Cdc14 phosphatase directs centrosome re-duplication at the meiosis I to meiosis II transition in budding yeast

Background: Gametes are generated through a specialized cell division called meiosis, in which ploidy is reduced by half because two consecutive rounds of chromosome segregation, meiosis I and meiosis II, occur without intervening DNA replication. This contrasts with the mitotic cell cycle where DNA replication and chromosome segregation alternate to maintain the same ploidy. At the end of mitosis, cyclin-dependent kinases (CDKs) are inactivated. This low CDK state in late mitosis/G1 allows for critical preparatory events for DNA replication and centrosome/spindle pole body (SPB) duplication. However, their execution is inhibited until S phase, where further preparatory events are also prevented. This “licensing” ensures that both the chromosomes and the centrosomes/SPBs replicate exactly once per cell cycle, thereby maintaining constant ploidy. Crucially, between meiosis I and meiosis II, centrosomes/SPBs must be re-licensed, but DNA re-replication must be avoided. In budding yeast, the Cdc14 protein phosphatase triggers CDK down regulation to promote exit from mitosis. Cdc14 also regulates the meiosis I to meiosis II transition, though its mode of action has remained unclear. Methods: Fluorescence and electron microscopy was combined with proteomics to probe SPB duplication in cells with inactive or hyperactive Cdc14. Results: We demonstrate that Cdc14 ensures two successive nuclear divisions by re-licensing SPBs at the meiosis I to meiosis II transition. We show that Cdc14 is asymmetrically enriched on a single SPB during anaphase I and provide evidence that this enrichment promotes SPB re-duplication. Cells with impaired Cdc14 activity fail to promote extension of the SPB half-bridge, the initial step in morphogenesis of a new SPB. Conversely, cells with hyper-active Cdc14 duplicate SPBs, but fail to induce their separation. Conclusion: Our findings implicate reversal of key CDK-dependent phosphorylations in the differential licensing of cyclical events at the meiosis I to meiosis II transition.


Introduction
Meiosis is a specialized cell division, which generates gametes. In the canonical mitotic cell cycle, ploidy is maintained by alternating S and M phases. In contrast, during meiosis, chromosome duplication in S phase is followed by two consecutive chromosome segregation phases, meiosis I and meiosis II, to generate gametes with half the ploidy of the parental cell. Therefore, in addition to distinct modifications to the chromosome segregation machinery, meiosis requires a re-wiring of cell cycle controls (reviewed in Duro & Marston, 2015). Progression through the cell cycle is driven by cyclin-dependent kinases (CDKs), in association with distinct cyclin subunits. CDK activity is low in G1, but upon cell cycle entry, activation of S-phase and mitotic CDKs in turn promote DNA replication followed by spindle assembly and chromosome segregation. Following completion of chromosome segregation, CDKs are inactivated, triggering spindle disassembly and the return to G1 (mitotic exit) (Stegmeier & Amon, 2004). This state of low CDK activity in G1 allows for the re-licensing of DNA replication origins and centrosomes/spindle pole bodies (SPBs), events that must be restricted to once per cell cycle. In budding yeast, the Cdc14 phosphatase triggers CDK inactivation through multiple mechanisms to promote exit from mitosis and return to G1 (Jaspersen et al., 1998;Visintin et al., 1998;Zachariae et al., 1998). Cdc14 is regulated by its localization: for the majority of the cell cycle it is sequestered in the nucleolus through association with its inhibitor, Cfi1/Net1 (Shou et al., 1999;Visintin et al., 1999). Upon chromosome segregation at anaphase onset, the Cdc Fourteen Early Anaphase Release (FEAR) network promotes Cdc14 release from the nucleolus into the nucleus; later in anaphase the Mitotic Exit Network (MEN) maintains Cdc14 release throughout the cytoplasm (Pereira et al., 2002a;Stegmeier & Amon, 2004;Yoshida et al., 2002). While FEAR-dependent Cdc14 release promotes successful completion of chromosome segregation, only MEN-dependent Cdc14 release is sufficient to trigger exit from mitosis, leading to spindle disassembly and entry into G1 (Yellman & Roeder, 2015).
The state of low CDK activity in G1 is permissive for the assembly of pre-replicative complexes, the later firing of which requires S phase CDKs (reviewed in (Blow & Dutta, 2005; Drury & Diffley, 2009)). This separation of pre-RC assembly and firing into differential CDK activity states ensures that DNA replication is strictly restricted to once per cell cycle. Similarly, centrosome/SPB duplication must occur exactly once in each cell cycle. Yeast SPBs are microtubule-organising centres, composed of at least 19 proteins, forming three layers that are assembled into a cylindrical organelle embedded in the nuclear envelope (reviewed in Jaspersen & Winey, 2004). SPB duplication is initiated in G1 by extension of the half-bridge (Byers & Goetsch, 1974;Vallen et al., 1994), which protrudes from the central SPB layer and consists of Kar1, Mps3, Sfi1 and Cdc31. The full bridge structure serves as the site for new SPB assembly (Jaspersen et al., 2002;Kilmartin, 2003;Spang et al., 1993;Spang et al., 1995), so that at the close of G1, two sideby-side SPBs are physically connected by a full-bridge (Byers & Goetsch, 1974). S phase CDK activity severs the bridge structure triggering SPB separation. Recent  The meiosis I to meiosis II transition requires specialized cell cycle controls. Uniquely, chromosome segregation at meiosis I exit is followed not by a DNA replication phase, but by a second chromosome segregation phase, meiosis II. Importantly, centrosome/SPBs must be re-licensed at meiosis I exit to permit this additional segregation event, yet DNA replication origins must not be re-set to avoid over-duplication of chromosomes. How this is controlled is not clear. One potential contributing factor is the retention of partial CDK activity between meiosis I and meiosis II. Indeed, there is evidence to suggest that CDKs are only partially downgraded between meiosis I and II in Xenopus oocytes and fission yeast (Iwabuchi et al., 2000;Izawa et al., 2005). However, how these global alterations in CDK activity impinge on the differential re-licensing of DNA replication origins and SPBs has not been investigated.
In budding yeast, the Cdc14 phosphatase plays a prominent role in the meiosis I to meiosis II transition (Buonomo et al., 2003;Marston et al., 2003). Following meiosis I chromosome segregation, cdc14-1 mutants disassemble the spindle, only to reassemble a single spindle that directs segregation of some chromosomes in a meiosis II-like manner (Bizzari & Marston, 2011). The result is binucleate, rather than tetranucleate, cells with a mixed complement of chromosomes (Sharon & Simchen, 1990). Furthermore, ectopic activation of Cdc14 is also detrimental to meiosis. Depletion of the regulatory subunit of protein phosphatase 2A, Cdc55, results in premature release of Cdc14 from the nucleolus in meiosis and a block to spindle assembly, so that nuclear division largely fails (Bizzari

Amendments from Version 1
The revised version incorporates textual changes and corrections suggested by the reviewers. This includes revisions to Figure 3, Figure 4 and Figure 7 to correct the error bars, and also replacing Figure 5 and Figure 6 to be a higher resolution.

