Impaired oocyte maturation and ovulation in membrane progestin receptor (mPR) knockouts in zebrafish

Abstract

Accumulating evidence suggest that membrane progestin receptor α (mPRα) is the membrane receptor mediating nongenomic progestin signaling that induces oocyte maturation in teleost. However, the involvement of other members of mPR family in oocyte maturation is still unclear. In this study, we found impaired oocyte maturation in zebrafish lacking mPRα1, mPRα2, mPRβ, or mPRγ2. In contrast, no difference was observed in oocyte maturation in the single knockout of mPRγ1, mPRδ, or mPRε. To study possible redundant functions of different mPRs in oocyte maturation, we generated a zebrafish line lacking all seven kinds of mPRs (mprs^−/−). We found oocyte maturation was further impaired in mprs^−/−. In addition, oocyte ovulation delay was observed in mprs^−/− females, which was associated with low levels of nuclear progestin receptor (Pgr), a key regulator for ovulation. We also found reduced fertility in mprs^−/− female zebrafish. Furthermore, eggs spawned by mprs^−/− females were of poor quality.

Introduction

Meiosis resumption, i.e. final oocyte maturation is trigged by binding and signaling of progestin to its membrane receptor located at surface of oocytes in fish and amphibians (Nagahama and Yamashita, 2008; Thomas, 2012; Zhu et al., 2008). Membrane progestin receptor α (mPRα) has been suggested to mediate this process (Thomas, 2008; Zhu et al., 2008). The mPRα is a G protein-coupled receptor (GPCR)-like protein that has a high affinity for progestin and is capable of signaling like GPCRs (Pace and Thomas, 2005; Thomas et al., 2007; Zhu et al., 2003a; Zhu et al., 2003b). The mPRα belongs to a progestin and adipoQ receptor superfamily (PAQR) (Tang et al., 2005; Thomas, 2012; Thomas et al., 2007). Five mPR paralogues (mPRα (Paqr7), mPRβ (Paqr8), mPRγ (Paqr5), mPRδ (Paqr6), and mPRε (Paqr9)) have been identified in this family in all vertebrates. Zebrafish has two additional mPR paralogs (mPRα2 (Paqr7b), mPRγ2 (Paqr5b)), likely due to teleost specific genome duplication (Glasauer and Neuhauss, 2014). Studies conducted so far were focused on the roles and signaling of mPRα or mPRα2. These studies suggest mPRα mediates rapid nongenomic signaling of progestin to induce oocyte maturation (Aizen et al., 2018; Hanna and Zhu, 2011; Thomas, 2008; Zhu et al., 2008). However, the involvement of other members of mPRs are still unclear.

In addition, progestins also signal through nuclear progestin receptor (Pgr) and progesterone receptor membrane components (Pgrmcs) regulating physiological processes such as oogenesis and ovulation (Kubota et al., 2016; Lydon et al., 1995; Wu et al., 2019; Wu and Zhu, 2019; Wu et al., 2018; Zhu et al., 2015). Limited studies have suggested these progestin receptors also regulate physiological processes by interacting and regulating each other (Aizen et al., 2018; Thomas et al., 2014; Wu and Zhu, 2019). Previous studies suggested Pgrmc1 regulates mPRα expression and transport mPRα to cell surface in oocytes (Aizen et al., 2018; Wu et al., 2018). In addition, activation of mPRs leads to transactivation of PGR-B in human myometrium cells (Karteris et al., 2006). These results suggest cross talks between different families of progestin receptors.

This study attempts to fill a knowledge gap by generating mPR mutants for all known mPR paralogues to investigate their roles in vivo. We also generated a total mutant line that lacking all seven mPR paralogues in zebrafish, and found these mPRs are critical for oocyte maturation, ovulation and fertility. Furthermore, we observed lacking mPRs also lead to poor-quality eggs in zebrafish.

Materials and Methods

Animals

The zebrafish (Danio rerio) strain used in this investigation, the Tübingen strain, was initially obtained from the Zebrafish International Resource Center, then propagated in our lab at East Carolina University following previously published guidelines (see (Zhu et al., 2015). All the animal care and use protocols were approved by Institutional Animal Care and Use Committee (IACUC) at East Carolina University.

TALEN assembly and in vitro synthesis of TALEN mRNAs

We designed and assembled TALEN molecules using the unit assembly method detailed in Huang et al. (Huang et al., 2011). Using mPRα2 (Paqr7b) as an example (Supplemental Fig. 2), candidate TALEN target sites were identified using the following parameters: (1) nucleotide T was at position 0; (2) length of the spacer and nucleotides that bound to forward or reverse TALEN proteins were between 16 and 22 bp; and (3) a restriction endonuclease site was identified near the center of the spacer for convenient mutation detection and mutation rate estimation. The target was selected near the beginning of the coding sequence (Supplemental Fig. 2; forward target site sequence: ACTGTGATGGTGAGCG, reverse target site sequence: TATGTGTGACTCACGG, and spacer sequence with a NcoI restriction enzyme site was in bold). All assembled TALEN vectors were confirmed using Sanger sequencing.

