In D. melanogaster, a set of ‘X:A numerator genes’, which includes sisterlessA (sisA), determines sex by controlling the transcription of Sex-lethal (Sxl). We characterized sisA from D. pseudoobscura and D. virilis and studied the timing of sisA and Sxl expression with single cell-cycle resolution in D. virilis, both to guide structure-function studies of sisA and to help understand sex determination evolution. We found that D. virilisSISA shares 58% amino acid identity with its melanogaster ortholog. The identities confirm SISA as an atypical bZIP transcription factor. Although virilisSISA can substitute for melanogasterSISA, the protein is not fully functional in a heterologous context. The putative sisA regulatory sequence CAGGTAG is a potential ‘numerator box,’ since it is shared with the other strong X:A numerator gene, sisB, and its target, SxlPe. Temporal and spatial features of sisA and SxlPe expression are strikingly conserved, including rapid onset and cessation of transcription in somatic nuclei, early cessation of sisA transcription in budding pole cells and persistent high-level sisA expression in yolk nuclei. Expression of sisA and Sxl is as tightly coupled in virilis as it is in melanogaster. Taken together, these data indicate that the same primary sex determination mechanism exists throughout the genus Drosophila.

Although sexual dimorphism is a ubiquitous feature of metazoans, the genetic programs that determine sex ratios differ remarkably among species (see Graves, 1996; Hodgkin, 1992) and can differ even among members of the same species (Hilfiker-Kleiner et al., 1994; Kallman, 1968; Miura et al., 1996). This rapid diversification of developmental mechanisms makes sex determination an appealing subject for comparative molecular analysis aimed at understanding how genetic pathways evolve, especially since sex determination mechanisms are particularly amenable to experimental manipulation. Interspecific comparisons are also useful for understanding structure-function relationships for sex determination genes and their products.

For these reasons, we have compared the sex-determination gene sisterlessA (sisA) among three species that span the breadth of the Drosophila radiation: virilis, pseudoobscura and melanogaster. The common ancestor of these species is estimated to have lived 40-60 Myr ago (Powell and DeSalle, 1995; Russo et al., 1995). We have also included a comparison of the temporal and spatial aspects of the expression pattern for Sex-lethal (Sxl), the regulatory target of sisA. Sxl is the master regulatory gene that heads a gene hierarchy controlling all aspects of sexual development (reviewed in Cline and Meyer, 1996).

Sex in D. melanogaster is determined by the effect on Sxl expression of the zygotic dose of at least four ‘X:A numerator genes,’ one of the most influential being sisA (reviewed in Cline and Meyer, 1996). For chromosomal females (2X2A), the double dose of these X-linked genes launches the female developmental program by transiently activating a Sxl ‘establishment’ promoter, Pe. By contrast, chromosomal males (1X2A) have only a single copy of these four genes and therefore fail to generate a level of numerator gene products sufficient to activate SxlPe. The male developmental program follows by default in the absence of SXL protein.

The pulse of SXL protein generated in females by the action of the numerator elements on SxlPe engages a positive feedback loop on Sxl pre-mRNA splicing after SxlPe shuts off and a ‘maintenance’ promoter, SxlPm, turns on. Primary transcripts from SxlPm, unlike those from SxlPe, are generated in both sexes but are processed into functional mRNA only in the presence of SXL, an RNA-binding protein. By imposing the productive splicing mode for Sxl pre-mRNA, SXL protein ensures its own continued synthesis throughout development. Appropriate sexual differentiation and X-chromosome dosage compensation follow as a consequence of the effects of SXL on a variety of switch genes whose sex-specific functions are more specialized. Although males activate SxlPm, they lack the early pulse of SXL protein made in response to a 2X dose of numerator genes. Consequently, males splice SxlPm-derived RNA in a non-productive fashion and downstream switch gene targets are expressed in their male mode by default.

In D. melanogaster, sisA encodes a transcription factor that resembles proteins of the basic-leucine zipper (bZIP) class in a number of key respects, but which lacks other features shared by most bZIP proteins (Erickson and Cline, 1993). By determining whether these defining similarities were evolutionarily conserved, we sought to ascertain their significance and thereby assess the validity of the bZIP classification for SISA. In addition, we tested the functional consequences of aspects of the amino acid sequence that were not conserved.

Unlike the other X:A numerator genes, sisA is not expressed in somatic cells after the blastoderm stage, nor anywhere else after mid-embryogenesis (Erickson and Cline, 1993). It is, however, highly expressed in yolk nuclei, suggesting that any non-sex-specific function it might have is likely to be related to yolk utilization. We wondered whether sisA in other species would also exhibit strong expression in yolk nuclei. If sisA was available for recruitment as a very early acting sex-determination signal element because of a prior involvement in yolk utilization very early in development, one would expect sisA expression in yolk nuclei to be at least as conserved evolutionarily as its expression in somatic nuclei.

In D. melanogaster, sisA is among the earliest genes expressed. Its transcription begins in nuclear cycle 8 before any of the rapidly dividing syncytial nuclei of the young embryo have reached the egg cortex (Erickson and Cline, 1993). Cycle 8 is before the germline (i.e. the pole cells) segregates from the soma, and it is considerably before somatic cells first form in cycle 14 and most embryonic gene expression begins (reviewed in Foe et al., 1993; Pritchard and Schubiger, 1996). A striking feature of sisA expression is its abrupt cessation in those nuclei that bud off to form pole cells. This shutting off of sisA is the earliest indication of fundamental differences between sex determination in the soma and germline of D. melanogaster. Would such an early onset of sisA transcription in the embryo and such an abrupt shut off in the precursors of the germline be characteristic of other species as well?

The early onset of transcription for sisA in D. melanogaster allows for an early onset of transcription for its regulatory target, Sxl. Moreover, cessation of sisA expression is paralleled by the subsequent silence of SxlPe. Transcription of Sxl in female somatic nuclei begins in nuclear cycle 12, only four short cell cycles after sisA and still prior to the stage at which somatic cells first form (Erickson and Cline, 1993). Would other species determine their sex as early as melanogaster and would they display such a close coupling between regulator and target?