REVISED
Since FEAR-dependent Cdc14 release appears insufficient to trigger CDK inactivation (Stegmeier et al., 2002;Yellman & Roeder, 2015), it is likely that the critical role of Cdc14 at the meiosis I to meiosis II transition is to reverse the phosphorylation of key substrates.
Here we investigate the role of Cdc14 in executing the meiosis I to meiosis II transition. Our findings suggest that a critical role of Cdc14 at meiosis I exit is to re-license SPB duplication. This re-licensing ensures assembly of a pair of spindles for a second round of nuclear division at meiosis II. Conversely, premature Cdc14 activation prevents SPB separation. We provide evidence that Cdc14 associates with the SPB in meiosis and that this localization is important for permitting the duplication cycle. Our data suggest that the critical function of Cdc14 at the meiosis I to meiosis II transition is to reverse key phosphorylations to enable SPB re-duplication.

Materials and methods
Yeast strains and plasmids Yeast strains used in this study were generated using standard genetic methods and are given in Table 1. pCLB2-3HA-CDC55

Fluorescence microscopy
To visualize chromosomes and SPBs labelled with fluorescent proteins in fixed cells, 100 μl of meiotic culture was added to eppendorf tubes containing 10 μl of 37% formaldehyde and incubated for 8-10 mins at room temperature. Cells were spun down, washed with 1 ml of 80% ethanol and resuspended in 20 μl of 1 μg/ml DAPI before microscopy. Indirect immunofluorescence of meiotic spindles was carried out as previously described (Bizzari & Marston, 2011).
Imaging of live cells at isolated timepoints was performed on ~1mm deep 2% agarose slides. In total, 100 μl of meiotic culture was spun down and resuspended in 5 μl of sporulation (SPO) media (0.3% Potassium acetate) and 2-3 μl of cell suspension was added to each agarose pad. The slide was covered with a glass coverslip and sealed with a molten mixture of vasoline:lamalin:paraffin (1:1:1) before microscopy.
Live-cell meiotic movies were generated using CellASIC® ONIX Y04D Microfluidics plates (Merck Millipore). All chambers on the plate were washed three times with 500 μl of SPO media before 200 μl of SPO media plus 1mM β-estradiol was added to chambers 1-6. Plates were then pre-incubated at 30°C 30 mins.

Electron microscopy
For sample preparation, 3 ml of culture from meiotic cell cycle time-course was vacuum filtered through a 0.45 μm Millipore filter. The cell paste was rapidly frozen under high pressure in a Wohlwend Compact 02 High Pressure Freezer. Frozen cell pellets were then freeze substituted in acetone containing 2% (w/v) osmium tetroxide and 0.1% (w/v) uranyl acetate at -80°C. Samples were slowly warmed to room temperature over three days. After washing cells twice in acetone, samples were embedded in Epon 812 resin (Hexion) through multiple changes of diluted resin with acetone (1:3, 1:1 and 3:1). Three more changes using undiluted Epon 812 resin were carried out over two days before resin was polymerised at 60-70°C overnight. Epon blocks were serially sectioned at a thickness of 70 nm and stained with 2% (w/v) uranyl acetate in sterile water for 8 mins, and then in Reynolds' lead citrate for 3 mins. Sections were viewed on a Philips CM120 transmission electron microscope, and images were collected with a Gatan Orius CCD camera and processed using ImageJ v1.47.

Immunoprecipitation
Meiotic cells were harvested and washed with sterile water by centrifugation at 4000 rpm for 6 mins. Cells were resuspended in 0.2x cell volume of sterile water before drop-freezing in liquid nitrogen. Lysates were centrifuged at 4000 rpm for 5 mins at 4°C and the supernatant was transferred to a new 50 ml falcon tube. Immunoprecipitation was performed by adding 5 mg of rabbit IgG-coupled Dynabeads or 18 mg of M2 αFLAG-coupled Dynabeads per 30 g lysed yeast, and the lysates were rotated at 4°C for 2 h. Lysates were then washed five times in cold buffer without inhibitors and then transferred to a 1.5 ml eppendorf tube with 1 ml buffer. Residual buffer was removed, 25 μl of 1x NuPAGE® LDS sample buffer was added, samples were boiled at 100°C for 5 mins before 5 μl of β-mercaptoethanol was added and samples were boiled for a further 5 mins, spun down at 13000 rpm for 5 mins and loaded onto a precast NuPAGE® 8-12% Bis-Tris gel (Novex). Bands were visualized after staining using the Pierce silver staining kit (Thermo Scientific).

