The balance between acetylation and deacetylation controls Smad7 stability.

Transforming growth factor beta (TGFbeta) regulates multiple cellular processes via activation of Smad signaling pathways. We have recently demonstrated that the inhibitory Smad7 interacts with the acetyl transferase p300 and that p300 acetylates Smad7 on two lysine residues. These lysine residues are critical for Smurf-mediated ubiquitination of Smad7, and acetylation protects Smad7 from TGFbeta-induced degradation. In this study we demonstrate that Smad7 interacts with specific histone deacetylases (HDACs) and that the same HDACs are able to deacetylate Smad7. The interaction with HDACs is dependent on the C-terminal MH2 domain of Smad7. In addition, HDAC1-mediated deacetylation of Smad7 decreases the stability of Smad7 by enhancing its ubiquitination. Thus, our results demonstrate that the degradation of Smad7 is regulated by the balance between acetylation, deacetylation and ubiquitination, indicating that this could be a general mechanism to regulate the stability of cellular proteins.

Transforming growth factor beta (TGF␤) regulates multiple cellular processes via activation of Smad signaling pathways. We have recently demonstrated that the inhibitory Smad7 interacts with the acetyl transferase p300 and that p300 acetylates Smad7 on two lysine residues. These lysine residues are critical for Smurfmediated ubiquitination of Smad7, and acetylation protects Smad7 from TGF␤-induced degradation. In this study we demonstrate that Smad7 interacts with specific histone deacetylases (HDACs) and that the same HDACs are able to deacetylate Smad7. The interaction with HDACs is dependent on the C-terminal MH2 domain of Smad7. In addition, HDAC1-mediated deacetylation of Smad7 decreases the stability of Smad7 by enhancing its ubiquitination. Thus, our results demonstrate that the degradation of Smad7 is regulated by the balance between acetylation, deacetylation and ubiquitination, indicating that this could be a general mechanism to regulate the stability of cellular proteins.
Transforming growth factor beta (TGF␤) 1 is a member of the TGF␤ superfamily of cytokines that regulate multiple cellular processes including extracellular matrix production, cell growth, apoptosis, and differentiation. Dysfunction of TGF␤ signaling has been implicated in various human disorders ranging from vascular diseases to cancer progression (for a recent review see Ref. 1). TGF␤ exerts its cellular effects via formation of a heteromeric complex of type I and type II serine/ threonine kinase receptors. The type II receptor phosphorylates and activates the type I receptor, which in turn phosphorylates and activates the receptor-activated Smads (Smad2 and Smad3) in their C-terminal SSXS motif. The activated Smads then interact with Smad4 and translocate into the nucleus where they act as transcription factors together with co-activators and co-repressors (2).
The subfamily of inhibitory Smads consists of Smad6 and Smad7, which are immediate early target genes of TGF␤ (3)(4)(5)(6). Smad7 is a nuclear protein in resting cells (7); after TGF␤ stimulation, Smad7 translocates out of the nucleus to the plasma membrane in a manner that is dependent on the Smurf E3-ubiquitin ligases (8 -10). At the cell membrane, the Smad7-Smurf complex interacts with the activated receptors and inhibits TGF␤ signaling by blocking the interaction between the receptor-activated Smads and the activated receptors (3,6). In addition, Smurf ubiquitinates both the receptors and Smad7, thereby inducing the degradation of both the receptors and Smad7 (8,9).
