The differentiation of Drosophila germ cells is a useful model for studying mechanisms of cell specification. We report the identification of a gene, stonewall, that is required for germ cell development. Mutations in stonewall block proper oocyte differentiation and frequently cause the presumptive oocyte to develop as a nurse cell. Eventually, germ cells degenerate apoptotically. Stonewall is a germ cell nuclear protein; Stonewall has a DNA binding domain that shows similarities to the Myb and Adf-1 transcription factors and has other features that suggest that it is a transcription activating factor. We suggest that Stonewall transcriptional regulation is essential in cystocytes for maturation into specialized nurse cells and oocyte.

Cells are the targets for a broad range of both extrinsic and intrinsic differentiation factors. Molecular pairs of diffusible ligands and membrane bound receptors (Schüpbach and Roth, 1994), cell contact dependent interactions (Artavanis-Tsakonas et al., 1995), and cascades of intracellular signaling molecules (Wasserman, 1993) are only a fraction of the biochemical signals that regulate and/or induce differentiation in cells. Very often the intracellular end of a signaling pathway is a transcription factor that responds to the signals to regulate the expression of a specific set of genes

We and others are studying the oogenic germ cell lineage as a model for cell differentiation because it offers special genetic and cytological advantages (see Spradling, 1993, for review) and because of intrinsic interest in understanding how germ cells develop. We have focused our efforts on events that occur very early in the oogenic germ cell lineage by identifying mutations in genes required for four steps of germ cell differentiation; stem cells into cystoblasts, cystoblasts into cystocytes, one cystocyte into an oocyte and the remaining cystocytes into nurse cells (Fig. 1A; also see below). Some of the molecules discovered to date have proved to be germ cell-specific differentiation factors (McKearin and Ohlstein, 1995; Christerson and McKearin, 1994; King and Storto, 1988; Oliver et al., 1987) while others have proved to be more ubiquitous proteins used in novel ways in germ cells (Lin et al., 1994; Ran et al., 1994).

Fig. 1.

The germ cell lineage and formation of the egg chamber. In A, the pattern of cystocyte divisions is presented in schematic form. One of the sixteen cystocytes will develop as the oocyte while its sisters will become nurse cells. B shows a drawing of a typical germarium with germarial regions labeled. The anterior end of the germarium is marked by a stack of somatic cells called the terminal filaments (tf). The egg chamber in Region 3 is called an oogenic stage 1 chamber; it will be pushed from the germarium into more posterior regions of the ovary for maturation.

Fig. 1.

The germ cell lineage and formation of the egg chamber. In A, the pattern of cystocyte divisions is presented in schematic form. One of the sixteen cystocytes will develop as the oocyte while its sisters will become nurse cells. B shows a drawing of a typical germarium with germarial regions labeled. The anterior end of the germarium is marked by a stack of somatic cells called the terminal filaments (tf). The egg chamber in Region 3 is called an oogenic stage 1 chamber; it will be pushed from the germarium into more posterior regions of the ovary for maturation.

The Drosophila egg chamber is made up of sixteen interconnected germline cells, the cystocytes, surrounded by a monolayer of somatic follicle cells (Fig. 1; for a recent review of Drosophila oogenesis, see Spradling, 1993). The syncytium of cystocytes is the product of four successive nuclear divisions of a single cystoblast, with incomplete cytokinesis during each mitosis. Once the cluster of cystocytes is complete, somatic cells from the surrounding tissue encircle the cluster, forming an epithelial monolayer of follicle cells. These early events of germ cell differentiation take place in a substructure of the ovary, the germarium (Fig. 1B).

Maturation of the egg chamber involves dramatic specializations of both the germ cells and follicle cells; this report will focus on the differentiation of the former. One cystocyte develops as the oocyte. This cell becomes transcriptionally quiescent and its genome is committed to meiosis, forming stable synaptonemal complexes and arresting in meiotic prophase in Region 2b of the germarium (Koch et al., 1966). Its fifteen interconnected sisters become nurse cells, undergoing many rounds of endoreduplication to become highly polyploid; these transcriptionally active polyploid cells produce biosynthetic products for the oocyte.

The cystocyte that becomes the oocyte is always one of the two oldest in the syncytium and, because the planes of cystocyte divisions are regulated and stereotypical, always develops from a cystocyte with four intercellular connections or ring canals (Brown and King, 1964). Cystocyte mitoses are accompanied by the growth of an unusual cellular organelle, the fusome, from a spherical dot to a highly branched structure that eventually penetrates all 16 cystocytes. Apparently the fusome is responsible for fixing the planes of cystocytes divisions and elegant studies from Lin and Spradling (1995) have suggested that it may be a critical element in determining which cystocyte becomes the oocyte.

One mutation, egalitarian, results in all cystocyte genomes initially forming synaptonemal complexes and subsequently all losing those complexes and developing as nurse cells (Carpenter, 1994). Null mutations in a second gene, Bicaudal D (BicD), also causes egg chambers to form with 16 nurse cells and no oocyte (Ran et al., 1994), apparently by disrupting the formation of a polarized array of microtubules within the syncytium. Normally the polarized microtubule network directs the accumulation of many mRNAs and proteins to the oocyte and this polarized transport is probably a critical element of oocyte differentiation (Theurkauf et al., 1993).

Present data suggests that BicD and egalitarian probably act close to the time of the oocyte determining event(s) (Ran et al., 1994; Carpenter, 1994). However, a cadre of other genes must be responsible for executing the determination decision and principal among these downstream genes might be transcription factors that regulate the nuclear events that produce differentiated oocytes and nurse cells. In this paper we report the identification of a candidate for such a germ cell maturation effector, stonewall. A null mutation affects only oogenesis, producing a limited set of germ cell phenotypes that indicate disruptions in cystocyte maturation. The gene is transcribed and the protein expressed in young cystocytes; the protein is nuclear and is similar to a family of helix-turn-helix transcriptional activating proteins.

Fly stocks and genetic crosses

All fly stocks used were maintained under standard culture conditions. w1118 served as a wild-type control in all experiments. The following deficiency stocks were used to test complementation of stwl P alleles: Df(3L)fz-M21 (70D2-3;70E4-5) was from the Indiana Drosophila Stock Center while Df(3L)D5 (70D2-4) and Df(3L)D5-R+6 (70C770D6) were kindly supplied by A. Carpenter. Descriptions of stocks can be found in Flybase Stock Lists.

Isolation of mutants

Two stwl alleles, fs(3)ry1 and fs(3)ry2, were isolated from a collection of 1019 P[ry+] insertion stocks (Berg and Spradling, 1991). Five additional P[lacZ; ry+] stwl alleles were recovered from a collection of approx. 8000 P-insertion lines (Karpen and Spradling, 1992). These were recognized as stwl alleles because they (1) carried insertions at 70D-E and (2) were female sterile, male fertile as homozygotes. An eighth allele (stwlRK) was kindly provided by R. Kelley (personal communication). All of these failed to complement the stwl6 allele.

P-elements were mobilized by standard means (Cooley et al., 1988) using a P[Δ2-3; ry+]99B, Sb chromosome (Robertson et al., 1988). Excised derivatives of the parental stwlP allele were recognized as ry flies in the F1 generation. These were used to establish balanced stocks and the new alleles were tested for homozygous phenotypes.

Microscopy and histochemistry

Ovary dissections and fixation was performed as described by Christerson and McKearin (1994). X-gal staining of enhancer trap encoded β-gal activity was performed as described by Cooley et al. (1992). For visualization of DNA using 4′,6-diamidino-2-phenylindole (DAPI), ovaries were fixed and then incubated with a 1 μg/ml solution of the reagent in PBS for 10 minutes. Ovaries were washed several times in PBS and mounted in 25% glycerol. The SPIF reagent (sonicated phenylenediamine-derived intense fluorochrome) was prepared according to Lundell and Hirsh (1994). Ovaries were fixed and rinsed several times in PBS+0.3% Triton X-100 to improve penetration of the reagent. The SPIF reagent was diluted 1:1 with PBS and used as a mounting medium for the ovaries. The fluorophore was visualized using a BioRad MRC600 confocal laser scanning microscope in the FITC channel. Rhodamine-labeled phalloidin was reacted with fixed ovaries as described by the manufacturer (Molecular Probes, Inc.).

Ovarian morphology was viewed on a Zeiss Axiophot microscope using DIC optics. Confocal images were collected from a Zeiss Axioplan microscope fitted with the Biorad MRC600 system and stored as digitized images. Images were either captured as prints on film or assembled directly in Adobe Photoshop for photographic presentation.

TdT-mediated dUTP-biotin nick end labeling (TUNEL) was carried out as described by Gavrieli et al. (1992) with modifications by K. Foley (Cooley lab; personal communication) and this laboratory. Ovaries were dissected in Drosophila Ringers and ovariole tips were teased apart. The tissue was fixed in 4% paraformaldehyde/1× PBT (1× PBS + 0.1% Tween 20) and heptane (1:3, vol/vol). After extensive washes in 1× PBS + 0.3% TritonX-100 (1× PBTx), ovaries were incubated in terminal transferase solution (1x TdT Boehringer Mannheim buffer; 2.5 mM CoCl2; 60 μM biotin-16-UTP; 200 nM dNTPs; 15 units TdT, final reaction volume 50 μl) for 3 hours at 37°C. Tissue was washed in 1× PBTx and blocked in 1× PBTx + 1.5% BSA for 1 hour at room temperature (RT). Incorporated biotin-dUTP was detected by incubation in streptavidin-texas red (1:2000; Molecular Probes, Inc.) in 1× PBTx + 1.5% BSA at RT in the dark for 2 hours. After extensive washing, tissue was mounted in 25% glycerol/1× PBTx and viewed by epifluorescence.

