dj-1β regulates oxidative stress, insulin-like signaling and development in Drosophila melanogaster

Abstract

DJ-1 (or PARK-7) is a multifunctional protein implicated in numerous pathologies including cancer, sterility and Parkinson disease (PD). The popular genetic model Drosophila melanogaster has two orthologs, dj-1: α and β. Dysfunction of dj-1β strongly impairs fly mobility in an age-dependent manner. In this study, we analyze in detail the molecular mechanism underlying the dj-1β mutant phenotype. Mitochondrial hydrogen peroxide production, but not superoxide production, was increased in mutant flies. An increase in peroxide leak from mitochondria causes oxidative damage elsewhere and explains the strong reduction in mobility caused by dj-1β mutation. However, at the same time, increased levels of hydrogen peroxide activated a pro-survival program characterized by (1) an alteration in insulin-like signaling, (2) an increase in mitochondrial biogenesis and (3) an increase in the de-acetylase activity of sirtuins. The activation of this pro-survival program was associated with increased longevity under conditions of moderate oxidative stress. Additionally, the dj-1β mutation unexpectedly accelerated development, a phenotype not previously associated with this mutation. Our results reveal an important role of dj-1β in oxidative stress handling, insulin-like signaling and development in Drosophila melanogaster.

Keywords: :

• Drosophila
• dj-1β
• lifespan
• mitochondria
• oxidative stress

Introduction

DJ-1 (PARK7) is a member of the DJ-1/Pfpl superfamily with homologs in nearly all organisms. DJ-1 has been associated with a variety of human pathologies, including sterility, cancer and Parkinson disease.¹ It is abundantly expressed throughout the body with no clear tissue- or organ-specific roles. DJ-1 is found in both the cytosol, nucleus and is also associated with mitochondria.² Its cellular location is influenced by the cellular redox state and oxidative damage.³ This suggests an important role for DJ-1 in preserving cellular homeostasis.

The association of DJ-1 and mitochondrial dysfunction was inferred from studies in DJ-1-deficient mice and primary neuronal cultures derived from these mice that consistently revealed a striking sensitivity to stress-inducing stimuli, including hydrogen peroxide, rotenone and 1-methyl-4-phenyl 1,2,3,6-tetrahydropyridine (MPTP). All of these oxidative stimuli have the capacity to impair normal mitochondrial function through the inhibition of respiratory complex I.⁴ DJ-1 has recently been described as a direct regulator of complex I.⁵ Indeed, knockdown of DJ-1 decreases complex I activity, which would explain why its mutation results in a phenotype similar to rotenone or MPTP administration. The current view is that DJ-1 is a oxidative stress-response protein that defends the cell against reactive oxygen species (ROS) and mitochondrial damage. Despite this, an understanding of DJ-1 biochemical function is still missing.¹

Drosophila has two different dj-1 genes: dj-1α, expressed predominantly in testis, and the ubiquitiously expressed dj-1β. Mutations in both of these orthologs,⁶ or in dj-1β alone,⁷ cause a Parkinsonian-like syndrome in flies, characterized by a strong decrease in mobility and an increased sensitivity to oxidative stress. We, and others, have shown that mutations in dj-1β (or the analogous gene in mammals) result in increased H₂O₂ generation in isolated mitochondria.^8,9 However, it remains unknown if the increased H₂O₂ has an impact on oxidative damage in vivo, or if it just affects cellular signaling. Moreover, it is not known if higher levels of H₂O₂ are caused by an increase in the generation of superoxide or by a decrease in the efficiency of H₂O₂ detoxification systems.

The alternative oxidase (AOX) is able to bypass blockades in both complex III and IV in fungi and plants.¹⁰ When AOX is ectopically expressed in Drosophila melanogaster, it decreases mitochondrial H₂O₂ and rescues the phenotype caused by mutations in dj-1β.⁹ However, mutation of dj-1β does not shorten lifespan,¹¹ arguing against a direct role of DJ-1 in the regulation of oxidative stress in vivo. This also raises doubts as to the pathological mechanism implicated in Parkinson disease, associated with DJ-1 mutation.

In this study, we have measured levels of H₂O₂ and superoxide production in isolated mitochondria. We demonstrate that only H₂O₂ levels are increased in mutant flies. Using mass spectrometry, we show that protein damage is increased in vivo as a consequence of the increase in mitochondrial H₂O₂ leak. Moreover, different cellular pathways related to the regulation of longevity are also altered. Strikingly, we found that insulin-like signaling (IIS) is altered, mitochondrial biogenesis is increased, and the de-acetylase activity of sirtuins is increased. We present evidence that deficiency of dj-1β gene function initiates a pro-survival program. This program overcomes the negative longevity effects of oxidative stress, allowing mutant flies to live longer under conditions where oxidative damage is moderately elevated. Unexpectedly, we also found a previously undescribed effect of the dj-1β mutation: a reduction in developmental time. In summary, our data clearly demonstrate a major role for dj-1β in oxidative stress handling, insulin-like signaling and development in Drosophila melanogaster.

