The Histone Variant H2A.Z C-Terminal Domain Has Locus-Specific Differential Effects on H2A.Z Occupancy and Nucleosome Localization

We provide evidence that the Saccharomyces cerevisiae C-terminal region of histone variant H2A.Z can mediate its special function in performing gene regulation by interacting with effector proteins and chaperones. These functional interactions allow H2A.Z not only to incorporate to very specific gene regulatory regions, but also to facilitate the gene expression process. ABSTRACT The incorporation of histone variant H2A.Z into nucleosomes creates specialized chromatin domains that regulate DNA-templated processes, such as gene transcription. In Saccharomyces cerevisiae, the diverging H2A.Z C terminus is thought to provide the H2A.Z exclusive functions. To elucidate the roles of this H2A.Z C terminus genome-wide, we used derivatives in which the C terminus was replaced with the corresponding region of H2A (ZA protein), or the H2A region plus a transcriptional activating peptide (ZA-rII′), with the intent of regenerating the H2A.Z-dependent regulation globally. The distribution of these H2A.Z derivatives indicates that the H2A.Z C-terminal region is crucial for both maintaining the occupation level of H2A.Z and the proper positioning of targeted nucleosomes. Interestingly, the specific contribution on incorporation efficiency versus nucleosome positioning varies enormously depending on the locus analyzed. Specifically, the role of H2A.Z in global transcription regulation relies on its C-terminal region. Remarkably, however, this mostly involves genes without a H2A.Z nucleosome in the promoter. Lastly, we demonstrate that the main chaperone complex which deposits H2A.Z to gene regulatory region (SWR1-C) is necessary to localize all H2A.Z derivatives at their specific loci, indicating that the differential association of these derivatives is not due to impaired interaction with SWR1-C. IMPORTANCE We provide evidence that the Saccharomyces cerevisiae C-terminal region of histone variant H2A.Z can mediate its special function in performing gene regulation by interacting with effector proteins and chaperones. These functional interactions allow H2A.Z not only to incorporate to very specific gene regulatory regions, but also to facilitate the gene expression process. To achieve this, we used a chimeric protein which lacks the native H2A.Z C-terminal region but contains an acidic activating region, a module that is known to interact with components of chromatin-remodeling entities and/or transcription modulators. We reasoned that because this activating region can fulfill the role of the H2A.Z C-terminal region, at least in part, the role of the latter would be to interact with these activating region targets.

must be remodeled in order to make DNA accessible. Several fundamental mechanisms can alter chromatin structure, including post-translational modifications of histones, which may modify their properties or interacting partners; ATP-dependent chromatin remodeling; and the replacement of replicative histones by histone variants which change the histone composition of nucleosomes. The latter mechanism creates specialized chromatin domains that can be permissive or not permissive to transcription (1).
H2A.Z, a histone variant of H2A, can be incorporated into nucleosomes in order to create such a specialized chromatin domain. H2A.Z is conserved among diverse eukaryotic organisms, from yeast to mammals, and replaces replicative H2A histone in 5% to 10% of all nucleosomes (2). In Saccharomyces cerevisiae, H2A.Z plays roles in the regulation of gene expression, DNA repair, maintenance of heterochromatic-euchromatic boundaries, chromosome segregation, and resistance to genotoxic stress (3)(4)(5)(6)(7). In general, the three-dimensional structure of H2A.Z-containing nucleosomes is similar to that of H2A-containing nucleosomes. However, there are subtle differences in the specific regions that differentiate structures of nucleosomes containing H2A versus H2A.Z, and this may explain their functional differences (8). The main divergence between these structures resides in the C-terminal docking domain of H2A.Z, which shows less than 40% sequence identity with the corresponding region of H2A.
In yeast, H2A.Z-nucleosomes are found on most gene promoters (9)(10)(11)(12)(13)(14)(15)(16), and this genome-wide localization is conserved in Drosophila, chicken, plants and mammals (17)(18)(19)(20)(21)(22). H2A.Z is almost exclusively incorporated in the 11 nucleosome in the direction of transcription (16,23). Several studies have reported H2A.Z-dependent positive and negative regulation of gene transcription (3)(4)(5)(6). It is thought that these specific H2A.Z nucleosomes regulate transcription by their particular positioning (9), by being less stable (12), and/or by contacting specific components of the transcription machinery (5). The acidic surface on H2A.Z-containing nucleosomes may constitute a binding platform for interacting partners such as chromatin-remodeling complexes or transcription coactivators (8). Moreover, H2A.Z is predominantly found in the promoters of inactive or weakly transcribed genes, and the H2A.Z-containing nucleosomes are thought to prepare these promoter areas to be disassembled for transcriptional initiation. Indeed, the assembly of the transcription pre-initiation complex (PIC) has been proposed to evict H2A.Z from gene promoters (24). More recent evidence suggests that H2A.Z eviction is dependent on RNA polymerase II and the Kin28/Cdk7 kinase, which phosphorylates serine 5 on the carboxy-terminal domain of the Pol II subunit, Rpb1 (25). Indeed, the association of H2A.Z with gene promoters is gradually lost following the induction of transcription in yeast and in mammals (4,5,9,26). During transcription elongation, nucleosomes in gene bodies are also remodeled to allow the passage of RNA polymerase II. Finally, there is evidence that H2A.Z eviction from gene bodies is also performed by histone chaperones, such as FACT and Spt6 (27,28).
