Drosophila Gain-of-Function Mutant RTK Torso Triggers Ectopic Dpp and STAT Signaling

Corresponding author: Willis X. Li, University of Rochester Medical Center, 601 Elmwood Ave., KMRB 2-9641, Rochester, NY 14642., willis_li{at}urmc.rochester.edu (E-mail)

Communicating editor: T. SCHÜPBACH

	ABSTRACT

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Overactivation of receptor tyrosine kinases (RTKs) has been linked to tumorigenesis. To understand how a hyperactivated RTK functions differently from wild-type RTK, we conducted a genome-wide systematic survey for genes that are required for signaling by a gain-of-function mutant Drosophila RTK Torso (Tor). We screened chromosomal deficiencies for suppression of a gain-of-function mutation tor (tor^GOF), which led to the identification of 26 genomic regions that, when in half dosage, suppressed the defects caused by tor^GOF. Testing of candidate genes in these regions revealed many genes known to be involved in Tor signaling (such as those encoding the Ras-MAPK cassette, adaptor and structural molecules of RTK signaling, and downstream target genes of Tor), confirming the specificity of this genetic screen. Importantly, this screen also identified components of the TGFß (Dpp) and JAK/STAT pathways as being required for Tor^GOF signaling. Specifically, we found that reducing the dosage of thickveins (tkv), Mothers against dpp (Mad), or STAT92E (aka marelle), respectively, suppressed tor^GOF phenotypes. Furthermore, we demonstrate that in tor^GOF embryos, dpp is ectopically expressed and thus may contribute to the patterning defects. These results demonstrate an essential requirement of noncanonical signaling pathways for a persistently activated RTK to cause pathological defects in an organism.

RECEPTOR tyrosine kinases (RTKs) play many essential roles in normal development as well as in pathogenesis. The RTK family includes a large number of single-transmembrane cell surface receptors, including receptors for peptide ligands such as epidermal growth factor (EGF), platelet-derived growth factor (PDGF), and fibroblast growth factor. RTK overactivation, which can result from ligand overabundance or ligand-independent constitutive activating mutations, has been linked to many cancers and other human diseases (reviewed by ROBERTSON et al. 2000 ). Despite extensive studies, we are still far from understanding how regulated activation of RTKs induces precise patterns of gene expression and how constitutive activation of RTKs causes aberrant gene expression patterns and pathological conditions. Previous models suggest that a constitutively activated RTK hyperactivates a canonical downstream signal transduction pathway, the Ras/Raf/MEK/mitogen-activated protein kinase (MAPK) signaling cassette, and the resulting change in signaling duration and/or intensity is sufficient to cause a qualitative perturbation of gene expression patterns (MARSHALL 1995 ; GREENWOOD and STRUHL 1997 ; SEWING et al. 1997 ; WOODS et al. 1997 ; GHIGLIONE et al. 1999 ). However, an alternative model is that the effects of high RTK signaling may require the participation of additional signaling pathways that may not be essential for the functions of the RTK under physiological conditions. According to this scenario, signaling from a highly activated RTK might cross-activate components of a pathway that is not normally involved in or important for the RTK response, or it might induce the ectopic expression of a ligand that triggers such a noncanonical signaling cascade.

We have studied the Torso (Tor) RTK pathway in the early Drosophila embryo to address the issue of how overactivation of an RTK may result in aberrant gene expression. The early Drosophila embryo offers a unique opportunity for studying signaling pathway interactions because many pathways operate simultaneously during development and these pathways can potentially interact with each other. Tor is most homologous to the PDGF receptor and is responsible for determining embryonic terminal cell fates (reviewed by DUFFY and PERRIMON 1994 ). The Tor protein is uniformly distributed along the cell membrane at the multinucleated, syncytial blastoderm stage of the embryo. However, it is activated only at the two polar regions of the early embryo by spatially restricted ligands (CASANOVA and STRUHL 1993 ). As a result, its downstream target genes, tailless (tll) and huckebein (hkb), are expressed only at the anterior and posterior terminal regions of the embryo (PIGNONI et al. 1990 , PIGNONI et al. 1992 ; WEIGEL et al. 1990 ; BRONNER and JACKLE 1991 ). tll is necessary and sufficient for the development of the posterior structure Filzkörper (see Fig 1A; STEINGRIMSSON et al. 1991 ). Embryos from tor mutant mothers show no posterior tll expression and are missing the Filzkörper and other posterior structures. In contrast, embryos from tor^GOF mothers exhibit expanded tll expression domains and enlarged, or occasionally ectopic, Filzkörper material (LI et al. 2002 ). Three alleles of tor^GOF have been isolated and each contains a point mutation in the extracellular, ligand-binding domain, presumably favoring ligand-independent dimerization of the receptor (SPRENGER and NUSSLEIN-VOLHARD 1992 ). Furthermore, studies of the transcriptional regulation of tll have led to the proposal that Tor controls tll expression by derepression, relieving the repressors bound to the tll promoter via phosphorylation by MAPK. This allows tll activation by transcriptional activators that have yet to be unidentified and are presumably ubiquitously present in the embryo (LIAW et al. 1995 ).

View larger version (79K): In this window In a new window Download PPT slide   Figure 1. Classification of Tor gain-of-function mutant phenotypes associated with torY9. Larval cuticle preparations of embryos laid by females with indicated genotypes are shown. Solid arrows point to the Filzkörper. Open arrows indicate the head skeletons and mouth hook. Triangles indicate the positions of visible ventral denticle bands. (A) A wild-type larva has eight abdominal denticle bands (A1–A8). Note the position and morphology of the Filzkörper (arrow) and head skeletons. (B) Class I cuticles are typical of embryos from torY9/torY9 mothers; usually only a hollow shell of the vitelline membrane can be found in the cuticle preparations. Ventral denticle bands and head skeletons are usually completely absent. The Filzkörper, if present, appears diffused and mislocalized. (C) Class II cuticles are representative of embryos from torY9/+ mothers raised at 18°. These embryos have remnants of defective head skeletons and one or two partial ventral denticle bands. The Filzkörper usually appears normal in structure. (D) Class III embryos, found in the majority of torY9/+ flies raised at 21°, have normal head skeleton and Filzkörper. About four ventral denticle bands are usually present. (E) Class IV cuticles are nearly normal, with only minor defects in the central body region. Some of these embryos will hatch.

