Overexpression of an antioxidant enzyme improves male mating performance after stress in a lek-mating fruit fly

Abstract

In many species, courtship displays are reliable signals of male quality, and current hypotheses suggest that these displays allow females to choose males with high cellular function. Environmental stressors generate excess reactive oxygen species (ROS) that impair cellular function, and thus antioxidant pathways that remove ROS are probably critical for preserving complex sexual behaviours. Here, we test the hypothesis that enhanced antioxidant activity in mitochondria preserves mating performance following oxidative stress. Using a transgenic approach, we directly manipulated mitochondrial antioxidant activity in the Caribbean fruit fly, Anastrepha suspensa, a lek-mating species with elaborate sexual displays and intense sexual selection that is also a model for sterile insect technique programmes. We generated seven transgenic lines that overexpress mitochondrial superoxide dismutase (MnSOD). Radiation is a severe oxidative stressor used to induce sterility for sterile insect programmes. After radiation treatment, two lines with intermediate MnSOD overexpression showed enhanced mating performance relative to wild-type males. These improvements in mating corresponded with reduced oxidative damage to lipids, demonstrating that MnSOD overexpression protects flies from oxidative stress at the cellular level. For lines with improved mating performance, overexpression also preserved locomotor activity, as indicated by a laboratory climbing assay. Our results show a clear link between oxidative stress, antioxidant capacity and male performance. Our work has implications for fundamentally understanding the role of antioxidants in sexual selection, and shows promise for using transgenic approaches to enhance the field performance of insects released for area-wide pest management strategies and improving performance of biological control agents in general.

1. Introduction

The ability to accurately assess mate quality is a critical function for animals. Many species have evolved ornaments and other elaborate behaviours as indicators of male quality [1]. The current understanding of condition-dependent traits is that they provide an honest signal of physiological and cellular function [2]. Reactive oxygen species (ROS) are essential cell signalling molecules, but an excess of ROS compromises cellular function and causes oxidative stress [3]. Many environmental challenges (e.g. temperature, desiccation, hypoxia-reperfusion) lead to ROS accumulation [3,4], which in turn causes oxidative damage and reduces male quality [5–7]. Mitochondria are the primary source of ROS production, and initial ROS release following stress can trigger a positive feedback loop of sustained ROS release, resulting in long-lasting effects on organismal function [8]. Thus, current hypotheses suggest mitochondrial function is a key mechanism that underscores phenotypic variation in sexual displays [9], and on evolutionary time scales mitonuclear interactions may be critical determinants of life-history traits [10].

Oxidative stress is pervasive in nature, and antioxidant defences play a critical role in maintaining the vitality of cells, ultimately mediating whole-organism performance [1,5–7]. Organisms have evolved a suite of antioxidant defence systems to detoxify ROS and preserve cellular function in the face of an oxidative challenge [11]. In natural systems, current hypotheses about the role of antioxidants in mating performance mostly stem from a large body of work on the role of dietary carotenoids in sexual ornamentation (reviewed by Catoni et al. [12] and Blount [13]). Carotenoids are the main pigments used in vertebrate sexual ornaments and have antioxidant activity, so long-standing hypotheses suggest these carotenoids provide a direct link between antioxidant function and mate choice. However, carotenoids represent a minor component of an animal's total antioxidant capacity [14], and there is conflicting evidence regarding a physiological role for carotenoids in defence against oxidative stress. For example, white canaries that carry a recessive mutation resulting in extremely low levels of tissue carotenoids are no more susceptible to oxidative challenges than wild-type canaries [15]. Thus, endogenous antioxidants that represent the majority of an organism's antioxidant capacity are likely to play a substantial role in mediating condition-dependent sexual traits through their roles in maintaining cellular vitality in stressful environments.

