Defects in Meiotic Recombination Delay Progression Through Pachytene in Mouse Spermatocytes

During meiosis, recombination, synapsis, chromosome segregation and gene expression are coordinately regulated to ensure successful execution of this specialised cell division. In many model organisms, checkpoint controls can delay meiotic progression to allow defects or errors in these processes to be repaired or corrected. Mouse spermatocytes possess quality control checkpoints that eliminate cells with persistent irreparable defects in chromosome synapsis or recombination, and here we show that a spermatocyte checkpoint regulates progression through pachytene to accommodate delays in meiotic recombination. We have previously show that the appearance of early recombination foci is delayed in Tex19.1-/- spermatocytes during leptotene/zygotene, but some Tex19.1-/- spermatocytes still successfully synapse their chromosomes. Therefore, we have used autosomally synapsed Tex19.1-/- mouse spermatocytes to assess the consequences of delayed recombination on progression through pachytene. We show that these pachytene spermatocytes are enriched for early recombination foci. This skew is not accompanied by cell death and likely reflects delays in the generation and/or maturation of recombination foci. Moreover, patterns of axis elongation, chromatin modifications, and histone H1t expression are also all skewed towards earlier substages of pachytene suggesting these events are co-ordinately regulated. Importantly, the delay in histone H1t expression in response to loss of Tex19.1 does not occur in a Spo11 mutant background, suggesting that histone H1t expression is being delayed by a recombination-dependent checkpoint. These data indicate that a recombination-dependent checkpoint operates in mouse spermatocytes that can alter progression through pachytene to accommodate spermatocytes with some types of recombination defect.


Introduction
Meiosis is a central feature of the life cycle of sexually reproducing organisms that requires coordinated regulation of multiple distinct processes including recombination, chromosome synapsis, chromosome segregation and changes in gene expression, to generate haploid germ cells [1][2][3]. The correct temporal execution of some of these processes relies on substrate-product relationships between them, for example meiotic recombination generates aligned homologous chromosomes that are substrates for chromosome synapsis. In addition feedback and feedforward loops help to ensure robust execution of some meiotic processes [4]. One of the feedback systems operating in mouse spermatocytes restricts DNA double strand break (DSB) formation to the asynapsed regions of chromosomes [5]. Evidence for this feedback system can be seen in spermatocytes that have reduced activity of SPO11, a subunit of the endonuclease that generates meiotic DSBs [5]. These hypomorphic Spo11 spermatocytes generate fewer DSBs during leptotene causing reduced homologous chromosome synapsis, but a feedback mechanism allows the resulting asynapsed chromosomal regions to accumulate high densities of DSBs during late zygotene, presumably in order to stimulate the homology search in these regions and promote their synapsis [5].
In addition to feedback controls, checkpoints are also an important component of meiotic regulation. Checkpoints monitor and co-ordinate meiotic events, and can provide a quality control mechanism to eliminate cells that do not execute some aspects of meiosis correctly [4]. Persistent defects in chromosome synapsis in mouse spermatocytes can activate a checkpoint that triggers cell death during pachytene [6]. This synapsis checkpoint is caused by asynapsed chromosomes sequestering the transcriptional silencing machinery away from the heterologous sex chromosomes, causing defective meiotic sex chromosome inactivation (MSCI) and inappropriate expression of sex chromosome-encoded gene products [7]. Moreover, oocytes and spermatocytes also possess meiotic -Page 4 -checkpoints that trigger prophase cell death in response to recombination defects, although there are differences in the signalling pathways involved between the sexes [8][9][10]. The tight coupling between recombination and synapsis means that many meiotic recombination mutants activate the synapsis checkpoint in spermatocytes [6], making it difficult to study the effects of perturbed recombination on meiotic progression in the absence of confounding effects of asynapsis. However, a recombination checkpoint has been proposed to operate in spermatocytes carrying moderate severity mutations in Trip13 [10]. Trip13 encodes an AAA+ ATPase implicated in regulating HORMAD proteins [11], and moderate severity Trip13 mutants successfully synapse their chromosomes but have defects in the maturation of recombination foci [12,13]. These spermatocytes accumulate markers of unrepaired DNA damage and die through activation of a DNA damage checkpoint [10,12,13].
In addition to chromosome synapsis and DNA repair checkpoints operating in prophase, a spindle assembly checkpoint can delay the onset of anaphase I in response to defects in chromosome alignment in mouse meiosis [14]. This checkpoint appears to be more sensitive in spermatocytes than oocytes [15,16]. Furthermore, even when this checkpoint is activated, mouse oocytes with misaligned chromosomes can eventually proceed into anaphase I after a delay, whereas similar defects induce a strict metaphase I arrest and cell death in spermatocytes [17,18]. Thus, checkpoint activation can delay cell cycle progression to allow the cells to repair defects before progressing to the next stage of the cell cycle, and cells with persistent irreparable defects eventually continue through the checkpoint or initiate cell death depending on the context [4,19]. Although the pachytene asynapsis and DNA damage checkpoints have primarily been demonstrated to cause apoptosis and meiocyte death in mice, it is not clear whether these checkpoints are also able to coordinate meiotic progression to accommodate suboptimal cells that have encountered reparable lesions or delays. Tex19.1 was originally isolated as a testis-expressed gene, and encodes a mammal-specific protein that interacts with the E3 ubiquitin ligase UBR2 [20,21]. Tex19.1 plays a role in repressing retrotransposon expression in germ cells and hypomethylated somatic tissues [22,23], but potentially has additional targets that cause defects in meiotic recombination [24]. We have recently shown that Tex19.1 -/spermatocytes have reduced numbers of early recombination foci during leptotene and that, similar to hypomorphic Spo11 spermatocytes [5], additional early recombination foci are generated during zygotene in these mutants [24]. This early recombination defect likely causes the defects in homologous chromosome synapsis seen around two-thirds of pachytene Tex19.1 -/spermatocytes [24]. However, the remaining one-third synapse their autosomes completely. These autosomally synapsed Tex19.1 -/spermatocytes therefore provide a system to study how delays in recombination might influence progression through pachytene in the absence of the confounding effects of asynapsis.
Here we show that progression through pachytene is perturbed in autosomally synapsed Tex19.1 -/spermatocytes. We show that the progression, timing or duration of recombination, chromosome axis elongation, chromatin modifications, and histone H1t expression are co-ordinately altered in Tex19.1 -/spermatocytes, and that the altered progression of Tex19.1 -/spermatocytes through pachytene depends on Spo11. These findings indicate that meiotic recombination is monitored by a checkpoint in spermatocytes that can delay aspects of progression through pachytene to accommodate cells with altered recombination kinetics.

