Females With a Mutation in a Nuclear-Encoded Mitochondrial Protein Pay a Higher Cost of Survival Than Do Males in Drosophila

Abstract

Males and females age at different rates in a variety of species, but the mechanisms underlying the difference is not understood. In this study, we investigated sex-specific costs of a naturally occurring mildly deleterious deletion (DTrp85, DVal86) in cytochrome c oxidase subunit 7A (cox7A) in Drosophila simulans. We observed that females and males homozygous for the mutation had 30% and 26% reduced Cox activity, respectively, compared with wild type. Furthermore, 4-day-old females had 34%–42% greater physical activity than males. Greater physical activity in mutant females was correlated with a 19% lower 50% survival compared with wild-type females. Mutant and wild-type males had equal survival. These data suggest that females paid a higher cost of the mutation than did males. The data demonstrate linking population genetics and structural modeling to experimental manipulations that lead to functional predictions of mitochondrial bioenergetics and organism aging.

IT is not clear why males and females age at different rates in a variety of species, but a mechanistic understanding of these differences is key to understanding the process of aging. Among many mammalian species, females tend to be longer lived than males. Females also tend to have reduced incidence of disease associated with both aging and mitochondrial dysfunction (1–3). In contrast to mammals, females of the fly Drosophila simulans typically have shorter life span and greater mitochondrial DNA density and efficiency of ATP production compared with males (4). These data suggest that the rate of aging and its association with mitochondrial function may differ between mammals and Drosophila. In this study, we determine the sex-specific costs of a naturally occurring mutation that mildly affects function of the mitochondrial protein, cytochrome c oxidase (Cox). The use of induced mutations to understand differences in survival has a rich history in Drosophila [eg, (5–7)], but few studies have exploited naturally occurring variants to understand the influence of mildly deleterious changes (8,9).

The oxidative phosphorylation (OXPHOS) system is composed of the four enzyme complexes (complexes I–IV) that make up the mitochondrial respiratory chain and the ATP synthase (complex V) (10). The energy released during electron transport along the respiratory chain is used to form a proton gradient across the inner mitochondrial membrane that drives phosphorylation of ADP to ATP. Aside from complex II, each respiratory chain complex is composed of both mitochondrial DNA– and nuclear DNA–encoded protein subunits. In humans, mitochondrial disorders of nuclear origin include OXPHOS disorders, defects in nuclear-encoded mitochondrial proteins that maintain mitochondrial DNA integrity, and mitochondrial disorders with secondary effects on the OXPHOS system. The clinical phenotypes of mitochondrial disease are highly variable in humans (11). Tissues having a high demand for bioenergy are most commonly affected, notably heart and skeletal muscle, the central nervous system and sensory epithelia, yet the specific phenotypes are not understood.

The goal of this study was to investigate the sex-specific physiological costs of a naturally occurring two amino acid deletion (DTrp85, DVal86) in cytochrome c oxidase subunit 7A (cox7A) of the mitochondrial electron transport chain (12) in D simulans. The nuclear-encoded cox7A gene produces a protein that is imported into the mitochondrion and forms a subunit of complex IV (cytochrome c oxidase, or Cox). The chromosomal location of this gene (GD18537) is 3R:17,095,012-17,095,384(−). The D simulans gene is orthologous to Drosophila melanogaster CG18193 located at 84F13-84F13, 3R: 4,169,402-4,170032(+).

Cox represents the terminal complex of the respiratory chain and is hypothesized to be a control point for the rate of electron flow through the entire chain (13). To date, only one naturally occurring mutation in a nuclear-encoded Cox gene has been identified and associated with human mitochondrial disease (14). It is likely that Cox subunit mutations causing subclinical effects go unnoticed, whereas highly deleterious mutations are fatal early in development (15).

In mammals, Cox is composed of 13 subunits, 10 of which are encoded by nuclear genes. In Drosophila, nine nuclear-encoded genes and four tissue specific isoforms are reported (12,16). Melvin and colleagues (12) modeled the tertiary structure of the cox7A mutation by homology to the bovine cox7A structure and predicted that the two amino acid deletion would decrease the function of Cox. This predicted result remained untested in males but has previously been observed in females (8). Also, in females homozygous for the cox7A two amino acid deletion, there is increased messenger RNA expression of five genes in four complexes of the electron transport chain, implying a coordinate upregulation of genes in response to the deletion (8).

