outstretched

Transcriptional Regulation

Notch (N) signal is activated at the dorsoventral (DV) border of the Drosophila eye disc and is important for growth of the eye disc. In this study, the Pax protein Eyg is shown to be a major effector mediating the growth promotion function of N. eyg transcription is induced by N signaling occurring at the DV border. Like N, eyg controls growth of the eye disc. Loss of N signaling can be compensated by overexpressing eyg, whereas loss of the downstream eyg blocks the function of N signaling. In addition, N and eyg can induce expression of upd, which encodes the ligand for the Jak/STAT pathway and acts over long distance to promote cell proliferation. Loss of eyg or N can be compensated by overexpressing upd. These results suggest that upd is a major effector mediating the function of eyg and N. The functional link from N to eyg to upd explains how the localized Notch activation can achieve global growth control (Chao, 2004).

Notch is activated at the DV boundary of the early eye disc. This equatorial N signal then activates eyg expression at the transcriptional level. When N signal is reduced, eyg expression is reduced. When N signal is elevated, eyg expression is induced. Induction of eyg expression occurs at the DV border between the dorsal Dl-expressing and the ventral Ser-expressing cells. Furthermore, when the upstream N signal is blocked, overexpression of eyg can rescue the growth defect in the eye, whereas increasing N signaling cannot rescue the eye-growth defect caused by the downstream eyg gene. This analysis shows that the induction of eyg by N is dependent on the ligands Dl and Ser, and involves the effector Su(H) and the antagonist Hairless. Thus, the localized activation of N signal is transmitted to the induction of a transcription factor, Eyg, which then promotes cell proliferation. A recent paper (Dominguez, 2004) has come to the same conclusion (Chao, 2004).

Eyg is a transcription factor, so must activate the transcription of some genes that promote cell proliferation. Upd is reported to act through the Jak/STAT signaling pathway to promote cell proliferation. upd expression is dependent on eyg and N signaling. Furthermore, when the upstream N signaling or eyg is reduced, overexpression of upd can rescue the growth defect. The overgrowth effect due to overexpression of the upstream N or eyg is blocked when the downstream upd is defective. The results suggest that upd is a major effector for the growth promotion by N and eyg (Chao, 2004).

These results have demonstrated the functional link from Notch to eyg to upd in the promotion of eye growth. The link to upd solved a long-standing problem. N signaling is activated locally at the border between the dorsal Dl-expressing cells and the ventral Ser-expressing cells. How does a localized activation of N signal promote cell proliferation throughout the entire eye disc? The finding of eyg as the major mediator of N function did not solve the problem, since Eyg is a transcriptional factor and is expected to affect target gene expression autonomously. The link from eyg to upd provided a solution, because Upd is a diffusible signaling molecule. Upd protein can distribute over a long distance and exert long-range non-autonomous effect to promote cell proliferation (Tsai, 2004). So the localized N activation can locally activate eyg, which then turns on upd expression, probably through a short-range signal. The Upd signal then acts over a long range to promote cell proliferation in the early eye disc (Chao, 2004).

Although N activates eyg, and eyg activates upd, these transcriptional activation may be direct or indirect. When novel DV borders were created by ectopic expressing Dl or Ser, eyg is induced non-autonomously at the border of these clones. It is also noted that in Su(H) mutant clones, mutant cells at the border of the clone can still express eyg-lacZ. These observations suggest that N may induce a short-range signal, which then activates eyg expression. Alternatively, the apparent non-autonomous induction may be due to perdurance of the reporter protein in cells that were once close to the clone border. The induction of upd by eyg also may be indirect. Clonal expression of eyg also induced upd expression non-autonomously. In addition, based on RNA in situ hybridization, eyg expression in the eye disc does not extend to the posterior margin, so does not overlap with the expression domain of upd (Tsai, 2004). These observations suggested that the effect of eyg on upd expression may be indirect. However, an eyg enhancer trap line showed reporter expression extending to the posterior margin (Dominguez, 2004). Thus, the possibility that Eyg can directly activate the expression of upd cannot be excluded (Chao, 2004).

The activation of eyg and upd are context dependent. N^act does not induce eyg expression in antenna and wing discs. In the eye disc, N^act induces eyg expression only in the region anterior to the MF, and not within the wg expression domain in the lateral margin. Similarly, N^act and eyg can only induce upd expression at the margin, but not in the center of the eye disc. N^act induces upd at the posterior margin but not lateral margins, while eyg can induce upd in the lateral margins but not in the posterior margin. The context dependence indicates that additional factors are involved to determine the specificity of induction (Chao, 2004).

In a late third instar eye disc, eyg is expressed in an equatorial domain that does not overlap with the disc margin, so eyg cannot induce upd. In early eye disc, eyg expression domain comes closer to the posterior margin. Thus, the induction of upd by eyg is likely at second instar, which is consistent with the timing of upd expression (Chao, 2004).

Although eyg plays an important role in mediating the growth-promoting N signal, it is probably not the only effector. In the eyg^M3-12 null mutant, ey>N^act does not rescue the endogenous eye field, but can still induce proliferation to provide the antennal disc and an extra eye field. Thus, N can induce proliferation by an eyg-independent mechanism. The effect on antenna and on eye seems to be separate, because ey>N^act can induce a large antenna disc with duplicate or triplicate antennal field without rescue of the eye disc. Because N can induce upd, but not eyg, in the posterior margin, the induction of upd can also be through an eyg-independent mechanism (Chao, 2004).

N^act can induce overgrowth in the central domain of the eye disc. In this case, eyg, but not upd, is induced. In addition, the overgrowth does not extend much beyond the clone. Ectopic eyg in the central domain also induces proliferation without inducing upd. In eyg¹ mutant, there is no upd-lacZ expression in eye disc, but the eye is only slightly reduced. These results suggest that the N signaling and eyg can induce local proliferation independent of upd (Chao, 2004).

