JNK/FOXO-mediated Neuronal Expression of Fly Homologue of Peroxiredoxin II Reduces Oxidative Stress and Extends Life Span*

Activation of c-Jun N-terminal kinase (JNK) signaling in neurons increases stress resistance and extends life span, in part through FOXO-mediated transcription in Drosophila. However, the JNK/FOXO target genes are unknown. Here, we identified Jafrac1, a Drosophila homolog of human Peroxiredoxin II (hPrxII), as a downstream effecter of JNK/FOXO signaling in neurons that enhances stress resistance and extends life span. We found that Jafrac1 was expressed in the adult brain and induced by paraquat, a reactive oxygen species-generating chemical. RNA interference-mediated neuronal knockdown of Jafrac1 enhanced, while neuronal overexpression of Jafrac1 and hPrxII suppressed, paraquat-induced lethality in flies. Neuronal expression of Jafrac1 also significantly reduced ROS levels, restored mitochondrial function, and attenuated JNK activation caused by paraquat. Activation of JNK/FOXO signaling in neurons increased the Jafrac1 expression level under both normal and oxidative stressed conditions. Moreover, neuronal knockdown of Jafrac1 shortened, while overexpression of Jafrac1 and hPrxII extended, the life span in flies. These results support the hypothesis that JNK/FOXO signaling extends life span via amelioration of oxidative damage and mitochondrial dysfunction in neurons.


Activation of c-Jun N-terminal kinase (JNK) signaling in neurons increases stress resistance and extends life span, in part through FOXO-mediated transcription in Drosophila.
However, the JNK/ FOXO target genes are unknown. Here, we identified Jafrac1, a Drosophila homolog of human Peroxiredoxin II (hPrxII), as a downstream effecter of JNK/FOXO signaling in neurons that enhances stress resistance and extends life span. We found that Jafrac1 was expressed in the adult brain and induced by paraquat, a reactive oxygen species-generating chemical. RNA interference-mediated neuronal knockdown of Jafrac1 enhanced, while neuronal overexpression of Jafrac1 and hPrxII suppressed, paraquat-induced lethality in flies. Neuronal expression of Jafrac1 also significantly reduced ROS levels, restored mitochondrial function, and attenuated JNK activation caused by paraquat. Activation of JNK/FOXO signaling in neurons increased the Jafrac1 expression level under both normal and oxidative stressed conditions. Moreover, neuronal knockdown of Jafrac1 shortened, while overexpression of Jafrac1 and hPrxII extended, the life span in flies. These results support the hypothesis that JNK/FOXO signaling extends life span via amelioration of oxidative damage and mitochondrial dysfunction in neurons.
FOXO transcription factors are key regulators of growth, metabolism, life span, and stress resistance in various organisms, including Drosophila (1,2). FOXO is regulated by the insulin signaling pathway and the stress-induced JNK 4 signaling pathway (3,4). Oxidative stress activates the stress-responsive JNK, which promotes FOXO nuclear localization and upregulates expression of antioxidant proteins (5,6).
In Drosophila, neuronal activation of JNK/FOXO signaling confers resistance to oxidative stress and extends life span (4,7). Neurons are particularly susceptible to oxidative damage because of their high levels of ROS production and relatively low levels of antioxidant enzymes (8). Thus, activation of the JNK/FOXO pathway in neurons may extend life span through up-regulation of anti-oxidative stress genes. However, little is known regarding the JNK/FOXO target genes in neurons.
Thiol-reducing systems are important reducers of many oxidative stressors, such as peroxide (9). Peroxiredoxin (Prx), also called thioredoxin peroxidase, eliminates hydroperoxide with thioredoxin as an immediate hydrogen donor and reduces ROS levels (10). Among six distinct mammalian Prxs (I-VI), Prx II is exclusively expressed in the brain (11), suggesting that Prx II may play an important role in response to oxidative stress in neurons. However, the regulation of PrxII expression in neurons has not been elucidated. In this study, we demonstrated that neuronal expression of Jafrac1, a Drosophila homologue of human Prx II (hPrxII), was regulated by JNK/FOXO signaling, promoted resistance to oxidative stress, and extended the life span of the flies.
