To facilitate HIV-1 integrase detection during the course of cell infection, we took advantage of an infectious HIV-1 viral clone carrying IN tagged at the C-terminus with the HA epitope (HIV-1_IN-HA 68 ). Epitope fusion at the C-terminus of IN disrupted the Vif open reading frame. However, viral replication is Vif-independent in SupT1 human lymphocytic cells used in our study 69 . Using this virus, we first analyzed IN stability during the early steps of HIV-1 replication. Indeed, the ubiquitin degradation pathway targets IN to the proteasome 70 , which may account for its short half-life in transiently transfected cells 70 71 . SupT1 cells were exposed to HIV-1 for 2 h and harvested at 2 h, 4 h and 6 h p.i. At each time whole cell extracts (WCE) were prepared and viral proteins CA, MA and IN-HA were detected by Western blotting. By optimizing cell infection conditions in order to maximize cell-virus contact surface, incoming CA, MA and IN-HA were readily detectable in cell extracts of infected cells. Infection of cells with a heat inactivated virus prevented detection of viral proteins in cells extracts, indicating that we specifically detected intracellular-associated viral proteins rather than virus absorbed at the cell surface (Figure 1A). Strikingly, the amount of IN-HA dramatically decreased during the 6 h period of time, whereas amounts of CA remained stable (Figure 1A and 1B). To ensure that the HA-tag did not interfere with IN properties, we compared the stability of HA-tagged and non-tagged IN proteins. SupT1 cells were infected either with wild type HIV-1 or HIV-1_IN-HA, and IN was detected in cell lysates at different times post-infection using an anti-IN antibody. Regardless of the presence of a C-terminus HA-tag fusion, IN levels in cell lysates rapidly decreased to become barely detectable at 6 h p.i. (Figure 1B). In addition, levels of late reverse transcription product (LRT), 2-LTR circles and integrated viral DNA were similar at 6 h p.i. for both viruses indicating that the HA tag does not interfere with IN functions during early steps of virus replication (data not shown). Treatment of cells with proteasome inhibitor MG-132 resulted in rapid stabilization of IN-HA in infected cells indicating that once IN entered in the cell, a subset of the protein is actively degraded in a proteasome-dependent manner (Figure 1C). We further analyzed early steps of viral replication of HIV-1_IN-HA. Quantitative PCR analysis indicated that the maximum amount of integrated provirus was reached at 10 h p.i (Figure 1D). Taken together, our results showed rapid degradation of IN protein upon cell infection, coinciding with detection of integrated proviral forms at 6 h p.i. Therefore, we decided to focus our study on early time points (i.e. 2 h p.i. and 6 h p.i.) to monitor the dynamic of IN-containing complexes.

The temporal dynamic of IN interactions with cellular cofactors during early steps of HIV replication was characterized by co-immunoprecipitation experiments. Interactions between IN-HA and host partners TNPO3 and LEDGF/p75 were readily detected at 6 h p.i. (Figure 2). To further assess the rapid changes of IN-associated complexes during the early stages of HIV-1 replication, we performed biochemical fractionation of infected cell lysates. First, fractions from size-exclusion chromatography of WCE were analyzed by Western blotting against TNPO3 and LEDGF/p75. In non-infected cells, the relatively broad elution profile of TNPO3 might reflect its interactions with several different proteins/cargos (fractions 17 to 33, Figure 3A). In contrast, a major peak around 300 KDa (Fractions 22–28, Figure 3A) was observed for LEDGF/p75. In addition, a small portion of LEDGF/p75 was found in higher molecular weight fractions (fractions 17–19, Figure 3A). Similar profiles were observed in infected cells (data not shown), indicating that HIV-1 infection does not trigger major changes in the distribution of LEDGF/p75 or TNPO3-associated complexes.

We then addressed the distribution of viral proteins early following infection. At 2 h p.i., IN-HA was detected in 2 complexes of different size and composition (Figure 3B). The high molecular weight complex (IN complex I), which was excluded from gel separation range (>1.3 MDa), was the most abundant one (fractions 15–18, Figure 3B). Viral proteins capsid (CA) and matrix (MA) were mostly found as free molecules (fractions 36–44, Figure 3B and C), except for a small proportion co-eluting with IN complex I (fractions 15–17 for CA and fraction 17–20 for MA, Figure 3B).

