Regulation of S-Adenosylhomocysteine Hydrolase by Lysine Acetylation*

Background: S-Adenosylhomocysteine hydrolase (SAHH) regulates methyltransferase reactions and is acetylated on two lysines. Results: We prepared acetylated semisynthetic SAHH and determined the impact of acetylation on structure and activity. Conclusion: Lys acetylation of SAHH changes hydrogen bonding patterns in its NAD+ binding regions and reduces its catalytic activity. Significance: Acetylation of SAHH may serve to regulate global alterations in cellular methylation. S-Adenosylhomocysteine hydrolase (SAHH) is an NAD+-dependent tetrameric enzyme that catalyzes the breakdown of S-adenosylhomocysteine to adenosine and homocysteine and is important in cell growth and the regulation of gene expression. Loss of SAHH function can result in global inhibition of cellular methyltransferase enzymes because of high levels of S-adenosylhomocysteine. Prior proteomics studies have identified two SAHH acetylation sites at Lys401 and Lys408 but the impact of these post-translational modifications has not yet been determined. Here we use expressed protein ligation to produce semisynthetic SAHH acetylated at Lys401 and Lys408 and show that modification of either position negatively impacts the catalytic activity of SAHH. X-ray crystal structures of 408-acetylated SAHH and dually acetylated SAHH have been determined and reveal perturbations in the C-terminal hydrogen bonding patterns, a region of the protein important for NAD+ binding. These crystal structures along with mutagenesis data suggest that such hydrogen bond perturbations are responsible for SAHH catalytic inhibition by acetylation. These results suggest how increased acetylation of SAHH may globally influence cellular methylation patterns.

DNA. It has been established previously that genetic knockdown of SAHH or its chemical inhibition by 3-deazaneplanocin A leads to accumulation of AdoHcy and generalized product inhibition of AdoMet-dependent methyltransferases (2,3). One of the enzymes that seems particularly sensitive to AdoHcy feedback inhibition is EZH2, the histone lysine methyltransferase component of the Polycomb repressor complex 2 responsible for Lys 27 methylation of histone H3, a signature mark of heterochromatic gene silencing (2)(3)(4). Recent studies indicate that SAHH inhibition by 3-deazaneplanocin A can reactivate developmental genes in acute myeloid leukemia cells, halting cellular proliferation (3,4).
SAHH is a highly conserved, NAD ϩ -dependent, homotetrameric enzyme that shows Ͼ70% amino acid sequence identity from yeast to human (18). The functional tetramer of SAHH is composed of four identical ϳ48-kDa subunits each of which contains an N-terminal substrate binding domain, a catalytic domain, and a C-terminal tail. Unlike typical NAD ϩ -dependent dehydrogenases, where each cofactor is bound by one monomer, NAD ϩ binding to SAHH requires contributions from neighboring subunits via their C-terminal tails (19 -22). The catalytic mechanism of SAHH involves NAD ϩ -mediated oxidation of the 3Ј-hydroxy group to the ketone followed by ␤-elimination of homocysteine, generating an exo-methylene intermediate (Fig. 1). Stereospecific Michael addition of water to the exo-methylene intermediate followed by NADH-mediated reduction of the 3-ketone furnishes adenosine and regenerates NAD ϩ (Fig. 1), which is tightly bound and serves a catalytic role in the overall reversible reaction (1,23,24).
Whereas SAHH is recognized as a key enzyme in controlling gene expression and a potential drug target for the design of antiviral (25,26), antiparasitic (27), and antiarthritic agents (28), the cellular regulation of SAHH is poorly understood. Several years ago, mammalian SAHH was reported in a proteomics study to be acetylated on two C-terminal Lys resides, 401 and 408 (29), the latter of which is highly conserved. However, the consequences of these acetylations were not explored. Classical site-directed mutagenesis is commonly used to analyze the role of post-translational modifications including acetyl-Lys, but can be misleading because of imprecise mimicry with the standard amino acid substitutes for unacetylated Lys (Arg) and acetyl-Lys (Gln) (30 -32). Here we use expressed protein ligation, a technique for protein semisynthesis, to generate sitespecifically acetylated SAHH and determine the effects of acetylation on enzymatic kinetics. Both acetylation of Lys 401 and Lys 408 inhibit the activity of SAHH. We also show using x-ray crystallography that both Lys 401 and Lys 408 acetylation of SAHH perturb local hydrogen bonding interactions, and we propose that these perturbations may influence catalysis. The significance of acetylation-induced changes in SAHH structure and function are discussed in the context of a potential global impact on cellular methylation status.