These assembled TALEN vectors were linearized with Not I, gel extracted, and purified using the QIAquick gel extraction kit according to manufacturer’s specifications (Qiagen, Germantown, MD), and mRNAs were transcribed using SP6 mMACHINE kit (Ambion, Austin, TX). The transcribed mRNAs were stored at −80°C until use. Immediately prior to microinjection, mRNA was diluted into workable concentrations (100 ng/μl) with nuclease-free water, and mixed with an equal volume of 0.5% phenol red solution (Sigma P0290, Sigma-Aldrich, St. Louis, MO).

CRISPR/Cas9 target design and preparation of Cas9 RNA and sgRNAs

Using mPRα1 (Paqr7a) as an example, we identified 5’GG-(N₁₈)-NGG3’ target sequences in exon 1 of mPRα1 (GGCTCTGGTATAGTAGTTACGGG) (Supplemental Fig. 1). Syntheses of Cas9 RNA and single guide RNAs (sgRNAs) were based on a protocol from Chen’s lab (Jao et al., 2013). For Cas9 transcripts (nls-zCas9-nls RNA), a template plasmid (pCS2-nls-zCas9-nls) was linearized by NotI digestion, then purified using a QIAprep column (Qiagen, Germantown, MD). Capped Cas9 mRNA (nls-zCas9-nls) was synthesized using mMESSAGE mMACHINE SP6 kit (Fisher Scientific, Hampton, NH) and purified using RNeasy Mini kit (Qiagen). For sgRNAs, template plasmids were linearized by BamHI digestion and purified using a QIAprep column. The gRNAs were generated by in vitro transcription using MEGAshortscript T7 kit (Fisher Scientific). The size and quality of the resulting gRNA was confirmed by electrophoresis using a 2% (wt/vol) formaldehyde agarose gel.

Establishment of zebrafish mutant lines (mprs^−/−)

Screening and generation of mutant zebrafish lines followed protocols established previously (Zhu et al., 2015). To generate a founder population (F0), fertilized eggs were collected within 5 minutes of natural spawning between wildtype fish from their crossing tanks, which were set up the previous night. Microinjection was performed on newly fertilized zebrafish embryos at either the one or two-cell stages. Approximately 1 nl of 100 ng/μl of sgRNA and 150 ng/μl of Cas9 mRNA (or 1 nl of 100 ng/ul of TALEN transcripts) were co-injected into the embryos using a glass microcapillary pipette attached to a micromanipulator under a stereomicroscope (Leica MZ6, Wetzlar , Germany). The injection was driven by compressed N2 gas, under the control of a PV820 Pneumatic PicPump (World Precision Instrument, Sarasota, FL). For comparison, and to estimate mutagenesis efficiency, embryos without microinjection were designated as wildtype and used as controls. A pool of genomic DNA was extracted from 30 well-developed wildtype or CRISPR/Cas9-gRNA-microinjected (TALEN microinjected) embryos two days post-fertilization (dpf) using a HotSHOT method (Meeker et al., 2007). Mutation rates were estimated by comparing band intensities of digested PCR products to intensities of undigested PCR products using T7 endonucleases I assay. The PCR products were cleaned through a Qiagen column, cloned into a TA cloning vector (Chen et al., 2009), and potential mutant clones were selected for DNA sequencing analysis to confirm the presence of a frame-shifting mutation (Zhu et al., 2015).

To identify germline-transmitted mutations, remaining F0 founder embryos were raised to adulthood and outcrossed with wildtype fish. Genomic DNA from each cross was extracted from 30 randomly selected individuals, and they were pooled as F1 embryos, and the status of the target site was analyzed via PCR amplification, T7 Endonucleases I assay, and DNA sequencing as described above. The remaining F1 embryos with identified frame shifted mutations were raised to adulthood and were genotyped individually. Genomic DNA was extracted from part of the caudal fin of adult fish in a 50 μl hot alkaline solution and analyzed as above. Heterozygous F1 adults carrying the same frameshift mutant allele were crossed with each other, which yielded wildtype, heterozygous, and homozygous F2 fish that were further characterized genetically and physiologically. The total knockout for mPRs was obtained by crossing different mutant lines.

