The Set1 N-terminal domain and Swd2 interact with RNA polymerase II CTD to recruit COMPASS

Abstract

Methylation of histone H3 lysine 4 (H3K4) by Set1/COMPASS occurs co-transcriptionally, and is important for gene regulation. Set1/COMPASS associates with the RNA polymerase II C-terminal domain (CTD) to establish proper levels and distribution of H3K4 methylations. However, details of CTD association remain unclear. Here we report that the Set1 N-terminal region and the COMPASS subunit Swd2, which interact with each other, are both needed for efficient CTD binding in Saccharomyces cerevisiae. Moreover, a single point mutation in Swd2 that affects its interaction with Set1 also impairs COMPASS recruitment to chromatin and H3K4 methylation. A CTD interaction domain (CID) from the protein Nrd1 can partially substitute for the Set1 N-terminal region to restore CTD interactions and histone methylation. However, even when Set1/COMPASS is recruited via the Nrd1 CID, histone H2B ubiquitylation is still required for efficient H3K4 methylation, indicating that H2Bub acts after the initial recruitment of COMPASS to chromatin.

Fig. 1. The Set1 N-terminal domain is needed for RNApII interaction and proper H3K4 methylation.

a Schematic of Set1 constructs used, with numbers below diagram indicating the amino acid residues from the wild-type proteins. A FLAG-tag at the N-terminus of all constructs is not shown. b Full length Set1 (Set1) and N-terminal truncated mutants (SΔ100, SΔ200) were transformed into set1Δ cells. Whole cell lysates were separated by SDS-PAGE and analyzed by immunoblotting using the indicated antibodies. Lysates from cells transformed with empty vector plasmid (set1Δ) served as a negative control. TATA-binding protein (TBP) and histone H3 were used as loading controls. c FLAG-tagged full-length (Set1) or Set1Δ200 (SΔ200) were expressed in set1Δ cells. Set1 proteins were immunoprecipitated with anti-FLAG beads (IP:FLAG) from whole cell lysates and analyzed by immunoblotting using antibodies for Ser5P (S5P CTD: 3E8) or total Rpb1 (POL II: 8WG16). Immunoprecipitations from cells transformed with empty vector plasmid (set1Δ) served as a negative control. Bottom panels show input samples, and upper panels show proteins bound to FLAG beads after precipitation. Source data are provided as a Source data file.

Fig. 2. The Set1 N-terminal region binds the Rpb1 CTD.

a Schematic of Set1 fragments fused to GAL4 BD for yeast two hybrid assay. b GAL4 BD–Set1 fusions were co-expressed with GAL4 AD–CTD in Y2H strain PJ69-4A, which has both HIS3 and ADE2 reporters for Gal4 activation. Plasmid expressing only the GAL4 AD was used as a negative control. Cells were grown for 2 days on synthetic complete (SC) media plates lacking the indicated amino acids. Five or 20 mM of 3-aminotriazol (3AT) was added to media as indicated to increase HIS3 selection stringency. Cells were also replica plated to -Ade or -Ade, -His plates to monitor the ADE2 reporter. c COMPASS containing FLAG-tagged Set1 or a derivative lacking residues 1–229 was incubated with purified RNApII. Bottom panels show input samples, and upper panels show proteins bound to FLAG beads after precipitation. Source data are provided as a Source data file.

Fig. 3. Swd2 enhances the Set1–CTD Y2H interaction.

Gal4 BD fused to full length Set1 (Set1) or amino acids 1–236 (Set1–F1, see Fig. 2) were expressed in Y2H strain PJ69-4A carrying wild-type (SWD2) or deletion (swd2Δ) alleles of SWD2. They were combined with a Gal4 AD fusion to the Rpb1 CTD (CTD) or the parent vector (Empty). Cells were spotted on the indicated selective plates. Source data are provided as a Source data file.