Here, we experimentally test the relationship between antioxidant function, oxidative stress and mating performance in the Caribbean fruit fly (caribfly), Anastrepha suspensa. This species uses a lek-mating system that involves an elaborate series of visual displays, courtship songs, male–male fighting and pheromone release [16]. Male courtship in caribflies is energetically demanding and requires high levels of physiological function [17,18]. Caribflies are amenable to laboratory rearing and genetic transformation [19], allowing for direct manipulation of antioxidant function to test hypotheses regarding the role of endogenous antioxidants in maintaining cellular function and sexual displays following stress. In this species, hormetic-conditioning treatments that improve male mating performance after stress are correlated with increased activity of two endogenous antioxidant enzymes, suggesting these pathways may be involved in maintaining courtship behaviour by attenuating oxidative stress that reduces cellular vitality in stressful environments [20].

Beyond impacts on fundamental understanding of the role of antioxidants and mitochondrial function in the face of stress, our work also has implications for the practical improvement of biological control agents. Sterile insect technique (SIT) is an environmentally friendly pest control strategy in which mass-reared males are sterilized with ionizing radiation and released to suppress wild populations [21]. Caribflies were once controlled with field SIT releases, and although those programmes have ended, A. suspensa is routinely used to evaluate strategies for improving SIT in other pests. Several of their close relatives (e.g. A. ludens) are used in active SIT programmes, and caribflies are also a useful model for other economically important fruit flies as well (e.g. [20,22–24]). The radiation used for sterilization in SIT causes oxidative stress that reduces the mating competitiveness and lifespan of sterilized males, thereby hindering the performance of these biological control agents and increasing economic costs associated with SIT [20,25–28]. Numerous transgenic strategies have been developed to improve SIT, for example, conditional lethality strains that obviate the need for radiation sterilization and genetic sexing strains that eliminate unwanted females [29,30]. Our work stands out from these previous studies because we are specifically focused on transgenic approaches that improve male performance. Overexpression of antioxidants may have additional benefits for SIT and other types of biological control where ionizing radiation is not used, for example, by improving longevity [31] and/or preventing declines in muscular performance [32] in insects used for augmentative biological control. Even if genetic sterilization eventually replaces the need for radiation as a sterilizing agent, any insect released for population control requires successful field performance, and our work demonstrates the potential of using transgenic strategies to target specific deficits in the performance of beneficial released insects.

2. Methods and materials

(a). Insect husbandry and germline transformation

Wild-type and transgenic colonies of the Caribbean fruit fly, Anastrepha suspensa, were maintained under standard laboratory conditions [19]. Larvae were reared on an artificial diet containing, in g l⁻¹, 4 agar, 75 dextrose, 75 torula yeast, 75 wheat germ, 0.6 cholesterol, 1.3 sodium benzoate, 1.3 methyl 4-hydroxybenzoate and 1.0 butyl 4-hydroxybenzoate at 27°C, 14 : 10 L : D, and allowed to pupariate on vermiculite. Adults were fed 3 : 1 sugar : yeast hydrolysate and water ad libitum, and eggs were laid onto red fabric cloths placed on top of the cage.

PiggyBac transformation vectors for mitochondrial superoxide dismutase (MnSOD) (GenBank: MK840869) and Cu/ZnSOD (GenBank: MK840870) were generated using restriction cloning (figure 1a; see details in the electronic supplementary material). The MnSOD transformation vector (500 ng µl⁻¹) and a piggyBac transposase helper plasmid (200 ng µl⁻¹) were co-injected into wild-type embryos as described previously [19]. Additional details on the generation of transgenic lines are found in the electronic supplementary material. qPCR was used to estimate the number of genomic insertion events for each transgenic line. The quantity of DsRed sequence was measured relative to a known single-copy gene, Pros26 [33] using primers DsRed-F, DsRed-R, Pros26-F and Pros26-R (electronic supplementary material, table S1).

Figure 1.

