An EMC–Hpo–Yki axis maintains intestinal homeostasis under physiological and pathological conditions Free

Stem cells in adult tissues constantly proliferate to produce differentiated cells to maintain tissue homeostasis. Disruption of adult stem cell proliferation and differentiation often leads to various diseases, including cancer. Intestinal epithelial cells have a high transformation rate and form a hot spot for the development of malignant colorectal cancer, one of the leading causes of cancer deaths (Aran et al., 2016; Schwitalla et al., 2013). Therefore, in-depth study of the mechanisms of intestinal stem cell (ISC) proliferation and differentiation is a key step to understand tissue homeostasis maintenance and clarify the pathogenesis of colorectal cancer.

Studies have proved that the adult Drosophila intestine is an excellent model for studying tissue homeostasis and tumorigenesis (Banerjee et al., 2019; Kim et al., 2020; Lucchetta and Ohlstein, 2012). The Drosophila intestine is functionally equivalent to the mammalian intestine. They are highly similar in development, cell type, and genetic regulation (Casali and Batlle, 2009; Wang and Hou, 2010). An ISC in the adult Drosophila intestine normally undergoes asymmetric division to produce a new ISC and a daughter cell called an enteroblast (EB) or an enteroendocrine progenitor cell (EEP) (Chen et al., 2018; Zeng and Hou, 2015). Delta (Dl), one of the ligands of Notch, is specifically expressed in ISCs (Ohlstein and Spradling, 2007). Activation of Notch signaling in EBs leads to further differentiation into enterocytes (ECs), whereas low levels of Notch signaling cause EEPs to divide once to generate two enteroendocrine cells (EEs) (Chen et al., 2018; Noah and Shroyer, 2013). In addition to Notch signaling, the proliferation and differentiation of ISCs are also regulated by other evolutionarily conserved signaling pathways, such as JAK/STAT, EGFR, Hpo, Wnt and BMP (Gervais and Bardin, 2017; Guo et al., 2016; Joly and Rousset, 2020; and references therein). Originally identified in Drosophila and evolutionarily conserved from fruit fly to mammals, the Hpo pathway plays an important regulatory role in cell proliferation and fate determination to control organ growth, regeneration, and tumorigenesis. In the canonical Hpo kinase cascade, the Hpo–Sav (Salvador) complex phosphorylates and activates the Wts (Warts)–Mats (Mob as tumor suppressor) complex, and activated Wts in turn phosphorylates and inactivates Yki (Yorkie) (Misra and Irvine, 2018; Zheng and Pan, 2019). Upstream regulators of the Hpo signaling pathway were recently found to form functionally antagonizing condensates to mediate Hpo signaling activation (Wang et al., 2022). Previous studies found that Hpo signaling negatively regulates ISC proliferation and intestinal homeostasis in Drosophila (Karpowicz et al., 2010; Ren et al., 2010; Shaw et al., 2010). However, how Hpo signaling is regulated in intestines is not yet understood.

In eukaryotes, most protein synthesis and folding processes occur in the endoplasmic reticulum (ER). The evolutionarily conserved ER membrane protein complex (EMC) plays an important role in these processes (Jonikas et al., 2009). Analysis of the structure of human EMC showed that it contains nine subunits [EMC1, EMC2, EMC3, EMC4, EMC5 (MMGT1), EMC6, EMC7, EMC8/9 (EMC8 and EMC9 are ~40% identical), EMC10] (Alvira et al., 2021; O'Donnell et al., 2020). The EMC performs multiple functions in organisms by mediating the insertion of the transmembrane domain into the lipid bilayer and acting as a chaperone to aid the folding of other types of membrane proteins in the ER (Bai et al., 2020; Hegde, 2022; Volkmar and Christianson, 2020). EMC deficiency affects various physiological functions of organisms, such as ER stress, lipid homeostasis, autophagy, male fertility, defective phototransduction, and photoreceptor degeneration (Volkmar and Christianson, 2020). However, whether and how the EMC functions in ISC proliferation regulation remain unexplored.

In this study, we investigated the role of the EMC in ISC proliferation under physiological and pathological conditions in adult Drosophila intestine. We find that the EMC is required for ISC proliferation and intestinal homeostasis. Our biochemical and genetic studies reveal that the EMC interacts with and stabilizes Hpo to curb excessive ISC proliferation, thereby maintaining intestinal homeostasis. Importantly, the EMC is compromised in tunicamycin (TM)-treated animals, leading to Hpo destruction and excessive ISC proliferation due to activated Yki. Thus, our data uncover the mechanism underlying the role of the EMC in stem cell proliferation control and tumorigenesis.

In order to identify endogenous factors affecting the homeostasis of adult Drosophila intestine, we conducted a large-scale RNA interference (RNAi) screen using the esgGal4, UAS-GFP, tubGal80^ts (esg^ts) driver, which is expressed in progenitors (ISCs and EBs) in the posterior midgut (Liu et al., 2021; Zhao et al., 2022a,b). The screen identified different subunits of the EMC as candidates for the regulatory control of ISCs. Compared with control intestines, in which esg⁺ cells were frequently observed in isolation or in clusters fewer than three cells, knockdown of different subunits of the EMC (EMC1, EMC3, EMC5, EMC6 and EMC7) led to a dramatic increase in the number of esg⁺ cells and intestinal homeostasis disruption (Fig. 1A-G). The percentage of esg⁺ cells (i.e. the number of esg⁺ cells out of total cells) in these intestines was dramatically increased accordingly (Fig. 1H). As knockdown of these examined components of the EMC showed similar defects, we thus focused on EMC1 and EMC3 for further study. Quantitative RT-PCR (qRT-PCR) and western blot (WB) results confirmed the knockdown efficiency of the RNAi constructs (Fig. S1A-C). RNAi knockdown efficiency and EMC antibody specificity were further confirmed by the immunofluorescence staining results (Fig. S1D-I). To confirm that the defects are specific to EMC depletion, co-expression of a UAS-EMC1-Flag transgene with EMC1^RNAi or a UAS-EMC3-Myc transgene with EMC3^RNAi completely rescued the defects caused by EMC1 or EMC3 depletion (Fig. 1I,J). Moreover, overexpression of EMC1 or EMC3 in progenitors did not cause significant defects (Fig. S2). We further examined the specific role of the EMC in ISCs and EBs by depleting the EMC in ISCs and EBs using DlGal4 and GBE+Su(H)Gal4, respectively. Our data showed that knockdown of EMC specifically in ISCs or EBs showed less-severe defects compared with those of esg^ts>EMC^RNAi intestines (Fig. S3). Thus, these data indicate that, although the EMC does play a role in ISCs and EBs, the observed defects in esg^ts>EMC^RNAi intestines are the combined effects of simultaneous depletion of EMC components in both ISCs and EBs. To confirm the RNAi results, we utilized EMC1 and EMC3 mutants. The size of EMC1^GT1 and EMC3^e02662 mutant ISC mosaic analysis with a repressible cell marker (MARCM) clones was significantly increased compared with control clones (Fig. 1K-N, Fig. S4A-G; Lee and Luo, 2001). Further, the number of Dl⁺ cells and Pros⁺ cells changed in EMC1/3 ISC MARCM clones (Fig. S4H,I). Moreover, a significant increase in the number of mitotic cells was observed in EMC1/3-depleted intestines, suggesting that the dramatic increase in esg⁺ cells was caused by excessive ISC proliferation (Fig. 1O-R). Altogether, these data indicate that the EMC negatively regulates ISC proliferation under physiological conditions.

