Introduction

Notch signaling is an evolutionarily conserved pathway in metazoans that mediates local cell-cell interactions through the membrane-tethered ligand Delta (Dl) and the membrane receptor Notch. This pathway transduces signals from the cell surface to the nucleus, regulating the transcription of target genes ^1,2. Dl-Notch signaling controls various cell fates and developmental processes, with mutations in the pathway implicated in numerous human diseases, including congenital defects and cancer ^3–5. Given that Notch can transduce, amplify, and consolidate molecular differences to influence cell fate decisions, it is crucial to understand how the strength of the Notch signaling pathway is precisely regulated to ensure proper cell fate determination and tissue development.

Notch signaling is known as a major regulator of cell fate decisions in both mammalian and Drosophila intestinal stem cell (ISC) lineages, as its activity determines the binary fate of intestinal progenitor cells. A high level of Notch activity promotes the specification of absorptive enterocytes, while low or absent Notch activity favors the specification of secretory cell types.^6,7. The Drosophila intestinal stem cell (ISC) lineage in the adult midgut is relatively simple, comprising only one type of secretory cell: enteroendocrine cells (EEs). This simplicity makes it an attractive experimental system for studying Notch signaling in cell fate decisions during epithelial renewal ⁶. The fly ISCs, which specifically express Dl, periodically generate two types of committed daughter cells: enteroblasts, which experience high levels of Notch activation and are primed for the specification of absorptive enterocytes, and enteroendocrine progenitor cells (EEPs), which receive low levels of Notch activation. Typically, EEPs undergo one round of mitosis before terminal differentiation, resulting in the formation of EE pairs ^8–11. The loss of Notch signaling results in a complete blockage of enteroblast differentiation, leading to the accumulation of ISCs, EEPs and EEs in the intestinal epithelium, producing an “EE tumor” phenotype ^8,10. It is suggested that the expression level of Dl in individual ISCs varies, leading to differential levels of Notch activation in the immediate daughter cells ⁹. This variation in Dl expression may result from dynamic bidirectional Dl-Notch signaling between ISCs and EEPs, which in parallel or in conjunction with an intrinsic feedback mechanism in ISCs, guides a periodic activation of an EE fate inducer Scute, thereby periodic generation of EEPs from ISCs ^10,12,13. The transient activation of Scute induces irreversible Prospero expression in EEPs through a Phyl-Sina-Ttk69 regulatory cascade, initiating terminal differentiation into EEs ^10,14–16. During the period of Scute inactivation, ISCs only generate enterocyte-committed enterobalsts by default via Dl-Notch mediated lateral inhibition ^8,17. In newly enclosed flies, increased lipid intake from food can alter the membrane trafficking of Dl, subsequently increasing the frequency of EE specification from ISCs ¹⁸. Thus, both intrinsic and extrinsic mechanisms are involved in regulating Dl-Notch signaling to orchestrate stem cell fate decisions.

To better understand fate regulation in the Drosophila ISC lineage, here we conducted a forward mosaic genetic screen in the adult Drosophila midgut, focusing on the right arm of the second chromosome to identify genes whose mutations could disrupt ISC fate decisions. From this screen, we identified several complementation groups that result in EE tumor phenotypes. While most of the identified gene loci correspond to known components of the Notch signaling pathway, we discovered one gene locus not previously associated with Notch signaling. Further analyses revealed a specific GSL metabolite that regulates Notch signaling and cell fate decisions in the intestinal epithelium, and this function appears to be conserved from Drosophila to mice.

Results

A mosaic genetic screen identifies GlcT as a tumor suppressor in the adult Drosophila midgut

