| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Gamete Biology |
The Jackson Laboratory, Bar Harbor, Maine 04609
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
gametogenesis, meiosis, oocyte development, spermatid, spermatogenesis
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
In male mice, spermatogenesis begins 4 days after birth with the differentiation of germ cells into the type A spermatogonial stem cells capable of continual mitotic self-renewal [6]. Further differentiation and subsequent mitotic divisions produce type A and type B spermatogonia that proceed to the first meiotic prophase, at which point they are called spermatocytes. During meiosis, DNA is replicated, homologous chromosomes align, recombination and synapsis occur, and the reductional and equational divisions follow. Subsequent to meiosis, haploid spermatids mature to spermatozoa during the process of spermiogenesis. Finally, mature spermatozoa are released into the lumen of the seminiferous tubule, a process called spermiation, and proceed to the epididymis, where motility and fertilization potential develop.
Oogenesis in mice is initiated prenatally after the migration of PGCs to the genital ridge, where they associate with pregranulosa cells [7]. Meiotic division in female germ cells begins around Embryonic Day 13, signaling the transition from oogonia to oocytes [3]. The oocytes then enter a prolonged arrest in diplonema of the first meiotic prophase, known as the dictyate stage. Growth of the oocyte during the dictyate stage is coordinated with the development and growth of the ovarian follicle. At this time, the nucleus or germinal vesicle (GV) and the nuclear membrane are intact. The dictyate stage is maintained until a preovulatory surge of luteinizing hormone causes GV breakdown (GVBD) and resumption of meiosis. A polar body is extruded during the first meiotic division. After meiosis I, the haploid oocyte then assembles a second meiotic spindle and arrests at metaphase II, awaiting fertilization.
Historically, the genes controlling mammalian gametogenesis have been difficult to elucidate, and researchers have relied on spontaneous mutations or strategies of targeted mutagenesis. The nature of the phenotypic consequence of mutations affecting gametogenesis—sterility or subfertility—is difficult to analyze genetically in humans (with the exception of genes on the Y chromosome [8]). Even though mice are experimentally tractable, spontaneous infertility mutations go unnoticed during routine colony maintenance, because nonproductive matings are common. Hence, the collection of spontaneous mutations causing sterility in mice has been limited to those that also confer a visible phenotype, such as skeletal fusions with sterility (sks), dominant white spotting (Kitw), and t haplotypes [9–11]. Fortuitous transgene insertions disrupting genes required for fertility have been another limited, but informative, source of infertility mutations [12–19]. To date, approximately 250 genes are known to be involved (in males and/or females) in the successful production of offspring, a good proportion of which function during gametogenesis [20].
In contrast to the mammalian system, genetic studies in other eukaryotic model organisms have led to the identification of many genes involved in various aspects of gamete production. For example, the genetic and experimental simplicity of fungi such as Saccharomyces cerevisiae has been exploited to conduct screens for mutations affecting all key steps of meiosis: recombination, DNA metabolism, synaptonemal complex formation, double-strand break (DSB) formation/repair, and cross-over interference and cohesion. Subsequent gene targeting in mouse embryonic stem (ES) cells has enabled investigators to elucidate the roles of the mammalian homologues of these genes. These include members of the RAD52 epistasis group (Rad51 [21, 22], Rad52 [23], Rad54 [24], and Dmc1 [25, 26]) and genes encoding structural proteins and repair enzymes expressed during meiosis, such as Mlh1 [27], Mlh3 [28], Msh4 [29], Msh5 [30], Sycp3 [31], and Hspa2 [32].
Exploitation of cross-species homology has been an invaluable strategy, but mouse homologues of many of the key yeast meiosis-specific genes that are involved in other aspects of meiotic control and chromosome metabolism have not been found, despite the availability of whole genomic sequences and extensive expressed sequence tag sets. Examples of such genes include NDT80, which activates a meiotic transcriptional program [33, 34]; REC102, which participates in the meiotic DSB formation that is required for recombination [35, 36]; and SAE2, which is involved in strand resection at the sites of DSBs [37]. It is possible that orthologues of such genes exist, but the degree of divergence is so high that definitive relationships cannot be drawn. This may also be the case for SYCP1 and SYCP2, which have been suggested to be potential orthologues of the yeast Red1 and Zip1 synaptonemal complex proteins, respectively [38, 39]. In this case, the main similarities lie in the overall protein structures rather than at the sequence level [40]. Therefore, other means will be required to identify the mouse functional orthologues of many yeast meiosis genes (if, indeed, they exist).