Next, we characterized the identity of these extra esg⁺ cells in esg^ts>EMC^RNAi intestines. We first stained the esg^ts>EMC^RNAi intestines with antibodies against Dl (an ISC marker) and Pros (an EE marker) after 7 days at 29°C. The number and percentage of Dl⁺ cells in esg^ts>EMC^RNAi intestines were significantly increased compared with those in control flies, whereas the number and the percentage of EEs were only affected in intestines depleted of some EMC subunits (Fig. 2A-H). The increase of Dl⁺ cells was independently verified with a Dl antibody, with Dl-lacZ and Dl-GFP reporters in esg^ts>EMC1^RNAi and esg^ts>EMC3^RNAi intestines (Fig. 2I-L, Fig. S5). Consistent with this, knocking down EMC components specifically in ISCs also resulted in accumulation of Dl⁺ cells; interestingly, although the number of mature EE cells was reduced, many Dl>GFP⁺Dl⁺Pros⁺ cells were observed in these intestines, indicative of EEP identity (Fig. S3A-H). Furthermore, the number and percentage of EBs [by GBE+Su(H)-lacZ] were also dramatically increased in esg^ts>EMC1^RNAi and esg^ts>EMC3^RNAi intestines compared with those in control flies (Fig. 2M-Q). Although none of the esg⁺ cells in control flies expressed the mature EC marker Pdm1 (Nubbin), many esg⁺ cells within the esg>GFP clusters expressed Pdm1 in esg^ts>EMC^RNAi intestines, indicative of intestinal homeostasis loss (Fig. 2R-U). Collectively, these data show that intestinal homeostasis is disrupted as a result of excessive ISC proliferation in EMC-defective intestines.

We then examined whether ectopic expression of the EMC could inhibit ISC proliferation. Ectopic expression of EMC1 or EMC3 in progenitors led to no obvious defects under physiological conditions (Fig. S2). We further examined whether ectopic expression of EMC components could inhibit ISC proliferation when ISCs are under rapid proliferation caused by ectopic signaling activation and/or under stress conditions. The results showed that overexpression of EMC1 or EMC3 partially inhibited the excessive ISC proliferation caused by ectopic JAK/STAT and EGFR signaling (Fig. S6A-O). However, we found that overexpression of EMC1 or EMC3 could not inhibit tissue repair induced by dextran sulfate sodium (DSS) administration and rapid ISC proliferation due to defective Notch signaling (Fig. S6P-W).

We then examined the expression pattern of the EMC in adult Drosophila intestines using EMC1- and EMC3-specific polyclonal antibodies (Fig. S1C-I) (Satoh et al., 2015). We found that both EMC1 and EMC3 are mainly expressed in progenitors, which is consistent with the defects observed upon their absence in progenitors (Fig. S7A,B). EMC1 and EMC3 are also expressed in other intestinal cell types, such as ECs and EEs (Fig. S7C-F), indicating that they may function in those intestinal cells as well. Interestingly, we found that the protein levels of EMC1 and EMC3 were largely unaffected when JAK/STAT and EGFR signaling were ectopically activated (Fig. S8).

Consistent with their ER localization, transiently expressed EMC1-GFP and EMC3-Myc localized in the ER in progenitors (Fig. S9) (Jonikas et al., 2009). As EMC components were first identified based on their ability to interact with ER-associated degradation (ERAD) components (Christianson et al., 2012; Jonikas et al., 2009), we then explored whether EMC depletion caused ER stress in progenitors. We examined the activation of the three distinct branches of ER stress: ATF6/Hsc, IRE1/Xbp1 and PERK (PEK)/eIF2α (Hetz and Papa, 2018; Yoo et al., 2017). No significant changes were observed in the levels of Hsc3 (Hsc70-3), Xbp1 and p-eIF2α in EMC component-knockdown guts compared with controls (Fig. S10). These data indicate that ER stress is not activated in the absence of the EMC in progenitors and that the observed intestinal homeostasis disruption in these intestines is unlikely to be a consequence of ER stress activation.

