Tex19.1 promotes Spo11-dependent meiotic recombination in mouse spermatocytes

Meiosis relies on the SPO11 endonuclease to generate the recombinogenic DNA double strand breaks (DSBs) required for homologous chromosome synapsis and segregation. The number of meiotic DSBs needs to be sufficient to allow chromosomes to search for and find their homologs, but not excessive to the point of causing genome instability. Here we report that the mammal-specific gene Tex19.1 promotes Spo11-dependent recombination in mouse spermatocytes. We show that the chromosome asynapsis previously reported in Tex19.1-/- spermatocytes is preceded by reduced numbers of recombination foci in leptotene and zygotene. Tex19.1 is required for normal levels of early Spo11-dependent recombination foci during leptotene, but not for upstream events such as MEI4 foci formation or accumulation of H3K4me3 at recombination hotspots. Furthermore, we show that mice carrying mutations in Ubr2, which encodes an E3 ubiquitin ligase that interacts with TEX19.1, phenocopy the Tex19.1-/- recombination defects. These data suggest that Tex19.1 and Ubr2 are required for mouse spermatocytes to accumulate sufficient Spo11-dependent recombination to ensure that the homology search is consistently successful, and reveal a hitherto unknown genetic pathway promoting meiotic recombination in mammals.


Introduction
Recombination plays key roles in meiosis and gametogenesis through facilitating the pairing and reductional segregation of homologous chromosomes, and by increasing genetic variation in the next generation. Meiotic recombination is initiated when programmed DNA double strand breaks (DSBs) are generated during the leptotene stage of the first meiotic prophase. Meiotic DSBs recruit a series of recombination proteins visualised cytologically as recombination foci, and initiate a search for homologous chromosomes thereby promoting homologous chromosome synapsis during zygotene. Recombination foci continue to mature while the chromosomes are fully synapsed in pachytene, and eventually resolve into crossover or non-crossover events. Crossovers exchange large tracts of genetic information between parental chromosomes, increasing genetic diversity in the population. Furthermore, these crossovers, which physically manifest as chiasmata, hold homologs together after they desynapse in diplotene and help to ensure that homologous chromosomes undergo an ordered reductional segregation at anaphase I [1,2].
Meiotic DSBs have a non-random distribution across the genome, and their frequency and location play an important role in shaping the recombination landscape [2,3]. In male mice, a few hundred meiotic DSBs are generated during leptotene, around 20-25 of which mature into crossovers. The positions of meiotic DSBs across the genome are determined by PRDM9, a histone methyltransferase that mediates trimethylation of histone H3 lysine 4 (H3K4me3) at recombination hotspots [4,5]. Meiotic DSBs are generated by an endonuclease that comprises SPO11 and TOPOVIBL subunits [2,3,6]. In mice, mutations in Spo11 result in fewer DSBs during leptotene and zygotene, and defects in the pairing and synapsis of homologous chromosomes [7][8][9]. The overall amount of SPO11 activity appears to be dynamically controlled at multiple levels during meiotic prophase. At the RNA level, Spo11 is alternatively spliced into two major isoforms whose relative abundance changes as meiotic prophase proceeds [10][11][12]. There also appears to be regulation of SPO11 activity at the protein level: negative feedback mechanisms acting through the DNA damage-associated protein kinase ATM prevent excessive Spo11-dependent DSBs from being generated during meiosis, potentially limiting any genetic instability caused by errors arising during repair of the DSBs and meiotic arrest caused by unrepaired DSBs [13]; and chromosome synapsis feeds back to locally inhibit SPO11 activity in chromosomal regions that have already synapsed during zygotene [14].