In addition to RNA polymerase II-promoters, previous analyses of genome-wide location studies of H2A.Z have reported the presence of H2A.Z at a variety of genomic loci, including transcribed sequences (tRNAs, small nucleolar RNAs, rRNAs, and Ty elements) (9,12,14) and non-transcribed elements (the mating-type silent cassette, HZADs [Htz1-activated domains], telomeric regions, centromeres, replication origins, and double-stranded DNA breaks) (9-12, 14, 29-32). However, we note that there are also conflicting results on subsets of these localizations (10,12). Although the role of H2A.Z in the maintenance of heterochromatic-euchromatic boundaries, chromosome segregation and DNA repair, have been investigated, the recruitment mechanisms of H2A.Z to the various genomic loci are still unclear, and the function of the H2A.Z C-terminal docking domain in these processes remains to be investigated.
Notwithstanding these studies, the involvement of the H2A.Z C-terminal docking domain has not been studied with regard to genome-wide localization and transcriptional control. Thus, in order to determine the role of the H2A.Z C-terminal docking domain in the functions of H2A.Z-containing nucleosomes, we used chimeric proteins derived from H2A.Z which bore modified C-terminal regions (Fig. 1A). Surprisingly, the results showed that the genome-wide localization of H2A.Z at many gene promoters is not dependent on its C-terminal region. However, modifications of the H2A.Z C-terminal docking domain leads to a loss of occupancy at these promoters and a slight shift of the position of the H2A.Z variant nucleosome toward the 12 nucleosome. However, the C terminus is required for correct positioning of H2A.Z nucleosomes near origins of replication, centromeres and promoters of snoRNA genes. The data also demonstrate that the special function of H2A.Z in regulation of the global transcriptome is dependent on its C-terminal region. Remarkably, a modification of the H2A.Z C terminus mostly affects the expression of genes lacking a H2A.Z nucleosome in the promoter, and several biological processes are affected. Therefore, our study also provides evidence for a global role of the H2A.Z C-terminal docking domain in regulation of gene transcription by coupling specific occupancy at gene promoters with the transcriptional outcome.

RESULTS
A modification of the C-terminal domain of H2A.Z affects its localization to certain loci and reduces its nucleosomal incorporation genome-wide. Previous reports have shown that the H2A.Z C-terminal docking domain is essential for the incorporation of H2A.Z within chromatin and the role of H2A.Z in gene expression on a limited number of loci, but the necessity of this region regarding its genome-wide location and gene regulation has not been documented. To study the functional role of the H2A.Z C-terminal docking domain, we wanted to test whether replacing it with a transcriptional activating region could fulfill the same role. In fact, such H2A.Z chimeras have been used in the past, namely, the ZA and ZA-rII9 fusion proteins (5,33). Briefly, the ZA fusion protein was obtained by replacing the C-terminal region of H2A.Z, which includes the M6 region (amino acids 97 to 134), with the corresponding region of H2A (amino acids 91 to 132) (Fig. 1A). Larochelle and Gaudreau (33) hypothesized that the H2A.Z C-terminal region could harbor a function reminiscent of a transcriptional activating region and therefore created a chimeric protein named ZA-rII9 by fusing the ZA protein to the Gal4 acidic transcriptional activating region (amino acids 840 to 881) (Fig. 1A). We reasoned that this transcriptional activating domain could, at least to a certain degree, mimic the function of the H2A.Z C-terminal docking domain and therefore complement the transcriptional phenotypes of Dhtz1 cells. Previous data showed that the expression of the ZA chimera protein affects the expression of H2A.Z-regulated genes, namely, GAL1, GAL7, GAL10, and PUR5, and renders the cell more sensitive to genotoxic stress compared to H2A.Z-expressing cells (5,33). Interestingly, the expression of the ZA-rII9 fusion protein allowed us to partially rescue the activation of the genes described above and the resistance to genotoxic stressors (33). Thus, these H2A.Z derivatives are a useful tool in order to dissect and study the functional role of the H2A.Z C-terminal region.
