U.S. flag

An official website of the United States government

NCBI Bookshelf. A service of the National Library of Medicine, National Institutes of Health.

Riddle DL, Blumenthal T, Meyer BJ, et al., editors. C. elegans II. 2nd edition. Cold Spring Harbor (NY): Cold Spring Harbor Laboratory Press; 1997.

Cover of C. elegans II

C. elegans II. 2nd edition.

Show details

Section IIIMeiosis

A. Cytology of Meiotic Prophase

C. elegans chromosomes undergo all of the classically described stages of meiotic prophase in preparation for meiosis I, the reductional division. In both males and hermaphrodites, germ-cell nuclei representing all stages of meiotic prophase are arranged in temporal order along the distal-proximal axis of the adult gonads. As nuclei enter the extended prophase that precedes the first meiotic division, homologous chromosomes partially condense and pair, becoming aligned and physically associated, or synapsed, in a side-by-side configuration along their entire lengths. Alignment of homologs is complete at the pachytene stage, with the chromosomes localized to the periphery of the nucleus (Fig. 2a).

Figure 2. Stages of meiotic prophase.

Figure 2

Stages of meiotic prophase. ( a,b ) Chromosomes in unfixed dissected germ-cell nuclei stained with Hoechst 33342; (c,d) oocyte chromosomes fixed and (more...)

Electron micrographs of sectioned pachytene nuclei show normal-appearing synaptonemal complexes (SC), the highly ordered proteinaceous structures located at the interface of synapsed homologs (Goldstein and Slaton 1982). Hermaphrodite pachytene nuclei contain six tripartite SCs (corresponding to five autosomal bivalents and one X chromosome bivalent) composed of two lateral elements (each 35 μm width) flanking a striated central element (20 μm width). One end of each SC is attached to the nuclear envelope. Males, which have only a single X chromosome, contain five SCs, and the univalent X chromosome is present in a condensed heterochromatic state (Goldstein 1982).

Later in prophase, at the diplotene stage (Fig. 2b), the chromosomes desynapse but remain condensed and are held together by chiasmata. The chromosomes detach from the nuclear envelope and continue to condense as nuclei move into the diakinesis stage of meiotic prophase. In developing oocytes, nuclear volume increases as chromosome condensation proceeds, allowing individual bivalents to be visualized and counted (Fig. 2c). Wild-type oocytes reliably have six bivalents at diakinesis, consistent with all six homolog pairs having undergone a crossover (Villeneuve 1994). The chiasmata can no longer be readily distinguished at this stage due to the extremely condensed state of the bivalents, but their presence can be inferred from the fact that the homologs remain attached; when chiasmata are absent due to failure in crossing over, unattached univalent chromosomes are observed (Fig. 2d) (Villeneuve 1994; see below). In developing spermatocytes, the nuclear volume is much more compact, making it difficult to distinguish individual chromosomes or bivalents.

At diakinesis (and at metaphase I), the homologous chromosomes that constitute the bivalent are associated in an apparent end-to-end configuration (Nigon and Brun 1955; Herman et al. 1979; Albertson and Thomson 1993). in situ hybridization studies have shown that either end can be oriented toward the outside. For most chromosomes, which end is outside appears to be random, but chromosome I exhibits a bias such that the right ends (which contain the ribosomal DNA gene cluster) are more frequently located on the outside (Albertson and Thomson 1993). The apparent end-to-end association of homologs at this stage has been taken to reflect terminalization of the chiasmata, although there is debate in the literature over whether terminalization actually occurs in organisms for which meiosis I is the reductional division (Jones 1987). The end-to-end appearance of the bivalents at late diakinesis might instead be a consequence both of the original distribution of crossover events and of the progressive condensation of the chromosomes that occurs as prophase proceeds: Nearly 90% of autosomal crossovers occur in the terminal thirds of the chromosomes (Barnes et al. 1995; see below), and condensation results in an estimated three- to tenfold shortening in chromosome length (concomitant with an increase in width) between the pachytene stage and late diakinesis. Thus, lateral projections observable at diplotene may be absorbed into the width of the bivalent by late diakinesis, resulting in bivalents that appear terminally associated. Whether or not terminalization occurs, the inside/outside orientation of the bivalents is most likely determined by the initial position of the crossover, with the end nearest the crossover becoming the inside end. If this hypothesis is correct, then the inside/outside orientation at diakinesis should be predictable for any bivalent in which the location of crossovers is constrained by regional crossover suppression. This prediction has been borne out for a number of different rearrangements, an example of which is shown in Figure 3 (Albertson et al. 1995; D.G. Albertson, unpubl.; see below).

