Competing Activities of Heterotrimeric G Proteins in Drosophila Wing Maturation

Drosophila genome encodes six alpha-subunits of heterotrimeric G proteins. The Gαs alpha-subunit is involved in the post-eclosion wing maturation, which consists of the epithelial-mesenchymal transition and cell death, accompanied by unfolding of the pupal wing into the firm adult flight organ. Here we show that another alpha-subunit Gαo can specifically antagonize the Gαs activities by competing for the Gβ13F/Gγ1 subunits of the heterotrimeric Gs protein complex. Loss of Gβ13F, Gγ1, or Gαs, but not any other G protein subunit, results in prevention of post-eclosion cell death and failure of the wing expansion. However, cell death prevention alone is not sufficient to induce the expansion defect, suggesting that the failure of epithelial-mesenchymal transition is key to the folded wing phenotypes. Overactivation of Gαs with cholera toxin mimics expression of constitutively activated Gαs and promotes wing blistering due to precocious cell death. In contrast, co-overexpression of Gβ13F and Gγ1 does not produce wing blistering, revealing the passive role of the Gβγ in the Gαs-mediated activation of apoptosis, but hinting at the possible function of Gβγ in the epithelial-mesenchymal transition. Our results provide a comprehensive functional analysis of the heterotrimeric G protein proteome in the late stages of Drosophila wing development.


Introduction
G protein-coupled receptors (GPCRs) represent the most populous receptor family in metazoans. Approximately 380 non-olfactory GPCRs are encoded by the human genome [1], corroborated by ca. 250 GPCRs in insect genomes [2,3], making 1-1.5% of the total gene number dedicated to this receptor superfamily in invertebrates and mammals. GPCRs transmit their signals by activating heterotrimeric G protein complexes inside the cell. A heterotrimeric G protein consists of a GDP-bound a-subunit and a bc-heterodimer. Ligandstimulated GPCR serves as a guanine nucleotide-exchange factor, activating the GDP-to-GTP exchange on the Ga-subunit. This leads to dissociation of the heterotrimeric complex into Ga-GTP and bc, which transmit the signal further inside the cell [4].
The band c-subunit repertoire of the Drosophila genome is reduced as compared with that of mammals: only two Gc and three Gb genes are present in flies (Table 1). Gc30A and Gb76C are components of the fly phototransduction cascade and are mostly expressed in the visual system [5,6]. Gc1 and Gb13F have been implicated in the asymmetric cell divisions and gastrulation [7,8], while the function of Gb5 is as yet unknown.
Despite the fact that bc can activate signal effectors [9], the main selectivity in GPCR coupling and effector activation is provided by the Ga-subunits [10]. Sixteen genes for the a-subunits are present in the human genome, and six in Drosophila. All human Ga-subunit subgroups are represented in Drosophila (Table 1): Gai and Gao belonging to the Gai/o subgroup; Gaq belonging to the Gaq/11 subgroup; Gas belonging to the Gas subgroup, and concertina (cta) belonging to the Ga12/13 subgroup [10]. Additionally, Drosophila genome encodes for Gaf which probably represents an insect-specific subfamily of Ga-subunits [11].
Multiple functions have been allocated to different heterotrimeric G proteins in humans and flies [12], see Table 1. For example, in Drosophila development cta is a crucial gastrulation regulator [13], Gao is important for the transduction of the Wnt/Frizzled signaling cascade [14,15], and Gai controls asymmetric cell divisions during generation of the central and peripheral nervous system [7] (the later in cooperation with Gao [16,17]). Gaq is the Drosophila phototransduction Ga-subunit, but probably has additional functions [18]. Pleotropic effects arise from defects in Gas function [19], while the function of Gaf has not yet been characterized.
Among the developmental processes ascribed to the control by Gas are the latest stages of Drosophila wing development. Newly hatched flies have soft and folded wings, which during the 1-2 hours post-eclosion expand and harden through intensive synthesis of components of the extracellular matrix. These processes are accompanied by epithelial-mesenchymal transition and apoptosis of the wing epithelial cells, producing a strong but mostly dead adult wing structure [20,21,22]. Expression of the constitutively active form of Gas leads to precocious cell death in the wing epidermis, which results in failure of the closure of the dorsal and ventral wing sheets and accumulation of the hemolymph inside the wing, producing wing blistering [22,23]. Conversely, clonal elimination of Gas leads to autonomous prevention of the cell death. Kimura and co-workers have performed an extensive analysis of the signaling pathway controlling apoptosis at late stages of wing development [22]. They provide evidence suggesting that the hormone bursicon, synthesized in the head of post-eclosion Drosophila and secreted in the hemolymph, activates a GPCR rickets on wing epithelial cells, which signals through Gas to activate the cAMP-PKA pathway, culminating at the induction of apoptosis [22]. However, the identity and importance of the bc subunits in bursicon signaling, as well as possible involvement of other Ga proteins remained outside of their investigation. There also remain some uncertainties as to the phenotypic consequences of elimination of the bursicon-Gas-PKA pathway in wings [21,22,24].
