Zwint-1 is required for spindle assembly checkpoint function and kinetochore-microtubule attachment during oocyte meiosis

The key step for faithful chromosome segregation during meiosis is kinetochore assembly. Defects in this process result in aneuploidy, leading to miscarriages, infertility and various birth defects. However, the roles of kinetochores in homologous chromosome segregation during meiosis are ill-defined. Here we found that Zwint-1 is required for homologous chromosome segregation during meiosis. Knockdown of Zwint-1 accelerated the first meiosis by abrogating the kinetochore recruitment of Mad2, leading to chromosome misalignment and a high incidence of aneuploidy. Although Zwint-1 knockdown did not affect Aurora C kinase activity, the meiotic defects following Zwint-1 knockdown were similar to those observed with ZM447439 treatment. Importantly, the chromosome misalignment following Aurora C kinase inhibition was not restored after removing the inhibitor in Zwint-1-knockdown oocytes, whereas the defect was rescued after the inhibitor washout in the control oocytes. These results suggest that Aurora C kinase-mediated correction of erroneous kinetochore-microtubule attachment is primarily regulated by Zwint-1. Our results provide the first evidence that Zwint-1 is required to correct erroneous kinetochore-microtubule attachment and regulate spindle checkpoint function during meiosis.


Zwint-1 knockdown accelerates passage through meiosis I.
To determine the function of Zwint-1 in meiosis, we microinjected long double-stranded RNA (dsRNA) targeting Zwint-1 into germinal vesicle (GV) stage oocytes and cultured them for 13 hours in the presence of 3-isobutyl-1-methylxanthine (IBMX) to block GV breakdown (GVBD). The level of Zwint-1 mRNA was significantly decreased by dsRNA injection (Fig. 1a). Knockdown of Zwint-1 on the protein level was also confirmed by immunostaining (Fig. 1b). This result also demonstrates that Zwint-1 is localized at the kinetochore during meiosis, consistent with the observations in somatic cells 13,15,18,19 . Zwint-1-knockdown oocytes underwent polar body extrusion (PBE), with no significant differences observed between control and knockdown oocytes (Fig. 1c). However, the timing of PBE was accelerated by 1 hour in Zwint-1-depleted oocytes compared with the control oocytes, suggesting that depletion of Zwint-1 results in precocious anaphase onset, followed by premature PBE. Given that anaphase onset and subsequent PBE are regulated by anaphase-promoting complex/cyclosome (APC/C)-mediated cyclin B degradation 20,21 , we monitored the cyclin B level after injection of mRNA encoding cyclin B-GFP. Consistent with the observed premature PBE, cyclin B levels also exhibited an early decrease in Zwint-1-knockdown oocytes (Fig. 1d), indicating accelerated passage through MI.
Acceleration of meiosis I is due to a failure to recruit Mad2 at kinetochore in Zwint-1-knockdown oocytes. Since the SAC controls the timing of PBE by regulating APC/C activity 22-24 , we next asked whether the accelerated PBE in Zwint-1-depleted oocytes is a result of the failure to activate the SAC. To test this possibility, oocytes were cultured with nocodazole from 4 hours after GVBD to activate the SAC. While less than 20% of all oocytes exhibited PBE in the control group, about 85% of all Zwint-1-depleted oocytes showed PBE (Fig. 2a). This phenotype was specific to Zwint-1-knockdown, as it was rescued by overexpression of human Zwint-1 (Fig. 2a). To further confirm the inactivation of the SAC, the localizations of Mad2 and BubR1 were determined after Zwint-1 knockdown. Interestingly, BubR1 remained at kinetochores, while Mad2 no longer localized to kinetochores in Zwint-1-depleted oocytes (Fig. 2b,c). These results suggest that Zwint-1 selectively regulates the kinetochore recruitment of SAC proteins during meiosis, consistent with a previous report in somatic cells 19 . Therefore, our results demonstrate that Zwint-1 is required to localize Mad2 at kinetochores, but not BubR1.