Results

dj-1β regulates mitochondrial H₂O₂ generation

Previously, we have demonstrated that mutation of dj-1β increases production of H₂O₂ in isolated mitochondria when using homovanillic acid as a fluorescent dye.¹¹ In the present study, we repeated our measurements (Fig. 1A) using a more sensitive probe (Amplex Red), obtaining the same results. H₂O₂ production was higher in mitochondria from dj-1β mutant flies when using pyruvate + proline as a substrate (H₂O₂ produced by complex I + III), but not when using sn-3-P-glycerate + rotenone (H₂O₂ produced by complex III). Using a new technique,¹² we measured mitochondrial superoxide generation using flow cytometry with MitoSOX™. We found that dj-1β mutant mitochondria produced less superoxide than controls with either complex I- and complex III-linked substrates (Fig. 1B). The different effects of the dj-1β mutation on superoxide and H₂O₂ production implicates dj-1β in H₂O₂, but not superoxide detoxification. This is supported by our previous data showing a reduction in the expression levels of genes involved in H₂O₂ detoxification, but no change in superoxide-handling enzymes.¹¹

Figure 1. Effects of dj-1β mutation on mitochondrial physiology. (A) Levels of H₂O₂ production (nmol/min mg prot) shown as mean ± SEM. Between four to six replicate experiments per group. (B) Levels of superoxide production (AUF, arbitrary units of fluorescence) shown as mean ± SEM. Between four to six replicate experiments per group. (C) Mitochondrial oxygen consumption state 3 respiration (nmol O₂/s mg prot) shown as mean ± SEM. Between three to six replicate experiments per group. (D) ATP content (µmol/mg protein) in fly homogenates shown as mean ± SEM, 10 replicate experiments per group. (E) Activity of individual flies recorded as average activity per hour (mean ± SEM) over a 48 h period is shown. Between 34–38 individual flies per group. (F) BNE gels of mitochondrial protein extracts from control and mutant flies. In gel, activity of complex I and IV is shown separately. p < 0.05 indicated by *. m, male; f, female.

A decrease in the levels of superoxide (with high levels of H₂O₂) can be caused by partial inhibition of the electron transport chain (ETC). In order to test this possibility, we measured mitochondrial respiration using high-resolution respirometry. We detected a 30% decrease in mitochondrial respiration in the mutant flies either with complex I- or with complex IV-linked substrates (Fig. 1C). This inhibition of respiration was accompanied by decreased ATP levels, although only in females (Fig. 1D). The unchanged ATP levels in males might be explained by the fact that their motor activity is more greatly decreased by the mutation (Fig. 1E): moving less implies a decreased energy expenditure. It has been previously shown that DJ-1 binds directly to complex I in human cells, and that its knockdown decreases complex I activity.⁵ We therefore investigated if the same phenomenon occurs in Drosophila. We separated the respiratory complexes using Blue Native Electrophoresis, and then we measured the activity of complex I and IV. We did not observe major differences either in the organization or in the activity of the respiratory complexes (Fig. 1F). This clearly suggests that the inhibition of respiration is caused by an indirect mechanism, e.g., resulting from oxidative stress, rather than by a decrease level of complex I.

dj-1β regulates oxidative stress in vivo

Because of the different effects of dj-1β on different types of ROS, we studied levels of oxidative damage in vivo. We measured five different markers of protein damage, representing three different types of modification: (1) glutamic semialdehyde (GSA) and (2) aminoadipic semialdehyde (AASA) as protein carbonyl markers; (3) carboxyethyl-lysine (CEL) and (4) carboxymethyl-lysine (CML) as glycoxidative markers and (5) malondialdehyde-lysine as a marker of lipoxidative damage (MDA-lys). These markers show a good correlation, both with mitochondrial generation of H₂O₂ and with the sensitivity to lipoxidation of biological membranes.¹³ In addition, the levels of these markers are markedly increased in several neurodegenerative diseases.¹⁴ The five markers were more abundant (independent of their chemical provenance) in mutant flies (Fig. 2A). In order to confirm that mitochondria are both the source and target of damage, we repeated our measurements using isolated mitochondria. Again, we observed higher levels of oxidative damage in mitochondria from mutant flies (Fig. 2B). It is noteworthy that levels of oxidative damage in dj-1β mutant flies were higher than in any other wild-type strain studied in our laboratory (Pamplona and Sanz unpublished results). The fact that all markers of damage were increased indicates that the defect in H₂O₂ detoxification disturbs normal metabolism, increasing oxidative damage in vivo. This has an important physiological correlate in the serious deterioration in fly mobility (Fig. 1E). A loss of climbing capacity is a characteristic feature of older flies. Flies deficient in dj-1β thus manifest at least one feature of aging, even though neither mean nor maximum lifespan was decreased. In fact, dj-1β mutant flies are long-lived compared with controls (Fig. 3A). Differences in genetic background are not responsible for this, since the effects were still present after 11 generations of backcrossing (Fig. 3B).