Identification of mammalian genes that encode molecules important for gametogenic processes that do not occur in yeast, such as stem cell development, spermiogenesis, oogenesis, and somatic cell/germ cell interactions, might be identified through comparisons with higher organisms, such as flies and worms. However, these organisms differ from mammals so profoundly in certain aspects of germ cell development that clear orthologues of many key genes may not be shared [41].
Because of the complex interaction of soma and germ tissues, the number of genes responsible for gametogenesis in mammals likely is much greater than the number in lower organisms. Currently, few accurate estimates are available regarding the total number of genes necessary for the completion of gametogenesis in mammals. A high number of sexually dimorphic transcripts also likely exist. Serendipity and targeted deletion have generated a starter pool of gametogenic mutants (
250), but these will not be sufficient to disclose fully the genetic control of the process. Clearly, other strategies will be needed to identify all the genes required for gametogenesis in mammals.
The development of methods for random mutagenesis of the mouse genome makes it possible to consider whole-genome, forward genetic screens for mutations that affect gametogenesis. In mice, the chemical ethylnitrosourea (ENU) is reported to be the most potent and experimentally efficient germ line mutagen [42], and it has been exploited to conduct genome-wide screens for mutations affecting various phenotypes, such as behavior, embryonic development, general anatomy, and blood chemistry [43–48]. However, to date, no genome-wide ENU screens have been reported that are designed specifically to recover infertility mutations in mice.
An alternative approach for generating mice with random mutations of their genomes is via chemical mutagenesis of ES cells [49]. Here, we report the isolation of a collection of mouse infertility mutations generated by both techniques. The mutations affect a broad range of gametogenic stages, and several of the mutations have been mapped genetically. These data permit estimates regarding the number of additional infertility mutations that one can expect to uncover by further screening and of the total number of gametogenesis-specific genes in mice. Based on these data, we also discuss optimal strategies for mutation screens.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Male C57BL/6J (B6) mice (age, 8–10 wk) obtained from The Jackson Laboratory were given three weekly intraperitoneal injections of 80 or 85 mg/kg body weight of ENU. The preparation and quantitation of ENU was done as described previously [44]. After the final treatment, 4 wk were allowed to pass before the males were mated to C3H/HeBFeJ-Rw/+ mice. Mice that were sterile for at least 10 wk and then successfully regained fertility were used to produce G1 progeny. In the present study, only mice not inheriting the rump-white (Rw) inversion were used (mice inheriting Rw were employed in a screen to detect mutations specifically on chromosome 5 [50]). The G1 males sired by the ENU-treated male mice are heterozygous for a wild-type genome. For all practical purposes, each G1 male is unique, carrying an entirely different set of mutations, even if derived from the same treated animal. This is a result of random (and, presumably, unique) mutagenesis of each type A spermatogonial cell in the injected male. The G1 males were mated to C3H females to produce G2 females. The G2 females were then backcrossed to the G1 males to produce G3 animals (Fig. 1A).
|
ES Cell Mutagenesis
Three stocks of CJ7 (129S1/Sv-p+Tyr+Kitl+ or 129) ES cells were ethyl methansulfonate (EMS)-mutagenized previously in this lab with treatments of 300, 400, and 500 µg/ml of EMS (Sigma, St. Louis, MO) for 16 h [49]. Chimeras were generated by injection of 10–15 nonclonal ES cells, from the batch cultures, into B6 blastocysts. It is therefore possible that the ES cell-derived progeny can carry unrelated sets of mutations as a consequence of being products of different ES cells. However, we previously presented evidence that the germ line of most chimeras derived from heavily mutagenized ES cells arose from a single ES cell [49]. Chimeras were crossed to B6-producing agouti male offspring (G1). In this scenario, G1 progeny of each chimera are likely to contain overlapping subsets of mutations. Therefore, multiple G1 animals were obtained from each chimera to capture mutations existing in both sets of alleles (Fig. 1B). Each was then used to produce several G2 female progeny that were backcrossed to their sires, as indicated in Figure 1B, to produce G3 animals.
