The Drosophila EGF receptor (DER) is activated by secreted Spitz to induce different cell fates in the ventral ectoderm. Processing of the precursor transmembrane Spitz to generate the secreted form was shown to be the limiting event, but the cells in which processing takes place and the mechanism that may generate a gradient of secreted Spitz in the ectoderm were not known. The ectodermal defects in single minded (sim) mutant embryos, in which the midline fails to develop, suggested that the midline cells contribute to patterning of the ventral ectoderm. This work shows that the midline provides the site for Spitz expression and processing. The Rhomboid and Star proteins are also expressed and required in the midline. The ectodermal defects of spitz, rho or Star mutant embryos could be rescued by inducing the expression of the respective normal genes only in the midline cells. Rho and Star thus function non-autonomously, and may be required for the production or processing of the Spitz precursor. Secreted Spitz is the only sim-dependent contribution of the midline to patterning the ectoderm, since the ventral defects observed in sim mutant embryos can be overcome by expression of secreted Spitz in the ectoderm. While ectopic expression of secreted Spitz in the ectoderm or mesoderm gave rise to ventralization of the embryo, increased expression of secreted Spitz in the midline did not lead to alterations in ectoderm patterning. A mechanism for adjustment to variable levels of secreted Spitz emanating from the midline may be provided by Argos, which forms an inhibitory feedback loop for DER activation. The production of secreted Spitz in the midline, may provide a stable source for graded DER activation in the ventral ectoderm.

Gradients of diffusible polypeptides play cardinal roles in the determination of a diverse array of cell fates. A variety of mechanisms have been identified for the generation of such gradients. For example, the tissue that produces the diffusible factor may be insensitive to it, while the neighboring tissue is competent to respond to the signals emanating from the producing cells. This mechanism was shown to take place in the case of the Drosophila Hedgehog protein which is produced by the engrailed-expressing cells, and diffuses to pattern the neighboring cells (Heemskerk and DiNardo, 1994). In the wing imaginal disc, a stripe of dpp-expressing cells serves as a source for long-range influence of the secreted protein, which may induce the expression of target genes in a concentration-dependent manner (Nellen et al., 1996; Lecuit et al., 1996). Alternatively, an inactive precursor molecule is produced and its site of activity is defined by the distribution of molecules which take part in processing. For example, the maternal Spätzle precursor is distributed as an inactive protein in the perivitelline fluid surrounding the Drosophila embryo. Only upon proteolytic processing that is restricted to the ventral vitelline membrane, will it activate the Toll receptor in the ventral and ventrolateral regions of the embryonic ectoderm (reviewed in Chasan and Anderson, 1993).

In the Drosophila ectoderm, the initial determination of fates is induced by the maternal Dorsal pathway (Chasan and Anderson, 1993). Consequently, the embryo is subdivided into distinct domains of gene expression along the dorsoventral axis, including the precursors of the mesoderm, ventral ectoderm, dorsal ectoderm and amnioserosa. Furthermore, within the ventral ectoderm, the nuclear levels of the Dorsal protein may contribute to the induction of different neuro-blast cell fates. In the same region, the zygotic Drosophila EGF receptor (DER) pathway further specifies and refines ventral cell fates. In DER/flb mutant embryos, ventral cell identities are replaced by lateral fates (Raz and Shilo, 1993). Elucidation of the DER signaling pathway has been facilitated by the identification of the spitz group genes, which give rise to similar phenotypes and interact genetically with DER (Mayer and Nüsslein-Volhard, 1988; Sturtevant et al., 1993; Raz and Shilo, 1993). The spitz gene encodes a TGFα-like molecule which is produced as a transmembrane precursor (Rutledge et al., 1992). Overexpression in the ectoderm of secreted Spitz, but not of the transmembrane precursor, results in ventralization of the embryo. This observation suggests that the levels of secreted Spitz determine the extent of DER activation in the ectoderm (Schweitzer et al., 1995b).

Different cell fates are monitored in the ventral ectoderm. (1) The ventralmost cells, which experience the highest level of DER activation, express one set of markers, including orthodenticle (otd). (2) The ventrolateral cells express other markers such as Fasciclin III (Raz and Shilo, 1993). It is thus possible that secreted Spitz is present in a gradient, with the highest concentrations in the ventral midline.

Since the Spitz precursor is ubiquitously expressed (Rutledge et al., 1992), the gradient of DER activation could be regulated by restricted processing of Spitz. Two molecules that may participate in Spitz production or processing were identified as members of the spitz group: Rhomboid (Rho) and Star are novel transmembrane proteins containing seven and one transmembrane domains, respectively (Bier et al., 1990; Kolodkin et al., 1994). Genetic epistasis experiments suggested that both proteins function upstream to Spitz, since their mutant phenotypes can be overcome by expression of secreted Spitz (Schweitzer et al., 1995b). However, since high levels of secreted Spitz were used, it was not possible to say whether physiological levels of secreted Spitz would indeed be able to bypass the requirement for Rho and Star. Furthermore, the possible mechanism of action in facilitating Spitz processing and relevant sites of activity of Rho and Star are not known.

It is interesting to note that the single minded (sim) gene, which is required for the development of the midline cells, has also been assigned to the spitz group (Mayer and Nüsslein-Volhard, 1988). Sim is a bHLH-PAS transcription factor that is expressed exclusively in the mesectoderm and midline, and is responsible for the induction of midline cell fates (Nambu et al., 1990, 1991). In sim mutant embryos, ventral cell fates are missing, as monitored by cuticle preparations (Mayer and Nüsslein-Volhard, 1988) and lack of expression of ventral markers (Kim and Crews, 1993). An indication for the role of Sim in the context of the spitz group is provided by the observation that spitz, rho and Star are expressed in the midline (Rutledge et al., 1992; Bier et al., 1990; Kolodkin et al., 1994; Heberlein et al., 1993). The expression of rho was shown to be dependent upon Sim (Nambu et al., 1990). Presumably, expression of spitz and Star also requires the function of Sim.

