Effects of Oxidative Stress on Behavior, Physiology, and the Redox Thiol Proteome of Caenorhabditis elegans

Strains and culture conditions

The Bristol strain N2 (wild-type), and the original isolate VC289 prdx-2(gk169)II were provided by the Caenorhabditis Genetics Center. The prdx-2 strain was backcrossed three times to strain N2. Strains were maintained and cultured under standard conditions using E. coli OP50 as a food source (32). For large-scale C. elegans cultivation, see Supplementary Data (available online at www.liebertonline.com/ars).

Oxidative stress treatment

Synchronized worms at day 0 of adulthood were collected by centrifugation in M9 medium and washed. One hundred microliters of worms (∼30,000 worms) was incubated in 2 ml M9 with the indicated concentration of H₂O₂ at room temperature in a rotating roller drum for 30 min. After treatment, the worms were collected by centrifugation and the oxidant was washed away with M9 medium. The animals were seeded on fresh nematode growth medium plates and immediately singled. Singling and scoring was performed blinded for the day 0 evaluation.

Lifespan, movement, and brood size analysis

All behavioral studies were conducted, scored, and statistically analyzed as previously described (15). Details are found in Supplementary Data.

ATP measurements

One hundred microliters of synchronized young adults was treated with the indicated concentrations of H₂O₂ in 1.8 ml of M9 media for 30 min. After the treatment, worms were washed with M9 medium, shock-frozen in liquid N2, thawed, transferred into boiling 4 M Guanidinium-HCl, and boiled for 15 min. After centrifugation (30 min, 4°C, 13,200 g), the protein concentration of the supernatant was determined using the BioRad assay (Bio-Rad Laboratories). Samples were diluted 1:200 into ATP buffer (40 mM HEPES-KOH and 4 mM magnesium sulfate, pH 7.8) and mixed with the same volume of assay buffer containing 200 nM luciferase (Roche), 140 μM luciferin (Biotium Inc.), 0.1 mg/ml BSA, 200 mM KH₂PO₄, 50 mM glycylglycine, and 0.4 mM ethylenediaminetetraacetic acid, pH 7.8. Chemiluminescence was measured using a FLUOstar Omega (BMG Labtech). ATP concentrations were determined according to an ATP standard curve and normalized over protein content.

Sample preparation for OxICAT

To prepare protein extracts for the redox proteomic analysis, 50–100 μl stress-treated worms were harvested onto 10% (w/v) trichloracetic acid (TCA) and shock-frozen in liquid nitrogen to induce breaking and lysis; after thawing, they were homogenized for 2 min with Power Gen 125 (Fisher Scientific). The TCA-treated protein extract from C. elegans was centrifuged (13,000 g, 4°C, 30 min) and the resulting pellet was washed once with 500 μl 10% (w/v) TCA and once with 200 μl 5% (w/v) TCA. The supernatant was completely removed and the sample was transferred to the anaerobic chamber for the first OxICAT alkylation step. All subsequent steps of the OxICAT labeling were performed according to the published protocol (30). Mass spectrometry (MS) was performed at the Michigan Proteome Consortium. Data were analyzed as described (30) and details can be found in Supplementary Data.

Peroxide treatment leads to reversible behavioral defects in C. elegans

The cellular accumulation of ROS has been attributed to many different physiological and pathological alterations, yet the precise effects that sublethal concentrations of oxidants such as H₂O₂ exert on the physiology of multicellular organisms have not been well defined. We therefore decided to investigate the effects of specific oxidants on the behavior and redox proteome of C. elegans. We chose C. elegans because it is a very well-characterized organism that can be easily synchronized, and it has a multitude of age-related behavioral traits, such as mobility, brood size, and pharyngeal pumping, that can be quantitatively assessed along with size and lifespan. In previous oxidative stress resistance tests in C. elegans, the minimal lethal oxidant concentrations have been determined, but investigations of physiological effects and/or potential recovery have remained relatively unexplored (28).