To address how the EMC regulates ISC proliferation, we performed co-immunoprecipitation (co-IP) and liquid chromatography-tandem mass spectrometry (LC-MS/MS) experiments to identify EMC1-interacting proteins by transiently expressing Flag-tagged EMC1. Most of the EMC subunits were identified, indicative of successful IP-MS (Fig. S11). Among the candidate interactors, we identified Hpo, the core protein kinase in the Hpo signaling pathway. Previous studies have shown that the Hpo pathway is involved in intestinal homeostasis regulation and tumorigenesis (Jin et al., 2013; Karpowicz et al., 2010; Poernbacher et al., 2012; Xu et al., 2019). We performed co-IP experiments to confirm the association between the EMC and Hpo, which showed that EMC1 and Hpo reciprocally associate with each other in Drosophila S2R+ cells (Fig. 3A,B). Furthermore, in vivo co-IP results showed that both transiently expressed and endogenous EMC1 reciprocally associate with transiently expressed and endogenous Hpo, respectively (Fig. 3C-E). The co-IP results showed that transiently expressed EMC3 also associates with Hpo in vivo (Fig. 3F). Consistent with this, endogenous EMC1 and EMC3 colocalize with endogenous Hpo in vivo (Fig. 3G-H″). Attempts to establish the interaction domains between EMC1 and Hpo were hindered as truncated proteins of EMC1 and Hpo were unstable and could not be detected when expressed in S2R+ cells, thus preventing examination of domain interactions between EMC1 and Hpo. Further, transiently expressed EMC components also colocalized with Hpo (Fig. S12). These data demonstrate that the EMC associates with Hpo in vivo. As the activity of Hpo is regulated by phosphorylation modification (Boggiano et al., 2011; Poon et al., 2011), we wondered whether the phosphorylation status of Hpo would affect its association with the EMC. It is interesting to note that both Hpo-T195A (phosphorylation dead form of Hpo) and Hpo-T195E (phosphorylation mimic form of Hpo) associate with EMC1, indicating that Hpo associates with the EMC independently of its phosphorylation status (Fig. 3I) (Jin et al., 2012). Altogether, these data show an association between the EMC and Hpo, suggesting that the EMC may control ISC proliferation through Hpo.

The defects observed in EMC-defective intestines mimic those of hpo-depleted and yki^CA (constitutively active yki)-expressing intestines, indicating that the EMC may function through the Hpo signaling pathway to control ISC proliferation (Fig. S13). To prove this speculation, we examined the expression of several downstream targets of Hpo signaling. First, we found that expression levels of anti-apoptotic Diap1 (Death-associated inhibitor of apoptosis 1) were significantly increased in esg^ts>EMC^RNAi intestines compared with control (Fig. 4A-G) (Huang et al., 2005; Yang and Choi, 2021). Second, expression levels of bantam (ban) microRNA (by ban-lacZ) were significantly increased in esg^ts>EMC^RNAi intestines compared with control intestines (Fig. 4H-K). Consistent with this, knockdown of EMC components led to a significant decrease in the levels of bantam sensor (Fig. S14A-D) (Brennecke et al., 2003). Third, the expression of CycE was significantly increased in esg^ts>EMC^RNAi intestines (Fig. S14E-H). Fourth, a series of protein kinase cascade reactions led to the final phosphorylation of Yki in the Hpo signaling pathway; phosphorylated Yki (p-Yki) is inactive as it cannot enter the nucleus to regulate the expression of downstream target genes (Avruch et al., 2012; Huang et al., 2005). The levels of p-Yki were significantly reduced when EMC3 was systemically depleted, whereas the levels of total Yki protein remained unchanged, indicating that Yki signaling is activated in the absence of the EMC (Fig. 4L). These data indicate that the EMC likely functions through the Hpo signaling pathway to regulate ISC proliferation. Further supporting this conclusion, simultaneous knockdown of each of these Yki downstream target genes or ectopic expression of a bantam sponge in esg^ts>EMC^RNAi intestines partially suppressed ISC proliferation (Fig. S15) (Huang et al., 2014, 2005; Ren et al., 2013). Collectively, these observations indicate that Hpo signaling is inactivated and Yki is activated in the absence of the EMC, thereby leading to increased ISC proliferation and disruption of intestinal homeostasis.

How does the EMC affect the Hpo signaling pathway? We investigated whether EMC knockdown caused any changes in Hpo protein levels. We examined the levels of endogenous Hpo protein using a reporter carrying a fosmid expressing GFP-tagged Hpo under its endogenous promoter (Sarov et al., 2016). Compared with control, the levels of Hpo protein in progenitors were significantly reduced upon EMC knockdown (Fig. 4M-O). Furthermore, western blot data showed that the levels of Hpo protein were diminished upon systemic EMC3 depletion, whereas the transcription levels of hpo were largely unaffected (Fig. 4P,Q). These data indicate that the EMC associates with and stabilizes Hpo protein to limit ISC proliferation under physiological conditions, rather than affecting hpo transcription. We further examined the fate of Hpo protein in the absence of the EMC. Interestingly, the levels of Hpo protein were significantly restored when the function of the 26S proteasome was further compromised (by depleting Rpn1) in esg^ts>EMC^RNAi intestines, suggesting that Hpo protein is finally degraded by the 26S proteasome in the absence of the EMC (Fig. 4O,R-R″). Knockdown of EMC3 did not alter lysosomal morphology, supporting the notion that Hpo is degraded by the 26S proteasome, but not degraded by the autophagy/lysosome pathway (Fig. S16). Altogether, these data show that the EMC positively regulates the Hpo signaling pathway by stabilizing Hpo protein.

We further examined whether the EMC is also required in ECs for intestinal homeostasis control. We depleted EMC components in ECs using the EC-specific driver Myo1AGal4^ts. We observed a striking accumulation of esg⁺ cells in Myo1A^ts>EMC^RNAi intestines compared with control (Fig. 5A-G). In agreement with the drastic increase of esg⁺ cells, the number of esg⁺ cells undergoing mitosis was dramatically increased, suggesting that absence of EMC components in ECs leads to non-autonomous ISC proliferation (Fig. 5H-N). Consistent with this, a significant increase of ISC-like cells (by Dl staining) was observed in these EMC-deficient midguts, whereas no significant changes in the number of Pros⁺ cells were observed (Fig. S17A-E). These data show that the EMC also functions in ECs to control midgut homeostasis.