Paraquat Treatment-To investigate the effect of oxidative stress in the Drosophila model, adult flies (5 days old) were exposed in 20 mM paraquat for 24 h. The flies were kept in vials containing 1 ml of 1.3% agar for 6 h for starvation before the paraquat treatment. The flies were then transferred to vials containing a 22-mm filter paper disk soaked with 20 mM paraquat (methyl viologen, Sigma) in 5% sucrose solution. The data are presented as means Ϯ S.E. The statistically significant differences were examined using the Student's t test (Microsoft Excel) and p Ͻ 0.05 was accepted as statistically significant.
Measuring ROS Levels in Drosophila-To measure the intercellular ROS level in Drosophila, we used the method described by Strayer et al. (15). Non-fluorescent 2,7-dichlorofluorescein di-acetate (Molecular Probes) is a cell permeable dye and can be converted into 2,7-dichlorofluoroscein by interacting with hydrogen peroxide (16). Flies (3 days old) were treated with 20 mM paraquat for 24 h and collected in tubes containing 500 l of PBST (PBS containing 0.1% Tween 20). The flies were then homogenized, and 100 l of each supernatant was transferred into a 96-well plate. After adding 50 M 2,7-dichlorofluorescein diacetate to the samples, the fluorescence intensity was measured every 5 min for 15 min using a fluorescence microplate reader (FLUOstar Optima, BMG Laboratory) and the fluorescence intensity (excitation 485 nm and emission 640 nm) was quantified. Three independent experiments with 50 flies in each experiment were performed.
ATP Assay and mtDNA PCR Analysis-Total ATP production was measured as previously described (17). Heads of 10 2-day-old flies were dissected and homogenized in extraction buffer (100 mM Tris and 4 mM EDTA, pH 7.8) followed by quick-freezing in liquid nitrogen and boiling for 3 min. The samples were then centrifuged to collect the supernatant and mixed with luminescent solution (Enliten kit, Promega, Madison, WI). The luminescence was measured using a luminometer (FLUOstar Optima, BMG Laboratory), and the results were compared with standards. The ATP level was measured relative to total protein concentration. For mtDNA PCR, total DNA was extracted from 2-day-old flies and subjected to PCR amplification with primer sets for various target genes (supplemental Table 1). The genomic DNA level of rp49 of each sample was used as the loading control. Results are expressed as -fold change relative to control.
Semi-quantitative Reverse Transcription-PCR Analysis-For reverse transcription-PCR analysis, 1 g of total RNA was used with the oligo(dT) primer and avian myeloblastosis virus reverse transcriptase (Roche Biochemical) to generate first strand cDNA. Next, 1 l of the cDNA was subjected to PCR amplification with the primer sets for various target genes (supplemental Table 1). PCR conditions were 94°C for 5 min, followed by 30 cycles of 94°C for 30 s, 55°C for 30 s, 72°C for 1 min, and a final extension at 72°C for 10 min using a Thermal Cycler (Applied Bioscience). PCR products were resolved in 1.5% agarose gels and visualized by ethidium bromide staining. The rp49 gene was used as a control.
Protein Analysis and Immunostaining-Western blot analyses were performed as described previously (18). The antibodies against JNK (Santa Cruz Biotechnology, Santa Cruz, CA) and phospho-JNK (Promega) were used to detect JNK activation. The antibody against dFOXO (a gift from O. Puig) was used to detect FOXO translocation to the nucleus. To detect pJNK and FOXO in vivo, the third instar larvae were fed with 20 mM paraquat for 24 h, and then cuticles were dissected and fixed in 4% paraformaldehyde in PBS for 1 h at room temperature and incubated with the pJNK antibody (1:200) or FOXO antibody (1:500) and, subsequently, with Alexa Fluor 594-conjugated anti-rabbit IgG (1:200, Molecular Probes). Fluorescence images were acquired using an Axiovert 200M microscope (Carl Zeiss, Germany).
Life Span Assay-For longevity experiments, 1-and 2-dayold adult male or female flies were collected (10 per vial), transferred to fresh medium every 2 days, and scored for survivors. The starting population for each genotype was 100 flies. Three replicates were tested for each genotype. Survival, l x , was estimated as N x /N 0 , where N x is the number of flies alive at the beginning of each census interval, and N 0 is the initial cohort size. Significant differences in survival between pairs of cohorts were tested using the log-rank test.

