Skip to main content
PMC home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2014 Apr 18.
Published in final edited form as: Aging Cell. 2003 Apr;2(2):123–130. doi: 10.1046/j.1474-9728.2003.00044.x

Testing an ‘aging gene’ in long-lived Drosophila strains: increased longevity depends on sex and genetic background

Christine C Spencer 1, Christine E Howell 1, Amber R Wright 1, Daniel E L Promislow 1
PMCID: PMC3991309  NIHMSID: NIHMS571973  PMID: 12882325

Summary

Molecular advances of the past decade have led to the discovery of a myriad of ‘aging genes’ (methuselah, Indy, InR, Chico, superoxide dismutase) that extend Drosophila lifespan by up to 85%. Despite this life extension, these mutants are no longer lived than at least some recently wild-caught strains. Typically, long-lived mutants are identified in relatively short-lived genetic backgrounds, and their effects are rarely tested in genetic backgrounds other than the one in which they were isolated or derived. However, the mutant's high-longevity phenotype may be dependent on interactions with alleles that are common in short-lived laboratory strains. Here we set out to determine whether one particular mutant could extend lifespan in long-lived genetic backgrounds in the fruit fly, Drosophila melanogaster. We measured longevity and resistance to thermal stress in flies that were transgenically altered to overexpress human superoxide dismutase (SOD) in the motorneurones in each of 10 genotypes. Each genotype carried the genetic background from a different naturally long-lived wild-caught Drosophila strain. While SOD increased lifespan on average, the effect was genotype- and sex-specific. Our results indicate that naturally segregating genes interact epistatically with the aging gene superoxide dismutase to modify its ability to extend longevity. This study points to the need to identify mutants that increase longevity not only in the lab strain of origin but also in naturally long-lived genetic backgrounds.

Keywords: aging, epistasis, lifespan, stress resistance, superoxide dismutase, thermal tolerance

Introduction

Recent molecular advances have given new impetus to the search for single genes that increase lifespan. By identifying these so-called ‘aging genes’, we move one step closer to understanding the fundamental genetic basis of aging. In the fruit fly, Drosophila melanogaster, one common way to identify aging genes is through the manipulation of P-elements in M-type (P-element free) strains (e.g. Lin et al., 1998; Parkes et al., 1998; Rogina et al., 2000; Tatar et al., 2001). Since the global spread of P-elements in D. melanogaster in the 1950s, there are no longer any M-type strains in the wild. As a result, the M-type strains used by biogerontologists to carry out aging gene research have spent decades in the lab adapting to the laboratory environment. Adaptation to lab conditions has two unintended consequences for aging research. First, lab strains have evolved unnaturally short lifespan that could bias the search for genes that increase lifespan (Promislow & Tatar, 1998; Sgro & Partridge, 2000). Second, given the relatively few M-type strains available for transgenic manipulations, researchers have typically measured the effects of aging genes in a single genetic background. This approach limits our ability to assay how genetic background interactions might affect longevity.

These two problems, the evolution of shortened lifespan and the effects of genetic background, apply to almost any lab strain or model organism. In the case of fruit flies, lab strains are typically kept in 2-week culture, where adults are allowed to lay eggs on fresh medium for 1–2 days and then are discarded at the age of 5–7 days. Two-week culture selects for high early fecundity and, at least in lab populations, fecundity is strongly negatively correlated with survival (Fowler & Partridge, 1989; Trevitt & Partridge, 1991; Chapman et al., 1993; Cordts & Partridge, 1996; Prowse & Partridge, 1997; Sgro & Partridge, 2000). At the same time that lab culture increases selection on early fecundity, it dramatically reduces the strength of selection to remove deleterious genes expressed later than about 7 days (Promislow & Tatar, 1998). In the absence of selection pressures, deleterious germ-line mutations with effects confined to later than 7 days will gradually accumulate, leading to shorter lifespan (Medawar, 1952). In fact, several experiments have now shown that in common D. melanogaster strains, such as Canton-S, Oregon-R, and w1118, evolution in lab culture has increased early fecundity, reduced age of maturity and shortened longevity (Matos et al., 2000; Sgro & Partridge, 2000; Linnen et al., 2001; Matos & Avelar, 2001; Houle & Rowe, 2003). In light of these factors, it may be that aging genes simply restore longevity in lab strains to its original duration, but do not extend it beyond what has previously been observed (Linnen et al., 2001; Promislow & Tatar, 1998).