In this work, we investigate the sites in which production and processing of Spitz are required in the embryo. The normal function of Spitz appears to be non autonomous: expression of the Spitz precursor only in the midline is sufficient for patterning the ventral ectoderm. Rho and Star also function non autonomously and are required in the midline. Facilitating the expression of spitz, rho and Star is the only sim-dependent contribution of the midline to patterning the ventral ectoderm, since the mutant sim ectodermal defects can be overcome by expression of secreted Spitz in the ectoderm. These results suggest a mechanism for generating a graded distribution of secreted Spitz, which may subsequently give rise to graded activation of DER in the ectoderm.

Probes and antibodies

The otd probe was kindly provided by R. Finkelstein and the FasIII monoclonal antibodies by T. Volk. Standard procedures were used for RNA in situ hybridization and antibody staining. Enhancer trap lines were stained with X-gal or rabbit anti-β-Gal antibodies (Cappel). Balancer chromosomes carrying lacZ-expressing constructs were identified by X-gal staining.

Plasmid constructs

The UAS-Star construct was generated by introducing a 3.3 kb EcoRI fragment containing the entire coding sequence of the Star cDNA into pUAST. The UAS-rho construct was generated by introducing the 1.05 kb DdeI fragment of the Rho cDNA, containing the entire coding region, into the EcoRI-BglII sites of the pUAST vector (Brand and Perrimon, 1993).

Fly strains

Strains carrying the following mutant chromosomes were used: spiIIT25, spiOE92, rhoΔ38, SIIN23, flb1F26, simH9and snaII6. The C50.1S1 enhancer trap line in the disco locus, and a sim enhancer trap line were used. The following Gal4 inducer lines were used: twist-Gal4 (2nd chromosome), sim-Gal4 (2nd chromosome), rho-Gal4 (2nd chromosome), Kr-Gal4/TM3 (3rd chromosome) and 69B (3rd chromosome). The following UAS lines were used: UAS-sspi4a (2nd chromosome), UAS-rho24-2 (2nd chromosome), UAS-rho11-1(3rd chromosome), UAS-mspi7b (1st chromosome) and UAS-S4-3 (3rd chromosome).

To test the effects of induced expression in mutant backgrounds the following lines were constructed: (1) w; sim-Gal4, spiOE92/CyO lacZ (2) w, UAS-mspi7b; spiOE92/CyO lacZ (3) w; sim-Gal4; rhoΔ38/TM3 lacZ (4) w; UAS-rho24-2; rhoΔ38/TM3 lacZ (5) w; sim-Gal4, SIIN23/ CyO lacZ (6) w; SIIN23/CyO lacZ; UAS-Star4-3 (7) w; rho-Gal4; simH9 /TM3 lacZ (8) w; UAS-sspi4a; simH9/TM3 lacZ (9) w; SIIN23/ CyO lacZ; 69B. The flb1F26 mutant embryos were grown at 29°C. Embryos were collected at 1 hour intervals at 25°C and aged until stage 11 (6.5 hours after egg lay).

Non-autonomous activity of secreted Spitz and Rhomboid

In order to express secreted Spitz, lines carrying a UAS-secreted spitz construct were generated. Upon expression of secreted Spitz in the ectoderm, the tissue in which DER is activated, ventralization of the embryo was observed (Schweitzer et al., 1995b). We next examined whether secreted Spitz expression in neighboring tissues would result in a similar outcome. To do so, secreted Spitz expression was induced by twist-Gal4, which is expressed in the mesoderm and midline (Baker and Schubiger, 1996) (see scheme in Fig. 1A). In these embryos, normal invagination of the mesoderm takes place, as monitored by FasIII expression in the visceral mesoderm (not shown). Ventral cell fates in the ectoderm were followed by expression of otd, which normally marks the ventralmost cells (Wieschaus et al., 1992; and Fig. 1B). A dramatic ventralization of the embryos is observed; up to 8-10 cell rows express otd on each side of the midline (Fig. 1C), comparable to the result obtained following expression of secreted Spitz in the ectoderm. This experiment demonstrates that secreted Spitz can diffuse readily between the mesoderm and ectoderm. Although the mesoderm spreads over the entire ectoderm, expression of otd was detected only in the ventral ectoderm. This is likely to result from the fact that by the time secreted Spitz was induced, dorsal cell fates have already been established by the Dpp pathway (see Discussion).

Fig. 1.

Secreted Spitz or Rhomboid in the mesoderm induce ventralization of the ectoderm. The non-autonomous effects on ectodermal patterning of secreted Spitz or Rho expressed in the mesoderm were examined. (A) A scheme of a cross section of a germband-extended embryo (stage 10/11), showing the different embryonic tissues and the expression domains of otd and fasIII. (B) otd expression in wild-type embryos is restricted to 1 –2 cell rows on each side of the midline. (C) In twist-Gal4/UAS-secreted spitz embryos or (D) twist-Gal4/UAS-rho embryos, otd expression is expanded to 8-10 cell rows on each side of the midline. Anterior is to the left in all figures.

Fig. 1.

Secreted Spitz or Rhomboid in the mesoderm induce ventralization of the ectoderm. The non-autonomous effects on ectodermal patterning of secreted Spitz or Rho expressed in the mesoderm were examined. (A) A scheme of a cross section of a germband-extended embryo (stage 10/11), showing the different embryonic tissues and the expression domains of otd and fasIII. (B) otd expression in wild-type embryos is restricted to 1 –2 cell rows on each side of the midline. (C) In twist-Gal4/UAS-secreted spitz embryos or (D) twist-Gal4/UAS-rho embryos, otd expression is expanded to 8-10 cell rows on each side of the midline. Anterior is to the left in all figures.

The Spitz precursor is normally expressed in the ectoderm and mesoderm (Rutledge et al., 1992), while Rho is excluded from the mesoderm due to transcriptional repression by Snail (Ip et al., 1992; Nambu et al., 1990). Rho was considered to function as a facilitator of DER signaling and, in view of its multiple transmembrane domains, it was assumed to be required in the same cells as DER. While epistasis studies have suggested that Rho acts upstream to the production of secreted Spitz (Schweitzer et al., 1995b), it was still possible to argue that the high levels of secreted Spitz used may overcome the requirement for facilitation of DER signaling by Rho. Since secreted Spitz expressed in the mesoderm can affect cell fate specification in the ectoderm, it was possible to test whether Rho would also be able to function non-autonomously. Indeed, when Rho was ectopically expressed in the mesoderm and midline, a similar degree of ectodermal ventralization was observed (Fig. 1D). Expression of Star in the mesoderm by twist-Gal4 did not give rise to any phenotype (not shown).

