Tubular sclerosis complex gene products TSC1 and TSC2 have evolutionarily conserved roles in cell growth from Drosophila to mammals. Here we reveal important roles for TSC1/2 in regulating intestinal stem cell (ISC) maintenance and differentiation of the enteroendocrine cell lineage in the Drosophila midgut. Loss of either the Tsc1 or Tsc2 gene in ISCs causes rapid ISC loss through TORC1 hyperactivation, because ISCs can be efficiently rescued by mutation of S6k or by rapamycin treatment. In addition, overexpression of Rheb, which triggers TORC1 activation, recapitulates the phenotype caused by TSC1/2 disruption. Genetic studies suggest that TSC1/2 maintains ISCs independently of nutritional status or Notch regulation, probably by inhibiting cell delamination. We show that Tsc1/Tsc2 mutant ISCs can efficiently produce enterocytes but not enteroendocrine cells, and this altered differentiation potential is also caused by hyperactivation of TORC1. Reduced TORC1–S6K signaling by mutation of S6k, however, has no effect on ISC maintenance or cell lineage differentiation. Our studies demonstrate that hyperactivation of TORC1 following the loss of TSC1/2 is detrimental to stem cell maintenance and multiple lineage differentiation in the Drosophila ISC lineage, a mechanism that could be conserved in other stem cell lineages, including that in humans.
The intestinal epithelium is a rapidly renewing tissue and the residing intestinal stem cells (ISCs) are responsible for this high capacity of epithelial regeneration (Barker et al., 2008). Many intestinal malfunctions may be caused by improper regulation of ISCs, yet how ISCs are regulated in vivo is poorly understood. ISC in the Drosophila midgut has recently emerged as an attractive model to study ISC regulation in vivo, because of cell lineage and regulatory similarities to those of mammals (Casali and Batlle, 2009; Karpowicz and Perrimon, 2010; Jiang and Edgar, 2012). Individual midgut ISCs are scattered along a thin layer of basal lamina (Micchelli and Perrimon, 2006; Ohlstein and Spradling, 2006) that separates the single-layered epithelium from the surrounding visceral muscle layer. Multiple signaling pathways, including those involving Notch, Wnt, JAK/STAT, EGFR and insulin are involved in the regulation of ISC maintenance, proliferation and/or differentiation (Jiang and Edgar, 2012). ISCs, the only dividing cells in the epithelium, can be identified by their specific expression of the Notch ligand, Delta (Dl). Upon each asymmetric division of an ISC, it produces a new ISC and another daughter, known as an enteroblast (EB), which will undergo terminal differentiation into either a polyploidy enterocyte (EC) or a diploid enteroendocrine (ee) cell (Fig. 1A) (Micchelli and Perrimon, 2006; Ohlstein and Spradling, 2006). It is suggested that Dl from ISCs directly activate Notch in EBs to promote their terminal differentiation, and differential levels of Notch activation in EBs determine their binary cell fate choice, with high Notch activation favoring EC differentiation and low Notch activation favoring ee cell differentiation (Ohlstein and Spradling, 2007).
ISC division also changes in response to stress conditions such as injury and aging (Biteau et al., 2008; Choi et al., 2008; Amcheslavsky et al., 2009; Jiang et al., 2009), and is affected by nutrition availability (McLeod et al., 2010; O'Brien et al., 2011). Insulin/PI3K signaling is essential to coordinate cell growth with nutritional conditions in Drosophila (Britton et al., 2002). Consistent with this function, ISCs with a compromised insulin signaling pathway become less prolific, even after severe injury (Amcheslavsky et al., 2009). Tubular sclerosis complex human disease gene products TSC1 and TSC2 have an evolutionarily conserved role in cell growth control from Drosophila to mammals (Pan et al., 2004). They form a protein complex as a GAP (GTPase activating protein) to inhibit Rheb (Ras homolog enriched in brain), a small G protein that inhibits target of rapamycin (TORC1); TORC1 is known as master regulator of cell growth and metabolism (Wullschleger et al., 2006). Apart from its role in cell growth, TSC1/2–TORC1 signaling has emerged as an important regulator for stem cells (Russell et al., 2011). TSC1/2 functions unconventionally in Drosophila ISCs to positively regulate their proliferation (Amcheslavsky et al., 2011), but whether the TSC1/2–TORC1 pathway has a role in regulating ISC maintenance or multiple lineage differentiation is unclear.
