Cellular TRIM33 restrains HIV-1 infection by targeting viral integrase for proteasomal degradation

Abstract

Productive HIV-1 replication requires viral integrase (IN), which catalyzes integration of the viral genome into the host cell DNA. IN, however, is short lived and is rapidly degraded by the host ubiquitin-proteasome system. To identify the cellular factors responsible for HIV-1 IN degradation, we performed a targeted RNAi screen using a library of siRNAs against all components of the ubiquitin-conjugation machinery using high-content microscopy. Here we report that the E3 RING ligase TRIM33 is a major determinant of HIV-1 IN stability. CD4-positive cells with TRIM33 knock down show increased HIV-1 replication and proviral DNA formation, while those overexpressing the factor display opposite effects. Knock down of TRIM33 reverts the phenotype of an HIV-1 molecular clone carrying substitution of IN serine 57 to alanine, a mutation known to impair viral DNA integration. Thus, TRIM33 acts as a cellular factor restricting HIV-1 infection by preventing provirus formation.

Fig. 1

Degradation of retroviral integrases by the host ubiquitin-proteasome system. a Schematic representation of the HIV-1 IN stability assay. HeLa cells were cotransfected with plasmids expressing Flag-tagged IN or EGFP; 48 h later, cells were treated with cycloheximide (CHX) in the presence and absence of the proteasome inhibitor MG132. b Measurement of HIV-1 IN stability. Flag-IN-expressing HeLa cells were treated with CHX (30 μg/ml) with or without MG132, followed by protein detection at different time points using an anti-Flag antibody. Tubulin levels served as a loading control and EGFP levels as a transfection and specificity control. c Quantification of HIV-1 IN protein levels in the presence and absence of MG132 in CHX-treated cells at the indicated time points. Data are mean ± SEM; n = 3 independent experiments. d Schematic representation of the MLV IN stability assay. e Measurement of MLV- IN stability. Experiments were performed as in panel c. f Quantification of MLV-IN protein levels. Data are mean ± SEM; n = 3 independent experiments. g Knock-down of cellular factors to assess HIV-1 IN levels. HeLa cells were transfected with a plasmid expressing pFlag-IN and either transfected 24 h in advance with siRNAs against selected cellular factors or treated, 48 h later, with MG132. h Representative experiment showing the levels of HIV-1 IN after knock-down of selected cellular factors by RNAi, as indicated. Cell treatment with MG132 served as a positive control. IN levels were analyzed in whole cell lysates by immunoblotting with anti-Flag antibody. Immunoblotting for Tubulin served as a loading control. The bottom panel shows a western blotting for EGFP in a representative experiment to show lack of effect of any of the tested siRNAs on EGFP levels. i Quantification of HIV-1 IN levels in whole cell extracts of cells in which the indicated factors were knocked down by siRNA transfection. Data are normalized for Tubulin levels and expressed as fold over cell treatment with non-targeting NT2 siRNA control. Data are mean ± SEM; n = 3 independent experiments; *P < 0.05; **P < 0.01; one-way ANOVA

