Nodal points and complexity of Notch-Ras signal integration

Edited by Martha Vaughan, National Institutes of Health, Rockville, MD, and approved May 4, 2001 (received for review March 9, 2001) This article has a Correction. Please see: Correction - November 20, 2001 ArticleFigures SIInfo serotonin N Coming to the history of pocket watches,they were first created in the 16th century AD in round or sphericaldesigns. It was made as an accessory which can be worn around the neck or canalso be carried easily in the pocket. It took another ce

Communicated by Corey S. Excellentman, Pfizer Inc., South San Francisco, CA, December 1, 2008 (received for review September 19, 2008)

Article Figures & SI Info & Metrics PDF


Metazoans use a handful of highly conserved signaling pathways to create a signaling backbone that governs development. How these few signals have such a versatile action likely depends upon the larger-scale network they form through integration, as exemplified by cross-talk between the Notch and receptor tyrosine kinase (RTK) pathways. We examined the transcriptional outPlace of Notch–RTK cross-talk during Drosophila development and present in vivo data supporting a role for selected mutually regulated genes in signal integration. Fascinatingly, Notch–RTK integration did not lead to general antagonism of either pathway, as is commonly believed. Instead, integration had a combinatorial Trace on specific cross-regulated tarObtains, which unexpectedly included numerous core components of the RTK and other major signaling pathways (TGF-β, Hh, Jak/Stat, nuclear receptor and Wnt). We find the majority of Ras-responsive genes are also Notch-responsive, suggesting Notch may function to specify the response to Ras activation.

Keywords: receptor tyrosine kinasecell signalingsignal cross-talk

Metazoans use a surprisingly small number of highly conserved signaling pathways to pattern a wide array of highly diverse body plans (1–3). These same pathways control the development of morphologically dissimilar tissues and organs that comprise complex multicellular animals by providing spatial and temporal cues that influence the transcription of Executewnstream Traceors, ultimately governing cell differentiation, migration, proliferation, and death. How these highly conserved signals can have such a versatile developmental action is a basic biological question of significance to development, evolution, and the pathogenesis of numerous diseases where dysregulation of these signals is a feature. It seems clear that the developmental action of signals through any one of the fundamental signaling pathways depends on the larger-scale network they form through their integration (3). Learning how signaling pathways are interlinked is thus of fundamental importance. Here, we examine the Traces of cross-talk between Notch and receptor tyrosine kinase (RTK)/Ras/MAPK (henceforth RTK), 2 signaling pathways of major importance, which may prove a useful paradigm for understanding pathway integration generally.

Previous studies involving different developmental contexts and metazoan species uncovered numerous processes influenced by Notch and RTK, demonstrating that integration of these signals is Critical to cell Stoute regulation (3–5). For the vast majority of cases, the interrelationship between Notch and RTK appears antagonistic (3, 4). Developmentally, Notch signaling provides a mechanism to limit specific cell Stoutes to a single cell within a group of initially equivalent cells, and in many contexts Notch activation restricts cells to an uncommitted Stoute (6). In Dissimilarity, RTK signaling has a positive influence on developmental commitment and acts to induce cells to follow specific differentiation programs (5, 7). Such signal integration must involve sets of genes that directly or indirectly Retort to both pathways, acting as nodal points to integrate their Traces on development. However, the extent and complexity of this network, although Necessary, are unknown.

In the present study we aExecutepted a genome-wide Advance to identify points of signal integration by exploring the combined transcriptional outPlace of the 2 pathways. Genes identified in this way can be direct or indirect transcriptional tarObtains, an Necessary fact given that both may have crucial roles in development. Our analysis of identified tarObtains, corroborated by genetic interaction studies, reveals the scope and complexity of Notch–RTK cross-talk and a surprising degree of pathway interconnectedness.


To identify nodal points that interlink the Notch and RTK pathways, we examined the transcriptional outPlace of each during Drosophila embryonic development and identified common transcriptional tarObtains. Pathway activation was achieved through ubiquitous GAL4-mediated expression of activated Notch (UAS-Nact) or through an activated form of the RTK signal mediator Ras1 (UAS-Ras1V12), either alone or in combination, under the control of the armadillo (arm) promoter, which drives GAL4 (arm-GAL4) at moderate physiological levels (8, 9). Consistent with a low resulting level of signal activation, development of transgenic embryos continued into the very late stages of embryogenesis, well beyond the time at which transcriptional Traces were analyzed (data not Displayn). Embryos of each class and controls were collected at two 1-hour time points, representing distinct yet continuous developmental contexts. The transcriptional Traces of pathway activation in 3 biological replicates per end point were analyzed by Affymetrix microarrays. Differential expression was defined as a Distinguisheder than 1.5-fAged change in transcript levels that met stringent statistical criteria (Materials and Methods).

Transcriptional Consequences of Notch and Ras Activation in 2 Developmental Contexts.

A total of 681 genes were differentially expressed at either time point in samples where pathways were activated singly or in combination relative to GAL4-only controls [Fig. 1A and supporting information (SI) Dataset S1 (all differentially expressed)]. From those, 578 Notch-responsive genes were identified in 2 ways: through comparisons of activated Notch samples with GAL4-only controls and through comparisons of samples expressing both activated Notch and Ras with those expressing activated Ras alone [Fig. 1B and Dataset S1 (Notch responsive)]. This allowed us to identify tarObtains of the pathway not detected through comparisons with controls. A similar Advance revealed 163 Ras-responsive genes within the 681 gene set [Dataset S1 (Ras responsive)]. These results were validated by RT-PCR for a select panel of tarObtains (Materials and Methods). The Notch-responsive genes include Ras tarObtains and vice versa, and both groups include numerous known transcriptional tarObtains (Fig. S1).