Spermatocytes
We have previously shown that Tex19.1 -/spermatocytes have defects in the initiation of Spo11dependent early recombination foci [24]. The number of DMC1-containing recombination foci in Tex19.1 -/spermatocytes is reduced to around 30% of those present in littermate controls during leptotene, but increases to reach 87% of control levels during zygotene [24]. Thus, a significant number of recombination foci are being generated during zygotene rather than leptotene in Tex19.1 -/spermatocytes. These early recombination defects are accompanied by synapsis defects in some Tex19.1 -/pachytene spermatocytes, but the remaining fully synapsed pachytene spermatocytes must have possessed sufficient recombination foci to promote a successful homology search, and do not have any confounding autosomal asynapsis that would interfere with analysing the effects of delayed recombination on meiotic progression. We have therefore used these autosomally synapsed pachytene Tex19.1 -/nuclei to investigate whether delayed recombination might subsequently affect progression through the pachytene stage of meiosis.
Pachytene nuclei with complete autosomal synapsis were identified by immunostaining for synaptonemal complex components, and autosomally asynapsed pachytene nuclei, which are  Figure 1D), comparable to other studies [13,26]. Interestingly, autosomal RPA and DMC1 foci in autosomally synapsed Tex19.1 -/pachytene spermatocytes were significantly increased by 28% and 95% respectively ( Figure 1D). Violin plots of these data (Supporting Figure S1) suggest bimodal distributions for both RPA and DMC1, with one population of pachytene nuclei containing relatively high numbers of RPA or DMC1 foci, and a second population containing few or no RPA or DMC1 foci. The increase in RPA foci number in autosomally synapsed Tex19.1 -/pachytene spermatocytes appears to primarily reflect a shift towards the abundant foci population, whereas the increase in DMC1 foci number reflects both a shift towards the abundant foci population and an increase in the number of DMC1 foci in that population (Supporting Figure S1). RAD51 foci were not significantly increased in Tex19.1 -/spermatocytes ( Figure 1D), although it is not clear if this reflects a genuine difference between RAD51 and its meiotic homolog DMC1 rather than technical differences in antibody sensitivity. Regardless, the increased frequency of DMC1 and RPA foci in pachytene Tex19.1 -/spermatocytes indicates that, despite successful autosomal synapsis, these cells possess an abnormally high frequency of recombination foci participating in early stages of meiotic recombination.
To verify our analysis of early recombination intermediates in pachytene Tex19.1 -/spermatocytes we examined the behaviour of γH2AX (Figure 2A), a marker for unrepaired DNA damage [27,28].
In pachytene, γH2AX is typically detected as a strong cloud of staining over the asynapsed sex chromosomes, and as condensed axial foci (S-foci) on autosomes that co-localise with recombination foci [28]. The number of autosomal γH2AX S-foci decreases during pachytene and are undetectable in diplotene [28]. As distinguishing individual S-foci becomes somewhat subjective when there are large numbers of adjacent γH2AX S-foci on the same axis, we grouped  Figure   2B). Loss of Tex19.1 does not have any detectable effect on the frequency of Spo11-independent γH2AX foci in spermatocytes [24], suggesting that the increase in the number of pachytene spermatocytes with abundant γH2AX S-foci represents a difference in behaviour of Spo11dependent meiotic DSBs in these mutants. Therefore, in addition to containing elevated levels of early recombination foci, autosomally synapsed Tex19.1 -/pachytene spermatocytes more frequently contain abundant γH2AX S-foci suggesting there is more unrepaired DNA damage in these nuclei.

