Key features of Drosophila diapause include ovarian arrest and post-diapause recovery of fertility (the ability to produce any progeny) and fecundity (the number of progeny produced). Whereas the majority of studies of Drosophila diapause focus on arrest of oogenesis, we chose to quantify diapause recovery by assessing the ability of newly eclosed flies to undergo 35 days of diapause, recover, and produce viable progeny, which is a more stringent test of success (Figure 1A). Additionally, in pilot experiments, post-diapause fecundity exhibited sufficient variation to be useful for a GWAS.

DGRP lines (n=193) were allowed to develop under non-diapause conditions (25°C and 12:12 L:D). Virgin females were collected and transferred to diapausing conditions (10°C and 8:16 L:D) for 5 weeks. Subsequently, flies were shifted to 18°C for 1 day, followed by incubation at 25°C in a fresh vial of fly food for recovery. We chose 5 weeks based on pilot studies that showed that nearly all DGRP lines showed excellent survival at 5 weeks in diapause conditions while exhibiting sufficient variation in post-diapause fecundity to carry out GWAS. Beyond 5 weeks, fecundity was low, and there was insufficient variation for a GWAS.

We evaluated the ability of post-diapause flies to reproduce by mating individual females with two Canton-S (CS) male control (non-diapausing) flies for 4 days. Afterward, parents were removed, and progeny were allowed to develop for 12 more days. We then measured post-diapause fecundity by counting the number of adult progeny that eclosed. A minimum of 20 female flies per DGRP line were tested. Non-diapause fecundity was measured by setting crosses immediately upon collecting virgin flies. The average non-diapause and post-diapause 4-day fecundity of DGRP lines is shown in Figure 1B, arranged in ascending order of non-diapause fecundity. In general, most lines showed lower fecundity post-diapause compared to newly eclosed flies in optimal conditions, as previously reported for one inbred wild strain (Tatar et al., 2001). However, there were interesting exceptions. For example, lines 309, 738, and 853 exhibited better fecundity post-diapause compared even to young flies in optimal conditions (Figure 1—source data 1). The DGRP lines exhibit variability in fecundity both post-diapause and in non-diapause conditions (Figure 1B; Durham et al., 2014). Therefore, we normalized post-diapause to non-diapause fecundity (Figure 1C, Figure 1—source data 1). The broad-sense heritability for normalized post-diapause fecundity was 0.51 (see Materials and methods). The lines that showed poor non-diapause fecundity that improved after diapause (e.g. 362, 737, 822) suggested that there might be a genetic trade-off between the two. However, the overall correlation (Figure 1D) between non-diapause and post-diapause fecundity was 32.3% (R²=0.3228, Pearson’s correlation coefficient, r=0.5682), suggesting a positive rather than negative correlation overall between post-diapause and non-diapause fecundity. The frequency distributions of 4-day fecundity for the non-diapause (Figure 1E) and normalized fecundity (Figure 1F) conditions are shown. Although fecundity is lower post-diapause than in young flies, even in CS controls, it is higher than for flies maintained in non-diapause conditions for 35 or 42 days (Figure 1G, Figure 1—source data 3).

The normalized fecundity scores (average of normalized post-diapause fecundity [individual post-diapause 4-day fecundity/average non-diapause 4-day fecundity]) served as the basis for conducting a GWAS using the DGRP2 web tool (Huang et al., 2014; Mackay et al., 2012). A total of 546 genetic variants, encompassing single nucleotide polymorphisms (SNPs) and insertions/deletions, were identified as associated with diapause fecundity (Figure 2A and B, Figure 2—source data 1). When a variant is located within 1 kb up- or downstream of an annotated gene, or within the gene, it is considered potentially associated with that gene. We thus identified 291 candidate diapause-associated fecundity genes. Notably, 40 out of the 291 genes had previously been reported to be associated with diapause (Figure 2—source data 2) either through functional analysis (e.g. insulin receptor and couch potato [cpo]) (Kankare et al., 2012; Kubrak et al., 2014; Schmidt et al., 2008; Sim and Denlinger, 2013; Zhang and Denlinger, 2011), due to changes in gene expression (e.g. expanded [ex] and Laminin A [LanA]) (Zhao et al., 2016), or associated with cold tolerance (β-Tub97EF; Myachina et al., 2017; Figure 2—source data 2). Eighty-nine genes were associated with more than one diapause-associated variant (e.g. at least two different SNPs) (Figure 2C).

