Ovarian follicle formation in Drosophila melanogaster requires stall (stl) gene function, both within and outside the ovary, for follicle individualization, stalk cell intercalation, and oocyte localization. We have identified the stl transcript as CG3622 and confirmed the presence of three alternatively spliced isoforms, contrary to current genome annotation. Here we show that the gene is expressed in both ovarian and brain tissues, which is consistent with previous evidence of an ovary nonautonomous function. On the basis of amino acid sequence, stl encodes a metalloprotease similar to the "a disintegrin and metalloprotease with thrombospondin" (ADAMTS) family. Although stl mutant ovaries fail to maintain the branched structure of the fusome and periodically show improperly localized oocytes, stl mutants do not alter oocyte determination. Within the ovary, stl is expressed in pupal basal stalks and in adult somatic cells of the posterior germarium and the follicular poles. Genetically, stl exhibits a strong mutant interaction with Delta (Dl), and Dl mutant ovaries show altered stl expression patterns. Additionally, a previously described genetic interactor, daughterless, also modulates stl expression in the somatic ovary and may do so directly in its capacity as a basic helix-loop-helix (bHLH) transcription factor. We propose a complex model of long-range extraovarian signaling through secretion or extracellular domain shedding, together with local intraovarian protein modification, to explain the dual sites of Stl metalloprotease function in oogenesis.

---

AN emerging picture of the regulation of oogenesis in Drosophila includes multiple, diverse molecular and cellular mechanisms that take place in the ovary itself, as well as a growing number of regulatory processes that act from outside the ovary to coordinate the external/internal environmental conditions with the founding and development of the oocyte. In the ovary this process requires molecular communication between soma and germline for proper cell fate determination, adhesion, and migration and organization of follicular structure; it begins in the germarium, at the anterior end of each of the ovary's 15–20 oocyte assembly line structures (called ovarioles) (Figure 1A) (for reviews, see KING 1970; SPRADLING 1993). Here, 2–3 germline stem cells (GSCs) divide asymmetrically to produce daughter cystoblasts, while maintaining stem cells within the molecular niche. A specialized organelle, the spectrosome/fusome, anchors the GSC mitotic spindle to direct the axis of division and subsequently divides to be inherited by the cystoblast (supporting information, Figure S2). Following four rounds of mitosis with incomplete cytokinesis, the 16 germline cystocyte daughters are connected by elongated fusomes through actin-rich ring canals (LIN et al. 1994; ROPER and BROWN 2004). Of the 16 cystocytes, 1 retains the most fusome material and differentiates into the oocyte (LIN and SPRADLING 1995): Its nucleus remains diploid in preparation for meiosis. The remaining 15 cells of each germline cyst become nurse cells: Each nucleus becomes polyploid to produce sufficient nutrients for the oocyte. The germline cyst travels toward the posterior of the germarium where somatic stem cells lie laterally (NYSTUL and SPRADLING 2007). Here, a controlled number of somatic progeny encapsulate each germline cyst with an epithelial monolayer (MARGOLIS and SPRADLING 1995). The germline cyst and its somatic epithelium bud off from the germarium as a discrete follicle and are separated subsequently from the next formed egg chamber by a somatic stalk. This process of follicle individualization requires regulated somatic cell proliferation, cell fate determination, stalk cell recruitment, differential adhesion, and cell migration; many of the genes that contribute to these processes have been identified, and functions both inside and outside the ovary have been described (reviewed in BASTOCK and ST. JOHNSTON 2008; BERG 2008; GRUNTENKO and RAUSCHENBACH 2008).

View larger version (62K): In this window In a new window Download PPT slide   FIGURE 1.— The stl null phenotype. (A) Wild-type ovariole structure includes the separation of follicles by interfollicular stalks. (B) A stla16 ovariole from a young female (1 day posteclosion) lacks any individualized follicles or interfollicular stalks: Multiple germline cysts are enclosed in a single somatic epithelium. (C) A whole ovary from an older stlph57/stlpa49 female (10 days old) shows severe cell degeneration. This phenotype was observed for all homozygous and heteroallelic stl genotypes. (D) Df(2R)d02208-f07572/stla16 ovarioles show the same severe phenotype as stl point mutations, confirming their identification as null alleles. The nuclei are stained with DAPI to visualize cellular organization. In this and all subsequent figures, scale bar is 50 µm, and anterior is positioned at the top or the left of each panel.

 
Control of follicle formation within the ovary requires multiple cell signaling and adhesion pathways. For example, daughterless (da) regulates cell proliferation and apoptosis in the germarium, as well as stalk cell recruitment at the budding follicle border (SMITH et al. 2002). In the newly formed follicle, Notch (N) and Delta (Dl) induce anterior polar cell fate, and the anterior polar cells subsequently signal through the Janus kinase/signal transducer and activator of transcription (JAK/STAT) pathway to initiate stalk cell differentiation (MCGREGOR et al. 2002; TORRES et al. 2003). Further, the stalk induces the posterior polar cell fate of the younger follicle, leading to upregulation of adhesion molecules, such as DE cadherin [shotgun (shg)] and β-catenin [armadillo (arm)], in the oocyte and posterior somatic cells. Homophilic interactions and cell sorting ultimately position the oocyte to the posterior of the egg chamber (GODT and TEPASS 1998; GONZALEZ-REYES and ST. JOHNSTON 1998). Mutations in many of the genes that control these events disrupt follicle separation, resulting in the packaging of multiple germline cysts within a single somatic epithelium (RUOHOLA et al. 1991; CUMMINGS and CRONMILLER 1994; MCGREGOR et al. 2002). This phenotype is shared by stall (stl) mutants, in which multicyst ovarioles lack interfollicular stalk structures as early as pupal ovary development (BAKKEN 1973; SCHUPBACH and WIESCHAUS 1991; TWOROGER et al. 1999; SMITH et al. 2002; WILLARD et al. 2004).

While many of the ovarian regulators of oogenesis have been described in detail, less is known about the extraovarian control of this process. Hormonal input has been shown to affect oocyte production and maturation through germarial cell proliferation and cell death, follicle apoptosis, and yolk protein synthesis and uptake: Insulin, juvenile hormone (JH), and 20-hydroxyecdysone (20E) affect egg development in response to the nutritional environment of the fly (SOLLER et al. 1999; DRUMMOND-BARBOSA and SPRADLING 2001; LAFEVER and DRUMMOND-BARBOSA 2005). In addition, neural influences on follicle formation and maturation have been shown to involve the activity of the neurotransmitters, serotonin and dopamine (WILLARD et al. 2006). Finally, on the basis of mitotic clonal analysis and ovary transplantation, we identified stall as an additional important extraovarian regulator of follicle morphogenesis; however, the molecular identity of the stl gene product was then unknown (WILLARD et al. 2004).

Here, we extend our analysis of stl's control of oogenesis by further defining the stl mutant phenotype, elaborating on the gene's functional interactions, and identifying its molecular product as a metalloprotease with distinguishing similarities to a disintegrin and metalloprotease with thrombospondin (ADAMTS) domain proteins. We show that, although stl mutant ovaries fail to maintain the fusome within early germline cysts and contain a moderate number of mislocalized oocytes, stl does not alter oocyte determination. We also address a genetic interaction between stl and Dl in follicle individualization and oocyte polarity. Finally, the identification of Stl as a metalloprotease is a critical leap in the study of ovarian follicle formation. These evolutionarily conserved enzymes participate in a wide range of biological processes, and the characterization of Stall as an ADAMTS offers a new approach to considering the roles of these proteins in oogenesis.

Drosophila stocks:

Flies were maintained on molasses–cornmeal–yeast medium at 25°. Stocks used in this study are listed in Table 1.

