Abstract

The Saccharomyces cerevisiae Yta7 protein is a component of a nucleosome bound protein complex that maintains distinct transcriptional zones of chromatin. We previously found that one protein copurifying with Yta7 is the yFACT member Spt16. Epistasis analyses revealed a link between Yta7, Spt16, and other previously identified members of the histone regulatory pathway. In concurrence, Yta7 was found to regulate histone gene transcription in a cell-cycle-dependent manner. Association at the histone gene loci appeared to occur through binding of the bromodomain-like region of Yta7 with the N-terminal tail of histone H3. Our work suggests a mechanism in which Yta7 is localized to chromatin to establish regions of transcriptional silencing, and that one facet of this cellular mechanism is to modulate transcription of histone genes.

DIFFERENT regulatory mechanisms are involved in the cell cycle control of gene transcription. Many of these pathways are centered around the regulation of chromatin and the histone epigenetic status (Berger 2007; Li et al. 2007). ATP-dependent and -independent chromatin remodeling provides for the physical rearrangement of nucleosomes and accessibility to underlying DNA (Cairns 2005). Furthermore, histones can be dynamically cycled out of chromatin to define chromosomal features such as active promoters and chromatin boundary elements (Dion et al. 2007; Jamai et al. 2007; Rufiange et al. 2007). They can even be replaced by histone variants to regulate transcription (Li et al. 2007). A more combinatorial process of transcriptional regulation is the post-translational modification (PTM) of histones (Goldberg et al. 2007; Kouzarides 2007). Certain histone PTMs serve as binding sites for recruitment of “effector” protein complexes that modulate gene transcription. For example, the Yng1 component from the NuA3 histone acetyltransferase binds to lysine 4 trimethylated histone H3 to provide for histone H3 lysine 14 acetylation and transcriptional activation (Taverna et al. 2006). Transcriptional regulation through chromatin is undoubtedly a complex system of events that remain to be fully explored.

In previous work, we described a novel transcriptional role for the yeast tat-binding analog 7 (Yta7) (Tackett et al. 2005). We found that Yta7 maintains the transition region (or barrier) between heterochromatin and euchromatin upstream of the silent HMR locus (Tackett et al. 2005). Deletion of the YTA7 gene results in spreading of the silent state from HMR and silencing of surrounding genes, demonstrating that Yta7 can regulate distinct transcriptional states. Additional work has shown that Yta7 functions in a similar manner downstream of the silent region at HMR (Jambunathan et al. 2005). Yta7 contains both an AAA ATPase domain and a bromodomain-like region (Yta7bd). An affinity-tagged version of Yta7 shows avid copurification of all core histones, suggesting that activity may be propagated through nucleosome association (Tackett et al. 2005). Yta7bd can mediate the association with histones; however, the precise region of histone interaction and post-translational modification status has not been established (Jambunathan et al. 2005). Yta7 also appears to be needed for transcription of repetitive DNA sequences in Caenorhabditis elegans, where its deletion leads to embryonic lethality (Tseng et al. 2007). Additionally, Yta7 has been linked to telomere maintenance and has shown an interaction with Rad53 upon DNA damage (Askree et al. 2004; Smolka et al. 2005). These studies suggest that Yta7 may be associating with chromatin to modulate transcription and to function in numerous cellular pathways.

In previous work, we observed that Yta7 copurifies with transcriptional and chromatin modifying proteins (Tackett et al. 2005). Intriguingly, we noted that Spt16 is one of many proteins that copurified with Yta7 (Tackett et al. 2005). The Spt16 protein is involved in transcription and nucleosome regulation via its role with the yFACT complex (Formosa et al. 2002). Spt16 appears to be essential for displacement of histones to promote initiation of transcription (providing for passage of RNA polymerase) and for reassembly of nucleosomes after elongation (Formosa et al. 2002; Biswas et al. 2005). Since Yta7 regulates zones of transcription and Spt16 plays a crucial role in transcription, we sought to explore the functional significance of the observed Yta7/Spt16 copurification.

MATERIALS AND METHODS

Yeast strains:

Yeast strains are listed in Table 1. Gene replacements of YTA7 with LEU2 were done by PCR amplification of LEU2 from the pUG73 plasmid with 45 bp of sequence upstream and downstream of the YTA7 gene, genomic incorporation by homologous recombination and conformation by PCR. Gene replacements of YTA7 with KAN were done by PCR amplification of KAN from yta7∷KAN BY4742 (Open Biosystems) genomic DNA with 100 bp of sequence upstream and downstream of the YTA7 gene, incorporation into the genome by homologous recombination and confirmation by PCR. To make strain ATY146, we used marker swap plasmid M4755 to change the KAN marker from a bar1∷KAN BY4741 strain (Open Biosystems) to a LEU2 marker (Voth et al. 2003). For strains ATY223 and ATY232, YTA7 was internally tagged by insertion of PrA or Myc9 as described (Gauss et al. 2005).

View this table:
TABLE 1

S. cerevisiae strains used in this study

Epistasis analysis:

Temperature sensitivity: cells were grown to saturation at 25°, 10-fold serially diluted, spotted on YEPD plates and incubated at various temperatures (25°, 30°, 32°, 34°, and 37°). Damage assays: cells were grown to saturation at 25°, 10-fold serially diluted and spotted on YEPD plates containing methyl methanesulfonate (MMS, 0.01, 0.03, 0.1, and 0.3%), hydroxyurea (0.1 and 0.2 m) or 6-azauracil (6-AU, 75, 100, and 200 μg/ml). All serial dilutions were performed in triplicate.

