Mouse Chd4-NURD is required for neonatal spermatogonia survival and normal gonad development

Testis development and sustained germ cell production in adults rely on the establishment and maintenance of spermatogonia stem cells and their proper differentiation into spermatocytes. Chromatin remodeling complexes regulate critical processes during gamete development by restricting or promoting accessibility of DNA repair and gene expression machineries to the chromatin. Here, we investigated the role of Chd4 and Chd3 catalytic subunits of the NURD complex during spermatogenesis. Germ cell-specific deletion of chd4 early in gametogenesis, but not chd3, resulted in arrested early gamete development due to failed cell survival of neonate undifferentiated spermatogonia stem cell population. Candidate assessment revealed that Chd4 controls expression of dmrt1 and its downstream target plzf, both described as prominent regulators of spermatogonia stem cell maintenance. Our results show the requirement of Chd4 in mammalian gametogenesis pointing to functions in gene expression early in the process.


Introduction
Defects in gametogenesis are a leading cause of infertility and an important cause of birth defects associated with aneuploidy. Insights into the mechanisms underlying testis formation, including spermatogonia stem cell, are necessary to improve the outcomes of common gonad developmental diseases.
In mice, spermatogenesis begins from isolated germ cells called spermatogonia A-singles (As) that undergo to a series of mitotic divisions to produce spermatogonia known as paired (Apr) and aligned (Aal) that contains chains of 4 to 16 cells anchored by intercellular bridges as a result of incomplete cytokinesis [1,2]. At this point, spermatogonia cells start a process of differentiation [A1, A2, A3, A4 or intermediate (In)], formation of type B spermatogonia, and then transition to pre-leptotene cells that initiate the series of meiotic divisions that ultimately originate spermatozoa.
Chromatin undergoes extensive remodeling during gametogenesis, leading to altered gene expression and chromosome organization, and ultimately controlling obligatory developmental transitions, such as the conversion from undifferentiated to differentiated spermatogonia, spermatogonia commitment to meiosis, and meiotic progression [3,4]. These transitions are accompanied by changes in the structural properties of meiotic chromosomes (monitored by Hi-C), ultimately revealing how the chromosome structure influences fundamental meiotic processes, such as recombination and transcription [5]. The NURD (NUcleosome Remodeling and Deacetylase) is a prominent chromatin modifying complex that functions to control gene

Open Access
Epigenetics & Chromatin *Correspondence: pezzar@omrf.org 1 Cell Cycle and Cancer Biology Research Program, Oklahoma Medical Research Foundation, Suite B305. 825 NE 13th street, Oklahoma City, OK 73104, USA Full list of author information is available at the end of the article expression via chromatin remodeling and histone deacetylation [6,7]. The NURD complex contains two highly conserved and widely expressed catalytic subunits, Chd3/Mi-2α (chromodomain-helicase-DNAbinding 3) and Chd4/Mi-2β, which are members of the Snf2 superfamily of ATPases [6][7][8][9][10]. NURD plays a central role in various developmental and cellular events, such as controlling the differentiation of stem cells, maintaining cell identity, and responding to DNA damage [10][11][12]. In testis, Chd5 is required for normal spermiogenesis and proper spermatid chromatin condensation [13], while Chd3/4 has been described to localize at the X-Y pseudoautosomal region, the X centromeric region, and then spreads into the XY body chromatin [14,15]. Although the role of Chd5 has been well defined, the requirements and mechanisms of Chd4 and Chd3 in mouse gametogenesis and testis development is not completely understood. In a recent work, siRNA knockdown of Chd4 in primary cultures followed by spermatogonia transplantation revealed a loss of the regenerative capability of these cells. Interpretation of RNA-seq data obtained from spermatogonia siRNA treated versus control cell cultures revealed global transcription changes, including genes possibly involved in cell self-renewal [16]. Questions remain unanswered regarding the effect of chd3 and chd4 deletion in germ cells and testis development.
In this study, we report that testis-specific specific deletion of chd4 is essential for testis development and sustained germ cell production, while chd3 deletion results in no apparent phenotype. This germ cellspecific deletion of chd4 results in the developmental arrest of undifferentiated spermatogonia in neonatal mice progressing to a Sertoli-only phenotype. Our studies of selected Chd4 target genes and subsequent cytological and expression analysis show that Chd4 control dmrt1 gene expression and downstream targets, such as plzf. We propose these results suggest a possible mechanism by which Chd4 contributes to early germ cell development regulating genes that are required for survival/maintenance of spermatogonia cells.