Isolation of D. virilis and D. pseudoobscura genomic clones

D. virilis sisA was recovered from a λ EMBL3 genomic library provided by G. Rubin (UC Berkeley). Plaque lifts were hybridized with a full-length D. melanogaster sisA cDNA probe in 6× SSC, 0.2% SDS, 5× Denhardt’s solution, 100 μg/ml ssDNA at 50°C. Washes were in 1× SSC at 55°C. Hybridizing phage were re-screened with a D. melanogaster l(1)10Bb cDNA probe in 6× SSC, 0.25% non-fat dry milk at 65°C. Washes were in 0.2× SSC, 0.1% SDS at 65°C. D. pseudoobscura sisA was recovered from a λ EMBL3 genomic library provided by A. Berry and R. Lewontin (Harvard University). The library was screened with a probe encoding the last 87 residues of melanogaster sisA. Conditions for the primary and counter-screen were as described for l(1)10Bb. Other techniques were standard (Sambrook et al., 1989).

Sequences and sequence analysis

GenBank accession numbers for sisA from D. melanogaster, D. virilis and D. pseudoobscura are AF046044, AF045585 and AF045586, respectively, and AF045587 for l(1)10Bb of melanogaster. A portion of the D. virilis SxlPe region (Genbank #AF046045) was amplified from genomic DNA using primers CGCTTAACTAGCCTCTAATCG and GACAGCGACCAGAGTTGTCAG. P. Gergen (SUNY Stony Brook) generously shared unpublished 5′ runt sequence. Sequence analyses were performed with IBI and GCG software. Percentage identity was calculated excluding residues opposite gaps from consideration.

P-element-mediated germline transformation and genetic analysis

NcoI and BglII sites were introduced into D. virilis sisA at the start of translation and in the 3′ UTR by PCR amplification with primers: Dv4Nco CAACTGACcATGGAACAGAG and Dv3Bgl GTAATGTAGAtCtCCAGCGGC. The D. virilis coding sequence was inserted into the melanogaster gene at the NcoI and BamHI sites. This replaces the melanogaster sisA-coding region and 31 bp of 3′ sequence with the virilis coding region and 14 bp of 3′ sequence. A 6.4 kb NheI fragment containing the chimeric gene and l(1)10Bb was cloned into the P-element transformation vector Casper. The plasmid carries the same regulatory information as P[A10, w+mC, sisA+, l(1)10Bb+] described in Erickson and Cline (1993), and germline transformants were obtained by the same techniques. Alleles and balancer chromosomes are listed at http://flybase.bio.indiana.edu.

Embryo in situ hybridization

D. virilis embryos were collected at 25°C on our standard cornmeal/molasses/agar food smeared with live yeast paste. Embryo fixation, preparation and hybridization with digoxigenin-labeled RNA probes, followed Lehmann and Tautz (1994) except that proteinase K concentration and incubation time were reduced for virilis embryos, and antibody incubations were in 0.1 M maleate, 0.15 M NaCl, 2% Boehringer Mannheim Blocking Reagent, 0.1% Tween 20, pH=7.5. Embryos were treated with DAPI (0.5 μg/ml) to stain nuclei. D. virilis sisA probes contained the coding region and 13 bases of 3′ sequence.

D. virilis Sxl exon E1 was cloned by PCR amplification of genomic DNA using mismatch primers virE1Bam5′ TttagGATCcTTTAA-ATAGCTACGCAC and virE1Xba3′ GGAGAATGCAtTtCTAGAG-CCTGG. The product was cloned into the BamHI and XbaI sites of pBluescript KS(+). The 252 bp insert includes the 198 bp Sxl exon E1 and 19 bases of 3′ intron sequence. For comparative purposes, embryos were collected and processed together until the hybridizations. Embryos were staged according to Foe and Alberts (1983). Progress through the nuclear cycle was estimated by nuclear size and chromosome morphology (Edgar et al., 1994) as revealed by DAPI staining.

Nascent Sxl and sisA transcripts were identified as focused dots of staining within the nuclei (Pritchard and Schubiger, 1996; Rothe et al., 1992; Shermoen and O’Farrell, 1991). For X-linked genes such as sisA and Sxl, the number of nuclear dots corresponds to the number of X chromosomes. As described in Barbash and Cline (1995), the dots were sometimes twinned, likely reflecting separation of sister chromatids. Such twinning could be clearly distinguished from the well-separated dots representing homologous chromosomes. The earliest sisA transcripts showed a punctate nuclear staining that made it difficult to identify nuclear dots prior to cycle 11. In cycles 11 and 12, it was usually possible to distinguish male and female embryos based on sisA-stained nuclear dots; however, the dots could rarely be seen in all nuclei and were always less intense than those observed for Sxl. This may reflect the more than ten times greater length of the Sxl transcription unit.

For mRNAs, the Sxl exon E1 probe clearly discriminates between SxlPe-derived and SxlPm-derived species, since only SxlPe mRNA contains this exon; however, discrimination between nascent transcripts is not as clear, since E1 is included in the first intron of SxlPm transcripts (Bopp et al., 1996). Hence probe E1 should detect nascent transcripts from SxlPm that have not yet made their exon 1-2 splice. All cycle 12 embryos that stained with probe E1 exhibited two nuclear dots in at least some nuclei and were therefore female. In cycle 13 and early cycle 14, females could be identified not only by the presence of two intense nuclear dots per nucleus, but also by a cytoplasmic signal. While similarly staged male embryos did show very faint single dots in their nuclei, they exhibited no cytoplasmic Sxl mRNA. In late cycle 14 females, cytoplasmic Sxl mRNA levels declined and intense nuclear dots disappeared. At this point, both sexes displayed very faint nuclear dots that are not apparent in photographs. These faint dots seem likely to represent SxlPm-derived primary transcripts, although we cannot exclude the possibility that they reflect a low basal level of SxlPe expression in both sexes.

A three-species comparison of sisA gene and protein structure

To clone sisA from D. virilis and D. pseudoobscura, we took advantage of the close linkage discovered in D. melanogaster between sisA and l(1)10Bb (Erickson and Cline, 1993), a small gene that is highly conserved from plants to animals (Hla et al., 1995). We correctly assumed that synteny would be maintained in virilis for two genes only 2 kb apart in melanogaster. Consequently, we first screened virilis and pseudoobscura genomic libraries with a melanogaster sisA cDNA probe, then determined which phages were likely to carry the sisA ortholog by screening positives with the melanogaster l(1)10Bb gene. The sisA regions were sequenced and compared. Because the structure and much of the sequence of sisA was highly conserved, identification of the orthologs was unambiguous. For all three species, the sisA-coding region is small and there are no introns.

Outside the coding region, nucleotide sequence similarity was apparent only in a few upstream regions (Fig. 1). Given the rate of sequence divergence in the genus Drosophila, these conserved regions are likely to be important for regulation (Hartl and Lozovskaya, 1995). We wondered whether any of these putative regulatory sequences might be shared with other X:A numerator elements or with their target, Sxl. The heptamer CAGGTAG was most promising in this regard. Within sisA, this sequence was found on the sense strand at a corresponding position less than 200 bp upstream of the transcription start site in all three species (Figs 1, 2). In melanogaster and pseudoobscura, but not virilis, another copy was found on the antisense strand closer to the start site.