Transcriptional assays:

Cells were grown to mid-log phase and blocked with either 50 nm α-factor (bar1Δ and yta7Δ bar1Δ) or 15 μg/ml nocodazole [wild-type (BY4742), yta7Δ, and hir2Δ] for 2.5–3 hr at 30°. α-Factor showed >90% blocking efficiency, while nocodazole showed >80%. To release the cell cycle block, cells were washed twice in ice cold water and resuspended in 30° prewarmed YEPD media. Time point samples were taken for budding index counting, RNA extraction and flow cytometer analysis.

Percent new buds was calculated as the percentage of cells forming new buds relative to the total population. Cell shape was classified as unbudded single cells, newly formed small buds or large buds. Cells arrested in nocodazole were considered large budded, while those arrested in α-factor were classified as unbudded single cells. We plotted the formation of newly formed small buds as this is a rough visual representation of the onset of histone gene transcription and S phase (Driscoll et al. 2007).

Hot acidic phenol extraction of RNA was performed as previously described (Tackett et al. 2005). Reverse transcription was conducted with the Invitrogen SuperScript first strand kit according to the providers' protocol. Transcription levels of the histone genes were determined by real-time PCR. ACT1 was chosen as a housekeeping gene. CLN1 mRNA was used as a marker of the cell cycle. Primer sequences are listed in Table 2. Primers to differentiate HHF1 and HHF2 transcripts could not be designed due to sequence similarity, thus both transcripts were simultaneously monitored with a single primer set. Specificity of primers for HTA1, HTA2, HTB2, HHT1, and HHT2 transcripts was determined by real-time PCR. To measure primer specificity, cDNA was prepared from genomic deletion strains (hta1Δ, hta2Δ, htb2Δ, hht1Δ, and hht2Δ) and analyzed by real-time PCR with the primers corresponding to the deleted gene. Nonspecific amplification with each primer set was measured relative to ACT1 mRNA. Relative mRNA levels in Figures 3 and 4 were corrected for primer specificity. We did not determine the specificity of HTB1 transcript primers because systematic deletion of the HTB1 gene produces inviable cells (Giaever et al. 2002).

View this table:
TABLE 2

Sequences of the primers used for real-time PCR

Flow cytometer samples were fixed in 70% ethanol and stored at 4°. After sonication in 200 mm sodium citrate, cells were treated with RNase A at 37°. SYTOX green (Molecular Probes/Invitrogen) was added to a final concentration of 2 μm. Samples were analyzed with a Beckman Coulter EPICS XL-MCL flow cytometer.

Chromatin immunoprecipitation:

Chromatin immunoprecipitation (ChIP) was performed as previously reported using an anti-H3K56ac antibody (Upstate/Millipore 07-677) (Taverna et al. 2006). ChIP to general histone H3 (Abcam Ab1791) was performed to control for nucleosome occupancy (Taverna et al. 2006). Real-time PCR was used to compare the enriched levels of the intergenic region between the HTA1 and HTB1 gene pair relative to ACT1.

High resolution ChIP-chip analysis of Yta7:

ChIP-chip analysis of Yta7-Myc was performed in triplicate with dye-swapping as reported (Ren et al. 2000; Taverna et al. 2006). Tiled microarrays covering the entire Saccharomyces cerevisiae genome were utilized (50 bp resolution, NimbleGen Systems). Microarray data were independently analyzed with both SignalMap (NimbleGen Systems) and Mpeak version 2.0 software (http://www.stat.ucla.edu/∼zmdl/mpeak/). Binding regions were considered relevant if >2 SD from the mean.

We utilized a strain containing an internal Myc9 tag that was inserted at amino acid position 1200 of Yta7 (Gauss et al. 2005). The location of the tag was determined via two different approaches analyzing four different positions: N-terminal, insertion at position 106 (prior to the AAA ATPase and bromodomains), insertion at position 1200 (near the C terminus), and C-terminal. Each tagged strain was tested for maintenance of barrier activity at HMR and for immunopurification of histones (Tackett et al. 2005). The results showed that only the insertion of the tag at amino acid 1200 preserved barrier maintenance and provided for histone immunopurification (A. J. Tackett and B. T. Chait, unpublished observations).

Yta7bd binding to histones and histone H3 peptide mimics:

Amino acids 999–1101 composing the Yta7 bromodomain-like region (Yta7bd) with a N-terminal GST fusion were expressed, purified, and confirmed by mass spectrometry (Taverna et al. 2006).

For histone binding studies, 40 μg of GST-Yta7bd (or GST control) was incubated with 5 μg of acid-extracted T. thermophila histones (purified as described in Taverna et al. 2007) and 50 μg of BSA (5 μl from a 10 mg/ml New England BioLabs stock) in binding buffer (20 mm HEPES, pH 7.9, 25% glycerol, 150 mm NaCl, 1.5 mm MgCl2, 0.2 mm EDTA, 1% Triton X-100, 10 mm sodium butyrate, 1:100 Sigma yeast protease inhibitor cocktail) in a final volume of 500 μl for 1 hr at room temperature. GST-Yta7bd or GST alone was collected for 1 hr at room temperature with EZview red glutathione affinity gel (Sigma), washed three times in binding buffer with 300 mm NaCl, washed once with low salt buffer (4 mm HEPES, pH 7.9, 10 mm NaCl), and copurifying histones were released by heating to 95° in SDS–PAGE sample buffer. Bound histones were resolved with 4–20% SDS–PAGE and individually visualized by immunoblotting.