Spermatocytes
We next tested whether the increased frequency of early recombination intermediates and unrepaired DNA damage in pachytene Tex19.1 -/spermatocytes is accompanied with any changes in the frequency of spermatocytes containing markers of late recombination foci. Autosomally synapsed pachytene Tex19.1 -/spermatocytes were scored for the presence or absence of MLH1 ( Figure 2C), a component of late recombination foci that marks crossover recombination events [29]. 61% of control pachytene nuclei displayed MLH1 foci associated with autosomal axes, however this was reduced to just 20% among pachytene Tex19.1 -/nuclei ( Figure 2D). Thus, the population of autosomally synapsed pachytene Tex19.1 -/spermatocytes has more early recombination foci, is enriched for nuclei containing large numbers of γH2AX S-foci, and is depleted for nuclei containing late recombination foci.

Autosomally Synapsed Tex19.1 -/-Spermatocytes Establish Functional MSCI
The apparent enrichment of earlier substages in the population of pachytene Tex19.1 -/spermatocytes could reflect altered kinetics of recombination foci maturation during pachytene, or death of pachytene spermatocytes depleting later substages from the pachytene population. We have previously reported that loss of Tex19.1 causes cell death in some spermatocytes [22], but this is an expected consequence of the synapsis defects, which can cause MSCI failure and apoptosis during pachytene [6,7]. However, the analyses performed in Figures 1 and 2 specifically selected pachytene nuclei that had no autosomal asynapsis. Cell death within the autosomally synapsed population ought to result in fewer cells progressing to diplotene, however the proportion of autosomally synapsed pachytene : diplotene spermatocytes is not significantly altered in Tex19.1 -/mice (Supporting Figure S2). Although the simplest interpretation of these data is that there is no significant cell death in autosomally synapsed Tex19.1 -/pachytene spermatocytes, it is possible that MSCI is failing and causing cell death in a subset of fully synapsed pachytene Tex19.1 -/cells, and that delays in pachytene progression in the surviving cells masks any change the autosomally synapsed pachytene : diplotene ratio. We therefore assessed whether MSCI is occurring normally in the absence of autosomal asynapsis in Tex19.1 -/spermatocytes. HORMAD1 has a role in recruiting the transcriptional silencing machinery to asynapsed chromosomes [30], and defects in HORMAD dissociation from synapsed axes [11] could trigger MSCI failure in synapsed pachytene cells.
Moreover, γH2AX immunostaining [27,31] suggests that the sex body itself is forming normally in autosomally synapsed Tex19.1 -/synapsed pachytene nuclei (Figure 2A, asterisks). To test whether MSCI is being functionally established we immunostained meiotic chromosome spreads for RBMY, a Y-chromosome-encoded protein silenced by MSCI during pachytene [32]. RBMY was readily detected in asynapsed pachytene Tex19.1 -/spermatocytes containing asynapsed autosomes but not suggesting that MSCI is occurring normally in autosomally synapsed spermatocytes in the absence of Tex19.1. Thus, autosomally synapsed Tex19.1 -/pachytene spermatocytes do not have defects in MSCI that might trigger cell death in mid-late pachytene in the absence of asynapsis. Taken together, these data suggest that the enrichment of early recombination markers in autosomally synapsed Tex19.1 -/spermatocytes likely represents progression through meiotic prophase being altered in at least some of these cells.
Immunostaining control spermatocytes for ubH2A revealed a cloud of staining at the sex chromosomes ( Figure 4C), similar to that previously reported [35,36], in 87% of control pachytene spermatocytes ( Figure 4D). A cloud of ubH2A staining was also detected at the sex chromosomes in autosomally synapsed Tex19.1 -/pachytene spermatocytes ( Figure 4C), however the proportion of cells with such staining was reduced to 74% ( Figure 4D). Thus, the delay in maturation of recombination in autosomally synapsed Tex19.1 -/pachytene spermatocytes is accompanied by more widespread changes in progression through pachytene.
In addition to the specific establishment of H2A mono-ubiquitylation at the sex body, the patterns of total ubiquitylation also vary as spermatocytes progress through pachytene [37,38]. Ubiquitylation can be broadly monitored using the FK2 monoclonal antibody which recognises both mono-and poly-ubiquitylation [39]. Therefore we investigated the progression of ubiquitylation in Tex19.1 -/by immunostaining spermatocyte chromosome spreads with the FK2 antibody. Consistent with previous reports, FK2 staining in control pachytene spermatocytes is strikingly enriched at the sex chromosomes [37,38]. FK2 staining is initially restricted to the chromosome axes in the XY body in early pachytene, then extends throughout the XY body chromatin in mid and late pachytene [37,38].
Although both control and Tex19.1 -/autosomally synapsed pachytene spermatocytes exhibit enriched FK2 staining on the sex chromosomes ( Figure 5A), the frequency of pachytene nuclei with axial XY staining characteristic of early pachytene is higher in autosomally synapsed Tex19.1 -/spermatocytes ( Figure 5B). Thus, similar to the recombination markers, γH2AX, histone H1t, and ubH2A analyses, sex chromosome ubiquitylation patterns suggest that autosomally synapsed We also analysed the FK2 staining patterns on autosomes in pachytene Tex19.1 -/spermatocytes.  Figure 5D). Therefore the altered patterns of ubiquitylation in Tex19.1 -/spermatocytes are not limited to ubH2A at the sex body, but extend to more general ubiquitylation at the sex chromosomes and autosomes. Similar to the altered recombination profiles and histone H1t marker analysis, the perturbed ubiquitylation patterns suggest that the population of autosomally synapsed Tex19.1 -/spermatocytes is enriched for earlier substages of pachytene.

