Skip to main page content
Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
. 2017 May 15;36(10):1364-1378.
doi: 10.15252/embj.201696050. Epub 2017 Apr 24.

A Second Wpl1 Anti-Cohesion Pathway Requires Dephosphorylation of Fission Yeast Kleisin Rad21 by PP4

Affiliations
Free PMC article

A Second Wpl1 Anti-Cohesion Pathway Requires Dephosphorylation of Fission Yeast Kleisin Rad21 by PP4

Adrien Birot et al. EMBO J. .
Free PMC article

Abstract

Cohesin mediates sister chromatid cohesion which is essential for chromosome segregation and repair. Sister chromatid cohesion requires an acetyl-transferase (Eso1 in fission yeast) counteracting Wpl1, promoting cohesin release from DNA We report here that Wpl1 anti-cohesion function includes an additional mechanism. A genetic screen uncovered that Protein Phosphatase 4 (PP4) mutants allowed cell survival in the complete absence of Eso1. PP4 co-immunoprecipitated Wpl1 and cohesin and Wpl1 triggered Rad21 de-phosphorylation in a PP4-dependent manner. Relevant residues were identified and mapped within the central domain of Rad21. Phospho-mimicking alleles dampened Wpl1 anti-cohesion activity, while alanine mutants were neutral indicating that Rad21 phosphorylation would shelter cohesin from Wpl1 unless erased by PP4. Experiments in post-replicative cells lacking Eso1 revealed two cohesin populations. Type 1 was released from DNA by Wpl1 in a PP4-independent manner. Type 2 cohesin, however, remained DNA-bound and lost its cohesiveness in a manner depending on Wpl1- and PP4-mediated Rad21 de-phosphorylation. These results reveal that Wpl1 antagonizes sister chromatid cohesion by a novel pathway regulated by the phosphorylation status of the cohesin kleisin subunit.

Keywords: WAPL; Protein Phosphatase 4; cohesin; fission yeast.

Figures

Figure 1
Figure 1. A genetic screen for eso1‐H17 suppressors identifies pph3, encoding the catalytic subunit of PP4

The eso1‐H17 mutant is thermosensitive for growth. After 3 days at 36.5°C, spontaneous suppressors emerged as white colonies on an otherwise background of dead cells coloured dark red by the vital dye phloxine B.

Among 59 suppressors, most (51) were allelic to wpl1 or eso1. Two mapped to pds5, three to psc3 and two to psm3. A mutation was genetically linked to eso1 (20% recombinants) on the right arm of chromosome 2 and mapped to the SPBC26H8.05c open reading frame (pph3) by tiling array hybridization and sequencing. The mutation (R111W) is located near to the annotated active site (H114) of the phosphatase.

Cell growth assays showing that all mutants are eso1‐H17 suppressors at 36°C.

Cell growth assay showing that pph3 or psy2 deletion (encoding PP4 catalytic and regulatory subunits, respectively) suppress the thermosensitive growth phenotype of eso1‐H17.

Deletion of pph3 or psy2 allows cell growth in the absence of the otherwise essential eso1 gene.

The deletion of wpl1 bypasses eso1 requirement more efficiently and is epistatic on PP4 deletion mutants.

Figure 2
Figure 2. Wpl1 anti‐cohesion function requires PP4

Wpl1 induction in eso1Δ G2 cells induces loss of sister chromatid cohesion and aberrant mitosis in a PP4‐dependent manner. Cycling cdc25‐22 cells (˜80% G2 cells) were shifted to 37°C to prevent mitotic entry, and Wpl1‐MYC was induced at the indicated time‐points. Cells were released into mitosis by shifting the temperature back to 25°C, DNA was stained with DAPI, and aberrant mitoses (arrows) were scored. Scale bar 10 μm.

Wpl1‐MYC was induced for 2 h after which time Wee1‐as8 was inhibited with 3BrBPP1 to override the cdc25‐22 arrest. Cells progressed into M phase at 37°C and were arrested at metaphase by the thermosensitive APC mutation cut9 ts. Cells were fixed, tubulin stained to visualize the mitotic spindle and sister chromatid cohesion was monitored by FISH using a probe proximal to the centromere of chromosome 2. Scale bar 2 μm. ***P < 0.0001 two‐sided Fisher's exact test with α < 0.05. The number of metaphase cells examined is indicated within the graph.

Figure EV1
Figure EV1. Deletion of pph3, psy2 or both similarly reduced the occurrence of aberrant mitoses in eso1Δ cells upon Wpl1 induction

Scheme of the experiment. Cycling cdc25‐22 cells were shifted to 37°C, and Wpl1‐MYC was induced (TET) or not (DMSO) 1 h after the temperature shift. Cells were released into mitosis 2 h later by shifting the temperature back to 25°C, DNA was stained with DAPI, and aberrant mitoses were scored.

