Imaginal disk growth factors are Drosophila chitinase-like proteins with roles in morphogenesis and CO₂ response

Abstract

Chitinase-like proteins (CLPs) are members of the family 18 glycosyl hydrolases, which include chitinases and the enzymatically inactive CLPs. A mutation in the enzyme's catalytic site, conserved in vertebrates and invertebrates, allowed CLPs to evolve independently with functions that do not require chitinase activity. CLPs normally function during inflammatory responses, wound healing, and host defense, but when they persist at excessive levels at sites of chronic inflammation and in tissue-remodeling disorders, they correlate positively with disease progression and poor prognosis. Little is known, however, about their physiological function. Drosophila melanogaster has 6 CLPs, termed Imaginal disk growth factors (Idgfs), encoded by Idgf1, Idgf2, Idgf3, Idgf4, Idgf5, and Idgf6. In this study, we developed tools to facilitate characterization of the physiological roles of the Idgfs by deleting each of the Idgf genes using the CRISPR/Cas9 system and assessing loss-of-function phenotypes. Using null lines, we showed that loss of function for all 6 Idgf proteins significantly lowers viability and fertility. We also showed that Idgfs play roles in epithelial morphogenesis, maintaining proper epithelial architecture and cell shape, regulating E-cadherin and cortical actin, and remarkably, protecting these tissues against CO₂ exposure. Defining the normal molecular mechanisms of CLPs is a key to understanding how deviations tip the balance from a physiological to a pathological state.

Introduction

Chitinase-like proteins (CLPs) are secreted glycoproteins with properties of cytokines and growth factors (Kawamura et al. 1999; Kzhyshkowska et al. 2016; Guan et al. 2020; Lu et al. 2022). Normally, CLPs aid in inflammatory responses, wound healing, and defense against parasites, but they also have a dark side (Sutherland et al. 2014; Di Rosa et al. 2016; Mazur et al. 2021; Pinteac et al. 2021). Excess levels of CLPs are associated with diseases that exhibit chronically inflamed tissues, such as cancer, asthma, and arthritis, and there they promote the disease process rather than mitigate it (Lee et al. 2011; Qureshi et al. 2011; Di Rosa et al. 2016; Kzhyshkowska et al. 2016; Mazur et al. 2021; Pinteac et al. 2021).

The enzymatically active members of the family 18 glycosyl hydrolases hydrolyze the glycosidic bonds of chitin, a polysaccharide present in plant, fungal, bacterial, and animal species (Samac et al. 1990; Fuhrman et al. 1992; Miyashita and Fujii 1993; Watanabe et al. 1993; Tharanathan and Kittur 2003). CLPs, on the other hand, lack hydrolytic activity due to substitution of a key amino acid (glutamic acid) in the protein's catalytic domain (e.g. with glutamine in Drosophila Idgfs and human CHI3L2, or with leucine in human CHI3L1; Kirkpatrick et al. 1995; Recklies et al. 2002; Varela et al. 2002). This mutation was facilitated by duplication of ancestral chitinases, followed by subsequent mutations and further duplications (Bussink et al. 2007). Over evolutionary time, CLPs acquired new functions that do not require chitinase activity.

Extensive research in humans has identified four CLPs: chitinase 3 like 1 (CHI3L1), chitinase 3 like 2 (CHI3L2), oviductal glycoprotein 1 (OVGP1), and stabilin-interacting CLP (SI-CLP). Most studies have focused on the expression patterns of CLPs and their clinical relevance as potential biomarkers of disease or as targets for therapy (Mazur et al. 2021; Pinteac et al. 2021). Analyses of the associated molecular mechanisms are limited to a few studies identifying binding partners, receptors, or signaling pathways (Fusetti et al. 2003; Kzhyshkowska et al. 2004, 2006; Shao et al. 2009; Zhang et al. 2009; Francescone et al. 2011; Areshkov et al. 2012; He et al. 2013; Libreros et al. 2013; Low et al. 2015; Malik et al. 2015; Guan et al. 2020), but these studies relied on in vitro cell culture, binding assays, immunohistochemistry, or biochemical methods rather than whole-animal models. To understand the molecular mechanisms of CLP function in physiological and pathological contexts, we need to analyze function in vivo, preferably in a model organism that would allow definitive functional tests.

The 6 Drosophila Idgfs are encoded by genes located at four sites in the genome: Idgf1, Idgf2, and Idgf3 are clustered together on the left arm of chromosome 2 (2L), Idgf4 lies on the X chromosome, while Idgf5 and Idgf6 reside ∼1.7 Mb apart on the right arm of chromosome 2 (2R; Fig. 1a and Supplementary Table 1; Kirkpatrick et al. 1995; Kawamura et al. 1999; Zurovcova and Ayala 2002). Although now defined with a common nomenclature, the proteins were originally identified through different routes. Idgf6 (previously called DS47) was isolated from media from cultured Schneider line-2 (S2) cells, which have macrophage-like properties (Kirkpatrick et al. 1995). Idgf1, Idgf2, Idgf3, and Idgf4 were isolated from conditioned media from imaginal disk cell culture (Clone 8 cells; Kawamura et al. 1999). Idgf5 was identified by computer search of the Drosophila genome (Zurovcova and Ayala 2002).

Fig. 1.

Idgf complete knockouts are not entirely lethal. a) Diagram shows cytological locations of Idgf genes and structure of transcript isoforms, including 5′ and 3′ untranslated regions, coding sequences, and introns. Black arrows on transcripts indicate the orientation of transcription in the plus (right-pointing) or minus (left-pointing) direction. Arrowheads indicate approximate cut sites for deleting each Idgf gene. In the sxtΔ mutant, the deletion of Idgf2 and Idgf3 spans the entire sequence from the Idgf2 5′ upstream cut site to the Idgf3 3′ downstream cut site, deleting both genes and the 400 base pairs between them. b) Homozygous and heterozygous sxtΔ mutants have low hatch rates relative to control (w¹¹¹⁸). Dots represent egg-laying assays sampled from 3 independent biological replicates with 350–852 eggs analyzed per replicate. Error bars indicate 95% confidence intervals. Significance is based on pairwise t-tests with P-values adjusted for multiple comparisons (Benjamini–Hochberg). ** P ≤ 0.01, **** P ≤ 0.0001. c) Phenotypes exhibited by adult offspring of homozygous sxtΔ parents (i.e. F2 adults). Top row, dorsal view: 40% of males (n = 48) and 71% of females (n = 31) had etched tergites (arrow). Of the same flies, 60% of males and 48% of females exhibited abnormal wing postures: wings were held out and down (“droopy” wings) or held out and up. Middle row, ventral view: rarely (<5% of flies), sxtΔ mutants displayed dark cuticle patches (arrow) reminiscent of melanotic clots. Bottom row: 5% of male wings (n = 78 wings) and 74% of female wings (n = 43 wings) displayed ectopic wing veins (arrow). w¹¹¹⁸ flies displayed none of these phenotypes (n = 20 males and 20 females). Anterior is up, distal is to the right.

Idgf proteins exhibit about 50% amino acid identity with each other and 15–25% amino acid sequence homology to family 18 glycosyl hydrolases, which include the chitinases (Kawamura et al. 1999; Varela et al. 2002). Consistent with their identification as secreted proteins, Idgfs and their human orthologs have an N-terminal signal sequence and an N-linked glycosylation sequence (Nyirkos and Golds 1990; Arias et al. 1994; Kirkpatrick et al. 1995; Hu et al. 1996; Renkema et al. 1998; Kawamura et al. 1999; Meng et al. 2010; Schimpl et al. 2012). The proteins form a characteristic barrel structure composed of 8 parallel beta sheets surrounded by 8 anti-parallel alpha helices (Varela et al. 2002).

Idgfs promote growth, proliferation, polarization, and motility in cultured cells (Kawamura et al. 1999). They participate in wound healing, protect against infection by nematodes (Kucerova et al. 2016; Broz et al. 2017), function in detoxification of cells (Broz et al. 2017), and participate in extracellular matrix organization required for molting (Pesch et al. 2016). To accomplish these functions, the Idgfs are not co-regulated, but rather, exhibit distinct expression patterns and respond to different stimuli. In the embryo, Idgf RNA is localized in cells adjacent to invaginating cells during tissue morphogenesis (Kawamura et al. 1999). In larvae and adults, Idgfs are expressed by hemocytes and the fat body (functional equivalent of the liver) and secreted into the hemolymph (Kirkpatrick et al. 1995; Kawamura et al. 1999; Khush and Lemaitre 2000; Meister et al. 2000; Carton and Nappi 2001; Irving et al. 2001). Hemocytes and fat body both participate in innate immune responses. In the ovary, Idgfs are expressed in dynamic patterns specific to the different cell types of the egg chamber (Zimmerman et al. 2017). Both over- and under-expression in egg chamber cells disrupt epithelial tube morphogenesis (Zimmerman et al. 2017; Espinoza and Berg 2020). Thus, Idgfs function in diverse processes in different tissues.

In this study, we focused on characterizing the physiological roles of the Idgfs by precisely deleting each of the Idgfs using the CRISPR/Cas9 system and assessing the effects of complete loss of function in various tissues and different life stages. We found that flies lacking all 6 Idgfs have low viability and fertility. Defects during germ cell formation or gonad formation in these mutants likely contribute to the observed low fertility. We detected numerous cuticle defects in adults, abnormal epithelial morphogenesis in egg chambers, and segmentation defects in embryos. Finally, we found that Idgfs regulate E-cadherin and cortical actin and protect epithelia against CO₂ exposure.

