INTRODUCTION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Meiosis is essential to sexual reproduction in all multicellular organisms because it is the process whereby the chromosome complement is precisely divided in half. The fusion of the two gametes at fertilization creates a complete diploid genome. Meiotic crossing-over is the most important mechanism for ensuring the proper segregation of homologous chromosomes at meiosis I (Hawley, 1988). Crossovers, and the resulting chiasmata, link and orient homologous chromosomes so that they segregate properly. Crossing-over also increases the genetic variation between progeny and parents. A failure to produce a crossover between a pair of homologous chromosomes can result in nondisjunction, and the consequent aneuploidy in most organisms causes zygotic lethality.

In Drosophila females, the process of meiotic recombination occurs within the context of a developing oocyte. The steps involved in meiotic recombination happen early, shortly after the oocyte finishes the premeiotic S-phase (Carpenter, 1979). After this stage, the oocyte begins a developmental program of growth and definition of cell polarity. Thus, it is expected that there will be a molecular link between the proteins intimately involved in meiotic recombination and others required in a regulatory role for oocyte differentiation. These regulatory processes ensure that meiotic recombination is initiated and completed within a specific time frame. Delays in this process can have disastrous consequences on development of the oocyte (Ghabrial and Schupbach, 1999).

Genetic studies have shown that the number and distribution of crossovers are tightly regulated. The "precondition defective" class of genes in Drosophila was originally defined as those that reduce crossing-over and alter the distribution of the residual crossovers (Sandler et al., 1968). mei-218 mutants are an example of this class; >90% of all crossing-over during meiosis is eliminated and the residual crossovers are abnormally distributed (Carpenter and Sandler, 1974; McKim et al., 1996). In contrast, the frequency of gene conversion is not reduced, demonstrating that the initiation of recombination (double-strand break [DSB] formation) still occurs (Carpenter, 1982, 1984). Thus, mei-218 and others in its class are required specifically to generate the crossovers from a DSB event. Previous experiments failed to detect any mitotic, zygotic, or oogenesis phenotypes or sensitivity to methyl methanesulfonate and x-ray mutagenesis in mei-218 mutants. By these criteria, mei-218 mutant defects are limited to female meiosis (Baker et al., 1976, 1978; Lutken and Baker, 1979). These results, combined with the specific requirement for crossing-over during meiosis, suggest that mei-218 encodes a meiosis-specific gene product.

We have previously described the cloning of mei-218 (McKim et al., 1996) and found that it is part of a dicistronic message with another gene, mei-217 (Liu et al., 2000). Despite the requirement of these genes for most crossover events, their sequence has not provided insights into protein function because homologues do not exist in the databases. Therefore, to learn more about the function of mei-218, we have determined the expression patterns of the RNA and protein product, and we have sequenced the homologues from two other Drosophila species. Analysis of the transcription pattern in Drosophila melanogaster shows a specific program of meiotic gene expression. Based on the drastic effects of mei-218 mutants on crossing-over, one prediction was that MEI-218 would be a nuclear protein. However, we found by immunocytochemical analyses that MEI-218 can only be detected in the cytoplasm. The protein localization patterns suggest that MEI-218 has a vital regulatory role in meiotic crossing-over.

	MATERIALS AND METHODS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Isolation of RNA, Reverse Transcriptase (RT)-PCR Analysis, and in Situ Hybridization

Total RNA was collected from dissected ovaries or testis by grinding the tissue in 50% RNA lysis buffer (0.3 M sodium acetate, 4 mM EDTA, 50 mM Tris-HCl, pH 9.0, 1% SDS)/50% acid phenol followed by two extractions in acid phenol. mRNA was isolated from dissected ovaries using the Poly(A)pure Isolation kits (Ambion, Austin, TX). RT-PCR was carried out using the single tube methodology and reagents from Invitrogen (Carlsbad, CA) or Roche Molecular Biochemicals (Summerville, NJ ). The location of primers is shown in Figure 1. Digoxygenin-labeled RNA probes for in situ hybridization were made from the linearized mei-218 cDNA clone pHA-15 using the Roche Molecular Biochemicals RNA-labeling kit and hybridized as described by Tautz and Pfeifle (1989).

