Non-apoptotic caspase-dependent regulation of enteroblast quiescence in Drosophila

Caspase malfunction in stem cells often instigates the appearance and progression of multiple types of cancer, including human colorectal cancer. However, the caspase-dependent regulation of intestinal stem cell properties remains poorly understood. Here, we demonstrate that Dronc, the Drosophila ortholog of caspase-9/2 in mammals, limits the proliferation of intestinal progenitor cells and prevents the premature differentiation of enteroblasts into enterocytes. Strikingly, these unexpected roles of Dronc are non-apoptotic and have been uncovered under experimental conditions without basal epithelial turnover. A novel set of genetic tools have also allowed us to correlate these Dronc functions with its specific accumulation and transient activation in enteroblasts. Finally, we establish that the Dronc-dependent regulation of enteroblast quiescence, largely relies on the fine-tuning of the Notch and Insulin-TOR signalling pathways. Together, this data provides novel insights into the caspase-dependent but non-apoptotic modulation of enteroblast differentiation in non-regenerative conditions. These findings could improve our understanding regarding the origin of caspase-related intestinal malignancies, and the efficacy of therapeutic interventions based on caspase-modulating molecules.

conditions without basal epithelial turnover. A novel set of genetic tools have also allowed us to 24 correlate these Dronc functions with its specific accumulation and transient activation in 25 enteroblasts. Finally, we establish that the Dronc-dependent regulation of enteroblast quiescence, 26 largely relies on the fine-tuning of the Notch and Insulin-TOR signalling pathways. Together, this data 27 provides novel insights into the caspase-dependent but non-apoptotic modulation of enteroblast 28 differentiation in non-regenerative conditions. These findings could improve our understanding 29 regarding the origin of caspase-related intestinal malignancies, and the efficacy of therapeutic 30 interventions based on caspase-modulating molecules. 31 32 INTRODUCTION 33 Caspases are cysteine-dependent aspartate-specific proteases commonly associated with the 34 implementation of apoptosis [1]. Despite this canonical function, an emerging body of evidence is 35 also attributing a non-apoptotic and regulatory role to these enzymes in a wide variety of cell types, 36 including stem cells [2]. However, the caspase-dependent control of intestinal stem cell properties 37 and their contribution to human colon cancers remain poorly characterised. Equally unclear are the 38 molecular mechanisms controlling the differentiation of the intermediate intestinal precursors 39 referred to as enteroblasts (EBs), in experimental situations without associated tissue-damage and 40 regeneration. Addressing these questions could improve our understanding regarding the origin and 41 progression of intestinal tumours associated with caspase malfunction. 42 The evolutionary conservation of gene function and ease of gene manipulation in Drosophila 43 melanogaster have been routinely exploited to uncover many genetic networks and cellular 44 processes connected with human diseases [3]. Accordingly, important discoveries regarding 45 intestinal stem cell biology and caspases have been obtained using this model organism [4]. The 46 caspases are expressed as pro-enzymes that only become fully active after one or more steps of 47 proteolytic processing [1,2,[5][6][7][8]. In Drosophila, the apoptosis programme is initiated by the 48 upregulation of different pro-apoptotic proteins (Hid, Reaper, Grim and/or Sickle) [7][8][9][10], which 49 counteract molecular effects of the Drosophila inhibitors of apoptosis, DIAP-1 [11,12] and -2 [13, 50 14]. In pro-apoptotic conditions, the main Drosophila initiator caspase, referred to as Dronc (Death 51 regulation Nedd2-like caspase; caspase-2/9 orthologue in mammals) can interact molecularly with 52 Dark-1 (Apaf-1) forming a protein complex termed the apoptosome. These events facilitate the full 53 activation of Dronc [15][16][17][18], which subsequently leads to the cleavage of the effector caspases 54 (Death caspase-1 , DCP-1 (caspase-7); the death related ICE-like caspase, drICE (caspase-3); Death 55 associated molecule related to Mch2 caspase, Damm and the Death executioner caspase related to 56 Apopain/Yama, Decay). Upon cleavage, functional effector caspases disrupt all of the essential 57 subcellular structures leading to cell death [2,6]. Intriguingly, in a previous report we uncovered a 58 stereotyped pattern of non-apoptotic caspase activation in the Drosophila intestine of unknown 59 origin and functional relevance [19]. 