Materials and methods

Fly stocks

Bloomington stocks: w¹¹¹⁸ (3605), Df(2L)ED1102 (24113), Df(1)ED6989 (9056), Df(2R)BSC337 (24361), Df(2R)Exel6064 (7546). VDRC stock: Idgf6^{fTRG01331.sfGFP-TVPTBF} (Sarov et al. 2016). Stocks generated in this study and used in the experiments: Idgf^1Δ, Idgf^2Δ, Idgf^3Δ, Idgf^4Δ, Idgf^5Δ, Idgf^6Δ, Idgf^2–3Δ, Idgf^{(1Δ, 2–3Δ)}, and Idgf^4Δ; Idgf^{(1Δ, 2–3Δ, 6Δ, 5Δ)} (sxtΔ). All Idgf mutants generated in this study, including all double and triple nulls, are listed in Supplementary Table 2. Flies were maintained on standard molasses food at 25°C.

Reagents and materials

Ovaries

• Dissecting dishes and forceps

• PBS, 10×, (combine NaCl 80 g, KCl 2 g, Na₂HPO₄ 14.4 g, KH₂PO₄ 2.4 g in 1 l of H₂O, pH to 7.4)

• PBT, 1% (vol/vol) or 0.1% Triton X-100 diluted in 1× PBS

• EBR, 10× (modified Ephrussi–Beadle Ringer's solution) 1.3 M NaCl, 47 mM KCl, 19 mM CaCl₂, and 100 mM HEPES pH to 6.9

• Paraformaldehyde, 4% (wt/vol), dilute 16% (wt/vol) EM-grade paraformaldehyde (Fisher Scientific, #50980487) in 1× PBS or PBT (defined above) at a ratio of 1:4

• Western blocking reagent (WBR; Sigma-Aldrich, #11921673001)

Embryos

• Flow Buddy–Benchtop flow regulator (Genesee, #59-122B)

• Nylon mesh (Elko Filtering Co., #03-110/47)

• Egg collection baskets: Falcon tube (15 or 50 ml) with the bottom cutoff and nylon mesh attached to bottom (e.g. melt the bottom edge slightly with a Bunsen burner and press on the nylon mesh)

• Double-sided sticky tape

• Whatman paper

• Pipette controller (e.g. Millipore Sigma, #BR25900)

• Pasteur pipettes, glass, 3.5 inches

• Petri dishes: Olympus 35 mm × 10 (Genesee Scientific, #32-103)

• Plastic fly bottles punched with holes (to make egg laying chambers when attached to apple juice plates)

• 30-Gauge syringe needles + plastic syringes

• PBTA (1× PBS, 1% BSA, 0.05% or 0.1% Triton X-100, 0.02% sodium azide)

• PBT [1× PBS with 0.1% (vol/vol) Triton X-100]

• Paraformaldehyde, 4% (wt/vol), dilute 16% (wt/vol) EM-grade paraformaldehyde (Fisher Scientific, #50980487) in 1× PBS or PBT at a ratio of 1:4

• 37% formaldehyde (e.g. Sigma-Aldrich, #252549)

• Apple juice plates, 175 ml H₂O, 25 ml 4× apple juice, 5 g of agar 1.2 g of methyl paraben. Heat in microwave until agar dissolves. Pour into Petri dish lids. Attach these plates to the plastic fly bottles using rubber bands.

• Bleach

• Heptane saturated with 37% formaldehyde: Combine equal volumes of heptane and 37% formaldehyde, shake the mixture vigorously for 15 s. Let the solution settle into the 2 phases. Prepare the solution the day before it is to be used, shaking the vial or bottle periodically throughout the day. The saturated heptane is the upper phase in the 1:1 heptane:formaldehyde stock. Store at room temperature (RT), protected from light, up to several months

• 1× embryo wash solution: 0.7% NaCl, 0.05% Triton X-100, H₂O

• Normal goat serum (NGS; Fisher Scientific, #10-000-C)

• Hoyer's mounting medium: Hoyer's stock is 30 g gum Arabic, 200 g chloral hydrate, 20 g glycerol, and 50 ml H₂O. Mounting medium is 4 parts Hoyer's stock:1 part H₂O:1 part lactic acid

• Vectashield mounting medium (Vector Laboratories, #H-2000-2)

Wings

• AquaPolymount mounting medium (Polysciences Inc, #18606-20)

• Ethanol

Generation of Idgf mutant fly lines

Deletions of the entire locus or most of the locus for each of Idgf1, Idgf2, Idgf3, Idgf4, Idgf5, and Idgf6 were created using CRISPR/Cas9-catalyzed homology-directed repair. For each deletion, we designed and injected 3 vectors: 2 guide RNAs and 1 donor containing a Ds-Red eye reporter to aid in screening newly transformed flies. Using the FlyCRISPR algorithm (tools.flycrispr.molbio.wisc.edu/targetFinder), we designed guide RNAs targeting sites upstream and downstream of each gene. Guide RNAs were synthesized as 5′-phosphorylated oligos (Operon), annealed, and ligated into the BbsI sites of the pU6-BbsI-chiRNA plasmid (Addgene #45946). For the donor vector, we PCR-amplified ∼1 kb homology arms from genomic DNA of the injection strain and cloned them into pHD-DsRed-attP (Addgene #51019). We co-injected guide RNA plasmids and donor vectors (Rainbow Transgenics) into y vas-Cas9 w¹¹¹⁸ embryos (Bloomington 55821) and screened F1 flies for germline transmission of the DsRed reporter, followed by single-fly PCR and sequencing. The Idgf null strains were outcrossed to w¹¹¹⁸; +; + for 3 generations to remove background effects and remove the vas-Cas9 transgene. All single Idgf-null lines are homozygous viable and fertile. See Supplementary Table 2 for information on strains carrying multiple null alleles.

Hatch rates, larval lethality, and pupal lethality

Eclosing male and female adults were collected over a few days, mated, and aged on apple juice plates with a dab of wet yeast paste for 2 days before egg collections. Eggs were counted and allowed to develop for >24 h at 25°C to allow hatching. Subsequently, unhatched eggs were counted to determine the hatch rates. Larval lethality was assessed by allowing adults to lay on apple juice plates and allowing the eggs to hatch. All living and dead 1st instar larvae were counted, and the live larvae were transferred to vials with standard fly food. The number of subsequent pupae represents the number of transferred larvae that survived larval development. Larval lethality was calculated by subtracting the number of pupae from the total number of larvae (live transferred + dead on plates). Pupal lethality was calculated by counting the number of empty pupal cases vs the number of dead pupae on the side of the vial. Care was taken to keep the food sufficiently hydrated to encourage pupariation on the sides of the vials rather than in the food.

Fertility assay

Virgins and males were collected over 2 days and aged for 2 days in vials with dry yeast. Equal numbers of males and females of the appropriate genotypes were crossed in embryo collection bottles on apple juice plates with a dab of wet yeast paste. After a 12-h pre-collection, embryos were collected for 4 h and aged for 1.5–2 h, after which they were washed off the plate, using embryo wash and water, into a collection basket made by sealing nylon mesh to a cutoff 50-ml Falcon tube. The basket was placed into a 100 × 50 Kimax low top beaker. Freshly prepared 50% bleach was added and swirled for 90 s while viewing dechorionation under a dissecting microscope. The basket was removed from the bleach and rinsed thoroughly with embryo wash and ddH₂O. The embryos were rinsed into a 35-mm Petri dish with embryo wash and examined on an inverted phase contrast microscope. Embryos that were visibly developing (Wieschaus and Nüsslein-Volhard 1986) were scored as fertilized. The percentage of fertilized embryos was calculated from averaging counts from 2 collections each from 2 biological replicates.

Tissue processing and immunohistochemistry

Ovaries

For CO₂ experiments in Figs. 4 and 5, adult females were exposed to a 1-min pulse of 100% CO₂ by inserting the CO₂ needle into a glass vial. Thirty minutes after CO₂ exposure, the flies were killed by placing them in 100% ethanol. Ovaries were dissected in EBR, fixed for 20 min in 4% (wt/vol) paraformaldehyde, permeabilized for 1 h in PBS with 1% Triton X-100, and blocked for 1 h in WBR. Ovaries were washed in PBT, incubated on a nutator O/N at 4°C in PBT with mouse anti-Broad (DSHB 25E9.D7, 1:500), rat anti-E-cadherin (DSHB DCAD2-c, 1:50) and 5% NGS, washed in PBT, incubated for 1 h at RT in Alexa Fluor goat anti-mouse (Invitrogen A-11001, 1:200), goat anti-rat (Invitrogen A-21247, 1:200), rhodamine phalloidin (Abcam ab235138, 1:500), and DAPI (1 µg/ml), washed in PBT, mounted on slides, and imaged on a Leica SP8X LSM.

Fig. 4.

sxtΔ mutants exhibit defects in cell morphology but not patterning of DA-forming cells. Images are projections of Z-stacks. a–c′) Stage 10B egg chambers. Top row (a–c): High expression of Broad marks the roof cells in the DA primordium, DAPI indicates the nuclei. Patterning is normal in sxtΔ mutant exposed to CO₂. Lower row (a′–c′): E-cadherin delineates the apico-lateral cell junctions in follicle cells. Epithelial morphology and E-cadherin localization are disrupted in sxtΔ mutant DA-forming patches after a single pulse of 100% CO₂ exposure. Insets show magnified views outlined by the white rectangles. d–g) Stage 12 egg chambers. Filamentous actin is visualized with rhodamine-phalloidin. DA roof cells (white rectangles) form a wider appendage and are less constricted in the sxtΔ mutant (e) compared with control (w¹¹¹⁸) (d) egg chambers. f) As shown in w¹¹¹⁸ egg chambers, basal cell membranes normally exhibit bright spots of actin localization (arrow). g) In sxtΔ mutants, filamentous actin is reduced in cell membranes after CO₂ exposure. Arrowheads in the sxtΔ + CO₂ image indicate chorion autofluorescence. Images are representative of 68 egg chambers: w¹¹¹⁸ no CO₂ (not shown) (n = 7 Stage 10B and 6 Stage 12 egg chambers), w¹¹¹⁸ plus CO₂ (n = 11 Stage 10B and 9 Stage 12), sxtΔ no CO₂ (n = 10 Stage 10B and 7 Stage 12), sxtΔ plus CO₂ (n = 10 Stage 10B and 12 Stage 12). Scale bars = 50 µm.

