A Native Peptide Ligation Strategy for Deciphering Nucleosomal Histone Modifications
2003; Elsevier BV; Volume: 278; Issue: 18 Linguagem: Inglês
10.1074/jbc.m301445200
ISSN1083-351X
AutoresMichael A. Shogren‐Knaak, Christopher J. Fry, Craig L. Peterson,
Tópico(s)Advanced biosensing and bioanalysis techniques
ResumoPost-translational modifications of histones influence both chromatin structure and the binding and function of chromatin-associated proteins. A major limitation to understanding these effects has been the inability to construct nucleosomes in vitro that harbor homogeneous and site-specific histone modifications. Here, we describe a native peptide ligation strategy for generating nucleosomal arrays that can harbor a wide range of desired histone modifications. As a first test of this method, we engineered model nucleosomal arrays in which each histone H3 contains a phosphorylated serine at position 10 and performed kinetic analyses of Gcn5-dependent histone acetyltransferase activities. Recombinant Gcn5 shows increased histone acetyltransferase activity on nucleosomal arrays harboring phosphorylated H3 serine 10 and is consistent with peptide studies. However, in contrast to analyses using peptide substrates, we find that the histone acetyltransferase activity of the Gcn5-containing SAGA complex is not stimulated by H3 phosphorylation in the context of nucleosomal arrays. This difference between peptide and array substrates suggests that the ability to generate specifically modified nucleosomal arrays should provide a powerful tool for understanding the effects of post-translational histone modifications. Post-translational modifications of histones influence both chromatin structure and the binding and function of chromatin-associated proteins. A major limitation to understanding these effects has been the inability to construct nucleosomes in vitro that harbor homogeneous and site-specific histone modifications. Here, we describe a native peptide ligation strategy for generating nucleosomal arrays that can harbor a wide range of desired histone modifications. As a first test of this method, we engineered model nucleosomal arrays in which each histone H3 contains a phosphorylated serine at position 10 and performed kinetic analyses of Gcn5-dependent histone acetyltransferase activities. Recombinant Gcn5 shows increased histone acetyltransferase activity on nucleosomal arrays harboring phosphorylated H3 serine 10 and is consistent with peptide studies. However, in contrast to analyses using peptide substrates, we find that the histone acetyltransferase activity of the Gcn5-containing SAGA complex is not stimulated by H3 phosphorylation in the context of nucleosomal arrays. This difference between peptide and array substrates suggests that the ability to generate specifically modified nucleosomal arrays should provide a powerful tool for understanding the effects of post-translational histone modifications. N-(9-fluorenyl)methoxycarbonyl high pressure liquid chromatography dithiothreitol matrix-assisted laser desorption ionization-time of flight Histone proteins, the core protein components of nucleosomes, play a central role in controlling a wide range of DNA-dependent processes, including gene expression, DNA repair, and DNA replication (1Wu J. Grunstein M. Trends Biochem. Sci. 2000; 25: 619-623Abstract Full Text Full Text PDF PubMed Scopus (312) Google Scholar). How histones modulate these processes appears to be dictated in part by their wide range of post-translational modifications, including acetylation, phosphorylation, methylation, ribosylation, and ubiquitination (1Wu J. Grunstein M. Trends Biochem. Sci. 2000; 25: 619-623Abstract Full Text Full Text PDF PubMed Scopus (312) Google Scholar, 2Strahl B.D. Allis C.D. Nature. 2000; 403: 41-45Crossref PubMed Scopus (6507) Google Scholar). Many of these modifications are found to occur simultaneously, suggesting the potential for sequential or combinatorial interactions that may act synergistically or antagonistically to regulate downstream biological effects (2Strahl B.D. Allis C.D. Nature. 