Molecular Identification of virilizer, a Gene Required for the Expression of the Sex-Determining Gene Sex-lethal in Drosophila melanogaster

Corresponding author: Rolf Nothiger, Zoological Institute of the University of Zürich, Winterthurerstrasse 190, 8057 Zürich, Switzerland., rolnot{at}zool.unizh.ch (E-mail)

Communicating editor: T. SCHÜPBACH

	ABSTRACT

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Sex-lethal (Sxl) is a central switch gene in somatic sexual development of Drosophila melanogaster. Female-specific expression of Sxl relies on autoregulatory splicing of Sxl pre-mRNA by SXL protein. This process requires the function of virilizer (vir). Besides its role in Sxl splicing, vir is essential for male and female viability and is also required for the production of eggs capable of embryonic development. We have identified vir molecularly and found that it produces a single transcript of 6 kb that is ubiquitously expressed in male and female embryos throughout development. This transcript encodes a nuclear protein of 210 kD that cannot be assigned to a known protein family. VIR contains a putative transmembrane domain, a coiled-coil region and PEST sequences. We have characterized five different alleles of vir. Those alleles that affect both sexes are associated with large truncations of the protein, while alleles that affect only the female-specific functions are missense mutations that lie relatively close to each other, possibly defining a region important for the regulation of Sxl.

THE gene Sex-lethal (Sxl) controls sex determination, dosage compensation, and oogenesis in Drosophila melanogaster. Its state of activity is set around blastoderm stage in somatic cells by the primary sex-determining signal, which is formed by the ratio of X chromosomes to sets of autosomes (X:A ratio; for review see CLINE and MEYER 1996 ). If this ratio is 1.0 (XX:AA), Sxl is activated and a female develops. Alternatively, a ratio of 0.5 (X:AA) leaves Sxl inactive, and male development ensues. SXL is an RNA-binding protein that controls splicing of the transformer (tra) pre-mRNA, its downstream target in sex determination, by binding to and thereby blocking an optimal splice acceptor site (SOSNOWSKI et al. 1989 ; INOUE et al. 1990 ). As a consequence, a cryptic acceptor site located more downstream is used. This splice creates an open reading frame (ORF) such that a functional protein is produced in females only. The downstream target of Sxl in dosage compensation is male specific lethal-2 (msl-2). The production of MSL-2 is blocked in the presence of SXL, thus preventing hyper-transcription of the two X chromosomes. The negative control of msl-2 by SXL occurs at the level of translation and of splicing (BASHAW and BAKER 1997 ; KELLEY et al. 1997 ; GEBAUER et al. 1998 ).

The regulation of Sxl itself is complex and occurs in two steps. First, activation of the gene is transcriptional and relies on an establishment promoter (P_e), which is transcribed only around blastoderm stage in response to a female X:A ratio (XX:AA). Transcripts derived from P_e give rise to early SXL protein. Transcription from this promoter ends shortly after blastoderm stage when a constitutive promoter (P_m) becomes active in both sexes. From now on, Sxl expression is post-transcriptionally regulated. In the presence of early SXL, pre-mRNA derived from P_m is female-specifically spliced by skipping exon 3. This exon contains STOP codons in all three reading frames, and therefore functional late SXL cannot be produced if the exon is present in the mRNA (BELL et al. 1988 ; BOPP et al. 1991 ; KEYES et al. 1992 ). Late SXL is used continuously to splice Sxl pre-mRNA in the female mode, and an autoregulatory feedback loop is established. SXL directs exon skipping by cooperatively binding to multiple uridine-rich stretches within and around the regulated exon (SAKAMOTO et al. 1992 ; HORABIN and SCHEDL 1993A , HORABIN and SCHEDL 1993B ). A male primary signal (X:AA) does not activate P_e, and early SXL is not produced. As a consequence, exon 3 is retained in the Sxl mRNAs when P_m becomes active in males.

Besides Sxl itself, three other genes are needed for Sxl autoregulation. These are sans fille (snf; ALBRECHT and SALZ 1993 ), female-lethal-2-d [fl(2)d; GRANADINO et al. 1990, 1992], and virilizer (vir; HILFIKER and NOTHIGER 1991 ; HILFIKER et al. 1995 ; SCHUTT et al. 1998 ). In all three genes, female-specific mutations exist that affect the expression of Sxl, but most alleles are lethal to both sexes, suggesting also a more general vital role. snf encodes a nuclear protein with functional and sequence similarity to the mammalian U1A and U2B'' snRNP proteins. SXL physically interacts with SNF via its RRM 1 (RNA recognition motif 1; SAMUELS et al. 1998 ). Therefore, snf is probably directly involved in splice site recognition (FLICKINGER and SALZ 1994 ). fl(2)d encodes a novel nuclear protein with an aminoterminal HisGlu-rich domain that is often found in transcriptional regulators (PENALVA et al. 2000 ). FL(2)D is expressed at all stages of development in both sexes. The third gene is virilizer (vir). Three phenotypic classes of alleles are known. One temperature-sensitive allele, vir^1ts, transforms XX animals into intersexes at the restrictive temperature of 29°; this mutation shows that vir is involved in sex determination. A different allele, vir^2f, is an XX-specific lethal that interferes with dosage compensation and sex determination. Alleles in the third and largest class kill both sexes in the third larval instar, implying a vital function of vir unrelated to sex. In cell clones, the lethal alleles cause sexual transformation of XX cells into male structures, and in trans with vir^2f, they cause female-specific lethality, suggesting that they lack the function required for female-specific expression of Sxl. Indeed, in XX animals mutant for vir^2f, Sxl pre-mRNA is spliced in the male mode, indicating that vir is required for the female-specific splicing of Sxl transcripts (HILFIKER et al. 1995 ).

