Progestin and Nuclear Progestin Receptor Are Essential for Upregulation of Metalloproteinase in Zebrafish Preovulatory Follicles

Ovulation requires proteinases to promote the rupture of ovarian follicles. However, the identity of these proteinases remains unclear. In our previous studies using RNA-seq analysis of differential expressed genes, we found significant down-regulation of five metalloproteinases: adam8b (a disintegrin and metalloproteinase domain 8b), adamts8a (a disintegrin and metalloproteinase with thrombospondin motif 8a), adamts9, mmp2 (matrix metalloproteinase 2), and mmp9 in the nuclear progestin receptor knockout (pgr−/−) zebrafish that have failed to ovulate. We hypothesize that these metalloproteinases are responsible for ovulation and are regulated by progestin and Pgr. In this study, we first determined the expression of these five metalloproteinases and adamts1 in preovulatory follicles at different times within the spawning cycle in pgr−/− and wildtype (wt) zebrafish and under varying hormonal treatments. We found that transcripts of adam8b, adamts1, adamts9, and mmp9 increased drastically in the preovulatory follicular cells of wt female zebrafish, while changes of adamts8a and mmp2 were not significant. This increase of adam8b, adamts9, and mmp9 was significantly reduced in pgr−/−, whereas expression of adamts1 was not affected in pgr−/− zebrafish. Among upregulated metalloproteinases, adamts9 mRNA was found to be expressed specifically in follicular cells. Strong immunostaining of Adamts9 protein was observed in the follicular cells of wt fish, and this expression was reduced drastically in pgr−/−. Interestingly, about an hour prior to the increase of metalloproteinases in wt fish, both Pgr transcript and protein increased transiently in preovulatory follicular cells. The results from in vitro experiments showed that adamts9 expression markedly increased in a dose, time and Pgr-dependent manner when preovulatory follicles were exposed to a progestin, 17α,20β-dihydroxy-4-pregnen-3-one (DHP). Taken together, our results provide the first evidence that upregulation of adamts9 occurs specifically in preovulatory follicular cells of zebrafish prior to ovulation. Progestin and its receptor (Pgr) are essential for the upregulation of metalloproteinases.


INTRODUCTION
Ovulation is essential for successful reproduction to occur. In vertebrates, ovulation is triggered by a surge of luteinizing hormone (LH) and mediated by progesterone (P4) (1)(2)(3)(4)(5). Ovulation also requires upregulation of proteolytic enzymes (6). These proteinases promote the rupture of follicles releasing mature oocytes. Studies have shown correlations between increased expression of proteolytic enzymes and elevated levels of P4 or its receptor (a.k.a. the progestin receptor, PGR). For instance, induction of transcripts of Adamts1 (a disintegrin and metalloproteinase with thrombospondin motifs 1) by human chorionic gonadotropin (hCG), a popular substitute for LH, was reduced in PGR knockout (Pgr −/− ) mice (7). A synthetic PGR agonist, R5020, could reverse the inhibitory effect of mifepristone (RU486) on the LH induced expression of ADAMTS1 and ADAMTS9 mRNA in granulosa cells of cattle (8). In a teleost medaka, mmp15 (matrix metalloproteinase 15) was found to be upregulated by Pgr within LH exposed follicles (9). Further, plasminogen activator and several other MMPs were also reported to be regulated by PGR during ovulation (3,10,11). Though the relationships between PGR and these proteases have implicated their involvement in the ovulation process, our knowledge of the regulation and functions of these proteases is extremely limited. So far, knockout studies of these metalloproteinases in mice have provided little information on the functions of these proteases, mainly due to null mice either dying in utero or exhibiting no observable defects (12)(13)(14)(15)(16). One exception is ADAMTS1 knockout mice that were found to be sub-fertile, but ultimately they were still able to ovulate (17). By comparison, anovulatory PGR knockout mice were completely infertile (18). Furthermore, ADAMTS1 is not expressed in cumulus cells (19), ADAMTS1 gene does not have P4 receptor response element (20) and is believed not to be regulated by Pgr in high mammalian species (personal communication). These studies suggest that there may be critical protease(s) other than ADAMTS1 necessary for ovulation that has not yet been identified.
Zebrafish is an established model for studying gene functions and signaling pathways in conserved ovarian events such as oogenesis, oocyte maturation, and ovulation (21)(22)(23). Our previous results show that pgr −/− female zebrafish were unable to ovulate demonstrating the conserved function of Pgr in ovulation from fish to mice (18,24,25). Furthermore, compared to pgr −/− zebrafish, a genome-wide analysis of transcripts revealed conserved signaling pathways and higher expression levels for adamts8a, adamts9, mmp2, mmp9, and adam8b (a disintegrin and metalloproteinase domain 8b) in the follicular cells of wildtype (wt) fish (26). We hypothesized that some of these genes are likely obligatory for follicular rupture and ovulation in zebrafish. In this study, we aimed to elucidate the expression and hormonal regulation of these proteinases in zebrafish ovarian follicles. We first determined the expression changes of five abovementioned metalloproteinases and adamts1. We found dramatic increases of adam8b, adamts1, adamts9, and mmp9 in the follicular cells prior to ovulation.
Interestingly, expression of adamts9, adam8b, and mmp9 were significantly reduced in pgr −/− fish, whereas that of adamts1 was not affected. Then, we found that 17α,20β-dihydroxy-4-pregnen-3-one (DHP) stimulated adamts9 expression in preovulatory follicles in a dose, time and Pgr-dependent manner. Our study suggests progestin and Pgr are critical for the upregulation of metalloproteinases prior to ovulation in zebrafish.

