The DOT1L-MLLT10 complex regulates male fertility and promotes histone removal during spermiogenesis

Spermatogenesis is a complex developmental process that produces sperm. It consists of three phases: mitotic, meiotic and haploid. During the mitotic phase, spermatogonial stem cells renew to sustain life-long production of sperm, then differentiate and enter meiosis. During meiosis, homologous chromosomes pair and undergo meiotic recombination. Meiotic germ cells undergo one round of DNA replication but two successive rounds of cell divisions, resulting in haploid round spermatids. During the haploid phase (i.e. spermiogenesis), round spermatids undergo dramatic structural changes to become spermatozoa, including nuclear elongation and condensation, histone-to-protamine replacement, formation of flagella and shedding of the cytoplasm.

Dramatic chromatin remodeling occurs during spermiogenesis. Most histones in spermatids are replaced by sperm-specific transition proteins and then by sperm-specific protamines (Rousseaux et al., 2008). Protamines are highly basic proteins of lower molecular weight than histones. As a result, the histone-to-protamine replacement causes much more compact chromatin in sperm. Global histone acetylation precedes histone replacement (Meistrich et al., 1992; Oliva and Mezquita, 1982). GCN5, a histone acetyltransferase, acetylates histones in spermatids (Luense et al., 2019). BRDT, a testis-specific bromodomain protein, binds to acetylated histones to induce reorganization of chromatin when ectopically expressed in somatic cells (Pivot-Pajot et al., 2003). Genetic studies have shown that both GCN5 and BRDT play a role in the histone-to-protamine replacement during spermiogenesis (Gaucher et al., 2012; Luense et al., 2019; Shang et al., 2007). Therefore, histone acetylation enhances chromatin accessibility and facilitates nucleosome eviction during spermiogenesis.

Like histone acetylation, histone methylations also play important roles in chromatin remodeling and gene expression. DOT1L is the sole methyltransferase for histone H3 at lysine 79 (H3K79) and is responsible for mono-, di and tri-methylation (Feng et al., 2002; Lacoste et al., 2002; Ng et al., 2002; van Leeuwen et al., 2002). Dot1, the yeast orthologue of DOT1L, methylates H3K79 in the context of nucleosomes and chromatin (Lacoste et al., 2002; van Leeuwen et al., 2002). In addition, ubiquitylation of H2B at lysine 120 directly stimulates intranucleosomal methylation of H3K79 by human DOT1L (McGinty et al., 2008). In budding yeast, Dot1 functions in telomere silencing and in meiotic checkpoint control (San-Segundo and Roeder, 2000). In mouse, DOT1L is essential for embryonic development and early erythropoiesis (Feng et al., 2010; Jones et al., 2008). In mouse embryonic stem (ES) cells, DOT1L is important for establishment of heterochromatin structures at telomeres, centromeres and pericentromeres (Jones et al., 2008; Malla et al., 2021 preprint). The phosphorylated C-terminal domain of RNA polymerase II recruits DOT1L, which leads to H3K79 methylations in actively transcribed genes (Kim et al., 2012). Several DOT1L-associated proteins have been identified in cancer cells (Vlaming and van Leeuwen, 2016). In MLL leukemia, DOT1L interacts with AF9, MLLT10 (i.e. AF10) and ENL. MLLT10 regulates the H3K79 methyltransferase activity in leukemic cells. The fusion of MLL-MLLT10 leads to malignant transformation of myeloid progenitor cells and upregulation of HoxA9 in leukemia (Deshpande et al., 2014; Okada et al., 2005). Therefore, DOT1L plays important roles in development and leukemic transformation.