In this article, we investigate the molecular structure and function of vir. In contrast to the complex effects of vir mutations, the gene produces only a single transcript that encodes a large protein. We use the structure of the putative VIR protein to speculate about the molecular mechanism by which vir affects the expression of Sxl and other vital genes.

	MATERIALS AND METHODS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Drosophila stocks and mutagenesis:
Mutations and chromosomes used in this work are described in LINDSLEY and ZIMM 1992 and FlyBase. Additional vir alleles were isolated in a powerful selective F₁ screen that makes use of the male-specific lethality of Sxl^M1, which is rescued by vir^2f (HILFIKER et al. 1995 ). Females of genotype y cm Sxl^M1; vir^2f bw/SM5 were crossed to Adh^D pr cn males that were either irradiated with 40 Gy (Philips MG 160, 2 mm Al-filter, 25 cm distance, 150 kV, 14 mA, 8 min) or fed with diepoxybutan (DEB; REARDON et al. 1987 ). Parents were discarded after 5 days to prevent clustering of new mutations. New vir^- alleles were easily recovered because males, which all receive Sxl^M1 from their mothers, can survive only if they inherit a newly generated vir^- allele from the father in trans with the maternal vir^2f. Thus, progeny were screened for rare XY males of genotype y cm Sxl^M1; vir^2f bw/Adh^D pr cn vir*, where "*" indicates a treated chromosome. Each single male was then crossed to pr cn ix tra2 vir^1ts bw/CyO females, and offspring of genotype XX; pr cn ix tra2 vir^1ts bw/Adh^D pr cn vir* were inspected for the sex-transforming effect of vir* at 29°. Balanced stocks were established and used for genomic Southern analysis (SOUTHERN 1975 ).

Animals used to analyze the vir^1ts, vir^2f, vir⁴, vir²², and vir²³ alleles were obtained as follows: For the two alleles vir^1ts and vir^2f, it was possible to collect homozygous adults (only males in the case of vir^2f) for genomic DNA isolation. The homozygous-lethal vir⁴, vir²², and vir²³ alleles were crossed to Df(2R)130, vir (HILFIKER et al. 1995 ) and recovered as hemizygous third instar larvae. Identification of these hemizygotes was possible because in this cross, all vir alleles were balanced over T(2;3)ap^Xa carrying Blackcell (Bc) as a dominant larval marker.

Southern and Northern analysis:
For Southern analysis, genomic DNA was isolated from adult flies or third instar larvae. Radiolabeled probes (FEINBERG and VOGELSTEIN 1983 ) were prepared from restriction endonuclease fragments and used for hybridization. For Northern analysis, total RNA was isolated according to CHOMCZYNSKI and SACCHI 1987 and was separated on formaldehyde/agarose gels (SAMBROOK et al. 1989 ). Blots contained 30 µg of RNA per lane. Radiolabeled probes were prepared as described for Southern analysis.

In situ hybridization on embryos:
RNA probes used for in situ hybridization were transcribed in vitro from cloned cDNAs with T3 or T7 RNA polymerases according to the manufacturer's instructions (Boehringer Mannheim, Indianapolis). Treatment of the embryos was as described by TAUTZ and PFEIFLE 1989 .

Reverse transcription and PCR amplification:
Total RNA (0.1 µg) was reverse transcribed with Superscript (GIBCO BRL, Gaithersburg, MD) according to the manufacturer's instructions. Aliquots were then amplified by PCR with the appropriate primers.

Construction and screening of libraries:
For walking purposes, genomic libraries based on phages and cosmids were prepared from Oregon-R DNA. High-molecular-weight genomic DNA was partially digested with Sau3A, ligated to EMBL-3 arms (FRISCHAUF et al. 1983 ) or pWE15 (WAHL et al. 1987 ), and packaged in vitro using Gigapack-Gold extracts (Stratagene, La Jolla, CA). All libraries were screened prior to amplification according to the method of BENTON and DAVIS 1977 .

Expression of a VIR-6 x HIS fusion protein in tissue culture cells:
Constructs used for the expression of tagged VIR contained the genomic sequence starting with the triplet following the putative ATG start codon and extending to the last coding triplet. The sequence GSHHHHHH (6 x HIS tag) was fused to the last carboxy-terminal amino acid of VIR by PCR. Expression was driven by the Drosophila tubulin promoter. Schneider cells were transfected with calcium phosphate and harvested after 3 days. For in situ analysis, cells were fixed in formaldehyde, washed three times with PBS/Triton X-100, and incubated with an antibody against the HIS-tag (QIAGEN, Chatsworth, CA). For Western analysis, cells were lysed in a solution containing 8 M urea/50 mM Tris pH 6.8. SDS-PAGE and all following steps were done as described by SAMBROOK et al. 1989 .

