We began this study by looking at the pattern of restriction effect of Trim5α isolated from diverse primate species with a panel of different retroviruses. Remarkably, we found that HIV-1 and FIV had an identical pattern of restriction in that both were strongly restricted by rhesus Trim5α, had slight sensitivity to human and Tamarin Trim5α, and were resistant to Titi Trim5α (data not shown). Similar data was also reported by another group 23. Because Trim5α restriction, like the CypA interaction, is dictated by CA, this led us to examine if FIV is also similar to HIV-1 in its interaction with CypA.

The TrimCyp protein from owl monkeys strongly restricts HIV-1 because the C-terminal CypA portion of the protein binds to HIV-1 CA, while the N-terminal portion of the protein leads to premature uncoating or degradation of incoming virions 13141718. We therefore asked whether FIV is also sensitive to TrimCyp restriction as an indirect measure of CA-CypA recognition. Cells expressing TrimCyp were generated and infected with VSV-G pseudotyped wild-type HIV-1, a HIV-1 G89V mutant (CA with a G89V mutation does not bind CypA 2), and FIV (Fig. 1). Infections were done in the presence or absence of CsA to verify the dependence of the restriction on CypA function. Because infections were done with viruses that encode GFP (see Methods), the number of infected cells could be directly analyzed by flow cytometry (Fig. 1). Consistent with previous reports 1718, HIV-1 is sensitive to TrimCyp restriction and the restriction can be neutralized by treatment of cells with CsA (Fig. 1A), while the negative control, the HIV-1 G89V mutant, is not restricted by TrimCyp (Fig. 1B). We found that FIV behaves similar to HIV-1 in that it was restricted by TrimCyp, and this restriction is reversed by CsA (Fig. 1C).

These results with TrimCyp suggested that FIV does indeed interact with CypA. In order to test this directly, we used an assay to biochemically detect binding of CypA to FIV CA (Fig. 2). Thus, 293T cells were transiently transfected with a plasmid that expressed a fusion protein between GST and human CypA (GST-CypA) or with GST alone. Cell lysates were then incubated with glutathione beads to partially purify the GST proteins from the cell extracts. After washing to remove unbound proteins, the beads were incubated with FIV virions, and the GST-bound proteins were run on SDS-PAGE gels and blotted with anti-CA antibodies. HIV-1 WT and G89V were used as the positive and negative controls for this assay, respectively. HIV-1 wild type associates with CypA because it is pulled down by the GST-CypA fusion protein, whereas the HIV-1 G89V mutant does not bind to GST-CypA (data not shown). On the other hand, FIV behaves similar to HIV in that partially purified GST-CypA, but not GST alone, is able to bind FIV in vitro (Fig. 2A). This binding is dose-dependent (adding more FIV virions increases the amount of CA that is pulled-down, Fig. 2B, top), and is sensitive to CsA (Fig. 2B, bottom). Thus, by both genetic (TrimCyp restriction) and biochemical (GST-CypA pull-downs) assays, we demonstrate that FIV, like HIV-1, interacts with CypA.

The glycine at position 89 and the proline at position 90 on HIV-1 capsid are the most critical target amino acids of CypA binding 2. Since we observed that FIV also interacts with CypA, we next tried to identify if amino acids in a similar region of CA of FIV are also critical for the CypA recognition. By aligning the capsid amino acid sequence of FIV with that of HIV-1, we noticed that FIV capsid contains five prolines on the loop between the predicted alpha-helices 4 and 5 (Fig. 3A). Each of these five proline residues was individually mutated to alanine. In addition, arginine at position 89 was also mutated to alanine because this position corresponds to the critical glycine at position 89 in the HIV CA (Fig. 3A).

These constructs were then co-transfected to 293T cells with VSV-G and GFP reporter plasmids to generate VSV-G pseudotyped FIVs that were used to infect CRFK or CRFK-TrimCyp cells. The infectivity of each virus was normalized to give an amount of virus that resulted in about the same number of infected CRFK cells. Similar to wild-type FIV, the mutations P84A, P85A, P88A, P92A, and R89A are also restricted by TrimCyp (Fig. 3B). Moreover, in each case, the restriction by TrimCyp is abrogated by the addition of CsA. These data indicate that mutating each of these residues did not destroy a functional interaction between CypA and the viral CA. However, the P90A mutant is not sensitive to the TrimCyp restriction and the infectivity is not affected by the addition of CsA. This indicates that amino acid P90 is critical for a functional interaction between FIV CA and CypA (Fig. 3B).

In addition to the TrimCyp assay, we also applied the GST-CypA pull down assay to test the interaction of CypA and FIV mutants. The FIV wild type and mutant virions were incubated with the GST-CypA and the interactions were analyzed by Western Blotting (Fig. 3C). Wild-type FIV and other mutants are sensitive to the TrimCyp restriction and are pulled down by the GST-CypA. On the other hand, the P90A mutant, which is not sensitive to the TrimCyp restriction, is also not pulled down by GST-CypA (Fig. 3C). These data demonstrate that amino acid P90 is a critical target for CypA binding to FIV CA.

