Abstract

   DEAD-box RNA helicase 3 (DDX3) and its homologs play a vital role in
   translation initiation by unwinding secondary structures of selected
   mRNAs. The human DDX3 gene is located on the sex chromosomes, so there
   are DDX3X and DDX3Y. DDX3X is ubiquitously expressed in almost all
   tissues and critical for embryonic development, whereas DDX3Y is only
   expressed in the testis and essential for male fertility. Drosophila
   belle (bel) is the single ortholog of DDX3, and mutations in bel cause
   male and female infertility. Using Drosophila bel mutants and Ddx3x
   conditional knockout (cKO) mice, we confirmed the pivotal role of DDX3
   in female fertility and ovarian development. Drosophila bel mutants
   exhibited female infertility and immature egg chambers. Consistently,
   oocyte-specific Ddx3x knockout in mice resulted in female infertility
   and impaired oogenesis. We further found that immature egg chambers in
   Drosophila bel mutants and impaired follicular development in
   oocyte-specific Ddx3x cKO mice were caused by excessive apoptosis. We
   also identified a set of DDX3 target genes involved in oocyte meiosis
   and maturation and demonstrated that DDX3 is involved in their
   translation in human cells. Our results suggest that DDX3 is critical
   for female fertility via translational control in oogenesis.

   Subject terms: Mechanisms of disease, Infertility

Introduction

   The human DEAD-box protein DDX3 plays a crucial role in cellular RNA
   metabolism. DDX3 and its homologs have been implicated in multiple
   cellular processes, including translation initiation [[38]1–[39]6]. The
   yeast DDX3 homolog, Ded1, is required for translation of bulk mRNAs
   [[40]1]. However, human DDX3 may affect the translation of only a
   subset of mRNAs. We previously reported that DDX3 is required for the
   translation of selected mRNAs that contain a long or structured 5′ UTR
   in human cells [[41]3, [42]6]. Consistently, depletion of DDX3 in
   HCT116 cells represses translation of mRNAs with complex 5′ UTRs, such
   as those with high GC content or containing the cytosine-enriched
   regulator of translation (CERT) motif [[43]5]. Given that the helicase
   activity of DDX3 is required for its function in translation [[44]3],
   DDX3 may facilitate ribosome scanning by resolving secondary structures
   in the 5′ UTR of selected mRNAs.

   DDX3 is a multifunctional protein implicated in a variety of biological
   processes, including cell cycle control [[45]6–[46]8], tumorigenesis
   and cancer progression [[47]9–[48]12], cell apoptosis [[49]13–[50]15],
   viral replication [[51]16–[52]18], innate immunity [[53]19–[54]21], and
   neurodevelopment [[55]22–[56]24]. The human DDX3 genes are located on
   both the X and Y chromosomes, so there are DDX3X and DDX3Y. Although
   DDX3X and DDX3Y share ~92% of the protein sequence identity and can
   functionally complement each other in translation [[57]25], they show
   distinct expression patterns and biological functions [[58]26]. DDX3X
   is ubiquitously expressed in almost all tissues and critical for
   embryonic development [[59]7, [60]27], whereas DDX3Y is only expressed
   in the testis, especially in spermatogonia and early spermatocytes, and
   essential for male fertility [[61]28]. The most common male infertility
   cases are caused by deletions in the Y chromosome’s azoospermia factor
   (AZF) regions [[62]29]. Deletion analysis of the Y chromosome revealed
   three common deletions: AZFa, AZFb, and AZFc [[63]30]. AZFa deletion
   causes the most severe azoospermia phenotype, a complete absence of
   germ cells in the seminiferous tubules of the testis, and Sertoli
   cell-only (SCO) syndrome in humans [[64]31]. The human AZFa region
   contains three genes: USP9Y, DDX3Y, and UTY [[65]32]. Uty knockout in
   mice did not affect fertility [[66]33]. Loss of the human USP9Y gene
   did not cause male infertility [[67]34], suggesting that USP9Y is not
   the causative gene. It has been reported that DDX3Y has a higher
   mutation rate in SCO syndrome than the other two genes in the AZFa
   region [[68]35]. Overexpression of DDX3Y in AZFa-deleted iPSCs restored
   germ cell formation [[69]36]. Therefore, DDX3Y is regarded as the major
   gene in AZFa deletion-induced spermatogenic failure [[70]35].

