Reproductive isolation that initiates speciation is likely caused by incompatibility among multiple loci in organisms belonging to genetically diverging populations. Laboratory C57BL/6J mice, which predominantly originated from Mus musculus domesticus, and a MSM/Ms strain derived from Japanese wild mice (M. m. molossinus, genetically close to M. m. musculus) are reproductively isolated. Their F1 hybrids are fertile, but successive intercrosses result in sterility. A consomic strain, C57BL/6J-ChrXMSM, which carries the X chromosome of MSM/Ms in the C57BL/6J background, shows male sterility, suggesting a genetic incompatibility of the MSM/Ms X chromosome and other C57BL/6J chromosome(s). In this study, we conducted genomewide linkage analysis and subsequent QTL analysis using the sperm shape anomaly that is the major cause of the sterility of the C57BL/6J-ChrXMSM males. These analyses successfully detected significant QTL on chromosomes 1 and 11 that interact with the X chromosome. The introduction of MSM/Ms chromosomes 1 and 11 into the C57BL/6J-ChrXMSM background failed to restore the sperm-head shape, but did partially restore fertility. This result suggests that this genetic interaction may play a crucial role in the reproductive isolation between the two strains. A detailed analysis of the male sterility by intracytoplasmic sperm injection and zona-free in vitro fertilization demonstrated that the C57BL/6J-ChrXMSM spermatozoa have a defect in penetration through the zona pellucida of eggs.
SPECIATION is often initiated by reproductive isolation, which prevents gene flow across the two diverging populations and thereby accelerates genetic differentiation toward speciation. Therefore, understanding of the speciation process requires elucidation of the genetic mechanisms underlying reproductive isolation. Hybrid sterility, defined as sterility occurring in F1 hybrids, is one form of reproductive isolation (Haldane 1922). It is caused by genetic incompatibility among multiple loci. Alleles of the interacting loci are fixed differentially in two diverging populations, and an allele causes harmful effects when it is transferred to the other population (Dobzhansky 1937; Muller 1940; Orr 1996). Hybrid breakdown is another form of reproductive isolation. It is defined as inviability or sterility in the F2 or later generations of interspecific or intersubspecific crosses, while F1 hybrids are viable and fertile. Hybrid breakdown, like hybrid sterility, is caused by genetic incompatibility among multiple loci. Unlike hybrid sterility, hybrid breakdown behaves as a recessive trait (Muller 1940; Orr 1993). As F1 hybrids are fertile in the case of hybrid breakdown, the heterozygous state of alleles at interacting loci is not sufficient to cause inviability or sterility. These defects occur only when alleles of single loci or combinations of loci become homozygous in the F2 or later generations. Because its recessive phenotype hampers genetic analysis, hybrid breakdown has not been extensively studied.
Reproductive isolation has been studied in a number of animals, including Drosophila flies (Perez et al. 1993), Nasonia wasps (Gadau et al. 1999), Chorthippus parallelus grasshoppers (Shuker et al. 2005), and plants such as Mimulus monkey flowers (Christie and MacNair 1984) and Oryza sativa rice (Harushima et al. 2001). Hybrid sterility among Drosophila species has been investigated for >80 years. Recent studies were intended to map the full set of genes responsible for hybrid sterility (e.g., Harushima et al. 2001; Presgraves 2003). Despite the progress of DNA technologies and statistical tools, only a small number of hybrid sterility genes or hybrid breakdown genes have been characterized. The difficulty in identifying the responsible genes is attributable to the fact that the reproductive isolation arises from the cumulative effect of a number of genes with moderate-to-large effects (Wu et al. 1996; Hochstenbach and Hackstein 2000).
The house mouse, Mus musculus, is a complex species composed of several subspecies. It is thought that commonly used laboratory strains were derived predominantly from the M. m. domesticus subspecies, which is indigenous to Western Europe and the Mediterranean basin, with some small contributions from Asian subspecies, mostly M. m. musculus (Ferris 1982; Bishop et al. 1985; Bonhomme et al. 1987; Yonekawa et al. 1988; Moriwaki 1994; Sakai et al. 2005), whose range extends through northern Asia and eastern Europe. Hybrid sterility has been experimentally observed in intersubspecific crosses between M. m. domesticus and M. m. musculus (Forejt and Ivanyi 1975; Forejt et al. 1991; Trachtulec et al. 1997; Storchová et al. 2004; Britton-Davidian et al. 2005; Vyskočilová et al. 2005), as well as interspecific crosses between M. musculus and M. spretus (Guénet et al. 1990; Pilder et al. 1991, 1993, 1997; Pilder 1997; Elliott et al. 2001). In Europe, the two subspecies, M. m. domesticus and M. m. musculus, form a hybrid zone representing secondary contact of genetically divergent populations. Partial reproductive isolation between the two subspecies may reduce the introgression of marker genes across the hybrid zone (Tucker et al. 1992; Boursot et al. 1993; Dod et al. 1993; Sage et al. 1993; Payseur et al. 2004; Payseur and Hoekstra 2005; Raufaste et al. 2005).
