Requirement of the 3′-UTR-dependent suppression of DAZL in oocytes for pre-implantation mouse development

Functional oocytes are produced through complex molecular and cellular processes. In particular, the contribution of post-transcriptional gene regulation mediated by RNA-binding proteins (RBPs) is crucial for controlling proper gene expression during this process. DAZL (deleted in azoospermia-like) is one of the RBPs required for the sexual differentiation of primordial germ cells and for the progression of meiosis in ovulated oocytes. However, the involvement of DAZL in the development of follicular oocytes is still unknown. Here, we show that Dazl is translationally suppressed in a 3′-UTR-dependent manner in follicular oocytes, and this suppression is required for normal pre-implantation development. We found that suppression of DAZL occurred in postnatal oocytes concomitant with the formation of primordial follicles, whereas Dazl mRNA was continuously expressed throughout oocyte development, raising the possibility that DAZL is dispensable for the survival and growth of follicular oocytes. Indeed, follicular oocyte-specific knockout of Dazl resulted in the production of normal number of pups. On the other hand, genetically modified female mice that overexpress DAZL produced fewer numbers of pups than the control due to defective pre-implantation development. Our data suggest that post-transcriptional suppression of DAZL in oocytes is an important mechanism controlling gene expression in the development of functional oocytes.


Introduction
The production of functional oocytes is an essential process in the female ovary, by which genetic information is continuously passed to the next generations. For successful oocyte development, gene expression needs to be precisely regulated according to the developmental stages and environmental cues such as gonadotropic hormones. Oocytes that fail to regulate proper gene expression are degenerated during their development or are unable to proceed with embryonic cleavage even if they are fully developed [1]. Therefore, unveiling the mechanisms controlling the quality of oocytes is a crucial issue to understand the molecular basis of female reproduction.
Post-transcriptional gene regulation mediated by RNA-binding proteins is an important molecular mechanism involved in this process. An evolutionarily conserved and well-documented post-transcriptional event is the translational suppression and storage of maternal mRNAs with shorter poly (A) tails [2]. Although the genome is actively transcribed and proteins are produced during oocyte growth, transcription become inactive in full-grown oocytes and maternal mRNAs are used for protein synthesis in early zygote development [3]. These processes are orchestrated by a battery of relevant RNA-binding proteins, including cytoplasmic polyadenylation element binding proteins (CPEB), maskin, and other germ cell-specific RNA-binding proteins [4,5,6,7]. In addition to the maturation process, germ cell-specific RNA-binding proteins are also responsible for multiple processes in oogenesis. For instance, CPEB1 and Pumilio1 are involved the progression of meiotic prophase I in the embryonic ovary [8,9], and MSY2 is involved in follicle development after birth in mice [10,11], suggesting the significant contribution of post-transcriptional gene regulation throughout oogenesis.
Deleted in azoospermia-like (DAZL) is a member of the evolutionarily conserved DAZ family of RNA-binding proteins that acts as a translational activator in mice [12]. Biochemical and structural analyses showed that DAZL binds to the U-rich region of its target's 3 0 -UTR [13,14,15], and genetic analyses revealed that DAZL is indispensable for gametogenesis in both males and females [16,17,18]. As DAZL is reportedly required for sexual differentiation of primordial germ cells, progression of meiotic prophase I in embryonic female germ cells [19], and for the progression of meiosis in maturating oocytes [20], it is believed that DAZL is involved in female germ cell development throughout oogenesis. However, the role of DAZL in follicular oocytes remains unknown because Dazl-deficient oocytes die due to the failure of meiotic progression in the embryonic ovary [16,21]. Moreover, although the previous immunohistochemical analysis demonstrated that DAZL was expressed in both embryonic and follicular oocytes in postnatal ovaries [16], it was also noted that DAZL signals were not detectable by western blotting in ovaries 1 to 2 weeks after birth [22]. Therefore, further analysis is required to clarify this contradiction.
In this study, we investigated DAZL expression in embryonic and postnatal ovaries, and found that DAZL was translationally suppressed in a 3 0 -UTR-dependent manner in follicular oocytes. Genetic analysis by knocking out the Dazl gene in a follicular oocyte-specific manner indicated that Dazl is dispensable for follicular growth, maturation, and fertilization. On the other hand, the 3 0 -UTR-dependent suppression of DAZL in follicular oocytes is required for the progression of normal pre-implantation development. Our data clarify the previously ambiguous expression pattern of DAZL in the postnatal ovary, and simultaneously demonstrate the significance of the post-transcriptional suppression of DAZL in follicular oocytes.