Mutations in erupted, the Drosophila ortholog of mammalian tumor susceptibility gene 101, elicit non-cell-autonomous overgrowth caused by increased Notch-mediated signaling and ectopic expression of the Notch target gene unpaired

The reproducible pattern of organismal growth during metazoan development is the product of genetically controlled signaling pathways. Patterned activation of these pathways shapes developing organs and dictates overall organismal shape and size. Patches of tissue that are mutant for the Drosophila Tsg101 ortholog, erupted, cause dramatic overexpression of adjacent wild-type tissue. Tsg101 proteins function in endosomal sorting and are required to incorporate late endosomes into multivesicular bodies. Drosophila cells with impaired Tsg101 function show accumulation of the Notch receptor in intracellular compartments marked by the endosomal protein Hrs. This causes increased Notch-mediated signaling and ectopic expression of the Notch target gene unpaired (upd), which encodes the secreted ligand of the JAK-STAT pathway. Activation of JAK-STAT signaling in surrounding wild-type cells correlates with their overgrowth. These findings define a pathway by which changes in endocytic trafficking can regulate tissue growth in a non-cell-autonomous manner (Moberg, 2005). Tsg101 possesses the ability to bind monoubiquitinated substrates. These substrates are predicted to be the ubiquitinated cytoplasmic tails of membrane bound proteins, and this interaction is predicted to deliver cargos to the lysosome via multivesicular bodies (Moberg, 2005 and reference therein).

The Notch receptor has two properties that implicate it in a pathway by which ept mutations non-cell-autonomously promote tissue growth. (1) The restricted activation of Notch in cells along the dorsoventral (D/V) boundary of the eye imaginal disc is required for growth of the entire eye. (2) Ub-dependent endocytosis plays an important role in regulating Notch activity in vivo. In mammalian cells, ubiquitination and endocytosis contributes to Notch1 activation, and, in Drosophila, there is evidence to suggest that the ubiquitin ligase Deltex may be required for endocytosis-dependent Notch activation. Further, alleles of the endosomal sorting gene Hrs, the homolog of yeast Vps27, affect Notch localization in imaginal disc cells, indicating that Notch is a physiological target of the MVB pathway (Moberg, 2005).

In light of these observations, ept mosaic eye discs were stained with an antibody specific to the Notch cytoplasmic domain (anti-N^cyto). Notch protein is detected in wild-type eye discs most prominently in a stripe of cells within the morphogenetic furrow (MF) and is concentrated at the apical cell surface. In contrast, ept cells contain elevated levels of Notch. This increase occurs in ept clones throughout the eye disc, but it is most apparent in clones that lie within or posterior to the MF. Moreover, the Notch in ept cells accumulates in nonnuclear, intracellular puncta that also stain positive for Ub, and for the endosomal protein Hrs. Together, these data indicate that ept mutations block the routing of ubiquitinated cell surface proteins, among them Notch, in an Hrs-positive endosomal compartment (Moberg, 2005).

Notch is normally processed in cells by a series of cleavage events required for receptor maturation and presentation at the cell surface, and for ligand-stimulated activation of the Notch pathway. Because ubiquitination and endocytosis have been shown to affect Notch cleavage, attempts were made to determine if ept mutations also affect Notch processing. Eye-antennal discs composed of ept mutant cells [ept/M(3)] or FRT80B control cells (FRT80B/M(3)) were generated by the eyFLP/Minute technique. Immunoblot of tissue extracts with the anti-N^cyto antibody confirms that Notch levels are increased considerably in eye-antennal discs composed of ept mutant cells, and shows that ept mutant cells are enriched in a ~120 kDa form of Notch. The molecular identity of this fragment has not been determined, but its size appears similar to certain processed forms of Notch. Indeed, while no one form of Notch predominates in wild-type cells, this species appears to be the most abundant Notch species in ept cells (Moberg, 2005).

To examine Notch activation, clones of ept mutant cells were generated in the presence of the Notch-inducible transgene E(spl)mβ-CD2, a Suppressor of Hairless (Su(H))-dependent transcriptional reporter that has been used to detect equatorial Notch activation in the developing eye. Posterior to the MF, CD2 expression is detected in the interommatidial cells, and outlines a single cell from each photoreceptor cluster in a mirror-image pattern along the equator. Thus, in addition to equatorial activation, the reporter detects Notch activation in postmitotic interommatidial cells, and in the R3-R4 cell fate choice. In ept mutant clones, reporter activity is strongly elevated. The degree of activation exceeds that observed in wild-type eye discs, and it does not appear to depend upon the location of ept cells within the disc, occurring on either side of the MF and in the antennal disc. Some ept cells within a single optical section appear not to activate the Notch reporter. However, in most of these cases, CD2, which localizes to cell membranes, can be detected in a focal plane slightly offset from that of the nuclear green fluorescent protein (GFP). Thus, these data show that defects in Notch regulation in ept cells are accompanied by ectopic and excessive activation of the Notch pathway (Moberg, 2005).

The requirement for Notch in eye disc growth has been linked to its ability to induce expression of the eyegone (eyg) gene at the D/V boundary of the eye disc. eyg encodes a Pax6-like transcription factor (Eyg) required for disc growth, and, like Notch, ectopic expression of eyg is able to induce growth nonautonomously. Consistent with its effect on Notch, it was found that ept mutant cells express elevated levels of Eyg compared to surrounding wild-type cells. Thus, Eyg may function downstream of Notch within ept cells to promote the growth of surrounding cells in a manner similar to its normal growth-promoting role at the D/V boundary (Moberg, 2005).

Recent work suggests that the unpaired (upd) gene may be an important growth regulatory target of Notch. upd encodes the secreted ligand (Upd) of the Domeless (Dome) receptor, which signals through the JAK-STAT pathway. JAK-STAT signaling is implicated in many processes during Drosophila development, including the control of cell proliferation, cell motility, stem cell renewal, and planar cell polarity. upd is required for normal growth of the eye, and ectopic expression of upd in the larval eye nonautonomously promotes cell proliferation and produces enlarged and misshapen eyes similar to those observed in ept mosaics. Significantly, Notch is both necessary and sufficient to activate upd transcription along the posterior margin of the eye disc (Moberg, 2005).

When ept mosaic eye discs were stained with an anti-Upd antiserum, a dramatic increase was observed in the level of Upd protein in ept mutant cells compared to adjacent wild type cells. Consistent with a transcriptional link between Notch and upd, Upd protein accumulation appears coincident with expression of the Notch reporter, and ept mosaic eye-antennal discs contain clones of cells expressing very high levels of upd mRNA. Together, these observations suggest that Notch, perhaps acting via Eyg, promotes ectopic upd expression in ept mutant cells (Moberg, 2005).