View larger version (29K): [in this window] [in a new window]   Figure 1.   Developmental analysis of mei-218 expression. (A) The mei-217/mei-218 region showing the structure of the dicistronic message (Liu et al., 2000), the location of primers used in RT-PCR, and the regions used to generate antibodies. The MEI-217-coding region is gray and the MEI-218 is black. The A5 antibody was generated against a 391-amino acid peptide from the 3'-end of mei-218, whereas the PX antibody was generated against a 315-amino acid peptide from the middle of mei-218. The FLAG epitope was fused to the beginning of the mei-218-coding region. In the hsp83::mei-218FLAG constructs, the upstream region of the normal mei-218 transcript, including the mei-217-coding region, is replaced by the hsp83 promoter. For the hsp83::mei-218FLAG-SV40 experiments, the 3'-UTR was replaced by the SV40 3'-UTR. (B) RT-PCR reveals that mei-218 is transcribed in embryos, larvae, and testis, all tissues in which no mutant phenotype has previously been observed.

Generation of Polyclonal Anti-MEI-218 Antibodies

The A5 (391 amino acids) and PX (315 amino acids) fragments were subcloned from mei-218 cDNA (Figure 1) into the Novagen (Madison, WI) pET-30c or pET-30b vectors and expressed in Escherichia coli. The proteins were purified under denaturing conditions over Ni²⁺-binding columns using His-binding Ni-nitrilotriacetic acid resins (Qiagen, Valencia, CA, or Novagen). After electroelution, proteins were concentrated and used to raise antibodies in rats (Pocono Rabbit Farm & Laboratory, Canadensis, PA). Antibodies were affinity purified in 1-ml aliquots using Affi-Gel 10 agarose beads (Bio-Rad, Hercules, CA). Beads (1.5 ml) were washed with 3 ml cold, 10-mM sodium acetate pH 4.5, then they were washed two times with 5 ml of 50 mM-HEPES, pH 7.5, before overnight incubation at 4°C with 5 ml antigen protein (dialyzed into 50-mM HEPES, pH 7.5). Resin was then blocked with 100 µl 1-M ethanolamine, pH 8.0, for 1 h at room temperature. The beads were washed into the column with 50 mM HEPES, pH 7.5, and rinsed with PBST (PBS + 0.1% Tween-20) until A₂₈₀ = 0. 1 ml sera was diluted to 5 ml with PBST and incubated in a column, rocking for 2-4 h at room temperature. The column was washed several times with PBST until A₂₈₀ = 0. Antibody was eluted with 0.2 M glycine, pH 2.7, in 800-µL fractions, into 200 µL of 1 M Tris, pH 8.0.

Transgenic Constructs

The hsp83::mei-218 construct was originally described by McKim et al. (1996). For the FLAG-tagged version (hsp83::mei-218^FLAG), PCR was used to introduce an EcoRI site in place of the mei-218 ATG. The plasmid pFLAG83 is a derivative of pBluescript containing the hsp83 promoter upstream of a sequence encoding an initiator ATG and the FLAG tag (MDYKDDDDK). The mei-218-coding region was fused in-frame to the FLAG tag using the EcoRI site introduced by PCR. The entire construct was cloned with KpnI and NotI into pCasper4 for transformation into Drosophila (Rubin and Spradling, 1982). A derivative with the simian virus (SV)40 3'-untranslated region (UTR) (hsp83::mei-218^FLAG-SV40) was constructed using PCR to introduce an XbaI site after the mei-218 stop codon and removing the 3'-UTR in the process. The SV40 3'-UTR was cloned out of pCasPeR-AUG--Gal using XbaI - SalI and cloned downstream of mei-218. As before, the entire construct was cloned with KpnI and NotI into pCasPeR 4.

The X-chromosome nondisjunction frequency in each transgenic was determined by crossing mei-218⁶; hsp83::mei-218^FLAG females to C(1:Y), v f B males.

