MEIOSIN retains a STRA8-independent activity that contributes to meiotic gene activation across vertebrates

In this study, Shimada et al. show that the HMG-domain containing MEIOSIN can function independently of STRA8, and has a core function in meiotic gene activation and meiotic entry. The bHLH-dependent MEIOSIN-STRA8 interaction further reinforces the meiotic transcriptional program and facilitates meiotic progression.

Abstract

Meiotic entry in vertebrates has been viewed as a transcriptional switch driven by the MEIOSIN–STRA8 axis, but whether this represents the ancestral regulatory logic of meiosis has remained unclear. Here we combine comparative genomics with genetic and single-cell analyses in zebrafish and mice to show that MEIOSIN retains an intrinsic STRA8-independent activity. We identify Meiosin orthologs in zebrafish and hagfish, vertebrate lineages that lack Stra8, and show that these proteins retain the HMG domain but have lost the bHLH domain required for the canonical MEIOSIN–STRA8 interaction. In zebrafish, meiosin is transiently induced at meiotic entry in both sexes, yet its loss selectively disrupts the female germline, causing failure of oogenesis and female-to-male sex reversal. In mice, MEIOSIN lacking the bHLH domain still initiates key features of meiotic entry and partially activates meiotic target genes, although it fails to support full meiotic progression. Together, these findings support a model in which HMG-containing MEIOSIN provides a conserved core activity for meiotic gene activation, whereas STRA8 and the bHLH-mediated MEIOSIN–STRA8 interaction reinforce the efficiency and robustness of this program in lineages that retain the canonical module.

Meiosis is the defining developmental transition that enables gamete formation and sexual reproduction. In vertebrates, this transition is triggered by a dedicated transcriptional program that activates meiotic genes as germ cells exit the mitotic cycle and commit to meiotic prophase. In mice, this switch depends on the transient co-expression of Stimulated by Retinoic Acid 8 (STRA8) and Meiosis Initiator (MEIOSIN), which together activate a broad set of meiotic genes in both male and female germ cells (Oulad-Abdelghani et al. 1996; Menke et al. 2003; Baltus et al. 2006; Anderson et al. 2008; Ishiguro et al. 2020; Feng et al. 2021). Both proteins contain a basic helix–loop–helix (bHLH) domain and an HMG or HMG-like domain, and together form a central regulatory axis for meiotic entry in mammals (Baltus et al. 2006; Kojima et al. 2019; Ishiguro et al. 2020). However, whether this two-factor module represents a universally conserved vertebrate mechanism—or a derived configuration built upon an older core system—has remained unresolved.

This question is particularly important because STRA8 appears to be absent from multiple fish lineages, including zebrafish, despite the continued occurrence of meiosis in these species (Rodríguez-Marí et al. 2013; Pasquier et al. 2016). Although Meiosin orthologs have been reported in several vertebrates, their distribution, domain architecture, and functional dependence on STRA8 remain poorly defined outside mammals. This leaves two major gaps in our understanding: when the MEIOSIN–STRA8 axis emerged during vertebrate evolution, and how meiotic entry is executed in lineages that lack STRA8. Here we address these questions through a vertebrate-wide comparative analysis of Meiosin and Stra8 orthologs, coupled with functional dissection in zebrafish and mice. We show that HMG-containing MEIOSIN exists in zebrafish and hagfish despite the absence of STRA8, that zebrafish meiosin is required for female but not overtly for male fertility, and that mouse MEIOSIN retains a STRA8-independent, bHLH-independent capacity to initiate meiotic gene activation. Together, these findings identify MEIOSIN as an evolutionarily conserved core regulator of meiotic entry and position STRA8 as a later reinforcing component of the vertebrate meiotic initiation program.

In addition to the shared mechanism in both sexes, meiotic entry in mice is further shaped by sex-specific regulation. STRA8 binds Retinoblastoma family proteins (RB1 and p107) through its LXCXE motif independently of MEIOSIN (Shimada et al. 2023). In female germ cells, this interaction sequesters RB proteins from E2F and derepresses S-phase-associated genes, coupling premeiotic DNA replication to meiotic gene activation. Thus, the STRA8–RB and STRA8–MEIOSIN interactions together ensure the timing and fidelity of meiotic entry in oocytes.

MEIOSIN orthologs containing both bHLH and HMG domains have been identified in mammals, birds, reptiles, amphibians, and several fishes (Ishiguro et al. 2020). In contrast, STRA8 orthologs are present in some fish species, such as Southern catfish and Atlantic salmon, but absent from many teleost genomes (Dong et al. 2013; Li et al. 2016; Skaftnesmo et al. 2021). In salmon, stra8 disruption increases germ cell apoptosis, yet spermatogenesis can still proceed, implying the existence of compensatory mechanisms (Skaftnesmo et al. 2021). These observations raise the possibility that the canonical mammalian MEIOSIN–STRA8 module is not the only vertebrate solution for meiotic entry.

Results

Meiosin orthologs retained by vertebrate species that lack Stra8

In mice, MEIOSIN contains both a bHLH domain and an HMG domain (Ishiguro et al. 2020). Our TBLASTN searches using the mouse HMG domain identified MEIOSIN orthologs in coelacanth and other vertebrates, and subsequent cDNA cloning confirmed a Meiosin ortholog in cloudy catshark (Fig. 1A), that is co-expressed with Stra8 in the dorsal portion of adult testis (Supplemental Fig. S1A,B). These data support broad conservation of HMG-containing MEIOSIN across vertebrates (Supplemental Fig. S2A). STRA8 is also predicted to contain a bHLH domain and an HMG-like domain (Baltus et al. 2006; Kojima et al. 2019), and STRA8 orthologs are broadly conserved in vertebrates (Supplemental Fig. S2B; Pasquier et al. 2016). However, STRA8 orthologs have not been identified in several fish lineages, including zebrafish and hagfish (Fig. 1A; Rodríguez-Marí et al. 2013; Pasquier et al. 2016). We therefore asked whether Meiosin orthologs are retained in these lineages despite the apparent loss of Stra8.

Figure 1.

MEIOSIN orthologs identified in zebrafish and hagfish that lack STRA8. (A) Distribution of Stra8 and Meiosin orthologs mapped onto the vertebrate phylogeny. An asterisk (*) in the bHLH column indicates the presence of a bHLH domain in Meiosin. The phylogeny was generated in R using the Open Tree of Life database (Sanchez Reyes et al. 2021). (B) Domain organization of MEIOSIN orthologs from mouse, hagfish, and zebrafish. Predicted ribbon models generated by AlphaFold3 are shown below. Helices are colored as indicated in the schematics above. (C) Superimposed structures of the HMG box from mouse (green), zebrafish (magenta), and hagfish (cyan) MEIOSIN orthologs. Numbers indicate amino acid positions in each species.

We identified a potential inshore hagfish Meiosin ortholog (ENSEBUG00000016420.1) encoding an HMG-containing protein on the genomic contig FYBX02009704.1 (Yu et al. 2024). Sequence alignment and phylogenetic analysis supported its relationship to jawed vertebrate Meiosin genes (Supplemental Fig. S2C). This hagfish Meiosin was specifically expressed in adult testis, similar to Sycp3, but not in other examined tissues (Supplemental Fig. S3A).

We next identified a candidate zebrafish ortholog at the si:ch211-191j22.7 locus on chromosome 21 (Supplemental Fig. S3B). Because this genomic region contained ambiguous sequence information, we confirmed its transcript sequence by de novo assembly of public zebrafish testis RNA-seq data (SRR5378555) followed by RT–PCR and 5′ RACE (Supplemental Fig. S3B,C). This yielded a full-length cDNA encoding an HMG-containing protein, confirming that the si:ch211-191j22.7 locus corresponds to zebrafish meiosin (Supplemental Fig. S2A). Southern blotting showed that zebrafish meiosin is present as a single-copy gene despite the teleost genome duplication (Supplemental Fig. S3D). RT–PCR detected expression specifically in adult testis, but not in other adult tissues examined (Supplemental Fig. S3E), indicating that zebrafish meiosin is germ cell-enriched, as in mice.

Notably, both zebrafish and hagfish MEIOSIN proteins retain a conserved HMG domain but lack an obvious bHLH domain (Fig. 1B; Supplemental Fig. S2A). AlphaFold3 predictions likewise identified the HMG domain but not a bHLH domain in either protein (Fig. 1B). In contrast, the predicted three-dimensional structure of the HMG domain was highly conserved among mouse, hagfish, and zebrafish MEIOSIN proteins (Fig. 1C). Thus, MEIOSIN is retained in zebrafish and hagfish, two vertebrate lineages that lack STRA8, and in both cases the protein has lost the bHLH domain while preserving the HMG domain (Fig. 1A). Phylogenetic relationships suggest that bHLH loss occurred independently in these lineages rather than being secondarily gained elsewhere (Fig. 1A). These findings raised two questions: whether zebrafish meiosin still contributes to meiotic entry without a bHLH domain, and how it functions in the absence of STRA8. Together, these findings identify previously unrecognized Meiosin orthologs in zebrafish and hagfish and show that loss of the bHLH domain correlates with the absence of STRA8 in these lineages.

Zebrafish meiosin marks meiotic entry in both sexes

To define the germ cell populations that express zebrafish meiosin (zf meiosin), we performed single-cell RNA sequencing (scRNA-seq) on adult WT testes (Fig. 2A–C; Supplemental Fig. S4). UMAP analysis resolved 14 transcriptionally distinct germ cell clusters (Fig. 2A; Supplemental Fig. S4C; Supplemental Data 1). Marker expression identified cluster 8 as spermatogonial stem cells (SSCs) as suggested by the high expression id4, nanos2 and dnd1, clusters 3/5/9/6 as meiotic prophase cells, and clusters 2/10/13/7/11 as spermatids (Fig. 2B,C). Based on these markers (Qian et al. 2022; Sposato et al. 2024; Burgos-Ruiz et al. 2026), we inferred a developmental trajectory from clusters 8 to 11.