We performed chemical mutagenesis on chromosome 2R (carrying FRTG13/42B) and used the MARCM system to induce mutant clones in the adult midgut (Supplementary Figure 1). We screened approximately 10,000 lethal lines and identified totally 12 mutations across six complementation groups that exhibited a “small cell tumor” phenotype, characterized by a significant accumulation of diploid cells within the clones (Supplementary Figure 1 and supplementary Table 1). Co-staining with the ISC marker Dl and the EE cell marker Pros revealed that most of these tumors contained an excessive number of ISCs and EEs, the EE tumor phenotype commonly observed in Notch mutant clones ^11,19. Further genetic mapping showed that three complementation groups were associated with genes known to function in the Notch pathway: mam, Gmer, and O-fut1 ^20–22. One complementation group, comprising two alleles (EA30 and E230), mapped to the GlcT gene (see details in Methods), while the remaining two complementation groups, each consisting of one allele, remain undetermined (Supplementary Table 1). GlcT encodes glucosylceramide synthase, an enzyme that catalyzes the formation of glucosylceramide, the core component of glycosphingolipids (GSLs). Both alleles carried one or more missense mutations resulting in the amino acid replacements E292K (for EA30) and L395P/S397T (for E230, Figure S1D). The two mutations produced a similar tumor phenotype in terms of tumor size and the composition of EEs in the tumors (Figure 1B-D). The large polyploid cells were still present in the tumors, indicating that the multipotency of ISCs was retained, but the differentiation of ISCs may have been biased towards the EE fate commitment. Additionally, GlcT^Δ8, a previously characterized null allele of GlcT ²³, exhibited a similar EE tumor phenotype (Figure 1B-D). Expressing a UAS-GlcT transgene in the mutant clones fully prevented the tumor phenotype (Figure 1E, F). Furthermore, depleting GlcT in progenitor cells using RNAi with esg-GAL4^tsalso led to overproliferation of ISCs and excessive generation of EEs in the midgut epithelium (Figure 1G-I). Taken together, these data suggest that GlcT acts as a tumor suppressor in adult Drosophila midgut. Loss of GlcT induces ISC over-proliferation and excessive EE differentiation, resulting in an EE tumor or “neuroendocrine tumor (NET)” phenotype.

Figure 1.

Genetic analysis of the GSL synthesis pathway components reveals specific tumor-suppressive activity for egh, in addition to GlcT

GlcT is involved in the initial step in the GSL biosynthesis pathway, which synthesizes glucosylceramide by transferring glucose to ceramide (Figure 2A). The loss of GlcT could result in the accumulation of ceramide or the absence of glucosylceramide and the downstream GSL metabolites. Ceramide accumulation has been linked to promoting cell apoptosis ²⁴, which could potentially explain the ISC over-proliferation phenotype since apoptotic cells can induce compensatory ISC proliferation ²⁵. However, we did not observe a significant increase in apoptotic cells within the mutant clones (Figure 2B). Moreover, the overexpression of p35, an inhibitor of cell apoptosis, failed to suppress or alleviate the tumor phenotype (Figure 2B). Ceramidase (CDase) catalyzes ceramide into shingosine and free fatty acid, and overexpressing CDase can in theory downregulate ceramide levels. We therefore introduced CDase expression in the mutant clones, but this intervention also did not alleviate the tumor phenotype (Figure 2B). Hence, the EE tumors developed from GlcT mutant clones is unlikely caused by the accumulation of ceramide.

Figure 2.

In the GSL biosynthesis pathway in Drosophila, glucosylceramide undergoes sequential modifications catalyzed by Egghead (Egh) and Brainiac (Brn), which add the second (Mactosyl-) and third (GlcNAc-) glycosyl residues, respectively (Figure 2A) ^26,27. This leads to the synthesis of Mactosylceramide (MacCer) and GlcNAcβ1-3Manβ1-4Glcβ1Ceramide. Complex GSL biosynthesis involves the participation of β4-N-Acetylgalactosyltransferase (β4GalNAcTA/TB) and α1,4-Galactosyltransferase (α4GT1/2) in the fourth and fifth steps of GSL sugar chain elongation ^28,29. These enzymes catalyze the formation of GalNAcβ1-4GlcNAcβ1-3Manβ1-4Glcβ1Ceramide and GalNAc-α1-4GalNAcβ1-4GlcNAcβ1-3Manβ1-4Glcβ1Ceramide, respectively.

To understand the consequences of mutations in these GSL biosynthesis pathway genes, we generated mutant MARCM clones for each of the above-mentioned genes and examined their effects. Interestingly, mutations in egh, assessed using three different loss-of-function alleles, consistently resulted in cell over-proliferation and excessive EE phenotypes, similar to those caused by the loss of GlcT (Figure 2E-G). In contrast, none of the mutant clones of brn, β4GalNAcTA, or α4GT1 exhibited any noticeable abnormalities in clone size or percentages of EEs, when compared to the control group of wild-type clones (Figure 2E-G). Therefore, GlcT and Egh, but not other enzymes in the GSL biosynthesis pathway, exhibit specific tumor suppressive activity in the adult Drosophila midgut.