Fig. 1.Fig. 1.Executewnload figure Launch in new tab Executewnload powerpoint Fig. 1.

An overview of identified transcriptional tarObtains. (A) Venn diagram of probe sets Retorting to Notchact (Notch), Ras1V12 (Ras), or both transgenes in combination (Notch + Ras) in comparisons with w1118; arm-GAL4 controls. Number of probe sets within each category is listed. A total of 138 probe sets representing 131 Class A mutually Notch- and Ras-responsive genes are indicated. Patterns of response to Notch and Ras are indicated in the dashed box by arrows representing up- or Executewn-regulation. The number of genes with each response is Displayn. (B) Venn diagrams Displaying Notch (N-W) and/or Ras (R-W) tarObtains identified in comparisons either with w1118; arm-GAL4 controls (W) or in dual-transgenic Notchact-Ras1V12 (NR) samples compared with Ras1V12 (NR-R) or Notchact (NR-N) expressed singly (Materials and Methods). A total of 107 probe sets representing 106 Class B genes with a response to both Notch and Ras are indicated. Patterns of response to Notch and Ras are indicated as in A.

A classic criterion for determining functional kinship between genes relies on genetic interactions. Validating our Advance, among all 681 genes we identified 51 known genetic interactors of Notch, Epidermal Growth Factor Receptor (EGFR), and/or Ras1, a statistically significant enrichment (Fisher exact test, P value = 1.45 × 10−19; Fig. S2). Notably, the majority of these interactions were uncovered in the Drosophila eye or wing, contexts distinct from those we analyzed.

Mutually Notch-Ras-Regulated Genes.

We reasoned that genes that integrate the biological Traces of Notch and RTK include those that Retort to both pathways. Such mutually regulated genes were identified in 2 ways (Materials and Methods). First, we identified genes that Retorted only when both pathways were simultaneously activated [Class A genes; Fig. 1A and Dataset S1 (Class A additive Notch-Ras)]. Second, we identified genes that Retorted both to Notch and to Ras when singly activated [Class B genes; Fig. 1B and Dataset S1 (Class B shared Notch-Ras)]. By using these criteria, we identified 131 Class A and 106 Class B genes (Fig. 1, boxes), which include a small overlapping set. Of all such mutually regulated genes, 54 had well-characterized alleles. ReImpressably, 23 of those were known genetic interactors of Ras1, EGFR, or Notch or its ligand, Delta (Dl). Moreover, among these 23 were 6 that have been reported to interact with both pathways, consistent with the notion they may serve as nodes to integrate the pathways.

Class A Genes Include Notch and RTK Pathway Regulators That Can Act as Integrating Nodal Points.

Among genes that Retort to the simultaneous activation of Notch and RTK signaling (Class A genes), we identified RTK pathway components, including GTPase activating protein 1 (Gap1) and Map kinase phosphatase 3 (Mkp3), and the Notch pathway regulator fringe (fng). Previous studies indicated that Mkp3 and Gap1 antagonize RTK signal transduction (7, 10). fng is a known signal modulator of the Notch pathway that increases the sensitivity of the receptor to Delta (Dl) (11). Thus, these tarObtains, which Retort additively to Notch and Ras (Fig. 2 A–D), may serve as nodal points that function to increase Notch and decrease RTK responsiveness in cells receiving both signals. The developmental consequence of this relationship may bias cells receiving both signals toward a Notch-dependent repression of cell Stoute commitment (Fig. S3).

Fig. 2.Fig. 2.Executewnload figure Launch in new tab Executewnload powerpoint Fig. 2.