We set up a genome-wide localization assay (chromatin immunoprecipitation [ChIP] with DNA microarray, ChIP-chip) to map H2A.Z and the H2A.Z derivatives across the yeast genome. ChIP assays were performed using yeast strains bearing hemagglutinin (HA) epitopes on H2A.Z, ZA, and ZA-rII9. In each case, a ChIP for histone H3 (using an anti-H3 antibody) was hybridized together with the H2A.Z, ZA, or ZA-rII9 ChIPs to control for nucleosomal density. Genomic data visualization using the UCSC Genome Browser allowed us to determine that the H2A.Z, ZA, and ZA-rII9 proteins are distributed across the yeast genome in a non-random manner (see Fig. 1B for a map of chromosome III). The wild-type (WT) H2A.Z genome-wide location data obtained here agree and overlap very strongly with previously obtained localization results (6, 9-13, 15, 16, 36). In general, the ZA and ZA-rII9 peaks overlap the H2A.Z enrichment peaks, suggesting that, globally, all three proteins follow a similar non-random genomic distribution (Fig. 1B). Comparison of the localization signals revealed that the signal and enrichment peaks are diminished for the ZA fusion protein compared to H2A.Z and ZA-rII9 ( Fig. 1C and D), suggesting at first sight that modification of the H2A.Z C-terminal docking domain does not grossly affect the localization of the enrichment peaks, but rather the level of incorporation of this protein within chromatin. To better appreciate this difference in the binding levels of the H2A.Z derivatives, we demonstrated binding enrichment relative to the transcription start site (TSS) of average genes, normalized either to H3 (Fig. S1A) or to input (Fig. S1B). A closer inspection of our genomic data suggests that H2A.Z is predominantly localized within promoter regions, as expected, and that the distributions of ZA and ZA-rII9 are also similar in that aspect ( Fig. 1C and D). An enrichment of H2A.Z, ZA, and ZA-rII9 was observed upstream of every open reading frame (ORF), as shown in Fig. 1D, and very few genes show an enrichment of H2A.Z, ZA, and ZA-rII9 in the coding region. Intergenic regions where two genes converge, for example, SOL2 and ERS1 (Fig. 1D, dotted lines), did not show an enrichment of H2A.Z, ZA, and ZA-rII9. Intergenic regions between two divergent genes, for example, YCR087C-A and ABP1 (Fig. 1D, dashed lines), often contain two separate enrichment peaks of H2A.Z, ZA, and ZA-rII9, supporting the idea that each promoter presents its own enrichment in H2A.Z. H2A.Z is involved in several cellular processes and multiple groups have reported an association of H2A.Z with particular chromosome elements and other transcribed elements (6, 9-12, 14, 29, 30, 32, 35). Given the evidence for these other localizations of H2A.Z, we verified whether its localization at centromeres, replication origins, HZAD (Htz1-activated domain) genes, tRNA genes, and small nucleolar RNA (snoRNA) genes was dependent on its C-terminal region. To do this, we profiled H2A.Z, ZA, and ZA-rII9 localizations (log 2 ratio IP/H3) over these chromosome elements and the TSS of the respective gene classes using the Versatile Aggregate Profiler (VAP) tool (40). We found that H2A.Z is enriched around centromeres and replication origins, as reported previously, but the H2A.Z distribution pattern at these sites changes significantly when the C-terminal docking domain is modified ( Fig. 2A and B), which may affect the protein's role in chromosome segregation and DNA replication. Moreover, H2A.Z is enriched around the TSS of HZAD, tRNA, and snoRNA genes ( Fig. 2C to E), suggesting that the role of H2A.Z in transcription is not limited to Pol II-transcribed genes but also includes Pol III-transcribed genes. Furthermore, replacement of the H2A.Z C-terminal region with the corresponding region from H2A causes a significant loss of the H2A.Z derivative over the HZAD and tRNA gene promoters ( Fig. 2C and D). In addition, on the snoRNA promoters, there is a clear and distinct shift in localization for the H2A.Z derivatives (Fig. 2E). Interestingly, the addition of the Gal4 transcriptional activating domain partially restores the localization over the HZAD genes and tRNA genes ( Fig. 2C-D), but not the effect on the snoRNA genes (Fig. 2E). We performed paired t tests to ensure that the differences in binding efficiency observed among the various derivatives were significant, and in most cases (except for HZAD versus ZA-rII9), P values were well below 0.01, suggesting that all differences were significant. The results of these t tests are shown in Fig. S2. We conclude that the H2A.Z C-terminal docking domain has a role in the efficient recruitment of H2A.Z not only to promoters of protein-coding genes, but also to promoters of noncoding RNAs and on other chromosomal loci. Furthermore, its localization near centromeres, replication origins, and snoRNA genes is absolutely dependent on the C-terminal docking domain. The replacement of the C-terminal region decreases the incorporation level of H2A.Z at target promoters. As mentioned above, H2A.Z is preferentially localized at promoters and gene regulatory regions, and this localization is conserved from yeasts to mammals (9)(10)(11)(12)(13)(14)(15)(16)(17)(18)(19)(20)(21). Although it was originally suggested that H2A.Z is incorporated in the 21 and 11 nucleosomes flanking the nucleosome-free region, more recent data have suggested that H2A.Z is almost exclusively incorporated in the 11 nucleosome in the direction of transcription (16,23). Moreover, using strains