PMCCPMCCPMCC

Search tips
Search criteria 

Advanced

 
Logo of molcellbPermissionsJournals.ASM.orgJournalMCB ArticleJournal InfoAuthorsReviewers
 
Mol Cell Biol. 2007 April; 27(7): 2661–2675.
Published online 2007 February 5. doi:  10.1128/MCB.01098-06
PMCID: PMC1899900

An Acetylation/Deacetylation-SUMOylation Switch through a Phylogenetically Conserved ψKXEP Motif in the Tumor Suppressor HIC1 Regulates Transcriptional Repression Activity[down-pointing small open triangle]

Abstract

Tumor suppressor HIC1 (hypermethylated in cancer 1) is a gene that is essential for mammalian development, epigenetically silenced in many human tumors, and involved in a complex pathway regulating P53 tumor suppression activity. HIC1 encodes a sequence-specific transcriptional repressor containing five Krüppel-like C2H2 zinc fingers and an N-terminal BTB/POZ repression domain. Here, we show that endogenous HIC1 is SUMOylated in vivo on a phylogenetically conserved lysine, K314, located in the central region which is a second repression domain. K314R mutation does not influence HIC1 subnuclear localization but significantly reduces its transcriptional repression potential, as does the mutation of the other conserved residue in the ψKXE consensus, E316A, or the overexpression of the deSUMOylase SSP3/SENP2. Furthermore, HIC1 is acetylated in vitro by P300/CBP. Strikingly, the K314R mutant is less acetylated than wild-type HIC1, suggesting that this lysine is a target for both SUMOylation and acetylation. We further show that HIC1 transcriptional repression activity is positively controlled by two types of deacetylases, SIRT1 and HDAC4, which increase the deacetylation and SUMOylation, respectively, of K314. Knockdown of endogenous SIRT1 by the transfection of short interfering RNA causes a significant loss of HIC1 SUMOylation. Thus, this dual-deacetylase complex induces either a phosphorylation-dependent acetylation-SUMOylation switch through a ψKXEXXSP motif, as previously shown for MEF2, or a phosphorylation-independent switch through a ψKXEP motif, as shown here for HIC1, since P317A mutation severely impairs HIC1 acetylation. Finally, our results demonstrate that HIC1 is a target of the class III deacetylase SIRT1 and identify a new posttranslational modification step in the P53-HIC1-SIRT1 regulatory loop.

HIC1 (hypermethylated in cancer 1) is located in 17p13.3 in a region frequently hypermethylated or deleted in many types of prevalent human tumors (59). HIC1 is a tumor suppressor gene since its enforced expression by stable transfection in various cancer cell lines results in a significant decrease in their clonogenic survival levels (59) and since heterozygous Hic1+/ mice develop, after 70 weeks, many different spontaneous malignant tumors (13). Moreover, animal models using Hic1 and p53 double-heterozygous-knockout mice have shown that the epigenetically silenced gene Hic1 cooperates with the mutated tumor suppressor gene p53 in determining cancer prevalence, progression, and spectrum (11). Indeed, HIC1 is a direct P53 target gene through a P53-responsive element recently identified (7). Finally, a regulatory feedback loop between HIC1 and P53 has recently been deciphered in which HIC1 directly represses the transcription of SIRT1, which deacetylates and thereby inactivates P53 (12).

HIC1 is essential for normal mammalian development and could also be implicated in a contiguous-gene syndrome, the Miller-Dieker syndrome, a severe form of lissencephaly accompanied by developmental anomalies (61). Together with displaying perinatal death and a reduction in overall size, Hic1/ mouse embryos display other developmental anomalies resembling those found in patients with Miller-Dieker syndrome (9).

The HIC1 protein is a sequence-specific transcriptional repressor containing three main functional domains: a conserved protein-protein interaction domain called BTB/POZ (Broad Complex, Tramtrack, and bric à brac/poxviruses and zinc finger) at the N terminus, five Krüppel-like C2H2 zinc fingers near its C terminus, and a central region which is not well conserved among the HIC1 proteins from various species (Fig. (Fig.1)1) (14, 16). HIC1 binds specifically to DNA through its zinc finger domain that recognizes the consensus sequence 5′-C/GNGC/GGGGCAC/ACC-3′ centered on a GGCA core motif (42). The BTB/POZ domain is a conserved structural motif found mainly in transcription factors, actin binding proteins, and substrate-specific adapters of CUL-3-based ubiquitin ligases (54). Crystal structures of the PLZF and BCL6 BTB/POZ domains have demonstrated that this domain is a tightly intertwined obligate dimer with a conserved dimerization interface (54). The homo-/heterodimerization as well as the oligomerization properties of the BTB/POZ domain are essential for the biological function of these proteins. The HIC1 BTB/POZ domain negatively regulates DNA binding of the full-length protein to a single site, whereas the oligomerization of the protein via this domain mediates cooperative DNA binding to multiple sites (42). Finally, the BTB/POZ domain is essential for the function of transcriptional repressors by directly recruiting nuclear corepressor (SMRT, N-CoR, or B-CoR)/histone deacetylase (HDAC) complexes, as shown for the human PLZF and BCL6 proteins (19, 38). Previously, we have shown that the HIC1 BTB/POZ domain is an autonomous transcriptional repression domain insensitive to trichostatin A (TSA), a specific inhibitor of class I and class II HDACs (14). The HIC1 central region is also an autonomous transcriptional repression domain that contains the short phylogenetically conserved sequence GLDLSKK, is involved in the recruitment of the CtBP corepressor (Fig. (Fig.1),1), and represses transcription in a TSA-sensitive manner (2, 16).

FIG. 1.
Identification of a new conserved motif in the central repression domain of HIC1: a SUMOylation consensus. Schematic structure of the human HIC1 protein. The BTB/POZ domain and the five C2H2 zinc fingers are represented as dotted and gray boxes, respectively. ...

During the last 5 years, SUMOylation has emerged as a new and important versatile modification of numerous nuclear proteins (23, 28, 49, 57). SUMO (small ubiquitin-related modifier) is an 11-kDa polypeptide structurally related to ubiquitin and covalently conjugated to lysine residues within target proteins in the consensus ψKXE motif, where ψ is a large hydrophobic residue and X is any amino acid. This modification involves coordinated action, in a manner similar to the ubiquitination, of an E1 SUMO-activating enzyme, a unique E2-conjugating enzyme (Ubc9), and an E3 ligase, which promotes the transfer of SUMO from Ubc9 to the target protein. The protein inhibitor of activated STAT (PIAS) family proteins (48), nucleoporin protein RanBP2 (41), and the Polycomb protein Pc2 (31) have been identified as SUMO E3 ligases with some substrates specificity. Recently, the histone deacetylase HDAC4 has been shown to promote SUMOylation of MEF2 transcription factors independently of its deacetylase activity (26, 27, 63). However, the underlying mechanisms for this remain elusive since HDAC4 has been proposed to be a bona fide E3 ligase in vivo (64) but not in vitro (27) or to promote the phosphorylation of MEF2 on an adjacent serine residue to upregulate SUMOylation (27, 63). SUMOylation is a highly dynamic modification, and multiple proteases (Sentrin-specific proteases [SENP]) are able to remove SUMO from its specific substrates (28). In contrast to ubiquitination, which generally marks proteins for rapid degradation, SUMOylation is involved in the regulation of DNA binding activity, transcriptional activity, nuclear sublocalization, and assembly of multiprotein complexes (23, 49, 57). The essential role of SUMO modification in a number of biological processes has been well established. However, it must be stressed that even though a small proportion of the substrate is modified and SUMO can be rapidly deconjugated, SUMOylation affects the long-term fate of the modified substrate. Several models have been proposed to solve these SUMO enigmas (28). Lysine residues are subject to several posttranslational modifications, including acetylation, methylation, ubiquitination, and SUMOylation. In some instances, the same lysine residue can be competitively targeted by two different modifications, resulting in distinct functional outputs, such as protein stability for the ubiquitination/SUMOylation of IκBα (17) or transcriptional activity for the acetylation/SUMOylation of the Sp3 (46) or MEF2 transcription factors (27). Recently, bioinformatics analyses identified a subset of SUMO consensus sites called the PDSM (phosphorylation-dependent SUMOylation motif), which associates a classical SUMO consensus site with an adjacent proline-directed phosphorylation site, ψKXEXXSP (29, 63), as for example, in the major MEF2 isoforms (26). In the case of MEF2A, a phosphorylation-regulated SUMOylation-acetylation switch (SAS) of this motif is essential for postsynaptic differentiation (51).

