HDAC3 promotes meiotic apparatus assembly in mouse oocytes by modulating tubulin acetylation

Acetylation of the lysine ε-amino group is a dynamic post-translational modification regulated by the opposing activities of lysine acetyltransferases and histone deacetylases (HDACs) (Seto and Yoshida, 2014; Zhan et al., 2017). To date, the HDAC family includes 18 members divided into Class I (HDACs 1-3 and 8), Class II (HDACs 4-7 and 9, 10), Class III (Sirt1-7) and Class IV (HDAC11) (Horiuchi et al., 2009). Class I, II and IV HDACs are zinc-dependent enzymes collectively referred to as classical HDACs, whereas Class III HDACs, commonly referred to as Sirtuins, require NAD⁺ for their activity (Haberland et al., 2009; Yang and Seto, 2008). HDACs, as a vast family of enzymes, have been reported to play crucial roles in diverse biological processes, largely through their effects on transcription and chromatin conformation. Among them, Class I HDACs are ubiquitously expressed nuclear enzymes and conserved in a wide range of species (Mahlknecht et al., 1999; Vidal and Gaber, 1991; Yang et al., 1997). Conditional targeting of HDAC1 and HDAC2 in mice has been shown to lead to female infertility (Ma et al., 2012). In particular, HDAC2 regulates chromosome segregation and kinetochore function via H4K16 deacetylation during oocyte maturation (Ma and Schultz, 2013). In addition, HDAC6 appears to be an essential factor modulating oocyte maturation and meiotic apparatus assembly (Ling et al., 2017; Zhou et al., 2017). HDAC3 is required for the proliferation of adult neural stem/progenitor cells by modulating cyclin-dependent kinase levels (Jiang and Hsieh, 2014). HDAC3 was found to localize to the mitotic spindle of HeLa cells and participate in the establishment of interactions between kinetochore and microtubules (Ishii et al., 2008). Recently, it has shown that deletion of HDAC3 in adult beta cells improves glucose tolerance via increased insulin secretion (Remsberg et al., 2017). However, the function of HDAC3 in meiosis remains unknown.

Poor oocyte quality is a major cause of female infertility. In mammals, oocytes are arrested within ovarian follicles at the diplotene stage of the first meiotic prophase, which is also called germinal vesicle (GV) stage. These fully grown oocytes resume meiosis upon hormone stimulation. During this process, accurate control of spindle assembly and chromosome movement is essential for producing a healthy oocyte. Kinetochores provide the principal attachment of chromosomes to microtubules, while spindle assembly checkpoint (SAC) is the mechanism that monitors and responds to kinetochore-microtubule interaction. Unfaithful chromosome segregation during meiosis could lead to aneuploidy in the oocyte. Egg aneuploidy is a leading cause of spontaneous abortions, birth defects and developmental disabilities in humans (Ma et al., 2014). Although numerous molecules have been proposed to affect spindle/chromosome organization in oocyte meiosis, the underlying pathways that regulate the meiotic structure remain to be explored.

In the present study, we investigated the potential roles of HDAC3 in mouse oocytes by knockdown and overexpression assays. Our results indicate that HDAC3 depletion adversely influences meiotic maturation, especially spindle assembly and chromosome organization.

To explore the role of HDAC3 in meiosis, we first examined its distribution at different stages of the mouse oocyte by immunostaining and confocal microscopy. HDAC3 is predominantly distributed in the nucleus of GV oocytes (Fig. 1A, arrowhead). However, accompanying meiotic resumption, HDAC3 resides in the cytoplasm, with signals accumulating on the meiotic spindle region from pre-metaphase I to metaphase II. Double staining was performed to confirm the spatial relationship between HDAC3 and the spindle. HDAC3 is indeed colocalized with the meiotic spindle in mouse oocytes, as evidenced by the overlapping fluorescence signals (Fig. 1B). In addition, to confirm whether the HDAC3 antibody staining pattern is specific, exogenous Myc-HDAC3 was ectopically expressed in mouse oocytes, and then metaphase cells were labeled with anti-Myc antibody. Myc-HDAC3 is distributed in the cytoplasm, with strong signals on the spindle region (Fig. S1, arrowheads), similar to the localization of endogenous HDAC3 protein.

