(A) Comparison of the Sir protein complements of S. cerevisiae and S. bayanus. Percent identity (ID) and similarity (sim) for each orthologous pair of proteins, as determined by BLASTP alignments, is indicated above each S. bayanus ortholog. Protein lengths in numbers of amino acids (a.a.) are given to the right of each schematic. Black boxes indicate known domains within each protein (to approximate scale), with domain names indicated below the S. cerevisiae orthologs. OIR, ORC-Interacting Region; BAH, Bromo-Adjacent Homology domain; PAD, Partitioning and Anchoring Domain; CC, Coiled-Coil. (B) Percent identities of orthologous S. cerevisiae and S. bayanus proteins as reported by BLASTP. Histogram shows percent identities of orthologous S. cerevisiae and S. bayanus proteins based on BLASTP alignments. The distribution of percent identity of orthologous protein pairs, in bins of 5 percent increments, is plotted versus the number of orthologous pairs in each bin. The bin containing Sir4 (45% identity) is indicated with an arrow.

(A) Silencing of the Sb-HMR::URA3 reporter gene (top panel) or the Sc-HMR::URA3 reporter gene (bottom panel) in S. cerevisiae/S. bayanus hybrids was assayed by growth on selective media. For each strain, a 10-fold dilution series of yeast cells was spotted onto medium counter-selective for URA3 expression (FOA), selective for URA3 expression (CSM/-Ura), or rich medium (YPD). Schematics at left show the configurations of the salient features of two species' HMR loci in each hybrid strain: silencers (ovals), mating-type cassette homology regions (blue boxes), HMR a1 ORF (red arrow), and URA3 ORF (green arrow). Gray oval indicates the presumed location of the Sb-HMR-I silencer. Presence or absence (Δ) of the S. cerevisiae (Sc) and S. bayanus (Sb) SIR4 alleles are indicated to the right of schematics. See for complete strain genotypes. (B) Silencing of the Sb-HMR::URA3 reporter gene in wild-type, SIR4/sir4Δ, or sir4Δ/sir4Δ S. bayanus diploids.

(A) Sc-SIR4 was unable to silence Sb-HMR in S. bayanus. Top row: Silencing of Sb-HMR::URA3 in an S. bayanus haploid strain bearing Sc-SIR4 integrated in place of Sb-SIR4. Bottom row: Control showing that the Sb-sir4Δ::Sc-SIR4 replacement allele could supply silencing function to Sc-HMR in an S. cerevisiae/S. bayanus hybrid. (B) RNA analysis of HMR::URA3 reporters in S. cerevisiae/S. bayanus hybrids and S. bayanus diploids. URA3 amplification values were normalized to those of actin (ACT1) for each strain. Error bars show standard deviations (n = 3).

(A) Left: Sir4 IP/Input ratios, normalized to control regions within each experiment, for the Sc-HMR-E silencer. Center: Normalized IP/Input ratios for the Sb-HMR-E silencer. Right: Normalized IP/Input ratios for the Sb-HMR::URA3 ORF. “Sc-Sir4” or “Sb-Sir4” labels indicate which species' Sir4 protein was examined by ChIP. Species' identities of both SIR4 alleles in each strain are given in parentheses, with the allele bearing the 13x-Myc tag indicated in red: (Sc/Sc), JRY9062; (Sb/Sb), JRY9063; (Sc/Sb), JRY9064 (see for complete strain genotypes). Dashed lines indicated IP/Input ratio of non-silenced control regions. Error bars indicate the standard error of the mean of all 100 bp windows covering a region. See for non-normalized IP/Input ratios and for a description of data processing. (B) ChIP-Seq profiles of Sc-Sir4 (JRY9062), Sb-Sir4 (JRY9063), and the “No tag” control (JRY9054) at two S. cerevisiae telomere regions. The ratio of IP/Input read counts for each base of a telomeric region is plotted. Diagrams indicate salient genetic features of two telomeres (see key at left) with X elements (yellow boxes), Y′ elements, and terminal repeats (TR) containing Rap1 binding sites, labeled above. TELXV-L (left panel) has an X-element-only end, whereas TELVIII-R (right panel) has an X-Y′ end. The TELVIII-R Y′ element spans nucleotide positions 556986–562456, with two helicase-encoding ORFs located between positions 558014 and 562047 (www.yeastgenome.org). For the ORFs within this Y′ element, Sc-Sir4 had a mean IP/Input ratio of 1.2, and the “No tag” control had a mean IP/Input ratio of 0.9 (the mean IP/Input ratio for all non-silenced regions, genome wide, was approximately 0.7 for both Sc-Sir4 and Sb-Sir4 ChIPs).