Mass spectrometry
Protein bands were excised from Coomassie-stained NuPAGE® 8-12% Bis-Tris gels and washed alternatingly with 50 mM ammonium bicarbonate and acetonitrile solutions until Coomassie staining was removed. Gel pieces were treated with 10 mM DTT in 50 mM ammonium bicarbonate for 30 mins at 37°C, then DTT was removed and samples were washed with acetonitrile. A total of 55 mM iodoacetamide in 50 mM ammonium bicarbonate was added to the gel slices, and these were incubated at room temperature in the dark for 20 mins. After washing again with 50 mM ammounium bicarbonate and acetonitrile, gel pieces were incubated with trypsin for 15 mins on ice, and then samples were transferred to 37°C for overnight digestion. The following morning, digestion reactions were treated with 0.1% (w/v) trifluoroacetic acid and left for 15 mins to allow peptides to diffuse from the gel. Samples were then passed through an equilibrated StageTip consisting of two layers of Empore Disks C18 within a pipette tip ( To establish the importance of Cdc14 in spindle disassembly following meiosis I, we determined the time from anaphase I onset until spindle breakdown in live cells with impaired Cdc14 function. Securin (Pds1-tdTomato) degradation was used a marker for anaphase I onset and the time taken for the meiosis I spindle (GFP-Tubulin) to completely disassemble after Pds1 proteolysis was measured in individual cells. In the wild type example ( Figure 1A), spindle disassembly was observed 40 min after anaphase I onset, after which meiosis II spindles formed (note that Pds1-tdTomato is not visualised in meiosis II cells, presumably due to slow maturation of the fluorophore (Matos et al., 2008)). Spindle disassembly occurred 45.7 min after anaphase I, on average ( Figure 1B) and was observed in 100% of wild type cells ( Figure 1C). In ~82% cdc14-1 mutant cells, anaphase I spindles broke down and a new spindle did not assemble ( Figures  , spindle reassembly at the presumptive time of meiosis II was observed in a fraction of slk19Δ and spo12Δ cells, though the extent to which this occurred varied between the different mutants for reasons that are unclear. Taken together, these findings indicate that, while FEAR and Cdc14 appear to work together to ensure that two spindles are produced during meiosis II, Cdc14 may promote timely meiosis I spindle disassembly through a FEAR-independent mechanism. Nevertheless, spindle disassembly invariably occurs in cdc14-1 mutants, raising the possibility that, in contrast to the critical requirement for Cdc14 for exit from mitosis, Cdc14 is not absolutely required for CDK down-regulation at meiosis I exit. Consistently, degradation of the major meiosis I cyclin Clb1 is not obviously delayed in cdc14-1 mutants (Bizzari & Marston, 2011). Similarly, a recent study observed timely meiosis II spindle disassembly following inactivation of Cdc14 using the distinct cdc14-3 temperaturesensitive allele (Argüello-Miranda et al., 2017). Although we cannot completely rule out retention of partial activity by the temperature sensitive Cdc14 proteins, together these findings suggest that Cdc14 is more critical for spindle disassembly in mitosis than meiosis.
Cdc14 associates with the SPB at the meiosis I to meiosis II transition We took an unbiased approach to identify cellular processes targeted by Cdc14 to regulate the meiosis I to meiosis II transition. Following its release from the nucleolus during anaphase I, Cdc14 is expected to associate with, and dephosphorylate, substrates that facilitate the transition to meiosis II. We reasoned that identification of Cdc14 interacting partners in both wild type cells and pCLB2-CDC55 cells, in which Cdc14 is ectopically released from the nucleolus, would inform on the processes it regulates. Anti-FLAG immunoprecipitates from wild type and pCLB2-3HA-CDC55 cells harvested 4 h after induction of meiosis and carrying CDC14-3FLAG were analysed by mass spectrometry. Despite similar Cdc14 peptide counts in wild type and pCLB2-3HA-CDC55 samples, we observed a lower Cfi1/Net1 peptide count in the latter sample, consistent with premature release of Cdc14 from the nucleolus in Cdc55deficient cells ( Figure 2A). Interestingly, however, the predominant class of proteins identified in both samples were components of the yeast centrosome/spindle pole body (SPB) (Figures 2A and B).
To determine the timing of Cdc14 association with the SPB during meiosis, we imaged live cells carrying CDC14-GFP and the SPB marker, SPC42-tdTomato undergoing meiosis. As previously reported, in wild type cells, Cdc14 is sequestered in the nucleolus throughout prophase I and metaphase I of meiosis and, accordingly, we did not observe co-localization with SPBs at these stages  Figure 3A). During anaphase I, however, concomitant with its release from the nucleolus, Cdc14-GFP was detected at the SPB ( Figure 3A, arrows). To confirm the timing of Cdc14 association with the SPB we determined the ratio of intensity of Cdc14-GFP and Spc42-tdTomato fluorescence ( Figure 3B). This revealed the strongest association of Cdc14-GFP with the SPB in anaphase I, with a weaker association in anaphase II ( Figure 3B). Interestingly, Cdc14-GFP localized asymmetrically, generally associating with just one of the two SPBs in anaphase I, or two of the four SPBs during anaphase II, with no detectable SPB association during metaphase I or metaphase II ( Figure 3C).
We, and others, previously showed that ectopic release of Cdc14 prevents nuclear division in pCLB2-3HA-CDC55 cells (Bizzari & Marston, 2011; Kerr et al., 2011). To determine whether premature association of Cdc14 with the SPB could underlie this phenotype, we induced pCLB2-3HA-CDC55 cells carrying CDC14-GFP and SPC42-tdTomato to undergo meiosis and categorised cells based on the localization of Cdc14-GFP: nucleolar sequestration (class 1); partial release (class 2) or complete release (class 3) (Figures 3D and E). Cdc14-GFP was detected at the SPB in virtually all pCLB2-3HA-CDC55 cells where Cdc14-GFP was either completely or partially released from the nucleolus ( Figure 3E). Note that the vast majority of pCLB2-3HA-CDC55 cells contain only a single Spc42-tdTomato foci, therefore it was not possible to address whether Cdc14-GFP remains asymmetric in these cells.  Figures 5A and B). As expected, wild type cells produced two, then four Spc42-tdTomato foci concomitant with the appearance of binucleate and tetranucleate cells ( Figure 5A). In contrast, and consistent with the observed single nuclear division, cdc14-1 cells produced a maximum of two Spc42-tdTomato foci ( Figure 5B), as did cells lacking the two FEAR activators, Spo12 and Slk19 ( Figures 5C and D). Upon depletion of Cdc55, a single Spc42-tdTomato focus was observed in the majority of cells ( Figure 5E), indicating a failure in the first round of SPB duplication or separation. Furthermore, this lack of SPB duplication/separation prior to meiosis I in Cdc55depleted cells was a consequence of ectopic Cdc14 activation, since cdc14-1 pCLB2-3HA-CDC55 cells produced two Spc42-tdTomato foci, similar to cdc14-1 single mutant cells ( Figure 5F). Consistent with a requirement for Cdc14 at SPBs, a large fraction of bub2Δ and bfa1Δ cells completed only a single meiotic division with only 2 Spc42-tdTomato foci ( Figures 6A-C). During mitosis, Bub2/Bfa1 is asymmetrically localized on SPBs, but this asymmetry is broken in response to defective spindle positioning, in a manner dependent on Kin4 and Bmh1 (Gryaznova We found, however that Kin4 and Bmh1 are dispensable for either the asymmetric SPB localization of Cdc14 ( Figure 4D) or the execution of two meiotic divisions ( Figures 6E-F), at least under normal conditions. These findings indicate that localization of Cdc14 at the SPB is important for the successful execution of the meiosis I to meiosis II transition.
Cdc14 is essential for SPB duplication at the meiosis I to meiosis II transition To determine whether cdc14-1 and pCLB2-3HA-CDC55 mutants are defective in SPB duplication or separation, we initially used   quantitative fluorescence microscopy. We measured the total intensity of the central SPB component Spc42-tdTomato within each cell ( Figure 7A). First, we examined cells progressing from meiotic entry (induced by resuspension in sporulation medium) into a prophase I block (by preventing NDT80 expression) (Carlile & Amon, 2008). At meiotic entry, the majority of cells are in G1 phase of the cell cycle and expected to contain 1 SPB with a full bridge that is already licensed/competent for assembly of a new SPB alongside it. In meiosis, separation of these SPBs occurs only at prophase I exit. Consistent with this idea, we observed an approximately 1.5 fold increase in Spc42-tdTomato intensity as wild type cells progressed from G1 into the prophase I arrest ( Figure 7B). Upon Cdc14 inactivation (cdc14-1) or Cdc55 depletion (pCLB2-3HA-CDC55) at meiotic entry we observed a similar increase in Spc42-tdTomato intensity in prophase I, suggesting that Cdc14 and Cdc55 are not required for the assembly of a new SPB, once licensing has occurred ( Figure 7B). Next we examined the intensity of Spc42-tdTomato as cells progressed from the prophase I block into the meiotic divisions ( Figures 7C-E).
In wild type cells, overall Spc42-tdTomato intensity was slightly increased in the cells with 2 Spc42-tdTomato foci, compared to the cells with 1 Spc42-tdTomato focus suggesting that SPB reduplication had occurred between meiosis I and meiosis II, though interestingly the increase was much less than twofold (~x1.18) ( Figure 7C). Spc42-tdTomato intensity was greatly increased in cells with 4 Spc42-tdTomato foci, perhaps in preparation for SPB maturation that occurs in meiosis II and which is important for spore formation (Neiman, 2011). In cdc14-1 cells, the overall  intensity of Spc42-tdTomato foci did not increase, even though cells produced 2 Spc42-tdTomato foci (x1.07) ( Figure 7D), suggesting that SPB re-duplication failed to occur prior to SPB separation. The single Spc42-tdTomato focus of pCLB2-3HA-CDC55 cells also did not increase in Spc42-tdTomato intensity throughout the timecourse (x0.98), suggesting that both SPB separation and re-duplication fail to occur upon Cdc55 depletion.
To gain further insight into how Cdc14 influences SPB morphogenesis during meiosis, we compared the composition of the SPB in wild type and cdc14-1 cells by quantitative mass spectrometry. SPBs were purified from wild type and cdc14-1 cells undergoing meiosis and carrying SPC42-3FLAG by immunoprecipitation using anti-FLAG antibodies ( Figure 7F). Comparison of relative peptide intensities for three biological replicates ( Figure 7G)  These results are consistent with the idea that a major function of Cdc14 at the meiosis I to meiosis II transition is to enable half bridge extension, thereby allowing SPB re-duplication.
To examine SPB morphogenesis more directly, we analyzed cdc14-1 and pCLB2-3HA-CDC55 meiotic cells by electron microscopy. As predicted by the quantitative fluorescence microscopy, SPB re-duplication was not observed in cdc14-1 mutants (n=8) and cells arrested with two unduplicated SPBs. We observed late meiosis II events in 3/8 cells. In the example shown ( Figure 8A) a long spindle connects two unduplicated SPBs. Assembly of the outer plaque and vesicles are apparent at one of the SPBs (SPB 1, white arrow), indicating that the cell is in a late stage of meiosis II, though outer plaque formation has not been initiated at the other SPB (SPB 2, white arrow). In pCLB2-3HA-CDC55 cells, also as predicted from our quantitative fluorescence microscopy, two side-by-side SPBs connected by a half bridge were invariably observed (n=6; Figure 8B). However, we found no evidence of over-duplication of SPBs in pCLB2-3HA-CDC55 cells. This suggests that Cdc14 must be held inactive during early meiosis to allow SPB separation and acts in a licensing step, rather than as an assembly factor.

