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

The incorporation of histone variant H2A.Z, a conserved H2A variant, into nucleosomes creates specialized chromatin domains that regulate DNA-templated processes, including transcription. H2A.Z-containing nucleosomes may poise gene regulatory regions for transcription, but mechanistic details are lacking. 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 made use of derivatives where the C-terminus was replaced with the corresponding region of H2A (ZA protein), or the H2A region plus a transcriptional activating peptide, with the idea of regenerating the H2A.Z-dependent regulation globally. An assessment of the genome-wide distribution of these H2A.Z derivatives shows that the H2A.Z C-terminal region is crucial for both maintaining the occupation level of H2A.Z and the proper positioning of the targeted nucleosomes. Interestingly, the specific contribution on incorporation efficiency vs nucleosome positioning varies enormously, depending on loci analyzed. For the ZA protein, positioning changes most importantly near origins of replication and snoRNA gene promoters, while incorporation is affected on all other sites. Furthermore, the role of H2A.Z in global transcription regulation also depends on its C-terminal region. Remarkably however, the latter mostly involves genes without a H2A.Z nucleosome in the promoter. We profiled H2A.Z and chimeric proteins derived from H2A.Z that bear modified C-terminal regions, namely ZA and ZA-rII', all tagged with HA, in yeast Saccharomyces cerevisiae by ChIP-chip on tiling arrays. All H2A.Z derivatives ChIP samples were labeled with Cy5 and hybridized against H3 ChIPs or Input DNA (labeled with Cy3). The microarrays used were described before (Jeronimo and Robert, 2014). The data were normalized using the Limma Loess method, and replicates were combined as described previously (Ren et al. 2000). The data were subjected to one round of smoothing using a Gaussian sliding window with a standard deviation of 100 bp to generate data points in 10-bp intervals (Guillemette et al. 2005). Each experiment was done in duplicate.