Reproductive Isolation in Hybrid Mice Due to Spermatogenesis Defects at Three Meiotic Stages

Ayako Oka,*,† Akihiko Mita,† Yuki Takada,‡ Haruhiko Koseki‡ and Toshihiko Shiroishi†,1 *Transdsciplinary Research Integration Center, Research Organization of Information and Systems, Toranomon, Tokyo, Japan 105-0001, †Mammalian Genetics Laboratory, National Institute of Genetics, Mishima, Shizuoka, Japan 411-8540 and ‡RIKEN Research Center for Allergy and Immunology, Yokohama, Kanagawa, Japan 230-0045 Manuscript received May 17, 2010 Accepted for publication June 27, 2010

ABSTRACT Early in the process of speciation, reproductive failures occur in hybrid animals between genetically diverged populations. The sterile hybrid animals are often males in mammals and they exhibit spermatogenic disruptions, resulting in decreased number and/or malformation of mature sperms. Despite the generality of this phenomenon, comparative study of phenotypes in hybrid males from various crosses has not been done, and therefore the comprehensive genetic basis of the disruption is still elusive. In this study, we characterized the spermatogenic phenotype especially during meiosis in four different cases of MSM MSM reproductive isolation: B6-ChrX , PGN-ChrX , (B6 3 Mus musculus musculus-NJL/Ms) F1, and (B6 3 Mus spretus)F1. The ﬁrst two are consomic strains, both bearing the X chromosome of M. m. molossinus;in B6-ChrXMSM, the genetic background is the laboratory strain C57BL/6J (predominantly M. m. domesticus), while in PGN-ChrXMSM the background is the PGN2/Ms strain purely derived from wild M. m. domesticus. The last two cases are F1 hybrids between mouse subspecies or species. Each of the hybrid males exhibited cell-cycle arrest and/or apoptosis at either one or two of three distinct meiotic stages: premeiotic stage, zygotene-to-pachytene stage of prophase I, and metaphase I. This study shows that the sterility in hybrid males is caused by spermatogenic disruptions at multiple stages, suggesting that the responsible genes function in different cellular processes. Furthermore, the stages with disruptions are not correlated with the genetic distance between the respective parental strains.

HEN animals from genetically diverged popula- animals include Drosophila (Joly et al. 1997), stickle- W tions hybridize, complete or partial sterility is back fish Pungitius (Takahashi et al. 2005), cavio- often observed in the F1 hybrids or in their descend- morph rodent Thrichomys (Borodin et al. 2006), ants. This phenomenon belonging to postzygotic re- house musk shrew Suncus (Borodin et al. 1998), wallaby productive isolation accelerates irreversible genetic Petrogale (Close et al. 1996), and genus Mus (Forejt divergence by preventing free gene flow across the and Ivańyi 1974; Matsuda et al. 1992; Hale et al. 1993; two diverging populations, and thereby plays a pivotal Yoshiki et al. 1993; Kaku et al. 1995; Gregorova´ and role in speciation. Sexual dimorphism is a general Forejt 2000; Elliott et al. 2001, 2004; Good et al. feature of reproductive isolation (Wu and Davis 1993; 2008). Although reproductive isolation by spermato- Laurie 1997; Orr 1997; Kulathinal and Singh 2008). genic impairment is a well-known phenomenon, its In mammals, impairment is much more severe in males underlying genetic mechanism and molecular basis than in females, and in general the heterogametic sex is have remained elusive. The Dobzhansky–Muller model, more sensitive to interspecific and intersubspecific which infers that hybrid sterility or inviability is caused genetic incompatibility. This phenomenon is well known by deleterious epistatic interactions between nuclear as Haldane’s rule (Haldane 1922; Laurie 1997; Orr genes derived from their respective parent species or 1997). subspecies (Dobzhansky 1936; Muller 1942), is widely In many animals, the reproductive isolation is caused accepted in animals and plants and is also applicable to by spermatogenic disruptions characterized by reduced the sterility of hybrid animals in F2 or backcross number of germ cells and small testis size. These generations, so-called hybrid breakdown, in which the genes causing postzygotic reproductive isolation are rr Supporting information is available online at http://www.genetics.org/ partially recessive (O 2005). cgi/content/full/genetics.110.118976/DC1. The genetic incompatibility between house mouse Available freely online through the author-supported open access subspecies is an ideal animal model for studying the option. early stage of speciation. Two subspecies of mouse, Mus 1Corresponding author: Mammalian Genetics Laboratory, National Institute of Genetics, Yata 1111, Mishima, Shizuoka, Japan 411-8540. musculus domesticus and M. m. musculus, diverged from E-mail: [email protected] their common ancestor 0.3–1.0 MYA (Yonekawa et al.

Genetics 186: 339–351 (September 2010) 340 A. Oka et al.

1980; Moriwaki 1994; Bonhomme and Gueńet 1996; sperms. We extended the same analysis to three other Boursot et al. 1996; Din et al. 1996). M. m. domesticus cases of reproductive isolation. Another consomic strain ranges across western Europe and the Middle East, PGN-ChrXMSM has an MSM/Ms-derived X chromosome whereas M. m. musculus ranges throughout eastern in the genetic background of the PGN2/Ms strain Europe and northern Asia (Bonhomme and Gueńet derived from wild mice (M. m. domesticus). PGN- 1996). The two subspecies meet in a narrow hybrid ChrXMSM males produce a small number of dysfunc- zone, which is most likely maintained by a balance tional sperms as was the case with B6-ChrXMSM males, but between dispersal and selection against hybrids (Hunt the former males show apoptosis mainly at metaphase of ´ and Selander 1973; Bonhomme and Guenet 1996; meiosis I. Furthermore, we examined F1 hybrid males Payseur et al. 2004). M. m. domesticus also displays from intersubspecific cross of (B6 3 M. m. musculus- reproductive isolation from the Japanese wild mouse, NJL/Ms) and interspecific cross of (B6 3 M. spretus). M. m. molossinus, which originated from hybridization of These F1 hybrid males exhibited apoptosis at metaphase M. m. castaneus and M. m. musculus and its nuclear I and at the zygotene-to-pachytene stage of prophase I. genome is predominantly derived from M. m. musculus As a whole, the postzygotic reproductive isolation in (Yonekawa et al. 1980, 1988; Moriwaki 1994; Sakai mice is caused by disruptions at a minimum of three et al. 2005). To investigate the reproductive isolation different spermatogenic stages. between M. m. domesticus and M. m. molossinus, we previously constructed a consomic strain B6-ChrXMSM (Oka et al. 2004). This strain has the X chromosome MATERIALS AND METHODS from the MSM/Ms strain, which is derived from M. m. Mouse strains: Construction of the consomic strain B6- molossinus, in the genetic background of the laboratory ChrXMSM was described in our previous report (Oka et al. strain C57BL/6J (B6), which is predominantly derived 2004). The strain is maintained in the animal facility of the from M. m. domesticus (Moriwaki 1994). F1 hybrid National Institute of Genetics (NIG). For maintenance of B6- MSM animals between B6 and MSM/Ms strains are fully ChrX , females that have the MSM-derived X chromosome MSM heterozygously are selected by genotyping with 12 X-linked fertile. On the contrary, B6-ChrX shows male-specific microsatellite markers and backcrossed to B6 males at each sterility characterized by a reduced sperm number and backcross generation. From these backcrosses, males with the dysfunction of the sperm, including abnormal mor- MSM-derived X chromosome and males with the B6-derived X phology and low motility, indicating that B6-ChrXMSM is chromosome are obtained at a frequency of 12.5% each. a model of hybrid breakdown in animals (Oka et al. Because the reproductive phenotype of males with the B6- derived X chromosome (B6-ChrXB6) is almost identical to that 2004, 2007). Our previous study indicated that the of B6 males (Oka et al. 2004), we used the B6-ChrXB6 males as abnormal morphology of the sperm head results from controls in this study. the genetic incompatibility between MSM/Ms-derived The proximal and distal microsatellite markers used for X-linked genes and B6 genes on autosomes including the maintenance are DXMit89 at 10.0 Mb and DXMi160 at MSM chromosomes 1 and 11 (Oka et al. 2007). 162.5 Mb, respectively. Thus, B6-ChrX strain possibly has B6 genome in the proximal side beyond DXMit89 and the distal In this study, to understand the genetic mechanism of side beyond DXMit160 (supporting information, Figure S1). To reproductive isolation in mice, we first undertook in- clarify the origin of the pseudoautosomal region, which starts depth characterization of phenotype for each B6- from the middle of Mid1 gene at 165.2 Mb toward telomere, in ChrXMSM male especially during meiosis. Meiosis is a B6-ChrXMSM strain, we designed several primer sets for PCR, special cell division that produces four haploid cells which can distinguish between B6 and MSM/Ms alleles. The primers are: X163466600 Forward (F) primer 59-CAAGCTGTG after one round of chromosome replication and two CTGAATATCTGGTG-39 and reverse (R) primer 59-GCAGCCA rounds of chromosome segregation. During meiosis, GGCTTAAGGTAGGACC-39); X164174029 F-primer 59-CGTCA homologous chromosomes pair, synapse, undergo TCTCTCAAAGCTCTTCAC-39 and R-primer 59-GAAGATCTG crossing over, and achieve bipolar attachment to the TAATGTTTTACTGGG-39); X164783477 F-primer 59-CACTTG spindle to segregate one set of chromosomes to each AGGGAGAAATGTATTGTC-39 and R-primer 59-CCCTAGAGG TCTATATGAAGCC-39); X165318225 F-primer 59-GTCCTGTA daughter cell. Homologous recombination is initiated ATCCACCTTCACTGG-39 and R-primer 59-CATTCACAAAATA during the leptotene stage of meiotic prophase I with TCTAGTGCTCC-39); X165319677 F-primer 59-AAATGAGTA the formation of DNA double-strand breaks (DSBs), TGGAATGACCCAGC-39 and R-primer 59-CACAGCCTAGGAA which are repaired immediately during the zygotene CCATGCTGGCC-39. Genotyping of B6-ChrXMSM showed that stage or after crossing over of homologous chromo- the MSM genome remained only at marker X163466600, and oeder the other regions were completely substituted by B6 genome, somes during the pachytene stage (R 1997; indicating that the pseudoautosomal region of B6-ChrXMSM Tarsounas and Moens 2001). strain is derived from the B6 strain (Figure S1). During the first wave of spermatogenesis, most M. spretus strain (BRC no. RBRC00208) was provided by mitotic spermatogonia in the B6-ChrXMSM testes fail to RIKEN BRC (supported by the National Bio-Resource Project initiate meiotic DNA replication. Some proportion of of MEXT, Japan). The NJL/Ms inbred strain was derived from wild M. m. musculus mice trapped in Northern Jutland, those spermatogonia that enter into meiosis are again Denmark, and maintained by NIG. PGN-ChrXMSM strain was arrested and eliminated by apoptosis at the pachytene constructed by the same strategy used in the construction of stage, resulting in the production of a small number of B6-ChrXMSM. The inbred strain PGN2/Ms was derived from Meiotic Disruptions in Hybrid Males 341