DAZL is post-transcriptionally suppressed in postnatal oocytes
To examine Dazl/DAZL expression in detail, we performed quantitative reverse transcriptionpolymerase chain reaction (RT-qPCR) and western blotting analyses using ovaries from embryos until juvenile stages (Fig 1A and 1B). RT-qPCR data showed that Dazl mRNA was constantly expressed and expression differences were less than 2-fold among all stages investigated ( Fig 1A). On the other hand, DAZL expression was markedly changed during oocyte development ( Fig 1B). Although it was abundant in embryonic ovaries, with the strongest expression at embryonic day (E) 15.5, the expression declined in newborn ovaries. Afterwards, DAZL expression further declined and was hardly detectable in 1-and 2-week ovaries, which is consistent with previous descriptions [22].
As DAZL expression was significantly decreased in the newborn ovary and onward, we next asked whether this reduction in DAZL was correlated with the formation of primordial follicles. In the embryonic ovary, oocytes are connected with each other by intercellular bridges (ICBs). Within a few days after birth, ICBs are broken and each oocyte is enclosed by pregranulosa cells, resulting in the formation of primordial follicles [23]. Thus, we examined the expression changes of DAZL in perinatal ovaries by immunostaining DAZL together with a granulosa cell marker, forkhead box protein L2 (FOXL2), from E18.5 to 1W ovaries. Strong DAZL signals were observed in most oocytes until the day of birth (P0), when a large number of oocytes were still connected with each other. However, at one day after birth (P1), its expression began to decrease in some oocytes (Fig 1C, open arrowheads). The expression of DAZL was further decreased at two days after birth (P2), at which point, many oocytes formed primordial follicles (Fig 1C, yellow arrowheads) and exhibited weaker expression than cysticoocytes ( Fig 1C, white arrowheads). Thereafter, the weakened DAZL expression was observed in 1-week ovaries. These data suggest that DAZL is decreases in oocytes coinciding with the development of primordial follicles.

DAZL is not required for oogenesis in the postnatal ovary
Both immunostaining and western blotting analyses revealed that DAZL was decreased in oocytes shortly after birth, which raised the possibility that DAZL is dispensable for follicular development. To test this possibility, we used conditional Dazl knockout (cKO) mice (Fig 2A). Dazl flox mice were crossed with a postnatal oocyte-specific Cre mouse line, Gdf9-iCre, which expresses improved Cre recombinase from P2 oocytes [25]. The Dazl gene was successfully disrupted by Gdf9-iCre, as evidenced by RT-qPCR and western blotting, in which both Dazl mRNA and DAZL protein were hardly detectable in Dazl cKO ovaries ( Fig 2B). We also confirmed that our Dazl KO (Dazl 1lox/1lox ) mouse line recapitulated the phenotype of previous Dazl knockout females (S1 Fig) [16]. On histological analysis, Dazl cKO ovaries as well as control ovaries contained both primordial and growing follicles ( Fig 2C). Notably, cKO ovaries did not have any significant differences in the number of primordial or growing follicles ( Fig  2D). These data suggest that DAZL is dispensable for the survival and growth of follicular oocytes.
In order to evaluate the reproductive capability of Dazl cKO oocytes, we next crossed Dazl cKO females with wild-type (WT) males. We found that Dazl cKO females were fertile and produced a normal number of pups ( Fig 2E). The average litter size delivered from Dazl cKO females (11.3±0.62) was almost identical with that from WT (12.0±2.83). We also confirmed that all progeny delivered by Dazl cKO females were heterozygotes for the Dazl 1lox allele (n = 258). These results were surprising because a previous report stated that Dazl knockdown in MII oocytes results in the defective progression of the oocyte to zygote transition [20]. However, MII oocytes derived from Dazl cKO females did not have abnormal spindle morphology (S2 Fig). These data indicate that DAZL is not required for the maturation of oocytes or subsequent fertilization. . Mvh (also known as Ddx4) was used as a normalizer because this gene is continuously expressed in germ cells from E12.5 to secondary follicles [24]. The vertical axis represents relative expression level of Dazl to Mvh. Error bars represent S.D. (B) Western blotting analysis for DAZL from E12.5 to 2W ovaries. Two bands representing DAZL were detected. As the molecular weight of DAZL is estimated to be 33KDa, the lower band corresponds with this. However, both bands disappear in Dazl knockout ovaries [25], indicating that both two bands are DAZL signals. Quantification of western data is shown below. The vertical axis represents relative DAZL expression level normalized by MVH. (C) Immunofluorescence analysis for DAZL (green) and FOXL2 (magenta) from E18.5 to P2, and 1W ovaries. White arrowheads indicate cystic oocytes, and yellow arrowheads indicate primordial follicles. Open arrowheads indicate oocytes showing weaker DAZL expression. Scale bar, 30μm. https://doi.org/10.1371/journal.pgen.1007436.g001

