We reasoned that Vpx -which depletes SAMHD1- should rescue the early IFN-α block to HIV-1 transduction, if this block was the result of SAMHD1 activity. Monocytes (Figure  1A) or MDMs (Figure  1B) from two donors were treated for 24 hours with high doses of IFN-α or IFN-γ (10 000 units/ml), thereafter cells were incubated with empty or Vpx-containing viral-like particles (VLP X- or VLP X+) and infected with a GFP-encoding HIV-1 virus. The percentage of infected (GFP positive) cells was measured 3 days post-infection. HIV-1 transduction was inhibited by IFN-α in both cell types and to a lesser extent by IFN-γ though the respective effects of both IFN were difficult to assess because of the low transduction efficiencies. Vpx dramatically increased HIV-1 transduction as expected 45 and was able to slightly rescue HIV-1 transduction from the IFN-α inhibitory effects in correlation with SAMHD1 degradation (Figure  1A and B and Additional file 1: Figure S1). At lower doses of IFN-α, the rescue from the IFN-α block by Vpx was still weak (Figure  1C, 1000 U/ml; Figure  2A, 100 U/ml) despite undetectable levels of SAMHD1 (Figure  1D, 1000U/ml IFN-α or γ, one representative donor over three donors). Lower doses of IFN-γ did not significantly affect HIV-1 transduction or Vpx helper effect (Figure  1C), which led us to pursue our study only with IFN-α. Of note, APOBEC3A levels were efficiently increased by IFN-α as previously shown 50 51 , and were not decreased by Vpx (Figure  1D). No change in SAMHD1 levels was observed in a time-course experiment where macrophages were treated with IFN-α or IFN-γ for different times, in contrast to APOBEC3A levels used as a control (Figure  1E).

These preliminary results argue for the existence of IFN-α induced anti-HIV interfering factors other than SAMHD1, but did not rule out a role of SAMHD1 in the early innate response in macrophages. In case SAMHD1 contributed to the IFN-α-block, we expected the helper effect of Vpx toward HIV-1 to be magnified under IFN-α treatment, which should be measurable by an enhancement of the transduction ratio “+ Vpx” over “- Vpx” (Vpx helper effect). Transduction efficiencies using HIV-1 GFP-expressing vectors were very low, namely in the presence of IFN-α, and precluded the determination of such ratios (Figure  1). We therefore transduced cells with a VSV-G pseudotyped HIV-1 luciferase virus, which enabled accurate quantification. Transduction was gradually inhibited with increasing doses of IFN-α, as well as the helper effect of Vpx toward HIV-1 transduction (Figure  3A). In parallel, SAMHD1 degradation was obtained at every single dose of IFN-α, while levels of APOBEC3A and ISG15, a second interferon-induced protein, were increased (Figure  3B). These results suggest that SAMHD1, which is well expressed at a basal level, is unlikely an early specific IFN-α-induced factor in macrophages. We further transduced IFN-α-primed macrophages with Vpx + VLP to trigger SAMHD1 proteasomal degradation and subsequently infected the cells with a Vpx-deleted GFP-encoding SIVmac virus, which is sensitive to SAMHD1 restriction. Unlike in control cells, where Vpx rescue was efficient, Vpx was unable to enhance SIV transduction in IFN-α treated cells even at low doses of IFN-α as previously reported (Figure  3C) 19 52 53 .

Altogether, our results suggest that SAMHD1 does not specifically contribute to the IFN-α-block toward HIV-1 transduction in macrophages.

