Phosphorylation of Rec8 cohesin complexes regulates mono-orientation of kinetochores in meiosis I

Cohesin-Mis4 but not cohesin-Pds5 is excessively loaded at the core centromere in moa1∆ cells

In meiosis, Rec8 cohesin complexes are loaded at the core centromere to establish cohesion and mono-orientation of kinetochores. In moa1Δ cells, Rec8 cohesin localization increases at the core centromere although sister chromatid cohesion is abolished at this site (Yokobayashi & Watanabe, 2005; Sakuno et al, 2009). To explore the regulatory mechanisms of Rec8 cohesin at the core centromere, we examined the localization of Rec8, Mis4, and Pds5 by ChIP assay. The data revealed that in moa1Δ cells not only the localization of Rec8 but also that of Mis4 increases at the core centromere whereas Pds5 does not (Fig 1B). Based on these ChIP results, we hypothesized that the Mis4-dependent cohesin loading at the core centromere is active in moa1Δ cells, but the process of converting cohesin from the Mis4-bound into Pds5-bound form might be hampered. This might be one reason why cohesion is not established or maintained at the core centromere.

Identification of Plo1 phosphorylation sites in Rec8 required for mono-orientation

Recent cryo-EM studies have revealed how the cohesin complex binds to its loader and DNA (Collier et al, 2020; Higashi et al, 2020; Shi et al, 2020). The studies suggest that the major chromatin binding pocket of the cohesin complexes is formed by the heterodimer of Psm1-Psm3 head domains, Mis4, and the N-terminal helical domain of Rec8 that binds to the coiled-coil region of Psm3 (Fig 1A, square). We explored the possibility that the putative Mis4/Pds5-binding surface of Rec8 (111–225 a.a.) might be the target of regulation by Moa1-Plo1, which may facilitate the exchange from Mis4 into Pds5. We identified 11 polo-kinase consensus N/Q/E/D-X-S/T sites and four non-consensus S/T in this domain, which were phosphorylated by Plo1 in vitro and some of them were detected also in vivo (Fig S1A–C). To examine the significance of Rec8 phosphorylation by Plo1, we replaced endogenous rec8⁺ with the non-phosphorylatable mutant rec8-15A, in which all 15 serines/threonines in the putative Mis4/Pds5-binding surface of Rec8 were replaced with alanines (Fig 2A). We then examined meiotic chromosome segregation of the mutant. We monitored GFP fluorescence associated with a lacO array integrated at the centromere of one homolog of chromosome I (imr1::GFP). In moa1Δ cells, a minority population (11%) of cells underwent equational segregation at meiosis I (because of defects in mono-orientation), whereas the majority underwent reductional segregation due to the presence of chiasmata and tension exerted across homologs, as reported previously (Miyazaki et al, 2017) (Fig 2B). Moreover, 20% of the cells with reductional segregation at meiosis I underwent nondisjunction of sister chromatids in meiosis II, suggesting that 40% of the reductional population underwent random segregation at meiosis II which is originated from loss of cohesion (a defect in cohesion protection) in anaphase I. In contrast, rec8-15A cells showed nearly normal chromosome segregation in both meiosis I and II (Fig 2B). Because a mono-orientation defect at the core centromere is greatly suppressed by chiasmata and pericentromeric cohesion protection, we removed these two effects by deleting the recombination gene rec12 (the Spo11 homolog in fission yeast) and by introducing a protection defective mutation in Rec8 (rec8-2A: S449A, S450A) (Ma et al, 2021) (Fig 2B). In rec12Δ cells, merotelic attachment of mono-oriented (fused) kinetochores often causes bi-orientation of homologs at metaphase I but sister chromatids largely fail to separate in the following anaphase because of protection of cohesion at the pericentromere, resulting in only rare (<2%) equational segregation (Sakuno et al, 2011) (Fig 2C). However, in rec12Δ rec8-2A cells, the separation increases to 12% because of the loss of protection (Fig 2C and D). Deletion of moa1 in this background, by abolishing mono-orientation, leads to >80% of cells showing equational segregation. Remarkably, introducing the rec8-15A mutation into rec12Δ rec8-2A cells increased equational segregation to 36% (Fig 2D), suggesting that phosphorylation at some or all 15S/T sites in Rec8 is contributing at least partly to establishing mono-orientation, most likely by promoting cohesion at the core centromeres.