For histone H3 peptide mimic binding studies, streptavidin-coated Dynabeads M280 (250 μg) were incubated with 1.2 μg of biotinylated histone H3 peptide mimics (Upstate/Millipore or in-house synthesized). Peptide amounts were standardized by dot-blotting and visualization with streptavidin-HRP. Binding assays were performed by incubation of Yta7bd with the peptide coated beads for 1 hr at room temperature in 20 mm HEPES, pH 7.4, 150 mm NaCl, 0.2% Triton X-100. Beads were washed three times with 20 mm HEPES, pH 7.4, 300 mm NaCl, 0.2% Triton X-100. Proteins were resolved by 4–12% SDS–PAGE and visualized by Coomassie staining. Binding was quantified with ImageJ software (http://rsb.info.nih.gov/ij/).

Full-length Yta7 binding to histone H3 peptides:

Yta7-PrA was immunopurified on IgG-coated Dynabeads as previously described except that the concentration of NaCl was raised from 0.3 to 1.5 m such that cellular histones no longer copurified (Tackett et al. 2005). Yta7-PrA was released from the resin with a nondenaturing elution peptide (Strambio-De-Castillia et al. 2005; Tackett et al. 2005). Binding studies of Yta7-PrA were performed following the same protocol as for Yta7bd binding experiments with the following biotinylated peptides: H3 1-20, H3 48-63, and H3 48-63 K56ac. Bound Yta7-PrA was visualized by immunoblotting for the PrA tag.

Immunopurification of Yta7-PrA and associated histones:

Purification of Yta7-PrA with in vivo associated histones on IgG-coated Dynabeads was done as previously reported (Gauss et al. 2005; Tackett et al. 2005). Sodium butyrate (50 mm) was included during immunopurifications to inhibit histone deacetylases.

Mass spectrometric analysis of H3K56ac:

The gel band containing histone H3 (that purified with Yta7-PrA) was chemically treated with d6-acetic anhydride and digested with trypsin as previously reported (Dilworth et al. 2005; Tackett et al. 2005). The d6-acetic anhydride treatment converts all unmodified lysines to triply deuterated acetyl-lysines (Tackett et al. 2005); thus allowing one to differentiate unmodified lysines (+45 Da) from in vivo acetylated lysines (+42 Da) for a quantitative measurement of histone acetylation (Tackett et al. 2005). A mass spectrum of the histone H3 peptides was collected with a MALDI-prOTOF mass spectrometer (PerkinElmerSciex). Peptides were confirmed by tandem mass analysis with a vMALDI-LTQ ion trap mass spectrometer (Thermo). To quantify the level of acetylation for H3K56, m-over-z software was used to extract the monoisotopic peak areas for the heavy and light versions of the histone H3 53-63 peptide. Percent acetylated was the percentage of light area relative to the total percentage of light and heavy areas. Peptides identified are listed in Table 3.

View this table:
TABLE 3

Peptides from histone H3 copurifying with Yta7-PrA that were identified by MALDI mass spectrometry

RESULTS

Yta7 activities overlaps with the activities of other known histone regulators:

Spt16, a subunit of the yFACT complex, was found to coenrich with Yta7 (Tackett et al. 2005). The yFACT complex plays an important role during the cell cycle, helping the passage of polymerases through chromatin and nucleosome reassembly (Formosa et al. 2002). When subunits of the yFACT are mutated the reassembly of nucleosomes can still be handled by the Hir/Hpc complex, but simultaneous mutation of proteins from each complex leads to strong synthetic defects (Formosa et al. 2002). For instance, a combination of hir1Δ spt16-11(T8281I P859S) caused synthetic temperature sensitivity of the cells, demonstrating dependence of the yFACT pathway with the Hir pathway (Sutton et al. 2001; Formosa et al. 2002). In consequence, we decided to investigate the effect of double deletions of the YTA7 gene with the point mutation spt16-11 on cell growth, as well as with complete deletion of two genes encoding proteins in the Hir/Hpc complex: HIR1 and HIR2. The spt16-11 mutant was chosen because cells carrying the genomic deletion of SPT16 are inviable.

Additionally, the genes encoding Asf1 and Sas3 were deleted in WT and yta7Δ strains and cells were subjected to restrictive growth conditions. Asf1 is a histone chaperone protein that is widely involved in replication-coupled nucleosome assembly, response to DNA damage, repression of histone gene transcription and heterochromatic silencing (reviewed in Mousson et al. 2007). Asf1 mediates its activity through interactions with proteins including the Hirs (Sutton et al. 2001). The protein Sas3 is a component of the NuA3 histone acetyltransferase complex and interacts with Spt16, suggesting a role in transcription and/or replication (John et al. 2000; Taverna et al. 2006). Sas3 also copurifies with Yta7 (Tackett et al. 2005).

Cells were plated in serial dilutions and subjected to increasing temperatures or damaging agents: hydroxyurea (stalls DNA replication), methyl methanesulfonate (DNA damaging agent) or 6-azauracil (inhibits transcriptional elongation) (Figure 1). The 6-AU screen did not show any synthetic defects (data not shown). Cells containing a deletion of the RTT109 gene served as a sensitive control for the DNA damage agents (Driscoll et al. 2007; Han et al. 2007).

Figure 1.—

Yta7 activity overlaps with the activities of other described histone regulators. (A) Temperature sensitivity. Indicated cells were 10-fold serially diluted and incubated at 25° or 34° for the indicated days (d). (B) Hydroxyurea (HU) sensitivity. Cells were 10-fold serially diluted and incubated at 25° for the indicated days. (C) Methyl methanesulfonate (MMS) sensitivity. Cells were 10-fold serially diluted and incubated at 25° for the indicated days.