M. m. domesticus mice trapped in Pigeon, Canada, and main- (1:1000), which was generously provided by Dr. Handel tained by NIG. (Inselman et al. 2003), and mouse monoclonal anti-XLR Histology and detection of apoptosis: Testes were placed (1:300), which was generously provided by Dr. Garchon in Bouin’s fixative and then embedded in paraffin. Sections (Escalier and Garchon 2000). For the immunocytochemis- (6 mm) were stained with hematoxylin and eosin (HE). try of gH2AX, we used anti-SYCP3 antibody labeled by Zenon Apoptotic cells were visualized by TUNEL assay (In Situ Cell rabbit IgG labeling kits (Molecular Probes). Death Detection kit, AP; Roche) following the manufacturer’s protocol. To examine the progression of the first-wave spermatogenesis, 50 tubules were randomly chosen from a RESULTS section of each male and were classified on the basis of the cell MSM types in the innermost layer of epithelium. Germ cells in the Failure to initiate meiotic S phase in B6-ChrX epithelium were classified according to the book by Russell testes: We previously reported variation in the degree of et al. (1990). Three individuals from each strain were used for testicular defects across seminiferous tubules within the evaluation. individual adult B6-ChrXMSM males, ranging from rela- Immunohistochemistry: For immunohistochemistry, testes tively normal spermatogenesis to loss of all germ cells were fixed in Mildform 20NM (Wako) and embedded in ka paraffin. After deparaffinization, antigen retrieval was per- except spermatogonia (O et al. 2004). To learn when formed by boiling the sections in 0.01 m citrate buffer, pH 6.0, the disruption of spermatogenesis starts, we examined in a microwave for 5 min. After blocking with 10% fetal calf testicular histology of B6-ChrXMSM males at various stages serum (FCS) in phosphate-buffered saline (PBS), the sections in the first wave of spermatogenesis, which progresses were incubated with primary antibody for 1 hr at room synchronously in mice. After the several rounds of temperature (RT). After washing, the sections were incubated with Alexa-Fluor-488 conjugated secondary antibody at RT mitosis, type B spermatogonia enter the meiotic prelep- for 30 min. Then, after washing, the sections were counter- totene stage, during which meiotic DNA replication stained with Hoechst 33258 in PBS. Primary antibodies used occurs, at around 10 days post partum (dpp) (Ellis were rabbit polyclonal anti-gH2AX (1:50; Upstate), rabbit et al. 2004). Then, meiotic spermatocytes enter the polyclonal anti-SYCP3 (1:50; Novus Biologicals), mouse mono- prolonged prophase, which takes 10–12 days (Ellis clonal anti-PLZF (1:50, Santa Cruz), and rat monoclonal anti- CBX1/HP1b (1:50, Abcam). et al. 2004). The first meiotic prophase is divided into For bromodeoxyuridine (BrdU) staining, we injected mice five stages: leptotene, zygotene, pachytene, diplotene, intraperitoneally with 40 mg of BrdU (Sigma) per kilogram and diakinesis stages. Pachytene stage is the longest one, body weight. After 1 hr, testes were fixed in Mildform 20NM continuing for 5–7 days, and the first pachytene sperma- and processed in paraffin. We detected BrdU incorporation tocytes are observed at around 14 dpp (Ellis et al. 2004; with monoclonal antibody to BrdU (Sigma), incubated samples ohen with biotinylated secondary antibody and streptavidin-HRP, C et al. 2006). MSM and then detected with diaminobenzidine. Histological examination of B6-ChrX testes by light Testicular cell spread preparations: For spermatocyte microscopy revealed no detectable difference from those preparations, 18- or 19-day-old mice were used. After removal of control littermates at 8 dpp (data not shown). of tunica albuginea, testes were minced in 10% FCS/PBS and Perceptible changes in the constitution of germ cells m large tissues were removed with 70- m nylon mesh. Collected within the tubules in B6-ChrXMSM testes became visible at cells were washed twice with 10% FCS/PBS and then fixed for 1 min in a solution of 1.85% formaldehyde and 0.05 m sucrose 10 dpp (Figure 1A). At 14 dpp, a clear difference in in PBS. After washing with 10% FCS/PBS, cells were pelleted histology was noted between B6-ChrXMSM and control and resuspended in a small volume of 10% FCS/PBS (,20 ml/ testes: meiotic spermatocytes in the seminiferous tubules testis). Cells were then smeared on APS-coated glass slides were abundant in control animals, but rarely seen in the (Matsunami) and air dried. Specimens were used for immu- B6-ChrXMSM animals (Figure 1A). At 19 dpp, many nostaining immediately after the preparation. MSM For immunocytochemistry, slides were immersed in 0.5% tubules in the B6-ChrX testes still contained no cells Triton-X 100/PBS for 10 min, then washed with PBS and engaged in meiosis (Figure 1A). The disruption of incubated with 10% FCS/PBS for 1 hr at RT to block spermatogenesis became more severe in the B6-ChrXMSM nonspecific binding. Primary antibodies were applied and testes at 32 dpp. Some tubules contained surviving incubated either overnight at 4° or 1 hr at RT. After washing spermatocytes and round spermatids, but none of the with 0.1% Triton-X 100 in PBS (PBST), slides were incubated with Alexa-Fluor-conjugated secondary antibodies for 30 min mature sperm cells that were observed in most tubules of at RT. Following PBST washes, slides were mounted in the control testes (Figure 1A). These results show that Vectashield medium (Vector Laboratories). Staging of pro- spermatogenesis of B6-ChrXMSM mice is affected in the phase spermatocytes was based on SYCP3 staining on auto- meiotic phase. We quantified the progression of sper- shley urner somes and the sex chromosomes (A et al. 2004; T matogenic development by counting tubules containing et al. 2004). The substaging of pachytene spermatocytes was based on H1t staining (Yang et al. 2008). Fluorescent signals germ cells of the various developmental stages in the were detected under a Carl Zeiss laser scanning confocal innermost layer of seminiferous epithelium. As summa- microscope (LSM510). Images of 50–80 cells for each male rized in Figure 1B, delayed timing of emergence of were captured for quantitative analysis. Primary antibodies developing germ cells was detected in the B6-ChrXMSM used were rabbit polyclonal anti-SYCP3 (1:200; Novus Bio- testes at all ages tested. logicals), goat polyclonal anti-ATR (1:20; Santa Cruz), rabbit polyclonal anti-gH2AX (1:200; Upstate), mouse monoclonal To examine whether mitotic spermatogonia are MSM anti-RAD51 (1:20, Abcam), mouse monoclonal anti-MLH1 affected in B6-ChrX , we performed immunohisto- (1:20; BD Pharmingen), guinea pig polyclonal anti-H1t chemical analysis of the testicular sections with anti- 342 A. Oka et al.

