The subcortical maternal complex protein Nlrp4f is involved in cytoplasmic lattice formation and organelle distribution

Mammalian oocyte-to-embryo transition (OET) comprises numerous specific events, including maternal RNA clearance, organelle rearrangement, epigenetic reprogramming and zygotic genome activation in early embryonic development (Li et al., 2013; Lu et al., 2017). The low transcriptional activity before zygotic genome activation indicates that maternal-effect genes play essential roles in early embryonic development (Li et al., 2013, 2010). Since the initial identification of the maternal-effect genes Nlrp5 (also known as Mater) and Hsf1 (Christians et al., 2000; Tong et al., 2000), dozens of maternal-effect genes have been reported (Condic, 2016; Li et al., 2010). Yet, how the maternal-effect genes regulate early embryonic development remains poorly understood.

We first described the subcortical maternal complex (SCMC) as a 669-2000 kDa molecular weight complex composed of several proteins encoded by maternal-effect genes; these proteins include Ooep (also known as Floped), Nlrp5, Tle6, Khdc3 (also known as filia) and Zbed3, and plausibly also Nlrp2 and Padi6 (Li et al., 2008; Lu et al., 2017). Based on their direct interactions and roles in the stability of other SCMC proteins, the complex might include three core proteins (Ooep, Nlrp5 and Tle6) (Lu et al., 2017). Depletion of each SCMC core protein leads to decreased expression of other SCMC proteins and results in arrested development at the two-cell stage, suggesting an essential role of this complex in mouse early development and female fertility (Li et al., 2008; Tong et al., 2000; Yu et al., 2014). The SCMC is involved in multiple processes during the mouse OET, including cytoskeleton reorganization, organelle redistribution and cell division (Gao et al., 2018; Lu et al., 2017; Tashiro et al., 2010; Yu et al., 2014). The SCMC genes and this complex might be conserved in other mammals, including humans (Bebbere et al., 2014; Lu et al., 2017; Zhu et al., 2015). Recently, mutations of several human genes encoding the SCMC proteins NLRP5, TLE6, KHDC3L, NLRP2 and PADI6 were reported to be related to human reproductive disorders (Alazami et al., 2015; Docherty et al., 2015; Lu et al., 2017; Mu et al., 2019; Parry et al., 2011; Rezaei et al., 2016). However, details of how the SCMC regulates mammalian embryogenesis and fertility remain unclear.

Several decades ago, cytoplasmic lattices (CPLs), also known as cytoplasmic sheets, paracrystalline arrays, bilaminar lamellae or plaques, were reported to specifically persist in mammalian oocytes and preimplantation embryos (Capco and McGaughey, 1986; Hadek, 1966; Weakley, 1966; Zamboni, 1970). CPL fibers are absent in primary oocytes, gradually appear as oocytes grow, become abundant in fully grown oocytes and persist until the blastocyst stage (Gallicano et al., 1991; Wassarman and Josefowicz, 1978; Yurttas et al., 2008; Zamboni, 1970). CPLs undergo large spatial rearrangements during fertilization, embryo compaction and blastocyst formation (Capco and McGaughey, 1986; Gallicano et al., 1991). CPLs might be the specific structure of the cytoskeleton, including intermediate filaments or other unknown components (Capco et al., 1993; Gallicano et al., 1994). Padi6, also known as ePAD, is the first oocyte-specific protein identified to localize to CPLs in the mouse (Wright et al., 2003). Genetic depletion of Padi6 results in the absence of CPLs, arrested development at the two-cell stage and female infertility (Esposito et al., 2007). Padi6 is reported to associate with Nlrp5 or Ooep (two SCMC proteins), which have been shown to localize to CPLs in mouse oocytes (Kim et al., 2010; Tashiro et al., 2010). CPLs are absent in oocytes from Ooep and Nlrp5 null females (Kim et al., 2010; Tashiro et al., 2010). These results suggest a close relationship between the SCMC and CPLs.

