Meiotic recombination and chromosome segregation in Drosophila females.

Key Words double-strand break, meiotic recombination, Drosophila, crossing over, oogenesis, synaptonemal complex

Abstract

In this review, we describe the pathway for generating meiotic crossovers in Drosophila melanogaster females and how these events ensure the segregation of homologous chromosomes. As appears to be common to meiosis in most organisms, recombination is initiated with a double-strand break (DSB). The interesting differences between organisms appear to be associated with what chromosomal events are required for DSBs to form. In Drosophila females, the synaptonemal complex is required for most DSB formation. The repair of these breaks requires several DSB repair genes, some of which are meiosis-specific, and defects at this stage can have effects downstream on oocyte development. This has been suggested to result from a checkpoint-like signaling between the oocyte nucleus and gene products regulating oogenesis. Crossovers result from genetically controlled modifications to the DSB repair pathway. Finally, segregation of chromosomes joined by a chiasma requires a bipolar spindle. At least two kinesin mot or proteins are required for the assembly of this bipolar spindle, and while the meiotic spindle lacks traditional centrosomes, some centrosome components are found at the spindle poles.

INTRODUCTION

Drosophila females, like the majority of sexually reproducing organisms, generate crossovers between homologous chromosomes to direct segregation at the first meiotic division. This important function of crossovers is likely related to the three other important features of meiosis. First, the number and distribution of crossovers along a chromosome is nonrandom and under genetic controls. Second, the meiotic chromosomes have a unique structure, including the synaptonemal complex (SC) that forms between homologs during meiotic prophase, and an assembly of cohesion proteins that regulate separation of sister chromatids at the two meiotic divisions. Third, the segregation of homologous chromosomes requires the formation of a bipolar spindle.

The SC is a proteinaceous structure with remarkably similar structure in a variety of organisms (Figure 1). Although the structure of the SC has been conserved, its relationship to recombination has diverged significantly. This review concentrates on recent studies that provide insight into the relationship between the SC and recombination in Drosophila melanogaster females, and the mechanisms by which chiasmata function to segregate the homologs. Since Drosophila males lack meiotic recombination, including crossing over (93) and gene conversion (30), this review focuses on meiosis in Drosophila females. We also give only scant coverage to the extensive and rich background of classical genetic studies in Drosophila, because they are elegantly described in three previous reviews (6, 73, 111).

Our goal is to review some of the recent findings that are explaining many of the classical genetic observations on the molecular and cytological level. Since the publication of a similar review on meiosis almost 10 years ago (55), several major advances have been made.

1. The SC has two roles: It is required for initiating meiotic recombination events and for the repair of DSBs into crossovers.

2. Identification and cloning of genes required for initiating meiotic recombination have shown that double-strand break formation (DSB) is a conserved mechanism.

3. Defects in DSB repair cause reductions in crossing over and abnormal oogenesis.

4. Crossing over may be regulated by controlling which mechanism of DSB repair is utilized. This could involve the generation of an intermediate that only forms in the crossing over pathway.

5. Segregation occurs on an acentrosomal spindle, although this spindle may have a defined structure or complex of proteins at the spindle poles.

THE GENETICS OF MEIOTIC RECOMBINATION AND HOMOLOG DISJUNCTION IN DROSOPHILA FEMALES

Most of the meiotic mutants in Drosophila were isolated because they increased the frequency of X-chromosome nondisjunction. Fortunately, with only four major chromosomes, D. melanogaster females are fertile and produce abundant euploid progeny even if there is random segregation of all chromosomes. The first genetic screens to isolate meiotic mutants were performed in Larry Sandler's laboratory by testing for X-chromosome nondisjunction (5, 112). More recently, genetic screens on a larger scale have been performed using P-element mutagenesis (117) or EMS mutagenesis (K. McKim, unpublished results; J. Sekelsky, unpublished results). Some of the mutants identified in these and other screens are listed in Tables 1 to 3.

There Are Multiple Pathways for Segregating Homologs

CHIASMATE SEGREGATION The canonical pathway for chromosome segregation is based on the ability of chiasmata to hold and orient homologs on the meiotic spindle (Figure 2). A single medially placed crossover appears ideal for this process. Not only is that where the majority of crossovers are located (73), but there is evidence that distally or proximally located crossovers, or the presence of multiple crossovers, may fail to ensure or even hinder segregation (67, 90).

ACHIASMATE SEGREGATION Several observations reveal the presence of additional mechanisms that influence the segregation of chromosomes. For instance, the Drospohila fourth chromosomes segregate correctly, even though they never form a chiasma. Also, heterozygosity for a single balancer, which suppresses crossing over on one chromosome, does not increase the frequency of nondisjunction. These observations are explained by the presence of systems that are responsible for the segregation of achiasmate chromosomes. Further analysis has led to a model in which there are two achiasmate segregation systems (54, 56). One achiasmate system is based on heterochromatic homology (the homologous achiasmate system). These pairings persist until the meiosis I spindle assembles and may be instrumental in the orientation of achiasmate homologs (34). In contrast, the nonhomologous achiasmate system segregates chromosomes without centromeric alignment and is based on parameters such as shape and size. Genetic assays for defects in the achiasmate system may actually facilitate the identification of spindle-associated proteins. Mutations in several genes have phenotypes that are specific to the achiasmate system (Table 3). Many of these genes may also be important for chiasmate segregation, but the achiasmate system is more sensitive to mild spindle defects.