G3 Screening
The G3 mice produced by both strategies were screened for potential gametogenetic defects in two ways (Table 1). The majority of the screens concentrated on the identification of defects in males. In all ES cell experiments and in two of the ENU screens, potential male infertility mutants were detected by examining ductus deferens sperm for defects in general morphology, motility, and quantity. In ENU experiment 3, males were fertility tested by mating. Matings in experiments 1 and 3 also assessed female fertility. To assess sperm presence, general morphology, and general motility, 8-wk-old male G3 animals were killed. The testes, epididymis, and ductus deferens were removed. The epididymis and ductus deferens were placed in PBS (pH 7.4) at 37°C and 5% CO2. The tissue was macerated with scissors to release sperm and allowed to sit for 10 min. After incubation, 10 µl of the solution were analyzed on a hemocytometer. Samples were scored for sperm presence, movement, and general sperm appearance.
|
Oocyte Culture and Assessment of Meiotic Progression
Ovaries were removed from females 44 h after treatment of the mice with eCG and manually disrupted to release oocytes. The released oocytes were cultured as described previously to test for spontaneous resumption of meiosis [51, 52]. The oocytes were examined using a stereomicroscope after 15 h of culture and assessed for GVBD, indicative of the resumption of meiosis, and the presence of a polar body, which is usually characteristic of progression to metaphase II.
Histology
Testes were fixed in Bouin solution for more than 24 h before being embedded in paraffin. Sections (thickness, 5 µm) were cut and stained with hematoxylin and eosin. Digital images were acquired with a charge-coupled device camera.
Genetic Mapping
In the case of mutations derived from ENU treatment of B6 males, a genome-wide set of microsatellite markers, polymorphic between B6 and C3H and representing every autosome (at least three markers per chromosome), was used to assay DNA from affected G3 animals by polymerase chain reaction (PCR). In the case of mutations derived from EMS treatment of 129 ES cells, a similar genome-wide set of microsatellite markers, polymorphic between B6 and 129, was used to assay DNA from affected G3 animals by PCR. Linkage was implicated by the association of phenotype with homozygosity for B6 (ENU) or 129 (EMS) alleles, respectively. For some cases, in which a higher degree of polymorphism was necessary, heterozygotes were bred to Mus castaneus mice, and F1 animals were intercrossed to generate F2 progeny for higher-resolution mapping.
Genotyping
Tail tips (2 mm) were lysed by incubation overnight at 55°C in 200 µl of PBND buffer (50 mM KCl, 10 mM Tris [pH 8.3], 2.5 mM MgCl2, 0.1 mg/ml of gelatin, 0.45% [v/v] NP40, and 0.45% Tween 20) and 2 µl of proteinase K (10 mg/ml). The lysate (1.5 µl) was directly used as template in subsequent standard PCR reactions. The PCR products were separated by agarose gel electrophoresis on a horizontal, 4.0% MetaPhor gel (Cambrex, Rockland, ME).
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Two strategies were taken to induce mutations randomly throughout the mouse genome. One involved traditional ENU mutagenesis, in which male C57BL6/J (B6) mice were treated with ENU and used to initiate a classical, three-generation breeding scheme to obtain third-generation (G3) offspring that were potentially homozygous for induced mutations (Fig. 1). The other strategy entailed the production of mice derived from ES cells treated with the mutagen EMS. Chimeras produced from these cells were used to initiate a three-generation mating scheme, as shown in Figure 1, right panel.