EXPERIMENTAL PROCEDURES
Reagents-Mercaptoethylsulfonate was purchased from Sigma. All Fmoc-derivatives were from EMD (Billerica, MA). Other reagents were at the highest grade obtainable from commercial sources.
Peptide Synthesis and Purification-The C-terminal fragment of SAHH (amino acid 396 -432 with 396 replaced by Cys,  CAHLGKLNVKLTKLTEKQAQYLGMSCDGPFKPDHYRY) with or without acetyl groups at Lys 401 and Lys 408 were synthesized using the standard Fmoc solid phase synthesis strategy on a PS3 peptide synthesizer from Protein Technologies (Tucson, AZ) on 0.1-0.2 mmol scale. Fmoc-protected amino acids were mixed with 4 eq of the coupling reagent 1-(bis(dimethylamino)methylene)-1H-1,2,3-triazolo[4,5-b]pyridinium 3-oxid hexafluorophosphate and double coupled to Wang resin (Novabiochem) containing the protected C-terminal amino acid. The Fmoc protecting group was removed by addition of 20% piperidine in dimethylformamide. A total of four peptides were produced including unacetylated, two singly acetylated, and the bi-acetylated forms. After completion, peptides were cleaved from the resin using reagent K (TFA/thioanisole/water/ phenol/ethane dithiol (82.5:5:5:5:2.5, v/v)) for 3-4 h at room temperature. After precipitation with ice-cold diethyl ether, peptides were purified by reversed phase HPLC using a preparative C 18 column and a gradient of increasing acetonitrile versus water, each containing 0.05% trifluoroacetic acid. Pure fractions were pooled, concentrated by rotor vapor, lyophilized, and stored at Ϫ20°C.
Purified peptides were run on an analytical C 18 reversedphase HPLC column (Agilent Eclipse XDB-C18) to confirm purity. Peptide powder was dissolved in HPLC aqueous buffer A (water containing 0.05% trifluoroacetic acid). The elution was monitored at 214 nm and carried out at a flow rate of 1 ml/min over a multistage gradient: 5-60% buffer B (acetonitrile containing 0.05% TFA) for 0 -65 min, 60 -100% B for 65-70 min, 100% B for 70 -80 min, 5-100% B for 80 -84 min, and 5% B for 84 -94 min. Synthetic peptides were shown to be Ͼ95% pure by analytical HPLC and the correct structures were confirmed by MALDI-TOF mass spectrometry (Applied Biosystems, Voyager DE-STR).
DNA Plasmid Construction-The full-length human SAHH open reading frame (Invitrogen, IOH14308) was subcloned into pTXB1 vector (New England Biolabs) using NdeI and SapI restriction sites. The pTXB1 vector encodes the GyrA intein from Mycobacterium xenopi followed by the chitin binding domain tag for affinity chromatography. DNA sequencing identified a mutation at Asp 86 in the commercial cDNA, which was replaced with an Asn at this position using QuikChange mutagenesis (Agilent) by standard protocols to make it identical to human S-adenosylhomocysteine hydrolase sequence (AAA51681) in GenBank TM . The mutant proteins corresponding to N27K, D293A, and D422K were also prepared using QuikChange mutagenesis.