Fig. 1.

Conserved nucleotide sequences in the sisA promoter region. The 5′ region of sisA is shown for D. melanogaster (mel), D. pseudoobscura (pse) and D. virilis (vir). (+1) identifies the presumptive transcription start sites, and (AUG) the translation starts. The alignment emphasizes the largest continuous blocks of conserved sequence. Identities are shown only in regions with extended similarity. A conserved heptamer is tentatively labeled ‘numerator box’ due to its occurrence in functionally related genes (see Fig. 2).

Fig. 1.

Conserved nucleotide sequences in the sisA promoter region. The 5′ region of sisA is shown for D. melanogaster (mel), D. pseudoobscura (pse) and D. virilis (vir). (+1) identifies the presumptive transcription start sites, and (AUG) the translation starts. The alignment emphasizes the largest continuous blocks of conserved sequence. Identities are shown only in regions with extended similarity. A conserved heptamer is tentatively labeled ‘numerator box’ due to its occurrence in functionally related genes (see Fig. 2).

Fig. 2.

Occurrence of a conserved heptamer sequence near the promoters of sisA, sisB and Sxl, three genes involved in the primary sex-determination signal. When CAGGTAG occurs on the antisense strand, its complement is shown italicized on the sense strand. (Abbreviations as in Fig. 1, and sub, D. subobscura).

Fig. 2.

Occurrence of a conserved heptamer sequence near the promoters of sisA, sisB and Sxl, three genes involved in the primary sex-determination signal. When CAGGTAG occurs on the antisense strand, its complement is shown italicized on the sense strand. (Abbreviations as in Fig. 1, and sub, D. subobscura).

Among the sex determination genes, the expression of sisB most closely matches that of sisA. Within the 543 bases of upstream sequence published for D. melanogaster sisB, the heptamer appears twice on the antisense strand at positions comparable to those conserved in sisA (Fig. 2). The gene runt is a weaker numerator element whose expression occurs slightly later and in a more spatially restricted domain than that of sisA or sisB. Within the 440 bases of 5′ sequence available for runt in melanogaster, pseudoobscura and virilis, there are no occurrences of this heptamer or its complement.

Two copies of the heptamer occur in conserved positions near the transcription start site in Sxl that is activated by the numerator elements (Fig. 2). Approximately 80 base pairs upstream of the SxlPe start, the heptamer is found on the antisense strand in D. melanogaster and D. subobscura and on the sense strand in D. virilis. Approximately 50 bases further upstream, the heptamer (allowing one mismatch in melanogaster) is repeated on the sense strand in all three species.

The deduced amino acid sequence of SISA is compared in Fig. 3 for the three species. The three proteins (DV-SISA, DP-SISA and DM-SISA) are similar in size, charge and hydrophilicity. DM-SISA shares 58% amino acid identity with DV-SISA and 55% with DP-SISA. It appears that SISA has two structural domains: a conserved bipartite N-terminal region of unknown function and a C-terminal dimerization and DNA-binding region containing an unusual basic leucine zipper (bZIP) motif. These domains are separated by a non-conserved ‘spacer’ of varying size.

Fig. 3.

Amino acid sequences of D. melanogaster, pseudoobscura and virilisSISA protein. Identities are blocked. Alphabetic positions (a-g) denote residues in the leucine zipper. The basic region consensus sequence is shown (B=R or K; ψ=L, I, V or M).

Fig. 3.

Amino acid sequences of D. melanogaster, pseudoobscura and virilisSISA protein. Identities are blocked. Alphabetic positions (a-g) denote residues in the leucine zipper. The basic region consensus sequence is shown (B=R or K; ψ=L, I, V or M).

The most conserved part of SISA is the bZIP motif. The features that SISA shares with other bZIP family members are preserved, but so are the features that are so unusual for a protein of this class. Among these atypical features is the absence of characteristic single or paired alanines located three and/or four residues downstream of the invariant asparagine in the basic region (Hurst, 1995). Also atypical is the pattern of amino acids in the ‘a’ position of the dimerization helix. The ‘a’ and ‘d’ residues serve as the dimerization interface and are, in most bZIP proteins, large hydrophobic amino acids. ‘d’ is usually leucine, consistent with the L•L•I•L sequence found for SISA. ‘a’ generally contains other hydrophobic residues, but in SISA, the invariant sequence of ‘a’ residues is A•R•E•G. This arrangement, combined with the lack of helix-stabilizing residues in the ‘e’ and ‘g’ positions, suggests 1993), in agreement with results of two-hybrid analyses (Liu and Belote, 1995). The positions in the zipper domain that diverge most, ‘c’ and ‘f’, are predicted to lie on the outside surface of the helix and thus not to participate in dimer formation.

D. virilis SISA protein functions suboptimally in D. melanogaster

To determine if the D. virilisSISA protein is capable of providing sisA function in D. melanogaster, we expressed DV-SISA under the control of melanogaster regulatory sequences and assayed its function in transgenic animals. The melanogaster sisA protein-coding region was replaced with that for virilis within a P-element construct previously shown to provide full sisA function (see Materials and Methods). Data in Table 1 document that the SISA protein sequence divergence between these two species has impaired but not eliminated the ability of the protein to function in the heterologous situation. A single copy of a DV-SISA expressing chimeric transgene fully rescued females homozygous for the hypomorphic, female-specific lethal allele sisA1; however, the chimeric transgene was not fully effective at rescuing hemizygous sisA1 females, a more stringent test of sisA function (Cline, 1988). Consistent that SISA is unlikely to form stable homodimers (Krylov et al., 1994; Vinson et al., with the results for females, the chimeric transgenes also displayed sisA+ activity in males (not shown). This was manifested as a male-specific lethal effect of the transgene in combination with duplications of sisB+ that was somewhat weaker than the male-lethal effect previously reported for the DM-SISA parental transgene.

Table 1.

D. virilisSISA functions suboptimaly in D. melanogaster

D. virilisSISA functions suboptimaly in D. melanogaster
D. virilisSISA functions suboptimaly in D. melanogaster

Conservation of sisA expression

In D. melanogaster, the early onset of sisA transcription, the extinction of sisA transcription in pole cells and the abrupt cessation of sisA transcription during nuclear cycle 14 are all reflections of this gene’s participation in sex determination through its regulation of Sxl. The persistence of sisA transcription in melanogaster yolk nuclei suggested that the gene might also have functions that are not sex-specific. We found that all four of these unusual features have been conserved over the >40 Myr separating D. melanogaster and D. virilis.

