|Home | About | Journals | Submit | Contact Us | Français|
Posttranslational modifications of the HIV-1 Tat protein have emerged as critical regulatory mechanisms that fine-tune interactions of the viral transactivator with TAR RNA and cellular cofactors. Here, we identify the lysine methyltransferase Set7/9 (renamed KMT7) as a novel co-activator of HIV transcription. Set7/9-KMT7 associates with the HIV promoter in vivo and monomethylates lysine 51, a highly conserved residue located in the RNA-binding domain of Tat. Knockdown of Set7/9-KMT7 suppresses Tat transactivation of the HIV promoter, but does not affect the transcriptional activity of methylation-deficient Tat (K51A). Set7/9-KMT7 itself binds TAR RNA and forms a complex with Tat and the positive transcription elongation factor P-TEFb. Our findings uncover novel RNA-binding properties of Set7/9-KMT7 and demonstrate a positive role of Tat methylation in early steps of the Tat transactivation cycle.
The Tat protein of human immunodeficiency virus 1 (HIV-1) plays an essential role in HIV gene expression by promoting efficient transcriptional elongation of viral transcripts. During the early steps of HIV transcription when Tat is absent, cellular RNA polymerase II initiates transcription at the HIV promoter located in the 5′ LTR, but elongation is impaired (Kao et al., 1987; Toohey and Jones, 1989). When Tat is produced, transcriptional elongation becomes highly effective, resulting in the sustained production of full-length viral transcripts and high-level viral replication.
Tat binds to a specific RNA stem-loop structure called TAR that forms spontaneously at the 5′ ends of nascent viral transcripts (Barboric and Peterlin, 2005). Tat binding to TAR RNA involves a highly conserved arginine-rich motif (ARM), located between amino acids (aa) 49 and 57 in Tat, and the positive transcription elongation factor b (P-TEFb). Tat and the Cyclin T1 component of P-TEFb bind TAR RNA cooperatively and induce phosphorylation of the C-terminal domain of RNA polymerase II by the cyclin-associated kinase CDK9, a critical step to improve the elongation competence of the polymerase complex (Herrmann and Rice, 1993; Kim et al., 2002; Wei et al., 1998; Zhu et al., 1997).
Tat also recruits the SWI/SNF chromatin remodeling complex to the HIV promoter (Agbottah et al., 2006; Ariumi et al., 2006; Mahmoudi et al., 2006; Tréand et al., 2006) and interacts with several histone-modifying enzymes, including histone acetyltransferases p300/CBP (renamed KAT3B/KAT3A), p300/CBP-associated factor (PCAF)/human GCN5 (KAT2B/KAT2A), TAF1 (former TAFII250) and Tip60 (KAT5) (Benkirane et al., 1998; Col et al., 2001; Creaven et al., 1999; Deng et al., 2000; Hottiger and Nabel, 1998; Weissman et al., 1998). While it was originally believed that interactions with histone-modifying enzymes serve to relieve the elongation block imposed by the nucleosomal organization of the HIV promoter (Verdin et al., 1993), growing evidence shows that Tat itself is a target. Tat is acetylated by p300-KAT3B, human GCN5-KAT2A and PCAF-KAT2B, deacetylated by the NAD+-dependent deacetylase SIRT1, ubiquitinated by the E3 ubiquitin ligase Hdm2, and methylated by the protein arginine methyltransferase PRMT6 and the lysine methyltransferase SETDB1-KMT1E (Boulanger et al., 2005; Brès et al., 2003; Col et al., 2001; Kiernan et al., 1999; Ott et al., 1999; Pagans et al., 2005; Van Duyne et al., 2008).
Known Tat modifications can either positively (acetylation/deacetylation) or negatively (methylation) influence Tat transcriptional activity by regulating interactions of Tat with TAR RNA and Cyclin T1. Acetylation of Tat also generates a specific interaction domain with the PCAF bromodomain (Brès et al., 2002; Dorr et al., 2002; Kaehlcke et al., 2003; Mujtaba et al., 2002) and enhances Tat binding to the Brg-1 subunit of the SWI/SNF chromatin-remodeling complex (Mahmoudi et al., 2006). Tat ubiquitination also enhances Tat transactivation, although the molecular mechanism associated to this function is not yet known (Brès et al., 2003).