In the ENU experiments, a total of 2033 G3 animals were screened, representing 102 mutagenized gametes (Table 1). Eight heritable phenodeviants were isolated from the ENU screens. Five of these (swm2, swm6, mei4, 9a, and 12d) have been mapped genetically. Two (mei7 and 11b) have yet to be mapped, and one (swm4) was lost. In comparison, EMS mutagenesis yielded five heritable mutations identified among 540 G3 males produced by 10 chimeras: mei1, mei2.5, mei2.7, Sgdp, and gcd2 (germ cell deficient) (Table 1). Each of the three new mutations, along with mei1 and Sgdp, which were recovered in previous ES cell screens [53] (experiments 4 and 5), were mapped. A summary of the mutation phenotypes and their map positions is presented in Table 2. Histological sections from each mutant (except for swm4, which will not be discussed further) are presented in Figures 2–5. Mutations with seemingly similar phenotypes are grouped as outlined below.
|
|
|
Preliminary Phenotype Characterization
Group 1 includes gcd2 and Sgdp. Both cause a phenotype of severe germ cell depletion, indicating a possible defect early in germ line differentiation. The gcd2 was initially identified as an autosomal recessive mutation that caused male sterility (Fig. 2, C and D). Subsequently, we found that gcd2 causes germ cell depletion and sterility in both sexes. Histological analysis of mutant testes showed that spermatogenesis occurs in some seminiferous tubules but not others. Preliminary results suggest that the gcd2 phenotype is approximately 65% penetrant, possibly affected by strain background modifiers segregating in the mapping crosses. The gcd2 has been mapped to chromosome 2.
The spermatogonial depletion (Sgdp) mutation, derived from ES cell mutagenesis [49], causes a subset of seminiferous tubule segments to become agametic, similar to Sertoli Cell-only (SCO) syndrome in humans (Fig. 2, E and F). Heterozygous males are sterile and do not transmit the mutation. Female heterozygotes may be sterile, semifertile (having small litters of approximately three or four offspring), or yield normal-size litters, depending on genetic background. Because the mutation was transmitted from a male chimera derived from EMS-mutagenized ES cells to a daughter and propagated through her, we speculate that the defect is not intrinsic to germ cells. Rather, Sgdp may function primarily in somatic cell types, such as Sertoli cells and granulosa cells, so that the mutation was "rescued" from the chimera by wild-type function provided via host blastocyst-derived cells. Thus, this mutation may affect germ cell-somatic cell interactions. It has been mapped to chromosome 15.
Group 2 includes 11b, mei7, and mei4. Mutations in this group cause phenotypic consequences during the stages of meiosis. The mutation 11b causes sterility only in females. Although the oocytes have normal morphology, they fail to undergo GVBD (data not shown).
The mei7 causes sterility only in males. Spermatogenesis is arrested early in meiotic prophase I (Fig. 3, A and B). By early to midprophase I, cell death is observed. In some cases, early prophase I nuclei appear large and swollen. This phenotype is similar to that described for the autosomal recessive mutation microrchidia [16].
|
The mei4 was derived from the ENU mutagenesis of a B6 male. Males and females homozygous for the mei4 mutation are sterile. Affected males have extremely small testes and seminiferous tubules that lack postmeiotic cell differentiation (Fig. 3, C and D). Female homozygotes have reduced numbers of oocytes and ovarian dysmorphology (data not shown). Preliminary evidence suggests that mei4 animals are unable to complete meiosis, arresting at the first meiotic division. This mutation maps to the central portion of chromosome 14.
Group 3 includes mei2.5, mei2.7, 12d, and swm6. These four mutations are grouped because the germ cells of affected males apparently develop beyond meiosis but do not produce normal, motile sperm. The mei2 mutations were derived from mutagenized ES cells. Homozygous males are sterile, and histological analysis of testes revealed that the majority of spermatocytes undergo meiotic arrest around the pachytene stage (Fig. 4, A–D). However, seminiferous tubules sometimes contain postmeiotic spermatids, indicating that the mutation does not cause an absolute block to meiotic progression. The two mutations (mei2.5 and mei2.7) were originally thought to have been a single mutation, because they arose in the same family and presented similar phenotypes. However, they were later found to be separate mutations mapping to the proximal portions of chromosomes 5 (mei2.5) and 7 (mei2.7). Testicular size in both mutants varies depending on the genetic background.
|
The mutations 12d and swm6 have similar phenotypes. The swm series of mutations (sperm without motility) cause spermiogenetic defects. The mutants show either an absence or greatly diminished numbers of mature sperm (Fig. 4, E–H). Spermatogenesis in affected males progresses through meiosis, resulting in the production of round spermatids. The acrosome and sperm head appear developed; however, very few intact epididymal spermatozoa are seen. These mutations have been mapped to chromosomes 16 (12d) and 5 (swm6).