We initiated our studies by exposing synchronized C. elegans on day 0 of adulthood to different H₂O₂ concentrations for 30 min. We then removed the oxidant and monitored immediate and long-term behavior. Our goal was to find oxidant concentrations that affect the majority of the worm population without killing them; this information would be used for subsequent quantitative redox proteomic analysis to reveal oxidative stress-specific target proteins in C. elegans. We found that short-term exposure of C. elegans to 6 or 10 mM H₂O₂ fulfills these criteria. These treatments are nonlethal (survival rate >95% compared to control group) and do not affect the mean lifespan of the organism (Fig. 1A). At the same time, they cause very distinct behavioral defects in the majority of animals. The most obvious and immediate effects of H₂O₂ treatment that we observed were mobility defects, which were expressed as decreased thrashing in H₂O₂-supplemented liquid media (data not shown). This movement defect remained after washing the worms and seeding them on oxidant-free food plates. We found that highly impaired movement, ranging from no movement to slow and uncoordinated motility, occurred in ∼20% of the worms treated with 6 mM H₂O₂ and in ∼60% of worms treated with 10 mM H₂O₂ (Fig. 1B). Importantly, within <48 h after treatment, the vast majority of worms regained fast, sinusoidal body movement, which they maintained until the typical age-related motility decline set in (Fig. 1B). This result suggests that young worms have effective antioxidant systems that reverse the effects of exogenous oxidative stress and promote their recovery.

Another consequence of exposing C. elegans to H₂O₂ stress was a significant reduction in the total progeny production. This decrease was not due to differences in the length of the fertile period but instead due to a severe decline in egg production in the first 3 days after the oxidative stress treatment (Fig. 1C). As observed with the mobility defects, the animals recovered from the H₂O₂ treatment and showed a progeny production very similar to the progeny production of the control group after 72 h (Fig. 1C). In addition, we found that peroxide-treated animals showed reduced pharyngeal pumping (Fig. 1D), a significant decrease in body length (Fig. 1E), and a significantly reduced growth rate (Fig. 1F).

Peroxide stress has been shown to lead to a massive decline in intracellular ATP levels due to the oxidative inactivation of key enzymes involved in energy-generating pathways, that is, glyceraldehyde-3-phosphate dehydrogenase (GAPDH) and ATP synthases (16). Because decreased ATP levels might explain at least some of the observed behavioral changes in C. elegans, we measured intracellular ATP concentrations after 30 min of exposure to peroxide stress. As shown in Figure 1G, we found that peroxide-treated worms do indeed suffer from a significant drop in intracellular ATP levels. These results suggest that a change in the cellular energy charge might be at least in part responsible for some of the behavioral changes observed in peroxide-treated worms.

Peroxiredoxin 2 promotes recovery from peroxide stress-induced motility and egg-laying defects

One important player in the H₂O₂ detoxification system of C. elegans is the 2-Cys peroxiredoxin 2 (PRDX-2) (35), a highly abundant protein that constitutes ∼0.5% of the total C. elegans proteome (Supplementary Fig. S1; see Supplementary Data available online at www.liebertonline.com/ars). Deletion of prdx-2 has been shown to increase the sensitivity of C. elegans toward exogenous peroxide treatment and to cause a significant decrease in the lifespan of these worms, especially at lower cultivation temperatures (35) (Fig. 2A, B). Moreover, the phenotypes of prdx-2 mutant worms, especially after the larval stages (18, 35), were highly reminiscent of the behavioral changes that we observed when we treated wild-type strains with H₂O₂ (Fig. 2C–F). Adult prdx-2 deletion strains show a reduced brood size, and remain significantly smaller during their adult lifespan. These results not only indicate that prdx-2 worms suffer from endogenous oxidative stress but also suggest that short-term peroxide treatment can be used to mimic reversible peroxide stress in vivo.

