Genetic dissection of stress-induced reproductive arrest in Drosophila melanogaster females

Abstract

By genetic manipulations, we study the roles played by insulin-producing cells (IPCs) in the brain and their target, the corpora allata (CA), for reproductive dormancy in female Drosophila melanogaster, which is induced by exposing them to a combination of low temperature (11°C), short-day photoperiod (10L:14D) and starvation (water only) for 7 days immediately after eclosion (dormancy-inducing conditions). Artificial inactivation of IPCs promotes, whereas artificial activation impedes, the induction of reproductive dormancy. A transcriptional reporter assay reveals that the IPC activity is reduced when the female flies are exposed to dormancy-inducing conditions. The photoperiod sensitivity of reproductive dormancy is lost in pigment-dispersing factor (pdf), but not cry, mutants, suggesting that light input to IPCs is mediated by pdf-expressing neurons. Genetic manipulations to upregulate and downregulate insulin signaling in the CA, a pair of endocrine organs that synthesize the juvenile hormone (JH), decrease and increase the incidence of reproductive dormancy, respectively. These results demonstrate that the IPC-CA axis constitutes a key regulatory pathway for reproductive dormancy.

Fig 1. Reproductive dormancy is enhanced by the combination of low temperature, short-day photoperiod and nutritional deficiency.

(A) A diagram showing the progression of ovarian development in 7-day-old w¹¹¹⁸ flies exposed to four different rearing conditions (indicated at the far left-hand side) after eclosion. The majority of the oldest egg chambers in a paired ovary are at stage-14 (St. 14) under 16L:8D/25°C/ad lib feeding conditions (top), St. 9 under 10L:14D/11°C/ad lib feeding conditions (second to the top), St. 8 under 16L:8D/11°C/starved conditions (second to the bottom) or St. 7 under 10L:14D/11°C/starved conditions (bottom). Typical egg chambers and ovaries in flies kept under the indicated conditions are depicted in the right-hand side panels. The vertical broken line separates the previtellogenic (left of the line) and vitellogenic stages (right of the line). Scale bars: 30 μm (left-hand side panels) and 500 μm (right-hand side panels). Yolks are encircled by broken red lines. (B and C) The proportion of flies with ovarian arrest at 11°C in w¹¹¹⁸ flies fed a normal diet (B) and those starved (C) under 16L:8D, 10L:14D and DD conditions. The means ± SEM are shown. ns, no significance; **P<0.01 by the two-tailed Fisher’s exact test (see S1 Table for numerical data). (D) Diagram showing the experimental procedure for evaluating reproductive dormancy. The experimental design shown here was adopted for all experiments. Light illumination (~620 lux) of shorter than ~5 min in duration was applied 2–3 times during the pupal stage to check the developmental status, and when necessary, the temperature was down-shifted to 15°C for 2–3 days to synchronize the emergence of flies from different batches of pupae. The animals were maintained under the constant darkness (DD) until eclosion, because DD is unlikely to prime flies to exhibit arrested vitellogenesis at the adult stage. (E) Time course of egg maturation after eclosion. The proportion of flies carrying egg chambers containing at least one egg with a yolk (ordinate, %) is plotted as a function of time after eclosion (abscissa, days) in flies kept at 11°C with the nutrient-deficient medium under three different photoperiodic conditions after eclosion. ** indicates that the difference is statistically significant at P<0.01 when the value for 10L:14D is compared to that for DD and that for 16L:8D by the two-tailed Fisher’s exact test. †† indicates that the difference is statistically significant at P<0.01 when the value for 10L:14D is compared to that for 16L:8D but not to that DD. The number of flies examined was 100 for Day 0 to Day 2 and 125 for Day 3 to Day 7 (see S1 Table for numerical data). (F) Feeding and temperature conditions immediately after eclosion affect adult lifespan. Shown are the survival curve constructed for female flies kept under the dormancy-inducing conditions throughout the adult stage (open circles) and that of female flies which were fed at 25°C for two days after eclosion and then kept under the dormancy-inducing conditions (filled circles). The number of flies examined was 42 (open circles) and 20 (filled circles). The mean (M) and median (m) lifespan data for control (M: 8.9 days; m: 9 days) and test (M: 11.7 days; m: 12 days) groups were used to estimate the statistical significance of the two groups by the log-rank test (P<0.0001) (see S1 Table for numerical data).

Fig 2. IPC activity levels are correlated with the incidence of reproductive dormancy.