   Mouse Ddx3x mRNA is predominantly expressed in the ovary and embryo
   [[71]37], implying that Ddx3x may play a role in the ovary and embryo.
   Conventional knockout of Ddx3x in mice causes early embryonic lethality
   [[72]27]. Mouse Ddx3x protein is expressed in germinal vesicle (GV)
   oocytes, and its expression is remarkably increased in metaphase II
   oocytes [[73]7]. This finding suggests a role for mouse Ddx3x in oocyte
   development. Drosophila belle (bel), the single ortholog of human DDX3,
   is required for larval growth and can functionally substitute for the
   yeast Ded1 in vivo [[74]38], suggesting the functional conservation of
   DDX3 homologs. It has been reported that Drosophila bel mutants are
   male- and female-sterile [[75]38]. Ectopic expression of bel ATPase
   mutants results in decreases in female fertility and ovary mass
   [[76]39]. A recent study showed that DDX3 is required for ovarian
   development and oocyte maturation in Locusta migratoria [[77]40]. These
   results reveal an evolutionarily conserved role of DDX3 in the
   development of germ cells. However, the molecular mechanisms of DDX3 in
   gametogenesis (spermatogenesis and oogenesis) remain largely unclear.

   Here, we used Drosophila bel mutants and Ddx3x conditional knockout
   (cKO) mouse models to confirm the importance of DDX3 in female
   fertility and ovarian development. We previously identified many target
   mRNAs whose translation is regulated by DDX3 in HeLa cells [[78]6].
   Pathway enrichment analysis showed that ten candidate DDX3 target genes
   are involved in oocyte meiosis, oocyte maturation, and the
   gonadotropin-releasing hormone (GnRH) signaling pathway. Therefore, we
   assume that DDX3 is critical for female fertility via translational
   control in oogenesis.

Results

Drosophila bel mutants are female-sterile

   Drosophila bel is the single ortholog of human DDX3. It has been
   reported that Drosophila bel mutations cause male and female
   infertility [[79]38, [80]39]. We first need to verify the experimental
   results. The bel^neo30 is a hypomorphic mutation with a P-element
   insertion at the first intron of the bel gene. The bel^6 is a lethal
   mutation with a point mutation that causes a premature stop codon. To
   verify whether Drosophila bel is required for female fertility, the
   bel^neo30/TM6B heterozygotes and the bel^6/TM6B heterozygotes were
   crossed to produce the trans-heterozygous allelic combination
   bel^neo30/bel^6. Western blot analysis showed that the expression of
   bel/DDX3 protein was markedly decreased in Drosophila bel^neo30/bel^6
   female mutants (Fig. [81]1A). Compared to the w^1118 strain and
   heterozygous mutations (bel^neo30/TM6B and bel^6/TM6B), Drosophila
   bel^neo30/bel^6 females laid a lot fewer eggs, and did not produce any
   viable adults (Fig. [82]1B). The results indicate that Drosophila
   bel^neo30/bel^6 mutants are female-sterile.

Fig. 1. Mutations in Drosophila bel cause female infertility and ovarian
agenesis.

   [83]Fig. 1
   [84]Open in a new tab

   A Adult females were lysed in 1× RIPA Lysis Buffer and subjected to
   western blot analysis using antibodies against DDX3 and α-tubulin
   proteins. Detection of α-tubulin served as a loading control. B Three
   virgin females were mated with two w^1118 males for 5 days. Fecundity
   was assessed by the number of eggs laid within 24 h. Fertility was
   evaluated by the number of offspring after 19 days of ovulation. The
   strain w^1118 served as a control. Data are shown as mean and standard
   deviation (n = 10/group). C Adult females were fed diets containing
   yeast for 3 days and then dissected to collect their ovaries. Body
   length and ovary size were measured and recorded under a dissecting
   microscope. Data are shown as mean and standard deviation
   (n = 10/group). Statistical significance was determined using the
   Student’s t-test (ns not significant; **p < 0.01; ****p < 0.0001).

