Site-specific Acetylation of the Proteasome Activator REGγ Directs Its Heptameric Structure and Functions*

Background: REGγ mediates degradation of numerous proteins. How REGγ activity is regulated remains unclear. Results: REGγ acetylation at lysine 195 promotes its activity by enhancing monomeric interactions and heptameric formation of the REGγ molecules. Conclusion: Site-specific acetylation of REGγ is important for its structural architecture and enzymatic function. Significance: Our discovery provides basis for additional venue to intervene the REGγ proteasome function. The proteasome activator REGγ has been reported to promote degradation of steroid receptor coactivator-3 and cyclin-dependent kinase inhibitors p21, p16, and p19 in a ubiquitin- and ATP-independent manner. A recent comparative analysis of REGγ expression in mouse and human tissues reveals a unique pattern of REGγ in specific cell types, suggesting undisclosed functions and biological importance of this molecule. Despite the emerging progress made in REGγ-related studies, how REGγ function is regulated remains to be explored. In this study, we report for the first time that REGγ can be acetylated mostly on its lysine 195 (Lys-195) residue by CREB binding protein (CBP), which can be reversed by sirtuin 1 (SIRT1) in mammalian cells. Site-directed mutagenesis abrogated acetylation at Lys-195 and significantly attenuated the capability of REGγ to degrade its target substrates, p21 and hepatitis C virus core protein. Mechanistically, acetylation at Lys-195 is important for the interactions between REGγ monomers and ultimately influences REGγ heptamerization. Biological analysis of cells containing REGγ-WT or REGγ-K195R mutant indicates an impact of acetylation on REGγ-mediated regulation of cell proliferation and cell cycle progression. These findings reveal a previously unknown mechanism in the regulation of REGγ assembly and activity, suggesting a potential venue for the intervention of the ubiquitin-independent REGγ proteasome activity.

The proteasome activator REG␥ (also known as PA28␥, PSME3, Ki antigen) belongs to the REG or 11 S family of proteasome activators that have been shown to bind and activate the 20 S proteasome (1,2). REG␥ has been reported to promote degradation of some important regulatory proteins such as steroid receptor coactivator-3 and cyclin-dependent kinase inhibitors p21, p16, and p19 in a ubiquitin-and ATP-independent manner (3)(4)(5). Moreover, REG␥ facilitates the turnover of tumor suppressor p53 by promoting MDM2-mediated p53 ubiquitination (6) and regulating p53 cellular distribution (7). Furthermore, REG␥ is overexpressed in some cancers (8,9) and is linked to multiple cancer-related pathways (10). A unique expression pattern of REG␥ in cell specific manner has been documented, suggesting undisclosed functions and biological importance of this molecule (11). Despite recent progress made in this field, how REG␥ is regulated in mammalian cells is largely unknown.
Post-translational modification is an important process in regulating protein structures and functions. Acetylation occurs as a co-translational and post-translational modification of histones and non-histone proteins such as p53 and tubulins (12). In fact, proteomic studies have identified thousands of acetylated mammalian proteins (13,14), of which chromatin proteins and enzymes are highly represented. Acetylation commonly occurs at a lysine residue and can affect protein nuclear localization, stability, transcriptional activity, DNA binding, and interactions with other proteins and cofactors (12,15), indicating that acetylation has a considerable impact on protein functions. Several studies suggest that acetylation can alter protein structures or protein-protein interactions (16 -18). For example, acetylation of KLF5 transcription factor enhances its interaction with Smad4 to promote transcription of target genes (16). Thompson et al. (17) demonstrate that acetylation of the putative inhibitory loop of p300 may open the locked gate and activate its acetyltransferase activity.
Protein acetylation is a reversible process that is governed by the opposing actions of histone acetyltransferases and histone deacetylases. CBP 4 and p300 (E1A binding protein p300) possess strong histone acetyltransferase activity and act on both histone and non-histone proteins (19,20). Histone deacetylases are classified into four classes and two families: classical (classes I, II, and IV) and Sir2 (silent information regulator 2)-related protein (sirtuin) families (class III) (21). Among the seven members of mammalian sirtuins (SIRT1-7), SIRT1 is the most studied and strongly implicated in cellular regulation through its deacetylase activity (22).
In this study, we illustrate that acetylation of REG␥ at the lysine 195 residue by CBP is important for the degradation of REG␥ substrates, such as p21 and HCV core proteins. However, SIRT1, a deacetylation enzyme, can interact with REG␥ and remove acetylation group at Lys-195, attenuating REG␥ activity. Further study reveals that blocking acetylation at Lys-195 significantly reduces interactions between REG␥ monomers and ultimately influences the formation of heptamer. Finally, functional analysis in cells containing REG␥-WT or REG␥-K195R mutation has validated the crucial role of acetylation in REG␥-mediated regulation of cell proliferation and cell cycle progression.