RESULTS
Drosophila Jafrac1 Is an Ortholog of Human PrxII-Based on amino acid sequence similarity, we identified a Drosophila peroxiredoxin II gene (Jafrac1, CG1633, Dpx-4783) that belongs to the 2-Cys peroxiredoxin subfamily. The amino acid sequence of the Jafrac1 protein shows significant homology to hPrxII, including conserved cysteine motifs (Fig. 1A), and the overall amino acid similarity of 83% between Jafrac1 and hPrxII (157/ 188). There are six Drosophila homologs of human peroxiredoxins that share at least 60% amino acid identity. Among these, the molecular characters of Jafrac1 are most similar with hPrxII, including protein size, the number of conserved cysteine residues, and subcellular localization (supplemental Table  2). Jafrac1 mRNA is expressed throughout development, and in the third instar larvae, Jafrac1 mRNA is abundant in brain, imaginal discs, and Malpighian tubules (supplemental Fig. S1).
Neuronal Expression of Jafrac1 or hPrxII ROS-induced Lethality-The expression of Jafrac1 was induced in wildtype flies treated with 20 mM paraquat for 24 h (Fig. 1B). To test the effect of Jafrac1 overexpression on oxidative stress, we generated transgenic fly lines carrying the Jafrac1 or hPrxII genes under the control of the UAS promoter. Reverse transcription-PCR (Jafrac1) and Western blot analysis (hPrxII) confirmed that both lines exhibited Gal4dependent up-regulation of Jafrac1 transcripts and hPrxII proteins, respectively (supplemental Fig. S2). Ubiquitous expression of Jafrac1 using Actin5C-Gal4 demonstrated that Jafrac1 expression reduced oxidative stress-induced lethality (Fig.  1C). We next tested whether neuronal overexpression of Jafrac1 is sufficient to confer resistance to paraquat treatment. Using elav-Gal4, which drives expression in all postmitotic neurons (19), and Cha-Gal4, which drives expression primarily in cholinergic neurons (20), we found that elavϾJafrac1 and ChaϾJafrac1 adult flies exhibited significantly reduced paraquat-induced lethality, comparable to Actin5CϾJafrac1 flies (Fig. 1C). Interestingly, Jafrac1 overexpression in glial cells using repo-Gal4 was not protective. In contrast, neuronal knockdown of Jafrac1, but not knockdown of Jafrac1 in glial cells, sensitized flies to paraquat-induced lethality (data not shown). These results suggest that Jafrac1 plays a protective role against oxidative stress in neurons.
To test the role of Jafrac1 in ROS metabolism, we measured ROS levels using a 2,7-dichlorofluorescein-DA dependent fluorescence  ATP levels in flies exposed to 20 mM paraquat. Flies with neuronal overexpression of Jafrac1 (elavϾJafrac1) or hPrxII (elavϾ hPrxII) exhibit higher ATP levels compared with wild-type controls (WT), whereas ATP levels were reduced in the Jafrac1 G1104 mutant and neuronal Jafrac1 inhibition (elavϾJafrac1-RNAi) flies. Data are expressed as mean Ϯ S.E. from four independent experiments (n ϭ 10 per each experiment; *, p Ͻ 0.05; **, p Ͻ 0.001; Student's t test). Copy numbers of mitochondrial marker genes (Co I: cytochrome c oxidase subunit I; Co III: cytochrome c oxidase subunit III; and Cyt C: cytochrome c) in the flies after paraquat treatment. The copy number reduction was significantly aggravated in the Jafrac1 G1104 mutant and neuronal Jafrac1 inhibition (elavϾJafrac1-RNAi) flies and restored by neuronal overexpression of Jafrac1 or hPrxII (elavϾJafrac1 or elavϾhPxII). Mean Ϯ S.E. from three independent experiments (n ϭ 5 per each experiment; *, p Ͻ 0.05; **, p Ͻ 0.001; Student's t test).
assay. When wild-type flies were exposed to 20 mM paraquat, intercellular ROS levels were significantly increased in a time-dependent manner: 3-fold and 3.5-fold after 6-and 12-h exposures, respectively. The intracellular ROS levels were dramatically reduced by neuronal overexpression of Jafrac1 or hPrxII (elavϾJafrac1 or elavϾhPrxII) (Fig. 1D). Conversely, flies with neuronal knockdown of Jafrac1 (elavϾJafrac1-Ri) showed increased ROS levels compared with wild-type control flies (Fig. 1D). These results indicate that both Jafrac1 and hPrxII can modulate intracellular ROS levels in Drosophila.
