|Home | About | Journals | Submit | Contact Us | Français|
This is an Open Access article distributed under the terms of the Creative Commons Attribution Non-Commercial License (http://creativecommons.org/licenses/by-nc/2.0/uk/) which permits unrestricted non-commerical use, distribution, and reproduction in any medium, provided the original work is properly cited.
The tails of core histones (H2A, H2B, H3 and H4) are critical for the regulation of chromatin dynamics. Each core histone tail is specifically recognized by various tail binding proteins. Here we screened for budding yeast histone H4-tail binding proteins in a protein differential display approach by two-dimensional gel electrophoresis (2DGE). To obtain highly enriched chromatin proteins, we used a Mg2+-dependent chromatin oligomerization technique. The Mg2+-dependent oligomerized chromatin from H4-tail deleted cells was compared with that from wild-type cells. We used mass spectrometry to identify 22 candidate proteins whose amounts were reduced in the oligomerized chromatin from the H4-tail deleted cells. A Saccharomyces Genome Database search revealed 10 protein complexes, each of which contained more than two candidate proteins. Interestingly, 7 out of the 10 complexes have the potential to associate with the H4-tail. We obtained in vivo evidence, by a chromatin immunoprecipitation assay, that one of the candidate proteins, Pwp1p, associates with the 25S ribosomal DNA (rDNA) chromatin in an H4-tail-dependent manner. We propose that the complex containing Pwp1p regulates the transcription of rDNA. Our results demonstrate that the protein differential display approach by 2DGE, using a histone-tail mutant, is a powerful method to identify histone-tail binding proteins.
A nucleosome comprises ~200 bp of DNA, an octameric core of two copies each of histones H2A, H2B, H3 and H4, and the linker histone H1. The core histone proteins contain histone-fold domains and histone-fold extensions responsible for histone–histone and histone–DNA interactions, and N-terminal tail domains. The crystal structure of the nucleosome core particle revealed that the N-terminal tails are predominantly external to the particle and highly flexible, suggesting that they interact with DNA/histones, and also serve as targets for chromatin binding proteins and enzymatic functions (1). Consistent with the idea that the histone-tails mediate the key nucleosome–nucleosome interactions that are essential for chromatin structures in vivo, the histone-tails were shown to be critical for both the intramolecular folding of nucleosomal arrays and the fiber–fiber interactions observed in vitro (2). The locations and interactions of the H3 tail domain are dependent upon the degree of the condensation of a nucleosomal array, and alterations in the tail-interactions may elaborate different structural and functional states of chromatin (3). Thus, the core histone tail domains are essential determinants of chromatin fiber dynamics. In the chromatin context in vivo, the tails electrostatically interact with DNA. In addition, histone tails also interact with other tails, other histone domains and non-histone proteins.
The histone-tail-binding proteins associate with specifically modified tails and induce various chromatin functions. In yeast, repressor proteins, such as Sir3p, and transcriptional activators, such as the bromodomain-containing protein Bdf1p, interact with the histone H4 tail (4–7). The N-terminal tail of histone H4 has four acetylated lysines: K5, K8, K12 and K16. Of these, only K16 is strongly correlated with a specific regulatory function in yeast, where its acetylation state regulates the extent of heterochromatin silencing. K16 is acetylated by Sas2p and deacetylated by Sir2p, and the balance of these opposing activities determines the extent of silencing spreading from telomeres (8,9). The spreading is in part owing to Sir3p's preferential association with the H4-tails that have been deacetylated at K16. In contrast, Bdf1p preferentially associates with acetylated H4-tails. Bdf1p competes with a Sir2p deacetylase for binding to the acetylated H4-tails (7). Besides acetylation, methylated H3-tails are recognized by repressor proteins, such as heterochromatin protein 1 (HP1) and Polycomb, in higher eukaryotes (10–12). The chromodomains of HP1 and Polycomb associate with H3-tails methylated at K9 and K27, respectively, leading to changes in DNA template accessibility and higher-order chromatin structures.
The molecular interaction partners of the histone tails in different states of modification are being elucidated and have been shown to play key roles in the regulation of all nuclear processes, leading to an ‘epigenetic code' or ‘histone code' hypothesis (13). Based on this hypothesis, histone-tail-binding proteins have been isolated by affinity columns using histone tail peptides with specific modifications. This approach identified various factors, such as a nucleosome remodeling and deacetylase (NuRD) complex, hSNF2H, an inhibitor of acetyltransferase (INHAT) complex, a WD40-repeat protein WDR5, and 14-3-3 isoforms of phosphospecific binding proteins as H3-tail-binding proteins (14–18). These identifications indicated that the peptide columns are a powerful approach to screen for the tail-binding proteins. However, the screening is probably biased, since the binding to the peptides is performed in vitro. Thus, many tail-binding proteins may still be unidentified. Therefore, it is worthwhile to perform the screening for the tail-binding proteins by in vivo binding.
Here we screened for potential histone H4-tail associated proteins using a protein differential display approach by two-dimensional gel electrophoresis (2DGE) comparing H4 tail deleted chromatin with wild-type chromatin. We have found that a WD40-repeat protein, Pwp1p, interacts with chromatin through the H4-tail, which confirms that our approach is effective for identifying histone-tail-associated proteins.