Plasmid Constructs and Site-directed Mutagenesis-The mammalian expression vector pCDNA5/FRT/TO (Invitrogen) was modified to express REG␥ or FLAG-tagged REG␥ at the N terminus. HA-tagged REG␥ and HCV core-173 constructs were generated in the pSG5 vector. pCDH-CMV-EF1-REG␥ was constructed by inserting a digested PCR fragment into the lentivirus expression vector pCDH-CMV-EF1-Puro. GSTtagged REG␥ was generated in pGEX-4T-1 vector. pPAL7-REG␥ was constructed into pPAL7 vector. His-SIRT1 was generated in pET28a vector. pCDNA3.1-p21 was generated into the pCDNA3.1 vector. pCDNA FLAG-CBP was kindly provided by Dr. Qin Feng (Department of Molecular and Cellular Biology, Baylor College of Medicine), pCDNA3 FLAG-SIRT1, pCDNA3 SIRT1, and pCDNA3 SIRT1 H363Y were provided by Dr. Qiang Tong (Departments of Pediatrics, Medicine, Molecular Physiology & Biophysics, Baylor College of Medicine). Lysine-to-arginine mutations in REG␥ or its FLAG/HA-tagged versions were generated by site-directed mutagenesis at residues Lys-6, Lys-14, and Lys-195. All of the constructs were verified by DNA sequencing.
Mass Spectrometry-The HEK293 FLAG-REG␥ inducible cells were treated with doxycycline 1 g/ml for 48 h to induce highly expressed FLAG-REG␥. The cells were lysed with lysis buffer (50 mM Tris-HCl (pH 7.5), 150 mM NaCl, 1 mM EDTA, 0.5% Nonidet P-40). FLAG-REG␥ was immunoprecipitated from precleared cell lysates by incubation with anti-FLAG M2 Affinity Gel overnight at 4°C. The immunoprecipitates were washed three times with NETN buffer (20 mM Tris-HCl (pH 8.0), 100 mM NaCl, 1 mM EDTA, 0.5% Nonidet P-40). The washed beads were boiled with SDS-PAGE loading buffer. FLAG-REG␥ were resolved by SDS-PAGE and stained with Coomassie Blue. FLAG-REG␥ protein bands were cut, destained, and digested with trypsin or V8 protease overnight in the NH 4 HCO 3 buffer. Mass spectrometry was performed as described previously (5). Recombinant REG␥ protein was purified by Profinity eXact TM Protein Purification System (Bio-Rad). Recombinant REG␥ was resolved by SDS-PAGE and stained with Coomassie Blue. After in-gel digestion and peptide extraction, the peptides were analyzed on an Orbitrap Elite spectrometer connected to an EASY-nLC 1000 UPLC system using a nanoelectrospray ion source (Thermo Scientific).
Immunoprecipitation and Detection of REG␥ Acetylation in Cells-To detect REG␥ acetylation in cells, HEK293/A549/ HeLa cells were harvested and lysed in the lysis buffer described above. The lysates were centrifuged, and the supernatants were incubated with 1 g of anti-REG␥ antibody overnight at 4°C with 1 g of rabbit IgG (Santa Cruz Biotechnology) as a control. Each sample was incubated with protein A/G Plus-agarose beads (Santa Cruz Biotechnology) for 3 h. Immunoprecipitated REG␥ was resolved on a 10% SDS-PAGE gel and analyzed by Western blotting with anti-AcK antibody. The nitrocellulose membrane was stripped by the Restore TM Western blot Stripping Buffer (Thermo) and then probed with anti-REG␥ antibody. For reciprocal immunoprecipitation, HEK293 cell lysate was incubated with anti-AcK antibody overnight at 4°C and then incubated with protein A/G plus-agarose beads for 3 h. The immunoprecipitates were resolved on a 10% SDS-PAGE gel and analyzed by Western blotting with anti-REG␥ antibody. FLAG-tagged REG␥ was immunoprecipitated from 293 REG␥inducible cells with anti-FLAG M2 affinity gel. FLAG-REG␥ was eluted with 200 g/ml FLAG peptide for 1 h. FLAG-REG␥ was incubated with anti-AcK antibody overnight at 4°C in 80 l of buffer (50 mM Tris, 137 mM NaCl, 1 mM EDTA, 10 mM NaF, 0.1 mM Na 3 VO 4 , 1 mM DTT, 10% glycerol, 0.5% Nonidet P-40, pH 7.8), together with 50 mM sodium butyrate, 6.6 M trichostatin A (TSA), 10 mM nicotinamide (NAM), and protease inhibitors. The mixture was incubated with protein A/G-agarose beads for 3 h. The supernatant unbound REG␥ and the immunoprecipitated acetylated REG␥ was resolved on a SDS-PAGE gel and analyzed by Western blotting with anti-REG␥ antibody. For deacetylase inhibition, the HEK293 FLAG-REG␥ inducible cells were treated with 1 g/ml doxycycline (for 48 h), 6.6 M TSA, and 10 mM NAM for 6 h prior to cell harvest. Cells were lysed in FLAG lysis buffer (50 mM Tris, 137 mM NaCl, 1 mM EDTA, 10 mM NaF, 0.1 mM Na 3 VO 4 , 1% Nonidet P-40, 1 mM DTT, 10% glycerol, pH 7.8) containing fresh protease inhibitors (Roche Applied Science), 6.6 M TSA, and 10 mM NAM. Cell extracts were immunoprecipitated with anti-FLAG M2 affinity gel. REG␥ acetylation was analyzed by Western blotting with a Pan-anti-AcK antibody.
