Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 7-Day Trial for You or Your Team.

Learn More →

Kinetic Mechanism of the Histone Acetyltransferase GCN5 from Yeast

Kinetic Mechanism of the Histone Acetyltransferase GCN5 from Yeast THE JOURNAL OF BIOLOGICAL CHEMISTRY Vol. 275, No. 29, Issue of July 21, pp. 22048 –22055, 2000 © 2000 by The American Society for Biochemistry and Molecular Biology, Inc. Printed in U.S.A. Kinetic Mechanism of the Histone Acetyltransferase GCN5 from Yeast* Received for publication, April 5, 2000, and in revised form, May 2, 2000 Published, JBC Papers in Press, May 12, 2000, DOI 10.1074/jbc.M002893200 Kirk G. Tanner‡, Michael R. Langer, Youngjoo Kim, and John M. Denu§ From the Department of Biochemistry and Molecular Biology, Oregon Health Sciences University, Portland, Oregon 97201-3098 The transcriptional coactivator GCN5 from yeast Recent reports have demonstrated that the acetylation of specific lysine side chains in the amino termini of core histones (yGCN5) is a histone acetyltransferase that is essential for activation of target genes. GCN5 is a member of a large plays a crucial role in the transcriptional activation of specific family of histone acetyltransferases that are conserved target genes (2– 4). The yeast GCN5 (yGCN5) HAT shows a between yeast and humans. To understand the molecular strong preference for acetylation of lysine 14 of histone H3 in mechanisms of histone/protein acetylation, a detailed ki- vitro with reported broader specificity in vivo. In vivo yGCN5 netic analysis was performed. Bi-substrate kinetic analy- has been reported to acetylate lysine 9 and 18 of histone H3, sis using acetyl-coenzyme A (AcCoA) and an H3 histone residues 8 and 16 of histone H4, and lysine residues in the synthetic peptide indicated that both substrates must amino-terminal tail of histone H2B, albeit to a much lesser bind to form a ternary complex before catalysis. Product degree than lysine 14 of histone H3 (5, 6). Models have been inhibition studies revealed that the product CoA was a proposed in which these site-specific acetylations in the amino competitive inhibitor versus AcCoA. Desulfo-CoA, a dead- termini of histones lead to altered nucleosomal chromatin end inhibitor, also demonstrated simple competitive inhi- structure by disrupting histone-DNA contacts and histone-hi- bition versus AcCoA. Acetylated (Lys14Ac) H3 peptide dis- stone contacts (7). Enrichment of acetylation on specific lysine played noncompetitive inhibition against both H3 peptide residues suggests that differential acetylation within distinct and AcCoA. These results support a sequential ternary chromatin loci may play a key role in transcriptional regulation complex (ordered Bi-Bi) kinetic mechanism, where Ac- (2– 4). Several classes of HATs have been identified: p300/CBP CoA binds first, followed by H3 histone. Acetylated (8) cytosolic HAT 1 (9), P/CAF (10), TAFII250 (11), and SRC-1 (Lys14Ac) H3 product is released first, and CoA is the last (12). Among HAT enzymes, yGCN5, which is a P/CAF family product to leave. Also, two methods were developed to member, has been the most thoroughly characterized. yGCN5 measure the binding affinities of AcCoA/CoA for GCN5. eCoA, is typically found in two high molecular mass complexes, SAGA Employing the fluorescent CoA analog etheno-CoA ( -etheno-CoA), a K for eCoA of 5.1 6 1.1 mM was deter- (1.8 MDa) and Ada (0.8 MDa). The SAGA complex activates 1-N mined by fluorescence anisotropy. This value was similar transcription by association with acidic activation domains of value of 8.5 6 2.6 mM for AcCoA obtained using to the K various transcription factors and results in the acetylation of (inhibition constant) of equilibrium dialysis and to the K i nucleosomal histones H3 and H2B (13). The Ada complex has mM for CoA obtained from steady-state kinetic assays. 6.7 also been shown to acetylate nucleosomal histone H3 and his- Together, these data suggest that the acetyl moiety of tone H2B (14). AcCoA contributes little to the binding energy. Recently, several crystallographic and nuclear magnetic res- onance (NMR) molecular models have been solved for the cat- alytic domains of yGCN5, P/CAF, and a Tetrahymena homo- The histone acetyltransferase (HAT) GCN5 from Saccharo- logue (p55) of yGCN5 (15–17). These structures demonstrated myces cerevisiae catalyzes the transfer of an acetyl moiety from that the high degree of sequence homology between these en- acetyl-CoA (AcCoA) to the e-amino group of lysine 14 of histone zymes is manifested by a similar overall structure. The binary H3 (Scheme 1). GCN5 was originally identified as a transcrip- complex of P/CAFzCoA and the ternary complex of Tetrahy- tional activator that was necessary to promote maximal levels mena p55zCoAzH3 peptide revealed that residues contacting of GCN4-dependent transcription (1). CoA in the active site are highly conserved. In addition, a biochemical study (18) of yGCN5 enzymatic activity implicated yGCN5 the conserved glutamic acid 173 as a general base catalyst in AcCoA 1 histone H3| - 0 CoA 1 acetylated histone H3 the GCN5 HAT reaction, deprotonating the e-amino group of SCHEME 1 lysine 14 in histone H3. Elucidating the mechanism of HAT catalysis will provide a basis for understanding the link between histone acetylation * The costs of publication of this article were defrayed in part by the and gene activation. In this study we have systematically de- payment of page charges. This article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to termined the overall kinetic mechanism, developed methods for indicate this fact. measuring substrate/product binding affinities (fluorescence ‡ Supported by Postdoctoral Fellowship T32 DK07680. anisotropy and equilibrium dialysis), and determined the order § Supported by American Cancer Society Grant RPG-97-175-01-TBE of substrate binding and the order of product release by em- and National Institutes of Health Grant GM 59785-01. To whom cor- respondence should be addressed: Oregon Health Sciences University, ploying the product inhibitors CoA and acetylated (Lys14Ac) Dept. of Biochemistry and Molecular Biology, 3181 SW Sam Jackson H3 peptide and the dead-end inhibitor desulfo-CoA. The results Park Rd., Portland, OR 97201-3098. Tel.: 503-494-0644; Fax: 503-494- are consistent with a fully ordered Bi-Bi kinetic mechanism, 8393; E-mail: [email protected]. where AcCoA is the first substrate to bind, and CoA is the last The abbreviations used are: HATs, histone acetyltransferases; yGCN5, yeast GCN5; AcCoA, acetyl-coenzyme A; eCoA, 1-N -etheno-CoA. product released. 22048 This paper is available on line at http://www.jbc.org This is an Open Access article under the CC BY license. Kinetic Mechanism of Histone Acetyltransferase GCN5 22049 EXPERIMENTAL PROCEDURES v 5 V 3 @S#/@K ~1 1 I/@K ! 1 @S## (Eq. 4) m m is Materials—All chemicals were of the highest grade commercially v 5 V 3 @S#/@K ~1 1 I/K ! 1 [S]~1 1 I/K !# (Eq. 5) m m is ii available. Histone H3 peptide, ARTKQTARKSTGGKAPPKQLC, and the corresponding acetylated H3 peptide (Lys14Ac), corresponding to v 5 V 3 @S#/@K 1 @S#~1 1 I/K !# (Eq. 6) m m ii the 20 amino-terminal residues of human histone H3 and an additional carboxyl-terminal cysteine, were synthesized by the Protein Chemistry Determination of the Dissociation Constant for AcCoA Binding to Core Lab at The Baylor College of Medicine. Acetyl-CoA was purchased yGCN5 via Equilibrium Dialysis—Equilibration was performed using from Roche Molecular Biochemicals. Calf thymus histones were pur- Dispo-Equilibrium Dialyzers (Amika Corp.), which contains two 75-ml chased from Calbiochem. [ H]Acetyl-CoA (1.88 Ci/mmol) was from NEN chambers separated by a 5-kDa cut-off dialysis membrane. Equilibra- Life Sciences Products. P81 phosphocellulose disks were from Life tion conditions were 50 mM Tris, 50 mM Bis-Tris, 100 mM sodium Technologies, Inc. Dispo-Equilibrium Dialyzers were from Amika Corp. acetate, pH 7.5, and 32.5 mM yGCN5. The K value for AcCoA in the eCoA, 1-N -etheno-CoA was from Sigma. All other reagents were from presence of yGCN5 was determined by transferring 10 – 400 mM AcCoA Sigma or Fisher. (20 – 40 cpm H/pmol) into the buffer chamber and yGCN5 into Expression and Purification—The catalytic domain (amino acids 99 – the sample chamber. After 48 h of equilibration on a level shaker, 262) of yGCN5 was recombinantly expressed by isopropyl-B-D-thioga- samples were recovered from each chamber and counted by liquid lactopyranoside induction for 12 h at 25 °C and purified from BL21-DE3 scintillation to determine the amount of radioactivity in the buffer bacteria as described previously (18, 19). After cation exchange chro- chamber ([AcCoA] ) and the sample chamber ([AcCoA] 1 free free matography on S-Sepharose, fractions (assessed by SDS-polyacryl- [AcCoAzyGCN5]). The calculated [AcCoA] was subsequently sub- free amide gel electrophoresis) were pooled and concentrated and then sub- tracted from the calculated concentration [AcCoA] 1 free jected to size-exclusion chromatography on G-75 Sephadex. Purity was [AcCoAzyGCN5] to determine the concentration of bound AcCoA. The analyzed by SDS-PAGE, and fractions were concentrated and stored at data were presented in hyperbolic form (Equation 7). 220 °C until use. Protein concentrations were determined by the method of Bradford (20). @AcCoA z yGCN5# 5 ~@yGCN5] z @AcCoA] !/~K 1 @AcCoA] ) (Eq. 7) tot free d free Enzymatic Assays for yGCN5—yGCN5 histone acetyltransferase ac- Steady-state Fluorescence Anisotropy Measurements for eCoA Bind- tivity was monitored continuously using a Multiskan Ascent microplate ing to yGCN5—eCoA was titrated with increasing [yGCN5]. [eCoA] was reader (LabSystems, Franklin, MA) as described previously (19). determined spectrophotometrically by using a molar extinction coeffi- Briefly, the CoASH generated in the HAT reactions was continuously 21 21 cient of 5.6 mM cm at 275 nm. The anisotropy measurements were measured by using a coupled enzyme system with pyruvate dehydro- performed on a Photon Technologies QM-1 steady-state fluorescence genase. The CoASH-dependent oxidation of pyruvate was accompanied spectrophotometer equipped with a thermostat. eCoA was excited at by the reduction of NAD to NADH, which was measured spectrophoto- NADH 21 21 305 nm (3-nm slits), and emission was monitored at 405 nm with 20-nm metrically at 340 nm (e 5 6220 M cm ). The HAT assay reaction slits for both the parallel and perpendicular components. Measure- mixtures (70 ml) contained 0.2 mM NAD, 0.2 mM thiamine pyrophos- ments were performed in 50 mM Tris, pH 7.5, 25 °C, using a 200-ml phate, 5 mM MgCl ,1mM dithiothreitol, 2.4 mM pyruvate, and 0.03 quartz cuvette. A concentrated solution of yGCN5 (in 1-ml increments) units of pyruvate dehydrogenase (1 unit of dehydrogenase is defined by was added to the cuvette containing 0.5 mM eCoA and mixed using a the manufacturer (Sigma) to be the conversion of 1.0 mmol of b-NAD to small magnetic stirring bar. Anisotropy increases were measured after b-NADH/min at pH 7.4 at 30 °C). The assay buffer was 100 mM sodium 1–2 min. The anisotropy r of eCoA was calculated from the parallel Ivv acetate, 50 mM Bis-Tris, and 50 mM Tris and pH 7.5. All assays were and perpendicular Ivh polarized fluorescence intensities measured performed at 25 °C and were initiated by the addition of yGCN5. The upon parallel excitation according to Equation 8, rates were analyzed continuously for up to 5 min, and background rates resulting from the spontaneous formation of CoA were subtracted from r 5 ~Ivv 2 G z Ivh!/~Ivv 1 2 z G z Ivh! (Eq. 8) the initial velocities derived from the yGCN5-catalyzed reactions. Un- der these conditions, the coupled enzyme reaction does not limit the where G is the ratio of the parallel Ivh and perpendicular Ivh polarized observed initial velocities (19). Also, only the linear portion of the fluorescence intensity measured upon parallel excitation to correct for kinetic traces were used to determine the initial rates. These were the different efficiencies of the parallel and perpendicular polarization typically linear for up to 5 min. Stock solutions of the H3 peptide or fluorescence detectors. Data were fitted to Equation 9 described in acetylated (Lys14Ac) H3 peptide were prepared fresh daily in the pres- Richards et al. (22), ence of 5 mM dithiothreitol. Alternatively, yGCN5 activity was meas- ured using [ H]AcCoA, the P81-filter binding assay as described previ- A 5 ~~A 2 A ! z @yGCN5#/K 1 @yGCN5#! 