Exposure to males, but not receipt of sex peptide, accelerates functional ageing in female fruit flies
Abstract
- Increased exposure to males can affect females negatively, reducing female life span and fitness. These costs could derive from increased mating rate and also harassment by males. Additionally, early investment in reproduction can increase the onset or rate of senescence in reproductive traits. Hence, there is a tight link between reproduction and ageing.
- Here, we assess how mating and encounter rate with males impacts declines in female functional traits that are not directly involved in reproduction. In Drosophila melanogaster fruit flies, exposure to males and mating reduces female life span through harassment and receipt of seminal proteins, including sex peptide (SP). We manipulated the intensity of female exposure to males and regularly assessed female stress responses and recorded physiological traits over her lifetime.
- Both mating itself and increased exposure to males accelerate declines in female climbing ability and starvation resistance. However, this is not related to changes in female body mass or fat storage. Moreover, these declines are not driven by the receipt of SP.
- Our results suggest some synchrony in senescence across traits in response to female exposure to males; however, this is not universal, as we did not find this for physiological traits. Synchrony in senescence has been theorized but little supported in the literature. It is clear that ageing is a multifaceted trait; to understand environmental impacts on ageing rates, we must measure more than life span and indeed measure senescence in multiple traits. Specifically, our work shows that we must identify which female traits are sensitive to elevated mating activity to understand the impact of antagonistic interactions between the sexes on female ageing patterns.
A plain language summary is available for this article.
1 INTRODUCTION
Under the evolutionary theories of ageing, a weakening of natural selection forces later in life results in the less effective purging of late-acting mutations. Here, genes either accumulate mutations with age or these genes have pleiotropic effects, that is were selected for as they increase reproduction at younger ages even if they have deleterious effects at older ages (Gaillard & Lemaître, 2017; Kirkwood, 2005; Kirkwood & Rose, 1991; Williams, 1957). Apart from late-acting genes with deleterious effects, classic theory suggests that ageing results from a trade-off between resource allocation to reproduction rather than somatic maintenance. In this latter scenario, the resources invested into reproduction are not available for somatic maintenance (see Maklakov & Immler, 2016). These early ideas have been refined and widely discussed. For example, it is predicted that there would be genetic correlations between early and late-life fitness but these might not always be negative (e.g., Maklakov, Rowe, & Friberg, 2015) and the unappreciated costs of germ line maintenance challenge the fecundity-life span trade-off (Maklakov & Immler, 2016). Elevated rates of reproduction have been shown to decrease life span in a range of species (Chapman, Liddle, Kalb, Wolfner, & Partridge, 1995; Helle & Lummaa, 2013; Maynard Smith, 1958; Partridge & Farquhar, 1981; Tatar, Carey, & Vaupel, 1993). Furthermore, the hypothesized trade-off between early and late-life reproduction has been supported by various studies of wild vertebrates (Bouwhuis, Charmantier, Verhulst, & Sheldon, 2010; Nussey, Kruuk, Donald, Fowlie, & Clutton-Brock, 2006; Reed et al., 2008), likewise the influence of early reproductive effort on late-life survival and late-life body condition (Beirne, Delahay, & Young, 2015; Lemaître et al., 2015). Hence, elevated early reproductive effort not only shortens life span, but can also impact age-specific changes in reproductive traits (Lemaitre & Gaillard, 2017; Nussey et al., 2006). In females, this is variously measured through traits such as egg production, inter-birth interval and offspring weight and survival (Hayward et al., 2013; Nussey et al., 2006; Reed et al., 2008). However, what is rarely measured is the impact that mating has on senescence of functional traits not directly associated with reproduction.