The strains used in this study were a wild type strain (WT), ENY012 [isogenic with PKY501 (19) except for SIR3-FLAG], a histone H4 N-terminal tail (amino acids 4–28) deleted strain (H4-tailΔ), ENY017 [isogenic with PKY813 (19) except for SIR3-FLAG], ENY091 (isogenic with PKY501 except for TAF14-FLAG and PWP1-Myc), ENY092 (isogenic with PKY813 except for TAF14-FLAG and PWP1-Myc), ENY098 (isogenic with PKY501 except for PWP1-FLAG), and ENY102 (isogenic with PKY813 except for PWP1-FLAG). The tags were fused to the genes with a PCR-based one-step in vivo tagging strategy, using the FLAG tag vector p3FLAG-KanMX (20) and pFA6a-13Myc-TRP1 (21). Transformations of yeast strains were performed by the lithium chloride method. Transformed cells were plated onto YPD plates containing 200 μg/ml of G418 disulfate (Nacalai Tesque) or synthetic medium plates lacking tryptophan. Integration of the tag-encoding DNA was confirmed by PCR, followed by western blot analyses with anti-FLAG M2 (Sigma) and anti-Myc-tag monoclonal antibodies (Cell Signaling Technology, 9B11, #2276).
Yeast cells were grown in 200 ml of YPD medium at 30°C to an A600 of ~1.0. Cells were spheroplasted according to Liang and Stillman (22), with some modifications. Cells were incubated at room temperature for 5 min in 2 ml of prespheroplasting buffer [100 mM PIPES (pH 9.4) and 10 mM DTT], followed by an incubation in 2 ml of spheroplasting buffer [50 mM KH2PO4/K2HPO4 (pH 7.5), 0.6 M sorbitol and 10 mM DTT] containing either 2 mg/ml (for wild-type strain) or 6 mg/ml (for tail deleted strain) of Yeast Lytic Enzyme (ICN) at room temperature for 10 min with occasional mixing, until the OD600 of a 1:100 dilution of the cell suspension (in water) dropped to <25% of the value before digestion. Sphereoplasts were washed with 2 ml of ice-chilled wash buffer [100 mM KCl, 50 mM HEPES–KOH (pH 7.5), 2.5 mM MgCl2 and 0.4 M sorbitol], and then were pelleted at 3000 r.p.m. for 5 min in a microcentrifuge at 4°C. Crude cell nuclei were prepared from the spheroplasts according to Alfieri and Clark (23), with some modifications. The spheroplasts were lysed by vigorous resuspension with a pipette in 2 ml of spheroplast lysis buffer 18% (wt/vol) Ficoll 400, 0.2% Triton X-100, 40 mM potassium phosphate (pH 7.5), 2 mM Na-EDTA, 0.5 mM Na-EGTA, 0.5 mM spermidine hydrochloride, 0.15 mM spermine hydrochloride, 10 mM 2-ME and Complete protease inhibitor cocktail (Roche)]. The lysate was layered over 4 ml of a sucrose step gradient [65% (wt/vol) sucrose, 40 mM potassium phosphate (pH 7.5), 2 mM Na-EDTA, 0.5 mM Na-EGTA, 0.5 mM spermidine hydrochloride, 0.15 mM spermine hydrochloride, 10 mM 2-ME and Complete protease inhibitor cocktail (Roche)] and spun at 14 000 r.p.m. for 30 min in a microcentrifuge at 4°C. The nuclear pellet was resuspended in 0.8 ml of nuclei lysis buffer [50 mM Tris–HCl (pH 8.0), 5 mM Na-EDTA, 5 mM 2-ME and Complete protease inhibitor cocktail (Roche)]. RNase (32 μl, DNase free, 10 mg/ml) was added, and after incubating for 30 min on ice, the mixture was spun at 10000 r.p.m. for 3 min at 4°C. This step of the nuclear lysis was repeated once. The lysed nuclei were resuspended in 1 ml of MN buffer [10 mM Tris–HCl (pH 8.5) and 2 mM CaCl2], kept on ice for 15 min, and spun at 10000 r.p.m. for 3 min at 4°C. This washing step with the MN buffer was repeated once. The lysed nuclei were resuspended in 1 ml of MN buffer again, and were incubated at room temperature for 5 min. Micrococcal nuclease (MNase) (5 μl, 10000 U/ml, Worthington) was added. The digestion was performed for 1 min at room temperature, stopped by adding 30 μl of 100 mM EGTA and cooled on ice for 15 min. The supernatant was recovered by centrifugation at 10000 r.p.m. for 3 min at 4°C. The lysed nuclei pellet was resuspended in 0.5 ml MN buffer with 3 mM EGTA and kept on ice for 15 min. The supernatant was recovered by centrifugation at 10000 r.p.m. for 3 min at 4°C and was combined with the first supernatant (MNase eluted chromatin). Western blot analyses were performed with anti-H3 histone (Abcam, ab1791), anti-H2A histone (Abcam, ab4075), anti-H4 histone (Cell Signaling Technology, #2592), anti-BDF1 (gift from S. Roeder), anti-PGK1 (Molecular Probes, A-6457), anti-FLAG (Sigma) and anti-Myc antibodies (Cell Signaling Technology, 9B11, #2276) by an ECL Plus detection system (Amersham). The detection and quantitation of the signals were performed by a lumino-imager (LAS-1000plus, Fujifilm). Purified histones were purchased from Sigma (type II-A from calf thymus).