Zebrafish Husbandry
The wt zebrafish used in this study are a Tübingen strain initially obtained from the Zebrafish International Resource Center and propagated in our lab. Pgr gene knockout lines used in this study were generated and characterized previously (24,27

Collection of Stage I-IV Follicles
Thirty mature females from wt or pgr −/− were collected at 08:00 (1 h before lights turned on) from group housed tanks in conditions described as in section Zebrafish Husbandry. These fish were deeply anesthetized in a lethal dose of MS-222 (300 mg/L buffered solution) for 10 min. To ensure death, the spinal cord and blood supply behind the gill cover were cut off using sharp scissors. Ovaries were removed immediately and placed in a 90-mm petri dish containing 60% L-15 media (Sigma, in 15 mM HEPES, pH 7.2). Thereafter, ovaries were cut into ∼4 mm 3 pieces, and then transferred to a 15-ml centrifuge tube. Individual oocytes were separated by pipetting up and down using a disposable glass pipette with a polished 3 mm opening. Hereafter, we will use "follicles" to specifically refer to follicular cells and their enclosed oocytes to distinguish them from follicular cells or defolliculated oocytes (i.e., denuded oocytes). Individual follicles of various sizes ( Table 1) were separated into five stages according to an established oocyte classification system in zebrafish with modification (28). We further subdivided stage IV follicles into two stages, immature stage IVa (before germinal vesicle breakdown, i.e., GVBD) and mature stage IVb (after GVBD has occurred but prior to ovulation). Though the outside appearance of stage IVb follicles is transparent, same as stage V, distinguishing stage IVb follicles from stage V mature ova is straightforward since stage IVb mature follicles are scattered around the ovary, surrounded by follicular cells and other immature follicles. Whereas, stage V ovulated oocytes are grouped together and pushed to the posterior part of the ovary in vivo (26). Stage V ovulated oocytes were not collected for two reasons: (1) Our targeted genes are expressed mainly in the follicular cells, and stage V mature oocytes do not have follicular cells; (2) It was not possible to collect stage V oocytes from pgr −/− fish due to their inability to ovulate.

Collection of Follicular Cells and Denuded Oocytes From Stage IV Follicles During a Spawning Cycle
In group housing conditions, zebrafish skip spawning frequently in part due to intense competition for space and food. To increase and monitor individual spawning, we set up multiple spawning tanks with a pair of approximate 3-month old mature male and female wt fish for each tank. Everyday around 22:00 (1 h prior to lights off), water in spawning tanks was replaced with clean water along with an inner tank insert that allows fertilized eggs to drop through to the bottom to prevent the eating of the eggs by the adults. Spawning and release of fertilized eggs (∼150 embryos/day) were visually confirmed and recorded for each pair of fish every morning. At 12:00, fish were transferred to a new spawning tank with clean water but without an insert, so they could access the commercial fish food and newly hatched brine shrimp. Fish food and brine shrimp were supplemented every 3 h. In this setup and enhanced feeding condition, the majority of the pairs (7-8 out of 10 pairs) spawn almost daily. At 1 week following the setup of spawning, the mature female zebrafish of the pairs were sacrificed, their ovaries removed, and stage IV follicles were collected as described in section Zebrafish Husbandry. Stage IV follicles from wt fish were collected at four different time points: 13:00 (stage IVa, from fish that skipped spawning that morning); 21:00 (stage IVa, 2 h before lights off); 06:00 (stage IVa, onset of oocyte maturation and 3 h before lights on); and 08:00 (stage IVb, after oocytes have matured but before ovulation, 1 h before lights on). To determine the effect of Pgr knockout, stage IVb follicles at 08:00 were also collected from 3-month old virgin female pgr −/− fish that were paired with fertile wt male at the same time. Follicular cells (collected from ∼100 follicles/fish) and their enclosed oocytes (collected from ∼10 follicles/fish) were separated from stage IV follicles using a pair of small glass needles. Respective samples were pooled and homogenized immediately in RNAzol solution according to an established procedure (26,29).