Mutations in genes involved in regulating early stages in meiotic recombination in mammals might be expected to phenocopy Spo11 -/mutants to some extent in having reduced numbers of DSBs in leptotene, and arrest at pachytene with chromosome asynapsis. One group of genes that is required for chromosome synapsis in mouse spermatocytes, but whose mechanistic role in meiosis is poorly defined, is the germline genome defence genes [15]. These genes are involved in suppressing the activity of retrotransposons in developing germ cells, and mutations in many of them cause defects in progression through the pachytene stage of meiosis [15]. Mutations in one of these germline genome defence genes, Mael, which encodes a conserved component of the piRNA pathway, causes -Page 4 - de-repression of retrotransposons and a considerable increase in Spo11-independent DNA damage [16]. The Spo11-independent DNA damage generated in these mutants could potentially reflect the activity of the retrotransposon-encoded endonucleases that generate nicks or breaks in the host DNA to mediate mobilisation of these genetic elements [16]. In contrast, spermatocytes carrying mutations in the DNA methyltransferase accessory factor Dnmt3l also de-repress retrotransposons, but have relatively normal levels of DSBs that are aberrantly distributed across the genome [17][18][19].
The germline specificity in expression of at least a subset of the germline genome-defence genes is achieved through tissue-specific promoter DNA methylation [20]. One of the most methylation sensitive of these genes is Tex19. 1 [20,21]. Tex19.1 was originally identified in a screen for testisspecific genes [22], and is one of two rodent paralogs of this mammal-specific gene family [23].
Although TEX19.1 was described as being a nuclear factor with potential roles in maintenance of In this study we elucidate why loss of the germline genome defence gene Tex19.1 results in chromosome asynapsis in male meiosis. We show that loss of Tex19.1 generates a meiotic phenotype distinct from either Mael -/or Dnmt3l -/mutants. Rather loss of Tex19.1 phenocopies hypomorphic Spo11 mutants and impairs Spo11-dependent recombination during the leptotene stage of meiotic prophase. Furthermore, we show that mice lacking the TEX19.1-interacting protein UBR2 phenocopy the recombination defects seen in leptotene Tex19.1 -/spermatocytes. These data show that Tex19.1 and Ubr2 are required for sufficient SPO11-dependent recombination to ensure robust identification and synapsis of homologous chromosomes in meiotic spermatocytes.

Chromosome Asynapsis in Tex19.1 -/-Spermatocytes is not Caused by Primary Defects in Synaptonemal Complex Assembly
Tex19.1 is a DNA methylation-sensitive germline genome defence gene whose expression is primarily restricted to germ cells and pluripotent cells in the embryo [20,[22][23][24]. We and others have previously reported that Tex19.1 -/males have defects in spermatogenesis on a mixed genetic background, and that around 50% of pachytene spermatocytes in Tex19.1 -/testes have asynapsed chromosomes, but the molecular explanation for this defect remains unknown [15,24,25]. Synapsis requires the accurate and timely execution of a number of events in the preceding stages of the first meiotic prophase, including the generation of meiotic DNA double-strand breaks (DSBs) in leptotene, followed by homolog pairing and assembly of the synaptonemal complex (SC) in zygotene [1]. To investigate the molecular basis for the chromosome asynapsis in pachytene Tex19.1 -/spermatocytes we sought to test whether each of these events occurs normally in the absence of Tex19.1. First, we confirmed that the meiotic chromosome asynapsis phenotype persists in Tex19.1 -/spermatocytes after backcrossing onto an inbred C57BL/6 genetic background: 65.4% ± 1.3 of Tex19.1 -/pachytene spermatocytes from three animals were asynapsed in this genetic background, significantly higher than the 3.7% ± 1.3 of Tex19.1 +/±control pachytene spermatocytes that were asynapsed (Student's t-test, p<0.001) (Fig 1A). This is similar to the asynapsis in 50% of Tex19.1 -/pachytene nuclei previously reported for a mixed genetic background [24,25]. To assess whether the chromosome asynapsis in Tex19.1 -/spermatocytes represents defects in SC assembly rather than pairing of homologous chromosomes, we scored the configuration of the asynapsed chromosomes in these asynapsed pachytene Tex19.1 -/nuclei. Defects in assembly of the SC transverse filaments results in asynapsed chromosomes that are aligned in their homolog pairs whereas defective recombination or pairing between homologous chromosomes manifests as isolated asynapsed single chromosomes, partial synapsis between non-homologous chromosomes, and incomplete synapsis between homologous chromosomes [7,8,28,29]. Asynapsed chromosomes in Tex19.1 -/spermatocytes are present in multiple configurations consistent with defects in recombination or homolog pairing, but do not present as asynapsed aligned homolog pairs (Fig 1A, Fig 1B).