Figure 3. Orientation and segregation of meiotic chromosomes.

Figure 3

Orientation and segregation of meiotic chromosomes. The axial orientation of the meiotic chromosomes at metaphase I and II was determined using fluorescent in situ hybridization with a probe for (more...)

In oocytes, meiosis pauses in diakinesis, and the oocyte nucleus does not go on to complete the meiotic divisions until after ovulation and fertilization have occurred (for a description of oocyte maturation and ovulation, see Schedl, this volume); similar delay or arrest during meiotic prophase is a common feature of oocyte meiosis in animals. Spermatocytes, in contrast, proceed immediately into the first and second meiotic divisions.

B. Cytology of Metaphase I and II

At metaphase I of meiosis in both spermatogenesis and oogenesis, the meiotic bivalents appear highly condensed, and ultrastructural studies have failed to differentiate a kinetochore structure on the chromosome. However, in certain species of the plant genus Luzula, in which nonlocalized kinetochores have been demonstrated in mitosis (Braselton 1971; Bokhari and Godward 1980), a single, localized kinetochore-like structure has been distinguished on meiotic chromosomes (Braselton 1981). By using different fixation or staining procedures, it might also be possible to reveal kinetochore-like structures on the meiotic chromosomes of C. elegans. In electron micrographs of C. elegans meiotic chromosomes, microtubules appear to insert directly into the chromatin (Albertson and Thomson 1993), but it is currently not known where the microtubules attach or if there is a specific site of attachment.

At metaphase I and II, the chromosomes congress to the spindle equator with an outer ring of five chromosomes surrounding a central sixth chromosome. At metaphase I in males, the central chromosome was identified as the univalent X chromosome from reconstruction of electron micrographs of serial sections through metaphase I spindles. The autosomal bivalents formed the outer ring, with each half bivalent joined at the spindle equator by a chromatin thread, possibly corresponding to a chiasma. The central location of the X was also suggested by light microscopic observations of metaphase I. In hermaphrodites and at metaphase II in both sexes, the X chromosome cannot be distinguished morphologically from the autosomes (Albertson and Thomson 1993). We speculate that the observed differential placement of the X chromosome at metaphase I may be a consequence of unspecified mechanisms operating to ensure proper disjunction of the unpaired X in males (Hodgkin et al. 1979), which do not function in XO animals that have been transformed into hermaphrodites by a her-1 mutation (Hodgkin 1980).

1. The Meiotic Spindle

Ultrastructural studies revealed that the meiotic spindle apparatus is organized differently in spermatogenesis and oogenesis (Albertson and Thomson 1993). In spermatogenesis, centrioles are present in the poles of the spindle, and the spindle appears similar to mitotic metaphase spindles. Segregation of the chromosomes to the spindle poles in the two maturation divisions results in four sperm, each with a centriole closely associated with the chromatin. The sperm contributes centrioles to the oocyte, and following fertilization, the sperm centrosome divides and organizes the first mitotic cleavage spindle (Albertson 1984b).

The meiotic maturation divisions of oogenesis take place in the oocyte after fertilization. At both meiotic divisions, the spindles are acentriolar and have a barrel-shaped morphology typically seen in oogenesis of many organisms (for review, see McKim and Hawley 1995). Formation of the meiotic spindles has been studied in fixed specimens by indirect immunofluorescent staining with antitubulin antibodies (Albertson and Thomson 1993). The early stages of meiotic spindle formation appear to proceed without well-defined organizing centers, since microtubules are first seen in an apparently disorganized array around the chromosomes. As the chromosomes congress, two well-defined poles at the ends of an elongated spindle can be distinguished. At this stage, a maximum pole-to-pole length of 13 μm was measured. Progression toward metaphase results in shortening of the spindle to less than one third of the initial length, reaching a minimum pole-to-pole length of 3−4 μm at anaphase. Prior to metaphase, the spindles are oriented parallel to the surface of the embryo, but by anaphase, they rotate to lie perpendicular to the surface with one end of the spindle closely apposed to the membrane. At telophase, the blocks of chromatin separate to a maximum of 3−4 μm, leaving all antitubulin staining material lying between them.