We then explored whether the defects observed in Myo1A^ts>EMC1^RNAi intestines were also the consequences of defective Hpo signaling. The expression of Hpo downstream target genes (ban, Diap1 and CycE) was significantly increased in Myo1A^ts>EMC^RNAi intestines, indicating that the EMC regulates Hpo signaling in ECs as well (Fig. S17F-K). Consistent with this, the transcription levels of multiple Hpo target genes were significantly increased in Myo1A^ts>EMC^RNAi intestines, as determined by qRT-PCR (Fig. 5O,P). We further investigated whether the EMC also regulates cell proliferation through Hpo signaling in other tissues. Previous studies have shown that Hpo signaling regulates the proliferation of wing imaginal disc cells (Huang et al., 2005; Justice et al., 1995), we examined whether the EMC also regulates the proliferation of wing imaginal disc cells. The size of EMC1 and EMC3 mutant MARCM clones in wing discs was significantly increased compared with that of control clones, indicating that the EMC may also negatively regulate the proliferation of wing imaginal disc cells through Hpo signaling (Fig. S18A-D). To verify this conclusion, we systemically knocked down EMC components using tub^ts and detected the transcription levels of Hpo downstream target genes by qRT-PCR. Our results showed that the transcription levels of Hpo target genes (CycE and Myc) were significantly increased in these animals (Fig. S18E). Collectively, these data indicate that it is likely that the EMC generally functions through Hpo signaling for intestinal homeostasis control and cell proliferation.

We further determined whether the defects observed upon EMC depletion are caused by defective Hpo. Several lines of evidence support the notion that defective Hpo is the cause of the defects observed in EMC-defective intestines. First, further compromising the function of the 26S proteasome in esg^ts>EMC^RNAi intestines significantly restored the levels of Hpo protein and effectively suppressed the increased number of esg⁺ cells observed in EMC-defective intestines (Fig. 4M-O,R, Fig. S19A). Second and importantly, co-expression of Hpo-Flag completely suppressed the defects observed in esg^ts>EMC^RNAi intestines (Fig. 6A-F, Fig. S19B-D). Consistent with this, the number of esg⁺, GBE+Su(H)-lacZ⁺, and Dl⁺ cells in esg^ts>EMC^RNAi; Hpo-Flag intestines was completely restored (Fig. 6A-H, Fig. S19B-E). Meanwhile, the significant increase of mitotic ISCs was also completely suppressed in these intestines (Fig. S19F-I). Interestingly, ectopic expression of the phosphorylated mimic form of Hpo (Hpo-T195E) completely restored the defects observed in EMC-defective intestines, whereas ectopic expression of the phosphorylated dead form of Hpo (Hpo-T195A) failed to rescue the defects resulting from EMC deletion, suggesting that phosphorylated Hpo, but not phosphorylation-dead Hpo, activates downstream Hpo signaling in the absence of EMC components (Fig. S19J-V) (Jin et al., 2012). Moreover, ectopic expression of Hpo effectively reduced the size of EMC mutant ISC MARCM clones (Fig. 6I-N). Finally, the size of EMC3^RNAi ISC MARCM clones was effectively suppressed in the yki mutant (Fig. S20A-D). In addition, further knockdown of yki in EMC1 or EMC3 mutant clones significantly reduced the size of EMC mutant ISC MARCM clones (Fig. 6N-P, Fig. S20D-G). Moreover, co-expression of yki^RNAi, or treating the flies with verteporfin (VP), a small molecule that inhibits TEAD–YAP [Scalloped (Sd)–Yki] association (Liu-Chittenden et al., 2012), also effectively suppressed the defects observed in esg^ts>EMC3^RNAi intestines (Fig. 6Q-V). Consistent with the notion that activated Yki functions downstream of the EMC, overexpression of EMC components could not inhibit the defects induced by yki^CA (Fig. S20H-P). Altogether, these data demonstrate that the EMC restricts excessive proliferation of ISCs to maintain intestinal homeostasis by associating with and stabilizing Hpo protein under physiological conditions.

As the EMC is highly conserved, we investigated whether the function of the EMC in ISC proliferation is evolutionarily conserved. Ectopic expression of human EMC1 (hEMC1) effectively rescued the increased size of Drosophila EMC1 mutant ISC MARCM clones (Fig. 6I,J,W,Y). Similarly, ectopic expression of human EMC3 (hEMC3) also effectively rescued the increased size of Drosophila EMC3 mutant ISC MARCM clones (Fig. 6I,K,X,Y). These results show that the function of the EMC in ISC proliferation is evolutionarily conserved.

The abovementioned results demonstrate the essential role of an EMC–Hpo–Yki axis in ISC proliferation and tissue homeostasis control under physiological conditions. Finally, we investigated whether the EMC is involved in certain pathological or stressed conditions. TM, a canonical ER inhibitor that blocks vital functions of certain transmembrane protein families, causes accumulation of unfolded proteins in the ER and induces ER stress (Stengel et al., 2020; Yang et al., 2022). We fed flies with TM and examined whether TM administration caused any defects in these intestines. Along with changes in ER morphology (possibly indicative of ER stress), the number of esg⁺ cells was dramatically increased in TM-treated animals compared with mock-treated flies (Fig. 7A-B′, Fig. S21A-C). Furthermore, the number of mitotic ISCs was significantly increased in these TM-fed intestines (Fig. 7C-E). These results indicate that intestinal homeostasis is disrupted in these TM-treated intestines as a result of increased ISC proliferation. As the defects observed in TM-treated intestines mimic those of esg^ts>EMC^RNAi intestines, we wondered whether TM administration would accordingly affect the protein levels of EMC components. Interestingly, we found that the endogenous protein levels of different EMC subunits in TM-treated intestines were dramatically decreased compared with mock-treated animals (Fig. 7F-H, Fig. S21D-F). The dramatic decrease of EMC protein levels was further verified by western blot (Fig. 7I). Furthermore, the protein levels of ectopically expressed EMC1 were also dramatically decreased in TM-treated intestines (Fig. S21G-I). These data indicate that TM treatment likely affects the stability of the EMC, thereby resulting in ISC hyper-proliferation.