In this paper, we demonstrate that HIC1 is SUMOylated in vivo and in vitro on a phylogenetically conserved M/VK314XEP motif located in the central repression domain of HIC1. The mutation of this motif does not affect the punctate nuclear localization of HIC1 in transient transfection assays. In contrast, K314R or E316A mutations significantly impair the repression potential of HIC1, as does the overexpression of SSP3 deSUMOylase. Furthermore, HIC1 is acetylated by P300/CBP. Strikingly, the SUMOylation-deficient K314R mutant is acetylated at a significantly lower level than wild-type (wt) HIC1 is, suggesting that this lysine is a target for both SUMOylation and acetylation. HDAC4 promotes SUMOylation on this K314 residue of HIC1, whereas SIRT1, a class III HDAC, as well as other class I or II HDACs deacetylates it. In addition, SIRT1 knockdown by RNA interference results in a partial loss of HIC1 SUMOylation. Finally, the mutation of the proline residue in the ψKXEP motif has little effect on HIC1 SUMOylation but severely impairs its acetylation.

Thus, HIC1 transcriptional repression activity is positively controlled by two types of deacetylases, SIRT1 and HDAC4, which increase its deacetylation and SUMOylation, respectively, on the same phylogenetically conserved lysine residue. Since this lysine residue is not contained within a PDSM/SAS motif as it is in the MEF2 transcription factors (63), we thus propose that ψKXEP represents a core motif qualifying as a phosphorylation-independent SUMOylation/acetylation switch. In addition, we have shown for the first time that HIC1 is a target of the class III deacetylase SIRT1 and we have thus identified a new posttranslational modification step in the P53-HIC1-SIRT1 regulatory loop.

MATERIALS AND METHODS

DNA constructs.

The pcDNA3-FLAG-HIC1 expression vector and the 5XHIRE reporter gene have been previously described (16, 42). The pcDNA3 FLAG-HIC1 K314R, E316A, and P317A were derived from the wild-type FLAG-HIC1 vector by PCR mutagenesis. Construction details on the wild-type and K314R, E316A BTB-CR-Gal4 nuclear localization signal-hemagglutinin (HA) chimeras are available upon request. All vectors were verified by nucleotide sequencing.

PIASXα/ARIP3, PIASXβ/Miz1, and PIASγ (40) were kindly provided by Merlin Crossley (University of Sydney, Sydney, New South Wales, Australia). pCMV5-Myc-PIAS1 (45) and pCMV-FLAG-Pc2 (32) were kindly provided by S. H. Lin (Hong Kong University of Science and Technology, Hong Kong, China) and D. Wotton (University of Virginia, Charlottesville, VA), respectively, through the courtesy of Merlin Crossley. The expression vector for the SUMO-specific protease SSP3 (24) was kindly provided by Ron Hay (University of St. Andrews, St. Andrews, Scotland). The expression vectors for wt SIRT1 and its catalytically inactive mutant (H363Y) (34) were kindly provided by Tony Kouzarides (University of Cambridge, United Kingdom), as were expression vectors for Gal4-P300, HA-CBP, and FLAG-PCAF.

Cell culture and transfections.

Cos-7, DAOY, RK13 (rabbit kidney), and U2OS cell lines were maintained in Dulbecco's medium supplemented with 10% fetal calf serum. EcR-CHO clone 6, a stably transfected CHO cell line with inducible FLAG-HIC1 expression under the control of five modified ecdysone response elements, has been previously described (16). FLAG-HIC1 expression was induced by the ecdysone analogue ponasterone A (Invitrogen) at 10 μM for 48 h.

Cells were transfected in OptiMEM (Gibco-BRL) by the polyethyleneimine method, as previously described (16), either in 100-mm dishes (in vivo interaction) with 2.5 μg of DNA or in 12-well plates (repression assay and confocal microscopy) with 500 ng of DNA. Cells were transfected for 6 h and then incubated in fresh complete medium for 48 h before being processed for the relevant assay.

In vitro and in vivo SUMOylation assays.

In vitro SUMOylation assays were carried out as previously described (17, 50) by using 4 μl of a standard reticulocyte lysate (in vitro-translated substrate kit; Promega) per 20-μl reaction mixture.

The in vivo SUMOylation assay was carried out as previously described (50). Briefly, Cos-7 cells were transfected with 1.25 μg of pcDNA3, pcDNA3-FLAG HIC1, or pcDNA3-FLAG HIC1 K314R expression vectors and with 1.25 μg of pSG5-His6-SUMO-1 or empty control vectors using the polyethyleneimine method. At 48 h posttransfection, the cells were washed in ice-cold phosphate-buffered saline (PBS), harvested directly in 1 ml of Gua8 buffer (6 M guanidine-HCl, 100 mM NaCl, 10 mM Tris, 50 mM NaH2PO4 [pH 8.0]), briefly sonicated, and centrifuged. Clarified extracts were incubated for 2 h with 15 μl (packed volume) of Ni-agarose affinity beads (TALON metal affinity resin; Clontech). Bound proteins were washed three times in Gua8 buffer, twice in Urea6.5 buffer (8 M urea, 100 mM NaCl, 50 mM NaH2PO4 [pH 6.5]), and once in cold phosphate-buffered saline before being eluted by boiling in Laemmli loading buffer and separated by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE). To detect FLAG-HIC1 proteins, Western blots were performed with the anti-FLAG M2 monoclonal antibody (Sigma). For the E3 ligase assays with limiting amounts of SUMO-1, 125 ng of pSG5-His6-SUMO-1 and 125 ng of expression vectors for each E3 ligase were used (40).

For the SUMOylation assay in the presence of short interfering RNAs (siRNAs), 293T cells were first transfected in six-well plates with 2.2 pmol of SIRT1 siRNA (Ambion) or of negative control siRNA (Eurogentec) by using 8 μl of Interferin (Polyplus transfection) per well. Twenty-four hours after, cells were transfected for the SUMOylation assay with 900 ng of pcDNA3-FLAG-HIC1 and 900 ng of His-SUMO-1 expression vectors using 5 μl of Exgen 500 (Euromedex) transfection reagent per well. Finally, 24 h after this second transfection, cells were rinsed and harvested in 500 μl of cold PBS. Fifty microliters was directly lysed in Laemmli buffer (Input). After centrifugation, cells were resuspended in 1 ml of Gua8 buffer and the SUMOylation assay was performed as described above.

Immunoprecipitation assays.

Cells were lysed directly in 1 ml of radioimmunoprecipitation assay buffer (RIPA; 150 mM NaCl, 20 mM Tris-HCl [pH 7.4], 1% Triton X-100, 0.1% SDS, 0.5% Na-deoxycholate) containing 20 mM N-ethylmaleimide (NEM; Sigma) to minimize deSUMOylation. After centrifugation, supernatants were incubated for 2 h with 4 μl of the anti-HIC1 (325) polyclonal antibody. Then, protein A-Sepharose beads (Amersham Biosciences) were added for 1 h. The beads were washed first with P1 buffer (RIPA) and then with P2 buffer (RIPA and TNE [100 mM Tris-HCl {pH 7.4}, 0.1 M NaCl, 1 mM EDTA], vol/vol), P3 buffer (TNE with NaCl 0.5 M), and P4 buffer (TNE). Finally, proteins were eluted by boiling in Laemmli loading buffer and separated by SDS-PAGE before Western blotting.

Western blotting and antibodies.

Proteins were fractionated by SDS-PAGE and transferred onto nitrocellulose membranes. After 1 h of blocking in PBSM (PBS with 5% milk), the membranes were incubated overnight at 4°C with specific primary antibodies in PBSTM (PBS with 0.1% Tween 20 and 5% milk) and washed three times for 10 min with PBSN (PBS with 0.1% IGEPAL). The membranes were next incubated for 1 h at room temperature with secondary antibodies coupled to peroxidase (Amersham) (diluted 1/10,000) in PBSM, washed two times for 10 min with PBSN, rinsed with PBS, and revealed with a Western blot chemiluminescence reagent kit (Amersham).

Antibodies against various portions of the C-terminal part of HIC1 (325 and 2563) have been previously described (16). To detect SUMO-1, we used the monoclonal antibody 21C7 (catalog no. 33-2400; Zymed) raised against full-length SUMO-1. Anti-FLAG M2 (catalog no. F3165; Sigma) and anti-HA (Babco) are monoclonal antibodies directed against the epitope tag. The pan anti-acetyl-lysine is a monoclonal antibody (Ac-K-103) from Cell Signaling. To detect SIRT1, we used the monoclonal antibody 2G1/F7 (catalog no. 05-707; Upstate).

The secondary antibodies were anti-rabbit and anti-mouse immunoglobulins and a horseradish peroxidase-linked whole antibody from Amersham.

Confocal microscopy.

Cos-7 cells were cultured on coverslips in 12-well plates and transfected as described above. Twenty-four hours after transfection, they were washed, fixed for 20 min in cold 3% paraformaldehyde, permeabilized in 0.1% Triton X-100 for 5 min, saturated for 30 min in 300 μl of PBS with 10% goat serum, incubated for 30 min with primary antibody diluted 1/500 in PBS with 10% goat serum, and incubated in the dark for 30 min with the secondary antibody diluted 1/200 in PBS with 10% goat serum. Between each stage, they were washed three times for 5 min in PBS. Then, they were placed inverted on a drop of Immuno-Fluore mounting medium on a slide. The slides were stored in the dark at 4°C and visualized with a Leica TCS-NT confocal microscope.