To assess the effects of exogenous oxidative stress in worms lacking PRDX-2, we exposed synchronized prdx-2 mutant worms on day 0 of adulthood to our short-term peroxide stress treatment. Although we did not detect any change in the mean adult lifespan, we did notice a reproducible increase in the initial mortality rate and a decreased minimal lifespan (Fig. 3A). This result suggests that lack of PRDX-2 might lead to a prolonged period of peroxide stress in worms, which leads to early death in a susceptible subfraction of the animals. We also analyzed movement and progeny production of prdx-2 mutants upon peroxide treatment. Although exogenous peroxide stress treatment caused only a slight but reproducibly more severe immediate motility defect in the prdx-2 mutant strain as compared to wild-type C. elegans (compare Fig. 1B with Fig. 3B), recovery in the 10 mM H₂O₂-treated group was severely impaired in the mutant worms, with many worms never regaining their original mobility (Fig. 3B). These results suggest that lack of PRDX-2 lengthens the exposure of cellular macromolecules to the damaging effects of peroxide stress. Very similar results were obtained when we analyzed the effects of peroxide stress on the progeny production of prdx-2 worms; brood size dropped to <20% of the values obtained in the untreated prdx-2 mutants and never recovered (Fig. 3C). In contrast to mobility, progeny production, and pharyngeal pumping of prdx-2 worms, which were all additionally affected by exogenous oxidative stress treatment (Fig. 3B–D), neither size, growth rate, nor intracellular ATP levels of prdx-2 mutant animals were further reduced by exogenous peroxide treatment (Fig. 3E–G). These results suggest that some cellular processes might already be maximally affected by the oxidative stress that is caused by the absence of PRDX-2.

Quantitative redox proteomics identifies redox-sensitive C. elegans proteins

Peroxide stress conditions affect physiological processes, presumably through the oxidative modification of specific cellular targets. We therefore decided to use OxICAT to reveal particularly stress-sensitive proteins whose redox-regulated function might be responsible for some of the observed behavioral and physiological effects of peroxide stress. The OxICAT technology is based on the differential trapping of reduced and oxidized cysteines with two versions of the isotope coded affinity tag (ICAT): an isotopically light ¹²C-form (i.e., light ICAT) and a 9 Da heavier, isotopically heavy ¹³C-form (i.e., heavy ICAT) (30) (for details, see Fig. 4). All in vivo-reduced cysteines are alkylated with the light ICAT reagent, whereas all in vivo-oxidized cysteines are, upon their reduction, labeled with the heavy ICAT reagent. After tryptic digest and enrichment using affinity chromatography, the ICAT-labeled peptides are separated by liquid chromatography and identified by MS. Because light- and heavy-labeled peptides are chemically identical, the relative ion intensities of the two peaks represents the relative abundance of reduced and oxidized protein species in cells, making this method independent of absolute protein amounts. This makes the OxICAT technique uniquely suited to quantitatively describe changes in the thiol status of hundreds of individual proteins in a single experiment.