As with RTKs in mammals, Tor signaling in Drosophila is relayed by the Ras-MAPK pathway. Embryos lacking Draf gene activity have phenotypes identical to those of tor null embryos, resulting in the complete absence of the posterior tll expression and posterior structures (MELNICK et al. 1993 ; LI et al. 1997 ). Drosophila Ras1 and Draf are structurally and functionally homologous to mammalian H- or K-Ras and Raf-1, respectively, as human Ras and Raf-1 can at least partially substitute these Drosophila proteins during development (AMBROSIO et al. 1989 ; LU et al. 1993 ; CASANOVA et al. 1994 ; W. X. LI and J. LI, unpublished data), suggesting a conservation in this portion of the RTK signaling pathway. In addition, mutations in the Ras-MAPK cassette molecules completely abolish the effects of tor^GOF on embryos (AMBROSIO et al. 1989 ). Therefore, genetically the Ras-MAPK cassette is essential for both Tor and Tor^GOF molecules and appears to be the major, if not the sole, downstream signaling pathway that conveys Tor signals in wild-type embryos. Whether additional pathways are required for Tor^GOF remains unclear.

To systematically examine whether additional signaling pathways are required for Tor^GOF, we conducted a genetic screen using a set of contiguous chromosomal deficiencies to determine what genomic regions, and subsequently what genes in these regions, may participate in signaling by a Tor^GOF mutant, Tor^Y9. Genetic screens for modifiers of Tor or other RTK downstream components (such as Draf of Ras1) have previously been conducted and have contributed to the identification of a number of essential components of RTK or Ras/Raf signaling (SIMON et al. 1991 ; DOYLE and BISHOP 1993 ; DICKSON et al. 1996 ; KARIM et al. 1996 ; LI et al. 2000 ). We focused on the genomic regions that are not known to encompass genes required for RTK signaling because these regions might contain genes encoding novel components of RTK signaling or, more interestingly to this study, components of other signaling pathways that are particularly important for the phenotypic effects of the Tor^Y9 mutation or of gain-of-function forms of RTKs in general. Here we show that in addition to the Ras-MAPK cassette and other known components of the classical RTK pathway, the STAT (STAT92E) and Dpp [a transforming growth factor ß (TGFß) homolog] pathways are required by Tor^GOF to exert its effects on embryos. We have further investigated the involvement of STAT92E in Tor^GOF signaling and found that STAT92E is essential for the functions of Tor^GOF and is likely directly activated by Tor^GOF (LI et al. 2002 ). However, surprisingly, we found that STAT92E is dispensable for wild-type Tor to induce its target gene tll (LI et al. 2002 ). Therefore, it appears that STAT92E is particularly important for the activity of the gain-of-function mutant RTK. Dpp functions as a morphogen in the early embryo to direct the establishment of the embryonic dorsoventral (D/V) axis. Consistent with such a role, it is expressed mainly in the dorsal central region of the embryo, where Tor signaling is absent (RAY et al. 1991 ). We found that mutations in components of the Dpp pathway suppressed the phenotypes associated with tor^Y9, suggesting that Dpp signaling contributes to the defects caused by Tor^Y9. In addition, we found that dpp expression is ectopically induced in tor^Y9 embryos. These results suggest that, in contrast to wild-type RTK that mainly requires the Ras-MAPK cassette, gain-of-function RTK requires multiple additional signaling pathways to cause dramatically different biological consequences.

	MATERIALS AND METHODS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Drosophila strains and genetics:
Fly stocks: Fly stocks include tor^Y9/CyO (viable and fertile due to the presence of suppressors in this particular stock; KLINGLER et al. 1988 ) and dpp^H46/CyO23. The CyO23 chromosome carries an extra copy of dpp⁺ as a transgene in addition to the endogenous dpp⁺, thus duplicating the dpp locus (WHARTON et al. 1993 ). All other stocks, including the Bloomington deficiency kit stocks, are from the Bloomington Drosophila Stock Center (FLYBASE 1998 ) or as described in LI et al. 2000 , LI et al. 2002 .

Genetic screen: To screen for suppressors of tor^Y9 by autosomal Df stocks, virgin females of tor^Y9/CyO were crossed to males of each individual Df stock (Df/Balancer) in vials. From the F₁ progeny of each of the above crosses, one to five virgin females of the tor^Y9/+; Df/+ genotype (identified by the absence of Balancer and CyO) were crossed to wild-type males and tested for fertility at 18°. Normally tor^Y9/+ females are completely sterile at 18° because they lay eggs of mostly class II phenotype that fail to hatch (see RESULTS and Fig 1C). The presence of hatched larvae in eggs laid by tor^Y9/+; Df/+ females would indicate that the sterility of tor^Y9/+ females is dominantly suppressed by a particular Df. Suppression is further confirmed by retesting and examination of cuticle preparations. To screen for X-chromosomal Df stocks, Df/Balancer females were used to cross with tor^Y9/CyO males. Certain CyO chromosomes exhibited suppression of tor^Y9 and were not used as a control. Embryos collected from tor^Y9/CyO flies crossed to wild type (Oregon R) were used as a control.

Effects of the zygotic dosage of dpp on tor^GOF phenotypes: To test the effects of changing the dosage of dpp on tor^Y9 phenotypes, we crossed tor^Y9/+ females to dpp^H46/CyO23 (carrying Dp[dpp⁺]) males at 18° and examined the cuticles of the embryos. The resulting embryos from this cross should have either one or three zygotic doses of dpp⁺, since they should have a zygotic genotype of either dpp^H46/+ or CyO23/+ (two copies of the endogenous dpp⁺ and the duplication carried by the CyO23 chromosome) regarding the dpp locus. The embryos from such a cross are clearly divided into two phenotypic classes (class I and class III; see text) in an 1:1 ratio (see RESULTS and Fig 4A and Fig B). This is in contrast to the control embryos (tor^Y9/+ females crossed to +/+ males) that showed mostly class II phenotypes (not shown; see Fig 1C). To confirm this interpretation, tor^Y9/+ females were crossed to +/CyO23 males, which resulted in embryos zygotically containing either two or three copies of dpp⁺. Embryos from such a cross exhibited class I and II phenotypes in about equal ratios, suggesting that the class I phenotype is due to the presence of three copies of dpp⁺.

