Neonicotinoid-induced pathogen susceptibility is mitigated by Lactobacillus plantarum immune stimulation in a Drosophila melanogaster model

Pesticides are used extensively in food production to maximize crop yields. However, neonicotinoid insecticides exert unintentional toxicity to honey bees (Apis mellifera) that may partially be associated with massive population declines referred to as colony collapse disorder. We hypothesized that imidacloprid (common neonicotinoid; IMI) exposure would make Drosophila melanogaster (an insect model for the honey bee) more susceptible to bacterial pathogens, heat stress, and intestinal dysbiosis. Our results suggested that the immune deficiency (Imd) pathway is necessary for D. melanogaster survival in response to IMI toxicity. IMI exposure induced alterations in the host-microbiota as noted by increased indigenous Acetobacter and Lactobacillus spp. Furthermore, sub-lethal exposure to IMI resulted in decreased D. melanogaster survival when simultaneously exposed to bacterial infection and heat stress (37 °C). This coincided with exacerbated increases in TotA and Dpt (Imd downstream pro-survival and antimicrobial genes, respectively) expression compared to controls. Supplementation of IMI-exposed D. melanogaster with Lactobacillus plantarum ATCC 14917 mitigated survival deficits following Serratia marcescens (bacterial pathogen) septic infection. These findings support the insidious toxicity of neonicotinoid pesticides and potential for probiotic lactobacilli to reduce IMI-induced susceptibility to infection.

IMI exposure results in an increased abundance of indigenous Acetobacter and Lactobacillus spp. in D. melanogaster. Xenobiotics, including environmental toxins such as pesticides, often come into direct contact with the intestinal microbiota via oral exposure, and can alter its composition and/or function 35,36 . Thus, we sought to determine if IMI exposure could alter the composition of dominant bacterial genera in D. melanogaster. Third-instar WT Canton-S larvae reared on 10 µM IMI-containing media harboured significantly more Acetobacter and Lactobacillus spp. (t tests, t = 5.933, df = 28, P < 0.0001 and t = 6.734, df = 28, P < 0.0001, respectively) compared to vehicle controls (Fig. 2). However, there was no significant difference in the ratio of Acetobacter to Lactobacillus spp. between IMI-exposed and vehicle flies. This suggested that IMI could be affecting the innate immune function of D. melanogaster and thus their ability to regulate microbiota populations.
It was found that antibiotic-treated WT flies exposed to lethal concentrations of imidacloprid (100 µM) had significantly increased overall survival (log-rank [Mantel-Cox], chi-square = 25.54, P < 0.0001) and less early time point deaths (Gehan-Breslow-Wilcoxon test, chi-square = 27.09, P < 0.0001) in comparison to WT vehicle flies exposed to imidacloprid but not treated with antibiotics (Supp. Fig. 1). Rel −/− mutants exposed to 100 µM imidacloprid followed a similar trend with antibiotic treated flies having significantly increased overall survival (log-rank [Mantel-Cox], chi-square = 11.75, P = 0.0006) and reduced early time point deaths (Gehan-Breslow-Wilcoxon test, chi-square = 10.21, P = 0.0014) compared to flies not treated with antibiotics. These results suggested that artificially regulating microbial populations could mitigate the effects of IMI.