Confocal Microscopy

A fixation method based on buffer A (Belmont et al., 1989) was used most often and is briefly summarized below. Ovaries were dissected from 1-d-old virgins in 1× Robbs solution and fixed for 10 min in 500 µL of buffer A (150 mM piperazine-N,N'-bis[2-ethanesulfonic acid], pH 7.4, 0.8 M KCl, 200 mM NaCl, 20 mM EDTA, 5 mM ethyleneglycol-bis(-aminoethyl ether)-N,N'-tetraacetic acid, final pH 7.0) plus 4% formaldehyde. Ovaries were washed for 15 min each: two times in BAT (buffer A + 0.1% Triton X-100) and two times in BAT-NGS (BAT + 10% 60 mg/ml normal goat serum). Ovaries were incubated in primary antibody (1:50 or 1:100 anti-MEI-218, 1:20 or 1:50 M5 anti-FLAG monoclonal [Kodak IBI, New Haven, CT], 1:30 anti-ORB 6H4 and 4H8 [Lantz et al., 1994] or 1:50 nuclear lamin) in BAT-NGS at 4°C overnight. Ovaries were then washed with four changes over 2 h in BAT plus 0.2% bovine serum albumin, followed by a 30-min wash in BAT-NGS while the secondary (all 1:250 dilutions; horse anti-mouse fluorescein isothiocyanate [Vector Laboratories, Burlingame, CA] for anti-FLAG and anti-ORB, goat anti-rat CY3 [Amersham Pharmacia Biotech, Piscataway, NJ] or goat anti-rat fluorescein isothiocyanate [Vector] for anti-MEI-218) was preabsorbed against methanol-fixed embryos. Secondary antibodies were added with RNase if necessary (final concentration 5 µg/ml) and left to incubate at room temperature for 4 h, followed by a 30-min wash in BAT. Ovaries were left in 1× buffer A overnight at 4°C. If tertiary labeling was used, the incubation with the secondary antibody (10 µg of biotinylated anti-rat [Vector ] ) was followed by four 30-min BAT-bovine serum albumin washes and a 30-min BAT-NGS wash. The tertiary antibody, 1:500 strepavidin-Cy3, was left to incubate overnight at 4°C. The ovaries were then washed four times over 1 h with BAT before DNA stain was added. Ovary chromosomes were stained with Hoechst, 4,6-diamidino-2-phenylindole, Yo-Pro (Molecular Probes, Eugene, OR) or propidium iodide, washed for 15 min in BAT, quickly rinsed in 1× buffer A, and mounted in Vectashield (Vector). Images were collected on an LSM 510 (Zeiss, Thornwood, NY) or TCS SP (Leica, Wetzlar, Germany) confocal microscope.

Cloning and Sequencing mei-218 from Other Species

Genomic phage libraries were screened with either a probe for mei-218 (Drosophila yakuba) or the neighboring gene RpS5 (Drosophila virilis). Previous attempts to use an mei-218 clone to probe the D. virilis library were unsuccessful because, at low stringency, clones carrying repetitive sequences were isolated that did not contain the mei-218 gene. A probe for RpS5 was used because this gene is highly conserved and adjacent to the mei-218-coding region (McKim et al., 1996). Subclones containing mei-218 sequences were identified by Southern blot of phage DNA and then sequenced using the University of Medicine and Dentistry of New Jersey core sequencing facility. The sequence was analyzed using the Wisconsin Package, version 10.0 (Genetics Computer Group, Madison, WI). The accession number for the mei-217/mei-218 D. virilis sequence is AF426408 and for D. yakuba is AF426409.

	RESULTS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Overview of Oocyte Development and Meiosis

The Drosophila ovariole is composed of two prominent regions: the germarium and the vitellarium. The germarium comprises stem cells that give rise to cystoblasts that undergo four incomplete divisions to produce a 16-cell cyst in which all of the cells are connected through ring canals (Figure 3; de Cuevas et al., 1997). In the 16-cell cyst, there are two cells that have four ring canals, whereas the other 12 cells have three, two, or one ring canal(s). Both of the four-ring canal cells form complete synaptonemal complexes (SC) and recombination nodules (RN). Both SC and RN are cytological markers that the cell is in the pachytene stage of meiosis, during which it is believed that meiotic recombination occurs. However, only one of these cells is maintained in meiosis as the oocyte, whereas the other 15 become polyploid nurse cells. The pro-oocyte, the four-ring canal cell that maintains meiosis, moves to the posterior end of the 16-cell cyst as it continues out of the germarium. The vitellarium consists of the successive stages of oocyte development leading up to the mature stage 14 oocyte, at which meiosis arrests in metaphase.

mei-218 Transcription Is Not Limited to the Drosophila Ovary

Meiotic prophase occurs in the germarium; thus, it is the region where we would expect meiotic recombination genes to be expressed (Carpenter, 1979). We used in situ hybridization to examine the expression pattern of mei-218 transcription in the ovary using an antisense RNA probe. In the germarium, RNA expression is not localized to any one cell, although it is highly enriched in regions 2 and 3 where the oocyte is in meiotic prophase. After this stage there is a rapid reduction of transcript, beginning in stage 2 of the vitellarium (Figure 2). This pattern seems to reflect a program of meiosis-specific gene expression.