Next, we asked if phosphorylation of Mis4 might contribute to mono-orientation by facilitating cohesin conversion from the Mis4-bound into Pds5-bound (Fig 1A). Structural analyses indicate that the middle and C-terminal region of Mis4 directly interact with the kleisin subunit, Psm1, Psm3-K105K106 and DNA, and play a key role in cohesin loading and translocation (Kikuchi et al, 2016; Collier et al, 2020; Higashi et al, 2020; Shi et al, 2020) (Fig 1A, square). We identified five polo-kinase consensus sequences in these domains (Fig 2E). Furthermore, we examined the in vitro phosphorylation by Plo1 using recombinant Mis4 protein and found that the N terminus of Mis4 is preferentially phosphorylated by Plo1 (Fig S2A–C). This domain contains nine polo-kinase consensus sequences (Fig S2D). Consequently, we introduced alanine substitutions in totally 14 polo-kinase consensus sequences in Mis4 (Fig 2E). However, mis4-14A cells showed only 16% equational segregation compared with WT cells (12%) in the rec12Δ rec8-2A background, implying that few defects in mono-orientation were caused by mis4-14A (Fig 2D).

To delineate the interaction of Rec8 with Mis4 or Pds5, we examined their interactions by two-hybrid assay using the N-terminal fragment of Rec8 (1–245 a.a.) that includes the Mis4/Pds5-binding surface. Rec8-N indeed interacted with Pds5 but little with Mis4. Remarkably, the interaction between Rec8 and Pds5 was reduced in non-phosphorylatable Rec8-15A but robust in Rec8-15D/E which carries phosphomimetic replacements (Fig 2F). Although these results are consistent with the hypothesis that phosphorylation of Rec8 by Moa1-Plo1 facilitates converting Mis4-bound Rec8 cohesin complex into Pds5-bound (Fig 1A), further detailed analysis of each phosphorylation site of Rec8 and Mis4/Pds5 might be required for drawing firm conclusion.

Phosphorylation in the kleisin-Psm3 exit/entry gate

Mis4 entraps DNAs near the kleisin-Psm3 gate (Collier et al, 2020; Higashi et al, 2020; Shi et al, 2020). Once Psm3-K105K106 acetylation takes place during DNA replication (Moldovan et al, 2006), Mis4 is readily replaced by Pds5 because Pds5 but not Mis4 binds acetylated Psm3 (Petela et al, 2021). Therefore, it is possible that the kleisin-Psm3 gate and nearby Psm3-K105K106 acetylation domain might also be the target of regulation by Moa1-Plo1 (Fig 1A, square). We examined if Psm3 is phosphorylated by Plo1 in vitro and found phosphorylation of T182 in the N terminus and S1001 in the C terminus, both locate in the coiled-coil region of the DNA exit gate (Figs 3A and S3A–E), which directly interacts with the N-terminal domain of kleisin (kleisin’s NTD) (Figs 1A and 3B). Indeed, cells expressing Psm3-2A (alanine substitution at T182 and S1001) showed mono-orientation defects (20%) albeit mildly, compared with WT (12%) (Fig 3C), suggesting that phosphorylation of these sites might contribute to establishing cohesion at the core centromeres. We analyzed phosphomimetic mutants psm3-T182E or psm3-S1001D. Although neither mutation showed defects in meiotic chromosome segregation in rec12⁺ cells, psm3-ED (T182E and S1001D) caused equational segregation in >70% cells (Fig 3D). This is a reminiscent of rec8Δ which also causes equational rather than random segregation at meiosis I because the residual Rad21 cohesin sustains centromeric cohesion (Yokobayashi et al, 2003). We assayed sister chromatid cohesion in prophase I arrested cells by monitoring GFP fluorescence associated with a lacO array integrated at the centromere of one homolog of chromosome I (imr1::GFP) and at the arm region of one homolog of chromosome II (cut3::GFP). The results indicate that cohesion defects in psm3-ED cells is a reminiscence of rec8Δ (Fig 3E). Thus, we thought that Rec8 cohesin function was impaired in psm3-ED cells. Accordingly, immunofluorescence examination indicated that the localization of Rec8 was largely lost in meiotic psm3-ED cells (Fig 3F). Furthermore, Western blot assay indicated that the amount of Rec8 protein was reduced in psm3-ED cells but not in WT and psm3-2A cells (Fig S4), suggesting that Rec8 protein is not stable in psm3-ED cells. However, psm3-ED did not affect Rad21 localization during vegetative growth and cells were not sensitive to thiabendazole (TBZ) (Fig 3G and H), indicating that Rad21 cohesin function is intact in psm3-ED cells. These results are consistent with the scenario that phosphorylation at Psm3-T182 and -S1001 might open the DNA-entry/exit gate only when it is formed by Rec8’s NTD but not by Rad21’s NTD. We speculate that the Psm3 phosphorylation at the gate may play a role in transiently loosening Rec8-Psm3 gate to facilitate establishment of cohesion at the core centromere.