(A) ChIP-qPCR analysis of Sc-Sir4 versus Sb-Sir4. For each primer set, the IP/Input ratios for Sc-Sir4 (JRY9062), Sb-Sir4 (JRY9063), and the “No-tag” control (JRY9054) are shown. Error bars show standard deviations (n = 3). (B) ChIP-Seq analysis of Sc-Sir4 versus Sb-Sir4 association on an S. bayanus contig containing subtelomeric sequence (GenBank accession number AACG02000166). Hybrid strains used in this analysis were identical to those used in and Sc-Sir4, JRY9062; Sb-Sir4, JRY9063; No tag, JRY9054. Per-base IP/Input ratios, determined as in , are plotted versus contig position. We note that the terminal TG_1–3 repeats are not present in the current S. bayanus genome assembly . (C) RNA expression analysis of putative S. bayanus subtelomeric genes. Quantitative RT-PCR was performed on RNA isolated from S. bayanus wild-type (Sb-SIR4), Sb-sir4Δ::Sb-SIR4 (Sc-SIR4), or sir4Δ strains. The actin gene (ACT1) served as a euchromatic control gene. Error bars show standard deviations (n = 3).

(A) Schematic diagram depicts replacement of Sc-HMR by Sb-HMR::URA3 in S. cerevisiae, creating the Sc::(Sb-HMR::URA3) allele. Diagonal lines depict cross-overs for the HMR allele swap, with other genetic features of the two HMR loci as in . A hygromycin-resistance marker (Hyg^R) was inserted 3 kb to the right of Sb-HMR to allow targeted recombination. Shown above is percent identity (BLASTN) between S. cerevisiae and S. bayanus for the two silencers, the two cassette homology regions (light blue boxes), and the HMR a1 gene (promoter plus ORF). Note that the silencer sequences show no significant alignment by BLAST. (B) Silencing of the Sc::(Sb-HMR::URA3) reporter in SIR4/sir4Δ S. cerevisiae diploids (first row), in Sc-sir4Δ/Sb-SIR4 S. cerevisiae/S. bayanus hybrids (second row), and in Sc-sir4Δ::Sb-SIR4/Sb-SIR4 S. cerevisiae/S. bayanus hybrids (third row). Note that the change in silencing between the first and second rows could be seen only on FOA, with similar growth on CSM/-Ura. This likely reflects Sb-SIR4 dosage sensitivity, as seen in the original hybrid diploids (). (C) Control strains showing expected silencing functions of Sc::(Sb-HMR::URA3) and Sc-sir4Δ::Sb-SIR4 replacement alleles in interspecies hybrids. Silencing of the Sb-HMR::URA3 reporter gene, located in its native S. bayanus chromosomal context, in Sc-sir4Δ/Sb-SIR4 hybrids (top row), and in Sc-sir4Δ::Sb-SIR4/Sb-sir4Δ hybrids (bottom row). Note that silencing of Sb-HMR::URA3 in these hybrids was equivalent to Sc::(Sb-HMR::URA3) silencing in (B), indicating that the functions of Sb-HMR and Sb-SIR4 were largely unaffected by S. cerevisiae chromosomal context.

A multiple alignment of five species' HMR-E sequences was produced by ClustalW. Note the strong conservation of the Rap1 and Abf1 binding sites (all 7 bp of the Abf1 consensus site were conserved), with poor conservation of the ORC binding site and intervening sequences. Asterisks indicate a nucleotide is conserved across all species. Adopted and modified from .

