The piRNA protein Asz1 is essential for germ cell and gonad development in zebrafish and exhibits differential necessities in distinct types of germ granules

Abstract

Germ cells are essential for fertility, embryogenesis, and reproduction. Germline development requires distinct types of germ granules, which contains RNA-protein (RNP) complexes, including germ plasm in embryos, piRNA granules in gonadal germ cells, and the Balbiani body (Bb) in oocytes. However, the regulation of RNP assemblies in zebrafish germline development are still poorly understood. Asz1 is a piRNA protein in Drosophila and mice. Zebrafish Asz1 localizes to both piRNA and Bb granules, with yet unknown functions. Here, we hypothesized that Asz1 functions in germ granules and germline development in zebrafish. We generated asz1 mutant fish to determine the roles of Asz1 in germ cell development. We show that Asz1 is dispensable for somatic development, but essential for germ cell and gonad development. asz1^-/- fish developed exclusively as sterile males with severely underdeveloped testes that lacked germ cells. In asz1 mutant juvenile gonads, germ cells undergo extensive apoptosis, demonstrating that Asz1 is essential for germ cell survival. Mechanistically, we provide evidence to conclude that zygotic Asz1 is not required for primordial germ cell specification or migration to the gonad, but is essential during post-embryonic gonad development, likely by suppressing the expression of germline transposons. Increased transposon expression and mis-organized piRNA granules in asz1 mutants, argue that zebrafish Asz1 functions in the piRNA pathway. We generated asz1;tp53 fish to partially rescue ovarian development, revealing that Asz1 is also essential for oogenesis. We further showed that in contrast with piRNA granules, Asz1 is dispensable for Bb granule formation, as shown by normal Bb localization of Buc and dazl. By uncovering Asz1 as an essential regulator of germ cell survival and gonadogenesis in zebrafish, and determining its differential necessity in distinct germ granule types, our work advances our understanding of the developmental genetics of reproduction and fertility, as well as of germ granule biology.

Author summary

Germ cells undergo a highly dynamic developmental program that begins in the early embryo and continues through juvenile and adult life. Identifying functional regulators and deciphering the developmental mechanisms of germ cells are critical for advancing our understanding of fertility and reproduction, as well as their associated diseases. Here, we identified Asz1 as an essential regulator of germ cell and gonad development in zebrafish. We demonstrate that zygotic Asz1 is dispensable for the specification and migration of primordial germ cells in the embryo, but is necessary for germ cell survival in the developing gonad, likely by protecting them from transposable elements. Upon loss of asz1, expression of germline transposons was induced, and piRNA granules were mis-organized, suggesting a conserved role for zebrafish Asz1 in the piRNA pathway. We further show that Asz1 is required for proper oogenesis. However, unlike RNA-protein (RNP) complexes of germ granules of the piRNA pathway, Asz1 was not required for RNP complexes of the Balbiani body in differentiating oocytes, revealing its differential necessity in distinct types of germ granules. In mice, Asz1 was shown to be essential for spermatogenesis but not oogenesis, but its functions in human gonads are unclear. Our work reports the functional requirements of Asz1 in both sexes in zebrafish, contributes to our knowledge of developmental reproduction biology, and sheds new light on the complexity of RNP complex assemblies in germ granules.

Introduction

Germ cells are essential for fertility, embryonic development, and reproduction. In animals, germ cells undergo a highly dynamic developmental program. Primordial germ cells (PGCs) are first specified in the early embryo and then migrate to the developing gonad. In the gonads, PGCs give rise to germline stem cells that in turn produce oogonia or spermatogonia which later initiate differentiation by the induction of the meiosis program. Sex determination mechanisms in the gonad direct early germ cells to initiate either oogenesis or spermatogenesis, for the generation of functional egg or sperm cells, through intricate mechanisms of differentiation and morphogenesis. Identifying regulators and deciphering the mechanisms by which they control the dynamic multi-step differentiation of germ cells is essential for better understanding of their development, as well as fertility and reproduction.

In zebrafish, the regulation of various steps in germ cell development requires three types of germ granules, which contains RNA-Protein (RNP) complexes. First, embryonic germ granules that contain germ cell fate determinants, termed germ plasm, initially specify PGCs in the embryo, and then maintain their germ fate and are required for their proper migration to the gonad [1]. The germ plasm includes the RNP components Ddx4, Dead-end (Dnd), and Bucky ball (Buc) [2,3]. A second type of germ granules form during gametogenesis in oogonia/spermatogonia and early differentiating oocytes/spermatocytes in the ovary and testis [4]. Gamete germ granules are localized perinuclearly and contain the PIWI interacting RNAs (piRNA) pathway machinery [5], which protects germline nuclei from retrotransposons by processing and cleaving their transcripts [4–8]. In most animals [5,7], including zebrafish [9,10], gamete germ granules associate with clustered mitochondria and were collectively referred to as nuage. Zebrafish gamete germ granules contain the piRNA proteins Ziwi, Zili, as well as Ddx4, and are essential for repression of transposon expression [10].

A third type of RNP complexes form in an oocyte membraneless organelle, which is conserved from insects to humans, and called the Balbiani body (Bb) [11–16]. The Bb establishes the oocyte animal-vegetal polarity, by specifying the vegetal pole of the egg and future embryo, which is key for embryonic development [17]. Bb RNP granules include the germ plasm components as well as dorsal determinants [18]. The Bb delivers the germ plasm and dorsal determinants to the oocyte vegetal pole from which they will later act to specify the germline and to establish the embryonic dorsal-ventral axis [18]. In oogenesis, following an early symmetry-breaking event that polarizes Bb granule localization around the centrosome [13], the Bb forms adjacent to the oocyte nucleus and then translocate to the oocyte cortex [19]. At the cortex, the Bb disassembles, docking its RNP granules to the cortex, and thus specifying this region as the oocyte vegetal pole [19]. A fundamental understanding of the regulation of Bb RNP complexes is lacking, and the Bucky ball (Buc) protein is the single known essential Bb protein in any species [13,20]. In buc mutants the Bb fails to form, resulting in radially symmetrical eggs and embryonic lethality [20]. Only two other Bb proteins, Tdrd6a and Rbpms2, were shown to affect Bb morphology [21,22]. The Tdrd6a protein interacts with Buc and enhances its granule formation, and the Rbpms2 protein localizes to the Bb, but neither are essential for Bb formation [21,22].

Germ granules are thought to form dynamic molecular condensates, as known for germ plasm and piRNA granules in germ cells and embryos in invertebrates, such as P granules in C. elegans and polar granules in Drosophila [23], as well as in somatic mammalian tissues, such as stress granules and P bodies [24]. Accumulating knowledge in invertebrates begin to uncover regulators and decipher mechanisms that control the molecular condensation of granules [25–27], as well as generate the understanding of their hierarchical formation of homotypic and heterotypic RNP complexes [5]. However, our knowledge of the structure, the full repertoire of RNA/protein components, and the regulatory proteins of germ plasm-, piRNA-, and Bb granules is still lacking in vertebrates, including zebrafish.

The Four ankyrin repeats (ANK), a sterile alpha motif (SAM), and leucine zipper 1 protein, named Asz1 (previously called GASZ), is a germ cell specific protein that localizes to germ granules [28]. In mice gonads, Asz1 is localized to the cytoplasm of pachytene spermatocytes and early spermatids, and is expressed during all stages of oogenesis [28]. In both oocytes and spermatocytes, Asz1 localizes to the perinuclear germ granule RNPs, which contain the machinery of the piRNA pathway [5]. Asz1 co-localizes with the PIWI protein MILI and is required for piRNA processing [29], and it functions similarly in Drosophila [30,31]. piRNA granules associate with clustered mitochondria and are collectively referred to as nuage. Asz1 is important for nuage formation, mitochondrial clustering and transposon repression due to its ability to interact with the mitochondria outer membrane [32,33]. Loss of Asz1 in mice results in increased retrotransposon transcription and male sterility but has no apparent phenotypes in females [29].

In zebrafish, Asz1 was shown to localize in oogonia and early oocytes to germ granules that contain the piRNA proteins Ziwi and Zili, as well as Ddx4 [13], and in zebrafish and Xenopus, Asz1 localizes to the Bb as well [13,20,34]. Zebrafish Asz1 was shown to undergo concomitant localization dynamics with Bb granules during symmetry-breaking [13], and its localization to the mature Bb was dependent on Buc [20]. Based on the localization of Asz1 to the germ granules and the Bb, and considering its roles in mice [29], we hypothesized that Asz1 could regulate germ granule dynamics, as well as germ cell and gonad development in zebrafish.

Here, to test this hypothesis, we generated asz1 mutant fish and determined the roles of Asz1 in germ cell and gonad development. We provide evidence to show that loss of Asz1 exhibit no apparent somatic developmental defects but is essential for germ cell survival and gonad development. asz1^-/- fish developed exclusively as sterile males that completely lacked germ cells. Rare rescue of ovarian development in asz1;tp53 double mutant fish revealed that Asz1 is essential for oogenesis. Mechanistically, we show that loss of Asz1 results in defects in the organization of piRNA germ granules, de-repression of transposon expression, and extensive germ cell apoptosis, which likely demonstrates a conserved role for Asz1 in the piRNA pathway in zebrafish. In contrast, we provide evidence concluding that zygotic Asz1 is not essential in germ plasm granules, and is dispensable for Bb formation. Our work establishes that Asz1 is essential for germ cell and gonad development in both females and males in zebrafish, and reveals differential necessities for this protein in distinct types of germ granules.

Results

Loss-of Asz1 function induces defective germ cell and gonad development

Asz1 is a germline specific piRNA granule protein in the mouse, and a strong candidate regulator of germ granule biology, as well as germ cell and gonad development in zebrafish. However, the functions of Asz1 in zebrafish have not been addressed. We confirmed by RT-PCR analysis that zebrafish asz1 is expressed specifically in ovaries and testes, with maternal deposition of transcripts in early embryos (S1A Fig). To determine the roles of Asz1 in zebrafish, we generated fish carrying a loss-of-function allele of asz1.

