|Home | About | Journals | Submit | Contact Us | Français|
Histone N-terminal domains are frequent targets of posttranslational modifications. Multiple acetylated lysine residues have been identified in the N-terminal domain of H2B (K6, K11, K16, K17, K21, and K22), but little is known about how these modifications regulate transcription. We systematically mutated the N-terminal domain of histone H2B, both at known sites of lysine acetylation and elsewhere, and characterized the resulting changes in genome-wide expression in each mutant strain. Our results indicate that known sites of lysine acetylation in this domain are required for gene-specific transcriptional activation. However, the entire H2B N-terminal domain is principally required for the transcriptional repression of a large subset of the yeast genome. We find that the histone H2B repression (HBR) domain, comprised of residues 30 to 37, is necessary and sufficient for this repression. Many of the genes repressed by the HBR domain are located adjacent to telomeres or function in vitamin and carbohydrate metabolism. Deletion of the HBR domain also confers an increased sensitivity to DNA damage by UV irradiation. We mapped the critical residues in the HBR domain required for its repression function. Finally, comparisons of these data with previous studies reveal that a surprising number of genes are coregulated by the N-terminal domains of histone H2B, H3, and H4.
In eukaryotic cells, DNA is packaged with histones and other proteins into chromatin (5). The principal packaging unit is the nucleosome, which consists of approximately 147 bp of DNA wrapped around a protein core of one histone H3-H4 tetramer and two histone H2A-H2B dimers (21). Because of their close association with DNA in the nucleosome, histones play integral roles in DNA-templated processes, such as DNA transcription, replication, and repair. Each of the four histone proteins is covalently modified at multiple residues, principally in their flexible N-terminal domains (9, 15). Histone modifications, such as lysine acetylation, lysine methylation, and serine phosphorylation, profoundly affect transcription initiation by regulating the association of transcriptional regulatory proteins with DNA (2, 10, 30).
While modifications in the N-terminal domains of histone H3 and H4 have been extensively studied, relatively little is known about how N-terminal modifications in histone H2A and H2B regulate transcription. In Saccharomyces cerevisiae, histone H2B is acetylated at six lysine residues in its N-terminal domain (K6, K11, K16, K17, K21, and K22), most likely by the Gcn5 histone acetyltransferase (24, 28). H2B-K16 has been shown to be hypoacetylated in regions of subtelomeric heterochromatin due to the actions of the Hda1 histone deacetylase (23). It is unclear, however, whether there is a functional requirement for histone H2B hypoacetylation in subtelomeric heterochromatin. The histone H2B N-terminal domain does not appear to regulate telomeric silencing (29, 32).
Deletion of amino acids 30 to 37 in the H2B N-terminal domain, which is lethal in some strain backgrounds (19), results in a loss of repression of basal transcription from the GAL1 promoter (11). Mutations in the histone H2B N-terminal domain have shown genetic interactions with the SPT4, SPT5, and SPT6 genes (18) and the SNF5 gene (19), which encodes a component of the SWI/SNF nucleosome-remodeling complex (12). Biochemical experiments have shown that depletion or acetylation of the histone H2A-H2B dimer facilitates transcription in vitro (8, 17).
To discover how acetylated lysines and other residues in the histone H2B N-terminal domain regulate genome-wide transcription, we systematically mutated this domain in S. cerevisiae and characterized the subsequent changes in genome expression using whole-genome oligonucleotide arrays. Our results indicate that acetylated lysine residues in histone H2B are required for the transcriptional activation of a number of genes involved in NAD biosynthesis and vitamin metabolism. In contrast, our results also indicate that a conserved domain located in amino acids 30 to 37 in the histone H2B N-terminal tail is required for the repression of 8.6% of yeast genes, including many genes located in telomeric and subtelomeric heterochromatin. This H2B repression (HBR) domain and the H2B acetylated lysine residues appear to have opposing effects on the transcriptional regulation of a number of genes. We mapped the critical residues in the HBR domain that are required for its repressive function. Finally, we showed that the N-terminal domains of the core histone proteins have overlapping functions in transcriptional regulation.