Cells harboring the spt16-11 mutation exhibited sensitivity to both temperature and HU treatments as previously described (Formosa et al. 2002) (Figure 1, A and B). Additional deletion of the YTA7 gene increased the growth defect observed for both treatments. This result suggests that both Yta7 and Spt16 are acting in two functionally overlapping pathways. No effect was observed for sas3Δ and sas3Δ yta7Δ mutants under any stress condition (data not shown). However, deletion of the ASF1 gene showed a cell growth defect under temperature and HU conditions. Further deletion of the YTA7 gene increased the HU effect only, indicating that Yta7 and Asf1 have overlapping function during times of DNA synthesis. Furthermore, deletion of the YTA7 gene induced growth defects in hir1Δ and hir2Δ cells in response to temperature increase, HU and MMS—suggesting that Hir and Yta7 pathways are functionally overlapping. Our results imply (1) a role of Yta7 in DNA damage response and (2) that Yta7 operates through a functionally overlapping pathway with the Asf1/Hir and Spt16-yFACT complexes.

Yta7 is a repressor of histone gene transcription:

Since we observed a genetic interaction with known histone gene transcriptional regulators (Hirs, Asf1, and Spt16) and since Yta7 is known to regulate transcriptional zones of chromatin via barrier activity, we investigated whether Yta7 played a role in transcriptional regulation of the histone genes. To assay for cell-cycle-dependent transcriptional defects, we synchronized wild-type and yta7Δ cells with α-factor (G1-phase arrest) or nocodazole (M-phase arrest) and cultures were sampled at various times throughout a full cell cycle.

We first verified that the cells were growing in synchrony. Measurement of the late G1 cyclin CLN1 transcript levels revealed that after release from each block, the cells progressed in a similar manner (Figure 2, A and C). However, cells containing the YTA7 gene deletion showed a defect in CLN1 transcription at the 90 and 100 min time points post-nocodazole release, which could indicate a lost of synchrony toward the end of the first cell cycle (Figure 2C). Following the cell cycle through the formation of new buds also showed a similar progression of the cells after release from the blocking agents (Figure 2, B and D), with a slight delay in the accumulation of buds at 80 min post-release from nocodazole for yta7Δ cells (Figure 2D). Thus, the cells appeared to be synchronized relative to expression of a G1 cyclin and formation of buds. However, release of yta7Δ cells from the nocodazole block did show CLN1 transcription and budding defects at times consistent with M phase (∼90 min post-release).

Figure 2.—

Timing of CLN1 gene transcription and new bud formation in α-factor synchronized bar1Δ and yta7Δ bar1Δ cells and in nocodazole synchronized wild-type and yta7Δ cells. Cells were blocked with α-factor (A and B) or nocodazole (C and D), released, and samples were taken at the indicated times. For each time point, the amount of late G1-cyclin CLN1 transcript relative to ACT1 transcript was determined by real-time PCR (A and C). Error bars are the standard deviation. The formation of new buds was also monitored (B and D).

For the time course transcriptional analyses of yta7Δ cells with either a nocodazole or an α-factor block, the peak of histone transcription appeared as expected, after the peak of CLN1 gene transcription in accordance with previous studies (Figure 2, A and C; Figure 3; Figure 4A) (Hereford et al. 1981). However, the transcriptional analysis of all the histone genes revealed strikingly different effects of YTA7 gene deletion with the different cell cycle blocks (Figures 3 and 4A). After α-factor block and release, yta7Δ cells revealed a significant higher activation of transcription for all the histone genes at the 20 and 30 min time points (Figure 3). Even though gene transcription seemed to start at similar times (∼20 min post-release), there was an apparent 20-min window where gene transcription in yta7Δ cells was higher than in wild-type, but then gene transcription decreased in a similar manner to the control strain (50 and 60 min time points). This α-factor blocking experiment was repeated three times with identical results. We also varied which culture (yta7Δ or wild-type cells) was sampled first at each time point, avoiding sampling discrepancies. One explanation of the apparent early transcription of each histone gene in yta7Δ cells is that Yta7 serves as a cell-cycle-dependent transcriptional repressor of the histone genes.

Figure 3.—

Yta7 regulates histone gene transcription. bar1Δ (solid circles) and yta7Δ bar1Δ (open squares) cells were blocked with α-factor, released, and samples were taken at the indicated times. For each time point, cDNA was prepared and the amount of indicated transcript (relative to ACT1 transcript) was determined by real-time PCR. Darkly shaded regions emphasize the increased histone gene transcription upon gene deletion of YTA7. Error bars are the standard deviation. Asterisks show significant differences determined by one-tailed t-testing (P < 0.05).

Figure 4.—

Nocodazole blocking alters the observed yta7Δ transcriptional defect in histone gene regulation. (A) Wild-type (solid circles) and yta7Δ (open squares) cells were blocked with nocodazole and released. Samples were taken at the indicated times and analyzed as in Figure 3. Darkly shaded regions emphasize the increased histone gene transcription upon gene deletion of YTA7. Error bars are the standard deviation. (B) HTB1 gene transcription is upregulated upon deletion of HIR2. Wild-type (solid circles) and hir2Δ (open squares) cells were blocked with nocodazole and released. Samples were taken at the indicated times and analyzed as in Figure 3. Error bars are the standard deviation. Asterisks show significant differences determined by one-tailed t-testing (P < 0.05).