While recent research has shed some light on the implications for aging studies of evolution in lab culture (Promislow & Tatar, 1998; Matos et al., 2000; Sgro & Partridge, 2000; Linnen et al., 2001), the use of lab strains presents a second critical and rarely considered problem. By using a single strain to isolate and study a newly discovered aging gene, we fail to consider genetic background effects. Genes for complex traits such as aging rarely act alone to influence the trait. Complex traits may be influenced by several to hundreds of genes (Falconer & Mackay, 1996) that may interact with one another to yield unpredictable results. For example, susceptibility to the human inflammatory bowel diseases (e.g. Crohn's disease) is influenced by several quantitative trait loci (QTL) of large effect, and at least two of these interact epistatically (Cho et al., 1998). In addition to these genes of major effect, the QTL are linked with at least 17 other genomic regions that do not show up as individual QTL, implying that several genes below their QTL detection threshold also affect disease susceptibility. With the rapid growth in studies of gene and protein networks (e.g. Wagner, 2000; Wagner & Fell, 2001; Brazhnik et al., 2002), it is now clear that all but the simplest traits are probably influenced by hundreds of interacting genes.

In most cases, studies of aging genes deliberately eliminate any potential confounding effects of epistasis by using a single, homogeneous genetic background. Only two Drosophila aging gene studies have examined how different genetic backgrounds can influence the effect of an aging gene. Sun & Tower (1999) simultaneously overexpressed Drosophila superoxide dismutase in two lab strains with different inherent longevities. Although the study only looks at two backgrounds, superoxide dismutase overexpression in the shorter-lived strain had a more dramatic increase in longevity than in the long-lived strain, which on average had a decrease in lifespan. In a study identifying the aging gene I'm-not-dead-yet (Indy), Rogina et al. (2000) determined that Indy increased lifespan in the short-lived Shaker, Hyperkinetic and drop dead strains by 40–80%, but only extended lifespan by 15% in a strain laboratory-selected for increased longevity. These studies imply that aging genes may not increase lifespan as much in long-lived strains as in short-lived lab strains. But aside from this example, we know little about the ability of aging genes to extend longevity in naturally long-lived populations. This is of obvious importance if we are to understand the genetic basis of longevity in human populations.

Here we examined how different genetic backgrounds affected the ability of the superoxide dismutase (SOD) gene to increase longevity and stress resistance in the fruit fly, D. melanogaster. SOD reduces intracellular damage caused by reactive oxygen species, a by-product of cellular metabolism that can build to toxic levels if not removed (Frank, 1985). Toxic reactive oxygen levels cause intracellular damage that is hypothesized to be a putative cause of aging (Orr & Sohal, 1994; Parkes et al., 1998; Sun & Tower, 1999). Furthermore, environmental stressors such as temperature fluctuations and exposure to toxins also increase cellular reactive oxygen concentrations. Because SOD functions as an antioxidant, overexpression of SOD should lead to increased stress resistance in addition to extended lifespan.

By overexpressing the human SOD (hSOD) enzyme in adult Drosophila motorneurones, Parkes et al. (1998) demonstrated the efficacy of hSOD to increase longevity and stress resistance. Using these same hSOD overexpression and control lines, we crossed the hSOD construct into 10 wild-caught, naturally long-lived genetic backgrounds to determine how the effects of hSOD overexpression on lifespan and resistance to thermal stress may depend on interactions with alleles that segregate in natural populations.

Results

The results presented here are based on analysis of lab strain control and SOD overexpression flies obtained from G. Boulianne and colleagues (Parkes et al., 1998) at the University of Toronto (UT strains). Strains from G. Boulianne assayed in our lab are referred to below as (UGA) UAS-HS strains. Males from UAS-HS strains were crossed to females from 10 distinct wild-caught, inbred strains to create 10 pairs of experimental strains, each pair consisting of an overexpression strain and a standard expression strain of SOD in each genetic background. We then assayed virgin flies of each sex for life expectancy and for thermal tolerance. The genotypes (Fig. 1), experimental design and statistical analyses are described in detail in the Experimental procedures below.

Fig. 1.

Fig. 1

Open in a new tab

Genotypes for overexpression and control levels of superoxide dismutase (SOD). (a) Genotype of UAS-HS control and overexpression flies obtained from Parkes et al. (1998) containing two P-elements. The P(w+)UAS-hSOD1 insert on the second chromosome carries a GAL4-sensitive upstream activating sequence and a human SOD1 gene. On the third chromosome, P[GAL4] is expressed only in adult motorneurones, thereby activating expression of the hSOD1 gene. (b) The UAS-HS overexpression strain used at UGA was homozygous for the hSOD1 P-element. The third chromosome carries GAL4 and a null mutation (X39) at the Drosophila SOD locus; it is balanced to prevent recombination. The control strain lacks GAL4. (c) Experimental strains result from a cross between UAS-HS males from b and females from 10 recently wild-caught, inbred isofemale strains. Here, + represents wild-caught wild-type alleles. The first three chromosomes of D. melanogaster are shown. Male genotypes are given in a and b; female genotypes are given in c. Numbers on the right represent the number of Drosophila SOD and human SOD alleles being expressed for each genotype. w = white, P = P-element, + = wild-type.