The Japanese wild mouse M. m. molossinus is known to have emerged from the hybrid population of M. m. musculus and M. m. castaneus, a southeast Asian subspecies (Yonekawa et al. 1988; Moriwaki 1994; Sakai et al. 2005). In BAC clone-end sequence analysis of an inbred stain, MSM/Ms, which was derived from M. m. molossinus, ∼1% of nucleotides differed from those in the standard laboratory strain C57BL/6J (B6) (Abe et al. 2004). Moreover, most of the different nucleotides were shared by M. m. molossinus and M. m. musculus, indicating that the MSM/Ms strain is likely derived from M. m. molossinus and is genetically close to M. m. musculus. Currently, we are constructing a full set of consomic strains using MSM/Ms as the donor and B6 as the recipient strain. In each consomic strain, a whole chromosome is substituted by the MSM/Ms-derived counterpart in the genetic background of B6. One of these consomic strains, C57BL/6J-ChrXMSM, which carries an MSM/Ms-derived X chromosome in the B6 background, shows male sterility due to abnormal sperm morphology and low sperm motility (Takagi et al. 1994; Oka et al. 2004). This sterility most likely results from genetic incompatibility between MSM/Ms-derived X-linked genes and genes on the B6 autosome(s) and/or the Y chromosome. Because the F1 hybrids of the B6 and MSM/Ms strains are completely fertile, this sterility is a form of hybrid breakdown. In our previous study, we mapped three X-linked quantitative trait loci (QTL) responsible for the abnormal sperm-head morphology of C57BL/6J-ChrXMSM (Oka et al. 2004). They were designated Sha1 (Sperm head anomaly 1), Sha2, and Sha3 on the basis of their locations at the proximal, middle, and distal portions of the X chromosome.
In this study, we aimed to identify the other chromosomal regions responsible for the hybrid breakdown observed in the C57BL/6J-ChrXMSM males by genomewide linkage analysis and QTL analysis. When the sperm-head anomaly was used as a trait, we successfully detected three QTL on chromosomes 1 and 11, which were designated Ilx1 (Interacting locus with the X chromosome 1), Ilx2, and Ilx3. Then we introduced the candidate region of the MSM/Ms chromosomes into the C57BL/6J-ChrXMSM strain to confirm the genetic interaction between the Ilx genes and the X chromosome. The resulting mice failed to show significant restoration of sperm-head shape, but exhibited partially restored male fertility. Finally, to explore the biological functions of the Ilx and Sha genes in the male reproductive system, we characterized in detail the reproductive failure in C57BL/6J-ChrXMSM males.
MATERIALS AND METHODS
The C57BL/6J (B6) strain was purchased from the Jackson Laboratory (Bar Harbor, ME) and has been maintained at the animal facility of the National Institute of Genetics (NIG, Mishima, Japan). The MSM/Ms strain was established and maintained at NIG (Moriwaki 1994; Bonhomme and Guénet 1996). Consomic strains were constructed and are maintained at NIG. The construction of the consomic strain C57BL/6J-ChrXMSM was described in our previous report (Oka et al. 2004). For the maintenance of the C57BL/6J-ChrXMSM strain, females carrying a nonrecombinant MSM/Ms-derived X chromosome were selected in each backcross generation and mated with B6 males for the subsequent backcross generation.
Inheritance of B6- or MSM/Ms-derived alleles was determined by the following 144 MIT microsatellite markers located at ∼20-cM intervals on each chromosome: D1Mit316, -373, -213, -214, -46, -49, -338, -288, -16, -206, -459; D2Mit1, -362, -431, -8, -11, -15, -46, -22, -409, -266, -145; D3Mit60, -264, -305, -64, -170, -9, -11, -42, -14, -200, -19; D4Mit149, -104, -4, -242, -9, -40, -13, -51, -146; D5Mit1, -176, -5, -9, -241, -95, -29, -51, -102; D6Mit205, -268, -274, -96, -22, -104, -12, -259, -15; D7Mit21, -267, -228, -200, -148, -222, -7, -12, -362; D8Mit149, -64, -294, -7, -211, -13, -59, -128, -99, -194, -263, -346, -151; D10Mit279, -4, -186, -150, -103; D11Mit71, -294, -236, -155, -89, -70, -52, -302, -69; D12Mit59, -109, -4, -28, -18; D13Mit205, -55, -264, -16, -17, -138, -27, -36, -196, -35; D14Mit132, -11, -44, -203, -225, -267; D15Mit174, -224, -136, -270, -105, -73, -77, -40; D16Mit144, -4, -157, -105, -49; D17Mit29, -9, -53, -41, -123, -65; D18Mit15, -150, -142, -4; D19Mit32, -39, -10, -1, -33. PCR primers were made by RIKAKEN (Nagoya, Japan) or Invitrogen Japan K.K. (Tokyo), using information from the database of the Whitehead Institute/MIT Genome Database (http://www-genome.wi.mit.edu/). Information on length polymorphisms of the microsatellite markers was obtained from the Mouse Microsatellite Database of Japan (http://www.shigen.nig.ac.jp/mouse/mmdbj/top.jsp). Genomic DNA was extracted from ear and liver. The PCR-amplified DNA was separated by electrophoresis on agarose gels and stained with ethidium bromide.