Chd4 expression during mouse germ cell development
To investigate a potential role for Chd4 in germ cells, we assessed Chd4 expression in newborn and adult mouse testes by immunofluorescence. Plzf is expressed in undifferentiated spermatogonia [17,18]. We observed that Chd4 was highly expressed in spermatogonia cells (marked by Plzf, aka Zbtb16) and in Sertoli cells (marked by Sox9) (Fig. 1A and Additional file 1: Fig. S1). In agreement with a previous report [14], immunosignal of Chd4 was detected in late-pachytene stages (Fig. 1A, selected area on adult mouse top panel). In sum, Chd4 is detected in spermatogonia, Sertoli cells, and primary spermatocytes.
We confirmed that Chd4 is expressed at pre-meiotic stages of male gamete development by analyzing Chd4 protein levels by western blot in enriched fractions (see "Materials and methods" for details) of undifferentiated and differentiating spermatogonia (obtained from wild-type 7 dpp mice and using Thy1.2+ and c-Kit+ affinity columns, respectively) (Fig. 1B). Chd4 was also detected in the affinity column flow-through (FT, which was enriched mostly in Sertoli cells, here demonstrated by immunoblotting of Sox9). The level of cell population enrichment was assessed by western blot and markers specific for undifferentiated spermatogonia (Plzf ) and differentiating spermatogonia (Stra8) (Fig. 1B). In addition, we analyzed cell fractions enrichment by immunofluorescence (Additional file 2: Fig. S2).

Composition of Chd4-NURD complexes during spermatogenesis
NURD function is influenced by its subunit composition [19,20]. To determine whether the expression of NURD composition might change during spermatogenesis, we analyzed the levels of representative NURD subunits in enriched fractions of undifferentiated and differentiating spermatogonia, as well as in the flow-through (enriched in Sertoli cells) after spermatogonia enrichment. NURD subunits Hdac2a, Mta1, Rbbp4, Rbbp7 and Mbd2 were present in enriched fractions of undifferentiated (Thy1.2+) and differentiating (c-KIT+) spermatogonia (Fig. 1B).  To determine the composition of Chd4-NURD complexes, we used co-immunoprecipitation analysis to uncover NURD subunits that interact with Chd4 in enriched fractions of Thy1.2 and c-Kit spermatogonia cells together and the flow-through. The NURD subunits Hdac2a, Mta1, and Rbbp4/Rbbp7, but not Mbd2, coimmunoprecipitated with Chd4 from both spermatogonia and flow-through fractions (Fig. 1C). Although Mbd2 was enriched in FT (Sertoli) and reduced in spermatogonia fractions (Fig. 1B), we could not observe Mbd2 interaction with Chd4 by immunoprecipitation in any of these two fractions (Fig. 1C). This could be explained because of a transient interaction between Mbd2 and the Chd4-containing NURD complex. Hdac2a coimmunoprecipitated with Chd4 from wild type and chd3 −/− spermatogonia cells (Additional file 3: Fig. S3B). These data suggest that: (i) Chd4 forms a NURD complex independently of Chd3, (ii) that loss of Chd3 does not perturb Chd4-NURD complex formation in spermatogonia cells.

