m6A potentiates Sxl alternative pre-mRNA splicing for robust Drosophila sex determination

N6-methyladenosine (m6A) is the most common internal modification of eukaryotic messenger RNA (mRNA) and is decoded by YTH domain proteins. The mammalian mRNA m6A methylosome is a complex of nuclear proteins that includes a stable heterodimer [METTL3 (methyltransferase-like 3) and METTL14], WTAP (Wilms tumour 1-associated protein) and KIAA1429. Drosophila has corresponding homologues named Ime4 (Inducer of meiosis 4) , Mettl14 (Methyltransferase-like 14 ), the Wilms tumour 1-associated protein Female-lethal (2)d (Fl(2)d) and Virilizer (Vir). In Drosophila, fl(2)d and vir are required for sex-dependent regulation of alternative splicing of the sex determination factor Sex lethal (Sxl). However, the functions of m6A in introns in the regulation of alternative splicing remain uncertain. This study shows that m6A is absent in the mRNA of Drosophila lacking Ime4. In contrast to mouse and plant knockout model, Drosophila Ime4-null mutants remain viable, though flightless, and show a sex bias towards maleness. This is because m6A is required for female-specific alternative splicing of Sxl, which determines female physiognomy, but also translationally represses male-specific lethal 2 (msl-2) to prevent dosage compensation in females. The m6A reader protein YT521-B decodes m6A in the sex-specifically spliced intron of Sxl, as its absence phenocopies Ime4 mutants. Loss of m6A also affects alternative splicing of additional genes, predominantly in the 5' untranslated region, and has global effects on the expression of metabolic genes. The requirement of m6A and its reader YT521-B for female-specific Sxl alternative splicing reveals that this hitherto enigmatic mRNA modification constitutes an ancient and specific mechanism to adjust levels of gene expression (Haussmann, 2016).

In mature mRNA the m6A modification is most prevalently found around the stop codon as well as in 5' untranslated regions (UTRs) and in long exons in mammals, plants and yeast. Since methylosome components predominantly localize to the nucleus, it has been speculated that m6A localized in pre-mRNA introns could have a role in alternative splicing regulation in addition to such a role when present in long exons. This prompted an investigation of whether m6A is required for Sxl alternative splicing, which determines female sex and prevents dosage compensation in females. A null allele of the Drosophila METTL3 methyltransferase homologue Ime4 was induced by imprecise excision of a P element inserted in the promoter region. The excision allele Δ22-3 deletes most of the protein-coding region, including the catalytic domain, and is thus referred to as Ime4null. These flies are viable and fertile, but both flightless and this phenotype can be rescued by a genomic construct restoring Ime4. Ime4 shows increased expression in the brain and, as in mammals and plants, localizes to the nucleus (Haussmann, 2016).

Following RNase T1 digestion and 32P end-labelling of RNA fragments, m6A was detected after guanosine (G) in poly(A) mRNA of adult flies at relatively low levels compared to other eukaryotes, but at higher levels in unfertilized eggs. After enrichment with an anti-m6A antibody, m6A is readily detected in poly(A) mRNA, but absent from Ime4null flies (Haussmann, 2016).

As found in other systems, and consistent with a potential role in translational regulation, m6A was detected in polysomal mRNA, but not in the poly(A)-depleted rRNA fraction. This also confirmed that any m6A modification in rRNA is not after G in Drosophila (Haussmann, 2016).