bodies against PLZF and CBX1/HP1b, which are localized to nuclei of undifferentiated type A and type B spermatogonia, respectively (Hoyer-Fender et al. 2000; Buaas et al. 2004; Costoya et al.2004).The antibodies used for immunostaining in this study are listed in Table 1. Control and B6-ChrXMSM testes showed no obvious differences in the number of undifferentiated type A and type B spermatogonia at 14 dpp (Figure S2). We further investigated proliferation of spermatogonia and spermatocytes in the testes at 19 dpp by analyzing the incorporation of BrdU. We distinguished proliferative spermatogonia from spermatocytes on the basis of size and number of BrdU-positive nuclei in the seminiferous tubules: tubules containing preleptotene spermatocytes were recognized as those with a larger number of small positive nuclei. In the control testes, BrdU incorporation was detected in both spermatogonia and preleptotene spermatocytes (Figure 2A). In contrast, in the B6-ChrXMSM testes, BrdU signals were detected mainly in spermatogonia, and rarely in the preleptotene spermatocytes (Figure 2B). Although most of tubules in the B6-ChrXMSM testes were aberrant, the frequency of tubules containing BrdU-positive germ cells was not significantly different between control and the B6-ChrXMSM testes (31.8 and 27.0% for the two control males; 36.1 and 28.9% for the two B6-ChrXMSM males), suggesting relatively normal proliferation of spermatogonia in B6-ChrXMSM. We next investigated whether B6-ChrXMSM germ cells enter meiosis. Immunohistochemistry was performed on testis sections at 14 dpp with antibody against the phosphorylated form of histone H2AX, gH2AX. Since gH2AX is observed in chromatin immediately after the induction of DSBs at the leptotene and zygotene stages and on the sex chromosomes at the pachytene stage (Mahadevaiah et al. 2001), we used gH2AX as a marker of early spermatocytes. Almost all tubules of the control testes contained spermatocytes at the leptotene-to- zygotene stage or the pachytene stage (Figure 2C). By contrast, in the B6-ChrXMSM testes, a small proportion of tubules showed gH2AX-positive spermatocytes at the leptotene-to-zygotene stage but very few tubules showed spermatocytes at the pachytene stage (Figure 2D). We then quantified frequency of early spermatocytes by detecting SYCP3-positive cells. SYCP3 is a component of the synaptonemal complex that physically tethers homologous chromosomes and is localized in different manners from leptotene stage to metaphase I. SYCP3 therefore igure F 1.—Testicular histology and spermatogenic devel- serves as a marker of early spermatocytes. We performed opment in B6-ChrXMSM testes during first-wave spermatogenesis. (A) Hematoxylin and eosin (HE) staining of cross immunocytochemical analysis with SYCP3 antibody for MSM sections of testes from 10, 14, 19, and 32 dpp from control cell spreads from B6-ChrX and control testes at 18 or B6-ChrXB6 males (left) and B6-ChrXMSM males (right). The 19 dpp and quantified the ratio of SYCP3-positive early B6-ChrXMSM testes showed few cells engaged in meiosis spermatocytes to all testicular cells stained by Hoechst and vacuolation in the seminiferous tubules at ages after 14 dpp. Scale bar: 100 mm. (B) The proportion of developing germ cells in the innermost layer of seminiferous tubules in control (left) and B6-ChrXMSM (right) males during tocytes were grouped together as pre-pachytene spermato- first-wave spermatogenesis. Leptotene and zygotene sperma- cytes. Meiotic Disruptions in Hybrid Males 343

TABLE 1 B6-ChrXMSM testes, proliferation and differentiation are Marker proteins used for immunoassays normal prior to type B spermatogonia, but initiation of meiotic DNA replication is prevented in the majority of Protein Feature marked type B spermatogonia, allowing only a small proportion of the spermatogonia to enter into meiosis. PLZF Undifferentiated type A spermatogonia Synaptic errors and apoptosis at pachytene stage in CBX/HP1b Type B spermatogonia MSM gH2AX DSBs (leptotene-zygotene); XY body the B6-ChrX spermatocytes: We next characterized (pachytene) the early meiotic spermatocytes found in the B6- SYCP3 Synaptonemal complex ChrXMSM testes. During prophase I, synapsis of the (leptotene-metaphase I) homologous chromosomes can be monitored by local- ATR XY body (pachytene); unrepaired DSBs ization of SYCP3 (Cohen et al. 2006; Hamer et al. 2006). on unsynapsed autosomes (pachytene) The lateral axes of synaptonemal complexes are formed XLR XY body (pachytene) during the leptotene stage, and synapsis between H1t Mid-to-late pachytene RAD51 Early recombination nodules on DSBs homologous chromosomes occurs during the zygotene (leptotene-pachytene) stage and becomes complete prior to the pachytene MLH1 Late recombination nodules on DSBs stage (Cohen et al. 2006). Unsynapsed and synapsed (mid-to-late pachytene) chromosomes can be distinguished by fine vs. thick fibers of SYCP3. We observed that the pattern of SYCP3 signals in B6-ChrXMSM was similar to that in the control 33258. As shown in Figure 2E, almost half (46.9%) of cells spermatocytes during leptotene to zygotene (Figure in the control testes were SYCP3 positive, whereas only a S3A). Because pachynema is a prolonged stage, we small proportion (6.4%) of cells in the B6-ChrXMSM testes divided it into two substages (early and mid-to-late) on were SYCP3 positive. All these data indicate that in the the basis of the testis-specific H1 variant H1t, which

Figure 2.—Failure in initiation of meiotic S-phase and reduced proportion of meiotic spermatocytes in B6-ChrXMSM testes. BrdU incorporation into control (A) and B6-ChrXMSM (B) testes at 19 dpp was detected by anti-BrdU antibody. In control testis, BrdU signals were detected in type A spermatogonia, intermediate spermatogonia (In), type B spermatogonia (SB) and preleptotene spermatocytes (preL). In B6-ChrXMSM testis, anti-BrdU antibody detected type A and B spermatogonia but not preleptotene spermatocytes. (C and D) Meiotic spermatocytes were indicated by gH2AX signals in testes at 14 dpp. In control testis (C), leptotene to zygotene spermatocytes (LZ) and pachytene spermatocytes with XY bodies (P) were detected. In contrast, a few leptotene to zygotene spermatocytes were detected in B6- ChrXMSM testis (D). Signals are gH2AX (green) and Hoechst (gray). (E) Cell spreads from control (top) and B6-ChrXMSM (bottom) testis at 19 dpp. Anti-SYCP3 antibody (red) stained spermatocytes at meiotic prophase I. All nuclei were stained by Hoechst (blue). Only a few cells were SYCP3 positive in B6-ChrXMSM testis. Frequency of spermatocytes at prophase I in control and B6-ChrXMSM testes at 18–19 dpp (right graph). We counted 1731 and 1473 cells from four control and B6-ChrXMSM males, respectively. The difference in frequency was signiﬁcant by t-test. Scale bars: 200 mm. 344 A. Oka et al.

