Abstract
Free full text
Functional and physical interaction between the histone methyl transferase Suv39H1 and histone deacetylases
The histone methyl transferase Suv39H1 is involved in silencing by pericentric heterochromatin. It specifically methylates K9 of histone H3, thereby creating a high affinity binding site for HP1 proteins. We and others have shown recently that it is also involved in transcriptional repression by the retinoblastoma protein Rb. Strikingly, both HP1 localisation and repression by Rb also require, at least in part, histone deacetylases. We found here that repression of a heterologous promoter by Suv39H1 is dependent on histone deacetylase activity. However, the enzymatic activity of Suv39H1 is not required, since the N-terminal part is by itself a transcriptional repression domain. Coimmunoprecipitation experiments indicated that Suv39H1 can physically interact with HDAC1, -2 and -3, therefore suggesting that transcriptional repression by Suv39H1 could be the consequence of histone deacetylases recruitment. Consistent with this interpretation, the N-terminal transcriptional repression domain of Suv39H1 bound the so-called ‘core histone deacetylase complex’, composed of HDAC1, HDAC2 and the Rb-associated proteins RbAp48 and RbAp46. Taken together, our results suggest that a complex containing both the Suv39H1 histone methyl transferase and histone deacetylases could be involved in heterochromatin silencing or transcriptional repression by Rb.
Eukaryotic nuclei contain condensed chromatin, the heterochromatin, which is mostly transcriptionally inactive. This silencing is believed to be mediated by high order chromatin structure. Chromatin structure and function are controlled, at least in part, through the posttranslational modifications of nucleosomal histones. The most studied histone modification is acetylation of lysines. Acetylation largely correlates with transcriptional activation, whereas histone deacetylation is associated with transcriptional repression (1). Consequently, from yeast to human, histone deacetylases are important for transcriptional silencing. The NAD-dependent Sir2 enzyme is involved in silencing of telomeric sequences and mating type locus in Saccharomyces cerevisiae (2,3). In Schizosaccharomyces pombe, Clr3, which deacetylates specifically K14 of histone H3, is also important for silencing of mating type locus, telomeric and centromeric chromatin (4,5). Although genetic studies are more difficult in mammals, it is known that histones within silenced chromatin are largely hypoacetylated, as is the case for pericentric heterochromatin (6).
Strahl and Allis proposed that the various histone modifications form a code (7). According to this ‘histone code’ hypothesis, a precise combination of modifications leads to a particular functional consequence for chromatin function, through the recruitment of a specific factor. An obvious consequence of this hypothesis is that the various histone modifications work in an interdependent manner to regulate chromatin function. Consequently, our interest in the histone modifications other than acetylation has recently been renewed. Many studies have now shown that histone methylation on lysines, a modification that has been known for years, is also important for transcriptional silencing (8,9). Indeed, the first described histone methyl transferase (10), the Suv39H1 enzyme, is the mammalian homologue of S.pombe Clr4 and Drosophila Su(Var)3.9, two proteins that are involved in silencing by pericentric heterochromatin (11–13). This enzyme methylates K9 from histone H3 (10), thereby creating a high affinity binding site for proteins of the HP1 family (14,15). Consistent with this model, the activity of Suv39H1 or its orthologues is required for the proper localisation of HP1 (or its orthologues) in heterochromatin (4,14,15). Interestingly, localisation of HP1 is also dependent on histone deacetylases. This was shown by genetic studies in S.pombe (4), and also in mammalian cells, by the use of trichostatin A (TSA), a potent inhibitor of histone deacetylases (16). Thus, Suv39H1 and histone deacetylases function together to mediate proper HP1 localisation and silencing through heterochromatin.