NLRP subfamily proteins, containing an N-terminal pyrin domain (PYD), a NACHT nucleotide-binding domain and C-terminal leucine-rich repeats (LRRs), are well known for their roles in innate immunity, where they participate in the assembly of inflammasomes (Martinon et al., 2007; Ratsimandresy et al., 2013; Zambetti et al., 2012). These subfamily proteins also play important roles in female reproduction in mammals (Kuchmiy et al., 2016; Mahadevan et al., 2017; Monk et al., 2017; Murdoch et al., 2006; Tong et al., 2000). Nlrp4f was initially identified as an oocyte-specific gene similar to Nlrp5 in the mouse (Dadé et al., 2004). Later, Nlrp4f was identified as one of the downstream targets of Figla, an oocyte-specific transcription factor (Joshi et al., 2007). However, the physiological function of Nlrp4f remains unknown. Here, we characterized Nlrp4f as a novel component associated with the SCMC and CPLs. Genetic ablation of Nlrp4f showed that maternal Nlrp4f is required for mouse preimplantation development and female fertility. Furthermore, investigation of the relationship between the SCMC and CPLs using SCMC mutants indicated that the SCMC is involved in CPL formation.

CPLs are resistant to Triton X-100 extraction and are absent in Nlrp5 null oocytes (Gallicano et al., 1991; Kim et al., 2010). Using Nlrp5 null oocytes as a negative control, we extracted mouse oocytes with Triton X-100 to identify new components of CPLs (see Materials and Methods). The insoluble and soluble fractions were separated by SDS-PAGE and stained with Coomassie Brilliant Blue (Fig. 1A). Two protein bands (1 and 2) were found in the insoluble fraction of Nlrp5^+/+ and Nlrp5^+/− but not Nlrp5^−/− oocytes and were analyzed by tandem mass spectrometry (Fig. 1A). We identified Nlrp5 with 45 peptides as the primary protein in band 1, and Nlrp4f with 33 peptides in band 2 as a novel component associated with CPLs. To confirm the mass spectrometry results, we extracted normal mouse germinal vesicle (GV) stage oocytes with Triton X-100 and examined the insoluble fraction by western blot for Nlrp4f and other known CPL-associated proteins, including Nlrp5, Ooep and Zbed3. Similar to the other related proteins of CPLs, Nlrp4f was resistant to Triton X-100 extraction (Fig. 1B,C). These data suggest that Nlrp4f is a novel component associated with CPLs.

Using mouse anti-Tle6 antibody to precipitate the lysates of mouse oocytes, combined with mass spectrometry, we previously identified 23 potential SCMC proteins including Tle6, Nlrp5, Ooep, Khdc3 and Zbed3 (Gao et al., 2018), as well as Nlrp4f with 14 peptides (Fig. S1). Immunostaining showed that Nlrp4f colocalized with the SCMC protein Nlrp5 in the subcortex and cytoplasm of mouse oocytes, especially after extraction (Fig. 1D). Co-immunoprecipitation (Co-IP) showed that both Tle6 and Ooep antibodies specifically precipitated Nlrp4f, as well as Ooep and Tle6, in normal mouse ovarian lysates (Fig. 1E). Similar to the decreased level of SCMC proteins (Gao et al., 2018; Li et al., 2008; Yu et al., 2014), Nlrp4f was also decreased in the oocytes with disruption of SCMC core proteins Nlrp5, Tle6 or Ooep (Fig. 1F,G). These results identify Nlrp4f as a novel component of the SCMC in mouse oocytes.

Quantitative reverse transcription PCR (qRT-PCR) showed that Nlrp4f mRNA was highly expressed in mouse ovaries (Fig. 2A). Compared with fully grown oocytes, Nlrp4f mRNA persisted at high levels in one-cell stage embryos but decreased in two-cell stage embryos (Fig. 2B). Western blots showed that Nlrp4f was primarily detected in mouse ovaries, but not in other tissues (Fig. 2C). In mouse oocytes and preimplantation embryos, Nlrp4f protein remained at high levels until the morula stage (Fig. 2D). Immunofluorescence showed that Nlrp4f protein was localized in the cytoplasm of oocytes and preimplantation embryos, and concentrated at their subcortex (Fig. 2E).These data suggest that Nlrp4f is specifically expressed in oocytes and early embryos in the mouse.

To investigate the role of Nlrp4f, we generated Nlrp4f knockout mice by targeting its exon 2 using CRISPR/Cas9 technology (Fig. S2A). DNA sequencing showed that a male founder was obtained with a mutation at the Nlrp4f gene locus (Fig. S2A). This founder was used to establish the Nlrp4f mutant mouse line. Nlrp4f^−/− mice grew into adults and appeared mostly normal. The results of qRT-PCR revealed that Nlrp4f mRNA was absent in oocytes from Nlrp4f^−/− females (Fig. S2B). Western blots and immunostaining confirmed the successful ablation of Nlrp4f protein in oocytes from Nlrp4f^−/− females (Fig. S2C,D).