Here we report a novel role of TSC1/2 in maintaining Drosophila ISCs and regulating the binary lineage choice during ISC differentiation through inhibition of TORC1 activation. This function is important regardless of the nutritional status and is independent of Notch regulation. Our studies suggest that the TSC1/2–TORC1 pathway could have a general role in promoting stem cell maintenance and multiple lineage differentiation in many stem cell lineages, including that in humans.
TSC1/2 is required for ISC maintenance
Tsc1 and Tsc2 are essential for viability, and to study their function in the adult midgut, we generated positively marked epithelial clones that were homozygous mutant for Tsc1 or Tsc2 (Tsc2 is also known as gigas, gig) by using the mosaic analysis with a repressible cell marker (MARCM) system (Lee and Luo, 1999), as previously described (Lin et al., 2008; Lin et al., 2010; Xu et al., 2011). Briefly, females of appropriate genotypes, aged 4–7 days, were collected and heat-shock treated to induce mitotic recombination to generate GFP-marked wild-type or mutant ISCs. Because ISCs in the epithelium are the only dividing cells, the initially marked cells would be either ISCs or EBs. Normally, ISC-marked clones persist over time whereas most of the transient EB clones are lost after 1 week of clone induction, as the differentiated cells are replaced approximately twice a week. Flies were dissected and the midguts were immunostained and analyzed at day 4, 7 and 14 after clone induction (ACI). Quantification of the number of ISC-containing clones (or ISC clones, for simplicity) for each genotype at each time point indicated Tsc1 and gig mutant ISCs are gradually lost over time. In the wild-type control (FRT82B and FRT80B), 86–100% of ISC clones present on day 4 were maintained on day 14 ACI (Fig. 1B,C). However, almost all Tsc1R453X, Tsc1Q87X and gig192 homozygous mutant ISC clones present on day 4 were no longer observed on day 14 ACI, with only 0.7, 4.7 and 1.4% ISC clones remaining, respectively, for each allele (Fig. 1D–I; Fig. 2H). The remaining ISC clones on day 14 generally contained fewer cells, indicating Tsc mutant ISCs are also underproliferative, as previously reported (Amcheslavsky et al., 2011). Some remaining clones no longer contained ISCs, suggesting a recent loss of ISC within the clone (Fig. 1J). The mutant ISCs were also larger, which is consistent with a general role of TSC in cell growth control. These observations suggest that TSC1 and TSC2 are critically required for ISC maintenance and also involved in the regulation of ISC division and growth.
ISC maintenance defects indicate that TSC1 and TSC2 could be required for ISC survival. However, a TUNEL labeling assay, which marks fragmented DNA, showed that none of the Tsc1R453X (n = 49) and Tsc1Q87X (n = 45) mutant ISCs underwent apoptosis (Fig. 1K). Because many Tsc mutant clones on day 14 ACI no longer contain ISC but do have differentiated cells, the ISCs are most probably lost by differentiation without self-renewal.
TSC1/2 maintains ISCs through inhibiting TORC1/S6K signaling
It is well established that TSC inhibits target of rapamycin (TOR) in controlling cell growth and proliferation, but it is not clear whether its role in ISC maintenance is also through this pathway. In one approach to address this question, we asked whether overexpressing Rheb, a positive regulator of TORC1 downstream of TSC, could lead to ISC loss. esgGal4 is an ISC- and EB-specific driver in the posterior midgut (Micchelli and Perrimon, 2006), and by using the GAL4/UAS and Gal80ts elements (Brand and Perrimon, 1993; McGuire et al., 2004), Rheb expression could be temporally controlled in ISCs and EBs. Flies of genotypes: esgGal4, UAS-GFP/+; Gal80ts/+ (control) and esgGal4, UAS-GFP/UAS-rheb; Gal80ts/+ (experiment) were cultured at the restrictive temperature (29°C) for 2 weeks before being dissected and stained. In the control midgut, GFP-positive cells were evenly distributed in the epithelium (Fig. 2A) and were about 31.2% of total epithelial cells (Fig. 2D). In the experimental midgut, however, there were considerably fewer GFP-positive cells (Fig. 2B), accounting for 10.4% of the total epithelial population (Fig. 2D). Quantification of the ISC population marked by Dl confirmed that many ISCs are lost over time when Rheb is overexpressed (Fig. 2E). Some of the remaining GFP-positive cells in Rheb overexpressing midgut were Dl-positive and diploid cells (Fig. 2C), indicating that they were ISCs. They were also larger (Fig. 2C), which is in agreement with a role of Rheb in regulating cell growth. In another approach to address whether TSC1/2 functions through TORC1 in regulating ISC maintenance, we treated flies carrying Tsc1-mutant GFP clones with rapamycin, a TORC1-specific inhibitor. In contrast to the untreated cells, there were a significant number of GFP clones still maintained in the posterior midgut even on day 14 ACI (Fig. 2F).
