Key points are not available for this paper at this time.
Within chromatin, reversible acetylation of core histones is critical for transcriptional activation of eukaryotic target genes. The recent identification of intrinsic histone acetyltransferase (HAT) catalytic activity from a number of transcriptional co-activators (including yeast GCN5, p300/CBP, P/CAF, and TAFII250), has underscored the importance of protein acetylation in transcriptional control. The GCN5 family is the prototype for a diverse group of at least four distinct human HATs families. Although there is now a clear link between in vivo HAT catalytic activity and gene activation, little is known about the molecular mechanisms of histone acetylation. Herein, we report the first detailed biochemical study that probes the catalytic mechanism and the function of invariant glutamic acid 173 within the GCN5 family of HATs. Our results suggest that the HAT reaction involves the formation of a ternary complex (histones, acetyl-CoA, and enzyme) where the ε-amino group of histone lysine residues directly attacks the bound acetyl-CoA. The acetylation reaction requires deprotonation of the ε-amino group prior to nucleophilic attack. Employing site-directed mutagenesis, chemical modification, steady-state, and pH-dependent rate analysis, it is demonstrated that glutamic acid 173 is an essential catalytic residue, acting as a general base catalyst by deprotonating the histone substrate. Within chromatin, reversible acetylation of core histones is critical for transcriptional activation of eukaryotic target genes. The recent identification of intrinsic histone acetyltransferase (HAT) catalytic activity from a number of transcriptional co-activators (including yeast GCN5, p300/CBP, P/CAF, and TAFII250), has underscored the importance of protein acetylation in transcriptional control. The GCN5 family is the prototype for a diverse group of at least four distinct human HATs families. Although there is now a clear link between in vivo HAT catalytic activity and gene activation, little is known about the molecular mechanisms of histone acetylation. Herein, we report the first detailed biochemical study that probes the catalytic mechanism and the function of invariant glutamic acid 173 within the GCN5 family of HATs. Our results suggest that the HAT reaction involves the formation of a ternary complex (histones, acetyl-CoA, and enzyme) where the ε-amino group of histone lysine residues directly attacks the bound acetyl-CoA. The acetylation reaction requires deprotonation of the ε-amino group prior to nucleophilic attack. Employing site-directed mutagenesis, chemical modification, steady-state, and pH-dependent rate analysis, it is demonstrated that glutamic acid 173 is an essential catalytic residue, acting as a general base catalyst by deprotonating the histone substrate. Histone acetyltransferases (HATs) 1The abbreviations used are: HAT, histone acetyltransferase; DEPC, diethyl pyrocarbonate; EDAC, 1-ethyl-3(3-dimethylaminopropyl) carbodiimide.catalyze acetyl group transfer from acetyl-CoA to the ε-amino group of specific lysine residues within histone amino termini. Recent identification of intrinsic HAT catalytic activity from a number of transcriptional co-activators, including yeast GCN5 (1Brownell J.E. Zhou J. Ranalli T. Kobayashi R. Edmondson D.G. Roth S.Y. Allis C.D. Cell. 1996; 84: 843-851Abstract Full Text Full Text PDF PubMed Scopus (1286) Google Scholar), p300/CBP (2Bannister A.J. Kouzarides T. Nature. 1996; 384: 641-643Crossref PubMed Scopus (1533) Google Scholar), P/CAF (3Yang X.J. Ogryzko V.V. Nishikawa J. Howard B.H. Nakatani Y. Nature. 1996; 382: 319-324Crossref PubMed Scopus (1317) Google Scholar), and TAFII250 (4Mizzen C.A. Yang X.J. Kokubo T. Brownell J.E. Bannister A.J. Owen-Hughes T. Workman J. Wang L. Berger S.L. Kouzarides T. Nakatani Y. Allis C.D. Cell. 1996; 87: 1261-1270Abstract Full Text Full Text PDF PubMed Scopus (623) Google Scholar), has prompted the evaluation of histone acetylation in transcriptional control. Reversible acetylation of core histones is critical for transcriptional activation of eukaryotic target genes (5Brownell J.E. Allis C.D. Curr. Opin. Genet. Dev. 1996; 6: 176-184Crossref PubMed Scopus (468) Google Scholar, 6Wade P.A. Pruss D. Wolffe A.P. Trends Biochem. Sci. 1997; 22: 128-132Abstract Full Text PDF PubMed Scopus (410) Google Scholar, 7Grunstein M. Nature. 