Construct design and molecular characterization of transgenic lines. (a) Design of transgenic construct for overexpressing MnSOD. (b) Transcript abundance of MnSOD in wild-type (WT) and the seven transgenic lines overexpressing MnSOD. The SOD1 line is an additional control line transformed with a vector that contains Cu/ZnSOD in lieu of MnSOD, to demonstrate that overexpression is specific to MnSOD. (c) Enzyme activity of MnSOD in wild-type and transgenic flies. (d) Correlation between enzyme activity and mRNA expression in wild-type and transgenic lines. In (b–d), bars and symbols represent mean expression or activity while whisker bars represent standard error of the mean, n = 4–5 pools of males for each genotype. In (b,c), groups are arranged from lowest to highest expression/activity, and different letters represent significant differences between groups following ANOVA and Tukey's HSD, p < 0.05. (Online version in colour.)

(b). SOD transcript expression

qPCR was used to measure the transcript expression of both MnSOD and cytosolic superoxide dismutase (Cu/ZnSOD) in each line. Expression of Cu/ZnSOD was measured as a control to indicate the specificity of MnSOD overexpression. A single Cu/ZnSOD overexpressing line was also included as a control in the gene expression and enzyme activity experiments to further evaluate the specificity of the transgene. All flies used for transcript abundance and enzyme activity were snap frozen in liquid nitrogen, and each experiment consisted of at least four independent replicates. In all lines, gene expression was measured in both 3-day-old adult males and pharate adult males. In a subsequent experiment, we measured expression in a single line (line 3.9) with robust overexpression to assess the developmental trajectory of overexpression and the functionality of the exogenous Drosophila heat shock promoter. While this line was not one of our focal lines in later experiments (see §3e,f), we elected to use it in these early evaluations of transgene function because it had consistently high mRNA expression and enzyme activity in preliminary experiments. Flies from line 3.9 were sampled as eggs, wandering larvae, pupae, pharate adults and 3-day-old adult males and females. Adult were also heat shocked for 2 h at 37°C, and flies were sampled immediately after the heat shock and after 2 h recovery at 25°C to test the extent to which heat shock could further increase MnSOD expression.

RNA was extracted from pools of four flies using the Direct-zol RNA MiniPrep kit (Zymo) according to manufacturer's protocols. RNA purity and concentration were measured using a Nanodrop, and the iScript cDNA Synthesis Kit (Bio-Rad) was used to generate cDNA from 1 µg RNA from each sample. Each qPCR reaction contained 2 µl cDNA, 2 µl of each primer (at 250 nM), 10 µl iTaq Universal SYBR Green Mastermix and 4 µl water. The relative quantity of MnSOD and Cu/ZnSOD transcripts was calculated using the 2-ΔCt method, as in [34], with RP49 (GenBank: MK847891) as the reference gene. Primer sequences for MnSOD, Cu/ZnSOD and RP49 can be found in electronic supplementary material, table S1.

(c). SOD enzyme activity

Enzyme activity of MnSOD and Cu/ZnSOD was measured using the Superoxide Dismutase Assay Kit (Cayman Chemical), according to the manufacturer's protocol. Total SOD activity was measured in protein extracts, and MnSOD activity was measured by adding 3 mM KCN to inhibit Cu/ZnSOD. Additional details for the SOD assay can be found in the electronic supplementary material. As with gene expression, enzyme activity was measured in adult and pharate adult males of all lines, as well as across development and following heat shock in line SOD2-3.9 (see details above).

(d). Mate choice tests

We evaluated the precopulatory mating success of transgenic and wild-type males in mate choice tests under laboratory conditions [20]. Copulation success of males was assessed by exposing one wild-type virgin female to two virgin males. Twelve days after eclosion, males were marked with a small dot of water-based paint on the thorax and released into plastic cages the day prior to the experiments. On the following day, females were released into cages with two marked males and observed for up to 4 h during the afternoon. At least 50 pairs spread across at least three cohorts were measured for each of the four combinations tested: (1) untreated transgenic males paired with untreated wild-type males, (2) irradiated transgenic males paired with irradiated wild-type males, (3) irradiated wild-type and transgenic males paired with untreated wild-type males, and (4) untreated wild-type and transgenic males paired with irradiated wild-type males.