Here we describe a comprehensive functional analysis of the Drosophila heterotrimeric G protein proteome using loss-of-function and overexpression experiments. We show that loss of Gas but not any other Ga-subunit leads to the failure of wing expansion after fly hatching. We also show that Gao, but not another Ga, can compete with Gas and thus antagonize its function. Finally, we identify the Gb13F and Gc1 as the bc subunits of the heterotrimeric Gs complex responding to the epithelial-mesenchymal transition and cell death-promoting signal.

Results
Gao, but not other Ga-subunits, in its GDP-loaded state prevented post-eclosion wing unfolding in Drosophila In the course of our studies of the role of the Gao subunit of heterotrimeric G proteins in the Wnt and PCP signaling in Drosophila wing development [15] we came across an observation that overexpression of Gao in Drosophila wings often led to the failure of wing expansion after fly hatching from the pupal case. Using the X-chromosome-located MS1096-Gal4 driver line, we found that ca. 80% of the aged adult female flies and 90% of male flies had folded wings characteristic of the freshly eclosed flies -a phenomenon never observed with wild-type animals (Fig. 1A, B, Table 2). MS1096-Gal4 drives strong expression in the dorsal domain and weaker expression in the ventral domain of the developing larval and pupal wing [25,26,27].
A similar overexpression of other Ga-subunits, Gas, Gai, or Gaq, did not produce this effect ( Fig. 1C-E), suggesting that Gao was unique in its ability to prevent wing expansion post-eclosion. Interestingly, the activated Q205L mutant form of Gao, which stays constantly bound to GTP [15,16], could not induce the folded wing phenotype (Fig. 1F). These data suggest that the GDP-, but not the GTP-loaded, form of Gao upon overexpression binds and sequesters a specific protein required for the proper posteclosion wing development.

Proteomic analysis identifies very few proteins discriminatively interacting with Gao in its GDP vs GTP form
In order to identify the protein(s) which might be sequestered by the overexpression of the wild-type (mostly GDP-loaded), but not the GTP-loaded form of Gao during post-eclosion wing expansion, we performed a proteomic analysis of Gao-binding partners which would bind specifically to its GDP-or its GTPloaded states, but not to both forms. To this end, we bacterially expressed wild-type Gao and immobilized it on CNBr-sepharose. These procedures resulted in Gao which was approximately 50%  [15]; control of asymmetric cell divisions in the sensory organ lineage [17]; feeding behavior [42]; learning and memory [43]; heart development [44,45]; axonal growth/guidance [44]; blood-brain barrier formation [46] Gai G protein ai subunit 65A control of the asymmetric cell divisions in the neuroblast and sensory organ lineages [7]; blood-brain barrier formation [46]; Hedgehog signal transduction [47] Gas G protein sa 60A 2 Gas G as (72%) larval growth [19]; establishment of the neuro-muscular synapse [48]; post-eclosion wing maturation [22] Gaf G protein a 73B 1 Gaf none (40% to Gas) none described Gaq (77%) phototransduction [18]; olfaction [31] concertina cta 1 Ga12/13 Ga13? (55%) gastrulation [13] Gb subunits Gb13F 1 Gb1 (83%) control of the asymmetric cell divisions in the neuroblast and sensory organ lineages [7]; gastrulation [7]; heart development [45] Gb76C Gbe 1 none? (43% to Gb1) phototransduction [5] Gb5 1 G b5 (68%) none described active as determined in the GTP-binding assays [16]. The matrix was then preloaded with GDP or GTPcS (a non-hydrolysable GTP analog) and used to apply Drosophila head extracts. After washing, proteins retained were eluted with Urea and resolved on SDS-PAGE ( Fig. 2A). We could identify three bands which bound preferentially to either nucleotide form of Gao: two in the GTPcSmatrix (ca. 53 kDa and 71 kDa), and one in the GDP-matrix (ca. 37 kDa). These findings could be confirmed by high resolution protein separation using 2D-PAGE with DIGE labeling [28]. The three proteins from Drosophila head extracts discriminatively bound to either nucleotide form of Gao in our experiment ( LC-MSMS after trypsin in-gel digestion was used to identify these three proteins. The 