Skip to main content
NCBI home page
eLife logoLink to eLife
. 2025 Jun 11;14:e100204. doi: 10.7554/eLife.100204

Single-cell transcriptomes of zebrafish germline reveal progenitor types and feminization by Foxl2l

Chen-wei Hsu 1,2,†, Hao Ho 3,4,†, Ching-Hsin Yang 2, Yan-wei Wang 2, Ker-Chau Li 3,4,✉, Bon-chu Chung 1,2,5,6,✉
Editors: John Postlethwait7, Didier YR Stainier8
PMCID: PMC12289312  PMID: 40497446

Abstract

Zebrafish is an important organism for genetic studies, but its early germ cell types and the mechanism of sex differentiation have not been fully characterized. Here, we profiled single-cell transcriptomes and charted a developmental trajectory going from germline stem cells, through early, committed, and late progenitors, to premeiotic and meiotic cells. We showed that the transcription factor Foxl2l was expressed in the progenitor, directing progenitor differentiation toward oocytes. CRISPR-Cas9-mediated mutation of foxl2l produced 100% male fish with normal fertility. Another single-cell profiling of foxl2l-/- germ cells revealed the arrest of germ cell development at the stage of progenitor commitment. Concomitantly, nanos2 transcript (germline stem cell marker) was elevated together with an increase of nanos2+ germ cells in foxl2l mutants, indicating the acquisition of a novel stem cell state. Thus, we have identified developmental stages of germ cells in juvenile zebrafish and demonstrated that zebrafish Foxl2l drives progenitor germ cells toward feminization and prevents them from expressing nanos2.

Research organism: Zebrafish

Introduction

Sex differentiation is an important yet complicated subject. The mammalian sex is determined by the XX/XY system, in which the SRY gene on the Y-chromosome plays a pivotal role (Berta et al., 1990). Females lack SRY. They rely on genes such as RSPO1, FOXL2, and WNT4 to promote female differentiation and to antagonize male differentiation (She and Yang, 2017). In contrast to mammals, fishes have more complicated schemes of sex differentiation. Wild zebrafish strain Nadia uses the female-dominant ZZ/ZW system, but domesticated laboratory strains have no sex chromosomes (Wilson et al., 2014). Furthermore, male and female zebrafish are morphologically indistinguishable at the larval stage, creating an obstacle for the study of fish sex differentiation. Zebrafish oocytes start to develop at around 21 dpf (Tong et al., 2010), while male differentiation appears later at around 30 dpf (Luzio et al., 2021; Uchida et al., 2002). The sex differentiation mechanism in zebrafish remains unclear.

Germ cell abundance and the existence of oocytes play essential roles in the differentiation of zebrafish gonads. Zebrafish containing reduced numbers of germ cells (nr0b1 mutants) or no germ cells (dnd morphants) develop into males (Chen et al., 2016; Slanchev et al., 2005). Mutations of genes (figla, sycp3, or bmp15) involved in oocyte development also result in female-to-male sex reversal (Dranow et al., 2016; Pan et al., 2022; Qin et al., 2018). These results indicate that intrinsic signals from germ cells are required for the female fate. Germ cell types during and after sex differentiation stages have been well characterized by scRNAseq (Liu et al., 2022; Wilson et al., 2024). However, the regulation and subtypes of germ cells before meiosis in zebrafish have not been fully characterized.

One transcription factor that promotes female differentiation is Foxl2l (forkhead box L2-like). Medaka (Oryzias latipes) foxl2l, named foxl3 (Liu et al., 2022), encodes a germline intrinsic factor that suppresses spermatogenesis and initiates oogenesis in XX gonads (Kikuchi et al., 2020). XX medaka fish with a foxl3 mutation become hermaphrodites with both oocyte and functional sperm in adult XX gonads (Nishimura et al., 2015). Zebrafish foxl2l is a marker of progenitors. Its disruption leads to 100% male fish (Liu et al., 2022), signifying the role of Foxl2l in female development. However, the mechanism of Foxl2l action remains poorly understood.

In this study, we conducted single-cell RNA sequencing (scRNAseq) analysis to dissect the developmental trajectory of germ cells in zebrafish during the critical sex determination stage. In addition to the known germline stem cells and meiotic germ cells, we have further identified three subpopulations of progenitor cells: early progenitor, committed progenitor, and late progenitor. We conducted another scRNAseq of foxl2l mutant and proved that Foxl2l controls the development of committed progenitors to late progenitor, which is indispensable for female differentiation. We further show that Foxl2l suppresses stemness-related genes in progenitors.

Results

Developmental trajectory of zebrafish germ cells

Germ cells play important roles in zebrafish sex differentiation, but we still have much to learn about their development. We set out to investigate their development during the initial stage of sex differentiation at 26 dpf (Luzio et al., 2021; Uchida et al., 2002). Fluorescent germ cells were isolated by cell sorting from piwil1:EGFP transgenic fish, which labels all stages of germ cells with EGFP (Leu and Draper, 2010; Ye et al., 2019; Figure 1—figure supplement 1). The cells were used for scRNAseq using the 10X Genomics platform (Figure 1A). After data preprocessing, we obtained single-cell transcriptomes for clustering analysis. The germ cells were partitioned into 14 clusters and visualized on UMAP (Figure 1B). Trajectory analysis identified a linear structure for germ cell development, and pseudotime analysis ordered these clusters as W1 to W14 (Figure 1C). Marker genes from each cluster were found, and their expression profiles were displayed in a heatmap, showing high expression of marker genes in neighboring clusters in a successive pattern (Figure 1—figure supplement 2). We also showed the expression profiles of the top 37 genes with the highest specificity within each cluster out of a total of 1419 marker genes on UMAP separately (Figure 1—figure supplement 3).

Figure 1. Single-cell transcriptome landscape of zebrafish germ cell development.

(A) Flowchart for sample collection and RNA sequencing. (B) Uniform Manifold Approximation and Projection (UMAP) visualization of wild-type (WT) germ cells in 14 clusters. (C) Pseudotime shown by the color bar and the trajectory shown by a red line. (D) Dot plot showing the expression of top marker genes for each cluster. (E) Violin plots displaying S and G2/M cell-cycle scores for each cluster. (F) UMAP visualization of marker gene expression for nanos2, foxl2l, and rec8a. (G,H) Identification of germ cell types by RNA fluorescent in situ hybridization (FISH) of marker genes, nanos2 and foxl2l in (G), rec8a and foxl2l in (H), in 26 days post fertilization (dpf) gonads. Scale bar: 10 µm.

Figure 1.

Figure 1—figure supplement 1. Germ cell sorted by fluorescence-activated cell sorting (FACS).

Figure 1—figure supplement 1.

Sorting of fluorescent germ cells from 26 days post fertilization (dpf) gonad by FACS. Cells within the rectangular or irregular box were selected. Pseudocolors indicate four different gating, P1 to P4. After gating, germ cells from wild-type (WT) (+/+) or foxl2l10i/10i (-/-) with piwil1:egfp transgenic background were collected.
Figure 1—figure supplement 2. Analysis of wild-type (WT) clusters.

Figure 1—figure supplement 2.

Expression of marker genes (Y-axis) in each WT cluster (X-axis) in a heatmap. Genes that reach the criteria (fraction ≥0.8, specificity ≥0.2, q-value ≤0.01) are defined as marker genes. Expression value is visualized with a range between –2 and 2.
Figure 1—figure supplement 3. Expression profiles of top and known marker genes in wild-type (WT) germ cells at 26 days post fertilization (dpf).

Figure 1—figure supplement 3.

Expression of representative WT marker genes shown by Uniform Manifold Approximation and Projection (UMAP). The scales of expression levels are shown at the right of each UMAP.
Figure 1—figure supplement 4. Expression profiles of top and known wild-type (WT) markers extracted from the database.

Figure 1—figure supplement 4.

Expression of representative WT marker genes in 40 days post fertilization (dpf) germ cells obtained from Liu et al., 2022, as shown by Uniform Manifold Approximation and Projection (UMAP). The representative WT markers were selected from our 26 dpf single-cell RNA sequencing (scRNAseq) analysis. The scales of expression levels are shown at the right of each UMAP.
Open in a new tab

Using known markers, we annotated zebrafish germ cells to include three broad categories: germline stem cells (GSCs), progenitors (Prog), and meiotic germ cells. Cells in cluster W1 were identified as GSCs because they express nanos2 (Figure 1D; Beer and Draper, 2013; Cao et al., 2019a). Clusters W3 to W6 were identified as progenitors because they have high cell-cycle scores (Figure 1E), and progenitors are known to divide synchronously forming germline cysts (Bertho et al., 2021). Cluster W2 lies between W1 and W3 with reduced nanos2 expression, indicating they transit from GSC to progenitor (G-P). Cells in clusters W8 to W14 are enriched with genes highly expressed in meiotic oocytes (sycp1, sycp3, dmc1, buc, and zp3e) (Gautier et al., 2013; Pan et al., 2022; Marlow and Mullins, 2008; Wassarman and Litscher, 2021; Yoshida et al., 1998). Cluster W7 lies between progenitors and meiotic cells; it has a high S-phase score typical of Pre-leptotene (Pro-Lept) at the premeiotic S-phase (Figure 1E).

A portion of progenitors (W4 to W6) expressed the progenitor gene foxl2l (Figure 1D), suggesting them as progenitor subtypes. We defined cluster W3 as early progenitors (Prog-E) characterized by the absence of foxl2l, cluster W4 as committed progenitors (Prog-C) characterized by the expression of foxl2l, cluster W5 as Prog-C(S) characterized by the expression of both foxl2l and typical S-phase genes, zgc:153405 and si:ch211-113a14.24 (orthologs of linker histone H1 genes), and cluster W6 as the late progenitor (Prog-L) characterized by the expression of foxl2l and genes for meiotic entry such as rec8a (ortholog of REC8) and meiosin (ortholog of MEIOSIN, the meiosis initiator) (Figure 1D).

In addition to our own data, we also extracted scRNAseq data of 40 dpf female germ cells from a public database (GSE191137) (Liu et al., 2022). Following the same data analysis procedure, the developmental trajectory and marker gene expression profiles were found consistent with ours (Figure 1—figure supplement 4), suggesting the validity of our data and the finer developmental transitions of germ cells in juvenile gonads as charted above.

Histological identification of germ cell progenitors

Zebrafish progenitors are round in shape with one to three nucleoli at the center of the nucleus, and morphologically indistinguishable from GSC (Tong et al., 2010). Our single-cell transcriptome analysis identified markers nanos2, foxl2l, and rec8a, which mark different subtypes of germ cells (Figure 1F). This can be verified by double fluorescence in situ hybridization (FISH). GSC was only positive for nanos2 (foxl2l-nanos2+). Prog-E was negative for both nanos2 and foxl2l (foxl2l-nanos2-), whereas both Prog-C and Prog-L were foxl2l+nanos2- (Figure 1G). Double FISH also detected Prog-C as foxl2l+ rec8a-, Prog-L as foxl2l+ rec8a+, and pre-leptotene cells as foxl2l- rec8a+ (Figure 1H). The FISH result is consistent with the scRNAseq data, which showed little overlap between foxl2l and nanos2 expression in different types of germ cells (Figure 1—figure supplement 3).

Expression of foxl2l becomes female biased during sexual differentiation

We examined the temporal expression of foxl2l by in situ hybridization during the sexual differentiation period. We used the Nadia strain for its defined ZZ/ZW (male/female) sex chromosomes, facilitating the distinction of their sexes at the juvenile stage (Wilson et al., 2014). The foxl2l expression was similar between ZZ and ZW gonads at 11 dpf before apparent sex differentiation (Figure 2A). Upon female differentiation at 22 dpf, foxl2l expression was decreased in the male ZZ gonad. The female-biased expression became more obvious at 30 dpf. In the adult ovary at 5 months post fertilization, foxl2l was highly expressed in cystic germ cells but was not observed in testis (Figure 2A).

Figure 2. Foxl2l is required for female differentiation.

(A) In situ hybridization detects foxl2l transcripts in indifferent gonads at 11 days post fertilization (dpf) (a, a’). Expression of foxl2l becomes female biased when female differentiation initiates at 22 dpf (b, b’), and becomes female-specific at the initiation of male differentiation at 30 dpf (c, c’). In adults, foxl2l is expressed in cystic cells in ovary but not in testis (d, d’). mpf: months post fertilization. All samples were collected from Nadia strain. Scale bars are 50 µm in (a, a’), 100 µm in (b, b’, c, c’), 20 µm in (d, d’). (B) Domain structure of wild-type (WT) and the predicted mutant of Foxl2l with 10 base pairs insertion (10i) or 22 base pairs insertion (22i) generated by CRISPR/Cas9 system. (C) Sex ratios of WT (+/+), foxl2l+/22i heterozygous (+/-), and foxl2l22i/22i homozygous (-/-) mutants at 4 months of age reveal that foxl2l mutant fish are all males. The total numbers of fish in each genotype and gender are labeled in white. Dot: sex ratio of each batch. (D) All foxl2l10i/10i mutant males (-/-) are fertile. Dot: individual fish used in the test, n=3 in WT and heterozygous mutants, n=6 in homozygous mutants. (E) The foxl222i/22i homozygous (-/-) mutants lack meiotic oocytes at 20 dpf shown by histological staining. Black dashed circle: germ cell cysts with one to three nucleoli in each cell. Yellow dashed circle: meiotic oocyte. Scale bars represent 20 µm. (F) The foxl2l22i/22i mutants (-/-) express male markers (tekt1 or odf3b) detected by RNA fluorescence in situ hybridization (FISH) at 28 dpf. Scale bars are 10 µm.

Figure 2.

Figure 2—figure supplement 1. Generation and the phenotype analysis of foxl2l mutant.

Figure 2—figure supplement 1.

(A) The foxl2l sequence of wild-type (WT) and two foxl2l mutant lines, 10 base pairs insertion (10i) and 22 base pairs insertion (22i). Gray letters represent gRNA targeting region. Red letters represent stop codons. (B) Sequencing results of WT and two foxl2l mutant lines, 10i and 22i. The inserted sequences are marked with horizontal orange lines.
Open in a new tab

All foxl2l mutants fail to enter female meiosis and become fertile males

To investigate the function of foxl2l, we generated foxl2l mutants using CRISPR-Cas9. Two resultant mutant lines, foxl2l10i/10i and foxl2l22i/22i, contained DNA insertions (Figure 2—figure supplement 1), causing a premature stop codon disrupting the forkhead (FH) domain of Foxl2l (Figure 2B). All foxl2l mutants became males (Figure 2C), and all foxl2l mutant males were fertile (Figure 2D). We previously also showed that the mutant and wild-type (WT) testes are similar in histology (Yang et al., 2025).

We further examined the histology of mutant gonads. At 20 dpf, diplotene oocytes with large perinucleolar nuclei were observed in WT gonads, but only non-meiotic cystic germ cells were found in foxl2l mutants (Figure 2E). At 28 dpf, small and round germ cells that express spermatogenic markers, tekt1 and odf3b (Nishimura et al., 2015), were found in mutant but not WT gonads (Figure 2F). Therefore, in foxl2l mutants, germ cells fail to enter female meiosis and develop into fertile males. This indicates that foxl2l is required for female but not for male development.

The development of foxl2l mutant germ cells is halted at the committed progenitor stage

We conducted scRNAseq from 26 dpf foxl2l-/- germ cells to investigate their developmental stages via transcriptomic profiling. The expression profiles of the top 37 WT marker genes in mutant cells were visualized via UMAP (Figure 3—figure supplement 1A). Marker genes corresponding to WT clusters of GSC, G-P, Prog-E, and Prog-C were upregulated in distinct subsets of mutant cells, whereas markers associated with stages beyond Prog-L were expressed at consistently low levels. Accordingly, we inferred the developmental stages of mutant cells by aligning their transcriptomic profiles with those of WT cells (Figure 3—figure supplement 1B, C).

Pie charts showed that most of the mutant cells were annotated as GSC, G-P, Prog-E, or Prog-C, with only a few as Prog-C(S), the stage preceding Prog-L differentiation (Figure 3A). None of the mutant cells were at Prog-L or meiotic stages (Figure 3A). This indicates that mutant germ cells cannot progress beyond Prog-C. To verify this statement, we used fluorescent staining and detected Prog-L marker rec8a in cystic germ cells and meiotic marker Sycp3 in meiotic cells of the control (Figure 3B). Both rec8a and Sycp3 staining were missing in foxl2l mutant gonads. These data show that mutant germ cells do not enter the state of Prog-L. The components of WT cells from GSC to Prog-C stages were examined more closely (Figure 3—figure supplement 2A). The data of these cells were used for further analysis with mutant germ cells.

Figure 3. Germ cells in foxl2l mutants are arrested at the stage of committed progenitor with aberrant gene expression.

(A) Pie charts showing the proportion of mutant cells classified into different developmental stages. (B) Absence of the Prog-L marker rec8a and meiotic marker Sycp3 in foxl2l mutant gonads. (Left) RNA fluorescence in situ hybridization (FISH) and (right) immunofluorescence staining of wild-type (WT) (+/+), foxl2l+/10i (+/-), and foxl2l10i/10i (-/-) gonads at 28 days post fertilization (dpf) (left) and 21 dpf (right), respectively. The numbers at the bottom right corner indicate that all three gonads in each genotype have the same staining pattern. Scale bars represent 10 µm. (C) Uniform Manifold Approximation and Projection (UMAP) visualization of integrated cells grouped into eight clusters. Circles with black outline: WT cell. Circles with gray outline: mutant cell. (D) Bar graphs showing the proportions of developmental stages for WT and mutant cells in each integrated group. Prominent stages with statistical significance were marked by asterisks (*). (E) Dot plot showing the expression of marker genes (Y-axis) in the mutant cells across integrated groups (X-axis). (F) foxl2l10i/10i (-/-) gonads contain an increased number of foxl2l+ cells at 26 dpf. Left: in situ hybridization of the gonads with foxl2l. Right: quantification of the staining data. One dot in the graph represents one gonad, n=6 in each genotype.

Figure 3.

Figure 3—figure supplement 1. Transcriptome analysis of mutant germ cells.

Figure 3—figure supplement 1.

(A) Expression of 27 wild-type (WT) top marker genes and mutant top marker genes (in orange box) in mutant cells shown by Uniform Manifold Approximation and Projection (UMAP). The scales of the scores are shown at the right. (B) Flowchart of the inferring developmental stages of mutant cells. (C) Mutant cells shown in UMAP. Color of each mutant cell represents annotated developmental stage.
Figure 3—figure supplement 2. Transcriptome analysis of integrated groups.

Figure 3—figure supplement 2.

(A) Wild-type (WT) germ cells in clusters W1 to W5 visualized by Uniform Manifold Approximation and Projection (UMAP). (B) The flowchart of the co-clustering analysis. (C) Split-violin plots showing the distribution of cells expressing four top germline stem cell (GSC) markers in integrated groups I1 and I2.
Figure 3—figure supplement 3. The flowchart of data processing and cell-cycle scoring for wild-type (WT) and mutant single-cell RNA sequencing (scRNAseq) transcriptome.

Figure 3—figure supplement 3.

Figure 3—figure supplement 4. The flowchart of cell clustering analysis for wild-type (WT) cells.

Figure 3—figure supplement 4.

Open in a new tab

Altered developmental program in foxl2l mutant

The mutant germ cells exhibited heterogeneity, encompassing diverse cell types. To better understand mutant subtypes and their relationship with WT cells, we conducted a co-clustering analysis of the full transcriptomic profiles of both WT cells (GSC to Prog-C(S) stages) and mutant cells, as outlined in Figure 3—figure supplement 2B. This approach does not assume that mutant cells correspond directly to specific WT stages; instead, it enables the identification of potentially novel germ cell types. This analysis also provides insight into the relationships between WT and mutant cells, resulting in eight integrated groups (I1-I8) (Figure 3C). WT and mutant cells with similar transcriptomic profiles were grouped together within the same integrated cluster.

We examined the distribution of cells at different stages across different integrated groups (Figure 3D). Prominent stages were identified after calibrating against their prevalence rates within the entire WT or mutant cell population. For WT cells, we found WT GSC is prominent in groups I1 and I2, G-P is prominent in group I4, Prog-E is prominent in group I5, Prog-C is prominent in group I7, and Prog-C(S) is prominent in group I8 (Figure 3D).

For mutant cells, groups I5-I8 had two prominent stages that are similar to those in WT (Figure 3D). This indicates that WT and mutant cells that have high transcriptome similarity likely belong to the same developmental stage. Groups I1 and I2 were dominated by GSC, but the expression patterns of top GSC marker genes were slightly different with higher expression of nanos2 in I1 but of cbx7a, psmb13a, and dnmt3bb.1 in I2 (Figure 3—figure supplement 2C). This suggests that GSCs can be further divided into two finer stages, GSCI and GSCII.

Groups I3 and I6 are different from the rest of the groups. They contained few WT cells and were classified as GSC for I3 and as Prog-E for I6 mutant cells (Figure 3D). In addition to expressing the GSC marker, I3 mutant cells expressed zgc:158463; I6 mutant cells expressed tp53 and zgc:158463 in addition to Prog-E markers (Figure 3E). This indicates that I3 and I6 mutant cells expressed different genes from WT GSCs and progenitors, potentially leading to a unique cell fate.

Furthermore, the Prog-C marker foxl2l was expressed in mutant cells of I1, I5, I7, and I8, which were mostly classified as GSC or Prog-E in addition to Prog-C (Figure 4E). The number of foxl2l+ cells detected by FISH was also increased in mutant gonads compared to WT (Figure 3). This result indicates that mutant foxl2l mRNA remains stable. The aberrant foxl2l expression in some earlier GSC and Prog-E, the accumulation of foxl2l-expressing cells in the mutant, and the failure to find clusters representing more mature stages indicate that progenitor development is impaired in foxl2l mutants.

Figure 4. Identification of differentially expressed genes (DEGs) in wild-type (WT) and foxl2l mutants.

(A) Uniform Manifold Approximation and Projection (UMAP) visualization of integrated cells, with six matched stages of WT and mutant cells displayed in different colors. Unmatched cells are labeled in gray. Circles with a black outline: WT cells. Circles with gray outline: mutant cells. (B) Heatmap showing the fold change (FC) between the mutant versus WT (mutant/WT) for all DEGs at each stage. Developmental stages are displayed on the X-axis, and gene names are displayed on the Y-axis. (C) Split-violin plots showing the distribution of cells expressing top DEGs at different development stages (X-axis) between WT and mutant. Asterisks in (B) and (C) indicate significant differences in gene expression between WT and mutants.