We have recently shown that Smad7 interacts with the transcriptional co-activator p300, a protein acetyl transferase (11). p300-mediated acetylation transfers the acetyl moiety from acetyl coenzyme A to the amino group of a lysine residue of the acceptor protein. Acetylation is a dynamic process and the balance between acetylated and non-acetylated histones has major effects on chromatin structure and transcription (for a recent review see Ref. 12). Histones H3 and H4 are acetylated on specific lysine residues in their N-terminals, thereby relaxing the nucleosomal structure and allowing transcription. It has been demonstrated that non-histone proteins such as p53 (13), E2F (14), YY1 (15), NFB (16), SREBP (17), and Smad7 (11) are acetylated. Acetylation can have multiple effects, including changes in protein-DNA interactions (18) or proteinprotein interactions (19). The side chain of lysine residues is also targeted by another posttranslational modification, ubiquitination. Polyubiquitination modulates protein function by inducing proteasome-dependent degradation. Protein acetylation can also affect protein stability, because it has been demonstrated that acetylation prevents ubiquitination of the same lysine residues (11, 20 -22).
Acetylation is a reversible process, and removal of the acetyl group is catalyzed by histone deacetylases ((HDACs) for a recent review see Ref. 23). Four classes of mammalian HDACs have been described. The class I HDACs, consisting of HDAC1, -2, -3 and -8, are homologs of the yeast RPD3 protein, whereas the class II HDACs, HDAC4, -5, -6, -7, -9, and -10, are related to yeast HDA1. Class III HDACs, the sirtuins, are homologs of the yeast protein Sir2. In contrast to members of the class I and II HDACs, class III HDACs require NAD ϩ for their deacetylase activity (24) and are not affected by the well known HDAC inhibitor trichostatin A (TSA) (25). Based on its sequence, the most recently described HDAC, HDAC11 (26), has been placed in a separate group of HDACs (27). Class II HDACs have been shown to shuttle between the cytosol and the nucleus in response to extracellular stimulation (25,28,29), whereas class I HDACs are predominantly nuclear proteins. All classes of HDACs have been shown to deacetylate histones. In addition, it was recently shown that HDAC1 interacts with the ubiquitin ligase MDM2 and that HDAC1-mediated deacetylation of the tumor suppressor p53 promotes MDM2-mediated ubiquitination and degradation of p53 (20). Thus, HDAC-mediated deacetylation may be a novel mechanism to regulate the ubiquitination and degradation of some acetylated proteins.
In this study, we demonstrate that Smad7 interacts with spe-* This work was supported in part by grants from the Swedish Cancer Foundation (to E. G.) and the Swedish Research Council (to J. E.). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
‡ Research Fellow of the Royal Swedish Academy of Sciences through a grant from the Knut and Alice Wallenberg Foundation. To whom correspondence should be addressed. Tel.: 4618160404; Fax: 4618160420; E-mail: johan.ericsson@licr.uu.se. 1 The abbreviations used are: TGF␤, transforming growth factor ␤; HDAC, histone deacetylase; siRNA, short interfering RNA; TSA, trichostatin A; E3, ubiquitin-protein isopeptide ligase; GST, glutathione S-transferase; HA, hemagglutinin. cific HDACs. The HDACs that associate with Smad7 also deacetylate the protein. We also demonstrate that endogenous Smad7 interacts with HDAC1 and map the interaction to the C-terminal MH2 domain of Smad7. siRNA-mediated inactivation of endogenous HDAC1 enhances the acetylation and steady-state levels of endogenous Smad7. In addition, HDAC1-mediated deacetylation of Smad7 promotes the ubiquitination and degradation of Smad7. Thus, our results indicate that the balance between acetylation and deacetylation of Smad7 can control its degradation through the ubiquitin-proteasome pathway.

EXPERIMENTAL PROCEDURES
Cell Culture-All tissue culture media and antibiotics were obtained from Invitrogen and Sigma Genosys. Human embryonic kidney epithelial 293, 293T, and HepG2 cells were from American Type Culture Collection. Cells were maintained at 37°C in Dulbecco's modified Eagle's medium supplemented with 10% fetal calf serum, sodium pyruvate (1 mM), non-essential amino acids (1ϫ), 50 units/ml penicillin, and 50 g/ml streptomycin in 5% CO 2 . For overnight starvation, cells were incubated in Dulbecco's modified Eagle's medium supplemented with 0.5% fetal calf serum, sodium pyruvate (1 mM), non-essential amino acids (1ϫ), 50 units/ml penicillin, and 50 g/ml streptomycin.
