Accelerated cell cycles enable organ regeneration under developmental time constraints in the Drosophila hindgut

Summary

Individual organ development must be temporally coordinated with development of the rest of the organism. As a result, cell division cycles in a developing organ occur on a relatively fixed time scale. Despite this, many developing organs can regenerate cells lost to injury. How organs regenerate within the time constraints of organism development remains unclear. Here, we show the developing Drosophila hindgut regenerates by accelerating the mitotic cell cycle. This process is achieved by decreasing G1 length and requires the JAK/STAT ligand Unpaired-3. Mitotic capacity is then terminated by the steroid hormone ecdysone receptor and the Sox transcription factor Dichaete. These two factors converge on regulation of a hindgut-specific enhancer of fizzy-related, a negative regulator of mitotic cyclins. Our findings reveal how the cell cycle machinery and cytokine signaling can be adapted to accomplish developmental organ regeneration.

Graphical Abstract

eTOC blurb

In response to organ injury during development, an organism may pause overall growth to enable repair. Cohen et al. find a different type of response in fly larvae, where the injured organ (fly large intestine) speeds up cell production to regenerate the injured tissue within the normal developmental time window.

Introduction

Development is subject to exquisite temporal regulation (Nüsslein-Volhard and Wieschaus, 1980, Lagha et al., 2012). This control synchronizes development of the organ with development of the organism. For example, in Drosophila, systemic hormone signals coordinate growth of specific tissues with the organism-wide progression through metamorphosis (Karim and Thummel 1992; Riddiford et al. 2000; Kozlova and Thummel 2003).

In spite of this tight temporal regulation, many developing tissues have a striking capacity to regenerate after injury. Examples of such developmental regeneration in animals include Xenopus tadpole tails, Drosophila imaginal discs, and the myocardium and digit tips of mammals (Halme et al., 2010; Illingworth, 1974; Porrello et al., 2011; Slack et al., 2004; Smith-Bolton et al., 2009). If developing tissues are unable to repair an injury, long-term abnormalities may arise. For example, pediatric traumatic brain injury causes neural abnormalities (Imms et al., 2019; Taylor et al., 2017) and reduced brain function (Ganesalingam et al., 2011; Lindsey et al., 2019). In bone, deformities occur after unresolved pediatric facial fractures (<5 years; Singh and Bartlett, 2004; Wheeler and Phillips, 2011). Thus, while development is subject to time constraints, the capacity for organ repair and regeneration plays important roles across evolution.

In Drosophila, one developmental injury repair paradigm involves extending the time needed to complete organism development. Drosophila larval imaginal discs produce many adult organs. Imaginal cells undergo compensatory divisions to regenerate disc-derived tissues after injury (Hariharan and Serras 2017). In the wing disc, these extra regenerative divisions do not occur within normal organismal developmental timing. Rather, wing injury at 2^nd and early 3^rd instar larval stages (hereafter L2 and L3) releases systemic cues that activate a regeneration checkpoint. This checkpoint delays animal development, allowing the wing extra time to repair (Halme et al., 2010; Hussey et al., 1927; Simpson et al., 1980; Smith-Bolton et al., 2009). Once development passes the period of checkpoint activation, the wing disc can no longer regenerate (Halme et al., 2010; Smith-Bolton et al., 2009). Currently, existing models have not addressed how organs complete tissue repair within developmental time constraints.

The Drosophila hindgut contains three main segments- the pylorus, ileum, and rectum. These segments are present in both the larval and adult gut. During metamorphosis, the normally developing hindgut undergoes a regeneration of sorts. As documented 85 years ago (Robertson, 1936), the larval ileum undergoes histolysis, and is replaced by regenerative activity from adult hindgut precursors. Cell division from the larval pylorus, which occurs during metamorphosis, expands the adult pylorus and generates the adult ileum (Cohen et al., 2020; Fox and Spradling, 2009; Sawyer et al., 2017; Takashima et al., 2008; Yang and Deng, 2018). This development is not disrupted by severe, acute injury to the wandering 3^rd instar larval (hereafter L3W) pylorus. Our previous lineage labeling suggested that pyloric cells remaining after injury undergo additional cell divisions to regenerate the adult hindgut (Cohen et al., 2018). Nevertheless, the capacity of the pylorus to respond to injury by regenerative cell divisions is terminated before adulthood. Unlike the larval pylorus, adult pylorus injury results in endocycles, where cells replicate DNA without mitosis, leading the pylorus to regenerate through polyploidization and hypertrophy (Cohen et al., 2018; Fox and Duronio, 2013; Fox and Spradling, 2009; Losick et al., 2013; Sawyer et al., 2017). This injury response switch from larval mitosis to adult endocycles requires the Anaphase Promoting Complex/Cyclosome (APC/C) activator Fizzy-related (Fzr), also known as Cdh1 (Cohen et al., 2018). Such Fzr-dependent termination of injury-induced mitosis in the hindgut suggests that progression through development may limit the mitotic capacity of this tissue.

Here, we report that unlike the wing disc, the hindgut can regenerate without delaying animal development. Instead, hindgut regeneration after late-stage larval injury occurs by accelerating the mitotic cell cycle. This enables the injured hindgut to regenerate within the normal developmental time frame. Cell cycle acceleration occurs by shortening G1 phase length. Unpaired3 (Upd3), a JAK/STAT pathway cytokine, is necessary and sufficient for accelerated injury cell cycles. We further find that fzr expression is sufficient to terminate mitotic activity in the pylorus, and this termination depends on developmental activation of a fzr enhancer. This enhancer is activated by the receptor for the steroid hormone Ecdysone. This fzr enhancer contains binding sites for the Sox transcription factor Dichaete, which (like fzr) terminates injury-induced hindgut mitosis. Our findings reveal how developmental and injury signaling converge to define a developmental time window when accelerated mitotic cycles regenerate an injured organ.

Results

The Drosophila hindgut can regenerate without delaying metamorphosis

To understand how mitotic organ regeneration is accomplished in the hindgut despite time constraints imposed by development (metamorphosis), we considered two possible models. In one model, hindgut injury delays organism-wide development (Fig1A, Model1) until injury is repaired, allowing time for additional regenerative cell cycles. In a second alternative model, injury does not delay development (Fig1A, Model2). Instead, additional compensatory mitotic cell cycles occur during normal developmental time. To distinguish between these models, we acutely injured the larval hindgut and assessed whole-animal development progression (Fig1A’ “18C”, Methods). For tissue injury, we used a temporally and spatially regulated system, where Gal4 is inhibited by temperature sensitive (ts) Gal80 at the permissive temperature (18C). Upon shifting to 29C, Gal4 is expressed in the hindgut by the brachyenteron (byn) promoter to induce the UAS-apoptotic genes head involution defective (hid) and reaper (rpr) (Fox and Spradling, 2009). We induced injury at L2, L3 (Fig1A’, green “29C”), and L3W stages (Fig1A’, yellow “29C”). To assess if whole animal developmental progression timing changes after injury, we measured pupation onset timing in animals +/− hid and rpr transgenes.