Strikingly, both enzymes are also involved in transcriptional repression of specific promoters at eukaryotic loci. The artificial recruitment of histone deacetylases or Suv39H1 to a heterologous promoter results in transcriptional repression (17–19). Furthermore, it has been demonstrated that histone deacetylases are directly recruited on some natural promoters to mediate reversible transcriptional repression. One example is the regulation of E2F-responsive genes. These genes encode proteins that are required for cell progression into S phase (20). E2F-regulated promoters are activated at the end of G1 and at the G1/S transition by the E2F transcription factor, which is composed of heterodimers between a E2F protein and a DP protein. These heterodimers bind specifically to E2F sites. In G0 and at the beginning of G1, E2F-containing promoters are repressed by the members of the retinoblastoma protein (Rb) family, which are recruited by a physical interaction with the E2F protein (21). We and others have proposed that transcriptional repression by Rb and its cousins involves the recruitment of histone deacetylases (HDAC1, HDAC2 or HDAC3) to E2F-containing promoters (22–25). Subsequently, it has been demonstrated that histones present on E2F-regulated promoters evolve during G1 from a hypoacetylated to a hyperacetylated state (26). Finally HDAC1 is physically present on these promoters in G0 but not at the G1/S transition (27). Recently, we and others have shown that Suv39H1 could also be involved in transcriptional repression of E2F-responsive promoters (18,28). Suv39H1 interacts physically with Rb and functions as a corepressor of E2F activity. Although it has not been directly shown, Suv39H1 is likely to be recruited to the E2F-regulated promoter, since the histone H3 present on the cyclin E promoter is methylated at K9 (28). Thus, the cell cycle-dependent regulation of E2F-containing promoters could involve both Suv39H1 and histone deacetylases.
Here, we show that histone deacetylases are required for repression of a heterologous promoter by Suv39H1. Furthermore, we found that Suv39H1 and histone deacetylases physically interact at endogenous levels. Suv39H1 interacts with histone deacetylases through its N-terminal domain, which functions as a transcriptional repression domain. Taken together, these results suggest that the functional cooperation between Suv39H1 and histone deacetylases could be due, at least in part, to their physical interaction.
Cell culture and transfection
U2OS cells were grown in Dulbecco’s modified Eagle’s medium supplemented with antibiotics (1%) and fetal calf serum (10%). For luciferase reporter assays, 100 000 cells were seeded in 6-well plates. Cells were transfected by calcium/phosphate coprecipitation using standard procedures. Each transfection mix included 100 ng of pCMV-β-galactosidase (pCMV-βGAL) to monitor transfection efficiency. Either 24 or 32 h after transfection, cells were lysed using 300 µl of reporter lysis buffer (Promega). TSA (Sigma) was added where indicated (500 ng/ml) 24 h after transfection and cells were harvested 8 h later. Ten microlitres of lysates were then assayed for luciferase activity using the luciferase assay reagent (Promega) in a luminometer (Berthold). β-galactosidase activity was measured using a Galacto-light kit (Tropix), according to the manufacturer’s instructions.
Plasmids
To express Gal4 fusion proteins in mammalian cells, the corresponding deletion mutant of Suv39H1 was introduced into the pCMV glutathione S-transferase (GST) empty vector (a kind gift from Dr T. Kouzarides), which expressed the GAL4 DNA binding domain (amino acids 1–147) from a CMV promoter. Gal4-luciferase reporter vector and pCMV-βGAL standardisation vector have been described previously (18). The plasmid expressing GST–Suv39H1 82–412 fusion protein in Escherichia coli was a kind gift from Dr T. Jenuwein (10). The plasmid expressing GST–Suv39H1 4–110 in E.coli was constructed by inserting the corresponding part of Suv39H1 into pGEX2TKp vector. Details of constructions are available upon request.