REG␥ Acetylation and Deacetylation Assays in Vitro-FLAG-tagged CBP was expressed in HEK293T cells and immunoprecipitated by anti-FLAG M2 affinity gel. FLAG-CBP was eluted with 200 g/ml FLAG peptide (Sigma) for 1 h. GSTtagged REG␥ was expressed and purified as described previously (5). 5 g of GST-REG␥ was incubated with 3 g of FLAG-CBP in 30 l of histone acetyltransferase buffer (250 mM Tris-HCl, pH 8.0, 500 M EDTA, 5 mM DTT, 50% glycerol, 50 mM sodium butyrate, 6.6 M TSA, 10 mM NAM, and protease inhibitors) with or without addition of 5 mM acetyl-CoA. After 3 h at 30°C, samples were resolved on SDS-PAGE and analyzed by Western blot. In vitro deacetylation assays were performed as follows. FLAG-REG␥ was immunoprecipitated from HEK293 FLAG-REG␥ inducible cells transfected with CBP, and the enriched REG␥ proteins were eluted with FLAG peptide. Acetylated FLAG-REG␥ was incubated with recombinant His-SIRT1 in 50 l of deacetylation buffer (50 mM Tris-HCl, pH 8.0, 50 mM NaCl, 4 mM MgCl 2 , 0.5 mM DTT, 0.1 mM PMSF, 5% glycerol, 0.02% Nonidet P-40, and protease inhibitors) in the presence or absence of 50 M NAD ϩ for 2 h at 30°C. The reactions were subjected to SDS-PAGE and analyzed by Western blot.
Transfections and REG␥ Activity Detection-Plasmid CBP (1 g) was transiently transfected into HEK293 cells for 32 h using FuGENE HD DNA transfection reagent (Roche Applied Science). REG␥ acetylation levels were detected as described above. Additional experiments were performed in the HEK293 FLAG-REG␥-inducible cells by expressing CBP in the presence or absence of SIRT1, followed by determination of REG␥ acetylation. In HeLa cells, SIRT1-WT and the deacetylase mutant, SIRT1-H363Y, were ectopically expressed for 32 h, and protein expressions were examined by Western blotting using anti-SIRT1, anti-REG␥, and anti-p21. For RNA interference, HEK293 cells were transfected with 60 nM ON-TARGET plus siRNA specific for human CBP using Lipofectamine TM 2000 transfection reagent (Invitrogen). After 72 h of transfection, cells were harvested and lysed for Western blot analysis of CBP, REG␥, and p21 protein levels. The p21 and/or HCV core protein levels were measured by Western blotting to estimate REG␥ activity. In the cells, the p21 or HCV core construct was either co-transfected with a control vector or REG␥-WT, REG␥-K6R, REG␥-K14R, REG␥ K195R, or REG␥ K195Q. HEK293 REG␥-inducible cells were incubated with 100 mg/ml cycloheximide with or without additional of doxycycline (1 g/ml, for 48 h) at indicated time for p21 protein half-life analysis.
Co-immunoprecipitation and Protein-Protein Interaction Analysis-HEK293 cells were cultured in the presence of 20 M MG132 for 8 h before harvest. REG␥ was immunoprecipitated from precleared cell lysates with 1 g of anti-REG␥ overnight at 4°C, with 1 g of rabbit IgG in the control group. Each sample was incubated with protein A/G Plus-agarose beads for 3 h. The immunoprecipitates were washed three times with immunoprecipitation (IP) washing buffer (50 mM Tris, 137 mM NaCl, 1 mM EDTA, 10 mM NaF, 0.1 mM Na 3 VO 4 , 1 mM DTT, 10% glycerol, pH 7.8), and the co-immunoprecipitated proteins were analyzed by Western blotting with a SIRT1 antibody. The nitrocellulose membrane was stripped by the stripping buffer and then probed with anti-REG␥ as references. For the reciprocal co-immunoprecipitation, HEK293 cell lysates were incubated with a SIRT1 antibody followed by detection of REG␥ in the immunoprecipitates with Western blot analysis. For interactions between REG␥ monomers, HEK293 cells were transfected with HA-tagged and/or FLAG-tagged REG␥-WT or K195R mutant for 48 h, along with MG132 (20 M for 8 h) before harvest. The cell lysates were incubated with anti-FLAG M2 affinity gel overnight at 4°C. The coimmunoprecipitated HA-REG␥ were examined by Western blot with an anti-HA antibody.
Analysis of REG␥ Heptamer by Native PAGE and FPLC-The HEK293 FLAG-REG␥-WT or K195R inducible cells were cultured in the presence of doxycycline (1 g/ml, for 48 h), 6.6 M TSA, and 10 mM NAM (for 6 h) prior to harvest. Following cell lysis, the supernatants were prepared in NativePAGE TM sample buffer (Invitrogen) and resolved by NativePAGE TM Novex® Bis-Tris gel system (Invitrogen). Transferred PVDF membranes were stained with Ponceau S to display the Native-Mark TM unstained protein standard (Invitrogen). After decolorization, the PVDF membrane was probed with a FLAG antibody to detect the REG␥ heptamer complex. HEK293 cells were transiently transfected with 1 g of FLAG-REG␥-WT, K195R, K195Q, or K188F (a heptamerization-defective mutant) into a six-well plate for 32 h, REG␥ heptamer complexes were detected by native PAGE as described above. To investigate endogenous REG␥ heptamerization, HEK293 FLAG-REG␥-inducible cells cultured in the presence or absence of TSA, NAM, or resveratrol were fractionated through a Superose 6 column. The elution profiles of REG␥ were monitored by Western blotting. The FPLC was performed as described (5).
MTS Assay and Cell Cycle Analysis-MTS assay was performed by seeding HEK293 FLAG-REG␥-inducible cells in a 96-well plates at 3 ϫ 10 3 cells per well and were cultured for 5 days. Doxycycline (1 g/ml) was added to cells at day 1. Cells were incubated with MTS solution at 37°C for 2 h, and the absorbances (490 nm) were measured and analyzed. Cell cycle analysis was carried out by estimating DNA contents with flow cytometry. HEK293 FLAG-REG␥-inducible cells were fixed in ice-cold 70% ethanol, incubated overnight at Ϫ20°C, and stained with propidium iodide/Triton X-100-containing RNase A solution for 15 min at 37°C. Cell cycle analysis was performed by BD CantoII Cell Analyzers. The data were analyzed using Flowjo software.

