To explore the function of Cln3 in Drosophila midgut, we previously generated a Cln3 mutant line named Cln3 ^ΔMB1 by imprecise excision of a transposable element (Tuxworth et al., 2011). This allele exhibites a deletion of 1,534 bp and can be viewed as a null (Tuxworth et al., 2011). This homozygote mutant survived to adulthood and showed no phenotype in the midgut epithelium (a model of cell lineage was shown in Figure 1A) from the following aspects: (Lerner et al., 1995): escargot ⁺ (esg ⁺, a promotor activating in stem and progenitor cells) cells (Figures 1B–D); (Nita et al., 2016) Delta⁺ (Dl⁺, the ligand of Notch signaling) intestinal stem cells (ISCs) (Figures 1B, E); (Cooper, 2003) Prospero⁺ (Pros) enteroendocrine cells (EEs) (Figures 1C, F); (Mirza et al., 2019) polyploid enterocytes (ECs) (Figures 1B, C, G). Also, phosphorylated histone H3 (pH3, marks mitotic cells) staining revealed no difference in ISC proliferation between Cln3 ^ΔMB1 mutant and wild type flies (Figure 1H). These data suggested that Cln3 is not a necessary gene under homeostasis condition in the Drosophila midgut.

Since we previously demonstrated that Cln3 ^ΔMB1 mutant flies showed no phenotype in normal condition but were hypersensitive to oxidative stress (Tuxworth et al., 2011), we asked whether Cln3 also functions only under stress condition in the midgut. To test this, we fed Drosophila with bleomycin (Figure 2A) and Erwinia carototovovora carototovovora 15 (Ecc15) (Supplementary Figure S1), which induced DNA break of ECs and infection of the midgut, respectively, and subsequent activation of ISCs to replenish the damage of epithelium (Guo et al., 2013). We first tested the RNA expression level of Cln3 in the midgut and found it upregulated after bleomycin and Ecc15 treatment (Figure 2B; Supplementary Figure S1A). This suggested that Cln3 may play a role in the regeneration process. Then, we checked the changes of ISC behavior. As a result, while the percentage of esg ⁺ and Dl⁺ cells increased in wild type flies after bleomycin and Ecc15 treatment, Cln3 ^ΔMB1 mutant failed to exhibit this increase (Figures 2C–E; Supplementary Figures S1B–D). Moreover, pH3 staining revealed fewer proliferative stem cells in the Cln3 ^ΔMB1 mutant (Figure 2F; Supplementary Figure S1E). These results indicated that Cln3 plays a role in the proliferation of ISCs during regeneration. Since both of bleomycin and Ecc15 treatment revealed similar results, we used only bleomycin for further investigation.

To further support this, we used mosaic analysis with a repressible cell marker (MARCM) system (Lee and Luo, 1999), in which only active ISCs are able to generate a clone. Results revealed that compared to wild type clones, Cln3 ^ΔMB1 mutant clones were smaller (Figures 2G, H), suggesting weaker mitotic ability of mutated ISCs. Noticed that both wild type clones and mutant clones contained Pros⁺ EEs (Figure 2G), which further supported the conclusion that Cln3 did not affect ISC differentiation.

Next, we sought to figure out which cell type is responsible for this failure of ISC activation. We first tested the efficiency of two Cln3 RNAi lines. Results revealed that both of these two RNAi lines driven by whole body expressed tub-gal4 exhibited over 60% of Cln3 mRNA depletion (Figure 3A), suggesting high RNAi efficiency. Then, we used different kinds of cell type specific-gal4 to drive Cln3 RNAi (stem and progenitor cell-specific esg-gal4; ISC-specific esg-gal4, NRE-gal80, tub-gal80 ^ts (shortened as ISC ^ts -gal4); EE-specific 386Y-gal4; EC-specific Mex-gal4 and muscle-specific How-gal4). These results combinely suggested that the failure of ISC activation during regeneration was cell-autonomous, because only esg-gal4 and ISC ^ts -gal4 driven Cln3 RNAi led to the failure of the increase of esg ⁺ cells, Dl⁺ cells and pH3⁺ cells upon bleomycin treatment (Figures 3B–F).