In situ hybridizations

The hybridizations were performed on whole tissue as detailed in Tautz and Pfeifle (1989) with modifications (Christerson and McKearin, 1994). All probes used were generated by digoxigenin-labeling of singlestranded PCR products from cloned cDNAs (Patel and Goodman, 1992). The orb probe spanned +17 to +1401 of orb cDNA (Lantz et al., 1992).

Molecular biology

All cloning and hybridization techniques were carried out as described by Sambrook et al. (1989) with exceptions noted below. Randomhexamer primed probes were generated according to Feinberg and Vogelstein (1984). Hybridization conditions and washes for Southern and northern blots were based on conditions described by Church and Gilbert (1984) as modified by McKearin and Spradling (1990).

The phage library of stwl6 DNA was constructed in the following manner. Genomic DNA was isolated from stwl6 flies and digested to completion with BamHI. Restriction fragments were cloned into EMBL3 arms, packaged with Gigapack extracts according to manufacturers instructions (Stratagene Corp.) and screened with a DNA probe corresponding to the 5′ P-end.

A 2.7 kb BamHI-PstI fragment which hybridizes only to the 3.7 kb transcript was used to screen a 0to 3-hour emybryonic cDNA library (Brown and Kafatos, 1988). A single clone was isolated and its identity was confirmed by Southern hybridization and restriction enzyme mapping. An internal PvuII fragment derived from this cDNA was used to probe the same cDNA library and a second cDNA clone was recovered.

RT-PCR

The RT-PCR reactions were carried out essentially as described by Kawasaki (1990), with the following modifications. Total RNA was isolated from 10 w1118 flies (or 20 stwl Δ95 flies) as described by McKearin and Spradling (1990). The DNA in these samples was digested with RNAse-free DNAse. For the annealing reaction, we used 200 ng of an antisense oligonucleotide corresponding to +439 to +410 of the stwl cDNA and 20 μg of RNA (Vt= 10 μl). After annealing, 16 μl of reverse transcriptase buffer was added (50 mM Tris, pH 8.4; 50 mM potassium acetate; 15 mM MgCl2; 10 mM DTT; 2 mM dNTPs; 1 μl AMV reverse transcriptase). The reaction was incubated at 48°C for 30 minutes; an additional 1 μl of reverse transcriptase was then added and incubation continued for another 30 minutes. The resulting product was then used in a standard PCR amplification (2 μl of cDNA per reaction).

Germline transformation

Injections into Drosophila embryos was performed essentially as described by Spradling (1986).

The stwl minigene was constructed by ligation of the stwl cDNA to 1.7 kb of upstream genomic DNA at a common SacI site. This was cloned into the transformation vector, pCaSpER4 (Pirrotta, 1988) and coinjected into w1118 embryos with pΔ2-3, a source of transposase (Robertson et al., 1988). Transformants were identified in the second generation as w+ adults. These chromosomes were crossed into a w1118; stwl1 background and assayed for rescuing activity.

P[hsp70-stwl; w+] was constructed by ligation of a NotI/HindIII fragment made blunt at the HindIII site to pCaSpER-hs (Thummel and Pirrotta, 1992) digested with HpaI/NotI. This construct was transformed in a similar manner to the minigene.

Heat shock treatments began with a warm-up step from 25°C to 38°C. Animals were kept at 38°C for an additional 30 minutes. For adults, these treatments were repeated on three successive days.

DNA sequencing and sequence analysis

The stwl cDNA was subcloned into Bluescript (KS+). Dideoxy sequencing of both strands of the clone was carried out on either single or double-strand templates using T3 and T7 promoter oligonucleotides or stwl primers.

Analysis of sequence data was performed using the GCG sequence analysis software package on the VAX computer at UT-Southwestern. Database searches were carried out using the BLAST program (Altschul et al., 1990). Predictions of secondary structure were made using the PhD program (Rost et al., 1994). Helical wheel plots were created using the GCG program HelicalWheel.

Production of polyclonal sera against Stwl protein

A His-tag fusion protein (pET28b; Invitrogen) contained sequences from the 3.5 kb cDNA, beginning at an EcoRI site at +1814. This protein was expressed in BL21(Lys3) and purified on an agarosenickel column according to manufacturer’s instructions. The protein was then dialyzed to remove urea and lyophilized. The protein was resuspended at 75 mg/ml in complete Freund adjuvent and injected intracardially. Rats were boosted at 14-day intervals. Animals were killed after the fourth bleed.

Stwl antisera and immunohistochemistry

Immunoblots were performed as described previously (Christerson and McKearin, 1994). In vitro transcription/translation of stwl and bam cDNAs was done according to manufacturer’s instructions using the TnT kit (Promega). Antisera were used at 1:2000 to 1:5000 dilutions and detected using secondary antibodies conjugated to alkaline phosphatase and the chemiluminescent substrate CSPD (Tropix). Three rat antisera were tested and all gave similar results. To assure that even low levels of mutant Stwl proteins could be detected, we determined empirically that 7 stwl ovaries, which produce trivial amounts of yolk proteins, yield as much non-yolk protein as 1 wild-type ovary.

Immunolocalization on w1118 and stwl mutant ovaries was performed as described previously (McKearin and Ohlstein, 1995).

Immunolocalizations in whole adult and larval preparations were performed similarly. Cuticles were opened at the midline to allow penetration of reagents and then treated like ovary preparations. AntiHts antisera for detecting the fusome were kindly supplied by H. Lipshitz and used at 1:100 dilutions. Orb monoclonal antibody 4H8, described by Lantz et al. (1994), was used at 1:40. BicD monoclonal antibody 1B11, described by Suter and Steward (1991) was used at 1:100. Stwl antisera were used at 1:5000 dilution. Antibodies were detected using Cy3-labeled secondary antibodies at 1:500 dilution (Jackson).

Immunodepletion experiments to test the specificity of Stwl antisera were performed as follows: Stwl antisera were diluted 1:5000 in PBT/0.1% BSA and incubated with 0.5 μg of either Stwl fusion protein (see antibody preparation) or Bam fusion protein (McKearin and Ohlstein, 1995) for 1 hour on a rocker prior to addition to tissue.

Identification of stwl mutations

We identified eight independent stwl P alleles (Table 1) from several collections of P-element mutagenesis screens (Materials and Methods). Chromosomal in situ hybridization showed that each carried a P-transposon at 70D-E; subsequent molecular analysis revealed that stwl5 carried 2 elements separated by ∽3 kb (not shown). Homozygous or hemizygous combinations produced essentially identical female sterile phenotypes; females failed to lay eggs, and ovaries remained rudimentary. Remarkably, stwl1–7 insertions are clustered within a single base-pair of one another (see Fig. 5 legend); stwl8 has not yet been sequenced.

Table 1.

stwl alleles recovered from several independent Pelement mutageneses

stwl alleles recovered from several independent Pelement mutageneses
stwl alleles recovered from several independent Pelement mutageneses

Chromosomal deletions that removed bands 70D-E were tested for complementation of stwl2 and stwl6. Df(3L)fzM21, deleted for 70D2–3;70E4–5, failed to complement either stwl allele. stwl locus cytological position within the 70D-E interval was further refined by complementation with two smaller deficiencies that affect Dichaete (Df(3L)D). Df(3L)D5, which removes 70D2–4, complemented both stwl alleles, while Df(3L)D5R+6, which removes 70C7-70D6, failed to complement. Thus the stwl locus lies between 70D4–70D6.

Additional proof that the P-elements caused stwl phenotype was obtained by transposase-dependent excision of a stwl Pelement (Materials and Methods). All flies that carried a molecularly wild-type stwl locus, reflecting precise transposon excision, were fertile. Thus there was 100% correspondence between precise excision of P[lacZ; ry+]70D-E and restoration of fertility, indicating that P-element insertions were responsible for the stwl phenotypes.

X-gal histochemistry with stwl alleles that carried enhancertrapping transposons which act as reporters for nearby transcriptional enhancers (Mlodzik and Hiromi, 1992) showed that β-gal expression was limited to germ cells (Fig. 2A). β-gal activity was first detectable in cystocytes in germarial Region 1, which contains mitotically active cystocytes, continued through Region 2b when cystocyte differentiation into nurse cells and oocytes is underway (Koch et al., 1966) and decayed between Region 3 and stage 3 egg chambers. β-gal activity was again detectable in nurse cells and oocytes of stage 10−11 egg chambers.

Fig. 2.

