Induction of indora expression in pole cells by the mesoderm is required for female germ-line development in Drosophila melanogaster

Germ-line development involves the interaction between germline cells and adjacent somatic cells. Studies in a number of organisms have shown that mesodermal cells of the adult ovary and testis support the differentiation of germ-line cells into mature gametes (Kimble and White, 1981; Tres and Kierszenbaum, 1983; Abe, 1988; Lin and Spradling, 1993; Henderson et al., 1994; Tax et al., 1994; Miura et al, 1996). Despite the importance of the somatic cells to gametogenesis, the role of these cells in germ-line development during embryogenesis is not well understood. In many animal groups, the germ line is proposed to be specified autonomously by maternal factors localized in the germ plasm, a histologically distinct region of the egg cytoplasm (Beams and Kessel, 1974; Eddy, 1975). Experimental studies in Xenopus and Drosophila have demonstrated that factors capable of directing germ-line development are localized in the germ plasm (Illmensee and Mahowald, 1974; Ikenishi et al., 1986). However, germ plasmbearing cells do not necessarily differentiate into functional germ cells. Wylie et al. (1985) reported that in Xenopus embryos, germ plasm-bearing cells isolated during their migration toward the genital ridge are unable to become germ line when implanted into ectopic sites. Thus, germ-line development is not simply the consequence of autonomous function of germ plasm, but is also influenced by cellular environment (Wylie et al., 1985).

In Drosophila, the germ plasm is localized in the posterior pole region of early embryos and partitioned into the germline progenitors, or pole cells (Mahowald, 1962, 1968; Underwood et al., 1980; Technau and Campos-Ortega, 1986; Hay et al., 1988). During gastrulation, the pole cells move along the dorsal surface of the embryos, along with the posterior midgut rudiment. They then migrate through the midgut epithelium toward the overlying mesodermal layer, where they associate with the dorsolateral mesoderm in parasegments (PS) 10-12 (Boyle and DiNardo, 1995; Boyle et al., 1997; Broihier et al., 1998; Moore et al., 1998a). The dorsolateral mesoderm in PS 10-12 is specified as somatic gonadal precursors (SGPs), and they coalesce with the associating pole cells to form the embryonic gonads (Boyle and DiNardo, 1995; Boyle et al., 1997; Moore et al., 1998a). The physical proximity of pole cells to the mesodermal cells raises the question as to whether germ-line development could be partly controlled by the mesodermal cells. Indeed, several mutations which abolish the gonadal mesoderm affect pole cell migration (Broihier et al., 1998; Moore et al., 1998a, b). For example, in tinman (tin), zinc finger homeodomain protein-1 (zfh-1) double-mutant embryos where the dorsolateral mesoderm is ablated, pole cells pass through the midgut epithelium, but subsequently they are dispersed around the midgut (Broihier et al., 1998; Moore et al., 1998b). Furthermore, in abdominal-A (abd-A) mutants which fail to specify the dorsolateral mesoderm into SGPs, pole cells associate with the dorsolateral mesoderm, but are later released from the mesoderm and disperse throughout the embryo (Brookman et al., 1992; Cumberledge et al., 1992; Boyle and DiNardo, 1995; Greig and Akam, 1995; Broihier et al., 1998; Moore et al., 1998b). These mesodermal cells may induce gene expression in pole cells resulting in their development into functional germ cells. However, such downstream genes have remained elusive.

Here, we describe the identification of a novel transcript called indora (idr), which is expressed in pole cells within the embryonic gonads. Antisense experiments, which reduce the idr mRNA in embryos, reveal that idr is required for germ-line development in females. Furthermore, we present evidence showing that idr expression in pole cells is induced nonautonomously by the dorsolateral mesoderm, although it does not require the specification of the dorsolateral mesoderm as SGPs. These data strongly suggest that the induction of idr expression in pole cells by the mesodermal cells is essential for proper development of the germ line.