REG␥ Is Acetylated in Mammalian
Cells-Recent studies reveal a variety of physiological functions of REG␥ in growth, cell proliferation, and cancer progression (9 -11, 23, 24). How REG␥ function is regulated remains to be elucidated. Our previous proteomic analysis (5) indicates REG␥ as a protein with different post-translational modification. To determine whether REG␥ could be an acetylated protein, we carried out IP of REG␥ in lysates from different mammalian cells followed by Western blot analysis using a general anti-acetyl-lysine (anti-AcK) antibody. In HEK293, endogenous REG␥ is clearly recognized in the IP enriched sample by the anti-AcK antibody (Fig.  1A). Similarly, in human cancer cell lines, including A549 and HeLa, REG␥ protein acetylation can also be detected (Fig. 1, B and C). Alternatively, we performed a reciprocal IP analysis with the anti-AcK antibody followed by Western blotting using anti-REG␥. The result showed a single band at the molecular weight identical to the size of REG␥, indicative of an acetylated REG␥ (Fig. 1D). To further test our observation, we treated cells with histone deacetylase inhibitors TSA and NAM (25,26) to enhance cellular levels of acetylated proteins. In the doxycycline-induced FLAG-REG␥ over-expressing HEK293 cells, we found a notable increase FIGURE 1. REG␥ is acetylated in mammalian cells. Endogenous REG␥ in HEK293 (A), A549 (B), or HeLa cells (C) were immunoprecipitated with anti-REG␥ antibody. Acetylation of REG␥ was detected by immunoblotting with an anti-AcK antibody. D, a reciprocal immunoprecipitation was performed with the anti-AcK antibody using lysates from HEK293 cells, and immunoprecipitated REG␥ was examined by immunoblotting with anti-REG␥ antibody. E, HEK293 cells inducibly expressing FLAG-REG␥ were cultured in the presence of doxycycline (1 g/ml) for 48 h, with 6.6 M TSA and 10 mM NAM for 6 h before harvesting the cells. FLAG-REG␥ was immunoprecipitated by FLAG M2 affinity gel. REG␥ acetylation was examined as in A-C. Asterisk refers to nonspecific bands.
in the acetylation of FLAG-REG␥ with TSA/NAM treatment (Fig. 1E). Taken together, these results suggest that REG␥ can be acetylated in mammalian cells.
Lys-195 Is a Major Acetylation Site in REG␥-To identify the potential acetylated residues, doxycycline-induced FLAG-REG␥ in HEK293 cells was immunoprecipitated with FLAG beads, resolved by SDS-PAGE, and analyzed by MS/MS (LC-MS/MS). MS/MS spectrum showed acetylation at three lysine residues at positions 6 (Lys-6), , and 195 (Lys-195), respectively ( Fig. 2A and data not shown). This was endorsed in part by a different mass spectrometry study with identification of acetylation at Lys-195 (13). The mass score (score difference average) for REG␥ acetylation at 195 is above the mean average of nearly 2000 acetylated proteins identified (13), reflecting relative abundance of this site specific acetylation. As a control, the recombinant REG␥ protein purified from Escherichia coli The precursor ion m/z showed a mass shift of 42.01 Da, b3, b4, b7-10, and y3-7, y9, y10 fragment ions were found in MS/MS spectrum. The acetylated peptide hits were filtered by 1% false discovery rate at protein, peptide, and site level. C, FLAG-tagged REG␥ was purified from 293 REG␥-inducible cells by FLAG peptide. Acetylated REG␥ was immunoprecipitated by anti-AcK antibody. The unbound REG␥ and bound acetylated REG␥ were examined by immunoblotting with anti-FLAG antibody. D, HEK293 cells inducibly expressing FLAG-REG␥ were treated with 1 g/ml doxycycline for 48 h. FLAG-REG␥ WT and mutants were immunoprecipitated with anti-FLAG M2 affinity gel, and acetylation status was examined by immunoblotting with anti-AcK antibody. E, in HEK293 cells, transiently expressed FLAG-REG␥-WT and FLAG-REG␥ K195R were immunoprecipitated with anti-FLAG M2 affinity gel, and acetylations were determined by immunoblotting with anti-AcK antibody.
was resolved by SDS-PAGE and cut out for LC-MS/MS analysis. To our surprise, bacterially generated recombinant REG␥ was also acetylated at Lys-195 (Fig. 2B). Sequence analysis of REG␥ from multiple species reveals that the Lys-195 site and surrounding residues are highly conserved across the animal kingdom ( Fig. 2A, lower panel), with Lys-6 and Lys-14 slightly less conserved among vertebrates ( Fig. 2A, upper panel). To define the ratio of acetylated REG␥ in total molecules, we purified FLAG-tagged REG␥ from 293 REG␥-inducible cells followed by the reciprocal immunoprecipitation with anti-AcK antibody. We found that about half of the REG␥ molecules are acetylated (Fig. 2C, lanes 4 and 5) compared with the unbound REG␥ protein (Fig. 2C, lanes 2 and 3). To analyze acetylation, normally Lys3 Gln (KQ) substitution is used to mimic lysine acetylation, whereas Lys3 Arg (KR) substitution is to eliminate acetylation target site without neutralization of the positive charges (27)(28)(29). Therefore, we generated doxycycline-inducible REG␥-expressing HEK293 cells, including acetylation-defective mutations of K6R, K14R, or K195R in FLAG-REG␥. In these stable cell lines, the K195R mutation significantly reduced REG␥ acetylation levels, whereas K6R and K14R mutations had little impact on the overall acetylation in REG␥ (Fig. 2D). A similar result was found in HEK293 cells transiently expressing FLAG-REG␥-K195R mutant, which was poorly recognized by the anti-AcK antibody compared with FLAG-REG␥-WT (Fig.  2E). These results emphasize that Lys-195 is a major acetylation site in REG␥, reflecting a potentially important role of this lysine residue in modulating REG␥ activity.