Set7/9-KMT7, a lysine monomethyltransferase, was originally identified as an enzyme that modifies K4 in histone H3 (Nishioka et al., 2002; Wang et al., 2001), but was later found to methylate nonhistone proteins. Its known substrates include the TAF10 component of the TFIID complex, the tumor suppressor p53 and estrogen receptor α, all of which are positively influenced by Set7/9-KMT7-mediated modifications (Chuikov et al., 2004; Ivanov et al., 2007; Kouskouti et al., 2004; Kurash et al., 2008; Subramanian et al., 2008). However, methylation of DNA methyltransferase 1 (DNMT1) was found to destabilize and inactivate this factor (Estève et al., 2009; Wang et al., 2009). The p65 subunit of NF-κB is also methylated by Set7/9-KMT7, although different methylation sites and different functions associated to Set7/9-KMT7 mediated methylation of p65 have been reported (Ea and Baltimore, 2009; Yang et al., 2009). Notably, Set7/9-KMT7 acts as an important co-activator of NF-κB-dependent gene expression in monocytes and pancreatic beta cells and plays a critical role in the regulation of p65 expression during hyperglycemia (Brasacchio et al., 2009; Deering et al., 2008; El-Osta et al., 2008; Li et al., 2008). PCAF-KAT2B was recently identified as a substrate of Set7/9-KMT7 but the function of PCAF-KAT2B methylation remains unknown (Masatsugu and Yamamoto, 2009).
The emerging importance of Set7/9-KMT7 as a central regulator of gene expression in diseases, such as cancer and diabetes, prompted us to investigate whether the enzyme also plays a role in HIV transcription. Our data identify Tat as a novel substrate of Set7/9-KMT7 and demonstrate that monomethylation by Set7/9-KMT7 defines a positive step in the Tat transactivation cycle.
To test whether Tat is methylated by Set7/9-KMT7, full-length synthetic Tat protein (aa 1–72) was incubated with recombinant Set7/9-KMT7 enzyme and radiolabeled S-adenosyl-L-methionine (SAM). Reactions were resolved by gel electrophoresis and developed by autoradiography. Tat was methylated in response to increasing amounts of Set7/9-KMT7 (Figure 1A). No spontaneous methylation was observed with SAM alone. As expected, Set7/9-KMT7 also methylated histone H3, a known substrate of Set7/9-KMT7. In contrast, methylation by Set7/9-KMT7 was not observed with recombinant IκBα (Figure 1A). The lysine methyltransferase G9a did not methylate Tat, but methylated histones as reported (Figure 1B) (Tachibana et al., 2001). These results show that Tat is a specific in vitro substrate of Set7/9-KMT7.
To map the site of methylation in Tat, we generated short synthetic Tat peptides and subjected them to in vitro methylation assays. Methylation by Set7/9-KMT7 was observed with one peptide (aa 45–58), corresponding to the Tat ARM (Figure 1C). Analysis by MALDI TOF mass spectrometry showed that the Tat ARM is monomethylated by Set7/9-KMT7: the molecular mass of the peptide increased by 14 Da, corresponding to the addition of a single methyl group (Figure 1E and Figure S1A). Mass spectrometry further revealed that recombinant Set7/9-KMT7 methylates the Tat ARM without addition of exogenous SAM, possibly by utilizing SAM co-purified from E. coli (Figure S1B).
The Tat ARM contains two lysines, K50 and K51. Both residues are strictly conserved among HIV-1 isolates. To determine which lysine is methylated by Set7/9-KMT7, we generated ARM peptides containing alanine substitutions at position K50, K51, or both. Methylation by Set7/9-KMT7 was abrogated when K51 or both lysines were mutated while mutation of K50 alone had no effect, indicating that K51 is the target of Set7/9-KMT7 in the Tat ARM (Figures 1D and E).