Fig. 2

High-throughput, siRNA-based screening to identify cellular factors regulating HIV-1 integrase stability. a Workflow for the siRNA-based screening. Cellular fluorescence, as surrogate of IN levels, was analyzed by automated, high-content fluorescent microscopy. b Results of screening. The graphs show the log10 values of the fold change of EGFP-positive cells over control in the two replicate screenings (R1 and R2). The dotted lines show 2x increase over Control (pool of results using 4 non-targeting siRNAs and mock-transfected cells). The 6 siRNAs in red are those that were in the top 10 in both screenings. The 4 siRNAs in green are those that were among the top 10 in one of screening while anyhow showing an effect ≥2 fold over control in the other screening. The effect of MG132 is shown in blue. c Confirmation of effective silencing of Pin1, TRIM33, FBOX28, RNF31, RNF125, RFPl3, and DTX by immunoblotting with the respective antibodies. Cells transfection of non-targeting siRNA1 (siNT1) was used as a control. HSC70 served as a loading control. d Representative immunoblot showing the levels of HIV-IN after knock-down of the top 10 E3 ligases from the screening. HeLa cells were transfected with siRNAs against the identified factors or with a siRNA against Pin1, followed by transfection of Flag-IN and EGFP. Forty-eight hour later the levels of IN were assessed by anti-Flag immunoblotting. Tubulin served as a loading control; β-catenin was used to confirm effect of the MG132 treatment. The bottom panel shows a western blotting for EGFP in a representative experiment to show lack of effect of any of the tested siRNAs on EGFP levels. e Quantification of the levels of HIV-1 IN after knock-down of the top 10 E3 ligases identified by the screening. Experiments were performed as in panel d. IN levels are expressed after normalization for tubulin and as fold over siNT1. Data are mean ± SEM; n = 3 independent experiments; *P < 0.05; **P < 0.01; one-way ANOVA. f Representative high-content microscopy images showing of EGFP-IN-expressing cells after depletion of four cellular ubiquitin-conjugation factors (TRIM33, FBXO28, siDTX2, and siUBE2J2) or Pin1

Fig. 3

Cellular TRIM33 is only TIF1 family member that negatively regulates HIV-1 integrase stability. a Schematic representation of domain organization of TIF1 subfamily members. b Effect of anti-TRIM24 siRNAs. The immunoblots show the levels of TRIM24, TRIM33, TRIM28, and HIV-1 IN in HeLa cells transfected with individual siRNAs against TRIM24 (#4, #5, #6, and #7), their pool (Pool) or non-targeting siRNA-2 (siNT2). MG132 was used to assess the effect of inhibiting proteasome-mediated protein degradation. c Effect of anti-TRIM28 siRNAs. Same as in panel b using siRNAs against TRIM28. d Effect of anti-TRIM33 siRNAs. Same as in panel b using siRNAs against TRIM33. e Stability of HIV-1 IN protein in siRNA-treated cells. Representative immunoblots showing the levels of HIV-1 IN after HeLa cell transfection with siRNAs against TRIM33, Pin1, and LEGF/p75 at different times and treatment with cycloheximide (30 µg/ml) to block protein degradation. Tubulin served as a loading control. f Quantification of HIV-1 IN levels. Experiments were performed as in panel e. IN amounts are normalized to Tubulin and expressed as percent over IN at time = 0 for each treatment. Data are mean ± SEM; n = 3 independent experiments. g Effects of siRNAs on the levels of HIV-1 and MLV IN proteins. The figure shows representative immunoblots using lysates of HeLa cells transfected to express either Flag-IN from HIV-1 (upper part) or MLV-IN (lower part)and siRNAs against the indicated factors. Tubulin and HSC70 served as loading controls. h Quantification of the effects of siRNAs against the indicated factors or treatment with MG132 on the levels of HIV-1 IN. IN levels are expressed after normalization for tubulin and as fold over non-targeting siRNA siNT1. Data are mean ± SEM; n = 3 independent experiments; *P < 0.05; **P < 0.01; one-way ANOVA. i Same as in panel g for MLV-IN. Normalization was for cellular HSC70. Data are mean ± SEM; n = 3 independent experiments; *P < 0.05; one-way ANOVA