The nocodazole block and release experiment revealed a different trend (Figure 4A). For each histone gene, a prolonged transcription in yta7Δ cells was apparent after the peak (80–90 min post-release). This defect in transcription was of similar magnitude to that observed upon the gene deletion of the known histone repressor Hir2 at 80 min post-release (Figure 4B). Unlike the α-factor blocking experiment, a uniform defect in transcriptional regulation was not evident. For example, the HTA1 and HTB1 and HTA2 and HTB2 gene pairs are known to be regulated in a similar manner (Hereford et al. 1981). In our nocodazole blocking experiment, the defect observed upon deletion of YTA7 showed a similar defect for HTA1 and HTA2 and HTB1 and HTB2 (Figure 4A). Since we observed prolonged histone gene transcription and a differential histone gene transcriptional regulation (relative to the α-factor blocking experiment), we decided to take a more detailed look into possible cell cycle defects in each of the blocking experiments.

In Figure 2, we showed that wild-type and yta7Δ cells had similar CLN1 transcript levels and new bud formation, indicative of synchronous growth. To investigate cell synchrony in a more detailed manner, we monitored DNA content by flow cytometry (Figure 5). Following α-factor block, cells were released and grew in synchrony (Figure 5, A and B). This indicated that the observed early transcriptional effect is likely not a consequence of a cell cycle progression defect (Figure 3). However, the nocodazole block and release exhibited slightly different results. YTA7 gene deletion impaired the ability of the cells to exit from M phase or recover from the nocodazole block (Figure 5, C and D). In Figure 5D, the yta7Δ cells showed a delay in the transition from 2N to 1N content relative to wild-type. This led to two important conclusions. First, the prolonged and variable histone gene transcription observed in Figure 4A is likely a product of asynchrony, and that the true effect of YTA7 deletion on histone gene transcription is observed in the highly synchronized α-factor blocking experiment in Figure 3. Second, the Yta7 protein could have an undefined role during M phase. This could be consistent with Yta7 playing some chromosome specific role as it is a member of a histone bound protein complex during this time of the cell cycle (Tackett et al. 2005). However, a previous study showed that Yta7 is not involved in chromosome segregation during M phase (Dubey and Gartenberg 2007). Altogether, we believe that the data presented in Figures 3–5 support Yta7 functioning as a histone gene transcriptional repressor.

Figure 5.—

FACS analysis of DNA content for bar1Δ and bar1Δyta7Δ or wild-type and yta7Δ cells. (A) bar1Δ and bar1Δyta7Δ cells were blocked with α-factor, released, and samples were taken at the indicated times for FACS analysis. (B) Histograms of FACS data from A showing 1N and 2N DNA contents as a percentage of gated cells. (C) Wild-type and yta7Δ cells were blocked with nocodazole, released, and samples were taken at the indicated times for FACS analysis. The asterisks highlight the key differences between the WT and yta7Δ cells. (D) Histograms of FACS data from C showing 1N and 2N DNA contents as a percentage of gated cells.

Yta7 associates with histone loci:

In order to better understand the mechanism by which Yta7 regulates histone gene transcription, we utilized a chromatin immunopurification with high resolution DNA microarray readout strategy (ChIP-chip) to determine whether the Yta7 protein associates with these loci (Ren et al. 2000; Taverna et al. 2006). We chose this technique over conventional ChIP because we were interested in observing the distribution profile of Yta7 across the loci at high resolution, and because uncovering additional binding sites could provide further information for defining the mechanism of Yta7 activity. The ChIP-chip analysis of Yta7-Myc showed significant binding at 177 discrete regions (supplementary Table S1). We did not observe any obvious trend toward the type of genes Yta7 was associating with.

We show the array data as a composite profile of the Yta7-associated DNA occupancy (relative occupancy) plotted as a function of relative coordinates on an average gene flanked with intergenic regions as was previously done in Taverna et al. (2006) (Figure 6A). Figure 6A revealed that on average the Yta7 protein had preferential association with intergenic regions of chromatin as well as with the extreme 5′- and 3′-ends of an ORF. In contrast, a component of the NuA3 histone acetyltransferase complex, Yng1, was recently showed to be preferentially enriched at the 5′ half of an ORF, while minimal association is observed in the intergenic regions (Taverna et al. 2006). The averaged localization of Yng1 is consistent with the role of NuA3 in stimulating transcription via H3K14 acetylation, which is a PTM localized to 5′-ends of actively transcribing genes (Pokholok et al. 2005). Our data in Figure 6A suggest that the mechanism of Yta7 activity may be mediated through non ORF associations and that Yta7 binding to intergenic regions could be a mechanism for transcriptional silencing.

Figure 6.—

The Yta7 protein localizes to histone gene loci. ChIP-chip analysis of Yta7-Myc from asynchronous cells using 50-bp resolution tiled arrays (NimbleGen Systems). (A) Yta7 associated DNA occupancy is plotted as a function of relative coordinates on an average gene flanked with intergenic regions. (B) Shown is Yta7-Myc binding at the four histone loci: HTB1 and HTA1, HHF1 and HHT1, HHT2 and HHF2, and HTA2 and HTB2. The log2 intensity ratio of immunoprecipitated to control DNA is plotted as a function of genomic position. Error bars are the standard deviation of triplicate array analyses.