The MNase eluted chromatin was chilled on ice for 15 min, and after centrifugation at 15 000 r.p.m. for 15 min in a microcentrifuge, the supernatant was recovered. A MgCl2 solution was added to 12 mM, and after chilling on ice for 15 min, the mixture was spun at 15 000 r.p.m. for 15 min. The oligomerized chromatin was recovered in the precipitate. In Figure 2A, the MgCl2 solution was added to achieve the desired final salt concentration. The DNA was purified, electrophoresed on a 2% agarose gel and stained with ethidium bromide.
The oligomerized chromatin (30–50 μg) was dissolved in 125 μl of rehydration buffer [7 M urea, 2 M thiourea, 4% CHAPS, 0.1% bromophenol blue, 0.5% IPG buffer (Amersham) and 1.5 μl DeStreak Reagent (Amersham)] and applied to 7 cm isoelectric focusing (IEF) strips (Amersham Immobiline DryStrip pH 3–10 NL, 7cm). Strip rehydration and IEF were performed in an Ettan IPGphor II isoelectric focusing system (Amersham) according to the manufacturer's manual. The focused strips were used immediately for the second dimension. The second dimension was performed on a NuPAGE 4–12% Bis–Tris ZOOM Gel (Invitrogen), according to the manufacturer's recommendations. Gels were stained with a Silver Stain MS kit (Wako Pure Chemical). Selected protein spots were excised and in-gel digested with trypsin. The peptides extracted from the gels were separated on a PepMap C18 (75 μm × 15 cm, LC Packings) column using a capillary HPLC and directly applied to a Micromass hybrid quadrupole time-of-flight (Q-Tof micro) mass spectrometer (Micromass, Manchester, UK). Data derived from the MS/MS spectra were used to search the NCBI non-redundant protein database with the Mascot program (Matrix Science).
Chromatin immunoprecipitation (ChrIP) and PCR were performed as described previously (24,25), with some modifications. Immunopreciptiation of Pwp1p-FLAG or Sir3p-FLAG was done in FA lysis buffer containing 20 mM NaCl2 with 25 μl of anti-FLAG M2-agarose beads (50% slurry, Sigma) for 125 μl of whole cell extract. The immunoprecipitate was washed three times in the same buffer, and the elution was done by the addition of FLAG peptide (Sigma) to 200 μg/ml. PCR was performed with a 1/250 portion of immunoprecipitated DNA within a linear range of amplification. The linear range of amplification was confirmed by serial 2.5-fold dilutions of an input DNA as templates (Figure 6A and 6C). The products were electrophoresed on a 6% polyacrylamide gel and were stained with SYBR Gold (Invitrogen). The signals of the PCR products were quantified with a fluoro-imager (LAS-1000plus, Fujifilm). To calculate the relative fold enrichment of Pwp1p or SIR3p binding, we first normalized each signal to the loading control of CUP1, and then normalized the signal to an input DNA. The primer sequences used in this study are available upon request.
Yeast cells were treated with a lytic enzyme to remove their cell walls and were converted to spheroplasts. The spheroplasts were lysed using Ficoll and Triton X-100, and the cell nuclei were precipitated by centrifugation. The nuclei were lysed by resuspension in a low ionic strength buffer. The genomic chromatin was undamaged and still attached to the nuclear structures. The nuclear pellet still contained some ribosomes, which co-sedimented with the nuclei (23). RNase was used to remove the ribosomes and RNA from the lysed nuclei. The genomic chromatin was then released by digestion with MNase from the lysed nuclei (the MNase-eluted chromatin fraction) (Figure 1A). As shown in Figure 1B, we detected nucleosomal ladders with the MNase digestion, indicating that soluble chromatin was eluted from the purified nuclei by the MNase digestion. Without MNase, no nucleosomal ladder was detected, indicating that no chromatin was eluted from the nuclei. We also confirmed the equal elution of chromatin at different genomic regions by quantitating the amounts of DNA at selected heterochromatin and euchromatin regions (data not shown).
MNase digestion of chromatin may also cause the release of non-chromatin associated proteins from the nuclei. Therefore, proteins that are not associated with chromatin may be included in the MNase-eluted chromatin fraction. These contaminating proteins should be removed for effective screening of the chromatin-associated proteins.
Chromatin condenses and oligomerizes in the presence of Mg2+ ions. This property is called Mg2+-dependent oligomerization (2,26). The addition of increasing concentrations of Mg2+ ions to an unfolded array of nucleosomes in a low salt buffer causes a series of conformational changes and leads to the formation of a highly folded state. In addition to the intra-fiber conformational changes, inter-fiber-interactions also occur, resulting in the oligomerization between the fibers. The oligomerized chromatin is easily precipitated by centrifugation. We investigated whether the Mg2+-dependent oligomerization of the MNase-eluted chromatin separates the chromatin-associated proteins from the non-associated proteins. As shown in Figure 2A, a Mg2+ ion titration was conducted. After the addition of various Mg2+ ion concentrations and centrifugation, the DNA was purified from both the supernatant and precipitate, fractionated on an agarose gel, and visualized by ethidium bromide staining. A larger amount of DNA was precipitated with higher concentrations of the Mg2+ ions. We found that most of the chromatin was oligomerized and fractionated into the precipitate with 12 mM Mg2+ ions. We used this concentration for our preparation of the oligomerized chromatin fraction.