Consistent with the hypothesis that m6A plays a role in sex determination and dosage compensation, the number of Ime4null females was reduced to 60% compared to the number of males, whereas in the control strain female viability was 89%. The key regulator of sex determination in Drosophila is the RNA-binding protein Sxl, which is specifically expressed in females. Sxl positively auto-regulates expression of itself and its target transformer (tra) through alternative splicing to direct female differentiation. In addition, Sxl suppresses translation of msl-2 to prevent upregulation of transcription on the X chromosome for dosage compensation; full suppression also requires maternal factors. Accordingly, female viability was reduced to 13% by removal of maternal m6A together with zygotic heterozygosity for Sxl and Ime4 (Ime4Δ22-3 females crossed with Sxl7B0 males, a Sxl null allele). Female viability of this genotype is completely rescued by a genomic construct or by preventing ectopic activation of dosage compensation by removal of msl-2. Hence, females are non-viable owing to insufficient suppression of msl-2 expression, resulting in upregulation of gene expression on the X chromosome from reduced Sxl levels. In the absence of msl-2, disruption of Sxl alternative splicing resulted in females with sexual transformations displaying male-specific features such as sex combs, which were mosaic to various degrees, indicating that Sxl threshold levels are affected early during establishment of sexual identities of cells and/or their lineages. In the presence of maternal Ime4, Sxl and Ime4 do not genetically interact (Sxl7B0/FM7 females crossed with Ime4null males, 103% female viability. In addition, Sxl is required for germline differentiation in females and its absence results in tumorous ovaries. Consistent with this, tumorous ovaries in Sxl7B0/+;Ime4null/+ daughters from Ime4null females or heterozygous Sxl7B0 females (Haussmann, 2016).

Furthermore, levels of the Sxl female-specific splice form were reduced to approximately 50%, consistent with a role for m6A in Sxl alternative splicing . As a result, female-specific splice forms of tra and msl-2 were also significantly reduced in adult females (Haussmann, 2016).

To obtain more comprehensive insights into Sxl alternative splicing defects in Ime4null females, splice junction reads were examined from RNA-seq. Besides the significant increase in inclusion of the male-specific Sxl exon in Ime4null females, cryptic splice sites and increased numbers of intronic reads were detected in the regulated intron. Consistent with reverse transcription polymerase chain reaction (RT–PCR) analysis of tra, the reduction of female splicing in the RNA sequencing is modest, and as a consequence, alternative splicing differences of Tra targets dsx and fru were not detected in whole flies, suggesting that cell-type-specific fine-tuning is required to generate splicing robustness rather than being an obligatory regulator. In agreement with dosage-compensation defects as a main consequence of Sxl dysregulation in Ime4null mutants, X-linked, but not autosomal, genes are significantly upregulated in Ime4null females compared to controls (Haussmann, 2016).

Furthermore, Sxl mRNA is enriched in pull-downs with an m6A antibody compared to m6A-deficient yeast mRNA added for quantification. This enrichment is comparable to what was observed for m6A-pull-down from yeast mRNA (Haussmann, 2016).

To map m6A sites in the intron of Sxl, an in vitro m6A methylation assay was employed using Drosophila nuclear extracts and labelled substrate RNA. m6A methylation activity was detected in the vicinity of alternatively spliced exons. Further fine-mapping localized m6A in RNAs C and E to the proximity of Sxl-binding sites. Likewise, the female-lethal single amino acid substitution alleles fl(2)d1 and vir2F interfere with Sxl recruitment, resulting in impaired Sxl auto-regulation and inclusion of the male-specific exon. Female lethality of these alleles can be rescued by Ime4null heterozygosity, further demonstrating the involvement of the m6A methylosome in Sxl alternative splicing (Haussmann, 2016).

Next, alternative splicing changes was globally analyzed in Ime4null females compared to the wild-type control strain. A statistically significant reduction in female-specific alternative splicing of Sxl was observed. In addition, 243 alternative splicing events in 163 genes were significantly different in Ime4null females, equivalent to around 2% of alternatively spliced genes in Drosophila. Six genes for which the alternative splicing products could be distinguished on agarose gels were confirmed by RT-PCR. Notably, lack of Ime4 did not affect global alternative splicing and no specific type of alternative splicing event was preferentially affected. However, alternative first exon (18% versus 33%) and mutually exclusive exon (2% versus 15%) events were reduced in Ime4null compared to a global breakdown of alternative splicing in wild-type Drosophila, mostly to the extent of retained introns (16% versus 6%), alternative donor (16% versus 9%) and unclassified events (14% versus 6%). Notably, the majority of affected alternative splicing events in Ime4null were located to the 5' UTR, and these genes had a significantly higher number of AUG start codons in their 5' UTR compared to the 5' UTRs of all genes. Such a feature has been shown to be relevant to translational control under stress conditions. (Haussmann, 2016 and references therein).