Preparation of Semisynthetic SAHH Proteins-The truncated SAHH-intein-CBD fusion protein was produced in Escherichai coli BL21(DE3) liquid cultures grown at 37°C in LB media in shaker flasks on a liter scale, followed by induction at A 600 ϭ 0.5 with 0.5 mM isopropyl thiogalactoside at 16°C for 16 -20 h before pelleting the cells by centrifugation (5000 ϫ g). Cells were resuspended in lysis buffer containing 25 mM HEPES, 500 mM NaCl, 1 mM EDTA, 1 mM Tris(2-carboxyethyl)phosphine, 10% glycerol, pH 7.5) and then lysed by passage through a chilled French press cell. The cell lysate maintained at 4°C was centrifuged for 30 min (27,000 ϫ g) to remove cell debris and insoluble materials and then the supernatant was loaded onto a column containing chitin beads (5 ml/liter of E. coli culture) and incubated at 4°C for 30 min. After the SAHH-intein-CBD fusion protein was bound to chitin beads, the column was drained by gravity and washed with 20 column volumes of wash buffer I (50 mM HEPES, 500 mM NaCl, 0.1 mM EDTA, pH 7.5) followed by wash buffer II (50 mM HEPES, 150 mM NaCl, 0.1 mM EDTA, pH 7.5). The column was then quickly equilibrated with 200 mM 2-mercaptoethanesulfonate and protease inhibitor (Roche Complete mini) in cleavage buffer (0.2 M potassium phosphate, 0.1 mM EDTA, pH 7.5). After the addition of 5 ml (per liter E. coli culture) of cleavage buffer, the column was purged with N 2 gas and sealed. Upon apparent completion of the cleavage reaction as gauged by Coomassie-stained 10% SDS-PAGE (ϳ40 h at room temperature), the eluent was collected in 2-ml fractions and the column was further eluted with 1-2 more column volumes of cleavage buffer. Fractions that contained significant SAHH levels were concentrated 5-10fold and subjected to buffer exchange (0.2 M potassium phosphate, 0.1 mM EDTA, pH 6.5) using an Amicon protein concentrator (10-kDa cutoff) to minimize thioester hydrolysis.
The appropriate 37-mer SAHH N-Cys C-terminal peptide containing either Lys or acetyl-Lys at positions 401, 408, or both locations was dissolved in water and then diluted in ligation buffer (0.2 M potassium phosphate, 0.1 mM EDTA, 200 mM mercaptoethylsulfonate, 1 mM NAD ϩ , protease inhibitor, pH 7.0) to make a peptide solution with a final concentration ϳ500 M. The peptide solution was cleared by centrifugation at 20,627 ϫ g in a microcentrifuge for 10 min to remove any precipitation. The concentrated, truncated SAHH C-terminal thioester protein (amino acid 1-395) prepared as described above was added to the N-Cys peptide solution to achieve a final SAHH protein concentration of about 50 M, ϳ1:10 protein:peptide ratio. The tube containing the ligation reaction was sealed after blanketing with N 2 gas. Ligation progress was routinely monitored every 12-16 h by Coomassie-stained 10% SDS-PAGE until ϳ90% completion (about 36 h at room temperature). If observed, precipitation during the ligation reaction was removed by centrifugation every 12-16 h.
Upon completion of the ligation reaction, the ligation mixture was loaded onto a Superdex 200 size exclusion chromatography column for further purification using purification buffer (20 mM K 2 HPO 4 , 100 mM NaCl, 1 mM EDTA, 1 mM DTT, pH 7.2). The FPLC flow rate was 0.4 ml/min and 1.5-ml fractions were collected. Gel filtration standard (Bio-Rad) was run under the same conditions. Fractions that were Ͼ95% pure by stained 10% SDS-PAGE were pooled and concentrated using an Amicon protein concentrator (10 kDa cutoff) to Ͼ1 mg/liter and dialyzed into final dialysis buffer (10 mM K 2 HPO 4 , 1 mM EDTA, pH 7.2) for 48 -72 h with two or more buffer exchanges in a dialysis cassette (Slide-A-Lyzer) with a 20-kDa molecular mass cutoff. Semisynthetic SAHH was aliquoted and flash frozen using liquid N 2 and stored at Ϫ80°C. Typical yields were Ͼ2 mg of semisynthetic SAHH/liter of E. coli culture.
Mass Spectrometric Analysis of Semisynthetic SAHH-Semisynthetic SAHH forms were dialyzed against mass spectrometry buffer (20 mM NH 4 HCO 3 , pH 8.0) using a Slide-A-Lyzer mini dialysis unit (20 kDa mass cutoff). Protein was then spotted on a sample plate using the sandwich method with sinapinic acid (10 mg/ml in 40% acetonitrile, 0.1% TFA) as matrix. MALDI-TOF (Applied Biosystems, Voyager DE-STR) spectra were collected with calibration using bovine serum albumin as standard.