Immunoprecipitations and GST pull-downs
For immunoprecipitation experiments, HeLa nuclear extracts (Computer Cell Culture Center, Belgium) were diluted to 1 ml using washing buffer (50 mM Tris pH 8.0, 150 mM NaCl, 5 mM MgCl2, 0.5% NP-40) supplemented with CaCl2 (2.5 mM), DNase I and protease inhibitors (Complete, Roche Diagnostics). Nuclear extracts were first precleared using 20 µl of protein A/protein G beads at 4°C on a rotating wheel. The immunoprecipitating antibody (1 µg) was then added. After 1 h at 4°C, protein A/protein G beads (10 µl) were added and incubation was continued for 1 h at 4°C on a rotating wheel. Beads were then extensively washed and were either assayed for histone deacetylase or for histone methyl transferase activity, or subjected to western blot analysis. For GST pull-downs, beads harbouring the various bacterially expressed GST-fusion proteins were prepared by incubating E.coli extracts with glutathione beads as described (29). The concentration of fusion proteins was estimated by SDS–PAGE and coomassie staining, and we systematically used a higher amount of control protein (GST) in pull-down experiments. Beads (10 µl) were then incubated with HeLa nuclear extracts (precleared on glutathione beads) in 1 ml of washing buffer (supplemented as above) for 2 h at 4°C on a rotating wheel. Beads were then washed extensively using washing buffer and analysed as described above.
Western blots and antibodies
Transfer of SDS–PAGE gels to nitrocellulose membranes and western blots were performed using standard procedures. We described previously the anti-Suv39H1 antibody (18). The anti-HDACs antibody was the anti-HDAC3 antibody from Transduction Laboratories. This antibody cross-reacts with HDAC1 and HDAC2, and was used in western blots or immunoprecipitation (30). The anti-RbAp48 antibody was the anti-RB-BP antibody from Transduction Laboratories. We used this antibody in western blots. It recognises both RbAp48 and RbAp46 (29).
Histone deacetylase and methyl transferase assays
Beads from GST pull-downs or immunoprecipitations were washed twice in Tris-buffered saline (TBS) (50 mM Tris pH 8.0, 150 mM NaCl), then resuspended in 35 µl of TBS. TSA (100 ng/ml) was added or not, and beads were incubated for 10 min at 4°C. 3H-labelled acetylated nucleosomal histones [30 000 c.p.m., purified from Jurkat cells as previously described (29)] were then added, and deacetylation reaction was performed at 37°C for 2 h on a rotating wheel. Reactions were stopped by the addition of 35 µl of acid mix (0.24 M acetic acid, 1.4 M HCl). Labelled acetate was then extracted using 700 µl of ethyl acetate, and counted on a scintillation counter. Histone methyl transferase assays were performed as described (18), using a histone H3-derived peptide (sequence: ARTKQTARKSTGGKAPRKQLATKA).
Transcriptional repression by Suv39H1 does not require its catalytic domain
Suv39H1 is known to function as a transcriptional repressor (18,19). To test whether this repression could be mediated through its histone methyl transferase SET domain, we constructed some deletion mutants of Suv39H1 (Fig. (Fig.1A).1A). We then transiently transfected U2OS cells with a reporter construct in which the luciferase encoding gene is under the control of Gal4 sites, together with vectors allowing the expression of these various deletion mutants fused to the Gal4 DNA binding domain. This experiment leads to the targeting of various parts of Suv39H1 to an artificial promoter. As already shown (18,19), targeting Suv39H1 full length to a heterologous promoter resulted in transcriptional repression (Fig. (Fig.1B).1B). We found that a deletion mutant containing only the first 332 amino acids from Suv39H1 and lacking key residues for the enzymatic activity (10) repressed transcription as efficiently as full length. Deletion mutant analysis of this N-terminal part of Suv39H1 indicated that it contains two independent transcriptional repression domains, one located in the N-terminus (from amino acid 4 to 195), and the other located between amino acids 195 and 332. By themselves, these two transcriptional repression domains were fully efficient, since both mutants repressed transcription as strongly as Suv39H1 full length. Note that it is impossible to test using this artificial targeting assay whether the enzymatic activity of Suv39H1 is sufficient for transcriptional repression, because the C-terminal repression domain (between amino acids 195 and 332) is fully included within the catalytic domain (10). Nevertheless, these data indicate that Suv39H1 possesses transcriptional repression properties independent of histone methyl transferase activity.