View larger version (41K): In this window In a new window Download PPT slide   Figure 2. Genomic regions and candidate genes required for the functions of TorY9. Chromosomal arms are represented by solid bars with cytological divisions indicated by numbers. Chromosomal deficiencies screened are represented by boxes that span the genomic region deleted in the deficiency. Open boxes represent the screened deficiencies that did not suppress torY9. Solid boxes indicate those that suppressed the maternal effect of torY9 heterozygotes. Enhancers are not indicated. Candidate genes in each genomic region are indicated.

View larger version (160K): In this window In a new window Download PPT slide   Figure 3. Representative cuticle phenotypes of torY9 and a suppressor mutant combination. Representative cuticles from eggs laid by the indicated trans-heterozygotes at 18° are shown. These cuticles fall into class III or IV categories (see text and Table 3 for more detail). (A) Draf11-29/+; torY9/+. (B) torY9/+; Ras1C40B/+. (C) cswKN27/+; torY9/+. (D) torY9/leoP1188. (E) torY9/+; tllL10/+. (F) torY9/+; hkbA321R1/+. (G) torY9/tkv7. (H) torY9/Madk00237. (I) torY9/+; mrl06346/+. (J) torY9/torXR1. Interestingly, torXR1, a null allele, did not completely suppress torY9. Most of the torY9/ torXR1 larvae exhibited collapsed head skeletons similar to those of torXR1/torXR1 (not shown).

View larger version (155K): In this window In a new window Download PPT slide   Figure 4. Involvement of dpp in the effects of torGOF mutations on embryos. Lowering and increasing the copy number of dpp+ suppressed (A and C) and enhanced (B and D) torY9 and tor4021 phenotypes, respectively. (E) Wild-type patterns of dpp expression. (F) dpp expression in torY9 embryos. Note that dpp mRNA is nearly uniformly expressed in torY9 embryos. (G) Elimination of zygotic tll [by Df(3R)tll-e and Df(3R)tll-g trans-heterozygotes] did not completely suppress torY9. Note the defective head skeletons (arrow) and a reduced number of the ventral denticle bands. (H) Additional reduction in the maternal Mad gene product (using Madk00237) resulted in a nearly complete suppression of torY9. Note the appearance of head skeletons (arrow) and a complete set of the ventral denticle bands. This phenotype is nearly identical to that of tll zygotic null embryos.

A similar strategy was used to test the effects of changing the zygotic dosage of dpp on the tor⁴⁰²¹ phenotype. tor⁴⁰²¹/+ females were crossed to dpp^H46/CyO23 males at 22° (intermediate temperature) and the resulting embryos were analyzed for cuticle phenotypes.

Examination of embryos:
Cuticle preparations were performed according to a standard protocol with minor modifications. Embryos were dechorionated with 50% Clorox, washed extensively with 0.1% Triton, mounted in Hoyer's, and photographed in dark-field optics. In situ hybridization for dpp mRNA was performed according to a standard protocol using digoxigenin-incorporated antisense RNA probes made from a PCR fragment of a 500-bp dpp coding region according to the supplier's protocol. Stained embryos were photographed using Normaski optics.

	RESULTS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Dosage- and temperature-dependent embryonic phenotypes associated with tor^Y9, a tor^GOF allele, and the design of a genetic screen:
tor^Y9 is a cold-sensitive dominant gain-of-function allele of tor that is associated with a point mutation W317R in the extracellular domain of Tor, presumably causing ligand-independent dimerization of the Tor^Y9 molecules (SPRENGER and NUSSLEIN-VOLHARD 1992 ). To make use of tor^Y9 for a genetic screen, we characterized the temperature and dosage dependency of the tor^Y9 embryonic phenotypes in detail (Fig 1 and Table 1).

First, we found that tor^Y9 is dosage sensitive. Embryos from homozygous mothers (tor^Y9/tor^Y9) exhibit the strongest gain-of-function phenotypes, while those from the heterozygous (tor^Y9/+) or hemizygous (tor^Y9/tor^-) females show progressively weaker phenotypes (Table 1). In cuticle preparations of embryos from tor^Y9/tor^Y9 mothers, only a hollow shell of vitelline membrane can usually be found. No remnants of ventral denticle bands or the head skeleton are visible. The pair of Filzkörper, if present, is diffused as two balls of light-colored material usually located in the center of the embryo (Fig 1B). We designate such a phenotype as class I. In contrast, the majority of embryos from tor^Y9/tor^- show less severe cuticle phenotypes and fall into the class IV category (see below and Table 1).

Second, eggs from tor^Y9/+ females show a dramatic temperature-dependent phenotypic series. At 18°, tor^Y9/+ females lay eggs that exhibit the most severe defects. In cuticle preparations, the majority of these embryos show remnants of the head skeleton, visible as brownish-colored material at the anterior of the egg. Essentially all the central structures are undifferentiated except for an occasional strip of ventral denticle band with discernible bristle patterns (Fig 1C). The Filzkörper is clearly identifiable and often is a little enlarged. This phenotype is less severe than the class I phenotype and we designate these embryos as class II. When the temperature is raised to 21°, most eggs show improved differentiation such that they exhibit intact head skeleton, Filzkörper, and more structures in the central region. Usually about four ventral denticle bands are identifiable in these embryos, which we designate as class III (Fig 1D). When tor^Y9/+ females are kept at 25°, there is a significant improvement in the development of the embryos such that many of them hatch and become nearly morphologically normal crawling larvae. The majority of these larvae show only minor defects in the central regions. The phenotypes range from missing one or two ventral denticle bands to indistinguishable from wild-type larvae. These are class IV embryos (Fig 1E).

The temperature and dosage-dependent phenotypes of tor^Y9 indicate that the strength of this allele is modifiable, suggesting that it might be sensitive to the dosage of other molecules that are required for Tor^Y9 to exert its effects on embryos. If true, the tor^Y9 allele would be suitable for conducting a dosage-sensitive modifier genetic screen. We tested the possible suppression of tor^Y9 phenotypes by reducing the dosage of Draf and Ras1 genes by half. Indeed, Draf or Ras1 heterozygotes strongly suppressed the tor^Y9 phenotype. At 18°, halving the maternal dosage of Ras1 or Draf in tor^Y9/+ females resulted in embryos that exhibited mostly class III phenotype while a significant number of the embryos exhibited class IV phenotype and hatched into larvae (Fig 3A; data not shown for Ras1). Therefore, tor^Y9 appears suitable for conducting a modifier screen for dosage-sensitive genes affecting its function.

