Establishment of DNA-DNA Interactions by the Cohesin Ring

Abstract

The ring-shaped structural maintenance of chromosome (SMC) complexes are multi-subunit ATPases that topologically encircle DNA. SMC rings make vital contributions to numerous chromosomal functions, including mitotic chromosome condensation, sister chromatid cohesion, DNA repair, and transcriptional regulation. They are thought to do so by establishing interactions between more than one DNA. Here, we demonstrate DNA-DNA tethering by the purified fission yeast cohesin complex. DNA-bound cohesin efficiently and topologically captures a second DNA, but only if that is single-stranded DNA (ssDNA). Like initial double-stranded DNA (dsDNA) embrace, second ssDNA capture is ATP-dependent, and it strictly requires the cohesin loader complex. Second-ssDNA capture is relatively labile but is converted into stable dsDNA-dsDNA cohesion through DNA synthesis. Our study illustrates second-DNA capture by an SMC complex and provides a molecular model for the establishment of sister chromatid cohesion.

Keywords: Cohesin; S. pombe; SMC complexes; second-DNA capture; sister chromatid cohesion.

Graphical abstract

Figure 1

Second-DNA Capture by the Fission Yeast Cohesin Ring (A) Schematic of the second-DNA capture assay (protocol 1) and a gel image showing input and recovered DNA from the assay performed with the indicated substrates. ds, dsDNA; ss, ssDNA; rc, relaxed circular; c, circular. All reactions were carried out in the presence of ATP and an ATP regenerating system. 16.7% or 25% of input free dsDNA or ssDNA are shown. (B) Schematic of the second-DNA capture assay (protocol 2) and a representative gel image. 25% of input DNA is shown. (C and D) Quantification of the assays in (A) and (B), respectively, performed with WT and 1B3B cohesin. The means and standard deviations from three independent experiments are shown. (E) Quantification of second-DNA capture, using protocol 1B, in the absence or presence of the indicated adenosine derivatives. The means and standard deviations from three independent experiments are shown. See also Figure S1 for gel images that include supernatant fractions, reactions using 1B3B cohesin, titration of components, ATP competition, and an assay using ΦX174 ssDNA.

Figure S1

ssDNA, but not dsDNA, Is a Substrate for Second-Strand Capture, Related to Figure 1 (A and B) Extended gel images of Figures 1A and 1B, showing both beads bound and supernatant fractions. 25% of the supernatant fractions are shown. (C) Protocol 2 reactions were carried out using the indicated protein concentrations (each of cohesin, Mis4-Ssl3 and Psc3). The gel image and quantification of captured dsDNA are shown. (D) A typical gel image of inputs and products of a second-DNA capture reaction following protocol 1B, performed with WT cohesin. (E) Gel images showing second-DNA capture by 1B3B cohesin, containing Walker B mutations in both Psm1 and Psm3, using a protocol 1 reaction. (F) As (E), but the reaction followed protocol 2. (G) Competition of ATP with ADP or non-hydrolyzable ATP-γ−S. As in Figure 1E (top gel image), but an additional reaction was performed in which 0.25 mM of ATP was present in all reactions that were then supplemented by additional nucleotides. The ability of ADP and ATP-γ−S to compete with ATP demonstrates that both nucleotides are able to bind cohesin, but that their hydrolysis is required for second-DNA capture. (H) Cohesin mediates second-DNA capture irrespective of sequence homology. Protocol 2 reactions were carried out using pSKsxAS ssDNA (partially homologous to the free dsDNA substrate) or ΦX174 virion ssDNA (of unrelated sequence) as substrates. The graph presents means and standard deviations from three independent experiments.

Figure 2

Second-DNA Capture Results in Topological Embrace (A–C) (A) Schematic and (B) gel images of the DNA-release experiments by restriction enzyme cleavage of captured DNA or (C) cohesin cleavage by TEV protease following ssDNA to dsDNA conversion. The gel images show input and recovered DNA at each of the indicated steps. nc, nicked circular; L, linear; WT, wild-type cohesin; Rad21-TEV, TEV protease cleavable cohesin; TEV, TEV protease; S, supernatant; B, beads bound fraction. See also Figure S2 for further experiments that confirm the topological nature of second-DNA capture.

Figure S2

Second-DNA Capture is Topological in Nature, Related to Figure 2 (A) Cohesin must topologically embrace dsDNA to mediate second-DNA capture. Protocol 1 reactions were carried out with ‘closed’ (C) or ‘linear’ (L) topology dsDNA beads. The gel image and graph show recovery of free ssDNA. The graph shows means and standard deviation from three independent experiments (WT cohesin) or the range of recovered ssDNA from two independent experiments (1B3B cohesin). (B) A DNA release experiment as shown in Figure 2A was carried out using 1B3B cohesin. (C) Schematic of a DNA release experiment following protocol 2 s DNA capture. The ssDNA substrate was converted to dsDNA by DNA synthesis following capture. Then either of the two circular DNAs was digested with unique restriction enzymes, PstI (DNA beads) or BglII (free dsDNA). Recovered DNAs at the indicated stages of the experiment were analyzed by agarose gel electrophoresis. Input, bead bound (B) and supernatant (S) fractions are shown. (D) TEV cleavage of cohesin following second-DNA capture using protocol 2. The gel shows a representative image of input and recovered DNA, using WT and TEV cleavable (21TEV) cohesin, without or with TEV protease (TEV) treatment. After TEV protease treatment, the beads were washed with high salt buffer and recovered DNA was analyzed. The graph depicts the means and standard deviations from three independent experiments.