We then examined fertility by mating. Nlrp4f^−/− males exhibited normal fertility after mating for three months (Fig. 3A). However, compared with Nlrp4f^+/+ females, Nlrp4f^−/− female mice produced fewer pups (Fig. 3A). We then performed Hematoxylin and Eosin (H&E) staining for Nlrp4f^+/+ and Nlrp4f^−/− ovaries. The results showed that primordial, growing and mature follicles, as well as corpora lutea (CL), were present in both Nlrp4f^+/+ and Nlrp4f^−/− ovaries (Fig. 3B). We also examined the ratio of nonsurrounded nucleolus (NSN)-type and surrounded nucleolus (SN)-type fully grown oocytes from Nlrp4f^+/+ and Nlrp4f^−/− ovaries. Compared with controls, the ratio of NSN or SN oocytes from Nlrp4f^−/− females did not change significantly (Fig. S3A,B). Furthermore, similar numbers of metaphase II (MII) oocytes were recovered from Nlrp4f^+/+ and Nlrp4f^−/− females after superovulation with gonadotrophins (Fig. S4A,B). This suggests that Nlrp4f plays a role in female fertility, but is not essential for mouse ovulation or male fertility.

Next, we isolated preimplantation embryos from Nlrp4f^+/+ and Nlrp4f^−/− females after mating with normal males. Similar numbers of one- and two-cell stage embryos were recovered from Nlrp4f^+/+ and Nlrp4f^−/− females at embryonic day (E)0.5 and E1.5 (Fig. 3C,D). The loss of maternal Nlrp4f had mininal effect on the total numbers of embryos at E2.5 and E3.5 (Fig. 3D). However, the majority of embryos from Nlrp4f^−/− females displayed a delay in embryo compaction and blastocyst cavitation (Fig. 3C,D; Fig. S5). We also isolated two-cell embryos from normal and Nlrp4f^−/− females after treatment with gonadotrophins and cultured these embryos in vitro for 60 h. The cultured embryos also displayed similar developmental delays (Fig. 3E). Thus, Nlrp4f is required for mouse preimplantation development.

The distribution of organelles such as endoplasmic reticulum (ER) and mitochondria is linked to early embryo and female fertility (Fernandes et al., 2012; Gao et al., 2018). Thus, we examined organelle distribution in zygotes from Nlrp4f^+/+ and Nlrp4f^−/− females using ER-Tracker and MitoTracker to label the ER and mitochondria, respectively. In control zygotes from Nlrp4f^+/+ females, the ER and mitochondria were concentrated around the male and female pronucleus during interphase and aggregated around the mitotic spindle following nuclear envelope break down (NEBD) (Fig. 4A). However, in zygotes from Nlrp4f^−/− females, the ER and mitochondria were concentrated around male and female pronucleus, and also aggregated in the subcortical region (Fig. 4A). Live imaging using MitoTracker confirmed that depletion of Nlrp4f led to the disorganized distribution of mitochondria in the zygote (Fig. 4B; Movies 1 and 2).

We also examined the distribution of ER and mitochondria in oocytes from Nlrp4f^+/+ and Nlrp4f^−/− females. Here, the organelles radially distributed around the nucleus at GV, around the spindles after germinal vesicle break down (GVBD), and also around the spindles at MII oocytes (Fig. 4C). However, in Nlrp4f^−/− oocytes, they localized around both the nucleus and cortical region at the GV stage, became concentrated in the subcortical region after GVBD and were dispersed throughout the cytoplasm at the MII stage (Fig. 4C). Live imaging confirmed that depletion of Nlrp4f resulted in defective mitochondrial distribution during mouse oocyte maturation (Fig. 4D; Movies 3 and 4). These data suggest that Nlrp4f is involved in organelle distribution or redistribution during mouse oocyte maturation and zygote development.

Microtubules (MT) and acetylated tubulin are involved in organelle redistribution during mouse oocyte maturation (Gao et al., 2018; Kan et al., 2011). Thus, we examined the expression of acetylated tubulin in Nlrp4f^−/− oocytes using western blotting. This revealed a decrease in acetylated α-tubulin in Nlrp4f^−/− GV and MII oocytes (Fig. 5A). Treatment with the deacetylase inhibitor trichostatin A (TSA) significantly increased the level of acetylated α-tubulin in oocytes from both Nlrp4f^+/+ and Nlrp4f^−/− females (Fig. 5B; Fig. S6A). However, TSA treatment did not rescue the disorganized distribution of organelles in Nlrp4f^−/− oocytes (Fig. 5C). Immunostaining showed that spindle MTs formed in both Nlrp4f^−/− and control oocytes after GVBD, but the astral-like MTs were longer in Nlrp4f^−/− oocytes than those in Nlrp4f^+/+ oocytes following GVBD (Fig. 5D). However, nocodazole treatment did not dramatically affect the subcortical distribution of organelles in 64% (30/47) of Nlrp4f^−/− oocytes after GVBD (Fig. 5E; Fig. S6B), suggesting that the lengthened microtubules are not the predominant cause of organelle localization to the subcortex in Nlrp4f^−/− oocytes.

