Juvenile hormone regulation of Drosophila aging

Abstract

Background: Juvenile hormone (JH) has been demonstrated to control adult lifespan in a number of non-model insects where surgical removal of the corpora allata eliminates the hormone's source. In contrast, little is known about how juvenile hormone affects adult Drosophila melanogaster. Previous work suggests that insulin signaling may modulate Drosophila aging in part through its impact on juvenile hormone titer, but no data yet address whether reduction of juvenile hormone is sufficient to control Drosophila life span. Here we adapt a genetic approach to knock out the corpora allata in adult Drosophila melanogaster and characterize adult life history phenotypes produced by reduction of juvenile hormone. With this system we test potential explanations for how juvenile hormone modulates aging.

Results: A tissue specific driver inducing an inhibitor of a protein phosphatase was used to ablate the corpora allata while permitting normal development of adult flies. Corpora allata knockout adults had greatly reduced fecundity, inhibited oogenesis, impaired adult fat body development and extended lifespan. Treating these adults with the juvenile hormone analog methoprene restored all traits toward wildtype. Knockout females remained relatively long-lived even when crossed into a genotype that blocked all egg production. Dietary restriction further extended the lifespan of knockout females. In an analysis of expression profiles of knockout females in fertile and sterile backgrounds, about 100 genes changed in response to loss of juvenile hormone independent of reproductive state.

Conclusions: Reduced juvenile hormone alone is sufficient to extend the lifespan of Drosophila melanogaster. Reduced juvenile hormone limits reproduction by inhibiting the production of yolked eggs, and this may arise because juvenile hormone is required for the post-eclosion development of the vitellogenin-producing adult fat body. Our data do not support a mechanism for juvenile hormone control of longevity simply based on reducing the physiological costs of egg production. Nor does the longevity benefit appear to function through mechanisms by which dietary restriction extends longevity. We identify transcripts that change in response to juvenile hormone independent of reproductive state and suggest these represent somatically expressed genes that could modulate how juvenile hormone controls persistence and longevity.

Figure 1

Corpora allata knockout of adult Drosophila reduces juvenile hormone. A) Adult corpora allata (yellow cells) marked by Aug21-Gal4 driving UAS-GFP (Aug21, GFP). B) Total JH measured from adults by GC-MS per wet mass (total body). Genotypes with intact corpora allata: Aug21, GFP; UAS-NIPP1, wildtype (WT, [w¹¹¹⁸]). Corpora allata knockout (CAKO): Aug21,GFP; UAS-NIPP1. Error bars are standard deviation; significant difference indicated by * (ANOVA, P < 0.000,1).

Figure 2

Egg production in response to corpora allata knockout and rescue by JHA. A) Daily per capita fecundity of mated females in cages. Control genotypes with intact corpora allata: Aug21, GFP and UAS-NIPP1. Corpora allata knockout (CAKO): Aug21,GFP; UAS-NIPP1. B) Daily per capita fecundity of mated females in cages: wildtype (w¹¹¹⁸) and CAKO mated females exposed to vehicle control (EtOH) or to methoprene (JHA). In (A) and (B) error bars are standard deviation. C) Total eggs per mated, individually held females (age 2 to 10 days). Whisker plots show individual values, group means (wildtype, 764 eggs; CAKO, 365 eggs), one standard deviation (wildtype, 74 eggs; CAKO, 140.6 eggs) and range. D) Distribution of ovariole maturation scored by oocyte stage from mated females aged 10 days; wildtype and CAKO females, exposed to vehicle or JHA. The percentage is among ovarioles of one ovary, averaged among 842 (wildtype no JHA), 71 (wildtype with JHA), 849 (CAKO no JHA) and 780 (CAKO with JHA) females per group. Oocyte stage categories: <6, the most mature egg chamber in ovariole has no yolk; 5/6, the most mature egg chamber in ovariole has little yolk; Discont. (discontinuous), ovariole contains one mature (or hypertrophied mature) egg chamber but next most mature egg chamber contains no yolk (stage <6); Matureovariole has a continuous sequence of mature and maturing egg chambers with at least one chamber between stage 7 and 12.