Germinal vesicle breakdown (GVBD) assay

To determine the sensitivities of oocytes to a maturation inducing steroid, 17α,20β-dihydroxyprogesterone (DHP), oocytes were isolated and incubated with DHP (Hanna and Zhu, 2011; Pang and Thomas, 2009). Gravid female zebrafish were euthanized humanely, following which the ovaries were dissected out then washed several times in 60% Leibovitz L-15 medium (Sigma-Aldrich), ensuring that the individual oocytes were carefully prepared without damaging the surrounding follicular cell layers per previously established protocols (Hanna and Zhu, 2011; Liu et al., 2017). Fully-grown immature oocytes of the same size (diameter of 550–650 μm) were selected and randomly distributed into the wells of a 24-well plate (~30 oocytes/1 ml medium/well), then treated with DHP dissolved in ethanol. Final concentrations of DHP and ethanol in the incubation medium are 5 nM and 0.1%, respectively. As controls, pure ethanol was added into control wells containing a same number of oocytes collected at the same time from the same group of individual fish, and the rates of GVBD were also recorded. Incubation continued for up to 5 hours, with GVBDs being scored each half hour during the incubation period. All experiments were repeated five times to confirm the results.

Separation of oocytes according to their size and developmental stages

Oocyte maturation (~6:00) and ovulation (~8:00) in zebrafish typically occur prior to the onset of (day) light, while spawning occurs within 1 h following the onset of light. Adult females (n = 7) from each mutant genotype were euthanized at 9:30 am, thirty minutes after laboratory lights were turned on, by placing each fish in a lethal dose of MS-222 (300 mg/L buffered solution) for 10 minutes, then severing the spinal cord and blood supply using IACUC approved procedures. Oocyte maturation and ovulation in zebrafish typically occurs sequentially prior to the onset of lights, while spawning occur within 1 hr following the onset of lights (Liu et al., 2018). The ovaries of each fish were then immediately dissected out and rinsed in 60% L-15 media (Sigma-Aldrich) containing 15mM HEPES (pH=7.2). Oocytes of various sizes were isolated from the ovaries using fine forceps. The diameter of each oocyte was measured with a stereo microscope (SZX7, Olympus, Japan), and recorded. The developmental stages of oocytes were divided into five stages based on morphological criteria (Selman et al., 1993) with slight modification: stage I (<140 μm) and II (140–340 μm) previtellogenic oocytes; stage III early vitellogenic oocytes (340–690 μm); stage IV late vitellogenic oocytes (690–730 μm) that are further divided into IVa and IVb two stages, IVa is maturational competent fully grown immature oocytes (IVa), IVb is matured oocytes that underwent oocyte maturation but haven’t yet gone through ovulation (IVb); and stage V ovulated oocytes (730–750 μm), i.e. ovulated eggs with no follicular cells attached.

RNA isolation and real-time quantitative PCR

Based on the size standard mentioned above, total RNA was isolated from stage III (21:00), stage IVa (6:00), and stage IVb (08:00) oocytes using the RNAzol reagent (Molecular Research Center, Cincinnati, OH) according to a modified protocol (Liu et al., 2017), and reverse transcribed to cDNA using High Capacity cDNA Reverse Transcription Kit (Thermo Fisher). The amount and purity of the RNA was determined using a Nanodrop 2000 (Thermo Fisher). Real-time quantitative PCR (qPCR) was performed using the SYBR green with C1000 Touch Thermal Cycler (Bio-Rad). PCR efficiency was calculated from the equation of efficiency (EFF) = 10(−1/slope) − 1 and authentic PCR products were confirmed by analyses of gel electrophoresis, DNA sequencing, and the melting curve. Real-time PCR data was analyzed using the absolute quantitation method, expressed as copies/μg RNA, and was determined using Ct values of samples and a standard curve from serial known concentrations of plasmids containing different cDNA fragment of target genes. Primers sequences can be found in our previous studies (Liu et al., 2018; Wu and Zhu, 2019).

Western blotting

Expression of nuclear progestin receptor (Pgr) in the fully-grown stage IVa immature oocytes was determined by the Western blot using a previously developed polyclonal antibody for Pgr (Hanna et al., 2010). In brief, stage IVa oocytes were collected from adult zebrafish following protocols listed previously (Hanna & Zhu, 2011; Liu et al., 2017). The total protein from ten stage IVa oocytes, collected directly from freshly sacrificed fish, was sonicated in 100 μl of 1x SDS sample buffer (62.5mM Tris-Cl pH 6.8, 2% SDS, 10% glycerol, 100mM Dithiothreitol) on ice for about 10 short bursts (Sonic Dismembrator, Fisher Scientific). Samples were then immediately boiled for 10 minutes and stored in −20 °C freezer until the start of Western blot analysis. 10 μL of each sample was loaded onto 8% SDS PAGE gel and transferred to a nitrocellulose membrane. The membrane was first pre-incubated for 3 hours with a blocking solution containing 5% BSA (Albumin from bovine serum, Sigma A7906) in TBST (50mM Tris, 100mM NaCl, 0.1% Tween 20, pH 7.4), then with a primary antibody (Pgr, 1:250 dilution; α-Tubulin (Sigma, T6074), 1:3000 dilution) in the 1% BSA blocking solution overnight. The following day, the membrane was washed five times for a period of 5 minutes each with 1x TBST, incubated for 2 hours with horseradish peroxidase conjugated secondary antibody (1:3000 dilutions, goat anti-rabbit antibody for Pgr detection or goat anti-mouse antibody for α-Tubulin), and finally washed five times for a period of 5 minutes each with 1x TBST. The membranes were developed using Super Signal West Extended Dura Substrate (Pierce, Rockford, IL) in plastic wraps, then visualized using a Fluor Chem 8900 imaging station (Alpha Innotech, San Leandro, CA). Protein size was determined by comparison between a biotinylated protein ladder (Cell Signaling Technology, Danvers, MA) and a pre-stained protein ladder (Fermentas, Waltham, MA). Finally, ImageJ was used to estimate relative densitometries (Schneider et al., 2012).