Fig. 5.

Cortical actin and E-cadherin respond differently to CO₂ in sxtΔ egg chambers. a) Fluorescent intensity in Stage 10B roof cells (square) and Stage 12 floor cells (square). Measurements were made apically in w¹¹¹⁸ and sxtΔ egg chambers in both E-cadherin and actin channels. Shown are examples of E-cadherin in w¹¹¹⁸ egg chambers. Scale bar = 50 µm. b) Comparison of w¹¹¹⁸ and sxtΔ fluorescence intensity plot profiles averaged across 3 cell membranes, from the center of 1 cell to the center of the neighboring cell, averaged over at least 6 egg chambers for each genotype and each exposure regime as indicated. Plots show mean (inner thicker plot line) and 95% confidence limits (outer thinner plot lines). c) Integrated area under the curves in (b) and (b′) averaged over all measurements. Error bars show 95% confidence limits. *P ≤ 0.05, Student's t-test, Benjamini–Hochberg adjusted P-values.

Embryos requiring manual devitellinization and phalloidin staining

Females were aged for 2 days with wet yeast paste. Males and females were mated and allowed to lay on apple juice plates O/N or for a few hours to obtain a range of desired stages. All aspects of this experiment prior to dechorionization were done at 25°C. For CO₂ experiments only, embryos were exposed to 100% CO₂ for 2 min (Stages 8–11) or 6 min (Stages 3–5) by inserting the CO₂ needle through a hole in the plastic fly bottle. The CO₂ was humidified by flowing the gas through a bubbler while the flow rate was held precisely at 3.5 L per minute using the Flow Buddy. Using a paint brush and embryo wash, embryos were washed from plates into a basket and rinsed with embryo wash using a squirt bottle, then dechorionated in 50% bleach (1:1 bleach:H₂O) for 2 min by submerging the mesh end of the basket in a Falcon tube with the bleach. The dechorionated embryos were rinsed with embryo wash in the basket, then transferred to a 1.5 ml Eppendorf tube by squirting embryo wash onto the outside end of the mesh and into the Eppendorf tube. For hand devitellinization, embryos were fixed for 40 min in 1 ml of heptane saturated with 37% formaldehyde. Using a pipette tip with the tip cutoff and pre-treated in PBTA, embryos were transferred to Whatman paper for ∼30 s to allow the heptane to evaporate. The Whatman paper was placed embryo side down on the tape, gently tapped until embryos were stuck to the tape, and the embryos were immediately covered with PBS. Vitelline membranes were manually removed using 30-Gauge syringe needles. Using a pipette controller attached to a glass Pasteur pipette pre-treated with PBTA, embryos were transferred to PBTA (PBS with 0.1% Triton X-100, 1% BSA and 0.02% sodium azide) for staining or to PBTA (with 0.05% Triton X-100) for storage at 4°C. Stage 3–5 embryos were incubated with DAPI (1 µg/ml), and rhodamine phalloidin (Abcam ab235138, 1:500) for 1 h, and washed in PBTA (with 0.05% Triton X-100) for 1 h, rinsed 3× in PBS, and mounted in Vectashield mounting medium. Stage 8–12 embryos were incubated with rat anti-E-cadherin (DSHB DCAD2, 1:50) and mouse anti-Engrailed (DSHB 4D9) concentrate (1:20) or supernatant (1:50) in PBTA (with 0.1% Triton X-100) and 5% NGS O/N at 4°C, washed 2× 30 min in PBTA (with 0.1% Triton X-100), incubated for 1 h at RT in PBTA with 0.1% Triton X-100, 5% NGS, DAPI (1 µg/ml), rhodamine phalloidin (1:500), Alexa Fluor goat anti-mouse 488 and anti-rat 647, washed 2× 30 min in PBTA with 0.1% Triton X-100, rinsed 3× (quick rinses) in PBS, and mounted on slides with Vectashield mounting medium.

Embryos not requiring phalloidin staining

Embryos in Supplementary Figs. 3, d, d′, e, and e′ and 4, a and a′) were dechorionated and fixed in 1:1 heptane:4% paraformaldehyde for 20 min. The paraformaldehyde layer was removed and replaced with an equal amount of 100% methanol and shaken vigorously for 30 s to remove the vitelline membranes. All liquid was removed, and embryos were rinsed with methanol. Embryos were rehydrated by removing ∼1/4 of the methanol and replacing it with PBT. This rehydration process was repeated 2×, and then the embryos were rinsed 3× in PBT. Embryos were incubated in PBT with 5% NGS, rabbit anti-Vasa (Paul Lasko, 1:2,000) and mouse anti-Engrailed (DSHB 4D9, 1:50) O/N at 4°C on a nutator. Primary antibody solution was removed and embryos were washed 3× with PBT over an hour. Secondary antibody incubation was performed at RT for 1 h with Alexa Fluor (Invitrogen) goat anti-mouse (1:200) and goat anti-rabbit (1:200) in PBT and 5% NGS.

Embryos expressing Idgf6-GFP in Supplementary Fig. 3, f and g were dechorionated and rinsed 3× in PBS. PBS was removed and 1 ml of heptane saturated with 37% formaldehyde was added. The tube was shaken vigorously for 15 s and allowed to stand for 5 min at RT. To devitellinize the embryos, the saturated heptane was removed and replaced with 0.5 ml of methanol and 0.5 ml of heptane. The tube was shaken vigorously for 15 s and allowed to stand for 1 min. After devitellinization, the heptane and methanol were removed, along with the embryos that did not sink, but leaving the embryos that did sink at the bottom of the tube. One milliliter of methanol was added and embryos were rehydrated by performing 5-min washes in the following series of methanol:PBS solutions: 80%:20%, 60%:40%, and 20%:80% followed by washing for 30 min in PBS. PBS was removed and embryos were incubated on a nutator O/N at 4°C in PBTA (with 0.1% Triton X-100), rat anti-Vasa (DSHB anti-vasa-c, 1:500), and 5% NGS. The antibody solution was removed, and the embryos were rinsed 3× in PBTA (with 0.05% Triton X-100) and washed for 30 min in PBTA (with 0.05% Triton X-100), rinsed 3× in PBTA (with 0.05% Triton X-100), and incubated on a nutator for 2 h at RT in PBTA (with 0.1% Triton X-100), 5% NGS, and Alexa Fluor (Invitrogen) goat anti-rat (1:400). Antibody solution was removed, embryos were rinsed 3× in PBTA (with 0.05% Triton X-100) and washed for 30 min in PBTA (with 0.05% Triton X-100), rinsed 3× in PBS, mounted on slides in AquaPolymount, and imaged on a Leica SP8X LSM.

Wing area measurements

Wings were dissected in either PBS or 70% EtOH, transferred to 100% EtOH, and mounted in AquaPolymount. For wing measurements, images were acquired on a Nikon Microphot-FX microscope equipped with an AmScope MU1203-FL digital camera. Wing blade areas were measured using the segmented line tool in the ImageJ software (Version 1.53q) to outline the wing blade. The area within the outline was calculated in ImageJ using Analyze > measure. High resolution wing images in Fig. 1 were acquired on a Leica Inverted Multidimensional Widefield microscope.

Embryo/larval cuticle preps

Embryos and 1st instar larvae were washed from apple juice egg collection plates with embryo wash into egg collection baskets and dechorionated (2–3 min in 50% bleach). Embryos were rinsed and transferred to 1.5 ml Eppendorf tubes. Embryos were devitellinized by adding 1:1 heptane:methanol and shaking vigorously for several minutes or by vortexing. The heptane and un-devitellinized embryos floating at the interphase were removed and the devitellinized embryos were rinsed 3 times in methanol. Embryos were mounted in Hoyer's mounting medium and baked overnight at 60°C to dissolve internal tissue. The embryos and larvae were imaged on a Leica Inverted Multidimensional Widefield microscope.

Dorsal appendage phenotype analysis

Eclosing adults were collected for a few days and aged for 2 days on apple juice plates with wet yeast paste. Plates were changed daily. Flies were exposed to either 100% CO₂ or 100% N₂ by inserting the CO₂ (N₂) needle through a hole in the plastic bottle for 1-min (CO₂) or 5-min (N₂). The CO₂ was humidified by flowing the gas through a bubbler. For the 20% CO₂, flies were continuously exposed to CO₂ in a regulated CO₂ chamber for 1.5–2 days or 7–8 days. Timing of egg collection after gas exposure was according to the regimes indicated in Fig. 3 and Supplementary Fig. 5. Eggs were rinsed into a mesh basket with embryo wash, then transferred to an Eppendorf tube. Eggs were mounted in Hoyer's medium and baked O/N at 60°C. The slides were imaged on a Nikon Labophot-2 microscope equipped with an AmScope MU1203-FL digital camera. Dorsal appendages (DAs) were scored blind and categorized phenotypically as normal/mild, moderate, or severe. Fisher's exact test was used for the statistical analysis.

Fig. 3.

Idgfs protect against CO₂ exposure. a) Effect of hypoxia on DA phenotype. Air is replaced with 100% N₂ for 5 min to induce hypoxia. DA phenotype is not significantly affected by hypoxia. b) Expressing a single copy of a wild-type Idgf6 transgene rescues the DA Idgf^6Δ phenotype induced by a single, 1-min pulse of 100% CO₂. The sxtΔ phenotype is partially but not significantly rescued (P = 0.23, Fisher's exact test). Proportions of defects peak at 8–10 h after the CO₂ pulse. The Idgf^6Δ deletion is transheterozygous with a deficiency chromosome to cover potential background mutations. ****significance (P ≤ 0.0001, Fisher's exact test), NS = not significant. c) Hourly assessment of DA phenotype following a 1-min pulse of 100% CO₂. DA defects in eggs laid by sxtΔ females peak at 7–10 h after CO₂ exposure.