Figure 3

mRNA abundance for fat body related genes. All plots: wildtype (w^DAH/w¹¹¹⁸) relative to CAKO, and both genotypes exposed to vehicle (genotype name alone in label) or JHA; A-C) Yolk proteins (yp1, yp2, yp3), D) larval serum protein (LSP1), E) fat body protein 1 (FbP1). Error bars show 95% confidence intervals among three biological replicates.

Figure 4

Adult survival of control and CAKO genotypes, and response to JHA. A) CAKO (in w¹¹¹⁸ background) relative to three control genotypes with intact corpora allata (Aug21, GFP; UAS-NIPP1, wildtype (WT(w¹¹¹⁸)). B) CAKO (in w^DAH background) relative to wildtype (w^DAH). C) CAKO (in w^DAH/w¹¹¹⁸ background) exposed to vehicle (EtOH) relative to CAKO treated with JHA, and compared to wildtype (w^DAH/w¹¹¹⁸) with EtOH or JHA. D) CAKO (in w^DAH/w¹¹¹⁸ background) with Ovo^D1 mutation relative to control w^DAH/w¹¹¹⁸ with Ovo^D1 mutation; females of both cohorts produce no eggs. In each panel, CAKO overall mortality is significantly less than each control (log rank test, P < 0.0001) while there are no significant differences among controls within a trial.

Figure 5

Adult survival in response to dietary restriction. CAKO and wildtype in w¹¹¹⁸/w^DAH background. Median lifespan of female cohorts (upper and lower 95% confidence interval) maintained on sugar-cornmeal diets with yeast at 1%, 2%, 4%, 8% and 16%. Relative risk ratios (Cox proportional hazard mortality) do not statistically differ among flies at 2%, 8% and 16%, or among those at 1% and 4% (Additional file 1).

Figure 6

Adult female stress survival. CAKO and w¹¹¹⁸/w^DAHwildtypecontrol, treated with vehicle (EtOH) or JHA while A) exposed to hydrogen peroxide, or B) fasted.

Figure 7

mRNA abundance from females with and without corpora allata, and interactions with dominant sterileOvo^D1. CAKO and wildtype are in w^DAH background. Ovo^D1 and CAKO; Ovo^D1 are in w¹¹¹⁸/w^DAH background. A) Group 1; genes repressed by JH: mRNA of 52 genes induced in females without corpora allata (CAKO) in both the fertile and sterile (Ovo^D1) genotypes. Group 2; genes induced by JH: mRNA of 42 genes decreased in females without corpora allata (CAKO) in both the fertile and sterile (Ovo^D1) genotypes. B) Heat map of overlap genes within each group. C) Gene Ontology for most enriched categories within each group.

Figure 8

mRNA abundance via qPCR for genes responsive to CAKO and JHA. Left column: CAKO and wildtype in w¹¹¹⁸/w^DAH background, Ovo^D1 and CAKO; Ovo^D1 in w¹¹¹⁸/w^DAH background. Right column:wildtype (w¹¹¹⁸/w^DAH) relative to CAKO; exposed to vehicle (genotype name alone in label) or JHA. A, B) Odorant binding protein 99b (Obp99b). C, D) Anachronism (ana). E, F) Johan 25Bii (Jon25Bii). G, H) Krupple homolog 1 (Kr-h1). I, J) Takeout (to). Error bars show 95% confidence intervals.

Figure 9

mRNA abundance of Drosophila insulin-like peptides (dilp) in response to CAKO while interacting withOvo^D1. CAKO and wildtype are in w¹¹¹⁸/w^DAH background. Ovo^D1 and CAKO; Ovo^D1 are in a w¹¹¹⁸/w^DAH background. A) dilp2. B) dilp3. C) dilp5. D) dilp6. Error bars show 95% confidence intervals.