2000; 403: 41-45Crossref PubMed Scopus (6507) Google Scholar). For instance, previous studies have used small peptide substrates to demonstrate that phosphorylation of serine 10 within histone H3 enhances the ability of yeast Gcn5p to acetylate lysine residues within the same histone N-terminal domain (3Lo W.S. Trievel R.C. Rojas J.R. Duggan L. Hsu J.Y. Allis C.D. Marmorstein R. Berger S.L. Mol. Cell. 2000; 5: 917-926Abstract Full Text Full Text PDF PubMed Scopus (461) Google Scholar, 4Cheung P. Tanner K.G. Cheung W.L. Sassone-Corsi P. Denu J.M. Allis C.D. Mol. Cell. 2000; 5: 905-915Abstract Full Text Full Text PDF PubMed Scopus (669) Google Scholar). This dual histone mark appears to be key for transcriptional controlin vivo (3Lo W.S. Trievel R.C. Rojas J.R. Duggan L. Hsu J.Y. Allis C.D. Marmorstein R. Berger S.L. Mol. Cell. 2000; 5: 917-926Abstract Full Text Full Text PDF PubMed Scopus (461) Google Scholar, 4Cheung P. Tanner K.G. Cheung W.L. Sassone-Corsi P. Denu J.M. Allis C.D. Mol. Cell. 2000; 5: 905-915Abstract Full Text Full Text PDF PubMed Scopus (669) Google Scholar).A major technical limitation for testing various mechanistic aspects of individual or combinations of histone modifications is the inability to generate nucleosomes in vitro that harbor specific histone modifications. Small peptides encompassing a histone N-terminal tail that contain very specific and homogeneous modifications can be synthesized, and such substrates have played a key role in understanding the potential roles of histone modifications (3Lo W.S. Trievel R.C. Rojas J.R. Duggan L. Hsu J.Y. Allis C.D. Marmorstein R. Berger S.L. Mol. Cell. 2000; 5: 917-926Abstract Full Text Full Text PDF PubMed Scopus (461) Google Scholar, 4Cheung P. Tanner K.G. Cheung W.L. Sassone-Corsi P. Denu J.M. Allis C.D. Mol. Cell. 2000; 5: 905-915Abstract Full Text Full Text PDF PubMed Scopus (669) Google Scholar, 5Zegerman P. Canas B. Papin D. Kouzarides T. J. Biol. Chem. 2002; 277: 11621-11624Abstract Full Text Full Text PDF PubMed Scopus (210) Google Scholar). However, peptides lack the more complex determinants of nucleosomes, such as DNA and multiple histone tails, which are likely to play an important role in their biological functions. Similarly, nucleosomes purified from extracts or treated enzymatically can provide more physiologically relevant substrates, but such nucleosomes lack homogeneous patterns of modifications.Here we describe a native chemical ligation strategy that permits the reconstitution of nucleosomal arrays that harbor a wide range of individual or combinations of histone modifications, including serine phosphorylation, lysine acetylation, and lysine methylation. In this method, solid phase peptide synthesis is used to generate a histone N-terminal tail domain that contains modified amino acids at any desired location. Native ligation chemistry is used to assemble full-length recombinant histone, which is reconstituted into histone octamers and subsequently assembled into nucleosomal arrays. We describe the first test of this method in which we have engineered model nucleosomal arrays harboring a phosphorylated serine at position 10 of each histone H3 protein.RESULTS AND DISCUSSIONGenerating full-length histone proteins by native ligation chemistry requires two components: (i) an N-terminal protein fragment that terminates with a C-terminal thioester, and (ii) a C-terminal protein fragment that begins with an N-terminal cysteine residue (Fig.1) (13Blaschke U.K. Silberstein J. Muir T.W. Methods Enzymol. 2000; 328: 478-496Crossref PubMed Google Scholar). Mixing these two components initiates a two-step chemical reaction producing a full-length ligated product in which the N-terminal fragment is linked to the C-terminal fragment via a canonical peptide bond (13Blaschke U.K. Silberstein J. Muir T.W. Methods Enzymol. 2000; 328: 478-496Crossref PubMed Google Scholar).Synthesis of N-terminal Histone H3 Peptide FragmentsSince most histone H3 post-translational modifications occur within the first 28 amino acids (2Strahl B.D. Allis C.D. Nature. 