Figure 2.

scRNA-seq analysis of male and female germ cells in zebrafish. (A) UMAP clustering of scRNA-seq transcriptomes from testicular germ cells isolated from adult WT zebrafish. Arrow indicates the inferred developmental trajectory. See also Supplemental Figure S4 for the UMAP of whole testicular cells. Germ cells were separated from somatic cells (encompassing Sertoli cells, Leydig cells, and hemocytes) using established marker genes. (B) UMAP plots showing mRNA abundance of meiosin and key marker genes across spermatogenic germ cells. Marker genes: id4, dnd1, and nanos2, spermatogonial stem cells; moto, spermatogonia and meiotic-entry cells; rec8b and dmc1, meiotic prophase spermatocytes; odf3b, spermatids. (C) Dot plot showing average scaled expression (color gradient) and the fraction of cells with detectable expression (dot size) for key developmental marker genes. (D) Serial paraffin sections of adult testis stained with hematoxylin and eosin (HE), anti-Meiosin detected with an AP-conjugated secondary antibody, anti-Sycp3 detected with an Alexa Fluor 488-conjugated secondary antibody, or control rabbit IgG detected with an AP-conjugated secondary antibody. Yellow and red dashed lines indicate seminiferous tubules containing preleptotene–leptotene (PL-L), leptotene/zygotene/pachytene-stage spermatocytes with extended Sycp3-positive chromosome axes (L-P) spermatocysts, respectively. Scale bar, 10 µm. (E) Adult testis sections stained for Sycp3 and Meiosin. Dashed lines indicate spermatocysts at the indicated meiotic stages as in D. Enlarged images of a seminiferous tubule (L-P) are shown on the bottom. Scale bar, 10 µm. (F) UMAP clustering of scRNA-seq transcriptomes from ovarian germ cells isolated from Tg(vas::EGFP) females at 54 and 85 dpf. Arrow indicates the inferred developmental trajectory. See also Supplemental Figure S5 for the UMAP of whole ovarian cells. Germ cells were separated from somatic cells (encompassing Follicle cells, Theca cells, Stromal cells and hemocytes) using established marker genes. (G) UMAP plots showing mRNA abundance of meiosin and key marker genes across ovarian germ cells. Marker genes: id4, stem cells; foxl2l, oocyte progenitors; meioc (moto), rec8a, rec8b, and dmc1, meiotic-entry/prophase oocytes; zp3b, mature oocytes. (H) Dot plot showing average scaled expression (color gradient) and the fraction of cells with detectable expression (dot size) for the marker genes shown in G. (I) Serial sections of 42 days postfertilization (dpf) juvenile ovary stained for Foxl2l (red), Meiosin (green), and DAPI. Arrowheads indicate zf Meiosin-positive cells; arrows indicate strongly Foxl2l-positive cells. Scale bar, 10 µm. (J) Overlap between genes enriched in meiosin-positive male germ cell clusters (clusters 12 and 1) and female germ cell clusters (clusters 4 and 5). Gene ontology enrichment analysis of the shared gene set is shown below. See also Supplemental Data 3.

The zebrafish ortholog of mouse Meioc (moto) (Bowen et al. 2012; Abby et al. 2016; Soh et al. 2017; Pfaltzgraff et al. 2024; Kawasaki et al. 2025), enriched in clusters 0, 12, and 1, spanning the transition from SSCs to early meiotic prophase. Zf meiosin was specifically induced in clusters 12 and 1, placing its expression in a narrow window at meiotic initiation. Zf meiosin was specifically induced in clusters 12 and 1, placing its expression in a narrow window at meiotic initiation. Consistent with the scRNA-seq data, immunostaining detected zf Meiosin-positive cells among spermatocytes assigned to the preleptotene/leptotene-transition region (Fig. 2D,E). Thus, in the male germline, zf meiosin marks the mitosis-to-meiosis transition.

We next analyzed ovarian germ cells from Tg(vas::EGFP) females at 54 and 85 dpf (Fig. 2F–H; Supplemental Fig. S5). UMAP analysis identified 12 ovarian germ cell clusters (Fig. 2F; Supplemental Fig. S5C; Supplemental Data 2). Most clusters corresponded to maturing oocytes, but cluster 4 expressed the GSC marker id4 together with the oocyte progenitor marker foxl2l (Nishimura et al. 2015; Liu et al. 2022; Ren et al. 2024), suggesting that it contains GSCs and early oocyte progenitors. Cluster 5 appeared to represent a mixed population of late progenitors and early meiotic cells, whereas clusters 0 and 7 corresponded to meiotic prophase cells (Fig. 2G,H).

Unlike the male germline, where rec8b predominated, both rec8a and rec8b were expressed in early female clusters, indicating activation before and during early meiotic prophase. Strikingly, zf meiosin was transiently upregulated in clusters 4 and 5, coincident with meioc (moto) and overlapping with id4 and foxl2l.

To further validate the expression relationship among zf meiosin, id4, and foxl2l in female GSC/progenitor populations, we also analyzed a publicly available scRNA-seq data set from 40 dpf zebrafish ovaries (Liu et al. 2022). In this data set, zf meiosin mRNA was detected in an id4-positive population and in cells associated with, or following, foxl2l expression. Notably, a subset of zf meiosin- and id4-positive cells did not express foxl2l (Supplemental Fig. S6). These observations suggest that zf meiosin is induced within the GSC/progenitor compartment and near the transition toward meiotic/oogenic differentiation. Consistent with this interpretation, immunostaining showed that zf Meiosin appeared preferentially in cells with weak Foxl2l expression, whereas its signal was reduced in strongly Foxl2l-positive cells (Fig. 2I), further supporting zf Meiosin induction around the transition out of the progenitor state.

Genes enriched in zf meiosin-positive clusters from both sexes were associated with ribosome biogenesis and translation (Fig. 2J; Supplemental Data 3). Because failure of meiotic progression in zebrafish oogenesis causes female-to-male sex reversal (Takemoto et al. 2020; Blokhina et al. 2021; Imai et al. 2021; Olaya et al. 2025), the early female-specific transcriptional state accompanying meiosin expression may help couple meiotic entry to commitment to the female pathway.

Thus, zebrafish meiosin is transiently induced at the mitosis-to-meiosis transition in both sexes, with female expression initiating at an unusually early stage in the GSC/progenitor compartment.

Zebrafish Meiosin is essential for the female germline

To test the function of zebrafish Meiosin, we generated a CRISPR–Cas9 frameshift allele by targeting exon 4 and isolating an 11 bp net insertion mutant (Fig. 3A,B). Immunostaining confirmed complete loss of zf Meiosin signal in homozygous mutants, hereafter referred to as meiosin KO (Fig. 3C).

Figure 3.

Phenotypes of meiosin knockout zebrafish. (A) Structure of the zebrafish meiosin locus. The protein-coding sequence is shown in green. The sgRNA target site is indicated by a red bar. (B) Frameshift mutation allele of meiosin generated by CRISPR–Cas9. A 2 bp region was replaced by a 13 bp insertion, resulting in a net insertion of 11 bp (meiosin⁺¹¹). (C) Adult testis sections from WT and meiosin KO fish stained for Sycp3 (green), zf Meiosin (magenta), and DAPI (blue). (Sz) spermatozoa. Scale bar, 50 µm. (D) Sex ratios of offspring derived from intercrosses of meiosin heterozygotes. Numbers within bars indicate the number of individuals analyzed. (E) Fertility of WT (meiosin^+/+), heterozygous (meiosin^+/−), and knockout (meiosin^−/−) males. Each male was paired with a WT female. (F) Survival rates of embryos derived from WT (meiosin^+/+), heterozygous (meiosin^+/−), and knockout (meiosin^−/−) males. The percentages of embryos that survived and developed normally at 5 dpf after the crosses in E are shown for each paternal genotype. (G) HE-stained sections of juvenile gonads from WT and meiosin KO zebrafish at 35 dpf. Gonads are outlined by yellow dashed lines. Red arrows indicate stage IB oocytes larger than 20 µm in diameter. Numbers represent the proportion of individuals exhibiting the phenotypes shown. Scale bar, 20 µm. (H) Juvenile gonad sections from WT and meiosin KO fish (27 dpf) stained for Sycp3 or Sycp1 (green), together with Ziwi (magenta) and DAPI (blue). White arrows indicate meiosin KO oocytes showing synaptonemal complex formation. Numbers represent the proportion of individuals positive for Sycp3 or Sycp1. Scale bar, 10 µm. (I) Embryos derived from paired matings between meiosin heterozygous females and WT males are shown. Representative embryos exhibiting normal (upper panels) and abnormal (middle panels) development at each time point are presented. Normal embryogenesis at 1.5, 2.5, 5, and 7 h postfertilization (hpf) corresponds to the 8 cell, 256 cell, 30% epiboly, and 75% epiboly stages, respectively. Abnormal development is distinguishable from unfertilized embryos (bottom panel) at 5 hpf. Numbers in the inset indicate 5 hpf embryos in each category, pooled from two independent paired matings. At 5 hpf, 78/110 embryos showed normal development, 24/110 showed abnormal development, and 8/110 were unfertilized eggs. Thus, ∼24% of the fertilized embryos showed abnormal development detectable at 5 hpf. Scale bars, 100 µm.

In laboratory zebrafish strains, phenotypic sex is not determined by a simple sex-chromosome genotype (Aharon and Marlow 2022). Instead, gonadal sex fate is strongly influenced by germ cell number, and mutations that reduce or eliminate germ cells can result in sex reversal (Slanchev et al. 2005; Siegfried and Nüsslein-Volhard 2008). Juvenile gonads initially form early oocytes, and the subsequent maintenance or loss of these oocytes is closely associated with ovarian versus testicular development. Accordingly, mutations that impair early oocyte development or meiotic progression can lead to female-to-male sex reversal (Rodríguez-Marí et al. 2010; Takemoto et al. 2020; Blokhina et al. 2021; Imai et al. 2021; Olaya et al. 2025). This developmental context is important for interpreting the meiosin mutant phenotype, because defects in meiotic initiation, early oocyte progression, or maintenance of the female pathway can all ultimately result in loss of ovarian development.