Mating can affect female life span as part of the costs of mating inflicted by sexually antagonistic interactions (Arnqvist & Nilsson, 2000; Chapman et al., 1995). In some species, females are already affected after a single mating, as seen in the seed beetle Acanthoscelides obtectus (Maklakov, Kremer, & Arnqvist, 2005) with males altering female ageing rates to benefit their own genetic interests. In the fruit fly Drosophila melanogaster, a model in studies of ageing, effects of mating on life span are well known, particularly in females (Flatt, 2011). In D. melanogaster females, mating and increased egg production can increase susceptibility to oxidative stress (Rush et al., 2007; Salmon, Marx, & Harshman, 2001), a much invoked driver of ageing (Kregel & Zhang, 2007). The sources of the effects of mating on female life span have been identified both as male harassment (Partridge & Fowler, 1990) and the repeated receipt of seminal fluid transferred from males during mating (Chapman et al., 1995; Fowler & Partridge, 1989), though even just the perception of male pheromones reduces female life span (Gendron et al., 2014). Constant exposure to males causes females to have a shorter total and reproductive life span (Edward, Fricke, Gerrard, & Chapman, 2011) indicating that male exposure alters female life-history traits. Alterations of female physiology and behaviour are driven by components of the male seminal fluid. The best studied of these seminal proteins, sex peptide (SP), influences various female traits (Ram & Wolfner, 2007). Some of these are directly related to reproductive effort, such as increasing egg production and reducing willingness to remate (Chapman et al., 2003), but SP also manipulates female nutritional decisions (Ribeiro & Dickson, 2010), increases immune responses (Peng, Zipperlen, & Kubli, 2005), increases activity and reduces sleep (Isaac, Li, Leedale, & Shirras, 2010). Female susceptibility to male-induced costs of mating is influenced by protein (likely crucial for egg production) available to them in their diet (Chapman & Partridge, 1996; Fricke, Bretman, & Chapman, 2010). Given this plethora of phenotypes, it is perhaps unsurprising that SP reduces female life span and so overall fitness, and is a much-cited example of sexual conflict (Fricke, Wigby, Hobbs, & Chapman, 2009; Smith, Clarke, Boone, Fricke, & Chapman, 2017; Wigby & Chapman, 2005).
Understanding functional senescence, the decline in physical functioning with age, can elucidate how different traits contribute to the gross ageing phenotype and are obviously of great concern when assessing “health span.” The move towards understanding health span requires knowledge of whether traits differ in the onset and rate of senescence and their responsiveness to factors that are known to alter senescence patterns (Martin, Bergman, & Barzilai, 2007; Promislow, Fedorka, & Burger, 2006). Theory predicts that natural selection should act most strongly on those functions that impact the risk of death most strongly (reviewed by Gaillard & Lemaître, 2017). Natural selection should then promote a stronger synchronicity in senescence patterns among traits (Maynard Smith, 1962; Williams, 1957). However, this is largely not borne out by empirical studies, which show asynchrony between traits in senescence in humans, laboratory and wild animals (Bansal, Zhu, Yen, & Tissenbaum, 2015; Grotewiel, Martin, Bhandari, & Cook-Wiens, 2005; Hayward et al., 2015; Nussey, Froy, Lemaitre, Gaillard, & Austad, 2013; Walker & Herndon, 2010). There is therefore a need to uncover the environmental and genetic factors which contribute to this variation in senescence among traits (Nussey et al., 2013).
Functional senescence in D. melanogaster has been measured in a variety of traits including resistance to various stressors, climbing ability, immune and memory function, though the age of the onset of senescence in different traits ranges from ~10 to 100 days (reviewed by Grotewiel et al., 2005). Social environment can affect functional senescence as same-sex social contact versus isolation reduces the speed of decline in climbing ability in females (but not males) (Leech, Sait, & Bretman, 2017). Here, we aimed to assess whether intensified mating interactions accelerate functional ageing in females by measuring not just life span but also climbing ability and starvation resistance, plus potential underpinning physiological traits, that is fat content and body mass. If mating per se causes more rapid functional declines then we would expect virgins to decline more slowly, but females intermittently or constantly exposed to males to show similar patterns. However, if number of matings and/ or harassment by males plays a role in functional senescence, then females intermittently exposed to males will decline more slowly than those constantly held with males. If the receipt of SP is part of the underlying mechanism of reproduction-induced functional senescence, then females mated to males that do not produce SP should show slower declines than females receiving SP.
2 MATERIALS AND METHODS
2.1 Fly culturing
For these sets of assays, Dahomey wild-type individuals were used. This strain was collected in the 1960s in Dahomey (now Benin) and has ever since been cultivated at 25°C and 60% RH on a 12:12-light: dark cycle in the laboratory in large, cage cultures in overlapping generations. All experiments were conducted under these standard conditions using vials containing 7 ml Sugar–Yeast (SY) medium (Bass et al., 2007) with excess live yeast granules unless stated otherwise. We allowed the parental generation to lay eggs on agar-grape juice plates (50 g agar, 600 ml red grape juice, 42.5 ml nipagin (10% w/v solution), 1.1 L water) supplemented with yeast paste. The following day first-instar larvae were collected at 100 larvae per vial, and 10 days later virgin females and males were collected on ice. Adults were stored in same-sex groups of 20 per vial until used in the experiment when they were 4 days post-eclosion.