The wild-type strain used was Oregon R. The tudor (tud), abd-A mutant and snail (sna), twist (twi) double-mutant stocks were kindly provided by the Bloomington Drosophila Stock Center. The wunen (wun), infra-abdominal 4 (iab-4) and tin, zfh-1 double-mutant stocks were kindly provided by Drs Ken Howard, Shigeru Sakonju and Ruth Lehmann, respectively. Homozygous mutant embryos were collected from balanced heterozygous stocks: abd-A^m1/TM3 flies; cn¹sna¹⁸twi³bw¹sp¹/CyO flies; Df(2R)wun^GL/SM6 Roi flies; iab-4³⁰²/TM3 flies; tinδ^GC14zfh-1^75.26/TM3 Sb flies. Stocks were maintained on a standard medium at 25°C (except for tud mutant stock, which was maintained at 18°C).

Total RNA was isolated from embryos (at 0-24 hours after egg laying; AEL) from tud^wc/tud^wc and tud^wc/CyO mothers and used as templates for mRNA differential display reactions performed as described previously (Liang and Pardee, 1992; Nakamura et al., 1996). Amplified products were separated on 5% acrylamide gels containing 7 M urea. The bands of interest were cut from the gels. Following re-amplification, the fragments were cloned into pBluescript KS.

A Drosophila cDNA library from embryos at 12-24 hours AEL (Brown and Kafatos, 1988) was screened, using the cDNA fragments obtained from the differential display screen as a probe. Two independent cDNAs (1.5 and 0.6 kb cDNA) were isolated and sequenced using a T7 Sequencing Kit (Pharmacia). Database searches were done using the BLAST program (Altschul et al., 1990).

Northern blot hybridization was done according to Sambrook et al. (1989). Total RNA was isolated from Oregon R embryos at 12-24 hours AEL, second instar larvae, male and female adults. Poly(A)⁺RNA was separated by mRNA Purification Kit (Pharmacia Biotech). Approximately 5 μg of poly(A)⁺RNA per lane was separated on a 1% agarose gel containing 2.2 M formaldehyde and transferred onto a nylon membrane (GeneScreen Plus, DuPont). The filter was hybridized with a DNA probe made from the 1.5 kb cDNA. Autoradiography was performed for 7 days at –80°C with an intensifying screen.

The 5′ ends of the shorter cDNA were extended using the RACE (Rapid Amplification of cDNA Ends) technique as described by Frohman et al. (1988). The first strands of the cDNAs were synthesized from 3 μg of poly(A)⁺ RNA extracted from 13±2 hour AEL embryos using primer 1 (5′-TGGGGTCTGCGACAAGGCGG-3′, nucleotides 1181-1200 of the 1.5 kb cDNA). The cDNAs were amplified using primer 2 (5′-CAAGGCGGAGCGCGATGGCTGC-3′, nucleotides 1167-1187) and primer 3 (5′-GCTCTAGACGCGATG-GCTGCCACACGAC-3′, nucleotides 1159-1178). The amplified cDNAs were cloned into pBluescript SK and sequenced.

For the analysis of idr RNA levels in the embryos from the AS-idr transformant lines and w– control line, total RNA was extracted from 11±1 hour AEL embryos with or without heat treatment during 5±1 to 11±1 hours AEL, using the acid guanidine-phenol-chloroform method (Chomczynski and Sacchi, 1987). Then, 3 μg of poly(A)⁺ RNA was loaded per lane. The RNA was electrophoresed on an agarose gel containing formaldehyde, and blotted onto a GeneScreen Plus filter (DuPont). The filter was hybridized with a DNA probe corresponding to nucleotides 351-759 of the 1.5 kb full length cDNA. The filter was reprobed with a 0.2 kb rp49 cDNA as a loading control (O’Cornell and Rosbash, 1984). Signal intensity was determined using BAS 2000 (Fujix).