All meiosin KO offspring derived from heterozygous intercrosses developed as males, whereas WT and heterozygous siblings developed as both sexes (Fig. 3D). This fully penetrant sex-reversal phenotype indicates that zf Meiosin is required for female germline development and/or maintenance of the ovarian pathway. In contrast, meiosin KO males produced spermatozoa and displayed fertility comparable to WT males under standard laboratory conditions (Fig. 3C,E), indicating that zf Meiosin is not overtly required for male fertility in this setting. Furthermore, embryos produced by meiosin mutant males crossed to WT females developed normally under standard laboratory conditions (Fig. 3F).

Histological analysis revealed no obvious abnormality in adult mutant males. In juvenile gonads, however, stage IB oocytes were absent in meiosin KO animals at 28–35 dpf (Fig. 3G), indicating a failure before or during the transition to the follicular stage, when oocytes normally enter and remain in dictyate arrest (Selman et al. 1993). Notably, a subset of Sycp3-positive oocytes showed Sycp1-positive synaptonemal complex formation (Fig. 3H), suggesting that mutant oocytes retain a limited capacity to acquire early meiotic-prophase features, although they fail to develop further. Thus, zf Meiosin is not absolutely required for the initial acquisition of early meiotic-prophase features, but is required for efficient progression or stabilization of early oocytes beyond this initial stage. We therefore interpret the complete female-to-male sex reversal as a likely consequence of impaired oocyte development and/or insufficient maintenance of the early female germline program, as observed in mutants with defective meiotic progression, rather than as direct evidence that Meiosin acts as a primary sex-determination factor (Rodríguez-Marí et al. 2010; Takemoto et al. 2020; Blokhina et al. 2021; Imai et al. 2021; Olaya et al. 2025). In addition, early embryos derived from meiosin heterozygous females crossed with WT males frequently exhibited abnormal embryogenesis associated with unequal cleavage (Fig. 3I). Approximately 24% of fertilized embryos showed abnormal development detectable at 5 hpf, suggesting that Meiosin function may be dosage-sensitive in the female germline. Together, these findings indicate that although zf Meiosin is induced at the mitosis-to-meiosis transition in both sexes, it is not absolutely required for spermatogenesis under standard laboratory conditions. We therefore conclude that zf Meiosin is associated with meiotic entry in both sexes, but is genetically essential for female oocyte progression and maintenance of the ovarian pathway.

The MEIOSIN–STRA8 interaction depends on bHLH domains

The zebrafish phenotype indicates that a MEIOSIN ortholog can be retained and function in a vertebrate lineage lacking both STRA8 and the MEIOSIN bHLH domain. However, this observation alone does not establish that all vertebrate MEIOSIN orthologs have equivalent STRA8-independent functions. We therefore next examined the role of the bHLH domain in the canonical mammalian MEIOSIN–STRA8 system.

To define the role of the bHLH domains in the canonical mammalian system, we examined the interaction between mouse MEIOSIN and STRA8. Mouse MEIOSIN contains an N-terminal bHLH domain and a C-terminal HMG domain (Ishiguro et al. 2020), whereas STRA8 contains a putative bHLH domain (Baltus et al. 2006), a glutamate-rich region, an HMG-like domain, and an LXCXE motif required for RB binding (Supplemental Fig. S7A; Kojima et al. 2019; Shimada et al. 2023).

Because bHLH domains frequently mediate protein–protein interactions, and because STRA8 nuclear localization depends in part on MEIOSIN in preleptotene spermatocytes (Ishiguro et al. 2020), we tested whether the two proteins interact through their bHLH regions. In 293T cells, anti-FLAG immunoprecipitation of STRA8-3×FLAG-HA recovered WT MEIOSIN, but not MEIOSIN lacking the bHLH domain (Supplemental Fig. S7A). Conversely, STRA8 lacking the bHLH domain failed to recover MEIOSIN. In contrast, deletion of the MEIOSIN HMG domain did not abolish the interaction. These results identify the bHLH domains as the principal determinants of the MEIOSIN–STRA8 interaction.

To define the physiological significance of this domain, we generated a Meiosin-ΔbHLH mouse line lacking the MEIOSIN bHLH region (Supplemental Fig. S7B). Meiosin-ΔbHLH males developed smaller testes than controls and lacked postmeiotic spermatids and sperm, closely resembling Meiosin-KO testes (Supplemental Fig. S7C). Female mutants showed ovarian degeneration and a marked reduction in mature follicles, again resembling Meiosin-KO animals (Supplemental Fig. S7D). Thus, the MEIOSIN bHLH domain is required for normal meiosis in both sexes.

In vivo, STRA8 localization in postnatal day 17 (PD17) Meiosin-ΔbHLH testes was shifted toward the cytoplasm, although some nuclear signal remained, unlike the near-complete loss of nuclear STRA8 in Meiosin-KO (Fig. 4A). This residual localization contrasts with the interaction defect observed in 293T cells and suggests that in spermatocytes, MEIOSIN may still weakly support STRA8 localization through indirect or context-dependent mechanisms.

Figure 4.

The bHLH domains of STRA8 and MEIOSIN are required for full meiotic progression in mice. (A) Seminiferous tubule sections from WT, Meiosin-ΔbHLH, and Meiosin-KO testes at PD17 stained for STRA8, MEIOSIN, and DAPI. Enlarged views are shown on the right. (B) Seminiferous tubule sections at PD17 stained as indicated. SYCP3 and SYCP1 mark the synaptonemal complex, and γH2AX marks DNA double-strand breaks. Scale bar, 20 µm. Quantification of meiotic prophase stages is shown below (see also Supplemental Fig. S7E). (pL) preleptotene, (Z) zygotene, (P) pachytene. n: the number of cells examined. (C) Seminiferous tubule sections at PD17 stained as indicated. Enlarged views are shown on the right. (ML) mitotic-like H3S10P+/SYCP3+ cells, (L/Z) leptotene/zygotene. Quantification of seminiferous tubules containing SYCP3+/pH3+ cells among SYCP3+ tubules is shown on the right for control (n = 6), Meiosin-ΔbHLH (n = 4), and Meiosin KO (n = 3) testes (mean ± SD). (n.s.) not significant (P = 0.947), (***) P = 0.000185. Scale bar, 20 µm. n: the total number of tubules examined. (D) Seminiferous tubule sections at PD17 subjected to TUNEL assay together with SYCP3 immunostaining. Scale bar, 20 µm. Quantification of TUNEL-positive tubules among SYCP3-positive tubules is shown below. n: the total number of tubules examined. (E) Seminiferous tubule sections at PD17 from WT, Meiosin-ΔbHLH/Stra8-null, and Stra8-null testes stained as indicated. Scale bar, 20 µm. (F) Model summarizing meiotic gene activation in WT, Meiosin-ΔbHLH, Meiosin-KO, and Stra8 KO germ cells.

Together, these results establish the bHLH domains as the principal determinants of the MEIOSIN–STRA8 interaction and demonstrate their requirement for full meiotic function in mice.

MEIOSIN can trigger meiotic entry without its bHLH domain

Despite the gross similarity between Meiosin-ΔbHLH and Meiosin-KO gonads, closer examination revealed clear cellular differences. In Meiosin-ΔbHLH testes, at least a subset of SYCP3-positive cells progressed to a more advanced zygotene-like state, as judged by SYCP1-positive synaptonemal complex formation and broad γH2AX staining (Fig. 4B; Supplemental Fig. S7E). These cells did not progress beyond pachytene stage, but they clearly advanced further than Meiosin-KO cells.

Consistent with our previous work, Meiosin-KO spermatocytes failed to undergo the mitosis-to-meiosis transition and instead accumulated as mitotic-like cells marked by pH3 and diffuse, patchy SYCP3 staining (Fig. 4C; Ishiguro et al. 2020). In contrast, such mitotic-like cells were not readily detected in Meiosin-ΔbHLH testes, indicating that the transition into the meiotic state had occurred. However, both Meiosin-ΔbHLH and Meiosin-KO testes contained abundant TUNEL-positive tubules (Fig. 4D), showing that germ cells entering prophase are subsequently eliminated by apoptosis.

A similar partial meiotic phenotype has been reported in Stra8 KO testes (Mark et al. 2008; Ishiguro et al. 2020). To test whether the residual activity observed in Meiosin-ΔbHLH testes depends on STRA8, we analyzed Meiosin-ΔbHLH/Stra8-null testes. Cells with leptotene/zygotene-like features, including partial SYCP1 staining and broad γH2AX signals, were still detected (Fig. 4E), suggesting that bHLH-deficient MEIOSIN is associated with partial activation of early meiotic-like features even in the absence of STRA8. However, the double-mutant phenotype was more severe than that of Meiosin-ΔbHLH alone, indicating that STRA8 contributes to the residual activity of bHLH-deficient MEIOSIN or otherwise facilitates progression of the early meiotic program.

Although definitive separation from MEIOSIN-independent pathways would require additional genetic comparisons, such as analysis of Meiosin/Stra8 double-null mutants, these data support the view that the MEIOSIN bHLH domain is not absolutely required for the initial activation of some meiotic-prophase features, but is required for robust execution and completion of the meiotic program (Fig. 4F).

The MEIOSIN bHLH domain is required for full transcriptional reprogramming

To define the molecular basis of this partial phenotype, we performed scRNA-seq on spermatogenic cells from WT, Meiosin-ΔbHLH, and Meiosin-KO testes at postnatal day (PD) 15 to define the cellular states affected by loss of MEIOSIN function or deletion of the bHLH domain (Fig. 5; Supplemental Fig. S8). UMAP analysis identified 15 transcriptionally distinct clusters spanning spermatogenic development (Fig. 5A; Supplemental Fig. S8E; Supplemental Data 4). Cluster 10 corresponded to SSCs, whereas clusters 0, 12, and 9 represented transitional populations in which Meiosin is transiently induced (Fig. 5B).