We next analyzed the proteins in the supernatant and the precipitate produced by the fractionation with the 12 mM Mg2+ ions. The proteins in both fractions were detected by silver staining. Most of the DNA was in the precipitate, as described above. In contrast, about a half of the proteins remained in the supernatant (Figure 2B). We speculated that the proteins in the supernatant would not be associated with the chromatin because they did not fractionate into the precipitate, where most of the DNA fractionated. This speculation is supported by the observation of the distinct composition of each fraction. For example, core histones, fundamental chromatin components, were fractionated into the precipitate and became major bands in the precipitate, as identified with histone markers. Lower amounts of histones were present in the supernatant, which was consistent with less DNA in the supernatant. On the other hand, protein bands between 37 and 75 kDa were abundant in the supernatant. This suggests that chromatin and its associated proteins, but not proteins that do not associate with chromatin, were specifically fractionated into the precipitate by the chromatin oligomerization.
To confirm our speculation, we performed a western blot analysis (Figure 2C). Histones H2A, H3 and H4 were detected in the precipitate but not in the supernatant, confirming the result from the silver stain. Together with the DNA distribution described above, this indicates that basic components of chromatin, such as DNA and core histones, are precipitated by the Mg2+-dependent oligomerization. To investigate whether non-histone chromatin proteins are also present in the precipitate, we analyzed Sir3p and Bdf1p. A western blot analysis revealed that both Sir3p and Bdf1p were present only in the precipitate. This result suggests that the non-histone chromatin proteins, in addition to the basic chromatin components, also fractionated into the precipitate. Finally, we assessed the distribution of an abundant cytoplasmic marker protein, Pgk1p. In contrast to the chromatin proteins analyzed above, we detected Pgk1p only in the supernatant. This indicates that Pgk1p contaminated the MNase-eluted fraction, but was excluded by the Mg2+-dependent oligomerization, and suggests that the non-chromatin associated proteins are not precipitated by the Mg2+ ions. We concluded that the oligomerized chromatin fractionation by the Mg2+ ions is useful to isolate the chromatin proteins.
We assumed that the H4-tail associated-proteins would be absent from the chromatin lacking the H4-tails. We compared the proteins of the oligomerized chromatin from the H4-tailΔ with those from the WT. Silver stained gels revealed that about equal amounts of H3, H2A and H2B were present in both oligomerized chromatin fractions (Figure 3A). In the H4-tail deleted chromatin, a band corresponding to the tail-deleted H4 (indicated as H4Δ on the left in Figure 3A) was present instead of the wild-type H4 band and was detected by western blotting with an anti-H4 antibody against the C-termini of histone H4 (data not shown). We further analyzed histones H2A and H3 by western blotting (Figure 3B), and observed equal amounts of H2A and H3 in the oligomerized chromatin fractions from both cells.
To elucidate whether our strategy would be useful for identifying H4-tail associated proteins, we again analyzed Sir3p and Bdf1p, which both interact with the H4-tails (4–7). Although the total amount of either Sir3p or Bdf1p in each strain was not changed (data not shown), lower amounts of both proteins were present in the oligomerized chromatin of the H4-tail deleted cells (Figure 3B compare WT and H4-tailΔ). This suggests that the amounts of H4 tail-binding proteins, such as Sir3p and Bdf1p, are decreased in the oligomerized chromatin fraction prepared from the H4 tail-deleted cells.
The proteins associated with the oligomerized chromatin from the WT and H4-tailΔ strains were then resolved by 2DGE and visualized by silver staining (Figure 4). The amount of the chromatin was measured by western blot analysis with the anti-H3 antibody, and then equal amounts of the oligomerized chromatin from the WT and H4-tailΔ strains were applied to the 2D gels. Over 350 protein spots were detected. We selected 36 spots representing proteins whose amounts were reduced in the fraction from H4-tail deleted cells. A total of 39 proteins were identified by the MS from the 36 spots (Figure 4).
We excluded the predicted contaminating proteins (Figure 4 in gray), which appeared non-specifically in previously reported large-scale experiments of protein complexes (27–29), from the 39 identified proteins. The predicted contaminants are abundant proteins in the cell, such as a ribosomal protein, a chaperone protein, and a metabolic enzyme. We also subtracted yeast transposons and L-A virus proteins. Consequently, we found 22 proteins whose amounts were reduced in the oligomerized chromatin fraction of the H4-tailΔ, as candidates for the H4 tail-associated proteins (Table 1). The steady state transcription levels of the genes that encode these 22 proteins did not significantly differ between the WT and H4-tailΔ strains (data not shown). The largest change was a 1.82-fold decrease in FPR4 expression. This suggests that the observed reduction in the amount of the proteins is not owing to a reduction of the gene expression.