Blocking Acetylation at  Attenuates REG␥ Activity-To understand the biological consequence of REG␥ acetylation, expressions of p21 and a truncated HCV core-173 protein, well known substrates of the REG␥ proteasome (3,4,30), are utilized to evaluate REG␥ activity. We analyzed the ability of REG␥ acetylation-defective mutants K6R, K14R, and K195R to promote degradation of p21 and HCV core-173 in HEK293 cells as well as the human lung cancer cell line H1299. The latter cell line lacks p53 expression and therefore avoids the impact of REG␥ on p21 degradation through down-regulation of p53 (6). Although REG␥ mutations at Lys-6 or Lys-14 had minor effect to attenuate REG␥-mediated degradation of p21 protein (data not shown), REG␥ mutation at Lys-195 significantly reduced the capacity of REG␥ to degrade HCV core-173 and p21 proteins in H1299 cells (Fig. 3, A, lane 3, and B, lane 3). As controls, REG␥-WT retained the activity to promote degradation of HCV core-173 and p21 proteins (Fig. 3, A, lane 2, and B, lane 2). In the REG␥ inducible HEK293 cells, we also obtained results similar to what we observed in H1299 cells (data not shown), indicating that acetylation of REG␥ occurs in different cell types. Next, we generated an acetylation-mimetic REG␥ K195Q mutant, which is capable of accelerating p21 degradation in comparison with REG␥ K195R in H1299 cells, but to a less extent compared with REG␥-WT (Fig. 3B). To understand the action of endogenous REG␥ acetylation mutants in cells without wild type REG␥, we generated REG␥ derivatives in lentivirus, including REG␥-WT, K195R, K195Q, along with a vector control, and stably integrated these constructs in REG␥ Ϫ/Ϫ MEF cells. Despite that stable integration of REG␥-WT in REG␥ Ϫ/Ϫ MEF cells does not function efficiently in its degradation of endogenous p21 probably due to compensation mechanisms, the stable cells expressing REG␥-WT and REG␥-K195Q in the parental REG␥ Ϫ/Ϫ MEFs had comparable effects on p21 degradation. However, REG␥-K195R showed a remarkable inhibition of p21 protein degradation (Fig. 3C). Even though the Lys3 Gln substitution does not always faithfully mimic the acetylation status of the lysine residue (29,31), our data indicate that acetylation at Lys-195 is crucial for maintaining REG␥ activity in its degradation of target proteins.
To explore whether acetylation of REG␥ at Lys-195 affects its function in the turnover of substrate protein p21, we used REG␥-inducible HEK293 cells treated with cycloheximide for indicated time periods. As expected, p21 degradation was expedited when overexpressing REG␥ WT (Fig. 3D). However, induced overexpression of REG␥ K195R mutant had no significant impact on the decay rate of p21 (Fig. 3E). In a parallel experiment, induced overexpression of REG␥ K195Q could effectively decrease p21 half-life as REG␥ WT did (Fig. 3F). Taken together, these results indicate that acetylation at Lys-195 is critical for retaining REG␥ activity.
CBP and SIRT1 Reversely Regulate REG␥ Acetylation and Activity-CBP and p300 are transcriptional coactivators with intrinsic histone acetyltransferase activity (32)(33)(34)(35)(36) to regulate gene expression. To test whether REG␥ could be acetylated by these histone acetyltransferases, CBP was transiently expressed into HEK293 cells. We found significantly increased acetylation level of REG␥ with CBP co-transfection (Fig. 4A). In addition, the recombinant GST-tagged REG␥ protein could be acetylated by FLAG-tagged CBP in the presence of acetyl-CoA in vitro (Fig. 4B). Next, we investigated whether REG␥ is a substrate of the Sir2 families in mammalian cells. Among SIRT1-SIRT7, FLAG-tagged SIRT1 showed strong interaction with GFPtagged REG␥ (data not shown). Furthermore, we found robust interactions between endogenous SIRT1 and REG␥ in HEK293 cells by IP analysis using the anti-REG␥ antibody (Fig. 4C). Similarly, a reciprocal IP with anti-SIRT1 antibody detected endogenous co-IP of REG␥ and SIRT1 in HEK293 cells (Fig. 4D), indicating that REG␥ is a potential target of SIRT1. In the inducible FLAG-REG␥-WT expressing HEK293 cells, augmented CBP expression enhanced acetylation in the FLAG-REG␥, whereas co-expressing CBP and SIRT1 blocked the effect by CBP, reflecting a causal relation between SIRT1 and REG␥ deacetylation (Fig. 4E, left panel). In contrast, transient overexpressing CBP failed to significantly enhance REG␥ acety-lation in the FLAG-REG␥-K195R-inducible HEK293 cells and co-expressing SIRT1 and CBP further diminished REG␥ acetylation (Fig. 4E, right panel). In addition, we found an obviously increased acetylation level of REG␥ after SIRT1 knocking down in 293T cells (Fig. 4F). As expected, when acetylated FLAG-REG␥ was incubated with recombinant His-SIRT1 and NAD ϩ , REG␥ acetylation level was obviously reduced in vitro (Fig. 4G). These results strongly suggest that CBP and SIRT1 mainly target Lys-195 for acetylation/deacetylation in REG␥, although we cannot exclude their regulation in other residues in REG␥.