Because S6K functions downstream of TORC1 in controlling protein synthesis, we tested whether TSC1/2 also functions through S6K to control ISC maintenance by analyzing the behavior of S6k and gig double mutant ISC clones (as S6k and gig are located on the same chromosome arm). As a control, S6k mutant ISCs did not show any maintenance defects, as ISC clones were properly maintained on day 14 ACI. Interestingly, S6k gig double mutant GFP clones were also properly maintained on day 14 ACI (Fig. 2G). Quantitative analysis of ISC clones indicated that S6k gig double mutant ISC clones had a similar maintenance rate to that of the wild-type ISC clones (Fig. 2H). Therefore, disruption of S6k could fully rescue the gig-mutation-induced ISC loss phenotype. We conclude that TSC1/2 inhibits the TORC1/S6K pathway to maintain ISCs.
TSC1/2 is required for ISC maintenance independently of nutritional status
TSC1/2 also serves as an important sensor to the environment to coordinate cell growth and metabolism. Therefore, it could function to sense the cellular environment, such as nutritional status to control stem cell maintenance. To test this hypothesis, the loss rate of Tsc1 mutant ISCs in flies with different diets was compared. All of the genetic experiments so far used flies fed on a rich food (regular food and wet yeast paste added to the food surface). Two additional sets of experiment used flies fed with a less rich ‘diet’ food (regular food except the yeast content was 0.1× or 0.3×) to compare with rich food ones. As a control, the wild-type ISC clones from flies treated with diet food showed similar maintenance rate to those of flies treated with rich food from day 4 to day 14 ACI. Tsc1 mutant ISC clones from flies on diet food also showed similar loss rate to that with rich food (Fig. 3A). These observations suggest that TSC1/2 is continuously required to prevent precocious ISC differentiation regardless of nutritional status. We also observed that Tsc1 mutant ISCs grow significantly large regardless of nutritional status (Fig. 3B–E). Indeed, it has been demonstrated that loss of Tsc1/2 or activation of Rheb renders cells resistant to amino acid starvation (Gao and Pan, 2001; Saucedo et al., 2003).
The role of TSC1/2 in ISC maintenance is not through Notch regulation
It has been shown that ectopic Notch activation in ISCs invariably promotes their differentiation into ECs (Micchelli and Perrimon, 2006; Ohlstein and Spradling, 2006). The observation that Tsc1/2 mutant ISCs are lost by adopting the EC fate indicates that the loss of TSC1/2 in ISC could lead to Notch activation, which subsequently promotes ISC differentiation into ECs. Su(H)m8-lacZ, a Notch activation reporter gene, is normally expressed in EBs, but not in ISCs (Micchelli and Perrimon, 2006; Ohlstein and Spradling, 2007). In Tsc1 mutant ISCs, which still maintain Dl expression, Su(H)m8-lacZ expression was not detectable (Fig. 4A). However, Su(H)m8-lacZ expression could be detected in Tsc mutant cells that lacked Dl expression (Fig. 4B), suggesting that the loss of Tsc1 mutant ISCs is accompanied by Notch activation. To functionally determine the requirement of Notch activation in the loss of Tsc mutant ISCs, we again used the MARCM system and generated NRNAi ISC clones and Tsc1 NRNAi double mutant ISC clones and examined their behavior. NRNAi clones were maintained and expanded in the epithelium over a 3-week period. Tsc1 NRNAi double-mutant ISC clones, however, were gradually lost from the epithelium, as only 11.3% ISC clones were maintained by 21 days ACI compared with 94.8% maintenance rate for NRNAi single clones (Fig. 4C–E). This result suggests that the loss of Tsc1 mutant ISCs is not caused by Notch activation (this was confirmed by cellular analysis described later), and TSC1/2–TORC1 signaling maintains ISCs independently of Notch regulation.