Using the gene list obtained from the GWAS, we performed a gene network analysis using the GeneMANIA application within Cytoscape (Shannon et al., 2003; Warde-Farley et al., 2010). This application facilitates the generation of network predictions based on a combination of known physical and genetic interactions, co-localization, co-expression, shared protein domains, pathway data, and predicted functional relationships between genes. Additionally, it identifies subnetworks of genes related by functional Gene Ontology (GO). Enriched subnetworks are identified by dividing the number of genes from the input set by the total number of genes associated with a specific gene ontology. The subnetwork analysis is presented in Figure 2D, Figure 2—source data 1. Two primary classes of GO terms are significantly enriched in the diapause GWAS set compared to the genome as a whole: nervous system development and the reproductive system/gonad development. These two categories are consistent with the idea that changes in the environment are sensed by the nervous system, and the information is relayed to the reproductive system.

We also compared the list of diapause-associated genes to results of other Drosophila GWAS (Figure 3, Figure 3—source data 1). Intriguingly, the most significant overlap was with genes identified as associated with olfactory behavior (Horváth and Kalinka, 2018), which has not previously been described as important for diapause. The second most significant overlap was with genes identified in non-diapause fecundity and lifespan (Durham et al., 2014). Less significant overlaps were found with genes associated with circadian rhythm (Harbison et al., 2019), chill coma recovery (Mackay et al., 2012), development time associated with lead toxicity (Zhou et al., 2016), alcohol sensitivity (Morozova et al., 2015), starvation resistance (Mackay et al., 2012), mean food intake (Garlapow et al., 2015), sleep (Harbison et al., 2013), neurotoxin methylmercury tolerance (Montgomery et al., 2014), brain wiring regulators (Li et al., 2020), and death due to traumatic brain injury (Katzenberger et al., 2015). Overlaps with genes associated with courtship, viability in lead toxicity (Zhou et al., 2016), effects of genetic architecture and Wolbachia status on starvation (Huang et al., 2014), aggression (Shorter et al., 2015), ER stress (Chow et al., 2013), leg development (Grubbs et al., 2013), startle response (Mackay et al., 2012), and embryo development time (Horváth et al., 2016) were not statistically significant. In summary, both analyses implicated the nervous system, and potentially the olfactory system in particular, in the regulation of post-diapause fecundity.

To assess the functional significance of the top candidates, we conducted an RNAi screen. First, we evaluated the effectiveness of Gal4 under diapausing conditions. As a control, we crossed Mat-α-tub-Gal4, which drives expression in the female germline to UASp-F-tractin.tdTomato to assess the effectiveness of Gal4-mediated expression at 10°C. We compared tdTomato expression in flies maintained at different temperatures for 3 weeks. Compared to 25°C (Figure 4A) and 18°C (Figure 4B), flies at 10°C exhibited as high or higher expression of tdTomato (Figure 4C).

To evaluate the effectiveness of Gal4-mediated RNAi at 10°C, we used Mat-α-tub-Gal4 to drive zpg (zero population growth, aka INX4) RNAi and assessed the extent of Zpg knockdown by immunostaining with an anti-Zpg antibody. We used Mat-α-tub-Gal4 because Zpg is expressed in the germline, and we chose to target Zpg because of the availability of the anti-Zpg antibody. Zpg is expressed from the earliest stages of germline development in the germarium (Figure 4E and E’). Mat-α-tub-Gal4 is not expressed detectably in the germarium but is expressed in early egg chambers (Figure 4A–C). So, to quantify the knockdown efficiency, we measured the level of Zpg in stage 3 egg chambers relative to the germarium staining, as an internal control. Remarkably, comparison of Mat-α-tub-Gal4>zpg RNAi flies kept at different temperatures for 3 weeks revealed that Gal4-mediated RNAi was as effective at 10°C, as it was at 18°C and 25°C (Figure 4D–H’; Figure 4—source data 1).

Fifteen GWAS hits for which multiple RNAi lines were available were selected for RNAi screening (Figure 4—source data 2). We used tub-Gal4 because it is widely expressed, including in neurons. We measured non-diapause and post-diapause fecundity. For the five RNAi lines that proved lethal, we used tubulin-Gal80^ts to suppress the RNAi expression during development and diapause and then shifted the flies to 30°C to inactivate the Gal80 and drive the RNAi during recovery (Figure 4I), and measured post-diapause fecundity. In this way, we could test specifically for a function in post-diapause recovery (Figure 4—source data 2). RNAi against two genes, Defective-proboscis-extension-response interacting protein-γ (Dip-γ) and Scribbler (sbb), significantly reduced post-diapause fecundity (Figure 4J).