Clonal overexpression of upd induces localized tissue outgrowths and deregulates the division of surrounding cells. This mitogenic activity is linked to induction of cyclin D, and to accelerated progression through the G1 phase of the cell cycle. ept mutant clones can produce phenotypes quite similar to clonal overexpression of upd. In one example of an ept clone, lower half of the disc appeared morphologically normal, while the other half, despite being composed largely of wild-type cells, was misshapen and enlarged. This localized effect correlated with proximity to a large ept mutant clone expressing Upd. Similar hyperplastic growth was associated with clones of upd-expressing cells in the antennal disc. The patterns of BrdU incorporation in ept mosaic eye discs are disorganized, and the number of BrdU-labeled nuclei increases in proximity to Upd-expressing ept mutant cells in the eye and antenna. This aberrant cell proliferation occurs in GFP-positive wild-type cells. Hence, the growth-promoting activity of ept mutations is likely mediated by a diffusible extracellular signal like Upd (Moberg, 2005).

Receipt of the Upd signal via Domeless initiates a signaling cascade that activates a transcription factor encoded by the stat92E gene. stat92E encodes the Drosophila ortholog of the mammalian signal transducers and activators of transcription (STAT) family of transcriptional regulators, which function in diverse processes such as immunity and oncogenesis, and is the only member of this gene family in Drosophila. Heterozygosity for a stat92E loss-of-function allele (stat92E⁰⁶³⁴⁶) strongly suppresses the nonautonomous eye overgrowth associated with mosaicism for ept mutations, such that ept-mosaic;stat92E⁰⁶³⁴⁶/+ eyes are comparable in size to control FRT80B mosaic eyes. Thus, nonautonomous overgrowth elicited by ept mutations is sensitive to the genetic dosage of the Upd-responsive transcription factor stat92E. In light of the effect on Upd, these data strongly indicate that the growth-promoting activity of ept mutant cells requires Upd-dependent activation of the JAK-STAT pathway in adjacent tissue (Moberg, 2005).

ept mutant clones in mosaic eye discs are small and survive poorly into adulthood. It is possible that this is the result of cell competition, a process by which slow-growing cells in the vicinity of wild-type cells are eliminated. If so, then the poor survival of ept cells might be rescued by eliminating competing cells. Therefore the growth characteristics were examined of ept/M(3) discs, which are composed almost entirely of cells lacking Tsg101 function. ept/M(3) animals reach the larval 'wandering' stage 4 days later than control larvae, and, when they do, they are enlarged. A small fraction of these animals pupate and die before becoming pharate adults. The remainder die as giant larvae containing high levels of Upd (Moberg, 2005).

Allowing ept mutant cells to grow in epithelia lacking wild-type cells also uncovers a context-dependent cell-autonomous overgrowth phenotype. Rather than surviving poorly as they do in mosaic discs, ept/M(3) eye discs overgrow into large masses that lack normal disc morphology. These masses are composed of folded and convoluted sheets of cells fused together, and they often include a distended sac-like structure. The overgrowth phenotypes of ept/M(3) animals and discs do not reflect an increased rate of growth: control L3 larvae are the same size as ept/M(3) larvae of the same temporal age, and the ept/M(3) eye discs, while mispatterned, are not obviously increased in size. Thus, the ept/M(3) masses are the result of an extended larval phase, and a failure of the disc to stop growing when it reaches the appropriate size. Thus, cells lacking Tsg101 may be unable to respond to signals that normally sense and restrict organ size (Moberg, 2005).

Targets of activity

The genetic and molecular data regarding outstretched and its relationship to Hopscotch and Stat92E are all consistent with the predicted role of Os as a ligand that activates the JAK signaling cascade. To directly investigate this hypothesis, os was expressed in Drosophila cells, which were then assayed for tyrosine phosphorylation of Hop. The cell line chosen for this experiment is the Clone 8 (Cl.8) line, derived from developing wing imaginal disc. For cells to respond to Os by phosphorylating Hop, it was hypothesized that some transmembrane receptor would be required to bind Os by an extracellular domain, and be associated with Hop on the intracellular domain. Since no such receptor has yet been identified in flies, cells were chosen that were derived from a tissue known to be responsive to such a signal. It has been shown that Om(1E) protein from Drosophila ananassae, overexpressed in the wing disc, causes defects (Juni, 1996), suggesting that a receptor for this Os homolog must be present in the wing discs of D. ananassae. Thus, the D. melanogaster wing disc-derived Cl.8 cell line seemed a likely candidate to express a receptor for Os. To show Os-dependent tyrosine phosphorylation of Hop, anti-Hop immunoprecipitates from os-transfected cells were prepared and tested for reactivity with the anti-phosphotyrosine antibody 4G10. Although Hop protein is detectable in all samples, Hop is tyrosine phosphorylated only in immunoprecipitates prepared from os-transfected cells. Transfection of cells with os lacking a signal sequence does not result in Hop phosphorylation, consistent with the notion that Os is required extracellularly for signaling to occur. To further demonstrate that extracellularly provided Os is necessary and sufficient to observe Hop phosphorylation, Cl.8 cells were cocultured with S2 cells transiently transfected with Os. After thorough removal of the nonadherent S2 cells, Hop immunoprecipitates were prepared from Cl.8 cell lysates and analyzed. Hop phosphorylation is only seen when Cl.8 cells are cultured in the presence of os-transfected S2 cells. Identical results are obtained when Cl.8 cells are grown in the presence of conditioned medium taken from os-transfected 293T cells. These data are consistent with the hypothesis that Os is an extracellular ligand that binds a membrane-bound receptor to activate the JAK signaling pathway (Harrison, 1998).

The determination of sexual identity in Drosophila depends upon a system that measures the X chromosome to autosome ratio (X/A). This system relies upon the unequal expression of X-linked numerator genes in 1X and 2X nuclei. The numerators activate a special Sex lethal promoter, Sxl-Pe, in 2X/2A nuclei, but not 1X/2A nuclei. By multimerizing a conserved Sxl-Pe sequence block, a gain-of-function promoter, Sxl-Pe_GOF, is generated that is inappropriately active in 1X/2A nuclei. GOF activity requires the X-linked unpaired (upd) gene, which encodes a ligand for the Drosophila JAK/STAT signaling pathway. upd also functions as a numerator element in regulating wild-type Sxl-Pe reporters. The JAK kinase, Hopscotch, and the STAT DNA-binding protein, Marelle, are also required for Sxl-Pe activation (Jinks, 2000).