Figure1. The Drosophila hindgut can regenerate without delaying metamorphosis.

(A) Schematic illustrating two models for regeneration under developmental time constraint. Lightning bolts=injury. (A’) Injury timing and methods used. Upper panel- injury using Gal80ts. Lower panel- injury using DEMISE. HS= heat shock, C=Celsius, L=larval instar, W=wandering. (B) Time from 3^rd instar larva to pupation onset in L2-early L3 hindguts +/− injury. Data represent mean ± SEM, N≥20 animals/condition, at least two replicates. Two-way ANOVA (interaction p<0.05), Sidak’s multiple comparisons. (C-D) Time from L3W, +/− injury to (C) pupation onset or (D) adult stage. Data represent mean ± SEM, N≥20 animals per condition, at least two replicates. Two-way ANOVA (not significant). (E-F) Mitotic figures 24hpi in the (E) uninjured pylorus or (F) pylorus injured at L3W. Phospho-HistoneH3 (green), and nuclei (DAPI, magenta). Yellow hashed line=developing pylorus. (G) Quantification of pyloric mitotic index (MI) +/− L3W injury. Data represent mean ± SEM, N≥6 animals per condition, and visualized as symbols, at least two replicates. Unpaired two-tailed t-tests. (H) Cell number in developing pylori +/− L3W injury. Data represent mean ± SEM, N≥6 animals per condition, at least two replicates. Unpaired two-tailed t-tests reported in graph. Analysis of recovery slope using ANCOVA- p<0.05 for 24–72hr slope. Scale bars 20μm. Asterisk number= p value (see statistical analysis in Methods).

We first injured the hindgut at L2-early L3, as previous experiments found a whole-organism developmental delay when the wing is injured at this stage. Similar to the wing, L2-early L3 hindgut acute injury significantly delays development by ~24hours (h, Fig1B, FigS1A–C), similar to the rpr-induced injury delay in L2-early L3 wing discs (Halme et al., 2010). Supporting this finding, L2-early L3 hindgut injury leads to expression of DILP 8 (Drosophila Insulin-Like Peptide 8, FigS1D–E) a known regulator of organism-wide developmental delay after imaginal disc injury (Colombani et al., 2012; DaCrema et al., 2021; Garelli et al., 2012). As with our previous experiments with L3W injury, L2-early L3 hindgut injury leads to wholescale adult hindgut regeneration (FigS1C). These findings are consistent with a model where L2/early L3 hindgut injury acts similarly to wing injury at this stage, causing DILP8 release and organism-wide animal development delay.

We previously found that, unlike in the wing, the hindgut regenerates when injury is induced later, at L3W (Cohen et al., 2018). We therefore tested if L3W hindgut injury delays organismal development. In contrast to L2-early L3 injury, L3W hindgut injury does not delay pupation onset or cause DILP8 release (Fig1C, FigS1F). However, it remained possible that L3W hindgut injury delays progression through pupation. However, we find there is no delay in fly eclosion after L3W hindgut injury (Fig1D). These results indicate that hindgut injury at different larval stages results in different responses. L2-early L3 hindgut injury delays animal development, while L3W injury leads to hindgut regeneration in concert with development.

We next looked at the cellular level to determine how the injured L3W hindgut fully regenerates without delaying development. As the pylorus is the site of hindgut regenerative activity, we closely examined this region. We assayed cell death/number and the frequency of cells with the mitotic marker Phospho-Histone H3 (PH3, Methods) during metamorphosis in animals injured at L3W. To dually control injury induction and transgene expression, we used our previously established Dual-Expression-Method-for-Induced-Site-specific-Eradication (DEMISE) system (Cohen et al., 2018; Fox et al., 2020). DEMISE uses heat shock (HS)-induced Flippase (FLP) to excise a stop cassette in a UAS-rpr cell death transgene. This allows for site specific mosaic injury in any tissue. DEMISE injury is controlled by HS, while Gal4/UAS transgene activation is again controlled by temperature-sensitive Gal80. L3W animals were raised at 29C, heat shocked at 37C to injure (Fig1A’ “29C DEMISE”, “HS”), then shifted to 29C to recover and (when applicable) express other transgenes.

Using our L3W injury protocol, at 16 hours post-injury (hpi), we observe widespread pyknotic nuclei and reduced pyloric cell number (FigS1G–G’, Fig1G, 16hpi). During the 16h required to induce apoptosis and cell death, we see no increase in pyloric mitosis. This suggests that inducing rpr in the hindgut does not immediately lead to mitosis. However, at 24hpi, injured animals have an accelerated rate of cell number increase and a higher mitotic index (Fig1E–G). The developmental period of hindgut mitosis is not impacted by injury (occurring 72–96h after L3W, Fig1G, FigS1H–J), and hindgut cell number is restored in the normal developmental timeframe (Fig1H, FigS1I–J).

We next identified the number of additional mitotic cell cycles that the L3W pylorus undergoes after injury to regenerate the hindgut. Without injury, our previous lineage tracing showed that L3W pyloric cells which produce the adult pylorus undergo ~2–3 divisions in the first 48h of metamorphosis (yielding 5–6 cell clones in adults), while pyloric cells that generate the adult ileum undergo 1 division to make the adult ileum (yielding 2 cell clones in adults, Cohen et al., 2018; Fox and Spradling, 2009; Sawyer et al., 2017). After L3W injury, our lineage tracing indicated a 3-fold increase in adult pyloric clone size (~16 cell clones) and a doubling of ileal clone size (4 cell clones, Cohen et al., 2018). Given our prior counts of adult pyloric and ileal cells (Cohen et al., 2018; Fox and Spradling, 2009; Sawyer et al., 2017) and our lineage observations, we conclude that our injury protocol causes adult hindgut progenitors, on average, to undergo 2–3 additional mitotic cycles. Consistent with this, we see a 2–3x fold increase in mitotic index at 24hpi (Fig1G, 2.8±0.4% vs 8.1±1.6%). This indicates that additional divisions occur during the time of developmental hindgut divisions. Given that more pyloric cells are produced in the same amount of time after injury, our data argue that the increased mitotic index does not simply reflect prolonged M-phase. Additionally, by revisiting our previous lineage tracing data (Cohen et al., 2018), we ruled out the model that injury activates a dormant, injury-responsive pool of mitotic pyloric cells. Specifically, we find no change in clone number after injury (FigS1K). Taken together, our results pinpoint a timeframe in late larval and pupal development when pyloric cells undergo additional divisions without delaying animal development to compensate for injury.