The progression of oocyte maturation includes meiotic resumption indicated by germinal vesicle breakdown (GVBD), microtubule organization in meiosis I (MI), extrusion of the first polar body (Pb1), and then progression to metaphase II (MII) in preparation for fertilization. To explore the role of HDAC3, fully grown oocytes were microinjected with specifically designed HDAC3 siRNAs, with a negative siRNA injected as control. After injection, the oocytes were arrested at GV stage for 20 h with milrinone to degrade mRNA. Based on western blotting results, we found that siRNA #2 led to the most substantial reduction in HDAC3 protein in oocytes (Fig. 2A). qRT-PCR assay confirmed that siRNA #2 could efficiently decrease endogenous Hdac3 mRNA levels in mouse oocytes (Fig. 2B). Similarly, siRNA knockdown also resulted in a significant decrease in HDAC3 immunofluorescent signal in oocytes (Fig. S2). Therefore, siRNA #2 was used in subsequent knockdown experiments. Of note, we found that the siRNA-injected oocytes displayed wide variation in HDAC3 staining (see images I and II in Fig. S2), which might contribute to distinct oocyte phenotypes (see below).

After 3 h of culture, both control and HDAC3 knockdown (KD) oocytes resumed meiosis normally, as evidenced by the similar GVBD rate (Fig. 2D). However, only 42.5% of HDAC3 KD oocytes extruded Pb1 after 14 h of culture, a significant decrease compared with controls (Fig. 2C arrowheads, 2E). In addition, a high frequency of HDAC3 KD oocytes underwent symmetric division, displaying the ʻ2-cell like' phenotype (Fig. 2C, asterisks). Specifically, nuclear staining and quantitative analysis revealed that 39.8% of HDAC3 KD oocytes were arrested at MI, which was significantly higher than that of control oocytes (Fig. 2F). To examine why the rest of the oocytes are still able to progress through meiosis, HDAC3 staining was performed on siRNA-injected oocytes with or without Pb1. As shown in Fig. S3, HDAC3 intensity was significantly higher in MII oocytes than in MI-arrested oocytes, indicating that the different phenotypes are associated with the degree of HDAC3 depletion. These observations strongly suggest that HDAC3 is required for orderly oocyte maturation and meiotic divisions.

The specific distribution of HDAC3 on the spindle and its effects on maturation progression prompted us to investigate whether HDAC3 depletion in oocytes influences the meiotic apparatus. Control and HDAC3 KD oocytes were immunolabeled with anti-tubulin antibody to visualize the spindle and counterstained with propidium iodide (PI) to visualize chromosomes. Significantly, HDAC3 KD elevated the percentage of spindle defects and chromosome misalignment in metaphase oocytes (Fig. 3A,B), which showed elongated/multipolar spindles (arrows) and scattered chromosomes (arrowheads). These phenotypes contrasted sharply with control oocytes showing the typical barrel-shaped spindle and well-aligned chromosomes at the equator. Together, these findings indicate that HDAC3 KD oocytes are unable to properly assemble the spindle and align the chromosomes in meiosis.

Accurate chromosome alignment and segregation depend on the proper attachment of kinetochores to microtubules emanating from opposite spindle poles (Tauchman et al., 2015). Given the spindle/chromosome disorganization in HDAC3 KD oocytes, we examined whether kinetochore-microtubule (K-MT) attachments were compromised correspondingly. MI oocytes were labeled with CREST (also known as Ss18l1) antibody to detect kinetochores, with anti-tubulin antibody to visualize spindle, and co-stained with Hoechst 33342 for chromosomes (Fig. 4A). We found that the majority of normal oocytes presented amphitelic K-MT attachments (every kinetochore attached to one pole; chromosomes labeled 1 and 2 in Fig. 4A,B). By contrast, the frequency of K-MT misattachments in HDAC3 KD oocytes was greatly increased relative to control cells, including lost attachments (kinetochore attached to neither of the poles; chromosome 3), merotelic attachment (one kinetochore attached to both poles; chromosome 4), as well as undefined attachment (chromosome 5). These erroneous K-MT attachments are likely to be the major factor contributing to the chromosome alignment failure observed in HDAC3 KD oocytes, influencing the establishment of stable chromosome biorientation.