Assaying by in situ hybridization, we first detected D. virilis sisA transcripts in nuclear cycle 8 (Figs 4, 5). At this point, the nuclei are migrating outwards to the cortex. Although differences were noted among nuclei in the time of onset, all nuclei had begun to transcribe sisA within one cell cycle. Initially, sisA transcripts were found primarily within nuclei. Unambiguous cytoplasmic signals arose only later during cycle 10. Other Drosophila genes that turn on at a comparably early stage display similar behavior (Pritchard and Schubiger, 1996). Nuclei near the posterior pole expressed lower levels of sisA mRNA than nuclei elsewhere. All somatic and yolk nuclei continued to express sisA in the subsequent cleavage cycles. In contrast, expression was extinguished in the germline precursors, the pole cells, during cycle 9. Transcripts were readily apparent in prepole cell nuclei as the pole buds formed, but disappeared midway through bud formation (Fig. 5E,F).

Fig. 4.

Time course of sisA expression in D. virilis. Transcripts were detected by in situ hybridization with a sisA probe. Precellular embryos are from cycles 5, 8, 9, 10, 12, 13 and 14. Early gastrula (gast) and germ-band-extended (gbe) embryos are shown.

Fig. 4.

Time course of sisA expression in D. virilis. Transcripts were detected by in situ hybridization with a sisA probe. Precellular embryos are from cycles 5, 8, 9, 10, 12, 13 and 14. Early gastrula (gast) and germ-band-extended (gbe) embryos are shown.

Fig. 5.

Details of sisA expression in D. virilis. (A) Nomarski image of cycle 8 embryo expressing sisA. (B) Bright-field surface image of a cycle 9 embryo. (C,D) Magnified surface images of cycle 12 embryos. Nascent sisA transcripts are represented by dots of staining in the nuclei: one dot in (C) males, two dots in (D) females. (E,F) Sequential optical sections through the posterior of a cycle 9 embryo undergoing pole cell formation. Embryo was illuminated with both UV and visible light to reveal DAPI-stained nuclei (pale blue) and sisA staining. Nuclear sisA staining is visible in somatic nuclei and in early pole buds (shorter arrows). Note the absence of sisA staining in the most mature pole buds (blue, longer arrows).

Fig. 5.

Details of sisA expression in D. virilis. (A) Nomarski image of cycle 8 embryo expressing sisA. (B) Bright-field surface image of a cycle 9 embryo. (C,D) Magnified surface images of cycle 12 embryos. Nascent sisA transcripts are represented by dots of staining in the nuclei: one dot in (C) males, two dots in (D) females. (E,F) Sequential optical sections through the posterior of a cycle 9 embryo undergoing pole cell formation. Embryo was illuminated with both UV and visible light to reveal DAPI-stained nuclei (pale blue) and sisA staining. Nuclear sisA staining is visible in somatic nuclei and in early pole buds (shorter arrows). Note the absence of sisA staining in the most mature pole buds (blue, longer arrows).

Transcription of sisA was maintained during the subsequent cycles of the syncytial blastoderm (Fig. 4). For somatic nuclei, the level of sisA mRNA peaked in late cycle 12 or early cycle 13. By this point, the sisA regulatory target, SxlPe, has become active (see below). In some embryos, focused dots of sisA hybridization could be seen in the nuclei. Such dots represent nascent transcripts (Pritchard and Schubiger, 1996; Rothe et al., 1992; Shermoen and O’Farrell, 1991). Although hybridization dots could not be seen in all embryos, or even in all the nuclei of any single embryo, embryos displaying nuclear dots fell into two classes: individuals with only a single dot per nucleus and embryos with two (Fig. 5C,D). Presumably, this reflects an X-chromosome location for sisA in D. virilis with females carrying two alleles and males only one.

After cycle 13, nuclear dots disappeared, signaling the cessation of sisA transcription. The cytoplasmic hybridization signals begin to disappear from the periphery of the embryo first, but by early to mid cycle 14, the last traces of sisA mRNA had disappeared from all surface nuclei. As had been the case with D. melanogaster, no subsequent somatic or germ-cell transcription of sisA was observed in virilis embryos.

In D. virilis as in D. melanogaster, sisA transcripts persisted in the yolk nuclei until the demise of those nuclei 10 to 12 hours after fertilization (Fig. 4). The primary yolk nuclei derive from 70-100 nuclei that fail to migrate with the majority of nuclei toward the periphery of the embryo during cycles 8 and 9. Instead they fall back into the central yolk mass and eventually become polyploid (Foe and Alberts, 1983; Foe et al., 1993). During cycle 14, sisA transcript levels in these nuclei increased abruptly and remained high thereafter. During normal embryogenesis, some nuclei drop from the periphery into the surface of the yolk mass to become ‘secondary’ yolk nuclei, perhaps as a consequence of errors in replication or division (Frenz and Glover, 1996; Sullivan et al., 1993). Although not highly polyploid, these nuclei also express high levels of sisA. Probable secondary yolk nuclei can be seen in Fig. 4 (cycle 14) near the periphery and at the anterior pole of the embryo. Their behavior suggests that persistence of sisA transcription is likely to be a consequence of environmental differences between the central yolk mass and the yolk-free periphery, rather than polyploidy per se.

The timing of SxlPe expression is conserved in D. virilis

The observation that the timing of sisA expression is essentially identical between D. virilis and D. melanogaster leads to the question of whether the timing of expression of SxlPe, the target of sisA, is equally conserved. It had been shown previously that Sxl is expressed in a sex-specific manner at the cellular blastoderm stage in both D. virilis and D. subobscura (Bopp et al., 1996; Penalva et al., 1996), and for subobscura this expression had been shown to be due to SxlPe; however, the temporal resolution of that analysis was relatively crude. Consequently, we performed in situ hybridization experiments with an exon E1 probe for SxlPe products, analyzing the young embryos cell cycle by cell cycle. We found that the temporal regulation of SxlPe, like that of sisA, is virtually identical between these two distant Drosophila species.