Abnormal chromosome alignment and high incidence of aneuploidy in Zwint-1-knockdown oocytes. Given that defects in SAC activity increase the risk of aneuploidy 25 , we hypothesized that loss of Zwint-1 causes misaligned chromosomes during meiosis and, consequently, aneuploidy. To test this hypothesis, oocytes injected with dsRNA were cultured to the MetI stage and the positions of homologous chromosomes were determined. Whereas chromosomes were well-aligned at the metaphase plate in control oocytes, Zwint-1-knockdown oocytes exhibited an increased incidence of chromosome misalignment (Fig. 3a,b). To further confirm chromosomal defects in Zwint-1-knockdown oocytes, we imaged oocytes overexpressing histone H2B-mCherry ( Fig. 3c and Supplementary Movie S1). In addition to premature PBE, homologous chromosomes were misaligned and separated improperly, showing lagging chromosomes. Moreover, sister chromatids remained misaligned during MetII stage following homologous chromosome segregation, suggesting the occurrence of aneuploidy in Zwint-1-knockdown oocytes. Indeed, we observed a high incidence of aneuploidy in Zwint-1-knockdown oocytes (Fig. 3d).
We next examined kMT attachment, since misaligned chromosomes and aneuploidy are often caused by inappropriate kMT attachment. Microtubules that are stably attached to kinetochores are cold-stable, whereas unattached microtubules are immediately disassembled after cold treatment 26 . Thus, we performed cold treatment with oocytes at the MetI stage and observed the resultant spindle morphology. Whereas distinct clustered kMTs were observed in control oocytes, unattached kinetochores and misaligned chromosomes were detected in Zwint-1-depleted oocytes (Fig. 3e). These results suggest that Zwint-1 is required for correct kMT attachment, thereby ensuring chromosome alignment on the metaphase plate during MI. Taken together, our data demonstrate that Zwint-1 is essential for accurate homologous chromosome segregation by regulating kMT attachment, since disruption of Zwint-1 function resulted in abnormal chromosome segregation and aneuploidy. Zwint-1 is required for chromosome alignment and kMT attachment independently of Mad2. Mad2 delays the onset of anaphase, allowing time for error correction 27,28 . Since Zwint-1 knockdown accelerated the passage through meiosis I due to the failure of Mad2 recruitment at kinetochores, we speculated that the chromosome misalignment induced by Zwint-1 knockdown is a result of insufficient time to repair erroneous kMT attachment as a consequence of Mad2 deficiency. To test this hypothesis, oocytes were cultured with MG132, a proteasome inhibitor, to delay anaphase onset, thereby allowing more time for error correction. kMT attachment was then assessed after cold treatment. Importantly, lagging chromosomes were still observed in Zwint-1-knockdown oocytes, whereas chromosomes were well-aligned in oocytes injected with anti-Mad2 antibodies (Fig. 4a,b). Moreover, the spindles were clustered and persisted after cold treatment in oocytes injected with anti-Mad2 antibodies, Normalized expression of Zwint-1 was quantified from two independent experiments and shown to the right panel (*p = 0.0098). (b) Immunostaining of Zwint-1-knockdown oocytes. Oocytes were fixed at 6-7 h after GVBD and immunostained with anti-Zwint-1 antibody. Kinetochores and DNA were stained with anti-centromere antibody (ACA) and DAPI, respectively. Shown are representative of three independent experiments with at least 30 oocytes. Scale bar, 10 μ m. Quantification of Zwint-1 signals is shown in the right of images (*p < 0.0001). The number of oocytes analyzed is shown above the bars. (c) Timing of polar body extrusion (PBE) was determined in Zwint-1-knockdown oocytes. Data are mean ± SEM from three independent experiments with the indicated number of oocytes (*p < 0.05). The mean time at which half of oocytes completed PBE was 6 h 58 min for Zwint-1 knockdown and 8 h 5 min for control. (d) Oocytes depleted of Zwint-1 were injected with mRNA encoding cyclin B-GFP, and GFP levels were measured every 20 min up to 8 h. Quantification of cyclin B levels is shown. For each oocyte, the fluorescence intensity was normalized to the intensity recorded 4 h after GVBD. Data are mean ± SEM from one experiment with the indicated number of oocytes (*p < 0.05). Note that the cyclin B degradation was initiated around 7 h after GVBD for control and 6 h after GVBD for Zwint-1-knockdown oocytes.
whereas Zwint-1-knockdown oocytes exhibited faint arrays of microtubules and unattached kinetochores (Fig. 4a). This observation is not an off-target effect of antibody injection, because antibody-injected oocytes were able to overcome nocodazole-mediated MetI arrest (Fig. 4c). Similar results were obtained when either Zwint-1 or Mad2 was depleted using siRNA, further supporting that chromosomal defects in Zwint-1 knockdown oocytes are not simply due to Mad2 deficiency ( Supplementary Fig. S1).
