A genome-wide association study implicates the olfactory system in Drosophila melanogaster diapause-associated lifespan extension and fecundity

Abstract

The effects of environmental stress on animal life are gaining importance with climate change. Diapause is a dormancy program that occurs in response to an adverse environment, followed by resumption of development and reproduction upon the return of favorable conditions. Diapause is a complex trait, so we leveraged the Drosophila Genetic Reference Panel (DGRP) lines and conducted a genome-wide association study (GWAS) to characterize the genetic basis of diapause. We assessed post-diapause and non-diapause fecundity across 193 DGRP lines. GWAS revealed 546 genetic variants, encompassing single nucleotide polymorphisms, insertions, and deletions associated with post-diapause fecundity. We identified 291 candidate diapause-associated genes, 40 of which had previously been associated with diapause, and 89 of which were associated with more than one SNP. Gene network analysis indicated that the diapause-associated genes were primarily linked to neuronal and reproductive system development. Similarly, comparison with results from other fly GWAS revealed the greatest overlap with olfactory-behavior-associated and fecundity-and-lifespan-associated genes. An RNAi screen of selected candidates identified two neuronal genes, Dip-γ and Scribbler, to be required during recovery for post-diapause fecundity. We complemented the genetic analysis with a test of which neurons are required for successful diapause. We found that although amputation of the antenna had little to no effect on non-diapause lifespan, it reduced diapause lifespan and post-diapause fecundity. We further show that olfactory receptor neurons and temperature-sensing neurons are required for successful recovery from diapause. Our results provide insights into the molecular, cellular, and genetic basis of adult reproductive diapause in Drosophila.

Keywords: D. melanogaster; diapause; dormancy; fecundity; genetics; genomics; lifespan; resilience; stress.

Figure 1.. Quantification of diapause in Drosophila Genetic Reference Panel (DGRP) lines measured as the ratio of post-diapause to non-diapause fecundity.

(A) Schematic of experimental workflow. (B) Average number of progenies produced (fecundity) in a 4-day individual female fly mating experiment of DGRP lines either as non-diapausing (yellow) or after a 35-day post-diapause (blue) virgin flies. Each dot represents the average fecundity, and the line above represents the standard error. DGRP line numbers are indicated wherever the post-diapause fecundity exceeds the non-diapause fecundity. (C) Normalized post-diapause fecundity average of (individual post-diapause fecundity/mean non-diapause fecundity) of each DGRP line. (D) Correlation of post-diapause to non-diapause fecundity. Pearson’s r=0.5682. r²=0.3228. (E–F) Frequency distribution of DGRP lines fecundity under non-diapause (E) and of the normalized post-diapause fecundity (F). (G) Average 4-day fecundity of single female flies, each crossed with two young Canton-S male flies, aged for 1, 35, or 42 days in non-diapause conditions or kept in diapause conditions for 35 or 42 days followed by recovery. One-way ANOVA and Tukey’s multiple comparison test, compact letter display shows comparisons. n is the number of individual female fly fecundity measured, and whiskers represent the smallest and largest values within 1.5× the interquartile range (IQR).

Figure 2.. Genome-wide association of Drosophila diapause.

(A) Manhattan plot for genome-wide association distribution. The position of each point along the y-axis indicates –log₁₀(p-value) of association of a single nucleotide polymorphism (SNP), insertion, or deletion. Points above the blue line have a p-value<1e^–5. The red line represents Bonferroni-corrected p-value = 4.8e^–8. (B) Q-Q plot of p-values from the Drosophila Genetic Reference Panel (DGRP) single variant genome-wide association study (GWAS) with the red line representing expected p-value and observed p-values deviating (black dots) from expected. (C) Numbers of genetic variants and candidate genes associated with diapause according to the GWAS. (D) Subnetworks from the Cytoscape analysis showing q-value (using the Benjamini-Hochberg procedure) for each subnetwork identified.

Figure 3.. Common genes to diapause-genome-wide association study (GWAS) hits and other behavior-associated genes.

(A–T) Venn diagrams illustrate the intersection of genes associated with diapause identified through GWAS (diapause-GWAS) with genes from other behavior-related gene lists obtained from various studies. The diapause-GWAS gene set is represented as the first set throughout the figure, while subsequent sets represent different behavior-related gene lists identified in separate studies. The percentage of common genes compared to the total genes from different respective behavior-associated gene lists is provided for each Venn diagram. p-Values of overlap to the diapause gene list determined by Fisher’s exact tests are also provided. Venn diagrams are arranged in the order of p-values.

Figure 4.. RNAi-mediated loss-of-function study to identify genes involved in diapause.