The accumulation of D. virilis SxlPe transcript across the whole embryo is shown in Fig. 6, while Fig. 7 illustrates the time course of ‘nuclear dot’ development reflecting the synthesis of SxlPe transcripts. A SxlPe signal was first detected as light dots of staining in interphase nuclei of cycle 12 embryos (Fig. 7A). The fact that SxlPe activation is female-specific in virilis was established by the observation that as the dots darkened, they ultimately occurred only in pairs, reflecting the XX (female) karyotype of the Sxl-positive embryos. Nevertheless, SxlPe was not necessarily activated simultaneously on both X chromosomes within a single nucleus. Individual early cycle 12 embryos exhibited mosaic patterns of nuclear dots in which some nuclei expressed Sxl from both X chromosomes, some from only one X, and others from neither (Fig. 7A). Slightly later in cycle 12, and in cycle 13 and 14 embryos, the nuclear dots became much more intense and these intense dots were always found in pairs (Fig. 7B-D). Cytoplasmic Sxl transcripts first appeared in late cycle 12 females (Fig. 6). They reached maximum levels shortly after the peak in nuclear dot staining was reached – late cycle 13 to early cycle 14. In the nuclei of cycle 13 and 14 embryos that exhibited no cytoplasmic accumulation of exon-E1-containing mRNA, only a single faint dot of staining was visible (Fig. 7F). This dot is likely to represent unspliced SxlPm transcripts in males (see Materials and Methods).

Fig. 6.

Time course of SxlPe expression in D. virilis. SxlPe-derived transcripts were detected by in situ hybridization with an exon E1 probe. The embryo in interphase of cycle 11 (11i) has not stained. The embryo in early interphase 12 (12ei) is a female that had just initiated SxlPe expression and exhibits nuclear staining only (see also Fig. 7A). The later interphase 12 (12i); interphase 13 (13i); anaphase 13 (13a); and early cycle 14 (14) embryos are females exhibiting nuclear and cytoplasmic Sxl staining. The gastrulating embryo (gast) is female. The cycle 14 female represents the darkest staining observed and is similar in stage to the cycle 14 male.

Fig. 6.

Time course of SxlPe expression in D. virilis. SxlPe-derived transcripts were detected by in situ hybridization with an exon E1 probe. The embryo in interphase of cycle 11 (11i) has not stained. The embryo in early interphase 12 (12ei) is a female that had just initiated SxlPe expression and exhibits nuclear staining only (see also Fig. 7A). The later interphase 12 (12i); interphase 13 (13i); anaphase 13 (13a); and early cycle 14 (14) embryos are females exhibiting nuclear and cytoplasmic Sxl staining. The gastrulating embryo (gast) is female. The cycle 14 female represents the darkest staining observed and is similar in stage to the cycle 14 male.

Fig. 7.

Nascent transcripts from SxlPe in D. virilis detected by in situ hybridization with an exon E1 probe. Surface views: (A) Interphase cycle 12 female near the onset of SxlPe expression. (B) Interphase cycle 12 female. (C) Interphase/prophase cycle 13 female. (D) Early cycle 14 female. (E) Late cycle 14 female. (F) Male in interphase of cycle 13. The single faint dots likely reflect unspliced transcripts initiated from SxlPm (see Materials and Methods).

Fig. 7.

Nascent transcripts from SxlPe in D. virilis detected by in situ hybridization with an exon E1 probe. Surface views: (A) Interphase cycle 12 female near the onset of SxlPe expression. (B) Interphase cycle 12 female. (C) Interphase/prophase cycle 13 female. (D) Early cycle 14 female. (E) Late cycle 14 female. (F) Male in interphase of cycle 13. The single faint dots likely reflect unspliced transcripts initiated from SxlPm (see Materials and Methods).

As cycle 14 continued, shut off of SxlPe was signaled by the disappearance of clear-cut nuclear dots. Degradation of transcripts made prior to this point was evidenced by the rapid loss of cytoplasmic staining thereafter. The shut-off of SxlPe and degradation of Sxl mRNA appear to be non-uniform, since a series of weakly defined stripes became apparent in late blastoderm and early gastrula embryos (Fig. 6), just as had been observed for D. melanogaster (Barbash and Cline, 1995; Erickson and Cline, 1993). Loss of nuclear dots occurred first in the lighter staining areas (not shown). As was true for melanogaster, transcripts derived from SxlPe in D. virilis were not evident in pole cells or in embryos at later stages, regardless of sex (not shown).

The present study adds the strong X:A numerator element sisterlessA to the list of Drosophila sex determination genes for which interspecific comparisons have been made. That list now includes three genes of the primary sex determination signal (sisA, sisB, runt), the target of the primary signal (Sxl), three regulatory genes downstream of Sxl that are specifically involved in somatic sexual differentiation (tra, tra2 and dsx), and two regulatory targets of Sxl that control dosage compensation (msl-1 and msl-2) (Bone and Kuroda, 1996; Bopp et al., 1996; Botella et al., 1996; Chandler et al., 1997; Marin et al., 1996; O’Neil and Belote, 1992; Penalva et al., 1996; Pepling and Gergen, 1995; Raymond et al., 1998).

A potential ‘numerator box’: promoter sequence similarities among numerator elements sisA and sisB and their target, Sxl

One aim of this study was to use DNA sequence conservation to identify potential regulatory regions that allow genes to function as X:A numerator elements. The heptomeric sequence CAGGTAG, which is conserved in all three sisA genes examined here, is a promising candidate. It appears in pairs less than 200 bp upstream of the transcription start site not only for sisA, but also for sisB and Sxl (Fig. 2). These three genes have in common an unusually early onset of expression that allows the embryo to establish X-chromosome dosage compensation by the time general transcription begins. Perhaps this sequence provides for such an early start of transcription.

Conservation of the unusual bZIP functional domain of SISA protein

The protein-coding sequences of sisA display a commonly observed pattern: highly conserved regions separated by completely diverged sequences (Fig. 3). Conservation of the bZIP domain confirms the validity of this motif assignment, despite the occurrence of non-canonical residues in key positions. The N-terminal region of the protein was also found to be conserved, notwithstanding the fact that the N-terminal 15 amino acids can be deleted with only modest effects on in vivo function (Erickson and Cline, 1993). An intriguing feature of the melanogaster protein was a run of six glycine-serine pairs hypothesized to serve as a flexible spacer linking two functional domains. Although this run is not conserved, it occurs in a region that is notable for its variable length and lack of conservation, consistent with such a spacer role.

The level of amino acid sequence conservation for sisA protein within the genus (55-68% identity) is comparable to that seen with the Drosophila sex determination genes sisB and tra-2, is less than that for runt and Sxl (80-90%), and is considerably greater than that for tra (36%). Thus sisA, like most other sex-signal genes of Drosophila, does not display the remarkably rapid evolution observed for the mammalian sex-signal gene, SRY (Pamilo and O’Neill, 1997).

