Bruce suppresses autophagy-regulated caspase activity and wing tissue growth in Drosophila

Structurally, the executioner caspases, Dcp-1 and Drice, are similar, each comprising catalytic P20 and P10 domains and a highly disordered N-terminal prodomain (Figure 1A). First, we assessed the ability of two executioner caspases to trigger apoptosis. Using the WP-Gal4 driver, which is expressed in the wing pouch region of the wing imaginal disc (Obata et al., 2014), overexpression of Dcp-1 resulted in a wingless phenotype whereas the overexpression of Drice did not (Figure 1B). In wing imaginal discs, Dcp-1 overexpression induces nuclear condensation (Figure 1C), indicative of apoptotic cell death. To further characterize the type of cell death, we examined whether overexpressed Dcp-1 is activated through cleavage. By western blotting, we found that Dcp-1 is cleaved upon overexpression (Figure 1—figure supplement 1A). We also found that simultaneous overexpression of the caspase inhibitor p35 suppressed Dcp-1 cleavage (Figure 1—figure supplement 1A), suggesting that Dcp-1 is cleaved by caspases. Cleavage of overexpressed Dcp-1 is further confirmed by anti-cleaved Dcp-1 antibody staining in wing imaginal discs (Figure 1—figure supplement 1B), indicating that overexpressed Dcp-1 is cleaved and thereby activated. To evaluate whether overexpressed Dcp-1 shows typical executioner caspase activity, we employed two executioner caspase activity probes: GC3Ai (Schott et al., 2017; Zhang et al., 2013) and CD8::PARP::VENUS (Williams et al., 2006). GC3Ai becomes fluorescent upon caspase-mediated cleavage (Figure 1—figure supplement 1C). In wing imaginal discs, Dcp-1 overexpression resulted in strong GC3Ai fluorescence, suggesting that it exerts typical executioner caspase activity (Figure 1—figure supplement 1D). CD8::PARP::VENUS generates a neo-epitope that can be detected by anti-cleaved PARP antibody upon caspase-mediated cleavage (Figure 1—figure supplement 1E). Consistently, in wing imaginal discs, Dcp-1 overexpression led to a strong anti-cleaved PARP signal, further supporting that it induces typical executioner caspase activity (Figure 1—figure supplement 1F). Consistent with these results, Dcp-1 overexpression resulted in TUNEL-positive cell death, reflecting DNA strand breaks that are a hallmark of apoptosis (Figure 1D). Overall, these results suggest that Dcp-1 overexpression is sufficient to induce typical executioner caspase activity-dependent apoptotic cell death, consistent with our previous observations in olfactory receptor neurons (Muramoto et al., 2025).

Figure 1 with 2 supplements see all

Download asset Open asset

Figure 1—source data 1

Data used for graphs presented in Figure 1A, H.

https://cdn.elifesciences.org/articles/109876/elife-109876-fig1-data1-v1.xlsx

Download elife-109876-fig1-data1-v1.xlsx

Next, we examined the apoptotic potential of fluorescently tagged caspases in Drosophila S2 cells expressing mNeonGreen (mNG)-fused Dcp-1 or Drice. Dcp-1 expression caused a marked reduction in mNG intensity (Figure 1—figure supplement 2A–C) and a decreased number of mNG-positive cells (Figure 1—figure supplement 2A, D). These effects required Dcp-1 catalytic activity, as the mutation of the catalytic cysteine to glycine restored both mNG intensity and cell number (Figure 1—figure supplement 2A–D). In contrast, Drice expression neither reduced mNG intensity nor decreased the number of mNG-positive cells, resembling the catalytically inactive mutant (Figure 1—figure supplement 2A–D). Taken together, these results demonstrate that Dcp-1, but not Drice, is activated upon overexpression and induces apoptosis in multiple cell types.

Executioner caspases are typically activated by cleavage mediated by the initiator caspase, Dronc, which is activated through the assembly of the Dark-dependent apoptosome (Figure 1E; Nakajima and Kuranaga, 2017; Shinoda and Miura, 2024). We previously proposed that Dcp-1 can be activated independently of Dark and Dronc (Shinoda et al., 2019). To test this hypothesis further, we analyzed genetic modifiers of the Dcp-1-induced wingless phenotype. A heterozygous mutation in the H99 deficiency, which encompasses the pro-apoptotic reaper, hid, and grim (RHG) genes, rescued the wingless phenotype caused by Dcp-1 overexpression (Figure 1F, H). Similarly, expression of the caspase inhibitor p35 or knockdown of RHG genes also rescued this phenotype (Figure 1G, H), suggesting that Dcp-1 induces apoptosis in a catalytic activity-dependent manner under the control of RHG proteins. In contrast, knockdown of other core apoptotic components, including Dark, Dronc, and Drice, or overexpression of the dominant-negative form of Dronc (Dronc^DN) failed to suppress the wingless phenotype (Figure 1G, H). These findings support our earlier observation that Dcp-1 can be activated independently of the canonical apoptosome composed of Dronc and Dark to mediate non-lethal functions (Shinoda et al., 2019). Collectively, these results indicate that, at least under overexpression conditions, Dcp-1 activation proceeds independently of the apoptosome, in contrast to that of Drice.

Proximal proteins of caspases have emerged as potential regulators of executioner caspase activity (Muramoto et al., 2025). We have previously demonstrated that the proximal protein patterns of Dcp-1 and Drice differ even within the same tissue, such as wing imaginal discs (Shinoda et al., 2019). Based on this, we sought to identify Dcp-1-specific regulatory factors among its proximal proteins. First, we examined the tissue distribution of Drosophila caspases using previously established Caspase::V5::TurboID knock-in fly lines (Shinoda et al., 2019). Expression patterns were analyzed in larval tissues, including the salivary glands, fat bodies, and brains (Figure 2—figure supplement 1A), with confirmed expression in wing imaginal discs (Figure 2A) as well as in adult tissues, including the thoraxes, ovaries, and testes (Figure 2—figure supplement 1B). Drice was ubiquitously expressed in all examined tissues. In contrast, Dcp-1 expression was restricted to the larval brains, imaginal discs, and adult ovaries and testes.