2.2 Sex-peptide knockout mutants
Mutant stocks were maintained as bottle cultures (70 ml of SY food in 1-L bottles). We crossed virgin Δ130/TM3,Sb,ry females with SP0/TM3,Sb,ry males. Δ130/SP0 male offspring (SP0) from this cross do not produce SP (Liu & Kubli, 2003). As a genotype matched control, we crossed Δ130/TM3,Sb,ry females with SP0,SP+/TM3,Sb,ry males and the resulting Δ130/SP0,SP+ sons (SP+) produce and transfer SP at mating. The Δ130/TM3,Sb,ry stock was backcrossed for three generations into the Dahomey wild-type genetic background and chromosomes 1, 2 and 4 of the other two stocks for four generations.
2.3 Experimental set-up
2.3.1 Male exposure treatment
Wild-type females were assigned at random to one of three male exposure treatments. Females either encountered no males during their lifetime and remained virgin were continuously held with males or experienced an intermittent exposure regime. In the intermittent exposure treatment, females were held for three consecutive days with males and were held alone the remainder of the week. Once a week the batch of males used was discarded in both male exposure regimes and exchanged with a fresh batch of 4- to 5-day-old males to account for age-related declines in male courtship and mating behaviour. All flies were moved to new vials with fresh media twice a week.
Against the backdrop of these three male exposure treatments, we then performed three independent assays to test different functional aspects throughout female life span to measure a female's ability to maintain functional integrity while paying the cost of mating.
2.3.2 Negative geotaxis assay
Negative geotaxis or startle-induced climbing is a standard assay of locomotor senescence in flies (Jones & Grotewiel, 2011). When tapped to the bottom of a cylinder, flies “escape” by climbing upwards, a response which becomes progressively slower with age (Arking & Wells, 1990). In this assay, females were tested for their ability to climb up 8 cm in an empty vial. Once a week females from the three exposure treatments were transferred individually into empty vials, allowed to adjust for 5 min, and then, the females were banged down to the bottom of the vial. Immediately afterwards females were observed, we recorded whether they (a) tried to climb in the first place and (b) the time it took them to cross the 8-cm line, up to a maximum of 180 s.
We started with 180 females and first tested them at 4 days post-eclosion and afterwards assigned them randomly to a male exposure treatment (n = 60 per treatment). Thus for the first measurement, all females were virgins. For this assay, females were either held in pairs with one male (continuous treatment and intermittent treatment during the male exposure time) or individually (virgin treatment and intermittent females during the no male time). We then tested each female once a week for her ability to perform this negative geotaxis task. We daily checked for female survival and recorded the day of death. When fewer than 20 females within one of the male exposure treatments remained alive, we stopped assaying the females of that particular treatment but continued with the others.
2.3.3 Starvation resistance assay
We started with a total of 300 females per male exposure treatment and held females either in groups of 10 (virgin treatment and intermittent exposure during the no male time) or in groups of five females and five males (continuous exposure and intermittent exposure when with males) per vial. During this time, females were maintained on standard SY medium with live yeast grains added ad libitum. Once a week when females were transferred to new vials, we made sure to reshuffle female groups to avoid common vial effects. We did not record female survival in the male exposure treatments.
For the starvation assay, once a week a subgroup of 30 random females from each of the three male exposure treatments were put on agar-only food without any yeast added. Females were kept individually in these vials. We performed daily survival checks at roughly 24-hr intervals and recorded how long females survived to assay their starvation resistance. This sequence was repeated weekly until fewer than 30 females remained per treatment to perform this assay. Again, we performed the assay for the first time when females were 4 days post-eclosion and before assigning them to a male exposure treatment.