View this table: In this window In a new window   TABLE 1 Drosophila stocks used in this study

 

Genetic and molecular analysis of stl:

Bloomington Deficiency Kits (2003) for chromosomes X, 2, and 3 were crossed to stl^a16, and ovaries were dissected from doubly heterozygous adult progeny. Ovaries were 4',6-diamidino-2-phenylindole (DAPI) stained and scored for percentage of ovarioles exhibiting follicle formation defects to identify dominant genetic interactions. To locate the stl transcription unit, meiotic recombination frequencies were determined between stl and l(2)06496 and between stl and l(2)k06908 on the basis of segregation of the P{w⁺}markers. Both stl^a16 and stl^pa49 were used to calculate the distances in map units. Male recombination was performed with insertions in l(2)rG270, CG3732, CG3875, l(2)k17002, ppa, jbug, blw, asrij, and Nop60B opposite cn stl^pa49 bw or cn stl^ph57 bw via 2,3 transposase. Deletions were produced by mobilization between Exelixis insertion pairs, P{WH}f06717 to P{XP}d06151 and P{XP}d02208 to P{WH}f07572, as screened by the loss of w⁺. Ovaries of the sterile Df/stl^a16 females were dissected and DAPI stained; fly carcasses were used for genomic DNA analysis to confirm the genotype. The specific primer pairs for genomic PCR were GACGCATGATTATCTTTTACGTGAC and ATGATTCGCAGTGGAAGGCT for the P insertion and TTGCCTTTGTTCTACGCTCTC and GCCCAAGAACACGACGATAA for the flanking genomic region. To sequence candidate genes within the deleted region, RNA was isolated from cn bw (the parental strain for stl^pa49 and stl^ph57), stl^a16, stl^pa49, stl^ph57, stl^awk26, and stl^wu40 females with Trizol (GIBCO-BRL, Gaithersburg, MD), Phase-lock gel (Eppendorf), and poly(A) selection kits (Sigma, St. Louis). cDNA was produced by reverse transcription (Promega, Madison, WI) and amplified by PCR [Invitrogen (Carlsbad, CA) Platinum Taq High Fidelity]. Amplification of CG3622 cDNA was accomplished with the forward primers GGCCGGTTGTTAATTCTTCA (isoform A), GCCACCAATGCCACAAAT (isoform B), and CAGAGAAGCCGTCATCATCA (isoform C) and the reverse primer GAACTTTCTCGTTCGCACTG (common to all isoforms). Sequencing of both strands of these PCR products was performed in overlapping fragments with the following primers: GGCCGGTTGTTAATTCTTCA (F1A), GATCGTAGTGCAGCGGATCT (R1), GCCACCAATGCCACAAAT (F1B), CTTGACGACGCCTCACTGTA (R1B), CAGAGAAGCCGTCATCATCA (F1C60), CACCGTCGTTTGGCTTTAAT (F1C), GGCCCATTCCCATCTTCT (RTBCR), GTGCTGAAGCGCCTAGAGAT (F2), GAGATCCCGACAGATTTCCA (R2), GAGCAAGATGGAAATCTGTCG (F3), GGTGGAATTATCGCAGGAGA (R3), CGCTTCTCCTGCGATAATTC (F4), CACCTGCCATACACGCAATA (R4), GGTGTGGAGAGCATCCAACT (F5), and GAACTTTCTCGTTCGCACTG (R5). To confirm the identified mutations, independently isolated genomic DNA was amplified by PCR and sequenced with the above primers as well.

5' rapid amplification of cDNA ends:

The Ambion FirstChoice RLM–rapid amplification of cDNA ends (RACE) kit was used on total RNA from white Canton-S adult flies. The resulting adaptor-ligated cDNA was amplified with kit adaptor primers and a gene-specific primer to a common exon TTGACGACGCCTCACTGTAG (3622RACEouter); a nested primer to a common exon CTCTTGTCCACCCTCGCTTGGTAGAG (3622 RACEinnerLong); and nested primers to the unique exon for each isoform AATGCCATTGTAGTGCAAGCTCCTCT (RACERCLong), CATCGGTGGCATTTGTGGCATTGGTG (RACERBLong), and GATCGAGGCGTGAAGAATTAACAACC (RACERALong). Products were gel purified (QIAGEN, Valencia, CA) and sequenced to identify the 5'-UTR.

Tissue staining and analysis:

Ovaries were fixed in 4% paraformaldehyde and stained with DAPI, as described previously (CUMMINGS and CRONMILLER 1994). Antibodies for tissue staining included rabbit anti-Vasa 1:1000, rabbit anti-Mio 1:1000, rabbit anti-C(3)G 1:1000, mouse anti-Hts (1B1) 1:10, mouse anti-Bic-D (Bicaudal-D 4C2) 1:10, and mouse anti-Orb (Orb4H8) 1:10. [The last three antibodies were obtained from the Developmental Studies Hybridoma Bank (DSHB) maintained by the University of Iowa, Department of Biological Sciences.] Whole tissue immunostaining was performed as previously described (CRONMILLER and CUMMINGS 1993). Secondary antibodies were used at a concentration of 1:300 [FITC-conjugated goat anti-mouse and TRITC-conjugated goat anti-rabbit (Jackson Immunoresearch Laboratories)]. Stained ovaries were visualized on a Zeiss (Thornwood, NY) Axioskop microscope. Images were captured by an Olympus Magnafire digital camera and false colored in Adobe Photoshop.

In situ hybridization:

In vitro transcription plasmids were constructed by ligation of 699 bp of stl isoform A (of which only 100 bp are unique to A) or 635 bp of isoform B/C into pGEM-T Easy. Riboprobes were digoxigenin labeled via in vitro transcription and quantified by chemiluminescent dot blot. The hybridization protocol (WAHLI and BRAISSANT 1998) was modified as follows: Ovaries were dissected in 1x PBS, fixed for 10 min in 4% paraformaldehyde, treated with 0.1% active diethylpyrocarbonate (DEPC) for 20 min, and hybridized in 400 ng/ml riboprobe at 58° for 40 hr. After NBT/BCIP staining was stopped with TE buffer, ovaries were stored in 1x PBS at 4° (without dehydration) and mounted in 50% glycerol. Tissues were analyzed under DIC.

In silico analysis:

The Stl protein sequences were analyzed by SignalP 3.0 (BENDTSEN et al. 2004) for signal peptide sequence prediction (http://www.cbs.dtu.dk/services/SignalP/), by TMHMM v2.0 and Phobius for transmembrane domain and signal peptide prediction (http://www.cbs.dtu.dk/services/TMHMM/ and http://phobius.sbc.su.se), by NetNGlyc 1.0 for N-glycosylation consensus site prediction (http://www.cbs.dtu.dk/services/NetNGlyc/), by BLAT and Evoprinter for evolutionary conservation sequence analysis among Drosophila species (http://genome.ucsc.edu/cgi-bin/hgBlat and http://evoprinter.ninds.nih.gov/), and by ClustalW2 to align ADAMTS molecules of multiple species (http://www.ebi.ac.uk/Tools/clustalw2/index.html).

The stl mutant phenotype reflects a complete failure of follicle individualization (Figure 1, A and B) (BAKKEN 1973; SCHUPBACH and WIESCHAUS 1991). While the earliest defects appear in the pupal gonad (WILLARD et al. 2004), the disruption of ovarian morphology is complete even in newly eclosed adults. Subsequently, ongoing deterioration of ovarian structure as females age results in the grossly aberrant stl mutant phenotype that includes widespread cell degeneration (Figure 1C). To verify that these ovarian defects, which are common to all of the five extant stl mutant alleles, compose the null phenotype, we generated a deletion chromosome (see MATERIALS AND METHODS), examined the phenotype of hemizygous mutant flies (stl/Df), and found that it was indistinguishable from that of any mutant homozygotes (stl/stl) (Figure 1D). Using these five null alleles, we were able to identify the stl transcript and fill a critical gap in our understanding of the regulation of oogenesis.

stall encodes an ADAMTS metalloprotease:

Multiple approaches, including meiotic recombination, male recombination, deficiency mapping, and candidate gene sequencing of mutational lesions, were used to identify stl as CG3622 at 59B2 (Figure 2, A and B). The meiotic recombination frequencies between stl^a16 and l(2)06496 [II-102.0] and between stl^a16 and l(2)k06908 [II-104.5] were 1.5% (n = 1850) and 3.7% (n = 2000), respectively; this mapping placed stl at 100.5–100.8 and was confirmed with stl^pa49. Male recombination with multiple P insertions further refined the stl location to an 87-kb genomic region between blw and asrij. Of the 21 predicted coding regions in this interval, sequencing revealed only one gene with mutational lesions in each of the five extant stl alleles: This identified stl as CG3622.