Discussion
The existence of two consecutive rounds of chromosome segregation without an intervening S phase is a characteristic feature of meiosis that underlies sexual reproduction. Unique, yet poorly understood, controls allow a second round of spindle formation, but prevent a second round of DNA  (Figure 9). Overall, our findings show that Cdc14 is required to re-license SPB duplication between meiosis I and meiosis II and that its retention in the nucleolus during early meiosis is required to allow SPB separation during meiosis I.

SPB localization of Cdc14
Bub2/Bfa1-dependent association of Cdc14 with the SPB is not unique to meiosis, indeed it has been observed in early mitosis (  activity is presumably required at both SPBs during anaphase I to trigger their duplication, thereby ensuring production of a pair of spindles in meiosis II. While Cdc14 is detectable by microscopy only on 1 SPB during anaphase I, we speculate that undetectable levels of Cdc14 on the other SPB are sufficient to trigger SPB duplication. This however, raises the question of how and why Cdc14 is more concentrated on a single SPB, particularly considering that Bub2/Bfa1 is itself symmetrically localized. The origin and significance of the asymmetric localization of Cdc14 at the SPB during anaphase I therefore remain unexplained.

Control of the cell cycle at meiosis I exit
Although Cdc14 is essential for mitotic exit, accumulating evidence suggests that Cdc14 plays a lesser role in CDK downregulation at meiosis I exit. We found that spindle assembly is only slightly delayed in cells with impaired Cdc14 activity and cyclin destruction appears to occur on schedule in Instead it is likely that cyclin degradation upon APC Cdc20 activation at anaphase I onset initiates meiosis I exit. Understanding how this is regulated to ensure step-by-step release of cohesion, spindle elongation and spindle disassembly at meiosis I is an important priority for the future.

Data availability
Source data for mass spectrometry results from Figures  Author contributions CF and AM conceived the study and designed the experiments. CF carried out the research. JZ and JR analysed mass spectrometry data. CF and AM prepared the manuscript. All authors have approved the final content.

Competing interests
No competing interests were disclosed. The centrosome is a fascinating organelle; it serves as a scaffolding platform for numerous cell cycle regulators in addition to the major microtubule organising centre, which plays an essential role in chromosome segregation by acting as the spindle pole. During the mitotic cell cycle, a rigorous licensing mechanism that involves phosphorylation and dephosphorylation of the centrosome components restricts the number of centrosome duplications to exactly one per cell cycle. Physiological importance of this regulation is underlined by the fact that tumour cells are often associated with abnormally amplified centrosomes. The timing of the centrosome duplication in many organisms coincides with S phase, where DNA replication takes place. Common cell cycle regulators, such as CDKs, trigger both centrosome duplication and DNA replication events.