Fig. 2
Chd4 and chd3 gene targeting design and testis developmental defects in chd4 mutant mice. A, B Testis specific Cre knockout strategy for deletion of chd4 and chd3. See description in "Materials and methods". C H&E-stained paraffin testis sections of wild type, ddx4-chd4 −/− , and ddx4-chd3 −/− mice. D Quantification of testis weight for wild type and homozygous knockout mice is shown. Images and testis weigh measurements were obtained independently from three mice. E Chd3 transcription levels were measured in ddx4-chd3 −/− total testis and compared to wild-type littermates (2 months old mice, n = 3). Chd4 transcription level in enriched fractions of spermatogonia obtained from 7 dpp ddx4-chd4 −/− and control wild-type litter mate mice (mice, biological replicate n = 3). *** represents P < 0.0001 (two-tailed Student t test) de Castro et al. Epigenetics & Chromatin (2022) 15:16 Analysis of 2-month-old ddx4-chd4 −/− testes revealed a total loss of germ cells (marked by Tra98) in seminiferous tubules (Fig. 3B). No developing gametes were observed, including cell types at early stages (e.g., spermatogonia) (Fig. 3A, B). Sertoli cells develop normally in ddx4-chd4 −/− mice, consistent with the specific loss of Chd4 in germ cells at early stages of development. We did not observe differences in germ cell development between wild type and ddx4-chd4 wt/− , and between wild type and ddx4-chd3 −/− mice (Additional file 4: Fig. S4A, B), consistent with their similar testis sizes (Fig. 2C, D).
We conclude that deletion of Chd3 has no apparent effect on gamete development. However, germ cell specific deletion of Chd4 results in severe male and female germ cell developmental defects, possibly originated at premeiotic stages of development.

Chd4 is required for neonate spermatogonia survival
The severe phenotypes observed in ddx4-chd4 −/− mice (Figs. 2 and 3) prompted us to investigate spermatogonial differentiation during testis development in newborns. Testis sections from 9 dpp ddx4-chd4 −/− mice stained with H&E showed a markedly reduced number of germ cells (Fig. 4A) as well as differences in cell composition, compared to those from age-matched wild-type mice. To analyze this in detail, we examined the presence of cells and Tra98 (to mark germ cells). P pachytene cells, Se sertoli cells, Rs rounded spermatids. C H&E-stained histological sections of wild type and ddx4-chd4 −/− ovaries. F follicles, CL corpora lutea expressing Stra8 (Fig. 4A), which marks differentiating spermatogonia, Sycp3 and γH2AX which are markers of primary spermatocytes and Tra98, a marker for germ cells (Additional file 5: Fig. S5). Whereas tubules from 9 dpp wild-type mice contained cells expressing Tra98 (45 ± 10, n = 66 seminiferous tubules counted obtained from 3 mice) and Stra8 (18.8 ± 8.4, n = 36 obtained from 3 mice), tubules from ddx4-chd4 −/− mice showed a near absence of cells expressing these markers (Tra98 4.6 ± 3, n = 60 obtained from 3 mice, P < 0.0001, t test; Stra8 2.5 ± 3.5, n = 42 obtained from 3 mice, P < 0.0001, Student t test) (Fig. 4A). Testes sections from 9 dpp ddx4-chd4 −/− mice also showed a reduction in primary spermatocytes expressing the meiotic prophase I markers Sycp3 and γH2AX compared to those from 9 dpp wild-type mice (Additional file 5: Fig. S5A, B). Together, the results further suggest that testis defects in ddx4-chd4 −/− mice begin early, during pre-meiotic stages of postnatal development, leading to an absence of germ cells in adults.
To pinpoint when the testes defects originate in ddx4-chd4 −/− mice, we stained testes sections from 1 to 21 dpp mice for the expression of Tra98 (all germ cells). We observe that as general trend, in all analyzed stages, the number of Tra98 positive cells was reduced in ddx4-chd4 −/− mice compared to wild type, progressing to total absence of germ cells (Fig. 4B, C and Additional file 6: Fig. S6). Both Plzf-positive and Tra98-positive cells were substantially reduced in chd4 −/− testis compared to wildtype testis at 4 dpp and 7 dpp (Fig. 4C). We observed equal numbers of Sox9-positive (Sertoli) cells in testes from wild type and ddx4-chd4 −/− mice at 3, 4, and 7 dpp (Fig. 4C), as expected for the specific loss of Chd4 in spermatogonia cells.
Given that Chd4 may act as a regulator of cell-cycle progression, we then examined whether the rapid loss of Plzf-positive neonate spermatogonia in ddx4-chd4 −/− testes was due to altered proliferative activity. We conducted EdU incorporation study to test this possibility. 4 dpp mice were injected with EdU and analyzed 3 h later, after which we assayed its incorporation in Plzf-positive spermatogonia in whole mounts of seminiferous tubules (Additional file 7: Fig. S7A). We found that spermatogonia cell proliferation (Plzf/EdU + cells) in wild type and ddx4-chd4 −/− is proportionally the same (Additional file 7: Fig. S7C). In addition, reduced amount of total Plzf-positive cells was found in ddx4-chd4 −/− compared to wild type in the whole-mounting experiment (Additional file 7: Fig. S7B).
To determine whether cell death contributed to the loss of ddx4-chd4 −/− neonate spermatogonia (Fig. 4C), we performed TUNEL assay in paraffin embedded testis sections of wild type and ddx4-chd4 −/− 4 days old mice. At this age the testis is mostly constituted by spermatogonia and Sertoli cells, which can be easily distinguished by DAPI nuclear staining patterns. We found a significant increase in the percentage of apoptotic germ cells, but not Sertoli cells in ddx4-chd4 −/− testis compared to wildtype mice (Fig. 4D).
We conclude that the possible cause of spermatogonia failure in ddx4-chd4 −/− mice is in the survival/maintenance of neonate undifferentiated spermatogonia.