60 The Drosophila intestine comprises of a subset of intestinal stem cells (ISCs), responsible for the 61 renewal of the epithelial intestine [20][21][22]. ISCs can also differentiate upon demand as either 62 intermediate progenitor cells termed enteroblasts (EBs) or fully differentiated secretory cells called 63 enteroendocrine cells (EEs) [23]. The EBs rarely, if ever divide but can terminally differentiate as 64 mature absorptive cells referred to as enterocytes (ECs) [23]. Throughout the last two decades, an 65 abundant body of literature has emerged describing many of the genetic factors controlling the 66 proliferation and differentiation of ISCs into EBs; however, the differentiation pathway of EBs to ECs 67 remains less well characterised. Notch signalling is one of the instrumental signalling cascades 68 permitting the entry of the EBs into the EC differentiation programme. The interaction of the 69 extracellular domain of the Notch receptor with its ligands (either Delta or Serrate) facilitates the 70 activation of this evolutionary conserved signalling cascade [24]. Notch activation culminates with 71 the release and subsequent translocation into the nucleus of the Notch intra-cellular domain 72 (Notch Intra ) [25,26]. The interaction of the Notch intra fragment with several transcription factors 73 governs the transcriptional response in a highly cell-specific manner [27]. Low levels of Notch-74 signalling promote the self-renewal of ISCs, whilst elevated Notch activation stimulates the 75 conversion of ISCs into EB [28]. In addition to other transcriptional effects, high levels of Notch in EBs 76 repress the expression of the tuberous sclerosis protein complex 1 and 2 (TSC-1 and TSC-2) [29]. 77 These proteins are negative regulators of the Insulin-TOR pathway, and therefore naturally limit 78 cellular growth [29]. Conversely, insulin-TOR pathway upregulation in response to Notch activation 79 instigates the entry of EBs into the EC differentiation programme [29,30] Robust non-apoptotic caspase activation pattern in the Drosophila intestine independent of cellular 95 turnover 96 97 Using a highly-sensitive caspase activity sensor (Drice-based-sensor-QF; DBS-S-QF), we previously 98 reported the presence of a stereotyped pattern of non-apoptotic caspase activation in the adult 99 posterior midgut of Drosophila ( Fig 1A) [19]. Following this initial observation, we sought to 100 investigate the potential correlation of this caspase activation with intestinal homeostasis at the 101 cellular level, monitoring the dynamics of cell proliferation and differentiation. To that end, we 102 utilised the ReDDM cell lineage tracing system [34]. This system employs the combined expression of 103 a short-lived GFP-marker and a semi-permanent Histone-RFP-labelling to readily visualise the 104 turnover of the intestinal cells [34]. The short-lived GFP labelling co-exists with the Histone-RFP 105 marker within all undifferentiated intestinal progenitor cells (ISCs and EBs), expressing the Gal4 106 protein under the regulation of the esg promoter (esg-expressing cells, Fig 1B) [34]. However, the 107 silencing of the esg promoter during differentiation stimulates the rapid degradation of the GFP 108 signal whilst the Histone-RFP remains. The stability of the Histone-RFP bound to the DNA, and the 109 absence of GFP signal, allows the identification of differentiated ECs and EEs after the Gal4 protein 110 production ceases [34]. The incorporation of a Gal80 thermosensitive repressor of the Gal4 protein,111 under the regulation of the Tubulin promoter (TubG80 ts ) facilitates the spatial and temporal control 112 of the ReDDM system [34]. We exploited this dual labelling system to distinguish undifferentiated 113 intestinal progenitors (expressing GFP and RFP) from their progeny ( and cell replenishment would be taking place (Fig 1G-I). Furthermore, P35-expressing intestines 130 were cellularly and morphologically equivalent to the non-P35 expressing controls (compare Fig 1B  131 and 1H). These results unambiguously confirmed the quiescent status of the intestines maintained 132 under our experimental regime, and the non-apoptotic nature of the caspase activation detected 133 with the DBS-S-QF reporter [19]. Paradoxically, it also raised the question of whether there was a 134 functional requirement for the described non-apoptotic pattern of caspase activation, since the 135 epithelial integrity of the gut was unaffected by the overexpression of P35. 