Figure 3.—Synaptic errors and cell-cycle arrest at early pachytene stage in B6-ChrXMSM spermatocytes. (A) SYCP3 and H1t double immunostaining of control and B6-ChrXMSM spermatocytes at early and late pachytene substages. The early pachytene spermatocyte from B6-ChrXMSM showed aggregation of SYCP3 (arrows). (B) SYCP3 and ATR double immunostaining of control and B6-ChrXMSM early pachytene spermatocytes. ATR localized to the sex chromosomes in the control spermatocyte. Note the dot- ted ATR signals along asynaptic regions on autosomes in the B6- ChrXMSM spermatocyte. (C) Fre- quencies of aggregation and asynapsis at each substage. We counted 1035 pachytene spermatocytes from nine control testes and 1008 from seven B6-ChrXMSM testes. (D) Proportion of early pachytene and mid-to-late pachytene spermatocytes, which were subtyped by staining with anti- H1t antibody. We counted 1463 pachytene spermatocytes from seven control males and 1334 from ﬁve B6-ChrXMSM males. All P-values were evaluated by t-test.

replaces somatic H1 on chromatin at mid-pachytene during early pachytene substage. We then estimated and thus serves as a marker of mid-to-late pachytene the proportion of spermatocytes in the two pachytene spermatocytes (Moens 1995; Cobb et al. 1999). In most substages. Compared with the control B6-ChrXB6, B6- early pachytene spermatocytes in B6-ChrXMSM males, ChrXMSM had a significantly reduced proportion of the SYCP3 signals were observed as thick fibers as observed mid-to-late pachytene spermatocytes (Figure 3D). This in control males, but B6-ChrXMSM spermatocytes also suggests that cell-cycle arrest occurred during the early frequently exhibited aggregation of SYCP3 (Figure 3A). pachytene substage in the B6-ChrXMSM males. This SYCP3 aggregation was also found in mid-to-late To examine whether DSBs and recombination nod- pachytene B6-ChrXMSM spermatocytes, although less ules are properly formed in B6-ChrXMSM males, we frequently (Figure 3C). We then immunostained the immunostained the early spermatocytes with antibodies pachytene spermatocytes with antibody against the against gH2AX and RAD51 as markers of DSBs and early kinase ATR, which during pachytene stage localizes on recombination nodules, respectively. We found no the sex chromosomes, but it also persists on unrepaired obvious differences in the patterns of gH2AX and DSBs specifically on asynaptic autosomes (Keegan et al. RAD51 signals between control and the B6-ChrXMSM 1996; Moens et al. 1999). Control pachytene spermato- testes in leptotene, zygotene, and pachytene spermatocytes showed ATR only on the sex chromosomes, cytes (Figure S3, A and B). We next immunostained the whereas some proportion of B6-ChrXMSM pachytene pachytene spermatocytes with antibody against MLH1, a spermatocytes showed ATR on the asynaptic autosomes protein that functions in meiotic crossing over and DNA (Figure 3B). Such spermatocytes with asynapsis in- mismatch repair in the late recombination nodules cluded both spermatocytes with and without signals on from the mid-pachytene substage (Ashley et al. 2004). the sex chromosomes. Frequencies of spermatocytes Again, we found no significant difference in the pattern with SYCP3 aggregation and asynapsis were significantly of MLH1 foci between control males and B6-ChrXMSM higher in B6-ChrXMSM than in the control, although the (Figure S3, A and B). variance between individuals was relatively large (Figure During pachytene stage, a specialized meiotic chro- 3C). Those aberrant chromosomes were prominent matin domain named the XY (sex) body is formed, in Meiotic Disruptions in Hybrid Males 345

Figure 4.—Increased apoptosis at the pachytene stage in B6-ChrXMSM testes. (A) Incidence of apoptosisintestesat15 and 19 dpp. Occasional apoptosis was observed in both control and B6- ChrXMSM testes. In B6- ChrXMSM testis at 19 dpp, extensive apoptosis was observed in the tubules with meiotic spermatocytes. Scale bar: 100 mm. (B) The frequency of seminiferous tubules with apoptotic cells during first-wave spermatogenesis. Fifty tubules for each of three males were counted at each age tested. Apoptotic cells were judged by the condensed staining of nuclei by HE staining. which X and Y chromosomes are transcriptionally si- somal consomic strain, PGN-ChrXMSM. Because the lenced. To evaluate the proper development of pachy- classical laboratory inbred strains, including B6, have tene spermatocytes, we examined the XY body formation contributions of multiple subspecies to their genome, it with antibodies against three marker proteins, ATR, was uncertain whether meiotic defects of the B6- gH2AX, and XLR, which during pachytene stage all ChrXMSM males indeed reflect the reproductive isolation accumulate exclusively in the XY body (Calenda et al. between M. m. domesticus and M. m. molossinus. Thus, we 1994; Escalier and Garchon 2000; Reynard et al. constructed the consomic stain PGN-ChrXMSM with the 2007). Most of the spermatocytes of the B6-ChrXMSM X chromosome from the MSM/Ms strain in the genetic males properly formed an XY body, although the marker background of the PGN2/Ms strain, which is derived proteins were also frequently detected in autosomal purely from wild mice (M. m. domesticus). PGN-ChrXMSM regions (Figure S4, A and B). males showed phenotypes similar to those of B6- Next, we examined whether the spermatogonia ChrXMSM males: sterility with reduced testis weight and arrested prior to the meiotic entry and spermatocytes abnormal morphology of mature sperm (Table S1, arrested at pachytene stage are subject to apoptosis by Table S2, and Figure S6). This result indicated that using the TUNEL assay on testicular sections at 15 and PGN-ChrXMSM males show reproductive isolation like 19 dpp. At 15 dpp, occasional TUNEL-positive cells were that of B6-ChrXMSM males. Immunohistochemical anal- present in some tubules of both control and B6- ysis with anti-SYCP3 antibody revealed that the number ChrXMSM testes (Figure 4A). This suggests that most of of SYCP3-positive early spermatocytes was almost iden- the arrested B6-ChrXMSM spermatogonia at the pre- tical in PGN-ChrXMSM testes and in control PGN- meiotic stage did not induce apoptosis. At 19 dpp, ChrXPGN and B6-ChrXB6 testes (Figure 5A). This numerous apoptotic cells were detected in the tubules suggests that early spermatocytes in PGN-ChrXMSM, in the B6-ChrXMSM testes, but not in control testes unlike those in B6-ChrXMSM, are not arrested or elimi- (Figure 4A). Apoptosis was rarely observed in severely nated. On the other hand, histological analysis of defective tubules lacking meiotic spermatocytes (Figure HE- and TUNEL-stained testes showed that the PGN- S5A). Most of the apoptotic spermatocytes were found ChrXMSM testes have increased apoptosis in spermato- in tubules at epithelial stage IV, as judged by the cytes at epithelial stage XII, which is equivalent to presence of mitotic intermediate spermatogonia and metaphase I, rather than stage IV (Figure 5B). early type B spermatogonia (Figure S5B). We compared The second case of reproductive isolation is F1 hybrid numbers of tubules containing apoptotic cells in sec- males from an intersubspecific cross of B6 and the NJL/ tions of testes at various ages stained with HE. The Ms strain derived from east European wild mice (M. m. frequency of tubules containing apoptotic cells became musculus). These hybrids show male sterility with massive higher in the B6-ChrXMSM testes than in the control loss of germ cells during meiosis (Kaku et al. 1995). testes starting at 19 dpp during first-wave spermatogen- Their testicular histology revealed extensive apoptosis in esis (Figure 4B). spermatocytes at the zygotene and pachytene stages Meiotic cell-cycle arrest in other cases of reproduc- (Figure 6A). Immunohistochemical analysis showed tive isolation: We further investigated the meiotic that the (B6 3 NJL/Ms) F1 hybrid males had a phenotypes in three other cases of reproductive iso- significantly decreased proportion of SYCP3-positive lation. The first case is males of a second X chromo- early spermatocytes (Figure 6B). We observed a smaller 346 A. Oka et al.

