Reversible Thiol Oxidation Inhibits the Mitochondrial ATP Synthase in Xenopus laevis Oocytes

Oocytes are postulated to repress the proton pumps (e.g., complex IV) and ATP synthase to safeguard mitochondrial DNA homoplasmy by curtailing superoxide production. Whether the ATP synthase is inhibited is, however, unknown. Here we show that: oligomycin sensitive ATP synthase activity is significantly greater (~170 vs. 20 nmol/min−1/mg−1) in testes compared to oocytes in Xenopus laevis (X. laevis). Since ATP synthase activity is redox regulated, we explored a regulatory role for reversible thiol oxidation. If a protein thiol inhibits the ATP synthase, then constituent subunits must be reversibly oxidised. Catalyst-free trans-cyclooctene 6-methyltetrazine (TCO-Tz) immunocapture coupled to redox affinity blotting reveals several subunits in F1 (e.g., ATP-α-F1) and Fo (e.g., subunit c) are reversibly oxidised. Catalyst-free TCO-Tz Click PEGylation reveals significant (~60%) reversible ATP-α-F1 oxidation at two evolutionary conserved cysteine residues (C244 and C294) in oocytes. TCO-Tz Click PEGylation reveals ~20% of the total thiols in the ATP synthase are substantially oxidised. Chemically reversing thiol oxidation significantly increased oligomycin sensitive ATP synthase activity from ~12 to 100 nmol/min−1/mg−1 in oocytes. We conclude that reversible thiol oxidation inhibits the mitochondrial ATP synthase in X. laevis oocytes.

Allen [15] posits that: oocytes safeguard mtDNA homoplasmy by repressing OXPHOS to curtail superoxide production. In support, OXPHOS is repressed in oocytes compared to sperm in diverse phyla from jellyfish to mice [16][17][18][19][20]. Repressed OXPHOS is associated with lower mitochondrial free In-house bred X. laevis were maintained at the European Xenopus Resource Centre (EXRC) at 18 °C and fed daily on trout pellets [47]. Following ethical approval (#OLETHSHE1500), X. laevis oocytes were harvested, defolliculated with collagenase, and stored at -80 °C for biochemical analysis. In line with the ARRIVE guidelines [48], biological variability was accounted for by obtaining samples from three different adult females.

F1-Fo ATP Synthase Assay
Mitochondria were isolated by differential centrifugation wherein oocytes (n = 10) were lysed in STE buffer (250 mM sucrose, 200 mM Tris-HCL, 2 mM EDTA, pH 7.2) supplemented with a protease inhibitor tablet, 1% fatty acid free BSA and 100 mM N-ethylmaleimide (NEM) to block reduced thiols for 10 min on ice. Lysates were centrifuged at 700× g for 10 min at 4 °C, before the supernatant was centrifuged at 7000× g for 10 min at 4 °C. After discarding the supernatant, the mitochondrial pellet was resuspended in STE with fresh 10 mM 1-4-Dithiothreitol (DTT) or without (control) for 30 min on ice. Mitochondria were pelleted and washed (3 × 1 min in BSA free STE) to remove excess DTT, before being treated with 50 µg/mL alamethicin to permeabilise the inner membrane to ATP [49], 1 µM diphenyleneiodonium to prevent complex I oxidising NADH by inhibiting the prosthetic flavin mononucleotide group, and 300 nM antimycin A to block complex III. In the reduced group, TCEP (2 mM) was used to maintain a reducing environment (e.g., prevent vicinal dithiols reforming disulfide bonds after reduction). TCEP is preferable to DTT for maintaining a reducing environment because DTT can autoxidise to produce superoxide in the presence of transition metals [12].
F1-Fo ATP synthase activity was assessed by monitoring ATP hydrolysis in the presence of a glycolytic pyruvate kinase (PK), lactate dehydrogenase (LDH), and phosphoenolpyruvate (PEP) system to regenerate ATP. ATP hydrolysis was followed as the decrease in NADH absorbance at 340 nm (extinction coefficient: 6.22 Mm -1 cm -1 ) using a plate reader. Mitochondria were analysed in duplicate in a reaction buffer containing (400 µM NADH, 1 mM PEP, 20 U/mL LDH, 15 U/mL PK, and 2.5 mM ATP in 200 mM Tris, 2 mM MgCl2, 200 µM EDTA, pH 8.0). ATP hydrolysis was followed for 2 min without mitochondria to account for spontaneous ATP hydrolysis. Mitochondria were added and NADH absorbance was monitored for every 15 s 10 min. All wells were then spiked with 6 µM oligomycin and NADH absorbance was followed for 5 min to determine specific F1-Fo ATP synthase activity. After determining protein content using a Bradford assay and subtracting oligomycin insensitive absorbance, F1-Fo ATP synthase activity (nmol/min -1 /mg -1 ) was calculated using the following equation: (Δ 340/min × 1000)/[(extinction coefficient × sample volume) × protein concentration)] [50].