Figure 4.

Figure 4—figure supplement 1. Expression of differentially expressed genes (DEGs) in wild-type (WT) and mutant germ cells.

Figure 4—figure supplement 1.

Split-violin plots showing the distribution of WT and mutant cells expressing DEGs at different developmental stages (X-axis).
Open in a new tab

Differential gene expression between WT and foxl2l mutant germ cells

We further evaluated the effect of foxl2l mutation on gene expression throughout development. We compared WT and mutant cells at the same prominent stages within each integrated group, which is defined as a stage-matched set, color-coded GSCI, GSCII, G-P, Prog-E, Prog-C, and Prog-C(S) (Figure 4A). Figure 4B shows all the differentially expressed genes (DEGs) of mutants relative to WT (mutant/WT). Prog-C(S) contained the largest number of DEGs, consistent with the failure of mutant Prog-C(S) cells to differentiate further to Prog-L (Figure 4B). We also unexpectedly detected foxl2l transcripts in mutant Prog-E (Figure 4C). Concomitantly, the expression of stk31 was decreased and that of ptmab was increased upon transition to Prog-C in the WT cells; yet these genes were steadily expressed in the mutant progenitors (Figure 4—figure supplement 1). This pinpoints the defect in foxl2l mutant germ cells during the transition from Prog-E to Prog-C.

The top DEGs, dmrt1, dele1, and id1, and the GSC marker nanos2 were further analyzed (Figure 4C). The id1 and nanos2 transcripts were low in the WT progenitors but became abundant in mutant Prog-C (Figure 4C). The expression of dmrt1 and dele1 in WT was steady throughout the developmental stages (Figure 4C), but dmrt1 was abnormally high, and dele1 was abnormally low in both mutant GSCI and Prog-C/Prog-C(S). These results indicate multiple defects of gene expression in the foxl2l mutants and suggest diverse roles of Foxl2l, as an inhibitor for dmrt1 but activator for dele1.

Co-expression of foxl2l with dmrt1, nanos2, or id1 in foxl2l mutants

The expression patterns of foxl2l along with nanos2, id1, and dmrt1 at the early stages were further examined. The nanos2 and foxl2l genes were distinctly expressed among WT cells, but co-expressed in some mutant cells (Figure 5A). Double FISH also detected cells expressing both nanos2 and foxl2l, and the proportion of nanos2+foxl2l+ cells among foxl2l+ cells was increased in the mutant (Figure 5B). Moreover, the number of nanos2+ cells was increased in mutants at 21 dpf during the early stage of female differentiation (Figure 5C).

Figure 5. Aberrant co-expression of foxl2l with dmrt1, id1, and nanos2.

(A) Uniform Manifold Approximation and Projection (UMAP) visualization of dmrt1+, nanos2+, or id1+ cells and foxl2l+ cells. Left: wild-type (WT) cells in clusters W1 to W5 (+/+). Right: foxl2l mutant (-/-) cells. Yellow dots indicate cells co-expressing both genes. (B) Increased proportions of foxl2l mutant cells co-expressing foxl2l with nanos2 or id1. (Left) Double RNA fluorescence in situ hybridization (FISH) detecting nanos2 or id1 with foxl2l in WT (+/+) or foxl2l10i/10i homozygous mutant (-/-) gonads at 26 days post fertilization (dpf). Stained images are shown with (top panels) or without (bottom panels) DAPI in each set. Asterisks represent the expression of id1 in somatic cells. Scale bars represent 5 µm. (Right) Quantitation of the proportion of double positive cells in each gonad is shown. One dot represents the data from one gonad, n=3 in each genotype of each graph. (C) Increased numbers of nanos2-expressing cells in foxl2l10i/10i homozygous mutant gonad (-/-). One dot represents the data from one gonad, n=3 in each genotype of each age. (D) Increased number of nanos2+foxl2l- germline stem cells (GSCs) in foxl2l10i/10i homozygous mutant gonad (-/-). One dot represents the data from one gonad, n=3 in each genotype.

Figure 5.

Figure 5—figure supplement 1. Very few germ cells express foxl2l together with nanos2, id1, or dmrt1 in 40 days post fertilization (dpf) wild-type (WT) gonad.

Figure 5—figure supplement 1.

Expression profile of foxl2l with nanos2, id1, or dmrt1 in 40 dpf germ cells shown in Uniform Manifold Approximation and Projection (UMAP). The data was obtained from a published database (GSE191137). The expression values are visualized by colors ranging between 4 and 0.
Open in a new tab

The co-expression of the GSC marker, nanos2, in the foxl2l-expressing Prog-C in mutants raised the hypothesis that GSCs may continue the GSC program during the differentiation into progenitors. To validate this hypothesis, we examined the expression profiles of the GSC marker genes, nanos2, in all mutant cells and found that nanos2 expression was low in most mutant cells classified as G-P and Prog-E (mutant cells of I4 to I6 and I8) (Figure 3E). Other GSC markers including cbx7a, psmb13a, and dnmt3bb.1 also showed low expression in those mutant cells mostly classified as different types of progenitors (mutant cells of I4 to I8) (Figure 3E). This result indicates that the GSC program in mutant cells does not continue when GSCs are differentiated into progenitors; instead, some foxl2l-expressing Prog-C may acquire some GSC properties by upregulating nanos2 in mutants.

To further investigate whether the upregulation of GSC marker nanos2 in mutant Prog-C alters cell property, we counted the number of nanos2+foxl2l- GSC cells and found it was increased in mutants at 26 dpf compared with WT (Figure 5D). Since foxl2l is not expressed in WT GSCs, the increased GSC number in foxl2l mutants may be due to either direct reversion from Prog-C to GSCs or the proliferation of GSCs triggered by mutant Prog-C. Another gene, id1, was barely expressed in WT germ cells but was abundant in some foxl2l+mutant germ cells (Figure 5A). Double FISH further detected an increased proportion of id1+foxl2l+ cells in the mutant (Figure 5B). Id1 is involved in the maintenance of stemness in various organs (Kantzer et al., 2022; Ying et al., 2003; Zhang et al., 2014), consistent with the hypothesis that mutant Prog-C reverts to the GSC property.

Dmrt1 is a male regulator in many vertebrate species (Ge et al., 2017; Matson et al., 2011; Smith et al., 2009). In WT, very few cells co-expressed foxl2l and dmrt1. However, many mutant cells expressed both foxl2l and dmrt1 (Figure 5A). Similar results can also be observed in scRNAseq data extracted from a public database (GSE191137), which showed very few 40 dpf female germ cells co-expressed foxl2l with nanos2, id1, or dmrt1 (Figure 5—figure supplement 1).

Discussion

In this article, we have analyzed the development of zebrafish germ cells and elucidated the mechanism controlling sex differentiation in zebrafish by analyzing foxl2l mutants. The loss of functional Foxl2l leads to gene dysregulation, a halt of Prog-C development, aberrant upregulation of GSC marker gene nanos2, and eventual male development. These results show that Foxl2l is required for the differentiation of Prog-C to Prog-L, and therefore guides the progenitor toward the female pathway. Foxl2l probably also prevents progenitors from reverting to the stem cell fate as depicted in Figure 6.

Figure 6. Schematic illustration of the function of Foxl2l during germ cell development in juvenile gonad.

Figure 6.

Open in a new tab

During the sex determination period, germ cells develop following the trajectory from germline stem cell (GSC), GSC to progenitor (G–P) transition, early progenitor (Prog-E), committed progenitor (Prog-C), late progenitor (Prog-L) to the female meiotic stage. Foxl2l is essential for the maturation of Prog-C and the ensuing Prog-L development. In foxl2l mutant, germ cells are arrested at Prog-C. Some mutant Prog-C cells (mProg-C) express nanos2. The absence of functional Foxl2l eventually triggers male development. Blue symbols indicate alternative development in foxl2l mutants. Red arrows indicate the developmental trajectory of females.

Developmental trajectory of germ cells

The process of zebrafish early germ cell development during the sex differentiation period is poorly understood. Here with scRNAseq analysis, we have depicted the developmental trajectory of germ cells and further characterized GSCs and progenitors. GSCs can be split into two subtypes: GSC-I (nanos2hicbx7alo) and GSC-II (nanos2locbx7ahi). GSCs can self-renew intermittently through two types of cell division to ensure continuous gamete maintenance (Saito et al., 2007). One type of cell division is symmetric, producing two identical stem cells. The other type is asymmetric division producing two different daughter cells, one stem cell and the other cell preparing for further differentiation. The two subtypes of GSCs may represent two different daughter cells derived from asymmetric division. This speculation needs to be validated by further analysis. When GSCs lose stem cell property, they differentiate into progenitors.

Progenitors include three types: Prog-E (foxl2l-nanos2-), Prog-C (foxl2l+rec8a-), and Prog-L (foxl2l+rec8a+). We show here that Prog-L is essential for female differentiation. Prog-C is the gate before female differentiation and will stay at S-phase temporarily before developing into Prog-L. In Prog-L, genes such as rec8a and meiosin are highly expressed for the preparation of meiotic entry. Our study delineates the developmental process of zebrafish germ cells during sex differentiation.

A type of zebrafish germ cell progenitor possesses female identity at juvenile age

In many vertebrate species, sex determination is induced by sex-determining genes expressed in supporting cells (Berta et al., 1990; Matsuda et al., 2002). The fate of germ cells is determined following the instruction of supporting cells. In mammals, pre-Sertoli cells express the male sex-determining gene Sry to activate a cascade of male-related genes, including Sox9 and Amh, and to trigger male fate decision (She and Yang, 2017). In female supporting cell precursors, the absence of Sry expression enables the expression of female genes and female fate determination. In contrast, germ cell abundance is required for female development in zebrafish, and the early depletion of germ cells at the embryonic stage results in male development (Campbell et al., 2015; Dai et al., 2017; Draper et al., 2007; Slanchev et al., 2005; Zhou et al., 2018). Our study further determines early establishment of zebrafish germline feminization in Prog-L. Foxl2l drives the differentiation of Prog-C to Prog-L and, therefore, dictates female development in juvenile gonads.

Medaka Foxl2l controls genes in meiosis and folliculogenesis (Kikuchi et al., 2020). Zebrafish Foxl2l may have the same functions as shown by our previous RNAseq result (Yang et al., 2025) and the expression of foxl2l in Prog-L (Figure 1). However, there are differences regarding the mutant phenotypes in zebrafish and medaka. Medaka foxl2l (termed foxl3) mutant gonads still contain some meiotic oocytes (Nishimura et al., 2015). Zebrafish foxl2l deficiency, however, leads to germ cell arrest at the Prog-C stage as shown here (Figure 3A). No meiotic oocytes can be detected in zebrafish foxl2l-/- gonads. The difference in foxl2l mutant phenotypes between zebrafish and medaka raises two possibilities. First, zebrafish Foxl2l has an additional role in driving progenitor differentiation. Second, medaka Foxl2l plays a similar role in progenitor differentiation as in zebrafish; however, medaka may have a stronger female genetic tendency that prevents full blockage of female differentiation and meiosis in foxl3-/- gonads.

Direct male differentiation from indifferent gonads

Zebrafish testis development has been an unsolved issue. Zebrafish testis can be derived from ovary-like tissue that undergoes oocyte apoptosis during sex differentiation (Uchida et al., 2002). Here, we show that germ cells are arrested in the middle of progenitor development followed by direct male differentiation in foxl2l mutant. This indicates that zebrafish testis can be differentiated directly from indifferent gonads without going through the female phase. Our current result is also consistent with the reports showing direct male differentiation in zebrafish (Pan et al., 2022; Tong et al., 2010). The wild zebrafish strain, Nadia, with ZZ sex chromosomes undergoes direct male development without going through female-to-male transition (Wilson et al., 2024). Therefore, in addition to female-to-male transition, zebrafish males can develop directly similar to most gonocharists whose males and females are derived directly from undifferentiated gonads.

Foxl2l suppresses genes in male development and stemness

Our scRNAseq analysis has identified nanos2, dmrt1, and id1, which may be involved in cell-type transition in foxl2l mutants. Nanos2 functions in germ cell development in various species (Cao et al., 2019a; Kusz et al., 2009; Tsuda et al., 2003). NANOS2 also suppresses meiosis in male mice (Suzuki and Saga, 2008). Therefore, the aberrantly induced nanos2 in Prog-C of foxl2l mutant may participate in the suppression of meiotic genes in addition to the induction of stemness.

Id1, inhibitor of DNA binding 1, is a transcription regulator that lacks a DNA binding domain. It can bind basic helix-loop-helix proteins to prevent their binding to DNA (Benezra et al., 1990). Id1 is essential for homeostasis and the maintenance of stem cells in several cell types (Hong et al., 2011; Kantzer et al., 2022; Ying et al., 2003; Zhang et al., 2014). Although the function of Id1 in the testis remains unclear, its paralog Id4 regulates the self-renewal of spermatogonial stem cells in the testis (Sablitzky et al., 1998; Wang et al., 2018). Thus, the enhanced id1 expression in the mutant Prog-C germ cells may inhibit genes controlling differentiation and guide cells toward stem cell fate in the foxl2l mutant.

Dmrt1 regulates male differentiation in many animals. In Nile tilapia (Oreochromis niloticus), Dmrt1 and Foxl2l antagonize each other to determine germline sex (Dai et al., 2021). Zebrafish appear to behave similarly. As a male-promoting gene (Webster et al., 2017), zebrafish dmrt1 is aberrantly upregulated in foxl2l mutants at the Prog-C, Prog-C(S), and GSCI stage (Figure 4C). This aberrant dmrt1 expression may direct foxl2l mutants toward male development.

We show here that Foxl2l regulates genes important for progenitor differentiation from Prog-C to Prog-L (Figure 4B); it also suppresses GSC gene nanos2 and antagonizes male development by suppressing dmrt1. In foxl2l mutants, most of the DEGs were detected in Prog-C/Prog-C(S), consistent with the failure of the Prog-C differentiation to Prog-L. Foxl2l facilitates the expression of oocyte genes to ensure female differentiation.

Our DEG analysis identified both upregulated and downregulated genes in foxl2l mutants (Figure 4B), suggesting that Foxl2l can activate or suppress transcription. The dual role of Foxl2l is comparable with that of other Fox family proteins. The forkhead transcription factor L2 (forkhead box L2, FOXL2) is highly expressed in granulosa cells required for female differentiation, including meiosis and folliculogenesis (Georges et al., 2014). In mammalian granulosa cells, Foxl2l activates Fst to facilitate folliculogenesis (Kashimada et al., 2011). However, in undifferentiated pre-granulosa cells, FOXL2 suppresses granulosa differentiation and proliferation by repressing StAR, CYP11A, CYP19, and CCD2 (Kuo et al., 2011; Kuo et al., 2012; Pisarska et al., 2004; Pisarska et al., 2010). In juvenile zebrafish gonads, the activation and suppression of genes involved in progenitor development by Foxl2l results in eventual gonad differentiation.

Altered development of foxl2l mutant germ cells

We have detected aberrantly enriched tp53 and zgc:158463 in distinct foxl2l mutant cells in I3 or I6. Tumor suppressor protein p53, encoded by TP53, induces cell-cycle arrest and apoptosis. Mutation of genes such as zar1, fancl, and brca2 causing apoptosis of either early germ cells or developing oocytes results in eventual male development. This mutant phenotype can be rescued by additional tp53 mutation in these mutants (Rodríguez-Marí et al., 2010; Rodríguez-Marí et al., 2011; Shive et al., 2010). However, additional mutation of tp53 fails to rescue the sexual-reversal phenotype in other mutants, including zar1, figla, vasa, and rbpms2, that disrupt either early germ cell development or the beginning of female meiotic entry (Bertho et al., 2021; Hartung et al., 2014; Kaufman et al., 2018; Miao et al., 2017; Qin et al., 2018). Our foxl2l mutation results in the disruption of early germ cell development, and tp53-/-foxl2l-/- double mutant fails to rescue the all-male phenotype (Yang et al., 2025). In addition to inducing apoptosis, p53 also regulates the differentiation of airway epithelial progenitors (McConnell et al., 2016), embryonic stem cells (Lin et al., 2005), and brown adipocytes (Molchadsky et al., 2013). Thus, p53 may participate in the alternative differentiation of mutant germ cells. We have also identified an uncharacterized gene, zgc:158463, which is specifically expressed in foxl2l mutant cells in I3 and I6. Its high expression suggests that zgc:158463 may be involved in the male differentiation of mutant cells. This assumption still needs to be tested by more functional studies.

Materials and methods

Animals

WT zebrafish Danio rerio, TL and Nadia strains, piwil1:egfp transgenic (Pan et al., 2022) and foxl2l-/- mutant fish were maintained in system water (28.5°C, pH 7.2, conductivity 500 µS/cm) according to the standard protocol (Westerfield, 2000). For embryos, fertilized eggs were raised in 100 mm Petri dish with 100 eggs per dish until 4 dpf. Larvae were then transferred to a 1 l beaker with 60 larvae in 600 ml system water and fed with paramecia until 12 dpf. At 8 dpf, larvae density was reduced to a half and fed with brine shrimps twice daily. At 21 dpf, larvae were transferred to the system tank. For sample collection, the growth of larvae was determined according to both age and body length (Nüsslein-Volhard and Ralf, 2002). Zebrafish were handled under the guidelines of the Institutional Animal Care and Utilization Committee of Academia Sinica (IACUC: 10-10-084) and National Laboratory Animal Center of National Applied Research Laboratories (IACUC: NLAC-113-M-004-R1).

Generation of foxl2l mutant

The gRNA sequences targeting the FH domain of foxl2l were designed using CHOPCHOP (https://chopchop.cbu.uib.no) and CRISPRscan (https://www.crisprscan.org). Target sites with high efficiency, low off-target effect, and the sequence of NGG at the 3’ end were chosen. To get higher knockout efficiency, two gRNAs with 50 pg each and 200 pg of TrueCut Cas9 Protein v2 (Life Technologies, Cat. No. A36498) were co-injected into the embryo at the two-cell stage. F0 fish were crossed with WT fish to generate stable mutant lines. To determine the genotype, the fin of F1 offspring was heated in 50 mM NaOH at 95°C followed by the addition of 1/10 volume of 1 M Tris-HCl pH 8 on ice for 5 min before DNA amplification by PCR (F, 5’-AGTAAACCTGAAGCACACCTGG-3’, R, 5’-CATCCCTTTTTGTTCTTCTCGT-3’) using Fast-RunTM 2x Taq Master Mix without Dye (Protech Technology Enterprise CO., Cat. No. PT-TMM228-D). The size difference between WT and mutant alleles was determined from capillary electrophoresis, run by QIAxcel DNA Screening Kit (QIAGEN, Cat. No. 929004), detected by High Performance Nucleic Acid Analyzer (eGENE HDA-GT12) and analyzed by QIAxcel BioCalculator software.

RNA analysis and plasmid

Tissue RNA was extracted by TRIzol Reagent (Invitrogen, Cat. No. 15596026). cDNA was reverse-transcribed by Maxima First Strand cDNA Synthesis kit (Fermentas Int, Cat. No. K1641) from 1 µg of RNA.

For probe synthesis, templates were either constructed into vector (for foxl2l, nanos2, and id1) or amplified by PCR (for rec8a, obf3b, and tekt1). For construction, cDNA form ovary (for foxl2l and nanos2 probe) or 24 hpf larvae (hours post fertilization) (for id1) (Addgene) were amplified by PCR (foxl2l: F, 5’-CTTTCCACCTGTACCGTGCG-3’, R, 5’-CAGTCAGCACCGAGGTTTGC-3’; nanos2: F, 5’-GACGGATCCCATGGGCAAAACACACCTAAAACA-3’, R, id1: F, 5’-TGGTGAACTGTCATCGCACT-3’, R, 5’-AGCGTTCACATCATATGGCA-3’) using Taq DNA polymerase (Roche, Cat. No. 11146165001). Fragments of foxl2l and id1 were cloned into pGEMT-easy plasmid by TA cloning, while the nanos2 fragment was cloned into pCS2+ plasmid by BamHI and KpnI. Plasmid was cut, and the antisense probe was synthesized by in vitro transcription with Dig RNA Labeling Mix (Roche, Cat. No. 11277073910) or Fluorescein RNA labeling mix (Roche, Cat. No. 11685619910) together with T7 RNA polymerase (Roche, Cat. No. 10881775001), SP6 RNA polymerase (Roche, Cat. No. 10810274001), or T3 RNA polymerase (Roche, Cat. No. 11031163001) (foxl2l: NcoI, SP6; nanos2: BamHI, T3; id1: SacII, SP6). For rec8a, obf3b, and tekt1, cDNA containing T7 promoter was amplified by PCR (rec8a: F, GAGTATTTAGGTGACACTATAGACAATTCCCCCTCAGCAACC, R, GAGTAATACGACTCACTATAGGGGATGCACCGGTGATTTGTGC; obf3b: F, 5’-GAGTATTTAGGTGACACTATAGGGGGCAACTGGAATGAATAA-3’, R, 5’-GAGTAATACGACTCACTATAGGGACTACGACCGCTGAAGGAGA-3’; tekt1: F, 5’-GAGTATTTAGGTGACACTATAGGGAGGATCCAGGACATCAAA-3’, R, 5’-GAGTAATACGACTCACTATAGGGCCTTCTCGGCTTTGCTAATG-3’) using Fast-RunTM 2x Taq Master Mix without Dye (Protech Technology Enterprise CO., Cat. No. PT-TMM228-D). After purification by QIAquick PCR Purification Kit (QIAGEN, Cat. No. 28106), antisense probe was synthesized by in vitro transcription as described above.