We then examined Tat methylation in cells. Rabbits were immunized with synthetic peptides corresponding to the monomethylated Tat ARM. The resulting antiserum (α-meARM) specifically recognized ARM peptides carrying a monomethyl group at K51, while no crossreactivities with unmodified, di- or trimethylated ARM peptides were observed (Figure 2A). The antiserum also recognized full-length synthetic Tat (aa 1–72) after in vitro methylation by Set7/9-KMT7, but did not react with the unmodified protein (Figure 2B). ARM peptide and Tat levels in these reactions were visualized by immunoblotting with streptavidin-horseradish peroxidase conjugate (SA-HRP) that recognized the biotin label attached to the N terminus of the peptides (Figures 2A and B).
Next, we examined 293 cells transfected with expression vectors for Tat. We expressed wild-type or mutant Tat proteins in which K51 had been changed to alanine (K51A). Tat proteins were isolated from cell lysates via their FLAG epitope tags and examined by western blotting with α-meARM antibodies. Tat methylation was detected in samples expressing wild-type, but not mutant, Tat, demonstrating that Tat is methylated at K51 in cells (Figure 2C). Expression of wild-type and mutant Tat proteins was equivalent as confirmed by western blotting with α-FLAG antibodies (Figure 2C).
When we coexpressed Set7/9-KMT7 with Tat, Tat methylation increased, indicating that Set7/9-KMT7 can methylate Tat in cells (Figure 2D). Recognition of Tat was blocked when α-meARM antibodies were preincubated with monomethylated ARM peptides before western blotting. No effect was observed after incubation with unmodified ARM peptides, excluding crossreactivity of the antiserum with unmodified Tat (Figure 2D).
To examine whether Tat is also monomethylated in CD4+ T cells, a natural host cell of HIV-1, we used the Jurkat-derived A2 cell line. A2 cells harbor a latent minimal HIV-1 genome that, upon stimulation with phorbol 12-myristate-13-acetate (PMA) or tumor necrosis factor α (TNF-α), express Tat and enhanced green fluorescent protein (GFP) from the HIV long terminal repeat (LTR) (Jordan et al., 2003). Methylated Tat was detected in TNF-α-stimulated A2 cell lysates after immunoprecipitation with α-FLAG antibodies, demonstrating that Tat is methylated under conditions that mimic physiological conditions of infection (Figure 2E).
To investigate the biological role of Set7/9-KMT7 during HIV transcription, we transfected HeLa cells with an HIV LTR luciferase construct and expression vectors for Tat and Set7/9-KMT7. Wild-type, but not catalytically inactive Set7/9-KMT7 (H297A), functioned in synergy with Tat in the transactivation of the HIV LTR (Figure 3A). Set7/9-KMT7 overexpression also activated HIV LTR activity by ~ 2-fold in the absence of Tat. In parallel, we performed reporter assays with a luciferase construct containing the elongation factor 1 α (EF-1α) promoter, which was driving Tat expression in these experiments (Figure 3A). No effect of Set7/9-KMT7 overexpression was observed on the transcriptional activity of the EF-1α promoter, excluding that Set7/9-KMT7 activated Tat expression in these experiments.
We then introduced siRNA oligonucleotides specific for Set7/9-KMT7 into HeLa cells to downregulate endogenous Set7/9-KMT7 expression before transfection of the HIV LTR luciferase reporter and Tat. Set7/9-KMT7 expression was efficiently downregulated 72 h after siRNA transfections (Figure 3B). At this time, Tat transactivation of the HIV LTR was ~3-fold reduced, confirming that cellular Set7/9-KMT7 expression is necessary for full Tat transactivation (Figure 3B).
The transcriptional activity of mutant Tat (K51A) was ~3-fold reduced as compared to wild-type Tat, and no further reduction was observed after knockdown of Set7/9-KMT7 (Figure 3B). In addition, no effect of the siRNAs was observed when the HIV LTR reporter was expressed in the absence of Tat, supporting the model that endogenous Set7/9-KMT7 activates HIV transcription through K51 methylation in Tat (Figure 3B). Similar results were obtained with a Tat mutant which contained an arginine at position 51 (TatK51R), excluding that the transcriptional defect of the K51 mutant was caused by the change in charge instead of the lack of methylation (Figure 3C). Of note, the addition of a methyl group to a lysine residue does not neutralize the positive charge of this position. We verified by western blotting that wild-type and mutant Tat proteins were expressed equally in these experiments, and that no difference in Tat expression was observed when Set7/9-KMT7 expression was downregulated (Figures 3B and C).