Consistent with its structure in the mouse, zebrafish Asz1 is a 480 amino acid protein that contains three ankyrin repeats at the N-terminus, a SAM domains in the middle of the protein, and a trans-membrane domain at the C-terminus (S1D Fig) [34]. We used CRISPR/Cas9 to generate a loss-of-function allele, using two independent gRNAs. We targeted exon 6 at the end of the ankyrin repeats (S1D Fig) with gRNA1, and exon 2 at the middle of ankyrin repeats (S1D Fig) with gRNA2, and screened for alleles bearing an early termination codon. Despite multiple screening of many F1 families in the F2 and F3 generations, gRNA2 only produced various mismatch mutations that were not predicted to alter the Asz1 protein sequence and function. In contrast, gRNA1 resulted in a four base-pair deletion, generating a premature STOP codon at amino acid 243 (S1B and S1D Fig). PCR analysis of asz1 genomic locus at the region of the gRNA1 sequence, followed by high-resolution gel electrophoresis, confirmed the mutation (S1C Fig).

This allele, asz1^huj102 (referred to as asz1^- for simplicity; Methods), is predicted to generate either a truncated protein that lacks the SAM and TM domains and is thus non-functional, or no protein at all in case the transcript undergoes nonsense mediated decay. To distinguish between these two possibilities, we examined asz1 transcripts by RT-PCR. Below, we determine that loss of asz1 results in complete loss of germ cells, but that residual germ cells are still present at 4 wpf. Since lack of germ cells precludes their gene expression analysis, we performed this analysis on wt and mutant gonads at 4wpf, and monitored the presence of germ cell by using primers for the germ cell specific marker ddx4. While asz1 transcripts were clearly detected in wt and heterozygous gonads, they were almost completely abolished in homozygous asz1 gonads, despite substantial detection of residual vasa expression, which confirms the presence of germ cells in these samples (S1E Fig). As expected, asz1 transcripts were also abolished from adult gonads (S1E Fig). These results suggest that mutant transcripts undergo nonsense mediated decay.

We next tested for phenotypes in asz1 mutants. Sibling wild type (wt), asz1^+/-, and asz1^-/- fish were always raised under identical conditions and fish from all genotypes successfully developed to adulthood similarly. Consistently with the gonad specific expression of asz1, asz1^-/- mutant fish seemed to develop normally with no apparent somatic phenotypes (Figs 1A and S3). As a more accurate measure, the standard length (SL) in zebrafish is used as a parameter of fish growth in adulthood and during post-embryonic development, and is defined as the distance between the fish snout and the base of its tail [35]. The SL of sibling wt, and asz1^-/- fish were similar during post-embryonic development at 4 and 6 weeks post-fertilization (wpf), well as at 3 months post-fertilization (mpf), with statistically insignificant variation between sibling genotypes (Figs 1E and S3).

Fig 1. Loss of Asz1 results in germ cell loss and underdeveloped gonads and asz1^-/- fish develop as sterile males.

A. Representative adult fish at 3 months post-fertilization and their corresponding gonads of the indicated genotypes. Ruler grades are 1mm, showing similar standard lengths (SL) of fish from all genotypes (S3A Fig), but much smaller gonads of asz1^-/- compared to wt ovaries and testes. B-C. Representative confocal images of wt ovaries and testes, labeled with Ddx4 (green), mAb414 (red) and DAPI (blue) in B, and with Acetylated tubulin (green), β-Catenin (red) and DAPI (blue) in C. Insets show single and merge channels of magnifications of the yellow boxes in zoomed out images. Scale bars are 50 μm and 10 μm in zoomed out and inset magnification images, respectively. In C, yellow arrowheads indicate zygotene cilia in spermatocytes, and white arrowheads indicate flagella of mature sperm. For each B and C labeling, n = 12 ovaries and 12 testes. D. asz1^-/- gonads labeled for, top: Ddx4 (green), mAb414 (red) and DAPI (blue), bottom: Acetylated tubulin (green), β-Catenin (red) and DAPI (blue), as in B-C. n = 6 gonads. E. Dimorphic external sex criteria in wt, and their phenotypes in asz1^-/- fish. Left panels show larger abdomen in females (arrowheads). Middle panels show the anal fin (right images are zoomed-in magnifications), with more pigmented stripes (white bracket) in the male. Right panels show the larger genital pore in the female (arrowheads). asz1^-/- fish exhibit typical male anatomy of all criteria. The number of analyzed fish is indicated in the plot in F. F. A plot showing the representative sex ratios in each genotype from two independent clutches as determined by the external sex criteria in E. n = number of fish. Bars are mean ± SD. G. Progeny embryos of crosses between wt females and either wt, asz1^+/-, or asz1^-/- males as indicated, at 4 hpf. Fertilized embryos exhibit an opaque animal pole which results from cellularization during cleavage stages (white arrowheads), while unfertilized eggs exhibit a transparent acellularized animal pole (yellow arrowheads). H. Dot plots showing the fertilization rates per mating from the crosses in G. n = number of cross pairs. Bars are mean ± SD.

To test for potential germline loss-of-function phenotypes, we analyzed gonads of wt, asz1^+/-, and asz1^-/- juvenile and adult fish (Figs 1 and S2). We first analyzed the general morphology of the gonads, as well as visualized germ cell and gonad organization by DAPI labeling followed by confocal microscopy (S2 Fig). We analyzed developing gonads at 6 weeks post-fertilization (wpf) (S2 Fig), as well as adult gonads at 3 months post fertilization (mpf) (Figs 1A–1D and S2).

Wt and heterozygous fish exhibited ovaries and testes with normal morphology as expected. Wt and heterozygous ovaries were at the normal size range in both juvenile and adult stages (Figs 1A and S2A–S2I). Juvenile ovaries contained developing oocytes (S2B and S2F Fig), and adult ovaries contained normal premature opaque (yellowish) st. III oocytes (Figs 1A, 1F, S2L and S2P) that were surrounded by follicle cells, as detected by DAPI (S2L and S2P Figs). Wt and heterozygous testes exhibited normal size range at both stages (Figs 1A, S2C, S2G, S2M and S2Q), as well as developing seminiferous tubules in juveniles (S2D and S2H Fig), and fully grown tubules filled with mature sperm in adults, as detected by DAPI (S2N and S2R Fig). Adult testes also showed a characteristic milky color, indicative of the presence of sperm (Figs 1A, S2M and S2Q). In sharp contrast, homozygous asz1^-/- fish exhibited much thinner, smaller and transparent gonads that appeared like threads, with no apparent oocytes or milky color, suggesting the absence of germ cells (Figs 1A, S2I and S2S). DAPI staining confirmed that homozygous gonads did not seem to contain either oocytes or sperm, and appeared as “empty” undeveloped gonads that only contained somatic and stromal cells (S2J and S2T Fig). Based on these observations, we conclude that loss of Asz1 function causes defective germ cell and gonad development. In all analyses throughout our work, asz1 heterozygous fish were indistinguishable from wt fish (S2 Fig) and are thus referred to as wt.

To better understand the consequences of asz1 loss in gonads, we analyzed adult gonads labeled for additional cellular markers. We co-labeled gonads with the germ cell specific marker Ddx4 [36,37] and with mAb414. mAb414 detects Phenylalanine-Glycine (FG)-repeat proteins in the nuclear pore complexes on the nuclear envelope, as well as a signal that co-localizes with Asz1, Ddx4, and other piRNA machinery proteins in perinuclear piRNA granules in zebrafish gonads [13,38]. Ddx4-positive germ cells with normal perinuclear granules (either oocytes or spermatocytes) were abundant in wt gonads (insets are zoomed in magnification of yellow boxes) (Fig 1B). However, Ddx4-positive cells were completely absent from asz1^-/- gonads (Fig 1B), confirming the loss of germ cells in mutants.

We next co-labeled gonads with acetylated tubulin and β-catenin (Fig 1C). Acetylated tubulin labels the zygotene cilium in zygotene oocytes and spermatocytes [39], marking developing germ cells in meiotic prophase, as well as the flagella in mature sperm (white and yellow arrowheads in Fig 1C). β-catenin marks Adherens junctions on cell membranes and is used to visualize cell boundaries and tissue morphology. Wt testes showed normal cilia (yellow arrowheads) and flagella (white arrowheads) structures in spermatocytes and mature sperm, respectively (Fig 1C), that were organized normally in clear seminiferous tubules. Additionally, wt ovaries showed normal oocytes at various stages of development, as detected by β-catenin (insets are zoomed in magnification of yellow boxes), including ciliated zygotene oocytes as detected by acetylated tubulin (Fig 1C). In contrast, homozygous mutant gonads, completely lacked either zygotene cilia or flagella, further demonstrating the lack of differentiating oocytes and spermatocytes and the lack of mature sperm (Fig 1D), consistently with the absence of Ddx4-positive cells (Fig 1D). Moreover, β-catenin labeling revealed cellular gaps, that appeared as empty holes in the tissue, which could indicate an earlier loss of germ cells. These results demonstrate striking gonad morphological defects, as well as complete loss of germ cells in adult asz1 mutant fish, concluding that Asz1 is essential for gonad and germ cell development in zebrafish.

asz1 mutants develop as sterile males

The complete lack of germ cells in asz1 mutants made it challenging to determine the sex of asz1 mutant fish. In zebrafish mutations that lead to early germ cell loss can result in male sex bias or sex reversal to males [40–42]. Oocytes are required for maintaining the ovary fate in zebrafish by yet unknown mechanisms [43] This maintenance perishes with oocyte loss and gonads adopt a testis fate [43] The thin morphology of asz1 mutant gonads in adult and juvenile stages, resembled testes morphology (Figs 1A, S2J and S2T). Moreover, β-catenin labeled mutant gonads exhibited gaps or holes in the tissue (Fig 1D) that could represent abnormal seminiferous tubules that lacked germ cells and appeared as empty tubules, as previously shown in ddx4 mutant fish [44], and suggesting that asz1 mutant gonads develop as defective testes.

To definitively determine the sex of asz1 mutants, we used three established external sex dimorphic criteria in zebrafish: 1) a larger abdomen in the female (which contains the ovary with large premature and mature oocytes) (Fig 1E), 2) a dimorphic color of the anal fin which is composed of yellow and black strips in males, and of pale and black stripes in females, and 3) the genital pore, which is located at the ventral side of the abdomen anterior to the anal fin and posterior to the anus, and is larger and protrudes more substantially from the body in females compared to males [45]. We thus scored fish sex based on those external criteria. As shown in pooled data from two representative clutches (Fig 1F), wt (n = 32) and asz1^+/- (n = 31) fish exhibited approximately 1:1 female-male sex ratios as expected. However, 100% of asz1^-/- fish (n = 30) exhibited all three male-characteristic dimorphic criteria (Fig 1F), including smaller abdomen, more significant yellow color strips, and a male-typical genital pore (Fig 1E). Therefore, based on gonad morphology and external sex criteria, we conclude that loss of Asz1 function results in a complete male bias.