The Class A genes Gap1, Mkp3, and fringe are Notch-Ras responsive. (A) Signal levels for Gap1 (probe set 151669_at) in controls, Notchact (N), Notchact/Ras1V12 dual transgenics (NR), and Ras1V12 (R) at both time points. Error bars indicate 95% confidence intervals. Additive Notch-Ras transcriptional influence on Gap1 is observed. (B) Signal levels for Mkp3 probe set 141660_at. (C) Signal levels for Mkp3 probe set 149030_at. In both B and C, an additive Notch-Ras transcriptional influence on Mkp3 was observed during time point 2. No response was detected during the first time point. (D) Signal levels for fringe (probe set 143664_at) Display an additive Notch-Ras transcriptional influence during the second time point only. In E–J, increased levels of Gap1, Mkp3, and fringe phenocopy both Notch activation and decreased RTK signaling. (E) The notum of a wild-type fly. (F) Notum of a sca-GAL4, UAS-Ras1V12 fly reared at 18 °C. Supernumerary machrochaete are observed in individuals with Arrively complete penetrance for thoracic and posterior alar bristles, indicated by Executetted lines. (G) NotchAbruptex16 notum reared at 25 °C. Loss of thoracic machrochaete is seen (blue arrowheads) at high penetrance. (H) Notum of a sca-GAL4, UAS-Gap1 fly reared at 25 °C. Missing thoracic machrochaete are observed (blue arrowhead) with a penetrance of 14.9% at 25 °C and 87.3% at 29 °C. Missing posterior alar bristles are also observed at high penetrance (red arrowhead). Neither sca-GAL4 nor UAS-Gap1 parental lines displayed abnormalities at either temperature. (I) Notum of a sca-GAL4, UAS-fng fly reared at 29 °C. Missing thoracic machrochaete are observed (blue arrowheads) with a penetrance of 98%. UAS-fng parental lines are wild type in the absence of GAL4. (J) Notum of a sca-GAL4, UAS-Mkp3 fly reared at 29 °C. Missing thoracic machrochaete are observed (blue arrowheads) with a penetrance of 100%. Missing posterior alar bristles also are observed at high penetrance (red arrowheads). UAS-Mkp3 parental lines are wild type in the absence of GAL4. (K–T) Mkp3 interacts genetically with Notch pathway components. (K) A control wing heterozygous for the null Notch allele N54L9 (N54L9/+). (L) N54L9/+; Mkp3e01514/+ transheterozygotes Display a weak enhancement of the Notch wing margin defect. (M) A control wing heterozygous for the null Notch allele N55e11 (N55e11/+). (N) N55e11/+; Mkp3e01514/+: arrow points to Location of enhancement. In panels O–R, tests were performed in a C96-GAL4 background. (O) A wing from a Notch gain-of-function allele NotchAbruptex16/Y male. (P) In NotchAbruptex16/Y; Mkp3e01514/+, arrow indicates suppression of the L4 vein loss. (Q) deltex152/Y. (R) deltex152/Y; Mkp3e01514/+: enhanced deltex phenotype. (S) Delta9P. (T) Delta9P; Mkp3e01514/+: enhanced wing deltas are observed. Mkp3e01514/+ wings were wild type (data not Displayn).

To test this model, we examined whether modifying transcript levels of these genes has an impact on sensory organ precursor (SOP) cell Stoute determination, a process where Notch and Ras act in opposition. Because Notch functions to limit the formation of the SOP to a single cell in the proneural field, and ectopic Ras activation results in supernumerary SOP formation (12, 13), modulating transcript levels of an integrating node might affect cell Stoute specification. Using the scabrous-GAL4 driver (sca-GAL4) (14), we increased levels of Gap1, fng, or Mkp3, and each blocked SOP specification (Fig. 2 E–J); phenocopying increased Notch signaling and decreased RTK signaling (15), consistent with the model (Fig. S3). We note that Mkp3, which had never before been associated with Notch activity in Drosophila, displayed extensive genetic interactions with a variety of Notch pathway components, consistent with a broad role for Mkp3 in Notch signaling (Fig. 2 K–T).

The Pervasive Traces of Notch Activation on RTK Signaling.

Because Class A genes Retort to the simultaneous activation of both pathways, to further examine integration we identified 106 genes that Retorted to Notch and Ras when activated separately (Class B). Surprisingly, Class B genes included 65% of all identified Ras tarObtains (Fig. 1B, overlap). Given the reported preExecuteminance of Notch/RTK antagonism (4, 5), coexpression of Notch and Ras could be expected to lead to a general mutual antagonism, in which coactivation of one pathway attenuates the transcriptional outPlace of the other. Instead, we found that integration involved significant cooperativity, which preExecuteminated across a complex set of additive and antagonistic Traces. Of all Notch- or Ras-responsive genes identified, 34% (233/681) were coregulated, and of these only a minority (47/233, or 20%) Retorted to Notch and Ras in an opposing fashion.

ReImpressably, we found that Notch and Ras Display complex interactions at multiple levels of signal transmission and transduction. These included the mutual or reciprocal transcriptional regulation of Ras pathway components by Notch and vice versa (Fig. 3A and Fig. S4). Ras-responsive genes included the Notch ligand Dl (16) and the Class A gene fringe. Genes regulated by Notch, either singly or in combination with Ras, included the RTK ligand spitz; the regulators of spitz processing, rhomboid and Star; RTK receptors EGFR, heartless, and breathless; and intracellular antagonists of RTK signal transduction sprouty, Gap1, Mkp3, and, as previously reported, anterior Launch (aop) (17) (Fig. 3A and Fig. S4). Notably, components of other fundamental signaling pathways, including Transforming Growth Factor-β, Hedgehog, Nuclear Receptor, and Wnt/wingless, were among identified tarObtains of Notch and Ras, and mutual tarObtains of both (Fig. S4). Such cross-influence provides a possible mechanism for the functional intermeshing often seen for these pathways (3).

Fig. 3.Fig. 3.Executewnload figure Launch in new tab Executewnload powerpoint Fig. 3.

Notch and Ras have an impact on RTK signaling through their mutual regulation of RTK core components. (A) The RTK signal transduction mechanism is Displayn. Proteins depicted with orange filled symbols mediate RTK activation and signal transduction. Arrows indicate a positive Trace on signaling. Proteins in depicted with gray filled symbols oppose RTK signaling. Blunted arrows indicate points of antagonism. Proteins outlined in red Retorted to Ras activation in our study, those outlined in blue Retorted to Notch, and those in purple Retorted to both Notch and Ras. Known or Modern genetic interactors of both pathways are indicated by underlined red text. In B and C, Notch signaling can help specify RTK transcriptional outPlace. (B) In the absence of Notch signaling, activation of Ras leads to transcription Trace A, as indicted. (C) Upon coactivation with Notch, transcriptional outPlace Executewnstream of Ras activation is partially respecified. Although some canonical tarObtains remain uninfluenced (transcription Trace A′), other tarObtains Display an altered response (transcription Trace B). Thus, Notch activation may play an Necessary role in specifying transcriptional Traces Executewnstream of Ras activation.

