SIRT1 negatively regulates HDAC1-dependent transcriptional ... - Nature

Oncogene (2008) 27, 3384–3392

& 2008 Nature Publishing Group All rights reserved 0950-9232/08 $30.00 www.nature.com/onc

ORIGINAL ARTICLE

SIRT1 negatively regulates HDAC1-dependent transcriptional repression by the RBP1 family of proteins O Binda1, C Nassif1 and PE Branton1,2,3 1 Department of Biochemistry, McGill University, Montreál, Que´bec, Canada; 2Department of Oncology, McGill University, Montreál, Que´bec, Canada and 3McGill Cancer Center, McGill University, Montreál, Que´bec, Canada

Both RBP1 and the highly related protein BCAA play a role in the induction of growth arrest and cellular senescence via mechanisms involving transcriptional repression. While investigating the transcriptional repression activities of RBP1, we observed a genetic link between RBP1 and SIR2. Further work uncovered an interaction between RBP1 family proteins and the mammalian homologue of SIR2, SIRT1. Interestingly, the HDAC-dependent transcriptional repression domain of RBP1 proteins, termed R2, is necessary and sufficient for the interaction with SIRT1. In vitro and in vivo binding studies indicated that the p33ING1b and p33ING2 subunits of the mSIN3A/HDAC1 complex are responsible for the recruitment of SIRT1 to the R2 domain. To investigate the biological relevance of this interaction, we used the sirtuin activator resveratrol and the sirtuin inhibitor sirtinol in transcriptional repression assays and demonstrated that SIRT1 activity negatively regulates R2-mediated transcriptional repression activity. We therefore propose a novel mechanism of class I HDAC regulation by a class III HDAC. Explicitly, SIRT1 is recruited by ING proteins and inhibits R2-associated mSIN3A/HDAC1 transcriptional repression activity. Oncogene (2008) 27, 3384–3392; doi:10.1038/sj.onc.1211014; published online 14 January 2008 Keywords: transcription; RBP1; BCAA; SIRT1; HDAC1; mSIN3A

Introduction Histone deacetylation has a pivotal role in the regulation of epigenetic events leading to transcriptional silencing. A number of processes linked to cell cycle control and tumourigenesis appear to be regulated by histone deacetylases, including cell cycle progression, cellular differentiation, senescence and apoptosis. Indeed, since a number of tumour suppressors have been reported to be epigenetically silenced in cancer, Correspondence: Professor PE Branton, Department of Biochemistry, McGill University, McIntyre Medical Building, Room 702, 3655 Promenade Sir William Osler, Montreál, Que´bec, Canada H3G 1Y6. E-mail: [email protected] Received 4 September 2007; revised 29 November 2007; accepted 3 December 2007; published online 14 January 2008

considerable effort is being focused on the targeting of histone deacetylase (HDAC) activity as part of cancer therapy (Baylin et al., 2001; Rountree et al., 2001; Minucci and Pelicci, 2006). Three classes of HDACs have been characterized thus far (Khochbin et al., 2001; Thiagalingam et al., 2003). The zinc-binding class I and II HDACs share a similar mechanism of action and the common property of being sensitive to the histone deacetylase inhibitor trichostatin A (Finnin et al., 1999). The class I histone deacetylases are related to the yeast Rpd3 protein while the class II are related to HDA1. HDAC1 is a class I deacetylase that is found in association with the mSIN3A complex (Hassig et al., 1997; Laherty et al., 1997). The class III HDACs share homology within their catalytic core domain with the yeast NAD þ -dependent histone deacetylase Silent Information Regulator 2 (SIR2) and generally referred to as sirtuins (reviewed in Blander and Guarente, 2004). In a number of model organisms including yeast (Kaeberlein et al., 1999), nematode (Tissenbaum and Guarente, 2001) and Drosophila (Rogina and Helfand, 2004), SIR2 activity is required for the extended lifespan phenotype induced by calorie restriction (reviewed in Guarente, 2005; Guarente and Picard, 2005; Longo and Kennedy, 2006). In mammals, the SIR2 homologue SIRT1 is the best characterized sirtuin. Its activities range from histone deacetylation (Braunstein et al., 1993; Imai et al., 2000; Vaquero et al., 2004; North et al., 2005) and transcriptional silencing (Shore et al., 1984; Parsons et al., 2003; Vaquero et al., 2004) to regulation of the transcriptional activities of p53 (Luo et al., 2001; Vaziri et al., 2001; Langley et al., 2002) and p300 (Bouras et al., 2005) via deacetylation. The retinoblastoma-binding protein 1 (RBP1, also known as ARID4A) and the breast carcinoma-associated antigen (BCAA, also known as ARID4B) are transcriptional repressors (Lai et al., 1999a, b; Fleischer et al., 2003; Binda et al., 2006) found in association with the mSIN3A/HDAC1 complex (Lai et al., 2001; Skowyra et al., 2001; Kuzmichev et al., 2002; Fleischer et al., 2003; Meehan et al., 2004; Nikolaev et al., 2004; Doyon et al., 2006). RBP1 and BCAA repress transcription in both HDAC-independent (R1) and HDACdependent (R2) manners (Lai et al., 1999a; Fleischer et al., 2003; Binda et al., 2006). The R1 and R2 domains have been functionally characterized in both human