The EEC tumor phenotype is a result of MacCer deficiency

As Egh is necessary for the biosynthesis of MacCer, the specific phenotype resulted from the loss GlcT and Egh but not Brn could be attributed to the absence of MacCer. To test this hypothesis, we fed flies carrying GlcT mutant clones with Lactosylceramide (LacCer), an analog of MacCer that functions similarly to MacCer in mammalian cells. Remarkably, LacCer significantly suppressed the overgrowth and excessive EE phenotype in the anterior midgut clones (Figure 3B, C). The suppression effect was also significant, though relatively mild, in the posterior midgut (Figure 3B, C). This difference may be attributed to substantial absorption or metabolism of LacCer prior to its entry into the posterior midgut. Therefore, among the various GSL metabolites, MacCer and LacCer appear to possess specific tumor-suppressive activities in the Drosophila midgut.

Figure 3.

Loss of GlcT leads to reduced activation of Notch signaling in ISC progenies

In adult Drosophila midgut, the proliferation and differentiation of ISCs are regulated by multiple signaling pathways. Among these, the EGFR/Ras/MAPK and JAK/STAT signaling pathways play significant roles in promoting ISC proliferation during intestinal renewal and regeneration following tissue damage or infection ^30,31. Additionally, Notch signaling not only regulates ISC proliferation but also controls fate decisions between EE/EC during ISC differentiation ^19,32,33. Loss of Notch leads to increased ISC proliferation and excessive EE generation, phenotype similar to those caused by GlcT mutations. Therefore, we examined whether the loss of GlcT affects the activities of these signaling pathways.

Normally, among the intestinal cells in the midgut epithelium, progenitor cells including ISCs and enteroblasts, exhibit low but specific activation of EGFR and JAK/STAT activities ^30,31,34. The growth of GlcT mutant clones appeared to cause a general increase of EGFR and JAK/STAT activities within the intestinal epithelium, as many differentiated cells also exhibited EGFR or JAK/STAT signaling activation (Supplementary Figure 2), which is likely due to stress responses induced by tumor growth ³⁵. Nevertheless, we observed that pERK, a direct indicator of EGFR/Ras/MAPK signaling activity, was not significantly increased in GlcT mutant cells compared to cells outside the clones (Supplementary Figure 2). Similarly, the nuclear expression of STAT92E, which reflects JAK/STAT pathway activation, did not show significant changes in GlcT mutant cells compared to normal cells outside the clones (Supplementary Figure 2). However, the expression of NRE-lacZ, a reporter for Notch pathway activation, was specifically down-regulated in GlcT mutant clones. Most cells within the clones exhibited lower LacZ expression levels compared to NRE-lacZ positive cells outside the clones (Figure 4A, B). Previous studies have shown that while the loss of function of Notch completely blocks EC differentiation from ISCs, a reduction in Notch activity does not necessarily hinder EC differentiation from ISCs. Instead, it may lead to a dose-dependent increase in the ratio of differentiated EEs to ECs, depending on the severity of Notch loss ¹⁹. Therefore, the increased EE to EC ratio and the reduced NRE-lacZ expression observed in GlcT mutant clones collectively suggest that the mutant phenotype may be caused by a reduction in Notch signaling activity.

Figure 4.

In the ISC lineage, Notch signaling not only guides EC/EE fate decisions, but is also participated in regulating EE subtype diversity. Typically, each EEP undergo one round of asymmetric cell division to yield a pair of distinctive EEs, an allatostatin C (AstC)⁺ class I EE and a tachykinin (Tk)⁺ class II EE ¹⁰. The loss of Notch specifically compromises Mirr-dependent specification of class II EE generation, resulting in all EEs in Notch mutant clones becoming AstC⁺ class I EEs ^16,36. We found that in GlcT mutant clones, virtually all EEs were positive for AstC but TK (Figure 4C), further supporting the idea that Notch signaling is compromised in these clones.

We next tested whether forced activation of Notch in GlcT mutant clones could suppress the EE tumor phenotype. To achieve this, we expressed N^intra, an active form of Notch, in GlcT mutant clones. As expected, this resulted in the complete suppression of cell proliferation in GlcT mutant clones and promoted the differentiation of these cells into ECs, leading all clones to become single-cell clones that often contained a polyploid cell (Figure 4D). This finding suggests that N^intra is epistatic to GlcT in the signaling transduction pathway, indicating that GlcT may regulate Notch signaling at the ligand/receptor level.

