Fig 1.
Expressing Rec8 in place of Scc1 impairs the mitotic cell cycle and sister chromosome cohesion.
(A) A diagram of the mitotic and meiotic cohesin complexes. Mitotic cohesin holds replicated sister chromosomes together in mitosis. We investigated the ability of the meiotic cohesin to support mitosis. (B) The fitness of Rec8-expressing cells is 55% of a wild-type strain expressing Scc1. The Rec8-expressing strain expressed a fluorescent marker (PACT1-mCitrine) and was competed against wild type. The fitness of Rec8-expressing cells relative to wild type was calculated as changes in the ratio of these two strains over multiple generations. In the right panel, the darker gray points represent the values of three biological replicates, and the thinner gray bar represents one standard deviation on each side of the mean of these measurements. (two-tailed Student t test, **p < 0.01). (C) The Rec8-expressing strain cannot maintain sister chromosome cohesion in mitosis. All the strains (PSCC1-REC8, PSCC1-SCC1, and scc1Δ) carried a PGAL1-SCC1 copy integrated in the genome to allow the acute effect of altered kleisin expression to be analyzed. To examine sister chromosome cohesion in a single cell cycle, Scc1 expression was switched off in G1-arrested populations by transferring cells to YEP containing 2% raffinose and α-factor. Cells were released to YPD containing benomyl to resume cell cycle and held in mitosis. Two different patterns of sister chromosome cohesion are shown: A budded cell with a single GFP dot represents functional sister chromosome cohesion; a budded cell with two GFP dots represents lack of sister chromosome cohesion. A single GFP dot is marked with a white arrow. At least 100 cells were imaged at each time point in each experiment. Three biological repeats were performed at each time point and for each strain; the right panel showed mean and standard deviation for the wild type (blue), Rec8-expressing (red), and scc1Δ (brown) strains. The scale bar is 5 μm. Data associated with this figure can be found in S1 Data. (D) The Rec8-expressing strain progressed through S phase faster and through mitosis slower. All strains were cultured as in Fig 1C, but cells were released in YPD to allow completion of the first cell cycle and entry into the second. Samples were collected at the indicated time points to examine DNA content by flow cytometry. Cell cycle profiles at 30 and 180 minutes are labeled in red and orange, respectively. For the cells expressing Scc1, cells are predominantly in S phase at 30 minutes, predominantly in G2 and mitosis from 45 to 75 minutes, and return to G1 at 90 minutes. Cells that express Rec8 have a long delay in mitosis, which is abolished by removing the spindle checkpoint. For examining the progression through mitosis, the timepoints from 75 to 180 minutes are highlighted by a square bracket on the y-axis. Chr.5-GFP, Chromosome 5-green fluorescent protein; GFP, green fluorescent protein; Rec8, recombination 8; Scc1, sister chromosome cohesion 1; Smc, structural maintenance complex; YEP, yeast extract and peptone; WT, wild type; YPD, yeast extract, peptone, and dextrose.
Fig 2.
Rec8 is unstable and shows reduced binding to mitotic chromosomes.
(A) The Rec8 protein level is lower than Scc1 in mitosis. Both SCC1 and REC8 were expressed from the SCC1 promoter and fused to a triple hemagglutinin tag (3xHA) at their C termini in a strain that also carried PGAL1-SCC1. To follow kleisin protein levels in a single mitotic cycle, expression from the GAL1 promoter was repressed and cells were released from a G1 arrest and allowed to proceed through the cell cycle as in Fig 1D. Cells were collected at the indicated time points, and cell extracts were obtained by alkaline lysis prior to analysis by western blotting. Hxk1 was used as a loading control. In the blot, the upper arrow marks the size of full-length protein and the bottom arrow marks the size of cleavage product. The three colored points represent the values of three biological replicates, and the dark gray bar represents one standard deviation on each side of the mean of these measurements. The statistical significance between data from the Rec8-expressing strain and wild type was calculated by two-tailed Student t test, **p < 0.01. Raw images associated