SIRT1 regulates HDAC1-dependent repression O Binda et al

3385

(Lai et al., 1999a; Binda et al., 2006) and murine RBP1 (O Binda and PE Branton, unpublished) as well as human BCAA (Binda et al., 2006). Although R1 and R2 are unique to RBP1 and BCAA, the R1 domain encompasses an A/T-rich interacting domain (ARID; Herrscher et al., 1995), a domain phylogenically conserved in eukaryotes (Kortschak et al., 2000), which, in the case of RBP1, displays sequence nonspecific low DNA-binding affinity (Patsialou et al., 2005). The R1 domain of RBP1 and BCAA are the only instances in the literature of ARID-mediated transcriptional repression activity (Binda et al., 2006). Interestingly, adjacent to the carboxy terminal end of the ARID, two SUMOylation sites mediate strong transcriptional repression in RBP1 and BCAA (Binda et al., 2006). Sequence analysis (O Binda, unpublished observation) of the other ARID proteins (Wilsker et al., 2005) reveal that ARID1A and ARID1B have a conserved putative CKxE (see Hay, 2005 for review) SUMOylation site (LKPP) about 50 residues downstream of their ARID; ARID3B has an IKKE sequence about 60 residues from its ARID; ARID5A has an LKGE motif adjacent to the ARID and an IKIE motif about 30 residues further away; ARID5A putative SUMOylation motif is conserved in ARID5B IKGE motif; while JARID1B is the only Jumonji/ARID protein with a potential SUMOylation motif (IKIE) about 60 residues after its ARID. Therefore, SUMOylation of sequences adjacent to the ARID may be involved in regulating its function(s), presumably DNA-binding affinity or transcriptional repression. Both RBP1 and BCAA can associate with the mSIN3A/HDAC1 complex via a direct interaction between the SAP30 subunit and the R2 region (Lai et al., 2001; Binda et al., 2006). Overexpression of either of these two RBP1 family members leads to cell growth inhibition and induction of senescence (Binda et al., 2006), a cellular event also controlled by the mSIN3A/ HDAC1 complex subunits p33ING1b (Garkavtsev et al., 1996; Garkavtsev and Riabowol, 1997) and p33ING2 (Pedeux et al., 2005). Induction of this phenotype by RBP1 and BCAA requires the presence of an intact R1 region as well as the R2 HDAC-dependent transcriptional repression domain (Binda et al., 2006). Preliminary studies in yeast on transcriptional repression by RBP1 (A Lai, BK Kennedy and PE Branton, unpublished) suggested a genetic link with SIR2. Here we demonstrate for the first time that association of the class III HDAC, SIRT1, with the R2 region of RBP1 and BCAA proteins alleviates their HDAC-dependent transcriptional repression activity. These findings suggest a novel mechanism of regulation of the class I deacetylase HDAC1 transcriptional repression activity by the class III deacetylase SIRT1. Results BCAA and RBP1 associate with SIRT1 Endogenously expressed RBP1 and SIRT1 were immunoprecipitated from a nuclear protein extract using

Figure 1 BCAA and RBP1 associate with SIRT1. (a) Endogenously expressed RBP1 and SIRT1 were immunoprecipitated from an extract of mouse brain nuclear proteins using the LY11 (Upstate, Billerica, MA, USA) and 2G1/F7 (Upstate) mouse monoclonal antibodies, respectively. The mouse monoclonal antibody RK5C1 (Santa Cruz, Santa Cruz, CA, USA) was used as a negative control. The immunoprecipitates were analysed by immunoblotting using either LY11 or a-SIR2 (Upstate). (b) H1299 cells were co-transfected with plasmids expressing Flag alone or Flag-SIRT1 and HA-RBP1 or HA-BCAA. Whole cell extracts were immunoprecipitated using a-Flag M2-agarose and analysed by immunoblotting using either the a-HA or the a-Flag M2 antibodies.

specific antibodies. Figure 1a shows that the mouse monoclonal antibody a-Gal4 (RK5C1), used as a negative control, did not co-immunoprecipitate RBP1 or SIRT1 (first lane), but that SIRT1 was co-immunoprecipitated with RBP1 (second lane; LY11). Likewise, RBP1 was co-immunoprecipitated with a SIRT1-specific antibody (third lane; 2G1/F7), demonstrating an interaction in vivo between RBP1 and SIRT1. To demonstrate an interaction between SIRT1 and the RBP1-related protein BCAA, H1299 cells were transfected with plasmid DNAs expressing full-length HA-tagged BCAA or RBP1 along with Flag-SIRT1 or Flag alone. Figure 1b shows that both RBP1 family members are detected specifically in Flag-SIRT1 immunoprecipitates (lanes 3 and 4), but not with the control immunoprecipitates (lanes 1 and 2). In mammals, there are seven members in the sirtuin family of proteins (see Michan and Sinclair, 2007 for review). SIRT1, SIRT6 and SIRT7 are nuclear whereas SIRT2, SIRT3, SIRT4 and SIRT5 are cytoplasmic, although a small pool of nuclear SIRT2 exists (North et al., 2003; Bae et al., 2004) that probably accumulates via regulation of its nuclear export (Wilson et al., 2006). To determine if sirtuins other than SIRT1 associate with RBP1 proteins, co-immunoprecipitation experiments were conducted on whole cell protein extracts from Oncogene


3386

cells transiently expressing either HA-RBP1 or HABCAA, and Flag-SIRT1-7 (see Supplementary Figure S1). As expected, the cytoplasmic sirtuins did not co-immunoprecipitate with either BCAA or RBP1. However, SIRT4 expression being very low, as reported elsewhere (North et al., 2005), a possible interaction could not be neglected. Furthermore, neither SIRT6 nor SIRT7 nuclear sirtuins associated with RBP1 family members. These results suggested that SIRT1 may perform a specialized function towards RBP1 and BCAA that cannot be compensated by the other nuclear sirtuins. In addition, they indicated that the RBP1 family members associate with two distinct classes of HDACs: the class I, present in the mSIN3A/HDAC1

complex and an hitherto uncharacterized class IIIcontaining complex. The R2 region of BCAA and RBP1 associates with SIRT1 To determine which region of RBP1 proteins was required for SIRT1 association, a panel of amino- and carboxy-terminal deletion mutant forms of RBP1 and BCAA as well as the BCAAMCF-7 isoforms (Binda et al., 2006) were tested (see Supplementary Figure S2 for an illustration of RBP1/BCAA derivatives and a summary of the results). Co-immunoprecipitation experiments showed that the association with SIRT1 was dependent

Figure 2 The R2 region of BCAA and RBP1 associates with SIRT1. (a) Illustration of RBP1 and BCAA most relevant features with their position in amino acid residues. Immunoprecipitation and immunoblotting studies were conducted as in Figure 1b except that truncated forms of HA-RBP1 and HA-BCAA were expressed. (b) Immunoprecipitation studies were conducted as in panel (a), but the R2 region was in vitro transcribed and labelled with biotinylated lysine-tRNA during in vitro translation, then detected using HRP-conjugated streptavidin. Oncogene