To determine whether Dip-γ and sbb are required specifically in neurons, we crossed nSyb-Gal4, a pan-neuronal driver, to validated UAS-Dip-γ RNAi and UAS-sbb RNAi lines (Davis et al., 2014; Shimozono et al., 2019). Neuron-specific knockdown of Dip-γ (Figure 4K, Figure 4—source data 3) caused as severe a defect in post-diapause fecundity as ubiquitous RNAi, in contrast to glial knockdown using Repo-Gal4 (Figure 4L, Figure 4—source data 3). Pan-neuronal RNAi of sbb (Figure 4K) also significantly inhibited post-diapause fecundity more than glial RNAi (Figure 4L). We conclude that Dip-γ and sbb are required in neurons for successful post-diapause fecundity, consistent with the enrichment of this gene class in the diapause GWAS.

While we were able to show significant functional effects of Dip-γ and sbb, GWASs by their nature identify many genes with small effects, which are too small to detect individually. Furthermore, the overlap with genes associated with olfactory behavior led us to complement the genetic analysis by testing whether the antenna and neurons within it are required for successful diapause. We removed the antenna from CS flies and measured post-diapause fecundity. As a control for the surgery, we removed the arista, which is an appendage from the antenna. Removal of the antenna but not the arista reduced post-diapause fecundity compared to unmanipulated controls (Figure 5A, Figure 5—source data 1). Removal of the antenna but not the arista also reduced the number of germline stem cells (GSCs) post-diapause (Figure 5B). These results suggest that sensory cells in the antenna are important for successful recovery post-diapause.

Another important diapause trait is lifespan extension. So, we tested the effect of removing the antenna on post-diapause and non-diapause lifespan (Figure 5C and D). Antenna removal had little effect on non-diapause lifespan (median survival 72 days w/o antenna vs 75 days with antenna) (Figure 5C). In contrast, antennaless flies exhibited a substantial reduction in diapause lifespan compared to flies with antennae (Figure 5C). We considered the possibility that the precipitous deaths of many antennaless flies at 10°C (Figure 5C) might be a consequence of impaired wound healing. So, we repeated the experiments, allowing the flies 2 weeks to recover at 25°C prior to moving them to the diapause-inducing temperature of 10°C (Figure 5D). Two weeks at 25°C substantially prevented the early post-surgery death. Yet, antennaless flies still had shorter lifespans (median survival 95 days) compared to intact flies (median survival 142 days), supporting the importance of the antenna in lifespan extension (see Discussion). A potential caveat for the latter experiment is that 2 weeks at 25°C might impair diapause induction. However, flies with antennae exhibited very similar lifespan extension at 10°C with or without the 2-week wound-healing period at 25°C (compare dark green lines in Figure 5C and D), mitigating that concern. We conclude that the antennae are likely important for lifespan extension and post-diapause fecundity.

Drosophila antennae are involved in various sensory modalities, including olfaction. To identify which neurons are required for diapause, we suppressed neuronal activity in subsets of cells by expressing the tetanus toxin light chain protein (TNT), which selectively cleaves the neuronal isoform of fly synaptobrevin. Orco is the virtually universal co-receptor for odorant receptors of the OR class (Montell, 2021), so we used Orco-Gal4 to drive expression. We found that blocking neuronal transmission in the Orco-expressing neurons decreased the post-diapause/non-diapause fecundity ratio (Figure 5E, Figure 5—source data 3), further implicating odor perception in successful diapause recovery (see Discussion).

Ionotropic receptors (IRs) are another class of odorant receptor (Montell, 2021). Ir25a is broadly expressed, and inhibiting Ir25a-expressing neurons by expressing TNT reduced the post-diapause/non-diapause fecundity ratio compared to Gal4s expressed in other IR-expressing cells such as Ir8a and Ir84a, which are involved in odor perception, and Ir76b, which is involved in contact chemosensation (Figure 5F, Figure 5—source data 3). We also found that inhibiting neurons expressing Ir21a, mutations in which disrupt warm and cool avoidance behaviors (Budelli et al., 2019), dramatically impaired recovery from diapause. Blocking Ir21a neuronal transmission prevented flies from exiting diapause, as evidenced by 100% mortality of the post-diapause female flies moved to recovery conditions (n=25). Expressing TNT in the ‘hot cell’ neurons (Ni et al., 2013), which respond to heating, using HC-Gal4, also significantly reduced the post-diapause fecundity (Figure 5F, Figure 5—source data 3).