We then examined the potential mechanisms underlying the loss of Tsc1 NRNAi ISC clones. Because Tsc1/2 is not essential for ISC survival, it is unlikely that their loss is caused by cell death. Indeed, we did not observe any significant increase of apoptotic cells in Tsc1 NRNAi ISC clones (data not shown). At 2 weeks after clone induction, when significant numbers of Tsc1 NRNAi ISC clones were eliminated, we found that many mutant cells had detached from the epithelium and were in the lumen. Some mutant cells were delaminated as single cells (Fig. 5A), whereas some mutant cells were delaminated in clusters (Fig. 5B). These cells still expressed Dl and remained diploid, suggesting that they remain as undifferentiated cells during delamination (Fig. 5A′,B′). Cell delamination is usually accompanied by downregulation of cell adhesion molecules, so we therefore examined the expression of DE-cadherin, a major Drosophila cadherin molecule involved in cell–cell adhesion. Interestingly, DE-cadherin was significantly downregulated in Tsc1 NRNAi double mutant clones compared with NRNAi single mutant clones (Fig. 5C,D). This alteration of cell adhesion in Tsc1 NRNAi double mutant clones is also consistent with the observation that the mutant cells within the clones tend to separate from each other (Fig. 5D). These data suggest that the loss of Tsc1 NRNAi double mutant ISCs is the result of loss from the epithelium by delamination, not apoptosis.
TSC1/2–TORC1–S6K signaling regulates ee cell differentiation
Apart from the role for TSC1/2 in ISC maintenance, quantification of the two types of differentiated cells in Tsc1/2 mutant ISC clones suggests that the mutant ISCs are biased in producing either ECs or ee cells. Normally, an ISC produces approximately one ee cell for every nine ECs produced (Ohlstein and Spradling, 2007). For wild-type ISC clones on day 7 ACI, there are 100% of them contained at least one polyploid EC, and 44% (n = 74) of them contained at least one ee cell, which can be recognized by nuclear Prospero (Pros) expression (Fig. 6A; Table 1). However, for Tsc1Q87X and Tsc1R453X mutant ISC clones on day 7 ACI, only 7% (n = 60) and 0% (n = 38) of them contained ee cells, despite all of them containing ECs. Notably, on day 10 ACI none of the mutant ISC clones contained ee cell anymore (Tsc1Q87X, n = 43; Tsc1R453X, n = 31), indicating that TSC1/2 is essential for ee cell differentiation, and some ee cell formation observed on day 7 ACI in mutant ISC clones is possibly caused by the perdurance of gene products (Fig. 6B; Table 1). These observations also indicate that the mutant ISCs have differentiated into ECs but not ee cells after their last division.
The percentages of ISC clones with EC/ee were calculated by dividing the number of EC/ee-containing ISC clones examined at a given time point by the total number of ISC clones examined. In this experiment, an ISC clone was defined as a contiguous GFP clone that contains at least one ISC.
EC, enterocyte; ee, enteroendocrine cell.
Next we asked whether this altered binary fate choice in Tsc1/2 mutant clones is caused by hyperactivation of TORC1–S6K signaling. For Tsc1 mutant ISC clones, treatment with rapamycin not only rescued the ISC maintenance defect, but also rescued the ee cell differentiation defect (Fig. 6C; Table 1). Similarly, S6k gig double mutant ISCs also had no obvious defects in lineage differentiation, as ee cells and ECs were properly differentiated in the clones (Fig. 6E; Table 1). As a control, S6k single mutant ISCs did not show any maintenance and differentiation defects, as differentiated ee cells and ECs were similarly observed in these clones, which were indistinguishable to the wild-type clones (Fig. 6D; Table 1). Therefore, we conclude that TSC1/2 functions to inhibit the TORC1–S6K pathway to both maintain ISCs and allow proper ee cell differentiation.