Figure 2.dj-1β mutation increases oxidative damage in and outside mitochondria. (A) Levels of protein damage in fly homogenates mean ± SEM is shown. Between three to seven replicate experiments per group. (B) Levels of protein damage in fly mitochondria mean ± SEM is shown. Between three to six replicate experiments per group. GSA, glutamic semialdehyde; AASA, aminoadipic semialdehyde; CEL, carboxyethyl-lysine; CML, carboxymethyl-lysine; MDA-lys, malondialdehyde-lysine. Data are average ± SEM of at least three independent samples per group. p < 0.05 indicated by *. m, male; f, female.

Figure 3. Effect of dj-1β mutation on lifespan. (A) Survival of wild-type strains (DAH and OR) and dj-1β mutant flies (DJGE) at 25°C. n = 80 flies per experiment. Median (maximum) lifespan in days, males DAH 41 (51), males OR 65 (72), males DJGE 58 (60); females DAH 74 (79), females OR 81 (86), females DJGE 86 (86); DAH flies live significantly shorter than DJGE or OR (Log-Rank Test p < 0.001). (B) Survival of wild-type (DAH), dj-1β mutant flies (DJGE) and dj-1β mutant flies backcrossed for 11 generations into DAH (DJGE (DAH)) at 25°C. n = 80 flies. Median (maximum) lifespan in days, males DAH 35 (39), males DJGE 56 (74), males DJGE (DAH) 48 (63); females DAH 74 (79), females DJGE 79 (81), females DJGE(DAH) 77 (81); DAH flies live significantly shorter than DJGE or DJGE (DAH) (Log-Rank Test p < 0.001). (C) Survival of wild-type strains (DAH and OR) and mutant flies (DJGE) at 29°C. n = 200 flies per experiment. Median (maximum) lifespan in days, males DAH 38 (40), males OR 42 (47), males DJGE 36 (38); females DAH 46 (49), females OR 49 (56), females DJGE 46 (49); DJGE male flies live significantly shorter than DJGE or OR (Long Rank Test p < 0.001). No differences in lifespan between DJGE and DAH females. (D) Survival of wild-type (DAH) and dj-1β mutant flies (DJGE) at 25°C in high (HC) and low caloric- (LC) content diets. Male DAH 40 (54), male DJGE 65 (77), male DAH DR 40 (61), male DJGE DR 77 (89); female DAH 49 (68), female DJGE 77 (93), female DAH DR 74 (89), female DJGE DR 98 (108). DJGE flies live significantly longer than DAH in both diets (Log-Rank Test p < 0.001). m, male; f, female.

dj-1β regulates lifespan by activating of pro-survival pathways

If oxidative stress is so highly increased in dj-1β mutant flies as to cause a serious deterioration in motor function at early ages, why is lifespan not decreased? One possibility is that mitochondrial oxidative stress is not a key factor in lifespan determination, as previously proposed.¹⁵ Another alternative is that an increase in H₂O₂ triggers an adaptive response that extends lifespan. That has been described both in worms and in flies related to glucose restriction¹⁶ or superoxide dismutase overxpression.¹⁷ In relation to that, oxidative stress activates Jun-N-terminal kinase (JNK), which represses the expression of the insulin-like peptides (Ilp) Ilp2 and Ilp5.¹⁸ Consistent with this, decreased expression of Ilp2, -3 and -5 is a characteristic of long-lived flies under dietary restriction (DR).¹⁹ We measured the expression levels of Ilp2 and Ilp5 and observed a decrease in the dj-1β mutant background expression levels (Fig. 4A). Conversely, a large increase was seen in Ilp3, which was previously observed to be unresponsive to externally applied oxidative stress.¹⁸

Figure 4. The dj-1β mutation activates pro-survival pathways in Drosophila melanogaster. (A) Q-PCR of Ilp’s 2, 3 and 5 from dj-1β mutant and control flies, normalized to Actin 88F. Data are presented as mean ± SEM (six replicate experiments). (B) Mitochondrial copy number assayed by Q-PCR of 16S rRNA gene from dj-1β mutant flies and control flies, normalized to levels of Rpl32. Data are presented as mean ± SEM (four to six replicate experiments). (C) Western blots analysis of complex I (NDUFS3 subunit) and loading control PDH (E1α subunit). Data are presented as mean ± SEM (four replicate experiments). (D) Sirtuin activity in homogenates of dj-1β mutant flies and control flies (five replicate experiments). p < 0.05 indicated by *. m, male; f, female. M, dj-1β mutant; C, control.

