PP2A-Cdc55 phosphatase regulates actomyosin ring contraction and septum formation during cytokinesis

Eukaryotic cells divide and separate all their components after chromosome segregation by a process called cytokinesis to complete cell division. Cytokinesis is highly regulated by the recruitment of the components to the division site and through post-translational modifications such as phosphorylations. The budding yeast mitotic kinases Cdc28-Clb2, Cdc5, and Dbf2-Mob1 phosphorylate several cytokinetic proteins contributing to the regulation of cytokinesis. The PP2A-Cdc55 phosphatase regulates mitosis counteracting Cdk1- and Cdc5-dependent phosphorylation. This prompted us to propose that PP2A-Cdc55 could also be counteracting the mitotic kinases during cytokinesis. Here we show that in the absence of Cdc55, AMR contraction and the primary septum formation occur asymmetrically to one side of the bud neck supporting a role for PP2A-Cdc55 in cytokinesis regulation. In addition, by in vivo and in vitro assays, we show that PP2A-Cdc55 dephosphorylates the chitin synthase II (Chs2 in budding yeast) a component of the Ingression Progression Complexes (IPCs) involved in cytokinesis. Interestingly, the non-phosphorylable version of Chs2 rescues the asymmetric AMR contraction and the defective septa formation observed in cdc55∆ mutant cells. Therefore, timely dephosphorylation of the Chs2 by PP2A-Cdc55 is crucial for proper actomyosin ring contraction. These findings reveal a new mechanism of cytokinesis regulation by the PP2A-Cdc55 phosphatase and extend our knowledge of the involvement of multiple phosphatases during cytokinesis. Supplementary Information The online version contains supplementary material available at 10.1007/s00018-022-04209-1.


Introduction
Cytokinesis is the final event of the cell cycle and mediates the physical separation of mother and daughter cells. It is a highly ordered and regulated process that is conserved among eukaryotes. Cytokinesis must be spatially and temporally coordinated with sister chromatids segregation to avoid problems in chromosome segregation -failures can lead to aneuploidies and are associated with tumorigenesis [1]. Detailed mechanistic studies are important to understand cytokinesis relevance in normal cellular functions as well as its impact on human diseases.
In fungi and animal cells, the cytokinetic machinery comprises two major elements: septin and contractile actomyosin rings. In Saccharomyces cerevisiae, the process takes place at the division site, a narrow region linking mother and daughter cells known as the bud neck. At the end of G1, the Cdc42 GTPase polarizes actin cables and patches towards the new bud site, and the septin ring is assembled [2,3]. The septin ring is reorganized later in S phase and becomes an hourglass-like structure [4] that will split into two rings in anaphase [5][6][7]. Septins serve as an anchor for various cytokinesis-related proteins including the type II myosin heavy chain, Myo1 [8]. Myo1, together with actin filaments and the essential and regulatory myosin light chains Mlc1 and Mlc2, shape the actomyosin ring (AMR) [5,[8][9][10]. The AMR leads to furrow ingression through its contraction and constriction [11,12]. Cytokinetic regulatory proteins such as Iqg1, Hof1, Inn1 and Cyk3 are recruited at the AMR [6,[13][14][15], which mediate the activation of the chitin synthase II, Chs2 [6,[16][17][18]. Chs2 synthesizes the primary septum (PS), concomitantly with AMR constriction [19,20], and drives the second part of its contraction [21]. These proteins are part of the Ingression Progression Complexes (IPCs) [17,22], which coordinates primary septum formation, actomyosin ring contraction and ingression of the plasma membrane at the division site [16,17,[23][24][25][26].
Once the PS has been synthesized, the two secondary septa (SS) will be formed by Chs3, glucan synthases (Fks1) and mannosyltransferases on both sides of the PS [20,27,28]. Finally, cytokinesis is completed when the daughter cell synthesizes hydrolases and chitinases to hydrolyze the remaining cell wall structures in between SS [29].
Recent work demonstrates that multiple phosphatases shape the phosphoproteome during mitotic exit [48,49], pointing out the contribution of other phosphatases such as PP2A-Rts1 and PP2A-Cdc55 [48]. These observations suggest that PP2A phosphatases could also be required for cytokinesis. Here, we propose that PP2A-Cdc55 regulates cytokinesis through the dephosphorylation of the IPC protein, Chs2. We have found that the PP2A-Cdc55 participates in the regulation of the phosphorylation state of Chs2 and dephosphorylates Chs2 in vivo and in vitro. Primary septum formation and actomyosin ring contraction are asymmetric in the absence of PP2A-Cdc55, supporting a role for PP2A-Cdc55 in cytokinesis regulation. Interestingly, the non-phosphorylable version of Chs2, chs2-S133A rescues the asymmetric AMR contraction and the defective septa formation in cdc55∆ mutant cells. The phospho-mimetic chs2-S133E promotes a longer Hof1 contraction time in wild-type cells, similarly to the absence of Cdc55, suggesting that hyperphosphorylation of Chs2 alters Hof1 residence time at the division site. In conclusion, PP2A-Cdc55 regulates AMR contraction and PS formation via the dephosphorylation of Chs2 during cytokinesis.

Defective septum formation and AMR contraction in the absence of Cdc55
It was described that PP2A-Cdc55 is localized at the division site during cytokinesis [50]. However, neither a role for PP2A-Cdc55 during cytokinesis was demonstrated nor the molecular mechanism by which PP2A-Cdc55 could regulate cytokinesis. This prompted us to study the cytokinesis phenotypes of the cdc55Δ mutant cells.
Septins act as scaffold platforms for many proteins and impose a diffusion barrier for regulating cell polarity, cell remodeling and cytokinesis [51]. Septins localization and organization are highly dynamic through the cell cycle [52,53]. We wondered whether septin structures were defective in the absence of Cdc55. Septins were visualized in situ by immunofluorescence staining of Cdc11 and Shs1-HA in asynchronous cells. The septin structures (septin rings, collar and double rings) were indistinguishable between cdc55Δ and wild-type cells (Fig. S1), suggesting that septin dynamics might not be greatly affected by the absence of Cdc55.