The lack of germ cells in asz1 mutants predicted that adult males are sterile. To determine fertility, we tested whether asz1^-/- males can fertilize wt eggs. We crossed wt females with either wt, asz1^+/-, or asz1^-/- males (Fig 1G–1I) and determined the fertilization rates at 4 hours post fertilization (hpf) (Fig 1G and 1H). Wt and asz1^+/- males crossed to wt females resulted in normal 75%±5% and 73%±7 fertilization rate, respectively, as evident by the more opaque color of the animal pole that results from cellularization during cleavage stages, as opposed to the transparent single animal cell in non-fertilized eggs (n = 20 pairs of wt female and wt male, n = 20 pairs of wt female and asz1^+/- male) (Fig 1G and 1H). In contrast, in crosses of wt females with asz1^-/- males (n = 20 pairs), normal eggs were laid, but none (0%) were fertilized (Fig 1G and 1H). These results concluded that asz1 mutant males are sterile as expected from their lack of sperm. Altogether our data establishes that the Asz1 protein is essential for germ cell and gonad development, and that its loss is detrimental to fertility and reproduction.

Asz1 is essential for early gonad and germ cell development in the juvenile fish

To understand the developmental defects in asz1 mutant gonads, we sought to determine the timeframe during gonad and germ cell development at which Asz1 functions are essential. We thus backtracked asz1 phenotypes to juvenile stages and analyzed developing gonads at 6wpf. Zebrafish sex determination occurs in juvenile stages, and at 6wpf, both developing ovaries and testes can be captured, as was the case in wt fish (Fig 2A). However, we found that the developmental defects in asz1 mutants were already detected at these stages, and asz1^-/- gonads appeared as extremely smaller, thinner, and underdeveloped gonads compared to wt (Fig 2A).

Fig 2. Asz1 is essential for germ cell and gonad development in juvenile post-embryonic stages.

A. Representative juvenile gonads from fish at 6 wpf of the indicated genotypes. Ruler grades are 1mm, showing much smaller, thread-like, gonads of asz1^-/- compared to wt ovaries and testes. The distribution of gonad morphology from two representative clutches is plotted in B. n = number of gonads. Bars are mean ± SD. C. Representative confocal images of ovaries and testes of the indicated genotypes as in A, labeled with Ddx4 (green), mAb414 (red) and DAPI (blue). Insets show single and merge channels of magnifications of the yellow boxes in zoomed out images. Scale bars are 50 μm and 10 μm in zoomed out and inset magnification images, respectively. Wt ovaries (n = 16) and testes (n = 14) exhibited normal developing oocytes and early spermatocytes, as well as oogonia and spermatogonia as indicated by Ddx4, with normal perinuclear piRNA granules, as indicated by mAb414 signals (white arrowheads in wt ovary). D. Representative confocal images of ovaries and testes of the indicated genotypes as in A, labeled with cCaspase3 (green), mAb414 (red) and DAPI (blue). Insets show single and merge channels of magnifications of the yellow boxes in zoomed out images. Scale bars are 50 μm and 10 μm in zoomed out and inset magnification images, respectively. White arrowheads indicate cCaspas3-positive apoptotic cells. n = 10 ovaries and 10 testes (wt). E. asz1^-/- gonads labeled as in C-D. Top: Ddx4 (green), mAb414 (red) and DAPI (blue). asz1^-/- gonads (n = 13) had only few Ddx4-positive germ cells which exhibited abnormal mAb414 signal that appeared as coalesced granules (white arrowheads in asz1^-/- gonads). Bottom: Acetylated tubulin (green), β-Catenin (red) and DAPI (blue), as in C-D. n = 6 gonads. F. The number of apoptotic cells per gonad is plotted and is non-statistically significant between genotypes. n = number gonads. Bars are mean ± SD.

To monitor germ cell development, we labeled gonads with Ddx4 and mab414. Ddx4 labeling detected numerous germ cells at various developmental stages in wt testes (n = 12 testes) (Fig 2C), as well as in wt ovaries (n = 16 ovaries) (Fig 2C). Strikingly, in asz1 mutant gonads, only few germ cells were detected (n = 13 gonads) (Fig 2E). Furthermore, in contrast with the detection of mAb414 in normally organized perinuclear granules in wt gonads (Fig 2C, white arrowheads in C) [13], in the few existing germ cells in asz1^-/- gonads, mAb414 signal was detected as an abnormal and polarized large aggregate in the cell (Fig 2E, white arrowheads in E). The mis-organized mAb414 signal could indicate a piRNA granule defect in asz1 mutants. However, Ddx4-positive granules remained normally perinuclear (Fig 2C).

The plot in Fig 2B summarize these phenotypes. Based on the Ddx4 and mAb414 analyses above, we categorized all analyzed gonads as “normal testes” or “normal ovaries”, if they showed normal number and developmental stages and organization of spermatocytes or oocytes, respectively (as shown in wt in Fig 2A and 2C). We defined thinner gonads with no or very few germ cells, as “abnormal gonads”, which considering the male bias in asz1 mutants likely represent underdeveloped testes (as shown in asz1^-/- in Fig 2E). While wt (n = 16) and asz1^+/- (n = 16) gonads developed as either normal ovaries or normal testes (Fig 2D), with no detected abnormal gonads, asz1^-/- gonads (n = 15) were all classified as "abnormal gonads" and did not include any normal ovaries or testes (Fig 2D). These results conclude that Asz1 is essential for early germ cell development already in the juvenile gonad.

The striking and early loss of germ cells in 6 wpf asz1^-/- gonads suggested that germ cells could be lost by apoptosis. To monitor apoptosis, we labeled gonads from all three genotypes with the apoptosis marker cleaved Caspase3 (cCaspas3) (Fig 2D and 2E, white arrowheads). We detected a similar number of cCaspase-positive apoptotic cells in wt (8±3 apoptotic cells per gonad, n = 5 gonads), asz1^+/- (7±4 apoptotic cells per gonad, n = 5 gonads), and asz1^-/- (8±7 apoptotic cells per gonad, n = 6 gonads) gonads, with statistically insignificant differences (Fig 2F). This suggest that germ cells are either lost by another mechanism, or that they have already been lost by apoptosis earlier in gonad development. We therefore next analyzed gonads at earlier developmental stages.

Asz1 is essential for germ cell survival in early gonad development

To further backtrack the loss of germ cells and the developmental defects in the asz1 mutants, we analyzed gonads closer to their developmental onset, at 4wpf (Fig 3). Prior to sex determination in zebrafish, gonads develop as ovaries, and upon sex determination they either continue to develop as ovaries, or convert to testis development [44]. At 4wpf, preceding sex determination, all gonads still develop as early ovaries [46]. Indeed, wt and asz1^+/- gonads all appeared as normal early developing ovaries (Fig 3A and 3B). However, asz1^-/- gonads were again much smaller and thinner compared to siblings (Fig 3A and 3B), demonstrating that Asz1 functions are required already during early gonad development.

Fig 3. Asz1 is essential for germ cell survival during early gonad development.

A. Wt and asz1^-/- gonads at 4 wpf as shown on Figs 1 and 2. At 4 hpf, prior to sex determination, all gonads are still developing as ovaries as shown in wt and asz1^+/-. asz1^-/- exhibit much smaller thread like gonads (white arrows in F) already at 4hpf. Yellow arrows in F indicate surrounding bright fat tissue. B. Distribution of gonad phenotypes as shown in A. n = number of gonads. C-F. Confocal images of wt (C-D) and asz1^-/- (E-F) gonads, labeled for mAb414 and Ddx4, as well as β-Catenin and cCaspase3, as shown as in Figs 1 and 2. Developing ovaries in wt contain oogonia and early differentiating oocytes (Ddx4, green) with normally organized perinuclear granules (mAb414, red), but much fewer germ cells in asz1^-/- that also exhibit coalesced mAb414 signals (white arrowheads in I). n = 6 gonads per genotype. cCaspase3 labeling (green) of wt, asz1^+/-, and asz1^-/- gonads. Cells with coalesced mAb414 signals in asz1^-/- gonads (F), which are likely abnormal germ cells, are positive for cCaspase3 (white arrowheads in F). n = 5 gonads per genotype. Scale bars in are 50 μm and 10 μm in zoomed out and inset magnification images, respectively. G-H. asz1^+/+;Tg(ddx4:GFP) and asz1^-/-;Tg(ddx4:GFP) gonads labeled for GFP (magenta), cCspas3 (cyan), and DAPI (blue), showing germ cell apoptosis. Scale bars are 15 μm. I. The number of Ddx4-positive germ cells per gonad is plotted for each genotype. n = number of gonads. Bars are mean ± SD. J. The number of cCaspase3-positive cells are per gonad is plotted for each genotype. n = number of gonads. Bars are mean ± SD. K. % of Ddx4+ germ cells that are cCaspas3+ is plotted. n = number of gonads. Bars are mean ± SD.

To examine the cellular phenotypes at these stages, we labeled gonads with Ddx4 and mAb414. To test for potential germ cell loss at 4 wpf, we determined the number of germ cells per gonad in each genotype, by counting Ddx4-positive cells. Wt (n = 6) and asz1^+/- (n = 6) gonads contained 920±470 and 820±220 germ cells per gonad, respectively (Fig 3I). In sharp contrast, asz1^-/- gonads (n = 6) contained only 90±90 germ cells per gonad (Fig 3I), further confirming the early loss of germ cells in this mutant. Interestingly, in existing germ cells in asz1^-/- gonads, mAb414 labeling exhibited the same abnormal polarized aggregation phenotype as in 6 wpf gonads. Such aggregates were not detected in wt and asz1^+/- gonads, which showed normal perinuclear granules (Fig 3D and 3F).