Figure 5.

scRNA-seq analysis of WT, Meiosin-ΔbHLH, and Meiosin KO spermatogenic germ cells. (A) UMAP representation of scRNA-seq transcriptomes from spermatogenic germ cells isolated from PD15 WT (Alavattam et al. 2024), Meiosin-ΔbHLH, and Meiosin KO testes. Cells are colored by genotype (far left) or cluster identity (right). Arrow indicates the inferred differentiation trajectory based on marker gene expression. UMAP representation of scRNA-seq transcriptomes from testicular cells of WT, Meiosin-ΔbHLH, and Meiosin KO mice. The analysis identified 15 clusters, numbered 0–14. See also Supplemental Figure S7 for the UMAP of whole testicular cells. Germ cells were separated from somatic cells (encompassing Sertoli cells, Leydig cells, and hemocytes) using established marker genes. (B) UMAP plots (left) and violin plots (right) showing mRNA abundance of key developmental genes in spermatogenic germ cells. Black dots indicate medians. Marker genes: Gfra1, spermatogonial stem cells; Meiosin, meiotic initiation (highlighted in pink); Dmc1, early meiotic prophase; Hsf5, late meiotic prophase. (C) Proportion of WT, Meiosin-ΔbHLH, and Meiosin-KO germ cells in each cluster. (D) Metascape analysis of the top 100 enriched genes in clusters 2, 0, and 8, which show genotype-biased cell distributions. Bars indicate enrichment scores (log₁₀(P-value)). See also Supplemental Figure S7E and Supplemental Data 4 for the remaining clusters. (E) Violin plots showing average expression of differentially expressed genes in clusters 0, 9, 12, 8, and 13, comparing Meiosin-ΔbHLH and WT spermatocytes. Genes upregulated in Meiosin-ΔbHLH are shown on the left, and genes downregulated in Meiosin-ΔbHLH are shown on the right, together with their corresponding Metascape analyses. Black dots indicate medians. See also Supplemental Data 5.

UMAP visualization revealed clear genotype-biased distributions among these germ cell clusters (Fig. 5C). Cluster 0 was preferentially enriched in Meiosin-ΔbHLH germ cells, whereas cluster 2 was preferentially enriched in Meiosin-KO germ cells, indicating that the two mutant conditions produce distinct cellular states. Cluster 12 was detected in both WT and Meiosin-ΔbHLH germ cells and may represent a transitional leptotene-like state. In contrast, cluster 9, in which Dmc1 expression reached its peak, was mainly observed in WT germ cells and likely corresponds to an early meiotic prophase population around the leptotene-to-zygotene transition. These observations suggest that Meiosin-ΔbHLH germ cells do not efficiently progress into the cluster 9/Dmc1-high meiotic prophase state, but instead accumulate in cluster 0 as an abnormal early meiotic-prophase-like population. In addition, cluster 8 was selectively enriched in Meiosin-ΔbHLH germ cells and may correspond to the most advanced zygotene-like population detected by cytological analysis (Fig. 4B).

As expected, Meiosin-KO cells accumulated in cluster 2, consistent with a failure to initiate the meiotic program. This cluster was enriched for apoptosis-related genes (Fig. 5D). In contrast, Meiosin-ΔbHLH cells were distributed across clusters 0, 12, and 8, indicating partial entry into the meiotic program. However, they failed to reach late meiotic prophase, as shown by the absence of cells in cluster 14, where Hsf5 is strongly induced (Fig. 5B,C). Consistent with this interpretation, clusters 0 and 8, which were enriched for Meiosin-ΔbHLH cells, retained expression of meiotic genes (Fig. 5D), confirming partial activation of the meiotic transcriptional program.

Nevertheless, this transition was incomplete. Genes normally associated with early spermatogenic growth, particularly those involved in RNA metabolism and translation, remained inappropriately elevated in Meiosin-ΔbHLH cells. Conversely, genes involved in chromatin regulation and meiotic prophase progression were reduced relative to WT (Fig. 5E; Supplemental Data 5). Thus, Meiosin-ΔbHLH cells enter a meiotic state but fail to execute the full transcriptional switch required for orderly meiotic progression.

Together, these findings indicate that the bHLH domain is dispensable for entry into the meiotic state per se, but is required for the full transcriptional reprogramming that supports meiotic progression.

MEIOSIN activates target genes independently of bHLH, but less efficiently

We next asked how the MEIOSIN bHLH domain contributes to direct target gene activation. Previously defined MEIOSIN-upregulated genes (Ishiguro et al. 2020) were preferentially expressed during meiotic prophase, but showed distinct timing: some, such as Dmc1, were induced immediately after Meiosin expression began, whereas others, such as Hsf5, were induced later, after Meiosin expression had already declined (Fig. 5B). We therefore classified these targets into early and late groups (Fig. 6A,B; Supplemental Data 6).

Figure 6.

MEIOSIN activates target genes independently of the bHLH domain, but less efficiently. (A) UMAP plots showing average mRNA abundance of MEIOSIN-upregulated genes (left) and violin plots showing their average expression in clusters 0, 12, 9, 8, and 13 (right). Black dots indicate medians. P-values were calculated using a two-sided Wilcoxon rank-sum test. (B) Metascape analysis of early and late MEIOSIN-upregulated genes. See also Supplemental Data 6. (C) Median average expression of early and late MEIOSIN-upregulated genes in each cluster for Meiosin-ΔbHLH and WT cells. Statistical significance is indicated as P-values (early genes: *P < 0.05, **P < 0.01, ***P < 0.001; late genes: ^†P < 0.05, ^††P < 0.01, ^†††P < 0.001). Pink shading indicates the period of Meiosin expression. (D) Genomic distribution of MEIOSIN-bound peaks associated with MEIOSIN-upregulated genes. (E) DNA sequence motifs enriched in MEIOSIN-bound peaks associated with MEIOSIN-upregulated genes. Gray-shaded motifs were excluded from subsequent analyses. (F) Violin plot showing the number of MEIOSIN-binding motifs per gene among MEIOSIN-upregulated genes. P-values were calculated using a two-sided Wilcoxon rank-sum test. (G) Average expression of MEIOSIN-upregulated genes classified by the number of MEIOSIN-binding motifs per gene. P-values were calculated using a two-sided Wilcoxon rank-sum test.

Meiosin-ΔbHLH cells still activated both groups of MEIOSIN targets, but early genes were induced less strongly than in WT cells (Fig. 6C). Late genes appeared broadly preserved in cluster 0/12/9 (Fig. 6C), although their strongest induction normally occurred in cluster 14, which was absent in Meiosin-ΔbHLH testes (Fig. 5C). These results indicate that the bHLH domain is not absolutely required for transcriptional activation by MEIOSIN, but markedly enhances the amplitude and fidelity of target gene induction.

To examine how this residual activity is encoded, we reanalyzed our previously published MEIOSIN ChIP–seq data (Ishiguro et al. 2020) and found that 91.4% of MEIOSIN-binding peaks associated with genes upregulated in this study lie near transcription start sites (Fig. 6D). Motif analysis identified two prominent nonrepetitive enriched motifs—TGWGGDRR and CTCTAACATGCAGCT—(Fig. 6E). Early MEIOSIN target genes contained more MEIOSIN-binding motifs than late targets, and among early genes, those with three or more motifs were expressed more strongly than those with fewer motifs in both WT and Meiosin-ΔbHLH cells (Fig. 6F). Notably, expression of ≥3-motif genes in Meiosin-ΔbHLH cells approached that of ≤2-motif genes in WT cells (Fig. 6G) suggesting that multiple binding sites can partially compensate for reduced MEIOSIN activity in the absence of the bHLH domain.

We then compared WT, Meiosin-KO, and Stra8-null testicular scRNA-seq data sets (Supplemental Fig. S9A–C). While Meiosin-KO cells accumulated in cluster 2 and Stra8-null cells in cluster 0—both corresponding to stages prior to meiotic entry (Supplemental Fig. S9D,E; Supplemental Data 7). Although both mutants arrested before full meiotic progression (Supplemental Fig. S9B), Stra8-null cells expressed MEIOSIN-upregulated genes at higher levels than Meiosin-KO cells (Supplemental Fig. S9F), indicating that MEIOSIN can activate target genes independently of STRA8. These results suggest that, although STRA8 is required for robust and coordinated meiotic gene activation, some MEIOSIN-associated target-gene expression can persist in its absence.

To identify factors that might support this STRA8-independent activity, we generated mice carrying both 3xFLAG-HA tags at the C-terminus of MEIOSIN and performed tandem-affinity purification of endogenously tagged MEIOSIN followed by mass spectrometry (Supplemental Fig. S10A,B). In addition to STRA8, we identified components of the TIP60/NuA4 histone acetyltransferase complex, including TRRAP, RUVBL1, and RUVBL2, as MEIOSIN-associated proteins (Supplemental Fig. S10C,D; Supplemental Data 8). AlphaFold-based predictions further suggested that intrinsically disordered regions of MEIOSIN acquire structure upon complex formation with TRRAP and RUVBL1 in mouse (Supplemental Fig. S10E,F) and, to a more limited extent, in zebrafish (Supplemental Fig. S10G,H). In mouse MEIOSIN, the region predicted to become structured and to lie in close proximity to TRRAP is retained in the MEIOSIN-ΔbHLH mutant protein (Supplemental Fig. S10F), suggesting that the bHLH domain is not required for the potential interaction between MEIOSIN and TRRAP. The corresponding region of zebrafish Meiosin was also predicted to be positioned near TRRAP; however, this region was not predicted to acquire a similarly defined structure, and its amino acid sequence is not highly conserved between mouse and zebrafish (Supplemental Fig. S10F–H). These observations raise the possibility that MEIOSIN may functionally cooperate with TIP60/NuA4-related chromatin regulatory machinery independently of its bHLH domain. Further molecular analyses will be required to determine how MEIOSIN interacts with chromatin regulatory complexes and how these interactions contribute to meiotic target gene activation.

These observations support a model in which MEIOSIN has intrinsic transcriptional activity that can be potentiated by chromatin-associated cofactors even in the absence of STRA8.

