Essential Roles for Ecdysone Signaling During Drosophila Mid-Embryonic Development

See allHide authors and affiliations

Science  26 Sep 2003:
Vol. 301, Issue 5641, pp. 1911-1914
DOI: 10.1126/science.1087419


Although functions for the steroid hormone ecdysone during Drosophila metamorphosis have been well established, roles for the embryonic ecdysone pulse remain poorly understood. We show that the EcR-USP ecdysone receptor is first activated in the extraembryonic amnioserosa, implicating this tissue as a source of active ecdysteroids in the early embryo. Ecdysone signaling is required for germ band retraction and head involution, morphogenetic movements that shape the first instar larva. This mechanism for coordinating morphogenesis during Drosophila embryonic development parallels the role of ecdysone during metamorphosis. It also provides an intriguing parallel with the role of mammalian extraembryonic tissues as a critical source of steroid hormones during embryonic development.

Ecdysone triggers the programmed cell death of obsolete larval tissues and signals cell shape changes in the imaginal discs during Drosophila metamorphosis, transforming the body plan of the insect from a crawling larva into an adult fly. (“Ecdysone” in this paper refers to physiologically active ecdysteroids.) Ecdysone signaling is mediated by a heterodimer of two nuclear receptors, EcR (NR1H1) and the Drosophila RXR ortholog USP (Ultraspiracle, NR2B4) (1). Characterization of EcR and usp mutants at the onset of metamorphosis has revealed similar lethal phenotypes, indicating that these factors act together to transduce the hormone signal (1). A high titer ecdysone pulse also occurs midway through embryonic development, peaking during germ band retraction (GBR) (2). GBR, dorsal closure (DC), and head involution (HI) compose the major morphogenetic movements that form the body plan of the first instar larva. Ecdysone is required for cuticle deposition during late embryogenesis (3, 4), but functions for ecdysone at earlier stages of embryogenesis have been difficult to assess because ecdysteroids, as well as EcR mRNA and protein, are deposited maternally (3, 5). EcR zygotic null mutants die during embryogenesis with minor cuticular defects (6), and germline clones of EcR null mutants arrest during oogenesis, resulting in female sterility (7).

Immunofluorescent staining with common region and isoform-specific antibodies revealed that both EcR and USP are widely expressed throughout embryogenesis (fig. S1). To investigate when these receptors are activated, we used the GAL4-LBD (ligand binding domain) system, which accurately reflects ecdysone signaling through EcR-USP at the onset of metamorphosis (8). hs-GAL4-EcR or hs-GAL4-USP was combined with a GAL4-dependent nuclear lacZ reporter gene (UAS-nlacZ) to determine when and where the GAL4-LBD fusion protein can be activated by its ligand in the embryo. No activation was seen at stage 11, when the germ band is completely extended and the ecdysone titer is low (Fig. 1, A and D) (2). Activation was first evident at stage 12, when GBR begins, and was high at stage 13 (Fig. 1, B and E), when the ecdysone titer peaks, the germ band is retracted, and the dorsal surface of the embryo is covered by the amnioserosa. GAL4-EcR and GAL4-USP displayed similar patterns of activation, first detectable in the amnioserosa and remaining high in this tissue through the completion of DC at stages 14 and 15 (Fig. 1, C and F). More widespread activation was seen at later stages of embryogenesis (Fig. 1, C and F) (9). Endogenous EcR-USP activation was determined using a multimerized hsp27 Ecdysone Response Element driving lacZ expression (7xEcRE-lacZ) (5), although other nuclear receptors can also bind to this EcRE. The pattern of 7xEcRE-lacZ expression was similar to that of GAL4-EcR and GAL4-USP in the amnioserosa and was detectable in other tissues at stages 13 and 14 (Fig. 1, G to I). These observations are consistent with the specific and transient high-level expression of another ecdysone-dependent reporter, DR3-Fbp-1-lacZ, in the amnioserosa at stages 12 and 13 (10). As a control, we used antibodies against GAL4 to show that GAL4-EcR and GAL4-USP are widely expressed in both the amnioserosa and epidermis at stages 11 to 13. In addition, hs-GAL4 can drive strong uniform expression of β-galactosidase at stage 9, which persisted into stages 11 to 15 (9). Finally, embryos cultured with exogenous 20-hydroxyecdysone, a physiologically active ecdysteroid, displayed efficient and widespread 7xEcRE-lacZ expression at stage 9 (Fig. 2), when EcR-USP is normally silent (Fig. 1). Taken together, these observations indicate that the amnioserosa contains a high concentration of active ecdysteroids during mid-embryogenesis.

Fig. 1.