Mouse DOT1L is essential for self-renewal of spermatogonial stem cells (Lin et al., 2022). Inhibition of DOT1L leads to downregulation of HoxC cluster transcription factors in spermatogonial stem cells (Lin et al., 2022). In contrast, oocyte-specific inactivation shows that DOT1L is dispensable for oogenesis and that maternal DOT1L is not required for mouse development (Liao and Szabo, 2020). In mouse spermatocytes, DOT1L and H3K79 methylation exhibit dynamic localization patterns (Ontoso et al., 2014). Intriguingly, both DOT1L and H3K79me2 are abundant in diplotene spermatocytes; however, DOT1L accumulates in the XY body but H3K79me2 mark is excluded from the XY body, suggesting that DOT1L may play a role in male meiosis in mouse (Ontoso et al., 2014). Here, we report that DOT1L forms a complex with MLLT10 in mouse testis. DOT1L and MLLT10 localize to the XY chromatin in late meiotic spermatocytes and early round spermatids in an inter-dependent manner. Conditional inactivation of either Dot1l or Mllt10 results in depletion of H3K79me2 in germ cells, reduced sperm count and decreased male fertility. We also find that H3K79me2 is abundant in round and elongating spermatids. Unexpectedly, our study shows that histones are retained in Dot1l-deficient or Mllt10-deficient epididymal sperm, suggesting that H3K79 methylation promotes histone-to-protamine replacement during spermiogenesis.

To enable systematic identification of DOT1L-associated proteins in testis, we generated Dot1l-3×FLAG mice using the CRISPR/Cas9 genome editing approach (Fig. S1). The adult homozygous FLAG-tagged mice were fertile, showing that the C-terminal FLAG tagging did not affect the DOT1L function. For immunoprecipitation (IP) with anti-FLAG antibody, protein extracts were prepared from wild-type (no FLAG tag) and Dot1l-3×FLAG testes at postnatal day 26, which contain spermatogonia, spermatocytes and round spermatids. Immunoprecipitated proteins were identified by mass spectrometry. As expected, DOT1L had the highest number of peptides identified in the DOT1L-3×FLAG testis IP (Fig. 1A). In addition to DOT1L, MLLT10 and MLLT6 ranked second and third, respectively, in the number of peptides in the DOT1L-3×FLAG testis IP but had no peptides in the wild-type testis IP (Fig. 1A). MLLT10 is a known DOT1L-binding protein in leukemic cells (Deshpande et al., 2014; Okada et al., 2005). MLLT6 (also known as AF17) competes with MLLT3 (also known as AF9) for binding to DOT1L in epithelial cells (Reisenauer et al., 2009). However, none of these MLLT proteins has been studied in germ cells. To validate the DOT1L-MLLT10 interaction, we first performed co-IP in testis with the anti-FLAG antibody followed by western blotting analysis (Fig. 1B). As expected, the DOT1L-3×FLAG was immunoprecipitated with the anti-FLAG antibody. MLLT10 was co-immunoprecipitated only in the DOT1L-3×FLAG testis IP, demonstrating that MLLT10 is associated with DOT1L in testis (Fig. 1B). Reciprocal co-IP confirmed the DOT1L-MLLT10 association in testis (Fig. 1C).

We next examined the expression of DOT1L and MLLT10 in mouse tissues (Fig. 1D). Surprisingly, DOT1L was abundant in testis but not detected in other somatic tissues that were examined. Likewise, MLLT10 was abundant in testis and present at a low level in brain, but not in other tissues. In contrast, H3K79me2 was present in all tissues. We next investigated their expression in postnatal developing testes (Fig. 1E). In the first wave of spermatogenesis after birth, different germ cells appear at defined ages (Fig. 1E). DOT1L was present at low levels at postnatal days 7, 15 and 21, but its abundance increased at postnatal day 28, when elongated spermatids first appeared. The abundance of MLLT10 was sharply increased at postnatal day 21, when round spermatids first appeared. These results suggest that DOT1L and MLLT10 may play an important role in germ cells of late stages during spermatogenesis.