A phosphorylation located adjacent to the Psm3 acetylation loop

The in vitro Plo1 phosphorylation assay further identified Psm3-S110 fitting with the polo-kinase consensus sequence (Figs 4A and S3C). Mutagenesis revealed that psm3-S110A (alanine substitution at the S110) causes mono-orientation defects (50%) compared with WT (12%) in meiotic cells (Fig 4B). Because Psm3-S110 positions at the anti-parallel β-sheets structure that extrudes a peptide loop containing K105 and K106 residues of Psm3 (Psm3’s KKD loop) (Kouznetsova et al, 2016), the phosphorylation at S110 might structurally affect Psm3’s KKD loop, which plays a key role in cohesin loading and releasing function. Because psm3-S110D (phosphomimetic substitution at the S110) also causes mono-orientation defects (Fig 4B), we assume that the phosphorylation may occur only transiently. Importantly, either mutation does not affect the mitotic function of cohesin as the mutants show no defects in growth on a TBZ plate (Fig 4C). These results suggest that the phosphorylation at Psm3-S110 specifically regulates Rec8 cohesin at centromeres most likely depending on Moa1-Plo1. Moreover, even in meiosis, chromosome segregation in meiosis I and meiosis II are mostly normal in these mutants (Fig 4D), indicating that in contrast to the core centromere, pericentromeric cohesion, which is essential for correct meiosis II, is largely intact. Mono-orientation defects of psm3-S110A and psm3-S110D are exposed only in the rec12Δ rec8-2A background. Intriguingly, in budding yeast and human Smc3, the polo-kinase consensus sequence is also found at the identical parallel β-sheets but in the opposite side (Fig S5A and B), which might similarly affect Smc3’s KKD loop structure if the residue is phosphorylated by polo-like kinase.

To explore the relation among the phosphorylation on Rec8 and Psm3, we combined rec8-15A and psm3-3A (S110A, T182A, and S1001A), and examined mono-orientation defects in rec12Δ rec8-2A background. Accordingly, rec8-15A psm3-3A double mutant showed more defects in mono-orientation than either rec8-15A or psm3-3A mutant (Fig 4E), suggesting that the phosphorylation on Rec8 and Psm3 cooperatively act to establish cohesion at the core centromeres. When wpl1⁺ is deleted in the rec8-15A, psm3-3A, and rec8-15A psm3-3A mutants, mono-orientation defects were suppressed in psm3-3A and rec8-15A psm3-3A mutants but not in rec8-15A (Fig 4E). These results support the notion that the psm3-3A mutation locating near the Rec8-Psm3 gate and Psm3-KKD-loop enhanced the cohesin release pathway, which depends on Wpl1.

