1. Cell Biology
  2. Developmental Biology and Stem Cells
Download icon

Oxidative stress induces stem cell proliferation via TRPA1/RyR-mediated Ca2+ signaling in the Drosophila midgut

  1. Chiwei Xu
  2. Junjie Luo
  3. Li He
  4. Craig Montell
  5. Norbert Perrimon  Is a corresponding author
  1. Harvard Medical School, United States
  2. University of California Santa Barbara, United States
  3. University of California, Santa Barbara, United States
  4. Howard Hughes Medical Institute, Harvard Medical School, United States
Research Article
  • Cited 5
  • Views 1,423
  • Annotations
Cite as: eLife 2017;6:e22441 doi: 10.7554/eLife.22441

Abstract

Precise regulation of stem cell activity is crucial for tissue homeostasis and necessary to prevent overproliferation. In the Drosophila adult gut, high levels of reactive oxygen species (ROS) has been detected with different types of tissue damage, and oxidative stress has been shown to be both necessary and sufficient to trigger intestinal stem cell (ISC) proliferation. However, the connection between oxidative stress and mitogenic signals remains obscure. In a screen for genes required for ISC proliferation in response to oxidative stress, we identified two regulators of cytosolic Ca2+ levels, transient receptor potential A1 (TRPA1) and ryanodine receptor (RyR). Characterization of TRPA1 and RyR demonstrates that Ca2+ signaling is required for oxidative stress-induced activation of the Ras/MAPK pathway, which in turns drives ISC proliferation. Our findings provide a link between redox regulation and Ca2+ signaling and reveal a novel mechanism by which ISCs detect stress signals.

https://doi.org/10.7554/eLife.22441.001

Introduction

Multipotent intestinal stem cells (ISCs) are responsible for tissue homeostasis in the adult Drosophila midgut (Jiang and Edgar, 2011; Micchelli and Perrimon, 2006; Ohlstein and Spradling, 2006). Similar to the mammalian digestive epithelium, the activity of ISC proliferation and differentiation increases upon tissue damage (Amcheslavsky et al., 2009). Although signaling pathways controlling ISC activity, such as Ras/MAPK (Jiang et al., 2011), Hippo/Yki (Karpowicz et al., 2010; Shaw et al., 2010), and JAK/Stat (Jiang et al., 2009), play critical roles in gut homeostasis, the mechanisms modulating these pathways under different physiological and pathological conditions are not well understood. In particular, it is unclear how ISCs sense microenvironment signals under tissue damage conditions to upregulate the activity of mitogenic pathways such as Ras/MAPK. The simplicity of the Drosophila midgut and the ease of performing genetic screens provide a powerful system to study how stem cells sense microenvironment cues and adjust their activities accordingly.

The concentration of reactive oxygen species (ROS) is emerging as a critical microenvironment signal that regulates stem cell activity in neuronal and glial progenitors (Liu et al., 2009; Smith et al., 2000; Tsatmali et al., 2005), as well as in mammalian hematopoietic stem cells (HSCs) (Tothova et al., 2007) and Drosophila hematopoietic progenitors (Owusu-Ansah and Banerjee, 2009). In most cases, high ROS concentration stimulates, at least in the short term, the activity of stem cells by driving them out of an otherwise quiescent status. In mammals, oxidative stress is associated with intestine damage and inflammatory bowel disease (Rezaie et al., 2007). In the Drosophila midgut, high ROS levels are produced in response to different types of tissue damage (Ha et al., 2005; Hochmuth et al., 2011; Lee et al., 2013). Moreover, oxidants such as paraquat and hydrogen peroxide are commonly used to induce tissue damage (Chatterjee and Ip, 2009). Although high ROS levels in ISCs are sufficient to trigger ISC proliferation (Hochmuth et al., 2011), the connection between oxidative stress and mitogenic signals in ISCs remains obscure.

We performed an in vivo RNAi screen to identify novel components required for ISC proliferation in response to tissue damage. In the screen, we identified two cation channels, transient receptor potential A1 (TRPA1) and ryanodine receptor (RyR), which are required for ISC self-renewal and damage-induced proliferation. We show that these channels are required for Ca2+ homeostasis in ISCs and mediate increases in cytosolic Ca2+ in response to oxidative stress and ISC proliferation. Our results are consistent with a recent study (Deng et al., 2015) showing that high cytosolic Ca2+ in ISCs can stimulate proliferation. Finally, we characterize the signaling events downstream of cytosolic Ca2+ increase and demonstrate that Ras/MAPK mediates the effects of Ca2+ on ISC proliferation.

Results

TRPA1 and RyR are required for stem cell damage response and self-renewal

From an RNAi screen for new transmembrane proteins required for ISCs proliferation, we identified TRPA1 and RyR as candidate hits required for ISC pool size maintenance and damage-induced ISCs proliferation (see Materials and methods). We confirmed that the lines producing a consistent phenotype (v37249, BL31504 for TRPA1; BL29445 for RyR) effectively knockdown the corresponding transcripts, while the lines that did not score as hits from our screen (BL31384, BL36780 for TRPA1; BL31540, BL31695 for RyR) cannot achieve efficient knockdown (Figure 1—figure supplement 1C–D).

TRPA1 is a Ca2+-permeable cation influx channel (Wang et al., 2013). RyR is an endoplasmic reticulum (ER) cation channel that releases ER Ca2+ into the cytosol (Santulli and Marks, 2015). RNAi-mediated knockdown of either trpA1 or RyR in adult ISCs completely blocked proliferation induced by the tissue-damaging agent paraquat or bleomycin (Figure 1A–C and A’-C’, G). Depletion of trpA1 or RyR causes a gradual loss of the stem cell population, with a significant decrease in esg+ cell density undetectable until 7 days of RNAi expression (Figure 1—figure supplement 1E). The RNAi lines do not cause stem cell apoptosis, because cell death cannot be detected with anti-cleaved-caspase 3 staining in the midgut (Figure 1—figure supplement 2A–C), and the anti-apoptotic gene p35 (Hay et al., 1994) cannot rescue the mitosis defect caused by trpA1 RNAi or RyR RNAi (Figure 1—figure supplement 2D–I,D’–I’). In contrast, mitosis activity in stem cells expressing trpA1 RNAi or RyR RNAi can be regained by forced cytosolic calcium influx with RNAi against SERCA (Figure 1D–G), an ER Ca2+ ATPase whose inhibition triggers Stim/Orai channel (SOC)-mediated Ca2+ entry into the cytosol (Clapham, 2007). Altogether, these results place cytosolic calcium signaling genetically downstream of TRPA1 and RyR during ISC proliferation. Consistent with the results obtained with RNAi, CRISPR/Cas9-mediated knockout of trpA1 or RyR in adult ISCs suppressed damage-induced proliferation (Figure 1H). In addition, at the organ level, inhibition of ISC proliferation by trpA1 RNAi is associated with a shortening in midgut length (Figure 1I). Further, while wild-type midguts maintain their proper length after feeding paraquat or bleomycin for 2 days, midguts with trpA1 RNAi expression in the ISCs exhibit accelerated shortening following tissue damage. This damage-induced midgut shortening phenotype is reminiscent to the gut phenotype observed after depletion of the esg+ cell population when the cell death gene reaper (rpr) is expressed in ISCs for 4 days (Figure 1—figure supplement 3), suggesting the importance of ISC activity for maintaining tissue size.

Figure 1 with 3 supplements see all
The calcium channels TRPA1 and RyR are required for damage-induced ISC proliferation.

(A–C) Midguts overexpressing Luc RNAi, trpA1 RNAi or RyR RNAi in ISCs using EGT (esgGal4 UAS-GFP tubGal80ts) for 6d, were fed the oxidant agent paraquat for the last 2d (A’–C’) or co-expressed with SERCA RNAi (D–F). Midguts are stained for the mitosis marker phosphohistone H3 (pH3). Luc RNAi is used as control. ts stands for temperature sensitive tubGal80ts that allows for temporal control of genetic manipulation. GFP is used to report the expression pattern of esgGal4. (G) Quantification of mitosis (pH3+ cell number) of midguts expressing Luc RNAi, trpA1 RNAi or RyR RNAi in ISCs under normal condition, 2 mM paraquat feeding, 25 µg/ml bleomycin feeding, or co-expression with SERCA RNAi. N > 5 midguts are quantified per genotype per treatment. Data are represented as mean ± SEM. Double asterisks indicate a p value of less than 0.01. (H) Mitosis quantification of midguts with constitutive expression of sgRNA targeting trpA1 or RyR, and targeted Cas9 expression in ISCs for 7d (bleomycin feeding for the last 2d). Flies with the same genetic background but only empty insertional landing site are used as the control for sgRNA. Data are represented as mean ± SEM. Note that we generally observed that CRISPR/Cas9-mediated knockout in vivo is not working as efficiently as RNAi, probably because not all DNA damages by CRISPR/Cas9 result in frame-shift mutations. N > 6 midguts are analyzed per genotype. Data are represented as mean ± SEM. (I) Length quantification of midguts expressing Luc RNAi (control) or trpA1 RNAi for 7d, with the last 2d feeding on normal food, paraquat, or bleomycin. The average midgut length of control flies feeding on normal food is used for normalization. N > 5 midguts are analyzed per genotype per treatment. Data are represented as mean ± SEM.

https://doi.org/10.7554/eLife.22441.002

To determine whether TRPA1 and RyR act in a cell-autonomous manner, we generated GFP+ clones originating from ISCs that are deficient for either trpA1 or RyR using the MARCM method (Wu and Luo, 2006). MARCM clones generated from ISCs expressing trpA1 RNAi (Figure 2A’), or homozygous for the loss-of-function allele trpA11 (Kwon et al., 2008) (Figure 2B’), or RyRk04913 (Sullivan et al., 2000) (Figure 2C’) had a smaller average size compared to wild-type clones 10 days after clonal induction (Figure 2A–C, trpA1 RNAi and RyRk04913 clone sizes are quantified in Figure 2D), suggesting that TRPA1 and RyR are required autonomously for clonal expansion. Although smaller than wild-type clones, MARCM clones expressing trpA1 RNAi or losing RyR can survive and differentiate into Pdm1+ enterocytes (ECs) (Figure 2—figure supplement 1A–D). We also performed lineage tracing with the EGT F/O system (Jiang and Edgar, 2009), which relies on induced expression of Flp in the stem cells to excise transcriptional STOP cassettes flanking act::Gal4, and found that stem cells expressing trpA1 RNAi can still differentiate into mature Pros+ enteroendocrine cells (EEs) or Pdm1+ ECs (Figure 2—figure supplement 1E–H). Furthermore, while Dl+ cells (Figure 2D–F and L) are dramatically decreased by trpA1 RNAi or RyR RNAi expression for 5 days, there was no significant change in the density of Su(H)Gbe+ or Pros+ cells (Figure 2G–I and M). The Notch pathway reporter Su(H)GbeGFP is indicative of ISC differentiation activity because Notch signaling is high when ISCs differentiate, especially for the EC lineage (Micchelli and Perrimon, 2006; Ohlstein and Spradling, 2007). Altogether, our data suggest that trpA1 RNAi or RyR RNAi mainly affects ISC self-renewal and proliferation in the adult gut, rather than ISC differentiation or cell death.

Figure 2 with 1 supplement see all
TRPA1 and RyR are required for ISC self-renewal but not differentiation.

(A–C, A’–C’) MARCM clones deficient for TRPA1 or RyR are analyzed along with their corresponding control genotypes (A for A’, B for B’, C for C’) 10d after clone induction. Cell numbers per clone for MARCM clones expressing trpA1 RNAi, or lacking both wild-type RyR alleles, are quantified in (J). Randomly picked N = 35 clones from the posterior regions of five guts per genotype are analyzed. Data are represented as mean ± SEM. (D–F) ISC marker Dl-lacZ stainings of midguts expressing Luc RNAi, trpA1 RNAi or RyR RNAi for 5d in ISCs. The relative Dl+ cell density in the posterior midgut region is quantified in (K). Dl+ cell number is divided by the imaged area size for density calculation. N > 5 midguts per genotype are analyzed. Data are represented as mean ± SEM. (G–I) Immunostaining of midguts expressing Luc RNAi, trpA1 RNAi or RyR RNAi for 5d in ISCs for Notch activity reporter Su(H)GbeGFP and EE marker Prospero (pros). The relative Su(H)Gbe+ cell density in the posterior midgut region is quantified in (L). Su(H)Gbe+ cell number is divided by the imaged area size for density calculation. N > 4 midguts per genotype are analyzed. Data are represented as mean ± SEM. (M) Quantificatioin of EE density in the posterior midgut region for midguts expressing Luc RNAi, trpA1 RNAi or RyR RNAi in ISCs for 3d, 5d, and 7d. For each time point, EE density is normalized to the average value of midguts expressing Luc RNAi. N > 4 midguts per genotype are analyzed. Data are represented as mean ± SEM.

https://doi.org/10.7554/eLife.22441.007

TRPA1 and RyR expression in the midgut

The trpA1 gene is expressed as multiple mRNAs (Figure 1—figure supplement 1A), four of which have been characterized previously (Zhong et al., 2012). Compared to the TRPA1-A and -B isoforms, expressed in the fly head, TRPA1-C and -D isoforms are expressed in the midgut (Figure 1—figure supplement 1B) and have different N-terminal residues that might confer sensitivity to different stimuli. Based on anti-TRPA1 staining, TRPA1 appears to be expressed in stem cells (Figure 3A–B). Using a Gal4 fused to the C terminal of trpA1 transcripts via CRISPR/Cas9-mediated knock-in at the endogenous locus (Figure 3—figure supplement 1), we could also detect trpA1 expression in stem cells in the anterior region of the posterior midgut (Figure 3C). Consistently, stem cell depletion reduces trpA1 total mRNA expression in the midgut by more than 50% (Figure 3—figure supplement 2A). Because the signal-to-noise ratio for anti-TRPA1 staining is low, and trpA1Gal4CP2A might not fully recapitulate the endogenous expression pattern, we cannot exclude the possibility that trpA1 is also expressed in other cell types. In a recent study, for example, TRPA1-C has been identified to regulate defecation of food-borne pathogens in the EEs (Du et al., 2016). However, TRPA1 is mainly required in ISCs for proliferation, because knockdown of trpA1 in self-renewing ISCs, rather than in differentiating ISCs (Su(H)Gbe+ cells, also called enteroblasts (EBs)) or visceral muscles, blocks damage-induced proliferation (Figure 3—figure supplement 2B). Further, an additional piece of evidence that TRPA1 protein functions in the stem cells comes from the ex vivo calcium-imaging experiments with GCamP6f reporter, for which the intensity of green fluorescence indicates cytosolic calcium levels (Chen et al., 2013). While the TRPA1 agonist AITC (allyl isothiocyanate (Bautista et al., 2005) can induce calcium influx in ISCs carrying a wild-type allele of trpA1, such response is abolished in flies losing both alleles of wild-type trpA1 (Figure 3E–F).