Group 4 includes swm2 and 9a. Histological examination of seminiferous tubules from mutant animals of this group reveals development, albeit abnormal, of sperm heads and flagella, thus distinguishing them from the previous class of mutants (Fig. 5). In swm2 animals, most sperm lack flagella, although a few possess flagella and are capable of some movement. Nearly aspermic, mutant tubules contain a disorganized mass of cell bodies near the lumen (Fig. 5, A and B). The swm2 has been mapped to chromosome 7. Mutants from the 9a family lack epididymal sperm. Seminiferous tubules of 9a homozygotes contain intact spermatozoa (Fig. 5, C and D), but only round cells are found in the epididymis (data not shown). The 9a family shows considerable phenotypic variation, depending on the age and genetic background of the animals. This mutation has been mapped to mouse chromosome 2.
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Substantial Numbers of Mutations Will Be Recovered
Regarding screens of ENU-mutagenized mice in which animals were tested on the basis of obvious defects in sperm flushed from the ductus deferens, mutations were identified in 4 of 82 families screened (experiments 1 and 2 in Table 1). In experiment 3, which tested G3 animals of both sexes for fertility, each of the recovered mutations would have been identified by sperm screening with the exception of 11b, which is female-specific. Considering that fertility testing requires more space and time compared to postmortem examination of gametes (each animal must be test-mated in an individual box vs. group housing in the case of gamete inspection) and that only one mutation of 60 families screened (experiments 1 and 3) was identifiable only by fertility testing, the trade-off between ease and speed of screening versus the goals of the experiment should be considered. For example, if the goal is to identify genes that are required for both male and female gamete production, it would be more expeditious to screen sperm of G3 males exclusively. If the goal is to identify oogenesis-specific genes or infertility mutations that cause subtle morphological defects, fertility testing of both sexes is essential.
Estimating the Number of "Infertility Genes" in Mice
If we assume that the mutation rate in these experiments was close to reported estimates for the level of ENU used in the present study (3 x 80–85 mg/kg), approximately one per 1000 loci [54, 55], and that approximately 35 000 genes are found in mice [56], then each G1 animal would inherit approximately 35 recessive mutations. Furthermore, if we assume that each of these mutations was captured in at least one G2 daughter, that sufficient G3 animals were generated to guarantee obtaining at least one homozygote for each of these mutations, and that the rate of sterility mutations recovered is at least 8.0% of families, then we can predict from the ENU screen that 1 in 437 genes is required for fertility in mice (80 total). It is currently known that more than this number of genes is required for gametogenesis, suggesting that this method yields a gross underestimate [20]. Similar calculations based on the EMS mutation rates and the number of sterility mutations recovered in males would yield low-end estimates of 300 (1 in 120) to 1100 (1 in 30) genes required for gametogenesis in mice. These numbers are almost certainly an underestimation for the reasons described below.
First, null mutations in genes that are required for viability in addition to fertility, such as Brca1 [57] and, probably, Rad51 [21, 22, 58], are not recoverable in these screens. Second, mice with visible defects, such as spotting, runting, or dysmorphologies, were omitted from the screens performed here. This would eliminate the detection of mutations such as W (Kit) or those of Wnt7a [59]. Third, as discussed later, the actual mutation load present in treated animals in different experiments and within different strains is not clear [60], so the assumption of a 1 in 1000 mutation rate for the ENU-treated B6 males may be excessive. Finally, the screens were not performed to a sufficient depth to obtain homozygotes for each of the mutations existing in each G1 male, and most of the experiments did not screen for female infertility. In sum, one can estimate that in mice, probably at least 80–1100 genes, when mutated, cause infertility in the absence of other observable defects.
