Cell lines of human (Homo sapiens, HeLa cells), African green monkey (Cercopithecus aethiops, Vero cells), rhesus macaque (Macaca mulatta, CMMT cells), and owl monkey (Aotus trivigartus, OMK cells) origin were maintained in Dulbecco's modified Eagle's medium containing 10% fetal bovine serum and antibiotics.

Vesicular stomatitis virus glycoprotein (VSV-G)-pseudotyped retroviral vectors were generated by co-transfecting 10-cm plates of 293T cells with 10 μg of pVSV-G, 10 μg of Gag-Pol expression plasmid and 10 μg of green fluorescent protein (GFP)-expressing retroviral vector, using the ProFection calcium phosphate kit (Promega). MLV vectors were made with pCFG2-eYFP and pHIT60 (Mo-MLV, a NB-tropic strain), pCIG3N (N-tropic MLV), or pCIG3B (B-tropic MLV). HIV-1 vectors were prepared with pCSGW (GFP vector) and p8.91 (HIV-1 Gag-Pol). All MLV- and HIV-derived plasmids were kindly provided by J.P. Stoye (National Institute for Medical Research, London, UK). SIVmac vectors were generated with pAd-SIV3+ and GAE-CMV-GFP, which were kindly provided by F.L. Cosset (ENS Lyon, France). The HIV SCA Gag-Pol plasmid was a generous gift from J. Sodroski (Dana Farber Cancer Institute, Boston, MA, USA) and was co-transfected with pCSGW, pVSV-G and pRev (10:10:8:2 ratio) to produce HIV-1 SCA viral particles. All viral stocks were titrated on Mus dunni tail fibroblast (MDTF) cells and analyzed by FACS. The multiplicity of infection (MOI) is defined as MOI = -2 ln(1-fp), where fp is the fraction of GFP-positive MDTF cells. HeLa, Vero, CMMT or OMK cells were then transduced with increasing doses of GFP-encoding retroviral vectors and the percentage of transduced GFP-positive cells was determined by FACS 48 h post-transduction 78.

Cyclosporin A (CsA, Sigma) was prepared at 1 mM in ethanol and diluted further in culture medium before use. For all experiments, IFN-α, IFN-β or IFN-γ were used at 1000 units/ml (U/ml). Human recombinants IFN-α2 is from Schering-Plough, IFN-β is from Cytotech SA and IFN-γ is from Roussel Uclaf. For treating MDTF cells, we used universal type I IFN (PBL Biomedical Laboratories) at 200 U/ml, which was found to be active on most mammalian cells.

Total RNA was extracted using RNeasy Mini Kit (Qiagen) and cDNA were prepared using a Oligo(dT) primer and SuperScript III Reverse Transcriptase (Invitrogen). Quantitative PCR were performed in duplicates using 1 μl of cDNA on a Roche LightCycler, using the LightCycler Fast Start DNA Master SYBR Green 1 kit (Roche). Primers were synthesized by Eurogentec. TRIM5α: TTGGATCCTGGGGGTATGTGCTGG (forward) and TGATATTGAAGAATGAGACAGTGCAAG (reverse). GAPDH: GGGAAACTGTGGCGTGAT (forward) and GGAGGAGTGGGTGTCGCTGTT (reverse). CypA: AGTGGTTGGATGGCAAGC (forward) and GATTCTAGGATACTGCGAGCAAA (reverse). TRIMCyp: CAGAAGTCCAACGCTACTGGG (forward) and CTTGCCACCAGTGCCATTATGG (reverse). PCR reactions were carried out with a denaturation step of 10 min at 95°C followed by forty-five cycles of 10 s at 95°C, 5 s at annealing temperature (55°C for cyclophilin A, 60°C for TRIM5α, TRIMCyp and GAPDH) and 20 s amplification at 72°C. Quantifications of cDNAs were determined in reference to a standard curve prepared by amplification of serial dilutions of PCR product or plasmids containing matching sequences. Analyses were performed using the second-derivative-maximum method provided by the LightCycler quantification software, version 3.5 (Roche Diagnostics).

Cells untreated or treated with IFN-α, β or γ at 1000 U/ml (or with 200 U/ml of universal type I IFN in the case of MDTF cells) for 20 min were lysed with lysis buffer (20 mM Tris HCl pH7.5, 400 mM NaCl, 1% Triton, 1 mM EDTA, 50 mM KCl and 5 mM β-Mercaptoethanol). Cells extracts (100 μg) were resolved by sodium dodecyl sulphate polyacrylamide gel electrophoresis (SDS-PAGE) and western blotted using anti-phosphorylated Stat1 (Tyr 701) rabbit antibodies (Santa Cruz Biotechnology, Inc.) or an anti-actin mouse mAb (Calbiochem).