Preparation of Full-length, Recombinant SAHH-Full-length (amino acids 1-432) WT recombinant SAHH (r-SAHH) and N27K and D293A mutants were expressed in an intein containing vector using the conditions for the truncated protein as described above. Protein purification of the SAHH-intein-CBD fusion protein immobilized on chitin resin was similar to the truncated protein above except in place of mercaptoethylsulfonate containing buffer, 50 mM DTT cleavage buffer (50 mM HEPES, 150 mM NaCl, 0.1 mM EDTA, pH 7.5) was used. Incubation with DTT cleavage buffer for 16 h at room temperature led to the production of purified WT and mutant r-SAHH (Ͼ90% pure by Coomassie-stained 10% SDS-PAGE), flash frozen, and stored at Ϫ80°C until reconstitution at Ͼ1 mg/ml.
Reconstitution of SAHH/NAD ϩ -Reconstitution of the NAD ϩ form of SAHH was carried out as described previously (33). Briefly, SAHH was treated with 80% (v/v) ammonium sulfate in reconstitution buffer (25 mM K 2 HPO 4 , 5 mM DTT, 1 mM EDTA) to precipitate the purified SAHH over three successive cycles (pH 2.9, 2.9, and then pH 7.2). The precipitated protein was collected by centrifugation and dissolved in reconstitution buffer at pH 7.2. NAD ϩ (1 mM) was incubated with 0.1 mM SAHH overnight at 4°C. The reconstituted NAD ϩ form of SAHH was further purified by a Superdex 200 column (Amersham Biosciences Sciences, 10/300GL) as described above.
SAHH Enzymatic Assays-Measurement of SAHH activity in the forward (hydrolytic) direction was performed using an indirect, continuous assay by quantifying the production of homocysteine free thiol using Ellman's reagent (5,5Ј-dithiobis(nitrobenzoic acid) ) (33)(34)(35) in the presence of adenosine deaminase to prevent reaction reversal. To 200 l of the enzyme solution containing 0.15-0.3 M SAHH and 0.8 units of adenosine deaminase (Roche Applied Science) in assay buffer (50 mM potassium phosphate buffer, pH 7.2, 1 mM EDTA, 1 mM NAD ϩ ) was added 50 l of varying concentrations of AdoHcy (15.6 -1000 M) with 250 M 5,5Ј-dithiobis(nitrobenzoic acid) in the assay buffer. The reaction mixture was mixed rapidly and monitored at 412 nm continuously at 37°C using a UV spectrophotometer (Beckman DU640) containing a Peltier temperature controller. The initial rate was obtained by calculating the slope after the reaction reached 37°C and entered a linear region after a short lag phase (Ͻ2 min). An extinction coefficient of 13,600 M Ϫ1 cm Ϫ1 for 5-thio-2-nitrobenzoate, the product of the 5,5Ј-dithiobis(nitrobenzoic acid) reaction, was used to calculate the amount of homocysteine formed (33)(34)(35).
Measurement of SAHH activity in the reverse (AdoHcy synthesis) direction was performed using a direct, discontinuous assay by quantifying AdoHcy formation from adenosine and homocysteine by reversed phase HPLC (36). SAHH (ϳ0.1 M) was incubated with 1 mM adenosine and 5 mM homocysteine in 200 l of assay buffer at 37°C for 0, 10, 20, 40, and 60 min and the reaction was quenched by addition of 10 l of 6 N trifluoroacetic acid. Quenched reaction mixtures were filtered through an Amicon microconcentrator (3 kDa cutoff) to remove the protein precipitate. The resulting mixtures were diluted 1:1 in HPLC aqueous buffer A (water containing 0.05% trifluoroacetic acid) and analyzed by HPLC using a C 18 reversed-phase column (Agilent Eclipse XDB-C18). The elution was carried out at a flow rate of 1 ml/min over a multistage gradient: 0.5-7% buffer B (acetonitrile containing 0.05% trifluoroacetic acid) for 0 -45 min, 7-100% B for 45-50 min, and 100% B for 50 -60 min. The peaks corresponding to AdoHcy and adenosine were monitored at 254 nm. The concentration of AdoHcy was determined by calculating the percentage of AdoHcy in the reaction mixture at specific time points with less than 20% substrate conversion. The k cat values were obtained from the slopes from plots of [AdoHcy]/[E] versus time using linear regression curves. All reaction rates were measured in duplicate on two separate occasions and replicate measurements typically agreed within 20%.