PP2A-Cdc55 regulates bud morphology through actin polarization and cell-wall synthesis [54]. Actin cytoskeleton polarization at the site of bud emergence is triggered at the Start by the kinase activities of Cln1,2-Cdc28. Upon Clb2-Cdc28 kinase inactivation at the end of mitosis, the actin cytoskeleton is directed to the neck for the formation of the AMR [9,39]. We envisage the possibility that PP2A-Cdc55 influences also the actin cytoskeleton during cytokinesis. To examine this possibility, actin filaments were stained with rhodamine-labelled phalloidin in cells progressing through mitosis and cytokinesis in cdc55∆ cells. Actin signal was found to be depolarized in metaphase-arrested cells and localized at the bud neck in anaphase-cytokinesis in wildtype and cdc55Δ cells (Fig. S2a). However, in a subpopulation of cdc55Δ cells, actin polarization at the new bud was observed before cytokinesis had been completed (Fig. S2b).
We observed actin signals simultaneously at the division site during cytokinesis and at the new bud site in 12.5% of cdc55Δ mutant cells, whereas there was no premature actin re-polarization at the new bud site in wild-type cells (Fig.  S2b). The premature actin localization indicates that actin re-polarization of the next cell cycle occurs before cell division in a fraction of cdc55Δ cells. Therefore, PP2A-Cdc55 could act to prevent actin re-polarization until cytokinesis is completed.
Cytokinesis in budding yeast is accomplished by the concerted action of the actomyosin contractile ring (AMR) and the septum formation. We investigated how these two processes occur in the absence of Cdc55. First, to investigate whether the actomyosin contractile ring is functional, we analyzed the localization of the Myo1-tdTomato fusion protein in cells progressing through mitosis and cytokinesis after synchronous release from the metaphase arrest by Cdc20 depletion. It has been described that Myo1 localizes to the division site immediately after budding. Accordingly, in metaphase-arrested cells Myo1-tdTomato was localized at the bud neck in control cells. Similarly, Myo1-tdTomato was also detected at the bud neck during metaphase in cdc55Δ mutant cells. In anaphase, the Myo1 signal is reduced, reflecting the contraction of the AMR, until the signal becomes a single dot and finally disappears (Fig. 1a). The dynamics of AMR contraction were similar in the two strains, but we observed that AMR contraction was asymmetric with respect to the centripetal axis in 86% (N = 69) of cdc55Δ cells (Fig. 1b). To check whether the asymmetric Myo1 signal is not due to an adaptive mechanism of the cdc55Δ deletion mutant, we investigated the Myo1 contraction after inducing the Cdc55 degradation during metaphase using an auxin-degradation system [55]. Upon Cdc55 degradation, Myo1 contraction was also mostly asymmetric (86% of cells) (Fig. 1c). The results indicate that the lack of PP2A-Cdc55 activity promotes the asymmetric contraction of Myo1.
We next examined whether Myo1 asymmetry could also be detected in the inactive version of Cdc55 (cdc55-ED). We synchronized cells at the metaphase-anaphase transition by Cdc20 depletion and analyzed the contraction of the Myo1-tdTomato. We observed that 65% of cdc55-ED mutant cells showed an asymmetric Myo1-tdTomato signal upon contraction (Fig. 1c). This result indicates that, similar to the absence of Cdc55, the non-functional Cdc55 results in the alteration of AMR contraction. Overall, we conclude that Myo1 recruitment and time of localization at the division site are not altered in the absence of Cdc55, although AMR contraction is defective in the absence of PP2A-Cdc55 activity. This asymmetric localization has been reported before in some IPCs mutants [26] and suggests that PP2A-Cdc55 is required for efficient cytokinesis.
To assess whether the asymmetric AMR contraction phenotype was not affected by the synchronization method, we performed the assay synchronizing cells in G1 by alpha factor in absence of Cdc55. cdc55∆ cells enter mitosis with a delay due to compromised Cdk1 activity because of inhibitory Cdc28-Y19 phosphorylation [56]. To correct for this delay, we introduced the cdc28_Y19F allele, which is refractory to Cdk1 inhibition. The cdc55Δ cells containing cdc28_Y19F progressed normally through the cell cycle [57]. Again, we observed that Myo1-tdTomato constriction was asymmetric in 80% of the cdc55Δ cdc28_Y19F cell population (Fig. S3). Therefore, the asymmetric Myo1 signal in the absence of Cdc55 was observed independently of the synchronization method used.
To study septum formation, chitin deposition at the division site was analyzed by in vivo staining with calcofluor white and measured the fluorescence intensity of the incorporated calcofluor on living cells containing Myo1-tdTomato as a control for cytokinesis progression. We arrested cells at metaphase by Cdc20 depletion, released them into mitosis and took images 40-50 min after the release when we found cells at cytokinesis. Calcofluor intensity was then measured and quantified in wild-type and cdc55Δ cells (Fig. 1d). There was a 45% reduction in the intensity of the calcofluor staining in the absence of Cdc55 compared to control cells. Chitin is incorporated in the primary septum (PS) and secondary septa (SS); therefore, we cannot distinguish at which septum the reduction occurs. For this reason, we repeated the calcofluor staining in cells containing a deletion for CHS3, the chitin synthase responsible for secondary septa formation [20]. As previously described, chs3Δ cells showed wider bud necks and cytokinetic defects provoking the formation of chains of cells [58]. A reduction of 70% in the calcofluor intensity was measured in the chs3Δ background in the absence of Cdc55 (Fig. 1e). Interestingly, we noticed that 82% (N = 22) of the cells showed asymmetric Myo1 contraction in chs3Δ cells (Fig. 1c), indicating defects in cytokinesis [26]. The above results indicate that primary and secondary septa formation are reduced in cells lacking PP2A-Cdc55 activity, suggesting that PP2A-Cdc55 has a role regulating septum formation.
Primary septum formation and membrane invagination occur concomitantly. To determine whether ingression of the plasma membrane could occur in cdc55Δ cells, we performed time-lapse video microscopy of cells expressing the small G-protein Ras2 fused to 3 copies of GFP to study plasma membrane dynamics. The visualization of the plasma membrane at the site of division revealed no cytoplasm connection between mother and daughter cells after cytokinesis, confirming that cytoplasmic division is resolved in control and cdc55Δ cells (Fig. S4). In conclusion, cdc55Δ cells have an asymmetric AMR contraction and a reduction in septa formation, but nevertheless manage to complete cell division.