We next examined apoptosis in 4wpf gonads, as labeled by cCaspase3. We counted cCaspas3-positive cells per gonad in each genotype, and found that apoptosis levels were abnormally and statistically significantly higher in asz1^-/- gonads. asz1^-/- mutant gonads (n = 5) contained as high as 42±3 apoptotic cells per gonad, compared to only 17±11 and 25± 7 apoptotic cells per gonad in wt (n = 5), and asz1^+/- gonads (n = 5), respectively (Fig 3D, 3F and 3J). We confirmed the apoptosis of germ cells by co-labeling for cCaspas3 and Ddx4. Since both cCaspase3 and Ddx4 antibodies are rabbit in origin, we crossed asz1 heterozygous fish to Tg(ddx4:GFP) fish, in-crossed them to obtain asz1^-/-; Tg(ddx4:GFP), and co-labeled for cCaspase3 and GFP (using anti-GFP mouse antibody). In asz1^+/+;Tg(ddx4:GFP), cCaspas3 was detected in 17.4±2% of Ddx4+ cells (n = 4 gonads; Fig 3G and 3K), consistently with the low levels of apoptosis we generally detected in wt juvenile ovaries here (Fig 3J) and previously [39]. By contrast, in asz1^-/-; Tg(ddx4:GFP) gonads 42±24% of Ddx4+ germ cells were positive for cCaspas3 (n = 5 gonads; Fig 3H and 3K), confirming extensive germ cell apoptosis upon loss of asz1.

To further determine the developmental stage of germ cell loss in asz1 mutants, we analyzed 3 wpf gonads. At 3 wpf, the number of Ddx4+ cells was indistinguishable between wt (n = 6 gonads) and asz1^-/- gonads (n = 10 gonads; Fig 4A and 4B). Thus, germ cells are lost after 3 wpf, and apoptosis is likely continuous during the fourth week of development since vast reduction in their number is already evident at 4 wpf and they are completely abolished thereafter.

Fig 4. Zygotic Asz1 is not required for PGC specification and migration.

A. asz1^+/+;Tg(ddx4:GFP) and asz1^-/-;Tg(ddx4:GFP) gonads at 3 wpf, labeld for Ddx4 (magenta) and DAPI (blue), showing similar number of Ddx4+ germ cells. Scale bars are 15 μm. B. The number of Ddx4+ germ cells from A, is plotted. n = number of gonads. Bars are mean ± SD. C. Ddx4 labeling in 24 hpf wt, asz1^+/-, and asz1^-/- embryos shows normal localization of PGCs to the presumptive gonad region (arrowheads). D. The number of PGCs per embryo as shown in C is plotted for each genotype and shows non-significant variation across genotypes. n = number of embryos. Bars are mean ± SD. E. Ddx4 labeling (green) in 5 dpf wt, asz1^+/-, and asz1^-/- larvae. Zoom in images on the presumptive gonad regions are shown. F. The number of PGCs per larvea as shown in E is plotted for each genotype and shows non-significant variation across genotypes. n = number of embryos. Bars are mean ± SD. G. Ddx4 labeling in 7 dpf wt, asz1^+/-, and asz1^-/- larvae shows normal localization of PGCs to the forming gonad region. Bottom images are zoom-in magnifications of the boxes in larvae in the top panels. H. The number of PGCs per larva as shown in G is plotted for each genotype, and shows non-significant variation across genotypes. n = number of embryos. Bars are mean ± SD.

These results conclude that upon loss of Asz1 functions, germ cells are cleared by apoptosis at ~4 wpf, and that Asz1 is essential for their survival and for gonad development. While we did not rule out other possible mechanisms for the loss of germ cells, the extensive rates of apoptosis we observe at 4 wpf strongly suggest that apoptosis is a major mechanism contributing to their loss.

Zygotic Asz1 is dispensable for PGC specification and migration to the gonad

Prior to germ cell differentiation in the early developing gonad, germ cells and somatic cells of the gonad develop from different embryonic origins. Germ cell precursors, called primordial germ cells (PGCs), are specified by maternal inheritance of germ plasm during early embryonic development, and the gonad develops at later stages at the urogenital ridge [47]. PGCs migrate to the future developing gonad and populate it [47]. In zebrafish, PGCs migrate to the region of the future developing gonad by 24 hours post-fertilization and populate it by 7 days post-fertilization [1,15]. Our experiments above timed the loss of germ cells in asz1 mutants to 3-4 wpf, suggesting that PGC specification and migration, prior to gonad development, are normal.

To confirm this, we labeled PGCs using the germline specific marker Ddx4, in wt, asz1^+/-, and asz1^-/- embryos at 24 hpf (Fig 4C and 4D), as well as in larvae at 5 (Fig 4E and 4F) and 7 dpf (Fig 4G and 4H). We counted Ddx4-positive PGCs per embryo at 24 hpf (Fig 4C; lateral view is shown), detecting 23±6, 19±6, and 22±6 PGCs per embryo in wt (n = 22 embryos), asz1^+/- (n = 31 embryos), and asz1^-/- (n = 9 embryos) embryos, respectively, with no statistically significant differences between the genotypes (Fig 4D). The number of PGCs per larvae at 5 and 7dpf was consistently similar between genotypes with no statistically significant differences (Fig 4F and 4H). At 5 dpf, we detected 14±2, 13±3, and 14±3 PGCs per larvae in wt (n = 6), asz1^+/- (n = 6), and asz1^-/- (n = 6), respectively (Fig 4E and 4F). At 7 dpf, we detected 21±3, 21±4, and 21±2 PGCs per larvae in wt (n = 17), asz1^+/- (n = 17), and asz1^-/- (n = 14) larvae, respectively (Fig 4G and 4H). Thus, in zygotic asz1 mutants, PGCs seem to be specified at normal numbers [48] and to properly migrate to the gonad, demonstrating that zygotic Asz1 is not a major functional regulator of PGC specification and migration in zebrafish. These experiments confirm that in asz1 mutants, germ cells are lost directly in the gonad during post-embryonic development at ~4 wpf.

Fig 7. Schematics of the germline developmental requirement for Asz1 functions.

A. Wt juvenile gonads contain developing germ cells and give rise to ovaries and testes and fertile adult fish. Upon loss of Asz1, germ cells in juvenile gonads are lost by apoptosis, resulting in underdeveloped testes-like gonads that completely lack germ cells in the adult, and leading to development of fish as sterile males. B. Asz1 is essential for suppression of transposon expression very likely by acting in the piRNA pathway as shown in Drosophila and mice (blue and orange arrows). Loss of Asz1 also results in mis-organization of piRNA granules as detected by Ziwi and a yet unknown FG-repeat protein (maroon) of these granules. C. In contrast with piRNA granules, Asz1 (black) localizes to the Bb (light blue), but upon its loss, Bb granules are normally intact. These observations reveal differential necessities for Asz1 in distinct types of germline granules in zebrafish.

Asz1 is essential for germline transposon silencing in the gonad

We next addressed the potential defects which may lead to germ cell apoptosis upon loss of asz1. Having determined that Asz1 is essential directly in the developing gonad and considering the known roles of Asz1 in piRNA processing in Drosophila and mice [29–31], we hypothesized that Asz1 could function similarly in zebrafish.

We first examined the organization of piRNA germ granules in gonads. At 4 and 6 wpf asz1-/- gonads, the organization of piRNA granules appeared defected: mAb414 exhibited abnormal aggregation, while Ddx4 granules seemed normal (Figs 2 and 3). To more definitively monitor piRNA granule organization, we used the piRNA protein Ziwi. Since Ziwi and Ddx4 antibodies are both rabbit in origin, we co-labeled asz1^+/+;Tg(ddx4:GFP) and asz1^-/-;Tg(ddx4:GFP) for GFP and Ziwi, as above. We analyzed Ziwi organization in GFP+ germ cells. We reasoned that potential defects that may lead to apoptosis at 4 wpf, likely precede this timeframe, and therefore we analyzed gonads at 3 wpf, when germ cell number is still normal in asz1 mutants and towards the peak of their loss by apoptosis at 4 wpf.

In asz1^+/+;Tg(ddx4:GFP), Ziwi showed normal and specific localization to perinuclear granules in 97.5% of GFP+ germ cells (n = 164 germ cells in 5 gonads; Fig 5A and 5B), as expected [4]. By contrast, in asz1^-/-;Tg(ddx4:GFP) normal Ziwi localization to perinuclear granules was detected in only 26% of germ cells. Instead, Ziwi was either not detected (62%), or over-aggregated in large patches and spread over the cytoplasm in 12% of germ cells (n = 179 germ cells in 7 gonads; Fig 5A and 5B). Thus, Asz1 is required for normal piRNA granule organization, consistently with its requirements in nuage assembly in mice [32,33]. Whether Asz1 is directly or indirectly required for Ziwi localization to- and/or maintenance in granules, remains to be determined.

Fig 5. Asz1 is essential for piRNA granule organization ad repression of transposon expression.

A. 3 wpf asz1^+/+;Tg(ddx4:GFP) and asz1^-/-;Tg(ddx4:GFP) gonads are labeled for Ddx4 (green), Ziwi (red), ad DAPI (blue). Images are: top - merge images, middle - Ziwi channel, bottom - zoom-in images of white dashed boxes. Normal Ziwi localization to perinuclear granules is shown in asz1^+/+;Tg(ddx4:GFP) gonads. asz1^-/-;Tg(ddx4:GFP) gonads exhibit mostly undetected Ziwi granules (left), or over-aggregated and/or cytoplasmic Ziwi localization (right). Scale bars are 15 μm. B. Frequencies of Ziwi phenotypes from A are plotted. n = number of germ cells. Bars are mean ± SD. C. RT-qPCR analysis of the expression of germline transposons in 5 wpf gonads from each genotype (n = 50 gonads per genotype), as normalized for RNA loading by β-act as well as for germ cells by vasa expression (see methods). Bars are mean ± SD from biological duplicates, each made with technical replicates. Statistical analyses were tested by ANOVA.

To functionally test whether Asz1 is required for the piRNA pathway in zebrafish, we next inquired whether loss of asz1 results in de-repression of retrotransposon. Several germline specific transposon in zebrafish, including the long terminal repeat (LTR) transposons LTR2 and GypsyDR2, the non-LTR transposons DIRS1a, Ngaro, and L1-5, and the DNA transposons Polinton and EnSpmnN1, were shown to be upregulated in ziwi and zili mutant gonads [4,10]. We therefore monitored the expression of those germline specific transposons in wt, asz1^+/-, and asz1^-/- gonads. We extracted total RNA from wt, asz1^+/- and asz1^-/- gonads and analyzed the expression levels of the above germline specific transposons by RT-qPCR. Because of the small size of developing gonads, and in particular of the underdeveloped asz1^-/- gonads, we extracted RNA from 50 gonads per genotype to obtain sufficient RNA yield.

