Many mycophagous Drosophila species have adapted to tolerate high concentrations of mycotoxins, an ability not reported in any other eukaryotes. Although an association between mycophagy and mycotoxin tolerance has been established in many Drosophila species, the genetic mechanisms of the tolerance are unknown. This study presents the inter- and intraspecific variation in the mycotoxin tolerance trait. We studied the mycotoxin tolerance in four Drosophila species from four separate clades within the immigrans-tripunctata radiation from two distinct locations. The effect of mycotoxin treatment on 20 isofemale lines per species was studied using seven gross phenotypes: survival to pupation, survival to eclosion, development time to pupation and eclosion, thorax length, fecundity, and longevity. We observed interspecific variation among four species, with D. falleni being the most tolerant, followed by D. recens, D. neotestacea, and D. tripunctata, in that order. The results also revealed geographical variation and intraspecific genetic variation in mycotoxin tolerance. This report provides the foundation for further delineating the genetic mechanisms of the mycotoxin tolerance trait.

Fly isofemale lines

Four species were included in this study: D. falleni, D. recens, D. neotestacea, and D. tripunctata. Adult flies were collected by net sweeping on fermented banana baits, tomato baits, and mushroom baits over the summer months of 2017-2019 from two distant locations: Great Smoky Mountain National Park near Gatlinburg, TN (hereafter referred to as GSM) and Little Bay de Noc in Escanaba in The Upper Peninsula of Michigan (hereafter referred to as ESC). These two sites are approximately 1400 km apart. Multiple sites were used for fly collection within each location spanning over 3-5 square kilometers. The species and the sex of the captured flies were identified, and isofemale lines were set up by adding one wild-caught female with one wild-caught male from the same species and location and collecting their progeny (David 2005). The established isofemale lines were maintained on a diet of Carolina Biological Formula 4- 24 Instant Drosophila Medium supplemented with finely ground, freeze-dried Agaricus bisporus mushrooms (Oregon mushrooms, OR) at a ratio of 33.28:1 w/w, and a dental roll was added to the food vial as a pupation site. The standard conditions for maintenance and experiments were 22°C and a 14 h:10 h (L:D) photoperiod at 60% humidity. The authors note here that the isofemale lines were maintained in the laboratory for at least over a year before the experiments were conducted.

2.2 Mycotoxin tolerance assays

Basic food was prepared by mixing 28.3 g freeze-dried A. bisporus mushrooms (Oregon mushrooms, Oregon) with 941.9 g Carolina 4-24 Instant Drosophila Medium and grinding them together into a fine powder. For mycotoxin tolerance assays, clean glass vials were filled with 250 mg of basic food. The natural-toxin mix was provided by Dr. Clare Scott-Chialvo (Scott Chialvo et al. 2020), which contained methanol as eluent. To account for this methanol, one mL of 0.56% methanol solution was added to the control vials containing 250 mg of basic food. The mycotoxin vials were prepared by adding 1 mL of the natural-toxin mix (100 µg/mL of known amatoxins) to the vials containing 250 mg of basic food. Both control and mycotoxin vials were weighed and subjected to vacuum evaporation for 96 hours at room temperature to remove methanol from the vials. The loss in weight (in grams) in the vials was replenished with the appropriate amount (in mL) of sterile distilled water. The optimal duration of vacuum evaporation was identified using preliminary studies and 96 hours of vacuum evaporation showed survivorship that was comparable to vials without methanol.

Water agar plates were prepared using 15 g Bacto Agar (Sigma Aldrich) in 500 mL of distilled water and adding Tegosept to a final concentration of 0.1% and poured into 30 mm Petri-plates. These plates snugly fit the plastic bottles that were used to make egg- laying chambers. Tiny holes were punched into these plastic bottles for aeration. Equal amounts of dry yeast and freeze-dried mushroom powder were mixed together with autoclaved distilled water to prepare a paste (prepared fresh daily). A drop of this paste was applied to the water agar plate. Recently eclosed males and females of each isofemale line were transferred to egg-lay chambers and allowed to oviposit at 22°C and a 14 h:10 h (L:D) photoperiod at 60% humidity. The next day, the plates were replaced with fresh plates, and the water agar plates with oviposited eggs were allowed to hatch at 22°C and a 14 h:10 h (L:D) photoperiod at 60% humidity. The hatched first-instar larvae were used for the experiments. Pilot studies were performed to identify the optimal larval density for each species. As a result, fifteen first-instar larvae were added to each vial in the case of D. falleni, D. recens, and D. tripunctata, whereas 20 first-instar larvae were added to each vial for D. neotestacea. The experiments were conducted in triplicates. Each experiment was conducted on consecutive days to generate three replicates for each of the ten isofemale lines/location/species for two treatments (control and mycotoxin).