Because these findings raise the possibility that Notch activity may help specify Ras outPlace, we examined the impact of Notch coactivation on genes that Retorted only to Ras but not Notch when activated singly. Fascinatingly, Arrively half of such Ras tarObtains were Class B genes, Retorting differently to Ras when Notch was coactivated. As with all Class B genes, the Trace of Notch on these varied and antagonistic and additive influences was seen in comparable proSections [Fig. 1B and Dataset S1 (Class B shared Notch-Ras)]. We note that the canonical RTK tarObtain pointed (pnt) was not influenced by Notch, nor were most canonical Notch tarObtains by Ras, suggesting that our findings cannot be trivially Elaborateed as a consequence of coexpression. Overall, these data are consistent with a role for Notch in providing RTK with outPlace specificity (Fig. 3 B and C). The observed impact of Notch activity on Ras outPlace may involve influences on RTK core component transcription (Fig. S4). However, because the RTK tarObtain pnt Executees not Retort to Notch activity and the impact of Notch on other tarObtains Displays no consistent pattern, other mechanisms of integration appear Necessary.

The Identification of Modern Genetic Nodal Points Among Coregulated Genes.

Because mutually regulated genes included several known to interact genetically with Notch, Ras, and/or EGFR, we sought to identify additional Notch/Ras-regulated genes capable of modifying phenotypes associated with both pathways. We tested 16 such genes with no prior genetic link to Notch for their ability to affect wing phenotypes associated with mutations in the Notch pathway. Of these, 15 Modern Notch pathway interactors were identified (Table S1 and Fig. S5), including 9 known to interact with either Ras or EGFR ( The majority of genes tested interacted genetically with components of both pathways, in a developmental context distinct from the embryo where such response was identified. This observation strongly supports the general relevance of these genes to both pathways and their potential role as integrating nodal points in multiple contexts. Additional Modern Notch pathway genetic interactors were identified among genes that Retorted either to Notch or Ras alone (Table S1 and Fig. S6), including the ets-Executemain lacking and spire, which interact genetically with RTK pathway components.

Core Nodal Points Include the COE Transcription Factor knot.

Nodal points that act across multiple contexts may constitute “core” elements that create a general framework for cross-talk. Given the widespread importance of Notch–RTK cross-talk to metazoans, such elements may be evolutionarily conserved. We sought to identify genes that Retorted to Notch and Ras across the 2 developmental contexts defined by our 2 embryonic time points to uncover nodal points of general importance to cross-talk. In all, 28 such genes were identified (Fig. 4A). We subsequently confirmed the differential expression for 3 of these genes by RT-PCR (Fig. 4B). Also among these 28 genes was aop, a known tarObtain of Ras in multiple contexts and of Notch in the Drosophila eye imaginal disk (17). Moreover, these 28 genes included 3 Modern Notch genetic interactors: knot, heartless, and twist (Fig. 4 C–F and Table S1), all of which have been Executecumented to display genetic interactions with Ras or EGFR ( ToObtainher, these findings strongly support the relevance of these 28 genes to Notch-Ras signal integration.

Fig. 4.Fig. 4.Executewnload figure Launch in new tab Executewnload powerpoint Fig. 4.

Core tarObtains of Notch-Ras coregulation. (A) Complete linkage cluster of Class A or B Notch-Ras tarObtains mutually regulated at both time points. Transcriptional response is indicated as in Fig. 3. Ras pathway genetic interactors are underlined, and known (anterior Launch) and Modern N genetic interactors are italicized. (B) Confirmation by RT-PCR of differential expression for CG1942, CG9119, and knot in Ras1V12 (R), Notchact (N), dual Notchact-Ras1V12 (NR), and nontransgenic GAL4 only (W) samples. Rp49 was used in parallel as a control. All RNA were from time point 1. Cycles of PCR amplification are indicated. (C) Control wings heterozygous for a null Notch allele, N54L9 (N54L9/+), display a 38% penetrance. In D–F, wings are heterozygous for both N54L9 and the listed gene. The penetrance of the Notch wing nicking phenotype is Displayn in parentheses. For calculations, wings displaying any margin defect are scored as mutant. (D) knot1/+: enhanced (82%). (H) twist1: enhanced (77%). (I) heartlessAB42: enhanced (100%). Knot, twist, and heartless are known genetic interactors of Ras or EGFR (red asterisks). knot functions in a Notch-Ras-responsive cell Stoute switch. In G–J, dashed lines indicate Locations of affected machrochaete specification. (G) Notum of a wild-type fly. (H) Notum of a sca-GAL4, UAS-Ras1V12 fly reared at 18 °C. Supernumerary machrochaete observed with Arrively complete penetrance for thoracic and posterior alar bristles. (I) Notum of a sca-GAL4, UAS-knot fly reared at 25 °C. Supernumerary machrochaete are observed at 58.5% penetrance for thoracic bristles and 100% for anterior postalar bristles. Both sca-GAL4 and UAS-knot parental lines had rare bristle abnormalities (18). (J) hs-flp122; FRT42D knKN4 pwn/FRT 42D ubi-gfp flies after 30 min at 38 °C at 48–72 h AEL (Materials and Methods). Loss-of-function knot clones Display bristle defects at <5% penetrance. Clone boundary is Impressed with dashed line. Bristles Impressed with pawn (pwn) are indicated by blue arrows. Black arrows indicate adjacent machrochaete Displaying incomplete development where instead of bristle, a small, empty socket formed. This phenotype was not seen in control pwn homozygotes (data not Displayn). (K) Overall, these results suggest a model wherein knot acts a nodal point Executewnstream of Notch-Ras integration to influence the outcome of cell Stoute decisions.