DAZL is suppressed in a 3 0 -UTR dependent manner
In embryonic male germ cells, Dazl is post-transcriptionally suppressed in a 3 0 -UTR -dependent manner by a male-specific RNA-binding protein, NANOS2 [26]. As DAZL decreases in postnatal ovaries, it is possible that Dazl is also post-transcriptionally suppressed in a 3 0 -UTR -dependent manner by unidentified mechanisms in follicular oocytes. In order to test this possibility, we used our bacterial artificial chromosome (BAC)-carrying transgenic mouse line, in which the FLAG tag was inserted at the C-terminus of Dazl and the Dazl 3 0 -UTR was flanked with Frt sequences (Dazl 3F, Fig 3A upper) [26]. The significance of the Dazl 3 0 -UTR for its expression was assessed by crossing the BAC transgenic female with a Rosa-Flp male (Dazl 3F; Flp, Fig 3A lower). RT-qPCR showed that the amount of Flag-Dazl mRNA was increased in Dazl 3F;Flp ovaries after birth ( Fig 3B). However, the effect of removing the 3 0 -UTR was not clear because the difference in Flag-Dazl mRNA expression levels between Dazl 3F and Dazl 3F;Flp was less than 2-fold, and the total Dazl expression level (Flag-Dazl +endogenous Dazl) was not changed between Dazl 3F and Dazl 3F;Flp except in the P0 ovary (Fig 3B and 3C). In contrast to the small increase in the mRNA level, FLAG-DAZL expression was greatly increased after birth (Fig 3D). Although FLAG-DAZL (filled arrowheads) decreased in Dazl 3F ovaries from P0 onward, which was consistent with the reduction in endogenous DAZL (open arrowhead), its expression was continuously observed in P0, 1W, and 2W ovaries when the 3 0 -UTR was removed. Quantification of FLAG-DAZL expression revealed that its expression increased 20-fold in Dazl 3F;Flp at P0 (Fig 3E). The results of western blotting were also supported by immunostaining. Both total-and FLAG-DAZL expression was strongly observed in Dazl 3F;Flp ovaries ( Fig 3F and S3 Fig), whereas their expression levels in WT and Dazl 3F ovaries were comparable with those in Dazl cKO ovaries. Furthermore, strong DAZL expression was observed in all stages of follicular oocytes in Dazl 3F;Flp ovaries (S3 Fig). These data indicate that DAZL is post-transcriptionally suppressed in a 3 0 -UTR-dependent manner in follicular oocytes.