To investigate which C. elegans proteins are sensitive to oxidative modification, we exposed a synchronized population of ∼100,000 young adult wild-type C. elegans N2 or prdx-2 deletion mutants to our previously established oxidative stress regimen and quantified the thiol oxidation status of proteins before and after the stress treatment. We found a considerable number of ICAT-labeled peptides whose masses changed by 9 or 18 Da, indicating that they contain either one or two cysteine residues whose redox status changed upon peroxide treatment. We only considered C. elegans peptides to be redox sensitive if their stress-induced changes in thiol oxidation status were reproducibly >1.5-fold and the oxidized population exceeded 20%, the thresholds we previously used to identify redox-sensitive proteins in E. coli (30). Both the extent of oxidation as well as the total number of identified oxidation-sensitive proteins could be underestimates because only ∼60% of the treated worms showed behavioral changes upon stress treatment (Fig. 1B). The remaining 40% of worms revealed no obvious defects, suggesting that their redox proteome might not be affected. We then used MS/MS analysis to identify the peptides and the oxidation-sensitive cysteines. We found a total of 40 different proteins whose thiol oxidation status changed at least 1.5-fold and up to 9-fold upon peroxide stress treatment in wild-type and/or prdx-2 mutant strains (Table 1). Of those, 22 proteins were found to be oxidation sensitive in both strain backgrounds, confirming their general oxidation sensitivity. Noteworthy, many of these proteins showed very similar absolute oxidation levels in the two strains, both before and after stress treatment. We identified only two proteins whose cysteines were found to be approximately threefold more oxidized in nonstress-treated prdx-2 worms compared to wild-type worms: the highly conserved Cys307 of the Hsp70 homolog heat shock protein 1 (HSP-1) and Cys204 of the small subunit ISW-1 of the chromatin remodeling factor (Table 1). This increase in thiol oxidation might reflect subcellular, localized changes in the peroxide levels of prdx-2 mutant worms.

All other redox-sensitive peptides that are listed in Table 1 were detected either only in wild-type or prdx-2 deletion strains. That we did not detect these peptides in both strain backgrounds might be due to differences in protein expression levels or due to limitations of the liquid chromatography/MS analysis. The fact, however, that we confirmed so many of the peptides to be sensitive to peroxide stress in both strain backgrounds makes us confident that the listed proteins are indeed peroxide sensitive.

Protein translation is a major target of oxidative modifications

Proteins involved in translation appear to be among those most heavily targeted by H₂O₂ stress (Table 1), a result that agrees well with previous observations that protein translation is a peroxide-sensitive process in vivo (40). We identified peroxide-sensitive cysteines in six small (S3A, S14, S17, S19, S21, and S28) and three large (L4, L7, and L22) ribosomal subunits, in the polyadenylate-binding protein (PAB)-1, and in two elongation factors, EFT-2 and EF-1α (EFT-4). One of the most peroxide-sensitive cysteines that we identified is the lone Cys182 of the large ribosomal subunit protein L7; the oxidation status of this cysteine changed from 6% to over 50% upon peroxide treatment (Fig. 4, inset). Although this cysteine is highly conserved in eukaryotes, no studies have addressed its role in L7 function. It is conceivable that formation of sulfenic acid or of an intermolecular disulfide with either the small tripeptide glutathione (i.e., S-glutathionylation) or another protein leads to conformational and functional changes in this important component of the eukaryotic translation machinery.

The small subunit S21 (RPS-21) is another ribosomal protein that we found to contain a highly peroxide-sensitive cysteine (i.e., Cys57) in both wild-type and prdx-2 mutant strains (Fig. 5A). Its activity is required for embryonic and germline development and for the overall health of the animal (www.wormbase.org, release WS204, 29 July 2009). We identified two Cys57-containing RPS-21 peptides in wild-type worms that only differ in their respective cleavage site. Cys57 was oxidized to either 39% or 42% depending on the respective peptide (Table 1 and Fig. 5A), a result that nicely illustrates the accuracy of our OxICAT technique. Similarly, we identified two Cys745-containing peptides of the EFT-2 whose oxidation status changed from either 7% to 31% or from 6% to 27% upon peroxide treatment (Fig. 5A). This result strongly suggests that this cysteine, which is highly conserved and located in one of the three highly mobile transfer RNA mimicking C-terminal domains of EFT-2, is peroxide sensitive. This notion is consistent with previous studies, which revealed that EFT-2 in Jurkat cells contains peroxide stress-sensitive cysteine(s) (3) and that EFT-2 in macrophages and smooth muscle cells undergoes S-nitrosylation reactions (13). While the peroxide-sensitive cysteine(s) in EFT-2 have not yet been located, Cys567 has been identified as a target of S-nitrosylation (13). We found that a significant proportion of EFT-2 is reproducibly and significantly oxidized at the equally conserved Cys745. Additional studies are necessary to determine if this oxidative cysteine modifications plays indeed a regulatory role in EFT-2 function. Finally, we identified Cys149, one of three conserved cysteines in the PAB-1, to be sensitive to peroxide-mediated oxidation. Expression of PAB-1 has very recently been shown to suppress peroxide-induced cell death in NIH/3T3 cells, suggesting that it has an oxidative-stress protective function in vivo (33).