with Fig 2A can be found in S1 Raw Image. Data associated with this figure can be found in S1 Data. (B) The instability of Rec8 in mitosis depends on separase activity. Cells were grown to log phase in YPD at 30 ˚C and held in mitosis by addition of benomyl. To check protein stability, cycloheximide was added to the cultures to inhibit protein synthesis. Cells were collected every 30 minutes to examine protein levels by western blotting. Both Scc1 and Rec8 were detected by an anti-HA antibody. The darker gray points represent the values of at least three biological replicates, and the thinner gray bar represents one standard deviation on each side of the mean of these measurements. The half-life of Rec8 is increased in cells expressing a temperature-sensitive mutant of Esp1 (esp1-1) (two-tailed Student t test, **p < 0.01). Raw images associated with Fig 2B can be found in S1 Raw Image. Data associated with this figure can be found in S1 Data. (C) The level of chromosome-bound Rec8 is lower than that of Scc1 in mitosis. Strains were released from a G1 arrest, proceeded synchronously through one cell cycle, and then were arrested in mitosis in YPD containing benomyl. The chromosome-bound kleisin protein was immunoprecipitated using an anti-HA antibody. Chromatin lysates were prepared from wild type, a Rec8-expressing strain, and a strain without HA tag as negative control. The level of chromosome bound kleisin at the known cohesin binding sites was measured by the amount of DNA that associated with the immunoprecipitated kleisin. DNA was measured by qPCR and expressed as the fraction of material compared to the total chromatin lysate (shown in the y-axis). Four genomic loci on Chromosome 4 are shown. The colored points represent the values of five biological replicates and the dark gray bar represents one standard deviation on each side of the mean of these measurements. The statistical significance between data from the Rec8 strain and wild type was calculated by two-tailed Student t test, **p < 0.01. The bottom panel shows the ChIP-Seq data of Scc1 and Rec8 at the corresponding cohesin binding sites, under the same conditions. Data associated with this figure can be found in S1 Data. (D) ChIP-Seq analysis of Scc1 and Rec8 binding in mitosis. Sample preparation and immunoprecipitation were done as in Fig 2C. The immunoprecipitated DNA bound by Scc1 or Rec8 were examined by whole genome sequencing. The amount of immunoprecipitated DNA is expressed as reads per million (RPM) calibrated to the reference, Schizosaccharomyces pombe genomic DNA immunoprecipitated by the Scc1 ortholog, Rad21. The calibrated signal of ChIP-Seq data (relative to a control S. pombe sample) representing the degree of enrichment of kleisin (Scc1 in blue and Rec8 in red) and the difference in enrichment between the two kleisins (gray: Scc1’s signal is more than Rec8’s; orange: Rec8’s signal is more than Scc1’s) is visualized by the Integrated Genomic Viewer [33]. Chromosomes 3 and 15 are shown as examples of chromosomes with different sizes and different degrees of peri-centromeric Rec8 binding, with an expanded view of 20 kb DNA on each side of a centromere in the bottom panel. (E) Rec8 loads poorly on G1 chromosomes compared to Scc1. To overexpress kleisins in G1, the PGAL1-SCC1-HA and PGAL1-REC8-HA strains were arrested in YEP containing 2% galactose and α-factor. ChIP and qPCR were performed as described in Fig 2C. The darker points represent the values of three biological replicates and the darker gray bar represents one standard deviation on each side of the mean of these measurements. The statistical significance between data from the Rec8-expressing strain and wild type was calculated by two-tailed Student t test, **p < 0.01. Data associated with this figure can be found in S1 Data. CEN, centromere; ChIP-Seq, chromatin immunoprecipitation sequencing; CYH, cycloheximide; Esp1, extra spindle pole bodies 1; GAL1, galactose metabolism 1; HA, hemagglutinin; Hxk1, hexokinase; qPCR, quantitative polymerase chain reaction; Rad21, radiation sensitive 1; Rec8, recombination 8; Scc1, sister chromosome cohesion 1; YEP, yeast extract and peptone; YPD, yeast extract, peptone, and dextrose.
Fig 3.
Experimental evolution improves the fitness of Rec8-expressing populations.