Repression assays.

The repression assays were carried out as previously described with minor modifications (16). Two hundred nanograms of Gal4 plasmid and 250 ng of reporter plasmid were cotransfected. The β-OS lacZ vector (50 ng) was cotransfected in each assay to correct for variations in transfection efficiency. Forty-eight hours after transfection, cells were rinsed in PBS and lysed with Luc assay buffer (25 mM glycylglycine [pH 7.8], 15 mM MgSO4, 4 mM EGTA, 1% Triton X-100). Luciferase and β-galactosidase activities were measured by using beetle luciferin (Promega) and the Galacto-Light kit (Tropix), respectively, with a chemiluminometer (Berthold).

Docking model of the MKHEP peptide in the catalytic domain of SIRT1.

The structures of some low-energy conformers of the MKHEP peptide were obtained using the minimization and random searching tools within the SYBYL software (Tripos, Inc., St. Louis, MO). The Tripos force field was mainly used except for torsional coordinates and electrostatic interactions, for which the procedure described by Meziane-Tani et al. was used (39). In particular, partial atomic charges were calculated using quantum chemical methods based on the density functional theory (B3LYP hybrid functional and the 6-31G* basis set). For that purpose, Jaguar software was used. To derive atomic charges, the molecular electrostatic potential was fitted to a set of point charges located at the atomic centers also reproducing the dipole moment. Among the low-energy structures of MKHEP, the one displaying an extended conformation of the acetylated side chain for the lysine residue was chosen according to structural information given in the 1SZC file of the Protein Data Bank (http://www.rcsb.org). Flexible docking was performed for MKHEP and 1SZC using the GOLD 3.1 software (30). The most stable docking model was selected according to the best-scored structure predicted by the Goldscore function (30).

RESULTS

In vivo SUMOylation of HIC1.

By sequence comparisons of HIC1 proteins from various species, we have previously identified, in the central repression domain of HIC1, a conserved GLDLSKK motif that allows the recruitment of the CtBP corepressor (16). When cloned zebrafish (2) or in silico-deduced Fugu rupibres HIC1 sequences were included in these analyses, a second phylogenetically conserved peptide, YRWM/VKXEP, containing a potential SUMOylation consensus site (ψKXE) became obvious in the central region (Fig. (Fig.11).

To demonstrate that endogenous HIC1 proteins are covalently modified by SUMO in vivo, we first used our previously described CHO cell line with inducible human FLAG-HIC1 expression under the control of ecdysone-responsive elements (16). Upon induction by ponasterone A, two bands can be observed in the inducible HIC1 cell line (Fig. (Fig.2A,2A, lane 3), but not in the parental EcR-CHO cell line (lane 2), after immunoprecipitation with anti-HIC1 antibodies and Western blot analyses with the monoclonal FLAG antibodies. The major band migrated with an apparent molecular mass of ca. 83 kDa, corresponding to HIC1 proteins, whereas the slower-migrating band showed a ca. 14-kDa increase in molecular mass, consistent with the addition of a single SUMO molecule. Indeed, the same bands were observed in Cos-7 cells cotransfected with FLAG-HIC1 and SUMO-1 expression vectors (Fig. (Fig.2A,2A, lane 1). To more directly demonstrate that these slower-migrating HIC1-immunoreactive forms were SUMO-1-modified HIC1 proteins, the same blot was stripped and probed with anti-SUMO-1 monoclonal antibodies. As shown in Fig. Fig.2A2A (lanes 1′ and 3′), the SUMO-1 antibodies recognized the upper band in the transfected Cos-7 cells and in the induced EcR FLAG-HIC1 CHO cells.

FIG. 2.
HIC1 is SUMOylated in vivo. (A) In vivo SUMOylation of HIC1 in a stable inducible cell line. Cos-7 cells were transfected with expression vectors for FLAG-HIC1 and SUMO-1 (lane 1) as a positive control, whereas the parental EcR-CHO cell line (lane 2) ...

Using a similar experimental strategy, we next demonstrated that endogenous HIC1 proteins immunoprecipitated from the medulloblastoma cell line DAOY (35) are also SUMOylated, as shown by the presence of SUMO-1-immunoreactive species in the specific HIC1 immunoprecipitates (Fig. (Fig.2B,2B, lane 2′). Finally, direct lysis of the DAOY cells in loading buffer allowed the detection of HIC1 and its SUMOylated form (Fig. (Fig.2C),2C), as recently described for C/EBPα (47).

Thus, endogenous HIC1 proteins are subject to modification by SUMO-1 in vivo.

The phylogenetically conserved lysine 314 in the HIC1 central region is the target residue for SUMO-1 modification.

To demonstrate that the K314 residue identified in silico is indeed the target residue that is modified by SUMO, we mutated this residue to arginine in the full-length HIC1 protein. The resulting FLAG-HIC1 K314R point mutant and the wild-type FLAG-HIC1 proteins were in vitro translated and labeled with [35S]methionine in reticulocytes lysates. Then, their ability to undergo SUMO-1 modification was analyzed in an in vitro modification assay in the presence or absence of recombinant SUMO-1 with a fraction of HeLa cells as a source of E1 activity and purified Ubc9 (E2) as previously described (17, 50). A slower-migrating band was obtained with only the wild-type HIC1 protein (Fig. (Fig.3A,3A, compare lanes 2 and 5).

FIG. 3.
The evolutionarily conserved lysine 314 is the major SUMO-1 modification site in HIC1 in vitro and in vivo. (A) In vitro-translated, [35S]-labeled, full-length FLAG-HIC1 (lanes 1 to 3) or FLAG-HIC1 K314R (lanes 4 to 6) was analyzed by SDS-PAGE directly ...

To confirm SUMO-1 modification of HIC1 on K314 in vivo, Cos-7 cells were transfected by either the wild-type FLAG-HIC1 or the FLAG-HIC1 K314R point mutant with or without expression vectors for His-tagged SUMO-1 (His-SUMO-1). His-SUMO-1-modified proteins were affinity purified on nickel-agarose beads and analyzed by Western blotting. The transfection of FLAG-HIC1 or FLAG-HIC1 K314R alone resulted in the pull down of a significant amount of HIC1 and HIC1 K314R proteins (Fig. (Fig.3B,3B, lanes 9 and 11), most likely due to interactions via the C2H2 zinc fingers, as previously described for Tramtrack (36). In the His-SUMO-1-transfected cells, a single SUMOylated band was observed with the wt HIC1 protein (Fig. (Fig.3B,3B, lane 10). Consistent with the in vitro results, the FLAG-HIC1 K314R point mutant could no longer be modified by SUMO-1 (Fig. (Fig.3B,3B, lane 12). As a further confirmation, a point mutation of the conserved glutamic acid residue in the ψKXE consensus, E316A, also abolishes the SUMOylation of HIC1 at K314 (Fig. (Fig.3C,3C, lane 4). Similarly, the overexpression of SENP2/SSP3 deSUMOylase also results in the disappearance of the slower-migrating band, demonstrating that this species corresponds to a SUMOylated form of HIC1 (Fig. (Fig.3D,3D, lane 2).

Thus, the phylogenetically conserved lysine 314 in the central repression domain is the target for the in vivo SUMO-1 modification of HIC1.

PIAS family members act as E3 ligases for HIC1.

Proteins the PIAS family of have been widely reported to function as E3 ligases in the SUMO-1 conjugation pathway (48). To determine whether PIAS proteins could act as specific E3 ligases for HIC1 in vivo, we used the Cos-7 overexpression system in the presence of a limiting amount (1/10 of the amount used in standard assays) (Fig. (Fig.3B)3B) of the His-tagged SUMO-1 expression vector (40). Under these experimental conditions, the SUMOylated form of HIC1 became almost undetectable (Fig. (Fig.4A,4A, compare lanes 1 and 2). However, upon the cotransfection of PIAS expression vectors, PIAS1 or PIASXα/ARIP3, but not PIASγ or PIASXβ/MIZ1 (Fig. (Fig.4A,4A, lanes 3 to 6), restored HIC1 SUMOylation.

FIG. 4.
PIAS family members (but not Pc2) are E3 ligases for HIC1 in vivo. (A) Cos-7 cells were transfected with 1.25 μg of FLAG-HIC1 and 1.25 μg His-SUMO-1 expression vectors (lane 1) or 125 ng of His-SUMO-1 expression vector alone (lane 2) or ...