Various Hormone Treatments of Stage IVa Fully-Grown Immature Follicles in vitro
Stage IVa follicles (>650 µm) with visible germinal vesicles (GV) were collected from females at ∼05:30. Briefly, intact follicles with no obvious damage were selected and transferred into a 24well tissue culture plate containing 60% L15 medium (25 follicles per well). These follicles were incubated for 2.5 h at 25 • C with DHP (1-1,000 nM), testosterone (T, 500 nM), or RU486 (0.01-10 µM), alone or in combination. An exposure time of 2.5-h was selected to conduct the various hormone treatment based on results from our time course experiments (up to 6 h). Follicles that underwent final oocyte maturation, indicated by transparent yolk and GVBD, were easily determined under a dissecting microscope at the end of incubation. Excluding broken follicles, the number of transparent follicles that completed GVBD were counted and presented as a percentage of the total follicles. Thereafter, all the follicles were collected and homogenized immediately in RNAzol for qPCR analyses of gene expression, or in 1X SDS sample buffer for Western blot analyses of protein expression.

RNA Extraction, Reverse Transcription, and Real-Time Quantitative PCR
Total RNA was extracted using RNAzol (MRC, Cincinnati, Ohio, USA) and a Qiagen RNeasy kit. According to the manufacturer's protocol, BAN solution (4-bromoanisole) was added to purify the RNA and eliminate genomic DNA following the first precipitation step with water. After the second precipitation, an equal volume of cold 100% ethanol was added. The mixture was then loaded onto a RNeasy free spin column, centrifuged (8,000 g, for 30 s), washed twice with 650 µL 75% ethanol, and eluted in RNase-free water. We used 15-30 µL depending on the initial amount of sample. The approximate concentration and purity of samples were determined using a Nanodrop 2000 Spectrophotometer. RNA samples with concentrations >100 ng/µL, OD 260/280 >1.8, and OD 260/230 >1.3 were retained for further analyses. Reverse transcription was performed using SuperScript III Reverse Transcriptase and 0.5 µg of total RNA from each sample in a reaction volume of 10 µL, per manufacturer's instructions (Invitrogen, Carlsbad, CA). Gene expression was determined by quantitative real time PCR (qPCR)

Gene symbol
Accession number  (28). Stage IV follicles were further sub-divided into stage IVa and stage IVb. Stage IVa follicles were fully-grown immature follicles prior to oocyte maturation. Stage IVb follicles were follicular cells enclosed mature oocytes that have gone through oocyte maturation, but not yet ovulated (also see Table 1 for detail). Different letters (uppercase for wt; lowercase for pgr −/− ) indicate significant differences among different follicular stages in wt or pgr −/− fish. Significant differences between wt or pgr −/− fish at the same developmental stage of follicles are indicated by asterisks. *p < 0.05, *****p < 0.00001. N = 6. The metalloproteinases examined were 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, matrix metalloproteinase 9, respectively.
To avoid genomic DNA contamination, forward and revise primers were designed to be in two different exons of each target gene. Authentic single qPCR product for each gene was confirmed by melting curve analyses, gel electrophoresis, and sequencing. The absolute transcript levels, expressed as copies/µg total RNA, were determined using Ct-values of samples and standard curves corresponding to different target genes generated from known serial plasmid concentrations (10 2 -10 7 copies/µL). We did not use the comparative Ct method for this study because house-keeping-genes including actb1, actb2, and ef1a vary between different developmental stages of follicles (data not shown).

