Functional Analysis of H2B-Lys-123 Ubiquitination in Regulation of H3-Lys-4 Methylation and Recruitment of RNA Polymerase II at the Coding Sequences of Several Active Genes in Vivo*

Previous biochemical studies have demonstrated that Lys-123 ubiquitination of histone H2B is globally required for up-regulation of mono-, di, and trimethylation of Lys-4 of histone H3. However, recent studies have implicated H2B-Lys-123 ubiquitination in the regulation of di- and trimethylation, but not monomethylation, of H3-Lys-4 in vivo. Using a formaldehyde-based cross-linking and chromatin immunoprecipitation assay, we show that H3-Lys-4 trimethylation, but not dimethylation, is up-regulated by H2B-Lys-123 ubiquitination in vivo at the coding sequences of a set of transcriptionally active genes such as ADH1, PHO84, and PYK1. Both the ubiquitination of H2B-Lys-123 and the methylation of H3-Lys-4 are dispensable for recruitment of RNA polymerase II to the coding sequences of these genes, and hence, their transcription is not altered in the absence of these covalent modifications. However, recruitment of RNA polymerase II to the coding sequence of a galactose-inducible gene, GAL1, is significantly reduced in the absence of H2B-Lys-123 ubiquitination but not H3-Lys-4 methylation. Consistently, transcription of GAL1 is altered in the H2B-K123R point mutant strain. Finally, we show that H3-Lys-4 methylation does not regulate H3-Lys-9/14 acetylation. Collectively, our data reveal a “trans-tail” regulation of H3-Lys-4 tri- but not dimethylation by H2B-Lys-123 ubiquitination, and these modifications are dispensable for transcription of a certain set of genes in vivo.

The eukaryotic genome is packaged into chromatin that is an array of nucleosomes; in each nucleosomes, 146 bp of DNA is wrapped around an octamer of core histone proteins, H2A, H2B, H3, and H4 (1). Chromatin is a dynamic structure that modulates the access of regulatory factors to the genetic material. Thus, transcription and other cellular processes that require access to DNA are regulated by chromatin structure, and therefore, a precise coordination and organization of events in opening and closing the chromatin is crucial for these cellular processes to happen normally.
Post-translational modifications (e.g. acetylation, phosphorylation, ubiquitination, and methylation) of the histones affect chromatin structure directly by altering DNA-histone interactions within and between nucleosomes, thus changing higher order chromatin structure (2). Histone modifications are, therefore, the key determinants in defining the active and repressed states of the chromatin (3)(4)(5)(6)(7)(8). For example, methylation of H3-Lys-4 is correlated with gene activation, whereas H3-Lys-9 methylation results in repression and heterochromatin formation. Furthermore, histone acetylation is generally correlated with gene activation, although there are exceptions to this rule.
Histone methylation on specific Lys or Arg residues is carried out by CARM1, PRMT1, and SET domain-containing enzymes (9,10). The CARM1 and PRMT1 mediate Arg methylation, whereas Lys is methylated by the SET domain-containing enzymes (9,10). In Saccharomyces cerevisiae, Arg methylation of histones has not been demonstrated, but the N-terminal tail of histone H3 is methylated at Lys-4 and Lys-36. Lys methylation of histones in vivo occurs in three states: mono-, di, and trimethyl. Set1p, a component of the COMPASS complex, is involved in mono-, di-, and trimethylation of Lys-4 of H3 (11)(12)(13)(14). H3-Lys-4 methylation is important for rDNA silencing (11,15) and gene activation (8,16). On the other hand, H3-Lys-36 methylation that is mediated by Set2p is involved in transcriptional repression (17). Recently, S. cerevisiae Dot1p (a non-SET domain protein) and the related human protein have been identified as histone methylase that specifically methylates Lys-79 within the globular domain of histone H3 (18 -21). The loss of Dot1p or point mutation of Lys-79 weakens the interaction of Sir2p and Sir3p with the telomeric region in vivo (20,21), and hence, impairs telomeric silencing (22).
Recently, it has been shown that H2B-Lys-123 in yeast gets mono-ubiquitinated (23) by Rad6p (an ubiquitin carrier protein (E2) ubiquitin-conjugating enzyme) and Bre1p (an ubiquitinprotein isopeptide ligase (E3) ubiquitin ligase). Mono-ubiquitination of H2B-Lys-123 is an important prerequisite for global mono-, di, and trimethylation of Lys-4 and Lys-79 of H3 in a "trans-tail" mode of cross-talk (24 -28). Such a regulatory link between histone ubiquitination and methylation is provided by the proteasomal ATPases Rpt6p and Rpt4p because mutation in the respective genes disrupts Lys-4 and Lys-79 methylation of H3 but not H2B ubiquitination (29). Consistently, the proteasome components are recruited to the active gene via H2B ubiquitination and reconfigure chromatin for access of Set1p and Dot1p during transcription (29).