there was a significant upregulation of TotA expression in IMI-exposed WT Canton-S flies subjected to 37 °C heat stress compared to vehicle WT Canton-S flies at 37 °C (two-way ANOVA, P < 0.0001; Fig. 4B). These results suggested exposure to IMI reduced overall survival and exacerbated Imd pathway activation during heat stress. IMI-exposed D. melanogaster are more susceptible to septic infection with S. marcescens. Since the Imd pathway is critical to responding to Gram-negative bacterial infections 17 , similar comparative experiments were performed to elucidate the effect of IMI exposure on pathogenicity of S. marcescens NCIMB 11782 to D. melanogaster (common infection model). WT Canton-S and Rel −/− flies that were subjected to septic infection displayed median survivals of 8.5 and 4 h, respectively. Rel −/− flies that were infected had significantly decreased overall survival (log rank [Mantel-Cox], chi-square = 46.26, df = 1, P < 0.0001) and more early time point deaths (Gehan-Breslow-Wilcoxon test, chi-square = 42.63, df = 1, P < 0.0001) compared to infected WT Canton-S flies (Fig. 5A). Alternatively, IMI-exposed WT Canton-S subjected to septic infection exhibited an intermediate phenotype with a median survival of 7 h. Notably, infected IMI-exposed WT Canton-S flies had significantly reduced overall survival (log rank [Mantel-Cox], chi-square = 20.84, df = 1, P < 0.0001) and more early time point deaths (Gehan-Breslow-Wilcoxon test, chi-square = 16.87, df = 1, P < 0.0001) compared to infected vehicle-exposed WT Canton-S flies (Fig. 5A). IMI-exposed adult WT Canton-S flies were found to have significantly increased pathogen loads at 3 and 6 h (t tests, t = 3.306, df = 16, P = 0.0045 and t = 2.472, df = 16, P = 0.0250, respectively) after S. marcescens NCIMB11782 septic infection compared to infected vehicle-exposed WT Canton-S flies (Fig. 5B). To evaluate if oral infection resulted in similar trends, third-instar larvae were orally infected with S. marcescens Db11. IMI exposure resulted in significantly increased pathogen burden at 6, 12, and 24 h (t tests, t = 2.083, df = 35, P = 0.0446 and t = 5.629, df = 24, P < 0.0001, and t = 2.504, df = 23, P = 0.0198 respectively; Fig. 5C) in WT Canton-S larvae relative to vehicle-exposed controls. IMI-exposed WT Canton-S flies that were subjected to septic infection with S. marcescens NCIMB 11782 exhibited significantly elevated TotA gene expression compared to WT Canton-S flies that were infected but not exposed to IMI (two-way ANOVA, P = 0.0046; Fig. 5D). Likewise, Dpt (Imd downstream antimicrobial peptide) expression was also significantly increased in infected WT Canton-S flies exposed to IMI compared to infected WT Canton-S flies not exposed to IMI (two-way ANOVA, P = 0.0045; Fig. 5E). No significant difference was seen in Dpt expression between 10 µM IMI-exposed and vehicle-fed D. melanogaster (Supp. Fig. 2). These results suggested IMI exposure increased susceptibility to a bacterial pathogen and exacerbated IMI-induced Imd pathway activation in response to infection.
Lp39 can tolerate, but not bind or metabolize IMI. Several species of Lactobacillus were screened to identify a candidate that did not directly interact with IMI. All Lactobacillus species that were tested failed to demonstrate any notable metabolism or binding of IMI (0.1 mg/ml ≈ 391.1 µM) during 24 h of co-incubation (Fig. 6A). It was determined Lp39 could grow unimpaired at high concentrations of IMI (1 mg/mL ≈ 3911 µM) with no significant differences in growth compared to vehicle (Fig. 6B). Thus, Lp39 was chosen for further investigation due to its probiotic and immunostimulatory properties in both D. melanogaster and honey bees 37,38 . 4. IMI exposed D. melanogaster respond similarly to heat stress as Imd pathway mutants. (A) Survival curves for newly eclosed Imd pathway mutants (Rel −/− ) and WT Canton-S exposed to heat stress (37°) with or without concurrent exposure to IMI. Data are displayed from at least 3 independent experiments (15-25 flies reared in separate food vials on different occasions for each experiment). Statistical analyses are representation of comparisons made to heat-stressed WT Canton-S controls using the log-rank (Mantel-Cox) test. (B) TotA gene expression of newly eclosed D. melanogaster that were heat stressed (37°) with or without concurrent exposure to IMI compared to non-heat stressed (25°) controls. All samples were taken 6 h after experimental start time. Gene expression was quantified using RT-qPCR and is relative to vehicle flies not exposed to heat stress. Means ± standard deviations (two-way ANOVA) from 3 biological replicates (each consisting of 10 flies) with triplicate technical repeats are shown. ****p < 0.0001.