Expression of Spitz precursor in the midline patterns the ventral ectoderm

We have shown above that under artificial situations, Spitz and Rho can induce patterning of the ectoderm, in a non-autonomous manner. In view of the ectodermal defects of sim mutant embryos and the expression of spitz, rho and Star in the midline, it is plausible that during normal development the midline is the source of secreted Spitz to the ectoderm. The capacity of the Spitz precursor to rescue the spitz mutant phenotype in the ectoderm was tested, following induction of UAS-membrane spitz exclusively in the midline (by a sim-Gal4 construct). FasIII is normally expressed in 4-5 cell rows on each side of the midline in the thoracic segments (Raz and Shilo, 1993; Fig. 2A). In spitz mutant embryos, no expression of FasIII is detected in the ventral ectoderm. Upon induction of the normal spitz gene in the midline, partial rescue was observed and up to 4 rows of cells express FasIII (Fig. 2C). The incomplete rescue may be accounted for by the fact that expression of target genes induced by sim-Gal4 (e.g. UAS-lacZ) is observed only from stage 9/10 (not shown), which may be later than the time in which spitz normally begins to be expressed in the midline. Thus, induction by sim-Gal4 may provide expression levels that are lower than the level required for normal activity.

Fig. 2.

Spitz, Rhomboid and Star are required in the midline to pattern the ventral ectoderm. (A) In wild-type embryos, expression of FasIII is observed in 4-5 cell rows on each side of the midline of the thoracic parasegments. (B) In spiIIT25 mutant embryos, no expression of FasIII in the ectoderm is detected. (C) In spiOE92 homozygous embryos in which the expression of UAS-membrane spitz was induced by sim-Gal4, partial rescue of the phenotype is observed, such that up to 4 cell rows on each side express FasIII. The two spi alleles used are null alleles. (D) rhoΔ38 mutant embryo. (E) Induction of UAS-rho by sim-Gal4 in rhoΔ38 mutant embryos resulted in partial rescue of FasIII expression. (F) SIIN23 mutant embryo. (G) Induction of UAS-Star by sim-Gal4 in SIIN23 mutant embryos gave rise to partial rescue.

Fig. 2.

Spitz, Rhomboid and Star are required in the midline to pattern the ventral ectoderm. (A) In wild-type embryos, expression of FasIII is observed in 4-5 cell rows on each side of the midline of the thoracic parasegments. (B) In spiIIT25 mutant embryos, no expression of FasIII in the ectoderm is detected. (C) In spiOE92 homozygous embryos in which the expression of UAS-membrane spitz was induced by sim-Gal4, partial rescue of the phenotype is observed, such that up to 4 cell rows on each side express FasIII. The two spi alleles used are null alleles. (D) rhoΔ38 mutant embryo. (E) Induction of UAS-rho by sim-Gal4 in rhoΔ38 mutant embryos resulted in partial rescue of FasIII expression. (F) SIIN23 mutant embryo. (G) Induction of UAS-Star by sim-Gal4 in SIIN23 mutant embryos gave rise to partial rescue.

Rhomboid and Star in the midline are sufficient to pattern the ventral ectoderm

In view of the capacity of the Spitz precursor in the midline to pattern the ventral ectoderm shown above, it is possible that additional components of the spitz group would also function in the midline. rho and Star mutant embryos show ectodermal defects similar to spitz mutants (Mayer and Nüsslein-Volhard, 1988; and Fig. 2B,D,F). Epistasis experiments have shown that Rho and Star may function before or during Spitz processing (Schweitzer et al., 1995b). To test the possibility that Rho and Star are required in the midline, the rescue of rho or Star mutant phenotypes was examined following expression of the normal respective gene only in the midline.

While rho mutant embryos display no expression of FasIII in the ventral ectoderm, expression of rho in the midline was able to partially rescue this defect. In the thoracic segments of the rescued embryos, 1-4 cell rows expressing FasIII can be detected on each side of the midline (Fig. 2D,E). A similar experiment was carried out by expressing Star in the midline, in Star mutant embryos. Again, 1-3 cell rows expressing FasIII can be monitored on each side of the midline (Fig. 2F,G).

Secreted Spitz is the only sim-dependent element contributed by the midline to pattern the ectoderm

We have shown that Spitz is required predominantly or exclusively in the midline. Therefore, sim mutants, which are devoid of a functional midline, fail to activate DER and pattern the ectoderm. To test if the early development of the midline itself requires the DER pathway, midline markers were monitored in DER/flb mutant embryos. sim-lacZ expression shows no aberrations in midline patterning (Fig. 3A). The normal development of the midline is thus independent of the zygotic patterning of the ectoderm through the DER pathway.

Fig. 3.

Secreted Spitz is epistatic to the sim mutant phenotype. (A) Midline development does not depend on the DER pathway, as monitored by the expression of sim-lacZ in flb1F26 mutant embryos. (B) simH9 mutant embryos show significantly diminished expression of FasIII. (C) Induction of UAS-secreted spitz by rho-Gal4 in the entire neuroectoderm of simH9 mutant embryos is epistatic and results in expanded expression of FasIII. Note: due to the stability of Gal4, the early expression of rho in the entire neuroectoderm is retained until stage 12.

Fig. 3.

Secreted Spitz is epistatic to the sim mutant phenotype. (A) Midline development does not depend on the DER pathway, as monitored by the expression of sim-lacZ in flb1F26 mutant embryos. (B) simH9 mutant embryos show significantly diminished expression of FasIII. (C) Induction of UAS-secreted spitz by rho-Gal4 in the entire neuroectoderm of simH9 mutant embryos is epistatic and results in expanded expression of FasIII. Note: due to the stability of Gal4, the early expression of rho in the entire neuroectoderm is retained until stage 12.

It is possible that, in addition to Spitz, the midline may provide other factors that are essential for patterning the ectoderm. This option can be tested experimentally. If the midline is contributing additional sim-dependent factors, induction of secreted Spitz in the ectoderm of sim mutant embryos should not be able to induce the appearance of ventral cell fate markers. Conversely, if the midline is providing to the ectoderm only a source of secreted Spitz, expression of secreted Spitz in the ectoderm to activate the DER pathway should be epistatic to the sim mutant phenotype.