(A) Schematic of the experimental evolution of 15 independent populations forced to use Rec8 in mitosis for 1,750 generations. All the evolved populations increased fitness during evolution (shown in the right panel). The mean relative fitness of each population during evolution is shown as an individual colored line. The average fitness of all 15 evolved populations is shown as a black line. The fitness of wild type relative to the Rec8-expressing ancestor is indicated as a black dashed line. Two evolved populations showing fitness near wild type are labeled P13 and P15. (B) The fitness of 15 evolved populations relative to the Rec8-expressing ancestor at generations 375 and 1,750. The relative fitness of each evolved population was measured by competing it against a fluorescently labeled ancestor in YPD. The darker gray points represent the values of three biological replicates, and the thinner gray bar represents one standard deviation on each side of the mean of these measurements. Genes in the three functional modules that mutate in at least six evolved populations at generation 1,750 are shown. Mutations in the different functional modules are color-coded: transcriptional mediator complex (red), G1-to-S cell cycle regulators (blue), and cohesin and its regulators (black). Genes mutated twice within the same population were marked with an asterisk (*). The black dashed line shows the fitness of wild type relative to the Rec8-expressing ancestor. The details of these mutations are available in S1 and S2 Tables. (C) Summary of functional modules that had acquired fixed mutations in more than six populations at generation 1,750. The x-axis shows the number of populations that acquire a mutation in any specified gene at generation 375 and 1,750. The y-axis shows mutated genes grouped by their functions: genes involved in the transcriptional mediator complex in blue, cohesin related genes in black, and the G1-to-S cell cycle regulators in orange. Mutations causing early stop codon are shown in dark red and single nucleotide changes are shown in pink. (D) Summary of changes in chromosomal copy number of all 15 evolved populations. The copy number of each chromosome was calculated by normalizing the median read depth of each chromosome to the median read depth over the entire genome. The results of 15 evolved populations at generations 375 and 1,750 are shown here: gray marks one copy, dark red marks two copies, and pink marks 1.25–1.75 copies, suggesting part of the population was disomic. The results of five ancestors are shown in S6A Fig. Data associated with Fig 3A–3D can be found in S1 Data. Rec8, recombination 8; YPD, yeast extract, peptone, and dextrose.
Fig 4.
Reconstructing individual evolved mutations increases the fitness of the Rec8-expressing ancestor.
(A) The effect of single evolved mutations and deletion of genes encoding subunits of the Cdk8 complex on fitness of the Rec8-expressing ancestor. (B) The effect of deleting CLN2 and MBP1 on fitness of the Rec8-expressing ancestor. (C) The effect of single evolved mutations in genes that encode other cohesin components or separase and an extra copy of SCC3 on fitness of the Rec8-expressing ancestor. In 4A-4C, each single evolved mutation was reconstructed in the Rec8 ancestral strain used in the evolution experiment. The relative fitness of reconstructed strains to the ancestor was measured by competing it against a fluorescently labeled Rec8 ancestor. The darker gray points represent the values of three biological replicates, and the thinner gray bar represents one standard deviation on each side of the mean of these measurements. The fitness of the wild-type strain, labeled as SCC1, is shown in each panel. The statistical significance between data from the Rec8-expressing strain and each mutation-reconstructed strain was calculated by two-tailed Student t test, **p < 0.01. (D) esp1 evolved mutations (esp1-P8 and esp1-P15) are hypomorphic. A CEN plasmid carrying ESP1 was transformed into a wild-type strain, two ancestors (Anc), and two evolved populations that had acquired esp1 mutations (P8 and P15) at generation 375. Cells were subjected to 10-fold serial dilutions and spotted on YPD plates to assay growth. Cells transformed with an empty plasmid (pRS413) served as control. (E) The effect of deleting one copy of SCC3 on fitness of the evolved populations with disomic Chromosome 9 at generation 1,750. Data associated with Fig 4A, 4B and 4C, and 4E can be found in S1 Data. Cdk8, cyclin-dependent kinase 8; CEN, centromere; CLN2, cyclin 2; esp1, extra spindle pole bodies 1; MBP1, mlul-binding protein; Rec8, recombination 8; SCC, sister chromosome cohesion; YPD, yeast extract, peptone, and dextrose.
Fig 5.
Adaptive genetic changes improve sister chromosome cohesion in Rec8-expressing cells.