Next, to study the defects in septa formation we investigated the cytokinetic structure by transmission electron microscopy (TEM). We synchronized cells at the metaphaseanaphase transition by Cdc20 depletion and captured images 40-50 min after the synchronous release. Cells synthesizing the primary septum were identified and the structure of the septum was then examined. In wild-type cells, we observed PS formation from the cell wall following membrane invagination (Fig. 2a), as expected. Once the PS was finished, the two SS were formed on both sides of the PS Y1596 (N = 86) were arrested in metaphase by Cdc20 depletion and released into anaphase by Cdc20 re-induction. Calcofluor was added to visualize the septa upon the metaphase release. Student's unpaired t-test was carried out using the Prism5 program ( Fig. 2b). Finally, the PS is degraded, and the cells physically separated (Fig. 2c). Conversely, in cdc55∆ cells, the PS emerged from only one side of the division site ( Fig. 2d). From the cells performing primary septum formation, we quantified 85% (N = 25) with asymmetric PS in absence of Cdc55. Some cells showed aberrant, thicker structures with diverse morphologies that resemble the secondary septa-like structures commonly referred as remedial septum (Fig. 2e). The remedial septum was first described in IPCs mutant cells, i.e., myo1∆, chs2∆ and chs3∆ cells [24]. They are chitin structures reminiscent of SS that allow cells to complete cytokinesis when the PS is defective [24,59]. Strikingly, in cdc55∆ cells, the mother and daughter cells are finally able to separate physically, probably upon formation of the remedial-like septum. To further investigate this hypothesis, we repeated the experiments using a chs3∆ background in which no secondary septa is formed ( Fig. 2g-i). Similar results were obtained, whence the cdc55∆ chs3∆ cells had asymmetric PS formation and a remedial-like septum was formed ( Fig. 2j-l). These results suggest that remedial-like septum formation enables cytoplasm separation and cell division in the absence of Cdc55.
Since AMR constriction and septum formation are simultaneous events, the asymmetric PS formation detected by EM is likely to promote the asymmetric AMR contraction observed in cdc55∆ cells or vice versa. In summary, our results showed that PP2A-Cdc55 has a role in regulating AMR contraction and PS formation.

PP2A-Cdc55 regulates the dephosphorylation of Chs2
The Ingression Progression Complexes (IPCs) coordinate AMR constriction, plasma membrane ingression and septum formation [17,18]. We envisaged that PP2A-Cdc55 may have a role during cytokinesis through the dephosphorylation of IPCs proteins [60]. We initially investigated the genetic interactions between cdc55∆ and IPCs mutants. We prepared double mutants with cdc55∆ and degron-conditional mutants to induce protein degradation for the IPCs subunits [17,61]. No differences in cell growth were observed on control plates (Fig. 3a). However, the viability of the double mutants cdc55∆ hof1-aid and cdc55∆ td-cyk3-aid was impaired under restrictive conditions (presence of auxin) (Fig. 3a). The synthetic sick interactions found between cdc55∆ and hof1-aid or td-cyk3-aid degron mutants indicate that Cdc55 is functionally related to the IPC subunits, suggesting that Cdc55 is involved in the regulation of cytokinesis.
To determine whether the IPCs proteins physically interact with PP2A-Cdc55, we performed co-immunoprecipitation experiments between Cdc55 and Chs2. We synchronized cells in metaphase as described above, and co-purification of Chs2 during anaphase and cytokinesis were detected in Cdc55 immunoprecipitates (Fig. 3b). Taken together, these results suggest that Chs2 forms a complex with PP2A-Cdc55 during progression through mitosis and cytokinesis.
To study the functional link between Cdc55 and the IPCs, we studied whether PP2A-Cdc55 regulates the dephosphorylation of IPC proteins. We analyzed the phosphorylation status of the IPC proteins in the absence of Cdc55 ( Fig. 3c and S5a-d) and observed an increase of Chs2 phosphorylation levels during cytokinesis. First, we synchronized cells at metaphase and released them synchronously into anaphase to visualize the phosphorylation of Chs2 during mitosis and cytokinesis. Mitotic spindle was stained by tubulin immunofluorescence and used as a marker of cell-cycle progression. Cdc14 release from the nucleolus was also determined as control of mitosis * m e t a p h a s e a n a p h a s e -G 1 (% cells) The double mutants cdc55 hof1 and cdc55cyk3 are synthetically sick. Serial dilutions ofW303, Y844, Y1728, Y1730, Y1747 and Y1749 strains were spotted on YPD plates with and without 1.5 mM auxin (IAA). Cells were grown at 25ºC for 2-3 days. (b) PP2A-Cdc55 forms a complex with Chs2. Strain Y1567 was synchronized into anaphase progression by Cdc20 depletion and re-addition. Strain Y1318 without the Pk-tag was used as a negative control. Protein extracts were prepared at metaphase, anaphase, and cytokinesis and Cdc55 was immunoprecipitated using the anti-Pk antibody. Copurification of Chs2 was analyzed by western blot. (c) PP2A-Cdc55 regulates Chs2 dephosphorylation. Strains Y1318 and Y1315 were arrested in metaphase and released into anaphase by Cdc20 depletion and re-addition. Chs2 phosphorylation was analyzed by western blot in Phos-tag gels. Unspecific bands were detected for Chs2 and marked as (ª). Pgk1 levels were used as a loading control. Mitosis progression was followed by analyzing the anaphase spindle elongation by in situ immunofluorescence. At least 100 cells were scored at each time point. Quantification of the western blots was performed using Fiji Software and means and SEMs are represented. Student's unpaired t-test analyses were carried out using the Prism5 program.