As CPLs might also be associated with organelle distribution in oocytes (Gao et al., 2018; Kan et al., 2011), we investigated CPLs in the ovaries of Nlrp4f^+/+ and Nlrp4f^−/− females using transmission electron microscopy (TEM). As expected, CPLs were abundant in control oocytes from Nlrp4f^+/+ ovaries (Fig. 5F). However, CPLs were largely absent in oocytes from Nlrp4f^−/− ovaries (Fig. 5F), suggesting that Nlrp4f is required for CPL formation.

Although the SCMC core protein Nlrp5, Tle6 or Ooep was necessary for the stability of Nlrp4f (Fig. 1F,G), depletion of Nlrp4f did not affect expression levels of SCMC proteins, including Nlrp5, Tle6, Ooep and Zbed3 (Fig. S7A,B). Furthermore, depletion of Nlrp4f did not affect the interactions of these SCMC proteins (Fig. S7C). These results suggest that Nlrp4f is not a core component of the SCMC.

The absence of CPL formation in Nlrp4f null oocytes is reminiscent of oocytes depleted of Zbed3, which is also not a core component of the SCMC (Fig. 5F) (Gao et al., 2018). These results suggest that integrity of the SCMC or any of the SCMC proteins is required for the formation of CPLs. To test this, we analyzed CPLs in oocytes depleted of other SCMC proteins. We focused on Tle6, a core SCMC protein, and on Khdc3, another non-core SCMC protein. Our results revealed that the CPLs were largely absent in Tle6^−/− and Khdc3^−/− oocytes (Fig. 6A,B). We also observed the aberrant distribution of organelles in Khdc3^−/− oocytes (Fig. S8). Considering the absence of CPLs and the aberrant distribution of organelles in the oocytes with disrupted Nlrp4f, Khdc3 or Tle6, as well as Zbed3, Nlrp5, Ooep or Padi6 (Fig. 4, Fig. 5F; Fig. S8) (Gao et al., 2018; Kan et al., 2011; Kim et al., 2010; Tashiro et al., 2010; Yurttas et al., 2008), we propose that the SCMC is required for CPL formation and organelle distribution in mouse oocytes.

Although CPLs were absent in Nlrp4f and Zbed3 null oocytes, the SCMC was not affected in these oocytes (Fig. 5F; Fig. S7) (Gao et al., 2018), suggesting that the SCMC, or at least its core complex, is the upstream regulator of CPL formation. To test this, we examined the expression of SCMC proteins in normal ovaries at postnatal day (P)1, P2, P3, P5, P10, P17 and P21. Although mRNA levels of the SCMC genes Nlrp4f, Ooep, Nlrp5, Tle6 and Zbed3 were highly expressed in the ovary when mice were born (Joshi et al., 2007; Li et al., 2008), these SCMC proteins were at low levels in the oocytes of normal mouse ovaries at this stage, as shown by immunostaining and western blotting (Fig. 6C,D). Importantly, these SCMC proteins were specifically co-precipitated by Tle6 antibody in normal ovarian lysates after P3 (Fig. 6E). CPLs were sporadically observed in the oocytes of ovaries at P5-P10, but not observed in oocytes of ovaries at P3 (Wassarman and Josefowicz, 1978; Yurttas et al., 2008; Zamboni, 1970). Thus, our results suggest that the SCMC forms prior to CPLs in the oocytes of mouse ovaries, further supporting the idea that the SCMC is involved in CPL formation.