In situ hybridization to whole-mount embryos was carried out according to the method described by Tautz and Pfeifle (1989) with modification as described by Kobayashi et al. (1998). Digoxigenin (DIG)-labeled RNA probes were synthesized with T7 or T3 RNA polymerases in the presence of digoxigenin-UTP (Boehringer-Mannheim), using the 1.5 kb cDNA as a template. Following in situ hybridization, immunoperoxidase staining with an antibody against Vasa (a gift from Drs A. Nakamura and P. F. Lasko) was carried out as described previously (Simpson-Brose et al., 1994). The stained embryos were observed under a compound microscope (DMRB, Leica) equipped with Nomarski optics and were staged according to Campos-Ortega and Hartenstein (1985).

In situ hybridization to polytene chromosomes was performed as described by Engels et al. (1986). Biotinylated probe was synthesized from the 1.5 kb cDNA using a Random Primed DNA Labeling kit (Boehringer Mannheim Biochemica).

The fragment corresponding to nucleotides 974-1507 of the 1.5 kb idr cDNA was inserted between the EcoRI and XbaI sites of pCaSpeRhs (Thummel and Pirrotta, 1992) to generate AS-idr, which expresses an antisense idr RNA under the control of the hsp70 promoter. AS-idr was coinjected into w– embryos with a helper plasmid, pπ25.7wc. Independent w⁺ transformants were inbred to establish homozygous stocks. As control flies, we used w– flies without AS-idr.

Transformant flies and the control flies were allowed to lay eggs on a standard culturing medium in plastic vials at 25°C for 2 hours. The embryos developed to adults at 25°C. Some embryos were heattreated by putting the vials in a 29°C incubator during the assigned stages. After eclosion, the resulting adult females were cultured at 25°C for 3-5 days, and were dissected in order to observe their ovaries under a compound light microscope.

idr was identified as a gene expressed in pole cells. Using mRNA differential display technique (Liang and Pardee, 1992; Nakamura et al., 1996), we screened for RNA species that are present in wild-type embryos but absent or rare in the embryos derived from tud mutant females, which fail to form pole cells (Nakamura et al., 1996). From more than 1000 RNA species, we identified a cDNA clone corresponding to mRNA which is detected predominantly in pole cells by whole-mount in situ hybridization. This cDNA hybridized with 1.5 and 0.8 kb mRNAs on northern blots (Fig. 1A). We isolated cDNA clones corresponding to these idr mRNAs from an embryonic cDNA library (Fig. 1B). The nucleotide sequence of the 1.5 kb cDNA predicts a protein of 131 amino acids (Fig. 1B). The amino acid sequence showed no significant homology to any known proteins. The putative Idr protein is highly basic (calculated isoelectric pH is 10.1) with a calculated relative molecular mass of 16×10³. The nucleotide sequence of the cDNAs corresponding to the 0.8 kb mRNA was identical to the 3′-untranslated region of the longer cDNA (Fig. 1B). Using the RACE technique, we isolated cDNA fragments corresponding to the 5′ region of the shorter mRNA and determined their nucleotide sequences. These cDNAs initiated at around nucleotide 958 of the 1.5 kb cDNA, suggesting the posssibility that the shorter mRNA was produced by alternative promoter usage. In situ hybridization with a fragment specific to the 1.5 kb transcript (351-759 nt in Fig. 1B) revealed that the transcript was expressed in pole cells within the gonads (data not shown). We mapped the location of the gene encoding the idr transcript to 42A on the right arm of the second chromosome by in situ hybridization to polytene chromosomes.

During normal development, the expression of idr transcripts became discernible in pole cells at the embryonic stage 14, when pole cells are incorporated into the gonads (Fig. 2B). Expression persisted in pole cells until the completion of embryonic development (Fig. 2C, D). We also examined the expression of idr in testes and ovaries using wholemount in situ hybridization, but idr expression was undetectable in the adult germ line (data not shown). However, we cannot exclude the possibility that a trace amount of idr mRNAs is expressed in somatic cells as well as in germ line throughout most of the life cycle, because northern blot analysis revealed that idr transcripts were detectable from late embryogenesis to adulthood (Fig. 1A).