Figure 2 with 3 supplements see all

Download asset Open asset

Figure 2—source data 1

Uncropped raw blot images of Figure 2A, D, E.

https://cdn.elifesciences.org/articles/109876/elife-109876-fig2-data1-v1.zip

Download elife-109876-fig2-data1-v1.zip

Figure 2—source data 2

Uncropped annotated blot images of Figure 2A, D, E.

https://cdn.elifesciences.org/articles/109876/elife-109876-fig2-data2-v1.zip

Download elife-109876-fig2-data2-v1.zip

Figure 2—source data 3

Data used for graphs presented in Figure 2I.

https://cdn.elifesciences.org/articles/109876/elife-109876-fig2-data3-v1.xlsx

Download elife-109876-fig2-data3-v1.xlsx

Next, we compared the proximal protein profiles of Drosophila executioner caspases, based on our previous observation that Dcp-1 and Drice are associated with distinct sets of proximal proteins, potentially reflecting differences in their activation mechanisms (Shinoda et al., 2019). Streptavidin staining, used to visualize proteins biotinylated via TurboID-mediated proximity labeling, revealed distinct patterns of Dcp-1 and Drice expression across tissues (Figure 2A, Figure 2—figure supplement 1A, B). To directly compare their proximal landscapes in the same cellular context, we focused on the imaginal discs. NeutrAvidin-based purification of biotinylated proteins, followed by MS, identified 818 proteins (Figure 2B, C). As expected, the proximal proteomes of Dcp-1 and Drice were distinct. BubR1 and Fasciclin 3 isoform G, previously identified as Drice-proximal in adult brain and ovary tissues (Muramoto et al., 2025; Shinoda et al., 2023), were enriched in the Drice-associated proteome (Figure 2C, Supplementary file 1), suggesting that Drice maintains its proximity to specific partners across tissues. Unlike overexpressed Dcp-1, endogenous Dcp-1 expressed in wing imaginal discs exists in its full-length form (Figure 2—figure supplement 2A), suggesting that the identified Dcp-1 proximal proteins are associated with the full-length pro-form. Consistently, DIAP-1, which is shown to interact with Dcp-1 and Drice only after exposure of the IAP-binding motif (IBM) at the neo N-terminus of the large executioner caspase subunit following cleavage (Tenev et al., 2005), is not identified in the Dcp-1 proximal protein list (Figure 2C, Supplementary file 1). To validate the TurboID-MS result, we compared our Dcp-1 proximal protein list with the previously reported 24 Dcp-1 interactors identified in Drosophila l(2)mbn cultured cells by immunoaffinity purification followed by MS (Choutka et al., 2017). Although the cell types and methods are different, 15 proteins are identified in both analyses (Figure 2—figure supplement 2B, Supplementary file 2). Among these, eight proteins were identified as highly enriched Dcp-1 proximal proteins (Dcp-1/Control fold change ≧2; Figure 2—figure supplement 2B, Supplementary file 2). Importantly, SesB, one of the best-characterized Dcp-1 interactors located in mitochondria (DeVorkin et al., 2014), was also identified in our TurboID-MS dataset (Figure 2—figure supplement 2B, Supplementary file 2), indicating that our TurboID-MS analysis successfully identified Dcp-1 proximal proteins. Among the top Dcp-1-enriched proximal proteins was the chaperonin containing TCP1 subunit 8 (CCT8). To validate these findings, we generated an antibody against CCT8 (Figure 2D). Western blotting confirmed the presence of CCT8 in Dcp-1::V5::TurboID samples but not in Drice::V5::TurboID samples (Figure 2E). Taken together, these results show that, even when expressed at comparable levels in the same tissue, Dcp-1 and Drice are associated with distinct sets of proximal proteins, which is consistent with their differential regulation and functional specificity.

To identify potential regulators of Dcp-1 activation, we performed an RNAi screen targeting eight genes that were selectively proximal to Dcp-1 relative to Drice and wild-type controls (Dcp-1/Drice fold change >1.5, Dcp-1/Control fold change >10 and PSM ≧2; Supplementary file 1), using two independent RNAi lines per gene (Figure 2F). Under normal conditions, RNAi-mediated knockdown of the chaperonin components CCT2 and CCT8 resulted in a wingless phenotype, whereas knockdown of other genes had no effect (Figure 2—figure supplement 3A). Under Dcp-1 overexpression conditions, knockdown of Fkbp59 or Sirt1 partially suppressed Dcp-1-induced cell death (Figure 2G, Figure 2—figure supplement 3B). The requirement for Sirt1 was further confirmed using a trans-heteroallelic combination of two previously described Sirt1 deletion alleles, Sirt1^2A-7-11 (Furuyama et al., 2004; Xie and Golic, 2004) and Sirt1^4.5 (Newman et al., 2002), which significantly rescued the wingless phenotype (Figure 2H and I). These results indicate that the Dcp-1-proximal protein Sirt1 is required for Dcp-1 activation upon overexpression.

Sirt1, also known as dSir2, is a lysine deacetylase that promotes the expression of autophagy-related genes via Atg8a deacetylation under starvation (Jacomin et al., 2020). Multiple studies have linked Dcp-1 to the regulation of autophagy. Dcp-1, together with the Bcl-2 family proteins Debcl and Buffy, but not Drice, has been shown to be required for starvation-induced autophagy in Drosophila l(2)mbn cells (Hou et al., 2008), and Dcp-1 is also required for autophagy in the ovary (Choutka et al., 2017; DeVorkin et al., 2014; Hou et al., 2008). More recently, the BH3-only protein Sayonara (Synr) was reported to induce Dcp-1-dependent cell death, which is regulated by Debcl, Buffy, and autophagy (Ikegawa et al., 2023). In contrast, a gain-of-function modifier screen revealed that autophagy antagonizes Dcp-1-induced cell death (Kim et al., 2010). It is also reported that Dcp-1 suppresses autophagy through cleavage-mediated inactivation of Acinus (Haberman et al., 2010; Nandi et al., 2014).

Based on these reports, we examined whether Synr, Debcl, Buffy, and core autophagy components regulate Dcp-1 activation. Knockdown of Synr did not suppress Dcp-1-induced cell death, whereas knockdown of Debcl or Buffy significantly rescued this wingless phenotype (Figure 3A, B). Similarly, knockdown of the autophagy-related genes Atg2 and Atg8a suppressed Dcp-1-induced cell death (Figure 3A, B), indicating that Debcl, Buffy, and autophagy are required for Dcp-1 activation. Of note, knockdown of any of the genes tested did not affect wing morphology in the absence of Dcp-1 overexpression (Figure 3—figure supplement 1A), suggesting the specific requirement for Dcp-1 activation. We further performed knockdown of autophagy-related genes including FIP200/Atg17, which is required for the initiation of autophagosome formation; Atg9, which is required for autophagosomal membrane nucleation; Atg5, which is required for autophagosomal membrane expansion through the Atg12-Atg5-Atg16 ubiquitin-like conjugation system; and Stx17, which is required for autophagosome-lysosome fusion (Umargamwala et al., 2024). While knockdown of these genes alone did not affect wing morphology in the absence of Dcp-1 overexpression (Figure 3—figure supplement 1B), knockdown of each of these genes suppressed the Dcp-1-induced wing ablation phenotype (Figure 3—figure supplement 1C), supporting a requirement for canonical autophagy components across multiple stages of autophagosome biogenesis in Dcp-1 activation. Consistent with the requirement of autophagy for Dcp-1 activation, induction of autophagy, as judged by the accumulation of the formation of autophagosomes marked by mCherry-tagged Atg8a puncta, was observed upon Dcp-1 overexpression (Figure 3C). These results further support the idea that autophagy promotes Dcp-1 activation.