(e). Biochemical measurement of oxidative stress

To test the extent to which MnSOD overexpression reduces irradiation-induced oxidative damage, we measured two indices of oxidative damage in males. The thiobarbituric acid reactive substances (TBARS) assay measures the abundance of malondialdehyde (MDA), a product of lipid peroxidation, while the protein carbonyls assay measures oxidative damage to proteins. Assays were conducted according to [20]; see the electronic supplementary material, for additional details.

(f). Climbing assay

We used a climbing assay as a proxy for muscle performance in the two high-performing lines (SOD2-5.2 and SOD2-6.1). Irradiated flies were treated at 70 Gy as pharate adults, 2 days prior to emergence. For the climbing assay, male flies were placed in the bottom of a 33 cm long by 1.9 cm diameter clear acrylic tube, and flies were observed until they either reached the 33 cm mark on the tube or 5 min had expired. For each fly, we scored whether it made it to the top of the tube, and the number of times it fell while crawling. For each group, we measured 10 individual flies, and each fly was tested three times. For each group, we calculated the average number of falls per fly and the proportion of the total of the 30 trials for each line in which a fly made it to the top.

(g). Longevity measurements

The lifespan of irradiated and untreated males was monitored daily to determine the effect of MnSOD overexpression on transgenic flies' longevity. Briefly, 10 newly emerged transgenic or wild-type males were transferred into plastic cages containing five wild-type females, and the number of deaths was recorded daily until all died. During the experiments, flies had free access to hydrolysed protein diet (three parts of sugar: one part of protein) and water. A randomized block design was applied, and at least four blocks with four cages each were carried out to evaluate the longevity of irradiated and untreated males.

(h). Statistical analyses

Statistical analyses were conducted in JMP 12 (SAS), SAS or R. For all experiments, we used α = 0.05. To analyse qPCR data, we conducted ANOVAs on the ΔCt values, and pairwise comparisons were made using t-tests, with a Tukey's multiple comparisons adjustment. Enzyme activity data were analysed with a least-squares regression model with genotype as a fixed effect and experimental block as a random effect. For the heat shock experiment, we fit a linear mixed model with genotype, treatment and their interaction as main effects, with block as a random effect. The developmental enzyme activity data were analysed by fitting a linear model with genotype, developmental stage and their interaction as main effects, with block as a random effect, where a block included all samples analysed together on the same microplate. Mate choice tests were analysed with logistic regression, fitting the proportion mating success of each pairwise comparison as a function of genotype, with block as a random effect. Cox regression was used to compare the longevity of transgenic and wild-type males. To analyse oxidative stress data (TBARS and protein carbonyls), we fitted TBARS or carbonyls levels as a function of irradiation status, genotype, sampling time and their interactions, with block as a random effect. Linear contrasts were used to compare levels of each variable.

3. Results

(a). Generating and characterizing MnSOD overexpression lines

Using piggyBac germline transformation, we initially created 12 transgenic lines, and segregation analysis indicated that six lines carried a single autosomal insertion while one line carried a Y-linked, male-only insertion (electronic supplementary material, table S2). These seven lines were retained for subsequent experiments. After breeding the autosomal lines to homozygosity, real-time quantitative PCR (qPCR) analysis of genomic copy number confirmed the segregation analysis and all lines showed good viability compared to wild-type (electronic supplementary material, figure S1).