Genome wide Chd4 chromatin binding
To further investigate the mechanism of Chd4 requirement in neonate spermatogonia, we aimed to identify genes that directly interact with Chd4. As an approximation, we generated genome-wide chromatin-binding profiles of Chd4 by ChIP-seq in whole 7 dpp testis and in enriched fractions of spermatogonia from 7 dpp wildtype testes (Thy1.2+ and c-Kit+ cells) (Fig. 5). Heatmaps and direct visualization of profiles revealed a prominent enrichment of Chd4 in gene regulatory elements (Fig. 5A, B and Additional file 8: Table S1), suggesting that Chd4 regulates gene transcription in neonate spermatogonia. We observed a good correlation between both ChIP-seq profiles, with most peaks obtained from enriched fractions of spermatogonia from 7 dpp wild-type testes also present in chromatin-binding profiles from whole 7 dpp testis. To gain greater insight into the genes potentially regulated by Chd4, we annotated shared peaks of Chd4 to the closest gene and performed functional analysis using DAVID platform [22]. Among the most significantly enriched terms obtained using Genome Ontology Biological Processes (GO-BP, FDR < 0.05) we found: regulation of transcription, regulation of DNA replication, chromatin organization/remodeling, and regulation of apoptotic processes (Additional file 9: Table S2). On an attempt to narrow down genes that could be directly regulated by Chd4, and that may explain the observed cellular phenotype, we search for genes related to stem cell survival/differentiation with a Chd4 peak near the promoter. We found Dmrt1, a known gene acting in spermatogonia cell maintenance/survival [23]. We noted other genes with recognized spermatogonia function, such as Foxo3, involved in cell self-renewal and differentiation [24]; Rest, involved in survival of PGCs [25]; and Mettl14, which inactivation causes depletion of SSCs possibly by dysregulation of transcripts required for spermatogonia proliferation/differentiation [26] (Additional file 10: Table S3).

Chd4, Dmrt1, and Plzf work together in a regulatory axis involved in spermatogonia cell survival
Our results show that Chd4 is required for spermatogonia maintenance/survival. We then reason that Chd4 may interact with genes that have been described to work in spermatogonia maintenance. Indeed, Dmrt1 has been show to function in spermatogonia stem cells maintenance, and this function seems to be mediated by direct regulation of plzf expression, another transcription factor required for spermatogonia maintenance [23]. To test our hypothesis, we first performed dmrt1 RT-PCR analysis in enriched fractions of spermatogonia using an exon 4-5 specific probe. We observed that dmrt1 expression was substantially reduced in chd4 knockout testis compared to wild type (Fig. 5E). We also immunostained paraffined testes sections from 1, 4, and 7 dpp mice for the presence of Dmrt1 (Fig. 6A, B). Compared to wild type, ddx4-chd4 −/− mutant showed a clear reduction in Dmrt1 immunosignal.
Recent work described that Dmrt1 controls plzf expression, which is a transcription factor required for spermatogonia maintenance [23]. We then tested the effect of chd4 depletion in Plzf expression. Plzf immunostaining of testes sections from 1, 4, and 7 dpp mice revealed a significantly reduction of this protein in chd4 knockout cells with respect to wild type (Fig. 6C, D). This is consistent with a model by which Chd4 control of dmrt1 and its downstream targets influences cell survival/maintenance (Fig. 6E). We concluded that Chd4 participates in the maintenance/survival of neonate spermatogonia stem cell possibly through transcriptional regulation of genes participating in these critical processes. We note, however, that the dramatic phenotype observed in ddx4-chd4 −/− spermatogonia likely reflect CHD4 targeting a wide spectrum of genes participating in different pathways.