Identification of genomic regions and genes required for the effects of tor^Y9 on embryonic development:
Using a set of nearly contiguous deficiency (Df) stocks, we screened for genes that, when in a heterozygous state, could dominantly suppress tor^Y9 (see MATERIALS AND METHODS for more detail). Among a total of 206 Df stocks representing 70% of the Drosophila genome, we identified 31 deficiencies, which define 26 genomic regions, that dominantly suppressed the maternal-effect sterility of tor^Y9/+ females to various degrees (Fig 2 and Table 2). To determine which genes in each genomic region are responsible for the suppression, we selected candidate genes for each deficiency region and used available mutant alleles to test their ability to suppress the sterility and embryonic defects associated with tor^Y9 at 18° in trans-heterozygotes. Such testing allowed us to determine genes responsible for the suppression of tor^Y9 in the identified genomic region. These genes should be required for the biological functions of Tor^Y9. Listed in Table 2 are the candidate genes we have confirmed by testing the indicated mutant alleles. Cuticle phenotypes and classes that confirmed the suppression of tor^Y9 for representative genes are shown in Fig 3 and Table 3. Most of the genomic regions identified in our screen for suppressors of tor^Y9 encompass loci that are known to function positively in Tor signaling (17/26; Fig 2 and Table 2), validating the specificity of this screen.

Among the genes that behaved as strong suppressors of tor^Y9 are those encoding the Ras-MAK signaling cassette components, such as Ras1, Draf, Downstream of Raf1 (Dsor1; encoding a MEK; TSUDA et al. 1993 ), and rolled (rl; encoding a MAPK) (BRUNNER et al. 1994 ). These genes are located at cytogenetic positions 85D21, 3A1, 8D3, and h41, respectively, and are disrupted by the corresponding Dfs listed in Table 2 (FLYBASE 1998 ). In addition, we also identified genes encoding components that modulate the Ras-MAPK cassette in RTK signaling as suppressors of tor^Y9. These genes include corkscrew (csw; encoding a SHP-2 protein tyrosine phosphatase; PERKINS et al. 1992 ), downstream of receptor kinase (drk; encoding an SH2/SH3 adaptor molecule) (SIMON et al. 1993 ), son of sevenless (sos; encoding the Ras guanyl-nucleotide exchange factor; SIMON et al. 1991 ), leonardo (leo; encoding 14-3-3; KOCKEL et al. 1997 ; LI et al. 1997 ), 14-3-3 (CHANG and RUBIN 1997 ), hsp83 (encoding an HSP90 homolog; VAN DER STRATEN et al. 1997 ), and tor itself. Finally, many of the known downstream target genes of Tor or transcription factors involved in Tor signaling were identified in our genetic screen. These genes include the immediate Tor target genes tailless (tll) and huckebein (hkb), and additional transcription factors implicated downstream of Tor function, such as tailup (tup; STRECKER et al. 1991 ), lilliputian (lilli; TANG et al. 2001 ; WITTWER et al. 2001 ), and brachyenteron (byn; SINGER et al. 1996 ).

The most important results of this screen relate to the genes not previously known to be involved in RTK signaling. Testing of the mutant alleles led to the surprising discovery that two components of the Dpp pathway and one component of the JAK/STAT pathway are required for Tor^Y9 functions (see below). Since these genes were not identified in the numerous previous genetic screens, it is thus possible that they might be particularly important for gain-of-function mutant RTK (such as Tor^Y9).

We were unable to identify a gene(s) responsible for some of the Df regions that suppressed tor^Y9 (question marks in Fig 2 and Table 2). These regions may contain novel RTK signaling components or novel molecules especially required for Tor^Y9 signaling. Identification of such genes is ongoing.

Involvement of Dpp and STAT92E in signaling by Tor^GOF:
Testing of candidate genes allowed us to identify three signaling molecules not previously known to be required for the function of Tor^GOF. Mothers against dpp (Mad; a Smad homolog; NEWFELD et al. 1996 ), located at cytogenetic position 23D3, is among the genes deleted by both Df(2L)23C;23E3-6 and Df(2L)S2590, which suppressed tor^Y9 (Table 2). We tested Mad^k00237, a P-element allele, and found that it behaved similarly to the deficiencies and suppressed tor^Y9 in trans-heterozygotes (Fig 3H; Table 3). thickveins (tkv; encoding a TGFß type I receptor homolog; BRUMMEL et al. 1994 ; NELLEN et al. 1994 ; PENTON et al. 1994 ), located at 25D1-2, is disrupted by the two overlapping deficiencies, Df(2L) sc19-4 and Df(2L)cl-h3, both of which modestly suppressed tor^Y9 (Table 2). We tested tkv⁷, a strong allele of tkv, and found that embryos derived from tor^Y9/tkv females exhibit significantly less severe phenotypes than do those derived from tor^Y9/+ females, suggesting that tkv⁷ dominantly suppressed tor^Y9 (Fig 3G; Table 3). The identification of two genes in the Dpp pathway as dominant suppressors of tor^Y9 indicates that the Dpp pathway might be required for tor^Y9 signaling. In addition to the two components of the Dpp pathway, we found that STAT92E [also known as marelle (mrl), it encodes the Drosophila STAT and is located at 92E1–2; HOU et al. 1996 ; YAN et al. 1996 ] is responsible for the suppression of tor^Y9 by Df(3R)H-B79. (Fig 3I; Table 2 and Table 3). We have further investigated the involvement of STAT92E in Tor^GOF signaling and demonstrated that STAT92E is essential for the functions of Tor^GOF and is likely activated directly by Tor^GOF (LI et al. 2002 ).