Immunoprecipitations and Immunoblotting-Cells were lysed in icecold lysis buffer (50 mM HEPES, pH 7.2, 150 mM NaCl, 1 mM EDTA, 1 mM EGTA, 20 mM NaF, 2 mM NaVO 4 , 1% (w/v) Triton X-100, 10% (w/v) glycerol, 1 mM phenylmethylsulfonyl fluoride, 10 mM sodium butyrate, and 1% Aprotenin) and cleared by centrifugation. Immunoprecipitations were performed by adding the appropriate antibodies plus protein G-Sepharose beads, followed by incubation for 3 h at 4°C. The immune complexes were washed three times with lysis buffer, once with 0.5 M NaCl, and once with water. SDS sample buffer (400 mM Tris-HCl, pH 8.8, 4% SDS, 1 M sucrose, 10 mM EDTA, and 0.01% bromphenol blue) was added, and the precipitates were heated. The samples were resolved by SDS-PAGE and transferred to nitrocellulose membranes (Millipore, Bedford, MA). After blocking in phosphate-buffered saline with the addition of 5% bovine serum albumin, the membranes were incubated with the appropriate antibodies, washed with phosphate-buffered saline containing 0.05% Triton X-100, and incubated with horseradish peroxidase-coupled secondary antibodies. After washing, the blots were visualized with Western blotting chemiluminescence luminol reagent (Santa Cruz Biotechnology). For reprobing, blots were incubated in stripping buffer (0.2 M NaOH, 0.5 M NaCl) at room temperature for 10 min and washed extensively with Tris-buffered saline before blocking.
In Vivo Ubiquitination-Cells were lysed in Ub-lysis buffer (1% SDS, 1 mM EDTA, 0.5 mM Tris-HCl, pH 7.4) and sonicated. The samples were boiled for 10 min and diluted 10 times with lysis buffer. After clearing by centrifugation, immunoprecipitations were performed as described.
Generation of Recombinant Proteins and in Vitro Pull-down Assays-GST-Smad7 and GST-HDAC1 fusion proteins were expressed in BL21 (DE3pLysS) and purified according to standard protocols. The amount of protein on the beads was estimated by Coomassie staining of SDS-PAGE gels, and equal amounts of protein were used in the binding assays. Proteins were in vitro translated using the T7 TNT kit (Promega) and 35 S-labeled methionine and cysteine (Promix, Amersham Biosciences) and incubated with GST fusion proteins prebound to glutathione beads at 4°C for 3 h in the presence of bovine serum albumin (1 mg/ml) to avoid unspecific binding. The samples were treated as described for immunoprecipitates and resolved by SDS-PAGE. The gel was incubated in enhancer (Amplify, Amersham Biosciences) for 15 min, dried, and exposed to PhosphorImager overnight.
Deacetylation Assay-293T cells were transfected with Myc-Smad7 and HA-p300 using the reagents and methods described above. Acetylated Myc-Smad7 was immunoprecipitated, and the precipitates were washed extensively. After washing, HDAC buffer (10 mM Tris, pH 7.5, 10 mM NaCl, 10% glycerol, 5 mM MgCl 2 ) was added. The precipitates were aliquotted and used in deacetylation assays in the absence or presence of HeLa nuclear extract. To inactivate endogenous HDACs, HeLa nuclear extracts were preincubated 45 min on ice with TSA (final concentration 100 nM). The samples were incubated at 30°C for 2 h, SDS sample buffer was added, and the samples were separated by SDS-PAGE followed by immunoblotting. The preparation of HeLa nuclear extracts has been described elsewhere (31).