Figure 3 with 1 supplement see all

Download asset Open asset

Figure 3—source data 1

Data used for graphs presented in Figure 3B.

https://cdn.elifesciences.org/articles/109876/elife-109876-fig3-data1-v1.xlsx

Download elife-109876-fig3-data1-v1.xlsx

To assess the specificity of these regulators, we analyzed the eye ablation phenotype caused by overexpression of Hid, which triggers apoptosis via canonical apoptotic signaling. As expected, this phenotype was rescued by p35 expression or knockdown of RHG genes and partially suppressed by inhibition of the Dark-, Dronc-, and Drice-dependent apoptosome pathways (Figure 3D), confirming its dependence on canonical apoptosis (Figure 1E). In contrast, Hid-induced apoptosis was not rescued by RNAi against Dcp-1 (Figure 3D), nor by knockdown of Sirt1 or Fkbp59 (Figure 3E). Furthermore, knockdown of Debcl, Buffy, and autophagy genes failed to rescue Hid-induced apoptosis (Figure 3F). Taken together, these results indicate that Sirt1, Debcl, Buffy, and autophagy components are not general regulators of apoptosis but function as specific activators of Dcp-1.

Next, we examined the molecular link between autophagy and Dcp-1 activation. Caspase activation is negatively regulated by IAPs. In Drosophila, DIAP-1 suppresses caspase activity through ubiquitination (Ditzel et al., 2008; Figure 1E). In addition to DIAP-1, Drosophila has another giant IAP, Bruce, which is degraded via autophagy during oogenesis (Nezis et al., 2010). Bruce inhibits autophagy under nutrient-rich conditions (Hou et al., 2008), suggesting mutual negative regulation of Bruce and autophagy. Bruce selectively inhibits reaper-induced cell death, but not hid-induced apoptosis, suggesting a degree of pathway specificity (Domingues and Ryoo, 2012). These observations led us to focus on Bruce as a potential selective regulator of Dcp-1 activation.

The structure of human Bruce/BIRC6 has recently been resolved using cryo-electron microscopy (Dietz et al., 2023; Ehrmann et al., 2023; Hunkeler et al., 2023; Liu et al., 2024), revealing several conserved domains: linker domain (LD), baculoviral IAP repeat (BIR), WD40, β-X1, β-X2, carbohydrate-binding module (CBM), ubiquitin-like (UBL), and ubiquitin-conjugating (UBC) (Figure 4A). Human Bruce/BIRC6 ubiquitinates the activated executioner caspases caspase-3 and caspase-7, but not their non-processed proforms in vitro (Dietz et al., 2023). The BIR domain mediates interactions with ubiquitination substrates by recognizing N-terminal IBMs (Berthelet and Dubrez, 2013). The consensus IBM sequence is defined as A(K/T/V/I)(P/A/E)(F/E/I/S/Y) (Vaux and Silke, 2005). We identified IBM-like sequences at the N-termini of the P20 domains of both Dcp-1 and Drice (Figure 4B, C). To assess the potential interaction between Bruce and these caspases, we used AlphaFold3 to predict the structure of the complexes between Bruce and the IBM-like motifs of Dcp-1 and Drice (Abramson et al., 2024). Consistent with the known architecture of BIRC6, AlphaFold3 predicted that the BIR domain of Bruce interacts with LD, WD40, β-X1, and β-X2 domains (Figure 4—figure supplement 1A, A’, B, B’). For structural prediction, we used a Bruce fragment spanning residues 1–1960, including the BIR domain and its associated Zn²⁺-binding site (Figure 4A), and 15-residue peptides from the Dcp-1 and Drice cleavage sites encompassing their IBM-like motifs (Figure 4B, C). The predicted structures revealed a high-confidence interaction between Dcp-1 and the Bruce BIR domain. All five AlphaFold3 models showed confident pLDDT scores (70–90) for the N-terminal residues of Dcp-1, corresponding to the IBM-like motif (Figure 4D, E). The predicted binding mode resembled the canonical BIR-IBM interaction observed in XIAP BIR3-Smac (Wu et al., 2000; Figure 4—figure supplement 1C, C’). In contrast, although Drice appeared to bind to Bruce in the predicted models, the pLDDT scores were consistently low (<70) across the interaction interface, including the four N-terminal residues, indicating low confidence in the predicted interaction (Figure 4F, G).

Figure 4 with 2 supplements see all

Download asset Open asset

Figure 4—source data 1

Uncropped raw blot images of Figure 4H, J.

https://cdn.elifesciences.org/articles/109876/elife-109876-fig4-data1-v1.zip

Download elife-109876-fig4-data1-v1.zip

Figure 4—source data 2

Uncropped annotated blot images of Figure 4H, J.

https://cdn.elifesciences.org/articles/109876/elife-109876-fig4-data2-v1.zip

Download elife-109876-fig4-data2-v1.zip

To validate these predictions, we performed co-immunoprecipitation assays on Drosophila S2 cells. We found that only the cleaved form of Dcp-1, which forms a head-to-tail dimer composed of the P20 and P10 domains (Julien and Wells, 2017), and not the full-length protein, interacted with the N-terminal BIR domain of Bruce following apoptosis induction (Figure 4H), consistent with the observation that human Bruce/BIRC6 only targets cleaved executioner caspases for ubiquitination in vitro (Dietz et al., 2023). Conversely, neither the cleaved nor full-length forms of Drice showed detectable interactions with the Bruce BIR domain (Figure 4H), corroborating the AlphaFold3 predictions.

To further examine whether endogenous Bruce interacts with activated Dcp-1, we generated CRISPR/Cas9-mediated knock-in flies in which an mStayGold::C4 linker::V5-tag (Ando et al., 2024) was inserted into the N-terminus of Bruce (Figure 4—figure supplement 2A). Bruce has been reported to function in oogenesis and spermatogenesis in the ovaries and testes, respectively (Arama et al., 2007; Arama et al., 2003; Hou et al., 2008; Kaplan et al., 2010; Nezis et al., 2010). However, its expression in other tissues has not been thoroughly characterized. Using the knocked-in line, we confirmed that Bruce was expressed in the ovaries (Figure 4—figure supplement 2B) and testes (Figure 4—figure supplement 2C), consistent with previous reports (Arama et al., 2007; Arama et al., 2003; Hou et al., 2008; Kaplan et al., 2010; Nezis et al., 2010). Western blot analysis revealed Bruce expression in both adult males and females (Figure 4—figure supplement 2D). We also detected Bruce expression in wing imaginal discs (Figure 4—figure supplement 2E). This signal was abolished by the RNAi-mediated knockdown of Bruce, confirming the specificity of the signal (Figure 4—figure supplement 2F). These findings indicate that Bruce not only functions in the gonads but can also exert Dcp-1 regulatory functions in wing imaginal discs. We overexpressed Dcp-1::VENUS using WP-Gal4 in the background of the established mStayGold::C4 linker::V5-tag knock-in-tagged Bruce and conducted co-immunoprecipitation analysis (Figure 4I). Following immunoprecipitation with an anti-V5 tag antibody, we found that cleaved Dcp-1 signal was enriched by co-immunoprecipitation (Figure 4J), suggesting that full-length Bruce interacts with activated Dcp-1. Consistently, Bruce was not identified as a proximal protein of full-length Dcp-1 in our TurboID-MS analysis (Figure 2C, Supplementary file 1). Taken together, these findings suggest that Bruce specifically suppresses activated Dcp-1 but not Drice and serves as a branching point for a Dcp-1-specific, non-canonical regulatory pathway.

