In Situ Activation Pattern of Drosophila EGF Receptor Pathway During Development

See allHide authors and affiliations

Science  22 Aug 1997:
Vol. 277, Issue 5329, pp. 1103-1106
DOI: 10.1126/science.277.5329.1103


Signaling cascades triggered by receptor tyrosine kinases (RTKs) participate in diverse developmental processes. The active state of these signaling pathways was monitored by examination of the in situ distribution of the active, dual phosphorylated form of mitogen-activated protein kinase (ERK) with a specific monoclonal antibody. Detection of the active state of the Drosophilaepidermal growth factor receptor (DER) pathway allowed the visualization of gradients and boundaries of receptor activation, assessment of the distribution of activating ligands, and analysis of interplay with the inhibitory ligand Argos. This in situ approach can be used to monitor other receptor-triggered pathways in a wide range of organisms.

Receptor-tyrosine kinases participate in diverse biological processes and trigger, by way of Ras, a sequential activation of protein kinases called the mitogen-activated protein (MAP) kinase signaling cascade (1). Several RTKs have been identified inDrosophila. Whereas some RTKs control a single developmental decision, DER functions in numerous developmental processes (2-7). Regulation of DER activation is often achieved by the restricted processing of an activating ligand, Spitz (8, 9), and induction of the secreted inhibitory protein Argos (10).

MAP kinase (ERK) is activated by dual phosphorylation of threonine and tyrosine residues by MEK (1). A monoclonal antibody, termed diphospho-ERK (dp-ERK), was raised against a dually phosphorylated 11–amino acid peptide that constitutes the vertebrate ERK activation loop (11-13). All 11 residues are conserved in the single Drosophila ERK homolog Rolled (14), raising the possibility of cross-reactivity.

To study recognition specificity, we tested the antibody onDrosophila Schneider S2 cells expressing DER. Incubation with secreted Spitz (sSpitz) resulted in the detection of a single 44-kD polypeptide by dp-ERK antibody (Fig.1A). A general polyclonal antibody to ERK detected a similarly sized polypeptide in the presence or absence of Spitz. Immunohistochemical staining of induced cells detected activated ERK (Fig. 1B) (15).

Figure 1

Detection of dp-ERK in Schneider cells and embryos. (A) S2 cells expressing DER (S2:DER) were incubated with secreted Spitz (sSpitz). A 44-kD band was resolved by SDS-PAGE and detected with the anti–dp-ERK. The general antibody to ERK detected a similar molecular weight band with equal intensities before and after incubation with sSpitz. sSpitz was also induced in embryos containing K25 HS-Gal4/UAS-secreted Spitz(HS-sSpi). After heat shock, accumulation of the 44-kD band was detected by anti–dp-ERK, whereas the general antibody to ERK detected this band with or without sSpitz. (B) Immunohistochemical staining of S2 cells expressing DER showed accumulation of dp-ERK after incubation with sSpitz. (C) dp-ERK staining was monitored in Kr-Gal4/UAS-secreted Spitz embryos in which sSpitz was induced only in the central part of the embryo, and a uniform pattern was observed in this region (arrow shows anterior border ofKr domain; the posterior border is not seen in the image). Staining of a wild-type embryo at a similar stage (stage 11) with the general antibody to ERK showed a ubiquitous pattern throughout the embryo.

In embryos, activation of DER by ubiquitous sSpitz (HS-sSpi) resulted in the accumulation of a single 44-kD polypeptide detected by dp-ERK antibody (Fig. 1A). Immunohistochemical staining of embryos with the general antibody to ERK displayed ubiquitous staining throughout embryogenesis in accordance with the high maternal contribution of therolled gene. In contrast, dp-ERK displayed a specific staining pattern. Local activation of DER in the central segments led to staining only in this part of the embryo (Fig. 1C) (16).

Because ERK is a common junction for RTK pathways, the dp-ERK staining pattern represents the composite pattern of RTK signaling during development and can be correlated with activity of the known RTKs: Torso, DER, Heartless, and Breathless (17). Here we examine the dynamic DER-induced dp-ERK patterns in the embryo and imaginal discs.

Activation of DER is triggered by processing of Spitz, which is regulated by two membrane-spanning proteins, Rhomboid (Rho) and Star (8, 18, 19). Expression of Rho is tightly controlled (18, 20), and its ectopic induction leads to deregulated activation of DER (9,21). It was not known whether the initial expression of Rho in the neuroectoderm before gastrulation leads to DER activation, because no phenotypes in spitz, rho, orDER mutants were assigned to this phase. dp-ERK distribution during the early stages of embryogenesis was identical to the Rho expression pattern. This early zygotic dp-ERK pattern was absent in embryos mutant for spitz, rho, or DERgenes, demonstrating that it is generated by Spitz-induced DER activation (Fig. 2, A to C).