Western Blot-Western blots were performed using dry transfers to PVDF via an Iblot dry blotting system (Invitrogen) and blocked with 5% bovine serum albumin in Tris-buffered saline in Tween 20 for 1 h at room temperature. Primary antiacetyl-lysine antibody (ab21623, Abcam) was applied using a 1:2000 dilution and incubated overnight at 4°C. Secondary anti-rabbit IgG horseradish peroxidase-linked antibody (GE Healthcare, NA934V) was used at 1:10,000 dilution. Enhanced chemiluminescence Western blotting detection reagents (Amersham Biosciences) were employed in combination with a Western blot imager (Syngene PXI 6).
Protein Crystallization-Reconstituted acetylated semisynthetic SAHH was concentrated to 10 -15 mg/ml and dialyzed against crystallization buffer (10 mM Tris, 1 mM EDTA, pH 7.2). Crystals of acetylated semisynthetic SAHH were grown by the hanging-drop method by mixing 2 l of SAHH with 1 l of a well solution containing 200 mM sodium formate, 5-20% PEG 3350, and 100 mM BisTris, pH 6.0 -6.5. SAHH AcK 408 crystals were grown without seeding. Bi-acetylated SAHH was crystallized by microseeding with seeds prepared from SAHH AcK 408 crystals. All crystals appeared in less than a week and grew to dimensions of 150 ϫ 150 ϫ 50 m 3 .
X-ray Crystallographic Data Collection and Structure Determination-Both bi-acetylated and AcK 408 crystals were cryoprotected in a stepwise fashion. The crystals were first stabilized by dropwise addition of 20% PEG 3350, 200 mM sodium formate, 100 mM BisTris, pH 6, and 1 mM NAD ϩ and then cryoprotected by dropwise addition of the same buffer supplemented with 12% glycerol. Stabilized crystals were harvested and flash frozen in liquid nitrogen. Data were collected at beamline 7.2 at the Stanford Synchrotron Radiation Light Source and processed with HKL2000 (37). Molecular replacement was performed with PHASER (38) using the structure of human SAHH as a search model (39) and diffraction data extending to 2.8 Å. Two independent subunits of SAHH were found in the crystallographic asymmetric unit with a final translational z-score of 66.4. Model building and structure refinement were performed iteratively using COOT (40) and PHENIX (41), respectively.

RESULTS
SAHH Semisynthesis-To generate unacetylated and 401-and 408-acetylated forms of SAHH, we elected to use expressed protein ligation, a protein semisynthesis method that involves replacing the natural C terminus of a protein of interest with a synthetic peptide (Fig. 2) (42). One requirement of expressed protein ligation is the need for a cysteine at the ligation junction (42). As there was not a natural Cys conveniently located, we examined several sites for Cys replacement. We found that replacement of Glu 396 with Cys was relatively well tolerated for expression and catalytic properties as reflected in the parameters of semisynthetic SAHH compared with WT recombinant SAHH (see below, Table 1).