For the generation of foxl2l mutant lines, the pT7-gRNA-foxl2l#1 and pT7-gRNA-foxl2l#2 plasmids containing oligo against two different target sites of foxl2l were constructed. After annealing oligo pairs (foxl2l#1: F, 5’-TAGGAGCTGGATGAATGAAACG-3’; R, 5’-AAACCGTTTCATTCATCCAGCT-3’; foxl2l#2: F, 5’-TAGGAGCCACGTACGAATAAGG-3’, R, 5’-AAACCCTTATTCGTACGTGGCT-3’) foxl2l#1 and foxl2l#2 fragments were cloned into pT7-gRNA (Addgene) through one-step digestion and ligation, respectively. Reagents (400 ng of pT7-gRNA, 0.25 µM annealed oligo), 1x NEBuffer 3.1 (New England Biolabs, Cat. No. B7203), 1x T4 DNA ligase buffer (Promega Co., Cat. No. C126B), 1.5 U T4 DNA ligase (Promega Co., Cat. No. M180A), 5 U BsmBI (New England Biolabs, Cat. No. R0580), 3 U BglII (New England Biolabs, Cat. No. R0144S), and 6 U SalI (New England Biolabs, Cat. No. R0138S) were incubated with the following condition: 20 min in 37°C and 15 min in 16°C for three cycles, followed by 10 min in 37°C, 15 min in 55°C, 15 min in 80°C, and cooling in 4°C, for plasmid construction. Then, the plasmids were linearized by BamHI, and gRNA was synthesized by MegaShortscript T7 Transcription Kit (Thermo Fisher Scientific, Cat. No. AM1354).

RNA in situ hybridization

Whole-mount in situ hybridization is performed using a published protocol (Thisse and Thisse, 2008). Larvae or tissues were fixed in 4% paraformaldehyde (PFA) overnight at 4°C before the removal of head, tail, and intestine. The dissected samples went through dehydration, rehydration, permeation, hybridization, antibody incubation, and staining. For permeation, larvae were treated with 10 µg/ml Proteinase K (Roche, Cat. No. 03115887001) for 10 min while adult tissues were treated for 40 min. For probe hybridization, samples were incubated with 1 ng/µl digoxigenin (DIG)-labeled probe at 70°C overnight. For color staining in situ hybridization, the antibody incubation was performed with 1:5000 diluted Anti-DIG-AP antibody (Roche, Cat. No. 11093274910) at 4°C overnight. Samples were stained with BM-purple (Roche, Cat. No. 11442074001). FISH was performed as described (Brend and Holley, 2009) with 555 Styramide Kit (AAT Bioquest Inc, Cat. No. 45027) and 488 Tyramide Conjugate (Biotium Inc, Cat. No. 92171) followed by DAPI staining with 1:20,000 dilution. For photography, samples were mounted in 85% glycerol.

Photography

Photography was performed using AxioImager Z1 upright microscope (Carl Zeiss Inc) with AxioCam HRc digital camera. Image was processed by Axiovision 4.7 software (Carl Zeiss Inc). For fluorescent samples, photography was performed by LSM 780 Confocal Microscope (Carl Zeiss Inc).

Mating test

Adult females with different genotypes were mated with WT males individually from 4 months of age. Ten matings were performed with 7- to 14-day intervals. To avoid individual differences, the first three matings were considered as practice. Only the following seven matings were included in the experimental counting. Fertilization rate was calculated as the number of fertilized eggs/total number of eggs at 24 hpf. Each experimental group contained three to seven mating pairs.

Histology

Larvae, ovary, and testis were fixed in Bouin’s solution at 4°C overnight before six washes in phosphate-buffered saline (PBS) for 30 min each, serial dehydration from 50% to 100% ethanol and incubation with xylene and paraffin twice each. Then, the samples were embedded in paraffin before being sectioned with 6 µm thickness. The sectioned samples were stained with hematoxylin and eosin for histological observation.

Immunofluorescence

Larvae were fixed in 4% PFA overnight before removal of head, tail, and intestine. The dissected samples went through three washes of PBS with 0.2% Triton X-100 (PBST) for 5 min each, blocking with normal goat serum for 1 hr at room temperature, incubating with primary antibody (Vasa [GeneTEX, Cat. No. GTX128306, RRID:AB_2847856], 1:200 dilution; Sycp3 [Abcam, Cat. No. ab150292], 1:100 dilution) 4°C overnight, three washes in PBST for 30 min each and incubating with 1:200 diluted secondary antibody (Alexa Fluor 546 Donkey anti-Rabbit IgG [Invitrogen, Cat. No. A-10040, RRID:AB_2534016]; Alexa Fluor 488 Goat anti-Rabbit IgG [Invitrogen, Cat. No. A-11034, RRID:AB_2576217]) together with 1:20,000 diluted DAPI. Finally, the samples were mounted in 85% glycerol for photography.

Transcriptome analysis of scRNAseq

Germ cells from 26 dpf WT or foxl2l10i/10i zebrafish in piwi1:EGFP transgene background were collected for scRNAseq. At this juvenile age, fish do not have apparent sexual characteristics. To isolate EGFP+ germ cells, the trunk of 20 fish was collected and dissociated in 0.4% collagenase (Worthington Biochemical, Cat. No. LS004188) and 0.5% trypsin (Life Technologies, Cat. No. 15400-054) in 1 ml PBS at 28°C for 60 min. After dissociation, cells were supplemented with fetal bovine serum (FBS) (Life Technologies, Cat. No. 10437028), centrifuged at 600×g, 4 min, and the pellet was suspended in FACS Pre-Sort Buffer (BD Biosciences, Cat. No. 563503) before being filtered through cell strainers. 1 mg/1 μl of propidium iodide (Sigma-Aldrich, Cat. No. P4170) was added to the eluate to stain dead cells, and around a 75% survival rate was obtained in both genotypes. EGFP+ cells were collected into 60% Leibovitz’s L-15 medium (Thermo Fisher Scientific, Cat. No. 11415064) by FACSAria III Cell Sorter (BD Biosciences) and supplemented with 10% FBS. Total 4781 cells for foxl2l+/+ and 8673 cells for foxl2l10i/10i samples were obtained upon counting by a cell sorter. After centrifugation at 300×g for 5 min and resuspension in PBS with 0.04% bovine serum albumin (Sigma-Aldrich), the recovered cell number was 3231 in foxl2l+/+ and 4019 in foxl2l10i/10i. Finally, the single-cell cDNA libraries were constructed by Chromium Next GEM Single Cell 3’ Library Construction Kit v3.1 kit (10x Genomics) following the manufacturer’s protocols with 12 cycles for cDNA amplification and 14 cycles for library amplification. Sequencing was performed by NextSeq 500 (Illumina) with 150 cycles of 28×91 paired-end reads. The raw base call (BCL) files were transferred to FASTQ files by the cellranger mkfastq pipeline from Cell Ranger (10x Genomics). The reads were mapped to the zebrafish genome GRCz11, and gene expression was obtained by cellranger count from Cell Ranger (10x Genomics).

Statistical analysis

All the quantitative data is presented as mean ± SD. The data was analyzed by Student’s t-test using GraphPad Prism 8.

Single-cell transcriptome analysis

Count matrix generation

Raw count matrices were generated by Cell Ranger v3.1.0 (Zheng et al., 2017) using the zebrafish reference genome GRCz11 and reference transcriptome ensemble release v101. The count matrices were then stored as Seurat objects v3.2.3 (Stuart et al., 2019).

Quality control

A two-step quality control approach was conducted individually for each sample. First, we conducted quality control at the cell level. Cells with UMI counts <10,000 or mitochondrial gene expression percentages ≥10% were marked as low-quality cells. Multiplet detection was performed on the remaining cells using the R package scds (Bais and Kostka, 2020). Cells with a hybrid score ≥ the outer fence were marked as multiplets. Second, we performed quality control at the cluster level to screen out the potential low-quality cells that passed the hard thresholds at the cell level. A low-resolution clustering was performed using Seurat v3.2.3. Clusters containing more than 80% low-quality cells or multiplets were marked as low-quality clusters. Together, cells that were marked as low quality, multiplets, and cells in low-quality clusters were excluded from the subsequent analysis.

Data preprocessing and cell-cycle scoring

The count matrices of the remaining 769 cells from the WT and 2399 cells from the mutant were merged and normalized using the R package scran (Lun et al., 2016). After normalization, gene expression data of cells from WT and mutant samples were organized into individual Seurat objects. The function CellCycleScoring in Seurat v3.2.3 was used for cell-cycle scoring (Figure 3—figure supplement 3). Briefly, all genes were first sorted by their mean expression and split into bins. For each S- and G2M-specific gene, relative expression was calculated by comparison with other genes in the same bin. The S-score and G2/M-score were calculated as the average of the relative expression of all genes in the lists of S- and G2M-specific genes.

Cell clustering analysis of WT cells

We used the FindVariableFeatures in Seurat v3.2.3 to identify the top 2500 highly variable genes in each sample. The principal components (PCs) of these highly variable genes on scaled expression matrices were derived after their cell-cycle scores were regressed out (Figure 3—figure supplement 4). We used the leading 30 PCs for UMAP embedding. The Leiden algorithm was applied to identify cell clusters. The optimal resolution, determined by the highest average silhouette score, was chosen from a range of feasible values.

Trajectory and pseudo-time analysis of WT cells

The data matrix, along with the cell clusters and UMAP embedding, was converted to the CDS format and subsequently analyzed using Monocle 3 (Cao et al., 2019b) to obtain the trajectory that encompasses all clusters. We then used the GSC marker nanos2 to identify the root and the order of the trajectory and to estimate the pseudo-time for each WT cell.

Cluster marker identification

We used the top_markers in Monocle 3 (Cao et al., 2019b) to identify the cluster markers whose high expression discriminates cells in the target cluster from the other cells with a high degree of specificity. For each cluster, genes that satisfy the following criteria: (i) specificity ≥0.2, (ii) q-value <0.01, and (iii) having nonzero counts in ≥80% cells in the cluster were selected as markers. Marker genes were ranked based on the specificity within WT clusters, and the top 2–4 markers in each WT cluster were selected as top markers. Total 37 top markers were shown in Figure 1D.

Inferring the developmental stages to mutant cells

The function TransferData in Seurat v3.2.3 was utilized to infer the developmental stage from the WT cells to the mutant cells. Initially, PCs were derived from the WT cells based on the expression of their cluster markers. Mutant cells were then projected to the same PCs. The PCs and the projected PCs were used to identify the mutual nearest neighbors across mutant cells and WT cells. These mutual nearest neighbors were considered anchors and scored based on the consistency across the neighborhood structure of each dataset. Lastly, the developmental stage of each mutant cell was determined using a weighted voting mechanism. This process considers the stages of WT cells in every anchor, with votes weighted according to the anchor score and the proximity between the mutant cell and its corresponding anchor mutant cell.

Co-clustering analysis of mutant cells and WT cells from GSC to Prog-C

To integrate mutant cells and WT cells from GSC to Prog-C, the function DataIntegration in Seurat v3.2.3 was applied. Briefly, both datasets were projected to a shared space by performing canonical correlation analysis. Mutual nearest neighbors were then identified as anchors, representing pairs of mutant and WT cells sharing a similar expression pattern. Anchors were scored based on the consistency across the neighborhood structure of each dataset. For each WT cell, a correction vector was created based on the proximity to every anchor and the anchor score. The WT expression matrix was corrected accordingly and integrated with that of mutant cells for the same clustering procedure as previously described. Mutant-specific marker genes were identified using the top_markers function with the following criteria: (i) specificity ≥0.25, (ii) Q-value <0.01, (iii) presence in ≥80% mutant cells for each integrated group, and (iv) not included in the WT cluster marker gene list.

Composition analysis

Because each stage has its own overall rate of prevalence in the entire cell population, we adjusted for the unevenness of stage prevalence in determining how prominent a stage is in an integrative group. Fisher’s exact test was used to find out which integrative groups were enriched with what stages of cells. Prominent stages for an integrative group are defined as those with significantly higher proportions of cells than their prevalence rates. Specifically, we calculated the proportions of developmental stages for both WT and mutant cells in each group and then conducted Fisher’s exact tests for each stage within each group. The resulting p-values were adjusted by the Benjamini-Hochberg method for controlling for false discovery rates, which were set at 0.05. The stages with significant enrichment are referred to as the prominent stages after prevalence rate calibration, or prominent stages for short.

Differential expression analysis

To identify DEGs during development, we compared gene expression levels between mutant and WT cells at the same prominent stage within each integrated group. Other cells not included in the comparison are shown in gray on UMAP (Figure 4A). The Wilcoxon rank sum test was performed using the function FindMarkers in Seurat v3.2.3. Genes that satisfy the following criteria: (i) |log2(fold-change)|≥log2(2.5) and (ii) q-value<0.05 were selected as DEGs. The DEGs in each stage were ordered by the fold change.

Reprocessing of the public scRNAseq dataset from Liu et al

The processed data were retrieved from the designated repository (GEO accession number: GSE191137). To visualize the expression of GSCs and progenitors of WT cells in 40 dpf, we selected a subset of cells with sufficient UMI counts (≥7000) from clusters 0 and 2 and lacked zp3.2 gene expression. After regressing out the cell-cycle scores, we used the leading 30 PCs for UMAP embedding.

Acknowledgements

We would like to thank the help of Flow Cytometry Core (Institute of Biomedical Sciences), Genomics Core and Bioinformatics Core (Institute of Molecular Biology Academia Sinica) in scRNAseq study, and Taiwan Zebrafish Core Facility (NSTC 112-2740-B-400-001), Imaging Core (Institute of Molecular Biology Academia Sinica) for confocal imaging. This work was funded by grants from National Health Research Institutes (NHRI-EX107-10506SI) and National Science and Technology Council (NSTC 113-2313-B-492-001-, MOST 108-2311-B-001-038-MY3) to BcC.

Funding Statement

The funders had no role in study design, data collection and interpretation, or the decision to submit the work for publication.

Contributor Information

Ker-Chau Li, Email: kcli@stat.ucla.edu.

Bon-chu Chung, Email: mbchung@sinica.edu.tw.

John Postlethwait, University of Oregon, United States.

Didier YR Stainier, Max Planck Institute for Heart and Lung Research, Germany.

Funding Information

This paper was supported by the following grants:

  • National Health Research Institutes NHRI- EX107-10506SI to Bon-chu Chung.

  • National Science and Technology Council MOST 108-2311-B-001-038-MY3 to Bon-chu Chung.

  • National Science and Technology Council NSTC 113-2313-B-492-001- to Bon-chu Chung.

Additional information

Competing interests

No competing interests declared.

Author contributions

Conceptualization, Data curation, Formal analysis, Validation, Investigation, Visualization, Methodology, Writing – original draft, Writing – review and editing.

Resources, Data curation, Software, Formal analysis, Validation, Investigation, Visualization, Methodology.

Data curation, Validation, Investigation, Methodology, Writing – review and editing.

Data curation, Formal analysis, Investigation, Methodology, Writing – review and editing.

Resources, Software, Formal analysis, Supervision, Investigation, Visualization, Methodology, Project administration, Writing – review and editing.

Conceptualization, Formal analysis, Supervision, Funding acquisition, Investigation, Writing – original draft, Writing – review and editing.

Ethics

Zebrafish were handled under the guidelines of Institutional Animal Care and Utilization Committee of Academia Sinica and National Center for Biomodels, National Applied Research Laboratories (IACUC: 10-10-084, and NLAC-113-M-004-R1).

Additional files

MDAR checklist
Supplementary file 1. The list and the expression of marker genes in each cluster.
elife-100204-supp1.xlsx (782KB, xlsx)

Data availability

Data for single-cell transcriptomic data analysis are available with NCBI GEO (https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE173718). Analysis code is available at Figshare (https://doi.org/10.6084/m9.figshare.26314126.v1).

The following datasets were generated:

Chung BC. 2025. Single cell transcriptomes of zebrafish germline reveal progenitor types and feminization by Foxl2l. NCBI Gene Expression Omnibus. GSE173718

Ho H, Hsu C, Yang CH, Wang Y, Li KC, Chung B. 2025. Codes of "Single-cell transcriptomes of zebrafish germline reveal stemness suppression and progenitor feminization by Foxl2l". figshare.

The following previously published dataset was used:

Liu Y. 2021. Single-cell transcriptome reveals insights into the development and function of the zebrafish ovary. NCBI Gene Expression Omnibus. GSE191137