Down-regulation of TRIM5α or TRIMCyp expression by siRNA was performed by transfecting cells with siRNA oligos (Dharmacon) using HiPerFect Transfection Reagent (Qiagen) according to manufacturer's instructions. siRNAs targeting human TRIM5α were H1 (GGUCAUUUGCUGGCUUUGU) and HA2 (GCACUGUCUCAUUCUUCAA). For African green monkey (agm) TRIM5α silencing, we used HA2 and A3 (GCCUUACGAAGUCUGAAAC). In order to silence TRIMCyp expression in OMK cells, we used a siRNA targeting CypA (CypA: GGGUUCCUGCUUUCACAGA). Control siRNA transfections were performed using a luciferase control siRNA (Dharmacon).

OMK cells untreated or treated with 1000 U/ml of IFN-β were transduced 24 h later with GFP-expressing retroviral vectors pre-treated with DNase (Roche DNase I RNase-free 200 U/ml, 10 mM MgCl₂, 1 h at room temperature) and harvested 6 h later. Total DNA was extracted using DNeasy Mini Kit (Qiagen) and quantified by measuring OD at 260 nm. Reverse transcripts were detected by PCR on 100 ng of total DNA using primers to GFP: TACGGCAAGCTGACCCTGAAG (forward) and ACGAACTCCAGCAGGACCATG (reverse).

To investigate whether type I and type II IFNs can enhance TRIM5 (TRIM5α or TRIMCyp) expression in various primate cells, we decided to compare cell lines of human (HeLa), African green monkey (Vero), rhesus macaque (CMMT) and owl monkey (OMK) origin.

First, we compared the constitutive expression of TRIM5α (HeLa, Vero and CMMT cells) or TRIMCyp (OMK cells) transcripts in the different cell lines by quantitative RT-PCR (Figure 1B). We found that OMK cells contain approximately 1.5 × 10⁷ copies/μg total RNA of TRIMCyp transcripts, whereas HeLa and Vero cells contain 3 × 10⁷ transcripts encoding TRIM5α. Among the four cell lines tested, CMMT cells express the highest number of TRIM5 transcripts, with a concentration of 9 × 10⁷ copies of TRIM5α per μg of total RNA.

In order to verify if non-human primate cells are responsive to human IFN treatment, we looked at Stat1 phosphorylation, which is an early post-receptor signaling element of both type I and type II IFN pathways, and can thus serve as a positive-indicator of IFN signaling 26. For this purpose, CMMT, Vero and OMK cells were treated with 1000 U/ml of human IFN-α, β or γ for 20 min and the activation of Stat1 was estimated by western blotting using anti-phosphorylated Stat1 antibodies. As shown in Figure 1C, although the different cell lines do not equally respond to the three IFNs, both human type I (α, β) and II (γ) IFNs can induce Stat1 phosphorylation in non-human primate cells as well as in human cells (HeLa cells). As a control, the same assay was performed on MDTF cells treated with 200 U/ml of universal type I IFN (Fig 1C), which is active on most mammalian cells. MDTF cells are devoid of any post-entry retroviral restricting activity, since they are Fv1-null 2728, do not encode a TRIM5 ortholog, like all other mouse cells (reviewed in 29), and were found to express an inactive APOBEC3G protein30.

Next, we stimulated the primate cell lines with 1000 U/ml of IFN-α, β or γ for 8 h and determined TRIM5α or TRIMCyp mRNA amounts by quantitative RT-PCR. Results were normalized to GAPDH mRNA levels. Figure 1D shows the ratio of TRIM5α or TRIMCyp expression in IFN treated/untreated cell, which is referred to as "Fold increase". We found that type I IFN enhances TRIM5α mRNA expression with induction folds ranging from 3.5 (CMMT) to 7.6 (HeLa) for IFN-α, and from 2.8 (CMMT) to 8.8 (HeLa) for IFN-β. In contrast, IFN-γ treatment does not significantly affect TRIM5α expression in any cell line, with fold induction values ranging from 1.4 (Vero) to 1.9 (CMMT). These results are consistent with previous observations made in human cells 25. In OMK cells, stimulation by type I IFN leads to a 4.4 (IFN-α) to 6.6 (IFN-β) fold increase of TRIMCyp mRNA expression, whereas type II IFN does not have any significant effect (around 1.3 fold increase).