The ecdysone receptor is activated in the amnioserosa. Histochemical staining for β-galactosidase expression in heat-treated hs-GAL4-EcR/+; UAS-nlacZ/+ (A to C), hs-GAL4-USP/+; UAS-nlacZ/+ (D to F), or 7xEcRE-lacZ (G to I) embryos is shown. Activation was not detectable at stage 11 (A, D, and G), before the ecdysone pulse, but was strong in the amnioserosa at stage 13 [(B), (E), and (H), arrows], persisting through stages 14 and 15 [(C), (F), and (I), arrows]. Longer histochemical staining of these later stage embryos revealed widespread activation outside of the amnioserosa (9). All embryos are oriented anterior to the left and dorsal side up, except for (B), (E), and (H), where the embryos are shown from a dorsal-lateral perspective. The activation pattern in the amnioserosa of stage 13 7xEcRE-lacZ embryos was most intense at the anterior and posterior ends of this tissue [(H), arrow points to posterior patch], similar to the pattern of GAL4-USP [(E), arrow], but more pronounced. Activation in the amnioserosa was uniform at stage 14 and 15 in all genotypes [(C), (F), (I), arrows]. Patterns shown were consistent among several hundred embryos examined.

Fig. 2.

Widespread activation of the ecdysone receptor in embryos cultured with 20-hydroxyecdysone (20E). Zero- to 4-hour 7xEcRE-lacZ embryos were cultured with 5 × 10–6 M 20E or with vehicle control (1% ethanol) and histochemically stained for β-galactosidase expression. Sixty percent (n = 67) of the embryos cultured in the presence of 20E showed premature and widespread 7xEcRE-lacZ expression at stages 9 to 11, ranging from (B) moderate to (C) high expression. (A) Control embryos showed no activation (n = 88). Embryos are oriented anterior to the left and dorsal side up.

The amnioserosa is an extraembryonic tissue that does not contribute to the embryo proper, but which is nevertheless essential for both GBR and DC (1113). To test whether ecdysone signaling is required for these morphogenetic events, we inactivated both maternal and zygotic EcR functions by expressing either of two different dominant-negative EcR constructs during early embryogenesis: hs-GAL4-EcR or UAS-EcR-F645A (8, 14). As an independent means of disrupting ecdysone signaling, we used a temperature-sensitive ecdysone-deficient mutant, ecd1 (15), under conditions where both maternal and zygotic ecdysone levels are reduced. Expression of an EcR dominant-negative construct, or maintaining ecd1 mutants at a nonpermissive temperature, resulted in embryonic lethality (Fig. 3A) with highly penetrant defects in GBR, HI, and cuticle deposition (Fig. 3B). Weak GBR defects included malformed mouthhooks and a dorsal shift in the position of the posterior spiracles (Fig. 3D), whereas stronger phenotypes included a complete block in GBR and an anterior cuticular hole, indicating defective HI (Fig. 3E). No ecd1 mutant embryos arrested development before GBR, suggesting that this is the earliest embryonic function for ecdysone. Embryonic lethality caused by heat-treating hs-GAL4-EcR/+ embryos at 3 to 5 hours after egg laying increased from 60 to 96% in a heterozygous EcR mutant background, indicating that this dominant-negative construct interfered with endogenous EcR function. Both the penetrance and expressivity of the GBR defect also increased under these conditions (16). To determine whether ecdysone signaling is required in the amnioserosa and/or the germ band for GBR, we expressed UAS-EcR-F645A under the control of the amnioserosa-specific GAL4 drivers C381 and LP-1 or the ectodermal-specific GAL4 driver 69B. These crosses led to embryonic and first instar larval lethality with low-penetrance GBR defects (9). Upon repeating these experiments in a sensitized EcR genetic background, 23 and 15% of the embryos that express UAS-EcR-F645A under control of 69B-GAL4 or C381-GAL4, respectively, displayed GBR defects; no GBR defects were detected in control embryos (16). Thus, ecdysone signaling appears to be required in both the amnioserosa and germ band for GBR to proceed. Mutations in several genes that are required for GBR (Egfr, hnt, srp, and ush) cause premature amnioserosa cell death (11, 12, 17). We observed a similar effect when ecdysone signaling was downregulated; however, GBR defects were observed at the same penetrance when we expressed the GAL4-EcR dominant negative in a Df(3)H99 mutant background to block embryonic cell death (16). Therefore, as was shown in studies of the amnioserosa-specific Hindsight transcription factor (18), more than amnioserosa cell survival is required for GBR to occur.

Fig. 3.

Ecdysone signaling is required for germ band retraction, head involution, and cuticle deposition. (A and B) w1118 (control), hs-GAL4-EcR, or hs-GAL4/+; UAS-EcR-F645A/+ (EcR-F645A) embryos were maintained at either 25°C (–heat) or heat-treated at 3 to 5 hours after egg laying (+ heat). Embryos from ecd1 mutant flies maintained at either the permissive (–heat) or nonpermissive (+ heat) temperature were also analyzed for either lethality (A) (n = 200 to 300 in at least two independent experiments) or embryonic lethal phenotypes (B) (n = 100 to 120) (16). (C to E) The lethal phenotypes of embryos expressing the GAL4-EcR dominant-negative range from relatively mild defects in GBR and HI (D) to a strong block in these morphogenetic pathways (E), in comparison to a control w1118 embryo (C).