To investigate further whether Hpo is affected in TM-treated intestines, we examined the protein levels of Hpo in TM-fed intestines. The protein levels of endogenous Hpo (detected by Hpo-GFP and Hpo antibody, respectively) were dramatically reduced upon TM feeding compared with mock-treated animals (Fig. 7J-L, Fig. S21J-L). The dramatic decrease of Hpo-GFP levels was further verified by western blot (Fig. 7M). These data indicate that the protein levels of Hpo are significantly decreased when EMC protein levels are diminished in TM-treated animals. Consistent with the dramatic reduction of Hpo protein, levels of p-Yki were significantly reduced in TM-treated animals, indicative of activated Yki signaling (Fig. 7M-P). Furthermore, the expression of Hpo/Yki downstream target genes was significantly increased in TM-fed flies compared with mock-treated animals (Fig. 7Q). These data indicate that Yki signaling is activated upon TM administration as a result of defective EMC and Hpo. Finally, we examined whether compromised EMC/Hpo is responsible for the defects observed in TM-treated intestines. We found that ectopic expression of EMC/Hpo significantly suppressed the defects observed in TM-treated intestines, indicating that compromised EMC/Hpo is mainly responsible for the defects caused by TM administration (Fig. 7C,D,R-T, Fig. S21M-P,S). Moreover, inhibition of Yki function by either yki^RNAi or VP administration significantly suppressed the defects observed in TM-treated intestines (Fig. 7C,D,U-W, Fig. S21Q-S). Altogether, these results show that the EMC–Hpo–Yki regulatory axis is mainly responsible for the defects observed in TM-treated animals.

Activation of signaling pathways regulating adult stem cell proliferation must be precisely controlled to maintain tissue homeostasis. Here, we demonstrate that the EMC reduces excessive proliferation of ISCs to maintain intestinal homeostasis under physiological conditions. Our biochemical and genetic data show that the EMC associates with and stabilizes Hpo protein to limit ISC proliferation and maintain intestinal homeostasis. The EMC–Hpo–Yki axis is also required in differentiated enterocytes for intestinal homeostasis and in wing imaginal discs for cell proliferation. Importantly, EMC protein levels are diminished in TM-treated intestines, leading to intestinal homeostasis disruption due to compromised Hpo and subsequent Yki activation. Taken together, our results show that the EMC protects Hpo from destruction to restrain ISC proliferation and maintain intestinal homeostasis under physiological and pathological conditions.

The EMC functions as chaperone and insertase in protein synthesis, folding and transport in the ER and is related to a variety of physiological functions (Chitwood and Hegde, 2019). Previous studies mainly focused on the role of the EMC in formation of the correct transmembrane structure (topology) of transmembrane proteins or the function of the EMC in lipid membrane transfer (Chitwood and Hegde, 2019; Chitwood et al., 2018). In the absence of the EMC, the signal anchor of many nascent chains cannot be inserted to generate the correct topology, resulting in protein misfolding and degradation (Chitwood et al., 2018). Saccharomyces cerevisiae strains lacking multiple components of the EMC exhibit reduced transfer of phosphatidylserine from the ER to the mitochondria, leading to a significant reduction of phosphatidylserine and its derivative phosphatidylethanolamine levels in the mitochondria (Lahiri et al., 2014). Whether and how the EMC is involved in ISC proliferation control and intestinal homeostasis maintenance remain unexplored. Aside from its known functions in affecting the topological formation, folding or transport of various membrane proteins (Volkmar and Christianson, 2020), our results show that the EMC affects the stability of proteins destined for the cytoplasm, such as Hpo, directly or indirectly, to regulate ISC proliferation and intestinal homeostasis. These data suggest that the EMC may affect various aspects of cytoplasmic protein processing in the ER, probably by facilitating the folding, processing and/or transport of these proteins.

Proteomic studies indicate a close connection between EMC and ERAD components, which is responsible for ubiquitylation and degradation of mis-folded proteins (Christianson et al., 2012; Jonikas et al., 2009; Richard et al., 2013). Consistent with previous reports, no obvious ER stress responses (ER^UPR) in progenitors were observed in the absence of the EMC. These data indicate that affected proteins (including Hpo) could be effectively processed by the ERAD pathway and the 26S proteasome in the absence of the EMC; therefore, affected proteins are not accumulated in the ER and ER^UPR is not induced. Alternatively, as our recent study showed that no ER stress was observed in progenitors when the key components of the ERAD pathway were compromised (Liu et al., 2021), it is possible that the ER of progenitors can tolerate temporary accumulation of mis-folded and or unfolded proteins, e.g. Hpo, which can be slowly removed from the ER without ER^UPR induction.

Previous studies have showed that the Hpo signaling pathway plays an important role in cell proliferation, tissue growth and regeneration, and cancer occurrence (Huang et al., 2005; Pan, 2010; Zhang et al., 2009). The Hpo–Sav–Wts pathway restrains cell proliferation and tissue growth by inactivating Yki through phosphorylation. However, how Hpo is regulated and transduces upstream signals remain poorly understood, except that Hpo can be directly phosphorylated by the sterile 20-like kinase Tao at residue T195 in the Hpo activation loop (Boggiano et al., 2011; Jin et al., 2013; Xu et al., 2019). Our results show that Hpo is associated with and stabilized by the EMC, thereby ensuring the growth-suppressive cascade to limit ISC proliferation under physiological conditions. Interestingly, the EMC associates with Hpo regardless of its phosphorylation status at residue T195. Complete rescue of the defects observed in esg^ts>EMC^RNAi intestines by wild-type Hpo and Hpo-T195E show that diminished Hpo is responsible for the defects caused in the absence of the EMC. The notion is further supported by the observation that Hpo-T195A totally failed to rescue the defects observed in esg^ts>EMC^RNAi intestines. Thus, our study identifies the EMC as a component of the regulatory network controlling of the Hpo–Sav–Wts pathway, and therefore stem cell proliferation, tissue homeostasis maintenance, and tumorigenesis.