Both chromosomal copies of the histone H2A and H2B genes were deleted sequentially in the yeast strain Z1256 (20), a strain derived from W303. The HTA2-HTB2 gene pair was replaced by the LEU2 selectable marker using PCR-mediated gene disruption (1). The resulting yeast strain was transformed with plasmid pMP011, and then the HTA1-HTB1 gene pair was replaced with HIS3 to give the strain PY013. Gene deletions were verified by PCR amplification and sequencing. Plasmids containing wild-type or mutant HTB1 alleles were transformed into PY013, and plasmid pMP011 was removed by growth on medium containing 5-fluoroorotic acid. Yeast strains used in this study are listed in Table Table11.
For each genome-wide expression experiment, three mutant and three wild-type yeast cultures were grown in yeast extract-peptone-dextrose (YPD) medium to a final optical density at 600 nm of 0.5 to 0.7 and then harvested as described previously (13).
Plasmid pMP011 was constructed as follows. The wild-type HTA2-HTB2 genes were PCR amplified from yeast genomic DNA (Invitrogen) using primers OLZ13 and OLZ14 and ligated into plasmid pCR2.1-TOPO (TOPO TA cloning kit; Invitrogen) to give plasmid pMP010. pMP010 was digested with KpnI and XhoI (New England Biolabs), and the resulting 2.1-kb fragment containing HTA2-HTB2 was ligated into the pRS316 vector (27) to give plasmid pMP011.
The wild-type HTA1-HTB1 genes were PCR amplified from yeast genomic DNA (Invitrogen) using primers OLZ11 and OLZ12. The resulting 2.1-kb PCR product was digested with EcoRI and SpeI and ligated into the pRS314 vector (27) to give plasmid pMP002.
The HTB1 lysines 6, 11, 16, 17, 21, and 22 were changed to glycine using a two-step strategy. First, site-directed mutagenesis (QuikChange kit; Stratagene) was used to introduce silent mutations in HTB1 in plasmid pMP002 at serine 1 (TCT to AGC) using primers OMP19 and OMP20 and at serine 24 (TCC to AGT) using primers OMP21 and OMP22. These mutations introduced unique AfeI and SpeI restriction sites in HTB1 to give plasmid pMP028. Second, pMP028 was digested with AfeI and SpeI to excise nucleotides 4 to 67 of HTB1. Complementary oligonucleotides (OMP17 and OMP18), encoding HTB1 amino acids 2 to 24, with lysines 6, 11, 16, 17, 21, and 22 mutated to glycine, were annealed and ligated into the digested pMP028 to give plasmid pMP030.
All other HTB1 mutations were generated from plasmid pMP002 by site-directed mutagenesis (QuikChange kit; Stratagene). The complete list of mutagenic primer sequences is available at http://wyrick.sbs.wsu.edu/histoneH2B/. All mutations were confirmed by DNA sequencing.
Total RNA was isolated from each yeast culture and used to prepare cDNA and biotin-cRNA, as described previously (13). The cRNA was hybridized to a single S98 genome microarray and scanned following standard protocols (Affymetrix). Intensities were captured using GeneChip software (Affymetrix), and a single raw expression level for each gene was determined. Complete data sets are available at http://wyrick.sbs.wsu.edu/histoneH2B/.
The data from each chip were normalized using GeneChip software (version 5; Affymetrix), as no global changes in mRNA levels were detected. A modified version of the error model analysis method (22) was used to identify differentially expressed genes in the histone H2B mutant strains. Briefly, a triplicate array error model was used to calculate an X statistic for each gene present on the array. The X statistic was calculated using the following formula:
where a1 and a2 are the average fluorescence intensities of the gene in the three wild-type experiments and the three mutant experiments, respectively, and s1 and s2 are the standard deviations in the fluorescence intensities for the three wild-type experiments and three mutant experiments, respectively. The f factor was empirically chosen so that the X statistic fits a Gaussian distribution (for these experiments, f = 0.27).
In order to estimate the significance (P value) of the change in mRNA levels, the X statistic was fit to a Gaussian distribution. To estimate the mean and variance of this Gaussian distribution independent of the number of differentially expressed genes in the particular experiment, we conducted a similar error model analysis on all possible permutations of the three wild-type and three mutant data sets. The average variance (σ) and mean (μ) of the X values across all permutations were used to calculate the significance (P value) of the change in mRNA levels, using the following equation:
where erf is the error function (i.e., cumulative normal distribution). The modified triple array error model analysis was implemented as a macro in Excel (Microsoft). A P value cutoff of 0.001 was used to identify differentially expressed genes.
