Molecular basis of global promoter sensing and nucleosome capture by the SWR1 chromatin remodeler

Abstract

The SWR1 chromatin remodeling complex is recruited to +1 nucleosomes downstream of transcription start sites of eukaryotic promoters, where it exchanges histone H2A for the specialized variant H2A.Z. Here, we use cryoelectron microscopy (cryo-EM) to resolve the structural basis of the SWR1 interaction with free DNA, revealing a distinct open conformation of the Swr1 ATPase that enables sliding from accessible DNA to nucleosomes. A complete structural model of the SWR1-nucleosome complex illustrates critical roles for Swc2 and Swc3 subunits in oriented nucleosome engagement by SWR1. Moreover, an extended DNA-binding α helix within the Swc3 subunit enables sensing of nucleosome linker length and is essential for SWR1-promoter-specific recruitment and activity. The previously unresolved N-SWR1 subcomplex forms a flexible extended structure, enabling multivalent recognition of acetylated histone tails by reader domains to further direct SWR1 toward the +1 nucleosome. Altogether, our findings provide a generalizable mechanism for promoter-specific targeting of chromatin and transcription complexes.

Keywords: +1 nucleosome; DNA sliding; H2A.Z; SWR1; chromatin remodeler; cryo-EM; histone acetylation; histone exchange; histone reader; promoter.

Figure 1.. Cryo-EM structure of SWR1 bound to free DNA

(A) Cryo-EM structure of the DNA-bound SWR1 complex, in three different orientations. See Figures S2A–S2J for corresponding cryo-EM map and modeling details. (B) Superposition of the DNA in the SWR1-DNA complex with the nucleosomal SHL2 DNA from the SWR1-nucleosome structure, after aligning to the ATPase domain of Swr1. (C) Domain organization of the Swr1^ATPase, with the Snf2-family specific elements – Suppressor Helix (SuppH), ‘Gating’ helix, and Brace – colored in bright green, pink, and orange, respectively. Swr1 residue numbering is shown above the domain map, and the extent of the two main N-terminal and C-terminal lobes that comprise the ATPase structure are depicted at the top. The ‘insert’ domain incorporates into the RuvBL1/2 core and is not part of the ATPase. (D) Comparison of the conformations of the Snf2-family specific ATPase elements – SuppH, Gating, and Brace – observed in the DNA-bound Swr1^ATPase (left) and nucleosome-bound Swr1^ATPase (right). The color scheme is the same as in (C). The two structures are aligned by their RecA2 ATPase subdomains. (E) Superposition of the Swr1^ATPase domains from the DNA-bound and nucleosome-bound structures after aligning to the RuvBL core, colored according to the key below. (F) Comparison of the DNA path along the surface of Swc6 in the DNA-bound (cyan) and nucleosome-bound (green) states. The Swc6 surface is colored according to electrostatic potential, from −10 (red) to +10 (blue) kcal/(mol•e). See also Figures S1, S2, S3, Table 1, and Video S1.

Figure 2.. Swc2 organizes the SWR1 core and orients it on the nucleosome.

(A) Full model of the nucleosome-bound SWR1 core. All elements are modeled from cryo-EM data, except for the Swc2-H2A.Z-H2B subcomplex (region I, top), which was predicted using AlphaFold2. Roman numerals are used to label different regions of the Swc2 subunit. Transparent surfaces depict solvent-accessible molecular surfaces, not experimental cryo-EM density. (B) Domain organization of the Swc2 subunit, with different regions labeled with roman numerals as in (A). Modeled regions are colored in bright yellow and intervening disorded loops are colored in light yellow with grey outlines. Contacts between Swc2 and other proteins or DNA are shown below, using the same color scheme as (A). (C-D) Closeup (C) and sequence alignment (D) of the primary nucleosome-interacting region of Swc2 (region III). Conserved features are highlighted with matching colors on the Swc2 model (C) and sequence alignment (D). The H2A-H2B dimer is colored according to electrostatic potential, from −10 (red) to +10 (blue) kcal/(mol•e). Clustal consensus symbols are shown below the sequence alignment: * identical, : strongly similar, . weakly similar. (E) Superposition of the Swc2^throttle in the nucleosome-bound (yellow) and DNA-bound (orange) states. The C-terminal end of the Swc2^throttle that interacts with SHL1 DNA in the nucleosome-bound state is colored according to electrostatic potential, from −10 (red) to +10 (blue) kcal/(mol•e). (F) Closeup of the interactions made by the Swc2 Arp6-anchor and adjacent Swc2 C-terminal region (CTR). See also Figures S1 and S4, Table 1, and Video S1.