Tor^GOF induces ectopic dpp expression:
To further understand the involvement of the Dpp pathway in Tor^Y9 signaling, we examined whether the dpp gene itself is required by Tor^Y9 to cause embryonic defects. Since the dpp locus is haplo-insufficient, one cannot test the maternal effects of dpp by using a Df or a null allele in a heterozygous female. However, the requirement for dpp in the early embryos is purely zygotic. Therefore, the effects of varying the copy number of the paternal dpp gene on tor^Y9 embryonic phenotype could be tested. At 18°, embryos that are zygotically dpp^H46/+ (containing one copy of dpp⁺) and from tor^Y9/+ females exhibited class III phenotype (suppressed tor^Y9; Fig 4A), while those containing three copies of dpp⁺ showed mostly class I phenotype (enhanced tor^Y9; Fig 4B; see MATERIALS AND METHODS). This is in contrast to the embryos with two copies of dpp⁺ (wild type regarding the dpp locus), which showed mostly class II phenotypes at 18° (see Fig 1C). Therefore, increasing and decreasing the copy number of dpp⁺ enhances and suppresses tor^Y9 phenotype, respectively.

To investigate whether the dpp pathway is generally involved in Tor^GOF signaling or is specific for tor^Y9, we tested genetic interactions between dpp and another tor^GOF allele, tor⁴⁰²¹ (SPRENGER and NUSSLEIN-VOLHARD 1992 ), which is the strongest extant tor^GOF allele and is associated with a Tyr-to-Cys amino acid change in the extracellular domain of Tor, presumably causing ligand-independent dimerization (SPRENGER and NUSSLEIN-VOLHARD 1992 ). Similar to tor^Y9, tor⁴⁰²¹ is cold sensitive and its phenotypes are most severe at low temperatures (e.g., 18°; SPRENGER and NUSSLEIN-VOLHARD 1992 ). We examined embryos from tor⁴⁰²¹ females with different paternal dpp⁺ copy numbers (see MATERIALS AND METHODS) and found that, at an intermediate temperature (22°), increasing and decreasing the copy number of dpp⁺ enhanced and suppressed tor⁴⁰²¹ phenotypes, respectively (Fig 4C and Fig D). At 22°, embryos from tor⁴⁰²¹/+ females differentiated only rudimentary cuticular structures and exhibited mostly class I phenotype (not shown). Reducing dpp⁺ dosage by half in tor⁴⁰²¹/+ embryos resulted in mostly class II type cuticles (Fig 4C). Conversely, a 50% increase in dpp⁺ dosage enhanced the phenotypes of tor⁴⁰²¹/+ embryos so that most of them did not differentiate any discernible cuticles (Fig 4D). Therefore, changes of dpp⁺ dosage modify the phenotypes of at least two tor^GOF alleles, suggesting that the Dpp pathway is generally involved in Tor^GOF signaling.

Since dpp affects tor^GOF phenotypes in a dosage-dependent manner, we reasoned that dpp might be ectopically induced in tor^GOF embryos, triggering aberrant activation of the Dpp pathway, which in turn contributes to the patterning defects associated with tor^GOF. This reasoning led us to examine the dpp mRNA expression patterns in tor^Y9 embryos. Consistent with the above speculation, embryos derived from tor^Y9/+ females indeed showed levels higher than those of wild type as well as ectopic dpp mRNA expression (Fig 4E and Fig F).

Both ectopic tll and dpp may account for the full biological effects of Tor^Y9:
It has been proposed that the induction of ectopic expression of the Tor target gene tll in the central region of the embryo by Tor^GOF accounts for most, if not all, of the phenotypic defects associated with tor^GOF (KLINGLER et al. 1988 ; STRECKER et al. 1989 ). This is based on the suppression of the segmentation defects associated with tor^GOF alleles by zygotic tll mutations. It has been shown that tor^GOF; tll/tll embryos exhibited phenotypes that were similar to tll/tll embryos, including restoration of the ventral denticle bands in the central region that would have been disrupted by tor^GOF (KLINGLER et al. 1988 ). We retested the suppression of tor^GOF by tll zygotic null embryos and found that tor^Y9; tll/tll embryos (n = 34) exhibited severely defective head skeletons and a reduced number of denticle bands (Fig 4G). Such phenotypes were more similar to tor^Y9 embryos than to tll/tll embryos, because tll/tll embryos retain head skeletons and anterior ventral denticle bands (not shown; also see KLINGLER et al. 1988 ). This suggests that tll zygotic null mutations cannot completely suppress tor^Y9 phenotype. To assess the contribution of the Dpp pathway to tor^GOF phenotypes, we tested the effects of reducing the maternal Mad gene dosage on tor^GOF; tll/tll embryos. We found that the phenotypes of tor^Y9/Mad; tll/tll embryos (n = 27) were nearly indistinguishable from those of tll/tll embryos (Fig 4H), indicating that the tor^Y9 phenotype was more completely suppressed by eliminating zygotic tll expression and simultaneously reducing Dpp signaling. Therefore, the ectopic tll expression induced by Tor^GOF cannot account for all the embryonic defects caused by Tor^GOF. The finding that the Dpp pathway is required for the tor^GOF phenotypes suggests that both dpp and tll are essential for the defects associated with tor^GOF.

Involvement of Dpp and STAT92E in gain-of-function mutant EGF receptor signaling:
To investigate whether the Dpp and STAT92E pathways are also involved in gain-of-function mutant EGF receptor (EGFR) signaling, we tested mutations in components of these pathways for their ability to suppress the phenotypes of a gain-of-function mutation in Egfr, Ellipse (Elp or Egfr^Elp). Egfr^Elp encodes a hyperactive EGFR molecule, causing a rough eye phenotype and the appearance of extra wing veins (BAKER and RUBIN 1989 ; BRUNNER et al. 1994 ). By comparing the phenotypes of Egfr^Elp/+ flies with those of trans-heterozygotes for Egfr^Elp and mutant alleles of the Dpp and STAT92E pathways, we found that the extra wing vein and rough eye phenotypes were dominantly suppressed by stat92E, but not by Mad or tkv (Fig 5; not shown). This indicated that STAT92E is involved in EGFR^Elp signaling, although we cannot rule out the involvement of the Dpp pathway simply by the lack of genetic interactions. Therefore, it appears that STAT92E not only is involved in Tor^GOF signaling but also may be involved in signaling by at least another gain-of-function mutant RTK, EGFR^Elp.