Previous reports showed that CsA decreases spreading infections of FIV in cells and in animals 2425. However, comparisons with HIV-1 were not done. It has been shown that the infectivity of HIV-1 drops 2–5 fold when the CypA/CA interaction is blocked by CsA in Jurkat T cells in single-cycle infection experiments 789. To examine the role of CypA in FIV infection in the same cell type, Jurkat T cells were infected with VSV-G-pseudotyped FIV in the presence or absence of CsA (Fig. 4). As expected, the infectivity of wild type HIV-1 was reduced by the addition of CsA (Fig. 4, top). We obtained reduced infections of FIV in Jurkat cells relative to CRFK cells presumably due to human Trim5α restriction of FIV 23. However, similar to HIV-1, the infectivity of FIV decreased further by 2–5 fold when CsA was present during infection (Fig. 4, middle). The P90A mutant of FIV which fails to bind cyclophilin A (Fig. 3), in contrast, was not sensitive to CsA treatment (Fig. 4, bottom). These results indicate that FIV, like HIV-1, requires the endogenous cyclophilin A in target cells for optimal infection.

We aligned the amino acids in the region between the prediced alpha-helices 4 and 5 of CA from a number of different lentiviruses, and noticed that SIVagmTAN, similar to FIV, has 5 prolines, and that the length of the SIVagmTAN loop (9 amino acids) is the same as HIV-1 (Fig. 5A). We therefore tested whether SIVagmTAN also interacts with CypA using the TrimCyp assay (Fig. 5B). Consistent with other reports 56, SIVmac, as a negative control, is not sensitive to the TrimCyp. On the other hand, SIVagmTAN behaves like HIV-1 and FIV. The infectivity of SIVagmTAN is strongly restricted by TrimCyp protein, and the restriction can be counteracted by the treatment of CsA (Fig. 5B). This result suggests that the CA of SIVagmTAN is also recognized by the CypA, and that the ability of CypA to recognize lentiviral capsids is widespread, although not universal, among lentiviruses.

293T and CRFK (Crandall Feline Kidney) cells were cultured in Dulbecco's modified Eagle medium with 10% fetal bovine serum (FBS). Jurkat T cells were grown in RPMI with 10% FBS. 293T cells were used for generation of vesicular stomatitis virus G protein (VSV-G) pseudotyped lentiviruses and for production of GST-CypA fusion proteins. CRFK cells expressing TrimCyp protein from different species were generated as described previously 30.

2.5 × 10⁵ cells/ml of 293T cells were plated in 2 ml/well in a 6-well plate 16 hours prior to transfection. The VSV-G pseudotyped HIV-1 WT-GFP, HIV-1 G89V-GFP mutant, HIV-1 WT-Luc, SIVmac-Luc, and SIVagmTAN-Luc were generated as described previously 3031. Plasimds pFGinSin and pFP93 32 used to generate FIV vectors and virions were gifts of Eric Poeschla (The Mayo Clinic College of Medicine, Rochester). For generating VSV-G pseudotyped FIV WT-GFP, 0.75 μg of enhanced green fluorescent protein transfer vector pFGinSin was co-transfected with 0.4 μg of pL-VSV-G, 0.1 μg of pCMV-tat, and 0.75 μg of pFP93 to 293T cells by FuGene 6 transient transfection. For generating the VSV-G pseudotyped mutant FIVs, the same method was applied but replacing pFP93 with pFP93 proviral DNA with mutations of P84A, P85A, P88A, P90A, P92A, and R89A. The mutations were generated with the QuickChange Site-directed mutagenesis kit by following the instruction of manufacturer's protocol (Stratagene). Culture media were collected on 48 hours, 72 hours, and 96 hours after transfection, and were passed through a 0.2 μm filter (Nalgene) for harvesting the viruses. FIV P84A-GFP and FIV P90A-GFP were ultra-centrifuged in an SW28 rotor at 23000 rpm for 1.5 hours for concentrating viruses 10-fold and 100-fold, respectively. All viruses were aliquoted in 1.5 ml micro-tubes and frozen at -80°C until use. The FIV viral titer was tested by infecting CRFK cells with different volumes of the original viral stocks; the amount of virus causing 20% GFP+ (20% infected cells) was defined as 0.2 CRFK infectious equivalents.