Life expectancy

The UAS-HS overexpression flies outlived the standard expression controls in both sexes (males: t35.5 = –2.27, P = 0.0297; females: t35.3 = −3.63, P = 0.0009; Fig. 2). UAS-HS control males had life expectancy equivalent to a similar stock in Parkes et al. (1998; UGA: 43.4 ± 3.3 days, UT 45.1 ± 1.7 days, t26 = 0.33, P = 0.745). Note that SOD overexpression genotypes from the two studies are not exactly comparable because the flies described in Parkes et al. (1998) had a single copy of the hSOD1 gene, while flies in our study carried two copies (Fig. 1a,b). The UT flies had 15% longer lifespans, though the difference was not significant, most likely due to low cohort number in the UGA strain (male mean lifespan, UGA: 55.4 ± 0.1 days, n = 2; UT: 63.7 ± 1.7 days, n = 25; t25 = 1.38, P = 0.181).

Fig. 2.

Fig. 2

Open in a new tab

Life expectancy. Female (a) and male (b) mean lifespan in days for control (open bars) and SOD overexpressed (shaded bars) genotypes from the UGA UAS-HS control strains and 10 wild-caught, inbred genetic backgrounds. Bars are +1 standard error. *P < 0.05, **P < 0.005.

In the wild-caught experimental strains, we found that mean lifespan was significantly affected by sex (females live longer than males) and SOD expression (SOD increases lifespan) (Table 1). A significant SOD × sex interaction indicated that the SOD expression level altered lifespan differently in the two sexes (Table 1). Higher levels of SOD were far more likely to increase lifespan in females (F1,9.1 = 19.87, P = 0.002) than in males (F1,9.04 = 1.35, P = 0.275).

Table 1.

Life expectancy: Mixed model PROC MIXED results for SOD effect and sex modelled as fixed effects (F statistics) and genetic background (GB) modelled as a random effect.

Fixed effects F d.f. P
SOD 7.05 1,12.7 0.0201
Sex 14.22 1,11.2 0.0030
SOD × sex 4.07 1,87.3 0.0466
Random effects vs. Reduced model Likelihood ratio d.f. P
GB None 6.4 1 0.0057
GB, GB × sex GB 1.4 1 0.1184
GB, GB × SOD GB 3.5 1 0.0307
GB, GB × sex, GB × SOD GB, GB × SOD 2.3 1 0.0647
GB, GB × sex, GB × SOD GB, GB × sex 4.4 1 0.0180
Open in a new tab

The likelihood ratio, the difference in the –2*REML log likelihood scores for the full vs. reduced model, tests for the significant contribution of random effects to genetic variance for mean lifespan in SOD overexpression vs. control flies. It follows a χ2 distribution with degrees of freedom (d.f.) equal to the number of factors by which full and reduced models differ. P-values are one-tailed. All full and reduced models include all fixed effects, and the effect added for each model is highlighted in bold type

Interestingly, while the effect of SOD on lifespan was significant among females but not males, the effect on the two sexes was highly correlated across genetic background (Fig. 3, Pearson's r = 0.842, n = 10, P = 0.002). This last result points to the possibility that some backgrounds were more responsive than others to the effects of increased SOD expression. This is further supported by a likelihood ratio test in PROC MIXED (SAS Institute, 2001), which indicated that the different genetic backgrounds contribute significantly to the observed genetic variance. Furthermore, there was a significant SOD × genetic background interaction, but no evidence of a sex × genetic background interaction (Table 1). The full model did not include the three-way (SOD × sex × genetic background) interaction because no variance was explained by this term.

Fig. 3.

Fig. 3

Open in a new tab

Correlation between the sexes of the change in lifespan caused by SOD overexpression. Correlation by sex of the change in mean lifespan between SOD overexpression and control flies within each genetic background. Data are mean difference in days for each genetic background.

Because one of our primary interests was to determine if SOD increased lifespan within each genetic background, we compared each overexpression line to its respective control line within each sex using estimate statements in PROC MIXED (SAS Institute, 2001). For each of 10 wild-caught lines, the offspring of flies crossed with standard SOD flies lived longer than the ‘long-lived’ UAS-HS overexpression flies (Fig. 2a,b). For females, SOD overexpression significantly increased lifespan in six of 10 genetic backgrounds (Fig. 2a). Among males, SOD overexpression flies in only one of 10 genetic backgrounds had a significantly different life expectancy from their control (Fig. 2b). SOD overexpression only significantly increased lifespan in both sexes in a single genetic background (73). In four genetic backgrounds (lines 22, 37, 38, 64), SOD overexpression did not alter life expectancy in either sex; in others (lines 9, 18, 33, 40, 46), only females showed a significant increase in lifespan. Surprisingly, in two genetic backgrounds (lines 22, 38), male control flies outlived their overexpression counterparts, although this difference was not significant. In these lines, female lifespan was not affected by SOD overexpression.