2. Genes Involved in Formation of the Acentriolar Meiotic Spindle

The assembly of the acentriolar spindle takes place within a cytoplasm that will shortly undergo mitosis on a conventional centriolar spindle. This raises questions as to how assembly of these two different spindle structures is regulated. Genetic analysis is providing insights into both the process and the regulatory pathways associated with acentriolar spindle formation and coordination of mitosis and meiosis in a common cytoplasm. The mei-1 gene is required for spindle formation in the female germ line, since loss-of-function mutations in this gene disrupt meiotic spindle formation (Mains et al. 1990a, b). However, mei-1 gene activity must be eliminated prior to metaphase, since dominant gain-of-function alleles of mei-1(ct46) disrupt mitotic divisions (Clandinin and Mains 1993). Molecular cloning and sequencing of mei-1 revealed that the protein product is a member of a family of ATPases that share a highly conserved nucleotide-binding site (Clark-Maguire and Mains 1994a). Immunolocalization of the protein in metaphase I meiotic spindles revealed staining distributed throughout the spindle, with the highest concentration at the spindle poles. In mei-1 loss-of-function (lf) mutants, antitubulin staining at metaphase I remains diffuse, whereas no MEI-1 immunostaining is detected, suggesting that MEI-1 may play a part in promoting the organization of spindle poles (Clark-Maguire and Mains 1994b). Two genes have been identified that restrict mei-1 activity to meiotic divisions: mel-26 appears to be a postmeiotic inhibitor of mei-1 activity, whereas mei-2 is required to localize mei-1 to meiotic spindles (Clark-Maguire and Mains 1994b).

3. Alignment and Segregation of Chromosomes in Meiosis

Observations on a variety of organisms with holocentric chromosomes indicate that in a given species, alignment of the meiotic bivalent on the metaphase I spindle takes one of two possible orientations (White 1973). The bivalent can lie parallel to the equator of the spindle (equatorial orientation) or it can be aligned with the long axis parallel to the spindle pole axis (axial orientation). When bivalents orient axially, sister chromatids segregate to the same pole at anaphase I, so that the first meiotic division is reductional, as occurs in meiosis in species with monocentric chromosomes.

In C. elegans, the meiotic chromosomes adopt the axial orientation at both metaphase I and II (Albertson and Thomson 1993; D.G. Albertson, unpubl.). Because the bivalents are small and highly condensed, in situ hybridization was used to label one end of the chromatids of linkage groups I, II, or V. If the bivalent were to adopt the axial orientation, then the expectation would be that some metaphase figures should contain chromosomes in which the labeled ends are proximal to the spindle poles and not located on the spindle equator. If the chromosomes adopt the equatorial orientation, then the labeled ends should always lie on the equator. The axial orientation was demonstrated by the observation that in some metaphase I and II figures, the labeled ends of the chromatids were seen proximal to the spindle poles.

The monocentric centromere ensures the orderly disjunction of the chromatids by performing two functions in meiosis. First, it provides the site of attachment of the spindle microtubules. Second, it holds the sister chromatids together through meiosis I and then directs their splitting in meiosis II. In the absence of any cytologically visible centromeric structure in meiosis, the question arises as to the nature of the meiotic centromere on C. elegans chromosomes. It has so far not been possible to identify with certainty the site of attachment of microtubules to the meiotic bivalent because of the density of microtubules in the meiotic spindle (Albertson and Thomson 1993). However, it seems likely that attachment is made at the poleward end of the chromatids in metaphase I, and therefore this end of the chromatid performs one of the functions of the monocentric centromere at meiosis I. Does this end of the chromosome also hold the sister chromatids together and so perform both of the classical functions of the centromere? For the heterozygous translocation of chromosome I, hT3(I;X), the answer is yes (D.G. Albertson, unpubl.). Since the bivalent of the heterozygous translocation adopts a fixed orientation in diakinesis, the poleward ends of the chromatids are also fixed at metaphase I. It was therefore possible to ask whether the chromatid ends that were poleward at metaphase I also held the chromatids together in metaphase II. As shown in Figure 3, labeling one end of the chromatids by in situ hybridization revealed that the end of the chromatid that had been poleward at metaphase I was always at the equator at metaphase II. Therefore, one end of the sister chromatids performs both functions of the monocentric centromere. However, it is worth noting two differences in the behavior of these holocentric chromosomes compared to monocentric chromosomes. First, either end of the chromatid can perform the centromeric functions in metaphase I. Second, at metaphase II, the opposite end of the chromatid is proximal to the spindle pole compared to metaphase I, and so presumably spindle attachment takes place at different ends of the chromatids in the two divisions.