Neuronal Expression of Jafrac1 or hPrxII ROS-induced Mitochondrial Dysfunction-To test whether Jafrac1 has a protective effect on mitochondrial function under oxidative stress, we measured ATP levels in flies with neuronal overexpression of Jafrac1 or hPrxII after paraquat treatment. In the control flies, 20 mM paraquat treatment resulted in a 50% reduction in the ATP level. Neuronal overexpression of Jafrac1 or hPrxII (elavϾJafrac1, elavϾhPrxII) markedly restored ATP production, whereas the reduction of ATP levels after paraquat treatment was enhanced in loss-of-function Jafrac1 mutants (Jafrac1 G1104 ) and flies with neuronal knockdown of Jafrac1 (elavϾJafrac1-Ri) ( Fig. 2A).
We next quantified mitochondrial abundance by measuring the mitochondrial DNA (mtDNA) copy number. Treatment with 20 mM paraquat caused a marked reduction in the levels of mtDNA, and this reduction in mtDNA levels induced by paraquat treatment was restored by Jafrac1 or hPrxII overexpression in neurons (Fig. 2B). The reduction in mtDNA levels after paraquat treatment was enhanced in loss-of-function Jafrac1 mutants (Jafrac1 G1104 ) and flies with neuronal knockdown of Jafrac1 (elavϾJafrac1-Ri). These data indicate that neuronal expression of Jafrac1 or hPrxII restores oxidative stress-induced mitochondrial function.
Neuronal Expression of Jafrac1 Inhibits Oxidative Stress-induced JNK Activation-Oxidative stress can activate the stressresponsive JNK/FOXO signaling pathway (3,4,6,7). To examine JNK and FOXO activation by oxidative stress in cholinergic neurons, we marked cholinergic neurons with green fluorescent protein (ChaϾGFP) and immunostained with the phos-pho-JNK (pJNK) and FOXO antibody to detect the active form of JNK and FOXO, respectively, in the neuromuscular junctions of third instar larvae. Activated JNK was observed in the cholinergic neurons treated with 20 mM paraquat (Fig.  3B, yellow), but not in the controls (Fig. 3A). Similarly, FOXO protein was localized in the nuclei of cholinergic neurons treated with 20 mM paraquat (Fig. 3D, arrow), but not in the controls (Fig. 3C).
Jafrac1 expression (ChaϾ Jafrac1) reduced the number of pJNK-positive neurons (Fig. 3E). Western blot analysis following treatment with 20 mM paraquat revealed that the treatment induced activation of JNK, but overexpression of Jafrac1 in neurons (elavϾJafrac1 and ChaϾJafrac1) suppressed JNK activation (Fig. 3F). These results suggest that oxidative stress-induced JNK activation can be suppressed by Jafrac1 in neurons.
Peroxiredoxin (Prx) enzymes modulate oxidative stress via its evolutionary conserved cysteine (Cys) residues (21). Under oxidative stress condition, cysteine sulfinic acid of Prx is oxidized. To test whether a Drosophila Prx homolog, Jafrac1, is also oxidized under oxidative stress condition, we performed Western blot analysis with oxidized Prx specific antibody, anti-Prx-SO3. In wildtype flies, the level of oxidized Prx homolog was increased after 20 mM paraquat for 24 h, indicating that the oxidation of Cys residue is conserved in Jafrac1. Increase of oxidation of Jafrac1 was enhanced in a hemizygous mutation of JNKK (Hep 1 /Y) or transallelic mutation of dFOXO 21/25 compared the level of wild-type control flies after paraquat treatment, presumably by increased ROS levels in these mutant backgrounds (Fig. 3G). These results further support that Jafrac1 modulates oxidative stress in Drosophila.   (Fig. 4A).
Next, we tested the effect of genetic manipulation of FOXO on Jafrac1 gene expression. To induce FOXO expression in neurons in the adult stage, we used the RU486-inducible elav-Gal4 driver (elavGS-Gal4). Neuronal overexpression of wild-type FOXO (elavGSϾdFOXO ϩRU) or the insulin-insensitive nuclear form of FOXO (elavGSϾdFOXO.TM ϩRU) increased the expression level of Jafrac1 by Ͼ2-fold compared with the controls, whereas the expression of Jafrac1 was reduced in a FOXO mutant (Foxo 21/25 ) (Fig. 4B). In addition, the expression of Jafrac1 was over 3-fold increased in wild-type control flies treated with 20 mM paraquat for 24 h, but this increase was significantly reduced in a hemizygous mutation of JNKK (Hep 1 /Y) or transallelic mutation of dFOXO 21/25 (Fig. 4C). These results indicate that JNK/FOXO signaling regulates Jafac1 expression in neurons.