In comparison, recent studies have postulated the genetic contributions to gametogenesis in other eukaryotes. In Caenorhabditis elegans, it is estimated that 1416 transcripts are expressed in the germ line [61]. These data, from a microarray containing 11 917 genes, suggest that approximately 12% of the genes in the genome are used for germ line development. Similarly, in the budding yeast Saccharomyces cerevisiae, a core group of approximately 900 genes of 6220, or 14%, were identified as being regulated during meiosis [62]. Because of the complexity of germ-soma interactions in mammals and a much larger genome, the number of genes expressed and required for germ cell development is probably much higher.
Comparative Efficiencies of Whole-Animal Mutagenesis (ENU) vs. ES Cell (EMS) Mutagenesis
Several issues should be considered when drawing conclusions about the relative efficiency of traditional whole-animal ENU mutagenesis compared to ES cell mutagenesis. These include the degree and certainty to which the genome is mutagenized, the number of animals that must be screened to detect the mutations, and the time and effort required to conduct the screens.
In the ENU experiments described here, offspring derived from a total of 102 G1 males, representing 102 mutagenized haploid genomes, produced eight genetic phenodeviants resulting in failed gametogenesis (1 in 12.75 genomes). In the ES cell approach, seven chimeras, representing at least 14 haploid mutagenized genomes, produced four heritable male-specific mutations (1 per 1.75 chimeras, or 1 per 3.5 genomes). If we add to these numbers the previously reported results of three chimeras derived from the same mutagenized ES cell batch as used in experiment 4 [49], which yielded the meiotic mutation mei1 [53], then the totals for ES mutagenesis would be one mutation per two chimeras (four haploid genomes). On the basis of these results, chemical mutagenesis of ES cells appears to have induced an approximately 3-fold higher load of mutations per haploid genome than ENU mutagenesis of whole mice. The degree to which these results can be generalized, however, is not certain.
The data presented here reflect ENU mutagenesis of C57BL/6J mice; other strains may be more highly mutable. On the other hand, the present data regarding ES cell mutagenesis reflect only two batches of EMS-mutagenized cells, which were shown to have a mutation rate at the Hprt locus of 1/4524 and 1/1200 [49]. Using ENU treatment of ES cells, Chen et al. [63] induced inactivating mutations at Hprt at a frequency as high as 1/455. In practice, a point exists at which higher mutation rates may complicate matters. For example, the mutations mei2.5 and mei2.7 arose from the same chimera, and because the phenotypes were so similar, they were believed to be identical until mapping inconsistencies led to the realization that they were independent. Thus, we believe that the mutation load of the cells used here was either close to optimal or slightly higher than desirable for the breeding paradigm employed.
The chemically induced mutational load appears to be higher in ES cells than in spermatogonia of ENU-treated mice, but the ENU screens were more efficient in terms of the effort required for performing the mutational screens. This greater efficiency resulted from differences in the mating schemes. With ENU, the strategy was to produce four G2 daughters from each G1 male and eight G3 progeny from each G1 x G2 mating. This approach has a 94% probability that at least one G2 female will inherit a given mutation present in the G1. In the case of a sterility mutation affecting both sexes, the chance of at least one of the eight G3 offspring being a homozygote is 90%, but this same chance is only 68% if it affects only one sex (assuming half the pups [n = 4] are of the affected sex). However, because a 69% probability exists that two or more G2s will inherit a given mutation, the odds are better than 90% that a sex-specific mutation would be detected among the litters of G2 females. Given a screen conducted in this way, each family (representing one mutagenized haploid genome) involves five matings (G1 male to wild-type female to produce G2 offspring, and then the G1 crossed to each G2), 38 mice, and three generations.
The crosses performed with chimeras derived from mutagenized ES cells involved many more animals per haploid genome (a total of 133 G3s per chimera, as opposed to 20 G3s per G1 male in the ENU screens). This is because each of multiple G1 offspring from an individual chimera was used to initiate families consisting of one to four G2s and several G3s per G2. An alternative strategy would have been to produce just one G1 male from each chimera and then treat it as in the ENU screen. This would have sacrificed the recovery of half the mutations residing in any ES cell-derived spermatogonium, but it would have minimized the diminishing returns from pursuing them all.