Although H2B-Lys-123 ubiquitination is found to globally up-regulate H3-Lys-4 mono-, di, and trimethylation (24 -28), it was not known whether such a regulatory link is present at the active gene loci in vivo. Recent studies have demonstrated a trans-tail mode of cross-talk between H2B-Lys-123 ubiquitination and H3-Lys-4 di-and trimethylation, but not monomethylation, in vivo (30 -33). We have extended these studies to several active genes in S. cerevisiae to determine whether such a regulatory link exists at their chromosomal loci, using a formaldehyde-based in vivo cross-linking and chromatin immunoprecipitation (ChIP) 3 assay. Our results reveal that H2B-Lys-123 ubiquitination regulates H3-Lys-4 trimethylation, but not dimethylation, at the coding sequences of the PHO84, ADH1, and PYK1 genes in vivo. Furthermore, we show that both H2B-Lys-123 ubiquitination and H3-Lys-4 methylation are dispensable for recruitment of RNA polymerase II to the coding sequences of these genes in vivo.

EXPERIMENTAL PROCEDURES
Plasmids-The plasmid pRS416 (34) was used in the PCRbased gene disruption.
For the studies at the GAL1 gene, the yeast strains were first grown in YPR (yeast extract containing peptone plus 2% raffinose) to an A 600 of 0.9 and then transferred to YPG (yeast extract containing peptone plus 2% galactose) for 90 min at 30°C prior to formaldehyde cross-linking. The yeast strains were grown in YPD (yeast extract containing peptone plus 2% dextrose) to an A 600 of 1.0 at 30°C for the studies at the PHO84, ADH1, and PYK1 genes.
ChIP Assay-The ChIP assay was performed as described previously (36,37). Briefly, yeast cells were treated with 1% formaldehyde, collected, and resuspended in lysis buffer. Following sonication, cell lysates (800 l of lysate from 100 ml of yeast culture) were precleared by centrifugation, and then 100 l of lysate was used for each immunoprecipitation. Immunoprecipitated (IP) protein-DNA complexes were treated with protease K, the cross-links were reversed, and the DNA was purified. The IP DNA was dissolved in 20 l of TE 8.0 (10 mM Tris-HCl, pH 8.0, and 1 mM EDTA), and 1 l of IP DNA was analyzed by PCR. The PCR reactions (a total of 23 cycles) contained [␣-32 P]dATP (2.5 Ci for each 25-l reaction), and the PCR products were detected by autoradiography after separation on a 6% polyacrylamide gel. As a control, "input" DNA was isolated from 5 l of lysate without going through the immunoprecipitation step and suspended in 100 l of TE 8.0. To compare the PCR signal arising from the IP DNA with that from the input DNA, 1 l of input DNA was used for PCR analysis. Serial dilutions of input and IP DNA were used to assess the linear range of PCR amplification as described previously (38). The PCR data presented in this report are within the linear range of PCR analysis.
All the ChIP experiments were repeated at least three times, and consistent results were obtained. Primer pairs used for PCR analysis were as follows: Autoradiograms were scanned and quantitated by the NIH Image 1.62 program. IP DNAs were quantitated and presented as the ratio of IP to input (%IP). For mutant strains, DNA immunoprecipitated relative to wild-type (WT) was presented as %WT.
Chromatin Double Immunoprecipitation (ChDIP) Assay-To determine the levels of H2B-Lys-123 ubiquitination at the active gene loci, we performed a ChDIP assay as described previously (30,33). Briefly, 400 l of lysate from 50 ml of yeast culture was first immunoprecipitated using anti-FLAG antibody (Sigma) and protein A/G plus agarose beads. Following elution of anti-FLAG IP with FLAG peptide (Sigma), the eluate was immunoprecipitated using an anti-HA antibody, and the IP DNA sample was dissolved in 10 l of TE 8.0, of which 1 l was used for PCR analysis. The input DNA was isolated from 5 l of lysate and suspended in 100 l of TE 8.0, of which 1 l was used for PCR analysis.

Whole Cell Extract (WCE) Preparation and Western Blot Analysis-
For global biochemical analysis of H3-Lys-9/14 acetylation in the SET1 deletion mutant and its isogenic wild-type equivalent, the yeast strains were grown in 100 ml of YPD up to an A 600 of 1.0, and then harvested cells were lysed and sonicated to prepare the WCE with solubilized chromatin following the protocol as described previously for the ChIP assay (36,37). The WCE was run on SDS-polyacrylamide gel and then analyzed by Western blot. The anti-Lys-9/14 di-acetylated H3 (Upstate Biotechnology, Inc.) and anti-H3 (Abcam, Inc.) antibodies were used for Western blot analysis.