To test whether Set7/9-KMT7 activates Tat in infected T cells, we introduced Set7/9-KMT7 siRNAs into Jurkat A2 cells by Amaxa nucleofection. To induce Tat and GFP expression, cells were treated with PMA three days after siRNA transfection, when Set7/9-KMT7 expression was downregulated by 80% (Figure 4A). The rate of GFP+ cells was decreased by ~50% in Set7/9-KMT7 knockdown cells, confirming a positive role of Set7/9-KMT7 in HIV gene expression (Figure 4A).
We also examined the role of Set7/9-KMT7 in primary CD4+ T cells and generated lentiviral vectors expressing different shRNAs against Set7/9-KMT7 (shSet7/9-1 and shSet7/9-2). These vectors also express the mCherry protein under the control of the EF-1α promoter, which allows for the identification of successfully transduced cells (Grskovic et al., 2007; Ventura et al., 2004). Both shRNAs efficiently suppressed Set7/9-KMT7 expression after infection of Jurkat cells followed by sorting of mCherry+ cells (Figure 4B). They also caused a 40% reduction of Set7/9-KMT7 expression in unsorted, primary CD4+T cells in which ~40% of the cells were mCherry+ (Figure S2). These unsorted infected CD4+ T cells were subsequently infected with a molecular clone of the viral isolate HIVNL4-3 (HIV-R7/E−/GFP). This virus contains the GFP open reading frame in place of nef, allowing identification of infected cells by flow cytometry, and a frameshift mutation in the env gene, thereby restricting infection to a single cycle (Jordan et al., 2003). Knockdown of Set7/9-KMT7, as marked by mCherry expression, caused a 40% (shSet7/9-1) and 60% (shSet7/9-2) reduction in GFP expression as compared to cells expressing a non-targeting shRNA control (Figure 4B). No effect of the Set7/9-KMT7 shRNAs was observed when cells were infected with an HIV-based lentiviral vector expressing GFP from the EF-1α promoter instead of the HIV LTR (Figure 4B). These data confirm that Set7/9-KMT7 regulates proviral gene expression during HIV infection in primary CD4+ T cells while other steps of the viral life cycle such as reverse transcription, nuclear import and integration are unaffected.
To examine whether Set7/9-KMT7 is physically recruited to the HIV promoter during active gene expression, chromatin was prepared from Jurkat A2 cells treated with PMA or the solvent control and was immunoprecipitated with antibodies directed against endogenous Set7/9-KMT7. Real-time PCR analysis of the immunoprecipitated material showed that Set7/9-KMT7, while present at the uninduced HIV LTR at low concentrations, was 5-fold enriched in response to PMA (Figure 5A). No association of Set7/9-KMT7 with the GFP open reading frame was observed in uninduced cells; a slight signal was detected in response to PMA (Figure 5A). These data demonstrate that Set7/9-KMT7 associates in vivo with the HIV promoter and is specifically enriched during active transcription.
Since the enhancement of Set7/9-KMT7 recruitment to the HIV LTR coincides with Tat expression, we tested whether Tat and Set7/9-KMT7 interact. Tat expression was induced in A2 cells after treatment with TNF-α, and cellular lysates were subjected to immunoprecipitation with Set7/9-KMT7 antibodies. Tat was detected in the immunoprecipitated material by western blotting with FLAG antibodies, demonstrating that Tat binds endogenous Set7/9-KMT7 expressed in T cells (Figure 5B). This interaction was independent from the methylation status of K51, since the K51A mutant of Tat coimmunoprecipitated with endogenous Set7/9-KMT7 as efficiently as wild-type Tat in transfected 293 cells (Figure 5C). In vitro binding reactions of recombinant Set7/9-KMT7 with synthetic Tat protein demonstrated that Set7/9 and Tat may bind directly (Figure S3).
K51 lies in the RNA-recognition motif of Tat and forms critical interactions with TAR RNA (Anand et al., 2008). To examine how methylation by Set7/9-KMT7 influences the interaction between Tat and TAR RNA, we performed gel mobility shift assays with synthetic Tat, recombinant Set7/9-KMT7 and radiolabeled TAR RNA as a probe. Surprisingly, addition of Set7/9-KMT7 alone efficiently shifted the TAR RNA probe, pointing to new RNA-binding properties of Set7/9-KMT7 (Figure 6A).