Role of DAZL suppression in female reproduction
To investigate the role of 3 0 -UTR-dependent DAZL suppression in female reproduction, we crossed BAC transgenic females with WT males when female mice reached 6 weeks old. Each pair was kept in a breeding cage until female mice became 30 weeks old, and the number of pups delivered during this period was counted. We found that BAC transgenic females were fertile regardless of the presence or absence of the Dazl 3 0 -UTR (Fig 4A). The number of total pups was slightly lower by in Dazl3F females (39.8±5.5, n = 5) compared with control females (53.1±9.1, n = 7). Interestingly, Dazl 3F;Flp females produced less than half the normal number of pups (18.6±11.3, n = 5). As 3FLAG-DAZL protein rescued the germless phenotype in Dazl -/mice [26], it is unlikely that the observed litter size reduction was caused by the expression of 3FLAG-DAZL. Thus, these results suggest that DAZL overexpression results in litter size reduction. We next analyzed the number of deliveries and the number of pups in each delivery. The number of pups in each delivery was fewer by Dazl 3F and Dazl 3F;Flp mice (Fig Western blotting analysis for WT and Dazl cKO MII oocytes (upper panels), and RT-qPCR analysis for Dazl in 1 and 5W ovaries using a primer set that amplifies only the WT allele (lower graph, n = 3). Error bars represents S.D. Significance level of changes are indicated (two-tailed student's t-test; ÃÃ P < 0.005). (C) Periodic acid-Schiff (PAS) staining of control (left) and cKO (right) ovaries at 5W after birth. PrF, primordial follicle; PF, primary follicle; SF, secondary follicle; and AF, antral follicle. Scale bars, 100μm. (D) Follicle counting analysis of control (white bar, n = 9) and Dazl cKO (black bar, n = 6) ovaries at 5 weeks after birth. Error bars represents S.D. ns, no significant difference between control and Dazl cKO ovaries (two-tailed student's t-test). (E) Average litter size of control (n = 3) and Dazl cKO females (n = 5). Error bars represent S.D. ns, same as in (D).
https://doi.org/10.1371/journal.pgen.1007436.g002 4B), but the number of deliveries was not significantly different among genotypes (Fig 4C). These results suggest that the reduced female fecundity was due to defects during follicular development, fertilization, or zygote development after fertilization, but not to the shortened reproductive lifespan.

Excess DAZL is deleterious for pre-implantation development
To determine the cause of the litter size reduction in the DAZL overexpressing females, we examined the development of oocytes, fertilization, and pre-implantation development. Histological analysis revealed that BAC transgenic ovaries did not have significantly different numbers of primordial, primary, secondary or antral follicles compared with WT ovaries (Fig 5A  and 5B). We next asked whether ovulation normally occurs by counting the number of onecell embryos ovulated. However, the number was not significantly different among the genotypes ( Fig 5C). These data suggest that folliculogenesis and subsequent ovulation proceeds normally even in BAC transgenic females. Thus, to examine whether these ovulated eggs developed normally, we measured the proportion of blastocysts by flushing E3.5 embryos from oviducts. We cultured the collected embryos for a further two days and then counted the embryos because the different timing of sexual behavior in each mouse pair influences the progression of early embryonic development (Fig 5D). We found that only 56.1% of embryos derived from Dazl 3F;Flp females developed into blastocysts, whereas more than 97.3 and 97.1% embryos derived from control and Dazl 3F females became blastocysts, respectively.