Drosophila diapause is a fascinating life history trait that confers resilience in stressful environments and extends reproductive potential and organismal longevity. Our GWAS identified nearly 300 candidate genes associated with an important but understudied feature of diapause: the ability to recover and reproduce successfully post-diapause. The most striking finding is that genes associated with neural development are highly overrepresented in the diapause set. We further confirmed that at least two such genes, Dip-γ and sbb, are essential for post-diapause fecundity.

When we started this project, little was known about the neural control of diapause. Consistent with our finding reported here that neural genes are enriched in the diapause GWAS, and our recent finding that circadian activity and sleep are dramatically altered in diapause (Meyerhof et al., 2024), two recent studies have reported that low temperature affects two subsets of circadian neurons, DN3s, and sLNvs, which in turn impact reproductive arrest (Hidalgo et al., 2023; Meiselman et al., 2022). sLNv neurons secrete the neuropeptide pigment dispersing factor (PDF) onto insulin-producing cells (IPCs), which secrete insulin-like peptides, which in turn promote JH production. At 10°C, PDF mRNA and protein levels are reduced in sLNv neurons, reducing JH production, which arrests vitellogenesis. In parallel, cool-temperature-induced reduction in DN3 activity disinhibits cholinergic ASTC-R2 neurons, which in turn inhibit vitellogenic egg chamber development. These neurons do not synapse onto the IPCs, and the mechanism by which the ASTC-R2 neurons affect oogenesis is unknown. Thus, there are at least two parallel mechanisms by which cool temperatures affect neural activity and contribute to reproductive arrest. It is not yet clear how cool temperatures reduce PDF mRNA or how PDF mRNA and protein levels rebound during recovery.

We found that the neuronal transcription factor Sbb is required post-diapause for recovery, so it is interesting to speculate that Sbb could affect PDF mRNA abundance. Sbb functions predominantly as a transcriptional repressor, so it could promote PDF expression in recovery by inhibiting a repressor of PDF. Alternatively, PDF reduces levels of the transcriptional co-activator Eyes Absent (Eya) in the IPCs, so reduced PDF leads to elevated Eya protein levels during diapause. It is possible that Sbb represses Eya transcription to facilitate recovery post-diapause. A third possibility is that Sbb functions in the DN3-ASTC-R2 pathway. We have also found that circadian rhythms and sleep behavior are dramatically altered in flies at 10°C (Meyerhof et al., 2024). It will be of interest to determine if Dip-γ and sbb are required for those behavioral effects at low temperatures, in addition to the effects on post-diapause fecundity described here.

We complemented our genetic analysis with an investigation of the cells that contribute to diapause. The overlap between olfactory-behavior-associated genes and diapause-associated genes inspired us to test if the antenna is required, and we found that the antenna is required both for post-diapause fecundity and for lifespan extension. We conclude that cells in the antenna transmit important information for successful diapause. Furthermore, we found that inhibiting olfactory receptor neurons by expressing TNT with Orco-Gal4 caused a similar effect on post-diapause fecundity as antenna removal. This is interesting because Meiselman et al. found that the antenna was not required for accumulation of mature eggs post-diapause (Meiselman et al., 2022). Our measurement of fecundity required not only that mature eggs developed in the ovary but also that they were fertilized, laid, and could develop into adults. So, what is required for production of viable progeny in addition to mature eggs? Since both feeding and mating are suppressed during diapause, the results suggest that the antenna may be required to promote these behaviors, consistent with the known role of the antenna in providing sensory information important for flies to seek food and mates (Montell, 2021; Tunstall et al., 2012). The antenna may be dispensable for relieving the vitellogenesis block in the ovary but may specifically be required for animals to find sufficient food and mates post-diapause to support maximal fecundity.

Consistent with the possible role of the olfactory system in diapause traits, in a screen of 505 gene trap lines, Tunstall et al. identified 16 genes that were expressed in ORNs, three of which are present in our diapause-associated gene set: tai, cpo, and sbb (Tunstall et al., 2012). Cpo is a known diapause-associated gene (Kankare et al., 2012), and we show here that sbb is also required for post-diapause fecundity. Tunstall et al. showed that tai is required for normal olfactory-mediated attraction to food. The olfactory system also regulates longevity in Caenorhabditis elegans (Alcedo and Kenyon, 2004), as well as Drosophila (Libert et al., 2007), and the precise effects depend on environmental factors, including temperature and availability of food and mates (Allen et al., 2015), consistent with our findings on the role of olfaction in post-diapause lifespan extension and fecundity.