Peroxide treatment targets proteins involved in protein homeostasis

Oxidative modification and inactivation of proteins involved in protein translation is considered to be an effective strategy to rapidly downregulate new protein synthesis under stress conditions that cause harm to newly synthesized and existing proteins. In contrast, molecular chaperones, such as bacterial Hsp33 (20) or eukaryotic PRDX (21), which are involved in maintaining protein homeostasis, use oxidative thiol modifications as a mechanism to rapidly activate their chaperone function. This effectively prevents the aggregation of stress-unfolded proteins and increases oxidative stress resistance (27). Here, we identified several additional chaperones whose function might be redox regulated. One of these chaperones is the Hsp70 homolog HSP-1, a highly conserved family of ATP-dependent chaperones whose members play important roles in de novo protein folding, stress-induced unfolding, protein transport, and protein degradation [reviewed in Liberek et al. (31)]. We found that Cys307, one of the three highly conserved cysteines in eukaryotic Hsp70s, becomes significantly oxidized in both wild-type and prdx-2 deletion strains (Table 1 and Fig. 5B). S-glutathionylation of at least one Hsp70 homolog has been previously shown to increase its in vitro chaperone function, suggesting that oxidative cysteine modification might serve as post-translational regulation of Hsp70's chaperone activity (14). It remains now to be determined whether oxidation of the highly conserved Cys307 is involved in this regulation.

In addition to HSP-1, we also identified one conserved cysteine in the delta subunit of the TCP/TriC chaperonin complex (i.e., CCT-4) to be highly sensitive to peroxide-mediated oxidation (Fig. 5B). The TCP/TriC complex supports de novo folding of actin and myosin, as well as numerous other β-sheet-rich multidomain proteins (47). The oxidation-sensitive cysteine that we identified is part of a highly conserved Cys-X₃-Cys motif, which is common for redox-sensitive and/or metal-binding proteins (24). We also identified a peroxide-sensitive cysteine in the zeta subunit CCT-6 of the TCP/TriC chaperonin complex, suggesting that several members of this complex might be redox sensitive.

In addition to identifying these potentially redox-sensitive chaperones, we also noted that peroxide stress targets a set of cysteine-containing proteins involved in protein targeting and degradation. One of these proteins is the class I ubiquitin-conjugating enzyme UBC-2, whose active-site cysteine was found to become significantly oxidized upon peroxide stress treatment (Table 1). The C. elegans ortholog UBC-2, which is encoded by the essential let-70 gene, has been implicated in the ubiquitin-mediated degradation of many short-lived proteins (50). UBC-2 belongs to the highly conserved class of E2-conjugating enzymes. All members of this class share the presence of one active-site cysteine (e.g., Cys85 in UBC-2), which catalyzes the transfer of small peptidic modifiers, such as ubiquitin or SUMO, from specific E1-activating enzymes to protein substrates via the engagement of distinct E3 ligases. These post-translational modifications modify and affect protein–protein interactions, localization, activity, and/or stability of hundreds of different proteins in eukaryotic cells [for recent review, see Ye and Rape (49)]. Our study clearly demonstrates that the redox state of the active-site cysteine of UBC-2 changes from a predominantly reduced state before oxidative stress treatment to a >30% oxidized state upon treatment (Fig. 5B). This result is in excellent agreement with previous reports on the redox-regulated activity of both ubiquitin- and SUMO-conjugating enzymes E2 (4, 19). Because oxidation of the active-site cysteine is known to cause the inactivation of the conjugating enzyme, and by extension, the complete pathway, these results suggest significant changes in the proteostasis network during oxidative stress conditions in C. elegans.