Evaluation of sperm-head morphology for QTL analysis:
Epididymal spermatozoa were collected as previously described (Oka et al. 2004). The sperm suspension was washed three times and resuspended in 0.1 m ammonium acetate (pH 9.0) by centrifugation (×1500 g, 6 min). The final suspension was spread onto glass slides and air dried. The slides were stained with 0.22% Coomassie brilliant blue R250 (Fluka Chemie AG, Buchs, Switzerland) in 50% methanol/10% glacial acetic acid for 2 min (Jeffs et al. 2001). The severity of sperm-head anomaly was evaluated by a scoring method reported previously (Oka et al. 2004); in this study, we renamed it the sperm-head morphology index (SHMI). For quantitative evaluation of sperm-head morphology, we also employed a shape analysis software package, SHAPE (Iwata et al. 1998; Iwata and Ukai 2002; http://cse.naro.affrc.go.jp/iwatah/shape/), which uses a standardized elliptic Fourier descriptor (Kuhl and Giardina 1982; Furuta et al. 1995). The coefficients of the standardized elliptic Fourier descriptors were calculated from each of the 850 sperm-head samples, representing a total of 170 male progeny (five sperm heads for each male). Subsequently, these coefficients were applied to principal component analysis. We used the average of the first and second principal component (PC1 and PC2) scores for each male as a trait in the interval mapping.
Genomewide linkage analysis was performed to find associations between a genetic marker and a trait by Fisher's exact test. Because the aim of the genomewide linkage analysis was to screen suggestive chromosomal regions, we assumed a threshold level of P = 0.10. Single-marker analysis and interval mapping were performed to localize QTL governing sperm-head morphology using QTL Cartographer software (http://statgen.ncsu.edu/qtlcart/cartographer.html; Zeng 1993, 1994; Basten et al. 1997). The hypotheses for testing were as follows: H0, no QTL effect at any positions, namely, a = 0 (where a is the additive effect of a putative QTL); and H1, a QTL effect exists at the tested position, namely, a ≠ 0. The LOD score was defined as −log10(L0/L1), where L0 is the maximum likelihood under the null hypothesis H0, and L1 is the maximum likelihood under the alternative hypotheses H1. The threshold level was determined by WinQTLCart with a permutation test using 1000 permutations; the significance level was 0.05.
Intracytoplasmic sperm injection:
B6 females for intracytoplasmic sperm injection (ICSI) were purchased from Clea Japan (Tokyo). CD-1 females, used as foster mothers, were purchased from Charles River Japan (Yokohama, Japan). Eggs were obtained from 3- to-12-week-old mice after superovulation. At 16–20 hr after human chorionic gonadotropin (hCG) injection, eggs with cumulus mass were recovered from the ampullae with M2 medium (Quinn et al. 1982) containing 0.1% hyaluronidase (Sigma Chemical). The denuded eggs were rinsed thoroughly and maintained in M16 medium (Whittingham 1971) for up to 2 hr at 37° under 5% CO2. Epididymal spermatozoa were obtained from males, diluted in Toyoda–Yokoyama–Hoshi (TYH) medium (Toyoda et al. 1971), and placed on ice to maintain them at 17°–20° until use. ICSI was performed according to the procedures by Kimura and Yanagimachi (1995) with some modifications. Sperm were injected into eggs at room temperature (25°–27°) rather than at the lower temperatures specified previously. A small quantity of sperm suspension was removed and placed in a small drop of Hepes-buffered Chatot–Ziomek–Bavister medium containing 12% (w/v) polyvinyl pyrolidone (Wako Pure Chemical Industries, Osaka, Japan) and mixed well. A single spermatozoon was first drawn into the injection pipette from the tail and pulled back. This procedure was repeated until the head–midpiece junction (neck) was at the tip of the injection pipette; where this was achieved, the junction was then separated by repeated Piezo pulses (Kimura and Yanagimachi 1995). The sperm head was redrawn into the injection pipette and injected into an egg. The injected eggs were transferred into M2 medium and maintained at 17°–20° by cooling in a water bath for 15–40 min. The eggs were then placed into M16 medium and cultured at 37° under 5% CO2. Fertilized eggs were transferred into the oviducts of pseudopregnant CD-1 females on day 1 of pseudopregnancy.