2000; 403: 41-45Crossref PubMed Scopus (6507) Google Scholar), standard Fmoc-based, solid-phase peptide synthesis can be used to create histone N-terminal domains (1–31) that contain a wide range of modifications in this region. For these experiments we chose to study serine or phosphoserine incorporated at position 10 of H3. To install the desired C-terminal thioester residue, a strategy of selectively exposing and modifying the peptide C terminus was utilized (14Futaki S. Sogawa K. Maruyama J. Asahara T. Niwa M. Hojo H. Tetrahedron Lett. 1997; 38: 6237-6240Crossref Scopus (112) Google Scholar) (Fig. 1). Initial solid-phase peptide synthesis on an acid-hypersensitive resin allows cleavage of the peptides from the resin with weak acid, selectively unmasking the C terminus, which can then be selectively derivatized to generate a thioester. Side chain deprotection with strong acid followed by HPLC was sufficient to generate the desired unphosphorylated H3 peptide fragment in pure form (Fig. 2 A). Because some dephosphorylation occurred during the synthetic process, an additional metal affinity purification step was used to obtain homogeneous, phosphorylated H3 peptide (7Muszynska G. Dobrowolska G. Medin A. Ekman P. Porath J.O. J. Chromatog. 1992; 604: 19-28Crossref PubMed Scopus (100) Google Scholar). The identity of the peptides was confirmed by electrospray mass spectrometry, and their purity demonstrated by analytical HPLC (data not shown).Figure 2Phosphorylated and unphosphorylated histone H3 produced by native chemical ligation is competent for incorporation into histone octamers. A, SDS-PAGE analysis of histone H3 ligation. A ligation reaction (lane 4) containing synthetic unphosphorylated N-terminal histone H3 peptide (lane 2) and cysteine-containing C-terminal histone H3 protein (lane 3) analyzed on an 18% SDS-PAGE gel stained with Coomassie Blue. Lane 1 contains wild type recombinant histone H3. The highest molecular weight band shows the expected mass when analyzed by MALDI-TOF mass spectral analysis (adjacent graph). Right panel shows an SDS-PAGE of recombinantXenopus H3 (lane 5), the purified, ligated, unphosphorylated H3 (T32C-H3; lane 6) and phosphorylated H3 (S10Phos-H3; lane 7) proteins. B, SDS-PAGE analysis of histone octamers containing ligated histone H3. Column fractions from the gel filtration purification of recombinant histone octamers assembled with either T32C-H3 or S10Phos-H3 analyzed on an 18% SDS-PAGE stained with Coomassie Blue. Note that octamers assembled from non-ligated histones elute in the same fractions and thatXenopus histones H2A and H2B are not resolved under standard SDS-PAGE conditions. C, mass spectral analysis of histone octamers containing S10Phos-H3. The masses of the constituent components of histone octamers assembled with recombinantXenopus H2A, H2B, H4, and S10Phos-H3 were determined by MALDI-TOF mass spectral analysis and correspond to the expected mass of each histone subunit. The largest molecular weight histone, S10Phos-H3, is enlarged in the left inset to demonstrate the homogeneity of phosphorylation. For comparison, the right inset shows a mixture of roughly equal amounts of S10Phos-H3 and T32C-H3 histone.View Large Image Figure ViewerDownload Hi-res image Download (PPT)C-terminal H3 Protein FragmentSince the C-terminal fragment of histone H3 must contain an N-terminal cysteine residue, we used site-directed mutagenesis of a Xenopus histone H3 bacterial expression construct to replace the threonine codon at position 32 with a cysteine codon. Importantly, this T32C substitution is not expected to disrupt nucleosome structure since Thr-32 is located outside the structured domain of the histone octamer (15Luger K. Mader A.W. Richmond R.K. Sargent D.F. Richmond T.J. Nature. 1997; 389: 251-260Crossref PubMed Scopus (6795) Google Scholar) and is not well conserved among eukaryotes. Because the N-terminal methionine of recombinantly expressed proteins is not readily removed, a minimal Factor Xa cleavage site was also introduced directly N-terminal to the cysteine at position 32. Consequently, cleavage of purified, recombinant T32C histone H3 with Factor Xa exposes the N-terminal cysteine. Reversed-phase HPLC was used to isolate the desired C-terminal H3 protein fragment (32–135). The identity and purity of this fragment was confirmed by SDS-PAGE (Fig. 2 A) and MALDI-TOF mass spectral analysis (data not shown).H3 Protein LigationFull-length unphosphorylated (T32C-H3) and Ser-10 phosphorylated histone H3 (S10Phos-H3) polypeptides were generated by mixing the N-terminal peptide and C-terminal protein fragments under denaturing conditions in the presence of thiol reagents (9Dawson P.E. Churchill M.J. Ghadiri M.R. Kent S.B.H. J. Am. Chem. Soc. 1997; 119: 4325-4329Crossref Scopus (245) Google Scholar, 13Blaschke U.K. Silberstein J. Muir T.W. Methods Enzymol. 