TM inhibits GlcNAc phosphotransferase (GPT), thereby inhibiting the transfer of N-acetylglucosamine-1-phosphate from UDP-N-acetylglucosamine to dolichol phosphate in the first step of N-linked glycosylation of glycoprotein synthesis. TM has been widely used in the study of glycoprotein synthesis in various biological systems. TM treatment causes accumulation of unfolded proteins in the ER and induces ER stress (Huang et al., 2017; Stengel et al., 2020; Yang et al., 2022). The dramatically decreased protein levels of EMC components in TM-treated intestines can be explained in two ways: (1) some subunits of the EMC are N-glycosylated proteins and TM treatment blocks their glycosylation and proper folding, thus unfolded and/or mis-folded EMC components are degraded; (2) the EMC is somehow affected in TM-treated cells in which ER^UPR is activated and cleared by ERAD or other pathways. The second possibility can be excluded by the observations that when the animals were treated with thapsigargin, a microsomal Ca²⁺-ATPase inhibitor and widely used ER-stress inducer, ISCs underwent rapid proliferation and intestinal homeostasis was disrupted; however, the protein levels of EMC components were unaffected in these thapsigargin-treated animals. In support of the first possibility, a previous proteomic study identified six N-glycans in yeast EMC1 and EMC7 (Bai et al., 2020). These data support the notion that TM administration likely blocks N-glycosylation modification of EMC subunits, which in turn de-stabilizes Hpo, leading to Yki activation and increased ISC proliferation. Meanwhile, a previous study found that TM treatment of mouse liver tissue caused tumors and decreased phosphorylation levels of YAP1 (Yki) (Wu et al., 2015). However, it was not clear how TM administration affects the Hpo signaling pathway. Here, we find that TM treatment primarily affects the stability of the EMC, finally leading to Yki activation due to Hpo destruction. Interestingly, ectopic expression of EMC/Hpo or blocking Yki activation could not completely suppress the defects observed in TM-fed animals. These results indicate that, in addition to the EMC, TM also affects additional factors, which play minor roles in ISC proliferation regulation.

Our results here reveal the crucial relationship between EMC deficiency and tumorigenesis in Drosophila adult intestines, indicating that the EMC may be implicated in tumorigenesis by stabilizing Hpo protein to constrain stem cell proliferation and differentiation. Preliminary analyses of the ICGC (International Cancer Genome Consortium) database showed that aberrant expressions and/or mutations of EMC genes are observed in a variety of human tumors (Figs S22 and S23), suggesting that aberrant EMC may be responsible for the initiation, development and/or metastasis of these tumors. Given that the function of EMC and Hpo/Yki signaling in stem cell regulation is evolutionarily conserved, our findings identify EMC as a potential tumor suppressor.

Flies were maintained on standard media at 25°C. Crosses were raised at 18°C in humidity-controlled incubators, or as otherwise noted. Flies hatched in 18°C incubators (2-3 days old) were picked and transferred to a 29°C incubator, unless otherwise specified. Flies were transferred to new vials with fresh food every day, and dissected at the time points specified in the text. In all experiments, only the female posterior midgut was analyzed. Information for alleles and transgenes used in this study can be found either in FlyBase, TRiP stock center at Tsinghua University (THU) or as noted: esgGal4, UAS-CD8-GFP, tubGal80^ts (esg^ts; gift from N. Perrimon, Harvard Medical School, USA), esgGal4, UAS-CD8-RFP, tubGal80^ts, EMC1^RNAi [FlyBase HMS01055; THU1416; Vienna Drosophila Resource Center (VDRC) 9408], EMC3^RNAi (FlyBase HMS02242; THU4204), EMC5^RNAi (VDRC 48615), EMC6^RNAi (VDRC 49583), EMC7^RNAi (VDRC 8141), EMC1^GT1 (BDSC, 12852), FRT40A-EMC3^e02662 (Kyoto Stock Center, 114504), FRT40A-EMC3/dPob^Δ4 (gift from A. Satoh) (Satoh et al., 2015), bantam sensor (gift from D. Chen) (Brennecke et al., 2003), bantam-lacZ (gift from J. Shen, China Agricultural University, China), yki^RNAi (FlyBase HMS00041/THU0579), FRT42D-yki^B5 (gift from L. Zhang, Shanghai Jiao Tong University, China), UAS-KDEL-GFP [Bloomington Drosophila Stock Center (BDSC), 9898 and 9899], UAS-KDEL-RFP (BDSC, 30910), Dl-lacZ (Dl⁰⁵¹⁵¹), Dl-GFP (BDSC, 59819), UAS-Hpo-Flag, UAS-Flag-Hpo-T195A, UAS-Flag-Hpo-T195E (gift from L. Zhang) (Jin et al., 2012), Hpo-GFP (gift from J. Paster; VDRC 318388) (Sarov et al., 2016), Diap1-lacZ, Rpn1^RNAi (THU1562), hpo^RNAi (FlyBase HMS00006; THU551), UAS-yki^CA-V5 (BDSC, 28817), UAS-His-RFP (gift from Y. Song, Peking University, China). lacZ (β-galactosidase)^RNAi (VDRC, 51446), luc (luciferase)^RNAi (BDSC, 31603) and w (white)^RNAi (BDSC, 33623; from TRiP at Harvard Medical School) were used as control. No significant differences were observed when lacZ^RNAi, luc^RNAi and w^RNAi were induced in different intestinal cell types.

To address gene function in progenitors, esgGal4, UAS-CD8-GFP, tubGal80^ts (esg^ts) was used. The crosses (unless stated otherwise) were maintained at 18°C to bypass potential requirements during early developmental stages. Progeny with the desired genotypes were collected after eclosion (2-3 days old) and maintained at 29°C to inactivate Gal80^ts before dissection and immunostaining. Flies were transferred to new vials with fresh food every day. If possible, several dsRNA or shRNA lines were tested for each gene, and one or two RNAi lines were used for detailed study. The time points at which the flies are analyzed/dissected are indicated in the text.

Clonal analyses were achieved using the MARCM system (Lee and Luo, 2001). The ISC clones were induced by heat shocking 3- to 5-day-old adult flies at 37°C for 60 min. The flies were maintained in an incubator at 25°C and transferred to new vials with fresh food every day. The sizes of the marked clones were assayed at 6 days after clone induction (6D ACI; clones from at least ten midguts for each genotype were assayed).