Grant information
In meiosis, however, centrosome duplication must be uncoupled from DNA replication, because, no DNA replication must occur at the transition from meiosis I (MI) to meiosis II (MII). Although the mechanism of uncoupling of the centrosomal and DNA replication cycles in meiosis is mostly unclear, studies in oocytes and the fission yeast found significant residual CDK activities at MI -MII transition, Xenopus suggesting that these CDK activities may inhibit DNA replication, hence, decoupling the centrosome cycle from the DNA replication cycle. However, further supporting evidence for it or the molecular basis that directs the meiotic centrosome cycle has still been missing.
In this manuscript, Fox used the budding yeast as a model and showed that Cdc14 phosphatase is et al one of the essential regulators of the meiotic centrosome duplication. By exploiting a conditional cdc14-1 mutant, an absolute requirement of Cdc14 function in meiotic centrosome duplication was clearly demonstrated through extensive live and fixed cell imaging of the cells undergoing MI-MII synchronously. Observation using electron microscopy provided further supporting evidence.
The manuscript is very well written, the quality of the data (which are mostly very clearly presented) is high and sound data interpretations have been provided. Therefore, I would like to approve this study. A few minor points are listed below to help further clarify the authors' message.
Culturing condition to inactivate Cdc14: It was not too clear which temperature (and how long) was used to inactivate Cdc14 in the cdc14-1 ts mutant. Or, does mutant show its meiotic phenotype at an intermediate temperature 30 cdc14-1 C? It would be really helpful if the authors would clarify this point. In these figures, the Cdc14-GFP signal at the SPB is presented. If I understood correctly, in the pCLB2-3HA-CDC55 setting, SPB duplication does not occur and only one SPB can be found in each cell. If this would be the case, it would be helpful to clearly state so, thus, "asymmetric Cdc14 localisation" is not relevant in this experimental setting.
In the "Metaphase II" panel, it would be helpful to have additional small arrowheads in order to indicate SPBs (Spc42-CFP signal-positive dots) without Bfa1-tdTomato signal, so that asymmetric Bfa1-tdTomato localization is easily recognisable. Bfa1-tdTomato localization is easily recognisable.
Page 12, 2 paragraph: It would be helpful to dub the SPB as "Spc42-tdTomato foci" in this paragraph. For example, instead of saying "two SPB" or "four SPB", say "two Spc42-tdTomato foci" or "Four Spc42-tdTomato foci". This is because until we see Fig. 7 and 8, we do not know whether a "Spc42-tdTomato focus" represents a non-duplicated SPB (I would call this one SPB) or non-separated SPBs (I would call this two SPBs).  We used 30C, which we believe is efficient at largely inactivating Cdc14 (see comments to other reviewers) as vegetative cells fail to grow at this temperature. We have added this information to the methods. Fig. 3D and 3E: In these figures, the Cdc14-GFP signal at the SPB is presented. If I understood correctly, in the pCLB2-3HA-CDC55 setting, SPB duplication does not occur and only one SPB can be found in each cell. If this would be the case, it would be helpful to clearly state so, thus, "asymmetric Cdc14 localisation" is not relevant in this experimental setting.
We added a statement to this effect in this results section Page 8 final paragraph: "…but during metaphase II" -> asymmetrically symmetrically This has been corrected. In the "Metaphase II" panel, it would be helpful to have additional small arrowheads in order to indicate SPBs (Spc42-CFP signal-positive dots) without Bfa1-tdTomato signal, so that asymmetric Bfa1-tdTomato localization is easily recognisable.
Page 12, 2 paragraph: It would be helpful to dub the SPB as "Spc42-tdTomato foci" in this paragraph. For example, instead of saying "two SPB" or "four SPB", say "two Spc42-tdTomato foci" or "Four Spc42-tdTomato foci". This is because until we see Fig. 7 and 8, we do not know whether a "Spc42-tdTomato focus" represents a non-duplicated SPB (I would call this one SPB) or non-separated SPBs (I would call this two SPBs). This is a very good point. We have changed the text and figure accordingly. In addition, would it be possible to indicate the size/band that corresponds to the Cdc14 protein in the gel in panel F? There is no obvious band corresponding to Cdc14, a 61 kd protein and as we are unable to confirm that any particular band is Cdc14 we prefer not to annotate this on the gel. This is an interesting study addressing an important question, i.e. how yeast cells license SPB duplication at the MI/MII transition to be able to form bipolar spindles and support chromosome segregation during reductional division.

None
Previous work had shown that the Cdc14 phosphatase must be properly controlled to permit assembly of two MII spindles through an unknown mechanism. In addition, the SPB half-bridge protein Sfi1 has been recently identified as a critical target of Cdc14 in mitosis in order to license SPB duplication.
In this manuscript the authors show that Cdc14 associates with SPBs also in meiosis. Cdc14 localization at SPBs is asymmetric and requires the Bub2/Bfa1 complex. Cdc14, while not being strictly required for exit from MI (i.e. spindle disassembly), is required for SPB duplication and proper nuclear division in MII. Conversely, premature Cdc14 activation in cells lacking Cdc55 prevents SPB separation in MI. Finally, although SPBs do not duplicate upon Cdc14 inactivation, composition of the SPB remains mostly unchanged.
Overall, experiments are carefully executed and conclusions are mostly supported by the experimental data. I have, however, a couple of main issues and a few minor points that the authors might want to address.

Main points:
The incubation of mutant cells at 30°C, which might be a semi-permissive temperature, cdc14-1 raises the question as to what extent Cdc14 is actually inactivated in these cells. Wouldn't it have been better to use a tighter system to inactivate Cdc14 (e.g. expressing from the CDC14 CLB2 promoter) ? It is even possible that the Cdc14-1 protein is prematurely released, thereby tempering the effects of temperature inactivation. Has nucleolar release of the Cdc14-1 protein ever been checked at 25°C/30°C ? This would also be relevant to interpret correctly the data in Fig. 1B and C, where cells do cdc14-1 not behave exactly like FEAR mutants.
Quantification of SPB signals during meiotic progression (Fig. 7A-E) is an essential piece of data to support the conclusion that cells are defective in SPB duplication upon entry into MII. I cdc14-1 wonder if a slow-folding fluorescent protein, such as Spc42-tdTomato, is actually a suitable marker for this kind of analyses.
A related issue concerns the way these data are presented, which I find a bit confusing. For instance, I do not understand why there are gaps in the plot of wild type cells at 0.75-1h and 1.5-1.75h (Fig. 7C), since fluorescence intensities are measured on movies.
Direct inspection of SPBs by EM, like the authors did, should greatly help understanding what is going on in the mutant. However, knowing how many MII cells display the phenotype cdc14-1 shown in Fig. 8A is an essential piece of information that is currently missing. Another interesting information is whether SPB over-duplication was ever found in cells. pCLB2-CDC55 3. 1.

5.
6. information is whether SPB over-duplication was ever found in cells.

pCLB2-CDC55
Finally, if it were possible to inactivate Cdc14 before commitment to meiosis, would the authors expect a defect in SPB duplication also in MI? The likely possibility that Cdc14 is a crucial licensing factor in ALL kinds of divisions should perhaps be mentioned in the discussion.
Since the role of Cdc14 in licensing SPB duplication in mitosis has been linked to Sfi1 dephosphorylation, it would have been nice to check the phosphorylation state of Sfi1 in wild type and cells during meiosis. This kind of information might even be extracted from the mass cdc14-1 spec data from Fig. 7F.
Along the same line, it would have been interesting to check if the published unphosphorylatable mutant can suppress the SPB duplication defect of cells. This would have offered sfi1 cdc14-1 strong experimental support to the model presented in Fig. 9.

Minor points
It is unclear why in the movies of cells (Fig. 1) spindles do not reassemble in the second cdc14-1 meiotic division from the two unduplicated SPBs (like for example in Fig. 8A).
The abnormal spindle behaviour reported in Fig. 1C is not commented in the text. Surprisingly, have a milder phenotype than FEAR mutants in this respect, although has a more cdc14-1 cdc14-1 pronounced defect in spindle disassembly.
It is not clear why Spo21 is in the list of Tap-Cdc14 interacting proteins with 0 peptides identified.
Cdc14 is reported to localize at SPBs in most cells where Cdc14 is partially or pCLB2-CDC55 completely released from nucleolus. Yet, in the example cell #2 of Fig. 3D Cdc14 is not visible on the SPB. A better representative image could be selected. Fig. 5 and Fig. 6 look pixeled.

Graphs in
In Fig. 7C there is a label within the graph that belongs to the IPs underneath. Also, the graphs (C, D, E) are mislabeled relative to the legend.
In Fig. 7F two prominent bands appear in the Spc42-Flag IP from wt cells, while one of the two is much decreased in cells. Which one of the two is Spc42? cdc14-1 No competing interests were disclosed.