Native Blotting
Mitochondrial membranes were solubilised with 5 µL of 20% dodecyl lauryl maltoside (DDM) in PBS. After centrifuging samples at 16,250× g for 12 min at 4 °C, 0.1% Ponceau S was added to the supernatant to establish a dye front [51]. High resolution clear native polyacrylamide gel electrophoresis (Hr-CNPAGE) was performed as described by Wittig and colleagues [51] using a cathode (50 mM Tricine, 7.5 mM Imidazole, 0.02% DDM, and 0.05% sodium deoxycholate, pH ~7) and anode buffer (25 mM Imidazole, pH ~7). Electrophoresis was performed at 4 °C for 90 min to prevent band broadening and to ensure sufficient protein transfer [52]. Gels were transferred onto a low autofluorescence 0.45 µM polyvinylidene fluoride (PVDF) membrane at 100 V for 60 min in transfer buffer (50 mM Tricine, 7.5 mM Imidazole, pH 7) at 4 °C. Membranes were blocked with 5% non-fat dry milk (NFDM) in PBS for at least 60 min. Primary/secondary antibody incubations and fluorescent detection are described below (see Redox Mobility Shift Assay).

In-Gel ATP Hydrolysis Assay
Mitochondrial membranes were solubilised with DDM and Hr-CNPAGE was performed as described above. After alkylating reduced thiols with NEM (100 mM), reversibly oxidised thiols were treated with (reduced) or without (control) 3.5 mM TCEP for 15 min. Oligomycin controls (500 nM) were loaded in parallel. An equal amount of protein (5 µg) was loaded onto a precast 4-15% gradient gel. Native gels were incubated with buffer (35 mM Tris, 270 mM Glycine, pH 8.3) for 2 h, before being incubated with assay buffer (35 mM Tris, 270 mM Glycine, 14 mM MgSO4, 0.2% Pb(NO3)2, 2.8mM ATP, pH 8.3) for 2 h at room temperature. The reaction was stopped with 50% methanol for 30 min before membranes were scanned. Images were inverted to visualise the white bands and densitometry was performed using VisionWorks TM software (Analytik Jena, Germany).

Catalyst-Free TCO-Tz Immunocapture Coupled to Redox Affinity Blotting
Aliquots (5 µL) of anti-ATP-α-F1 primary antibody were incubated with 10 mM transcyclooctene polyethylene glycol 4 (PEG) N-hydroxysuccinimide (TCO-PEG4-NHS) for 30 min at 4 °C to label primary amines. To prevent unwanted labelling of primary amines in the sample, excess TCO-PEG4-NHS was quenched by adding 5 mM Tris for 15 min. A 45 min total reaction time should ensure NHS hydrolysis outcompetes the carbodiimide reaction thereby preventing adventitious sample labelling. Mitochondrial membranes were prepared as described below (see Catalyst-free TCO-Tz Click PEGylation), except TCO-PEG3-maleimide (TCO-PEG3-NEM) was substituted for 5 mM Biotin-dPEG ® 3-maleimide (Sigma, UK, Biotin-dPEG ® 3-MAL). Biotin-dPEG ® 3-MAL labelled mitochondrial membranes were incubated with TCO-PEG4-NHS labelled primary antibody for 30 min on ice, before being placed in a spin cup containing 85 µL 6-methyltetrazine substituted agarose beads to initiate the catalyst-free Inverse Electron Demand Diels Alder (IEDDA) Click reaction for 90 min on ice. Clicked samples were washed in PBS supplemented with 0.05% DDM for 1 min before being centrifuged for 1 min at 1000× g (washing was repeated five times). Using a spin cup increases purity by removing contaminants with high stringency. Western blotting was performed as described below (see Redox Mobility Shift Assay), expect Streptavidin Alexa Fluor™ 647 (ThermoFisher, UK, 1:500 in TBST for 60 min at room temperature) was used to detect reversibly oxidised subunits on an Analytik Jena (Analytik Jena, Germany) scanner using the appropriate filters (excitation: 600-645 nm; emission: 607-682 nm).