Scientific RepoRts | 7: 2703 | DOI:10.1038/s41598-017-02806-w Lp39 supplementation mitigates IMI-induced D. melanogaster survival deficits following infection. WT Canton-S flies that were subjected to septic infection with S. marcescens NCIMB 11782 had a median survival of 2.5 h, which was increased to 7 h when supplemented with Lp39. Notably, flies in these experiments died more rapidly to infection in comparison to earlier septic infection experiments. These discrepancies are believed to be due to uncontrollable biological variations that occur in different batches of flies. Importantly though, trends between IMI-exposed infected flies and infected vehicle-fed controls remained similar to previous findings. Lp39 supplementation significantly increased overall survival (log rank [Mantel-Cox], chi-square = 33.09, df = 1, P < 0.0001) and reduced early time point deaths (Gehan-Breslow-Wilcoxon test, chi-square = 40.71, df = 1, P < 0.0001) in infected flies (Fig. 7A). Similarly, IMI-exposed WT Canton-S flies had a median survival of 2 h after septic infection, which increased to 4 h when supplemented with Lp39. Lp39 significantly increased overall survival (log rank [Mantel-Cox], chi-square = 35.77, df = 1, P < 0.0001) and reduced early time point deaths (Gehan-Breslow-Wilcoxon test, chi-square = 28.66, df = 1, P < 0.0001) in IMI-exposed and infected flies. These findings suggested that Lp39 reduced D. melanogaster susceptibility to infection regardless of concurrent IMI exposure, although to a lesser extent (Fig. 7A). ± standard deviation (unpaired, two-tailed t-tests) of 9 biological replicates (n = 36 for each group). (C) Pathogen load of S. marcescens Db11 during oral infection of WT larvae was determined by plating surface-sterilized whole larvae homogenates on LB with 100 μg/ml streptomycin medium. CFU per larvae obtained at each time point represents the mean ± standard deviation (unpaired, two-tailed t-tests) of 9 biological replicates for each time point (n = 36 total for each group). (D,E) TotA and Dpt gene expression of newly eclosed WT flies that were subjected to septic injury with or without S. marcescens NCIMB 11782 infection and with or without concurrent exposure to IMI. All samples were taken 6 h after experimental start time. Gene expression was quantified by RT-qPCR and is relative to vehicle flies that were subjected to septic injury with a sterile needle. Means ± standard deviations (two-way ANOVA) from 3 biological replicates (each consisting of 10 flies) with triplicate technical repeats are shown. *p < 0.05 **p < 0.01, ***p < 0.001, ****p < 0.0001.
Scientific RepoRts | 7: 2703 | DOI:10.1038/s41598-017-02806-w To determine the effect of Lp39 supplementation on Imd pathway activity, Dpt-RFP flies were used to determine Dpt gene expression. Dpt-RFP flies orally-supplemented with Lp39 had significantly increased expression of Dpt compared to vehicle-fed Dpt-RFP flies (two-way ANOVA, P < 0.0001; Fig. 7B). No significant difference was seen in Lp39-supplemented negative control WT Canton-S flies compared to vehicle (two-way ANOVA, P = 0.9992). Furthermore, qualitative observation of dissected midguts demonstrated increased Dpt-RFP expression in flies that had been orally-supplemented with Lp39 relative to vehicle-fed controls (Fig. 7C).

