The Insulin-Like Proteins dILPs-2/5 Determine Diapause Inducibility in Drosophila

Abstract

Diapause is an actively induced dormancy that has evolved in Metazoa to resist environmental stresses. In temperate regions, many diapausing insects overwinter at low temperatures by blocking embryonic, larval or adult development. Despite its Afro-tropical origin, Drosophila melanogaster migrated to temperate regions of Asia and Europe where females overwinter as adults by arresting gonadal development (reproductive diapause) at temperatures <13°C. Recent work in D. melanogaster has implicated the developmental hormones dILPs-2 and/or dILP3, and dILP5, homologues of vertebrate insulin/insulin-like growth factors (IGFs), in reproductive arrest. However, polymorphisms in timeless (tim) and couch potato (cpo) dramatically affect diapause inducibility and these dILP experiments could not exclude this common genetic variation contributing to the diapause phenotype. Here, we apply an extensive genetic dissection of the insulin signaling pathway which allows us to see both enhancements and reductions in egg development that are independent of tim and cpo variations. We show that a number of manipulations dramatically enhance diapause to ~100%. These include ablating, or reducing the excitability of the insulin-producing cells (IPCs) that express dILPs-2,3,5 employing the dilp2,3,5-/- triple mutant, desensitizing insulin signaling using a chico mutation, or inhibiting dILP2 and 5 in the hemolymph by over-expressing Imaginal Morphogenesis Protein-Late 2 (Imp-L2). In addition, triple mutant dilp2,3,5-/- females maintain high levels of diapause even when temperatures are raised in adulthood to 19°C. However at 22°C, these females all show egg development revealing that the effects are conditional on temperature and not a general female sterility. In contrast, over-expression of dilps-2/5 or enhancing IPC excitability, led to levels of ovarian arrest that approached zero, underscoring dILPs-2 and 5 as key antagonists of diapause.

Fig 1. Insulin-producing cells (IPCs) regulate reproductive dormancy in D. melanogaster.

(A) Ablation of IPCs with dilp2>hid,rpr and Insp3>hid,rpr significantly enhances diapause at both 11 and 28 days compared with controls even under a long summer photoperiod (LD16:8). (B) Hyperexcitability of IPCs decreases diapause frequency (NaChBac) whereas reducing excitability (Ork1) significantly enhances diapause also compared to the non-conducting control NOrk1. Blue 11 days, red 28 days. Numbers within bars represent replicates (see Methods). Mean ± SD, ***p<0.001. (ANOVA was performed using the arcsin transformation but the results are presented as percentages for simplicity).

Fig 2. dilps-2/5 redundantly inhibit diapause.

(A) Combined dilps-2/5 null mutations promote diapause induction (Df(3L)dilp1-5^-/- and dilp2,3,5^-/-) at both 11 and 28 days. Among three dilps expressed in IPCs, only dilp2^-/- and dilp5^-/- single mutations modestly enhance diapause frequency. Hypomorphic mutations in chico and InR also significantly enhance diapause as does the over-expression of sImp-L2 using the neuropeptide cell driver c929. (B) Similarly RNAi driven by dilp2 promoter significantly enhances diapause induction. Mean ± SD, ANOVA on arcsin transformations, *p<0.05, **p<0.01, ***p<0.001.

Fig 3. Over-expression of dilp2 and dilp5 dramatically reduces diapause levels.

dilps-2/5 over-expression from early (dilp2(p)>dilp2 and dilp2(p)>dilp5) or late (dilp2>dilp2 and dilp2>dilp5) larval stages inhibit diapause at 12°C under short photoperiod (LD8:16). Mean ± SD, ANOVA on arcsin transformations, ***p<0.001. The same result is observed with ectopic drivers (see text for details).

Fig 4. Combinations of dilp mutations maintain diapause under more favourable conditions.

Under warmer temperatures and longer photoperiods (LD16:8), conditions favourable to growth, Df(3L)dilp1-5^-/- and Df(3L)dilp1-5/dilp2,3,5^- mutants remain in diapause at high levels even at 19°C compared to controls, but ovaries mature at 23°C. Mean ± SD, ANOVA on arcsin transformations, **p<0.01.

Fig 5. Regulation of FoxO and dilp2/5 in diapausing females.

(A) FoxO.RE-Luciferase reporter gene assay. Reporter activity in diapausing (12°C) versus non-diapausing flies (23°C) is shown (p = 0.032, t-test). Reporter activity is higher in flies abdomen at 12°C. Y-axis: Luciferase activity (fold change). (B) qPCR of dilps-2/5 mRNA levels in two different ‘wild-type’ genotypes, FoxO.RE.Luciferase used in A and dilp2>+. dilps-2/5 expression levels of diapausing (12°C) versus non-diapausing flies (23°C) are shown (ratio: diapausing/non-diapausing). dilps-2/5 are up-regulated in diapausing flies. Y-axis: mRNA levels (fold change). Dotted line indicates the expression levels in non-diapausing flies. Mean ± SE, *p<0.05, ***p<0.001 are based on t-test.

Fig 6. Schematic summary of dILP2-5 signaling effects on diapause.

(A) dILP2-5 signaling enhancement, obtained either through IPCs sensitization (NaChBac expression) or dilp2-5 over-expression (within IPCs or in other endocrine tissues, corpus allatum, corpus cardium and fat bodies) propels ovarian development even under diapause-inducing conditions (low temperature, 12°C, and short photoperiod). (B) Impairment of dILP2-5 signaling through Ork1 expression within IPCs, insulin signaling mutants (both dilp, InR and chico mutants), IPCs ablation, dilp2-5-RNAi, or over-expressing the dILPs inhibitor Imp-L2, consolidates diapause preventing further ovarian maturation.