The numerators most important for turning on Sxl are sis-a and sis-b (scute). They are expressed throughout the embryo, and mutations in both can have quite pronounced effects on Sxl-Pe activity. However, neither of these numerators is critical for the gain-of-function activity of the Sxl-Pe_GOF promoter. Instead, the two numerators that contribute most to Sxl-Pe_GOF activity are the segmentation genes runt and upd. At the syncytial blastoderm stage, run is expressed in a broad central domain, and it is in this region that Sxl activation is defective in 2X/2A run mutants. Except for a dorsal crescent in the head, the upd expression domain closely coincides with that of run. It is in this same central run-upd domain that the highest levels of Sxl-Pe_GOF promoter activity are observed. Moreover, in both run and upd mutant males, Sxl- Pe_GOF promoter activity is severely impaired. From these findings, it can be inferred that the multimerized 72 bp fragment contains cis-acting targets for run and upd action (Jinks, 2000).

Since Upd is a secreted ligand, it is unlikely that it interacts directly with sequences in the 72 bp fragment. Instead, the data suggests that Upd acts by turning on a Drosophila JAK/STAT signaling cascade consisting of the Hop protein kinase and the Mrl transcription factor. In this model, the extracellular Upd ligand would activate the Drosophila JAK protein Hop. Hop would in turn phosphorylate the D-STAT homolog Mrl, which would then enter the nucleus and activate Sxl-Pe. That the Mrl protein is critical for the activity of Sxl-Pe_GOF is demonstrated by the dramatic reduction in beta-galactosidase expression seen in both 1X/2A and 2X/2A embryos derived from homozygous mrl^- germline clones (Jinks, 2000).

Protein Interactions

The JAK/STAT signaling pathway plays important roles in vertebrate development and the regulation of complex cellular processes. Components of the pathway are conserved in Dictyostelium, Caenorhabditis, and Drosophila, yet the complete sequencing and annotation of the D. melanogaster and C. elegans genomes has failed to identify a receptor, raising the possibility that an alternative type of receptor exists for the invertebrate JAK/STAT pathway. domeless (dome) codes for a transmembrane protein required for all JAK/STAT functions in the Drosophila embryo. This includes its known requirement for embryonic segmentation and a newly discovered function in trachea specification. The DOME protein has a similar extracellular structure to the vertebrate cytokine class I receptors, although its sequence has greatly diverged. Like many interleukin receptors, DOME has a cytokine binding homology module (CBM) and three extracellular fibronectin-type-III domains (FnIII). Despite its low degree of overall similarity, key amino acids required for signaling in the vertebrate cytokine class I receptors are conserved in the CBM region. DOME is a signal-transducing receptor with most similarity to the IL-6 receptor family, but it also has characteristics found in the IL-3 receptor family. This suggests that the vertebrate families evolved from a single ancestral receptor that also gave rise to dome (Brown, 2001).

JAK/STAT signaling was first identified in vertebrates as mediating the response to some cytokines and growth factors. Ligand binding induces receptor homo- or hetero-dimerization and subsequent signal transduction. The receptors lack a tyrosine kinase domain but associate with cytoplasmic tyrosine kinases of the JAK family. After receptor dimerization, JAK phosphorylates a tyrosine residue on the receptor, and cytoplasmic STAT is recruited to the complex. JAK then phosphorylates STAT, which dimerizes, translocates to the nucleus, and induces gene transcription. In Drosophila, one JAK encoded by hopscotch (hop), one STAT encoded by stat92E, and one ligand encoded by unpaired (upd) have been identified, but no receptor has been found. Mutations for either hop, stat92E, or upd result in an identical, characteristic segmentation phenotype (Brown, 2001).

Mutations in stat92E affect the posterior spiracles, part of the respiratory apparatus of the larva. In a screen for P elements insertion mutations that give a phenotype similar to stat92E, domeless was identified. The six alleles, three strong (dome²¹⁷, dome⁴⁴¹, and dome⁴⁶⁸) and three weak (dome³²¹, dome⁴⁰⁵, and dome³⁶⁷), all affect the shape of the posterior spiracles, with the strongest leading to a loss of the characteristic dome shape. Mobilization of the P element reverts both the lethality and the phenotype, confirming that the insertions cause the observed defects (Brown, 2001).

A database sequence search using DNA flanking the P elements identified an expressed-sequence tag (EST) encoding a putative transmembrane protein. The finding that expression of this cDNA rescues the dome spiracle phenotype confirms that this cDNA encodes the dome gene (Brown, 2001).

The 4.8 kb dome encodes a 1282 amino acid protein with a putative signal peptide of 23 amino acids and a transmembrane domain. The extracellular region contains five fibronectin-type-III (FnIII) domains, of which two have similarity to the cytokine binding module (CBM) found in the vertebrate cytokine receptor class I family. No invertebrate receptors of this family have been described, despite the sequencing projects in Drosophila and Caenorhabditis being complete. The vertebrate cytokine receptor family comprises more than 20 different receptors that signal through the JAK/STAT pathway. The CBM present in the vertebrate interleukin receptors is typically composed of two FnIII domains that contain a set of four conserved cysteine residues in the N-terminal domain and a WSXWS motif in the C-terminal domain. Dome contains these features, but the C-terminal domain of Dome has an incomplete WSXWS motif (NTXWS). Dome has 18% identity to LIFR and 26% identity to CNTFR and is within the typical range of sequence similarity limits for all cytokine receptors. Interestingly, Dome also has some characteristics of the IL-3 receptor family. These characteristics include an alternating region of hydrophobic residues (YXLXVRVR) in the CBM-C domain and the incomplete WSXWS motif, present only in IL-3Ralpha. The intracellular region of Dome is rich in both serine and threonine (16%) as well as proline (11%) and has an acidic region; features shared by the IL-2Rß receptor and GM-CSFR. Like other receptors of this class, Dome lacks a kinase domain. Although the sequence similarity of Dome with the vertebrate cytokine class I receptors is low, the shared characteristics suggest that domeless may encode the elusive Drosophila JAK/STAT receptor (Brown, 2001).

Tests were performed to see whether dome and stat92E interact genetically. Zygotic stat92E homozygotes have a very mild spiracle phenotype due to the persistence of maternally expressed RNA. Despite this, the weak dome³⁶⁷ phenotype is strongly enhanced by stat92E mutants, suggesting that both genes are in the same genetic pathway (Brown, 2001).

RNA in situ hybridization shows that, similar to hop (JAK) and stat92E, dome is expressed maternally. At later stages dome expression appears ubiquitous, although there is some variation in the levels of expression. At stage 11 the tracheal pits show more intense expression, whereas at stage 14, higher expression is detected in the posterior spiracles, gut, and head (Brown, 2001).