1 A (Eq. 9) bound free d free ously (18). where A is the anisotropy measured at a given concentration of yGCN5, Bi-substrate Kinetic Measurements—The Bi-substrate kinetic analy- and A and A are the anisotropies of free and bound eCoA, free bound sis was performed at AcCoA concentrations spanning 1–20 mM and H3 respectively. This simplified equation can be used when the dissociation peptide or calf thymus histone concentrations spanning 25–730 mM. The constant is $10-fold higher than the fluorescent ligand concentration, yGCN5 HAT activity was monitored via the coupled-enzyme spectro- since the concentration of yGCN5 bound to eCoA becomes negligible in photometric assay using either pyruvate dehydrogenase or a-ketoglut- comparison to the total yGCN5 concentration (22). arate dehydrogenase. The data were fitted to three possible kinetic mechanisms: the sequential (ternary complex) mechanism equation RESULTS (Equation 1), the ping-pong (covalent intermediate) mechanism equa- Effect of Ionic Strength on yGCN5-catalyzed HAT Activity— tion (Equation 2), and the equilibrium-ordered equation (Equation 3), using the algorithms of Cleland (21) and the computer program Kin- Given the highly charged characteristic of the nucleosomal etAsyst (IntelliKinetics, State College, PA) and a nonlinear least complex, it is likely that ionic strength will greatly influence squares analysis. the catalytic activity of these enzymes. Therefore, the effect of ionic strength on the yGCN5 HAT reaction was examined by v 5 V 3 @A# 3 @B#/~~K 3 K ! 1 ~K 3 @B#! m ia b ma varying NaCl concentrations between 0 and 1 M, and the re- 1 ~K 3 @A#! 1 ~@A# 3 @B#!! (Eq. 1) sulting HAT activity was monitored in the radioactive P81- mb filter binding assay as described under “Experimental Proce- v 5 V 3 @A# 3 @B#/~~K 3 @B#! 1 ~K 3 @A#! 1 ~@A# 3 @B#!! (Eq. 2) m a b dures.” The resulting yGCN5 activity was plotted versus the log of the ionic strength and fitted to a line using least squares v 5 V 3 @A# 3 @B#/~~K 3 K ! 1 ~K 3 @A#! 1 ~@A# 3 @B#!! (Eq. 3) m a b b analysis (Fig. 1). A direct linear relationship between the log of Steady-state Inhibition—The product inhibitors CoA and acetylated the ionic strength and inactivation of yGCN5 HAT activity was (Lys14Ac) H3 peptide, as well as the substrate analog desulfo-CoA observed. There was an approximate 7-fold decrease of yGCN5 (lacking only the terminal sulfhydryl of CoA) were used in steady-state HAT activity between 0.15 M NaCl and 1 M NaCl. inhibition studies. To investigate whether competitive (Equation 4), Bi-substrate Kinetics—Yeast GCN5 catalyzes the transfer of noncompetitive (Equation 5), or uncompetitive (Equation 6) inhibition an acetyl moiety from AcCoA to the e-amino side chain of Lys was observed, the data were fitted to the respective inhibition equations 14 of histone H3 (Scheme 1). It was first necessary to determine based on the algorithms defined by Cleland (21) using a nonlinear least squares analysis. the basic kinetic mechanism (sequential or ping-pong) utilized 22050 Kinetic Mechanism of Histone Acetyltransferase GCN5 TABLE I Summary of kinetic and equilibrium constants for yeast GCN5 HAT Substrates K k /K k m cat m cat a 5 21 21 AcCoA 2.5 6 1.4 mM 6.8 3 10 M s 21 a 1.7 6 0.12 s a 3 21 21 H3 peptide 0.49 6 0.08 mM 3.5 3 10 M s Inhibitors K K K ii is d CoA 6.7 6 5.1 mM Desulfo-CoA 33 6 1.3 mM Etheno-CoA 5.1 6 1.1 mM a a AcH3 peptide (vs. CoA) 2.2 6 0.40 mM 1.9 6 0.15 mM a a AcH3 peptide (vs. H3) 2.5 6 1.3 mM 2.1 6 0.90 mM Experiments were performed in triplicate. The reported value is an average 6 S.D. FIG.1. Effect of ionic strength on yGCN5-catalyzed HAT activ- ity. NaCl concentrations were varied between 0 –1 M, and the resulting yGCN5 HAT activity was monitored in the spectrophotometric-coupled enzyme assay using pyruvate dehydrogenase. The conditions of the assay were 50 mM Tris, 50 mM Bis-Tris, 100 mM sodium acetate, pH 7.5, 25 °C, 75 nM yGCN5, and 10 mg/ml calf thymus histones. The resulting yGCN5 activity was plotted versus log of the ionic strength and fitted to a line using least squares analysis. FIG.3. CoA exhibits competitive inhibition toward AcCoA dur- ing yGCN5-catalyzed HAT activity. The data were presented in double-reciprocal form, where 1/velocity is plotted versus 1/[AcCoA] at a fixed concentration of 100 mM H3 peptide and various fixed concentra- tions of CoA: filled diamonds,0 mM; filled circles,50 mM; open diamonds, 100 mM; filled triangles, 300 mM; and open squares, 500 mM. The radio- active P81-filter binding assay was used to monitor activity. The reac- tion conditions were 50 mM Tris, 50 mM Bis-Tris, 100 mM sodium FIG.2. Ternary complex formation between yGCN5, AcCoA, acetate, pH 7.5, and 75 nM yGCN5 at 25 °C and a total reaction volume and H3 histone peptide. Double-reciprocal plots of 1/velocity versus of 55 ml. The experiment was performed in triplicate with a represent- 1/[H3 peptide] at several fixed concentrations of AcCoA are shown. Data ative plot displayed. were fitted to the equation for a sequential mechanism as described under “Experimental Procedures.” The coupled spectrophotometric as- say using pyruvate dehydrogenase was utilized to monitor H3 acetyla- demonstrated that yGCN5 utilizes a sequential (ternary com- tion at pH 7.5, 25 °C. The AcCoA concentrations were as follows: open plex) mechanism. Here, a synthetic H3 peptide (ARTKQTARK- boxes, 1.1; filled upright triangles, 2.0; open circles, 4.3; filled diamonds, 8.6; open triangles, 17.1; and filled inverted triangles,34 mM). The STGGKAPPKQLC) was employed as a substrate in lieu of calf concentration of yGCN5 was 70 nM, and H3 peptide concentrations thymus histones since this substrate provides a better defined spanned 25–730 mM. When the data were fitted to a covalent interme- system to analyze the steady-state kinetic parameters. diate (ping-pong) mechanism, the sum of the squares of the residuals To distinguish between a ternary complex (sequential) mech- was .100-fold larger than that for the sequential mechanism, which was .2-fold greater than an equilibrium ordered mechanism (described anism and a ping-pong (covalent-intermediate) mechanism, under “Experimental Procedures”). The experiment was performed in initial velocity steady-state kinetic parameters were obtained triplicate with a representative plot displayed. for AcCoA and H3 peptide using AcCoA concentrations span- ning 1–20 mM and H3 peptide concentrations spanning 25–730 by yGCN5 before a detailed investigation into the order of mM. Analysis of these data by plotting 1/velocity against 1/[H3 substrate binding and the order of product release could be peptide] at different fixed concentrations of AcCoA resulted in performed. CoA-dependent transferases are known to utilize an intersecting line pattern (Fig. 2) that is characteristic of a one of two distinct mechanisms to catalyze acetyl group trans- ternary complex (sequential) mechanism. In contrast, a ping- fer. One mechanism involves acetyl transfer from CoA to an pong (covalent intermediate) mechanism is typically character- enzyme side chain nucleophile before transfer to the amine ized by a parallel line pattern. The identification of a ternary substrate (23). In a subsequent step, the enzyme then transfers complex was consistent with our previously published report this acetyl group to the acceptor amine substrate. The alterna- (18) using a histone preparation derived from calf thymus, thus tive mechanism involves direct acetyl transfer from AcCoA to validating the use of H3 peptide as an excellent analog of the amine substrate acceptor without the formation of a cova- histone protein. The basic steady-state parameters derived lent enzyme intermediate (24). The latter mechanism requires from this Bi-substrate analysis are summarized in Table I. that both substrates and enzyme must form a ternary complex Employing acetylated (Lys14Ac) synthetic H3 peptide as a before catalysis can occur. In a previous study (18) we per- substrate for yGCN5, it was determined that lysine 14 of H3 formed a Bi-substrate kinetic analysis with core histone pro- peptide is the only residue of H3 peptide acetylated. No other teins (H3, H2A, H2B, and H4) derived from calf thymus and Lysine (Lys-4, Lys-9, or Lys-18) of H3 peptide was found to be Kinetic Mechanism of Histone Acetyltransferase GCN5 22051 SCHEME 2. Proposed kinetic mechanism of yGCN5. acetylated above background levels when using either the ra- hibition mechanism. This indicates that at high concentrations dioactive P81-filter binding assay or the coupled enzyme spec- of AcCoA, the inhibition by CoA is overcome, and the V max trophotometric assay (data not shown). This is an important values approach identical levels. CoA was found to be a linear result since acetylation at a secondary site to lysine 14 of H3 competitive inhibitor (Fig. 3) versus AcCoA at 100 mM H3 pep- peptide would significantly complicate the initial velocity ki- tide, with a K of 6.7 6 5.1 mM, similar to the value obtained for is netic analysis. (Fig. 2) the K for AcCoA (4.5 6 3.6 mM). The replot of the slopes Several distinct kinetic mechanisms are possible for enzyme from the double-reciprocal plot as a function of [CoA] was linear systems that employ two substrates and produce two products. and yielded similar inhibition constants (analysis not shown). Our previous study (18) combined with the Bi-substrate anal- These data indicated that AcCoA and CoA compete for the ysis described above have strongly argued against a ping-pong same form of yGCN5 and the same mutually exclusive binding (covalent intermediate) type mechanism and demonstrated site. In a sequential mechanism involving two substrates and that yGCN5 followed a sequential Bi-Bi mechanism. The bind- two products, this requires that AcCoA and CoA bind to free ing of substrates and release of products can be random, fully enzyme; thus, AcCoA is the first substrate to add, and CoA is ordered, or a combination of both. Product inhibition studies the last product to leave. In a complimentary experiment, CoA can easily distinguish among the various possibilities. did not inhibit against H3 peptide when AcCoA was present at Product Inhibition—To distinguish among the various pos- saturating levels (data not shown). These results are consistent sible kinetic models for substrate binding and product release, with CoA and H3 peptide binding to different forms of yGCN5 the products of the yGCN5-catalyzed HAT reaction were used and are consistent with the kinetic mechanism in which AcCoA to inhibit the yGCN5-catalyzed HAT reaction. To correctly binds first, followed by binding of H3 peptide and subsequent analyze product inhibition data, the reversibility of the yGCN5- transfer of the acetyl group from AcCoA to H3 and an ordered catalyzed reaction was examined. This was accomplished by a release of products with acetylated (Lys14Ac) H3 peptide re- high performance liquid chromatography assay using reversed leasing before the release of CoA (Scheme 2). phase chromatography to directly monitor the formation of To further demonstrate that H3 peptide and core histones deacetylated H3 peptide from a reaction of yGCN5, CoA, and operate through the same kinetic mechanism, CoA inhibition acetylated (Lys14Ac) H3 peptide. Analysis of the time points studies were carried out using calf thymus core histones as the from a reaction carried out at pH 7.5 and 25 °C with high amine nucleophile substrate (Fig. 4). As with H3 peptide (Fig. concentrations of acetylated (Lys14Ac) H3 peptide and CoA 3), CoA is a pure competitive inhibitor against AcCoA when calf with a corresponding non-enzyme control were subjected to thymus core histones are used. Again, this indicated that Ac- reversed phase chromatography. The estimated turnover num- CoA is the first substrate to add, and CoA is the last product to ber under this set of conditions (1.0 mM CoA, 1.0 mM acetylated be released. By default, H3 peptide binds second and acetylated (Lys14Ac) H3 peptide) was 0.00002 s (Lys14Ac) H3 peptide is the first to leave. The inhibition con- , ;100,000-fold lower than the k value (1.7 s ) determined for the forward reac- stants for CoA are summarized in Table I. Taken together with cat the fact that k tion under saturating substrates. Thus, yGCN5 is capable of values are similar between the two sources of cat 21 21 being a histone deacetylase as is predicted from thermodynam- substrate (k 5 1.7 s , k 5 0.74 s ) and that cat(H3) cat(histones) ics, albeit much less efficiently. both substrates (H3 peptide and calf thymus histones) generate Products of the yGCN5-catalyzed HAT reaction are classified intersecting line patterns in Bi-substrate kinetics, this strongly as