To address this question, we analyzed the levels of H3-Lys-4 diand trimethylation at a subset of the active genes such as PHO84, ADH1, and PYK1 in the H2B-K123R mutant (where H2B ubiquitination does not occur) and its isogenic wild-type strains, using a ChIP assay in S. cerevisiae. We confined our ChIP experiments to the coding sequences or open reading frames (ORFs) of these genes since previous studies (30,33,(41)(42)(43)(44)(45)(46) have demonstrated that the nucleosomes at the coding sequences are both methylated at H3-Lys-4 and ubiquitinated at H2B-Lys-123. Two sets of ORF-specific primer pairs were used that, as shown in Fig. 1A, could distinguish binding to the 5Ј-end (ORF1) and 3Ј-end (ORF2) of the coding sequence. The average size of the PCR products at the ORF1 and ORF2 of the above genes was confirmed to be around 150 bp by agarose gel electrophoresis. Fig. 1B shows the size FIGURE 1. Analysis of H3-Lys-4 di-and trimethylation and H2B-Lys-123 ubiquitination at the coding sequences of the PHO84, ADH1, and PYK1 genes. A, the PCR primer pairs located at the 5Ј-(ORF1) and 3Ј-(ORF2) ends of the coding sequences or ORFs of the PHO84, ADH1, and PYK1 genes. B, analysis of the PCR product at the PYK1 ORF by agarose gel electrophoresis. C, analysis of H3-Lys-4 di-and trimethylation (2Me-K4 and 3Me-K4, respectively) at the coding sequences of PHO84, ADH1, and PYK1. The wild-type yeast strain (YKH045) expressing FLAG-tagged H2B and HA-tagged ubiquitin was grown at 30°C in 1% yeast extract containing 2% peptone plus 2% dextrose (YPD). Formaldehyde-based in vivo cross-linking and ChIP was performed as previously described (36,37). Primer pairs located at two different regions in the ORFs of the PHO84, ADH1, and PYK1 genes (see "Experimental Procedures" and panel A) were used for PCR analysis of the IP DNA samples. Immunoprecipitations were performed using rabbit polyclonal antibodies against H3 di-methyl Lys-4 (Upstate Biotechnology, Inc.) and trimethyl Lys-4 (Abcam, Inc.). The ratio of immunoprecipitate over the input is indicated below each band as %IP. D, a ChDIP (chromatin double immunoprecipitation) assay to analyze H2B-Lys-123 ubiquitination at the PHO84 coding sequence in wild-type (YKH045) and H2B-K123R mutant (YKH046) strains expressing FLAG-tagged H2B and HA-tagged ubiquitin (ϩHA-Ub). The yeast cells were grown and cross-linked as described in panel C. The first immunoprecipitation was performed by an anti-FLAG antibody (Sigma) and then eluted by FLAG peptide (Sigma). The eluate was again immunoprecipitated by an anti-HA antibody (Santa Cruz Biotechnology, Inc.). Primer pairs located at two different regions of the PHO84 ORF were used for PCR analysis of the IP DNA samples. The percentage of DNA immunoprecipitated relative to wild-type (%WT) is indicated below each band of the mutant strains. of the PCR product using a specific primer pair targeted to the PYK1 ORF1. The other PCR products also appeared at the same position in the agarose gel (data not shown).
To determine the effect of H2B-Lys-123 ubiquitination on H3-Lys-4 methylation at the active gene loci, we first analyzed the levels of H3-Lys-4 di-and trimethylation at the coding sequences of PHO84, ADH1, and PYK1, using a ChIP assay. Fig.  1C shows that the coding sequences of these genes were both di-and trimethylated in the H2B wild-type strain. The ratio of immunoprecipitate over the input is indicated below each band as %IP. The antibodies used in the ChIP assay were specific to di-and trimethyl Lys-4 of H3 since these antibodies did not generate the ChIP signals at the coding sequences that were devoid of H3-Lys-4 di-and trimethylation in the SET1 deletion mutant strain (see Fig. 4A).