(d) Chs2 is dephosphorylated by PP2A-Cdc55. Metaphase arrested cells of the strain Y695 were used to purify the PP2A-Cdc55 complex by TAP purification. Strain Y824 without the TAP epitope was used to prepare a Cdc55-free purification and used as a negative control. Chs2-HA was purified from metaphase arrested cells of strain Y1318. Purified Chs2-HA was incubated with or without the PP2A-Cdc55 complex at the indicated times and the Chs2 phosphorylation levels were detected by western blot using the anti-phospho Ser/Thr antibody. Representative images of one phosphatase assay are shown. Protein levels were quantified using Fiji software. Quantifications of the remaining Chs2 phosphorylation signal normalized to the amount of Cdc55 and Chs2 are shown. Means and SEMs of three phosphatase assays are represented. Student's unpaired t-test analysis was carried out using the Prism5 program. (e) PP2A-Cdc55 dephosphorylates Chs2 in vitro. Strains Y824, Y1652 and Y1653 were arrested in metaphase by Cdc20 depletion. Cdc55 and cdc55-ED were purified by immunoprecipitation with HA antibody. Streptag-Chs2-1-629 purified from E. coli and previously phosphorylated by Cdk1-Clb2 in a radioactive kinase assay was used as substrate.
Representative images of the phosphatase assays are shown. Radioactive signals were detected using a multi-purpose imaging plate in a Typhoon FLA950 apparatus (GE healthcare). Protein levels were quantified using Fiji software. Quantifications of the remaining phosphorylation signals normalized with respect to the amount of immunopurified HA-Cdc55 or HA-cdc55-ED are shown. Means and SEMs of three phosphatase assays are represented. Student's unpaired t-test analysis was carried out using the Prism5 program progression. Cdc14 was prematurely released in metaphase in absence of Cdc55 as previously published [57]. During the time-course, Chs2 protein presented slow migration isoforms at metaphase and early anaphase in control cells (Fig. 3c meta-20 min). From anaphase until early G1 ( Fig. 3c; 30-50 min), two Chs2 isoforms were detected. Since Chs2 presented many different migrating bands, we performed an alkaline phosphatase experiment to check which corresponds to phosphorylation. In presence of alkaline phosphatase, the slower Chs2 migrating bands collapse into the faster-migrating band indicating that the upper bands are phosphorylation events (Fig.  S5e). Remarkably, in cdc55∆ mutant cells the slowermigrating isoforms are maintained throughout the timecourse, indicating that Chs2 was not efficiently dephosphorylated during late anaphase/cytokinesis ( Fig. 3c; 40-70 min). These results suggest that Chs2 phosphorylation levels are altered in the absence of Cdc55 during cytokinesis.
Later, to explicitly test whether Chs2 is dephosphorylated by PP2A-Cdc55, we measured PP2A-Cdc55 phosphatase activity against Chs2 after PP2A-Cdc55 purification from metaphase-arrested cells when PP2A-Cdc55 is active [57]. Immunopurified-Chs2 was incubated with purified PP2A-Cdc55 and Chs2 dephosphorylation was visualized using an anti-phospho Ser/Thr antibody (Fig. 3d). A Cdc55-free extract incubated with Chs2 for 40 min was used as a negative control. A 60% reduction in the Chs2 phosphorylation signal was measured after 40 min of incubation in presence of PP2A-Cdc55. As a negative control, the phosphatase assay was done in the same conditions using separase (Esp1) as substrate. Immunopurified-Esp1 was incubated with purified PP2A-Cdc55 and the phosphorylation levels of Esp1 visualized with the anti-phospho Ser/Thr antibody (Fig. S5f). Esp1 phosphorylation signal was maintained during the experiment indicating that PP2A-Cdc55 is not able to dephosphorylate Esp1 in the same conditions used for Chs2. This result suggests that PP2A-Cdc55 dephosphorylation occurs specifically over Chs2.
Finally, we repeated the PP2A-Cdc55 phosphatase assay using a recombinant fragment of Chs2 (1-629) purified from E. coli and previously phosphorylated by Cdk1-Clb2. In vitro 32 P-phosphorylated Chs2 substrate was incubated with Cdc55 immunoprecipitated from control cells expressing Cdc55 and cdc55-ED inactive version of PP2A-Cdc55 [62]. A reduction of 84% in the Chs2 phosphorylation signal was detected in the presence of control Cdc55. The cdc55-ED mutant versions were not able to dephosphorylate Chs2 significantly (Fig. 3e). These results strongly suggest that Chs2 is dephosphorylated by PP2A-Cdc55 in vivo and in vitro.

PP2A-Cdc55 regulates IPCs symmetric localization and residence time at the division site
IPCs proteins form a complex whose function is to coordinate AMR contraction, plasma membrane ingression and PS formation [17,18]. For this reason, we examined whether the absence of Cdc55 disturbs the IPCs subunits proper or timely localization at the division site. To do this, we studied the localization of the IPCs subunits by time-lapse microscopy in the absence of Cdc55. We arrested cells in metaphase by Cdc20 depletion and captured images every 2 min after synchronous release into anaphase by Cdc20 reinduction. Iqg1, Inn1, Hof1, Cyk3 and Chs2 have contracted asymmetrically at the division site in cdc55Δ cells (Fig. 4a), consistent with the asymmetry of Myo1 described above. These findings indicate that the lack of PP2A-Cdc55 activity provokes a defective contraction of the AMR, probably as a consequence of the defective PS formation or vice versa.
We then investigated the contraction and residence time of the IPCs subunits (Fig. 4b). No differences were detected in the residence time at the bud neck in Iqg1-GFP and Inn1-GFP proteins ( Fig. 4b and Fig. S6a-b). However, we observed that Hof1-GFP and Cyk3-GFP signals took longer to complete contraction and to disappear in cdc55∆ cells ( Fig. 4b and Fig. S6c-d), further indicating that cytokinesis is affected in absence of PP2A-Cdc55. The longer localization of Hof1 and Cyk3 at the division site could be a result of the impaired AMR contraction.
On the contrary, the Chs2-GFP residence time in cdc55Δ mutant cells was reduced compared to wild-type cells ( Fig. 4b and S6e). Therefore, this result suggests that PP2A-Cdc55 regulates Chs2 localization dynamics, probably through the dephosphorylation of Chs2.

Timely Chs2 dephosphorylation by PP2A-Cdc55 is required for proper AMR contraction and PS formation
Our results suggest that PP2A-Cdc55 regulates Chs2dependent processes during cytokinesis, such as PS formation and AMR contraction, and that participates in the dephosphorylation of Chs2. Therefore, PP2A-Cdc55 might contribute to cytokinesis regulation by controlling the phosphorylation levels of Chs2.