Taken together, these results indicate that disruption of the MEIOSIN bHLH domain broadly reduces MEIOSIN-dependent transcriptional activity rather than creating a clear separation-of-function state affecting a discrete subset of target genes. Nevertheless, the partial induction of MEIOSIN-upregulated genes in Meiosin-ΔbHLH cells, together with the higher expression of these genes in Stra8-null cells than in Meiosin-KO cells, suggests that some MEIOSIN-associated transcriptional activity can persist when the canonical bHLH-mediated MEIOSIN–STRA8 interaction is impaired. Robust and coordinated target-gene induction, however, requires full activity of the MEIOSIN–STRA8 axis.

Discussion

A conserved HMG-containing MEIOSIN core supports meiotic gene activation with STRA8-dependent reinforcement

Our findings suggest that HMG-containing MEIOSIN constitutes a conserved core component of the meiotic-entry program. However, because the absence of detectable Stra8 in zebrafish and hagfish may reflect lineage-specific secondary losses, we do not infer from phylogenetic distribution alone that a STRA8-independent module represents the ancestral vertebrate state. Rather, our functional data indicate that MEIOSIN retains partial activity when the canonical bHLH-mediated MEIOSIN–STRA8 interaction is disrupted (Fig. 4). These findings support a model in which STRA8 reinforces, amplifies, and stabilizes an underlying MEIOSIN-dependent transcriptional program in lineages that retain both factors. This framework may explain how meiotic programs can still be initiated or proceed in species that lack STRA8, and why bHLH-deficient MEIOSIN retains partial function in mice.

Because we have not generated an HMG-deletion allele, our data do not directly establish that the HMG domain alone is sufficient for MEIOSIN function. We therefore refer to the HMG domain as the most consistently conserved architectural feature of MEIOSIN and as a candidate core functional element. Direct genetic dissection of the HMG domain will be required to determine whether it contributes to a transcriptional program that is separable from, or broadly coupled to, the bHLH-mediated MEIOSIN–STRA8 module.

The evolutionary logic of this model is further supported by the domain architecture of MEIOSIN orthologs. MEIOSIN proteins in zebrafish and hagfish retain a conserved HMG domain and broadly disordered regions, but lack the bHLH domain (Fig. 1; Supplemental Fig. S2A). In other vertebrate lineages that retain STRA8, MEIOSIN orthologs generally also retain the bHLH domain. One parsimonious interpretation is that the ancestral vertebrate MEIOSIN possessed both domains and that Stra8 was lost secondarily in the zebrafish and hagfish lineages. Under this scenario, after STRA8 loss, the MEIOSIN bHLH domain—whose principal role is to mediate the canonical MEIOSIN–STRA8 interaction—may have become dispensable and been subsequently lost independently in these lineages. However, as our analysis did not include more distant species, further study is required to determine whether hagfish lost Stra8, as zebrafish did, or whether hagfish lineage branched off from other jawed vertebrates before Stra8 was acquired during evolution. Thus, the absence of the MEIOSIN bHLH domain in zebrafish and hagfish is unlikely to represent a primitive state, but may instead reflect secondary simplification that independently happen during evolution in zebrafish and hagfish.

The fission yeast meiotic regulator Ste11 (Sugimoto et al. 1991) shows a similar combination of an HMG domain and extended nonglobular regions, suggesting that HMG-based control of meiotic entry may reflect a deeply conserved regulatory solution. Our study therefore identifies the HMG domain as the minimal conserved architectural feature required for MEIOSIN function and places the bHLH-dependent interaction with STRA8 as a derived enhancement rather than the ancestral core. We propose that HMG-containing MEIOSIN represents an evolutionarily conserved core module for meiotic entry, whereas STRA8 and the bHLH-mediated MEIOSIN–STRA8 interaction act as lineage-specific amplifiers of this ancestral program.

STRA8 strengthens, but does not define, meiotic initiation

Our data refine the functional relationship between MEIOSIN and STRA8. In the prevailing mammalian framework, the two factors are often viewed as a single obligate module. However, the present results indicate that MEIOSIN retains intrinsic activity in the absence of both STRA8 and its own bHLH domain (Figs. 4–6). This residual activity is sufficient to trigger aspects of meiotic entry and to induce a subset of target genes, especially when multiple MEIOSIN-binding motifs are present near promoters.

The combined Meiosin-ΔbHLH; Stra8-null mutant exhibited a phenotype broadly similar to that of the Stra8-null mutant alone (Fig. 4E). Therefore, this genetic comparison does not by itself establish a clear separation between STRA8-dependent and STRA8-independent MEIOSIN functions. Rather, it indicates that disruption of the bHLH-mediated MEIOSIN–STRA8 interaction impairs the overall activity of the meiotic-entry program. Although deletion of the bHLH domain strongly impaired MEIOSIN–STRA8 interaction in the heterologous co-immunoprecipitation assay (Supplemental Fig. S7A), we cannot exclude the possibility that STRA8 and bHLH-deficient MEIOSIN still cooperate indirectly or through weak, context-dependent mechanisms in in vivo germ cells. In species that lack STRA8, such as zebrafish and hagfish, MEIOSIN may therefore require alternative cofactors or chromatin contexts to achieve sufficient activity. Thus, STRA8 is not the sole trigger of meiotic entry; rather, it enhances the efficiency, magnitude, and completeness of MEIOSIN-dependent transcription.

This interpretation helps reconcile previous genetic observations. In some mouse strain backgrounds, Stra8 mutants can still progress into early meiotic prophase (Mark et al. 2008; Ishiguro et al. 2020), consistent with the idea that MEIOSIN alone can partially activate the meiotic program, leading to phenotypic variation among different mouse strains. Such phenotypic variation may reflect differences in chromatin state, cofactor availability, or promoter architecture that modulate the threshold for MEIOSIN function.

Our data further suggest that MEIOSIN regulates transcription through both direct and indirect mechanisms. We found that MEIOSIN target genes can be categorized into early and late groups based on their expression dynamics (Fig. 6). Notably, late genes remain upregulated even after Meiosin expression has ceased, indicating that at least part of the downstream transcriptional program is maintained indirectly, likely through secondary regulatory factors induced by MEIOSIN.

The strong enrichment of NuA4/TIP60 components among MEIOSIN-associated proteins (Supplemental Fig. S10) further suggests that chromatin remodeling and histone acetylation help determine the competence of germ cells to respond to MEIOSIN.

The germ cell specificity of both MEIOSIN and STRA8 function is also consistent with this view. Even if MEIOSIN has intrinsic transcriptional activity, it likely operates only in a permissive chromatin environment established in germ cells at the onset of meiotic differentiation. A key next step will therefore be to define the epigenetic features that license MEIOSIN to activate meiotic genes and to understand how STRA8 and other cofactors increase the robustness of this response.

This framework explains how meiotic entry can still occur in the absence of STRA8, while also accounting for the enhanced efficiency and robustness conferred by the MEIOSIN–STRA8 complex in species that retain both factors.

A reduced MEIOSIN-dependent module for early oocyte development in zebrafish

The zebrafish data reveal a striking sexual asymmetry. Although meiosin is induced at the mitosis-to-meiosis transition in both sexes (Fig. 2), its loss selectively disrupts the female germline (Fig. 3), causing arrest of oogenesis and complete female-to-male sex reversal. However, the mutant phenotype does not allow us to conclude that Meiosin acts solely at the step of meiotic induction in females. In laboratory zebrafish strains, juvenile gonads initially form early oocytes, and maintenance or loss of these oocytes is closely associated with ovarian versus testicular development. Because early oocyte-like cells and SYCP1-positive nuclei are still detected in meiosin mutant juvenile gonads (Fig. 3H), zf Meiosin is unlikely to be absolutely required for the initial formation of oocytes or for the acquisition of early meiotic-prophase features. Rather, the absence of stage IB oocytes and the fully penetrant female-to-male sex reversal suggest that zf Meiosin is required for robust early oocyte development and maintenance of the ovarian pathway.

It remains difficult to determine whether the primary defect reflects impaired meiotic progression, defective oocyte maturation, or failure to maintain the female developmental program, because these processes are tightly coupled in zebrafish gonad development. Moreover, because common laboratory zebrafish strains lack a simple chromosomal sex marker, we cannot retrospectively distinguish individuals that would otherwise have developed as females from those that would have developed as males. Nevertheless, adult meiosin KO males produced spermatozoa and were fertile under standard laboratory conditions (Fig. 3), indicating that zf Meiosin is not overtly required for male fertility in this setting.

One possibility is that the female pathway imposes a narrower developmental window during which early meiotic progression must be coordinated with commitment to oogenesis. In this context, reduced or delayed activation of meiotic genes may compromise the female program and redirect development toward the male pathway. In contrast, the male germline may be buffered by broader developmental timing, redundant regulators, or a chromatin environment that more readily supports residual or alternative meiotic regulatory activity. The finding that Meiosin is induced in the GSC/progenitor compartment in females (Fig. 2G,I), earlier than in males, is consistent with the idea that early meiotic/oogenic differentiation is more tightly integrated with lineage commitment in the female germline.

Our mouse genetic dissection provides a possible framework for understanding how such a reduced system could operate. Because some MEIOSIN-associated transcriptional activity persists when STRA8 is absent or when the MEIOSIN bHLH domain is deleted (Fig. 6), zebrafish MEIOSIN may rely on alternative cofactors, promoter architectures, or permissive chromatin states to achieve sufficient activity in the absence of STRA8. Identifying these cofactors and chromatin features that potentiate zebrafish MEIOSIN will be essential for understanding how early oocyte development and meiotic gene activation are supported in vertebrate lineages that lack the canonical MEIOSIN–STRA8 module.