1997; 389: 349-352Crossref PubMed Scopus (2400) Google Scholar). By alleviating the transcriptionally repressive interaction between positively charged histone amino termini and negatively charged DNA and thus providing direct access of the transcription machinery, hyperacetylation of distinct loci within chromatin is thought to give rise to gene activation (reviewed in Refs. 5Brownell J.E. Allis C.D. Curr. Opin. Genet. Dev. 1996; 6: 176-184Crossref PubMed Scopus (468) Google Scholar, 6Wade P.A. Pruss D. Wolffe A.P. Trends Biochem. Sci. 1997; 22: 128-132Abstract Full Text PDF PubMed Scopus (410) Google Scholar, 7Grunstein M. Nature. 1997; 389: 349-352Crossref PubMed Scopus (2400) Google Scholar). However, several aspects of this model have proven to be more complex. Some evidence supports the concept that the amino termini participate primarily in protein-protein interactions rather than protein-DNA interactions (reviewed in Ref. 8Hansen J.C. Tse C. Wolffe A.P. Biochemistry. 1998; 37: 17637-17641Crossref PubMed Scopus (216) Google Scholar). It has been suggested that the physical consequence of histone tail acetylation is the disruption of structured domains in the amino termini, rather than the strict neutralization of positive charge (reviewed in Ref. 8Hansen J.C. Tse C. Wolffe A.P. Biochemistry. 1998; 37: 17637-17641Crossref PubMed Scopus (216) Google Scholar). Enrichment of acetylation on specific lysine residues suggests that differential acetylation within distinct chromatin loci may play a key role in regulation (5Brownell J.E. Allis C.D. Curr. Opin. Genet. Dev. 1996; 6: 176-184Crossref PubMed Scopus (468) Google Scholar). Also, the notion that nonhistone proteins may be the actual targets of many HATs has been proposed (9Imhof A. Yang X.J. Ogryzko V.V. Nakatani Y. Wolffe A.P. Ge H. Curr. Biol. 1997; 7: 689-692Abstract Full Text Full Text PDF PubMed Scopus (535) Google Scholar, 10Gu W. Roeder R.G. Cell. 1997; 90: 595-606Abstract Full Text Full Text PDF PubMed Scopus (2177) Google Scholar), because acetylation of the general transcription factors by various HATs (9Imhof A. Yang X.J. Ogryzko V.V. Nakatani Y. Wolffe A.P. Ge H. Curr. Biol. 1997; 7: 689-692Abstract Full Text Full Text PDF PubMed Scopus (535) Google Scholar) and the acetylation of the tumor suppressor p53 by p300 (10Gu W. Roeder R.G. Cell. 1997; 90: 595-606Abstract Full Text Full Text PDF PubMed Scopus (2177) Google Scholar) have been demonstrated in vitro. Despite the explosion of recent reports linking transcription and reversible protein/histone acetylation, the molecular mechanisms remain elusive. At least four distinct human HAT families have been identified (11Kouzarides T. Curr. Opin. Genet. Dev. 1999; 9: 40-48Crossref PubMed Scopus (590) Google Scholar). The prototypical GCN5 family of HATs constitutes a large group of enzymes found in such diverse organisms as yeast,Tetrahymena, and humans (5Brownell J.E. Allis C.D. Curr. Opin. Genet. Dev. 1996; 6: 176-184Crossref PubMed Scopus (468) Google Scholar). The GCN5 family exhibits specificity toward lysine 14 of H3 and lysines 8 and 16 of H4 (12Kuo M.H. Brownell J.E. Sobel R.E. Ranalli T.A. Cook R.G. Edmondson D.G. Roth S.Y. Allis C.D. Nature. 1996; 383: 269-272Crossref PubMed Scopus (507) Google Scholar). GCN5 enzymes are members of a superfamily ofN-acetyltransferases that appear to share a common structural core (13Neuwald A.F. Landsman D. Trends Biochem. Sci. 1997; 22: 154-155Abstract Full Text PDF PubMed Scopus (387) Google Scholar). Although the sequence identity is limited among these diverse enzymes, this region is predicted to form a similar acetyl-CoA binding structure (13Neuwald A.F. Landsman D. Trends Biochem. Sci. 1997; 22: 154-155Abstract Full Text PDF PubMed Scopus (387) Google Scholar, 14Dutnall R.N. Tafrov S.T. Sternglanz R. Ramakrishnan V. Cell. 1998; 94: 427-438Abstract Full Text Full Text PDF PubMed Scopus (196) Google Scholar, 15Wolf E. Vassilev A. Makino Y. Sali A. Nakatani Y. Burley S.K. Cell. 1998; 94: 439-449Abstract Full Text Full Text PDF PubMed Scopus (189) Google Scholar, 16Hickman A.B. Klein D.C. Dyda F. Mol. Cell. 1999; 