To assess the consequences of prolonged MI, oocytes were released from MG132-mediated inhibition and the rates of aneuploidy were determined. Consistent with the observations that prolonging the duration of MI rescued Mad2 inhibition-induced defects but did not restore Zwint-1-knockdown-mediated defects in chromosome alignment and kMT attachment, the aneuploidy rate was dramatically reduced in anti-Mad2 antibody-injected oocytes, whereas Zwint-1-knockdown oocytes still had a high incidence of aneuploidy (Fig. 4d,e). This result suggests that chromosome missegregation and aneuploidy induced by Mad2 inhibition are simply due to insufficient time for erroneous kMT attachment to be repaired. However, defects following Zwint-1 knockdown is not rescued by prolonging the duration of MI. Therefore, our data suggest that aneuploidy mediated by Zwint-1 knockdown is not a direct consequence of Mad2 deficiency, although the acceleration of MI after Zwint-1 knockdown is possibly dependent on the loss of SAC activity. Zwint-1 is required for Aurora C kinase to correct kMT attachment. Since Aurora B/C kinase activity is well-established to regulate kMT attachment and chromosome segregation 8,10,12,29,30 , we next asked whether inhibition of Zwint-1 affects Aurora B/C kinase activity, thereby producing defects in   chromosome alignment and SAC activation. To this end, we examined the localization and activity of Aurora C kinase. Aurora C kinase activity was assessed by measuring the phosphorylation of histone H3 serine 10 (H3S10ph), one of the best-known substrates of Aurora B/C kinase. Notably, no significant differences were observed in either Aurora C localization or the level of phosphorylated H3S10 ph in Zwint-1-knockdown oocytes compared with control oocytes (Fig. 5a,b). This result suggests that impaired SAC activation and kMT attachment in Zwint-1-knockdown oocytes does not result from disrupted Aurora C kinase activity.
To investigate relationship between Zwint-1 and Aurora C kinase, oocytes were treated with Aurora B/C kinase inhibitor ZM447439 and chromosome alignment was examined. To prevent premature PBE caused by ZM447439 treatment, oocytes were cultured with MG132. Similar to Zwint-1 knockdown, the chromosome misalignment was observed in ZM447439 treated oocytes. Importantly, the defects were repaired when the oocytes were washed out from ZM447439 in oocytes injected with dsEGFP. However, the chromosome misalignment was not restored in Zwint-1-knockdown oocytes even after ZM447439 washout (Fig. 5c,d). These results imply that Zwint-1 is required for Aurora C kinase to correct chromosome misalignment during meiotic maturation.

Discussion
Accurate chromosome segregation is essential for maintaining genomic integrity. One of the key steps in this process is kinetochore assembly. However, the functions of kinetochore proteins during oocyte meiosis have not been extensively studied. Here, we demonstrated that Zwint-1 plays a pivotal role in homologous chromosome segregation by regulating kMT attachment and the SAC during meiotic maturation.
We found that Zwint-1 knockdown accelerated anaphase I onset by abrogating the activation of SAC, leading to premature PBE. Interestingly, BubR1 was still localized to kinetochores in Zwint-1-knockdown oocytes, while Mad2 was not. Consistent with this observation, BubR1, but not Mad2, has been shown to remain at kinetochores in mitotic cells depleted of Zwint-1 19 . This result implies that the kinetochore localizations of BubR1 and Mad2 are differentially regulated during meiosis as in mitosis. This conclusion is supported by the finding that BubR1 localization does not depend on Ndc80, a component of the KMN network, whereas Mad2 is directly associated with Ndc80 31,32 . Therefore, it is likely that two independent pathways are present in mouse oocytes for recruiting SAC proteins to unattached kinetochores.