References

  1. Bais AS, Kostka D. scds: computational annotation of doublets in single-cell RNA sequencing data. Bioinformatics. 2020;36:1150–1158. doi: 10.1093/bioinformatics/btz698. [DOI] [PMC free article] [PubMed] [Google Scholar]
  2. Beer RL, Draper BW. nanos3 maintains germline stem cells and expression of the conserved germline stem cell gene nanos2 in the zebrafish ovary. Developmental Biology. 2013;374:308–318. doi: 10.1016/j.ydbio.2012.12.003. [DOI] [PubMed] [Google Scholar]
  3. Benezra R, Davis RL, Lockshon D, Turner DL, Weintraub H. The protein Id: a negative regulator of helix-loop-helix DNA binding proteins. Cell. 1990;61:49–59. doi: 10.1016/0092-8674(90)90214-y. [DOI] [PubMed] [Google Scholar]
  4. Berta P, Hawkins JR, Sinclair AH, Taylor A, Griffiths BL, Goodfellow PN, Fellous M. Genetic evidence equating SRY and the testis-determining factor. Nature. 1990;348:448–450. doi: 10.1038/348448A0. [DOI] [PubMed] [Google Scholar]
  5. Bertho S, Clapp M, Banisch TU, Bandemer J, Raz E, Marlow FL. Zebrafish dazl regulates cystogenesis and germline stem cell specification during the primordial germ cell to germline stem cell transition. Development. 2021;148:dev187773. doi: 10.1242/dev.187773. [DOI] [PMC free article] [PubMed] [Google Scholar]
  6. Brend T, Holley SA. Zebrafish whole mount high-resolution double fluorescent in situ hybridization. Journal of Visualized Experiments. 2009;1:1229. doi: 10.3791/1229. [DOI] [PMC free article] [PubMed] [Google Scholar]
  7. Campbell PD, Chao JA, Singer RH, Marlow FL. Dynamic visualization of transcription and RNA subcellular localization in zebrafish. Development. 2015;142:1368–1374. doi: 10.1242/dev.118968. [DOI] [PMC free article] [PubMed] [Google Scholar]
  8. Cao Z, Mao X, Luo L. Germline stem cells drive ovary regeneration in zebrafish. Cell Reports. 2019a;26:1709–1717. doi: 10.1016/j.celrep.2019.01.061. [DOI] [PubMed] [Google Scholar]
  9. Cao J, Spielmann M, Qiu X, Huang X, Ibrahim DM, Hill AJ, Zhang F, Mundlos S, Christiansen L, Steemers FJ, Trapnell C, Shendure J. The single-cell transcriptional landscape of mammalian organogenesis. Nature. 2019b;566:496–502. doi: 10.1038/s41586-019-0969-x. [DOI] [PMC free article] [PubMed] [Google Scholar]
  10. Chen S, Zhang H, Wang F, Zhang W, Peng G. nr0b1 (DAX1) mutation in zebrafish causes female-to-male sex reversal through abnormal gonadal proliferation and differentiation. Molecular and Cellular Endocrinology. 2016;433:105–116. doi: 10.1016/j.mce.2016.06.005. [DOI] [PubMed] [Google Scholar]
  11. Dai X, Shu Y, Lou Q, Tian Q, Zhai G, Song J, Lu S, Yu H, He J, Yin Z. Tdrd12 Is essential for germ cell development and maintenance in zebrafish. International Journal of Molecular Sciences. 2017;18:1127. doi: 10.3390/ijms18061127. [DOI] [PMC free article] [PubMed] [Google Scholar]
  12. Dai S, Qi S, Wei X, Liu X, Li Y, Zhou X, Xiao H, Lu B, Wang D, Li M. Germline sexual fate is determined by the antagonistic action of dmrt1 and foxl3/foxl2 in tilapia. Development. 2021;148:dev199380. doi: 10.1242/dev.199380. [DOI] [PubMed] [Google Scholar]
  13. Dranow DB, Hu K, Bird AM, Lawry ST, Adams MT, Sanchez A, Amatruda JF, Draper BW. Bmp15 Is an oocyte-produced signal required for maintenance of the adult female sexual phenotype in zebrafish. PLOS Genetics. 2016;12:e1006323. doi: 10.1371/journal.pgen.1006323. [DOI] [PMC free article] [PubMed] [Google Scholar]
  14. Draper BW, McCallum CM, Moens CB. nanos1 is required to maintain oocyte production in adult zebrafish. Developmental Biology. 2007;305:589–598. doi: 10.1016/j.ydbio.2007.03.007. [DOI] [PMC free article] [PubMed] [Google Scholar]
  15. Gautier A, Goupil AS, Le Gac F, Lareyre JJ. A promoter fragment of the sycp1 gene is sufficient to drive transgene expression in male and female meiotic germ cells in zebrafish. Biology of Reproduction. 2013;89:89. doi: 10.1095/biolreprod.113.107706. [DOI] [PubMed] [Google Scholar]
  16. Ge C, Ye J, Zhang H, Zhang Y, Sun W, Sang Y, Capel B, Qian G. Dmrt1 induces the male pathway in a turtle species with temperature-dependent sex determination. Development. 2017;144:2222–2233. doi: 10.1242/dev.152033. [DOI] [PubMed] [Google Scholar]
  17. Georges A, Auguste A, Bessière L, Vanet A, Todeschini AL, Veitia RA. FOXL2: a central transcription factor of the ovary. Journal of Molecular Endocrinology. 2014;52:R17–R33. doi: 10.1530/JME-13-0159. [DOI] [PubMed] [Google Scholar]
  18. Hartung O, Forbes MM, Marlow FL. Zebrafish vasa is required for germ-cell differentiation and maintenance. Molecular Reproduction and Development. 2014;81:946–961. doi: 10.1002/mrd.22414. [DOI] [PMC free article] [PubMed] [Google Scholar]
  19. Hong SH, Lee JH, Lee JB, Ji J, Bhatia M. ID1 and ID3 represent conserved negative regulators of human embryonic and induced pluripotent stem cell hematopoiesis. Journal of Cell Science. 2011;124:1445–1452. doi: 10.1242/jcs.077511. [DOI] [PubMed] [Google Scholar]
  20. Kantzer CG, Yang W, Grommisch D, Vikhe Patil K, Mak KHM, Shirokova V, Genander M. ID1 and CEBPA coordinate epidermal progenitor cell differentiation. Development. 2022;149:22. doi: 10.1242/dev.201262. [DOI] [PMC free article] [PubMed] [Google Scholar]
  21. Kashimada K, Pelosi E, Chen H, Schlessinger D, Wilhelm D, Koopman P. FOXL2 and BMP2 act cooperatively to regulate follistatin gene expression during ovarian development. Endocrinology. 2011;152:272–280. doi: 10.1210/en.2010-0636. [DOI] [PMC free article] [PubMed] [Google Scholar]
  22. Kaufman OH, Lee K, Martin M, Rothhämel S, Marlow FL. rbpms2 functions in Balbiani body architecture and ovary fate. PLOS Genetics. 2018;14:e1007489. doi: 10.1371/journal.pgen.1007489. [DOI] [PMC free article] [PubMed] [Google Scholar]
  23. Kikuchi M, Nishimura T, Ishishita S, Matsuda Y, Tanaka M. foxl3, a sexual switch in germ cells, initiates two independent molecular pathways for commitment to oogenesis in medaka. PNAS. 2020;117:12174–12181. doi: 10.1073/pnas.1918556117. [DOI] [PMC free article] [PubMed] [Google Scholar]
  24. Kuo FT, Bentsi-Barnes IK, Barlow GM, Pisarska MD. Mutant Forkhead L2 (FOXL2) proteins associated with premature ovarian failure (POF) dimerize with wild-type FOXL2, leading to altered regulation of genes associated with granulosa cell differentiation. Endocrinology. 2011;152:3917–3929. doi: 10.1210/en.2010-0989. [DOI] [PMC free article] [PubMed] [Google Scholar]
  25. Kuo FT, Fan K, Bentsi-Barnes I, Barlow GM, Pisarska MD. Mouse forkhead L2 maintains repression of FSH-dependent genes in the granulosa cell. Reproduction. 2012;144:485–494. doi: 10.1530/REP-11-0259. [DOI] [PubMed] [Google Scholar]
  26. Kusz KM, Tomczyk L, Sajek M, Spik A, Latos-Bielenska A, Jedrzejczak P, Pawelczyk L, Jaruzelska J. The highly conserved NANOS2 protein: testis-specific expression and significance for the human male reproduction. Molecular Human Reproduction. 2009;15:165–171. doi: 10.1093/molehr/gap003. [DOI] [PubMed] [Google Scholar]
  27. Leu DH, Draper BW. The ziwi promoter drives germline-specific gene expression in zebrafish. Developmental Dynamics. 2010;239:2714–2721. doi: 10.1002/dvdy.22404. [DOI] [PubMed] [Google Scholar]
  28. Lin T, Chao C, Saito S, Mazur SJ, Murphy ME, Appella E, Xu Y. p53 induces differentiation of mouse embryonic stem cells by suppressing Nanog expression. Nature Cell Biology. 2005;7:165–171. doi: 10.1038/ncb1211. [DOI] [PubMed] [Google Scholar]
  29. Liu Y, Kossack ME, McFaul ME, Christensen LN, Siebert S, Wyatt SR, Kamei CN, Horst S, Arroyo N, Drummond IA, Juliano CE, Draper BW. Single-cell transcriptome reveals insights into the development and function of the zebrafish ovary. eLife. 2022;11:e76014. doi: 10.7554/eLife.76014. [DOI] [PMC free article] [PubMed] [Google Scholar]
  30. Lun ATL, McCarthy DJ, Marioni JC. A step-by-step workflow for low-level analysis of single-cell RNA-seq data with Bioconductor. F1000Research. 2016;5:2122. doi: 10.12688/f1000research.9501.2. [DOI] [PMC free article] [PubMed] [Google Scholar]
  31. Luzio A, Santos D, Monteiro SM, Coimbra AM. Zebrafish male differentiation: Do all testes go through a “juvenile ovary” stage? Tissue & Cell. 2021;72:101545. doi: 10.1016/j.tice.2021.101545. [DOI] [PubMed] [Google Scholar]
  32. Marlow FL, Mullins MC. Bucky ball functions in Balbiani body assembly and animal-vegetal polarity in the oocyte and follicle cell layer in zebrafish. Developmental Biology. 2008;321:40–50. doi: 10.1016/j.ydbio.2008.05.557. [DOI] [PMC free article] [PubMed] [Google Scholar]
  33. Matson CK, Murphy MW, Sarver AL, Griswold MD, Bardwell VJ, Zarkower D. DMRT1 prevents female reprogramming in the postnatal mammalian testis. Nature. 2011;476:101–104. doi: 10.1038/nature10239. [DOI] [PMC free article] [PubMed] [Google Scholar]
  34. Matsuda M, Nagahama Y, Shinomiya A, Sato T, Matsuda C, Kobayashi T, Morrey CE, Shibata N, Asakawa S, Shimizu N, Hori H, Hamaguchi S, Sakaizumi M. DMY is a Y-specific DM-domain gene required for male development in the medaka fish. Nature. 2002;417:559–563. doi: 10.1038/nature751. [DOI] [PubMed] [Google Scholar]
  35. McConnell AM, Yao C, Yeckes AR, Wang Y, Selvaggio AS, Tang J, Kirsch DG, Stripp BR. p53 regulates progenitor cell quiescence and differentiation in the airway. Cell Reports. 2016;17:2173–2182. doi: 10.1016/j.celrep.2016.11.007. [DOI] [PubMed] [Google Scholar]
  36. Miao L, Yuan Y, Cheng F, Fang J, Zhou F, Ma W, Jiang Y, Huang X, Wang Y, Shan L, Chen D, Zhang J. Translation repression by maternal RNA binding protein Zar1 is essential for early oogenesis in zebrafish. Development. 2017;144:128–138. doi: 10.1242/dev.144642. [DOI] [PubMed] [Google Scholar]
  37. Molchadsky A, Ezra O, Amendola PG, Krantz D, Kogan-Sakin I, Buganim Y, Rivlin N, Goldfinger N, Folgiero V, Falcioni R, Sarig R, Rotter V. p53 is required for brown adipogenic differentiation and has a protective role against diet-induced obesity. Cell Death and Differentiation. 2013;20:774–783. doi: 10.1038/cdd.2013.9. [DOI] [PMC free article] [PubMed] [Google Scholar]
  38. Nishimura T, Sato T, Yamamoto Y, Watakabe I, Ohkawa Y, Suyama M, Kobayashi S, Tanaka M. Sex determination. foxl3 is a germ cell-intrinsic factor involved in sperm-egg fate decision in medaka. Science. 2015;349:328–331. doi: 10.1126/science.aaa2657. [DOI] [PubMed] [Google Scholar]
  39. Nüsslein-Volhard C, Ralf D. Zebrafish. New York: Oxford University Press; 2002. [DOI] [Google Scholar]
  40. Pan YJ, Tong SK, Hsu CW, Weng JH, Chung BC. Zebrafish establish female germ cell identity by advancing cell proliferation and meiosis. Frontiers in Cell and Developmental Biology. 2022;10:866267. doi: 10.3389/fcell.2022.866267. [DOI] [PMC free article] [PubMed] [Google Scholar]
  41. Pisarska MD, Bae J, Klein C, Hsueh AJW. Forkhead l2 is expressed in the ovary and represses the promoter activity of the steroidogenic acute regulatory gene. Endocrinology. 2004;145:3424–3433. doi: 10.1210/en.2003-1141. [DOI] [PubMed] [Google Scholar]
  42. Pisarska MD, Kuo FT, Bentsi-Barnes IK, Khan S, Barlow GM. LATS1 phosphorylates forkhead L2 and regulates its transcriptional activity. American Journal of Physiology. Endocrinology and Metabolism. 2010;299:E101–E109. doi: 10.1152/ajpendo.00534.2009. [DOI] [PMC free article] [PubMed] [Google Scholar]
  43. Qin M, Zhang Z, Song W, Wong QWL, Chen W, Shirgaonkar N, Ge W. Roles of Figla/figla in juvenile ovary development and follicle formation during zebrafish gonadogenesis. Endocrinology. 2018;159:3699–3722. doi: 10.1210/en.2018-00648. [DOI] [PubMed] [Google Scholar]
  44. Rodríguez-Marí A, Cañestro C, Bremiller RA, Nguyen-Johnson A, Asakawa K, Kawakami K, Postlethwait JH. Sex reversal in zebrafish fancl mutants is caused by Tp53-mediated germ cell apoptosis. PLOS Genetics. 2010;6:e1001034. doi: 10.1371/journal.pgen.1001034. [DOI] [PMC free article] [PubMed] [Google Scholar]
  45. Rodríguez-Marí A, Wilson C, Titus TA, Cañestro C, BreMiller RA, Yan Y-L, Nanda I, Johnston A, Kanki JP, Gray EM, He X, Spitsbergen J, Schindler D, Postlethwait JH. Roles of brca2 (fancd1) in oocyte nuclear architecture, gametogenesis, gonad tumors, and genome stability in zebrafish. PLOS Genetics. 2011;7:e1001357. doi: 10.1371/journal.pgen.1001357. [DOI] [PMC free article] [PubMed] [Google Scholar]
  46. Sablitzky F, Moore A, Bromley M, Deed RW, Newton JS, Norton JD. Stage- and subcellular-specific expression of Id proteins in male germ and Sertoli cells implicates distinctive regulatory roles for Id proteins during meiosis, spermatogenesis, and Sertoli cell function. Cell Growth & Differentiation. 1998;9:1015–1024. [PubMed] [Google Scholar]
  47. Saito D, Morinaga C, Aoki Y, Nakamura S, Mitani H, Furutani-Seiki M, Kondoh H, Tanaka M. Proliferation of germ cells during gonadal sex differentiation in medaka: Insights from germ cell-depleted mutant zenzai. Developmental Biology. 2007;310:280–290. doi: 10.1016/j.ydbio.2007.07.039. [DOI] [PubMed] [Google Scholar]
  48. She ZY, Yang WX. Sry and SoxE genes: How they participate in mammalian sex determination and gonadal development? Seminars in Cell & Developmental Biology. 2017;63:13–22. doi: 10.1016/j.semcdb.2016.07.032. [DOI] [PubMed] [Google Scholar]
  49. Shive HR, West RR, Embree LJ, Azuma M, Sood R, Liu P, Hickstein DD. brca2 in zebrafish ovarian development, spermatogenesis, and tumorigenesis. PNAS. 2010;107:19350–19355. doi: 10.1073/pnas.1011630107. [DOI] [PMC free article] [PubMed] [Google Scholar]
  50. Slanchev K, Stebler J, de la Cueva-Méndez G, Raz E. Development without germ cells: the role of the germ line in zebrafish sex differentiation. PNAS. 2005;102:4074–4079. doi: 10.1073/pnas.0407475102. [DOI] [PMC free article] [PubMed] [Google Scholar]
  51. Smith CA, Roeszler KN, Ohnesorg T, Cummins DM, Farlie PG, Doran TJ, Sinclair AH. The avian Z-linked gene DMRT1 is required for male sex determination in the chicken. Nature. 2009;461:267–271. doi: 10.1038/nature08298. [DOI] [PubMed] [Google Scholar]
  52. Stuart T, Butler A, Hoffman P, Hafemeister C, Papalexi E, Mauck WM, III, Hao Y, Stoeckius M, Smibert P, Satija R. Comprehensive integration of single-cell data. Cell. 2019;177:1888–1902. doi: 10.1016/j.cell.2019.05.031. [DOI] [PMC free article] [PubMed] [Google Scholar]
  53. Suzuki A, Saga Y. Nanos2 suppresses meiosis and promotes male germ cell differentiation. Genes & Development. 2008;22:430–435. doi: 10.1101/gad.1612708. [DOI] [PMC free article] [PubMed] [Google Scholar]
  54. Thisse C, Thisse B. High-resolution in situ hybridization to whole-mount zebrafish embryos. Nature Protocols. 2008;3:59–69. doi: 10.1038/nprot.2007.514. [DOI] [PubMed] [Google Scholar]
  55. Tong S-K, Hsu H-J, Chung B. Zebrafish monosex population reveals female dominance in sex determination and earliest events of gonad differentiation. Developmental Biology. 2010;344:849–856. doi: 10.1016/j.ydbio.2010.05.515. [DOI] [PubMed] [Google Scholar]
  56. Tsuda M, Sasaoka Y, Kiso M, Abe K, Haraguchi S, Kobayashi S, Saga Y. Conserved role of nanos proteins in germ cell development. Science. 2003;301:1239–1241. doi: 10.1126/science.1085222. [DOI] [PubMed] [Google Scholar]
  57. Uchida D, Yamashita M, Kitano T, Iguchi T. Oocyte apoptosis during the transition from ovary-like tissue to testes during sex differentiation of juvenile zebrafish. The Journal of Experimental Biology. 2002;205:711–718. doi: 10.1242/jeb.205.6.711. [DOI] [PubMed] [Google Scholar]
  58. Wang M, Liu X, Chang G, Chen Y, An G, Yan L, Gao S, Xu Y, Cui Y, Dong J, Chen Y, Fan X, Hu Y, Song K, Zhu X, Gao Y, Yao Z, Bian S, Hou Y, Lu J, Wang R, Fan Y, Lian Y, Tang W, Wang Y, Liu J, Zhao L, Wang L, Liu Z, Yuan R, Shi Y, Hu B, Ren X, Tang F, Zhao XY, Qiao J. Single-Cell RNA Sequencing analysis reveals sequential cell fate transition during human spermatogenesis. Cell Stem Cell. 2018;23:599–614. doi: 10.1016/j.stem.2018.08.007. [DOI] [PubMed] [Google Scholar]
  59. Wassarman PM, Litscher ES. Zona pellucida genes and proteins: Essential players in mammalian oogenesis and fertility. Genes. 2021;12:1266. doi: 10.3390/genes12081266. [DOI] [PMC free article] [PubMed] [Google Scholar]
  60. Webster KA, Schach U, Ordaz A, Steinfeld JS, Draper BW, Siegfried KR. Dmrt1 is necessary for male sexual development in zebrafish. Developmental Biology. 2017;422:33–46. doi: 10.1016/j.ydbio.2016.12.008. [DOI] [PMC free article] [PubMed] [Google Scholar]
  61. Westerfield M. The Zebrafish Book. A Guide for the Laboratory Use of Zebrafish (Danio rerio) Eugene: University of Oregon Press; 2000. [Google Scholar]
  62. Wilson CA, High SK, McCluskey BM, Amores A, Yan Y, Titus TA, Anderson JL, Batzel P, Carvan MJ, Schartl M, Postlethwait JH. Wild sex in zebrafish: loss of the natural sex determinant in domesticated strains. Genetics. 2014;198:1291–1308. doi: 10.1534/genetics.114.169284. [DOI] [PMC free article] [PubMed] [Google Scholar]
  63. Wilson CA, Batzel P, Postlethwait JH. Direct male development in chromosomally ZZ zebrafish. Frontiers in Cell and Developmental Biology. 2024;12:1362228. doi: 10.3389/fcell.2024.1362228. [DOI] [PMC free article] [PubMed] [Google Scholar]
  64. Yang CH, Wang YW, Hsu CW, Chung BC. Zebrafish Foxl2l functions in proliferating germ cells for female meiotic entry. Developmental Biology. 2025;517:91–99. doi: 10.1016/j.ydbio.2024.09.013. [DOI] [PubMed] [Google Scholar]
  65. Ye D, Zhu L, Zhang Q, Xiong F, Wang H, Wang X, He M, Zhu Z, Sun Y. Abundance of early embryonic primordial germ cells promotes zebrafish female differentiation as revealed by lifetime labeling of germline. Marine Biotechnology. 2019;21:217–228. doi: 10.1007/s10126-019-09874-1. [DOI] [PMC free article] [PubMed] [Google Scholar]
  66. Ying QL, Nichols J, Chambers I, Smith A. BMP induction of Id proteins suppresses differentiation and sustains embryonic stem cell self-renewal in collaboration with STAT3. Cell. 2003;115:281–292. doi: 10.1016/s0092-8674(03)00847-x. [DOI] [PubMed] [Google Scholar]
  67. Yoshida K, Kondoh G, Matsuda Y, Habu T, Nishimune Y, Morita T. The mouse RecA-like gene Dmc1 is required for homologous chromosome synapsis during meiosis. Molecular Cell. 1998;1:707–718. doi: 10.1016/s1097-2765(00)80070-2. [DOI] [PubMed] [Google Scholar]
  68. Zhang N, Yantiss RK, Nam HS, Chin Y, Zhou XK, Scherl EJ, Bosworth BP, Subbaramaiah K, Dannenberg AJ, Benezra R. ID1 is a functional marker for intestinal stem and progenitor cells required for normal response to injury. Stem Cell Reports. 2014;3:716–724. doi: 10.1016/j.stemcr.2014.09.012. [DOI] [PMC free article] [PubMed] [Google Scholar]
  69. Zheng GXY, Terry JM, Belgrader P, Ryvkin P, Bent ZW, Wilson R, Ziraldo SB, Wheeler TD, McDermott GP, Zhu J, Gregory MT, Shuga J, Montesclaros L, Underwood JG, Masquelier DA, Nishimura SY, Schnall-Levin M, Wyatt PW, Hindson CM, Bharadwaj R, Wong A, Ness KD, Beppu LW, Deeg HJ, McFarland C, Loeb KR, Valente WJ, Ericson NG, Stevens EA, Radich JP, Mikkelsen TS, Hindson BJ, Bielas JH. Massively parallel digital transcriptional profiling of single cells. Nature Communications. 2017;8:14049. doi: 10.1038/ncomms14049. [DOI] [PMC free article] [PubMed] [Google Scholar]
  70. Zhou L, Feng Y, Wang F, Dong X, Jiang L, Liu C, Zhao Q, Li K. Generation of all-male-like sterile zebrafish by eliminating primordial germ cells at early development. Scientific Reports. 2018;8:1834. doi: 10.1038/s41598-018-20039-3. [DOI] [PMC free article] [PubMed] [Google Scholar]

Editor's evaluation

John Postlethwait 1

Findings have important significance for understanding the role of the Foxl2-family of proteins in the early stages of germ cell maturation. Evidence to support the conclusions are solid and use state-of-the-field single cell transcriptomics. Results will interest biologists investigating mechanisms of reproduction, causes of infertility, and the mechanisms by which gonads embark on becoming a testis or an ovary.

Decision letter

Editor: John Postlethwait1
Reviewed by: Yong Zhu2

Our editorial process produces two outputs: i) public reviews designed to be posted alongside the preprint for the benefit of readers; ii) feedback on the manuscript for the authors, including requests for revisions, shown below. We also include an acceptance summary that explains what the editors found interesting or important about the work.

Decision letter after peer review:

Thank you for submitting your article "Single cell transcriptomes of zebrafish germline reveal stemness suppression and progenitor feminization by Foxl2l" for consideration by eLife. Your article has been reviewed by 3 peer reviewers, and the evaluation has been overseen by a Reviewing Editor and Didier Stainier as the Senior Editor.

The reviewers have discussed their reviews with one another, and the Reviewing Editor has drafted this to help you prepare a revised submission.

Essential Revisions:

1) Evidence that committed progenitors revert to germline stem cells (GSCs) is weak. Functional data are required to draw this conclusion. The phrase 'reveal stemness suppression' should be removed from the title and alternative explanations should appear in the text. The claim that suppressing stemness is necessary for female differentiation is not well supported, for example, cells might initiate the committed progenitor program but fail to shut off and continue to express part of the GSC program.

2) Essential literature that places the work in context must be discussed, with several cases noted in each specific review.

3) Alternative explanations to the statement that committed progenitor stage is "the gate toward female determination" must be discussed, for example, that the late progenitor stage is the key stage for female determination.

4) Consider alternatives to the conclusion that the progenitors are differentiated because it is based solely on the expression of foxl2l, which is initially expressed in the juvenile ovary state that lab strains develop through.

5). Consider alternative explanations for the differences in the action of Foxl2(Fox3) in zebrafish and medaka.

6) Methods require more details on how fish were selected for scRNA-seq.

Reviewer #1 (Recommendations for the authors):

Specific points.

1) Page 7 lines 258-261. The authors conclude that the committed progenitors revert to GSCs based on the coexpression of nanos2 and foxl2l nanos2 and later on the same page line 262, based on expression of id1 in mutants but not WT. Without functional data demonstrating that the progenitors revert to an early state, alternative interpretations should be considered. For example, can the authors exclude the alternative possibility that the cells initiate the committed progenitor program but do not shut off and continue to express the GSC program such that the coexpression of both programs blocks differentiation? It would help to discuss these findings in the context of Fox family functions. Some Fox family members, Foxl2 and FoxPs for example, are known to be both activators and repressors of transcription or act primarily as repressors. Relevant to this work, repressive activity of Foxl2 has previously been reported in the mammalian ovary (Pisarska et al. Endocrinology 2004, Pisarska Am J. Phys Endo. Metabolism 2010, Kuo Reproduction 2012, Kuo Endocrinology 2011, as well as several more recent publications). In this context interfering with Foxl2 repression was proposed to accelerate follicle recruitment and cause premature ovarian failure.

2) Discussion Page 9 lines 289-293. This section is not entirely clear. First GSC-I and GSC-II subtypes are introduced, then it is stated that "all GSCs self -renew" and a Saito 2007 paper is cited. This is followed by a description of "two types of daughter cells derived from GSC self-renewal". Does the conclusion that "all GSCs renew" still hold in light of the two GSC populations? Would a division that produces a daughter that cannot renew be considered "self-renewal" or instead a non-renewing division?

3) Discussion Page 9 Line 295-7. The authors conclude that the committed progenitor stage is "the gate toward female determination" and that they "stay at S-Phase temporarily before differentiation". Is this conclusion based solely on RNA expression? Also, in several species, including zebrafish, meiotic entry has been correlated with ovary development. Can the possibility that the late progenitor stage, the stage when meiotic genes are detected and a stage missing in foxl2l mutants, is the key stage for female determination be excluded?

4) The same comment as above pertains to Page 9 lines 281-284 of the discussion.

5) Page 3 Line 64 states that "the developmental stages of germ cells in zebrafish remain unclear". This is somewhat misleading as recently published single cell datasets have analyzed and defined the stages of germ cell development, including during and after sexual differentiation. This should be clarified, and the relevant published work should be cited. (Liu et al. eLife 2022, Wilson et al. Front Cell Dev Bio 2024.)

6) Page 3 Line 70 might be confusing because it states that "XX medaka fish with foxl3 mutation become hermaphroditic processing functional sperms". Please clarify whether they also produce oocytes or are genetically female "XX" but make sperm.

