Mutational analysis of the Sir3 BAH domain reveals multiple points of interaction with nucleosomes

Abstract

Sir3, a component of the transcriptional silencing complex in the yeast Saccharomyces cerevisiae, has an N-terminal BAH domain that is crucial for the protein's silencing function. Previous work has shown that the N-terminal alanine residue of Sir3 (Ala2) and its acetylation play an important role in silencing. Here we show that the silencing defects of Sir3 Ala2 mutants can be suppressed by mutations in histones H3 and H4, specifically, by H3 D77N and H4 H75Y mutations. Additionally, a mutational analysis demonstrates that three separate regions of the Sir3 BAH domain are important for its role in silencing. Many of these BAH mutations also can be suppressed by the H3 D77N and H4 H75Y mutations. In agreement with the results of others, in vitro experiments show that the Sir3 BAH domain can interact with partially purified nucleosomes. The silencing-defective BAH mutants are defective for this interaction. These results, together with the previously characterized interaction between the C-terminal region of Sir3 and the histone H3/H4 tails, suggest that Sir3 utilizes multiple domains to interact with nucleosomes.

FIG. 1.

(A) A screen for second-site suppressors of the Sir3 A2G silencing defect. Expected results for wild-type Sir3, the Sir3 A2G mutant, and a putative suppressor (sup X) of Sir3 A2G are depicted. Strain XRY36 was used for the screen. 5-FOA^r, 5-FOA resistant. (B) H3 D77N and H4 H75Y mutants can suppress the silencing defects of some Sir3 Ala2 mutants. MATα sir3Δ sir1Δ strains carrying the indicated alleles of Sir3 Ala2 in combination with either a plasmid encoding wild-type (WT) H3/H4 (top panels) or H3 D77N/H4 (middle panels) or H3/H4 H75Y (bottom panels) were assayed for silencing at HML and HMR, using strains VSY29, VSY30, and VSY31, respectively. Silencing at HML::URA3 was assessed on 5-FOA medium, and silencing at HMR by a mating assay. Silencing at a telomere was assessed by pink colony color on low-adenine medium.

FIG. 2.

Sir3 occupancy at silent regions increases in the presence of the H3 D77N mutant. (A) Anti-Sir3 antibodies were used in ChIP experiments to determine localization of Sir3 at HMR E and HMR-a1 (histograms). A schematic representation of the HMR locus and the regions analyzed is shown. The bar graphs show average Sir3 occupancy with vector, wild-type Sir3, or Sir3 A2G plasmids in strains with wild-type H3 (VSY29), H3 D77N (VSY30), or H4 H75Y (VSY31). (B) A schematic representation of TEL-VIR and the regions analyzed is shown. Sir3 occupancy at 0.67 kb and 5 kb from the telomere is presented in the bar graphs. Strains used were the same as for the experiment whose results are shown in panel A. The data shown are the averages and standard deviations of the results of two PCR amplifications of one representative IP. Sir3 localization in the presence of wild-type histone H3, the histone H3 D77N mutant, or the histone H4 H75Y mutant is depicted. WT, wild type.

FIG. 3.

Suppression by histone H3 D77N mutant is not mimicked by a loss of K79 methylation. (A) Strains W303-1a, UCC7014 (dot1Δ), VSY38 (wild-type H3), and VSY39 (H3 D77N) were assayed for levels of dimethyl K79 and H3 by Western blotting. The lower band marked by an asterisk in the Western blot for dimethylated K79 (me2 K79) is a nonspecific band recognized by the antibody and used as an internal loading control. α, anti. (B) Genetic interaction between dot1Δ and H3 D77N mutations. A DOT1 sir3Δ strain (VSY41) and a dot1Δ sir3Δ strain (VSY43) carrying various plasmid-borne alleles of Sir3 in combination with either a wild-type histone H3/H4 plasmid or a histone H3 D77N/H4 plasmid were assayed for telomeric silencing by growth on 5-FOA plates. WT, wild type.

FIG. 4.

(A) Schematic representation of the PCR mutagenesis screen used to isolate mutants with mutations in the Sir3 BAH domain. In vivo recombination between PCR-mutagenized DNA fragments and a gapped plasmid generated mutants with mutations in the Sir3 BAH domain. The mutants were screened for a telomeric silencing defect by using a TEL-URA3 reporter and identifying 5-FOA-sensitive colonies. For details, see Materials and Methods. 5-FOA^s, 5-FOA sensitive; 5-FOA^r, 5-FOA resistant; GAD, Gal4AD. (B) The telomeric silencing defect in strain PYY4 is shown for four representative mutants by lack of growth on 5-FOA medium. Sir3 Y207C is the only mutant that showed some growth on 5-FOA. (C) The results of mating tests in strains JCY3 and JCY4 show that some of the BAH domain mutants have a silencing defect at HML, while none are defective in silencing at HMR. WT, wild type.

FIG. 5.

(A) Suppression of the silencing defect of the Sir3 BAH domain mutants is due to the H3 D77N and H4 H75Y mutations. Results for six representative Sir3 mutants are depicted. Silencing was measured at HML::URA3 by growth on 5-FOA and at HMR by mating in the wild-type H3/H4 (strain VSY29), the H3 D77N mutant (VSY30), and the H4 H75Y mutant (VSY31). (B) Suppression of the silencing defect at HMR of some of the Sir3 BAH domain mutants by a histone H4 K16R mutant (strain PYY12). WT, wild type.

FIG. 6.

Sir3 BAH domain binds to nucleosomes. (A) The N terminus of Sir3 is important for binding to nucleosomes. Wild-type (W303-1a) nucleosomes were mixed with GST, Sir3^1-219-GST, Sir3^{A2G 1-219}-GST, or Sir3^{A2Q 1-219}-GST bound to glutathione-Sepharose beads. One-third of the nucleosomes retained by the GST-tagged protein on the beads was visualized by Western blotting with anti-H3 antibody. “Input Nucl” refers to 1% of nucleosomes used for the binding reaction. (B) N-terminal acetylation of Sir3 stimulates binding to nucleosomes. Wild-type (W303-1a) nucleosomes were incubated with increasing concentrations of wild-type Sir3^1-219-GST purified from ARD1 (W303-1a) and ard1Δ (JCY5) strains. One-third of the nucleosomes retained on the beads was visualized as described for panel A. (C) Loss-of-function Sir3 BAH domain mutants are unable to bind to nucleosomes, whereas the D205N mutant binds well. Nucleosomes from W303-1a were incubated with wild-type Sir3^1-219-GST purified from E. coli and with the indicated Sir3 mutants. One-fourth of the nucleosomes retained on the beads was visualized as described for panel A. “Input Nucl” refers to 0.5% of nucleosomes used for the binding reaction. The asterisk indicates the position of a Sir3-GST degradation product. WT, wild type; α, anti.

FIG. 7.

Location of the mutated residues in Sir3 BAH domain and in the nucleosome. (A) Sir3 BAH domain crystal structure (Protein Data Bank accession number 2FVU) was used to pinpoint the location of the mutated residues by using PyMOL. The region 1 mutant is colored orange, region 2 mutants are colored blue, and region 3 mutants are colored red. (B) The crystal structure of the nucleosome from Xenopus laevis (Protein Data Bank accession number 1KX5) was used to visualize the location of the suppressor mutations in histones H3 and H4 by using PyMOL. H3 is colored blue, H4 yellow, and H2A and H2B green. Residues depicted as spheres are H3 D77 in light pink, H3 K79 in brown, H4 H75 in red, and H4 K16 in purple.