The experiment showed that secreted Spitz in the ectoderm is indeed epistatic to the sim mutant phenotype, and the extent of ventralization observed is comparable to the one monitored following a similar expression of secreted Spitz in wild-type embryos (Fig. 3C). This experiment demonstrates that the sim-dependent contribution of the midline to the ectoderm is confined to the production and processing of Spitz. In sim mutant embryos, Spitz, Rho and Star have been eliminated only in the midline; yet, no expression of the ventralmost markers, such as otd, and only rudimentary expression of ventrolateral markers is observed in these embryos (Kim and Crews, 1993; and Fig. 3B). Thus, the severity of the sim phenotype implies that the major, if not the exclusive, source of secreted Spitz emanates from the midline.

Rhomboid is the only limiting component of the spitz group in the ectoderm

The demonstration that the activity of several spitz group genes is normally restricted to the midline was surprising, in view of the fact that the known essential components are also expressed in the ectoderm at the early stages of embryogenesis. spitz has maternal transcripts, and the zygotic expression is detected in the ectoderm, mesoderm and midline (Rutledge et al., 1992). Star expression was detected in the ectoderm and the midline (Kolodkin et al., 1994; Heberlein et al., 1993). Finally, rho is expressed in a highly dynamic pattern. The initial zygotic transcription of rho is induced by Dorsal and Twist and repressed by Snail. As a result, rho transcripts are detected in the neuroectoderm but are excluded from the mesoderm. The expression of rho recedes from the neuroectoderm towards the midline, such that by stage 10 only the midline expression is retained (Bier et al., 1990; Nambu et al., 1990; Ip et al., 1992).

We wanted to test whether spitz, rho or Star are limiting in the ectoderm. Alternatively, another unknown component required for Spitz processing may be absent from the ectoderm. Each of the three genes was expressed ectopically, in the ectoderm of wild-type embryos. No alterations in ventral cell fates were detected following expression of membrane Spitz (Schweitzer et al., 1995b). In the case of Star, ubiquitous ectopic expression in Star mutant embryos resulted in an almost complete rescue of the ventral defects (Fig. 4F). No aberrant phenotypes were observed following ectopic expression of Star in wild-type embryos (not shown). In contrast, ectopic expression of rho by ectodermal inducers such as rho-Gal4 or Kr-Gal4 resulted in an expansion of ventral cell fates, similar to the one observed following ectopic expression of secreted Spitz (Fig. 4A-D). The effects of Rho expression are dependent upon a functional DER pathway and are not observed in DER/flb mutant embryos (Fig. 4E). We can therefore conclude that, at the time in which we induced ectopic expression, Rho is the only limiting component in the ectoderm, with respect to the machinery required to regulate and carry out processing of Spitz.

Fig. 4.

Ectopic Rho in the ectoderm induces ventralization. (A) Induction of UAS-rho expression by rho-Gal4 results in expanded expression of otd. (B) Induction of UAS-rho expression by Kr-Gal4 leads to expanded expression of FasIII in parasegments T2 (arrow) and T3, but not in T1 (arrowhead). (C) An enhancer trap in the disco locus is expressed in the lateral ectodermal cells. The dorsal border of the disco patch is shown by an arrow, and a more dorsal dot expressing disco by an arrowhead. (D) Induction of rho expression by rho-Gal4 results in expansion of the ventral fates at the expense of lateral fates. The dorsal borders of disco expression are retained, however (arrow and arrowhead). (E) Induction of ectopic ventral fates by the expression of rho in the ectoderm is dependent upon a functional DER pathway. When carried out in flb1F26 mutant embryos, no expression of FasIII is detected. (F) The Star mutant ectodermal phenotype is rescued by ubiquitous induction of UAS-Star expression by the 69B-Gal4 construct. No ectopic cell fates are induced by ectopic S expression, either in wild-type (not shown) or in mutant S embryos.

Fig. 4.

Ectopic Rho in the ectoderm induces ventralization. (A) Induction of UAS-rho expression by rho-Gal4 results in expanded expression of otd. (B) Induction of UAS-rho expression by Kr-Gal4 leads to expanded expression of FasIII in parasegments T2 (arrow) and T3, but not in T1 (arrowhead). (C) An enhancer trap in the disco locus is expressed in the lateral ectodermal cells. The dorsal border of the disco patch is shown by an arrow, and a more dorsal dot expressing disco by an arrowhead. (D) Induction of rho expression by rho-Gal4 results in expansion of the ventral fates at the expense of lateral fates. The dorsal borders of disco expression are retained, however (arrow and arrowhead). (E) Induction of ectopic ventral fates by the expression of rho in the ectoderm is dependent upon a functional DER pathway. When carried out in flb1F26 mutant embryos, no expression of FasIII is detected. (F) The Star mutant ectodermal phenotype is rescued by ubiquitous induction of UAS-Star expression by the 69B-Gal4 construct. No ectopic cell fates are induced by ectopic S expression, either in wild-type (not shown) or in mutant S embryos.

Patterning of the ventral ectoderm can adjust to variable levels of secreted Spitz from the midline

Having shown that the midline is the major source of Spitz activity, we wanted to test the sensitivity of the system to variable levels of secreted Spitz. One way to increase the amount of secreted Spitz in the midline is to induce the expression of the secreted Spitz construct by sim-Gal4 in wild-type embryos. Under these conditions, no expansion of ventral cell fates was detected, as monitored by the expression of otd (Fig. 5A). Another procedure to increase the level of secreted Spitz in the midline is to examine snail mutant embryos, in which the midline was shown to be dramatically expanded at the expense of the mesoderm (Nambu et al., 1990; Leptin, 1991; Rao et al., 1991). Up to 10 rows of the ventralmost cells express midline markers in the blastoderm stage embryo. Eventually these cells delaminate and the neighboring cells develop as neuroectoderm. Again, under these conditions, normal expression of otd was monitored (Fig. 5B).

Fig. 5.

Patterning of the ventral ectoderm is retained after elevating secreted Spitz levels in the midline. No alterations in the normal pattern of otd expression were monitored following an increase in the level of secreted Spitz emanating from the midline. (A) Induction of UAS-secreted Spitz by sim-Gal4 in wild-type embryos. (B) snail mutant embryos in which the midline is expanded at the expense of the mesoderm.