7) On Page 9 lines 305-307, the authors discuss prior working showing that loss of germ cells leads to male development and that germ cells are required for female development. Then lines 307-309 state that this work shows some progenitors are already sexually differentiated. First, the stages compared are completely different. The early work looks at the primordial germ cells and their loss in the first few days of development before a gonad forms and this work looks after a gonad has formed and when sex is being determined. The second concern is that the conclusion that the cells are differentiated is based on the expression of foxl2l, which is expressed in the juvenile ovary state that lab strains have been shown to develop through (Wilson et al. Front Cell Dev Bio 2024). While it is fair to state that some cells express ovary markers, are they really differentiated? For example, in the context of the foxl2l mutant, GSCs and early progenitors inappropriately express foxl2l, but the mutants develop as males. Please clarify.

8) Page 9 lines 315-318 compares medaka and zebrafish foxl2l mutants and seems to suggest that Foxl2l is required for meiosis in medaka but has a different role in zebrafish. Is it possible that foxl2l plays a similar role in repressing the earlier developmental programs of GSCs and early progenitors in both cases? Is it possible that continued expression of these early programs interferes with activation of meiotic genes? Might this account for the absence of the late progenitor stage in foxl2l mutants since it is defined by expression of foxl2l and meiotic genes?

9) Page 10 lines 328-330. Direct differentiation of testis without going through a female phase is discussed. Recent work showing direct differentiation in wild strains should also be discussed and cited (Wilson et al. Front Cell Dev Bio 2024).

10) Page 10 lines 359-360. The authors state "Foxl2l may ensure female differentiation by preventing stemness and antagonizing male development." It is unclear why suppressing stemness would be necessary for female differentiation since female zebrafish have stem cells as do male zebrafish. It seems likely that turning off the GSC and early differentiation programs is important for allowing expression of meiosis and oocyte differentiation genes.

11) Page 11 370-374. "Thus, p53 may assist alternative differentiation of mutant germ cells". What is the evidence that p53 is involved in differentiation of germ cells? Sex bias has not been associated with the published p53 mutants. Is it possible that p53 eliminates germ cells that are simultaneously expressing programs of multiple stages? While p53 has been shown to be important for ovary to testis transformation in mutant contexts in adults, it is dispensable in mutants that disrupt ovary development at earlier stages. Please discuss in the context of, and cite, the relevant ovary literature Rodriguez-Mari et al. PLoS Gen 2010, Shive PNAS 2010, Hartung et al. Mol. Reprod. Dev 2014, Miao Development 2017, Kaufman et al. PLoSGen 2018, Bertho et al. Development 2021.

12) Figure 2D. It would be helpful to show the individual fish points or to present the data as a violin plot.

13) Figure 2F. The numbers are low here only 3-4 gonads are examined for each genotype and marker and in the mutants only 1/4 expresses the "testis" markers.

14) Figure 6 model. What is the evidence that prog C becomes GSCs again/revert versus continuing to express earlier programs?

Reviewer #2 (Recommendations for the authors):

The evidence that foxl2l cells are arrested at the committed progenitor stage (a possible novel phenotype) is based on the absence of expression of rec8a and Sycp3. Another explanation is that these gonads have already differentiated as testes. Is rec8a also expressed in male germ cells or is it female germ cell specific? Germ cells in testes are usually delayed in entering meiosis (and thus expressing Sycp3) relative to females so it is not clear if this represents the expected trajectory of testis vs. ovary or a novel developmental state caused by absence of foxl2l. Was expression of any male-specific genes identified in the mutant data set (e.g. tekt1 or odf3b) which could indicate these germ cells were already committed to spermatogenesis? Also, by "arrested" do the authors mean that they are in a cell cycle arrest? If so, this can be assayed with BrdU or EdU labeling.

L99: incomplete sentence.

L128: The publication associated with this dataset should be referenced (it's referenced in the figure legend but should also be referenced in main text).

L138: Using a negative result (absence of staining) to argue for a particular cell stage (prog-E) is problematic. To bolster this argument the authors should consider pointing out that this result is consistent with the scRNA-seq data that predicts little to no overlap between nanos2 and foxl2l expressing cells. Perhaps adding the nanos2 and foxl2l and rec8a UMAPs to Figure 1 would help.

L148 and Figure 2A. Wilson et al. (2024) showed that the gonads of Nadia ZZ fish appear to directly develop into testis without producing oocytes during the bipotential stage. The data in Figure 2A showing foxl2l expression in the ZZ gonads is therefore unexpected. Can the authors comment on this?

L154: "…barley observed in testis." Should this be "…not observed…", or is there low but detectable expression of foxl2l in the testis (not clear from Figure 1D)?

L170: It is not clear from the Methods how the fish were selected for scRNA-seq. With the piwil1:egfp line it is possible to distinguish wild-type males from females at the 27 dpf timepoint used in this analysis based on the size of the gonad (large=female, small=male). Were only fish with large gonads used for WT and foxl2l mutant scRNA-seq? If fish were chosen randomly, it would be expected that germ cells from testes would also be included. When do foxl2l mutant gonad first start looking like testes? Is there a delay in testis development relative to wild-type?

Figure 2B: Including age-matched wild-type testes in Figure 2B may help clarify if the foxl2l gonads are different or similar to a normal male gonad at this stage of development.

Figure 2D: I would suggest replacing si:ch211-191j22.7 with meiosin, since this is the clear ortholog.

Figure 3B: It is hard to see the morphology of the nuclei in these images. Grey scale is perfect for the DAPI channel, but it would be helpful to increase the apparent signal intensity using an image processing program like Photoshop so that it is possible to better assess the morphology of nuclei.

Figure 3C: I am having a hard time understanding the meaning of Figure 3C. For the WT cells, the nanos2+ cells are restricted to the W1 subcluster (Figure 1D), indicating that the GSCs are homogeneous with respect to gene expression. However, when WT cells are clustered with mutant cells the WT GSCs now partition to two different subclusters (I1 and I2). When the "control" cells are not clustering tightly, it calls into question how much this clustering represents reality. How do the authors rationalize this?

L194: Figure S7B is unfortunately not helping me understand how the data was processed. What is meant by "transfer developmental stage?" Does this require/assume that the mutant cells be a specific stage that can be mapped directly onto wildtype stages? I wonder if forcing the mutant cells to cluster with wildtype cells is somehow skewing the results. How different is the clustering if only mutant cells are used?

Figure 3E: nanos2 and cbx7a are listed twice. It is likely they should be psmb13a and dnmt3bb.1, respectively. Was expression of any male-specific gene identified in the mutant data set (e.g. tekt1 or odf3b) which could indicate these germ cells were already committed to spermatogenesis?

Sup Figure S8A: It would be helpful if this figure also included the UMAP shown in Figure 3C for reference (with I1-I7 labels). S8A is hard to interpret given the altered gene expression pattern of mutant cells. Some genes fit the proposed cell stages (e.g. dnmt3bb.1), but others are not as clear (e.g. nanos2, dhrs1).

L210: "This suggests that GSCs can be further divided into two finer stages, GSCI and GSCII." If you only see this subdivision in mutants, then it is hard to argue that these are representative stages present in WT.

Line 307: It is not clear what this statement means. It appears to imply that foxl2l-expressing cells are sexually differentiated. However, in Figure 2A it is shown that foxl2l is expressed in germ cells of ZZ fish during the bipotential stage, yet these germ cell presumable would have differentiated as sperm, not oocytes.

L302-322: The authors need to spell out the evidence that supports these statements. It could be argued that the difference is not in the role of foxl2l between Medaka and zebrafish, but the fact that medaka has a strong genetic sex determination mechanism while zebrafish does not. Thus, the apparent difference in phenotype is because medaka cannot sex reverse as easily as zebrafish.

L331: For completeness, please add a reference to Wilson et al., (2024; PMID: 38529407)

L356: The authors argue that one role of Foxl2l is to suppress Dmrt1 expression in progenitor cells, yet in the foxl2l mutant germ cells, Dmrt1 is upregulated in all mutant cell stages, from GSC to prog-C cells (i.e. even in cells that do not normally express foxl2l expression). It is possible that the upregulation in all cells is because the gonads from which these cells were isolated have already begun to differentiate as a testis and the germ cells are now more male-like than female-like. This would also suggest the possibility that male and female GSC's do not express equivalent amounts of Dmrt1.

Reviewer #3 (Recommendations for the authors):

The major conclusion, i.e., title and short title may be misleading and may cause confusion. Stemness is the ability of a cell to perpetuate/continue/preserve its lineage, to give rise to differentiated cells. With foxl2l, progenitor cells committed cell (Prog-C) will continue to develop into late progenitor (Prog-L). In addition, there is no strong evidence that Prog-C was able to dedifferentiate back into GSC in foxl2l mutants. Prog-C could be simply lost in foxl2l mutant, and increased nanos2 cells could be due to increased proliferation of GSC. Therefore, "Foxl2l suppresses stemness" could be misleading. Is there size and morphological difference between prog-C and prog-L?

Lines 23-24, 64-65 and in other parts of the text "germ cell types … remain elusive" "developmental stages of germ cells in zebrafish remain unclear." Probably should be specified as early germ cell types. Five developmental stages of germ cells (oocytes) in zebrafish are well defined by Kelly Selman since 1993. In fact, this manuscript was focusing on fine classification and separation of different germ cell types prior to stage I, which should be highlighted or emphasized, because this is not possible until scRNAseq became available.

The label D for figure 1D is too far, should be close to the sub-panel.

It is hard to view two different symbols illustrated in Figures 3C and 4A, suggest to separate symbols into side-by-side figures.

The cell numbers for mutant in figure 3D are much higher than those in WT, why?

Why are there two nanos2 lines (lines 1 & 3) in the same GSC group with different expression profiles in Figure 3E? nanos2 expression was not listed for the mutants in figure 3E. Y-axis had label for mutant, but does not have the label for WT, not was explained in the figure legend.

Lines 171-173, without foxl2l, which is a transcriptional factor, the gene expression would change. So, is it still validate using wildtype marker genes to define developmental stages found in foxl2l. Can different types of early germ cells be separated using scRNAseq data from the foxl2l mutant? Can pseudotime plots be used for the foxl2l mutant? If so, what looks like?

Abnormal increased expression of nonfunctional foxl2l (figure 3F) in foxl2l mutant suggests a feedback mechanism.

Co-expression and increased expression of nanos2/id1 with foxl2l in GSC in foxl2l mutant is interest, as these two genes are expressed separately in WT. Can overexpress of nanos2 and id1 lead to loss of oocytes, or dedifferentiation of oocytes (Prog-C) back to PGCs, and female to male sex reversal?

Guest Editor (suggestions for the authors)

The Hsu et al. manuscript presents valuable data that extend earlier single cell transcriptomic data on zebrafish gonads to help us better understand various initial stages of germ cell development in zebrafish and the role of Foxl2l in oogenesis. Data are well presented. Some of the conclusions do not seem to be well supported by the data, or alternatives are sometimes not considered.

The authors investigated development of zebrafish germ cells by isolating Piwil1-expressing cells and performing scRNA-seq in wild types at 26 days post fertilization, which augments published single cell transcriptomic studies of whole zebrafish gonads at 40 dpf, and they performed scRNA-seq experiments on mutants lacking foxl2l activity. These data, validated by in situ hybridization studies, allowed the identification of new stages of germ cell development. Studies showed that foxl2l mutants failed to develop meiotic oocytes, and that mutant germ cells reached a modified committed progenitor stage (Prog-C) but did not become late progenitors (Prog-L). The cells the authors called Prog-C cells in mutants, unlike wild types, expressed some GSC markers, suggesting that they maybe shouldn't be called Prog-C cells in mutants. From this result, the text concluded that committed progenitors then reverted to germ cell stem cells (GSC), which then directly developed into spermatogenic cells, and hence the animals became males, concluding that this was a direct development into males.

Alternative hypotheses exist for these data. The expression of genes normally expressed by GSCs (e.g., nanos2) in what otherwise seem to be committed progenitors was interpreted as a reversion of Prog-Cs to GSCs. An equally possible hypothesis is that the normally GSC-expressed genes never turned off and that the mutant 'Prog-C' never fully developed. Both explanations need to be presented and arguments assessed for and against each of these explanations.

Additionally, the conclusion that there was direct male differentiation in foxl2l mutants without going through the female phase has alternative explanations consistent with the data. This direct male development argument assumes that the Prog-C stage has not already embarked on a female pathway. The text did not demonstrate that Prog-C, or even the earlier stages like Prog-E, had not already begun to be female. The very name 'committed progenitor cells' suggests that these cells are already committed to the female pathway in wild types.

These authors and others have shown that one of the earliest steps in female gonad development in zebrafish is an increase in the number of primordial germ cells compared to zebrafish that will become males, by 14 days post fertilization. The cell division marker analysis reported here suggests that those PGC proliferation steps are already occurring in the cells studied here. So do the 'Proliferation' arrows shown in Figure 6 for both the Prog-E and Prog-C stages. Thus, these stages would already be female-oriented germ cells. And thus, when the reversion of Prog-C towards Prog-C-to-GSC happens, what that really represents is a reversion of female pathway cells towards uncommitted GSCs that then embark on a male pathway, as happens in many zebrafish individuals due to stress or temperature or other features. Thus, the data presented here shows that the male pathway is taken after germ cells have embarked on the female pathway and turned on foxl2l, but subsequently thwarted from becoming a meiotic oocyte due to the lack of Foxl2l in the mutants.

The title (Single cell transcriptomes of zebrafish germline reveal stemness suppression and progenitor feminization by Foxl2l) should be revised because the conclusion that Foxl2l suppresses 'stemness' is not well supported. The data do show that foxl2l mutants have cell types that co-express GSC markers like nanos2 with genes usually expressed later in female germ cell development, like sycp1 and sycp3 (Figure S6), but that does not clearly mean that these mutant cells are in fact stem cells.

The 'allmarkers' files are not included in the supplementary data, making it impossible to see how genes were selected for display. These files must be added to the supplementary data.

Some suggestions for improving clarity in the text and figures.

P2 'factor, Foxl2l, is expressed in the progenitors committed to the ovary fate.'. Consider saying '…in the germ cell progenitors committed to the oocyte fate..'

P2 'Another single-cell profiling of foxl2l-/- germ cells reveals the arrest of early progenitors.'

Text here uses both present and past tenses. For describing the results of experiments done in the past, past tense might be better, as the text often uses in 'mediated mutation of foxl2l produced 100% male fish' and 'in stem cells) was elevated'. So maybe use 'revealed' instead of 'reveals'. And again could say '…early germ cell progenitors.' Just to be precise. Likewise, '…of nanos2 (germ cell stem cell)…'.

In Key Words, consider changing 'stem cell' to 'germ cell stem cell', and 'germ cell progenitor', which might better direct interested folks to this paper. Consider changing 'female' as a keyword, maybe it's too generic, to 'female-to-male sex change', which might get more focused hits.

P2 'increase of nanos2+ germ cell in foxl2l'. Use here plural, 'germ cells'…

P3 'contrast to higher vertebrates,'. Consider saying what specific taxon is meant here rather than 'higher'. Talking about higher and lower animals comes from a pre-scientific view of the scale of life.

P3 'system, but domesticated laboratory strains have no sex chromosome'. Use plural, 'chromosomes'.

P3 'the differentiation of zebrafish gonad. Zebrafish'. Plural, 'gonads'

P3 'of germ cells (nr0b1 mutants) or no germ cell (dnd morphans) develop'. Plural, and spelling, 'or no germ cells dnd morphants

P3 'However, the developmental stages of germ cells in zebrafish remain unclear.'

I think the developmental stages have been clear, but the molecular genetics that moves cells through these stages is unclear, and the authors' work subdivides certain stages.

P3 'and initiates oogenesis in XX gonad (Kikuchi'. Plural, 'XX gonads'

P3 'with foxl3 mutation become hermaphroditic possessing'. Several changes: 'with a foxl3 mutation become hermaphrodites possessing functional sperm'

P3 'and proved that Foxl2l controls the development of committed progenitor,'. Plural, 'progenitors'.

P4 'Germ cells play important roles in zebrafish sex differentiation, but their development is poorly studied.'

Maybe more accurate to say: 'Germ cells play important roles in zebrafish sex differentiation, but we still have much to learn about their development.'

P4 '2002. Fluorescent germ cells were isolated from piwil1:EGFP transgenic fish by cell sorting'

The reader needs to know when piwil1 turns on in germ cell development with respect to the timing of oogonia vs. spermatogonia commitment. We need to know if by using the transgene we miss early but important stages in germ cell development.

Supplementary Figure S2 must have a mechanism for the reader to identify the genes in the rows. They would be too small to read if written on the figure so it should be in an Excel file in the supplement.

P4. 'expression profiles of the top 37 out of a total of 1491 marker genes'

Top based on what criterion? P-adj? P? Log2 fold change? Most specific 2 or 3 per cluster? Or something else? And what defines a 'marker gene'?

P4 'markers (sycp1, sycp3, dmc1, buc and zp3e). Are buc, which helps assemble germplasm, and zp3e, an eggshell component, \involved in meiosis? Consider rephrasing.

Suppl Figure 3 and 4. Although most of the markers show very similar patterns, thus cross-validating both datasets, it looks like for the markers CU467646.3, si:zfos-1505d6.3, and zp3e(zgc:171779), that the developmental pathway extended much further in the Liu data than the Hsu data. This isn't a huge concern, but it should be mentioned and discussed if there is a methodological difference that could explain this difference. Maybe it's just the age difference?

P5 'Zebrafish progenitors are round in shape'. Progenitors of zebrafish oocytes…

P5 'et al., 2010). To characterize these three progenitor subpopulations,'

I'm not sure which three are referred to here. When adding in GSC, we get four. Please clarify.

. 'et al., 2010). To characterize the GSC, Prog-E, Prog-C, and Prog-C(S) progenitor subpopulations,'

P5 'hybridization (FISH) and detected GSC as positive'. Make plural GSCs.

P5 Characterization of the three stages of progenitor subtypes using nanos2, foxl2l, and rec8 was useful.

P5 'premature stop codon disrupting the FH domain of Foxl2l'. Spell out 'forkhead'.

Figures 2C and 2D. These would be more impactful if error bars were displayed on the graphs and numbers of trials or numbers of individuals were shown.

P5 'non-meiotic cystic germ cells were found in foxl2l mutant (Figure 2E).'

Plural: mutants

P5 'foxl2l mutant germ cells are arrested at committed progenitor.'. 'foxl2l mutant germ cells are arrested at the committed progenitor stage.'

P5 '(Supplementary Figure S6A). The expression of the 37 top WT marker genes'. Again in a couple of words, explain the definition of 'top' genes more precisely.

P5 'found that these marker genes were expressed at similar stages as they represented for WT cells.'. The text here needs to say that WT and mutant cells were co-clustered. Also, it would be interesting to see how the pseudotime navigates its way through this blob of cells.

Figure 3D. Label the horizontal axis as cluster number. Also, consider changing 'dominant' in the legend to 'predominant'.

P6 'integrated groups (Figure 3D). Dominant stages were obtained'. Define' Dominant stages'.

P6 'calibration against their prevalence rates in the entire cell population.'

Does 'entire cell population' mean all of the mutant cells? or all of the mutant plus WT cells?

P6 'For mutant cells, while groups I5-I8 had two dominant stages,'

If dominant means the one that's present most frequently, then more than one shouldn't be dominant unless there's a tie. So 'dominant' needs a clear definition.

P6 'that WT and mutant cells with high cell-to-cell similarity belong to similar

developmental stage.'

'that WT and mutant cells that have high transcriptome similarity likely belong to the same developmental stage.'

P6 'Groups I3 and I6 are particularly unique.'

'Groups I3 and I6 are unique.' Unique means the only one of its kind. So there can't be gradations of being unique.

Figure 3E. Should the group at the bottom labeled 'Mutant', be called something like 'Mutant-specific cell types'? Because all of the data in the table refer to mutant cells, right?

P6 'mutant cells additionally expressed zgc:158463'

In addition to what? So mutant cells expressed all the genes of WTs in I3 but in addition zgc:158463 which wasn't expressed in the WT cells in I3? The same question for I6 and tp53. Please clarify.

P6 'Furthermore, the Prog-C marker foxl2l was expressed in mutant cells of I1,'

So that means that the mutant mRNA had normal stability? Wasn't degraded by nonsense mediated decay?

P6 'The aberrant foxl2l expression in some earlier GSC and Prog-E plus the accumulation of foxl2l expressing cells in the mutant indicate that progenitor development is impaired in foxl2l mutants.'

'in the mutants and the failure to find clusters representing more mature stages indicate that progenitor development is impaired in foxl2l mutants' [could that be added or not?]

P7 'throughout the development'. Change to: 'throughout development'

P7 'Stratified by integrated groups, we compared'. Check to see if there's a better word than 'stratified' to use here. As written, it means that 'we', the authors, were arranged into layers by integrated groups.

P7 'WT and mutant cells at the same dominant stages'. Each time that 'dominant' is used, see if 'predominant' would be a better word.

Figure 4A. The triangles are very difficult to see. Would some other visualization work better?

P7 'Figure 4B shows genes that were differentially expressed between WT and mutants.'

Are these all of the genes that were differentially expressed between these two genotypes? Or a subset? And if a subset, then how were these specific genes selected to display?

In addition, be explicit about the comparison. Does red mean that the WT or the mutant is overexpressing? E.g., is Dmrt1 overexpressed in WT or in mutant?

Figure 4B. The gradient in the main portion of the figure should match the order shown in the scale. The darkest red indeed is on top, but then for the blues, the darkest is on top, not at the bottom as in the scale.

Figure 4c. If id1 and nanos2 are marker genes for GSCs and disappear at later stages in WTs, and if their expression doesn't disappear in mutants, then the mutant cells don't advance to the stage where expression of these genes normally disappear, right? So there wouldn't be any mutant cells at stages G-P to Prog-C(S) because id1 and nanos2 are still expressed in mutant oocytes in these columns in the figure? Or at least these would be variants of the wild type clusters.

In addition, it looks like foxl2l inhibits its own expression at early stages, because without foxl2l activity, the expression of foxl2l is much higher in mutants and WTs in GSCI to Prog-E columns.

P7 'The expression of Dmrt1 and dele1 in WT was steady throughout the developmental stages (Figure 4C), but Dmrt1 was abnormally high and dele1 was abnormally low in both mutant GSCI and Prog-C/Prog-C(S).'