Figure 3 with 2 supplements see all
trpA1 and RyR expression in the midgut.

(A–B) Anti-TRPA1 immunostaining of midguts expressing Luc RNAi or trpA1 RNAi in ISCs. The channel of anti-TRPA1 signal is shown below the merged image. Note that anti-TRPA1 staining has high background signal. While we could detect TRPA1 expression in the ISCs, we could not exclude the possibility that it is also expressed in other cell types. (C) trpA1Gal4CP2A-driven expression of mCherry is co-stained with ISC marker esglacZ. (D) RyRGal4R14G09 (enhancer region of RyR locus) driven expression of membrane-localized CAAX-mCherry is co-stained with ISC marker esgGFP. (E–F) Imaging midguts missing one or both alleles of wild-type trpA1 with GCaMP6f expression in ISCs. The same guts are imaged again after exposure to 0.015% AITC in the imaging buffer for 5 min (E’–F’). The images are acquired on a Keyence microscope.

https://doi.org/10.7554/eLife.22441.010

RyR appears to be ubiquitously expressed at low levels in the midgut epithelium, based on the expression pattern of RyRGal4R14G09 (Figure 3D), where Gal4 is under the control of 2.92 kb putative enhancer region sequence of RyR (Pfeiffer et al., 2008).

TRPA1-D in the midgut senses the oxidative stress associated with tissue damage

Oxidative stress has previously been observed in tissue damage conditions involving JNK activation (Hochmuth et al., 2011) and pathogen infection (Ha et al., 2005; Lee et al., 2013). The source of oxidative stress is generally attributed to an imbalance in the formation of ROS and impaired antioxidant machinery in unhealthy tissues (Hochmuth et al., 2011; Lih-Brody et al., 1996). Staining with dihydro-ethidium (DHE) (Owusu-Ansah et al., 2008) confirmed that ROS levels in the midgut increase following tissue damage induced by feeding flies with paraquat (Figure 4—figure supplement 1B), a widely used oxidant that interferes with electron transfer to catalyze formation of superoxide-free radicals (Bus and Gibson, 1984), or bleomycin (Figure 4—figure supplement 1C), a chemotherapy drug that cleaves DNA via an unresolved mechanism potentially related to ROS production (Hecht, 2000). In addition, induction of midgut stress by JNK pathway activation or massive EC apoptosis by rpr expression can also lead to increased ROS and ISC expansion (Figure 4—figure supplement 1F–J).

Both TRPA1 and RyR are associated with redox sensing (Guntur et al., 2015; Hamilton and Reid, 2000; Sun et al., 2011). Guntur et al. found that TRPA1-C and TRPA1-D isoforms are H2O2-sensitive and that their endogenous expressions in the ring gland, or ectopic expression in other tissues, confer UV light sensitivity via sensing UV light-induced H2O2 production. To directly measure whether gut-specific TRPA1 isoforms can respond to paraquat, we used two-electrode voltage clamp recordings to measure the channel activities of TRPA1-C and -D (Wagner et al., 2000). In this experiment, we expressed similar amount of mRNA for TRPA1-C or -D isoform in the Xenopus oocytes, clamped the membrane potential at −80 mV, and measured the current flowing through the ion channels in the oocyte membrane. The current value is proportional to the opening probability of the ion channels. We added paraquat and AITC (used as the positive control) into the bath solution of oocytes sequentially to trigger the opening of the TRPA1-C or -D channel and compare their strength and dynamics. TRPA1-C barely displayed any response to paraquat (Figure 4A), whereas expression of TRPA1-D induced a slow increase in current (Figure 4B). Consistently, overexpression of TRPA1-D, but not TRPA1-C, stimulated ISCs expressing trpA1 RNAi to proliferate in response to paraquat (Figure 3C–F).

Figure 4 with 1 supplement see all
TRPA1 can respond to oxidant agent and its activation stimulates ISC activity.

(A–B) Oocyte clamp measurement of TRPA1-C or TRPA1-D channel activities in response to paraquat. The same amounts of mRNAs for TRPA1-C or TRPA1-D were injected into Xenopus oocytes. 4 mM paraquat is present in the buffer during the period marked with the gray line on the top. Near the end of the experiment, the TRPA1 agonist AITC was added as a positive control (indicated with arrowhead). 2 mM and 10 mM paraquat were also tested and gave consistent results. (C–E) Midguts, expressing trpA1 RNAi alone, or together with TRPA1-C or TRPA1-D in the ISCs for 7d, with the last 2d feeding paraquat (C’–E’), are stained for the mitosis marker pH3. Nucleus-localized GFP (nlsGFP) is used as the control for transgenic expression. We have also tested UAS-ultraGFP (Yang and Tower, 2009) that consists of multiple copies of UAS-2xEGFP as control and obtained the same results as UAS-nlsGFP with regard to stem cell activity (except that ultraGFP reporter is much brighter than nlsGFP). (F) Mitosis quantification of midguts expressing trpA1 RNAi together with TRPA1-C or TRPA1-D under normal and paraquat-feeding conditions. N > 4 midguts per genotype per treatment are analyzed for quantification. UAS-ultraGFP is used as the control for transgenic expression. Data are represented as mean ± SEM. Note that TRPA1-C or TRPA1-D are still targeted by trpA1 RNAi, which might limit their capacity to rescue trpA1 RNAi. (G) Mitosis quantification of midguts expressing CncC RNAi alone, or together with trpA1 RNAi in the ISCs for 7d. N > 7 midguts are analyzed for each genotype. Data are represented as mean ± SEM. (J) Mitosis quantification of midguts overexpressing TRPA1-C, TRPA1-D, or TRPA1-A in ISCs for 7d. N > 5 midguts per genotype are analyzed. Data are represented as mean ± SEM.

https://doi.org/10.7554/eLife.22441.014

Previously, it has been reported that knockdown of antioxidant protein CncC in ISCs can induce mitosis (Hochmuth et al., 2011). Our observation that trpA1 RNAi suppresses the hyper-proliferation status induced by CncC RNAi (Figure 4G) places TRPA1 genetically downstream of the pathway controlling ROS for ISC proliferation. Further, overexpression of TRPA1-D or conventional TRPA1-A isoform (active at 29°C) in stem cells can induce mitosis significantly, suggesting that increased TRPA1-calcium signaling is sufficient to drive ISC proliferation (Figure 4H).

TRPA1 and RyR are required for normal and ROS-induced cytosolic Ca2+ levels in ISCs

Interestingly, TRPA1 and RyR are not critically required in the midgut epithelium for development, because flies expressing trpA1 RNAi or RyR RNAi with esgGal4 throughout development [that is, in adult midgut progenitors (AMPs) during larval stages and in adults ISCs, see (Jiang and Edgar, 2009) can survive to adulthood without gross defects when they are fed with normal food]. For young adult flies of these genotypes, we used the ultrasensitive GCaMP6s fluorescent reporter (Chen et al., 2013) to examine cytosolic Ca2+ concentrations in their ISCs. While the number of GFP+ (i.e. high cytosolic Ca2+ concentration) ISCs in wild-type midguts increases significantly 10 min after adding paraquat to a final concentration of 4 mM (Figure 5A, A’ and D), expression of trpA1 RNAi or RyR RNAi in ISCs (driven by esgGal4 throughout development) not only reduced the basal levels of cytosolic Ca2+ in stem cells (Figure 5A–C) but also inhibited the ROS-mediated increase in cytosolic Ca2+ (Figure 5A’–C’). The observed reduction of high Ca2+ ISCs is not due to decrease in the total number of ISCs, as total ISC number or density in young adult flies is not yet significantly affected by trpA1 RNAi or RyR RNAi (Figure 5—figure supplement 1).

Figure 5 with 2 supplements see all
TRPA1 and RyR are required for ROS-mediated Ca2+ increases in ISCs.

(A–C) Imaging of midguts expressing the GCaMP6s reporter alone or together with trpA1 RNAi or RyR RNAi in ISCs. The same guts are imaged again after exposure to 4 mM paraquat in the imaging buffer for 10 min (A’–C’). The images are acquired on a Keyence microscope. (D) Quantification of high Ca2+ (GFP+) stem cells within the posterior midgut region. N > 5 midguts are analyzed for each genotype. Data are represented as mean ± SEM. (E–G) Numeric kinetics of high Ca2+ stem cells in midguts expressing trpA1 RNAi or RyR RNAi in ISCs. Wild-type midguts expressing GCaMP6s alone serve as control. Time-lapse confocal imaging is used for the analysis (see Supplementary Movies). The average number of high Ca2+ stem cells before drug treatment is set as the basal level of 0 for individual imaging experiments. Final concentration of 4 mM paraquat (E), 0.03% AITC (F), 4 µM thapsigargin (G) are added at 150 s. At least three replicates of each genotypes/treatments are shown in the figure, with different colors labeling different genotypes.

https://doi.org/10.7554/eLife.22441.017

Defining the average number of high Ca2+ stem cells before drug treatment as a baseline, we used the confocal time-lapse GCaMP6s imaging (Videos 16) and analysis to identify the kinetics of high Ca2+ in stem cells. Consistent with the TRPA1 channel kinetics as measured in the voltage clamp experiments in oocytes (Figure 4A–B), treatment with 4 mM paraquat caused a slow and continuous increase in the average number of high Ca2+ ISCs (Figure 5E), whereas the TRPA1 agonist AITC triggered a sharp increase followed by a quick drop in the number of high Ca2+ ISCs (Figure 5F). Finally, responses to both paraquat and AITC were dependent on TRPA1 and RyR (Figure 5E–F), whereas treatment with the SERCA inhibitor thapsigargin, as a positive control, led to increased Ca2+ in ISCs expressing trpA1 RNAi or RyR RNAi (Figure 5G).

Video 1
Calcium imaging of ISCs in response to paraquat.

Midguts expressing GCaMP6s reporter in ISCs are dissected and imaged in adult hemolymph-like (AHL) buffer. Z-stack images were acquired with 10 s interval. A maximal intensity z-projection was shown in the movie. A final concentration of 4 mM oxidant agent paraquat was added at 150 s (the 15th frame of the movie).

https://doi.org/10.7554/eLife.22441.021
Video 2
Calcium imaging of ISCs expressing trpA1 RNAi in response to paraquat.

Midguts expressing GCaMP6s reporter together with trpA1 RNAi in ISCs were dissected and imaged in AHL buffer. Z-stack images were acquired with 10 s interval. A maximal intensity z-projection was shown in the movie. A final concentration of 4 mM oxidant agent paraquat was added at 150 s (the 15th frame of the movie).

https://doi.org/10.7554/eLife.22441.022
Video 3
Calcium imaging of ISCs expressing RyR RNAi in response to paraquat.

Midguts expressing GCaMP6s reporter together with RyR RNAi in ISCs were dissected and imaged in AHL buffer. Z-stack images were acquired with 10 s interval. A maximal intensity z-projection was shown in the movie. A final concentration of 4 mM oxidant agent paraquat was added at 150 s (the 15th frame of the movie).

https://doi.org/10.7554/eLife.22441.023
Video 4
Calcium imaging of ISCs in response to AITC.

Midguts expressing GCaMP6s reporter in ISCs are dissected and imaged in AHL buffer. Z-stack images were acquired with 10 s interval. A maximal intensity z-projection was shown in the movie. A final concentration of 0.03% TRPA1 agonist AITC was added at 150 s (the 15th frame of the movie).

https://doi.org/10.7554/eLife.22441.024
Video 5
Calcium imaging of ISCs expressing trpA1 RNAi in response to AITC.

Midguts expressing GCaMP6s reporter together with trpA1 RNAi in ISCs were dissected and imaged in AHL buffer. Z-stack images were acquired with 10 s interval. A maximal intensity z-projection was shown in the movie. A final concentration of 0.03% TRPA1 agonist AITC was added at 150 s (the 15th frame of the movie).

https://doi.org/10.7554/eLife.22441.025
Video 6
Calcium imaging of ISCs expressing RyR RNAi in response to AITC.

Midguts expressing GCaMP6s reporter together with RyR RNAi in ISCs were dissected and imaged in AHL buffer. Z-stack images were acquired with 10 s interval. A maximal intensity z-projection was shown in the movie. A final concentration of 0.03% TRPA1 agonist AITC was added at 150 s (the 15th frame of the movie).

https://doi.org/10.7554/eLife.22441.026

Using GCaMP5G as the Ca2+ reporter and tdTomato as an internal control, we found that trpA1 RNAi blocked the increase of cytosolic Ca2+ following treatment with 2 mM paraquat (Figure 5—figure supplement 2A,A’, B,B’). Similar results were observed with another reporter GCaMP6f (data not shown).

In addition, we used TRIC (transcriptional reporter of intracellular Ca2+) (Gao et al., 2015) to examine cytosolic Ca2+ concentration in ISCs in vivo on fixed midguts. With TRIC lexGFP and a conventional UAS-RFP reporter, we found that the percentage of high Ca2+ ISCs among all stem cells increased from 23% to 63% (Figure 5—figure supplement 2C,C’, F) when flies were fed 2 mM paraquat for 1 day. In adult midguts expressing trpA1 RNAi or RyR RNAi in adult ISCs for 6 days, high Ca2+ ISCs showed little, if any, increase following paraquat feeding (Figure 5—figure supplement 2D,D’, E,E’, F). Unlike GCamP reporters, a reduction in basal levels of high Ca2+ ISCs by trpA1 RNAi or RyR RNAi was not observed with TRIC reporter, which might be due to differences in the stability and detection threshold of reporters as well as the time frame of knockdown for the two sets of experiments.