Binding of Set7/9-KMT7 to TAR RNA was successfully competed with unlabeled TAR RNA and, like Tat, required the bulge region of the TAR stem (Figure 6A and Figure S4A). We excluded the possibility that the shift was caused by a contamination of the Set7/9-KMT7 preparation with synthetic Tat and repeated the experiments with several different Set7/9-KMT7 preparations (data not shown). We also successfully supershifted the complex with α-Set7/9-KMT7 antibodies, confirming that it was, indeed, a complex formed by Set7/9-KMT7 and TAR (Figure 6B). No shift was observed when a truncated Set7/9-KMT7 (aa 110–366) was incubated with TAR RNA, indicating that the TAR-binding domain is located in the N terminus of the enzyme, outside the catalytic Set domain (Figure 6C). Preparation procedures for full-length and N-terminally truncated Set7/9-KMT7 proteins were identical, and integrity of both proteins was verified by coomassie staining (data not shown).
Interestingly, the main complex formed by Set7/9-KMT7 and TAR RNA showed the same mobility as the Tat-TAR RNA complex despite different molecular weights of Tat (72 amino acids) and Set7/9-KMT7 (366 amino acids) (compare Figure 6D, lanes 4 and 5). Combination of Tat and Set7/9-KMT7 did not result in the formation of a slower migrating complex although, in general, formation of an additional, slightly slower migrating complex was observed in the presence of Set7/9-KMT7 independently from Tat (Figures 6A–D).
We observed a synergistic increase in TAR binding when a suboptimal Tat concentration was combined with Set7/9-KMT7 (Figure 6D, compare lanes 7 and 3). This increase in binding was obvious when a loop-mutant TAR RNA probe was examined. Intact loop sequences of TAR RNA are required for the formation of the trimolecular complex of Tat-TAR-P-TEFb (Wei et al., 1998). Tat alone efficiently interacted with loop-mutant TAR RNA as expected (Figure 6D, lanes 17–20). In contrast, Set7/9-KMT7 did not bind the loop-mutant probe, demonstrating that, unlike Tat, it requires intact bulge and loop sequences of TAR for efficient binding (Figure 6D, lane 21). However, Tat binding to TAR RNA was increased in the presence of Set7/9-KMT7, supporting a model where Set7/9-KMT7 enhances the RNA-binding properties of Tat independently from its physical interaction with TAR RNA (Figure 6D, lanes 22-24).
Because in vivo Tat binding to TAR occurs in the presence of P-TEFb, we also performed gel shift experiments with recombinant P-TEFb (Cyclin T1/CDK9). The Cyclin T1 subunit interacts with the transactivating domain in Tat and the loop region of TAR and triggers formation of a large TAR ribonucleoprotein complex (Figure 6D, lanes 10–12). Formation of this large complex was slightly enhanced when Set7/9-KMT7 was added to the gel shift reactions (Figure 6D, lanes 14–16) and depended on intact loop sequences in TAR, irrespective of whether Set7/9-KMT7 was added to the reaction or not (Figure 6D, lanes 26–28 and lanes 30–32).
Addition of P-TEFb did not affect the complex formed by TAR RNA and Set7/9-KMT7 in the absence of Tat, indicating that any engagement of Set7/9-KMT7 in the Tat-TAR-P-TEFb complex depended on Tat (Figure 6D, lane 13). Because of the large size of the Tat-TAR-P-TEFb complex, it was difficult to assess whether migration of the complex changed in the presence of Set7/9-KMT7 or α-Set7/9-KMT7 antibodies (Figure 6D, and data not shown). We therefore performed coimmunoprecipitation experiments to examine whether Set7/9-KMT7 physically engages in complex formation with Tat and Cyclin T1. Endogenous Set7/9-KMT7 was immunoprecipitated in 293 cells expressing HA-tagged Cyclin T1. Cyclin T1 only coimmunoprecipitated with Set7/9-KMT7 when Tat was coexpressed, demonstrating that Set7/9-KMT7 cooperatively binds Tat and Cyclin T1 (Figure 6E). Accordingly, Set7/9 expressed in murine cells lacking a Cyclin T1 protein that can interact with Tat only synergized with Tat when human Cyclin T1 was coexpressed (Figure S4B). Collectively, these data support a model where Set7/9 promotes the recruitment of Tat and P-TEFb to TAR RNA.