Although present in a minor proportion, low molecular weight (IN complex II) IN-HA complex peaked in fractions 23–25 (around 440 KDa). At 6 h p.i., IN-HA was still found in IN complex I, but shifted readily to IN complex II, while both CA and MA were only detected in smaller peaks (fraction 36–44) (Figure 3C). Noteworthy, IN-HA detection required longer exposure times at 6 h than at 2 h p.i., due to its rapid degradation upon cell infection. Together, our results indicate that shortly after infection, IN associates with at least two distinct complexes.

To detect viral DNA potentially associated with IN complexes at these time points, we performed real-time PCR on DNA extracted from fractionated WCE (Figure 3D). At 2 h and 6 h p.i., viral DNA was detected in fractions 13 to 19 containing the IN complex I (Figure 3D). Elution of the viral cDNA in the void volume of the column (>1.3 MDa) is consistent with its estimated molecular weight (~6.4 MDa for the 9.7 kbp genome). In vitro integration activities of IN-containing complexes eluted from the column were also measured at 6 h p.i. To ensure optimal infection conditions, SupT1 cells were infected with VSV-G pseudotyped HIV-1_IN-HA and the ability of viral DNA to integrate into target plasmid in vitro was quantified by real-time PCR. Concomitant with the detection of viral DNA and IN in fractions containing complex I (Figure 4A and 4B), maximal PIC activity was also reached in these fractions (Figure 4C). Thereby, these data suggest that between 2 h and 6 h p.i. a significant portion of IN accumulates in a low molecular weight complex devoid of viral DNA.

To assess the role of reverse transcription in the accumulation of IN complex II, fractionation of infected cell extracts were conducted in the presence or absence of the non-nucleoside reverse transcriptase inhibitor Nevirapine. Nevirapine treatment did not affect the kinetics of IN degradation (Figure 5A). Furthermore, upon reverse transcription inhibition, IN was still detected in complex II (Figure 5B and 5C), suggesting that the accumulation of IN in low molecular weight complexes does not require the maturation of the reverse transcription complex (RTC) into a PIC.

Next, we tested whether IN complex II could be the result of a post-integration event leading to the release of DNA-free IN from integrated intasomes. SupT1 cells were infected with a virus harboring a catalytically inactivated IN_D116A. At 2 h p.i., cell extracts were fractionated by size exclusion chromatography and collected fractions were analyzed by Western blotting. IN from HIV-1_{IN D116A-HA} eluted in two separate complexes (I and II) that were indistinguishable from the complexes obtained with WT IN (Figure 6), indicating that accumulation of IN complex II is not a post-integration event.

We then investigated whether the distribution of the fractions containing IN was dependent on LEDGF/p75 expression. A LEDGF/p75-knock-down SupT1 cell clone (TL34) and its control polyclonal cell line counterpart (TC3) 72 were infected with VSV-G pseudotyped HIV-1Δenv-Luc. As expected, knock-down of endogenous LEDGF/p75 yielded a 10-fold decrease of viral infectivity, quantified by the viral encoded luciferase activity (Figure 7A). Then, WCE from TC3 or TL34 cells infected with HIV-1_IN-HA virus for 2 h were fractionated by size exclusion chromatography and collected fractions were analyzed by Western blotting. Consistently, we observed similar IN complexes (I and II) in TC3 lysates as in wild type SupT1 lysates (compare Figure 7B with 3B). In sharp contrast, in TL34 cells knocked down for LEDGF/p75 we observed a shift of IN complex II towards a lower molecular weight complex (IN complex III) with a molecular mass around 150 KDa (fractions 27–31) (Figure 7C). Regarding the ability of IN to form stable tetramers when expressed in human cells 51 , we decided to further characterize IN complex III. We thus pooled fractions 30–32 and performed proteins cross-linking with increasing amount of the cross-linker ethylene glycol bis-succinimidylsuccinate (EGS). Addition of 0.25 to 2 mM of EGS yielded cross-linked complexes of 60, 90 and 130 KDa, with a strong predominance of the latter band at the higher concentrations of cross-linker agent (Figure 7D). Thus, depletion of LEDGF/p75 led to the accumulation of IN complex III that contains tetramers of IN.