Since the HIC1 central region contains a CtBP binding site close to the SUMOylation site, we also investigated whether Pc2, the E3 ligase for CtBP (32), could also modulate the SUMOylation of HIC1 in vivo. For this investigation, we used immunoprecipitation instead of Ni affinity pulldown since Pc2 contains a polyhistidine stretch and could thus bind to Ni beads. As shown in Fig. Fig.4B,4B, lane 4, Pc2 does not induce the SUMOylation of HIC1. In addition, the deletion of the GLDLSKK motif of HIC1 responsible for the recruitment of CtBP (16) or the overexpression of CtBP1 or CtBP2 has no effect on the SUMOylation of HIC1 in vivo (data not shown).

These results demonstrate that some PIAS proteins, but not Pc2, stimulate the SUMOylation of HIC1 in vivo.

SUMOylation of lysine 314 is not required for the punctate nuclear localization of HIC1.

Since SUMOylation has been implicated in the subnuclear localization of the targeted protein (10), we analyzed HIC1 localization by immunofluorescence confocal microscopy. In transiently transfected Cos-7 cells, wt FLAG-HIC1 and FLAG-HIC1 K314R proteins localized to similar punctate nuclear structures, typical of many overexpressed BTB/POZ proteins (Fig. (Fig.5A)5A) (16, 18). Thus, the mutation of the key lysine residue in the SUMOylation site does not significantly alter subcellular localization of HIC1. We next coexpressed FLAG-HIC1 or FLAG HIC1K314R with SUMO-1 and performed double-labeling experiments with the rabbit anti-HIC1 (325) polyclonal antibodies and the mouse anti-SUMO-1 monoclonal antibody. All of the nuclear dots containing the ectopically expressed wild-type FLAG-HIC1 proteins were also labeled by the SUMO-1 antibodies, indicating that they contain SUMOylated HIC1 proteins (33). Consistent with the in vivo labeling assay (Fig. (Fig.3B),3B), the expression of the FLAG-HIC1 K314R mutant did not result in the recruitment of SUMO-1 to the HIC1 nuclear dots (Fig. (Fig.5B,5B, bottom panels).

FIG. 5.
wt FLAG-HIC1 but not FLAG-HIC1 K314R colocalizes with SUMO-1 on nuclear dots in transfected Cos-7 cells. (A) The mutation of the SUMOylation consensus does not impinge on the subnuclear localization of ectopically expressed HIC1 proteins. Cos-7 cells ...

Thus, these results confirm that lysine 314 is the target for in vivo SUMOylation of HIC1 and further demonstrate that this modification did not significantly modify the subnuclear localization of HIC1, at least not in transient transfection experiments.

SUMOylation of lysine 314 positively regulates the transcriptional repression activity of HIC1.

The SUMOylation site is located in the central region which is an autonomous repression domain (16). To assess the functional consequences of the SUMO-1 modification of HIC1, we next compared the transcriptional repression potentials of wild-type and SUMOylation-deficient K314R HIC1 proteins. As a first approach, we decided to use the Gal-4 fusion assay, which has been widely used to analyze the effects of SUMOylation on various transcription factors. However, preliminary results demonstrated that a chimera containing the HIC1 central region (residues 144 to 422 of the human HIC1 protein) fused downstream of the Gal-4 DNA binding domain (Gal4-CR) was not SUMOylated, although it was nuclear and able to efficiently interact with CtBP (data not shown) (16). These results indicated that a heterologous dimerization domain, namely, the Gal-4 DNA binding domain, could not fully substitute the BTB/POZ domain for SUMOylation of HIC1, whereas it can for the recruitment of CtBP (16).

We thus constructed two other chimeras, BTB-CR-G4 and BTB-CR-G4 K314R, in which the HIC1 BTB/POZ domain and the central region (CR), corresponding to residues 1 to 422 of the human HIC1 protein, are fused to a C-terminal Gal-4 DNA binding domain (DBD), thus mirroring the structure of the wild-type HIC1 protein with its C-terminal Krüppel-like C2H2 zinc fingers (Fig. (Fig.11 and and6A).6A). In vivo SUMOylation assays with Cos-7 cells demonstrated that the wild-type BTB-CR-Gal4 chimera was efficiently SUMOylated in contrast to the K314R point mutant (Fig. (Fig.6B,6B, lanes 5 and 6).

FIG. 6.
SUMO-1 modification modulates HIC1-mediated transcriptional repression. (A) Schematic structure of the two HIC1 BTB-CR-Gal4 chimeras. Numbers refer to human HIC1 residues. The BTB/POZ domain, the CR, and the Gal4 DBD domain are represented as dotted, ...

RK13 cells were then cotransfected with a reporter plasmid pG5-luc (containing five copies of a Gal4 binding site in front of a luciferase gene driven by a minimal herpes simplex virus thymidine kinase [tk] promoter: 5XGal4-tk-luc) and either the empty Gal4 vector or the vectors carrying the wild-type or the mutant K314R BTB-CR-Gal4 chimeras. Both chimeras repressed luciferase gene expression relative to the basal level obtained with the Gal4 DBD alone, G4. However, the K314R chimera repressed transcription almost 1.8-fold less efficiently than did the wild-type HIC1 chimera (Fig. (Fig.6C).6C). The same effect was obtained, albeit to a lesser extent, when the repression of the full-length HIC1 proteins, either the wild type or the K314R point mutant, was assessed in U2OS cells on an SV40-luc reporter gene containing five copies of the HIC1 responsive element (5XHiRE) (42) (data not shown). Similarly, a BTB-CR-G4 E316A chimera was not SUMOylated in vivo in Cos-7 cells (data not shown) and was severely impaired in its repression potential (Fig. (Fig.6D).6D). Since K314R or E316A point mutations are sufficient to abolish the in vivo SUMOylation of HIC1 (Fig. 3B and C), these results strongly support the idea that this covalent modification in the central region of HIC1 positively regulates its transcriptional repression potential.

In keeping with these results, the cotransfection of an expression vector for SENP2/SSP3 (24) which is able to deSUMOylate HIC1 (Fig. (Fig.3D)3D) has no significant effects on the basal level obtained with the Gal4 DBD alone but induces a decrease of two-thirds in the transcriptional repression mediated by the wt BTB-CR-Gal4 chimera (Fig. (Fig.6E).6E). Interestingly, the effect of SSP3 on the wt chimera is more pronounced than that induced by the K314R mutation (Fig. (Fig.6C).6C). Furthermore, the overexpression of SENP2/SSP3 also affects the transcriptional repression induced by the non-SUMOylable K314R chimeras (Fig. (Fig.6E).6E). An explanation could be that SENP2/SSP3 induces not only the deconjugation of HIC1 itself but also that of other SUMOylated proteins involved in HIC1-mediated repression.

Taken together, these results demonstrate that the SUMOylation of the central region of HIC1 positively regulates its transcriptional repression potential.

Lysine 314 is also subject to reversible acetylation.

Several covalent posttranslational modifications, such as ubiquitination, SUMOylation, methylation, and acetylation, can target lysine residues. Recently, the transcriptional activity of some transcription factors, such as Sp3, MEF2, and PLAG, has been shown to be regulated by reversible acetylation or SUMOylation targeting the same lysine residue (6). To test whether this holds also true for HIC1 lysine 314, we first asked whether HIC1 could be acetylated by histone acetyltransferases. As shown in Fig. Fig.7A,7A, coexpression in Cos-7 cells of the histone acetyltransferase CBP or p300 can induce HIC1 acetylation (Fig. (Fig.7A,7A, lanes 2 to 4), whereas PCAF is unable to do so (lane 5). Relative to wild-type HIC1, the K314R mutant is less acetylated by p300 in the presence of general HDAC inhibitors, TSA (an inhibitor of the class I and II HDACs), and nicotinamide (an inhibitor of the class III NAD+-dependent Sir2 family deacetylases), either alone (data not shown) or in combination (Fig. (Fig.7B,7B, compare lanes 4 and 6). Collectively, these data suggest that HIC1 is acetylated on several lysine residues and, further, that the conserved lysine 314, which is the only SUMO acceptor site, can also be modified by acetylation.

FIG. 7.
HIC1 is acetylated on several lysine residues, including K314. (A) HIC1 was expressed alone (lane 2) or with the indicated acetyltransferase (lanes 3 to 5) by transient transfection in Cos-7 cells. Cells were treated with 300 nM TSA for 24 h before lysis ...

Lysine 314 is deacetylated by SIRT1 but not by HDAC4.

HIC1 interacts with the deacetylase SIRT1, and this complex binds to and represses the transcription of the SIRT1 promoter (12). SIRT1 is also known to deacetylate several substrates, including P53 (34) and MEF2D (27). In the latter case, a lysine subject to reversible acetylation catalyzed by CBP and SIRT1 is also SUMOylated. Moreover, this SUMOylation is enhanced by HDAC4 by a still poorly understood mechanism (26, 27, 63, 64). These observations led us first to investigate whether HIC1 can interact with SIRT1 and/or HDAC4 and whether this interaction can be regulated by SUMOylation. Using transient transfection assays with Cos-7 cells, we first demonstrated that wt HIC1 interacts with both SIRT1 and HDAC4 (Fig. 8A and B, lanes 4), albeit in a SUMO-independent manner, since the wt and the K314R mutant proteins interact equally well with SIRT1 and HDAC4 (Fig. 8A and B, lanes 6).