3387

on the presence of the R2 region (Figure 2a), suggesting that SIRT1 associates with RBP1 proteins through the R2 region. To determine whether the R2 domain alone can bind SIRT1, a fragment of BCAA corresponding to the R2 domain was in vitro translated and incubated with lysates from cells expressing either Flag-SIRT1 or Flag alone. Upon a-Flag immunoprecipitation, R2 was detected only in precipitates containing Flag-SIRT1 (Figure 2b). Thus R2 is necessary and sufficient for mediating association with SIRT1. SIRT1 is recruited to the RBP1 proteins via the ING subunits of the mSIN3A/HDAC1 complex SIRT1 could interact with the R2 region of RBP1 proteins either directly or indirectly via an R2-associated protein such as a subunit of the mSIN3A/HDAC1 complex. We have shown previously that the mSIN3A/ HDAC1 complex is recruited to R2 through a direct interaction with the SAP30 subunit (Lai et al., 2001; Binda et al., 2006). Thus, if the SIRT1 interaction with R2 occurs via the mSIN3A/HDAC1 complex, depletion of SAP30 might reduce the amount of R2-bound SIRT1. To determine if SIRT1 association with the R2 region requires the presence of SAP30, the latter was knocked down by RNA interference methods. Figure 3 shows that treatment with SAP30-specific siRNA significantly reduced the amount of SAP30 protein in the cells. In addition, RBP1 lost the capacity to associate with SIRT1 with decreased levels of SAP30. These observations suggested that the interaction between R2 and SIRT1 occurs via the mSIN3A/HDAC1 complex, as SAP30 is the sole subunit known to make direct contact with the R2 region.

Figure 3 SIRT1 is recruited to RBP1 via the mSIN3A/HDAC1 complex. H1299 cells were co-transfected with the indicated siRNAs and a DNA vector expressing HA-RBP1. The a-HA immunoprecipitates from whole cell extracts were resolved by SDS–PAGE and analysed by immunoblotting using a-HA, a-SIRT1, or a-SAP30 antibodies.

The tumour suppressor inhibitor of growth protein p33ING1b associates with SIRT1 (Kataoka et al., 2003) and both p33ING1b and p33ING2 are found in association with the mSIN3A/HDAC1 complex (Skowyra et al., 2001; Kuzmichev et al., 2002; Vieyra et al., 2002; Doyon et al., 2006). It is therefore plausible that these two ING proteins could bridge SIRT1 to the R2 region of RBP1 proteins. To confirm that ING proteins associate with the RBP1 proteins, cells were transfected with cDNAs expressing Flag-tagged ING proteins and their ability to interact with RBP1 proteins was examined by coimmunoprecipitation. Figure 4a and Supplementary Figure S3 show that indeed BCAA and RBP1 co-immunoprecipitated with both p33ING1b and p33ING2 whereas, the isoform p24ING1c, which has a distinct, shorter amino terminal region, did not. This is in agreement with previous studies, which demonstrated that both p33ING1b (Skowyra et al., 2001; Kuzmichev et al., 2002) and p33ING2 (Doyon et al., 2006) are inherent components of the mSIN3A/HDAC1 complex, whereas the p24ING1c isoform is not incorporated in the complex (Skowyra et al., 2001). To address whether ING proteins could directly associate with SIRT1, in vitro binding assays were conducted with GST, GST-HDAC1 and GST-SIRT1 affinity purified recombinant proteins. Figure 4b confirms that in vitro translated SAP30 could bind directly to HDAC1 (Zhang et al., 1998). Interestingly, all ING proteins tested were found to bind directly to both HDAC1 and SIRT1. These results were then confirmed in vivo in co-immunoprecipitation studies. Figure 4c shows that both ING1 isoforms p24ING1c and p33ING1b associated in vivo with HDAC1 and SIRT1. In addition, although p33ING2 expression was lower than that of ING1 isoforms, p33ING2 was also seen to associate with both HDAC1 and SIRT1 (data not shown). These findings imply that p33ING1b and p33ING2 are capable of bridging SIRT1 to the mSIN3A/HDAC1 complex by associating directly with the SAP30 subunit via their conserved amino terminal region (p33ING1b residues 1– 125; Kuzmichev et al., 2002) and to deacetylases via the PHD-containing carboxy terminal three-quarter (Figures 4b and c). Interestingly, p24ING1c, which lacks the first 69 amino terminal residues, fails to associate with either RBP1 or BCAA (Figure 4a), but retains the capacity to bind HDACs (Figures 4b and c). This suggests that the conserved region 1 (Doyon et al., 2006) may mediate important protein-protein interactions for the assembly of the mSIN3A/HDAC1 complex. The repression activity of the R2 region is negatively regulated by SIRT1 To decipher the functional role of SIRT1 interaction with the R2 transcriptional repression domain, luciferase reporter assays were conducted in the presence or absence of resveratrol, a sirtuin activator (Howitz et al., 2003), using a Gal4 DNA binding domain fused to R2 and the Gal4-responsive reporter G5TKluc. Figure 5a shows that increasing amounts of resveratrol (0, 10, 50, 100 mM) resulted in proportional relief of repression Oncogene


3388

Figure 4 p33ING1b and p33ING2 allow SIRT1 recruitment to R2. (a) DNA vectors expressing Flag-tagged INGs were co-transfected with a plasmid expressing HA-BCAA in HeLa cells. Whole cell extracts were immunoprecipitated using a-Flag M2-agarose and analysed by immunoblotting using either the HA.11 or the a-Flag M2 antibodies, as indicated. (b) In vitro transcribed and translated SAP30 and ING proteins were incubated with GST, GST-HDAC1 or GST-SIRT1. Complexes were pulled-down using glutathione sepharose 4, washed extensively, resolved by SDS–PAGE, transferred to a PVDF membrane and analysed by immunoblotting using HRP-conjugated streptavidin (Promega). (c) Flag-tagged HDAC1 and SIRT1 were expressed in HeLa cells along with HAtagged ING1 isoforms. Whole cell extracts were immunoprecipitated using a-Flag M2-agarose and analysed by immunoblotting using either the HA.11 or the a-Flag M2 antibodies.