The relationship between TSC1/2 and Notch in multiple lineage differentiation
Notch does not function downstream of TSC1/2 in regulating ISC maintenance, but it could function upstream of TSC1/2 for multiple lineage differentiation. To test this hypothesis, we examined the Tsc1 NRNAi double-mutant ISC clones that still remained in the epithelium on day 14 ACI. As a control, NRNAi clones developed into both Dl-positive ISC-like and Pros-positive ee cell-like tumors (Fig. 7A) with ∼1∶1 ratio (84 ISC-like tumors and 84 ee cell-like tumors out of 168 tumors observed), as previously reported (Micchelli and Perrimon, 2006; Ohlstein and Spradling, 2006). Notably, there were two apparent phenotypic differences between NRNAi ISC clones and Tsc1 NRNAi double mutant ISC clones. Although Tsc1 NRNAi ISC-like cells were Dl positive, they were very enlarged (Fig. 7B). This further demonstrates a role of TSC1/2 in regulating ISC growth. Importantly, these cells remained diploid (Fig. 7C), and were also negative for the EC cell marker Pdm-1 (Fig. 7D), demonstrating that TSC1/2 mutant ISCs are not able to differentiate further without Notch, suggesting that Notch is in parallel or downstream of TSC1/2 in regulating EC differentiation. Moreover, ee cell-like tumors were barely observed in Tsc1 NRNAi double mutant clones (only 4 ee-like tumor observed out of 58 tumors; Fig. 7A,B), further supporting a role of TSC1/2 in regulating ee cell differentiation independent of Notch.
TSC1/2–TORC1 signaling has important roles in cell growth control in metazoans. In this study, we have demonstrated a novel role of TSC1/2 in controlling ISC maintenance and multiple lineage differentiation in the Drosophila midgut. ISCs with disrupted Tsc1 or Tsc2 genes cannot be maintained. Our genetic analyses demonstrated that the ISC loss is caused by TORC1 hyperactivation, which is based on several lines of evidence. First, lineage tracing studies suggest that Tsc1/2 ISCs differentiate into ECs without self-renewal, leading to their elimination. Secondly, overexpression of Rheb, a TORC1 activator, also induces ISC loss. Thirdly, treatment with rapamycin, a TORC1-specific inhibitor, efficiently rescues the loss of Tsc1/2 mutant ISCs. In addition, mutation of S6k also rescues the loss of Tsc1/2 mutant ISCs. We have also demonstrated that, in agreement with known roles of TSC1/TSC2 in cell growth, Tsc1/2 mutant or Rheb overexpressed epithelial cells, including ISCs, show cell overgrowth phenotypes. Therefore, in addition to a general role in cell growth control, TSC–TORC1 signaling is also important for ISC maintenance in the Drosophila midgut.
The TSC1/2–TROC1 and INR–TOR pathways have been implicated in controlling temporal coordination of cell growth and cell differentiation to pattern complex tissues during development (Bateman and McNeill, 2004). Cell growth and cell differentiation could also be temporally controlled in adult regenerative tissues, but the mechanisms are poorly understood. Our studies of TSC1/2 and insulin signaling pathway requirements in the Drosophila midgut epithelium provide two possible mechanisms. First, activation of insulin signaling in ISCs by PI3K activation does not affect ISC maintenance and differentiation (data not shown), indicating that the role of TSC1/2 in ISC maintenance is specific to TSC1/2–TORC1 signaling. Second, S6k mutant ISC-derived clones show slightly retarded cell growth, yet their ISC maintenance and differentiation are largely normal, indicating that cell growth and cell differentiation can be uncoupled. Consistently, Tsc1 and N double mutant ISCs are very large, yet their ISC fate is still maintained.
Our data shows that although S6K signaling is not essential for ISC maintenance and lineage differentiation, TORC1 hyperactivation through S6K, caused by Tsc1 or Tsc2 disruption, is responsible for the ISC loss phenotype. It is shown that TSC2 is activated during starvation, when AMPK induces TSC2 phosphorylation and activation and thereby inhibits toRC1 activity (Inoki et al., 2003). We demonstrate that Tsc1 mutant ISCs are lost at a similar rate when flies are cultured in nutrient-rich or -poor conditions, indicating that TSC1/2 is constantly required to prevent TORC1 hyperactivation and maintain ISCs, regardless of nutritional status/cellular energy levels.