To determine the phenotype caused by eliminating domeless maternal and zygotic products, germ line clones were induced. Maternal and zygotic dome embryos have segmentation defects identical to those reported for mutations in the stat92E and hop (JAK) mutants. Defects include the deletion of the A5 and most of the A4 denticle belts, partial or total fusion of A6 to A7, and a variable reduction of the thoracic and the A8 segments. These phenotypes are also observed in Df(1)osUE69, which deletes the ligand, upd. The segmentation defects in stat92E, upd, and hop have been shown to be due to the abnormal expression of pair rule genes. In dome germ line clones, the expression of even-skipped is affected in stripes 3 and 5, as described for the other members of this pathway (Brown, 2001).

To further investigate whether dome has the genetic characteristics expected of the JAK/STAT receptor, dome interactions with upd, the known JAK/STAT ligand, were tested. To do this, advantage was taken of the fact that when the h-GAL4 line is used for ectopic expression of upd in the embryo, the result is abnormal head formation in 81% of the embryos. When upd is expressed ectopically in dome zygotic mutant embryos, this proportion is reduced to 16%. This result is consistent with dome being necessary to transduce the upd signal (Brown, 2001).

To find out if the intracellular domain of Dome is required for its function, UAS constructs were made in which the putative intracellular domain was deleted. One construct, UAS-dome^DeltaCYT, contains the extracellular and transmembrane portion of the protein and should be membrane bound. The other, UAS-dome^DeltaTMCYT, contains only the extracellular part and might be secreted. Neither of these proteins is able to rescue the dome zygotic phenotype, proving that the Dome intracellular domain is required for signal transduction. Because both forms still contain the cytokine binding domain, they must have the potential to titrate the ligand and act as signaling antagonists. This has been shown to be the case for the soluble form of gp130, the signal-transducing subunit of many vertebrate cytokine class I receptors. Consistent with this, when a maternal GAL4 was used for expression of UAS-dome^DeltaTMCYT or UAS-dome^DeltaCYT at early stages of development, approximately 50% of the larvae acquired segmentation defects. The most frequent defects were deletions and fusions of A4 and A5 segments, the segments more sensitive to loss of JAK/STAT function, but stronger defects were also observed. These phenotypes are increased if the mothers are also heterozygous for a hop allele, further proving the central role of dome in JAK/STAT signaling (Brown, 2001).

In embryos lacking both maternal and zygotic dome function, the trachea does not develop. Because such an extreme tracheal defect has not been described for other mutants of the pathway, whether the trachea is similarly affected in stat92E germ line clones was examined. In zygotic and maternal stat92E mutants, the trachea is mostly absent. This cannot be a result of the abnormal segmentation because the trachea forms in paternally rescued embryos that still have segment defects. The tracheal system forms from ten pairs of tracheal pits arising in segments from T2-A8. The pits can be identified at early stages by the expression of three genes: trachealess (trh), ventralveinless (vvl; also known as drifter), and knirps (kni). These three genes are activated in the trachea independently of one another. trh and vvl are then required for the expression of all known tracheal genes (except kni) and for their own maintenance from stage 13 onward. To discover at what stage of tracheal development the JAK/STAT pathway is required, the expression of trh, kni, and vvl was studied in dome or stat92E maternal and zygotic mutants. In both types of mutant embryos, neither trh nor kni is expressed, whereas early vvl expression is not affected. Becasue trh is essential for tracheal development, its loss from the tracheal pits is the likely cause for the tracheal defects observed. These results show that STAT92E is the earliest transcription factor required for trachea specification (Brown, 2001).

The evidence presented here indicates that domeless encodes a receptor of the Drosophila JAK/STAT pathway and shows that the pathway is conserved in invertebrates. The previous failure to detect the receptor was due to sequence similarity being restricted to a few critical amino acids in otherwise quite common protein domains. Sequence comparison with vertebrate receptors reveals that the structure and sequence of Dome are most similar to those of CNTFR and LIFR (of the IL-6 receptor family) but that Dome also has some characteristics of the IL-3 receptor family. This suggests that the vertebrate family of receptors evolved from a single ancestral receptor that also gave rise to dome. The identical nature of the mutant phenotypes of dome, hop, and stat92E suggests that Dome transduces all extracellular signals activating HOP and STAT. Vertebrate IL-6 receptors generally function as heterodimers, whereas some receptors, such as CNTFR, lack the intracellular transducing domain and act exclusively as the ligand binding partner, recruiting the signal-transducing component of the receptor complex. It is interesting to note that although the Dome intracellular domain is essential for signal transduction, Dome is most similar to CNTFR. The Drosophila genome appears to have no other protein with significant similarity to other receptor members. Dome might therefore form homodimers that can function both as ligand binding and signal transducers. However, given the low sequence conservation between dome and the vertebrate receptors, the possibility that Dome may form a complex with another yet-unidentified partner cannot be discarded. Future biochemical experiments should confirm if, as with the vertebrate receptors, Drosophila JAK binds directly to Dome (Brown, 2001).

To study the interaction between Mom/Domeless and Unpaired, their cDNAs were subcloned into epitope-tagged mammalian expression vectors. 293T cells were cotransfected with V5-tagged Upd and HA-tagged Mom and expression was detected by immunofluorescence and Western blot with specific mouse monoclonal anti-tag antibodies. To examine the direct binding of Upd to Mom, 293T cells were transfected with Upd-V5. The ligand was released to the medium by treating the cells with heparin. Subsequently, the concentrated conditioned medium was applied to 293T cells nontransfected and transfected with HA-Mom and with a truncated form containing the N-terminal domain, Mom-N. Of note, given that there are no specific antibodies available for Mom, and that both anti-HA and anti-V5 antibodies have the same animal origin, (which prohibits double staining), an indirect approach was used to ascertain evidence of their presence in transfected cells. Since it is known that cells transfected with two DNAs will incorporate both at the same time, 293T cells were transfected with HA-Mom along with Stat92E. Therefore, cells stained with rabbit anti-STAT antibody should be those also expressing Mom. As expected, V5 staining was detected in cells containing Stat92E and transfected with Mom or Mom-N. These data show that Upd can be detected in 293T cells only when Mom is present, which indicates a physical interaction between these two molecules (Chen, 2002).