New Methods
An alternative-breeding scheme afforded by ES cell mutagenesis could provide great advantage by reducing the number of generations required to render mutations homozygous. This can be accomplished in one of two ways. First, daughters of male chimeras could be backcrossed to the chimera as long as the chimera produces sperm exclusively or predominantly derived from ES cells and not from the host blastocyst. This may be ensured by using tetraploid host blastocysts [64] or only chimeras judged to have germ lines that are essentially 100% ES cell-derived (using visible genetic markers and assessing litters).
Second, sibling offspring (G1) of a chimera can be intercrossed. Both approaches are dependent on the chimera transmitting a genome derived from a single mutagenized clone. In previous work, evidence was presented that the germ lines of chimeras produced from nonclonal mutagenized ES cell cultures consisted primarily, if not exclusively, of a single ES cell clone [49]. The evidence was that when progeny of most chimeras were intercrossed, the litters were extremely small or nonexistent. This was attributed to a high load of recessive mutations, because these animals were highly fecund when outcrossed to wild-type mice. In more recent work, this problem was alleviated by generating chimeras from ES cells bearing a lower mutation load, enabling the recovery of mutants by chimera-to-daughter matings and by intercrosses of G1s (unpublished results). Not only do these approaches save an entire generation of breeding time, but the mutation load is less diluted than with the three-generation schemes.
Recommendations for Future Experiments
From these initial experiences, we can make some recommendations and observations regarding strategy for conducting these screens in the most efficient manner. Most important is the ability to map genetically new mutations using affected progeny produced by the G1 and G2 parents. Maintenance of mutant stocks carrying segregating, unmapped recessive sterility mutations is difficult, because carriers can be identified only by progeny testing (mating potential heterozygotes to known heterozygotes, then testing progeny for sterility to reveal whether the parent in question was, indeed, a carrier). On the identification of sterile G3 mice, the only known carriers are the G2 mother and the G1 father. If either dies before the mutation is mapped, then one must conduct progeny testing of animals that may or may not be carriers. Hence, it is important to obtain sufficient numbers of G3 animals to map new mutations genetically. Mapping generally can be accomplished with eight or more affected animals by using techniques (e.g., interval haplotype analysis) that require genome screens involving a total of only 40 or so markers [48, 65].
Achieving this goal is expedited if more than one G2 carrier exists. Specifically, with four G2 females, the odds are 11/16 (69%) that two or more will inherit a particular mutation from the G1 sire. This is in contrast to a 50% probability that two or more G2s would be carriers if only three females were available. With at least two G2 carriers, the likelihood that it will be possible to map a new mutation solely with G2 animals increases. For example, if the G1 x G2 cross yields eight pups on average and a particular mutation affects both sexes, then each litter would have two affected animals on average. With two G2 carriers, each would need to produce two litters to yield a total of eight pups, which would be sufficient to obtain linkage. Otherwise, with one affected animal, four litters would need to be obtained (which would take a long time, or which might not happen at all) to get eight affected animals. Regarding mutations affecting only one sex, each G1 x G2 cross would produce a single affected G3 per litter of eight, thus necessitating a total of at least eight litters to achieve a sufficient number of affected offspring to obtain map positions. Again, this is not reasonable given even two G2 carriers. However, it should be considered that with mutations affecting a single sex, there will be fertile homozygotes of the opposite sex that become invaluable breeders. This fact was exploited with some of the mutations discussed here. In sum, when seeking to identify and map sterility mutations on a genome-wide basis, it is probably best to employ a strategy involving the generation of a single G3 litter from four or more G2 females rather than generating multiple G3 litters from a smaller number of G2 females. Similar issues apply to ES cell mutagenesis.
Random Mutagenesis Will Yield Mutations Affecting Multiple Stages of Gametogenesis
One class, including Sgdp and gcd2, affects germ cell survival or proliferation. In both cases, seminiferous tubules become depleted of germ cells, resembling a SCO phenotype. Whether the loss of spermatogonial cells can be attributed to a failure in migration or proliferation of PGCs or to a defect in spermatogonial and, possibly, oogonial maintenance is currently being investigated.