Next, we investigated whether Bruce regulated Dcp-1 activity. Upon overexpression of Bruce (Figure 5A), we observed marked suppression of Dcp-1 overexpression-induced cell death (Figure 5B, C). Although DIAP-1 overexpression also inhibited Dcp-1 activation, this suppressive effect was stronger following Bruce overexpression (Figure 5B, C). We examined the specificity of Bruce in the regulation of other forms of cell death. As mentioned earlier, the BH3-only protein Synr induces Debcl/Buffy, autophagy, and Dcp-1-dependent cell death (Ikegawa et al., 2023). To investigate whether Bruce was involved in Debcl/Buffy-autophagy-Dcp-1-mediated cell death, we examined the role of Bruce in Synr-induced cell death. Overexpression of Bruce, but not DIAP-1, specifically suppressed the wrinkled wing phenotype caused by Synr-induced cell death (Figure 5D). In contrast, Bruce overexpression failed to rescue the eye ablation phenotype independent of autophagy caused by Hid overexpression, whereas DIAP-1 fully rescued this phenotype (Figure 5E). Taken together, these results suggest that Bruce and DIAP-1 regulate distinct cell death pathways, with Bruce acting as a selective inhibitor of autophagy-regulated Dcp-1- and Synr-dependent cell death.

Figure 5 with 1 supplement see all

Download asset Open asset

Figure 5—source data 1

Data used for graphs presented in Figure 5C, G, I, K.

https://cdn.elifesciences.org/articles/109876/elife-109876-fig5-data1-v1.xlsx

Download elife-109876-fig5-data1-v1.xlsx

We have previously reported that Dcp-1 activity can promote imaginal tissue growth independent of the core apoptosis signaling pathway, including Dronc (Shinoda et al., 2019). To investigate the role of Bruce in imaginal tissue growth, we analyzed its effect on wing imaginal discs using WP-Gal4, which drives expression throughout the wing pouch. Overexpression of p35 suppressed wing tissue growth (Figure 5F, G), whereas inhibition of canonical apoptotic signaling by Dronc^DN and DIAP-1 had no effect (Figure 5F, G). In contrast, Bruce overexpression suppressed wing growth to an extent similar to p35 overexpression, whereas Bruce knockdown tended to increase wing size, although the difference was not statistically significant (Figure 5F, G). We confirmed these results using Engrailed-Gal4, which is expressed in the posterior compartment of the wing imaginal disc. p35 overexpression suppressed wing tissue growth (Figure 5H, I). In this context, the inhibition of canonical apoptosis signaling by Dronc^DN and DIAP-1 had a minor but negative effect on wing size (Figure 5H, I). Bruce overexpression suppressed wing growth, whereas Bruce knockdown increased wing size (Figure 5H, I), indicating that Bruce negatively regulates wing tissue growth. We also found that knockdown of Debcl and Buffy resulted in reduced wing size (Figure 5—figure supplement 1A), suggesting that Debcl and Buffy also participate in this process. Finally, to enhance the sensitivity of the assay, we overexpressed the dominant-negative form of the insulin receptor (InR^DN) using WP-Gal4, which significantly reduced wing size (Figure 5J, K). Under this sensitized condition, simultaneous Bruce knockdown significantly enhanced wing tissue growth (Figure 5J, K). Overall, these results indicate that Bruce suppresses wing tissue growth.

In this study, we identified a distinct regulatory mechanism governing the activation of the Drosophila executioner caspase, Dcp-1, in sharp contrast to its paralog Drice. Although both caspases share structural similarities and are expressed in overlapping tissues, Dcp-1, but not Drice, is activated upon overexpression in an apoptosome-independent, autophagy-regulated manner. Previous studies have implicated Dcp-1 in the induction of autophagy (Choutka et al., 2017; DeVorkin et al., 2014; Hou et al., 2008). Our findings extend this view by showing that autophagy is required for Dcp-1 activation, suggesting a positive feedback loop between Dcp-1 and autophagy during apoptosis. Although the precise downstream targets of the Bcl-2 family proteins, Debcl and Buffy, remain unknown, their roles in autophagy regulation (Hou et al., 2008) suggest that they influence Dcp-1 activity through autophagic pathways. Additionally, we identified Bruce, a giant inhibitor of apoptotic proteins, as a Dcp-1-specific suppressor. Functionally, Bruce limits wing tissue growth, suggesting that Dcp-1 activation contributes to growth regulation in a non-lethal manner (Shinoda et al., 2019). Taken together, our results demonstrate that the regulatory mechanisms of Dcp-1 and Drice diverge at the level of IAPs and that Dcp-1 activation is regulated by autophagy, enabling specific non-lethal functions that promote tissue growth independent of canonical apoptotic signaling (Figure 5L). It is important to note that Dcp-1 can also suppress autophagy through cleavage of Acinus (Haberman et al., 2010; Nandi et al., 2014). Although this may seem contradictory, the ability of caspases to function in a non-lethal manner requires strict prevention of excessive activation. Such suppression may represent a key mechanism that enables non-lethal caspase activity.

We have previously shown that another executioner caspase, Decay, also participates in promoting imaginal tissue growth (Shinoda et al., 2019). However, the upstream regulatory mechanisms of Decay remain unknown. Importantly, Decay has been shown to mediate Hid-induced cell death in the DIAP1- and apoptosome-independent manner in differentiating photoreceptors and accessory cells of the eye (Leulier et al., 2006). In addition, although not required for cell death, Decay accounts for most of the caspase activity during metamorphic midgut programmed cell death, which is executed by autophagy (Denton et al., 2009). Thus, similar to Dcp-1, Decay might be an executioner caspase that can be regulated independently of the canonical apoptosome-mediated pathway, potentially involving autophagy–Bruce axis, and thereby contributing to the regulation of tissue growth.

It is also important to understand how cells determine the usage of canonical versus alternative caspase activation pathways depending on physiological or environmental conditions. In mammals, two major apoptotic pathways are activated by distinct initiator caspases, caspase-8 and caspase-9, in response to different upstream signals (Newton et al., 2024). Similarly, in Drosophila, while most developmental cell death is driven by canonical apoptotic signaling involving Dronc (Xu et al., 2005), alternative caspase activation pathways may be preferentially engaged in response to environmental stimuli, such as starvation or heat shock.