TABLE 1 Steady-state kinetic parameters for semisynthetic SAHHs (s-SAHH) and unacetylated, wild-type recombinant SAHH (r-SAHH) and r-SAHH mutants
Standard error values are indicated with Ϯ and obtained from nonlinear curve fitting parameters. To perform expressed protein ligation, cDNA for the C terminally deleted form of SAHH (amino acids 1-395, t-SAHH) was subcloned in-frame to a gyrase intein-chitin binding domain containing vector, allowing for thioester generation at the C terminus of the recombinant protein fragment of SAHH. Separately, solid phase synthesis was performed to prepare highly purified N-Cys containing 37-mer synthetic peptides corresponding to the SAHH C-terminal residues amino acids 396 -432 with or without side chain acetyl groups on Lys 401 or Lys 408 (see Fig. 3). Ligation of the recombinant SAHH thioester fragment of amino acids 1-395 (t-SAHH) to the N-Cys peptides proceeded smoothly in the presence of 1 mM NAD ϩ cofactor, which enhanced the yield of the desired semisynthetic proteins, presumably by stabilizing the SAHH-folded protein structure. The four semisynthetic SAHH proteins, s-SAHH (unacetylated), AcK 401 -s-SAHH (acetylated at Lys 401 ), AcK 408 -s-SAHH (acetylated at 408), and bi-AcK 401/408s-SAHH (acetylated at both Lys 401 and Lys 408 ) were further purified by size exclusion chromatography (Fig. 4), demonstrating that each was tetrameric. In contrast, truncated SAHH (t-SAHH) was monomeric. Coomassie-stained SDS-PAGE and mass spectrometry showed that each of the semisynthetic SAHH proteins was Ͼ90% pure (Fig. 5). In addition, the presence of the acetyl-Lys modifications in AcK 401 -s-SAHH, AcK 408 -s-SAHH, and bi-AcK 401/408 -s-SAHH but not s-SAHH was confirmed by Western blot with anti-acetyl-Lys antibody (see Fig. 5B).  NOVEMBER 7, 2014 • VOLUME 289 • NUMBER 45

JOURNAL OF BIOLOGICAL CHEMISTRY 31365
Enzymatic Analysis of Semisynthetic SAHH Proteins-The catalytic activity of SAHH was measured in both the forward (hydrolase) and reverse (AdoHcy synthesis) directions. To determine SAHH activity in the thermodynamically disfavored forward direction, it is necessary to add adenosine deaminase to breakdown adenosine as it is formed to prevent the back reaction (18,(33)(34)(35)(36). Measurement of SAHH activity in the forward direction was performed by monitoring homocysteine production quantified via its free thiol by reaction with Ellman's reagent in a continuous assay. Measurement of the reverse reaction was performed using a direct, discontinuous assay that permits simultaneous quantification of AdoHcy and adenosine. Although more robust than the measurement of free thiol that can air-oxidize, this HPLC assay is less sensitive and in our hands was only amenable to k cat measurement with saturating substrate concentration because measurements at low substrate concentration show poor signal-to-noise. These experiments showed that the kinetic parameters of unacetylated semi-  synthetic SAHH (s-SAHH), which contains a E396C replacement, were very similar to fully recombinant WT SAHH (r-SAHH), with k cat (0.79 s Ϫ1 versus 1.04 s Ϫ1 , r-SAHH versus s-SAHH) and Ado-Hcy K m (13.9 versus 12.9 M, r-SAHH versus SAHH) values that are within 30% (see Table 1).
Interestingly, the presence of acetylation at either Lys 401 (AcK 401 -s-SAHH, k cat of 0.34 s Ϫ1 ) or Lys 408 (AcK 408 -s-SAHH, k cat of 0.30 s Ϫ1 ) in SAHH conferred 3-fold reduction in k cat relative to unacetylated s-SAHH, with less than a 30% change in Ado-Hcy K m among these enzyme forms (Table 1, Fig. 6). Doubly acetylated SAHH (bi-AcK 401/408 -s-SAHH) showed a similar 3-fold k cat reduction and a 2-fold increase in SAH K m . The k cat effects were corroborated by analyzing the reverse reactions (Table 1, Fig. 6). Thus, the catalytic efficiency (k cat /K m ) of SAHH was reduced by acetylation at either lysine position and the largest effect, 5-fold reduction, was observed in the dually acetylated enzyme.