Transcriptional repression by Suv39H1 does not require its enzymatic activity. (A) Schematic representation of Suv39H1. We have indicated the N-terminal chromodomain (hatched box), the SET domain (black box) and the adjacent cysteine-rich region (grey box). Also indicated is the catalytic domain, which includes the cysteine-rich region, the SET domain and the C-terminus of the protein (10). (B) U2OS cells were transiently transfected using 2 µg of GAL4-luciferase reporter vector, 100 ng of pCMV βGAL and 200 ng of an expression vector for the indicated GAL4 fusion protein (expressed from pCMVGT vector). The experiment was performed in duplicate. Luciferase and β-galactosidase activities were measured 24 h after transfection. Results were standardised for transfection efficiency by dividing luciferase activity by β-galactosidase activity. (C) U2OS cells were transiently transfected using 2 µg of GAL4-luciferase reporter vector, 100 ng of pCMV βGAL and 200 ng or 2 µg (as indicated by the height of the open triangle) of an expression vector for the indicated Gal4 fusion protein (expressed from pCMVGT vector). The amount of CMV promoter was kept constant using empty vector. The control sample (with 2 µg of pCMVGT) (Gal4) was performed in duplicate. Luciferase and β-galactosidase activities were measured 24 h after transfection. Results were standardised for transfection efficiency by dividing luciferase activity by β-galactosidase activity. In the lower panel, the expression of Gal4-Suv39H1 fusion proteins was monitored by western blot using an anti-myc antibody (9E10, Roche Diagnostics).
The N-terminus of Suv39H1 contains a chromodomain, which is a domain found in many chromatin-related proteins, a chromodomain. To investigate whether this domain was responsible for the N-terminal transcriptional repression properties of Suv39H1, we constructed a deletion mutant expressing Suv39H1 4–110 fused to the Gal4 DNA binding domain. Transfection of this vector led to an efficient transcriptional repression of the cotransfected Gal4-luciferase reporter vector (Fig. (Fig.1C).1C). Strikingly, it repressed transcription as efficiently as the Suv39H1 full-length protein. Since both fusion proteins were similarly expressed (Fig. (Fig.1C,1C, bottom), this result strongly suggests that the first 110 amino acids of Suv39H1 are fully competent for transcriptional repression.
Transcriptional repression through Suv39H1 requires histone deacetylases
We hypothesised that this transcriptional repression could be mediated through histone deacetylases, since histone deacetylases and histone methyl transferases are functionally linked in silencing through heterochromatin formation and transcriptional repression by Rb. To test this hypothesis, we used TSA, a known specific inhibitor of histone deacetylases. TSA treatment consistently induced a slight decrease of luciferase activity in the absence of Gal4-Suv39H1 fusion protein (Fig. (Fig.2B),2B), which is likely to reflect an effect of TSA on the basal promoter driving luciferase expression. However, transcriptional repression by Suv39H1 is largely abolished in the presence of TSA (Fig. (Fig.2B).2B). Indeed, fold repression by the Gal4-Suv39H1 fusion protein went from ~10-fold in the absence of TSA down to 1.5-fold in the presence of TSA. This decrease is not due to a defect in Gal4-Suv39H1 fusion protein expression, since it is expressed even more in the presence of TSA (Fig. (Fig.2B).2B). Since transcriptional repression by Suv39H1 is not dependent upon the amount of protein produced (e.g. Fig. Fig.1C),1C), this result strongly suggests that it requires histone deacetylases.
Transcriptional repression by Suv39H1 requires a histone deacetylase. U2OS cells were transiently transfected using 2 µg of GAL4-luciferase reporter vector, 100 ng of pCMV βGAL and 200 ng of the GAL4-Suv39H1 expression vector where indicated. The amount of promoters in the transfection was kept constant using the pCMVGT empty vector. TSA was added, or not, as indicated, 24 h after transfection. Luciferase and β-galactosidase activities were measured 8 h later. Note that we did not divide luciferase by β-galactosidase, because TSA treatment strongly activated the CMV promoter (data not shown, see also the expression of CMV-driven Gal4-Suv39H1 in the lower panel). The expression of Gal4-Suv39H1, which contains three myc tags, was assayed by western blot using the anti-myc antibody (9E10, Roche Diagnostics) (lower panel).