Neuronal Expression of Jafrac1 or hPrxII Extends Life Span in Flies-
Because it is well established that activation of JNK/FOXO signaling increases life span (2,4,7), we examined the role of neuronal Jafrac1, a target gene of JNK/FOXO signaling, in the control of the fly life span. Neuronal overexpression of Jafrac1 or hPrxII in neurons (elavϾJafrac1 or elavϾhPrxII) significantly increased life span, while neuronal knockdown of Jafrac1 (elavϾJafrac1-Ri), as well as the loss-of-function mutation (Jafrac1 G1104 ), caused a reduction in life span (Fig. 5A). To test whether neuronal expression of Jafrac1 in the adult stage is sufficient to extend life span, we used the RU486-inducible elav-Gal4 driver (elavGS-Gal4) to express Jafrac1 in adult neurons (22). It has been reported that RU486 feeding does not affect life span of flies (2). Expression of Jafrac1 in adult neurons extended life span by 26% in females (Fig. 5B) and 29% in males (Fig. 5C), compared with the control flies.

DISCUSSION
Multiple lines of evidence point to the activation of the JNK/ FOXO pathway as a common cellular response to oxidative damage across animal phyla (4,6,7,13,23). In Drosophila, JNK confers tolerance to oxidative stress and extends life span by inducing a protective gene expression program. Increased JNK activity in neurons is sufficient to promote stress tolerance and extend life span in flies (4,7). However, whether this effect is due to the specific protection of neurons against oxidative damage, or whether JNK activation in neurons may induce a humoral response that regulates longevity systemically, is unclear (1,5).
In this study, we demonstrated that JNK/FOXO signaling is required for the expression of Jafrac1 in brains under both normal and oxidative stressed conditions (Fig. 4). There are two putative FOXO consensus binding sites (RWWAACA) in the promoter region of Jafrac1 (data not shown), suggesting that the transcription factor FOXO may bind to the Jafrac1 promoter and directly activate Jafrac1 transcription. We also demonstrated that neuronal knockdown of Jafrac1 enhances, and neuronal overexpression of Jafrac1 reduces, ROS-induced lethality. Furthermore, the neuronal knockdown of Jafrac1 shortened, while overexpression of Jafrac1 extended, the life span of the flies. These results support the hypothesis that, in Drosophila, the JNK/FOXO pathway protects neurons from oxidative stress and extends life span by induction of antioxidant genes, including Jafrac1 in neurons (Fig. 5D).
Peroxiredoxins (Prxs) are identified by their ability to neutralize cellular hydroperoxides in mammals (24). A family of five Prx genes has been identified and characterized in D. melanogaster (25). All Drosophila Prxs have peroxidase activities, and their expressions are induced by oxidative stress. Prx overexpression enhances resistance to oxidative stress by hydrogen peroxide and paraquat in cultured Drosophila cells (25,26). In Drosophila, overexpression of Jafrac1 has been shown to counteract the enhanced susceptibility of immune-regulated catalase knockdown flies to natural infections (14). Moreover, mitochondrial peroxiredoxin (Dpx-5037, mTPx) has been reported to restore wild-type life span in a Drosophila model for Friedreich's ataxia (27).
We have demonstrated that neuronal expression of Jafrac1 and hPrxII significantly reduces the ROS level and restores mitochondrial function in paraquat-treated flies (Figs. 1D and 2). Several studies in Drosophila show that expression of specific mitochondrial proteins can increase resistance to oxidative stress as well as extend life span (28 -30), suggesting that mitochondrial function plays an important role in determining life span. Collectively, Jafrac1 or hPrxII may extend life span by acting as a guardian for neuronal mitochondria under age-associated oxidative stress conditions. Furthermore, because mitochondrial dysfunction is associated with many neurodegenerative diseases (8), induction of Jafrac1/PrxII in neurons may also be protective against age-associated neurodegenerative diseases.