GlcT regulates the endocytic trafficking of Delta

The observations above collectively suggest that full activation of Notch signaling in the ISC lineage may require a specific GSL: Mac/Lac-Cer. As transmembrane proteins, the proper internalization and recycling of Dl and Notch are important for the proper activation of Notch signaling ^37,38. Specific lipids have been implicated in the formation of micro-domains on the cell membrane, which play a role in regulating cell-cell signaling ³⁹. Therefore, we asked whether the loss of MacCer caused by GlcT mutation could disrupt Notch signaling by affecting the stability, endocytosis, or recycling of either Dl or Notch.

Knocking down GlcT in ISCs using Dl-GAL4^ts cased an increase in EEs in the midgut epithelium (Figure 5 A, B). However, when GlcT was specifically knocked down in EBs using NRE-GAL4^ts, there was no noticeable increase in EEs (Figure 5 A, B). This suggests that GlcT is specifically required in ISCs and implies its potential role on the regulation of Dl. It is well-known that the internalization of Dl in the signaling-sending cell is necessary for Notch activation in the signaling-receiving cell ³⁷. To investigate whether the endocytosis process of Dl is altered in GlcT mutant ISCs, we performed an ex vivo Dl internalization assay, following the previously described protocol ⁴⁰. In this assay, we cultured and incubated the freshly dissected gut with anti-Dl antibodies, and the endocytosis of the antibody-labeled Dl was visualized after fixation and immunostaining with secondary antibodies. Interestingly, during the 3-hour period, we observed a significant difference in the subcellular localization pattern of Dl between normal and GlcT mutant ISCs (Figure 5 C). In GlcT mutant ISCs, the majority of labeled Dl had already been internalized, while in wild-type ISCs, it remained on the cell membrane. This observation suggests that the loss of GlcT causes either reduced stability of Dl on the cell membrane or accelerated internalization of Dl from the cell membrane.

Figure 5.

To further characterize the endocytic trafficking of Dl in GlcT mutant ISCs, we performed conditional depletion of GlcT and examined the co-localization of internalized Dl with the early endosome marker Rab5 and the late endosome marker Rab7, respectively. In comparison to control ISCs, GlcT-RNAi ISCs exhibited a significant increase in early endosomes containing Dl (Figure 5D, F). However, the percentage of late endosomes containing Dl remained similar between wild-type and GlcT-RNAi ISCs (Figure 5E, F). These observations suggest that the internalized Dl in GlcT-RNAi ISCs experiences a delay in early endosomes, which might cause a delay in Dl recycling and consequently a reduction in Dl-Notch signaling activity (Figure 7E).

GlcT shows tissue specificity in regulating Notch signaling activity

Notch signaling has pleiotropic function and is known to be participated in the development of diverse tissue and organs. We therefore asked whether the discovered mechanism above is generally utilized in controlling Notch signaling in diverse tissues and organs. In the wing disc of third instar larvae, Notch signaling is activated by ligands Dl and Serrate at the dorsal-ventral boundary to induce the expression of Cut ⁴¹. We found that in GlcT mutant clones spanning the dorsal-ventral boundary region of the wing disc, the expression of Cut was not significantly altered (Figure 6A). This indicates that GlcT is not essential for regulating Notch-mediated Cut expression in this specific context.

Figure 6.

During oogenesis in the Drosophila ovary, Notch signaling regulates differentiation and transition of follicle cells from the mitotic cell cycle to the endocycle during stage 6 ^42,43. We found that in the developing egg chambers with germline mutations of GlcT, the patterning of the surrounding follicle cells remained largely normal (Figure 6B). However, we observed an abnormal accumulation of Dl protein in cytoplasmic vesicles of the mutant germline cells (Figure 6C), indicating that there is a defect in Dl trafficking in GlcT mutant germline cells. However, this defect is not sufficient to impact Dl-Notch signaling to a degree where follicle cell development is adversely affected. Collectively, these observations suggest that the role for GlcT in Notch signaling does not appear to be a universal mechanism across all developmental contexts. Even in situations where this mechanism is involved, the extent to which GlcT affects Dl-Notch signaling may vary in different developmental scenarios.

Transient loss of UGCG in mouse small intestine causes reduced number of ISCs and increased number of goblet cells

As both the Notch signaling and the metabolic pathway of GSLs are conserved from Drosophila to mammals, we asked whether this mechanism is conserved in the mouse small intestine to regulate the proliferation and differentiation of ISCs. Notch is known to control the absorptive versus secretory cell fate of mammalian intestinal progenitors as well, and the inhibition of Notch signaling in the intestinal epithelium of mouse small intestine leads excessive number of goblet cells ^44,45. However, unlike in Drosophila, Notch positively regulates ISC self-renewal in mice. The inhibition of Notch causes ISC loss, and the excessive activation of Notch promotes the proliferation of intestinal progenitor cells ^45,46. We therefore performed conditional knockout of UGCG, which encodes the mammalian glucosylceramide synthase, in the mouse small intestine and examined the consequences.