To confirm that the asynapsis phenotype in Tex19.1 -/spermatocytes does not represent a primary defect in SC assembly, we quantified the effect of Tex19.1 on the number of SC fragments assembled independently of recombination in a Spo11 -/genetic background [30]. Spo11 -/spermatocytes arrest with a zygotene-like SC configuration with complete axial element formation but limited synapsis [7,8]. Spo11 -/-Tex19.1 +/± and Spo11 -/-Tex19.1 -/spermatocytes are able to assemble similar amounts of SC in this assay (Fig 1C, Fig 1D), suggesting that loss of Tex19.1 does not severely impair recombination-independent SC assembly. Taken together, these data suggest that the chromosome asynapsis in Tex19.1 -/spermatocytes is likely primarily caused by defects in meiotic recombination and/or homolog pairing rather than a direct defect in SC assembly.  (Fig 2). Interestingly, the number of RPA foci is not statistically different from zygotene control nuclei (Fig 2), which could potentially reflect RPA foci being a later marker of recombination than RAD51 and DMC1 [34]. The differential behaviour of RAD51 and DMC1 foci in Tex19.1 -/spermatocytes suggests that the generation, repair, or maturation kinetics of recombination foci is perturbed in the absence of Tex19.1.
Meiotic recombination is initiated during leptotene [9], therefore we next investigated whether loss of Tex19.1 might perturb recombination foci frequency at this earlier stage of meiotic prophase.
Counts of RPA, DMC1 and RAD51 foci in leptotene nuclei revealed a severe reduction in the frequency of each of these in the absence of Tex19.1 (Fig 3). The numbers of RPA foci, DMC1 foci and RAD51 foci in leptotene Tex19.1 -/spermatocytes were reduced to 63%, 30%, and 60% of those present in control spermatocytes (Fig 3). Thus, loss of Tex19.1 results in reduced numbers of -Page 8 - The reduced numbers of recombination foci in Tex19.1 -/spermatocytes could potentially decrease the efficiency of the DSB-dependent homology search and contribute to chromosome asynapsis in this mutant. Analysis of Spo11 hypomorphs suggests that reduced numbers of meiotic DSBs impairs the initiation of synapsis and manifests as reduced numbers of SC fragments during late leptotene/early zygotene stages [14]. We therefore analysed the extent of synapsis in zygotene Tex19.1 -/nuclei to assess whether the initiation of synapsis might similarly be impaired in these mutants. Chromosome spreads were immunostained with axial and central element SC markers and the percentage synapsis assessed in each zygotene nucleus (Supporting Fig S1A). In the absence of

Tex19.1 -/-Spermatocytes Have Reduced Amounts of Spo11-Dependent Recombination
The reduced number of RPA, DMC1 and RAD51 foci in leptotene Tex19.1 -/spermatocytes might reflect fewer Spo11-dependent DSBs in these cells, or defects in the processing and resection of those DSBs to form the single-stranded DNA ends that recruit RPA, DMC1 and RAD51, or generate γH2AX occurs in response to Spo11-dependent DSB formation [9], and is not impaired in spermatocytes proposed to be defective in subsequent processing of those DSBs [35]. We therefore tested whether loss of Tex19.1 affects γH2AX abundance in leptotene spermatocytes. In both control and Tex19.1 -/leptotene nuclei, γH2AX is present as a diffuse cloud of staining over regions of the nucleus (Fig 4A). Interestingly, quantification of the γH2AX signal showed that the amount of γH2AX in leptotene Tex19.1 -/nuclei was around half that in Tex19.1 +/± controls ( Fig 4B). Taken together, the reduced numbers of recombination foci and the reduced intensity of γH2AX immunostaining in Tex19.1 -/spermatocytes suggests that loss of Tex19.1 likely causes defects in early stages of Spo11-dependent recombination, or accelerated repair of SPO11-induced DNA damage.
