EGFR-dependent TOR-independent endocycles support Drosophila gut epithelial regeneration

Following gut epithelial damage, epidermal growth factor receptor/mitogen-activated protein kinase (EGFR/MAPK) signalling triggers Drosophila intestinal stem cells to produce enteroblasts (EBs) and enterocytes (ECs) that regenerate the gut. As EBs differentiate into ECs, they become postmitotic, but undergo extensive growth and DNA endoreplication. Here we report that EGFR/RAS/MAPK signalling is required and sufficient to drive damage-induced EB/EC growth. Endoreplication occurs exclusively in EBs and newborn ECs that inherit EGFR and active MAPK from fast-dividing progenitors. Mature ECs lack EGF receptors and are refractory to growth signalling. Genetic tests indicated that stress-dependent EGFR/MAPK promotes gut regeneration via a novel mechanism that operates independently of Insulin/Pi3K/TOR signalling, which is nevertheless required in nonstressed conditions. The E2f1 transcription factor is required for and sufficient to drive EC endoreplication, and Ras/Raf signalling upregulates E2f1 levels posttranscriptionally. We illustrate how distinct signalling mechanisms direct stress-dependent versus homeostatic regeneration, and highlight the importance of postmitotic cell growth in gut epithelial repair.


GFP marked cells driven by Su(H)GBE-Gal4
were digested into single cell suspensions. Cell volumes for the GFP+ EBs and pre-ECs in each sample were measured using the forward scatter parameter of FACS. Error bars represent standard deviations. The Student's T-test was used to determine statistical significance. (****: p<0.0001).

Supplementary Figure 2 | Transcript levels of MAP kinase components in different cell types of the adult midgut
100 midguts/sample were digested by Collagenase and Trypsin into single cell suspensions. Then GFP-marked ISCs, EBs or ECs were purified by FACS using GFP expressed by the Dl-Gal4, Su(H)GBE-Gal4, or MyoIA-Gal4 drivers, respectively. mRNA was purified from each cell type and sequenced by Hiseq 2000. The absolute transcription value was normalized to RPKM. See (40) for protocol and further data. The heatmap illustrates selected genes in the EGFR/Map kinase pathway. Colors indicate mRNA gene expression levels (red: high, green: low). (Raw data from FlyGutSeq). (Z-score scale is indicated in the up-left corner of figure).

Supplementary Figure 3 | Ras/Raf activity in ECs promotes endocycle progression
Overexpression of Ras V12 or Raf GOF by Myo1A-Gal4 ts increased DNA re-replication in ECs. Flies were shifted to 29°C for 1 day to induce the indicated transgenes. After dissection, midguts were incubated with EdU in vitro for 2h, and then fixed. Midguts were stained with DAPI (blue) and incorporated EdU was fluorescently labeled (red). (a) Wild-type control guts displayed very little EdU incorporation. Overexpression of Ras V12 or Raf GOF strongly induced endoreplication, as indicated by extensive EdU incorporation. (b) Analysis of EC ploidy by FACS. Midguts were dissected from Ras V12 overexpressing and control flies prior to digestion with collagenase and trypsin to generate a single cell suspension. Nuclei were purified from the cells before staining with DAPI, a specific DNA-binding dye, and then analyzed by FACS. Compared to controls, Ras V12 overexpression resulted in more nuclei with higher DNA copy number. Left: control (blue), middle: overexpression of Ras V12 (red) by Myo1A-Gal4, right: Merge. All experiments were repeated three times.

Supplementary Figure 4 | EGFR, InR and Ras function assayed in midgut MARCM clones
EGFR co , InR 339 , or Ras85D ΔC40B mutant clones were made using the MARCM system. Several days after clone induction, midguts were dissected. DNA was stained with DAPI (blue). Clones were marked with GFP (green), and are outlined by yellow dashed lines. All experiments were repeated three times.
(a) FRT control and EGFR co mutant clones 3 days after induction. (b) 2 days after clone induction flies were infected by P.e. for 24h. (c) Quantification of DNA/nucleus for GFP+ clone cells from a-b. 80 GFP+ nuclei from 10 midguts were scored for each genotype. Error bars represent standard deviations and the Student's T-test was used to determine statistical significance. (****: p<0.0001, ns: p>0.05). Loss of EGFR resulted in lower ploidy only after P.e. infection.
(d) FRT control and InR 339 mutant clones. (e) 2 days after clone induction, FRT control and InR mutant flies were infected by P.e. for 24h. InR -/mutant clones were arrested as one or two diploid cells under normal conditions, but grew into large clones with polyploid ECs after P.e. infection.
(f) 5 day old Ras85D C40B null mutant clones generated using the MARCM system. DAPI stain shows DNA content (blue), and clones were marked with GFP (green). PDM1 antibody staining indicates differentiated ECs (red). Many Ras -/mutant cells are PDM1 positive.

Supplementary Figure 5 | Ras activity bypasses TOR to induce cell division, endoreplication and clonal growth
(a) TOR P mutant clones either expressing or not expressing Ras V12S35 were generated using the MARCM system. All flies were fed Rapamycin for two days prior to clone induction to ensure strong suppression of TOR activity. 5 days after clone induction, with continuous exposure to Rapamycin, midguts were harvested and labeled with DAPI (blue). Clones are marked with GFP (green). Clonal boundaries are outlined with yellow dashed lines. (b) Quantification of DNA content of clone cells by measuring the integrated DAPI intensity per nuclues for a. 80 GFP+ nuclei from clone in 10 midguts per genotype were counted. (c) Quantification of clone areas. Error bars represent standard deviations and the Student's T-test was used to determine statistical significance. (****: p<0.0001).Ras V12S35 overexpression enabled midgut cells to divide, grow, and endoreplicate in the absence of TOR activity. This experiment was repeated three times.

Supplementary Figure 6 | Deletion either E2f1 or Myc in midgut cell clones blocks both cell growth and endoreplication
Myc 3 or E2f1 7172 null mutant clones were generated in the midgut using the MARCM technique. 7 days after clone induction, midguts were dissected, fixed and stained for DNA (blue) with DAPI. GFP (green) marks mutant cells. E2f1 samples were stained with PDM1 antibody to detect EC differentiation. (a) FRT control and Myc mutant clones. Myc -/clones arrest as single small diploid cells. (b) E2f1 mutant clones. Some of E2f1 clone cells are PDM1-positive and therefore are differentiated ECs, despite their very small size. All experiments were repeated three times.

Supplementary Figure 7 | Ras/Raf signaling induces E2f1 transcriptional activity
The hs-Flp Act>CD2>Gal4 driver ("flip-out Gal4") was used to force widespread expression of Ras V12S35 or Raf GOF in flies carrying the PCNA-GFP reporter, which is E2f1 responsive, or the PCNA-GFP E2f1 reporter, which lacks E2f1 binding sites in its promoter and is not E2f1-responsive. 3 days after clone induction, midguts were dissected, fixed, and analyzed. The midguts were labeled with DAPI to visualize DNA (blue), anti-CD2 antibody to indicate non-clone cells (red) and anti-GFP (green) to detect PCNA-GFP reporter expression. Overexpression of Ras V12S35 or Raf GOF induced PCNA-GFP reporter transcription, indicating stimulation of E2f1 transcriptional activity. Consistent with this conclusion, transcription failed to be induced in flies with the E2f1 insensitive reporter, PCNA-GFP E2f1 . This experiment was repeated three times.