Discussion
In this work, we examined the potential function of two critical NURD catalytic subunits, Chd4 and Chd3, in spermatogenesis. Our data suggest that a Chd4-NURD (but not Chd3-NURD) complex controls neonate spermatogonia development at early stages of testis development. Germline deletion of chd4, but not chd3, results in a severe loss of germ cells specifically at early stages of the testis cord development. Chd4 deletion affects spermatogonia, with the first obvious consequences in undifferentiated spermatogonia. Our results agree with a recent report, in which using an in vitro approach, knockdown of Chd4 in spermatogonia cultures followed by cell transplantation in cell-ablated recipient testis, resulted in reduced viability of undifferentiated spermatogonia [16].
Most CHDs are expressed in testis; however, their insertion into the NURD complex seems to be developmentally regulated, with apparent different patterns of expression during gametogenesis. Chd5 [8][9][10] has been shown to be expressed and required to compact chromatin in postmeiotic stages of spermatogenesis [13,27]  results show that Chd4 is expressed and functions at premeiotic stages of gametogenesis. The functions of Chd4-NURD in neonate spermatogonia development and male gametogenesis are further revealed by our ChIP-seq (Fig. 5) and cytological analysis showing that Chd4 binds to the promoter of dmrt1 and regulates its expression (Figs. 5 and 6). Dmrt1 function in maintenance of spermatogonia stem cells has been proposed to be mediated by direct regulation of Plzf gene expression, another transcription factor required for spermatogonia maintenance [23]. In agreement with this possibility, we observed that the amount of Plzf was significantly reduced in chd4 knockout versus wild-type cells. In sum, our work provides evidence of a regulatory axis in which Chd4 may control important genes involved in spermatogonia stem cell maintenance and survival. The effect we observed with in vivo Chd4 deletion on dmrt1 expression in spermatogonia (and possibly downstream genes, such as sohlh1) is in contrast with a recent report showing an increase in dmrt1 and sohlh1 expression measured by single cell RNA-seq in chd4 siRNA knockdown cultured spermatogonia cells [16]. These contrasting results, and some of their implications, such as on the primary function of Chd4 as an activator of genes involved in spermatogonia survival, may be assigned to differences in the experimental system or experimental approaches utilized. Additional work will be required to test sohlh1 expression level in Chd4 knockout spermatogonia, another dmrt1 direct target involved in cell survival and differentiation [23,28,29], as well as in the stra8 gene, the latter which precocious expression is detected in dmrt1 knockout mice. We note that the dramatic cellular phenotype we observed after Chd4 depletion in spermatogonia may be more accurately explained by Chd4 activity targeting several genes and different pathways.
All experiments conformed to relevant regulatory standards guidelines and were approved by the Oklahoma Medical Research Foundation-IACUC (Institutional Animal Care and Use Committee).

Real-time PCR
Total RNA was isolated from adult testis or from enriched fractions of spermatogonia with the Direct-zol RNA MiniPrep Plus kit (Zymo Research). RNA (2.0 μg) was oligo-dT primed and reverse-transcribed with the high-capacity RNA-to-cDNA kit (Applied Biosystems). Exon boundaries of chd4 and chd3 were amplified using TaqMan Assays (Applied Biosystems) as directed by the manufacturer using Beta-2 macroglobulin as standard. TaqMan Mm01190896_m1 (chd4), Mm01332658_m1 (chd3), Mm00437762_m1 (Beta-2 microglobulin), and Mm00443809_m1 (dmrt1). Gene expression was normalized with respect to wild type with wild-type expression levels considered to be 1.