Competing Interests:
I have read this submission. I believe that I have an appropriate level of expertise to confirm that it is of an acceptable scientific standard, however I have significant reservations, as outlined above.
Author Response 11 Feb 2017 , University of Edinburgh, UK

Adele Marston
Thank you for your approval in principle of our article and you helpful suggestions, here we have 1.

1.
Thank you for your approval in principle of our article and you helpful suggestions, here we have addressed your reservations: Main points: The incubation of mutant cells at 30°C, which might be a semi-permissive cdc14-1 temperature, raises the question as to what extent Cdc14 is actually inactivated in these cells. Wouldn't it have been better to use a tighter system to inactivate Cdc14 (e.g. expressing from the promoter) ? It is even possible that the Cdc14-1 protein is CDC14 CLB2 prematurely released, thereby tempering the effects of temperature inactivation. Has nucleolar release of the Cdc14-1 protein ever been checked at 25°C/30°C ? This would also be relevant to interpret correctly the data in Fig. 1B and C, where cdc14-1 cells do not behave exactly like FEAR mutants.
Please see response to comment by Soni Lacefield. We agree we cannot completely rule out partial inactivation of Cdc14.
Quantification of SPB signals during meiotic progression (Fig. 7A-E) is an essential piece of data to support the conclusion that cells are defective in SPB duplication upon cdc14-1 entry into MII. I wonder if a slow-folding fluorescent protein, such as Spc42-tdTomato, is actually a suitable marker for this kind of analyses.
Because of these and other concerns (see response to Soni Lacefield) we used EM as a completely independent way to examine SPB duplication in cdc14-1 mutants. We believe the EM data provides convincing evidence for our conclusions, while the fluorescence intensity measurements provide some supporting data.
A related issue concerns the way these data are presented, which I find a bit confusing. For instance, I do not understand why there are gaps in the plot of wild type cells at 0.75-1h and 1.5-1.75h (Fig. 7C), since fluorescence intensities are measured on movies.
The data for individual cells was lined up so that the 1h timepoint is defined as the point at which 2 SPBs are first observed. The different coloured lines represent 1, 2 or 4 SPBs. We have clarified this point in the figure legend.
Direct inspection of SPBs by EM, like the authors did, should greatly help understanding what is going on in the mutant. However, knowing how many MII cells display the phenotype cdc14-1 shown in Fig. 8A is an essential piece of information that is currently missing. Another interesting information is whether SPB over-duplication was ever found in cells.

pCLB2-CDC55
Overall we observed SPBs in 8 cdc14-1 cells by EM. Out of these, evidence of late meiosis II events (i.e. spore formation/vesicles) was observed in 3 of these. The image shown in Figure 8A is one example of two in which a spindle could be followed between 2SPBs. The other example was spread over >8 sections (spore formation undetermined) so the image shown was favoured. The third example showed the spore wall forming around a nucleus so SPBs were not connected by spindles in that case. No evidence of overduplication of SPBs was observed in pCLB2-CDC55 cells (n=6).

1.
We have included this information in the relevant section in the paper. All EM images are available https://osf.io/g5cmh/ ( ). here: Marston, 2016 Finally, if it were possible to inactivate Cdc14 before commitment to meiosis, would the authors expect a defect in SPB duplication also in MI? The likely possibility that Cdc14 is a crucial licensing factor in ALL kinds of divisions should perhaps be mentioned in the discussion.
The licensing event that enables SPB duplication during meiosis I would be expected to occur in the last stages of vegetative growth i.e. at mitotic exit prior to commitment to meiosis. So yes, it would be required in the same way that Cdc14 is required to license SPBs during vegetative growth.
2.. Since the role of Cdc14 in licensing SPB duplication in mitosis has been linked to Sfi1 dephosphorylation, it would have been nice to check the phosphorylation state of Sfi1 in wild type and cells during meiosis. This kind of information might even be extracted from the mass cdc14-1 spec data from Fig. 7F.
Unfortunately we did not detect Sfi1 phosphorylation in our quantitative mass spec analysis. We attempted to examine Sfi1 phosphorylation on western blots and phostag gels but since the time period where Sfi1 dephosphorylation is expected to be observed (anaphase I) is very short, we were unable to obtain sufficient time resolution to make firm conclusions about the effect of Cdc14 inactivation of Sfi1 phosphorylation status.
3. Along the same line, it would have been interesting to check if the published unphosphorylatable mutant can suppress the SPB duplication defect of cells. This sfi1 cdc14-1 would have offered strong experimental support to the model presented in Fig. 9.
We agree that this would be a very interesting experiment but despite extensive efforts we were unable to generate yeast strains carrying the Sfi1 phosphonull mutants and were therefore sadly unable to do this experiment. Diploid cells carrying the mutations as the sole copy of Sfi1 were extremely sick and would not enter meiosis. Other approaches were also attempted, for example expression of Sfi1 only in meiosis but due to both biological and technical reasons we have so far been unable to address this important point.

Minor points
It is unclear why in the movies of cells (Fig. 1) spindles do not reassemble in the cdc14-1 second meiotic division from the two unduplicated SPBs (like for example in Fig. 8A). For reasons we do not understand, spindles reassemble in a fraction of cdc14-1 cells, depending on the growth condition. In live cell imaging (Fig. 1), about 10% of cells were observed to reassemble spindles, however, the fraction is likely to be higher in flasks (Fig. 8A).
The abnormal spindle behaviour reported in Fig. 1C is not commented in the text. Surprisingly, have a milder phenotype than FEAR mutants in this respect, although has a more cdc14-1 cdc14-1 pronounced defect in spindle disassembly.
We have included a sentence raising this point in the results section.
2. It is not clear why Spo21 is in the list of Tap-Cdc14 interacting proteins with 0 peptides identified.

It is not clear why Spo21 is in the list of Tap-Cdc14 interacting proteins with 0 peptides identified.
For completeness, all SPB proteins were included in the list, regardless of whether they were identified in the mass spectrometer.
3. Cdc14 is reported to localize at SPBs in most cells where Cdc14 is partially or pCLB2-CDC55 completely released from nucleolus. Yet, in the example cell #2 of Fig. 3D Cdc14 is not visible on the SPB. A better representative image could be selected.
Cell #3 shows a cell where Cdc14 is on the SPB. Cell #2 is an example of where the diffuse released Cdc14 signal overlaps with the SPB. As these are the criteria that were scored, these examples are appropriate. Fig. 5 and Fig. 6 look pixeled.

Graphs in
We have uploaded new versions of these figures that should correct this problem. Fig. 7C there is a label within the graph that belongs to the IPs underneath. Also, the graphs (C, D, E) are mislabeled relative to the legend.