In conclusion, we identified several Plo1 phosphorylation sites in the cohesin subunits, Rec8 and Psm3, non-phosphorylatable mutations of which cause defects in mono-orientation in meiosis I. We argue that the phosphorylation of Rec8 and Psm3 at the core centromeres is at least partly mediated by kinetochore localized Moa1-Plo1 that regulates Rec8 cohesin complexes to establish cohesion required for mono-orientation (Yokobayashi & Watanabe, 2005; Sakuno et al, 2009). Although previous studies suggest that Rec8 cohesin dependent cohesion at the core centromere is important for mono-orientation in various organisms including plants and animals (Watanabe, 2012; Kim et al, 2015; Ogushi et al, 2021), it has been elusive in budding yeast, the only of the studied organisms having a point centromere. However, a recent study in budding yeast suggests that cohesin indeed plays an important role in establishing mono-orientation (Barton et al, 2022). Thus, we anticipate that the molecular mechanisms of meiosis-specific kinetochore regulation revealed in fission yeast might be conserved in a wide range of eukaryotic organisms, including human.

Schizosaccharomyces pombe strains and media

Unless otherwise stated, all media and growth conditions were as described previously (Moreno et al, 1991). Complete medium (YE), minimal medium (EMM), minimal medium without nitrogen source (EMM-N) and sporulation-inducing media (SPA) were used. All strains used in this study are described in Table S1. Deletion of rec8⁺, moa1⁺, rec12⁺, mei4⁺, and wpl1⁺; tagging of rec8⁺, rad21⁺, and pds5⁺ by GFP; and tagging of mis4⁺ by PK were carried out according to the PCR-based gene targeting method for S. pombe using the kanMX6 (kanR), hphMX6 (hygR), natMX6 (natR), bsdMX6 (bsdR), and aurMX6 (aurR) genes as selection markers. We used imr1::GFP and cut3::GFP markers to monitor cohesion of sister chromatids (Tomonaga et al, 2000; Sakuno et al, 2009). To introduce mutants in rec8⁺, psm3⁺, and mis4⁺ genes, we cloned fragments of these genes into plasmid, respectively, and mutagenized using whole plasmid mutagenesis, correct mutagenesis was confirmed by PCR and/or DNA sequencing.

Synchronous cultures of fission yeast meiotic cells

For meiosis microscopic observation, logarithmically growing cells were collected, suspended in 20 mg/ml Leucine, spotted on sporulation-inducing medium (SPA), and incubated at 28°C. For chromosome segregation assay, imr1-GFP was observed in the mes1⁺ or mes1-B44 mutant that arrests at prophase II (after 16 h). For sister chromatid cohesion assay, imr1-GFP and cut3-GFP were observed in mei4Δ mutant that arrests at prophase I (after 24 h).

In vitro phosphorylation assay of Rec8, Mis4, and Psm3

The recombinant GST or MBP fusion proteins (Rec8, Mis4, and Psm3) were produced in Escherichia coli using the pGEX4T-2 and pGEX4T-1 vector (GE Healthcare). E. coli was lysed in lysis buffer (40 mM Tris–HCl [pH 7.5], 0.5 mM EDTA, 150 mM NaCl, and 0.5% Triton X-100), and GST fusion proteins were purified on glutathione–Sepharose (GE Healthcare) and MBP fusion proteins were purified by amylose resin (NEB). Recombinant Plo1 kinase and substrates were incubated with kinase buffer (50 mM Tris–HCl [pH 8.0], 150 mM NaCl, 10 mM MgCl₂, 0.5% Triton X-100, and 5 mM dithiothreitol) at 30°C for 30 min in the presence of (γ-³²P) ATP or cold ATP. The incorporation of the radioactive phosphate group was visualized by means of autoradiography, and protein loading was analyzed by staining with Coomassie brilliant blue. Phosphopeptides were analyzed by liquid chromatography coupled with tandem mass spectrometry (LC–MS/MS).

In vivo phosphorylation assay of Rec8

Protein bands were excised and in-gel digested using trypsin. The peptides were analyzed with an Orbitrap-Fusion Lumos mass spectrometer coupled to an UltiMate 3000 HPLC equipped with an EASY-Spray nanosource (Thermo Fisher Scientific). Raw data were processed using MaxQuant v1.6.0.1 and searched against a UniProt extracted S. pombe FASTA file amended to include common contaminants, with phosphorylation (STY), acetylation (K), and methylation (KR) being selected as variable modifications. The proteingroup.txt, phosphoSTY.txt output tables were imported into Perseus software for further processing. All intensity values were log2 transformed.