2.2.1 Development time, thorax length measurements, and survival

The vials were checked daily to record the time to pupation, survival to pupation, time to eclosion, and survival to eclosion. The eclosed flies were collected within 24 hours by light CO 2 anesthesia, sexed, and placed laterally to measure the thorax length. The thorax’s anterior margin length to the scutellum's posterior tip was measured and recorded as the thorax length. The thorax length of the eclosed flies was measured to the nearest 0.025 mm with an Olympus SZX16 dissection microscope fitted with an Olympus DP72 camera, using the ImageScan software (Hasson et al. 1992). The eclosed females were used for the fecundity assays, and the eclosed males were used for the longevity assays. The experiments were terminated after ensuring that no new flies had emerged for four consecutive days.

2.2.2 Fecundity assays

Only female flies were used for the fecundity assays. Female flies eclosed from the mycotoxin tolerance assay vials were labeled appropriately and maintained individually in food vials for three days as virgins. They were then transferred individually into a fresh food vial with three 3-day-old virgin males from the laboratory stocks of the same isofemale line. These parent flies were transferred to a new vial every three days. After 15 days, the adult flies were removed. The offspring of the females that survived the full 15 days were counted to provide an estimate of fecundity. We note that this assay cannot be used to evaluate egg-to-adult survival (Dyer and Hall 2019).

2.2.3 Longevity assays

Only male flies were used for the longevity study. Male flies eclosed from the mycotoxin tolerance assay vials were maintained individually in tiny 5-mL glass vials containing approximately 250 mg of the basic food used to create the mycotoxin tolerance assay vials. The vials were checked every alternate day to record any dead flies, and the remaining flies were transferred to fresh food vials every 2-3 days.

2.3 Statistical Analyses

All statistical analyses were done using R version 3.6.1 (https://www.r- project.org/foundation/) and R Studio version 2021.09.2+382 (https://www.rstudio.com/). We used the linear-mixed model (LMM) and the generalized linear-mixed effects model (GLMM), implemented in R package 'lme4' (Bates et al. 2015), to determine the independent variables that can explain the variation in survival, development time, body size, fecundity, and longevity. We modeled pupal and survival to eclosion using a binomial linear-mixed model with the logistic link function. We used the linear-mixed model to model development time, fecundity, longevity, and thorax lengths. The development time and thorax lengths were analyzed for each sex separately. To detect whether mycotoxin tolerance shows interspecific variation, we first fitted a GLMM that includes the main effects, the two-way interactions, and the three-way interaction of the species, the treatment, and the location as the fixed effects. The likelihood test (LRT) was used to test if the three interaction was significant. The final model includes the main effects and the two-way interactions of the species, the treatment, and the three-way interaction only if the three-way interaction is significant (in other words, the p-value from the LRT for the three-way interaction is less than 0.05). In all models, the isofemale lines and the replicate vials were included as the random effects. To check the sufficiency of the model, the scatter plots of the deviance residuals against the predicted values were generated, and the dispersion parameter was estimated based on the ratio of the sum of squared deviance residuals and the degrees of freedom of the model if the binomial linear-mixed model was used. To evaluate the effect of toxin treatment, we fitted either a binomial linear-mixed or linear-mixed model to assess whether the treatment affects the seven gross phenotypes. In all models, the main effect of the treatment was the only fixed effect, and the isofemale lines and the replicate vials were included as the random effects. The analysis was conducted for each of four species and seven gross phenotypes separately. Among seven gross phenotypes, the development time of eclosion and the thorax length of eclosion were analyzed for the males and the females separately. Therefore, 36 p-values were obtained. To account for the multiple testing adjustment, both the p-value adjusted using the Bonferroni correction and the false discovery rate (FDR) calculated with the Benjamini-Hochberg method (Benjamini and Hochberg, 1995) were presented. The FDR < 0.05 was used as a cut-off for significant results. To identify the extent of tolerance for each isofemale line, the binomial linear-mixed model was performed on each isofemale line with the replicate vial as a random effect, and the lines were segregated based on their p-values. High-tolerance lines were identified as those in which no significant difference in survival between the control and the mycotoxin treatments was observed or where the survival was significantly higher in the mycotoxin treatment. Isofemale lines with significantly low survivorship on the mycotoxin treatment (p-value < 0.05) were categorized as low-tolerance lines. For the scope of this study, high tolerance is defined as the ability of an isofemale line to survive in the presence of the natural-toxin mix (100 µg/mL of known amatoxins). For intraspecific variation, before model fitting, we pruned the data to exclude isofemale lines where only one data point was observed per treatment. This exclusion allowed us to eliminate data that could not estimate variation within an isofemale line. We assessed whether the main effects: a) isofemale line, b) treatment (presence or absence of mycotoxin), c) location, and d) interactions between the main effects affect the seven gross phenotypes (survival to pupation, survival to eclosion, development time to pupation and eclosion, thorax length, fecundity, and longevity) in each species. The isofemale lines and the replicate vials were included in the analysis as random effects.

R version 3.6.1 (https://www.r-project.org/foundation/)

R Studio version 2021.09.2+382 (https://www.rstudio.com/).

lme4 (Bates et al. 2015)