FIG. 8.
wt HIC1 and the K314R mutant interact with SIRT1 and HDAC4. (A) Cos-7 cells were mock transfected (lane 1) or transfected with expression vectors for FLAG-tagged HIC1 proteins (1.25 μg; wild type or K314R mutant) and SIRT1 (1.25 μg) (lanes ...

As to the functional consequence of these interactions, ectopic expression of SIRT1 dramatically reduces the p300-induced acetylation of HIC1 (Fig. (Fig.9A,9A, lanes 2 and 3), whereas the overexpression of an inactive SIRT1 enzyme (SIRT1 H363Y) (34) only slightly affects HIC1 acetylation (Fig. (Fig.9A,9A, lanes 2 and 4). By contrast, the overexpression of HDAC4 has no effect on the acetylation level of HIC1 (Fig. (Fig.9B,9B, lanes 2 and 3).

FIG. 9.
Class III deacetylase SIRT1 has HIC1 deacetylase activity but not class II deacetylase HDAC4. (A) Cos-7 cells were transfected with expression vectors for FLAG-tagged HIC1 proteins (1.25 μg) in the absence (lane 1) or presence (+; lanes ...

These experimental data showing that HIC1 is a target of the deacetylase SIRT1 were confirmed by a structural analysis using the Sir2-NAD-H4 complex, Saccharomyces cerevisiae (Protein Data Bank code 1SZC), as a template for the catalytic domain of human SIRT1. In that model, the acetylated lysine side chain of the HIC1 MKHEP peptide appears to be close to the H135 catalytic residue and a little bit far from F79. It is also close to the important residues D43, F44, and R45. In particular, the model displays a strong hydrogen bond between the carbonyl oxygen atom of the acetylated lysine side chain of MKHEP and the 3′ hydroxyl group of the ribose attached to the nicotinamide moiety of NAD (Fig. (Fig.9C9C).

Thus, these results demonstrate that while HIC1 interacts with both HDAC4 and SIRT1, only the latter functions as a HIC1 deacetylase.

HDAC4 upregulates the SUMOylation of HIC1 on lysine 314.

In MEF2D, the same lysine residue is a target for reversible acetylation mediated by CBP and SIRT1 and for SUMOylation potentiated by HDAC4 (27). Our results thus prompted us to investigate whether HDAC4 was also able to upregulate the SUMOylation of HIC1 on lysine 314. To that end, we used the Cos-7 overexpression system in the presence of a limiting amount of the His-tagged SUMO-1 (34) in conjunction with a bona fide E3 ligase for HIC1, PIAS1 (Fig. (Fig.4A),4A), or HDAC4. Under these conditions, PIAS1 (but not PIASγ, used as a relevant negative control) (Fig. (Fig.4A)4A) and, more importantly, HDAC4 restored HIC1 SUMOylation (Fig. 10A, lanes 3 to 5). Thus, HDAC4 is able to promote SUMOylation on HIC1 lysine 314, which is also a target for acetylation by CBP (Fig. (Fig.7)7) and deacetylation by SIRT1 (Fig. (Fig.9).9). To confirm this potential interplay between these two lysine-targeting modifications, we hypothesized that the levels of HIC1 SUMOylation and acetylation would be inversely correlated. Indeed, the SUMOylation of HIC1 K314 is increased when an active SIRT1 enzyme is overexpressed (Fig. 10B, lane 2) and decreased in the presence of its specific inhibitor nicotinamide (NIA) (Fig. 10B, lane 5). Interestingly, although HDAC4 is not able to deacetylate HIC1 and, hence, K314 (Fig. (Fig.9B,9B, lane 3), TSA significantly decreased the SUMOylation level of HIC1 on K314 (Fig. 10B, lane 4), indicating that other class I and/or II HDACs can also directly or indirectly regulate the SUMOylation of HIC1. To confirm that SIRT1 directly controls HIC1 SUMOylation, we performed a SUMOylation test with HEK293T cells transfected beforehand with SIRT1 siRNA or control siRNA. SIRT1 siRNA transfection caused the efficient knockdown of SIRT1 and affected levels of SUMOylated HIC1 proteins (Fig. 10C). This effect is specific since the SUMOylation of RanGAP1, the most abundant SUMO-1 substrate (4), is unaffected (Fig. 10C, bottom panel).

FIG. 10.
SUMOylation of HIC1 K314 is enhanced by HDAC4 and SIRT1 but through different mechanisms. (A) HDAC4 stimulates SUMO-1 conjugation to HIC1 lysine 314. Cos-7 cells were transfected with 1.25 μg of FLAG-HIC1 and 1.25 μg His-SUMO-1 expression ...

Thus, as in the case of MEF2D, SIRT1 and HDAC4 are involved in the SUMOylation and deacetylation of the same lysine residue of HIC1.

The ψKXEP motif is a proline-dependent SUMOylation-acetylation switch motif.

In striking contrast with MEF2, HIC1 lysine 314 is not embedded in a phosphorylation-dependent SUMOylation motif or SUMOylation-acetylation switch ψKXEXXSP motif but in a ψKXEP motif, with the proline conserved from human to zebrafish (Fig. (Fig.1).1). Since the G/SKXXP motif is a consensus motif for acetylation (44) and since half of the SAS (51) or PDSM (63) motifs contain a proline residue adjacent to the glutamic acid residue, our results emphasize the potential role of this proline residue in the acetylation/SUMOylation switch (Fig. (Fig.11).11). To functionally address this role, proline 317 in the ψKXEP motif of the full-length HIC1 protein was mutated into alanine. The resulting FLAG-HIC1 P317A mutant was compared with wt FLAG-HIC1 for its ability to undergo SUMO-1 modification and acetylation. The SUMOylation of the FLAG-HIC1 P317A was not significantly affected (Fig. 12A, lanes 2 and 4). By contrast, in the presence of P300 and HDAC inhibitors, the acetylation level of this mutant is significantly decreased (Fig. 12B, lanes 4 and 6).

FIG. 11.
KXEP: a core motif for coordinated acetylation and SUMOylation? Consensus motifs for which modifications have been fully validated by functional assays (44, 51, 63) are indicated. We noticed that in all known examples of coordinated SUMOylation and acetylation, ...
FIG. 12.
Mutation of proline 317 in the KXEP motif affects the acetylation but not the SUMOylation of HIC1. (A) The SUMOylation of the wt HIC1 protein or of the P317A mutant (a point mutation of the conserved proline residue in the HIC1 ψKXEP SUMOylation ...

Thus, our results demonstrate that the ψKXEP motif is an acetylation/SUMOylation switch motif.

DISCUSSION

Here, we show the dynamic modification of HIC1 by SUMO on a single phylogenetically conserved lysine residue, K314, located in the HIC1 central region which, in addition to the BTB/POZ domain, is a second autonomous repression domain. Furthermore, this lysine is also targeted by acetylation by CBP/p300. We have identified the deacetylases SIRT1 and HDAC4 as regulators of the interplay between these two posttranscriptional modifications that affect the transcriptional repression activity of HIC1.

In contrast with the BTB/POZ domain, the zinc fingers and, to a lesser extent, the carboxyl-terminal end, the majority of the HIC1 central region has not been subjected to a strong selection pressure, except for some small peptidic motifs which are perfectly conserved from zebra fish to human. One of these, the GLDLSKK peptide, has been previously identified as a variant of the PXDLS consensus motif required for the interaction with the CtBP corepressor (16). In this study, we have shown that a second motif, M/VKXEP, is a consensus SUMOylation site. Based on these results, we predict that the others also contribute to transcriptional repression in various ways, including the recruitment of HIC1 partners, the stabilization of the protein, and/or the modulation of DNA binding activities. Notably, these motifs, T163PVI, ELY200A, and S381EETGSSE, are centered on residues potentially subject to another regulatory posttranscriptional modification, phosphorylation.

Using the RONN prediction program (http://www.strubi.ox.ac.uk), the HIC1 central region appears to be natively disordered except for the CtBP interaction motif. However, this region is inserted between two highly structured regions which could impose a strict folding to this otherwise disordered region. Previously, we have shown that the dimerization of the BTB/POZ domain was essential to the creation of an interface for the optimal binding of CtBP though the GLDLS motif but could be replaced by a heterologous dimerization domain, such as the Gal-4 DNA binding domain (16). For the SUMOylation of HIC1, the situation is more complex since a Gal-4-central region chimera is not SUMOylated (data not shown), although it still interacts with CtBP. Whereas the BTB/POZ domain is not directly involved in the recruitment of CtBP, it could be directly involved in the SUMOylation of HIC1. Indeed, yeast two-hybrid screens with the BTB/POZ domain identified the unique E2-SUMO-conjugating enzyme Ubc9 as an HIC1 partner (S. Deltour, unpublished results). In addition, the HIC1 BTB/POZ contains a potential SUMO binding motif, V/I-X-V/I-V/I or V/I-V/I-X-V/I (52). However, crystallographic studies indicate that this motif, which is also found in BCL6, is buried in its structure, strongly suggesting that it may not be functional (G. Privé, personal communication) (54).