Oncogene

Figure 5 SIRT1 activity negatively regulates R2-mediated repression. (a) HeLa cells were co-transfected with pG5TKluc, phRL RSVluc, and Gal4 or Gal4-R2. The cells were either treated with dimethyl sulfoxide or increasing amount of resveratrol, as indicated on the graphic, for a period of 24 h. Luciferase activity was arbitrarily set at 100% for the Gal4-alone control. (b) This experiment was conducted as in panel (a) except that the cells were treated with sirtinol, instead of resveratrol. (c) This experiment was conducted as in panel (a) except that a Gal4-HDAC1 expressing plasmid was included. The results are shown as ‘Fold Repression’ on the graphic for easier comparison between R2 and HDAC1 repression. The experiments for this figure were conducted at least thrice in duplicate.

mediated by R2, such that at a concentration of 100 mM, the repression activity of R2 was completely alleviated. On the contrary, using the sirtuin inhibitor sirtinol (Grozinger et al., 2001) at 100 mM, R2-mediated repression was significantly enhanced (approximately threefold induction; Figure 5b). The Gal4 and Gal4-R2 protein levels were not significantly affected by the drug treatments (data not shown). These results suggested that SIRT1 activity inhibits the mSIN3A/HDAC1


3389

transcriptional repression activity associated with the R2 region. HDAC1 was recently shown to be acetylated (Qiu et al., 2006) and thus could be a target of the SIRT1 deacetylase activity recruited to the mSIN3A/HDAC1 complex. The highly acetylated form of HDAC1 is found in repressed chromatin, while the deacetylase found on transcriptionally active chromatin has a lower level of acetylation (Qiu et al., 2006). Furthermore, this post-translational modification appears to regulate HDAC1 histone deacetylase activity in vitro (Qiu et al., 2006). We therefore overexpressed Gal4-tagged HDAC1 in the context of the transcriptional repression experiments described above. Figure 5c shows that as with R2, HDAC1-mediated transcriptional repression activity can be alleviated by resveratrol whereas sirtinol significantly enhanced HDAC1 repression activity.

Discussion We describe for the first time in this report the association of the class III deacetylase SIRT1 with the mSIN3A/HDAC1 complex. In previous studies involving biochemically purified mSIN3A/HDAC1 complexes, SIRT1 has never been reported as an inherent subunit of the complex. The SIRT1 deacetylase has previously been linked to other transcriptional regulators such as p53 and p300 and may therefore be involved in the regulation of the transcriptional repression activities of RBP1 proteins. SIRT1 was found in association with both full-length RBP1 and BCAA (Figure 1). Interestingly, the interaction with SIRT1 required the R2 region (Figure 2a), which has already been shown to associate with class I HDACs (Lai et al., 2001). Furthermore, SIRT1 was the sole sirtuin found to associate with RBP1 proteins. Because R2 is known to associate with the mSIN3A/ HDAC1 complex via a direct interaction with the SAP30 subunit (Lai et al., 2001; Binda et al., 2006) and, as shown here, reduction of SAP30 levels by RNAi considerably reduced the association of SIRT1 with RBP1, it is likely that SIRT1 is recruited to RBP1 and BCAA via the mSIN3A/HDAC1 complex. In vitro binding assays demonstrated that the tumour suppressors p33ING1b and p33ING2 provide a direct link between SIRT1 and the mSIN3A/HDAC1 complex subunits HDAC1 and SAP30, suggesting that the association of SIRT1 with the R2 region of RBP1 and BCAA occurs through the ING subunits of the mSIN3A/HDAC1 complex. Recently, HDAC1 was shown to be acetylated and this post-translational modification appeared to be involved in the targeting of HDAC1 to heterochromatic regions of the nucleus, thus suggesting a possible regulatory mechanism via localization of histone deacetylase activity (Qiu et al., 2006). As a consequence, it was postulated that SIRT1 may negatively regulate the mSIN3A/HDAC1 repression activity associated with R2 via a mechanism involving deacetylation of HDAC1.

Accordingly, chemical agonists and antagonists of SIRT1 affected HDAC1-mediated transcriptional repression in a manner similar to R2 (Figure 5) without significantly affecting the expression level of HDAC1 (data not shown). These results suggest that SIRT1 may regulate not only the transcriptional repression activity of R2, but also that of HDAC1 or HDAC1-associated factors. We are aware that studies on transcriptional repression using the Gal4 reporter system may not accurately reflect chromatin regulation by RBP1 and BCAA. Unfortunately, there are no endogenous genes known to be regulated by either RBP1 or BCAA. In addition, neither RBP1 nor BCAA have been shown convincingly to be involved in the regulation of endogenous E2Fdependent promoters. Furthermore, resveratrol inhibits cellular proliferation in numerous cell lines (Lu and Serrero, 1999; Sgambato et al., 2001) and sirtinol irreversibly affects the cell cycle by inducing senescence (G1-like cell cycle exit; Ota et al., 2006, 2007). Therefore, even if studies were performed on endogenous E2Fdependent genes it would be difficult to discern between the cell cycle arrest effects and E2F-dependent transcriptional regulation by RBP1/BCAA, if any. In fact, we did conduct such a study with the mSIN3A-regulated gene cyclin B1 (Dannenberg et al., 2005), and found that its protein level was greatly reduced by sirtinol (data not shown); however, as discussed above, we were concerned that the sirtinol treatment may not have had an RBP1/BCAA-specific effect on cyclin B1 expression, but rather induced an indirect effect due to the G1-arrest. Unlike the other nuclear sirtuins, SIRT6 and SIRT7, SIRT1 is excluded from DAPI- and Hoechst 33342stained chromatin (heterochromatin; Michishita et al., 2005; Bordone et al., 2006). Thus SIRT1 appears to associate with euchromatin despite its reported role in the formation of heterochromatin (Vaquero et al., 2004). These contradictory observations may reflect dual functions associated with SIRT1. Nonetheless, the euchromatin localization of SIRT1 correlates with our observation that SIRT1 activity negatively regulates HDAC1-dependent transcriptional repression. It has been reported that SIRT1 activity can overcome p53/PML-induced senescence (Langley et al., 2002). We have previously shown that the R2 region is necessary for induction of senescence by the RBP1 proteins (Binda et al., 2006) and since SIRT1 activity is deleterious to R2 repression activity, it may also negatively regulate RBP1-induced senescence. The observation that inhibition of SIRT1 by either sirtinol or splitomicin induces senescence in H1299 (O Binda, unpublished observation; Ota et al., 2006), HUVEC (Ota et al., 2007) and MCF-7 (Ota et al., 2006) cells correlates with increased R2-mediated repression and RBP1-induced senescence, thereby providing indirect evidence suggesting that SIRT1 activity may antagonize the growth arrest phenotype caused by RBP1 family protein overexpression. A recent study, has reported that SIRT1 is gradually lost in aging cells (Sasaki et al., 2006). So, while SIRT1 levels fade away as the cells age, HDAC1-dependent transcriptional repression could be Oncogene