View larger version (131K): [in this window] [in a new window]   Figure 2.   Transcription of mei-218 in wild-type, egl, and orb mutants by in situ hybridization using antisense mei-218 RNA probes. In all cases, a sense probe failed to detect a signal. In each image the anterior end of the ovariole or testis is to the left. (A) In wild-type ovaries, mei-218 RNA appears at a high concentration in the germarium and then is reduced in abundance early in the vitellarium. Whereas the induction of transcription is consistent, the reduction in the vitellarium is variable between ovarioles. Localization to the vitellarium oocyte (arrow) is sometimes visible. (B) mei-218 transcript in the testis. Unlike oogenesis, the transcript does not show a preference for any particular stage in spermatogenesis. (C) In egl mutant ovaries, the transcripts appear at approximately the same time as in wild type but persist longer in the vitellarium. (D) In orbF343 mutant ovaries, the mei-218 transcript was detected at a much lower level. This image is at a lower magnification to show low levels of RNA in at least four ovarioles.

Despite its meiosis-specific mutant phenotype, mei-218 is transcribed in all tissues that we have studied. RT-PCR analysis has shown that mei-218 RNA is expressed in tissues in which no mutant defects in phenotype have been observed, including embryos, larvae, and testis (Figure 1). Although the male products appear larger, sequencing of one of these products failed to detect a difference. Consistent with these results, we have observed mei-218 RNA by in situ hybridization in testis (Figure 2). These transcripts are also translated; in the testis we have detected MEI-218 protein using antibodies to the native protein (E.A. Manheim and K.S. McKim, unpublished results).

Localization of the mei-218 Protein within Germarial 16-Cell Cysts

To analyze the mei-218 protein expression patterns, we used three different antibodies: two antibodies raised against E. coli-expressed MEI-218 fragments and the M5 antibody that recognizes the protein from a FLAG epitope-tagged transgene (Figure 1). The epitope-tagged constructs of mei-218 were generated using an hsp83 promoter to express a full-length mei-218 4.2-kb cDNA fused at the amino terminus to the FLAG epitope (hsp83::mei-218^FLAG). In all experiments, expression from the hsp83 constructs was observed without heat shock because transcription from this promoter occurs at a high level without heat shock in the ovary (Ding et al., 1993). We were able to detect the epitope-tagged protein expressed from the hsp83 promoter but not the endogenous protein, by Western blotting using either anti-FLAG or anti-MEI-218 antibodies (E.A. Manheim and K.S. McKim, unpublished results). The cause of this difference can be attributed to the different promoters used. The native protein is limited to the germarium and early vitellarium, whereas the hsp83-driven FLAG-tagged protein was expressed at high levels in the germarium and most vitellarium stages. The inability to detect native MEI-218 on a western blot correlates with the notion that small amounts of protein are necessary for wild-type function, and it is subjected to a tightly restricted period of expression.

We performed additional controls by immunofluorescence to ensure that the affinity-purified anti-MEI-218 antibodies were specific to MEI-218. First, the preimmune sera did not recognize MEI-218. Second, the signal from the purified antibody was either reduced or eliminated by incubation with an excess of purified PX antigen (10- or 100-fold molar excess, respectively) but not by a 100-fold molar excess of an unrelated protein also expressed in bacteria and containing the same 6× HIS tag. Furthermore, the same signal was seen with antibodies raised against two different MEI-218 antigens (Figure 3). Finally, qualitatively similar results were obtained using the epitope-tagged version of MEI-218.