2000; 328: 478-496Crossref PubMed Google Scholar). Note the S10Phos-H3 polypeptide also contains the T32C amino acid change. Successful chemical ligation was monitored by SDS-PAGE, which detected formation of a protein product that co-migrates with recombinant wild-type Xenopus histone H3 (Fig.2 A). Further analysis of the T32C-H3 reaction mixture by MALDI-TOF revealed that the mass of the largest molecular weight species corresponds to the expected mass for full-length T32C histone H3 (Fig. 2 A). Because the ligation reactions were performed under conditions of excess peptide and because the reaction proceeds to roughly 50% completion, ion exchange chromatography was used to purify the ligated products to homogeneity (Fig. 2 A).Octamer and Nucleosome AssemblyTo generate histone octamer containing either the ligated T32C-H3 or S10Phos-H3, the ligated H3 proteins were denatured and mixed with denatured, recombinantXenopus histones H4, H2A, and H2B in equal ratios, and the renatured histone octamers were isolated by gel filtration as described previously (8Luger K. Rechsteiner T.J. Richmond T.J. Methods Mol. Biol. 1999; 119: 1-16PubMed Google Scholar). No significant differences in the efficiency of octamer reconstitution were detected when ligated T32C-H3 or S10Phos-H3 histones were substituted for wild-type H3 (Fig. 2 B). To make certain that S10Phos-H3 was not dephosphorylated during the course of octamer assembly or purification, MALDI-TOF mass spectral analysis was performed on the purified octamer (Fig. 2 C). Each of the four different histone polypeptides was detected, and the largest molecular weight component corresponds to the expected mass of the singly phosphorylated histone H3. Moreover, no dephosphorylated H3 was detectable, confirming that this histone octamer is homogeneous for H3 serine 10 phosphorylation.Chicken erythrocyte or recombinant Xenopus histone octamers were used to assemble model nucleosomal arrays using a DNA template composed of eleven head to tail repeats of a 208-bp 5 S rRNA gene from Lytechinus variegatus (the 208–11 template; see Fig. 3 A; (16Logie C. Peterson C.L. EMBO J. 1997; 16: 6772-6782Crossref PubMed Scopus (166) Google Scholar)). Each 5 S repeat can rotationally and translationally position a nucleosome afterin vitro salt dialysis reconstitution, yielding a positioned array of nucleosomes (17Dong F. Hansen J.C. van Holde K.E. Proc. Natl. Acad. Sci. U. S. A. 1990; 87: 5724-5728Crossref PubMed Scopus (169) Google Scholar). This 208–11 template also contains a uniqueSalI restriction enzyme site in the central repeat of the array (see Fig. 3 A; (16Logie C. Peterson C.L. EMBO J. 1997; 16: 6772-6782Crossref PubMed Scopus (166) Google Scholar)). To monitor nucleosome assembly, we digested the reconstituted nucleosomal arrays with EcoRI. Since each 5 S rDNA repeat in the 208–11 array template is bordered byEcoRI restriction sites (Fig. 3 A), cleavage releases either a 208-bp free DNA fragment or a mononucleosome that can be identified due to its slower mobility after native gel electrophoresis. Comparison of the resolved digestion products (Fig.3 B) indicates that octamers containing either ligated T32C-H3 or S10Phos-H3 can be incorporated into the DNA template with efficiencies similar to that of chicken erythrocyte or wild-type (Thr-32) recombinant octamers (Fig. 3 B and data not shown). Additionally, increasing ratios of ligated T32C-H3-containing octamer to DNA template results in an expected increase in array saturation as demonstrated by the increase of mononucleosome and oligonucleosome products relative to free DNA. Together these results suggest that octamers generated from ligated histones are competent to form nucleosomal arrays.Figure 3Nucleosomal arrays containing serine 10 phosphorylated or unphosphorylated histone H3 are equivalently remodeled by SWI/SNF. A, schematic of the 208–11 nucleosomal array template. The