Heterospecific functioning of sex-determination genes

Genomic fragments from D. virilis that included tra or tra-2 provided function to D. melanogaster, but the level of activity was below that of endogenous alleles (Chandler et al., 1997; O’Neil and Belote, 1992). Similarly, sisB from D. subobscura provided sex-determination activity to melanogaster (Botella et al., 1996), but it is not clear how close to wild-type that activity might have been. Because these heterospecific transgenes carried both regulatory and structural information from the foreign species, the basis for their suboptimal performance remains unclear. In our study of the functioning in D. melanogaster of SISA from D virilis, only the protein-coding sequence was foreign. Nevertheless, the results resembled those with tra and tra2: the foreign protein provided some function, but less than the endogenous gene. This result implies that the sequence divergence among the SISA proteins is functionally significant.

The temporal and spatial pattern of sisA expression is conserved

We show here that the many distinctive features of the sisA expression pattern described for D. melanogaster are conserved in D. virilis. Not only were sisA transcripts first detected in cycle 8 embryos from both species, all nuclei expressed sisA by the end of that cell cycle. In contrast, Pritchard and Schubiger (1996), studying two other early expressed Drosophila genes ftz and Kr, found transcriptional activation to be more gradual, with different nuclei in the same embryo beginning transcription during different cell cycles. Whether this reflects fundamental differences in the mechanisms by which these early genes are activated remains to be determined.

In part because so little is known about the role of the yolk nuclei and the factors that control expression of their genes, our observation that persistent sisA expression in yolk nuclei is a conserved feature of this gene does not yet allow us to distinguish between the possibility that sisA has an important function in yolk utilization, and the alternative possibility that persistent yolk expression is simply a consequence of the way the gene is activated for sex determination. Ner and Travers (1994) noted that the chromatin environment of the early Drosophila embryo is similar to that of the yolk nuclei in being rich in HMG-d protein and deficient in histone H1. Hence it is possible that factors that allow sisA to be transcribed during the nuclear multiplication stage might lead to its continued expression in the yolk simply because the chromatin environment of those nuclei remains unchanged. Arguing against such an ‘environmental’ explanation, however, is the fact that sisB, which begins transcription as early as sisA, does not exhibit persistent high-level expression in the yolk nuclei (Erickson and Cline, 1993). If sisA expression in yolk nuclei is functionally significant, it will be necessary to examine species beyond the genus Drosophila to determine whether this might be the ancestral function predating sisA’s recruitment as a sex-determination signal element.

The behavior of sisA in the pole cells, the precursors to the germline, has implications for how those cells remain transcriptionally inactive prior to germ band extension. It had been suggested that this quiescence might be due to pole cells being derived from transcriptionally inactive nuclei that are kept inert until the germ band extends (Ner and Travers, 1994; Williamson and Lehmann, 1996). This explanation is inconsistent with the behavior of sisA, which establishes that prepole cell nuclei, and even budding pole cells, transcribe genes that will be silent after bud formation is complete. Since pole cell nuclei remain in mitotic synchrony with the other embryonic nuclei for some time after bud formation is complete (Foe et al., 1993), the sisA shut-off seems not to be a consequence of some obvious change in the mitotic cycle. There is evidence that Drosophila and C. elegans both keep early germ cells transcriptionally silent by preventing phosphorylation of the carboxy-terminal domain of RNA polymerase II (Seydoux and Dunn, 1997). The behavior of sisA suggests that active Pol II is either inactivated or excluded from the nuclei as pole cells form. sisA might serve as a useful molecular marker in a search for genes regulating such events.

Correlation between sisA expression and SxlPe activation, and the conservation of X-chromosome counting mechanisms

The fact that sisA and SxlPe exhibit precisely the same temporal and spatial expression patterns in two species that span the breadth of the Drosophila radiation argues that the same primary mechanism for determining sex exists throughout this genus. In both species, SxlPe is first activated in cycle 12 diplo-X embryos and shuts off during cycle 14. The pattern of Sxl RNA hybridization dots in early cycle 12 embryos shows that the activation of SxlPe must be a stochastic process that occurs independently not only in each nucleus, but even on each X chromosome. Such behavior favors the hypothesis that the rate-limiting step in Sxl activation is assembly of an active transcription complex on the promoter itself, rather than formation of active heterodimeric transcription factors in the nucleoplasm. There seems to be a difference between the initial activation of SxlPe and its reactivation after mitosis: nascent transcript dots were not apparent on anaphase and telophase chromosomes at the ends of cycles 12 and 13, but when they reappeared in interphase diplo-X embryos in cycles 13 and 14, they were always seen as two dots per nucleus. The temporal pattern of SxlPe activity is consistent with a direct effect of SISA protein on SxlPe in both virilis and melanogaster, since in both species SxlPe activity begins when the sisA mRNA level is near a very transient maximum and ceases shortly after sisA mRNA disappears.

Although the initial events in sex determination for D. virilis and D. melanogaster appear identical, primary events in sex determination for other Diptera are clearly very different (Hilfiker-Kleiner et al., 1994; Meise et al., 1998; Puchalla, 1994; Saccone et al., 1998; Traut, 1994). Is the Drosophila mechanism a relatively recent evolutionary innovation, or might it represent instead an ancient mechanism for Dipteran sex determination that has been lost in some lineages? In this connection, it is interesting to note that the Sciaridae and Cecidomyiidae, Dipterans that diverged from the Drosophila lineage some 200 Myr ago (see Sommer et al., 1992), may possess an X chromosome counting mechanism for sex determination that operates as early as that in Drosophila (de Saint Phalle and Sullivan, 1996; Gerbi, 1986; Metz, 1938; Stuart and Hatchett, 1991).

Similarities between sex determination mechanisms in genera as diverse as Sciara and Drosophila may be purely coincidental, but those among species within the genus Drosophila clearly are not. Advancing to the next level of sophistication in understanding the extent of similarities among Drosophila species – for example, is sisA as important a numerator element in virilis as in melanogaster? – is likely to require bona fide genetic analysis in these other species. Recent advances in transgenic techniques (Lohe and Hartl, 1996; Lozovskaya et al., 1996) make such an approach less formidable than it seemed not long ago.

We thank S. Feist, L. Subrahmanyan, J. Walker and M. Bell for help with library screening and sequencing, Y. Wei for assistance with transgenes, A. Berry, R. Lewontin and G. Rubin for genomic libraries, B. Meyer and L. Megna for comments on the manuscript, and B. de Saint Phalle for discussions about Sciara. Research support was from US NIH grant GM23468 (T. W. C.), ACS grant RPG-97-079-01-DB (J. W. E.), and a Searle Community Trust Award (J. W. E.).