DR or interruption in IIS signaling are characterized by simultaneous activation of several pro-survival pathways in parallel.²⁰ As we have shown, IIS signaling is altered in dj-1β mutant flies. For this reason, we decided to investigate other pathways associated with longevity. Mitochondrial biogenesis is increased during DR,²¹ and overexpression of the Drosophila PGC-1 homolog Spargel extends lifespan.²²We measured mitochondrial DNA (mtDNA) levels and found them to be increased in dj-1β mutants (Fig. 4B). Additionally, we investigated the expression of complex I, measuring the concentration of one of its subunits (NDUFS3) by western blotting. Complex I expression was higher in both males and female flies in dj-1β mutant flies (Fig. 4C), although the differences were statistically significant only in females. Mitochondrial biogenesis is stimulated by a decrease in ATP production. This finding is therefore consistent with the inhibition of respiration in the mutant flies. Strikingly, proportionate decrease in respiration was similar to the increase in mtDNA copy number (approximately 20%). Moreover, when mitochondrial respiration was analyzed in homogenates, no differences were observed between control and mutant flies (data not shown). Under these experimental conditions, oxygen consumption is measured as the combination of the respiratory efficiency and the mitochondrial mass (mitochondrial number). The increase in the amount of mitochondria in the mutants partially compensates for the loss of respiratory3 activity.

An increase in activity of the sirtuin SIR2 is considered to be enough for lifespan extension both in worms²³ and flies²⁴ and essential for longevity extension mediated by DR,²⁵ although these findings have recently been questioned.²⁴ We found that sirtuin activity was increased in dj-1β mutant flies (Fig. 4D), consistent with other reports of increased activity after oxidative damage.²⁶ Interestingly, sirtuin activity in dj-1β mutant flies was higher than in any wild-type strain in our laboratory (data not shown).

The upregulation of pro-survival pathways may explain the surprising longevity of dj-1β mutant flies. The effects of increased oxidation on lifespan would thus be compensated by downstream activation of global regulators of longevity such as sirtuins. If this is true, a further increase in oxidative stress beyond a critical threshold should be able to overcome these pathways and reduce longevity in the dj-1β mutant flies. To test this hypothesis, we cultured controls and mutant flies at 29°C. Under such conditions, dj-1β mutant flies were observed to have shorter lifespans than controls (Fig. 3C), demonstrating that the activation of pro-survival pathways can be overcome by increased levels of ROS.

Alterations to mitochondrial biogenesis and sirtuin activity in dj-1β mutant flies are as predicted, if regulated by IIS in a similar manner to DR. We therefore investigated whether dj-1β mutant flies are still responsive to DR. We cultured mutant and control flies in diets with high or low caloric content. In both diets, mutant flies lived longer than control flies, demonstrating that the mutants remain responsive to DR (Fig. 3C). This indicates that functional deficiency of dj-1β does not induce a condition equivalent to DR, and also supports the notion that oxidative stress is not essential for lifespan extension during DR.

dj-1β regulates the rate of development

We observed that the phenotype associated with the dj-1β mutation appears at a younger age in each new generation, indicating that this mutation could be conferring some type of adaptive advantage. We also observed that mutant flies developed and eclosed faster than controls. We quantified this phenomenon by counting the number of flies that eclosed after 5 d of mating. No difference in the total number of male and female flies was observed (Fig. 5A); however, there were slightly fewer males. More than 50% of mutant flies had emerged by day 4, while wild-type flies did not reach the same levels until day 5. The phenotype is quite interesting, since dj-1β mutant flies develop faster than any other wild-type strain available in our laboratory (data not shown). To the best of our knowledge, this is the first time that a mutation has been described that accelerates so significantly the rate of development in Drosophila. Alterations in development can, in principle, be caused by a dysregulation of IIS. In dj-1β mutant flies, we found a strong increase, approximately 6- and 3-fold in males and females, respectively, in the expression of Ilp3, whereas the expression of other Ilps was lower (Fig. 4A). Developmental differences appear when larvae stop feeding and crawl to become pupae (Fig. 5B). Interestingly, the transition between larvae to pupae is characterized by an increase in Ilp3 expression.²⁷ Our observations (Fig. 4A) suggest a tight correlation between these molecular and physiological phenomena and support a causal relationship.

Figure 5. The dj-1β mutation accelerates development. (A) Number of eclosing flies per day, 10 different replicate experiments in control and dj-1β mutant flies. (B) Representative images of those taken daily during developmental time experiments. Stages from pupariation to eclosion are shown.