Genes within 40 kb of a telomere were pooled and ordered according to their distance from a telomere. A sliding 50-gene window was moved along the ordered gene list in 10-gene steps. The fraction of genes up-regulated in each mutant and the average distance from the telomere were plotted for each 50-gene window. The statistical significance of the enrichment of up-regulated genes in telomere-proximal (0 to 10 kb from a telomere end) and subtelomeric (10 to 20 kb from a telomere end) regions for each data set was calculated using the hypergeometric probability distribution.
A hypergeometric probability model was used to calculate the statistical significance of the overlap between the lists of up- and down-regulated genes in the various histone mutants. The resulting P values were numerically calculated using the Mathematica software package (Wolfram Research).
Significant functional enrichments were identified in the lists of up- and down-regulated genes using the Saccharomyces Genome Database Gene Ontology Term Finder tool at http://www.yeastgenome.org.
Yeast growth conditions and RNA isolations were performed as described above. RNA was treated with RNase-free DNase I (Roche) to remove genomic DNA contamination. cDNA was synthesized using a cDNA synthesis kit (Roche) according to the manufacturer's instructions. Synthesized cDNA (2 μl) was amplified by multiplex PCR using primers OMP87 and OMP88 to amplify a 483-bp fragment of GCY1 and primers OMP94 and OMP95 to amplify an 837-bp fragment of ACT1 (for primer sequences, see http://wyrick.sbs.wsu.edu/histoneH2B/). The following PCR cycling conditions were used: 30 s at 94°C, 1 min at 47°C, and 1 min at 72°C with a 10-min extension at 72°C after 30 cycles. PCR products were electrophoresed on a 1% agarose gel and visualized after ethidium bromide staining. Pictures of gels were taken using a GelDoc EQ imager (Bio-Rad), and band intensities were quantified using Quantity One software (Bio-Rad). GCY1 expression levels were normalized using the ACT1 intensity as a loading control, since ACT1 expression was not significantly altered in any of the histone H2B mutants.
Triplicate cultures of mutant and wild-type strains were grown in YPD medium to mid-log phase and then serially diluted and plated (from 105 to 102 cells per plate). Plates were treated with 0, 50, or 100 J/m2 of UV light (primarily 254 nm). Following treatment, plates were incubated for 2 days at 30°C in the dark. Colonies were counted and compared to untreated (0 J/m2) controls.
The data discussed in this publication have been deposited in NCBI's Gene Expression Omnibus (GEO; http://www.ncbi.nlm.nih.gov/geo/) and are accessible through GEO Series accession numbers GSE3802, GSE3803, GSE3804, GSE3805, and GSE3806.
To investigate the contributions of the histone H2B N-terminal domain in transcription regulation, we constructed mutations or deletions of key residues in this domain and characterized the resulting changes in global mRNA levels using whole-genome Affymetrix oligonucleotide arrays. For each microarray experiment, triplicate mRNA samples from mutant and wild-type strains were profiled. The microarray data were analyzed using a modified triple array error model (see Materials and Methods) to calculate the significance (P value) of the observed change in mRNA levels. Genes with a P value of less than 0.001 were deemed to be differentially expressed. Examples of the error model analysis are shown in Fig. Fig.1.1. As a control, we used this method to analyze triplicate sets of microarray data from identical wild-type strains. As shown in Fig. Fig.1A,1A, only 6 out of 6,063 genes were significantly changed (P < 0.001) in the wild type-to-wild type comparisons, indicating a low false positive rate.
To investigate the role of acetylated lysine residues in the N-terminal domain of histone H2B in transcription regulation, we profiled the genome-wide expression levels in a histone H2B mutant in which K6, K11, K16, K17, K21, and K22 were mutated to glycine. Analysis of triplicate samples revealed that, out of 6,063 genes measured, the mRNA levels of 2 genes were up-regulated and the mRNA levels of 55 genes were down-regulated in the H2B K6,11,16,17,21,22G mutant relative to the wild type (Fig. (Fig.1C).1C). The list of down-regulated genes included many genes involved in NAD biosynthesis (P = 2.6 ×10−7) and vitamin metabolism (P = 1.4 × 10−6), including the genes BNA1, BNA2, BNA4, BNA5, THI2, THI11, and SNO1. These results indicate that in standard growth conditions the acetylated lysine residues in the H2B N-terminal domain are required for the transcriptional activation of a small set of genes, particularly those that function in NAD biosynthesis and vitamin metabolism.