Fig. 4

Cellular TRIM33 binds HIV-1 integrase and regulates its degradation by poly-ubiquitination. a Endogenous TRIM33 interacts with HIV-1 IN in vivo. HEK293T cells were transfected with the plasmids indicated on the top of the panel and immunoprecipitated with anti-Flag M2 beads, followed by immunoblotting with anti-TRIM33 antibody (upper blot). The lower three blots show the levels of expression of the transfected proteins in cell lysates. b Schematic representation of the HIV-1 IN domains used for the pull-down assay in panel c. NTD N-terminal domain, CCD catalytic core domain, CTD C-terminal domain. c TRIM33 binds the HIV-1 IN C-terminal domain (CTD) in in vitro pull-down assay. The indicated fragments of the HIV-1 IN protein fused to GST or GST alone were incubated with in vitro translated [³⁵S]-TRIM33, extensively washed, and then resolved by SDS-PAGE. The middle pictures shows the gel exposed to a phosphoimager from a representative experiment along with the quantification of the amount of bound proteins expressed as a percentage of radiolabeled input. The lower panel shows the protein gel stained with Coomassie blue to visualize proteins. Quantification in the upper graph reports mean ± SEM; n = 3 independent experiments; **P < 0.01; one-way ANOVA. d HIV-1 IN is degraded through Lys48-linked polyubiquitination. HEK293T cells were transfected with Flag-IN and HA-Ubiquitin for 48 h, as indicated. The ubiquitination profile of IN was visualized after immunoprecipitation with anti-Flag M2 beads using anti-ubiquitin, anti-K48, anti-K63, anti-HA, and anti-Flag antibodies. e TRIM33 Inhibition reduces IN polyubiquitination. HEK293T cells were co-transfected with Flag-IN and HA-Ubiquitin for 36 h, as indicated. IN polyubiquitination was visualized after immunoprecipitation with anti-Flag M2 beads using an anti-IN antibody. f Quantifications of the amount of polyubiquitinated IN protein (Ubi-IN) in whole cell extracts, normalized over total IN. Experiments were performed as in panel e

Fig. 5

Mapping the TRIM determinants determining integrase stability. a Schematic representation of the TRIM33 protein and its catalytically inactive mutants in the RING and PHD domains [RING(CA) and PHD(CA) respectively]. b Integrity of the TRIM33 RING motif is essential for HIV-1 IN ubiquitination. HEK293T cells were transfected with combinations of plasmids expressing HA-ubiquitin, EGFP-IN, Flag-TRIM33 and its catalytically inactive mutants (RING and PHD), and pcDNA3 as an empty vector. Cells were harvested after 5 h treatment with proteasome inhibitor and cell lysates were immunoprecipitated with anti-GFP antibody. Ubiquitin-conjugated IN (Ubi-IN) was detected by immunoblotting with anti-HA antibody (upper picture). The lower pictures show the levels of expression of the transfected plasmids in whole cell lysates control expression of transfected proteins after immunoblotting with the respective antibodies. c Quantifications of the amount of ubiquitinated IN protein (Ubi-IN) in whole cell extracts, normalized over total IN. Experiments were performed as in panel b. Data are mean ± SEM; n = 3 independent experiments; *P < 0.05; **P < 0.01; one-way ANOVA. d Representative immunoblot showing the stability of HIV-IN in HeLa cells stably expressing an shRNA against TRIM33 (TRIM33 KD) and transfected to overexpress wild-type TRIM33 or the C125A/C128A catalytically inactive (RING(CA). Cells were co-transfected with Flag-IN and EGFP, as transfection and specificity control; 48 h later, the levels of IN were assessed by anti-Flag immunoblotting. Tubulin served as a loading control; β-catenin was used to confirm effect of MG132 treatment; knock-down and overexpression of TRIM33 variants were verified by anti-TRIM33 immunoblotting. e Quantification of the effects of TRIM33 knock down and overexpression on the levels of HIV-1 IN. Results of three independent experiments (mean ± SEM) performed as in panel d are expressed, after normalization for Tubulin, as a percentage of time = 0 for each treatment