MnSOD was significantly overexpressed relative to wild-type in adult males of all transgenic lines (figure 1b), resulting in significantly elevated MnSOD enzyme activity in five of the lines (figure 1c). MnSOD transcript abundance in transgenic lines ranged from 1.8- to 4.5-fold higher than wild-type. MnSOD transcript abundance and enzyme activity were highly correlated (figure 1d), indicating that mRNA derived from the transgene was translated into a functional protein. MnSOD was also significantly overexpressed at the pharate adult stage, the developmental stage that is irradiated for SIT sterilization (electronic supplementary material, figure S2a,b). As in adults, MnSOD transcript abundance and activity were highly correlated in pharate adults (electronic supplementary material, figure S2c), and lines that had high expression and activity as pharate adults also had high expression and enzyme activity as adults (electronic supplementary material, figure S2d,e). Overexpression was specific for the mitochondrial isoform of SOD, as the transgene did not elevate transcript abundance or enzyme activity of the cytosolic Cu/ZnSOD in both adults and pharate adults (electronic supplementary material, figure S2f–i). In line SOD2-3.9, MnSOD was significantly overexpressed at the pupal stage and beyond (electronic supplementary material, figure S3a,b), and overexpression was specific to the mitochondrial version of SOD (electronic supplementary material, figure S3c). While heat shock treatment caused a slight increase in MnSOD mRNA levels (electronic supplementary material, figure S3d), heat shock failed to drive an increase in MnSOD activity (electronic supplementary material, figure S3e), indicating the heat shock response element of the Drosophila hsp70 promoter was not capable of driving higher levels of MnSOD activity in these lines despite high baseline activity. Overall, these experiments indicate that we successfully created lines that overexpress MnSOD, and transgene activity is highest at the times required to protect against damage from sterilizing irradiation even in the absence of additional stimulation of the transgenic promoter by heat shock.

(b). MnSOD overexpression enhances mating performance of irradiated flies

To test the extent to which male mating performance was enhanced by MnSOD overexpression, we conducted a series of female mate choice assays in the laboratory. When both wild-type and transgenic males were irradiated at 70 Gy, the recommended sterilization dose for tephritid SIT, and competed against each other in head-to-head assays with a single female, two of the lines (SOD2-5.2 and SOD2-6.1) outperformed wild-type males, with transgenic males 1.5 times more likely to be chosen by the female (χ² = 6.246, d.f. = 1, p = 0.012 for SOD2-5.2; χ² = 4.499, d.f. = 1, p = 0.033 for SOD2-6.1,). The mating success of the other five transgenic lines was indistinguishable from wild-type after irradiation (figure 2a). However, when both males were unirradiated, transgenic males underperformed relative to their wild-type counterparts (figure 2b; χ² = 14.143, d.f. = 1, p = 0.0002), indicating a cost to being transgenic under non-stressful conditions. Taken as a whole, the relative mating competitiveness of transgenic males improved after irradiation (nominal logistic model, χ² = 13.270, d.f. = 1, p = 0.0003), pointing to a specific, protective effect of MnSOD overexpression only when flies were stressed. Mate choice results for other combinations of males (irradiated transgenic males versus unirradiated wild-type and unirradiated transgenic males versus irradiated wild-type) are shown in electronic supplementary material, figure S4. While MnSOD overexpression improved mating performance, irradiation at 70 Gy still resulted in sterility. For each of the eight lines (wild-type + seven transgenic), a minimum of 50 eggs were collected from each of 10 crosses. In these 80 crosses, no viable eggs were observed after irradiation. By contrast, control crosses consisting of an untreated wild-type male with an untreated female yielded viable offspring in 7/10 crosses.

Figure 2.

MnSOD overexpression improves mating performance in irradiated flies. Males from the indicated genotypes were paired up with wild-type males and introduced to virgin wild-type females. (a) Irradiated transgenic males versus irradiated wild-type males. (b) Unirradiated transgenic males versus unirradiated wild-type males. In each panel, the grey portion of the stacked bar indicates the mating success of transgenic males, while the white portion indicates the mating success of wild-type. Bars represent the mean mating probability of each group relative to the wild-type male used in that particular experiment (i.e. untreated males in (a), irradiated males in (b)). Whisker bars represent s.e.m., n = 3–7 separate trials per comparison, with each trial consisting of at least five mate choice assays. Asterisk (*) indicates a significant difference from 50% mating success.