Ovarian development is impaired in Drosophila bel mutants

   In Drosophila, body size is one of the most determining traits in
   sexual selection [[85]41]. The size of the Drosophila body was measured
   and recorded under a dissecting microscope. There was no significant
   change in the mean length of the Drosophila body in bel^neo30/bel^6
   female mutants compared to the w^1118 strain or single-allele mutations
   (bel^neo30/TM6B and bel^6/TM6B) (Fig. [86]1C). To understand why
   Drosophila bel^neo30/bel^6 mutants are female-sterile, we dissected
   adult females to observe their ovaries. An adult female has two
   ovaries. Notably, we observed that the size and volume of the ovaries
   were much smaller in bel^neo30/bel^6 female mutants compared to the
   w^1118 strain or single-allele mutations (Fig. [87]1C). The results
   suggest that Drosophila bel/DDX3 plays a critical role in ovarian
   development.

Apoptosis results in immature egg chambers in Drosophila bel mutants

   To explore why Drosophila bel^neo30/bel^6 mutants lose female
   fertility, dissected ovaries were stained with DAPI and observed using
   a fluorescence microscope. We did not find mature oocytes in
   bel^neo30/bel^6 (Fig. [88]2A). The disruption of nuclei in nurse cells
   (NCs) of egg chambers was noticeable (Fig. [89]2A). We therefore
   speculate that apoptosis of egg chambers may occur in the development
   of the bel^neo30/bel^6 ovaries, thus resulting in the arrest of oocyte
   maturation. To test this hypothesis, we performed immunofluorescence
   (IF) and terminal deoxynucleotidyl transferase-mediated dUTP nick-end
   labeling (TUNEL) staining to detect apoptosis in Drosophila ovaries.
   Cleaved Dcp-1(cDcp-1) in Drosophila, a homolog of mammalian caspase 3,
   is a specific marker of apoptosis [[90]42]. Apoptotic cells can be
   labeled with cDcp-1 and TUNEL in Drosophila tissues [[91]43]. The
   results of IF and TUNEL staining showed that massive apoptosis of egg
   chambers occurs in the bel^neo30/bel^6 ovaries compared to the w^1118
   ovaries (Fig. [92]2B, C). Therefore, mutations in Drosophila bel
   resulted in apoptosis of egg chambers.

Fig. 2. Immature egg chambers and apoptosis in Drosophila bel^neo30/bel^6
mutants.

   [93]Fig. 2
   [94]Open in a new tab

   Adult females (w^1118 and bel^neo30/bel^6) were fed diets containing
   yeast for 3 days and then dissected to collect their ovaries. Ovaries
   were fixed with 4% formaldehyde and washed with 1× PBS containing 0.3%
   Triton X-100. A DAPI staining shows the presence of nuclei in
   Drosophila ovaries. NCs: nurse cells. B Ovaries were permeabilized with
   1% Triton X-100. Immunofluorescence staining was performed using
   antibodies against cleaved Dcp-1 (cDcp-1), commonly used as a marker of
   apoptosis in Drosophila. DAPI staining shows the presence of nuclei in
   ovaries. C Ovaries were permeabilized with 1% Triton X-100. TUNEL assay
   was performed to detect apoptotic DNA fragmentation in Drosophila
   ovaries. DAPI staining shows the presence of nuclei in ovaries. Scale
   bars, 100 μm.