The bulk of the γH2AX generated in spermatocytes reflects the generation of Spo11-dependent meiotic DSBs, however small amounts of γH2AX are generated independently of Spo11 in these cells [9,[36][37][38][39]. The extent of the decrease in γH2AX abundance in Tex19.1 -/spermatocytes is arguably more consistent with reduced abundance of Spo11-dependent DSBs, but it is possible that loss of Tex19.1 also affects Spo11-independent γH2AX generated during leptotene. To test directly whether loss of Tex19.1 affects Spo11-independent γH2AX we quantified γH2AX abundance as well as DMC1 foci in Spo11 -/-Tex19.1 -/double mutant spermatocytes. The relatively low levels of γH2AX present in Spo11 -/spermatocytes typically manifests as a pseudo sex body, a cloud of γH2AX associated with a subset of asynapsed axes undergoing meiotic silencing of unsynapsed chromatin [38,39]. In addition to the pseudo sex body, smaller additional flares of chromosome axis-associated γH2AX staining termed L-foci are also present [36,37]. Spo11 -/-Tex19.1 -/spermatocytes displayed similar γH2AX staining patterns and similar numbers of γH2AX L-foci as Spo11 -/-Tex19.1 +/± controls (Fig 4C, Fig 4D). Thus, pseudo sex body formation and Spo11independent γH2AX L-foci frequency are independent of Tex19. 1 Tex19.1 impairs DMC1 foci frequency in a wild-type Spo11 background (Fig 2, Fig 3), loss of Tex19.1 has no detectable effect on DMC1 foci frequency in a Spo11 -/mutant background (Fig 4E,   Fig 4F). Thus, loss of Tex19.1 appears to reduce the amount of Spo11-dependent recombination present in spermatocytes. In this respect the Tex19.1 -/phenotype bears some resemblance to  (Fig 4C, Fig 4D) or axis-associated RAD51 foci (Supporting Fig S2). Therefore, Tex19.1 and the piRNA pathway component Mael appear to have different effects on Spo11-independent DNA damage in meiotic spermatocytes. SPO11 is locally regulated in the nucleus, and feedback controls are thought to allow SPO11 to continue to generate DSBs on asynapsed regions of the chromosomes in late zygotene [14]. Spo11 hypomorphs are still able to generate DSBs on asynapsed chromatin [14]. To assess whether asynapsed chromatin is similarly able to accumulate high levels of DSBs in Tex19.1 -/mutants, we counted the number of RPA foci associated with the sex chromosomes, which remain largely asynapsed during pachytene. In the population of pachytene Tex19.1 spermatocytes that successfully synapse all their autosomes, sex chomosomes were still able to accumulate similar numbers of RPA foci as control pachytene nuclei (Supporting Fig S1C, Supporting Fig S1D). Thus, Tex19.1 has sexually dimorphic effects on progression through meiotic prophase, and in contrast to its effects on spermatocytes, loss of Tex19.1 does not cause defects in chromosome synapsis in female meiosis [26]. Nevertheless, it is possible that loss of Tex19.1 could still cause a reduction in the number of early recombination foci in female meiosis that might not be sufficient to result in chromosome asynapsis. We therefore analysed RAD51 foci in E14.5 Tex19.1 -/foetal oocytes to test whether loss of Tex19.1 affects recombination in female meiosis. However, the number of RAD51 foci in late leptotene Tex19.1 -/oocytes is not significantly different from late leptotene Tex19.1 +/± littermate controls (Supporting Fig S3). Therefore Tex19.1 is not required for accumulation of RAD51 foci in female meiosis and has a sexually dimorphic role in early meiotic recombination.