We next examined whether the increase in TRIM5α or TRIMCyp mRNA following type I IFN treatment can influence the cell permissivity to retroviral infections. For this purpose, VSV-G-pseudotyped, GFP-encoding retroviral vectors derived from HIV-1, N-MLV and NB-MLV were prepared by transfection of 293T cells (see methods) and titrated on MDTF cells. HeLa, CMMT, Vero or OMK cells were stimulated with IFN-α, β or γ and challenged with increasing doses of GFP-encoding retroviral vectors 24 h later. MDTF cells were stimulated with universal type I IFN and challenged with the same dose of virus. The percentage of transduced GFP-positive cells was determined by FACS 48 h post-transduction. Figure 2 shows the results obtained when MDTF, HeLa, CMMT and Vero cells were challenged with either a small (MOI of 0.5 on MDTF cells) or a large (MOI of 5 on MDTF cells) amount of virus.

In the absence of IFN treatment, N-MLV is restricted in HeLa and Vero cells, whereas HIV-1 is blocked in CMMT and Vero cells. As expected, NB-MLV infection is not affected in any cell line. Furthermore, IFN-treatment of MDTF cells, which are devoid of any restriction factor, does not affect their capacity to be transduced by any retroviral vector. We did not observe any dramatic effect on retroviral restriction levels in IFN-treated primate cells at this low multiplicity of infection. IFN-β was found to be the most effective in enhancing TRIM5α-mediated restriction, in accordance with the results of expression enhancement. Following IFN-β treatment, we observed a low but reproducible enhancement of N-MLV restriction in HeLa and Vero cells, as well as a higher restriction towards HIV-1 infection in CMMT cells. This moderate effect of IFN at low MOI is probably due to the fact that the constitutive amount of TRIM5α in the cells is sufficient to restrict a small amount of virus, whereas at higher MOI, it can be predicted that an increased expression of TRIM5α would prevent its saturation by incoming viruses, thus enhancing restriction. Indeed, when a higher dose of virus was used (MOI of 5 on MDTF cells), we observed a much higher restriction efficiency of N-MLV in HeLa and Vero cells treated with type I IFN (Figure 2). In contrast, no significant increase of HIV-1 restriction can be observed in Vero or CMMT cells.

Together, our observations confirm that type I IFN can modulate retroviral restriction in diverse primate cells. However, we did not observe a direct and systematic correlation between the levels of up-regulation of TRIM5α expression and the enhancement of restriction activity. Indeed, the IFN-induced reduction of infectivity of TRIM5α-sensitive viruses seems to be dependent of diverse parameters, such as the multiplicity of infection and the basal restriction activity of cells against a given virus.

In OMK cells, treatment with type I IFN was found to have a greater influence on retroviral permissivity compared to other primate cell lines (Figure 3). IFN-β treatment, in particular, induces a 3-fold decrease of HIV-1 infectivity at low multiplicity of infection (MOI of 0.5), which rises to 26-fold at high MOI (MOI of 5). More surprisingly, stimulation with type I IFN also led to a reduced susceptibility of OMK cells to NB-MLV and N-MLV infection, although TRIMCyp is known to be inefficient to restrict MLV strains, since MLV CA proteins do not interact with CypA. Indeed, we found that IFN-β induces a 3.9 (at high MOI) to 5.6 (at low MOI) fold increase of restriction towards N-MLV and a 5.5 (at high MOI) to 5.6 (at low MOI) fold increase of anti-NB-MLV restriction activity in OMK cells (Figure 3 and Table 1). In order to determine whether IFN initiates a wide and unspecific antiviral response in OMK cells, we tested two other viruses which are known to be insensitive to TRIMCyp-mediated restriction, B-MLV and SIVmac (Figure 3) 171822. Interestingly, B-MLV permissivity is also affected by IFN treatment, whereas OMK cells remain fully permissive to SIVmac. The fact that SIVmac is unaffected by this IFN-induced block proves that the loss of infectivity we observed with MLV is not the consequence of an increased cell mortality due to IFN treatment. Since IFN-β gave more contrasted results than IFN-α, we decided to pursue our study with IFN-β only. Table 1 summarizes the effect of IFN-β on retroviral restriction in all tested primate cell lines. Results are expressed as fold restriction, which represents the ratio of untreated to IFN-treated cells to be transduced by the GFP-encoding retroviral vectors. As shown, IFN-β significantly increases (fold restriction > 2) the restriction of N-MLV in HeLa and Vero cells, whereas it induces a wide-spectrum anti-retroviral activity in OMK cells affecting HIV-1, NB-MLV, N-MLV and B-MLV but not SIVmac.

