Structure and Conformational Dynamics of a COMPASS Histone H3K4 Methyltransferase Complex

Abstract

The methylation of histone 3 lysine 4 (H3K4) is carried out by an evolutionarily conserved family of methyltransferases referred to as complex of proteins associated with Set1 (COMPASS). The activity of the catalytic SET domain (su(var)3-9, enhancer-of-zeste, and trithorax) is endowed through forming a complex with a set of core proteins that are widely shared from yeast to humans. We obtained cryo-electron microscopy (cryo-EM) maps of the yeast Set1/COMPASS core complex at overall 4.0- to 4.4-Å resolution, providing insights into its structural organization and conformational dynamics. The Cps50 C-terminal tail weaves within the complex to provide a central scaffold for assembly. The SET domain, snugly positioned at the junction of the Y-shaped complex, is extensively contacted by Cps60 (Bre2), Cps50 (Swd1), and Cps30 (Swd3). The mobile SET-I motif of the SET domain is engaged by Cps30, explaining its key role in COMPASS catalytic activity toward higher H3K4 methylation states.

Keywords: COMPASS; MLL; cryo-EM; epigenetics; histone methylation.

Figure 1.. Architecture of core COMPASS.

(A) Views of the overall 4.0-Å cryo-EM density map of COMPASS with subunits rendered in different colors (Cps40 in pink, Cps50 in blue, Cps30 in green, SET762 in orange, Cps60 in purple, Cps25 in dark red). (B) Superimposition of 4.0-Å and 4.4-Å cryo-EM maps shown in orange and purple, respectively. (C) Structural model of the COMPASS in the same orientations and colors as shown in (A). (D) A schematic model of COMPASS structural dynamics adapts to the chromatin environment. See also Figures S1–S4 and S7.

Figure 2.. Cps50 serves as the assembly and regulatory hub for COMPASS.

(A) Interaction map of Cps50 with its partners. Regions 1–4 span sequences as follows: 1. 348–351aa; 2. 352–366aa; 3. 367–376aa; 4. 387–409aa. (B) Cryo-EM density map (colored in blue) of the Cps50 C-terminal region, which contacts most of the COMPASS subunits. (C and D) Close-up views of Cps50 interaction sites with Cps30 and SET domain (C), and Cps60 and SET domain (D). The residues involved in protein-protein interactions modeled according to EM map features, homologous structures and biochemical data are shown as stick in (C) and (D). The Cps50 residues involved in interactions are shown in their corresponding EM density (wire frame). (E and F), Representative mutagenesis analysis of H3K4 methylation profiles from yeast cells with mutated Cps50 (E) or Cps60 (F), respectively. See also Figures S4–S6.

Figure 3.. Interface details between Cps50 and its neighboring subunits.

(A) The Cps50 protein (shown as blue ribbon) connects COMPASS subunits (surface rendering). (B to D) Zoom-in views of boxed areas 1–3. (B) The conserved Cps50 N-terminus interacts with Cps30 mainly through the outer sides of 4^th and 5^th β sheets, and also associates with the stretched region (342–347aa) emanating from the Cps50 WD40 domain. (C) A short conserved motif in the Cps50 C-terminus is engulfed in the groove formed between the Cps30 5^th and 6^th β-sheets. (D) The extended Cps50 C-terminus is embraced by the Cps30 and Cps50 WD40 domains, and the Cps40 elongated helix. See also Figures S4–S6.

Figure 4.. Set1 is highly coordinated in COMPASS.

(A) Schematic (left) of SET762 with Win-motif (red) and SET domain (orange), and structure model (right) of SET762 with directly associated subunits in surface representation (right). The regions 762–798 and 851–924aa are not assigned in the structure due to intrinsic flexibility. (B and C) Zoom-in views at the interfaces between SET762 and its partners. (B) The n-SET helix region and Win-motif interact with Cps40 and Cps30, respectively. The Win-motif residues involved in interactions are shown in their corresponding EM density. (C) The catalytic SET domain is engaged in a quadripartite interface with Cps30, Cps50 and Cps60. (D) Mutation of the Win-motif region does not have an observable effect on COMPASS activity in vivo.

Figure 5.. Cps40 is important for COMPASS integrity and function.

(A) Interface among n-SET, Cps30 and Cps40. The helical segment of n-SET leans against the Cps30 WD40 domain and stabilized by the Cps40 cleft formed between its N-terminal helical stack and the C-terminal elongated helices. (B) In vivo analysis of the n-SET region important for COMPASS activity. Western blotting detection of H3K4 methylation in cells expressing full-length Set1, N-terminally truncated variants of Set1 protein (SET762, SET802 and SET830). (C) Zoom-in view of the interface between Cps40 and Cps50. (D) Recombinant MtCps50 proteins with mutations corresponding to interfacial residues on ScCps50 shown in (C) were pulled down by GST-CtCps40 protein. MtCps50: K67,W88, R105 are relative to ScCps50: R73, W94, R112. (E) Effects of ScCps40 mutants on COMPASS activity.

Figure 6.. Cps30 establishes COMPASS tri-methylation activity.

(A) View of the COMPASS structure focusing on Cps30 interactions with SET762. SET762 is divided into sub-domains and colored as in Figure 4C. Zoom-in view of Cps30-SIM loop contacts with SET-I motif (right). Conserved residues in Cps30-SIM in their corresponding EM density (wire frame), with the most proximal interacting residues from SET-I motif and Win-motif shown in stick. (B) Western blot analysis of In vivo H3K4 methylation profiles of Cps30 mutant cells targeting its Win-motif binding site and SIM integrity. (C) The SET-I region (residues 983–992) of ScSet1 is replaced with residues 2336–2344 of Drosophila melanogaster Trr (dTrr), and the resulting chimeric Set1(FL(dTrr^SET-I)) mutant cells are subjected to western blot analysis. See also Figure S7.