Discussion
This study demonstrated that the Imd pathway was necessary for D. melanogaster larvae and adult pro-survival in response to IMI. Relish loss-of-function (Rel −/− ) flies and larvae had significantly increased susceptibility to IMI-induced toxicity, compared to WT Canton-S controls. In addition, flies and larvae with loss-of-function genotypes in factors upstream of Relish, including Hop2 −/− , Upd1 −/− , and Upd2/3 −/− , displayed significantly increased susceptibility to IMI compared to their respective genetic backgrounds. The honey bee has several homologues with high amino acid sequence similarity to the D. melanogaster Relish protein such as the predicted nuclear factor NFkB p110 subunit isoform X1 (XP_006562282.1), dorsal (NP_001011577.1), and Relish itself (ACT66913.1). These findings and the highly-conserved nature of innate immunity suggest that translation of findings using this D. melanogaster model are relevant to honey bees and will be amenable for future investigation of pesticide interventions in honey bees 18,39 .
Neonicotinoid pesticides, such as IMI, have been shown to downregulate NF-kB insect immune responses at sub-lethal concentrations due to upregulation of the NF-kB negative regulator gene CG1399 present in D. melanogaster 12 . Compared to non-exposed controls, WT Canton-S flies reared on IMI experienced significantly increased mortality when exposed to 37 °C heat stress or S. marcescens septic infection (Figs 4A and 5A, respectively). WT Canton-S larvae and flies reared on IMI also demonstrated significantly increased S. marcescens pathogen burden following oral and septic infection relative to controls, respectively. This increased pathogen burden in IMI exposed WT Canton-S flies was associated with significantly increased Dpt expression (antimicrobial peptide), but it was not protective against early mortality from septic infection. This finding is explained, in part, by the fact that S. marcescens has been shown to resist innate immune response in D. melanogaster 30 . Furthermore, flies reared on IMI demonstrated significant upregulation of the JAK/STAT stress response gene TotA following heat stress and S. marcescens septic infection (Figs 4B and 5D). TotA induction is critically important for responding to a variety of stressful stimuli such as bacterial infection and heat stress 21,22 . Imd-activation and overall stress was synergistically higher when IMI and other harmful stimuli were encountered simultaneously by D. melanogaster. These results are supported in other insect studies, which have shown reduction in melanotic encapsulation of IMI-exposed Asian longhorned beetles (Anoplophora glabripennis) infected with the entomopathogenic fungus, Metarhizium brunneum. In addition, increased microsporidia Nosema apis pathogenicity 40,41 and deformed wing virus replication 12 have been reported in neonicotinoid-exposed honey bees.
Although IMI has been shown to alter microbial ecosystems in soil 42,43 , the effect of neonicotinoids on the insect intestinal microbiota has yet to be elucidated. This study demonstrated that WT Canton-S flies exposed to IMI had significantly increased absolute abundance of intestinal Acetobacter and Lactobacillus genera. These findings are suggestive of a potentially altered immune function, as commensal-gut mutualism is dependent on antimicrobial peptide regulation by the innate immune system 32,44 . Since the gut ecosystem is important for regulating numerous aspects of insect physiology including growth 19 , lifespan 45 , and mating behavior of insects 46 , the consequences of IMI-induced disruption of the insect microbiota merits further study.
Imd pathway activation can prime D. melanogaster to mitigate infection 20, 47 , thus we hypothesized that targeting this pathway with beneficial microorganisms could mitigate dysregulated immunity observed in D. melanogaster during IMI challenge. Administration of probiotic L. plantarum has been shown to increase survival, lessen gut epithelial damage, and reduce bacterial translocation associated with S. marcescens infection 48 . The present finding that oral Lp39 supplementation significantly mitigated IMI-induced susceptibility of WT Canton-S flies to S. marcescens septic infection mortality relative to unsupplemented controls, suggest that Lp39-mediated protection was likely host-mediated, as there would have been minimal direct interaction between S. marcescens (septic infection) and Lp39 (oral supplementation). Interestingly, Lp39 supplemented WT Canton-S flies exposed to IMI also displayed significantly better overall survival and less early time point deaths compared to flies not exposed to IMI following S. marcescens infection. This could be partially explained by the finding of significantly increased expression of the Imd pathway response antimicrobial gene, Dpt, following Lp39 supplementation. This aligns with past studies demonstrating the immunogenic potential of lactobacilli DAP-type peptidoglycan in modulating the Imd pathway [49][50][51] .