Figure 2

DER-dependent ERK activation during embryogenesis. (A) At stage 7, dp-ERK was detected in the ventral ectoderm (ve) but not in the future mesoderm (me). (B) At stage 8 after mesoderm invagination, staining persists in the ventral ectoderm (arrow) and was also observed in the dorsal ectoderm (arrowheads). (C) Once the initial contact of the mesoderm with the ectoderm is established, dp-ERK was also detected in the mesodermal cells immediately adjacent to activated ectodermal cells. This is presumably induced by diffusion of sSpitz to the mesoderm. The staining in the ventral ectoderm declines by stage 9. (D) At stage 8 dp-ERK was monitored in the ventral ectoderm, the dorsal-anterior and dorsal-posterior folds (df), the cephalic furrow, the head folds (hf), and the mesoderm. (E) Except for the mesodermal staining, all of the above aspects are induced by DER, because these patterns were absent inrho Δ 38 null mutants. However, the dorsal furrows (arrowheads) and cephalic furrow (cf) invaginated normally. (F) At stage 10, two prominent DER-dependent dp-ERK patterns were observed: the tracheal placodes (tp) and ventral ectoderm on both sides of the ventral midline (m). (G) Both of these patterns were eliminated in aspitz OE92 mutant embryo. (H) Higher magnification of a stage 10 embryo shows graded distribution of dp-ERK, in three to four cells on each side of the midline. (I) Inargos l Δ 7mutant embryos, this activation pattern was expanded to four to five rows of cells. (J) At stage 12, dp-ERK was detected in the four malpighian tubules (mt) (arrowheads). (K) Malpighian tubule staining intensity appears uniform and persists until stage 15. (L) The early malpighian tubule staining is DER-dependent, because it was absent in spi OE92mutants. (M) At stage 11, the founder cells of the chordotonal organs (c) can be detected in each segment byrho-lacZ expression (red). (N) Dp-ERK (green) is excluded from the founder cells and was observed in a ringlike structure surrounding them. (O) At stage 15, dp-ERK (green) was detected in the midline glial cells (mg) marked by the enhancer trap line AA142 (red). Staining in unidentified neurons (n) was also observed. (P) No dp-ERK was observed in acsw LE120 mutant embryos lacking both maternal and zygotic components. Shown is a stage 10 embryo.

At stages 7 to 9, the dp-ERK pattern is driven by Rho expression in the cephalic furrow, three head furrows, and the two dorsal folds. This pattern was absent in rho or DER mutants (Fig. 2, D and E), although no defects in furrow formation were detected. Whereas localized DER activation could provide the spatial and temporal cue to initiate furrows, another pathway may function in parallel.

The earliest known zygotic function of DER in patterning the ventral ectoderm at stage 9/10 is induced by Spitz, which is processed in the midline cells (9) to produce an active ligand. The extent of sSpitz diffusion and the levels of DER activation induced in the different ectodermal cell rows were not known. A graded dp-ERK pattern was observed in three to four cell rows on each side of the ventral midline, with the strongest staining displayed by cells adjacent to the midline (Fig. 2, F and H). This ectodermal staining was absent in spitz, rho, or DER mutant embryos (Fig. 2G), as well as in single minded mutant embryos that lack a functional midline. Thus, the midline is the only source of ligand for the ventral ectoderm at stage 9/10.

An inhibitory ligand, Argos, modulates the pattern of DER activation.argos transcription is induced by the DER pathway, thus constituting a negative-feedback loop. argos expression in the ventral-most cells, followed by diffusion, was postulated to terminate or reduce DER signaling in the ventrolateral cells (10). In argos mutant embryos an expanded pattern of dp-ERK was observed, with an intensity similar to that normally observed in the cells adjacent to the midline (Fig. 2I). This correlates with the expansion of ventral-most cell fates monitored in this mutant background (10). Thus, Spitz can diffuse up to five cell rows on each side of the midline.

At stage 11, DER-dependent dp-ERK staining was observed in the chordotonal organs. These stretch receptor organs in the peripheral nervous system are initially induced as clusters of three cells, with additional cells recruited by activation of the DER pathway (18, 22). This dp-ERK pattern was absent inrho mutant embryos. Double staining with arho-lacZ reporter demonstrated that dp-ERK was present in a ringlike structure surrounding the Rho-expressing cells (Fig. 2, M and N). This may be consistent with the production of sSpitz in the primary cluster in order to recruit additional cells through the DER pathway. It is not clear why the chordotonal cells that express Rho are refractory to ERK activation. DER-induced dp-ERK staining was also observed in the tracheal placodes and malpighian tubules (Fig. 2, F, J, and K). Finally, the DER pathway is essential for the development and viability of the midline glial cells (23). Prominent dp-ERK staining was observed at stage 15 in these cells (Fig. 2O).