expressing the ZA fusion protein or other variant H2A.Z proteins, previous experiments have documented a lower occupancy in nucleosomes at specific gene promoters, such as those of PHO5, PUR5, and GAL (33)(34)(35), suggesting that the incorporation of H2A.Z into nucleosomes at gene promoters is dependent on its C-terminal region. We also measured the occupancy level and mapped the respective proteins over yeast promoters using the VAP tool (log 2 ratio IP/Input) (40). The enrichment data were aligned for all genes with respect to a reference point, the TSS, and showed that H2A.Z is enriched in a specific area in the promoter regions relative to the coding regions, as expected (Fig. 3A). The detection of ZA-rII9 in this area is significantly reduced with respect to H2A.Z; nevertheless, the addition of the transcriptional activating region to the ZA protein retains the occupation at gene promoters genome-wide at a significant Locus-Specific Effects of the H2A.Z C Terminus Microbiology Spectrum level (Fig. 3A). However, we observed a strongly decreased signal for the ZA protein over these promoters (Fig. 3A), indicating that the occupation level of H2A.Z at promoters is dependent on its C-terminal region. In addition to simple occupancy, we also profiled the positioning of H2A.Z, ZA, and ZA-rII9 nucleosomes and plotted them with respect to the positioning of total nucleosomes (data set from Luk et al. [41]). This allowed us to compare their position with the 11 nucleosome. Given that the 11 nucleosome encompasses the TSS of yeast genes, our data for the WT H2A.Z are in full agreement with previous results, strongly suggestng that H2A.Z enrichment at gene promoters is an incorporation of H2A.Z in the 11 nucleosome. However, the ZA-rII9 peak clearly shifted toward the 12 nucleosome (Fig. 3A, blue versus black curves). In this analysis, due to the extremely low ZA protein incorporation, conclusive determination of the precise sub-localization of the ZA protein was not possible. Taken together, these results support a model in which the H2A.Z C-terminal region is involved in the overall incorporation level of H2A.Z into nucleosomes as well as the precise positioning of the H2A.Z-containing nucleosome on yeast promoters. Profiling of the H2A.Z, ZA, and ZA-rII9 distributions over promoters using VAP allowed us also to determine the specific number of genes with at least a 2-fold enrichment (IP/ H3 log 2 ratio $ 1.00) of H2A.Z, ZA, or ZA-rII9 at the TSS. We identified 2,240 genes that showed an enrichment of H2A.Z at the TSS and referred to them as "H2A.Z-associated genes." These genes represent 44% of the yeast verified ORFs and are potentially transcriptionally regulated by H2A.Z. We also identified 1,128 ZA-rII9-associated genes (22% of verified ORFs) and 370 ZA-associated genes (7% of verified ORFs). Comparison of the pool of H2A.Z-associated genes with the ZA-associated genes showed that only 4% of ZA-associated genes are specific to this protein and 96% are in the pool of H2A.Z-associated genes (Fig. 3B, upper graph). These data reinforce the conclusion that the replacement of the H2A.Z C-terminal region by the corresponding region from H2A causes a loss of occupancy within chromatin rather than a gross mislocalization of the protein.
Similarly, of the 1,128 ZA-rII9-associated genes, only 10% did not overlap the H2A.Z-associated gene pool (Fig. 3B, bottom graph). Therefore, the addition of the Gal4 transcriptional activating domain to the ZA fusion protein partially restores promoter occupancy within chromatin but is not sufficient to complement it to WT H2A.Z levels.
A previous H2A.Z localization study showed that the majority of H2A.Z associated promoters did not contain classical TATA boxes (12). Given that almost all promoter localizations of the protein chimaeras studied here are found in the pool for the WT H2A.Z protein, the data predicted that the fusion proteins preferentially associate with TATA-less promoters. First, our results confirmed that most of the promoters of H2A.Zassociated genes are TATA-less (74%), and most of them also are TAF1 (TFIID)-enriched (70%) (Fig. 3C, left charts). As predicted above, the promoters of ZA-rII9-and ZA-associated genes also are preferentially TATA-less (77% and 83%, respectively) and TFIIDenriched (also 73% and 85%; Fig. 3C, middle and right charts). These results suggest that the main downstream effector on H2A.Z-occupied promoters could be TFIID. However, the fact that the promoters of ZA-associated genes are preferentially TATAless and TFIID-enriched also suggests that the targeting of H2A.Z to its specific promoters is independent from its C-terminal region.
The H2A.Z C-terminal domain is important for positive and negative regulation of transcription of genes. Given that we observed an enrichment of H2A.Z at the promoters of 44% of ORFs and that this incorporation is affected by replacement of the H2A.Z C-terminal region, we hypothesized that the replacement of the H2A.Z C-terminal region deregulates the expression of H2A.Z-associated genes. To verify this hypothesis, we performed RNA sequencing (RNA-seq) to analyze the effect of replacement of the H2A.Z C terminus on gene expression. We compared the transcriptomes of Dhtz1 cells and Dhtz1 cells complemented with a copy of H2A.Z, ZA, or ZA-rII9. We identified genes from the Dhtz1, ZA, and ZA-rII9 strains showing differential expression with an adjusted P value of ,0.05 and an absolute log 2 -fold change of .1.00 compared to the WT strain.