We analyzed the cellular localization of the IN complexes at 2 h p.i. Cytosolic (CE) and nuclear extracts (NE) prepared from infected control (TC3) or LEDGF/p75-knock-down SupT1 cells (TL34) were fractionated. LEDGF/p75 was only detected in nuclear fractions of control cells, indicative of absence of NE contamination of the CE (Figure 8A). In both TC3 and TL34 cells, IN complex I was distributed between cytosolic and nuclear fractions, indicating that PICs reached the nuclear compartment as early as 2 h p.i. (Figure 8A and 8B). In sharp contrast, IN complex II was strictly associated with nuclear fractions and could not be detected in cytosolic extracts of TC3 cells (Figure 8A). Concordantly, in absence of LEDGF/p75, IN complex III was also only detected in nuclear fractions (Figure 8B). Taken together, our data indicate that IN complex II associates with nuclear fractions in a LEDGF/p75-dependent manner.

To further explore the mechanisms of the accumulation of IN in low molecular weight complexes, we took advantage of an IN class II mutant virus HIV-1_{IN Q168A-HA} that was first identified to be defective for LEDGF/p75 interaction 52 . Mutation of the Q168 residue of IN impairs its ability to form tetramers, resulting in a decrease of its concerted integration activity 43 73 . First, the stability of the mutant IN _Q168A was assessed during infection. WCE from SupT1 cells infected with HIV-1_{IN Q168A-HA} were harvested at 2 h, 4 h, 6 h and 8 h p.i. and viral proteins MA and IN were detected by Western blotting. IN_Q168A levels decreased rapidly during the early phase of HIV-1 infection suggesting that, similar to the WT IN, this mutant was degraded by the proteasome (Figure 9A). Next, the interaction between IN and LEDGF/p75 was assessed in cells infected with HIV-1 encoding wild type IN devoid of the HA tag, IN-HA or IN_Q168A-HA. As expected, the Q168A mutation ablated the LEDGF/p75-IN interaction (Figure 9B). Surprisingly, fractionation of infected WCE indicated that IN_Q168A failed to accumulate in low molecular weight complexes at 2 h and 6 h p.i. (Figure 9C). However, the high molecular weight complexes eluting in the void volume of the column (IN complex I) is still associated both with the cytoplasm and the nuclear fractions of the cells (Figure 9D). Taken together, these results indicated that the nuclear accumulation of IN complex II is impaired for a mutant defective for LEDGF/p75 interaction and tetramerization.

293 T cells were grown in DMEM plus glutamine, antibiotics and 10& decomplemented-FCS (foetal calf serum) (GibcoBRL, Invitrogen). SupT1, TC3 and TL34 cells were grown in RPMI 1640 plus glutamine, antibiotics and 10& decomplemented-FCS (foetal calf serum) (GibcoBRL, Invitrogen). Virus stocks were generated by transfecting 293 T cells with either wild type or IN-HA tagged Lai molecular clone (HIV-1_IN-HA 68 ) using Fugene 6 reagent (Roche). HIV-1_{IN Q168A-HA} and HIV-1_{IN D116A-HA} were generated by site-directed mutagenesis using HIV-1_IN-HA as a template. Mutations were confirmed by sequencing and subcloned back into the PflMI sites of the WT HIV-1_IN-HA. HIV-1 Δenv Luc has been previously described 84 . VSV-G pseudotyped viruses were produced by co-transfecting HIV-1_IN-HA or HIV-1 Δenv Luc with plasmid pMD.G 85 . Twenty-four hours post-transfection, cells were washed with PBS and supernatants were collected at 48 h and 72 h post-transfection. The supernatants were 0.45-μm-filtered and ultracentrifuged at 150,000 g for 1 h 30 minutes at 4°C. Virus pellets from 30 ml of supernatants were resuspended in 75 μl of PBS. An equal volume of RPMI 1640 medium (GIBCO, Invitrogen) supplemented with 10& fetal calf serum and antibiotics was added, and virus stocks were stored at −80°C. HIV-1 CAp24 antigen was quantified by ELISA (Innotests HIV Antigen mAb, Innogenetics, France).

