dose of particles. The cells treated with the particles carrying SIVmac L4/5 and those treated with particles carrying SIVmac L6/7 also showed enhanced susceptibility to HIV-1 infection (Fig. 4A, left). The cells treated with particles carrying both SIVmac L4/5 and L6/7 showed only slight enhancement of HIV-1 susceptibility (Fig. 4A, left; p = 0.007 compared by means of paired t test using all data points with the WT particle treated cells). Similarly, the cells treated with SIVmac particles showed only minor enhancement in HIV-1 susceptibility (Fig. 4A, left). Hamster TK-ts13 cells which lack TRIM5α expression, on the other hand, showed no difference in HIV-1 susceptibility among cells treated with various HIV-1 derivatives or SIVmac particles (Fig. 4A, right). As shown in Fig. 4B, similar results were obtained when we used a GFP-expressing virus with WT HIV-1 capsid (HIV-1-WT-GFP). These results indicate that both HIV-1 L4/5 and L6/7 are important for CA binding to antiviral factor(s) in Rh cells. As described previously 20, HIV-1-WT-GFP could induce infection in only small numbers of LLC-MK2 cells. In contrast, more TK-ts13 cells were infected with HIV-1-WT-GFP than with HIV-1-L4/5-GFP. It is thus possible that CypA is a supporting factor for HIV-1 replication in hamster cells as well as in human cells.

Endogenous TRIM5α seems to be a likely candidate for the antiviral factor saturated by a high dose of HIV-1 particles (Fig. 4A and 4B). To confirm this, we assessed the ability of WT and mutant HIV-1 particles to saturate the intrinsic restriction factor in the presence or absence of functional TRIM5α. The dominant negative effect of an over-expressed TRIM5 mutant lacking SPRY domain 43 was used to suppress the function of cell endogenous TRIM5α. As shown in Fig. 4C, the infection of a recombinant SeV expressing TRIM5 without the SPRY domain caused marked enhancement of HIV-1-L4/5S-GFP virus infection without prior particle treatment (crosses vs. asterisks). This indicates that this dominant negative TRIM5 mutant successfully suppressed the restriction activity of endogenous TRIM5α. Treatment with the WT HIV-1 particles also saturated the restriction factors in the cells infected with the empty vector virus (parental Z strain of SeV), while the additional effect of the dominant negative mutant TRIM5α remained unclear (Fig. 4C left, white vs. black squares). These results suggest that the intrinsic factors saturated by the WT particles were mainly endogenous TRIM5α. In contrast to the effect of the WT particle treatment, the effect of the dominant negative TRIM5 mutant on HIV-1 infection was evident when we used particles with SIVmac L4/5 and L6/7 (Fig. 4C, right, white vs. black diamonds, p = 0.007, paired t test). These findings suggest that the diminished capability of particles with SIVmac L4/5 and L6/7 to saturate restriction factors was mainly due to their loss of interaction with TRIM5α. We, therefore, concluded that the ability of HIV-1 with SIVmac L4/5 and L6/7 to bind to TRIM5α is diminished in LLC-MK2 cells.

To verify the effect of additional replacement of HIV-1 L6/7 with that of SIVmac in primary CM cells, we prepared PBMCs from CM and removed CD8+ cells by means of magnetic beads. The cells were then stimulated for 1 day with 1 μg/ml of PHA-L. NL-DT5R6/7S showed more efficient replication than did the parental NL-DT5R in these cells and reached its peak titer 8 days after infection (Fig. 5A). For prolonged stimulation, CD8-depleted CM PBMCs were first stimulated with 1 μg/ml of PHA-L for 2 days and then with human IL2 100 U/ml for 2 more days. In these cells, NL-DT5R with HIV-1 L6/7 did not grow at all. On the other hand, NL-DT5R with SIVmac L6/7 (NL-DT5R6/7S) grew in CM primary cells in response to prolonged stimulation by PHA and IL-2 to reach titers, similar to those attained in cells with short stimulation, up to 8 days after infection (Fig. 5A and 5B). Furthermore, NL-DT5R6/7S continued to grow to much higher titers and reached its peak titer 16 days after infection; this higher peak may be due to better proliferation of these cells than those cells receiving short term stimulation (Fig. 5B). These results confirmed that the replicative capability of HIV-1 in CM cells was augmented by the additional replacement of L6/7 of CA with the corresponding sequence from SIVmac.