Figure 5.—Increased apoptotic spermatocytes at metaphase I in PGN-ChrXMSM testis. (A, left) Cell spreads from PGN-ChrXMSM testis at 19 dpp were immunostained with anti-SYCP3 antibody (red) and counterstained by Hoechst (blue). (A, right graph) Frequency of SYCP3-positive spermatocytes in control and PGN-ChrXMSM testes at 18–19 dpp. The difference was not sig- niﬁcant by t-test. (B) Testicular histology of control PGN-ChrXPGN (top left) and PGN-ChrXMSM (top right) males at 24 dpp. TUNEL staining of adult control PGN-ChrXPGN (bottom left) and PGN-ChrXMSM (bottom right) testes. In- creased apoptosis of metaphase I spermatocytes was observed at epithelial stage XII in PGN- ChrXMSM testis (arrowheads in top right and positive nuclei in bottom right). Scale bars: 200 mm.

proportion of pachytene spermatocytes relative to zygo- at epithelial stage XII, equivalent to metaphase I (Figure tene spermatocytes in the (B6 3 NJL/Ms) F1 testes 6A). We further carried out immunohistochemical anal- compared to control testes (data not shown). Since ysis of (B6 3 M. spretus)F1 hybrid males with antibodies these testes contained only a small number of H1t- directed against several meiotic marker proteins. The pro- positive mid-to-late pachytene spermatocytes, we exam- portion of SYCP3-positive early spermatocytes in (B6 3 ined the marker proteins only at the early pachytene M. spretus)F1 hybrid males was not significantly differ- substage. Immunostaining with anti-ATR and anti- ent from that of control B6-ChrXB6 males (Figure 6B). gH2AX antibodies showed that these markers of XY Immunostaining with ATR and gH2AX antibodies on bodies rarely accumulated in the F1 hybrid spermato- cell spreads from the F1 hybrid testes showed a slight but cytes (Figure 6C). Furthermore, their pachytene sper- significant decrease in the frequency of XY body matocytes showed a significantly elevated frequency formation, especially at the early pachytene substage of persistent RAD51 foci and a lack of MLH1 foci on (Figure 6C). Furthermore, the axes of (B6 3 M. spretus) asynaptic axes (Figure 6C). These results suggest that F1 hybrid pachytene spermatocytes displayed persistent DSBs were not repaired and homologous crossing over RAD51 foci and lack of MLH1 foci (Figure 6C). This was not completed in nearly all zygotene-to-pachytene indicates that DSBs were not repaired and homologous spermatocytes in (B6 3 NJL/Ms) F1 hybrid males. crossing over was not completed in the small population The last case of reproductive isolation is F1 hybrid of spermatocytes in the F1 hybrid, although the sperma- males from the interspecific cross of B6 and M. spretus.It tocytes were mainly impaired at metaphase I. was reported that (B6 3 M. spretus)F1 hybrid males are sterile, showing spermatogenic breakdown at the first DISCUSSION meiotic metaphase and only a small number of sperm, which are dysfunctional (Matsuda et al. 1991, 1992). Figure 7 summarizes the spermatogenic stages at Histological analysis of the F1 hybrid testes revealed a which four different cases of reproductive isolation significant increase of apoptosis in spermatocytes, mainly undergo cell-cycle arrest and apoptosis. Meiotic Disruptions in Hybrid Males 347 348 A. Oka et al.

For B6-ChrXMSM, we showed that the number of type B males showed relatively mild synaptic defect and no spermatogonia is unchanged, but most of them fail to persistence of unrepaired DSBs. Thus, cell-cycle arrest initiate meiotic DNA replication, suggesting that cell- and apoptosis at pachytene stage in this strain less likely cycle arrest occurs at the stage prior to the meiotic entry. resulted from pachytene checkpoint response and are Premeiotic cell-cycle arrest in spermatogonia has sel- rather attributable to the incompetence of meiosis; the dom been reported. One report showed that the early pachytene spermatocytes cannot complete cellular knockout mutant of Cdk inhibitor p27, which negatively events due to the genetic incompatibility and may be regulates the G1/S transition, has a small proportion of hard to reach the mid-to-late pachytene stage. spermatogonia that fail to undergo meiotic entry A recent study demonstrated that one of the genes (Beumer et al. 1999). Thus, B6-ChrXMSM is the first responsible for the reproductive isolation between the example in which spermatogonia are predominantly M. m. musculus-derived PWD strain and B6 is the histone subject to cell-cycle arrest at the premeiotic stage. In this H3 lysine 4 trimethyltransferase Prdm9 (Mihola et al. study, during the first wave of spermatogenesis, the 2009). Both (PWD 3 B6) F1 spermatocytes and Prdm9 number of meiotic spermatocytes in B6-ChrXMSM males knockout spermatocytes fail to form XY bodies and was reduced to less than one-seventh of those in control undergo apoptosis at pachytene stage (Mihola et al. males at 18 and 19 dpp. However, we previously 2009), similar to the phenotype we observed in the observed that the number of mature sperms in adult (B6 3 NJL/Ms) F1 hybrids. Epigenetic regulation B6-ChrXMSM males was approximately one-third of con- induced by the genetic incompatibility was shown to trol males (Oka et al. 2004). Thus, the premeiotic cell- cause the meiotic defects. It will be interesting to see cycle arrest in adult B6-ChrXMSM males may be recovered whether checkpoint responses are involved in these partially. meiotic defects. Three of the types of hybrid males, B6-ChrXMSM, The third stage at which spermatocytes of the hybrid (B6 3 NJL/Ms) F1, and (B6 3 M. spretus)F1, exhibited males are arrested is the first meiotic metaphase. arrest at the zygotene-to-pachytene stage of meiosis I. A Apoptosis at this stage occurs in spermatocytes in most prominent cell-cycle arrest at this stage occurs in PGN-ChrXMSM and, to a much greater extent, in in- the testes of intersubspecific (B6 3 NJL/Ms) F1 hybrids. terspecific (B6 3 M. spretus)F1 hybrids. During both They show extensive arrest and apoptosis at the late mitotic and meiotic metaphase, the faithful separation zygotene and early pachytene stages. Early studies of of chromosomes requires accurate connections be- budding yeast, Saccharomyces cerevisiae, revealed that tween chromosomes and microtubules, with kineto- persistent unrepaired DSBs or asynapsis triggers meiotic chores playing the most significant roles. Some of the arrest at the pachytene stage (Roeder and Bailis 2000). spindle checkpoint proteins, which monitor the attach- This response to the meiotic defects, the pachytene ment of microtubules to kinetochores in mitosis, are checkpoint, is known to function in many animals known to function in meiosis (Yin et al. 2008). Moreover, including mice. In various mouse meiotic mutants, cell-cycle arrest and apoptosis at metaphase of meiosis cell-cycle arrest and apoptosis at zygotene-to-pachytene I have been observed in mutant mice having meiotic are associated with unrepaired DSBs, asynapsis, and impairments such as lack of chiasmata and dissociation possibly lack of the XY body, suggesting that the of homologous autosomes and XY chromosomes pachytene checkpoint monitors these defects (Pittman (Baker et al. 1996; Lipkin et al. 2002; Mark et al. et al. 1998; S. S. de Vries et al. 1999; Yuan et al. 2000; 2008). Previous studies revealed that genetic divergence Turner et al. 2004; F. A. de Vries et al. 2005). Thus, it between the parental species in the pseudoautosomal is likely that zygotene and pachytene spermatocytes in region of the XY chromosomes causes dissociation of (B6 3 NJL/Ms) F1 males are subject to the pachytene these chromosomes in (B6 3 M. spretus)F1 spermato- checkpoint-dependent apoptosis, because the sperma- cytes (Matsuda et al. 1991, 1992; Hale et al. 1993). This tocytes showed severe asynapsis and unrepaired DSBs, dissociation is caused by asynapsis between sex chromo- which were represented by the extensive RAD51 foci somes during pachytene (although X and Y chromo- that persisted throughout the pachytene stage. On the somes are still close to each other) and continues other hand, pachytene spermatocytes in B6-ChrXMSM through metaphase I (Matsuda et al. 1991, 1992; Hale