Catalyst-Free TCO-Tz Click PEGylation
We amended whole-cell TCO-Tz Click PEGylation for mitochondria [25]. Oocytes (n = 10) were lysed in STE buffer supplemented with a protease inhibitor tablet, 1% fatty acid free BSA and 100 mM NEM) to block reduced thiols for 10 min on ice. Lysates were centrifuged at 700× g for 10 min at 4 °C, before the supernatant was centrifuged at 7000× g for 10 min at 4 °C. Mitochondrial pellets were resuspended in STE with 5 mM TCEP for 30 min on ice. After washing to remove excess TCEP, mitochondria were resuspended in 5 mM TCO-PEG3-NEM to label newly reduced thiols [36]. Alamethicin (50 µg/mL) was added to ensure TCO-PEG3-NEM could permeate the inner mitochondrial membrane. Mitochondria were lysed in PBS (pH 7.3) supplemented with 1.5% (v/v) 20% DDM. After removing insoluble material by centrifugation (14,000× g for 5 min at 4 °C), the supernatant was incubated with 5 mM 6-methyltetrazine 5 kDa PEG (Tz-PEG5) for 90 min on ice. The catalyst-free IEDDA Click reaction was terminated by adding Laemmli buffer (4% SDS, 20% Glycerol, 0.004% Bromophenol blue and 125 mM Tris HCl, pH 6.8) supplemented with 100 mM DTT before samples were denatured at 80 °C for 5 min.

Statistical Analysis
Oligomycin sensitive F1-Fo ATP synthase activity data were analysed using independent Student's t-tests with alpha ≤ 0.05. TCO-Tz Click PEGylation data were analysed by paired Student's t-tests with alpha ≤ 0.05. Statistical analysis was performed on GraphPad Prism (GraphPad Software, USA). Data are presented as Mean and standard deviation (±).

F1-Fo ATP Synthase Activity is Significantly Greater in Testes Compared to Oocytes
Allen [15] proposes that: oocytes repress OXPHOS to safeguard mtDNA homoplasmy by curtailing superoxide production. If so, F1-Fo ATP synthase activity should be greater in the soma compared to the female germline. Sperm, a major cell type in testes, should sacrifice mtDNA homoplasmy by practicing OXPHOS [15]. Isolating sperm without contaminating somatic tissue (i.e., sertoli cells) is problematic; a situation abetted by a five orders of magnitude difference in mitochondrial number between sperm (10 2 ) and oocytes (10 7 ) in X. laevis [55]. Accordingly, we assessed mitochondrial F1-Fo ATP synthase activity in oocytes compared to testes. We measured ATP hydrolysis using a glycolytic ATP regenerating system in isolated mitochondria treated with alamethicin [49] ( Figure 1A). Using alamethicin to render the inner mitochondrial membrane permeable to ATP placed the rate-limiting step on the F1-Fo ATP synthase by eliminating the influence of the proton pumps and ATP/ADP carrier [22]. Oligomycin sensitive F1-Fo ATP synthase activity is significantly (p ≤ 0.0001) lower in oocytes compared to testes (oocyte: 18.81 ± 15.38; testes: 177.1 ± 37.82 nmol/min -1 /mg -1 ; Figure 1B).

Figure 1.
Oligomycin sensitive F1-Fo ATP synthase activity is significantly greater in testes compared to oocytes. (A) The F1-Fo ATP synthase hydrolysis ATP to ADP. Pyruvate kinase regenerates ATP by using phosphoenolpyruvate (PEP) to phosphorylate ADP to ATP. Lactate dehydrogenase reduces pyruvate to lactate using NADH derived electrons. F1-Fo ATP synthase activity is followed by monitoring the loss of NADH absorbance at 340 nm. (B). Oligomycin sensitive F1-Fo ATP synthase activity is higher significantly (p ≤ 0.0001) in testes (n = 6) compared to oocytes (n = 6) in X. laevis. Statistical significance is indicated by an asterix as assessed by an independent Student's t-test. (C). Native ATP-α-F1 blot image showing the F1-Fo ATP synthase is fully assembled in X. laevis oocytes (n = 4). A minor fraction is present as an F1 subcomplex. Each n is the weighted mean of 10 oocytes.

The F1-Fo ATP Synthase is Assembled in Oocytes
Sieber and colleagues [45] identified a Coomassie stained band on a native gel that may correspond to assembled F1-Fo ATP synthase in X. laevis oocytes. To immunologically confirm F1-Fo ATP synthase assembly, we performed a native blot against the ATP-α-F1 subunit [51,52]. Native blotting reveals the F1-Fo ATP synthase is assembled in X. laevis oocytes ( Figure 1C). F1-Fo ATP synthase disassembly is, therefore, unlikely to explain low oligomycin sensitive ATP hydrolysis in oocytes. Low activity is also unlikely to be attributable to low abundance because ATP-β-F1 content is estimated to be 7.1 µM in X. laevis oocytes [56]. F1-Fo ATP synthase assembly and abundance implies latent enzyme capacity that may be realised by reversing inhibitory thiol oxidation.

Several F1-Fo ATP Synthase Subunits are Reversibly Oxidised
If reversible thiol oxidation is inhibitory, then F1-Fo ATP synthase subunits must be reversibly oxidised. Annotated genome data reveals the F1-Fo ATP synthase contains 18 thiols in X. laevis [42] (see Table 1). After excluding mitochondrial leader sequences, the F1-Fo ATP synthase likely contains 10-11 thiols depending on whether the L or S chromosome copy of ATP-y-F1 is expressed. Available structures suggest cysteine residues in ATP-α-F1 (C 244 , C 294 ), ATP-γ-F1 (C 100 , C 173 ), OSCP (C 139 ), and subunit c (C 149 ) are likely fully and/or partially solvent exposed in catalytic state 3A [57] (Figure 2). Assuming a similar structure in X. laevis, C 104 in subunit b may also be solvent exposed. To determine whether F1-Fo ATP synthase subunits are reversibly oxidised, we used a catalyst-free TCO-Tz immunocapture approach coupled to redox affinity blotting ( Figure 3A). Since the ATP-α-F1 antibody recognises the assembled complex, we used a heterobifunctional TCO-PEG4-NHS linker to form a stable carbodiimide bond with primary amines in the ATP-α-F1 antibody. After labelling reversibly oxidised samples with Biotin-dPEG ® 3-MAL, samples were incubated with TCO-PEG4-NHS labelled ATP-α-F1 antibody before 6-methyltetrazine substituted agarose was used to capture the synthase. Reversibly oxidised F1-Fo ATP synthase subunits were detected by Western blot using a streptavidin conjugated fluorophore [36]. The key lists the cysteine residue by amino acid number for the bovine enzyme with the equivalent X. laevis residue in brackets in bold. A coloured asterix denotes the predicted position of additional cysteine residues in X. laevis. Additional cysteine residues are highlighted in bold in the key if they are likely to be solvent exposed. No structural data is available for subunit g and C3. Catalyst-free TCO-Tz immunocapture coupled to redox affinity blotting reveals discrete bands at ~100, 50, 37, 30, 20-25, and ≤ 10 kDa. Additionally, a distorted band at ~10 kDa was observed, which may reflect how DDM and SDS interact with hydrophobic proteins. Based on a theoretical profile constructed from Table 1 ( Figure 3B), observed bands likely correspond to ATP-α-F1, ATP-γ-F1, OSCP, subunit b and g ( Figure 3C). For OSCP and subunit g, the observed band reflects reversible C 139 and C 96 oxidation, respectively, because they contain a single thiol. After excluding mitochondrial leader sequences, the distorted band is likely an Fo subunit (e.g., subunit c and/or C3). The unassigned ~100 kDa band may represent an interacting protein, hydrophobic aggregate (likely heat induced), and/or crosslinked subunits. Proteomic profiling of the captured complex will be reported elsewhere. F1-Fo ATP synthase subunits are reversibly oxidised in oocytes.  Table 1. (C). Representative image of an experimentally observed reversibly oxidised subunit profile alongside a molecular weight (MW) ladder. Arrows indicate the predicted identity of the observed bands. The image shows several F1-Fo ATP synthase subunits are reversibly oxidised. An unpredicted band at 100 kDa was observed (see main text). Clickable TCO-Tz immunocapture coupled to redox affinity blotting was performed on five pools of 10 X. laevis oocytes. Each lane represents the weighted mean of 10 oocytes.