Cytosolic Ca2+ induces Ras/MAPK activity

The biological functions of TRPA1, RyR, and cytosolic Ca2+ signaling have been extensively studied in neurons, where it has been established that membrane depolarization and Ca2+ influx can activate Ras/MAPK (Randlett et al., 2015a; Rosen et al., 1994). Thus, we next asked whether high Ca2+ in ISCs similarly activates Ras/MAPK. To do this, we stained midguts with antibodies against diphospho-extracellular signal-regulated kinase (dpErk), which is a measure of Drosophila Ras/MAPK activity (Gabay et al., 1997). Adult midguts expressing SERCA RNAi in ISCs for 2 days, which at that stage do not present signs of ISCs expansion, showed induced MAPK activity (Figure 6A–B), and expression of trpA1 RNAi (Figure 6C) or RyR RNAi (Figure 6D) decreased the levels of MAPK activity. Consistent with our observation that ISC Ca2+ induction by SERCA inhibition does not depend on trpA1 or RyR (Figure 5G), SERCA RNAi could override trpA1 RNAi (Figure 6G) or RyR RNAi (Figure 6H) to activate MAPK. To test whether the increased dpErk signal we observed with SERCA RNAi merely reflects an hyper-proliferative status for ISCs, we combined SERCA RNAi with knockdown of Yorkie (Yki), the Hippo pathway transcriptional factor required for damage-induced ISC proliferation (Karpowicz et al., 2010; Ren et al., 2010; Shaw et al., 2010; Staley and Irvine, 2010). Although Yki RNAi could inhibit ISC proliferation caused by SERCA RNAi (Figure 7—figure supplement 1F–G), it could not block MAPK activity (Figure 6F). On the other hand, Ras1 (Drosophila Ras) RNAi could block SERCA RNAi-induced MAPK activity (Figure 6E). Altogether, our data suggest that Ras/MAPK activity is likely the direct target of high cytosolic Ca2+ in ISCs, rather than the consequence of cross-activation by other proliferation-related pathways such as Yki.

Figure 6 with 2 supplements see all
High cytosolic Ca2+ is necessary and sufficient to activate Ras/MAPK in ISCs.

(A–D) Midguts expressing Luc RNAi, SERCA RNAi, trpA1 RNAi, or RyR RNAi in ISCs for 2-3d are stained for the Ras/MAPK activity marker dpErk. The channel of dpErk signal is shown below the merged image. Note trpA1 RNAi group was imaged on a different date, but alongside Luc RNAi which exhibited the same level of pErk signal as the control group presented in A. (E–H) Midguts expressing SERCA RNAi together with Ras1 RNAi, Yki RNAi, trpA1 RNAi, or RyR RNAi in ISCs for 3d are stained for dpErk. The channel of dpErk signal is shown below the merged image. Note trpA1 RNAi group in G was imaged on a different date together with C. (I–J) Midguts expressing the TRIC lexGFP reporter of intracellular calcium concentration and Luc RNAi or trpA1 RNAi Luc in ISCs for 5d, with the last 2d feeding paraquat (I’–J’), are stained for dpErk. The channel of dpErk signal is shown to the right of the merged image. The TRIC lexGFP reporter consists of lexop-GFP, and two split modules of lexA whose assembly is dependent on intracellular calcium concentration.

https://doi.org/10.7554/eLife.22441.027

To further examine the relationship between Ca2+ levels and MAPK activity, we performed dpErk staining of midguts expressing the TRIC reporter in ISCs. Paraquat feeding for 1 day led to an increase in the number of cells with high dpErk signals, a majority of which are high Ca2+ ISCs (Figure 6I,I’). Moreover, paraquat induction of MAPK activity appears to require high Ca2+ in ISCs, because such response could be inhibited by trpA1 RNAi expression in ISCs (Figure 6J,J’). It should be noted that the best timing to detect autonomous activation of dpErk by high Ca2+ in the ISCs is before they enter the hyper-proliferative stage. Once ISCs start to expand massively following prolonged high Ca2+ induction, pErk returns to normal levels in many ISCs and is non-autonomously induced in some ECs, resulting in a diffusive activation pattern (Figure 6—figure supplement 1A–C).

In neurons, a common mechanism by which Ca2+ regulates Ras/MAPK is through Ca2+-sensitive Src kinases (Cullen and Lockyer, 2002), in a process faster than any transcriptional response (Randlett et al., 2015b; Rosen et al., 1994). Interestingly, overexpression of Src (Src42A or Src64B in Drosophila) in ISCs triggers ectopic MAPK activation and proliferation (Figure 6—figure supplement 2A–C), even in the presence of trpA1 RNAi (Figure 6—figure supplement 2E–H). Moreover, we could observe significant inhibition of SERCA RNAi-induced ISC proliferation by Src64B RNAi (Figure 6—figure supplement 2I) despite the likelihood of functional redundancy between Src64B and Src42A (Ma et al., 2013; Tateno et al., 2000). These data suggest that Src might mediate the activation of Ras/MAPK by cytosolic Ca2+ in ISCs.

Ras/MAPK activity is the major effector of cytosolic calcium for ISC proliferation

As a universal intracellular signal, increased Ca2+ might trigger a wide range of responses. Consistent with previous results (Deng et al., 2015; Jiang et al., 2011), ISCs expressing the active forms of many Ca2+ effector proteins, such as Ras1, Raf, phosphatase Calcineurin A1 (CanA1), CREB-regulated transcriptional co-activator (CRTC), and cyclic-AMP response element binding protein B (CrebB), displayed an hyper-proliferation phenotype (Figure 7A–E). To determine which Ca2+ effector proteins are sufficient for Ca2+-induced ISC proliferation, we expressed trpA1 RNAi in ISCs together with each putative Ca2+ effector protein. Strikingly, MAPK activation via expression of active Ras1 (Ras1A), constitutively active Raf (Rafgof), or secreted form of spi is epistatic to trpA1 RNAi for ISCs proliferation (Figure 7A–B and A’–B’, 7F). Consistently, MAPK activation is epistatic to RyR RNAi (Figure 7—figure supplement 3). By contrast, trpA1 RNAi effectively blocked the mitogenic activity of constitutively active CanA1 (CanA1ca) (Dijkers and O'Farrell, 2007), CRTC (Kim et al., 2016), or active CrebB (CrebBact) (Ganguly-Fitzgerald et al., 2006) (Figure 7C–E and C’–E’, F). Furthermore, while Ras1 RNAi could suppress SERCA RNAi-induced ISC proliferation, neither CanA1 RNAi nor CrebB RNAi could (Figure 7—figure supplement 1C–G), and CRTC RNAi could only confer moderate suppression of ISC hyper-proliferation (Figure 7—figure supplement 1E–F). Results of these genetic epistasis analyses, together with examination of changes in pErk signal intensity (Figure 6), indicate that MAPK activity is both necessary and sufficient for high Ca2+-triggered ISC proliferation.

Figure 7 with 3 supplements see all
Ras/MAPK activity, not CanA/CRTC/CrebB, is sufficient for ISC proliferation in the absence of TRPA1.

(A–E) Midguts over-expressing calcium responsive signaling molecules such as active Ras1 (Ras1A), constitutively active Calcineurin A1 (CanA1ca), CREB-regulated transcription coactivator (CRTC), or active form of Creb (Crebact) in ISCs for 5d are stained for the mitosis marker pH3. The expansion of esg>GFP signal by CanA1ca is reduced by trpA1 RNAi, although not quite down to the wild-type level. This could be due to the different kinetics of two reagents, as CanA1ca may take effect sooner than trpA1 RNAi. (A’–E’) Midguts expressing trpA1 RNAi together with calcium-responsive signaling molecules in ISCs for 5d are stained for the mitosis marker pH3. (F) Mitosis quantification of midguts expressing GFP, Ras/ Raf, RTK ligands Spi/ Pvf1, CanA1ca, CRTC, Crebact, alone or together with trpA1 RNAi for genetic epistasis analysis. N > 5 midguts are analyzed for each genotype. Data are represented as mean ± SEM. Although it has been reported that pvf1 overexpression can increase ISC population (Bond and Foley, 2012), we could barely detect mitotic effect of Pvf1 in young adult flies.

https://doi.org/10.7554/eLife.22441.031

Previous studies have documented that receptor tyrosine kinase (RTK) EGFR is required for MAPK activity and ISC proliferation (Jiang et al., 2011). We found that EGFR RNAi completely inhibited ISCs proliferation caused by SERCA RNAi (Figure 7—figure supplement 1B,F), indicating that MAPK activation by Ca2+ requires EGFR signaling. Interestingly, as transcriptional regulation of ligands and receptors by their own pathways is commonly observed (see for example Ammeux et al., 2016), we examined by RT-qPCR the expression of EGFR ligands, as well as other RTK ligands, in midguts with cytosolic Ca2+ levels altered in ISCs. Interestingly, transcription of the EGFR ligand spi and the Pvr ligand pvf1 (Bond and Foley, 2012) were suppressed by trpA1 RNAi and up-regulated in midguts expressing SERCA RNAi in ISCs for 1 day to mimic early stages of Ca2+ induction (Figure 7—figure supplement 2). Altogether, these results suggest that positive RTK signaling feedback loops may play a role in the activation of Ras/MAPK in the context of Ca2+ signaling.

Discussion

ROS induces ISC proliferation via cation channel-mediated Ca2+ signaling

We have found that the two cation channels TRPA1 and RyR are critical for cytosolic Ca2+ signaling and ISC proliferation. Under homeostatic conditions, the basal activities of TRPA1 and RyR are required for maintaining cytosolic Ca2+ in ISCs to ensure their self-renewal activities and normal tissue turnover (Figure 8A). Agonists, including but not limited to low levels of ROS, could be responsible for the basal activities of TRPA1 and RyR. Under tissue damage conditions, increased ROS stimulates the channel activities of TRPA1 to mediate increases in cytosolic Ca2+ in ISCs (Figure 8B). As for RyR, besides its potential to directly sense ROS, it is known to act synergistically with TRPA1 in a positive feedback mechanism to release more Ca2+ from the ER into the cytosol upon sensing the initial Ca2+ influx through TRPA1 (Pan et al., 2016).

Connection between ROS, intracellular calcium, and stem cell activity.

(A) Under tissue homeostasis conditions, basal levels of ROS and other unidentified stimuli can activate TRPA1 and RyR channels at low levels to allow low levels of cytosolic Ca2+, autocrine EGFR ligand Spi, and Ras/MAPK activity required for stem cell self-renewal and minimal level of proliferation. Solid arrows indicate mechanisms that are either well-characterized (green) or demonstrated in this study; dashed arrows indicate unclear mechanisms inferred from literature or this study. (B) Under various tissue damage conditions, ROS levels dramatically increase and activate TRPA1, especially the D isoform. RyR channel can be activated either directly by ROS or by the initial Ca2+ influx through TRPA1, allowing further calcium release from the ER to the cytosol. High levels of cytosolic Ca2+ activate Ras/MAPK signaling via Src, and further amplify Ras/MAPK signaling via autocrine Spi-EGFR signaling. High EGFR-Ras/MAPK activity triggers ISC proliferation.

https://doi.org/10.7554/eLife.22441.036

Our findings on the role of Ca2+ signaling for ISC proliferation are consistent with a recent study showing that high levels of cytosolic Ca2+ correlate with ISC proliferation, and that a prolonged increase in cytosolic Ca2+ concentration is sufficient to trigger ISC proliferation (Deng et al., 2015). Here, we demonstrate that TRPA1 senses damage-induced ROS; that knockdown of trpA1 or RyR efficiently reduces Ca2+ induction in the ISCs caused by ROS; and that cytosolic Ca2+ induction is required for damage-induced ISC proliferation.

Regulation of cytosolic Ca2+ in ISCs

Previously, Deng et al. identified L-glutamate as a signal that can activate metabotropic glutamate receptor (mGluR) in ISCs, which in turn modulates the cytosolic Ca2+ oscillation pattern via phospholipase C (PLC) and inositol-1,4,5-trisphosphate (InsP3). Interestingly, L-glutamate and mGluR RNAi mainly affected the frequency of Ca2+ oscillation in ISCs, while their influence on cytosolic Ca2+ concentration was very weak. Strikingly, the number of mitotic cells induced by L-glutamate (i.e. an increase from a basal level of ~5 per midgut to ~10 per midgut) is far less than what has been observed in tissue damage conditions (depending on the severity of damage, the number varies from ~20 to more than 100 per midgut following damage) (Amcheslavsky et al., 2009; Chatterjee and Ip, 2009; Jiang et al., 2011; Karpowicz et al., 2010). Consistent with this, in our screen for regulators of ISC proliferation in response to tissue damage, we tested three RNAi lines targeting mGluR (BL25938, BL32872, and BL41668, which was used by Deng et al.), and none blocked the damage response in ISCs, suggesting that L-glutamate and mGluR do not play a major role in damage repair of the gut epithelium.

We find that ROS can trigger Ca2+ increases through the redox- sensitive cation channels TRPA1 and RyR under damage conditions. In particular, we demonstrate using voltage-clamp experiments that the TRPA1-D isoform, which is expressed in the midgut, is sensitive to the oxidant agent paraquat. In addition, the results of previous studies have demonstrated the direct response of RyR to oxidants via single channel recording (Terentyev et al., 2008) and showed that RyR could amplify TRPA1-mediated Ca2+ signaling through the Ca2+-induced Ca2+ release (CICR) mechanism. Interestingly, expression of oxidant- insensitive TRPA1-C isoform in the ISCs also exhibits a tendency to induce ISC proliferation (Figure 4H), indicating that ROS may not be the only stimuli for TRPA1 and RyR under physiological conditions. Possible other activators in the midgut may be irritant chemicals, noxious thermal/mechanical stimuli, or G-protein-coupled receptors (Fowler and Montell, 2013; Shen et al., 2011; Zhong et al., 2012).