A shortened G1 phase accelerates regenerative mitotic cell cycles during hindgut metamorphosis

We next examined if any cell cycle phase is shortened to facilitate extra divisions during hindgut developmental regeneration. We assessed the cell cycle using Fly-FUCCI, where two markers (GFP-E2F1_1–230 and RFP-CycB_1–266) distinguish G1, S and G2/M cells (Zielke et al., 2014), and used DEMISE to induce injury (Fig2A, Methods). Upon injury, the % of G1 pyloric cells at both 16 and 24hpi decreases (fewer GFP-E2F1_1–230 +, RFP-CycB_1–266 - cells, Fig2B–F, yellow dotted lines). By completion of hindgut development (as determined in Fig1G–H and FigS1H), all cells are in G1/G0 regardless of injury (Fig2F). Given the persistence of GFP-E2F1_1–230 in nearly all pyloric cells, we did not distinguish between S and G2. Compared to uninjured animals, injured pylori increase in S/G2 cells immediately after injury (Fig2G, 16hpi). Our results indicate that injury causes a doubling in pyloric cells leaving G1 to enter S/G2 (Fig2F–G, 16hpi), followed by a doubling in mitotic index soon after (Fig1E–G, 24hpi). Therefore, by shortening G1 after L3W injury, the pyloric cell cycle accelerates, enabling compensatory cell divisions within normal developmental time constraints.

Figure2. A shortened G1 phase accelerates regenerative mitotic cell cycles during hindgut metamorphosis.

(A) Schematic illustrating the combination of DEMISE and Fly-FUCCI after injury. X= cell death. (B-E’) Insets of 16hpi (B-C’) or 24hpi (D-E’) developing pyloric cells +/− L3W injury . E2F1.1–230 (green), CycB.1–266 (magenta/white). Yellow hashed outline=G1 cells. (F-G) Quantification of (F) G1 and (G) S/G2 cells at different developmental times +/− L3W injury. Data represent mean ± SEM, N≥6 animals per condition, visualized as symbols, at least two replicates. Unpaired two-tailed t-tests. (H) Schematic of experiment to prolong G1/S with non-degradable dacapo (dap^CRL4). (I-J’) Pylorus of adults expressing hindgut specific dap^CRL4 , (I-I’) no injury or (J-J’) L3W injury. Nuclei (DAPI, white) are marked. Yellow box= location of inset in following panel. (K) Quantification of adult pyloric cell number after dap^CRL4 expression +/− L3W injury. N=7 animals per condition, two replicates, visualized as symbols. Data represent mean ± SEM. ANOVA, Tukey multiple comparisons test. Scale bars (B-E’, I’, J’) 10μm, (I,J) 50μm. Asterisk number= p value (see statistical analysis in Methods).

To further explore if accelerated cell cycling in the injured L3W pylorus may involve G1/S regulation, we manipulated the G1/S regulator Dacapo (Dap). During the cell cycle, Dap inhibits the formation of cyclin-E/cdk2 complex. This cyclin-E/cdk2 inhibition negatively regulates G1/ S-phase entry (Lane et al., 1996, 2000; de Nooij et al., 1996, 2000; Secombe et al., 1998). At G1/S, Dap is degraded by the ubiquitin ligase CRL4^CDT2 (Bivik Stadler et al., 2019; Higa et al., 2006; Swanson et al., 2015). To experimentally prolong the G1/S transition, we ectopically expressed a dap transgene lacking the CRL4^CDT2recognition site (dap^CRL4) during metamorphosis (Fig2H). This non-degradable Dap was previously established to delay G1/S (Bivik Stadler et al., 2019). Without injury, dap^CRL4 hindguts do not have any phenotype or cell number loss (Fig2I–I’). This suggests that cells of the uninjured developing hindgut have sufficient time to undergo mitotic cycles even when the G1/S transition is experimentally prolonged. In contrast, dap^CRL4 animals injured at L3W have severely reduced adult pyloric cell number (Fig2J–K). This observation further supports a model whereby G1 length control enables larval/early pupal pyloric cells to accelerate the mitotic cell cycle to accomplish organ regeneration.

Hindgut regeneration during metamorphosis requires the JAK/STAT cytokine Upd3

In addition to injury-specific cell cycle regulation, we postulated that injury-specific cell signaling also plays a role in developmental hindgut regeneration. Our previous work found the JAK/STAT pathway ligand Unpaired3 (Upd3, IL6-like) is expressed by injured hindgut pyloric cells, and is necessary for adult injury endocycles (Sawyer et al., 2017). Further, Upd3 plays a role in injured Drosophila hemocytes (Pastor-Pareja et al., 2008) and adult midgut intestinal stem cells (Biteau et al., 2008; Jiang et al., 2009; Osman et al., 2012). To test if, as in adult hindgut injury, upd3 is induced during larval hindgut injury, we examined upd3-LacZ expression at L3W. Following injury, upd3-LacZ is elevated throughout the pylorus (Fig3A–C), especially in the anterior region that undergoes the most extensive apoptosis (Cohen et al., 2018; Fox and Spradling, 2009; Sawyer et al., 2017).

Figure3. Hindgut regeneration during metamorphosis requires the JAK/STAT cytokine Upd3.

(A-B) Expression of upd3-lacZ 16h following L3W, (A) no injury and (B) injury. (C) Line profile quantification of upd3-lacZ in L3W+16hr pylori. Each line represents an average of 10 animals per condition across three replicates. Transparent line= SEM. (D-E) Mitotic figures 24hpi of L3W pylorus in DEMISE expressing flies in (D) wild-type and (E) upd3 heterozygous animals. Phospho-HistoneH3 (green), and nuclei (DAPI, white). Yellow hashed outline= developing pylorus. (F) Quantification of whole hindgut (pylorus + ileum) mitotic index 24h following L3W injury in wild-type or upd3^Δ genetic background. Wild-type data is duplicated from Fig1G. Data represent mean ± SEM, N≥8 animals per condition across 3 replicates, visualized as symbols. (G-J) Hindgut morphology in adults. (G) uninjured wild-type, (H) uninjured hemizygous upd3^Δ and (I,J) upd3^Δ/+ animals recovered from L3W injury. Nuclei (DAPI, white) and false coloring show regions of the adult hindgut: Pylorus (green) and Ileum (cyan). Pairwise stitching was performed to show the entire hindgut. (K) Quantification of adult hindgut (pylorus and ileal) cell number after byn-Gal4>UAS-hid, UAS-rpr injury to the L3W pylorus in wild-type and upd3 mutants. Data represent mean ± SEM, N≥7 animals per condition, visualized as symbols, at least two replicates. ANOVA, Tukey multiple comparisons test. Scale bars (A-B) 10μm, (D-E) 20μm, (G-J) 50μm. Asterisk number= p value (see statistical analysis in Methods).