We expected Vpx to enhance the dNTP pool to the same extent with or without IFN-α since SAMHD1 was degraded in both conditions. This assumption was directly explored by quantifying dNTP levels in MDMs treated or not with IFN-α. A 100U/ml dose of IFN-α was deliberately used to severely inhibit HIV-1 transduction without overwhelming the transduction pathway (Figure  2A). IFN-α did not affect the basal pool of intracellular dATP and dTTP, representative of the four dNTP (Figure  2B, one out of three representative donor data is shown). As expected, dATP and dTTP levels were drastically enhanced in the presence of Vpx or a 2 mM deoxynucleoside (dNs) treatment (Figure  2B). When cells were primed with IFN-α, neither Vpx nor dNs efficiently restored HIV-1 transduction (Figure  2A), though dNTP increase was still efficient in both conditions (Figure  2B). Thus, on the one hand, Vpx efficiently rescued dNTP levels under IFN-α treatment in correlation with SAMHD1 degradation despite poorly rescuing HIV-1 transduction. On the other hand, bypassing SAMHD1 dNTP hydrolase activity by providing dNs did not allow rescuing HIV-1 transduction from the IFN-α block (Figure  2A and B). Altogether, these results reinforce our conclusion that SAMHD1 does not contribute to the early interferon block to HIV-1 transduction and further suggest that dNTP levels are likely not targeted by IFN-α to establish this block in macrophages.

Because activated CD4+ T lymphocytes contain a higher amount of dNTP than macrophages, we wondered whether IFN-α would affect the dNTP pool in primary T cells. HIV-1 transduction rates were lower in primary CD4+ T cells (Figure  4A) than in Jurkat or CEM T cell lines (respectively 83% and 40%, data not shown). A 1000 U/ml dose of IFN-α had a modest effect on the transduction rate, ranging from a 2 to 8-fold decrease (Figure  4A). Despite this reduction, dNTP levels were not affected (Figure  4B).

Altogether, our results suggest that nucleotide depletion does not contribute to the early IFN block toward HIV-1 transduction in macrophages or in activated CD4+ T lymphocytes.

Phorbol 12-myristate 13-acetate (PMA)-treated THP-1 cells recapitulate the MDMs phenotype in many aspects and notably model MDMs in their sensitivity to Vpx-containing VLP and Vpx-deleted SIV viruses. Surprisingly, Vpx was unable to rescue HIV-1 transduction from the IFN-α-block in these cells (Figure  5A and C, top) unlike in primary MDMs. This lack of rescue correlated with the inability of Vpx to induce the degradation of SAMHD1 in IFN-α-treated THP-1 cells (Figure  5A and C, bottom). Similar results were obtained with IFN-γ. The levels of β-catenin, a very unstable protein, remained unchanged under IFN-α or IFN-γ treatments, suggesting that the inhibition of Vpx-mediated SAMHD1 degradation was not the result of an alteration of the proteasomal degradation pathway (Figure  5A, bottom). No rescue from the block was observed in SAMHD1-depleted THP-1 cells 28 further demonstrating that SAMHD1 did not contribute to the IFN block in THP-1 cells (Figure  5B). We further wondered why SAMHD1 was resistant to Vpx-mediated degradation in THP-1 cells. We asked whether this effect was the result of an inhibition of VLP entry using a β-lactamase-Vpr (BlaM-Vpr) virion-based fusion assay. This test consists in the use of viruses containing a BlaM-Vpr protein chimera. The delivery of BlaM-Vpr in the target cell is monitored by the enzymatic cleavage of a fluorogenic substrate of β-lactamase. 30% of decrease of viral entry was observed in the presence of very high doses of IFN-α but not at lower doses that remain sufficient to inhibit Vpx-mediated SAMHD1 degradation (Figure  5D and data not shown). Thus, a defect in viral entry was not responsible for the inhibition of Vpx-mediated SAMHD1 degradation in THP-1 cells. We further wondered whether an effect of IFN-α on the SAMHD1 promoter was at stake. To address this question, we stably transfected cells with a retroviral vector encoding HA-SAMHD1 under the control of the CMV promoter. Interestingly, expression of exogenous SAMHD1 in THP-1 cells, which already express endogenous SAMHD1, enhanced the restriction against HIV-1 by a 2-fold (Figure  5C, top). HA-SAMHD1 was degraded in the presence of Vpx and this process was inhibited by IFN-α (Figure  5C, bottom). This result suggests that inhibition of Vpx-mediated SAMHD1 degradation does not depend on an effect of IFN-α on the endogenous SAMHD1 promoter. Furthermore no change in the levels of DCAF1 or DDB1, subunits of the Cul4-based ubiquitin ligase used by Vpx, was observed under IFN treatment (data not shown). Finally, we asked whether IFN-α treatment could perturb SAMHD1 localization. Both endogenous SAMHD1 and the stably expressed HA-tagged-protein were present in the nucleus in agreement with previous reports 54 55 56 (Figure  5E). Priming with IFN-α did not perturb this localization.