To screen for proteins regulated by the PP2A-Cdc55 phosphatase we studied the phosphoproteome in absence of Cdc55 by a quantitative analysis based on SILAC labeling. Phosphopeptides were enriched by Tish enrichment as described [63] and identified by LC-MS/MS. The screening revealed that a phosphopeptide corresponding to Chs2 protein was hyperphosphorylated in cdc55∆ mutant cells. The Chs2 peptide contained one Cdk1 minimal S/ TP site: S133 (Fig. 5a). The phosphosite was detected with the highest confidence (pRS site probability 99.9% and p-value < 0.00001) and was also identified in a second phospho-proteomic study for cdc55∆ mutant cells [49]. The S133 is one of the 6 Cdk1 consensus sites (S/T-P) previously reported to be phosphorylated by Cdk1 [64]. Chs2's residues S14, S60, S69 and S100 containing the SP consensus sequence were described to be efficiently dephosphorylated by Cdc14, and the phosphomimetic mutant at those four sites, chs2-S4E, promotes Chs2 constitutive localization in the ER [41,47]. Cdc14 is active as a phosphatase and prematurely released from the nucleolus in cdc55∆ mutant cells [57,65]; therefore, the S133 hyperphosphorylation is a consequence of the lack of PP2A-Cdc55 activity since Cdc14 is not defective in absence of Cdc55.
To investigate whether the S133 is a critical Chs2 phosphorylation site, we introduced a non-phosphorylable version of Chs2, GAL1-chs2-S133A-YFP, in wild-type cells and analyzed the phosphorylation levels in Chs2-YFP immunoprecipitates using the anti-phospho Ser/Thr antibody. Chs2 phosphorylation was detected in control cells expressing the wild-type version of Chs2-YFP (Fig. 5b). By contrast, the signal was greatly reduced in the immunoprecipitation of the non-phosphorylable version chs2-S133A (Fig. 5b). This result suggests that the S133 of Chs2 is phosphorylated.
If the cytokinetic defects observed in cdc55∆ mutant cells are a consequence of the hyperphosphorylation of Chs2, the non-phosphorylable Chs2 version should rescue the asymmetric Myo1 contraction and PS formation. For this reason, we introduced the non-phosphorylable version GAL1-chs2-S133A-YFP in wild-type and cdc55∆ mutant strains. Cells were synchronized by Cdc20 depletion and re-addition and GAL1-CHS2 versions were induced prior to metaphase release. First, we followed Myo1-tdTomato and Chs2-YFP dynamics during mitosis and cytokinesis by time-lapse microscopy. Myo1 signal was asymmetric in the presence of the control Chs2-YFP (80% of cells; N = 25), as expected in a cdc55∆ background (Fig. 5c). Remarkably, in the chs2-S133A non-phosphorylable version Myo1 became symmetric in 90% of the cells (N = 31) (Fig. 5c). This indicates that the non-phosphorylable Chs2 mutant rescued the asymmetric localization of Myo1 in absence of Cdc55. Next, we wondered whether the chs2-S133A version also rescued the longer Hof1 contraction time observed in cdc55∆ cells (see Fig. 4b) by investigating the Hof1-mCherry dynamics. Hof1 contraction was similar in Chs2 and chs2-S133A in absence of Cdc55 (Fig. 5d), indicating that the delay in Hof1 contraction in absence of Cdc55 is not due to Chs2 phosphorylation. The Hof1 constriction times are longer, compared to Fig. 4b, presumably due to Chs2 overexpression. Finally, we investigated the cytokinetic structure by transmission electron microscopy (TEM). Upon introduction of Chs2 and chs2-S133A in cdc55∆ cells, PS formation became symmetric, supporting that Chs2 activity is impaired in absence of Cdc55. Nevertheless, PS structure was aberrant (curved) in presence of the control Chs2 and remedial septum formation was still detected (Fig. 5e). By contrast, PS structure was fully recovered and similar to WT cells upon introduction of the non-phosphorylable chs2-S133A in cdc55∆ cells. In conclusion, the non-phosphorylable chs2-S133A version rescued the Myo1 asymmetric signal and the defective PS formation, suggesting that the S133 is the main dephosphorylated residue regulated by PP2A-Cdc55. Finally, to investigate whether the hyperphosphorylation of Chs2 is sufficient to alter AMR symmetric contraction we introduced the phospho-mimetic chs2-S133E version in wild-type cells (Fig. 5c). No significant increase in the asymmetric Myo1 signal in the presence of chs2-S133E was observed suggesting that Chs2 phosphorylation is not sufficient to promote asymmetric AMR contraction. Nevertheless, an increase in the Hof1 residence time was observed in the phospho-mimetic chs2-S133E version compared to the control Chs2 and chs2-S133A (Fig. 5d), similarly to what we observed in cdc55∆ mutant cells (compare WT and cdc55∆ in Fig. 4b). These results suggest that PP2A-Cdc55 has a role in cytokinesis regulating Chs2 phosphorylation.
Overall, our findings demonstrate that the lack of PP2A-Cdc55 activity provokes an asymmetric contraction of the AMR and a defective PS formation that depends on the tight regulation of Chs2 phosphorylation. Therefore, Chs2 dephosphorylation by PP2A-Cdc55 is essential for maintaining proper AMR contraction and septum formation; both are required for an efficient cytokinesis.

PP2A-Cdc55 regulates Chs2 phosphorylation and its localization dynamics at the division site
The localization and function of the IPCs subunits (Iqg1, Myo1, Hof1, Cyk3, Inn1 and Chs2 proteins), which coordinate AMR contraction, plasma membrane ingression and PS formation, are tightly regulated by phosphorylation [43,44,66]. To date, only two phosphatases, Cdc14 and PP2A-Rts1, are known to have a role in cytokinesis [67], and it has been described that Cdc14 dephosphorylates several subunits of the IPCs proteins such as Chs2, Inn1 and Hof1 [39,41,44,46]. A putative role for PP2A-Cdc55 in cytokinesis was also proposed because of its elongated morphology and multinucleated cells at low temperatures in cdc55∆ cells [68]. In addition, PP2A-Cdc55 is localized at the division site during cytokinesis [50]. However, the role of PP2A-Cdc55 during cytokinesis was not demonstrated and the molecular mechanism by which PP2A-Cdc55 regulates cytokinesis is still unknown.