Discussion
In this study, we demonstrated that DAZL expression is post-transcriptionally suppressed in a 3 0 -UTR-dependent manner in postnatal oocytes. Although DAZL has been thought to function in postnatal oocytes, our data suggest that DAZL is not required for postnatal oocyte development. Supporting this idea, analysis of conditional Dazl knockout mice revealed that   It was previously reported that DAZL was expressed in growing oocytes [16], but a later study stated that DAZL was not detectable in the postnatal ovary [22]. Therefore, it has been unclear whether DAZL plays a role in follicular oocytes. Our results answered this question; DAZL expression is suppressed in follicular oocytes and is dispensable for oogenesis after birth. Interestingly, this suppression coincides with the formation of primordial follicles. As Dazl mRNA was continuously expressed in oocytes regardless of developmental stage, it is likely that post-transcriptional gene regulatory mechanisms are altered between cystic oocytes and follicular oocytes. Importantly, DAZL suppression requires its 3 0 -UTR, suggesting the presence of some mechanisms regulating DAZL expression in postnatal oocytes. In general, post-transcriptional regulation is conducted by microRNAs and RNA-binding proteins [27]. However, it was reported that the function of microRNA is globally suppressed in oocytes and early embryonic development [28]. Thus, it is possible that Dazl expression is regulated by some RNA-binding proteins (RBPs). One possible candidate RBP for DAZL suppression is CPEB1, a mammalian ortholog of Xenopus CPEB. CPEB acts as both a translational activator and suppressor of its target mRNAs depending on its phosphorylation state [29,30]. CPEB1 is expressed in postnatal oocytes and promotes the translation of Dazl in MII oocytes [20], thus it may suppress Dazl in follicular oocytes. Further expression and functional analyses, including the phosphorylation state, of CPEB1 are required to address this question.
Our oocyte-specific Dazl KO females exhibited no ovarian developmental defect and the MII oocytes had no spindle abnormalities. Furthermore, the Dazl cKO females produced normal numbers of pups. These observations were inconsistent with Chen and colleague's results that DAZL depletion in MII oocytes results in defective spindle formation in meiosis II [20]. One possible explanation for this contradiction is the method of gene depletion. We used the Cre-loxP system for Dazl cKO in vivo, whereas Chen et al. used morpholino knock-down in MII oocytes. A recent zebrafish report found that approximately 80% of phenotypes induced by morpholino do not correlate with mutant phenotypes induced by ZFN, TALEN or CRISPER/Cas9; therefore, the above-mentioned knock-down phenotype may have emerged due to indirect effects [31]. Alternatively, it is possible that some system that compensates for DAZL function works in Dazl-cKO MII oocytes because a previous study reported that the activation of a compensation system rescued deleterious mutations, which was not observed after translational or transcriptional knockdown [32]. Further analysis is required to evaluate the contribution of RNA-binding proteins for the progression of meiosis II.
Although DAZL expression is suppressed after birth, introducing the BAC transgenic allele in the Dazl +/+ background reduced litter size even in the presence of the 3 0 -UTR (Fig 4). As a previous study reported that Dazl dosage in females influences their litter size, and Dazl +/− females produced more pups than Dazl +/+ females [22], the slight reduction in litter size by our Dazl 3F mice may be attributed to the dosage effect. However, our histological and embryo culture experiments did not reveal any abnormalities in Dazl 3F mice. In addition, we were unable to observe obvious differences in resorption after implantation. One possible explanation is that insertion of the BAC transgene influences female reproduction.
Our results suggest that the suppression of Dazl translation in follicular oocytes is required for producing the proper number of progeny. However, why excess DAZL expression causes defective pre-implantation development remains still unclear. DAZL has been implicated in embryos. Scale bars, 100μm. (Right) Proportion of embryos that developed to each stage up to blastocysts. Embryos collected from pregnant females at E3.5 were counted. Delayed embryos were cultured for an additional two days and ones that reached the blastocyst stage were added. Control (n = 75), Dazl 3F (n = 69) and Dazl 3F;Flp (n = 56). https://doi.org/10.1371/journal.pgen.1007436.g005 Role of post-transcriptional suppression of Dazl in oocytes the positive regulation of translation [14,33,34], thus it is possible that the observed defect may be due to abnormal translational promotion. Alternatively, it is also possible that excess DAZL abnormally suppresses its target RNAs because DAZL works as a component of stress granules, cytoplasmic RNP granules involved in translational suppression or mRNA storage, in the testis [35,36]. Therefore, it is likely that suppression of DAZL expression in follicular oocytes is an important molecular mechanism for controlling proper gene expression.

Ethics statement
All mouse experiments were approved by the Animal Experimentation Committee at the National Institute of Genetics (approval number  and Yokohama National University (approval number 2017-09) and conducted under the Regulations for Animal Experiments at the National Institute of Genetics, Research Organization of Information and Systems and the guideline at Yokohama National University.