competitive inhibitors if increasing concentrations of prod- suggests that H3 peptide and core histones follow the same uct decrease the apparent k /K for the varied substrate, kinetic mechanism. cat m Desulfo-CoA as a Dead-end Inhibitor—To further establish uncompetitive if they decrease the apparent k , and noncom- cat petitive if they decrease both. Evaluation of the patterns are that CoA acts as a pure competitive inhibitor, we utilized the obtained by observing increases in the slopes (apparent K CoA derivative desulfo-CoA in steady-state inhibition studies. k ), the intercepts (apparent 1/k ), or both, respectively, for Desulfo-CoA (which lacks only the terminal sulfhydryl of CoA) cat cat is a dead-end inhibitor since it forms a non-productive complex lines in double-reciprocal plots of 1/v versus 1/[S] obtained at increasing concentrations of product (and a constant concen- with yGCN5 and H3 peptide. Initial velocities were obtained, tration of the second substrate). and the data were plotted in double-reciprocal form with 1/ve- CoA as a Product Inhibitor—CoA was evaluated as a product locity versus 1/[AcCoA] at several fixed concentrations of des- inhibitor against both AcCoA and H3 peptide. The experiment ulfo-CoA. Desulfo-CoA was found to be a linear competitive was performed using the radioactive P81-filter binding assay inhibitor (Fig. 5) versus AcCoA at 100 mM histone H3 peptide, since CoA itself is a substrate for the coupled enzyme in the with a K of 33 6 1.3 mM (Table I). The binding affinity for CoA is spectrophotometric assay (19). Initial velocities were deter- and desulfo-CoA were not too dissimilar, with K values of 6.7 is mined, and the data were plotted in double-reciprocal form mM and 33 mM, respectively. The absence of the terminal sulf- with 1/velocity versus 1/[AcCoA] at several fixed concentrations hydryl in desulfo-CoA must account for the 4-fold higher K is of CoA. A series of double-reciprocal straight line plots inter- value for desulfo-CoA compared with CoA. The deletion of the sected on the 1/velocity ordinate, indicating a competitive in- terminal sulfhydryl likely results in a loss of Van der Waals 22052 Kinetic Mechanism of Histone Acetyltransferase GCN5 FIG.6. Determination of the dissociation constant (K ) for Ac- FIG.4. CoA exhibits competitive inhibition toward AcCoA dur- CoA binding to yGCN5 via equilibrium dialysis. The data are ing yGCN5-catalyzed HAT activity. The radioactive P81-filter bind- presented in hyperbolic form, where [AcCoAzyGCN5] is plotted versus ing assay was used to determine the effect of the product CoA on [AcCoA] . Equilibration conditions were 50 mM Tris, 50 mM Bis-Tris, free yGCN5-catalyzed HAT activity. The data are presented in double- 100 mM sodium acetate, pH 7.5, and '32.5 mM yGCN5. Equilibration reciprocal form, where 1/v is plotted versus 1/[AcCoA] at a fixed concen- was performed using Dispo-Equilibrium Dialyzers (Amika Corp.), tration of 150 mM calf thymus core histones and various fixed concen- which contain two 75-ml chambers separated by a 5-kDa molecular trations of CoA: filled diamonds,0 mM; open circles, 300 mM; filled mass cut-off dialysis membrane. The K value for AcCoA in the pres- diamonds, 500mM. The reaction conditions were 50 mM Tris, 50 mM ence of yGCN5 was determined by transferring 10 – 400 mM AcCoA Bis-Tris, 100 mM sodium acetate, pH 7.5, and 75 nM yGCN5 at 25 °C (20 – 40 cpm H/pmol) into the buffer chamber and yGCN5 into and a total reaction volume of 55 ml. the sample chamber. After equilibration (48 h) on a level shaker, samples were recovered from each chamber and counted by liquid scintillation to determine the amount of radioactivity in the buffer chamber ([AcCoA] ) and in the sample chamber ([AcCoA] 1 free free [AcCoAzyGCN5]) in order to determine the amount of bound AcCoA. The data were fitted to: [AcCoAzyGCN5] 5 ([yGCN5] 3 [AcCoA] )/ 0 free (K 1 [AcCoA] ). The experiment was performed in triplicate, with a d free representative plot displayed. value obtained from enzyme kinetic methods (K 5 6.7 is(CoA) mM). Steady-state Fluorescence Anisotropy Measurements for eCoA Binding to yGCN5—Changes in intrinsic protein fluorescence are often utilized to follow protein ligand interactions. How- ever, we observed no significant change in the fluorescent prop- erties of yGCN5 (l or quantum yield) upon binding AcCoA. max Therefore, we sought an alternative fluorescent approach, flu- orescence anisotropy (25). Steady-state fluorescence anisotropy FIG.5. Desulfo-CoA as a product inhibitor of the yGCN5-cata- measurements for eCoA (1,N -etheno-CoA) binding to yGCN5 lyzed HAT reaction. yGCN5 HAT activity was monitored in the radioactive P81-filter binding assay. Double-reciprocal plots of were carried out to determine the dissociation constant (K ). 1/velocity versus 1/[AcCoA] at several fixed concentrations of desulfo- eCoA contains a fluorescent derivative of adenine (containing CoA. Data were fitted to the equation for competitive inhibition as two additional carbon atoms) that has been utilized to study described under “Experimental Procedures.” The desulfo-CoA concen- many ATP-, NAD-, and CoA-dependent enzymes (26 –30). Two trations were as follows: filled diamonds,0 mM; open circles, 100 mM; open diamonds, 300 mM, and filled triangles, 500 mM. The reaction separate x-ray structures of CoA and GCN5 complexes demon- conditions were 50 mM Tris, 50 mM Bis-Tris, 100 mM sodium acetate, pH strated that there are no significant interactions between 7.5, at 25 °C and 75 nM yGCN5 and a total reaction volume of 55 ml. GCN5 and the adenine moiety of CoA (15, 16). We therefore reasoned that eCoA could be successfully employed as a rele- interactions with residues in the hydrophobic CoA binding site vant CoA analog. As expected, upon binding yGCN5, eCoA (15, 16). Together, these data indicate that CoA or desulfo-CoA exhibited a change in fluorescence emission polarization (ani- only bind to the free form of the enzyme and are consistent with sotropy). We used this change in polarization to generate a the kinetic mechanism of Scheme 2. binding curve and to determine the dissociation constant (Fig. Equilibrium Dialysis—Using equilibrium dialysis, the disso- 7). Using this analysis, we calculated a dissociation constant ciation constant for AcCoA binding to yGCN5 was determined. (K )of5.1 6 1.1 mM. A control experiment in which excess Varying concentrations of [ H]AcCoA were transferred into the AcCoA was added resulted in the loss of the change in fluores- buffer chamber, and 32.5 mM yGCN5 was transferred into the cence emission polarization (data not shown). The K value is sample chamber. After equilibration, samples were recovered in excellent agreement with the K of 6.7 mM obtained for CoA is from each chamber and were counted by liquid scintillation. (Fig. 3), suggesting that eCoA is a valid replacement for CoA in The data were analyzed as described under “Experimental these binding studies. Procedures” and the Fig. 6 legend. A representative binding Acetylated (Lys14Ac) H3 peptide as a Product Inhibitor—To experiment is shown in Fig. 6. The average value from three complete the product inhibition studies, the product inhibitor separate experiments yielded a K value of 8.5 6 2.6 mM and an acetylated (Lys14Ac) H3 peptide was examined as an inhibitor enzyme concentration of 33.6 6 2.5 mM. These data are in versus both H3 peptide and AcCoA. As predicted by the kinetic excellent agreement with the calculated enzyme concentration model (Scheme 2), both effects were found to be noncompetitive (32.5 mM determined by the method of Bradford (20)) and the K (mixed type), indicating that the substrate/inhibitor pairs bind i Kinetic Mechanism of Histone Acetyltransferase GCN5 22053 DISCUSSION to different enzyme forms and that there is a reversible con- nection between the two points of binding (Fig. 8). Interest- Ternary Complex Formation—The enzyme kinetic mecha- ingly, the slope effects (reversible connection) were less dra- nism of two substrate reactions can be directly probed by si- matic, suggesting that the reversible connection is quite weak. multaneously varying both substrates (31). It is well estab- Our observation that the reverse reaction is extremely slow lished that a double-reciprocal plot analysis that generates an (about 0.00002 s ), ;100,000-fold lower than for the forward intersecting line pattern suggests a ternary complex mecha- reaction (1.7 s ), would provide a reasonable explanation for nism, and a parallel line pattern is characteristic of a ping-pong the weaker reversible connection. The steady-state kinetic in- (covalent intermediate) mechanism. This approach has been previously used to predict the behavior of other well character- hibition parameters for acetylated (Lys14Ac) H3 peptide are summarized in Table I. These results, in conjunction with the ized acetyltransferases (23, 24). Here, the resulting Bi-sub- strate intersecting line pattern with yGCN5 (Fig. 2) is consist- CoA and desulfo-CoA inhibition data, are consistent with ent with a ternary complex mechanism. These data are yGCN5 catalyzed HAT activity following kinetic Scheme 2. consistent with a direct attack of lysine on AcCoA. Although it is impossible to exclude the existence of a covalent enzyme intermediate by this approach, there has been no direct or indirect evidence for such an intermediate. Using quench-flow trapping approaches, we were unable to detect a covalent acetylated enzyme intermediate when yGCN5 was rapidly mixed with AcCoA. If an intermediate is formed in the yGCN5 reaction, it must occur after both AcCoA and H3 peptide are bound, and it must break down before either acetylated H3 peptide or CoA is released. However, our data strongly suggest that yGCN5 forms a ternary complex with AcCoA and H3 peptide before catalysis. In this study, we employed a synthetic peptide substrate (ARTKQTARKSTGGKAPPKQL) corresponding to the amino- terminal 20 amino acids of histone H3. Previously, it was demonstrated that commercial preparations of calf thymus cores histones also generated an intersecting-line pattern from similar Bi-substrate kinetic analyses (18). Moreover, the k cat FIG.7. Steady-state fluorescence anisotropy measurements values obtained using either substrate are similar. Together, for eCoA binding to GCN5. eCoA was titrated with increasing these results indicate that both the H3 peptide and H3 protein [GCN5]. Measurements were performed in 50 mM Tris buffer, pH 7.5, (within the core octamer) follow the same kinetic mechanism 25 °C, using a 200-ml cuvette. Concentrated yGCN5 (in 1-ml increments) was added to the quartz cuvette containing 0.5 mM eCoA and mixed and that synthetic H3 peptide is a valid and relevant substrate using a small magnetic stirring bar. Anisotropy increases were meas- in detailed mechanistic studies. In fact, the synthetic peptide ured after 1–2 min. The anisotropy r of eCoA was calculated from the substrate affords many advantages over a heterogeneous prep- parallel Ivv, and perpendicular Ivh polarized fluorescence intensities aration of histones derived from calf thymus. First, the syn- were measured upon parallel excitation as described under “Experi- mental Procedures.” Data were fitted to the following equation, de- thetic H3 peptide is completely unacetylated. The acetylation scribed in Richards et al. (22): A 5 ((A 2 A )z[yGCN5]/K 1 state of histones derived from calf thymus is ill-defined and bound free d [yGCN5]) 1 A , where A is the anisotropy measured at a given free likely varies among preparations. Second, the non-enzymatic concentration of yGCN5, and A and A are the anisotropies of free free bound acetylation reaction rate is greatly reduced with H3 peptide eCoA. This simplified equation can be used when the disso- and bound since there are only four possible lysine acetylation sites com- ciation constant is $10-fold higher than the fluorescent ligand concen- tration, since the concentration of yGCN5 bound to eCoA becomes negligible in comparison to the total yGCN5 concentration (22). The fit yielded the values A 5 0.029 6 0.004, A 5 0.151 6 0.010, and K K. G. Tanner, M. R. Langer, Y. Kim, and J. M. Denu, unpublished free bound d 5 5.1 6 1.1 mM with r 5 0.986. observation. FIG.8. Inhibition of GCN5 by the product acetylated H4 (Lys14Ac) peptide. Data are presented in double-reciprocal form, where 1/v is plotted versus 1/[H3] (left panel) and 1/[AcCoA] (right panel) at various fixed concentrations of acetylated H3 (Lys14Ac) peptide (circles,0mM; triangles, 1.0 mM; and squares, 2.0 mM). The yGCN5 concentration was 70 nM, [AcCoA] was 65 mM (left panel), and [H3] was 233 mM (right panel). The reaction conditions were 50 mM Tris, 50 mM Bis-Tris, 100 mM sodium acetate, pH 7.5, and 25 °C. Each data set were fitted to a noncompetitive inhibition equation (KinetAsyst, IntelliKinetics, State College, PA). The experiments were performed in triplicate, with a representative plot displayed. 