CPLs were first discovered as an abundant structure in mammalian oocytes and early embryos (Gallicano et al., 1991; Hadek, 1966; Weakley, 1966, 1968; Zamboni, 1970). However, the roles and compositions of these structures remain mysterious. Recently, Padi6 was identified as an oocyte- and embryo-specific protein localized in the CPLs of mouse oocytes (Wright et al., 2003). Depletion of Padi6 resulted in the absence of CPLs and arrested development at the two-cell stage, suggesting that CPLs play a role in development and female fertility (Esposito et al., 2007). More recently, the SCMC core protein Nlrp5 or Ooep was reported to localize to CPLs and be required for CPL formation and two-cell embryo development (Li et al., 2008; Tashiro et al., 2010; Tong et al., 2000). However, disruption of proteins Zbed3, Nlrp4f or Khdc3 also resulted in the loss of CPLs, but the phenotypes were less severe for preimplantation development and female fecundity (Fig. 3, Fig. 5F) (Gao et al., 2018; Zheng and Dean, 2009). These results suggest that disruption of CPL formation is not the major cause of arrested development at the two-cell stage and infertility in Ooep, Nlrp5, Tle6 and Padi6 null females. Consistent with this, a recent report shows that CPL formation may not be related to arrested two-cell development in the mouse (Longo et al., 2018).

CPLs occupy large spaces of cytoplasm in mouse oocytes and early embryos (Capco and McGaughey, 1986; Gallicano et al., 1991; Zamboni, 1970). Some CPLs have been observed to surround the organelles in mouse oocytes (Gao et al., 2018; Kan et al., 2011). These results suggest that CPLs play a role in maintaining the localization of organelles in mouse oocytes. Consistent with this, the absence of CPLs is related to defects in organelle distribution in oocytes with disruption of SCMC proteins Nlrp4f, Zbed3, Khdc3, Nlrp5, Ooep, Tle6 or Padi6 (Fig. 4) (Gao et al., 2018; Kan et al., 2011; Kim et al., 2010; Tashiro et al., 2010; Yurttas et al., 2008). Furthermore, CPLs contain intermediate filaments in mammalian oocytes and early embryos (Capco et al., 1993; Gallicano et al., 1994). The intermediate filaments directly or indirectly interact with other two cytoskeletons comprising microfilaments and microtubules, and regulate the dynamics of cytoskeletons in various cell types (Huber et al., 2015). Thus, CPLs might regulate microfilaments and microtubules in mouse oocytes and early embryos. Consistently, the absence of CPLs is related to the disorganized formation of microfilaments and microtubules in oocytes and embryos with disrupted SCMC genes (Fig. 5) (Gao et al., 2018; Kan et al., 2011; Yu et al., 2014). Thus, CPLs might control organelle redistribution through the cytoskeleton in mouse oocytes and early embryos.

Three SCMC-related proteins, Padi6, Ooep and Nlrp5, are localized to the CPLs of mouse oocytes (Tashiro et al., 2010; Wright et al., 2003). Disruption of all examined SCMC proteins, including Padi6, Nlrp5, Ooep, Tle6, Khdc3, Zbed3 and Nlrp4f, leads to loss of CPLs in mouse oocytes (Fig. 5F, Fig. 6A,B) (Esposito et al., 2007; Gao et al., 2018; Kim et al., 2010; Tashiro et al., 2010). These results suggest that the SCMC is closely related to CPLs or that they are the same structures in mouse oocytes (Kim et al., 2010; Tashiro et al., 2010). However, depletion of Zbed3 and Nlrp4f does not affect the stability of other SCMC proteins (including Nlrp5, Ooep and Tle6) nor the interactions of these proteins (Fig. S7) (Gao et al., 2018), suggesting that the SCMC, or at least the SCMC core complex, is distinct from the CPLs and involved in CPL formation. Consistent with this, SCMC proteins are expressed and interact before CPLs form during mouse oogenesis (Fig. 6C-E) (Wassarman and Josefowicz, 1978). How the SCMC regulates CPL formation remains unknown. One possibility is that all SCMC proteins control CPL formation via a common regulatory factor or pathway. Consistent with this, CPLs are preserved by glutaraldehyde but not by fixatives such as osmium tetroxide, formaldehyde and other short chain aldehydes (Weakley, 1966, 1968).