3390

relieved from SIRT1 inhibition and lead to senescence, as observed with p53/PML (Langley et al., 2002). The NCOR and SMRT corepressor complexes, like mSIN3A/HDAC1, also rely on HDAC activity for transcriptional silencing. They have been found in association with HDAC3 (Li et al., 2000; Yoon et al., 2003), HDAC4 (Huang et al., 2000; Fischle et al., 2002) and HDAC5 (Huang et al., 2000). However, many subunits of the mSIN3A/HDAC1 complex, essentially HDAC1, SAP30 and mSIN3A, can be found in biochemically purified NCOR (Laherty et al., 1998) and SMRT (Underhill et al., 2000) complexes. Furthermore, both NCOR and SMRT associate with SIRT1 through the repression domain RD1 (Picard et al., 2004), which is the main binding surface for mSIN3A (Heinzel et al., 1997; Nagy et al., 1997), suggesting that SIRT1 may be recruited to NCOR/SMRT via the mSIN3A complex. Importantly, SIRT1 was also shown to negatively regulate the transcriptional activity of PPAR-g via its NCOR cofactor (Picard et al., 2004). It therefore appears that SIRT1 activity can damper both transcriptional activation (PPAR-g) and transcriptional repression (RBP1/BCAA). The class I and III HDACs RPD3 and SIR2 are known in lower organisms such as yeast and flies to have opposite roles in the regulation of life span (Kim et al., 1999; Rogina et al., 2002; Rogina and Helfand, 2004). Our results suggest that not only do these two classes of HDACs have opposite functions, but also that SIRT1 actually regulates HDAC1-mediated transcriptional silencing. Furthermore, because of previous results from our lab and the implication of SIRT1 in preventing senescence, we believe that SIRT1 and RBP1 proteins may be important regulators in the equilibrium between ageing and longevity.

Materials and methods Cell lines and DNA transfections H1299 cells (ATCC CRL-5803) and HeLa cells (ATCC CCL-2) were maintained in Dulbecco’s modified Eagle medium (DMEM; Invitrogen, Burlington, ON, Canada) supplemented with 10% fetal bovine serum (FBS; Cansera, Etobicoke, ON, Canada), 100 U ml 1 Penicillin G, 100 mg ml1 Streptomycin and 2 mM Glutamine (Invitrogen). DNA transfections were conducted using TranIT LT1 (Mirus, Madison, WI, USA) using a DNA:LT1 ratio of 1:3. Plasmids RBP1 and BCAA cDNA expression vectors were previously described (Binda et al., 2006). pcDNA3-SIRT1-Flag was obtained from Fuyuki Ishikawa (Takata and Ishikawa, 2003). The Flag-tagged SIRT2-7 mammalian expression vectors were obtained from Eric Verdin (North et al., 2005). The cDNAs of p33ING1b and p24ING1c were cloned from human total RNA by reverse transcription (RT)–PCR into pCMV 3 Flag (Stratagene, La Jolla, CA, USA) using EcoRI and XhoI restriction sites, while p33ING2 cDNA was subcloned from pFLAG-CMV-6c-ING2 (Pedeux et al., 2005) by PCR into pCMV 3 Flag (Stratagene) using EcoRI and XhoI restriction sites. The INGs cDNAs were also subcloned into pET-33b( þ ) Oncogene