For standard immunostaining, intestines were dissected in 1× PBS (10 mM NaH₂PO₄/Na₂HPO₄, 175 mM NaCl, pH 7.4), and fixed in 4% paraformaldehyde for 25 min at room temperature. Samples were rinsed, washed with 1× PBT (0.1% Triton X-100 in 1× PBS) and blocked in 3% bovine serum albumin in 1× PBT for 45 min. Primary antibodies were added to the samples and incubated at 4°C overnight. The following primary antibodies were used: mouse mAb anti-Dl [1:50; C594.9B, developed by S. Artavanis-Tsakonas, Developmental Studies Hybridoma Bank (DSHB)], mouse mAb anti-Prospero (1:100; MR1A, developed by C. Q. Doe, DSHB), rabbit anti-β-galactosidase (1:5000, Cappel Laboratories, now part of MP Biochemicals, 55978), mouse anti-β-galactosidase (1:1000; Cell Signaling Technology, 14B7, 2372), rabbit anti-PH3 (1:2000; pSer10, Millipore, 07-424), rabbit anti-GFP (1:1000; Abcam, ab290), rabbit anti-Pdm1 (1:200; gifts from B. Xiao and X. Yang) (Dai et al., 2020; Yeo et al., 1995), rabbit anti-EMC3 (1:500; this study), rat anti-EMC3 (1:100; gift from A. Satoh) (Satoh et al., 2015), rabbit anti-EMC1 (1:500; this study) and mouse anti-Flag (1:1000; Sigma-Aldrich, 2EL-1B11, MAB3118). Secondary antibodies were incubated for 2 h at room temperature. DAPI (Sigma-Aldrich, 0.1 μg/ml) was added after secondary antibody staining. The samples were mounted in mounting medium (70% glycerol containing 2.5% DABCO). All images were captured using a Zeiss LSM780 inverted confocal microscope, and were processed in Adobe Photoshop and Illustrator.

The EMC1 coding region was cloned into the EcoRI site of an attB UAST-Flag vector to generate UAS-EMC1-FLAG. The EMC1 coding region was cloned into the EcoRI and XhoI sites of an attB UAST-GFP vector to generate UAS-EMC1-GFP. The EMC3 coding region was cloned from LD37839 (Drosophila Genomics Resource Center) into the EcoRI and XbaI sites of an attB UAST-Myc vector to generate UAS-EMC3-Myc. The EMC3 coding region was cloned into the EcoRI and XhoI sites of an attB UAST-GFP vector to generate UAS-EMC3-GFP. The hpo coding region was cloned into the EcoRI and XhoI sites of an attB UAST-GFP vector to generate UAS-Hpo-GFP. The coding region of human EMC1 (hEMC1) was cloned into the XhoI and XbaI sites of an attB UAST-V5 vector to generate UAS-hEMC1-V5. The coding region of human EMC3 (hEMC3) was cloned into the EcoRI and XbaI sites of an attB UAST-Myc vector to generate UAS-hEMC3-Myc. Transgenic flies were obtained by germline transformation using ϕX31-mediated site-specific integration with attP site at 86F or 36B.

Female adult flies at age 3 or 4 days were used to perform DSS feeding experiments. Flies were cultured in an empty vial containing chromatography paper wet with 3% DSS (MP Biomedicals, 0216011080) in 5% glucose solution with heat-inactivated yeast.

S2R+ cells were cultured in 10 ml of heat-inactivated FBS medium (Hyclone, SH30278.02; Gibco, 16000-044) supplemented with penicillin-streptomycin solution (Hyclone, SV30010.01) at a ratio of 1:100 in a 25°C incubators for 6-16 h. Transfection was carried out by polyethylenimine (PEI) transfection method when the cell density reached 2-4×10⁶ cells/cm. Two 1.5 ml Eppendorf tubes were added with 500 μl complete medium (Hyclone, SH30278.02), one tube with plasmid (∼2-5 μg) and the other tube with PEI (the ratio of PEI to plasmid is 3:1). After mixing, the medium containing PEI was slowly added into the plasmid medium, and then left for 15∼25 min at room temperature after full mixing. Finally, the above mixture was added drop by drop into the cell culture dish to be transfected, gently shaken and mixed, and cultured in the incubator. The medium was replaced with pre-warmed, fresh complete medium 6-8 h after transfection. The cells were harvested 48-72 h after transfection. Three plasmids were co-transfected for the experiments shown in Fig. 3A (Armadillo-Gal4, UAS-EMC1-Flag and UAS-Hpo-GFP) and three plasmids were co-transfected for the experiments shown in Fig. 3B (Armadillo-Gal4, UAS-EMC1-GFP and UAS-Hpo-Flag). ‘Arm>’ is used to indicate that Armadillo-Gal4 drives the expression of co-transfected UAS plasmids to save space.

To generate polyclonal antibodies, two peptides of Drosophila EMC1 (EQKPRGDVKLLQVSGFADDSSDTAAA and SGSIVEMPWHLLDPRRPIASTTQGREE), one peptide of Drosophila EMC3 (FNNEETGYFKTQKRAPVAQ), and two peptides of Drosophila Yki (cNPPSSHKPDDLEWYKIN and cMQTVHKKQRSYDVISPIQL) were synthesized. The peptides were used to immunize two rabbits each, and the antisera were affinity purified by BAM BIOTECH.

Fly tissues were lysed in RIPA buffer [50 mM Tris-HCl, pH 8.0, 150 mM NaCl, 5 mM EDTA, pH 8.0, 0.5% Triton X-100, 0.5% NP-40, 0.5% sodium deoxycholate, and complete protease inhibitor cocktail tablets (Roche)] on ice for 30 min. After centrifugation (13,000 g for 20 min), lysates were diluted tenfold with RIPA buffer and subjected to immunoprecipitation using anti-FLAG M2 affinity gel (A2220; Sigma-Aldrich). Immunocomplexes were collected by centrifugation (1000 g for 1 min) and washed with 1 ml of RIPA buffer three times. Negative control immunoprecipitations (IPs) were performed for each IP experiment. For western blotting, immunoprecipitated proteins were separated by SDS-PAGE and then blotted onto PVDF membranes. Membranes were stained with primary antibody overnight at 4°C, followed by washing. PVDF membranes were incubated with secondary antibodies conjugated to horseradish peroxidase, then the membranes were scanned using a GE ImageQuant LAS4000mini Luminescent Image Analyzer. Mouse anti-Flag (1:1000; Sigma-Aldrich, 2EL-1B11, MAB3118), mouse anti-Myc (1:1000; Sigma-Aldrich, 9E10), rabbit anti-GFP (1:4000; Abcam, ab290), rabbit anti-Yki S168 (p-Yki) (1:5000; gift from D. Pan) (Dong et al., 2007), rabbit anti-Yki (1:1000; this study), rabbit anti-phospho-MST1 (Thr183)/MST2 (Thr180) (1:1000; Cell Signaling Technology, p-Hpo, 3681S), rabbit anti-MST1/STK4 (1:1000; Abclonal, A8043), rabbit anti-EMC3 (1:1000; this study), rat anti-dPob (1:1000; gift from A. Satoh) (Satoh et al., 2015), rabbit anti-EMC1 (1:500; this study) and mouse monoclonal anti-α-Tubulin or anti-β-Tubulin (1:1000; Abbkine, ABP0128) antibodies were used. HRP-conjugated goat anti-rabbit and anti-mouse secondary antibodies were used (1:10,000; ZSGB-BIO, ZB-2301 and ZB-2305).