Six of the top 10 binding sites from the ChIP-chip analysis were pairings of histone genes (HTB1 and HTA1, HHF1 and HHT1, HHT2 and HHF2) (Figure 6B, supplemental Table S1). In addition to these histone loci, the other genomic pair of histone genes (HTA2 and HTB2) was also a significant Yta7 binding site (Figure 6B and supplemental Table S1). Consistent with the average plot (Figure 6A), the HTB1 and HTA1 histone locus showed peak Yta7 association at the intergenic region, even though the Yta7 protein was also binding to the genes themselves (Figure 6B). Yta7 bound the other histone loci HHF1 and HHT1, HHT2 and HHF2, and HTA2 and HTB2 with a similar profile except that peak association occurred within the ORFs. Since the transcription of all histone genes is similarly modified in yta7Δ cells (Figure 3), the exact nature of this slightly different binding profile among the loci remains to be elucidated. Our observed interaction of the Yta7 protein with the histone loci is suggestive of a direct transcriptional regulatory mechanism.

Yta7 interacts with chromatin via the N-terminal tail of histone H3:

The above results show that Yta7 associated with and transcriptionally regulated histone genes. Yta7 has also been shown to immunopurify with histone proteins (Tackett et al. 2005). One chromatin-based regulative pathway that precedes histone gene transcription is the acetylation of histone H3 Lys56 (H3K56ac) at histone loci (Xu et al. 2005). The Yta7 protein also contains a bromodomain-like region that shows histone association, but the PTM specificity has not been determined (Jambunathan et al. 2005). Therefore, we sought to investigate if the Yta7 protein association with histones was PTM dependent, and particularly H3K56ac dependent.

We purified a recombinant version of the Yta7 bromodomain-like region (Yta7bd). We first investigated whether the recombinant version of Yta7bd specifically bound to a particular histone. An Yta7bd-GST fusion or GST alone was incubated with acid-extracted histones, and interacting proteins were isolated on glutathione resin. Histones associated with Yta7bd were visualized by immunoblotting (Figure 7A). GST alone showed minimal nonspecific affinity for the histones when comparing the input (I) and bound (B) samples. However, the GST-Yta7bd fusion protein showed significant binding of histone H3 as well as minor affinity for H2B.

Figure 7.—

Yta7bd interacts with histone H3. (A) Yta7 protein preferentially interacts with histone H3. Yta7bd-GST fusion, or GST alone, was incubated with purified histones and isolated on glutathione resin. Histones associated with Yta7 were visualized by immunoblotting. Input (I) and bound (B) lanes are shown. (B) Yta7bd interacts with the N-terminal tail of histone H3. Recombinant Yta7bd, or GST alone, was incubated with streptavidin Dynabeads coated with biotinylated peptides representing different regions and modification states of histone H3 (methylation and acetylation). Bound Yta7bd was resolved by SDS–PAGE and visualized by Coomassie staining. (C) Quantification of Yta7bd binding from B. Yta7bd binding is shown as a percentage relative to H3 1-20 binding. (D) Full-length Yta7 protein interacts with the N-terminal tail of histone H3. Purified Yta7-PrA was incubated with streptavidin Dynabeads coated with biotinylated peptides representing different regions and modification states of histone H3. Bound Yta7-PrA was resolved by SDS–PAGE and visualized by immunoblotting for the PrA tag. (E) Yta7 associated histone H3 was acetylated at Lys56. Yta7-PrA was immunopurified and coenriching proteins were resolved by SDS–PAGE/Coomassie staining. Copurifying histone H3 was treated in-gel with d6-acetic anhydride to isotopically mark unmodified lysines and then digested with trypsin. Shown is a section of a mass spectrum containing the H3K56 peptide as either acetylated (K56ac) or unacetylated (K56). In accordance to the monoisotopic peak areas, H3K56 was 33% acetylated.

We then investigated if the Yta7bd interaction with histone H3 was directed by H3K56ac (or another histone H3 PTM). Binding of Yta7bd to a series of differentially modified histone H3 peptide mimics was performed (Figure 7B). We observed that Yta7bd did not associate with a peptide corresponding to the region of histone H3 containing H3K56 or H3K56ac (Figure 7B, lanes 13–16). However, Yta7bd showed a preferential interaction with the N-terminal 20-amino-acid tail of histone H3 (Figure 7B, lanes 3 and 4). The interaction of Yta7bd with the N-terminal tail was favored for the unmodified or singly modified peptide as demonstrated by the decreased association observed with the addition of two or three PTMs (Figure 7B, lanes 5–6 and 17–20 vs. lanes 7–12). Addition of H3K4me3 reduced the binding capacity >30%, and further acetylation reduced it ∼50–70%, while single acetylation caused a binding reduction of ∼20–40% (Figure 7C). The ability of Yta7bd to preferentially bind the unmodified histone H3 tail and to tolerate various single PTMs suggests that the Yta7 protein interaction with chromatin is mediated by a general interaction with the N-terminal tail of histone H3. This is not surprising because the bromodomain-like region of Yta7 is missing amino acids that are required for acetyl-lysine specificity in other bromodomains, but does retain many of the amino acids needed for binding to the histone peptide backbone (Jambunathan et al. 2005). These results suggest that the Yta7 protein associates with chromatin through an interaction with the N-terminal tail of histone H3. It is possible that Yta7 binding could be stimulated by combinations of different PTMs. A similar analysis with full-length Yta7 also showed enrichment for the N-terminal tail of histone H3 and no stimulated binding with an H3K56ac peptide (Figure 7D).