Phenotype of stwl mutant ovaries and expression of the stwl gene. (A) A micrograph of a stwl1/+ ovariole which has been stained with X-gal. Cystocytes in Region 2 of the germarium (upper left) express β-galactosidase strongly. X-gal staining decreases dramatically in post-germarial egg chambers but is again expressed strongly in nurse cells of stage 10 egg chambers. Nuclear localization of β-galactosidase activity is due to a nuclear localization signal that was fused to β-gal during cloning. The inset in the lower left presents a high magnification view of a stwl1/+ germarium reacted with X-gal to reveal the expression of β-gal from the enhancer trap P-element. The boundary between Region 1 and 2a is marked by an arrow. Note that the stwl enhancer trap reporter, lacZ, is transcriptionally active in Region 1 cystocytes. In B, a segment of a wild-type ovariole that includes the germarium through stage 8 chamber is shown for comparison with the mutant stwl phenotype seen in C. Note in C that chamber maturation to approximately stage 4 (penultimate chamber) appears superficially normal. However, the last egg chamber in the stwl ovariole shows the cellular blebbing that is one indication of the terminal stwl phenotype of germ cell degeneration. The inset shows a typical degenerating stwl chamber stained with SPIF (see Materials and Methods); note that nurse cell chromosomal DNA is highly condensed and, in some cases, fragmented (arrow).

Fig. 2.

Phenotype of stwl mutant ovaries and expression of the stwl gene. (A) A micrograph of a stwl1/+ ovariole which has been stained with X-gal. Cystocytes in Region 2 of the germarium (upper left) express β-galactosidase strongly. X-gal staining decreases dramatically in post-germarial egg chambers but is again expressed strongly in nurse cells of stage 10 egg chambers. Nuclear localization of β-galactosidase activity is due to a nuclear localization signal that was fused to β-gal during cloning. The inset in the lower left presents a high magnification view of a stwl1/+ germarium reacted with X-gal to reveal the expression of β-gal from the enhancer trap P-element. The boundary between Region 1 and 2a is marked by an arrow. Note that the stwl enhancer trap reporter, lacZ, is transcriptionally active in Region 1 cystocytes. In B, a segment of a wild-type ovariole that includes the germarium through stage 8 chamber is shown for comparison with the mutant stwl phenotype seen in C. Note in C that chamber maturation to approximately stage 4 (penultimate chamber) appears superficially normal. However, the last egg chamber in the stwl ovariole shows the cellular blebbing that is one indication of the terminal stwl phenotype of germ cell degeneration. The inset shows a typical degenerating stwl chamber stained with SPIF (see Materials and Methods); note that nurse cell chromosomal DNA is highly condensed and, in some cases, fragmented (arrow).

Phenotypic analysis of stwl mutants

In order to characterize the stwl phenotype more rigorously, we produced a null allele by transposase-induced imprecise excision of the P-element from stwl1 (Materials and Methods). One allele, stwlΔ95, represents the null phenotype; the molecular biology and immunochemistry that identifies Δ95 as a null allele will be presented in a later section.

In stwlΔ95 females, egg chamber growth (compare Fig. 2B and 2C) and nurse cell chromosome morphological changes (not shown) arrested between stages 4 and 7 (King, 1970). Examination of degenerating Δ95 chambers showed extensive blebbing in germ cells together with highly condensed and fragmented nurse cell and oocyte chromosomal DNA (Fig. 2C, inset). Since these are features of apoptotic cell death, we carried out additional tests for apoptosis by staining stwl ovaries with acridine orange and TUNEL labeling (Gavrieli et al., 1992; Materials and Methods). Both of these assays confirmed that the terminal phenotype for stwl egg chambers is germ cell apoptosis.

We found that β-gal from stwl enhancer traps (Fig. 2A), stwl RNA (not shown), and even Stwl protein (see below) was expressed in cystocytes in germarial Region 1, suggesting that Stwl might be required during early cystocyte stages. We were therefore surprised that stwl germ cells developed to oogenic stages 4-6 or rarely, stage 7. There are, however, precedents for oogenic mutations that act quite early during germ cell differentiation but that manifest a much later terminal phenotype. For example, BicD (Ran et al., 1994) and egalitarian (Schüpbach and Wieschaus, 1991) are required near the time that cystocyte divisions are completed but that show egg chamber arrest and subsequent degeneration at oogenic stage 6-7. We therefore assayed stwlΔ95 germ cells’ differentiation from cystoblasts to nurse cells/oocyte with probes for developmental landmarks of the germ cell lineage.

Immunohistochemical detection of Bam and Hts proteins, markers for early stages of cystocyte differentiation (McKearin and Ohlstein, 1995; Lin et al., 1994), suggested that initial stages of cyst formation were unaffected in stwl mutants. We used the dynamic morphological changes of the fusome to follow the pattern of cystocyte divisions. Reaction of anti-Hts antibodies, which stain the fusome, showed that the Δ95 cystocytes properly execute precisely four mitoses. We used anti-BamC antibodies, which recognize cytoplasmic Bam protein in cystocytes of 2-, 4and 8-cell clusters (McKearin and Ohlstein, 1995) as a second measure of proper cyst formation. The reaction of anti-BamC antibodies with Δ95 ovaries revealed that BamC recapitulates wild-type expression in the mutant.

Since initial germ cell differentiation into cystocytes appeared normal, we turned our attention to assays for nurse cell and oocyte differentiation. Light microscopic examination and confocal microscopy with DNA-specific fluorophores (DAPI and SPIF; see Materials and Methods) revealed that 38% of stwlΔ95 egg chambers contained 16 polyploid nurse cell nuclei and no oocyte nucleus (compare Fig. 3A,B with 3C,D). The clonal relatedness of germ cells in these 16 nurse cell chambers was confirmed by double labeling with rhodaminephalloidin (Materials and Methods), which binds to F-actin in the ring canals of intercellular bridges, and noting that they contained 15 ring canals. Thus, these germ cells had undergone the usual 4 cystocyte divisions that would produce 15 nurse cells and one oocyte but the presumptive oocyte sister had not become meiotic and had instead become polyploid.

Fig. 3.

stwl mutations block oocyte differentiation. A shows a projected Z-series of confocal images of a wildtype stage 8 and part of a stage 5-6 egg chamber stained with the DNAspecific dye, SPIF (Materials and Methods). Arrows indicate oocyte nuclei in order to point out the relative sizes and staining intensities between oocyte and nurse cell nuclei. B shows the same image as A in which the nurse cell and oocyte nuclei are outlined. C presents an approx. stage 6 stwlΔ95 chamber with 16 nurse cell nuclei stained with SPIF for comparison with wild type. Each of the nuclei has been circled in D to highlight the chromatin of each germ cell.

Fig. 3.

stwl mutations block oocyte differentiation. A shows a projected Z-series of confocal images of a wildtype stage 8 and part of a stage 5-6 egg chamber stained with the DNAspecific dye, SPIF (Materials and Methods). Arrows indicate oocyte nuclei in order to point out the relative sizes and staining intensities between oocyte and nurse cell nuclei. B shows the same image as A in which the nurse cell and oocyte nuclei are outlined. C presents an approx. stage 6 stwlΔ95 chamber with 16 nurse cell nuclei stained with SPIF for comparison with wild type. Each of the nuclei has been circled in D to highlight the chromatin of each germ cell.

Approximately 50% of Δ95 chambers contained 15 nurse cells and a single oocyte. The oocyte could be identified by its smaller nuclear volume and the appearance of four ring canals on its membrane. However, very often oocyte nuclear chromatin (i.e. the karyosome; King, 1970) did not have fully wild-type characteristics. For example, it was often larger than a wild-type karyosome and stained more intensely than its counterpart in wild-type oocytes of comparable developmental stage, suggesting partial polyploidy.

Eight percent of Δ95 chambers contained more than the usual sixteen germ cells; the most common supernumerary state was 32 cells. This suggested that cystocytes had undergone one additional round of mitosis in germarial Region 2; consistent with this conclusion, these egg chambers usually contained a single oocyte with five rhodamine-phalloidin stained ring canals.

The remaining small percentage of Δ95 chambers contained fewer than 16 germ cells. Chambers of this type were frequently adjacent to one another and, in some cases, the number of ring canals within the subnumerary chambers did not correlate with the number of germ cell nuclei. Often nuclei and ring canals could be found in the pinched region between adjacent chambers indicating that these chambers were not completely separated from one another. These observations are most consistent with breaking cystocyte clusters during follicle cell encystment rather than formation by fewer than 4 cystocyte divisions.

Oocyte differentiation within stwl cysts

The fact that we commonly found defective oocytes in stwl chambers suggested that oocyte determination and/or differentiation might be disrupted. Theurkauf et al. (1993) have shown that one early sign of oocyte differentiation is the formation of a polarized microtubule network that directs the asymmetric localization of several mRNAs and proteins into the oocyte. We used a probe for a localized transcript, orb, to test the efficiency of asymmetric transport in stwl mutant cysts. In wildtype germaria, orb transcripts accumulate in one cell, the presumptive oocyte, of each cyst as early as Region 2a (Fig. 4A; Lantz et al., 1992; Christerson and McKearin, 1994). Most often (approx. 90%) we found that orb transcript became concentrated in a single cell of stwl mutant cysts. Localization of orb transcripts was maintained in post-germarial cysts but was considerably more variable as chambers aged and most often stage 4 stwl egg chambers showed little or no asymmetric accumulation (Fig. 4B).