CRFK or CRFK cells expressing TrimCyp proteins were plated 16 hours before infection at a density of 8 × 10⁴ cells/ml on a 12-well (1 ml/well) or a 24-well (0.5 ml/well) plate. Jurkat T cells for the CypA-dependency experiment were prepared before infection at a density of 2.5 × 10⁵ cells/ml on a 12-well (1 ml/well) plate. The VSV-G pseudotyped viruses were added to target cells in the presence of 20 μg/ml DEAE/Dextran, and spinoculated at 1200 relative centrifugal force (rcf) at room temperature for 2 hours 33. The infected cells were placed at 37°C incubator for 48 hours until the analysis of the infectivity tested with flow cytometry (for viruses expressing GFP) or luminometer (for viruses expressing luciferase). The HIV-1 WT-GFP, HIV-1 G89V-GFP, and all FIVs (WT and mutants) infected cells was harvested and fixed with 1% paraformaldehyde for 1 hour at 4°C. The fixed cells were washed with PBS and re-suspended with 300 μl of PBS supplemented with 5% calf serum, and subjected to FACScan (Becton Dickinson) for the analysis of the infectivity. The HIV-1 WT-Luc, SIVmac-Luc, and SIVagmTAN-Luc infected cells were washed with PBS and lyzed with 80 μl of the cell culture lysis buffer (Promega), and 10 μl of the lysates were used for the measurement of luciferase activity (infectivity) with the Luciferase assay system (Promega) and luminometer.

The mRNA from 293T cells were extracted with RNeasy kit (Qiagen). The human CypA was cloned from the human mRNA by using the OneStep RT-PCR (Qiagen) with primer sets 5'-GGATCCACGGTTCAGGTGGTTCTGGAGGTTCAGGAGTCAACCCCACCGTGTTC-3' (forward) and 5'-CTCGAGTTATTCGAGTTGTCCACAGTCAGCAATGGT-3' (reverse). The cyclophilin cDNA was used to replace huTrim5α in the plasmid pEF/GST-huTrim5α 34 (a kind gift of Jeremy Luban, Columbia University, New York). To generate GST-CypA protein in human cells, 1 μg of pEF/GST-CypA or pEF/GST were transiently transfected to 293T cells with FuGene 6 (Roche). 72 hours after transfection, one 6-well plate of cells were harvested and lysed with 1.6 ml of GST buffer (50 mM Tris pH 8.0, 150 mM NaCl, 0.5% NP40, 0.1% SDS, and supplemented with complete protease inhibitor). The lysates were centrifuged at 13200 rpm for 1 minute, and the supernatants were harvested. To balance the glutathione-sepharose beads (BD Biosciences) for optimal binding, 200 μl of beads were centrifuged at 5000 rpm for 5 minutes and washed with 800 μl of GST buffer for 3 times. The beads were re-suspended with 200 μl of GST buffer and incubated with 800 μl of GST or GST-CypA supernatants at 4°C on a slowly rotated shaker for 1 hour. After incubation, beads were centrifuged at 5000 rpm for 5 minutes and washed three times with 800 μl of GST buffer. Viruses for the pull down assay (HIV-1 WT, FIV WT, and FIV mutants) were prepared by concentrating the same amount of virions (about 1 ml) into 200 μl of PBS. The concentrated virions then were incubated with the GST or GST-CypA bound beads at 4°C on a slowly rotated shaker for 1 hour. Beads were centrifuged and washed with 800 μl of GST buffer for 3 times. The beads then were re-suspended with 60 μl of GST buffer and frozen at -80°C until the Western blot analysis.

Lysates from the GST-pull down assay were mixed with 20 μl of sample buffer (2% SDS, 1% 2-mercaptomethanol, 1% glycerol, 65 mM Tris-hydrochloride 34). 30 μl of the mixture were loaded on SDS-10% polyacrylamide gels. After electrophoresis, the proteins were transferred to polyvinylidene difluoride membranes. The membranes were blocked for 30 minutes at room temperature with 5% fat-free milk in PBS, and then incubated with a 1:1000 dilution of a mouse anti-FIV capsid monoclonal antibody (clone 2C1, from NIH AIDS Research and Reference Reagent program) overnight at 4°C. The membranes were washed 10 minutes for 3 times in PBS containing 0.1% Tween 20 and then incubated with a 1:10000 dilution of a horseradish peroxidase-conjugated anti-mouse monoclonal antibody (Santa Cruz Biotechnology) for 30 minutes at room temperature. The membranes were washed 3 times for 30 min, and the bound antibody was detected with the ECL detection system (Amersham).

Lentiviral Gag sequences were from GenBank. The accession numbers are: HIV-1 (M19921), HIV-2 (X05291), SIVcpz (AF115393), SIVmac (AY588946), SIVsm (AF334679), SIVagmTAN (U58991), and FIV (M25381). The CA sequences from these lentiviruses were aligned by the ClustalX software 3536. The computer modeling of N-terminal CA structure from SIVmac and SIVagmTAN was performed with the SWISS-MODEL server [37, 38, 39] by using the CA crystal structure of HIV-1 which bound to human CypA (PDB:1AK4) 4. The predicted CA structures were viewed and displayed by DeepView (Swiss Pdb-Viewer).