View larger version (182K): In this window In a new window Download PPT slide   Figure 5. Test of genetic interaction with a gain-of-function allele of Egfr. Wings of flies with indicated genotypes are shown. (A) Wild-type wing. Wings from EgfrElp homozygous (B) or heterozygous (C) flies exhibit extra vein tissues (arrow). This phenotype is dominantly suppressed by stat92E (D) and Ras1 (G), moderately suppressed by Draf (H), but not suppressed by Mad (E) and tkv (F).

	DISCUSSION

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

To understand the signaling downstream of a constitutively activated RTK, we genetically screened for genes required for the functions of Tor^Y9, a gain-of-function mutant RTK. Results from this screen demonstrated that, in addition to the evolutionarily conserved Ras-MAPK signaling cassette, gain-of-function RTK requires multiple noncanonical pathways to exert its full biological effects. In particular, we found that Tor^GOF possibly directly causes STAT92E activation (LI et al. 2002 ) and triggers the Dpp pathway as a result of a global derepression of dpp expression via unrestricted activity of Tor^Y9. Taken together, these results demonstrate that a gain-of-function mutant RTK is capable of utilizing multiple signaling pathways to augment its potency.

We have analyzed the possible role of STAT92E in mediating wild-type Tor functions and found that mutations in mrl have minimal, if any, effect on the expression pattern of the Tor target gene tll (LI et al. 2002 ). These results suggest that STAT92E is particularly important for Tor^GOF yet may not be required for the functions of wild-type Tor or that signaling from wild-type Tor to STAT92E branches off and does not feed into tll. Dpp signaling, on the other hand, is essential for the patterning of the D/V axis of the early embryo. dpp is expressed in the dorsal 40% of the egg along the D/V axis and is repressed in the ventral and lateral regions by a nuclear Dorsal gradient that has the highest concentration in the ventralmost region (ROTH et al. 1989 ; RUSHLOW et al. 1989 ; RAY et al. 1991 ). In contrast, Tor is activated in 15% of the egg length in the posterior region along the anterior/posterior axis. There is a minimal overlap between the expression domains of dpp and Tor activation in the posterior domain of wild-type embryos. The posterior portion of dpp expression is dependent on Tor to antagonize the repression by Dorsal (RAY et al. 1991 ). Consistent with this model, dpp mRNA is expressed in the whole posterior region (including ventral) of the early embryo (Fig 4E; also see RAY et al. 1991 ). This terminal dpp expression domain is absent in tor mutant embryos (RAY et al. 1991 ), suggesting that the ability to activate dpp expression is intrinsic to Tor. However, the bulk of dpp expression does not appear to require input from Tor signaling. Consistent with this, tor mutant embryos exhibit normal D/V patterning in the central region of the embryo. On the basis of these observations, we infer that Tor is not required for the regulation of dpp expression in the central region of the wild-type embryo. In contrast, Tor^GOF possesses ligand-independent activity and is not confined to the terminal regions. It has been shown that, Tor, via the Ras-MAPK signaling pathway, antagonizes the function of Dorsal as a transcriptional repressor (RUSCH and LEVINE 1994 ; HADER et al. 2000 ), and Dorsal normally represses dpp expression in the ventral region. Activation of Tor throughout the whole embryo thus would allow derepression of dpp globally (Fig 6). This is analogous to the derepression of tll throughout all regions of the embryo. However, we have shown that STAT92E is required in the induction of the ectopically induced tll (LI et al. 2002 ). It remains to be investigated whether the induction of ectopic dpp requires STAT92E activation.

View larger version (20K): In this window In a new window Download PPT slide   Figure 6. Model for the involvement of STAT and the Dpp pathway in mediating TorGOF signaling. In wild-type embryos, Tor is active only in the two poles due to spatially restricted ligand supply and induces tll expression by inactivating the repressors (R) or derepression (see text for more details). In addition, Tor is able to induce dpp at least in the posterior region by similar derepression mechanisms. The expression of dpp in the dorsal central region of the embryo is independent of Tor but is dependent on the lack of active nuclear repressor Dorsal that represses dpp (see text). In torGOF embryos, TorGOF, due to its ligand-independent ac-tivity, is active in all regions of the embryo and is able to antagonize repressors (R) of both tll and dpp via either MAPK or other unidentified mechanisms (not shown in the model). In addition, TorGOF is capable of activating STAT, which is able to directly bind to the promoter regions of tll and therefore enhances its transcription.

	ACKNOWLEDGMENTS

We thank Christina Ficicchia, Healani Calhoun, and Russell LaFrance for assistance; the Bloomington Drosophila Stock Center (Bloomington, IN) for providing the deficiency kit stocks and various strains; and Drs. Dirk Bohmann, Hucky Land, and Vladic Mogila for comments on the manuscript. J.L. is a recipient of the Wilmot Cancer Research Fellowship from the James P. Wilmot Foundation. This study was supported by a Howard Hughes Medical Institute Research Resources Program (grant 53000237) and a grant from the National Institutes of Health (R01 GM65774) to W.X.L.

Manuscript received October 10, 2002; Accepted for publication January 29, 2003.

	LITERATURE CITED

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

AMBROSIO, L., A. P. MAHOWALD, and N. PERRIMON, 1989  Requirement of the Drosophila raf homologue for torso function. Nature 342:288-291.[Medline]

BAKER, N. E. and G. M. RUBIN, 1989  Effect on eye development of dominant mutations in Drosophila homologue of the EGF receptor. Nature 340:150-153.[Medline]

BRONNER, G. and H. JACKLE, 1991  Control and function of terminal gap gene activity in the posterior pole region of the Drosophila embryo. Mech. Dev. 35:205-211.[Medline]

BRUMMEL, T. J., V. TWOMBLY, G. MARQUES, J. L. WRANA, and S. J. NEWFELD et al., 1994  Characterization and relationship of Dpp receptors encoded by the saxophone and thick veins genes in Drosophila. Cell 78:251-261.[Medline]

BRUNNER, D., N. OELLERS, J. SZABAD, W. H. BIGGS, III, and S. L. ZIPURSKY et al., 1994  A gain-of-function mutation in Drosophila MAP kinase activates multiple receptor tyrosine kinase signaling pathways. Cell 76:875-888.[Medline]

