Centromere-like regions in the budding yeast genome

Abstract

Accurate chromosome segregation requires centromeres (CENs), the DNA sequences where kinetochores form, to attach chromosomes to microtubules. In contrast to most eukaryotes, which have broad centromeres, Saccharomyces cerevisiae possesses sequence-defined point CENs. Chromatin immunoprecipitation followed by sequencing (ChIP-Seq) reveals colocalization of four kinetochore proteins at novel, discrete, non-centromeric regions, especially when levels of the centromeric histone H3 variant, Cse4 (a.k.a. CENP-A or CenH3), are elevated. These regions of overlapping protein binding enhance the segregation of plasmids and chromosomes and have thus been termed Centromere-Like Regions (CLRs). CLRs form in close proximity to S. cerevisiae CENs and share characteristics typical of both point and regional CENs. CLR sequences are conserved among related budding yeasts. Many genomic features characteristic of CLRs are also associated with these conserved homologous sequences from closely related budding yeasts. These studies provide general and important insights into the origin and evolution of centromeres.

Figure 1. Formation of Centromere-Like Regions revealed by ChIP–Seq.

Cse4 (red), Mif2 (blue), Ndc10 (green) and Ndc80 (orange) bind to the same discrete regions outside centromeres in Cse4 OP (bottom panels), but not in WT (top). (A–B) Highest-confidence sites are CEN-proximal; examples include CLRs 9 kb away from CEN3 (A), and 15 kb from CEN13 (B). Asterisks above signal tracks denote location of CLRs. (C–D) A few sites are far from CENs; examples include CLRs 239 kb away from CEN2 (C) and 267 kb from CEN10 (D). Signal tracks are scaled relative to the number of uniquely-mapping reads. Control samples (immunoprecipitates from untagged strains) are shown in grey. Open reading frames (ORFs) are depicted by purple boxes. Horizontal scale bars represent 0.5 kb.

Figure 2. Quantitation of protein binding in Cse4 OP strains.

(A) Kinetochore proteins show a broader distribution at centromeres when Cse4 is overproduced. Shown is ChIP-Seq signal for kinetochore proteins in Cse4 OP (blue) compared to WT (red) at CEN2 (left), CEN5 (middle) and CEN10 (right). Aggregated signal plots depict the log ratio of read enrichment for four kinetochore components, centered at the CEN, on log 2 scales. (B) ChIP-qPCR confirms the presence of kinetochore proteins at CLRs in Cse4 OP, not in WT. Individual protein enrichments for 6 CLRs were averaged and compared to CEN3 binding levels. Normalized enrichment ratios (means in arbitrary units (a.u.)+/−SEM) were plotted on a log 10 scale. A normalized enrichment of 1 indicates no enrichment over a negative control region not enriched for kinetochore proteins. (C) ChIP-qPCR in a cohesin-deficient mcd1-1 background highlights CLR formation in Cse4 OP despite the abrogation of the pericentric intramolecular C loop. Individual protein enrichments for 6 CLRs were averaged and compared to CEN3 binding levels. Normalized enrichment ratios (means in arbitrary units (a.u.)+/−SEM) were plotted on a log 10 scale.

Figure 3. CLRs confer centromere function on plasmids and chromosomes.

(A) Top: Representation of ARS, CLR and CEN plasmids. Bottom: Transformation plates from strains carrying different plasmids. (B) Doubling times in selective medium of strains carrying different plasmids (Means+/−SEM). (C) Fraction of plasmid-containing cells after growth in non-selective media for strains bearing various plasmids (Means+/−SEM). P-values were computed using MCMC simulations (* p<0.05, ** p<0.01, *** p<0.001,. p = 0.07) (B–C). (D) Top: Depiction of pedigree assay. Unbudded mother cells contain a conditional CEN3 (ON raffinose; OFF galactose) or not (ON raffinose or galactose). Bottom: Fraction of budded daughter cells 12 h after transfer of mother cells to plates containing galactose (Means+/−SEM). ON and OFF indicate the presence of conditional CEN3. For the four left-most bars, similar values were obtained for strains with conditional CEN3 grown in raffinose (ON). A time course is shown in Figure S4. (E) Fraction of cells that segregated properly their GFP-labelled chromosome 3, as visualized by the presence of a GFP dot in each cell (Means+/−SEM). P-values have been calculated using Fisher's Exact Test (* p<0.05, ** p<0.01, *** p<0.001) (D–E).

Figure 4. Genomic features of CLRs.

(A) Comparison of CLRs and CENs. Average value refers to the fraction of CLRs or CENs located within 25 kb of a CEN or within 5 kb of an ARS, tRNA or retrotransposon. Mean intergenic indicates the average length of intergenic regions. Max AT represents the mean AT content of the most AT-rich 90-bp stretch of DNA. (B) CLRs can be separated from other genomic regions, using Discriminant Analysis (DA). Scores of CLRs (blue) and negative control regions (red) are plotted according to their discriminant function scores (Table 2). (C) Centromere proximity is a major contributor to variability among CLRs, as revealed by Principal Component Analysis (PCA). Scores of CEN-proximal (<25 kb, blue) and CEN-distal (>25 kb, red) CLRs are plotted relative to the first and second principal components (Table 3), along with a 95% confidence ellipse. (D) Conservation of CLR sequences among organisms with point CENs (blue), but not with fungi bearing regional CENs (black). Nucleotide blast (Blastn) was performed for 23 CLRs and 160 random intergenic regions. Mean BLAST scores are reported, with the percent of hits with a score over 45 (E<0.05) in parentheses. (E) Given our data and the confinement of CLR sequences to budding yeast bearing point centromeres, we proposed a modified version of the current model of centromere evolution (originally postulated in [1]), from regional to point CENs, to account for CLRs. CLRs would represent evolutionary remnants from regional CENs. Some AT-rich CEN repeats would have diverged but still retained the ability to bind Cse4 and other kinetochore proteins weakly, giving rise to the low-affinity CLRs observed in this study.