Curiously, deletion of the H2B N-terminal domain did not recapitulate this transcriptional activation defect. The H2B Δ3-32 deletion eliminates most of the H2B N-terminal domain, including all six of the known acetylated lysine residues. Analysis of triplicate samples revealed that the mRNA levels of 17 genes were up-regulated and the mRNA levels of 14 genes were down-regulated in the histone H2B Δ3-32 mutant relative to the wild type (Fig. (Fig.1C).1C). Comparison of the H2B K6,11,16,17,21,22G and Δ3-32 data sets indicates that three of the genes down-regulated in the H2B K6,11,16,17,21,22G mutant (BNA4, SNO1, and DAL1) are up-regulated in the H2B Δ3-32 mutant. These data suggest that other residues in the H2B N-terminal domain have an opposite (i.e., repressive) effect on transcription and that the deletion of these residues compensated for the loss of the acetylated lysine residues in the H2B Δ3-32 mutant.
To investigate this hypothesis further, we examined the functional role of the entire histone H2B N-terminal domain by using oligonucleotide arrays to profile the genome-wide expression changes in an H2B Δ3-37 mutant strain. Surprisingly, the data show that the mRNA levels of 650 genes were up-regulated and the mRNA levels of 68 genes were down-regulated in the H2B Δ3-37 mutant relative to the wild type (Fig. (Fig.1C).1C). This result indicates that, as a whole, the histone H2B N-terminal domain functions primarily to repress transcription and regulates the transcription of 11.9% of the yeast genome. It is notable that the deletion of five additional residues (amino acids 33 to 37) in the histone H2B N-terminal domain resulted in the derepression of almost 2 orders of magnitude more genes than was observed in the H2B Δ3-32 mutant.
Residues 30 to 37 in the histone H2B N-terminal domain have been previously shown to repress the basal transcription of the URA3 reporter gene when fused to the GAL1 promoter (11), suggesting that this domain may have a general role in transcriptional repression. To test whether the deletion of this subdomain is responsible for the large number of up-regulated genes in the histone H2B Δ3-37 mutant, we profiled the genome-wide expression changes in an H2B Δ30-37 mutant strain. Analysis of the data indicates that the mRNA levels of 521 genes were up-regulated and the mRNA levels of 81 genes were down-regulated in the H2B Δ30-37 mutant relative to the wild type (Fig. (Fig.1C).1C). Figure Figure2A2A shows that 79% of the genes up-regulated in the H2B Δ30-37 mutant were also up-regulated in the H2B Δ3-37 mutant, a highly significant overlap (P = 4.5 × 10−354). Figure Figure2B2B shows that 43% of the down-regulated genes in the H2B Δ3-37 mutant were also down-regulated in the H2B Δ30-37 mutant (P = 1.5 × 10−38).
We also investigated whether the H2B Δ3-37 and H2B Δ30-37 mutations caused similar magnitudes of change in mRNA levels. Scatterplot analysis was used to compare the changes in mRNA levels in the two mutant strains for all yeast genes (Fig. (Fig.2C).2C). Examination of the scatterplot shows that for most genes the H2B Δ3-37 and H2B Δ30-37 mutations caused similar changes in mRNA levels, independent of the analysis method used. As a control, we also compared the change (n-fold) data of the H2B Δ3-37 mutant to those of the H2B Δ3-32 mutant, as these two mutations have very different effects on genome-wide expression according to the error model criteria. As shown in Fig. Fig.2D,2D, the magnitude in the change (n-fold) in mRNA levels for most genes is clearly dissimilar in the H2B Δ3-37 and H2B Δ3-32 mutants. In summary, we find that similar numbers of genes were up- and down-regulated in the H2B Δ3-37 and H2B Δ30-37 mutants, most of these genes were shared, and the mRNA levels of these genes were altered to similar degrees between the mutant data sets.