2.3.4 Triglyceride assay
Starvation resistance is directly linked to lipid reserves (Lee & Jang, 2014), and triacylglyceride (TAG) is the major energy storage molecule in the fat body (Arrese & Soulages, 2010). Hence, we here directly measured TAG content in females. For this, we repeated the design of the starvation resistance assay with one exception; instead of weekly starving females, we snap froze 30 females in liquid nitrogen and subsequently estimated the amount of triglycerides stored. We followed the protocol for the coupled colorimetric assay for triglycerides as outlined by Tennessen, Barry, Cox, and Thummel (2014). With this assay, we compared the stored triglycerides across ages and treatments. The triglycerides are macromolecules bound to proteins and together form lipoproteins. We always pooled five females per sample and measured their wet weight (on a Sartorius MC 410S model at a resolution of 0.1 mg) for an estimate of body mass before they were treated according to protocol. We followed the protocol with a few minor exceptions; for example after the homogenization step (Step 3 in protocol by Tennessen et al., 2014), samples were centrifuged at 17,950 g at 4°C for 5 min. After preparing the samples and adding the glycerol standard and for half of them the triglyceride reagent (Sigma: T2449), we incubated tubes for 45 min at 37°C (step 7 in protocol by Tennessen et al., 2014) before measuring the colorimetric intensity of the sample at 540 nm in a Tecan Reader. Tennessen recommends normalization to an internal parameter to accurately reflect TAG levels across different conditions (here age), and we here normalize to body mass as this has been done before (e.g., Hildebrandt, Bickmeyer, & Kuehnlein, 2011) and allows for comparisons across studies.
2.3.5 Sex-peptide treatment
To test potential mechanistic underpinnings of female functional ageing responses to male exposure, we performed a further set of experiments to specifically test whether receipt of SP mediates responses in females experiencing high costs of mating. Thus, we repeated the two stress–response assays (negative geotaxis and starvation resistance) exactly as described above (Sections 2.3.2 and 2.3.3) in terms of sample sizes, starting age and sampling points. The only change implemented was that females were continuously exposed to either males lacking SP (SP0 treatment) or SP transferring control males (SP+ treatment) instead of wild-type males.
2.4 Data analysis
Data were analysed using R v 3.3.1. For the male exposure experiment, we had two hierarchical questions, firstly what was the overall effect of male exposure (including mating), and then for those females that mated, what was the effect of different amounts of exposure to males. As such, our approach was to analyse all three treatments first, and where an effect of treatment was found to then analyse the data without the virgin treatment. Survival data from females used in the negative geotaxis-climbing assay was analysed using Kaplan–Meier log-rank tests. Functional senescence data were analysed using GLMMs or GLMs as appropriate, using the package lme4 (using maximum likelihood rather than REML). Terms were subtracted from the maximal model by Analysis of Deviance (AOD) and by assessing the change in AIC (see Tables S1–S4). Senescence in climbing ability was analysed as time to reach 8 cm, with those individuals that tried but failed in the time allowed given a value of 180 s. On only four occasions across the two experiments did the fly not try to climb (each in different treatments), and of those that tried only ~6% failed to reach 8cm within the time limit (53/913 trials in the male exposure experiment, 21/284 trials in the SP experiment). Climbing time was used as the response variable in a GLMM with male exposure treatment as a factor, age and life span (to account for selective disappearance, e.g., Hayward et al. (2015)) as covariates, and the random effect of fly identity to account for repeated measures. Age was fitted both as linear and quadratic functions in the models. A difference in senescence rate between our treatments will be indicated by a significant age × treatment interaction term. Response variables that required flies to be sacrificed (starvation resistance, body mass and TAG per mg) did not yield repeated measures or life span data; hence, we used GLMs with age as a covariate and male exposure treatment as fixed factors. We initially assessed whether a linear or quadratic effect of female age within treatments was most appropriate, and where this was the case for at least one treatment, used the quadratic term in the full model (see Tables S1 and S3). For parameter estimates from the best supported model, see Tables S2 and S4.
3 RESULTS
3.1 Effect of exposure to males on life span and climbing ability
The amount of exposure to males significantly affected female survival (Kaplan–Meier log-rank test 
= 121.639, p < 0.001) with virgin females surviving the longest (Figure 1a). Comparing only females that mated showed that females intermittently exposed to males survived longer than those continuously exposed (
= 18.141, p < 0.001; Figure 1a). We assessed the effect on climbing ability using time to reach 8 cm. The maximal model contained the interaction between male exposure treatment, female age and life span (with fly identity included as a random factor). Removal of the three-way interaction compared to a model with all pair-wise interactions increased the AIC (see Table S1), and an AOD showed that a model without the three-way interaction was significantly worse (
= 7.304, p = 0.007). This significant interaction remained when comparing only females that mated (
= 8.490, p = 0.004). This suggests that the way climbing ability is affected by female age differs between treatments, with females in the constant male exposure treatment becoming worse at the climbing task more quickly with the exception of the final assay (Figure 1b). Furthermore, this interaction is affected by female life span. To illustrate this, we plotted life span and change in climbing time (day 32–day 4 assay), noting that there is a negative relationship only in the constantly exposed treatment (Figure S1). This could indicate selective disappearance within the constantly exposed treatment, as those females that lived longer showed less of a decline in climbing ability with age. In sum, both mating per se and the amount of exposure to males affect female life span and locomotor senescence.