Figure 3.. Structure of the Swc3-Swc2 subcomplex and interactions with nucleosome and linker DNA.

(A-B) Domain organization (A) and full structural model (B) of the Swc3-Swc2 subcomplex. “Dock”, docking domain; PH, pleckstrin homology; LH, long helix; HTH, helix-turn-helix; LC, linker cradle; NH, nucleosome helix; CTR, C-terminal region; βW, β-wedge. (C) Location of the Swc3-Swc2 subcomplex within the structure of the nucleosome-bound core of SWR1 (left), and (right) closeup view of the interaction between Swc3-Swc2 and the nucleosome, including the H3 N-terminal tail and nucleosomal DNA at SHL1. Lysine residues in the Swc3 nucleosome helix (NH) that are near the SHL1 DNA are shown as blue spheres. (D) Isolated view of the nucleosome-binding components of SWR1, dashed line represents the boundary between regions observed in the cryo-EM structure (left) and regions predicted and extrapolated from the cryo-EM structure. (E) Isolated view of Swc3 binding to nucleosomal and NFR DNA, with Swc3 colored by surface electrostatics from −10 (red) to +10 (blue) kcal/(mol•e). (F) EMSA with bacterially expressed Swc3 LH fragment and free DNA fragments of 20–100 basepairs (bp) in length, resolved by native PAGE. Right, Quantification of EMSA fraction bound for each DNA fragment bound versus Swc3 LH concentration is plotted on the right. (G) EMSA of purified native SWR1 complex from WT (top) or swc3Δ mutant (bottom) strains with a 244 bp DNA fragment, resolved by native agarose gel. Quantification of EMSA fraction bound versus SWR1 concentration is plotted on the right. See also Figures S4 and S5.

Figure 4.. Swc3 is required for promoter-specific H2A.Z deposition in cells.

(A) Purified endogenous SWR1 complexes from WT, swc3Δ, and swc2(CTRΔ) yeast strains, resolved by SDS-PAGE. Red arrow indicates the Swc3 band, which is missing in both mutant complexes (red arrow). (B) Histone exchange assay with purified WT and mutant SWR1 complexes (see Methods for details). Unreacted nucleosomes containing two H2A-H2B dimers are separated from reacted nucleosomes containing one or two copies of the tagged H2A.Z-H2B dimer (AZ or ZZ, respectively) via native PAGE. (C) Illustration of the SWR1 construct used for single particle tracking, harboring a Swr1-HaloTag fusion. (D-E) Chromatin-bound fractions (D) and corrected residence time of the chromatin-bound fraction (E) of Swr1-HaloTag, derived from live cell SPT in WT and mutant SWR1 cells (see Methods for details). (F-H) Averaged distribution of strand-separated 5’ ends of ChIP-exo sequencing tags (exonuclease stop sites) of Swr1 (F) or Htz1 (G-H) in WT or mutant SWR1 strains are distributed around +1 nucleosome midpoints for 4779 constitutive genes (F-G) or the indicated gene subset (H). The x-axis intervals are oriented such that transcription proceeds to the right, with same-strand data oriented with 5’ to 3’ from left to right and opposite-strand data inverted (right to left is 5’ to 3’). Data are normalized by setting a 50 bp interval in the NFR (from about −100 to −150 bp from the +1 nucleosome midpoint) to be equal across samples. Plot values are provided in Table S5. See also Figure S6.