The SAC is a ubiquitous safety device that monitors K-MT interactions and ensures accurate chromosome segregation (Overlack et al., 2015). Once the K-MT attachments are disturbed, the anaphase-inhibitory signal is triggered by the SAC (Tauchman et al., 2015). Taking into account the impaired K-MT attachments and MI arrest in HDAC3 KD oocytes, we proposed that the SAC might be provoked when HDAC3 is depleted. To test this, control and HDAC3 KD oocytes were immunolabeled for BubR1 (also known as Bub1b), an integral component of the checkpoint complex, to evaluate SAC activity. In control cells, BubR1 was detected on unattached kinetochores during pre-metaphase, and then completely disappeared once kinetochores had become properly attached to microtubules at metaphase I (Fig. 5). Remarkably, the BubR1 signal on kinetochores was dramatically increased in those HDAC3 KD oocytes arrested at MI (Fig. 5A,B, arrowhead), implying that SAC was activated. Collectively, these findings suggest that the SAC surveillance mechanism might be a major pathway mediating the effects of HDAC3 KD on meiotic progression in oocytes.

In normal cells, SAC delays anaphase onset until all chromosomes are well aligned at the equatorial plane and attached to the spindle apparatus. Dysregulated SAC is thought to be the major driving force of aneuploidy generation. Because of the high frequency of spindle/chromosome defects in HDAC3 KD oocytes, we examined whether the incidence of aneuploidy was accordingly increased. MII oocytes were processed for chromosome spreading and kinetochore labeling. Representative images of euploidy and aneuploidy are shown in Fig. 6A. Quantitative analysis (Fig. 6B) further revealed that HDAC3 KD resulted in a ∼3- to 4-fold increase in the incidence of aneuploidy eggs compared with controls. To confirm this karyotypic phenotype, we also counted chromosome number using a monastrol-based in situ spreading technique (Duncan et al., 2009; Lane et al., 2010), and similar results were obtained (Fig. S4). These findings suggest that HDAC3 KD disrupts spindle/chromosome organization and provokes SAC during meiosis, which in turn elevates the incidence of aneuploidy.

Tubulin is one of the most abundant non-histone proteins subject to acetylation, which occurs on lysine 40 of the α-tubulin subunit (Zilberman et al., 2009). Tubulin acetylation was first identified in flagellar axonemes of Chlamydomonas reinhardtii (Eddé et al., 1991). Subsequent studies demonstrated that this modification is also present in basal bodies and a subset of cytoplasmic microtubules (Piperno et al., 1987). Furthermore, tubulin acetylation was shown to be mainly on stable microtubules resistant to depolymerization induced by cold shock (Li and Yang, 2015). Of note, HDAC3 is able to modulate tubulin acetylation in a human prostate cancer cell line (Bacon et al., 2015). Hence, the involvement of HDAC3 in meiotic apparatus assembly prompted us to consider tubulin as a potential target mediating this process.

We found that the acetylation levels of α-tubulin were dramatically increased in HDAC3 KD mouse oocytes as compared with controls, whereas overexpression of HDAC3 significantly reduced the fluorescence intensity of acetylated tubulin (Fig. 7A,B). These results indicate that the acetylation status of tubulin depends on HDAC3 expression in mouse oocytes.

It has been shown that the substitution of lysine (K) with glutamine (Q) mimics an acetylated amino acid state, whereas substitution with arginine (R) mimics deacetylation (Gal et al., 2013; Zhang et al., 2015). To investigate whether tubulin acetylation affects meiotic structure in oocytes, we constructed site-specific mutants (K-to-Q and K-to-R) targeting lysine 40 (K40) of tubulin. The mRNA encoding these tubulin mutants was microinjected into oocytes and the effects on spindle/chromosome organization evaluated. As shown in Fig. 8A,B, tubulin-K40Q resulted in an almost 3-fold increase in the incidence of spindle/chromosome defects compared with the unmutated tubulin (tubulin-WT) control. Moreover, tubulin-K40Q also efficiently induced activation of the SAC in meiosis as compared with the tubulin-K40R mutant (Fig. 8C,D). These observations suggest that tubulin acetylation is essential for spindle assembly and chromosome congression in meiotic oocytes.