Altogether, the concentration of cytosolic Ca2+ in ISCs appears to be regulated by a number of mechanisms/inputs including mGluR and the ion channels TRPA1 and RyR. Although mGluR might make a moderate contribution to cytosolic Ca2+ in ISCs, TRPA1 and RyR have a much stronger influence on ISC Ca2+ levels. Thus, it appears that the extent to which different inputs affect cytosolic Ca2+ concentration correlates with the extent of ISC proliferation.

Signaling events downstream of cytosolic Ca2+ in ISCs

Although, as a universal intracellular signal, cytosolic Ca2+ controls a plethora of cellular processes, we were able to demonstrate that cytosolic Ca2+ levels regulate Ras/MAPK activity in ISCs. Specifically, we found that trpA1 RNAi or RyR RNAi block Ras/MAPK activation in stem cells, and that forced cytosolic Ca2+ influx by SERCA RNAi induces Ras/MAPK activity. Moreover, Ras/MAPK activation is an early event following increases in cytosolic Ca2+, since we observed increased dpErk signal in stem cells expressing SERCA RNAi before they undergo massive expansion, and when we co-expressed Yki RNAi to block proliferation. It should be noted that a more variable pattern of pErk activation was observed with prolonged increases of cytosolic Ca2+, suggesting complicated regulations via negative feedback, cross-activation, and cell communication at late stages of Ca2+ signaling. This might explain why Deng et al. failed to detect pErk activation after 4 days induction of Ca2+ signaling (Deng et al., 2015). Previously, Ras/MAPK activity was reported to increase in ISCs, regulating proliferation rather than differentiation, in regenerating midguts (Jiang et al., 2011), which is consistent with our findings about TRPA1 and RyR.

The CanA1/CRTC/CrebB pathway previously proposed to act downstream of mGluR-calcium signaling (Deng et al., 2015) is not likely to play a major role in high Ca2+-induced ISC proliferation, as we tested multiple RNAi lines targeting CanA1 or CrebB and none of them suppressed SERCA RNAi-induced ISC proliferation. In support of this model, we also found that the active forms of CanA1/ CRTC/ CrebB cannot stimulate mitosis in ISCs when their cytosolic Ca2+ levels are restricted by trpA1 RNAi, whereas mitosis induced by the active forms of Ras or Raf is not suppressed by trpA1 RNAi.

Cross-talk between Ca2+ signaling and EGFR

Prior to our study, it has been shown that paracrine ligands such as Vn from the visceral muscle, and autocrine ligands such as Spi and Pvf ligands from the stem cells, can stimulate ISC proliferation via RTK-Ras/MAPK signaling (Bond and Foley, 2012; Jiang et al., 2011; Xu et al., 2011). We found that multiple RTK ligands in the midgut are down-regulated by trpA1 RNAi expression in the ISCs, including spi and pvf1 that can be induced by SERCA RNAi. Further, we demonstrated that high Ca2+ fails to induce ISC proliferation in the absence of EGFR. As spi is induced by EGFR-Ras/MAPK signaling in Drosophila cells (Ammeux et al., 2016), and DNA binding mapping (DamID) analyses from our lab and others indicate that spi might be a direct target of transcriptional factors downstream of EGFR-Ras/MAPK in the ISCs (Jin et al., 2015) (Doupé and Perrimon, unpublished), the autocrine ligand Spi might therefore act as a positive feedback mechanism for EGFR-Ras/MAPK signaling in ISCs.

In summary, our study identifies a mechanism by which ISCs sense microenvironment stress signals. The cation channels TRPA1 and RyR detect oxidative stress associated with tissue damage and mediate increases in cytosolic Ca2+ in ISCs to amplify and activate EGFR-Ras/MAPK signaling (Figure 8). In vertebrates, a number of cation channels (Monteith et al., 2007), including TRPA1 and RyR (Abdul et al., 2008; Büch et al., 2013; Du et al., 2014), have been associated with tumor malignancy. Our findings, unraveling the relationship between redox-sensing, cytosolic Ca2+, and pro-mitosis Ras/MAPK activity in ISCs, could potentially help understand the roles of cation channels in stem cells and cancers, and inspire novel pharmacological interventions to improve stem cell activity for regeneration purposes and suppress tumorigenic growth of stem cells.

Materials and methods

Drosophila stocks and culture

The following strains were obtained from the Bloomington Drosophila Stock Center: Dl-lacZ (BL11651), y v; attP2 (landing site only, BL36303), UAS-Luc RNAi (BL31603), UAS-trpA1 RNAi (BL31504, BL31384, BL36780), UAS-RyR RNAi (BL29445, BL31540, BL31695), UAS-SERCA RNAi (BL25928, BL44581), trpA11 (BL26504), UAS-trpA1-A (BL26263, also called UAS-TrpA1.K), UAS-Ras1 RNAi (BL29319), UAS-Yki RNAi (BL31965), UAS-Src42A RNAi (BL44039, validated by Sopko et al. (2014)), UAS-Src64B RNAi (BL51772), UAS-EGFR RNAi (BL25781, BL31525, BL31526), UAS-CanA1 RNAi (BL25850, phenotypically validated by Lee et al. (2016)), UAS-CrebB RNAi (BL63681, BL29332), UAS-CRTC RNAi (BL28886), UAS-rpr (BL5823), UAS-Hepca (BL6406), UAS-Src42Aca (BL6410), UAS-Src64B (BL8477), UAS-GCaMP6s (BL42749), UAS-GCaMP6f (BL42747), UAS-RFP LexAop2-GFP; UAS-MKII::nlsLexADBD, UAS-p65AD::CaM; UAS-p65AD::CaM, tubGal80ts (TRIC lexGFP reporter, BL62828), tubGal80ts (on the X chromosome, BL7016), tubGal80ts[10] (BL7108), tubGal80ts[20] (BL7019), tubGal80ts[7] (BL7018), UAS-mCherry (BL52268), UAS-mCherryCAAX (BL59021), RyRGal4R14G09 (with cloned enhancer region, BL48662), FRT19A (BL1744), FRT42D; FRT82B (BL8216), FRT42D w+ (BL1928), FRT42D; ry605 (BL1802), FRT2A (BL1997). Strains from Vienna Drosophila RNAi Center: yw; attP (landing site only, v60100), UAS-trpA1 RNAi (v37249) (Zhong et al., 2012), UAS-Ras1 RNAi (v106642), UAS-Yki RNAi (v104523). Strains from Perrimon lab stock: w1118, UAS-Luciferase, UAS-nlsGFP, UAS-Ras1A, UAS-Rafgof, esgGal4 UAS-GFP tubGal80ts (EGT), esgGal4, tubGal4, 24BGal4, DaGal4, esgGFP (generated by David Doupé), Su(H)GbeGFP (generated by Li He), UAS-spi (secreted form, validated in Ghiglione et al. (2002)). trpA1Gal4CP2A was generated by Junjie Luo. FRT42D RyRk04913 was generated from DGRC111172 (Kyoto Stock Center). trpA11 FRT2A and the stock with a deletion of trpA1, Df4415, were obtained from C. Montell lab, DlGal4 and Su(H)Gal4 from Steven Hou, UAS-Cas9.P2 from Fillip Port, UAS-2xEGFP [m5B29]; UAS-2xEGFP [m6B1] (UAS-ultraGFP) from Fabio Demontis, UAS-CncC RNAi from Dirk Bohmann, UAS-trpA1-C (also referred to as trpA1(A)10b) and UAS-trpA1-D (also referred to as trpA1(A)10a) (Guntur et al., 2015) from Paul Garrity, UAS-CanA1 RNAi fb5 and UAS-CanA1ca (constitutively active) (Wong et al., 2014) from Patrick H. O'Farrell, UAS-pvf1 from Edan Foley, UAS-CRTC and UAS-tdTomato-P2A-GCaMP5 from Henri Jasper, UAS-CrebBact from Jerry Yin. EGT F/O system is from Bruce Edgar: EGT; UAS-Flp, Act>Stop>Gal4. For MARCM induction, hsFlp tubGal80 FRT19A/ FM7; tubGal4 UAS-mCD8::GFP/ TM3 was obtained from Ben Ohlstein, hsFlp tubGal4 UAS-GFP-myc-nls;; tubGal80 FRT2A/ TM6B from Gary Struhl, and yw hsFlp UAS-GFP tubGal4; FRT42D tubGal80 from Huaqi Jiang.

Flies were reared on standard cornmeal/agar medium in 12: 12 light: dark cycles at 25°C unless noted otherwise. Fly food was changed every 2 days to keep fresh. Conditional expression in adult flies using tubGal80ts was achieved by maintaining flies at 18°C until 4–7 days after eclosion, and then shifting young adults to 29°C. For MARCM experiments, flies were maintained at 18°C until 3–5 days after eclosion, heat-shocked at 37°C for 1 hr, and then maintained back at 18°C before dissection and analysis.

See Supplementary file 1 for detailed genotypes of flies used in each figure.

Infection and compound feeding

Standard cornmeal/agar medium were melted at 70 ~ 80°C and mixed with a final concentration of 25 µg/ml bleomycin (Calbiochem #203408) or 2 mM paraquat (Sigma Aldrich #856177) to prepare food that can induce tissue damage in the fly midgut. For pathogen infection, Pseudomonas entomophila (from Tony IP) was grown in LB media at 29°C until reaching OD600 = 1, and Pseudomonas aeruginosa 14 (from Chrysoula Pitsouli) was grown in LB media at 37°C until reaching OD600 = 3. Flies were transferred to vials containing a cotton ball impregnated with 5 ml of water solution composed of 20% Pseudomonas culture media and 4% sucrose.

RNAi screen for regulators ISC pool size and damage responses

With a luciferase reporter that we have previously developed as a sensitive surrogate for stem cell pool size (Markstein et al., 2014), we performed an RNAi screen to identify novel components regulating ISC activity. In brief, midguts expressing luciferase together with RNAi targeting individual trans-membrane and receptor proteins (Hu et al., 2015) in adult ISCs for 9 days, with the last 2 days feeding either normally or on bleomycin food to induce tissue damage. Ten midguts were collected per genotype/ treatment, snap frozen, and stored at −80°C. After samples from all experiment groups were collected, the midguts were homogenized at 4°C using bead blender in Glo Lysis Buffer (Promega) with Halt protease inhibitor (Pierce 87786) and 2 mM trypsin inhibitor benzamidine. The lysate were centrifuged at 8000 rpm for 5 min to sediment the debris. And the supernatants were measured for luciferase activity using the Steady-Glo luciferase assay system (Promega). Luciferase activities were normalized to the value of spiked-in control groups (yv; attP2 for TRiP stocks, v60100 for Vienna stocks, and w1118 for NIG stocks). In particular, we noticed that trpA1 RNAi (BL31504), trpA1 RNAi (v37249), and RyR RNAi (BL29445) reduced EGT>luciferase activity by 74%, 88%, and 80%, respectively under normal feeding condition, and blocked damage-induced EGT>luciferase expression. We validated the top candidates from the screen by fluorescence imaging.

Staining and fluorescence imaging

Drosophila midguts were dissected in PBS, fixed in 4% paraformaldehyde PBS solution for 30 min, rinsed and kept for 1 hr in PBST solution (PBS with 0.1% BSA, 0.3% Triton X-100) containing 5% Normal Donkey Serum. Midguts were stained overnight at 4°C with the following primary antibodies in PBST solution: rabbit anti-pH3 (Millipore #06–570; 1:3000), mouse anti-GFP (Invitrogen A11120; 1:300), rabbit anti-GFP (Invitrogen A6455; 1:500), rabbit anti-RFP (Life Technologies R10367; 1:500), rabbit anti-β-galactosidase (Cappel; 1:6000), rabbit anti-dpErk1/2 (Cell Signaling #4370; 1:500), rabbit anti-Pdm1 (from Xiaohang Yang; 1:500), rat anti-TRPA1 (from Paul Garrity; 1:100). Fly midguts were then washed three times in PBST and incubated with secondary antibodies and DAPI (1:2000) in PBST for 2 hr in the dark at room temperature. Secondary antibodies were donkey anti-rabbit, anti-mouse, and anti-rat IgGs conjugated to Alexa488, Alexa594 and Alexa647 (used at 1:1000, Molecular Probes). Fly guts were then washed three times in PBST, mounted in Vectashield (Vector Laboratories). For anti-TRPA1 staining, primary antibody incubation was done for 1 day and secondary antibody incubation overnight at 4°C. Images of posterior midgut area were captured with a Zeiss LSM780 confocal microscope equipped with 40x oil lens. A z-stack of 10–20 images covering the epithelium from the apical to the basal side were acquired and shown in max projection. All images were adjusted and assembled using NIH ImageJ.

For counting mitotic cells or MARCM clone size, an epifluorescence microscope was used to examine the stained midguts. For the quantification of esg+, Pros+, Dl+ or Su(H)Gbe+ cell density, as well as GCamP+ cell number, images of the majority of posterior midgut region were acquired on a Keyence microscope.

For measuring midgut length, images of the whole midgut were captured with an epifluorescence microscope. The middle line of the midgut was measured using NIH ImageJ, starting from the end of the proventriculus to the border between the posterior midgut and hindgut, where the malphighian tubules connect to the gut.

For detection of ROS via DHE staining, we followed the previously described protocol (Hochmuth et al., 2011). In brief, midguts were dissected and handled throughout in Schneider’s medium (HyClone). After incubation in 30 µM DHE (Invitrogen) for 5 min in the dark at room temperature, midguts were washed three times and mounted in Schneider’s medium. Images were captured immediately with a Zeiss LSM780 confocal microscope (543 nm excitation, 550–610 nm detection). All images were adjusted and assembled using NIH ImageJ.