Materials and methods

Animals

Mice

Meiosin ΔbHLH mutant knock-in (Meiosin^ΔHLH-KI) mice and other knockout and knock-in mice were congenic with the C57BL/6 background (age: postnatal day 17, 4 and 8 weeks old). Wild type Stra8-3xFLAG-HA-p2A-GFP knock-in (Stra8wt-3FH KI), Stra8-null GFP knock-in (Stra8-null GFP-KI), Meiosin KO mice (age: postnatal day 17, 4 and 8 weeks old) were generated as described in our previous studies (Ishiguro et al. 2020; Shimada et al. 2023). Male mice were used for histological analysis of testes, testis immunostaining, and scRNA-seq experiments at the indicated ages (age: postnatal day 15, and 8 weeks old). Female mice were used for histological analysis of the ovaries (age: 8 weeks old). Whenever possible, each knockout animal was compared to littermates or age-matched nonlittermates from the same colony, unless otherwise described. Animal experiments were approved by the Institutional Animal Care and Use Committee (Kumamoto University approval F28-078, A2022-001, and A2024-028, Chiba University approval A7-197).

Fish

Male and female hagfish were purchased from Shinshomaru. Zebrafish (Danio rerio) were maintained under standard conditions as described by Westerfield (1995). The wild-type AB line (see ZFIN at http://www.zfin.org/action/genotype/view/ZDB-GENO-980210-28) was used to generate the meiosin-like KO zebrafish line by the CRISPR-Cas9 mutagenesis based on published protocols (Hwang et al. 2013; Chen et al. 2017). Template DNA for single-guide RNA (sgRNA) synthesis was prepared by PCR using a primer specific to meiosin exon 4 (5′-taatacgactcactataGGTGCAAACAAGCTCCTCAAgttttagagctagaa-3′), a universal reverse primer (5′-AAAAGCACCGACTCGGTGCCACTTTTTCAAGTTGATAACGGACTAGCCTTATTTTAACTTGCTATTTCTAGCTCTAAAAC-3′), and T4 DNA polymerase.

After purification of the template DNA, sgRNA was transcribed in vitro with a MEGAscript T7 kit (Ambion) and purified with a MEGA clean-up kit (Ambion). Wild-type AB embryos were injected at the 1 or 2 cell stage with 2.3 nL of a mixture of 10 pmol/µL Cas9 NLS protein (abm) and 100 ng/µL meiosin sgRNA. Founders were backcrossed with AB fish, and the F1 siblings were screened by genotyping using a specific primer pair; meiosin P4F (5′-ACAGACTAAAGACTGCGTGAGAT-3′) and meiosin P4R (5′-TGACGAGTTGCTTTGTCTGTTG-3′).

The mutations were verified by Sanger sequencing after cloning the PCR products into the pCRII vector using a TOPO TA cloning kit (Invitrogen). Heterozygous meiosin knockout fish carrying an 11 bp frameshift insertion in exon 4 were intercrossed to obtain homozygous mutant fish. Sanger sequencing primers are M13 Fwd (-20) (5′-GTAAAACGACGGCCAG-3′) and M13 Rev (5′-CAGGAAACAGCTATGAC-3′).

Generation and genotyping of Meiosin-ΔbHLH knock-in mice

ΔbHLH mutant Meiosin knock-in (Meiosin^ΔbHLH KI) mouse was generated by introducing Cas9 protein (Nippon Gene 317-08441), tracrRNA (FASMAC GE-002), synthetic crRNA (FASMAC) and ssODN into C57BL/6N fertilized eggs using electroporation. The synthetic crRNAs were designed to target GAGAACAATTACCTGTCAGT (agg) in the Meiosin Exon 2, and TCATGTCAATGTTTCTCTGG (agg) in the Meiosin Exon 4.

The ssODN (5′-TTGCCATTGCTGCTGTGGAGCTTGAGCAAGGCCTTGGTCATGTCAATGTTCCTGTCAGTAGGACTTGGAGATCTAGCTCTGTGCTGATCAGAGAAACATG-3′) was used as a homologous recombination template.

The electroporation solutions contained 10 µM tracrRNA, 10 µM synthetic crRNA, 0.1 µg/µL Cas9 protein, 1 µg/µL ssODN in Opti-MEM I reduced serum medium (Thermo Fisher Scientific 31985062). Electroporation was carried out using the super electroporator NEPA 21 (Nepa Gene) on Glass Microslides with round wire electrodes, 1.0 mm gap (BTX 45-0104). Four steps of square pulses were applied (1, three times of 3 mS poring pulses with 97 mS intervals at 30 V; 2, three times of 3 mS polarity-changed poring pulses with 97 mS intervals at 30 V; 3, five times of 50 mS transfer pulses with 50 mS intervals at 4 V with 40% decay of voltage per each pulse; 4, five times of 50 msec polarity-changed transfer pulses with 50 msec intervals at 4 V with 40% decay of voltage per each pulse).

The Meiosin^ΔbHLH KI allele (341 bp) was identified by PCR using the following primers; Gm4969-13926F (5′-GTCCTATTTTAGGAACTCTAGGTGC-3′) and Meiosin delHLH-1R (5′-CAGGAAGGGGAAGGTAGCCAATTGG-3′). The wild-type allele (212 bp) was identified by PCR using the following primers; Gm4969-13926F and Gm4969-14136R (5′-GTAGAGAACAATTACCTGTCAGTAGG-3′). The PCR amplicons were verified by sequencing. To avoid potential mosaicism of the mutant allele, F0 mutant founders were backcrossed with C57BL/6N mice to establish heterozygous lines. Because homozygous Meiosin^ΔbHLH KI mice were infertile, the line was maintained as heterozygotes. The Meiosin^ΔbHLH KI allele encodes MEIOSIN lacking 10–70 aa.

Generation of Meiosin-3xFLAG-HA mutant knock-in mouse and genotyping

3xFLAG-HA was inserted by introducing Cas9 protein (317-08441; Nippon Gene), tracrRNA (GE-002; FASMAC), synthetic crRNA (FASMAC) and ssODN into C57BL/N fertilized eggs using electroporation. The synthetic crRNAs were designed to target CACCCCTAACTGCCAGTGAG (AGG) in the Meiosin Exon 14. The ssODN (5′-ATGCTAAGCAGGTGGCAGAGCCCAGCAGGACTGCCCgTCTCATTAGGCGTAGTCGGGCACGTCGTAGGGGTAtccCTTGTCATCGTCATCCTTGTAATCGATGTCATGATCTTTATAATCACCGTCATGGTCTTTGTAGTCtccCTGGCAGTTAGGGGTGGGCAGTGAGGTGAGGCCCAAAGACAC-3′) was used as a homologous recombination template. The electroporation solutions contained 10 µM tracrRNA, 10 µM of synthetic crRNA, 0.1 µg/µL of Cas9 protein, ssODN (1 µg/µL) in Opti-MEM I reduced serum medium (31985062; Thermo Fisher Scientific). Electroporation was carried out using the Super Electroporator NEPA 21 (NEPA GENE) on Glass Microslides with round wire electrodes, 1.0 mm gap (45-0104; BTX). Four steps of square pulses were applied (1, three times of 3 mS poring pulses with 97 mS intervals at 30 V; 2, three times of 3 mS polarity-changed poring pulses with 97 mS intervals at 30 V; 3, five times of 50 mS transfer pulses with 50 mS intervals at 4 V with 40% decay of voltage per each pulse; 4, five times of 50 mS polarity-changed transfer pulses with 50 mS intervals at 4 V with 40% decay of voltage per each pulse).

Using following primers, wildtype allele (647 bp) and the Meiosin-3FH KI allele (732 bp) were identified by PCR; Gm4969-29318F (5′-CACTAAGGCACGCAGGTTCAGCCGC-3′) and Gm4969-29964R (5′-GAGGTAAAGGGTTTGAGTTCAGACC-3′).

Transfection and cell culture

293T cells were cultured in DMEM, supplemented with 10% FBS, 2 mM l-glutamine, 100 U/mL penicillin and 100 µg/mL streptomycin at 37°C with 5% CO₂. For STRA8-MEIOSIN interaction analyses, 3.6 × 10⁶ 293T cells were transfected with the indicated expression vectors together with pCAG-PBase using Lipofectamine 2000 and cultured for 2 days. The plasmids used in this study are as follows: pBRPBCAG-STRA8-3FH-IN (WT STRA8-3xFLAG-HA), pBRPBCAG-STRA8-3FH dHLH-IN (dHLH STRA8-3xFLAG-HA) encoding STRA8 lacking 17–84 amino acids, pBRPBCAG-Gm4969-IN (WT MEIOSIN), pBRPBCAG-Gm4969 dHMG-IN (dHMG MEIOSIN) encoding MEIOSIN lacking 484–534 amino acids, and pBRPBCAG-Gm4969 dHLH-IN (dHLH MEIOSIN) encoding MEIOSIN lacking 10–70 amino acids.

Preparation of cell extracts and immunoprecipitation

Chromatin-bound and -unbound extracts were prepared as described previously (Tani et al. 2022). Briefly, testicular cells were suspended in low salt extraction buffer (20 mM Tris-HCl at pH 7.5, 100 mM KCl, 0.4 mM EDTA, 0.1% TritonX100, 10% glycerol, 1 mM β-mercaptoethanol) supplemented with complete protease inhibitor (Roche). After homogenization, the soluble chromatin-unbound fraction was separated after centrifugation at 100,000g for 30 min. The chromatin bound fraction was extracted from the insoluble pellet by high salt extraction buffer (20 mM HEPES-KOH at pH 7.0, 400 mM KCl, 5 mM MgCl₂, 0.1% Tween20, 10% glycerol, 1 mM β-mercaptoethanol) supplemented with Complete Protease Inhibitor. The solubilized chromatin fraction was collected after centrifugation at 100,000g for 30 min at 4°C.

Immuno-affinity purification was performed with anti-FLAG M2 monoclonal antibody-coupled magnetic beads (Sigma-Aldrich) from the testis chromatin-bound fractions of MEIOSIN-3xFLAG-HA-p2A-GFP knock-in mice. For negative control, mock immuno-affinity purification was done from the testis chromatin-bound and -unbound fractions from the age-matched wild type mice. The beads were washed with high salt extraction buffer for chromatin-bound proteins and low salt extraction buffer for chromatin-unbound proteins. The anti-FLAG-bound proteins were eluted by 3xFLAG peptide (Sigma-Aldrich). The second immuno-affinity purification was performed using anti-HA 5D8 monoclonal antibody-coupled magnetic agarose beads (MBL M132-10). The bead-bound proteins were eluted with 40 µL of elution buffer (100 mM Glycine-HCl at pH 2.5, 150 mM NaCl), and then neutralized with 4 µL of 1 M Tris-HCl (pH 8.0).