71 kDa protein was found to be the Heat-shock 70 kDa protein cognate 3 (gene name: Hsc70-3), the 53 kDa protein was identified as Tubulin b1-chain (gene name: b-Tubulin at 56D), and the 37 kDa protein exclusively binding to Gao-GDP -as the Guanine nucleotide-binding protein subunit b-1 (gene name: Gb13F subunit). While tubulins have previously been found to physically bind Ga-subunits [29,30], binding of Hsc70-3 to a G protein has not been reported before. As for the Gb13F subunit, the interaction of GDP-loaded Gao with the bc heterodimers is expected. However, we initially did not suspect that sequestration of bc by overexpressed Gao could be the reason for the wing unfolding defects, as other Ga-subunits would also be expected to sequester bc, and yet were ineffective in preventing wing unfolding (Fig. 1).  Gb13F and Gc1, but not other Gb/c subunits, are required for the post-eclosion wing unfolding To test whether the post-eclosion Gao-overexpression phenotype was due to sequestration of Gbc, we first aimed at rescuing the Gao phenotype by providing more bc. To this end, we co-expressed Gao, Gb13F, and Gc1 by the MS1096-Gal4 driver line. Indeed, we found an overwhelming rescue of the wing expansion defect if Gb13F/Gc1 were co-overexpressed: only 3% of aged female wings and 32% of the male wings now remained folded, as compared to 79% and 93% of female and male flies, respectively, overexpressing Gao alone (Table 2).
Next, to address the question whether Gbc heterodimers were necessary for the post-eclosion wing development, we expressed RNAi lines targeting Gb13F, Gb5, Gb76C, Gc1, or Gc30A by MS1096-Gal4. As shown on Fig. 3A-C, RNAi against Gc1, but not Gc30A, prevented wing expansion similarly to that induced by Gao overexpression (Fig. 1B). When RNAi lines targeting the three Gb-subunits were expressed, RNAi against Gb13F, but not Gb5 or Gb76C, was found to prevent wing expansion (Fig. 3D-F). Flies homozygous mutant for the Gb76C gene also showed no defects in wing development (data not shown). Other phenotypes of the downregulation of Gb13F and Gc1 suggested the role of this Gbc heterodimer in the process of asymmetric cell divisions [17], Wnt signaling [14], and planar cell polarity (not shown). Altogether, our results point to a simple model in which overexpression of the wild-type Gao, but not Gai, Gas, or Gaq, sequestered Gb13F/Gc1 required for the post-eclosion wing expansion in Drosophila.
Gb13F/Gc1 constitute with Gas the heterotrimeric G protein complex required for the post-eclosion wing expansion We supposed that Gao competed for Gb13F/Gc1 with another Ga-subunit, thus inactivating a heterotrimeric G protein complex required for the proper wing expansion. To investigate the nature of this Ga subunit outcompeted by Gao, we systematically removed all other Ga proteins by using loss-of-function mutations or targeted RNAi expression. RNAi-targeted downregulation was employed to target Gai, Gaq, Gaf, and Gas (Fig. 4A-F); of these constructs, those targeting Gaq and Gai were previously shown efficient in downregulating target gene expression [16,31]. Concertina was removed using the null allele [13]. Similar elimination of Gao is not possible due to the requirement of this G protein for cell viability in the wing [15]. Gao can be specifically uncoupled from GPCRs using the expression of pertussis toxin [29]; such whole wing expression of pertussis toxin does not result in any visible defects in wing expansion [16].
Out of all Ga tested, elimination of Gas from the wing produced the wing unfolding defect similar to that induced by overexpression of Gao or downregulation of Gb13F/Gc1 (Fig. 4E). In contrast, elimination of other Ga proteins in the wings did not produce visible defects (Fig. 4). Thus, we concluded that among different Ga-subunits only elimination of Gas led to the wing unfolding defect. In agreement with this, we found that cooverexpression of Gas together with Gao strongly suppressed the ability of the latter to produce the folded wing phenotype (Table 2). Thus, the heterotrimeric G protein complex, consisting of the Gas, Gb13F, and Gc1 subunits is required for the proper signaling regulating wing expansion post-eclosion, and can be antagonized by Gao.