Zona-free in vitro fertilization:
For analysis of zona-free in vitro fertilization (IVF), 2-month-old B6 females were superovulated by ip injection of 5 units of pregnant mare serum gonadotropin (Teikoku Hormone), followed 46–48 hr later by 5 units of hCG (Teikoku Hormone). Eggs containing a cumulus mass were recovered from the ampullae with TYH medium containing hyaluronidase solution at room temperature. The eggs were incubated for several minutes until the cumulus cells separated. Then the eggs were washed in fresh TYH medium three times for hyaluronidase treatment and transferred into acidic Tyrode solution (Sigma Chemical) at room temperature. As soon as the zona pellucida was dissolved (1–2 min), the eggs were transferred to fresh TYH medium under paraffin oil and incubated at 37° under 5% CO2 until insemination. Spermatozoa were collected from the cauda epididymides, dispersed into TYH medium under paraffin oil, and incubated for 60 min at 37° under 5% CO2 for capacitation. An adequate volume of sperm suspension was added into a drop of TYH-containing eggs. The appearance of pronuclei was assessed 4 hr after the insemination. On the next day, two-cell-stage embryos were transferred to a microdrop of Whittingham medium (Whittingham et al. 1972). Embryos were individually cultured for 2–3 days to avoid fusion and transferred into the uterus of pseudopregnant females.
To map autosomal genes responsible for male sterility in C57BL/6J-ChrXMSM, a genomewide linkage analysis was performed. First, females heterozygous for the X chromosome, C57BL/6J-ChrXMSM/XB6, were mated to MSM/Ms males, and the female progeny homozygous for the MSM/Ms-derived X chromosome (XMSM) were selected after genotyping for X-linked markers (Figure 1). Such homozygous females were subsequently crossed to C57BL/6J-ChrYMSM males that carry the MSM/Ms-derived Y chromosome in the B6 background.
From the above cross, we obtained 98 male progeny, all of which had the MSM/Ms-derived X and Y chromosomes and were heterozygous for various combinations of MSM/Ms-derived autosomal segments. They showed a distribution of sperm-head morphology, ranging from normal to abnormal. Our previous study indicated that male sterility in C57BL/6J-ChrXMSM is associated with abnormalities in sperm-head morphology (Oka et al. 2004). Therefore, we quantified in the male progeny the SHMI value, which correlates with the severity of the sperm-head anomaly. Because the hybrid breakdown of C57BL/6J-ChrXMSM males is a recessive trait, we expected that heterozygosity of chromosomal regions that interact with the X chromosome could restore fertility. Among the 98 progeny, we selected 13 males with normal sperm morphology (restored group) and 11 males that had the most severely affected sperm (nonrestored group). For these 24 males, we carried out genomewide genotyping at 144 microsatellite markers on 19 autosomes. Then we tested the statistical significance of different independent sites, comparing restored and nonrestored males for heterozygosity and homozygosity by Fisher's exact test. The tests revealed significantly low P-values at the markers listed in Table 1.
In the large regions with low P-values on chromosomes 1 and 11, M/B heterozygotes were observed more often than B/B homozygotes in the restored group; conversely, M/B heterozygotes were less frequent in the nonrestored group.
QTL analysis for the Y chromosome:
By performing the above cross and obtaining male progeny from C57BL/6J-ChrYMSM males, we were able to examine the possibility that the Y chromosome contributes to abnormal sperm morphology. We then generated male progeny carrying the B6-derived Y chromosome from a cross of the F1 female progeny and B6 males (Figure 1). We obtained 72 males with the B6-derived Y chromosome and 98 males with the MSM/Ms-derived Y chromosome. To test whether one or more Y-linked genes interact with the X chromosome, we performed a single-marker analysis for the SHMI trait with all 170 male progeny. This result indicated that the Y chromosome made little, if any, contribution to the sterility of the C57BL/6J-ChrXMSM males (LOD score of 0.37). Thus, combined data from all male progeny with the B6-derived and MSM/Ms-derived Y chromosomes were used for the subsequent QTL analysis of autosomes.