Discussion
We advance knowledge of reproductive biology by showing, for the first time, that the F1-Fo ATP synthase is inhibited in X. laevis oocytes. Repressed F1-Fo ATP synthase activity could impair OXPHOS by uncoupling Δp from ATP synthesis, but may enhance oocyte viability by constraining ATP hydrolysis. In considering Allen's theory [15], low F1-Fo ATP synthase activity would curtail superoxide production to safeguard mtDNA homoplasmy provided the proton pumps were inhibited. In support, cytochrome c oxidase (i.e., complex IV) activity is repressed and H2O2 levels are low in X. laevis oocytes [23,46]. Complete proton pump inhibition is, however, unlikely because oocyte mitochondria sustain an Δp and still produce some superoxide [23,63]. Protein binding to shield mtDNA and/or selectively curtailing superoxide production at single site (e.g., complex I) may, therefore, be required to safeguard mtDNA homoplasmy. Follow-up studies are required to determine proton pump activity and superoxide production in X. laevis oocytes using state-of-the-art tools (e.g., MitoNeoD [64]).
Existing literature has firmly established that F1-Fo ATP synthase activity is redox regulated in somatic mitochondria [26][27][28][29][30][31][32][33]. Current understanding is, however, restricted to isolated organelles and/or disease models. Whether redox regulation plays a physiological role is, therefore, unclear. We make a major novel contribution by showing that: chemically reversing protein thiol oxidation significantly increases F1-Fo ATP synthase activity in X. laevis oocytes. Our result defines a novel physiological role for mitochondrial reversible thiol oxidation in reproductive biology. Using a redox switch to inactivate the synthase during oogenesis would only imperil mtDNA homoplasmy if any damage sustained was unrepaired. Importantly, cells can transduce redox signals without sustaining oxidative macromolecule damage [65]. The ability of a redox switch to hold the F1-Fo ATP synthase inactive informs several hypotheses. For example, a protein thiol could regulate the metabolic switch from a reliance on fermenting glucose to OXPHOS in the developing X. laevis retina [66]. If a such a developmental Warburg phenotype is redox regulated, it may help rationalise how mitochondrial oxidative stress rewires metabolism in cancer [67].
Unravelling how reversible thiol oxidation inhibits the F1-Fo ATP synthase in oocytes relies on using redox proteomics and site-directed mutagenesis to identify the redox switch(es) (i.e., subunit (s) and site (s)) [68][69][70]. Disambiguating reversible modification type is also important because Sglutathionylation and disulfide bonds may inhibit the enzyme by different mechanisms [39]. Biological precedent exists: ATP-α-F1 S-glutathionylation seems to electrostatically repel nucleotide binding by introducing a bulky negative charge whereas intermolecular disulfide bonds between ATP-α-F1 and ATP-γ-F1 may impair conformational flexibility [30]. Reversible thiol oxidation in Fo is likely to inhibit catalysis by disrupting the ability to bind protons and/or rotate the c-ring [31,[39][40][41]71]. For example, reversible oxidation of subunit c at C 84 could impede proton transport owing to its proximity to glutamate 58 [72,73]. Reversible thiol oxidation may also protect the F1-Fo ATP synthase from irreversible inactivation secondary to sulfinic and sulfonic acid formation [25]. If reversible thiol oxidation impacts inhibitor binding and/or action, then it may lead to oligomycin sensitive enzyme activity being underreported. Given reversible thiol oxidation can activate several enzymes, it is unwise to assume reversible thiol oxidation is always inhibitory. Intriguingly, a redox code may exist wherein the biological outcome varies according to the number of thiols and subunits modified, occupancy (percent oxidation), and modification type [65].
From a structural perspective, S-glutathionylation could lock the F1-Fo ATP synthase in a monomeric state by impeding the formation of dimers, especially if they involve intermolecular disulfide bonds [31]. Alternatively, a negative charge may electrostatically repel dimer interfaces. Thiols in subunit e and g also regulate the stability and formation of oligomers, which dictate inner mitochondrial membrane topology [74][75][76]. For example, F1-Fo ATP synthase dimers shape cristae to create an efficient proton sink for ATP synthesis [77]. Perhaps, Fo redox state underlies the lack of mature cristae in oocytes [19]. Teixeira and colleagues [78] found that OXPHOS is dispensable for germline stem cell differentiation but F1-Fo ATP synthase dimers are essential because they are required for cristae maturation. The results of their study and the present work raise the possibility of at least two redox switches: (1) to prohibit dimers; and (2) to inhibit catalysis. Two redox switches would endow mitochondria with the capacity to differentiate cristae without a catalytically active F1-Fo ATP synthase. That is, to uncouple cristae differentiation from OXPHOS [78].