Next, we tested if upd3 is required for pyloric cells to increase in mitotic index 24h after L3W injury. Unlike in wild-type injured animals, the mitotic index of injured heterozygous null mutant upd3^Δ/+ females does not increase after injury and resembles wild-type, uninjured animals (Fig3D–F). upd3 is on the X-chromosome (Wieschaus et al., 1984), and uninjured upd3^Δ hemizygous animals (hereafter upd3^Δ) are viable and fertile with no obvious developmental defects, including in the hindgut (Osman et al., 2012; Sawyer et al., 2017) (Fig3G,H). We tested the requirement of upd3 in hindgut development after injury by assessing adult hindgut morphology (3–7 days post eclosion, Methods). In contrast to the recovery observed after L3W injury in wild-type animals (Fig1H, FigS1J), acute L3W injury in upd3^Δ animals leads to severe shortening, cell loss and/or complete abolishment of the adult hindgut (Fig3I,J).100% of injured upd3^Δ animals die <4 days following eclosion. Notably, injury-specific hindgut abnormalities and lethality also occur in upd3^Δ/+ females (Fig3I–K), suggesting that correct upd3 levels are essential for the L3W injury response. Together, our results support a role for Upd3 in the L3W injured hindgut to signal additional, injury-mediated cell divisions within developmental time constraints.

Fizzy-related activation by Ecdysone Receptor terminates the period of regenerative mitotic cycles

Having identified Upd3 as regulating regenerative pupal hindgut mitotic cycles, we next sought to find regulation that terminates this finite response. We previously found an important role for the endocycle regulator Fizzy-related (Fzr) in switching the hindgut injury response from mitotic cycles to endocycles. Fzr activates the Anaphase Promoting Complex/Cyclosome (APC/C) and plays a conserved role to degrade mitotic cyclins (Sigrist and Lehner, 1997; Zachariae et al., 1998). In the injured hindgut, Fzr is required to terminate mitotic cycles in favor of endocycles (Cohen et al., 2018). We next tested if fzr is sufficient to terminate the accelerated mitotic cycles in the injured L3W pylorus. While high levels of UAS-fzr expression (29C) are detrimental to hindgut development, we observe no developmental phenotypes upon mild UAS-fzr expression (25C) without injury (Fig4A). Upon injury, L3W injured pyloric cells expressing UAS-fzr (25C) reenter the cell cycle and become polyploid, as assayed by reduced cell number and a significant increase in nuclear area (Fig4B–D). We then compared this mitosis termination function of Fzr to that of Upd3. In contrast to our results with UAS-fzr, expressing UAS-upd3 starting at L3W (Fig4E–H) increases pyloric tissue area (Fig4G) due to a large cell number increase (Fig4H). These results highlight a role for Fzr in terminating mitotic cycles, whereas Upd3 is a permissive signal for either mitotic cycles or endocycles (see Discussion). In agreement, expressing UAS-upd3 in the adult hindgut (using adult-onset 13e02>Gal4) causes endocycles in the pylorus, observed through increased nuclear size but not increased cell number (FigS2A–E).

Figure4. Fizzy-related is sufficient to terminate regenerative mitotic cycles.

UAS-fzr expression after (A-A’) no injury or (B-B’) L3W injury. Nuclei (DAPI, white) are marked. Yellow box= inset location in following panel. (C-D) Quantification of adult (C) pyloric cell number and (D) nuclear area following UAS-fzr expression +/−L3W injury. N≥7 animals per condition across two replicates, visualized as symbols in graph. Data represent mean ± SEM. ANOVA, Tukey multiple comparisons test. (E-F) Pylorus of adults expressing hindgut specific (E) UAS-GFP or (F) UAS-upd3 during the first 48h of morphogenesis. Nuclei (DAPI, white). (G-H) Quantification of adult pylori (G) tissue area and (H) cell number following expression of UAS-GFP or UAS-upd3 during the first 48h of pupation. Data represent mean ± SEM, N=7 animals per condition, visualized as symbols, at least two replicates. Unpaired two-tailed t-test. Scale bars (A,B,E,F) 50μm, (A’,B’,) 10μm. Asterisk number= p value (see statistical analysis in Methods).

To next pinpoint when APC/C^Fzr terminates injury-mediated pyloric mitoses (Fig5A), we injured animals at several time points beginning at L3W and examined recovered adults. To increase temporal resolution, we raised and recovered animals at 18C, approximately doubling developmental time from 25C (Methods). To distinguish between mitotic cycles and endocycles, we measured cell number of recovered adult pylori after injury at 6 distinct times. Prior to mid-pupation, compensatory pyloric cell mitosis restores adult hindgut cell number (Fig5B–C, E). The ability to undergo regenerative divisions before mid-pupation is consistent with our cell cycle marker analysis (Fig1–2). Beginning at mid-pupation, cell number is abruptly no longer restored after injury and recovered nuclei are larger, indicating a switch to endocycles (Fig5D–E). We note that the end of injury-mediated mitotic cycles corresponds with the normal termination of mitotic cell cycles, at approximately 96h after pupa formation (APF) at 18C (FigS1H). To verify that the change in injury response coincides with the onset of compensatory endocycles, we measured cell nuclear area, an indicator of ploidy, before and after mid-pupation. Nuclear area doubles when the hindgut is injured after mid-pupation (FigS3A–B). Together, our data support the existence of a developmental window beginning at L3W and ceasing 72–96 hrs after (at 18C) in which pyloric cells can undergo rapid regenerative mitotic cycles.

Figure5. Fizzy-related enhancer activation by Ecdysone Receptor during termination of regenerative mitotic cycles.

(A) Schematic of the injury response APC/C^fzr-dependent switch from mitosis to endocycles. Ecdysone peaks shown. (B-D) Adult pyloric nuclei following (B) no injury, (C) injury induced 72h after pupation or (D) injury induced at 120h after pupation. Developmental stages represent time after pupation at 18C. Nuclei (DAPI, white). (E) Quantification of adult pyloric cell number in animals recovered from injury at different stages (=recovered adult cell number). Stages represent time after pupa formation (APF) at 18C, WPP= white pre-pupa, indicative of pupation onset. Data represent mean ± SEM, N≥6 animals per condition, visualized as symbols. At least two replicates. ANOVA, Dunnett multiple comparisons test against uninjured controls (F) Relative motif scores and motif location in fzr enhancer region as obtained from JASPAR core. Motifs were ranked with respect to the strength of their prediction score, with the strongest scoring motif ranked highest (1, 0.95) . The motif rank is then used to identify the motifs throughout (G) modENCODE data identifying whole-animal EcR binding to fzr via ChIP-seq data. LacZ represents the location of the fzr ^G0418 trap. TSS= transcriptional start site, ATG= translational start site. Thin lines= first and second introns, thick bars= exons (blue exon is retained in the coding sequence). Genomic locations are reported according to the modENCODE Browser. EcR motif locations are indicated and numbered according to the rank in panel F. (H) Schematic of experimental design to induce hindgut-specific, EcR.DN clones throughout metamorphosis. (I-J’) Expression of the fzr ^G0418 enhancer trap in adult pylori of (I-I’) GFP+ only clones or (J-J’) GFP+ EcR.DN clones. Nuclei (DAPI, magenta), clones (GFP, green), fzr ^G0418 (anti-Beta-Galactosidase, white). (K) Quantification of % fzr ^G0418 positive cells in adult clones expressing GFP or EcR-DN in anterior or posterior pylorus. Data represent mean ± SEM, N≥9 individual clones per condition, visualized as symbols, at least two replicates. ANOVA, Tukey multiple comparisons test. Scale bars (B-D) 10μm, (I-J’) 20μm. Asterisk number= p value (see statistical analysis in Methods).