View larger version (45K): In this window In a new window Download PPT slide   FIGURE 2.— stl is CG3622, encoding an ADAMTS metalloprotease. (A) Diagram of the transposable element insertions used in the genetic and molecular analysis of stl. P-element insertions (short triangles) were used to map stl by sequential male recombination analyses, delimiting the stl gene location to an 87-kb region between blw and asrij (arrows). PBac insertions (tall triangles) were used to generate molecularly defined deficiency chromosomes that confirmed both the stl gene location and the null phenotype. Within the 87-kb region, 21 candidate genes were sequenced from five independent stl mutant strains, leading to the identification of stl as CG3622 (B, red). (C) Semiquantitative RT–PCR analysis of stl expression in whole adults, dissected ovary tissue, or adult heads shows the presence of all three predicted isoforms, with the long isoform C most highly expressed. The amplified bands shown are 2985, 3445, and 3552 bp long for isoforms A, B, and C, respectively, and each band includes the coding region of its isoform. (D) Diagram of the three alternately spliced isoforms of CG3622 that were confirmed by RT–PCR and sequencing. The diagram includes previously undefined 5'-UTR for isoform C, as determined by 5' RACE. Dark gray represents the protein-coding region, while light gray indicates the untranslated regions. Colored lines mark the positions of stl mutational lesions, all of which fall within the protein-coding region (triangle, missense; asterisk, nonsense). (E) The Stl amino acid sequence, which identifies the protein as a member of the ADAMTS family. The protein domains are indicated as follows: prodomain, lavender highlight; reprolysin metalloprotease domain, green highlight; disintegrin-like domain, yellow highlight; thrombospondin (TS-1) repeat, purple highlight; cysteine-rich domain, orange highlight; and the C-terminal spacer region, blue highlight. This sequence shows the most N-terminal region unique to Stl-C (-C: 60 amino acids), the unique portion of Stl-B (-B: 15 amino acids), the region common to B and C (286 amino acids), and the remainder, which corresponds to the entire Stl-A isoform (-A: starting at M301/346) and is common to all three isoforms. The predicted signal sequence in Stl-C is italicized, the Zn-binding metalloprotease consensus sequence is boxed with a dotted line, potential furin cleavage sites are shown in boldface type (RxK/RR), and possible N-glycosylation sites are underlined (NxS/T). Finally, the positions of the molecular lesions of the stl point mutations are indicated by the boxed amino acids (color coordinated with the line indicators in D): red, stlawk26; yellow, stlwu40; purple, stla16; black, stlph57; and green, stlpa49.

 
Although the genome annotation of CG3622 varied over the course of our work, we confirmed three alternative splice forms of stl mRNA in wild-type tissues (Figure 2C). By RT–PCR and sequencing analyses, all three isoforms were found to share five common exons at their 3' ends (Figure 2, D and E). Isoforms B and C also shared three middle exons that were excluded from isoform A, while 5' exons for the three isoforms were found to be partially overlapping and/or unique. We used 5' RACE to characterize the 5'-UTRs for isoforms B and C, although low expression levels of isoform A apparently prevented precise determination of this transcript's 5' end by this method. Sequence analysis of these three stl mRNAs would predict proteins with 1136 aa (Stl-C), 1091 aa (Stl-B), or 790 aa (Stl-A). Moreover, the predicted amino acid sequences of all three Stl mRNA isoforms identified the conceptual protein of the gene as a metalloprotease, with many structural features in common with the ADAM and ADAMTS family of these proteins (Figure 2E).

The Stl protein is most similar to members of the ADAMTS proteins. ADAM and ADAMTS proteins are related in their enzymatic activity: They are metalloproteases that bind zinc ions to catalyze proteolytic cleavage of extracellular matrix (ECM) components. However, multiple structural domain differences between these protein families distinguish their associated substrates, cellular functions, and localization. Structural motifs of the canonical, secreted ADAMTS proteins include a prodomain, a Zn-binding and catalytic protease domain, a disintegrin-like domain, a central thrombospondin repeat (TSR) domain, a cysteine-rich region, a spacer region, and variable C-terminal TSRs (KUNO et al. 1997). The membrane-associated ADAMs differ from ADAMTSs primarily by the inclusion of a transmembrane domain and the absence of recognizable thrombospondin repeats. These differences appear to distinguish between membrane and extracellular sites where these proteins carry out their proteolytic functions. Importantly, the Stl protein contains a highly conserved Zn-binding and metalloprotease consensus sequence (Figure 2E, dotted outline), and this functional domain is most similar to the reprolysin domain of the ADAMTS family (KUNO et al. 1997). On the basis of overall sequence homology, the presence of a central thrombospondin repeat, and the secretion signal sequence in the long C isoform (Figure 2E, italics), the conceptual Stl protein appears most like the ADAMTS proteins (CAL et al. 2001; NICHOLSON et al. 2005).

As has been found for other ADAM/ADAMTS proteins (e.g., ADAM12 and ADAM28: GILPIN et al. 1998; ROBERTS et al. 1999; HOWARD et al. 2000), not all isoforms of Stl are likely to be secreted, on the basis of their N-terminal amino acid sequence differences (Figure 2E). By SignalP analysis, the longest isoform, Stl-C, contains a characteristic signal sequence (D-score = 0.723). For comparison, the closest mammalian ortholog, ADAMTS-16, possesses a strong signal sequence in both human (D-score = 0.733) and mouse (D-score = 0.650) and is predicted to be secreted (CAL et al. 2002). In contrast, neither Stl-A nor Stl-B Signal P analysis gave evidence of an unambiguously identifiable signal peptide (isoform A, D-score = 0.301; isoform B, D-score = 0.199) (NIELSEN et al. 1997; BENDTSEN et al. 2004); however, further study of their N-terminal sequences did suggest possible transmembrane positioning of these isoforms. The hydrophobicity plot of Stl isoform A was consistent with a single transmembrane domain at the N terminus (amino acids 13–32), and such a domain within the first 60 amino acids of a protein could act as a signal peptide (as defined by the transmembrane domain prediction programs, Phobius and TMHMM). The same amino acid sequence of Stl isoform B gave a far less convincing transmembrane prediction, with that sequence located much farther from the N terminus (amino acids 314–333); however, the sequence of Stl-B, like that of other ADAMTS proteins, contains a strong furin cleavage site at amino acid position 266–269 (Figure 2D, text in boldface type), and cleavage at this site would position a weak putative transmembrane domain much closer to the N terminus (amino acids 43–62). In the absence of any disrupting mutational lesions within the putative signal sequence/transmembrane domains (see below), further experiments would be necessary to verify the cellular localization of Stl.

Detection of Stl functional domains through mutational analysis and identification of sequence conservation:

To discriminate functionally among the three isoforms of Stl and to search for domains that are critical to the gene's roles during oogenesis, we identified the mutational lesions of all extant mutant alleles. While the mutational lesions for all five stl alleles were found to be distributed throughout the coding region of the gene (Table 2), the nucleotide changes associated with two of these mutations suggest that at least one of the longer protein isoforms (Stl-B and/or Stl-C) is required for oogenesis. A missense mutation in stl^wu40 was found to alter Gly₂₁₄ of Stl-C (Gly₁₆₉ of Stl-B), and although this residue has not been identified as part of a previously defined functional domain, it is conserved in the Stl orthologs from insects, rodents, and humans (Figure S1). Another mutant allele, stl^awk26, was found to contain two lesions, including a nonsense mutation that would severely truncate both Stl-B and Stl-C, leaving only 101 or 146 amino acids, respectively. The second lesion in this allele would affect only the Stl-A peptide, and since this change represents a conservative missense mutation in a nonconserved residue (T₉₉I, Figure S1), it is not clear whether the lesion would handicap any function associated with Stl-A. The mutational lesions of the remaining stl alleles were found to delete or alter defined functional domains that would be shared by all Stl isoforms. Two simple alleles, stl^ph57and stl^pa49, contained single nonsense mutations to truncate the protein(s). The more severe of these truncation alleles (stl^ph57) would delete the disintegrin-like, sole thrombospondin and cysteine-rich domains, as well as the variable C-terminal spacer region. The less severe truncation (stl^pa49) would remove only the C-terminal spacer. The identical null phenotypes that are associated with both of these truncation alleles affirm the importance of the C-terminal spacer region, which mediates substrate specificity in other ADAMTS proteins (FLANNERY et al. 2002; ZHENG et al. 2003; MAJERUS et al. 2005). Finally, stl^a16 was found to be the most polymorphic strain, and in this mutant we found several missense mutations. One of these mutations changed the highly conserved middle histidine (His_145/446/491) that defines the catalytic domain, and two mutations altered nonconserved amino acids at the extreme C terminus (Figure S1). Although the significance of the two distal mutations is unknown, mutation of the indispensable histidine illustrates the importance of Zn binding and protease activity for Stl function.