Thermal tolerance

We examined stress resistance in UAS-HS control flies and our experimental lines by exposing flies to 37 °C for between 20 min and 30 min (see Experimental procedures). Male and female UAS-HS overexpression flies had significantly higher post-stress survival than UAS-HS controls (F1,50 = 179.90, P < 0.0001). In the experimental lines, analysis in PROC MIXED revealed that SOD expression significantly affected thermal tolerance in females (F1,9 = 16.60, P = 0.0028) but not in males (F1,9 = 3.46, P = 0.0960). In both sexes, there was significant variability among genetic backgrounds as shown by a likelihood ratio test, but the interaction of SOD by genetic background was not significant (Table 2).

Table 2.

Thermal tolerance by sex: Likelihood ratio test for significant contribution of random effects (GB and GB × SOD) to genetic variance for thermal tolerance in SOD overexpression vs. control flies (see Table 1).

Sex Full model vs. Reduced model Likelihood ratio d.f. P
Female SOD, GB SOD 29.6 1 2.66 × 10–8
SOD, GB, GB × SOD SOD, GB 0 1 0.500
Male SOD, GB SOD 20.8 1 2.55 × 10–6
SOD, GB, GB × SOD SOD, GB 0.3 1 0.292
Open in a new tab

The effect added for each model is highlighted in bold type. P-values are one-tailed. GB = genetic background

Discussion

On average, an increase in SOD expression in motorneurones increased life expectancy in naturally long-lived flies and in many genetic backgrounds, suggesting that the life-extending capacity of SOD is not limited to short-lived lab strains. However, we found significant variation in the effects of SOD overexpression among different genetic backgrounds. Our results point to the need to examine closely the way in which genes with major effects on aging interact with other genes within the genome. Clearly, there are many genes that influence aging, and we must be careful about drawing conclusions regarding the effects of any one gene without considering the influence of the genetic background in which it functions.

Previous studies lend further support to the important role that epistasis may play in modifying the effect of aging genes. One of the best-known examples comes from work on a well-characterized genetic pathway known to influence longevity in the nematode, Caenorhabditis elegans. In this species, daf-2, daf-12 and other genes in the dauer pathway are known to influence aging through a series of complex interactions (Larsen et al., 1995; Gems et al., 1998). For example, when the effects of various daf-2 and daf-12 allelic combinations were compared, the increase in longevity ranged from 7.8 to 22.3%. At a more general level, studies in Drosophila have identified longevity QTL whose effects differ depending on the sex (Leips & Mackay, 2000; Nuzhdin et al., 1997) or genetic background (Leips & Mackay, 2000) in which they occur. While these Drosophila lifespan QTL point to the importance of genetic background, these studies did not look at specific aging genes (though the observed QTL do contain putative aging genes, including genes for stress resistance, such as SOD, catalase and heat shock protein hsp70A; Leips & Mackay 2000).

As expected, we found that females with SOD overexpression resisted thermal stress better than their controls, but the same was not true for males. There was significant genetic variation for the ability to resist thermal stress between different genetic backgrounds; however, we did not find any evidence for a genetic background × SOD interaction. The lack of an interaction may be due in part to the fact that our thermal tolerance assay was very noisy, and so had little statistical power. We did find, however, that females that overexpressed SOD displayed enhanced thermal tolerance when compared to males.

Two possible factors might account for differences in the effect of SOD between sexes and among genetic strains. First, genotypes or sexes might have similar sensitivities to a given concentration of the SOD enzyme but might differ in their actual concentrations of SOD. The effect of the GAL4-upstream activating sequence system used to drive SOD overexpression may differ between different sexes or genotypes, with some flies showing little or no increase. Alternatively, SOD overexpression might cause levels of SOD to increase by similar amounts among different strains or sexes, but flies of different genotypes might differ in their sensitivity to these similar concentrations of the SOD enzyme. To test between these two possibilities, future studies should measure genetic variation in SOD enzyme levels as well as phenotypic responses to SOD overexpression.

Perhaps the most striking result from this study was the different effect SOD overexpression had on the males vs. the females for both lifespan and thermal tolerance. Numerous studies have found sex-specific effects on gene × genotype interactions. Studies on Drosophila in several labs have identified sex-specific variation in longevity (Nuzhdin et al., 1997; Leips & Mackay, 2000), male genotype × female genotype interactions for ovariole number (Wayne et al., 1997) and sperm competitive ability (Clark et al., 1999), and Y chromosome effects on fitness (Chippindale & Rice, 2001). In a study of genetic variation for fitness in the flour beetle, Tribolium, Wade (2000) found significant deme × sire effects on fitness among daughters but not sons.