Here, we investigate the sex-specific influence of a two amino acid deletion in cox7A on Cox enzyme activity and survival in males and females. Few studies have investigated the influence of nuclear-encoded mitochondrial mutations on sex-specific variation (17); however, this is likely to change in the future as results from the Fly Atlas project can be directly linked with stocks from the Drosophila Gene Disruption Project (18,19). Tang and colleagues (5) considered effects of the heterozygous optic atrophy 1 (dOpa1) mutation on life span in male and female D melanogaster. Optic atrophy 1 is a dynamin-like GTPase located in the inner mitochondrial membrane, and mutations in OPA1 are associated with autosomal dominant optic atrophy. The OPA1 GTPase plays an important role in mitochondrial fusion, cristae remodeling, and apoptosis. Tang and colleagues (5) found that the heterozygous dOpa1 mutation caused shortened life span, increased susceptibility to oxidative stress, and elevated production of ROS in the whole Drosophila. Antioxidant treatment partially restored life span in the male dOpa1 mutants but had no effects in the females. In contrast to nuclear-encoded mitochondrial mutations, a large number of studies have suggested that mitochondrial DNA may be a source of sex-specific variation because it is maternally inherited (20–23).

We assay physical activity as mean walking speed and survival of flies as physiological consequences of the Cox deletion mutation. A reduction in Cox activity may impair electron transport and possibly ATP production unless the organism responds to the bioenergetic consequences of the predicted reduction in Cox activity. Increased mitochondrial mass is observed in mouse models of mitochondrial disease and is hypothesized to be a mechanism of compensating for reduced OXPHOS function and maintenance of ATP homeostasis (24). Mutations in mitochondrial genes are frequently characterized by dysfunction of energy production (11) and have been associated with disease and aging in humans and in model organisms (25,26).

MATERIALS AND METHODS

Flies

The D simulans fly lines were constructed from a heterozygous isofemale line, HW01, collected in Honolulu, Hawaii, in November 2004. For this study, our goal was to isolate genotypes that were homozygous for either the presence or the absence of a six-nucleotide deletion in cox7A but were as identical as possible at all other loci. To achieve this goal, we employed a six-step strategy (8). Briefly the six steps were: Step 1, to reduce heterozygosity, the wild-caught isofemale HW01 fly line was inbred for five generations; Step 2, to allow recombination to reduce the linkage block around the cox7A mutation, the five-generation inbred line was maintained in population cages in the laboratory for 12 months; Step 3, to cure flies of Wolbachia infection (27), flies were treated with tetracycline; Step 4, sibling mating was employed to construct five 11-generation inbred lines; Step 5, pairs of homozygous normal and homozygous mutant lines were constructed; Step 6, the five 11-generation inbred lines were assayed for starvation resistance (28) and the two lines that showed 50% starvation resistance most similar to that of the 5-generation inbred HW01 isofemale line were kept. The final step exposed slightly deleterious mutations that accumulated during inbreeding or were linked with the deletion. In a previous study, we report on the Cox activity following genetic transfer of the mutation to multiple other backgrounds (8).

Prior to studies, cox7A normal and mutant flies were raised for two generations at constant density in 250-mL bottles on instant Drosophila media (Carolina Biological, Burlington, NC) at 23 ± 1°C, 50% relative humidity, and 12-hour light:12-hour dark daily cycle (29). To produce study flies, parental flies were released into population cages that contained solid oviposition resources (4% agar, 10% molasses supplemented with yeast paste) and were allowed to lay eggs for 4 h. Eggs were collected following Clancy and Kennington (30) and placed at a density of about 200 eggs per bottle into 250-mL bottles containing instant Drosophila media.

To avoid confounding the cost of harboring the cox7A mutation with that of mating and reproduction, we used unmated flies in all experiments. Unmated flies were collected from bottles within 2 hours of emergence from pupae and were sorted by sex on ice. Females and males were housed separately, in 25- × 95-mm vials containing instant Drosophila media at a density of 10–12 flies per vial. After 3 days, vials containing females were checked for the presence of larvae, an indicator that at least one fly in the vial was mated. Vials containing larva were discarded.