Discussion

Our results reveal a specific role for dj-1β in the regulation of mitochondrially generated ROS. dj-1β influences the level of hydrogen peroxide, but not that of superoxide levels. This difference indicates that dj-1β is not directly involved in the generation of ROS at the level of the ETC, but rather at the level of ROS detoxification. Consistent with this, we previously reported that antioxidant genes responsible for peroxide detoxification are downregulated in mutant flies, but genes related to superoxide detoxification are expressed at similar levels as in wild-types flies.¹¹ This supports a role for dj-1β as a regulator of genes involved in peroxide, but not superoxide detoxification, and would explain our ROS findings in isolated mitochondria. Moreover, dj-1β mutant flies are extremely sensitive to drugs that target glutathione metabolism,²⁸ which is a central component of the defense against peroxides. All these observations support a central role for dj-1β in the detoxification of mitochondrial peroxides. At the moment, we cannot say if the protein acts as a peroxiredoxin or as a regulator of antioxidant genes. In the literature, both possibilities have been suggested.^8,11,28 However, whatever the mechanism of action, it is important to highlight that dj-1β is not involved in superoxide detoxification. In fact, a lack of superoxide detoxification creates a different phenotype with regard to mitochondrial physiology. Sod2 knockdown specifically affects proteins with iron-sulfur clusters, e.g., complexes I and II and aconitase located in the mitochondrial matrix.²⁹ Such a phenotype was not observed dj-1β mutants flies.

One important difference between hydrogen peroxide and superoxide is reactivity. H₂O₂ can diffuse far away from its origin, whereas superoxide cannot go through lipid membranes and has a much more limited area of action. This suggests that a decreased detoxification of mitochondrial H₂O₂ will increase oxidative stress both inside and outside of mitochondria, and this is exactly what we observed when we measured protein damage. The damage elicited is nonspecific, as shown by high levels of all markers of damage, independent of their chemical origin. The fact that oxidative damage was higher in dj-1β mutants than in any other wild-type strain in our laboratory [including Canton-S, Oregon-R and Dahomey (DAH)] suggests a central role of this protein in the regulation of oxidative stress. The increase in hydrogen peroxide is associated with decreased ETC activity, reducing its capacity for ATP production. The ultimate effect is to limit the available energy, consistent with our finding of decreased locomotor activity in young flies. Decreased activity and high levels of mitochondrial ROS are typical traits of older flies, which are also expressed in young dj-1β mutant flies.¹¹ Alterations in different types of ROS (superoxide vs. H₂O₂) have distinct effects on physiology and lifespan in different organisms. In worms, an increase in superoxide levels does not affect lifespan³⁰ or even extends it,¹² whereas, in flies, an increase in superoxides is associated with a shorter lifespan³¹ and muscle degeneration.³² However, in flies, H₂O₂ can exert a positive effect on longevity through direct administration,³³ mutation of dj-1β (this study) or overexpression of sod1³⁴ or sod2.³⁵ Extension of lifespan as a consequence of increase in mitochondrial oxidative stress has been called mitohormesis,³⁶ and the extension is normally explained by a long-term increase in antioxidant levels.³⁶ However, increased administration or overexpression of the same upregulated antioxidants do not clearly extend lifespan either in worms or mammals (reviewed in ref. 37). Our results indicate that the activation of pro-survival mechanism(s) related to IIS signaling and growth is a more feasible explanation for mitohormesis.³⁸ Surprisingly, decreasing H₂O₂ in the mitochondrial matrix through overexpression of catalase shortens Drosophila lifespan.³⁹ Data related to the effect of dj-1β on longevity are contradictory. In flies, mutation of dj-1β has been reported both to decrease and to have no effect on longevity;^6,40 whereas, no change in lifespan was seen in mice, in spite of higher levels of mitochondrial H₂O₂.⁸ Differences in longevity are sometimes confounded by genetic background effects that have not been properly controlled. In our case, differences between controls and mutants cannot be attributed to background effects, since mutant flies are long-lived before and after backcrossing into the Dahomey reference strain over 11 generations. In our laboratory, the longest-lived wild-type strain is Oregon-R (OR). OR is also the strain with the lowest mitochondrial ROS production both in vitro¹¹ and in vivo (Pamplona and Sanz, unpublished results). Differences in longevity between the OR and DAH strains are independent of temperature. However, the lifespan increase associated with the dj-1β mutation was not observed when temperature was increased. On the other hand, mutant flies liver longer when grown on diets of both high and low caloric content. The effect of temperature, but not DR, is related to oxidative stress in Drosophila lifespan extension.¹⁵ The fact that the survival advantage of dj-1β mutants was lost at high temperature, but not when caloric content was altered, is consistent with the inference that dj-1β is a regulator of oxidative stress in vivo. Moreover, when dj-1β mutant flies were cultured at high temperatures, they became immobilized at the bottom of the vial, whereas wild-type flies move faster. This supports the idea that age-related loss of mobility is caused by a ROS-mediated mechanism.