Figure 6.—Meiotic phenotypes of intersubspecific and interspecific F1 hybrid males. (A) HE-stained testicular histology of (B6 3 NJL/Ms) F1 (top left) and (B6 3 M. spretus)F1 (top right) at 19 dpp. (B6 3 NJL/Ms) F1 testis contained tubules with apoptotic zygotene spermatocytes (ap) and tubules lacking spermatocytes (*). Increased apoptosis of metaphase I was observed at epithelial stage XII in (B6 3 M. spretus)F1 testis. TUNEL-stained testicular histology of (B6 3 NJL/Ms) F1 (bottom left) and (B6 3 M. spretus)F1 (bottom right) at 24 dpp. Scale bars: 50 mm. (B) Ratio of SYCP3-positive meiotic spermatocytes to all testicular cells for each male at 18–19 dpp. Reduction of SYCP3-positive spermatocytes in (B6 3 NJL/Ms) F1 testes was significant by t-test. (C) Immunostaining of control, (B6 3 M. spretus)F1 and (B6 3 NJL/Ms) F1 pachytene spermatocytes with anti-ATR, gH2AX, RAD51, and MLH1 antibodies (indicated in green). All spermatocytes were double stained with anti-SYCP3 antibody (red). Graphs at the side indicate the quantification of signal-positive spermatocytes at early and mid-to-late pachytene stages. * indicates significant difference by t-test (P , 0.05). Meiotic Disruptions in Hybrid Males 349

Figure 7.—Summary of meiotic arrest in four different cases of reproductive isolation. Horizon- tal arrows indicate the progression of spermatogenic development. Downward arrows (blue) indicate points of disruptions (arrest and/ or apoptosis). Orange horizontal lines indicate defective spermiogenesis. Thickness of blue and orange lines represents the proportion of spermatocytes subjected to arrests and/or apoptosis.

et al. 1993). In this study, we observed that (B6 3 M. the pachytene stage (Storchova´ 2004). Interspecific spretus)F1 pachytene spermatocytes successfully form incompatibility in the (B6 3 M. spretus)F1 hybrid causes XY bodies that appear to contain normal amounts of partial arrest and apoptosis at the pachytene stage, but ATR and gH2AX. This suggests that the dissociated XY extensive apoptosis was observed at metaphase I as well. chromosomes can be packaged into XY bodies and are In a substantially long period of time for speciation, it is protected from pachytene checkpoint monitoring. conceivable that mutations responsible for the repro- Thus, we infer that the XY chromosome dissociation ductive isolation have continuously accumulated in the finally triggers the spindle checkpoint, and the sperma- diverging populations, which may stochastically in- tocytes are eventually eliminated by apoptosis. crease the chance that mutations responsible for re- This study showed that male sterility in hybrid mice productive defects at early meiotic stages, i.e., the from various crosses is caused by disruptions at a premeiotic stage, occur. Thus, the absence of obvious minimum of three stages during meiosis. In addition correlation between the genetic distance and the stages to meiotic disruptions, consomic strains such as B6- of the meiotic defects, which was observed in this study, ChrXMSM and PGN-ChrXMSM show impairments during may imply that these house mice are in the early stage of spermiogenesis, terminal differentiation stage during speciation. spermatogenesis, resulting in the malformation of We thank M. A. Handel for kindly providing H1t antibody and H. J. sperm heads. These results indicate that genes respon- Garchon for providing XLR antibody. We thank J. Turner for sible for reproductive isolation do not function in a providing useful advice on experimental design and methodology particular specific process, rather in different multiple and D. G. de Rooij for helpful comments on histology. We thank processes. In many cases of mouse reproductive iso- members of the Shiroishi lab for providing critical comments on early versions of the manuscript. We are grateful to A. Yamakage, F. lation, meiotic defects are predominantly observed Kobayashi, H. Nakazawa, I. Okagaki, and the staff of the animal facility (Matsuda et al. 1991; Kaku et al. 1995; Forejt 1996). of National Institute of Genetics for assistance with mouse husbandry. This is possibly due to the high intricacy of meiosis, This study was supported in part by a Grant-in-Aid for Scientific which includes many cellular events such as the in- Research on Priority Areas (‘Genome Science’) from the Ministry of duction of DSBs, homologous recombination, repair of Education, Culture, Sports, Science and Technology of Japan. This study was also supported in part by the Transdisciplinary Research DSBs, and segregation of homologous chromosomes. Integration Center, Research Organization of Information and Notably, the genetic distance between the two paren- Systems. This article is contribution no. 2514 from the National tal strains was not explicitly correlated with the stages at Institute of Genetics, Mishima, Japan. which cell-cycle arrest and apoptosis occur. For instance, intersubspecific incompatibility in B6-ChrXMSM causes arrest both at the premeiotic stage and the pachytene LITERATURE CITED stage, but similar intersubspecific incompatibility in shley aeth reemers ack de MSM A , T., A. P. G ,L.B.C ,A.M.H and D. G. PGN-ChrX causes extensive apoptosis at metaphase Rooij, 2004 Correlation of meiotic events in testis sections I. In the case of another consomic stain B6-XPWD, which and microspreads of mouse spermatocytes relative to the mid- has an X chromosome of wild M. m. musculus-derived pachytene checkpoint. Chromosoma 113: 126–136. Baker, S. M., A. W. Plug,T.A.Prolla,C.E.Bronner,A.C.Harris PWD strain in the genetic background of B6 strain, it et al., 1996 Involvement of mouse Mlh1 in DNA mismatch re- was reported that meiotic disruption occurs mainly at pair and meiotic crossing over. Nat. Genet. 13: 336–342. 350 A. Oka et al.