Figure 3

Topological but Labile ssDNA Embrace by the Cohesin Ring (A) Gel images and quantification of cohesin-loading assays using ssDNA or dsDNA as the substrate. Mis4-Ssl3 (MS) or Pds5-Wapl (PW) were added in the presence or absence of ATP. The graph shows the means and standard deviation from three independent experiments. (B) Following loading, the recovered material was challenged with NaCl and EDTA. The gel image and graph show DNA recovery at the indicated stages of the experiment. Means and standard deviations from three independent experiments are given. (C) Schematic and outcome of the dsDNA-to-ssDNA conversion experiment using E. coli exonuclease III (exoIII). Supernatant (S) and beads-bound (B) fractions were analyzed after the NaCl and EDTA chase. The graph indicates means and standard deviations from three independent experiments. (D) Specificity of second-ssDNA capture. Gel images and quantification of the protocol 1B second-DNA capture experiments in the presence of indicated ratio of nicked circular dsDNA competitor. The graph shows the means and standard deviation from three independent experiments. See also Figure S3, showing ssDNA stimulation of the cohesin ATPase and a control that released ssDNA remains circular.

Figure S3

Further Characterization of ssDNA-Cohesin Interactions, Related to Figure 3 (A) ssDNA and dsDNA binding of cohesin were analyzed by electrophoretic mobility shift experiments using the indicated cohesin concentrations and single or double stranded pBluescript as the substrate. (B) ssDNA is topologically entrapped by cohesin. A schematic of the DNA-release experiment following ssDNA to dsDNA conversion is shown together with a gel image of the input and recovered DNAs at the indicated stages. (C) Cohesin releases circular ssDNA. The released DNA from cohesin following dsDNA to ssDNA conversion, as shown in Figure 3D, was treated with E. coli exonuclease I that digests linear but not circular ssDNA. No detectable digestion was observed, suggesting that released ssDNA remained circular. As a control for the effectiveness of exonuclease I treatment, heat-denatured nicked circular DNA was treated with exonuclease I in the same way. Linear, but not circular ssDNA was readily digested under these conditions. Note that two nicks were present in the circular DNA before denaturation. This generated two linear ssDNA, only the longer one of which is visible on the gel. (D) ssDNA stimulates the cohesin ATPase. Mis4-Ssl3 and DNA-dependent ATP hydrolysis by cohesin was measured in the presence of indicated proteins and DNAs.

Figure 4

ssDNA-to-dsDNA Conversion Establishes Stable DNA-DNA Cohesion (A) Time course of second-DNA capture in a protocol 2 assay, comparing WT and 1B3B cohesin. The percentage of free dsDNA captured on ssDNA beads is plotted over time. (B) Same as (A), but the second-DNA capture incubation proceeded for 5 or 30 min in the absence or presence of an ATP-regenerating system (ATP-RG). The indicated additions were made 5 min into the second-DNA capture incubation. (C) Schematic of the experiment to convert ssDNA-to-dsDNA following second-DNA capture to test stabilization against NaCl and EDTA treatment. The gel image shows input and the recovered and released DNAs at the indicated stages of the experiment. The means and standard deviations from three independent experiments are shown in each panel.

Figure 5

Acetyl-Acceptor Lysines and the Cohesin Loader Promote Second-DNA Capture (A) Protocol 2 experiments were carried out with the indicated order of additions, demonstrating a strong preference of reaction order during second-DNA capture. (B) Protocol 1B was used to test the ability of the indicated cohesin loading cofactors to promote second-DNA capture. Recovered DNA was analyzed by agarose gel electrophoresis and quantified. (C) WT and Psm3^K106Q (KQ) cohesin was used in a protocol 2 experiment. An aliquot was taken after the first dsDNA loading incubation to confirm comparable levels of loading by carrying out the reaction in low-salt condition (15 mM NaCl) before performing second-DNA capture on ssDNA beads followed by agarose gel electrophoresis and quantification. The means and standard deviations from three independent experiments are shown in each panel. See also Figure S4 for experiments that confirm the Mis4 requirement for second-DNA capture and the contribution of the acetyl-acceptor lysines to cohesin loading onto ssDNA.