High levels of H₂O₂ from mitochondria can activate an adaptive response (mitohormesis) extending lifespan through the preservation of vital tissues.⁴¹ Consistent with this, dj-1β mutant flies exhibited decreased expression of Ilp2 and Ilp5 (IIS signaling), increased levels of mtDNA and complex I (mitochondrial biogenesis) and more activity of sirtuins in fly homogenate. Independently, it has been shown that JNK is activated by mitochondrial H₂O₂.⁴² Once active, JNK specifically reduces expression levels of Ilp2,¹⁸ as also happens in dj-1β mutant flies. Strikingly, the knockout of Ilp2 extends Drosophila lifespan.⁴³ Downstream in the IIS pathway, it was found that knockdown models for dj-1 (α or β) exhibit a decrease in AKT phosphorylation,⁴⁴ which is an essential step in the activation of a pro-survival program mediated by FOXO translocation to the nucleus.⁴⁵ Our results, and those of others, indicate that a reduction in AKT phosphorylation is a common, non-cell-autonomous response when oxidative damage is increased by internal or external causes.^18,46 The activation of such pro-survival pathways related with IIS signaling and growth can explain the positive effects of mitohormesis on lifespan independently of an upregulation in expression of antioxidants.⁴¹ We must remember that antioxidants by themselves do not extend lifespan, whereas the activation of these other pathways do without a consistent changing in antioxidant defenses.

It is reasonable that an increase in oxidative stress can be interpreted by the cell as a signal to activate pro-survival pathways. For example, an increase in mitochondrial biogenesis is essential for lifespan extension during DR, both in Drosophila²¹ and in mouse.⁴⁷ Either ROS or a partial inhibition of mitochondrial respiration can promote mitochondrial biogenesis by activating PGC-1α⁴⁸ or the AMPK signaling pathway,⁴⁹ respectively. It is important to note that higher mitochondria density can confer benefits without changing the total level of oxidative damage,¹⁵ and the overexpression of spargel (homolog of PGC-1) extends Drosophila lifespan. On the other hand, decreased respiration activates a survival-promoting strategy that is present in yeast,¹⁰ worms,⁵⁰ flies⁵¹ and even mice.⁵² An interruption of normal electron flow can increase or decrease longevity depending upon downstream consequences. For example, in worms, mutations in complex I, III or IV increase lifespan,¹² whereas disruption of complex II shortens lifespan.⁵³ Something similar occurs in Drosophila; when specific subunits of complex I, III, IV and V, but not II, are knocked down, lifespan can be extended.⁵⁴ From this point of view, a mild inhibition of respiration caused by more oxidative stress (dj-1β mutants) would be equivalent to a mild reduction in the levels of respiratory complexes I, III or IV (RNAi lines).⁵⁴

High levels of oxidative stress damage DNA, promoting the mobilization of sirtuins,²⁶ which can protect against genetic and epigenetic modifications of the genome.⁵⁵ dj-1β mutant flies have the highest levels both of oxidative damage and sirtuin activity compared with all wild-type strains in our laboratory. This should protect the mutant flies against genetic and epigenetic modifications, contributing to their extended lifespan. In the past, Sir2 was considered a major regulator of longevity and an instrumental gene for the prolongation of lifespan mediated by DR.²⁵ However, these results have recently been called into question.²⁴ It is possible that Sir2 does not directly regulate longevity, but that the protective effects of Sir2 or other sirtuins can nevertheless help overcome oxidative damage caused by the dj-1β mutation. In fact, knockdown of Sir2 in Drosophila fat body reduces resistance to starvation,⁵⁶ stress that initially is associated with an increase in oxidative stress.⁵⁷

It is generally accepted that the activation of a single pathway is insufficient to extend lifespan. For example, DR and other genetic strategies that increase lifespan are characterized by simultaneous changes in the expression of several different genes.⁵⁸ In the future, it would be interesting to determine whether the relevant pathways are regulated by the same master mechanism or by different signals. Different tissues are more sensitive to varying kinds of damage. For example, muscle is particularly sensitive to ROS, and sarcopenia is very sensitive to oxidative stress.⁵⁹ Alternatively, damage to muscle can activate a signal that leads to the preservation of other tissues. The dj-1β mutant fly is a clear example of how high levels of oxidative damage can be compatible with a long lifespan, a rare phenotype in Drosophila.³⁷

During our experiments, we observed that the mutant phenotype appeared earlier in each successive generation. Additionally, mutant flies eclosed more quickly than controls. This suggests that the mutation confers some adaptative advantage under our laboratory conditions. Strikingly, more rapid development was accompanied by a several-fold higher expression level of Ilp3 in mutant flies. Fly development is strictly regulated by the expression of specific llps.⁶⁰ For example, the expression of Ilp3 is increased prior to the pupariation when the non-feeding state starts.²⁷ Coincidentally, this is the moment when differences in the developmental rate between mutant and wild-type flies became evident. Increased Ilp3 expression can result from a compensatory mechanism induced by the decreased expression of other Ilps.^18,61 ROS have been proposed as regulators of development,⁶² and H₂O₂ promotes cellular proliferation,⁶³ differentiation⁶⁴ or senescence.⁶⁵ The effects of H₂O₂ depend on cell type, cell cycle stage and also concentration. The fact that dj-1β mutant flies develop faster than controls suggests that ROS may indeed be an important regulator of development. In fact, there is an increase in H₂O₂ levels during the transition from feeding to non-feeding states in Drosophila larvae.⁶⁶ Moreover, the administration of acetyl cysteine in C. elegans¹² reduces lifespan in long-lived mutants with overproduction of ROS and also slows development. Antagonist pleiotropy suggests that aging results from the accumulation of mutations that are not subject to natural selection, because the accumulation is only detrimental at an advanced age. These mutations may have a beneficial effect at earlier stages. The dj-1β mutation may be a good example of this phenomenon. It promotes more rapid development, but as a trade-off, it impairs mobility at an earlier age. In an environment where resources are abundant (our laboratory conditions), this is advantageous and can be favorably selected. This would explain why the phenotype appears at an earlier age in each successive generation.