Western blot cell lysates
Total testis or enriched cells fractions were lysed in icecold protein extraction buffer containing 0.1% Nonidet P-40, 50 mM Tris-HCl, pH 7.9, 150 mM NaCl, 3 mM MgCl 2 , 3 mM EDTA, 10% glycerol, 1 mM DTT, 1 mM PMSF and protease inhibitors (ThermoFisher Scientific, A32965) followed by sonication (3 pulses of 10 s) using micro ultrasonic cell disrupter (Kontes). The relative amount of protein was determined measuring absorbance at 260 nm using NanoDrop 2000c spectrophotometer (ThermoFisher Scientific). Proteins were solubilized with 2× sample buffer (4% SDS, 160 mM Tris-HCl, pH 6.8, 20% glycerol, 4% mM β-mercaptoethanol, and 0.005% bromophenol blue) and 30 µg/lane of sample were separated by 4-15% gradient Tris-acetate SDS-PAGE and electro transferred to PVDF membrane (Santa Cruz Biotechnology, sc-3723). The blots were probed with individual primary antibodies, and then incubated with HRP-conjugated goat anti-mouse or rabbit antibody as required. In all blots, proteins were visualized by enhanced chemiluminescence, and images acquired using Western Blot Imaging System c600 (Azure Biosystems). ImageJ software were used for quantification of non-saturated bands and α-tubulin were used for normalization. Antibodies used are detailed in Additional file 11: Table S4.

Histology and immunostaining
Testes and ovaries were dissected, fixed in 10% neutralbuffered formalin (Sigma) and processed for paraffin embedding. After sectioning (5-8 µm), tissues were positioned on microscope slides and analyzed using hematoxylin and eosin using standard protocols. For immunostaining analysis, tissue sections were deparaffinized, rehydrated and antigen was recovered in sodium citrate buffer (10 mM Sodium citrate, 0.05% Tween 20, pH 6.0) by heat/pressure-induced epitope retrieval. Incubations with primary antibodies were carried out for 12 h at 4 °C in PBS/BSA 3%. Primary antibodies used in this study are detailed in Additional file 11: Table S4, following three washes in 1× PBS, slides were incubated for 1 h at room temperature with secondary antibodies. A combination of fluorescein isothiocyanate (FITC)-conjugated goat anti-rabbit IgG (Jackson laboratories) with Rhodamine-conjugated goat anti-mouse IgG and Cy5-conjugated goat anti-human IgG each diluted 1:450 were used for simultaneous triple immunolabeling. Slides were subsequently counterstained for 3 min with 2 µg/ml DAPI containing Vectashield mounting solution (Vector Laboratories) and sealed with nail varnish. We use Zen Blue (Carl Zeiss, inc.) for imaging acquisition and processing.

Enrichment of spermatogonia populations
Our procedure of cell enrichment followed [31,32]. Briefly, testis from 7 dpp mice (or any other indicated age) were removed from mice and placed in a Petri dish containing Dulbecco's Modified Eagle Medium (DMEM without phenol red). After detachment of the tunica albuginea, the seminiferous tubules were loosen using forceps and incubated in a 15 mL tube containing DMEM containing 1 mg/mL of collagenase, 300 U/ mL of hyaluronidase and 5 mg/mL DNAse I (StemCell Technologies) under gentile agitation for 10 min. The seminiferous tubule clumps were pelleted by gravity and the cell suspension containing interstitial cells was discarded. The tubules were then incubated of with 0.05% Trypsin-EDTA solution (Mediatech Inc) for 5 min and the reaction was stopped by adding 10% volume of 10% BSA in PBS. Single cell suspension was obtained by mechanical resuspension followed by filtration through a 40-µm-pore-size cell strainer and dead cells were removed using Dead Cell Removal Kit (Miltenyi Biotec 130-090-101). Differentiating c-KIT+ neonate spermatogonia cells were magnetically labeled with CD117 (c-KIT+) MicroBeads (Miltenyi Biotec 130-091-224) and isolated using MS columns (Miltenyi Biotec 130-042-201) according to manufacturer's instructions. After the depletion of the c-KIT+ cells, the population of undifferentiated neonate spermatogonia cells were separated using CD90.2 (THY1.2+) MicroBeads (Miltenyi Biotec 130-121-278). Relative enrichment of cell populations was evaluated by STRA8 (c-Kit fractions) or PLZF (THY1.2 fractions) western blots (Fig. 1B). After c-kit and THY1.2 separation, the flow-through mostly contained Sertoli cells (SOX9 positive, Additional file 3: Fig.  S3). The number of cells obtained from a pool of 4 mice testis at 7 dpp was approximately 3.43 × 10 5 in THY1.2 fractions and 5.71 × 10 5 in c-Kit fractions.