The molecular mass of the SCMC is larger than the total mass of the first four proteins identified (Ooep, Nlrp5, Tle6 and Khdc3), suggesting additional proteins in this complex (Li et al., 2008, 2013; Lu et al., 2017). Padi6 might also be involved in the SCMC (Kim et al., 2010; Li et al., 2008). Recently, another maternal-effect gene, Nlrp2, has been reported to be associated with the SCMC, based on the in vitro interactions of Nlrp2 and SCMC proteins, their localization and similar phenotypes (Mahadevan et al., 2017). We have recently characterized Zbed3 as a novel component of the SCMC in mouse oocytes and early embryos (Gao et al., 2018). The SCMC may be conserved in other mammals, including humans (Bebbere et al., 2014; Lu et al., 2017; Zhu et al., 2015). These studies have prompted us to propose that the SCMC is a functional module in the mammalian OET (Lu et al., 2017). Here, we have identified Nlrp4f as a novel protein of the SCMC, further supporting the notion of a functional module of the SCMC. The first NLRP subfamily protein identified in the SCMC was Nlrp5 (Li et al., 2008; Tong et al., 2000). Recently, Nlrp2 was reported to be associated with the SCMC (Kuchmiy et al., 2016; Mahadevan et al., 2017). Similar to the expression patterns of SCMC genes, the transcripts of many NLRP subfamily members, such as Nlrp14, Nlrp4a-4g and Nlrp9a-9c, are specifically expressed in ovarian tissue, oocytes and preimplantation embryos (Hamatani et al., 2004). Thus, some other NLRP proteins might also function in oocytes and early embryos through the SCMC.

In summary, we have characterized Nlrp4f as a novel component of the SCMC in the mouse. The absence of CPLs is related to the disorganized distribution of organelles in oocytes in response to depletion of any SCMC genes, suggesting that the SCMC functions in regulating organelle distribution, probably by the CPLs in mouse oocytes and female fertility. Accumulating evidence suggests that mutations of human SCMC genes are related to female reproductive disorders, including recurrent embryonic loss, molar pregnancies and imprinting disorders (Amoushahi et al., 2019; Begemann et al., 2018; Lu et al., 2017; Mu et al., 2019; Nguyen et al., 2018; Wang et al., 2018). Further studies of phenotypes in Nlrp4f and other SCMC gene mutants and exploration of the relationships between the SCMC and its proteins, CPLs and oocyte quality will contribute to an understanding of human OET and the pathogenesis of reproductive diseases.

All animal maintenance and manipulations were performed according to the guidelines of the Animal Care and Use Committee of the Institute of Zoology, Chinese Academy of Sciences.

Normal and Nlrp4f^−/− females aged 6-8 weeks were stimulated with 5 IU pregnant mare serum gonadotrophin (PMSG). After 46-48 h, oocytes were recovered by scraping the surface of the ovaries with a 26-gauge needle in M2 medium. MII oocytes were collected at 13-15 h after additional stimulation with 5 IU human chorionic gonadotrophin (hCG). Zygotes, two-cell embryos, morulae and blastocysts from normal and Nlrp4f^−/− mice were collected at 24, 48, 72 and 96 h after hCG stimulation, respectively.

For in vitro culture, two-cell embryos were collected from Nlrp4f^+/+ and Nlrp4f^−/− females mated with normal fertile males. These two-cell embryos were cultured in KSOM medium (Millipore, MR-121-D) at 37°C in 5% CO₂. The development of embryos was examined every 4 h and the developmental rates of the embryos were calculated.

Equal numbers of oocytes (∼500 GV oocytes) were collected from Nlrp5^+/+, Nlrp5^+/− and Nlrp5^−/− mice and extracted with buffer containing 100 mM NaCl, 3 mM MgCl₂, 300 mM sucrose, 10 mM PIPES (pH 6.8), 0.5% Triton X-100 and 1× Complete Protease Inhibitor Cocktail (Roche) for 30 min at room temperature. After extraction, the oocytes were manually picked up as the insoluble fraction. The remains contained the soluble components of oocytes in the buffer. These samples were separated by SDS-PAGE and stained with Coomassie Brilliant Blue (SimplyBlu SafeStain, Invitrogen, LC6060) according to the manufacturer's protocol. The specific bands in Nlrp5^+/+ (wild type) and Nlrp5^+/−, but not Nlrp5^−/−, oocytes were collected for mass spectrometry (NanoLC-LTQ-Orbitrap XL, Thermo Finnigan).