Frequency of aberrant mitoses.

Figure 3
Figure 3. Wpl1 activity is coupled with PP4‐dependent de‐phosphorylation of Rad21

The regulatory PP4 subunit Psy2 co‐immunoprecipitates Wpl1 from total protein extracts.

Psy2‐FLAG co‐immunoprecipitates all three core cohesin subunits from total protein extracts.

Rad21 is hyper‐phosphorylated in pph3, pds5 and wpl1 deletion mutants but not in psm3 acetyl‐mutants.

In vitro treatment with λ phosphatase of Rad21‐9PK immunoprecipitated from the indicated strains shows that Rad21 mobility shift is due to phosphorylation.

Western blot analysis of total protein extracts showing that tet07‐wpl1‐MYC induction in G2 cells triggers PP4‐dependent Rad21 de‐phosphorylation. A cdc25‐22 wpl1‐MYC control strain was included (right panel) showing that the amount of Wpl1‐MYC produced from tet07‐wpl1‐MYC after 2 h is similar to that produced from wpl1‐MYC.

Source data are available online for this figure.
Figure EV2
Figure EV2. Identification of Rad21 residues more frequently phosphorylated in the absence of PP4

Scheme of the experiment. Rad21‐9PK was immuno‐purified from pph3 + and pph3Δ cells, digested into peptides and analysed by LC‐MS. Intensity versus retention time profiles were generated from which area under the curve (AUC) were calculated. Relative peptide amount in pph3 + versus pph3Δ is proportional to AUC. Peptide and phosphosite identification were determined by LC‐MS/MS.

Phosphorylated residues identified in Rad21‐PK purified from pph3 + and pph3Δ cells.

PISS164S165 and S216VHS219DNQSQIS226 are more frequently phosphorylated in the absence of Pph3. Quantitative analysis was performed from three independent experiments. See also Dataset EV1.

Figure 4
Figure 4. Screen for rad21 phospho‐mutants suppressors of eso1‐H17

A combinatorial rad21 DNA library was synthesized that incorporates either an alanine (non‐phosphorylatable), a glutamic acid (phospho‐mimicking) or a serine codon for each of the 12 sites shown in the diagram. These include serine residues more frequently phosphorylated in the absence of Pph3 (S164 S165, S216 S219 S226) and their neighbouring serine residues. S314 and S315 were included as they were detected as phosphorylated only in the absence of Pph3 although with low confidence (only in one experiment and two peptides with an ambiguity on the position of the phosphorylated residue, see Dataset EV1).

Workflow of the selection procedure. The rad21 DNA library was cloned in the pREP41 vector carrying the down‐regulated version of the nmt promoter and the LEU2 selection marker (Basi et al, 1993). The library was transformed into a leu1‐32 eso1‐H17 recipient strain. Transformants were allowed to grow for 24 h at 25°C and then shifted to 32°C to select for clones able to grow at the restrictive temperature for eso1‐H17. Plasmid DNA was recovered from individual clones and transformed again into the eso1‐H17 strain to ensure that suppression was conferred by plasmid DNA. The region surrounding the variable codons was PCR amplified and used to replace the wild‐type rad21 + allele to generate rad21 phospho‐mutants.

Primary amino acid sequence at each of the 12 variable positions encoded by the rad21 phospho‐alleles. Eleven alleles did not modify the restrictive temperature of eso1‐H17. Two mutants enhanced the thermosensitive phenotype (enhancer) and 7 rad21 phospho‐alleles behaved as eso1‐H17 suppressors. A, alanine; S, serine; E, glutamic acid.

Cell growth assays showing the suppressor/enhancer phenotypes of the selected phospho‐mutants.

Cell growth assays showing that a single phospho‐mimicking residue at position 163 (S163E) is sufficient to partially suppress the thermosensitive phenotype of eso1‐H17 and the level of suppression is increased with an additional phospho‐mimicking residue at position 164 and/or 165.

Figure 5
Figure 5. The phosphorylation status of Rad21 S163 S164 regulates Wpl1 anti‐cohesion function

Cell growth assays showing that rad21‐S163E164E and rad21‐S163A164A behave as eso1‐H17 suppressor and enhancer, respectively. Deletion of the pph3 gene is epistatic on rad21‐163E164E, and conversely, the suppression by pph3Δ is reduced by rad21‐S163AS164A.

wpl1Δ is epistatic on rad21 phospho‐alleles for eso1‐H17 suppression.

Cen2FISH on metaphase cells. Wpl1‐MYC was induced in cdc25‐22 wee1‐as8 cut9 ts eso1Δ cells as in Fig 2B, and sister chromatid cohesion was monitored by FISH in metaphase‐arrested cells using the cen2 proximal probe. The number of metaphase cells examined is indicated. ***P < 0.0001 two‐sided Fisher's exact test with α < 0.05.