Physical interaction between Suv39H1 and histone deacetylases
Since transcriptional repression by Suv39H1 is dependent upon histone deacetylase activity (Fig. (Fig.2),2), we tested the possibility of a physical interaction between Suv39H1 and histone deacetylases. We immunoprecipitated endogenous Suv39H1 from HeLa nuclear extracts and we measured histone deacetylase activity in the immunoprecipitates. We found (Fig. (Fig.3A)3A) that immunoprecipitation of Suv39H1 led to the coimmunoprecipitation of histone deacetylase activity, which was inhibited by TSA. This coimmunoprecipitation was specific, since it was not seen when the control preimmune serum (PI) was used. Thus, Suv39H1 physically interacts with a histone deacetylase.
Suv39H1 and histone deacetylases physically interact. (A) HeLa nuclear extracts (400 µl) were immunoprecipitated using the indicated serum (Suv: anti-Suv39H1; PI: preimmune serum). Immunoprecipitates were then assayed for histone deacetylase activity in duplicate. (B) Hela nuclear extracts (400 µl) were immunoprecipitated using the indicated serum. In lane 4, the anti-Suv39H1 serum was preincubated with the peptides (1 µg each) against which the antibody was raised. In lane 1 (Inp), 5 µl of HeLa nuclear extracts were directly loaded. (C) HeLa nuclear extracts (50 µl) were immunoprecipitated using the indicated antibody (HDAC: anti-HDACs; Irr: the anti-HA 12CA5 antibody, Roche Diagnostics). Immunoprecipitates were then assayed for histone methyl transferase activity (HMT activity) using a peptide derived from the histone H3 N-terminal tail as a susbstrate.
Transcriptional repression by Rb is believed to be dependent upon the histone deacetylases HDAC1, HDAC2 or HDAC3 (31). We recently showed that it also involves the Suv39H1 histone methyl transferase (18,28). This led us to test whether the histone deacetylase activity coimmunoprecipitated with Suv39H1 could also be due to these histone deacetylases. By western blot experiments, we found that HDAC1–3 histone deacetylases were indeed detected in Suv39H1 immunoprecipitates (Fig. (Fig.3B,3B, lane 2). Again, this coimmunoprecipitation was specific, since it was not seen with the preimmune serum (lane 3), and was inhibited by the peptides against which the antibody was raised (lane 4). Note that a faint band comigating with HDAC3 could be detected in control immunoprecipitates (lane 3). This faint band is most likely due to the immunoglobulin heavy chain of the immunoprecipitating antibody. Because of this band, the amount of HDAC3 coimmunoprecipitated with Suv39H1 (lane 2) is likely to be overestimated.
We then intended to perform the converse immunoprecipitation; however, our anti-Suv39H1 antibody does not work well in western blots and we were unable to directly test the presence of Suv39H1 in HDACs immunoprecipitates. We thus decided to detect Suv39H1 by its enzymatic activity. We found that, as expected, histone methyl transferase activity is coimmunoprecipitated by the anti-HDACs antibody, but not by the control antibody (Irr) (Fig. (Fig.3C).3C). Since the anti-HDACs antibody is specific for HDAC1-3, this result confirms the results shown in Figure Figure3B.3B. Taken together, results shown in Figure Figure33 indicate that endogenous Suv39H1 and HDAC1-3 are physically associated in live cells.
The N-terminus of Suv39H1 physically interacts with histone deacetylases
Transcriptional repression by Suv39H1 is mediated, at least in part, by its N-terminus (Fig. (Fig.1).1). We thus intended to test whether the Suv39H1 N-terminus interacts with histone deacetylases. We expressed in E.coli and purified a fusion protein between GST and the N-terminal domain of Suv39H1 (Suv39H1 4–110). We used beads harbouring this fusion protein in GST pull-down experiments using HeLa nuclear extracts. Bound proteins were then assayed for histone deacetylase activity (Fig. (Fig.4A).4A). When beads harbouring the N-terminus of Suv39H1 fused to GST (GST-Suv39H1 4–110) were used, we could detect histone deacetylase activity bound to the beads. This interaction was specific, since no activity was found using control GST beads, although a higher amount of fusion proteins was present on the beads (data not shown). These results indicate that Suv39H1 interacts with histone deacetylases through its N-terminus.
