Insulin signaling in the long-lived reproductive caste of ants

Abstract

In most organisms, reproduction is correlated with shorter life span. However, the reproductive queen in eusocial insects exhibits a much longer life span than that of workers. In Harpegnathos ants, when the queen dies, workers can undergo an adult caste switch to reproductive pseudo-queens (gamergates), exhibiting a five-times prolonged life span. To explore the relation between reproduction and longevity, we compared gene expression during caste switching. Insulin expression is increased in the gamergate brain that correlates with increased lipid synthesis and production of vitellogenin in the fat body, both transported to the egg. This results from activation of the mitogen-activated protein kinase (MAPK) branch of the insulin signaling pathway. By contrast, the production in the gamergate developing ovary of anti-insulin Imp-L2 leads to decreased signaling of the AKT/forkhead box O (FOXO) branch in the fat body, which is consistent with their extended longevity.

Fig. 1.. Phenotypic plasticity in the ant Harpegnathos.

(A) Harpegnathos female workers retain reproductive potential, which is suppressed by the queen pheromone. Removal of the queen pheromone drives some nonreproductive workers to start dueling and become reproductive pseudo-queens, called gamergates (caste transition). The gamergates can transition anew to nonreproductive revertants in the presence of queen pheromones (policing and caste reversion). (B) Life span of ants during the caste transition and reversion. (Left) Survival curves of nonreproductive workers (W, black; n = 291) versus reproductive gamergates (G, pink; n = 40) during the W-to-G transition. (Right) Survival curves of revertants (R, green; n = 47) versus worker nestmates (W^R, black; n = 36) during the G-to-R reversion. Gamergates derived from 3 months of transition (pink box) were subsequently subjected to reversion (green box). P values from Log-rank (Mantel-Cox) test are indicated in the plots. (C) Ovary development during the caste transition and reversion. (Top) Schematic of an ovariole within a gamergate ovary comprising a germarium and different stages of developing egg chambers (ECs). The sixth EC, which only contains a large oocyte without nurse cells, is not shown. (Bottom) Immunofluorescence (IF) staining of (top left) ovaries of worker and (bottom left) revertant and (right) a single ovariole of a gamergate with Phalloidin (green) and 4′,6-diamidino-2-phenylindole (DAPI) (blue). The numbers of developing ECs are indicated in the gamergate panel. (D and E) Quantifications of (D) the average number of developing ECs per ovariole and (E) the surface area of the largest EC in each ovariole of worker (W), gamergate (G), and revertant (R). P values are from Kruskal-Wallis test and multiple comparisons (***P < 0.001, n = 5 individuals). Bars and error bars represent mean ± SEM.

Fig. 2.. Ins-related gene expression in different castes and its induction of oogenesis through IIS-MAPK pathways.

(A to C) RNA abundance of DEGs in the (A) central brain, (B) abdominal fat body, and (C) ovary in workers (W; black), gamergates (G; light gray), and revertants (R; dark gray). Data are from four biological replicates per caste. P values from EdgeR are indicated. Hpd-1, 4-hydroxyphenylpyruvate dioxygenase; Fasn, fatty acid synthase; Elovl, fatty acid elongase; Scd, desaturase; Far, fatty acyl-CoA reductase; Vg, vitellogenin; LogCPM, log counts per million. (D) Localization of (left) Ins mRNA in the worker and (right) gamergate brains by means of ISH. Ins mRNA is located in the IPCs (indicated with yellow arrows) located between mushroom bodies (MB). AL, antennal lobe. (E to G) Yolky oocyte production in (E) water- versus Ins-injected ants, (F) low- versus high-dueling ants, and (G) Ins-injected versus Ins and U0126 co-injected ants. In (E) and (G), ovary development is represented by the number of yolky oocytes per individual in all workers regardless of their dueling activity. Bars and error bars represent mean ± SEM (*P < 0.05, **P < 0.01, ***P < 0.001). (H) Western blot analysis of phosphorylated AKT (P-AKT), total AKT (T-AKT), phosphorylated MAPK (P-MAPK), total MAPK (T-MAPK), and tubulin levels in the worker fat body treated with different concentrations of synthetic Ins peptide (0, 5, 10, and 20 μM). (I and J) Quantification of fold changes in (I) relative P-AKT/T-AKT and (J) P-MAPK/T-MAPK. Relative levels of AKT and MAPK phosphorylation were normalized to the control fat body. n = 8 individuals; P values are from Kruskal-Wallis test with multiple comparisons. Bars and error bars represent mean ± SEM.