View this table: In this window In a new window   TABLE 2 Molecular lesions in stl alleles

 
We extended our analysis of stl structure and function by comparing the genome database sequences of Stl orthologs for eight Drosophila species (melanogaster, sechellia, simulans, yakuba, erecta, ananasssae, pseudoobscura, virilis, and grimshawi). BLAT search and EvoPrinter revealed four regions of high similarity (Table 3): the hypothetical transmembrane domain of Stl-A, the metalloprotease consensus, a small portion of the disintegrin-like domain, and a significant region of the C terminus. Although the signal sequence of isoform C is not conserved among all eight Drosophila, it is extremely similar in the six most closely related species. Taken together with the characterization of all extant stl alleles as functional nulls and the diversity of the mutational lesions associated with these alleles, the strong sequence conservation suggests that all of these regions of the protein are critical to its function. Moreover, the importance of these domains is evident in cell biological and biochemical studies of mammals as well (reviewed in ROCKS et al. 2008).

View this table: In this window In a new window   TABLE 3 Conserved sequences in Stl orthologs among eight Drosophila species

 

stl mutant effects in early oogenesis:

Because the specific cellular processes that are regulated by stl during early oogenesis have not been identified, we examined critical aspects of follicle morphogenesis, namely oocyte determination/localization and somatic cell differentiation/organization, in stl mutants. In the earliest stages of oogenesis, germ cell behavior and oocyte specification depend upon spectrosome/fusome function, and although we found altered fusome structure in stl mutant ovaries, oocyte determination was not affected. In the ovaries of newly eclosed stl mutant females, spectrosomes and fusomes were present and appropriately branched in the majority of germaria (Figure S2, A and B); however, in 4-day-old females fusomes were often not detected, and, when present, they appeared unbranched (Figure S2 C). In older females, both organelles were completely absent (Figure S2 D). Thus, fusome instability might be a primary consequence of stl loss-of-function. Consequently, if stl function directly regulates fusome maintenance, we would expect oocyte determination to be disrupted in stl mutant ovaries, since the fusome mediates this process (LIN and SPRADLING 1995). We used oocyte-specific molecular markers to examine oocyte determination in stl mutants and found no significant defects: For several different markers, both the relative distributions and expression levels were unaltered, even in severely defective mutant ovarioles (Figure S2, E–K). For example, Orb, C(3)G, Bic-D, and Mio were all appropriately restricted to a single cell per germline cyst (Figure S2, G and K, and data not shown). In 20% of stl ovarioles, morphological defects did include mispositioned oocytes, such that the oocyte of the oldest germline cyst was not found at the posterior end of the cyst (e.g., Figure S2 K). However, these mislocalized oocytes still expressed oocyte-specific markers and exhibited the characteristic oocyte-specific diploid nucleus. Thus, with apparently normal oocyte determination in stl mutants, altered fusome structure and stability are most likely indirect effects of the severely abnormal morphology.

Delta is a dominant enhancer of the stall mutant follicle formation phenotype:

Since previous genetic interaction analyses successfully identified only one other gene potentially involved in stl-mediated regulation of follicle morphogenesis, namely daughterless (da) (SMITH and CRONMILLER 2001), we screened deficiency chromosomes for dominant enhancement of the stl phenotype (see Table S1 for a listing of the specific deletions tested). This search yielded three regions of interest. The first cytological region (30D; 31F) included da, confirming the previously described noncomplementation of stl by da point mutations and underscoring the question of the molecular relationship between stl and da function (see below). The second region (44F10; 45E1) was found to be genetically complex, and analysis of overlapping deficiencies in this interval failed to distinguish a single interacting locus (Figure S3). Finally, the third region (91F1–2; 92F3–6) identified stat92E and Delta as putative interactors, and since both of these genes had previously been connected to da regulation of follicle formation (CUMMINGS and CRONMILLER 1994; SMITH et al. 2002), we examined their interactions with stl in greater detail. The stat92E–stl interaction was confirmed with a stat92E null mutation; however, this interaction turned out to be specific to the stl^a16 allele. When examined molecularly, the stl^a16 chromosome was found to carry a mutation in a nearby, stl-unrelated, putative STAT-binding site. It is possible that this binding site regulates expression of another gene that functions during early oogenesis, thereby accounting for the coincidence of the ovary interaction phenotype. The Dl–stl interaction, however, was not merely allele specific for either gene.

To investigate the potential regulatory link between stl and the N-Dl signaling pathway, we expanded our examination of stl genetic interactions. We found significant phenotypic interactions between stl and multiple Dl alleles; however, surprisingly, we observed no genetic interaction between stl and Notch (N), which is generally believed to partner universally with Dl in its various developmental functions. Three strong alleles, Dl^X, Dl⁹, and Dl³, displayed oocyte mislocalization and severe follicle formation defects when doubly heterozygous with various recessive alleles of stl (Figure 3, A–C), and all of these interactions were statistically significant (Figure 3D). Interactions with a Dl^– deletion [Df(3R)Dl-BX12] were also significant (e.g., Figure 3D), although for both this deletion and the Dl^X allele, considerable variation in the frequency of defects among individual females suggested complications associated with both of these genetic backgrounds. For example, the frequency of defects in Dl^X–stl^ph57 heterozygotes ranged from 15.4 to 82.4%, and even the Dl^X heterozygous controls showed substantial variation (2.6–48.8%). Nevertheless, positive genetic interactions for all of these Dl–stl allelic combinations were obvious, and the only Dl allele that did not exhibit a dominant interaction with any stl allele tested was the weak recessive mutation, Dl⁰⁵¹⁵¹ (Figure 3D). Neither Ser allele tested demonstrated any enhancement of stl in the ovary. Surprisingly, we also observed no synergistic phenotypes between stl and Notch (N). While this result could indicate that N expression or N signaling is stronger than that of Dl during oogenesis, making N less likely to show a genetic interaction when reduced by only one genetic copy, it is also possible that stl is involved in an N-independent function of Dl. Further experiments will be necessary to distinguish between these alternative explanations.

View larger version (97K): In this window In a new window Download PPT slide   FIGURE 3.— Combined reduction of Dl and stl results in dominant follicle formation defects, but reduction of N and stl has no effect. (A–C) DAPI-stained ovary samples, which show the dominant mutant interaction of stl, Dl double heterozygotes. (A) stlpa49/+; Dl9/+ ovarioles display defects in follicle separation, resulting in compound follicles and compound ovarioles not seen in either heterozygote alone. (B) stlph57/+; Dl9/+ ovarioles show that defects are not allele specific. (C) A stlph57/+; DlX/+ ovariole that contains compound follicles also shows a mislocalized oocyte; the arrow points to the diploid oocyte nucleus that is located centrally among the nurse cells. This ovariole also contains a partially dispersed border cell cluster (arrowhead) that contains a higher than normal number of cells, probably contributed by more than one set of polar cells in this multicyst ovariole. (D) The frequency of follicle defects in stlph57 plus Dl, N, or Ser double heterozygotes, expressed as the percentage of ovarioles that contained compound/morphologically disrupted follicles. Control genotypes were Dl, Df, N, or Ser single heterozygotes (see Table 1 for complete genotypes). Significant increases over the controls are in boldface type (a, P << 0.01; b, P < 0.05). The numbers of ovarioles scored are in brackets.

 

stl is expressed in head and ovarian tissue:

Because stl has been shown to have both ovary autonomous and nonautonomous functions (WILLARD et al. 2004), we looked at wild-type stl mRNA expression in the ovary, as well as in nonovarian tissues that could be involved in long distance regulation of ovarian follicle formation. By RT–PCR, we detected all three alternative splice forms in adult ovaries and heads; however, the relative abundance of the three isoforms was different. We measured high levels of stl-C in all tissues, but comparatively low levels of stl-A and especially low amounts of stl-B in heads (Figure 2C). To identify where stl was expressed in these tissues, we carried out in situ hybridization. Unfortunately, expression in the adult brain, as the primary component of the head sample, was too low or in too few cells to be detected by this method, although brain expression was detectable by the more sensitive real-time RT–PCR (Figure S4). Other components of the nervous/neuroendocrine system in which we were unable to detect stl mRNA by in situ hybridization included the corpus allatum, the corpus cardiacum, and the ventral nerve cord. In contrast, we detected expression of stl in both pupal and adult ovaries, beginning at 48 hr after puparium formation (APF). At this stage of ovary development, stl was expressed in the somatic basal stalk cells of each ovariole (Figure 4A), consistent in timing with the first stl mutant defects, which develop between 48 and 60 hr APF (WILLARD et al. 2004). In the more mature ovaries of pharate adults, stl was confined to follicular poles and was no longer present in the basal stalk cells (Figure 4B, arrows). Just before eclosion, stl was expressed strongly in the somatic cells of region 3 of the germarium, restricted to the polar cells of the most mature follicle (Figure 4C), and was conspicuously absent from the interfollicular stalks (Figure 4D, arrow). The strong expression of stl in region 3 of the germarium persisted in adult ovaries (Figure 4E); these cells play critical roles during follicle individualization, stalk formation, and oocyte localization. Finally, a peculiarity of stl mRNA distribution was the apparent apico-lateral localization at the plasma membrane of the somatic epithelium (Figure 4E, arrow). We were not able to account for this subcellular bias, although its absence from negative controls (data not shown) and its loss specifically in da mutant ovaries (Figure 4K) suggest that it was not a staining artifact.