Spawning and fertility

After all zebrafish lines reached their maturity at ~ 4-month-old, ten homozygous mutant female fish were crossed with fertility confirmed wildtype males. Production of offspring for each genotype was recorded daily for a two-week period after a 14-day acclimatization period. Gonadosomatic index (GSI) = (weight of ovary/weight of body)*100%.

Chorion and yolk width measurement

Embryos at 30% epiboly stage were selected randomly and the length between two tips of cells was recorded as the width of yolk, and length of widest part of the chorion as width of chorion. In total, 7 individual pairs of wt or mprs^−/− (30 embryos for each individual) were recorded.

Environmental scanning electron microscope (ESEM)

To determine the submicroscopic structure difference on the surfaces of embryos, 16 to 32-cell stage (approximately 2 hours post fertilization) embryos were fixed in 4% paraformaldehyde overnight. Samples were then dechorionated and scanned by ESEM.

Histological examination and percentage of follicles at different stages

Three females from each genotype were anesthetized by submersion in a lethal dose of MS-222 (300 mg/L buffered solution) for 10 minutes, then severed the spinal cord and blood supply using procedures approved by the IACUC at 8 am. Ovaries were quickly removed and histological examination was performed following protocols described previously (Sullivan-Brown et al., 2011). Briefly, fresh ovaries were fixed overnight in 4% PFA, processed and embedded in JB-4 resin, cut into 5 μm sections, and stained with hematoxylin and eosin. The stages of oocytes were divided according to size and morphological characteristics described in a previous section. Each stage of oocytes was expressed as a percent of the total number of oocytes from both ovaries of each female used.

Statistical analysis

All the results were analyzed using GraphPad Prism 7.0a (San Diego, CA, US) and presented as mean ± SEM. Significant differences among paired treatment groups were determined using Student’s t-test. GVBD rates at different time points were analyzed by one-way analysis of variance (ANOVA) followed by Tukey’s multiple comparisons test. Statistical significance was set at p <0.05.

Results

Impaired oocyte maturation in individual mPR mutants

We generated mutant lines for each mPR paralogue, and conducted germinal vesicle breakdown (GVBD) assays in each of mPR mutant. Fully-grown immature stage IVa oocytes from mprα1^−/− (paqr7a^−/−), mprα2^−/− (paqr7b^−/−), mprβ^−/− (paqr8^−/−), and mprγ2^−/− (paqr5b^−/−) showed impaired oocyte maturation in response to DHP stimulation, but not those from mprγ1^−/− (paqr5a^−/−), mprδ^−/− (paqr6^−/−), and mprε^−/− (paqr9^−/−) (Fig. 1; supplemental tables 2 & 3). Among single mutants, mprα2^−/− (paqr7b^−/−) oocytes exhibited the most significant reduced oocyte maturation (Fig. 1B). In all single mutants, spontaneous oocyte maturation still could happen at low rates when treated with ethanol only (Fig. 1).

Figure 1. Inhibition of germinal vesicle breakdown (GVBD, i.e. final oocyte maturation) in stage IVa fully-grown immature oocytes in the knockouts of individual membrane progestin receptor (mPR) paralogues in comparison to those from wildtype (wt).

Stage IVa immature oocytes were exposed to 5 nM exogenous 17α,20β-dihydroxy-progesterone (DHP) or vehicle (0.1% ethanol) and incubated at room temperature for up to 5 hours, rate of GVBD were determined every 30 minutes. (A) mprα1^−/− (paqr7a^−/−); (B) mprα2^−/− (paqr7b^−/−); (C) mprβ^−/− (paqr8^−/−; (D) mprγ2^−/− (paqr5a^−/−); (E) mprγ12^−/− (paqr5b^−/−); (F) mprδ^−/− (paqr6^−/−); (G) mprε^−/− (paqr9^−/−). The results shown as average (mean ± SEM) of data from five representative experiments. *p < 0.05; **p < 0.01; ****p < 0.0001.