DNA template utilized is composed of eleven 208-bp head-to-tail repeats of a 5 S rDNA gene. Flanking the repeats are EcoRI restrictions sites. The central repeat contains a unique SalI restriction site. B, characterization of nucleosomal arrays containing ligated histones. Nucleosomal arrays were digested with EcoRI, electrophoresed on a native 4% polyacrylamide gel, and then stained with ethidium bromide. T32C-H3 nucleosomal arrays prepared with increasing ratios of histone octamer to template DNA are compared. Nucleosome free DNA is labeled Naked, while mononucleosomes are labeledNuc. C, SWI/SNF-mediated remodeling of nucleosomal arrays containing ligated histone H3. ATP-dependent remodeling assays were performed with purified yeast SWI/SNF complex (2–6 nm), SalI (2.5 units/μl), and radiolabeled nucleosomal arrays (1 nm) containing recombinant Xenopus H3 (filled circles), T32C-H3 (open circles), or S10Phos-H3 (filled squares). Assays were initiated with ATP (solid lines) and compared with control reactions lacking ATP (dashed lines). DNA was extracted from samples taken over the course of the remodeling assay and subjected to gel electrophoresis. The fraction of uncut array (top graph) was determined by quantification of cut and uncut DNA template using a PhosphorImager and ImageQuant version 1.2 (Molecular Dynamics). The amount of array remodeled (bottom graph) was calculated from the difference between the amount of cut array in the presence and absence of ATP. Assays were performed three independent times with a representative experiment shown.View Large Image Figure ViewerDownload Hi-res image Download (PPT)SWI/SNF-dependent Remodeling of S10Phos-H3 Nucleosomal ArraysRestriction enzyme accessibility assays were performed to further characterize the nucleosomal arrays containing ligated histones and to investigate whether ATP-dependent nucleosome remodeling by yeast SWI/SNF might be influenced by H3-serine 10 phosphorylation. Radiolabeled nucleosomal arrays containing chicken octamer, recombinant wild-type Xenopus H3, and ligated T32C-H3 and S10Phos-H3 were prepared and characterized byEcoRI digestion (data not shown). Arrays with a similar degree of nucleosome saturation were then subjected to digestion withSalI, which yields a quantitative measurement of the accessibility of DNA within the central nucleosome (Fig. 3 A) (16Logie C. Peterson C.L. EMBO J. 1997; 16: 6772-6782Crossref PubMed Scopus (166) Google Scholar). In the absence of ATP, nucleosomal arrays reconstituted with ligated histones exhibited nearly identical occlusion of theSalI site compared with arrays reconstituted with wild-type recombinant or chicken octamers (Fig. 3 and data not shown). Thus, the DNA wrapped onto ligated histone octamers does not appear to be inherently more accessible to restriction enzymes. Furthermore, when ATP was added to initiate SWI/SNF-dependent remodeling, we found that all of the arrays were excellent substrates for SWI/SNF action, showing similar kinetics over different enzyme and array concentrations (Fig. 3 and data not shown). For each array, SWI/SNF action enhanced SalI digestion kinetics by over 30-fold. Since this remodeling assay appears to measure the rate of nucleosome movements by ATP-dependent remodeling enzymes (18Jaskelioff M. Gavin I. Peterson C.L. Logie C. Mol. Cell. Biol. 2000; 20: 3058-3068Crossref PubMed Scopus (93) Google Scholar), these data suggest that H3 serine 10 phosphorylation does not grossly alter the ability of SWI/SNF to mobilize nucleosomes.Histone Acetyltransferase AssaysPrevious studies with peptide substrates demonstrated that H3 serine 10 phosphorylation enhances the histone acetyltransferase activities of recombinant Gcn5p and native Gcn5-containing HAT complexes, such as SAGA (3Lo W.S. Trievel R.C. Rojas J.R. Duggan L. Hsu J.Y. Allis C.D. Marmorstein R. Berger S.L. Mol. Cell. 2000; 5: 917-926Abstract Full Text Full Text PDF PubMed Scopus (461) Google Scholar, 4Cheung P. Tanner K.G. Cheung W.L. Sassone-Corsi P. Denu J.M. Allis C.D. Mol. Cell. 2000; 5: 905-915Abstract Full Text Full Text PDF PubMed Scopus (669) Google Scholar). To investigate whether Ser-10 phosphorylation within the context of nucleosomal arrays also stimulates the HAT activity