Fig. 5.

Patterning of the ventral ectoderm is retained after elevating secreted Spitz levels in the midline. No alterations in the normal pattern of otd expression were monitored following an increase in the level of secreted Spitz emanating from the midline. (A) Induction of UAS-secreted Spitz by sim-Gal4 in wild-type embryos. (B) snail mutant embryos in which the midline is expanded at the expense of the mesoderm.

While ectopic expression of secreted Spitz in the ectoderm or mesoderm gave rise to dramatic ventralization of the ectoderm, an increase in the level of secreted Spitz in the midline had no effect. Judging by the level of induction of UAS-lacZ by sim-Gal4, the activity of sim-Gal4 appears comparable to that of the ectodermal or mesodermal Gal4 lines that were used (not shown). We discuss below how Argos, which provides a negative feedback loop for the DER pathway, may be responsible for adjustments to higher levels of secreted Spitz emanating from the midline.

The midline is the source of secreted Spitz

This work has shown that the midline is the source of secreted Spitz for activating DER in the ventral ectoderm. In the absence of functional midline cells (in sim mutant embryos), ventral cell fates are not induced. It is formally possible that secreted Spitz triggers DER in the midline and the activation of DER induces the production of unknown secreted proteins that will diffuse, relay the signal and activate a different signaling pathway in the ectoderm. However, we believe this is not the case, since there is evidence that activation of DER normally takes place in the ectoderm: expression of argos is induced directly by the DER pathway in embryos and cells. argos is expressed, in a DER-dependent manner, in the ventralmost ectodermal cells (Golembo et al., 1996; Gabay et al., 1996). We can therefore suggest that, during normal development, after the midline is formed, it serves as a source for producing secreted Spitz, which will diffuse to the ventral ectoderm. Activation of DER, expressed in all ectodermal cells, will pattern the ventral ectoderm. Production of the midline depends upon cues from the Dorsal protein and the early zygotic genes twist and snail (Nambu et al., 1990; Leptin, 1991; Rao et al., 1991). It does not depend however, on the DER pathway.

The sim-dependent function of the midline in terms of patterning the ventral ectoderm seems to be restricted to Spitz processing, since ectopic expression of secreted Spitz is epistatic to the sim mutant phenotype. This work has shown that the Spitz precursor, Rho and Star function non-autonomously in patterning the ectoderm. They are normally expressed in the midline and their expression only in the midline is sufficient to rescue the ectodermal defects of the respective mutant embryos. Rho and Star may be required for the production or processing of the Spitz precursor. The same components are required later in embryogenesis for differentiation of the midline itself (Klämbt et al., 1991).

The production of secreted Spitz in the midline may generate a stable source of gradient for the ectoderm. The actual shape of the secreted Spitz gradient may depend upon several factors, such as the stability of the protein, the diffusion rate, dilution of the protein at a distance from its local source and the extent to which the Spitz protein will be trapped by DER on the ectoderm or by the extracellular matrix. Diffusion of Spitz has also been demonstrated in the eye imaginal disc (Freeman, 1994). While this work has shown that Rho and Star are required in the midline, their mode of action in facilitating Spitz processing remains an open question. They appear to play non-overlapping functions in the process, since both seem to be required in the midline. The ability to obtain ventralization following expression of Rho in the mesoderm suggests that Star or a molecule analogous to Star is normally expressed in the mesoderm.

Why processing of Spitz is restricted to the midline is puzzling, in view of the fact that the known components necessary for Spitz processing are also expressed in the ectoderm, at least at the early stages of embryogenesis. Induction of Rho in the ectoderm gave rise to ectopic ventral fates, suggesting that, at the time of induction or later on, it is the only component limiting or missing for effective Spitz processing. It is possible that at the earlier time in which Rho is normally expressed in the ectoderm, its contribution to Spitz processing is marginal. Alternatively, at this stage, the cells may not be competent to process Spitz, or to respond to secreted Spitz which may be produced in the ectoderm.

The expansion of ventral fates that can be obtained by hyper-activation of the DER pathway (by ectopic expression of secreted Spitz or Rho, or by removal of inhibitory molecules such as Argos), never exceeds 8-10 cell rows on each side of the midline. This result does not necessarily reflect the limits of diffusion of secreted Spitz or the cells in which components of the DER pathway are present, but rather the normal temporal distinction between the establishment of dorsal and ventral fates in the embryo. At 2.5 hours of embryogenesis, the Dpp pathway is triggered in the dorsal ectoderm, and the ventral ectoderm remains naive due to the inhibitory activity of Short gastrulation (Francois et al., 1994). Only later on, presumably after the processing machinery of Spitz begins to function in the midline, does the DER pathway pattern the ventral ectoderm (Raz and Shilo, 1993). Thus, ectopic activation of DER can alter only the fate of cells in the ventral and lateral ectoderm, which have not been previously patterned by Dpp. Under such conditions, the dorsal borders of lateral markers like disco retain their normal position (Fig. 4D).

Adjustments to variable levels of secreted Spitz

The ventral ectoderm is exquisitely sensitive to ectopic expression of secreted Spitz when it is provided uniformly, either in the ectoderm or from the mesoderm. Yet, no alterations in patterning were detected when the levels of secreted Spitz were elevated in the midline. We believe that these observations have a profound implication on the way in which graded signaling of the DER pathway is maintained. A central consideration in the regulation of DER activation is not only how the pathway is triggered, but also how signaling is terminated. Graded activation of DER generates a gradient of enzymatic activity of the DER tyrosine kinase and the cytoplasmic kinases triggered by it. If allowed to function over extended time periods, the consequences of DER activation would be similar in all ectodermal cells, regardless of the initial activation levels. Therefore, a special mechanism has evolved in order to define a sharp time window of signaling and to preserve the effects of graded DER activation.

Argos is a secreted polypeptide containing a single EGF repeat (Freeman et al., 1992b), which was shown to inhibit the capacity of Spitz to trigger DER autophosphorylation (Schweitzer et al., 1995a). The mechanism of Argos activity has been highlighted by the observation that transcription of argos is induced by the DER pathway (Golembo et al., 1996; Gabay et al., 1996). Thus, activation of DER triggers the expression of Argos, which is responsible in turn for shutting off the pathway. The effects of Argos are physiologically important not for the cells in which it is produced, but rather in the neighboring cells, which do not express it. Argos thus functions as a signal of lateral inhibition, to restrict the duration or level of DER signaling in the neighboring cells and preserve the graded effects of DER signaling.