136 137 138 Dronc prevents the premature differentiation of intestinal progenitor cells 139 140 The lack of cellular and morphological phenotypes linked to the overexpression of P35 could suggest 141 a negligible functional requirement for the non-apoptotic caspase activation previously described. 142 However, P35 overexpression only prevents the activity of effector caspases, and therefore a 143 potential function of the initiator caspases could have been overlooked. To address this question, we 144 created a new Dronc knockout allele using genome engineering protocols [37]. This resulted in a 145 Dronc allele (Dronc KO ) which contained an attP integration site immediately after the Dronc 146 promoter and within the 5'UTR of the gene (Appendix Fig 2A and B). As with previously described 147 Dronc null alleles [17,38], the new mutant was homozygous lethal during early pupal development, 148 and it failed to genetically complement other Dronc mutations (Appendix Fig 2C). The heterozygous 149 insertion of a wild-type Dronc cDNA into the Dronc attP-site gave rise to fertile adult flies that 150 appeared largely similar to their wild-type siblings (Appendix Fig 2D- with the EC maker Pdm-1, and the co-localisation between the GFP and Pdm1 (inset in Fig 2B and 2E  174 and Appendix Fig 3D). These phenotypes also worsened over time (Fig 2C-E). To discard any 175 unspecific/detrimental effect linked to the Suntag-HA-cherry peptide, flies expressing a WT version 176 of this caspase member tagged with Suntag-HA-Cherry (Dronc KO-FRT Dronc-GFP-Apex FRT-DroncWT-Suntag-HA-Cherry ) 177 failed to show any of the previously described phenotypes (Appendix Fig 2I and 3E). Collectively, 178 these results suggested that Dronc insufficiency in the intestinal progenitor cells causes gut 179 hyperplasia and the premature entry in the EC differentiation program. Since our data was collected 180 in experimental conditions without cellular turnover, Dronc seemed to restrain EBs in a quiescent 181 state. To further characterise the differentiation features of Dronc mutant progenitor cells, we next 182 analysed the expression profile of metabolic enzymes highly enriched in ECs [39]. Interestingly, we 183 found that several of these genes were transcriptionally downregulated in our mutant conditions, 184 while others were upregulated (Appendix Fig 3F-H). This irregular expression of EC markers 185 confirmed the premature and/or defective differentiation status of Dronc-mutant progenitor cells. 186 187 The Dronc-dependent quiescent-state of intestinal progenitor cells requires its enzymatic activity but 188 does not involve the effector caspases activity. Next, we explored whether these functions could be performed by the primary substrates of 205 Dronc in many cellular contexts, the effector caspases (drIce, Dcp-1, Decay and Damm). Since the 206 overexpression of two copies of P35 did not alter the cellular or morphological features of the gut, 207 our previous experiments already discarded a potential enzymatic requirement of effector caspases. 208 However, a potential role of these proteins acting as scaffolding partners could still exist. To 209 investigate this possibility, we simultaneously targeted the expression of all of these downstream 210 caspase members using validated RNAi lines [41]. This experimental design prevents the previously 211 described functional redundancy between the effector caspases [42]. This set of experiments failed 212 to replicate the excess of proliferation and premature differentiation phenotypes observed in Dronc 213 LOF conditions (Appendix Fig 4). Together, these results strongly argue in favour of a Dronc specific 214 regulation of progenitor cell properties through its enzymatic activity, but fully independent of 215 effector caspases and the apoptosis programme (see discussion). 