Inhibition of HDAC Activity Increases the Acetylation of Smad7-
We have recently shown that Smad7 is acetylated on lysine residues 64 and 70 by p300 (11). During our studies we observed that activation of the TGF␤ receptor resulted in a reduction in the acetylation of Smad7 (11), suggesting that Smad7 is deacetylated in response to TGF␤ signaling. To test the possibility that Smad7 is a substrate for HDACs, 293T cells were transiently transfected with Smad7 either alone or together with p300 and incubated in the absence and presence of the HDAC inhibitor TSA. Following lysis of the cells, Smad7 was immunoprecipitated and resolved on SDS-polyacrylamide gel, and the acetylation of Smad7 was monitored by anti-acetyl lysine antibodies. As previously reported, Smad7 was acetylated by p300 (Fig. 1A). Interestingly, the acetylation of Smad7 was enhanced in cells treated with TSA (Fig. 1A, compare lanes  2 and 4). The same result was obtained using an antibody specific for the major acetylation site in Smad7 (Fig. 1B). These results indicate that Smad7 is actively deacetylated in vivo. To further test the possibility that Smad7 is a HDAC substrate, Smad7 was expressed in 293T cells together with p300 and immunoprecipitated. The immunoprecipitated material was divided into three aliquots. One aliquot was left untreated, one aliquot was incubated with nuclear extract from HeLa cells, and the third aliquot was incubated with HeLa nuclear extract pretreated with TSA to inhibit endogenous HDACs. After incubation, the reactions were stopped by the addition of SDS sample buffer. The samples were resolved by SDS-PAGE, and the acetylation of Smad7 was detected by Western blotting. The incubation of Smad7 with HeLa nuclear extract significantly reduced the acetylation of Smad7 (Fig. 1C). Furthermore, preincubating the HeLa nuclear extracts with TSA blocked the effect on the acetylation of Smad7, indicating that the deacetylation of Smad7 is catalyzed by a TSA-sensitive process. Taken together, these results suggest that acetylated Smad7 is a potential HDAC substrate.
Smad7 Interacts with Specific HDACs-To analyze whether Smad7 is associated with specific members of the HDAC family of proteins, we performed co-immunoprecipitation experiments in transiently transfected 293T cells, using Myc-tagged Smad7 and FLAG-tagged HDACs (HDAC1-HDAC6). Among the class I HDACs, both HDAC1 and HDAC3 interacted well with Smad7, whereas only a weak interaction was detected between Smad7 and HDAC2 ( Fig. 2A). Among the class II HDACs, both HDAC5 and HDAC6 interacted with Smad7, whereas we were unable to detect any interaction between Smad7 and HDAC4 ( Fig. 2A). The interactions between Smad7 and HDAC5 or HDAC6 were consistently weaker than the interaction between Smad7 and HDAC1 or HDAC3. This difference made us focus on the interaction between Smad7 and HDAC1. GST pull-down experiments using GST-HDAC1 and in vitro translated Smad7 showed that these two proteins interact in vitro and that the interaction most probably is direct (Fig. 2B). To verify that Smad7 and HDACs interact under physiological conditions, the interaction between endogenous proteins was investigated. As illustrated in Fig. 2C, Smad7 antibodies immunoprecipitated endogenous HDAC1 from 293T whole-cell lysates, whereas an unrelated antibody was unable to do so.