SupT1 cells (2 × 10⁸) were infected with 200 μg of CAp24 (corresponding to a multiplicity of infection-MOI- of 5, as quantified by real time PCR) of HIV-1 or HIV-1 _IN-HA virus in a total volume of 500 μl for 2 h at 37°C. When indicated, 20 μM of MG-132 was added along the course of infection. Two hours later, cells were washed three times in 25 ml of PBS and resuspended in RPMI 1640 medium supplemented with 10& fetal calf serum and antibiotics at a final concentration of 1.5 × 10⁶ cells/ml. At different time post infection cells were harvested, washed twice with 25 ml of PBS and lysed in 3 cell pellet volumes of lysis buffer (20 mM Tris–HCl pH 8.0, 0.3 M KCl, 5 mM MgCl₂, 10& (v/v) glycerol, 0.1& tween 20, 1 mM PMSF and protease inhibitor cocktail from Sigma). Cell lysis was completed by two successive rounds of freeze-thaw, then incubated for 30 min at 4°C on rotating wheel. Two successive centrifugation steps at 16,000 g for 30 min at 4°C allowed complete removal of insoluble materials. The collected supernatant corresponding to soluble proteins within the cells was called whole cell extracts (WCE). Cytosolic and nuclear extracts were obtained by sequential cell fractionation. Cells were first washed in buffer A (20 mM HEPES-KOH pH 7.9, 1.5 mM MgCl₂, 10 mM KCl, 1 mM PMSF and protease inhibitor cocktail from Sigma) and lysed 5 min on ice in 2 cell pellet volumes in buffer A and 0.1& (v/v) NP-40. After centrifugation of the cell lysate at 10,000 g for 5 min, the supernatant corresponding to the cytosolic extract (CE) was collected and the pellet was resuspended in 3 volumes of buffer B (20 mM HEPES-KOH pH 7.9, 1.5 mM MgCl₂, 0.5 M NaCl, 25& (v/v) glycerol, 1 mM PMSF and protease inhibitor cocktail from Sigma). After 15 minutes incubation on ice, the nuclear extract (NE) was collected by centrifugation for 15 minutes at 14,000 g. CE and NE were stored at −80°C.

The assay is based on the quantification of integration events of viral cDNA into a parental vector (pTZ-19R) 8 with the following modifications. SupT1 cells were infected for 2 h with a VSV-G pseudotyped HIV-1_IN-HA virus and WCE were prepared at 6 h p.i. as previously described. WCE (3 mg) were injected into Superdex 200 10/300 GL gel filtration column. Each fraction was collected and treated with RNase A (final concentration at 20 μg/ml) for 20 minutes at RT in a final volume of 250 μl. Integration reaction assay was performed at 37°C for 45 minutes by mixing each fraction with 1 μg of target plasmid pTZ-19R in integrase reaction buffer (INRB: 20 mM HEPES-KOH pH 7.4, 150 mM KCl, 1 mM MgCl₂, 4& glycerol, and 1 mM DTT added just before starting the reaction) in a final volume of 350 μl. For each fraction, negative control was performed by omitting the target plasmid in the integration reaction assay. The reaction was stopped by adding 0.5& SDS, 8 mM EDTA and 0.5 mg/ml proteinase K for 1 h at 56°C. DNA was then extracted with phenol and phenol:chloroform:isoamylalcohol 25:24:1, using glycogen as carrier, and stored at −20°C. Integration events were quantified using a two-step PCR reaction. In the first round PCR, integrated HIV-1 sequences were amplified using the HIV-1 specific LM667 primer 78 together with primers annealing to opposite strand of pTZ-19R in tail-to-tail fashion. TZ2414 primer sequence is 5′- GTTGTTCCAGTTTG GAACAAGAGTC-3′. TZ2413 primer sequence is 5′- ACTCAACCCTATCTCGGTCTATTC-3′. To evidence background PCR signal, TZ2414 and TZ2413 primers were omitted from each integration reaction. We performed the second nested-PCR steps using conditions previously described 78 . Results shown are indicated as relative units of specific signal (signal from PCR with target and with TZ2414/TZ2413 primers) minus unspecific signal (signal from PCR without target or without TZ2414/TZ2413 primers). The copy number of integrated DNA was determined in reference to a standard curve obtained by concomitant two steps amplification of serial dilution of the standard pNLX-HIVLTR vector. This construct was generated by amplifying the LTR region (n.t. 1–702, numbering based on NL4.3 sequence) and cloning into BamH1 and Pst1 restriction endonuclease sites of pTZ-19R.