View larger version (72K): [in this window] [in a new window]   Figure 3.   MEI-218 protein is in the cytoplasm of the germarium and early vitellarium stage oocytes. In all images the tip of the germarium is pointed toward the top left. (A) Schematic of the germarium, showing the stages of meiosis relative to the stages of cyst development. The green shading represents ORB, and the SC in the four-ring canal cells is shown in red. (B) Low magnification view of a wild-type female ovariole showing MEI-218 (red) detected with the PX antibody relative to the chromosomes (blue) and ORB (green) localization. The germarium region 2b and 3 cysts and the first vitellarium cyst, stage 2, are labeled. This is one of the examples where MEI-218 shows distinct localization to the oocyte in stage 3 and 4 cysts. In other ovarioles, however, the protein gradually disappears in early vitellarium stages without the transient oocyte localization. (C) Similar MEI-218 staining was observed using the A5 antibody. Only the MEI-218 staining is shown. (D) MEI-218 (red) does not overlap with the nuclear lamin (green), showing that it is outside the nucleus. This is a higher magnification view of regions 2 and 3 of a hsp83::mei-218FLAG germarium. Because the PX antibody was used, both the endogenous and transgene protein are observed. Localization of the MEI-218FLAG to the oocyte is visible in region 3 (arrow).

We also examined 12 mei-218 mutants for effects on protein levels or localization. Using the anti-MEI-218 antibodies, we looked at alleles 1, 4, 5, 6, 6-7, 8, hfnd, g2, g4, g9, j1, and j2 and found no significant differences in expression or staining pattern from the wild type. These alleles have not been molecularly characterized, but because none of these mutations result in a detectable loss of MEI-218, they are likely not protein-null mutations.

Native MEI-218 Demonstrates a Distinct Pattern of Localization within the Cytoplasm The most striking feature of MEI-218 localization was that it was entirely cytoplasmic (Figure 3). In most experiments MEI-218 was observed at a low but detectable level in region 1 of the germarium. Higher levels of MEI-218 were observed in region 2a and accumulated to their highest levels in regions 2b and 3. This increase in MEI-218 corresponds well with the stage at which the oocyte enters meiosis, region 2a. Rather than clear localization to a single cell, MEI-218 was found in the cytoplasm of several cells within the 16-cell cyst, with more protein in some cells than others. Double-labeling experiments with an antibody to nuclear lamin (Harel et al., 1989) revealed MEI-218 clearly adjacent to, but not overlapping, the lamin (Figure 3). MEI-218 appeared punctate and was not uniform within the cytoplasm, perhaps reflecting accumulations in subcellular compartments. In most of our experiments with the MEI-218 antibodies, weak staining was observed in region 1, suggesting that mei-218 is expressed at a low level in premeiotic cells. This staining also was reduced or abolished by incubation with excess antigen.

View larger version (40K): [in this window] [in a new window]   Figure 4.   MEI-218 localization relative to the development of the germarium; in each image the anterior end of the germarium is toward the top left, and the bottom panels show only the MEI-218 staining. (A) ORB (green), MEI-218 (red), and DNA (blue) localization in wild-type ovaries (orange color reflects protein overlap). In wild type, both ORB and MEI-218 are visible in region 2a. Later, MEI-218 does not have the same oocyte specificity as ORB. (B) MEI-218 (red) relative to cortical actin (green) and DNA (blue), demonstrating the localization of MEI-218 relative to the cell boundaries.

MEI-218^FLAG Shows a Distinct Cytoplasmic Staining Pattern Genetic experiments assaying nondisjunction frequencies were performed to test whether the hsp83::mei-218^FLAG constructs produced functional protein. All transgenic lines tested rescued the mei-218 nondisjunction phenotype (Table 1). Similar transgenic constructs that lacked the FLAG epitope tag also reduced nondisjunction to wild-type levels. Furthermore, we were able to visualize MEI-218 in the ovaries of the hsp83::mei-218^FLAG transgenic flies using the anti-FLAG M5 antibody, as well as anti-MEI-218 antibodies, further demonstrating that the transgenics produced functional protein.

View larger version (41K): [in this window] [in a new window]   Figure 5.   Localization of MEI-218 expressed from the hsp83::mei-218FLAG transgene is qualitatively similar to wild type but with specific localization to the oocyte. The anterior ends of the germaria are pointed toward the top left of the frame, and the bottom row shows only the MEI-218 staining. (A) MEI-218FLAG was detected using the M5 antibody (red) and DNA stain (blue). The arrow is pointing to localized staining in the cell most likely to be the presumptive oocyte in region 2a. (B) hsp83::mei-218+ germarium stained with the PX antibody to MEI-218 (red). In this section oocyte-specific staining is visible early in region 2 (arrow). The endogenous MEI-218 is difficult to observe because the genetic background expresses MEI-218 at a low level (see RESULTS and Figure 6). (C) A volume projection of an eglPV27; hsp83::mei-218FLAG germarium showing uniform MEI-218 staining.