We further explored whether IR-expressing ORNs might also be required post-diapause by inhibiting subsets of IR-expressing cells with TNT. The most remarkable finding was that expressing TNT with Ir21a-Gal4 resulted in post-diapause death, indicating that Ir21a-Gal4-expressing cells are essential for successful exit from dormancy. Ir21a is an IR expressed in cells sensitive to cooling (cool cells). Animals mutant for Ir21a fail to detect cooling and also fail to avoid both warm and cool temperatures (Budelli et al., 2019). That cells required for temperature sensation are essential for diapause is perhaps not surprising; however, it is at first glance puzzling that inhibiting the Ir21a-expressing cells affects diapause because Ir21a-expressing cells are located in the arista, which when amputated did not affect post-diapause fecundity in our experiments. Similarly puzzling is the observation that inhibition of the ‘hot cells’, which are also located in the arista in close proximity to the cool cells, also impairs post-diapause fecundity. One possible explanation for these results could be that cool cells and hot cells inhibit one another such that eliminating cool cell activity actually disinhibits hot cells and vice versa. In fact, Budelli et al., 2019, found hot cell spiking in Ir21a mutants, as well as defects in avoidance of both warm and cool temperatures. If inhibiting cooling-sensitive cells hyperactivates warming-sensitive cells and vice versa, then eliminating both cell types by removal of the arista might be less harmful than inactivating individual cell types.

In summary, we report the first GWAS of diapause, using the stringent assay of the ability to enter diapause, survive, exit diapause, and produce fertile offspring. This analysis has revealed a key role for the olfactory system, as well as specific genes and cells required for diapause lifespan extension and post-diapause fecundity. Additionally, the 291 candidate diapause-associated genes identified here provide many intriguing candidates for future study.

The majority of the experiments were carried out using the inbred, sequenced lines of the DGRP (Huang et al., 2014). Diapause trait quantified in 193 DGRP lines was used in the GWAS analysis. Please refer to Figure 1—source data 1. The RNAi lines and Gal4 stocks were from BDSC and VDRC (v44077, v330386, 27049, v10268, v40698, 32034, v101659, v36166, 40889, 27508, v991, v992, v21052, v100412, 54811, v51936, v106911, 44661, 53315, 54037, 57840, v104551, v17767, 63013, 56867, 80461, v104056, 35785, v60102, 9331, 5138, 7019, 23292, 41731, 41728, 51311, 41750, 28838, 35607, 7063, 58989, 51635, 64349, and 7415). Hot cell-Gal4 (Gallio et al., 2011) is a gift from Zuker lab, and Ir21a-Gal4 (Ni et al., 2016) is a gift from Craig Montell lab.

For the non-diapause fecundity assessment, a minimum of 15 newly eclosed, not more than 6-hr-old virgin females of each strain were collected under mild CO₂ anesthesia. These females were individually placed into fly food vials containing cornmeal, molasses, agar medium, and yeast. In individual female fly crosses, one female from the DGRP line and two male flies of the same genotype (CS) were added to each vial.

In the post-diapause group, a minimum of 20 newly eclosed, not more than 6-hr-old virgin female flies of each strain were collected and promptly transferred to a cold room under diapause conditions (10°C and 8L:16D) for 5 weeks. After this diapause induction period, the flies were shifted to 18°C and 12L:12D for 1 day for temperature acclimation. Subsequently, the flies were transferred to new vials with fresh food medium dusted with yeast and placed at 25°C and 12L:12D for an additional day. Similar to the non-diapause crosses, individual female fly crosses were set up. The flies were allowed to mate and lay eggs for 4 days at 25°C under a normal photoperiod of 12L:12D. After 4 days, the parental flies were discarded, and the resulting progeny were counted on the 16th day from the initial mating start date. The progeny count was limited to the number of adult flies and any black pupae produced by a single female during these 16 days. If female flies failed to survive diapause, the line was given the lowest score of zero, indicating an inability to diapause. Those who died during the 4-day cross-period were not considered for scoring. In cases of low replicates due to deaths during diapause, recovery, or cross, the process was repeated to obtain sufficient replicates. The fecundity of flies can vary with the age of fly food, dryness of fly food, age of male flies used to assess the female fecundity, crowding in the bottle, etc. To control all these variables, we have always used fresh fly food (1-week-old) and 1-week-old male flies to reduce the variation. Also, we control the number of flies in collection bottles to 30–40 – about 25–30 females and 10 males and flipped every 3–4 days to a new bottle. We have also observed the fly vials regularly to prevent dryness of food and added water whenever necessary.