Fluorescence microscopy

All cell fluorescence microscopy was performed using a Nikon ECLIPSE Ti2-E inverted microscope with photometrics PRIME 95B camera. This microscope was controlled by NIS-Elements software. Thirteen z sections (spaced by 0.4 μm each) of the fluorescent images were converted into a single two-dimensional image by maximum intensity projection. Image J software (NIH) was used to adjust brightness and contrast and to render maximum projection images.

Two-hybrid assays

The yeast two-hybrid system (BD) was used according to Clontech’s instructions. We amplified the ORFs of pds5⁺ and mis4⁺ genes by PCR and cloned them into pGBK-T7, a Gal4 DNA-binding domain–based bait vector, amplified rec8⁺, rec8-15A, rec8-15D/E, eso1⁺, and ssl3⁺ and cloned into pGAD-T7, a Gal4 activation domain–based prey vector. Y2HGold and Y187 strains were used for transformation of pGBK-T7 and pGAD-T7, respectively. SD-tryptophan and SD-leucine plates were used as selective medium. Positive transformants of Y2HGold and Y187 were mixed on YPD plate to get diploid cells and streaked on SD-Trp-Leu plate to get single colonies; checked on nutrition-restricted plates (SD-Trp-Leu-His-Ade).

Serial dilution analyses

TBZ was dissolved in DMSO as a stock solution at 20 mg/ml and stored at room temperature. Sterilized YE agar medium (115°C, 20 min) was cooled down (about 60°C) and added with TBZ to prepare the plates. For serial dilution plating assays, 10-fold dilutions of a mid-log phase culture (OD660 = 0.3∼0.6) were plated on the indicated medium and grown for 2 d at 30°C.

Chromatin immunoprecipitation assays

For the chromatin immunoprecipitation analysis, we used h⁺/h⁻ diploid cells. mei4Δ was used to arrest at prophase I. Cells were grown in YE liquid to a density of 1 × 10⁷ cells/ml at 30°C, then resuspended in EMM liquid medium including NH₄Cl (EMM + N) and grown to a density of 0.5 × 10⁷ cells/ml at 30°C, then resuspended in EMM medium lacking NH₄Cl (EMM − N) at a density of 1 × 10⁷ cells/ml for 9 h at 30°C. ChIP assays were carried out as described previously (Yokobayashi & Watanabe, 2005). The sequences of primers used are as follows: cnt primers (5′-ATCTCATTGCTATTTGGCGAC-3′, 5′-GCGTTTCTTCGGCGAAATGC-3′), imr primers (5′-CCTTTACTGGAAAATTGTCG-3′, 5′-GCTGAGGCTAAGTATCTGTT-3′), dg primers (5′-TTTTCAGCGAGACATGTACC-3′, 5′-TCATAAAGCAACACTGGGTG-3′), dh primers (5′-TGAATCGTGTCACTCAACCC-3′, 5′-CGAAACTTTCAGATCTCGCC-3′), zfs1 primers (5′-CCGGTTGAAAGGAATAGACT-3′, 5′-TTTCTTGCCTGAGATACCGT-3′), and mes1 primers (5′-CGAAGGCTACTTTCATGCCA-3′, 5′-CGTACATTCAGACTGTTGAAC-3′). After getting the qPCR results, we divided the value of cnt, imr, dg, and dh by the mean value of zfs1 and mes1 because the zfs1 and mes1 regions are basically free of cohesin and the detected values belong to background (Yokobayashi & Watanabe, 2005; Goto et al, 2017). The final values were obtained by subtracting the value calculated from control strains without the tag.

Statistical analysis

All the data replicates were applied and analyzed using GraphPad Prism version 9.5.1 (GraphPad Software). To estimate the significant differences between two groups, t test was performed. To estimate the significant differences between ≥3 groups, one-way ANOVA was performed.

Author Contributions

• Y Liu: data curation, software, formal analysis, and investigation.

• Y Min: data curation and methodology.

• Y Liu: methodology.

• Y Watanabe: conceptualization, data curation, formal analysis, supervision, funding acquisition, validation, investigation, methodology, and writing—original draft.

Conflict of Interest Statement

The authors declare that they have no conflict of interest.