Our findings in Figure 7 revealed that Yta7 association with histones was H3K56ac independent. However, we hypothesized that these two histone gene transcriptional regulation signals could simultaneously coexist on chromatin to provide for independent, overlapping transcriptional regulation. To test whether H3K56ac and Yta7 could be found simultaneously on chromatin, we isolated an ex vivo complex of full-length Yta7 with copurifying histones as previously reported (Tackett et al. 2005). We performed a quantitative PTM analysis of histone H3 using isotope labeling and MALDI-mass spectrometry (Tackett et al. 2005). The analysis of the H3 peptides mass spectrum revealed peptides that corresponded to 41% sequence coverage of histone H3 (Table 3). Of particular interest, we observed peptide 53-63, which contained Lys56 (Figure 7E). We observed two peaks that corresponded to the unacetylated and the acetylated peptide. These peptides were confirmed by tandem MS. The peak areas from the mass spectrum were used to determine that the population of histone copurifying with Yta7 was 33% acetylated at H3K56, thus demonstrating that Yta7 could associate with both H3K56 acetylated or unmodified chromatin. Other detected histone H3 peptides did not show significant acetylation (Table 3). These data suggest that Yta7 and H3K56ac can coexist on a given piece of chromatin.

The data in Figure 7 show that Yta7 binding to chromatin is not stimulated by H3K56ac, but the two entities can coexist. We wanted to test whether acetylation of H3K56 was dependent on Yta7. To monitor for this dependence, we utilized ChIP to H3K56ac at the intergenic region between HTA1 and HTB1 (Figure 8A) (Xu et al. 2005). H3K56ac ChIP on wild-type and yta7Δ cells showed no significant difference (P-value = 0.19). Cells harboring a H3K56R mutation were used as a control (Recht et al. 2006). In addition to performing ChIP to this one region of known H3K56ac, we correlated our genomewide Yta7 ChIP-chip data to the reported sites of H3K56ac, using only enrichments ≥2 SD from the mean (Xu et al. 2005). Only 6.8% of the 177 Yta7-binding sites correlated to sites of H3K56ac (Figure 8B). This low correlation is in agreement with results from Figure 7B that showed no direct affinity of Yta7bd for H3K56ac. The results in Figures 7 and 8 suggest that Yta7- and H3K56ac-mediated regulation of histone gene expression are independent.

Figure 8.—

Yta7 does not direct or correlate with general H3K56 acetylation. (A) H3K56ac is not dependent on Yta7. ChIP to histone H3K56ac was performed in wild-type, yta7Δ, and H3K56R cells. Real-time PCR was used to measure enrichment at the intergenic region between HTA1 and HTB1 relative to ACT1. ChIP to unmodified H3 was used to control for nucleosome occupancy. Error bars are the standard deviation of the mean. The asterisk shows the significant difference determined by t-testing (P < 0.01). (B) Genomewide Yta7 binding does not correlate with H3K56ac. Over the 177 sites identified from ChIP-chip analysis of Yta7, only 6.8% were overlapping with previously reported H3K56ac-binding sites (Xu et al. 2005).

DISCUSSION

Chromatin remodeling and epigenetic regulation are tightly linked to histone gene expression. Maintaining the proper dosage level of histone transcripts is a crucial cellular mechanism that is protected through multiple pathways. These histone gene regulatory pathways exhibit various levels of redundancy and crosstalk of protein interactions. Our work describes how Yta7, a protein initially identified as a component of a barrier chromatin complex, is also involved in histone gene expression and is connected to other histone regulatory pathways.

S. cerevisiae contains two copies of the four core histones, which are transcribed during S phase. These copies exist as pairs on chromosomes 2, 4, and 14. Preceding transcriptional activation of the histone genes is a rapid cycle of acetylation on histone H3 lysine 56 (H3K56ac) (Xu et al. 2005). H3K56ac coats the promoter and extends into the open-reading frame regions of these genes in a Spt10-dependent manner, and is needed for proper histone gene transcription (Xu et al. 2005). In addition to H3K56ac, histone gene transcription is regulated by the Asf1/Hir complex and Spt16 containing yFACT complex pathways (Sherwood et al. 1993; Spector et al. 1997; Sutton et al. 2001; Formosa et al. 2002; Recht et al. 2006; Mousson et al. 2007). Asf1/Hir and yFACT complexes appear to function through separate pathways to regulate histone gene transcription, with both being needed for proper S-phase transcription of histone genes. The Asf1/Hir complex functions to maintain cell-cycle-dependent and hydroxyurea-induced repression of histone gene transcription (Sutton et al. 2001). The Spt16 yFACT complex aids in transcriptional initiation and provides for the proper reassembly of chromatin after polymerase passage (Formosa et al. 2002). Mutations of the Asf1/Hir and yFACT transcriptional regulators alter, but do not abolish S-phase histone gene transcription, suggesting that other mechanisms of gene regulation exist. Our work links the Yta7 protein to one of these novel mechanisms of histone gene transcriptional regulation. We favor the hypothesis that Yta7 functions in concurrence with a multi-pathway cellular program that regulates histone gene transcription during the cell cycle. Evidence for our hypothesis is demonstrated at the level of cellular pathway crosstalk, transcriptional control and chromatin association.