In order to know whether idr is required for germ-line development, we attempted to reduce idr function in embryos by antisense idr expression. We constructed a fusion gene (AS-idr) which expresses an antisense idr RNA under the control of the heat shock protein 70 (hsp70) promoter. This fusion gene was introduced into flies by P element-mediated transformation, and two independent homozygous viable lines (AS16 and AS23) were established. We found that varying percentages of females from the two transformant lines showed sterility even in the absence of heat treatment (Table 1). Approximately 2% and 6% of the females were sterile in AS16 and AS23 lines, respectively. These sterile females had bilateral agametic ovaries that lacked differentiated ovarioles and developing egg chambers. Unilateral agametic ovaries were also found in 7% and 18% of AS16 and AS23 females, respectively (Table 1). In contrast, no sterility was observed in control females lacking AS-idr. These results show that AS-idr has a moderate effect on female fertility at 25°C.

If this phenotype is caused by antisense idr expression from AS-idr, it is expected that heat treatment would cause a stronger phenotype. Indeed, the sterility was enhanced by heat treatment at 29°C during a period from embryonic stage 10 to eclosion (Table 1). Of the females from heat-treated AS23 line, 21.8% had agametic ovaries, an increase from 6.8% obtained without heat treatment. However, of the AS23 females heat-treated from stage 15 to eclosion, only 8.1% were sterile (total number of females counted; n=148). Thus, the enhancement of the female sterility was no longer discernible when heat treatment was initiated after stage 15. Similar result was obtained when AS23 females were heat-treated after embryogenesis until their eclosion; only 5.8% of the resulting females were sterile (n=69). These results strongly suggest that the heat induction of antisense idr expression during embryonic stages 10-15 has a marked effect on female fertility. Northern blot analysis revealed that antisense idr expression by heat treatment during these stages reduced idr mRNA to undetectable levels in embryos (Table 1). In the absence of heat treatment, idr mRNA remained detectable, but at a reduced level, when compared with the control embryos lacking the AS-idr construct. The extent of idr mRNA reduction in these embryos correlates well with the severity of the sterile phenotype (Table 1). Taken together, our results demonstrate that the expression of antisense idr RNA causes a reduction of idr mRNA in embryos, leading to female sterility.

We then focused on the mechanism by which idr expression is regulated in pole cells during embryogenesis. We noted in normal embryos that almost all pole cells incorporated within the gonads expressed the idr transcript. In contrast, a small fraction of pole cells which remained outside the gonads never expressed idr mRNA (Fig. 2E). This suggests that idr expression is intimately linked to gonad formation, which involves the coordinate development of pole cells and the mesodermal cells becoming somatic components of the gonads (Brookman et al., 1992; Warrior, 1994; Boyle and DiNardo, 1995; Moore et al., 1998a). This is also supported by analyses in wunen (wun) mutant embryos. In this mutant, pole cells are unable to migrate properly from the midgut toward the overlying mesodermal layer (Zhang et al., 1996a). Consequently, only a few pole cells associate with the gonadal mesoderm to form the embryonic gonads, and the remaining pole cells are dispersed throughout embryos (Fig. 3B-E) (Zhang et al., 1996a). In these embryos, idr expression was discernible in pole cells that successfully assembled within the embryonic gonads, but not in those cells that remained outside the gonads (Fig. 3C-E).

The above observation suggests that correct migration of pole cells into the mesodermal layer is necessary for idr expression. We next examined idr expression in sna, twist double-mutant embryos, which completely lack mesoderm (Leptin, 1991; Ray et al., 1991). In the absence of mesoderm, pole cells normally pass through the midgut epithelium but subsequently are dispersed around the midgut due to the failure of gonad formation (Fig. 3A; Jarglarz and Howard, 1994; Warrior, 1994). In these embryos, there was severe disruption of idr expression (Fig. 3A), as none of the pole cells in stage 14-16 embryos expressed idr. Failed expression was not caused by loss of general transcription, as zygotic transcription from the vasa gene is discernible in the dispersed pole cells (Van Doren et al., 1998; our unpublished result).