U6-sgRNA for CRISPR/Cas9-mediated conditional knockout

Two pairs of short guide RNAs (sgRNAs) targeting the coding region of trpA1 or RyR were designed using the ‘Find CRISPRs' online tool (http://www.flyrnai.org/crispr2/) (Housden et al., 2016), and cloned into the double sgRNA vector pCFD4 (Port et al., 2014) for site- specific insertion into attP2 on the third chromosome. We obtained transgenic flies ubiquitously (driven by U6 promoters) expressing sgRNAs with the following seed sequences:

U6-sgtrpA1: CACACGATCCACGTCACTGT & CTGTAGACTCCGTTGTAGAG

U6-sgRyR: TTCGTGTTCCATCTGTACAA & ACAAGAACGTGCCGCCGGAT

RT-qPCR and trpA1 isoform analysis

Total RNA was extracted from 15 to 20 midguts or 15 heads or 5 larvae (all females) by TRIZOL reagent (Thermo Fisher), converted to cDNA template after DNase I treatment and purification by QIAGEN RNeasy kit. The cDNA was analyzed by real-time quantitative PCR using SYBR Green or by regular PCR (for trpA1 isoform typing) with GAPDH and rp49 as an internal control. RT-qPCR data were analyzed with Bio-Rad CFX Manager software. Primers used for RT-qPCR or regular PCR (for trpA1 isoform typing) are shown below:

trpA1-All_FW: AGACGCTCATTAAAGTGCTGA

trpA1-All_RV: AGAAGGACAAGTGGTTCGGT

trpA1-CD_FW: CTTCAGGATATTGCGGGCG

trpA1-AB_FW: GTGGACTATCTGGAGGCGG

trpA1-AB/CD_RV: CCTTCGCATTAAAGTCGCCA

trpA1-AD_FW: TGCCGGCTTTGAATACCATG

trpA1-BC_FW: CTTTGGCCGTGGTGAACA

trpA1-AD/BC_RV: GCCAGGTGAAAGTACTTGCC

RyR_1_FW: GATCGGGACACAGGACATTG

RyR_1_RV: GACAGCTTGTCGTTCGACG

RyR_2_FW: TGCCTGACCATACCGAGTACA

RyR_2_RV: GCCACCAGTCCACTTGGTT

GAPDH_FW: CCAATGTCTCCGTTGTGGA

GAPDH_RV: TCGGTGTAGCCCAGGATT

rp49_FW: ATCGGTTACGGATCGAACAA

rp49_RV: GACAATCTCCTTGCGCTTCT

esg_FW: ATGAGCCGCAGGATTTGTG

esg_RV: CCTCCTCGATGTGTTCATCATCT

ε-Trypsin_FW: TAGACACCGGGATACCTCACATCA

ε-Trypsin_RV: TCCAGCATTCGCGAGATCCGTATT

Pdm1_FW: AGCTGTCCTAACGAGTTCCG

Pdm1_RV: ACATCGCGCATATTTGTGTCAA

pros_FW: TTTGACCGGAGATGGTGACG

pros_RV: GGTCGTTCCTGCCCAGTTTC

spi_FW: CGCCCAAGAATGAAAGAGAG

spi_RV: AGGTATGCTGCTGGTGGAAC

krn_FW: CGTGTTTGGCAACAACAAGT

krn_RV: TGTGGCAATGCAGTTTAAGG

vn_FW: AAGCATCACAAAAGACGTTC

vn_RV: CCGCATCGGAGGAACTATTGA

pvf1_FW: GCGCAGCATCATGAAATCAACCG

pvf1_RV: TGCACGCGGGCATATAGTAGTAG

pvf2_FW: TCAGCGACGAAACGTGCAAGAG

pvf2_RV: TTTGAATGCGGCGTCGTTCC

pvf3_FW: TCGTGAAGAGCAGTAAGCATCG

pvf3_RV: AGGTGCAACTCAGTATGGTGG

dilp3_FW: GCAATGACCAAGAGAACTTTGGA

dilp3_RV: GCAGGGAACGGTCTTCGA

upd1_FW: CAGCGCACGTGAAATAGCAT

upd1_RV: CGAGTCCTGAGGTAAGGGGA

upd2_FW: GACTCTTCTCCGGCAAATCAGA

upd2_RV: TGCTATCGCTGAGGCTCTCG

CRTC_FW: GCAGCAGTACGCAACAAACC

CRTC_RV: TTCACGTCGCTATTGAAGCCC

CanA1_FW: CCGCCAGTGGAAACAAACAG

CanA1_RV: GCGGAAGTGGAACATCATCG

Two-electrode clamp recordings in Xenopus oocytes

TRPA1 currents in Xenopus laevis oocytes were recorded as previously described (Kwon et al., 2010). Xenopus ovaries were dissected and treated with 1.5 mg/ml collagenase A (Roche) and 1 mg/ml trypsin inhibitor (Roche) in OR2 Ringers (100 mM NaCl, 2 mM KCl, 1 mM MgCl2, 5 mM HEPES, pH = 7.5). The deflocculated oocytes were recovered at 18°C for 12 hr in OR3 medium (pH = 7.5): 50% Leibovitz’s media L-15 (Sigma L1518), 13 mM HEPES, 90 μg/ml gentamicin, 90 μg/ml Fungizone, 90 μg/ml penicillin/streptomycin. trpA1-C and trpA1-D (Kang et al., 2011) mRNAs were generated from pCS2 or pOX expression vectors with the mMessage mMachine kit (Ambion) and diluted in DEPC-treated water at the concentration of 1 μg/μl. Stage V-VI oocytes were injected with 50 nl prepared mRNA or water, and incubated at 18°C for 3 days for optimal expression. Two electrode voltage clamp recordings were performed at −80 mV with TURBO TEC-03X system and Cellwork software (NPI Electronics, Tamm, Germany), using ND96 buffer (96 mM NaCl, 1 mM MgCl2, 4 mM KCl, and 5 mM HEPES, pH = 7.6) with 1.8 mM CaCl2. In each experiment, we used 4 mM paraquat to treat the oocytes expressing TRPA1-C or -D, and added 0.01% AITC near the end of the recordings as a positive control.

Imaging cytosolic Ca2+ with GCaMP reporters

Cytosolic Ca2+ in stem cells was monitored ex vivo using UAS-GCaMP6f (Figure 3E,E’, F,F’), UAS-GCaMP6s (Figure 5 and Supplementary Movies) or UAS-tdTomato-P2A-GCaMP5G (Figure 5—figure supplement 2A,A’, B,B’). In the latter case, tdTomato was used as an internal control. Young adult flies (4–7 days after eclosion) were first incubated at 29°C for 3 days to boost reporter expression before experiments. Intact midguts were dissected and handled in adult hemolymph-like (AHL) buffer. 1–2 intact midguts were placed in single wells of eight-well clear bottom cell culture chamber slides, gently oriented with Nylon mesh, and immersed in 200 µL AHL buffer for imaging. Images of the posterior midgut area (the region close to the middle midgut) were captured with a Zeiss LSM780 confocal microscope equipped with 10x lens. A z-stack of eight images covering one layer of the epithelium was recorded every 10 s for UAS-GCaMP6s, or every 20 s for UAS-tdTomato-P2A-GCaMP5G. Images before and after drug treatment are shown in z-max projection in Figure 5—figure supplement 2, while the time-lapse z-max projection movies for GCaMP6s are shown in Videos 16. Paraquat (Sigma Aldrich #856177), AITC (Sigma Aldrich #377430), and thapsigargin (Sigma Aldrich T9033) were dissolved in AHL buffer as 5x stock, and 50 µL drug stock was added to 200 µL imaging buffer to obtain desired final concentration. To quantify stem cells with high Ca2+ levels in most of the posterior region, Z-stack images were acquired on a Keyence microscope (Figure 5). All images were adjusted, assembled, and analyzed using NIH ImageJ.

Statistical methods

Statistical analyses were performed using Microsoft Excel (except for qPCR experiments, which were analyzed with Bio-Rad CFX Manager software). Average value and standard error of the mean (SEM) were calculated and shown in the figures. For comparison of data groups, p values were determined by Student’s two-tail t-test with unequal variances, except for the comparison before and after drug treatment in Figure 5D, where the p value was determined by two-tail paired t-test. Double asterisks (**) label a p value that is less than 0.01. The p value of more than 0.1 is labeled as ‘not significant’ (n.s.) or not labeled. Sample sizes were chosen empirically based on the observed effects and listed in the figure legends.

References

  1. 1
  2. 2
  3. 3
  4. 4
  5. 5
  6. 6
  7. 7
  8. 8
  9. 9
  10. 10
  11. 11
  12. 12
  13. 13
  14. 14
  15. 15
  16. 16
  17. 17
  18. 18
  19. 19
  20. 20
  21. 21
    Mechanism of activation of the Drosophila EGF receptor by the TGFalpha ligand gurken during oogenesis
    1. C Ghiglione
    2. EA Bach
    3. Y Paraiso
    4. KL Carraway
    5. S Noselli
    6. N Perrimon
    (2002)
    Development 129:175–186.
  22. 22
  23. 23
  24. 24
  25. 25
    Expression of baculovirus P35 prevents cell death in Drosophila
    1. BA Hay
    2. T Wolff
    3. GM Rubin
    (1994)
    Development 120:2121–2129.
  26. 26
  27. 27
  28. 28
  29. 29
  30. 30
  31. 31
  32. 32
  33. 33
  34. 34
  35. 35
  36. 36
  37. 37
  38. 38
  39. 39
  40. 40
  41. 41
  42. 42
  43. 43
  44. 44
  45. 45
  46. 46
  47. 47
  48. 48
  49. 49
  50. 50
  51. 51
  52. 52
  53. 53
  54. 54
  55. 55
  56. 56
  57. 57
  58. 58
  59. 59
  60. 60
  61. 61
  62. 62
  63. 63
  64. 64
  65. 65
  66. 66
  67. 67
  68. 68
  69. 69
  70. 70
  71. 71
  72. 72
  73. 73
  74. 74
  75. 75
  76. 76
  77. 77
  78. 78
  79. 79
  80. 80

Decision letter

  1. Bruce Edgar
    Reviewing Editor; University of Utah, United States

In the interests of transparency, eLife includes the editorial decision letter and accompanying author responses. A lightly edited version of the letter sent to the authors after peer review is shown, indicating the most substantive concerns; minor comments are not usually included.

Thank you for submitting your article "Oxidative stress induces stem cell proliferation via TRPA1/RyR mediated Ca2+ signaling in the Drosophila midgut" for consideration by eLife. Your article has been favorably evaluated by Jonathan Cooper (Senior Editor) and three reviewers, one of whom, Bruce Edgar (Reviewer #1), is a member of our Board of Reviewing Editors. The following individual involved in review of your submission has agreed to reveal their identity: W. Daniel Tracey (Reviewer #2).

The reviewers have discussed the reviews with one another and the Reviewing Editor has drafted this decision to help you prepare a revised submission.

General assessment:

Xu et al. demonstrate that two conserved regulators of intracellular calcium, TrpA1 and RyR, are required for stem cell stress responses in the Drosophila gut, and provide evidence that one of these effectors, the Ca2+ channel TRPA1, may be activated by reactive oxygen species (ROS). They present an extensive and coherent set of tests supporting a model in which ROS trigger Ca2+ release via TRPA1 in ISCs, and that the Ca2+ in turn activates Ras/Mapk signaling to promote intestinal stem cell (ISC) divisions. The reviewers agreed that the experiments were generally performed to a high standard and that most of the conclusions were convincingly supported. However, no mechanistic insight was offered as to how ROS might activate TRPA1, or how Ca2+ might activate Ras, though similar effects have been reported before in other cell types. Given that other studies of the fly gut have concluded that Ras/Mapk activation in ISCs is effected non-cell autonomously, via the EGFR and ligands produced by other cells, the authors' proposal that Ras is activated cell autonomously by Ca2+ is novel, interesting, and provocative.

Essential revisions:

1) Two reviewers noted that authors' proposal that the effect of ROS via Ca2+ on ERK and ISC division is cell autonomous, is not well supported and runs counter to many previous publications suggesting that ERK is activated by stress-dependent EGFR ligands non-cell autonomously supplied from other cells in the epithelium. The authors should either provide additional data supporting their cell-autonomous mechanism and ruling out non-autonomous mechanisms, or they should alter the model they present in Figure 7. In any case the autonomous vs. non-autonomous models should be carefully discussed.

2) Along the same lines, reviewer #3 notes that further data should be provided to rule out the role of CrebB and CanA as effectors of Ca2+, and to support the authors proposal that ERK is the principal Ca2+ effector.

3) Reviewer #2 notes a substantial body of relevant published work that was not cited or discussed (i.e. Du and Guntur papers). The manuscript should be revised to address these deficiencies.

4) Two reviewers commented on the weakness of data showing the TRPA1 activation is sufficient to promote ISC divisions, and reviewer #2 suggests an experiment to test this using a heat-activatable TRPA1 variant. If this experiment is possible, the authors should try it, or present some other type of data further supporting the sufficiency of TRPA1 mediated Ca2+ release to promote ISC division. This might also address point #1 above.

5) Reviewers 1and 3 discussed several issues with conclusions that could be clarified by counting ISCs and other cell types (EEs), from experiments that were already performed. This additional analysis should be provided if possible.