View larger version (87K): In this window In a new window Download PPT slide   FIGURE 4.— The stl expression pattern is altered in da and Dl mutants. (A–E) Detection of stl mRNA in the ovary by in situ hybridization. (A) Wild-type pupal ovaries (48–60 hr APF) show expression of stl mRNA in the somatic basal stalks. (B–D) In pharate adult ovaries, stl expression is downregulated in the basal stalks (B, arrows), increases in region 3 of the germarium and at the follicular poles (C, arrowhead), and is not detected in interfollicular stalks (D, arrow). In adult ovaries (E), stl mRNA remains abundant in region 3 of the germarium, and lower levels are also detectable at the somatic-germ cell border of more mature follicles (arrow). Although no longer restricted to polar cells, mRNA levels remains slightly elevated at the poles of each follicle. Detection of stl mRNA in stl (F and G), Dl (H), and da (I and J) mutant ovaries shows altered expression. In stlph57 pupal ovaries (F), stl is still strongly expressed in the basal stalks, and in mutant defective adult ovarioles (G), stl is also expressed normally in cells corresponding to region 3 of the germarium. In DlX heterozygous ovarioles (H), there is no expression of stl mRNA in the germarium, ectopic expression in interfollicular stalks, and high polar region expression that persists aberrantly in later follicle stages. In dalyh mutant ovaries, stl mRNA expression is severely reduced. In pupal ovaries (I), fewer basal stalk cells express stl, and the levels are significantly lower than wild type (arrows); adult ovarioles (J) also show severely reduced levels of stl expression. (K) Ovarioles hybridized with a sense riboprobe control demonstrate the absence of nonspecific staining. (L) A diagram of the upstream and 5'-end structure of the stl gene (length shown, 5700 bp) displays the locations (vertical hatch marks) of E-box sequences, as potential Da-binding target sites. Transcribed regions are indicated relative to the E boxes below the map: stl-C 5'-UTR (red), stl-B 5'-UTR (green), and stl-A 5'-UTR (yellow). The full transcripts are not illustrated.

 

stl expression in the ovary is independent of the gene's extraovarian function:

Because the stl mutant phenotype becomes apparent early during pupal oogenesis and results principally from a loss of extraovarian gene function (WILLARD et al. 2004), we examined whether a primary consequence of that loss was the instability of stl-expressing cells in the ovary itself. The expression pattern of stl in both pupal and adult mutant ovaries was indistinguishable from that of wild-type tissues (Figure 4, F and G). In mutant pupal ovaries, stl was still strongly expressed in the basal stalks. Further, in spite of the gross morphological disruption of the adult mutant ovary, a strong band of stl expression, corresponding to cells found at region 3 of the wild-type germarium, was still present. In the absence of any interfollicular stalks, this dense population of stl-expressing cells spreads across the width of the defective ovariole. Thus, the extraovarian and/or early ovarian function of stl does not appear to be required for the maintenance of stl-expressing ovarian cells in the adult.

stl expression in the ovary is regulated by Delta and daughterless:

The strong mutant interactions that stl exhibited with Dl and da suggested that these genes might function in the same regulatory pathway as stl. Since da encodes a known transcription factor, we examined the possibility that Dl and da function upstream of stl, at least in the ovary. Indeed, we found that stl expression was disrupted in both Dl and da mutant ovaries. In ovaries from Dl^X heterozygous females, it was primarily the pattern of stl mRNA distribution that was altered (Figure 4H). Most ovarioles exhibited the expected stl expression pattern in the germarium and at the poles of early stage follicles. While the levels of expression appeared to be normal early, high levels of expression at the follicular poles persisted even in the later stages, when stl expression was found to decrease normally (compare Figure 4E and 4H). There was also often ectopic expression in interfollicular stalks. We did not observe, however, the distinct apico-lateral localization of the stl mRNA that characterized wild-type expression in individualized follicles. And, consistent with our failure to detect any genetic interaction between stl and N, the stl expression pattern was not significantly changed in N mutant ovaries (data not shown). In ovaries from da^lyh mutant females, it was primarily the levels of stl mRNA expression that were altered (Figure 4, I and J): stl expression was dramatically reduced in both pupal and adult ovaries. Since the Da protein has been identified as a helix-loop-helix (HLH) transcription factor, this result suggested that Da could be directly involved in the regulation of stl transcription. This possibility is supported by the presence of multiple HLH target sequences (E boxes, "CAnnTG") in and around the stl transcription unit (Figure 4L). Direct transcriptional regulation of stl by Da, particularly at the extraovarian site(s) of stl function, would account for the strong synergistic mutant interaction between these genes.

The molecular identification of the Drosophila Stall (Stl) protein as a metalloprotease reveals a previously unrecognized mechanism through which the metabolically demanding process of oogenesis can be modulated by the physiological condition of the fly. Previously known mechanisms of significant external regulation of oogenesis include the neurosecretory and hormonal systems: Responses to nutritional input through insulin-like peptides trigger long-range control of ovarian stem cell division and apoptotic checkpoints, while JH and 20E regulate yolk protein synthesis and uptake by ovarian follicle cells (SOLLER et al. 1999; DRUMMOND-BARBOSA and SPRADLING 2001; LAFEVER and DRUMMOND-BARBOSA 2005). Further, pharmacological studies have discovered nervous system regulation of follicle formation and vitellogenic follicle survival/maturation points that are mediated by the neurotransmitters, dopamine and serotonin, as well as by hormones (JH and 20E) (WILLARD et al. 2006). Finally, extraovarian regulation of follicle individualization and stalk formation has been associated with stl gene function, on the basis of the analysis of stl mutant somatic clones (WILLARD et al. 2004). The identification of the stl gene product as a member of the ADAMTS family of metalloproteases and the discovery that the gene is expressed in the brain together suggest a biochemical model in which Stl enzymatic activity is required to initiate a long-range neurosecretory signaling pathway. Moreover, the predicted structural differences of the three Stl isoforms, together with their comparisons to other ADAMTS family members, hint at ways in which stl functions both locally (autonomously) and remotely (nonautonomously) to regulate ovarian follicle formation.

Metzincins (metalloproteases that require zinc ions for activation), such as ADAMs and ADAMTSs, have been associated with multiple cellular roles, with considerable variation in both substrate specificity and subcellular sites of action (reviewed in SEALS and COURTNEIDGE 2003; PORTER et al. 2005). Generally implicated in cell adhesion and migration through ECM remodeling, many of these enzymes cleave ECM components, such as procollagen, aggrecan, versican, and brevican (ABBASZADE et al. 1999; TORTORELLA et al. 1999). Non-ECM functions include ectodomain shedding of transmembrane cell signaling molecules and the activation of cell surface molecules. For example, transmembrane ADAM-dependent cleavage can target growth factors, cytokines, and cell surface receptors (MASSAGUE and PANDIELLA 1993; BLACK et al. 1997; REDDY et al. 2000), and secreted ADAMTS-dependent cleavage of von Willebrand factor is essential for platelet aggregation (LEVY et al. 2001; ZHENG et al. 2001). While both families affect the local cellular environment, ADAM proteins are more likely to influence long-range signaling through release of extracellular cell signaling peptides.