PBMCs from the blood of anonymous donors (obtained in accordance with the ethical guidelines of the Institut Cochin, Paris) were obtained by Ficoll density-gradient separation. Monocytes were isolated by positive selection on CD14 magnetic microbeads (Miltenyi Biotec) and CD4+ T cells were isolated by positive selection with CD4 magnetic microbeads (Milteny Biotec), followed by culture in BD Primaria flasks with R10 medium (RPMI-1640 GlutaMAX-I, 10 mM HEPES, pH 7.2, 1 mM sodium pyruvate, 1% (vol/vol) nonessential amino acids and 10% (vol/vol) heat-inactivated FCS) and antibiotics. MDMs were obtained from monocytes by culture of the cells for 7–9 d with granulocyte-macrophage colony-stimulating factor (10 ng/ml) and macrophage colony-stimulating factor (20 ng/ml). The primary CD4+ T cells were activated with phytohemaglutinin A (PHA-L; 3.0 mg/mL; Sigma-Aldrich) and interleukin-2 (IL2; 50 IU/mL; PeproTech) for 2 days prior to treatment with 1000 U/ml of IFN-α for 24 h. Human monocytic THP-1 cells and lymphoid Jurkat and CEM cells were maintained in R10 medium. THP-1 cells were differentiated with a 65 nM phorbol 12-myristate 13-acetate (PMA) treatment for 24 h. A THP-1-HA-SAMHD1 polyclonal cell line was obtained after cell transduction with viral particles expressing HA-SAMHD1 from the pLenti vector 27 and puromycin selection (2 μg/ml for 2 weeks). A monoclonal cell line was further developed by plating one cell expressing HA-SAMHD1 per well with the help of a flow cytometer. SAMHD1-depleted THP-1 cells were previously described 28 .

VLP and viruses were produced in 293 T cells co-transfected by the calcium-phosphate method according to Berger et al. 45 63 . Vpx-containing and control SIVmac VLP were produced by transfection of 293 T cells with pSIV3⁺ Δvpr (VLPs X⁺) or pSIV3⁺ Δvpx Δvpr (VLP X⁻) 52 along with a plasmid encoding for vesicular stomatitis virus glycoprotein (VSV-G) in a ratio of 10:1. For the SIVmac-GFP reporter virus, the minimal SIVmac genome pGAE 1.0, in which GFP expression is driven from the cytomegalovirus promoter, was added to the packaging and VSV-G encoding plasmids in a ratio of 5:5:1 45 63 . The HIV-1-GFP and HIV-1-Luc viruses were produced by transfection of cells with pRRLsin.eGFP and pCTS.Luc plasmids, respectively, with pCMVΔR8.91 packaging and VSV-G plasmids, in a ratio of 5:5:2 45 63 64 . β-lactamase-Vpr (BlaM-Vpr) containing particles were produced by transfection of cells with the pCMV-BlaM-Vpr plasmid along with the pCMVΔR8.91 packaging and the VSV-G encoding vectors 65 66 . The culture supernatants were collected 48 h and 72 h after transfection and passed through 0.45-μm pore filters. The viral particles were then concentrated in 10% polyethylene glycol 6000 (PEG-6000) (Sigma) containing 300 mM NaCl and titrated by quantification of HIV-1 p24 and SIV p27 capsid using an enzyme-linked immunosorbent assay (ELISA) (ZeptoMetrix Corp.). Blam-Vpr containing particles were concentrated by ultracentrifuged through a 25% sucrose cushion in TNE buffer (100 mM NaCl, 10 mM Tris-HCl, pH 7.4, and 1 mM EDTA) at 150000 X g for 1 h. All viruses were normalized by single-cycle infectivity assays on HeLa cells.

Cells were transduced as described before for infection of dendritic cells 63 . Cells were preincubated with VLP for 2 h, then, when required, transduced with the indicated virus. The minimal amount of VLP (X⁺) required for maximal helper effect for HIV-1 infection was determined by functional titration. After preincubation, cells were infected for 2 h with the indicated reporter virus at a multiplicity of infection (MOI) of 1. Where indicated, infected cells were treated for 20 h with deoxynucleosides, which consisted of a mixture of dA (D8668), dC (D0776), dG (D0901) and dT (T1895; all from Sigma-Aldrich) as previously described 27 .