Figure 5.. Architecture of the N-SWR1 subcomplex and role in acetyl-histone readout

(A) Domain layout of the Swr1 subunit showing the division between the N-terminal (light grey) and C-terminal (dark grey) regions that form the basis for the assembly of the N-SWR1 (green background) and C-SWR1 (mauve background) subcomplexes, respectively. Contacts with other subunits in the complex are depicted with colored bars for N-SWR1, and the C-SWR1 subunit architecture is approximated with colored shapes. (B) Full structural model of the SWR1 complex, divided into the N-SWR1 integrative model (left, green background) and C-SWR1 cryo-EM structure (right, mauve background). Dashed lines indicate disordered regions. AcK, acetyl-lysine. (C) Alternative view of the N-SWR1 model including only components directly associated with the Swr1 HSA region. (D) Isolated view of the subcomplex comprising Swc5, Actin, Swr1 HSA, and Swr1 C-terminus. (E) AlphaScreen interaction assay performed with Swr1 and nucleosome core particles containing specific acetylation marks (see Methods for details). The tetra-acetylated (tetra-ac) H3 and H4 nucleosomes correspond to H3K(4,9,14,18)ac and H4K(5,8,12,16)ac acetylation patterns, respectively. Error bars indicate standard deviation of 3 replicates, and the horizontal green bar marks the average ± standard deviation obtained for the unmodified nucleosomes. * p value < 0.05, ** p value < 0.01 (two-tailed unpaired t-test with unequal variance). (F) Cryo-EM densities corresponding to the C-SWR1 and N-SWR1 modules in the nucleosome-bound complex. The N-SWR1 region was separately lowpass filtered to 30 Å (see Figure S3B). (G) Composite model of the full nucleosome-bound SWR1 complex, generated by approximate fitting of the main structured part of N-SWR1 (shown in (C)) into the low-resolution N-SWR1 cryo-EM envelope (shown in (F)), and superposing this onto the C-SWR1 structure (from Figure 2A). The recognition of acetylated H3 (orange) and H4 (magenta) tails by Yaf9 and Bdf1 reader domains, respectively, is conceptually illustrated. See also Figure S7.

Figure 6.. Mechanism of promoter sensing and nucleosome capture by SWR1

(A) The SWR1 complex encounters the NFR via 3D diffusion. A typical chromatin architecture for a constitutive yeast gene is depicted, with a 150 bp NFR separating upstream and downstream gene bodies ~2 kbp in length. An arrow marks the transcription start site (TSS) and direction. Gene body nucleosomes are modeled with 20 bp of linker DNA between them. (B) SWR1 engages the NFR through the Swr1^ATPase domain, stabilized through additional DNA interactions with Swc6 and likely Swc3. SWR1 can undergo bidirectional 1D diffusion on the NFR, towards the +1 or −1 nucleosome. Acetylation patterns specific to the +1 nucleosome are recognized by the reader domains within the Bdf1 and Yaf9 subunits of SWR1, directing the complex towards the +1. (C) SWR1 encounters the +1 nucleosome, and Swc2 recognition of the H2A-H2B acidic patch and other nucleosomal features likely drives the transition of the Swr1^ATPase from binding the NFR to binding SHL2 of the nucleosome. Swc3 engages SHL1 and the H3 N-terminal tail. Acetyl histone binding by Bdf1 and Yaf9 anchor SWR1 to the nucleosome during this transition. (D) The binding of the Swr1^ATPase at SHL2 in a closed conformation results in DNA bulging at SHL3.5 that destabilizes histone-DNA interactions to facilitate subsequent DNA unwrapping, while binding of Swc6 to SHL6 induces further unwrapping of the NFR-distal side of the nucleosome. The anchoring of Swc3 near the dyad through DNA and histone interactions stabilize the unwrapped state of the nucleosome.