We investigated the expression and subcellular localization of DOT1L and MLLT10 by immunofluorescence of testis sections (Fig. 2A). Consistent with our western blot data in juvenile testes (Fig. 1E), DOT1L and MLLT10 were expressed in late spermatocytes and early spermatids. DOT1L was abundant in the nuclei of diplotene spermatocytes, metaphase spermatocytes, secondary spermatocytes, round spermatids and step 9 elongating spermatids (Fig. 2B). DOT1L was not detected in spermatids at step 11 and beyond. The MLLT10 expression pattern mostly overlapped with DOT1L, except that MLLT10 was also abundant in mid- to late-pachytene spermatocytes (Fig. 2B). H3K79me2 was present from late pachytene spermatocytes through step 11 elongating spermatids. Notably, the H3K79me2 signals increased in intensity in elongating spermatids (step 9-11) and then disappeared abruptly in condensing spermatids (step 12) (Fig. 2B,C). Sex chromosomes form the XY body and are γH2AX positive in pachytene and diplotene spermatocytes. Strikingly, although DOT1L did not accumulate in the XY body in pachytene spermatocytes, it accumulated in the XY body in diplotene spermatocytes. DOT1L was also concentrated in a peri-chromocenter area in round spermatids, which had a characteristic intermediate DAPI staining (weaker than the chromocenter) and was previously shown to be the sex chromosome (Namekawa et al., 2006). MLLT10 began to accumulate on the XY body in late pachytene cells and had the same localization pattern on sex chromosomes in diplotene cells and round spermatids as DOT1L. In metaphase I spermatocytes, DOT1L and MLLT10 localized to one pair of chromosomes, which were likely to be sex chromosomes. Nuclear spread analysis confirmed the localization pattern of DOT1L and MLLT10 in spermatocytes (Fig. S2A,B). In secondary spermatocytes, DOT1L and MLLT10 also occupied one subnuclear region, which likely corresponds to sex chromosomes (Fig. 2B). Unlike DOT1L and MLLT10, H3K79me2 was excluded from the XY body in diplotene cells (Fig. 2B, Fig. S2C). However, like DOT1L and MLLT10, H3K79me2 also had increased signals in the peri-chromocenter region (sex chromosome) in round spermatids. The largely overlapping localization patterns of DOT1L and MLLT10 in male germ cells were consistent with the DOT1L-MLLT10 interaction in testis. These results suggest a possible role for DOT1L and MLLT10 in meiosis and spermiogenesis.

We have previously demonstrated that DOT1L is essential for self-renewal of spermatogonial stem cells by inactivating Dot1l using the Ddx4-Cre (Lin et al., 2022). To circumvent the loss of spermatogonial stem cells, we inactivated Dot1l using the Stra8-Cre, which is expressed in differentiating spermatogonia before meiotic entry (Lin et al., 2017). We generated Dot1l^fl/−Stra8-Cre conditional knockout (referred to as Dot1l^cKO) mice. We found that the adult Dot1l^cKO testis was smaller (Fig. 3A). The testis weight of adult Dot1l^cKO males was significantly reduced (Fig. 3B). The sperm count of Dot1l^cKO males was sharply decreased (Fig. 3C). Mating tests showed that, out of six Dot1l^cKO males tested, two males were infertile and the remaining four males produced fewer and smaller litters (Fig. 3D,E). Western blot analysis revealed that DOT1L was absent in Dot1l^cKO testes (Fig. 3F). H3K79me2 was substantially lower in Dot1l^cKO testes than in Dot1l^fl/+ testes but not absent (Fig. 3F). Immunofluorescence of testis sections showed that the focal localization of DOT1L in metaphase spermatocytes was lost in Dot1l^cKO testes (Fig. S3A). H3K79me2 was absent in germ cells but still present in somatic cells in Dot1l^cKO testes, as expected (Fig. S3B). Histological analysis of Dot1l^cKO testes revealed no spermatogenic arrest (Fig. 3G). In addition, Dot1l^cKO testes lacked defects in self-renewal of spermatogonial stem cells. This was expected, because Stra8-Cre was not expressed in spermatogonial stem cells. On closer examination by histology, we observed an increase in apoptotic spermatocytes in stage XII tubules from Dot1l^cKO testes (Fig. 3G). Indeed, TUNEL assays showed a significant increase in the number of apoptotic spermatocytes in stage XII tubules from Dot1l^cKO testes (Fig. 3H). Therefore, the increased apoptosis in spermatocytes at stage XII could be at least partially responsible for reduced sperm production in Dot1l^cKO males.

We examined meiotic progression in Dot1l^cKO testes. The composition of prophase I spermatocytes was comparable between Dot1l^cKO and control testes (Fig. S4A). We did not observe defects in chromosomal synapsis in Dot1l^cKO spermatocytes. RPA, a single-strand DNA-binding heterotrimer (RPA1, RPA2 and RPA3), localizes to DSBs (Shi et al., 2019). The number of RPA2 foci was similar between Dot1l^cKO and control spermatocytes (Fig. S4B). We next examined the XY body. SCML2 localizes to the XY body in spermatocytes (Hasegawa et al., 2015; Luo et al., 2015). SCML2 still localized to the XY body in both pachytene and diplotene spermatocytes from Dot1l^cKO testes (Fig. S5A). H3K9me2, a silencing histone modification, accumulated at the XY body in both Dot1l^cKO and control spermatocytes (Fig. S5B). As in control spermatocytes, H3K4me3 and RNA polymerase II were excluded from the XY body in Dot1l^cKO pachytene and diplotene spermatocytes, showing that meiotic sex chromatin inactivation (MSCI) remained intact (Fig. S5C,D). These analyses showed that meiotic progression was apparently normal in Dot1l^cKO testes.