The majority of the 163 differentially alternatively spliced genes in Ime4 females are broadly expressed (59%), while most of the remainder are expressed in the nervous system (33%), consistent with higher expression of Ime4 in this tissue. Accordingly, Gene Ontology analysis revealed a highly significant enrichment for genes involved synaptic transmission (Haussmann, 2016).

Since the absence of m6A affects alternative splicing, m6A marks are probably deposited co-transcriptionally before splicing. Co-staining of polytene chromosomes with antibodies against haemagglutinin (HA)-tagged Ime4 and RNA Pol II revealed broad co-localization of Ime4 with sites of transcription, but not with condensed chromatin-visualized with antibodies against histone H4. Furthermore, localization of Ime4 to sites of transcription is RNA-dependent, as staining for Ime4, but not for RNA Pol II, was reduced in an RNase-dependent manner (Haussmann, 2016).

Although m6A levels after G are low in Drosophila compared to other eukaryotes, broad co-localization of Ime4 to sites of transcription suggests profound effects on the gene expression landscape. Indeed, differential gene expression analysis revealed 408 differentially expressed genes where 234 genes were significantly upregulated and 174 significantly downregulated in neuron-enriched head/thorax of adult Ime4null females. Cataloguing these genes according to function reveals prominent effects on gene networks involved in metabolism, including reduced expression of 17 genes involved in oxidative phosphorylation. Notably, overexpression of the m6A mRNA demethylase FTO in mice leads to an imbalance in energy metabolism resulting in obesity (Haussmann, 2016).

Next, tests were performed to see whether either of the two substantially divergent YTH proteins, YT521-B and CG6422, decodes m6A marks in Sxl mRNA. When transiently transfected into male S2 cells, YT521-B localizes to the nucleus, whereas CG6422 is cytoplasmic. Nuclear YT521-B can switch Sxl alternative splicing to the female mode and also binds to the Sxl intron in S2 cells. In vitro binding assays with the YTH domain of YT521-B demonstrate increased binding of m6A-containing RNA. In vivo, YT521-B also localizes to the sites of transcription (Haussmann, 2016).

To further examine the role of YT521-B in decoding m6A Drosophila strain YT521-BMI02006 was analyzed, where a transposon in the first intron disrupts YT521-B. This allele is also viable, and phenocopies the flightless phenotype and the female Sxl splicing defect of Ime4null flies. Likewise, removal of maternal YT521-B together with zygotic heterozygosity for Sxl and YT521-B reduces female viability and results in sexual transformations such as male abdominal pigmentation. In addition, overexpression of YT521-B results in male lethality, which can be rescued by removal of Ime4, further reiterating the role of m6A in Sxl alternative splicing. Since YT521-B phenocopies Ime4 for Sxl splicing regulation, it is the main nuclear factor for decoding m6A present in the proximity of the Sxl-binding sites. YT521-B bound to m6A assists Sxl in repressing inclusion of the male-specific exon, thus providing robustness to this vital gene regulatory switch (Haussmann, 2016).

Nuclear localization of m6A methylosome components suggested a role for this 'fifth' nucleotide in alternative splicing regulation. The discovery of the requirement of m6A and its reader YT521-B for female-specific Sxl alternative splicing has important implications for understanding the fundamental biological function of this enigmatic mRNA modification. Its key role in providing robustness to Sxl alternative splicing to prevent ectopic dosage compensation and female lethality, together with localization of the core methylosome component Ime4 to sites of transcription, indicates that the m6A modification is part of an ancient, yet unexplored mechanism to adjust gene expression. Hence, the recently reported role of m6A methylosome components in human dosage compensation further support such a role and suggests that m6A-mediated adjustment of gene expression might be a key step to allow for the development of the diverse sex determination mechanisms found in nature (Haussmann, 2016).