SUMOylation has been shown to regulate protein function in different manners. First, it can affect subcellular localization of several proteins, such as, for example, PML in the so-called nuclear bodies. In transient transfection assays, HIC1, like many BTB/POZ proteins, displays a punctate nuclear localization and the endogenous HIC1 proteins have been recently associated with “HIC1 bodies” in medulloblastoma DAOY cells or in WI38 primary human fibroblasts (55). Mutant K314R also displays a punctate nuclear localization, indicating that SUMOylation might not be essential for this localization. Second, SUMOylation may regulate protein-protein interactions. For example, the SUMOylation of two Drosophila BTB/POZ proteins, Mod(mdg4)2.2 and CP190, two components of the gypsy chromatin insulator, does not affect their ability to bind to chromatin per se but instead inhibits long-range interactions of insulator proteins and hence the establishment of chromatin domains (8). The majority of known SUMO substrates are transcription factors or cofactors, and in most cases, this modification is associated with transcriptional repression (22, 57). For example, SUMOylation negatively regulates the activity of transcriptional activators, such as Sp3, p300, c-jun, c-myb, or Elk1. In that last case, SUMOylation results in the recruitment of HDAC2 to promoters and, hence, transcriptional repression of Elk1 target genes (62). The SUMOylation of transcriptional repressors is also required for their optimal silencing activity as shown for BKLF (40). In the case of HIC1, SUMOylation likely plays both direct and indirect roles in repression. A direct role is inferred from the mutation of the HIC1 SUMOylation site which reduces transcriptional repression, whereas an indirect role is supported by the similar effects exhibited by the deSUMOylase SENP2/SSP3 on the repression mediated by wt HIC1 and the non-SUMOylable mutant, suggesting that SUMOylation affects HIC1 partners involved in repression (Fig. (Fig.6E).6E). SUMOylation could thus play a transient role in the formation of multiprotein complexes also involved in HIC1-mediated repression, as recently shown also for human p66α and p66β, which are components of the Mi-2/NuRD complex associated with HDACs (25). Notably, the HIC1 central region containing the SUMOylation site is an HDAC-dependent repression domain (16, 53).

Finally, lysine can be targeted by multiple modifications in a competitive and regulated way. The first example has been reported for the NF-κB inhibitor IκBα, whose SUMOylation blocks polyubiquitination on the same residue and, thus, subsequent ubiquitin-dependent proteasomal degradation (17). Recently, several studies have demonstrated that the transcriptional activity of MEF2D is regulated by interplay between acetylation and SUMOylation on the same residue (27). This regulation by different lysine modifications is mediated by two distinct deacetylases belonging to different classes, namely, the class II, TSA-sensitive HDAC4 and the class III, NAD+-dependent deacetylase SIRT1. These two deacetylases interact and form a dual-deacetylase complex with restricted and complementary properties since SIRT1, but not HDAC4, can deacetylate MEF2 and in which HDAC4 enhances MEF2 SUMOylation by a still poorly understood mechanism relying on an N-terminal coiled-coil domain (64). Here, we have shown that the transcriptional repressor HIC1 is also a target for this dual-deacetylase complex which can thus regulate, via similar mechanisms, two different families of transcription factors endowed with clearly different functional properties, activation and repression. In MEF2, this acetylation/SUMOylation motif is an extended version of the canonical ψKXE motif, ψKXEXXSP. Bioinformatics analysis has shown this motif to be conserved in numerous proteins, with most involved in transcriptional regulation (29, 63). Furthermore, functional assays with MEF2, GATA-1, and heat shock factors have demonstrated that phosphorylation of the adjacent serine residue regulates the SUMOylation of the lysine, hence the names phosphorylation-dependent SUMOylation motif and SUMOylation-acetylation switch (26). Within PDSM motifs, the spacing between the SUMO target and phosphorylation site appears to be crucial since searches using alternate spacing did not reveal SUMO substrates or transcriptional regulators (29). HIC1, which is also SUMOylated and acetylated on the same residue, lacks the adjacent SP motif found in MEF2 factors (Fig. (Fig.1).1). Strikingly, however, HIC1 shares with MEF2 a conserved proline residue adjacent to the glutamic acid residue (KXEP versus KXEPXSP) (Fig. (Fig.11).11). A G/SKXXP motif has been shown to be a consensus site for the acetylation of the Brm proteins by CBP/p300 (44) (5). These data suggest that the spectrum of modifications potentially affecting a given lysine residue could be defined by the presence/absence of several adjacent residues, including a glutamic acid, a proline, and/or a serine-proline motif (Fig. (Fig.11).11). This would give rise to (i) a classical SUMOylation-only ψKXE motif (ii) a ψKXEXXSP motif whose lysine could be targeted by a phosphorylation-dependent SUMOylation, as shown for GATA-1 (29, 63), (iii) a ψKXEP motif whose lysine could be targeted by SUMOylation or acetylation as shown for NF-IL-6β, also known as C/EBPδ (60) and HIC1 (this study), and (iv) a ψKXEPXSP motif whose lysine could be targeted by acetylation and phosphorylation-dependent SUMOylation (26). Interestingly, a ψKXEP motif is also found in Evi-1 and conserved in the C/EPBα, -β, and -γ proteins (57), but in these cases, the acetylation of this particular lysine has not been demonstrated. Sp3 is another well-documented example for SUMOylation/acetylation on the same residue, but here the motif contains three glutamic acid residues, KEEEP (6, 43, 46). Interestingly, half of the described SAS motifs contain an adjacent proline residue (51).

Recently, it has been shown that HIC1 forms a transcriptional repression complex with SIRT1 and that this complex directly binds the SIRT1 promoter to repress its transcription (12). The HIC1 BTB/POZ domain interacts with SIRT1 (12), but the repression mediated by this isolated domain in the context of a Gal4 chimera is inhibited neither by TSA, a specific inhibitor of class I and class II HDACs (14, 15), nor by NIA, an inhibitor of NAD+-dependent class III deacetylases, such as SIRT1 (our unpublished results). Our results thus provide the first mechanistic clues to this HIC1/SIRT1 interaction. SIRT1 is not involved in the repression mediated by the isolated BTB/POZ domain, but by deacetylating HIC1, SIRT1 favors its SUMOylation and, thus, the establishment of optimal transcriptional repression. The acetyl-peptide binding specificity among sirtuins remains unclear (44) and SIRT1 was reported to have no substrate specificity in vitro (3). In vivo, SIRT1 can deacetylate a still-growing list of proteins with a great versatility in the amino acid sequence surrounding the acetylated lysine, such as RHKK(Ac)382L for p53 (56, 37), RK(AC)360LKK for androgen receptor (20), KYKK(Ac)379 for the BTB/POZ transcriptional repressor BCL6 (1), and YWMK(Ac)314HEP for HIC1. BCL6 acetylation decreases its transcriptional repression activity (1) by impairing its physical association with MTA-3, a cell-type-specific subunit of the Mi-2/NURD corepressor complex (21). The interaction between CtBP and some corepressors can also be negatively regulated by lysine acetylation of the PXDLSXK motif, as shown for E1A and RIP140 (58). The HIC1 central region also interacts with CtBP through a conserved GLDLSKK motif which could be targeted by acetylation. In that case, acetylation and deacetylation by SIRT1 and/or TSA-sensitive HDACs would be global regulators of the transcriptional repression mediated by the central region by modulating its SUMOylation and the recruitment of CtBP.

Deciphering the complex interplay between different modifications and their influence on the recruitment of repression complexes will be necessary to better understand the repression mechanisms brought about by HIC1 on its target genes and hence its tumor-suppressive properties.

Acknowledgments

We thank M. Crossley, R. Hay, T. Kouzarides, S. H. Lin and D. Wotton for their generous gifts of reagents. We are indebted to Bénédicte Grasland and Aurélie Bauduin for the construction and the initial analyses of the Gal4-HIC1 and HIC1-Gal4 chimeras. We thank Brian Rood and Alexis Verger for comments and critical reading of the manuscript and Gil Privé for helpful discussions on the BTB/POZ structure.

This work was supported by funds from the CNRS, the Pasteur Institute, the Ligue Nationale contre le Cancer (Comité Interrégional du Septentrion), the EEC “Rubicon,” and the Association pour la Recherche contre le Cancer (to J.S. and D.L.). N. Stankovic-Valentin was supported by a fellowship from the Ministère de la Recherche et de la Technologie and by the Association pour la Recherche contre le Cancer.

Footnotes

[down-pointing small open triangle]Published ahead of print on 5 February 2007.