In this study, we found sex-specific differences in how SOD altered longevity in each of our genetic backgrounds, increasing lifespan in six female backgrounds but only in one male background. Possible explanations for sex-specific longevity differences include endocrinological differences in male and female fly brains (where SOD is expressed), the residual evolutionary effects of differential costs of reproduction in males vs. females, or SOD interactions with the sex chromosomes. While our experimental flies were held as virgins, which rules out differential costs of reproduction, our results provide room for speculation on sex-specific differences. Our experimental males received their Y chromosome from the UAS-HS parental stock, so if Y × SOD interactions occurred in our experimental flies, they should be a constant across all genetic backgrounds. However, this does not preclude lifespan-altering interactions with genes in the wild-caught half of the genome.

The finding that the effect of an aging gene can depend on genetic background points to a challenge for single-gene studies. Previous studies that demonstrate how laboratory culture shortens lifespan (Matos et al., 2000; Sgro & Partridge, 2000) suggest that using these short-lived strains may lead to biases in genetic studies of aging (Promislow & Tatar, 1998; Matos et al., 2000; Sgro & Partridge, 2000; Linnen et al., 2001). Mutants that increase longevity in lab strains may simply restore the wild-type function of a trait that had changed in the lab under the combined pressures of selection and mutation. A gene that restores lifespan is not without interest but may tell us little about how genes function to alter longevity in the wild.

The isofemale strains that we examined here had experienced relatively little selection in the lab and were still relatively long-lived. Isofemale strains preserve much of the standing genetic variation in a natural population (Hoffmann & Parsons, 1988). Thus, we were able to observe the effect of an aging gene on different natural genetic backgrounds. This approach might serve as a useful model for thinking about the influence of genetic or pharmacological manipulations in human populations, where epistatic interactions with unexamined segregating alleles may lead to unanticipated consequences (Crabbe, 2001; Phillips et al., 2002).

We now need to develop a better understanding of the specific kinds of interactions that modify the influence of aging genes. Interactions between a single gene and genetic background (gene × genotype epistasis) can take various forms: the aging gene may interact with one or a few other loci, with hundreds of loci, or with specific linkage groups or chromosomes such as the sex chromosomes. As results become available from large-scale studies of gene-expression profiles, we will be able to begin constructing a clearer picture of the complex networks of gene, protein and metabolite interactions that influence the aging process.

Experimental procedures

Parental strains

We obtained SOD control and overexpression strains (described in Parkes et al., 1998) from G. Boulianne at the University of Toronto (UT). The genotype w1; P(w+)UAS-hSOD1/P(w+)UAS-hSOD1; P(w+)GAL4+, SODx39/TM3, Sb uses a GAL4 regulated upstream activating sequence (UAS) to overexpress a human SOD1 gene in the motorneurones of adult Drosophila melanogaster, leaving metabolic rate unaltered (Parkes et al., 1998). The accompanying SOD control genotype w1; P(w+)UAS-hSOD1/P(w+)UAS-hSOD1; SODx39/TM3, Sb contains the same UAS-hSOD1 insert, but it is not expressed because GAL4 is absent. Both strains were created from a w1118 recipient strain, a white-eyed base stock, and carry a D. melanogaster SODx39 internal deletion that precludes synthesis of the Drosophila SOD enzyme, and a wild-type D. melanogaster SOD allele on the TM3 balancer. We used these strains (denoted UAS-HS) as controls to ensure that the lifespan effects observed at UT could be replicated in our lab. The UAS-HS strains also provided the male parents in the experimental cross described below. Females for the experimental cross were created from wild-caught isofemale lines collected from Watkinsville, GA, USA, in August 2000. By inbreeding full siblings from these wild-caught lines for 14 generations, we generated 10 lines that were homozygous at 95% of their loci (Falconer & Mackay, 1996).

Experimental lines

We expanded the parental strains to large population size and crossed males from the UAS-HS control and overexpression lines with virgin females from each of the 10 wild-caught inbred lines to generate fully heterozygous offspring. We used Stubble as a marker to select flies carrying the third chromosome containing the SODx39 null allele and, in the case of SOD overexpression, the P(w+)GAL4 activator. Male offspring obtained their Y chromosome from the UAS-HS parent and did not carry the w allele. Female offspring carried one paternally inherited X chromosome with a w allele. Aside from these intersexual differences, all offspring within a wild-caught inbred maternal strain were nearly identical. We collected virgin male and female offspring and established mortality cages (see below) of 150 flies per cage, with three replicates per sex for every genotypic combination. Using the same protocol, we established two replicate cages/sex of UAS-HS control and overexpression strains, for a total of 140 cages and 21 000 flies.