The 3xFLAG-HA tagged version of WT STRA8, and the STRA8^{ΔHLH E} proteins were immunoprecipitated with anti-FLAG M2 monoclonal antibody-coupled magnetic beads (Sigma-Aldrich) from 293T cell extracts expressing Stra8 ^wt-3FH, and Stra8^ΔHLH-3FH, respectively. The immunoprecipitated proteins were run on SDS-PAGE and immunoblotted. Immunoblotting for TRRAP detection was performed in Tris-glycine buffer containing 20% methanol, and 0.05% SDS. Immunoblot image was developed using ECL prime (GE HealthCare) and captured by FUSION Solo (VILBER).

Mass spectrometry

Mass spectrometry was performed as previously described (Tani et al. 2022). Briefly, the immunoprecipitated proteins were run on 4%–12% NuPAGE gel (Thermo Fisher Scientific) by 1 cm from the well and stained with SimplyBlue (Thermo Fisher Scientific) for in-gel digestion. The gel containing the proteins was excised and cut into pieces ∼1 mm size. The proteins in these gel pieces were then reduced with DTT (Thermo Fisher Scientific), alkylated with iodoacetamide (Thermo Fisher Scientific), and digested overnight at 37°C using trypsin and Lysyl endopeptidase (Promega) in a buffer containing 40 mM ammonium bicarbonate, pH 8.0. The resulting peptides were analyzed using an Advance UHPLC system (ABRME1ichrom Bioscience) connected to a Q Exactive mass spectrometer (Thermo Fisher Scientific), with the raw mass spectra processed using Xcalibur (Thermo Fisher Scientific). The raw LC-MS/MS data were analyzed against the SwissProt protein/translated nucleotide database, restricted to Mus musculus using Proteome Discoverer v 1.4 (Thermo Fisher Scientific) with the Mascot search engine version2.5 (Mascot Science). A decoy database comprising either randomized or reversed sequences from the target database was used to estimate false discovery rate (FDR), and the Percolator algorithm was used to evaluate false positives. The search results were filtered against a global FDR of 1% to achieve a high level of confidence. Proteins that detected in the control sample were removed from further analysis.

Proteins identified by LC–MS/MS are shown after excluding those also detected in anti-FLAG/ anti-HA immunoprecipitates from wild-type testis extracts. The proteins identified by more than two distinct peptide hits are listed together with their Swiss-Prot accession numbers, gene symbols, amino acid lengths, Mascot scores, and numbers of peptide hits.

Antibodies

The following antibodies were used for immunoblot (IB) and immunofluorescence (IF) studies in mice: rabbit anti-H3S10P (IF; 1:2000; Abcam: ab5176), rabbit anti-SYCP1 (IF; 1:1000; Abcam ab15090), rabbit anti-TRRAP (WB; 1:1000; Abcam ab73546), rat anti-SYCP3 and guinea pig anti-SYCP3, rabbit and rat anti-STRA8, rabbit and guinea pig anti-MEIOSIN N-terminal (amino acids 1–224) and rabbit and guinea pig anti-MEIOSIN C-terminal (amino acids 405–589) as described previously (Ishiguro et al. 2020). Rabbit anti-zf Vasa (IF; 1:200; GeneTex GTX128306) was used for cloudy catshark. For zebrafish samples, the following antibodies were used: Rabbit anti-zf Meiosin (IF: 1:300; WB: 1:1000; this study), Mouse anti-zf Foxl2l (IF, 1:100, this paper). Rabbit anti-Ziwi (IF, 1:500) (Houwing et al. 2007; Kawasaki et al. 2025), guinea pig anti-zf Sycp3 (IF; 1:200) (Imai et al. 2021), rabbit anti-zf Sycp3 (IF; 1:200) (Saito et al. 2011), and rat anti-zf Sycp1 (IF; 1:200) (Saito et al. 2014).

Histological analysis

For hematoxylin and eosin staining, mouse testes, epididymis and ovaries were fixed in Bouin solution, and embedded in paraffin. Sections were prepared on CREST-coated slides (Matsunami) at 6 µm thickness. Cloudy catshark and zebrafish testes, and zebrafish juvenile bodies were fixed in 4% paraformaldehyde for overnight at 4°C or Bouin solution, and embedded in paraffin. Sections were prepared on CREST slides (Matsunami) at 4 µm. The slides were deparaffinized and stained with hematoxylin and eosin.

Immunofluorescence microscopy of mouse testis

For immunofluorescence staining of mouse testes and ovaries, tissues were fixed in 4% PFA for several hours at 4°C, cryoprotected sequentially in 10%, 20%, and 30% sucrose at 4°C, embedded in Tissue-Tek O.C.T. compound (Sakura Finetek) and frozen at −80°C. Cryosections were prepared on the MAS-GP type A-coated slides (Matsunami) at 6 µm thickness, and then air-dried before staining. Sections were blocked for 30 min in 3% BSA containing 0.1% TritonX100 in TBS (TBS-T), and incubated with the primary antibodies in a blocking solution at 4°C. After three washes in TBS-T, sections were incubated for 90 min at room temperature with Alexa-dye-conjugated secondary antibodies (1:1000; Invitrogen). DNA was counterstained with Vectashield mounting medium containing DAPI (Vector Laboratory). TUNEL assays were performed using MEBSTAIN Apoptosis TUNEL Kit Direct (MBL 8445).

Immunofluorescence microscopy of fish testis and juvenile body

Cloudy catshark and zebrafish testes, and zebrafish juvenile bodies, were fixed in 4% paraformaldehyde overnight at 4°C, embedded in paraffin, and sectioned at 4 µm on CREST slides (Matsunami). Sections were deparaffinized in xylene and rehydrated through a graded ethanol series (100%, 90%, 80%, 70%, and 50%), followed by four washes in PBS-0.1% Tween-20 for 5 min each. Antigen retrieval was performed in 1× Immuno Saver (Nisshin EM) at 98°C for 45 min, after which sections were washed once in PBS-0.5% Triton X-100 for 15 min and three times in PBS-0.1% Tween-20 for 5 min each. After blocking with 2% BSA-PBS for 30 min at room temperature, sections were incubated with primary antibodies diluted in the blocking solution or Can Get Signal Immunostain Solution (Toyobo) overnight at room temperature. Following three washes in PBS-0.1% Tween-20, sections were incubated with Alexa dye–conjugated secondary antibodies (1:200 for anti-zf Sycp1 and anti-zf Sycp3; 1:500 for anti-zf Meiosin and anti-zf Foxl2l) for 2 h at room temperature. After three washes in PBS-0.1% Tween-20, sections were counterstained with 0.5 µg/mL DAPI for 15 min and mounted in ProLong Glass (Invitrogen).

Imaging

Immunostaining images were captured with DeltaVision (GE HealthCare). The projection of the images was processed with the SoftWorx software program (GE HealthCare). All images shown were Z-stacked. For counting seminiferous tubules and embryonic ovaries, immunostaining images were captured with BIOREVO BZ-X710 (KEYENCE), and processed with BZ-H3A program. XY-stitching capture by 40× objective lens was performed for multiple-point color images using DeltaVision. Images were merged over the field using SoftWorx software program. Bright field images were captured with OLYMPUS BX53 fluorescence microscope and processed with CellSens standard program.

Antibody production

Polyclonal antibody specific for zebrafish Sycp2 was described in the previous study (Takemoto et al. 2020). Polyclonal antibodies against zebrafish Meiosin (amino acids 1–407) were generated by immunizing rabbits. Polyclonal antibody against zebrafish Foxl2l (amino acids 1–260) was generated by immunizing mouse. His-tagged recombinant proteins were produced by inserting cDNA fragments in-frame with pET19b (Novagen) in Escherichia coli strain BL21-CodonPlus (DE3)-RIPL (Agilent), solubilized in a denaturing buffer (6 M HCl-Guanidine, 20 mM Tris-HCl pH 7.5) and purified by Ni-NTA (QIAGEN) under denaturing conditions. The antibodies were affinity-purified from the immunized serum with immobilized antigen peptides on CNBr-activated Sepharose (GE HealthCare).

5′-Full RACE

5′ Full RACE was performed using 5′-Full RACE Core Kit (TaKaRa Bio). Total RNA was isolated from the adult testis of zebrafish using TRIzol (Thermo Fisher Scientific). First cDNA was synthesized using Superscript III from total RNA with a 5′ end-phosphorylated zf Meiosin primer, zf Meiosin-5′-RACE-pR; 5′-AGTTGGAGTCTGTAG-3′. The hybrid DNA-RNA was degraded by treatment with RNase H at 30°C for 1 h. After ethanol precipitation, the single stranded cDNA was ligated using T4 RNA ligase. The ligated product was amplified using PrimeSTAR MAX (TaKaRa) with following primers; zfMeiosin_S1: 5′-ACAGACAAAGCAACTCGTCAC-3′ and zfMeiosin_A1: 5′-AGTCTGTGAGTCCCTCTTATTG-3′. Amplified product was further amplified using MightyAmp (TaKaRa) with following primers; zfMeiosin_S2: 5′-GTTGAAGACAGTGAAGCATCAC-3′ and zfMeiosin_A2: 5′-TGCGACTGTAAGTGTTCATG-3′. The PCR product was cloned into pCR2.1 using a TOPO TA Cloning Kit (Thermo Fisher Scientific) and sequenced.

Southern blot

Ten micrograms of zebrafish genomic DNA was digested by restriction enzymes and run on 0.8% agarose gel electrophoresis followed by transfer to Hybond N+ membrane (GE) in 0.4 N NaOH. The membrane was subjected to hybridization in QuickHyb (Agilent 201220) with ³²P-labeled probe that encompasses genomic region of meiosin-like Exon2–Exon5 at 65°C. Autoradiography image was acquired by the Typhoon FLA7000 Biomolecular imager (Cytiva).