These data indicate that the subdomain located at residues 30 to 37 in histone H2B is necessary for the transcriptional repression of 8.6% of the yeast genome. The list of up-regulated genes included many genes involved in vitamin metabolism (23 genes; P = 1.0 × 10−8) and carbohydrate metabolism (38 genes; P = 9.5 × 10−7); up-regulated genes in these functional categories are listed in Table Table2.2. We refer to this subdomain as the histone H2B repression (HBR) domain.
A cursory examination of the H2B mutant data sets indicates that a number of genes, particularly those that function in NAD biosynthesis and vitamin metabolism (e.g., BNA1, BNA2, BNA4, THI2, THI11, and SNO1), are down-regulated in the H2B K6,11,16,17,21,22G mutant but up-regulated in the H2B Δ30-37 (or Δ3-37) mutant (see Table S1 in the supplemental material). Figure Figure3A3A shows that 36% of the down-regulated genes in the H2B K6,11,16,17,21,22G mutant are up-regulated in the H2B Δ30-37 mutant, a significant overlap (P = 1.0 × 10−8). In contrast, significant overlaps were not observed when the sets of genes up-regulated (or down-regulated) in both pairs of mutants were compared (P > 0.05). A similar result is obtained if we compare the H2B K6,11,16,17,21,22G and H2B Δ3-37 data sets. Figure Figure3B3B shows that 47% of the down-regulated genes in the H2B K6,11,16,17,21,22G mutant are up-regulated in the H2B Δ3-37 mutant (P = 6.6 × 10−12). These data suggest that the transcriptional activation function of the H2B acetylated lysine residues is opposed by the repression function of the HBR domain. Analysis of the H2B Δ3-37 mutation, which eliminates the six acetylated lysine residues as well as the HBR domain, indicates that the repression function of the HBR domain outweighs or is epistatic to the activation function of the H2B acetylated lysine residues.
Chromosome display (33) of the H2B Δ30-37 genome-wide expression data revealed that many of the up-regulated genes clustered adjacent to yeast telomeres (data not shown). Twenty-seven percent of genes located in telomere-proximal regions (defined as a location within 10 kb of a telomere end) showed increased mRNA levels in the H2B Δ30-37 strain compared to a genome-wide average of 8.6% (P = 2.5 × 10−8). A statistically significant enrichment of up-regulated genes (18%; P = 9.6 × 10−4) was also observed in subtelomeric heterochromatin regions, which have been previously defined as a location 10 to 20 kb from a telomere end (13, 33).
In Fig. Fig.4,4, the fraction of genes that showed increased mRNA levels in the H2B Δ30-37 mutant strain was plotted versus distance from a telomere end. This analysis indicates that the HBR domain is required for the repression of a large fraction of the genes located 5 to 15 kb from a telomere end. A similar analysis of the H2B Δ3-32 data demonstrates that this region of the H2B N-terminal domain is not required for telomere-proximal gene repression (Fig. (Fig.4).4). In summary, the HBR domain regulates the silencing of telomere-proximal and subtelomeric genes.
Previous studies have shown that chromatin structure and histone modifications play vital roles in the damage and repair of DNA lesions (7, 31). We tested whether any of the histone H2B N-terminal deletion mutants exhibited an enhanced sensitivity to UV-induced DNA damage. Each histone H2B mutant strain was treated with various doses of UV light (0, 50, and 100 J/m2), and the number of colonies formed by surviving cells was scored on YPD plates. As shown in Fig. Fig.5,5, the yeast strains lacking the HBR domain (H2B Δ3-37 and Δ30-37) exhibited a ~10-fold-lower survival rate than the wild-type control when irradiated with a dose of 50 J/m2 and a ~100-fold-lower survival rate when irradiated with a dose of 100 J/m2. In contrast, the H2B Δ3-32 mutant strain exhibited a UV sensitivity similar to that observed for the wild-type control. In summary, these results indicate that deletion of the HBR domain in histone H2B leads to enhanced sensitivity to UV-induced DNA lesions.