The fact that Dmrt1 was abnormally high in mutants suggests that a role of foxl2l activity is to inhibit expression of Dmrt1.

The fact that dele1 was low in mutant Prog-C(S) cells in mutants, suggests that foxl2l activity encourages expression of dele1.

Those conclusions, if authors agree with them, might have more impact here than what's written: 'These results indicate multiple defects of gene expression in the mutants.'

P7 'Co-expression of foxl2l with Dmrt1, nanos2 or id1 in foxl2l mutant'. Consider changing to '…in foxl2l mutants', or to '…in foxl2l mutant germ cells'.

Figure 5A legend. Say explicitly that cells expressing both genes are represented by yellow. It's obvious of course from the scale grid but would help some readers to say it explicitly.

P7 'increased in the mutant (Figure 5B). Moreover, the number of nanos2+ cell'

Plural: …of nanos2+ cells…'

P7 '(Figure 5C). To further investigate whether GSC is affected in mutants' '…whether GSCs are affected…'

P7 'counted the number of nanos2+foxl2l- cells, and found that the number increased dramatically in mutants at 26 dpf'. increased dramatically in mutants compared to WTs? Or increased in mutants at 26 dpf compared to some other time in development?

P7 'GSC cells in mutant together'. Plural: 'GSC cells in mutants together'

P7 'GSC cells in mutant together suggest that some Prog-C may revert to GSC in', What is the evidence that they develop to Prog-C and revert rather than that they never made it fully to Prog-C and simply maintained some of the gene expression properties of the earlier stages? These are two competing hypotheses, and both should be mentioned.

Figure 6 legend. Mention 'G-P transition' in the legend or remove it from the sketch. The word 'Proliferation' in black on a dark gray field is hard to read.

P9 'they undergo G to P transition for further differentiation into progenitors.'

Define 'G-to-P'.

P8 'Progenitors include three types: Prog-E, Prog-C, and Prog-L.' It would be good if the text here could state the transcriptomic features that distinguish these three phases from each other like the text did for the two kinds of GSCs.

P9 'The fate of germ cells is determined following the instruction of the supporting cells.' Compare to germ cell sex determination in mammals.

P9 'deficiency, however, leads to cell arrest at Prog-C as shown in this article.'. 'deficiency, however, leads to cell arrest at Prog-C as shown by results presented here.

P9 'commitment to the female fate as shown here. It indicates'. What is the antecedent of 'It'?

P10 'Id1, inhibitor of DNA binding 1, is a transcription regulator that lacks DNA…'.'…that lacks a DNA …'

P10 'stem-cell genes (id1 and nanos2) and male gene Dmrt1.' First, '…and the male gene Dmrt1'

Second, are there other early male genes that are suppressed by Foxl2l action besides Dmrt1?

P10 'The alternative development of foxl2l mutant germ cell'. Plural: '…mutant germ cells'

eLife. 2025 Jun 11;14:e100204. doi: 10.7554/eLife.100204.sa2

Author response

Essential Revisions:

1) Evidence that committed progenitors revert to germline stem cells (GSCs) is weak. Functional data are required to draw this conclusion. The phrase 'reveal stemness suppression' should be removed from the title and alternative explanations should appear in the text. The claim that suppressing stemness is necessary for female differentiation is not well supported, for example, cells might initiate the committed progenitor program but fail to shut off and continue to express part of the GSC program.

We have removed the term “stemness suppression” from the text and added a discussion of alternative explanations.

2) Essential literature that places the work in context must be discussed, with several cases noted in each specific review.

We have removed the term “stemness suppression” from the text and added a discussion of alternative explanations.

3) Alternative explanations to the statement that committed progenitor stage is "the gate toward female determination" must be discussed, for example, that the late progenitor stage is the key stage for female determination.

We agree that the existence of Prog-L is required for female differentiation. Our results show that foxl2l drives the differentiation of Prog-C to Prog-L, and therefore guides progenitor toward female pathway. We have clarified our statement about the importance of Prog-L in female differentiation.

4) Consider alternatives to the conclusion that the progenitors are differentiated because it is based solely on the expression of foxl2l, which is initially expressed in the juvenile ovary state that lab strains develop through.

We do not think that all progenitors are sexually differentiated. Prog-E and GSC probably do not have a sexual identity. Only Prog-L is stepping towards female meiosis. Prog-C differentiate into Prog-L, and commits germ cells into the female pathway. We derived this conclusion not only based on Foxl2 expression, but also based on our whole-genome transcriptomic analysis and foxl2l mutant phenotype. Furthermore, foxl2l is not “initially expressed in the juvenile ovary state that lab strains develop through”. We have examined the expression of foxl2l in the wild Nadia strain and detected the same pattern as that in the lab strains during the larval stage (Figure 2). Our result showed that at 26 dpf, foxl2l is not expressed in WT meiotic oocytes, instead it is expressed in two type of progenitor, Prog-C/Prog-C(S) and Prog-L, and its level is reduced after meiotic entry. Our data show that Foxl2l drives the differentiation of Prog-C to ProgL, and the existence of Prog-L is required for female differentiation.

5). Consider alternative explanations for the differences in the action of Foxl2(Fox3) in zebrafish and medaka.

Medaka Foxl2l promotes meiotic entry and the expression of meiotic genes. Zebrafish Foxl2l also has the same functions as shown by our previous RNAseq result (Dev Biol, 517, 91-99, 2024) and the expression of zebrafish foxl2l in Prog-L (Figure 1D). Additionally, zebrafish foxl2l deficiency leads to activation of nanos2 in mutated Prog-C. It is unknown whether nanos2 is activated in progenitor cells of XX foxl2l-/- medaka. Furthermore, unlike the absence of meiotic germ cells in mutant zebrafish, in foxl2l-/- medaka, meiotic germ cells are still present in XX juvenile gonads. These differences raise two possibilities. First, Foxl2l in medaka and zebrafish share a similar role in regulating progenitor differentiation. The difference in mutant phenotypes can be due to a stronger female tendency in medaka that prevents complete blockage of female differentiation and meiosis in foxl2l mutants. Second, zebrafish Foxl2l has an additional role in driving progenitor differentiation. We have now included the discussion of the alternative explanations about the difference in the action of Foxl2l in these two species.

6) Methods require more details on how fish were selected for scRNA-seq.

We have added more details about the collection of fish used for scRNAseq in the text.

Reviewer #1 (Recommendations for the authors):

Specific points.

1) Page 7 lines 258-261. The authors conclude that the committed progenitors revert to GSCs based on the coexpression of nanos2 and foxl2l nanos2 and later on the same page line 262, based on expression of id1 in mutants but not WT. Without functional data demonstrating that the progenitors revert to an early state, alternative interpretations should be considered. For example, can the authors exclude the alternative possibility that the cells initiate the committed progenitor program but do not shut off and continue to express the GSC program such that the coexpression of both programs blocks differentiation? It would help to discuss these findings in the context of Fox family functions. Some Fox family members, Foxl2 and FoxPs for example, are known to be both activators and repressors of transcription or act primarily as repressors. Relevant to this work, repressive activity of Foxl2 has previously been reported in the mammalian ovary (Pisarska et al. Endocrinology 2004, Pisarska Am J. Phys Endo. Metabolism 2010, Kuo Reproduction 2012, Kuo Endocrinology 2011, as well as several more recent publications). In this context interfering with Foxl2 repression was proposed to accelerate follicle recruitment and cause premature ovarian failure.

Thanks for your insightful comment. Alternative interpretations are indeed needed. We have examined the expression of GSC markers including nanos2 and found that in the mutants, nanos2 or other GSC markers were not significantly upregulated in the GSC-to progenitor transition (G-P) and early progenitors (Prog-E) (Figure 4B). The expression of these GSC markers was also low in the integrated clusters when G-P and Prog-E stages were prominent (Figure 3D and Figure 3E). Only one GSC marker, nanos2, was high in mutant Prog-C. These results indicate that perhaps some mutant Prog-C acquires some GSC properties with the upregulation of nanos2 instead of a continuous GSC program. We have now included both interpretations and clarified our rationale about the mutant cells gaining a new GSC property. We have also included the discussion about the dual regulation of Foxl2 in the discussion.

2) Discussion Page 9 lines 289-293. This section is not entirely clear. First GSC-I and GSC-II subtypes are introduced, then it is stated that "all GSCs self -renew" and a Saito 2007 paper is cited. This is followed by a description of "two types of daughter cells derived from GSC self-renewal". Does the conclusion that "all GSCs renew" still hold in light of the two GSC populations? Would a division that produces a daughter that cannot renew be considered "self-renewal" or instead a non-renewing division?

Our original statement about the self-renewal in GSC was probably not accurate. Here we speculate that the two subtypes of GSC might represent two different daughter cells generated by asymmetric division of GSC. We have revised our discussion in the text.

3) Discussion Page 9 Line 295-7. The authors conclude that the committed progenitor stage is "the gate toward female determination" and that they "stay at S-Phase temporarily before differentiation". Is this conclusion based solely on RNA expression? Also, in several species, including zebrafish, meiotic entry has been correlated with ovary development. Can the possibility that the late progenitor stage, the stage when meiotic genes are detected and a stage missing in foxl2l mutants, is the key stage for female determination be excluded?

4) The same comment as above pertains to Page 9 lines 281-284 of the discussion.

We indicate that Prog-C “stay at S-pase temporarily” because we detected a group of Prog-C cells, (Prog-C(S)), that are rich in S-phase genes and nucleosome assembly (zgc:153405, si:ch211-113a14.24). We refer Prog-C as a gate before female differentiation because the foxl2l mutant cells are halted at Prog-C as shown by in situ hybridization, cell counting, and scRNAseq analyses. Prog-L as well as the meiotic stages are completely absent. We agree that the existence of Prog-L is important for the female differentiation, but the changes already start at Prog-C. We have revised our statement to make it more precise.

5) Page 3 Line 64 states that "the developmental stages of germ cells in zebrafish remain unclear". This is somewhat misleading as recently published single cell datasets have analyzed and defined the stages of germ cell development, including during and after sexual differentiation. This should be clarified, and the relevant published work should be cited. (Liu et al. eLife 2022, Wilson et al. Front Cell Dev Bio 2024.)

Thanks for your suggestion. We have cited the references and revised our statement.

6) Page 3 Line 70 might be confusing because it states that "XX medaka fish with foxl3 mutation become hermaphroditic processing functional sperms". Please clarify whether they also produce oocytes or are genetically female "XX" but make sperm.

XX medaka fish with foxl3 mutation produce both functional sperms and oocytes. Thanks for your suggestion. We have clarified this point.

7) On Page 9 lines 305-307, the authors discuss prior working showing that loss of germ cells leads to male development and that germ cells are required for female development. Then lines 307-309 state that this work shows some progenitors are already sexually differentiated. First, the stages compared are completely different. The early work looks at the primordial germ cells and their loss in the first few days of development before a gonad forms and this work looks after a gonad has formed and when sex is being determined.

Both previous studies and our study indicate the important role of germ cells in zebrafish sex differentiation during gonadal development. The earlier works show that the abundance of primordial germ cells contributes to sex differentiation. Our current finding further suggests the existence of female identify in some germ cells at the juvenile stage and discusses the importance of germ cell in sexual differentiation.

The second concern is that the conclusion that the cells are differentiated is based on the expression of foxl2l, which is expressed in the juvenile ovary state that lab strains have been shown to develop through (Wilson et al. Front Cell Dev Bio 2024). While it is fair to state that some cells express ovary markers, are they really differentiated?

We indicated that the Prog-L cells are differentiated based on the analysis of whole-transcriptomes, not simply by foxl2l expression. This analysis is furthermore backed up by functional analysis. In the absence of functional Foxl2l, Prog-L is absent and the mutant gonads eventually undergo male development. Therefore, we infer that Prog-L stage is required for female differentiation and Foxl2l is essential for the differentiation of Prog-C to Prog-L. On a different note about the expression of foxl2l, our in situ hybridization data show that foxl2l has the same expression pattern in the wild Nadia strain as in laboratory strains at the larval stage (Figure 2). Our analysis of 26-pdf scRNAseq further showed that foxl2l is expressed in two types of progenitors and becomes barely detected in the meiotic cell.

For example, in the context of the foxl2l mutant, GSCs and early progenitors inappropriately express foxl2l, but the mutants develop as males. Please clarify.

In foxl2l mutants, the mutant foxl2l mRNA is stable, but fails to produce functional protein, resulting in the impairment of female differentiation and progenitor development. We detected mutated foxl2l transcripts in mutant cells, but this does not mean that these cells contain Foxl2l. These mutants develop into males eventually.

8) Page 9 lines 315-318 compares medaka and zebrafish foxl2l mutants and seems to suggest that Foxl2l is required for meiosis in medaka but has a different role in zebrafish. Is it possible that foxl2l plays a similar role in repressing the earlier developmental programs of GSCs and early progenitors in both cases?

We agree that zebrafish and medaka Foxl2l share the same role to promote female meiosis. But we are not certain whether medaka Foxl2l represses early progenitor development. In foxl2l-/- zebrafish, meiotic germ cells are absent, while in foxl2l-/- medaka, XX gonads contain meiotic germ cells. The difference raises two possibilities. First, Foxl2l in medaka shares similar roles in progenitor differentiation and repressing GSC program mekada and as in zebrafish. The different phenotypes can be due to the stronger female tendency contributed by genomic background in medaka that prevents complete blockage of female differentiation and meiosis in foxl2l mutants. It would be interesting to know whether the GSC program is activated in progenitor cells of XX foxl2l-/- medaka. Second, zebrafish has additional role in driving progenitor differentiation. We have now included a discussion of the different phenotypes in these two species.

Is it possible that continued expression of these early programs interferes with activation of meiotic genes? Might this account for the absence of the late progenitor stage in foxl2l mutants since it is defined by expression of foxl2l and meiotic genes?

In zebrafish foxl2l mutants, early programs are not continuously activated. GSC marker gene expression is low at the G-P transition and Prog-E stages of foxl2l mutants (Figure 3D and 3E), indicating that in foxl2l-/- mutants, nanos2 reappears rather than continuous expresses from GSC to progenitor stages. In foxl2l mutants, late progenitors and the meiotic stages were completely lost. The loss of late progenitors in zebrafish foxl2l mutant is demonstrated by the analysis of the whole transcriptomes rather than merely foxl2l and meiotic genes only. Therefore, the absent of meiotic genes in foxl2l mutants might be due to the completely loss of late progenitors as well as meiotic cells. This is comparable with differential gene expression observed in Prog-C. The dysregulation of these genes can contribute to the block of Prog-C differentiation resulting in the loss of Prog-L. Moreover, we cannot rule out the possibility that nanos2 in foxl2l mutants might interfere with the activation of meiotic genes as well as other unknown genes. Therefore, it is possible that the absence of Prog-L in mutants is due to an interference of meiotic expression by the differentially expressed genes observed in mutant Prog-C.

9) Page 10 lines 328-330. Direct differentiation of testis without going through a female phase is discussed. Recent work showing direct differentiation in wild strains should also be discussed and cited (Wilson et al. Front Cell Dev Bio 2024).

Thanks for your suggestion. We have added the description and cited this paper.

10) Page 10 lines 359-360. The authors state "Foxl2l may ensure female differentiation by preventing stemness and antagonizing male development." It is unclear why suppressing stemness would be necessary for female differentiation since female zebrafish have stem cells as do male zebrafish. It seems likely that turning off the GSC and early differentiation programs is important for allowing expression of meiosis and oocyte differentiation genes.

development. It is true that we have not proved whether suppression of stemness is required for female differentiation. Yet as this reviewer pointed out, it is likely that turning off the GSC and early differentiation programs is important for allowing expression of genes in meiosis and oocyte differentiation, thus leading to final female differentiation. To avoid confusion, we revised the discussion and speculation about the possible actions of Foxl2l in the text.

11) Page 11 370-374. "Thus, p53 may assist alternative differentiation of mutant germ cells". What is the evidence that p53 is involved in differentiation of germ cells? Sex bias has not been associated with the published p53 mutants. Is it possible that p53 eliminates germ cells that are simultaneously expressing programs of multiple stages? While p53 has been shown to be important for ovary to testis transformation in mutant contexts in adults, it is dispensable in mutants that disrupt ovary development at earlier stages. Please discuss in the context of, and cite, the relevant ovary literature Rodriguez-Mari et al. PLoS Gen 2010, Shive PNAS 2010, Hartung et al. Mol. Reprod. Dev 2014, Miao Development 2017, Kaufman et al. PLoSGen 2018, Bertho et al. Development 2021.

Thanks for your comments and suggestion. The tp53-/-foxl2l-/- double mutant cannot alleviate the all-male phenotype of foxl2l-/- mutant (Dev Biol, 517, 91-99, 2024), indicating that the male development is not due to p53-mediated germ cell apoptosis. We have compared the relation of tp53 between these mutants and cited the suggested papers. The involvement of tp53 in germ cell differentiation is our assumption derived from high tp53 expression in foxl2l mutants in our scRNAseq result. Since p53 was enriched in certain mutant clusters, and the tp53 mutation fails to rescue the all-male phenotype, it is possible that tp53 expression in these mutant clusters may play roles other than inducing apoptosis. Although unknown in germ cell differentiation, p53 has a role in promoting the differentiation of airway epithelial progenitors and embryonic stem cells. We have pointed out in the Discussion that the notion about the role of p53 in germ cell differentiation is our assumption.

12) Figure 2D. It would be helpful to show the individual fish points or to present the data as a violin plot.

Thanks for your suggestion. We have included the individual fish points in Figure 2D and revised the figure legend.

13) Figure 2F. The numbers are low here only 3-4 gonads are examined for each genotype and marker and in the mutants only 1/4 expresses the "testis" markers.

We performed this staining experiment at 28 dpf. Since 28 dpf is the beginning of male development, it is expected that male morphology only starts to appear in a portion of the gonad. The number does not need to be high. The presence of the testis marker gene expression already indicates male differentiation.

14) Figure 6 model. What is the evidence that prog C becomes GSCs again/revert versus continuing to express earlier programs?

Thanks for your comments. We have examined the expression profile of GSC markers including nanos2, cbx7a, psmb13a and dnmt3bb.1 in mutant cells. All of them showed low expression in GSC-to-progenitor transition (G-P) stage and Prog-E stage in mutants (Figure 3D, 3E). In the DEG analysis, no GSC markers were upregulated in GSC-to-progenitor transition (G-P) or Prog-E. Only nanos2 were abruptly upregulated in Prog-C. This result indicates that the GSC program was shut off during the differentiation from GSC to progenitor. Therefore, it is more likely that in the absence of functional Foxl2l, Prog-C acquires some GSC property with upregulation of nanos2.

Reviewer #2 (Recommendations for the authors):

The evidence that foxl2l cells are arrested at the committed progenitor stage (a possible novel phenotype) is based on the absence of expression of rec8a and Sycp3. Another explanation is that these gonads have already differentiated as testes.

The arrest of committed progenitor stage in mutant is based on the whole transcriptomic profile (Figure 3A), rather than merely the expression of rec8a and sycp3. The foxl2l mutant germ cells failed to be classified as Prog-L or meiotic stages. Consistently, marker genes for Prog-L and meiotic stages including sycp3 and rec8a were consistently at low levels (Figure 3—Figure Supplement 1) (Figure 3B). The foxl2l-/- gonads will eventually become testes and their germ cells will be differentiated as sperm cells, but not at this early stage. At 26 dpf when scRNAseq was performed, we could not detect the expression of male genes including tekt1 and odf3b in our analysis. Therefore, the mutant cells were not yet differentiated into testis at 26 dpf.

Is rec8a also expressed in male germ cells or is it female germ cell specific? Germ cells in testes are usually delayed in entering meiosis (and thus expressing Sycp3) relative to females so it is not clear if this represents the expected trajectory of testis vs. ovary or a novel developmental state caused by absence of foxl2l.

The rec8a transcript can be detected in both testes and ovaries as shown before (DOI: 10.1371/journal.pgen.1009127). The male characteristics were detected in mutant gonads at a later stage, 28 dpf. Some mutant Prog-C cells additionally acquired a novel developmental state (nanos2+) caused by the absence of Foxl2l. Therefore, it is not a simple testis development from indifferent gonads, rather, the mutant germ cells went through a process involving the untimely GSC gene expression at 26 dpf.

Was expression of any male-specific genes identified in the mutant data set (e.g. tekt1 or odf3b) which could indicate these germ cells were already committed to spermatogenesis?

We did not detect transcripts of male markers including tekt1 and odf3b in mutant cells at 26 dpf by scRNAseq. Therefore, these mutant cells probably have not entered spermatogenesis at this stage.

Also, by "arrested" do the authors mean that they are in a cell cycle arrest? If so, this can be assayed with BrdU or EdU labeling.

The word “arrested” here means “blocked”. Cells cannot be further differentiated to the next step (Prog-L). It does not involve cell cycle arrest. We have revised the sentence to indicate the arrest of germ cell development rather than the arrest of cell cycles in the mutants.

L99: incomplete sentence.

L128: The publication associated with this dataset should be referenced (it's referenced in the figure legend but should also be referenced in main text).

L138: Using a negative result (absence of staining) to argue for a particular cell stage (prog-E) is problematic. To bolster this argument the authors should consider pointing out that this result is consistent with the scRNA-seq data that predicts little to no overlap between nanos2 and foxl2l expressing cells. Perhaps adding the nanos2 and foxl2l and rec8a UMAPs to Figure 1 would help.

Thanks for your suggestion. We have done the revision.

L148 and Figure 2A. Wilson et al. (2024) showed that the gonads of Nadia ZZ fish appear to directly develop into testis without producing oocytes during the bipotential stage. The data in Figure 2A showing foxl2l expression in the ZZ gonads is therefore unexpected. Can the authors comment on this?

The expression of foxl2l is dynamic during development. At the larval stage, foxl2l is expressed in all bipotential gonads in both ZZ and ZW. However, during the sexual differentiation stage, foxl2l expression gradually is decreased in the male gonads while sustained in the female gonad (ZW) at juvenile age. Many sex-specific genes are first expressed in bipotential gonads before their expression becomes sex-specific; these genes include Sox9 (Development 137:3921-3930, 2010), Dmrt1 (Nature 402:601-602, 1999), and zebrafish cyp19a1a (aromatase, Dev Biol, 344:849-856, 2010). Therefore, foxl2l is not the only gene that is expressed in early bipotential gonads, irrespective of the future sex identity of the gonad. The functional relevance of this early expression in all indifferent gonads has not been investigated.