	RESULTS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Molecular identification of the vir locus:
Complementation analysis of several deficiencies placed vir in region 59C3-59E1/2 (Fig 1A). The entry point for a genomic walk was provided by the most distal clone from the twist (twi) walk (THISSE et al. 1987 ). This clone spans the proximal breakpoint of Df(2R)bw-S46. The entire walk covers 180 kb distal from this breakpoint. Because none of the previously isolated vir alleles (HILFIKER et al. 1995 ) was detectable within the walk by whole genome Southern analysis, we induced additional vir alleles by X ray and DEB (see MATERIALS AND METHODS). In two independent experiments, 10 X-ray- and 18 DEB-induced alleles were recovered among a total of 45,000 F₁ progeny. All these alleles were homozygous lethal for XX and XY animals and also when trans-heterozygous over previously isolated lethal alleles of vir. Genomic DNA of these mutants was analyzed by genomic Southern blots to detect lesions associated with vir. Of these 28 alleles, three were found to harbor deletions in the 180-kb region. One deletion associated with vir¹⁵ removes 70 kb, whereas the two deletions associated with vir²² and vir²³ both remove 200 bp of DNA. The two latter lesions map to the same 10-kb genomic BamHI restriction fragment (Fig 1B).

View larger version (17K): In this window In a new window Download PPT slide   Figure 1. The isolation and molecular organization of vir. (A) The genomic walk that started near twist and was extended toward brown together with several deficiencies that were useful in localizing vir. A black bar denotes a deficient region. (B) The 10-kb rescue fragment and the two small deficiencies (black boxes indicate deleted sequences) vir22 and vir23 that disrupt the vir ORF. (C) The result of cDNA isolation and sequencing as well as the extent of genomic sequence analysis.

This BamHI fragment, when introduced by P-mediated germline transformation, was able to rescue the effect of the female-specific vir^1ts, vir^2f, and of the non-sex-specific alleles vir²² and vir²³ trans-heterozygous over a deficiency uncovering vir (data not shown). Therefore, this fragment harbors all functions of the vir gene. We used a 7.7-kb EcoRI subfragment of the rescue construct as a probe to identify transcripts and to isolate cDNAs. This probe detected two transcripts of 4.5 and 6 kb, respectively, on a Northern blot (Fig 2A). Three cDNAs were obtained (cVir1-3, Fig 1C), which were all colinear and correspond to the larger 6-kb transcript (Fig 2A, Fig 2). cVir-1 and cVir-2 span the region encompassing the deletions associated with vir²² and vir²³ (Fig 1B and Fig C).

View larger version (68K): In this window In a new window Download PPT slide   Figure 2. (A) Northern analysis. 1: Total RNA isolated from wild-type embryos 0- to 3-hr-old (left lane) and 3- to 17-hr-old (right lane) was hybridized to the vir-rescue fragment (Fig 1B). Detected transcripts are indicated by arrows and correspond to 6 kb and 4.5 kb, respectively. A probe specific for RP49 (O'CONNELL and ROSBASH 1984 ) was used to show that similar amounts of RNA were loaded. 2: Total RNA derived from female adults was hybridized to cVir-1 (Fig 1C). While the 10-kb rescue fragment detected two transcripts, cVir-1 detects only a 6-kb mRNA. 3: Total RNA isolated from female adults (left lane) and male adults (right lane) hybridized to cVir-1. Both sexes show the same 6-kb transcript. (B) PCR analysis of vir22. A pair of primers flanking the deletion of vir22 was used for analysis. 1: Total RNA derived from heterozygous vir22/CyO adult females. 2: cVir-1 in pBluescript. 3: cos89-20. (This cosmid contains vir.) 4: Genomic DNA derived from heterozygous vir22/CyO adult females. 5: Size marker, -DNA digested with EcoRI/HindIII. The resulting fragments were separated on an agarose gel containing ethidium bromide. Lanes 1 and 4 show three instead of two PCR products because wild-type and mutant strands hybridize to form two homo- and one heterodimer. When we denatured the probes before loading, only two bands were found on the gel (data not shown).

To test if the 6-kb transcript that corresponds to the isolated cDNAs is affected by the vir²² mutation, we performed RT-PCR on RNA isolated from vir²²/vir⁺ animals with primers flanking the deletion and compared the resulting products with those amplified from wild-type RNA, wild-type genomic DNA, and mutant genomic DNA. As shown in Fig 2B, there are shorter PCR products present in the reaction with RNA and genomic DNA from vir²²/vir⁺ heterozygous animals. The difference in mobility between the largest and the smallest fragment is 250 bp and corresponds well to the size of the deletion detected in vir²². We conclude that the 6-kb transcript is encoded by vir.

Three small introns were mapped by comparing the vir cDNAs with the genomic sequence. Two additional introns located upstream of cVir-1 were determined by RT-PCR. There is a long open reading frame of 5562 nucleotides (nt) starting from an ATG codon 66 nt upstream of intron 1. This site (CACAACAUG) has a good match with the consensus for Drosophila translation initiation (CAAAACAUG; CAVENER 1987 ) and is the first start codon in the vir ORF. The transcription start site must lie not more then 50 nt upstream of the ATG start codon because RT-PCR with primers further upstream did not yield amplification products indicating that this region is not present as RNA (data not shown). If this is correct, the vir transcript contains a small leader while the trailer consists of 204 nt and has no consensus (AATAAA) polyadenylation signal. Nevertheless, all the isolated vir cDNAs had the same 3' end and contained poly (A) tails.