Western Blotting
Total protein was extracted from 10 stage IVa or IVb follicles from newly sacrificed fish at various time points during an ovulatory cycle, or from in vitro incubation. Collected follicles were homogenized immediately by sonication in 100 µl of 1X SDS sample buffer (62.5 mM Tris-Cl pH 6.8, 2% SDS, 10% glycerol, 100 mM Dithiothreitol) with 0.1% protease inhibitor cocktail (Sigma, Saint Louis, MO) on ice for about 10 short bursts (1-2 s for each burst, Sonic Dismembrator, Fisher Scientific). Samples were then boiled immediately for 10 min and stored in −20 • C until analysis. Samples were loaded onto an 8% SDS PAGE gel (10 µL is equivalent to one follicle), separated by electrophoresis, and transferred to a nitrocellulose membrane. The membrane was pre-incubated for 2 h with a blocking solution containing 5% BSA (albumin

Statistical Analyses
For the comparison of two data sets (e.g., wt vs. pgr −/− , or DHP vs. vehicle control), unpaired students' t-test was used to determine significant differences. For multiple group comparison, one-way ANOVA followed by Tukey's test or Dunnett's test was used. All experiments were repeated at least three times, and the results of one representative experiment is shown in the following section.

Changes of Metalloproteinase Expression in Stage I-IV Follicles in Mature Female Zebrafish
We hypothesized that metalloproteinases required for ovulation should be expressed highly in late stage follicles such as stage IVa or IVb follicles. Therefore, we separated follicles according to their sizes and stages, and determined expression of six representative metalloproteinases in early (stage I) through late stage (IVb) follicles (Figure 1). Expression of adamts9 remained at extremely low levels (<300 copies/µg total RNA) and was sometimes undetectable from stage I through stage IVa fullygrown immature follicles; however, its expression increased drastically in stage IVb mature follicles in both wt and pgr −/− fish. This increased expression of adamts9 in pgr −/− (3,000-4,000 copies/µg total RNA) was only one-tenth of that in wt fish (∼40,000 copies/µg total RNA). Intriguingly, high expression of adam8b, adamts1, mmp2, and mmp9 were observed in all stages of follicles (>10,000 copies/µg total RNA), whereas expression of adamts8a remained at low levels (<2,000 copies/µg total RNA). Surprisingly, none of these five metalloproteinases (adam8b, adamts1, adamts8a, mmp2, mmp9) showed significant difference in their expression between pgr −/− and wt.

Pgr Mediates the Upregulation of Metalloproteinases in Preovulatory Follicular Cells
We then hypothesized that changes of metalloproteinases in late stage follicles, especially in the follicular cells, could be important for ovulation but their activity may be masked by differential expression and/or high levels of metalloproteinase transcripts in large oocytes when entire follicles (follicular cells and their enclosed oocytes) are used. Hence, we focused on changes of transcripts in stage IVa and IVb fully grown follicles. We also separated the follicular cell layers from their enclosed oocytes and determined gene expression in both cell types during the ovulatory cycle in wt female zebrafish (Figure 2). In follicular cells, expression of adam8b, adamts1, adamts9, and mmp9 remained relatively low at most times but increased significantly at about 1 h (08:00) prior to ovulation that occurred at around 09:00 when lights were turned on. The changes of adamts8a and mmp2 transcripts were not significant during the daily spawning cycle. In denuded oocytes, the expression of adamts8a and adamts9 was extremely low and nearly undetectable (<300 copies/µg total RNA). By contrast, expression of adam8b, adamts1, mmp2, and mmp9 in denuded oocytes was relatively high (>5,000 copies/µg total RNA) but did not significantly change during the ovulatory cycle. Because the expression of metalloproteinases was upregulated in preovulatory stage IVb follicles (mature but not ovulated follicles) sampled at 08:00, we hypothesized that Pgr is an upstream regulator for these metalloproteinases. We found significantly reduced expression of adam8b, adamts9, and mmp9 in follicular cells of pgr −/− fish compared to those in wt fish (Figure 3). Expression of adamts1 was not affected by Pgr knockout as ADAMTS1 is not regulated by PGR in high mammalian species. In denuded oocytes, the expression of all aforementioned metalloproteinases was not significantly different between wt and pgr −/− fish (data not shown).

Changes of Pgr Expression During Ovulatory Cycle
We hypothesized that upstream regulators of these metalloproteinases, especially Pgr would be upregulated prior to the upregulation of these metalloproteinases. As we predicated, expression of Pgr (transcript and protein) in follicular cells began to increase in the early morning (05:00, prior to maturation), reached a peak level in the middle of maturation (06:00-07:00), then decreased gradually thereafter (Figure 4).