As previously reported, the T7 promoter was added to the Cas9 encoding region using pX330 as the template (Yang et al., 2013). Cas9 PCR products were transcribed in vitro using mMESSAGE mMACHINE T7 ULTRA Kit (Ambion, AM1345). The Nlrp4f single guide RNAs (sgRNAs) were designed using the online CRISPR design tool (http://crispr.mit.edu/) and cloned into the pUC57-sgRNA expression vector. The sgRNA PCR templates were amplified for in vitro transcription with MAXIscript SP6/T7 In Vitro Transcription Kit, according to the manufacturer's protocol (Ambion, AM1322). Then, 50 ng/µl of Cas9 mRNA and 25 ng/µl of sgRNAs were mixed and injected into the cytoplasm of CD1 mouse zygotes. The surviving zygotes were transplanted into the oviducts of pseudopregnant females to produce the offspring. The offspring were genotyped by PCR with specific primers for Nlrp4f: forward primer, 5′-CTGAGGTCCCAGCTTGTGTC-3′; reverse primer, 5′-TTGGCCCTATGGTAGATGCG-3′.

For assessing the fertility of male mice, two-month-old Nlrp4f^+/+ females were mated with Nlrp4f^+/+ and Nlrp4f^−/− male mice by 2:1 co-caging. To test the fertility of female mice, two-month-old Nlrp4f^+/+ and Nlrp4f^−/− female mice were mated with normal fertile males by 2:1 co-caging. The number of pups per litter was recorded. The results were calculated after they had mated for three consecutive months.

Rabbit anti-Nlrp4f antibody was generated with a standard protocol from Abmart (Shanghai, China). In brief, two peptides selected from Nlrp4f were expressed in Escherichia coli Rosetta and purified with Ni²⁺ affinity columns. The purified peptides were mixed and used to immunize rabbits four times. The serum (antibody) from the immunized rabbits was purified with Protein A. The sequence of Nlrp4f peptide 1 was: LSNCSLSEQCWDYLSEVLRQNKTLSHLDISSNDLKDEGLKILCRSLILPYCVLESLCLSCCGITERGCQDLAEVLKNNQNLKYLHVSYNKLKDTGVMLLCDAIKHPNCHLKDLQLEACEITDASNEELCYAFMQCETLQTLNLMGNAFEV and that of Nlrp4f peptide 2 was: SKNIHHKLYQCLETLSGNAELQEQIDGMRLFSCLFEMEDEAFLVKAMNCMQQINFVAKNYSDFIVAAYCLKHCSTLKKLSFSTENVLNEGDQSYMEELLICWNNMCSVFVRSKDIQELRIKDTNFNEPAIRVLYESLKYPSFTLNKLVAN. Primary and secondary antibodies were used as in Table S1.

The ovaries from two-month-old females at different ages were fixed in Bouin solution overnight at room temperature for H&E staining or fixed with 4% paraformaldehyde (PFA) in PBS overnight at 4°C for immunofluorescence staining, then embedded in paraffin and sectioned. The sections were dewaxed, rehydrated and stained with H&E. For immunostaining, the rehydrated sections were treated for antigen retrieval and washed three times with PBS. The sections were blocked with 5% BSA for 1 h at room temperature and incubated with primary antibodies (Table S1) overnight at 4°C. After washing three times with PBS, the sections were incubated with secondary antibodies (Table S1) and Hochest 33342 for 1 h at room temperature and then mounted for imaging.

For whole-mount staining of oocytes and early embryos, the isolated oocytes and embryos were fixed with 4% PFA containing 0.5% Triton X-100 for 30 min at room temperature. The immunostaining was performed with primary and secondary antibodies as described above.

For classification of NSN-type and SN-type fully grown oocytes, the oocytes were fixed with 4% PFA, permeabilized with 0.5% Triton X-100, blocked with 5% normal donkey serum and the DNA stained with Hochest 33342. According to their chromatin configuration in the nucleus, oocytes were classified into three groups: nonsurrounded nucleolus (NSN) type with diffused chromatin, surrounded nucleolus (SN) type with a typical ring chromatin and middle type with diffused and half-ring chromatin. The images were acquired with a LSM780 (Zeiss).

For immunoblot, different tissues were collected from CD1 mice. Total proteins were extracted using RIPA lysis buffer containing 50 mM Tris-HCl (pH 7.5), 150 mM NaCl, 1% sodium deoxycholate, 1% Triton X-100, 0.1% SDS, 5 mM EDTA, 1 mM Na₃VO₄, 5-10 mM NaF and complete EDTA-free Protease Inhibitor Cocktail (Roche). The lysate was incubated on ice for 15 min and centrifuged at 12,000 rpm (13,800 g) for 15 min at 4°C. The supernatant was quantified using a BCA reagent kit (Beyotime, P0012-1). For oocyte samples, equal numbers of oocytes (50-200) were collected. Then, proteins were separated by 8-15% SDS-PAGE and transferred onto a PVDF membrane. The membrane was pretreated with 5% defatted milk for 1 h at room temperature and incubated with primary antibodies (Table S1) overnight at 4°C. After washing three times with PBS containing Tween-20, the membrane was incubated with horseradish peroxidase (HRP)-conjugated secondary antibodies (Table S1) for 1 h at room temperature. Signals were developed with SuperSignal West Pico Chemiluminescent Substrate (Thermo Fisher Scientific, 34080) and analyzed with Quantity One software (Bio-Rad Laboratories).