of Gcn5p, we first performed radioactive HAT assays using recombinant Gcn5p and nucleosomal arrays containing recombinant wild-type XenopusH3, T32C-H3, or S10Phos-H3. Importantly, our HAT assays contained 50 mm monovalent cation, which facilitates the nucleosomal HAT activity of rGcn5p (19Tse C. Georgieva E.I. Ruiz-Garcia A.B. Sendra R. Hansen J.C. J. Biol. Chem. 1998; 273: 32388-32392Abstract Full Text Full Text PDF PubMed Scopus (45) Google Scholar). The extent of 3H-acetate incorporation into the nucleosomal arrays was plotted as a function of time, and time points in the linear portion of the reaction were fit to determine initial velocities (Fig.4 A). Initial velocities for unphosphorylated nucleosomal array substrates containing recombinant wild-type Xenopus H3 and ligated T32C-H3 were nearly equivalent. In contrast, the initial velocity for the phosphorylated nucleosomal array substrate was increased greater than 2-fold relative to the unphosphorylated array, consistent with previous studies demonstrating enhanced rGcn5 HAT activity with phosphorylated peptide substrates.Figure 4Histone H3 serine 10 phosphorylation stimulates acetylation of nucleosomal arrays by Gcn5p but not by the Gcn5p-containing SAGA complex. A, comparison of Gcn5p HAT activity on phosphorylated and unphosphorylated nucleosomal arrays. Histone acetyltransferase assays were performed with recombinant Gcn5p (2.45 μm) and nucleosomal arrays (100 nmhistone H3) containing recombinant Xenopus H3 (filled circles), T32C-H3 (open circles), or S10Phos-H3 (filled squares). The extent of histone acetylation was determined at multiple timepoints by counting the incorporation of the tritiated acetate in a scintillation counter. Each assay was repeated at least three times, and initial velocities were calculated by plotting the average amount of radiolabel incorporation as a function of time. Error bars depict standard deviation for each time point. B, comparison of SAGA HAT activity on phosphorylated and unphosphorylated nucleosomal arrays. Assays and data analysis of SAGA were performed identically to those with Gcn5p above using SAGA (4.9 nm) and nucleosomal array (75 nm histone H3). C, comparison of SAGA HAT activity on phosphorylated and unphosphorylated nucleosomal arrays over a range of nucleosomal array concentration (12.5–625 nm histone H3 for arrays containing recombinant wild-type Xenopus H3; 12.5–250 nm histone H3 for arrays containing either T32C-H3 or S10Phos-H3). Assays were performed as described above except 9.8 nm SAGA was used with array concentrations containing 12.5–50 nm histone H3, while 4.9 nm SAGA was used with array concentrations containing 75–625 nmhistone H3. Initial velocities for each substrate concentration were calculated as described for Gcn5p. At least three independent experiments were performed at each array concentration, and the average initial velocity, normalized for SAGA concentration, was plotted as a function of histone H3 concentration. Initial velocities for nucleosomal arrays containing wild-type recombinant Xenopushistone H3 were fit to a saturation curve with a calculated Hill coefficient of 1.96 and depicted with a dashed line. Error bars show standard deviations for nucleosomal arrays containing wild-type recombinant Xenopus histone H3.D, comparison of SAGA histone specificity for phosphorylated and unphosphorylated nucleosomal arrays. HAT assays were performed with 14.8 nm SAGA and nucleosomal arrays (450 nmhistone H3) containing wild-type Xenopus H3 (lane 1), T32C-H3 (lane 2), or S10Phos-H3 (lane 3). Reactions were incubated for 30 min and then subjected to 15% SDS-PAGE gel. Gels were stained with Coomassie Blue to visualize histones, destained, soaked in ENHANCE (NEN), and then analyzed by radiography.View Large Image Figure ViewerDownload Hi-res image Download (PPT)To determine whether this increased activity toward phosphorylated nucleosomal arrays extended to a native Gcn5p-containing protein complex, we performed nucleosomal HAT assays using the multi-subunit SAGA complex (Fig. 4 B). Contrary to the increased HAT activity of Gcn5p on the phosphorylated array, the initial velocities for HAT activity of SAGA complex did not show a difference between the phosphorylated