By structural criteria alone, it is not easy either to classify Stl as strictly an ADAM vs. ADAMTS or to declare if/how each Stl isoform functions. By BLAST sequence analysis, the metalloprotease domain of Stl is most similar to that of human ADAMTS16. Notably, ADAMTS16 is also highly expressed in ovarian follicles, is induced hormonally by follicle stimulating hormone, has detectable proteolytic activity, and consists of alternatively spliced isoforms (GAO et al. 2007). However, both ADAMTS16 isoforms are predicted to be secreted. The Stl protein is expected to have proteolytic function, on the basis of a perfect Zn-binding metalloprotease consensus sequence and a disrupting mutation in that domain in stl^a16, but its primary sequence suggests isoform-specific processing. On the one hand, Stl-B resembles the ADAMs, which generally contain a transmembrane domain that organizes their metalloprotease domains outside the cell for interactions with other integral membrane proteins and the ECM. With a predicted transmembrane domain in the middle of the protein, Stl-B is not expected to be secreted, yet some aspects of Stl's structure suggest the protein may be processed more like an ADAMTS. These enzymes are generally expressed as inactive proprotein forms (zymogens), in which a long (230-aa) prodomain blocks the zinc-binding catalytic domain; following transport to the cell membrane as an inactive protein, furin-dependent prodomain removal results in enzymatic activation. ADAMTS prodomains are generally longer than those of ADAMs, and there are typically two furin recognition sites. Like an ADAMTS, Stl-B contains two furin target sites in its N terminus, and cleavage at the stronger site would yield a prodomain of 269 amino acids. Alternatively, ADAMTS1 is processed intracellularly and secreted as an active enzyme (LONGPRE and LEDUC 2004; KOO et al. 2006). In fact, the signal sequence of ADAMTS1 is so weak that the signal peptide prediction program, SignalP, fails to predict the protein's secretion. Because SignalP also failed to detect a signal sequence in Stl-B, it is possible that Stl-B is processed intracellularly and secreted in a manner similar to ADAMTS1. Finally, ADAMTS13 has an abnormally short prodomain (40 aa) and can be secreted and/or proteolytically active even in the absence of prodomain cleavage (MAJERUS et al. 2003). Thus, the shorter isoform Stl-A, which has no prodomain N terminal to its metalloprotease domain, could similarly be functional in the absence of any processing. Most definitive of the three isoforms, Stl-C contains a clear signal sequence and is overwhelmingly predicted to be secreted as a typical ADAMTS. In the end, there may not be a single mechanism by which Stl reaches its proper subcellular localization(s), and the apparent complexity of this issue could relate to the protein's function within vs. outside the ovary. The details of Stl function require further in vivo analysis to resolve.

Although Stl appears to be somewhat of a structural anomaly within the ADAMTS family, a comparison of known ADAMTS processing with the essential structural features of Stl that were identified by mutational lesions suggests possible mechanisms of regulation. First, several ADAMTSs undergo processing at the C terminus for regulated proteolytic activity, and the C-terminal spacer region has been shown to be important for substrate binding and the specificity of the enzyme's biological activity (ZHENG et al. 2003; MAJERUS et al. 2005). ADAMTS4 is cleaved extracellularly, thereby greatly increasing its aggrecanase activity (FLANNERY et al. 2002; GAO et al. 2002, 2004), and ADAMTS1 loses a portion of its C terminus, thereby losing affinity for heparin (RODRIGUEZ-MANZANEQUE et al. 2000). It is likely that the C-terminal spacer region is essential for Stl function as well: The least extreme nonsense allele, stl^pa49, truncates only the C-terminal spacer region (of all isoforms), and yet the mutant phenotype is as severe as if the metalloprotease domain itself were mutated. It is unlikely that the null phenotype of stl nonsense mutations results from nonsense-mediated decay of mutant transcripts, since amplification and sequencing of all stl alleles was performed on easily generated cDNA. However, it is possible that instability of truncated mutant protein results in stl loss-of-function; antibody analysis could address this issue. Next, ADAMTS proteins are frequently substrates for oligosaccharide modification, and these modifications affect membrane targeting and secretion. Generally, the propeptide has sites for N-glycosylation [e.g., ADAMTS9 (KOO et al. 2006)]. Stl-B and Stl-C have two potential N-glycosylation targets (NxS/T) within Stl-B's putative prodomain, and three additional sites within the thrombospondin and cys-rich regions are shared by all three Stl isoforms. Since N-glycosylation can affect protein folding and secretion efficiency [e.g., synaptotagmin 1 and inhibin/activin (HAN et al. 2004; ANTENOS et al. 2007)], it is possible that these sites in Stl either facilitate secretion in the absence of a strong signal peptide (Stl-B) or boost a weak signal sequence (Stl-A). Other ADAMTS oligosaccharide modifications, C-mannosylation and O-fucosylation, typically occur within the TSRs and can be important for function (e.g., ADAMTS13) (GONZALEZ DE PEREDO et al. 2002; PAAKKONEN et al. 2006; RICKETTS et al. 2007). Capacity for C-mannosylation is noteworthy in the context of Stl, because the solitary central TSR of Stl (with no C-terminal TSRs) is one of the major structural differences that marginalize Stl within the ADAMTS family and because C-mannosylation occurs only on secreted proteins. The Stl amino acid sequence contains C-mannosylation sites (WGEW) at the very beginning of its TSR and in the metalloprotease domain; however, there are no apparent O-fucosylation consensus sites. Such deviation from the "typical" ADAMTS profile further confounds the question of Stl's subcellular localization (secreted vs. not). It also could suggest that Stl is an evolutionary intermediate between ADAMs (no TSRs, no C-mannosylation, transmembrane rather than secreted) and ADAMTSs (multiple TSRs, with prominent C-mannosylation, generally secreted).

For now, stall represents a unique form of regulation of oogenesis. But does this gene function in the context of any known regulatory pathways? On the basis of dominant genetic interactions during follicle formation, stl, da, and Dl are all clearly functionally connected (CUMMINGS and CRONMILLER 1994); however, stl cannot be a simple common component in the previously described da/Dl oogenesis pathways. First, and most significantly, all of the essential requirements for da and Dl during oogenesis have so far been identified as functions within the ovary itself (LOPEZ-SCHIER and ST. JOHNSTON 2001; SMITH et al. 2002). In contrast, the most critical stl function originates outside the ovary, with only a minor role for the gene in the somatic ovary according to mitotic clonal analysis (WILLARD et al. 2004). It is possible that da and/or Dl could also have roles outside the ovary. Both have neurogenic functions earlier in Drosophila development (ALTON et al. 1988; CABRERA 1990; CRONMILLER and CUMMINGS 1993), so an additional function in the adult nervous system is feasible. However, at least for da, such a nonautonomous role in controlling oogenesis is improbable, since an ovary-specific mutant allele already displays the null phenotype for oogenesis (SMITH and CRONMILLER 2001). Second, a simple linear relationship between stl and da or Dl is unlikely because of plain differences in the cellular processes that are affected by these genes and/or some demonstrated genetic mechanisms for their functions. For example, although mutations in both stl and da affect follicle individualization, one reason for the da effect is disruption of the gene's normal regulation of cell proliferation within the germarium (SMITH et al. 2002); however, BrdU incorporation studies have shown that stl has no such activity (E. F. OZDOWSKI and C. CRONMILLER, unpublished data). Moreover, all previously described Dl functions involve its receptor Notch, and even da interacts genetically with both N and Dl (CUMMINGS and CRONMILLER 1994); yet, the stl genetic interaction with Dl does not extend to N. Perhaps this functional connection between stl and Dl in the ovary is related mechanistically to an N-independent Dl function that has been inferred from in vitro cell culture studies, which also demonstrated da involvement (MOK et al. 2005).

Finally, elucidating the specific cellular mechanism(s) of stl-mediated regulation of ovarian follicle morphogenesis will certainly require the identification of the substrate(s) of Stl metalloprotease activity. It is the extraovarian metalloprotease activity of Stl that is the critical one to discover, since it is that activity that leads to long-range communication with the ovary. While substrates of metalloproteases have generally been difficult to identify, genomewide yeast two-hybrid analyses have identified two potential physical interactions for the Stl protein: products of CG12517 and CG31357 (GIOT et al. 2003). These should certainly be examined for molecular contributions to the stl regulatory pathway. Otherwise, comprehensive biochemical approaches, together with genetic strategies, will probably be required to identify Stl's target(s).

We thank Maho Shibata for technical assistance with anti-Da immunofluorescence and appreciate the gifts of anti-Mio and anti-C(3)G antibodies from M. Lilly and anti-Vasa from P. Lasko. This work was supported by grants to C.C. from the Thomas F. Jeffress and Kate Miller Jeffress Memorial Trust and the National Institute on Drug Abuse. E.O. was supported by a National Institute of Child Health and Human Development training grant (HD07528-01). Y.M. was supported by a Harrison Undergraduate Research Award from The University of Virginia.

Supporting information is available online at http://www.genetics.org/cgi/content/full/genetics.109.107367/DC1.