Fig. 3.. Activity of the IIS-AKT and IIS-MAPK pathways in different tissues and the requirement of MAPK activity for reproduction.

(A to C) IIS activity is represented by phosphorylation of AKT and MAPK (P-AKT and P-MAPK, respectively), normalized by total AKT and MAPK (T-AKT and T-MAPK, respectively). (A) Western blot analysis of P-AKT, T-AKT, P-MAPK, and T-MAPK in (left) the fat body and (right) different parts of the dissected ovary [germarium, early egg chamber (EC), and late EC tissues from worker (W) versus gamergate (G)]. (B) Quantification of the relative P-AKT and P-MAPK levels in the fat body of W and G described in (A). P values are from unpaired Student’s t test (**P < 0.01, ***P < 0.001, n = 6 individuals). Bars and error bars represent mean ± SEM. (C) Immunofluorescence (IF) staining of transcription factor FOXO in the fat body as detected by Hs-FOXO antibody (green) and DNA staining with DAPI (blue). (Left) FOXO localization in the cytoplasm of worker fat body. (Right) FOXO localization in the nucleus of gamergate fat body. Nuclei were identified with DAPI and are indicated with arrows in the FOXO staining. (D) (Top) Averages of the yolky oocyte number per workers and quantitative RT-PCRs for vitellogenin (Vg) mRNA in abdominal fat body of workers 7 days after injection with either dimethyl sulfoxide (DMSO) or U0126 at different dosages (10, 100, and 1000 μM). (Bottom) Representative bright-field images of dissected ovaries. Data are from more than 10 biological replicates per condition. Bars and error bars indicate mean ± SEM. P values are from Kruskal-Wallis test with multiple comparisons. Rpl32 is used as a reference gene for data normalization.

Fig. 4.. Imp-L2 produced from gamergate ovaries preferentially antagonizes the IIS-AKT pathway.

(A) RNA abundance of ovarian lmp-L2 and ALS in workers (W), gamergates (G), and revertants (R). Bars and error bars represent mean ± SEM. LogCPM, log counts per million. (B) IF staining of insulin/IGF-binding proteins (ALS and Imp-L2) in the late stages of the egg chambers of a gamergate ovary as indicated with a pink box in the schematic at top. Magenta indicates ALS and Imp-L2 at top and bottom, respectively. DAPI is shown in cyan. (C and D) Western blot analyses showing the effects of recombinant Imp-L2 or ALS (produced in Sf9 cells) on IIS-AKT and IIS-MAPK pathways in the abdominal fat body tissue dissected from workers. Levels of P-AKT, T-AKT, P-MAPK, T-MAPK, and tubulin detected in the fat body as a function of treatment with Ins peptide minus/plus recombinant (C) Imp-L2 or (D) ALS protein. (E) A quantitative plot of phosphorylation levels of AKT and MAPK from the worker fat body incubated with or without Imp-L2 in the presence and absence of Ins peptide. Bars and error bars indicate median with interquartile range. P values are from Mann-Whitney test (*P < 0.05, n = 8 individuals). (F) Proposed models of reproduction-associated longevity in ants. (Left) A short-lived worker. (Right) A long-lived gamergate. The basal level of insulin (Ins) secreted from the worker brain is sufficient to activate the IIS-AKT pathway through insulin receptors (InRs) in the abdominal fat body, resulting in FOXO phosphorylation and a normal life span in workers. By contrast, the high level of Ins in the gamergate brain promotes ovary maturation and contributes to MAPK activation in the fat body. An unidentified receptor (left of InR1/2) may also contribute to MAPK activation. Additionally, RNA abundance of InR1/2 is decreased in the fat body of gamergate compared with that of worker (shown as lighter color), and the gamergate ovary produces Imp-L2 protein that antagonizes IIS-AKT signaling in the fat body. Together, this might lead to nuclear localization of FOXO and longevity in gamergates.