The presence of a putative signal sequence and five amino-linked glycosylation sites in Os suggests that it is post-translationally modified in the secretory pathway. To examine the possibility of amino-linked glycosylation, a mammalian expression plasmid containing os was transfected into human 293T cells. The cells were metabolically labeled with [35S]methionine in the presence or absence of tunicamycin, a potent inhibitor of amino-linked glycosylation. Os protein was recovered by immunoprecipitation, and the sizes of resultant Os proteins were compared. Several bands of 45-65 kD are seen in the untreated cells, whereas only the smallest 45-kD band is observed when the cells are treated with tunicamycin. The absence of the larger products in the treated cells suggests that the larger proteins are glycosylated forms. The presence of multiple forms of Os in the untreated cells most likely reflects partially glycosylated intermediates. These results suggest that Os contains a functional signal sequence that targets Os to the endoplasmic reticulum for glycosylation and secretion. Therefore, removal of the putative signal sequence should eliminate targeting of the Os protein to the endoplasmic reticulum and prevent glycosylation. As predicted, when the signal sequence was removed from Os, only the 45-kD band was observed (Harrison, 1998).

Comigration of the smallest Os protein species with a os product lacking a signal sequence suggests that the signal sequence is normally cleaved in the mature protein. To examine cleavage of the signal sequence, a plasmid was constructed in which a hemagglutinin (HA) epitope tag was fused to the amino terminus of Os. Previous work has shown that amino-terminal extensions on signal sequences do not abrogate in vivo function. The plasmid encoding the tagged Os was transfected into S2 cells and the resulting protein was precipitated with anti-Os antisera. The precipitated protein was then detected by Western blotting by use of either anti-Os or anti-HA antisera. The protein is detectable with the anti-Os serum, but not with the anti-HA antisera. These results suggest that the amino terminus of Os, including the HA tag, is proteolytically removed (Harrison, 1998).

The presence of modifications to Os that take place in the secretory pathway are consistent with the hypothesis that Os is a secreted protein. However, in initial experiments, Os could not be detected in the medium of transfected cells despite its high level of expression. Therefore, to determine the localization of Os, culture medium and extracellular matrix (ECM) were harvested separately from os-transfected S2 cells and tested for the presence of Os. To avoid contamination of the ECM fraction with cells and cell fragments, culture dishes were washed extensively prior to harvesting. The majority of Os protein is found to be associated specifically with the ECM, with only a small quantity free in the medium (Harrison, 1998).

Binding of proteins to the ECM is often mediated by glycosaminoglycans, such as heparan sulfate. Therefore, free heparin was added to the medium from os-transfected cells to determine whether heparan could prevent association of Os with the ECM. Addition of heparin releases nearly all of the Os protein into the medium. The ability of heparin to compete with ECM suggests that Os normally binds to the ECM through association with glycosaminoglycans. As expected, Os lacking a signal sequence is not secreted and cannot be detected in either the medium or ECM (Harrison, 1998).

Among the diverse cellular processes taking place during oogenesis, the delamination and migration of border cells (BCs), a group of anterior follicle cells, represent a powerful model to study cell invasion in a normal tissue. During stage 9 of oogenesis, BCs detach from the outer epithelium to invade the germline cyst compartment. The BC cluster contains two centrally located polar cells surrounded by approximately six outer border cells and undergoes a nearly 6-hour long posteriorward migration to reach the anterior part of the growing oocyte. Together with centripetal cells, they assemble the micropyle, a specialized structure required for sperm entry. domeless was isolated in a screen to identify genes essential in epithelial morphogenesis during oogenesis. The level of dome activity is critical for proper border cell migration and is controlled in part through a negative feedback loop. In addition to its essential role in border cells, dome is required in the germarium for the polarization of follicle cells during encapsulation of germline cells. In this process, dome controls the expression of the apical determinant Crumbs. In contrast to the ligand Upd, whose expression is limited to a pair of polar cells at both ends of the egg chamber, dome is expressed in all germline and follicle cells. However, Dome protein is specifically localized at apicolateral membranes and undergoes ligand-dependent internalization in the follicle cells. dome mutations interact genetically with JAK/STAT pathway genes in border cell migration and abolish the nuclear translocation of Stat92E in vivo. dome functions downstream of upd and both the extracellular and intracellular domains of Dome are required for JAK/STAT signaling. Altogether, the data indicate that Dome is an essential receptor molecule for Upd and JAK/STAT signaling during oogenesis (Ghiglione, 2002).

dome interacts genetically with the Stat92E and dpias ([a.k.a. Su(var)2-10 gene, Betz, 2001] a negative regulator of the JAK/STAT pathway) during BC migration, and dome phenotypes in ovaries are similar to those found in Stat92E and hop mutants. Furthermore, Stat92E nuclear localization is lost in dome mutant follicle cells, indicating that the mechanisms leading to Stat92E activation and subsequent nuclear translocation require dome. Since dome is epistatic to upd, the data indicate that dome is required downstream of upd and upstream of Stat92E for JAK/STAT signaling in egg chambers. Altogether, these results provide strong evidence that Dome is a receptor molecule for Upd during oogenesis (Ghiglione, 2002).

Dome is not uniformly distributed at the membrane but is restricted to apicolateral regions. Other receptor molecules have been shown to preferentially localize to apicolateral membranes, such as the EGF and Notch receptors, suggesting that the apical region is an active signaling interface for several receptors in follicle cells. Indeed, the apical localization of upd mRNA, membrane Dome and Dome-containing vesicles support a model in which ligand-receptor interactions take place apically in follicle cells, to activate the JAK/STAT pathway (Ghiglione, 2002).

Dome is a transmembrane protein with both extracellular and intracellular domains whose functions are unknown. The extracellular part contains a cytokine-binding module (CBM) and 3 fibronectin-type III domains likely participating in ligand binding, while the intracellular domain presumably interacts with Hop, through binding to one or several potentially phosphorylated tyrosines. Using truncated forms of Dome it has been shown that both the extracellular and intracellular domains are essential for BC migration and signal transduction. The dominant negative phenotypes that are observed are consistent with a model in which DomeDeltaCYT would titrate the ligand Upd, and DomeDeltaEXT titrates Hop, therefore inducing a dramatic reduction in signaling strength. Both constructs may also lead to the formation of non-functional Dome-Dome dimers by capturing the wild-type Dome protein in an inactive complex. Further biochemical work will be necessary to understand the molecular mechanisms underlying Dome signal transduction (Ghiglione, 2002).