The mechanism by which Dcp-1 is activated by itself remains an open and important question. In contrast to Drice, Dcp-1 is activated upon overexpression in an apoptosome-independent but autophagy-dependent manner. Unlike initiator caspases, executioner caspases such as Dcp-1 generally lack long N-terminal prodomains and require proteolytic cleavage to become catalytically active. Proximity-labeling analysis revealed that Dcp-1 was predominantly associated with the chaperonin subunit CCT8 (Figure 2E). Given that the overexpression of Dcp-1 increases its local concentration for activation, we propose that its accumulation in the chaperonin complex may serve as a local scaffold for caspase activation. This model implies a spatially restricted mode of regulation in which caspase activity is compartmentalized within the cell. In this model, we speculated that Dcp-1 activation occurs sporadically in subcellular regions enriched in activators or scaffolding molecules such as CCT8. Under these conditions, Dcp-1 activation may need to be tightly regulated to prevent its unintended proteolytic activity. Supporting this notion, we proposed that Bruce functions as a basal suppressor of Dcp-1 in non-lethal contexts, restraining caspase activity when full apoptotic signaling is not engaged.

Caspases are essential cysteine proteases that execute apoptotic cell death and are conserved from C. elegans to mammals (Shinoda and Miura, 2024). Multiple executioner caspases exist, including caspase-3 and caspase-7 in mammals and Dcp-1 and Drice in Drosophila. Although these enzymes are often functionally redundant in apoptosis, growing evidence indicates that each executioner caspase harbors distinct substrate specificities and fulfills non-overlapping roles (Boucher et al., 2012; Desroches and Denault, 2019; Shinoda et al., 2023; Shinoda et al., 2019; Walsh et al., 2008). In mammals, caspase-3 and caspase-7 exhibit high sequence similarity but differ in their substrate preferences. Caspase-7, for instance, contains polybasic residues in its exosite that enhance its ability to cleave RNA-binding proteins, a function not observed in caspase-3 (Boucher et al., 2012). Mechanistically, this exosite enables caspase-7 to interact directly with RNA, thereby promoting selective proteolysis of RNA-associated substrates (Desroches and Denault, 2019). Furthermore, we have recently demonstrated that caspase-7 is essential for early-peaking caspase activity at the plasma membrane during apoptosis, a process mediated by its N-terminal intrinsically disordered region, which facilitates rapid efferocytosis of apoptotic cells (Taira et al., 2025). In this study, we demonstrated that Drosophila executioner caspases Dcp-1 and Drice are also subject to distinct activation mechanisms for the specific non-lethal function of tissue growth. These findings underscore that the evolutionary expansion of executioner caspases does not merely provide functional redundancy for apoptotic execution but rather confers unique biochemical properties and regulatory roles under distinct cellular contexts. This diversification allows executioner caspases to participate in both lethal and non-lethal cellular processes, highlighting their broader functional repertoire beyond canonical apoptosis.

We raised the flies in standard Drosophila medium at 25°C. For in vivo biotin labeling experiments, flies were raised in standard Drosophila medium supplemented with 100 μM (+)-biotin (#029-08713, Wako). The fly strains used in this study are listed in Supplementary file 3. Genotypes corresponding to each figure are shown in Supplementary file 4.

Amino acid sequences were obtained from the FlyBase database (https://flybase.org/). Disordered regions were predicted using the PSIPRED protein sequence analysis workbench in DISOPRED3 (http://bioinf.cs.ucl.ac.uk/psipred/).

Adult female flies were anesthetized using CO₂. The wing and eye images were acquired using a Leica MZ16F fluorescent stereomicroscope (Leica Microsystems) equipped with a Leica DFC450 C CCD camera. Each wing was manually classified into one of four categories: Intact (no defect), Mild (≧50% of wing area intact), Severe (<50% but ≧5% of wing area intact), and Wingless (<5% of wing area intact).

For pAc5-Caspases-myc-mNeonGreen, the coding sequences of Dcp-1 and Drice were amplified using PCR from the corresponding pUASz-Caspase::V5::TurboID plasmids (Muramoto et al., 2025). The coding sequence of myc-tagged mNeonGreen was PCR-amplified from pUASz-Drice::myc::mNeonGreen plasmid (Muramoto et al., 2025). The two fragments were ligated into EcoRI/XhoI-digested pAc5_STABLE2_neo (González et al., 2011) using In-Fusion (#Z9648N, TaKaRa).

For pCold I-EcoRI-CCT8-EcoRI, the coding sequence of CCT8 was PCR-amplified from Drosophila cDNA, and the fragment was ligated into the EcoRI-digested pCold I DNA vector (#3361, TaKaRa) using In-Fusion.

For pAc5-Caspases^CG-myc, the coding sequences of Dcp-1 and Drice were PCR amplified from the corresponding pAc5-Caspases-myc-mNeonGreen plasmids. The fragments were ligated into EcoRI/BamHI-digested pAc5_STABLE2_neo (González et al., 2011) using In-Fusion. For pAc5-Bruce^1-900aa-3xFLAG, the first 900 aa of the coding sequences of Bruce were PCR-amplified from Drosophila cDNA, and the fragment was ligated into the EcoRI-digested pAc5-Fasciclin 3 isoform A-3xFLAG plasmids (Muramoto et al., 2025) using In-Fusion. For pAc5-3xFLAG-mRuby3, Drosophila codon-optimized mRuby3 was synthesized (Integrated DNA Technologies, Coralville, IA, USA) and ligated into the BamHI-digested pAc5-3xFLAG-Bub3 plasmids (Shinoda et al., 2023) using In-Fusion.

For mStayGold::C4 linker::V5-tag knock-in for Bruce, the guide RNA was identified using the CRISPR Optimal Target Finder tool available on flyCRISPR (http://targetfinder.flycrispr.neuro.brown.edu/) (Gratz et al., 2014). The DNA fragments for the guide RNA were subcloned into BbsI-digested pDCC6 vectors (#59985, Addgene) (Gokcezade et al., 2014) to establish the pDCC6-Bruce-sgRNA plasmid. The following primers were annealed to generate the DNA fragments for guide RNAs: 5′-CTTCGCTTGCTTGAGGACGCGCTGC-3′ and 5′-AAACGCAGCGCGTCCTCAAGCAAGC-3′. For the generation of the homology-directed repair template (pBac-mStayGold-c4-V5-Bruce-donor-3xP3-dsRed-polyA), DNA fragments of the 5′ homology arm were assembled into the BsaI site, and the DNA fragment of the 3′ homology arm with mStayGold::C4 linker::V5-tag was assembled into the SapI site of pBac[3xP3-DsRed_polyA_Scarless_TK] (Shinoda et al., 2019) using In-Fusion. The PAM sequence (TGG) of the gRNA-binding sites in the donor template was mutated (TGA) to prevent Cas9-directed cleavage following homology-directed repair. The mStayGold::C4 linker::V5-tag fragment was amplified from pRSETB/mStayGold(c4) (#212018; Addgene) (Ando et al., 2024) with V5-tag harboring primers. DNA fragments were amplified using PCR with the following primers: Bruce 5′ homology arm (forward primer, 5′-TGAAGGTCTCCTTAAAAGCCCGATTGGCGGCGCAGGAAGT-3′, and reverse primer, 5′-TCTTTCTAGGGTTAATCGAGTGCCCAGGGGGCGCGCGAGT-3′), Bruce 3′ homology arm-1 (forward primer, 5′-TCTTTCTAGGGTTAATTTTTGGCATTCACAGGCGATTGCT-3′, and reverse primer, 5′-TCCGGAGGTGCGCCGCGGCAATTTG-3′), mStayGold-C4 linker-V5-tag (forward primer, 5′-CGGCGCACCTCCGGAATGGTGTCTACAGGCGAGGAGCTGT-3′, and reverse primer, 5′-GGTGCTGTCCAGGCCCAGCAGGGGGTTGGGGATGGGCTTGCCCACGGCAGAAGCAGAAGGCTCGTGC-3′), 3′ homology arm-2 with mutated PAM (forward primer, 5′-GGCCTGGACAGCACCGCCACGGAGCAGCATCATCAGCAGCGCGTCCTCAA-3′, and reverse primer, 5′-GACGGCTCTTCATTAACTCTTAAATCAAATTTAAATTGAA-3′). The established plasmids were injected into yw; nos-Cas9(II-attP40) for CRISPR/Cas9-mediated transgenesis (BestGene Inc). Each DsRed-positive transformant was isogenized and confirmed using genomic PCR and sequencing.