Beumer, T. L., H. Kiyokawa,H.L.Roepers-Gajadien,L.A.van den Gregorova´, S., and J. Forejt, 2000 PWD/Ph and PWK/Ph inbred Bos,T.M.Lock et al., 1999 Regulatory role of p27kip1 in the mouse strains of Mus m. musculus subspecies: a valuable resource mouse and human testis. Endocrinology. 140: 1834–1840. of phenotypic variations and genomic polymorphisms. Folia Biol. Bonhomme, F., and J.-L. Gueńet, 1996 The laboratory mouse and 46: 31–41. its wild relatives, pp. 1577–1596 in Genetic Variants and Strains Haldane, J. B. S., 1922 Sex ratio and unisexual sterility in hybrid of the Laboratory Mouse, edited by M. F. Lyon. Oxford University animals. J. Genet. 12: 101–109. Press, Oxford. Hale, D. W., L. L. Washburn and E. M. Eicher, 1993 Meiotic ab- Borodin, P. M., M. B. Rogatcheva,A.I.Zhelezova and S. Oda, normalities in hybrid mice of the C57BL/6J 3 Mus spretus cross 1998 Chromosome pairing in inter-racial hybrids of the house suggest a cytogenetic basis for Haldane’s rule of hybrid sterility. musk shrew (Suncus murinus, Insectivora, Soricidae). Genome 41: Cytogenet. Cell Genet. 63: 221–234. 79–90. Hamer, G., K. Gell,A.Kouznetsova,I.Novak,R.Benavente et al., Borodin, P. M., S. C. Barreiros-Gomez,A.I.Zhelezova,C.R. 2006 Characterization of a novel meiosis-specific protein within Bonvicino and P. S. D’Andrea, 2006 Reproductive isolation the central element of the synaptonemal complex. J. Cell Sci. due to the genetic incompatibilities between Thrichomys pachyurus 119: 4025–4032. and two subspecies of Thrichomys apereoides (Rodentia, Echimyi- Hoyer-Fender, S., P. B. Singh and D. Motzkus, 2000 The murine dae). Genome 49: 159–167. heterochromatin protein M31 is associated with the chromocen- Boursot, P., W. Din,R.Anand,D.Darviche,B.Dod et al., ter in round spermatids and Is a component of mature sperma- 1996 Origin and radiation of the house mouse: mitochondrial tozoa. Exp. Cell Res. 254: 72–79. DNA phylogeny. J. Evol. Biol. 9: 391–415. Hunt, W. G., and R. K. Selander, 1973 Biochemical genetics of hy- Buaas, F. W., A. L. Kirsh,M.Sharma,D.J.McLean,J.L.Morris bridisation in European house mice. Heredity 31: 11–33. et al., 2004 Plzf is required in adult male germ cells for stem cell Inselman, A., S. Eaker and M. A. Handel, 2003 Temporal expres- self-renewal. Nat. Genet. 36: 647–652. sion of cell cycle-related proteins during spermatogenesis: estab- Calenda, A., B. Allenet,D.Escalier,J.Bach and H. Garchon, lishing a timeline for onset of the meiotic divisions. Cytogenet. 1994 The meiosis-specific Xmr gene product is homologous Genome Res. 103: 277–284. to the lymphocyte Xlr protein and is a component of the XY Joly, D., C. Bazin,L.W.Zeng and R. S. Singh, 1997 Genetic basis body. EMBO J. 13: 100–109. of sperm and testis length differences and epistatic effect on hy- Close, R. L., J. N. Bell,A.E.Dollin and H. R. Harding, brid inviability and sperm motility between Drosophila simulans 1996 Spermatogenesis and synaptonemal complexes of hybrid and D. sechellia. Heredity 78(4): 354–362. Petrogale (Marsupialia). J. Hered. 87: 96–107. Kaku, Y., Y. Kon,N.Takagi,T.Yamashita,M.Hayashi et al., Cobb, J., B. Cargile and M. A. Handel, 1999 Acquisition of com- 1995 Histological analysis of male hybrid sterility induced by petence to condense metaphase I chromosomes during sper- the Hst-1 gene in mice. J. Vet. Med. Sci. 57: 973–975. matogenesis. Dev. Biol. 205: 49–64. Keegan, K. S., D. A. Holtzman,A.W.Plug,E.R.Christenson,E.E. Cohen, P. E., S. E. Pollack and J. W. Pollard, 2006 Genetic anal- Brainerd et al., 1996 The Atr and Atm protein kinases associate ysis of chromosome pairing, recombination, and cell cycle con- with different sites along meiotically pairing chromosomes. trol during first meiotic prophase in mammals. Endocrine Rev. Genes Dev. 10: 2423–2437. 27: 398–426. Kulathinal, R. J., and R. S. Singh, 2008 The molecular basis of Costoya, J. A., R. M. Hobbs,M.Barna,G.Cattoretti,K.Manova speciation: from patterns to processes, rules to mechanisms. et al., 2004 Essential role of Plzf in maintenance of spermatogo- J. Genet. 87: 327–338. nial stem cells. Nat. Genet. 36: 653–659. Laurie, C. C., 1997 The weaker sex is heterogametic: 75 years of de Vries, S. S., E. B. Baart,M.Dekker,A.Siezen,D.G.de Rooij Haldane’s rule. Genetics 147: 937–951. et al., 1999 Mouse MutS-like protein Msh5 is required for Lipkin, S. M., P. B. Moens,V.Wang,M.Lenzi,D.Shanmugarajah proper chromosome synapsis in male and female meiosis. Genes et al., 2002 Meiotic arrest and aneuploidy in MLH3-deficient Dev. 13: 523–531. mice. Nat. Genet. 31: 385–390. de Vries, F. A., E. de Boer,M.van den Bosch,W.M.Baarends, Mahadevaiah, S. K., J. M. Turner,F.Baudat,E.P.Rogakou,P.de M. O oms et al., 2005 Mouse Sycp1 functions in synaptonemal Boer et al., 2001 Recombinational DNA double-strand breaks complex assembly, meiotic recombination, and XY body forma- in mice precede synapsis. Nat. Genet. 27: 271–276. tion. Genes Dev. 19: 1376–1389. Mark, M., H. Jacobs,M.Oulad-Abdelghani,C.Dennefeld,B.Fet´re Din, W., R. Anand,P.Boursot,D.Darviche,B.Dod et al., et al., 2008 STRA8-deficient spermatocytes initiate, but fail to 1996 Origin and radiation of the house mouse: clues from nu- complete, meiosis and undergo premature chromosome conden- clear genes. J. Evol. Biol. 9: 519–539. sation. J. Cell Sci. 121: 3233–3242. Dobzhansky, T., 1936 Studies on hybrid sterility. II. Localization of Matsuda, Y., T. Hirobe and V. M. Chapman, 1991 Genetic basis of sterility factors in Drosophila Pseudoobscura hybrids. Genetics 21: X-Y chromosome dissociation and male sterility in interspecific 113–135. hybrids. Proc. Natl. Acad. Sci. USA 88: 4850–4854. Elliott, R. W., D. R. Miller,R.S.Pearsall,C.Hohman,Y.Zhang Matsuda, Y., P. B. Moens and V. M. Chapman, 1992 Deficiency of X et al., 2001 Genetic analysis of testis weight and fertility in an and Y chromosomal pairing at meiotic prophase in spermato- interspecies hybrid congenic strain for chromosome X. Mamm. cytes of sterile interspecific hybrids between laboratory mice Genome 12: 45–51. (Mus domesticus) and Mus spretus. Chromosoma 101: 483–492. Elliott, R. W., D. Poslinski,D.Tabaczynski,C.Hohman and Mihola, O., Z. Trachtulec,C.Vlcek,J.C.Schimenti and J. Pazik, 2004 Loci affecting male fertility in hybrids between J. Forejt, 2009 A mouse speciation gene encodes a meiotic his- Mus macedonicus and C57BL/6. Mamm. Genome. 15: 704–710. tone H3 methyltransferase. Science 323: 373–375. Ellis, P. J., R. A. Furlong,A.Wilson,S.Morris,D.Carter et al., Moens, P. B., 1995 Histones H1 and H4 of surface-spread meiotic 2004 Modulation of the mouse testis transcriptome during chromosomes. Chromosoma 104: 169–174. postnatal development and in selected models of male infertility. Moens, P., M. Tarsounas,T.Morita,T.Habu,S.Rottinghaus Mol. Hum. Reprod. 10: 271–281. et al., 1999 The association of ATR protein with mouse meiotic Escalier, D., and H. J. Garchon, 2000 XMR is associated with the chromosome cores. Chromosoma 108: 95–102. asynapsed segments of sex chromosomes in the XY body of Moriwaki, K., 1994 Wild mice from the geneticist’s viewpoint, pp. mouse primary spermatocytes. Chromosoma 109: 259–265. 13–25 in Genetics in Wild Mice: Its Application to Biomedical Research, Forejt, J., 1996 Hybrid sterility in the mouse. Trends Genet. 12: edited by K. Moriwaki,T.Shiroishi and H. Yonekawa. Japan 412–417. Scientific Press and Karger, Tokyo. Forejt, J., and P. Ivańyi, 1974 Genetic studies on male sterility of Muller, H. J., 1942 Isolating mechanisms, evolution and tempera- hybrids between laboratory and wild mice (Mus musculus L.). ture. Biol. Symp. 6: 71–125. Genet. Res. 24: 189–206. Oka, A., A. Mita,N.Sakurai-Yamatani,H.Yamamoto,N.Takagi Good, J. M., M. A. Handel and M. W. Nachman, 2008 Asymmetry et al., 2004 Hybrid breakdown caused by substitution of the and polymorphism of hybrid male sterility during the early stages X chromosome between two mouse subspecies. Genetics 166: of speciation in house mice. Evolution 62: 50–65. 913–924. Meiotic Disruptions in Hybrid Males 351