The N-terminal transcriptional repression domain of Suv39H1 interacts with histone deacetylases. (A) HeLa nuclear extracts (100 µl) were subjected to a pull-down using beads harbouring ~1 µg of bacterially produced GST-Suv39H1 4–110 (Suv39H1 4–110) or control GST proteins. After extensive washing, beads were assayed for histone deacetylase activity in duplicate. (B) HeLa nuclear extracts (10 µl) were subjected to a pull-down using beads harbouring ~1 µg of bacterially produced GST-Suv39H1 4–110 (Suv39H1 4–110) or control GST proteins. After extensive washing, the presence of HDAC1, HDAC2 and HDAC3 on the beads was tested by western blot using the anti-HDACs antibody. The migration of HDAC1, HDAC2 and HDAC3 is indicated by the arrows. In lane 1, 0.5 µl of HeLa nuclear extracts were directly loaded.
To determine the identity of this histone deacetylase, we performed a western blot on the GST pull-downs. We found that both HDAC1 and HDAC2 were present on GST-Suv39H1 4–110 beads (Fig. (Fig.4B,4B, lane 2) but not on GST beads (lane 3). This result indicates that the N-terminus of Suv39H1 is able to bind the HDAC1 and HDAC2 histone deacetylases. Strikingly, neither GST-Suv39H1 4–110 (Fig. (Fig.4B),4B), nor a fusion protein harbouring most of Suv39H1 (GST-Suv39H1 82–412) (data not shown) could efficiently pull-down HDAC3 from HeLa nuclear extracts, although HDAC3 was well expressed in these nuclear extracts (Fig. (Fig.44B).
The N-terminus of Suv39H1 interacts with the core deacetylase complex
Careful analysis of the western blot shown in Figure Figure4B4B indicated that although HDAC2 was expressed slightly more than HDAC1 in HeLa nuclear extracts, an equivalent amount of both proteins were found on GST-Suv39H1 4–110 beads. Interestingly, HDAC1 and HDAC2 are found associated together (one molecule of HDAC1 for one of HDAC2) in the so-called ‘histone deacetylase core complex’ (32). This complex also contains one molecule of each of the Rb-associated proteins RbAp48 and RbAp46 (33,34). We thus tested whether Suv39H1 interacts with these two polypeptides by western blot using an antibody that recognises both RbAp48 and RbAp46. We found that RbAp46 and RbAp48 polypeptides were present in GST pull-downs performed with GST-Suv39H1 4–110 fusion protein (Fig. (Fig.5,5, lane 2), but not on control GST beads (Fig. (Fig.5,5, lane 3). In addition, both proteins appear to be present at approximately the same amount. Taken together, these data indicate that Suv39H1 interacts with the histone deacetylase core complex.
Suv39H1 can interact with the ‘histone deacetylase core complex’. The experiment shown in Figure 4B was reprobed using the anti-RbAp48 antibody (anti-RB-BP), which recognises both RbAp48 and the related RbAp46 protein. The migration of RbAp48 and RbAp46 is indicated by the arrows.
In this study, we found that Suv39H1 physically interacts with histone deacetylases in human cells. A recent study described a similar interaction between Suv39H1 and HDAC1 in Drosophila (35). Our results add substantial information to these data. Indeed, the fact that we detect this interaction in human cells indicates that it is conserved throughout evolution. In addition, we identified the chromatin-related RbAp48 and RbAp46 proteins as components of this complex. We also show that the interaction with histone deacetylases could be mediated by the N-terminus of Suv39H1, which contains the chromodomain.