As a control, we wanted to verify whether RNA expression levels would recapitulate previous results which showed that differential expression of Gal4-regulated genes is higher in cells harboring the ZA-rII9 construct compared to WT cells, and that the expression of known H2A.Z-regulated genes is downregulated in Dhtz1 cells (6, 35) (Fig. 4A). Using RNA-seq, 308 differentially expressed genes (log 2 -fold change $ 1.00 and P # 0.05) were identified in the Dhtz1 strain compared to the WT, 530 differentially expressed genes in the ZA strain, and 427 differentially expressed genes in the ZA-rII9 strain (Fig. 4B). While a slight bias toward downregulated genes (58%) was observed in the Dhtz1 strain, in the ZA-and ZA-rII9-expressing strains, 65% and 72%, respectively, of the differentially expressed genes were upregulated. This upregulation of genes in the ZA-rII9 strain would be expected because we inserted a transcriptional activating domain into the ZA fusion protein. Remarkably, however, a direct comparison of the identity of the differentially expressed genes in the Dhtz1, ZA, and ZA-rII9 strains allowed to determine that half or more of the differentially expressed genes were specific to each strain (Fig. 4C, left). This surprising pattern held true for downregulated and upregulated genes (Fig. 4C, middle and right). Therefore, these analyses show that the changes in the transcriptomes in the respective cells diverged considerably. Therefore, replacement of the H2A.Z C terminus causes a major shift of the expressed transcriptome compared to a WT or Dhtz1 strain. Given that H2A.Z is associated with several biological processesm such as transcription, DNA repair, chromosome segregation, chromatin silencingm and RNA splicing (7), we used the Metascape tool to identify significant enrichment of Gene Ontology terms among differentially expressed genes in Dhtz1, ZA, and ZA-rII9 strains compared to the WT. Although we found some enriched biological processes for differentially expressed genes in the ZA and ZA-rII9 strains compared to the WT, we detected no significant distinction between gene classes which would allow us to draw any convincing conclusion regarding the role of the H2A.Z C-terminal region.
H2A.Z and H2A.Z derivatives incorporation at the +1 nucleosome of genes does not correlate with differential gene expression. Previous studies have reported a negative correlation between H2A.Z occupancy at promoters and transcription rates in yeast (4,9,11,12). Because of this observation, it has been suggested that H2A.Z marks the promoters of inactive or weakly transcribed genes and presumably prepares nucleosomes for disassembly upon gene induction (4,5,9,26). However, the level of acetylated H2A.Z at promoters has been associated with actively transcribed genes (13). In contrast, several papers have reported that the H2A.Z incorporation level at gene promoters does not correlate with their transcription rates or RNA polymerase II occupancy (10,16,29). To investigate whether there was a correlation between the H2A.Z occupancy at the 11 nucleosome and differential gene expression in our data, we compared the data sets of genes with an enrichment of H2A.Z or H2A.Z derivatives at the promoter (ChIP-chip data) and the differentially expressed genes (RNA-seq data) (Fig. 5). The results from this comparison suggest that in the Dhtz1 strain, 65% of differentially expressed genes did not show an enrichment of H2A.Z at their promoter (Fig. 5A). Similarly, 82% and 96% of differentially expressed genes in the ZA-rII9-and ZA-expressing strains did not show an enrichment of ZA-rII9 or ZA at their promoter ( Fig. 5B-C). These results suggest that the association of the proteins studied here at specific promoters is inversely correlated with the expression level of the gene.