In addition to UBC-2, our study revealed at least one more redox-sensitive protein that plays a role in ubiquitin-mediated protein degradation: the highly conserved AAA-protein CDC-48 (i.e., p99, VCP). CDC-48 is an ATP-dependent molecular chaperone whose primary function appears to be the chaperoning of retro-translocated, ubiquitylated ER proteins to the proteasome [reviewed in Dreveny et al. (9)]. CDC-48 harbors several cysteine residues, of which only three are conserved between yeast CDC-48 and human VCP. One of these conserved cysteines is the highly oxidation-sensitive Cys105 that we identified in our study (Table 1). So far, it is unclear how oxidation of Cys105, which resides in the N-terminal substrate-binding domain of CDC-48, affects the functionality of the chaperone. The C-terminal ATPase activity of mammalian CDC-48 has been shown to be redox regulated (34), suggesting that CDC-48 is an intrinsically redox-sensitive protein. The last protein that we identified in this group of redox-sensitive proteins involved in proteostasis is the ER translocation channel protein SEC-61. It has been suggested that SEC-61 mediates the retro-translocation of ubiquitylated ER proteins into the cytoplasm and thereby the transfer to CDC-48 (22). These results suggest that peroxide stress conditions lead to the reversible downregulation of ubiquitin-mediated protein degradation. It remains to be tested whether this mechanism prevents the degradation of oxidatively modified proteins during sublethal oxidative stress conditions, which would eliminate the possibility of quickly regenerating their protein activity once reducing conditions have been restored.

Muscle contraction and growth rate are major targets of peroxide stress in C. elegans

Another large group of C. elegans proteins that is targeted by peroxide stress include muscle-specific proteins and enzymes involved in ATP homeostasis. This result agrees well with our earlier observations that peroxide-treated worms show a massive decline in motility (Fig. 1). We found increased cysteine oxidation in three different paralogs of the myosin class II heavy chain: Cys280 of the pharynx-specific LET-75, Cys702 of UNC-54, and the equivalent Cys708 of MYO-2 (Fig. 5C). The fact that we independently identified the same cysteine to be redox sensitive in two different myosin-isoforms (UNC-54, MYO-2) illustrates the general oxidation sensitivity of this specific cysteine residue. This is particularly significant because this cysteine is highly conserved from yeast to human myosin and is located adjacent to a second conserved cysteine, which has been proposed to be redox sensitive and involved in the ATPase activity of myosin (25). Oxidative inactivation of UNC-54, the major C. elegans myosin heavy chain required for locomotion and egg-laying, and MYO-2, which is exclusively expressed in the pharynx, might contribute to the observed peroxide-mediated defects in motility, pharyngeal pumping, and egg-laying. In addition to the three myosin class 2 paralogs, we also found MUP-2, a homolog of the invertebrate troponin T, the muscle-organizing protein DIM-1, and the intermediate filament IFC-2 (i.e., lamin) to contain at least one significantly peroxide-sensitive cysteine (Table 1 and Fig. 5C). Like myosin, troponin T is involved in muscle contraction and normal growth rate, whereas IFC-2 is required for movement, growth rate, body size, and body shape (www.wormbase.org, release WS204, 29 July 2009). It is therefore very likely that oxidative modification of one or more of these proteins contributes to the severe defects in movement, pharyngeal pumping, and/or growth rate observed in oxidatively stressed animals.