We suggest that the inhibitory feedback loop of Argos may account for our observations that, even when higher levels of secreted Spitz are emanating from the midline, normal patterning of the ventral ectoderm is retained. Under these conditions, the ectodermal cells closest to the midline will experience the highest level of DER activation. These cells will subsequently induce Argos, which will shut off signaling in the more lateral cells. As long as the midline is the only source of secreted Spitz, the Argos feedback mechanism is retained even when the absolute levels of secreted Spitz are altered. Thus, the system is extremely flexible in terms of adjusting to variable levels of secreted Spitz from the midline. The situation is completely different if all ectodermal cells are simultaneously encountering secreted Spitz (produced either in the ectoderm or mesoderm). Although Argos will be induced by all ectodermal cells (Golembo et al., 1996), by the time it is made the cells producing it are ventralized and refractive to Argos inhibition. Fig. 6 summarizes the model for Spitz processing in the midline and the role of Argos in retaining the effects of the secreted Spitz gradient.

Fig. 6.

A model for establishment and maintenance of graded DER activation in the ectoderm. Secreted Spitz is required to activate DER and pattern the ventral ectoderm. The major, if not the exclusive source of secreted Spitz is the midline. In these cells, the expression of spitz, rho and Star depends directly or indirectly on Sim (1), and provides the processing apparatus of the Spitz precursor (2). Secreted Spitz diffuses from the midline, and forms a gradient that is responsible for graded activation of DER in the ectoderm (3). In the cells closest to the midline, the highest level of DER activation will take place. Consequently, expression of argos will be induced (4). Diffusion of Argos to the neighboring ectodermal cells will terminate or reduce DER signaling, (5) and thus preserve the graded effects of DER activation. The induction of an inhibitory feedback loop provides a mechanism of adjustment to variable levels of secreted Spitz, as long as they are emanating from the midline.

Fig. 6.

A model for establishment and maintenance of graded DER activation in the ectoderm. Secreted Spitz is required to activate DER and pattern the ventral ectoderm. The major, if not the exclusive source of secreted Spitz is the midline. In these cells, the expression of spitz, rho and Star depends directly or indirectly on Sim (1), and provides the processing apparatus of the Spitz precursor (2). Secreted Spitz diffuses from the midline, and forms a gradient that is responsible for graded activation of DER in the ectoderm (3). In the cells closest to the midline, the highest level of DER activation will take place. Consequently, expression of argos will be induced (4). Diffusion of Argos to the neighboring ectodermal cells will terminate or reduce DER signaling, (5) and thus preserve the graded effects of DER activation. The induction of an inhibitory feedback loop provides a mechanism of adjustment to variable levels of secreted Spitz, as long as they are emanating from the midline.

Similar mechanism of DER activation in other tissues

The DER pathway is extremely pleiotropic, affecting a multitude of crucial developmental decisions. In most of these cases, the same DER signaling ‘cassette’ is used. It is interesting to examine whether the non-autonomous effects of DER ligands and their processing that were described in this work, also apply to other tissues.

In the ovary, the DER pathway is responsible for patterning both posterior and dorsal follicle cells (González-Reyes et al., 1995; Roth et al., 1995; Price et al., 1989). In this case, DER is triggered by Gurken, another TGFα homologue (Neuman-Silberberg and Schüpbach, 1993). gurken is expressed only in the germ cells and the transcript is localized to the oocyte nucleus. Gurken triggers DER in the follicle cells, thus acting in a non-autonomous manner. However, it is not known whether Gurken functions as a secreted or a membrane-anchored molecule, and how the expression of rho in the dorsal-anterior follicle cells is involved in DER signaling (Ruohola-Baker et al., 1993).

In the eye imaginal disc, Rho and Star are expressed in the same cells, namely R8, R2 and R5 (Freeman et al., 1992a; Kolodkin et al., 1994; Heberlein et al., 1993). These are the same cells in which Spitz is predominantly required (Freeman, 1994; Tio et al., 1994). Furthermore, the capacity of Spitz to diffuse has been demonstrated in this tissue (Freeman, 1994; Tio et al., 1994). Homozygous Star patches in the eye contain no R cells (Kolodkin et al., 1994). While loss of rho has only a very subtle phenotype in the eye, misexpression of Rho does induce hyperdifferentiation of cells (Freeman et al., 1992a; Freeman, 1994). Finally, Argos is produced by the differentiated cells in the eye disc, and has the capacity to diffuse. Reduction in Argos activity leads to hyperdifferentiation (Freeman et al., 1992b).

In the wing, the DER pathway is responsible for inducing both vein and intervein cell fates (Diaz-Benjumea and Garcia-Bellido, 1990; Diaz-Benjumea and Hafen, 1994). While DER is uniformly expressed, the expression of rho is restricted to the vein precursors (Sturtevant et al., 1994). Thus, processing of Spitz could be restricted to the veins, where Rho is produced, and provide a source of secreted Spitz gradient to the inter-vein regions. It is interesting to note that argos is also expressed in the vein cells (Sawamoto et al., 1994), presumably as a result of induction by high levels of activation of the DER pathway. Reduction in the level of Argos in the wing leads to expansion of the vein region (Sawamoto et al., 1994; Schweitzer et al., 1995a), again demonstrating that Argos is responsible for retaining the graded effects of DER signaling. The regulation of the DER pathway in the wing and ventral ectoderm appear to be highly similar. The major difference is that, in the wing disc, the cells expressing rho which are the likely source of the processed ligand, are induced by the DER pathway to become veins. Therefore, they are also the cells in which argos expression is observed. In the embryo, the midline which produces the secreted ligand is not affected by the DER pathway at the early stages, and the cells expressing argos are therefore the ventralmost ectodermal cells, in which DER activation is maximal.