Study flies homozygous for either the deletion or normal allele of cox7A were collected directly from their respective homozygote parent lines. Study heterozygote lines were constructed afresh for each study by crossing homozygote normal and homozygote mutant lines. To avoid bias, study heterozygote flies were collected from reciprocal crosses; one in which the female was homozygote normal and another in which the female was homozygote mutant. Equal numbers of flies from the crosses were pooled for assays.

Cytochrome c Oxidase (Cox) Activity

In this study, we compared the Cox activity of 4-day-old males and females. Cox activity declines with age in D simulans (31) and 4 days was the earliest time flies could be assayed while allowing for verification that females were unmated. Cox activity tissue extracts were obtained from 15 replicate single thoraces of inbred flies aged 4 days (8). Briefly, individual frozen thoraces were thawed on ice and homogenized for 10 seconds in 100 μL of ice-cold homogenization buffer (50 mM NaH₂PO₄, pH 7.1, containing 0.05% Tween-80) using a Kontes Pellet Pestle motor (Kimble Chase, Vineland, NJ). Homogenates were diluted to 200 μL with 50 mM NaH₂PO₄, pH 7.1, and cell debris was removed by centrifugation. The supernatant was diluted at a ratio of 1:20 and 40 μL aliquoted into four-microplate sample wells. Assays of tissue extracts that interrogate the maximum rate at which the COX enzyme reduces cytochrome c were initiated by adding 160 μL of reduced horse heart cytochrome c (50 μM) to each sample well following Melvin and Ballard (31). Data were analyzed by nested analysis of variance (ANOVA) using JMP software (SAS Institute, Cary, NC) (32).

Physical Activity

In flies, an increase in physical activity causes a reduction in survival (33). As a measure of physical activity, we compared the mean walking speed of normal, heterozygous and mutant male and female flies at 4 days using a Trikinetics (Waltham, MA) physical activity monitor (34). Briefly, five flies of each genotype were anesthetized with humidified CO₂ and placed into individual 5-cm-long 5-mm-diameter glass tubes that contained 2.5 mm of instant Drosophila media at one end. After inserting a fly, a cotton plug was inserted and positioned to restrict the end-to-end distance inside the tube to 4.5 cm. Tubes were placed in the activity monitor horizontally inside a 23 ± 1°C incubator with 12-hour light: 12-hour dark daily cycle. Flies were allowed 10 hours to recover from anesthesia and acclimate to the tubes before data collection began. Assays were run over an 8-day period with at least two replicates per line included on each assay day. The small diameter of the glass tube restricted flight so that the activity measured was that of the fly walking end to end in the tube. The number of times a fly walked through an infrared light beam bisecting the glass tube was recorded by DAM software (Trikinetics) Data were converted to walking speed in centimeters per hour by multiplying the number of beam crossings counted per hour by the length of the glass tube (4.5 cm). Very high walking speeds caused by a fly pausing for a prolonged period in the infrared light beam were removed, and data were analyzed by nested ANOVA using JMP software (32).

Survival

We compared the survival of normal, heterozygote, and mutant males and females. Briefly, survival of male flies of each cox7A genotype was determined at 23°C. Newly emerged flies were collected at 2-hour intervals, males were separated from females under light CO₂ anesthesia, and placed into five (females) or three (males) replicate demography cages provided with instant Drosophila media. Deaths were counted, dead flies removed, and fresh medium was provided every 2 days. Deaths during the first interval after transfer to demography cages were not included in calculations to avoid confounding deaths due to handling with “natural” deaths. Initial cohort size was calculated as the summed death observations across all ages beginning at the second 2-day interval. Initially for males, there were a total of 298 (60, 78, and 160) homozygote normal males, 421 (122, 132, and 167) heterozygote, and 414 (147, 164, and 103) homozygote-mutant flies. For females, there were a total of 416 (36, 41, 102, 59, and 178) homozygote normal, 453 (62, 97, 43, 75, and 184) heterozygote, and 419 (89, 48, 48, 124, and 112) homozygote-mutant flies.

The age at 50% survival was calculated from the survival data and compared by nested ANOVA with sex and genotype as fixed factors. The survival function and simultaneous 95% confidence intervals were calculated and analyzed by the log-rank test using JMP software (32).