BubR1 was recently shown to undergo autophosphorylation in a process dependent on centromere-associated protein E (CENP-E). CENP-E is a kinesin-like microtubule motor protein that is essential for chromosome congression during mitosis 33,34 . Moreover, CENP-E-dependent BubR1 autophosphorylation at unattached kinetochores is required for accurate chromosome segregation and SAC activation 33 . In this regard, it is possible that Zwint-1 regulates CENP-E-dependent BubR1 phosphorylation instead of BubR1 localization. Thus, Zwint-1 knockdown disrupts BubR1 phosphorylation without affecting BubR1 localization, thereby inhibiting the faithful segregation of homologous chromosomes during meiosis. It will be interesting to determine the phosphorylation status of BubR1 in Zwint-1-knockdown oocytes in future studies. Given that chromosome segregation in female meiosis I is notoriously error-prone 22,23,27 , it is also possible that oocytes enter anaphase onset without full inactivation of the SAC. Indeed, oocytes with incorrect kMT attachment can exist simultaneously with an active APC and meiotic progression [35][36][37][38] .
Aurora B/C kinases play a pivotal role in the correction of erroneous kMT attachment and SAC activation 8,10,12,29,30,39,40 . Hence, Aurora B/C kinases have been considered to be at the top of the signaling cascade that ensures accurate chromosome segregation. Since the phenotype of Zwint-1 knockdown is similar to that of Aurora B/C inhibition, it is possible to hypothesize that Zwint-1 disruption affects the activity of Aurora C kinase, thereby impairing SAC activation and kMT attachment during meiosis. However, this is not the case, since neither the localization nor the activity of Aurora C kinase was affected by Zwint-1 knockdown. However, we do note that chromosomal defects following ZM447439 treatment were not restored after removing the inhibitor in Zwint-1-knockdown oocytes, whereas the defects was rescued after the inhibitor washout in the control oocytes. These data, along with the observation that Aurora C kinase activity remains intact in Zwint-1-knockdown oocytes, indicate that Zwint-1 is a key component for Aurora C kinase-mediated correction of erroneous kMT attachment during meiosis. In addition, the erroneous kMT attachment induced by Mad2-mediated SAC inactivation was rescued when the MI duration was prolonged. Therefore, it is likely that the SAC, at least Mad2 associated portion, is not directly involved in the error correction process. Instead, this SAC simply extends the duration of MI to allow sufficient time for enough error correction to take place before anaphase onset. However, it could not be generalized because several checkpoint proteins including BubR1 and Mps1 are required for kMT attachment 22,41 .
The incidence of aneuploidy has been shown to be dramatically increased in aged oocytes 42 . Although the exact molecular mechanisms underlying this phenomenon are not yet known, one compelling hypothesis is age-dependent decline in SAC activity. Consistent with this hypothesis, SAC deficiencies in oocytes from older women have been proposed to contribute to aneuploidy 43 . Given that chromosome misalignment induced by Mad2 deficiency is restored when MI duration is prolonged, extension of MI duration may reduce the aneuploidy rate in aged oocytes. However, if the level of Zwint-1 also decreases with oocyte aging, like other SAC proteins, Aurora C kinase-mediated error correction may also be impaired in aged oocytes. In this case, prolonged incubation of aged oocytes may not reduce the aneuploidy rate. Therefore, it will be interesting to examine the level of Zwint-1 in aged oocytes in future studies.
Collectively, our data suggest that Zwint-1 is a key component that delivers signals from Aurora C kinase, which is localized at the inner centromeres, to processes occurring at the outer kinetochores, such as the recruitment of SAC proteins and the correction of erroneous kMT attachment. We also showed that depletion of Zwint-1 impairs the primary function of Aurora C kinase, which is the repair of erroneous kMT attachment. Our results provide the first evidence that Zwint-1 is required for Aurora C kinase to regulate kMT attachment and SAC activation during oocyte meiosis.