Fig. 4.

Oocyte differentiation in stwl cysts. A shows RNA in situ hybridization to a wild-type germarium with a probe for the orb transcript. Note that orb mRNA localizes to the oocyte in each cyst. B shows orb distribution in a stwlΔ95 ovariole. Note that orb mRNA usually localizes to a single cell, the presumptive oocyte, but that it occasionally accumulates inappropriately in two cells such as in the stage 2 chamber indicated by the double-headed arrow (shown at higher magnification in the inset). C presents an example of the distribution of Orb protein in a wild-type ovariole (also see Lantz et al., 1994). Orb abundance increases as cysts reach germarial Region 2a; the protein is distributed uniformly in newly formed 16-cell cysts but quickly begins to show concentration in one cell, the future oocyte, in each egg chamber. D shows the reaction of anti-Orb antibodies against a Δ95 ovariole. Although Orb protein abundance reaches approximately wild-type levels, the protein fails to become concentrated in any cystocyte. The asterisks in C and D mark the anterior end of the germaria.

Fig. 4.

Oocyte differentiation in stwl cysts. A shows RNA in situ hybridization to a wild-type germarium with a probe for the orb transcript. Note that orb mRNA localizes to the oocyte in each cyst. B shows orb distribution in a stwlΔ95 ovariole. Note that orb mRNA usually localizes to a single cell, the presumptive oocyte, but that it occasionally accumulates inappropriately in two cells such as in the stage 2 chamber indicated by the double-headed arrow (shown at higher magnification in the inset). C presents an example of the distribution of Orb protein in a wild-type ovariole (also see Lantz et al., 1994). Orb abundance increases as cysts reach germarial Region 2a; the protein is distributed uniformly in newly formed 16-cell cysts but quickly begins to show concentration in one cell, the future oocyte, in each egg chamber. D shows the reaction of anti-Orb antibodies against a Δ95 ovariole. Although Orb protein abundance reaches approximately wild-type levels, the protein fails to become concentrated in any cystocyte. The asterisks in C and D mark the anterior end of the germaria.

Fig. 5.

Restriction map and transcripts of the stwl locus. The schematic in A shows a simplified restriction map of the stwl locus (B = BamHI; H = HindIII; P = PstI) and the position of the mutagenic P-elements (filled triangle). The positions and approximate lengths of transcripts derived from the region are shown below the bar representing genomic DNA; transcripts are labeled α, β and γ. The position of the single intron in the α transcript is shown; the cDNA clone for α mRNA rescued stwl mutants and thus the α transcript represents the stwl mRNA (see text). DNA deleted by imprecise excision stwl alleles, stwlΔ61 and stwlΔ95, is shown above the genomic DNA; filled bars represent deleted DNA. (B) Expression of transcripts from the stwl locus. The northern blot contains poly(A)+ RNA from WT females (Lane 1; from 200 μg total RNA), WT males (Lane 2; from 250 μg total RNA), stwl1 (Lane 3; from 600 μg total RNA) and 0– 3 hour embryos (Lane 4; from 300 μg total RNA) probed with the 4.3 kb BamHI genomic fragment that carries stwl P-element insertion mutations. The probe recognizes 3 mRNA species in lane 1 at 3700, 1400 and 800 bases; the band at 800 bases was later shown to represent a doublet of mRNAs (transcripts γ1 and γ2 in Fig. 5A). The blot has been overexposed to reveal the low abundance 3700 nucleotide transcript.

Fig. 5.

Restriction map and transcripts of the stwl locus. The schematic in A shows a simplified restriction map of the stwl locus (B = BamHI; H = HindIII; P = PstI) and the position of the mutagenic P-elements (filled triangle). The positions and approximate lengths of transcripts derived from the region are shown below the bar representing genomic DNA; transcripts are labeled α, β and γ. The position of the single intron in the α transcript is shown; the cDNA clone for α mRNA rescued stwl mutants and thus the α transcript represents the stwl mRNA (see text). DNA deleted by imprecise excision stwl alleles, stwlΔ61 and stwlΔ95, is shown above the genomic DNA; filled bars represent deleted DNA. (B) Expression of transcripts from the stwl locus. The northern blot contains poly(A)+ RNA from WT females (Lane 1; from 200 μg total RNA), WT males (Lane 2; from 250 μg total RNA), stwl1 (Lane 3; from 600 μg total RNA) and 0– 3 hour embryos (Lane 4; from 300 μg total RNA) probed with the 4.3 kb BamHI genomic fragment that carries stwl P-element insertion mutations. The probe recognizes 3 mRNA species in lane 1 at 3700, 1400 and 800 bases; the band at 800 bases was later shown to represent a doublet of mRNAs (transcripts γ1 and γ2 in Fig. 5A). The blot has been overexposed to reveal the low abundance 3700 nucleotide transcript.

Occasionally (approx. 10% of germarial cysts) aberrant orb localization was observed; the most common manifestation of defective localization was orb mRNA accumulation in two cells (Fig. 4B, arrow and inset). These observations suggested that the process of oocyte selection or suppression of oocyte differentiation in all but one cystocyte was compromised in stwl cysts. Furthermore, assays which followed the progress of oocyte development indicated that oocyte differentiation was unstable. For example, BicD and Orb proteins which show asymmetric accumulation in the oocyte of Germarial Region 2b cysts (Suter and Steward, 1991; Lantz et al., 1994), failed to localize in all stwlΔ95 cysts. Fig. 4C and D compare Orb localization, which shows a steeper localization gradient than BicD, in wild-type and Δ95 cysts, respectively. These data indicated that the efficiency of maintaining asymmetric transport to the oocyte was compromised in stwl cysts.

Molecular genetics of the stwl locus

In order to clone the stwl locus, a fragment of DNA adjacent to the P-element in stwl6 was recovered from a phage library of completely digested DNA (Materials and Methods). This fragment recognized a 4.3 kb BamHI band on genomic Southern blots; the 4.3 kb BamHI band was altered in each stwl allele. Five cosmids spanning >50 kb (not shown) and which shared the 4.3 kb BamHI fragment were recovered from a wildtype genomic library (Fig. 5A). The insertion site for all stwl alleles mapped to the center of the 4.3 kb BamHI restriction fragment.

The polymorphic BamHI restriction fragment was used to probe northern blots containing poly(A)+ RNA from wild-type ovaries. Four transcripts (α = 3.7 kb, β = 1.4 kb, γ = 0.8 kb, later resolved into 2 transcripts) could be recognized (Fig. 5B). Transcripts α and β were present at approximately the same levels in RNA from females and males while the abundance of transcripts γ1 and γ2 was reduced in males. All three species were recovered from 0-3 hour embryo RNA samples. Hybridization to stwl1 poly(A)+ RNA from females showed reduced levels for all mRNAs. Simultaneous reduction of all mRNA species is most parsimoniously explained if α, β and γ transcripts are ovarian since the ovaries of stwl flies remain rudimentary. Elimination of hybridization to poly(A)+ RNA from gonadectomized carcasses confirmed the ovarian origin of transcripts α, β and γ (data not shown).

A 700 bp Pst1-HindIII genomic fragment (Fig. 5A), which is centered around the insertion site for stwlP alleles hybridized only to the 3700 nucleotides α transcript. In addition, only transcription unit α was physically disrupted by stwl P-element insertions. We therefore suspected that the α transcript represented the stwl gene, and screened a cDNA library for full length clones that corresponded to this mRNA (Materials and Methods). Two clones of approx. 3600 bases were recovered; each encodes a single long open reading frame (ORF).

The transcriptional complexity of the putative stwl locus made it necessary to identify the stwl gene unambiguously by germline transformation rescue. For this purpose, an α cDNA clone was joined to a 1.7 kb BamHI-PstI genomic fragment immediately adjacent to the putative α start site (Fig. 5A). We reasoned that, if this fragment contained the stwl promoter, it would direct transcription of α-cDNA in a pattern that recapitulated authentic stwl expression. Females that carried this putative ‘stwl minigene’ in a stwl mutant background were viable and fully fertile. Ovaries dissected from P[w+; stwl minigene]/+; stwl/stwl flies were wild type in all aspects of oogenesis.

Results of minigene rescue suggested that transcript α represented the authentic stwl gene. However the possibility remained that minigene rescuing activity was provided by one of the other transcripts ( β, γ1 or γ2) potentially encoded by the 1.7 kb of genomic DNA that was used as a promoter in the stwl minigene (Fig. 5A). This possibility was addressed with a second construct containing transcript α cDNA driven by the hsp70 promoter. Even without heat induction, this construct provided weak rescue of stwl mutants. An average of 14 eggs were laid per fly and 30% of these eggs hatched into viable larvae. Ovaries dissected from P[w+; hsp70-transcript α]/+; stwl/stwl females contained many fully developed egg chambers; these are never found in stwl/stwl ovaries.