216 217 The non-apoptotic function of Dronc is exclusively required in EBs 218 219 Although our previous experiments uncovered unknown functions of Dronc, they could not 220 discriminate whether these functions were ascribed to the ISCs, the EBs or both cell types. To 221 address this question, we specifically targeted the expression of Dronc in ISCs using the Delta-Gal4 222 driver [43]. As previously shown with esg-Gal4, we obtained a high excision efficiency (81.15%) of 223 the FRT-rescue cassette in ISCs by driving the expression of the Flippase recombinase with the Delta-224 Gal4 line (Appendix Fig 3O-P). However, Dronc deficiency in ISCs did not cause any cellular or 225 morphological alteration of the gut, and no-increase in the number and cell size of Delta-positive 226 cells (Fig 2H and I). To explore the possibility of increased differentiation into EE fate following Dronc 227 LOF, we quantified the number of small nuclei concomitantly the EE cell identity marker, Prospero. 228 No statistical differences were observed in this set of experiments between experimental and 229 control intestines (Appendix Fig 3Q). These results unambiguously indicated a specific functional 230 requirement for Dronc in EBs. They also explained the differentiation bias of progenitor cells towards 231 EC fate utilising the esg-Gal4 driver (see discussion). 232 233 The protein accumulation and transient activation of Dronc in EBs determines its functional 234 specificity 235 236 Our previous data suggested a functional requirement for Dronc in EBs, but the molecular origin of 237 such specificity remained elusive. To elucidate this question, we first explored whether the 238 transcriptional regulation of Dronc could be restricted to EBs, using a newly created Dronc  strain (Appendix Fig 2L). This fly line transcriptionally expresses Gal4 under the physiological 240 regulation of the Dronc promoter, and therefore is a bona fide transcriptional read out of the gene. 241 Dronc was widely transcribed in all of the intestinal cell subtypes (Appendix Fig 5A). This experiment 242 separated the specificity of the Dronc-dependent EB quiescence from the transcriptional regulation 243 of the gene. Next, we investigated whether the Dronc protein level could be differentially regulated 244 in EBs. Our laboratory has created multiple Dronc alleles tagged at the C-terminal with different 245 peptides (e.g HA, Cherry and GFP) that successfully rescue the insufficiency of Dronc amorphic 246 alleles. However, we failed to detect the physiological expression of Dronc using these fly strains in 247 most of the analysed tissues, including the intestinal system. Similar frustrating results were 248 obtained using the short repertoire of validated antibodies raised against the Dronc protein [44,45]. 249 To circumvent this technical issue, we created a new Dronc protein sensor. The new reporter line 250 ubiquitously expresses a catalytically inactive (C318A) and tagged template of Dronc under the 251 regulation of the Actin promoter. This Dronc mutant cannot induce apoptosis, however the tagging 252 at the C-terminus with a modified GFP and Myc facilitates its immunodetection (Appendix Fig 5B). 253 Importantly, the overexpression of this construct generates fertile adult flies without any noticeable 254 developmental or morphological defects. Furthermore, this reporter was able to recapitulate the 255 subcellular localisation of Dronc previously described in other tissues such as the salivary glands and 256 the wing imaginal discs (Appendix Fig 5C) [46]. Interestingly, our construct reported a noticeable 257 accumulation of Dronc protein levels within a subpopulation of progenitor cells despite the absence 258 of cellular turnover in our experimental conditions ( Fig 3A). Furthermore, Dronc accumulation 259 showed a striking overlap with the EB marker Su(H) (Fig 3B). Complementing these findings, we also 260 observed that the specific EB accumulation of Dronc disappeared in tissue-damaging conditions, 261 after exposure to paraquat (Appendix Fig 5D). Considering these findings, we decided to re-evaluate 262 our previous results regarding the activation of the DBS-S-QF sensor, combining this activity reporter 263 line with Su(H). These experiments demonstrated that most of the caspase activation in intestinal 264 progenitor cells was specifically ascribed to the EBs in our experimental conditions (Fig 3C and  265 Appendix Fig 5E). These results strongly correlated the non-apoptotic functions of Dronc with its 266 specific accumulation and transient activation in EBs.