mei-218 Protein Expression Levels Are Sensitive to Genetic Background

MEI-218 in egl and BicD Mutants The cytoplasmic nature of MEI-218 led us to investigate whether genes that have a role in germ line differentiation regulate its expression and localization. The egl and BicD genes are required to localize many factors to the oocyte and restrict oocyte development to a single cell (Theurkauf, 1994). In both egl^PV27 and BicD^R26 females, mei-218 was transcribed at a higher level than wild type and persisted at this high level into the vitellarium stages (Figure 2). Thus, as compared with wild type, in these mutant backgrounds mei-218 was turned on normally, but the RNA expression continued much later in oocyte development.

View larger version (64K): [in this window] [in a new window]   Figure 6.   Mutational analysis of MEI-218 expression. The anterior end of each germaria is toward the top left of the frame. (A) The MEI-218 staining pattern in an orbF343/TM3 germarium, an example of a stock with low protein levels detected by tertiary labeling. (B) In an orbF343 homozygote, variably low MEI-218 protein was observed but with clearly less localization to single cells. (C) Expression of MEI-218 in hsp83::mei-218FLAG-SV40 transgenic 6A-2, which contains the SV40 3'-UTR. MEI-218 (red) was detected using the M5 anti-FLAG antibody to distinguish the artificially expressed protein from the native MEI-218 (DNA is stained blue). This line rescues the mei-218 mutant nondisjunction phenotype and shows oocyte localization in the later stages of the germarium (anterior end toward the top left of frame). (D) The same image showing only the MEI-218 channel.

mei-218 expression in orb mutants The ORB protein is necessary to promote formation of the 16-cell cyst from the eight-cell stage as well as oocyte differentiation (Lantz et al., 1994). We studied the weaker orb^F303 and stronger orb^F343 alleles to ascertain whether the ORB protein was necessary for proper transcription, translation, or localization of MEI-218. Even although a 16-cell cyst is not formed in orb^F343 ovaries, by in situ hybridization a low level of mei-218 RNA was detected (Figure 2). In orb^F303 ovaries, in which the 16-cell cyst is formed with an abnormal oocyte that is located in the center of the cyst instead of at the posterior end, there were high levels of mei-218 transcript qualitatively similar to wild type. RT-PCR analysis confirmed that mei-218 mRNA was present in both alleles, and consistent with the in situ hybridization results, lower amounts of RNA were present in orb^F343 as compared with wild type or orb^F303 (E.A. Manheim and K.S. McKim, unpublished results). These results demonstrated that the generation of the 16-cell cyst, but not necessarily normal oocyte differentiation, is required for high levels of mei-218 transcription.

The Role of the 3'-UTR in mei-218 Expression

Several genes involved in early oocyte development and embryogenesis require their 3'-UTR for proper translation and localization (Lantz et al., 1994; St. Johnston, 1995). To investigate the relationship between the 3'-UTR, oocyte localization, and protein function, we generated a derivative of the hsp83::mei-218^FLAG in which the endogenous 3'-UTR was replaced with that from SV40 (hsp83::mei-218^FLAG-SV40). Unlike the hsp83::mei-218^FLAG transgenes, these transgenic constructs showed variable results in genetic rescue experiments. In some lines there was complete rescue of the mei-218 nondisjunction phenotype, whereas in other lines there was partial or no rescue at all (Table 1).

Immunofluorescence analysis of the hsp83::mei-218^FLAG-SV40 transformants that had full rescue of the mutant phenotype showed protein staining similar to the original hsp83 transgenes with the endogenous 3'-UTR (Figure 6). MEI-218 appeared in region 2 with enrichment in the pro-oocyte in region 2B; this oocyte localization continued into later stages of the vitellarium. On the other hand, in the transformants that showed a complete lack of rescue, there was very little MEI-218^FLAG-SV40 present, and it was difficult to discern if the protein localized. These results show that the MEI-218 protein has an inherent ability to localize to the oocyte. Therefore, the 3'-UTR may only be important for efficient translation of mei-218. Furthermore, position effects were probably responsible for the differences in rescue of the mutant phenotype. Although the hsp83::mei-218^FLAG transgenes with the endogenous 3'-UTR were also susceptible to position effects, as shown by protein levels assayed by western blot (data not shown), even in the hsp83::mei-218^FLAG line expressing the least amount of protein, there was enough to detect by immunofluorescence and rescue mei-218 mutants.