Cells were lysed in adequate volumes of M-PER buffer (Pierce) containing 150 mM NaCl and an anti-protease mixture (Sigma) 67 . Protein extracts were separated by SDS-PAGE electrophoresis, transferred onto PVDF membranes, and revealed by a chemiluminescence procedure (CDPStar®, Applied Biosystems). Signals were acquired by a LAS 3000 apparatus (Fujifilm) for further analysis using the Multigauge software (Fujifilm). The following antibodies were used: monoclonal anti-HA tag antibody [16B12], anti-SAMHD1 [1 F9], and anti-ISG15 [3E5] from Abcam; monoclonal anti-α-Tubulin and anti-Actin from Sigma, monoclonal anti-β-Catenin from BD Bioscience, and polyclonal anti-A3G Apo-C17 antibody that recognizes both A3G and A3A (recognized for its smaller size, 59 ) from the AIDS Reagents and Reference Program of the NIH.

The dNTP were quantified as previously described 40 . Briefly, cells were harvested and lysed in cold 65% aqueous methanol, vortexed vigorously, and heated to 95°C for 5 min. Cellular debris were removed from the sample through centrifugation at 13,000 rpm for 3 min. Supernatant was removed and dried in a speed vacuum. The dried samples were suspended in (50 mM Tris-HCl, pH 8.0, and 10 mM MgCl₂)_. Extracts were incubated with 200 fmol of primer/template. An 18-nucleotide primer labeled at the 5^′ end with 32 P (5^′-GTCCCTGTTCGGGCGCCA-3^′) was annealed at a 1:2 ratio to four different 19-nucleotide templates (5^′-NTGGCGCCCGAACAGGGAC-3^′), where ‘N’ represents dG, dC, dA or dC. The reactions contained 2 μl of dNTP cell extract, 4 μl of excess HIV-1 RT, 25 mM Tris–HCl, pH 8.0, 2 mM dithiothreitol, 100 mM KCl, 5 mM MgCl2, and 10 μM oligo(dT) to a final volume of 20 μL. Reactions were incubated for 5 min at 37°C and terminated using 10 μl 40 mM EDTA, 99% formamide and heated at 95°C for 5 min. For analysis, reaction products were separated on a 14% polyacrylamide-urea denaturing gel (SequaGel, National Diagnostics) and analyzed on a PhosphorImager (PerkinElmer). Product extension was quantified by densitometry (Quantity One) and dNTP amount was calculated based on accountability of the dNTP sample dilution factor, so that each sample volume was adjusted to obtain a signal within the linear range of the assay’s standard curve.

The BlaM-Vpr entry test was performed as described in 65 by measuring enzymatic cleavage of a β-lactamase substrate (CCF2/AM, Invitrogen) loaded in target cells. Briefly, cells were incubated with HIV-1 VLP containing the BlaM-Vpr fusion protein for 2 h at 37°C. Cells were then incubated in CO2-independent medium containing the CCF2/AM β-lactamase substrate at room temperature in the dark for 12 hours before fixation and analysis by flow cytometry.

THP-1 cell monolayers were differentiated with PMA for 24 h on coverslips, then treated or not with 1000 U/ml IFN-α for 24 h. After washing and fixing as decribed in 68 , cells were permeabilized for 3 + 15 min with PBT buffer (0.2% Tween 20, 1% bovine serum albumin in PBS). Subsequently, cells were incubated for 1 h at room temperature with primary antibodies in PBT buffer (Alexa Fluor® 488 anti-HA antibody (Invitrogen), or anti-SAMHD1 antibody [3 F5] (ab119751, Abcam). After three washes with PBS, when required, the cells were incubated for 30 min with secondary antibodies. Nuclei were stained with DAPI in mounting media (Vectashield; Vector Laboratories). Images were collected on a Zeiss Axio Observer.Z1 microscope using a 100X oil immersion objective, and analyzed with Metamorph 7 (Molecular Devices).