RESULTS

Cytochrome c Oxidase (Cox) Activity

The activity of cytochrome c oxidase was lowest in thoraces of 4-day females homozygous for the cox7A mutation. Compared with homozygote normal flies, Cox activity was reduced by 30% and 26% in homozygote mutant female and male flies, respectively; and by 10% in both female and male heterozygotes (Figure 1). Nested ANOVA shows that Cox activity is significantly influenced by genotype, F(2,212) = 18.26, p < .001, and sex, F(1,212) = 5.47, p = .02, but not by genotype × sex interaction, F(2,212) = 0.12, p < .88. Replicate lines did not differ in Cox activity, F(3,212) = 1.70, p = .17.

Figure 1.

Cytochrome c oxidase activity from thoraces of 4-day-old male and female Drosophila simulans. The two amino acid deletion mutation results in reduced Cox activity. We have previously reported the Cox activity of 4-day-old females (8). Bars represent mean Cox activity in nmoles cytochrome c oxidized/min/mg ± SE. Bars not connected by the same letter differ significantly according to Tukey’s honestly significantly different test with Q = 2.88 and α = 0.05.

Physical Activity

Homozygous females harboring the mutation had the highest mean walking speed, but this difference between mutant and normal females was not significant (Figure 2). Physical activity of females was higher than that of males for all genotypes with 34%, 42%, and 39% higher activity in mutant, heterozygote, and normal females, respectively, compared with males. Nested ANOVA shows a significant influence of sex, F(1,210) = 49.29, p < .001, and no influence of genotype, F(1,210) = 2.39, p = .09, or sex × genotype interaction, F(2,210) = 0.01, p = .99, on walking speed. Replicate lines did not differ, F(3,210) = 0.5, p = .6, in walking speed. Nested ANOVA also shows that replicates did not differ over the 8 days of assays, F(7,210) = 0.24, p = .93, therefore days were combined in the above analysis.

Figure 2.

Total physical activity calculated as mean walking speed of normal, heterozygous, and mutant Drosophila simulans. Survival of male and female flies homozygous or heterozygous for the cox7A (ΔTrp85, ΔVal86) is similar at 4 days of age. We have previously reported the physical activity of females (9). Bars represent mean ± SE. Bars not connected by the same letter differ significantly according to Tukey’s honestly significantly different test with Q = 2.38 and α = 0.05.

Survival

Females harboring the cox7A mutation had reduced 50% survival compared with wild-type females. No difference in 50% survival was observed in males of any genotype (Figure 3). Compared with homozygote normal females, age at 50% survival was reduced by 19% in homozygote mutant and by 20% in heterozygote females. Survival of homozygote normal females was 5% lower than that of homozygote normal males. Nested ANOVA shows that 50% survival was influenced by genotype, F(2,39) = 3.26, p = .049, and sex, F(1,39) = 21.86, p < .001, but not sex × genotype interaction, F(2,39) = 2.62, p = .09. Replicate lines did not differ in 50% survival, F(2,39) = 1.51, p = .23.

Figure 3.

Female flies suffer a greater cost of survival due to the cox7A (ΔTrp85, ΔVal86) mutation than do male flies when compared with the normal cox7A allele. We have previously reported the survival of females (9). Bars represent mean 50% survival ± SE. Bars not connected by the same letter differ significantly according to Tukey’s honestly significantly different test with Q = 3.00 and α = 0.05.

Survival functions differed significantly between females and males of all three genotypes (Figure 4). The survival function for homozygote normal females declines below that of males at age 42 days, and the simultaneous 95% confidence intervals of male and female survival overlap until age 48 days, become non-overlapping between ages 48–58 days, and overlap again in late life. Mean survival of homozygote normal females was 2.6% lower than that of homozygote males (44.1 ± 0.7 days and 45.3 ± 1.1 days, mean ± SE for females and males, respectively; log-rank χ² = 15.62, p < .001). The survival function of heterozygote females declines below that of males at age 35 days, and the simultaneous 95% confidence intervals become non-overlapping between sexes at age 36 days. Heterozygote females had a 19% reduction of mean survival compared with males (36.4 ± 0.4 days and 44.8 ± 0.7 days; log-rank χ² = 240.83, p < .001). Survival function of homozygote mutant females declines below that of males at age 34 days, and the 95% confidence intervals become non-overlapping between sexes at 49 days. Mean survival of homozygote mutant females was reduced by 16% compared with males (37.4 ± 0.6 and 44.4 ± 0.8: log-rank χ² = 186.02, p < .001).