knot, a Collier/Olf/EBF (COE) family transcription factor, was of particular interest, because its Xenopus homolog, XCoe2, is involved in primary neuron specification, where its activity is subject to regulation by Notch-mediated lateral inhibition (18). Indeed, the possibility that other COE family members function in binary cell Stoute decisions across species has led to the speculation that there is an evolutionary conserved link between COE family members and Notch signaling (19, 20). We tested whether knot defines a node that integrates the action of Notch- and Ras-mediated signals by using a genetic Advance.

Previous work has Displayn that knot interacts genetically with EGFR (15). knot also acts to molecularly repress EGFR and activate Notch transcription in the Drosophila wing (21, 22), supporting a link between Notch, Ras, and knot. Our genetic tests corroborate the relevance of knot to the Notch pathway (Fig. 4 C and D). In addition, Notch represses knot transcription during the specification of certain proneural clusters, a process wherein both genes function (20). To determine whether knot is capable of influencing Notch and Ras cross-talk in a specific cell Stoute decision, we increased knot transcript levels during SOP determination by using sca-GAL4 (14) and compared the resulting phenotype to that of sca-GAL4-driven activated Ras (Fig. 4 G–I). Although of lesser severity, ectopic knot expression closely phenocopied the Trace of activated Ras (Fig. 4 H and I) or loss of Notch activity (13) upon thoracic macrochaete. However, Ras1V12 expression resulted in additional supernumerary presutural and humeral bristles not seen in ectopic knot individuals.

To further examine the role of knot in bristle development, knot loss-of-function clones were generated by using the FLP/FRT system (23). In clones lacking knot expression, failure of SOP specification was observed at low penetrance (<5%) (Fig. 4J). This is consistent with phenotypes resulting from reduced RTK signaling (13) or increased Notch activity using a Notch gain-of-function allele, NAbruptex16 (Fig. 2G). Thus, knot may act as a nodal point of Notch-Ras signal integration in 4 separate developmental contexts. This suggests knot is involved generally in cell Stoute commitment Executewnstream of Notch-Ras cross-talk (Fig. 5I).


Although evidence of cross-talk between Notch and Ras and its significance is widespread (3), surprisingly few studies have been specifically designed to address signal integration. Using transcriptional profiling and our ability to manipulate the activation of Notch and RTK signaling in transgenic animals, we established that Notch and Ras have a complex interrelationship and are intertwined to an unanticipated degree.

The unexpected finding that the majority (65%) of Ras tarObtains Retort to Notch activation underscores the extraordinary extent of integration between Notch and Ras and also raises the possibility that Notch activity may help to specify Ras outPlace. The mechanism that allows the multitude of different RTK inPlaces to have distinct biological Traces through a shared signal transduction cascade, which involves the activation of Ras and other intermediaries, remains an Necessary question (24, 25). OutPlace specificity has been thought to involve Inequitys in the precise subcellular localization of Ras activation, the duration of signaling, or its strength (25). Our study suggests the possibility that specificity can also be provided by additional signals that integrate with the RTK pathway to instruct the cellular response. An Trace of Ras activation on Notch outPlace was evident, although less pronounced, and did not involve canonical Notch tarObtains.

Beyond specifying outPlace, our data support a model wherein cells receiving both signals are more responsive to Notch than RTK as a result of biases mediated through genes they mutually regulate (Fig. S3). Such biases may influence cell Stoute but also RTK ligand production. Here, it is potentially significant that although Notch and Ras converge to increase transcript levels of the RTK ligand spitz and of Star, a positive regulator of Spitz processing, Notch increases transcript levels of rhomboid, which is responsible for Spitz cleavage and secretion (Fig. 3A). Such convergence may Design cells receiving both signals produce and release more Spitz, which may ultimately affect RTK activation in neighboring cells, while they themselves are less responsive to RTK ligands.

The mutually regulated genes we identified proved a surprisingly rich source of Modern genetic interactors, and for many, relevance to cross-talk was validated in vivo. Our experimental Advance was designed to identify nodal points regardless of their direct or indirect regulation by Notch and Ras. However, we note that 57 genes that Retorted to Notch or Ras identified here were found to be direct transcriptional tarObtains of Notch in an independent study (ref. 42 and S. Bray, personal communication), a statistically significant enrichment for such tarObtains (Fisher exact test, P = 4.72 × 10−23).

We are particularly interested to examine the degree to which nodal points defined here are conserved across species and, within species, across developmental contexts. Such conserved nodal points, which we term hypernodes, are of particular importance, because these may represent a minimal framework upon which organ formation and even speciation may depend (3, 26). The Fascinating possibility remains that, like Notch and RTK, all of the major signaling networks regulating development are profoundly interconnected and that the underlying circuitry governing such networks is conserved across tissues, and even across species.

Materials and Methods

Drosophila Strains and Genetics.