In
We have corrected this. In their manuscript, Fox demonstrate that Cdc14, a key phosphatase that promotes CDK inactivation et al. and exit from mitosis in , also plays a pivotal role in promoting the re-licensing Saccharomyces cerevisiae of the SPBs at the meiosis I to meiosis II transition, thus ensuring the re-duplication of these structures between these two different meiotic phases. In order to fulfill this function, Cdc14 is loaded on a single SPB during anaphase I in a process that depends on the Bfa1/Bub2 complex. Once loaded on the SPB, Cdc14 likely promotes dephosphorylation of key substrates on this structure, thereby allowing the re-duplication of the SPBs in meiosis II. This newly suggested meiotic function of Cdc14 helps to shed light on how cells establish the particular pattern of chromosome segregation during meiosis, an essential cell division process by which gametes are generated. Overall, the experiments detailed in the manuscript are carefully designed, nicely presented and well executed. Also, and importantly, the final conclusions of the manuscript are properly sustained by the provided experimental data. Therefore, I support an 1.

5.
6. the manuscript are properly sustained by the provided experimental data. Therefore, I support an approved status for this article. In any case, I also indicate some minor corrections that would need to be introduced, as well as some suggestions that might help the authors to improve the manuscript and strengthen some of the conclusions: In page 8, when the authors indicate that ∆ and ∆ cells disassembled anaphase I slk19 spo12 spindles with a timing comparable to wild type cells, they only refer to Figure 1B, but Figure 1A should be also mentioned. Additionally, and regarding the data in Figure 1A, I would suggest to show images for other live cell experiments that are more representative of the previous statement. Although Figure 1B indeed shows that, as an average, time from anaphase I onset until spindle disassembly is similar for wild type, ∆ and ∆ cells, in the images shown in Figure 1A for slk19 spo12 the ∆ and ∆ cells the spindle disassembles, respectively, 20 min and 30 min after slk19 spo12 anaphase I onset, which is significantly earlier than for the wild type (40 min).
Also in page 8, it is later stated that "[…] the two component GAP Bfa1/Bfa1, […] localizes symmetrically at SPBs during metaphase I, anaphase I but symmetrically during metaphase II". This sentence should be corrected, since, as shown in Figure 1B, Bfa1 localization is asymmetric during metaphase II.
The localization of Cdc14-GFP is difficult to assess in Figure 3A, due to the nucleolar background and the appearance of other GFP foci that do not co-localize with the SPBs and whose nature it is not indicated ( , 1:15 time point in Figure 3A). Since the authors show different proteins that e.g. interact with Cdc14 on the SPBs, they might consider to use a different approach if they plan to further analyze this localization in the future. As such, a bimolecular fluorescence complementation assay (BiFC) could be helpful to determine the exact timing and pattern of localization of Cdc14 on the SPBs. The BiFC assay could facilitate tracking Cdc14 localization exclusively to the SPBs by means of the reconstitution of a fluorescent signal when the phosphatase and a SPB component, both tagged with different fragments of a fluorescent protein, interact.
In page 16, when the authors show a 1.5-fold increase in Spc42-tdTomato intensity as wild type cells progressed from G1 into prophase I, they only refer to Figure 7A, but it is in Figure 7B where the quantification is shown.
A shocking observation is that, despite Bfa1/Bub2 being symmetrically localized in anaphase I and Cdc14 activity being presumably required at both SPBs to trigger their duplication in meiosis II, the phosphatase is asymmetrically localized to only one SPB during anaphase I. This is an obvious caveat and, although I do appreciate that it would require further extensive analysis for this question to be solved, the authors might have speculated a little more extensively about the reasons for this asymmetry in the discussion of the manuscript.
A prediction of the model proposed by Fox . is that re-duplication of the SPBs should be et al restored in a mutant background by conditionally forcing Cdc14 loading on the SPBs at cdc14-1 the appropriate time ( , by temporarily expressing a fusion between Cdc14 and a SPB e.g. component). This is a relatively straightforward experiment that the authors might consider to carry out at some point to further strengthen their model. Furthermore, and coming back to the differential localization of Cdc14 on the SPBs during anaphase I, this experiment could be informative to address the importance of this asymmetry in future studies. Cdc14 could be fused either to Bfa1 (asymmetrically localized) or an integral SPB component (symmetrically localized), and then analyze the consequences of expressing both Cdc14 fusions during meiosis. We are grateful for your approval of our manuscript, kind comments and helpful suggestions, which we respond to here.
1. In page 8, when the authors indicate that ∆ and ∆ cells disassembled anaphase I slk19 spo12 spindles with a timing comparable to wild type cells, they only refer to Figure 1B, but Figure 1A should be also mentioned. Additionally, and regarding the data in Figure 1A, I would suggest to show images for other live cell experiments that are more representative of the previous statement. Although Figure 1B indeed shows that, as an average, time from anaphase I onset until spindle disassembly is similar for wild type, ∆ and ∆ cells, in the images shown in Figure 1A for slk19 spo12 the ∆ and ∆ cells the spindle disassembles, respectively, 20 min and 30 min after slk19 spo12 anaphase I onset, which is significantly earlier than for the wild type (40 min).
We have now included the reference to Figure 1A. Under the conditions we observed them, the ( Figure 1B) timing of spindle disassembly varied across all strains , we tried to show representative images to reflect this.
Also in page 8, it is later stated that "[…] the two component GAP Bfa1/Bfa1, […] localizes 2. symmetrically at SPBs during metaphase I, anaphase I but symmetrically during metaphase II". This sentence should be corrected, since, as shown in Figure 1B, Bfa1 localization is asymmetric during metaphase II.
The typo has been corrected.
3. The localization of Cdc14-GFP is difficult to assess in Figure 3A, due to the nucleolar background and the appearance of other GFP foci that do not co-localize with the SPBs and whose nature it is not indicated ( , 1:15 time point in Figure 3A). Since the authors show e.g. different proteins that interact with Cdc14 on the SPBs, they might consider to use a different approach if they plan to further analyze this localization in the future. As such, a bimolecular fluorescence complementation assay (BiFC) could be helpful to determine the exact timing and pattern of localization of Cdc14 on the SPBs. The BiFC assay could facilitate tracking Cdc14 localization exclusively to the SPBs by means of the reconstitution of a fluorescent signal when the phosphatase and a SPB component, both tagged with different fragments of a fluorescent protein, interact.
It is presently unclear what the non-nucleolar, non-SPB Cdc14-GFP foci represent. Although 1 It is presently unclear what the non-nucleolar, non-SPB Cdc14-GFP foci represent. Although outside the scope of the current manuscript, we appreciate the excellent suggestion to use a bimolecular complementation assay to probe the co-localization of Cdc14 with SPBs. It would be particularly interesting to determine whether Cdc14 can be detected at both SPBs using this system.
In page 16, when the authors show a 1.5-fold increase in Spc42-tdTomato intensity as wild type 4. cells progressed from G1 into prophase I, they only refer to Figure 7A, but it is in Figure 7B where the quantification is shown.
This has been corrected.