All established plasmids were sequenced by Eurofins, Inc The detailed plasmid DNA sequences are available upon request.

Drosophila S2 cells (#RCB1153, RIKEN BRC) were grown at 25°C in Schneider’s Drosophila medium (#21720001, Gibco) supplemented with 10% (vol/vol) heat-inactivated fetal bovine serum, 100 U/ml penicillin, and 100 μg/ml streptomycin (#168-23191, Wako). Mycoplasma contamination was assessed using the MycoStrip (#rep-mys-10, InvivoGen) and found to be negative.

Drosophila S2 cells (2.5 × 10⁵) seeded in 24-well plates were transfected with 200 ng of the desired plasmid using Effectene Transfection Reagent (#301427, QIAGEN) following the manufacturer’s protocol. Sixteen hours after transfection, cells were transferred to four-compartment glass-bottomed dishes (#627870, Greiner). After another 24 hr, images were captured using a Leica TCS SP8 confocal microscope (Leica Microsystems).

The mNeonGreen intensity was quantified using Fiji (ImageJ) software (NIH Image). Images were analyzed automatically using the ImageJ macro, as previously described (Fujisawa et al., 2020) with modifications. Briefly, the mNeonGreen channel was filtered by the ‘Gaussian Blur (sigma = 1)’, processed with ‘Auto Threshold (“Mean” method)’, processed with ‘Watershed’, and processed with ‘Analyze particles (size = 40–200 µm², circularity = 0.2–1.0)’ to define the region of interest. For a given region of interest, the average mNeonGreen intensity was measured.

Whole larvae without digestive tract and fat bodies, larval salivary glands, fat bodies, brains, wing discs, and whole adults, adult thoraxes, ovaries, and testes of desired genotype were dissected and lysed with RIPAi buffer (50 mM Tris-HCl, pH 8.0, 150 mM sodium chloride, 0.5% wt/vol sodium deoxycholate, 0.1% wt/vol sodium dodecyl sulfate, and 1.0% wt/vol NP-40) supplemented with cOmplete ULTRA EDTA-free protease inhibitor cocktail (#05892953001, Roche). Samples were homogenized and centrifuged at 20,000 × g, 4°C for 10 min. The supernatants were collected and snap-frozen in liquid nitrogen. Protein concentrations were determined using the BCA assay (#297-73101, Wako) following the manufacturer’s protocol. The samples were mixed with Laemmli buffer, boiled at 95°C for 5 min, and then subjected to western blotting analysis.

Each sample was separated using SDS–PAGE. The proteins were then transferred onto Immobilon-P polyvinylidene fluoride membranes (#IPVH00010; Millipore) for immunoblotting. Membranes were blocked with 4% skim milk diluted in 1× TBST. Immunoblotting was performed using the antibodies mentioned below, which were diluted with 4% skim milk. Signals were visualized using Immobilon Western Chemiluminescent HRP Substrate (#WBKLS0500; Millipore) and FUSION SOLO. 7S. EDGE (Vilber-Lourmat). Contrast and brightness adjustments were applied equally using Fiji (ImageJ) software (NIH Image).

The primary antibodies used included rabbit anti-GFP polyclonal antibody (1:2000, #598, MBL), mouse anti-V5 tag monoclonal antibody (1:1000, #R960-25, Invitrogen), mouse anti-FLAG M2 monoclonal antibody (1:5000, #F1804, Sigma), rabbit anti-CCT8 polyclonal antibody (1:5000, this study), rabbit anti-Myc-Tag (71D10) monoclonal antibody (1:5000, #2278, CST), rabbit anti-3xDYKDDDDK Tag (3xFLAG Tag) (E7C5T) monoclonal antibody (1:5000, #87537, CST), mouse anti-alpha tubulin (DM1A) monoclonal antibody (1:10,000, #T9026, Sigma), mouse anti-Histone H3 (1B1B2) monoclonal antibody (1:10,000, #14269, CST), mouse anti-Actin (AC-40) monoclonal antibody (1:5000; #A4700, Sigma), rabbit anti-alpha tubulin polyclonal antibody (1:2000, #ab15246, Abcam), and rabbit anti-eEF2 polyclonal antibody (1:2000, #ab33523, Abcam). The secondary antibodies used were goat anti-rabbit IgG horseradish peroxidase (HRP)-conjugated antibody (1:5000, #7074S, CST) and goat/rabbit/donkey anti-mouse IgG HRP Conjugate (1:5000, #W402B, Promega). The membranes for streptavidin blotting were blocked with 3% bovine serum albumin. Streptavidin blotting was performed using streptavidin–HRP (1:10,000, #SA10001; Invitrogen) diluted in 3% bovine serum albumin.

Two biological replicates were analyzed for each experimental condition to determine the abundance of Drice/Dcp-1::TurboID relative to the control.