Exogenous DHP Exposure Directly Upregulates adamts9 in Follicles in vitro
To determine if low expression of metalloproteinases found in pgr −/− zebrafish is due to direct effect of progestin and its receptor (Pgr) in the follicular cells, we chose an in vitro oocyte maturation assay (29), and focused on hormonal regulation of adamts9 because only adamts9 meets all following criteria: (1) Expressed specifically in follicular cells but not in the oocytes (Figures 2, 5); (2) Expressed differently between wt and pgr −/− (Figures 2, 3, and 5); and (3) Increases significantly prior to ovulation (Figures 1, 2).
Consistent with expression of adamts9 transcripts, expression of Adamts9 protein was also observed specifically in the follicular cells of zebrafish and was relative high in the follicular cells of wt than in pgr −/− fish in vivo (Figure 5). Exposure of stage IVa follicles to DHP in vitro significantly upregulated adamts9 expression in a dose dependent manner (Figure 6A). The increase of adamts9 correlated well with the occurrence of oocyte maturation in vitro (Figure 6B). DHP-induced adamts9 expression was transient (Figure 6C), peaking at 2 h post incubation with DHP (100 nM) when about 90% of stage IV oocytes had matured (Figure 6D), but then decreased gradually to basal levels after an additional 2 h of in vitro incubation.

RU486 Did Not Affect Expression of adamts9 but Triggered Oocyte Maturation in Follicles in vitro
Exposure to the mammalian P4 antagonist, RU486, did not block DHP-induced oocyte maturation (Figure 7A). RU486 also did not inhibit adamts9 expression induced by DHP ( Figure 7B). Intriguingly, at a high dosage (10 µM) by itself, RU486 induced oocyte maturation, but had no effect on the expression of adamts9 (Figure 7).

Effect of DHP on the Increased Expression of adamts9 Is Blocked in pgr −/− Fish in vitro
As expected, DHP-induced oocyte maturation was not impaired in pgr −/− fish ( Figure 8A) because DHP signaling is mediated via the membrane progestin receptor (mPR) at the oocyte surface (24,29,31,32). However, DHP-induced expression of adamts9 was completely blocked in pgr −/− fish in vitro (Figure 8B).

DHP Downregulates Pgr Protein
Because we observed downregulation of Pgr during oocyte maturation (GVBD) in vivo, we hypothesized this downregulation of receptor is progestin (DHP) specific but not GVBD specific. Therefore, we tested our hypothesis by examining direct effect of DHP, RU486 and T on Pgr expression in an in vitro incubation of follicles. All three steroids induced GVBD in preovulatory follicles (stage IVa) following an 2.5 h in vitro incubation (data not shown). As expected, DHP treatment induced significant reduction of Pgr protein expression, whereas treatments of RU486 or T had no such effect on the expression of Pgr protein, comparable to the vehicle control (Figure 9).