Our findings here (Fig4A–D) and our previous study shows that Fzr is necessary and sufficient to terminate mitosis in favor of endocycles in the pylorus. Further, we find that fzr expression is detectable in the adult (endocycle-prone), but not in the larval (mitosis-prone) pylorus (Cohen et al., 2018). We thus hypothesized that upstream fzr activation in the pylorus terminates rapid regenerative mitotic cycles. One such upstream candidate is the ecdysone steroid hormone. Ecdysone regulates developmental timing and progression (Riddiford et al., 2000), and Ecdysone level peaks (Fig5A) regulate mid-pupal developmental events. Recent studies in wing discs found that signaling through the Ecdysone Receptor (EcR) blocks regenerative capacity, after a prepupal Ecdysone pulse (Halme et al., 2010; Harris et al., 2016; Narbonne-Reveau and Maurange, 2019). However, unlike wing discs, we demonstrated that the pylorus continues to respond to injury after the prepupal ecdysone pulse, but terminates mitosis in favor of fzr-dependent endocycles (Cohen et al., 2018). We therefore explored if ecdysone signaling activates fzr in the hindgut, which would tie an animal developmental timing signal to terminating rapid injury mitotic cell cycles.

We first identified a candidate fzr enhancer region based on the location of a lacZ enhancer trap in the first intron of the fzr 5’ untranslated region. This trap is highly expressed in adult pylori (fzr^G0418, Cohen et al. 2018). We first computationally examined if ecdysone might activate fzr in the pylorus during development. We used JASPAR 2020 Scan motif analysis (Fornes et al., 2020), which compares available Insecta motif matrices in the JASPAR core to a given DNA sequence and returns transcription factor binding sites, as well as a relative prediction strength score. By querying a candidate fzr enhancer region surrounding the fzr^G0418 insertion, we found several strong EcR binding motifs (Fig5F, Methods). By analyzing modENCODE whole-animal ChIP-seq data (Celniker et al., 2009), we found that EcR directly binds the fzr enhancer region at mid-pupation, during the termination of rapid regenerative mitotic cycles (Fig5E 72APF, Fig5G, modEncode dataset 2642).

We next assessed the effect of ecdysone signaling on hindgut development and fzr expression. Flies expressing UAS-dominant-negative EcR (EcR.DN) throughout the developing hindgut die soon after eclosion due to several defects including a lack of larval ileum histolysis (FigS3C, D). Consequently, to ask if ecdysone signaling is required for fzr enhancer expression in the adult pylorus, we induced GFP-marked hindgut-specific UAS-EcR.DN clones (Fig5H, Methods). Clones were induced such that L3W heat shock removes the Gal80 repressor in a mosaic subset of cells. Flies with EcR.DN clones show no morphological abnormalities and are viable. We assayed fzr^G0418 expression in adult animals containing either GFP+ EcR.DN clones or GFP-only control clones (Fig5I–K). Control clones express fzr^G0418 without injury (Fig5I–I’), but clones expressing EcR.DN show a strong reduction in fzr^G0418 expression (Fig5J–J’). The loss of fzr ^G0418 expression is especially strong in EcR.DN clones of the anterior pylorus (Fig5K), whereas the directionality of pyloric divisions (anterior-posterior) may cause posterior pyloric cells to express UAS-EcR.DN for less time. Alternatively, ecdysone may primarily act in anterior pyloric cells. Together, our results suggest ecdysone signaling activates fzr in the developing hindgut, which links steroid signaling to termination of injury-mediated mitosis.

Dichaete is required to terminate regenerative mitotic cycles

To further dissect a role for chromatin-level regulation in terminating injury-responsive mitosis, we closely examined fzr regulatory sequences. We sought to find regions of the candidate fzr enhancer near the EcR binding sites that reproduce the temporal and spatial hindgut expression of fzr ^G0418. By analyzing modENCODE data, we confirmed that the region surrounding the EcR sites contains the active enhancer mark H3K27ac in adult males and females (Fig6A, modEncode datasets 844–845). To identify a specific fzr enhancer that might be responsible for hindgut activity, we cloned multiple fzr sequence fragments flanking various combinations of the five top EcR binding sites (Fig6A, Fig5F–G, Methods). Fragments were inserted in front of mCherry (pHPdestmCherry vector) and we derived transgenic lines. We note that flies expressing a fragment containing EcR motif sites 2–4 alone were not viable (>400 embryos injected) and die as embryos, whereas all other tested combinations (Fig6A) are viable. We assessed fragment activity in larval and adult guts (Methods). Of three tested constructs, we found one (fzr.B) that matches fzr^G0418 expression (Cohen et al., 2018), showing a pattern specific to the adult hindgut but not the larval hindgut (Fig6B–C’), and therefore behaves as an enhancer.

Figure6. Dichaete is required to terminate regenerative mitotic cycles.

(A) Schematic of constructs. Full enhancer region (fzr.full, 3kb). 1.5kb fragments of this enhancer were isolated and cloned in front of mCherry sequence. Fragments were randomly integrated into the Drosophila genome and N≥2 lines of each fragment were assessed for expression. TSS= transcriptional start site. EcR motifs are numbered as in Fig5F. (B-C’) Expression of fzr.B in (B-B’) the L3W pylorus and (C-C’) the adult pylorus. Nuclei (DAPI, white), mCherry (magenta or green). (D) JASPAR core motif analysis of non-overlapping region of fzr.B vs fzr.A identifies multiple potential binding sites for a Dichaete motif. Motifs were ranked in respect to strength of prediction score, with the strongest scoring motif ranked highest (1, 0.95). The motif rank is then used to identify the motifs throughout. (E-J) Adult pyloric nuclei, (DAPI-white) recovered after (E) no injury, (F) L3W injury (G) adult injury (H) adult injury in fzr RNAi (I-J) adult injury in Dichaete RNAi. (K) Quantification of recovered cell number after adult pyloric injury +/− Dichaete RNAi. Data represent mean ± SEM, N≥6 animals per condition, visualized as symbols, at least two replicates. ANOVA, Tukey multiple comparisons test. Scale bars (B-C’) 20μm, (E-J) 10μm. Asterisk number=p value (see statistical analysis section in Methods).

fzr.B contains the highest ranked EcR motif site (Fig6A site #1, Fig5F–G), adjacent to both the lacZ trap and the modENCODE pupal EcR ChIP-Seq peak. However, all three tested constructs (Fig6A) contain the highest ranked EcR binding site, yet not all express in the hindgut (Fig6B, FigS4A–D’). Rather, fzr.Full and fzr.A fragments express in different spatial patterns, in the Malpighian tubules and midgut enterocytes respectively (FigS4A–D’). The differential construct expression suggests that a strong EcR binding site is not sufficient to reproduce fzr ^G0418 expression, or that sequences in fzr.Full and fzr.A but not in fzr.B include fzr repressors. These results align with previous observations in wing discs, where EcR is responsible for changing chromatin architecture rather than directly activating genes (Ma et al., 2019; Uyehara et al., 2017).