It has been previously reported that in the new born mice with intestine-specific knockout of UGCG, the proliferation of enterocytes appeared largely normal. However, 3-4 days after intestinal-specific knockout of UGCG in adult mice, epithelial phenotypes including epithelial hyperproliferation, impaired lipid absorption and the detachment of enterocytes from the basal lamina were observed, and these phenotypes are considered to be consequences of gut barrier disruption ⁴⁷. To separate the primary from the secondary phenotypes, we repeated the experiment by generating UGCG^flox/flox; Vil-CreERT2 mice, and examined the phenotype at 48h post tamoxifen administration, an early timepoint with a hope that the secondary phenotypes have not yet developed.

Consistent with the previous findings, the conditional knockout mice showed a severe diarrhea phenotype and died several days following tamoxifen induction. Interestingly, we observed a few phenotypes at the 48h timepoint that was not observed before. We found that in the mutant intestine, there was a marked increase in crypt depth accompanied by a marked reduction in villus length (Supplementary Figure 3). In addition, an excess population of goblet cells was observed, dispersed along the crypt-villus axis (Figure 7C, D). By examining the expression of Olfm4, which is transcriptionally regulated by Notch in ISCs ⁴⁸, we found that the fluorescence intensity of Olfm4 staining was significantly reduced in UGCG mutant ISCs (Figure 7A, B). It is unclear how the crypt-villus morphology phenotype is developed in UGCG mutant epithelium, but the increased goblet cells and the reduced expression level of Olfm4 are consistent with a Notch phenotype, suggesting a conserved role for glucosylceramide synthase in modulating Notch signaling from Drosophila to mammals.

Figure 7.

Discussion

From an unbiased genetic screen on the right arm of chromosome II, here we have pinpointed four mutant loci that cause an excessive EE phenotype in the intestinal epithelium. Three of them encode known components of the Notch signaling (mam, Gmer, and O-fut1), and the remaining one, GlcT, turns out to be a new Notch regulator as well. This work further reinforces the notion that Notch as the key regulator of the secretory versus absorptive cell fate decisions from ISCs, and reveals an evolutionarily conserved link between GSL metabolism and Notch signaling in regulating intestinal homeostasis.

As GlcT encodes glucosylceramide synthase, which that catalyzes the formation of glucosylceramide from ceramide, we further analyzed egh and brn, which encodes enzymes for the subsequent sequential modifications of glucosylceramide by adding the second (Mactosyl-) and third (GlcNAc-) glycosyl residues, respectively. We found that loss of egh, but not brn, can give rise to a similar excessive EE phenotype. In addition, the phenotype caused by GlcT can be suppressed by feeding with LacCer, which is analogous to MacCer. Collectively, these observations demonstrate that the excessive EE phenotype caused by GlcT or egh mutation is a result of MacCer deficiency.

GSLs constitute a specific category of glycolipids localized on the outer leaflet of eukaryotic cell membranes, and studies have implicated their roles in regulating multiple cellular signaling pathways by serving on the lipid raft to facilitate signaling transductions ^49–52. In particular, there is an evolutionarily conserved phospholipid binding domain (C2 domain) on the Notch ligand that is known to interacts with GSLs, and this interaction facilitates ligand-receptor interactions and thereby promotes robust Notch signaling ^53–55. Our analyses of GlcT mutant cells in the Drosophila midgut suggest that MacCer may represent a key member of GSLs in facilitating Dl-Notch signaling transduction between intestinal stem cells (ISCs) and their immediate progeny to steer binary cell fate decisions. To do so, MacCer may help to modulate the endocytic trafficking of Dl, an event that is known to be essential for Notch activation. We found that Dl on membrane of GlcT mutant ISCs tends be rapidly endocytosed and the resultant endosomes appear to experience a delay in transition from early endosome to late endosome. Prior genetic investigations have linked glycosphingolipids synthesized by α4GT1 to the modulation of Dl ligand activity in vivo. While the loss of α4GT1 does not overtly affect Dl-Notch signaling, it can exacerbate the wing phenotype of haploinsufficient Notch mutants. Furthermore, the overexpression of α4GT1 has been shown to counteract signaling and endocytosis dysfunctions of Dl caused by the inhibition of the E3 ligase Mib1 or Neur activity ⁵⁰. Therefore, the robust genetic results presented in this study provide definitive evidence for a functional role of MacCer and its downstream derivatives in regulating Notch signaling. Future structural studies focusing on the interactions between GSLs and Dl should offer valuable insights into the role of GSLs in modulating Dl-Notch interactions and membrane trafficking processes.