Image analysis and quantification

For actin and E-cadherin intensity measurements in egg chambers, images were acquired as 8-bit images at the same laser intensities and detector settings on a Leica SP8X LSM. Using ImageJ, we measured E-cadherin and actin intensity across the cell membrane in 3 pairs of cells, from the center of 1 cell to the center of the neighboring cell, in at least 6 egg chambers for each genotype and compared egg chambers exposed to CO₂ to egg chambers not exposed to CO₂. Plot profiles across 5.0 µm were averaged over all egg chambers for each genotype and treatment. Total E-cadherin or actin was calculated by integrating the area under the mean-plot-profile curve using the trapezoidal rule. Data analyses were performed in Excel. Statistical analyses were performed in Excel and R.

For actin and E-cadherin intensity measurements in embryos, images of both genotypes and treatments were acquired as 16-bit images at the same laser intensities and detector settings on a Leica SP8X LSM. For each individual channel and optical plane being measured, background was subtracted, and a median filter was applied (rolling ball radius of 2.0) prior to measurements. For each genotype, we measured E-cadherin and actin fluorescent intensity as above, across the cell membrane in 6 pairs of cells per embryo, with n = 10 embryos for each treatment (no CO₂, +CO₂) within 3 bioreplicates, except for sxtΔ bioreplicate 3, no CO₂ (n = 6 embryos). Plot profiles across 3.5 µm were averaged over all embryos for each genotype and treatment. Total E-cadherin or actin was calculated by integrating the area under the mean plot profile curve using the trapezoidal rule. Data analyses were performed in Excel. Statistical analyses were performed in Excel and R.

Statistical analysis

All statistical analyses were performed in either Excel (Version 2209) or R (Version 4.0.2). Comparisons of categorical data (i.e. curly vs straight wing phenotypes and DA phenotypes) were performed in R using Fisher's exact test. t-tests were performed in Excel, and adjustments for multiple comparisons were performed in R using the Benjamini–Hochberg method.

Results

Idgf null lines

To begin to understand how Idgfs function, we developed Idgf null lines using the CRISPR/Cas9 system. We deleted the coding region and all or part of the untranslated regions of each Idgf, creating single, double, triple, and more mutant lines, as well as a line with all 6 deletions (w¹¹¹⁸ Idgf^4Δ; Idgf^{(1Δ, 2–3Δ, 6Δ, 5Δ)}, termed “sextuple mutant” and designated as sxtΔ; Fig. 1a, Supplementary Fig. 1, and Table 2). Since Idgf1 and Idgf6 reside in introns of CG5888 and N-Acetylgalactosaminyltransferase 9 (Pgant9), respectively, we designed the deletions in such a way as to minimize the impact on non-coding regions and leave all exons intact in CG5888 and Pgant9. Similarly, the Idgf2 deletion was a partial deletion to avoid removing any part of CG5888, which has a transcriptional start site in the first intron of Idgf2 (see FlyBase for detailed information about each Idgf locus). We confirmed the extent and structure of each deletion by sequence analyses (See Supplementary Data File 16 for sequence definitions, FASTA sequences uploaded to NCBI, GenBank accession numbers OP745290–OP745323). In the laboratory environment of rich food and benign growth conditions, we found no highly penetrant defects in any single mutant. sxtΔ flies, however, exhibited numerous phenotypes described below.

sxtΔ mutants have low viability and fertility

Viability and fertility depend on many developmental processes that require patterning, growth, and cell migration. We assessed the function of Idgfs in this context by quantifying hatch rates for offspring from sxtΔ homozygous flies and from sxtΔ flies carrying the second chromosome CyO balancer (w¹¹¹⁸ Idgf^4Δ; Idgf^{(1Δ, 2–3Δ, 6Δ, 5Δ)}/CyO). We compared these rates to hatch rates of offspring from w¹¹¹⁸ and w¹¹¹⁸ carrying the same second-chromosome balancer (w¹¹¹⁸; +/CyO) (Fig. 1b). See Supplementary Data File 1 for supporting data.

The expected proportions of genotypes for offspring from w¹¹¹⁸; +/CyO parents is 25% homozygous w¹¹¹⁸; +/+, 50% w¹¹¹⁸; +/CyO, and 25% homozygous w¹¹¹⁸; CyO/CyO. Lethality for CyO/CyO offspring occurs in late embryonic or early larval development (Kidwell 1972). For offspring of this w¹¹¹⁸; +/CyO background, we observed a hatch rate of 73% compared with a hatch rate of 64% in offspring from w¹¹¹⁸ Idgf^4Δ; Idgf^{(1Δ, 2–3Δ, 6Δ, 5Δ)}/CyO parents. The lower hatch rate for embryos derived from the sxtΔ/CyO background indicates that reduction of Idgfs moderately impacts embryonic development and/or the fertility of the male parents (i.e. more of the laid eggs were not fertilized). For the heterozygous offspring from this cross (embryos carrying the CyO chromosome), maternal and zygotic contributions of wild-type Idgfs come only from the CyO chromosome. In the sxtΔ homozygous embryos, all zygotic expression of Idgfs is lost. Consistent with maternal loading of Idgf transcripts (Zimmerman et al. 2017), maternal and zygotic loss of Idgfs is more detrimental: nearly 90% of eggs laid by sxtΔ females mated to sxtΔ males fail to hatch. In comparison, only about 12% of eggs laid by w¹¹¹⁸ parents fail to hatch. Surprisingly, some sxtΔ escapers survive to adulthood, and when crossed to w¹¹¹⁸ flies, these escapers are weakly fertile (Fig. 1b).

Heterozygous offspring from sxtΔ females crossed to w¹¹¹⁸ males have a higher hatch rate (26.5%) than embryos from a cross between sxtΔ homozygotes, presumably due to zygotic expression of wild-type Idgfs, and embryos from w¹¹¹⁸ females crossed to sxtΔ males, which have both maternal and zygotic Idgf expression in embryos, have a 52% hatch rate (Fig. 1b and Supplementary Data File 1). This percentage is lower than the hatch rate from the Idgf^4Δ; Idgf^{(1Δ, 2–3Δ, 6Δ, 5Δ)}/CyO cross, possibly reflecting greater fertility of the males carrying the CyO balancer, which expresses wild-type Idgfs. These results indicate that maternal and zygotic expressions of wild-type Idgfs contribute to embryonic development and adult fertility.

To determine what portion of laid eggs were either unfertilized or had arrested before any noticeable development had taken place, we examined embryos from crosses between homozygous w¹¹¹⁸ males and females, between sxtΔ males and females, between w¹¹¹⁸ females and sxtΔ males, and between sxtΔ females and w¹¹¹⁸ males (Supplementary Data File 15). Embryos were collected for 4 h and aged for 1.5–2 h, after which they were dechorionated, placed in a petri dish containing embryo wash, and examined on an inverted phase contrast microscope. 12% of all embryos resulting from the w¹¹¹⁸ × w¹¹¹⁸ cross were undeveloped, implying that the ∼12% of w¹¹¹⁸ embryos that fail to hatch (see above) are unfertilized or undeveloped. 37% of embryos from the sxtΔ × sxtΔ cross are unfertilized or undeveloped. This observation suggests that out of the remaining laid eggs, 13% hatch and 50% die as embryos (see above). 31% of embryos from sxtΔ males crossed to w¹¹¹⁸ females were undeveloped, indicating lower fertility in sxtΔ males compared with w¹¹¹⁸ males. sxtΔ females also were less fertile: 31% of embryos from w¹¹¹⁸ males crossed to sxtΔ females were undeveloped. This result could indicate a deleterious effect on very early development due to the lack of maternally contributed Idgfs; alternatively, sxtΔ females have a greater propensity to lay unfertilized eggs. Thus, embryonic lethality is higher in embryos resulting from crosses between sxtΔ × sxtΔ flies, w¹¹¹⁸ females × sxtΔ males, or sxtΔ females × w¹¹¹⁸ males.

In addition to losses due to undeveloped embryos or losses during embryonic stages, lethality occurred during larval development. 39% of control larvae (offspring of w¹¹¹⁸ parents carrying the CyO chromosome) died during larval development compared with 62% of sxtΔ larvae (offspring of parents carrying the CyO chromosome). CyO/CyO individuals die during late embryonic development or as 1st instar larvae (Kidwell 1972), so a portion of the larval lethality could have resulted from hatched CyO/CyO individuals. Larvae might also have succumbed due to handling when transferring animals from egg-laying plates to vials. Pupal lethality was 7% in both control and sxtΔ pupae. See Supplementary Data File 2 for supporting data.

To further explore the effect of loss of Idgfs on viability and development, we compared eclosion rates between homozygous (straight-winged) and heterozygous (curly winged) adult offspring of heterozygous parents (w¹¹¹⁸; +/CyO or w¹¹¹⁸ Idgf^4Δ; Idgf^{(1 Δ, 2–3Δ, 6Δ, 5Δ)}/CyO males and females). Considering the lethality of homozygous CyO/CyO flies, the expected Mendelian fractions of curly vs straight-winged adults should be 2/3 heterozygous (curly winged) and 1/3 homozygous (straight-winged) flies. Fisher's exact test demonstrated that the relative fractions of wing phenotypes in the sxtΔ background (73% curly, 27% straight) deviates significantly from the relative fractions of wing phenotypes in the w¹¹¹⁸ background (69% curly, 31% straight; P = 0.0091; Supplementary Fig. 2a and Data File 3). Furthermore, sxtΔ homozygotes are developmentally delayed compared with w¹¹¹⁸ homozygotes. The fraction of sxtΔ homozygotes eclosing each day relative to the total after 3 days lagged behind the w¹¹¹⁸ homozygotes (Supplementary Fig. 2b).