Recombination Hotspots
We next investigated whether the reduced frequency of Spo11-dependent recombination foci in leptotene Tex19.1 -/spermatocytes might reflect defects upstream of Spo11 in meiotic recombination. The requirements upstream of Spo11 for meiotic DSB formation are relatively poorly understood in mammals, however SPO11 activity likely depends on the recruitment of the conserved axis-associated protein MEI4 to the chromosomal axes in leptotene [41]. We therefore quantified MEI4 foci in leptotene Tex19.1 -/nuclei to test whether this event is perturbed by loss of Tex19.1. Control leptotene Tex19.1 +/± spermatocytes possess an average of 218 axis-associated MEI4 foci (Fig 5A, Fig 5B), similar but slightly lower than the average 309 foci per leptotene nucleus reported previously [41]. Leptotene Tex19.1 -/nuclei possess similar numbers of MEI4 foci to leptotene Tex19.1 +/± controls ( Fig 5A, Fig 5B). Thus, the reduced frequency of recombination foci seen in Tex19.1 -/leptotene spermatocytes appears to be a consequence of defects acting downstream or independently of MEI4 localisation to chromosome axes. Spo11 function is also influenced by the activity of the histone methyltransferase PRDM9, which targets SPO11 to recombination hotspots [2,4,5]. Mutations in Prdm9 result in reduced anti-H3K4me3 immunostaining in P14 spermatocytes, a failure to enrich H3K4me3 at Prdm9-dependent recombination hotspots, a reduction in recombination foci during early prophase, and meiotic chromosome asynapsis [5,42,43]. We therefore tested whether loss of Tex19.1 might impair Prdm9 function by assessing anti-H3K4me3 immunostaining intensity in leptotene nuclei. However, we could not detect a difference in the amount of anti-H3K4me3 immunostaining between Tex19.1 +/± and Tex19.1 -/leptotene nuclei ( Fig 5C, Fig 5D). To test whether the distribution of H3K4me3 rather than its total abundance might be altered in the absence of Tex19.1 we performed H3K4me3 chromatin immunoprecipitation (ChIP)-qPCR on P16 testes. H3K4me3 is enriched at transcriptional start sites (TSSs) of active genes in addition to meiotic recombination hotspots [44], and as expected both Tex19.1 +/± and Tex19.1 -/testes show enrichment of H3K4me3 at Gapdh and Polr2a active TSSs, but not at a Polr2a intragenic region ( Fig 5E). However, loss of Tex19.1 does not perturb the accumulation of H3K4me3 at Prdm9-dependent recombination hotspots ( Fig 5E).
Thus, the defects in Spo11-dependent recombination seen in Tex19.1 -/spermatocytes does not appear to be a downstream consequence of impaired Prdm9 activity. We generated Ubr2 -/mice carrying a premature stop codon in the N-terminal region of UBR2 within the UBR domain that binds N-end rule substrates. The Ubr2 -/mice analysed here have no detectable UBR2 protein in their testes (Supporting Fig S4A), a 68% reduction in testis weight (Supporting Fig S4B), and no detectable sperm in their epididymis (Supporting Fig S4C), consistent retrotransposon RNAs are derepressed in Ubr2 -/spermatocytes ( Fig 6A, Fig 6B).