(Novagen, Mississaga, ON, Canada) using SacI and XhoI for in vitro transcription/translation and in vitro binding assays. Luciferase assays Transcriptional repression assays were essentially conducted as described before (Binda et al., 2006). To assess SIRT1dependent biological function on R2-mediated repression, the cells were treated with resveratrol (Sigma, Oakville, ON, Canada), sirtinol (Sigma) or DMSO for 24 h prior to the luciferase assay measurements. RNAi depletion of SAP30 The siRNA duplexes were obtained from Dharmacon (Lafayette, CO, USA) and information on the targeted sequence can be provided upon request. H1299 cells were seeded at a density of 5 105 cells per 60 mm-diameter plate the day prior to transfection. Before transfection, the medium was replaced with 2 ml of fresh DMEM containing 10% FBS. Then the cells were transfected with 8.0mg of HA expression vector DNA using 24 ml TransIT LT1 in 500 ml serum-free DMEM. In parallel, the siRNAs were transfected using 7.5 ml of 20 mM stock (50 nM final concentration) and 15 ml TransIT TKO (Mirus) in 500 ml. The medium was changed 24 h posttransfection. The cells were rinsed with phosphate-buffered saline (PBS), harvested and lysed in 450 ml nuclear lysis buffer (Lai et al., 1999b) 48 h post-transfection. SAP30 knockdown was detected by immunoblotting using the rabbit polyclonal a-SAP30 antibody (Abcam ab15034). Immunoprecipitation Co-immunoprecipitation experiments were conducted either in H1299 or in HeLa cells. Briefly, 4.0 mg of HA plasmid and 4.0 mg of Flag plasmid DNAs were transfected into the cells. The cells were lysed 48 h post-transfection in nuclear lysis buffer supplemented with EDTA-free protease inhibitor cocktail (Roche, Laval, QC, Canada). a-Flag M2-agarose (Sigma) was used to immunoprecipitate the complexes (as previously described by Binda et al., 2006), which were resolved by SDS– polyacrylamide gel electrophoresis (SDS–PAGE), transferred to PVDF membrane and analysed by immunoblotting using either a-HA monoclonal antibody (HA.11 from Covance (Emeryville, CA, USA)) or HRP-conjugated a-Flag M2 (Sigma). To overcome the masking of the R2 signal by the IgG light chain, the studies on the interaction of SIRT1 and R2 were conducted as outlined above, but HA-tagged R2 was in vitro transcribed and translated from a pcDNA3 plasmid using the TNT Coupled Reticulocyte Lysate System (Promega, Nepean, ON, Canada) and labelled with biotinylated lysinetRNA (Transcend from Promega) for detection by HRPconjugated streptavidin (Promega). Recombinant protein purification GST, GST-HDAC1 and GST-SIRT1 were purified essentially as described previously (North et al., 2005; Binda et al., 2006). In vitro binding assays In vitro translation (IVT) was performed using the TNT Coupled Reticulocyte Lysate System (Promega) and biotinylated lysine-tRNA (Transcend tRNA; Promega) according to the manufacturer’s guidelines. Equivalent amounts of IVT products were incubated with 2.0 mg GST recombinant proteins in nuclear lysis buffer (100 mM KCl), washed six times with 1mL nuclear lysis buffer (100 mM KCl), resolved by Sodium dodecyl sulphate–polyacrylamide gel electrophoresis (SDS–PAGE), transferred to a PVDF membrane and analysed by immunoblotting using HRP-conjugated streptavidin (Promega).


3391 Acknowledgements We thank Laurent Huck and Peter Moffett for critically reviewing this manuscript, and Jean-Se´bastien Roy for

valuable discussions. This work was supported through grants from the Canadian Cancer Society, the National Cancer Institute of Canada and the Canadian Institutes for Health Research.

References Bae NS, Swanson MJ, Vassilev A, Howard BH. (2004). Human histone deacetylase SIRT2 interacts with the homeobox transcription factor HOXA10. J Biochem (Tokyo) 135: 695–700. Baylin SB, Esteller M, Rountree MR, Bachman KE, Schuebel K, Herman JG. (2001). Aberrant patterns of DNA methylation, chromatin formation and gene expression in cancer. Hum Mol Genet 10: 687–692. Binda O, Roy J-S, Branton PE. (2006). RBP1 family proteins exhibit SUMOylation-dependent transcriptional repression and induce cell growth inhibition reminiscent of senescence. Mol Cell Biol 26: 1917–1931. Blander G, Guarente L. (2004). The SIR2 family of protein deacetylases. Annu Rev Biochem 73: 417–435. Bordone L, Motta MC, Picard F, Robinson A, Jhala US, Apfeld J et al. (2006). Sirt1 regulates insulin secretion by repressing UCP2 in pancreatic b cells. PLoS Biol 4: e31. Bouras T, Fu M, Sauve AA, Wang F, Quong AA, Perkins ND et al. (2005). SIRT1 deacetylation and repression of p300 involves lysine residues 1020/1024 within the cell cycle regulatory domain 1. J Biol Chem 280: 10264–10276. Braunstein M, Rose AB, Holmes SG, Allis CD, Broach JR. (1993). Transcriptional silencing in yeast is associated with reduced nucleosome acetylation. Genes Dev 7: 592–604. Dannenberg J-H, David G, Zhong S, van der Torre J, Wong WH, DePinho RA. (2005). mSin3A corepressor regulates diverse transcriptional networks governing normal and neoplastic growth and survival. Genes Dev 19: 1581–1595. Doyon Y, Cayrou C, Ullah M, Landry A-J, Coˆte´ V, Selleck W et al. (2006). ING tumor suppressor proteins are critical regulators of chromatin acetylation required for genome expression and perpetuation. Mol Cell 21: 51–64. Finnin MS, Donigian JR, Cohen A, Richon VM, Rifkind RA, Marks PA et al. (1999). Structures of a histone deacetylase homologue bound to the TSA and SAHA inhibitors. Nature 401: 188–193. Fischle W, Dequiedt F, Hendzel MJ, Guenther MG, Lazar MA, Voelter W et al. (2002). Enzymatic activity associated with class II HDACs is dependent on a multiprotein complex containing HDAC3 and SMRT/N-CoR. Mol Cell 9: 45–57. Fleischer TC, Yun UJ, Ayer DE. (2003). Identification and characterization of three new components of the mSin3A corepressor complex. Mol Cell Biol 23: 3456–3467. Garkavtsev I, Kazarov A, Gudkov A, Riabowol K. (1996). Suppression of the novel growth inhibitor p33ING1 promotes neoplastic transformation. Nat Genet 14: 415–420. Garkavtsev I, Riabowol K. (1997). Extension of the replicative life span of human diploid fibroblasts by inhibition of the p33ING1 candidate tumor suppressor. Mol Cell Biol 17: 2014–2019. Grozinger CM, Chao ED, Blackwell HE, Moazed D, Schreiber SL. (2001). Identification of a class of small molecule inhibitors of the sirtuin family of NAD-dependent deacetylases by phenotypic screening. J Biol Chem 276: 38837–38843. Guarente L. (2005). Calorie restriction and SIR2 genes—towards a mechanism. Mech Ageing Dev 126: 923–928. Guarente L, Picard F. (2005). Calorie restriction—the SIR2 connection. Cell 120: 473–482. Hassig CA, Fleischer TC, Billin AN, Schreiber SL, Ayer DE. (1997). Histone deacetylase activity is required for full transcriptional repression by mSin3A. Cell 89: 341–347. Hay RT. (2005). SUMO: a history of modification. Mol Cell 18: 1–12. Heinzel T, Lavinsky RM, Mullen T-M, Soderstrom M, Laherty CD, Torchia J et al. (1997). A complex containing N-CoR, mSln3 and