Of the 22 potential proteins, two proteins, Arp4p and Isw1p, have already been shown to be involved in chromatin regulation via the H4 tail, and five proteins, Isw1p, Rsc8p, Pwp1p, Tif34p and YDL156p, contain a possible histone-tail-binding domain (SANT or WD40-repeat). Therefore, it is likely that our candidates are histone H4-tail-associated proteins. The interaction between Arp4p and histone-tails is supposed to contribute to the recruitment of an Arp4p-containing nucleosome acetyltransferase of histone H4 (NuA4) complex to nucleosomes (30,31). Isw1p, one of yeast imitation switch (ISWI) homologs, requires the H4-tail for its chromatin remodeling activity. ISWI probably recognizes a DNA-bound H4-tail, as a patch of three residues, R17H18R19, that contact nucleosomal DNA, to stimulate the ATPase activity (32). Isw1p and Rsc8p contain a SANT domain, which was identified based on its homology to the DNA binding domain of c-Myb and is also present in the SWI3, ADA2, nuclear receptor co-repressor N-CoR and transcription factor (TF)IIIB factors, and is envisaged as a histone-tail-binding module that regulates the chromatin remodeling activity of the associated complex (33). Pwp1p, Tif34 and YDL156Wp have a WD40-repeat domain, which functions in the interactions among proteins in many cellular complexes including chromatin-regulated proteins. Several chromatin-regulated proteins containing the WD40-repeat are known as histone-tail binding proteins (described in Discussion). More interestingly, a WD40-repeat of WDR5 was shown to interact with histone H3-tails (17).
Many proteins form complexes to carry out cellular functions. Chromatin is associated with many complexes. Therefore, a decrease in the association of a protein with chromatin means a decrease in the association of its complex with chromatin. We looked for the complexes that contain the 22 proteins, using the Saccharomyces genome database (http://www.yeastgenome.org). We found 10 complexes that contain more than two proteins among the 22 proteins. Each of the 22 proteins was included as one of the subunits of the 10 complexes, except for Cdc12p (Table 2). As described below, most of the complexes are probably involved in chromatin-mediated cellular functions.
Complex #1: Arp4p, Isw1p, Npl6p, Rpc40p, Rsc6p and Rsc8p are components of the yTAP-C116 complex, which was found in a large-scale identification of yeast multi-protein complexes using tandem-affinity purification (TAP)-epitope-tagging (27). TAP-tagged Isw1p, Isw2p or Hhp10p all yielded the yTAP-C116 complex, containing various components of a chromatin remodeling complex, such as an ISWI complex, a RSC (for remodels the structure of chromatin) complex, an INO80 ATPase complex and a NuA4 histone H4/H2A acetyltransferase complex (34). This suggests that the yTAP-C116 complex may be involved in a chromatin remodeling function. The presence of Arp4p and Isw1p indicates that the yTAP-C116 complex appears to associate with chromatin via an H4-tail.
Complex #2: The yTAP-C154 complex contains many subunits of RNA Polymerases I, II and III and may function in transcription (27). Our candidates, Dhh1p, Rpa43p, Rpc40p and Rsc8p, were present in this complex. Although its association with an H4-tail is not known, Rsc8p has a SANT domain, which was proposed as a histone tail interaction domain, as mentioned above.
Complex #3: The Nop58/Sik1 complex consists of three proteins, Nop58p, Sik1p and Nsr1p, which are all our candidate proteins. This complex is involved in ribosomal RNA (rRNA) processing and small nucleolar RNA (snoRNA) synthesis (29). Whether the complex is associated with chromatin has not been investigated.
Complex #4: We found Arp2p, Rpc40p and Tif34p, which are included in the NTG1 complex (28). Ntg1p and Ntg2p, the Saccharomyces cerevisiae homologs of endonuclease III from Escherichia coli, are both required for the efficient repair of spontaneous or induced oxidative DNA damage in yeast (35). Although we did not investigate whether the NTG1 complex interacts with chromatin, recent reports have firmly established a mechanistic link between DNA repair and the recruitment of histone modifiers or ATP-dependent chromatin-remodeling complexes (36). A WD40-repeat within Tif34p may play a role in recruiting the NTG complex to chromatin.
Complex #5: The yTAP-C228 complex contains Fpr4p and Pwp1p (27). Fpr4p has both peptidyl propyl cis–trans isomerase activity and histone chaperone activity. Fpr4p is required for the in vivo silencing of the gene expression at a ribosomal DNA (rDNA) locus (37). Pwp1p, a protein with a WD40-repeat, is involved in rRNA processing and associates with trans-acting ribosome biogenesis factors (38). Thus, it is likely that the yTAP-C288 complex associates with the rDNA chromatin, and we will show that Pwp1p actually binds to the rDNA chromatin in an H4 tail-dependent manner (described below).
Complex #6: Both Rsc8p and YDL156Wp are components of the yTAP-C152 complex (27). This complex probably functions in DNA repair, since it includes many proteins (10 out of 35 components) involved in this process. Rsc8p may associate with a histone-tail by its SANT domain. YDL156Wp contains a WD40-repeat domain that shares 39% amino acid identity with that of the human Xeroderma pigmentosum group E (XP-E) protein p48 (also known as DDB2). p48 is a damage-specific DNA binding protein, and may play a role in repairing DNA lesions by associating with damaged chromatin via its WD40-repeat domain (39,40).