In situ hybridization with vir antisense probes gives an unambiguous signal and is detectable in all somatic cells, but not in pole cells, at the formation of blastoderm. 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. 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. Because a sexually mixed population of embryos did not fall into two classes after in situ hybridization, it is concluded that male as well as female embryos ubiquitously express vir RNA at the same level. Furthermore, a single transcript of 6 kb is detected in RNA derived from male and female adults as well as in RNA isolated from embryos and larvae (Niessen, 2001).

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 is detected before blastoderm. This suggests that the maternal contribution to vir transcripts in the zygote is small. This is surprising because has been shown (Schutt, 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. Three alternative explanations must be considered: (1) the mother contributes only low levels of vir transcripts for embryonic development; (2) Vir protein, but not VIR mRNA, is deposited into the egg; (3) it is possible 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 (Niessen, 2001).

Effects of Mutation or Deletion

The gene virilizer is needed for dosage compensation and sex determination in females and for an unknown vital function in both sexes. In genetic mosaics, XX somatic cells mutant for vir differentiate male structures. One allele, vir2f, is lethal for XX, but not for XY animals. This female-specific lethality can be rescued by constitutive expression of Sxl or by mutations in msl (male-specific lethal) genes. Rescued animals develop as strongly masculinized intersexes or pseudomales. They have male-specifically spliced mRNA of tra, and when rescued by msl, also of Sxl. These data indicate that vir is a positive regulator of female-specific splicing of Sxl and of tra pre-mRNA (Hilfiker, 1995).

All the tested alleles that cause lethality to males and females are also deficient for the female-specific functions of dosage compensation and sex determination. These pathways are controlled by Sxl, and vir could thus be a regulator of Sxl. The role of Sxl is well established in the pathway of somatic sex determination. Its protein, besides regulating the splicing of its own pre-mRNA (autoregulation), promotes the female-specific splicing of the tra pre-mRNA (transregulation) by blocking the 3' splice site that is used in males. The most obvious and best understood aspect of vir is its role in sex determination, as demonstrated by XX;vir mutant cells forming male structures in genetic mosaics. The temperature-sensitive intersexual phenotype of X/X;vir1ts/vir1ts animals is rescued by the construct hs-tra-female that constitutively expresses the female mode of tra (Hilfiker, 1991). This indicates that vir acts upstream of tra. The new findings confirm the high hierarchical position of vir in the cascade of sex-determining genes and suggest that it plays a role in the regulation of Sxl and tra (Hilfiker, 1995).

These genetic arguments receive support from molecular analyses that show that surviving XX; vir2f mle pseudomales produce mRNAs of Sxl and tra that are male-specifically spliced. The male-specific products of tra could be the consequence of the male-specific expression of Sxl. The next experiments, however, indicate that vir is also directly engaged in the splicing of tra transcripts (Hilfiker, 1995).

Since SxlM/+;vir2f animals are efficiently rescued from the female lethal effect of vir2f, but remain sexually transformed pseudomales or strongly masculinized intersexes, it is concluded that the sex-transforming effect of vir mutations is not exerted via Sxl alone. When the female-specific function of Sxl is constitutively provided in XX;vir2f animals by SxlM or SxlcF1#19, the female-specific function of tra is still not guaranteed, but in addition appears to require an active product of vir. To test this conclusion the rescued animals were subjected to Western analysis. Substantial levels of Sxl protein are present in strong intersexes and pseudomales. Even animals with normal levels of Sxl, such as the intersexes, still splice the pre-mRNA of their tra gene largely in the male mode, as indicated by their strongly masculinized appearance. These intersexes are not mosaics of male and female cells, but display an intermediate sexual phenotype at the cellular level characteristic of flies expressing only low levels of female-specific products of tra. The results thus support the conclusion that vir is needed not only for female-specific expression of Sxl, but also of tra. This may also be the case for fl(2)d, although to a lesser extent, since XX animals mutant for fl(2)d are rescued by SxlM1 and are sexually transformed into intersexes at 29°C (Hilfiker, 1995).