REFERENCES

1. Bereshchenko, O. R., W. Gu, and R. Dalla-Favera. 2002. Acetylation inactivates the transcriptional repressor BCL6. Nat. Genet. 32:606-613. [PubMed]
2. Bertrand, S., S. Pinte, N. Stankovic-Valentin, S. Deltour-Balerdi, C. Guerardel, A. Begue, V. Laudet, and D. Leprince. 2004. Identification and developmental expression of the zebrafish orthologue of the tumor suppressor gene HIC1. Biochim. Biophys. Acta 1678:57-66. [PubMed]
3. Blander, G., J. Olejnik, E. Krzymanska-Olejnik, T. McDonagh, M. Haigis, M. B. Yaffe, and L. Guarente. 2005. SIRT1 shows no substrate specificity in vitro. J. Biol. Chem. 280:9780-9785. [PubMed]
4. Bossis, G., and F. Melchior. 2006. Regulation of SUMOylation by reversible oxidation of SUMO conjugating enzymes. Mol. Cell 21:349-357. [PubMed]
5. Bourachot, B., M. Yaniv, and C. Muchardt. 2003. Growth inhibition by the mammalian SWI-SNF subunit Brm is regulated by acetylation. EMBO J. 22:6505-6515. [PubMed]
6. Braun, H., R. Koop, A. Ertmer, S. Nacht, and G. Suske. 2001. Transcription factor Sp3 is regulated by acetylation. Nucleic Acids Res. 29:4994-5000. [PMC free article] [PubMed]
7. Britschgi, C., M. Rizzi, T. J. Grob, M. P. Tschan, B. Hugli, V. A. Reddy, A. C. Andres, B. E. Torbett, A. Tobler, and M. F. Fey. 2006. Identification of the p53 family-responsive element in the promoter region of the tumor suppressor gene hypermethylated in cancer 1. Oncogene 25:2030-2039. [PubMed]
8. Capelson, M., and V. G. Corces. 2006. SUMO conjugation attenuates the activity of the gypsy chromatin insulator. EMBO J. 25:1906-1914. [PubMed]
9. Carter, M. G., M. A. Johns, X. Zeng, L. Zhou, M. C. Zink, J. L. Mankowski, D. M. Donovan, and S. B. Baylin. 2000. Mice deficient in the candidate tumor suppressor gene Hic1 exhibit developmental defects of structures affected in the Miller-Dieker syndrome. Hum. Mol. Genet. 9:413-419. [PubMed]
10. Chakrabarti, S. R., R. Sood, S. Nandi, and G. Nucifora. 2000. Posttranslational modification of TEL and TEL/AML1 by SUMO-1 and cell-cycle-dependent assembly into nuclear bodies. Proc. Natl. Acad. Sci. USA 97:13281-13285. [PubMed]
11. Chen, W., T. K. Cooper, C. A. Zahnow, M. Overholtzer, Z. Zhao, M. Ladanyi, J. E. Karp, N. Gokgoz, J. S. Wunder, I. L. Andrulis, A. J. Levine, J. L. Mankowski, and S. B. Baylin. 2004. Epigenetic and genetic loss of Hic1 function accentuates the role of p53 in tumorigenesis. Cancer Cells 6:387-398.
12. Chen, W. Y., D. H. Wang, R. C. Yen, J. Luo, W. Gu, and S. B. Baylin. 2005. Tumor suppressor HIC1 directly regulates SIRT1 to modulate p53-dependent DNA-damage responses. Cell 123:437-448. [PubMed]
13. Chen, W. Y., X. Zeng, M. G. Carter, C. N. Morrell, R. W. Chiu-Yen, M. Esteller, D. N. Watkins, J. G. Herman, J. L. Mankowski, and S. B. Baylin. 2003. Heterozygous disruption of Hic1 predisposes mice to a gender-dependent spectrum of malignant tumors. Nat. Genet. 33:197-202. [PubMed]
14. Deltour, S., C. Guerardel, and D. Leprince. 1999. Recruitment of SMRT/N-CoR-mSin3A-HDAC-repressing complexes is not a general mechanism for BTB/POZ transcriptional repressors: the case of HIC-1 and γFBP-B. Proc. Natl. Acad. Sci. USA 96:14831-14836. [PubMed]
15. Deltour, S., S. Pinte, C. Guerardel, and D. Leprince. 2001. Characterization of HRG22, a human homologue of the putative tumor suppressor gene HIC1. Biochem. Biophys. Res. Commun. 287:427-434. [PubMed]
16. Deltour, S., S. Pinte, C. Guerardel, B. Wasylyk, and D. Leprince. 2002. The human candidate tumor suppressor gene HIC1 recruits CtBP through a degenerate GLDLSKK motif. Mol. Cell. Biol. 22:4890-4901. [PMC free article] [PubMed]
17. Desterro, J. M., M. S. Rodriguez, and R. T. Hay. 1998. SUMO-1 modification of IκBα inhibits NF-κB activation. Mol. Cell 2:233-239. [PubMed]
18. Dhordain, P., O. Albagli, S. Ansieau, M. H. Koken, C. Deweindt, S. Quief, D. Lantoine, A. Leutz, J. P. Kerckaert, and D. Leprince. 1995. The BTB/POZ domain targets the LAZ3/BCL6 oncoprotein to nuclear dots and mediates homomerisation in vivo. Oncogene 11:2689-2697. [PubMed]
19. Dhordain, P., O. Albagli, R. J. Lin, S. Ansieau, S. Quief, A. Leutz, J. P. Kerckaert, R. M. Evans, and D. Leprince. 1997. Corepressor SMRT binds the BTB/POZ repressing domain of the LAZ3/BCL6 oncoprotein. Proc. Natl. Acad. Sci. USA 94:10762-10767. [PubMed]
20. Fu, M., M. Liu, A. A. Sauve, X. Jiao, X. Zhang, X. Wu, M. J. Powell, T. Yang, W. Gu, M. L. Avantaggiati, N. Pattabiraman, T. G. Pestell, F. Wang, A. A. Quong, C. Wang, and R. G. Pestell. 2006. Hormonal control of androgen receptor function through SIRT1. Mol. Cell. Biol. 26:8122-8135. [PMC free article] [PubMed]
21. Fujita, N., D. L. Jaye, C. Geigerman, A. Akyildiz, M. R. Mooney, J. M. Boss, and P. A. Wade. 2004. MTA3 and the Mi-2/NuRD complex regulate cell fate during B lymphocyte differentiation. Cell 119:75-86. [PubMed]
22. Gill, G. 2005. Something about SUMO inhibits transcription. Curr. Opin. Genet. Dev. 15:536-541. [PubMed]
23. Gill, G. 2004. SUMO and ubiquitin in the nucleus: different functions, similar mechanisms? Genes Dev. 18:2046-2059. [PubMed]
24. Girdwood, D., D. Bumpass, O. A. Vaughan, A. Thain, L. A. Anderson, A. W. Snowden, E. Garcia-Wilson, N. D. Perkins, and R. T. Hay. 2003. P300 transcriptional repression is mediated by SUMO modification. Mol. Cell 11:1043-1054. [PubMed]
25. Gong, Z., M. Brackertz, and R. Renkawitz. 2006. SUMO modification enhances p66-mediated transcriptional repression of the Mi-2/NuRD complex. Mol. Cell. Biol. 26:4519-4528. [PMC free article] [PubMed]
26. Grégoire, S., A. M. Tremblay, L. Xiao, Q. Yang, K. Ma, J. Nie, Z. Mao, Z. Wu, V. Giguere, and X. J. Yang. 2006. Control of MEF2 transcriptional activity by coordinated phosphorylation and sumoylation. J. Biol. Chem. 281:4423-4433. [PubMed]
27. Grégoire, S., and X. J. Yang. 2005. Association with class IIa histone deacetylases upregulates the sumoylation of MEF2 transcription factors. Mol. Cell. Biol. 25:2273-2287. [PMC free article] [PubMed]
28. Hay, R. T. 2005. SUMO: a history of modification. Mol. Cell 18:1-12. [PubMed]
29. Hietakangas, V., J. Anckar, H. A. Blomster, M. Fujimoto, J. J. Palvimo, A. Nakai, and L. Sistonen. 2006. PDSM, a motif for phosphorylation-dependent SUMO modification. Proc. Natl. Acad. Sci. USA 103:45-50. [PubMed]
30. Jones, G., P. Willett, R. C. Glen, A. R. Leach, and R. Taylor. 1997. Development and validation of a genetic algorithm for flexible docking. J. Mol. Biol. 267:727-748. [PubMed]
31. Kagey, M. H., T. A. Melhuish, S. E. Powers, and D. Wotton. 2005. Multiple activities contribute to Pc2 E3 function. EMBO J. 24:108-119. [PubMed]
32. Kagey, M. H., T. A. Melhuish, and D. Wotton. 2003. The polycomb protein Pc2 is a SUMO E3. Cell 113:127-137. [PubMed]