Structure and subcellular localization of VIR protein:
The deduced amino acid sequence of VIR and the positions of identified motifs are presented in Fig 3. VIR is a large protein of 1854 amino acids (aa) with a calculated molecular weight of 210 kD. A transmembrane domain is predicted at the N terminus of VIR based on the TM-Predict program (HOFMANN and STOFFEL 1993 ). Furthermore, VIR contains a consensus nuclear localization signal (NLS) and several PEST domains, which predict that the protein is targeted to the nucleus and has a short half-life in vivo. Amino acids 779 through 805 and two other regions have a high probability of forming coiled coils (LUPAS et al. 1991 ). In addition, VIR contains a stretch of 144 aa that show homology to domain 6303 of the ProDom database (Fig 4; CORPET et al. 1999 ). Domain 6303 is defined by proteins that interact with nucleic acids such as helicases, hnRNPs, or translation initiation factors. Database searches with the conceptual amino acid sequence of VIR retrieved a human protein of 1795 amino acids. An alignment between VIR and this protein reveals that the two sequences have 28% identical and 40% similar residues. An alignment between VIR and the most conserved parts in the human protein is shown in Fig 5. The human protein contains the putative transmembrane domain found in VIR. This points to the importance of this stretch of amino acids for the function of both proteins.

View larger version (80K): In this window In a new window Download PPT slide   Figure 3. Deduced amino acid sequence of VIR. Predicted structural features are indicated as follows: The boxed region shows homology to domain 6303 of the ProDom database. Positions and effects of mutations in the five analyzed vir alleles are shown above the wild-type amino acid sequence. Affected positions are in boldface. A bar followed by the name of alleles vir22 or vir23 indicates where the protein sequence is interrupted because of a shift in reading frame. PEST sequences, the NLS, and the putative transmembrane domain (TM) are underlined. The region of amino acids with a high probability of forming a coiled coil is shown by arrows above the sequence. The genomic sequence of vir was submitted to GenBank and has the accession no. AF281363. vir corresponds to the gene with no. CG3496 in Gadfly.

View larger version (83K): In this window In a new window Download PPT slide   Figure 4. Amino acids 187 to 331 of VIR show homology to a domain found in several proteins involved in translation, RNA coating, and unwinding. Multiple alignment of VIR(187-331) with ProDom domain 6303. Conserved amino acids are shaded in black. Similar and less well-conserved residues are shaded in gray. The functions, where known, of the proteins in the alignment are: IF2_BORBU is an initiation factor from Borrelia burgdorferi. P90897_CAEEL and Q91372_XENLA are two DEAD-box helicases from Caenorhabditis elegans and Xenopus laevis, respectively. The Xenopus helicase is homologous to D. melanogaster VASA. Q15415_HUMAN is an azoospermia factor candidate and thought to be involved in spermatogenesis. It is predicted to bind to RNA. ROG_HUMAN/MOUSE are components of ribonucleosomes. Boxed amino acids indicate similarity between VIR and ROG_HUMAN/ROG_MOUSE.

View larger version (50K): In this window In a new window Download PPT slide   Figure 5. Alignment (CLUSTAL W, THOMPSON et al. 1994) between VIR and three highly conserved regions in the human protein with accession no. BAA92667 (GenBank). Identical positions are shaded in black and similar positions are shaded in gray. Note the high degree of conservation of residues 109 through 127. This is the predicted transmembrane domain of VIR.

To test whether VIR is indeed a nuclear protein if the size predicted for the conceptual translation product is found in vivo, we transiently expressed VIR as a fusion with a 6 x HIS tag in Drosophila Schneider cells. An antibody against the tag was used to detect the fusion product in situ and on Western blots prepared from total cell extracts. The results are shown in Fig 6. The VIR-6 x HIS fusion protein was found in the nucleus of transfected cells and migrated as a single band of slightly more than 200 kD on Western blots.

View larger version (53K): In this window In a new window Download PPT slide   Figure 6. Expression of a VIR-6 x HIS fusion protein in Schneider cells under the control of the Drosophila tubulin promoter. Cells were transfected and harvested, and a fraction was used for in situ staining while the rest of the cells were lysed and used for Western analysis. (A) 4',6-Diamidino-2-phenylindole staining to visualize the nuclei of the cells. (B) An antibody specific for the 6 x HIS part detects the fusion protein in the nucleus (secondary antibody was FITC labeled). (C) Western blot with total extracts of transfected cells and probed with the same antibody as in B. The fusion protein shows the expected size of 210 kD. Lane 1 contains a control extract. Lane 2 shows an extract derived from cells expressing the fusion protein.

Molecular analysis of vir mutations:
Initial analysis of >30 vir alleles showed that most of them represent lesions undetectable by whole genome Southern analysis (data not shown). Two exceptions were the alleles vir²² and vir²³, which were both associated with small deletions of genomic DNA. Another 5 alleles were also detectable at this level of analysis, but showed rather large deletions of >30 kb each.

To identify potentially relevant domains, we determined the lesions in the two female-specific alleles, vir^1ts and vir^2f, and in three strong alleles, vir⁴, vir²², and vir²³, which are recessive lethals for males and females. We isolated genomic DNA from adult flies homozygous for the sex-specific and viable vir^1ts or vir^2f alleles and from third instar larvae that were hemizygous for the lethal alleles vir⁴, vir²², or vir²³ over Df(2R)vir130 (HILFIKER et al. 1995 ). The effect of each allele can be seen in Fig 3. vir²³ is a 202-bp deletion, and vir²² is a deletion of 295 bp associated with an insertion of 2 bp. Both introduce frameshifts in the ORF so that the wild-type sequence is interrupted. The proteins produced by these two alleles contain 339 and 1261 amino acids of the wild-type conceptual translation product, respectively, followed by a short stretch of amino acids until the next STOP codon. The third lethal allele, vir⁴, is a single base exchange that leads to a nonsense mutation at position 1177 in the deduced amino acid sequence. All three lethal alleles are hence predicted to produce truncated proteins of differnt lengths. The two sex-specific allels, vir^1ts and vir^2f, are point mutations that lead to single amino acid exchanges. In vir^1ts, glutamic acid at position 1423 is mutated into lysine (1423 E-K), and in vir^2f, methionine 1283 is replaced by lysine (1283 M-K). The latter two point mutations lie relatively close together and could define a region in VIR that is required primarily for correct functioning in Sxl autoregulation.