Here, we describe a new function of PP2A-Cdc55 phosphatase in the regulation of cytokinesis. In summary, PP2A-Cdc55 regulates the dephosphorylation of Chs2 and it is involved in proper AMR contraction and septum formation. PP2A-Cdc55 dephosphorylates Chs2 in vivo and in vitro, indicating that it is a PP2A-Cdc55 target. The lack of dephosphorylation of Chs2 in absence of Cdc55 might be responsible for the defects seen in IPCs (Hof1) residence time (Fig. 4b) and AMR asymmetric contraction since the chs2-S133A (Fig. 5c, d) non-phosphorylable mutant rescues the asymmetric Myo1 signal in cdc55∆ mutant cells. Moreover, Chs2 overexpression is able to rescue the asymmetric PS formation although its morphology is still impaired, further suggesting that Chs2 activity is defective in absence of Cdc55. Only the introduction of the non-phosphorylable chs2-S133A is able to rescue proper PS formation (Fig. 5e). In contrast, the phospho-mimetic mutant chs2-S133E is not sufficient to induce an asymmetric Myo1 signal in control cells, although it provokes an increase of Hof1 residence time at the division site. These results suggest that timely dephosphorylation of Chs2 by PP2A-Cdc55 is crucial for proper AMR contraction and PS formation. Since the IPCs coordinate AMR constriction, plasma membrane ingression and septum formation, we propose that PP2A-Cdc55 is also involved in this efficient mechanism of regulation through Chs2 dephosphorylation (Fig. 6). This would explain the double mutants td-cyk3-aid cdc55Δ and hof1-aid cdc55Δ synthetic sick phenotypes, consistently with Hof1 and Cyk3 acting in parallel to activate Chs2, and in accordance with the cyk3Δ hof1Δ synthetic lethal phenotype [14,18,43,60] The role of PP2A-Cdc55 during cytokinesis In the absence of Cdc55, Myo1 was constricted and unanchored in a timely fashion, but the constriction was displaced from the central axis, becoming asymmetric (Fig. 1a). This phenotype is characteristic of IPCs mutants [6,26,52] and denotes a dysfunctional AMR.
On the other hand, Hof1 asymmetric PS formation was described in chs2∆ cells [6]; but not vice versa [6,17], suggesting that Hof1 also influences PS formation. Therefore, defective PS formation in absence of Cdc55 could also contribute to the asymmetry observed in the other IPCs proteins.
In addition, PP2A-Cdc55 regulates septum formation. We observed that chitin incorporation was severely reduced in cdc55∆ and cdc55∆ chs3∆ cells (Fig. 1d), indicating that Chs2 activity might be impaired. Interestingly, the electron microscopy study demonstrated that the PS appears only at one side of the division site during its formation in absence of Cdc55 (Fig. 2d, j). Later, a remedial-like septum that contains even cytosol fractions embedded in the septa is detected in absence of Cdc55 (Fig. 2f, l). The remedial septum formation was previously described in IPCs and MEN mutants with defective PS formation [13,18,69,70]. Chs3 glucan synthases and mannosyltransferases are responsible for the SS formation and the construction of the remedial septum [27,28,59]. However, the chs3Δ mutant is still viable and capable of dividing, suggesting that additional chitin synthases might be able to incorporate chitin in the absence of Chs3. This may explain why the remedial-like septum is still formed in the cdc55Δ chs3Δ double mutant (Fig. 2i, l).
Recently, it has been shown that multiple phosphatases shape the phospho-proteome during mitotic exit [48,49,71]. Notably, Hof1 and Chs2 phosphosites have been identified in cdc55 and cdc14 mutants [48]. We demonstrated that PP2A-Cdc55 regulates cytokinesis through the Chs2 dephosphorylation and it was reported that Cdc14-dependent dephosphorylation is required for Chs2 release from the ER [41], supporting different and specific contributions to cytokinesis of several phosphatases. We observed hyperphosphorylations of Hof1 and Inn1 during cytokinesis in absence of Cdc55 (Fig. S5a-b), suggesting that PP2A-Cdc55 might also regulate other IPCs proteins. Further investigations will be The PP2A-Cdc55 homologue in fission yeast, PP2A-Pab1 was described to be involved in sporulation, cell wall synthesis and cytokinesis [72] suggesting that the PP2A-Cdc55 role during cytokinesis might be evolutionarily conserved. Studying the role of PP2A in model organisms will help extend our knowledge of cytokinesis in mammalian cells.

Yeast strains, plasmids and cell-cycle synchronization procedures
All yeast strains used in this study were derivatives of W303 and are listed in Table S1. Epitope tagging of endogenous genes and gene deletions were performed by gene targeting using polymerase chain reaction (PCR) products [73].
Cell synchronization using α-factor and metaphase arrest by Cdc20 depletion and entry into synchronous anaphase by Cdc20 re-introduction was performed as previously described [74]. To obtain the pGAL-CHS2-S133A-YFP construct, the S133A mutation was introduced by DNA assembly using the pGAL1-CHS2-YFP backbone [47].
All the cdc55∆ mutant strains were freshly prepared by transformation or by crossing strains to avoid the accumulation of suppressor mutations. The introduction of the cdc55∆ deletion was determined by PCR-genotyping and observation of the slow growth and elongated morphology under the microscope.

Recombinant protein purification
All plasmids were freshly introduced into BL21 E. coli. Cells were grown in LB medium and protein expression was induced with 0.1 mM isopropyl β-D-1-thiogalactopyranoside (IPTG) at 25ºC overnight. Collected cells were washed with PBS1 and frozen for at least 30 min.