The Western data furthermore suggest that presence of Sxl protein still requires vir for the splicing of its own pre-mRNA: pseudomales have substantially less Sxl protein than females which suggests that the transcripts produced by the Sxl plus allele in these genotypes are efficiently spliced only in the presence of vir plus. Based on these results, vir belongs to the class of genes that are involved in the regulation of Sxl. Recessive mutations in the genes fl(2)d and snf have been shown to prevent female-specific splicing of Sxl pre-mRNA in XX animals; and, like vir2f, the female-specific mutant effects of both genes are rescued by SxlM mutations (Hilfiker, 1995 and references therein).

The sex-specific lethality of XX animals mutant for vir2f by itself is suggestive of a role for vir in dosage compensation. Interactions between vir2f and mutations in genes known to play a role in dosage compensation, such as Sxl and the male-specific lethal genes (msl genes) further support the proposed role of vir. XX animals with otherwise viable combinations of vir alleles, e.g. vir1ts/vir2f, die when they have only a single dose of Sxl plus. However, animals with two X chromosomes that are mutant for vir2f are rescued by constitutive expression of Sxl. These results indicate that vir mutations achieve their female-specific lethal effect via Sxl, which explains why males are not affected by these same allelic combinations (Hilfiker, 1995).

Animals with two X chromosomes and mutant for vir2f are also rescued, although with low frequency, by loss of function of msl genes. In view of the sex-transforming effect of mutations in vir, and since mutations in msl genes do not interfere with sexual differentiation, it is not surprising that the rescued animals are transformed into sterile pseudo-males. It is concluded that vir2f allows the inappropriate activity of the msl genes in XX zygotes, and that the affected animals then die as a result of overexpression of X-linked genes. Elimination of any one of the msl genes lowers X-chromosomal transcription towards a level that is lethal for XY animals, but is appropriate for a single X in XX animals. However, the fact that the rescue of XX;vir2f animals by mutations in msl genes is weak, is a warning that the regulatory network of dosage compensation is more complex. Sxl controls an early female-specific vital function that is not dependent on msl gene activity (Hilfiker, 1995 and references therein).

This interpretation is supported by the recent finding that the Mle and Msl-1 proteins bind to the polytene X-chromosomes in salivary glands and Malpighian tubules of XX animals mutant for vir2f or vir1ts; furthermore, these animals have the male-specifically acetylated form H4Ac16 of histone H4. Binding of the Mle and Msl-1 proteins to the X-chromosome and acetylation of histone H4 at lysine 16 normally occur only in males and are strongly, although not absolutely, correlated with hypertranscription of their single X-chromosome (Hilfiker, 1995).

How could vir participate in the regulation of the male specific lethal (msl) genes? Recent studies have shown that, of all the msl genes [mle; msl-1 and msl-3, only msl-2 is differentially expressed due to sex-specific splicing -- only males are capable of producing a functional protein; in females, the productive splice seems to be prevented by Sxl. Thus, it appears that the Sxl protein acts again, as in the regulation of its own mRNA and that of tra, by blocking a splice site in the msl-2 transcripts. Since the rescue of XX;vir2f animals by SxlM is not complete and since the SxlM1/Y males survive despite the presence of Sxl, it is concluded that the Sxl product requires vir function to efficiently prevent the male-specific processing of msl-2 transcripts (Hilfiker, 1995).

It was unexpected that XX;vir2f animals rescued by SxlM were transformed into pseudomales or strongly masculinized intersexes. A simple model in which vir acts above Sxl predicts that the rescued animals, due to the female-determining effect of Sxl, would be females; alternatively, if vir acted below Sxl, there would be no rescue of the lethality. The actual results are best interpreted by a model in which vir acts at both levels, upstream and downstream of Sxl, but ostensibly with differential effectiveness. Its function appears to be absolutely required for female-specific splicing of Sxl transcripts, but seems less important for the regulation of tra, and even less for msl-2 which is assumed to be the other target. This is inferred from the observation that the function provided by SxlM1 or SxlM4 in XX;vir2f animals is largely, but not completely, sufficient to prevent the inappropriate activity of the msl genes, but insufficient to make enough female-specific products of tra necessary for female sexual development (Hilfiker, 1995).