Since the honey bee genome has about half as many glutathione-S-transferases, cytochrome P450 (Cyp) monooxygenases, and carboxyl/cholinesterases as D. melanogaster 52 , it is challenging to directly extrapolate our findings. This is important since these enzymes are involved in detoxification of IMI. In particular, the global spread of Cyp6g1 overexpression in D. melanogaster has resulted in a high level of resistance to a variety of insecticides including IMI 53 . This study determined that 10 µM IMI (approximately 50x higher than what is generally observed in agricultural settings) 8 represented a sub-lethal dosage in D. melanogaster. However, the observed deficits of IMI-exposed D. melanogaster to additional stressors (e.g. infection, heat) are likely relevant to honey bees at environmental concentrations due to their previously reported sensitivity to neonicotinoid pesticides 3,5,10 .
In summary, this study has shown that: 1) the Imd pathway was necessary for promoting insect survival in response to IMI toxicity, 2) IMI exposure altered the insect microbiota, and 3) IMI exposure exacerbated insect susceptibility to both heat stress and bacterial infection. Furthermore, Lp39 mitigated IMI-induced susceptibility to a bacterial pathogen in D. melanogaster. These findings provide evidence for the insidious insect toxicity of neonicotinoid pesticides in combination with other environmental stressors and illustrate the prophylactic potential of lactobacilli to combat some of the predicted causes of colony collapse disorder. The extension of these findings to honey bees is promising given that colony supplementation with lactobacilli is affordable, feasible, and has already been shown to benefit honey bee colony growth 54 , microbiota composition 55 , and antimicrobial defense 18,20 . IMI-exposed D. melanogaster survival assays. Twenty-25 newly eclosed flies were anesthetized with CO 2 and randomly transferred into standard vials containing IMI-supplemented (10, 50, and 100 µM doses) food or vehicle (DMSO) at mid-light cycle 56 . Following anesthetization, flies were confirmed to be alive, and subsequently monitored daily (9 AM) for survival. Flies were transferred to fresh media every 3 d.
Culture-based D. melanogaster microbiota enumeration. D. melanogaster were surface sterilized with 70% ethanol and homogenized in 0.01 M PBS using a motorized pestle. Homogenates were then serially diluted and spot plated onto MRS and ACE (Acetobacter growth media containing 3 g/L proteose peptone no. 3 [catalog number: 211693; BD Difco], 5 g/L yeast extract [catalog number: 212750; BD Difco], and 25 g/L D-mannitol [catalog number: M9647; Signma-Aldrich]) agar, followed by anaerobic incubation at 37 °C for 48 h or aerobic incubation at 37 °C for 48 h to enumerate Lactobacillus spp. and Acetobacter spp., respectively. Colonies forming units on MRS and ACE were enumerated and confirmed to be Lactobacillus and Acetobacter spp., respectively, via microscopy analysis and 16S rRNA gene sequencing using the Applied Biosystems 3730 Analyzer platform at the London Regional Genomics Centre (Robart's Research Institute, London, Canada).

D. melanogaster eclosion assays.
Drosophila melanogaster eggs were collected on grape agar plates 57 .
For each vial ten first-instar larvae were transferred to D. melanogaster media with or without IMI (1 or 10 µM), incubated at the aforementioned conditions, and monitored daily for up to 16 d for eclosion. D. melanogaster heat stress assay. Newly eclosed female flies reared on media containing 10 μM IMI or vehicle were anesthetized with CO 2 and randomly transferred into fresh vials containing the same media they were reared on previously. Following anesthetization, flies were confirmed to be alive and then subsequently exposed to lethal heat stress (37 °C) and monitored hourly for survival.