Clearly, the presence of the mesoderm is needed to induce idr expression in pole cells. In order to determine the source of the mesodermal cue, we next analyzed idr expression in the absence of the mesodermal cells that make up the gonads. The origin and development of the somatic components of the gonads have been described (Brookman et al., 1992; Boyle and DiNardo, 1995; Boyle et al., 1997; Broihier et al., 1998; Moore et al., 1998b; Riechmann et al., 1998). The SGPs are specified from the dorsolateral mesoderm within PS 10-12 at stage 11. It has been reported that in tin, zfh-1 double-mutants no dorsolateral mesoderm is formed, which results in loss of SGPs (Broihier et al., 1998; Moore et al., 1998b). In these embryos, pole cells pass through the midgut epithelium, but subsequently they are dispersed around the midgut (Broihier et al., 1998). We found that idr expression was drastically reduced in tin, zfh-1 double-mutants (Fig. 3F). This result shows the requirement of the dorsolateral mesoderm for idr expression in pole cells.

We next asked whether the specification of the dorsolateral mesoderm as SGPs is needed to induce idr expression in pole cells. To examine this, we used the abd-A and iab-4 mutations. abd-A function is required in the mesodermal cells for the specification of SGPs (Brookman et al., 1992; Cumberledge et al., 1992; Boyle and DiNardo, 1995, Boyle et al., 1997; Broithier et al., 1998). In abd-A mutant embryos, pole cells pass through the midgut wall and are normally associated with the dorsolateral mesoderm (Broithier et al., 1998). However, they do not coalesce with the pole cells to form the gonads due to their failure to be specified as SGPs. Consequently, pole cells are released from the mesoderm and scattered throughout the embryo (Boyle and DiNardo, 1995; Broithier et al., 1998). In these embryos, the dispersed pole cells expressed idr during stages 14-16 (Fig. 3G). Furthermore, we found that a regulatory mutation in the abd-A locus, iab-4 also had no deleterious effect on idr expression (data not shown). Thus, the specification of the dorsolateral mesoderm as SGPs is dispensable for idr expression.

We identified a novel transcript, idr, which is expressed only in pole cells within the gonads. We provide evidence showing that the expression of idr is required for germ-line development. Reduction of idr mRNA by expression of antisense idr resulted in the failure of pole cells to produce functional germ line in females. Moreover, our results show that the heat induction of antisense idr expression during embryonic stage 10-15 enhances the female sterility, suggesting an important role of idr expression during these stages in germ-line development. Lastly, we have found that idr expression in pole cells depends on the presence of the dorsolateral mesoderm. These observations suggest that the mesodermal cells induce idr expression in pole cells, which is required for their development into functional germ cells.

We have shown that the reduction of idr mRNA to an undetectable level in embryos leads to female sterility without any deleterious effect on the viability. In these embryos, pole cell formation and their migration to the embryonic gonads appear to be unaffected (data not shown). These observations are compatible with the finding that idr expression is initiated after the completion of pole cell migration, and is detectable predominantly in germ line. A simple explanation for the female sterility is that pole-cell differentiation within the gonads is impaired by the reduction of idr function. In this study, we could not observe any defects in male fertility (data not shown). However, this observation is based on the analysis of embryos with reduced idr function caused by antisense idr expression. Thus, it remains possible that null mutations in idr locus may reveal additional functions in male sterility.

Our results demonstrate that the idr expression in pole cells is induced by the specialized mesoderm. The disruption of the dorsolateral mesoderm by the tin, zfh-1 double mutation drastically reduces idr expression in pole cells. The dorsolateral mesoderm within PS10-12 and their derivatives, SGPs are already associated with pole cells at stage 10-11, and remain in contact with pole cells until they coalesce as the gonads at stage 14 (Boyle et al., 1997; Broihier et al., 1998; Moore et al., 1998b). Therefore, we propose that the dorsolateral mesoderm produces a cue for idr expression and their association with pole cells is needed to induce idr expression in those cells. Alternatively, it is possible that the dorsolateral mesoderm within PS 4-9 and PS 13, which develop into fat body (Moore et al., 1998b; Riechmann et al., 1998), produce a diffusible cue acting on pole cells to induce idr expression. Our observations favor the former model. In wild-type embryos, the idr expression is detected only in pole cells that are associated with SGPs within the gonads, while those cells outside the gonads never express idr. A similar situation is observed in wun mutants. In these embryos, only a small fraction of pole cells which successfully colonized the gonads properly express idr, while the pole cells dispersed throughout embryos are unable to express it, even though they are present in fat body immediately in the vicinity of the gonads.