According to our immunoprecipitation data, the histone deacetylases associated with Suv39H1 could be HDAC1, HDAC2 and HDAC3 (Fig. (Fig.3).3). The interaction of Suv39H1 with HDAC1 and HDAC2 was confirmed by our GST pull-down experiments (Fig. (Fig.4).4). However, we did not find any evidence for binding of HDAC3 to Suv39H1 by GST pull-down experiments, although we used overlapping parts of Suv39H1 (Fig. (Fig.44 and data not shown). One possibility could be that HDAC3 does not directly interact with Suv39H1, so that its binding to exogenous bacterially-expressed Suv39H1 would require a bridging factor, which would be present in limited amount. This bridging factor could be HP1β, which represses transcription through histone deacetylases (36) and which binds Suv39H1 (13). Alternatively, the interaction between HDAC3 and Suv39H1 could be strictly dependent on a posttranslational modification of Suv39H1.
None of these histone deacetylases is enzymatically active in a purified recombinant form, suggesting that they require cofactors. Indeed, cofactors for the HDAC3 histone deacetylases have recently been described (37). Interestingly, we found that, besides histone deacetylases, Suv39H1 can also associate with RbAp48 and RbAp46 proteins (Fig. (Fig.5).5). These proteins are able to bind histones (38) and are present within many multimolecular complexes related to chromatin (39–42). In particular, in association with HDAC1 and HDAC2, they form the so-called ‘histone deacetylase core complex’ (32), which contains one molecule of HDAC1, HDAC2, RbAp46 and RbAp48. This complex is itself a component of larger complexes functioning in transcriptional regulation, such as Sin3-containing complexes (42) or the NURD complex (40,43). It is tempting to speculate that Suv39H1 is actually associated with one of these larger complexes.
What could be the role of this complex? Obviously, it is likely to function in mechanisms related to transcriptional repression. Indeed, targeting HDACs or Suv39H1 to a heterologous promoter leads to transcriptional repression (17–19). Furthermore, we have now shown that transcriptional repression by Suv39H1 is dependent on histone deacetylase activity (Fig. (Fig.2).2). Thus, if Suv39H1 is recruited to a specific promoter, it is likely to act as a transcriptional repressor, at least in part through the recruitment of histone deacetylases. What promoters are targeted by Suv39H1? To date, only two classes of promoters have been proposed to involve Suv39H1: promoters regulated by transcription factors containing krüppel-associated box domains (44) and E2F-regulated promoters (18,28). Strikingly, in the case of E2F-regulated promoters, the histone methyl transferase activity of Suv39H1 is required for transcriptional repression (18,28). Although we cannot formally rule out the possibility that the HDAC1/HDAC2 histone deacetylase complex recruits endogenous Suv39H1 to the Gal4 fusion protein, our results (Fig. (Fig.1)1) indicate that the histone methyl transferase activity of Suv39H1 is not required for repression of a heterologous promoter. The same part of Suv39H1, Suv39H1 1–332, represses transcription when targeted to an artificial promoter (Fig. (Fig.1),1), but does not cooperate with Rb to repress transcription of E2F-responsive genes (18). Thus, there is a major difference between E2F-responsive promoters and the artificial promoter we used in this study. The most likely possibility is that, in the case of E2F-regulated promoters, histone deacetylases recruitment is not dependent on Suv39H1 recruitment. Consistent with this interpretation, HDAC1 can directly interact with Rb (23). Consequently, on E2F-regulated promoters, the limiting step could be the recruitment of Suv39H1 histone methyl transferase activity, whereas in our experiments (Fig. (Fig.1)1) the limiting step would be the recruitment of histone deacetylase activity.
One immediate consequence of this hypothesis is that both enzymes would work cooperatively to repress transcription. Since transcriptional repression of E2F-responsive genes by Suv39H1 is most likely mediated through HP1 binding to the methylated histone H3 (28), it could be envisioned that HP1 recruitment also requires histone deacetylases. One possibility could be that histone deacetylases are required for the prior deacetylation of K9 before its methylation by Suv39H1, as already proposed (4), since K9 of histone H3 cannot be methylated or acetylated at the same time (10). Alternatively, HP1 recruitment could depend on K9 methylation and deacetylation of adjacent lysines. Consistent with this hypothesis, binding of Swi 6, the S.pombe homologue of HP1, to silenced loci requires both methylation of K9 and deacetylation of K14 of histone H3 (4).