vir2f rescues XY males from the male-specific lethal effect of SxlM1 and partially of SxlcF1#19, but not of SxlM4. These results can be understood by recalling that SxlM1, in contrast to SxlM4, is not unconditionally constitutive, and XY animals mutant for SxlM1 and snf1621 or fl(2)d1 can survive as males, some of which are even fertile. Surviving males of genotype SxlM1/Y;vir2f do produce some Sxl protein, but not enough to make vir function dispensable. This implies that SxlM1, in addition to the functions of snf plus and fl(2)d plus, also requires vir plus to become fully functional in XY animals, and hence that an active product of vir must also be present in males (Hilfiker, 1995).

If vir is active in males, as these results imply, then it is unlikely to play a discriminatory role in sex determination. Rather, it may be comparable to da and tra2, two genes whose products are present in both sexes, but serve sex-determining functions only in females (Hilfiker, 1995).

Most alleles of vir are zygotic lethals. Since the first screen, many more vir alleles have been generated, and still only two, vir1ts and vir2f, are truly female-specific. This indicates that most lesions induced in vir affect a general non sex-specific function which, however, is dispensable in somatic cells of genetic mosaics: homozygous mutant cells survive in clones and in transplanted imaginal discs (Hilfiker, 1995).

Female-specific alleles and regular lethal alleles are also known for snf and fl(2)d, two genes that are, like vir, involved in the female-specific splicing of Sxl. Whereas putative null alleles of vir and fl(2)d are viable in cell clones, snf null is required for normal cell proliferation. The gene snf codes for the Drosophila homolog of the mammalian U1A snRNP, a general component of the splicing machinery. The vital function of vir, fl(2)d and snf, however, cannot depend on Sxl as this gene is completely dispensable in males, and constitutive expression of Sxl does not rescue animals mutant for those alleles that abolish this vital function (Hilfiker, 1995).

If, as the results suggest, vir is involved in splicing of the pre-mRNAs of Sxl and tra, and perhaps msl-2, it could also participate in the splicing of pre-mRNAs of vital genes. Mutations in snf, in fl(2)d, and in vir, can cause lethality to both sexes. However, since clones of vir and fl(2)d, in contrast to those of snf null, are viable in males and females, neither vir nor fl(2)d can encode a general and absolutely necessary factor for cell growth and cell viability. It is not known whether vir is directly involved in the splicing of tra and msl-2 transcripts, or whether it somehow activates the Sxl protein (Hilfiker, 1995).

The continuous production of Sxl proteins in XX animals is maintained by autoregulation and depends on virilizer. In the soma, clones of XX cells lacking either Sxl or vir are sexually transformed and form male structures; in the germline, XX cells mutant for Sxl extensively proliferate, but are unable to differentiate. The role of vir in the germline has been studied by generating germline chimeras. XX germ cells mutant for vir, in contrast to cells mutant for Sxl, perform oogenesis. The early production of Sxl in undifferentiated germ cells is independent of vir; later, in oogenesis, expression of Sxl becomes dependent on vir. It is concluded that the early Sxl proteins are sufficient for the production of eggs whereas the later Sxl proteins are dispensable for this process. However, vir must be active in the female germline to allow normal embryonic development because maternal products of vir are required for the early post-transcriptional regulation of Sxl in XX embryos and for a vital process in embryos of both sexes (Schutt, 1998).

daughterless (da), hermaphrodite (her) and snf are known to have a maternal effect on the viability of daughters. XX embryos from da or her mutant mothers fail to initiate Sxl. vir is a further component that is maternally required for the regulation of Sxl in XX embryos. Unlike da and her, but similar to snf, it acts at the level of post-transcriptional control. The establishment promoter of Sxl is activated both in vir6/plus and vir2f/vir2f embryos derived from vir mutant mothers, but Sxl expression is not maintained. The lethality of vir6/plus embryos indicates that maternal product of vir is responsible for the autoregulation of Sxl during early development. In addition, maternal vir function is also required for viability of offspring of both sexes as shown by strong alleles disrupting the vital function. For such alleles, neither the autoregulation of Sxl nor the vital process can be rescued by a paternal vir plus allele. The reason may be that the zygotic gene is either not yet active or not yet sufficiently active (Schutt, 1998).