3.2 Effect of exposure to males on senescence of starvation resistance, body mass and body fat
We found that male exposure significantly affected senescence in female starvation resistance (female age2 × male exposure treatment: 
= 11.995, p = 0.002, Table S1), likely caused by the virgin treatment showing an initial increase in starvation resistance before declining (Figure 1c). When the virgin treatment was removed, an interaction between the linear effect of age and exposure treatment remained (female age × male exposure treatment: 
= 7.038, p = 0.008), with females held intermittently with males surviving longer under starvation until day 32. This suggests that mating per se and the amount of contact with males affected age-related decreases in starvation resistance. In order to assess whether this was due to differences in fat reserves between treatments, we assayed female body mass and triacylglyceride content (TAGs). While for body mass, there was a significant interaction between female age and male exposure treatment (F1, 84 = 5.670, p = 0.019), this pattern does not mirror that of the starvation assay as females held constantly with males were initially slightly heavier (Figure 1d) yet had lower starvation resistance (Figure 1c). Furthermore, accounting for body mass, there was no effect of female age or male exposure treatment on amount of TAGs, either as an interaction (F1, 84 = 0.405, p = 0.526) or as main effects (treatment: F1, 85 = 0.419, p = 0.519; age: F1, 85 = 0.824, p = 0.367; Figure 1e).
3.3 Effect of receipt of SP on life span and functional senescence
We then tested whether the results we had observed could be attributed to the receipt of SP. Female exposure to males that did or did not transfer SP had no effect on life span (Kaplan–Meier log-rank test: 
= 0.496, p = 0.481; Figure 2a). Likewise, we found no effect on senescence in climbing ability. A model using all data (with fails fixed to 180 s, the maximum observation period) showed no significant interactions (AOD model comparisons all p > 0.05, see Table S3), and no main effect of treatment (
= 0.018, p = 0.894) but a decline with age (
= 72.976, p < 0.0001) (Figure 2b). For starvation resistance, the interaction between female age and SP treatment was non-significant (
= 2.944, p = 0.088), and when this term was removed, there was a significant effect of age (comparing models with age to that with age plus the quadratic term of female age 
= 16.951, p < 0.0001) but not of treatment (
= 2.40, p = 0.123; Figure 2c). These results suggest that the effects of mating and male exposure on female functional ageing are not driven by the receipt of SP.
4 DISCUSSION
While previous studies have reported that exposure to males affects female survival and reproductive senescence, our main findings show that this effect also applies to functional senescence in climbing ability and starvation resistance. Our data suggest that female life span and rates of functional senescence are altered in response to both mating per se and amount of contact with males. However, these effects did not appear to be driven by differences in body mass or stored fats. Likewise, the patterns in functional senescence were not attributed to the receipt of SP, as this did not cause any difference in the decline seen in either climbing ability or starvation resistance, or indeed in survival.