Figure 6
Figure 6. Analysis of chromatin‐bound Rad21 after Wpl1 induction

Wpl1‐MYC was induced as in Fig 2B. The mis4‐367 mutation prevents further cohesin loading at 37°C.

Cen2FISH on metaphase cells. ***P < 0.0001 two‐sided Fisher's exact test with α < 0.05. The number of metaphase cells examined is indicated.

Cohesin (Rad21‐9PK) map at and around the centromere of chromosome 2 in cdc25‐22 arrested cells (Schmidt et al, 2009) was used to design primer pairs (vertical arrows) for chromatin immunoprecipitation (ChIP).

Efficiency of Wpl1‐dependent Rad21‐9PK release from DNA. The ratio ChIP TET/ChIP DMSO was calculated from four ChIPs pairs (mean ± SD).

Source data are available online for this figure.
Figure EV3
Figure EV3. Wpl1‐dependent cohesin removal is inhibited in a psm3 NN background

Cycling cells were shifted at 37°C, and Wpl1‐MYC was induced (TET) or not (DMSO) after 1 h. Cells were collected 2 h later and processed for Rad21‐9PK ChIP.

Rad21‐9PK binding at and around the centromere of chromosome 2. Note that the amount of DNA‐bound cohesin is reduced in a psm3 NN background, presumably because cohesin loading is reduced (Murayama & Uhlmann, 2015). The values are the mean ± SD from four ChIPs.

The ratio ChIP TET/ChIP DMSO indicates that psm3 NN prevents Wpl1‐mediated Rad21‐9PK release from DNA. The values are the mean ± SD from four ratios. For some sites, the ratio was not calculated as the ChIP values were close to background levels for psm3 NN.

Figure EV4
Figure EV4. Wpl1‐dependent Rad21 release from DNA is barely affected by rad21 phospho‐mutants

Wpl1‐dependent Rad21 release in a psy2 + background.

Wpl1‐dependent Rad21 release in a psy2Δ background.

Data information: The experiment was performed as in Fig 6, and the ratio ChIP TET/ChIP DMSO was calculated (mean ± SD from four ratios). The data for eso1Δ and eso1Δ psy2Δ were reported from Fig 6 for comparison.Source data are available online for this figure.
Figure 7
Figure 7. PP4 interacts with Wpl1 and de‐phosphorylates Rad21 on chromatin

Rad21 is de‐phosphorylated on chromatin in a Wpl1‐ and PP4‐dependent manner in the absence of Eso1. Wpl1 was induced in the indicated strains as in Fig 6. Samples were collected 2 h later, chromatin proteins were extracted and analysed by Western blotting with anti‐Rad21 antibodies.

Fractionation controls. Tubulin and histone H3 were used as markers for the soluble fraction (S) and the chromatin pellet (P), respectively.

PP4 interacts with Wpl1 on chromatin. Exponentially growing cells from the indicated strains were collected. Psy2‐FLAG was immunoprecipitated from the soluble fraction (S1) and from chromatin proteins (S2) solubilized by nucleic acid digestion of the chromatin pellet (P). Input, IP and fractionation controls were analysed by Western blotting using the indicated antibodies.

Model of Wpl1‐PP4 regulation of sister chromatid cohesion.

Source data are available online for this figure.
Figure EV5
Figure EV5. Psy2 binds chromatin and interacts with cohesin independently of Wpl1 and Pds5

Psy2‐FLAG binds chromatin. Nuclear spreads were made from cycling cells of the indicated strains and probed with anti‐FLAG and anti‐PK antibodies. Scale bar 1 μm.

Psy2‐FLAG is found at cohesin‐associated regions. Cycling cells from the indicated strains were processed for ChIP with anti‐FLAG antibodies. DNA enrichment was measured at centromeres (imr), the non‐transcribed spacer region within the rDNA gene cluster (NTS1) and 5 chromosomal arm sites. The values represent the enrichment calculated from one ChIP and two qPCRs.

Psy2 binds chromatin independently of pds5 and wpl1. Nuclear spreads were made from cycling cells of the indicated strains and probed with anti‐FLAG antibodies. The values are the mean fluorescence intensities from the indicated number of nuclei. Error bar = 95% confidence interval of the mean with α = 0.05. a.u., arbitrary units.

Psy2‐FLAG co‐immunoprecipitates Psm1 independently of wpl1, pds5 and pph3. Psy2‐FLAG was immunoprecipitated with anti‐FLAG antibodies from total protein extracts made from cycling cells of the indicated genotypes.

Wpl1‐13MYC co‐immunoprecipitates Rad21 independently of PP4. Total protein extracts were made from cycling cells of the indicated strains and immunoprecipitated with anti‐MYC antibodies.

Source data are available online for this figure.

Similar articles

See all similar articles

Cited by 7 articles

See all "Cited by" articles

Publication types

MeSH terms

LinkOut - more resources

Feedback