Global deletion of Mllt10 causes embryonic lethality (Deshpande et al., 2014). To study its role in germ cells, we generated Mllt10^fl/−Ddx4-Cre conditional knockout (referred to as Mllt10^cKO) mice. Ddx4-Cre begins to be expressed in germ cells at embryonic day 15 (E15) (Gallardo et al., 2007). Therefore, Mllt10 was expected to be deleted in all germ cells, including spermatogonia, in postnatal Mllt10^cKO testis. We found that adult Mllt10^cKO testis was slightly smaller than Mllt10^fl/− testis (Fig. 4A). Western blot analysis showed the absence of MLLT10 in Mllt10^cKO testis (Fig. 4B). Immunofluorescence showed that MLLT10 signals were absent in Mllt10^cKO testis (Fig. S3C). H3K79me2 was absent in germ cells but still present in somatic cells in Mllt10^cKO testis, demonstrating that MLLT10 is required for the methyltransferase activity of DOT1L in germ cells (Fig. S3D). The testis weight of Mllt10^cKO mice was lower than that of control (Fig. 4C). The sperm count of Mllt10^cKO males was reduced by 50% (Fig. 4D). Mating tests showed that although the number of litters was the same, the size of litters sired by Mllt10^cKO males was significantly lower (Fig. 4E,F). In conclusion, loss of MLLT10 results in subfertility in males.

Histological analysis of adult Mllt10^cKO testes did not reveal spermatogenic block or an apparent loss of germ cells (Fig. 4G). Germ cell-specific deletion of Dot1l using the Ddx4-Cre causes a failure in self-renewal of spermatogonial stem cells, resulting in Sertoli cell-only tubules in testes as early as postnatal day 40 (Lin et al., 2022). However, Sertoli cell-only tubules were not observed in Mllt10^cKO testes up to the age of 4 months, showing that MLLT10, unlike DOT1L, is dispensable for self-renewal of spermatogonial stem cells. Furthermore, there was no increase in apoptosis of spermatocytes in stage XII tubules or other tubules from Mllt10^cKO testes (Fig. 4H,I). We monitored chromosomal synapsis, XY body formation and meiotic DSB formation by surface nuclear spread analysis of spermatocytes (Fig. S6). The composition of prophase I spermatocytes was similar between Mllt10^fl/− and Mllt10^cKO testes (Fig. S6A). The XY body was formed in Mllt10^cKO pachytene and diplotene cells, evidenced by the localization of γH2AX and H3K9me2 (Fig. S6C,D). Chromosomal synapsis and formation of meiotic DSBs appeared to be normal in Mllt10^cKO pachytene cells (Fig. S6B,E). These analyses show that MLLT10 is dispensable for meiosis.

DOT1L and MLLT10 form a complex in testis (Fig. 1). We sought to address whether their abundance is dependent on each other by western blotting analysis in knockout testes. MLLT10 was readily detected in Dot1l^cKO testes but at a lower abundance than in Dot1l^fl/+ testes (Fig. 5A). The DOT1L abundance was much lower in Mllt10^cKO testes than in Mllt10^fl/+ testes (Fig. 5B). The H3K79me2 levels were reduced in both Dot1l^cKO testes and Mllt10^cKO testes (Fig. 5A,B). These data show that DOT1L and MLLT10 stabilize each other in testes and that MLLT10 also plays a role in facilitating DOT1L in H3K79 methylation in germ cells.