C. Genetic Organization of Chromosomes

The genetic length of the C. elegans chromosomes is independent of their physical length and is a reflection of the functional regulation of the meiotic process. Although the chromosomes range in size from 12 to 20 Mb of DNA, the genetic map of each of the six chromosomes is approximately 50 map units in length, consistent with approximately one crossover event per bivalent per meiosis (Brenner 1974; Barnes et al. 1995). Since crossing over between homologs is required to ensure their disjunction at the meiosis I division (Hawley 1988; Zetka and Rose 1992), recombination must be carefully regulated to guarantee that a limited number of crossover events are distributed so that each bivalent undergoes a crossover. This regulation is exemplified by the observation that when crossing over is eliminated from one portion of a chromosome (e.g., due to heterozygosity for a regional crossover suppressor), there is often a compensatory elevation of crossing over in other regions of the chromosome (McKim et al. 1988, 1993; Zetka and Rose 1992). In an extreme example, when crossing over was restricted to a small (normally 6-map-unit) interval on chromosome I in hermaphrodites heterozygous for two different rearranged chromosomes, the frequency of exchange in that interval was nearly the full amount expected for the entire length of the chromosome (Zetka and Rose 1992).

The C. elegans genetic map represents the synthesis of a large volume of data from 2-factor mapping experiments (measuring recombination frequency between genetic markers in hermaphrodites), 3-factor and multifactor mapping experiments (determining the relative order of genetic markers), and deficiency and duplication mapping experiments (testing for complementation of marker mutant phenotypes) (Edgely and Riddle 1993). The current genetic map, as well as the data used to construct it, can be accessed using ACeDB (Eeckman and Durbin 1995). A striking feature of the meiotic maps of each of the autosomes, but not the X chromosome, is a central region where the number of meiotic exchanges per unit length of DNA is greatly reduced relative to the genomic average of 300 kb per map unit (Greenwald et al. 1987; Prasad and Baillie 1989; Starr et al. 1989; Barnes et al. 1995). On chromosome I, for example, the number varies from as much as 1200 kb per map unit in the dpy-5 unc-13 region to as little as 100 kb per map unit in the unc-101 unc-54 region (Barnes et al. 1995; Zetka and Rose 1995b). These regions of low recombination have been called “gene clusters” because of their appearance on the genetic map (Brenner 1974), but they are also regions where the genes are physically closer together (see Sulston et al. 1992; Barnes et al. 1995; Waterston et al., this volume). Flanking the gene clusters are regions (referred to as the chromosome “arms") where exchange preferentially occurs. In some cases, these regions have been shown to contain recombinationally hot regions, which have crossover frequencies as much as tenfold higher than the genomic average (Hodgkin 1993; Pilgrim 1993; Clark-Maguire and Mains 1994a). A comparison of the frequency of crossing over and gene density across the entire genome has shown that most recombination occurs in gene-poor regions, and it has been proposed that recombination-promoting sequences exist at high density in the chromosome arms (Cangiano and LaVolpe 1993; Barnes et al. 1995).

The meiotic pattern of crossing over that gives rise to the genetic map is under genetic control and can be altered by a mutation in the rec-1 gene. Animals carrying the rec-1 mutation have the same total number of exchange events per chromosome each meiosis, but the placement of exchanges is altered, indicating that meiotic mechanisms determining the number of crossovers are separable from mechanisms determining their positions (Zetka and Rose 1995b). On chromosome I, for example, genetic distances expand across the medial gene cluster (Rose and Baillie 1979b; Rattray and Rose 1988) and contract near the right end of the chromosome (Zetka and Rose 1995b), resulting in a map that is more consistent with the physical distances than is the meiotic map. Despite causing a redistribution of autosomal crossover events during meiosis, the rec-1 mutation does not appear to have any effect on the fidelity of chromosome segregation or on growth and viability. This result suggests that the wild-type distribution of crossover events is neither required for faithful meiosis nor is it an indirect consequence of a feature of chromosome organization that serves some other essential function, raising questions about how the wild-type crossover distribution originated and how or why it is maintained. Mutations in several other genes also disrupt the distribution of meiotic exchanges (see below). Unlike rec-1 , however, these mutations cause an overall reduction in the amount of crossing over, leading to nondisjunction.

1. Factors Influencing Crossover Frequency

Crossing over occurs in all three C. elegans germ lines (oocyte and spermatocyte in the hermaphrodite, spermatocyte in the male), and the frequency of recombination between genetic markers can be affected by a variety of environmental and physiological factors. These include temperature, age, and sex of the parent (Rose and Baillie 1979a; Kim and Rose 1987; Zetka and Rose 1990). For several pairs of markers within medial gene clusters, and one genetic interval spanning a cluster/arm boundary, crossover frequencies in both males and hermaphrodites were observed to increase with temperature and to decrease with parental age. Furthermore, for both genetic intervals tested, recombination frequencies were higher in hermaphrodite spermatocytes than in oocytes. Since spermatocytes are produced in the fourth larval stage, before the switch is made to oocyte production, these differences may also be attributable to differences in parental age. The effects of temperature and age have not been measured for other regions of the chromosomes, however. It is possible that the observed increases and decreases in crossover frequencies within the gene clusters may be offset by compensatory changes in crossover frequencies on the chromosome arms, as has been observed in certain cases of crossover suppression (McKim et al. 1988). Because many factors can influence the frequency of crossing over between markers, standardized conditions are recommended for genetic mapping experiments and for experiments comparing recombination frequencies in different strains (Rose and Baillie 1979a).