Surprisingly, heat induction of P[w+; hsp70-stwl] animals did not improve the degree of rescue. This could be due to poor inducibility of the hsp70 promoter in germline cells (P. Schedl, personal communication). However an alternative explanation arose when we observed that heat induction of the hsp70-stwl transgene produced dominant lethality. A single short heat pulse delivered to third instar larvae was sufficient to kill them within 24 hours; adults were killed by 30 minute heat shocks delivered on three successive days. Thus we suspect that failure to improve rescue with the induced hsp70-stwl transgene is due to deleterious effects of ectopic Stwl expression in somatic cells.

The stwl gene product

Fig. 6 shows the sequence of the longest stwl cDNA and the predicted amino acid sequence of the single long ORF. A putative start site of translation is found at +162 and is followed by an ORF of 1043 amino acids; conceptual translation of this ORF predicts a protein of 113×103Mr with a pI of 10.2. In addition to a putative DNA binding domain (see below), the predicted Stwl protein (Fig. 7A) contains a highly acidic domain (aa 182– 237), a stretch of reiterations of Ser/Thr residues (aa 474 – 586) and several regions with a high content of basic amino acids (Fig. 7A).

Fig. 6.

Nucleotide and predicted protein sequence of Stwl. This figure presents the DNA sequence and predicted open reading frame of the longest stwl cDNA; the first nucleotide of a second cDNA clone corresponds to position 20. Numbers on the left apply to nucleotides while those on the right apply to amino acid residues. Amino acids that correspond to an acidic domain are printed in bold type and residues in a Ser/Thr domain are printed in italics. A consensus poly(A)+ addition signal is printed in bold, italic text at nucleotides 3586-3591. The P-element insertion site for stwl1 and stwl3-7 is between nucleotides 281 and 282; the P-element in stwl2 is between nucleotides 282 and 283.

Fig. 6.

Nucleotide and predicted protein sequence of Stwl. This figure presents the DNA sequence and predicted open reading frame of the longest stwl cDNA; the first nucleotide of a second cDNA clone corresponds to position 20. Numbers on the left apply to nucleotides while those on the right apply to amino acid residues. Amino acids that correspond to an acidic domain are printed in bold type and residues in a Ser/Thr domain are printed in italics. A consensus poly(A)+ addition signal is printed in bold, italic text at nucleotides 3586-3591. The P-element insertion site for stwl1 and stwl3-7 is between nucleotides 281 and 282; the P-element in stwl2 is between nucleotides 282 and 283.

Fig. 7.

Features of protein domains in the predicted Stwl sequence. (A) A schematic view of the predicted Stwl protein. A putative DNA binding domain (HTH) at the N-terminal end is labeled as is an acidic-rich domain (indicated by its net charge, –23) and a region rich in Ser and Thr amino acids (S/T; 46 S/T in 112 amino acids). The other regions of the predicted protein carry net positive charge which is shown above the schematic; numbers in parentheses are the number of H+R+K residues/total residues. In B, the amino acid sequence of the Stwl putative HTH motif is shown aligned with the HTH motif from the Drosophila Adf-1 transcription factor and related HTH regions from Drosophila c-Myb proteins (Adf-1 and c-Myb alignments taken from England et al., 1992). The assignment of residues to Helix 1 and Helix 2 together with the region linking the putative helices is based on data from England et al. (1992). An arrow in Helix 1 indicates the site of stwl Pelement insertions. (C) A display of segments of the putative DNA binding domains of Stwl and Adf-1 as helical wheel schematics. Hydrophobic residues are printed in bold, large font. Note that for both Stwl and Adf-1, hydrophobic residues cluster on one side of the predicted helix, presenting distinctly amphipathic faces.

Fig. 7.

Features of protein domains in the predicted Stwl sequence. (A) A schematic view of the predicted Stwl protein. A putative DNA binding domain (HTH) at the N-terminal end is labeled as is an acidic-rich domain (indicated by its net charge, –23) and a region rich in Ser and Thr amino acids (S/T; 46 S/T in 112 amino acids). The other regions of the predicted protein carry net positive charge which is shown above the schematic; numbers in parentheses are the number of H+R+K residues/total residues. In B, the amino acid sequence of the Stwl putative HTH motif is shown aligned with the HTH motif from the Drosophila Adf-1 transcription factor and related HTH regions from Drosophila c-Myb proteins (Adf-1 and c-Myb alignments taken from England et al., 1992). The assignment of residues to Helix 1 and Helix 2 together with the region linking the putative helices is based on data from England et al. (1992). An arrow in Helix 1 indicates the site of stwl Pelement insertions. (C) A display of segments of the putative DNA binding domains of Stwl and Adf-1 as helical wheel schematics. Hydrophobic residues are printed in bold, large font. Note that for both Stwl and Adf-1, hydrophobic residues cluster on one side of the predicted helix, presenting distinctly amphipathic faces.

Comparison of the Stwl ORF to other proteins identified a region of homology to the Drosophila transcription factor Alcohol dehydrogenase distal factor 1 (Adf-1; England et al., 1992). The aligned region (Fig. 7B) corresponds in part to a region of Adf-1 that Tjian and colleagues had previously identified as similar to c-Myb proteins (England et al., 1992). This domain contains a motif recognized by secondary structure algorithms as a helix-turn-helix (HTH) DNA binding motif found in a variety of bacterial and eukaryotic DNA binding proteins (Harrison and Aggarwal, 1990). The putative HTH domain in Stwl conserves amino acids that have been shown to be critical for Myb DNA binding and function (Frampton et al., 1991). Particularly significant are Trp39 and Trp57, which are apparently involved in packing the hydrophobic core of αhelices, and Gly47, which is thought to be critical for allowing the formation of a very tight turn between two helices.

Stwl and Adf-1 also share similarities of predicted secondary structure that are probably significant for function as DNA binding proteins. For example, Helix 2 is reported to be the ‘recognition helix’ of the HTH motif because it inserts into the major groove of B-form DNA (Harrison and Aggarwal, 1990). In Adf-1, Helix 2 forms an amphipathic helix with non-polar and hydrophobic residues clustered on one face, and charged and polar residues on the other. Interactions of the hydrophobic face are thought to be important for positioning the helix for insertion into the major groove while the charged face contains residues that make specific contacts with nucleotide bases in the major groove at the protein’s recognition site. Fig. 7C demonstrates by ‘helical wheel’ plotting that, like Adf-1 Helix 2, Stwl Helix 2 forms a strongly amphipathic helix.

Conservation of amino acid sequence between Stwl and Adf1 extends to the N-terminal side of the putative HTH domain (Fig. 7B). Although the significance of these similarities is not yet understood biochemically, G. Cutler reports (personal communication) that deletions in the Stwl-related region of the Adf-1 N terminus reduce or eliminate DNA binding.

Analysis of Stwl protein by immunoblots

Antibodies were raised against the C-terminal half of Stwl overexpressed in bacteria as His-tagged fusion proteins. The specificity of antisera was tested on immunoblots containing in vitro translated (IVT) Stwl proteins and ovarian protein extracts. Stwl antisera detected a band of approx. 150,000 Mr in IVT extracts that had been programmed with full-length stwl cDNA (Fig. 8, lane 1) but did not react with any proteins in IVT extracts programmed with bam mRNA (not shown). Although 150,000 Mr is larger than the calculated relative molecular mass of 113,000 for Stwl, the highly charged nature of the protein might account for aberrant migration. A band of similar relative molecular mass was found in extracts from wild-type, although ovarian Stwl reproducibly migrated slightly faster than IVT-Stwl (Fig. 8, lane 2).

Fig. 8.

Immunoblot analysis of Stwl antigens. Western blot showing reaction of rat anti-Stwl antisera to IVT-Stwl (Lane 1), wild-type ovarian extracts from 1.5 ovaries (Lane 2), Δ61 ovarian extracts from 70 ovaries (Lane 3) and Δ95 ovarian extracts from 70 ovaries (Lane 4). The position of molecular weight markers is indicated on the left side of the blot.

Fig. 8.

Immunoblot analysis of Stwl antigens. Western blot showing reaction of rat anti-Stwl antisera to IVT-Stwl (Lane 1), wild-type ovarian extracts from 1.5 ovaries (Lane 2), Δ61 ovarian extracts from 70 ovaries (Lane 3) and Δ95 ovarian extracts from 70 ovaries (Lane 4). The position of molecular weight markers is indicated on the left side of the blot.

As a further test of antibody specificity, we reacted antiStwl antisera with ovarian extracts from several stwl alleles that we suspected carried inactivating mutations. Immunoblots against protein isolated from the previously mentioned Δ95 allele showed that no immunoreactive Stwl was detectable. Molecular analysis of the Δ95 transcript, by RT-PCR, revealed that this allele produces a mRNA that carries both a deletion of wild-type sequences and an insertion of novel sequences. The net effect of this lesion is the introduction of a stop codon at amino acid 43 and a frameshift of the normal Stwl ORF. These investigations demonstrated that Δ95 is a complete loss-of-function for stwl. A second allele, Δ61, deletes DNA to the 5′ side of the P-element insertion site and removes the transcriptional start site as well as approx. 700 bp from the putative stwl promoter region (Fig. 5A). This allele produces low levels of transcript and immunoblots containing Δ61 extracts showed immunoreactive Stwl protein (Fig. 8, lane 3). Significantly, Δ61 protein migrates more rapidly than wild-type Stwl in denaturing gels suggesting that the protein might be truncated.

