Histone H3K18 & H3K23 acetylation directs establishment of MLL-mediated H3K4 methylation

Abstract

In an unmodified state, positively charged histone N-terminal tails engage nucleosomal DNA in a manner which restricts access to not only the underlying DNA but also key tail residues subject to binding and/or modification. Charge-neutralizing modifications, such as histone acetylation, serve to disrupt this DNA-tail interaction, facilitating access to such residues. We previously showed that a polyacetylation-mediated chromatin "switch" governs the read-write capability of H3K4me3 by the MLL1 methyltransferase complex. Here, we discern the relative contributions of site-specific acetylation states along the H3 tail and extend our interrogation to other chromatin modifiers. We show that the contributions of H3 tail acetylation to H3K4 methylation by MLL1 are highly variable, with H3K18 and H3K23 acetylation exhibiting robust stimulatory effects and that this extends to the related H3K4 methyltransferase complex, MLL4. We show that H3K4me1 and H3K4me3 are found preferentially co-enriched with H3 N-terminal tail proteoforms bearing dual H3K18 and H3K23 acetylation (H3{K18acK23ac}). We further show that this effect is specific to H3K4 methylation, while methyltransferases targeting other H3 tail residues (H3K9, H3K27, & H3K36), a methyltransferase targeting the nucleosome core (H3K79), and a kinase targeting a residue directly adjacent to H3K4 (H3T3) are insensitive to tail acetylation. Together, these findings indicate a unique and robust stimulation of H3K4 methylation by H3K18 and H3K23 acetylation and provide key insight into why H3K4 methylation is often associated with histone acetylation in the context of active gene expression.

Keywords: chromatin; epigenetics; histone acetylation; histone methylation; nucleosome.

Figure 1

Acetylation of H3K18 & H3K23 promotes H3K4me1.A, in vitro methylation assays with the recombinant MLL4 complex (10 nM) and a panel of designated nucleosomes (300 nM) reveal a robust stimulation of MLL4-mediated H3K4 methylation by H3K18ac & H3K23ac. Data points shown are technical replicates. n = 3. Error: SD. B, middle-down MS of HEK293 cells quantifying co-enrichment of H3K4me1 with distinct H3K18/23 acetylation states with (dark red) and without (light red) histone deacetylase inhibition (butyrate). Changes in site-specific N-terminal H3 tail acetylation states as a result of butyrate treatment are shown in Fig. S1. Data points shown are biological replicates. n = 3. Error: SD. C, quantification of total H3K18 and H3K23 acetylation states by middle-down mass spectrometry. MS data shown is reported in Table S1. Data points shown are biological replicates. n = 3. Error: SD. Significance was determined by unpaired Student’s t test. NS unless otherwise designated. ∗p < 0.05, ∗∗p < 0.005, ∗∗∗p < 0.001, ∗∗∗∗p < 0.0001.

Figure 2

Acetylation of H3K18 & H3K23 promotes H3K4me3.A, in vitro methylation assays with recombinant MLL1 (10 nM) and a panel of designated nucleosomes (300 nM) reveal a robust stimulation of MLL1-mediated H3K4 methylation by H3K18ac & H3K23ac. Data points shown are technical replicates. n = 3. Error: SD. B, middle-down MS of HEK293 cells quantifying co-enrichment of H3K4me3 with distinct H3K18/23 acetylation states with (dark blue) and without (light blue) histone deacetylase inhibition (butyrate). MS data shown is reported in Table S1. Data points shown are biological replicates. n = 3. Error: SD. Significance was determined by unpaired Student’s t test. NS unless otherwise designated. ∗p < 0.05, ∗∗p < 0.005, ∗∗∗p < 0.001.

Figure 3

Methyltransferases performing non-H3K4 H3 methylation are insensitive to acetylation.A–E, in vitro methylation assays with (A) recombinant G9a (913–1193; 10 nM), (B) recombinant SETDB1 (567–1291; 25 nM), (C) recombinant PRC2 (EED/EZH2/SUZ12/AEBP/RbAp48; 50 nM), (D) recombinant ASH1L (2046–2330; 50 nM), or (E) recombinant DOT1L (1–420; 15 nM) and a panel of designated nucleosomes (300 nM) reveal minimal influence of H3 acetylation on non-H3K4 lysine methylation. F, in vitro kinase assays with recombinant Haspin (465–798; 25 nM) and a panel of designated nucleosomes (EpiCypher) reveal minimal influence of H3 acetylation state on Haspin-mediated H3T3 phosphorylation. Auto: autophosphorylation (no substrate). Significance was determined by unpaired Student’s t test. NS unless otherwise designated. ∗p < 0.05, ∗∗p < 0.005, ∗∗∗p < 0.001. Data points shown are technical replicates. n = 3. Error: SD.

Figure 4

Model for H3{K18acK23ac}-mediated stimulation of MLL-mediated H3K4 methylation. In the unmodified state (left), the H3 N-terminal tail preferentially engages nucleosomal DNA in a manner that restricts access to H3K4 by MLL family methyltransferases. Acetylation of H3K18 and H3K23 (right) releases the H3 tail from nucleosomal DNA, permitting H3K4 access by MLL family methyltransferases. Created with BioRender.com.