Most genes of the somatic sex determination cascade are dispensable within germ cells. Sxl, although necessary for oogenesis, does not have a master regulatory function for sex determination in the germline, as indicated by the following observations. (1) Expression of Sxl in germ cells is first detected in 16- to 20-hour-old embryos. A male-specifically expressed gene, however, is already observed in germ cells of 10-hour-old embryos, suggesting that some aspect of sexual development must have been determined prior to expression of Sxl. (2) Sxlf4/Sxlf1 and Sxlf5/Sxlf1 larvae have female gonads by the criteria of size and morphology; only as adults do the germ cells form abnormal multicellular cysts. Sxl does not seem to control the sex-specific differentiation of germ cells in larvae, but is required later for oogenesis. (3) XY germ cells containing SxlM1 or SxlM4 are not feminized, but instead form fertile sperm (Schutt, 1998 and references therein).

The expression of Sxl in the germline depends on inductive signals from the gonadal soma and on an autonomous signal given by the germline X:A ratio, which is measured by elements different from those used to determine the somatic X:A ratio. Thus, initiation of Sxl is different in soma and germline. In a second phase, expression of Sxl in differentiating cysts becomes dependent on vir. This parallels the regulation of Sxl in somatic cells. In contrast to the soma, however, the consequences of misexpression of Sxl are different: germ cells are not sexually transformed (Schutt, 1998).

The different germline phenotypes of Sxl and vir mutations may be due to trivial reasons. A perdurance effect of Vir protein made in heterozygous stem cells prior to mitotic recombination could maintain the production of Sxl protein in vir clones. However, the tested females continue to lay eggs even two weeks after the induction of clones and after several rounds of cell divisions. By this time, Vir protein should have been diluted or eliminated and should no longer be able to regulate Sxl. A second possibility would be that the first 24 hours of development in heterozygous condition before the induction of clones determines the female fate of a germ cell irreversibly. In all previous experiments with ovaries mutant for Sxlf4 and with transplanted pole cells mutant for Sxlf1, the germ cells were homozygous mutant from the beginning. But even when Sxl clones were induced under the same conditions as the vir clones, the females remained sterile as in earlier experiments showing that Sxl is necessary after the induction of the clones. Therefore, the phenotypic differences of Sxl and vir mutants are caused by different requirements for the two genes (Schutt, 1998).

Based on these results, the following model is proposed. The initiation mechanism of Sxl in the germline is unknown. However, for female development, early Sxl protein is necessary in gonial cells at the tip of the germarium in the adult ovary. Later in oogenesis, the germ cells may become independent of the primary signals by the soma and the X:A ratio and maintain Sxl production by autoregulation. This later post-transcriptional regulation depends on vir. Late Sxl protein, however, is no longer necessary for the female differentiation of germ cells, once these cells have embarked on the oogenic pathway. It is not yet clear what other functions Sxl has in later stages of oogenesis. Sxl protein is redistributed during oogenesis. In stem cells and early cystoblasts in the germarium, Sxl is predominantly cytoplasmic. During the mitotic divisions, the level of cytoplasmic Sxl drops drastically. In the cluster of 16 cells, the protein becomes concentrated in the nuclei of the cystocytes. This transition of Sxl protein may reflect the early and late functions of Sxl. It is concluded that Sxl is required for a short period during the transition from stem cells to cystoblasts. Lack of Sxl during this phase results in tumorous cysts (Schutt, 1998).

snf and fl(2)d are two other genes involved in the autoregulation of Sxl in the soma. These two genes are also required for oogenesis. Mutant females produce no eggs, but instead form multicellular cysts. It can only be speculated why mutations in snf, fl(2)d and vir have different consequences in the germline. They may still all affect the autoregulation of Sxl, as they do in somatic cells, but do so at different times. It is possible that the production of early Sxl protein depends on a post-transcriptional regulation that requires snf and fl(2)d, whereas vir may only act later in oogenesis (Schutt, 1998).