Generation of oocyte-specific Ddx3x conditional knockout (cKO) mice

   As mentioned above, conventional knockout of Ddx3x in mice results in
   early embryonic lethality [[95]27]. We therefore used oocyte-specific
   Ddx3x cKO mice to study the importance of Ddx3x in oogenesis. A Ddx3x
   cKO mouse model, in which the Ddx3x gene is flanked with loxP sites
   using a CRISPR/Cas9-based gene editing, has been established (Fig.
   [96]3A). According to the Cre-loxP system [[97]44], gene deletion can
   be precisely controlled by spatiotemporally defined Cre recombinase
   expression [[98]45]. A transgenic mouse line Zp3-Cre in which Cre
   recombinase expression is controlled by the promoter of the zona
   pellucida 3 (Zp3) gene, which is exclusively expressed in the growing
   oocytes [[99]46]. The Zp3-Cre line is used to delete the specific gene
   in oocytes [[100]47]. To generate oocyte-specific Ddx3x cKO female mice
   (Ddx3x^loxP/loxP; Zp3-Cre), heterozygous Ddx3x^loxP mice were inbred to
   produce homozygous Ddx3x^loxP/loxP female mice and then crossed with
   Zp3-Cre male mice (Fig. [101]3B). Oocyte-specific Ddx3x cKO female mice
   were confirmed by PCR-based genotyping (Fig. [102]3C).

Fig. 3. Oocyte-specific Ddx3x cKO mice are generated and genotyped by PCR
analysis.

   [103]Fig. 3
   [104]Open in a new tab

   A The loxP site (ATAACTTCGTATAATGTATGCTATACGAAGTTAT), followed by an
   ApaI restriction site (GGGCCC), was inserted into intron 1 and intron
   10 of the Ddx3x gene using a CRISPR/Cas9-based gene editing strategy.
   The length of a PCR product with a loxP site will increase by 40 bp,
   and it can be cut into two fragments by the ApaI restriction enzyme. B
   Diagram of breeding strategy for generation of Cre-loxP-mediated
   oocyte-specific Ddx3x cKO mice. C Homozygous Ddx3x^loxP/loxP female
   mice were detected by PCR product (636 bp) compared to wild-type mice
   (596 bp). Zp3-Cre mice were detected by PCR product (~300 bp). Among
   the 13 female mice, 7 were oocyte-specific Ddx3x cKO mice
   (Ddx3x^loxP/loxP; Zp3-Cre) (red circle).

Female infertility in oocyte-specific Ddx3x cKO mice

   To examine whether oocyte-specific deletion of Ddx3x affects female
   fertility, the fertility test was carried out twice to reduce
   experimental errors. The number of offspring produced by mating
   10-week-old Ddx3x^loxP/loxP or Ddx3x^loxP/loxP; Zp3-Cre female mice
   with normal C57BL/6 male mice were counted. Male mice were randomly
   assigned to female mice. A female mouse is considered to be fertile if
   she gives birth to pups. The number of offspring from each pregnancy
   was recorded after birth. The average litter size was evaluated in
   Ddx3x^loxP/loxP; Zp3-Cre female mice compared to Ddx3x^loxP/loxP female
   mice. As expected, Ddx3x^loxP/loxP; Zp3-Cre female mice did not give
   birth to any pups, whereas Ddx3x^loxP/loxP female mice had an average
   of 7–8 pups per litter (Fig. [105]4A). The results indicated that
   oocyte-specific Ddx3x cKO mice are female-sterile.

Fig. 4. Female infertility and impaired ovarian development in
oocyte-specific Ddx3x cKO mice.