Impaired oocyte maturation in a mutant lacking all seven mPR paralogs (mprs^−/−)

To determine possible gene redundancy among mPR paralogs, a zebrafish line lacking all 7 mPRs (mprs^−/−) was generated. Immature stage IVa oocytes can be found in 4 out of 9 mprs^−/− females after lights were on; in contrast, only 1 out of 9 wt female possessed immature stage IVa oocytes at same time point. The number of stage IVa oocytes was higher in mprs^−/− females than wt, but was not significant. Abnormal higher frequency of stage IVa oocytes in mprs^−/−females indicates impaired oocyte maturation in vivo (Fig. 2A). This impaired oocyte maturation in vivo could be due to less sensitivity of fully-grown immature oocytes to progestin in mprs^−/−. In GVBD assays, fully-grown immature oocytes from mprs^−/− exhibited significantly reduced maturation over a 5-hour treatment (Fig. 2B–2D; supplemental table 4). The reduction of GVBD in mprs^−/− was significantly lower than those observed in single mPR mutants including mprα2^−/− (Supplemental Fig.8). Even being treated with DHP for 5 hours, only ~70% immature oocytes from mprs^−/− became matured. In contrast, around 100% of wildtype oocytes matured after treated with DHP for 3 hours. Interesting, around 20% of oocytes from wildtype underwent maturation spontaneously after treated with vehicle (ethanol) for 5 hours. In contrast, none of the oocytes from mprs^−/−matured spontaneously when treated with vehicle only (Fig. 2B).

Figure 2. Further reduction of progestin sensitivities in stage IVa fully-grown immature oocytes with all seven membrane progestin receptor paralogues mutated (mprs^−/−).

(A) Increased number of females with stage IVa oocytes found in ovaries from mprs^−/− in vivo. Ovaries were collected at 0.5 hours following lights on from nine females. Only one out of nine wildtype (wt) female had stage IVa oocytes, while 4 out of 9 mprs^−/− had stage IVa oocytes. (B) Further reduction in germinal vehicle breakdown (GVBD) in response to 5 nM exogenous 17α,20β-dihydroxy-progesterone (DHP) in vitro in stage IVa oocytes with all seven mPR paralogues (mprs^−/−) mutated. (C) Majority of stage IVa oocytes from wt underwent GVBD after a 3-hour incubation with 5 nM DHP in vitro. Oocyte become translucent following GVBD because of fusion of yolk protein that permits more lights going through. An immature oocyte failed undergoing GVBD is indicated by an arrow. (D) Many stage IVa oocytes from mprs^−/− did not undergo GVBD following treatment with DHP for 3 hours. Immature oocytes failed undergoing GVBD are indicated by arrows. Scale bar: 1mm. *p < 0.05; **p < 0.01; ***p < 0.001; ****p < 0.0001.

Ovulation delay in mprs^−/− females

In addition to impaired oocyte maturation, we also found impaired ovulation in mprs^−/− females in vivo. Significant numbers of matured, but not yet ovulated follicles (stage IVb follicles) were only found in the ovaries of mprs^−/− in vivo 30 minutes after lights were turned on but not in wt females (Fig. 3A). Since nuclear progestin receptor (Pgr) is the key mediator for ovulation, we determined Pgr expression in mprs^−/− oocytes. The transcript levels of pgr were lower in stage III oocytes (21:00) and stage IVb oocytes (8:00) in mprs^−/− compared to those in wt (Fig. 3B). Meanwhile, the protein level of Pgr was also significantly reduced in stage IVa follicles in the mprs^−/− (Fig. 3C). Therefore, we checked the upstream (lhcgr (luteinizing hormone receptor)) and downstream (proteinases) gene expressions of Pgr and found reduced expression of lhcgr, mmp2 (matrix metallopeptidase 2) in pre-ovulatory follicles in mprs^−/− (Fig. 4).

Figure 3. Reduced ovulation due to attenuated expression of nuclear progestin receptor (Pgr) in all seven membrane progestin receptor paralogues mutant (mprs^−/−) in vivo.

(A) Stage IVb oocytes were found in the ovaries of mprs^−/− 30 minutes after lights on but not in wildtype (wt) females in vivo. N=9. (B) Reduced expression of pgr transcripts in most advanced oocytes in mprs^−/− (stage III oocytes collected at 21:00, stage IVa oocytes collected at 6:00, stage IVb oocytes collected at 8:00. n=6). (C) Reduced expression of Pgr protein in the stage IVa oocytes from mprs^−/− collected at 6:00 compared to those from wt (n = 3). *p < 0.05; **p < 0.01.

Figure 4. Reduced expression of luteinizing hormone receptor (lhcgr) and metallopeptidases in most advanced oocytes in all seven membrane progestin receptor paralogues mutant (mprs^−/−; 21:00, stage III; 06:00, stage IVa; 08:00, stage IVb).