In summary, DJ-1β regulates oxidative stress, insulin signaling and development in Drosophila melanogaster. Since signaling pathways are highly conserved across different phyla, these data should prove useful to scientists studying the role of DJ-1 in human diseases such as Parkinson disease or cancer.

Materials and Methods

Fly stocks and lifespan studies

Wild-type strains used were our own stocks of Dahomey (DAH) and Oregon-R (OR). The loss-of-function dj-1β mutant fly or dj-1β^GE23381 (DJGE in all figures) has been previously described.⁷ Lifespan experiments were repeated in DJGE flies after 11 generations of backcrossing into DAH; both mitochondrial and nuclear DNA were in DAH background in DJGE (DAH) flies. Flies were maintained on standard media (1% agar, 1.5% sucrose, 3% glucose, 3.5% dried yeast, 1.5% maize, 1% wheat, 1% soya, 3% treacle, 0.5% propionic acid, 0.1% Nipagin). For experiments evaluating the effect of caloric content on lifespan, we used the diets described by reference 67 (high caloric-content diet: 2% agar, 15% dried yeast, 15% sucrose, 3% nipagin, 0.3% propionic acid; low caloric-content diet: 2% agar, 6.5% dried yeast, 6.5% sucrose, 3% nipagin, 0.3% propionic acid). Flies were collected using CO₂ anesthesia within 24 h of eclosion and then kept at a density of 20 flies (males and females separately) per vial at 25°C or 29°C in a controlled 12 h light:dark cycle. Flies were transferred to new vials every 2–3 d. Lifespan studies were performed with at least 80 flies per genotype. The number of dead flies was recorded every 2–3 d, and from this, the mean and maximum lifespan (the last 10% of surviving flies) were calculated. For the remainder of the experiments, 10-day-old flies were used.

Developmental assay

Five virgin females were crossed for five days with three males. Pictures of the vials were taken every day. After eclosion, the number of flies eclosing per day was counted for 10 d.

Mitochondrial O₂ consumption and measurement of mitochondrial H₂O₂ and superoxide production

Mitochondria were isolated as described⁶⁸ and used to measure ROS production and oxygen consumption. Mitochondrial oxygen consumption was measured by high-resolution respirometry using an oxygraph 2-K (Oroboros). Mitochondria were incubated in assay buffer (120 mM KCl, 5 mM KH₂PO₄, 3 mM Hepes, 1 mM EGTA, 1 mM MgCl₂, 0.2% bovine serum albumin, pH 7.2 at 25°C) and a combination of complex I- (5 mM pyruvate + 5 mM proline) or complex IV- (50 μM TMPD + 1 mM ascorbate + 10 μM Antimycin + 10 μM rotenone to inhibit complex III and complex I, respectively) linked substrates. State 3 respiration was initiated by adding 500 µM of ADP to the sample. In parallel, we assayed H₂O₂ production as previously described.⁶⁹ In order to measure superoxide, we adapted the method described by reference 12 in C. elegans to Drosophila. Briefly, isolated mitochondria were incubated with assay buffer containing complex I- (5 mM pyruvate + 5 mM proline) or complex III-linked [20 mM sn-3-glycerol-phosphate with and without rotenone (10 µM)] substrates and 5 μM MitoSox™ at 25°C for 10 min. Mitochondria were then sorted using an Accuri C6 Flow cytometer (BD Accuri Cytometers), and fluorescence was measured using the appropriate channel and software provided by the manufacturer.

Western blot analysis

Protein extraction, SDS-PAGE and western blotting were as described in reference 69. Antibodies used together with the appropriate secondary antibodies, were as follows: against the complex I NDUFS3 subunit (MitoSciences) used at 1:20,000; against the pyruvate dehydrogenase (PDH) E1α subunit (MitoSciences) used at 1:10,000; the secondary antibody was HRP-conjugated horse anti-mouse IgG (H⁺L) (Vector Laboratories), used at 1:10,000.