L154: "…barley observed in testis." Should this be "…not observed…", or is there low but detectable expression of foxl2l in the testis (not clear from Figure 1D)?

Thanks for your suggestion. The expression of foxl2l was not observed in the testis by ISH. We have done the revision.

L170: It is not clear from the Methods how the fish were selected for scRNA-seq. With the piwil1:egfp line it is possible to distinguish wild-type males from females at the 27 dpf timepoint used in this analysis based on the size of the gonad (large=female, small=male). Were only fish with large gonads used for WT and foxl2l mutant scRNA-seq? If fish were chosen randomly, it would be expected that germ cells from testes would also be included.

The sizes of male versus female gonads can be distinguished only when these gonads are carefully examined under a fluorescent microscope. To assure the quality and abundance of active germ cells for RNA isolation, one has to dissect fish trunks quickly without checking gonad size. Therefore, germ cells from both presumptive male and presumptive female gonads were included. This is why we did not find obvious male developmental trajectory from our WT database. In zebrafish, around ~80% of juveniles undergo female development first with oocyte development (Tissue and Cell 72:101545, 2021). Only ~ 20% of juveniles are presumptive male remaining as undifferentiated germ cell (gonocyte) before male development after 28 dpf. Our scRNAsq was done in germ cells of 26-dpf gonads, earlier than the initiation of typical male development after 28 dpf. Therefore, germ cells in these 20% presumptive male gonads were undifferentiated GSC or progenitor type, which may be quite similar to those cells from presumptive female gonads. Thus, it is reasonable that we did not detect male cell clusters in the WT scRNAseq dataset.

When do foxl2l mutant gonad first start looking like testes? Is there a delay in testis development relative to wild-type?

Spermatogenesis started to be observed in some foxl2l mutant gonads at 28 dpf, as shown by DAPI staining and the staining of male markers, tekt1 and odf3b (Figure 2F). This time point of male development is not delayed compared with that in WT.

Figure 2B: Including age-matched wild-type testes in Figure 2B may help clarify if the foxl2l gonads are different or similar to a normal male gonad at this stage of development.

We have published the histology analysis of mutant testis and showed that mutant testis has the same histology as the WT testis. Here we have cited this paper.

Figure 2D: I would suggest replacing si:ch211-191j22.7 with meiosin, since this is the clear ortholog.

Thanks for your suggestion. We have done the revisions.

Figure 3B: It is hard to see the morphology of the nuclei in these images. Grey scale is perfect for the DAPI channel, but it would be helpful to increase the apparent signal intensity using an image processing program like Photoshop so that it is possible to better assess the morphology of nuclei.

Thanks for your suggestion, we have revised figure 3B as suggested.

Figure 3C: I am having a hard time understanding the meaning of Figure 3C. For the WT cells, the nanos2+ cells are restricted to the W1 subcluster (Figure 1D), indicating that the GSCs are homogeneous with respect to gene expression. However, when WT cells are clustered with mutant cells the WT GSCs now partition to two different subclusters (I1 and I2). When the "control" cells are not clustering tightly, it calls into question how much this clustering represents reality. How do the authors rationalize this?

During clustering analysis, its granularity is influenced by the dataset size and is adjusted using a resolution parameter. To ensure robustness, we selected the optimal resolution by identifying the value that produced the highest average silhouette score within a feasible range. This led to the clustering of all GSCs into a single group in the WT dataset. When WT cells were co-clustered with mutant samples, the increased number of GSCs enabled higher-resolution clustering, leading to the identification of two distinct clusters, GSCI and GSCII. These clusters displayed varying levels of upregulation in marker genes: one subset of WT GSCs exhibited higher nanos2 and lower dnmt3bb.1 expression, while the other subset showed the reverse pattern (Figure 1—Figure Supplement 3). While this subdivision is more apparent in the co-cluster dataset, it is not unique to mutants. In both WT and co-cluster analyses, subsets of WT GSCs displayed distinct transcriptomic profiles, in which nanos2 and dnmt3bb.1 were high in different subsets of GSC cells (Figure 1—Figure Supplement 3). The high expression of nanos2 and dnmt3bb.1 in different subsets of WT GSCs can also be observed in 40-dpf ovaries in a public database (DOI: 10.7554/eLife.76014) shown below. Therefore, these findings support the presence of GSC subtypes in both WT and mutant samples.

L194: Figure S7B is unfortunately not helping me understand how the data was processed. What is meant by "transfer developmental stage?" Does this require/assume that the mutant cells be a specific stage that can be mapped directly onto wildtype stages? I wonder if forcing the mutant cells to cluster with wildtype cells is somehow skewing the results. How different is the clustering if only mutant cells are used?

We used the term "transfer developmental stage" directly from the name of the software TransferData in Seurat v3.2.3. This process aligns mutant cells with developmental stages previously defined in WT cells, based on transcriptomic similarities as described in the Materials and methods. The function TransferData in Seurat v3.2.3 allowed the transfer of the developmental stages from the WT cells to the mutant cells. Initially, principal components (PCs) were derived from the WT cells based on the expression of their cluster markers. Mutant cells were then projected to the same PCs. The PCs and the projected PCs were used to identify the mutual nearest neighbors across mutant cells and WT cells. These mutual nearest neighbors were considered anchors and scored based on the consistency across the neighborhood structure of each dataset. Lastly, the developmental stage of each mutant cell was determined using a weighted voting mechanism. This process considers the stages of WT cells in every anchor, with votes weighted according to the anchor score and the proximity between the mutant cell and its corresponding anchor mutant cell. This approach assumes that mutant cells can indeed be mapped to specific WT stages through their transcriptomic profiles.

To mitigate potential bias, we subsequently conducted a co-clustering analysis of WT and mutant cells, employing a data-driven methodology to identify both shared and unique germ cell subtypes. This co-clustering does not rigidly assign mutant cells to predefined WT stages; rather, it delineates the transcriptomic relationships between WT and mutant cells, facilitating the identification of potentially novel mutant subpopulations.

To further address your concern, we performed an independent clustering analysis using only mutant cells as suggested. The results were highly consistent with those from the co-clustering analysis (see the confusion matrix below; Adjusted Rand Index: 0.8). This independent analysis reinforces the robustness of our conclusion.

In response to your feedback, we have revised the term "transfer developmental stage" to "infer developmental stage" and updated Figure 3—Figure Supplement 2 to include detailed descriptions of each step for greater clarity.

Figure 3E: nanos2 and cbx7a are listed twice. It is likely they should be psmb13a and dnmt3bb.1, respectively.

Thank you for your insightful question. Indeed, the labels of two genes (nanos2 and cbx7a) were mistakenly duplicated in Figure 3E. We have replaced them with the correct gene names: pasmb13a and dnmt3bb.1.

Was expression of any male-specific gene identified in the mutant data set (e.g. tekt1 or odf3b) which could indicate these germ cells were already committed to spermatogenesis?

We examined the expression profiles of the male-specific genes tekt1 and odf3b within our dataset. In WT cells, both genes exhibited minimal expression prior to the onset of meiotic prophase. Specifically, tekt1 was virtually undetectable, while odf3b was expressed in only a limited subset of WT cells during the GSC and G-P transition stages [(refer to Author response image 1)].

Author response image 1.

Author response image 1.

Open in a new tab

In the mutant dataset, the expression levels of tekt1 and odf3b similarly remained low and were restricted to a small fraction of cells. Importantly, no statistically significant upregulation was observed when compared to their WT counterparts within the same co-clustered groups [(refer to Author response image 2)]. We, therefore, conclude that at 26 dpf, no sign of spermatogenesis can be detected in these germ cells.

Author response image 2.

Author response image 2.

Open in a new tab

Sup Figure S8A: It would be helpful if this figure also included the UMAP shown in Figure 3C for reference (with I1-I7 labels). S8A is hard to interpret given the altered gene expression pattern of mutant cells. Some genes fit the proposed cell stages (e.g. dnmt3bb.1), but others are not as clear (e.g. nanos2, dhrs1).

We are not sure about this reviewer’s point about supplementary Figure S8A. Our original supplementary Figure S8A shows violin plots of differentially expressed genes but does not show UMAP and does not contain dnmt3bb.1 and dhrs1genes. Thus, supplementary Figure S8A is unrelated to the UMAP in Figure 3C. We guess this reviewer may want to comment on the UMAP of the original supplementary Figure S6A. To improve clarity, we have reorganized the supplementary figures and added our analysis methods to enhance readers’ comprehension. The UMAP in the revised Figure 3—Figure Supplement 1C (original Figure S6A) shows the inferred developmental stage of mutant cells using the method shown in the revised. Figure 3—Figure Supplement 1B (original Figure S6C) showed that marker of early stage (GSC, G-P, Prog-E and Prog-C/Prog-C(S)) were upregulated in distinct subset of mutant cells. For example, nanos2 is highly expressed in a different subset from dhrs1. The other analysis is co-clustering analysis. The flow chart is shown in the new Figure 3—Figure Supplement 2B and the co-cluster result is shown in Figure 3C.

L210: "This suggests that GSCs can be further divided into two finer stages, GSCI and GSCII." If you only see this subdivision in mutants, then it is hard to argue that these are representative stages present in WT.

We acknowledge your concern regarding the subdivision of GSCs into GSCI and GSCII. While this subdivision is more apparent in the co-cluster dataset, it is not unique to mutants. In both WT and co-cluster analyses, subsets of WT GSCs displayed distinct transcriptomic profiles, in which nanos2 and dnmt3bb.1 were high in different subsets of GSC cells (Figure 1—Figure Supplement 3). The high expression of nanos2 and dnmt3bb.1 in different subsets of WT GSCs can also be observed in 40-dpf ovaries in a public database (DOI: 10.7554/eLife.76014) Author response image 3. These findings indicate that GSCI and GSCII subtypes exist in both WT and mutant samples, reflecting shared biological heterogeneity.

Author response image 3.

Author response image 3.

Open in a new tab

Author response image 4.

Author response image 4.

Open in a new tab

Line 307: It is not clear what this statement means. It appears to imply that foxl2l-expressing cells are sexually differentiated. However, in Figure 2A it is shown that foxl2l is expressed in germ cells of ZZ fish during the bipotential stage, yet these germ cell presumable would have differentiated as sperm, not oocytes.

Thanks for your comment. Foxl2l might play diverse roles during the bipotential stage and sex differentiation stage. In bipotential gonads, foxl2l is expressed in both female (ZW) and male (ZZ) gonad, although it is considered a female-specific gene (Figure 2A). Many sex-specific genes are initially expressed in bipotential gonads before they become sex-specific. For example, mouse Sox9 is expressed in both XX and XY gonads before testis differentiation (Nat Genet, 14:62-68, 1996). The same for Dmrt1 from mouse, chicken, and alligator (Nature, 402:601-602,1999). The roles of these genes in the initial indifferent gonads are still unknown at the moment.

L302-322: The authors need to spell out the evidence that supports these statements. It could be argued that the difference is not in the role of foxl2l between Medaka and zebrafish, but the fact that medaka has a strong genetic sex determination mechanism while zebrafish does not. Thus, the apparent difference in phenotype is because medaka cannot sex reverse as easily as zebrafish.

Thanks for your comment. We have toned down our speculation. We pointed out the difference in phenotypes between zebrafish and medaka and the possibilities that cause the difference, including the different roles of Foxl2l between these two species or the existence of sex chromosome in medaka.

L331: For completeness, please add a reference to Wilson et al., (2024; PMID: 38529407)

Thanks for your suggestion, we have done the citation.

L356: The authors argue that one role of Foxl2l is to suppress Dmrt1 expression in progenitor cells, yet in the foxl2l mutant germ cells, Dmrt1 is upregulated in all mutant cell stages, from GSC to prog-C cells (i.e. even in cells that do not normally express foxl2l expression). It is possible that the upregulation in all cells is because the gonads from which these cells were isolated have already begun to differentiate as a testis and the germ cells are now more male-like than female-like. This would also suggest the possibility that male and female GSC's do not express equivalent amounts of Dmrt1.

DEG analysis in Figure 4 showed that the upregulation of Dmrt1 is significant only in GSCI, Prog-C and Prog-C(S) in foxl2l mutants. We cannot find the upregulation of other male genes in mutant cells. Therefore, it is more likely the upregulation of Dmrt1 in some mutant Prog-C(S) and GSCI is due to the loss of functional Foxl2l rather than the initiation of the spermatogenic program.

Reviewer #3 (Recommendations for the authors):

The major conclusion, i.e., title and short title may be misleading and may cause confusion. Stemness is the ability of a cell to perpetuate/continue/preserve its lineage, to give rise to differentiated cells. With foxl2l, progenitor cells committed cell (Prog-C) will continue to develop into late progenitor (Prog-L). In addition, there is no strong evidence that Prog-C was able to dedifferentiate back into GSC in foxl2l mutants. Prog-C could be simply lost in foxl2l mutant, and increased nanos2 cells could be due to increased proliferation of GSC. Therefore, "Foxl2l suppresses stemness" could be misleading.

Our data show that nanos2 expression was low in mutant cells at the GSC-to-progenitor transition stage and Prog-E stage, but become upregulated only in Prog-C (Figure 4C), indicating that mutant Prog-C may acquire some GSC property with upregulation of nanos2. Moreover, in the mutant gonad, both the nanos2-expressing Prog-C and the foxl2l-nanos2+ GSC were increased compared with WT. These results raise one possibility that some nanos2-expressing mutant Prog-C may revert to GSC in foxl2l mutant. Although foxl2l is absent in WT GSC, it is also possible that the loss of Foxl2l triggers the proliferation of GSC. We have added this possibility to the Result. We agree to tone down our statement and remove the word “stemness suppression” from our title.

Is there size and morphological difference between prog-C and prog-L?

No. The size and the morphology were similar between Prog-C and Prog-L.

Lines 23-24, 64-65 and in other parts of the text "germ cell types … remain elusive" "developmental stages of germ cells in zebrafish remain unclear." Probably should be specified as early germ cell types. Five developmental stages of germ cells (oocytes) in zebrafish are well defined by Kelly Selman since 1993. In fact, this manuscript was focusing on fine classification and separation of different germ cell types prior to stage I, which should be highlighted or emphasized, because this is not possible until scRNAseq became available.

Thanks for your suggestion, we have done the revision.

The label D for figure 1D is too far, should be close to the sub-panel.

Thanks for your suggestion, we have revised figure 1D.

It is hard to view two different symbols illustrated in Figures 3C and 4A, suggest to separate symbols into side-by-side figures.

Thank you for the feedback. We have revised Figures 3C and 4A so that the symbols are easier to read.

The cell numbers for mutant in figure 3D are much higher than those in WT, why?

We do not know the exact reasons why we obtained more mutant germ cells than WT. The initial number of sorted cells for scRNA-seq was comparable between the two groups, with 3,231 WT cells and 4,019 mutant cells collected. Following sequencing, 1,677 WT and 2,947 mutant barcodes were identified as containing intact cells using the default settings of the 10x Cell Ranger pipeline. Subsequently, we employed a rigorous two-step quality control procedure to eliminate low-quality cells, such as damaged cells and doublets, as detailed in the "Quality Control" section. To ensure the robustness of our analysis, we retained 769 high-quality WT cells and 2,399 high-quality mutant cells for downstream analyses. The greater retention of mutant cells after quality control may reflect variations in cellular properties, such as viability or transcriptomic integrity between the WT and mutant samples. It can be random or because of the higher numbers of germ cells at the Prog-C stage in the mutant gonads due to germ cell arrest and accumulation.

Why are there two nanos2 lines (lines 1 & 3) in the same GSC group with different expression profiles in Figure 3E? nanos2 expression was not listed for the mutants in figure 3E. Y-axis had label for mutant, but does not have the label for WT, not was explained in the figure legend.

Thanks for pointing out our typographical error. The labels of two genes (nanos2 and cbx7a) were incorrectly duplicated in Figure 3E. We have replaced them with pasmb13a and dnmt3bb.1. The original label “mutant” in the Y-axis of Figure 3E is misleading. We have changed to “Mutant Specific Markers”, which indicates the expression of mutant-specific markers in the mutant cells of each integrated cluster (I1-I8). The expression nanos2 in the mutant cells of each integrated cluster (I1-I8) is shown in line 1.

Lines 171-173, without foxl2l, which is a transcriptional factor, the gene expression would change. So, is it still validate using wildtype marker genes to define developmental stages found in foxl2l.

While we recognize that the absence of Foxl2l may alter the transcriptional landscape in mutant cells and we indeed detected many changes in gene expression as shown in Figure 4, our analysis demonstrates that many WT marker genes remain expressed across distinct subsets of mutant cells (Supplementary Figure S6A). Additionally, we also detected many cells that uniquely expressed mutant-specific marker genes (Figure 3E). These findings support their utility in delineating mutant germ cell types despite the altered gene regulatory environment.

Furthermore, we acknowledge the presence of unique early germ cell types in the mutant population. To address this, we performed a co-clustering analysis of WT and mutant cells based on their comprehensive transcriptomic profiles. This approach does not assume a direct correspondence between mutant and WT developmental stages. Instead, it facilitates the identification of novel germ cell types in mutants while characterizing the relationship between WT and mutant cells. The eight integrated groups (I1-I8) reveal both shared and distinct transcriptomic features, providing deeper insights into the altered developmental program of foxl2l mutants (Figure 3C).

Can different types of early germ cells be separated using scRNAseq data from the foxl2l mutant? Can pseudotime plots be used for the foxl2l mutant? If so, what looks like?

Trajectory and pseudotime analyses depend on critical assumptions regarding the selection of root cells and the underlying structure of cellular trajectories, whether they follow a tree-like, linear, or circular pattern. The inferred trajectories are highly sensitive to parameter choices, which are often informed by biological hypotheses and thus carry an element of subjectivity. Most trajectory inference tools, including Monocle3, are optimized for tree-like structures and may require parameter adjustments to accurately model looped or cyclic trajectories as hypothesized for the mutant cells. A pseudotime plot for the mutant cells may thus appears meaningless. To ensure a rigorous and transparent analysis, we opted not to include potentially subjective trajectory and pseudotime results, presenting only similarity-based analysis in the main text.

Abnormal increased expression of nonfunctional foxl2l (figure 3F) in foxl2l mutant suggests a feedback mechanism.

Yes, I agree that there is a feedback mechanism for foxl2l expression although we did not emphasize this point in the text. Figure 3F shows the increased number of foxl2l+ cells in foxl2l mutant. We also found that foxl2l transcripts were aberrantly upregulated in only mutant Prog-E (Figure 4B). It is possible that the Foxl2l might have feedback regulation in Prog-E. It is also possible progenitor transit from Prog-E to Prog-C might be defective.

Co-expression and increased expression of nanos2/id1 with foxl2l in GSC in foxl2l mutant is interest, as these two genes are expressed separately in WT. Can overexpress of nanos2 and id1 lead to loss of oocytes, or dedifferentiation of oocytes (Prog-C) back to PGCs, and female to male sex reversal?

It is possible that overexpressing nanos2 or id1 triggers the dedifferentiation of Prog-C back to PGCs, or prohibits the development of oocytes. However, in the foxl2l mutant, the expression of quite a few genes was altered. Therefore, overexpress/knockout only one or two DEGs might not be sufficient to replicate these foxl2l mutant phenotypes.

Guest Editor (suggestions for the authors)

The Hsu et al. manuscript presents valuable data that extend earlier single cell transcriptomic data on zebrafish gonads to help us better understand various initial stages of germ cell development in zebrafish and the role of Foxl2l in oogenesis. Data are well presented. Some of the conclusions do not seem to be well supported by the data, or alternatives are sometimes not considered.

The authors investigated development of zebrafish germ cells by isolating Piwil1-expressing cells and performing scRNA-seq in wild types at 26 days post fertilization, which augments published single cell transcriptomic studies of whole zebrafish gonads at 40 dpf, and they performed scRNA-seq experiments on mutants lacking foxl2l activity. These data, validated by in situ hybridization studies, allowed the identification of new stages of germ cell development. Studies showed that foxl2l mutants failed to develop meiotic oocytes, and that mutant germ cells reached a modified committed progenitor stage (Prog-C) but did not become late progenitors (Prog-L). The cells the authors called Prog-C cells in mutants, unlike wild types, expressed some GSC markers, suggesting that they maybe shouldn't be called Prog-C cells in mutants. From this result, the text concluded that committed progenitors then reverted to germ cell stem cells (GSC), which then directly developed into spermatogenic cells, and hence the animals became males, concluding that this was a direct development into males.

Alternative hypotheses exist for these data. The expression of genes normally expressed by GSCs (e.g., nanos2) in what otherwise seem to be committed progenitors was interpreted as a reversion of Prog-Cs to GSCs. An equally possible hypothesis is that the normally GSC-expressed genes never turned off and that the mutant 'Prog-C' never fully developed. Both explanations need to be presented and arguments assessed for and against each of these explanations.

Thanks for your comments. We have contemplated alternative explanations for our results. Regarding the possibility of continued GSC programs in the mutants, we examined the expression profile of GSC markers including nanos2, cbx7a, psmb13a and dnmt3bb.1 in mutant cells. All of them showed low expression in GSC-to-progenitor transition (G-P) stage and Prog-E stage in the mutant (Figure 3D, 3E). In DEG analysis, no GSC markers were upregulated in GSC-to-progenitor transition (G-P) or Prog-E. In Prog-C, however, nanos2 was abruptly upregulated. This result indicates that the GSC program was shut off during the differentiation from GSC to progenitor. It was re-expressed in the mutant at the Prog-C stage. Therefore, it is more likely that in the absence of functional Foxl2l, Prog-C acquires some GSC property with upregulation of nanos2.

Additionally, the conclusion that there was direct male differentiation in foxl2l mutants without going through the female phase has alternative explanations consistent with the data. This direct male development argument assumes that the Prog-C stage has not already embarked on a female pathway. The text did not demonstrate that Prog-C, or even the earlier stages like Prog-E, had not already begun to be female. The very name 'committed progenitor cells' suggests that these cells are already committed to the female pathway in wild types.