We estimated life expectancy in 32-oz (0.95-L) plastic mortality cages provided with standard fly medium (Ashburner, 1989). Cages were held at 24 °C on a 12 : 12 h dark/light cycle and were removed to room temperature for less than 1 h every 2 days to provide fresh fly medium and remove and count all dead flies. To estimate life expectancy, we took the mean lifespan of the flies within each cage; the cage is the unit of replication. We estimated mean lifespan, or life expectancy, using WinModest (Pletcher, 1999).

Thermal tolerance

At the same time that we collected virgins for the mortality cages, we collected the same genotypes for thermal tolerance assays. These flies were held as virgins in single-sex holding cages of 150 flies each. After three weeks, we briefly chilled the flies to transfer them to 8-dram (29.6 mL) glass holding vials containing 5 mL fly medium. Each vial contained 10 flies, with five replicate vials of each sex and genotype. Immediately before the assay, we transferred the flies into empty 8-dram vials and sequestered the flies in the bottom third of the vial with a cotton plug. Vials were randomized, and the lower three-quarters of each vial was submerged in a 37 °C circulating water bath for 30 min for females and 20 min for males. Exposure times to achieve LD50 for each sex were determined in pilot experiments. Within each sex, we assayed all flies on a single day. Males and females were assayed on two consecutive days. After removal to room temperature, we transferred the flies back onto fly medium and held them at 24 °C. After 16 h, we determined the proportion of survivors within each vial. Proportional data were arcsine square-root transformed after Anscombe (1948), and then averaged within each genetic background × sex × SOD combination. Raw data are presented in Fig. 4. Here, the vial is the unit of replication.

Fig. 4.

Fig. 4

Open in a new tab

Thermal tolerance for experimental lines. Thermal tolerance in female (a) and male (b) controls and 10 wild-caught, inbred genetic backgrounds for SOD control and SOD overexpression genotypes. Error bars are +1 standard error.

Statistical analyses

UT and UGA UAS-HS strains were compared using a t-test in Excel (Sokal & Rohlf, 1995). Pearson's correlation coefficients were obtained from SPSS (10.0.5). To compare the effects of SOD, genetic background and sex on life expectancy, we used a mixed model analysis of variance using restricted maximum likelihood (REML) estimates of the variance component parameters (PROC MIXED in SAS 8.2). SOD, sex and the SOD × sex interaction were modelled as fixed effects and tested with F-statistics. Genetic background and its interaction terms were modelled as random effects. We used a likelihood ratio test to compare each model containing random effects with a model reduced by one random variable, testing each nested combination (see Table 1). The likelihood ratio test is given by twice the difference between nested models of the log likelihoods and is distributed as χ2 with degrees of freedom equal to the number of terms differing between each full and reduced model. Because this tested if the variance was significantly greater than zero (variances cannot be negative), the test was one-tailed and the significance value was half of the probability obtained from χdf2 (Littell et al., 1996). To estimate the mean squares correctly under the mixed model analysis of variance and the REML analysis, we used a Satterthwaite approximation to calculate the denominator degrees of freedom. Satterthwaite approximations can result in fractional degrees of freedom (SAS Institute, 2001).

For the thermal tolerance assay, we used the same analysis of variance and REML analysis described above with two changes. First, because the sexes were not assayed on the same day, sex was removed as a variable from the models. Second, the Satterthwaite correction was not used to analyse the thermal tolerance data because a variance component of zero for genetic background × SOD caused the degrees of freedom to artificially inflate in females when compared to males. To compare SOD overexpression flies with their control within each genetic background and sex, we estimated the variance associated with each genotype × SOD combination using ‘estimate’ statements in SAS (Littell et al., 1996).

Acknowledgments

We thank G. Boulianne for providing the UAS-HS strains, T. Haselkorn, the Promislow lab group and its undergraduate workers for assistance with experimental design and hours of fly pushing. We acknowledge M. Whitlock's lab group for invaluable comments on the manuscript, and heartily thank India and Andy Peters for their assistance with statistics. This work was funded by grants to D.P. from the National Institute on Aging (NIH grants AG21298 and AG14027) and to C.S. from the American Federation of Aging Research.