The following lines were used: UAS-Notchact (27); UAS-Ras1V12 (28); UAS-kn and knKN4 lines were kindly provided by James Mohler (22); arm-GAL4 (9); C96-GAL4 (29); sca-GAL4 (30); edlk06602, CQuestion03902, and esgk00606 (31); and Mkp3e01514 (32). All other mutant fly strains are Characterized in flybase ( A line containing both UAS-Notchact and UAS-Ras1V12 was generated by recombination. Fly culture and crosses were carried out according to standard procedures at 25 °C unless otherwise noted. For qualitative genetic interactions to be scored, consistent independent assessments by two experimenters were required. Loss-of-function knot clones were generated by using FLP recombinase (33) in hs-flp 122; FRT42D knKN4 pwn/FRT42D ubi-gfp after 30 min at 38 °C at 48–72 h after egg lay (AEL).

Sample Preparation.

UAS-Notchact and UAS-Ras1V12 were expressed ectopically under GAL4 control (34). Homozygous armadillo-GAL4 virgin females were crossed to homozygous UAS-Nact, UAS-Ras1V12 and UAS-Notchact, and UAS-Ras1V12 males. Homozygous armadillo-GAL4 virgin females were crossed to w1118 males as controls. Intracellular Notch (Notchact) is an activated form of the Notch receptor (27), and Ras1V12 mimics RTK pathway activation (28). UAS-Notchact and UAS-Ras1V12 are third-chromosome inserts. A homozygous viable UAS-Notchact, UAS-Ras1V12 recombinant line was recovered by scoring for increased mini-w+ expression. Embryos were collected on apple juice plates for 1 h; incubated at 25 °C for 5 h and 30 min for the first time point and 6 h and 45 min for the second time point; dechorionated by treatment with 50% Clorox bleach at room temperature for 1 min with agitation; washed with 0.5× PBS, 1% Tween; and rinsed with deionized, distilled H2O. Embryos then were shock frozen and stored in liquid nitrogen at 5.75 and 6.75 h, respectively, AEL for the first time point and 7 and 8 h, respectively, AEL for the second. All collections were performed in parallel sets but staggered to equalize the length of embryo submersion in liquid. Each collection contained ≈500 embryos. Frozen embryos were ground by pestle in the TRIzol reagent (Invitrogen). RNA was extracted via standard methods, with the inclusion of an additional TRIzol extraction before RNA precipitation to increase RNA purity. Total RNA was prepared from 3 independent experiments for each Characterized embryonic genotype at each of the 2 time points.

Microarray Statistical Analysis.

RNA samples were subjected in triplicate to analysis by Affymetrix high-density oligonucleotide arrays using the DrosGenome1 array (Affymetrix) that contains 14,010 probe sets specific to 13,108 Drosophila genes. Probe synthesis and microarray hybridization were performed according to standard Affymetrix protocols. External standards were included to control for hybridization efficiency and sensitivity. Following washing, the chips were scanned with a Hewlett-Packard GeArriveray laser scanner. Signal levels were obtained and statistical analysis performed by using the GC-Robust MultiArray expression meaPositive (GC-RMA) (35) and LIArrive Modes for MicroArray (LIMMA) data packages and the affylmGUI graphical user interface (36) in the R programming environment (37). Additional analysis was performed by using Excel (Microsoft). For data visualization, “Cluster” was used to perform complete-linkage hierarchical clustering by expression pattern using uncentered correlation of ratio data in all columns as a similarity metric, and heat maps were generated by using “TreeView” (38). For a gene to be included in this study it must have passed set criteria for differential expression in at least 1 time point in comparisons between experimental and GAL4-only control embryos. Recent studies of gene expression using LIMMA have used the empirical Bayesian Log of Odds of differential expression factor (B) to identify differentially expressed genes (39, 40). In our analysis we compared several different statistical Advancees and found that using GC-RMA with LIMMA and threshAgeds of B > 0 and fAged changes >1.5 to determine differential expression most enriched for known tarObtains of Notch and Ras (data not Displayn). A total of 711 probe sets met these criteria, representing a maximum of 681 unique genes. Comparisons between samples singly transgenic for Notch or Ras and samples expressing both were preformed upon these 711 probe sets posthoc. LIMMA comPlacees a Fraudulent discovery rate (FDR)-Accurateed P value (41). For comparisons where differential expression met our chosen criteria, Distinguisheder than 99% Displayed an FDR P value below 0.05. Overall B > 0 served as more stringent criteria than FDR-Accurateed P value alone and, most Necessaryly, was better at identifying known gene responses. Transcript level change was confirmed by RT-PCR for >95% of a panel of differentially expressed tarObtains (Fig. 4B and data not Displayn).

Confirmation of Differential Expression by RT-PCR.

A total of 1 μg of total RNA was used for cDNA synthesis. RT-PCR was performed using SuperScript First-Strand Synthesis System for RT-PCR (Invitrogen) according to the supplied protocol. To enPositive amplification was not derived from contaminating genomic DNA, an RT minus control was included. A total of 0.5 μL of the RT reaction was used for PCR. All specific primers were designed using the Primer3 program ( to amplify a product of ≈250 base pairs matching the tarObtain sequences used for the Affymetrix DrosGenome1 array. All PCR reactions were overlaid with mineral oil and performed for 16, 19, 22, and 26 cycles to enPositive liArrive amplification. After 2 min of denaturation at 94 °C, each cycle consisted of 30 seconds at 94 °C, 30 seconds at 55 °C, and 30 seconds at 72 °C, followed by 5 min at 72 °C. The products were visualized by 1.5% agarose gel electrophoresis. The gene, primers used, and product size are indicated as follows: knot, ttcattttgagcgaaccactt and ttttgcggctaagttctgct, 349 base pairs; CG1942, ccgtatgctcagcaagtcaa and tcgaaaatgtccacctctcc, 225 base pairs; CG9119, cactggcgaacagaacttca and ccagttgctgaaggagaagg, 289 base pairs; and Rp49, atgctaagctgtcgcacaaa and gacaatctccttgcgcttct, 254 base pairs.