Next, we determined whether the coding sequences of the PHO84, ADH1, and PYK1 genes were also ubiquitinated at Lys-123 of H2B, using a ChDIP (see "Experimental Procedures") assay (30,33) in the H2B wild-type and H2B-K123R mutant strains expressing FLAG-tagged H2B and HA-tagged ubiquitin. In the ChDIP assay, the first immunoprecipitation was performed using an anti-FLAG antibody, and the eluate was subsequently immunoprecipitated by an anti-HA antibody. The IP DNA was analyzed by PCR using primer pairs specific for ORF1 and ORF2 regions of the PHO84 coding sequence (Fig. 1A). If FLAG-tagged H2B is ubiquitinated at Lys-123, a PCR signal would be observed in the wild-type but not the H2B-K123R mutant strain. Fig. 1D indeed shows the presence of the PCR signals at the ORF1 and ORF2 of PHO84 in the H2B wild-type but not H2B-K123R mutant strain, thus demonstrating the specific association of Lys-123-ubiquitinated H2B with the PHO84 coding sequence. These results are in good agreement with previous immunoprecipitation studies (23,24,47) that demonstrated global ubiquitination of H2B at Lys-123 in the WCE of formaldehyde cross-linked yeast cells. Like that at PHO84, the coding sequences of the ADH1 and PYK1 genes were also ubiquitinated at Lys-123 of H2B (data not shown), consistent with previous ChDIP data (33,42).
Finally, we analyzed the role of H2B-Lys-123 ubiquitination in the regulation of H3-Lys-4 di-and trimethylation at the coding sequences of the PHO84, ADH1, and PYK1 genes. Fig. 2A shows that the levels of H3-Lys-4 trimethylation at the coding sequences of PHO84, ADH1, and PYK1 were significantly reduced or abolished in the H2B-K123R mutant strain (DNA immunoprecipitated relative to wild type is presented as %WT), consistent with previous biochemical and ChIP studies (24 -29, 32, 42). However, the levels of H3-Lys-4 dimethylation at the coding sequences of these genes were not altered in the H2B-K123R mutant strain, when input and IP DNAs were amplified within the linear range of PCR analysis ( Fig. 2A). Although PCR was performed within the linear range, we further diluted both the input and the IP DNAs (dimethyl Lys-4 of H3) by 5-fold and then carried out PCR analysis to determine whether H2B-Lys-123 ubiquitination regulates H3-Lys-4 dimethylation at these genes. Even after 5-fold dilution, the levels of H3-Lys-4 dimethylation remained unchanged in the H2B-K123R mutant strain (Fig. 2B). Thus, H2B-Lys-123 ubiquitination is dispensable for up-regulation of H3-Lys-4 dimethyla-tion, but not trimethylation, at the coding sequences of the PHO84, ADH1, and PYK1 genes, consistent with a recent study demonstrating the role of H2B-Lys-123 ubiquitination in the regulation of H3-Lys-4 tri-but not dimethylation (48). Similarly, Henry et al. (30) have also demonstrated that H3-Lys-4 trimethylation, but not dimethylation, at GAL1 is regulated by H2B-Lys-123 ubiquitination.
H2B-Lys-123 Ubiquitination Is Dispensable for Recruitment of RNA Polymerase II to the Coding Sequences of PHO84, ADH1, and PYK1-Recently, H2B-Lys-123 ubiquitination has been implicated in transcriptional elongation (42). Since the coding FIGURE 2. Regulation of H3-Lys-4 tri-but not dimethylation by H2B-Lys-123 ubiquitination at the coding sequences of the PHO84, ADH1, and PYK1 genes. A, H3-Lys-4 tri-but not dimethylation (3Me-K4 and 2Me-K4, respectively) at the coding sequences of the PHO84, ADH1, and PYK1 genes are significantly reduced or abolished in the absence of H2B-Lys-123 ubiquitination. The wild-type and H2B-K123R mutant yeast strains were grown, crosslinked, and immunoprecipitated as described in the legend for Fig. 1. Primer pairs located at two different regions in the ORFs of the PHO84, ADH1, and PYK1 genes were used for PCR analysis of the IP DNA samples. The PCR was performed using diluted input and IP DNAs within the linear range of PCR analysis. B, the PCR analysis of H3-Lys-4 dimethylation at the coding sequences of PHO84, ADH1, and PYK1 in the wild-type and H2B-K123R mutant strains, using 5-fold more diluted input and IP DNAs than those in panel A.
sequences of the PHO84, ADH1, and PYK1 genes are ubiquitinated at H2B-Lys-123, we thus asked whether transcriptional elongation of these genes is regulated by H2B-Lys-123 ubiquiti-nation. To address this question, we analyzed recruitment of RNA polymerase II to the coding sequences of PHO84, ADH1, and PYK1 in the H2B-K123R mutant and its isogenic wild-type strains. If ubiquitination of H2B-Lys-123 is essential for transcriptional elongation, recruitment of RNA polymerase II to the coding sequence would be significantly reduced or abolished in the H2B-K123R mutant strain, and hence, mRNA abundance would be altered in the absence of H2B-Lys-123 ubiquitination.