Single-marker analysis for significant markers:
To test the significance of the markers that showed low P-values in the genomewide linkage analysis, we performed single-marker analysis at the same markers and nearby markers on 88 randomly chosen progeny. The analysis showed a particularly high level of significance at chromosome 1- and chromosome 11-linked markers (Table 2). Among 88 mice, progeny homozygous for B6 alleles at either D1Mit16 or D11Mit294 showed ∼1.5-fold higher SHMI compared to heterozygous progeny. Progeny homozygous for B6 alleles at both D1Mit16 and D11Mit294 showed ∼2-fold higher SHMI than progeny heterozygous at both markers (mean SHMI ± SD: homozygous, 216.3 ± 74.8; heterozygous, 116.5 ± 73.2). Other significant markers were detected on chromosome 9.
Interval mapping on chromosomes 1 and 11:
To map precisely the responsible genes on chromosomes 1 and 11, we performed interval mapping with all 170 male progeny that were used in the analysis of the Y chromosome. For the interval mapping, we used not only SHMI, but also PC scores of contour shape of the sperm head as traits. In contrast to SHMI, which is based on observation, PC scores of contour shape are more objective. Thus, we thought that PC scores could verify the existence of QTL detected by SHMI on each chromosome. Two principal components extract different aspects of contour shape: PC1 corresponds to the degree of thickness and PC2 corresponds to the location of the hook at the apex of the sperm head (Figure 2). A wide and continuous spectrum of variation was recognized for all traits (Figure 3). Whereas PC scores showed normal distributions, SHMI values deviated from normal.
As summarized in Figure 4, results of the QTL analysis detected significant QTL on chromosomes 1 and 11. For the SHMI trait, the interval mapping program detected the highest LOD score (4.7) on chromosome 1 in the interval between D1Mit49 and D1Mit338, the region we named Ilx1. The second and third highest LOD scores (4.4 and 4.2) were detected on chromosome 11 in the interval between D11Mit52 and D11Mit302 (Ilx2) and in the interval between D11Mit294 and D11Mit236 (Ilx3), respectively. The additive effects of Ilx1, Ilx2, and Ilx3 are 62.6, 61.2, and 61.8, respectively. Using log-transformed SHMI, which gave a normal distribution of SHMI values, we obtained a similar result in the interval mapping (data not shown). When the PC1 score was used for the QTL analysis, the highest LOD score (5.8) was detected on chromosome 1 in the interval between D1Mit214 and D1Mit46. The analysis for PC2 detected a QTL on chromosome 11 in the interval between D11Mit236 and D11Mit155 (LOD score of 2.2).
To reevaluate the association of the genotypes of Ilx loci and sperm-head morphology, we compared the distribution of the PC1 score of the male progeny that have M/B heterozygous alleles at the markers D1Mit49, D11Mit302, and D11Mit294 with that of the complete set of progeny used for the QTL analysis. Figure 5 shows that the PC1 scores of males heterozygous for the three loci are significantly shifted to more positive values, representing a restoration of sperm-head morphology (mean PC1 ± SD: heterozygous males, 0.019 ± 0.024, n = 36; other progeny, including homozygous males, −0.004 ± 0.035, n = 121). Of these, some heterozygous males had exceptionally low PC1 scores. Significant differences were not detected between the heterozygous males and other progeny with respect to the PC2 score.
Partial restoration of fertility of C57BL/6J-ChrXMSM males by introduction of MSM/Ms chromosomes 1 and 11:
To confirm the genetic interaction between chromosomes 1 and 11 and the X chromosome, we introduced MSM/Ms-derived chromosome 1 (chromosome 1MSM) and chromosome 11 (chromosome 11MSM) into the C57BL/6J-ChrXMSM background. The mating scheme is illustrated in Figure 6. First, we generated F1 hybrids of C57BL/6J-Chr1MSM and C57BL/6J-Chr11MSM, which are heterozygous for chromosome 1MSM and chromosome 11MSM in the B6 genetic background; subsequently, the F1 hybrids were intercrossed. We genotyped relevant markers in the F2 males generated from this intercross to select males that were homozygous for MSM/Ms-derived alleles at least in the interval between D1Mit46 and D1Mit503 (162.3 Mbp from the centromere) on chromosome 1 while also harboring the nonrecombinant chromosome 11MSM homozygously. Then the selected males were crossed to heterozygous C57BL/6J-ChrXMSM/XB6 females. Progeny from this cross were always heterozygous for the MSM/Ms-derived relevant regions of chromosome 1 and chromosome 11. From these progeny, we obtained six males carrying a nonrecombinant XMSM chromosome. Hereafter, we refer to these males as C57BL/6J-Chr(1p,11,X)MSM (1p indicates partial chromosome 1).