We tested whether the meiotic defects in Ubr2 -/spermatocytes might resemble the asynapsis seen in Tex19.1 -/spermatocytes (Fig 1A, 1B). Chromosome spreads from Ubr2 -/testes confirm that this Ubr2 mutant allele causes defects in progression through meiotic prophase, and very few spermatocytes progress through pachytene into diplotene ( Fig 6C). Furthermore, around 40% of pachytene Ubr2 -/spermatocytes had at least one asynapsed autosome pair when staging SYCP3positive nuclei for meiotic progression under low magnification ( Fig 6C, Fig 6D). At higher magnification, 65.9% ± 2.5 Ubr2 -/pachytene nuclei from three animals have some autosomal asynapsis, compared to 11.3% ± 2.5 pachytene nuclei from three Ubr2 +/+ animals (p<0.001, Student's t-test). Like in Tex19.1 -/spermatocytes ( Fig 1A, Fig 1B), these asynapsed chromosomes are present in multiple configurations consistent with defects in recombination or homolog pairing ( Fig 6E). Similar to Tex19.1 -/spermatocytes, the asynapsis in Ubr2 -/spermatocytes is also associated with earlier defects in meiotic recombination. γH2AX abundance, DMC1 foci frequency and RAD51 foci frequency are reduced to around 50%, 52% and 58% respectively of those seen during leptotene in Ubr2 -/mutants (Fig 7), which contrasts with a previous report that γH2AX staining, and RAD51 and RPA foci frequency are unaffected in leptotene Ubr2 -/spermatocytes [48]. Consistent with the decrease in leptotene recombination foci frequency reported here, DMC1 and RAD51 foci frequency remain around 66% and 86% respectively of that seen in control spermatocytes during zygotene (Fig 7). As there appeared to be some qualitative similarity between the defects in recombination foci frequency in Ubr2 -/and Tex19.1 -/spermatocytes, we tested whether this meiotic recombination defect would be sufficient to delay or impair the initiation of  Fig S1B) and Spo11 hypomorphs [14], synapsis is delayed in the absence of Ubr2 (Supporting Fig S4E,   Supporting Fig S4F). These data suggest that the defect in progression to pachytene previously reported in Ubr2 -/mutants [46] may reflect loss of TEX19.1 protein and earlier defects in the meiotic recombination in these mutants. Furthermore, these data show that Ubr2 and Tex19.1 are both required to allow sufficient early recombination foci to accumulate to drive robust homologous chromosome synapsis in mouse spermatocytes.

Meiotic Recombination and Chromosome Synapsis in Meiosis
This study aimed to elucidate the mechanistic basis of the chromosome synapsis defect in male mice carrying mutations in the germline genome defence gene Tex19. 1 [24]. We have shown that the pachytene chromosome asynapsis in these mice, and in mice carrying mutations in the TEX19.1-interacting protein UBR2, is likely a downstream consequence of reduced meiotic recombination earlier in meiotic prophase. Wild-type mice generate around 10-fold more meiotic DSBs than there are chiasmata, and the large numbers of DSBs generated in leptotene and zygotene appear to be important to drive pairing and synapsis of homologous chromosomes [2,7,8]. Allelic series of Spo11 activity suggest that reducing the number of meiotic DSBs to around 50% of normal levels is sufficient to cause chromosome asynapsis [14,40]. The reduction in early recombination foci seen in leptotene Tex19.1 -/and leptotene Ubr2 -/spermatocytes, is similar to this threshold and could be sufficient to account for the chromosome asynapsis seen in these mutants. Notably, and there is some progression to post-meiotic spermatid stages in both these mutants. Thus thẽ 50% reduction in leptotene DSB frequency caused by loss of Tex19.1 or Ubr2 could be sufficient to cause the level of asynapsis present in these spermatocytes.
Interestingly, the frequency of recombination foci in zygotene Tex19.1 -/spermatocytes is closer to wild-type levels than that seen during leptotene, suggesting additional recombination foci are accumulating during zygotene that allow the Tex19.1 -/spermatocytes to catch up with wild-type cells. It is possible that DSB generation is delayed in Tex19.1 -/spermatocytes, or that repair of DSBs is accelerated in leptotene but not zygotene Tex19.1 -/spermatocytes, or that this compensation of the Tex19.1 recombination deficiency during zygotene reflects control mechanisms that regulate DSB frequency in meiotic cells [14]. An overall delay in germ cell development is probably not causing a delay in meiotic recombination relative to axial element assembly as previous analysis of gene expression profiles in P16 Tex19.1 -/testes does not exhibit enrichment of genes expressed in more immature germ cells such as spermatogonia or leptotene spermatocytes [24,45]. In hypomorphic Spo11 mice, DSBs are generated on asynapsed regions of the chromosomes during zygotene, potentially stimulating homology search and synapsis in these regions [14]. However, although any additional early recombination foci that accumulate in zygotene in Tex19.1 -/spermatocytes might be rescuing asynapsis to some degree, they are not sufficient to allow the majority of Tex19.1 -/spermatocytes to complete synapsis.