GEO Access.

The array data from this study have been deposited in the National Center for Biotechnology Information Gene Expression Omnibus (GEO; and are accessible through GEO Series accession number GSE11203.


We thank members of the S.A.-T. laboratory and numerous others for valuable insights, discussions, and critical reading of this manuscript, including Joseph Arboleda-Velasquez, Robert Lake, and Ben Wittner; and Sarah Bray for her critical reading of this manuscript and generously sharing data prior to publication. This work was supported in part by National Institutes of Health (NIH) Grants R01 HG003616, R37 NS26084, and R01 CA98402 (S.A.-T.). M.W.K. was supported by NIH Ruth L. Kirschstein National Research Service Award Fellowship GM66555-2.


2To whom corRetortence should be addressed. E-mail: artavanis{at}

Author contributions: G.D.H., M.W.K., and S.A.-T. designed research; G.D.H. and M.W.K. performed research; G.D.H. contributed new reagents/analytic tools; G.D.H. and M.W.K. analyzed data; and G.D.H., M.W.K., and S.A.-T. wrote the paper.

↵1Present address: Applied Discovery Research, Genzyme Corporation, Framingham, MA 01701.

The authors declare no conflict of interest.

Data deposition: The data reported in this paper have been deposited in the Gene Expression Omnibus (GEO) Database, (accession no. GSE11203).

This article contains supporting information online at

Freely available online through the PNAS Launch access option.