One intriguing observation from our study is the tissue-specific regulatory role of GlcT in Notch signaling. Its loss in ISCs significantly impairs Notch signaling, leading to the formation of EE tumors. However, when GlcT is lost in the germline cyst within the developing egg chamber, it only modestly disrupts the endocytic trafficking of Dl, without conspicuous effects on the differentiation and patterning of surrounding follicle cells. In the context of imaginal discs, the loss of GlcT appears to have no discernible impact on Notch signaling activity at all. This tissue-specific modulation of Notch signaling by GlcT suggests a potential variability in membrane lipid composition across different cell types. Such variability could render certain cells more susceptible to alterations in specific GSLs and their influence on Dl-Notch signaling transduction, while others remain unaffected. As Dl trafficking in Drosophila ISCs can be modulated by dietary lipid content, which impacts the duration and level of Notch activation to control the production of EEs in newly enclosed flies ¹⁸, the sensitivity of ISCs to specific GSL content may serve as a mechanism enabling ISCs to govern the fate of their progeny based on nutrient content.

Taken together, our study identified MacCer/LacCer as a tissue-specific regulator of Dl-Notch signaling by regulating the membrane localization and endocytic trafficking of Dl. As Notch signaling has pleiotropic roles in many developmental processes, diseases and cancer, the findings should contribute to our understanding of Notch signaling transduction and may help to open up new approaches for developing prevention and treatments against human diseases and cancer.

Materials and methods

Drosophila Stocks and mice

Drosophila melanogaster stocks were maintained on standard cornmeal molasses food in a 25°C incubator. Flies for all experiments were maintained at a constant 25°C with 15–25 flies per vial and tossed onto new food every 2–3 days. Drosophila stocks used in this study: hsflp, act-GAL4, UAS-GFP; FRT42B tub-GAL80; hsflp, act-GAL4, UAS-GFP; FRT42D tub-GAL80; and hsflp, Tub-GAL80, FRT19A; act-GAL4,UAS-GFP/Cyo ⁵⁶; hsflp; FRT42D, arm-lacZ/Cyo; esg-GAL4, UAS-GFP ⁵⁶; Su(H)GBE (NRE)-GAL4; and Dl-Gal4 ⁵⁷; Su(H)GBE-lacZ (NRE-lacZ, gift from Sarah Bray); Glc^EA30(generated in this study); Glc^E230(generated in this study); GlcT^{Δ8 23}; Tub-GAL80^{ts 58}; UAS-GlcT (generated in this study); GlcT-RNAi #1 (VDRC, id:44912); GlcT-RNAi #2 (VDRC, id:108064); UAS-CDase (gift from Tao Wang); UAS-p35 (BDSC, #6298); egh^A (BDSC, #52353); egh^B (BDSC, #52354); egh⁷ (BDSC, #77889); brn²²⁸ (BDSC, #7392); β4GalNAcTA-RNAi (VDRC, id:4867); α4GT1-RNAi (VDRC, id:2608); UAS-N^{intra 56}; UAS-GFP-Rab5 (BDSC, #43336).

Mice used in this study include floxed UGCG, which was generated as previously described ⁴⁷, and villin-CreERT2 ⁵⁹. All the research performance underwent strict ethical review and was approved by the Ethics Committee of the National Institute of Biological Sciences, Beijing. Mice were housed at the animal center in the National Institute of Biological Sciences in a Specific Pathogen Free (SPF) facility.

EMS Mutagenesis in Drosophila

3∼5 day-old, isogenized y w; FRT42B(G13) males were subjected to a 10-hour starvation period before being fed with EMS (30 mM in sucrose solution applied on filter papers) overnight. Following recovery, these males were mated with y w; Sco/Cyo virgin females. In the F1 offspring, y w; FRT42B */Cyo males were individually mated with several y w; Sco/Cyo virgin females. The y w; FRT42B */Cyo males and virgin females in the F2 generation were carefully selected and crossed with each other to establish balanced stocks for further genetic studies. Altogether, around 10,000 stocks harboring lethal mutations were obtained and screened.