In summary, these results indicate that loss of all 6 Idgfs leads to lower fertility, lower hatch rates, greater embryonic and larval lethality, and delays in the timing of embryonic, larval, and/or pupal development.

sxtΔ mutants lay fewer eggs, have smaller ovaries, and fewer germ cells

In addition to lower hatch rates, sxtΔ females also laid fewer eggs, averaging 5.4 eggs per female per day, compared with 14.3 eggs per female per day for control females (P = 0.018, Student's t-test, 20 females, two 24-h collections; Supplementary Fig. 3b and Data File 4). Consistent with producing fewer eggs, sxtΔ females of the same age and treated to the same nutrient-rich diet had smaller ovaries relative to w¹¹¹⁸ females. sxtΔ ovaries weighed about half as much (0.41 mg/ovary) as control ovaries (0.79 mg/ovary; n = 24 w¹¹¹⁸ ovaries and 22 sxtΔ ovaries; Supplementary Fig. 3, a and a′).

Small ovaries could result from gonads with fewer germ cells. Germ cells form at the posterior pole of the embryo and migrate through the gut, guided by attractive and repellent cues, to the somatic gonadal precursors; together these 2 cell types form the gonads (Starz-Gaiano et al. 2001; LeBlanc and Lehmann 2017). To assess whether small ovaries and low fertility were consequences of defects in gonad formation, we marked germ cells with anti-Vasa antibody in control (w¹¹¹⁸) and sxtΔ embryos (Supplementary Fig. 3, d, d′, e, and e′). In the mutants, average germ-cell numbers in 3-h old blastoderm embryos were significantly reduced relative to embryos from control flies (Supplementary Fig. 3c and Data File 5, P = 0.0016, Mann–Whitney U test, n = 12 w¹¹¹⁸ and 10 sxtΔ embryos). 16-h sxtΔ embryos had smaller gonads and exhibited germ cells outside of the gonads (Supplementary Fig. 3, e and e′, n = 8 w¹¹¹⁸ and 6 sxtΔ embryos). These defects potentially result from increased cell death or altered guidance cues, cell adhesion, or polarization. In support of the hypothesis that Idgfs facilitate population of germ cells into the gonads, we have observed a close association of Idgf6-GFP transgene expression with germ cells during embryonic Stages 4–7 (Supplementary Fig. 3, f and g, n = 5 Stage 4–7 embryos and n = 5 Stage 7–10 embryos).

Sextuple-mutant embryos lacking maternal and zygotic Idgfs exhibit segmentation defects

The insect body is organized in a series of segments, regulated by segmentation genes consisting of gap, pair rule, and segment polarity genes. For example, engrailed is a segment polarity gene that is expressed in the posterior 2 rows of each segment (Supplementary Fig. 4a) and participates in the development of denticle belts (DiNardo et al. 1985). Denticles are hook-like cuticle structures on the ventral side of each larval segment that provide traction for crawling (Supplementary Fig. 4b). During late embryogenesis, the larval epithelial cells that secrete the ventral cuticle have 1 of 2 fates: either they remain smooth, or they produce polarized, actin-based protrusions that form the denticles (Dilks and DiNardo 2010; Supplementary Fig. 4c). Maintenance of the segmental pattern of alternating smooth and denticle-producing stripes is regulated by the interactions between Wingless (Wg)-expressing cells and Engrailed (En)/Hedgehog (Hh)-expressing cells (Bejsovec and Martinez Arias 1991; Dougan and DiNardo 1992).

The pattern of En expression was disrupted in sxtΔ mutant embryos, both on a tissue level and within each cell (Supplementary Fig. 4, a′ and d′, n = 12 w¹¹¹⁸ and 10 sxtΔ embryos). Smooth cuticle overlapped denticle belts, parts of neighboring belts sometimes fused together, and single rows of denticles localized within bands of smooth cuticle (Supplementary Fig. 4, b′ and c′). 64% of sxtΔ embryos had disorganized denticle belts (n = 14 embryos) and 0% of w¹¹¹⁸ (n = 25 embryos) exhibited defects. En aberrantly localized to cell membranes (Supplementary Fig. 4, d and d′, and Data File 6) in 80% of sxtΔ embryos (n = 10) compared with 0% of w¹¹¹⁸ embryos (n = 12).

We explored the defects in larval cuticle by examining the segmental pattern of denticle-producing and non-denticle-producing cell shapes in late-stage embryos and discovered an array of altered cell shapes and sizes (Supplementary Fig. 4, c and c′, and Data File 7, n = 0/9 w¹¹¹⁸ and n = 11/14 sxtΔ embryos with defects). This phenotype could be due to patterning defects, or, alternatively, the cells may have achieved proper cell fate but could not maintain proper positioning within a segment due to loss of cell-junction integrity. Cell polarity might also have been affected.

Adult flies display a variety of cuticle defects and lesions

Adult F1 and F2 sxtΔ flies exhibited several visible phenotypes (Fig. 1c). We quantified these effects in progeny of sxtΔ parents (Supplementary Data File 8). 56% of flies displayed an abnormal wing posture, either droopy wings or wings that were held out and up, and 52% of flies displayed etched tergites (Fig. 1c, top row, n = 79 flies). Consistent with the role of Idgfs in immune responses (Kucerova et al. 2016), a small percentage of adults (<5%) had lesions resembling melanotic clots, which normally occur at the sites of wounds or at sites of pathogen invasion (Fig. 1c, middle row, arrow). Although we have not quantified this change, we have noticed a decrease in the frequency of these melanotic lesions, potentially due to improved culture conditions or to the appearance of suppressors in the sxtΔ/CyO strain. In addition, we found that on average, 30% of adult wings exhibited ectopic wing veins (Fig. 1c, bottom row, arrow, n = 121 wings), revealing defects in wing imaginal-disk patterning during larval and early pupal development. We observed a striking disparity between sxtΔ males (5% of wings) and sxtΔ females (74% of wings) displaying ectopic veins, potentially due to an X-chromosome suppressor.

Overall size of idgf mutant adults was not affected

As noted in the Introduction, CLPs have properties of growth factors (Shao et al. 2009; Lee et al. 2011; Areshkov et al. 2012). To assess whether loss of Idgfs compromises growth, we quantified fly size by measuring wing area, which correlates well with overall body size (Mirth and Shingleton 2012; Siomava et al. 2016). We compared average wing area in w¹¹¹⁸ and sxtΔ adults. We found no significant difference between these genotypes under our culture conditions (P = 0.65, Student's t-test, n = 20 w¹¹¹⁸ wings and 24 sxtΔ wings). Average wing area was ∼1.5 mm² (Supplementary Data File 9) in flies, regardless of whether they had homozygous parents or parents carrying the CyO balancer, which expresses wild-type Idgfs. This result implies that these Idgfs do not contribute to growth and proliferation of imaginal disks and histoblast cells during larval and pupal development. Consistent with this hypothesis, we have not detected any defects in bristle morphology among the thousands of sxtΔ flies we have analyzed.

Dorsal appendage tube formation is disrupted in sxtΔ mutants

Cell migration, polarization, growth, proliferation, and signaling govern the tissue movements that form morphologic structures such as epithelial tubes. We used our DA tube morphogenesis model to assess the effects of Idgf loss of function during oogenesis (Fig. 2a). DA tubes produce eggshell structures that facilitate gas exchange for the developing embryo, and they are made by a subset of follicle cells in late-stage egg chambers. The egg chamber consists of 15 germline-derived nurse cells, 1 oocyte, and a surrounding monolayer of somatic epithelial cells consisting posteriorly of columnar cells and anteriorly of squamous cells (termed “stretch” cells). The tubes develop from 2 patches of precursor cells in the follicular epithelium and, through cell intercalation and cell shape change, elongate by migrating anteriorly over the stretch cells and beneath the extracellular matrix (Dorman et al. 2004; Ward and Berg 2005; Osterfield et al. 2013). The tube cells secrete eggshell protein into the tube lumen, providing a readout for tube formation, similar to the way a mold forms a sculpture (DAs; Fig. 2a, right image). This highly conserved process of tube formation is a wrapping mechanism akin to mammalian neural tube formation (Osterfield et al. 2017).

Fig. 2.

Idgf mutants exhibit defects in DA tube formation. a) DA formation. Roof cells (blue) and floor cells (red) change shape and reorganize to make 2 DA tubes. Stretch cells, cut away to show nurse cells, guide tube elongation. DAs (arrowheads) of the laid egg bring air to the embryo that is developing inside the eggshell. b) Representative examples of DA morphology categorized into normal, moderate, and severe phenotypes. Scale bar = 100 µm. c) Moderate and severe DA phenotypes are significantly increased in eggs laid by Idgf6Δ/Df and sxtΔ females relative to the w¹¹¹⁸ control, indicating defects in tubulogenesis during oogenesis. Idgf1Δ/Df, Idgf2Δ/Df, Idgf3Δ/Df, Idgf4Δ/Df, and Idgf5Δ/Df single mutants do not significantly affect DA morphology [C(i)]. Expression of a wild-type Idgf6 transgene in sxtΔ mutant females partially rescues DA morphology [C(ii)]. Deficiency chromosomes specific to each mutant are listed in Supplementary Table 3. ***P < 0.001. Fisher's exact test, adjusted for multiple comparisons. Schematic diagrams in (a) are adapted with permission from Dorman et al. (2004).