vir mRNA is ubiquitously expressed in males and females and throughout development:
In situ hybridization with vir antisense probes on wild-type embryos detected a weak signal prior to blastoderm stage that was not distinguishable from controls hybridized with sense probes. A first unambiguous signal was detectable in all somatic cells, but not in pole cells, at the formation of blastoderm (Fig 7A and Fig B). Thereafter, vir mRNA remains expressed in all cells, although not at the same level, throughout embryogenesis. There appears to be a gradual decrease in signal intensity from early (gastrulation) toward late embryogenesis (Fig 7, C–E). This was also observed on Northern blots where RNA extracts from 0- to 3-hr-old embryos showed a stronger hybridization signal than RNA derived from 3- to 17-hr-old embryos (Fig 2A, Fig 1). Because a sexually mixed population of embryos did not fall into two classes after in situ hybridization, we conclude that male as well as female embryos ubiquitously express vir RNA at the same level. Furthermore, a single transcript of 6 kb was detected in RNA derived from male and female adults as well as in RNA isolated from embryos and larvae (Fig 2A, Fig 2 and Fig 3).

View larger version (57K): In this window In a new window Download PPT slide   Figure 7. In situ hybridization with digoxygenin-labeled sense and antisense RNA probes derived from cVir-1 (see Fig 1C). Prior to the blastoderm stage, signal intensity was the same after hybridization with sense or antisense probes (not shown). From blastoderm on, signals were found only in those embryos hybridized to antisense RNA. (A) Blastoderm embryo that was incubated with a sense probe. (B) Embryo at blastoderm that starts to express vir-mRNA in all somatic cells. Higher magnification of the posterior pole shows that there is no signal in the pole cells at that stage. (C) vir expression peaks during gastrulation. (D and E) Stage 13 and 17 embryos with decreasing levels of vir expression.

Our in situ hybridizations show that vir mRNA is expressed from blastoderm stage on throughout embryonic development. vir RNA levels appear highest at the beginning of gastrulation and then gradually decrease with increasing age. Little or no vir RNA was detected before blastoderm. This suggests that the maternal contribution to vir transcripts in the zygote is small. This is surprising because it was previously shown (SCHUTT et al. 1998 ) that removal of vir from the female germline results in abortive embryonic development, and this effect could not be rescued by a paternally contributed wild-type copy in the zygote. We see three alternative explanations: First, the mother contributes only low levels of vir transcripts for embryonic development. Second, VIR protein, but not vir mRNA, is deposited into the egg. And the third possibility is that the mother does not deposit any vir activity at all into the developing zygote, but instead maternal vir is required solely during oogenesis to build a functional oocyte and is not required in the early zygote itself.

	DISCUSSION

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

The developmental roles of vir:
Genetic analysis of vir revealed three functions of the gene, namely, a role in sex determination, in dosage compensation, and in a vital process unrelated to sex. These results raise the question of how vir performs these functions, in particular whether the gene might encode different protein variants. The effects of vir mutations on sex determination and dosage compensation are easily explained by the requirement of vir for the female-specific splicing of transcripts of Sxl, which controls both these processes (HILFIKER et al. 1995 ). XX animals mutant for the female-specific allele vir^2f express the gene Sxl in the male mode. Such animals could lack one protein variant that is necessary for female-specific splicing of Sxl pre-mRNA. vir alleles lethal for XX and XY animals might in addition miss another protein variant required for the vital function. Both variants would be missing in the latter class of mutants, which are deficient for the function required for Sxl autoregulation as well as for the vital function. Molecular analysis, however, shows that vir, despite its complex function, produces only a single transcript that probably encodes a single protein. The two XX-specific vir alleles, vir^2f (1283 M-K) and vir^1ts (1423 E-K), cause amino acid substitutions that lie relatively close to each other. These mutations may define a domain in VIR that is essential for female-specific splicing of Sxl pre-mRNAs, but not for the vital function of the protein. This hypothesis implies that a second domain exists that harbors the vital function. However, the genetic data are more in favor of a single protein with a single molecular function. Except for the two female-specific alleles, all the 34 vir mutations isolated so far affect the regulation of Sxl and the vital function. The three lethal alleles vir⁴, vir²², and vir²³ all cause large carboxy-terminal truncations of VIR. These truncations also remove the region mutated in vir^1ts and in vir^2f, which is consistent with the lethal alleles being also deficient in Sxl autoregulation (HILFIKER et al. 1995 ).

We thus favor the interpretation that different threshold requirements exist for vir in Sxl autoregulation and in the vital process, the former being more sensitive than the latter. The female-specific and the lethal mutations may represent weaker (hypomorphic) and stronger (amorphic) alleles of the gene.