Co-immunoprecipitation, kinase assays and phosphatase assays
Co-immunoprecipitation assays were performed using 10 8 yeast cells, which were resuspended in lysis buffer (50 mM HEPES-KOH pH 7.5, 70 mM KOAc, 5 mM Mg(OAc) 2 , 10% glycerol, 0.1% Triton X-100, 8 μg/ mL of protease inhibitors (leupeptin, pepstatin, aprotinin), 1 mM PMSF, 1.25 mg/ml of benzamidin, 1 tablet Fig. 5 Chs2 phospho-mutants at S133 regulate AMR contraction and Hof1 residence time at the division site. (a) Hyperphosphorylation of Chs2 in S133 was detected by mass spectrometry in absence of Cdc55. MS/MS spectrum of the Chs2 peptide containing the S133 phosphorylated from one replicate is shown. (b) Chs2 phosphorylation is greatly reduced in the chs2-S133A mutant. Strains Y1901 and Y1900 were arrested in metaphase and GAL1-CHS2 versions were induced by galactose addition for 3 h. Protein extracts were prepared and Chs2 and chs2-S133A purified using GFP-trap columns. Chs2 phosphorylation levels were detected by western blot using the antiphospho Ser/Thr antibody. Representative blots of one assay are shown (left panel). Western blot signals were quantified using Fiji software. Quantifications of the Chs2 phosphorylation levels normalized to the amount of Chs2 from three independent experiments are shown (right panel). Means and SEMs are represented. Student's unpaired t-test analysis was performed using the Prism5 program.  Roche)). Lysates were obtained by mechanical lysis using glass beads in a Bertin disrupter (6 cycles of 10 s at 5,000 rpm). Protein extracts were clarified by centrifugation and incubated with α-Pk clone SV5-Pk1 (AbSerotec) orα-HA clone 12CA5 (Roche) antibodies for 1 h. Protein extracts were then incubated for 1 h with protein A-conjugated Dynabeads (Life Technologies), after which the beads were washed with lysis buffer at incremental KOAc concentrations (100 mM, 120 mM, 150 mM, and 60 mM NaOAc). Co-immunoprecipitated proteins were eluted with loading buffer and protein copurification was visualized by western blot.
To purify the PP2A-Cdc55 complex, a strain with the tandem affinity protein (TAP) epitope, TAP-HA-Cdc55 was used. A Cdc55-free pull-down was prepared using a strain without the TAP-HA epitope tagged for Cdc55 and used as a negative control. Cells from 2 L cultures were arrested in metaphase by Cdc20 depletion, frozen in liquid nitrogen and stored at -80ºC. Cells were resuspended in lysis buffer (0.1 M NaCl, 50 mM Tris-HCl pH 7.5, 1.5 mM MgCl 2 and 0.15% NP-40, 8 μg/mL of protease inhibitors (leupeptin, pepstatin, aprotinin), 1 mM PMSF, 1.25 mg/ml of benzamidin, 1 tablet of complete protease inhibitor without EDTA (Roche)) and protein extracts prepared by mechanical lysis. Protein extracts were clarified by centrifugation at 20.000 g and incubated with 0.4 mL IgG-Sepharose beads for 1 h. The beads were extensively washed and incubated with 3 μL of TEV protease at 16ºC for 2 h to remove the CBP-HA-Cdc55 from the beads. The purified PP2A-Cdc55 was stored at -80ºC and used for the phosphatase assays in Fig. 3d and Fig. S5f.
The Chs2-HA and Esp1-HA were purified following the co-immunoprecipitation protocol as above but using Pierce HA-magnetic beads and eluted from the beads using 1 mg/ mL of HA peptide. For the phosphatase assays with purified PP2A-Cdc55, eluted Chs2 or Esp1 were incubated with phosphatase reaction buffer (50 mM Tris-HCl pH7.4, 0.1 mM EGTA, 1 mM β-mercaptoethanol, 1 mg/mL BSA) and the purified CBP-HA-Cdc55 at 30ºC for 20-40 min. Reactions were terminated by adding SDS-PAGE loading buffer. Proteins were separated by electrophoresis, transferred to PVDF membranes and phosphorylation levels were detected with the anti-phospho Ser/Thr antibody (Cell Signaling) using the Supersignal West Atto Ultimate Sensitivity Substrate (Pierce). Anti-HA western blots were then performed to analyze and quantify the amount of Chs2, Esp1 and Cdc55. The phosphorylation levels were normalized against Cdc55 and Chs2 or Esp1. Proteins were quantified using Fiji software [75] and values of the mean and SEM were calculated.
For radioactive 32 P kinase assays, Clb2-Pk immunoprecipitations were performed as above and incubated with α-Pk clone SV5-Pk1 (AbSerotec) antibody. Beads were washed with ten volumes of lysis buffer and twice with the kinase reaction buffer (50 mM Tris-HCl, pH7.4, 10 mM MgCl 2 , 1 mM DTT). The kinase reaction (50 mM Tris-HCl, pH7.4, 10 mM MgCl 2 , 1 mM DTT, 5 mM β-glycerophosphate, 25 μM ATP, 10 mCi/mL 32 gamma-ATP and 1 μg of Streptag-Chs2) was incubated at 30ºC for 30 min. Kinase assays were stopped by placing the tubes on ice. The supernatant containing the phosphorylated substrate was separated from the magnetic beads and stored at -80ºC. An aliquot of the kinase assay was mixed with SDS-PAGE loading buffer, proteins were separated by electrophoresis, transferred to nitrocellulose membranes and radioactivity detected in a Typhoon FLA950 (GE Healthcare). Immunopurified protein was quantified by western blot and the membrane was stained with Coomassie to detect the recombinant substrate. Proteins were quantified using Fiji software.
For in vitro phosphatase assays, cells containing HA 3 -Cdc55-ED and HA 3 -Cdc55 were collected from metaphase arrested cells. Immunoprecipitation was performed as above but without phosphatase inhibitors. Beads were washed twice with phosphatase buffer and incubated with Fig. 6 Model for PP2A-Cdc55 regulation of actomyosin ring (AMR) contraction, primary septum (PS) and secondary septa (SS) formation. Dephosphorylation of Chs2 by PP2A-Cdc55 regulates AMR contraction and primary septum formation resulting in an increased Hof1 residence time at the division site the phosphatase reaction (50 mM Tris-HCl pH7.4, 0.1 mM EGTA, 1 mM β-mercaptoethanol, 1 mg/mL BSA and the 32 P-Streptag-Chs2) at 30ºC for 40 min. Reactions were terminated by adding SDS-PAGE loading buffer. Proteins were separated by electrophoresis, transferred to nitrocellulose membranes and radioactivity was detected in a Typhoon FLA950 apparatus (GE Healthcare). Later, the membrane was used to analyze and quantify the amount of protein immunoprecipitated by western blot. Finally, the membrane was stained with Coomassie to detect the recombinant Chs2 substrate. Proteins levels were quantified using Fiji software and values of the mean and SEM were calculated.