22054 Kinetic Mechanism of Histone Acetyltransferase GCN5 pared with calf thymus histones, which contain four core his- binds to EzCoA. Also, acetylated (Lys14Ac) H3 peptide exhib- tones H3, H2A, H2B, H4 along with linker H1. Third, we are ited noncompetitive inhibition with respect to H3 peptide, in- able to use the well defined product, synthetic acetylated dicating that acetylated and unacetylated H3 peptide bind to (Lys14Ac) H3 peptide, as a product inhibitor for the yGCN5- different enzyme forms. That is, H3 peptide binds to EzAcCoA, catalyzed HAT reaction. These advantages have allowed us to whereas acetylated H3 peptide binds to EzCoA. Interestingly, completely define the kinetic mechanism, as described in the slope effects (reversible connection) were less dramatic, Scheme 2. suggesting that the reversible connection is quite weak. Our The recently solved structures (both x-ray and NMR) of observation that the reverse reaction is extremely slow GCN5 and GCN5 homologues have implicated the formation of (;0.00002 s ) would provide a reasonable explanation for the a ternary complex of GCN5, AcCoA, and H3 peptide (15–17). weaker reversible connection. The product CoA exhibited pure competitive inhibition toward AcCoA (Fig. 3), indicating that The molecular models revealed two pronounced clefts that are roughly orthogonal to one another. Two of the structures have AcCoA and CoA compete for the same enzyme form. In a either AcCoA or CoA bound to the smaller of the clefts, whereas sequential mechanism, this requires that acetyl-CoA and CoA another revealed a ternary complex of CoA bound to the bind to free enzyme; thus, AcCoA is the first substrate to add, smaller cleft and an 11-residue H3 peptide, KSTGGKAPRKQ, and CoA is the last product to be released (Scheme 2). As bound to the larger cleft. The fact that the acetyl moiety of predicted by the kinetic model, CoA does not inhibit against H3 AcCoA was not transferred to the enzyme in the binary com- when AcCoA is present at saturating levels (data not shown). plex of Tetrahymena GCN5zAcCoA is consistent with the ab- Consistent with CoA acting as a pure competitive inhibitor, we sence of a covalent acetylated enzyme intermediate formation utilized the CoA derivative desulfo-CoA in our steady-state during catalysis. Analysis of the ternary GCN5zCoAzH3 com- inhibition studies and found that desulfo-CoA also acts as a plex revealed that it is possible for both H3 peptide and AcCoA pure competitive inhibitor with respect to acetyl-CoA (Fig. 5). to be bound simultaneously. This dead-end complex should be Desulfo-CoA is a dead-end inhibitor since it forms a non-pro- similar to that of a ternary complex of GCN5 with both sub- ductive complex. Taken together with the fact that acetylated strates AcCoA and H3 peptide, since the majority of the (Lys14Ac) H3 peptide is a linear noncompetitive (mixed) inhib- CoAzGCN5 interactions involve the pyrophosphate group and itor against both H3 peptide and AcCoA (Fig. 8), this suggests pantetheine arm of CoA. The crystal structure predicts that that AcCoA is the first to bind yGCN5 followed by H3 peptide. there is ample room to accommodate the acetyl moiety of Ac- After transfer of the acetyl moiety from AcCoA to H3 peptide in CoA. In fact, the carbonyl of the acetyl group would be posi- the ternary complex, acetylated (Lys14Ac) H3 peptide must tioned closer to both the nucleophile e-amino group of lysine 14 leave first, followed by the subsequent release of CoA (Scheme and the proposed general base glutamic acid 173, which must 2). The data suggest that if H3 peptide is capable of binding to deprotonate the e-amino group of lysine 14 before attack on the yGCN5, it does so only weakly. carbonyl carbon of AcCoA. Indeed, we had demonstrated pre- Given our kinetic and biochemical evidence that this highly viously that the conserved glutamic acid 173 residue was acting conserved family of GCN5 HATs utilize a fully ordered Bi-Bi as a general base (18). This study revealed an ionization of a kinetic mechanism, it is likely that binding of AcCoA to yGCN5 residue that must be unprotonated for activity (pK of 8.3 6 induces a conformational change that then optimizes the 0.2). Using site-directed mutagenesis and pH activity profiles, pocket for H3 peptide binding. Consistent with this, a recently the observed ionization was assigned to Glu-173. We concluded solved structure of a GCN5-related N-acetyltransferase (GNAT) that Glu-173 was functioning as a general base, deprotonating family member, serotonin N-acetyltransferase complexed, with the e-amino group of lysine 14 from histone H3. From later a Bi-substrate analogue predicts that binding of AcCoA will structural studies, it was discovered that two large hydropho- induce a conformational change that exposes the binding site bic clefts converged at the location of glutamic acid 173. This for the amine substrate serotonin (32). This is in good agree- hydrophobic environment surrounding glutamic acid 173 is ment with the kinetic mechanism of serotonin N-acetyltrans- consistent with the observed elevated pK (16, 18). ferase proposed by Cole and co-workers (33). As we have dem- Product Inhibition Studies—The Bi-substrate kinetic analy- onstrated for yGCN5, serotonin N-acetyltransferase was shown sis described here strongly suggests that a ternary complex to obey an ordered Bi-Bi mechanism. Moreover, the crystal mechanism is utilized by GCN5. To determine the order of structure of the Tetrahymena GCN5zCoAzH3 peptide ternary substrate binding and the order of product release, product complex suggests that the binding of H3 peptide is dependent inhibition studies were undertaken. The type of product inhi- on AcCoA binding first (16). The binding of AcCoA was sug- bition observed is diagnostic for a particular kinetic scheme gested to widen the histone substrate binding groove. (31). Typically, one substrate and one product are varied at Equilibrium Binding Studies—Measuring the binding affin- saturating or subsaturating levels of the other substrate, and ities of substrates and products is an important goal toward the resulting double-reciprocal plots indicate the type of inhi- understanding the catalytic and kinetic mechanisms of the bition. The two HAT reaction products, acetylated (Lys14Ac) HAT enzymes. Therefore, we developed a steady-state fluores- H3 peptide and CoA, were employed as product inhibitors. The cence anisotropy assay for eCoA binding to yGCN5. eCoA (1,N - fully ordered mechanism of Scheme 2 was easily distinguished etheno-CoA), a fluorescent analogue of CoA, is identical to CoA from other possible random mechanisms. Acetylated (Lys14Ac) except for two additional carbon atoms within the adenine H3 peptide was examined as an inhibitor versus both H3 pep- moiety. The structures of the binary and ternary complexes of tide and AcCoA (Fig. 8). As predicted by the kinetic model GCN5 with CoA or AcCoA suggested that this would be a (Scheme 2), both effects were found to be noncompetitive logical approach since the majority of the contacts involve the (mixed type), indicating that the substrate/inhibitor pairs bind pyrophosphate group and pantetheine arm of AcCoA/CoA. In to different enzyme forms and that there is a reversible con- fact, the adenine base in these structures is highly disordered, nection between the two points of binding. Acetylated with no specific interactions with GCN5 residues (15, 16). The (Lys14Ac) H3 peptide demonstrated noncompetitive inhibition binding curve generated from these anisotropy measurements with respect to AcCoA, indicating that acetylated (Lys14Ac) H3 yielded a dissociation constant (K )of5.1 mM (Fig. 7). This peptide and AcCoA do not bind to the same form of yGCN5 (Fig. value is in good agreement with the K of 6.7 mM for CoA (Fig. is 8). That is, AcCoA binds to E, whereas acetylated H3 peptide 3), determined in the inhibition studies. By equilibrium dialy- Kinetic Mechanism of Histone Acetyltransferase GCN5 22055 (1997) Nature 389, 194 –198 sis, a dissociation constant (K )of8.5 6 2.6 mM was determined 13. Grant, P. A., Duggan, L., Cote, J., Roberts, S. M., Brownell, J. E., Candau, R., for AcCoA binding to free enzyme (Fig. 6). The fact that the Ohba, R., Owen-Hughes, T., Allis, C. D., Winston, F., Berger, S. L., and Workman, J. L. (1997) Genes Dev. 11, 1640 –1650 dissociation constants determined by these binding assays 14. Eberharter, A., Sterner, D. E., Schieltz, D., Hassan, A., Yates, J. R. R., Berger, were in good agreement suggests that AcCoA and CoA bind S. L., and Workman, J. L. (1999) Mol. Cell. Biol. 19, 6621– 6631 with similar binding affinities to yGCN5. The observation that 15. Clements, A., Rojas, J. R., Trievel, R. C., Wang, L., Berger, S. L., and Marmorstein, R. (1999) EMBO J. 18, 3521–3532 the binding affinities between substrate AcCoA and product 16. Rojas, J. R., Trievel, R. C., Zhou, J., Mo, Y., Li, X., Berger, S. L., Allis, C. D., CoA are similar predicts that the intracellular ratio of AcCoA/ and Marmorstein, R. (1999) Nature 401, 93–98 CoA would be a critical determinant for in vivo HAT activity. 17. Lin, Y., Fletcher, C. M., Zhou, J., Allis, C. D., and Wagner, G. (1999) Nature 400, 86–89 This possible regulatory mechanism for HATs would link tran- 18. Tanner, K. G., Trievel, R. C., Kuo, M. H., Howard, R. M., Berger, S. L., Allis, scriptional activation/repression with various cellular meta- C. D., Marmorstein, R., and Denu, J. M. (1999) J. Biol. Chem. 274, 18157–18160 bolic pathways that control the fluctuating concentrations of 19. Kim, Y., Tanner, K. G., and Denu, J. M. (2000) Anal. Biochem. 280, 308 –314 AcCoA/CoA. It will be interesting to determine whether this is 20. Bradford, M. M. (1976) Anal. Biochem. 72, 248 –254 a common feature of all HAT families. 21. Cleland, W. W. (1977) Adv. Enzymol. Relat. Areas Mol. Biol. 45, 273–387 22. Richards, J. P., Bachinger, H. P., Goodman, R. H., and Brennan, R. G. (1996) REFERENCES J. Biol. Chem. 271, 13716 –13723 23. Thompson, S., Mayerl, F., Peoples, O. P., Masamune, S., Sinskey, A. J., and 1. Georgakopoulos, T., and Thireos, G. (1992) EMBO J. 11, 4145– 4152 Walsh, C. T. (1989) Biochemistry 28, 5735–5742 2. Brownell, J. E., and Allis, C. D. (1996) Curr. Opin. Genet. Dev. 6, 176 –184 24. Lewendon, A., Murray, I. A., Shaw, W. V., Gibbs, M. R., and Leslie, A. G. (1994) 3. Wade, P. A., Pruss, D., and Wolffe, A. P. (1997) Trends Biochem. Sci. 22, Biochemistry 33, 1944 –1950 128 –132 25. Lundblad, J. R., Laurance, M., and Goodman, R. H. (1996) Mol. Endocrinol. 10, 4. Grunstein, M. (1997) Nature 389, 349 –352 607– 612 5. Grant, P. A., Eberharter, A., John, S., Cook, R. G., Turner, B. M., and 26. Basso, L. A., Engel, P. C., and Walmsley, A. R. (1997) Biochim. Biophys. Acta Workman, J. L. (1999) J. Biol. Chem. 274, 5895–5900 1340, 63–71 6. Zhang, W., Bone, J. R., Edmondson, D. G., Turner, B. M., and Roth, S. Y. (1998) 27. Klebl, B. M., and Pette, D. (1996) Anal. Biochem. 239, 145–152 EMBO J. 17, 3155–3167 28. Kasprzak, A. A., and Villafranca, J. J. (1988) Biochemistry 27, 8050 – 8056 7. Hansen, J. C., Tse, C., and Wolffe, A. P. (1998) Biochemistry 37, 17637–17641 29. Atkins, W. M., Cader, B. M., Hemmingsen, J., and Villafranca, J. J. (1993) 8. Bannister, A. J., and Kouzarides, T. (1996) Nature 384, 641– 643 Protein Sci. 2, 800 – 813 9. Dutnall, R. N., Tafrov, S. T., Sternglanz, R., and Ramakrishnan, V. (1998) Cell 30. Favilla, R., Martin, S. R., and Bayley, P. M. (1988) Biochim. Biophys. Acta 955, 94, 427– 438 321–329 10. Yang, X. J., Ogryzko, V. V., Nishikawa, J., Howard, B. H., and Nakatani, Y. (1996) Nature 382, 319 –324 31. Segel, I. H. (1993) Enzyme Kinetics, pp. 506 – 845, John Wiley & Sons, Inc., New York 11. Imhof, A., Yang, X. J., Ogryzko, V. V., Nakatani, Y., Wolffe, A. P., and Ge, H. (1997) Curr. Biol. 7, 689 – 692 32. Hickman, A. B., Klein, D. C., and Dyda, F. (1999) Mol. Cell 3, 23–32 33. De Angelis, J., Gastel, J., Klein, D. C., and Cole, P. A. (1998) J. Biol. Chem. 12. Spencer, T. E., Jenster, G., Burcin, M. M., Allis, C. D., Zhou, J., Mizzen, C. A., McKenna, N. J., Onate, S. A., Tsai, S. Y., Tsai, M. J., and O’Malley, B. W. 273, 3045–3050 http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Journal of Biological Chemistry Unpaywall