The regulation of Sxl in the germline is certainly more complex than in the soma, and the germline function of this gene is also less clear. A comparative analysis of the regulation and function of Sxl in soma and germline is therefore worthwhile and may contribute to an understanding of how a gene operates in different cell types (Schutt, 1998).

Initial analysis of >30 vir alleles shows that most of them represent lesions undetectable by whole genome Southern analysis. Two exceptions are the alleles vir22 and vir23, which are both associated with small deletions of genomic DNA. Another 5 alleles have also been detected at this level of analysis, but show rather large deletions of >30 kb each. To identify potentially relevant domains, the lesions were determined in the two female-specific alleles, vir1ts and vir2f, and in three strong alleles, vir4, vir22, and vir23, which are recessive lethals for males and females. Genomic DNA was isolated from adult flies homozygous for the sex-specific and viable vir1ts or vir2f alleles and from third instar larvae, hemizygous for the lethal alleles vir4, vir22, or vir23 over Df(2R)vir130 (Hilfiker, 1995). 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, vir1ts and vir2f, are point mutations that lead to single amino acid exchanges. In vir1ts, glutamic acid at position 1423 is mutated into lysine (1423 E-K), and in vir2f, 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 (Niessen, 2001).

The products of Sxl, tra, and tra-2 are known regulators of alternative splicing decisions in Drosophila but they are not essential for processes other than sex determination (and dosage compensation, in the case of Sxl) because males that are null for these genes are viable and appear phenotypically normal. However, additional genes [fl(2)d, virilizer, and l(2)49Db] are required for correct control of alternative splicing decisions by Sxl are also essential for viability in both sexes; hence, their products may also have roles in other alternative splicing events. To determine whether these include the control of Ubx alternative splicing, it was asked whether the Ubx isoform ratios are altered in heterozygotes for mutations in these genes. In contrast to the stability described in the preceding section, the Ubx splicing pattern is altered significantly when the expression or function of virilizer or fl(2)d is reduced. The strongest effect is observed with virilizer, using a loss-of-function allele (vir3) that is recessive lethal in both sexes. In heterozygous larvae the proportion of Ubx class I mRNAs declines while that of classes II and IV increases. The proportion of class I that contains the B element is not altered. The increase in classes II and IV indicates that inclusion of both mI and mII is reduced but that the effect on mI exceeds that on mII. Inclusion of mI is also reduced in adults, although the effect was weaker than in larvae (Burnette, 1999).


Burnette, J. M., Hatton, A. R. and Lopez, A. J. (1999). Trans-acting factors required for inclusion of regulated exons in the Ultrabithorax mRNAs of Drosophila melanogaster. Genetics 151: 1517-1529. 10101174

Haussmann, I. U., Bodi, Z., Sanchez-Moran, E., Mongan, N. P., Archer, N., Fray, R. G. and Soller, M. (2016). m6A potentiates Sxl alternative pre-mRNA splicing for robust Drosophila sex determination. Nature 540(7632): 301-304. PubMed ID: 27919081

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

Hilfiker, A., et al. (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. 8575302

Niessen, M., Schneiter, R. and Nothiger, R. (2001). Molecular identification of virilizer, a gene required for the expression of the sex-determining gene Sex-lethal in Drosophila melanogaster. Genetics 157: 679-688. 11156988

Schutt, C., Hilfiker, A. and Nothiger, R. (1998). virilizer regulates Sex-lethal in the germline of Drosophila melanogaster. Development 125: 1501-1507. 9502731

virilizer: Biological Overview | Regulation | Developmental Biology | Effects of Mutation

date revised: 30 April 2017

Home page: The Interactive Fly ©1997 Thomas B. Brody, Ph.D.

The Interactive Fly resides on the
Society for Developmental Biology's Web server.