Finally, we asked whether the non-acetylatable mimetic mutant of tubulin could rescue at least some of the phenotypic defects seen in HDAC3 KD oocytes. Tubulin-K40R was microinjected into HDAC3 KD oocytes and mature oocytes were then stained to evaluate spindle/chromosome organization. Tubulin-K40R significantly lowered the proportion of abnormal spindle/chromosomes in HDAC3 KD oocytes (Fig. 8E,F). These results suggest that non-acetylated tubulin-K40R is capable, at least in part, of preventing the meiotic defects that arise in the oocyte as a result of HDAC3 loss.

Taken together, we conclude that HDAC3 controls assembly of the meiotic apparatus in oocytes by maintaining the hypoacetylation state of tubulin.

HDAC3 is a crucial epigenetic modifying enzyme and regulator of the maintenance of chromatin structure and genome stability, ensuring normal cell viability (Bhaskara et al., 2010; Takami and Nakayama, 2000). In this study, we showed that HDAC3 predominantly accumulates on the spindle accompanying meiotic resumption (Fig. 2), which is in line with previous findings in somatic cells (Ishii et al., 2008). In support of this observation, we further discovered that HDAC3 KD in oocytes disrupts spindle assembly and chromosome alignment and impairs K-MT interaction. K-MT linkage is essential for the poleward chromosome movement in meiosis. The kinetochore is composed of a number of conserved protein complexes that assemble on the centromere, a specialized chromatin domain (Santaguida and Musacchio, 2009). The correct chromosome alignment to the metaphase plate requires bipolar attachment and tension between kinetochore and microtubule. It is therefore conceivable that the high percentage of spindle/chromosome abnormalities in HDAC3 KD oocytes is a consequence of K-MT misattachments. These attachment errors need be corrected in normal cells; otherwise, they would cause chromosome separation defects if they persisted until anaphase. Consistent with this notion, the frequency of aneuploidy was significantly increased in HDAC3-depleted oocytes relative to controls (Fig. 6). We propose that, in HDAC3 KD oocytes, compromised K-MT stability and the resulting meiotic defects contribute to the generation of aneuploid eggs observed in our experiments.

Regulation of microtubule dynamics and organization is essential for bipolar spindle assembly and function. Microtubules are essential cytoskeletal polymers that are made of repeating α/β-tubulin heterodimers (Nogales, 2000). Tubulin is one of the most abundant non-histone proteins subject to acetylation. It is worth noting that α-tubulin acetylation serves as a marker for the presence of stable microtubules and affects the activity of microtubule-associated proteins and microtubule-based motors (Chen et al., 2013). Acetylated microtubules have been detected in mouse oocytes (Schatten et al., 1988); however, their function(s) during meiosis remains unknown. Here, by constructing site-specific mutants targeting tubulin K40, we demonstrated that tubulin acetylation is essential for spindle assembly and chromosome congression in meiotic oocytes (Zhan et al., 2017) (Fig. 7). HDAC3 has been shown to indirectly modulate tubulin acetylation in human prostate cancer cells (Bacon et al., 2015). In our study, we observed the enrichment of HDAC3 protein on the spindle during oocyte maturation. Moreover, as revealed by our knockdown and overexpression experiments (Fig. 7), HDAC3 evidently controls the acetylation level of α-tubulin in oocytes. Hence, the evidence strongly suggests that α-tubulin may be an important deacetylation target of HDAC3 in mouse oocytes. Interestingly, recent reports show that treatment with tubastatin A, an HDAC6 inhibitor, also increases the acetylation level of α-tubulin in mouse oocytes (Ling et al., 2017; Zhou et al., 2017). It remains to be determined how α-tubulin hyperacetylation disrupts spindle assembly and chromosome movement during oocyte meiosis, potentially through affecting microtubule stability and K-MT interaction. Combined with our tubulin-K40R rescue experiments (Fig. 8), these data lead us to propose that HDAC3 depletion induces tubulin hyperacetylation, which in turn compromises microtubule stability and K-MT attachments, leading to spindle/chromosome organization failure and aneuploidy. Our data cannot exclude the possibility that HDAC3 might act on other targets in its function during oocyte maturation. Owing to the limitation in oocyte number and for technical reasons, we have not been able to analyze the relationship between α-tubulin acetylation and its stability in mouse oocytes. Further assays will be required to clarify these issues.