Reviewer #1:

In this paper Xu et al. demonstrate that two conserved regulators of intracellular calcium, TrpA1 and RyR, are required for stem cell stress responses in the Drosophila gut, and provide evidence that one of these effectors, the Ca2+ channel TRPA1, may be activated by reactive oxygen species (ROS). They present an extensive and coherent set of tests supporting a model in which ROS trigger Ca2+ release via TRPA1 in ISCs, and that the Ca2+ in turn activates Ras/Mapk signaling to promote ISC divisions. No mechanistic insight is offered as to how ROS might activate TRPA1, or how Ca2+ might activate Ras, though similar effects have been reported before in neurons (also without mechanistic explanation). The results are consistent with a recent report in Nature from Jasper's group (Deng, 2016). Given that other studies of the fly gut have concluded that Ras/Mapk activation in ISCs is effected non-cell autonomously, via the EGFR and ligands produced by other cells, the authors' proposal (Figure 7) that Ras is activated cell autonomously by Ca2+ is novel and interesting. However, this study does not definitively rule out non-cell autonomous signaling as a mechanism of ROS sensing or Ca2+ release, and so I think the model should be presented with reference to the rather different conclusions of previous work. Otherwise, apart from a few relatively minor issues with data analysis and interpretation, the paper is a nice story in principle appropriate for eLife. I list issues that need to be addressed in more detail below:

1) Subsection “TRPA1 and RyR are required for stem cell damage response and self-renewal”, third paragraph, Figure 2E-H: The authors show that the trpA1 and RyR RNAi's deplete Δ+ stem cells in 5 days. Nevertheless, they conclude that these RNAi's are not killing stem cells. This is not very convincing, and I think the number of stem cells should be counted using several assays, at several timepoints, since the issue of cell viability affects other critical results and the overall interpretation. For instance, in Figure 3 we see decreases in ISC mitoses. These results would be confounded if the total number of ISCs decreases during the experiment.

2) Similarly, in Figure 4 we see increases in high Ca2+ cells after paraquat treatment in controls, but not in animals expressing RNAi against TRPA1. (Subsection “TRPA1 and RyR are required for normal and ROS-induced cytosolic Ca2+ 146 levels in ISCs”, last paragraph, Figure 4—figure supplement 1). But these increases could be due to division of cells with high Ca2+, rather than to an increase in the fraction of ISCs with high Ca2+. This would not occur in the RNAi treated ISCs, since the RNAi blocks their division. To address this potentially confounding issue the authors should count total ISC numbers in these experiments, as well as numbers of high Ca2+ cells, and display the fraction of ISCs that are Ca2+ high.

3) As alluded to above, the regulation of Ras activity by Ca2+ is not explained mechanistically, and the authors' model that the effect of ROS via Ca2+ is autonomous to stem cells is based on correlations that don't rule out other explanations. Although TrpA1 and RyR are clearly necessary for ISC activation, and although other treatments that increase Ca2+ are sufficient to activate ISCs (e.g. SERCA-RNAi), this paper doesn't provide strong evidence that activation of TrpA1 or RyR (in this case by ROS) is sufficient to activate ISCs. Moreover the data on ROS activation of TrpA1, though extensive and essential to the model (Figure 7), is the weakest in the paper. Thus, although the authors' model is coherent and consistent with their data, it seems possible that the increase of Ca2+ and activation of ERK in ISCs actually involves signals from other cells in the epithelium, rather than a direct effect of ROS on TrpA1, as depicted in Figure 7. It would be interesting, for instance, to test whether the EGFR is required in this context (it is required for ISC responses to many stresses, but should not be required here). In the absence of experiments ruling out roles for other cell types, I feel the model figure (Figure 7) and Discussion should be modified to include the possibility that the oxidative damage signals are received by cells other than stem cells.

4) The dpERK stainings in Figure 5A-F should be repeated with anti-TrpA1 RNAi's. Only tests with RyR-RNAi are shown.

5) For most of the study, only TrpA1 is tested in detail. The manuscript would be better if the authors had carried through with a deeper analysis of RyR as well

Reviewer #2:

The study of Xu et al. finds an important role for the Drosophila TRPA1 channel and Ryanodine Receptor channels in Reactive Oxygen Species (ROS) induced proliferation of intestinal stem cells. The combined evidence in support of the role of these two channels in this biological process as presented in the study is very strong. With respect to dTRPA1, the study points to a specific isoform (dTRPA1-D) in the sensing of ROS and the study also finds that Ras/Mapk function downstream of Ca influx/release that are triggered by dTRPA1 and Ryr respectively. The strength of the advance in the study is that it ties calcium signaling to ISC proliferation specifically in the context of ROS triggered proliferation and the study also specifically ties the calcium to dTRPA1 and Ryr. The authors do discuss the fact that a prior study has tied calcium to ISC proliferation through mGluRs and dietary glutamate. But some of the prior literature on dTRPA1 channels (in the context of ROS signaling) has not been adequately credited by the current study. For example, the dTRPA1-D isoform has been previously shown to be activated by ROS (Drosophila TRPA1 isoforms detect UV light via photochemical production of H2O2.

Guntur AR, Gu P, Takle K, Chen J, Xiang Y, Yang CH.Proc Natl Acad Sci U S A. 2015 Oct 20;112(42):) through a dual oxidase pathway in the ring gland. As well, a role for dTRPA1 in sensing ROS in the gut (also through dual oxidase) has been previously reported: TrpA1 Regulates Defecation of Food-Borne Pathogens under the Control of the Duox Pathway. Du EJ, Ahn TJ, Kwon I, Lee JH, Park JH, Park SH, Kang TM, Cho H, Kim TJ, Kim HW, Jun Y, Lee HJ, Lee YS, Kwon JY, Kang K.PLoS Genet. 2016 Jan 4;12(1)). This latter study also finds that it is the dTRPA1-D isoform which is activated by ROS and the study has not been cited in the current manuscript. Although Guntur et al. is briefly cited in the current study, the analysis of ROS activation by dTRPA1in heterologous oocytes do not significantly extend those that were initially reported by the studies of Guntur and Du. As such the new heterologous expression analyses should thus be presented as important confirmatory findings (even if investigating different chemical compounds to generate the ROS) with a bit more credit/discussion given to the prior studies. The new biological role for this function- in promoting ISC proliferation in response to ROS- is what is important here, and this will be of great interest to the readers of eLife.

Experimental points to be addressed:

The anti-dTRPA1 staining does not provide a clear description of where dTRPA-C/D is actually expressed in the gut. Reporters for dTRPA1-C/D are available from several labs and these could be used to more clearly illustrate the expression pattern.

Reviewer #3:

In this manuscript, Xu et al. describe a novel mechanism controlling intestinal stem cell proliferation in Drosophila. They found that oxidative stress stimulates Ca2+ signaling by regulating TRPA1 and RyR. While the role of calcium signaling has previously been reported, this new study describes a different pathway causing changes in [Ca2+] and a different downstream effector cascade (Ras/ERK).

Experiments reported in this paper are generally well designed and properly controlled. They address an important question in the field, regarding the mechanisms that allow the integration of multiple signaling pathways in ISCs to control proliferation rates in response to stress (oxidative stress and tissue damage).

This study has the potential to be of interest for the large audience of eLife.

In my opinion, however, a few important points would need to be addressed to fully support the authors' main conclusions:

The authors show that TRPA1 and RyR knock-down results in a loss of Dl+ stem cells. Because they argue that these cells do not die of apoptosis and do not commit to the enteroblast lineage, I am confused about what the fate of these ISCs is, after these genetic manipulations. The authors should at least test whether it promotes enteroendocrine cells fate by monitoring the number of prospero-positive cell in these epithelia. Also, it is possible that the expression of Δ is lost in these stem cells? Other ISC markers such as Sanpodo would be useful to test this hypothesis. Finally, are the single cell trpA and RyR mutant clones Dl-positive?

Altogether, these experiments would greatly help clarifying the fate of ISCs in which trpA or RyR activities is impaired.

The authors use paraquat exposure as an oxidative stress in most of their experiments. While the results are compatible with a regulation of Ca2+ signaling by ROS via a cell autonomous mechanism (including the control of TRPA1 activity), relatively simple experiments could be performed to rule out a non-cell autonomous mechanism following paraquat-induced tissue damage.

Does the expression of antioxidant proteins in ISCs prevent ROS-induced Ca2+ signaling? Do ISC-specific genetic manipulations that have been shown to induce ROS production in ISCs and promote cell proliferation (such as ND42 RNAi) increase Ca2+ signaling?

In the last part of this manuscript, the authors try to exclude the role of CanA1 and CrebB downstream of Ca2+ release, arguing that Ras/ERK is the main Ca2+ effector in ISCs. This is an important point as it has been previously published that these factors are part of the signaling pathway that regulate ISC proliferation in response to [Ca] changes, and this new study challenges this view. I believe that this conclusion needs to be supported by additional data.

The experiment in Figure 6H needs to be quantified and the other genetic manipulations (multiple RNAi lines targeting CanA1 and CrebB mentioned as "not shown" in the Results and Discussion) should be included in the main figures. It would be essential here to demonstrate that CanA1 and CrebB are dispensable for ROS- and SERCARNAi-mediated proliferation as suggested by the authors' model.

[Editors' note: further revisions were requested prior to acceptance, as described below.]

Thank you for resubmitting your work entitled "Oxidative stress induces stem cell proliferation via TRPA1/RyR mediated Ca2+ signaling in the Drosophila midgut" for further consideration at eLife. Your revised article has been favorably evaluated by Jonathan Cooper (Senior Editor), a Reviewing Editor and two reviewers.

The consensus is that the manuscript has been improved, and we appreciate the new data you've included as well as your editorial efforts. However, there are two remaining issues that are quite significant, and really need to be addressed before acceptance, as outlined below:

1) One reviewer and the reviewing editor still believe the manuscript still falls short of establishing how, mechanistically, [Ca2+] activates ERK. A key issue is whether this happens directly, cell-autonomously, or through non-autonomous interactions between different cells in the gut epithelium. The old and new data do not resolve this issue, despite requests from reviewers. The model figure tries to cover both (or all possible) mechanisms and as such is a confusing conclusion – in fact the model as presented will only serve to obscure whatever the real mechanism is, not to resolve it. We would very much like to see a clear resolution of this issue, so that this paper can be seen as an obvious advance over the recent similar publication from Deng et al. A reviewer's specific comments on this point follow: "[…] the authors provide new data demonstrating that [Ca2+] controls the expression of Spitz and other RTK ligands. However, the authors appear to favor a model where [Ca2+] controls ERK activation directly through Src, but such activation is not demonstrated and the mechanism is unclear. As an example of this, the revised model still highlights a direct regulation of Ras/MAPK by Ca2+, and the authors discuss a scenario where Spitz induction is downstream of an initial direct activation of ERK rather than an integral part of the ERK activation mechanism. I think the two models (direct or autocrine/paracrine) could be easily be distinguished by testing whether EGFR and/or PVR are required for ISC proliferation induced by SERCARNAi and TRPA1-D overexpression, as suggested in the initial review. I believe that confirming that Ca2+ can activate ERK directly in ISCs (independently of a RTK signal) would greatly increase the impact of this study." Conversely, demonstrating the EGFR is essential for deregulated proliferation after Ca2+ signaling dysfunction would at least show that the mechanism is not simple and not cell autonomous.

2) A second, and related, issue is how your results and conclusions relate to the Deng et al. publication. We feel that you should show the relevant data you have and clarify how your conclusions are different from Deng's and why. On this issue, a reviewer commented: "some of the data required is not presented and some of the conflicting data not discussed. For example, the genetic interaction between SERCARNAi and CrebBRNAi (CrebBRNAi not being able to block SERCARNAi-induced proliferation) is still reported as "data not shown" in the revised version. This result is used as an important piece of data to support the authors' conclusion that ERK is the major effector of Ca2+ in the control of ISC proliferation. Deng et al. reported that CrebBRNAi expression causes ISC loss: do the authors here find something different? Also, Deng et al. reported that dpERK levels are not affected by SERCARNAi expression. This inconsistency should be at least noted and discussed." Although we feel these issues are less pressing than #1 above, we would appreciate some revisions to make it clear how your results and conclusions differ from those presented by Deng et al.

https://doi.org/10.7554/eLife.22441.039

Author response

Essential revisions: 1) Two reviewers noted that authors' proposal that the effect of ROS via Ca2+ on ERK and ISC division is cell autonomous, is not well supported and runs counter to many previous publications suggesting that ERK is activated by stress-dependent EGFR ligands non-cell autonomously supplied from other cells in the epithelium. The authors should either provide additional data supporting their cell-autonomous mechanism and ruling out non-autonomous mechanisms, or they should alter the model they present in Figure 7. In any case the autonomous vs. non-autonomous models should be carefully discussed.

Since it was reported that Ras activation increases Ca2+level in the ISCs (Deng et al., 2016), and EGFR ligand spi might act as a positive feedback signal for EGFR/Ras/MAPK pathway (Ammeux et al., 2016), the Ca2+mechanism and EGFR mechanism of Erk activation are not mutually exclusive but most likely intertwined. We have revised the model in Figure 7 to reflect the multiple layers of regulation for Ras/MAPK signaling (explained in subsection “Signaling events downstream of cytosolic Ca2+in ISCs”, second paragraph).

2) Along the same lines, reviewer #3 notes that further data should be provided to rule out the role of CrebB and CanA as effectors of Ca2+, and to support the authors proposal that ERK is the principal Ca2+ effector.

We have quantified the epistasis analysis to demonstrate that CanA1 RNAi cannot reduce ISC proliferation caused by SERCA RNAi (Figure 7—figure supplement 1E). In addition, we are presenting the quantification of experiments using the CanA1 RNAi line (“FB5”) from Dr. O’Farrell (Wong et al., 2014). Thus, we are very confident that SERCA RNAi- induced ISC proliferation does not depend on CanA1. In a modified screen to identify suppressor of SERCA RNAi, we have tested RNAi lines against calcium signaling effectors including CrebB (BL29332), CaMKII (BL29401, BL35330) and other calcineurin family proteins like CanB (BL27307), CanB2 (BL27270, BL38971), Pp2B-14D (BL25929), none of which rescued the over-proliferation phenotype.

3) Reviewer #2 notes a substantial body of relevant published work that was not cited or discussed (i.e. Du and Guntur papers). The manuscript should be revised to address these deficiencies.

The papers suggested have been cited and discussed in the revised manuscript (see subsection “TRPA1 and RyR expression in the midgut”, first paragraph and subsection “TRPA1-D in the midgut senses the oxidative stress associated with tissue damage”, second paragraph).

4) Two reviewers commented on the weakness of data showing the TRPA1 activation is sufficient to promote ISC divisions, and reviewer #2 suggests an experiment to test this using a heat-activatable TRPA1 variant. If this experiment is possible, the authors should try it, or present some other type of data further supporting the sufficiency of TRPA1 mediated Ca2+ release to promote ISC division. This might also address point #1 above.