2. Crossing Over and Segregation in Spermatocytes

Recombination frequencies in males have been measured along the entire length of chromosome I, yielding a male genetic map that is approximately one-third shorter than the standard chromosome I map derived from measurements of genetic recombination frequencies in hermaphrodites (Zetka and Rose 1990). Although double crossovers have been detected in C. elegans males, their frequency is not high enough to account for the reduced size of the male map (Hodgkin et al. 1979; Zetka and Rose 1995b). Unless there are undetected crossovers in regions at the tips of the chromosomes beyond known genetic markers, the length of the male map suggests that a significant fraction of chromosomes in male spermatocytes may be achiasmate. (Unfortunately, the frequency of achiasmate chromosomes in spermatocytes cannot be readily assessed cytologically as it can in oocytes due to the compact structure of the spermatocyte nucleus.) Nevertheless, males reliably segregate their homologous chromosomes, raising the possibility that in C. elegans males, a crossover per bivalent is not necessary for proper disjunction of homologs.

Mechanisms for ensuring the segregation of achiasmate chromosomes have been described for other organisms (Hawley et al. 1992), and several lines of evidence suggest that similar mechanisms may operate in C. elegans spermatocytes but not in oocytes. For example, in hermaphrodites mutant in either of two loci required specifically for normal X chromosome crossing over, the frequency of X chromosome nondisjunction is substantially higher in oocytes (which rely on chiasmata for X chromosome disjunction) than in spermatocytes, despite the fact that recombination appears to be strongly reduced in both gamete lines (Broverman and Meneely 1994; Villeneuve 1994). Unless there is a particularly high frequency of undetected crossovers near the X chromosome ends in these mutants, these data suggest that hermaphrodite spermatocytes may have the capacity to segregate achiasmate X chromosomes, at least when autosomal crossing over is normal. The existence of a recombination-independent segregation system is also supported by the observation that free duplications (Herman et al. 1979; McKim and Rose 1990) and extrachromosomal arrays exhibit a modest tendency to segregate from the lone X chromosome in male spermatocytes. There is, however, no direct evidence for a system of this type that would be stringent enough to ensure faithful segregation.

D. Cis- acting Chromosomal Features Involved in Pairing and Crossing Over

Genetic studies examining the meiotic behavior of chromosomal rearrangements have provided evidence that each of the six C. elegans chromosomes has a specialized chromosomal domain located near one end that has an important role in the pairing of homologous chromosomes. A key observation is that many reciprocal translocations behave as efficient crossover suppressors in C. elegans (Rosenbluth and Baillie 1981; Herman et al. 1982; Ferguson and Horvitz 1985; Fodor and Deak 1985; McKim et al. 1988 1993). In individuals heterozygous for such reciprocal translocations, crossing over readily occurs in chromosomal segments extending from one end of the chromosome to the translocation breakpoint, but it is strongly suppressed or eliminated from the breakpoint to the other end of the chromosome (Rosenbluth and Baillie 1981; McKim et al. 1988 1993). Thus, each of the half-translocations consistently recombines with, and segregates from, only one of the two normal sequence chromosomes (Rosenbluth and Baillie 1981; McKim et al. 1988). In the case of eT1(III;V), for example, one half-translocation (consisting of the left portion of chromosome III and the left portion of chromosome V) always crosses over with and segregates from the normal chromosome III, whereas the reciprocal half-translocation consistently crosses over with and disjoins from the normal chromosome V. At the same time, crossing over is essentially eliminated for the right portion of chromosome III and the left portion of chromosome V. The fact that crossing over is strongly suppressed along the entire length of these translocated segments suggests that the absence of crossing over may result from a failure of these segments to pair with their homologs. Moreover, the ability of the reciprocal segments to recombine with their normal homologs suggests that features which facilitate homologous pairing are asymmetrically distributed on the chromosomes. Several translocations exhibit similar asymmetries in exchange and segregation behavior, and the accumulated data are all consistent with the proposal that each of the six chromosomes has a single localized domain, probably near one end, that facilitates crossing over, presumably through involvement in the process of homolog pairing (Fig. 4) (McKim et al. 1988).