CASANOVA, J. and G. STRUHL, 1993  The torso receptor localizes as well as transduces the spatial signal specifying terminal body pattern in Drosophila. Nature 362:152-155.[Medline]

CASANOVA, J., M. LLIMARGAS, S. GREENWOOD, and G. STRUHL, 1994  An oncogenic form of human raf can specify terminal body pattern in Drosophila. Mech. Dev. 48:59-64.[Medline]

CHANG, H. C. and G. M. RUBIN, 1997  14-3-3 epsilon positively regulates Ras-mediated signaling in Drosophila. Genes Dev. 11:1132-1139.[Abstract/Free Full Text]

DICKSON, B. J., A. VAN DER STRATEN, M. DOMINGUEZ, and E. HAFEN, 1996  Mutations modulating Raf signaling in Drosophila eye development. Genetics 142:163-171.[Abstract]

DOYLE, H. J. and J. M. BISHOP, 1993  Torso, a receptor tyrosine kinase required for embryonic pattern formation, shares substrates with the sevenless and EGF-R pathways in Drosophila. Genes Dev. 7:633-646.[Abstract/Free Full Text]

DUFFY, J. B. and N. PERRIMON, 1994  The torso pathway in Drosophila: lessons on receptor tyrosine kinase signaling and pattern formation. Dev. Biol. 166:380-395.[Medline]

FLYBASE,, 1998  FlyBase: a Drosophila database. Flybase Consortium. Nucleic Acids Res. 26:85-88.

GHIGLIONE, C., N. PERRIMON, and L. A. PERKINS, 1999  Quantitative variations in the level of MAPK activity control patterning of the embryonic termini in Drosophila. Dev. Biol. 205:181-193.[Medline]

GREENWOOD, S. and G. STRUHL, 1997  Different levels of Ras activity can specify distinct transcriptional and morphological consequences in early Drosophila embryos. Development 124:4879-4886.[Abstract]

HADER, T., D. WAINWRIGHT, T. SHANDALA, R. SAINT, and H. TAUBERT et al., 2000  Receptor tyrosine kinase signaling regulates different modes of Groucho-dependent control of Dorsal. Curr. Biol. 10:51-54.[Medline]

HOU, X. S., M. B. MELNICK, and N. PERRIMON, 1996  Marelle acts downstream of the Drosophila HOP/JAK kinase and encodes a protein similar to the mammalian STATs. Cell 84:411-419. [Erratum. Cell 85 (2)].[Medline]

KARIM, F. D., H. C. CHANG, M. THERRIEN, D. A. WASSARMAN, and T. LAVERTY et al., 1996  A screen for genes that function downstream of Ras1 during Drosophila eye development. Genetics 143:315-329.[Abstract]

KLINGLER, M., M. ERDELYI, J. SZABAD, and C. NUSSLEIN-VOLHARD, 1988  Function of torso in determining the terminal anlagen of the Drosophila embryo. Nature 335:275-277.[Medline]

KOCKEL, L., G. VORBRUGGEN, H. JACKLE, M. MLODZIK, and D. BOHMANN, 1997  Requirement for Drosophila 14-3-3 zeta in Raf-dependent photoreceptor development. Genes Dev. 11:1140-1147.[Abstract/Free Full Text]

LI, W., E. M. SKOULAKIS, R. L. DAVIS, and N. PERRIMON, 1997  The Drosophila 14-3-3 protein Leonardo enhances Torso signaling through D-Raf in a Ras 1-dependent manner. Development 124:4163-4171.[Abstract]

LI, W., E. NOLL, and N. PERRIMON, 2000  Identification of autosomal regions involved in Drosophila Raf function. Genetics 156:763-774.[Abstract/Free Full Text]

LI, W. X., H. AGAISSE, B. MATHEY-PREVOT, and N. PERRIMON, 2002  Differential requirement for STAT by gain-of-function and wild-type receptor tyrosine kinase Torso in Drosophila. Development 129:4241-4248.

LIAW, G. J., K. M. RUDOLPH, J. D. HUANG, T. DUBNICOFF, and A. J. COUREY et al., 1995  The torso response element binds GAGA and NTF-1/Elf-1, and regulates tailless by relief of repression. Genes Dev. 9:3163-3176.[Abstract/Free Full Text]

LU, X., T. B. CHOU, N. G. WILLIAMS, T. ROBERTS, and N. PERRIMON, 1993  Control of cell fate determination by p21ras/Ras1, an essential component of torso signaling in Drosophila. Genes Dev. 7:621-632.[Abstract/Free Full Text]

MARSHALL, C. J., 1995  Specificity of receptor tyrosine kinase signaling: transient versus sustained extracellular signal-regulated kinase activation. Cell 80:179-185.[Medline]

MELNICK, M. B., L. A. PERKINS, M. LEE, L. AMBROSIO, and N. PERRIMON, 1993  Developmental and molecular characterization of mutations in the Drosophila-raf serine/threonine protein kinase. Development 118:127-138.[Abstract]

NELLEN, D., M. AFFOLTER, and K. BASLER, 1994  Receptor serine/threonine kinases implicated in the control of Drosophila body pattern by decapentaplegic. Cell 78:225-237.[Medline]

NEWFELD, S. J., E. H. CHARTOFF, J. M. GRAFF, D. A. MELTON, and W. M. GELBART, 1996  Mothers against dpp encodes a conserved cytoplasmic protein required in DPP/TGF-beta responsive cells. Development 122:2099-2108.[Abstract]

PENTON, A., Y. CHEN, K. STAEHLING-HAMPTON, J. L. WRANA, and L. ATTISANO et al., 1994  Identification of two bone morphogenetic protein type I receptors in Drosophila and evidence that Brk25D is a decapentaplegic receptor. Cell 78:239-250.[Medline]

PERKINS, L. A., I. LARSEN, and N. PERRIMON, 1992  corkscrew encodes a putative protein tyrosine phosphatase that functions to transduce the terminal signal from the receptor tyrosine kinase torso. Cell 70:225-236.[Medline]

PIGNONI, F., R. M. BALDARELLI, E. STEINGRIMSSON, R. J. DIAZ, and A. PATAPOUTIAN et al., 1990  The Drosophila gene tailless is expressed at the embryonic termini and is a member of the steroid receptor superfamily. Cell 62:151-163.[Medline]