To examine this hypothesis more directly, we ranked the genes into 10 groups based on their differential gene expression level and compared them to H2A.Z or H2A.Z derivative occupancy at the respective promoter ( Fig. 6A to C). Intriguingly, we found no correlation between the incorporation level of H2A.Z and H2A.Z derivatives   H2A.Z and H2A.Z derivatives incorporation at the 11 nucleosome of genes does not correlate with differential gene expression. (A to C) All genes were sorted by differential expression level (log 2 -fold change over H2A.Z strain) (gray bars), separated into 10 groups of decreasing (Continued on next page) Locus-Specific Effects of the H2A.Z C Terminus Microbiology Spectrum at promoters and differential gene expression levels, possibly suggesting that genes with very high differential expression (fold change over the WT strain) showed low levels of H2A.Z or H2A.Z derivative occupancy at promoters, and vice versa. However, when the H2A.Z and H2A.Z derivatives' occupancy at the TSS of all genes (log 2 ratio) was plotted against ranked differential gene expression values (log 2 -fold change over H2A.Z strain) ( Fig. 6D and F), no correlation between the occupancy at promoters and the expression level of the gene was observed, confirming previous studies (10,16,29). Thus, we propose that H2A.Z can be targeted and recruited to promoters of transcriptionally active and inactive genes, independently from its C-terminal region, but the transcriptional outcome of a specific gene will differ in a H2A.Z C-terminal-modified strain due to as-yet unknown mechanisms. ZA and ZA-RII9 require SWR1-C for deposition within chromatin loci. One important issue that needed to be addressed was whether or not the differential association of H2A.Z derivatives to chromatin would mean that they would no longer depend on physical association to SWR1-C, an important H2A.Z chaperone, and would therefore function through a different chromatin-depositing factor. To address this, we performed ChIP-seq experiments using HA-tagged versions of H2A.Z, ZA, and ZA-RII9 in wild-type cells or cells lacking SWR1 (Dswr1). We reasoned that if ZA, particularly ZA-rII9, functioned independently of SWR1-C, their association to gene regulatory regions would not be significantly affected by the loss of SWR1-C. Figure 7 shows the average enrichment of H2A.Z derivatives relative to the TSS of genes in either WT or Dswr1 cells. As expected, WT H2A.Z shows the strongest association to the TSS region of genes, followed by ZA-RII9 and then ZA. Importantly, and in all three cases, there is a loss of significant association of the H2A.Z derivatives to the TSS region of genes in the absence of the SWR1 complex. These results clearly demonstrate that differential association of H2A.Z derivatives to chromatin regions cannot be attributed to a problem with their association to the main H2A.Z chaperone complex.

DISCUSSION
The aim of this study was to determine whether the C terminus of H2A.Z affects the nucleosomal occupancy level of this alternative histone and/or the precise localization of these nucleosomes. While some directed analyses on a few loci have been reported, here, we analyzed these effects genome-wide and hence for all possible H2A.Z nucleosomes. For this purpose, we used two H2A.Z variants lacking the normal C terminus: one which the C terminus was replaced with that of H2A (ZA protein), and one which  Locus-Specific Effects of the H2A.Z C Terminus Microbiology Spectrum additionally contained an acidic activation domain (ZA-rII9 protein). We performed ChIP-chip on these two variant proteins and compared the results of the chimaera proteins with those of WT H2A.Z. First of all, the overall localization of the WT H2A.Z reassessed here coincides almost perfectly with previous data (9-12, 14, 29). Therefore, we are confident that the differential results with the chimaera adequately covered all concerned loci. Second, the incorporation efficiency globally appeared to be much reduced for the ZA protein, with a slight amelioration for the ZA-rII9 protein (Fig. 1). However, upon closer inspection, there are major differences in these results depending on the locus analyzed. RNA Pol II promoters: expected and unexpected effects. As expected (33), ZA protein occupancy in the 11 nucleosome after the nucleosome-free region is virtually undetectable, compared to H2A.Z. The rII' addition complements this loss somewhat, but not back to WT levels. Interestingly, overall, the replacement of the H2A.Z C-terminal region did not change the targeting on specific promoters: virtually all affected loci are a subset of the loci with H2A.Z. There was no correlation between the differential expression level of a gene and the occupancy level at the TSS (Fig. 4), as expected from previous studies (10,16,29). Given that steady-state RNA levels reflect a combination of transcription initiation rate and RNA turnover rates (16), we suggest that the occupancy level of H2A.Z at promoters does not predict the expression level of a gene and that H2A.Z marks the promoters of active and inactive genes. Multiple explanations could be provided on how the replacement of the H2A.Z C-terminal domain might affect the transcriptional outcome of a gene, for example, impaired recruitment of certain transcriptional mediators or components of the transcriptional machinery, such as TFIID, SAGA, and RNA Pol II (5,8). However, the 11 nucleosome enrichment peak of ZA-rII9 shifted significantly toward the 12 nucleosome position; due to its very low occupancy, it is not clear whether the ZA 11 nucleosome shifted (Fig. 3B). How the ZA-rII9 shift would affect local or global gene transcription remains to be determined. It is also possible that ZA and ZA-rII9 could affect chromatin dynamics and gene regulation via an indirect mechanism, i.e., independent from that elicited at the TSS of a gene.
RNA Pol III promoters and HZAD loci. The changes in localization patterns of H2A.Z derivatives over RNA Pol III-transcribed tRNA genes are remarkably similar to Pol II promoters; that is, the replacement of the H2A.Z C-terminal region caused a reduced occupancy of H2A.Z derivatives at the 11 nucleosome at tRNA promoters and the rII9 addition complements this loss almost back to WT levels.
HZAD loci are large areas encompassing at least several nucleosomes that are enriched in H2A.Z. Again, the ZA derivative is almost completely lost from these areas and the rII9 peptide does rescue this loss somewhat (Fig. 2C); thus, they behave very similar to RNA Pol II and RNA Pol III promoters.
snoRNA gene promoters. While snoRNA genes are transcribed by RNA polymerase II, there was clearly a different effect upon the replacement of the H2A.Z C-terminal region. First, the reduction in occupancy at the 11 nucleosome was similar for both H2A.Z derivatives, thus with no recovery for the ZA-rII9 protein. Second, the ZA-and ZA-rII9-nucleosomes appeared in two distinct peaks on either side of the TSS compared to one strong enrichment peak upstream of the TSS for the WT H2A.Z (Fig. 2E). At present, we do not have a good explanation for this surprisingly differential effect on snoRNA gene promoters.