We have performed the mitosis quantification of midguts overexpressing TRPA1 isoforms C and D, as well as the heat-activatable TRPA1 isoform A (Figure 4H). The data support the sufficiency of TRPA1 mediated Ca2+release to promote ISC division.

5) Reviewers 1and 3 discussed several issues with conclusions that could be clarified by counting ISCs and other cell types (EEs), from experiments that were already performed. This additional analysis should be provided if possible.

We have added the time-course quantification of ISC (Figure 1—figure supplement 1E) and EE (Figure 2M) cell density in the revised manuscript. Besides, we have used EGT F/O (flp out tracing of ISC progenies) and MARCM systems for lineage tracing, and presented the images of midguts stained for the terminal differentiation markers Pros (for EE) and Pdm1 (for EC) (Figure 2—figure supplement 1). Our data suggest that while TRPA1 and RyR are required for maintaining basal level of cytosolic Ca2+and ISC self-renewal, and do not affect ISC differentiation.

Reviewer #1:

[…] 1) Subsection “TRPA1 and RyR are required for stem cell damage response and self-renewal”, third paragraph, Figure 2E-H: The authors show that the trpA1 and RyR RNAi's deplete Δ+ stem cells in 5 days. Nevertheless, they conclude that these RNAi's are not killing stem cells. This is not very convincing, and I think the number of stem cells should be counted using several assays, at several timepoints, since the issue of cell viability affects other critical results and the overall interpretation. For instance, in Figure 3 we see decreases in ISC mitoses. These results would be confounded if the total number of ISCs decreases during the experiment.

The observation that anti-apoptotic p35 protein cannot rescue ISC proliferation defects (Figure 1—figure supplement 2D-I, D’-I’), and the added data of anti-cleaved caspase3 stainings (Figure 1—figure supplement 2A-C), demonstrate that trpA1 and RyR RNAi are not causing ISC apoptosis. We have taken the reviewer’s advice and counted ISC and EE density after 3d, 5d, 7d of trpA1 or RyR RNAi expression. Although the signal of Dl-lacZ, indicating ISC maintenance, reduces significantly after 5 days of RNAi expression (Figure 2D-F, 2K), the density of overall esg+ stem cell population exhibits a much more moderate decrease (Figure 1—figure supplement 1E). Moreover, ISC differentiation activity (measured by Notch reporter Su(H)Gbe-GFP, Figure 2G-I,L) and EE density (Figure 2M) remains relatively stable during the time frame for most of our experiments. Our data suggest that trpA1 or RyR RNAi affect ISC proliferation but not differentiation, therefore the ISC population undergo gradual loss as they differentiate.

2) Similarly, in Figure 4 we see increases in high Ca2+ cells after paraquat treatment in controls, but not in animals expressing RNAi against TRPA1. (Subsection “TRPA1 and RyR are required for normal and ROS-induced cytosolic Ca2+ 146 levels in ISCs”, last paragraph, Figure 4—figure supplement 1). But these increases could be due to division of cells with high Ca2+, rather than to an increase in the fraction of ISCs with high Ca2+. This would not occur in the RNAi treated ISCs, since the RNAi blocks their division. To address this potentially confounding issue the authors should count total ISC numbers in these experiments, as well as numbers of high Ca2+cells, and display the fraction of ISCs that are Ca2+high.

The increases of high Ca2+cells caused by paraquat treatment cannot be caused by expansion of cells with high Ca2+because the GCamP imaging lasted less than 20 min, not long enough for cell division. We have also stained flies used in Figure 5 (Figure 4 in original manuscript) with anti-GFP antibody (which recognizes GCamP6s) to visualize total ISC population. We did not observe significant changes in the total ISC number or density by trpA1 RNAi or RyR RNAi expression for the time frame we used to perform the GCamP imaging experiments (Figure 5—figure supplement 1).

3) As alluded to above, the regulation of Ras activity by Ca2+ is not explained mechanistically, and the authors' model that the effect of ROS via Ca2+ is autonomous to stem cells is based on correlations that don't rule out other explanations. Although TrpA1 and RyR are clearly necessary for ISC activation, and although other treatments that increase Ca2+ are sufficient to activate ISCs (e.g. SERCA-RNAi), this paper doesn't provide strong evidence that activation of TrpA1 or RyR (in this case by ROS) is sufficient to activate ISCs. Moreover the data on ROS activation of TrpA1, though extensive and essential to the model (Figure 7), is the weakest in the paper. Thus, although the authors' model is coherent and consistent with their data, it seems possible that the increase of Ca2+ and activation of ERK in ISCs actually involves signals from other cells in the epithelium, rather than a direct effect of ROS on TrpA1, as depicted in Figure 7. It would be interesting, for instance, to test whether the EGFR is required in this context (it is required for ISC responses to many stresses, but should not be required here). In the absence of experiments ruling out roles for other cell types, I feel the model figure (Figure 7) and Discussion should be modified to include the possibility that the oxidative damage signals are received by cells other than stem cells.

Mitosis quantification of midguts overexpressing TRPA1 isoforms C and D, as well as the heat-activatable TRPA1 isoform A (Figure 4H) demonstrates that TRPA1 gain-of-function can significantly increase ISC division. As the reviewer pointed out, previous published studies have demonstrated that EGFR and its ligands induced after tissue damage can be responsible for Ras/MAPK/Erk activation. We do not think that our mechanism is mutually exclusive with the EGFR-related mechanism. Besides the paracrine ligand vn, EGFR in the progenitor population (ISCs are EBs) can act autonomously by sensing the autocrine ligands spi and Krn (Xu et al., 2011). It was reported before in fly cell lines that spi might act as a positive feedback mechanism for EGFR/Ras/MAPK signaling (Ammeux et al., 2016). Further, on-going work in our lab has identified spi as a target of Ras/MAPK signaling in the ISCs (Doupe et al., unpublished). By RT-qPCR of midgut mRNA we found that the transcription of spi is suppressed by trpA1 RNAi and induced by SERCA RNAi in the ISCs (Figure 6—figure supplement 2). Therefore, TRPA1-mediated Ca2+ might also be required for autocrine activation of EGFR/Ras/MAPK via spi, through or in parallel with the Src-mechanism that we have alluded to. On the other hand, it was reported that Ras activation causes increases in cytosolic Ca2+ level (Deng et al., 2016). We hypothesize that the Ca2+ and EGFR mechanisms are intertwined, amplify each other for activation, and turn off each other if one of them gets restricted. The hypothesis is reflected in the revised model in Figure 8.

4) The dpERK stainings in Figure 5A-F should be repeated with anti-TrpA1 RNAi's. Only tests with RyR-RNAi are shown.

We have added the imaging experiments as suggested (Figure 6C,G). As control, wild type midguts (EGT>Luc RNAi) was processed alongside the experimental groups of Figure 6C and G, and the levels of pErk signal look the same as Figure 6A.

5) For most of the study, only TrpA1 is tested in detail. The manuscript would be better if the authors had carried through with a deeper analysis of RyR as well

We have added quantification of RyR RNAi effect on EE and EB population (Figure 2L-M); and added the observation that Ras1 is epistatic to RyR RNAi for ISC proliferation (Figure 7—figure supplement 2). Because RyR is not at the plasma membrane, we cannot perform the direct measurement of RyR response to ROS as we did with TRPA1, due to the limitation of our equipment.

Reviewer #2:

The study of Xu et al. finds an important role for the Drosophila TRPA1 channel and Ryanodine Receptor channels in Reactive Oxygen Species (ROS) induced proliferation of intestinal stem cells. The combined evidence in support of the role of these two channels in this biological process as presented in the study is very strong. With respect to dTRPA1, the study points to a specific isoform (dTRPA1-D) in the sensing of ROS and the study also finds that Ras/Mapk function downstream of Ca influx/release that are triggered by dTRPA1 and Ryr respectively. The strength of the advance in the study is that it ties calcium signaling to ISC proliferation specifically in the context of ROS triggered proliferation and the study also specifically ties the calcium to dTRPA1 and Ryr. The authors do discuss the fact that a prior study has tied calcium to ISC proliferation through mGluRs and dietary glutamate. But some of the prior literature on dTRPA1 channels (in the context of ROS signaling) has not been adequately credited by the current study. For example, the dTRPA1-D isoform has been previously shown to be activated by ROS (Drosophila TRPA1 isoforms detect UV light via photochemical production of H2O2.

Guntur AR, Gu P, Takle K, Chen J, Xiang Y, Yang CH.Proc Natl Acad Sci U S A. 2015 Oct 20;112(42):) through a dual oxidase pathway in the ring gland. As well, a role for dTRPA1 in sensing ROS in the gut (also through dual oxidase) has been previously reported: TrpA1 Regulates Defecation of Food-Borne Pathogens under the Control of the Duox Pathway. Du EJ, Ahn TJ, Kwon I, Lee JH, Park JH, Park SH, Kang TM, Cho H, Kim TJ, Kim HW, Jun Y, Lee HJ, Lee YS, Kwon JY, Kang K.PLoS Genet. 2016 Jan 4;12(1)). This latter study also finds that it is the dTRPA1-D isoform which is activated by ROS and the study has not been cited in the current manuscript. Although Guntur et al. is briefly cited in the current study, the analysis of ROS activation by dTRPA1in heterologous oocytes do not significantly extend those that were initially reported by the studies of Guntur and Du. As such the new heterologous expression analyses should thus be presented as important confirmatory findings (even if investigating different chemical compounds to generate the ROS) with a bit more credit/discussion given to the prior studies. The new biological role for this function- in promoting ISC proliferation in response to ROS- is what is important here, and this will be of great interest to the readers of eLife.

We cited the studies of Guntur et al. 2015 in the original manuscript but did not discuss it. Now we have cited and discussed both studies (subsection “TRPA1 and RyR expression in the midgut”, first paragraph and subsection “TRPA1-D in the midgut senses the oxidative stress associated with tissue damage”, second paragraph). Guntur et al. found that TRPA1-C and TRPA1-D isoforms are H2O2-sensitive and that their endogenous expressions in the ring gland, or ectopic expression in other tissues, confer UV light sensitivity via sensing UV light-induced H2O2 production. Although Du et al., 2016 reported that TRPA1-D mediates much stronger NaOCl-induced current than TRPA1-C, it seems inconsistent that they could rescue uracil-dependent defecation defects in trpA1 mutant flies with TRPA1-C overexpression but not with TRPA1-D. We found that TRPA1-D, rather than TRPA1-C, exhibits response to paraquat (Figure 4A-B), and that TRPA1-D overexpression in ISCs causes a stronger proliferation phenotype than TRPA1-C (Figure 4C-F,H).

Experimental points to be addressed:

The anti-dTRPA1 staining does not provide a clear description of where dTRPA-C/D is actually expressed in the gut. Reporters for dTRPA1-C/D are available from several labs and these could be used to more clearly illustrate the expression pattern.

The signal to noise ratio is not ideal for immunostaining with anti-dTRPA1 antibody in the midgut (Figure 3A-B). Personal communication with Dr. KyeongJin Kang (who used the antibody for the Du et al., 2016 paper) also confirmed that the antibody seems quite dirty. To confirm trpA1 expression, we inserted Gal4 at the C terminal of endogenous dTRPA1 with CRISPR/Cas9 (Figure 3—figure supplement 1). With this Gal4 line we could detect trpA1 expression in some ISCs in the posterior midgut areas that is close to the middle midgut region as well as the posterior end of the midgut, where ISC proliferation is most active (Figure 3C). We further demonstrated that for trpA1 mutant flies, there is no calcium influx in ISCs in response to the trpA1 agonist AITC (Figure 3E,E’,F,F’). Moreover, knockdown of trpA1 in Dl+ stem cells with Dl-Gal4 inhibits ISC proliferation (Figure 3—figure supplement 2B). These data serve as functional evidence that trpA1 is expressed in the ISCs.

Reviewer #3:

[…] In my opinion, however, a few important points would need to be addressed to fully support the authors' main conclusions:

The authors show that TRPA1 and RyR knock-down results in a loss of Dl+ stem cells. Because they argue that these cells do not die of apoptosis and do not commit to the enteroblast lineage, I am confused about what the fate of these ISCs is, after these genetic manipulations. The authors should at least test whether it promotes enteroendocrine cells fate by monitoring the number of prospero-positive cell in these epithelia. Also, it is possible that the expression of Δ is lost in these stem cells? Other ISC markers such as Sanpodo would be useful to test this hypothesis. Finally, are the single cell trpA and RyR mutant clones Dl-positive?

Altogether, these experiments would greatly help clarifying the fate of ISCs in which trpA or RyR activities is impaired.

A more careful quantification of esg+ (Figure 1—figure supplement 1E), Su(H)Gbe+ (Figure 2L), and Pros+ cell density (Figure 2M) has been performed. Besides, lineage-tracing experiments with EGT F/O system, or MARCM, suggest that stem cells deficient for trpA1 or RyR can still differentiate into mature progenies (Figure 2—figure supplement 1). These data support our current hypothesis that trpA1 or RyR knockdown allows for ISC differentiation but blocks ISC self-renewal and proliferation. Therefore, trpA1 or RyR RNAi will cause a gradual loss of esg+ stem cells as they differentiate. During the time frame for most of our experiments, total ISC density (esg+) exhibits very modest change with trpA1 or RyR RNAi expression in ISCs, even though the hallmark of ISC self-renewal (Dl+) and damage-induced ISC proliferation decrease significantly. Since the flies lack ISC self-renewal activity, the total ISC population will gradually decrease as the flies age, and eventually be unable to replenish the mature cell types during tissue turnover.

The authors use paraquat exposure as an oxidative stress in most of their experiments. While the results are compatible with a regulation of Ca2+ signaling by ROS via a cell autonomous mechanism (including the control of TRPA1 activity), relatively simple experiments could be performed to rule out a non-cell autonomous mechanism following paraquat-induced tissue damage.