We next examined whether the subcellular localization of MLLT10 was affected in Dot1l^cKO testes (Fig. 5C). MLLT10 accumulated on the sex chromosomes (XY body) in diplotene cells and formed one large blob (presumably sex chromosomes) in wild-type metaphase spermatocytes, secondary spermatocytes and round spermatids (Figs 2 and 5C). Notably, MLLT10 failed to localize to sex chromosomes in spermatocytes and round spermatids from Dot1l^cKO testes (Fig. 5C). DOT1L had a similar accumulation pattern on the sex chromosomes in wild-type spermatocytes and round spermatids, but failed to localize to the sex chromosomes from Mllt10^cKO testes (Fig. 5D). Therefore, these results demonstrate the inter-dependent localization of DOT1L and MLLT10 at the sex chromosomes from diplotene spermatocytes to round spermatids.

While performing sperm count, we noticed abnormal morphology in sperm from the Dot1l^cKO cauda epididymis (Fig. 6A,B). We observed three types of sperm in the Dot1l^cKO males. Although morphologically normal sperm (type 1) were present, some sperm (type 2) had head foldback onto the flagella and a minority of sperm had coiled flagella (type 3) (Fig. 6A,B). The normal sperm (type 1) and head foldback sperm (type 2) each accounted for about 50% of the sperm (Fig. 6C). The type 2 sperm appeared to have a round head morphology under bright field illumination. However, DAPI staining revealed that type 2 sperm had nuclear morphology similar to normal sperm (Fig. 6A,B). Type 3 sperm also had apparently typical nuclear morphology. Under a dissecting microscope, all three types of sperm from the Dot1l^cKO males appeared motile.

Most histones are replaced by protamines in sperm. H3K79me2 reached the highest abundance in elongating spermatids (step 9-11), in which histone-to-protamine replacement occurs (Fig. 2). Therefore, we sought to address whether histone-to-protamine exchange was affected in the Dot1l^cKO testis. Western blot analysis showed that histone H3 was substantially retained in epididymal sperm from the Dot1l^cKO males in comparison with the Dot1l^fl/+ (control) males (Fig. 6D). In addition, the abundance of histone H3 increased in epididymal sperm from Mllt10^cKO males (Fig. 6E). These results demonstrate that DOT1L and MLLT10 promote histone replacement during spermiogenesis.

Dramatic chromatin reorganization occurs during spermiogenesis. A major event is the replacement of most histones by protamines in elongating spermatids. Histone acetylation precedes the histone-to-protamine exchange in spermatids (Meistrich et al., 1992; Oliva and Mezquita, 1982). GCN5, a histone acetyltransferase, and BRDT, an acetylated histone reader, function in histone exchange in late spermatids (Gaucher et al., 2012; Luense et al., 2019; Shang et al., 2007). BRD4, another acetylated histone reader, localizes as a ring structure under the acrosome and has been postulated to facilitate histone removal (Bryant et al., 2015). The bromodomain in BRDT and BRD4 recognizes acetylated histones. In this study, we found that H3K79me2 reached the highest abundance in elongating spermatids from step 9 to 11. Histone H3 was retained in epididymal sperm from either DOT1L- or MLLT10-deficient mice. Our result is consistent with the histone retention in spermatids from DOT1L conditional knockout testes as reported by Malla et al. (2023). Therefore, in addition to histone acetylation, histone H3K79 methylation also plays an important role in histone removal in spermatids. Histone acetylation and histone H3K79 methylation are associated with active gene transcription and thus are features of open chromatin. In spermatids, the open chromatin state induced by histone acetylation and histone methylation possibly facilitates nucleosome eviction and thus histone removal.

Two proteins were previously reported to play a role in histone removal: CHD5 and RNF8. CHD5, a chromatin remodeler, plays a multi-faceted role in histone removal in spermatids (Li et al., 2014). CHD5 binds to unmodified histone 3 (H3) with its PHD fingers (Oliver et al., 2012). Loss of CHD5 compromises H4 acetylation and perturbs the expression of transition proteins and protamines. The role of RNF8, a ubiquitin E3 ligase, in histone removal in spermatids is contradictory. In one study, it was claimed that RNF8-dependent histone ubiquitylation induces H4K16 acetylation, resulting in nucleosome removal (Lu et al., 2010). However, another study concluded that RNF8 is required for neither H4K16 acetylation nor histone-to-protamine exchange (Abe et al., 2021). Here, we report that the DOT1L-MLLT10 complex is important, highlighting the multi-factorial regulation of this complex process.