Centromeres and origins of replication. There was clearly a different distribution of the H2A.Z derivatives in the chromatin surrounding centromeres and replication origins compared to WT H2A.Z ( Fig. 2A-B). For example, while H2A.Z is incorporated in the centromeric and pericentromeric nucleosomes, ZA-and ZA-rII9 are incorporated only on one side of the centromere. Moreover, the double peak of H2A.Z nucleosomes surrounding replication origins is reduced to one peak of ZA nucleosomes right at the center of the origin. This is interesting because yeast cells expressing the ZA fusion display a lower growth rate than WT H2A.Z and ZA-rII9 (33). Thus, the fact that ZA associates with the center of replicating origins could impede or slow down the recruitment of origin factors.
How could such disparate, locus-specific effects be explained? Previous papers have shown that the H2A.Z C-terminal region determines the association of H2A.Z with the Chz1 histone chaperone or the SWR1 complex for its deposition into chromatin (34,38,42,43). Therefore, the generally reduced occupancy signal of H2A.Z derivatives could be due to an impaired binding of H2A.Z derivatives to SWR-C or a problem in the deposition of H2A.Z derivatives. However, we were able to show that in yeast cells bearing a deletion in SWR1 (Dswr1), the binding of all H2A.Z chimeras was nearly abolished, suggesting that ZA and ZA-RII9 work primarily through SWR1-C for their genomic localization. This is also in line with a recent study by Brewis et al. (44) who used H2A-H2A.Z chimeras to demonstrate that the M6 region of H2A.Z was necessary and sufficient for interaction with SWR1-C. They also showed that the H2A.Z C-terminal region (including the M6 region) was not sufficient to rescue H2A.Z occupancy at specific gene promoters. This is consistent with the fact that we observed differences in localization with our ZA and ZA-rII9 derivatives. Moreover, anchor-away experiments showed that an impairment of the SWR-C mediated deposition of H2A.Z still results in a certain level of H2A.Z occupancy within chromatin (25), possibly explaining the observed occupancy of ZA and ZA-rII9 genome-wide.
Previous papers have also demonstrated that yeast and mammalian nucleosomes containing H2A.Z with mutations in their C terminus are less stable, i.e., the C terminus is required for stable retention of H2A.Z (22,36,37,45). Thus, ZA-or ZA-rII9-containing nucleosomes may be less stable and more easily lost from chromatin.
In addition to problems in deposition and nucleosome stability, the eviction mechanisms for H2A.Z derivatives from chromatin might vary depending on localization. While there is evidence that the INO80 remodeling complex contributes to H2A.Z eviction in general (37,46), its eviction at gene promoters is linked to transcription initiation, promoter escape, and early elongation of Pol II (25,47).
There is also evidence that H2A.Z is deposited via replication-dependent random incorporation into the genome and selectively depleted from transcription units (26)(27)(28)48). This hypothesis could provide a valid explanation for the presence or accumulation of H2A.Z in regions such as HZAD genes, origins of replication, and centromeres. In that case, the surrounding chromatin could influence the final localization of variant-containing nucleosomes, depending on chromatin remodelers attracted to the area via other histone modifications. These remodelers have significant conserved roles in nucleosome homeostasis (49). For example, at Pol II gene promoters, the position of the 11 nucleosome is set by the remodeling activities of Isw2, Isw1a, Ino80, and general gene regulatory factors (45,(50)(51)(52). Because HZADs and snoRNA genes are mostly Pol II-transcribed genes as well, we suggest that the precise localization of the H2A.Z-containing 11 nucleosome in those areas could be dependent on a similar set of factors. Isw1, Isw2, and Ino80 have also been reported to be localized to tRNA gene promoters (53,54), and INO80-C is present at various chromosome elements, including origins of replication and pericentric chromatin (54)(55)(56). Furthermore, specific histone tail modifications are known to attract other remodeling complexes, such as the SWI/SNF and the CHD families (49). While the localization of these complexes on the yeast genome is consistent with a role in regulating nucleosomes with H2A.Z, locus-specific regulation mechanisms remain unclear.
The C-terminal-modified H2A.Z variants used in this study do not, to our knowledge, represent physiological proteins in yeast. Nevertheless, we note that there is a naturally occurring splice isoform of the human H2A.Z-2 gene encoding a C-terminally truncated H2A.Z protein which is predominantly expressed in human brain, skeletal muscle, and liver tissues (36). This short form of human H2A.Z-2 protein is less stably bound to chromatin than the full-length protein (36), suggesting that the function of the H2A.Z C-terminal region in the regulation of the association of H2A.Z with nucleosomes is conserved and biologically relevant.