References

  1. Anscombe FJ. The transformation of Poisson, binomial and negative-binomial data. Biometrika. 1948;35:246–254. [Google Scholar]
  2. Ashburner M. Drosophila: a Laboratory Handbook Cold Spring Harbor. Cold Spring Harbor Laboratory Press; 1989. [Google Scholar]
  3. Brazhnik P, Fuente ADL, Mendes P. Gene networks: how to put the function in genomics. Trends Biotechnol. 2002;20:467–472. doi: 10.1016/s0167-7799(02)02053-x. [DOI] [PubMed] [Google Scholar]
  4. Chapman T, Hutchings J, Partridge L. No reduction in the cost of mating for Drosophila melanogaster females mating with spermless males. Proc. Royal Soc. London Series B-Biol. Sci. 1993;253:211–217. doi: 10.1098/rspb.1993.0105. [DOI] [PubMed] [Google Scholar]
  5. Chippindale AK, Rice WR. Y chromosome polymorphism is a strong determinant of male fitness in Drosophila melanogaster. Proc. Natl Acad. Sci. USA. 2001;98:5677–5682. doi: 10.1073/pnas.101456898. [DOI] [PMC free article] [PubMed] [Google Scholar]
  6. Cho JH, Nicolae DL, Gold LH, Fields CT, LaBuda MC, Rohal PM, et al. Identification of novel susceptibility loci for inflammatory bowel disease on chromosomes 1q, 3q, and 4q: Evidence for epistasis between 1p and IBD1. Proc. Natl Acad. Sci. USA. 1998;95:7502–7507. doi: 10.1073/pnas.95.13.7502. [DOI] [PMC free article] [PubMed] [Google Scholar]
  7. Clark AG, Begun DJ, Prout T. Female–male interactions in Drosophila sperm competition. Science. 1999;283:217–220. doi: 10.1126/science.283.5399.217. [DOI] [PubMed] [Google Scholar]
  8. Cordts R, Partridge L. Courtship reduces longevity of male Drosophila melanogaster. Anim. Behav. 1996;52:269–278. [Google Scholar]
  9. Crabbe JC. Use of genetic analyses to refine phenotypes related to alcohol tolerance and dependence. Alcoholism-Clin. Exp. Res. 2001;25:288–292. [PubMed] [Google Scholar]
  10. Falconer DS, Mackay TFC. Introduction to Quantitative Genetics. 4th edn. Essex; Longman: 1996. [Google Scholar]
  11. Fowler K, Partridge L. A cost of mating in female fruit flies. Nature. 1989;338:760–761. [Google Scholar]
  12. Frank L. Oxygen toxicity in eukaryotes. In: Oberley LW, editor. Superoxide Dismutase: Pathological States. CRC Press; Boca Raton: 1985. pp. 1–43. [Google Scholar]
  13. Gems D, Sutton AJ, Sundermeyer ML, Albert PS, King KV, Edgley ML, et al. Two pleiotropic classes of daf-2 mutation affect larval arrest, adult behavior, reproduction and longevity in Caenorhabditis elegans. Genetics. 1998;150:129–155. doi: 10.1093/genetics/150.1.129. [DOI] [PMC free article] [PubMed] [Google Scholar]
  14. Hoffmann AA, Parsons PA. The analysis of quantitative variation in natural populations with isofemale strains. Genetique, Selection. Evolution. 1988;20:87–98. doi: 10.1186/1297-9686-20-1-87. [DOI] [PMC free article] [PubMed] [Google Scholar]
  15. Houle D, Rowe L. Natural selection in a bottle. Am. Naturalist. 2003;161:50–67. doi: 10.1086/345480. [DOI] [PubMed] [Google Scholar]
  16. Larsen PL, Albert PS, Riddle DL. Genes that regulate both development and longevity in Caenorhabditis elegans. Genetics. 1995;139:1567–1583. doi: 10.1093/genetics/139.4.1567. [DOI] [PMC free article] [PubMed] [Google Scholar]
  17. Leips J, Mackay TFC. Quantitative trait loci for life span in Drosophila melanogaster: interactions with genetic background and larval density. Genetics. 2000;155:1773–1788. doi: 10.1093/genetics/155.4.1773. [DOI] [PMC free article] [PubMed] [Google Scholar]
  18. Lin YJ, Seroude L, Benzer S. Extended life-span and stress resistance in the Drosophila mutant methuselah. Science. 1998;282:943–946. doi: 10.1126/science.282.5390.943. [DOI] [PubMed] [Google Scholar]
  19. Linnen C, Tatar M, Promislow D. Cultural artifacts: a comparison of senescence in natural, laboratory-adapted and artificially selected lines of Drosophila melanogaster. Evol Ecol. Res. 2001;3:877–888. [Google Scholar]
  20. Littell RC, Milliken GA, Stroup WW, Wolfinger RD. SAS System for Mixed Models. SAS Institute Inc.; Cary, NC: 1996. [Google Scholar]