To analyze recruitment of RNA polymerase II to the coding sequences of the above active genes, we used the antibody (8WG16 from Covance) against the largest subunit (Rpb1p) of RNA polymerase II in the ChIP assay. The specificity of the anti-Rpb1p antibody was confirmed by analyzing recruitment of Rpb1p to the GAL1 coding sequence under inducing (galactose-containing growth medium) and non-inducing (raffinose-containing growth medium) conditions. Recruitment of Rpb1p to the non-transcribing GAL1 coding sequence was not observed using an anti-Rpb1p antibody under non-inducing condition (see Fig. 5D). However, recruitment of Rpb1p was detected at the transcribing GAL1 coding sequence under inducing condition (see Fig. 5D). Thus, the above anti-Rpb1p antibody is specific to the Rpb1p subunit of RNA polymerase II. We also show that Rpb1p was not recruited to the ADH1 upstream activating sequence (UAS) where activator Rap1p (repressor activator protein 1), but not RNA polymerase II, binds (Fig. 3A), further demonstrating the specificity of the anti-Rpb1p antibody. As a nonspecific antibody control at the coding sequence, we used an anti-TBP (TATA box-binding protein) antibody since TBP is recruited to the TATA box (core promoter) but not to the coding sequence (Fig. 3B).
Using the above anti-Rpb1p antibody, we show that recruitment of Rpb1p to the ORF1 and ORF2 regions of the coding sequences of PHO84, ADH1, and PYK1 was not altered within the linear range of PCR analysis in FIGURE 3. H2B-Lys-123 ubiquitination is dispensable for recruitment of RNA polymerase II to the coding sequences of the PHO84, ADH1, and PYK1 genes. A, analysis of recruitment of RNA polymerase II to the UAS and ORF1 of the ADH1 gene. Yeast cells were grown, cross-linked, and immunoprecipitated as described in the legend for Fig. 1. Immunoprecipitation was performed using the mouse monoclonal antibody 8WG16 (Covance) against the carboxyl-terminal domain of the RNA polymerase II large subunit (Rpb1p). B, analysis of recruitment of TBP to the core promoter and coding sequence of the ADH1 gene. Immunoprecipitation was performed using polyclonal antibody against TBP. C, recruitment of RNA polymerase II to the coding sequences of the PHO84, ADH1, and PYK1 genes is not dependent on H2B-Lys-123 ubiquitination. The wild-type and H2B-K123R mutant strains were grown, cross-linked, and immunoprecipitated as described in the legend for Fig. 1. The PCR was performed using diluted input and IP DNAs within the linear range of PCR analysis. D, the PCR analysis of recruitment of RNA polymerase II to the coding sequences of PHO84, ADH1, and PYK1 in the wild-type and H2B-K123R mutant strains, using 5-fold more diluted input and IP DNAs than those in panel C. E, transcription. Total cellular RNA was prepared from the wild-type or H2B-K123R mutant strain, and mRNA levels from the PHO84, ADH1, and PYK1 genes were quantitated by primer extension. The percentage mRNA level relative to wild-type (%WT) is indicated below. A schematic diagram of primer extension analysis is presented on the right side of this panel.
the H2B-K123R mutant strain when compared with isogenic wild-type equivalent (Fig. 3C). Additionally, the PCR analysis using further 5-fold diluted input and IP DNAs did not show any change of Rpb1p recruitment to the coding sequences of these genes in the H2B-K123R mutant strain (Fig. 3D). Consistently, transcription of these genes remained unchanged in the absence of H2B-Lys-123 ubiquitination (Fig. 3E). However, unlike PHO84, ADH1, and PYK1, the active GAL1 gene lost a significant amount of RNA polymerase II recruitment to its coding sequence in the H2B-K123R mutant strain (see Fig. 5A), consistent with previous studies at GAL1 (30,42). Such a reduced level of RNA polymerase II recruitment lowered transcription of GAL1 (see Fig. 5C). Taken together, H2B-Lys-123 ubiquitination is differentially required for recruitment of RNA polymerase II (and hence transcription) at the coding sequences in vivo.
H3-Lys-4 Methylation Is Dispensable for Recruitment of RNA Polymerase II to the Coding Sequences of PHO84, ADH1, and PYK1-We next analyzed the role of H3-Lys-4 methylation in recruitment of RNA polymerase II to the coding sequence and hence transcription. Methylation of histone H3 at Lys-4 is associated with active chromatin in a wide range of eukaryotic organisms (8, 45, 49 -53), and is correlated with gene expression (8,16,54,55). Histone H3 has been found to be dimethylated at Lys-4 in active euchromatic regions but not in silent heterochromatic sites, whereas trimethylated H3-Lys-4 is typically not detected at a locus prior to transcriptional activation (16). Consistently, recent studies have implicated H3-Lys-4 methylation in expression of several genes (8, 16, 43, 54 -56).