The GAL1-YFP-Chs2 and GAL1-YFP-Chs2-S133A were purified following the co-immunoprecipitation protocol as above but using GFP-Trap beads (Chromotek). Proteins were separated by electrophoresis, transferred to PVDF membranes and phosphorylation levels were detected with the anti-phospho Ser/Thr antibody (Cell Signaling) using the Supersignal West Atto Ultimate Sensitivity Substrate (Pierce). Anti-GFP western blots were then performed to analyze and quantify the amount of Chs2. The phosphorylation levels were normalized against total Chs2. Protein levels were quantified using Fiji software and values of the mean and SEM calculated.

Microscopy techniques
Synchronized cells for time-lapse experiments were deposited in chambers (Nunc Lab-Tek) containing concanavalin A/ PBS 1 mg/mL. Images were captured every 2 min. Different z-stacks at 0.7-μm intervals were taken and projected onto a single image per channel. A Zeiss-Apotome epifluorescence microscope with an HXP 120C fluorescent lamp and a Carl Zeiss Plan-Apochromat 63 × N.A 1.40 oil immersion lens were used. The filters used were Cy3, GFP, and DAPI. For the GFP-Ras2 experiments, the Apotome was disabled. For the Cyk3-GFP time-lapse experiments, a Carl Zeiss LSM880 confocal microscope, with a 63 × N.A. objective was used. Images were acquired using ZEN software. For the time-lapses in Fig. 5, a Leica SP8 confocal microscope with a 63X objective was used. Images were acquired using LAS software and quantified and processed using Fiji software. To determine the symmetry of the IPCs proteins, the bud neck length was measured, the central point was calculated and manually established whether the central point was inside (symmetry) or outside (asymmetry) the dot signal.
Calcofluor staining was performed with calcofluor white MR2 (Fluorescent brightener 28, Sigma) in living cells. Synchronized cells were incubated with 50 μg/mL calcofluor upon release from the metaphase arrest. Images were acquired, quantified and processed as for time-lapse experiments.
Phalloidin staining was performed on fixed cells. Cells collected during the time-course experiments were prefixed with PBS containing 3.7% formaldehyde and 0.1% Triton X-100 at 25ºC for 10 min. Cells were then washed with PBS1 and fixed with 3.7% formaldehyde at 25ºC for 1 h. After fixation, cells were washed twice with PBS1 and sedimented onto a multi-well slide previously incubated with poly-L-lysine. Cells were stained with a PBS solution containing 50 U/mL rhodamine-phalloidin R415 (Life Technologies) for 2 h. Cells were washed twice with PBS1. Mounting medium containing DAPI (Vectashield) was added, cells were covered with a coverslip and sealed with nail polish. Images were acquired, quantified and processed as for timelapse experiments.
For electron microscopy, cell samples were fixed with 0.2 M phosphate buffer (without salts) pH 7.4, containing 2.5% glutaraldehyde at 25ºC, for 1 h. Cells were washed three times with phosphate buffer without glutaraldehyde and rinsed with milli-Q water and post-fixed with 1% osmium tetroxide for 2 h. They were then dehydrated in an acetone series (10%, 20%, 30%, 40%, 60%, 80%, and 100%) for 15-20 min. Ultrathin sections of 60-nm thickness were obtained using a UC6 ultramicrotome (Leica Microsystems, Austria) and stained with 2% uranyl acetate and lead citrate. Sections were observed in a Jeol EM J1010 apparatus (Jeol, Japan) and images were acquired at 80 kV with a 1 k CCD Megaview camera.

SILAC labeling and phosphoproteomic analysis
Y859 (MATa lys2∆::TRP1, arg4∆::HIS3 MET-CDC20::LEU2) and Y858 (as Y859, but cdc55∆) cells were grown in free-methionine minimum media containing 13 C 6lysine and -arginine (heavy, Cambridge Isotope Laboratories Inc., US) or unmodified arginine and lysine (light), respectively. Protein extracts were prepared by mechanical lysis using glass beads in presence of protein inhibitors (Complete EDTA-free, Roche) and 2 × phosphatase inhibitors PhosStop (Roche). Cell lysates were mixed 1:1 and digestion with trypsin was performed. The phosphopeptides were pre-enriched using TiO 2 chromatography (GL Sciences Inc) followed by SIMAC purification [63]. The mono-phosphorylated peptide fraction from the SIMAC enrichment was further subjected to a second TiO 2 purification and pre-fractionated by HILIC chromatography (Hydrophilic Interaction Liquid Chromatography). The peptides were then analyzed by LC-MS/MS using a LTQ-Orbitrap Fusion Tribride mass spectrometer (Thermo Fisher Scientific). Peptide identification was performed using Proteome Discoverer v1.4.1.14 (Thermo Scientific) and search against the Swiss Prot/Uniprot Saccharomyces cerevisiae database. Peptide quantification from SILAC labels was performed with Proteome Discoverer v1.4.

Other techniques
Protein extracts for western blots were obtained by TCA protein extraction. For the phostag gels, 10 μM Phostag (Wako) 5-8% gels containing 10 mM MnCl 2 were used. Protein gels were washed with 1 mM EDTA before transferring protein to remove the manganese. Proteins were quantified using Fiji software, and values of the mean and SEM were calculated. To calculate the ratio of phosphorylated/dephosphorylated Chs2, the faster migration band corresponds to the dephosphorylated Chs2 (as shown in the alkaline phosphatase experiment Fig. S5e). The slower migration band observed in metaphase is considered the phosphorylated Chs2. For each triplicate, the signals were normalized against the higher signal in the blot. To assign the cells to metaphase, anaphase and cytokinesis we took into consideration the percentage of anaphase spindles and the FACS profiles (not shown). Time-points with intermediate percentages of anaphase spindles were not used in the quantification.

Statistical analysis
All experiments were done at least three times. Statistical analyses were performed with Prism5. Student's T-test was used to analyze the p-values of the different assays compared. A p-value < 0.05 was considered statistically significant, ****p < 0.0001, ***p < 0.001, **p < 0.01; *p < 0.05.