Primary spermatocyte enrichment
Synchronized pachytene spermatocytes from the first spermatogenic wave were purified as described in [33]. Briefly, 2 dpp mice were injected for 7 consecutive days with WIN 18,446 to arrest germ cells as spermatogonia. The next day (9 dpp), mice were injected with retinoic acid (RA) to induce their coordinated maturation. 2 Mice were killed at 13 days after RA injection (22 dpp). Testes were disaggregated using proteases and meiocytes were then purified by fluorescence activated cell sorting (FACS). Unlike Romer et al., after testes disaggregation and prior to FACS, cells were not washed (by centrifugation and resuspension) to avoid breakage of fragile pachytene cells. Instead, cells were sorted from the protease-containing buffer. Purity and stage of cells was assessed by immunofluorescence of chromosome spreads using anti SYCP1 and anti SYCP3 antibodies. More than 80% of cells were at pachytene stage.

Immunoprecipitation
Co-immunoprecipitation experiments were performed using testis of wild type or ddx4-chd3 −/− mouse (adult-2 months). After detunication, seminiferous tubules were loosen using forceps, washed twice with cold 1× PBS and lysed using ice-cold RIPA buffer (50 mM Tris pH 7.5, 150 mM NaCl, 1% NP40, 0.5% Deoxycholate) containing protease inhibitors (ThermoFisher Scientific, A32965), sheared using 23 G needle, incubated on ice for 15 min and centrifugated at 1000×g for 10 min at 4 °C. Supernatant were collected in a separate tube, the pellet was resuspended in RIPA buffer, disrupted by sonication (3 pulses of 10 s) and centrifuged 12,000×g. This second supernatant was combined with the previous one and protein concentration was determined. We used 1 mg of protein for each immunoprecipitation. Lysates were pre-cleared with protein G magnetic beads (BioRad, 161-4023) for 1 h at room temperature and incubated with rabbit anti-Chd3 (5 μg, Bethyl A301-220A), rabbit anti-Chd4 (2 μg, Abcam ab72418), or rabbit IgG (5 μg Jackson ImmunoResearch, 011-000-003). Lysates were rotated overnight at 4 °C and immune complexes were collected with protein G magnetic beads (2 h at 4 °C). Beads were washed 4 times with washing buffer (50 mM Tris pH 7.5, 150 mM NaCl, 0.1% TX100, 5% glycerol) and two times with PBS. Proteins were eluted by boiling the beads with 2× sample buffer and analyzed by SDS-PAGE as described above.

EdU-based proliferation assay
Mice at indicated age received subcutaneous injection of EdU (50 mg/kg) (Invitrogen, A10044) 3 h prior euthanasia. After that, testes were removed and processed for whole-mount immunohistochemistry. EdU was detected by incubation of testis samples with reaction mix (2 mM CuSO 4 , 50 mM ascorbic acid and 2 mM Alexa Azide conjugates (488 or 647) in PBS) for 3 h at room temperature.

Whole-mount seminiferous tubules
Immunohistochemistry of whole-mount seminiferous tubules was performed as described [34]. Briefly, after detachment of the tunica albuginea, the seminiferous tubules were loosen using forceps and incubated in a 15 mL tube containing DMEM containing 1 mg/mL of collagenase, 300 U/mL of hyaluronidase and 5 mg/mL DNAse I (StemCell Technologies) under gentile agitation for 10 min. The seminiferous tubules clumps were pelleted by gravity and the cell suspension containing interstitial cells was discarded. Seminiferous tubules were fixed for 4 h in 4% PFA (pH7.2 in PBS) at 4 °C. After extensively wash in PBS, the tubules were permeabilized with series of MeOH/PBS (25, 50, 75, 95%, and twice in 100% MeOH) for 15 min at room temperature, treated with MeOH: DMSO: H 2 O 2 (4:1:1), and rehydrated with MeOH/PBS (50, 25% and twice in PBS). Samples were incubated in ice-cold blocking solution PBSMT (PBS with 2% non-fat dry milk and 0.5% triton X-100) for 3 h and then over-night at 4 °C with indicated primary antibodies under gentle rotation. Seminiferous tubules were washed in PBSMT (5 × 1 h) and incubated with dye conjugated (Alexa488 or TRITC) goat anti-mouse or rabbit antibody as required. The tubules were mounted in raised coverslips glass slides.