Our data identify Set7/9-KMT7 as a new co-activator of Tat transactivation. Set7/9-KMT7 methylates K51 in Tat and activates Tat transactivation in a K51-dependent manner. In infected T cells, Set7/9-KMT7 is required for full activation of HIV gene expression and associates with the latent and activated HIV LTR in vivo. The transcriptional activity of methylation-deficient Tat is impaired and, unlike wild-type Tat, not affected by siRNA-mediated knockdown of Set7/9-KMT7. Importantly, knockdown of Set7/9-KMT7 also inhibits HIV proviral gene expression in primary CD4+ T cells. Mechanistically, we link the positive role of Set7/9-KMT7 in HIV transcription to an increase in Tat-TAR interaction and demonstrate that Set7/9-KMT7 itself can bind TAR and the Tat-P-TEFb complex.
Previously, a consensus recognition sequence was proposed for Set7/9-KMT7 substrates based on structural data derived from substrates that were known at that time (Couture et al., 2006). This consensus sequence predicts a lysine or arginine at position − 2 and a serine or threonine at position −1 (where 0 is the methylated lysine) and applies to many Set7/9-KMT7 substrates, including histone H3, p53, TAF10, estrogen receptor α and DNMT1. However, recently identified target lysines in the p65 subunit of NF-κB or the PCAF-KAT2B acetyltransferase do not align with this consensus sequence (Masatsugu and Yamamoto, 2009; Wang et al., 2009). In Tat, position −2 is occupied by R49 and is consistent with the proposed consensus sequence. However, position −1 in Tat is occupied by K50, which is similar to K314 in p65 when K315 is the target site (Yang et al., 2009). Tat residues R49, K50 and K51 are all strictly conserved among HIV isolates, underlining the importance of recognition of K51 by Set7/9-KMT7 in the HIV life cycle.
The finding that Set7/9-KMT7 has intrinsic TAR RNA-binding properties was unexpected. However, other lysine methyltransferases also bind single-stranded DNA and RNA (Krajewski et al., 2005). The fission yeast Set1 (KMT2) protein possesses an N-terminal canonical RNA-recognition motif (RRM) that is essential for its catalytic activity in vivo (Noma and Grewal, 2002). Although this RRM is not conserved in Set7/9-KMT7, we show that TAR RNA binding is also mediated via the N terminus of Set7/9. Based on these observations, we propose a model where low levels of Set7/9-KMT7 are continuously recruited to initiated HIV transcripts via direct interactions with loop and bulge sequences in TAR (Figure 7A).
After Tat expression, in vivo recruitment of Set7/9-KMT7 to the HIV LTR is enhanced. We show that Set7/9-KMT7 physically interacts with Tat and also forms a multimolecular complex with Tat and P-TEFb. We therefore speculate that during the elongation phase of HIV transcription, Set7/9-KMT7 becomes enriched at the HIV LTR via these interactions and methylates K51 in Tat (Figure 7B). Although it remains unclear from the gel shift experiments whether Tat methylation directly influences the formation of the Tat-TAR-P-TEFb complex, the finding that Set7/9-KMT7 enhances Tat-TAR binding independently from its interaction with TAR RNA supports such a model. In addition, we verified by western blot analysis that Tat in the gel shift reactions becomes efficiently methylated by Set7/9-KMT7 whether exogenous SAM is added or not to the reactions (data not shown).
Important evidence that Set7/9-KMT7 activates Tat function through K51 methylation comes from the experiments with the Tat K51A and K51R mutants. In our studies, the transcriptional activity of these mutants was ~3-fold decreased and were overall unresponsive to Set7/9-KMT7 knockdown, supporting the model that methylation by Set7/9-KMT7 is an important step in Tat transactivation. A similar decrease in transactivation with the Tat K51A mutant was previously reported by Kiernan and colleagues (Kiernan et al., 1999). Moreover, Bennasser and colleagues demonstrated that mutation of K51 in the context of an infectious clone of HIV-1 suppressed viral replication (Bennasser et al., 2005). However, the latter study did not find that the K51A mutation affected Tat transcriptional activity, but linked it instead to a new function of Tat as a suppressor of the cellular RNAi machinery (Bennasser et al., 2005). While our study underlines the importance of K51 for Tat transactivation, a potential involvement of K51 methylation in the new RNAi suppressor function of Tat is interesting and requires further studies.