PIGNONI, F., E. STEINGRIMSSON, and J. A. LENGYEL, 1992  bicoid and the terminal system activate tailless expression in the early Drosophila embryo. Development 115:239-251.[Abstract]

RAY, R. P., K. ARORA, C. NUSSLEIN-VOLHARD, and W. M. GELBART, 1991  The control of cell fate along the dorsal-ventral axis of the Drosophila embryo. Development 113:35-54.[Abstract]

ROBERTSON, S. C., J. A. TYNAN, and D. J. DONOGHUE, 2000  RTK mutations and human syndromes when good receptors turn bad. Trends Genet 16:265-271. [Erratum. Trends Genet. 16 (8): 368].[Medline]

ROTH, S., D. STEIN, and C. NUSSLEIN-VOLHARD, 1989  A gradient of nuclear localization of the dorsal protein determines dorsoventral pattern in the Drosophila embryo. Cell 59:1189-1202.[Medline]

RUSCH, J. and M. LEVINE, 1994  Regulation of the dorsal morphogen by the Toll and torso signaling pathways: a receptor tyrosine kinase selectively masks transcriptional repression. Genes Dev. 8:1247-1257.[Abstract/Free Full Text]

RUSHLOW, C. A., K. HAN, J. L. MANLEY, and M. LEVINE, 1989  The graded distribution of the dorsal morphogen is initiated by selective nuclear transport in Drosophila. Cell 59:1165-1177.[Medline]

SEWING, A., B. WISEMAN, A. C. LLOYD, and H. LAND, 1997  High-intensity Raf signal causes cell cycle arrest mediated by p21Cip1. Mol. Cell. Biol. 17:5588-5597.[Abstract]

SIMON, M. A., D. D. BOWTELL, G. S. DODSON, T. R. LAVERTY, and G. M. RUBIN, 1991  Ras1 and a putative guanine nucleotide exchange factor perform crucial steps in signaling by the sevenless protein tyrosine kinase. Cell 67:701-716.[Medline]

SIMON, M. A., G. S. DODSON, and G. M. RUBIN, 1993  An SH3-SH2-SH3 protein is required for p21Ras1 activation and binds to sevenless and Sos proteins in vitro. Cell 73:169-177.[Medline]

SINGER, J. B., R. HARBECKE, T. KUSCH, R. REUTER, and J. A. LENGYEL, 1996  Drosophila brachyenteron regulates gene activity and morphogenesis in the gut. Development 122:3707-3718.[Abstract]

SPRENGER, F. and C. NUSSLEIN-VOLHARD, 1992  Torso receptor activity is regulated by a diffusible ligand produced at the extracellular terminal regions of the Drosophila egg. Cell 71:987-1001.[Medline]

STEINGRIMSSON, E., F. PIGNONI, G. J. LIAW, and J. A. LENGYEL, 1991  Dual role of the Drosophila pattern gene tailless in embryonic termini. Science 254:418-421.[Abstract/Free Full Text]

STRECKER, T. R., S. R. HALSELL, W. W. FISHER, and H. D. LIPSHITZ, 1989  Reciprocal effects of hyper- and hypoactivity mutations in the Drosophila pattern gene torso. Science 243:1062-1066.[Abstract/Free Full Text]

STRECKER, T. R., M. L. YIP, and H. D. LIPSHITZ, 1991  Zygotic genes that mediate torso receptor tyrosine kinase functions in the Drosophila melanogaster embryo. Proc. Natl. Acad. Sci. USA 88:5824-5828.[Abstract/Free Full Text]

TANG, A. H., T. P. NEUFELD, G. M. RUBIN, and H. A. MULLER, 2001  Transcriptional regulation of cytoskeletal functions and segmentation by a novel maternal pair-rule gene, lilliputian. Development 128:801-813.[Abstract]

TSUDA, L., Y. H. INOUE, M. A. YOO, M. MIZUNO, and M. HATA et al., 1993  A protein kinase similar to MAP kinase activator acts downstream of the raf kinase in Drosophila. Cell 72:407-414.[Medline]

VAN DER STRATEN, A., C. ROMMEL, B. DICKSON, and E. HAFEN, 1997  The heat shock protein 83 (Hsp83) is required for Raf-mediated signalling in Drosophila. EMBO J. 16:1961-1969.[Medline]

WEIGEL, D., G. JURGENS, M. KLINGLER, and H. JACKLE, 1990  Two gap genes mediate maternal terminal pattern information in Drosophila. Science 248:495-498.[Abstract/Free Full Text]

WHARTON, K. A., R. P. RAY, and W. M. GELBART, 1993  An activity gradient of decapentaplegic is necessary for the specification of dorsal pattern elements in the Drosophila embryo. Development 117:807-822.[Abstract]

WITTWER, F., A. VAN DER STRATEN, K. KELEMAN, B. J. DICKSON, and E. HAFEN, 2001  Lilliputian: an AF4/FMR2-related protein that controls cell identity and cell growth. Development 128:791-800.[Abstract]

WOODS, D., D. PARRY, H. CHERWINSKI, E. BOSCH, and E. LEES et al., 1997  Raf-induced proliferation or cell cycle arrest is determined by the level of Raf activity with arrest mediated by p21Cip1. Mol. Cell. Biol. 17:5598-5611.[Abstract]

YAN, R., S. SMALL, C. DESPLAN, C. R. DEAROLF, and J. E. DARNELL, JR., 1996  Identification of a Stat gene that functions in Drosophila development. Cell 84:421-430.[Medline]

This article has been cited by other articles:

				K. L. Pepple, A. E. Anderson, B. J. Frankfort, and G. Mardon A Genetic Screen in Drosophila for Genes Interacting With senseless During Neuronal Development Identifies the Importin moleskin Genetics, January 1, 2007; 175(1): 125 - 141. [Abstract] [Full Text] [PDF]

				C.-J. Zettervall, I. Anderl, M. J. Williams, R. Palmer, E. Kurucz, I. Ando, and D. Hultmark A directed screen for genes involved in Drosophila blood cell activation PNAS, September 28, 2004; 101(39): 14192 - 14197. [Abstract] [Full Text] [PDF]