(A) Top panel: Silencing of the Sc::(Sb-HMR::URA3) replacement allele in S. cerevisiae MATα haploids bearing either the endogenous Sc-SIR4 gene or an integrated Sb-SIR4 gene (top panel). Bottom panel: Silencing of the Sc::(Sb-HMR::URA3) replacement allele in the absence of Sc-SIR1. (B) S. cerevisiae strains bearing the Sc::(Sb-HMR::URA3) replacement allele, and either the endogenous Sc-SIR4 gene or an integrated Sb-SIR4 gene, were transformed with plasmids encoding individual S. bayanus Sir1 paralogs and assayed for silencing function (FOA/-His, CSM/-His-Ura, or CSM/-His indicate silencing reporter media also selective for maintenance of plasmids bearing the HIS3 marker). Quantification of relative silencing function, based on growth on FOA/-His, is indicated at right. Fold-change comparisons were made relative to the Sc::(Sb-HMR::URA3) Sc-SIR4 strain bearing an empty vector (row 1). We note that the CEN/ARS plasmid itself appeared to enhance Sb-HMR silencing relative to the untransformed strains (compare , “empty vector” rows, to , rows 1 and 2). However, relative comparisons among transformed strains were still possible.

(A) ChIP analysis of Sc-Orc5 and Sc-Abf1 in S. cerevisiae at Sc-HMR-E versus Sb-HMR-E. Relative enrichment of silencer sequences was verified by comparison to amplification values for a positive control region, the ARS1 replication origin, and a negative control region in the SEN1 gene (unpublished data). Note log scale on y-axis. Error bars show standard deviations (n = 3). (B) Top panel: Silencing of the Sb-HMR::URA3 reporter gene in S. cerevisiae/S. bayanus hybrids each lacking a single allele of the RAP1, ORC1, or ABF1 genes. Bottom panel: Silencing of the Sb-HMR::URA3 reporter gene in S. bayanus diploids lacking one allele of RAP1 or ORC1. Note that to properly document the silencing differences shown in (B), the CSM/-Ura plates were photographed after 5 d, whereas in CSM/-Ura plates were photographed after 3 d.

(A) Ratios of nonsynonymous to synonymous divergence (dN/dS, or ω) computed in 102 bp windows every 3 bp along alignments of the SIR4 gene from five sensu stricto species (S. cerevisiae, S. paradoxus, S. mikatae, S. kudriavzevii, and S. bayanus). Horizontal lines show the value of ω for the median SIR4 window (red solid line), the median window of all genes (blue solid line), and the limits within which 95% of all ∼1,500 windows, sampled from across coding regions of the genome, fall (dashed lines). Diamonds indicate codons having a posterior probability of (ω>1)≥0.75 (corresponding to positions given in panel C). Each of these 11 rapidly evolving codons is labeled with the inferred ancestral amino acid at that position, the amino acid number, and the amino acid present in S. cerevisiae Sir4. Labeled boxes at top indicate the locations of functional domains. (B) Table summarizing statistics of PAML's M7 versus M8 evolutionary models for SIR2, SIR3, and SIR4. For each gene, the starting nucleotide alignments were generated using sequences from the five sensu stricto species used in panel A (see for further description of PAML analysis). NS, not significant. (C) Table summarizing a subset of the Bayes Empirical Bayes (BEB) analysis for model M8 of PAML. The identities of the 11 sites with posterior probability of (ω>1)≥0.75 are shown. Nucleotide (nt) and amino acid (a.a.) positions from S. cerevisiae SIR4 are given.

Although the site-specific DNA binding proteins, ORC, Rap1, and Abf1 were largely interchangeable between S. cerevisiae and S. bayanus, the Sc-Sir4 protein showed a striking inability to function on S. bayanus silencers. Cis-regulatory information in the S. bayanus silencer was specifically tuned to the Sb-Sir4 and Sb-Kos3 silencing proteins, as shown by reconstitution experiments in S. cerevisiae (). Loss of Kos3 and changes in Sir4 in the S. cerevisiae lineage were compensated by changes in the silencers, thereby maintaining robust silencing.