The analogies between patterning the ventral ectoderm of Drosophila by the DER pathway and the induction of vulval cell fates by the EGF receptor in C. elegans are striking. This work demonstrated that, in addition to the conserved machinery for activating the receptor and terminating signaling (see Golembo et al., 1996), the mechanisms for generating a graded source of the ligand are also similar. The C. elegans Lin-3, a TGF-α homologue, is produced only in the anchor cell (Hill and Sternberg, 1992). Processing of Lin-3 to generate a secreted molecule generates a localized source of ligand, which will induce Let-23, the EGF receptor, in the vulval precursor cells. The distance of a given vulval precursor cell from the anchor cell will dictate the amount of ligand it encounters, and the respective cell fate it will adopt, accordingly (Katz et al., 1995). In the Drosophila ventral ectoderm, processing of the ligand, Spitz, is restricted to the midline cells. This creates a stable gradient of secreted Spitz, which will trigger DER in the ectoderm and may induce graded cell fates accordingly.

We are grateful to U. Banerjee, E. Bier and R. Finkelstein for sending DNA clones, T. Volk for antibodies, and M. Akam, A. Brand, S. Cohen, S. Crews, M. Freeman, C. Klämbt, M. Leptin, M. Levine, C. Nüsslein-Volhard and N. Perrimon for fly lines. We are indebted to stimulating discussions, suggestions and comments by R. Schweitzer, and to an anonymous referee for extremely useful comments. We thank C. Klämbt and E. Schejter for critical reading of the manuscript. This work was supported by grants from the National Institutes of Health, The Council for Tobacco Research, and the Wolfson Fund to B. S.