ABBASZADE, I., R. Q. LIU, F. YANG, S. A. ROSENFELD, O. H. ROSS et al., 1999 Cloning and characterization of ADAMTS11, an aggrecanase from the ADAMTS family. J. Biol. Chem. 274: 23443–23450.[Abstract/Free Full Text]

ALTON, A. K., K. FECHTEL, A. L. TERRY, S. B. MEIKLE and M. A. MUSKAVITCH, 1988 Cytogenetic definition and morphogenetic analysis of Delta, a gene affecting neurogenesis in Drosophila melanogaster. Genetics 118: 235–245.[Abstract/Free Full Text]

ANTENOS, M., M. STEMLER, I. BOIME and T. K. WOODRUFF, 2007 N-linked oligosaccharides direct the differential assembly and secretion of inhibin {alpha}- and {beta}A-subunit dimers. Mol. Endocrinol. 21: 1670–1684.[Abstract/Free Full Text]

BAKKEN, A. H., 1973 A cytological and genetic study of oogenesis in Drosophila melanogaster. Dev. Biol. 33: 100–122.[CrossRef][Medline]

BASTOCK, R., and D. ST. JOHNSTON, 2008 Drosophila oogenesis. Curr. Biol. 18: R1082–R1087.[CrossRef][Medline]

BENDTSEN, J. D., H. NIELSEN, G. VON HEIJNE and S. BRUNAK, 2004 Improved prediction of signal peptides: SignalP 3.0. J. Mol. Biol. 340: 783–795.[CrossRef][Medline]

BERG, C. A., 2008 Tube formation in Drosophila egg chambers. Tissue Eng. Part A 14: 1479–1488.[CrossRef][Medline]

BLACK, R. A., C. T. RAUCH, C. J. KOZLOSKY, J. J. PESCHON, J. L. SLACK et al., 1997 A metalloproteinase disintegrin that releases tumour-necrosis factor-alpha from cells. Nature 385: 729–733.[CrossRef][Medline]

CABRERA, C. V., 1990 Lateral inhibition and cell fate during neurogenesis in Drosophila: the interactions between scute, Notch and Delta. Development 110: 733–742.[Abstract/Free Full Text]

CAL, S., J. M. ARGUELLES, P. L. FERNANDEZ and C. LOPEZ-OTIN, 2001 Identification, characterization, and intracellular processing of ADAM-TS12, a novel human disintegrin with a complex structural organization involving multiple thrombospondin-1 repeats. J. Biol. Chem. 276: 17932–17940.[Abstract/Free Full Text]

CAL, S., A. J. OBAYA, M. LLAMAZARES, C. GARABAYA, V. QUESADA et al., 2002 Cloning, expression analysis, and structural characterization of seven novel human ADAMTSs, a family of metalloproteinases with disintegrin and thrombospondin-1 domains. Gene 283: 49–62.[CrossRef][Medline]

CRONMILLER, C., and C. A. CUMMINGS, 1993 The daughterless gene product in Drosophila is a nuclear protein that is broadly expressed throughout the organism during development. Mech. Dev. 42: 159–169.[CrossRef][Medline]

CUMMINGS, C. A., and C. CRONMILLER, 1994 The daughterless gene functions together with Notch and Delta in the control of ovarian follicle development in Drosophila. Development 120: 381–394.[Abstract]

DRUMMOND-BARBOSA, D., and A. C. SPRADLING, 2001 Stem cells and their progeny respond to nutritional changes during Drosophila oogenesis. Dev. Biol. 231: 265–278.[CrossRef][Medline]

FLANNERY, C. R., W. ZENG, C. CORCORAN, L. A. COLLINS-RACIE, P. S. CHOCKALINGAM et al., 2002 Autocatalytic cleavage of ADAMTS-4 (Aggrecanase-1) reveals multiple glycosaminoglycan-binding sites. J. Biol. Chem. 277: 42775–42780.[Abstract/Free Full Text]

GAO, G., J. WESTLING, V. P. THOMPSON, T. D. HOWELL, P. E. GOTTSCHALL et al., 2002 Activation of the proteolytic activity of ADAMTS4 (aggrecanase-1) by C-terminal truncation. J. Biol. Chem. 277: 11034–11041.[Abstract/Free Full Text]

GAO, G., A. PLAAS, V. P. THOMPSON, S. JIN, F. ZUO et al., 2004 ADAMTS4 (aggrecanase-1) activation on the cell surface involves C-terminal cleavage by glycosylphosphatidyl inositol-anchored membrane type 4-matrix metalloproteinase and binding of the activated proteinase to chondroitin sulfate and heparan sulfate on syndecan-1. J. Biol. Chem. 279: 10042–10051.[Abstract/Free Full Text]

GAO, S., C. DE GEYTER, K. KOSSOWSKA and H. ZHANG, 2007 FSH stimulates the expression of the ADAMTS-16 protease in mature human ovarian follicles. Mol. Hum. Reprod. 13: 465–471.[Abstract/Free Full Text]

GILPIN, B. J., F. LOECHEL, M. G. MATTEI, E. ENGVALL, R. ALBRECHTSEN et al., 1998 A novel, secreted form of human ADAM 12 (meltrin alpha) provokes myogenesis in vivo. J. Biol. Chem. 273: 157–166.[Abstract/Free Full Text]

GIOT, L., J. S. BADER, C. BROUWER, A. CHAUDHURI, B. KUANG et al., 2003 A protein interaction map of Drosophila melanogaster. Science 302: 1727–1736.[Abstract/Free Full Text]

GODT, D., and U. TEPASS, 1998 Drosophila oocyte localization is mediated by differential cadherin-based adhesion. Nature 395: 387–391.[CrossRef][Medline]

GONZALEZ DE PEREDO, A., D. KLEIN, B. MACEK, D. HESS, J. PETER-KATALINIC et al., 2002 C-mannosylation and o-fucosylation of thrombospondin type 1 repeats. Mol. Cell. Proteomics 1: 11–18.[Abstract/Free Full Text]

GONZALEZ-REYES, A., and D. ST. JOHNSTON, 1998 The Drosophila AP axis is polarised by the cadherin-mediated positioning of the oocyte. Development 125: 3635–3644.[Abstract]

GRUNTENKO, N. E., and I. Y. RAUSCHENBACH, 2008 Interplay of JH, 20E and biogenic amines under normal and stress conditions and its effect on reproduction. J. Insect Physiol. 54: 902–908.[CrossRef][Medline]

HAN, W., J. S. RHEE, A. MAXIMOV, Y. LAO, T. MASHIMO et al., 2004 N-glycosylation is essential for vesicular targeting of synaptotagmin 1. Neuron 41: 85–99.[CrossRef][Medline]

HOWARD, L., R. A. MACIEWICZ and C. P. BLOBEL, 2000 Cloning and characterization of ADAM28: evidence for autocatalytic pro-domain removal and for cell surface localization of mature ADAM28. Biochem. J. 348(Pt. 1): 21–27.[CrossRef][Medline]

KING, R. C., 1970 The meiotic behavior of the Drosophila oocyte. Int. Rev. Cytol. 28: 125–168.[Medline]

KOO, B. H., J. M. LONGPRE, R. P. SOMERVILLE, J. P. ALEXANDER, R. LEDUC et al., 2006 Cell-surface processing of pro-ADAMTS9 by furin. J. Biol. Chem. 281: 12485–12494.[Abstract/Free Full Text]

KUNO, K., N. KANADA, E. NAKASHIMA, F. FUJIKI, F. ICHIMURA et al., 1997 Molecular cloning of a gene encoding a new type of metalloproteinase-disintegrin family protein with thrombospondin motifs as an inflammation associated gene. J. Biol. Chem. 272: 556–562.[Abstract/Free Full Text]

LAFEVER, L., and D. DRUMMOND-BARBOSA, 2005 Direct control of germline stem cell division and cyst growth by neural insulin in Drosophila. Science 309: 1071–1073.[Abstract/Free Full Text]

LEVY, G. G., W. C. NICHOLS, E. C. LIAN, T. FOROUD, J. N. MCCLINTICK et al., 2001 Mutations in a member of the ADAMTS gene family cause thrombotic thrombocytopenic purpura. Nature 413: 488–494.[CrossRef][Medline]

LIN, H., and A. C. SPRADLING, 1995 Fusome asymmetry and oocyte determination in Drosophila. Dev. Genet. 16: 6–12.[CrossRef][Medline]

LIN, H., L. YUE and A. C. SPRADLING, 1994 The Drosophila fusome, a germline-specific organelle, contains membrane skeletal proteins and functions in cyst formation. Development 120: 947–956.[Abstract]

LONGPRE, J. M., and R. LEDUC, 2004 Identification of prodomain determinants involved in ADAMTS-1 biosynthesis. J. Biol. Chem. 279: 33237–33245.[Abstract/Free Full Text]