How does TORC1/S6K signaling activation following TSC1/2 depletion lead to ISC loss? Our genetic and cellular analysis suggest that Tsc1/2 mutant ISCs differentiate into ECs without self-renewal, indicating that the ISC loss could be caused by cell-autonomous Notch activation. If it is true, the loss of Tsc1/2 mutant ISCs could be prevented by inhibiting Notch activation. However, our time-course clonal analysis data shows that the Tsc1 Notch double mutant ISC clones are unable to differentiate, but are still gradually lost from the epithelium by delamination. Tsc1 Notch double mutant ISCs still maintain the expression of the ISC marker Dl and the diploid genome, showing that Notch is required for the differentiation of Tsc1 mutant ISCs. These data suggest that TSC1/2 maintains ISCs through a mechanism independent of Notch pathway regulation, although Notch activation is required for the differentiation of Tsc1 mutant ISCs. This conclusion is consistent with the observation that in Tsc1 mutant clones, Notch activation does not precede ISC loss. We propose that TSC1/2 might function specifically in ISCs to prevent cell delamination, a poorly understood process that enterocytes may use for their regular turnover. We suggest that when Tsc1/2 is depleted, ISCs are prone to delaminate. However, because they are capable of enterocyte differentiation, they eventually adopt enterocyte fate instead. Tsc1 Notch double mutant ISCs are unable to differentiate and therefore will eventually be delaminated from the epithelium. ISC delamination and lineage differentiation could be a coordinated process, which is poorly understood and worth further investigation.
This study also reveals an important role of TSC1/2 in allowing ee cell differentiation by preventing hyperactivation of TORC1, as Tsc1/2 mutant ISCs rarely produce ee cells, and this altered binary fate choice can be efficiently rescued by S6k mutation or rapamycin treatment. Because high levels of Notch activity favors EC over ee cell differentiation, depletion of TSC1/2 may lead to hyperactivation of Notch, and consequently favors EC differentiation. However, Tsc1 and N double mutant ISCs generate many ISC-like cells but fail to generate ee-like cells, whose formation normally does not require Notch. These observations suggest that failed ee cell differentiation in Tsc1/2 mutant ISC clones is independent of Notch regulation but occurs through hyperactivation of TORC1–S6K signaling. S6k mutant ISC clones produce ee and EC lineages normally, suggesting that reduced TORC1 activity does not have a significant effect on the multiple lineage differentiation, but hyperactivation of which does. During the preparation of this manuscript, Kapuria et al. showed that the expression of Tsc2 is inhibited by Notch to allow EC differentiation (Kapuria et al., 2012). They found that TSC2 maintains ISCs in a nutrition-dependent manner, and Tsc2 Notch double mutant ISCs can terminally differentiate into ECs. It cannot be excluded that differences in food ingredients or strength of mutant alleles could lead us to different conclusions. TSC1/2 also maintains germline stem cells (GSCs) in the Drosophila ovary by preventing differentiation, and we found that TSC1/2 was required for GSC maintenance in a nutrition-independent manner as well (supplementary material Table S1). Possibly, the Tsc2 Notch double mutant clones they analyzed could be transient clones and therefore had sufficient Notch activity for EC differentiation.
In this study, we demonstrate a novel role for TSC1/2–TORC1 signaling in ISC maintenance and multiple lineage differentiation. A similar role for stem cell maintenance was observed for Drosophila ovarian GSCs and mammalian hematopoietic stem cells, although the underlying mechanisms could be different (LaFever et al., 2010; Sun et al., 2010). For example, DE-cadherin expression was not altered in Tsc1/2 mutant germline stem cells (Sun et al., 2010), or in Tsc1/2 mutant imaginal disc cells (supplementary material Fig. S1). Therefore, TSC1/2 could maintain stem cells through tissue-type- and context-dependent mechanisms. Given evolutionary conservation of TSC1/2 function, we suggest that it could serve as a general role in many stem cell lineages, including that in humans, to maintain tissue homeostasis and regeneration.
Materials and Methods
Drosophila stocks and culture
The following fly stocks were used: Tsc1Q87X, Tsc1R453X, gig192, gig56 (Ito and Rubin, 1999; Tapon et al., 2001), S6kI-1 (Montagne et al., 1999), Su(H)m8-lacZ (Furriols and Bray, 2001), UAS-Notch-dsRNA (Presente et al., 2002), UAS-Rheb (Patel et al., 2003). All flies were cultured at 25°C on standard media with fresh wet yeast paste added to the food surface, unless otherwise stated. Additional information on these alleles can be found in the FlyBase (http://flybase.org/).