33. Kirsh, O., J. S. Seeler, A. Pichler, A. Gast, S. Muller, E. Miska, M. Mathieu, A. Harel-Bellan, T. Kouzarides, F. Melchior, and A. Dejean. 2002. The SUMO E3 ligase RanBP2 promotes modification of the HDAC4 deacetylase. EMBO J. 21:2682-2691. [PubMed]
34. Langley, E., M. Pearson, M. Faretta, U. M. Bauer, R. A. Frye, S. Minucci, P. G. Pelicci, and T. Kouzarides. 2002. Human SIR2 deacetylates p53 and antagonizes PML/p53-induced cellular senescence. EMBO J. 21:2383-2396. [PubMed]
35. Lefebvre, T., S. Pinte, C. Guerardel, S. Deltour, N. Martin-Soudant, M. C. Slomianny, J. C. Michalski, and D. Leprince. 2004. The tumor suppressor HIC1 (hypermethylated in cancer 1) is O-GlcNAc glycosylated. Eur. J. Biochem. 271:3843-3854. [PubMed]
36. Lehembre, F., P. Badenhorst, S. Muller, A. Travers, F. Schweisguth, and A. Dejean. 2000. Covalent modification of the transcriptional repressor Tramtrack by the ubiquitin-related protein Smt3 in Drosophila flies. Mol. Cell. Biol. 20:1072-1082. [PMC free article] [PubMed]
37. Luo, J., A. Y. Nikolaev, S. Imai, D. Chen, F. Su, A. Shiloh, L. Guarente, and W. Gu. 2001. Negative control of p53 by Sir2α promotes cell survival under stress. Cell 107:137-148. [PubMed]
38. Melnick, A., G. Carlile, K. F. Ahmad, C. L. Kiang, C. Corcoran, V. Bardwell, G. G. Prive, and J. D. Licht. 2002. Critical residues within the BTB domain of PLZF and Bcl-6 modulate interaction with corepressors. Mol. Cell. Biol. 22:1804-1818. [PMC free article] [PubMed]
39. Meziane-Tani, M., P. Lagant, A. Semmoud, and G. Vergoten. 2006. The SPASIBA force field for chondroitin sulfate: vibrational analysis of d-glucuronic and N-acetyl-d-galactosamine 4-sulfate sodium salts. J. Phys. Chem. A. 110:11359-11370. [PubMed]
40. Perdomo, J., A. Verger, J. Turner, and M. Crossley. 2005. Role for SUMO modification in facilitating transcriptional repression by BKLF. Mol. Cell. Biol. 25:1549-1559. [PMC free article] [PubMed]
41. Pichler, A., A. Gast, J. S. Seeler, A. Dejean, and F. Melchior. 2002. The nucleoporin RanBP2 has SUMO1 E3 ligase activity. Cell 108:109-120. [PubMed]
42. Pinte, S., N. Stankovic-Valentin, S. Deltour, B. R. Rood, C. Guerardel, and D. Leprince. 2004. The tumor suppressor gene HIC1 (hypermethylated in cancer 1) is a sequence-specific transcriptional repressor: definition of its consensus binding sequence and analysis of its DNA binding and repressive properties. J. Biol. Chem. 279:38313-38324. [PubMed]
43. Ross, S., J. L. Best, L. I. Zon, and G. Gill. 2002. SUMO-1 modification represses Sp3 transcriptional activation and modulates its subnuclear localization. Mol. Cell 10:831-842. [PubMed]
44. Roth, S. Y., J. M. Denu, and C. D. Allis. 2001. Histone acetyltransferases. Annu. Rev. Biochem. 70:81-120. [PubMed]
45. Rui, H. L., E. Fan, H. M. Zhou, Z. Xu, Y. Zhang, and S. C. Lin. 2002. SUMO-1 modification of the C-terminal KVEKVD of Axin is required for JNK activation but has no effect on Wnt signaling. J. Biol. Chem. 277:42981-42986. [PubMed]
46. Sapetschnig, A., G. Rischitor, H. Braun, A. Doll, M. Schergaut, F. Melchior, and G. Suske. 2002. Transcription factor Sp3 is silenced through SUMO modification by PIAS1. EMBO J. 21:5206-5215. [PubMed]
47. Sato, Y., K. Miyake, H. Kaneoka, and S. Iijima. 2006. Sumoylation of CCAAT/enhancer-binding protein alpha and its functional roles in hepatocyte differentiation. J. Biol. Chem. 281:21629-21639. [PubMed]
48. Schmidt, D., and S. Muller. 2003. PIAS/SUMO: new partners in transcriptional regulation. Cell. Mol. Life Sci. 60:2561-2574. [PubMed]
49. Seeler, J. S., and A. Dejean. 2003. Nuclear and unclear functions of SUMO. Nat. Rev. Mol. Cell Biol. 4:690-699. [PubMed]
50. Seeler, J. S., A. Marchio, R. Losson, J. M. Desterro, R. T. Hay, P. Chambon, and A. Dejean. 2001. Common properties of nuclear body protein SP100 and TIF1α chromatin factor: role of SUMO modification. Mol. Cell. Biol. 21:3314-3324. [PMC free article] [PubMed]
51. Shalizi, A., B. Gaudilliere, Z. Yuan, J. Stegmuller, T. Shirogane, Q. Ge, Y. Tan, B. Schulman, J. W. Harper, and A. Bonni. 2006. A calcium-regulated MEF2 sumoylation switch controls postsynaptic differentiation. Science 311:1012-1017. [PubMed]
52. Song, J., Z. Zhang, W. Hu, and Y. Chen. 2005. Small ubiquitin-like modifier (SUMO) recognition of a SUMO binding motif: a reversal of the bound orientation. J. Biol. Chem. 280:40122-40129. [PubMed]
53. Stankovic-Valentin, N., A. Verger, S. Deltour-Balerdi, K. G. Quinlan, M. Crossley, and D. Leprince. 2006. A L225A substitution in the human tumour suppressor HIC1 abolishes its interaction with the corepressor CtBP. FEBS J. 273:2879-2890. [PubMed]
54. Stogios, P. J., G. S. Downs, J. J. Jauhal, S. K. Nandra, and G. G. Prive. 2005. Sequence and structural analysis of BTB domain proteins. Genome Biol. 6:R82. [PMC free article] [PubMed]
55. Valenta, T., J. Lukas, L. Doubravska, B. Fafilek, and V. Korinek. 2006. HIC1 attenuates Wnt signaling by recruitment of TCF-4 and β-catenin to the nuclear bodies. EMBO J. 25:2326-2337. [PubMed]
56. Vaziri, H., S. K. Dessain, E. Ng Eaton, S. I. Imai, R. A. Frye, T. K. Pandita, L. Guarente, and R. A. Weinberg. 2001. hSIR2SIRT1 functions as an NAD-dependent p53 deacetylase. Cell 107:149-159. [PubMed]
57. Verger, A., J. Perdomo, and M. Crossley. 2003. Modification with SUMO. A role in transcriptional regulation. EMBO Rep. 4:137-142. [PubMed]
58. Vo, N., C. Fjeld, and R. H. Goodman. 2001. Acetylation of nuclear hormone receptor-interacting protein RIP140 regulates binding of the transcriptional corepressor CtBP. Mol. Cell. Biol. 21:6181-6188. [PMC free article] [PubMed]
59. Wales, M. M., M. A. Biel, W. El Deiry, B. D. Nelkin, J. P. Issa, W. K. Cavenee, S. J. Kuerbitz, and S. B. Baylin. 1995. p53 activates expression of HIC-1, a new candidate tumour suppressor gene on 17p13.3. Nat Med. 1:570-577. [PubMed]
60. Wang, J. M., C. Y. Ko, L. C. Chen, W. L. Wang, and W. C. Chang. 2006. Functional role of NF-IL6β and its sumoylation and acetylation modifications in promoter activation of cyclooxygenase 2 gene. Nucleic Acids Res. 34:217-231. [PMC free article] [PubMed]
61. Wynshaw-Boris, A., and M. J. Gambello. 2001. LIS1 and dynein motor function in neuronal migration and development. Genes Dev. 15:639-651. [PubMed]
62. Yang, S. H., and A. D. Sharrocks. 2004. SUMO promotes HDAC-mediated transcriptional repression. Mol. Cell 13:611-617. [PubMed]
63. Yang, X. J., and S. Gregoire. 2006. A recurrent phospho-sumoyl switch in transcriptional repression and beyond. Mol. Cell 23:779-786. [PubMed]
64. Zhao, X., T. Sternsdorf, T. A. Bolger, R. M. Evans, and T. P. Yao. 2005. Regulation of MEF2 by histone deacetylase 4- and SIRT1 deacetylase-mediated lysine modifications. Mol. Cell. Biol. 25:8456-8464. [PMC free article] [PubMed]

Articles from Molecular and Cellular Biology are provided here courtesy of American Society for Microbiology (ASM)