Furthermore, we determined whether the influence of SIRT1 on REG␥ activity depends on its deacetylase activity. Transient overexpression of SIRT1-WT in HeLa cells significantly increased p21 protein level, whereas exogenously expressed SIRT1-H363Y, a deacetylase-defective mutant (37), failed to inhibit REG␥-dependent degradation of p21 (Fig. 4H). Consistently, we also found that silencing CBP in HEK293 cells enhanced p21 protein levels (Fig. 4I). Taken together, these results suggest that CBP and SIRT1 can regulate REG␥ activity through acetylation and deacetylation in REG␥ at specific sites.
Acetylation at Lys-195 Is Crucial for the Monomeric Interactions and Overall Structure of REG␥-Based on the discovery that acetylation in REG␥ influences its activity, we intended to address how this could be achieved. Previous studies demonstrate that functional REG␥ exists as a heptameric ring in cells (8,38,39). We tested whether acetylation is involved in regulation of the overall structure of REG␥. In FLAG-REG␥-WToverexpressing HEK293 cells, REG␥ heptamers were easily detected by native PAGE followed by membrane transferring and antibody blotting (Fig. 5A, lane 1). Interestingly, enhancing REG␥ acetylation by TSA and NAM treatment greatly increased the amount of REG␥ heptamer complexes (Fig. 5A,  lane 2). In FLAG-REG␥-K195R-overexpressing HEK293 cells, REG␥ heptamer formation was dramatically suppressed even in the presence of TSA and NAM treatment (Fig. 5A, lanes 3 and  4). Furthermore, we transiently expressed FLAG-REG␥-WT and corresponding acetylation mutants in HEK293 cells to examine REG␥ heptamerization by native PAGE. The data clearly showed that FLAG-REG␥-WT and the acetylation-mimetic FLAG-REG␥-K195Q mutant formed heptamers in cells (Fig. 5B, lanes 2 and 4). In contrast, the acetylation-defective FLAG-REG␥-K195R mutant and the heptamerization-defective FLAG-REG␥-K188F mutant (38) had poor REG␥ heptamer formation (Fig. 5B, lanes 3 and 5). The crystal structure of heptameric REG␣ (40), which is highly homologous to REG␥, suggests a parallel intermolecular interactions between helix 2 of one monomer with the helix 4 of the neighboring molecule. We then questioned whether acetylation affects association between REG␥ monomers. By transiently expressing a FLAGtagged REG␥ and an HA-tagged REG␥ construct or corresponding mutant constructs in HEK293 cells followed by IP and Western blot analysis (Fig. 5C), we found that REG␥-WT interacted better with each other, whereas REG␥-K195R monomeric interactions were compromised, suggesting a role of acetylation at this position in REG␥ protein-protein interactions. To further test the impact of acetylation on REG␥ monomeric interactions, we extended part of the above experiments described in Fig. 5C with additional treatment of either TSA/ NAM or SIRT1 co-transfection, alone or in combination. As expected, TSA/NAM treatment enhanced the interactions between FLAG-REG␥ and HA-REG␥ (Fig. 5D, compare lane 5 with lane 6), whereas SIRT1 attenuated their association in the presence or absence of TSA/NAM (Fig. 5D Moreover, we performed size exclusion chromatography using lysates from the inducible FLAG-REG␥-WT and FLAG-REG␥-K195R HEK293 cells. Based on the molecular standard, we found that majority of the FLAG-REG␥-WT formed heptamers with a peak at ϳ230 kDa (Fig. 5E, upper panel). Upon TSA/ NAM treatment and increased REG␥ acetylation, a change in fraction pattern occurred with a shift toward the peak fraction, indicating an increase in REG␥ heptamerization (Fig. 5E, middle panel). On the contrary, FLAG-REG␥-K195R expressing cells produced a stretched elution pattern ranging from monomers (ϳ 30 kDa), various degrees of oligomers, and reduced higher molecular weight fractions (heptamers) (Fig. 5E, lower  panel). Similar results were obtained following REG␥ deacetylation by resveratrol in the induced FLAG-REG␥-WT overex- . CBP and SIRT1 regulate REG␥ acetylation and activity. A, HEK293 cells were transfected with CBP for 32 h, and acetylation of endogenous REG␥ was examined by immunoblotting with anti-AcK antibody. B, the recombinant GST-REG␥ (5 g) were incubated with FLAG-CBP (3 g) in 30 l of histone acetyltransferase buffer for 3 h at 30°C. The acetylation level of GST-REG␥ was examined by immunoblotting. C, endogenous REG␥ protein in HEK293 cells was immunoprecipitated with anti-REG␥ antibody and IgG as a control. The coimmunoprecipitated SIRT1 protein levels were determined by immunoblotting with anti-SIRT1 antibody. D, endogenous SIRT1 protein in HEK293 cells was immunoprecipitated with an anti-SIRT1 antibody, and the coimmunoprecipitated REG␥ protein levels were determined by immunoblotting with anti-REG␥ antibody. E, CBP and SIRT1 were transfected into HEK293 cells inducibly expressing FLAG-REG␥-WT or FLAG-REG␥-K195R as indicated. Acetylation levels of REG␥ were examined by immunoblotting with anti-AcK antibody after immunoprecipitation of FLAG-REG␥ with anti-FLAG M2 affinity gel. F, endogenous REG␥ acetylation levels were examined in SIRT1 knocking down 293T cells by immunoblotting with anti-REG␥ antibody. G, acetylated FLAG-REG␥ was incubated with recombinant His-SIRT1 in 50 l of deacetylase buffer for 2 h at 30°C. REG␥ acetylation level was analyzed by immunoblotting. H, HeLa cells were transfected with SIRT1 or SIRT1-H363Y mutant for 32 h. Endogenous p21 and REG␥ protein levels were analyzed by immunoblotting with anti-p21 or anti-REG␥ antibody. I, HEK293 cells were transiently transfected with siRNA against CBP for 72 h. CBP knockdown efficiency and REG␥ p21 protein levels were examined by immunoblotting. CHX, cycloheximide; DOX, doxycycline; Ctrl, control. pressing cells (data not shown). Based on the results that acetylation-defective mutant significantly impaired its heptamerization and monomers association (Fig. 5, A-C), we examined whether REG␥-K195R mutant is unstable, which may be targeted for degradation or association with other protein complexes. Consequently, we found that acetylation-defective mutant REG␥-K195R degrades faster compared with REG␥-WT (Fig. 5F). Collectively, these results demonstrate acetylation at Lys-195 is crucial for REG␥ monomeric interactions and assembly of REG␥ heptameric complexes.