Normalized expression of all transposons examined and their relative expression to the expression of the germ cell specific marker ddx4 showed background levels in wt gonads, which were unaltered in asz1^+/- gonads (Fig 5C). In sharp contrast, asz1^-/- gonads exhibited higher expression levels of all transposons examined, ranging from mild upregulation in the case of L1-5, to ~3-10-fold higher expression of GypsyDR2, DIRS1a, Ngaro, Polinton, and EnSpmnN1 (Fig 5C). These results demonstrate that upon asz1 loss of function, retrotransposons are derepressed and over-expressed, concluding that Asz1 is required for germline transposon silencing, likely through the piRNA pathway, consistently with its functions in Drosophila and mice [29–31]. They further reveal dynamics whereby miss-regulation of transposon activity in asz1^-/- gonads is accompanied by extensive germ cell apoptosis, as is the case in ziwi loss of function mutant gonads in zebrafish [10]. Based on these data, we conclude that Asz1 is essential for early germ cell survival, as well as gonad development in zebrafish, and likely functions in the piRNA pathway, as in Drosophila and mice.

Partial rescue of ovarian development in asz1;tp53 double mutant fish reveals that Asz1 is essential for oogenesis

In zebrafish oocytes, Asz1 localizes to perinuclear piRNA granules together with the major piRNA enzymes Zili and Ziwi [13]. Asz1 also undergoes polarization dynamics with Bb granules during oocyte symmetry breaking [13], and later localizes to the mature Bb [13]. These observations suggest that Asz1 could function in oogenesis. However, in asz1 mutants, the severe and early loss of germ cells, together with gonad development as underdeveloped testes, precluded analyses in ovaries. To circumvent this issue, we attempted to rescue ovarian development by crossing the asz1 mutant to tp53 mutant fish. In zebrafish, the conversion of ovaries to testes during sex determination requires oocyte apoptosis [44]. In several mutants with severe oocyte defects that induce oocyte apoptosis, rescue from apoptosis on a tp53 mutant background rescued female and ovarian development [40], enabling functional studies in oogenesis. However, this is not always the case, and in some mutant lines loss of tp53 failed to rescue ovarian development [49].

We generated asz1^+/-;tp53^+/- double mutant fish and attempted to rescue ovarian development by in-crosses to generate asz1^-/-;tp53^-/- progeny for analyses at juvenile stages. We have performed 17 rounds of in-crosses between 40 individual double heterozygous fish. Progeny juveniles carried the two mutations in nine combinations of genotypes that segregated in the expected Mendelian ratios. We examined whether gonads in asz1^-/-;tp53^-/- progeny formed developing ovaries or testes, based on gonad morphology, as well as Ddx4 and mAb414 labeling (S4A Fig). Despite tremendous effort, at 5 wpf from 15 rounds of in-crosses, gonads of all asz1^-/-;tp53^-/- fish (n = 180 gonads) exhibited the asz1^-/- mutant phenotype, and appeared like underdeveloped testes (S4A and S4F Fig). However, in two in-cross rounds we could recover 6 gonads (4 and 2 per in-cross) that appeared as ovaries (S4A, S4D and S4E Fig; 2 ovaries from the cross in D, and 4 from the cross in E). All fish from all crosses developed normal and consistent SL (S4G–S4I Fig).

In the asz1;tp53 crosses that yielded juvenile progeny with rescued ovaries, all genotype combinations with wt or heterozygous asz1 exhibited normal ratios of ovaries and testes (S4A, S4D and S4E Fig). asz1^-/-, and asz1^-/-;tp53^+/- exhibited defective testes as shown thus far for asz1 single mutants (S4A, S4D and S4E Fig). However, in asz1^-/-;tp53^-/-, we detected three different gonadal morphologies of normal and defective juvenile ovaries and testes (S4A, S4D and S4E Fig), as we previously described [39]. 60% of gonads (n = 10) exhibited the normal asz1 defective testes morphology, where testes are thin and contain only a few germ cells, and we termed this category “defective testes” (S3G and S3H Fig). 20% of gonads appeared similar to the defective testes, but contained oogonia and no further progressing stages of oogenesis, which we termed “underdeveloped ovaries” (S3E and S3F Fig). This category could represent normal gonads which developed slower and are lagging behind. However, this would predict detection of this category in the normal wt distribution, which we never detected, supporting the interpretation that those are abnormal “underdeveloped ovaries”. The remaining 20% of gonads contained progressing oocytes and were termed “developing ovaries” (S3A–S3D Fig). The total of 6 rescued ovaries from both crosses described above, belong to this category. The SL of these fish at 5wpf was consistent between all nine genotypes (S4 Fig). Thus, loss of tp53 only rarely and partially rescued female development in asz1 mutant fish.

To substantiate our categorization of gonads as developing ovaries or testes, we performed RT-qPCR for sex-specific markers. We first confirmed the specific expression of female-specific genes (cdc20, sox11b, ypel) and male-specific genes (ccrng2, star) [48, 50–54] in adult females and males, respectively, at 3 mpf (S4B Fig). We next compared the expression of those genes in wt versus asz1^-/- juvenile gonads at 4 wpf (S4C Fig). Wt juvenile gonads, which appear as developing ovaries based on our morphological criteria, expressed female-specific genes as expected from developing ovaries (n = 50 gonads; S4C Fig). In contrast, asz1^-/- gonads, which appear as developing/under-developing testes based on our morphological criteria, expressed lower levels of female-specific genes and elevated levels of male-specific genes (n = 50 gonads; S4C Fig), molecularly confirming their conversion to a testis fate. Since RT-qPCR analyses required RNA yield from 50 gonads, we could not directly analyze gonads in our rescue experiments. However, this analysis validates our morphological criteria and demonstrates that our categorization above is reliable.

Despite the rare rescue, we analyzed the partially rescued asz1^-/-;tp53^-/- gonads to obtain insight on potential Asz1 functions in oogenesis. Interestingly, asz1^-/-;tp53^-/- developing ovaries contained oocytes with abnormal morphology, as detected by Ddx4 and mAb414 labeling (Figs 6 and S5). Wt, tp53^+/-, tp53^-/-, asz1^+/-, asz1^+/-;tp53^+/-, and asz1^+/-;tp53^-/- ovaries were normal and contained Ddx4-positive oogonia and progressing oocytes with appropriate morphology (Figs 6A–6I and S5). Oogonia and progressing oocytes exhibited mAb414-positive piRNA granules that were normally organized perinuclearly (Figs 6A–6I and S5). asz1^-/- and asz1^-/-;tp53^+/- gonads were all defective testes, with few germ cells and abnormal mAb414-positive piRNA granule distribution (Figs 6A–6I and S5), as shown for asz1^-/- thus far. However, asz1^-/-;tp53^-/- ovaries contained progressing oocytes of two defective types (n = 2 ovaries). Type I defective oocytes exhibited extremely abnormal morphology, where the cytoplasm and nucleus appeared collapsed (Figs 6E, 6F, 6I and S5). Type II defective oocytes in double mutant ovaries exhibited mAb414-positive granules that were either ectopically distributed in the cytoplasm or coalesced into a large aggregate (Figs 6G, 6H, 6I and S5, white arrowheads). We never detected neither phenotype in normal ovaries in the wt, or all other genotypes (Figs 6D and S5), demonstrating clear defects specifically in asz1^-/-;tp53^-/- ovaries.

Fig 6. Partially rescued asz1 ovaries reveal defective oogenesis, but normal Bb.

A-C. Ovaries of the indicated genotypes were labeled with Ddx4 (green), mAb414 (red), and DAPI (blue). Images show representative general morphology of gonads. Right panels are single and merged channels zoomed-in images of the yellow boxed regions in the left panels. Scale bars are 50 μm and 10 μm in zoomed out and inset magnification images, respectively. Ovaries of all genotypes exhibit normally developing oocytes and ovarian morphology, except the following. asz1^-/- and asz1^-/-; tp53^+/- gonads exhibited the defective testes morphology as shown in previous figures. asz1^-/-;tp53^-/- ovaries showed defective oocytes, with abnormal nuclear morphology mAb414 granule signals, as detailed in D-H below. n = 7-11 gonads per genotype, and 2 ovaries in asz1^-/-;tp53^-/-, see I below. Full panel of all genotypes is shown in S5 Fig. D-H. Images of normal oocytes in the wt (D), and representative defective oocytes in asz1^-/-;tp53^-/- ovaries (E-H) from A, showing mAb414 and DAPI in the top panels and Ddx4 in the bottom panels per oocyte. Two representative images of defective type I oocytes that exhibit abnormal, seemingly collapsed, morphology are shown in E-F. Two representative images of defective type II oocytes with ectopic mAb414 signals that appear as coalesced mis-organized cytoplasmic aggregates are shown in G-H (arrowheads), as opposed to perinuclear granules in wt (arrowheads in D). The distribution of these phenotypes in gonads is plotted in I. Scale bars are 10 μm. I. The number of defective type I (red) and type II (purple) per gonad is plotted for each genotype. Both phenotypes were only detected in asz1^-/-;tp53^-/- ovaries. n = number of gonads. J-K. The Buc protein (red) shows normal localization in the forming Bb in the nuclear cleft (mAb14, green; DAPI, blue). n = 5 wt ovaries and 2 asz1^-/-;tp53^-/- ovaries. Representative images of all gonads are shown in S6 Fig. L-M. The dazl mRNA (HCR-FISH, red) shows normal localization in the forming Bb in the nuclear cleft (cytoplasm labeled with DiOC6, green; DAPI, blue). n = 3 wt ovaries and 2 asz1^-/-;tp53^-/- ovaries. Representative images of all gonads are shown in S7 Fig. Scale bars in A-D are 10 μm.

Abnormal oocytes in asz1^-/-;tp53^-/- developing ovaries likely represent defective oocytes that were rescued from apoptosis, but fail to normally progress through oogenesis due to the loss of Asz1 functions. It is very likely that Asz1 functions in the piRNA pathway in oocytes as well (see discussion). Despite many efforts, we could only recover partial rescue with low number of asz1^-/-;tp53^-/- ovaries, which preclude further analyses. However, the above phenotypes conclude that Asz1 is essential for oogenesis.