Fig. 4. Nrd1 CID fusion to Set1Δ200 partially restores CTD-binding and H3K4 methylation.

a Schematic diagram of the NSΔ200 fusion protein consisting of the Nrd1 CID (amino acids 1–153) and Set1Δ200. A FLAG-tag at the N-terminus is not shown. The positions of two CID mutants that disrupt CTD binding are also shown. b, c FLAG-tagged full-length Set1 (Set1), Set1Δ200 (SΔ200), or Nrd1 CID fusions (NSΔ200, D70RSΔ200, and I130RSΔ200) were transformed into set1Δ cells. Whole-cell lysates (b) and proteins immunoprecipitated using anti-FLAG beads (c) were analyzed by immunoblotting using indicated antibodies. Cells transformed with empty plasmid (set1Δ) served as a negative control. d, e ChIP-Seq heat maps of H3K4me3 (d) and H3K4me2 (e) from strains carrying full-length Set1 (Set1), Set1Δ200 (SΔ200), or Nrd1CID–Set1Δ200 fusion (NSΔ200). After normalizing to an internal spiked-in sample of S. pombe chromatin, SPMR values were mapped for individual RNApII transcriptional units and ordered according to H3K4me3 values in Set1 cells. H3K4me levels within 1500 bps from transcription start site (TSS) are shown. ChIP-seq reads from two technical repeats were averaged and plotted. f, g FLAG-tagged full-length Set1 (Set1), Set1Δ200 (SΔ200), and Nrd1CID–Set1Δ200 (NSΔ200) were transformed into set1Δ or set1Δswd2Δ cells. f Whole-cell lysates or g anti-FLAG immunoprecipitates were separated by SDS-PAGE and analyzed by immunoblotting using the indicated antibodies. Histone H3 was used as a loading control. Source data are provided as a Source data file.

Fig. 5. The Swd2 WD40 domain is critical for COMPASS function.

a Swd2–HA or Swd2–F250A–HA were expressed in an swd2Δ deletion strain. Protein levels were analyzed by immunoblotting using the indicated antibodies. Actin and Histone H3 were used as loading controls. Note that samples were run on the same blot, but intervening lanes were removed (dashed line, full blots appear in Source data). b Plasmid-encoded Swd2–HA and Swd2–F250A–HA were expressed in FLAG-Set1 containing cells that also express the endogenous untagged SWD2 gene in the chromosome. Proteins were immunoprecipitated with anti-FLAG beads (IP:FLAG) and then immunoblotted for Set1 and Swd2-HA. c The Y2H interaction between the Gal4BD–Set1 and Gal4AD–CTD fusions was tested in strain PJ69-4A expressing wild-type (WT) or mutant (F250A) Swd2. The empty Gal4 vectors served as negative controls. Activation of the Gal4-activated HIS3 reporter was tested on plates lacking histidine, with the addition of 5 mM 3AT in the last panel for additional stringency. d Analysis of recombinant baculovirus-expressed Set1/COMPASS made with WT Swd2 or the Swd2–F250A mutant. SDS-PAGE/Coomassie blue staining shows total protein, while immunoblotting with anti-FLAG and anti-HA show Set1 and Swd2 levels, respectively. e H2Bub chromatin was subjected to in vitro histone methyltransferase assays with Swd2 WT and Swd2-F250A mutant-containing complexes shown in panel (d). H3K4 methylation status was monitored by immunoblotting with indicated antibodies. Histone H3 was used as an internal loading control. Source data are provided as a Source data file.

Fig. 6. Nrd1CID–Set1Δ200 fusion does not bypass the role of Paf1C and H2Bub in H3K4me.

a Wild-type Set1 (Set1), Set1Δ200 (SΔ200), and Nrd1CID–Set1Δ200 (NSΔ200) were transformed into set1Δ, set1Δpaf1Δ, or set1Δrtf1Δ cells. Protein levels and histone methylations in whole-cell extracts were analyzed by immunoblotting. TBP and Histone H3 were used as loading controls. Parallel experiments in set1Δbre1Δ cells showed the same loss of H3K4 methylation (Supplemental Fig. 6a). b Set1 association with RNApII was quantitated by immunoprecipitating Set1 using α-FLAG beads and probing for Rpb1 CTD Ser5P and Set1. Set1 and Ser5P signal intensities from each immunoblot were quantitated using ImageJ, and Ser5P signals were normalized to those for Set1. Circles show individual experimental values, and error bars show s.e.m. (n = 5). Values from wild type Set1 transformants were normalized to 1.0 for each experiment. Source data are provided as a Source data file.