histone deacetylase mediates transcriptional repression. Nature 387: 43–48. Herrscher RF, Kaplan MH, Lelsz DL, Das C, Scheuermann R, Tucker PW. (1995). The immunoglobulin heavy-chain matrix-associating regions are bound by bright: a B cell-specific trans-activator that describes a new DNA- binding protein family. Genes Dev 9: 3067–3082. Howitz KT, Bitterman KJ, Cohen HY, Lamming DW, Lavu S, Wood JG et al. (2003). Small molecule activators of sirtuins extend Saccharomyces cerevisiae lifespan. Nature 425: 191–196. Huang EY, Zhang J, Miska EA, Guenther MG, Kouzarides T, Lazar MA. (2000). Nuclear receptor corepressors partner with class II histone deacetylases in a Sin3-independent repression pathway. Genes Dev 14: 45–54. Imai S-i, Armstrong CM, Kaeberlein M, Guarente L. (2000). Transcriptional silencing and longevity protein Sir2 is an NADdependent histone deacetylase. Nature 403: 795–800. Kaeberlein M, McVey M, Guarente L. (1999). The SIR2/3/4 complex and SIR2 alone promote longevity in Saccharomyces cerevisiae by two different mechanisms. Genes Dev 13: 2570–2580. Kataoka H, Bonnefin P, Vieyra D, Feng X, Hara Y, Miura Y et al. (2003). ING1 represses transcription by direct DNA binding and through effects on p53. Cancer Res 63: 5785–5792. Khochbin S, Verdel A, Lemercier C, Seigneurin-Berny D. (2001). Functional significance of histone deacetylase diversity. Curr Opin Genet Dev 11: 162–166. Kim S, Benguria A, Lai C-Y, Jazwinski SM. (1999). Modulation of life-span by histone deacetylase genes in Saccharomyces cerevisiae. Mol Biol Cell 10: 3125–3136. Kortschak RD, Tucker PW, Saint R. (2000). ARID proteins come in from the desert. Trends Biochem Sci 25: 294–299. Kuzmichev A, Zhang Y, Erdjument-Bromage H, Tempst P, Reinberg D. (2002). Role of the Sin3-histone deacetylase complex in growth regulation by the candidate tumor suppressor p33ING1. Mol Cell Biol 22: 835–848. Laherty CD, Billin AN, Lavinsky RM, Yochum GS, Bush AC, Sun J-M et al. (1998). SAP30, a component of the mSin3 corepressor complex involved in N-CoR-mediated repression by specific transcription factors. Mol Cell 2: 33–42. Laherty CD, Yang W-M, Sun J-M, Davie JR, Seto E, Eisenman RN. (1997). Histone deacetylases associated with the mSin3 corepressor mediate mad transcriptional repression. Cell 89: 349–356. Lai A, Kennedy BK, Barbie DA, Bertos NR, Yang XJ, Theberge M-C et al. (2001). RBP1 recruits the mSIN3-histone deacetylase complex to the pocket of retinoblastoma tumor suppressor family proteins found in limited discrete regions of the nucleus at growth arrest. Mol Cell Biol 21: 2918–2932. Lai A, Lee JM, Yang W-M, DeCaprio JA, Kaelin Jr WG, Seto E et al. (1999a). RBP1 recruits both histone deacetylase-dependent and independent repression activities to retinoblastoma family proteins. Mol Cell Biol 19: 6632–6641. Lai A, Marcellus RC, Corbeil HB, Branton PE. (1999b). RBP1 induces growth arrest by repression of E2F-dependent transcription. Oncogene 18: 2091–2100. Langley E, Pearson M, Faretta M, Bauer U-M, Frye RA, Minucci S et al. (2002). Human SIR2 deacetylates p53 and antagonizes PML/ p53-induced cellular senescence. EMBO J 21: 2383–2396. Li J, Wang J, Wang J, Nawaz Z, Liu JM, Qin J et al. (2000). Both corepressor proteins SMRT and N-CoR exist in large protein complexes containing HDAC3. EMBO J 19: 4342–4350. Longo VD, Kennedy BK. (2006). Sirtuins in aging and age-related disease. Cell 126: 257–268. Oncogene