Asz1 is not essential for Balbiani body formation

Since Asz1 localizes to the Bb in both fish and frogs [13,28], and exhibits co-polarization with Bb granules in zebrafish [13], we wanted to examine whether Asz1 is required for Bb formation. To address this in asz1 loss of function ovaries, we analyzed Bb formation in all wt, asz1 and asz1;tp53 genotype combinations, including four of the asz1^-/-;tp53^-/- ovaries from the above rescue attempts. To obtain a representative view of Bb RNP granules, we examined the Buc protein, which localizes to the Bb and is essential for its formation [20], as well as the dazl Bb mRNA [13]. Buc and dazl localize to the Bb during its early formation in the nuclear cleft [20] and through the mature Bb [13,20].

We found that both Buc (n = 2 ovaries) and dazl (n = 2 ovaries) localized normally to the forming Bb in asz1^-/-;tp53^-/- ovaries, similar to wt (Figs 6J, 6K, 6L, 6M, S6 and S7), as well as all other genotypes (S6 and S7 Figs; n = 3-6 gonad per genotype for Buc and 3-5 gonads per genotype for dazl), except for asz1^-/-, and asz1^-/-;tp53^+/-. These latter genotypes developed as defective testes as described thus far and do not form the Bb, which is a female, oocyte-specific, organelle. These experiments demonstrate that Asz1 is not essential for the formation of the Bb. Since Asz1 localizes to the Bb, it is still possible that it is required for Bb RNP structure or functions that do not affect Buc and/or dazl localization to the Bb, or that it serves redundant functions. Interestingly, together with the likely roles of Asz1 in piRNA granules that we showed above (Figs 2–4), these experiments suggest differential necessities for Asz1 in distinct granules in the germline.

Discussion

In this work, we uncover the Asz1 protein as an essential regulator of germ cell and gonad development in zebrafish (Fig 7). We show that zygotic loss of asz1 results in severe loss of germ cells by apoptosis in developing gonads and establish that Asz1 is essential for piRNA pathway activity in zebrafish, similar to its roles in flies and mice.

Our work shows that in the absence of Asz1 functions, PGCs and germ cells develop normally during embryonic, larval, and post-embryonic development until 4 wpf. At 4 wpf we detected a peak of extensive germ cell apoptosis, as well as a sharp decrease in their number, and a complete loss of germ cells thereafter. These dynamics and the high rates of germ cell apoptosis at 4 wpf, suggest that apoptosis is the major cause for their loss. Thus, Asz1 is essential for germ cell survival and gonad development (Fig 7A). asz1^-/- fish develop exclusively as sterile males with underdeveloped testes that contain no germ cells (Fig 7A), since in zebrafish oocytes are required for maintaining the female fate, and upon their loss fish develop as males [43]. However, the development of apparent empty seminiferous tubules and lack of germ cells in males, precludes investigation of Asz1 functions directly in spermatogenesis.

Partial rescue of ovary development revealed that Asz1 is essential for oogenesis. We generated asz1^+/-;tp53^+/- fish and performed demanding attempts to rescue ovarian development and obtain asz1 loss-of-function ovaries. We performed 17 rounds of in-crosses between 40 individual double heterozygous fish and out of a total of 200 double homozygous asz1^-/-;tp53^-/- gonads of progeny at 5wpf, we could only recover 6 developing ovaries. Those 6 ovaries exhibited abnormal oocytes, demonstrating the need for Asz1 for proper oogenesis. Abnormal oocytes in asz1^-/-;tp53^-/- developing ovaries likely represent defective oocytes that were rescued from apoptosis, but fail to normally progress through oogenesis due to the loss of Asz1 functions.

We provide evidence to support that Asz1 is essential for piRNA pathway activity in zebrafish. Ziwi is an essential piRNA protein, and in ziwi mutant fish, germ cells are lost by apoptosis similarly to asz1, and transposon expression is de-repressed [4]. We found that Ziwi is mis-localized in asz1 mutants, and is either undetected or found as large aggregates and abnormally spread in the cytoplasm, demonstrating that Asz1 is required for piRNA granule organization. Consistently with Ziwi mis-localization, we detected a 2- to 10-fold increase in germ cell expression of LTR-, non-LTR- and DNA transposons upon loss of asz1 (Fig 5), providing functional proof for piRNA loss-of-function in the mutant. Asz1 function in the piRNA pathway in zebrafish is consistent with its known roles in piRNA processing in Drosophila and mice [29–31].

The extensive germ cell apoptosis and the elevated expression of transposons are concomitantly detected at 4 wpf, and are preceded by Ziwi mis-localization at 3 wpf. This may suggest that Asz1 is essential for germ cell survival, at least partly by protecting germ cells form miss-regulated transposon activity and apoptosis, through the piRNA pathway (Fig 7B). De-repression of transposon expression is thought to be correlated with apoptosis likely by induction of DNA damage. Loss of germ cells in the asz1 mutant induces sex conversion and fish develop as sterile males. Mutations in other piRNA components in zebrafish similarly leads to sterile male development, where germ cells are lost by apoptosis at pre-meiotic stages [10], or at early [4], or later [6] meiotic differentiation likely due to miss-regulation of transposon expression. However, a causative connection between de-repression of transposons, DNA damage, and apoptosis has not been demonstrated. We report that germ cells in asz1 mutants can be rescued and survive by loss of tp53. While this a partial and very rare rescue, this experiment may provide evidence to support the hypothesis that abnormally elevated transposon expression, DNA damage, and apoptosis are causatively linked.

While the lack of a sufficient number of rescued ovaries prevented direct testing of transposon expression by RT-qPCR, it is very likely that Asz1 functions in the piRNA pathway in oocytes as well. The germline specific transposons above were shown to be expressed similarly in both testes and ovaries in zebrafish [4]. Furthermore, silencing of germline transposons by the piRNA pathway is essential in zebrafish ovaries, as demonstrated by the loss of the piRNA enzymes zili and hen1, which did not severely abolish early germ cells in gonads, but resulted in later loss of abnormally differentiating oocytes [4,6]. These dynamics of piRNA loss-of-function are consistent with asz1 phenotypes: 1) loss of asz1 leads to severe loss of germ cell by apoptosis, likely by the up-regulation of transposon expression, similar to the ziwi mutant phenotype [4], 2) the rare cases of rescue of germ cells from apoptosis in asz1;tp53 double mutants allows oocyte survival, but reveals defects in oogenesis similar to the milder zili and hen1 phenotypes [4,6]. Despite many efforts, we could only recover partial rescue with low number of asz1^-/-;tp53^-/- ovaries, which preclude further analyses. Nevertheless, the above phenotypes conclude that Asz1 is essential for oogenesis.

Interestingly, asz1 loss-of-function affected distinct piRNA granule components differently (Fig 7B). In zebrafish, the mAb414 antibody marks Ddx4+ piRNA germ granules in ovaries and germplasm granules in PGCs (by detecting an unknown FG-repeat protein) [13,38]. In asz1^-/- developing gonads, mAb414 signal exhibited large coalesced aggregates in the cytoplasm of early pre-meiotic germ cells, instead of its normal perinuclear granule organization in the wt. A similar coalescence of mAb414-positive granules into large aggregates, and/or their ectopic distribution in the cytoplasm was detected in asz1^-/-;tp53^-/- oocytes. The piRNA processing protein Ziwi was either undetected, or spread in the cytoplasm, and/or formed large aggregates in asz1 mutants. Thus, Asz1 is required, directly or indirectly, to regulate Ziwi localization to- and/or maintenance in granules. Interestingly, Ddx4 still localized normally to perinuclear germ granules, in the same cells where mAb4141 was mis-localized.

These observations suggest specific structural defects of piRNA granules in the absence of Asz1 (Fig 7B). The nature of RNP complex composition and hierarchy in zebrafish germ granules is still unclear, and it is possible that Asz1 is required for the integrity and/or regulation of some complexes in piRNA granules, but not for others. Deciphering potential Asz1-Ziwi interactions in zebrafish, and identifying the yet unknown FG-repeat protein that is detected by mAb414 in zebrafish germ granules will help resolving this in the future.

Considering the roles of Asz1 in piRNA granules, as well as the importance of different types of germ granules for germline development, we tested the potential role of Asz1 in Bb granules. Asz1 localizes to the mature Bb in zebrafish and Xenopus [13,20,34], and undergoes concomitant polarization dynamics with Bb formation in zebrafish [13]. However, in asz1^-/-;tp53^-/- oocytes, the Buc protein and the dazl mRNA showed normal localization to the forming Bb, as well as to the mature Bb (Fig 6). These results demonstrated that the Asz1 protein is not essential for the localization of major Bb granule components to the Bb, or for major steps in Bb formation (Fig 7C). Since only a few Bb proteins have been identified so far, including Buc [20], Macf1 [55], Tdrd6a [56], and Rbpms [22], it is possible that 1) Asz1 is required for Bb structure/function downstream of Buc and dazl, and/or 2) Asz1 affects other components in ways that are not detected by Buc and dazl labeling, and/or 3) Asz1 functions redundantly with other factors in the Bb. It is also possible that the Asz1 protein does not function in the Bb, but uses the Bb as means for localization in the oocyte [19], towards potential maternal functions in the future embryo (see below). The exclusive development of asz1 mutants as sterile males, precludes further analyses of potential functions of the Bb and/or its components later in oogenesis or embryogenesis.

The differential necessity for Asz1 in piRNA granules versus Bb granules sheds interesting light on the regulation of distinct types of germ granules. The full repertoire of proteins and transcripts in both piRNA and Bb granules is far from being identified. Moreover, both homotypic and heterotypic complexes have been demonstrated to form in hierarchy in polar granules in the Drosophila germline [57–61]. However, whether RNP interactions form heterotypic or homotypic complexes or both, and their hierarchy of formation is not understood in zebrafish germline granules. A few proteins show overlapping and distinct localization and function between granule types. For example, Buc localizes to both germ granules in PGCs [3,62,63] and Bb granules [13,62] and is essential for both [20,62,64]. Tdrd6a localizes to all three granule types [56], is essential for PGCs, dispensable in the piRNA pathway, and modifies Buc aggregation and Bb morphology [56]. Here, we show that Asz1is required for the granule localization of some piRNA components (Ziwi, mAb4141-detected protein), but not others (Ddx4). Moreover, Asz1 localizes to both piRNA and Bb granules, but is only essential for piRNA function, adding an important piece to the puzzle of zebrafish germline granule organization.