5.
A shocking observation is that, despite Bfa1/Bub2 being symmetrically localized in anaphase I and Cdc14 activity being presumably required at both SPBs to trigger their duplication in meiosis II, the phosphatase is asymmetrically localized to only one SPB during anaphase I. This is an obvious caveat and, although I do appreciate that it would require further extensive analysis for this question to be solved, the authors might have speculated a little more extensively about the reasons for this asymmetry in the discussion of the manuscript.
We agree, this is indeed a surprising finding that remains unexplained. We expanded our discussion as suggested, though ultimately this is indeed a very puzzling observation for which we currently do not have an explanation.

A prediction of the model proposed by Fox
. is that re-duplication of the SPBs should be et al restored in a mutant background by conditionally forcing Cdc14 loading on the SPBs at cdc14-1 the appropriate time ( , by temporarily expressing a fusion between Cdc14 and a SPB e.g. component). This is a relatively straightforward experiment that the authors might consider to carry out at some point to further strengthen their model. Furthermore, and coming back to the differential localization of Cdc14 on the SPBs during anaphase I, this experiment could be informative to address the importance of this asymmetry in future studies. Cdc14 could be fused either to Bfa1 (asymmetrically localized) or an integral SPB component (symmetrically localized), and then analyze the consequences of expressing both Cdc14 fusions during meiosis.
We attempted this experiment using the GFP-GBP system. However, we were unable to obtain strains in which Cdc14 and SPB components were both tagged with these binding partners. Although other explanations are possible, these findings are consistent with the idea that forced constitutive targeting of Cdc14 to the SPB is lethal also in vegetative cells. Due to the lethality we is. were unfortunately unable to assess the effect on meios None Competing Interests: I have read this submission. I believe that I have an appropriate level of expertise to confirm that it is of an acceptable scientific standard, however I have significant reservations, as outlined above.
Author Response 11 Feb 2017 , University of Edinburgh, UK

Adele Marston
We are grateful for your approval in principle and constructive comments. Here we address your reservations.
Major considerations: 1. Does the Cdc14-1 allele result in a complete loss of function in meiosis? A discussion of whether this is a hypomorph or complete loss of function would be helpful, especially for thinking about the conclusion that "Cdc14 is not absolutely required for CDK down-regulation at meiosis I exit".
It is true that we cannot be sure that the cdc14-1 allele is completely inactivated during meiosis. Our attempts to deplete or degrade Cdc14 during meiosis have been unsuccessful and therefore temperature-sensitive alleles are the best tools we have. We find that spindle disassembly is impaired in cdc14-1 conditions in mitosis but not meiosis under similar temperature conditions, leading us to suggest a lesser requirement for Cdc14 in meiotic spindle disassembly. Recently, similar findings have been reported for meiosis II using a different temperature sensitive allele (Argüello-Miranda 2017). et al., We have revised the text to include the possibility that Cdc14 retains partial function using these alleles and included a reference to this recent study. Figure 1 C, why do only 50% of the cells form 2 spindles instead of all of the cells? The duration of the movies needs to be increased to ensure that there is enough time for meiosis II to occur. The duration of the movies was 12h, however not all cells complete meiosis under the imaging conditions used for this experiment. Rather than bias the data by scoring only cells that complete meiosis II, all cells that enter meiosis were included in the analysis. This experiment was performed at a time where we were still establishing optimal conditions for live cell imaging of cells undergoing meiosis. Although we recognise that we could improve on this with our optimised conditions (where essentially all cells complete meiosis), we feel that the data presented clearly shows the effect of different mutants on meiosis I spindle disassembly.

In
3. I am having difficulty reconciling the results in Fig 7 with the model in Fig. 9. Why does the Spc42 fluorescence remain the same throughout meiosis I (metaphase I and anaphase I) when the SPB should be duplicating during the time? It would be helpful to have a graph showing the change in Spc42 fluorescence, tracking individual cells, also marking the time of anaphase I and anaphase II.

SPB fluorescence stays the same throughout meiosis I (metaphase I and anaphase I) because
SPBs are already duplicated before metaphase I (i.e. in the ndt80D arrest -see Figure 7C). We do note, however, that a greater increase in fluorescence might be expected during the later stages of anaphase I in wild type cells. The source data for this experiment that would allow the reader to https://osf.io/g5cmh/ ( ).

make traces of individual cells is available at
Marston, 2016 The fluorescence analysis of Spc42-tdTomato is complicated by two factors that mean we cannot make stand alone conclusions from it. First, in other experiments (not shown) we concluded that Spc42 incorporation into the SPB is continuous and exchange occurs. Second, the SPB undergoes a remodelling and maturation during meiosis II. As a result of these concerns we performed electron 1.

4.
incorporation into the SPB is continuous and exchange occurs. Second, the SPB undergoes a remodelling and maturation during meiosis II. As a result of these concerns we performed electron cdc14-1 microscopy, which clearly show a failure in SPB re-duplication in mutants. We believe that the EM data unequivocally supports our conclusions, while the fluorescence intensity data offer supportive, but not definitive evidence.
2. Does Sfi1 phosphorylation change throughout meiosis? Does Sfi1 remain phosphorylated in the Cdc14-1 cells? Despite extensive efforts to examine Sfi1 phosphorylation in wild type and cdc14-1 mutants undergoing meiosis, the findings were inconclusive. Judging by the presence of multiple Sfi1 species on western blots, Sfi1 appears to be extensively phosphorylated during meiosis in wild type cells. Our model predicts transient dephosphorylation at the time of anaphase I, while phosphorylation should persist in cdc14-1 mutants. Unfortunately, using this method, we were unable to reproducibly visualize Sfi1 dephosphorylation during anaphase I during wild type cells, likely due to the very short time window in which it occurs and variable cell synchronization, even using the best methods currently available (NDT80 block-release).
A thought about the conclusions: The prematurely released Cdc14 in also 3.
pCLB2-CDC55 cells localized to SPBs but does not allow SPBs reduplication. Does this finding suggest that the Cdc14 SPB localization is required but not sufficient to promote SPB duplication if the SPBs have not first separated? SPBs undergo the first round of duplication in pCLB2-CDC55 cells but do not appear to be able to separate. The licensing model predicts that, following Cdc14 dephosphorylation of Sfi1, CDK activity is required to separate the SPBs. We suggest that Cdc14 at the SPB counteracts CDK activity, thereby preventing separation of SPBs. Therefore, we agree, SPB separation may be a pre-requisite for duplication.

Minor points:
The last sentence in the abstract should end in "meiosis II transition." It currently says meiosis I.

We have corrected this
The sentence in the last paragraph on page 8 should read, "An exception is the two component GAP Bub2/Bfa1 which localized symmetrically at SPBs during metaphase I, anaphase I by asymmetrically during metaphase II".

Corrected.
The reference to the white arrow in Figure 8 is right after talking about vesicle formation, but the arrow does not point at the vesicles. The arrows mark the SPBs, not the vesicles, as written in the figure legend. We made this clearer by writing "SPB1, white arrow", "SPB2, white arrow". On page 16, the reference to Fig. 7F is missing and one of the references to 7E should be to 7G. We have corrected this.

None
Competing Interests: