Mitochondrial Reactive Oxygen Species Generated at the Complex-II Matrix or Intermembrane Space Microdomain Have Distinct Effects on Redox Signaling and Stress Sensitivity in Caenorhabditis elegans

Abstract

Aims: How mitochondrial reactive oxygen species (ROS) impact physiological function may depend on the quantity of ROS generated or removed, and the subcellular microdomain in which this occurs. However, pharmacological tools currently available to alter ROS production in vivo lack precise spatial and temporal control.

Results: We used CRISPR/Cas9 to fuse the light-sensitive ROS-generating protein, SuperNova to the C-terminus of mitochondrial complex II succinate dehydrogenase subunits B (SDHB-1::SuperNova) and C (SDHC-1::SuperNova) in Caenorhabditis elegans to localize SuperNova to the matrix-side of the inner mitochondrial membrane, and to the intermembrane space (IMS), respectively. The presence of the SuperNova protein did not impact complex II activity, mitochondrial respiration, or C. elegans development rate under dark conditions. ROS production by SuperNova protein in vitro in the form of superoxide (O₂˙⁻) was both specific and proportional to total light irradiance in the 540–590 nm spectra, and was unaffected by varying the buffer pH to resemble the mitochondrial matrix or IMS environments. We then determined using SuperNova whether stoichiometric ROS generation in the mitochondrial matrix or IMS had distinct effects on redox signaling in vivo. Phosphorylation of PMK-1 (a p38 MAPK homolog) and transcriptional activity of SKN-1 (an Nrf2 homolog) were each dependent on both the site and duration of ROS production, with matrix-generated ROS having more prominent effects. Furthermore, matrix- but not IMS-generated ROS attenuated susceptibility to simulated ischemia reperfusion injury in C. elegans.

Innovation and Conclusion: Overall, these data demonstrate that the physiological output of ROS depends on the microdomain in which it is produced. Antioxid. Redox Signal. 31, 594–607.

Introduction

Mitochondria are subcellular organelles responsible for vital cellular processes, including the synthesis of ATP, and are an important source of reactive oxygen species (ROS) generation. The primary form of ROS, the superoxide anion (O₂˙⁻), has been shown to be generated from at least 11 sites associated with the mitochondrial electron transport chain (ETC), with differing contributions from each site depending on the metabolic conditions and/or pathological state (28, 30, 56). Since these sites of O₂˙⁻ generation are located at either side of the inner mitochondrial membrane, distinct levels of O₂˙⁻ and their subsequent radical species may be expected to occur in the matrix compartment relative to the intermembrane space (IMS) under different conditions. Moreover, different landscapes of ROS removal systems exist in each compartment. For example, there are multiple isoforms of the superoxide dismutase (SOD) enzyme that are differentially distributed throughout the cytosol and mitochondrial compartments, which remove O₂˙⁻ by dismutation into the membrane-permeable hydrogen peroxide (H₂O₂). Notably, the IMS contains an abundance of cytochrome c, which has a high affinity for O₂˙⁻, but this scavenging mechanism does not result in the formation of H₂O₂ (35). This is notable because moderate levels of localized H₂O₂ accumulation (i.e., oxidative eustress) in specific microdomains can initiate redox signaling cascades through mechanisms such as thiol oxidation (5). These redox signals can lead to hormesis via stress signaling responses, yet excessive or prolonged O₂˙⁻/H₂O₂ accumulation may lead to cellular damage via irreversible oxidative modifications (i.e., oxidative distress) (44). Therefore, the precise location of O₂˙⁻/H₂O₂ generation at a submitochondrial microdomain may be a key factor in eliciting a subsequent physiological effect, as per the third principle of the redox code (22). To date, however, our understanding of the effects of localized ROS generation in defined subcellular compartments (i.e., microdomains) is limited, since precise experimental control over O₂˙⁻/H₂O₂ generation at specific microdomains in vivo has remained a significant technical challenge.

Innovation

Mitochondrial redox biology is a rapidly expanding field, but suffers from lack of precise spatial and temporal experimental control, especially in physiological contexts. Optogenetics offers a platform to overcome these barriers, and using this approach we demonstrate differing physiological responses to equal amounts of reactive oxygen species (ROS) generated on either side of the inner mitochondrial membrane at the mitochondrial complex-II microdomain. These findings show that spatial and temporal discrimination can be achieved in vivo, and this discrimination may reveal novel mechanisms of mitochondrial redox physiology.

Optogenetics involves the use of light to activate genetically encoded photosensitive proteins to modify cellular function or elicit a signaling response. Optogenetics has typically been used to alter neuronal function via light-sensitive ion channels, but more recently, photosensitive ROS-generating proteins have been developed [reviewed elsewhere; (49, 52)]. One such example is SuperNova, a monomeric ∼29 kDa fluorescent protein capable of generating oxidants when the chromophore absorbs ∼580 nm photons through a process called photosensitization (46). SuperNova is thought to primarily generate O₂˙⁻, although its precise ROS-generating capabilities are uncharacterized (49). A key feature of photosensitive proteins such as SuperNova, is that they can be targeted to specific microdomains in vivo by expressing in fusion with an endogenous protein, such as the nuclear-encoded mitochondrial ETC subunits (55).

Mitochondrial complex-II (succinate dehydrogenase [SDH]) consists of four nuclear-encoded subunits: A, B, C, and D. Subunits A and B comprise the catalytic core (flavin binding and iron–sulfur clusters, respectively) located in the mitochondrial matrix, while C and D are integral membrane proteins that span the inner membrane, thus being in contact with the IMS. Complex-II has been shown to generate ROS under both physiological and pathological conditions (39), and is also involved in oxygen sensing (54). Complex-II is a central player in the succinate-driven reverse electron transfer (RET) toward complex-I under conditions of a highly reduced ubiquinone (Q) pool, which leads to bulk O₂˙⁻ generation that is a major contributor to the etiology of ischemia reperfusion injury (8, 26). Collectively, due to its physiological relevance and its all-nuclear-encoded subunits, mitochondrial complex-II/SDH is germane to investigating the intricate effects of ROS in vivo. To achieve this, an ideal model organism for optogenetic studies is the nematode Caenorhabditis elegans, due to its transparent tissues that allow for light transmission to photosensitive proteins, amenability to genetic editing, and well-conserved mitochondrial stress responses.

Here, we demonstrate that SuperNova generates O₂˙⁻ in response to light using high-performance liquid chromatography (HPLC) separation of dihydroethidium (DHE)-oxidation products. We then use clustered regularly interspaced short palindromic repeats/Cas9 endonuclease (CRISPR/Cas9) to insert the SuperNova coding sequence at the C-termini of the B and C subunits of mitochondrial complex-II (Cx-II) of C. elegans. This novel model not only provides unique insight into complex II ROS but also into stoichiometric amounts of ROS generated from different mitochondrial compartments. We aimed to determine whether specific ROS microdomains within mitochondria elicit distinct redox signals and protection against simulated ischemia reperfusion injury in the model organism C. elegans.

Results

Purified SuperNova generates O₂˙⁻ in a light-dependent manner in vitro

We first cloned and purified SuperNova protein to assess its photosensitization properties. SuperNova migrated via sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE) at the expected molecular weight of ∼29 kDa, and immunoblotting revealed specificity with a commercially available antibody against KillerRed, the dimeric version of SuperNova (Supplementary Fig. S1A). Purified SuperNova protein had an extinction coefficient of ɛ = 31,238 M⁻¹·cm⁻¹ and exhibited fluorescence excitation/emission peaks at 584/610 nm, respectively (Supplementary Fig. S1B), as expected (46). We then evaluated the ability of SuperNova to generate O₂˙⁻ in a light-dependent manner as depicted in Figure 1A. To detect O₂˙⁻, we used DHE, a probe that yields fluorescent reaction products in the presence of oxidants. The O₂˙⁻ selective fluorescent product is 2-hydroxyethidium (2-OH-E⁺), which is separated from the nonselective oxidation product ethidium (E⁺) using HPLC (59, 60). We found that the formation of 2-OH-E⁺ by SuperNova was light dependent and specific to O₂˙⁻, since this was attenuated in the presence of SOD enzyme (Fig. 1B). Formation of 2-OH-E⁺ by SuperNova was also proportional to total light irradiance in the 540–600 nm spectra (Fig. 1C). Since pH is known to influence fluorescent protein function (25) and because the mitochondrial matrix and IMS compartments differ by ∼1 pH unit (37), we measured 2-OH-E⁺ generation by SuperNova under buffer pH conditions resembling that of the mitochondrial matrix (pH 7.8), IMS (pH 6.8), or the cytosolic (pH 7.4) environments (36). Notably, we found no impact of buffer pH on 2-OH-E⁺ formation by SuperNova protein in response to light (Fig. 1D). Together, these data demonstrate that SuperNova protein generates O₂˙⁻ in a light-dependent manner regardless of the different physiological pH levels tested, in vitro.

FIG. 1.

Purified SuperNova protein generates superoxide in response to light, in vitro. (A) Schematic depicting ∼580 nm light-dependent superoxide (O₂˙⁻) generation by purified SuperNova protein in vitro that is detected with the probe DHE, with HPLC used to separate and measure the superoxide-specific reaction product, 2-OH-E⁺ that can be prevented by SOD. (B) SuperNova (0.25 mg/mL) induces 2-OH-E⁺ formation from DHE in response to light (5 min ∼580 nm irradiation, 6.7 mW/mm²), is attenuated in the presence of SOD, occurs in a (C) time-dependent manner (y = 448x + 636; r^{2 =} 0.999; p < 0.001); and is (D) unaffected by buffer pH levels corresponding to either the mitochondrial IMS (pH ∼6.8), cytosol (pH ∼7.3), or mitochondrial matrix (pH ∼7.8), as depicted by wedge shape from left to right. Values are expressed as mean ± SEM for n = 3 independent experiments; *p < 0.05 versus all others; n.s., not significant. 2-OH-E+, 2-hydroxyethidium; DHE, dihydroethidium; HPLC, high-performance liquid chromatography; IMS, intermembrane space; SOD, superoxide dismutase. Color images are available online.

Complex-II—SuperNova fusion proteins are expressed with correct localization and topology in C. elegans mitochondria

Next, we used CRISPR/Cas9 to generate fusion proteins with endogenous nuclear-encoded mitochondrial complex-II subunits succinate dehydrogenase subunit B (SDHB) and subunit C (SDHC) in C. elegans. At the C-terminus of each of these, we inserted the SuperNova construct (SDHB-1::SuperNova and SDHC-1::SuperNova). We also generated separate strains containing the red fluorescent protein mCherry (SDHB-1::mCherry and SDHC-1::mCherry) in place of SuperNova as an additional approach to confirm appropriate expression. Next, we generated a strain expressing green fluorescent protein (GFP) fused to mitofilin (IMMT, inner mitochondrial membrane protein 1, or C. elegans ortholog of mitofilin) (IMMT-1::GFP), a 76 kDa protein that is part of the mitochondrial contact site and cristae organizing system complex, which tethers the outer/inner mitochondrial membrane, but does not localize to the interior cristae regions (21). We confirmed with confocal microscopy that the IMMT-1::GFP signal colocalized with MitoTracker Red dye in live anesthetized C. elegans (Supplementary Fig. S2A). Next, to confirm the anticipated mitochondrial expression pattern of the complex-II fusion constructs depicted in Figure 2A, we crossed the IMMT-1::GFP strain with each of the other (mCherry and SuperNova) fluorescent protein strains. We observed the expected mitochondrial expression pattern across all tissues with colocalization between IMMT-1::GFP and each of SuperNova and mCherry fusion proteins for both the SDHB-1 and SDHC-1 constructs (Fig. 2B and Supplementary Fig. S2A).

FIG. 2.

Compartmentalized localization of mitochondrial complex-II fusion proteins in Caenorhabditis elegans. (A) Predicted fusion protein construct mitochondrial complex-II localization, demonstrated via (B) confocal microscopy of C. elegans expressing mitofilin GFP fusion protein (IMMT-1::GFP) to indicate the IMS with either of SDHB-1::mCherry or SDHC-1::mCherry (scale bars = 1 μm). (C) pro-K-protection assay on isolated mitochondrial mitoplasts demonstrates correct inner mitochondrial membrane-sidedness, with the matrix facing SDHB-1::mCherry construct protected from degradation by proteinase-K (lane 5) similar to the matrix marker protein (HSP60), relative to the IMS-facing SDHC-1::mCherry (lane 2), which shows the same susceptibility to pro-K as the IMS marker (IMMT-1::GFP). Triton-X dissolves inner membranes to allow access of pro-K to the matrix compartment. Images are representative of n = 3 independent experiments; uncropped images are shown in Supplementary Figure S2C. GFP, green fluorescent protein; IMMT, inner mitochondrial membrane protein 1 (C. elegans ortholog of mitofilin); pro-K, proteinase-K; SDHB, succinate dehydrogenase subunit B; SHDC, succinate dehydrogenase subunit C. Color images are available online.

To confirm the inner mitochondrial membrane topology (i.e., sidedness) of the fusion protein construct, we made mitoplasts and performed a protease-protection assay using the serine protease, proteinase-K (pro-K). Mitoplasts are isolated mitochondria whose outer membrane is ruptured, leaving the intact inner membrane exposed. IMMT-1::GFP was used as a marker of the IMS, and was readily cleaved with exposure to pro-K as anticipated (Fig. 2C; lanes 2 and 5). Heat shock protein of 60 kDa (HSP60) was used as a marker of the mitochondrial matrix, which was protected from pro-K until the inner membrane was dissolved with detergent. Note that only partial digestion of the overall HSP60 protein pool is to be expected in the 10-min digestion time (indicated by a small downward shift in molecular weight and abundance in lanes 3 and 6), whereas the fusion proteins only require cleavage of the linking peptide sequence between the fluorescent protein and the endogenous protein to result in a dramatic decrease in molecular weight. Consistent with the predicted topology, the SDHC-1 fusion protein (mCherry) was susceptible to pro-K digestion, suggesting that it was exposed to the IMS (lane 2), whereas the SDHB-1 construct was protected from degradation (lane 5) until the inner membrane was dissolved (lane 6), indicating its location within the matrix compartment. Profile plot analyses of confocal images were also consistent with these pro-K findings, which demonstrate that both mCherry and SuperNova fused to both SDHB-1 and SDHC-1 subunits were correctly localized to the mitochondrial matrix and IMS compartments, respectively (Supplementary Fig. S2A, B). Finally, we also confirmed by immunoblot that the overall protein expression levels of SDHB-1::SuperNova and SDHC-1::SuperNova were equal in whole-worm lysates (Supplementary Fig. S2D).

Photoactivated SuperNova at Cx-II does not impair basal complex-II function or cause oxidative stress in C. elegans

To test whether the physical presence of the SuperNova fusion protein alone impacted mitochondrial bioenergetics, we isolated mitochondria and assessed succinate/rotenone-supported state-2 and adenosine diphosphate (ADP)-stimulated state-3 oxygen consumption rates under dark conditions. There were no intrinsic defects to succinate-linked respiratory function due to the presence of the fusion proteins at complex-II compared with N2 wild type (Fig. 3A). Consistent with the lack of impact on basal mitochondrial bioenergetics, the presence of the fusion proteins also had no effect on whole organism development rate compared with C. elegans wild-type strain (N2) wild-type worms grown under dark conditions (Fig. 3B). We then sought to determine whether SuperNova affects complex-II enzyme activity after photoactivation. There were no changes in succinate-supported, malonate-sensitive complex-II enzyme activity of mitochondria isolated from SDHB-1::SuperNova or SDHC-1::SuperNova animals after photoactivation compared with dark-control conditions (Fig. 3C). Thus, altered mitochondrial bioenergetics can be excluded as a possibility for any downstream physiological effects of SuperNova photoactivation. We then asked whether photoactivated SuperNova-induced O₂˙⁻ leads to oxidative distress in whole worms. For this, we used Western blotting to probe for protein carbonylation, an irreversible oxidative posttranslational protein modification (12). We found no change in protein carbonyl levels as a result of 1 h light exposure relative to dark-controls for either of the SuperNova strains or N2 worms (Fig. 3D). In addition, there was no change in reduced/oxidized glutathione ratio (GSH:GSSG) after 1 h of light exposure (Supplementary Fig. S3D).

FIG. 3.

Complex-II SuperNova fusion constructs do not impair mitochondrial function under dark or light conditions or cause cellular oxidative stress in C. elegans. (A) The presence of either complex-II SuperNova fusion protein does not impact isolated mitochondria respiratory flux normalized to a marker of mitochondrial content (citrate synthase enzyme activity) relative to N2 (wild-type) under dark conditions. Succinate+rotenone (S/Rot)-supported state-2 leak state respiratory flux was followed by addition of ADP (0.2 mM) to stimulate state-3 oxidative phosphorylation, which was inhibited by malonate (Malo); n = 3 individual mitochondrial isolation experiments. (B) Body length of individual C. elegans measured at intervals from eggs through to adulthood to estimate animal development rate is unaffected by the presence of either SuperNova fusion construct under dark conditions; n = 6 worms/strain. (C) Succinate-linked, malonate-sensitive complex-II enzyme activity normalized to citrate synthase enzyme activity is unaffected by light-activated SuperNova (10 min ∼580 nm, 1 mW/mm²); n = 4 independent mitochondrial preparations analyzed in duplicate. (D) Protein carbonylation, indicating cellular oxidative stress, was unchanged by SuperNova immediately after whole worms were exposed to 1 h ∼580 nm light (+, 1 mW/mm²) or kept in the dark (-). Bars represent mean ± SEM, with individual responses overlaid for n = 3 independent experiments; uncropped images are shown in Supplementary Figure S3C. No statistically significant differences observed. ADP, adenosine diphosphate; N2, C. elegans wild-type strain. Color images are available online.

Taken together, these data suggest that SuperNova may generate ROS at rates that do not lead to oxidative distress. Instead, single-copy expression of SuperNova may permit low levels of compartmentalized ROS generation localized to either side of the inner mitochondrial membrane in response to light—characteristics that may be suitable for redox signaling.

Photoactivated SuperNova O₂˙⁻ generation at Cx-II induces redox signaling

Having established that photoactivated SuperNova expressed at single-copy levels does not cause oxidative distress, we sought to investigate whether this induces redox signaling in vivo and whether these responses are distinct when O₂˙⁻ is generated stoichiometrically at either side of the inner mitochondrial membrane. To this end, we focused our attention on SKN-1, the C. elegans functional ortholog of mammalian Nrf2 encoded by NFE2L2, a transcription factor responsible for orchestrating the cellular antioxidant response (1). Upon activation, SKN-1 can be released from WD40 repeat protein, DDB1- and CUL4- associated factor 11 homolog (WDR-23), a cytosolic-localized protein, allowing it to translocate to the nucleus and bind to and promote transcription from regions of DNA known as antioxidant response elements common to the promotor regions of phase II detoxification genes (6). To assess this, we used a transgenic reporter strain expressing the glutathione S-transferase-4 gene promoter sequence fused to GFP (gst-4p::GFP). We first confirmed that the expression of gst-4p::GFP increased in response to exposure to a chemical oxidant (paraquat [PQ]) in wild-type background reporter worms, and that this increase was completely prevented by RNA interference (RNAi) against skn-1, which demonstrates both the efficacy of the knockdown and the specificity of the response (Supplementary Fig. S4A, B). Then, in wild-type background reporter strain worms (lacking SuperNova), there was no change in gst-4p::GFP expression levels 24 h after exposure to 10 or 30 min of light (Fig. 4A, B). On the contrary, in worms also expressing SDHB-1::SuperNova, the level of gst-4p::GFP expression was significantly greater 24 h after exposure to 10 min of light (Fig. 4A, B). In the reporter strain worms crossed with SDHC-1::SuperNova, level of gst-4p::GFP expression was also greater 24 h after the light treatment, but this required a longer duration of exposure (30 min) to achieve an equivalent response (Fig. 4A, B). These SuperNova+light-dependent increases in gst-4p::GFP expression were absent in the presence of skn-1 RNAi (Supplementary Fig. S4C).

FIG. 4.

Spatial- and temporal-dependent redox signaling responses to ROS generated by complex-II SuperNova at either the IMS or matrix. (A) L4 stage C. elegans (CL2166) expressing the SKN-1 transcriptional GFP reporter (gst-4p::GFP) alone (wild-type background) or with either of the complex-II SuperNova fusion proteins were exposed to light (∼580 nm, 1 mW/mm²) for 0, 10, or 30 min to induce O₂˙⁻ generation, then imaged 24 h later as day-1 adults. (B) Quantification of mean GFP fluorescence per worm from 5 independent experiments, each consisting of 10–15 worms per group. *p < 0.05 strain effect; ^†p < 0.05 time effect; each point represents an individual worm, bars represent mean ± SD. gst-4p, glutathione S-transferase 4 promoter region; ROS, reactive oxygen species; SKN-1, skinhead-1 (C. elegans ortholog of Nrf2). Color images are available online.

Activation of Nrf2/SKN-1 is largely dependent upon its phosphorylation by PMK-1, the C. elegans ortholog of the mammalian p38 mitogen-activated protein kinase (p38 MAPK) (18). PMK-1 belongs to a highly conserved stress response pathway that transduces signals from stimuli (including oxidants) to promote cellular adaptations (10, 18). Thus, to determine the spatial- and temporal-dependent effects of ROS generation on the activation of this pathway, we measured PMK-1 phosphorylation status in whole worm lysates from N2 wild-type and SDHB-1/C-1::SuperNova-expressing worms immediately after 10, 30, or 60 min exposure to light. There was an overall trend for a time-dependent increase in PMK-1 phosphorylation with light exposure, and this was greater in SDHB-1::SuperNova worms compared with N2 wild-type worms at the 60 min time point (p < 0.05; Fig. 5 and Supplementary Fig S5).

FIG. 5.

PMK-1 phosphorylation in response to complex-II SuperNova-induced ROS generation. (A) Plates of day-1 adult wild-type (N2) C. elegans or those expressing Cx-II SuperNova fusion proteins were kept in the dark (“0”) or exposed to light (580 nm, 1 mW/mm²) for 60, 30, or 10 min (depicted by black wedge from left to right), then immediately snap frozen. Whole worm lysates were separated via SDS-PAGE. Representative immunoblot images are shown from one experiment for phosphorylated PMK-1, SuperNova (fused to complex-II subunits, detected with anti-KillerRed antibody), and mitochondrial ATP synthase F1 subunit alpha (ATP5A). (B) Quantification of phosphorylated PMK-1 expressed relative to ATP5A from n = 3 independent experiments. Values are expressed as mean ± SEM; *p < 0.05 for SDHB-1::SuperNova versus N2. PMK-1, mitogen-activated protein kinase (C. elegans ortholog of p38 MAPK); SDS-PAGE, sodium dodecyl sulfate polyacrylamide gel electrophoresis. Color images are available online.

Taken together, these data indicate that O₂˙⁻/H₂O₂ generated from the matrix compartment are particularly suited to promoting a SKN-1-mediated antioxidant transcriptional response that may occur in part via PMK-1 phosphorylation (18). The physiological implications of upregulating these stress pathways are the potential cellular adaptations that may provide greater resistance to a subsequent stressor.

Effects of matrix- or IMS-derived Cx-II O₂˙⁻ on simulated ischemia reperfusion injury in C. elegans

Organisms, including C. elegans, initiate stress signaling pathways to protect against injury. ROS plays a role in both the protection against and the pathology of ischemia reperfusion injury, which can be modeled in C. elegans using anoxia-reoxygenation (AR) (43). Protection against AR injury can be conferred by anoxic preconditioning, which consists of a short duration of anoxia before subsequent prolonged anoxia, and C. elegans can be preconditioned against AR injury via a number of conserved mechanisms (11). We first confirmed a robust protective effect of anoxic preconditioning against subsequent AR (Fig. 6A). Although the mechanisms underlying this phenomenon are complex and yet to be fully elucidated (53), one proposed mechanism involves upregulation of antioxidant capacity to better cope with the succinate-driven O₂˙⁻ burst during RET at reperfusion (7). To assess this, we first used worms with an SDHC-1 mutant allele known as mev-1 [succinate dehydrogenase cytochrome b560 subunit (C. elegans–methyl viologen sensitive mutant), or kn1) (20), which is known to constitutively generate elevated levels of ROS and also be sensitive to hyperoxygenation (19). We found that the mev-1 mutant was endogenously protected against AR injury (Fig. 6A). To begin to better understand whether this protective effect was due to altered ETC bioenergetics or oxidants per se, we tested whether protection against simulated AR injury was afforded by short-term exposure of wild-type N2 worms to PQ. PQ generates O₂˙⁻ via redox cycling at mitochondrial complex-I (9), although recent evidence demonstrates that cytochrome P450 oxidoreductase is also a site of PQ-mediated ROS production (41). We found that PQ treatment had a similar protective effect against simulated AR injury (Fig. 6A).

FIG. 6.

Matrix-localized mitochondrial ROS is protective against AR injury in C. elegans. (A) Survival rates of C. elegans 24 h after 18.5 h anoxia were higher with 4 h anoxia preconditioning performed 24 h before AR, or exposure to the mitochondrial O₂˙⁻ generating agent paraquat (PQ, 4 mM for 4 h). An SDHC-1 mutant strain [TK22, mev-1(kn1)] known to generate elevated levels of mitochondrial O₂˙⁻ was also endogenously protected from AR-induced death. Data are expressed as mean ± SEM for n = 4 independent experiments of ≥15 worms per plate performed in duplicate; *p < 0.05 versus N2 control. (B) Wild-type (N2) C. elegans or those expressing SuperNova protein in the mitochondrial matrix compartment (SDHB-1::SuperNova) or IMS (SDHC-1::SuperNova) were kept in the dark or pretreated with light for 60 min (∼580 nm, 1 mW/mm²), 24 h later exposed to 17 h of anoxia before being scored for survival 24 h after reoxygenation. Data are expressed as mean ± SEM for n = 8 plates of ≥15 worms per plate over n = 3 independent experiments. AR, anoxia reoxygenation. Color images are available online.

These findings suggest that mitochondrial ROS generation may be involved in protection against AR injury. However, the interpretation of the kn1 mutant experiment could be confounded by compensatory responses across the life span that may be expected from altered mitochondrial bioenergetics due to mutation of SDHC-1, which is a subunit in a key ETC enzyme. This greatly limits the ability to attribute the observed effects exclusively to ROS, per se. Therefore, using our optogenetic model, we sought to determine whether O₂˙⁻ acutely generated at complex-II microdomains by photoactivated SuperNova was protective against AR injury. Indeed, we observed that with 60 min of light pretreatment, a significantly greater fraction of SDHB-1::SuperNova worms were alive 24 h after the AR insult compared with the SDHC::SuperNova worms (Fig. 6B). Interestingly, there was no protective effect of O₂˙⁻ generated in the IMS against AR, as light-treated SDHC-1::SuperNova worm survival rates were not different from those of N2 wild-type worms (Fig. 6B). Collectively, these data suggest that mitochondria-derived oxidants generated from the matrix compartment may contribute to preconditioning against AR injury.

Discussion

Here, we report the novel findings, in that (i) SuperNova protein generates O₂˙⁻ in response to light; (ii) redox signaling but not oxidative distress occurs in response to O₂˙⁻ generated using optogenetics at the complex-II microdomain; (iii) this response is more prominent when generated in the mitochondrial matrix compartment relative to the IMS, despite equal opportunity for O₂˙⁻ generation in either compartment; and (iv) O₂˙⁻ generation in the mitochondrial matrix—but not the IMS—contributes to preconditioning against AR injury in C. elegans.

The primary oxygen-derived radical molecule, the O₂˙⁻ anion, is membrane impermeable due to its negative charge. However, it is rapidly dismutated by SOD enzymes at a rate constant of ∼10⁹ M⁻¹ s⁻¹ (24), and the resulting H₂O₂ can diffuse across lipid membranes and act as a second messenger by interacting with redox-sensitive target molecules (5). In particular, oxidation of thiol groups by H₂O₂ can lead to conformational changes or post-translational modifications to cysteine and methionine residues contained within kinases, phosphatases, or transcription factors of key regulatory pathways, resulting in the initiation of signal transduction (27). While it could be argued that O₂˙⁻ could be made anywhere within the cell, be dismutated into the membrane-permeable H₂O₂ to then diffuse throughout the cell to exert its effects (34), there is a higher probability of interaction with a target in proximity to its formation (44). Moreover, numerous oxidant removal systems are present in the cell (i.e., glutathione, thioredoxin [TRX] systems), which prevent global levels of oxidants from reaching toxic levels. Therefore, the spatial aspect of ROS formation is an important determinant in the specific redox signal generated, as outlined in the redox code (22).

An interesting finding of this study is that the matrix-initiated ROS signaling was more robust, which appears counterintuitive since the IMS is physically closer to the cytosol where redox-sensitive targets such as the SKN-1 transcription factor are located. This may be due to the unique scavenging capacities of each compartment resulting from the amount and localization of different SOD isoforms. The Mn-SOD isoform was shown to be present in rat liver mitochondrial matrix at ∼3.6 × 10^–6 M, and the Cu/Zn-SOD (copper/zinc SOD) in the IMS at a similar concentration (31). Yet, the IMS also contains a relatively high concentration of cytochrome c (∼0.7 mM), which is capable of scavenging O₂˙⁻ without yielding H₂O₂ (13). Moreover, it has been shown that IMS-localized CuZnSOD activity is redox dependent and may actually have lower rates of O₂˙⁻ dismutation activity (17). In C. elegans, SOD-2/-3 has been shown to localize to mitochondrial ETC supercomplexes (45), where SOD may modulate O₂˙⁻ levels generated by the ETC. In addition to scavenging of O₂˙⁻ by SODs, H₂O₂ is removed enzymatically by various peroxidases, including the glutathione peroxidases and peroxiredoxins/TRX systems, depending on the subcellular compartment (2, 29, 42). Peroxiredoxin, such as mitochondrial-localized PRDX-3 in C. elegans, has a low K_M for H₂O₂, and readily undergoes reversible oxidation, allowing for localized H₂O₂ accumulation, thus promoting the propagation of a redox signal (40, 57). Together, the mitochondrial matrix appears to be a more favorable environment for O₂˙⁻ dismutation and localized H₂O₂ accumulation compared with the IMS for the purposes of redox signaling. Indeed, our data show that the complex-II matrix microenvironment can make a relevant ROS signal, which seems reasonable when considering that 8 of the 11 endogenous O₂˙⁻ generation sites in the mitochondrial ETC release O₂˙⁻ to the matrix but only three to the IMS compartment (56).

SKN-1 transcriptional activity is regulated by various signals that modulate its expression levels, degradation, and subcellular localization (3). PMK-1 phosphorylation leads to nuclear localization of SKN-1, and this has been shown to occur in response to oxidants (18). ROS-mediated PMK-1 phosphorylation may occur via NSY-1, the C. elegans ortholog of ASK1 (apoptosis signal-regulating kinase 1), a MAP kinase upstream of PMK-1. In mice, ASK1 is negatively regulated under basal conditions by being bound to TRX, a key H₂O₂ scavenger protein (48). When critical cysteine residues in TRX are oxidized by H₂O₂, ASK1 can be released allowing its autophosphorylation, leading to downstream phosphorylation of p38 MAPK (15). Phosphorylation of SKN-1 by PMK-1 may prevent it from binding to the WD40 repeat protein WDR-23 in the nucleus, which normally marks it for proteasomal degradation, allowing it to bind to DNA instead (6). Also, SKN-1 has been shown to be bound to PGAM-5 (phosphoglycerate mutase family member 5), an outer mitochondrial membrane scaffold protein (32) whose activity may conceivably impact SKN-1 responses to mitochondria-derived ROS. However, it was recently shown that PMK-1 is sufficient but not necessary for SKN-1 activation in response to oxidative stress (58). Therefore, precisely how PMK-1 phosphorylation is involved in mediating the responses observed in this study requires further investigation.

Previous studies have used KillerRed (from which monomeric SuperNova was derived) in different cell types and subcellular compartments at high expression levels for the purposes of generating a large burst of localized ROS to cause oxidative distress, neuronal ablation, or cell death (16, 50). Expression using plasmid-based systems can result in multicopy and mosaic expression, thereby preventing an equal comparison in different compartments. CRISPR-based expression systems can overcome mosaic and overexpression to allow for stoichiometric comparisons between microdomains. While optogenetics is widely used in the field of neurobiology, these approaches are starting to be applied to ROS and mitochondrial biology (47, 55). Herein, we present an optogenetic approach to generate subtle, physiologically relevant levels of ROS (as opposed to oxidative distress) to study complex redox signaling networks in a living model organism.

It was interesting that the photoactivated SuperNova expressing worms were not as robustly protected from AR as the complex-II mutant (kn1) or PQ-treated worms. We speculate that the small effect for altered resistance to AR injury in photoactivated SDHB-1 relative to SDHC-1::SuperNova worms may be due to the short duration of light treatment, when considered in comparison with the life-long compensatory responses that likely occurs in mutant strains. Similarly, the greater protective effects of PQ treatment may be due to its longer treatment duration (4 h) and because residual PQ likely continued to generate O₂˙⁻ in the animals even after they were moved to regular plates, highlighting the lack of temporal control over ROS generation using a pharmacological approach. Among the numerous potential molecular mechanisms that regulate sensitivity to oxygen deprivation in C. elegans (38), it is possible that the greater activation of PMK-1 that we observed in photoactivated SDHB-1::SuperNova worms also played a role in protecting against AR, since it was previously shown in C. elegans that the NSY-1-dual specificity motigen-activated protein kinase-1-PMK-1 pathway modulates sensitivity to anoxia (14). Future studies are required to better understand the precise mechanism(s) that underlie these protective responses.

In summary, this study provides novel in vivo evidence, in the absence of confounding bioenergetic factors, that temporally and spatially distinct patterns of redox signaling occur depending on the mitochondrial site that the O₂˙⁻ is generated. In particular, we observed that a lower dose of matrix-generated O₂˙⁻ relative to IMS-generated O₂˙⁻ is sufficient to elicit redox signaling and stress resistance responses. These data provide important information toward understanding fundamental aspects of mitochondrial biology. Given that these subcellular organelles are responsible for vital cellular processes and physiological function, our novel findings may inform future studies that ultimately lead to improved therapeutic strategies for diseases involving mitochondrial oxidative stress.

Materials and Methods

Protein cloning and purification

SuperNova was purified essentially as described by Takemoto et al. (46). In brief, SuperNova (courtesy of Dr. Takeharu Nagai, Osaka University, Japan; Addgene plasmid No. 53234) was transformed into JM109(DE3) XJ autolysis (Zymo T3031) and grown at 18°C with shaking (80 rpm) for ∼65 h without induction. At optical density (OD) 0.8, the culture was cooled to ∼15°C and induced with 400 μM isopropyl β-D-1-thiogalactopyranoside (IPTG) and shaken at 200 rpm for 16 h at room temperature. Cultures were centrifuged at 3200 g for 10 min, washed with phosphate-buffered saline (PBS), and flash frozen. Cell lysate was incubated with nickel-charged affinity resin (Ni-NTA) agarose beads (Qiagen) for 45 min at 4°C. Bead–protein conjugate was washed three times with PBS (pH 7.4), followed by two washes in TN Buffer (TN: 50 mM Tris, 300 mM NaCl, pH 7.4). The beads were then washed with TN-10 Buffer (TN-10: 50 mM Tris, 300 mM NaCl, 10 mM Imidizole, pH 7.4). SuperNova was eluted with TN-100 Buffer (TN-100: 50 mM Tris, 300 mM NaCl, 100 mM imidizole, pH 7.4), and desalted with a PD-10 column. Protein concentration was determined by the Bradford assay (Pierce). SuperNova was concentrated using Amicon 10 kDa filters (Millipore) and flash frozen for storage at −80°C.

Cas9 protein was expressed and purified as described by Paix et al. (33). In brief, BL21(DE3)lyseS cells (NEB) expressing pHO4d-Cas9 (gift from Michael Nonet; Addgene plasmid No. 67881) were incubated at 37°C with shaking (225 rpm) overnight. Then, 2.5 mL culture was added to 500 mL lysogeny broth, including 0.1% v/v glucose and ampicillin, and grown to OD 0.5 at 37°C. Culture was chilled on ice for 5 min, before 0.4 mM IPTG was added, and the culture grown at room temperature for 24 h with shaking (225 rpm). The culture was pelleted at 2700 g for 8 min, then resuspended in 6 mL/g wet weight Buffer A [20 mM Tris pH 8.0, 250 mM KCl, 20 mM imidazole, 10% glycerol, 1 mM Tris(2-carboxyethyl)phosphine (TCEP)]. Next, 1 mM PMSF, protease inhibitors (Roche), and 1 mg/mL lysozyme were added. Cells were then sonicated with a needle-tip sonicator at 20% duty cycle for 45 min on ice. Lysate was centrifuged at 16,000 g for 30 min, then protein was incubated with Ni-NTA agarose beads (Qiagen) for 45 min at 4°C. Bead–protein conjugate was then transferred to 5 mL columns (Qiagen) and rinsed with 100 mL Buffer B (20 mM Tris 8.0, 800 mM KCl, 20 mM imidazole, 10% glycerol, 1 mM TCEP). Cas9 was then eluted with Buffer C [20 mM 2-[4-(2-hydroxyethyl)piperazin-1-yl]ethanesulfonic acid (HEPES) pH 8.0, 500 mM KCl, 250 mM imidazole, 10% glycerol]. Protein concentration was determined by Bradford assay (Pierce). DNA was then removed as follows: 1 mL Q sepharose/mL eluate was loaded onto a 5 mL Qiagen column and rinsed with 5 mL 1 M KCl/mL sepharose, then equilibrated with 5 mL Buffer C/mL sepharose. Eluate was then run through sepharose, and flow through discarded. Protein was then eluted with 20 mL Buffer D (20 mM HEPES pH 7.5, 500 mM KCl, 20% glycerol). Buffer exchange was then performed via dialysis in cassettes (Slide-a-lyzer) in Buffer D at 4°C. Cas9 was concentrated using Amicon 100 kDa filters (Millipore) to 25 mg/mL and flash frozen for storage at −80°C.

Optogenetic photoactivation

Photoactivation was performed using a light-emitting diode (LED) system emitting 540–600 nm wavelength through a liquid light guide (GYX module, X-Cite LED1; Excelitas, Waltham MA). For in vitro experiments, irradiance was 6.7 mW/mm², and for in vivo experiments irradiance was 1 mW/mm² on a single 35-mm diameter nematode growth media (NGM) plate for the indicated duration. Irradiance was determined using a calibrated thermopile detector (818P-010-12; Newport Corporation, Irvine, CA) and optical power meter (1916-R, Newport Corporation).

In vitro detection of superoxide via HPLC

SuperNova protein was irradiated in poly(methyl methacrylate) plastic cuvettes in the presence of the ROS probe DHE (0.1 mM, No. D11347; Invitrogen) and 0.1 mM diethylenetriaminepentacetic acid (DTPA) in mitochondrial respiration buffer (MRB, refer to Mitochondrial respiration and isolated enzyme activity assay section). To these samples, an equal volume of 200 mM perchloric acid/methanol was added, centrifuged at 17,000 g, and the supernatant transferred to an equal volume of 1 M potassium phosphate K⁺PO₄⁻ at pH 2.6. Samples were then separated via HPLC (Shimadzu) on a Polar-RP column (150 × 2 mm; 4 μm; Phenomenex) with mobile-phase solvents A and B comprised of 10% and 60% acetonitrile, respectively (both with 0.1% trifluoroacetic acid) a flow rate of 0.1 mL/min. Solvent gradient was as follows (values for solvent B): 0–5 min, 40%; increase to 100% by 25 min, 25–30 min, 100%; decrease to 40% by 35 min; 35–40 min, 40%. Fluorescence detection was used to identify peaks at known retention times for DHE (Ex 358, Em 440 nm), 2-OH-E⁺ (Ex 490, Em 567 nm), and E⁺ (Ex 490, Em 596 nm) were then quantified (Lab Solutions) against a 2-OH-E⁺ standard curve.

Culture of C. elegans and strains

C. elegans were grown on plates containing NGM and OP50 bacterial food using established laboratory techniques (4). All strains used are shown in Table 1 and were provided by the C. elegans Genetics Center (CGC, University of Minnesota, Minneapolis, MN). All alleles generated in this study were backcrossed to N2 at least four times.

Table 1.

List of Caenorhabditis elegans Strains

Strain	Genotype	Source
N2-Bristol	Wild-type	CGC
APW54	sdhb-1(jbmSi12 [sdhb-1::SuperNova]) II	This study
APW149	sdhc-1(jbmSi34 [sdhc-1::SuperNova]) III	This study
APW47	sdhc-1(jbm1 [sdhc-1::mCherry]) III	This study
APW49	sdhb-1(jbm4 [sdhb-1::mCherry]) II	This study
APW40	immt-1(jbmSi7 [immt-1::link::GFP]) X	This study
APW75	sdhb-1(jbmSi4 [sdhb-1::mCherry]) II; immt-1(jbmSi7 [immt-1::link::GFP]) X	This study
APW76	sdhc-1(jbmSi1 [sdhc-1::mCherry]) III; immt-1(jbmSi7 [immt-1::link::GFP]) X	This study
APW224	sdhb-1(jbm12 [sdhb-1::SuperNova]) II; immt-1(jbmSi7 [immt-1::link::GFP]) X	This study
APW225	sdhc-1(jbm34 [sdhc-1::SuperNova]) III; immt-1(jbmSi7 [immt-1::link::GFP]) X	This study
CL2166	dvIs19[pAF15(Pgst-4::GFP::NLS)	CGC
APW61	sdhb-1(jbmSi12 [sdhb-1::SuperNova]) II; dvIs19 [(pAF15) gst-4p::GFP::NLS]	This study
APW150	sdhc-1(jbmSi36 [sdhc-1::SuperNova]) III; dvIs19 [(pAF15) gst-4p::GFP::NLS]	This study
TK22	mev-1(kn1)III	CGC

All alleles generated in this study were backcrossed to N2 four times. Strains were provided by the CGC.

CGC, C. elegans Genetics Center.

C. elegans transgene construction with CRISPR/Cas9

Transgenic worms were generated via established C. elegans CRISPR/Cas9 gene editing protocols (33). In brief, repair template amplicons were designed with 35 nucleotide homology to either side of the cut site. All primers used for repair template amplicons, and target crRNA sequences are detailed in Tables 2 and 3, respectively. Repair templates were PCR amplified and purified using Qiagen MinElute kit. The injection mix was prepared consisting of (final concentrations) 25 mM KCl, 7.5 mM HEPES, 4 μg/μL tracrRNA, 0.8 μg/μL target crRNA, 0.8 μg/μL dpy-10 crRNA, 50 ng/μL dpy-10 ssODN, 2.5 μg/μL Cas9 enzyme, and nuclease-free water to a final volume of 10 μL. Cas9 enzyme was cloned and purified in-house as described in Materials and Methods section above. Injection mix was then centrifuged for 2 min at 17,000 g, and incubated at 37°C for 10 min before microinjection into the germline of ∼25 young adult worms per construct under a DIC microscope. After injection, worms were transferred to individual plates, then again on day 2. Progeny were screened for the dpy-10 dumpy/roller phenotype (or for fluorescence), then at least three independent homozygous lines were selected for subsequent genotyping to confirm that the coding region of SuperNova or mCherry was correctly inserted before the translational stop codon/3′ UTR of mev-1 (ortholog of mammalian sdhc-1) or sdhb-1 to create a translational fusion protein. The coding sequence for GFP was inserted before the translational stop codon/3′ UTR of immt-1 (aka mitofilin) yielding IMMT-1::GFP expressing worms.

Table 2.

List of Amplicon Repair Templates for CRISPR-Homology Directed Repair Editing

Strain	Repair template primers	Template	Product
APW49	F: GCTTACTGGATTCACATCGAAGCCAGCCGCTGAGCCATCTGCATTCATGGTCTCAAAGGGTGAAG	pCFJ90 (Addgene No. 19327)	mCherry
	R: CATCGAATCTAGGCAAAACAATGGAATTTTCTTATTTATACAATTCATCCATGCCACCTG
APW54	F: GCTTACTGGATTCACATCGAAGCCAGCCGCTGAGCCATCTGCATTCATGGGTTCAGAGGTCGGCC	SuperNova/pRSETB (Addgene No. 53234)	SuperNova
	R: CATCGAATCTAGGCAAAACAATGGAATTTTCTTATTTAATCCTCGTCGCTACCGATGGCG
APW47	F: CTTCAACTCTTGCCAGAACAAGAGCAACAAGACTGCtATGGTCTCAAAGGGTGAAGAAG	pCFJ90 (Addgene No. 19327)	mCherry
	R: GCGGAGTAAGAAAAAAGAAGGCGGAGCATCTGTGCTTATACAATTCAATCCATGCCACCTG
APW149	F: CTTCAACTCTTGCCAGAACAAGAGCAACAAGACTGCCGGAGCATCGGGAGCCTCAGGAGCATCGATGGGTTCAGAGGTCGGCCC	SuperNova/pRSETB (Addgene No. 53234)	SuperNova
	R: GAGTAAGAAAAAAGAAGGCGGAGCATCTGTGCCTAATCCTCGTCGCTACCGATG
APW40	F: CTCGCCCAGCTTCTTGTGGCTCACGCCGCCGTCTCATCGATTCGCTCAACTTATCCTCGTAGCCCTCGGACTCCTCGTAGCATGAGCAAGGGCGAGG	pFCK-Arch_GFP (Addgene No. 22217)	GFP
	R: TTTTAAAAAACGGTGACGCAAGACAATCAATTGTTTTACTTGTACAGCTCGTCCATGCCG

GFP, green fluorescent protein.

Table 3.

List of crRNA for CRISPR-Homology Directed Repair editing

Gene target	crRNA target sequence
sdhb-1	TTTCTTATTTAAAATGCTGA
sdhc-1	CAAGAGCAACAAGACTGCCT
immt-1	CTAATAAGTTGAGCGAATCG

RNAi

RNAi was performed as previously described (55). Worms were placed onto NGM plates seeded with HT115 Escherichia coli that transcribes double-stranded RNA (dsRNA) homologous to skn-1 (23).

Western blotting

Each sample lysate was prepared from a single 35 mm NGM plate containing ∼200 developmentally staged adult worms. These were collected into microcentrifuge tubes using room temperature M9 media (22 mM KH₂PO₄, 42 mM Na₂HPO₄, 86 mM NaCl, 1 mM MgSO₄), then shortly centrifuged. Excess M9 was removed before snap freezing in liquid nitrogen. Protein extraction was then performed with the addition of 30–50 μL modified RIPA buffer (50 mM Tris HCl, 150 mM NaCl, 20 mM EDTA, 0.5% sodium deoxycholate, 1% Triton-X, 0.1% SDS, and Halt™ Protease & Phosphatase Inhibitor cocktail [No. 78440; Thermo Fisher], pH 8.0) combined with mechanical homogenization by vortexing for 8 × 30 s on/off ice with 3 × 2 mm zirconin beads (No. 11079124ZX; BioSpec) to disrupt the worm cuticle. After centrifugation at 8000 g for 5 min at 4°C, supernatant was collected, and protein concentration determined by the Lowry assay. After equalizing protein concentration, loading buffer (100 mM Tris HCl, 10% v/v glycerol, 0.2% w/v bromophenol blue, 2% v/v β-mercaptoethanol) was added in equal volume. Samples were then loaded onto 10%–12% gels and separated by SDS-PAGE. Proteins were transferred to nitrocellulose membranes and blocked in 5% nonfat milk/Tris-buffered saline with tween (TBST) for 1 h at room temperature. Membranes were then incubated overnight at 4°C in the appropriate primary antibodies diluted 1:1000 in 5% BSA: anti-KillerRed (No. AB961; Evrogen), anti-GFP (No. 11814460001; Roche), anti-HSP60 (Developmental Studies Hybridoma Bank, University of Iowa, lot No. 20150409), anti-mCherry (No. PA5-34974, lot No. SH2437272B; Thermo Fisher), anti-ATP5a (No. ab14748, lot No. GR160041-12; Abcam), anti-Actin (No. ab14128, lot No. GR319639-1; Abcam), antiactive PMK-1 (No. V1211, lot No. 0000290588; Promega), and anti-2,4-dinitrophenylhydrazine for OxyBlot (No. S7150; Millipore Sigma). After washing in TBST, membranes were incubated in HRP-conjugated secondary antibodies: antirabbit IgG (No. 7074S; Cell Signaling) or antimouse IgG (No. 32430, lot No. RF234708; Thermo Scientific) for 1 h at room temperature. Finally, proteins were visualized using ECL (Clarity Western ECL Substrate; Bio Rad) for image capture (ChemiDoc; Bio Rad).

Mitochondrial respiration and isolated enzyme activity assay

Mitochondria were isolated from ∼200,000 day-1 adult C. elegans in mannitol, sucrose, (3-(N-morpholino) propanesulfonic acid), EDTA buffer (MSM-E) buffer (MSM-E: 220 mM mannitol, 70 mM sucrose, 5 mM 3-morpholinopropane-1-sulfonic acid, 2mM ethylene glycol-bis(2-aminoethylether)-N,N,N′,N′-tetraacetic acid (EGTA), pH 7.4 at 4°C), then resuspended in MRB (120 mM KCl, 25 mM sucrose, 5 mM MgCl₂, 5 mM KH₂PO₄, 1 mM EGTA, 10 mM HEPES, 1 mg/mL fatty-acid free bovine serum albumin (FF-BSA), pH 7.35 at 25°C) as described previously (55). Mitochondria were added to a clark-type electrode chamber (Hansatech Instruments, United Kingdom) at 1 mg/mL⁻¹ in MRB also containing 1 mg/mL⁻¹ FF-BSA. Substrates and inhibitors were then sequentially added: 5 mM succinate and 2 μM rotenone, followed by 0.4 mM ADP to induce state-3 oxidative phosphorylation, then 2.5 mM malonate to inhibit complex-II activity.

Frozen mitochondria isolates were used to measure enzyme activity spectrophotometrically. Mitochondria were freeze-thawed, then Complex II activity was determined as the succinate-driven, malonate-sensitive rate of 2,6-dichlorophenolindophenol reduction (ɛ = 21,000 M⁻¹ at 600 nm) (51). Citrate synthase activity was determined as the rate of 5,5′-disulfanediylbis(2-nitrobenzoic acid)-coenzyme A formation (ɛ = 13,600 M⁻¹ at 412 nm) (55).

Protease-protection assay for determination of inner mitochondrial membrane protein topology

Mitochondria were isolated as described above, and an aliquot of resuspended isolated mitochondria in MSM-E at 5 mg/mL was divided in half, pelleted at 7000 g for 5 min, then the supernatant was removed from each. To form mitoplasts, one of the two aliquots was resuspended in hypotonic swelling buffer (20 mM HEPES, 1 mM EGTA, pH 7.2 at 4°C), and the other in MSM-E to maintain intact mitochondria and incubated on ice for 10 min. Both were pelleted again at 7000 g for 5 min, then resuspended in MRB. Samples were then treated with proteinase-K (No. P81075; Biolabs) at 0.1 U·mg protein⁻¹ for 10 min without or with Triton-X (2.5% v/v), then digestion was inhibited with the serine protease inhibitor, phenylmethyl sulfonyl fluoride (PMSF, 40 mM). Sample loading buffer was added in equal volume, and the SDS-denatured protein lysates were resolved on a 15% gel via SDS-PAGE as described above.

Development rate assay

Day-1 adult worms were moved to an NGM plate and allowed to lay eggs for 30 min, then a single egg was transferred to each new NGM plate and placed in an incubator maintained at 20°C. At regular intervals, images were taken of each worm and body length along the midline determined using ImageJ.

Confocal microscopy

Day 1 adult worms were placed on a 2% agarose pad under anesthetic (0.1% tetramisole), and 1 μm image stacks were acquired using an Olympus FV1000 confocal microscope with a 100 × oil immersion objective (University of Rochester Light Microscopy Core Facility). Single Z-stack images were used to assess cross-section profile plots of large spherical hypodermal mitochondria coexpressing GFP and mCherry or SuperNova using ImageJ. Profile plot data were smoothed by three-point moving average of pixel intensity, then normalized to peak intensity of each profile. To confirm mitochondrial-specific expression of IMMT-1::GFP, these worms were imaged after 2 h incubation in 5 μM Mitotracker Red CMXRos dye (Invitrogen) in M9 media overnight, which preferentially accumulates in the mitochondrial matrix.

Epifluorescence microscopy for GFP transcriptional reporter experiments

Worms were placed on a 2% agarose pad under anesthetic (tetramisole, 0.1% w/v), then images were obtained through GFP and TexasRed filter sets on a fluorescence microscope (Nikon MVX10) equipped with a camera (Lumenera) and acquisition software (Infinity Analyze; Lumenera). GFP reporter expression was then quantified using ImageJ by tracing around individual worms to determine their mean fluorescence intensity.

Glutathione assay

Approximately 20,000 staged day-1 adult worms per plate were exposed to light (1 h, 1 mW/mm²) or kept in dark conditions, then collected in MES buffer [0.2 M 2-(N-morpholino)ethanesulfonic acid, 0.05 M phosphate, 1 mM EDTA, pH 6] and homogenized as described in the Western blotting method using zircon beads. Total and oxidized glutathione were then determined using a commercially available kit (703002, Cayman Glutathione Assay Kit) according to the manufacturer's directions.

AR sensitivity assay

All pretreatment was performed on developmentally synchronized day-1 adult worms. For optogenetic experiments, worms were subjected to 60-min light pretreatment or dark-control, then 1 h later placed at 26°C for 4 h in normoxic conditions in the dark. For anoxia preconditioning, worms were placed in a chamber maintained at <2 ppm O₂ (5%/95% H₂/N₂, palladium catalyst; Coy Lab Products) for the same 4 h period at 26°C. For PQ (N,N′-dimethyl-4,4′-bipyridinium dichloride) pretreatment, worms grown on regular NGM plates were moved to NGM plates containing 4 mM PQ for the 4 h pretreatment period at 26°C, then returned to fresh plates. In all groups, after pretreatment, worms were returned to dark/normoxia conditions in a 20°C incubator for 24 h. Then, day-2 adult worms (≥15 per plate in triplicate per condition, strain and time point) were subjected to anoxia at 26°C for a duration of 18 h unless otherwise indicated. Worms were then returned to normoxia at 20°C to recover for 22–24 h, after which they were scored for the number of surviving worms on each plate. Survival was defined as at least having pharyngeal pumping and side-to-side head movement in response to gentle touch.

Statistical analysis

Data were analyzed by one- or two-way ANOVA where appropriate, with Tukey's correction for multiple comparisons using GraphPad Prism (v7). All data are expressed as mean ± SEM unless otherwise indicated. Exact n for each experiment is reported in the respective figure legend. Significance was accepted at p < 0.05.

Abbreviations Used

2-OH-E⁺

2-hydroxyethidium

ADP

adenosine diphosphate

AR

anoxia reoxygenation

ASK1

apoptosis signal-regulating kinase 1

Cu/Zn-SOD

copper/zinc superoxide dismutase

DHE

dihydroethidium

E⁺

ethidium

EGTA

3-morpholinopropane-1-sulfonic acid, 2mM ethylene glycol-bis(2-aminoethylether)-N,N,N′,N′-tetraacetic acid

ETC

electron transport chain

FF-BSA

fatty-acid free bovine serum albumin

GFP

green fluorescent protein

gst-4p

glutathione S-transferase 4 promoter region

H₂O₂

hydrogen peroxide

HEPES

2-[4-(2-hydroxyethyl)piperazin-1-yl]ethanesulfonic acid

HPLC

high-performance liquid chromatography

HSP60

heat shock protein 60 kDa

IMMT

inner mitochondrial membrane protein 1 (C. elegans ortholog of mitofilin)

IMS

intermembrane space

IPTG

isopropyl β-D-1-thiogalactopyranoside

LED

light-emitting diode

MAPK

mitogen activated protein kinase

mev-1

succinate dehydrogenase cytochrome b560 subunit (C. elegans—methyl viologen sensitive mutant)

MRB

mitochondrial respiration buffer

MSM-E

mannitol, sucrose, (3-(N-morpholino)propanesulfonic acid), EDTA buffer

N2

C. elegans wild-type strain

NGM

nematode growth medium

Ni-NTA

nickel-charged affinity resin

NSY-1

mitogen-activated protein kinase, C. elegans ortholog of ASK1

O₂˙⁻

superoxide

OD

optical density

PBS

phosphate-buffered saline

PMK-1

mitogen-activated protein kinase (C. elegans ortholog of p38 MAPK)

PMSF

phenylmethylsulfonyl fluoride

PQ

paraquat (methyl viologen)

pro-K

proteinase-K

Q

ubiquinone

RET

reverse electron transfer

RNAi

RNA interference

ROS

reactive oxygen species

SDHB

succinate dehydrogenase subunit B

SDHC

succinate dehydrogenase subunit C

SDS-PAGE

sodium dodecyl sulfate polyacrylamide gel electrophoresis

SKN-1

skinhead-1 (C. elegans ortholog of Nrf2)

SOD

superoxide dismutase

TBST

Tris-buffered saline with tween

TCEP

Tris(2-carboxyethyl)phosphine

TRX

thioredoxin

WDR-23

WD40 repeat protein, DDB1- and CUL4-associated factor 11 homolog

Author Disclosure Statement

No competing financial interests exist.

Supplementary Material

Supplementary Figure S1

Supplementary Figure S2

Supplementary Figure S3

Supplementary Figure S4

Supplementary Figure S5