Samples were prepared as previously described (Shinoda et al., 2023). Briefly, 100 µg of biotinylated protein-containing lysate was subjected to FG-NeutrAvidin beads (#TAS8848 N1171, Tamagawa Seiki) purification. FG-NeutrAvidin beads (25 µl, approximately 500 µg) were washed three times with RIPA buffer. Benzonase-treated biotinylated protein samples suspended in 1 ml of RIPA buffer were incubated overnight at 4°C. Conjugated beads were magnetically isolated and washed with 500 µl of an ice-cold RIPA buffer solution, 1 M KCl solution, 0.1 M Na₂CO₃ solution, and 4 M urea solution. For western blot analyses, the purified samples were mixed with Laemmli buffer, boiled at 95°C for 5 min, and then subjected to western blot analysis. For LC–MS/MS analysis, the purified samples were washed with 500 µl ultrapure water (#214-01301, Wako) and 500 µl of 50 mM ammonium bicarbonate. The samples were then mixed with 50 µl of 0.1% Rapigest diluted in 50 mM ammonium bicarbonate, and 5 µl of 50 mM TCEP was subsequently added. Samples were incubated at 60°C for 5 min, and then 2.5 µl of 200 mM methyl methanethiosulfonate was added. One microgram sequence-grade modified trypsin (#V5111, Promega) was then added for on-bead trypsin digestion at 37°C, leaving the treatment for 16 hr. The beads were then magnetically isolated, and 60 µl of the supernatant was collected. Then, 3 µl of 10% trifluoroacetic acid was added to the supernatants, incubating them at 37°C for 60 min with gentle agitation. The samples were then centrifuged at 20,000×g, 4°C for 10 min. The peptides were then desalinated and purified using a GL-tip SDB (#7820-11200, GL Sciences) following the manufacturer’s instructions. The samples were speed-backed at 45°C for 30 min and dissolved in 25 µl of 0.1% formic acid. The samples were then centrifuged at 20,000 × g, 4°C for 10 min, and the supernatants were collected. The peptide concentrations were determined using the BCA assay. Finally, 500 ng of the purified protein was subjected to LC–MS/MS analysis.

Samples were loaded onto an Acclaim PepMap 100 C18 column (75 µm × 2 cm, 3 µm particle size and 100 Å pore size; #164946, Thermo Fisher Scientific) and separated on a nano-capillary C18 column (75 µm × 12.5 cm, 3 µm particle size; #NTCC-360/75-3-125, Nikkyo Technos) using the EASY-nLC 1200 system (Thermo Fisher Scientific). Elution conditions are listed in Supplementary file 5. The separated peptides were analyzed using QExactive (Thermo Fisher Scientific) in data-dependent MS/MS mode. The parameters for the MS/MS analysis are listed in Supplementary file 6. The collected data were analyzed using Proteome Discoverer (PD) 2.2 software with the Sequest HT search engine. The parameters for the PD 2.2 analysis are described in Supplementary file 7. Peptides were filtered at a false discovery rate of 0.01 using the Percolator node. Label-free quantification was performed based on the intensities of the precursor ions using a precursor-ion quantifier node. Normalization was performed using the total amount of peptides in all average scaling modes. MS proteomic data were deposited in the ProteomeXchange Consortium via the jPOST partner repository with the dataset identifier PXD067425 (Okuda et al., 2017).

Twenty-four Dcp-1 interactors identified by immune-affinity purification MS (Choutka et al., 2017) were individually searched among all proteins identified in our TurboID-mediated proteomics analysis using their corresponding UniProt accessions. Fifteen proteins were identified in both datasets, and their abundance ratios and #PSMs are listed in Supplementary file 2.

CCT8 was expressed in Escherichia coli BL21 (#9126, TaKaRa) transformed with pCold I-EcoRI-CCT8-EcoRI. Cells were grown at 37°C overnight, diluted to OD₆₀₀ 0.2, and cultured until OD₆₀₀ 0.5 at 37°C. Protein expression was induced by the addition of 1 mM IPTG at 15°C for 24 hr. Cells were centrifuged at 4000 × g, 4°C for 20 min and resuspended in phosphate-buffered saline (PBS). Cells were centrifuged at 4000 × g, 4°C for 10 min. Purification of 6xHis-tagged CCT8 was performed using the QIAexpress Ni-NTA Fast Start Kit (#30600, QIAGEN) according to the manufacturer’s protocol: purification of 6xHis-tagged proteins under denaturing conditions. Briefly, the cell pellets were lysed in Lysis Buffer (Denaturing Buffer adjusted to pH 8.0) with gentle agitation for 60 min. The sample was centrifuged at 9200 × g, 20°C for 30  min. The supernatant was applied to Fast Start column and washed twice with Wash Buffer (Denaturing Buffer adjusted to pH 6.3). Proteins were eluted using Elution Buffer (Denaturing Buffer adjusted to pH 4.5). Polyclonal antibodies were raised in rabbits using purified 6xHis-tagged CCT8 as an immunogen (Eurofins Inc).

Amino acid sequences were obtained from the FlyBase database (https://flybase.org/). The protein complex structures were predicted using AlphaFold3 (Abramson et al., 2024) on the AlphaFold server (https://alphafoldserver.com/). Molecular structures and figures were visualized and prepared using PyMOL (Schrödinger, LLC).

Drosophila S2 cells (2.0 × 10⁶) seeded in 6-well plates were transfected with the desired plasmid using Effectene Transfection Reagent. For live cell lysate, Drosophila S2 cells were lysed on ice using 550 µl of IP lysis buffer (50 mM Tris-HCl (pH 7.5), 150 mM NaCl, 0.5% NP-40) supplemented with cOmplete ULTRA EDTA-free protease inhibitor cocktail. For apoptotic cell lysate, the cells were treated with 10 μg/ml cycloheximide (#C7698, Sigma) for apoptosis induction. After 6 hr of incubation, cells were harvested and centrifuged at 800 × g, 4°C for 5  min and washed twice. Then, cells were lysed by using 550 µl of IP lysis buffer (50 mM Tris-HCl (pH 7.5), 150 mM NaCl, and 0.5% NP-40) supplemented with cOmplete ULTRA EDTA-free protease inhibitor cocktail. The cell lysates were centrifuged at 20,000 × g, 4°C for 5  min. Supernatants (50 μl) were collected as inputs. The remainder of the supernatant was incubated for 2 hr at 4°C with 20 μl of anti-DDDDK-tag mAb-Magnetic Agarose (#M185-10R, MBL) equilibrated with the IP lysis buffer. The beads were washed thrice in IP lysis buffer and boiled for 5 min at 95°C with 50 μl 1× Laemmli buffer. The samples were magnetically separated, and the supernatants were subjected to SDS–PAGE.

Fifty wing imaginal discs were dissected and lysed on ice using 550 µl of IP lysis buffer supplemented with cOmplete ULTRA EDTA-free protease inhibitor cocktail. Samples were homogenized and centrifuged at 20,000 × g, 4°C for 5 min. Supernatants (50 μl) were collected as inputs. The remainder of the supernatant was incubated for 2 hr at 4°C with 20 μl of anti-V5-tag mAb-Magnetic Agarose (#M167-10, MBL) equilibrated with the IP lysis buffer. The beads were washed thrice in IP lysis buffer and boiled for 5 min at 95°C with 50 μl 1× Laemmli buffer. The samples were magnetically separated, and the supernatants were subjected to SDS–PAGE.