268
Dronc regulates EB quiescence acting upstream of Notch and Insulin-TOR pathway 269 270 As stated, the Notch pathway is one of key signalling cascades involved in the regulation of cell 271 proliferation and differentiation in the Drosophila and the mammalian intestinal system [28,47,48]. 272 Furthermore, non-apoptotic protein-protein interactions have been described between Dronc and 273 Notch pathway regulators (e.g. Numb) in the Drosophila neuroblasts [40]. Considering these 274 precedents, we analysed through classical genetic epistasis the potential genetic interplay between 275 Dronc and the Notch pathway. The inhibition of Notch-signalling in intestinal progenitor cells 276 promotes the expansion of ISCs and EEs, whilst preventing the differentiation of EBs to ECs [49]. 277 Since Dronc LOF facilitates the premature differentiation of EBs, we first investigated whether 278 Notch-signalling deficiency would be able to revert the Dronc mutant phenotype. To that end, we 279 simultaneously targeted the expression of Dronc and Notch in progenitor cells using the esg- Gal4 280 driver. The epithelial features of intestines obtained from these experiments were equivalent to the 281 Notch LOF conditions, indicating that the characteristic increase in cell size and EC features linked to 282 the Dronc insufficiency were supressed (compare Fig 4A-C with Fig 2B-E, Appendix Fig 3A and I-N). 283 However, we did detect a mild increase in cell number in the Dronc LOF intestines, perhaps linked to 284 a potential rescue of apoptosis ( Figure 4C). Conversely, the ectopic activation of Notch-pathway in 285 Dronc mutant progenitor cells promoted their quick conversion into ECs and subsequent elimination 286 from the epithelia (Fig 4D-E). Furthermore, the depletion of intestinal precursors was quicker than 287 with the activation of Notch (N intra ) in the control genetic background (Fig 4F). These results 288 genetically located the function of Dronc upstream of the Notch-pathway, likely acting as a negative 289 regulator of the terminal differentiation programme of EBs to ECs. The Insulin-TOR pathway is 290 required downstream of the Notch-pathway to complete the terminal differentiation of EBs [29,30]. 291 Since our data suggested that Dronc LOF could boost Notch-signalling and ultimately differentiation, 292 we formally tested whether the Insulin-TOR pathway would be genetically downstream of the Dronc 293 and Notch-pathway. To that end, we concomitantly eliminated the expression of Dronc and the 294 Insulin receptor. As expected, the lack of Insulin-TOR signalling rescued the premature 295 differentiation phenotypes triggered by the Dronc LOF (Fig 4G-I) The specific accumulation and activation of Dronc in EBs promotes cellular quiescence 303 304 Over the past two decades, numerous Drosophila studies have utilised either environmental or 305 genetically-induced tissue damaging conditions to decipher the molecular factors controlling the 306 proliferation and differentiation of ISCs into EBs [32,[50][51][52]. However, the differentiation step of 307 EBs into fully differentiated ECs is less understood. We utilised experimental conditions, which 308 ensure the Drosophila intestine remains free of apoptosis and basal cellular turnover during at least 309 the first 7 days post ReDDM activation (Fig 1) to investigate the potential non-apoptotic role for the 310 caspases within intestinal progenitor cells. Strikingly, in this experimental setting we observed a 311 stereotypical pattern of caspase activation that appears to occur to a large extent in EBs (Fig. 3C, 312 Appendix Fig 5E and [19]). Our data demonstrates the correlation of this caspase activation with the 313 initiator caspase Dronc. Furthermore, we demonstrate that the accumulation and activation of 314 Dronc in EBs is essential to prevent the appearance of gut hyperplasia, as well as the entry of these 315 progenitor cells into the EC differentiation programme (Fig 2 and 3). These findings indicate that a 316 sophisticated genetic network controls the differentiation of EBs, and therefore the epithelial 317 homeostasis of the intestine does not rely exclusively on the regulation of ISCs. Based on this, the 318 fine-tuning of EB properties could be more relevant than previously thought for maintaining the 319 intestinal epithelial homeostasis. Importantly, caspase-9 deficiency in human intestinal precursors 320 results in excessive proliferation and poor differentiation [33]. Furthermore, these features are 321 considered a bad prognosis marker for human colon cancer [33]. Together, our findings could help to 322 better explain the origin of these human malignancies. 