To confirm the results obtained from the analysis of the array data, we used RT-PCR to examine the effects of the histone H2B mutations on the mRNA levels of the gene GCY1. The GCY1 gene was chosen because its expression is strongly up-regulated in the HBR deletion mutants, it is classified in the enriched carbohydrate metabolism functional category (see Table Table2),2), and its mRNA was readily detected by RT-PCR. In the array data, the mRNA levels of GCY1 are up-regulated 6.8-fold in the H2B Δ3-37 mutant (P = 3.3 × 10−10) and 5.4-fold in the H2B Δ30-37 mutant (P = 1.8 × 10−7) but are unchanged in the H2B Δ3-32 and K6,11,16,17,21,22G mutants (1.1- and 0.92-fold changes, respectively).
Figure Figure66 shows the results of RT-PCR analysis of GCY1 mRNA in these mutant strains. A significant increase in the mRNA levels of GCY1 is detected in the H2B Δ3-37 and Δ30-37 mutants relative to the wild type (8.5- and 8.7-fold changes, respectively), while there is not a significant change in GCY1 expression in the H2B Δ3-32 mutant (1.53-fold change) or the H2B K6,11,16,17,21,22G mutant (0.93-fold change). Hence, the RT-PCR data confirm the results obtained from the array analysis.
To determine the key residues in the HBR domain required for the repression of GCY1, we constructed a series of smaller deletions of the HBR domain and examined the resulting changes in the expression of GCY1 by RT-PCR. Figure Figure7A7A shows the subsequent changes in GCY1 mRNA levels for a series of progressively shorter histone H2B deletions moving in from the N terminus of the HBR domain (Fig. (Fig.7C).7C). The H2B Δ31-37 mutation, which restores a single lysine residue (K30) to the HBR domain, leads to a 2.4-fold reduction in GCY1 mRNA levels compared to the Δ30-37 mutation (P < 0.05). The GCY1 mRNA levels in the Δ32-37, Δ33-37, and Δ34-37 mutants are not significantly different from those in the Δ31-37 mutant (P > 0.05). However, the addition of K34 in the H2B Δ35-37 mutant leads to a significant decrease in GCY1 expression compared to either the Δ31-37 or Δ34-37 mutant (P < 0.05). Finally, the addition of A35 in the H2B Δ36-37 mutant leads to an additional small decrease in the GCY1 mRNA levels compared to the Δ35-37 mutant. The GCY1 mRNA levels in the H2B Δ36-37 mutant are indistinguishable from wild-type levels.
A similar trend was observed when testing a series of progressively shorter H2B deletions from the C terminus of the HBR domain (Fig. (Fig.7D).7D). Figure Figure7B7B shows the results of RT-PCR analysis of GCY1 expression levels in these H2B mutant strains. Note that the same RT-PCR data set was used for the H2B Δ31-37 and H2B Δ30-36 mutants (these mutants have identical protein sequences), so as previously observed, the Δ30-36 mutant leads to a significant decrease in GCY1 expression compared to the Δ30-37 mutant. The GCY1 expression level is not altered significantly further in the Δ30-35 and Δ30-34 mutants (P > 0.05). However, the addition of K34 in the H2B Δ30-33 mutant leads to a significant decrease in GCY1 expression compared to the Δ30-36 mutant (P < 0.05). The mRNA levels of GCY1 in the H2B Δ30-33 mutant are indistinguishable from wild-type levels. The H2B Δ30-34 mutant, which adds A35 to the HBR domain, results in a GCY1 expression level that is midway between the expression levels observed in the H2B Δ30-33 and Δ30-35 mutants; however, these mRNA changes are not statistically significant.
In summary, systematic deletions of the HBR domain identified at least two critical residues required for the repression of GCY1 transcription: (i) a flanking lysine residue (either K30 or K37) and (ii) K34. These data suggest that A35 may also play a role in GCY1 transcriptional repression.