Acetylation-defective Mutant REG␥ Impairs Cell Growth and Cell Cycle Progression-As a broad acting cyclin-dependent kinase inhibitor, p21 plays a central role in cell cycle regulation in many cells types (41,42). Recent studies indicate that REG␥ influences cell cycle through degradation of several cell cycle regulators, including p21, p16, and p19 (3,4). Thus, we reasoned that blocking REG␥ acetylation at Lys-195 may also affect cell proliferation and cell cycle progression. Using the HEK293 cells inducibly expressing FLAG-REG␥-WT and FLAG-REG␥-K195R for the cell proliferation assay, we observed that cells overexpressing FLAG-REG␥-K195R had significantly reduced growth rate at day 2 through day 5 (Fig.  6A). Next, these cells were also subjected to flow cytometry analysis to evaluate the impact of acetylation defect on cell cycle progression. Compared with FLAG-REG␥-WT expressing cells, FLAG-REG␥-K195R expressing cells had an increased population at the G 0 /G 1 phase, and a significantly decreased proportion of S phase cells (Fig. 6B), indicating a reduced cell cycle progression from G 0 /G 1 to S phase transition. Taken together, our results further substantiate an important role for REG␥ acetylation in cell growth and cell cycle regulation.

DISCUSSION
Emerging evidence leads to a renewed attention to the ubiquitin-independent REG␥-proteasome pathway. With the discovery of more and more substrates in this protein degradation pathway (3-6, 30, 43, 44), REG␥ is being recognized as an important regulator. Yet, the regulatory input that may alter the biological function of REG␥ remains poorly understood. In this study, we demonstrate that site-specific acetylation of REG␥ at Lys-195 is crucial for its monomeric interactions and formation of a functional heptameric complex. Modulating acetylation status at this lysine residue has profound impact on the activity of REG␥ in substrate protein degradation. Our results provide the first biochemical evidence for a role of acetylation in structural and functional regulation of the REG␥-proteasome complex.
In eukaryotic cells, acetylation is among the most common covalent modifications and ranks among the most important master switches similar to phosphorylation (45). It is now clear that prokaryotes have the capacity to acetylate both the ␣amino groups of N-terminal residues and the ⑀-amino groups of lysine side chains, suggesting that acetylation appears to be an FIGURE 5. Acetylation at Lys-195 is crucial for REG␥ heptamerization and interactions between REG␥ monomers. A, HEK293 cells inducibly expressing FLAG-REG␥ were cultured with 1 g/ml doxycycline for 48 h, in the presence or absence of TSA and NAM for 6 h before cells were harvested. REG␥ heptamerization was determined by native PAGE gel system with anti-FLAG antibody. B, FLAG-REG␥-WT, K195R, K195Q, or K188F was transfected into HEK293 cells for 32 h. Cell lysates were analyzed by native PAGE Gel system to examine REG␥ heptamerization. C, FLAG-REG␥ and HA-REG␥ constructs were transfected into HEK293 cells as indicated. Cell lysates was immunoprecipitated by FLAG M2 affinity gel. Co-immunoprecipitated HA-REG␥ was detected by immunoblotting with an anti-HA antibody. D, HEK293 cells were treated with TSA/NAM or TSA/NAM along with transient expression of SIRT1 as indicated, together with co-expression of different tagged REG␥ derivatives. Co-immunoprecipitated HA-REG␥ was detected by immunoprecipitating FLAG-REG␥ and immunoblotting with indicated antibodies. E, cell lysates from the FLAG-REG␥ inducible HEK293 cells were subjected to size exclusion chromatography for FPLC analysis. ancient reversible modification such as phosphorylation (46). Reversible protein acetylation provides key regulatory switches for cell signaling pathways, which has been shown to affect a diverse array of biochemical properties, including protein activity, protein stability, DNA/protein-protein interactions, and intracellular localization (12,15). We have defined lysine 195 in REG␥ as the mostly affected residue for acetylation/ deacetylation mediated by CBP/SIRT1. The REG␥ K195R mutant that can no longer be acetylated failed to promote its substrate degradation, whereas the K195Q mutant that mimics a constitutively acetylated state retained the capacity to degrade target proteins, which further correlates with REG␥-dependent regulation in cell growth and cell cycle progression. We also demonstrated reversible acetylationdeacetylation modification at the Lys-195 sites. Our results by no means exclude the possibility of weak acetylation on other sites or regulation by other histone acetyltransferases/ histone deacetylases. In fact, Lys-6 and Lys-14 in REG␥ can also be acetylated based on mass spectrometry and bioinfor-matics prediction. We have previously demonstrated that SUMOylation of REG␥ can be enhanced in the presence of PIAS1 at Lys-6, Lys-12, and Lys-14, which results in cytoplasmic distribution and stabilization of REG␥ (47). It is likely that competition by SUMOylation at Lys-6 and Lys-14 attenuates acetylation at these sites.