Identification of peroxide-sensitive thiols in ATPases and enzymes catalyzing transphosphorylation reactions

Like the defects in C. elegans motility, which became immediately obvious in oxidatively stressed animals, we measured a dramatic loss in intracellular ATP levels within 30 min of peroxide treatment in wild-type worms (Fig. 1G) and found decreased steady-state ATP levels in prdx-2-deficient worms (Fig. 3G). A similar decrease in cellular ATP levels was previously observed in peroxide-stress-treated bacteria and yeast and has, at least in part, been attributed to the oxidative inactivation of GAPDH and F-type ATPases (16). Although we could not detect the cysteine-containing active-site peptides of C. elegans GAPDH or F-type ATPase in our mass spectra, we found redox-sensitive cysteines in the A and B subunits of the V-type ATPase and in proteins involved in transphosphorylation reactions [e.g., nucleoside diphosphate kinase [NDPK] and creatine kinase] (Table 1 and Fig. 5C). V-type ATPases deplete cellular ATP levels by hydrolyzing ATP to acidify lysosomal and endosomal organelles, whereas NDPK transfers phosphoryl groups from ATP to other nucleoside diphosphates. Creatine kinase, on the other hand, generates ATP by transferring phosphoryl groups from the phosphocreatine pool of the muscle cells to adenosine diphosphate. Importantly, redox sensitivity has been previously reported for all three proteins, which agrees well with our observed results. It has been shown, for instance, that mammalian and plant V-type ATPases undergo reversible disulfide bond formation in response to H₂O₂ treatment, which leads to the inactivation of the proton pump (10, 45). Whereas some of the redox-sensitive cysteines in subunit A have already been identified (i.e., Cys 254 and Cys532) (10), we have now also discovered a highly conserved Cys33 in subunit B to be oxidation sensitive. Inactivation of C. elegans V-type ATPase has been recently shown to exert neuroprotective effects against necrosis (44), suggesting that oxidation-mediated inactivation of V-type ATPases might not only save already scarce ATP resources during oxidative stress, but may be a strategy to protect against necrosis.

We confirmed previous reports that the active-site Cys109 of NDPK is highly susceptible to oxidative modification (Table 1) and that the multiple activities of the mammalian NDPK homolog Nm23 in endocytosis and tumor suppression are redox regulated (29). In addition, we detected the nearby Cys117 to be highly oxidation sensitive as well (Table 1), suggesting that peroxide stress leads to intramolecular disulfide bond formation in NDPK. In the case of creatine kinase, we were unable to identify the peptide containing the active-site Cys283, which has also previously been reported to be redox sensitive (38). Instead, we detected a nearby second cysteine to be highly oxidation sensitive, implying again the possibility of an intramolecular disulfide bond. In either case, oxidative thiol modification has been demonstrated to cause the inactivation of both kinases, which might contribute to the observed changes in C. elegans' ATP homeostasis and/or serve as a protective measure to prevent further ATP depletion. Unfortunately, we did not detect the corresponding active-site cysteine-containing peptides in the mass spectra of our prdx-2 deficient worms, making it impossible for us to confirm whether oxidative modification of any of these proteins is responsible for the significantly decreased steady state ATP levels observed in prdx-2 deficient worms.

In addition to the redox-sensitive proteins that we identified to be involved in protein translation, protein homeostasis, muscle function, and ATP homeostasis, we also found redox-sensitive proteins involved in a variety of other functions, including metabolism (e.g., aldolase and aconitase), signal transduction (e.g., annexin), chromatin remodeling (e.g., ISW-1 complex), and heterochromatin formation (e.g., vigilin) (Table 1 and Fig. 5D). Again, some of these proteins have previously been shown to be redox sensitive (e.g., annexin) (6), whereas many others have not, to our knowledge, been reported to undergo reversible oxidative modifications (e.g., DIM-1, vigilin, and tubulin). However, the fact that we confirmed the oxidation-sensitive cysteines in many of the known redox-sensitive proteins makes us very confident that the newly identified proteins are redox sensitive as well. Potential functional or structural changes that occur upon the oxidative modification of these proteins will likely contribute to the behavioral and physiological changes that accompany oxidative stress in C. elegans and possibly higher eukaryotes.