(A) Expression of Gr28b.b-GAL4 in the brain (frontal view). (B-B”) Enlarged images of IPCs stained for Dilp3-lacZ (B and B”) and Gr28b.b-GAL4 (B’ and B”). 13 cells per brain were Dilp3-lacZ-positive and all of them expressed Gr28b.b-GAL4 as detected with UAS-mCD8::GFP. The brain was triply stained with an anti-GFP antibody for Gr28b.b-GAL4 (green), an anti-β-galactosidase antibody for the product of Dilp3-lacZ (red), and an anti-DN-cadherin (blue) for visualizing the brain structure. (C to H) The TRIC assay revealed that starved flies kept under a 10L:14D photoperiod exhibited a lower transcriptional reporter level than those kept under a 16L:8D photoperiod. Typical examples of IPCs (arrows) with expression of RFP (C) and GFP (C’ to F) in the TRIC assay. Labeling of target neurons (IPCs) with RFP was independent of the neural activity levels, whereas GFP expression reflected the level of neural activity because it was proportional to the amount of reconstituted split GAL4, which was correlated with the activity-dependent Ca²⁺ influx. All flies were exposed to a cold challenge (11°C) under the conditions of 10L:14D/ad lib feeding (C to C”), 16L:8D/ad lib feeding (D), 10L:14D/starvation (E) and 16L:8D/starvation (F) for 7 days. The boxed region in (C’) is enlarged in (C”). The labeling intensity of GFP was weaker in the starved flies kept under the 10L:14D photoperiod (E) than in the flies kept under the 16L:8D photoperiod (F). Scale bars: 100 μm (A); 50 μm (C”). (G) The brain region examined for quantification of the TRIC-signal intensity (I_TRIC). The GFP-labeling intensity was measured at the three brain areas indicated by dotted squares, areas a, b and c, and the signal intensity was calculated with the equation I_TRIC = a–(b + c) / 2, where a, b and c are the signal intensity at areas a, b and c, respectively. (H) Comparisons of I_TRIC between the flies kept under the conditions of 10L:14D/ad lib feeding, 16L:8D/ad lib feeding, 10L:14D/starvation and 16L:8D/starvation. The temperature was 11°C in all cases. The means ± SEM are shown. ns, no significance, **P < 0.01, ***P < 0.001, by the one-way ANOVA post hoc Tukey’s multiple comparisons test (see S1 Table for numerical data). The genotypes of flies used were 10 × UAS-IVS-mCD8::RFP, LexAop2-mCD8::GFP / w; UAS-MKII::nlsLexADBDo, UAS-p65AD::CaM / Gr28b.b-GAL4; UAS-p65AD::CaM / +. (I and J) The proportion of flies with ovarian arrest was decreased by activation (middle data set) and increased by inactivation (lower data set) of IPCs in comparison with the control (upper data set). Flies expressed GAL4 ((I) Dilp2-GAL4; (J) Dilp3-GAL4) alone (upper data set) or GAL4 together with UAS-TRPM8 (middle data set) or UAS-Kir2.1 (lower data set). The means ± SEM are shown. Statistical comparisons were made between the two photoperiodic conditions within the same genotype (*), between the GAL4-only control and TRPM8-test genotypes (†) or between the GAL-only and Kir2.1-test genotypes (§). ns, no significance, */†/§P<0.05; **/††/§§P<0.01; ***/†††/§§§P<0.001, by the two-tailed Fisher’s exact test (see S1 Table for numerical data).

Fig 3. Characteristics of reproductive dormancy in PDF and cry mutants.

(A and B) The proportion of flies with ovarian arrest was compared between the flies kept under a 16L:8D photoperiod and those kept under a 10L:14D photoperiod in heterozygotes and homozygotes for the mutations pdf (A) and cry (B). All flies were exposed to a cold-challenge (11°C) and starvation under the indicated photoperiod for 7 days. The means ± SEM are shown. ns, no significance, *P<0.05, by the two-tailed Fisher’s exact test. The number of flies examined in (A) and (B) was 125 and 100 for each test group (see S1 Table for numerical data).

Fig 4. Insulin signaling in the corpora allata regulated reproductive dormancy.

(A and B) Effects of overexpression or knockdown of JHAMT on the dormancy rate. Overexpression of wild-type JHAMT (A) reduced the proportion of flies with ovarian arrest, whereas JHAMT knockdown (B) increased it irrespective of the photoperiod. (C to I) Effects of Hmgcr (C), InR (D and E), PI3K (F and G), mTor (H), 4E-BP (I) manipulation on the proportions of flies with ovarian arrest. The proportions of flies with ovarian arrest under the 16L:8D and 10L:14D conditions are shown for the test and control genotypes. Flies carrying only GAL4 transgenes (upper two bars) and only UAS drivers (middle two bars) served as controls for (D)-(I). Examined were: Hmgcr mutant homozygotes (C), overexpression of constitutively active InR (InR-del; D), InR RNAi (E), constitutively active PI3K (PI3K^CAAX; F), dominant-negative PI3K (PI3K DN; G), dominant-negative mTOR (Tor DN; H) and hyperactive 4E-BP (Thor.LL; I). The means ± SEM are shown. Statistical comparisons were made between the two photoperiodic conditions within the same genotype (*), between the GAL4-only control and test genotypes (†) or between the UAS-only and test genotypes (§). ns, no significance; */†/§P<0.05; **/††/§§P<0.01; ***/†††/§§§P<0.001, by the two-tailed Fisher’s exact test. The number of flies examined in (A) was 100 (GAL4 only), 150 (UAS only) and 100 (GAL4 + UAS) for 16L:8D and 100 (GAL4 only), 125 (UAS only) and 100 (GAL4 + UAS) for 10L:14D. 100 flies were examined for each test group in (B) to (I) (see S1 Table for numerical data).

Fig 5. A model illustrating the molecular events likely underlying the regulation of reproductive dormancy in female D. melanogaster.

External (light and temperature) and internal (nutritional) information converge onto brain IPCs, which in turn stimulate or suppress synthesis of JH in corpora allata. JH acts on the ovary and other tissues to prevent reproductive dormancy and other biological changes associated with dormancy.