LOPEZ-SCHIER, H., and D. ST. JOHNSTON, 2001 Delta signaling from the germ line controls the proliferation and differentiation of the somatic follicle cells during Drosophila oogenesis. Genes Dev. 15: 1393–1405.[Abstract/Free Full Text]

MAJERUS, E. M., X. ZHENG, E. A. TULEY and J. E. SADLER, 2003 Cleavage of the ADAMTS13 propeptide is not required for protease activity. J. Biol. Chem. 278: 46643–46648.[Abstract/Free Full Text]

MAJERUS, E. M., P. J. ANDERSON and J. E. SADLER, 2005 Binding of ADAMTS13 to von Willebrand factor. J. Biol. Chem. 280: 21773–21778.[Abstract/Free Full Text]

MARGOLIS, J., and A. SPRADLING, 1995 Identification and behavior of epithelial stem cells in the Drosophila ovary. Development 121: 3797–3807.[Abstract]

MASSAGUE, J., and A. PANDIELLA, 1993 Membrane-anchored growth factors. Annu. Rev. Biochem. 62: 515–541.[CrossRef][Medline]

MCGREGOR, J. R., R. XI and D. A. HARRISON, 2002 JAK signaling is somatically required for follicle cell differentiation in Drosophila. Development 129: 705–717.[Abstract/Free Full Text]

MOK, L. P., T. QIN, B. BARDOT, M. LECOMTE, A. HOMAYOUNI et al., 2005 Delta activity independent of its activity as a ligand of Notch. BMC Dev. Biol. 5: 6.[CrossRef][Medline]

NICHOLSON, A. C., S. B. MALIK, J. M. LOGSDON, JR. and E. G. VAN MEIR, 2005 Functional evolution of ADAMTS genes: evidence from analyses of phylogeny and gene organization. BMC Evol. Biol. 5: 11.[CrossRef][Medline]

NIELSEN, H., J. ENGELBRECHT, S. BRUNAK and G. VON HEIJNE, 1997 Identification of prokaryotic and eukaryotic signal peptides and prediction of their cleavage sites. Protein Eng. 10: 1–6.[Abstract/Free Full Text]

NYSTUL, T., and A. SPRADLING, 2007 An epithelial niche in the Drosophila ovary undergoes long-range stem cell replacement. Cell Stem Cell 1: 277–285.[CrossRef][Medline]

PAAKKONEN, K., H. TOSSAVAINEN, P. PERMI, H. RAKKOLAINEN, H. RAUVALA et al., 2006 Solution structures of the first and fourth TSR domains of F-spondin. Proteins 64: 665–672.[CrossRef][Medline]

PORTER, S., I. M. CLARK, L. KEVORKIAN and D. R. EDWARDS, 2005 The ADAMTS metalloproteinases. Biochem. J. 386: 15–27.[CrossRef][Medline]

REDDY, P., J. L. SLACK, R. DAVIS, D. P. CERRETTI, C. J. KOZLOSKY et al., 2000 Functional analysis of the domain structure of tumor necrosis factor-alpha converting enzyme. J. Biol. Chem. 275: 14608–14614.[Abstract/Free Full Text]

RICKETTS, L. M., M. DLUGOSZ, K. B. LUTHER, R. S. HALTIWANGER and E. M. MAJERUS, 2007 O-fucosylation is required for ADAMTS13 secretion. J. Biol. Chem. 282: 17014–17023.[Abstract/Free Full Text]

ROBERTS, C. M., P. H. TANI, L. C. BRIDGES, Z. LASZIK and R. D. BOWDITCH, 1999 MDC-L, a novel metalloprotease disintegrin cysteine-rich protein family member expressed by human lymphocytes. J. Biol. Chem. 274: 29251–29259.[Abstract/Free Full Text]

ROCKS, N., G. PAULISSEN, M. EL HOUR, F. QUESADA, C. CRAHAY et al., 2008 Emerging roles of ADAM and ADAMTS metalloproteinases in cancer. Biochimie 90: 369–379.[CrossRef][Medline]

RODRIGUEZ-MANZANEQUE, J. C., A. B. MILCHANOWSKI, E. K. DUFOUR, R. LEDUC and M. L. IRUELA-ARISPE, 2000 Characterization of METH-1/ADAMTS1 processing reveals two distinct active forms. J. Biol. Chem. 275: 33471–33479.[Abstract/Free Full Text]

ROPER, K., and N. H. BROWN, 2004 A spectraplakin is enriched on the fusome and organizes microtubules during oocyte specification in Drosophila. Curr. Biol. 14: 99–110.[Medline]

RUOHOLA, H., K. A. BREMER, D. BAKER, J. R. SWEDLOW, L. Y. JAN et al., 1991 Role of neurogenic genes in establishment of follicle cell fate and oocyte polarity during oogenesis in Drosophila. Cell 66: 433–449.[CrossRef][Medline]

SCHUPBACH, T., and E. WIESCHAUS, 1991 Female sterile mutations on the second chromosome of Drosophila melanogaster. II. Mutations blocking oogenesis or altering egg morphology. Genetics 129: 1119–1136.[Abstract]

SEALS, D. F., and S. A. COURTNEIDGE, 2003 The ADAMs family of metalloproteases: multidomain proteins with multiple functions. Genes Dev. 17: 7–30.[Free Full Text]

SMITH, III, J. E., and C. CRONMILLER, 2001 The Drosophila daughterless gene autoregulates and is controlled by both positive and negative cis regulation. Development 128: 4705–4714.[Abstract/Free Full Text]

SMITH, III, J. E., C. A. CUMMINGS and C. CRONMILLER, 2002 Daughterless coordinates somatic cell proliferation, differentiation and germline cyst survival during follicle formation in Drosophila. Development 129: 3255–3267.[Abstract/Free Full Text]

SOLLER, M., M. BOWNES and E. KUBLI, 1999 Control of oocyte maturation in sexually mature Drosophila females. Dev. Biol. 208: 337–351.[CrossRef][Medline]

SPRADLING, A., 1993 Developmental Genetics of Oogenesis. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY.

TORRES, I. L., H. LOPEZ-SCHIER and D. ST. JOHNSTON, 2003 A Notch/Delta-dependent relay mechanism establishes anterior-posterior polarity in Drosophila. Dev. Cell 5: 547–558.[CrossRef][Medline]

TORTORELLA, M. D., T. C. BURN, M. A. PRATTA, I. ABBASZADE, J. M. HOLLIS et al., 1999 Purification and cloning of aggrecanase-1: a member of the ADAMTS family of proteins. Science 284: 1664–1666.[Abstract/Free Full Text]

TWOROGER, M., M. K. LARKIN, Z. BRYANT and H. RUOHOLA-BAKER, 1999 Mosaic analysis in the Drosophila ovary reveals a common hedgehog-inducible precursor stage for stalk and polar cells. Genetics 151: 739–748.[Abstract/Free Full Text]

WAHLI, W., and O. BRAISSANT, 1998 A simplified in situ hybridization protocol using non-radioactively labeled probes to detect abundant and rare mRNAs on tissue sections. Biochemica 1: 10–16.

WILLARD, S. S., E. F. OZDOWSKI, N. A. JONES and C. CRONMILLER, 2004 stall-mediated extrinsic control of ovarian follicle formation in Drosophila. Genetics 168: 191–198.[Abstract/Free Full Text]

WILLARD, S. S., C. M. KOSS and C. CRONMILLER, 2006 Chronic cocaine exposure in Drosophila: life, cell death and oogenesis. Dev. Biol. 296: 150–163.[CrossRef][Medline]

ZHENG, X., D. CHUNG, T. K. TAKAYAMA, E. M. MAJERUS, J. E. SADLER et al., 2001 Structure of von Willebrand factor-cleaving protease (ADAMTS13), a metalloprotease involved in thrombotic thrombocytopenic purpura. J. Biol. Chem. 276: 41059–41063.[Abstract/Free Full Text]

ZHENG, X., K. NISHIO, E. M. MAJERUS and J. E. SADLER, 2003 Cleavage of von Willebrand factor requires the spacer domain of the metalloprotease ADAMTS13. J. Biol. Chem. 278: 30136–30141.[Abstract/Free Full Text]

Communicating editor: T. SCHUPBACH

International Access Link

Online ISSN: 1943-2361  Print ISSN: 0016-6731
Copyright 2009 by the Genetics Society of America
phone: 412-268-1812   fax: 412-268-1813   email: genetics-gsa{at}andrew.cmu.edu