Figure S4

Mis4-Ssl3, but not Pds5-Wapl, Promote Second-DNA Capture, Related to Figure 5 (A) Protocol 1B reactions were initiated by cohesin and Mis4-Ssl3. Then the dsDNA beads were washed and increasing concentrations of Pds5-Wapl were included for second-DNA capture. A reaction in which Mis4-Ssl3 was added back is included for comparison. The graph shows quantification of recovered free ssDNA detected by agarose gel electrophoresis. (B) Protocol 1B reactions were carried out in the presence of the indicated loading cofactors as described in Figure 5B, but 1B3B cohesin was used. The graph shows means and the range of recovered ssDNA from two independent experiments. (C) Protocol 2 reactions were carried out in the presence of the indicated loading cofactors. The gel image shows recovery of free dsDNA on the ssDNA beads, the graph reports means and standard deviations from three independent experiments. (D) Acetyl-acceptor lysines on Psm3 contribute to ssDNA loading. DNA loading reactions were carried out with the indicated cohesin complexes using dsDNA or ssDNA as substrates. The graph shows means and standard deviations of the recovered DNA from three independent experiments.

Figure 6

RPA Competes with Cohesin for Second-ssDNA Capture Schematic of the experiment in which cohesin was first loaded onto dsDNA, then increasing amounts of RPA (WT) or RPA containing Rfa1^G78E (G78E) were added before transfer onto ssDNA beads. The gel image shows input and recovered DNAs, and the graph shows means and standard deviations from three independent experiments. See also Figure S5 for a ssDNA binding assay comparing RPA and RPA containing Rfa1^G78E.

Figure S5

ssDNA Binding by RPA and RPA Containing Rfa1^G78E, Related to Figure 6 (A) Purified hetero-trimeric RPA (WT) and RPA containing Rfa1^G78E (G78E) were analyzed by SDS-PAGE followed by Coomassie Blue staining. (B) ssDNA binding activity of WT and the mutant RPA complex was analyzed in an electrophoretic mobility shift experiment using single stranded (dT)₅₄ as the substrate.

Figure S6

Additional Characterization of Budding Yeast rfa1^G77E, Related to Figure 7 (A) Characterization of the budding yeast rfa1^G77E mutant. 10-fold serial dilutions of cultures of the indicated strains were spotted on YPD plates and grown at the indicated temperatures, or were spotted on YPD plates containing 100 mM HU that were incubated at 25°C. (B) Cell-cycle progression in synchronized cultures of the indicated strains from Figures 7A and 7B was monitored by FACS analysis of DNA content. (C) rfa1^G77E does not rescue the HU sensitive growth of ctf18Δ cells. 10-fold serial dilutions of the indicated budding yeast cell cultures were spotted on YPD plates containing the indicated concentrations of HU and incubated at 25°C. (D) Rescue of the centromeric chl4Δ cohesion defect by rfa1^G77E. Sister chromatid cohesion close to centromere 5 (- 12.6 kb) was analyzed in strains of the indicated genotypes. Cells were arrested in mitosis by depleting Cdc20 under control of the MET3 promoter by methionine addition. At least 100 cells were scored in each experiment. The graphs show means and standard deviations from 3 independent experiments. In parallel, metaphase-arrested cells from the experiment were released into anaphase by washing with and further growth in minimal medium lacking methionine to re-induce Cdc20 expression. The percentage of GFP-marked chromosomes that split 20 min after release into anaphase is shown. (E) rfa1^G77E does not suppress cohesion defects of chl1Δ or eco1-1 cells. Sister chromatid cohesion was analyzed as Figure 7A. At least 100 cells were scored in each experiment reported. The graphs show means and standard deviations from 3 independent experiments. (F) rfa1^G77E does not rescue Smc3 acetylation in chl1Δ cells. Smc3 acetylation was analyzed in synchronized cultures from the indicated strains in the experiment shown in panel (E) by western blotting.

Figure 7

RPA Impacts on Sister Chromatid Cohesion Establishment In Vivo (A) Effect of the rfa1^G77E mutation or RPA overexpression (RPA OE) on sister chromatid cohesion in ctf18Δ cells. Cells were synchronized and arrested in mitosis by nocodazole treatment. Sister chromatid cohesion at the GFP-marked URA3 locus was analyzed. Western blotting confirmed RPA overexpression. At least 100 cells were scored under each condition. The graph shows means and standard deviations from three independent experiments. (B) Smc3 acetylation was analyzed in synchronized cultures from the strains above by western blotting. The acetyl-Smc3 signal, normalized to tubulin and then to the WT signal at 90 min, was quantified in three independent repeats of the experiment. The means and standard deviations are shown. (C) The cohesin loader promotes sister chromatid cohesion establishment. Sister chromatid cohesion was monitored at indicated time points following release from G1 or HU under the indicated conditions and genotypes. (D) A model for the establishment of sister chromatid cohesion at the DNA replication fork. See the Discussion for details. See also Figure S6 for supporting genetic and cell-cycle analyses that explore the role of RPA in sister chromatid cohesion establishment.