Deficiency Mapping

The complete Bloomington Deficiency Kit for chromosome 2R (https://bdsc.indiana.edu/stocks/df/dfkit-info.html) was used for initial mapping, which helped to localize both the EA30 and E230 lethal mutations to a cytogenetic location between 58A3 and 58F1. Subsequent mapping with additional deficiency lines encompassing this region allowed the identification of two small deficiency lines, Df(2R)Exel7170 and Df(2R)Excel6078, which failed to complement both mutations. This refinement narrowed the critical region down to 58B2-58C1. There are 13 protein-coding genes within this region, and genomic sequencing data indicated that both lines carried missense mutations on the GlcT gene, as described in the text.

Genomic Sequencing and Analysis Method

As all mutants exhibited lethality in the adult stage, homozygous larvae were selected using a green balancer, and the extracted genomic DNAs were sequenced at BGI Genomics using a library with an insert size of ≤800bp. The sequencing platform employed was the HiSeq 2000. For sequence analysis, all mutants were compared to a reference genome. Subsequently, sequencing reads were assigned to each allele based on heterozygote SNPs retained using a custom Perl script. The candidate heterozygous sites were annotated using snpEff.

Mosaic Analysis with a Repressible Cell Marker (MARCM) Analysis

To generate MARCM clones ⁶⁰, flies of the desired genotype were raised and maintained at 25°C until they reached 7 days of age. Subsequently, they were subjected to a 60-minute heat shock in a water bath at 37°C. Following the heat shock, the flies were fed with standard food supplemented with yeast paste and this feeding regimen was repeated every 2 days before the flies were dissected for further analysis.

GAL4/GAL80^ts System

The GAL4/UAS and GAL80ts systems were utilized as previously described ^58,61. To inhibit the GAL4 system, the crosses were sustained at 18°C when employing temperature-sensitive GAL4-mediated RNAi or gene overexpression. Subsequently, F1 adult flies with the correct genotype were transitioned to 29°C to activate the GAL4 system, thereby triggering RNAi or gene overexpression.

Tamoxifen Treatment in Mice

Cre recombinase activity was induced via intraperitoneal injection of 2 mg tamoxifen (Sigma-Aldrich, T5648), and tissues were collected 48 hours post-injection. Cre-negative littermates served as controls.

Immunofluorescent Staining and Imaging

Adult female midguts were dissected in PBS and fixed for 30 minutes at room temperature in 4% paraformaldehyde. The fixed tissues were then dehydrated in methanol for 5 minutes followed by rehydration in a PBT solution (PBS containing 0.1% Triton X-100). The tissues underwent four washes with 0.1% Triton in 1×PBS (PBST) and were subsequently incubated with PBST and primary antibodies overnight at 4°C. After another round of washing, the samples were incubated for 2 hours with secondary antibodies in PBST. Finally, the samples were stained with DAPI for visualization of nuclei.

Mouse small intestine was dissected out and gently flushed with cold 1× PBS to remove fecal content, and fixed overnight at 4°C in 4% paraformaldehyde (PFA) prepared in 1× PBS. The tissue was then incubated in a 30% sucrose solution at 4°C for 24 hours, and embedded in OCT (Tissue-Tek) before frozen. Frozen sections of 15 μm thickness were prepared for further analysis. For immunostaining, tissue sections were permeabilized with 1% Triton X-100 (Sigma-Aldrich) in PBS for 1 hour at room temperature (RT), followed by blocking with a solution containing 0.2% Triton X-100, 1% bovine serum albumin (BSA), and 3% goat serum in PBS for 1 hour at RT before antibody staining. The sections were mounted in Vectashield mounting media with DAPI (Vector Laboratories) for imaging.

Primary antibodies used in this study: mouse anti-Pros (DSHB #MR1A; 1:300); mouse anti-Dl (DSHB Cat#C594.9B; RRID:AB_528194; 1:300); rabbit anti-AstC (gift from Dr. Dick Nassel; 1:300); rabbit anti-Tk (gift from Dr. Jan-Adrianus Veenstra; 1:300); Rabbit anti-pH3 (CST Cat# 9701; RRID:AB_331535; 1:500); mouse monoclonal anti-β-galactosidase (DSHB, #40-1a; 1:30); mouse anti-Cut (DSHB, #2B10; 1:20); mouse anti-Rab7 (DSHB, 1:300); rabbit anti-pERK (CST, 1:200); rabbit anti-STAT92E (gift from Dr. Zhaohui Wang, 1:300); rabbit anti-DCP-1 (CST Cat#9578S; 1:300); mouse monoclonal anti-EYA (DSHB, #10H6; 1:300); rabbit monoclonal anti-GFP (Invitrogen Cat#G10362; 1:300). Secondary antibodies used in this study include Alexa Fluor 568- or Cy5-conjugated goat anti-rabbit, anti-mouse IgGs (Molecular Probes, A11034-A11036, A10524; 1:300). For nuclei staining, DAPI (Sigma-Aldrich, 1 μg/ml) was used. The preparations were mounted in 70% glycerol, and the slides were stored at -20°C. Imaging was conducted using a Leica SP8 confocal microscope. All captured images were processed and compiled using Adobe Photoshop and Illustrator.