Finally, we show that zygotic Asz1 is dispensable for embryonic PGC specification and migration to the gonad, by demonstrating that the number of embryonic PGCs in asz1 mutants is consistent with the literature reports for the wt [48], and that by 7 dpf migration to the gonad was normally complete. It is thus likely that zygotic Asz1 is not a major functional regulator of PGC specification and migration in zebrafish. Nevertheless, in addition to the zygotic expression of asz1 as detected in embryos at 8 and 24 hpf, our expression studies showed that asz1 transcripts are detected in embryos at 1.5 and 5hpf (S1A Fig), which is likely due to maternal deposition. It is possible that maternal Asz1 is required for germ plasm regulation and early PGC specification and/or migration. However, the lack of sexually mature asz1^-/- females precludes further analyses of potential maternal Asz1 functions in PGCs.

Asz1 functions are conserved in mammals. Asz1 gain-of-function was shown to promote PGC differentiation form mouse and human embryonic stem cells, and to enhance the expression of PGC markers [65], while its loss-of-function in mice embryos down-regulated the expression of those markers [65]. In mice, Asz1 is essential in piRNA regulation in the male [29–32], but despite expression in oocytes [28], is dispensable for female fertility [29]. In humans, asz1 is specifically expressed in adult testes according to the human genome atlas, but its functions, in particular potential functions in developing ovaries, have not been determined. Our studies in zebrafish provide new insight into Asz1 functions that can be directly relevant for human reproduction.

In summary, our work demonstrates the function of the Asz1 protein in the zebrafish germline, and sheds new light on germ granule biology, as well as on the piRNA pathway and the Bb. By identifying a new regulator of germ cell and gonad development in zebrafish and deciphering its mechanistic requirements, our study contributes to advancing our understanding of fertility and reproduction.

Material and methods

Ethics statement

All animal experiments were supervised by the Hebrew University Authority for Biological Models, according to the Hebrew University School of Medicine Institutional Animal Care and Use Committee under ethics requests MD-18-15600-2,and accredited by the Association for Assessment and Accreditation of Laboratory Animal Care International.

Fish lines and gonad collections

Gonads were collected from juvenile and adult fish at indicated developmental stages. Gonad collection was done as in [13,66]. Briefly, fish were cut along the ventral midline and the lateral body wall was removed. For juvenile fish, the head and tail were removed and, the trunk pieces, with the exposed abdomen containing the ovaries were fixed in 4% PFA at 4°C overnight with nutation. Trunks were then washed in PBS and ovaries were finely dissected in cold PBS. For fish at ≤3 wpf, ovaries are very small and were therefore not dissected from the trunks, but were directly stained for and imaged in the trunks. For adult fish, gonads were directly removed and fixed. Gonads were washed in PBS and then either stored in PBS at 4°C in the dark or dehydrated and stored in 100% MeOH at -20°C in the dark. Gonads were imaged using a Leica S9i stereomicroscope and camera.

Fish lines used in this research are: TU wild type, asz1^huj102 (this work, see below), tp53^M214K [65], and Tg (vasa:GFP) [67].

Generation of asz1 mutant fish by Crispr/Cas9

The asz1^huj102 allele (refered to as asz1^- for simplicity) was generated by CRISPR/Cas9 as follows. 1-cell stage embryos were injected with 1nl of gRNA duplex (crRNA:tracrRNA, 250 ng/μL) and Cas9 protein (PNA-Bio CP01-50; 500ng/μL), as described in [68]. The crRNA and tracrRNA were manufactured by IDT. The asz1 locus was targeted by two gRNA sequences, designed by the CRISPRscan software [69]. gRNA1 was GTGCGGGAGTTGTTGGATGGTGG (including the PAM sequence) and targeted the middle of Ankyrin repeats domain in exon 2. gRNA2 was AGGGAAGATGGCAGCCGACATGG (including the PAM sequence) and targeted the end of the Ankyrin repeats domain in exon 6.

F0 embryos were raised to adulthood and crossed with wt fish to screen for germline transmission. To identify potential loss-of-function mutations DNA was extracted from F1 fish using the RAPD method (see below), followed by genomic PCR amplification using Phusion DNA polymerase, targeting the asz1 target sequence with primers - Forward: AATCGCAGTGCAGTATGGACA, Reverse: TCCTACTGTGATCTCTTACCGTCA, yielding a 170bp amplicon in the wt. PCR products were analyzed by both Metaphore gel electrophoresis (2% resolution), and by Sanger sequencing (S1 Fig). Positive F1 fish were outcrossed to raise individual heterozygous F2 families, which were then in-crossed to generate F3 fish with Mendelian segregation of asz1 alleles. From gRNA1, we identified a family with a mutation including a four base pair deletion, generating a premature STOP codon at amino acid 243 (S1D Fig). Despite numerous screening efforts in F0 through F3 fish, gRNA2 only produced various mismatch mutations that were not predicted to alter the Asz1 protein sequence and function.

Genotyping

Fish were anaesthetized in 0.02% of Tricaine (Sigma Aldrich, #A5040) in system water, and fin-clipped, followed by RAPD DNA extraction protocol [70]. Briefly, tails are dehydrated in MeOH and lysed and DNA is extracted in in RAPD buffer [70]. For asz1^huj102, genotyping was performed by PCR amplification followed by Metapore gel electrophoresis using the primers - Forward: AATCGCAGTGCAGTATGGACA, Reverse: TCCTACTGTGATCTCTTACCGTCA, yielding a 170bp amplicon in the wt and 166bp in the mutant (S1 Fig). For tp53, we performed KASP genotyping (LGC, Teddington, UK), using the following SNP sequence: GATTTACAAACTCTTTTTTTATTTTTATTTTATTATTATTATTATTATTATTATTTACATGAAATTGCCAGAGTATGTGTCTGTCCATCTGTTTAACAGTCACATTTTCCTGTTTTTGCAGCTTGGTGCTGAATGGACAACTGTGCTACTAAACTACATGTGCAATAGCAGCTGCATGGGGGGGA[T/A]GAACCGCAGGCCCATCCTCACAATCATCACTCTGGAGACTCAGGAGTAAGTACTGCATATTTGATTCCTCCTCTTGTGAACTGCTTTTTTAAATTTATTTTTTATTTTTTTGAT.

RNA extraction and RT-PCT

RNA was extracted from embryos, larvae, juvenile and adult gonads, and adult organs, using TRI-reagent, followed by phenol;chloroform extraction and precipitation, DNAaseI reaction and additional phenol;chloroform extraction and precipitation. cDNA was generated using random hexamer primers and the Superscript IV Reverse Transcriptase kit as per the manufacturer instructions. cDNA was then used in PCR amplification followed by gel electrophoresis, or by qPCR, using a StepOnePlus Real-Time PCR machine (Applied Biosystems) and SYBRgreen. Primers were:

Ddx4 For 5′-GGTCGTGGAAAGATTGGCCTG-3′,

Ddx4 Rev 5′-CAGCAGCCATTCTTTGAATATCTTC-3′,

GypsyDR2 For 5′-GAAATCACCTGTGCATTTAC-3′,

GypsyDR2 Rev 5′-ATGCAGACATTGGGTAAAGC-3′,

EnSpmN1 For 5′-GATTGGCCATTGTGTTCACATGC,

EnSpmN1 Rev 5′-GCTGTGACTGTCATAGGTTTACC-3′,

Ngaro For 5′-GGGAGCGATCGAGACCTACC,

Ngaro Rev 5′-CAATCATATCACGTGCTCCTCTCG-3′,

Polinton For 5′-CCTGACAATGTTGTCAGCCTG-3′,

Polinton Rev 5′-CATGAAAGCTAAGGGTATAACTCTG-3′,

DIRS1a For 5′-GGGTGCGTCACGCTTGC-3′,

DIRS1a Rev 5′-GTAACCTCGAACGTTCCCC-3′,

L1-5 For 5′-GCACAAAGGACAAATTCACTGGAC-3′,

L1-5 Rev 5′-GTCCACGTTTAGTATTACAGTTGC-3′,

LTR2 For 5′-GGTGTCGTTAGAATGCCCTTGAC-3′,

LTR2 Rev 5′-GGTTATACCTGTGGGTCACGTG-3′.

cx44 For 5′- CTGGAGACAGGAGCTACTAATCA-3’,

cx44 Rev 5′- ATTTTGCGTTTGCGTGGTCG-3’,

figá For 5’- AAAGCTGTGAAGAGGCGACA-3’,

figá Rev 5’- GCATCACTGGCACAATACGC-3’,

ypel3 For 5’- ACGAGCCTACCTGTTCAA-3’,

ypel3 Rev 5’- GTCCCAGCCATTATCCTT-3’,

cdc20 For 5’- TCGTGGAGCAAAGATGGCAA-3’,

cdc20 Rev 5’- GATGTGGTGGTCTGCTACCC -3’,

sox11b For 5’- CTCACTTCGAGTTCCCGGAC -3’,

sox11b Rev 5’- CCACCGAGGCAGAGTCAAAA -3’,

ccng2 For 5’- CCATATGATCTGACGGGGGC -3’,

ccng2 Rev 5’- AAACATCTGGTGCTGGTCGG -3’,

star For 5’- ATAAACCACATCCGAAGA 3’,

star Rev 5’- CCAGGCAGGACTTTACTC 3’,

Sex specific marker gene primer sequences are from [71].

Fluorescence immunohistochemistry (IHC) and RNA-FISH by HCR

For gonads, IHC was performed as in [13,66]. Briefly, gonads were washed 2 times for 5 minutes (2x5min) in PBT (0.3% Triton X-100 in 1xPBS; if stored in MeOH, gonads were gradually rehydrated first), then washed 4x20min in PBT. Gonads were blocked for 1.5-2 hours (hr) in blocking solution (10% FBS in PBT) at room temperature, and then incubated with primary antibodies in blocking solution at 4°C overnight. Gonads were washed 4x20min in PBT and incubated with secondary antibodies in fresh blocking solution for 2 hr, and were light protected from this step onward. Gonads were washed 4x20min in PBT and then incubated in PBT containing DAPI (1:1000, Molecular Probes), with or without DiOC6 (1:5000, Molecular Probes) for 50 min and washed 2x5min in PBT and 2x5min in PBS. All steps were carried out with nutation. Gonads were transferred into Vectashield (with DAPI, Vector labs). Gonads were finally mounted between two #1.5 coverslips using a 120 μm spacer (Molecular Probes).