Complex #7: The general transcription factor TFIID has 15 subunits, including two candidates, Taf11p and Taf14p (41). The higher eukaryotic TAF1 has acetyltransferase activity, two bromodomains, and kinase activity. Its yeast homolog, Taf1p, has the acetyltransferase activity but lacks the bromodomains and the kinase activity. Bdf1p, like the higher eukaryotic TAF1, has the kinase activity located on the carboxyl-terminal side of its bromodomains. The structural and functional similarities suggest that Bdf1p corresponds to the carboxyl-terminal region of the higher eukaryotic TAF1 (42). The bromodomain of Bdf1p binds to an acetyl-lysine of the histone-tail (6,7). As shown in Figure 3B, we found that the amount of Bdf1p was lower in the oligomerized chromatin fraction from the H4-tail deleted cells. Thus, the amount of the TFIID complex associated with chromatin may be decreased in the H4-tail deleted cells.
Complex #8: Arp4p and Taf14p are components of both the NuA4 histone H4/H2A acetyltransferase and INO80 chromatin remodeling complexes (31,43). These complexes associate with chromatin through Arp4p (43,44).
Complex #9: Arp2p interacts with Srp1p both physically and genetically (45). Arp2p was shown to play an important role in nuclear pore complex structure and functions. Srp1, a karyopherin α, forms a heterodimer with Kap95p, a karyopherin β, to mediate the import of a nuclear protein (46). Kap95p preferentially associates with a subset of highly transcribed genes (47). Active transcription is linked to an H4-tail modification, such as acetylation. It is likely that the complex containing Arp2p and Srp1p mediates the interaction between transcriptionally active chromatin and a nuclear pore complex.
Complex #10: The PKH1 complex (28), which is composed of four subunits, includes Lsp1p and Pil1p. They negatively regulate the 3-phosphoinositide-dependent protein kinase-like kinase Pkh1p and downstream signaling pathways (48). It is not known whether this complex interacts with chromatin. However, Pil1p is included in the yTAP-C7 complex, which was purified with TAP-tagged Scs2p (27). It is interesting that the deletion of SCS2 reduced the silencing at telomeric chromatin, which is also caused by an H4-tail deletion (49).
As described above, our potential proteins, except for Cdc12p, are contained in one of the 10 complexes that may be involved in chromatin regulation. Of the 10 complexes, 3 are suggested to interact with the H4-tail by known tail-binding proteins and 5 have the potential for H4-tail binding activity by the proposed histone-tail associated domains: Arp4p in the #1 and #8 complexes, Isw1p in the #1 complex and Bdf1p in the #7 complex; a SANT domain of Rsc8 in the #1, #2 and #6 complexes, and WD40-repeat proteins in the #4, #5 and #6 complexes. In future experiments, we will investigate whether the H4-tail is required for chromatin association in each complex.
Our three candidate proteins, Pwp1p, Tif34p and YDL156Wp, share a WD40-repeat domain, which is present in many chromatin-associated factors. Several proteins containing the WD40-repeat have binding activity to a histone tail, and one of the several proteins, WDR5, binds to histone H3-tail via its WD40-repeat domain (described in Discussion). We focused on Pwp1p and constructed yeast strains in which the endogenous copy of the PWP1 gene was modified to encode either 13 copies of a Myc epitope or 3 copies of a FLAG epitope at its C-terminus. PWP1 is an essential gene, but neither tagged strain displayed a growth defect, indicating that the tagged proteins are functional. We investigated whether the amount of Pwp1p was reduced in the H4-tail deleted chromatin by a western blot analysis using Myc-tagged Pwp1p (Figure 5). Equal amounts of the oligomerized chromatin from both the WT and H4-tailΔ strains were fractionated on the gels and analyzed by western blotting with the anti-H3 and anti-H2A antibodies. The most appropriate loading control in this experiment is abundance of histone H4, which could be detected with the anti-H4 antibody against the C-terminus of histone H4. Unfortunately, we were not able to use histone H4 as the loading control, since the antibody gave the less signal intensity from the tail-deleted H4 than that from wild-type histone H4 in our western blot (data not shown). Chromatin consists of an equal amount of each core histone. The integrity of nucleosome is not altered in this H4-tail deleted strain (50). Therefore, we chose histone H3, instead of histone H4, as a loading control, whose amount represents the amount of chromatin. The amount of Pwp1p was decreased in the H4-tail deleted chromatin in a similar manner as Bdf1p, which has been shown to interact with the H4-tail (Figure 5A). The reductions of Bdf1p and Pwp1p were calculated to be 89 and 69%, respectively, when the signals were quantified in comparison with the histone H3 signal. On the other hand, the reduction of H2A was not observed (Figure 5A and and5B).5B). The total amounts of Pwp1p in the cell were the same in the WT and H4-tailΔ strains (data not shown). Therefore, we concluded that the amount of Pwp1p in the oligomerized chromatin fraction is decreased by the H4-tail deletion.