Blue native-PAGE (BNE) and in-gel histochemistry

BNE was performed using NativePAGE^TM Novex^® 3–12% Bis-Tris Gels according to reference 69. For in-gel assay of complex I, gels were incubated in complex I activity buffer (2 mM Tris, 0.1 mg/ml NADH, 2.5 mg/ml Nitroblue tetrazolium chloride, pH 7.4) for 20 min, after which they were fixed in destain solution for 20 min. For in-gel assay of complex IV, gels were incubated in complex IV activity buffer [5 mg 3.3^′-diamidobenzidine tetrahydrochloride (DAB) dissolved in 9 ml phosphate buffer (0.05 M, pH 7.4), 1 nM catalase (20 μg/ml), 10 mg cytochrome c and 750 mg sucrose] for 2 h, after which they were fixed in destain solution for 20 min. All activity staining was performed at room temperature.

ATP measurement

Total ATP content was measured using the luciferin-luciferase-based ATP determination kit (Molecular Probes). One to five flies were homogenized in 100 µl 6M guanidinium chloride. Homogenates were centrifuged at 16,000 g_max for 5 min. Supernatants were dissolved 1:750 in TE buffer. Ten microliters of diluted sample were mixed with reaction solution containing luciferin-luciferase, and triplicates were measured by luminometry. A standard curve of different ATP dilutions was run in parallel, and results were normalized by protein content (mg) after application of the standard curve.

Fly activity

The locomotor activity of individual flies was measured in a Digitherm CircKinetics monitoring incubator (Tritech Research) at 25°C in a controlled 12:12 h dark:light cycle. Individual flies were put in capillaries with standard fly food, and their activity was monitored over 72 h using TriKinetics Activity Monitors (Trikinetics Inc.). Flies were acclimated for the first 24 h; hand activity was monitored during the next 48 h; the number of times that flies crossed the center of the vial per hour was counted and integrated using TriKinetics software.

RNA/DNA quantification

Isolation of mRNA, cDNA and q-RT-PCR synthesis was as described in full detail in reference 69. For mtDNA copy number, total DNA was isolated according to reference 69 and analyzed by q-PCR using the same conditions as used for mRNA analysis. Data were extracted and analyzed using Applied Biosystems StepOne software version 2.1. Primer sequences were 16S (forward: ttcgtccaaccattcattcc, reverse: tctaacctgcccactgaa), Act88F (forward: agggtgtgatggtgggtatg, reverse: cttctccatgtcgtcccagt) Ilp2 (forward: agcaagcctttgtccttcatctc, reverse: acaccatactcagcacctcgttg), Ilp3 (forward: tgtgtgtatggcttcaacgcaatg, reverse: cactcaacagtctttccagcaggg), Ilp 5 (forward: gaggcaccttgggcctattc, reverse:catgtggtgagattcggagc), Rpl32 (forward: gttcgatccgtaaccgatgtt, reverse: caccagtcggatcgatatgc).

Assay of sirtuin activity

Sirtuin activity was assayed in fly homogenates, based on the deacetylation of the substrate Fluor de Lys-SIRT1 (Enzo Life Sciences) according to reference 69.

Analysis of protein oxidative damage markers by mass spectrometry

The levels of markers, aminoadipic semialdehyde (AASA, oxidation), glutamic semialdehyde (GSA, oxidation), carboxymethyl-lysine (CML, glycoxidation), carboxyethyl-lysine (CEL, glycoxidation) and malondialdehydelysine (MDAL, lipoxidation) were determined by gas chromatography/mass spectrometry (GC/MS) according to reference 69.

Statistical analysis

Data were analyzed using GraphPad Prism 4 and the unpaired t-test was used for statistical testing of mutant and wild-type flies of the same sex. Lifespan data were analyzed using the Kaplan Meier Log-Rank Test. The level of statistical significance was established as p < 0.05.

Abbreviations:
MPTP	=	1-methyl-4-phenyl 1,2,3,6-tetrahydropyridine
AOX	=	alternative oxidase
AASA	=	aminoadipic semialdehyde
CEL	=	carboxyethyl-lysine
BNE	=	blue native-polyacrylamide gel electrophoresis
CML	=	carboxymethyl-lysine
DAH	=	Dahomey
DR	=	dietary restriction
DJGE	=	dj-1βGE23381
ETC	=	electron transport chain
GSA	=	glutamic semialdehyde
JNK	=	Jun-N-terminal kinase
Ilp	=	insulin-like peptides
IIS	=	insulin-like signalling
MDA-lys	=	malondialdehyde-lysine
mtDNA	=	mitochondrial DNA
PDH	=	pyruvate dehydrogenase
OR	=	Oregon-R
PD	=	Parkinson disease
ROS	=	reactive oxygen species

Acknowledgments

This work was supported by the Academy of Finland, the Tampere University Hospital Medical Research Fund, the European Research Council, the Spanish Ministry of Science and Innovation and the Autonomous Government of Catalonia.