Figure 4. Physical map locations of the HRR/pairing centers.

Figure 4

Physical map locations of the HRR/pairing centers. Each of the six C. elegans chromosomes is predicted to contain a specialized cis-acting chromosomal domain, termed the (more...)

What is the basis for crossover suppression in the reciprocal translocation heterozygotes? In many organisms, cytologically observable tetravalents form during meiotic prophase in reciprocal translocation heterozygotes, with each segment of the translocation chromosomes successfully pairing and forming crossovers with the corresponding segment of the normal sequence chromosomes (Fig. 5a). In C. elegans, the patterns of crossing over and segregation in balancer translocation heterozygotes are compatible with the absence of a tetravalent (Fig. 5b) and suggest instead that two separate bivalents may form, each of which includes a segment of homologous synapsis (where crossovers can occur) adjacent to a segment of nonhomologous synapsis (where crossovers cannot occur) (Rosenbluth and Baillie 1981). Reconstructions from electron microscope serial sections of pachytene nuclei from one translocation heterozygote suggest that this is in fact the case; six normal-appearing bivalents were observed, indicating that translocated chromosome segments had participated in nonhomologous synapsis (Goldstein 1986). Thus, crossing over is apparently suppressed because the translocated segments are sequestered away from their partners in a nonproductive synapsed configuration (Villeneuve 1994). The simplest interpretation of the genetic data is that only one of the two possible pairing configurations is adopted for a given translocation, despite the fact that in many cases, both configurations have comparable amounts of sequence identity. The dominant configuration is determined by the location of the specialized pairing domains involved in initiating an early event in the homolog pairing process.

Figure 5. Possible pachytene pairing configurations in reciprocal translocation heterozygotes.

Figure 5

Possible pachytene pairing configurations in reciprocal translocation heterozygotes. (a) Tetravalent structure predicted to form if DNA sequence identity alone were the sole determinant of the (more...)

Further evidence for these important pairing domains is the asymmetrical ability of chromosomal duplications derived from different ends of a chromosome to undergo meiotic exchange with a normal full-length chromosome. For example, of a pair of duplications that cover different ends of chromosome I (sDp1 and sDp2), only the duplication of the right portion of chromosome I can undergo exchange with the normal homolog (Rose et al. 1984). Similar behaviors have been observed for duplications of the X chromosome (Herman and Kari 1989) and chromosome IV (Rogalski and Riddle 1988). In all of these cases, only duplications that are derived from the region of the chromosome predicted to contain the pairing domain (based on translocation studies) can recombine with the normal homolog, whereas duplications of other chromosomal regions do not.

These domains have been termed “homolog recognition regions” (HRRs) to denote their role in promoting the pairing of homologous chromosomes (McKim et al. 1988, 1993). They have also been referred to as “meiotic pairing centers” (Villeneuve 1994), employing terminology applied to potentially analogous chromosomal domains in maize (Maguire 1986). Although there is agreement that these chromosomal domains have an important role in homolog pairing, it is not known whether they function in the initial step of homolog recognition or in a later aspect of the pairing process.

Experiments examining the meiotic behavior of X chromosomes carrying deletions of the X chromosome HRR/pairing center region have provided additional information about the homolog pairing process (Villeneuve 1994). In hermaphrodites heterozygous for deficiencies of the left end of the X chromosome and a normal-sequence X chromosome, most X chromosomes were able to form crossovers with and disjoin from their homologs. This was true even for deficiencies that apparently removed the entire 4−5-map-unit region to which the HRR/pairing center had been localized by previous studies using translocations and duplications (Herman et al. 1982; McKim et al. 1988; Herman and Kari 1989). Only when the region was deleted from both homologs was a high frequency of noncrossover X chromosomes and a concomitant high frequency of X chromosome nondisjunction observed. Thus, the HRR/pairing center can apparently function even when it is present on only one of the two homologs. This meiotic behavior is inconsistent with the predictions of models proposing that the information content for homolog recognition is restricted to this region of the chromosome. According to this class of model, deletion of the region from only one of the two homologs would be just as detrimental as deleting it from both homologs, in either case rendering the homologs unable to find each other. The data instead suggest that the chromosomes utilize information outside the HRR/pairing center region to identify their pairing partners. It is not known whether autosomes deleted for their HRR/pairing center regions would exhibit meiotic behavior similar to that of the X chromosomes with HRR/pairing center deletions, since no comparable studies have been carried out.