Fig. 6

TRIM33 inhibits HIV-1 infection. a Analysis of the effect of stable knockdown of TRIM33 in multiple round infection with wild-type HIV-1. SupT1 cells were first transduced with lentiviral vector expressing different anti-TRIM33 shRNAs; after selection with puromycin, cells were infected with wild-type HIV-1_BRU or HIV-1_BRUIN(S57A), followed by analysis of infection over time by p24 ELISA. b Western blotting analysis showing endogenous TRIM33 levels in SupT1 cells transduced with lentiviral vectors expressing the indicated shRNAs and selected with puromycin. Tubulin was used as a loading control. c. Stable knockdown of TRIM33 enhances wild type HIV-1 replication and rescues replication of the IN(S57A) mutant. The graph show the levels of p24 in control Sup1 cells or cells previously transduced with anti-TRIM33 shRNA #4, which efficiently knocks down TRIM33 expression, after infection with the two viruses. The results shown are representative or four independent experiments. d Quantification of the levels of integrated HIV-1 DNA in SupT1 cells, expressing an shRNA control or shRNA #4 against TRIM33, and infected with wild-type HIV-1_BRU or HIV-1_BRUIN(S57A), at day 3 after infection. Data are mean ± SEM; n = 3 independent experiments; **P < 0.01; n.s. not significant; one-way ANOVA. e Silencing of TRIM33 protects both wild-type IN and mutant IN(S57A) from degradation. Expression plasmids for the two IN proteins were transfected into HeLa cells in which TRIM33 was knocked down by RNAi; cells were then treated with cycloheximide 30 µg/ml for the indicated time points prior to lysis. The panel shows a representative immunoblots using antibodies against IN, TRIM33 and tubulin as a loading control. f Quantification of IN levels in the presence of cycloheximide for the indicated time points in the whole cell extracts of SupT1 cells treated with siRNAs against TRIM33 or not targeting and transfected with either wt IN or IN(S57A). After normalization for Tubulin, data (mean ± SEM; n = 3 independent experiments) are expressed as a percentage of time = 0 for each treatment

Fig. 7

Interaction of TRIM33 with integrase from infectious HIV-1. a Analysis of IN protein levels in TRIM33-knockdown (TRIM33 KD) cells during HIV-1 infection. Cells stably expressing an anti-TRIM33 shRNA were infected for 1 h with HIV-1; at the indicated time points after infection, cells were washed with PBS and the levels of IN were analyzed. b Western blot showing IN levels in control and TRIM33 depleted cells Beta-actin was used as a loading control; TRIM33 downregulation was verified using an anti-TRIM33 antibody. c TRIM33 interacts with IN during HIV-1 infection. SupT1 cells were infected with HIV-1 in the presence of MG132. Either TRIM33 or IN were immunoprecipitated with the respective antibodies, followed by immunoblotting to test the presence of the reciprocal protein. d Analysis of the effect of stable knockdown and overexpression of TRIM33 variants in multiple round infection with wild-type HIV-1. SupT1 cells stably expressing an hRNA targeting the TRIM33 3′-UTR (selected after transduction with a lentiviral vector and selection for puromycin) were transduced with lentiviral vectors expressing the TRIM33 coding sequence (cs) or the coding sequence of the RING(CA) mutant. After selection with hygromycin, cells were infected with wild type HIV-1_BRU, followed by analysis of infection over time by p24 ELISA. e Western blotting analysis showing TRIM33 levels after transduction with shRNA targeting 3′-UTR of TRIM33 alone or after transduction of wild-type or mutant TRIM33 (TRIMM33cs or RING(CA)cs. Tubulin was used as a loading control. f real-Time RT-PCR analysis showing TRIM33 expression after transduction with the shRNA targeting the 3′-UTR of TRIM33 alone or after transduction of wild type or mutant TRIM33. g Stable knockdown of TRIM33 enhances wild-type HIV-1 replication, while overexpression of its catalytically active cDNA reduces it. The graph shows the levels of p24 in control Sup1 cells or cells previously transduced with anti-TRIM33 shRNA targeting the 3′-UTR of the protein, with or without transduction with a lentiviral vector expressing the coding sequence of wild type TRIM33 (TRIM33cs) or that of the catalytically inactive RING(CA) mutant (RING(CA)cs), after infection with wild-type virus. The results are representative of three independent experiments