Pwp1p is localized in the nucleolus and is involved in rRNA processing (38,51). However, how Pwp1p localizes to the nucleolus is unknown. The localization of Pwp1p to the nucleolus raised the possibility that Pwp1p may associate with the chromatin at rDNA. We assessed whether Pwp1p associates with the chromatin at rDNA by a ChrIP assay using the FLAG-tagged Pwp1p strain. To confirm our ChrIP assay using a FLAG-tagged protein, we investigated enrichment of FLAG-tagged Sir3p at silent chromatin. FLAG-tagged Sir3p is specifically recognized by an anti-FLAG antibody (data not shown). The immunoprecipitated DNA was quantified by multiplex PCR with primers for an HMR silent chromatin region, an INO1 promoter region, and a loading control, a CUP1 region (52). Our PCR amplifications were in the linear range (Figure 6A lanes 1–5). To calculate the relative fold enrichment of Sir3p binding to chromatin, we first normalized each signal to the CUP1 loading control, and then compared its normalized signal with that of the input DNA used in each ChrIP assay. The fold enrichment values of an untagged strain indicated background binding. Consistent with the previous work (53), we found that Sir3p associates with the silent chromatin at the HMR region preferentially (3.1-fold), but not at the INO1 region (Figure 6A lane 10 and and6B).6B). We then used the anti-FLAG antibody to immunoprecipitate the FLAG-tagged Pwp1p-associated chromatin DNA. FLAG-tagged Pwp1p is specifically recognized by an anti-FLAG antibody (data not shown). We examined the Pwp1p binding to a 9.1 kb rDNA repeat unit with 16 sets of primers that spanned the rDNA repeat unit (Figure 6E). The immunoprecipitated DNA was quantified by multiplex PCR with two sets of the primers for the rDNA regions and one set for the CUP1 region (52). To calculate the relative fold enrichment of Pwp1p binding to the rDNA chromatin, we first normalized the rDNA regions to the CUP1 region, and then compared its normalized signal with that of input DNA. We generated a graphical representation of Pwp1p association across the rDNA repeat using the relative fold enrichment values for each DNA fragment (Figure 6E, closed circles). The fold enrichment values of the untagged strain indicated background binding (Figure 6E open circles). We found that Pwp1p only associated with a 5′ region of 25S rDNA, not at other regions. The ChrIP data used to obtain the graph at 3.1 kb (25S) and 8.1 kb (5S) positions and the linear amplification of the PCR is shown in Figure 6C and and6D.6D. We found that the immunoprecipitated DNA from the 25S rDNA region was enriched ~5.5-fold in the FLAG-tagged strain. In contrast, the immunoprecipitated DNA from the 5S rDNA and the INO1 regions were not significantly enriched in the FLAG-tagged strain (Figure 6C and and6D6D for the 5S rDNA region andFigure 6A and and6B6B for the INO1 region). These results indicated that Pwp1p binds preferentially to the chromatin at the 25S rDNA region.
We next investigated whether the Pwp1p binding to the 25S rDNA chromatin is dependent on the H4-tail. As shown in Figure 6C–E, no enrichment was observed at the 25S rDNA regions in the FLAG-tagged H4-tailΔ strain. Quantification showed that the fold enrichment of the Pwp1p binding to the 25S rDNA region in the FLAG-tagged and the untagged H4-tailΔ strains was 2.6-fold and 1.7-fold, respectively (Figure 6D). These results demonstrated that the level of the Pwp1p binding at the 25S regions in the H4-tailΔ strain is much reduced compared to that in the WT strain. Thus, we obtained in vivo evidence that Pwp1p associates with the 25S rDNA chromatin and its association is dependent on the N-terminal-tail of histone H4.
We performed the protein differential display approach in 2DGE with the H4-tail deleted mutant to identify the H4-tail-associated-proteins by MS and identified 22 candidate proteins, including the H4-tail binding proteins, Arp4p and Isw1p, and proteins with the SANT and WD40-repeat, which may be a histone tail-binding domain. We also provided in vivo evidence that one candidate, Pwp1p, associates with the rDNA chromatin in an H4 tail-dependent manner.
In previous reports, in order to identify protein complexes capable of binding specifically to histone H3 N-terminal tail, nuclear extracts were applied to affinity columns displaying either unmodified H3 tails or the same tails but differently methylated at specific residues. The NuRD complex binds specifically to unmodified H3 tails but not to K4 methylated tails (14). Two proteins, INHAT complex subunits SET and pp32, bind specifically to unmodified H3 tails but not to phosphorylated tails at T3 (16). hSNF2H and WDR5 were identified as H3 tail-binding proteins that associate with the methylated tail at K4 (15,17). The 14-3-3 isoforms also bind to the phosphorylated H3-tail at S10 (18). These protein identifications proved that the peptide affinity column is a powerful method to identify the tail-binding proteins. However, this approach is dependent on the preparation of a nuclear extract and the salt concentration of the binding buffer. Most importantly, the peptide conformation could be different from that in the chromatin context in vivo. Thus, the screening may be restricted, and some proteins would not be identified. By a yeast two-hybrid screen employing yeast genomic libraries in combination with a bait plasmid expressing a fragment of histone H4 (amino acids 1–59), only two proteins, Bdf1p and Hif1p, were identified (54,55). This indicates that the two-hybrid screen also has limitations. The authors suggested that a folding difference may exist between the full-length histone H4 and the portion of amino acids 1–59 (54). Again, the tail in this assay may have a different conformation as compared to that in the chromatin context. In this regard, our approach seems to be more suitable than the others to identify tail-binding proteins. The nucleosome ladder was detected by MNase digestion in our H4-tail mutant, showing the maintenance of the nucleosome structure in the absence of H4-tails (50 and data not shown). Therefore, our screening reflects the H4-tail binding ability in vivo.