The C-terminal Domain of Smad7 Interacts with HDAC1-Smad7 contains an N-terminal domain, which is conserved between the I-Smads and a C-terminal MH2 domain, which is conserved in all members of the Smad family. Smad7 also contains a linker region, which, via its PY motif, is critical for the interaction with the Smurf family of ubiquitin ligases (8,32). The acetylated lysine residues are located within the Nterminal domain of Smad7 (11). To determine the domain(s) in Smad7 that are responsible for its interaction with HDAC1, we used deletion mutants of Smad7 together with full-length HDAC1 in co-immunoprecipitation experiments. As seen in Fig. 3A, full-length Smad7 and the isolated MH2 domain (amino acids 261-246) interacted strongly with HDAC1, whereas no interaction was detected between HDAC1 and the N-terminal domain (amino acids 1-126) or linker region (amino acids 127-260) of Smad7. To further map the interaction between Smad7 and HDAC1, we performed co-immunoprecipita-  ). B, FLAGtagged, wild-type, or mutant (K64A) Smad7 was expressed in 293T cells in the absence or presence of HA-p300. Twenty-four hours posttransfection, cells were either left untreated or treated with TSA (200 ng/ml) for 12 h. Following immunoprecipitation of Smad7, samples were resolved by SDS-PAGE, and the acetylation of Smad7 was detected with anti-acetyl lysine antibodies directed against the major acetylation site in Smad7 (␣-AcK64). The amount of Smad7 in the immunoprecipitates was determined with anti-FLAG antibodies (␣-Flag). C, 6Myc-tagged Smad7 was expressed together with HA-tagged p300. Following immunoprecipitation of Smad7 with anti-Myc antibodies, the sample was divided and incubated alone, with HeLa nuclear extracts, or with HeLa nuclear extract preincubated with TSA (100 nM). After incubation, the samples were resolved by SDS-PAGE, and the amount of acetylated Smad7 was detected by Western blotting with anti-acetyl lysine antibodies (␣-AcK). The amount of Smad7 in the immunoprecipitates was determined with anti-Myc antibodies (␣-Myc).

FIG. 2. Smad7 interacts with specific deacetylases.
A, 6Myctagged Smad7 was expressed in the absence or presence of the indicated FLAG-tagged HDACs. Following immunoprecipitation of the HDACs with anti-FLAG antibodies, the samples were resolved by SDS-PAGE, and the amount of co-precipitating Smad7 was determined with anti-Myc antibodies (␣-Myc). The levels of Flag-HDAC1, -2, -3, -4, -5, and -6 in the immunoprecipitates and 6Myc-Smad7 in total cell lysates were determined by Western blotting. B, Smad7 was in vitro translated and used in GST pull-down assays with GST alone or GST-HDAC1. The amount of Smad7 was analyzed by phosphorimage analysis. C, 293T cell lysates were immunoprecipitated with an unrelated polyclonal antiserum (lane 1) or an antiserum directed against Smad7 (lane 2). The immunocomplexes were washed, and the proteins were separated by SDS-PAGE. Co-immunoprecipitating HDAC1 was detected by Western blotting using anti-HDAC1 antibodies. The amount of Smad7 in the immunoprecipitates was determined with anti-Smad7 antibodies (␣-Smad7). tion experiments using Myc-tagged Smad7 together with FLAG-tagged HDAC1, either full-length or three deletion fragments (amino acid residues 1-160, 161-483, and 321-483). As expected, Smad7 interacted with full-length HDAC1. Smad7 also interacted with the large central fragment of HDAC1, whereas no interaction was detected with the N-terminal or the C-terminal fragments of HDAC1 (Fig. 3B). The fragment of HDAC1 that interacted with Smad7 encompasses a large part of the catalytic domain of HDAC1 (33), further supporting the hypothesis that Smad7 is a substrate for HDAC1.