3: 23-32Abstract Full Text Full Text PDF PubMed Scopus (117) Google Scholar). Several x-ray structures of other CoA-dependent transferases in this superfamily have recently been reported (14Dutnall R.N. Tafrov S.T. Sternglanz R. Ramakrishnan V. Cell. 1998; 94: 427-438Abstract Full Text Full Text PDF PubMed Scopus (196) Google Scholar, 15Wolf E. Vassilev A. Makino Y. Sali A. Nakatani Y. Burley S.K. Cell. 1998; 94: 439-449Abstract Full Text Full Text PDF PubMed Scopus (189) Google Scholar, 16Hickman A.B. Klein D.C. Dyda F. Mol. Cell. 1999; 3: 23-32Abstract Full Text Full Text PDF PubMed Scopus (117) Google Scholar); however, the catalytic mechanism has not been established. Here, we report the first detailed biochemical study that probes the catalytic mechanism and the function of invariant glutamic acid 173 within the GCN5 family of HATs. Our results suggest that the HAT reaction does not involve the formation of a covalent acetyl-enzyme intermediate but rather involves the formation of a ternary complex (histones, acetyl-CoA, and enzyme) where the ε-amino group of histone lysine residues directly attacks the bound acetyl-CoA. We demonstrate that glutamic acid 173 is an essential catalytic residue, acting as a general base catalyst by deprotonating the histone substrate. All chemicals were of the highest grade commercially available. Calf thymus histones where purchased from Calbiochem. Amino acids 99–262 of yeast GCN5 corresponding to the HAT catalytic domain were overexpressed in BL21-DE3 bacteria. T7 based protein expression of GCN5 was induced for 3.5 h. Harvested cells were lysed by French pressure in 50 mm sodium phosphate (pH 8), 300 mm NaCl, 1 mm β-mercaptoethanol, 10% glycerol with protease inhibitors (1 mmphenylmethylsulfonyl fluoride, 10 μg/ml leupeptin, and 5 μg/ml aprotinin). GCN5 from clarified extract was adsorbed over a 4-ml column of Ni2+ affinity resin (Qiagen) and eluted with a 0–50 mm imidazole gradient in the above buffer. GCN5 eluted at 12 mm imidazole. Fractions with HAT activity where pooled and concentrated to a volume of 1–2 ml. Concentrated sample was then applied to a G-75 size exclusion column and eluted in a buffer of 30 mm HEPES (pH 8), 150 mm NaCl, 1 mmdithiothreitol, and 10% glycerol. E173Q was generated by site-directed mutagenesis and was purified using SP-Sepharose (Amersham Pharmacia Biotech), CoA-agarose (Sigma) and Superdex 75 (Amersham Pharmacia Biotech) gel filtration chromatography. Pure fractions (assessed by SDS-polyacrylamide gel electrophoresis) were pooled and stored at either 4 °C or −20 °C until use. Protein concentration was determined by the method of Bradford (28Bradford M.M. Anal. Biochem. 1976; 72: 248-254Crossref PubMed Scopus (216428) Google Scholar). Extensive development of the HAT assays were required for the detailed kinetic analysis. Assays were performed in 100 mm acetate, 50 mm Bis-Tris, and 50 mm Tris, a three-component buffer that maintains constant ionic strength at a function of pH. Briefly, the assay involves the determination of 3Hacetate transfer from 3Hacetyl-CoA to the histone substrate, catalyzed by GCN5. Labeled 3Hacetyl-CoA (NEN Life Science Products) was mixed with cold acetyl-CoA to generate a stock solution with 40–200 cpm/pmol specific activity. Final acetyl-CoA concentrations of 2–30 μm were employed depending on the assay. Histones were resuspended separately at ∼10 mg/ml in the assay buffer. The reactions (55 μl) were initiated by either the addition of 3Hacetyl-CoA or enzyme (final concentration, 20–100 nm). The reaction was allowed to proceed for 5–25 min at 22 °C before quenching the reaction by spotting the solution onto Whatman P81 (phosphocellulose) filter paper discs. The absorbed histones were then washed three times in 600 ml of 50 mmNaHCO3 buffer (pH 9). Discs were dried briefly in acetone, and the amount of tritium retained was determined by liquid scintillation counting. Initial velocities were determined by varying both substrate (histones and acetyl-CoA) concentrations. Usually, histone concentrations were varied (0.5–50 μm) at different fixed levels of acetyl-CoA (1–30 μm). Because the nonenzymatic histone acetylation reaction resulted in significant background counts, we developed a series of controls to take this into