Tex19.1 -/-Spermatocytes Have Reduced Chromosome Axis Elongation During Pachytene
The data on recombination and DNA damage markers, histone H1t and ubiquitylation together indicate that fully synapsed pachytene Tex19.1 -/nuclei may be delayed in their progression from early to mid-late pachytene. However, it is not clear whether the autosomally synapsed Tex19.1 -/spermatocytes that reached late pachytene were still delayed in their meiotic progression.
Progression through pachytene is reported to be associated with elongation of the autosomal chromosome axes between early and late pachytene [40], therefore we used chromosome axis length to assess whether late pachytene Tex19.1 -/nuclei are still delayed in aspects of their meiotic progression. The late recombination marker MLH1 was used to select for spermatocytes in late pachytene for this analysis. Comparison of total axis length between Tex19.1 +/± and Tex19.1 -/-MLH1-positive autosomally synapsed spermatocytes revealed a 6.5% reduction in length in the  Table 1). Comparison between individual and groups of axes ranked by size, as has been performed in similar analyses [13], demonstrated that all chromosome sizes were similarly affected ( Figure 6B, Table 1). A significant reduction in length was not achieved for the smallest group of axes, though this is likely due to a greater degree of measurement error relative to total length. Thus, either elongation of the autosomal chromosome axes is delayed in fully synapsed late pachytene Tex19.1 -/spermatocytes as part of a general perturbation of progression through pachytene in these mutants, or Tex19.1 -/spermatocytes have shorter axes independent of the other changes in progression through pachytene.

Axis Ranks
Mean group axis length (arbitrary units) Difference Autosomes were ranked by size and grouped as indicated. The mean total axis lengths for each group of autosomes is shown. Four animals were analysed for each genotype and a total of 50 and 55 nuclei analysed for Tex19.1 +/± and Tex19.1 -/respectively. Student's t-test was used to test for statistical signficance between genotypes (* indicates p < 0.05, ** indicates p < 0.01.) To investigate whether loss of Tex19.1 results in shorter chromosome axes independently of altered progression through pachytene we used the repair of γH2AX S-foci to monitor progression through pachytene. Tex19.1 +/± and Tex19.1 -/spermatocytes both elongated their chromosome axes as they progressed through pachytene and repaired γH2AX S-foci ( Figure 6C). However, when autosomally are compared, no significant differences in axis length are evident ( Figure 6C). Therefore, the shorter chromosome axes in MLH1-positive Tex19.1 -/pachytene spermatocytes likely reflects the perturbed progression of these spermatocytes through pachytene, and axis length appears to be more closely associated with repair of γH2AX S-foci than appearance of MLH1 foci during pachytene .

Delayed Progression Through Pachytene in Tex19.1 -/-Spermatocytes is Dependent on Spo11
Taken together, the data presented in this manuscript suggests that multiple aspects of progression delayed recombination that we have identified would potentially allow or reflect the repair of recombination-dependent lesions in these cells before they continue further in the meiosis.