Previous studies have characterized the genome-wide expression changes due to mutations in the N-terminal domains of histone H3 (13, 25) or histone H4 (4, 25, 26). In Fig. Fig.8A,8A, we use Venn diagrams to compare the list of genes up-regulated in a histone H2B Δ3-37 mutant with the lists of genes up-regulated in a histone H4 N-terminal deletion (25) and H3 lysine mutant (13). Figure Figure8A8A shows that 71% of the genes up-regulated in a histone H3 K4,9,14,18,23,27G mutant are also up-regulated in the histone H2B N-terminal deletion mutant (P = 5.8 × 10−85). The coregulated genes included many genes that function in vitamin metabolism (P = 1.0 × 10−8) or carbohydrate metabolism (P = 8.0 × 10−6) or that are located in telomeric (P = 0.00142) or subtelomeric (P = 3.2 × 10−5) regions. A significant overlap was also apparent between the genes up-regulated in the histone H4 Δ2-26 mutant and those up-regulated in the histone H2B N-terminal deletion mutant (P = 2.5 × 10−44). Similar overlaps were observed when the set of genes up-regulated by the H2B Δ30-37 mutant was compared to the H4 Δ2-26 and H3 K4,9,14,18,23,27G mutant data sets (Fig. (Fig.8B).8B). Finally, we also observed a significant overlap between the genes down-regulated in a histone H2A.Z (HTZ1) deletion strain (14) and those up-regulated in the H2B Δ3-37 or Δ30-37 mutant (data not shown).
Previous studies have suggested that the N-terminal domain of histone H2B regulates transcription, but little is known about which genes are regulated by this domain and which residues in this domain are required for this regulation. In the present study, we systematically mutated the H2B N-terminal domain in S. cerevisiae and characterized the subsequent changes in genome expression using whole-genome oligonucleotide arrays. First, we showed that acetylated lysine residues in histone H2B are required for the transcriptional activation of a small set of genes, particularly those that function in NAD biosynthesis or vitamin metabolism. Second, we found that, as a whole, the H2B N-terminal domain functions to repress the transcription of a relatively large subset of the yeast genome, including many genes that are activated by the H2B acetylated lysine residues. Third, this transcriptional repression function maps to the HBR domain, which is comprised of residues 30 to 37 in the histone H2B N-terminal tail. Deletion of the HBR domain leads to the induction of genes located in telomere-proximal heterochromatin and confers a UV sensitivity phenotype. Fourth, we mapped which residues in the HBR domain are required for the transcriptional repression of GCY1. Finally, we found that many of the genes repressed by the HBR domain are also regulated by the N-terminal domains of histone H3 and H4. Taken together, these findings define the functional roles of subdomains in the H2B N-terminal tail in genome-wide transcription.
Previous studies have identified six acetylated lysine residues (K6, K11, K16, K17, K21, and K22) in the yeast histone H2B N-terminal domain. However, relatively little is known about the role of these acetylated lysine residues in transcriptional regulation. Our results indicate that, under standard growth conditions, the acetylated lysine residues in histone H2B are required for the transcriptional activation of 55 genes, many of which function in NAD biosynthesis and vitamin metabolism. Our data do not distinguish whether the transcriptional activation of these genes is dependent on all six acetylated lysine residues (H2B K6, K11, K16, K17, K21, and K22) or a subset of them. Intriguingly, we find that many of the genes activated by the H2B acetylated lysine residues are repressed by the HBR domain. It is possible that acetylation of the H2B N-terminal lysine residues may act to partially overcome repression by the HBR domain.
Our results indicate that deleting the HBR domain in histone H2B (residues 30 to 37) alters the expression of ~10% of the yeast genome. Most of the affected genes are up-regulated, indicating that the primary function of the HBR domain is to repress gene expression. A very similar result is observed when the entire H2B N-terminal domain is deleted (Δ3-37), indicating that the key H2B N-terminal residues required for this transcriptional repression reside in the HBR domain. Previous studies have shown that this domain is required for the basal repression of the GAL1 promoter (11). Our data demonstrate that the HBR domain is required for the repression of 8.6% of the yeast genome, including genes that are involved in vitamin and carbohydrate metabolism or that are located in telomeric and subtelomeric regions of yeast chromosomes. In comparison, a general depletion of nucleosomes leads to the up-regulation of 15% of the yeast genome (33), suggesting that the HBR domain plays a relatively large role in transcription repression.
While our data indicate that the HBR domain plays an important role in transcriptional repression, the mechanism(s) by which the HBR domain regulates these genes is unknown. One possible model is that deletion of the HBR domain destabilizes the histone H2B protein or the H2A-H2B dimer, a possibility supported by biophysical studies (16). This model would predict that deletion of the HBR domain would cause a reduction in the levels of histone H2B (and H2A) protein in yeast chromatin, leading to the induction of gene transcription. We have tested this model by performing Western blot analysis using anti-histone H2B antibodies on chromatin isolated from the various histone H2B mutant and wild-type strains. Our results indicate that the levels of histone H2B protein are not significantly altered in any of the histone H2B mutants profiled in this paper (see Fig. S1 in the supplemental material).