For co-immunoprecipitation, ovaries were isolated from CD1 female mice of different ages. Total proteins were extracted using immunoprecipitation (IP) lysis buffer containing 25 mM Tris-HCl (pH 7.5), 150 mM NaCl, 1 mM EDTA, 1% NP-40, 5% glycerol, 1 mM Na₃VO₄, 2 mM NaF and complete EDTA-free Protease Inhibitor Cocktail (Roche). BSA (final concentration 1%) was also added to samples of GV oocytes to protect proteins from degradation. After centrifugation [12,000 rpm (13,800 g) for 15 min at 4°C], the lysates were incubated with mouse anti-Tle6 antibody or normal mouse IgG at 4°C for 3-4 h before Protein A/G magnetic beads (Selleck, B23202) were added. After an additional incubation for 3 h at 4°C, beads were washed five times with IP lysis buffer and eluted with 1× SDS loading buffer. The samples were separated by SDS-PAGE, and western blot performed using specific antibodies.

According to the manufacturer's protocols, mRNAs of different tissues were extracted using RNAzol reagent (Molecular Research Center, RN190) and mRNA of mouse eggs or embryos was isolated using Dynabeads mRNA DIRECT Micro Kit (Life Technologies, 61021). Complementary DNA (cDNA) was synthesized using PrimeScript RT Reagent Kit (Takara, RR037A). Quantitative RT-PCR was conducted with EvaGreen 2× qPCR MasterMix (Applied Biological Materials, MasterMix-S). The expression of target genes was normalized with Gapdh. Nlrp4f forward primer, 5′-TCATCCAACACTTGCTCCAGC-3′; Nlrp4f reverse primer, 5′-AAAGATGCCATCTTGTCTTCAGG-3′; Gapdh forward primer, 5'-CCCCAATGTGTCCGTCGTG-3'; and Gapdh reverse primer, 5'-TGCCTGCTTCACCACCTTCT-3'.

Before live-imaging, mouse oocytes and zygotes were cultured in M2 and KSOM medium, respectively, with MitoTracker Red CMXRos (1:10,000; Invitrogen, M7512) and ER-Tracker Blue-White DPX (1:5000; Invitrogen, E12353) at 37°C in 5% CO₂ for 60 min. The pictures of live imaging were captured by UltraVIEW-VoX (Perkin Elmer).

For drug treatment, Nlrp4f^+/+ and Nlrp4f^−/− GV oocytes were cultured in M2 medium supplemented with 3-isobutyl-1-methylxanthine (IBMX, 100 µg/ ml; Sigma-Aldrich, I7018) and with DMSO, TSA (2 µg/ ml for 12 h; Beyotime, S1893) or nocodazole (2.5 µg/ ml for 2 h at 37°C in 5% CO₂; Sigma-Aldrich, M1404). These arrested oocytes were washed three times with M2 medium to release IBMX inhibition, cultured to GVBD in M2 medium with DMSO, TSA or nocodazole, and labeled with MitoTracker Red CMXRos.

TEM was performed as previously described (Gao et al., 2018). Briefly, the ovaries were isolated from two-month-old females and fixed in buffer containing 2.5% glutaraldehyde, 4% PFA and 0.1 M Na-cacodylate for 24 h at 4°C. After gradient dehydration, the samples were embedded in LX112 resin and cut into ultrathin sections. Oocytes from the ovarian sections were imaged using TEM (Jeol, JEM-1230).

All experiments were performed at least three times with different samples. All statistics were analyzed by the Student's t-test in GraphPad Prism5 software and presented as mean±s.e.m. *P<0.05 was considered significant.

We thank Lijuan Wang, Shiwen Li, Xili Zhu and Hua Qin (Institute of Zoology Chinese Academy of Science) for technical assistance, and Dr Jurrien Dean at the National Institutes of Health in the USA, Dr Marcel van Duin at the Ferring Pharmaceuticals and Dr Nicolas Plachta at Agency for Science, Technology and Research (Singapore) for their critical review of the manuscript.