VIR is a nuclear protein:
Loss-of-function mutations in vir that cause male-specific splicing of the Sxl pre-mRNA in XX animals suggest a nuclear function of vir in the regulation of Sxl (HILFIKER et al. 1995 ). Consistent with this idea, VIR contains an NLS, and a VIR fusion protein locates to the nucleus in Drosophila tissue culture cells. This result, however, does not yet reveal the specific molecular role of the protein. The amino acid sequence of VIR contains a putative transmembrane domain at the amino terminus. Three regions in the protein have a high probability of forming a coiled coil, which, in other proteins, was demonstrated to serve as an interface for homo- or heterodimeric binding (LUPAS 1996 ). Furthermore, we found a region in VIR with similarity to domain 6303 of the ProDom database. This domain is found in several RNA and DNA helicases, in translation initiation factor 2 (IF2) from Borrelia, and in ribonucleoproteins. The member of the family with the highest similarity to VIR within this domain is a human RNA-binding protein (TrEMBL: Q15415) that is thought to be involved in RNA processing or translation (MA et al. 1993 ). However, there are also striking similarities between VIR and ROG_Human (boxed in Fig 4), which is the human hnRNP G. We tested whether domain 6303 is also present in other hnRNPs that share similar functions. To this end, we scanned 16 hnRNPs for the presence of domain 6303. Indeed, 4 other hnRNPs besides hnRNP G, namely, hnRNP A1, D1, R, and U, contain this domain. The proposed functions of these hnRNPs are diverse. hnRNP A1 is involved in RNA splicing, mRNA transport, and telomere biogenesis; hnRNP D1 is thought to be involved in transcription, while hnRNP U has a function in nuclear retention. No functions have yet been assigned to hnRNP G and R (for a review about hnRNPs see KRECIC and SWANSON 1999 ).

On the other hand, domain 6303 is not present in SXL, SNF, TRA, and TRA2, proteins that are involved in regulation of splicing. To test if this domain is preferentially present in proteins that interact with RNA, we derived a pattern from domain 6303 and used it to scan various data banks (using the PatternFind Server at ISREC, data not shown). We recovered 33 proteins that contained the pattern. Fifteen of them are indeed predicted to interact with RNA (e.g., rRNA methyltransferases, RNA helicases, snRNPs, poly(A)binding proteins) and three are predicted to bind to DNA. Because domain 6303 is also found in proteins without known nucleic acid binding ability and is not present in some proteins that do bind to nucleic acids, it is unlikely that this domain confers binding itself. Rather, it might act as an interface for protein interactions.

The prediction of a transmembrane domain, the nuclear localization of VIR, and the presence of domain 6303 suggest that VIR may be a member of a new class of splice regulators, such as IRE1P, a transmembrane protein that acts as an unconventional splice factor in the unfolded protein response pathway in yeast (WELIHINDA et al. 1999 ). However, no sequence homology exists between VIR and IRE1P. The characteristics of VIR lead us to speculate that the protein is located in the nuclear membrane where it may mediate mRNA transport. If this were correct, this process could be severely disrupted in strong vir mutants, with mRNAs of vital genes being affected, which would cause sex-unspecific lethality. However, even presumably amorphic alleles are not cell lethal in genetic mosaics (HILFIKER et al. 1995 ). In contrast, null alleles of snf, which encodes a component of the general splicing machinery, are incompatible with cell survival (FLICKINGER and SALZ 1994 ). Unlike the strong alleles of vir, the female-specific alleles (vir^1ts and vir^2f) may reduce effectiveness of the nuclear transportation system such that dose-sensitive processes begin to fail. That autoregulation of Sxl is indeed a sensitive system was recently demonstrated by STITZINGER et al. 1999 . Even mutations in aspartyl- as well as in tryptophanyl tRNA synthetase of Drosophila can act as maternal modifiers and disrupt the autoregulation of Sxl. However, we realize that the HIS-tagged VIR fusion protein was found in the nucleoplasm of Schneider cells (Fig 6) and not in the membrane. We think that this finding does not exclude the possibility of membrane insertion because expression might not be physiological in this experiment and staining from excess protein in the nucleoplasm could mask staining at the periphery. In addition, VIR could be localized to the nucleoplasm and to the membrane also in wild type. These two possibilities are compatible with our speculation about a function in nuclear transport.

Sxl is not the only gene that requires the function of vir. Female-specific expression of tra and msl-2 depends on vir even when SXL protein is present. This is seen in XX animals mutant for vir^2f, which can be partially rescued by the constitutive allele Sxl^M4. However, such females are strongly masculinized because the transcripts of tra are not efficiently spliced in the female mode, and their low survival rate points to only partial repression of msl-2 (HILFIKER et al. 1995 ). We cannot rule out that vir is directly involved in splicing of Sxl and tra pre-mRNA, but VIR is not an essential component of the general splicing machinery because even amorphic vir alleles allow growth and differentiation of imaginal discs. Instead, as recently proposed by PENALVA et al. 2000 for FL(2)D, VIR might be used to splice a subset of pre-mRNAs. Among those are the transcripts of Sxl and tra, which are both needed in females only. Another target is the primary transcript of the Ubx gene (BURNETTE et al. 1999 ). In the absence of VIR or FL(2)D, microexons mI and mII are not efficiently included in Ubx transcripts. There are probably other essential genes whose splicing depends on vir and fl(2)d. The sex-unspecific lethality associated with amorphic alleles of vir could be caused by the failure to splice the pre-mRNAs of such targets in the correct manner.

	ACKNOWLEDGMENTS

We are especially grateful to Daniel Bopp for many discussions and also for his contribution to some of the experiments. Many thanks also go to Nathalie Méthod for her help in the transformation of Schneider cells and for providing vector DNA. This work was supported by the Ernst Hadorn Stiftung, by the Julius Klaus Stiftung, by the Kanton Zürich, and by grant 31-47180.96 of the Swiss National Foundation.

Manuscript received September 6, 2000; Accepted for publication October 25, 2000.