Our transcriptomic analysis showed that both male genes (eg. tekt1, odf3b) and genes related to female meiosis (e.g., rec8a, meiosin) were barely detected in Prog-E and Prog-C, suggesting that male or female pathway has not yet been initiated in Prog-E and Prog-C. Furthermore, in Prog-L, the female meiotic genes start to express highly, suggesting that Prog-C is the commitment stage toward either female differentiation or remaining indifferent, followed by male differentiation in the future. This is why the term “committed progenitor” is used.

These authors and others have shown that one of the earliest steps in female gonad development in zebrafish is an increase in the number of primordial germ cells compared to zebrafish that will become males, by 14 days post fertilization. The cell division marker analysis reported here suggests that those PGC proliferation steps are already occurring in the cells studied here. So do the 'Proliferation' arrows shown in Figure 6 for both the Prog-E and Prog-C stages. Thus, these stages would already be female-oriented germ cells. And thus, when the reversion of Prog-C towards Prog-C-to-GSC happens, what that really represents is a reversion of female pathway cells towards uncommitted GSCs that then embark on a male pathway, as happens in many zebrafish individuals due to stress or temperature or other features. Thus, the data presented here shows that the male pathway is taken after germ cells have embarked on the female pathway and turned on foxl2l, but subsequently thwarted from becoming a meiotic oocyte due to the lack of Foxl2l in the mutants.

The title (Single cell transcriptomes of zebrafish germline reveal stemness suppression and progenitor feminization by Foxl2l) should be revised because the conclusion that Foxl2l suppresses 'stemness' is not well supported. The data do show that foxl2l mutants have cell types that co-express GSC markers like nanos2 with genes usually expressed later in female germ cell development, like sycp1 and sycp3 (Figure S6), but that does not clearly mean that these mutant cells are in fact stem cells.

Thanks for your insightful comments. Only Prog-L expressed female meiotic genes, indicating these cells already differentiated towards females. The earlier progenitors, Prog-E and Prog-C, are indifferent progenitors because they lack transcripts of both male and female marker genes. Prog-C is the commitment stage deciding either to initiate female differentiation or remain indifferent, waiting for future male differentiation. Here in our mutant study, we found foxl2l is critical for the development of Prog-C toward Prog-L. Therefore, in the foxl2l mutant, Prog-C fails to develop into Prog-L and fails to undergo female differentiation. Taken together, we suggest that in the foxl2l mutant, female differentiation is fully blocked, and some of the Prog-C start to express the GSC marker nanos2 through an unknown mechanism. We agree to remove the term “suppression of stemness”.

The 'allmarkers' files are not included in the supplementary data, making it impossible to see how genes were selected for display. These files must be added to the supplementary data.

We have now added the files in our supplements.

Some suggestions for improving clarity in the text and figures.

P2 'factor, Foxl2l, is expressed in the progenitors committed to the ovary fate.'. Consider saying '…in the germ cell progenitors committed to the oocyte fate..'

P2 'Another single-cell profiling of foxl2l-/- germ cells reveals the arrest of early progenitors.'

Text here uses both present and past tenses. For describing the results of experiments done in the past, past tense might be better, as the text often uses in 'mediated mutation of foxl2l produced 100% male fish' and 'in stem cells) was elevated'. So maybe use 'revealed' instead of 'reveals'. And again could say '…early germ cell progenitors.' Just to be precise. Likewise, '…of nanos2 (germ cell stem cell)…'.

In Key Words, consider changing 'stem cell' to 'germ cell stem cell', and 'germ cell progenitor', which might better direct interested folks to this paper. Consider changing 'female' as a keyword, maybe it's too generic, to 'female-to-male sex change', which might get more focused hits.

P2 'increase of nanos2+ germ cell in foxl2l'. Use here plural, 'germ cells'…

P3 'contrast to higher vertebrates,'. Consider saying what specific taxon is meant here rather than 'higher'. Talking about higher and lower animals comes from a pre-scientific view of the scale of life.

P3 'system, but domesticated laboratory strains have no sex chromosome'. Use plural, 'chromosomes'.

P3 'the differentiation of zebrafish gonad. Zebrafish'. Plural, 'gonads'

P3 'of germ cells (nr0b1 mutants) or no germ cell (dnd morphans) develop'. Plural, and spelling, 'or no germ cells (dnd morphants

P3 'However, the developmental stages of germ cells in zebrafish remain unclear.'

I think the developmental stages have been clear, but the molecular genetics that moves cells through these stages is unclear, and the authors' work subdivides certain stages.

P3 'and initiates oogenesis in XX gonad Kikuchi'. Plural, 'XX gonads'

P3 'with foxl3 mutation become hermaphroditic possessing'. Several changes: 'with a foxl3 mutation become hermaphrodites possessing functional sperm'

P3 'and proved that Foxl2l controls the development of committed progenitor,'. Plural, 'progenitors'.

P4 'Germ cells play important roles in zebrafish sex differentiation, but their development is poorly studied.'

Maybe more accurate to say: 'Germ cells play important roles in zebrafish sex differentiation, but we still have much to learn about their development.'

Thanks for your suggestion. We have done the revision.

P4 '2002). Fluorescent germ cells were isolated from piwil1:EGFP transgenic fish by cell sorting'

The reader needs to know when piwil1 turns on in germ cell development with respect to the timing of oogonia vs. spermatogonia commitment. We need to know if by using the transgene we miss early but important stages in germ cell development.

According to the previous study (doi: 10.1002/dvdy.22404), in piwil1:EGFP transgenic fish, EGFP is expressed in all stages of germ cell in gonad at larva and juvenile stages. In ovary, EGFP is highly expressed in oogonia and stage I oocyte. In testis, EGFP can be detected in all stage of spermatogenic cells. Therefore, to obtain all germ cells for analysis, we chose piwil1:EGFP transgenic fish for germ cell sorting. We have now described the cells that can be labeled by piwil1:EGFP in the text.

Supplementary Figure S2 must have a mechanism for the reader to identify the genes in the rows. They would be too small to read if written on the figure so it should be in an Excel file in the supplement.

Thank you for the suggestion. To enhance accessibility and readability, we have included a detailed list of the genes represented in Figure 1—Figure Supplement 2 as an Excel file, titled ["Supplementary_File_Genes_FigureS2.xlsx"], within the supplementary materials. This provides an efficient means for readers to access the information without compromising the clarity of the figure.

P4. 'expression profiles of the top 37 out of a total of 1491 marker genes'

Top based on what criterion? P-adj? P? Log2 fold change? Most specific 2 or 3 per cluster? Or something else? And what defines a 'marker gene'?

Thanks for your suggestion. For each cluster, genes that satisfy the following criteria: (i) specificity ≥ 0.2, (ii) Q-value < 0.01, and (iii) having non-zero counts in ≥ 80% of cells in the cluster were selected as markers. Marker genes were ranked based on the specificity within clusters, and the top 2 to 4 genes in each cluster were selected as top markers. We have revised the Result and defined criteria of marker genes and top marker genes in the “Cluster marker identification” section within “Materials and methods”

P4 'markers (sycp1, sycp3, dmc1, buc and zp3e). Are buc, which helps assemble germplasm, and zp3e, an eggshell component, \involved in meiosis? Consider rephrasing.

Thanks for your suggestion. We have done the revision.

Suppl Figure 3 and 4. Although most of the markers show very similar patterns, thus cross-validating both datasets, it looks like for the markers CU467646.3, si:zfos-1505d6.3, and zp3e(zgc:171779), that the developmental pathway extended much further in the Liu data than the Hsu data. This isn't a huge concern, but it should be mentioned and discussed if there is a methodological difference that could explain this difference. Maybe it's just the age difference?

Yes, the difference is due to the age difference. Our data was obtained from 26-dpf gonad while Liu’s data was obtained from 40-dpf ovary. The 40-dpf ovary contains more developing oocytes than 26-dpf gonad. Therefore the expression pattern of diplotene markers, CU467646.3, si:zfos-1505d6.3, and zp3e(zgc:171779), were extended much further in 40-dpf dataset than in 26-dpf dataset.

P5 'Zebrafish progenitors are round in shape'. Progenitors of zebrafish oocytes…

The progenitor type of germ cell (including Prog-E, Prog-C and Prog-L), which is generally called gonocyte previously. The germ cells are called oocytes only after meiosis begins. We therefore prefer not to change our text here to avoid confusion.

P5 'et al., 2010). To characterize these three progenitor subpopulations,'

I'm not sure which three are referred to here. When adding in GSC, we get four. Please clarify.

The three progenitor subpopulations indicate Prog-E, Prog-C and Prog-L. Thanks for your suggestion. We have revised the text to clarify this point.

'et al., 2010. To characterize the GSC, Prog-E, Prog-C, and Prog-C(S) progenitor subpopulations,'

P5 'hybridization (FISH) and detected GSC as positive'. Make plural GSCs.

P5 Characterization of the three stages of progenitor subtypes using nanos2, foxl2l, and rec8 was useful.

P5 'premature stop codon disrupting the FH domain of Foxl2l'. Spell out 'forkhead'.

Thanks for your suggestion. We have done the revision.

Figures 2C and 2D. These would be more impactful if error bars were displayed on the graphs and numbers of trials or numbers of individuals were shown.

Thanks for your suggestion. We have revised figures 2C and 2D accordingly.

P5 'non-meiotic cystic germ cells were found in foxl2l mutant (Figure 2E).'

Plural: mutants

P5 'foxl2l mutant germ cells are arrested at committed progenitor.'. 'foxl2l mutant germ cells are arrested at the committed progenitor stage.'

Thanks for your suggestion. We have done the revision.

P5 '(Supplementary Figure S6A). The expression of the 37 top WT marker genes'. Again in a couple of words, explain the definition of 'top' genes more precisely.

For each cluster, genes that satisfy the following criteria: (i) specificity ≥ 0.2, (ii) Q-value < 0.01, and (iii) having non-zero counts in ≥ 80% of cells in the cluster were selected as markers. Marker genes were ranked based on the specificity within WT clusters, and the top 2 to 4 markers in the rank in each WT cluster were selected as top markers. We have defined the “37 top WT marker genes” more clearly in the paragraph “Cluster marker identification” within the section “Materials and methods”.

P5 'found that these marker genes were expressed at similar stages as they represented for WT cells.'. The text here needs to say that WT and mutant cells were co-clustered. Also, it would be interesting to see how the pseudotime navigates its way through this blob of cells.

Thank you for your feedback. The mutant analysis was performed exclusively on mutant cells and did not involve co-clustering with WT cells. The UMAPs presented in Figure 3—Figure Supplement 1A (original Figure S6B) are derived solely from mutant cells and underscore this focus. We have clarified this distinction in the revised main text to ensure accuracy and transparency.

Figure 3D. Label the horizontal axis as cluster number. Also, consider changing 'dominant' in the legend to 'predominant'.

P6 'integrated groups (Figure 3D). Dominant stages were obtained'. Define' Dominant stages'.

Thanks for your suggestion. The term “dominant” may not be precise. We have now changed it to “prominent” as suggested. The prominent stage of an integrated group was defined as the stage with a significantly higher proportion of cells in the group among the entire WT or mutant population. The “prominent stages” were defined in the section “Composition analysis” within Materials and methods.

P6 'calibration against their prevalence rates in the entire cell population.'

Does 'entire cell population' mean all of the mutant cells? or all of the mutant plus WT cells?

Thanks for your suggestion. The “entire cell population” indicates either entire WT cells or mutant cells. We have done the revision.

P6 'For mutant cells, while groups I5-I8 had two dominant stages,'

If dominant means the one that's present most frequently, then more than one shouldn't be dominant unless there's a tie. So 'dominant' needs a clear definition.

Dominant” might not be accurate when more than one stages were mentioned. We have changed it to “prominent” stages.

P6 'that WT and mutant cells with high cell-to-cell similarity belong to similar developmental stage.'

'that WT and mutant cells that have high transcriptome similarity likely belong to the same developmental stage.'

P6 'Groups I3 and I6 are particularly unique.'

'Groups I3 and I6 are unique.' Unique means the only one of its kind. So there can't be gradations of being unique.

Figure 3E. Should the group at the bottom labeled 'Mutant', be called something like 'Mutant-specific cell types'? Because all of the data in the table refer to mutant cells, right?

Thanks for your suggestion. We have changed the term.

P6 'mutant cells additionally expressed zgc:158463'

In addition to what? So mutant cells expressed all the genes of WTs in I3 but in addition zgc:158463 which wasn't expressed in the WT cells in I3? The same question for I6 and tp53. Please clarify.

We have revised the text to clarify the statement.

P6 'Furthermore, the Prog-C marker foxl2l was expressed in mutant cells of I1,'

So that means that the mutant mRNA had normal stability? Wasn't degraded by nonsense mediated decay?

Yes, the mutant mRNA had normal stability. It was not degraded by nonsense-mediated decay.

P6 'The aberrant foxl2l expression in some earlier GSC and Prog-E plus the accumulation of foxl2l expressing cells in the mutant indicate that progenitor development is impaired in foxl2l mutants.'

'in the mutants and the failure to find clusters representing more mature stages indicate that progenitor development is impaired in foxl2l mutants' [could that be added or not?]

P7 'throughout the development'. Change to: 'throughout development'

P7 'Stratified by integrated groups, we compared'. Check to see if there's a better word than 'stratified' to use here. As written, it means that 'we', the authors, were arranged into layers by integrated groups.

Thanks for your suggestion. We have done the revision.

P7 'WT and mutant cells at the same dominant stages'. Each time that 'dominant' is used, see if 'predominant' would be a better word.

Thanks for your suggestion. To make a more precise statement, we change the term “dominant” into “prominent”

Figure 4A. The triangles are very difficult to see. Would some other visualization work better?

Thank you for the feedback. We have revised Figures 3C and 4A so that they are easier to read.

P7 'Figure 4B shows genes that were differentially expressed between WT and mutants.'

Are these all of the genes that were differentially expressed between these two genotypes? Or a subset? And if a subset, then how were these specific genes selected to display?

In addition, be explicit about the comparison. Does red mean that the WT or the mutant is overexpressing? E.g., is Dmrt1 overexpressed in WT or in mutant?

Figure 4B shows all differentially expressed genes. The fold changes were calculated by mutant/WT. The red color means the mutant is overexpressing.

Figure 4B. The gradient in the main portion of the figure should match the order shown in the scale. The darkest red indeed is on top, but then for the blues, the darkest is on top, not at the bottom as in the scale.

The genes in the heatmap were ordered by the fold change. In each matched stage, the differentially expressed genes (DEGs) are ordered by the absolute value of the fold change within the set of upregulated or downregulated genes, with the most significantly different genes positioned at the top of each matched stage. This ensures that both positively and negatively regulated genes are prominently displayed. The gradient order reflects this logic, where the darkest red indicates the largest positive fold change and the darkest blue indicates the largest negative fold change. We chose this representation to emphasize the most differentially expressed genes, regardless of the direction of change.

Figure 4c. If id1 and nanos2 are marker genes for GSCs and disappear at later stages in WTs, and if their expression doesn't disappear in mutants, then the mutant cells don't advance to the stage where expression of these genes normally disappear, right? So there wouldn't be any mutant cells at stages G-P to Prog-C(S) because id1 and nanos2 are still expressed in mutant oocytes in these columns in the figure? Or at least these would be variants of the wild type clusters.

nanos2 is a GSC marker gene while id1 is not. Id1 was barely detected in all types of germ cells in WT juvenile gonad. In the mutant, nanos2 was only expressed in mutant GSC, became barely detected at G-P and Prog-E stages, then was highly expressed at Prog-C stage. The discontinuous expression of nanos2 in mutant cells during developmental stages suggests that in the mutant, nanos2 is re-expressed in Prog-C without going through the steps of G-P and Prog-E. We still detected mutant cells that correspond to G-P and Prog-E as shown in Figure 3D.

In addition, it looks like foxl2l inhibits its own expression at early stages, because without foxl2l activity, the expression of foxl2l is much higher in mutants and WTs in GSCI to Prog-E columns.

The foxl2l transcripts were aberrantly upregulated in only mutant Prog-E. The expression of foxl2l transcripts in mutant GSCI is not significantly different from that in WT, and transcripts in mutant GSCII and P-G stage were really low. It is possible that Foxl2l may inhibit its own expression in Prog-E although we did not investigate this point further.

P7 'The expression of Dmrt1 and dele1 in WT was steady throughout the developmental stages (Figure 4C), but Dmrt1 was abnormally high and dele1 was abnormally low in both mutant GSCI and Prog-C/Prog-C(S).'

The fact that Dmrt1 was abnormally high in mutants suggests that a role of foxl2l activity is to inhibit expression of Dmrt1.

The fact that dele1 was low in mutant Prog-C(S) cells in mutants, suggests that foxl2l activity encourages expression of dele1.

Those conclusions, if authors agree with them, might have more impact here than what's written: 'These results indicate multiple defects of gene expression in the mutants.'

P7 'Co-expression of foxl2l with Dmrt1, nanos2 or id1 in foxl2l mutant'. Consider changing to '…in foxl2l mutants', or to '…in foxl2l mutant germ cells'.

Figure 5A legend. Say explicitly that cells expressing both genes are represented by yellow. It's obvious of course from the scale grid but would help some readers to say it explicitly.

P7 'increased in the mutant (Figure 5B). Moreover, the number of nanos2+ cell'

Plural: …of nanos2+ cells…'

P7 '(Figure 5C). To further investigate whether GSC is affected in mutants' '…whether GSCs are affected…'

Thanks for your suggestion. We have done the revision.

P7 'counted the number of nanos2+foxl2l- cells, and found that the number increased dramatically in mutants at 26 dpf'. increased dramatically in mutants compared to WTs? Or increased in mutants at 26 dpf compared to some other time in development?

Thanks for your suggestion, we have done the revision (compared with WT).

P7 'GSC cells in mutant together'. Plural: 'GSC cells in mutants together'

Thanks for your suggestion, we have done the revision.

P7 'GSC cells in mutant together suggest that some Prog-C may revert to GSC in', What is the evidence that they develop to Prog-C and revert rather than that they never made it fully to Prog-C and simply maintained some of the gene expression properties of the earlier stages? These are two competing hypotheses, and both should be mentioned.

We have examined the expression profiles of GSC markers including nanos2, cbx7a, psmb13a and dnmt3bb.1 in mutant cells. We found that the expression of these genes was low in mutants at GSC-to-progenitor transition stage and Prog-E stage. Their expression remained low after cells are differentiated into progenitors, and only nanos2 was abruptly upregulated in mutant Prog-C. Therefore, it is more likely that the mutant Prog-C acquires some GSC property with the upregulation of nanos2 instead of the maintenance of GSC property during the development from GSC to progenitor. Nevertheless, we followed the suggestion of this reviewer and mentioned both hypotheses in this section.

Figure 6 legend. Mention 'G-P transition' in the legend or remove it from the sketch. The word 'Proliferation' in black on a dark gray field is hard to read.

Thanks for your suggestion. We have revised the figure and the figure legend.

P9 'they undergo G to P transition for further differentiation into progenitors.'

Define 'G-to-P'.

Thanks for your suggestion. We have revised the sentence.

P8 'Progenitors include three types: Prog-E, Prog-C, and Prog-L.' It would be good if the text here could state the transcriptomic features that distinguish these three phases from each other like the text did for the two kinds of GSCs.

Thanks for your suggestion. We have done the revision to include transcriptomic features of these cell types.

P9 'The fate of germ cells is determined following the instruction of the supporting cells.' Compare to germ cell sex determination in mammals.

Thanks for your suggestion, we have done revision to describe the germ cell sex determination in mammals.

P9 'deficiency, however, leads to cell arrest at Prog-C as shown in this article.'. 'deficiency, however, leads to cell arrest at Prog-C as shown by results presented here.

Thanks for your suggestion, we have done the revision

P9 'commitment to the female fate as shown here. It indicates'. What is the antecedent of 'It'?

We have rearranged this paragraph to avoid ambiguity in our statement.

P10 'Id1, inhibitor of DNA binding 1, is a transcription regulator that lacks DNA…'.'…that lacks a DNA …'

Thanks for your suggestion, we have done the revision.

P10 'stem-cell genes (id1 and nanos2) and male gene Dmrt1.' First, '…and the male gene Dmrt1'

Second, are there other early male genes that are suppressed by Foxl2l action besides Dmrt1?

Thanks for your suggestion. Among the total 43 DEGs shown in Figure 4B, Dmrt1 is the only gene known for male differentiation in zebrafish and mammals.

P10 'The alternative development of foxl2l mutant germ cell'. Plural: '…mutant germ cells'

Thanks for your suggestion, we have done the revision.

Associated Data

    This section collects any data citations, data availability statements, or supplementary materials included in this article.

    Data Citations

    1. Chung BC. 2025. Single cell transcriptomes of zebrafish germline reveal progenitor types and feminization by Foxl2l. NCBI Gene Expression Omnibus. GSE173718 [DOI] [PMC free article] [PubMed]
    2. Ho H, Hsu C, Yang CH, Wang Y, Li KC, Chung B. 2025. Codes of "Single-cell transcriptomes of zebrafish germline reveal stemness suppression and progenitor feminization by Foxl2l". figshare. [DOI] [PMC free article] [PubMed]
    3. Liu Y. 2021. Single-cell transcriptome reveals insights into the development and function of the zebrafish ovary. NCBI Gene Expression Omnibus. GSE191137 [DOI] [PMC free article] [PubMed]

    Supplementary Materials

    MDAR checklist
    elife-100204-mdarchecklist1.docx (89.6KB, docx)
    Supplementary file 1. The list and the expression of marker genes in each cluster.
    elife-100204-supp1.xlsx (782KB, xlsx)

    Data Availability Statement

    Data for single-cell transcriptomic data analysis are available with NCBI GEO (https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE173718). Analysis code is available at Figshare (https://doi.org/10.6084/m9.figshare.26314126.v1).

    The following datasets were generated:

    Chung BC. 2025. Single cell transcriptomes of zebrafish germline reveal progenitor types and feminization by Foxl2l. NCBI Gene Expression Omnibus. GSE173718

    Ho H, Hsu C, Yang CH, Wang Y, Li KC, Chung B. 2025. Codes of "Single-cell transcriptomes of zebrafish germline reveal stemness suppression and progenitor feminization by Foxl2l". figshare.

    The following previously published dataset was used:

    Liu Y. 2021. Single-cell transcriptome reveals insights into the development and function of the zebrafish ovary. NCBI Gene Expression Omnibus. GSE191137

    Articles from eLife are provided here courtesy of eLife Sciences Publications, Ltd

    RESOURCES