323 Independently, our data supports the hypothesis that non-apoptotic caspase activation is key to 324 modulate fundamental stem cell properties, such as cell proliferation and differentiation, beyond 325 their role in apoptosis [2,53,54]. Indeed, considering the fast-growing list of examples indicating an 326 implication for the caspases in non-apoptotic functions [2,53,54], apoptosis could be the 327 phenotypically more apparent function of these enzymes, but not necessarily the primary and/or the 328 most relevant. Future evolutionary analysis of these novel non-apoptotic functions in primitive 329 organisms should clarify the primary function of caspases. 330 331 332 The Dronc-dependent EB quiescence relies on its enzymatic activity of Dronc and its ability to 333 modulate Notch signalling, but is independent of the apoptotic pathway 334 335 Our results indicate that the sole presence of Dronc is insufficient to fulfil all of its functions in EBs, 336 and therefore its enzymatic activity is required. Therefore, Dronc does not regulate EB quiescence 337 acting as a scaffold protein but as a proteolytic enzyme. Our experiments also suggest that either the 338 expression or activation of effector caspases is dispensable in order to ensure EB quiescence, but 339 instead an unknown substrate "X" of Dronc must exist ( Fig 5A). Importantly, our genetic epistasis 340 indicates that the Dronc-mediated cleavage of such a factor could directly or indirectly limit Notch-341 signalling. caspase effect helps to restrain the proliferation and differentiation of ISCs after tissue damage 355 (transition 1 in the model of Fig 5B, [32]). Our findings now demonstrate that Dronc is required to 356 control the timely entry of EBs into the EC differentiation programme (transition 2 model Fig 5B). 357 This new function of Dronc is unlikely to be correlated with expression of Brahma, since it is totally 358 independent of effector caspases. However, it is tightly connected with the regulation of Notch-359 signalling and of the Insulin-TOR pathway ( Fig 5A). Collectively, these results illustrate the ability of 360 caspases to modulate in multiple ways the homeostasis of different subpopulations of intestinal 361 precursors without causing apoptosis. In parallel, they suggest the presence of an unknown and 362 highly specific mechanism to activate caspases at sublethal thresholds in different intestinal cell 363 subpopulations.

365
Tumour suppressor role of caspases beyond apoptosis 366 367 The evasion of apoptosis is one of the hallmarks of tumour cells, and reasonably, the caspase 368 regulation of apoptosis is one of the main tumour suppressor mechanism linked to these enzymes 369 [55]. However, caspase activity in the intestinal system seems to be coupled to alternative tumour 370 suppressor mechanisms. Along these lines, caspases can block the excess of proliferation and 371 differentiation of intestinal stem cells [32,56]. This clearly prevents gut hyperplasia as well as 372 tumour prone conditions. Independently, our Drosophila findings indicate that caspases can limit the 373 differentiation of EBs ( Fig 5B). Importantly, these caspase effects are phenocopied by human 374 caspase-9 intestinal precursors and are not linked to apoptosis.  Fig 1A). 416 417 Paraquat treatment 418 The Drosophila were dry starved for 4 hours. Subsequently the flies were transferred to an empty fly 419 vial in which the fly food was replaced by a flug soaked in a solution of 5% sucrose and 6.5mM 420 Paraquat. Flies were left in this vial for 16 hours prior dissection. 421 422 Full description of experimental genotypes 423 424  w 1118 DBS-S-QF, UAS-mCD8-GFP, QUAS-tomato-HA (Fig 1 A, Fig 1B)

425
 w 1118 ; esg-Gal4 UAS-CD8-GFP / CYO; UAS-Histone-RFP TubG80 ts / TM6B (Fig 1 B,C We then subcloned the PCR products into the targeting vector pTV-Cherry. The 5' and 3' homology 496 arms were cloned into pTV-Cherry as a NotI-KpnI and BglII-AvrII fragments, respectively. The Dronc KO 497 behaves as previously described null alleles of the gene and is homozygous lethal during pupal 498 development. The molecular validation of the allele was also carried out by PCR. First, we extracted 499 the DNA from 10 larvae using the Quick genomic DNA prep protocol described in the link below. 500 http://francois.schweisguth.free.fr/protocols/Quick_Fly_Genomic_DNA_prep.pdf. 501 The sequence of the primers used for performing the PCR were: The PCRs was made using the Q5 High-Fidelity polymerase from New England Biolabs (NEB, 508 M0492L). 509