Sequence of mei-218 in D. virilis and D. yakuba

No known homologues of mei-218 have been found in any of the sequence databases. We have cloned and sequenced mei-218 from D. yakuba and D. virilis to study the evolutionary conservation of the gene and to identify the important domains for function. D. yakuba diverged from D. melanogaster ~10-15 million years ago, and D. virilis so diverged 40-60 million years ago (Powell and DeSalle, 1995). mei-218 is situated ~2 kb from the end of RpS5 (Figure 1), a highly conserved ribosomal protein. Using probes to both mei-218 and RpS5, phage clones containing mei-218 from D. yakuba and D. virilis were isolated. The use of the RpS5 probe was necessary to screen out false positives picked up by the mei-218 probe detected at low stringency. We obtained the complete sequence from the D. virilis gene and a partial sequence from D. yakuba.

With the complete sequence of the D. virilis gene, we could determine that the structural features of this locus are conserved (Figure 1). The dicistronic nature of mei-217 and mei-218 was conserved in D. virilis. In fact, the two proteins overlap in both species, and the distance from the mei-218 ATG to the stop codon of mei-217 is 20 bp in both species. There are eight introns located at the same positions, and two of these (nos. 2 and 3) are larger than the rest in both species (>100 bp). There are also conserved sequences within the first intron that may be involved in posttranscriptional regulation.

D. virilis MEI-217 shows high similarity (72.25%) and identity (63.87%) to the D. melanogaster protein, including sequences upstream of the MEI-217 ATG we had originally predicted (Liu et al., 2000). Based on the sequence of RT-PCR products and conservation in D. virilis, it now appears that the D. melanogaster MEI-217 open reading frame begins with an ATG 250 bp upstream of the previously predicted start codon. This is possible because of a splicing event between the two ATGs. One implication of these findings is that the mei-217^r1 mutation, originally thought to be in the 5'-UTR, is actually in the coding sequence and changes a Lys residue to a stop codon. A nonsense mutation was unexpected because this allele is a hypomorph. We also compared the promoter and 3'-UTR sequences in an effort to find regulatory sequences. The promoter region, defined as the sequences between the coding regions of mei-217 and RpS5, did not have significant homology. Only weak similarity was found in the 3'-UTR regions, and it was far less extensive than that observed for some other germ line transcripts such as nanos (Gavis et al., 1996).

The amino acid alignment corresponds approximately to visible domains in the mei-218 protein. Although not conserved and lacking known motifs, the protein can be divided into three domains: a basic amino-terminal domain, a central acidic domain, and a carboxyl-terminal hydrophobic domain (Figure 7). It was not possible to accurately align the amino acids of the basic domain of MEI-218 with the corresponding D. virilis sequence because of the poor homology. In addition, the D. virilis sequence is 191 amino acids shorter than D. melanogaster. However, the chemical properties of this region are conserved: in both proteins this region is basic, hydrophilic, and rich with Gln and Ser. In mei-218, the motif (Q)_2-5K is repeated six times and the D. virilis sequence is also Q-rich but with a different structure of repeats. These similarities show that the lack of conservation does not mean that this region is not important but signifies that the region's crucial features are chemical properties rather than the sequence of amino acids.

View larger version (15K): [in this window] [in a new window]   Figure 7.   Schematic showing the alignment of the mei-218 region in the three Drosophila species. The percentage of identity or similarity reflects a comparison of the species to D. melanogaster only. The D. yakuba sequence is not complete, and therefore, the amino acid coordinates are not shown.

The central region of both proteins again is not highly conserved but is slightly acidic and hydrophilic. The only highly conserved region is the more hydrophobic C-terminal region with 71.8% similarity and 63.8% identity. Similarly, this same pattern of conservation was found in D. yakuba with a more highly conserved C-terminal domain than the central domain.

	DISCUSSION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Meiosis in Drosophila occurs within a complex developmental context. The occurrence of meiosis at a specific time within a unique tissue implies that the genes required for meiotic recombination and oocyte differentiation are regulated by the same factors. Our observations of MEI-218 support this view and suggest that both transcriptional and posttranscriptional modes of regulation are important for meiotic recombination genes.