Therefore, our observations suggest that OMK cells express an IFN-inducible anti-retroviral protein other than TRIMCyp, which can interfere specifically with certain retroviruses.

In order to verify whether the increased restriction activity towards N-MLV infection in HeLa and Vero cells was due to the IFN-induced TRIM5α, cells were first transfected with a siRNA targeting TRIM5α, treated with IFN-β 24 h later, and finally challenged the next day with N-MLV at a MOI of 5 (as determined on MDTF cells). Three siRNA were used, which silence human (H1 and HA2) or African green monkey (HA2 and A3) TRIM5α expression (Figure 4A). A siRNA targeting luciferase (Luc) was used as a control. As shown in Figure 4B, the IFN-induced enhancement of N-MLV restriction in HeLa and Vero cells is almost completely lost when TRIM5α expression is down-regulated by either of the two siRNA, thus suggesting that TRIM5α is the main mediator of the IFN-induced N-MLV restriction. In contrast, no effect on B- or NB-MLV infectivity can be observed following TRIM5α silencing in HeLa or Vero cells, as expected (not shown).

We used the same strategy in order to confirm the existence of an IFN-induced TRIMCyp-independent retroviral restriction activity in OMK cells. For this purpose, we used a siRNA targeting CypA as well as the CypA C-terminal portion of TRIMCyp. The anti-Luc siRNA was used as control. Figure 5A shows the quantification of TRIMCyp expression in siRNA-transfected OMK cells by quantitative RT-PCR. As expected, OMK cells transfected with the TRIMCyp siRNA lost their capacity to restrict HIV-1, as compared to cells transfected with the Luc siRNA (Figure 5B). Next, we tested the effect of TRIMCyp silencing on retroviral restriction following IFN-β stimulation. One day after siRNA transfection, OMK cells were stimulated with 1000 U/ml of IFN-β and challenged with one of the four retroviruses found to be restricted: N-MLV, B-MLV, NB-MLV or HIV-1. All retroviral vectors were used at the same titer, corresponding to the viral dose that gives a MOI of 5 on MDTF cells. As shown in Figure 5C, the IFN-β-induced restriction of N-, B- and NB-MLV is not affected by TRIMCyp silencing, thus demonstrating that another mediator is involved in this anti-retroviral activity. In the case of HIV-1, the extinction of TRIMCyp expression prevents the IFN-induced enhancement of viral restriction, confirming the fact that TRIMCyp is the main mediator of the IFN-induced anti-HIV-1 activity in OMK cells.

To further address the question of the involvement of TRIMCyp in the IFN-induced anti-retroviral activities, we examined the effect of IFN-β stimulation on retroviral transduction in the presence of CsA, which is known to relieve the TRIMCyp-mediated restriction in OMK cells 14152122. OMK cells were stimulated with 1000 U/ml of IFN-β and challenged with retroviral vectors 24 h later in the presence of CsA. We found that IFN-induced restriction of HIV-1 was partially relieved by CsA, thus confirming that TRIMCyp is the main mediator of the IFN-induced anti-HIV restriction in OMK cells (Figure 6A). In contrast, CsA treatment had no effect on the IFN-induced activity against N-, B- and NB-MLV, thus confirming that TRIMCyp is not responsible for the IFN-induced block of MLV infection.

This was further confirmed by comparing the restriction profile of HIV-1, SIVmac and HIV-1 containing the CA protein of SIVmac (HIV SCA) following IFN-β stimulation of OMK cells. As shown in Figure 6B, the HIV SCA chimera virus is as insensitive to IFN-β treatment as SIVmac, compared to HIV-1.

Since IFNs can induce several antiviral mediators acting at different stages of viral replication, we next examined which step of the viral cycle is inhibited in response to IFN treatment of OMK cells.

First, we challenged naive or IFN-β-treated OMK cells with SIVmac, NB-MLV, B-MLV, HIV-1 or N-MLV and extracted total DNA 6 h post-transduction. Reverse transcripts were detected in cell extracts by PCR using GFP primers. In accord with the mode of action of TRIMCyp, we observed an inhibition of viral DNA synthesis in the case of HIV-1 infection, which is even more pronounced when cells are pretreated with IFN-β (Figure 7). In contrast, the amount of SIVmac reverse transcripts is comparable between naive and IFN-stimulated cells, as expected. Interestingly, the amount of reverse transcripts significantly decreases in cells treated with IFN-β following infection with NB-, B- or N-tropic MLV strains. We concluded from this experiment that IFN-β treatment interferes with an early step of the MLV replication cycle, prior to or during reverse transcription.