Then, we asked the mechanism underlying how Cln3 regulates ISC proliferation during regeneration. JAK/STAT signaling is well reported to regulate ISC proliferation (Jiang et al., 2009; Zhai et al., 2017; Osman et al., 2012). During regeneration, ECs that undergo apoptosis secreate the ligand of JAK/STAT signaling, Unpaired 3, which binds to the receptor Domeless expressed through the plasma membrane of ISCs, to activate ISC proliferation (Osman et al., 2012). To test whether Cln3 depletion has an impact on JAK/STAT signaling in ISCs, 10 × STAT-GFP reporter line was used (Bach et al., 2007). During regeneration, while the control group showed more cells with 10 × STAT-GFP positive signal, stronger 10 × STAT-GFP signal per cell, and higher GFP protein level, these changes were not significant in the Cln3 depletion group (Figures 4A–D). Moreover, we tested the mRNA level of two STAT-target genes, Socs36E and Dpp, which revealed similar results (Supplementary Figure S2A). These data suggested that JAK-STAT signaling was not able to switch on with Cln3 depletion.

To further verify the function of JAK/STAT signaling upon Cln3 depletion, we tested the genetic relationship between JAK/STAT signaling and Cln3. Apart from the function of reporting JAK/STAT signaling, Socs36E is also a vital negative regulator which mediates degradation of positive regulator STAT92E (Issigonis et al., 2009). We drove Socs36E RNAi to upregulate JAK/STAT signaling. During regeneration, while wild type flies exhibited increased esg ⁺ cells, Dl⁺ cells and pH3⁺ cells, these numbers reached a lower level in Cln3 depletion group (Figures 4E–H). In the Cln3 and Socs36E double depletion group, the esg ⁺ cell number, Dl⁺ cell number and pH3⁺ cell number showed the same level to Socs36E depletion group (Figures 4E–H). These data indicated that JAK/STAT signaling pathway functioned downstream of Cln3 in response to stress.

The fly stocks used in this study are listed as follows: wild type line: w ¹¹¹⁸ ;; (BDSC: 3605); Cln3 mutant line: ;;Cln3 ^ΔMB1 (Our lab); esg-GFP reporter line: ;esg-GFP/CyO; (From Allan Spradling); 10 × STAT-GFP reporter line: ;10 × STAT-GFP; (From Zheng Guo); tub-gal4 line: ;;tub-gal4/TM3.e.Sb (From Allan Spradling); esg-gal4 line: ;esg-gal4/CyO; (From Allan Spradling); ISC ^ts -gal4 line: Su(H)-lacZ;esg-gal4.UAS-GFP/Kr-GFP.CyO;tub-gal80ts.GBE-gal80/MKRS (From Zheng Guo); Mex-gal4 line: ;Mex-gal4; (From Zheng Guo); 386Y-gal4 line: w[*];; P{w[+mW.hs] = GawB}386Y (BDSC: 25410); How-gal4 line: w[*];; P{w[+mW.hs] = GawB}how[24B] (BDSC: 1767); Cln3 RNAi 1# line: TH201500324.S; Cln3 RNAi 2# line: THU1691; Socs36E RNAi line: THU1666; UAS-GFP reporter line: w[1118];; P{w[+mC] = UAS-GFP.nls}8 (BDSC: 4776). FRT79D line: ;;FRT79D (From Zheng Guo); MARCM79D tool line: yw.hsflp.tub-gal4::GFP/FM7;;FRT79D, tub-gal80/TM6B (From Zheng Guo).

All flies were cultured at 25°C at a 12 h light/dark cycle. For the Gal4/Gal80^ts system, the offspring were shifted to 29°C 5 days after eclosion for removing Gal80 inhibition and enabling Gal4 to drive UAS-linked transgene expression for another 5 days. For MARCM system, flies were raised at 25°C until 5 days after eclosion. Then the flies were transferred to a new vial and incubated in 37°C water bath for 1-h-heat shock. Flies were transferred back to 25°C for another 7 days for clone growth followed by dissection.