RIV-Dronc KO-FRT-DroncWT-GFP-Apex-FRT QF . Conditional Dronc allele followed by QF 534
We first modified the WT cDNA of Dronc adding a GFP-Apex2 chimeric fragment. This fragment was 535 appended in frame before the stop codon of Dronc in order to facilitate biochemical approaches not 536 used in this manuscript. The cDNA of Dronc and the GFP-Apex2 chimeric fragments were amplified 537 by PCR from the plasmids dronc KO-Dronc-WT and pCDNA-conexing-EGFP-APEX2 (addgene #49385), 538 respectively. The sequences of the primers used to that end were: 539 Forward primer 1: Cherry. 579 We first generated two point mutations through gene synthesis (Genewizz) in the wild-type cDNA of 580 Dronc that caused the following amino acid substitutions; C318A and E352A. This version of Dronc is 581 enzymatically inactive (C318A), and cannot be either processed during the proteolytic activation 582 steps of Dronc (E352A). This fragment was subcloned in PUC57-Dronc KO-Dronc-WT-Suntag-HA-Cherry as a BglII-583 XmaI fragment, thus replacing the wildtype version of Dronc by the mutated. Finally, the DNA 584 sequence was transferred to the RIV-Dronc KO FRT-DroncWT-GFP-Apex-FRT QF plasmid as an AvrII-ClaI fragment. 585 Homozygous flies expressing this mutant form of Dronc die during metamorphosis indicating this 586 allele behaves as previously described null alleles. 587  HA-Cherry. 589

RIV-Dronc
We generated a PCR product that deletes the CARD domain of Dronc using the following primers and 590 as template for the PCR RIV-Dronc KO  Cherry. 605 We generated a PCR product using as a template for the PCR RIV-Dronc KO-Dronc WT-Suntag-HA  The PCR product was subcloned in PUC57-Dronc KO-Dronc-Suntag-HA-Cherry as a NotI-EcoRI fragment, thus 614 inserting in frame the wild type version of Dronc in frame with the Suntag-HA-Cherry peptide. 615 Finally, the DNA sequence was transferred to the RIV-Dronc KO FRT-DroncWT-GFP-Apex-FRT QF plasmid as an 616 AvrII-PasI fragment. Heterozygous flies expressing this mutant form rescue the pupal lethality 617 associated with Dronc insufficiency. 618 We first generated one-point mutation through gene synthesis (Genewizz) in the wild-type cDNA of 621 Dronc that causes the following amino acid substitution C318A. This version of Dronc is enzymatically 622 inactive. In addition we appended a Suntag and a HA peptide sequence to the C-terminus, in frame  623 with the open reading frame (ORF) of Dronc. Downstream of the ORF we included the 3'UTR of 624 Dronc present in the genomic locus. Extra restriction sites were added at the 5' and 3' ends of the 625 construct to facilitate future subcloning projects. The entire construct was subcloned in PUC57 as a 626 Not-KpnI fragment. We then opened this vector with SmaI and NheI; this enzymatic digestion 627 eliminates the C-terminal tagging of Dronc (Suntag-HA) whilst retaining the 3'UTR. Using HiFi DNA 628 assembly, we inserted a PCR product that encodes for a modified version of GFP with a Myc tag 629 appended at the C-terminal end. The primers used to amplify the modified GFP-Myc were: 630 Forward primer 1: The template used to obtain the GFP-Myc PCR product was extracted from genomic DNA of flies 640 containing the construct UAS-GC3Ai [58]. The construct was finally subcloned as a NotI-XhoI 641 fragment in an Actin-polyA vector of the lab previously opened with NotI-PspXI. Sequence of the 642 plasmid will be provided upon request until the vector is deposited in a public repository. 643 644 Immunohistochemistry 645 646 Adult mated female Drosophila Intestines were dissected in ice-cold PBS. Following dissection, the 647 intestines were fixed by immersing for 6 seconds in wash solution (0.7% NaCl, 0.05% Triton X-100) 648 heated to approximately 90°C. Subsequently the intestines were rapidly cooled in Ice-cold wash 649 solution. The intestines were then rapidly washed in PBT (0.3%) before blocking for at least 1 hour in 650 1% BSA-PBT (0.3%). Primary antibodies were incubated overnight at 4°C and secondary antibodies at 651 room temperature for two hours, diluted in blocking solution. Primary antibodies used were: Goat 652 Anti