Baker
,
R.
and
Schubiger
,
G.
(
1996
).
Autonomous and nonautonomous Notch functions for embryonic muscle and epidermis development in Drosophila
.
Development
122
,
617
626
.
Bier
,
E.
,
Jan
,
L. Y.
and
Jan
,
Y. N.
(
1990
).
rhomboid, a gene required for dorsoventral axis establishment and peripheral nervous system development in Drosophila melanogaster
.
Genes Dev
.
4
,
190
203
.
Brand
,
A.
and
Perrimon
,
N.
(
1993
).
Targeted gene expression as a means of altering cell fates and generating dominant phenotypes
.
Development
118
,
401
415
.
Chasan
,
R.
and
Anderson
,
K. V.
(
1993
).
Maternal control of dorsal-ventral polarity and pattern in the embryo
. In
The Development of Drosophila melanogaster
(ed.
M.
Bate
and
A.
Martinez-Arias
), pp.
387
424
.
Cold Spring Harbor
:
Cold Spring Harbor Press
.
Diaz-Benjumea
,
F. J.
and
Garcia-Bellido
,
A.
(
1990
).
Behaviour of cells mutant for an EGF receptor homologue of Drosophila in genetic mosaics
.
Proc. R. Soc. Lond B. Biol. Sci
.
242
,
36
44
.
Diaz-Benjumea
,
F. J.
and
Hafen
,
E.
(
1994
).
The sevenless signalling cassette mediates Drosophila EGF receptor function during epidermal development
.
Development
120
,
569
578
.
Francois
,
V.
,
Solloway
,
M.
,
O’Neill
,
J. W.
,
Emery
,
J.
and
Bier
,
E.
(
1994
).
Dorsal-ventral patterning of the Drosophila embryo depends on a putative negative growth factor encoded by the short gastrulation gene
.
Genes Dev
.
8
,
2602
2616
.
Freeman
,
M.
(
1994
).
The spitz gene is required for photoreceptor determination in the Drosophila eye where it interacts with the EGF receptor
.
Mech. Dev
.
48
,
25
33
.
Freeman
,
M.
,
Kimmel
,
B. E.
and
Rubin
,
G. M.
(
1992a
).
Identifying targets of the rough homeobox gene of Drosophila: evidence that rhomboid functions in eye development
.
Development
116
,
335
346
.
Freeman
,
M.
,
Klämbt
,
C.
,
Goodman
,
C. S.
and
Rubin
,
G. M.
(
1992b
).
The argos gene encodes a diffusible factor that regulates cell fate decisions in the Drosophila eye
.
Cell
69
,
963
975
.
Gabay
,
L.
,
Scholz
,
H.
,
Golembo
,
M.
,
Klaes
,
A.
,
Shilo
,
B.-Z.
and
Klämbt
,
C.
(
1996
).
EGF receptor signaling induces pointed P1 transcription and inactivates Yan protein in the embryonic ventral ectoderm
.
Development
122
,
3357
3360
.
Golembo
,
M.
,
Schweitzer
,
R.
,
Freeman
,
M.
and
Shilo
,
B.-Z.
(
1996
).
Argos transcription is induced by the Drosophila EGF receptor pathway, to form an inhibitory feedback loop
.
Development
122
,
223
230
.
González-Reyes
,
A.
,
Elliot
,
H.
and
St. Johnston
,
D.
(
1995
).
Polarization of both major body axes in Drosophila by gurken-torpedo signalling
.
Nature
375
,
654
658
.
Heberlein
,
U.
,
Hariharan
,
I. K.
and
Rubin
,
G. M.
(
1993
).
Star is required for neuronal differentiation in the Drosophila retina and displays dosage-sensitive interactions with Ras1
.
Dev. Biol
.
160
,
51
63
.
Heemskerk
,
J.
and
DiNardo
,
S.
(
1994
).
Drosophila hedgehog acts as a morphogen in cellular patterning
.
Cell
76
,
449
460
.
Hill
,
R. J.
and
Sternberg
,
P. W.
(
1992
).
The gene lin-3 encodes an inductive signal for vulval development in C
.
elegans. Nature
358
,
470
476
.
Ip
,
Y. T.
,
Park
,
R. E.
,
Kosman
,
D.
,
Bier
,
E.
and
Levine
,
M.
(
1992
).
The dorsal gradient morphogen regulates stripes of rhomboid expression in the presumptive neuroectoderm of the Drosophila embryo
.
Genes Dev
.
6
,
1728
1739
.
Katz
,
W. S.
,
Hill
,
R. J.
,
Clandinin
,
T. R.
and
Sternberg
,
P. W.
(
1995
).
Different levels of the C. elegans growth factor LIN-3 promote distinct vulval precursor fates
.
Cell
82
,
297
307
.
Kim
,
S. H.
and
Crews
,
S. T.
(
1993
).
Influence of Drosophila ventral epidermal development by the CNS midline cells and spitz class genes
.
Development
118
,
893
901
.
Klämbt
,
C.
,
Jacobs
,
J. R.
and
Goodman
,
C. S.
(
1991
).
The midline of the Drosophila central nervous system: a model for the genetic analysis of cell fate, cell migration, and growth cone guidance
.
Cell
64
,
801
815
.
Kolodkin
,
A. L.
,
Pickup
,
A. T.
,
Lin
,
D. M.
,
Goodman
,
C. S.
and
Banerjee
,
U.
(
1994
).
Characterization of Star and its interactions with sevenless and EGF receptor during photoreceptor cell development in Drosophila
.
Development
120
,
1731
1745
.
Lecuit
,
T.
,
Brook
,
W. J.
,
Ng
,
M.
,
Callajea
,
M.
,
Sun
,
H.
and
Cohen
,
S. M.
(
1996
).
Two distinct mechanisms for long-range patterning by Decapentaplegic in the Drosophila wing
.
Nature
381
,
387
393
.
Leptin
,
M.
(
1991
).
twist and snail as positive and negative regulators during Drosophila mesoderm development
.
Genes Dev
.
5
,
1568
1576
.
Mayer
,
U.
and
Nüsslein-Volhard
,
C.
(
1988
).
A group of genes required for pattern formation in the ventral ectoderm of the Drosophila embryo
.
Genes Dev
.
2
,
1496
1511
.
Nambu
,
J. R.
,
Lewis
,
J. O.
,
Wharton
,
J. K. A.
and
Crews
,
S. T.
(
1991
).
The Drosophila single-minded gene encodes a helix-loop-helix protein that acts as a master regulator of CNS midline development
.
Cell
67
,
1157
1167
.
Nambu
,
J. R.
,
Franks
,
R. G.
,
Hu
,
S.
and
Crews
,
S. T.
(
1990
).
The single- minded gene of Drosophila is required for the expression of genes important for the development of CNS midline cells
.
Cell
63
,
63
75
.
Nellen
,
D.
,
Burke
,
R.
,
Struhl
,
G.
and
Basler
,
K.
(
1996
).
Direct and long-range action of a DPP morphogen gradient
.
Cell
85
,
357
368
.
Neuman-Silberberg
,
F. S.
and
Schüpbach
,
T.
(
1993
).
The Drosophila dorsoventral patterning gene gurken produces a dorsally localized RNA and encodes a TGF alpha-like protein
.
Cell
75
,
165
174
.
Price
,
J. V.
,
Clifford
,
R. J.
and
Schüpbach
,
T.
(
1989
).
The maternal ventralizing locus torpedo is allelic to faint little ball, an embryonic lethal, and encodes the Drosophila EGF receptor homolog
.
Cell
56
,
1085
1092
.
Rao
,
Y.
,
Vaessin
,
H.
,
Jan
,
L. Y.
and
Jan
,
Y. N.
(
1991
).
Neuroectoderm in Drosophila embryos is dependent on the mesoderm for positioning but not for formation
.
Genes Dev
.
5
,
1577
1588
.
Raz
,
E.
and
Shilo
,
B. Z.
(
1993
).
Establishment of ventral cell fates in the Drosophila embryonic ectoderm requires DER, the EGF receptor homolog
.
Genes Dev
.
7
,
1937
1948
.
Roth
,
S.
,
Neuman-Silberberg
,
F. S.
,
Barcelo
,
G.
and
Schüpbach
,
T.
(
1995
).
cornichon and the EGF receptor signaling process are necessary for both anterior-posterior and dorsal-ventral pattern formation in Drosophila
.
Cell
81
,
967
978
.
Ruohola-Baker
,
H.
,
Grell
,
E.
,
Chou
,
T. B.
,
Baker
,
D.
,
Jan
,
L. Y.
and
Jan
,
Y. N.
(
1993
).
Spatially localized rhomboid is required for establishment of the dorsal-ventral axis in Drosophila oogenesis
.
Cell
73
,
953
965
.
Rutledge
,
B. J.
,
Zhang
,
K.
,
Bier
,
E.
,
Jan
,
Y. N.
and
Perrimon
,
N.
(
1992
).
The Drosophila spitz gene encodes a putative EGF-like growth factor involved in dorsal-ventral axis formation and neurogenesis
.
Genes Dev
.
6
,
1503
1517
.
Sawamoto
,
K.
,
Okano
,
H.
,
Kobayakawa
,
Y.
,
Hayashi
,
S.
,
Mikoshiba
,
K.
and
Tanimura
,
T.
(
1994
).
The function of argos in regulating cell fate decisions during Drosophila eye and wing vein development
.
Dev. Biol
.
164
,
267
276
.
Schweitzer
,
R.
,
Howes
,
R.
,
Smith
,
R.
,
Shilo
,
B. Z.
and
Freeman
,
M.
(
1995a
).
Inhibition of Drosophila EGF receptor activation by the secreted protein Argos
.
Nature
376
,
699
702
.
Schweitzer
,
R.
,
Shaharabany
,
M.
,
Seger
,
R.
and
Shilo
,
B. Z.
(
1995b
).
Secreted spitz triggers the DER signaling pathway and is a limiting component in embryonic ventral ectoderm determination
.
Genes Dev
.
9
,
1518
1529
.
Sturtevant
,
M. A.
,
O’Neill
,
J. W.
and
Bier
,
E.
(
1994
).
Down-regulation of Drosophila Egf-r mRNA levels following hyperactivated receptor signaling
.
Development
120
,
2593
600
.
Sturtevant
,
M. A.
,
Roark
,
M.
and
Bier
,
E.
(
1993
).
The Drosophila rhomboid gene mediates the localized formation of wing veins and interacts genetically with components of the EGF-R signaling pathway
.
Genes Dev
.
7
,
961
973
.
Tio
,
M.
,
Ma
,
C.
and
Moses
,
K.
(
1994
).
spitz, a Drosophila homolog of transforming growth factor-alpha, is required in the founding photoreceptor cells of the compound eye facets
.
Mech. Dev
.
48
,
13
23
.
Wieschaus
,
E.
,
Perrimon
,
N.
and
Finkelstein
,
R.
(
1992
).
orthodenticle activity is required for the development of medial structures in the larval and adult epidermis of Drosophila
.
Development
115
,
801
811
.