The proportion of MI-arrested oocytes was significantly increased by HDAC3 KD (Fig. 2). The kinetochore has been shown to function as a signaling hub that coordinates chromosome attachment, SAC activity, and cell cycle progression from metaphase to anaphase. In the presence of unattached kinetochores, SAC proteins form a complex that inhibits anaphase onset (Tauchman et al., 2015). By confocal microscopy, we found that the BubR1 signal persists at metaphase chromosomes in HDAC3 KD oocytes compared with controls (Fig. 5). Interestingly, tubulin-K40R mutant expression was able to partly rescue these phenotypes in HDAC3 KD oocytes (Figs 7, 8). Considering the frequent K-MT misattachments, we reasoned that a metaphase-anaphase transition block arises in HDAC3 KD oocytes from the SAC activation.

In summary, we have discovered a novel pathway through which HDAC3 modulates the meiotic apparatus in oocytes. Our data support a model whereby HDAC3, most likely through deacetylating tubulin, promotes the establishment of K-MT interactions, consequently ensuring proper spindle morphology and accurate chromosome movement during oocyte maturation.

All chemicals and culture media were purchased from Sigma unless stated otherwise.

Female 4-week-old ICR mice (Charles River Laboratories China, Beijing, China) were used in all experiments. Experiments were approved by the Animal Care and Use Committee of Nanjing Agricultural University and were performed in accordance with institutional guidelines.

Primary antibodies: rabbit polyclonal anti-HDAC3 antibodies (Santa Cruz Biotechnology, sc-376957; 1:150); goat polyclonal anti-BubR1 (ab28193) and mouse monoclonal anti-Myc tag (ab18185) antibodies (Abcam; 1:250); mouse polyclonal anti-acetylated tubulin (Lys 40) (Sigma, T7451; 1:200); human anti-centromere CREST antibody (Antibodies Incorporated, 15-234; 1:500); mouse monoclonal anti-α-tubulin (A5441) and FITC-conjugated anti-α-tubulin (F2168) antibodies (Sigma, 76074; 1:300). Secondary antibodies: Cy5-conjugated donkey anti-human IgG (709-605-149) and FITC-conjugated donkey anti-goat IgG (705-095-147) antibodies (Jackson ImmunoResearch Laboratories; 1:500); FITC-conjugated goat anti-rabbit IgG (ZSGB-BIO, China, ZF-0311) and TRITC-conjugated goat anti-rabbit IgG (ZSGB-BIO, China, ZF-0316) antibodies (Thermo Fisher Scientific; 1:300).

At 46-48 h after injection of 5 IU pregnant mare serum gonadotropin (PMSG), fully grown cumulus-enclosed oocytes were collected from ovaries of 6- to 8-week-old female ICR mice. To obtain denuded oocytes, cumulus cells were removed by repeatedly pipetting. For in vitro maturation, oocytes were cultured in M16 medium under mineral oil at 37°C in a 5% CO₂ incubator.

Total RNA was extracted from 50 denuded oocytes using the Arcturus PicoPure RNA Isolation Kit (Applied Biosystems), and cDNA was generated with the QIAquick PCR Purification Kit (Qiagen). Purified PCR products were digested with FseI and AscI (NEB) and cloned into the pCS2⁺ vector with Myc tags. The vectors encoding the Myc-tubulin substitution mutants (K40Q and K40R) were generated with the QuikChange Site-Directed Mutagenesis Kit (Stratagene).