Immunoprecipitated proteins from whole-body samples of tub^ts>EMC1-Flag flies were precipitated with 25% trichloroacetic acid for at least 30 min on ice. Protein pellets were washed twice with 500 μl ice-cold acetone, air dried, and then resuspended in 8 M urea, 100 mM Tris, pH 8.5. After reduction [5 mM Tris (2-carboxyethyl) phosphine, 20 min at room temperature] and alkylation (10 mM iodoacetamide, 15 min in the dark at room temperature), the samples were diluted to 2 M urea with 100 mM Tris, pH 8.5 and digested with trypsin at 1/50 (w/w) enzyme/substrate ratio at 37°C for 16-18 h. Digestion was then stopped by addition of formic acid to a final concentration of 5%.

All samples were analyzed using an EASY-nLC 1000 system (Thermo Fisher Scientific) interfaced with a Q-Exactive mass spectrometer (Thermo Fisher Scientific). Peptides were loaded on a trap column (75 μm ID, 4 cm long, packed with ODS-AQ 12 nm-10 μm beads) and separated on an analytical column (75 μm ID, 12 cm long, packed with Luna C18 1.9 μm 100 Å resin) with a 60 min linear gradient at a flow rate of 200 nl/min as follows: 0-5% B for 2 min, 5-30% B for 43 min, 30-80% B for 5 min, 80% B for 10 min (B=0.1% formic acid in acetonitrile). Spectra were acquired in data-dependent mode: the top ten most intense precursor ions from each full scan (resolution 70,000) were isolated for HCD MS2 (resolution 17,500; NCE 27) with a dynamic exclusion time of 30 s. The AGC targets for the MS1 and MS2 scans were 3e6 and 1e5, respectively, and the maximum injection times for MS1 and MS2 were both 60 ms. Precursors with 1+, more than 7+ or unassigned charge states were excluded.

MS data were searched against a Uniprot Drosophila melanogaster protein database (database ID number UP000000803) using ProLuCID with the following parameters: precursor mass tolerance, 3 Da; fragment mass tolerance 20 ppm; peptide length, minimum six amino acids and maximum 100 amino acids; enzyme, trypsin, with up to three missed cleavage sites (Xu et al., 2015). Results were filtered by DTASelect requiring false discovery rate <1% at the peptide level and spectra count ≥2 (Tabb et al., 2002). Proteins identified from the negative control and Flag-EMC1 IP were contrasted using the ‘Contrast’ tool.

TM (APEXBIO, B7417) was diluted with 5% glucose solution to a final concentration of 50 μM. VP (MedChemExpress, HY-B0146) was diluted with 5% glucose solution to a final concentration of 50 μM. VP was added to the filter paper and was put into an empty vial. Flies were fed for 2-3 days in an incubator at 29°C. In the phenotypic rescue experiment, the corresponding genotypes were induced to express at 29°C for 7 days, and then the drug feeding experiment was carried out.

Total RNA was extracted from 30 flies or guts using TRIzol (Invitrogen) according to the manufacturer's instructions. RNA was cleaned using RNeasy (QIAGEN), and complementary DNA (cDNA) was synthesized using the iScript cDNA synthesis kit (Bio-Rad). Quantitative PCR was performed using the Taq Pro Universal SYBR QPCR Master Mix (Vazyme). qRT-PCR was performed in duplicate for each of three independent biological replicates. All results are presented as mean±s.e.m. of the biological replicates. The ribosomal gene RpL11 was used as the normalization control. See Table S1 for primer sequences.

All quantitative data were derived from multiple independent experiments. PH3 numbers were scored manually using a Zeiss Imager Z2/LSM780 microscope for the indicated genotypes. To determine the number of indicated cells per confocal image (the relative number), confocal images of 40× lens/1.0 zoom from a defined posterior midgut region of different genotypes indicated were acquired. The relative number of indicated cell types (such as progenitor cells and Dl⁺ cells) was quantified using ImageJ software. The percentage of indicated cells was calculated as the number of these cells per image out of the total cell number in this image. Fluorescence intensity was measured using ImageJ software for the indicated genotypes. The number of intestines scored is indicated in figures and figure legends. Statistical analysis was performed using unpaired Student's t-test or ordinary one-way ANOVA test using GraphPad Prism 8 software. Graphs were further modified using Adobe Photoshop and Illustrator. The following significance levels were used: *P<0.05, **P<0.01, ***P<0.001, ****P<0.0001.

We are grateful to Norbert Perrimon, Sarah Bray, Rongwen Xi, Jose Pastor, Xiaohang Yang, Xiaolin Bi, Akiko Satoh, Lei Zhang, Jie Shen, Dahua Chen, Xing Wang, Duojia Pan, Yan Song, Lei Xue and Yu Cai. for generous gifts of reagents, the Bloomington Stock Center, VDRC, NIG-FLY Center, TRiP at Harvard Medical School and the TsingHua Fly Center (THFC) for fly stocks, DSHB for antibodies, and Drosophila Genomics Resource Center for cDNA clones. We thank Duojia Pan for critical comments on the manuscript. We thank the College of Life Sciences, Capital Normal University, for providing some core facilities such as confocal microscope.

The peer review history is available online at https://journals.biologists.com/dev/lookup/doi/10.1242/dev.201958.reviewer-comments.pdf