We have shown that idr expression is independent of abd-A function. This suggests that the specification of the dorsolateral mesoderm as SGPs is dispensable for idr expression, since previous studies suggest that abd-A function is required for SGPs specification (Brookman et al., 1992; Boyle and DiNardo, 1995; Boyle et al., 1997; Moore et al., 1998b). It has been reported that even in the absence of abd-A function, pole cells normally migrate from the midgut into the dorsolateral mesoderm during stage 11 (Broihier et al., 1998; Moore et al., 1998b). The pole cells maintain this association until mid stage 12, and afterwards they disperse. Thus, initial association of pole cells with the dorsolateral mesoderm occurs in an abd-A-independent manner. This suggests that the association of pole cells with the dorsolateral mesoderm during stage 11-12 is sufficient to induce idr expression in pole cells.

It is unclear what defines the stage at which idr expression is initiated. Pole cells are already in contact with the dorsolateral mesoderm at stage 11 (Brookman et al., 1992; Boyle and DiNardo, 1995), but idr expression is undetectable until gonad coalescence at stage 14. It is unlikely that the coalescence itself defines the onset of idr expression, because pole cells normally initiated idr expression at stage 14 without being ensheathed by SGPs in abd-A mutants (Fig. 3G). Based on the recent finding that gene expression, which normally occurs in pole cells at stage 13-14, is repressed autonomously by Nanos until these stages (Kobayashi et al., 1996; Asaoka et al., 1998), we suggest that there may be a requirement for a similar repression mechanism to enable pole cells to properly express idr at the later stage.

Interactions between germ-line progenitors and the surrounding somatic cells might be a general phenomenon in animal development. For example, in the mouse embryo, the migrating primordial germ cells (PGCs) expressing c-kit receptor move along a track of mesodermal cells expressing a membrane-bound form of the Steel ligand (Matsui et al., 1990). The ligand acts as a growth factor on primordial germ cells and also a mediator of cell-cell adhesion (Matsui et al., 1990; Flanagan et al., 1991; Matsui et al., 1991). In Drosophila, the somatic signals acting on pole cells are not well understood. Recent genetic screens have identified several genes required in somatic tissues, rather than in pole cells, for correct migration of pole cells (Zhang et al., 1996a,b; Broihier et al., 1998; Moore et al., 1998a). These genes may ultimately affect the gene expression in pole cells. Interactions between pole cells and somatic cells can also be seen during sex determination of germ line. Staab et al. (1996) have reported that somatic masculinizing signals act on pole cells to induce male germ-line marker 1 (mgm1) in stage13 embryos. idr does not act in the pathway regulating sex determination as well as pole cell migration, because idr expression initiates after the completion of pole cell migration and is independent of sex (data not shown). Instead, we favor the idea that idr is required for the differentiation of pole cells within the gonads. Isolation of mutations in the idr locus, as well as the identification of the signaling molecules required for its expression, will be informative in elucidating the role of the soma and germ line interaction in Drosophila.

We thank Bloomington Drosophila Stock Center for mutant flies. Our thanks also go to Dr Ken Howard for the wun mutant, Dr Shigeru Sakonju for the iab-4 mutant, Dr Ruth Lehmann for tin, zfh-1 double mutant flies, Dr Paul Lasko for an antibody against VAS and all members of the Kobayashi laboratory for helpful comments on the manuscript. This work was supported in part by a Grant-in-aid from the Ministry of Education, Science and Culture, Japan, by Tsukuba Advanced Research Alliance-Project, by Toray Science foundation, by the Sumitomo Foundation, and by a Research Project for Future Program from Japan Society for the Promotion of Science.