The wing expansion defect is associated with, but is not caused by, prevention of cell death Clonal elimination of Gas results in failure of the cell death in the wing [22]. Indeed, while aged flies retained live GFP-and rhodamine phalloidin-stained cells only along the veins and wing margin (Fig. 5A, B), we found that the MS1096-Gal4-driven expression of Gao or RNAi constructs targeting Gb13F, Gc1, or Gas all similarly resulted in maintenance of live cells within the wing blade of well-aged flies (Fig. 5C-G). To better resolve the remaining live cells, we performed the nuclear staining with DAPI [22,24]. Young (ca. 1h-old) wild-type wings contain many DAPIpositive living cells (Fig. 5H), but aged wild-type wings showed DAPI staining only alone the veins (Fig. 5I). In contrast, wings of the Gao-overexpressing flies up to six days old were still filled with DAPI-positive living cells (Fig. 5J). These data clearly show that the wing expansion failure is associated with the failure of cell death. However, is prevention of the cell death sufficient to cause the folded wing phenotype? To investigate this possibility, we expressed the baculovirus apoptosis inhibitor p35 in the entire wing under the MS1096-Gal4 control. While apoptosis was efficiently prevented, wing expansion was normal in these wings (Fig. 5K). This data agrees with the similar observations obtained when p35 was expressed using other Gal4 drivers [21,22]. Cumulatively, our data suggest that apoptosis, being an important process during post-eclosion wing maturation, is unlikely to be the sole driving force behind wing expansion. Wing expansion seems more dependent on the epithelial-mesenchymal transition [21,24], or perhaps requires both processes to act in concert. Elimination of the components of the heterotrimeric Gs proteins apparently leads to both the failure of epithelial-mesenchymal transition and apoptosis, leading cumulatively to the wing expansion defect.

Overactivation of Gas by cholera toxin mimics expression of the constitutively active mutant form of Gas, not reproduced by overexpression of Gb13F/Gc1
Expression of the GTPase-deficient point mutant of Gas induces precocious cell death, which results in hemolymph accumulation between the two epithelial wing sheets and wing blistering [22,23], Fig. 6A. In mammalian systems Gas can be overactivated by cholera toxin, which covalently ADP-ribosylates a conserved arginine residue of the GTPase active center [32]. To test whether cholera toxin was also active against Drosophila Gas, we expressed the toxin in developing Drosophila wings, and found wing blistering induced by the toxin (Fig. 6B) similar to that induced by the constitutively activated Gas (Fig. 6A). These data not only extend the known similarity between mammalian and fly Gas, but they also demonstrate that targeted activation of the endogenous, not overexpressed, Gas is sufficient to overactivate the pathway and produce wing blistering.
Cholera toxin-mediated activation of Gas mimics that achieved by GPCR-mediated activation and results in production of GTPloaded Gas and free Gbc subunits. As the latter can induce signal transduction in some instances [9], we investigated the effects of direct co-overexpression of Gb13F/Gc1 in Drosophila wings using a number of Gal4 drivers lines. Gb13F or Gc1 subunits expressed alone were ineffective in inducing phenotypes (Fig. 6C, D). Despite the fact that co-overexpression of Gb13F and Gc1 could affect asymmetric cell divisions [17], Wnt/Frizzled signaling [14], planar cell polarity (data not shown), and venation (Fig. 6E), Gb13F/Gc1 was in no condition able to mimic the wing blistering phenotype induced by activation of Gas (Fig. 6E). We also boosted Gb13F/ Gc1 overexpression by combining two copies of the UAS-Gb13F, UAS-Gc1 transgenes, as well as by providing two copies of the Gal4 driver lines; these attempts also failed to produce the wing blistering phenotype. These results demonstrate that the Gbc heterodimer is required for the proper Gas signaling, but by itself plays only the passive, permissive role in the signal transduction leading to apoptosis.
To further prove that Gbc is not necessary for the execution of the apoptosis program once the activated Gas is released, we coexpressed Gas[GTP] with the wild-type Gao sequestering the Gbc subunits. We found that the potency of Gas[GTP] to induce wing blistering was not at all affected by such Gbc sequestration (Fig. 6F).