and unphosphorylated array substrates. Because this result was contrary to previous results that used SAGA and phosphorylated peptides (3Lo W.S. Trievel R.C. Rojas J.R. Duggan L. Hsu J.Y. Allis C.D. Marmorstein R. Berger S.L. Mol. Cell. 2000; 5: 917-926Abstract Full Text Full Text PDF PubMed Scopus (461) Google Scholar), a more extensive kinetic characterization was performed. To ensure that the HAT assays were not performed under saturating substrate concentrations, where differences in binding affinity would be largely masked, assays were repeated over a wide range of nucleosomal array concentrations for all three substrates. Plots of initial velocities of HAT activity as a function of array concentration (Fig. 4 C) show that SAGA exhibits saturation kinetics for all three nucleosomal array substrates. Strikingly, comparison of initial velocities of SAGA for phosphorylated and unphosphorylated arrays over a range of substrate concentrations confirms an absence of any significant difference in initial velocities. Furthermore, SDS-PAGE analysis of these HAT reactions demonstrates that the H3 subunit remains the preferred acetylation substrate (Fig. 4 D) and that no detectable difference in the level of acetylation is detectable among the three substrate arrays.The ability of Gcn5p and SAGA to efficiently utilize nucleosomes composed of either wild-type or ligated histones provides further support for the integrity of nucleosomal arrays reconstituted with histones generated by native chemical ligation. Moreover, the ability to generate homogeneous, phosphorylated nucleosomal arrays has allowed us to determine that the nucleosomal HAT activity of recombinant Gcn5p is enhanced by H3 phosphorylation and that the nucleosomal HAT activity of the native SAGA complex is insensitive to H3 phosphorylation. This latter result was particularly surprising since previous studies that used peptide substrates demonstrated stimulation of SAGA activity by H3 phosphorylation (3Lo W.S. Trievel R.C. Rojas J.R. Duggan L. Hsu J.Y. Allis C.D. Marmorstein R. Berger S.L. Mol. Cell. 2000; 5: 917-926Abstract Full Text Full Text PDF PubMed Scopus (461) Google Scholar). In the case of rGcn5p it is known that the phosphorylation-dependent stimulation of HAT activity is due to an increased affinity for the peptide substrate (6- to 10-fold; (3Lo W.S. Trievel R.C. Rojas J.R. Duggan L. Hsu J.Y. Allis C.D. Marmorstein R. Berger S.L. Mol. Cell. 2000; 5: 917-926Abstract Full Text Full Text PDF PubMed Scopus (461) Google Scholar, 4Cheung P. Tanner K.G. Cheung W.L. Sassone-Corsi P. Denu J.M. Allis C.D. Mol. Cell. 2000; 5: 905-915Abstract Full Text Full Text PDF PubMed Scopus (669) Google Scholar)). In the case of SAGA complex, the failure of H3 phosphorylation to stimulate nucleosomal HAT activity may be due to other subunits (e.g. Ada2p) contributing to a different mode of histone tail binding. In addition, SAGA may make additional contacts with nucleosomes that are simply not possible with small peptide substrates; for instance SAGA may interact with multiple histone tails, or the ability of SAGA to bind with high affinity to DNA (20Sendra R. Tse C. Hansen J.C. J. Biol. Chem. 2000; 275: 24928-24934Abstract Full Text Full Text PDF PubMed Scopus (38) Google Scholar) may also contribute to a different mode of nucleosomal substrate binding. We note that our biochemical studies are consistent with several recent reports that suggest that histone H3 phosphorylation is not always coupled to enhanced histone acetylation in vivo (21Thomson S. Clayton A.L. Mahadevan L.C. Mol. Cell. 2001; 8: 1231-1241Abstract Full Text Full Text PDF PubMed Scopus (172) Google Scholar,22Labrador M. Corces V.G. Genes Dev. 2003; 17: 43-48Crossref PubMed Scopus (46) Google Scholar).Concluding RemarksAlthough our studies have used native peptide ligation to generate homogeneous, phosphorylated nucleosomal arrays, this strategy has the potential for generating and analyzing nucleosomal arrays that harbor a wide range of histone post-translational modification. Importantly, the ligation chemistry is compatible with peptides that harbor specific acetylated or methylated lysine residues. Furthermore, nucleosomal arrays could be generated
Referência(s)