Earlier studies have demonstrated that the amnioserosa promotes GBR through integrin-mediated adhesion to the adjacent germ band, with the highest penetrance GBR defects observed in mutants lacking both maternal and zygotic βPS integrin, encoded by myospheroid (mys) (19, 20). Both the penetrance and expressivity of GBR defects were greatly enhanced when the hs-GAL4-EcR dominant negative was expressed in embryos that carry a reduced dose of mys, but not in control embryos, suggesting that ecdysone exerts its effects on GBR, at least in part, through integrins (Fig. 4).

Fig. 4.

Functional interaction between ecdysone signaling and βPS integrin in germ band retraction. Embryos from the respective crosses (16) were maintained at either 25°C or heat-treated at 3 to 5 hours after egg laying. The penetrance of total GBR defects (weak + strong) and strong GBR defects (fully extended germ band) are depicted. Forty to 80 embryos were examined for each genotype. A single copy of the GAL4-EcR dominant negative construct alone [+/GAL4-EcR(II)] generated relatively weak GBR defects in this experiment because two copies of the transgene are required to produce the strong phenotypes shown in Fig. 3.

The embryonic ecdysone titer begins to rise before the formation of the ring gland, the endocrine organ of the insect, suggesting that maternally deposited ecdysteroids contribute to the embryonic ecdysone pulse (2, 3). This is consistent with the maternal effect that we observe in our ecd1 studies, where a cross of ecd1 mutant females to wild-type males results in a high degree of GBR and HI defects in the offspring (16). Maternal ecdysteroids are stored as inactive conjugates in the yolk, which lies in close apposition to the amnioserosa (21, 22). We propose that these inactive ecdysteroids are converted into active forms of the hormone by enzymes that reside in the yolk and/or amnioserosa, resulting in a high concentration of active ecdysteroids in the amnioserosa, defining this tissue as a critical source of the hormone.

Mutations in dib and sad, genes that encode key cytochrome P-450s in the zygotic ecdysone biosynthetic pathway, have established a role for ecdysone signaling in HI, DC, and cuticle deposition (4, 23). Although we do not see penetrant DC defects when we disrupt ecdysone signaling, it is possible that this phenotype is masked by the earlier defect in GBR. Indeed, 5 to 8% of ecd1 mutant embryos, or embryos expressing EcR-F645A, display an anterior-dorsal hole that could be indicative of a DC defect (16). Alternatively, the absence of GBR defects in dib and sad mutants may be attributed to their exclusively zygotic function and the apparent dependence of GBR on maternally contributed ecdysteroids. Finally, although our functional studies support a key role for EcR in mid-embryonic morphogenetic movements, reduction of maternal and zygotic usp function results in embryonic lethality with minor cuticle defects and no reported defects in morphogenesis (24, 25). This could be due to the hypomorphic nature of two usp alleles used in these studies (26). Alternatively, EcR might exert its roles in embryonic morphogenesis independently of USP.

Two models have been proposed to explain the role of the amnioserosa in GBR, either as a source of one or more diffusible signals that trigger cell shape changes in the adjacent germ band (18) or through dorsoventral contraction of the amnioserosa promoting GBR via integrin-mediated adhesion between these tissues (19, 22). Our results unify these models by implicating ecdysone as a critical signal from the amnioserosa that is required for GBR, as well as suggesting that ecdysone directs this response through integrins. The observation that ecdysone signaling is required in both the amnioserosa and germ band for GBR suggests that the amnioserosa may not provide all of the force for this morphogenetic event and that ecdysone-triggered cell shape changes in the germ band may also contribute to GBR. Moreover, analogous ecdysone-triggered cell shape changes drive imaginal disc morphogenesis during metamorphosis (22, 27), with at least part of this genetic program required for GBR and HI during embryonic development (28). Genetic interactions between EcR and integrin subunits, including βPS integrin, have also been observed during adult wing morphogenesis during metamorphosis (29). Taken together, these observations raise the intriguing possibility that ecdysone functions in a parallel manner during embryogenesis and metamorphosis, triggering coordinated changes in cell shape that establish the basic body plan for the next phase in the life cycle. An endocrine signal in embryos also provides a mechanism to explain the temporal coordination of mid-embryonic morphogenetic events, defining a new ecdysone-dependent phase in the insect life cycle. Finally, this study provides a parallel with the function of mammalian extraembryonic tissues, where the placenta acts as a critical source of steroid hormones for the maintanence of pregnancy.

Supporting Online Material

Fig. S1

Materials and Methods


References and Notes

View Abstract

Navigate This Article