Does the expression of antioxidant proteins in ISCs prevent ROS-induced Ca2+ signaling? Do ISC-specific genetic manipulations that have been shown to induce ROS production in ISCs and promote cell proliferation (such as ND42 RNAi) increase Ca2+ signaling?

As previously reported, tissue damage can induce ligands expression from the microenvironment, for example the EGFR ligand vn from the visceral muscle, to stimulate ISC proliferation (Jiang et al., 2011). We do not think the non-cell autonomous mechanisms are mutually exclusive with the Ca2+ signaling mechanism we have identified. In the revised manuscript we have incorporated the known mechanism via EGFR into our model (Figure 8). Expression of the antioxidant protein CncC in ISC can reduce its proliferation (Hochmuth et al., 2011). Due to the complexity of the genotype we have not performed GCamP imaging to confirm if ROS-induction of Ca2+ signaling is decreased by CncC, or if CncC knockdown can induce Ca2+ in ISCs, but we have obtained indirect evidence that CnCC RNAi-induced ISC hyper-proliferation status is dependent on trpA1 (Figure 4G).

In the last part of this manuscript, the authors try to exclude the role of CanA1 and CrebB downstream of Ca2+ release, arguing that Ras/ERK is the main Ca2+ effector in ISCs. This is an important point as it has been previously published that these factors are part of the signaling pathway that regulate ISC proliferation in response to [Ca] changes, and this new study challenges this view. I believe that this conclusion needs to be supported by additional data.

The experiment in Figure 6H needs to be quantified and the other genetic manipulations (multiple RNAi lines targeting CanA1 and CrebB mentioned as "not shown" in the Results and Discussion) should be included in the main figures. It would be essential here to demonstrate that CanA1 and CrebB are dispensable for ROS- and SERCARNAi-mediated proliferation as suggested by the authors' model.

As the reviewer suggested, we quantified mitosis (Figure 7—figure supplement 1E) for genotypes used in Figure 7—figure supplement 1A-B (Figure 6H in original manuscript). Quantification of mitosis for an additional validated RNAi line against CanA1 (FB5 from Dr. O’Farrell, Figure 7—figure supplement 1F) also demonstrates that CanA1 RNAi cannot suppress hyper-proliferation caused by SERCA RNAi.

[Editors' note: further revisions were requested prior to acceptance, as described below.]

The consensus is that the manuscript has been improved, and we appreciate the new data you've included as well as your editorial efforts. However, there are two remaining issues that are quite significant, and really need to be addressed before acceptance, as outlined below:

1) One reviewer and the reviewing editor still believe the manuscript still falls short of establishing how, mechanistically, [Ca++] activates ERK. A key issue is whether this happens directly, cell-autonomously, or through non-autonomous interactions between different cells in the gut epithelium. The old and new data do not resolve this issue, despite requests from reviewers. The model figure tries to cover both (or all possible) mechanisms and as such is a confusing conclusion – in fact the model as presented will only serve to obscure whatever the real mechanism is, not to resolve it. We would very much like to see a clear resolution of this issue, so that this paper can be seen as an obvious advance over the recent similar publication from Deng et al. A reviewer's specific comments on this point follow: "[…] the authors provide new data demonstrating that [Ca2+] controls the expression of Spitz and other RTK ligands. However, the authors appear to favor a model where [Ca2+] controls ERK activation directly through Src, but such activation is not demonstrated and the mechanism is unclear. As an example of this, the revised model still highlights a direct regulation of Ras/MAPK by Ca2+, and the authors discuss a scenario where Spitz induction is downstream of an initial direct activation of ERK rather than an integral part of the ERK activation mechanism. I think the two models (direct or autocrine/paracrine) could be easily be distinguished by testing whether EGFR and/or PVR are required for ISC proliferation induced by SERCARNAi and TRPA1-D overexpression, as suggested in the initial review. I believe that confirming that Ca++ can activate ERK directly in ISCs (independently of a RTK signal) would greatly increase the impact of this study." Conversely, demonstrating the EGFR is essential for deregulated proliferation after Ca++ signaling dysfunction would at least show that the mechanism is not simple and not cell autonomous.

As suggested by the reviewers, we performed the genetic epistasis experiments and found that EGFR RNAi can suppress SERCA RNAi-induced ISC proliferation (Figure 7—figure supplement 1B and F). In addition, we have shown that activation of EGFR by spi over-expression even allows ISC to bypass the requirement of TRPA1 for proliferation (Figure 7F). While there are various RTK ligands that might activate Ras/MAPK, sometimes in a redundant fashion, the intensity of signal they trigger differs by several orders of magnitude. In particular, Spi is known as a strong EGFR ligand compared to weak non-autonomous ligand such as Vn (Austin et al., 2014). And spi expression is required in the ISCs for neoplastic proliferation caused by Notch RNAi (Patel et al., 2015), suggesting a non-redundant role of autocrine Spi-EGFR signaling in supporting ISC proliferation in some contexts. Ongoing work in our lab suggests that spi might be the target of transcription factors downstream of EGFR/Ras/MAPK pathway (DamID analysis by Doupé et al.) in the ISCs. Therefore, Spi could be involved in a positive feedback mechanism to enhance EGFR/Ras/MAPK signaling in the ISCs. We are now proposing that both EGFR and [Ca2+] are required for MAPK activation and ISC proliferation, and autocrine Spi-EGFR signaling might participate in amplifying MAPK activity, downstream of [Ca2+] (revised model in Figure 8). In the revised model, [Ca2+]-sensitive Src could be a possible mechanism for the initial stimulation of Ras/MAPK by [Ca2+], as the kinetics of which is considered to be much faster than transcriptional regulation in the neurons (Randlett et al., 2015; Rosen et al., 1994). We have added the mitosis quantification showing that Src is epistatic to trpA1 RNAi (Figure 6—figure supplement 2H), and that Src64B RNAi can partially suppress SERCA RNAi-induced ISC proliferation (Figure 6—figure supplement 2I).

Specifically, to address the issue raised by the reviewer we have now revised Figure 6—figure supplement 2, Figure 7, Figure 7—figure supplement 1, Figure 8, reassigned Figure 7—figure supplement 2, adjusted the figure legends accordingly, and added the following paragraphs to the main text:

“In neurons, a common mechanism by which Ca2+ regulates Ras/MAPK is through Ca2+ sensitive Src kinases (Cullen and Lockyer, 2002), in a process faster than any transcriptional response (Randlett et al., 2015; Rosen et al., 1994). […] These data suggest that Src might mediate the activation of Ras/MAPK by cytosolic Ca2+ in ISCs.”

“Previous studies have documented that receptor tyrosine kinase (RTK) EGFR is required for MAPK activity and ISC proliferation (Jiang et al., 2011). […] Altogether, these results suggest that positive RTK signaling feedback loops may play a role in the activation of Ras/MAPK in the context of Ca2+ signaling.”

“Cross-talk between Ca2+ signaling and EGFR

Prior to our study, it has been shown that paracrine ligands such as Vn from the visceral muscle, and autocrine ligands such as Spi and Pvf ligands from the stem cells, can stimulate ISC proliferation via RTK-Ras/MAPK signaling (Bond and Foley, 2012; Jiang et al., 2011; Xu et al., 2011). […] As spi is induced by EGFR-Ras/MAPK signaling in Drosophila cells (Ammeux et al., 2016), and DNA binding mapping (DamID) analyses from our lab and others indicate that spi might be a direct target of transcriptional factors downstream of EGFR-Ras/MAPK in the ISCs (Jin et al., 2015), the autocrine ligand Spi might act as a positive feedback mechanism for EGFR-Ras/MAPK signaling in ISCs.”

2) A second, and related, issue is how your results and conclusions relate to the Deng et al. publication. We feel that you should show the relevant data you have and clarify how your conclusions are different from Deng's and why. On this issue, a reviewer commented: "some of the data required is not presented and some of the conflicting data not discussed. For example, the genetic interaction between SERCARNAi and CrebBRNAi (CrebBRNAi not being able to block SERCARNAi-induced proliferation) is still reported as "data not shown" in the revised version. This result is used as an important piece of data to support the authors' conclusion that ERK is the major effector of Ca++ in the control of ISC proliferation. Deng et al. reported that CrebBRNAi expression causes ISC loss: do the authors here find something different? Also, Deng et al. reported that dpERK levels are not affected by SERCARNAi expression. This inconsistency should be at least noted and discussed." Although we feel these issues are less pressing than #1 above, we would appreciate some revisions to make it clear how your results and conclusions differ from those presented by Deng et al.

As requested by the reviewers, we have included the epistasis analysis of SERCA RNAi with CrebB RNAi and CRTC RNAi (Figure 7—figure supplement 1D-F). In addition, as the reviewers pointed out, Deng et al. claimed that dpERK levels are not affected by SERCA RNAi expression (in Extended Figure 9C from their paper, Author response image 1). To address the inconsistency, we found that after ISCs start massive expansion with prolonged high Ca2+ induction, pErk returns to normal levels in many ISCs (probably through some negative feedback mechanism) and is non-autonomously induced in some ECs (Figure 6—figure supplement 1), resulting in a diffusive and more variable pattern of activation. Such kinetics we observed might explain why Deng et al. failed to detect pErk induction, especially when they were imaging very small fields of 2-3 cells for midgut samples at late stages of Ca2+ signaling.

Specifically, to address the issue raised by the reviewer we have now added Figure 6—figure supplement 1, revised Figure 7—figure supplement 1, adjusted the figure legends accordingly, and added the following paragraphs to the main text:

“It should be noted that the best timing to detect autonomous activation of dpErk by high Ca2+ in the ISCs is before they enter the hyper-proliferative stage. Once ISCs start to expand massively following prolonged high Ca2+ induction, pErk returns to normal levels in many ISCs and is non-autonomously induced in some ECs, resulting in a diffusive activation pattern (Figure 6—figure supplement 1A-C).”

“Furthermore, while Ras1 RNAi could suppress SERCA RNAi-induced ISC proliferation, neither CanA1 RNAi nor CrebB RNAi could (Figure 7—figure supplement 1C-G), and CRTC RNAi could only confer moderate suppression of ISC hyper-proliferation (Figure 7—figure supplement 1E-F).”

https://doi.org/10.7554/eLife.22441.040

Article and author information

Author details

  1. Chiwei Xu

    Department of Genetics, Harvard Medical School, Boston, United States
    Contribution
    CX, Conceptualization, Data curation, Formal analysis, Investigation, Visualization, Methodology, Writing—original draft, Writing—review and editing
    Competing interests
    The authors declare that no competing interests exist.
  2. Junjie Luo

    1. Department of Molecular, Cellular and Developmental Biology, University of California Santa Barbara, Santa Barbara, United States
    2. Neuroscience Research Institute, University of California, Santa Barbara, Santa Barbara, United States
    Contribution
    JL, Investigation, Methodology, Writing—review and editing
    Competing interests
    The authors declare that no competing interests exist.
  3. Li He

    Department of Genetics, Harvard Medical School, Boston, United States
    Contribution
    LH, Software, Formal analysis, Methodology, Writing—review and editing
    Competing interests
    The authors declare that no competing interests exist.
  4. Craig Montell

    1. Department of Molecular, Cellular and Developmental Biology, University of California Santa Barbara, Santa Barbara, United States
    2. Neuroscience Research Institute, University of California, Santa Barbara, Santa Barbara, United States
    Contribution
    CM, Formal analysis, Funding acquisition, Writing—review and editing
    Competing interests
    The authors declare that no competing interests exist.
  5. Norbert Perrimon

    1. Department of Genetics, Harvard Medical School, Boston, United States
    2. Howard Hughes Medical Institute, Harvard Medical School, Boston, United States
    Contribution
    NP, Conceptualization, Resources, Supervision, Funding acquisition, Writing—original draft, Project administration, Writing—review and editing
    For correspondence
    perrimon@receptor.med.harvard.edu
    Competing interests
    The authors declare that no competing interests exist.
    ORCID icon 0000-0001-7542-472X

Funding

National Eye Institute (EY010852)

  • Craig Montell

National Institute on Deafness and Other Communication Disorders (DC007864)

  • Craig Montell

Howard Hughes Medical Institute

  • Norbert Perrimon

National Institute of General Medical Sciences (GM067761)

  • Norbert Perrimon

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

Acknowledgements

We thank the Transgenic RNAi Project (TRiP) and Bloomington Drosophila Stock Center (BDSC) for providing fly stocks; we thank Drs. Paul Garrity, Patrick O'Farrell, Henri Jasper, Jerry Yin, Ben Ohlstein, Gary Struhl, Huaqi Jiang, Dirk Bohmann, and Xiaohang Yang for sharing reagents and fly stocks; we thank the members of the Perrimon lab, especially Kwon Young, Yanhui Hu, David Doupé, Stephanie Mohr, Patrick Jouandin, and Leonard Rabinow, for advice on experiments and comments on the manuscript. Work in the Montell lab is supported by the NIDCD (DC007864) and the NEI (EY010852). Work in the Perrimon lab is supported by NIGMS (GM067761) and HHMI. N.P. is an Investigator of the Howard Hughes Medical Institute.

Reviewing Editor

  1. Bruce Edgar, Reviewing Editor, University of Utah, United States

Publication history

  1. Received: October 16, 2016
  2. Accepted: May 23, 2017
  3. Version of Record published: May 31, 2017 (version 1)

Copyright

© 2017, Xu et al.

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

Metrics

  • 1,423
    Page views
  • 376
    Downloads
  • 5
    Citations

Article citation count generated by polling the highest count across the following sources: Scopus, Crossref, PubMed Central.

Download links

A two-part list of links to download the article, or parts of the article, in various formats.

Downloads (link to download the article as PDF)

Download citations (links to download the citations from this article in formats compatible with various reference manager tools)

Open citations (links to open the citations from this article in various online reference manager services)

Further reading

    1. Cell Biology
    2. Chromosomes and Gene Expression
    Madhav Jagannathan et al.
    Research Article Updated
    1. Cell Biology
    2. Chromosomes and Gene Expression
    Susan A Gerbi
    Insight