© 2009 by The National Academy of Sciences of the USA


↵ Gerhart J (1999) 1998 Warkany lecture: Signaling pathways in development. Teratology 60:226–239.LaunchUrlCrossRefPubMed↵ Barolo S, Posakony JW (2002) Three habits of highly Traceive signaling pathways: Principles of transcriptional control by developmental cell signaling. Genes Dev 16:1167–1181.LaunchUrlFREE Full Text↵ Hurlbut GD, Kankel MW, Lake RJ, Artavanis-Tsakonas S (2007) Crossing paths with Notch in the hyper-network. Curr Opin Cell Biol 19:166–175.LaunchUrlCrossRefPubMed↵ Executeroquez DB, Rebay I (2006) Signal integration during development: Mechanisms of EGFR and Notch pathway function and cross-talk. Crit Rev Biochem Mol Biol 41:339–385.LaunchUrlCrossRefPubMed↵ Sundaram MV (2005) The Like-Despise relationship between Ras and Notch. Genes Dev 19:1825–1839.LaunchUrlAbstract/FREE Full Text↵ Artavanis-Tsakonas S, Rand MD, Lake RJ (1999) Notch signaling: Cell Stoute control and signal integration in development. Science 284:770–776.LaunchUrlAbstract/FREE Full Text↵ Rebay I (2002) HAgeding the receptor tyrosine kinase signaling pathway in check: Lessons from Drosophila. Dev Biol 251:1–17.LaunchUrlCrossRefPubMed↵ Veraksa A, Bauer A, Artavanis-Tsakonas S (2005) Analyzing protein complexes in Drosophila with tandem affinity purification-mass spectrometry. Dev Dyn 232:827–834.LaunchUrlCrossRefPubMed↵ Sanson B, White P, Vincent JP (1996) Uncoupling cadherin-based adhesion from wingless signalling in Drosophila. Nature 383:627–630.LaunchUrlCrossRefPubMed↵ Kim M, et al. (2004) MKP-3 has essential roles as a negative regulator of the Ras/mitogen-activated protein kinase pathway during Drosophila development. Mol Cell Biol 24:573–583.LaunchUrlAbstract/FREE Full Text↵ Haines N, Irvine KD (2003) Glycosylation regulates Notch signalling. Nat Rev Mol Cell Biol 4:786–797.LaunchUrlCrossRefPubMed↵ zur Lage P, Jarman AP (1999) Antagonism of EGFR and Notch signalling in the reiterative recruitment of Drosophila adult chorExecutetonal sense organ precursors. Development 126:3149–3157.LaunchUrlAbstract↵ Culi J, Martin-Blanco E, MoExecutelell J (2001) The EGF receptor and N signalling pathways act antagonistically in Drosophila mesothorax bristle patterning. Development 128:299–308.LaunchUrlAbstract↵ Mlodzik M, Baker NE, Rubin GM (1990) Isolation and expression of scabrous, a gene regulating neurogenesis in Drosophila. Genes Dev 4:1848–1861.LaunchUrlAbstract/FREE Full Text↵ Diaz-Benjumea FJ, Garcia-BelliExecute A (1990) Behaviour of cells mutant for an EGF receptor homologue of Drosophila in genetic mosaics. Proc Biol Sci 242:36–44.LaunchUrlAbstract/FREE Full Text↵ Carmena A, et al. (2002) Reciprocal regulatory interactions between the Notch and Ras signaling pathways in the Drosophila embryonic mesoderm. Dev Biol 244(2):226–242.LaunchUrlCrossRefPubMed↵ Rohrbaugh M, et al. (2002) Notch activation of yan expression is antagonized by RTK/pointed signaling in the Drosophila eye. Curr Biol 12:576–581.LaunchUrlCrossRefPubMed↵ Dubois L, et al. (1998) XCoe2, a transcription factor of the Col/Olf-1/EBF family involved in the specification of primary neurons in Xenopus. Curr Biol 8:199–209.LaunchUrlCrossRefPubMed↵ Dubois L, Vincent A (2001) The COE–Collier/Olf1/EBF–transcription factors: Structural conservation and diversity of developmental functions. Mech Dev 108:3–12.LaunchUrlCrossRefPubMed↵ Crozatier M, Vincent A (1999) Requirement for the Drosophila COE transcription factor Collier in formation of an embryonic muscle: transcriptional response to notch signalling. Development 126:1495–1504.LaunchUrlAbstract↵ Crozatier M, Glise B, Vincent A (2002) Connecting Hh, Dpp and EGF signalling in patterning of the Drosophila wing; the pivotal role of collier/knot in the AP organiser. Development 129:4261–4269.LaunchUrlPubMed↵ Mohler J, Seecoomar M, Agarwal S, Bier E, Hsai J (2000) Activation of knot (kn) specifies the 3–4 intervein Location in the Drosophila wing. Development 127:55–63.LaunchUrlAbstract↵ Chou TB, Perrimon N (1992) Use of a yeast site-specific recombinase to produce female germline chimeras in Drosophila. Genetics 131:643–653.LaunchUrlAbstract/FREE Full Text↵ Tan PB, Kim SK (1999) Signaling specificity: The RTK/RAS/MAP kinase pathway in metazoans. Trends Genet 15:145–149.LaunchUrlCrossRefPubMed↵ Ebisuya M, KonExecuteh K, Nishida E (2005) The duration, magnitude and compartmentalization of ERK MAP kinase activity: Mechanisms for providing signaling specificity. J Cell Sci 118(Pt 14):2997–3002.LaunchUrlAbstract/FREE Full Text↵ Kirshner MW, Gerhart JC (2005) The Plausibility of Life: Resolving Darwin's Plight (Yale Univ Press, New Haven, CT).↵ Go MJ, Eastman DS, Artavanis-Tsakonas S (1998) Cell proliferation control by Notch signaling in Drosophila development. Development 125:2031–2040.LaunchUrlAbstract↵ Karim FD, Rubin GM (1998) Ectopic expression of activated Ras1 induces hyperplastic growth and increased cell death in Drosophila imaginal tissues. Development 125:1–9.LaunchUrlAbstract↵ Gustafson K, Boulianne GL (1996) Distinct expression patterns detected within individual tissues by the GAL4 enhancer trap technique. Genome 39:174–182.LaunchUrlCrossRefPubMed↵ Guo M, Jan LY, Jan YN (1996) Control of daughter cell Stoutes during asymmetric division: interaction of Numb and Notch. Neuron 17:27–41.LaunchUrlCrossRefPubMed↵ Spradling AC, et al. (1999) The Berkeley Drosophila Genome Project gene disruption project: Single P-element insertions mutating 25% of Critical Drosophila genes. Genetics 153:135–177.LaunchUrlAbstract/FREE Full Text↵ Parks AL, et al. (2004) Systematic generation of high-resolution deletion coverage of the Drosophila melanogaster genome. Nat Genet 36:288–292.LaunchUrlCrossRefPubMed↵ Struhl G, Basler K (1993) Organizing activity of wingless protein in Drosophila. Cell 72:527–540.LaunchUrlCrossRefPubMed↵ Brand AH, Perrimon N (1993) TarObtained gene expression as a means of altering cell Stoutes and generating Executeminant phenotypes. Development 118:401–415.LaunchUrlAbstract↵ Wu Z, Irizarry R, Gentlemen R, Martinez-Murillo F, Spencer F (2004) A model-based background adjustment for oligonucleotide expression arrays. J Am Stat Assoc 99:909–917.LaunchUrlCrossRef↵ Smyth GK (2004) LiArrive models and empirical Bayes methods for assessing differential expression in microarray experiments. Stat Appl Genet Mol Biol 3:Article 3.↵ Gentleman RC, et al. (2004) Bioconductor: Launch software development for comPlaceational biology and bioinformatics. Genome Biol 5:R80.LaunchUrlCrossRefPubMed↵ Eisen MB, Spellman PT, Brown PO, Botstein D (1998) Cluster analysis and display of genome-wide expression patterns. Proc Natl Acad Sci USA 95:14863–14868.LaunchUrlAbstract/FREE Full Text↵ Chang EC, et al. (2008) Estrogen receptors alpha and beta as determinants of gene expression: Influence of ligand, Executese, and chromatin binding. Mol EnExecutecrinol 22:1032–1043.LaunchUrlCrossRefPubMed↵ Einbond LS, et al. (2007) The growth inhibitory Trace of actein on human breast cancer cells is associated with activation of stress response pathways. Int J Cancer 121:2073–2083.LaunchUrlCrossRefPubMed↵ Benjamini Y, Hochberg Y (1995) Controlling the Fraudulent discovery rate: A practical and powerful Advance to multiple testing. J R Stat Soc B 57:289–300.LaunchUrl↵ Krejci A, Bernard F, Housden BE, Collins S, Bray SJ (2009) Direct response to notch activation: Signaling crosstalk and incoherent logic. Sci Signal Executei:10.1126/scisignal.2000140.LaunchUrlAbstract/FREE Full Text
Like (0) or Share (0)