The pattern of epithelial markers in dome mutant cells indicates that the JAK/STAT pathway is active in all follicle cells, a notion that is reinforced by the wide expression of nuclear Stat92E. How is Dome activated during egg chamber development and does this activation follow the same profile at all stages? Given the restricted pattern of upd expression in the egg chamber and its dramatic effect upon overexpression, it is unlikely that Upd is able to signal long distances in the follicular epithelium of late stage egg chambers. Rather, a model by which the JAK/STAT pathway plays a pre-patterning function is favored, acting early during egg chamber development to activate DE-cadherin and Crumbs expression. This view is consistent both with the expression pattern of upd and the distribution of Dome-containing vesicles described in this study. The formation of endogenous vesicles can be promoted by Upd, and a gradient of such vesicles is present around polar cells. Strikingly, these vesicles, which likely indicate active signaling through Dome, are widespread at early stages and become more restricted later on. It is proposed that during early development, the Upd signal produced by anterior and posterior polar cells contributes to the differentiation of all follicle cells. At this stage, Upd would be more diffusible than later, as suggested by the pattern of Dome intracellular vesicles. The study of the mechanisms controlling Dome activation and Upd activity will require additional tools to directly detect Upd, as, for example, Upd-GFP fusion proteins (Ghiglione, 2002).

This study has revealed several new findings about the function of dome and the JAK/STAT pathway during oogenesis. Future work will help to understand how Upd and Dome initially interact at the cell surface and transduce the signal to downstream JAK/STAT pathway members (Ghiglione, 2002).

It is commonly accepted that activation of most signalling pathways is induced by ligand receptor dimerization. This belief has been challenged for some vertebrate cytokine receptors of the JAK/STAT pathway. This study addresses the question of whether DOME, the Drosophila receptor of the JAK/STAT pathway, can dimerize and whether the dimerization is ligand-dependent. To analyze DOME homo-dimerization, a ß-gal complementation technique was applied that allows the detection of protein interactions in situ. This technique has been used in cell culture but this is the first time that it has been applied to whole embryos. This technique, which has been rename ßlue-ßlau technique, can be used to detect DOME homo-dimerization in Drosophila developing embryos. Despite DOME being ubiquitously expressed, dimerization is developmentally regulated. The state of DOME dimerization was investigated in the presence or absence of ligand; DOME dimerization is not ligand-induced, indicating that ligand independent cytokine receptor dimerization is a conserved feature across phyla. The functional significance of ligand-independent receptor dimerization was further analyzed by comparing the effects of ectopic ligand expression in cells in which the receptor is, or is not, dimerized. Ligand expression can only activate STAT downstream targets or affect embryo development in cells in which the receptor is dimerized. These results suggest a model in which ligand-independent dimerization of the JAK/STAT receptor confers cells with competence to activate the pathway prior to ligand reception. Thus, competence to induce the JAK/STAT signalling pathway in Drosophila can be regulated by controlling receptor dimerization prior to ligand binding. These results reveal a novel level of JAK/STAT signalling regulation that could also apply to vertebrates (Brown, 2003).

A sensitized genetic screen to identify novel regulators and components of the Drosophila janus kinase/signal transducer and activator of transcription pathway

The JAK/STAT pathway exerts pleiotropic effects on a wide range of developmental processes in Drosophila. Four key components have been identified: Unpaired, a secreted ligand; Domeless, a cytokine-like receptor; Hopscotch, a JAK kinase, and Stat92E, a STAT transcription factor. The identification of additional components and regulators of this pathway remains an important issue. To this end, a transgenic line was generated where the upd ligand was misexpressed in the developing Drosophila eye. GMR-upd transgenic animals have dramatically enlarged eye-imaginal discs and compound eyes that are normally patterned. The enlarged-eye phenotype is a result of an increase in cell number, and not cell volume, and arises from additional mitoses in larval eye discs. Thus, the GMR-upd line represents a system in which the proliferation and differentiation of eye precursor cells are separable. Removal of one copy of stat92E substantially reduces the enlarged-eye phenotype. An F1 deficiency screen was performed to identify dominant modifiers of the GMR-upd phenotype. Nine regions have been identified that enhance this eye phenotype and two specific enhancers: C-terminal binding protein and Daughters against dpp. Twenty regions have been identified that suppress GMR-upd and 13 specific suppressors: zeste-white 13, pineapple eye, Dichaete, histone 2A variant, headcase, plexus, kohtalo, crumbs, hedgehog, decapentaplegic, thickveins, saxophone, and Mothers against dpp (Bach, 2003).

These results indicate that Upd and the JAK/STAT pathway control the size of the Drosophila eye. Heteroallelic hypomorphic combinations of upd result in a small adult eye, while ectopic misexpression of upd in the developing fly eye results in a greatly enlarged eye. This phenotype is specific to activation of the JAK/STAT pathway in the developing eye because reduction in the dose of stat92E or the eye-specific transcription factor glass results in suppression of the enlarged eye. The results suggest that ectopic misexpression of upd in the developing eye results in additional mitoses of precursor cells in the region of the eye disc anterior to the furrow. These additional cells are patterned normally by the morphogenetic furrow, resulting in increased numbers of ommatidia in GMR-upd discs (Bach, 2003).

The enlarged-eye phenotype observed by ectopic misexpression of an activated form of ras85D using the ey enhancer, ey-ras^V12, is the result of ectopic R7 cells and also appears very rough. The results indicate that the GMR-upd phenotype is distinct from the ey-ras^V12 because GMR-upd eyes are patterned normally, are not rough, and are not modified by ras85D mutations. The enlarged eyes observed with misexpression of the Drosophila InR using GMR-Gal4 results primarily from increased cell volume. The results indicate that in the Drosophila eye the JAK/STAT and InR pathways do not interact, at least when ectopically misexpressed. Reduction in doses in InR pathway genes, such as InR, Pten, and chico, do not modify the GMR-upd phenotype. Moreover, the GMR-upd phenotype results from increased cell numbers, not from increased cell volume. In fact, cells in GMR-upd adult eyes actually exhibit decreased cell volumes when compared to wild type. Interestingly, the enlarged-eye phenotype in GMR-upd shares similarities with that produced as a nonautonomous effect of expression of an activated form of Notch (N^intra) in the eye, with prominent dorsal outgrowths. This observation is also interesting in light of the fact that CtBP, which represses N pathway activity, was identified as an enhancer of GMR-upd. It is possible that CtBP represses Stat92E itself or negatively regulates transcriptional coactivation by Stat92E (Bach, 2003).