(A–C) Mat-ɑ-tub-Gal4 driving expression of UASp-F-tractin.tdTomato (red) at the indicated temperatures. Scale bars are 100 µm. (D) Quantification of zpg RNAi knockdown in stage 3 egg chambers normalized to the level of Zpg in the germarium (one-way ANOVA and Tukey’s multiple comparison test, compact letter display shows comparisons). The numbers (n) of stage 3 egg chambers quantified are shown, and whiskers represent the smallest and largest values within 1.5× the interquartile range (IQR). (E–H) Representative images of egg chambers stained with anti-Zpg antibody (green) from either control (no-knockdown) (E, E’) or knockdown of zpg (F-H’) driven by Mat-ɑ-tub-Gal4 at different temperatures. (E’–H’) are higher magnification, single channel views of the ovarioles shown in (E–H). Scale bars are 100 µm in (E–H) and 20 µm in (E’–H’). All flies in (A–H’) were kept at respective temperatures for 3 weeks. (I) Experimental design for RNAi knockdown specifically during recovery for the experiment shown in (J). The temperature-sensitive Gal80 repressor of Gal4 prevented RNAi expression during development and diapause. Incubation at 30°C during recovery inactivates Gal80, allowing Gal4-mediated RNAi knockdown. (J) Ubiquitous knockdown of Dip-γ or sbb with tub Gal4 specifically during recovery as shown in (I) significantly reduces post-diapause/non-diapause fecundity compared to the control (tubGal80^TS;tubGal4 >Ctrl RNAi #9331). (K) Pan-neuronal RNAi knockdown of Dip-γ and sbb with nSybGal4 significantly reduces post-diapause/non-diapause fecundity compared to the control (nSyb Gal4>Ctrl RNAi #54037). (L) Glia-specific knockdown of Dip-γ or sbb with Repo Gal4 causes little or no reduction in post-diapause/non-diapause fecundity (Control- Repo Gal4>Ctrl RNAi #54037). In (J–L), one-way ANOVA and Tukey’s multiple comparison test, compact letter display shows comparisons. n is the number of individual female flies tested, and whiskers represent the smallest and largest values within 1.5× the interquartile range (IQR).

Figure 5.. Neural control of diapause.

(A) Effect of antenna removal on recovery of fecundity post-diapause. Arista removal was used as a control for the surgery. n is the number of individual female flies tested. One-way ANOVA and Tukey’s multiple comparison test, compact letter display shows comparisons. Whiskers represent the smallest and largest values within 1.5× the interquartile range (IQR). (B) Effect of antenna removal on germline stem cell (GSC) recovery after 5 weeks of diapause. n is the number of germaria counted (there are typically 2–3 GSCs/germarium). One-way ANOVA and Tukey’s multiple comparison test with compact letter display to show comparisons. Error bars represent standard error. (C) Role of antenna in lifespan extension in diapause. Control flies were maintained at 25°C and diapause flies were moved to 10°C. Median survival for flies with intact antenna in diapause (+A_Diap) – 142.5 days; antennaless flies in diapause (-A_Diap) – 7 days; control with intact antenna in optimal conditions (+A_Ctrl) – 75 days; and antennaless flies in optimal conditions (-A_Ctrl) – 72 days. Survival curves are compared pairwise using the Log-rank (Mantel-Cox) test and Χ² values are: Diap±antenna = 98.89, Ctrl ±antenna = 11.69, Diap+antenna vs Ctrl+antenna = 163.5, and Diap – antenna vs Ctrl – antenna=0.5. n represents the number of flies used for the survival curve. (D) Control (+A_Diap) and antennaless (-A_Diap) flies were maintained at 25ºC for 2 weeks post-surgery to allow for wound healing and shifted to the diapausing conditions. Median survival for +A_Diap – 142 days; -A_Diap – 95 days. Survival curves are compared pairwise using the Log-rank (Mantel-Cox) test and Χ² values = 86. n represents the number of flies used for the survival curve. (E–F) Effects on diapause of inactivating neuronal transmission by driving tetanus toxin using UAS-TeTxLC.tdt, (E) in odor responsive neurons using the indicated Gal4 lines. Orco is a co-receptor for odorant receptors of the OR class. (F) Ir8a is a co-receptor involved in organic acid detection. Ir25a is a co-receptor involved in chemo- and thermo-sensation. Ir76b is involved in the detection of various amines and salt. Ir84a is involved in the detection of phenylacetic acid and male courtship behavior. Hot cells are heat-sensitive cells in the arista. In (E-F), one-way ANOVA and Tukey’s multiple comparison test, compact letter display shows comparisons. n is the number of individual female flies tested, and whiskers represent the smallest and largest values within 1.5× the interquartile range (IQR).