To evaluate the fertility of the C57BL/6J-Chr(1p,11,X)MSM males, we caged each male with 1–3 B6 females for a maximum of 47 days, using a total of 12 females. We found that 5 of the 6 C57BL/6J-Chr(1p,11,X)MSM males were able to impregnate the females, resulting in a total of 10 pregnancies. Their average litter size (2.2, n = 9) was much smaller than the average litter size of B6 mice (7.9, n = 13). These pregnancies are significant, because no pregnancies were observed after 10 C57BL/6J-ChrXMSM males were caged with 2 B6 females each for >47 days. Although C57BL/6J-Chr(1p,11,X)MSM males showed partial restoration of the fertility, we did not observe significant restoration of sperm-head morphology (data not shown).
Inability of C57BL/6J-ChrXMSM sperm to fertilize eggs:
Our previous study showed that C57BL/6J-ChrXMSM males are sterile, owing either to failure of fertilization or to defects in early development of the fertilized eggs before the two-cell stage. To identify the stage at which sterility occurs, we assessed the development of eggs collected from superovulated B6 females on the day of copulation with C57BL/6J-ChrXMSM males. As a control, we used male C57BL/6J-ChrXB6 littermates from crosses used to maintain C57BL/6J-ChrXMSM; these C57BL/6J-ChrXB6 males carried a B6-derived X chromosome in the B6 background. As shown in Table 3, the copulation rate was normal in C57BL/6J-ChrXMSM males, but 98.7% of eggs did not form any pronuclei, and 3 of 238 eggs (1.3%) formed only maternal pronuclei. On the next day, only 2 eggs (0.8%) developed to the two-cell stage; these presumably resulted from parthenogenesis, because we could not detect any paternal pronuclei the day before. The results indicate that C57BL/6J-ChrXMSM males fail in fertilization in vivo.
Restoration of fertilization efficiency of C57BL/6J-ChrXMSM sperm by ICSI:
We examined whether the C57BL/6J-ChrXMSM sperm have the potential to fertilize eggs. To do this, we performed ICSI with these sperm and B6 eggs. The proportion of eggs surviving after ICSI under these conditions was similar to that observed with B6 and C57BL/6J-ChrXB6 sperm (Table 4). Proportions of fertilized eggs and live offspring were also not significantly lower in ICSI with C57BL/6J-ChrXMSM sperm compared to B6 sperm by Student's t-test (fertilized eggs, P = 0.11; live births, P = 0.14).
As observed in our previous study, C57BL/6J-ChrXMSM spermatozoa showed very low motility. Using medium containing 12% polyvinyl pyrrolidone, which is of high viscosity, we could divide C57BL/6J-ChrXMSM spermatozoa into two groups, one with relatively high motility and the other with severely low motility. To examine whether fertilization efficiency is correlated with spermatic motility, we compared two groups isolated among sperm obtained from single animals. We injected the spermatozoa of each group into the eggs by ICSI. The two groups showed no difference in the rate of surviving eggs or in the rate of fertilized eggs (Table 5). Although the rate of live offspring was somewhat higher in the relatively high-motility group compared to the low-motility group, the difference was not significant (Student's t-test; P = 0.09). Even after freezing and thawing procedures, C57BL/6J-ChrXMSM spermatozoa showed no remarkable reduction in fertilization efficiency (three tests with 199 eggs injected: 81.9 ± 1.9% of eggs surviving; 72.4 ± 11.5% of eggs fertilized).
Zona-free IVF with C57BL/6J-ChrXMSM sperm:
C57BL/6J-ChrXMSM spermatozoa swam toward eggs in vitro, indicating that substances from eggs chemotactically attracted the spermatozoa. To test whether the defect in fertilization by C57BL/6J-ChrXMSM spermatozoa lies in the ability to penetrate the zona pellucida or to fuse with the plasma membrane of eggs, we performed zona-free IVF with C57BL/6J-ChrXMSM spermatozoa. Because of their extremely low motility, C57BL/6J-ChrXMSM spermatozoa were inseminated into eggs in numbers two- to sixfold higher compared to C57BL/6J-ChrXB6 spermatozoa. Of 853 eggs, we obtained 216 fertilized eggs by zona-free IVF with C57BL/6J-ChrXMSM spermatozoa (Table 6), demonstrating that the C57BL/6J-ChrXMSM spermatozoa have the ability to fertilize the zona-free eggs. The proportion of the embryos that developed to morula and blastocyst was as high in this group as it was among those generated by zona-free IVF with C57BL/6J-ChrXB6 spermatozoa. After 2–3 days of culture, 163 morula/blastocyst embryos generated by zona-free IVF with C57BL/6J-ChrXMSM spermatozoa were transferred into the uteri of pseudopregnant females, and 23 pups were obtained.