Oka, A., T. Aoto,Y.Totsuka,R.Takahashi,M.Ueda et al., back Pungitius pungitius (Pisces, Gasterosteidae). Zool. Sci. 22: 2007 Disruption of genetic interaction between two autosomal 35–40. regions and the X chromosome causes reproductive isolation be- Tarsounas, M., and P. B. Moens, 2001 Checkpoint and DNA-repair tween mouse strains derived from different subspecies. Genetics proteins are associated with the cores of mammalian meiotic 175: 185–197. chromosomes. Curr. Top. Dev. Biol. 51: 109–134. Orr, H. A., 1997 Haldane’s rule. Ann. Rev. Ecol. Syst. 28: 195–218. Turner, J. M. A., O. Aprelikova,X.Xu,R.Wang,S.Kim et al., Orr, H. A., 2005 The genetic basis of reproductive isolation: in- 2004 BRCA1, histone H2AX phosphorylation, and male mei- sights from Drosophila. Proc. Natl. Acad. Sci. USA 102: 6522– otic sex chromosome inactivation. Curr. Biol. 14: 2135–2142. 6526. Wu, C. I., and A. W. Davis, 1993 Evolution of postmating reproduc- Payseur, B. A., J. G. Krenz andM.W.Nachman, 2004 Differential tive isolation: the composite nature of Haldane’s rule and its ge- patterns of introgression across the X chromosome in a hybrid netic bases. Am. Nat. 142: 187–212. zone between two species of house mice. Evolution 58: 2064–2078. Yang, F., K. Gell,G.W.van der Heijden,S.Eckardt,N.A.Leu Pittman, D., J. Cobb,K.Schimenti,L.Wilson,D.Cooper et al., et al., 2008 Meiotic failure in male mice lacking an X-linked fac- 1998 Meiotic prophase arrest with failure of chromosome syn- tor. Genes Dev. 22: 682–691. apsis in mice deﬁcient for Dmc1, a germline-speciﬁc RecA homo- Yin, S., X. F. Sun,H.Schatten and Q. Y. Sun, 2008 Molecular in- log. Mol. Cell. 1: 697–705. sights into mechanisms regulating faithful chromosome separa- Reynard, L. N., J. M. Turner,J.Cocquet,S.K.Mahadevaiah,A. tion in female meiosis. Cell Cycle. 7: 2997–3005. Toure´ et al., 2007 Expression analysis of the mouse multi-copy Yonekawa, H., K. Moriwaki,O.Gotoh,J.Watanabe,J.I.Hayashi X-linked gene Xlr-related, meiosis-regulated (Xmr), reveals that et al., 1980 Relationship between laboratory mice and the sub- Xmr encodes a spermatid-expressed cytoplasmic protein, SLX/ species Mus musculus domesticus based on restriction endonucle- XMR. Biol. Reprod. 77: 329–335. ase cleavage sites. Jap. J. Genet. 55: 289–296. Roeder, G. S., 1997 Meiotic chromosomes: it takes two to tango. Yonekawa, H., K. Moriwaki,O.Gotoh,N.Miyashita,Y.Matsushima Genes Dev. 11: 2600–2621. et al., 1988 Hybrid origin of Japanese mice ‘‘Mus musculus molossi- Roeder, G. S., and J. M. Bailis, 2000 The pachytene checkpoint. nus’’: evidence from restriction analysis of mitochondrial DNA. Trends Genet. 16: 395–403. Mol. Biol. Evol. 5: 63–78. Russell, L. D., R. A. Ettlin,A.P.S.Hikim,E.D.Clegg Yoshiki, A., K. Moriwaki,T.Sakakura and M. Kusakabe, 1990 Histological and Histopathological Evaluation of the testis. 1993 Histological studies on male sterility of hybrids between Cache River Press, Clearwater, FL. laboratory and wild mouse strains. Dev. Growth Differ. 35: Sakai, T., Y. Kikkawa,I.Miura,T.Inoue,K.Moriwaki et al., 271–281. 2005 Origins of mouse inbred strains deduced from whole- Yuan, L., J. G. Liu,J.Zhao,E.Brundell,B.Daneholt et al., genome scanning by polymorphic microsatellite loci. Mamm. 2000 The murine SCP3 gene is required for synaptonemal com- Genome 16: 11–19. plex assembly, chromosome synapsis, and male fertility. Mol. Storchova´, R., S. Gregorova,D.Buckiova,V.Kyselova,P.Divina Cell. 5: 73–83. et al., 2004 Genetic analysis of X-linked hybrid sterility in the house mouse. Mamm. Genome 15: 515–524. Takahashi, H., T. Nagai and A. Goto, 2005 Hybrid male sterility between the fresh- and brackish-water types of ninespine stickle- Communicating editor: F. F. Pardo Manuel de Villena GENETICS

Supporting Information http://www.genetics.org/cgi/content/full/genetics.110.118976/DC1

Reproductive Isolation in Hybrid Mice Due to Spermatogenesis Defects at Three Meiotic Stages

Ayako Oka, Akihiko Mita, Yuki Takada, Haruhiko Koseki and Toshihiko Shiroishi

FIGURE S1.—Schematic diagram depicting the positions of the proximal and distal markers for genotyping of

B6-ChrXMSM strain. The pseudoautosomal region starts in the middle of Mid1. Yellow and gray indicate the MSM-derived and B6-derived regions, respectively. In the B6-ChrXMSM strain, the distal region beyond the marker X164174029 is derived from B6 strain, and the small region including marker X163466600 contains B6 genome in some individuals. A. Oka et al. 3 SI

FIGURE S2.—Spermatogonia in B6-ChrXMSM testes at 14 dpp. Anti-PLZF and CBX1/HP1 antibodies detected undifferentiated type A and type B spermatogonia, respectively. The number of each type of spermatogonium did not significantly differ between control and B6-ChrXMSM testes at 14 dpp. In bottom panels, CBX1/HP1 signals are indicated in green and nuclei are counter-stained by Hoechst (gray). Scale bar: 200 m. 4 SI A. Oka et al.

FIGURE S3.—Distribution patterns of DSBs and recombination nodules in B6-ChrXMSM spermatocytes. (A)

Immunocytochemical detection of DSB marker H2AX, early recombination nodule marker RAD51, and late recombination nodule marker MLH1. Control and B6-ChrXMSM spermatocytes showed no obvious differences in the localization patterns of these marker proteins. (B) Frequency of RAD51- and MLH1-positive spermatocytes was compared at the two pachytene substages, which were determined by absence or presence of H1t signal. No significant difference was detected between control and the B6-ChrXMSM spermatocytes by t-test. A. Oka et al. 5 SI

FIGURE S4.—XY bodies in pachytene spermatocytes of B6-ChrXMSM. (A) Immunocytochemical detection of XY body proteins ATR, H2AX and XLR at the two pachytene substages. Early and late pachytene nuclei were distinguished by absence or presence of H1t signal. Despite the defective synapsis of the early pachytene spermatocytes of B6-ChrXMSM, the

XY body marker proteins colocalized with the sex chromosomes. (B) Frequencies of spermatocytes with XY bodies, as detected by the marker proteins, at each pachytene substage. No significant differences in the frequency of any marker were detected by t-test. More than 80 spermatocytes were counted for each male. 6 SI A. Oka et al.

FIGURE S5.—Apoptosis at the pachytene stage in B6-ChrXMSM testes. (A) TUNEL-labeled section of B6-ChrXMSM testis at 19 dpp. The severely disrupted seminiferous tubules lacking spermatocytes displayed no increase in apoptosis. (B)

HE-stained tubules at epithelial stage IV in B6-ChrXMSM testes at 15 dpp. Arrow heads: apoptotic spermatocytes, black arrows: late intermediate spermatogonia, white arrows: early type B spermatogonia. Scale bars: 100 m. A. Oka et al. 7 SI

FIGURE S6.—Sperm head morphology of PGN-ChrXPGN (control) and PGN-ChrXMSM males in N4 backcross generations. Note that apical hooks of the sperm heads of PGN-ChrXMSM male are shortened as compared to

PGN-ChrXPGN sperm heads. 8 SI A. Oka et al.

TABLE S1

Fertility of PGN-ChrXMSM strain in backcross generations

Genotype of the X chromosome Generation PGN MSM

F1a 5 / 5 (100.0%)b 3 / 3 (100.0%)

N3 5 / 5 (100.0%) 6 / 9 (66.7%)

N4 5 / 5 (100.0%) 3 / 7 (42.9%)

N5 5 / 5 (100.0%) 1 / 5 (20.0%)

N6 3 / 3 (100.0%) 4 / 4 (0%)

a F1 males with PGN-derived X chromosomes were obtained from crosses between PGN x MSM

and F1 males with MSM-derived X chromosomes were obtained from crosses between MSM x

PGN.

b Fertile males per total tested males. Percentage of fertile males is indicated in parentheses. A. Oka et al. 9 SI

TABLE S2

Testis weight of PGN-ChrXMSM males in backcross generations

Genotype of the X chromosome

Generation PGN MSM

n TWa RTW b n TW RTW

F1 7 58.8 2.83 3 90.5 4.20

N3 8 90.2 3.76 10 63.7 2.78

N4 5 100.6 4.14 7 44.3 1.76

N5 7 89.6 3.93 8 68.6 3.14

N6 4 94.0 3.98 2 66.0 2.94 a Single average testis weight in milligrams. b Relative testis weight in milligrams of single average testis per gram of body weight.