We previously observed disruption of DA morphogenesis using Idgf1 and Idgf3 overexpression (Zimmerman et al. 2017; Espinoza and Berg 2020) and Idgf RNAi for each single mutant (Zimmerman et al. 2017). Since null mutations are the gold standard in genetics, we asked “What about complete knockout of Idgfs?” We compared the DA phenotypes of eggs laid by each single Idgf CRISPR-null mutant and the sxtΔ mutant to the DAs of eggs laid by control (w¹¹¹⁸) flies and categorized the phenotypes as mild, moderate, or severe by scoring the DAs while blinded to the genotypes (Fig. 2b). To limit the influence of potential background mutations, we assessed the phenotypes produced by single mutant genotypes Idgf1Δ through Idgf6Δ by placing each Idgf-deletion chromosome in trans to a deficiency chromosome generated independently by the Bloomington Stock Center (NIH P40OD018537; Ryder et al. 2004). Note that none of the deficiencies produced DA phenotypes on their own (Supplementary Table 3). The sxtΔ mutant was assessed homozygously due to the unfeasibility of creating flies heterozygous with deficiencies spanning each of the 6 deleted Idgf loci. Therefore, we cannot rule out the possibility that potential background mutations may influence homozygous sxtΔ embryonic and adult phenotypes.

Approximately 50% of eggs laid by the sxtΔ mutants and 18% of the Idgf6Δ mutants had moderate or severe phenotypes, but both of these percentages varied somewhat from experiment to experiment (see below). The other single mutants (Idgf^1Δ, Idgf^2Δ, Idgf^3Δ, Idgf^4Δ, and Idgf^5Δ) did not exhibit significant levels of defects [Fig. 2c (i)]. We also assessed the double mutant Idgf^2–3Δ and the triple mutant Idgf^{(1Δ,2–3Δ)}in trans to a deficiency uncovering the region. The triple mutant had a significant level of defects compared with eggs laid by w¹¹¹⁸ flies (P < 0.001, Fisher's exact test). The double mutant did not have a significant level of defects (P > 0.5, Fisher's exact test; Supplementary Data File 10). Expression of an Idgf6-GFP transgene in a sxtΔ mutant background partially rescued the sxtΔ DA phenotype [Fig. 2c (ii) and Supplementary Data File 10].

Overall, these results provide evidence for important physiologic roles of Idgfs in tissue morphogenesis, potentially including cell polarization, migration, adhesion, and signaling.

Idgfs protect against effects of CO₂ exposure

While analyzing the DA defects in eggs laid by Idgf mutants, we observed variability in penetrance and expressivity. Initially, we found differences in the percentage and severity of DA defects for eggs laid by Idgf^6Δ females. Because these 2 sets of experiments (Idgf^6Δ vs other single Idgf mutants) were performed by 2 different lab workers, we asked whether these differences were biological or whether the variability was due to a difference in technique between lab workers (e.g. age of flies, overcrowding, or CO₂ exposure). Intriguingly, we determined that there was variability in whether flies were exposed to CO₂ when they were transferred to fresh egg-laying plates. This observation raised the question as to whether loss of Idgf function rendered flies more sensitive to CO₂.

To compare CO₂ sensitivity between wild-type and mutant flies, we assessed DA morphology on eggs laid by w¹¹¹⁸ flies and Idgf mutants by using several different regimes of CO₂ exposure (Supplementary Data File 11). Treating adult females to 100% CO₂ for 60 s twice per day (∼every 12 h) for 3.5 days (Supplementary Fig. 5a) resulted in a significant increase in DA defects on the eggs laid by Idgf^2Δ through Idgf^6Δ and sxtΔ mutants, but not on the w¹¹¹⁸ eggs nor in Idgf^1Δ.

Since tissue and cellular responses to carbon dioxide depend on the length and severity of exposure (Helenius et al. 2009, 2016; Sharabi, Hurwitz, et al. 2009; Sharabi, Lecuona, et al. 2009), we asked if a prolonged exposure to a lower level of CO₂ would have a similar effect. We exposed sxtΔ mutant and w¹¹¹⁸ adult flies to either 20% CO₂ or 100% air and evaluated DA morphology at 1.5–2 days and 7–8 days of exposure. Exposure to CO₂ over either of these time periods did not significantly increase DA defects in eggs laid by sxtΔ mutants (Supplementary Fig. 5b).

We next asked whether the effect of CO₂ exposure was directly due to the CO₂ itself or due to hypoxia, the transient absence of oxygen. To distinguish between these 2 hypotheses, we exposed flies to 100% N₂ for 5 min followed by DA analysis at 8–10 h and 2 days after exposure. We did not observe a significant effect on DA morphology following this treatment, indicating that the increase in DA morphology defects was directly due to CO₂ exposure rather than hypoxia (Fig. 3a).

These results suggest that Idgf proteins protect tissues against acute CO₂ exposure. If this hypothesis is correct, expression of Idgf protein should rescue the DA mutant phenotype associated with high CO₂. To test this prediction, we used a GFP-tagged transgene, Idgf6-GFP, in which Idgf6 is fused in-frame at its C-terminus to superfolder GFP (Sarov et al. 2016). We expressed this transgene in the Idgf^6Δ and sxtΔ mutants and compared the DA phenotypes to the same mutants lacking the transgene. We exposed the flies to a single pulse of 100% CO₂ for 1 min and analyzed DA defects at 8–10 h and 2 days after exposure (Fig. 3b). While both mutants exhibited defects at 8–10 h after exposure, Idgf6-GFP expression restored the percentage of DA defects to pre-exposure levels in the Idgf^6Δ mutants, and it reduced the DA defects in the sxtΔ mutants but not to the level of statistical significance. After 2 days, the DA defects in both mutants recovered to pre-exposure levels. These results demonstrate that this gene—environment interaction is due to the loss of Idgfs. Significantly, expression of a single Idgf protein can rescue or ameliorate the DA phenotype in Idgf mutants.

To examine the time course following acute CO₂ exposure on a finer scale, we exposed adult sxtΔ flies to a single pulse of 100% CO₂ for 1 min, followed by hourly DA defect analysis (Fig. 3c). Defects peaked at 6–10 h after exposure and then recovered by ∼27 h after exposure.

In addition to the developmental roles for Idgfs in normoxia, these results support a hypothesis that Idgfs protect against morphogenetic disruption following acute CO₂ exposure.

Actin but not E-cadherin is disrupted in egg chambers after CO₂ exposure

How do Idgfs function in DA morphogenesis, and how does acute exposure to CO₂ impact that process? One hypothesis is that Idgfs are necessary for patterning the DA primordial cells. A second hypothesis is that Idgfs regulate cell morphology and epithelial tissue dynamics. To test these hypotheses, we exposed adult females to 100% CO₂ for 1 min and then waited for 20–30 min before dissecting, fixing, and staining the ovaries. We then examined patterning and cell morphology of the DA-forming cells using antibodies against Broad, a transcription factor that specifies the DA roof-forming cells, and against E-cadherin, an adhesion protein that outlines cell shapes (Dorman et al. 2004; Ward and Berg 2005; Ward et al. 2006).

Patterning was normal in both w¹¹¹⁸ and sxtΔ in both the exposed and unexposed egg chambers as revealed by 2 patches of ∼50, high-Broad-expressing cells residing just lateral to the dorsal midline (Fig. 4, a–c). Cell morphology in the DA patches, however, was more disorganized, and E-cadherin was more disrupted in the sxtΔ exposed egg chambers relative to the exposed w¹¹¹⁸ and unexposed sxtΔ egg chambers (Fig. 4, a′–c′, rectangles). At Stage 12, DA roof-cell primordia were wider, and cells were less constricted in the sxtΔ mutants relative to the DA cells in the exposed w¹¹¹⁸ egg chambers (Fig. 4, d and e, rectangles). After exposure to CO₂, sxtΔ DA-forming cells exhibited a reduction of filamentous actin from both the cell junctions and the basal surfaces (Fig. 4g). In contrast, cell shapes and actin levels looked normal in the w¹¹¹⁸ DA-forming cells (Fig. 4f).

We quantified E-cadherin and actin in the DA-forming follicle cells by measuring fluorescent intensities across cell membranes using ImageJ (Fig. 5 and Supplementary Fig. 6) and calculating the total area under the plot profiles (supporting data are provided in Supplementary Data File 12). To assess whether potential differences in intensity might be stage or cell-type specific, we analyzed roof cells, floor cells, and main-body follicle cells at Stage 10B, just prior to tube wrapping, and at Stage 12, during tube elongation (Fig. 5a and Supplementary Fig. 6a). E-cadherin intensity was lower in sxtΔ mutants relative to w¹¹¹⁸ control flies (P ≤ 0.05), but it was not significantly affected by CO₂ exposure (Fig. 5, b and c and Supplementary Fig. 6b). In contrast, actin levels were similar in w¹¹¹⁸ and sxtΔ mutants but dropped quickly in both samples upon exposure to CO₂ (P ≤ 0.05; Fig. 5, b′ and c′ and Supplementary Fig. 6b′). These results suggest that Idgfs regulate E-cadherin expression or stability in normoxia and that they impact actin dynamics in response to CO₂.

Actin is disrupted in Stage 8–12 embryos after CO₂ exposure

To ask whether Idgfs regulate E-cadherin and actin in other contexts, we examined E-cadherin and actin during embryonic development. We exposed w¹¹¹⁸ and sxtΔ embryos of all stages to 100% CO₂ for 2 min before immediately dechorionating, fixing, and staining (Supplementary Data File 13). Since embryos comprise numerous distinct cell types, we focused on cells near the surface that were relatively uniform in size and easy to measure. We chose ectodermal cells located between the mid-lateral cells and lateral edge of the germ band in abdominal segments 2–6 and measured these cells during the extended germ band stages (late Stage 8–early Stage 12). Under these conditions, we observed no significant difference between E-cadherin levels in w¹¹¹⁸ embryos or sxtΔ mutant embryos (Fig. 6, a and a′). Actin levels also did not differ significantly in w¹¹¹⁸ embryos between treatments (Fig. 6, b and b′), but actin levels were significantly lower in sxtΔ mutants exposed to CO₂ compared with unexposed embryos. Taken together, these results support our hypothesis that CO₂ exposure disrupts actin in embryonic epithelia.

Fig. 6.