Additional support for a critical role of K51 in Tat transactivation comes from a recent crystallization study of the equine infectious anemia virus (EIAV) Tat protein (Anand et al., 2008). This crystal structure predicts that K51 in HIV-1 Tat plays a central role in the interaction between Tat and TAR RNA. Our data indicate that addition of a methyl group to K51 could strengthen these interactions. In contrast, it was reported that K51 methylation by the H3K9 lysine methyltransferase SETDB1-KMT1E inhibited the formation of the TAR ribonucleoprotein complex (Van Duyne et al., 2008). Since SETDB1-KMT1E can function as a di- and trimethyltransferase, the possibility exists that addition of a single methyl group to K51 enhances Tat binding to TAR and P-TEFb, while addition of two or three methyl groups inhibits complex formation (Wang et al., 2003).
It remains unclear whether recruitment of Set7/9-KMT7 also serves to methylate histone H3K4 at the HIV LTR, since the enzyme cannot efficiently methylate histones when assembled into nucleosomes (Nishioka et al., 2002; Wang et al., 2001). HIV transcriptional elongation correlates with an increase in H3K4 trimethylation at the integrated HIV template (Zhou et al., 2004). Levels of H3K4 di- or trimethylation have been correlated with the recruitment of Set7/9-KMT7 to cellular promoters, including the Ins1/2, Slca2, MCP-1 and TNF-α genes, in support of the concept that Set7/9-KMT7 could act as a histone methyltransferase in vivo (Deering et al., 2008; Li et al., 2008). Notably, Set7/9-KMT7 itself is not a di- or trimethyltransferase; structural analysis of Set7/9-KMT7 revealed that the enzyme functions mainly as a monomethyltransferase (Xiao et al., 2003). Accordingly, we show that Tat K51 is monomethylated by Set7/9-KMT7.
The finding that Tat is monomethylated at K51 adds a new post-translational modification to a region in Tat that is a substrate of multiple histone-modifying enzymes (Hetzer et al., 2005). Tat residue K50 is the target of the p300-KAT3B and GCN5-KAT2A acetyltransferase activities, while R52 and R53 are methylated by PRMT6 (Col et al., 2001; Kiernan et al., 1999; Ott et al., 1999; Xie et al., 2007). The dynamics and hierarchy of these modifications are unknown. However, because of their close proximity within the Tat ARM, the occurrence and function of one modification is likely coupled to the modification status of a neighboring residue as reported for histones (Jenuwein and Allis, 2001). Since Set7/9-KMT7 is continuously present at the HIV promoter even before Tat is produced, we predict that Tat monomethylation at K51 occurs as an early event in the Tat transactivation cycle and serves to enhance the formation of the Tat/TAR/P-TEFb complex. Future studies will focus on dynamics of Tat methylation at the HIV LTR and its interplay with other Tat modifications.
Details for cells, reagents and plasmids are described in the Supplemental Experimental Procedures.
In vitro methylation reactions with synthetic Tat peptides, Set7/9-KMT7 enzyme and 3H-SAM were performed as described (Nishioka et al., 2002). Tat peptide reactions performed with nonradioactive SAM were analyzed by MALDI TOF mass spectrometry. Details for in vitro methylation assays are described in the Supplemental Experimental Procedures.
The α-meARM antibodies were generated after immunization of rabbits with K51-monomethylated ARM peptides. The same peptides were used for affinity purification of the antiserum. Details for the preparation and use of polyclonal α-meARM antibodies are described in the Supplemental Experimental Procedures.