Since the nucleosomes at the coding sequences of the PHO84, ADH1, and PYK1 genes are both di-and trimethylated at H3-Lys-4 (Fig. 1C), we thus hypothesized that H3-Lys-4 methylation might play a significant role in transcriptional elongation of these genes. To test this hypothesis, we analyzed recruitment of RNA polymerase II to the coding sequences of these genes in the SET1 deletion mutant. The deletion of SET1 resulted in complete loss of H3-Lys-4 di-and trimethylation at the coding sequences of the PHO84, ADH1, and PYK1 genes (Fig. 4A). However, the loss of H3-Lys-4 di-and trimethylation did not alter recruitment of RNA polymerase II to the coding sequences of PHO84, ADH1, and PYK1 within the linear range of PCR analysis (Fig. 4B), indicating the dispensability of H3-Lys-4 methylation in transcription of these genes. In fact, transcription of these genes was also not altered in ⌬set1 (Fig. 4C). These results are in good agreement with the effect of H2B-Lys-123 ubiquitination on  PHO84, ADH1, and PYK1 genes. A, the analysis of di-and trimethylation of H3-Lys-4 (2Me-K4 and 3Me-K4, respectively) at the coding sequences of the PHO84, ADH1, and PYK1 genes in ⌬set1. The wild-type and mutant yeast strains were grown, cross-linked, and immunoprecipitated as described in the legend for Fig. 1. B, recruitment of RNA polymerase II to the coding sequences of the PHO84, ADH1, and PYK1 genes is not dependent on H3-Lys-4 methylation. The PCR was performed using diluted input and IP DNAs within the linear range of PCR analysis. C, transcription. Total cellular RNA was prepared from the wild-type or ⌬set1 mutant strain, and mRNA levels from the PHO84, ADH1, and PYK1 genes were quantitated by primer extension. recruitment of RNA polymerase II (and hence transcription) at the coding sequences of PHO84, ADH1, and PYK1 genes. Since H2B-Lys-123 ubiquitination, which regulates H3-Lys-4 trimethylation, is not required for recruitment of RNA polymerase II to the PHO84, ADH1, and PYK1 coding sequences, H3-Lys-4 trimethylation does not seem to regulate recruitment of RNA polymerase II to these genes.

. H3-Lys-4 methylation is dispensable for recruitment of RNA polymerase II and transcription of the
Next, we analyzed whether GAL1 requires H3-Lys-4 di-and trimethylation for recruitment of RNA polymerase II to its coding sequence. Fig. 5B shows that recruitment of RNA polymerase II to the GAL1 coding sequence was not altered in the SET1 deletion mutant strain. Consistently, transcription of GAL1 remained unchanged in ⌬set1 (Fig. 5C). Further, although the similar levels of H3-Lys-4 di-and trimethylation were observed at the GAL1 coding sequence in both raffinose (non-inducing)-and galactose (inducing)-containing growth media, recruitment of RNA polymerase II to the coding sequence was observed only in galactosecontaining growth medium (Fig. 5D). Thus, the H3-Lys-4 methylation status does not regulate recruitment of RNA polymerase II to the GAL1 coding sequence.
H3-Lys-4 Methylation Does Not Regulate H3-Lys-9/14 Acetylation at the Coding Sequences of PHO84, ADH1, and PYK1-We next asked whether H2B-Lys-123 ubiquitination or H3-Lys-4 methylation regulates H3-Lys-9/14 acetylation. Histone acetylation is one of the better understood histone modifications. It is now generally accepted that hyperacetylated histones are mostly associated with the transcriptionally activate genomic regions at both local and global levels (46,57,58). However, it remained largely unknown until recently how histone acetylation communicates with other histone modifications. Ng et al. (27) have demonstrated that the deletion of the RAD6 gene, the protein product of which is essential for H2B-Lys-123 ubiquitination, does not affect the global level of H3-Lys-9/14 acetylation in yeast. Consistently, a ChIP study at the transcriptionally active GAL1 promoter demonstrated that H2B-Lys-123 ubiquitination is dispensable for H3-Lys-9/14 acetylation (41). On the other hand, H3 acetylation has been shown to be stimulated biochemically by H3-Lys-4 methylation in mammalian system (59). However, it is not known whether a similar regulatory link between H3-Lys-4 methylation and H3 acetylation is also present in yeast. Further, the role of H3-Lys-4 methylation for upregulation of H3 acetylation has not been analyzed at the active gene loci in vivo. To address this issue, we performed a ChIP assay to analyze the level of H3-Lys-9/14 acetylation at the 5Ј-ends (i.e. ORF1) of the coding sequences of PHO84, ADH1, and PYK1 in the SET1 deletion mutant, using the antibody specific for Lys-9/14-diacetylated H3. The specificity of the antibody against Lys-9/14-diacetylated H3 (Upstate Biotechnology, Inc.) was confirmed by analyzing the levels of H3-Lys-9/14 acetylation at the SAGA-dependent GAL1 promoter (39) in the GCN5 (Gcn5p subunit of SAGA has histone acetyl transferase activity) deletion mutant strain. The ChIP experiments using the above antibody showed drastic loss of H3-Lys-9/14 acetylation at the GAL1 promoter in the GCN5 deletion mutant strain (Fig. 6A), thus demonstrating the specificity of the antibody.