Meiotic Defects in Genome Defence Mutants
Tex19.1 is one of a group of germline genome defence genes which cause retrotransposon de-

Roles for Tex19.1 and Ubr2 in Meiotic Recombination
The data presented here suggest that both Tex19.1 and Ubr2 are required for sufficient meiotic recombination to drive robust chromosome synapsis in spermatocytes. We have shown defects in the number of early recombination foci, and in the amount of γH2AX present during leptotene in both these mutants. Further experiments are required to delineate which stage in early recombination is disrupted in these mutants. It is possible that SPO11 activity and DSB formation itself is reduced or delayed. Alternatively early stages in processing SPO11-dependent DSBs or signalling the SPO11-induced DNA damage could be perturbed in these mutants. Or SPO11dependent DSBs or recombination intermediates could be repaired more rapidly in the absence of UBR2 was previously suggested not to have a role in the initiation of meiotic recombination as it did not localise to recombination foci and was not required for normal recruitment of RAD51 or RPA to recombination foci during leptotene [48]. Immunocytologically-detectable enrichment at recombination foci is probably not a requirement for UBR2 to directly or indirectly influence the initiation of meiotic recombination. However, the effect of Ubr2 on recombination foci and γH2AX during leptotene reported here does contradict the previous description of the Ubr2 -/leptotene spermatocyte phenotype, although representative images and quantitative analysis of recombination foci in leptotene Ubr2 -/spermatocytes were not shown in that study [48]. Differences between mouse strain background or Ubr2 allele being studied may contribute to this, and the delay in synapsis initiation during zygotene (Supporting Fig S4F) could also complicate meiotic prophase

Mice
Tex19.1 -/animals backcrossed three times to a C57BL/6 genetic background were bred and genotyped as described [24]. Spo11 +/heterozygous mice [7] on a C57BL/6 genetic background [18] were inter-crossed with Tex19.1 +/mice. Animal experiments were carried out under UK Home Office Project Licence PPL 60/4424. Noon on the day that a plug was found was designated E0.5, day of birth was designated P1, and adult mice were typically analysed at between 6-14 weeks old. foci; 217 ± 7 and 217 ± 8 RPA foci for wild-type and heterozygous mice respectively, no significant difference by Mann-Whitney U test, n=40, 21, 10, 40 respectively). Therefore data from these control genotypes were pooled as Tex19.1 +/± to reduce animal use. Epididymal sperm counts were determined as described [24]. Ubr2 -/mice were generated by CRISPR/Cas9 double nickase- Recombination foci in leptotene and zygotene nuclei were imaged by capturing z-stacks using a piezoelectrically-driven objective mount (Physik Instrumente) controlled with Volocity software (PerkinElmer). These images were deconvolved using Volocity, a 2D image generated in Fiji [58], and analysed in Adobe Photoshop CS6. DMC1, RAD51 and RPA foci were counted as recombination foci when they overlapped a chromosome axis. To measure leptotene γH2AX or H3K4me3 signal intensity, nuclear area was delimited using the DAPI signal, and signal intensity in that area quantified and corrected for background non-nuclear signal in 16 bit grayscale images using Fiji software. To assess the extent of synapsis in zygotene nuclei (Supporting Fig S1,   Supporting Fig S4), the total length of completely assembled SC was estimated by SYCP1 or SYCE2 staining and expressed relative to the total length of SYCP3-containing axial/lateral element in that nucleus. For this and all immunocytological scoring, images were scored blind with respect to genotype by pooling control and knockout images, randomly assigning new filenames to each image, then decoding the filenames after scoring.

Histology and In Situ Hybridisation
Histology of Bouin's-fixed testes, and in situ hybridisation of MMERVK10C probes to Bouin'sfixed testis sections were performed as described [24].

qRT-PCR
RNA was isolated from macerated mouse testes using TRIzol (Invitrogen) and treated with Turbo DNAse (Ambion) to digest any genomic DNA contamination. 1µg DNAse-treated RNA was used to synthesise cDNA using Superscript III (Invitrogen). The cDNA was used as a template for qPCR using SYBR Select Master Mix (Applied Biosystems), and the relative quantity of RNA transcript calculated using the standard curve method as described by the supplier. The qPCR was performed 14. Kauppi