Largely, there is consensus in the phenomenon that mating reduces life span in female D. melanogaster (reviewed by Flatt, 2011), and this is confirmed in our data. Exposure to males seems to have an additive effect, in that females continuously exposed to males were more severely affected than those intermittently exposed, as in previous work (Chapman & Partridge, 1996; Edward et al., 2011). Edward et al. (2011) used a similar experimental set-up to us and measured female offspring production. Continuous exposure leads to reduced reproductive life span with a strong correlation with lifetime reproductive success and females having a high reproductive output early in life (Edward et al., 2011). What we now show is that this pattern is reiterated in the senescence of two of the non-reproductive traits we measured. Indeed, for both traits, the constantly exposed females start to show a more obvious decline at the second assay at 11 days post-eclosion, suggesting that there is some synchrony in senescence in these traits in response to exposure to males. It has been suggested that life span and health span are mechanistically connected (Rhodenizer, Martin, Bhandari, Pletcher, & Grotewiel, 2008) because longer-lived flies tend to have better climbing ability across ages (e.g., Gargano, Martin, Bhandari, & Grotewiel, 2005), though this is not always the case (Cook-Wiens & Grotewiel, 2002). In addition, we found no corresponding senescence in female body mass or body fat here, thus highlighting that not all aspects of female physiology were equally affected by male exposure. While it has been predicted that senescence should be observed as generalized deterioration rather than failure of single systems (Williams, 1957), this is largely not borne out by empirical work (recently reviewed by Gaillard & Lemaître, 2017). In general, these opposing examples are either studies of wild populations that are subjected to multiple environmental drivers of ageing (Hayward et al., 2015; Kervinen, Lebigre, Alatalo, Siitari, & Soulsbury, 2015; Massot et al., 2011) or laboratory studies that do not impose any particular pressures (Herndon et al., 2002). By applying a specific environmental driver of ageing in a controlled manner, this may allow us to dissect which traits are predominantly affected and assess whether asynchronicity is a general aspect or instead explained by different sensitivities to multiple environmental drivers. It would be fruitful to establish whether this is generally observed when other known determinants of ageing are manipulated.
Our finding that mating/ contact with males reduced starvation resistance is at odds with previous work. Multiple studies have found that mating increases female food intake and the size of the midgut, resulting in greater lipid storage and increasing starvation resistance (Jang & Lee, 2015; Reiff et al., 2015; Rush et al., 2007). Both our measures of functional senescence rely on energy reserves, but we found no evidence that lipid content was increased by multiple mating. However, these previous studies were not designed to assess senescence patterns, hence usually only measure resistance once, relatively early in life. Additionally, females were exposed to males in a very limited way, perhaps a single mating or interactions were allowed for just a few days. Here, instead we took a long-term approach where females mated repeatedly and were subject to male harassment, indicating that mating activity in the long-term results in the opposite pattern. While receipt of seminal proteins might induce higher feeding rates in young females (also seen in our constantly exposed females who were heavier in the beginning before declining, see Figure 1d), this might be countered by the harassment of females by males, therefore reducing feeding time and in the long-term energy stores. Also prior to being individualized for the starvation assay, individuals were held in groups of 10 until they were chosen at the appropriate test age, potentially leading to competition over food particularly as also larvae were present in the treatments with mated females. Combined this might limit female access to resources and explain the discrepancy between studies. Additionally, it may be that there is selective disappearance of the lighter females (Nussey, Coulson, Festa-Bianchet, & Gaillard, 2008), but we cannot test for this directly in these destructive sampling assays (e.g., by adding individual lifespan into the model).
While in general, body condition indices are used as a proxy for lipid content in animals, whether the two are tightly correlated is debated and further, whether either of these two measures influences fitness positively is not always clear cut (Wilder, Raubenheimer, & Simpson, 2016). At least in D. melanogaster, dietary composition has a strong impact on fat deposition and fecundity as well as life span (Lee & Jang, 2014; Skorupa, Dervisefendic, Zwiener, & Pletcher, 2008). TAG levels are strongly dependent on the amount of carbohydrates in the diet (Skorupa et al., 2008). On a balanced diet, the cellular composition of the female fat body is stable with age (Johnson & Butterworth, 1985) or can show a slight decrease in TAG levels (Skorupa et al., 2008). Here, we directly measured TAG levels in females as overall higher lipid reserve confers higher starvation resistant (Lee & Jang, 2014) and there is a strong link between lipid storage and egg production, as oocytes contain large amounts of lipids and are provisioned from the fat body (Arrese & Soulages, 2010). Curiously, we found females continuously exposed to males were least starvation resistant even though they had similar levels of TAGs. This suggests male exposure (including mating rate) might alter female fat metabolism. Mating status in redback spiders (Latrodectus hasselti) altered responses to food shortage, with mated females lowering their resting energetic rate in response to nutrient shortage, preserving energy, while virgin females maintained higher rates and had shorter life spans (Stoltz, Hanna, & Andrade, 2010). If similar metabolic mechanisms are at work in Drosophila, then we would expect male-exposed females to be more starvation resistant, but we found the opposite. Under starvation conditions, female Drosophila first use up the non-lipid fraction of their bodies before switching to lipids (Lee & Jang, 2014), but if anything continuously exposed females were slightly heavier. As we quantified the total amount of triglycerides, this did not allow us to distinguish whether it was dedicated to use in the ovaries and hence potentially not available for maintenance under starvation. This basic trade-off, where females under high male exposure invest more resources into egg production, might explain why these females were less resistant to starvation despite similar lipid energy reserves and this would be in line with the disposable soma theory (e.g., see Lemaître et al., 2015). The physiological dynamics that underpin life-history trade-offs deserve further scrutiny.