Overall, our study revealed that there is no global, all-encompassing mechanism by which the H2A.Z C terminus affects nucleosomal properties. It appears that incorporation, localization, and H2A.Z loss rates vary in a context-specific fashion. Therefore, the locus-specific mechanisms by which the C terminus of H2A.Z affects incorporation rates and localization need to be investigated by targeted and locus-specific approaches.

MATERIALS AND METHODS
Yeast strains and protein chimaera construction. The WT haploid strain W303 (MAT A) was used in this study. The Dhtz1 haploid strains was obtained by replacing the HTZ1 allele with a KMX cassette. The HA-H2A.Z, HA-ZA, and HA-ZA-rII9 constructs used in this study are as described previously (5,33). Briefly, all H2A.Z derivatives bear a HA tag inserted in the BglII site of the HTZ1 gene and were expressed from the ACT1 promoter. The ZA fusion was constructed by fusing amino acids 91 to 132 of H2A to the C-terminal region of amino acids 1 to 97 of H2A.Z (5). The ZA-rII9 fusion was generated by adding amino acids 840 to 881 of Gal4 to the C-terminal end of ZA (33). The plasmids were linearized and integrated at the endogenous URA3 locus of the Dhtz1 strain.
Chromatin immunoprecipitations, ChIP-chip, and ChIP-seq. ChIP experiments were performed in duplicates as previously described (5,9). Briefly, 50 mL of cells were grown in yeast extract-peptone-dextrose (YPD; 2% glucose) medium to an optical density at 600 nm (OD 600 ) of 0.6 and fixed with 1% formaldehyde for 10 min. Next, 350 mL of sonicated whole-cell extract was incubated with either anti-HA (12CA5; Roche) or anti-H3 (ab1791; Abcam) antibody coupled to magnetic beads (Dynal) overnight at 4°C with agitation. Immunoprecipitated DNA was used for genome-wide location analysis.
For the ChIP-chip, H2A.Z, ZA, and ZA-rII9 ChIP DNA, labeled with Cy5, was hybridized in competition with Cy3-labeled H3 ChIP or input DNA. The microarrays used for location analysis have been described previously (57) and were purchased from Agilent Technologies (Palo Alto, CA, USA). Samples were cohybridized on an array containing a total of 180,000 60-mer probes (including controls) covering the entire yeast genome with virtually no gaps between the probes.
For the ChIP-seq, ChIP experiments were performed in duplicates as previously described (57). For anti-HA ChIP-seq, 500 mL of whole-cell extract was incubated with 50 mL pre-equilibrated Pierce Anti-HA Magnetic Beads (Thermo Fisher Scientific [Waltham, MA], cat no. 88836) overnight at 4°C. Immunoprecipitated DNAs were purified and used as input into standard Illumina library construction.
ChIP-chip data analysis. The ChIP-chip data were normalized using the Limma Loess method, and replicates were combined using a weighted average method as described previously (58). The data were subjected to one round of smoothing using a Gaussian filter (SD = 100 bp). Aggregate profiles were generated using the VAP (40). All data sets for the ChIP-chip experiment were deposited into the NCBI Gene Expression Omnibus (GEO) database under accession no. GSE156492 (enter token gzozeeimrrqlrqd into the box).
ChIP-seq data analysis. ChIP-seq reads were analyzed using the GenPipes ChIP-seq pipeline (59). Briefly, the pipeline begins by trimming adaptors and low-quality bases using Trimmomatic v0.39 (60) and mapping the reads to a Saccharomyces cerevisiae R64-1-1 reference genome using Burrows-Wheeler Aligner (BWA) v0.7.17 (61). Reads are filtered by mapping quality and duplicate reads are marked using Sambamba v0.8.0 (62). Next, Homer v4.11 quality control routines were used to provide information and feedback regarding the quality of the experiment (63). Peak calls are executed by Model-Based Analysis of ChIP-seq (MACS) v2.2.7.1 (64), and annotation and motif discovery for narrow peaks are executed using Homer.
The trimmed reads were aligned using Tophat37 onto a genome composed of the S. cerevisiae S288C genome (sacCer3). DESeq2 was used to identify the differentially expressed genes. Yeast genes showing a differential expression with an adjusted P value of ,0.05 and having an absolute log 2 -fold change of .1.00 were considered differentially expressed. The Gene Ontology term enrichment analysis for differentially expressed genes was investigated using Metascape (https://metascape.org/gp/index.html #/main/step1). All data sets for the RNA-seq experiment were deposited into NCBI GEO database under accession no. PRJNA657613.
Data availability. The raw reads and processed files from the ChIP-chip, ChIP-seq, and RNA-seq experiments have been deposited in the NCBI Gene Expression Omnibus (GEO) database (http://www .ncbi.nlm.nih.gov/geo/) and are accessible under accession no. GSE156492 and PRJNA657613.

SUPPLEMENTAL MATERIAL
Supplemental material is available online only. SUPPLEMENTAL FILE 1, PDF file, 0.1 MB.