Low lhcgr and mmp2 expression can be observed in Stage IVb oocytes of mprs^−/−. In addition, expression of adamts1 was upregulated in Stage IVb oocytes in mprs^−/−. (A) lhcgr; (B) adam8b; (C) adamts1; (D) adamts8a; (E) adamts9; (F) mmp2; (G) mmp9. Asterisks indicate a significant difference of transcripts compared to wildtype (wt) at the same time point. adam8b, a distintegrin and metalloproteinase domain 8b; adamts1, a disintegrin and metalloproteinase with thrombospondin type 1 motif 1; adamts8a; adamts9; mmp2, matrix metalloproteinase 2; and mmp9. *p < 0.05; **p < 0.01.

Reduced fertility in mprs^−/− females

To evaluate the fertility of mprs^−/−, we conducted a continuous mating study using mature female fish (n=10) of 4-month-old with known fertile wildtype males over 2 weeks. The fecundity of mutant female zebrafish was remarkably compromised (mprs^−/−, 914.6 ± 105.8 fertilized embryos/2 weeks, p=0.0039) in comparison with wt females (1361 ± 83.88 fertilized embryos/2 weeks, n=10) (Fig. 5A). In addition, we found that mprs^−/−females (59.96 ± 7.69%) had low spawning frequency than those in wt females (71.43 ± 5.63%), though was not significant (Fig. 5B). Cross anatomical examination revealed an obvious decrease in the size of ovaries in mprs^−/− when compared with their wt littermates. Gonadosomatic index (GSI) was also smaller in mprs^−/− than wt (Fig. 5C). We observed mutant females always produced less embryos than wt females (Fig. 5D). In addition, less stage V oocytes were obtained from mprs^−/− than from wt (Fig. 5E). All these evidences suggested fertility is compromised in mprs^−/−females.

Figure 5. Reduced fertility in all seven membrane progestin receptor paralogues mutant (mprs^−/−) females.

(A) Fewer embryos were spawned in mprs^−/− females when compared to wildtype (wt) during a two-week examination. (B) Spawning frequency in mprs^−/− female was compromised, but not significantly. (C) Gonadosomatic index (GSI) of mprs^−/− female was smaller than wt. (D) Daily fertilized embryos released from mprs^−/− in comparison to wt (n=10). (E) Fewer Stage V oocytes were found in mprs^−/− ovaries. *p < 0.05; **p < 0.01.

Poor-quality of zygotes from mprs^−/− females

Not only the number, but also the quality of embryos is affected in mprs^−/−. Compared to embryos from wt incross, embryos from mprs^−/− females cross with wt males had smaller yolks (Fig. 6A). In addition, the embryos from mprs^−/− female were restricted in a smaller chorion with narrow space for embryonic development (Fig. 6B). The ratio of width of chorion to that of yolk was also significantly reduced in embryos from mprs^−/−females when compared to those from wt females (Fig. 6C). The yolk of some mutant embryos appeared dark and opaque under a dissecting microscope, and the cells divided in a disorganized pattern at a very early stage (8–16 cell stage) (Fig.6D). We found the yolk granules on the surface of early-stage embryos from mprs^−/− females displayed a disorganized and scrambled pattern with multiple holes (Fig. 6Ec and 6Ed) when compared to those from wt (Fig. 6Ea and 6Eb). Embryos from mprs^−/− males crossed with wt females showed normal appearance and development as those from wt (data not shown).

Figure 6. Malformed embryos from in all seven membrane progestin receptor paralogues mutant (mprs^−/−) females.

(A-C) Reduced yolk size, chorion width, and the ratio of chorion size to yolk size in the mprs^−/− embryos. 30% epiboly embryos were used (n=7). (D) Representative malformed embryos (16-cell stage) from mprs^−/− females with smaller yolk and chorion. (E) Disorganized and scrambled yolk granules on the surface of mprs^−/− early stage embryo (32-cell stage embryo). a & b) wt; c & d) mprs^−/−. Scale bars: a & c) 200 μm; b & d) 50 μm.

Abnormal early oogenesis in mprs^−/− females

In addition to impaired oocyte maturation and ovulation, we found mPRs also had roles in early oogenesis. Higher percentage of stage I oocyte was observed in mprs^−/− when compared to wt (Fig. 7A). Compared to wt (Fig. 7B), more stage I oocytes were observed in the ovaries of mprs^−/− (Fig. 7C).

Figure 7. Abnormal oogenesis in all seven membrane progestin receptor paralogues mutant (mprs^−/−) ovaries.

(A). Higher percentage of stage I oocytes was found in mprs^−/−. (B) & (C). Representative image of wildtype (wt) or mprs^−/− ovary. Scale bar: 1mm. *p < 0.05.