The ability to directly monitor the pattern of dp-ERK in the different mutant backgrounds provides a rapid and powerful tool to place genes in the signaling cascade with respect to ERK. The corkscrew(csw) gene, encoding an SH2 (SRC homology 2)-containing tyrosine phosphatase (24), is a case in point. Althoughcsw was originally identified by an embryonic phenotype resembling that of torso, subsequent work has suggested that it may also be required for signaling by other RTKs (24). By monitoring the dp-ERK pattern in embryos lacking both maternal and zygotic csw, we have shown that all aspects of dp-ERK staining are diminished (Fig. 2P). Thus, Corkscrew is required for signaling by all RTKs and converges into the signaling pathway upstream of ERK.

A dynamic dp-ERK pattern identified in imaginal discs can be correlated with known DER functions. In the third-instar wing imaginal disc, prominent dp-ERK staining was detected in the future vein regions expressing Rho (25), including veins L3, L4, and L5, and in a transverse band of cells that will give rise to the wing margin. dp-ERK staining was also observed in the notum, along the posterior margin, and in the anterior mesopleura (Fig.3) (26).

Figure 3

Dp-ERK in the wing imaginal disc. Third instar wing imaginal disc from larvae heterozygous for anargos l(3)5959 enhancer trap, stained with anti–β-Gal (red) and dp-ERK (green). (A) Composite image. (B) Position of wing vein primordia (L) and wing margin (m) are marked. (C) In addition to prominent dp-ERK staining in the vein primordia, staining was also detected at several sites of Vein expression including the notum (n), lateral anterior mesopleura (mp), and posterior margin (pm). Vein is a putative DER ligand (27). No dp-ERK was detected in the inter-vein 3-4 region, which is also a prominent site of Vein expression. Anterior is at the top.

The interplay between activating and inhibiting DER ligands was monitored during development of the eye imaginal disc. At the morphogenetic furrow, prominent dp-ERK staining was observed in the newly formed photoreceptor clusters. Staining was also detected in clusters posterior to the first row, but at a lower intensity (Fig.4, A and C). During ommatidial differentiation, successive cycles of photoreceptor recruitment through the DER pathway take place. Activation of DER by sSpitz gives rise to the induction of Argos expression, preventing Spitz from activating DER in the more distal cells by a mechanism termed remote inhibition (5). When the production of sSpitz overcomes this inhibition, a new cycle of activation is initiated.

Figure 4

Dp-ERK in the eye imaginal disc. (A) Third-instar eye imaginal disc showed the most prominent dp-ERK staining in the row of photoreceptor clusters immediately posterior to the morphogenetic furrow (arrowhead). More posterior ommatidial rows showed less staining. (B) In discs derived from larvae homozygous for the hypomorphicargos W11 mutation, the intensity of dp-ERK in the posterior rows was comparable to that of the most anterior one. (C) Eye disc derived from a heterozygousargos l(3)5959 enhancer trap line. Dp-ERK (green) is most prominent at the first row. (D) Argos expression (red) was observed after the first row and may thus account for the lower dp-ERK intensity in all posterior rows. (E and F) Enlargements showing that at the time of formation of the ommatidia, dp-ERK staining is strong and centrally located. As differentiation proceeds, the more posterior ommatidia show exclusion of dp-ERK from the central cells expressingargos and activation in a ring around each ommatidium. The rings expand as the ommatidia mature. Anterior is to the left.

Only in the first photoreceptor row, sSpitz activates DER in the absence of Argos. Once argos is induced, its expression persists throughout all stages of photoreceptor, cone, and pigment cell differentiation (10). Double staining of dp-ERK and Argos showed that Argos is present only in the region with weak dp-ERK staining (Fig. 4D). The causal role of Argos in reducing the dp-ERK intensity was demonstrated in eye discs homozygous for the hypomorphicargos W11 mutation. In this case, the posterior rows also showed strong dp-ERK staining (Fig. 4B).

In addition to the reduction in dp-ERK intensity posteriorly, a parallel alteration in its spatial distribution within each ommatidial cluster was observed. Only one to three centrally located cells were labeled in each cluster in the anterior row. As ommatidia matured, staining diminished in the central cells and was observed in the periphery of each cluster (Fig. 4, E and F). The postulated reiterative cycles of DER activation in the eye disc (5) may thus be visualized as expanding concentric rings of dp-ERK.

In accordance with the multiple functions of DER during development, many of the dp-ERK patterns are attributed to DER activation. The temporal and spatial correlation of DER-induced dp-ERK to Rho expression stands out. The only exception is the activation of DER in the ventral ectoderm at stage 10 that is induced by Spitz-processing machinery restricted to the midline. These findings point to Rho as the limiting element in activation of the DER pathway. Different ranges of diffusion were observed for Spitz in different biological contexts, highlighting the importance of molecules that may restrict or facilitate ligand diffusion in regulating the spatial pattern of receptor activation.

  • * To whom correspondence should be addressed. E-mail: lvshilo{at}


Stay Connected to Science

Navigate This Article