TFetian_CSHL_MOET25_conferencePoster.pdf

Proper gene expression in eukaryotes depends on the dynamic interplay between transcription and chromatin regulatory mechanisms. Nucleosomes can be co-transcriptionally altered by post-translational modifications, remodeling, repositioning or by site-specifically swapping out canonical histones with histone variants. Many studies have shed light on how nucleosomes are co-transcriptionally modified on a molecular level; however, the extent to which these modifications directly contribute to transcription and chromatin structure is largely unknown. Monoubiquitylation of H2B (H2Bub) on K123 in budding yeast (K120 in humans) is a conserved modification that is tightly coupled to RNA polymerase II activity through the Paf1C transcription elongation factor. A small domain within the Rtf1 subunit of Paf1C, named the Histone Modification Domain (HMD), directly interacts with the ubiquitin conjugase Rad6 leading to H2Bub stimulation. We found that disruption of the Rad6-Rtf1 interaction interface is synthetically lethal with the absence of H2A.Z, the H2A variant enriched at the +1 nucleosome of genes. This extreme outcome suggests that H2A.Z and H2Bub have important overlapping functions that would normally compensate for each other. Through an unbiased genetic suppressor screen in yeast, we found that multiple well-known chromatin regulators become toxic in the absence of H2Bub and H2A.Z, including the Set1 histone methyltransferase complex (Set1C) and the Chd1 chromatin remodeler. Set1C is responsible for H3K4 mono-, di- and tri-methylation, which have differential occupancy patterns on chromatin. The synthetic lethality between H2Bub and H2A.Z is suppressed by the loss of the catalytic subunit of Set1C (Set1), catalytically dead mutants of Set1, or the loss of the methylated lysine on H3 (H3K4). Importantly, in the absence of H2Bub and H2A.Z, SET1 deletion is epistatic with H3K4 substitution mutants indicating that there is no added advantage of Set1 loss if the methylation site is already compromised. These results suggest that Set1C-mediated toxicity is primarily due to residual H3K4 methylation. Interestingly, while all forms of H3K4me are reduced in the absence of H2Bub, H3K4me1 is the least affected. Follow-up experiments will probe genomic profiles of H3K4me1 to pinpoint the cause of the toxicity. We have also developed a rapid protein depletion system to probe the immediate genome-wide effects on transcription and chromatin upon losing H2Bub and/or H2A.Z. This system will be utilized to understand the functional interactions between H2Bub, H2A.Z and other chromatin players that go awry in their absence, including Set1C and Chd1. Collectively, our results underscore a tight balance between co-transcriptional chromatin modifications that is essential for cell viability.