Scientific RepoRts | 7: 2703 | DOI:10.1038/s41598-017-02806-w Adult D. melanogaster septic infection. Newly eclosed (within 2-4 d) female flies that were reared from eggs on media containing vehicle or 10 μM IMI were infected by septic pinprick (approximately 100 bacteria per fly) as described previously 58 . Briefly, overnight cultures of S. marcescens NCIMB 11782 were washed with 0.01 M phosphate-buffered saline (PBS). Washed cultures were pelleted by centrifugation at 5000 g for 5 min, and supernatants discarded. Glass microinjection needles were made using a glass micropipette puller and sterilized with 70% ethanol before being used. Microinjection needles were dipped in the S. marcescens NCIMB 11782 bacterial pellet prior to pin pricking D. melanogaster at the sternopleural plate of the thorax just under the attachment sites of the wings. Alternatively, vehicle groups were subjected to septic injury via pinprick with a sterile needle. Following infection, groups of 10 flies were transferred to new tubes containing the same media they were raised on and monitored hourly for survival at 25 °C. Flies collected 0 h, 1 h, 3 h and 6 h post-infection were surface sterilized with 70% ethanol, homogenized, and plated on LB agar for pathogen load. Colony forming units were enumerated following 48 h incubation at 22 °C (S. marcescens NCIMB 11782 colonies appear red under these conditions).
Larval D. melanogaster oral infections. Third-instar larvae, reared from eggs on media containing vehicle or 10 μM IMI, were orally infected with S. marcescens Db11 as described previously 60 . Larvae collected at 6 h, 12 h, and 24 h post-infection were surface sterilized with 70% ethanol, homogenized, and then plated on LB agar with 100 μg/mL streptomycin for overnight incubation at 37 °C to enumerate pathogen load.
Lactobacilli-mediated IMI metabolism/binding assay. Overnight bacterial subcultures were pelleted at 5000 g for 10 min. Pellets were washed and re-suspended in 50 mM HEPES (pH 6.8). Bacterial-buffer or buffer-alone solutions were incubated with IMI (100 ppm) protected from light at 37 °C for 24 h (unless otherwise stated) with shaking (200 rpm). Cultures were pelleted and supernatant collected then analyzed for IMI levels using liquid-chromatography mass spectrometer consisting of an Agilent 1290 Infinity High Performance Liquid Chromatography (HPLC) coupled to a Q-Exactive Orbitrap mass spectrometer (Thermo Fisher Scientific, Waltham, USA) with a heated electrospray ionization (HESI) source. Two μL of each sample and standard were injected into a ZORBAX Eclipse plus C18 2.1 × 50 mm × 1.6 micron column. Mobile phase (A) consisted of 0.1% formic acid in water and mobile phase (B) consisted of 0.1% formic acid in acetonitrile. The initial composition of 100% (A) was held constant for 0.5 minutes and decreased linearly to 0% over 3.5 minutes. Mobile phase A was held at 0% for 1.5 minutes then returned to 100% over 30 seconds. The system was re-equilibrated at 100% for 1 minute, for a total analysis time of 6.50 minutes.
The HESI source was operated under the following conditions: nitrogen flow of 17 and 8 arbitrary units for the sheath and auxiliary gas, respectively. Probe temperature and capillary temperature were 450 °C and 400 °C, respectively. Spray voltage was 3.9 kV. The S-Lens was set to 45. A full MS and MS/MS scanning were used to monitor the parent mass and fragment ions respectively. A full MS scan between the ranges of 50-400 m/z in positive mode at 70,000 resolution (AGC target and maximum injection time were 3e6 and 250 ms respectively) was utilized. The MS/MS was set to 17,500 resolution and normalized collision energy set to 35 (AGC target and maximum injection time of 5 × 10 5 and 65 ms, respectively).
Bacterial imidacloprid tolerance assay. Overnight Lp39 cultures were sub-cultured (1:100 dilution) into 96 well plates (Falcon, catalog number: 351177) containing MRS broth with or without the addition of IMI or vehicle (DMSO). Plates were incubated at 37 °C and read every 30 min for 24 h at a wavelength of 600 nm using a Labsystems Multiskan Ascent microplate reader.