How might an HRR/pairing center located near one end of a chromosome function to promote normal levels of crossing over along the entire length of the chromosome? Various related models have been proposed. One possibility is that the HRR/pairing center functions as an organizing center that nucleates the assembly of a bivalent which is competent to undergo crossing over, by acting as a binding site that loads trans-acting factors onto the chromosomes. These proposed trans-acting factors might function in the initial homolog recognition process, perhaps through involvement in a directional search for DNA homology that is initiated at the HRR/pairing center end of the chromosome. Alternatively, these trans-acting factors might function after homolog recognition, by promoting the initiation of homologous synapsis between prealigned chromosomes. In either case, the initial nucleation event would be followed by movement or distribution of factors along the length of the chromosomes, in a manner that would allow a chromosome with an intact HRR/pairing center to interact with a homolog from which the domain had been deleted. Initiation of synapsis in the vicinity of the HRR/pairing center would readily account for the observed and inferred synapsed configurations of various translocation heterozygotes. Furthermore, independent cytological evidence is consistent with a single site of synapsis initiation for the X chromosomes (Goldstein 1984). A variation on the above models is that the HRR/pairing center functions as a molecular address, targeting the end of the chromosome to a particular subnuclear location. This subnuclear localization would then serve to facilitate proper alignment and/or synapsis of homologous chromosomes.

Although the function of the HRR/pairing centers is required to promote normal levels of crossing over between homologs, it is unlikely that they are themselves the sites of initiation of recombination events (e.g., via the formation of double-strand breaks in the DNA), based on several considerations. First, crossovers frequently occur in regions of the chromosomes that are many megabases away from the HRR/pairing centers. Furthermore, in hermaphrodites heterozygous for some deletions (Rosenbluth et al. 1990), inversions (Zetka and Rose 1992), or insertions (McKim et al. 1993), crossovers have been observed to occur in intervals separated from the proposed HRR/pairing center by large stretches of nonhomologous DNA. It is difficult to envision how branch migration of hypothetical recombination intermediates could occur over such distances or through such obstacles. Moreover, initiation of a recombination event requires the presence of homologous DNA sequences on both homologs, whereas the analysis of deletion heterozygotes has suggested that the X chromosome HRR/pairing center can function even when it is present on only one of the two homologs (Villeneuve 1994).

Growing evidence indicates that multiple additional meiotic sites may be involved in alignment of homologous chromosomes and/or the initiation of recombination. In particular, studies examining crossover distribution in deficiency heterozygotes have suggested that the autosomes may have secondary sites, near the ends of the chromosomes opposite to the HRR/pairing centers, that participate in homolog alignment (Rosenbluth et al. 1990; McKim et al. 1993). Whereas these sites might be considered pairing sites, they are functionally distinct from, and subordinate to, the HRR/pairing centers.

Genetic and cytological experiments have provided evidence for chromosomal sites or domains in other organisms that may be analogous to the C. elegans HRR/pairing centers, although they differ in their chromosomal distribution. Whereas HRR/pairing center function appears to be concentrated to a single region on the small C. elegans chromosomes, the data suggest that there may be several pairing centers per chromosome in maize (Maguire 1986) and Drosophila (Hawley 1980).

E. Genes Required for Wild-type Level and Distribution of Meiotic Crossovers

Because pairing and crossing over between homologs are required to ensure their disjunction at the meiosis I division, mutations in genes involved in these processes can be identified among mutations causing a high frequency of meiotic nondisjunction. The chromosomal sex determination mechanism and reproductive lifestyle of C. elegans can be exploited to isolate such mutations in a straightforward fashion (Hodgkin et al. 1979; Herman et al. 1982; Villeneuve 1994). Since males (XO) arise among hermaphrodite (XX) self-progeny at a frequency of 0.2% due to spontaneous nondisjunction of the X chromosome in the hermaphrodite germ line (Hodgkin et al. 1979), C. elegans mutants with increased nondisjunction are readily identifiable as hermaphrodites that produce increased frequencies of male self-progeny. Such mutants have been termed him mutants, for high incidence of males (Hodgkin et al. 1979).