HIV LTR luciferase reporter and protein expression vectors were transfected into HeLa cells with lipofectamine (Invitrogen). Cells were harvested 24 h later and processed for luciferase assays (Promega). For RNAi experiments, HeLa cells were transfected with pooled Set7/9-KMT7 and control siRNAs (both Dharmacon) using oligofectamine (Invitrogen) and were re-transfected after 48 h with the HIV LTR luciferase construct, Tat expressing vectors and corresponding amounts of the empty vector. Cells were harvested 24 h later and processed for luciferase assays or western blotting. SiRNA-nucleofection of Jurkat A2 cells was performed as described (Mahmoudi et al., 2006). 72 h after nucleofection, cells were treated with PMA (Sigma; 2 ng/ml) or DMSO for 12 h. GFP expression was analyzed on a Calibur FACScan (Beckton Dickinson). P values (paired t test) were used for statistical analysis. Set7/9-KMT7 knockdown efficiency in A2 cells was quantified using the ImageJ software available at htpp:rsb.info.nih.gov/ij.
Activated CD4+ T cells were transduced with lentiviral vectors containing shRNAs against Set7/9-KMT7 (shSet7/9-1, shSet7/9-2) or the non-targeting shRNA control. These lentiviral vectors also express the mCherry protein under the control of the EF-1α promoter. Four days after infection, cells were reinfected with viral particles produced from the molecular clones HIV-R7/E−/GFP and pHR'-EF-1α/GFP. The percentage of GFP+ mCherry+ cells was monitored by flow cytometry three days after the second infection. Details for the viral production and infection are described in the Supplemental Experimental Procedures.
Chromatin from Jurkat A2 cells treated with PMA or DMSO was immunoprecipitated with α-Set7/9 antibodies. The immunoprecipitated material was quantified by real-time PCR with primers specific for the HIV LTR, GFP and β-actin. Details for the Chromatin immunoprecipitation experiments are described in Supplemental Experimental Procedures.
Jurkat A2 cells were stimulated with 10 ng/ml TNF-α (Biosource) for 18 h or were left unstimulated. Cells were lysed in IP buffer, and immunoprecipitated with α-Set7/9-KMT7 antibodies for 2 h at 4°C. Beads were extensively washed and analyzed by western blotting with α-Set7/9-KMT7 or monoclonal α-FLAG antibodies. For coimmunoprecipitations in 293 cells, cells were transfected with vectors expressing Tat/FLAG, Tat K51A/FLAG, or HA-Cyclin T1 as indicated using lipofectamine reagent, and immunoprecipitations were performed as described above.
TAR RNAs (WT, Δbulge and Δloop) were synthesized in in vitro transcription reactions with the Riboprobe system (Promega) as previously described (Kaehlcke et al., 2003). Transcripts were treated with 2 U of DNase I (Promega), extracted with a phenol:chloroform mixture and purified over an illustra MicroSpin G50 column (GE Healthcare). Gel-mobility reactions (16 μl final volume) were carried out in binding buffer (50 mM Tris, pH 7.4, 0.5 mM EGTA, 150 mM NaCl, 2% glycerol, 0.2% Tween 20, 0.5 mM DTT, 90 mM ZnSO4, 0.005% BSA and 100 μM ATP) and contained 2×104 cpm TAR probes/reaction and the indicated concentrations of Tat, recombinant Set7/9-KMT7 or Set7/9-KMT7 (aa 110-366), and recombinant Cyclin T1/CDK9 (Millipore). Supershift experiments were conducted in the presence of 7 μg of α-Set7/9-KMT7 antibodies or corresponding amounts of preimmune rabbit serum. Reactions were incubated for 30 min at 30°C and separated on a pre-run 4% Tris-glycine gel.
We thank Danny Reinberg, Katherine A. Jones, Matthew Spindler, Silke Wissing, Marielle Cavrois and members of the Verdin, Ott and Greene labs for sharing their reagents and expertise, Sarah Elmes from the UCSF Flow Core for cell sorting, and Robert Houtz, Lynnette A. Dirk, Shiv Grewal and Warner C. Greene for helpful discussions. We thank John Carroll, Alisha Wilson, Teresa Roberts and Chris Goodfellow for graphics, Gary Howard for editorial assistance, and Veronica Fonseca for administrative assistance. This work was supported by funds from the Gladstone Institutes, the University of California San Francisco–Gladstone Institute of Virology & Immunology Center for AIDS Research, the NIH and by grant number 106736-40-RFRL from amfAR, the American Foundation for AIDS Research.
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.