	LITERATURE CITED

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

ALBRECHT, E. B. and H. K. SALZ, 1993  The Drosophila sex determination gene snf is utilized for the establishment of the female-specific splicing pattern of Sex-lethal.. Genetics 134:801-807[Abstract].

BASHAW, G. J. and B. S. BAKER, 1997  The regulation of the Drosophila msl-2 gene reveals a function for Sex-lethal in translational control. Cell 89:789-798[Medline].

BELL, L. R., E. M. MAINE, P. SCHEDL, and T. W. CLINE, 1988  Sex-lethal, a Drosophila sex determination switch gene, exhibits sex-specific RNA splicing and sequence similarity to RNA binding proteins. Cell 55:1037-1046[Medline].

BENTON, W. D. and R. W. DAVIS, 1977  Screening lambda gt recombinant clones by hybridization to single plaques in situ. Science 196:180-182[Abstract/Free Full Text].

BOPP, D., L. R. BELL, T. W. CLINE, and P. SCHEDL, 1991  Developmental distribution of female-specific Sex-lethal proteins in Drosophila melanogaster.. Genes Dev. 5:403-415[Abstract/Free Full Text].

BURNETTE, J. M., A. R. HATTON, and A. J. LÓPEZ, 1999  Trans-acting factors required for inclusion of regulated exons in the Ultrabithorax mRNAs of Drosophila melanogaster.. Genetics 151:1517-1529[Abstract/Free Full Text].

CAVENER, D. R., 1987  Comparison of the consensus sequence flanking translational start sites in Drosophila and vertebrates. Nucleic Acids Res. 15:1353-1361[Abstract/Free Full Text].

CHOMCZYNSKI, P. and N. SACCHI, 1987  Single-step method of RNA isolation by acid guanidinium thiocyanate-phenol-chloroform extraction. Anal. Biochem. 162:156-159[Medline].

CLINE, T. W. and B. J. MEYER, 1996  Vive la différence: males vs females in flies vs worms. Annu. Rev. Genet. 30:637-702[Medline].

CORPET, F., J. GOUZY, and D. KAHN, 1999  Recent improvements of the ProDom database of protein domain families. Nucleic Acids Res. 27:263-267[Abstract/Free Full Text].

FEINBERG, A. P. and B. VOGELSTEIN, 1983  A technique for radiolabeling DNA restriction endonuclease fragments to high specific activity. Anal. Biochem. 132:6-13[Medline].

FLICKINGER, T. W. and H. K. SALZ, 1994  The Drosophila sex determination gene snf encodes a nuclear protein with sequence and functional similarity to the mammalian U1A snRNP protein. Genes Dev. 8:914-925[Abstract/Free Full Text].

FRISCHAUF, A. M., H. LEHRACH, A. POUSTKA, and N. MURRAY, 1983  Lambda replacement vectors carrying polylinker sequences. J. Mol. Biol. 170:827-842[Medline].

GEBAUER, F., L. MERENDINO, M. W. HENTZE, and J. VALCÁRCEL, 1998  The Drosophila splicing regulator Sex-lethal directly inhibits translation of male-specific-lethal 2 mRNA. RNA 4:142-150[Abstract].

GRANADINO, B., S. CAMPUZANO, and L. SÁNCHEZ, 1990  The Drosophila melanogaster fl(2)d gene is needed for the female-specific splicing of Sex-lethal RNA. EMBO J 9:2597-2602[Medline].

GRANADINO, B., A. SAN JUAN, P. SANTAMARIA, and L. SÁNCHEZ, 1992  Evidence of a dual function in fl(2)d, a gene needed for Sex-lethal expression in Drosophila melanogaster.. Genetics 130:597-612[Abstract].

HILFIKER, A. and R. NOTHIGER, 1991  The temperature-sensitive mutation vir^ts (virilizer) identifies a new gene involved in sex determination of Drosophila. Roux's Arch. Dev. Biol. 200:240-248.

HILFIKER, A., H. AMREIN, A. DÜBENDORFER, R. SCHNEITER, and R. NOTHIGER, 1995  The gene virilizer is required for female-specific splicing controlled by Sxl, the master gene for sexual development in Drosophila. Development 121:4017-4026[Abstract].

HOFMANN, K. and W. STOFFEL, 1993  TMbase—a database of membrane spanning protein segments. Biol. Chem. Hoppe-Seyler 374:166.

HORABIN, J. I. and P. SCHEDL, 1993a  Regulated splicing of the Drosophila Sex-lethal male exon involves a blockage mechanism. Mol. Cell. Biol. 13:1408-1414[Abstract/Free Full Text].

HORABIN, J. I. and P. SCHEDL, 1993b  Sex-lethal autoregulation requires multiple cis-acting elements upstream and downstream of the male exon and appears to depend largely on controlling the use of the male exon 5' splice site. Mol. Cell. Biol. 13:7734-7746[Abstract/Free Full Text].

INOUE, K., K. HOSHIJIMA, H. SAKAMOTO, and Y. SHIMURA, 1990  Binding of the Drosophila Sex-lethal gene product to the alternative splice site of transformer primary transcript. Nature 344:461-463[Medline].

KELLEY, R. L., J. WANG, L. BELL, and M. I. KURODA, 1997  Sex-lethal controls dosage compensation in Drosophila by a non-splicing mechanism. Nature 387:195-199[Medline].

KEYES, L. N., T. W. CLINE, and P. SCHEDL, 1992  The primary sex determination signal of Drosophila acts at the level of transcription. Cell 68:933-943. (erratum: Cell 69: following 572).[Medline].