Given these findings, we next sought to find transcription factors that terminate injury-induced mitosis in the pylorus. To identify candidates, we analyzed the adult hindgut-expressed fzr enhancer fragment for transcription factor binding sites. We analyzed the unique 725bp region of fzr.B compared to the non-hindgut expressing fzr.A. Using JASPAR Insecta scan analysis, we found four strong motif sites (relative score>0.8) of the Sox-domain-containing transcription factor Dichaete (D, fish-hook, Fig6D, Methods). Dichaete is implicated in many processes including embryonic hindgut development Sánchez-Soriano and Russell, 2000), embryo segmentation, and differentiation (Ma et al., 1998; Russell et al., 1996; Zhao and Skeath, 2002). Additionally, Dichaete is expressed in all larval hindgut cells and multiple imaginal discs, suggesting a role at metamorphosis (Mukherjee et al., 2000). However, the role of Dichaete in tissue injury remains unexplored.

We next tested if Dichaete regulates hindgut injury responses. Hindgut-specific RNAi of Dichaete, expressed throughout metamorphosis, does not cause morphological defects without injury (assayed by hindgut-specific expression of the membrane marker UAS-Moesin-GFP FigS4E). Using DEMISE, we confirmed pyloric injury in Dichaete RNAi animals by DCP1 staining (FigS4F). We then allowed injured flies to recover and measured pyloric cell number as previously (Cohen et al., 2018) to assess if mitosis or endocycles occur. In contrast to the mitotic response following wild-type L3W injury, wild-type injured adults increase the size/ploidy of pyloric nuclei (Fig6E–G). fzr RNAi in the adult pylorus enables cell number restoration without ploidy increase after adult injury, as previously described (Fig6H, Cohen et al., 2018). Using two separate constructs, Dichaete RNAi also restores adult pylorus cell number after injury, without a noticeable nuclear size increase (Fig6I–K). These results implicate Dichaete as a developmental regulator that terminates injury-mediated mitotic cycles. Together, our findings here identify a finite developmental period when the mitotic cycle of hindgut pyloric cells is accelerated to accomplish regeneration and define molecular regulation that initiate and terminate this response.

Discussion

Accelerating the cell cycle as a mechanism to regenerate organs within developmental time constraints

The connection between organism development and the capacity to regenerate has been observed in diverse tissues and organisms (Poss 2010; Seifert and Voss 2013; Yun 2015). Here, we use the Drosophila hindgut as a model to understand how regenerative mitotic cell cycles are coordinated with tissue development. We show that a developmental delay is not essential for hindgut regenerative mitotic cell cycles. Rather, L3W pylorus injury shortens G1 phase to accelerate the mitotic cell cycle. This increases regenerative cell divisions within a developmental time constraint. We further identify regulation that terminates this mitotic regenerative capacity. Our work reveals that the steroid hormone ecdysone receptor and the Sox family transcription factor Dichaete switch the cellular injury response from relying on mitotic cycles to relying on endocycles. Our data reveal a mechanism by which flexible cell cycle dynamics during a succinct period of development enable mitosis-based organ regeneration (Fig7).

Figure7. Model: The hindgut pylorus undergoes distinct injury responses that rely on cytokine regulation of cell cycle.

Hindgut injury at L2–L3 results in release of Dilp8 and developmental delay. Hindgut injury between L3W and the onset of APC/C^fzr activity (red rectangle on timeline) leads pyloric cells to undergo up to 2 additional mitotic cycles in order to restore cell number within the same organismal developmental time frame. These additional mitotic cycles occur through decreased time spent in G1 and require Upd3 cytokine. Following termination of mitotic cycles by Ecdysone receptor-mediated APC/C^fzr activation, hindgut injury induces pyloric endocycles. Upd3 is necessary and sufficient to induce endocycles following APC/C^fzr activation, and endocycles also require the SOX transcription factor Dichaete.

The ability of the late larval/early pupal hindgut to undergo mitotic regeneration without delaying development is distinct from the Drosophila wing disc, which cannot regenerate after pupation onset (Halme et al., 2010). Our findings here reveal several properties of regeneration within a time constraint. First, we show that mitotic regeneration under time constraints can occur by cell cycle acceleration. Such acceleration in the late larval/early pupal hindgut does not require prolonging the period when cells undergo mitosis. Second, our findings reveal G1 flexibility G1 as a mechanism to accelerate the cell cycle for regeneration. Without injury, we find that overexpressing non-degradable Dap to prolong G1/S does not lead to developmental phenotypes, but after injury this same transgene blocks regeneration. This suggests that cell cycle flexibility enables more cell cycles during regeneration. Third, our data suggest that injury-esponsive JAK/STAT signals provide an extrinsic cue that possibly acts upstream of Dap to enable quick progression through G1/S. Such signaling is likely dose sensitive, as our findings using heterozygous upd3, as well as ectopic upd3, suggest cytokine levels dictate the correct degree of injury-induced mitotic cell cycle acceleration. We note that previous studies in the Drosophila eye and wing disc found that misexpression of the JAK/STAT cytokine upd also leads to cell cycle acceleration (Bach et al., 2003; Rodrigues et al., 2012). Our work identifies Upd3 as regulating not only mitosis after injury, but also endocycles, in a manner that depends on developmental stage (Fig7). This observation suggests that a core injury-responsive pathway operates in different modes of tissue recovery, both mitosis and endocycles.

Our findings here are likely conserved in other contexts. Accelerated mitotic cycles under stress conditions have been identified in other stem cell/regenerative contexts, including the axolotl spinal cord (Rost et al., 2016) and mammalian hematopoietic progenitors, where shortened G1 may also be involved (Guo et al., 2014; Mende et al., 2015). Future work in the genetically tractable pylorus can provide a model of conserved molecular mechanisms and cell cycle control after tissue injury. Interesting future questions include if cell cycle rate tunes to the number of cells needed to be replaced by an injury.

A role for hormonal and transcriptional regulation in terminating mitotic regeneration

Our results find both hormonal and transcriptional cues terminate mitotic regeneration in the hindgut. We previously found the APC/C^fzr negatively regulates injury responsive mitotic cycles in the pylorus (Cohen et al., 2018). In the adult pylorus, fzr primes the tissue towards an endocycle-based injury response (Cohen et al., 2018). We now find the ecdysone steroid hormone receptor activates a fzr enhancer to terminate injury-mediated mitosis. Regulation of this enhancer by EcR directly links systemic hormone factors to an injury-mediated mitosis-to-endocycle switch.