Wing imaginal discs of the late third-instar larvae and testes and ovaries of the adults were fixed with 4% paraformaldehyde in PBS. The wing imaginal discs and testes samples were washed in PBS containing 0.1% Triton X-100 (PBST), and the ovary samples were washed in 1.0% PBST, blocked in PBST with 5% normal donkey serum (PBSTn) for 30 min, and incubated with primary antibodies in PBSTn overnight at 4°C. The samples were washed with PBST, incubated for 2 hr with secondary antibodies suspended in PBSTn, and washed again with PBST. The samples were mounted on glass slides with SlowFade Gold antifade reagent (#S36939, Invitrogen). Images were captured using a Leica TCS SP8 microscope (Leica Microsystems). Images were analyzed and edited using the Fiji (ImageJ) software.

The primary antibodies used were rabbit anti-HA (Y11) polyclonal antibody (1:100, #sc-805, Santa Cruz), mouse anti-HA (12CA5) monoclonal antibody (1:200, #ab1424, Abcam), rabbit anti-cleaved Drosophila Dcp-1 (Asp215) polyclonal antibody (1:200, #9578S, CST), rabbit anti-cleaved PARP (Asp214) (D64E10) monoclonal antibody (1:200, #5625S, CST), mouse anti-Myc-tag monoclonal antibody (1:500, #R950-25, Invitrogen), and mouse anti-V5 tag monoclonal antibody (1:100, #R960-25, Invitrogen). The secondary antibodies used were Cy3-conjugated donkey anti-rabbit IgG antibody (1:250, #711-165-152, Jackson Immuno Research), Alexa647-conjugated donkey anti-mouse IgG antibody (1:100, #A-31571, Thermo Fisher Scientific), Alexa488-conjugated donkey anti-rabbit IgG antibody (1:200, #A-21206, Thermo Fisher Scientific), Alexa647-conjugated donkey anti-rabbit IgG antibody (1:200, #A-31573, Thermo Fisher Scientific), and Alexa488-conjugated donkey anti-mouse IgG (1:500 (Myc) or 1:100 (V5), #A-21202, Thermo Fisher Scientific). Hoechst 33342 (8 µM; #H3570, Invitrogen) for nuclear staining and Phalloidin-iFluor633 (1:1,000, #ab176758, Abcam) for F-actin staining were used during the secondary antibody incubation.

TUNEL assay was performed using the ApopTag Red in situ apoptosis detection kit (#S7165, Millipore) according to the manufacturer’s instructions as previously described (Shinoda et al., 2019). Wing imaginal discs were dissected and fixed for 20 min at room temperature in 1% PFA in PBS. The samples were then washed with PBS three times, 5 min per wash. Next, the samples were re-fixed in ethanol:acetic acid, 2:1, for 5 min at −20°C; washed in PBS twice, 5 min per wash; incubated in Equilibration Buffer for 10 s; and incubated again in Working Strength TdT Enzyme at 37°C for 1 hr. The TdT reaction mix was replaced with Working Strength Stop/Wash Buffer and incubated for 10 min at room temperature. Then, the samples were washed three times with PBS, 1 min per wash, and incubated with Working Strength Rhodamine Antibody Solution with 8 µM Hoechst 33342 for 4 hr at room temperature. The samples were then washed four times in PBS, 2 min per wash. The samples were mounted on glass slides with SlowFade Gold antifade reagent. Images were captured using a Leica TCS SP8 microscope. Images were analyzed and edited using the Fiji (ImageJ) software.

Flies were raised, and their wings were mounted and digitized as previously described (Shinoda et al., 2019). Briefly, flies were placed on an acetic acid agar plate (2% agar, 1% sucrose, and 0.36% acetic acid) to lay eggs overnight. The plates were collected, and the eggs were incubated at 25°C for 24 hr. The 1st instar larvae were collected from the agar plate, and 30 larvae were placed in one vial to precisely control the larval density. These flies were raised at 25°C. After more than 2 days of hatching, the adult female flies were collected and submerged in 99.5% ethanol to dehydrate for 24 hr. Dehydrated flies were dissected to mount their wings using Euparal (#3C-239, Waldeck) with the dorsal side upward. Images were digitized using an upright microscope (Leica DM5000B; Leica Microsystems) equipped with a CCD camera. The wing size was measured using a Drosophila wing digital twin (https://datamarkin.com/models/drosophila-wings-beta) developed by Datamarkin. Areas 0, 1, and 2 were quantified as the anterior parts, whereas areas 4, 6, 7, and 8 were quantified as the posterior parts, respectively.

The data are presented as the means ± SEM. p-values were calculated using a one-way ANOVA, followed by Sidak’s multiple comparison test for selected pairs (Figure 1—figure supplement 2B–D; Figure 5K), or Dunnett’s multiple comparison test (Figure 5G, I; Figure 5—figure supplement 1A) using GraphPad Prism 8. NS: p ≧ 0.05, †p < 0.05.

All plasmids and the antibody generated in this study are available from the corresponding authors upon reasonable request. The transgenic fly line mStayGold::C4 linker::V5::Bruce^3xP3-DsRed is available from the Kyoto Stock Center (DGRC Stock Number 119866).

The funders had no role in study design, data collection, and interpretation, or the decision to submit the work for publication.

We would like to thank G W Davis, R Carthew, B Hay, C-H Cheng, H Richardson, M Suzanne, D W Williams, S K Yoo, E Baehrecke, H Steller, Y Hiromi, N Fujita, H D Ryoo, the FlyORF Zurich ORFeome Project, Bloomington Drosophila Stock Center, Vienna Drosophila Resource Center, Kyoto Stock Center, and NIG-FLY for providing the fly strains. We thank Datamarkin for providing the Drosophila wing digital twin tool. We thank Dr. S Hirayama, Dr. S Murata, and the One-Stop Sharing Faculty Center for Future Drug Discovery at the Graduate School of Pharmaceutical Sciences, University of Tokyo, for the LC–MS/MS analysis. We thank Miura’s lab members for their technical assistance and discussions; in particular, K Takenaga for preparing the fly food and R Takamoto for experimental support. We thank Editage (https://www.editage.com/) for the English language editing of this manuscript. This work was supported by grants from the Ministry of Education, Culture, Sports, Science, and Technology of Japan (KAKENHI Grant Numbers 19K16137, 21K15080, 22H05586, and 23K05747 to NS and 21H04774, 23H04766, 24H00567, and 25H01842 to MM), the Japan Science and Technology Agency (ACT-X, Grant Number JPMJAX2539), the Japan Agency for Medical Research and Development (AMED; Grant Number JP21gm5010001 to MM), and grants from the Takeda Science Foundation and Sumitomo Foundation to NS.

1. Preprint posted: November 5, 2025
2. Sent for peer review: December 4, 2025
3. Reviewed Preprint version 1: January 20, 2026
4. Reviewed Preprint version 2: August 7, 2026
5. Version of Record published: October 19, 2026

You can cite all versions using the DOI https://doi.org/10.7554/eLife.109876. This DOI represents all versions, and will always resolve to the latest one.

© 2026, Shinoda et al.

This article is distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use and redistribution provided that the original author and source are credited.