Mice
Mice were housed in a specific-pathogen-free animal care facility at the National Institute of Genetics (NIG). All experiments were approved by the NIG Institutional Animal Care and Use Committee and the animal experimental committee at Yokohama National University. The genetic background of mice used in this study was C57BL/6N (Clea Japan), except in the DAZL expression analysis and conditional Dazl knockout mice (mixed genetic background of ICR and C57BL/6N). The BAC-carrying transgenic mouse line was generated in a previous study [26]. The BAC transgenic mice were backcrossed with C57BL/6N at least 3 times. Dazl flox mice were generated from an ES cell line produced by the Knock Out Mouse Project (KOMP, Dazl tm2a (KOMP) Wtsi ).

Reverse transcription-quantitative polymerase chain reaction (RT-qPCR)
Total RNAs were isolated from whole gonads of wild-type and BAC transgenic mice at each stage by RNeasy Mini Kit (Qiagen). One hundred ng (1W to 5W) and 40 ng (E12.5 to P0) of total RNA were used for cDNA synthesis using Prime Script RT Reagent Kits with gDNA Erase according to the manufacturer's protocol (Takara). Real time PCR was performed with KAPA SYBR FAST qPCR kits using a thermal cycle dice real time system (Takara). The obtained data was normalized by Mvh.
The following primers were used for PCR amplification:

Immunostaining of MII oocytes
MII oocytes were fixed with MeOH at -20˚C for 3 minutes, washed with PBS-TX (0.1%Tri-tonX, PBS), and were then incubated with blocking solution (3%BSA, 0.1%TritonX, PBS) at 4 o C for 3 hours. Primary antibody reactions were performed with the following dilutions (Antiα-tubulin antibody, Sigma, 1:1000) at 4˚C overnight. After washing with PBS-TX, secondary antibody reaction was performed with the following dilutions (Alexa 488 Donkey anti-Rabbit, Life technologies, 1:1000) and DAPI at RT for 60 min. Then, oocytes were washed with PBS-TX. Fluorescent images were obtained using confocal microscopy FV1200 (Olympus).

Histological analysis
Histological analysis was carried out by PAS (Periodic acid-Schiff) staining according to the standard protocol. Briefly, ovaries were fixed in Bouin solution, embedded in paraffin wax, and sliced at 6-μm thickness. The sections were submerged in xylene, 100%, 90%, 70% ethanol, and distilled water at RT, and stained with PAS solution. Ovarian images were obtained with an inverted microscope BX 51 and 61(Olympus). Follicle stages were counted on every 5 sections.

Litter size investigation
Dazl +/+ , Dazl 3F, and Dazl 3F;Flp females at 6 weeks old were crossed with C57BL/6N males, and kept together until female mice reached 30 weeks old. The number of pups and deliveries was recorded. Pups were removed after counting the number and sex. Females that killed their pups were excluded from the analysis.

Collection of MII oocytes, 1-cell, 2-cell embryos and blastocysts
To obtain MII, 1-cell and 2-cell oocytes for western blotting, female mice were injected with PMSG (ASKA Pharmaceutical). Forty-eight hours after PMSG injection, mice were stimulated with hCG (ASKA Pharmaceutical) for 14 h and MII oocytes were collected. To obtain western blotting samples of 1-cell and 2-cell embryos, each female was crossed with a WT male after hCG injection. Eggs with obvious abnormalities were removed from experiments. One-cell embryos for investigation of ovulation number and pre-implantation development investigation were obtained from the ampulla of pregnant females at E0.5. Blastocysts for examining progression of early embryonic development were obtained by flushing oviducts at E3.5. Collected blastocysts were cultured for two days in KSOM medium (Ark resource).

Statistical analysis
Significance was assessed by the Student's t-test for differences between two samples. For quantitative analyses among multiple samples, significance was assessed using one-way ANOVA followed by Tukey HSD (Honest Significant Difference) test. Asterisks in figures indicate significance: Ã P < 0.05, ÃÃ P < 0.005.