Structure of Acetylated SAHH-We determined the crystal structures of semisynthetic SAHH bound to NAD ϩ with either one acetylation (AcK 408 ) or two acetylations (AcK 401/408 ) by molecular replacement and refine these structures using diffraction data extending to 2.6 and 2.3 Å, respectively ( Table  2, Fig. 7). Additional electron density was observed in the adenosine binding pocket and was assigned as adenosine, which likely copurified with SAHH. The electron density for the backbone atoms of the modified residues was clear, but the electron density for the side chains was weak (Fig. 8, A  and B) indicating that the side chains are poorly ordered. As the structure of the SAHH AcK 408 is nearly identical to FIGURE 6. Measurements of catalytic activity of semisynthetic SAHHs. A, representative steady-state kinetic plots for semi-synthetic SAHH forward reactions: (blue s-SAHH •), AcK 401 -s-SAHH (green ), AcK 408 -s-SAHH (red OE), bi-AcK 401/408 -s-SAHH (black ࡗ); B, representative time course plots for semisynthetic SAHH reverse reactions at saturating substrate concentrations: s-SAHH (blue •), AcK 401 -s-SAHH (red ), AcK 408 -s-SAHH (green OE), bi-AcK 401/408 -s-SAHH (black ࡗ); C, representative HPLC traces for semisynthetic SAHHs reverse reactions at saturating substrate concentrations. From top to bottom: enzyme reaction mixture at t ϭ 0 min, s-SAHH reaction mixture at t ϭ 20 min, and AcK 401 -s-SAHH reaction mixture at t ϭ 20 min. X, Y, and Z represent peaks for NAD ϩ , adenosine, and AdoHcy from left to right, respectively. SAHH AcK 401/408 , we will focus our discussion on the biacetylated structure.
Acetylated SAHH crystallizes with two monomers in the asymmetric unit and the tetramer is generated by crystallographic symmetry. This tetramer superimposes with the wildtype, unmodified tetramer with a root mean square deviation of 0.45 Å over 1712 C␣ atoms (Fig. 7A). No local perturbations in the backbone positions were observed at the sites either of semisynthetic ligation or acetylations (Figs. 7 and 8), but small structural changes were clearly observed for side chains near the sites of modifications. Acetylation of Lys 408 disrupts a hydrogen bond with Asp 422 , and both side chains adopt different conformations than when non-acetylated (Fig. 8C). Acetylation of Lys 401 results in a rearrangement of a hydrogen bonding network between Lys 401 , Asn 27 , and Asp 293 (Fig. 8D). In SAHH AcK 401/408 , the ⑀-amino group of AcK 408 makes hydrogen bonds to the backbone carbonyl of Asn 27 . In the non-acetylated structure, a water mediates this interaction. Additionally, Arg 34 adopts a different rotamer when Lys 401 is modified and, in one chain of the asymmetric unit, the guanidine group of Arg 34 is within hydrogen bonding distance of the carbonyl.
To explore the possibility that SAHH hydrogen bonding networks involving Asn 27 , Asp 293 , and Asp 422 contribute to the catalytic defects conferred by 401 and 408 Lys acetylation, we attempted to generate three SAHH recombinant mutant proteins: N27K, D293A, and D422K. Although D422K r-SAHH did not express sufficient soluble protein for reconstitution and purification, N27K and D293A r-SAHH were obtained. Both N27K and D293A r-SAHH showed 3-fold reductions in catalytic efficiency (k cat /K m ) compared with WT r-SAHH (see where I j is the intensity of an individual reflection, and ͗I͘ is the mean intensity for multiply recorded reflections. b The values in parentheses are for the highest resolution shell.
where F o is an observed amplitude and F c a calculated amplitude; R free is the same statistic calculated over a subset of the data that has not been used for refinement.

DISCUSSION
Protein lysine acetylation is now understood to be a very abundant post-translational modification in the cell, with thousands of proteins and specific sites being recently identified using mass spectrometry (29,43,44). A large number of these acetylation events have been linked to metabolic enzymes inside and outside of the mitochondria. The effects of the overwhelming majority of protein acetylation modifications have not yet been explored. In many studies, the effects of these acetyl modifications have been approximated by mutagenesis substitution involving Lys 3 Gln (45)(46)(47)(48)(49). Although sometimes informative, such Lys 3 Gln replacements can be misleading because of the structural differences between Gln and acetyl-Lys. In a small number of cases the precise contributions of specific, stoichiometric protein acetylation events have been characterized in detail with the authentic post-translational modification. Prior examples include cyclophilin (50), histone modifications (51-53), p53 (54), and acetyltransferase autoacetylation (55)(56)(57). By using expressed protein ligation on SAHH, we establish here, for the first time to our knowledge, the high resolution structural and functional consequences of acetylation at two independent sites in the same protein.