Even though mei-218 has a critical role in meiosis (it is required for >90% of meiotic crossing-over), the protein sequence has evolved rapidly. Our sequence analysis suggests that there are three separate domains within MEI-218 diverging at different rates. The sequences from D. yakuba and D. virilis show that the C terminus of MEI-218 is under selection within the genus of Drosophila. The other domains of mei-218, however, appear to be evolving at a rapid rate because they are less conserved and in some cases unrecognizable based on simple sequence comparisons. Begun and Whitley (2000) made a similar conclusion regarding the evolution of mei-218 based on a partial sequence of mei-218 from several isolates of Drosophila simulans.

Regulation of mei-218 RNA and Protein Expression

Transcriptional Regulation Our analysis of mei-218 shows a specific pattern of transcription in the ovary. We believe that activation of mei-218 transcription occurs when the 16-cell cyst forms. In support of this hypothesis, mei-218 transcription occurs at a very low level in a strong orb mutant that fails to make a 16-cell cyst. ORB is proposed to be a translational regulator (Lantz et al., 1992), and its effects on mei-218 expression may be indirect. For example, transcription of mei-218 and other genes required for meiotic recombination may occur at a high level only if a 16-cell cyst forms.

Posttranscriptional Regulation Although MEI-218 did not show specificity to the oocyte in the germarium in most genotypes, it did show a restricted pattern, including uneven distribution between cells in a cyst and punctate appearance within each cell. This MEI-218 localization may reflect subcellular structures such as the Golgi apparatus or endoplasmic reticulum. In addition, whereas the RNA is rapidly induced in region 2a, the protein is initially present at very low levels and then gradually accumulates to a peak in region 3. The pattern of protein accumulation followed by a rapid decline is similar to the observation of Carpenter (1979) that late RNs first appear in region 2a, accumulate to higher numbers in region 2b, and then decline in region 3. It is striking, however, that distinct enrichment to the oocyte is not required for native MEI-218 to function. A similar situation has been observed with BICD and EGL, which are required to restrict SC formation to the two four-ring canal cells before they are localized (Carpenter, 1994; Huynh and St. Johnston, 2000). Thus, it may be common that proteins required for oocyte differentiation may not demonstrate oocyte localization at the time they are first required.

Models for the Role of mei-218 in Meiotic Crossover Formation

The original DSB repair model proposed that gene conversion and crossing-over were alternative resolutions of the same intermediate whose fate depends on which strands of the Holliday junction are cut. Evidence from several organisms is inconsistent with this feature of the DSB model. Studies of a variety of organisms have shown that the alternative outcomes of the Holliday junction do not occur with equal frequency (Carpenter, 1990; Kleckner, 1996). The relative frequency of gene conversion to crossover events at the rosy locus of D. melanogaster is ~5:1 (Hilliker et al., 1988). These results suggest that the alternative outcomes, gene conversions and crossovers, result from different repair and resolution reactions. Thus, crossing-over may require a specific set of proteins to orchestrate a modified pathway of DSB repair that promotes crossover resolution of the Holliday junction but is not required for gene conversion.

The strong phenotype of mei-218 mutants, in which crossing-over but not gene conversion (Carpenter, 1982) or DSB formation (Liu et al., 2000; McKim et al., 2000) are reduced by >90%, is consistent with a meiosis-specific repair pathway for generating crossovers. We propose that crossing-over in Drosophila is the result of a mei-218-dependent meiosis-specific pathway of DSB repair and resolution. Cytoplasmic localization of MEI-218 is consistent with an indirect role in crossover formation, perhaps in the regulation of other gene products required for crossing-over. In this model, MEI-218 functions in the cytoplasm to regulate other proteins or possibly control the access of proteins involved in building late RNs and generating crossovers to the oocyte nucleus. A regulatory role is not without precedent: hypomorphic alleles of Sxl (K. Cook, personal communication; Bopp et al., 1999) and mei-P26 (Page et al., 2000), both genes with primary roles in germ line differentiation, have crossover phenotypes similar to mei-218. Therefore, it is possible that MEI-218 functions in the pathway between germ line differentiation and realization of meiotic recombination events. The lack of oogenesis defects in mei-218 mutants suggests that the position of mei-218 is after the point at which the meiotic and developmental pathways diverge.