Kinetic Mechanism of the Histone Acetyltransferase GCN5 from Yeast

Tanner, Kirk G.; Langer, Michael R.; Kim, Youngjoo; Denu, John M.
Journal of Biological Chemistry – Jul 1, 2000
Download PDF
8 pages

Loading next page...
 
/lp/unpaywall/kinetic-mechanism-of-the-histone-acetyltransferase-gcn5-from-yeast-IMnslXl1jE

References

References for this paper are not available at this time. We will be adding them shortly, thank you for your patience.

Publisher
Unpaywall
ISSN
0021-9258
DOI
10.1074/jbc.m002893200
Publisher site
See Article on Publisher Site

Abstract

THE JOURNAL OF BIOLOGICAL CHEMISTRY Vol. 275, No. 29, Issue of July 21, pp. 22048 –22055, 2000 © 2000 by The American Society for Biochemistry and Molecular Biology, Inc. Printed in U.S.A. Kinetic Mechanism of the Histone Acetyltransferase GCN5 from Yeast* Received for publication, April 5, 2000, and in revised form, May 2, 2000 Published, JBC Papers in Press, May 12, 2000, DOI 10.1074/jbc.M002893200 Kirk G. Tanner‡, Michael R. Langer, Youngjoo Kim, and John M. Denu§ From the Department of Biochemistry and Molecular Biology, Oregon Health Sciences University, Portland, Oregon 97201-3098 The transcriptional coactivator GCN5 from yeast Recent reports have demonstrated that the acetylation of specific lysine side chains in the amino termini of core histones (yGCN5) is a histone acetyltransferase that is essential for activation of target genes. GCN5 is a member of a large plays a crucial role in the transcriptional activation of specific family of histone acetyltransferases that are conserved target genes (2– 4). The yeast GCN5 (yGCN5) HAT shows a between yeast and humans. To understand the molecular strong preference for acetylation of lysine 14 of histone H3 in mechanisms of histone/protein acetylation, a detailed ki- vitro with reported broader specificity in vivo. In vivo yGCN5 netic analysis was performed. Bi-substrate kinetic analy- has been reported to acetylate lysine 9 and 18 of histone H3, sis using acetyl-coenzyme A (AcCoA) and an H3 histone residues 8 and 16 of histone H4, and lysine residues in the synthetic peptide indicated that both substrates must amino-terminal tail of histone H2B, albeit to a much lesser bind to form a ternary complex before catalysis. Product degree than lysine 14 of histone H3 (5, 6). Models have been inhibition studies revealed that the product CoA was a proposed in which these site-specific acetylations in the amino competitive inhibitor versus AcCoA. Desulfo-CoA, a dead- termini of histones lead to altered nucleosomal chromatin end inhibitor, also demonstrated simple competitive inhi- structure by disrupting histone-DNA contacts and histone-hi- bition versus AcCoA. Acetylated (Lys14Ac) H3 peptide dis- stone contacts (7). Enrichment of acetylation on specific lysine played noncompetitive inhibition against both H3 peptide residues suggests that differential acetylation within distinct and AcCoA. These results support a sequential ternary chromatin loci may play a key role in transcriptional regulation complex (ordered Bi-Bi) kinetic mechanism, where Ac- (2– 4). Several classes of HATs have been identified: p300/CBP CoA binds first, followed by H3 histone. Acetylated (8) cytosolic HAT 1 (9), P/CAF (10), TAFII250 (11), and SRC-1 (Lys14Ac) H3 product is released first, and CoA is the last (12). Among HAT enzymes, yGCN5, which is a P/CAF family product to leave. Also, two methods were developed to member, has been the most thoroughly characterized. yGCN5 measure the binding affinities of AcCoA/CoA for GCN5. eCoA, is typically found in two high molecular mass complexes, SAGA Employing the fluorescent CoA analog etheno-CoA ( -etheno-CoA), a K for eCoA of 5.1 6 1.1 mM was deter- (1.8 MDa) and Ada (0.8 MDa). The SAGA complex activates 1-N mined by fluorescence anisotropy. This value was similar transcription by association with acidic activation domains of value of 8.5 6 2.6 mM for AcCoA obtained using to the K various transcription factors and results in the acetylation of (inhibition constant) of equilibrium dialysis and to the K i nucleosomal histones H3 and H2B (13). The Ada complex has mM for CoA obtained from steady-state kinetic assays. 6.7 also been shown to acetylate nucleosomal histone H3 and his- Together, these data suggest that the acetyl moiety of tone H2B (14). AcCoA contributes little to the binding energy. Recently, several crystallographic and nuclear magnetic res- onance (NMR) molecular models have been solved for the cat- alytic domains of yGCN5, P/CAF, and a Tetrahymena homo- The histone acetyltransferase (HAT) GCN5 from Saccharo- logue (p55) of yGCN5 (15–17). These structures demonstrated myces cerevisiae catalyzes the transfer of an acetyl moiety from that the high degree of sequence homology between these en- acetyl-CoA (AcCoA) to the e-amino group of lysine 14 of histone zymes is manifested by a similar overall structure. The binary H3 (Scheme 1). GCN5 was originally identified as a transcrip- complex of P/CAFzCoA and the ternary complex of Tetrahy- tional activator that was necessary to promote maximal levels mena p55zCoAzH3 peptide revealed that residues contacting of GCN4-dependent transcription (1). CoA in the active site are highly conserved. In addition, a biochemical study (18) of yGCN5 enzymatic activity implicated yGCN5 the conserved glutamic acid 173 as a general base catalyst in AcCoA 1 histone H3| - 0 CoA 1 acetylated histone H3 the GCN5 HAT reaction, deprotonating the e-amino group of SCHEME 1 lysine 14 in histone H3. Elucidating the mechanism of HAT catalysis will provide a basis for understanding the link between histone acetylation * The costs of publication of this article were defrayed in part by the and gene activation. In this study we have systematically de- payment of page charges. This article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to termined the overall kinetic mechanism, developed methods for indicate this fact. measuring substrate/product binding affinities (fluorescence ‡ Supported by Postdoctoral Fellowship T32 DK07680. anisotropy and equilibrium dialysis), and determined the order § Supported by American Cancer Society Grant RPG-97-175-01-TBE of substrate binding and the order of product release by em- and National Institutes of Health Grant GM 59785-01. To whom cor- respondence should be addressed: Oregon Health Sciences University, ploying the product inhibitors CoA and acetylated (Lys14Ac) Dept. of Biochemistry and Molecular Biology, 3181 SW Sam Jackson H3 peptide and the dead-end inhibitor desulfo-CoA. The results Park Rd., Portland, OR 97201-3098. Tel.: 503-494-0644; Fax: 503-494- are consistent with a fully ordered Bi-Bi kinetic mechanism, 8393; E-mail: [email protected]. where AcCoA is the first substrate to bind, and CoA is the last The abbreviations used are: HATs, histone acetyltransferases; yGCN5, yeast GCN5; AcCoA, acetyl-coenzyme A; eCoA, 1-N -etheno-CoA. product released. 22048 This paper is available on line at http://www.jbc.org This is an Open Access article under the CC BY license. Kinetic Mechanism of Histone Acetyltransferase GCN5 22049 EXPERIMENTAL PROCEDURES v 5 V 3 @S#/@K ~1 1 I/@K ! 1 @S## (Eq. 4) m m is Materials—All chemicals were of the highest grade commercially v 5 V 3 @S#/@K ~1 1 I/K ! 1 [S]~1 1 I/K !# (Eq. 5) m m is ii available. Histone H3 peptide, ARTKQTARKSTGGKAPPKQLC, and the corresponding acetylated H3 peptide (Lys14Ac), corresponding to v 5 V 3 @S#/@K 1 @S#~1 1 I/K !# (Eq. 6) m m ii the 20 amino-terminal residues of human histone H3 and an additional carboxyl-terminal cysteine, were synthesized by the Protein Chemistry Determination of the Dissociation Constant for AcCoA Binding to Core Lab at The Baylor College of Medicine. Acetyl-CoA was purchased yGCN5 via Equilibrium Dialysis—Equilibration was performed using from Roche Molecular Biochemicals. Calf thymus histones were pur- Dispo-Equilibrium Dialyzers (Amika Corp.), which contains two 75-ml chased from Calbiochem. [ H]Acetyl-CoA (1.88 Ci/mmol) was from NEN chambers separated by a 5-kDa cut-off dialysis membrane. Equilibra- Life Sciences Products. P81 phosphocellulose disks were from Life tion conditions were 50 mM Tris, 50 mM Bis-Tris, 100 mM sodium Technologies, Inc. Dispo-Equilibrium Dialyzers were from Amika Corp. acetate, pH 7.5, and 32.5 mM yGCN5. The K value for AcCoA in the eCoA, 1-N -etheno-CoA was from Sigma. All other reagents were from presence of yGCN5 was determined by transferring 10 – 400 mM AcCoA Sigma or Fisher. (20 – 40 cpm H/pmol) into the buffer chamber and yGCN5 into Expression and Purification—The catalytic domain (amino acids 99 – the sample chamber. After 48 h of equilibration on a level shaker, 262) of yGCN5 was recombinantly expressed by isopropyl-B-D-thioga- samples were recovered from each chamber and counted by liquid lactopyranoside induction for 12 h at 25 °C and purified from BL21-DE3 scintillation to determine the amount of radioactivity in the buffer bacteria as described previously (18, 19). After cation exchange chro- chamber ([AcCoA] ) and the sample chamber ([AcCoA] 1 free free matography on S-Sepharose, fractions (assessed by SDS-polyacryl- [AcCoAzyGCN5]). The calculated [AcCoA] was subsequently sub- free amide gel electrophoresis) were pooled and concentrated and then sub- tracted from the calculated concentration [AcCoA] 1 free jected to size-exclusion chromatography on G-75 Sephadex. Purity was [AcCoAzyGCN5] to determine the concentration of bound AcCoA. The analyzed by SDS-PAGE, and fractions were concentrated and stored at data were presented in hyperbolic form (Equation 7). 220 °C until use. Protein concentrations were determined by the method of Bradford (20). @AcCoA z yGCN5# 5 ~@yGCN5] z @AcCoA] !/~K 1 @AcCoA] ) (Eq. 7) tot free d free Enzymatic Assays for yGCN5—yGCN5 histone acetyltransferase ac- Steady-state Fluorescence Anisotropy Measurements for eCoA Bind- tivity was monitored continuously using a Multiskan Ascent microplate ing to yGCN5—eCoA was titrated with increasing [yGCN5]. [eCoA] was reader (LabSystems, Franklin, MA) as described previously (19). determined spectrophotometrically by using a molar extinction coeffi- Briefly, the CoASH generated in the HAT reactions was continuously 21 21 cient of 5.6 mM cm at 275 nm. The anisotropy measurements were measured by using a coupled enzyme system with pyruvate dehydro- performed on a Photon Technologies QM-1 steady-state fluorescence genase. The CoASH-dependent oxidation of pyruvate was accompanied spectrophotometer equipped with a thermostat. eCoA was excited at by the reduction of NAD to NADH, which was measured spectrophoto- NADH 21 21 305 nm (3-nm slits), and emission was monitored at 405 nm with 20-nm metrically at 340 nm (e 5 6220 M cm ). The HAT assay reaction slits for both the parallel and perpendicular components. Measure- mixtures (70 ml) contained 0.2 mM NAD, 0.2 mM thiamine pyrophos- ments were performed in 50 mM Tris, pH 7.5, 25 °C, using a 200-ml phate, 5 mM MgCl ,1mM dithiothreitol, 2.4 mM pyruvate, and 0.03 quartz cuvette. A concentrated solution of yGCN5 (in 1-ml increments) units of pyruvate dehydrogenase (1 unit of dehydrogenase is defined by was added to the cuvette containing 0.5 mM eCoA and mixed using a the manufacturer (Sigma) to be the conversion of 1.0 mmol of b-NAD to small magnetic stirring bar. Anisotropy increases were measured after b-NADH/min at pH 7.4 at 30 °C). The assay buffer was 100 mM sodium 1–2 min. The anisotropy r of eCoA was calculated from the parallel Ivv acetate, 50 mM Bis-Tris, and 50 mM Tris and pH 7.5. All assays were and perpendicular Ivh polarized fluorescence intensities measured performed at 25 °C and were initiated by the addition of yGCN5. The upon parallel excitation according to Equation 8, rates were analyzed continuously for up to 5 min, and background rates resulting from the spontaneous formation of CoA were subtracted from r 5 ~Ivv 2 G z Ivh!/~Ivv 1 2 z G z Ivh! (Eq. 8) the initial velocities derived from the yGCN5-catalyzed reactions. Un- der these conditions, the coupled enzyme reaction does not limit the where G is the ratio of the parallel Ivh and perpendicular Ivh polarized observed initial velocities (19). Also, only the linear portion of the fluorescence intensity measured upon parallel excitation to correct for kinetic traces were used to determine the initial rates. These were the different efficiencies of the parallel and perpendicular polarization typically linear for up to 5 min. Stock solutions of the H3 peptide or fluorescence detectors. Data were fitted to Equation 9 described in acetylated (Lys14Ac) H3 peptide were prepared fresh daily in the pres- Richards et al. (22), ence of 5 mM dithiothreitol. Alternatively, yGCN5 activity was meas- ured using [ H]AcCoA, the P81-filter binding assay as described previ- A 5 ~~A 2 A ! z @yGCN5#/K 1 @yGCN5#! 