LacCer Treatment

The 2-3 days old female flies of the specified genotype were moved to fresh food layered with a filter paper soaked in a 5% sucrose solution containing 20 μM of LacCer (Sigma-Aldrich CAS#4682-48-8) and this transfer was repeated daily before dissection and analysis.

Antibody Uptake Assay

The anti-Dl antibody Uptake assay was conducted following a previously described protocol ⁴⁰. In brief, the Drosophila gut was dissected and placed in fresh medium containing anti-Delta. The dissecting dish was then placed in a humidified box and incubated at 25 °C for the necessary duration. Subsequently, the medium containing the antibody was removed, and a single wash with Schneider’s Drosophila medium was carried out. The midgut was then fixed directly in the dissecting dish for 20 minutes under a chemical fume hood by adding 1 ml of fixative. Following this, samples underwent three washes with PBT and were then incubated for 60 minutes with fluorescently labeled anti-mouse secondary antibodies in PBT on a rotating platform. After another three washes with PBT, the samples were mounted in 50% glycerol in PBS for analysis under microscopy.

Fluorescence Intensity Statistics

The Image J software (downloaded from https://imagej.nih.gov/ij/) was utilized to measure the fluorescence intensity of the captured images. The corrected fluorescence intensity is determined as the average fluorescence intensity of each cell’s nuclear region minus the average background fluorescence intensity.

Statistical analysis

All quantifications were presented in the form of mean ± SEM. GraphPad Prism 5 software (GraphPad Software Inc.) was used to calculate p-values using an unpaired Student’s t-test. The number of intestines used for calculations was labeled on the figures or in the related figure legends.

Acknowledgements

We thank Drs. Feng Yu and Sanduo Zheng for technical assistance, Drs. Dick Nassel, Jan-Adrianus Veenstra, Zhaohui Wang, and the Developmental Studies Hybridoma Bank (DSHB) for antibodies, Drs. Sarah Bray, Steven Hou, Akiko Satoh, Tao Wang, the Bloomington Drosophila Stock Center (BDSC) and Vienna Drosophila Resource Center (VDRC) for fly stocks, and Dr. Sylvie Robine for the villin-Cre^ERT2 mice. This work was supported by National Key Research and Development Program of China (2020YFA0803502 and 2017YFA0103602 to R.X.) from the Chinese Ministry of Science and Technology.

Additional files

Supplementary Data. Supplementary Figure 1. Cross scheme for the mosaic genetic screen on the 2R chromosome. (A) A schematic illustrating the principles of FLPase-induced mitotic recombination for generating homozygous mutant cells in heterozygous animals. (B) The cross scheme and workflow for the forward genetic mosaic screen conducted in this study. Abbreviation: EMS, ethyl methanesulfonate. Supplementary Figure 2. The EGFR/Ras/MAPK and JAK/STAT signaling activities in GlcT mutant clones. GlcT mutant clones stained with pERK (red, upper panels), a direct indicator of EGFR/Ras/MAPK signaling activity, and STAT92E (red, lower panels), whose nuclear localization is indicative of JAK/STAT signaling activity. Note that pERK was not significantly increased in GlcT mutant cells compared to normal cells outside of the clones, and STAT92E activity was also not obviously altered in GlcT mutant cells compared to normal cells outside the clones. Scale bar: 25 μm. Supplementary Figure 3. The Crypt-Villus morphology in UGCG CKO mice. The duodenal structure exhibits significant abnormalities in CKO mice Small intestine (duodenum) from control and villin-creERT2, UGCG^flox/flox mice stained with K20, which marks enterocytes and goblet cells, and Ki67, which marks proliferative cells. Note that the mutant intestine has shortened villi and elongated crypts. Dotted lines indicate the boundary between the crypt and the villus. Samples were collected at 48h after tamoxifen injection. Scale bar: 100 μm.