Discussion

Accumulating evidence suggest mPRα is the nongenomic progestin receptors involved in various physiological processes including oocyte maturation in several teleosts (Hanna et al., 2006; Hanna and Zhu, 2011; Shi et al., 2016; Tubbs et al., 2010; Zhu et al., 2003b). However, studies to date were focused mostly on mPRα. The functions of other mPR paralogues still remain unclear. Most critically, all the studies conducted to date were either using overexpression or knockdown approaches, evidence from knockouts are still lacking. To our knowledge, this is the first attempt to identify mPRs’ functions in reproduction in vivo using gene editing method to knockout each and all seven mPRs in one organism. Our study clearly demonstrated feasibility and advantages of alteration of multiple genes in a relative economic model, zebrafish. Our results clearly show that mPRs have functions in progestin induced oocyte maturation, ovulation, egg quality, and female fertility. Involvement of mPR signaling in meiosis resumption has been shown in other teleost species including goldfish (Tokumoto et al., 2006; Tokumoto et al., 2012), medaka (Roy et al., 2017), Atlantic croaker (Tubbs et al., 2010), European eel (Morini et al., 2017), and flounder (Shi et al., 2016). The involvement of mPRs in the regulation of ovulation, egg quality and female fertility were only first demonstrated in current study. Manipulation of mPRs signaling should affect female reproduction including oocyte maturation, ovulation and egg quality, which may have important implications for aquaculture. Progestin signaling is also important for various reproductive processes including endometrial preparation, follicular development, steroidogenesis, ovulation, implantation of fertilized egg, maintenance of pregnancy, breast development, and milk production in mammals (Dressing et al., 2010; King and Brucker, 2010; Macias and Hinck, 2012; Nilsson et al., 2006; Peluso, 2006; Peluso et al., 2008). Further investigation on the roles and signaling of mPRs will advance our understanding and regulation of these important processes in mammals.

The mPRα and mPRβ are the membrane progestin receptors that mediate progestin signaling and induce oocyte maturation. Firstly, mPRα and mPRβ display a high affinitive (Kd~5 nM), saturable, displaceable, single-binding site specific for DHP (Hanna et al., 2006; Thomas et al., 2014; Valadez-Cosmes et al., 2016). Secondly, mPRα and mPRβ proteins have been localized on the plasma membranes of granulosa cell, theca cells, and oocytes (Dressing et al., 2010; Hanna and Zhu, 2011; Thomas, 2003). In addition, prior to oocyte maturation, in response to a dramatic increase in progestins, mPRα and mPRβ are up-regulated in fully-grown immature oocytes (Hanna and Zhu, 2011; Thomas, 2003; Zhu et al., 2003b). Also, mPRs appear to function as plasma membrane-bound GPCRs mediating rapid actions of P4 via activation of an inhibitory G protein (Gi) and suppression of adenylyl cyclase activity and cAMP production (Hanna et al., 2006; Thomas et al., 2007). In a previous study, antisense microinjection of mPRα can block progestin induced oocyte maturation (Zhu et al., 2003b). Agreeing with this previous study, we also found impaired oocyte maturation in vivo and in vitro in this study. Overall, mPRs located on the oocyte membrane, bind progestin with high affinity, and are involved in the nongenomic action of progestin in oocyte maturation.

The functions and signaling of different membrane progestin receptors in oocyte maturation may not be identical. For example, signaling of zebrafish mPRα but not mPRβ was blocked by pertussis toxin, implying mPRβ activates a different pertussis toxin-insensitive G protein (Hanna et al., 2006). In addition, over-expression of mPRα in follicle-enclosed oocytes significantly increased the activity of MAPK, the production of cyclin B protein, and the number of oocytes that underwent oocyte maturation without exogeneous progestin, while over-expression of mPRβ alone had no such effect (Hanna and Zhu, 2011). We also investigated the different mPRs in oocyte maturation. Our results suggest that mPRα1 (Paqr7a), mPRα2 (Paqr7b), mPRβ (Paqr8), and mPRγ2 (Paqr5b), but not mPRγ1 (Paqr5a), mPRδ (Paqr6), and mPRε (Paqr9) play a role in oocyte maturation. In addition, the largest reductions in oocyte maturation were observed in oocytes from mprα2^−/− and mprγ2^−/− The mPRα2 was recognized as mPRα in zebrafish and other teleost species previously. Therefore, our results is consistent with those reported previously (Hanna et al., 2006; Hanna and Zhu, 2011; Shi et al., 2016; Zhu et al., 2003b). In contrast, the studies of mPRγ2 in oocyte maturation in vitro are still missing which need further investigation. Due to genetic compensation in zebrafish (El-Brolosy et al., 2019; Ma et al., 2019), even mPRδ (Paqr5a), mPRδ (Paqr6), and mPRε (Paqr9) seem dispensable for oocyte maturation, these mPRs may still play a role in oocyte maturation in zebrafish. Therefore, we generated a mutant line with lacking all seven mPRs (mprs^−/−). However, even in mPRs total mutant (mprs^−/−), oocyte maturation still can occur slowly and incompletely. Other progestin receptors besides mPRs may explain why oocyte maturation still can happen in mprs^−/− Besides mPRs, there are two progestin receptor families, Pgr and Pgrmcs. Around 5% of the Pgr can be associated with the plasma-membrane caveolae (Norman et al., 2004). Pgr’s palmitoylation, a post-translational modification that may be important for membrane anchoring of Pgr (Pedram et al., 2007). Acceleration of oocyte maturation by overexpression of Pgr in Xenopus and zebrafish oocytes also suggest the involvement of Pgr in meiosis resumption (Bayaa et al., 2000; Hanna et al., 2010; Tian et al., 2000). Pgr may participate in oocyte maturation through interacting with SRC tyrosine kinase and subsequent activation of the MAPK pathway (Zhang et al., 2008). However, we found acceleration of oocyte maturation in three different Pgr knockout lines (unpublished data). Therefore, the role of Pgr in oocyte maturation remain controversial. The Pgrmcs may also play important roles in oocyte maturation. Pgrmc1 antibody injection significantly lowered oocyte maturation percentages after 24 hours of culture (Luciano et al., 2010). Similar results were found in pgrmc1^−/− female zebrafish (Wu et al., 2018). But Pgrmc1 may not regulate oocyte maturation directly, as no evidence is available on the direct interaction between Pgrmc1 and G-protein. Pgrmc1 might facilitate expression and plasma localization of mPRα, which in turn regulates oocyte maturation (Thomas et al., 2014; Thomas et al., 2007; Wu et al., 2018). Interestingly, around 20% of oocytes from wt females spontaneously matured without the ligand, DHP, while none of the oocytes from mprs^−/−matured under same condition. In yeast, mPRs can constitutively transduce signals without ligand when expressed at high levels, but did require an agonist when expressed at low levels (Kupchak et al., 2009; Smith et al., 2008; Villa et al., 2009). Right before oocyte maturation, mPRs are up-regulated and/or activated and then the net production of the second messenger increases above the threshold necessary to initiate downstream signaling molecules that initiate oocyte meiosis in wt zebrafish (Hanna and Zhu, 2011). In contrast, spontaneous oocyte maturation could not occur due to lack of mPRs and their signals in mprs^−/−. Therefore, mPRs may be important for spontaneous oocyte maturation.