Using the above antibody, we show that the 5Ј-ends of the coding sequences of PHO84, ADH1, and PYK1 were acetylated at Lys-9/14 of H3 (Fig. 6B). The deletion of SET1 had no effect on the levels of H3-Lys-9/14 acetylation at the 5Ј-ends of the coding sequences of these genes (Fig. 6B), whereas H3-Lys-4 methylation was completely abolished (Fig. 4A). Thus, our data demonstrate the absence of a regulatory link between H3-Lys-4 methylation and H3-Lys-9/14 acetylation at several active genes in vivo.
We next asked whether H3-Lys-4 methylation globally regulates H3-Lys-9/14 acetylation in yeast. To address this question, we analyzed the levels of H3-Lys-9/14 acetylation in the WCE of the wild-type and SET1 deletion mutant strains. Fig. 6C shows that the global level of H3-Lys-9/14 acetylation remained unaltered in the SET1 deletion mutant strain, in contrast to FIGURE 5. Analysis of recruitment of RNA polymerase II to the GAL1 coding sequence in the H2B-K123R and ⌬set1 mutant strains. A, recruitment of RNA polymerase II to the GAL1 coding sequence is significantly reduced in the H2B-K123R mutant strain. The wild-type and H2B-K123R mutant strains expressing FLAG-tagged H2B and HA-tagged ubiquitin were first grown in raffinose-containing YPR growth media up to an A 600 of 0.9 and then shifted to galactose-containing YPG growth media for 90 min before treatment with formaldehyde. Immunoprecipitation was performed using the mouse monoclonal antibody 8WG16 (Covance) against the Rpb1p subunit of RNA polymerase II. The specific primer pair targeted to the GAL1 ORF (see "Experimental Procedures") was used for PCR analysis of the IP DNA samples. B, recruitment of Rpb1p to the GAL1 coding sequence is not altered in ⌬set1. Yeast cells were grown, cross-linked, and immunoprecipitated as described in panel A. C, transcription. Total cellular RNAs were prepared from the ⌬set1 and H2B-K123R mutant strains and their isogenic wild-type equivalents. The yeast strains were grown as described in panel A. The GAL1 mRNA levels in the wild-type and mutant strains were quantitated by primer extension. D, analysis of H3-Lys-4 methylation as well as recruitment of RNA polymerase II at the GAL1 coding sequence in raffinose-and galactose-containing growth media. Yeast cells were grown at 30°C in 1% yeast extract containing 2% peptone plus 2% galactose or raffinose up to an A 600 of 1.0 prior to formaldehyde cross-linking. Immunoprecipitations were performed as described in panel A and in Fig. 1. 2Me-K4 and 3Me-K4, di-and trimethylation of H3-Lys-4.
previous biochemical data in mammalian system (59). As a loading control, we show an equal amount of histone H3 in the WCEs of the wild-type and SET1 deletion mutant strains (Fig.  6C). Thus, taken together, H3-Lys-4 methylation does not regulate H3-Lys-9/14 acetylation globally as well as at the active gene loci in yeast.
Although both di-and trimethylation of H3-Lys-4 are regulated by H2B-Lys-123 ubiquitination at the PHO84 core promoter (33), only H3-Lys-4 trimethylation is affected in the absence of H2B-Lys-123 ubiquitination at the PHO84 coding sequence. However, the molecular basis for such a differential regulation of H3-Lys-4 methylation by H2B-Lys-123 ubiquitination at the PHO84 core promoter and coding sequence remains unknown. Previous studies have demonstrated that both di-and trimethylation of H3-Lys-4 are differentially controlled by COMPASS components (28). For example, purified COMPASS lacking Cps60p can mono-and dimethylate but is not capable of trimethylating H3-Lys-4 (28). Cps25p component of COMPASS is essential for di-and trimethylation of H3-Lys-4 (28). Cps30p drastically affects COMPASS-mediated mono-, di, and trimethylation of H3-Lys-4 (28). Perhaps the subunit composition or methyltransferase activity of COM-PASS at the promoter differs from that at the coding sequence, possibly because of the association of COMPASS with different transcription factors at the promoter and coding sequence, and hence, di-and trimethylation of H3-Lys-4 are differentially regulated by COMPASS at the PHO84 promoter and coding sequence. However, such a model remains to be tested.