That we could not attribute any of the senescence effects to the receipt of SP is curious, given the multitude of effects on female phenotypes that have previously been found (Chapman et al., 2003; Isaac et al., 2010; Peng et al., 2005; Ram & Wolfner, 2007; Ribeiro & Dickson, 2010) and the importance of SP in inducing female costs of mating (Wigby & Chapman, 2005). For example, receipt of SP increases female activity (Isaac et al., 2010), but we did not see evidence of this in our climbing assay at any time point. It is possible that we did not measure these at old enough ages, though these measurements were within the timeframe for females constantly exposed to males to show a difference to virgins. At least for the induction of egg-laying by SP, females were only receptive when very young and this rapidly declined with age (Fricke, Green, Mills, & Chapman, 2013); hence, SP-induced female post-mating responses are female age-dependent and tend to diminish with age. It seems therefore that factors other than SP are more important in determining female ageing phenotypes, either the other Sfps or the direct harassment by males (Partridge & Fowler, 1990). While SP is one component implicated in the costs of matings, other Sfps are toxic (Lung et al., 2002; Mueller, Page, & Wolfner, 2007) and it is the entirety of the Sfps in the ejaculate that mediates the negative effects of multiple mating (Chapman et al., 1995). Mating (Zhou, Mackay, & Anholt, 2014) and receipt of SP (Gioti et al., 2012) dramatically alter female gene expression, including those implicated in life span such as the TOR pathway (Gioti et al., 2012). Simply hearing courtship song alters female expression of Turandot genes, a family of stress–response genes (Immonen & Ritchie, 2012). Furthermore, ecdysone receptor and genes involved in germ line maintenance and gustation/odorant reception are candidates for female responses to continuous male exposure (Gerrard, Fricke, Edward, Edwards, & Chapman, 2013). These candidates represent potential hormonal and metabolic pathways that might influence resource allocation to the germ line. It would be interesting to test whether these pathways alter allocation to the germ line versus somatic maintenance and could potentially explain the heterogeneity in physiological versus functional senescence found here. For a general pattern to emerge though, a wider array of functional traits, representing a broader set of biological functions, should be screened. This approach can be extended to include males, to test for sex differences in functional ageing to further our understanding how investment in mating activities alters senescence patterns. This could also reveal whether, in addition to being implicated in interlocus sexual conflict over mating rate, these traits might be targets for intralocus sexual conflict over ageing profiles (Archer, Recker, Duffy, & Hosken, 2018). Hence, there remains much work to be done on the molecular mechanisms underpinning how this environmental variable (mating activity) can alter senescence in multiple traits and how the different non-reproductive traits are integrated to contribute to observed ageing phenotypes.
5 CONCLUSIONS
Overall, we here showed that female costs of mating due to intensified male exposure lead to accelerated functional ageing in female motor ability and resistance to starvation stress. This decline though was not underpinned by a matching decline in relevant physiological traits. Hence, while we found some synchronicity in ageing phenotypes in response to mating activity across traits, this was not universal. Understanding which traits contribute to the observed mating costs, are particularly affected by high mating effort and display high rates of functional ageing should give valuable insights into ageing patterns and integration of different traits to the ageing phenotype.
ACKNOWLEDGEMENTS
We thank Dolors Amorós Moya, Sonja Schindler and Kristina Wensing for help with the assays and Xavier Harrison for help with statistical analysis. We thank the editor, associate editor and two reviewers for their helpful comments. We have no conflict of interest to declare. C.F. was supported by the DFG Emmy Noether Program (FR 2973/1-1).
AUTHORS’ CONTRIBUTIONS
A.B. and C.F. conceived the idea and designed the experiment; C.F. conducted the experiment; A.B. analysed the data; and both authors wrote the manuscript.
DATA ACCESSIBILITY
Data are freely available from Dryad Digital Repository under https://doi.org/10.5061/dryad.9qb0dt6 (Bretman & Fricke, 2019).