Figure 4.

Survival function of female and male flies by genotype. Dark lines indicate survival function, shading indicates the lower and upper 95% confidence intervals of the survival function. We have previously reported the survival of females (9).

DISCUSSION

Understanding why males and females age at different rates has strong potential to give insight into basic gerontological processes. In this study, we explore the sex-specific consequences of a two amino acid deletion mutation in a nuclear gene, cox7A, that produces a protein that is imported into the mitochondrion. These data show that unmated females pay a greater survival cost of the mutation than do males. One possible explanation for this result is that females pay the higher bioenergetic cost. One bioenergetic cost may be increased physical activity. In this study, we observed that females have a significantly higher mean walking speed, a measure of activity, than do males (Figure 2). Previously, it has been shown that higher physical activity in flies causes a decrease in survival (33). Unmated Drosophila females also pay a bioenergetic cost of producing unfertilized eggs. It has been known for several decades that unmated females can lay a large number of unfertilized eggs (35), but the cost of such egg production is not well known. If it is the case that unmated females pay a bioenergetic cost, we predict that mated females harboring the mutation should have a still greater reduction in survival unless they have reduced physical activity. Mated females produce higher number of eggs than do unmated females (35), and this is expected to confer an even greater bioenergetic cost. The marked cost of reproduction in many insects is typically manifest as an increased death rate of mated relative to unmated females (36,37), which may be partitioned into the cost of mating and the cost of producing eggs (38). To understand the full impact of mutations that affect mitochondrial function and sex differences in life span, future experiments should include mated flies.

One alternate explanation for the high survival cost of the mutation in females may be that cox7A is more tissue specific in males than in females. The Fly Atlas records the expression of cox7A (Dsim\CG18193 or GD18537) as highly testis specific (19). In contrast, Ballard and colleagues (8) reported similar levels of expression of cox5A (GD18785) and GD18537 in the thoraces of unmated females (GD18785 is ubiquitously expressed). These latter workers also showed that females harboring the deletion show elevated levels of messenger RNA expression, suggesting a specific response to the deletion.

An alternate, or perhaps complimentary, explanation is that another isoform of cox7A (GD18537) exists but has not yet been documented in Drosophila. Searching of the Drosophila transcripts with tblastn (39) suggests that at least one and possibly two isoforms may exist. These are genes GD20874 (48% identical, E value = 2e⁻⁵) and CG34172 (no Dsim ortholog listed in FlyBase, 61% identical, E value = 2e⁻⁴). Like GD18537, GD20874 occurs at 84F13-84F13 and therefore mutations in both genes could be physically linked. As such, the reduction in Cox activity observed in the mutant flies may result from a linked mutation in GD20874 and not the deletion recorded in GD18537. The presence of a hitherto unreported isoform of cox7A could plausibly explain the Cox expression data if the primers and probe used by Ballard and colleagues (8) amplified both genes. Computational analysis of primers and probes suggest this is unlikely and direct sequencing show this is not the case. The plausible second isoform, CG34172, occurs at cytological position 22D1-22D1 in D melanogaster. CG34172 appears truncated, but it is possible that the last 30–40 amino acids are actually missing from the annotation. Fly Atlas (19) reports that CG34172 is highly expressed in head, heart, and carcass. Computational analysis of primers and probes also suggests that this gene was not amplified by Ballard and colleagues (8).

The influence of genes on phenotype has been a longstanding question in biology and is now one of the greatest challenges of the postgenomics era. Discovering the link between naturally occurring gene variants and organismal phenotype can provide insight into the molecular physiology that determines phenotype and the mechanisms by which populations evolve. Here, we show that linking population genetics and quaternary modeling (12) with experimental manipulations can lead to functional assessment of the influence of naturally occurring mildly deleterious amino acid mutations on mitochondrial bioenergetics (Cox activity) and aspects of the organismal phenotype (physical activity and survival). In this regard, the cox7A (ΔTrp85, ΔVal86) deletion identified in D simulans provides a superb system in which to explore the mechanisms by which organisms maintain energy homeostasis in response to disease and aging.

FUNDING

This work was supported by Australian Research Council Discovery grant DP110104542 to J.W.O.B.