   [106]Fig. 4
   [107]Open in a new tab

   A Ten-week-old Ddx3x^loxP/loxP female mice (n = 9) and Ddx3x^loxP/loxP;
   Zp3-Cre female mice (n = 10) were mated with wild-type C57BL/6 male
   mice for 10 days (1 female:1 male). Females were then separated from
   the males and allowed to rest for 21 days, the average gestation period
   in mice. The number of pups was counted after birth. Fertility tests
   were carried out twice. Data are shown as mean ± SD. B Ten-week-old
   Ddx3x^loxP/loxP female mice (n = 6) and Ddx3x^loxP/loxP; Zp3-Cre female
   mice (n = 6) were dissected to collect their uteruses (with ovaries).
   The weight of the body and uterus (with ovaries) was measured and
   recorded. The ratio of uterus weight to body weight of mice was
   calculated. Data are shown as mean ± SD. C Experimental mice are as
   described in (B). Mouse ovaries were observed under a dissecting
   microscope. Representative images of mouse ovaries were shown. D
   Experimental mice are as described in (B). Mouse ovaries were fixed in
   formalin, embedded in paraffin, and then sectioned at a thickness of 5
   μm. Ovarian sections were stained with hematoxylin and eosin (H&E).
   Representative images of H&E-stained ovarian sections were shown. E
   Experimental mice were prepared and treated as described in (D). The
   number of follicles in the ovarian sections at different stages
   (primordial, primary, secondary, and antral) was counted. The bar graph
   shows the number of follicles at different stages in the ovaries of
   oocyte-specific Ddx3x cKO mice (Ddx3x^loxP/loxP; Zp3-Cre) compared to
   control mice (Ddx3x^loxP/loxP). Data are shown as mean and standard
   deviation (n = 6/group). Statistical significance was determined using
   the Student’s t-test (ns not significant; *p < 0.05; **p < 0.01;
   ***p < 0.001).

Follicular development is impaired in oocyte-specific Ddx3x cKO mice

   There was no significant difference in appearance and body weight
   between Ddx3x^loxP/loxP and Ddx3x^loxP/loxP; Zp3-Cre female mice (Fig.
   [108]4B). To understand why oocyte-specific Ddx3x cKO mice lose female
   fertility, we dissected Ddx3x^loxP/loxP and Ddx3x^loxP/loxP; Zp3-Cre
   female mice to observe their ovaries. The weight of the uterus (with
   ovaries) was measured and recorded. The uterus index, defined as the
   ratio of uterus weight to body weight of mice, was calculated and
   compared to the control group. There was only a slight decrease but no
   significant difference in the uterus index of Ddx3x^loxP/loxP; Zp3-Cre
   female mice compared to Ddx3x^loxP/loxP female mice (Fig. [109]4B). In
   contrast, the size of the ovaries of Ddx3x^loxP/loxP; Zp3-Cre female
   mice were much smaller than that of Ddx3x^loxP/loxP female mice (Fig.
   [110]4C). The ovaries of Ddx3x^loxP/loxP and Ddx3x^loxP/loxP; Zp3-Cre
   female mice were collected, fixed in formalin, and embedded in paraffin
   for long-term storage. Follicular development was observed in
   hematoxylin and eosin (H&E)-stained ovarian sections. Notably, it is
   difficult to find secondary and antral follicles in the ovaries of
   Ddx3x^loxP/loxP; Zp3-Cre female mice (Fig. [111]4D). The ovarian tissue
   of Ddx3x^loxP/loxP; Zp3-Cre female mice showed the absence of corpus
   luteum (Fig. [112]4D). The accumulation of primary follicles in
   oocyte-specific Ddx3x cKO female mice suggested that deletion of Ddx3x
   inhibits the primary to secondary follicle transition (Fig. [113]4E).
   The results supported that Ddx3x is required for the development of
   ovaries and mature follicles in female mice. However, the molecular
   mechanisms of DDX3 in ovarian development and oogenesis remain to be
   further studied.

Extensive apoptosis occurs in the ovarian follicles of oocyte-specific Ddx3x
cKO female mice