Single-cell RNA-sequencing

Testes were collected from wild type zebrafish at adulthood. Ovaries were collected from Tg(vas::EGFP) females at 54 dpf. In addition, Ovaries were also collected from Tg(vas::EGFP, gata6SAGFF254A, UAS::NTR) females at 85 dpf. In this line, nitroreductase (NTR) is expressed in oocytes larger than 40 µm, allowing selective ablation of growing oocytes after metronidazole treatment. To enrich early ovarian germ cell populations, fish were subjected to two rounds of 8 mM metronidazole treatment in fish water before dissection: an initial 3 day treatment, followed by 2 days of recovery in fish water, and a second 3 day treatment performed 1 week before ovary collection at 85 dpf.

Single-cell suspensions were prepared by incubating testes and ovaries with Accutase (Innovative Cell Technologies, Inc.) for 5 min at 37°C. After incubation, DMEM with 10% FBS was added to block Accutase and aggregate was disrupted by pipetting. Then, cell suspensions were filtered through a 35 µm cell strainer sieve (BD Bioscience). Cells were collected by centrifugation and re-suspended in PBS containing 0.1% BSA, and dead cells and debris were removed using Cell Sorter SH800 (SONY). Collected cells were re-suspended in DMEM containing 10% FBS. Resulting ∼1000–2000 single-cell suspensions were loaded on Chromium Controller (10X Genomics Inc.). Single cell RNA-seq libraries were generated using Chromium Single Cell 3′ Reagent Kits v3 following manufacturer's instructions, and sequenced on an Illumina HiSeq X to acquire paired end 150 nt reads. The number of animals used, the total number of single cells captured from testes or ovaries, the mean read depth per cell, sequencing saturation (%), number of detected genes, median UMI counts per cell, and total number of cells before and after quality control are shown in Supplemental Figures S4, S5, and S8.

Sequencing data are available at DDBJ Sequence Read Archive (DRA) under the accession.

Statistical analysis of scRNA-seq

Fastq files were processed and aligned to the zebrafish transcriptome (GENCODE vM23/Ensembl 98) using the 10X Genomics Cell Ranger v 4.0 pipeline. Further analyses were conducted on R (ver.4.4.1) (R Core Team 2019) via RStudio (ver.2024.04.2+764) (RStudio Team 2018). Because the number of germ cells obtained from a single collection was limited, the data from the 54 dpf Tg(vas::EGFP) ovaries and the 85 dpf NTR-mediated oocyte-ablation ovaries were integrated for downstream scRNA-seq analysis. Quality assessment of scRNA-seq data and primary analyses were conducted using the Seurat package for R (v5.1.0) (Butler et al. 2018; Stuart et al. 2019). Contaminated somatic cells were excluded, and only the cells that expressed more than 200 genes were used for further analysis to remove the effect of low-quality cells. The scRNA-seq data were merged and normalized using SCTransform function built in Seurat. The dimensional reduction analysis and visualization of cluster were conducted using RunUMAP function built in Seurat. The clustering of cells was conducted using FindNeighbors and Findclusters built in Seurat with default setting. Determination of differentially expressed genes (DEGs) was performed using FindAllMarkers function built in Seurat with following options, only.pos = T, min.pct = 0.25 and logfc.threshold = 0.25, for the identification of marker genes in each cluster, and FindMarker function built in Seurat with default settings to characterize the arbitrary group of cells. RNA velocity analysis was conducted using the RNA velocyto package for python (Python 3.7.3) and R with default settings (La Manno et al. 2018), and visualized on UMAP plots built in Seurat.

Gene enrichment analysis

Gene enrichment analyses were performed using Metascape (Zhou et al. 2019) with default settings and the results were visualized using R. To characterize the feature of each cluster, top100 representative genes in each cluster were used.

Classification of MEIOSIN upregulating genes

Original MEIOSIN upregulating genes were reported in previous report (Ishiguro et al. 2020). The genes detected in scRNA-seq data were selected for further analysis. The genes that expressed in cluster 0, 12, 9, and 8 (E1) were determined using FindMarker function built in Seurat with following option, only.pos = T. Highly expressed genes in clusters 13 and 14 compared to clusters 0, 12, 9, and 8 were determined using FindMarker function built in Seurat, with a threshold P_val_adj < 0.05 (Late High). Among original MEIOSIN upregulating gene list, genes that also listed in Late High genes were defined as Late genes. Among genes commonly listed in both original MEIOSIN upregulating genes and E1 genes, genes that were not listed as Late genes were defined as Early genes.

RT-PCR

Total RNA was isolated from hagfish tissues using TRIzol (Thermo Fisher). Hagfish cDNA was generated from total RNA using Superscript III (Thermo Fisher) followed by PCR amplification using Quick-Taq (TOYOBO) and template cDNA. Cloudy catshark and zebrafish cDNAs were generated from total RNAs using PrimeScript RT-PCR kit (Takara) followed by PCR amplification using KOD Plus Neo (Toyobo) or GoTaq (Promega). Amplified PCR products were cloned into the pCRII vector using TOPO TA cloning kit (Invitrogen), and sequenced with M13 Fwd (-20) and M13 Rev primers.

Sequences of primers used for RT-PCR were as follows: Hagfish Ddx4_F1 (5′-GAAGTGCGCTCACTGTCAGAG-3′), Hagfish Ddx4_R1 (5′-CATGACCTCCTCTGCCAACAC-3′), Hagfish Sycp3_F1 (5′-ATGGCCTCTGATCAAGATGACATG-3′), Hagfish Sycp3_R1 (5′-TCCTGGTGCAGAGTCTCCAG-3′), Hagfish Gapdha_F (5′-CAGCAATGCTTCCTGCACTACC-3′), Hagfish Gapdha_R (5′-CGTTGTCGTACCATGACACCAAC-3′), Hagfish meiosin-like_F1 (5′-CCCAATGAACAGCCAGACCAAG-3′), Hagfish meiosin-like_R1 (5′-CTGTCCCGTACACAGAATCGGTAG-3′), Zf actb1_F1 (5′-GACGACCCAGACATCAGGGAGTG-3′), Zf actb1_R1 (5′-CACGGACAATTTCTCTTTCGGCTG-3′), Zf meiosin-like_F1 (5′-CCAGGGAGAAAGTCATTGGA-3′), Zf meiosin-like_R1 (5′-AGGCCGTACTGGTGTTTTTG-3′), Cloudy shark Tbp_F (5′-CGGGAACCTCGTACAACTGC-3′), Cloudy shark Tbp_R (5′-GGCAGCAGGCAAACGAAATG-3′), Cloudy shark Sycp3_F (5′-CACTGGGGAGCAGCTATGAGA-3′), Cloudy shark Sycp3_R (5′-GAAAAGCTTCTGTTGCTGCCG-3′), Cloudy shark Stra8_F (5′-ATGGAGAGCTCTGGAGATTG-3′), Cloudy shark Stra8_R (5′-TTACAGATCATCATCGAATGTCTC-3′), Cloudy shark Meiosin_F (5′-ATGGCCTCCATAATTGAGGTGAAGC-3′), and Cloudy shark Meiosin_R (5′-GTCAGGCAGCATCTTACAGTAGATGG-3′).

Fertilization tests and embryogenesis of zebrafish

To assess male fertility, three independent pair matings per genotype were performed, in which three meiosin^−/−, meiosin^+/− and wild-type sibling males were individually mated with a wild-type female in the same day. Each pair was subjected to two repeated matings. Statistical analysis was performed with the Mann–Whitney two-tailed test using GraphPad Prism 11 software.

To examine embryogenesis of meiosin^+/− female-derived embryos, embryos obtained from two independent paired crosses of meiosin^+/− and wild-type male were kept at 28°C. Embryos were observed and imaged at the indicated developmental stages using a Leica MX16 FA microscope equipped with a Leica DFC310 FX camera.

Molecular phylogenetic analysis

Protein sequences were collected from the NCBI and Ensembl databases. The deduced amino acid sequences were aligned using the MAFFT v7.505 (Katoh and Standley 2014) with the L-INS-i method. The aligned sequences were trimmed using trimAl v1.4.rev15 (Capella-Gutiérrez et al. 2009) with the “-gappyout” option to remove unreliably aligned sites. A maximum-likelihood tree was inferred using iqtree 1.6.1 (Nguyen et al. 2015) with the JTT+I+G4 model for Meiosin and the JTT+G4 model for Stra8. Ultrafast bootstrap resampling with 1000 replicates was performed to evaluate the confidence of the nodes.

Data availability

All data supporting the conclusions are present in the paper and the Supplemental Materials. Raw sequence data generated in this study are publicly available. Sequencing data have been deposited in the DDBJ Sequence Read Archive (DRA) under accession numbers PRJDB40438 for scRNA-seq data from Meiosin-KO, Meiosin-ΔbHLH, and Stra8-null mice, and PRJDB40439 for zebrafish adult testes and ovaries. The PD15 WT testis scRNA-seq data set used in this study was derived from our previously published data set (Alavattam et al. 2024), deposited in the DDBJ Sequence Read Archive under BioProject accession PRJDB16643. Sequence data for the zebrafish Meiosin gene are deposited in the National Center for Biotechnology Information-National Institutes of Health (NCBI-NIH) GenBank under accession numbers LC649921.

The raw MS data generated in this study are publicly available under the accession numbers JPST004665 and PXD079165 for JPOST and proteomeXchange, respectively.

Materials and resource availability

Mouse lines generated in this study have been deposited to Center for Animal Resources and Development (CARD ID3012-3015 for Meiosin ΔbHLH mutant mouse, CARD ID2999 for Meiosin-3xFLAG-HA knockin mouse). The antibodies are available upon request. Antibody availability is limited by the absence of an external centralized repository for its distribution and by the need to maintain the stock. We are willing to share antibodies upon reasonable request, with the requester covering processing and shipping costs.

Competing interest statement

The authors declare no competing interests.