Although the x-ray structure electron density for the side chains of the acetylated Lys residues was not strong, changes in the electron density for the rotameric conformations of several surrounding residues in acetylated SAHH were clear. Such altered rotameric conformations indicate changing hydrogen bonding patterns in acetylated SAHH. We propose that the reduction in hydrolase activity in acetylated SAHH may be mediated by propagation of these altered hydrogen bond networks to the dynamics of the protein-NAD ϩ interactions that in turn are transmitted to the catalytic site. A recent theoretical study of SAHH catalysis has highlighted the potential catalytic importance of Asp 293 (58), involved in the network of hydrogen bonds to the C terminus as shown in Fig. 8. As the Asn 27 -FIGURE 8. Experimental electron density and hydrogen bonding network of acetylated residues. Simulated annealing omit maps of bi-acetylated SAHH in which the atoms corresponding to the side chains of residues 401 and 408 were removed from the model. Shown is the electron density corresponding to the region near AcK 401 (A) and AcK 408 (B). The 2F o Ϫ F c electron density is contoured at 1 (blue), and the F o Ϫ F c electron density is contoured at 2 (green). The final model of SAHH AcK 401/408 is shown in stick representation. Comparison of non-acetylated (Ref. 39, PDB 1LI4) (shown in yellow sticks) and bi-acetylated SAHH (shown in white sticks) structures is shown for the region surrounding Lys 408 (C) and Lys 401 (D). A dashed black line marks residues within hydrogen bonding distance. Water is represented by red spheres. The rotamer of the wild-type Asn 27 has been modified from that deposited based on MolProbity (79) analysis. In the bi-acetylated structure, Asn 27 occupies two positions, each with partial occupancy.
Asp 293 hydrogen bond appears altered by Lys 401 acetylation, it is plausible that this change in interaction may result in reduction of hydrolase activity. In addition, the loss of the water molecule-mediated interaction of Asn 27 by Lys 401 acetylation may contribute to catalytic inhibition. The mutagenesis experiments involving Asn 27 and Asp 293 reported here confirm the contributions of these residues in catalysis.
The biological consequences of SAHH acetylation have yet to be determined. Identification of the SAHH acetyltransferase(s) and deacetylase(s) and the role of SAHH lysine acetylation potentially affecting the cellular stability and/or proteinprotein interactions of SAHH are important topics for future research. However, we are intrigued by the possibility that increased acetylation of SAHH could lead to a global decline in SAM-mediated methyl transfer reactions in the cell. The product of methyltransferase reactions, AdoHcy, is a general competitive inhibitor of AdoMet binding to the active sites of these enzymes (2, 18, 25, 59 -63). A crippling of SAHH by acetylation, as identified in our enzymatic data, could lead to a build-up of AdoHcy that in turn would impede most cellular methyltransferases. In this way, SAHH regulation could serve as a key node that facilitates the known connection between increased cellular acetyl-CoA concentrations that are associated with a reduction in histone lysine methylation (64 -66). Although the competitive dynamic between lysine acetylation and lysine methylation is a recognized concept (64 -66), the potential of the catalytically impaired acetylated SAHH serving as a contributing factor to reduced protein lysine methylation should now be considered. Although the 3-5-fold enzymologic effects of post-translational modification of SAHH are moderate, similar sized effects on other enzymes have been associated with biological consequences (67)(68)(69). Future cellular experiments will be needed to investigate the specific role of SAHH acetylation in contributing to this acetylation/methylation dynamic.
This study also highlights the utility of expressed protein ligation in the study of protein acetylation. Unnatural amino acid mutagenesis by nonsense suppression strategies (30,70), Cys derivatization (31,32,71), enzyme-catalyzed acetylation (56 -58, 72-74), and sortase-mediated semisynthesis (75) are viable strategies for site-specific introduction of acetyl-Lys or close mimics into proteins. Expressed protein ligation is especially attractive, however, when these modifications occur near the C terminus of proteins of interest (42, 55, 67-69, 76 -78). This is particularly relevant when multiple acetylation sites are clustered as is the case with SAHH.