The GMR-upd line was identified as a sensitized genetic background and an F₁ screen for dominant modifiers of the GMR-upd phenotype was performed using a set of overlapping deletions of the Drosophila genome. Twenty loci were identified that suppress and nine that enhance the enlarged-eye phenotype. The gene(s) in these deficiencies, responsible for the modification of the phenotype, may represent new components of or new interactors with the JAK/STAT pathway. Thirteen mutations were identified as Su(GMR-upd): zw13, crb, pie, D, His-2Av, kto, hdc, px, hh, dpp, tkv, sax, and Mad. In addition, two mutations were identified as En(GMR-upd): CtBP and Dad (Bach, 2003).

zw13 interacts genetically with the meiotic kinesin-like genes nod and ncd and encodes a poorly characterized protein with RNA-recognition motifs. Therefore, Zw13 may be important in regulating upd expression. crb was also identified as a suppressor of GMR-upd. Crb is a PDZ-containing protein involved in the establishment and maintenance of apical-basal polarity in epithelia. crb may suppress the GMR-upd phenotype by altering the localization of Dome and/or Upd or the signaling output of the JAK/STAT pathway in the eye (Bach, 2003).

Several transcription factors were identified as suppressors of GMR-upd: pie, D, His2Av, kto, px, and hdc. Pie is a nuclear protein that contains a PHD finger, which is a C4HC3 zinc-finger-like motif thought to facilitate chromatin-mediated transcriptional regulation. Eyes from pie homozygotes show irregular spacing of ommatidia, although the ommatidia have the normal array of photoreceptors. Notably, pie homozygous flies also have held-out wings, a phenotype shared by os flies and flies that overexpress full-length Dome. In embryonic segmentation, D directly regulates the expression of the pair-rule gene, even-skipped (eve), by binding to multiple sites located in downstream regulatory regions that direct formation of eve stripes 1, 4, 5, and 6. This overlaps with the function at Stat92E, which is needed for proper expression of eve stripes 3 and 5. Interestingly, fish and upd share related expression patterns and phenotypes. The early expression pattern of fish is almost identical to that of upd. Like upd, fish is also required in the hindgut, and the D held-out wing phenotype is very similar to that of os. His2Av belongs to the H2AZ variant subclass, which is involved in chromatin stability, chromatin remodeling, and transcriptional control. Given that mammalian STATs have been shown to mediate transcriptional changes within seconds of activation, it is possible that histone modification must be coordinated with transcriptional coactivation. Kto is the homolog of thyroid-hormone receptor associated protein (TRAP230), which was originally identified as part of the trithorax group, a large transcriptional coactivation complex. kto is involved in photoreceptor differentiation because homozygous mutant clones in the eye disc fail to develop into photoreceptors, although mutant cells can respond to Hh by expressing dpp. hdc encodes a nuclear factor involved in tracheal development, where it acts nonautonomously in an inhibitory signaling mechanism to determine the number of cells that will form unicellular sprouts in the trachea. Interestingly, it has been recently noted that stat92E is also required in tracheal development. However, whether hdc and stat92E interact, if at all, in this tissue is not known, nor is it understood whether any interaction exists in the eye disc. Px is a nuclear protein that, like Pie, contains a PHD zinc finger and is involved in venation in the wing. It is not known if px mutants exhibit an eye phenotype. Clearly, future work must focus on the elucidation of any biochemical interaction between Stat92E and these transcription/nuclear factors and also whether they regulate the transcription of a common set of genes required for growth of the eye disc (Bach, 2003).

The other modifiers identified in the modifier screen are genes in the Dpp pathway, specifically dpp, tkv, sax, mad, hh, and Dad. It was initially reasoned that upd may exerts its proliferative effects through hh or dpp. However, hh and dpp are expressed normally in GMR-upd. In addition, ectopic misexpression of hh or dpp in the os/os1A flies does not rescue the small-eye phenotype whereas upd does and ectopic expression of upd in flip-out clones does not induce hh. These results suggest that upd may not directly regulate dpp or hh expression. These data also suggest that Upd and Dpp and/or Hh may coregulate genes involved in the proliferation of eye precursor cells. This hypothesis is supported by observations in mammalian systems. The cytokines leukemic inhibitory factor and bone morphogenic protein 2 activate Stat3 and Smad1, respectively, and act synergistically in fetal neuroepithelial cultures to promote the differentiation of astrocytes from progenitor cells. The synergism requires functional Stat3 and Smad1. However, these proteins do not physically interact; rather, they both bind to p300/CBP to promote transactivation of target genes, such as glial fibrillary acidic protein, a marker of astrocyte differentiation (Bach, 2003).

In both mammals and flies, the JAK/STAT pathway plays an important role in the control of organ/tissue size. Stat5 knock-out mice are runted due to impaired growth-hormone signaling. Similarly, Socs-2 knock-out mice are significantly larger than their wild-type littermates, due to a lack of negative regulation of the growth-hormone pathway in vivo in the absence of the Socs-2 gene. Overexpression of an activated, constitutively dimerized STAT, c-Stat3, results in the formation of tumors in mice. Importantly, the only gain-of-function mutations in any JAK are found in Drosophila hop. hop^Tum-l and hop^T42 are independent point mutations that give rise to hyperactive Hop proteins, overproliferation and premature differentiation of Drosophila larval blood cells (a so-called fly 'leukemia'), melanotic tumors, and lethality. Overexpression of upd or hop in the developing Drosophila eye leads to a greatly enlarged eye due to an increase in the number of cells in the eye disc. In contrast, hypomorphic mutations in upd, for example, os or os/os1A, lead to a small adult eye (Bach, 2003).

Although proliferation is clearly a result of activation of the JAK/STAT pathway in mammals and Drosophila, little is known about how this pathway regulates the increase in cell number or the cell cycle. The data suggest that activation of the JAK/STAT pathway in the eye disc increases the number of cycling cells, possibly by shortening the G₁ phase or by regulating the G₂/M transition of the cell cycle. As a secreted molecule, Upd presumably acts in a cell-nonautonomous manner and may promote proliferation directly through activation of Hop and Stat92E. However, the observed proliferation in GMR-upd may in fact be due to the ability of Upd to induce another molecule that can also act cell nonautonomously. At the moment it is not possible to differentiate between these two possibilities. Nonetheless, the fact that more cells are observed in GMR-upd indicates that Upd may regulate genes involved in proliferation in the eye disc. In addition to the 15 modifiers of GMR-upd described here, several uncharacterized mutations have been identified that modify GMR-upd and may encode potentially novel molecules and uncover new functions of the JAK/STAT pathway. Given the high conservation between the Drosophila and mammalian JAK/STAT pathways, it is likely that the genes and functions uncovered in this screen will also be relevant to higher organisms (Bach, 2003).