Smad7 Is a HDAC Substrate-To investigate whether Smad7 is a substrate for the deacetylase activity of HDACs, 293T cells were transiently transfected with p300 and Smad7, either in the absence or presence of various HDACs (HDAC1-HDAC6). After immunoprecipitation, the acetylation of Smad7 was monitored with anti-acetyl lysine antibodies. The acetylation of Smad7 was drastically reduced in the presence of HDAC1, -3, and -6 (Fig. 4A). In contrast, only minor or no effects on the acetylation of Smad7 were seen when the protein was co-expressed with HDAC2, -4, or -5 (Fig. 4A). Thus, the HDACs that interacted the strongest with Smad7 in the coprecipitation assays were the most effective in catalyzing the deacetylation of Smad7 (compare Figs. 2A and 4A). To further study the effect of HDAC1 on the acetylation of Smad7, the acetylation of the protein was studied in response to short hairpin RNA-mediated inactivation of HDAC1. 293T cells were transfected with Smad7 with or without p300. As seen in Fig.  4B, the acetylation of Smad7 was significantly increased when the protein levels of endogenous HDAC1 were reduced, indicating that endogenous HDAC1 affects the acetylation of Smad7. To further confirm that the acetylation of Smad7 is regulated by endogenous HDACs, HepG2 cells were transfected with Flag-Smad7 in the absence or presence of a vector expressing short hairpin RNA for HDAC1. Twenty-four hours after transfection, TGF␤ was added to the cells to stimulate Following immunoprecipitation of Smad7 with anti-Myc antibodies, the samples were resolved by SDS-PAGE, and the level of acetylated Smad7 was detected by Western blotting with anti-acetyl lysine antibodies (␣-AcK). The amount of Smad7 in the immunoprecipitates was determined with anti-Myc antibodies (␣-Myc). The levels of HDAC1 in the total cell lysates were determined with HDAC1 antibodies (␣-HDAC1). C, HepG2 cells were transfected with FLAG-tagged Smad7 in the absence or presence of pSUPER-control (lane 1) or pSUPER-HDAC1 (lane 2). Following immunoprecipitation of Smad7 with anti-FLAG antibodies, the samples were resolved by SDS-PAGE, and the levels of acetylated Smad7 were detected by Western blotting with anti-acetyl lysine antibodies (␣-AcK64). The amount of Smad7 in the immunoprecipitates was determined with anti-FLAG antibodies (␣-Flag).
Smad7 turnover. Following immunoprecipitation of the transfected Smad7, the acetylation of the protein was analyzed with anti-acetyl lysine antibodies. Inactivation of endogenous HDAC1 enhanced the acetylation of Smad7 in HepG2 (Fig. 4C). Similar results were also observed in HeLa cells (data not shown).
HDAC1 Decreases the Half-life of Smad7 by Promoting Its Ubiquitination-For some proteins, the acetylation of specific lysine residues is able to prevent ubiquitination of the same residues, thereby preventing protein degradation (11,20,21). To determine whether HDAC1 could affect the half-life of Smad7, 293 cells were transfected with FLAG-tagged Smad7 in the absence or presence of co-transfected FLAG-tagged HDAC1. The cells were starved overnight and stimulated with TGF␤ for the indicated time periods in the presence of cycloheximide. In a separate experiment, TSA was added to cells expressing Flag-Smad7 to inhibit the activity of endogenous HDACs. Following lysis of the cells, the levels of Smad7 were determined by Western blotting. Co-expression of HDAC1 significantly decreased the half-life of Smad7 (Fig. 5A, compare the top and middle panels). Interestingly, the degradation of transiently transfected Smad7 was blocked in cells treated with TSA, indicating that endogenous HDACs regulate the stability of Smad7 (Fig. 5A, lower panel). To determine whether endogenous HDAC1 could influence the stability of Smad7, HepG2 cells were transfected with low amounts of Flag-Smad7 in the absence or presence of a vector expressing short hairpin RNA for HDAC1. As illustrated in Fig. 5B, the steady-state levels of Flag-Smad7 were enhanced in response to inactivation of endogenous HDAC1, suggesting that HDAC1 regulates the stability of Smad7. Similar results were obtained when the experiment was repeated in HeLa cells (data not shown). To determine whether HDAC1 could also affect the ubiquitination of Smad7, 293T cells were transfected with FLAG-tagged Smad7 and HA-tagged ubiquitin in the absence or presence of co-transfected HDAC1. After lysis of the cells, Smad7 was immunoprecipitated and the ubiquitination of Smad7 was determined by Western blotting. Co-expression of HDAC1 significantly increased the ubiquitination of Smad7 (Fig. 5C), indicating that deacetylation of Smad7 increases the amount of lysine residues available for subsequent ubiquitination.