Our data supports a model where developmentally timed systemic signals impact tissue injury responses. Similar to the pylorus, ecdysone limits the regenerative capacity of the Drosophila wing disc (Narbonne-Reveau and Maurange, 2019). In response to injury, retinoic acid and insulin-like peptide 8 are released from the wing disc to delay ecdysone production and organism development (Colombani et al., 2012; Halme et al., 2010; Katsuyama et al., 2015). In mammals, circulating factors impact regeneration (Avci et al., 2012; Conboy et al., 2005; Elabd et al., 2014; Hirose et al., 2019; Rebo et al., 2016). Notably, oxytocin and thyroid hormones regulate regenerative capacity (Avci et al., 2012; Elabd et al., 2014; Hirose et al., 2019). Our model supports the importance of hormonal control of injury cell cycles.

Interestingly, modENCODE data from L3W and early pupation (datasets 2640 and 3398 respectively) suggests EcR binds the fzr enhancer before termination of pyloric mitotic cycles. We propose that EcR binding to a fzr enhancer acts as a permissive regulator of pyloric injury endocycles, possibly by modifying chromatin (Ma et al., 2019; Uyehara et al., 2017). We further propose that the Sox transcription factor Dichaete is a direct effector of the switch to endocycles, recruited to fzr after EcR. Sox transcription factors are involved in regeneration of the zebrafish spinal cord (Guo et al., 2011) and mouse neurons (Jing et al., 2012). Further study of the hindgut can determine the interaction between Sox transcription and steroid regulation in injury biology.

Our results here suggest chromatin accessibility in a fzr enhancer fragment regulates an important injury response. Tissue regenerative enhancer elements have been found in both vertebrates and invertebrates. In the Drosophila wing disc, a wingless injury-responsive enhancer fragment is required for regeneration (Harris et al., 2016; Smith-Bolton et al., 2009), and ecdysone-induced factors alter chromatin around cell cycle genes (Ma et al., 2019; Uyehara et al., 2017). Tissue regeneration enhancer elements (TREE) have also been found in zebrafish and acoels, emphasizing the evolutionary connection between enhancer activation and regeneration (Gehrke et al., 2019; Kang et al., 2016). Our findings here emphasize the role of chromatin regulation in injury-responsive cell cycle programming.

Conclusion

In summary, this study highlights the Drosophila pylorus as an accessible model for studying how mitotic regeneration is coordinated and achieved within developmental tissue programming. Future work in this system will illuminate how systemic signals and chromatin changes can regulate cell cycle dynamics to control tissue injury responses.

Limitations of study

The injury responses in this study are based on our experimental system, wherein we use rpr or rpr + hid to genetically ablate cells. It is possible that other responses occur in the hindgut in response to different injuries, including those that involve non-apoptotic cell death. The developmental timing information that we report is subject to temperature variation and may also be strain-dependent.

STAR Methods

Resource Availability

Lead Contact

Further information and requests for resources and reagents should be directed to and will be fulfilled by the Lead Contact, Dr. Don Fox (don.fox@duke.edu).

Materials availability

Flies and expression plasmids for fzr.Full>mCherry, fzr.A>mCherry and fzr.B>mCherry generated in this study are available upon request. The DEMISE component UAS-FRT-stop-FRT-rpr was previously generated in our lab and is available upon request (Cohen et al., 2018).

Data and code availability

No datasets or code were generated for this study. Original data by modEncode is publicly available at http://www.modencode.org/ (Celniker et al., 2009) and datasets used are referenced in text. Motif analysis was done using JASPAR Core Insecta Scan and is publicly available at http://jaspar.genereg.net/ (Fornes et al., 2020).

Experimental model and subject details

Fly stocks and genetics

Full genotypes are described at flybase.org. Flies were raised on standard Drosophila media (Archon Scientific, Durham) at 25C unless reported otherwise. All adult dissections were performed 4–7 days following eclosion. With the exception of upd3Δ hemizygous males, data were collected and quantified from female flies. upd3Δ homozygous virgin females were used for all crosses to avoid rescue by maternal load. The following publicly available stocks were used in the study and their Bloomington Drosophila Stock Center number (BDSC#) provided: fzr^G0418 (#BDSC 12297), hsFLP12;Sco/CyO (#BDSC 1929) , UAS-EcR.B1-ΔC655.F645A (#BDSC 6869), upd3Δ (#BDSC 55728), UAS-D.RNAi#1^TRiP.JF02115 (#BDSC 26217), UAS-D.RNAi#2 ^{TRiP.HMS01150} (#BDSC 34672), ptub-FRT-Gal80-FRT (#BDSC 38881), UAS-dap.CRL4 (#BDSC 83335). Additionally, the following flies were used in the study: byn>Gal4, UAS-hid, UAS-reaper (Cohen et al., 2018; Fox and Spradling, 2009; Sawyer et al., 2017).

Method details

All UAS transgenes were induced by byn>Gal4 unless indicated. Unless indicated, all injury protocols were performed as previously described using the DEMISE injury system (hsFLP;UAS-FRT-stop-FRT-reaper, Cohen et al. 2018). We provide brief information on how we injured here. For wandering L3 (L3W) injury, flies were kept at 29C throughout development and collected at the wandering L3 stage. Larvae were then subjected to a 35 min 37C heat shock, a sub-lethal dose of injury. Flies were then shifted back to 29C to allow continued expression of transgenes and apoptotic genes. UAS-fzr expressing flies were shifted to 25C following injury due to lethality at higher temperatures (Fig4A–D). Injury for the following experiments: developmental delay (Fig1B–D), pupal injury (Fig5E) and morphology following injury of upd3Δ animals (Fig3A–C,G–K) was performed by regulating UAS-hid and UAS-rpr expression using the Gal80^ts repressor as previously described (Cohen et al., 2018). Briefly, in these experiments, flies were kept at 18C until the desired developmental stage and shifted to 29C for 16h. Flies were then transferred to 18C for measuring development or adult dissection 4–7 days following eclosion. Adult DEMISE injury was performed as previously described (Cohen et al., 2018). In short, 4–5 days after eclosion adults were shifted to 37C for 3×40 minutes with 1 hour delay between pulses. flies were then shifted to 29C for 3–4 days to allow expression of byn>UAS-rpr transgene in stochastically flipped cells and hindgut recovery. Adults were then dissected, imaged and assayed for pyloric cell number.

Expression of UAS-upd3 at the L3W stage (Fig4C–H) was performed by crossing byn>Gal4,tub>Gal80^ts and UAS-upd3 flies at 18C (Gal80^ts permissive temperature). Flies were shifted to 29C for 48h (Gal80^ts restrictive temperature, early pupation), allowing site specific UAS-upd3 hindgut expression. Flies were then returned to 18C until dissection at day 4–5 post eclosion. Adult UAS-upd3 expression (FigS2A–E) was performed by crossing the adult pylorus-specific Gal4 (13e02>Gal4, Tub>Gal80) at 18C. Adult flies (4–5 days post eclosion) were shifted to 29C for 7 days and dissected for imaging. upd3-lacZ expression following injury (Fig3A–B) was assessed by crossing upd3-lacZ reporter flies to byn>Gal4,Gal80^ts UAS-hid, UAS-reaper flies at 18C. Flies were shifted for 16h to 29C at L3W stage and dissected immediately for imaging. upd3-LacZ expression was analyzed by averaging 300-line profiles per animal (length of 50um from hindgut-midgut boundary), normalized against background intensity, and reporting the combined average and SEM across all animals per condition (Fig3C).