We have previously reported that DOT1L is essential for self-renewal of spermatogonial stem cells (Lin et al., 2022). In this study, we find that DOT1L plays additional roles during male meiosis and spermiogenesis. In Dot1l-deficient testis, the increased apoptosis in metaphase I spermatocytes is presumably caused by activation of the spindle assembly checkpoint, which is usually triggered by defects in chromosome segregation. However, chromosome synapsis and meiotic recombination appeared to be unaffected in Dot1l-deficient testis. In C. elegans, Dot-1.1 plays a role in the regulation of chromosome synapsis and meiotic recombination, but is not essential for either meiotic process (Lascarez-Lagunas et al., 2020). In budding yeast, Dot1 functions in the meiotic checkpoint control (San-Segundo and Roeder, 2000). Such a role for Dot-1.1 in meiotic checkpoint in C. elegans is not as clear-cut as in yeast (Lascarez-Lagunas et al., 2020). Although the methyltransferase activity of DOT1L is evolutionarily conserved, the role of DOT1L and its orthologues during meiosis differs between the single cell eukaryotes (yeast) and metazoans.

DOT1L interacts with MLLT10 in leukemic cells (Deshpande et al., 2014; Okada et al., 2005). Direct fusion of either DOT1L or MLLT10 to MLL (mixed lineage leukemia) leads to leukemic transformation (Okada et al., 2005). The methyltransferase activity of DOT1L depends on its binding partner MLLT10 in leukemic cells (Deshpande et al., 2014). We find that DOT1L and MLLT10 form a complex in testis. Although DOT1L is essential for self-renewal of spermatogonial stem cells (Lin et al., 2022), the current study did not identify a role for MLLT10 in self-renewal of spermatogonial stem cells, suggesting that the requirement of DOT1L in spermatogonial stem cells is independent of MLLT10. However, the localization of DOT1L and MLLT10 on the sex chromatin in spermatocytes and spermatids is inter-dependent. Importantly, MLLT10 is required for H3K79me2 in spermatocytes and spermatids. Furthermore, loss of either DOT1L or MLLT10 leads to histone retention in epididymal sperm. Therefore, like in leukemic transformation, the interaction between DOT1L and MLLT10 is essential for H3K79 methylation and histone removal during spermiogenesis.

Dot1l floxed (Dot1l^fl) mice and Mllt10 floxed (Mllt10^fl) mice were generated previously (Bernt et al., 2011; Deshpande et al., 2014). To inactivate Dot1l or Mllt10 specifically in germ cells, floxed mice were crossed with the Cre mouse strains Ddx4-Cre (stock number 006954, Jackson Laboratory) and Stra8-Cre (Gallardo et al., 2007; Lin et al., 2017). PCR genotyping primers were as follows: the Dot1l floxed allele (418 bp), p2 (5′ CCCAAAAGGGTCTTTTCACA 3′) and p3 (5′ ATGGGATTTCATGGAAGCAA 3′); the Dot1l deletion allele (620 bp), p1 (5′ CTCACAGTCACATACTACCTCTGAC 3′) and p3 (5′ ATGGGATTTCATGGAAGCAA 3′); Ddx4-Cre (240 bp), Vasa-Cre-1 (5′ CACGTGCAGCCGTTTAAGCCGCGT 3′) and Vasa-Cre-2 (5′ TTCCCATTCTAAACAACACCCTGAA 3′); the Mllt10 floxed allele (602 bp), Flox-F (5′ TGCTCGGATCAAAGCTTCC 3′) and Flox-R (5′ TCTGTCTCTGTCCCTCACAAC 3′); the Mllt10 deletion allele (250 bp), 1a (5′ CACAGCCTACTTCAAAGAAC 3′) and 4a (5′ ATTAGAGTCCATCCCACTTC 3′); and Stra8-Cre (400 bp), F (5′ ACTCCAAGCACTGGGCAGAA 3′) and R2 (5′ CGTTTACGTCGCCGTCCAG 3′). The Institutional Animal Care and Use Committee (IACUC) of the University of Pennsylvania approved all procedures.