For embryos, IHC was performed similarly, except that embryos were always stored in MeOH at -20C, and rehydrated before labeling, followed by five 5 minutes washes in PBT. Embryos were then treated with 5μg Proteinase K for 1 min, re-fixed in 4% PFA for 20min, and washed three times in PBT for 5 minutes, before blocking.

Primary antibodies used were Ddx4 (1:5000, [39]), mAb414 (1;1000, Abcam), Acetylated tubulin (1:200; Sigma-Aldrich), β-Catenin (1:1000; Sigma-Aldrich), cCaspase3 (1:300, Abcam), Buc (1:400, [19]), Ziwi (1:100, [10]), mouse anti-GFP (1:400) Thermo scientific). Secondary antibodies were used at 1:500 (Alexa-flour, Molecular Probes).

RNA-FISH was performed using the third generation DNA-HCR-FISH technique (Molecular Instruments) [72], as in [13,66] and following the company protoco)l, except for the hybridization temperature that was optimized for 33°C.

Confocal microscopy, image acquisition and processing

Images were acquired on a Zeiss LSM880 confocal microscope using a 40X lens. The acquisition setting was set between samples and experiments to: XY resolution = 1104x1104 pixels, 12-bit, 2x sampling averaging, pixel dwell time = 0.59sec, zoom = 0.8X, pinhole adjusted to 1.1μm of Z thickness, increments between images in stacks were 0.53μm, laser power and gain were set in an antibody-dependent manner to 7-11% and 400-650, respectively, and below saturation condition. Unless otherwise noted, shown images are partial Sum Z-projection. Acquired images were not manipulated in a non-linear manner, and only contrast/brightness were adjusted. All figures were made using Adobe Photoshop CC 2014.

In vitro fertilization (IVF)

IVF was performed as in [39]. Briefly, sperm were collected from anesthetized males into Hank’s solution and stored on ice until eggs were collected. Sperm from individual males of either wt or asz1^-/- strains, was used to fertilize eggs collected from control wt females. Anesthetized wt females were placed in a dish and squeezed for egg collection. 100-150ul sperm solution was added to the collected eggs and incubated for 20 seconds. 1ml of E3 containing 0.5% fructose was added to activate sperm, gently mixed and incubated for 2 minutes. 2 ml of E3 was added followed by 5 minutes incubation. The dish was then flooded with E3 and placed in a 28°C incubator until examination of fertilization rates.

Statistical analysis

All statistical analysis and data plotting was performed using the GraphPad Prism 7 software. Data sets were tested with two-tailed unpaired t-test, unless otherwise indicated. p-values were: *<0.05, **<0.01, ***<0.001, ****<0.0001, ns = not significant (>0.05).

Supporting information

S1 File. The numerical values of the data presented in graphs all figures are provided.

Data for each figure is found as an individual spreadsheet in this Excel file.

(XLSX)

S1 Fig. Asz1 expression and the asz1^huj102 mutant.

A. RT-PCR analysis showing specific expression of asz1 in ovaries and testes, with maternal deposition in early embryos, and detection of likely zygotic transcripts in 8 and 24 hpf embryos. -RT – no reverse transcriptase control; NTC – no DNA template control. B. wt and asz1^huj102 genomic sequence (from Sanger sequencing) showing a 4 bp deletion in asz1^huj102, leading to premature Stop codon at amino acid position 243 (D). C. PCR analysis of genomic DNA from wt, asz1^+/huj102 and asz1^{huj102/huji102} fish, followed by Metapore gel electrophoresis (with 2% electrophoresis resolution), showing distinct bands of the asz1 locus in the three genotypes (individual fish per genotype; two homozygous fish are shown). Amplicons are 170 and 166 bp in the wt and asz1^huj102 alleles, respectively. D. A scheme of the zebrafish Asz1 protein, showing the Ankyrin repeats, Sterile alpha motif, and transmembrane domain. The top arrow indicates the position of the premature STOP codon at amino acid position 243 in the asz1^huj102 allele. E. RT-PCR analysis of asz1 expression in wt, asz1^+/huj102 and asz1^{huj102/huji102} gonads at 5 wpf (n = 50 gonads per genotype) and 3 mpf (n = 6 gonads per genotype). β-act expression serves as loading control. Ddx4 expression marks the presence of germ cells at 5 wpf. asz1 transcripts are not detected in 3 mpf gonads. At 5 wpf, asz1 transcripts are only very weakly detected, despite vasa expression, demonstrating the likely decay of the asz1 mRNA by nonsense mediated decay. -RT – no reverse transcriptase control; NTC – no DNA template control.

(TIFF)

S2 Fig. Loss of Asz1 leads to severe gonad developmental defect and germ cell loss.

A, C, E, G, I. Representative brightfield images of wt, asz1^+/-, or asz1^-/- gonads at 6 wpf. Ruler grades are 1 mm. B, D, F, H, J. Confocal images of the same gonads from the brightfield images above, labeled for DAPI (greyscale). Wt and asz1^+/- gonads exhibited normal developing oocytes and early spermatocytes in ovaries and testes as indicated, as well as oogonia and spermatogonia as generally detected by germ and somatic cell nuclear morphology (DAPI). asz1^-/- gonads were much thinner with no clear detection of presumptive germ cells. Scale bars are 50 μm. K, M, O, Q, S. Representative brightfield images of adult wt, asz1^+/-, or asz1^-/- gonads at 3 mpf. Ruler grades are 1 mm. L, N, P, R, T. Confocal images of the same adult gonads from the brightfield images above, labeled for DAPI (greyscale). Wt and asz1^+/- gonads exhibited normal ovarian and testes morphology as generally detected by DAPI (oocyte show weaker DAPI signal than their surrounding follicle cells), while asz1^-/- gonads were much thinner with no clear detection of presumptive germ cells, and exhibited gaps in the tissue. Scale bars are 50 μm.

(TIFF)

S3 Fig. Normal somatic development in asz1 fish.

A. A plot of the SL of fish from all genotypes from Fig 1. n = number of fish. SL was not significantly different between genotypes. Bars are mean ± standard deviation (SD). B-C. Images of juvenile fish at 6 wpf from Fig 2 (B) and their SL (C). D-E. Images of juvenile fish at 4 wpf from Fig 3 (D) and their SL (E). In all panels, n = number of fish. SL was not significantly different between genotypes. Bars are mean ± standard deviation (SD).

(TIFF)

S4 Fig. Partial rescue of ovary fate in asz1;tp53 double mutant fish.

A. Representative images of the gonad categories in asz1^-/-;tp53^-/- fish. Top panels are brightfield images of the gonads in the bottom panels which are labeled with Ddx4 (green), mAb414 (red), and DAPI (blue). Scale bars are 50 μm. Developing ovaries are thicker, and contain oogonia and differentiating oocytes. Underdeveloped ovaries are thinner and contain oogonia with no further progressing oocytes. Defective testes are gonads that exhibit the asz1^-/- phenotypes, being very thin and containing very few germ cells with abnormal mAb414 signal. These categories are consistent with our previous description of the morphology of normal and defective juvenile ovaries and testes [39]. The distribution of gonad categories per genotype are plotted for the two independent crosses that yielded ovaries (D-E). N = number of gonads. B. RT-qPCR of female-specific (red) and male-specific (blue) marker genes in adult ovaries and testes (B) confirm the sex-specific expression of those markers. Bars are ± SD between independent experiments. C. RT-qPCR on wt and asz1-/- juvenile gonads exhibit female-specific marker gene expression in wt, but reduced female-marker expression and elevated male-specific marker gene expression. Bars are ± SD between independent experiments. F. The distribution of gonad categories in all crosses that did not yield ovaries in asz1^-/-;tp53^-/- fish. A total of 17 rounds of in-crosses between 40 asz1^+/-;tp53^+/- double heterozygous individual fish were performed. In two crosses (panels D-E) a total of 6 asz1^-/-;tp53^-/- ovaries were obtained. In the remaining 15 crosses (panel F), all 180 asz1^-/-;tp53^-/- gonads were defective testes. n = number of gonads. Bars are mean ± SD between independent crosses. G-I. SL of fish from crosses in D-I, respectively.

(TIFF)

S5 Fig. Supporting information for Fig 6.

Overview of gonads of all asz1;tp53 genotypes from the experiment in Fig 6A–6I, labeled for Ddx4 (green), mAb414 (red), and DAPI (blue). Right panels are single and merged channel zoom-in magnifications of the yellow boxes in the overview panels. Scale bars are 50 μm and 10 μm in zoomed out and inset magnification images, respectively.

(TIFF)

S6 Fig. Supporting information for Fig 6.

Overview of gonads of all asz1;tp53 genotypes from the experiment in Fig 6J and 6K, labeled for Buc (red), mAb414 (green), and DAPI (blue). Right panels are single and merged channel zoom-in magnifications of the yellow boxes in the overview panels. Scale bars are 50 μm and 10 μm in zoomed out and inset magnification images, respectively. All gonads (ovaries), show normal Buc localization in the forming Bb, except for the testes in asz1^-/-, and asz1^-/-;tp53^+/-. n = 3-6 gonad per genotype, except 2 asz1^-/-;tp53^-/- ovaries.

(TIFF)

S7 Fig. Supporting information for Fig 6.

Overview of gonads of all asz1;tp53 genotypes from the experiment in Fig 6L and 6M, labeled for dazl (red), DiOC6 (green), and DAPI (blue). Right panels are single and merged channel zoom-in magnifications of the yellow boxes in the overview panels. Scale bars are 50 μm and 10 μm in zoomed out and inset magnification images, respectively. All gonads (ovaries), show normal dazl localization in the forming Bb, except for the testes in asz1^-/-, and asz1^-/-;tp53^+/-. n = 3-5 gonad per genotype, except 2 asz1^-/-;tp53^-/- ovaries.

(TIFF)

Data Availability

All data are in the manuscript and/or supporting information files.

Funding Statement

This work was supported by a research grant from the Israel Science Foundation (No. 558/19 to YME). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.