Intercrossing between the B6 and MSM/Ms strains results in impaired reproduction in the F4–F5 generation; this phenomenon is an example of hybrid breakdown. When we constructed the C57BL/6J-ChrXMSM strain, we found that the males were sterile. Our previous study showed that genetic incompatibility between X-linked Sha genes and genes on some of the autosomes and/or on the Y chromosome causes male sterility in C57BL/6J-ChrXMSM mice (Oka et al. 2004). Male sterility is also observed in the consomic strain C57BL/6J-XPWD, in which the X chromosome of the B6 strain was substituted by the X chromosome of the PWD strain, which corresponds to M. m. musculus (Storchová et al. 2004). Another study of crosses between wild-derived outbred strains of M. m. musculus and M. m. domesticus implied the involvement of the X chromosome in male sterility (Britton-Davidian et al. 2005). These observations suggest that the X chromosome has a crucial role in reproductive isolation between M. m. musculus/M. m. molossinus and M. m. domesticus. A study from wild mouse populations in the European hybrid zone revealed that the genetic flow of X-linked loci across the hybrid zone is disproportionately reduced, indicating that reproductive isolation between M. m. musculus and M. m. domesticus is attributable to the X chromosome (Tucker et al. 1992; Dod et al. 1993; Payseur et al. 2004). Furthermore, a recent study reported that two X-linked diagnostic markers, Emd and Pola1, exhibit reduced introgression across the European hybrid zone (Payseur et al. 2004). Notably, these two markers are close to Sha2, which has the highest LOD score on the X chromosome.
Although the X chromosome clearly plays an important role in the divergence between M. m. musculus/M. m. molossinus and M. m. domesticus, reproductive isolation can occur without the participation of the X chromosome. Hst1 is the causative gene for hybrid sterility between C57BL/10 mice and wild M. m. musculus and has been mapped to chromosome 17 (Forejt and Ivanyi 1975; Forejt et al. 1991). Another study showed that hybrid sterility between the C57BL/10 and wild M. m. musculus obtained from a different location is attributable to the Hst1 gene and additional genes on autosomes (Vyskočilová et al. 2005). The female reproductive impairment observed in F2 progeny of the B6 and M. m. molossinus-derived strain, MOM, is likely to be caused by autosomal genes (Niwa-Kawakita 1994). If two segregating populations evolve reproductive isolation due to a single incompatibility between two genes, these species will not stop diverging at additional loci that can also cause hybrid sterility or hybrid breakdown (Orr and Coyne 2004). Thus, we can now identify a number of sites of genetic incompatibility between species or subspecies. Although we cannot determine whether the incompatibilities that we found were involved in the initial genetic divergence, we believe that it had an important role in restricting gene flow across the subspecies.
Recent genomic analyses showed that inbred laboratory mouse strains have a mosaic genomic structure, with the vast majority of segments derived from M. m. domesticus and a minority from the Asian subspecies M. m. musculus, M. m. molossinus, and M. m. castaneus (Wade et al. 2002; Abe et al. 2004; Sakai et al. 2005). On the basis of a comparative analysis of the MSM/Ms-derived BAC-end sequence covering 2.29% of the total genome of the B6 genome, we found that single nucleotide polymorphisms (SNPs) were unevenly distributed. We proposed that regions of the B6 genome with high SNP rates are derived from M. m. domesticus and those with low SNP rates, which occupy 5.2% of the genome, are related to M. m. musculus/M. m. molossinus (Abe et al. 2004). If the results from laboratory crosses between the subspecies reflect the evolutionary significance during speciation, the regions corresponding to the QTL should have M. m. domesticus alleles in the B6 genome. Interestingly, low-SNP-rate regions are exceptionally rare on the X chromosome, demonstrating a less significant contribution of M. m. musculus/M. m. molossinus to the X chromosomes of the B6 strain (Wiltshire et al. 2003; Abe et al. 2004; Sakai et al. 2005). A recent study tried to detect the genes responsible for the reproductive isolation between M. m. musculus and M. m. domesticus by a comparative analysis with SNP data from a database (Harr 2006a,b). Harr (2006a,b) made clear that highly differentiated regions in the genome should have failed to introgress across the subspecies historically and would therefore contain potential genes leading to hybrid unfitness. We are now conducting shotgun sequencing of the entire MSM/Ms genome. The comparative analysis using the B6 and MSM/Ms sequence data will detect chromosomal regions with a high density of SNPs where the responsible genes should exist.