Barbash
,
D. A.
and
Cline
,
T. W.
(
1995
).
Genetic and molecular analysis of the autosomal component of the primary sex determination signal of Drosophila melanogaster
.
Genetics
141
,
1451
71
.
Bone
,
J. R.
and
Kuroda
,
M. I.
(
1996
).
Dosage compensation regulatory proteins and the evolution of sex chromosomes in Drosophila
.
Genetics
144
,
705
713
.
Bopp
,
D.
,
Calhoun
,
G.
,
Horabin
,
J. I.
,
Samuels
,
M.
and
Schedl
,
P.
(
1996
).
Sex-specific control of Sex-lethal is a conserved mechanism for sex determination in the genus Drosophila
.
Development
122
,
971
82
.
Botella
,
L. M.
,
Donoro
,
C.
,
Sánchez
,
L.
,
Segarra
,
C.
and
Granadino
,
B.
(
1996
).
Cloning and characterization of the scute (sc) gene of Drosophila subobscura
.
Genetics
144
,
1043
51
.
Chandler
,
D.
,
McGuffin
,
M. E.
,
Piskur
,
J.
,
Yao
,
J.
,
Baker
,
B. S.
and
Mattox
,
W.
(
1997
).
Evolutionary conservation of regulatory strategies for the sex determination factor transformer-2
.
Mol. Cell Biol
.
17
,
2908
2919
.
Cline
,
T. W.
(
1988
).
Evidence that sisterless-a and sisterless-b are two of several discrete ‘numerator elements’ of the X/A sex determination signal in Drosophila that switch Sxl between two alternative stable expression states
.
Genetics
119
,
829
862
.
Cline
,
T. W.
and
Meyer
,
B. J.
(
1996
).
Vive la différence: males vs females in flies vs worms
.
Annu. Rev. Genet
.
30
,
637
702
.
de Saint Phalle,
B.
and
Sullivan
,
W.
(
1996
).
Incomplete sister chromatid separation is the mechanisms of programmed chromosome elimination during early Sciara coprophila embryogenesis
.
Development
122
,
3775
3784
.
Edgar
,
B. A.
,
Sprenger
,
F.
,
Duronio
,
R. J.
,
Leopold
,
P.
and
O’Farrell
,
P. H.
(
1994
).
Distinct molecular mechanisms regulate cell timing at successive stages of Drosophila embryogenesis
.
Genes Dev
.
8
,
440
452
.
Erickson
,
J. W.
and
Cline
,
T. W.
(
1993
).
A bZIP protein, SISTERLESS-A, collaborates with bHLH transcription factors early in Drosophila development to determine sex
.
Genes Dev
.
7
,
1688
1702
.
Foe
,
V. E.
and
Alberts
,
B. M.
(
1983
).
Studies of nuclear and cytoplasmic behavior during the five mitotic cycles that precede gastrulation in Drosophila embryogenesis
.
J. Cell Sci
.
61
,
31
70
.
Foe
,
V. E.
,
Odell
,
G. M.
and
Edgar
,
B. A.
(
1993
). Mitosis and Morphogenesis in the Drosophila embryo: Point and Counterpoint. In
The Development of Drosophila melanogaster.’ (M. Bate and A. Martinez-Arias, Eds
.), pp.
149
300
.
Cold Spring Harbor Press, Cold Spring Harbor
.
Frenz
,
L. M.
and
Glover
,
D. M.
(
1996
).
A maternal requirement for glutamine synthetase I for the mitotic cycles of syncytial Drosophila embryos
.
J. Cell Sci
.
109
,
2649
2660
.
Gerbi
,
S.
(
1986
). Unusual chromosome movements in Sciarid flies. InGerm- line-Soma Differentiation.’ (
W.
Henning
, Ed.), pp.
71
-
104
.
Springer-Verlag, Heidelberg
.
Graves
,
J. A.
(
1996
).
Mammals that break the rules: genetics of marsupials and monotremes
.
Ann. Rev. Genet
.
30
,
233
260
.
Hartl
,
D. L.
and
Lozovskaya
,
E. R.
(
1995
).
The Drosophila Genome Map: A Practical Guide.’ R.G. Landes, Austin
.
Hilfiker-Kleiner
,
D.
,
Dübendorfer
,
A.
,
Hilfiker
,
A.
and
Nöthiger
,
R.
(
1994
).
Genetic control of sex determination in the germ line and soma of the housefly, Musca domestica
.
Development
120
,
2531
2538
.
Hla
,
T.
,
Jackson
,
A. Q.
,
Appleby
,
S. B.
and
Maciag
,
T.
(
1995
).
Characterization of edg-2, a human homologue of the Xenopus maternal transcript G10 from endothelial cells
.
Biochim. Biophys. Acata
1260
,
227
229
.
Hodgkin
,
J.
(
1992
).
Genetic sex-determination mechanisms and evolution
.
BioEssays
14
,
253
261
.
Hurst
,
H.
(
1995
).
Transcription factors 1: bZIP proteins
.
Protein Profile
2
,
101
168
.
Kallman
,
K. D.
(
1968
).
Evidence for the existence of transformer genes for sex in the teleost Xiphophorus Maculatus
.
Genetics
60
,
811
828
.
Krylov
,
D.
,
Mikhailenko
,
I.
and
Vinson
,
C.
(
1994
).
A thermodynamic scale for leucine zipper stability and dimerization specificity: e and g interhelical interactions
.
EMBO J
.
13
,
2849
2861
.
Lehmann
,
R.
and
Tautz
,
D.
(
1994
). In situ hybridization to RNA. In
‘Drosophila melanogaster: Practical Uses in Cell and Molecular Biology’ (L. S. B. Goldstein and E. A. Fyrberg, Eds
.), Vol.
44
, pp.
576
597
.
Academic Press, San Diego
.
Liu
,
Y.
and
Belote
,
J. M.
(
1995
).
Protein-protein interactions among components of the Drosophila primary sex determination signal
.
Mol. Gen. Genet
.
248
,
182
189
.
Lohe
,
A. R.
and
Hartl
,
D. L.
(
1996
).
Germline transformation of Drosophila virilis with the transposable element mariner
.
Genetics
143
,
365
374
.
Lozovskaya
,
E. R.
,
Nurminsky
,
D. I.
,
Hartl
,
D. L.
and
Sullivan
,
D. T.
(
1996
).
Germline transformation of Drosophila virilis mediated by the transposable element hobo