Methods and Materials
Animals. All mice used in this study were 3-to 5-week-old CD-1 female mice (Koatech). Animals were maintained with food and water ad libitum under a 14-hour light/10-hour dark cycle. All procedures for mouse care and use were conducted in accordance with the guidelines and approved by the Institutional Animal Care and Use Committees of Sungkyunkwan University.
Oocyte collection and culture. Ovaries were isolated from 4-to 6-week-old CD-1 female mice 46-48 hours after intraperitoneal injection of 10 IU of pregnant mare's serum gonadotrophin (PMSG). Oocytes were collected from follicles and recovered in M2 medium supplemented with 200 μ M 3-isobutyl-1-methylxanthine (IBMX) to prevent meiotic resumption. For in vitro maturation, oocytes were washed and cultured in IBMX-free M2 or M16 medium in a 5% CO 2 atmosphere at 37 °C. All reagents and media were from Sigma, unless otherwise stated.
Cloning and in vitro mRNA synthesis. The full-length cDNA sequence encoding human Zwint-1 was provided by the Korea Human Gene Bank, Medical Genomics Research Center, KRIBB, Korea. mRNA for microinjection was prepared in vitro using an mMessage mMachine kit (Ambion), polyadenylated, and then purified with a Nucleospin RNA clean-up kit (Macherey-Nagel). For double stranded RNA, PCR products were used as a template for in vitro transcription using a MEGAscript Kit (Ambion). The primers used were: EGFP, ATTAATACGACTAACTATAGGGAGAATGGTGAGCAAGGGCGAG and ATTAATACGACTCACTATAGGGAGAGCTCGTCCATGCCGAGAG; Zwint-1, ATTAATACGACTC ACTATAGGGAGATGAGCAGAGCCCTGATGC and ATTAATACGACTCACTATAGGGAGATTGGC TGTAAGGGCTCGC.
Microinjection. Approximately 5-10 pl dsRNA, siRNA or mRNA was microinjected into the cytoplasm of oocytes using a FemtoJet microinjector (Eppendorf) with a Leica inverted microscope (DMIRB) equipped with a micromanipulator (Narishige). After injection, oocytes were cultured for 13 hours in M2 or M16 medium containing IMBX. The oocytes were then transferred to fresh medium and cultured under mineral oil at 37 °C in an atmosphere of 5% CO 2 in air. The small interfering RNAs (siRNAs) were purchased from a local company (Bioneer). The mixture of two siRNAs targeting either Zwint-1 or Mad2 was used. The siRNA sequences were as following: Zwint-1, CUAAUUCUCCAAGCUGAUU and UCAUGCUAACUUUGACAGU; Mad2, CUAAGAUCCGACAUUGAUA and GUGUUCUCCG UUCGAUCUA.
Time-lapse imaging. Oocytes injected with histone H2B-mCherry or cyclin B-GFP mRNA were placed in a heated chamber and fluorescence images were acquired using a Nikon Eclipse Ti inverted microscope equipped with a CCD cooled camera (DS-Qi1Mc, Nikon). For quantification of cyclin B level, GFP fluorescence intensity was quantified using ImageJ software (National Institutes of Health).
Nocodazole and cold treatment of oocytes. For nocodazole treatment, 400 nM nocodazole (Sigma) dissolved in dimethyl sulfoxide (DMSO) was added to the culture medium at the indicated time point; control oocytes were treated with an equivalent volume of DMSO.
For aneuploidy analysis, oocytes that had been pretreated for 2 hours in 200 μ M monastrol (Sigma) were fixed and stained with the indicated antibodies, as described previously 45 . Image analysis. Images were acquired using a confocal laser-scanning microscope (LSM 700; Zeiss) equipped with a C-Apochromat 63 × /1.2 water immersion objective. Optical sections were obtained at 1 μ m intervals and converted into maximum intensity projections. Data analysis was performed using ZEN 2010 LSM software (Zeiss) and ImageJ software (National Institutes of Health).

Statistical analysis.
Statistical analysis was performed with GraphPad Prism (GraphPad Software).
Data are representative of at least three independent experiments unless otherwise specified. The significance of differences between groups was analyzed by Student's t-test; p values < 0.05 were considered statistically significant.