We enriched chromatin proteins by the Mg2+-dependent oligomerization for an efficient identification of a histone-tail binding protein (Figure 2). However, silver stained 2D gels were not sensitive enough to identify Sir3p, whose abundance was reduced in western blot (Figure 3B). Total mixture mass spectrometry is more sensitive than the silver stain. This method will help to identify less-abundant proteins.
Our screening is based on in vivo binding activity. So a candidate protein may interact either directly or indirectly with an H4-tail. We identified Isw1p, whose fly homolog, ISWI, probably recognizes a DNA-bound H4-tail. ISWI does not interact stably with the GST–H4 tail fusion protein in the absence of DNA (32). This type of a binding protein cannot be identified by a widely used approach such as the peptide affinity column.
Histone tails are involved in folding higher-order chromatin structures (2). In the tail mutant strain, the chromatin structures may be altered at the higher-order folding level. The MNase digestion produced a nucleosomal ladder in the H4-tailΔ, and its MNase sensitivity was higher than that of the WT, suggesting a structural alteration beyond the nucleosomal level by the tail-deletion (50). If there is a protein that recognizes the higher-order chromatin structure, then the amount of this protein would also be decreased in our screening. Such a protein may be included among our candidate proteins. We cannot examine this possibility at present, as there is no biochemical method to assess the folding of higher-order chromatin in vivo.
The WD40-repeat is defined by a sequence repeat of ~40 amino acids, typically beginning with a glycine–histidine pair and ending with a tryptophan–aspartic acid pair. This motif is shared among over 30 functional families, which are involved in signal transduction, mRNA synthesis, RNA splicing, vascular trafficking, cytoskeletal assembly, control of transcription initiation complex assembly and a chromatin-regulated complex (56). Among the chromatin-regulated proteins, several WD40-repeat proteins are known as histone-tail binding proteins. For example, the transcriptional repressor proteins Tup1p, Groucho and transducin beta-like protein (TBL1)/TBL1-related protein (TBLR1) associate with a histone-tail via a domain other than the WD40-repeat domain (57–59). WDR5, a common component of three H3 K4 methyltransferase complexes (the mixed-lineage leukemia gene (MLL)1, MLL2 and hSet1), directly associates with a histone H3-tail di- and trimethylated at K4 via the WD40-repeat domain itself (17). We plan to investigate whether the three candidates (Pwp1p, Tif34p, and YDL156Wp) interact with the H4-tail directly or indirectly.
rDNA transcription accounts for ~60% of the transcription in a rapidly growing yeast cell. However, only about half of the ~150 copies of the rDNA are active at any given time, whereas the remaining copies are maintained in a silenced state. This ratio of active to inactive genes is stably propagated throughout the cell cycle and is independent of the transcriptional activity of the cell. The regulatory mechanism that controls the ratio of active to inactive rDNA genes is poorly understood, but rDNA silencing is one of the factors that establish and maintain the transcriptionally inactive rDNA genes. We showed that Pwp1p associates with the rDNA chromatin, and its association is dependent on the H4-tail. The importance of the H4-tail for rDNA regulation was also suggested by studies of a mammalian nucleolar remodeling complex (NoRC) and a yeast regulator of nucleolar silencing and telophase exit (RENT) complex, which both regulate rDNA silencing. The NoRC induces nucleosome sliding in an ATP- and histone H4 tail-dependent fashion, and the NoRC-directed rDNA repression requires the histone H4-tail (60,61). The RENT complex mediates the rDNA silencing and deacetylates the acetylated K16 of the H4-tail by Sir2p (52,62–64). These studies raised the possibility that the H4-tail binding protein that associates with the rDNA chromatin performs a role similar to that of Sir3p in the silencing at telomeric and HM loci. Pwp1p may be the factor that links the RENT complex to the rDNA chromatin through the H4-tail.
The yTAP-C288 complex consists of eight subunits: Act1p, Eno2p, Fpr4p, Nan1p, Pol5p, Pwp1p, Smc1p and YPL207Wp (27). Other subunits besides Pwp1p may also function in the rDNA regulation. For instance, Fpr4p binds to rDNA chromatin and regulates rDNA silencing (37); Nan1p associates with the RENT complex (63) and Pol5p is required for the synthesis of rRNA (65). In addition, Smc1p is a member of a ubiquitous family of chromosome-associated ATPases and plays a role in chromosome dynamics (66). Taken together, we propose that the yTAP-C228 complex associates with rDNA chromatin by the histone H4-tail and regulates rDNA transcription by modulating the chromatin structure.
We compared the Mg2+-dependent oligomerized chromatins isolated from H4 tail deletion cells and wild type cells by the protein differential display approach in 2DGE, and effectively identified histone-tail binding proteins. Other tail-deleted and point mutant cells are possible sources for the identification of tail-binding proteins for other core histones and specifically modified histones, respectively.
The authors are grateful to Michael Grunstein for the histone-tail-mutant strain, Toshio Tsukiyama for the p3FLAG-KanMX plasmid, Mark Longtine for the pFA6a-13Myc-TRP1 plasmid, and Shirleen G. Roeder for the anti-BDF1 antibody. This work was supported by a research grant endorsed by the New Energy and Industrial Technology Development Organization (NEDO) and by Grants-in-Aid from the Ministry of Education, Science, Sports and Culture of Japan (MEXT). Funding to pay the Open Access publication charges for this article was provided by MEXT.
Conflict of interest statement. None declared.