Although several genes expressed during spermatogenesis exist on the Y chromosome, our single-marker analysis showed that Y-linked genes, except in the pseudoautosomal region, are irrelevant to the hybrid breakdown seen in C57BL/6J-ChrXMSM males. This is consistent with the fact that the Y chromosome of laboratory mouse strains, including B6, originated from the Japanese wild mouse M. m. molossinus (Bishop et al. 1985; Nagamine et al. 1992; Tucker et al. 1992).
In this study, we employed two parameters to evaluate anomalies of sperm-head morphology. SHMI is an observational index based upon morphological features, including the length of the apical hook, the roundness of the posterior region, and the overall size (Oka et al. 2004). On the other hand, PC scores are more objective parameters. Captured images of sperm heads are subjected to Fourier transformation and principal component analysis, resulting in determination of scores that describe each sample. However, because PC scores do not provide descriptive information and do not directly assess size, we used PC scores only for verification of QTL on chromosomes 1 and 11. QTL analysis with PC scores successfully detected QTL on chromosomes 1 and 11, which confirmed that these QTL underlie differences in sperm-head morphology.
The PC score distribution of males heterozygous for M/B at the Ilx loci demonstrated the contribution of the Ilx loci to the sperm-head morphology. However, C57BL/6J-Chr(1p,11,X)MSM males did not show significant restoration of sperm-head morphology, suggesting that additional genes on other autosomes are required for normal morphogenesis of sperm. We observed that some of the male progeny used for the genomewide linkage analysis and the QTL analysis had normal sperm-head morphology. Such restored males should have had a set of MSM/Ms alleles at additional loci other than the three Ilxe loci, which properly interact with MSM/Ms alleles at the Sha loci and are necessary for complete restoration of normal morphogenesis. In addition to chromosomes 1 and 11, single-marker analysis revealed suggestive loci on chromosome 9. Thus, chromosome 9 is likely to interact with the X chromosome in determining fertility.
Although C57BL/6J-Chr(1p,11,X)MSM males failed to exhibit normal sperm-head morphology, they showed partial restoration of fertility. To our knowledge, this is the first experimental demonstration of genetic incompatibility causing male sterility in mammals. It also confirmed the validity of SHMI as an indicator of male fertility. Partial restoration of male fertility suggests that additional QTL on other chromosomes are relevant in normal fertility.
To understand the biological roles of the genetic interaction of Sha and Ilx loci in the male reproductive system, we investigated fertilization with C57BL/6J-ChrXMSM sperm in detail. Eggs collected from females that copulated with C57BL/6J-ChrXMSM males did not form paternal pronuclei and never developed to the two-cell stage. This demonstrates that C57BL/6J-ChrXMSM spermatozoa cannot fertilize eggs in vivo. The ICSI experiment in this study confirmed that C57BL/6J-ChrXMSM spermatozoa were able to produce normal offspring after injection into eggs. Spermatic motility did not correlate with the rate of egg survival after the injection. Moreover, the rates of egg survival and live births were not affected by the severity of the sperm-head anomaly. All of these results indicate that all C57BL/6J-ChrXMSM sperm possess normal genomic and epigenetic factors, which are necessary for embryonic development. The ability of C57BL/6J-ChrXMSM spermatozoa to fertilize eggs when the zona pellucida was removed suggests that their fertilization impairment stems from defects in their ability to penetrate the zona pellucida.
Sperm penetration through the zona pellucida is assumed to occur by the mechanical force of sperm motility, the action of lytic enzymes on the spermatozoa, and the interaction of proteins on the spermatic surface and glycoproteins of the zona pellucida. Although several molecules necessary for sperm–egg recognition have been identified, the molecular mechanism that ensures zona penetration is not well understood. Study of the Sha and Ilx genes may provide new insight into the molecular mechanisms underlying this process. In animals, fertilization in interspecific crosses is strictly restricted by several steps before fertilization. Many studies have reported that gamete-recognition proteins on sperm and egg evolve rapidly, creating barriers to cross-species fertilization (Vacquier 1998; Swanson and Aquadro 2002; Swanson and Vacquier 2002; Torgerson et al. 2002; Good and Nachman 2005). Future study is required to determine whether the Sha and Ilx genes, which underlie hybrid breakdown, do so by regulating the ability of sperm to penetrate the zona pellucida.
We thank J-L. Guénet and H. Yonekawa for useful discussion and comments on this work. This study was supported in part by a Grant-in-Aid for Scientific Research on Priority Areas (“Genome Science”) from the Ministry of Education, Culture, Sports, Science and Technology of Japan. This study was also supported in part by the Transdisciplinary Research Integration Center, Research Organization of Information and Systems. This article is contribution no. 2505 from the National Institute of Genetics, Mishima, Japan.
Communicating editor: C. A. Kozak
- Received July 5, 2006.
- Accepted October 11, 2006.
- Copyright © 2007 by the Genetics Society of America