Library preparation and sequencing
Spermatogonia-purified ChIP and its input libraries were prepared using an in-house method (see below), and sequenced single read 85 bp on an Illumina Next-Seq 500 instrument. Whole testes ChIP and its input libraries were prepared using ACCEL-NGS ® 1S PLUS DNA LIBRARY KIT (cat. # 10024) in conjunction with Swift unique dual indexing kit (cat. # X9096), following manufacturer's instructions, and sequenced as paired end 150 bp on an Illumina Novaseq 6000 instrument. Both sequencings were done at OMRF Clinical Genomics Core.
Before sequencing, samples were quantified by qPCR using Kapa library quantification kit (cat # KK4854), and size and quality of DNA were assessed using Agilent Tape station. 20  were diluted according to the DNA concentration (insert:adaptor molar ratio of 1:2, with maximal dilution of 1:50) and 1 µL of a specific diluted adaptor was added to each sample. 20 µL of ligation mix (30 U/µL T4 DNA ligase (cat. # NEB M0202L), and 1.58× ligase buffer) was added to 10 µL of the insert:adaptor mix and incubated for 30 min at 20 °C. DNA was purified using Qiagen Minelute kit using 7 volumes of PB buffer and eluted in 12 µL of EB buffer. Finally, DNA was amplified using the Kapa HiFi Hot Start library amplification kit (cat. # KK2621) according to manufacturer's instructions.

ChIP-seq data processing and analysis Alignment and quality filtering
For samples prepared using adaptase technology (whole testes and its input), reads 1 and 2 were trimmed 10 bases at the beginning using fastp [36] (version 0.23.2), as recommended by the library preparation kit manufacturer. Further processing was done in parallel for all samples. Reads were adapter-trimmed and qualitypruned using fastp with default settings. Then, reads were aligned to mm10 genome using BWA-MEM [37] (version 0.7.15) with default settings except for option '-M' for Picard compatibility. Picard (version 2.21.2, http:// broad insti tute. github. io/ picard/) and SAMtools (version 1.11) [38] were used to obtain mapping quality metrics, remove duplicates and filter reads. Only primary alignment reads that were not duplicated, properly paired, with a MAPQ > 30, and not placed in mitochondrial or unplaced-chromosomes were kept.

Peak calling
Peaks were called using MACS2 [39] (version 2.2.7.1) using broad mode and default values for the rest of the parameters.

Peak annotation and gene ontology analysis
Common peaks to both ChIPs were filtered by blackgreylist and then annotated using HOMER's annotate-Peaks.pl script (22, version 4.10). To assign a score to the common peaks, we averaged the −log(q value) for both peaks. In cases in which a peak from one experiment was intersected by more than one peak from the other experiment, the −log(q values) from the same experiment were averaged first. Annotated peaks can be found at Additional file 8: Table S1. Functional annotation analysis was done using DAVID with default parameters and the list of annotated genes in which the promoter was closer than 5000 bp. Further processing and plotting were done using R (https:// www.r-proje ct. org/, version 4.1.1) and packages within tidyverse (https:// joss. theoj. org/ papers/ 10. 21105/ joss. 01686). Functional analysis results can be found at Additional file 12: Table S5.

Heatmaps and IGV snapshots
Bigwig coverage tracks were generated using bamCoverage tool from deepTools (version 3.4.3, [40]) with a bin size of 10 bp, 40 bp smoothing and RPKM normalization. Average aggregate profiles and heatmaps were plot using deepTools as well (computeMatrix and plotHeatmap). Mapping statistics can be found in Additional file 12: Table S5.

Statistical analyses
Results are presented as mean ± standard deviation (SD). Statistical analysis was performed using Prism Graph statistical software. Two-tailed unpaired Student's t test