To generate MARCM clones, fly crosses were made to generate adults of the following genotypes:
hsflp122/+; Act-Gal4, UAS-GFP/+; Tub-Gal80 FRT82B/FRT82B;
hsflp122/+; Act-Gal4, UAS-GFP/+; Tub-Gal80 FRT82B/Tsc1Q87X FRT82B;
hsflp122/+; Act-Gal4, UAS-GFP/+; Tub-Gal80 FRT82B/Tsc1R453X FRT82B;
hsflp122,tub-gal4,UAS-GFP/+; tub-gal80 FRT80B/FRT80B;
hsflp122,tub-gal4,UAS-GFP/+; tub-gal80 FRT80B/gig192 FRT80B;
hsflp122,tub-gal4,UAS-GFP/+; tub-gal80 FRT80B/S6kI-1 FRT80B;
hsflp122,tub-gal4,UAS-GFP/+; tub-gal80 FRT80B/S6kI-1gig56 FRT80B;
hsflp122/UAS-Notch-dsRNA; Act-Gal4, UAS-GFP/+; Tub-Gal80 FRT82B/FRT82B; hsflp122/UAS-Notch-dsRNA; Act-Gal4, UAS-GFP/+; Tub-Gal80 FRT82B/Tsc1Q87X FRT82B;
hsflp122/+; Act-Gal4, UAS-GFP/Su(H)m8-lacZ; Tub-Gal80 FRT82B/Tsc1Q87X FRT82B.
Owing to differences in clone induction rate with different FRT sites, the FRT82B and FRT80B MARCAM clones were induced by 1 and 2 hour heat shock treatments, respectively, in order to obtain a comparable number of induced clones. For time course clonal analysis, flies with appropriate genotype was grouped randomly with 20–25 flies maintained in each vial after heat shock. Flies were transferred every 2 days to fresh food until dissection. For overexpression experiments using esg-Gal4 and Tub-gal80ts, fly crosses were done at 18°C, and newly enclosed flies of appropriate genotypes were shifted to 29°C for 2 weeks before dissection.
For Tsc1 mutant rescue experiments, 20 µl of 10 mM rapamycin (dissolved in ethanol) was directly added to the food surface for each vial. The liquid on the food was air-dried before the food was used. The flies were transferred every second day with fresh food and rapamycin.
The immunostaining procedure was as previous studies (Lin et al., 2008). The following antisera or dyes were used: mouse anti-Dl antibody [Developmental Studies Hybridoma Bank (DSHB); 1∶100]; mouse anti-Pros antibody (DSHB, 1∶300); rat anti-DE-cadherin antibody (DSHB, 1∶10); rabbit anti-Pdm-1 (a gift from Xiaohang Yang, Institute of Genetics, Zhejiang University, China; 1∶1000); rabbit anti-β-gal antibody (Cappel, 1∶6000); Alexa-Fluor-568-conjugated goat anti-mouse/rabbit and Alexa-Fluor-488-conjugated goat anti-rabbit/mouse secondary antibodies (Molecular Probes, 1∶300); DAPI (49,69-diamidino-2-phenylindole, Sigma; 0.1 mg/ml, 5 minute incubation). Images were captured using a Zeiss Meta 510 confocal microscope. All images were processed in Adobe Photoshop and Illustrator.
To study the effects of nutrition on ISC maintenance, the fly food was made with different amounts of yeast, as previously described (Layalle et al., 2008): For the 1× recipe: 10 g agar, 83 g corn flour, 60 g white sugar, 17 g inactivated yeast extract in 1 liter H2O. After the food was pasteurized at 100°C for 8 minutes, 2.5 g potassium sorbate was added. For 0.3× or 0.1× yeast food, 5.1 g or 1.7 g inactive yeast extract were added instead.
Apoptosis in the midgut was detected using an in situ cell death detection kit (Roche), as previously described (Lin et al., 2008).
We thank S. Bray, I. Hariharan, S. Hayashi, X. Yang, G. Thomas, T. Xu, the Bloomington Stock Center and Developmental Studies Hybridoma Bank for fly stocks and antibodies, H. Jasper and members of the Xi laboratory for comments and discussions.
Z.Q., P.S. and G.L. performed the experiments; Z.Q. and R.X. analyzed the data; R.X. wrote the paper.
This work was supported by the Chinese Ministry of Science and Technology through programs 973 [grant number 2011CB12700 to R.X.] and 863 [grant number 2007AA02Z1A2 to R.X.].