3392 Lu R, Serrero G. (1999). Resveratrol, a natural product derived from grape, exhibits antiestrogenic activity and inhibits the growth of human breast cancer cells. J Cell Physiol 179: 297–304. Luo J, Nikolaev AY, Imai S-i, Chen D, Su F, Shiloh A et al. (2001). Negative control of p53 by Sir2a promotes cell survival under stress. Cell 107: 137–148. Meehan WJ, Samant RS, Hopper JE, Carrozza MJ, Shevde LA, Workman JL et al. (2004). Breast cancer metastasis suppressor 1 (BRMS1) forms complexes with retinoblastoma-binding protein 1 (RBP1) and the mSin3 histone deacetylase complex and represses transcription. J Biol Chem 279: 1562–1569. Michan S, Sinclair D. (2007). Sirtuins in mammals: insights into their biological function. Biochem J 404: 1–13. Michishita E, Park JY, Burneskis JM, Barrett JC, Horikawa I. (2005). Evolutionarily conserved and nonconserved cellular localizations and functions of human SIRT proteins. Mol Biol Cell 16: 4623–4635. Minucci S, Pelicci PG. (2006). Histone deacetylase inhibitors and the promise of epigenetic (and more) treatments for cancer. Nat Rev Cancer 6: 38–51. Nagy L, Kao H-Y, Chakravarti D, Lin RJ, Hassig CA, Ayer DE et al. (1997). Nuclear receptor repression mediated by a complex containing SMRT, mSin3A, and histone deacetylase. Cell 89: 373–380. Nikolaev AY, Papanikolaou NA, Li M, Qin J, Gu W. (2004). Identification of a novel BRMS1-homologue protein p40 as a component of the mSin3A/p33ING1b/HDAC1 deacetylase complex. Biochem Biophys Res Commun 323: 1216–1222. North BJ, Marshall BL, Borra MT, Denu JM, Verdin E. (2003). The human Sir2 ortholog, SIRT2, is an NAD+-dependent tubulin deacetylase. Mol Cell 11: 437–444. North BJ, Schwer B, Ahuja N, Marshall B, Verdin E. (2005). Preparation of enzymatically active recombinant class III protein deacetylases. Methods 36: 338–345. Ota H, Akishita M, Eto M, Iijima K, Kaneki M, Ouchi Y. (2007). Sirt1 modulates premature senescence-like phenotype in human endothelial cells. J Mol Cell Cardiol 43: 571–579. Ota H, Tokunaga E, Chang K, Hikasa M, Iijima K, Eto M et al. (2006). Sirt1 inhibitor, sirtinol, induces senescence-like growth arrest with attenuated Ras-MAPK signaling in human cancer cells. Oncogene 25: 176–185. Parsons XH, Garcia SN, Pillus L, Kadonaga JT. (2003). Histone deacetylation by Sir2 generates a transcriptionally repressed nucleoprotein complex. PNAS 100: 1609–1614. Patsialou A, Wilsker D, Moran E. (2005). DNA-binding properties of ARID family proteins. Nucleic Acids Res 33: 66–80. Pedeux R, Sengupta S, Shen JC, Demidov ON, Saito S, Onogi H et al. (2005). ING2 regulates the onset of replicative senescence by induction of p300-dependent p53 acetylation. Mol Cell Biol 25: 6639–6648. Picard F, Kurtev M, Chung N, Topark-Ngarm A, Senawong T, Machado de Oliveira R et al. (2004). Sirt1 promotes fat mobilization in white adipocytes by repressing PPAR-g. Nature 429: 771–776. Qiu Y, Zhao Y, Becker M, John S, Parekh BS, Huang S et al. (2006). HDAC1 acetylation is linked to progressive modulation of steroid receptor-induced gene transcription. Mol Cell 22: 669–679. Rogina B, Helfand SL. (2004). Sir2 mediates longevity in the fly through a pathway related to calorie restriction. PNAS 101: 15998–16003.

Rogina B, Helfand SL, Frankel S. (2002). Longevity regulation by Drosophila Rpd3 deacetylase and caloric restriction. Science 298: 1745. Rountree MR, Bachman KE, Herman JG, Baylin SB. (2001). DNA methylation, chromatin inheritance, and cancer. Oncogene 20: 3156–3165. Sasaki T, Maier B, Bartke A, Scrable H. (2006). Progressive loss of SIRT1 with cell cycle withdrawal. Aging Cell 5: 413–422. Sgambato A, Ardito R, Faraglia B, Boninsegna A, Wolf FI, Cittadini A. (2001). Resveratrol, a natural phenolic compound, inhibits cell proliferation and prevents oxidative DNA damage. Mutat Res 496: 171–180. Shore D, Squire M, Nasmyth KA. (1984). Characterization of two genes required for the position-effect control of yeast mating-type genes. EMBO J 3: 2817–2823. Skowyra D, Zeremski M, Neznanov N, Li M, Choi Y, Uesugi M et al. (2001). Differential association of products of alternative transcripts of the candidate tumor suppressor ING1 with the mSin3/HDAC1 transcriptional corepressor complex. J Biol Chem 276: 8734–8739. Takata T, Ishikawa F. (2003). Human Sir2-related protein SIRT1 associates with the bHLH repressors HES1 and HEY2 and is involved in HES1- and HEY2-mediated transcriptional repression. Biochem Biophys Res Commun 301: 250–257. Thiagalingam SAM, Cheng K-H, Lee HJ, Mineva N, Thiagalingam A, Ponte JF. (2003). Histone deacetylases: unique players in shaping the epigenetic histone code. Ann NY Acad Sci 983: 84–100. Tissenbaum HA, Guarente L. (2001). Increased dosage of a sir-2 gene extends lifespan in Caenorhabditis elegans. Nature 410: 227–230. Underhill C, Qutob MS, Yee S-P, Torchia J. (2000). A novel nuclear receptor corepressor complex, N-CoR, contains components of the mammalian SWI/SNF complex and the corepressor KAP-1. J Biol Chem 275: 40463–40470. Vaquero A, Scher M, Lee D, Erdjument-Bromage H, Tempst P, Reinberg D. (2004). Human SirT1 interacts with histone H1 and promotes formation of facultative heterochromatin. Mol Cell 16: 93–105. Vaziri H, Dessain SK, Eaton EN, Imai S-i, Frye RA, Pandita TK et al. (2001). hSIR2SIRT1 functions as an NAD-dependent p53 deacetylase. Cell 107: 149–159. Vieyra D, Loewith R, Scott M, Bonnefin P, Boisvert F-M, Cheema P et al. (2002). Human ING1 proteins differentially regulate histone acetylation. J Biol Chem 277: 29832–29839. Wilsker D, Probst L, Wain HM, Maltais L, Tucker PW, Moran E. (2005). Nomenclature of the ARID family of DNA-binding proteins. Genomics 86: 242–251. Wilson J, Le V, Zimmerman C, Marmorstein R, Pillus L. (2006). Nuclear export modulates the cytoplasmic Sir2 homologue Hst2. EMBO Rep 7: 1247–1251. Yoon H-G, Chan DW, Huang Z-Q, Li J, Fondell JD, Qin J et al. (2003). Purification and functional characterization of the human N-CoR complex: the roles of HDAC3, TBL1 and TBLR1. EMBO J 22: 1336–1346. Zhang Y, Sun Z-W, Iratni R, Erdjument-Bromage H, Tempst P, Hampsey M et al. (1998). SAP30, a novel protein conserved between human and yeast, is a component of a histone deacetylase complex. Mol Cell 1: 1021–1031.

Supplementary Information accompanies the paper on the Oncogene website (http://www.nature.com/onc).

Oncogene