Prior to infection, viral stocks were treated 1 h at 37°C with 100 U per ml of DNAseI (Roche applied Science). SupT1 cells (6x10⁶) were infected with viral doses corresponding to 6 μg of HIV-1_IN-HA CAp24 antigen in 12-wells plates. At 2 h p.i., cells were washed twice in PBS. At 2 h, 4 h, 6 h, 8 h and 10 h p.i. cells were harvested, washed twice in PBS and DNA was extracted using the QIAamp Blood DNA Minikit (Qiagen). Quantifications of viral DNA were performed by real-time PCR using the LightCycler 480 system (Roche Applied Science). Primers, probes and PCR run conditions were described previously 84 . The copy numbers of HIV-1 late reverse transcription product (LRT) and 2-LTR circles were determined using standard curves obtained by amplification of cloned DNA containing the matched sequences. The copy number of integrated HIV-1 DNA was determined in reference to a standard curve generated by concomitant two-stage PCR amplification of a serial dilution of the standard HeLa HIVR7-Neo cell DNA 78 . Copy numbers of each viral form were normalized with the number of cells obtained by the quantification by PCR of the β-globin gene according to the manufacturer instructions (Roche Applied Science).

For co-immunoprecipitation experiments, 3 mg of WCE were mixed with 50 μl (50& slurry) of anti-HA affinity matrix (clone 3 F10, Roche Applied Science) supplemented with protease inhibitor cocktail and 1 mM PMSF for 3 h at 4°C on rotating wheel. Beads were washed three times with 15 volumes PBS-0.1& tween 20 for 5 minutes on a rotating wheel at 4°C. IN-HA complexes were directly resuspended in 1× loading sample buffer and boiled for 5 min.

Whole cell extracts were subjected to size exclusion chromatography at 4°C using an AKTA purifier system (GE Healthcare). Four mg of WCE were injected on a Superdex 200 10/300 GL column and 400 μl fractions were collected at a flow rate of 0.5 ml/min of WCE buffer. Proteins were precipitated over-night at 4°C in 10& trichloroacetic acid, washed twice in cold acetone and analyzed by Western blotting. Protein standards (Sigma) were fractionated under same conditions to estimate the size of the eluted protein complexes. When required, DNA was extracted from each fraction using the QIAamp Blood DNA Minikit (Qiagen).

Cross-linking reactions were performed on gel filtration fractions collected in 20 mM HEPES-KOH pH 7.2, 0.3 M KCl, 5 mM MgCl₂, 10& glycerol, 0.1& tween 20, 1 mM PMSF and protease inhibitor cocktail (Sigma). Fractions 30–32, corresponding to part of IN complex III, were pooled and proteins were cross-linked with 0.25 mM to 2 mM of the cross-linker ethylene glycol bis-succinimidylsuccinate (EGS, freshly prepared) for 30 minutes at room temperature. The reaction was stopped by adding Tris–HCl pH 7.5 at a final concentration of 50 mM for 15 minutes at room temperature.

Proteins were separated by 4-12& gradient SDS-PAGE (Invitrogen), transferred onto nitrocellulose membranes, and revealed by Western-blotting using the following antibodies as indicated: anti-CAp24 (AIDS Research and Reference Reagent Program), anti-IN (Santa Cruz), anti-MA (Hybridolab), anti-HA-HRP (clone 3 F10, Roche Applied Science), LEDGF/p75 (BD Bioscience), TNPO3 (Abcam), and α-tubulin (Sigma-Aldrich).