1. Mutations Preferentially Affecting the X Chromosomes

Mutations in several genes, him-1 , him-5 , and him-8 , preferentially affect the segregation of the X chromosomes. These mutations cause not only a reduction in crossing over on the X chromosomes (Hodgkin et al. 1979; Herman and Kari 1989; Broverman and Meneely 1994), but also an altered distribution of the crossover events that do occur (Broverman and Meneely 1994). Although the mutants exhibit reduced recombination over most of the length of the X chromosomes, they have normal or elevated levels of recombination in genetic intervals at the left end of X, near the region proposed to contain the HRR/meiotic pairing center (see above). Such an altered distribution of crossovers suggests that these mutations do not cause defects in the recombination machinery itself, which might be expected to produce a more uniform reduction in crossover frequency. Rather, these mutations may instead affect a function involved in regulating the formation of crossovers, most likely some aspect of the pairing process responsible for identification and alignment of homologous chromosomes in a configuration that is productive for crossover formation, or some feature of chromosome architecture that mediates access of the recombination machinery to its chromosomal DNA substrates. Independent evidence that him-8 may function in chromosome pairing comes from experiments showing that him-8 mutants exhibit elevated levels of intrachromosomal and/or unequal crossing over between tandemly duplicated segments of the X chromosome; these events are normally inhibited by homolog pairing (A.M. Villeneuve, unpubl.; K. Tanner et al., pers. comm.).

Of these three genes, only him-8 functions specifically in ensuring X chromosome segregation. Cytological analysis consistently reveals only a single pair of achiasmate chromosomes (the noncrossover X chromosomes) at diakinesis in him-8 oocytes (D.G. Albertson; A.M. Villeneuve; both unpubl.). In contrast, him-5 mutants exhibit some achiasmate autosomes (D.G. Albertson, unpubl.), as well as autosomal nondisjunction, and sterility at elevated temperatures (P. Meneely, pers. comm.). Furthermore, the isolation of several lethal him-1 alleles (Howell et al. 1987) indicates that the him-1 gene is required for some other essential function in addition to its role in meiosis.

2. Mutations Affecting All Chromosomes

Mutants defective for the segregation of the autosomes as well as the X chromosomes have been identified in either of two ways: (1) as him mutants that also produce a high frequency of inviable aneuploid zygotes (Hodgkin et al. 1979; A.M. Villeneuve, unpubl.) or (2) as apparent maternal-effect embryonic lethal mutants that produce a low frequency of anatomically normal, fertile adult survivors, many of which are male, that by chance received a euploid (or near euploid) chromosome complement (Kemphues et al. 1988a; J. Ahringer, pers. comm.). These include mutants defective in him-2 , him-3 , him-6 , emb-26 , and him-14 (Hodgkin et al. 1979; Kemphues et al. 1988a; A.M. Villeneuve, unpubl.), as well as at least ten new complementation groups identified by directly screening for meiotic mutants (A.M. Villeneuve, unpubl.).

In the vast majority of these mutants, high nondisjunction apparently results from a failure to form crossovers between homologous chromosomes. Cytological analysis of oocyte chromosomes revealed a high frequency of achiasmate chromosomes at diakinesis in most of the mutants examined (K. Kemphues, pers. comm.; D.G. Albertson; A.M. Villeneuve; both unpubl.). Furthermore, in all cases tested, this high frequency of achiasmate chromosomes correlates with a substantial reduction in crossover frequencies measured in genetic mapping experiments (McKim and Rose 1994; Zetka and Rose 1995b; K. Kemphues, pers. comm.; A.M. Villeneuve, unpubl.). This indicates that the absence of chiasmata late in meiotic prophase is likely due to defects in chiasma formation rather than chiasma maintenance.

An altered distribution of a reduced number of crossovers was detected in strains carrying presumptive partial loss-of-function mutations in several of these genes, him-3 , him-6 , him-14 , and him(me9) (McKim 1990; Zetka and Rose 1995b; L. Y.p and A. M. Villeneuve, unpubl.). Thus, in every case examined in sufficient detail, C. elegans mutations that cause a reduction in the overall frequency of crossovers also cause an altered distribution of the residual crossover events. This observation suggests that screens for C. elegans mutants that produce viable gametes with abnormal chromosome complements may preferentially yield mutations in genes involved in regulating the formation of crossovers (e.g., through effects on pairing and alignment of homologous chromosomes, chromosome architecture, or crossover interference) rather than in genes encoding components of the recombination machinery itself. This apparent bias might reflect a requirement for recombination enzymes in essential functions such as DNA replication or repair. Alternatively, some recombination defects may prevent the formation of functional gametes, perhaps by triggering arrest of the cell cycle in response to unresolved or defective recombination intermediates. Another possible interpretation of the altered distribution of residual crossovers is that two (or more) separate pathways exist for meiotic recombination and that a given mutation affects only one of the pathways.

Copyright © 1997, Cold Spring Harbor Laboratory Press.
Bookshelf ID: NBK20188

Views

  • PubReader
  • Print View
  • Cite this Page

Related Items in Bookshelf

Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...