To establish that endogenous HDACs regulate the deacetylation of endogenous Smad7, HepG2 cells were treated with or without TSA for 8 h. After lysis, Smad7 was immunoprecipitated and resolved by SDS-PAGE. As seen in Fig. 6A, the acetylation of endogenous Smad7 was increased after TSA treatment, suggesting that the acetylation of Smad7 is regulated by endogenous HDACs. To investigate the effect of en- FIG. 5. HDAC1 regulates the stability and ubiquitination of Smad7. A, FLAG-tagged Smad7 was expressed in 293 cells in the absence or presence of FLAG-tagged HDAC1. Twenty-four hours after transfection, cells were starved overnight. The following day, fresh starvation medium supplemented with cycloheximide (100 g/ml) and TGF␤ (5 ng/ml) was added to the cells. Where indicated, TSA (500 ng/ml) was also added to the cells. After incubation for the indicated times, cell lysates were prepared and analyzed by SDS-PAGE. The levels of Smad7 were detected by Western blotting with anti-FLAG antibodies. As a control for equal loading, the filters were probed for tubulin (lower panels). B, HepG2 cells were transfected with FLAGtagged Smad7 in the absence or presence of pSUPER-control (lane 1) or pSUPER-HDAC1 (lane 2). Following lysis, the samples were resolved by SDS-PAGE, and the protein levels of Smad7 were detected by Western blotting with anti-FLAG antibodies (␣-Flag). The amount of HDAC1 in the lysates was determined with anti-HDAC1 antibodies. The filters were probed for tubulin as a control for equal loading (lower panel). C, Myc-tagged Smad7 was expressed in 293T together with HA-tagged ubiquitin in the absence or presence of FLAG-tagged HDAC1. Thirty-six hours after transfection, cell lysates were prepared. Following immunoprecipitation (IP) of Smad7 with anti-Myc antibodies, the samples were resolved by SDS-PAGE, and the presence of ubiquitinated (Ub) Smad7 was detected with anti-Ub antibodies (␣-Ub). The levels of Smad7 in the immunoprecipitates and of HDAC1 in the cell lysates were determined with anti-Myc (␣-Myc) and anti-FLAG (␣-Flag) antibodies, respectively. The migration of an antibody-related band is indicated by an asterisk. dogenous HDAC1 on the acetylation of endogenous Smad7, gene silencing experiments were performed. HepG2 cells were either transfected with a nonspecific control siRNA or a HDAC1-specific siRNA. Lysates were prepared, and Smad7 was immunoprecipitated. Analysis by Western blotting revealed that inactivation of endogenous HDAC1 resulted in an increased acetylation of endogenous Smad7 (Fig. 6B). Inactivation of endogenous HDAC1 also increased the total amount of Smad7 that could be precipitated from HepG2 cells (Fig. 6B, middle panel). Similar results were also obtained when the experiments were repeated in 293T cells (data not shown). Taken together these data indicate that endogenous Smad7 is a substrate for endogenous HDAC1 and that HDAC1-mediated deacetylation affects the stability of Smad7. DISCUSSION In this study we presented evidence that Smad7 interacts with multiple HDACs, resulting in the deacetylation of Smad7, which enhances its ubiquitination and degradation, further strengthening the hypothesis that the acetylation of Smad7 is critical for its stability. We demonstrated that the acetylation of Smad7 is increased in the presence of the deacetylase inhibitor TSA in vivo, indicating that Smad7 is actively deacetylated by HDACs (Figs. 1A and 6A). An additional indication that Smad7 is a substrate for HDACs came from the observation that HeLa nuclear extracts contained an activity that could deacetylate Smad7 and that this activity was TSA-sensitive (Fig. 1C). TSA has been shown to inhibit both class I and class II HDACs, whereas class III HDACs are insensitive to TSA (25). Thus, our results indicate that class III HDACs are not involved in the deacetylation of Smad7; in accordance, treatment of cells with nicotinamide, an inhibitor of the class III HDACs, had no effect on Smad7 acetylation (data not shown).