1 A (Eq. 9) bound free d free ously (18). where A is the anisotropy measured at a given concentration of yGCN5, Bi-substrate Kinetic Measurements—The Bi-substrate kinetic analy- and A and A are the anisotropies of free and bound eCoA, free bound sis was performed at AcCoA concentrations spanning 1–20 mM and H3 respectively. This simplified equation can be used when the dissociation peptide or calf thymus histone concentrations spanning 25–730 mM. The constant is $10-fold higher than the fluorescent ligand concentration, yGCN5 HAT activity was monitored via the coupled-enzyme spectro- since the concentration of yGCN5 bound to eCoA becomes negligible in photometric assay using either pyruvate dehydrogenase or a-ketoglut- comparison to the total yGCN5 concentration (22). arate dehydrogenase. The data were fitted to three possible kinetic mechanisms: the sequential (ternary complex) mechanism equation RESULTS (Equation 1), the ping-pong (covalent intermediate) mechanism equa- Effect of Ionic Strength on yGCN5-catalyzed HAT Activity— tion (Equation 2), and the equilibrium-ordered equation (Equation 3), using the algorithms of Cleland (21) and the computer program Kin- Given the highly charged characteristic of the nucleosomal etAsyst (IntelliKinetics, State College, PA) and a nonlinear least complex, it is likely that ionic strength will greatly influence squares analysis. the catalytic activity of these enzymes. Therefore, the effect of ionic strength on the yGCN5 HAT reaction was examined by v 5 V 3 @A# 3 @B#/~~K 3 K ! 1 ~K 3 @B#! m ia b ma varying NaCl concentrations between 0 and 1 M, and the re- 1 ~K 3 @A#! 1 ~@A# 3 @B#!! (Eq. 1) sulting HAT activity was monitored in the radioactive P81- mb filter binding assay as described under “Experimental Proce- v 5 V 3 @A# 3 @B#/~~K 3 @B#! 1 ~K 3 @A#! 1 ~@A# 3 @B#!! (Eq. 2) m a b dures.” The resulting yGCN5 activity was plotted versus the log of the ionic strength and fitted to a line using least squares v 5 V 3 @A# 3 @B#/~~K 3 K ! 1 ~K 3 @A#! 1 ~@A# 3 @B#!! (Eq. 3) m a b b analysis (Fig. 1). A direct linear relationship between the log of Steady-state Inhibition—The product inhibitors CoA and acetylated the ionic strength and inactivation of yGCN5 HAT activity was (Lys14Ac) H3 peptide, as well as the substrate analog desulfo-CoA observed. There was an approximate 7-fold decrease of yGCN5 (lacking only the terminal sulfhydryl of CoA) were used in steady-state HAT activity between 0.15 M NaCl and 1 M NaCl. inhibition studies. To investigate whether competitive (Equation 4), Bi-substrate Kinetics—Yeast GCN5 catalyzes the transfer of noncompetitive (Equation 5), or uncompetitive (Equation 6) inhibition an acetyl moiety from AcCoA to the e-amino side chain of Lys was observed, the data were fitted to the respective inhibition equations 14 of histone H3 (Scheme 1). It was first necessary to determine based on the algorithms defined by Cleland (21) using a nonlinear least squares analysis. the basic kinetic mechanism (sequential or ping-pong) utilized 22050 Kinetic Mechanism of Histone Acetyltransferase GCN5 TABLE I Summary of kinetic and equilibrium constants for yeast GCN5 HAT Substrates K k /K k m cat m cat a 5 21 21 AcCoA 2.5 6 1.4 mM 6.8 3 10 M s 21 a 1.7 6 0.12 s a 3 21 21 H3 peptide 0.49 6 0.08 mM 3.5 3 10 M s Inhibitors K K K ii is d CoA 6.7 6 5.1 mM Desulfo-CoA 33 6 1.3 mM Etheno-CoA 5.1 6 1.1 mM a a AcH3 peptide (vs. CoA) 2.2 6 0.40 mM 1.9 6 0.15 mM a a AcH3 peptide (vs. H3) 2.5 6 1.3 mM 2.1 6 0.90 mM Experiments were performed in triplicate. The reported value is an average 6 S.D. FIG.1. Effect of ionic strength on yGCN5-catalyzed HAT activ- ity. NaCl concentrations were varied between 0 –1 M, and the resulting yGCN5 HAT activity was monitored in the spectrophotometric-coupled enzyme assay using pyruvate dehydrogenase. The conditions of the assay were 50 mM Tris, 50 mM Bis-Tris, 100 mM sodium acetate, pH 7.5, 25 °C, 75 nM yGCN5, and 10 mg/ml calf thymus histones. The resulting yGCN5 activity was plotted versus log of the ionic strength and fitted to a line using least squares analysis. FIG.3. CoA exhibits competitive inhibition toward AcCoA dur- ing yGCN5-catalyzed HAT activity. The data were presented in double-reciprocal form, where 1/velocity is plotted versus 1/[AcCoA] at a fixed concentration of 100 mM H3 peptide and various fixed concentra- tions of CoA: filled diamonds,0 mM; filled circles,50 mM; open diamonds, 100 mM; filled triangles, 300 mM; and open squares, 500 mM. The radio- active P81-filter binding assay was used to monitor activity. The reac- tion conditions were 50 mM Tris, 50 mM Bis-Tris, 100 mM sodium FIG.2. Ternary complex formation between yGCN5, AcCoA, acetate, pH 7.5, and 75 nM yGCN5 at 25 °C and a total reaction volume and H3 histone peptide. Double-reciprocal plots of 1/velocity versus of 55 ml. The experiment was performed in triplicate with a represent- 1/[H3 peptide] at several fixed concentrations of AcCoA are shown. Data ative plot displayed. were fitted to the equation for a sequential mechanism as described under “Experimental Procedures.” The coupled spectrophotometric as- say using pyruvate dehydrogenase was utilized to monitor H3 acetyla- demonstrated that yGCN5 utilizes a sequential (ternary com- tion at pH 7.5, 25 °C. The AcCoA concentrations were as follows: open plex) mechanism. Here, a synthetic H3 peptide (ARTKQTARK- boxes, 1.1; filled upright triangles, 2.0; open circles, 4.3; filled diamonds, 8.6; open triangles, 17.1; and filled inverted triangles,34 mM). The STGGKAPPKQLC) was employed as a substrate in lieu of calf concentration of yGCN5 was 70 nM, and H3 peptide concentrations thymus histones since this substrate provides a better defined spanned 25–730 mM. When the data were fitted to a covalent interme- system to analyze the steady-state kinetic parameters. diate (ping-pong) mechanism, the sum of the squares of the residuals To distinguish between a ternary complex (sequential) mech- was .100-fold larger than that for the sequential mechanism, which was .2-fold greater than an equilibrium ordered mechanism (described anism and a ping-pong (covalent-intermediate) mechanism, under “Experimental Procedures”). The experiment was performed in initial velocity steady-state kinetic parameters were obtained triplicate with a representative plot displayed. for AcCoA and H3 peptide using AcCoA concentrations span- ning 1–20 mM and H3 peptide concentrations spanning 25–730 by yGCN5 before a detailed investigation into the order of mM. Analysis of these data by plotting 1/velocity against 1/[H3 substrate binding and the order of product release could be peptide] at different fixed concentrations of AcCoA resulted in performed. CoA-dependent transferases are known to utilize an intersecting line pattern (Fig. 2) that is characteristic of a one of two distinct mechanisms to catalyze acetyl group trans- ternary complex (sequential) mechanism. In contrast, a ping- fer. One mechanism involves acetyl transfer from CoA to an pong (covalent intermediate) mechanism is typically character- enzyme side chain nucleophile before transfer to the amine ized by a parallel line pattern. The identification of a ternary substrate (23). In a subsequent step, the enzyme then transfers complex was consistent with our previously published report this acetyl group to the acceptor amine substrate. The alterna- (18) using a histone preparation derived from calf thymus, thus tive mechanism involves direct acetyl transfer from AcCoA to validating the use of H3 peptide as an excellent analog of the amine substrate acceptor without the formation of a cova- histone protein. The basic steady-state parameters derived lent enzyme intermediate (24). The latter mechanism requires from this Bi-substrate analysis are summarized in Table I. that both substrates and enzyme must form a ternary complex Employing acetylated (Lys14Ac) synthetic H3 peptide as a before catalysis can occur. In a previous study (18) we per- substrate for yGCN5, it was determined that lysine 14 of H3 formed a Bi-substrate kinetic analysis with core histone pro- peptide is the only residue of H3 peptide acetylated. No other teins (H3, H2A, H2B, and H4) derived from calf thymus and Lysine (Lys-4, Lys-9, or Lys-18) of H3 peptide was found to be Kinetic Mechanism of Histone Acetyltransferase GCN5 22051 SCHEME 2. Proposed kinetic mechanism of yGCN5. acetylated above background levels when using either the ra- hibition mechanism. This indicates that at high concentrations dioactive P81-filter binding assay or the coupled enzyme spec- of AcCoA, the inhibition by CoA is overcome, and the V max trophotometric assay (data not shown). This is an important values approach identical levels. CoA was found to be a linear result since acetylation at a secondary site to lysine 14 of H3 competitive inhibitor (Fig. 3) versus AcCoA at 100 mM H3 pep- peptide would significantly complicate the initial velocity ki- tide, with a K of 6.7 6 5.1 mM, similar to the value obtained for is netic analysis. (Fig. 2) the K for AcCoA (4.5 6 3.6 mM). The replot of the slopes Several distinct kinetic mechanisms are possible for enzyme from the double-reciprocal plot as a function of [CoA] was linear systems that employ two substrates and produce two products. and yielded similar inhibition constants (analysis not shown). Our previous study (18) combined with the Bi-substrate anal- These data indicated that AcCoA and CoA compete for the ysis described above have strongly argued against a ping-pong same form of yGCN5 and the same mutually exclusive binding (covalent intermediate) type mechanism and demonstrated site. In a sequential mechanism involving two substrates and that yGCN5 followed a sequential Bi-Bi mechanism. The bind- two products, this requires that AcCoA and CoA bind to free ing of substrates and release of products can be random, fully enzyme; thus, AcCoA is the first substrate to add, and CoA is ordered, or a combination of both. Product inhibition studies the last product to leave. In a complimentary experiment, CoA can easily distinguish among the various possibilities. did not inhibit against H3 peptide when AcCoA was present at Product Inhibition—To distinguish among the various pos- saturating levels (data not shown). These results are consistent sible kinetic models for substrate binding and product release, with CoA and H3 peptide binding to different forms of yGCN5 the products of the yGCN5-catalyzed HAT reaction were used and are consistent with the kinetic mechanism in which AcCoA to inhibit the yGCN5-catalyzed HAT reaction. To correctly binds first, followed by binding of H3 peptide and subsequent analyze product inhibition data, the reversibility of the yGCN5- transfer of the acetyl group from AcCoA to H3 and an ordered catalyzed reaction was examined. This was accomplished by a release of products with acetylated (Lys14Ac) H3 peptide re- high performance liquid chromatography assay using reversed leasing before the release of CoA (Scheme 2). phase chromatography to directly monitor the formation of To further demonstrate that H3 peptide and core histones deacetylated H3 peptide from a reaction of yGCN5, CoA, and operate through the same kinetic mechanism, CoA inhibition acetylated (Lys14Ac) H3 peptide. Analysis of the time points studies were carried out using calf thymus core histones as the from a reaction carried out at pH 7.5 and 25 °C with high amine nucleophile substrate (Fig. 4). As with H3 peptide (Fig. concentrations of acetylated (Lys14Ac) H3 peptide and CoA 3), CoA is a pure competitive inhibitor against AcCoA when calf with a corresponding non-enzyme control were subjected to thymus core histones are used. Again, this indicated that Ac- reversed phase chromatography. The estimated turnover num- CoA is the first substrate to add, and CoA is the last product to ber under this set of conditions (1.0 mM CoA, 1.0 mM acetylated be released. By default, H3 peptide binds second and acetylated (Lys14Ac) H3 peptide) was 0.00002 s (Lys14Ac) H3 peptide is the first to leave. The inhibition con- , ;100,000-fold lower than the k value (1.7 s ) determined for the forward reac- stants for CoA are summarized in Table I. Taken together with cat the fact that k tion under saturating substrates. Thus, yGCN5 is capable of values are similar between the two sources of cat 21 21 being a histone deacetylase as is predicted from thermodynam- substrate (k 5 1.7 s , k 5 0.74 s ) and that cat(H3) cat(histones) ics, albeit much less efficiently. both substrates (H3 peptide and calf