However, Gbc might have a separate function in the Gs signaling, namely the induction of the epithelial-mesenchymal transition sub-pathway. Indeed, while co-overexpression of Gas is capable of rescuing the folded wing phenotype induced by overexpression of Gao, the constitutively activated form of Gas is much less potent in performing such a rescue (Table 2). These data suggest that it is not the GTP-loaded Gas, but the free Gbc heterodimer, released from the heterotrimeric Gs complex upon rickets or other GPCR receptor activation, which is required for the epithelial-mesenchymal transition and wing expansion. This issue is further discussed in the next section.

Discussion
The soft folded wings of the young insect freshly hatched from the pupal case within 1-2 hours expand and harden, becoming a robust flight organ. This process is accompanied by epithelialmesenchymal transition and cell death of the wing epithelial cells [20,21]. Genetic dissection has revealed the function of the neurohormone bursicon and its wing epithelial receptor rickets in initiation of these processes [21,22,24]. The GPCR rickets couples to the heterotrimeric G protein Gs; the Gas-activated cAMP-PKA pathway culminates at the induction of apoptosis [22]. However, the overall phenotypic consequences of the loss of the Gs signaling pathway in post-eclosion wings were unknown, as well as the nature of the Gbc subunits of the heterotrimeric Gs complex responding to the bursicon-rickets signaling.
Here we have performed an extensive analysis of the heterotrimeric G protein subunits in these post-eclosion stages of wing maturation. We find that the whole-wing down-regulation of Gas results in the failure of wing expansion, demonstrating that this change in the shape of the wing is the major morphological outcome of the bursicon-rickets-Gs signaling. We also identify the Gb13F and Gc1 subunits as the other two constituents of the heterotrimeric Gs complex, as downregulation of Gas, Gb13F, or Gc1, but not any other Ga, Gb, or Gc subunits encoded by the Drosophila genome, each leads to the same folded wing phenotype.
We also show that Gao, but not any other Ga-subunit, can inhibit the wing expansion program through sequestration of the Gb13F/Gc1 heterodimer. The reason for the specificity of Gao over other Ga-subunits in antagonizing the Gs signaling is unclear. It is unlikely that differences in expression levels of the tested Gasubunits may account for the selective activity of Gao. Indeed, most overexpression experiments were done with the X-chromosomeinserted MS1096-Gal4 driver, which results in markedly higher expression levels in males than heterozygous female flies, producing a more penetrant folded wing phenotype in males overexpressing Gao (see Table 2). However, even in male flies overexpressing other Ga-subunits no instances of the folded wing phenotype could be seen. Furthermore, several independent insertions of the UAS-Ga transgenes were tested; while different Gao transgenes all produced the folded wing phenotype upon overexpression, other Ga constructs remained ineffective (data not shown).
Similarly, the different Ga-subunits possess a similar affinity towards the interaction with the Gbc heterodimer [33,34], not providing an explanation for a specific ability of Gao to antagonize the Gs-mediated post-eclosion pathway. We are thus tempted to propose that a previously uncharacterized biochemical mechanism may allow for a specific antagonism physiologically existing between the Gs-and Go-mediated signaling pathways. As liberation of high amounts of GDP-loaded Gao is predicted to be a consequence of activation of multiple Go-coupled GPCRs [33], and as Go is a heavily expressed G protein representing the major G protein species e.g. in the brain of flies and mammals [35,36], this specific ability of Gao to antagonize the Gs-mediated signaling may have physiological implications in other tissues and organisms than Drosophila wing. However, we would like to add that these speculations are based on the analysis of the overexpression data and must be treated with caution when translating them into physiological situations.
Only the GDP-loaded, but not the activated GTP-loaded form of Gao is effective in antagonizing Gs. We have performed a proteomics analysis of the Drosophila proteins which would discriminate between the two nucleotide forms of Gao, and revealed surprisingly few targets of this kind. While the chaperone Hsc70-3 and b1-tubulin preferentially interacted with the GTP-loaded Gao, Gb13F was found to specifically interact with Gao-GDP. These data suggest that many Gao-interaction partners do not discriminate between the two guanine forms of Gao. These findings are in agreement with our other experimental findings [16], as well as our mathematical modeling predicting that high concentrations of free (monomeric) signaling-competent Gao-GDP are produced upon activation of Go-coupled GPCRs [33].