An epistasis analysis provided evidence that Yta7 activity genetically interacts with other known cellular pathways that regulate histone gene transcription, namely the Asf1/Hir and Spt16/yFACT pathways (Figure 1). The synthetic nature of the observed growth defects indicated that Yta7, Asf1/Hir, and Spt16/yFACT have overlapping cellular functions. Thus, we identified a functional significance for our observed copurification of Yta7 and Spt16 proteins (Tackett et al. 2005). The synthetic growth defects with yta7Δ cells were detectable only under stressful growing conditions (temperature, DNA synthesis inhibition, and DNA damage), suggesting that Yta7 is linked to mechanisms of DNA repair. Under stress conditions, such as hydroxyurea treatment, cells shut down histone gene transcription (Sutton et al. 2001). Deletions in histone transcriptional regulators, such as Hir1 and Asf1, prevent the proper deactivation of histone gene transcription in the presence of hydroxyurea, thus highlighting the significance of the observed synthetic growth defects with yta7Δ cells under stress (Figure 1) (Sutton et al. 2001). Our MMS studies are in accordance with previous studies showing that Yta7 interacted with Rad53 under DNA damage conditions (Smolka et al. 2005). Moreover, the Yta7 human homolog ATAD2 was down-regulated in a colon cancer cell line arrested in S phase after 5-fluorouracil treatment, suggesting a role in DNA damage and repair (De Angelis et al. 2006). The epistasis analysis revealed that the cellular pathway(s) containing Yta7 were overlapping with known pathways that can modulate histone transcript levels, thereby linking Yta7 to a possible role in histone gene transcription.

A direct test of the involvement of Yta7 in the transcriptional regulation of histone gene expression demonstrated that Yta7 functioned as a transcriptional repressor (Figure 3). Deletion of the YTA7 gene led to elevated transcript levels for all histone genes prior to peak transcription. Histone gene transcription was then downregulated in a similar manner relative to wild-type cells. In this experiment, we observed a similar misregulation of transcription at each gene, which was consistent with Yta7 serving a similar repressive role at each locus (Figure 3). The areas under the peaks of histone gene transcription showed that deletion of the YTA7 gene produced more histone transcripts; however, the peak level was not higher—rather the peak was broader. With the decline of both peaks overlapping, Yta7 appeared to be involved in some early step of transcriptional regulation. The evidence in Figure 3 showing that Yta7 served as a repressor of histone gene transcription is in agreement with the epistasis analysis that linked Yta7 to the histone gene transcriptional regulatory network (Figure 1). As may be expected, double deletion of the repressor Yta7 with hydroxyurea-dependent histone gene repressors (namely Hir1 and Asf1) caused a synthetic growth defect in cells (Figure 1B). This growth defect could be the consequence of potentially toxic levels of free histone proteins as previously suggested (Gunjan and Verreault 2003; Sharp et al. 2005). Our time course transcriptional data did not indicate whether this repressive effect was direct or indirect. It is possible that Yta7 acts in an indirect manner to modulate histone transcript levels. Rather than acting as a direct repressor of gene transcription, Yta7 could play a role in the post-transcriptional turnover of histone mRNA (Hereford et al. 1982). However, since Yta7 has a putative bromodomain, transcriptional gene regulatory activity, and associates with chromatin, we favor a direct mechanism that would involve the Yta7 protein binding to the histone loci (Tackett et al. 2005).

To determine if the Yta7 protein indeed directly associated with the histone loci, we performed a series of in vivo and in vitro binding studies. Genomewide ChIP-chip of Yta7 revealed direct chromatin association at 177 novel sites (Figure 6). Six of the top 10 binding sites were the histone loci (Figure 6B). Yta7 binding was observed across each ORF and intergenic region at the histone loci. It is possible that such a binding profile could provide for formation of a chromatin state that represses transcription. Binding of Yta7 to the histone loci further strengthens the hypothesis that Yta7 is directly regulating histone gene transcription. Since Yta7 contains a histone H3 binding bromodomain-like region (Figure 7A), we investigated whether there was a particular histone H3 PTM code that promoted Yta7 association. Full-length Yta7 and Yta7bd were found to engage histone H3 at the N terminus with a preferential association in the absence of tested PTMs (Figure 7, B–D). Purification of in vivo associated histone H3, revealed significant levels of H3K56ac, but this PTM did not stimulate Yta7 protein association nor was H3K56ac dependent upon Yta7 (Figure 7, C and E; Figure 8). The in vivo purification of H3K56ac with Yta7 would be predicted because both are associated with the intergenic regions at histone loci (Figure 6; Xu et al. 2005). Our chromatin association studies are consistent with the hypothesis that Yta7 mediated transcriptional repression occurs through direct association at the histone loci.

The results presented shed light on the cellular program needed for transcriptional regulation of the histone genes. There appear to be multiple levels of regulation that provide for proper histone dosage, which is not surprising for a key cellular process. We find that Yta7 serves a key role in a previously uncharacterized branch of the histone gene transcriptional regulatory pathway. Our findings also provide for a broader understanding of the function of Yta7, which seems to regulate transcription not only at transition zones between silent and active chromatin, but also via a dynamic mechanism at other chromosomal regions.

Acknowledgments

We thank Lauren Blair and Cagdas Tazearslan for critical reading, Piotr Zimniak for access to real-time PCR instrumentation, Andrew Krutchinsky and Markus Kalkum for assistance with mass spectrometric hardware/software development, Cecile Artaud and Anastas Pashov for flow cytometry analysis, Tim Formosa for the spt16-11 strain, and Judith Recht for the H3K56R strain. The following National Institutes of Health (NIH) grants supported this work: P20RR015569 (A.J.T., co-investigator), GM63959 (C.D.A.), RR022220 and GM076547 (J.D.A.). A.J.T. acknowledges support from the Arkansas Biosciences Institute and recognizes mass spectrometric support from Sam Mackintosh and Chris Warthen funded through an NIH IDeA Network of Biomedical Research Excellence grant (P20RR016460).

Footnotes

  • Communicating editor: F. Winston

  • Received December 28, 2007.
  • Accepted February 22, 2008.

References

View Abstract