thymus histones) generate Products of the yGCN5-catalyzed HAT reaction are classified intersecting line patterns in Bi-substrate kinetics, this strongly as competitive inhibitors if increasing concentrations of prod- suggests that H3 peptide and core histones follow the same uct decrease the apparent k /K for the varied substrate, kinetic mechanism. cat m Desulfo-CoA as a Dead-end Inhibitor—To further establish uncompetitive if they decrease the apparent k , and noncom- cat petitive if they decrease both. Evaluation of the patterns are that CoA acts as a pure competitive inhibitor, we utilized the obtained by observing increases in the slopes (apparent K CoA derivative desulfo-CoA in steady-state inhibition studies. k ), the intercepts (apparent 1/k ), or both, respectively, for Desulfo-CoA (which lacks only the terminal sulfhydryl of CoA) cat cat is a dead-end inhibitor since it forms a non-productive complex lines in double-reciprocal plots of 1/v versus 1/[S] obtained at increasing concentrations of product (and a constant concen- with yGCN5 and H3 peptide. Initial velocities were obtained, tration of the second substrate). and the data were plotted in double-reciprocal form with 1/ve- CoA as a Product Inhibitor—CoA was evaluated as a product locity versus 1/[AcCoA] at several fixed concentrations of des- inhibitor against both AcCoA and H3 peptide. The experiment ulfo-CoA. Desulfo-CoA was found to be a linear competitive was performed using the radioactive P81-filter binding assay inhibitor (Fig. 5) versus AcCoA at 100 mM histone H3 peptide, since CoA itself is a substrate for the coupled enzyme in the with a K of 33 6 1.3 mM (Table I). The binding affinity for CoA is spectrophotometric assay (19). Initial velocities were deter- and desulfo-CoA were not too dissimilar, with K values of 6.7 is mined, and the data were plotted in double-reciprocal form mM and 33 mM, respectively. The absence of the terminal sulf- with 1/velocity versus 1/[AcCoA] at several fixed concentrations hydryl in desulfo-CoA must account for the 4-fold higher K is of CoA. A series of double-reciprocal straight line plots inter- value for desulfo-CoA compared with CoA. The deletion of the sected on the 1/velocity ordinate, indicating a competitive in- terminal sulfhydryl likely results in a loss of Van der Waals 22052 Kinetic Mechanism of Histone Acetyltransferase GCN5 FIG.6. Determination of the dissociation constant (K ) for Ac- FIG.4. CoA exhibits competitive inhibition toward AcCoA dur- CoA binding to yGCN5 via equilibrium dialysis. The data are ing yGCN5-catalyzed HAT activity. The radioactive P81-filter bind- presented in hyperbolic form, where [AcCoAzyGCN5] is plotted versus ing assay was used to determine the effect of the product CoA on [AcCoA] . Equilibration conditions were 50 mM Tris, 50 mM Bis-Tris, free yGCN5-catalyzed HAT activity. The data are presented in double- 100 mM sodium acetate, pH 7.5, and '32.5 mM yGCN5. Equilibration reciprocal form, where 1/v is plotted versus 1/[AcCoA] at a fixed concen- was performed using Dispo-Equilibrium Dialyzers (Amika Corp.), tration of 150 mM calf thymus core histones and various fixed concen- which contain two 75-ml chambers separated by a 5-kDa molecular trations of CoA: filled diamonds,0 mM; open circles, 300 mM; filled mass cut-off dialysis membrane. The K value for AcCoA in the pres- diamonds, 500mM. The reaction conditions were 50 mM Tris, 50 mM ence of yGCN5 was determined by transferring 10 – 400 mM AcCoA Bis-Tris, 100 mM sodium acetate, pH 7.5, and 75 nM yGCN5 at 25 °C (20 – 40 cpm H/pmol) into the buffer chamber and yGCN5 into and a total reaction volume of 55 ml. the sample chamber. After equilibration (48 h) on a level shaker, samples were recovered from each chamber and counted by liquid scintillation to determine the amount of radioactivity in the buffer chamber ([AcCoA] ) and in the sample chamber ([AcCoA] 1 free free [AcCoAzyGCN5]) in order to determine the amount of bound AcCoA. The data were fitted to: [AcCoAzyGCN5] 5 ([yGCN5] 3 [AcCoA] )/ 0 free (K 1 [AcCoA] ). The experiment was performed in triplicate, with a d free representative plot displayed. value obtained from enzyme kinetic methods (K 5 6.7 is(CoA) mM). Steady-state Fluorescence Anisotropy Measurements for eCoA Binding to yGCN5—Changes in intrinsic protein fluorescence are often utilized to follow protein ligand interactions. How- ever, we observed no significant change in the fluorescent prop- erties of yGCN5 (l or quantum yield) upon binding AcCoA. max Therefore, we sought an alternative fluorescent approach, flu- orescence anisotropy (25). Steady-state fluorescence anisotropy FIG.5. Desulfo-CoA as a product inhibitor of the yGCN5-cata- measurements for eCoA (1,N -etheno-CoA) binding to yGCN5 lyzed HAT reaction. yGCN5 HAT activity was monitored in the radioactive P81-filter binding assay. Double-reciprocal plots of were carried out to determine the dissociation constant (K ). 1/velocity versus 1/[AcCoA] at several fixed concentrations of desulfo- eCoA contains a fluorescent derivative of adenine (containing CoA. Data were fitted to the equation for competitive inhibition as two additional carbon atoms) that has been utilized to study described under “Experimental Procedures.” The desulfo-CoA concen- many ATP-, NAD-, and CoA-dependent enzymes (26 –30). Two trations were as follows: filled diamonds,0 mM; open circles, 100 mM; open diamonds, 300 mM, and filled triangles, 500 mM. The reaction separate x-ray structures of CoA and GCN5 complexes demon- conditions were 50 mM Tris, 50 mM Bis-Tris, 100 mM sodium acetate, pH strated that there are no significant interactions between 7.5, at 25 °C and 75 nM yGCN5 and a total reaction volume of 55 ml. GCN5 and the adenine moiety of CoA (15, 16). We therefore reasoned that eCoA could be successfully employed as a rele- interactions with residues in the hydrophobic CoA binding site vant CoA analog. As expected, upon binding yGCN5, eCoA (15, 16). Together, these data indicate that CoA or desulfo-CoA exhibited a change in fluorescence emission polarization (ani- only bind to the free form of the enzyme and are consistent with sotropy). We used this change in polarization to generate a the kinetic mechanism of Scheme 2. binding curve and to determine the dissociation constant (Fig. Equilibrium Dialysis—Using equilibrium dialysis, the disso- 7). Using this analysis, we calculated a dissociation constant ciation constant for AcCoA binding to yGCN5 was determined. (K )of5.1 6 1.1 mM. A control experiment in which excess Varying concentrations of [ H]AcCoA were transferred into the AcCoA was added resulted in the loss of the change in fluores- buffer chamber, and 32.5 mM yGCN5 was transferred into the cence emission polarization (data not shown). The K value is sample chamber. After equilibration, samples were recovered in excellent agreement with the K of 6.7 mM obtained for CoA is from each chamber and were counted by liquid scintillation. (Fig. 3), suggesting that eCoA is a valid replacement for CoA in The data were analyzed as described under “Experimental these binding studies. Procedures” and the Fig. 6 legend. A representative binding Acetylated (Lys14Ac) H3 peptide as a Product Inhibitor—To experiment is shown in Fig. 6. The average value from three complete the product inhibition studies, the product inhibitor separate experiments yielded a K value of 8.5 6 2.6 mM and an acetylated (Lys14Ac) H3 peptide was examined as an inhibitor enzyme concentration of 33.6 6 2.5 mM. These data are in versus both H3 peptide and AcCoA. As predicted by the kinetic excellent agreement with the calculated enzyme concentration model (Scheme 2), both effects were found to be noncompetitive (32.5 mM determined by the method of Bradford (20)) and the K (mixed type), indicating that the substrate/inhibitor pairs bind i Kinetic Mechanism of Histone Acetyltransferase GCN5 22053 DISCUSSION to different enzyme forms and that there is a reversible con- nection between the two points of binding (Fig. 8). Interest- Ternary Complex Formation—The enzyme kinetic mecha- ingly, the slope effects (reversible connection) were less dra- nism of two substrate reactions can be directly probed by si- matic, suggesting that the reversible connection is quite weak. multaneously varying both substrates (31). It is well estab- Our observation that the reverse reaction is extremely slow lished that a double-reciprocal plot analysis that generates an (about 0.00002 s ), ;100,000-fold lower than for the forward intersecting line pattern suggests a ternary complex mecha- reaction (1.7 s ), would provide a reasonable explanation for nism, and a parallel line pattern is characteristic of a ping-pong the weaker reversible connection. The steady-state kinetic in- (covalent intermediate) mechanism. This approach has been previously used to predict the behavior of other well character- hibition parameters for acetylated (Lys14Ac) H3 peptide are summarized in Table I. These results, in conjunction with the ized acetyltransferases (23, 24). Here, the resulting Bi-sub- strate intersecting line pattern with yGCN5 (Fig. 2) is consist- CoA and desulfo-CoA inhibition data, are consistent with ent with a ternary complex mechanism. These data are yGCN5 catalyzed HAT activity following kinetic Scheme 2. consistent with a direct attack of lysine on AcCoA. Although it is impossible to exclude the existence of a covalent enzyme intermediate by this approach, there has been no direct or indirect evidence for such an intermediate. Using quench-flow trapping approaches, we were unable to detect a covalent acetylated enzyme intermediate when yGCN5 was rapidly mixed with AcCoA. If an intermediate is formed in the yGCN5 reaction, it must occur after both AcCoA and H3 peptide are bound, and it must break down before either acetylated H3 peptide or CoA is released. However, our data strongly suggest that yGCN5 forms a ternary complex with AcCoA and H3 peptide before catalysis. In this study, we employed a synthetic peptide substrate (ARTKQTARKSTGGKAPPKQL) corresponding to the amino- terminal 20 amino acids of histone H3. Previously, it was demonstrated that commercial preparations of calf thymus cores histones also generated an intersecting-line pattern from similar Bi-substrate kinetic analyses (18). Moreover, the k cat FIG.7. Steady-state fluorescence anisotropy measurements values obtained using either substrate are similar. Together, for eCoA binding to GCN5. eCoA was titrated with increasing these results indicate that both the H3 peptide and H3 protein [GCN5]. Measurements were performed in 50 mM Tris buffer, pH 7.5, (within the core octamer) follow the same kinetic mechanism 25 °C, using a 200-ml cuvette. Concentrated yGCN5 (in 1-ml increments) was added to the quartz cuvette containing 0.5 mM eCoA and mixed and that synthetic H3 peptide is a valid and relevant substrate using a small magnetic stirring bar. Anisotropy increases were meas- in detailed mechanistic studies. In fact, the synthetic peptide ured after 1–2 min. The anisotropy r of eCoA was calculated from the substrate affords many advantages over a heterogeneous prep- parallel Ivv, and perpendicular Ivh polarized fluorescence intensities aration of histones derived from calf thymus. First, the syn- were measured upon parallel excitation as described under “Experi- mental Procedures.” Data were fitted to the following equation, de- thetic H3 peptide is completely unacetylated. The acetylation scribed in Richards et al. (22): A 5 ((A 2 A )z[yGCN5]/K 1 state of histones derived from calf thymus is ill-defined and bound free d [yGCN5]) 1 A , where A is the anisotropy measured at a given free likely varies among preparations. Second, the non-enzymatic concentration of yGCN5, and A and A are the anisotropies of free free bound acetylation reaction rate is greatly reduced with H3 peptide eCoA. This simplified equation can be used when the disso- and bound since there are only four possible lysine acetylation sites com- ciation constant is $10-fold higher than the fluorescent ligand concen- tration, since the concentration of yGCN5 bound to eCoA becomes negligible in comparison to the total yGCN5 concentration (22). The fit yielded the values A 5 0.029 6 0.004, A 5 0.151 6 0.010, and K K. G. Tanner, M. R. Langer, Y. Kim, and J. M. Denu, unpublished free bound d 5 5.1 6 1.1 mM with r 5 0.986. observation. FIG.8. Inhibition of GCN5 by the product acetylated H4 (Lys14Ac) peptide. Data are presented in double-reciprocal form, where 1/v is plotted versus 1/[H3] (left panel) and 1/[AcCoA] (right panel) at various fixed concentrations of acetylated H3 (Lys14Ac) peptide (circles,0mM; triangles, 1.0 mM; and squares, 2.0 mM). The yGCN5 concentration was 70 nM, [AcCoA] was 65 mM (left panel), and [H3] was 233 mM (right panel). The reaction conditions were 50 mM Tris, 50 mM Bis-Tris, 100 mM sodium acetate, pH 7.5, and 25 °C. Each data set were fitted to a noncompetitive inhibition equation (KinetAsyst, IntelliKinetics, State College, PA). The experiments were performed in triplicate, with a representative plot displayed. 