Smad7 interacted with both HDAC1 and -3, whereas the interaction with HDAC2 was weak ( Fig. 2A). Although HDAC1 and -2 have 85% amino acid sequence similarity (34) and are found in the same repressor complexes, it was recently shown that in the DT40 chicken B-cell line, disruption of either HDAC1 or -2 resulted in different protein expression patterns (35). This observation indicates that HDAC1 and -2 may have both specific and common interaction partners, which is consistent with the data on the differential interaction with Smad7 described in the present study.
The class II HDACs, HDAC5 and -6, were also able to interact with ( Fig. 2A) and deacetylate (Fig. 4A) Smad7, although to a lesser extent than the class I HDACs. In contrast, Bai and Cao (36) showed that Smad6, which interacts with HDAC1 and -3, failed to interact with any member of the class II HDACs tested.
The C-terminal MH2 domain of Smad7 interacted strongly with full-length HDAC1 (Fig. 3A), whereas no interaction was seen with the N-terminal domain or linker region. This is similar to Smad6, which also interacts with HDAC1 through its MH2 domain (36). The interaction between Smad7 and HDAC1 is most probably direct as GST-HDAC1 was able to pull down Smad7 (Fig. 2B). Interestingly, GST-Smad7 failed to interact with any of the HDACs tested (data not shown), indicating that Smad7 needs to be modified to interact with HDACs. Smad7 has been shown to be both acetylated (11) and phosphorylated (37). Acetylation of Smad7 is mediated by p300 and Smad7 is readily acetylated in vitro. However, in vitro acetylated GST-Smad7 still failed to associate with any of the HDACs tested indicating that this modification does not suffice for HDAC interaction (data not shown). The kinase(s) phosphorylating Smad7 is currently unknown preventing us from testing if this modification is critical for the interaction with HDACs in vitro. It was recently shown that the interaction between BCL-3 and HDAC1 is regulated by phosphorylation of BCL-3 and that HDAC1 negatively regulates BCL-3 stability (38). It is tempting to speculate that a similar mechanism regulates the interaction between HDAC1 and Smad7. The observation that Smad7 interacts with both transcriptional co-activators (p300) and co-repressors (HDACs) indicate that Smad7 could have a direct role in transcriptional regulation. Further studies will be necessary to analyze this possibility.
HDAC6 has been shown to associate with proteins involved in the ubiquitin signaling pathway (39). The link between acetylation and ubiquitination gained further support from the observations that the ubiquitination and degradation of various proteins such as Smad7 (11), SREBP (17), and c-Myc (40) were reduced following acetylation. In Smad7, mutation of the two acetylated residues prevented its polyubiquitination and degradation (11), suggesting that acetylation protects Smad7 from ubiquitination. In support of this hypothesis, co-expression of HDAC1 and Smad7 enhanced the ubiquitination of Smad7 and increased its degradation (Fig. 5). In addition, the acetylation as well as the protein levels of endogenous Smad7 were increased after siRNA-mediated inactivation of endogenous HDAC1 (Fig. 6B). We have shown previously that the acetylation of Smad7 is decreased in response to TGF␤-signaling (11). The reduction in acetylation is most probably the result of multiple, separate events. TGF␤ induces nuclear export of Smad7, thereby preventing p300-mediated acetylation. It is also possible that TGF␤ signaling enhances HDAC1-mediated deacetylation of Smad7. However, we were unable to detect any TGF␤ dependence in the interaction between Smad7 and HDAC1 (data not shown). Thus, further investigations regarding the mechanisms involved in TGF␤-dependent regulation of the acetylation and deacetylation of Smad7 are warranted. Together with our previous findings, the data presented in the current report support a model where the activity of p300 and one or more HDACs determine the stability of Smad7 (Fig. 7). It is an interesting possibility that the stability of other proteins is also regulated by the balance between acetylation, deacetylation, and ubiquitination.