(c). MnSOD overexpression protects against oxidative stress at the cellular level

Our initial screens above indicated that lines 5.2 and 6.1 experienced the largest improvements in mating performance, so the remainder of the experiments focused on these two lines. To confirm that MnSOD overexpression is protecting against oxidative damage as expected if oxidative damage mediates cellular vitality and thus male performance, lipid peroxidation and protein carbonylation were measured in males at four time points following irradiation: (1) immediately after irradiation as pharate adults, (2) 24 h after irradiation, (3) as 3-day-old adult males and (4) as 12-day-old male adults (the time at which mating performance peaks). In response to irradiation, MDA levels, an indicator of lipid peroxidation (e.g. [35]), were significantly elevated in wild-type flies across the four measured time points, particularly in 3-day-old adult males that were 5 days post-irradiation (figure 3a; Linear Model, F_1,69 = 25.91, p = 3 × 10⁻⁶). In unirradiated males, there were no significant differences in MDA levels across all four time points (electronic supplementary material, figure S5; Linear Model, F_1,69 = 0.5312, p = 0.47), although there was a slight increase in SOD2-5.2 males as 3-day-old adults. By contrast, there was a significant trend towards lower MDA levels in irradiated transgenic flies across the four time points (figure 3b,c; Linear Model, F_1,69 = 25.92, p = 2.9 × 10⁻⁶). Sterilized SOD2-5.2 males had significantly reduced MDA levels immediately after irradiation and 24 h later, while SOD2-6.1 males had lower MDA levels at the middle time points. Our other measure of oxidative damage, protein carbonylation, did not change in response to sterilizing irradiation, and there were no differences between wild-type and transgenic flies (electronic supplementary material, figure S6).

Figure 3.

MnSOD overexpression reduces oxidative damage following irradiation. (a) Irradiation significantly increases lipid peroxidation, as indicated by the TBARS assay. (b) TBARS levels in irradiated flies from wild-type and line SOD2-5.2. (c) TBARS levels in irradiated flies from wild-type and line SOD2-6.1. In all graphs, symbols represent mean and whisker bars represent s.e.m., n = 4 per group. The p-values indicate the p-values from a linear contrast across all four time points between wild-type and the indicated group, while an asterisk indicates a significant difference between a group and its respective control at a particular time point.

(d). MnSOD overexpression improves muscle performance of sterilized flies

For lines SOD2-5.2 and SOD2-6.1, we measured the extent to which MnSOD overexpression preserved muscle performance in males following oxidative stress. Muscle performance was assessed with a negative geotaxis assay, routinely used to measure deficits in neuromuscular performance in flies (e.g. [36]). Flies were tested two weeks after adult emergence, the time when mating performance peaks. When flies were irradiated, both transgenic lines were nearly twice as likely to reach the top of the arena as wild-type flies (figure 4a), and transgenic flies were only half as likely to fall while climbing (figure 4b). Together, these results indicate that MnSOD overexpression improves locomotor ability of irradiated flies.

Figure 4.

MnSOD overexpression improves climbing performance of irradiated flies, but fails to improve longevity. (a,b) Groups of males were placed into a 33 cm long acrylic tube and tapped to the bottom of the tube to start the trial. (a) The proportion of males reaching the top of the tube. (b) The number of times a fly fell during each climbing trial. (c,d) Longevity curves for (c) unirradiated and (d) irradiated males from each genotype. In (b), bars represent the mean and whisker bars represent SEM. In (a,b), an * indicates a significant different between transgenic and wild-type, based on a χ²-test (a) or ANOVA with Tukey's HSD post hoc comparisons (b). In (c,d), the p-values indicate the effects of genotype and treatment on longevity from a Cox regression analysis. (Online version in colour.)

(e). MnSOD overexpression fails to increase lifespan

According to the free radical theory of ageing [31], gradual accumulation of damage from excess ROS contributes to age-related declines. We hypothesized that MnSOD overexpression would enhance longevity of caribflies and thus increase their opportunities for mating in the field. In unirradiated flies, transgenic flies had reduced longevity, with median longevity being 3–14 days shorter than wild-type flies (figure 4c; Cox regression, p < 0.05). Irradiation reduced longevity by 11 days in wild-type flies (figure 4c,d; Cox regression, p < 0.05), which is consistent with the free radical theory of ageing. The longevity of irradiated transgenic flies was still significantly shorter than that of irradiated wild-type flies (figure 4d), although the differences were less pronounced than in unirradiated flies.