22054 Kinetic Mechanism of Histone Acetyltransferase GCN5 pared with calf thymus histones, which contain four core his- binds to EzCoA. Also, acetylated (Lys14Ac) H3 peptide exhib- tones H3, H2A, H2B, H4 along with linker H1. Third, we are ited noncompetitive inhibition with respect to H3 peptide, in- able to use the well defined product, synthetic acetylated dicating that acetylated and unacetylated H3 peptide bind to (Lys14Ac) H3 peptide, as a product inhibitor for the yGCN5- different enzyme forms. That is, H3 peptide binds to EzAcCoA, catalyzed HAT reaction. These advantages have allowed us to whereas acetylated H3 peptide binds to EzCoA. Interestingly, completely define the kinetic mechanism, as described in the slope effects (reversible connection) were less dramatic, Scheme 2. suggesting that the reversible connection is quite weak. Our The recently solved structures (both x-ray and NMR) of observation that the reverse reaction is extremely slow GCN5 and GCN5 homologues have implicated the formation of (;0.00002 s ) would provide a reasonable explanation for the a ternary complex of GCN5, AcCoA, and H3 peptide (15–17). weaker reversible connection. The product CoA exhibited pure competitive inhibition toward AcCoA (Fig. 3), indicating that The molecular models revealed two pronounced clefts that are roughly orthogonal to one another. Two of the structures have AcCoA and CoA compete for the same enzyme form. In a either AcCoA or CoA bound to the smaller of the clefts, whereas sequential mechanism, this requires that acetyl-CoA and CoA another revealed a ternary complex of CoA bound to the bind to free enzyme; thus, AcCoA is the first substrate to add, smaller cleft and an 11-residue H3 peptide, KSTGGKAPRKQ, and CoA is the last product to be released (Scheme 2). As bound to the larger cleft. The fact that the acetyl moiety of predicted by the kinetic model, CoA does not inhibit against H3 AcCoA was not transferred to the enzyme in the binary com- when AcCoA is present at saturating levels (data not shown). plex of Tetrahymena GCN5zAcCoA is consistent with the ab- Consistent with CoA acting as a pure competitive inhibitor, we sence of a covalent acetylated enzyme intermediate formation utilized the CoA derivative desulfo-CoA in our steady-state during catalysis. Analysis of the ternary GCN5zCoAzH3 com- inhibition studies and found that desulfo-CoA also acts as a plex revealed that it is possible for both H3 peptide and AcCoA pure competitive inhibitor with respect to acetyl-CoA (Fig. 5). to be bound simultaneously. This dead-end complex should be Desulfo-CoA is a dead-end inhibitor since it forms a non-pro- similar to that of a ternary complex of GCN5 with both sub- ductive complex. Taken together with the fact that acetylated strates AcCoA and H3 peptide, since the majority of the (Lys14Ac) H3 peptide is a linear noncompetitive (mixed) inhib- CoAzGCN5 interactions involve the pyrophosphate group and itor against both H3 peptide and AcCoA (Fig. 8), this suggests pantetheine arm of CoA. The crystal structure predicts that that AcCoA is the first to bind yGCN5 followed by H3 peptide. there is ample room to accommodate the acetyl moiety of Ac- After transfer of the acetyl moiety from AcCoA to H3 peptide in CoA. In fact, the carbonyl of the acetyl group would be posi- the ternary complex, acetylated (Lys14Ac) H3 peptide must tioned closer to both the nucleophile e-amino group of lysine 14 leave first, followed by the subsequent release of CoA (Scheme and the proposed general base glutamic acid 173, which must 2). The data suggest that if H3 peptide is capable of binding to deprotonate the e-amino group of lysine 14 before attack on the yGCN5, it does so only weakly. carbonyl carbon of AcCoA. Indeed, we had demonstrated pre- Given our kinetic and biochemical evidence that this highly viously that the conserved glutamic acid 173 residue was acting conserved family of GCN5 HATs utilize a fully ordered Bi-Bi as a general base (18). This study revealed an ionization of a kinetic mechanism, it is likely that binding of AcCoA to yGCN5 residue that must be unprotonated for activity (pK of 8.3 6 induces a conformational change that then optimizes the 0.2). Using site-directed mutagenesis and pH activity profiles, pocket for H3 peptide binding. Consistent with this, a recently the observed ionization was assigned to Glu-173. We concluded solved structure of a GCN5-related N-acetyltransferase (GNAT) that Glu-173 was functioning as a general base, deprotonating family member, serotonin N-acetyltransferase complexed, with the e-amino group of lysine 14 from histone H3. From later a Bi-substrate analogue predicts that binding of AcCoA will structural studies, it was discovered that two large hydropho- induce a conformational change that exposes the binding site bic clefts converged at the location of glutamic acid 173. This for the amine substrate serotonin (32). This is in good agree- hydrophobic environment surrounding glutamic acid 173 is ment with the kinetic mechanism of serotonin N-acetyltrans- consistent with the observed elevated pK (16, 18). ferase proposed by Cole and co-workers (33). As we have dem- Product Inhibition Studies—The Bi-substrate kinetic analy- onstrated for yGCN5, serotonin N-acetyltransferase was shown sis described here strongly suggests that a ternary complex to obey an ordered Bi-Bi mechanism. Moreover, the crystal mechanism is utilized by GCN5. To determine the order of structure of the Tetrahymena GCN5zCoAzH3 peptide ternary substrate binding and the order of product release, product complex suggests that the binding of H3 peptide is dependent inhibition studies were undertaken. The type of product inhi- on AcCoA binding first (16). The binding of AcCoA was sug- bition observed is diagnostic for a particular kinetic scheme gested to widen the histone substrate binding groove. (31). Typically, one substrate and one product are varied at Equilibrium Binding Studies—Measuring the binding affin- saturating or subsaturating levels of the other substrate, and ities of substrates and products is an important goal toward the resulting double-reciprocal plots indicate the type of inhi- understanding the catalytic and kinetic mechanisms of the bition. The two HAT reaction products, acetylated (Lys14Ac) HAT enzymes. Therefore, we developed a steady-state fluores- H3 peptide and CoA, were employed as product inhibitors. The cence anisotropy assay for eCoA binding to yGCN5. eCoA (1,N - fully ordered mechanism of Scheme 2 was easily distinguished etheno-CoA), a fluorescent analogue of CoA, is identical to CoA from other possible random mechanisms. Acetylated (Lys14Ac) except for two additional carbon atoms within the adenine H3 peptide was examined as an inhibitor versus both H3 pep- moiety. The structures of the binary and ternary complexes of tide and AcCoA (Fig. 8). As predicted by the kinetic model GCN5 with CoA or AcCoA suggested that this would be a (Scheme 2), both effects were found to be noncompetitive logical approach since the majority of the contacts involve the (mixed type), indicating that the substrate/inhibitor pairs bind pyrophosphate group and pantetheine arm of AcCoA/CoA. In to different enzyme forms and that there is a reversible con- fact, the adenine base in these structures is highly disordered, nection between the two points of binding. Acetylated with no specific interactions with GCN5 residues (15, 16). The (Lys14Ac) H3 peptide demonstrated noncompetitive inhibition binding curve generated from these anisotropy measurements with respect to AcCoA, indicating that acetylated (Lys14Ac) H3 yielded a dissociation constant (K )of5.1 mM (Fig. 7). This peptide and AcCoA do not bind to the same form of yGCN5 (Fig. value is in good agreement with the K of 6.7 mM for CoA (Fig. is 8). That is, AcCoA binds to E, whereas acetylated H3 peptide 3), determined in the inhibition studies. By equilibrium dialy- Kinetic Mechanism of Histone Acetyltransferase GCN5 22055 (1997) Nature 389, 194 –198 sis, a dissociation constant (K )of8.5 6 2.6 mM was determined 13. Grant, P. A., Duggan, L., Cote, J., Roberts, S. M., Brownell, J. E., Candau, R., for AcCoA binding to free enzyme (Fig. 6). The fact that the Ohba, R., Owen-Hughes, T., Allis, C. D., Winston, F., Berger, S. L., and Workman, J. L. (1997) Genes Dev. 11, 1640 –1650 dissociation constants determined by these binding assays 14. Eberharter, A., Sterner, D. E., Schieltz, D., Hassan, A., Yates, J. R. R., Berger, were in good agreement suggests that AcCoA and CoA bind S. L., and Workman, J. L. (1999) Mol. Cell. Biol. 19, 6621– 6631 with similar binding affinities to yGCN5. The observation that 15. Clements, A., Rojas, J. R., Trievel, R. C., Wang, L., Berger, S. L., and Marmorstein, R. (1999) EMBO J. 18, 3521–3532 the binding affinities between substrate AcCoA and product 16. Rojas, J. R., Trievel, R. C., Zhou, J., Mo, Y., Li, X., Berger, S. L., Allis, C. D., CoA are similar predicts that the intracellular ratio of AcCoA/ and Marmorstein, R. (1999) Nature 401, 93–98 CoA would be a critical determinant for in vivo HAT activity. 17. Lin, Y., Fletcher, C. M., Zhou, J., Allis, C. D., and Wagner, G. (1999) Nature 400, 86–89 This possible regulatory mechanism for HATs would link tran- 18. Tanner, K. G., Trievel, R. C., Kuo, M. H., Howard, R. M., Berger, S. L., Allis, scriptional activation/repression with various cellular meta- C. D., Marmorstein, R., and Denu, J. M. (1999) J. Biol. Chem. 274, 18157–18160 bolic pathways that control the fluctuating concentrations of 19. Kim, Y., Tanner, K. G., and Denu, J. M. (2000) Anal. Biochem. 280, 308 –314 AcCoA/CoA. It will be interesting to determine whether this is 20. Bradford, M. M. (1976) Anal. Biochem. 72, 248 –254 a common feature of all HAT families. 21. Cleland, W. W. (1977) Adv. Enzymol. Relat. Areas Mol. Biol. 45, 273–387 22. Richards, J. P., Bachinger, H. P., Goodman, R. H., and Brennan, R. G. (1996) REFERENCES J. Biol. Chem. 271, 13716 –13723 23. Thompson, S., Mayerl, F., Peoples, O. P., Masamune, S., Sinskey, A. J., and 1. Georgakopoulos, T., and Thireos, G. (1992) EMBO J. 11, 4145– 4152 Walsh, C. T. (1989) Biochemistry 28, 5735–5742 2. Brownell, J. E., and Allis, C. D. (1996) Curr. Opin. Genet. Dev. 6, 176 –184 24. Lewendon, A., Murray, I. A., Shaw, W. V., Gibbs, M. R., and Leslie, A. G. (1994) 3. Wade, P. A., Pruss, D., and Wolffe, A. P. (1997) Trends Biochem. Sci. 22, Biochemistry 33, 1944 –1950 128 –132 25. Lundblad, J. R., Laurance, M., and Goodman, R. H. (1996) Mol. Endocrinol. 10, 4. Grunstein, M. (1997) Nature 389, 349 –352 607– 612 5. Grant, P. A., Eberharter, A., John, S., Cook, R. G., Turner, B. M., and 26. Basso, L. A., Engel, P. C., and Walmsley, A. R. (1997) Biochim. Biophys. Acta Workman, J. L. (1999) J. Biol. Chem. 274, 5895–5900 1340, 63–71 6. Zhang, W., Bone, J. R., Edmondson, D. G., Turner, B. M., and Roth, S. Y. (1998) 27. Klebl, B. M., and Pette, D. (1996) Anal. Biochem. 239, 145–152 EMBO J. 17, 3155–3167 28. Kasprzak, A. A., and Villafranca, J. J. (1988) Biochemistry 27, 8050 – 8056 7. Hansen, J. C., Tse, C., and Wolffe, A. P. (1998) Biochemistry 37, 17637–17641 29. Atkins, W. M., Cader, B. M., Hemmingsen, J., and Villafranca, J. J. (1993) 8. Bannister, A. J., and Kouzarides, T. (1996) Nature 384, 641– 643 Protein Sci. 2, 800 – 813 9. Dutnall, R. N., Tafrov, S. T., Sternglanz, R., and Ramakrishnan, V. (1998) Cell 30. Favilla, R., Martin, S. R., and Bayley, P. M. (1988) Biochim. Biophys. Acta 955, 94, 427– 438 321–329 10. Yang, X. J., Ogryzko, V. V., Nishikawa, J., Howard, B. H., and Nakatani, Y. (1996) Nature 382, 319 –324 31. Segel, I. H. (1993) Enzyme Kinetics, pp. 506 – 845, John Wiley & Sons, Inc., New York 11. Imhof, A., Yang, X. J., Ogryzko, V. V., Nakatani, Y., Wolffe, A. P., and Ge, H. (1997) Curr. Biol. 7, 689 – 692 32. Hickman, A. B., Klein, D. C., and Dyda, F. (1999) Mol. Cell 3, 23–32 33. De Angelis, J., Gastel, J., Klein, D. C., and Cole, P. A. (1998) J. Biol. Chem. 12. Spencer, T. E., Jenster, G., Burcin, M. M., Allis, C. D., Zhou, J., Mizzen, C. A., McKenna, N. J., Onate, S. A., Tsai, S. Y., Tsai, M. J., and O’Malley, B. W. 273, 3045–3050

Journal

Journal of Biological Chemistry – Unpaywall

Published: Jul 1, 2000

There are no references for this article.