Quantification of cortical actin and E-cadherin intensity in Stage 8–12 embryos. a, b) Representative embryo images showing E-cadherin (a) and actin (rhodamine phalloidin) (b) comparing genotypes and CO₂ regime as indicated. Scale bar = 50 µm. (a′, b′) Comparison of w¹¹¹⁸ and sxtΔ fluorescence intensity measured across 6 cell membranes per embryo, from the center of 1 cell to the center of the neighboring cell. The colors represent different biological replicates. Each point in the graphs represents the average of the areas under the 6 plot profiles per embryo. The notches on the box plots represent the 95% confidence interval of the medians (horizontal lines), ± 1.57*IQR/sqrt(n). IQR = interquartile range. + signs = means. Non-overlapping notches are strong evidence that their medians significantly differ. sxtΔ embryos are offspring of homozygous mutant parents.

When analyzing actin dynamics in egg chambers, we noticed that some cell types were more sensitive to CO₂ exposure than others and that the effect of CO₂ exposure on actin perdured over a longer period in sxtΔ mutants compared with w¹¹¹⁸ controls. We therefore asked whether we might see more dramatic effects in embryos if we exposed embryos during cellularization, a period when actin function is crucial (Sokac et al. 2022). We also reasoned that embryonic cells might be protected from CO₂ exposure by the eggshell. We therefore exposed early stage embryos to 100% CO₂ and increased the exposure time to 6 min. In this experiment, embryos were allowed to age for 30 min before fixing them and staining with phalloidin. We measured actin fluorescence intensity across cell membranes in syncytial blastoderm embryos during Stages 3–5. Prior to cellularization in Stage 5, nuclei in the syncytial blastoderm embryo are surrounded by actin extensions (called pseudofurrows) at the periphery. In Stage 5, the actin furrows extend around the nuclei to form a single layer of cells surrounding the yolky core. At these stages, the cells are uniform in size and shape, facilitating uniformity in measurements. As we observed in egg chambers, loss of actin was significant in early embryos after exposure to CO₂ but did not differ significantly between w¹¹¹⁸ and sxtΔ (Fig. 7, a–c; Supplementary Data File 14).

Fig. 7.

Quantification of cortical actin intensity in Stage 3–5 embryos. a) Actin in early embryos comparing genotypes and CO₂ regime as indicated. b) Comparison of w¹¹¹⁸ and sxtΔ fluorescence intensity plot profiles averaged across 6 cell membranes, from the center of 1 cell to the center of the neighboring cell, averaged over at least 10 Stage 3–5 embryos for each genotype and each exposure regime as indicated. Plots show mean (inner thicker plot line) and 95% confidence limits (outer thinner plot lines). c) Integrated area under the curves in (b) averaged over all measurements. sxtΔ embryos are offspring of homozygous mutant parents. Error bars show 95% confidence limits. **P ≤ 0.01, Student's t-test, Benjamini–Hochberg adjusted P-values.

Discussion

We investigated the in vivo roles of Drosophila imaginal disk growth factors throughout development. Previous studies have sought to characterize the function of individual Idgfs using RNAi or in cell culture. We are the first to delete each of the Idgfs and develop fly lines with single or multiple deletions of the Idgfs, including flies null for all 6 Idgfs. Although individual knock-out strains lacked obvious visible defects, each Idgf could play roles in processes that enhance survival or fecundity under specific environmental conditions. These strains represent the first example of knocking out a complete gene family dispersed in the genome, and they offer the opportunity to explore the interplay between single gene functions and redundancy in a gene family. Using these tools, we found that loss of all Idgf function produces a variety of detrimental phenotypes.

In Idgf sxtΔ mutants, gonads form with significantly reduced germ cell numbers. Most germ cells are lost or die before they reach the somatic gonadal precursor cells, consistent with the low fertility we observed in sxtΔ adults. Normally, germ cells form at the posterior pole of the embryo and are taken into the midgut primordium during gastrulation. They initially form a tight cluster in the midgut, but subsequently extend protrusions toward the midgut cells, lose adhesion from each other, disperse, migrate individually through the midgut, and associate with somatic gonadal precursors cells to form the gonads (Review: Richardson and Lehmann 2010). To migrate, germ cells become polarized and extend protrusions toward the midgut cells. One hypothesis is that Idgfs participate in remodeling of cell junctions and actin dynamics in migrating cells, consistent with our finding that loss of all Idgfs disrupts E-cadherin and actin in certain contexts. A second hypothesis is that loss of Idgf function disrupts polarization and protrusiveness of primordial germ cells, consistent with a previous study demonstrating that Idgfs promote cell motility and protrusiveness in cell culture (Kawamura et al. 1999). Studies have identified key regulators of germ cell migration, such as the lipid phosphate phosphatases, Wunen and Wunen-2 (Hanyu-Nakamura et al. 2004), the small acidic protein, 14-3-3ε (Tsigkari et al. 2012), and pathways such as JAK/STAT (Li et al. 2003; Brown et al. 2006). CLPs are known to have properties of cytokines (Lee et al. 2011), raising the question of whether Idgfs may promote guidance cues for migrating germ cells. We have observed that Idgf6 associates with migrating germ cells from embryonic Stages 4–7; this association is lost by Stage 10. How Idgfs might interact with these components or otherwise participate in germ cell development will require further study.

Consistent with studies showing expression of Idgfs throughout development, we showed that hatch rates are severely reduced in sxtΔ homozygous embryos derived from sxtΔ parents. This effect of loss of all maternal and zygotic Idgfs was ameliorated by adding either maternal, zygotic, or both maternal and zygotically expressed Idgfs. Interestingly, offspring of parents where both parents carry a CyO chromosome in a sxtΔ background have a higher hatch rate than offspring of w¹¹¹⁸ females crossed to sxtΔ males, even though in the second scenario, all embryos receive both maternal and zygotically expressed Idgfs while in the first scenario, 1 quarter of the offspring carry 2 balancers and another quarter lack zygotic Idgfs. Furthermore, w¹¹¹⁸ females crossed to sxtΔ males lay more unfertilized or undeveloped embryos than w¹¹¹⁸ females crossed to w¹¹¹⁸ males. These results demonstrate that sxtΔ males are less fertile than their sxtΔ/CyO counterparts. One hypothesis is that the sxtΔ male's fertility is compromised by reduced numbers of germ cells in the gonad, consistent with our finding that germ-cell specification, migration, and/or survival is defective in sxtΔ embryos. We speculate that the small size of the ovaries in sxtΔ females may also result from defects in germ cell biology.

Further, we found that the proportions of homozygous vs heterozygous offspring produced by heterozygous parents deviate significantly from the expected Mendelian ratios, reflecting greater lethality of the homozygotes from embryonic through adult stages. Our observation that eclosion is delayed in sxtΔ mutants could imply that Idgfs function as growth hormones; loss of function could slow development. Nevertheless, we found no significant size difference between Idgf mutants and controls.

Consistent with disruption of embryonic development (as revealed by lower sxtΔ hatch rates), we observed abnormal expression of Engrailed protein and disorganized denticle belts in embryos, including a mislocalization of Engrailed at the cell membrane. Although secretion of Engrailed is well documented in some contexts (Joliot et al. 1998; Maizel et al. 2002; Punia et al. 2019), Engrailed functions as a transcription factor in the early embryo. Patterning is achieved through the expression of the segment polarity genes wingless (wg) and engrailed (en), which activates expression of hedgehog (hh). Within each segment, cells receiving the Wg signal produce smooth cuticle posteriorly, and cells receiving the Hh signal produce denticles anteriorly (Bejsovec and Martinez Arias 1991; Swarup and Verheyen 2012). The denticle-vs-smooth cell shapes are distinct: whereas the denticle-producing cells are rectangular and arranged in rows with the long edges oriented in the ventrolateral direction (similar to a staggered brick-wall arrangement), the smooth cells are larger and not arranged in rows (Hirano et al. 2009). Mislocalization or abnormal expression of En could cause a transformation of cell fate from non-denticle-producing to denticle-producing cells, resulting in the disorganized denticle belt phenotype we observed in sxtΔ embryos. Alternatively, cells in the epidermal epithelium may have altered cell adhesion and increased motility, allowing denticle-producing cells to invade into the non-denticle-producing cells.

Segmentation defects were apparent in adults in that sxtΔ flies exhibited etched tergites. Each abdominal segment in the adult epidermis develops from 3 bilateral pairs of groups of cells called histoblast nests. The anterior pair in each segment forms the tergite. Histoblast nests proliferate after pupariation and migrate by intercalation into the larval epithelial cells, which then extrude basally and die. The histoblast nests fuse and secrete the adult cuticle (Mangione and Martin-Blanco 2018; Nardi et al. 2018; Michel and Dahmann 2020; Prat-Rojo et al. 2020; Athilingam et al. 2021; Panzade and Matis 2021; Davis et al. 2022). Successful migration depends on Decapentaplegic (Dpp) and protrusive extensions at the leading edge of the histoblast nests (Ninov et al. 2007, 2010). Our findings suggest a role for Idgfs in histoblast cell proliferation, migration, or survival, but how Idgfs function in tergite formation is unknown.

Our previous studies showed that the stretch cells relay signals from the nurse cells to the DA cells to guide their movement and ensure tube closure. These signals rely on function of the SOX transcription factor Bullwinkle (Bwk) in the nurse cells (Rittenhouse dissertation, 1996. Bullwinkle, an HMG box protein, is required for proper development during oogenesis, embryogenesis and metamorphosis in Drosophila melanogaster, PhD, University of Washington) (Rittenhouse and Berg 1995), and the non-receptor tyrosine kinases Shark and Src42a in the stretch cells (Tran and Berg 2003). The specific signaling molecules and their targets are unknown. Bwk mutants produce short, wide, open tubes that create DAs resembling moose antlers (Rittenhouse and Berg 1995; Dorman et al. 2004). Shark and Src42A act downstream of Bwk to regulate DA morphogenesis; Shark loss of function enhances the Bwk phenotype (Tran and Berg 2003).