Immunolocalization in ovaries

Immunohistochemistry with Stwl antisera against wild-type ovaries revealed that Stwl is a nuclear protein expressed in germ cells from stem cells through nurse cell and oocyte nuclei of stage 7 egg chambers (Fig. 9A,B). Thus Stwl protein can be detected slightly earlier (i.e. in stem cells) than either β-gal activity from stwl enhancer trap alleles (Fig 2A) or stwl RNA (not shown) first appear. The anti-Stwl immunofluorescent signal in cystocyte nuclei in Germarial Region 1 is fainter than that seen in nuclei in Region 2 (Fig. 9B), suggesting that Stwl levels might be lower in mitotically active cystocytes than in post-mitotic nuclei. Both RNA in situ hybridization and enhancer trap activity indicate that stwl transcription is reinitiated in stage 8 11 germ cells but no immunoreactive Stwl can be detected in those cells. Thus it appears that transcripts accumulated in late stage egg chambers might not be translated during oogenesis.

Fig. 9.

Immunohistochemistry with Stwl antisera. (A) The reaction of stwl antisera against wild-type ovaries. A germarium is in the upper left and a stage 6 egg chamber in the lower right. Note that Stwl antigen is localized to the nuclei of germ cells; the images have been overexposed to show the outlines of the ovariole. (B) A higher magnification of a wild-type germarium which has been outlined manually. The filled arrow points to a putative germline stem cell while the open arrow indicates the boundary between Regions 1 and 2a. The reaction of Stwl antisera against a Δ95 ovariole is shown in C and demonstrates that no immunoreactive protein is present. This ovariole begins with a germarium in the upper left and ends with a degenerating stage 5 chamber in the lower right. D and E show the reaction of Stwl antisera in larval gut. Heatshocked wild-type larvae (D, section of gut and caecae) do not produce immunoreactive Stwl, but larvae transgenic for an hsp70-stwl gene (E) produce Stwl in abundance and the protein localizes to the nucleus. The inset in the upper left shows a section of the transgenic gut caecae where Stwl nuclear localization is particularly apparent.

Fig. 9.

Immunohistochemistry with Stwl antisera. (A) The reaction of stwl antisera against wild-type ovaries. A germarium is in the upper left and a stage 6 egg chamber in the lower right. Note that Stwl antigen is localized to the nuclei of germ cells; the images have been overexposed to show the outlines of the ovariole. (B) A higher magnification of a wild-type germarium which has been outlined manually. The filled arrow points to a putative germline stem cell while the open arrow indicates the boundary between Regions 1 and 2a. The reaction of Stwl antisera against a Δ95 ovariole is shown in C and demonstrates that no immunoreactive protein is present. This ovariole begins with a germarium in the upper left and ends with a degenerating stage 5 chamber in the lower right. D and E show the reaction of Stwl antisera in larval gut. Heatshocked wild-type larvae (D, section of gut and caecae) do not produce immunoreactive Stwl, but larvae transgenic for an hsp70-stwl gene (E) produce Stwl in abundance and the protein localizes to the nucleus. The inset in the upper left shows a section of the transgenic gut caecae where Stwl nuclear localization is particularly apparent.

Consistent with results obtained from immunoblots, incubation of anti-Stwl antisera with Δ95 ovaries showed complete elimination of immunopositive antigen (Fig. 9C). Immunopositive reaction could also be extinguished by preadsorption of Stwl antisera with bacterially overexpressed Stwl (not shown; see Materials and Methods) before incubation with wild-type ovaries; preadsorption with an unrelated bacterial fusion protein (Bam; McKearin and Ohlstein, 1995) had no effect on anti-Stwl immunoreaction.

As further confirmation of antisera specificity and to investigate the basis of dominant lethality caused by P[w+; hsp70stwl] transgenes, Stwl antibodies were incubated with heatshocked transgenic larvae. Stwl is undetectable in most somatic cells of wild-type (Fig. 9D) or non-heat-shocked transgenic (not shown) larvae. We have noted very faint staining of some somatic nuclei with anti-Stwl antisera but have not determined if this reaction represents Stwl or a related protein. However, after a brief heat induction, immunoreactive protein appeared in nuclei of most tissues examined; an example of such an experiment is shown in Figure 9E with 3rd instar larval gut cells. In heat-shocked transgenic salivary gland cells, Stwl can be found associated with the polytene chromosomes (not shown) suggesting that Stwl is DNA bound and not simply a nucleoplasmic protein.

Stwl is probably a transcriptional activating protein

The prediction by secondary structure algorithms of a helixturn-helix motif in Stwl was strengthened by finding amino acid conservation to helices α2 and α3 of the well characterized HTH domain of Myb oncoproteins (Fig. 7B; Frampton et al., 1991). Especially significant was the conservation of several residues that were shown by mutagenesis to be essential for Myb DNA binding activity (Frampton et al., 1991). Extensive studies of many HTH proteins have identified the second helix, sometimes termed the ‘recognition helix’, as the part of the HTH motif that inserts into the major groove of DNA to make specific contacts between amino acid side chains and nucleotides (Harrison and Aggarwal, 1990). Consistent with DNA binding activity, immunolocalization showed that Stwl is a germ cell nuclear protein.

Comparison of the Stwl sequence with other proteins revealed similarity to another Drosophila HTH protein, Adf1, a transcription factor that binds to a cis-activating element in the Adh gene distal promoter (Heberlein et al., 1985). Purification of the Adf-1 protein demonstrated that it binds to the promoters of several other genes including Antp and Dopa decarboxylase (Ddc), each of which contains a version of a consensus binding element (England et al., 1990). Stwl is more similar to Adf-1 than Myb; amino acid matches between Stwl and Adf-1 extended to the N terminal side of the HTH motif, and conservation of residues within HTH domains was greater than that found between Stwl and Myb (Fig. 7B). Secondary structural predictions of the sequences preceding the Stwl and Adf-1 HTH motifs suggest that these residues might form a third helix. Investigations of a large number of HTH proteins indicates that many form a third helix that can precede or follow the HTH motif helices. Mutagenesis studies (Frampton et al., 1991; Harrison and Aggarwal, 1990) have demonstrated that these third helices can also be involved in recognition site binding. Cutler reports that deletions in the sequences preceding the Adf-1 HTH motif eliminate DNA binding (personal communication); this region is extensively conserved between Stwl and Adf-1 (Fig. 7B) and, perhaps, is also important for Stwl DNA recognition.

In addition to amino acid identities and conservative substitutions, Stwl and Adf-1 share an important feature of predicted secondary structures: namely that Helix 2 in both is strongly amphipathic (Fig. 7C). The hydrophobic face of the helix is thought to be involved in protein-protein interactions that align the helix for presentation of the amino acids on the charged helix face to DNA (Harrison and Aggarwal, 1990). We expect that mutagenesis of residues that lie in this hydrophobic face could have dramatic effects on Stwl DNA binding activity and specificity.

Despite the considerable similarities between Stwl and Adf1, the two proteins do not share any significant sequence matches outside of their N-terminal domains. Stwl is a much larger protein and contains a large, strongly acidic region that might act as a transcriptional activating domain (Figs 6 and 7A); Adf-1 does not have any comparable acidic-rich domain (England et al., 1992). From analysis of many DNA proteins, blocks of highly acidic residues have been recognized as one of the major types of transcriptional activating domains (Mitchell and Tjian, 1989; Roberts and Green, 1994). In addition to an ‘acid patch’, Stwl contains a region with a high concentration of Ser/Thr residues (Figs 6 and 7A). Although the function of this region is unknown, we speculate that it could be the target for extensive phosphorylation similar to the Ser/Thr-rich carboxy terminal domain (CTD) of RNA Pol II (Conaway and Conaway, 1993). In this context, we have observed that Stwl from in vivo sources consistently migrated faster than Stwl produced by in vitro translation (Fig. 8); phosphorylation of this large Ser/Thr region could account for altered electrophoretic migration. An alternative explanation is that Stwl is subject to specific proteolysis in ovaries. Either modification might represent a means of controlling Stwl activity.

The stwl phenotype suggests early function in cystocytes

Our genetic and immunological investigations allowed an assessment of when stwl+ activity was required during oogenesis. Although Δ95 ovaries contain readily detectable amounts of stwl RNA, no Stwl protein could be detected either on immunoblots or by immunohistochemistry, suggesting strongly that the allele represents complete loss of Stwl function. The analysis of Δ95 by RT-PCR confirmed that this allele represented a molecular null.

The fact that the null allele Δ95 did not affect viability or male fertility and caused only ovarian phenotypes, together with observations that Stwl protein was restricted to germ cells, leads us to conclude that, in the adult fly, stwl+ is essential only in oogenic germ cells. We have found stwl transcripts in the embryo, probably as a maternally inherited RNA since levels are highest from 0-2 hours of embryogenesis (Clark and McKearin, unpublished data), and have also noted faint staining with Stwl antibodies in some somatic nuclei. Therefore, we cannot exclude the possibility that Stwl is utilized at other times during development.