KRECIC, A. M. and M. S. SWANSON, 1999  hnRNP complexes: composition, structure and function. Curr. Opin. Cell Biol. 11:363-371[Medline].

LINDSLEY, D. L., and G. G. ZIMM, 1992 The Genome of Drosophila melanogaster. Academic Press, San Diego.

LUPAS, A., 1996  Coiled coils: new structures and new functions. Trends Biochem. Sci. 21:375-382[Medline].

LUPAS, A., M. VAN DYKE, and J. STOCK, 1991  Predicting coiled coils from protein sequences. Science 252:1162-1164[Medline].

MA, K., J. D. INGLIS, A. SHARKEY, W. A. BICKMORE, and R. E. HILL et al., 1993  A Y chromosome gene family with RNA-binding protein homology: candidates for the azoospermia factor AZF controlling human spermatogenesis. Cell 75:1287-1295[Medline].

O'CONNELL, P. O. and M. ROSBASH, 1984  Sequence, structure and codon preference of the Drosophila ribosomal protein 49 gene. Nucleic Acids Res. 12:5495-5513[Abstract/Free Full Text].

PEÑALVA, L. O. F., M. F. RUIZ, A. ORTEGA, B. GRANADINO, and L. VICENTE et al., 2000  The Drosophila fl(2)d gene, required for female-specific splicing of Sxl and tra pre-mRNAs, encodes a novel nuclear protein with a HQ-rich domain. Genetics 155:129-139[Abstract/Free Full Text].

REARDON, J. T., C. A. LILJESTRAND-GOLDEN, R. L. DUSENBERY, and P. D. SMITH, 1987  Molecular analysis of diepoxybutane-induced mutations at the rosy locus of Drosophila melanogaster.. Genetics 115:323-331[Abstract/Free Full Text].

SAKAMOTO, H., K. INOUE, I. HIGUCHI, Y. ONO, and Y. SHIMURA, 1992  Control of Drosophila Sex-lethal pre-mRNA splicing by its own female-specific product. Nucleic Acids Res. 20:5533-5540[Abstract/Free Full Text].

SAMBROOK, J., E. F. FRITSCH and T. MANIATIS, 1989 Molecular Cloning: A Laboratory Manual. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY.

SAMUELS, M., G. DESHPANDE, and P. SCHEDL, 1998  Activities of the Sex-lethal protein in RNA binding and protein:protein interactions. Nucleic Acids Res. 26:2625-2637[Abstract/Free Full Text].

SCHUTT, C., A. HILFIKER, and R. NOTHIGER, 1998  virilizer regulates Sex-lethal in the germline of Drosophila melanogaster.. Development 125:1501-1507[Abstract].

SOSNOWSKI, B. A., J. M. BELOTE, and M. MCKEOWN, 1989  Sex-specific alternative splicing of RNA from the transformer gene results from sequence-dependent splice site blockage. Cell 58:449-459[Medline].

SOUTHERN, E. M., 1975  Detection of specific sequences among DNA fragments separated by gel electrophoresis. J. Mol. Biol. 98:503-517[Medline].

STITZINGER, S. M., A. PELLICENA-PALLE, E. B. ALBRECHT, K. M. GAJEWSKI, and K. M. BECKINGHAM et al., 1999  Mutations in the predicted aspartyl tRNA synthetase of Drosophila are lethal and function as dosage-sensitive maternal modifiers of the sex determination gene Sex-lethal.. Mol. Gen. Genet. 261:142-151[Medline].

TAUTZ, D. and C. PFEIFLE, 1989  A non-radioactive in situ hybridization method for the localization of specific RNAs in Drosophila embryos reveals translational control of the segmentation gene hunchback.. Chromosoma 98:81-85[Medline].

THISSE, B., M. EL MESSAL, and F. PERRIN-SCHMITT, 1987  The twist gene: isolation of a Drosophila zygotic gene necessary for the establishment of dorsoventral pattern. Nucleic Acids Res. 15:3439-3453[Abstract/Free Full Text].

THOMPSON, J. D., D. G. HIGGINS, and T. J. GIBSON, 1994  CLUSTAL W: improving the sensitivity of progressive multiple sequence alignment through sequence weighting, position-specific gap penalties and weight matrix choice. Nucleic Acids Res. 22:4673-4680[Abstract/Free Full Text].

WAHL, G. M., K. A. LEWIS, J. C. RUIZ, B. ROTHENBERGER, and J. ZHAO et al., 1987  Novel cosmid vectors for genomic walking, restriction mapping and gene transfer. Proc. Natl. Acad. Sci. USA 84:2160-2164[Abstract/Free Full Text].

WELIHINDA, A. A., W. TIRASOPHON, and R. J. KAUFMAN, 1999  The cellular response to protein misfolding in the endoplasmic reticulum. Gene Expr. 7:293-300[Medline].

This article has been cited by other articles:

				L. O. F. Penalva and L. Sanchez RNA Binding Protein Sex-Lethal (Sxl) and Control of Drosophila Sex Determination and Dosage Compensation Microbiol. Mol. Biol. Rev., September 1, 2003; 67(3): 343 - 359. [Abstract] [Full Text] [PDF]

				A. Ortega, M. Niksic, A. Bachi, M. Wilm, L. Sanchez, N. Hastie, and J. Valcarcel Biochemical Function of Female-Lethal (2)D/Wilms' Tumor Suppressor-1-associated Proteins in Alternative Pre-mRNA Splicing J. Biol. Chem., January 24, 2003; 278(5): 3040 - 3047. [Abstract] [Full Text] [PDF]