Adult D. melanogaster septic infection mitigation assays with Lp39 supplementation. Overnight cultures were pelleted at 5000 g for 10 min, washed with 0.01 M PBS, and concentrated 10-fold with 0.01 M PBS. Drosophila melanogaster food was supplemented with 100 μL (10 9 CFU) Lp39 when experimentally appropriate and allowed to air dry prior to usage. Drosophila melanogaster eggs were seeded on food media supplemented with Lp39 or vehicle (100 µL 0.01 M PBS) and reared until adult hood on this same media. Initial Lp39 supplementation occurred the day of egg seeding and additional supplementation followed every third day thereafter. Septic infections of supplemented flies were performed under the aforementioned experimental conditions. Scientific RepoRts | 7: 2703 | DOI:10.1038/s41598-017-02806-w being the most stably expressed reference gene (compared to GAPDH1, TBP, RpL13A, and ATUB84B) under experimental conditions in this study, and was thus used as the internal standard for normalization as per MIQE guidelines 61 .
cDNA was diluted 10-fold and used for qPCR reactions with the Power SYBR Green kit (Applied Biosystems, catalog number: 4368702). Reagent volumes for 20 µL reactions consisted of 10 µL Power SYBR (2x), 0.4 µL forward primer (10 µM stock), 0.4 µL reverse primer (10 µM stock), 4.2 µL nuclease-free H 2 O, and 5 µL cDNA. Reaction conditions were 95 °C for 10 min followed by 40 cycles of 95 °C for 15 sec and 60 °C for 1 min. qPCR was performed on a 7900HT Sequence Detection System (Applied Biosystems) and analyzed using SDS RQ 6.3 manager software (Applied Biosystems) and relative gene expression was calculated using the 2 −ΔΔct method 62 . PCR amplification was confirmed via melt-curve dissociation analyses to verify product size and check for non-specific amplification.
Fluorescent microplate reader. Newly eclosed Dpt-RFP D. melanogaster were exposed to cotton gauze soaked in 2 mL 5% sucrose containing 10 9 CFU of Lp39 or vehicle for 15 h at 25 °C. WT Canton-S flies exposed to identical experimental conditions were used as negative controls. Samples containing 10 flies were homogenized in 400 µL of 0.01 M PBS on ice using a BioSpec 3110BX Mini Beadbeater (Fisher Scientific, catalog number: NC0251414) with silica beads. Homogenates were centrifuged at 12,000 g for 20 minutes at 4° C. Supernatants (200 µL/sample) were added to Corning 96-well solid black polystyrene microplates for Dpt expression of the red fluorescent protein (RFP) reporter, mCherry. Plates were measured at an excitation wavelength/bandwidth of 587/9 nm and emission wavelength/bandwidth of 645/20 nm 63 using a BioTek Synergy2 microplate reader (Fisher Scientific, catalog number: 36-101-5201).

Statistical analyses.
All statistics were performed using GraphPad Prism 7.0 software. Data sets with unique values were tested for normality using the omnibus-based Shapiro-Wilk test, while data set with ties (two or more identical values) were tested for normality using the D' Agostino-Pearson test. Normally distributed data were statistically compared with unpaired, two-tailed t tests, one-way or analysis of variance (ANOVA) or two-way ANOVA as indicated. ANOVA tests were complemented with Tukey's (for data with one categorical variable) or Sidak's (for data with two categorical variables) multiple comparison tests when appropriate. Nonparametric data sets were statistically compared using Mann-Whitney and Kruskal-Wallis tests (with Dunn's multiple comparisons) when appropriate. Mantel-Cox tests were used to analyze overall D. melanogaster survival. Alternatively, Gehan-Breslow-Wilcoxon tests were used to analyze D. melanogaster survival with emphasis on early time point events. Multiple comparisons for Mantel-Cox and Gehan-Breslow-Wilcoxon tests were performed using the Bonferroni method.