If there is a system to segregate achiasmate homologs, why do crossoverdefective mutants cause nondisjunction? An important insight into this question comes from the observation that the relationship between the frequencies of X-chromosome nondisjunction and nonexchange tetrads is not linear. In fact, X-chromosome nondisjunction is proportional to the frequency of nonexchange tetrads cubed (6). The reason either achiasmate system fails in crossover-defective mutants is that it can accurately segregate only one pair of large chromosomes. When an achiasmate pair of X-chromosomes is present in the same oocyte as a pair of achiasmate autosomes, equal numbers of chromosomes are sent to each pole without respect for homology. If there are simultaneously two X and two 2nd chromosomes in the achiasmate pool, XX [left and right arrow] 22 segregation is just as likely as X2 [left and right arrow] X2, but XX2 [left and right arrow] 2 is less likely. Thus, when it comes to large chromosomes, the preference of the achiasm ate system is for the equal distribution of chromosome numbers. A detailed analysis of possible mechanisms for achiasmate segregation has been published by Hawley & Theurkauf (56).

FEMALE MEIOSIS IN THE DEVELOPMENTAL CONTEXT

The Oocyte Develops Alongside 15 Nurse Cells

Meiosis is one of the first programmed events in the development of the oocyte. Drosophila females have two ovaries, each comprised of several ovarioles containing chains of developing oocytes. At the anterior end of each ovariole is the germarium, where four rounds of incomplete mitotic divisions produce a 16-cell cyst with intercellular junctions termed ring canals. The germarium is divided into four regions based on the morphology of the 16-cell cysts (Figure 3, see color insert). The cysts move down the germarium as they mature such that a cyst in a more posterior position is in a later stage of meiotic prophase than a cyst in a more anterior position. Soon after the last premeiotic division (region 2a), meiotic prophase begins and recombination is initiated (21). Note that the position of a cyst within the germarium is only a rough guide to its meiotic stage (18). While the cysts are arrayed in order of developmental age, their absolute position in the germarium does not equate to a specific stage in mei otic prophase. This organization, however, makes it possible to compare oocytes at different stages of meiosis within a single germarium.

Cytological Descriptions of Meiosis in Drosophila

A series of electron microscopy studies by Carpenter laid the foundation for much of the current work on meiosis in Drosophila (18-22, 26, 27). Several cells enter meiosis in each 16-cell cyst, but only one will become the oocyte while the rest become nurse cells. Specifically, the SC first develops in the two cells with four ring canals, one of which will become the oocyte (18, 21). Later, SC develops to variable extents in the other cells that have fewer ring canals. Recombination is probably initiated in these pro-nurse cells as well (27, 76). Eventually, all cells except the oocyte exit the meiotic program and SC is maintained in only the oocyte. These studies also revealed that the SC is a dynamic structure; early in pachytene the SC progressively shortens and thickens and then towards the end of pachytene, lengthens (18, 21).

It was from Carpenter's studies that the significance of recombination nodules (RN) was first discerned (19, 21). Like most organisms, Drosophila females develop two temporally and morphologically distinct types of RN. Ellipsoidal RNs appear first and have a random distribution. Spherical RNs appear later and have a nonrandom distribution that resembles crossovers. In addition, the number of spherical RNs is reduced and their morphology is abnormal in mei-218 mutants, in which there is a specific failure in crossover formation (20). Based on these observations, it is reasonable to conclude that ellipsoidal RNs are the sites of early DSB repair while spherical RNs are those sites where a crossover will occur. Note that in Drosophila, both types of RN do not appear until after the homologs are fully synapsed (pachytene). This is later than some other organisms, such as the plants Lycopersicon esculentum, Allium cepa, and A. fistulosum, where early RNs first appear during zygotene (2, 121).

MOLECULES IN THE PATHWAY FOR EXCHANGE

Evidence that Double-Strand Breaks Initiate Meiotic Recombination

Double-strand breaks (DSB) were initially shown to be efficient initiators of recombination in meiotic cells of Saccharomyces cerevisiae (72). Extending this observation to other organisms was not trivial because of the inability to physically detect DSBs. Recent evidence from several experiments now strongly indicates that DSBs initiate meiotic recombination in Drosophila females.

mei-W68 ENCODES A spoil HOMOLOG One of the genes required for all meiotic recombination in Drosophila females, mei-W68 (85), encodes a homolog of spoil of S. cerevisiae (87). The spoil proteins, which are homologous to a novel type U topoisomerase from archaebacteria (topo6A), are proposed to be the enzymatic factor for DSB formation (62). Similarly, Spoil homologs have now been identified in many organisms, including Caenorhabditis elegans (33) and mammals (61, 109), lending support to the idea that most or all organisms initiate meiotic recombination with a DSB.

X-IRRADIATION INDUCES MEIOTIC CROSSING OVER When mei-W68 mutants were exposed to 4000R of radiation, crossing over was efficiently induced to almost 50% of the wild-type frequency, a 200-fold increase over unirradiated controls (R. Bhagat & K. McKim, unpublished). These results suggest that mei-W68 mutants lack meiotic recombination due to an absence of DSBs. …