Although we failed to generate a Lys-195-specific Ac antibody, the Pan-AcK antibody successfully detected acetylation at Lys-195 in REG␥-WT but not in REG␥-K195R mutant, suggesting high prevalence of this site-specific acetylation in mammalian cells. Based on our comparative analysis of score difference average (which is a standard for estimation of modification signal) for ϳ2000 of the acetylated proteins (13), we found the score for acetylation of REG␥ at Lys-195 is above the average score, indicating that acetylation of REG␥ at Lys-195 is around average levels compared with all proteins examined so far. In addition, our IP Western results with Pan-acetylation antibody suggest that about half of the molecules are acetylated. Because the REG␥ molecules form heptamers, it is likely that FIGURE 6. Acetylation mutation at Lys-195 in REG␥ reduces cell proliferation and mitigates cell cycle progression. A, HEK293 cells inducibly expressing FLAG-REG␥ were treated with doxycycline after cells were seeded in 96-well plates for 24 h. Absorbance was measured at indicated times. Data were analyzed as means Ϯ S.D. of spectrometric absorbance of three independent experiments. *, p Ͻ 0.05; **, p Ͻ 0.01; and ***, p Ͻ 0.001 represent the statistic comparisons between growth in HEK293 with FLAG-REG␥-WT and HEK293 with FLAG-REG␥-K195R. B, the HEK293 with FLAG-REG␥-WT and HEK293 with FLAG-REG␥-K195R were treated with doxycycline for 48 h, and DNA contents of the inducible cells in different cell cycles were analyzed by flow cytometry. Each bar indicates the distribution of the cell cycles. Data are reported as means Ϯ S.D. of three independent experiments. *, p Ͻ 0.05; **, p Ͻ 0.01; and ***, p Ͻ 0.001 indicate statistic differences between indicated groups. FIGURE 7. A working model simplifies the influences of acetylation on REG␥ assembly and function. CBP acetylates REG␥ at Lys-195, which promotes REG␥ to form a heptamer, resulting in augmented REG␥ activity. When SIRT1 binds with REG␥, it deacetylates REG␥ and inhibits REG␥ heptamerization, releasing monomers from REG␥ disassembly.
such an acetylation fraction may be enough for enhancing protein-protein interactions.
Given that REG␥ is highly homologous to REG␣, the predicted location of Lys-195 in REG␥ should be at the very C terminus of the helix 3 based on the structure of REG␣ (48). Facing the substrate interaction surface, Lys-195 should be easily accessed by enzymes such as SIRT1 to dynamically regulate the disassembly of REG␥. If interactions between REG␥ monomers only occur between helixes 2 and 4, we believe Lys-195 acetylation may induce favorable structure to facilitate this interaction. Acetylation in REG␥ remarkably enhances monomeric interactions and heptameric formation, which is consistent with previous reports that acetylation can enhance protein-protein interactions (49 -51). Whether acetylation of REG␥ at Lys-195 is a default or translationally coupled process remains to be investigated. Future crystal structure analysis of REG␥ may enable us to understand how this site-specific acetylation facilitates its protein-protein interactions.
Interestingly, the endogenously expressed acetylation-defective REG␥-K195R mutant dramatically impaired the heptameric complex, resulting in an elution pattern ranged from monomers, various degrees of oligomers, and reduced amount of heptamers (Fig. 5E). The results suggest that REG␥ indeed has intrinsic properties in self-association (48). It is likely that acetylation may accelerate the oligomerization processes of intracellular REG␥, which may be otherwise targeted for degradation or association with other protein complexes. In support of this idea, we found that acetylated REG␥ is more stable, whereas acetylation defective mutant degrades faster (Fig. 5F). Despite that cells stably or transiently expressing REG␥-K195R also have significant amount of endogenous wild type REG␥, we still observe significant impact of K195R mutant on proteinprotein interaction and proteolytic functions (Figs. 3 and 5).
To summarize our findings in this study, CBP acetylates REG␥ at Lys-195, which promotes its heptamerization and increases REG␥ activity in the degradation of targets proteins. In contrast, SIRT1 can bind with REG␥ and deacetylates REG␥, which inhibits heptamerization or trigger disassembly of REG␥, leading to inactivation of REG␥ capacity (Fig. 7). Our results provide a novel mechanism for the reciprocal regulation of REG␥ homeostasis by acetylation. As a potential druggable target, REG␥ activity may be modulated by histone acetyltransferase/SIRT inhibitors or activators.