Previous studies have shown that the H2B N-terminal tail may affect the folding of the DNA around the nucleosome particle (11). In this model, mutations in the HBR domain could disrupt repressive chromatin structures by altering the proper DNA folding in the nucleosome, leading to an increase in gene transcription. In support of this model, a high-resolution crystal structure of the Xenopus nucleosome shows that lysine and arginine residues in the Xenopus HBR domain (amino acids 24 to 31) interact with the DNA minor groove (3). However, previous studies have shown that the chromatin structure and nucleosome positioning of the endogenous GAL1 promoter (which is repressed by the HBR domain ) are not significantly altered in the histone H2B N-terminal deletion strains (11). Yet it is possible that the HBR mutations have a subtle effect on nucleosome positioning or dynamics, which would not be detected in this assay.
A third model is that the HBR deletion may eliminate sites of undiscovered posttranslational modifications which function to repress transcription. The HBR domain shows considerable sequence conservation from yeast to humans (Fig. (Fig.9).9). While the HBR domain contains an abundance of potential sites of acetylation, methylation, and ubiquitination, no known histone modifications have been identified in this domain. One report indicated that there may be a potential methylation site at lysine-34 in bovine histone H2B (equivalent to lysine-37 in yeast H2B) (34); however, this result has yet to be confirmed.
Deletion of the HBR domain also confers a UV sensitivity phenotype. While this phenotype is not as sensitive as one caused by mutations in the nucleotide excision repair pathway, its moderate sensitivity suggests that the HBR domain may be required for the repair of or recovery from DNA lesions caused by UV irradiation. Deletion of the HBR domain does not lead to the down-regulation of any known DNA repair genes (data not shown), suggesting that the loss of the HBR domain may directly affect the repair of UV-induced DNA lesions. Future studies will investigate this possibility.
A systematic deletion strategy was used to map the residues in the HBR domain responsible for the transcriptional repression of yeast genes. For each progressive HBR deletion mutant, the corresponding changes in the expression of GCY1, which is repressed by the intact HBR domain, were assayed by RT-PCR. The first conclusion from this experiment was that the retention of a single lysine residue from the HBR domain significantly restored the transcriptional repression of the GCY1 gene. While these data do not distinguish which lysine residue in the HBR domain (K30, K37, or another residue) is responsible for this repression, these data do indicate that the HBR lysine residues appear to be functionally redundant. Second, the addition of K34 also resulted in a significant increase in the repression of GCY1, leading to nearly wild-type GCY1 expression levels. These critical lysine residues could facilitate transcriptional repression by stabilizing the interaction of histone H2B with DNA, as suggested by the crystal structure of the nucleosome, or may be targets of repressive posttranslational modifications. K34 is of particular interest, since its surrounding sequence is reminiscent of other methylated and acetylated lysine residues (6).
In summary, this study has defined two regulatory regions in the histone H2B N-terminal tail: (i) the acetylated lysine residues, which are required for gene-specific transcriptional activation, and (ii) the HBR domain, which has a fairly general role in transcriptional repression. We have also identified the target genes regulated by these domains and mapped the critical residues in the HBR domain required for its repression function. Finally, we have shown that the N-terminal tails of histone proteins have overlapping and/or synergistic functions in repressing gene transcription.
We are grateful to Bill Davis, Lisa Gloss, Ray Reeves, Amy Rodriguez, and Michael Smerdon for helpful discussions and comments on the manuscript. We thank Richard Young for the generous gift of the yeast strain Z1256. We thank Jason Sikes for software and web support. We thank Julie Stanton and Yi Jin for performing the histone H2B Western blots.
This work was supported by American Cancer Society grant RSG-03-181-01-GMC. D.F. was supported by NIH grant ES04106 from the National Institute of Environmental Health Sciences (NIEHS) to Michael J. Smerdon.
†Supplemental material for this article may be found at http://mcb.asm.org/.