Similarly, assessments of non-diapause and post-diapause fecundity were conducted for the RNAi experiments. The post-diapause fecundity was normalized by dividing individual female fly fecundity by average non-diapause fecundity of the same genotype.

For both the non-diapause and the post-diapause, the fecundity of each strain was calculated based on the average number of progeny produced across all replicates, excluding cases in which the female and/or both males died during the 4-day mating window. To accurately assess the diapause capacity of each strain, we normalized it by dividing the individual post-diapause fecundity by the average of non-diapause fecundity for each DGRP line. The average of this normalized post-diapause fecundity was then utilized in our GWAS analysis pipeline (Mackay et al., 2012). This normalization eliminated the basal difference in fecundity among DGRP lines, allowing us to focus on the variability in diapause. We have also performed a one-way ANOVA to get the mean squares for between-group and within-group variances and calculated broad-sense heritability using the formula: H²=MS_G MSE/MS_G + (k–1) MS_E, where MS_G – mean square between groups and MS_E – mean square within groups and k – number of individuals per group. Using this formula, the broad-sense heritability for normalized post-diapause fecundity was found to be 0.51.

GWAS was carried out using the DGRP2 website. We conducted GO enrichment and network analyses based on the top variants (p<1e10⁻⁵) associated with the mean post-diapause fecundity/non-diapause fecundity score using the GeneMANIA application in Cytoscape (Shannon et al., 2003; Warde-Farley et al., 2010). The GO categories and q-values from the false discovery rate-corrected hypergeometric test for enrichment are reported. Additionally, coverage ratios for the number of annotated genes in the displayed network vs the number of genes with that annotation in the genome are provided. GeneMANIA estimates q-values using the Benjamini-Hochberg procedure.

Set Comparison Appyter was used for comparing the diapause-GWAS gene list to other behavior gene lists using 13,500 as the background gene list and 0.05 as the significance level to calculate the p-value (Clarke et al., 2021).

Immunofluorescence was performed using standard procedures. Briefly, adult female ovarioles were carefully dissected in phosphate-buffered saline (PBS) using a bent tungsten needle, pulling on the stalk region of older egg chambers to minimize damage to the germarium. Ovarioles prepared for immunostaining were fixed in 4% paraformaldehyde in PBS for 20 min. To prevent sample sticking and facilitate settling, 20 µl of PBST (PBS+0.2% Triton X-100) was added to the fixing solution. After fixation, ovarioles were washed three times for 10 min each in PBST and then incubated with primary antibodies at 4°C overnight. The following morning, ovarioles were washed twice for 15 min in PBST before incubation with secondary antibodies for at least 2 hr at room temperature. After the removal of secondary antibodies, the samples were washed three times in PBST for 10 min each. Hoechst was used as a nuclear stain and added to the second wash solution. The ovarioles were subsequently mounted in Vectashield and stored at 4°C until imaging. All antibody dilutions were prepared in PBST. Due to difficulties in settling diapause samples in the solutions, a 5 min hold step was introduced on a stand before removing solutions from the tube at each step, consistently followed for all controls as well. The anti-Zpg antibody (rabbit polyclonal) was used at a 1:20,000 dilution (Smendziuk et al., 2015). The antibodies used for identifying GSCs are Hts (1B1) and Vasa (DSHB) (Easwaran et al., 2022).

Virgin female flies were collected and moved to the conditions to assess their lifespan. In the antenna-removed fly lifespan measurement, before moving to the diapausing conditions, we gave 14 days at optimal conditions (25°C, 12L:12D photoperiod) to heal the wound caused by amputation of the antenna. Flies were kept in food vials with a density of 20 or fewer flies per vial. Deaths were censored daily, and all vials were flipped every other day for lifespan measurement at optimal condition (25°C, 12L:12D photoperiod) and once every month for lifespan measurement at diapause condition (10°C, 16L:8D photoperiod) to prevent desiccation and drying up of fly food. The survival plot was created in Prism using the Kaplan-Meier survival analysis. Survival curves are compared pairwise by the Log-rank (Mantel-Cox) test to obtain the χ²-values.