Our result also raises the question of the involvement of the complex between Suv39H1 and histone deacetylases in silencing through heterochromatin formation. Indeed, the Drosophila and S.pombe homologues of Suv39H1 are required for pericentric heterochromatin silencing (11,12). The action of these enzymes are thought to create high affinity binding sites for HP1 (4,14,15), resulting in the formation of heterochromatin. Suv39H1 itself is important for correct localization of mammalian proteins of the HP1 family in heterochromatin (14). Strikingly, histones in heterochromatin are largely underacetylated (6), and localisation of HP1 in heterochromatin is lost upon inhibition of histone deacetylases (16). This result thus suggests that Suv39H1 and histone deacetylases also cooperate for HP1 localisation and silencing through heterochromatin formation in mammalian cells. The physical interaction between both enzymes that we describe in this manuscript could be the molecular basis of this functional cooperation.
The authors wish to thank Drs T. Kouzarides and T. Jenuwein for materials and M. Briet for critical reading of the manuscript. This work was supported by a grant from the Ligue Nationale Contre le Cancer to D. Trouche, as a ‘Equipe Labellisée’. E.N. and L.V. are respectively recipients of a studentship and a fellowship from the Association de Recherche contre le Cancer (ARC).
Articles from Nucleic Acids Research are provided here courtesy of Oxford University Press
Full text links
Read article at publisher's site: https://doi.org/10.1093/nar/30.2.475
Read article for free, from open access legal sources, via Unpaywall:
https://academic.oup.com/nar/article-pdf/30/2/475/9901124/300475.pdf
Citations & impact
Impact metrics
Citations of article over time
Alternative metrics
Article citations
Werner syndrome RECQ helicase participates in and directs maintenance of the protein complexes of constitutive heterochromatin in proliferating human cells.
Aging (Albany NY), 16(20):12977-13011, 17 Oct 2024
Cited by: 0 articles | PMID: 39422615 | PMCID: PMC11552638
Interpreting the molecular mechanisms of RBBP4/7 and their roles in human diseases (Review).
Int J Mol Med, 53(5):48, 05 Apr 2024
Cited by: 1 article | PMID: 38577935 | PMCID: PMC10999228
Review
Free full text in Europe PMC
Suv39h1 Silencing Recovers Memory Decline in Scopolamine-Induced Amnesic Mouse Model.
Mol Neurobiol, 61(1):487-497, 26 Aug 2023
Cited by: 0 articles | PMID: 37626270
Hedgehog pathway orchestrates the interplay of histone modifications and tailors combination epigenetic therapies in breast cancer.
Acta Pharm Sin B, 13(6):2601-2612, 11 Mar 2023
Cited by: 3 articles | PMID: 37425067 | PMCID: PMC10326305
The HP1α protein is mandatory to repress the circadian clock and its output genes during the 12 h period of transcriptional repression.
Proc Natl Acad Sci U S A, 120(8):e2213075120, 15 Feb 2023
Cited by: 1 article | PMID: 36791098 | PMCID: PMC9974504
Go to all (107) article citations
Other citations
Data
Similar Articles
To arrive at the top five similar articles we use a word-weighted algorithm to compare words from the Title and Abstract of each citation.
Transcriptional repression by the retinoblastoma protein through the recruitment of a histone methyltransferase.
Mol Cell Biol, 21(19):6484-6494, 01 Oct 2001
Cited by: 141 articles | PMID: 11533237 | PMCID: PMC99795
The histone deacetylase HDAC3 targets RbAp48 to the retinoblastoma protein.
Nucleic Acids Res, 29(15):3131-3136, 01 Aug 2001
Cited by: 38 articles | PMID: 11470869 | PMCID: PMC55834
Methyl-CpG binding domain 1 (MBD1) interacts with the Suv39h1-HP1 heterochromatic complex for DNA methylation-based transcriptional repression.
J Biol Chem, 278(26):24132-24138, 23 Apr 2003
Cited by: 154 articles | PMID: 12711603
SU(VAR)3-9 is a conserved key function in heterochromatic gene silencing.
Genetica, 117(2-3):149-158, 01 Mar 2003
Cited by: 62 articles | PMID: 12723694
Review