account. For every GNC5 enzymatic reaction, a corresponding nonenzymatic control reaction was performed. The nonenzymatic rate was then subtracted from each sample, prior to fitting the data. When one substrate was varied at saturating levels of the second, the data were fitted to the following equation using KinetAsyst (IntelliKinetics, State College, PA.).v=(kcat·S)/(Km+S)Equation 1 When both substrates were varied, the data were fitted to the following equation.v=kcat·A·B/(Kia·Kb+Ka·B+Kb·A+A·B)Equation 2 Alternatively, the data were plotted in double reciprocal form (as in Fig. 1), and the kinetic parameters were determined from intercept and slope replots (17Segel I.H. Enzyme Kinetics. John Wiley 28: 5735-5742Crossref PubMed Scopus (84) Google Scholar). In a subsequent step, the enzyme then transfers this group to the acceptor substrate. The alternative mechanism involves direct acetyl transfer from CoA to the substrate acceptor, without the formation of a covalent enzyme intermediate (19Lewendon A. Murray I.A. Shaw W.V. Gibbs M.R. Leslie A.G. Biochemistry. 1994; 33: 1944-1950Crossref PubMed Scopus (62) Google Scholar). The latter mechanism requires that both substrates and enzyme must form a ternary complex before catalysis can occur. For yeast GCN5 HAT, these two mechanisms were distinguished from a steady-state kinetic analysis. Both substrate (histones and acetyl-CoA) concentrations were varied, and the resulting initial velocities were determined. Fig.1 displays the data in double reciprocal form where 1/velocity is plotted against 1/ histone at various fixed concentrations of acetyl-CoA. The observed intersecting line pattern is diagnostic for the mechanism in which both substrates must bind to the enzyme before catalysis (Fig. 1). These data argue against a mechanism involving initial acetylation of an enzyme residue, followed by acetyl transfer to lysine. Instead, the data suggest that the ε-amino group of lysine directly attacks the acetyl-CoA within the active site of the enzyme. Supporting evidence comes from the recent x-ray structure of the related HAT1/acetyl-CoA complex, which revealed that the acetyl group of acetyl-CoA was not transferred to the enzyme in the complex (14Dutnall R.N. Tafrov S.T. Sternglanz R. Ramakrishnan V. Cell. 1998; 94: 427-438Abstract Full Text Full Text PDF PubMed Scopus (196) Google Scholar). To date, there is no evidence for a covalent acetyl-enzyme intermediate. Consistent with our data (Fig. 1), similar intersecting line patterns were reported for the native rat liver nuclear HAT (20Wiktorowicz J.E. Campos K.L. Bonner J. Biochemistry. 1981; 20: 1464-1467Crossref PubMed Scopus (14) Google Scholar). Thus the GCN5 HAT family utilize a catalytic mechanism that this direct on acetyl-CoA by lysine. The HATs must the that at pH the ε-amino of the lysine substrate is and The ε-amino group of lysine must first be by an active site amino acid acting as a general base To critical active site we a of amino acid modifying to inactivate GCN5. modifying agents and iodoacetate, which are specific for and to inactivate GCN5 HAT activity. Although group has been reported with native HATs J.E. Allis C.D. Sci. S. A. PubMed Scopus Google Scholar) and rat liver J.E. Bonner J. J. Biol. Full Text PDF PubMed Google Scholar), we no of the HAT catalytic domain is It is that these in from of residues the catalytic In the reagent was at GCN5 catalytic activity in a (Fig. the importance of one or more a is for deprotonating the ε-amino group of a histone then the enzyme be active this is in the active ternary complex. The pH of the reaction rate the presence of an ionization that must be for To the of general base catalysis by an residue, we generated pH rate of the steady-state kinetic k cat andk cat/K The second rate constant k cat/K the reaction between enzyme and substrate. both substrate binding and thek cat/K pH the intrinsic a values of critical The k cat parameter the catalytic subsequent to complex formation and a values of the complex. Because of nonenzymatic histone acetylation at pH, the background reaction rate at each substrate concentration was determined and subtracted from each data The pH of cat/K andk cat revealed a critical ionization that must be for catalysis (Fig. the pH