For the synthesis of mRNA, the HDAC3/tubulin-pCS2⁺ plasmids were linearized by NotI. The capped cRNAs were in vitro transcribed with SP6 mMESSAGE mMACHINE (Ambion) and purified using the RNeasy Micro Kit (Qiagen) according to the manufacturer's instructions. Primers are listed in Table S1.

Microinjections of siRNA or mRNA with a Narishige microinjector were used to knock down or overexpress proteins in mouse oocytes, respectively. 2.5 pl 1 mM HDAC3 siRNA (5′-GACGGUACUAAAGUCACCUTT-3′; 5′-AGGUGACUUUAGUACCGUCTT-3′) was injected into oocytes for knockdown analysis, or an equivalent amount of negative siRNA (5′-UUCUCCGAACGUGUCACGUTT-3′; 5′-ACGUGACACGUUCGGAGAATT-3′; no significant sequence similarity to mouse, rat or human gene sequences) was injected as control. 10 pl mRNA solution (10 ng/µl) was injected into oocytes for overexpression analysis, with RNase-free PBS injected as control. After injection, oocytes were arrested at GV stage in medium with 2.5 µM milrinone for 20 h, and then were cultured in milrinone-free medium for maturation.

Oocytes were fixed in 4% paraformaldehyde for 30 min and permeabilized with 0.5% Triton X-100 for 20 min. After three washes, oocytes were blocked in 1% BSA in PBS for 1 h at room temperature. Sample were incubated overnight at 4°C with primary antibodies and then at room temperature for 1 h with secondary antibodies. To examine kinetochores, oocytes were colabeled with CREST antibody. Chromosomes were counterstained with Hoechst 33342 (blue, 1:300) or propidium iodide (red, 1:200) for 10 min. Oocyte samples were mounted on anti-fade medium (Vectashield) and examined under an LSM 710 laser scanning confocal microscope (Zeiss) equipped with 40× or 63× oil objectives. ImageJ software (National Institutes of Health) was used to quantify the intensity of fluorescence.

qRT-PCR was performed as described previously (Hou et al., 2015). Total RNA was isolated from 50 oocytes and cDNA was quantified using an ABI StepOnePlus Real-Time PCR System (Applied Biosystems). The fold change in gene expression was calculated using the ΔΔCt method with the housekeeping gene glyceraldehydes-3-phosphate dehydrogenase (Gapdh) as the internal control. The result was expressed as fold change relative to control. Primers are listed in Table S2.

One hundred oocytes were lysed in Laemmli sample buffer containing protease inhibitor at 100°C for 5 min. Proteins were separated by 10% SDS-PAGE, transferred to PVDF membranes by constant current at 60 V, and then blocked in PBST (PBS containing 0.1% Tween 20) with 5% fat-free milk for 1 h at room temperature. Membranes were then incubated with rabbit anti-HDAC3 antibody (1:300; Santa Cruz, sc-376957) overnight at 4°C. After three washes in PBST and incubation with anti-rabbit horseradish peroxidase-conjugated secondary antibodies (ZSGB-BIO, China, ZB-2301; 1:2000), the protein bands were analyzed with an ECL Plus Western Blotting Detection System. The membrane was reprobed with anti-β-actin antibody (Sigma, A5441; 1:5000) to provide a loading control.

Chromosome spreading was conducted as previously described (Wang et al., 2009). Oocytes were exposed to Tyrode's buffer (pH 2.5) to remove the zona pellucida, and then fixed in a drop of 1% paraformaldehyde with 0.15% Triton X-100 on a glass slide. Samples were labeled with CREST antibody (1:500) for 1 h to detect kinetochores, and chromosomes were counterstained with Hoechst 33342 (1:300). The laser scanning confocal microscope was used to examine chromosome numbers in oocytes.

Data are presented as mean±s.d. unless otherwise stated. Differences between two groups were analyzed by Student's t-test. Multiple comparisons between more than two groups were analyzed by one-way ANOVA using Prism 5.0 (GraphPad). P<0.05 was considered significant.