Examples of germ cell-deficiency phenotypes resulting from mutations in genes that affect PGC migration and/or proliferation include the dominant white spotting (W) and steel (Sl) loci, which encode the KIT receptor and its ligand, also known as stem cell factor, respectively [66, 67]. Mutations that affect the stem cell decision to divide or differentiate are jsd (juvenile spermatogonial depletion) [68] and a targeted mutation of FancC [69]. Mutations of this type have the potential to reveal new genes involved not only in reproduction and fertility but also in stem cell biology and cancer.
A second predominant class is the meiotic arrest mutants. Germ cell development in mei1 mutant males arrests at the zygotene-pachytene transition [53], when there appears to be a checkpoint for chromosome repair and synapsis defects. The mei4 mutant gonocytes progress further, to late meiosis I, in both males and females before undergoing arrest. Similar phenotypes are seen in the gene disruption mutants of Mlh1 [27] and Mlh3 [28]. The MEI4 may function as a regulator of sister chromatid cohesion or homologous chromosome exchange. The 11b female-specific mutation may affect a gene required for meiotic maturation, because oogenesis arrests just before GVBD. Disruption of cdc25b [70] and inhibition of phosphodiesterase 3 have a similar effect [71]. Because 11b has yet to be genetically mapped, it is not known if it encodes either of these genes or components of MPF (maturation-promoting factor).
Another class contains postmeiotic and spermiogenesis mutants. The process of spermiogenesis involves a lengthy, coordinated differentiation program, including repackaging of the haploid genome, sperm morphological development, and posttesticular (epididymal) maturation. Examples of genes involved in various stages of spermiogenesis include protein phosphatase 1 catalytic subunit gamma (Ppp1cc;) [72], which when disrupted causes early spermiogenetic defects in the round spermatid; JunD1, targeted mutations of which cause aberrant sperm head and flagellum development, leading to male infertility [73]; and Adam3, the disruption of which causes adhesion dysfunction during fertilization [74]. At the moment, the possible functions of the spermiogenesis mutants identified here are unknown. Identification of the underlying genes will likely be required to understand the precise nature of the defects.
A point worth mentioning is the relative robustness of infertility as a phenotypic screen. It is simple to assess animals for infertility. Furthermore, of the 11 mutations reported here, all but gcd2 were 100% penetrant at the age of fertility testing, which facilitates genetic mapping. (Systematic analyses have not been performed, but fertile females homozygous for gcd2 appear to undergo premature menopause.) Some of the other mutations, although 100% penetrant for infertility, show variable expressivity. As mentioned earlier, genetic background likely affects penetrance and expressivity, and this eventually may be exploited to discover modifier loci encoding other molecules in the same pathway.
One question that cannot yet be answered is the degree to which genes identified in forward genetic screens will be novel versus merely those that have not previously been subjected to targeted mutagenesis. The answer must await the cloning of the mutated genes responsible for the phenotypes. Presently, only mei1 has been identified, and this is a gene with previously unknown function (unpublished results). Based on map positions and ongoing positional cloning efforts directed at several of these new mutations, most or all clearly do not coincide with genes previously known to cause infertility when mutated.
The studies reported here demonstrate the feasibility and productivity of conducting forward genetic screens for infertility mutations in mice. At the moment, it appears likely that a substantial fraction of the mutations induced will be in genes with functions in gametogenesis that are not known or in genes that have not previously been subjected to targeted mutation. As the mutagenesis program expands, a clearer picture will emerge regarding the numbers of genes required for gametogenesis in mice, shedding light on the evolution of mammalian gametogenesis, clarifying essential gametogenic pathways, and identifying genes that underlie human infertility.
|
ACKNOWLEDGMENTS |
|---|
|
FOOTNOTES |
|---|
2 Correspondence: FAX: 207 288 6082; jcs{at}jax.org
Received: 2 June 2003.
First decision: 29 June 2003.
Accepted: 8 July 2003.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  T. Harris, B. Marquez, S. Suarez, and J. Schimenti Sperm Motility Defects and Infertility in Male Mice with a Mutation in Nsun7, a Member of the Sun Domain-Containing Family of Putative RNA Methyltransferases Biol Reprod, August 1, 2007; 77(2): 376 - 382. [Abstract] [Full Text] [PDF]
|
|