No evidence has shown mPRs located in nucleuses and functioned as transcription factors. However, nongenomic actions through mPRs may eventually affect the genomic action of progestin. Mutation of mPRs may affect nongenomic action through MAPK signaling pathway (Pace and Thomas, 2005), which can eventually affect gene transcription and translation, resulting in lower level transcription and translation of nuclear progestin receptor (Pgr). Studies have shown Pgr is essential for ovulation across different species (Kubota et al., 2016; Lydon et al., 1995; Zhu et al., 2015). Then, low expression levels of Pgr in mPRs total mutants cause the ovulation delay in vivo. Similar low Pgr level and impaired ovulation was found in pgrmc1/2^−/− (Wu and Zhu, 2019). The previous study also showed progestin membrane receptor-mediated pathways can regulate transactivation of Pgr resulting in an alteration in gene transcription (Karteris et al., 2006).

Ovulation normally began around 8:00 (one hour before lights on). In order for ovulation to occur, nuclear progestin receptor (Pgr) and its related signaling molecules need to be expressed at appropriate levels around 8:00. Although the Pgr transcript was not significantly different in mprs^−/− compared to its level in wt at 6:00 (Fig. 3B), the Pgr protein was clearly lower in mprs^−/− knockout than its expression in wt at 6:00 (Fig. 3C). The low transcript levels of genes including pgr and lhcgr at 21:00 (the night before spawning) and 8:00 (the morning prior to the spawning) may indicate overall low basal expression of genes in mprs^−/−. This low expression of Pgr may further reduced protein expression of Pgr and its downstream molecules important for ovulation. It would be ideal to determine the protein levels and enzyme activities of these genes, which will be the topics of future studies when research reagents and tools such as antibodies and enzymatic assays become available. Besides oocyte maturation and ovulation, we also found significant more stage I oocytes in mprs^−/−females (Fig. 7). This indicates impaired development in early oogenesis in mprs^−/−oocytes. Similar phenotype was also found in pgrmc1^−/−and pgrmc2^−/− mutant zebrafish (Wu et al., 2019; Wu et al., 2018). During early stages of oogenesis in Japanese huchen (Hucho perryi) and common carp (Cyprinus carpio), DHP significantly promotes DNA synthesis in the ovarian germ cells and acted directly on the initiation of the first meiosis of oogenesis (Miura et al., 2007). Pgrmcs and mPRs may play an important role in promoting early oogenesis; however, further studies are needed.

Overall, our results showed mPRs are important for oocyte maturation and ovulation. However, further studies are required to identify the additional membrane progestin receptors for oocyte maturation in zebrafish since oocyte maturation could still occur even all seven paralogs of mPR were mutated.

Highlights.

• Impaired oocyte maturation and ovulation in mPR female mutants

• Reduced egg quality and fertility in mPR female mutants

• mPRs regulate Pgr to affect ovulation