A guide RNA mapping to the last coding exon of the Dot1l gene was designed (Fig. S1). DNA oligos for sgRNA contained overhangs for ligation into the BbsI site in the px330 vector (Addgene 127875). sgRNA was generated through in vitro transcription with T7 RNA polymerase. The ssDNA repair template contained the 3×FLAG-coding sequence (Fig. S1). The sgRNA (50 ng/μl), CAS9 protein (100 ng/μl) and template ssDNA (100 ng/μl) were mixed according to a protocol described previously (Yang et al., 2014). This mixture was microinjected into mouse zygotes to produce founder mice at the PennVet Transgenic Core facility. Injected zygotes were cultured in vitro until the two-cell stage and were then transferred into oviducts of 0.5 day post-coitum pseudopregnant foster mothers. Around 19.5 days post coitum, recipient mothers delivered pups. Two positive founders were verified by Sanger sequencing and bred with DBA2 mice for germline transmission, which was successful. The Dot1l-3×FLAG mice were bred to homozygosity.

For sperm count, one cauda epididymis from each mouse was harvested and placed in a tissue dish with 1×PBS. The cauda was minced into small pieces with scissors, fixed in paraformaldehyde and counted on the Hausser Bright-Line hematocytometer. For each mating test, one 8-week-old control or conditional knockout male was housed with two 8-week-old wild-type females for 3 months. The litter number and size during this period were recorded.

Protein extracts prepared from 100 mg of 26-day-old wild-type and Dot1l^Flag/Flag testes were used for immunoprecipitation with a murine IgG₁ monoclonal anti-FLAG antibody covalently conjugated to agarose (A2220, Sigma). Proteins were eluted with FLAG peptides. Eluted proteins were run to 1 cm away from the loading well on an 8% SDS-PAGE gel and stained with Coomassie Brilliant Blue G-250 dye (Sigma). The stained gel was cut and sent for protein identification by mass spectrometry at the Wistar Proteomics Core Facility (Table S1). The peptide maps of DOT1L and MLLT10 are shown in Fig. S9.

For co-immunoprecipitation experiments followed by western blotting, 100 mg of 26-day or adult testes (wild type or Dot1l^Flag/Flag) were homogenized in 1 ml lysis buffer (50 mM Tris, 150 mM NaCl, 5 mM MgCl₂, 1 mM DTT, 0.5% deoxycholate and 1% Triton) with a cocktail of proteinase inhibitors. The lysate was precleared by incubating with agarose beads at 4°C for 2 h. Pre-cleared lysate was used for immunoprecipitation with anti-FLAG or anti-MLLT10 antibodies followed by western blotting with anti-DOT1L, anti-FLAG or anti-MLLT10 antibodies.

For histological analysis, testes were fixed in Bouin's solution, dehydrated through a gradient of ethanol, embedded with paraffin wax and sectioned. Sections were deparaffinized in xylene, rehydrated through a gradient of ethanol, and stained with Hematoxylin and Eosin. Sections were imaged on a Leica DM5500B microscope by DFC450 digital color camera (Leica Microsystems). For immunofluorescence including TUNEL analyses, testes were fixed in 4% paraformaldehyde (in 1×PBS) overnight at 4°C, dehydrated in 30% sucrose (in 1×PBS) overnight, embedded in optimal cutting temperature compound (6502, Thermo Scientific) and sectioned on a cryostat. Alternatively, testes were fixed in a fixative solution (30% formaldehyde, 15% ethanol and 5% glacial acetic acid) overnight. Fixed testes were dehydrated through a gradient of ethanol, embedded in paraffin wax and sectioned on a microtome. Paraffin wax-embedded slides were deparaffinized in xylene, rehydrated through a gradient of ethanol. Heat-induced epitope retrieval was performed for the rehydrated slides by incubation in the epitope retrieval buffer (1 mM EDTA buffer and 0.05% Tween-20 at pH 8.0) at 95°C for 20 min, excluding TUNEL assays. Epitope retrieval for TUNEL assays was performed by incubation with proteinase K (10 µg/ml) at room temperature for 10 min. Subsequently, slides were incubated with 0.5% Triton X-100 in 1×PBS at room temperature for 10 min. Surface nuclear spread analysis of spermatocytes has been described previously (Peters et al., 1997). In brief, seminiferous tubules were extracted from testes and incubated for 45 min to 1 h in the hypotonic buffer (50 mM sucrose, 30 mM Tris, 5 mM EDTA, 17 mM trisodium citrate dihydrate, 0.5 mM DTT and 1 mM phenylmethylsulfonyl fluoride). After hypotonic treatment, seminiferous tubules were minced into cell suspension in 100 mM sucrose and were then spread on a slide covered with paraformaldehyde solution containing Triton X-100 in a wet chamber overnight.