4. Discussion

Here we show that increasing the activity of a mitochondrial antioxidant decreases oxidative damage and increases male mating success and muscular performance after exposure to a substantial oxidative stressor. We generated seven transgenic lines of caribflies that overexpress MnSOD, a key mitochondrial enzyme that is one of the cell's first lines of defence against ROS [37,38]. After irradiation, which causes substantial oxidative stress, two lines with intermediate MnSOD overexpression levels showed enhanced mating performance relative to sterilized wild-type flies (figure 2a). When flies were not irradiated, transgenic males were less competitive for mates than their wild-type counterparts (figure 2b). Yet the overall competitiveness of all transgenic lines was higher following irradiation. The observation that transgenic flies were less competitive in non-stressful environments reflects the presumed costs of transgenesis [39], which include energetic costs associated with synthesizing the transgene and marker proteins, as well as costs of insertional mutagenesis. Additionally, ROS are important signalling molecules for cell growth and survival [40], so excess antioxidant capacity in transgenic flies may reduce performance in non-stressful environments.

As opposed to previous data on the role of antioxidants in mating performance, which have been based largely on correlational observations [12,13], we provide direct evidence that endogenous antioxidants are critical for maintaining physiological function and sexual performance under stressful conditions. These results are in line with current models suggesting that sexual displays allow females to discern males with higher relative cellular vitality [2]. While we predicted that lines with the highest expression of MnSOD would perform best after stress, we found that our two best performing lines had intermediate levels of increased antioxidant activity. Similar effects of MnSOD transgene dosage have been observed in Drosophila, where moderate overexpression improved lifespan while strong overexpression had negative effects on survival [41]. It is also possible that lines differed in their tissue-specific expression of MnSOD, which could explain variation in mate performance across lines. However, identifying the precise reasons why certain transgenic lines outperformed others is beyond the scope of this study.

Biochemical measures of oxidative damage confirmed that MnSOD overexpression reduced oxidative damage caused by irradiation. In wild-type flies, sterilizing radiation resulted in a substantial increase in lipid peroxidation levels, and the effects of irradiation were long lasting. Levels of the oxidized lipid MDA were nearly 60% higher in irradiated wild-type flies 24 h post-irradiation and nearly 80% higher 5 days post-irradiation than those in unirradiated wild-type flies (figure 3a). These same radiation treatments significantly lowered mating performance (figure 2), consistent with previous work on birds [42] and fish [43] demonstrating that accumulation of oxidative damage reduces male mating success. MnSOD overexpression attenuated radiation-induced lipid peroxidation, with levels of MDA significantly lower in transgenic flies following irradiation compared to wild-type flies (figure 3b,c). A similar reduction in lipid peroxidation was observed when we previously used anoxic treatments to boost antioxidant capacity prior to irradiation [20], indicating that our transgenic strategy recapitulated the phenotypic effects of anoxic hormesis.

A recent synthesis of the biomedical literature by Koch & Hill [9] proposes that sexually selected behavioural displays are dependent on mitochondrial function, and our experiments provide experimental evidence in support of this hypothesis. Our results are consistent with previous work that oxidative stress and condition-dependent sexual traits are tightly linked [5–7]. This previous work, mostly in birds, primarily consists of correlational analyses of dietary carotenoids, antioxidant capacity and sexual selection [14,44]. Our work extends these findings by developing a tractable insect system in which antioxidant function can be directly manipulated through genetic means. Furthermore, to our knowledge, this study is the first direct demonstration of a critical role for endogenous antioxidants in maintaining sexually selected behaviours after stress. Previous work from our group has demonstrated that low oxygen treatments boost antioxidant capacity and increase performance after stress [20,25,26], and here we show a precise genetic mechanism using targeted overexpression of MnSOD. The elaborate sexual displays of tephritid fruit flies [16,45], coupled with their genetic tractability, make this system an excellent model for rigorously testing the links between cellular redox balance, courtship displays and male performance.