In this study, we found that except for Idgf6, the single Idgf-null mutants did not cause a defective DA phenotype. The results for the single null mutants contrast with our previous studies using RNAi to knock down each Idgf individually (Zimmerman et al. 2017); we found that RNAi knockdown of each Idgf results in a partially penetrant DA phenotype featuring shorter, wider DAs. Shorter, wider DAs also occur when Idgfs are overexpressed or when elevated as in eggs from bwk mutant flies. Such "moose antler" DAs contrast with the moderate category of defective DAs in the null mutants, which have thin DAs. This result suggests that tissue-specific RNAi knockdown of single Idgfs actually induces a compensation process that over-expresses Idgfs, producing a gain-of-function phenotype.

Why do significant defects occur in single Idgf RNAi knockdowns, but not in single nulls? One difference in these scenarios is that RNAi knockdown is spatially and temporally restricted, occurring in a short window at precisely the time that DA tubes are formed (see below). In contrast, the deletions remove gene function throughout development, allowing a more balanced homeostasis process to occur. Functional redundancy from the remaining Idgfs or even from expression of non-Idgf genes could account for the incomplete penetrance or lack of a phenotype. Duplicated genes can maintain functional redundancy despite their divergence over evolutionary time (Dean et al. 2008; DeLuna et al. 2008; Kafri et al. 2008; Musso et al. 2008). Functional redundancy between paralogous genes can result in a strong phenotype when all of the paralogs are deleted and a weak or neutral phenotype when the genes are deleted singly (Gu et al. 2003; Dean et al. 2008; DeLuna et al. 2008; Kafri et al. 2008; Musso et al. 2008), similar to what we observed for the Idgf null mutants.

Compensation in single null mutants from the remaining wild-type paralogs or even from non-Idgf genes could be post-transcriptional, e.g. through increased translation or greater protein stability. Compensation may also occur through transcriptional adaptation, in which upregulation of compensating genes can be triggered by any of a number of mRNA decay pathways that target aberrant mRNAs (Garneau et al. 2007). A recent study demonstrated that transcriptional adaptation depends on the existence of a mutant mRNA, for which degradation can occur via different surveillance pathways, including non-stop, no-go, or non-sense-mediated decay, depending on the nature of the mutation. Upregulation can then occur in genes with sequence similarity to the degraded mRNA (Garneau et al. 2007; Rossi et al. 2015; El-Brolosy and Stainier 2017; El-Brolosy et al. 2019).

Except for Idgf^2Δ, the single null mutants cannot express a transcript. The Idgf2 deletion leaves intact the 5′-UTR, the first exon and intron, and part of the second exon, potentially allowing production of a mutant transcript and therefore transcriptional adaptation in the Idgf^2Δ and sxtΔ mutants, which carry the Idgf2 deletion. Degradation of the Idgf^2Δ transcript could not be via the non-sense-mediated decay (NMD) process, which requires a premature stop codon upstream of an intron. Degradation via a non-NMD pathway, however, could be occurring in either the Idgf^2Δ or sxtΔ mutant lines. Our RNAi experiments were performed in a wild-type background, so there would be no mutant mRNAs, and any adaptation would be on the translational or post-translational protein level (Torres et al. 2008; DeLuna et al. 2010; Donnelly and Storchova 2014; Ishikawa et al. 2017).

Although adaptation could be occurring in both the nulls and the knockdowns, the time frame for adaptation could also be a factor. In the RNAi study, we used the GAL4/UAS system to knockdown transcripts from each of the Idgf genes. We used a stretch-cell-specific Gal4 driver, which is not expressed until Stage 10 of egg chamber development, after patterning has occurred but just hours before tube formation. This timing may not allow enough time for optimal compensation to occur. On the other hand, Idgf-null mutant flies are mutant throughout their entire life cycle. Thus, the different results of our Idgf RNAi knockdown and null experiments could be due to different mechanisms of compensation. Discerning whether compensation is occurring and what contributes to the differences in phenotypes exhibited by overexpression, RNAi, and gene deletion will require further study.

Mutations can render cells less robust to endogenous or exogenous perturbations (Levy and Siegal 2008; Masel and Siegal 2009). Drosophila researchers commonly use CO₂ to anesthetize flies without considering the potential profound and long-lasting effects of CO₂ exposure. Previous studies have shown that hypercapnia can suppress the immune system and alter gene expression via the NF-κB pathway in mammalian cells (Cummins et al. 2010). Hypercapnia also reduces fertility (Helenius et al. 2009) and alters climbing and flight behavior in Drosophila (Bartholomew et al. 2015).

We demonstrated that exposing Idgf-null mutants to CO₂ enhances DA morphogenesis defects independently of hypoxia and induces loss of cortical actin in epithelial tissues during oogenesis and embryonic development. We did note molecular differences, raising the question of whether Idgfs regulate E-cadherin in normoxia and actin in response to CO₂. We saw that E-cadherin was disrupted in sxtΔ mutant egg chambers but was unaffected by CO₂ exposure in either egg chambers or embryos. In contrast, actin levels were unaffected in sxtΔ mutants vs controls, and they were significantly reduced upon CO₂ exposure in both strains (Figs. 5 and 7). Nevertheless, DA tubes proceeded to develop normally in the control but not in the sxtΔ mutants (Figs. 3c and 5a). It could be that the actin is also perturbed in wild-type tissue but recovers, whereas absence of Idgfs eliminates a similar recovery in sxtΔ tissue. The mechanisms leading to these novel outcomes are unknown.

One possible mechanism for this regulation is that Idgfs interact with the immune deficiency (IMD) pathway through the transcription factor Relish (Drosophila NF-κB). Previous studies show that Relish inhibits JNK signaling (Park et al., 2004). Activation of JNK signaling reorganizes the actin cytoskeleton, impacting developmental processes such as cell migration and dorsal closure (Jacinto et al. 2000; Kockel et al. 2001; Kaltschmidt et al. 2002; Homsy et al. 2006; Rudrapatna et al. 2014). Upon Relish loss, JNK activation causes upregulation of actin remodelers (Ramesh et al. 2021). Another possibility is that CO₂ exposure reduces pH at the cell membrane, thereby altering actin dynamics. pH is an important regulator in developmental processes such as differentiation (Ulmschneider et al. 2016) and tissue architecture (Grillo-Hill et al. 2015). Alternatively, the mechanism could involve a pathway independent of JNK signaling or pH.

A common thread among the various phenotypes produced by loss of all Idgfs is the disruption of E-cadherin and actin, an outcome that could interfere with cell migration, segmentation, and tissue morphogenesis. Consistent with this hypothesis, we recently identified eukaryotic translation initiation factor 3 subunit e (eIF3e) as a strong enhancer of Idgf3 overexpression (Espinoza and Berg 2020). eIF3e is part of a complex that regulates the redox enzyme, Mical, which functions to disassemble actin filaments and inhibit actin polymerization (Hung et al. 2010; Grintsevich et al. 2016, 2017, 2021). We are currently investigating the mechanistic role of Idgfs in this pathway.

Our investigation into the effects of complete loss of Idgfs has raised intriguing questions regarding Idgf function. What are the pathways? What are the precise mechanisms? What are the common threads among the various phenotypes? Our Idgf null lines are valuable tools that will facilitate studies providing new insights into the function of Idgfs in Drosophila. Our results suggest new avenues of investigation for the mechanisms of CLP function in cancer pathogenesis, autoimmune disease, and tissue remodeling disorders.

Data availability

Fly lines generated in our lab are available upon request. Sequences confirming the extent and structure of the Idgf deletions are available in GenBank (accession numbers OP745290–OP745323, release date 2022 December 14). The authors affirm that all other data necessary for confirming the conclusions of the article are present within the article, figures, tables, and in the supplementary data files: Data Files S1 through S16. Data File S1 contains hatch rate data, Data File S2 contains larval and pupal lethality data, Data File S3 contains eclosion data, Data File S4 contains number-of-eggs-laid data, Data File S5 contains germ-cell counts, Data File S6 contains counts of embryos with Engrailed at the cell membranes, Data File S7 contains disorganized-denticle-belt counts, Data File S8 contains adult cuticle-defect data, Data File S9 contains wing-area measurements, Data File S10 contains dorsal-appendage defect and rescue data, Data File S11 contains hypoxia and CO₂ exposure data, Data File S12 contains quantification of actin and E-cadherin in egg chambers, Data File S13 contains quantification of actin and E-cadherin in Stage 3–5 embryos, Data File S14 contains quantification of actin and E-cadherin in Stage 8–11 embryos, Data File S15 contains sxtΔ fertility data, and Data File S16 contains descriptions of the GenBank sequences.

Supplemental material available at GENETICS online.

Funding

NIH R01 GM079433.

Author contributions

C.A.B.: Project design; oversight of all experiments; analyses of mutant embryos and adult fertility; review and editing of manuscript.

L.G.S.: Generation of Idgf6Δ fly strain using CRISPR/Cas9; experiments and data analysis to examine Idgf6Δ DA defects; experiments to assess variability in phenotypes; review of manuscript.

A.E.S.: Generation of Idgf null fly strains using CRISPR/Cas9; experiments (hatch rates, DA morphogenesis, adult phenotype analysis, germ-cell migration phenotype, CO₂ response in egg chambers, patterning in egg chambers, actin, and E-cadherin measurements in egg chambers); imaging; data analyses; review of manuscript.

S.G.Z.: Writing and editing drafts, constructing figures, experiments (E-cadherin and actin in embryos, hatch rates, larval and pupal lethality, adult cuticle phenotypes, wing area measurements, embryonic and larval denticles, IHC, and phalloidin in embryos), imaging, and data analysis.

 

Communicating editor: D. Andrew