there is a corresponding in catalytic activity that above the critical ionization with a of cat and cat/K the the rate of catalysis for every pH toward a of These data the catalytic of an active site and are with a mechanism in which this may a proton from the ε-amino group of substrate. The catalytic domain of GCN5 has been to region between amino acids and (1Brownell J.E. Zhou J. Ranalli T. Kobayashi R. Edmondson D.G. Roth S.Y. Allis C.D. Cell. 1996; 84: 843-851Abstract Full Text Full Text PDF PubMed Scopus (1286) Google Scholar, R. Zhou Allis C.D. Berger S.L. J. 1997; PubMed Scopus Google Scholar). Within this of amino there are three invariant and indicated the of at but a of an M.H. Zhou J. Allis C.D. Dev. 1998; PubMed Scopus Google Scholar). Although at resulted in this was not among the in the study M.H. Zhou J. Allis C.D. Dev. 1998; PubMed Scopus Google Scholar, L. L. Berger S.L. Dev. 1998; PubMed Scopus Google Scholar). involving residues was one of the in vivo and in an role for L. L. Berger S.L. Dev. 1998; PubMed Scopus Google Scholar). To the role of a GCN5 protein a to was generated and as the catalytic base in the The steady-state kinetic parameters were determined and with the enzyme thek cat for the E173Q was than that determined for the enzyme at pH values for histones and acetyl-CoA were not different between μm) and protein μm). These data suggest that the E173Q does not the of protein to bind substrates but rather has a on in binding affinity suggests that the structural of the protein is for GCN5 and E173Q histone pH pH pH pH are the of the fitted parameters determined from least fitting from one data The were 100 mm acetate, 50 mm Bis-Tris, 50 mm Tris, and 22 °C. in a are the of the fitted parameters determined from least fitting from one data The were 100 mm acetate, 50 mm Bis-Tris, 50 mm Tris, and 22 °C. is the general base for deprotonating the lysine (Fig. one that at pH values where the concentration of lysine is the E173Q activity toward activity. Because the reaction requires an lysine to as the nucleophile within the active this can either be by active deprotonation by at pH base or by the pH to the histone lysine at pH the E173Q cat is than as with at pH 7.5 To this a pH rate of the E173Q was performed (Fig. The E173Q does not the ionization with a of 8 found with the enzyme but rather an ionization with a a of in k cat/K for a group that must be for activity. a group is with the ionization of the ε-amino group of histone lysines J. of and Scholar). To that the ionization in the E173Q pH is the ε-amino group from histone lysine the pH-dependent rate for the nonenzymatic reaction was determined (Fig. The nonenzymatic reaction be the ionization of histone lysines and an a of lysine The second rate constant for the reaction of acetyl-CoA and histones in the of GCN5 was as a function of pH and is in Fig. to the E173Q catalyzed reaction, the reaction rate for every pH a a of the rate of nonenzymatic histone acetylation the solution ionization of lysine the E173Q enzyme the solution ionization of histone lysine At pH where the of histone lysine residues are the rate of catalysis by E173Q values observed with enzyme. the for lysine deprotonation by is at pH, and the E173Q can function as as GCN5 (Fig. However, at pH the E173Q is as a with GCN5. In the we this to be as a transcriptional The recent x-ray structure of revealed that a glutamic acid is between two R. J. M. A. A. F. E. J. Biol. 1998; PubMed Scopus Google Scholar). within the the have this in acetyltransferase a as a general base to a proton from the group of (19Lewendon A. Murray I.A. Shaw W.V. Gibbs M.R. Leslie A.G. Biochemistry. 1994; 33: 1944-1950Crossref PubMed Scopus (62) Google Scholar). acid at this can as the general with catalytic activity (19Lewendon A. Murray I.A. Shaw W.V. Gibbs M.R. Leslie A.G. Biochemistry. 1994; 33: 1944-1950Crossref PubMed Scopus (62) Google Scholar). The of the glutamic acid within the active site resulted in a in the a of the general from a a of for to a for glutamic to our of for of GCN5, be to within a active site of the molecular of reactions to a of specific and the of specific HAT the for the role of protein/histone acetylation in the control of gene
Tanner et al. (Tue,) studied this question.