(A) Individual adaptive genetic changes partially improve sister chromosome cohesion. Deletions of genes in the Cdk8 complex, adaptive mutations in cohesin and its regulator, two copies of SCC3, and deletion of genes that regulate G1-to-S transition were reconstructed individually in the strain used for assaying sister chromosome cohesion. Cells were prepared as in Fig 1C, and the percentage of cells with two GFP dots in populations arrested in mitosis (150 minutes after releasing from G1) is shown. The darker gray points represent the values of three biological replicates, and the thinner gray bar represents one standard deviation on each side of the mean of these measurements. The statistical significance between data from the Rec8-expressing strain and each mutation-reconstructed strain was calculated by two-tailed Student t test, *p < 0.05, **p < 0.01. (B-F) Relative fitness and sister chromosome cohesion of double mutants are shown: ssn2Δ and smc1-P15 (B), mbp1Δ and smc1-P15 (C), ssn2Δ and mbp1Δ (D), ssn2Δ and esp1-P15 (E), mbp1Δ and esp1-P15 (F). The blue dashed line represents the expected fitness if two mutations contribute additively, and the shaded region represents the standard error of that expectation. Data associated with Fig 5A–5F can be found in S1 Data. Cdk8, cyclin dependent kinase 8; GFP, green fluorescent protein; Rec8, recombination 8; SCC3, sister chromosome cohesion 3.
Fig 6.
Delaying genome replication partially improves sister chromosome cohesion in Rec8-expressing cells.
(A) Cell cycle profiles of wild-type, the Rec8-expressing strain, and the Rec8-expressing strains carrying a single reconstructed mutation. Cells were released from a G1 arrest as described in Fig 1D. Flow cytometry profiles are shown at the indicated times and profiles at 30 minutes are labeled in red. (B) Quantitation of the fraction of replicating cells in strains carrying a single reconstructed mutation at 30 minutes after release from a G1 arrest. The replicating subpopulation was measured as the fraction of the population between the G1 peak and the G2/M peak. The darker gray points represent the values of three biological replicates, and the thinner gray bar represents one standard deviation on each side of the mean of these measurements. The statistical significance between data from the Rec8-expressing strain and each mutation-reconstructed strain was calculated by two-tailed Student t test, *p < 0.05, **p < 0.01. (C) The replication profiles of wild-type, the Rec8-expressing strain, and the Rec8-expressing strains with single reconstructed mutations (ssn3Δ, esp1-P15, or smc1-P15). Replication dynamics is expressed as Trep (shown on the y-axis), the time at which 50% of cells in a population complete replication at a given genomic locus. The mean replication profile of two experiments on one part of Chromosome 14 is shown. The replication profile of each strain is color-coded. An arrowhead represents a fired replication origin. We confirmed that the different strains exited from G1 at the same time by monitoring their budding index over time (S15 Fig). (D) Genetically and chemically perturbing genome replication improves sister chromosome cohesion. The Rec8-expressing rrm3Δ and clb5Δ clb6Δ strains were assayed for sister chromosome cohesion as described in Fig 1C. Hydroxyurea was added in YPD as Rec8-expressing cells entered the cell cycle. The percentages of cells with two GFP dots in mitotically arrested populations (150 minutes after G1) are shown. The darker gray points represent the values of three biological replicates and the thinner gray bar represents one standard deviation on each side of the mean of these measurements. The statistical significance between data from the Rec8-expressing strain and each mutant strain was calculated by two-tailed Student t test, *p < 0.05, **p < 0.01. (E) The effect of rrm3Δ and clb5Δ clb6Δ on the fitness of the Rec8-expressing strain. The darker gray points represent the values of three biological replicates, and the thinner gray bar represents one standard deviation on each side of the mean of these measurements. The statistical significance between data from the Rec8-expressing strain and each mutant strain was calculated by two-tailed Student t test, **p < 0.01. Data associated with Fig 6B, 6D, and 6E can be found in S1 Data. clb, cyclin B 5; esp1, extra spindle pole bodies 1; GFP, green fluorescent protein; HU, hydroxyurea; Rec8, recombination 8; rrm3, rDNA recombination mutation 3; SCC, sister chromosome cohesion; smc1, structural maintenance complex 1; ssn3, suppressor of snf1; YPD, yeast extract, peptone, and dextrose.
Fig 7.
Summary of the mechanism that allows budding yeast to use the meiotic kleisin, Rec8, for mitosis.
Yeast cells expressing Rec8 in place of Scc1 cannot build robust cohesion to hold sister chromosomes together before anaphase due to the weak association of cohesin with chromosomes and Rec8 protein instability. Rec8-expressing cells induce earlier firing of replication origins compared to what wild type does and start S phase earlier. After experimental evolution, adaptive mutations in different functional modules delay origin firing and improve sister chromosome cohesion of Rec8-expressing cells, potentially by allowing more time for Rec8-containing cohesin to load onto chromosomes prior to the passage of replication forks. Rec8, recombination 8; Scc1, sister chromosome cohesion 1; WT, wild type.