   We also performed immunohistochemistry (IHC) staining with DDX3
   antibody to detect oocyte-specific knockout of Ddx3x in the ovaries of
   Ddx3x^loxP/loxP; Zp3-Cre female mice. However, the IHC staining of
   Ddx3x in ovarian oocytes is too light to distinguish between
   Ddx3x^loxP/loxP and Ddx3x^loxP/loxP; Zp3-Cre female mice (Fig.
   [114]5A). Among the observed ovarian sections of oocyte-specific Ddx3x
   cKO female mice, only one antral follicle was seen (Fig. [115]5A).
   Notably, oocyte collapse/degeneration was observed in the ovaries of
   oocyte-specific Ddx3x cKO female mice (Fig. [116]5A). The phenotype of
   oocyte collapse/degeneration suggested that cells may undergo
   apoptosis. To examine whether oocyte-specific deletion of Ddx3x results
   in apoptosis in the ovarian follicles, we performed IHC and TUNEL
   staining to detect apoptosis in mouse ovarian sections. Cleaved caspase
   3 is a specific marker of apoptosis in mouse oocytes [[117]48]. IHC
   staining with cleaved caspase 3 antibody and TUNEL assay were performed
   to detect apoptotic cells in mouse ovarian sections. The results of
   cleaved caspase 3 and TUNEL staining showed that extensive apoptosis
   occurs in ovarian sections of Ddx3x^loxP/loxP; Zp3-Cre female mice
   compared to Ddx3x^loxP/loxP female mice (Fig. [118]5B, C). Therefore,
   loss of Ddx3x in mouse oocytes results in a female infertility
   phenotype due to apoptosis of ovarian follicles.

Fig. 5. Impaired oogenesis and extensive apoptosis occur in the ovaries of
oocyte-specific Ddx3x cKO mice.

   [119]Fig. 5
   [120]Open in a new tab

   10-week-old Ddx3x^loxP/loxP female mice (n = 6) and Ddx3x^loxP/loxP;
   Zp3-Cre female mice (n = 6) were dissected to collect their ovaries.
   Mouse ovaries were fixed in formalin, embedded in paraffin, and then
   sectioned at a thickness of 5 μm. A Ovarian sections were processed for
   immunohistochemistry (IHC) to stain endogenous mouse Ddx3x.
   Representative images of IHC-stained ovarian sections were shown. An
   antral follicle and nearby degenerated oocytes were enlarged in
   Ddx3x^loxP/loxP; Zp3-Cre ovarian sections. B Ovarian sections were
   processed for immunohistochemistry (IHC) to stain cleaved caspase 3.
   Representative images of IHC-stained ovarian sections were shown. Brown
   signals were frequently observed in Ddx3x^loxP/loxP; Zp3-Cre ovarian
   sections. C TUNEL assay was performed to detect apoptotic DNA
   fragmentation in mouse ovarian sections. DAPI staining shows the
   presence of nuclei in mouse ovaries. Scale bars, 100 μm.

DDX3-regulated candidate genes are involved in oogenesis

   Because it is difficult to collect mouse oocytes from oocyte-specific
   Ddx3x cKO female mice, we re-analyzed a set of translational targets of
   DDX3 that were previously identified in HeLa cells [[121]6]. To gain
   insight into the biological functions of DDX3, pathway enrichment
   analysis was performed using the DAVID Bioinformatics Resources
   software ([122]https://david.ncifcrf.gov/) to map genes to biological
   pathways defined by the Kyoto Encyclopedia of Genes and Genomes (KEGG)
   database. The result showed that a set of candidate genes are involved
   in oocyte meiosis, progesterone-mediated oocyte maturation, and the
   GnRH signaling pathway (Table [123]1). These candidate DDX3 target
   genes include APC1, CCNE1, P38γ, B56β, PP2Aβ, PKACα, FBXW11, RAC1,
   P38β, and GNAS (Table [124]1). Since these candidate genes have been
   implicated in oocyte meiosis and maturation (Table [125]2), we
   speculate that DDX3 has a role in oogenesis.

Table 1.

   Pathway enrichment analysis of candidate DDX3 target genes involved in
   oogenesis.
   KEGG pathway p Value Genes
   Oocyte meiosis 0.0045 APC1, CCNE1, P38γ, B56β, PP2Aβ, PKACα, FBXW11
   Progesterone-mediated oocyte maturation 0.0357 APC1, RAC1, P38γ, PKACα,
   P38β
   GnRH signaling pathway 0.0411 P38γ, PKACα, P38β, GNAS
   [126]Open in a new tab

Table 2.

   DDX3-regulated candidate genes involved in meiosis and oogenesis.
   Genes Effects on oogenesis References