.
Genetics
142
,
173
177
.
Marin
,
I.
,
Franke
,
A.
,
Bashaw
,
G. J.
and
Baker
,
B. S.
(
1996
).
The dosage compensation system of Drosophila is co-opted by newly evolved X chromosomes
.
Nature
383
,
160
163
.
Meise
,
M.
,
Hilfiker-Kleiner
,
D.
,
Dübendorfer
,
A.
,
Brunner
,
C.
,
Nöthiger
,
R.
and
Bopp
,
D.
(
1998
).
Sex-lethal, the master sex-determining gene in Drosophila, is not sex-specifically regulated in Musca domestica
.
Development
125
,
1487
1494
.
Metz
,
C. W.
(
1938
).
Chromosome behavior, inheritance and sex determination in Sciara
.
Am. Nat
.
72
,
485
520
.
Miura
,
I.
,
Ohtani
,
H.
,
Kashiwagi
,
A.
,
Hanada
,
H.
and
Nakamura
,
M.
(
1996
).
Structural differences between XX and ZW sex lampbrush chromosomes in Rana rugosa females
.
Chromosoma
105
,
237
241
.
Ner
,
S. S.
and
Travers
,
A. A.
(
1994
).
HMG-D, the Drosophila melanogaster homologue of HMG-1 protein, is associated with early embryonic chromatin in the absence of histone H1
.
EMBO J
.
13
,
1817
1822
.
O’Neil
,
M. T.
and
Belote
,
J. M.
(
1992
).
Interspecific comparison of the transformer gene of Drosophila reveals an unusually high degree of evolutionary divergence
.
Genetics
131
,
113
28
.
Pamilo
,
P.
and
O’Neill
,
R. J. W.
(
1997
).
Evolution of the Sry genes
.
Mol. Biol. Evol
.
14
,
49
55
.
Penalva
,
L. O.
,
Sakamoto
,
H.
,
Navarro-Sabate
,
A.
,
Sakashita
,
E.
,
Granadino
,
B.
,
Segarra
,
C.
and
Sánchez
,
L.
(
1996
).
Regulation of the gene Sex-lethal: a comparative analysis of Drosophila melanogaster and Drosophila subobscura
.
Genetics
144
,
1653
64
.
Pepling
,
M. E.
and
Gergen
,
J. P.
(
1995
).
Conservation and function of the transcriptional regulatory protein RUNT
.
Proc. Natl. Acad. Sci. USA
92
,
9087
9091
.
Powell
,
J. R.
and
DeSalle
,
R.
(
1995
).
Drosophila molecular phylogenies and their uses. InEvolutionary Biology’ (M. K. Hecht, R. J. MacIntyre and M. T. Clegg
, Eds.), Vol. 28. Plenum Press,
New York
.
Pritchard
,
D. K.
and
Schubiger
,
G.
(
1996
).
Activation of transcription in Drosophila embryos is a gradual process mediated by the nucleocytoplasmic ratio
.
Genes Dev
.
10
,
1131
1142
.
Puchalla
,
S.
(
1994
).
Polytene chromosomes of monogenic and amphogenic Chrysomya species (Calliphoridae, Diptera): analysis of banding patterns and in situ hybridization with Drosophila sex-determining gene sequences
.
Chromosomea
103
,
16
30
.
Raymond
,
C. S.
,
Shamu
,
C. E.
,
Shen
,
M. M.
,
Seifert
,
K. J.
,
Hirsch
,
B.
,
Hodgkin
,
J.
and
Zarkower
,
D.
(
1998
).
Evidence for evolutionary conservation of sex-determining genes
.
Nature
391
,
691
695
.
Rothe
,
M.
,
Pehl
,
M.
,
Taubert
,
H.
and
Jäckle
,
H.
(
1992
).
Loss of gene function through rapid mitotic cycles in the Drosophila embryo
.
Nature
359
,
156
159
.
Russo
,
C. A. M.
,
Takezaki
,
N.
and
Nei
,
M.
(
1995
).
Molecular phylogeny and divergence times of Drosophilid species
.
Mol. Biol. Evol
.
12
,
391
404
.
Saccone
,
G.
,
Peluso
,
I.
,
Artiaco
,
D.
,
Giordano
,
E.
,
Bopp
,
D.
and
Polito
,
L. C.
(
1998
).
The Ceratitis capitata homologue of the Drosophila sex-determining gene Sex-lethal is structurally conserved, but not sex-specifically regulated
.
Development
125
,
1501
1507
.
Sambrook
,
J.
,
Fritsch
,
E. F.
and
Maniatis
,
T.
(
1989
).
Molecular cloning: A laboratory manual.’
Cold Spring Harbor Laboratory Press, Cold Spring Harbor
,
New York
.
Seydoux
,
G.
and
Dunn
,
M. A.
(
1997
).
Transcriptionally repressed germ cells lack a subpopulation of phosphorylated RNA polymerase II in early embryos of Caenorhabditis elegans and Drosophila melanogaster
.
Development
124
,
2191
2201
.
Shermoen
,
A. W.
and
O’Farrell
,
P. H.
(
1991
).
Progression of the cell cycle through mitosis leads to the abortion of nascent transcripts
.
Cell
67
,
303
310
.
Sommer
,
R. J.
,
Retzlaff
,
M.
,
Goerlich
,
K.
,
Sander
,
K.
and
Tautz
,
D.
(
1992
).
Evolutionary conservation pattern of zinc-finger domains of Drosophila segmentation genes
.
Proc. Natl. Acad. Sci. USA
89
,
43
52
.
Stuart
,
J. J.
and
Hatchett
,
J. H.
(
1991
).
Genetics of sex determination in the Hessian fly, Mayetiola destructor
.
J. Hered
.
82
,
43
52
.
Sullivan
,
W.
,
Daily
,
D. R.
,
Fogarty
,
P.
,
Yook
,
K. J.
and
Pimpinelli
,
S.
(
1993
).
Delays in anaphase initiation occur in individual nuclei of the syncytial Drosophila embryo
.
Mol. Biol. Cell
4
,
885
896
.
Traut
,
W.
(
1994
).
Sex determination in the fly Megaselia scalaris, a model syhstem for primary steps of sex chromosome evolution
.
Genetics
136
,
1097
1104
.
Vinson
,
C. R.
,
Hai
,
T.
and
Boyd
,
S. M.
(
1993
).
Dimerization specificity of the leucine zipper-containing bZIP motif on DNA binding: prediction and rational design
.
Genes Dev
.
7
,
1047
1058
.
Williamson
,
A.
and
Lehmann
,
R.
(
1996
).
Germ cell development in Drosophila
.
Annu. Rev. Dev. Biol
.
12
,
365
391
.