Stwl was detectable in nuclei of all germ cells from stem cells through stage 7 and stwl egg chambers showed a range of defects that affected germ cell differentiation between postmitotic cystocytes and stage 7 germ cells. To define more precisely when stwl+ is required, we followed the progress of oogenesis in stwl null mutant germ cells using a variety of markers.

The presence of Stwl in all germ cells raised the possibility that stwl+ was required for the earliest stages of germ cell differentiation, namely stem cells and cystoblasts. However, stem cells continued to divide and produced appropriately differentiating cystoblasts in stwlΔ95 ovaries as evidenced by a continuous supply of germ cells in cysts. Cystoblast and cystocyte differentiation also appeared normal as assayed by accumulation of BamC in dividing cystocytes (McKearin and Ohlstein, 1995) and by the morphogenesis of the fusome (Lin et al., 1994). The fact that fusomes in stwl cells appeared normal was especially significant since elimination of the fusome has been correlated with poor oocyte differentiation (Yue and Spradling, 1992; Lin et al., 1994; Lin and Spradling, 1995).

The most striking phenotype in stwl chambers was the failure of oocyte differentiation. Morphologically, defects in the maintenance of oocyte identity were observed as sixteen nurse cell nuclei or fifteen nurse cells and one partially polyploid nucleus. These inappropriate levels of chromosome ploidy were detectable as early as stage 3 egg chambers. However, it was possible to find evidence of defective oocyte differentiation in much younger stwl cells. Although most often orb transcript localized to one pro-oocyte in Region 2a of stwl germaria, occasionally we found orb mRNA in 2 cystocytes, hinting that the mechanism(s) for oocyte selection was compromised. This conclusion was strengthened when antibodies against oocyte protein markers were used to test the efficiency of oocyte differentiation. Orb and BicD normally form accumulation gradients directed at the oocyte beginning in Region 2b cysts (Lantz et al., 1994; Suter and Steward, 1991). Both of these proteins remained distributed homogenously in stwl cysts, indicating that oocyte identity was poorly maintained as early as Region 2b cysts. These data also contribute support to the previously proposed hypothesis that polarized transport of mRNAs and proteins to the oocyte proceeds by independent mechanisms (Suter and Steward, 1991).

Although it appeared only occasionally, the production of egg chambers with 32 germ cells suggested that Stwl might have functions as early as mitotic cystocytes. When it was possible to count ring canals in these 32-cystocyte chambers accurately, the total number of ring canals and the appearance of germ cells with 5 ring canals indicated that stwl cystocytes sometimes undergo an extra round of mitosis. The low expressivity of the phenotype has prevented us from determining when the extra round of division takes place or even it happens at a reproducible point of cystocyte divisions. However, we postulate that Stwl is active in dividing cystocytes and its effects on cystocyte differentiation include maintaining an accurate counting mechanism that limits the number of mitoses to four.

The terminal phenotype for stwl germ cells was apoptotic death. Extensive cellular blebbing, DNA staining that showed condensed chromosomal DNA and TUNEL labeling of fragmented nuclear DNA in germ cells convincingly showed apoptosis between stages 4 and 7. Apoptosis has not been described in other female sterile mutations which produce 16 nurse cells and no oocyte, such as BicD and egalitarian (Ran et al., 1994; Schüpbach and Wieschaus, 1991), and our own observations of egalitarian ovaries suggest that germ cells do not degenerate apoptically (Clark and McKearin, unpublished data). Therefore failure to produce an oocyte within an egg chamber does not induce apoptosis in germ cells a priori; rather our data suggest that stwl+ is specifically necessary to prevent apoptosis.

A proposal for Stwl function

The diversity of stwl phenotypes, such as extra cystocyte division and failed oocyte differentiation, suggests that the protein is active at different times while variable expressivity indicates that Stwl-dependent gene expression is important but not essential for these aspects of germ cell differentiation. Since Stwl is probably a transcription factor, we propose that it regulates genes that function at several stages of cyst formation and oocyte development. Partial functional redundancy for Stwl protein might account for the variable severity of stwl phenotypes. Redundancy could take the form of another HTH-family protein that can partially substitute for Stwl or incomplete transcriptional inactivation of Stwl-responsive genes’ promoters due to combinatorial regulation, similar to that reported for several of the early zygotic transcription factors (Pankratz and Jackle, 1993; Benedyk et al., 1994; Gray et al., 1994).

We propose that in its role as an effector of oocyte differentiation, Stwl responds to oocyte determination signals and acts as a transcriptional regulator in nurse cells of genes required to promote oocyte differentiation. In the absence of stwl+, oocyte differentiation cannot be maintained efficiently and the presumptive oocyte escapes to develop as a nurse cell; eventually the lack of Stwl-dependent gene expression causes all germ cells to activate apoptosis. Thus we anticipate that isolation of Stwl target genes will identify molecules that promote oocyte development; testable classes of candidate genes might include proteins required for maintaining the integrity of the MT-network or meiotic proteins.

A. Carpenter and R. Kelley kindly provided fly stocks. R. Kelley generously made available a cosPNeo genomic library. H. Lipshitz graciously provided anti-Hts antibodies before publication. Many thanks to S. Miklausz of the UI Immunological Resources Center who provided expert technical assistance in raising anti-Stwl antibodies. M. Kuhn also provided excellent technical assistance for immunohistochemistry and fly culture. The authors extend sincere appreciation to G. Cutler (Tjian lab) for sharing results of mutagenesis experiments on Adf-1 before publication and useful conversations about HTH domains. Former and present members of the McKearin lab contributed valuable comments and criticism at various stages of this work. The authors thank L. Cooley, H. Krämer and S. Wasserman for comments on the various drafts of this manuscript and two anonymous reviewers for thoughtful suggestions for improving the manuscript. This work was supported by NIH grant GM45820 to D. M.