Drosophila were transferred from the medium to empty bottles for 2 h. The filter paper was cut into 3.5 × 6.0 cm pieces and treated with 5% (wt/vol) sucrose with 25 μg/mL bleomycin (BLM). Then, the moist papers were added to empty bottles for 24 h. BLM-1d group flies were dissected at this point, and the rest flies were transferred into a new standard medium without BLM for another 24 h before dissection. For Ecc15 infection, an infection solution was prepared by mixing an equal volume of 100× concentrated pellet from an overnight culture of Ecc15 (OD₆₀₀ = 200) with a solution of 5% sucrose. One piece of filter paper treated with this infection solution was placed into a bottle with standard food. After treatment with empty bottles for 2 h, Drosophila were transferred to the bottles with standard food and filter paper for 24 h followed by dissection. Filter paper treated with 5% sucrose were used as control.

Adult fly intestines were dissected at 4°C in PBS and fixed with a mixture of 100 µL 4% EM-grade paraformaldehyde fixation buffer and 100 µL n-Heptane for 20 min. The intestines were washed in 200 µL methanol for 2 times, 5 min each. They were subsequently rinsed in 200 µL PBS plus 0.1% Triton X-100 (Sigma, PBST) for 2 times, 5 min each. The tissues were then incubated with primary antibodies dissolved in 0.1% PBST for overnight at 4°C. Then they were washed with 0.1% PBST for 3 times, 5 min each before secondary antibodies and DAPI (1:1,000; Sigma) incubated for 2 h at room temperature. After secondary antibodies incubation, the intestines were washed with 0.1% PBST for 3 times, 5 min each. The sources and dilutions of antibodies used are listed as follows: Chicken anti-GFP (1:1000; ab13970, Abcam), Rabbit anti-GFP (1:1000; 50430-2-AP, Proteintech), Prospero antibody (1:200; MR1A, DSHB), Delta antibody, (1:50; C594,9B, DSHB), and Alexa Fluor secondary antibodies (1:2000; A11004, A11008, A11011, A11039, A32733 and A32933). Leica TCS-SP8 confocal microscope was used to acquire all immunofluorescence images. For each set of experiments, images were acquired as confocal stacks using the same settings.

50 adult midguts or 10 whole flies were collected into 4°C diethylpyrocarbonate (DEPC)-treated water-PBS solution. Samples were homogenized in RNA-easy Isolation Reagent (Vazyme, R701) for total RNA isolation and cDNA synthesis. RT-qPCR was performed on a CFX96 Touch Deep Well (Bio-Rad) using ChamQ Universal SYBR qPCR Master Mix (Vazyme, Q711). The reference standard group was Rp49. The expression levels were counted by the 2^−△△CT method. The primers used were listed as below:

Rp49-F: GCC​CAA​GGG​TAT​CGA​CAA​CA.

Rp49-R: GCG​CTT​GTT​CGA​TCC​GTA​AC.

CLN3-F: TCG​TCG​GGA​AGA​AAC​TGT​CAC.

CLN3-R: GAG​CTA​TAC​GGA​AAT​TCA​CCC​AA.

Socs36E-F: GCT​GCC​AGT​CAG​CAA​TAT​GT.

Socs36E-R: GAC​TGC​GGC​AGC​AAC​TGT.

Dpp-F: TGG​CGA​CTT​TTC​AAA​CGA​TTG​T.

Dpp-R: CAG​CGG​AAT​ATG​AGC​GGC​AA.

Immunofluorescence imaging results were analyzed based on z-stacks acquired with confocal microscopy. The fluorescence intensity of the region of interest (ROI) or cell was calculated using ImageJ software following the fomular: Integrated Density = Integrated Density of ROI or cell — Integrated Density of background region / Area of background region × Area of ROI or cell .

50 midguts were collected for protein extraction. The primary antibodies used for western blotting in this study were: anti-GFP (rabbit, 1:5000, 50430-2-AP, Proteintech); and anti-β-actin (rabbit, 1:2000, PA5-85271, Invitrogen). The secondary antibody was horseradish peroxidase-conjugated goat anti-rabbit (1:5000, Jackson ImmunoResearch Labs, Cat# 211-032-171).