  21. Matos M, Avelar T. Adaptation to the laboratory: comments on Sgro and Partridge. Am. Naturalist. 2001;158:655–656. doi: 10.1086/323591. [DOI] [PubMed] [Google Scholar]
  22. Matos M, Rose MR, Pite MTR, Rego C, Avelar T. Adaptation to the laboratory environment in Drosophila subobscura. J. Evol. Biol. 2000;13:9–19. [Google Scholar]
  23. Medawar PB. An Unsolved Problem in Biology. H.K. Lewis; London: 1952. [Google Scholar]
  24. Nuzhdin SV, Pasyukova EG, Dilda CL, Zeng ZB, Mackay TFC. Sex-specific quantitative trait loci affecting longevity in Drosophila melanogaster. Proc. Natl Acad. Sci. USA. 1997;94:9734–9739. doi: 10.1073/pnas.94.18.9734. [DOI] [PMC free article] [PubMed] [Google Scholar]
  25. Orr WC, Sohal RS. Extension of life-span by overexpression of superoxide dismutase and catalase in Drosophila melanogaster. Science. 1994;263:1128–1130. doi: 10.1126/science.8108730. [DOI] [PubMed] [Google Scholar]
  26. Parkes TL, Elia AJ, Dickinson D, Hilliker AJ, Phillips JP, Boulianne GL. Extension of Drosophila lifespan by overexpression of human SOD1 in motorneurons. Nature Genet. 1998;19:171–174. doi: 10.1038/534. [DOI] [PubMed] [Google Scholar]
  27. Phillips TJ, Belknap JK, Hitzemann RJ, Buck KJ, Cunningham CL, Crabbe JC. Harnessing the mouse to unravel the genetics of human disease. Genes Brain Behav. 2002;1:14–26. doi: 10.1046/j.1601-1848.2001.00011.x. [DOI] [PubMed] [Google Scholar]
  28. Pletcher SD. Model fitting and hypothesis testing for age-specific mortality data. J. Evol. Biol. 1999;12:430–439. [Google Scholar]
  29. Promislow DEL, Tatar M. Mutation and senescence: where genetics and demography meet. Genetica. 1998;103:299–314. [PubMed] [Google Scholar]
  30. Prowse N, Partridge L. The effects of reproduction on longevity and fertility in male Drosophila melanogaster. J. Insect Physiol. 1997;43:501–512. doi: 10.1016/s0022-1910(97)00014-0. [DOI] [PubMed] [Google Scholar]
  31. Rogina B, Reenan RA, Nilsen SP, Helfand SL. Extended life-span conferred by cotransporter gene mutations in Drosophila. Science. 2000;290:2137–2140. doi: 10.1126/science.290.5499.2137. [DOI] [PubMed] [Google Scholar]
  32. SAS Institute Inc. SAS, Version 8.2. SAS Institute Inc.; Cary, NC: 2001. [Google Scholar]
  33. Sgro CM, Partridge L. Evolutionary responses of the life history of wild-caught Drosophila melanogaster to two standard methods of laboratory culture. Am. Naturalist. 2000;156:341–353. [Google Scholar]
  34. Sokal RR, Rohlf FJ. Biometry. 3rd edn. W.H. Freeman; New York: 1995. [Google Scholar]
  35. Sun JT, Tower J. FLP recombinase-mediated induction of Cu/Znsuperoxide dismutase transgene expression can extend the life span of adult Drosophila melanogaster flies. Mol. Cell. Biol. 1999;19:216–228. doi: 10.1128/mcb.19.1.216. [DOI] [PMC free article] [PubMed] [Google Scholar]
  36. Tatar M, Kopelman A, Epstein D, Tu MP, Yin CM, Garofalo RS. A mutant Drosophila insulin receptor homolog that extends life-span and impairs neuroendocrine function. Science. 2001;292:107–110. doi: 10.1126/science.1057987. [DOI] [PubMed] [Google Scholar]
  37. Trevitt S, Partridge L. A cost of receiving sperm in the female fruit fly Drosophila melanogaster. J. Insect Physiol. 1991;37:471–475. [Google Scholar]
  38. Wade MJ. Epistasis as a genetic constraint within populations and an accelerant of adaptive divergence among them. In: Wolf JB, Brodie ED, Wade MJ, editors. Epistasis and the Evolutionary Process. Oxford University Press; Oxford: 2000. pp. 213–231. [Google Scholar]
  39. Wagner A. Robustness against mutations in genetic networks of yeast. Nature Genet. 2000;24:355–361. doi: 10.1038/74174. [DOI] [PubMed] [Google Scholar]
  40. Wagner A, Fell DA. The small world inside large metabolic networks. Proc. Royal Soc. London Series B-Biol. Sci. 2001;268:1803–1810. doi: 10.1098/rspb.2001.1711. [DOI] [PMC free article] [PubMed] [Google Scholar]
  41. Wayne ML, Hackett JB, Mackay TFC. Quantitative genetics of ovariole number in Drosophila melanogaster 1. Segregating variation and fitness. Evolution. 1997;51:1156–1163. doi: 10.1111/j.1558-5646.1997.tb03963.x. [DOI] [PubMed] [Google Scholar]

RESOURCES