Gao-mediated sequestration of Gb13F/Gc1 depletes the pool of the heterotrimeric Gs complexes. As only heterotrimeric Gabc, but not monomeric Ga proteins can efficiently bind and be activated by their cognate GPCRs [4,34], overexpression of Gao abrogates the rickets-Gs signaling. Phenotypic consequences of this abrogation are the failures of apoptosis and wing expansion. In contrast, expression of the constitutively activated form of Gas induces premature cell death and wing blistering [22,23]. We find that this phenotype can be also induced by expression of cholera toxin, revealing that the ability of cholera toxin to specifically overactivate Gas reported in mammalian systems [32] is reproduced with Drosophila proteins. These data also confirm that not only exogenously overexpressed, but also the endogenous Gas can induce the precocious cell death upon overactivation.
However, prevention of apoptosis is not sufficient to produce the folded wing phenotype (Fig. 5). Together with the observation that the constitutively active form of Gas is ineffective in rescuing the wing expansion defects produced by Gao overexpression (Table 2), these data suggest that the Gas-cAMP-PKA pathway culminating at apoptosis is not the sole signaling branch emanating from the bursicon-rickets GPCR activation. We propose that the second signaling branch initiated by the rickets-mediated dissociation of the heterotrimeric Gs complex is represented by the free Gbc subunits, signaling to epithelial-mesenchymal transition (Fig. 7). Such a double signaling impact mediated by the two components of the heterotrimeric G protein complex leads to initiation of two cellular programs -apoptosis and epithelialmesenchymal transition -which cumulatively result in wing expansion and solidification (Fig. 7), producing the adult flight organ. This two-fold response of the Drosophila wing to the maturation signal, mediated by the two components of the heterotrimeric G protein complex activated by the single hormone-responsive GPCR, provides an elegant paradigm for the coordination of signaling and developmental programs.

Histology
All wings were prepared from $1 day-old flies and mounted in GMM as described [15]. For GFP, as well as for rhodamine phalloidin (Molecular Probes) visualization after treatment as described for imaginal discs [15], wings were mounted in Moviol. Whole young flies (#1 day-old) were photographed through a Zeiss Stemi 2000 binocular using the Canon PowerShot G10 camera to visualize wing blistering. DAPI staining was performed after [21] with the following modifications: after fixation in 4% formaldehyde, the wings were successively treated at 17uC with chloroform for 1 h, heptane for 2 h, and 16PBS/0.2% Tween 20 for 3 h, prior to the overnight DAPI (1:10000, Sigma) staining at 17uC. These modifications aimed at increasing accessibility of DAPI to the folded aged wings.
Fifty mg of the precipitated proteins were labeled with CyDye DIGE Fluor minimal dyes according to the manufacturer recommendations (GE Healthcare Life Sciences). The samples were cup-loaded onto 24 cm pH 3-11 IEF strips and electrofocused with a total of 459000 Vh using an Ettan IPGphor II (both GE Healthcare Live Sciences). The strips were reduced and alkylated according to the manufacturer recommendations. The second dimension separation was performed on 10-15% linear gradient gels automatically casted using a2DEoptimizer (NextGen Sciences) and the gels were run in the Ettan Dalt II (GE Healthcare Live Sciences) at 25uC. The gels were scanned using a Typhoon 9200 scanner (GE Healthcare Life Sciences). The gel images were analyzed using SameSpots (Nonlinear Dynamics) involving automatic normalization and automatic background substraction.
After subsequent Coomassie staining, spots of interest were picked using GelPal (Genetix) and digested overnight at 37uC (19 ng trypsin (Promega) in 47 mM Tris pH 9.0). The peptides were analyzed using LC-MSMS (4000 Q TRAP, Applied Biosystems) and proteins were identified using Mascot (Matrix Science) searching the protein sequence database UNIPROT-15.6. Figure 7. A model of the action of components of the heterotrimeric Gs complex in wing maturation. The neurohormone bursicon acts on the Gs-coupled GPCR rickets expressed in wing cells. The GPCR activity leads to dissociation of the heterotrimeric Gs complex into GTP-loaded Gas and free Gbc-heterodimer. Gas [GTP] activates the cAMP-PKA pathway to promote apoptosis. Gbc, on the other hand, acts to induce the epithelial-mesenchymal transition. These two processes, acting in coordination, lead to post-eclosion wing expansion and solidification. Expression of the constitutively active Gas or cholera toxin stimulates the Gas-dependent branch in this signaling. Expression of Gao inhibits this signaling through sequestration of the Gbc-subunits. doi:10.1371/journal.pone.0012331.g007