Altschul
,
S. F.
,
Gish
,
W.
,
Miller
W.
,
Myers
,
E. W.
and
Lipman
,
D. J.
(
1990
).
Basic local alignment search tool
.
J. Mol. Biol
.
215
,
403
410
.
Artavanis-Tsakonas
,
S.
,
Matsuno
,
K.
and
Fortini
,
M. E.
(
1995
).
Notch signaling
.
Science
268
,
225
232
.
Benedyk
,
M. J.
,
Mullen
,
J. R.
and
DiNardo
,
S.
(
1994
).
odd-paired: a zinc finger pair-rule protein required for the timely activation of engrailed and wingless in Drosophila embryos
.
Genes Dev
.
8
,
105
117
.
Berg
,
C. A.
and
Spradling
,
A. C.
(
1991
).
Studies on the rate and sitespecificity of P element transposition
.
Genetics
127
,
515
524
.
Brown
,
E. H.
and
King
,
R. C.
(
1964
).
Studies on the events resulting in the formation of an egg chamber in Drosophila melanogaster
.
Growth
28
,
41
81
.
Brown
,
N. H.
and
Kafatos
,
F. C.
(
1988
).
Functional cDNA libraries from Drosophila embryos
.
J. Mol. Biol
.
203
,
425
437
.
Carpenter
,
A.
(
1994
).
egalitarian and the choice of cell fates in Drosophila melanogaster oogenesis
.
In CIBA Foundation Symposium 182, Germline Development
(ed.
J.
Marsh
and
J.
Goode
), pp.
223
254
.
New York
:
John Wiley and Sons Ltd
.
Christerson
,
L. B.
and
McKearin
,
D. M.
(
1994
).
orb is required for anteroposterior and dorsoventral patterning during Drosophila oogenesis
.
Genes Dev
.
8
,
614
628
.
Church
,
G. M.
and
Gilbert
.,
W.
(
1984
).
Genomic sequencing
.
Proc. Natl. Acad. Sci. USA
81
,
1991
1995
.
Conaway
,
R. C.
and
Conaway
,
J. W.
(
1993
).
General initiation factors for RNA polymerase II
.
Ann. Rev. Biochem
.
62
,
161
190
.
Cooley
,
L.
,
Berg
,
C.
and
Spradling
,
A. C.
(
1988
).
Controlling P element insertional mutagenesis
.
Trends Genet
.
4
,
254
258
.
Cooley
,
L.
,
Verheyen
,
E.
and
Ayers
,
K.
(
1992
).
chickadee encodes a profilin required for intercellular cytoplasm transport during Drosophila oogenesis
.
Cell
69
,
173
184
.
England
,
B. P.
,
Heberlein
,
U.
and
Tjian
,
R.
(
1990
).
Purified Drosophila transcription factor, Adh distal factor-1 (Adf-1), binds to sites in several Drosophila promoters and activates transcription
.
J. Biol. Chem
.
265
,
50865094
.
England
,
B. P.
,
Admon
,
A.
and
Tjian
,
R.
(
1992
).
Cloning of Drosophila transcription factor Adf-1 reveals homology to Myb oncoproteins
.
Proc. Natl. Acad. Sci. USA
89
,
683
687
.
Feinberg
,
A. P.
and
Vogelstein
,
B.
(
1984
).
Addendum, A technique for radiolabeling DNA restriction endonuclease fragments to high specific activity
.
Analyt. Biochem
.
137
,
266
267
.
Frampton
,
J.
,
Gibson
,
T. J.
,
Ness
,
S. A.
,
Doderlein
,
G.
and
Graf
,
T.
(
1991
).
Proposed structure for the DNA-binding domain of the Myb oncoprotein based on model building and mutational analysis
.
Protein Engineering
4
,
891
901
.
Gavrieli
,
Y.
,
Sherman
,
Y.
and
Ben-Sasson
,
S. A.
(
1992
).
Identification of programmed cell death via specific labeling of nuclear DNA fragmentation
.
J. Cell Biol
.
119
,
493
501
.
Gray
,
S.
,
Szymanski
,
P.
and
Levine
,
M.
(
1994
).
Short-range repression permits multiple enhancers to function autonomously within a complex promoter
.
Genes Dev
.
8
,
1829
1838
.
Harrison
S. C.
and
Aggarwal
,
A. K.
(
1990
).
DNA recognition by proteins with the helix-turn-helix motif
.
Ann. Rev. Biochem
.
59
,
933
969
.
Heberlein
,
U.
,
England
,
B.
and
Tjian
,
R.
(
1985
).
Characterization of Drosophila transcription factors that activate the tandem promoters of the Alcohol Dehydrogenase gene
.
Cell
41
,
965
977
.
Karpen
,
G. H.
and
Spradling
,
A. C.
(
1992
).
Analysis of subtelomeric heterochromatin in the Drosophila minichromosome Dp1187 by single P element insertional mutagenesis
.
Genetics
132
,
737
753
.
Kawasaki
,
E. S.
(
1990
).
Amplification of RNA
.
In PCR Protocols: A Guide to Methods and Applications
(ed.
M. A.
Innis
,
D. H.
Gelfand
,
J. J.
Sninsky
, and
T. J.
White
), pp.
21
27
.
San Diego, California
:
Academic Press
.
King
,
R. C.
(
1970
).
Ovarian Development in Drosophila melanogaster
.
New York
:
Academic Press
.
King
,
R. C.
and
Storto
,
P. D.
(
1988
).
The role of the otu gene in Drosophila oogenesis
,
BioEssays
8
,
18
24
.
Koch
,
E. A.
,
Smith
,
P. A.
and
King
,
R. C.
(
1966
).
The division and differentiation of Drosophila cystocytes
.
J. Morphol
.
121
,
55
70
.
Lantz
,
V.
,
Ambrosio
,
L.
and
Schedl
,
P.
(
1992
).
The Drosophila orb gene is predicted to encode sex-specific germline RNA-binding proteins and has localized transcripts in ovaries and early embryos
.
Development
115
,
75
88
.
Lantz
,
V.
,
Chang
,
J. S.
,
Horabin
,
J. I.
,
Bopp
,
D.
and
Schedl
,
P.
(
1994
).
The Drosophila orb RNA-binding protein is required for the formation of the egg chamber and estblishment of polarity
.
Genes Dev
.
8
,
598
613
.
Lin
,
H.
,
Yue
,
L.
and
Spradling
,
A. C.
(
1994
).
The Drosophila fusome, a germline specific organelle, contains membrane skeletal proteins and functions in cyst formation
.
Development
120
,
947
956
.
Lin
,
H.
and
Spradling
,
A. C.
(
1995
).
Fusome asymmetry and oocyte determination in Drosophila
.
Dev. Genet
.
16
,
6
12
.
Lundell
,
M. J.
and
Hirsh
,
J.
(
1994
).
A new visible light DNA fluorochrome for confocal microscopy
.
Biotechniques
16
,
434
440
.
McKearin
,
D. M.
and
Spradling
,
A. C.
(
1990
).
bag-of-marbles; A Drosophila gene required to initiate male and female gametogenesis
.
Genes Dev
.
4
,
2242
2251
.
McKearin
,
D. M.
and
Ohlstein
,
B.
(
1995
).
A role for the Drosophila Bag-ofmarbles protein in the differentiation of cystoblasts from germline stem cells
.
Development
121
,
2937
2947
.
Mitchell
,
P. J.
and
Tjian
,
R.
(
1989
).
Transcriptional regulation in mammalian cells by sequence-specific DNA binding proteins
.
Science
245
,
371
378
.
Mlodzik
,
M.
and
Hiromi
,
Y.
(
1992
).
The enhancer trap method in Drosophila, Its application to neurobiology
.
In Methods in Neuroscience
(ed.
P. M.
Cann
) Vol.
9
, pp.
397
414
.
Orlando, FL
.:
Academic Press
.
Oliver
,
B.
,
Perrimon
,
N.
and
Mahowald
,
A. P.
(
1987
).
The ovo locus is required for sex-specific germ line maintenance in Drosophila
.
Genes Dev
.
1
,
913
923
.
Pankratz
,
M. J.
and
Jäckle
,
H.
(
1993
).
Blastoderm segmentation
.
In The Development of Drosophila Melanogaster
(eds.
M.
Bate
and
A.
Martinez Arias
) Vol.
1
, pp.
467
516
.
Cold Spring Harbor, NY
:
Cold Spring Harbor Press
.
Patel
,
N. H.
and
Goodman
,
C. S.
(
1992
).
DIG-labeled single-stranded DNA probes for in situ hybridizaton
.
In Nonradioactive Labeling and Detection of Biomolecules
(ed.
C.
Kessler
), pp.
377
381
.
Berlin, Germany
:
Springer Verlag
.
Pirrotta
,
V.
(
1988
).
Vectors for P-element mediated transformation in Drosophila
.
In Vectors, a Survey of Molecular Cloning Vectors and Their Uses (eds. R. L. Rodrigues and D. T. Denhardt)
, pp.
437
456
.
Boston
:
Butterworth
.
Ran
,
B.
,
Boop
,
R.
and
Suter
,
B.
(
1994
).
Null alleles reveal novel requirements for Bic-D during Drosophila oogenesis and zygotic development
.
Development
120
,
1233
1242
.
Roberts
,
S. G. E.
and
Green
,
M.
(
1994
).
Activator-induced conformational change in general transcription factor TFIIB
.
Nature
371
,
717
720
.
Robertson
,
H. M.
,
Preston
,
C. R.
,
Philips
,
R. W.
,
Johnson-Schlitz
,
D. M.
,
Benz
,
W. K.
and
Engels
,
W. R.
(
1988
).
A stable genomic source of P element transposase in Drosophila melanogaster
.
Genetics
118
,
461470
.
Rost
,
B.
,
Sander
,
C.
and
Schneider
,
R.
(
1994
).
PHD--an automatic mail server for protein secondary structure prediction
.
Computer Applic. Biosciences
10
,
53
60
.
Sambrook
,
J.
,
Fritsch
,
E. F.
and
Maniatis
,
T.
(
1989
).
Molecular Cloning, A Laboratory Manual, Second Edition
Cold Spring Harbor, NY
:
Cold Spring Harbor Laboratory Press
.
Schüpbach
,
T.
and
Roth
,
S.
(
1994
).
Dorsoventral patterning in Drosophila oogenesis
.
Curr. Opin. Genet. Devel
.
4
,
502
507
.
Schüpbach
,
T.
and
Wieschaus
,
E. F.
(
1991
).
Female sterile mutations of the second chromosome of Drosophila melanogaster, II. Mutations blocking oogenesis or altering egg morphology
.
Genetics
129
,
1119
1136
.
Spradling
,
A. C.
(
1986
).
P element-mediated transformation
.
In Drosophila; A Practical Approach
(ed.
D.B.
Roberts
), pp.
175
199
.
Oxford
:
IRL Press
.
Spradling
,
A. C.
(
1993
).
Developmental genetics of oogenesis
.
In The Development of Drosophila melanogaster
(ed.
M.
Bate
and
A.
Martinez Arias
). pp.
1
70
.
Cold Spring Harbor, NY
:
Cold Spring Harbor Laboratory Press
.
Suter
,
B.
and
Steward
,
R.
. (
1991
).
Requirement for phosphorylation and localization of the Bicaudal-D protein in Drosophila oocyte differentiation
.
Cell
67
,
917
926
.
Tautz
,
D.
and
Pfeifle
,
C.
(
1989
).
A non-radioactive in situ hybridization method for localization of specific RNAs in Drosophila embryos reveals translational control of the segmentation gene hunchback
.
Chromosoma
98
,
81
85
.
Theurkauf
,
W. E.
,
Alberts
,
B. M.
,
Jan
,
Y. N.
and
Jongens
,
T. A.
(
1993
).
A central role for microtubles in the differentation of Drosophila oocytes
.
Development
118
,
1169
1180
.
Thummel
,
C.
and
Pirrotta
,
V.
(
1992
).
New pCaSpeR P element vectors
.
Dros. Inf. Serv
.
71
,
150
.
Wasserman
,
S. A.
(
1993
).
A conserved signal transduction pathway regulating the activity of the rel-like proteins dorsal and NF-kappa B
.
Molec. Biol. Cell
4
,
767
771
.
Yue
,
L.
and
Spradling
,
A.C.
(
1992
).
Hu-li tai shao, a gene required for ring canal formation during Drosophila oogenesis, encodes a homolog of adducin
.
Genes Dev
.
6
,
2443
2454
.

The nucleotide sequence of stonewall will appear in the GenBank Nucleotide Sequence Database under accession number U41367.