Injury to Dilp8-GFP reporter flies (FigS1D–F) was performed as in upd3-LacZ experiments at either L2–L3 or L3W injury. As dilp8-GFP flies are heterozygous and do not maintain selectable balancer at the pupal stage, all animals were scored for dilp8-GFP expression and expected genotype ratios were assessed in eclosed progeny. Accordingly, our data observed expected mendelian ratios of dilp8-GFP expression in animals injured at L2–L3 (6/9), while none of the injured L3W animals showed dilp8-GFP expression(0/12). We note that we identified an impact on dilp8-GFP fly survival following injury induction at the L3W stage.

Ecdysone receptor dominant negative clones were induced by flipping out a ptub-FRT-Gal80-FRT repressive cassette (Bohm et al., 2010). Flies containing (hsFLP/fzr^G0418; EcR.DN; byn>Gal4, UAS-NLSGFP, ptub-FRT-Gal80-FRT) were kept at 29C and subjected to a 30-minute heat shock at 37C at the wandering L3 stage. Induction of FLP by heat shock leads to removal of Gal80 cassette and stochastic expression of transgene of interest by byn>Gal4 throughout metamorphosis. Flies were then shifted back to 29C and dissected 4–7 days post eclosion.

Enhancer fragment cloning and motif analysis

Identification of H3K27ac and ecdysone receptor potential binding sites was based on publicly available data from modENCODE (Celniker et al., 2009). Motif analysis was performed using the JASPAR core Scan analysis tool with an 80% cutoff threshold (Fornes et al., 2020). The top 5 ranking non-overlapping motifs are reported. Candidate fzr enhancer fragments were isolated by PCR from w¹¹¹⁸ flies, sequenced and compared to a reference genome. The fragments were then inserted into the Gateway entry vector pDONOR221 (ThermoFisher Scientific) using Invitrogen Gateway BP Clonase II Enzyme Mix. Fragments were then inserted into a publicly available destination vector pHPdestmCherry (Addgene #24567, Boy et al. 2010), using the Invitrogen LR Clonase Enzyme Mix (ThermoFisher Scientific). The final construct was integrated at random genomic sites, and two homozygous viable lines were selected per fragment. All flies and constructs are available upon request.

Staining and image analysis

For all experiments, dissection, fixation and staining protocols were performed as previously described (Cohen et al., 2018; Sawyer et al., 2017). Briefly, this involves dissection in PBS, paraformaldehyde fixation, and blocking and staining with normal goat serum along with Triton-X. Measurement of animal pyloric nuclear area represents the average of N≥30 cells per pylorus. The following antibodies were used in this study: Beta-Galactosidase (Abcam, ab9361, 1:1000), DCP1 (Cell Signaling, Asp261, 1:500), Phospho-Histone H3 (Cell Signaling, #9706, 1:1000). Secondary antibodies were Alexa Fluor dyes (Invitrogen, 1:500). Tissues were mounted in Vectashield (Vector Laboratories Inc.). Images were obtained with an upright Zeiss AxioImager with Apotome.2 processing, inverted Leica SP5 or Andor Dragonfly Spinning Disk Confocal. Image analysis was performed using ImageJ (Schneider et al., 2012), including adjusting brightness/contract, Z projections, cell counts, and integrated density quantification. Image stitching (Fig3, FigS3C–D) was performed using ImageJ grid/collection stitching plugin (Schneider et al., 2012).

Quantification and statistical analysis

Quantification procedures are indicated in the results text and figure legends. All graphed data includes biological replicates and a total N value (# animals) of at least 6. Statistical analysis was performed using GraphPad Prism 8.2.0. Statistical tests and n values are detailed in figure legends and visualized on graphs. For additional clarity and exact values, all statistical tests and n values are also reported in attached statistical reporting excel sheet. P and adjusted P value reporting is as follows: (p>0.05, not significant); (p<0.05,*); (p<0.01,**); (p<0.001,***); (p<0.0001, ****).

KEY RESOURCES TABLE

REAGENT or RESOURCE	SOURCE	IDENTIFIER
Antibodies
Beta-Galactosidase	Abcam,	Cat: ab9361
DCP1 (Asp261)	Cell Signaling	Cat: 9578
Phospho-Histone H3	Cell Signaling	Cat: #9706,
Experimental Models: Organisms/Strains
w[67c23] P{w[+mC]=lacW}fzr[G0418]/FM7a	BDSC	#12297
P{ry[+t7.2]=hsFLP}12, y[1] w[*]; sna[Sco]/CyO	BDSC	#1929
w[1118]; P{w[+mC]=UAS-EcR.B1-DeltaC655.F645A}TP1	BDSC	#6869
w[*] upd3[Delta]	BDSC	#55728
y[1] v[1]; P{y[+t7.7] v[+t1.8]=TRiP.JF02115}attP2	BDSC	#26217
y[1] sc[*] v[1] sev[21]; P{y[+t7.7] v[+t1.8]=TRiP.HMS01150}attP2	BDSC	#34672
w[*]; Bl[1]/CyO; P{w[+mC]=alphaTub84B(FRT.GAL80)}3	BDSC	# 38881
y[1] w[1118]; PBac{y[+mDint2] w[+mC]=UASdap.CRL4}VK00027	BDSC	#83335
y[1] w[*]; Mi{y[+mDint2]=MIC}Ilp8[MI00727]	BDSC	#33079
upd3.1lacZ	Kindly provided by Huaqi Jiang (UTSW, Dallas, TX, USA)	N/A
UAS-upd3	Kindly provided by Huaqi Jiang (UTSW, Dallas, TX, USA)	N/A
hsFLP;UAS-FRT-stop-FRT-reaper	(Cohen et al., 2018)	N/A
byn>Gal4	(Singer et al., 1996)	N/A
UAS-hid, UAS-reaper	(Cohen et al., 2018; Fox and Spradling, 2009; Sawyer et al., 2017)	N/A
UAS-fzr	Kindly provided by Mary Lilly (NIH, Bethesda, MD, USA)	N/A
Recombinant DNA
pHPdestmCherry	Addgene, (Boy et al., 2010)	#24567
Software and Algorithms
modEncode	(Celniker et al., 2009)	http://www.modencode.org/
JASPAR motif analysis	(Fornes et al., 2020)	http://jaspar.genereg.net/

Highlights.

• The Drosophila hindgut regenerates from severe injury without delaying development

• This regeneration requires G1 phase shortening and Jak-Stat signaling

• Steroid signaling and a Sox transcription factor terminate mitotic regeneration