MnSOD overexpression also improved locomotor ability of male flies (figure 4a,b). Neuromuscular performance is critical for elaborate courtship displays [9], and for SIT to be successful, males must disperse from a release point and locate mates. Oxidative stress is known to reduce muscle force production and contribute to atrophy [32,46], and thus the rearing conditions and radiation doses used in SIT often reduce muscle performance, thereby limiting the efficacy of released males [27,47,48]. In addition to improving male performance, enhanced antioxidant capacity is predicted to increase longevity under the free radical hypothesis for ageing [31]. However, our results were not consistent with the free radical hypothesis for ageing. MnSOD overexpression failed to improve longevity in both fertile and sterile flies (figure 4c,d), and most transgenic lines in fact had reduced lifespans. While the free radical theory of ageing is widely touted, results in the literature are mixed. In Drosophila, SOD overexpression increases lifespan in some genotypes (e.g. [41]) but not others (e.g. [49]), while in C. elegans overexpression of Cu/ZnSOD enhances lifespan independent of oxidative damage [50], and deletion of MnSOD actually increases lifespan [51]. In our system, additional experiments with controlled genetic backgrounds are needed to fully test the link between antioxidant capacity and longevity, but our experiments do strongly suggest that ageing and MnSOD activity are not coupled in caribflies.

The results of this study also have implications for biological control programmes, especially SIT, an environmentally friendly control strategy that relies on sterilized males to suppress and/or eradicate pest populations. Global demands for food production and sustainable agriculture practices are increasing [52], and thus strategies like SIT that reduce dependence on chemical pesticides are likely to be relied on more heavily in the future. While SIT is effective at reducing pest populations, and in some cases is more cost effective than chemical control [53], the poor performance of sterilized males limits the potential economic efficiency of SIT. Poor male performance necessitates releasing large numbers of males (in some cases 100 : 1 ratio of released to wild males), which increases the cost of current programmes and can make it challenging to initiate and sustain new programmes. Here, we primarily targeted stress associated with irradiation, but oxidative stress may be a facet of many steps in the rearing, sorting and shipping of live biological control agents, and insect quality is still a challenge in programmes using alternative sterilization strategies such as genetic sterilization [39,48]. Our experiments demonstrate the potential of using transgenic approaches to enhance male quality and drive down the costs associated with SIT. Clearly, additional work is needed to further increase the positive performance benefits of antioxidant overexpression beyond the modest levels of improvement we have shown here to make a substantial impact on performance of biological control agents in the field. However, our current proof of concept suggests that with further optimization of parameters like recombination site, transgene copy number, tissue specificity of expression, etc., improvement of male sexual performance in sterile insect programmes is indeed possible. We also provide a proof of concept that similar transgenic strategies to improve physiological performance could also benefit other insect release programmes, such as augmentative biological control, managed pollinators and even gene-drive systems.

Data accessibility

Data are available from the Dryad Digital Repository: https://doi.org/10.5061/dryad.nk58h0m [54].

Competing interests

We declare we have no competing interests.

Funding

This work was supported by the AFRI Research Initiative Education and Workforce Development grant nos. 2015-67012-22793 and 2015-67012-25339 and by the Biotechnology Risk Research Grants Program Assessment grant no. 2017-33522-27068 from the USDA National Institute of Food and Agriculture to N.M.T., the Emmy Noether program of the German Research Foundation SCHE 1833/1-1 to M.F.S., USDA-NIFA-AFRI grant no. 2016-67013-25087 to A.M.H., and NSF-DEB 1639005, NSF-IOS 1257298, the Florida Agricultural Experiment Station, the Charles Steinmetz Endowment for Emerging Applications in Entomology, and IAEA/FAO CRP in Dormancy Management to Enable Mass-rearing and Increase Efficacy of Sterile Insects and Natural Enemies to D.A.H. V.S.D. was supported by a fellowship from the Brazilian Coordination for the Improvement of Higher Education Personnel (CAPES).