DISCUSSION
In the present study, we provide the first evidence that expression of adam8b, adamts1, adamts9, and mmp9 increased drastically within the preovulatory follicular cells of zebrafish, a basal vertebrate model. A significant increase of Pgr was required for this increase. Using follicles from pgr −/− fish, we showed that expression of adamts9, adam8b, and mmp9 was severely reduced in the knockout. One main cause for anovulation in pgr −/− fish is likely due to this dramatically reduced expression of these metalloproteinases, which was caused by the loss of progestin signal in the Pgr knockout fish. Our study indicates a direct regulation of metalloproteinase by progestin and Pgr in preovulatory follicles, which likely is essential for ovulation in zebrafish.
In this study, we found Pgr was expressed throughout the daily ovulatory cycle in late stage follicles and was significantly elevated at the onset of oocyte maturation in zebrafish. This increase of Pgr corresponds to the transient appearance and increase of PGR prior to ovulation in mammals (2,5,29,(33)(34)(35), supporting the idea of conserved regulation and roles for Pgr in vertebrate ovulation (24,26). Ligand-dependent down-regulation of PGR has been shown in multiple reproductive tissues in mammals (36,37). This down-regulation of PGR was suggested to be concomitant with its transcriptional hyperactivity (38). In fact, production of DHP in zebrafish ovarian tissue increased significantly at 3.5 h prior to lights turned on (39), which could contribute to the down-regulation of Pgr in zebrafish. In the present study, Pgr protein level was downregulated either in vitro by DHP exposure or in vivo likely by endogenous progestin. Our results support previous studies and indicate that the appropriate Pgr expression in preovulatory follicles is required for the increase of metalloproteinase expression in zebrafish.
Suppression of the stimulatory effect of DHP on metalloproteinase expression in Pgr knockout indicates an indispensable role of Pgr in metalloproteinase expression in zebrafish ovary. Interestingly, expression of adamts9 increased significantly during oocyte maturation in vivo in pgr −/− fish, though the magnitude of increase was greatly reduced in pgr −/− fish in comparison to those in wt fish. It has been suggested that LH and PGR have distinct but coordinate effects on transactivation of the Adamts1 gene in mice granulosa cells (40). Whether gonadotropin stimulates adamts9 and other metalloproteinase expression via Pgr-independent signaling pathway deserves further investigation. An interesting finding from the current study is failure of RU486 in inhibition of Pgrdependent and DHP induced adamts9 expression. One possible scenario is that RU486 has low or no binding affinity to zebrafish Pgr. A single amino acid substitution of glycine by cysteine in the hormone binding domain (HBD) of the human PGR abrogates binding of RU486 (41). Alignment of the zebrafish Pgr HBD with other species, including humans and chickens, has shown a glycine substitution by cysteine in the zebrafish Pgr (30,42). Alternatively, instead of acting as an antagonist, RU486 can act as partial agonist (43). Indeed, our study shows that high doses of RU486 can induce oocyte maturation, which suggests that RU486 acts as an agonist of progestin activating mPRα and/or Pgr signaling in zebrafish. Further studies are required for elucidation of actions and molecular mechanisms of RU486 in fish.
Most likely, several proteinases are required for the appropriate remodeling of the extracellular matrix (ECM) and basal membranes that lead to follicular rupture and ovulation in vertebrates. Knockout of PGR, or administration of P4 inhibitors, significantly reduced hCG-induced expression of Adam8 in mice granulosa cells (44). In agreement with this, we observed the expression of adam8b was also dramatically reduced in the follicular cells of pgr −/− zebrafish, suggesting a conserved signaling pathway for the regulation of this metalloproteinase in preovulatory follicular cells. Unfortunately, knockout of Adam8 in mice had no noticeable effect on fertility, which might be due to the overlapping roles of other proteinases as multiple metalloproteinases could target the same substrate (15). MMPs have also been postulated to play critical roles in ECM remodeling associated with ovulation (45). Transcript of MMP9 was elevated in rhesus monkey granulosa cells exposed to LH in vitro (46). Another study also suggested that MMP9 played a critical role in LH-induced steroidogenesis during ovulation in mouse granulosa cells (47). In line with these studies, we found significant increase of mmp9 expression in zebrafish follicular cells prior to ovulation, and this increase was significantly reduced in pgr −/− fish. Increase of Adamts1 transcripts were observed in mice granulosa cells of periovulatory follicles after LH stimulus, but not in small follicles or in denuded oocytes (7). Other studies suggest that ADAMTS1 may cleave versican, a proteoglycan located in the basal follicular region and COC matrix (17). This cleavage of versican promotes changes in the ECM of preovulatory follicle required for the successful ovulation of a fertilizable oocyte in mice. In concert with these findings, our results showed that expression of adamts1 increased remarkably in zebrafish follicular cells after oocyte maturation, supporting its role in ovulation. In human preovulatory granulosa cells, gene expression of ADAMTS9 exhibited a significant upregulation (40-fold) following hCG treatment (48). In cattle, mifepristone inhibited the effects of LH on the expression of ADAMTS9 transcripts in granulosa cells in vitro (8). In this study, we showed expression of adamts9 in zebrafish was extremely low in immature follicles (stage I-IVa), negligible in denuded oocytes, and dramatically increased to its peak level specifically in the follicular cells of mature follicles (stage IVb) prior to ovulation. Additionally, knocking out Pgr indeed severely reduced the expression of adamts9. These results suggest that Adamts9 may also be responsible for follicular rupture. In summary, the preovulatory increases of adam8b, mmp9, adamts1, and adamts9 in follicular cells support the requirement of multiple metalloproteinases for ovulation in zebrafish. Future studies including changes of protein contents and enzyme activities, and the effects of knockouts of these metalloproteinases in zebrafish are required to determine relative importance of each metalloproteinase in ovulation in zebrafish.