The slides were blocked with 10% goat serum in PBST at 37°C for 1 h followed by primary antibody incubation at 37°C overnight. The primary antibodies used for immunofluorescence were as follows: DOT1L (1:50, UP2565) (Lin et al., 2022), MLLT10 (1:200, HPA005747, Sigma), H3K79me2 (1:100, ab3594, Abcam), SYCP3 (1:500, ab97672, Abcam), SYCP2 (1:500) (Yang et al., 2006), SYCP1 (1:100, ab15090, Abcam), γH2AX (1:500, 16-202A, clone JBW301, Millipore), RPA2 (1:50; UP2436) (Shi et al., 2019), SCML2 (1:200; UP2323) (Luo et al., 2015), H3K9me2 (1:500, ab1220, Abcam), H3K4me3 (1:500, 39159, Active Motif) and POLII (1:50, sc-899, Santa-Cruz Biotech). Secondary antibodies were FITC- or Texas red-conjugated (1:200, Vector Laboratories, FI-1000 or TI-1000). TUNEL assays were performed with a TUNEL kit (TUNEL Enzyme, 11767305001, Roche) and the TUNEL Label kit (11767291910, Roche) at a ratio of 1:10. Mounting medium with DAPI (H-1200, Vectashield) was used for nuclear DNA counterstaining and fluorochrome preservation. Fluorescence images were captured with an ORCA Flash4.0 digital monochrome camera (Hamamatsu Photonics) on a Leica DM5500B microscope.

Sperm protein extract preparation and western blot analysis have been described previously (Luense et al., 2019; Zhou et al., 2012). Cauda epididymal sperm were collected from each male. Briefly, caudal epididymides were cut several times and incubated in 1×PBS solution at 37°C for 10 min to allow sperm to swim out. Sperm concentration was determined by counting with a hemocytometer. Sperm were collected by centrifugation at 800 g for 5 min at room temperature. The sperm pellet was resuspended in 100 μl of lysis buffer [20 mM Tris-HCl (pH 7.5), 1 mM MgCl₂, 1 mM CaCl₂, 137 mM NaCl, 10% glycerol, 1% NP-40, 300 nM TSA, benzonase (12.5 U/ml) and 1× protease inhibitors] and incubated at 4°C for 1 h. Alternatively, sperm were homogenized in 100 μl SDS-EDTA solution (1% SDS, 24 mM EDTA and 75 mM NaCl at pH 6.0) followed by centrifugation at 5000 g for 30 min at room temperature. An equal volume of the SDS-PAGE sample buffer [62.5 mM Tris (pH 6.8), 3% SDS, 10% glycerol, 5% β-mercaptoethanol and 0.02% Bromophenol Blue) was added to the saved supernatant,. The sample was heated at 95°C for 10 min and was used for western blot analysis.

Total protein extract from testis or cultured cells was prepared by homogenization in five volumes of protein extraction buffer [62.5 mM Tris-HCl (pH 6.8), 3% SDS, 10% glycerol and 5% 2-mercaptoethanol) and boiled at 95°C for 15 min followed by centrifugation at 12,000 g for 5 min. Samples were run in an SDS-PAGE gel and then transferred to the PVDF membrane. The membrane blot was blocked in 5% non-fat milk (in 1×PBST) at room temperature for 1 h. The primary antibodies used for western blot were as follows: FLAG (1:5000, F1804, Sigma), MLLT10 (1:1000, HPA005747, Sigma), DOT1L (1:1000, ab64077, Abcam), H3K79me2 (1:1000, ab3594, Abcam), ACTB (1:5000, A5441, Sigma), SYCP3 (1:500, 23024-1-AP, Proteintech), acetylated tubulin (1:2000, T7451, Sigma) and histone H3 (1:2000, 9715S, Cell Signaling Technology). Secondary antibodies were HRP-conjugated (1:5000, Cell Signaling Technology, 7076S or 7074S). ECL western blotting substrate (32106, ThermoFisher Scientific) was used for development.

We thank Fang Yang and Zhenlong Kang for comments on the manuscript. We thank Hsin Yao Tang and Thomas Beer at the Wistar Proteomics Core for mass spectrometry.

The peer review history is available online at https://journals.biologists.com/dev/lookup/doi/10.1242/dev.201501.reviewer-comments.pdf