H2B-Lys-123 ubiquitination has been implicated in transcriptional elongation on the basis of the growth defect of the yeast strain carrying H2B-K123R point mutation or RAD6 deletion in the growth media containing 6-aza uracil (42). Thus, one would expect an impairment of RNA polymerase II recruitment to the coding sequence in the absence of H2B-Lys-123 ubiquitination. However, it was not known whether H2B-Lys-123 ubiquitination is required for recruitment of RNA polymerase II to the coding sequences of the active genes. Recently, Xiao et al. (42) have demonstrated that Bre1p, which ubiquitinates H2B at Lys-123, is essential for recruitment of RNA polymerase II to the promoter as well as 5Ј-end of the coding sequence of the GAL1 gene, indicating the role of H2B-Lys-123 ubiquitination in recruitment of RNA polymerase II. Consistently, we show here that recruitment of RNA polymerase II to the GAL1 coding sequence is significantly reduced in the H2B-K123R mutant strain. However, H2B-Lys-123 ubiquitination is not required for recruitment of RNA polymerase II to the coding sequences of PHO84, ADH1, and PYK1, although their coding sequences are ubiquitinated at Lys-123 of H2B (33,42). Similarly, transcription of these genes remained unchanged in the H2B-K123R mutant strain when compared with the wild-type equivalent. However, there might be a subset of genes, like GAL1, FIGURE 6. H3-Lys-4 methylation does not regulate H3-Lys-9/14 acetylation in S. cerevisiae. A, analysis of H3-Lys-9/14 acetylation levels at the UAS and core promoter of the GAL1 gene in ⌬gcn5. The wild-type and GCN5 deletion mutant strains were grown, cross-linked, and immunoprecipitated as described in the legend for Fig. 5. Immunoprecipitation was performed using a rabbit polyclonal antibody against Lys-9/14 di-acetylated H3 (Ac-H3, Upstate Biotechnology, Inc.). B, analysis of H3-Lys-9/14 acetylation levels at the coding sequences of the PHO84, ADH1, and PYK1 genes in ⌬set1. The wild-type and SET1 deletion mutant strains were grown, cross-linked, and immunoprecipitated as described in the legend for Fig. 1. Immunoprecipitation was performed using a rabbit polyclonal antibody against Lys-9/14 di-acetylated H3 (Upstate Biotechnology, Inc.). C, H3-Lys-9/14 acetylation is not globally regulated by H3-Lys-4 methylation. The wild-type and SET1 deletion mutant strains were grown as described in the legend for Fig. 1, and then WCEs were prepared as in the ChIP assay (36,37). The WCEs were analyzed by Western blot using anti-Lys-9/14 di-acetylated H3 and anti-H3 (Abcam, Inc.) antibodies. that would require H2B-Lys-123 ubiquitination for recruitment of RNA polymerase II (and hence transcriptional elongation) at their coding sequences since previous studies have demonstrated the growth defect of H2B-K123R mutant strain in a synthetic complete medium containing 6-aza uracil (42). Furthermore, several other studies (30,60,61) have demonstrated that H2B-Lys-123 ubiquitination is involved in transcriptional repression as well as stimulation of the active genes. However, the molecular basis for such differential roles of H2B-Lys-123 ubiquitination in the regulation of transcription remains unknown. Perhaps the overall pattern of histone modifications, not a single modification at Lys-123 of H2B, determines the function of chromatin structure in transcription.
Our further in vivo analysis on the regulation of transcriptional elongation by H3-Lys-4 methylation reveals that recruitment of RNA polymerase II to the coding sequences of the PHO84, ADH1, and PYK1 genes is not altered in the SET1 deletion mutant strain when compared with the isogenic wild-type equivalent. Consistently, transcription of these genes is not changed in ⌬set1. Like that at PHO84, ADH1, and PYK1, recruitment of RNA polymerase II to the coding sequence of the GAL1 gene is not affected in the absence of H3-Lys-4 diand trimethylation. Thus, H3-Lys-4 methylation does not seem to play an active role in the regulation of transcriptional elongation of these genes. However, several studies (8,16,60,62) have demonstrated the role of Set1p or H3-Lys-4 methylation in transcription of a subset of genes, consistent with the fact that H3-Lys-4 methylation is associated with active chromatin in a wide range of eukaryotic organisms (8, 45, 49 -53) and is correlated with transcription (43, 54 -56). Currently, it is not known how H3-Lys-4 methylation can be involved in different transcriptional states. According to the "histone code" hypothesis (63), it is likely that H3-Lys-4 methylation in combination with other histone modifications determines the outcome on chromatin structure.