Sulfide quinone oxidoreductase contributes to voltage sensing of the mitochondrial permeability transition pore

Abstract

Pathological opening of the mitochondrial permeability transition pore (mPTP) is implicated in the pathogenesis of many disease processes such as myocardial ischemia, traumatic brain injury, Alzheimer's disease, and diabetes. While we have gained insight into mPTP biology over the last several decades, the lack of translation of this knowledge into successful clinical therapies underscores the need for continued investigation and use of different approaches to identify novel regulators of the mPTP with the hope of elucidating new therapeutic targets. Although the mPTP is known to be a voltage-gated channel, the identity of its voltage sensor remains unknown. Here we found decreased gating potential of the mPTP and increased expression and activity of sulfide quinone oxidoreductase (SQOR) in newborn Fragile X syndrome (FXS) mouse heart mitochondria, a model system of coenzyme Q excess and relatively decreased mPTP open probability. We further found that pharmacological inhibition and genetic silencing of SQOR increased mPTP open probability in vitro in adult murine cardiac mitochondria and in the isolated-perfused heart, likely by interfering with voltage sensing. Thus, SQOR is proposed to contribute to voltage sensing by the mPTP and may be a component of the voltage sensing apparatus that modulates the gating potential of the mPTP.

Graphical Abstract

Sulfide quinone oxidoreductase (SQOR) contributes to the voltage response of the mitochondrial permeability transition pore (mPTP). Normally, depolarization of the inner mitochondrial membrane (IMM) leads to mPTP opening. Pharmacological inhibition or genetic silencing of SQOR increases mPTP open probability at high ΔΨ, while excess SQOR shifts the gating potential of the mPTP such that it is closed at relatively lower ΔΨ.

Abbreviations

AA

antimycin A

AM

acetoxymethyl

ANT

adenine nucleotide translocator

ATR

atractyloside

BCA

bicinchoninic acid

BDM

2,3-butanedione monoxime

BSA

bovine serum albumin

CoQ

coenzyme Q

CsA

cyclosporine A

CypD

cyclophillin D

DB

decylubiquinone

ETC

electron transport chain

EV50

voltage at 50% sensitivity

FMRP

fragile X mental retardation syndrome

FXS

fragile X syndrome

GSH

glutathione

IMM

inner mitochondrial membrane

mito

mitochondria

mPTP

mitochondrial permeability transition pore

MYX

myxothiazol

NA

numerical aperature

oligo

oligomycin

OSCP

oligomycin sensitivity conferring protein

PAO

phenylarsine oxide

PT

permeability transition

ROS

reactive oxygen species

SQOR

sulfide quinone oxidoreductase

TMPD

N,N,N′,N′-tetramethyl-p-phenylenediamine

TMRE

tetramethylrhodamine ethyl ester

TPP⁺

tetraphenylphosphonium

VDAC

voltage-dependent anion channel

1 INTRODUCTION

Myocardial ischemia, traumatic brain injury, Alzheimer's disease, and diabetes are just a few of the seemingly disparate disease processes that are linked by dysregulation of a common entity: the mitochondrial permeability transition pore (mPTP, i.e., pore).^1-3 Given its mechanistic contribution to these diseases, the pore is considered a prime therapeutic target. The mPTP is thought to reside in the typically impermeable inner mitochondrial membrane (IMM) where it mediates the abrupt, non-selective, and non-specific increase in permeability to small ions and solutes of up to 1.5 kDa, a phenomenon known as the permeability transition (PT).^{2, 4} Pore opening can affect the mitochondrial membrane potential (ΔΨ), ATP and reactive oxygen species (ROS) production, and calcium homeostasis, and ultimately threatens cell viability.² Prolonged high-conductance opening of the pore is pathologic and is implicated in numerous disease processes.^{1, 2} Such pore opening results in collapse of the ΔΨ, mitochondrial matrix swelling, loss of pyridine nucleotides, outer membrane rupture, release of pro-apoptotic proteins, and cell death.^{2, 5} In contrast, low-conductance opening of the mPTP is thought to be physiological and necessary for cellular health, limiting the IMM permeability to ~300 Da.² Low-conductance pore opening, as a source of physiologic proton leak, causes controlled changes in ΔΨ and electron flux through the electron transport chain to regulate ROS signaling, calcium homeostasis, and cellular development and maturation.^6-9

A notable challenge to studying the pore is the lack of consensus on the molecular identity of the mPTP structure. The most recent data suggest that the F1/F0 ATP synthase (ATP synthase) and the adenine nucleotide translocator (ANT) are responsible for high- and low-conductance modes, respectively.¹⁰ Despite the continued debate regarding the pore-forming components, certain key characteristics of the mPTP are indisputable. First, a plethora of small molecule modulators, such as calcium, inorganic phosphate, fatty acids, and ROS, are known to stimulate mPTP opening, while adenine nucleotides (particularly ADP), magnesium, protons, and cyclosporine A (CsA) inhibit (or desensitize) the pore.^{6, 10} In addition, a variety of coenzyme Q (CoQ) analogs modulate mPTP open probability and CoQ deficiency or excess has been shown to result in tissue-specific differences in open and closed mPTP probability, respectively.^11-13 Second, the mPTP is known to be voltage-gated, whereby depolarization of the IMM causes pore opening.¹⁴ Classical studies indicate that the redox state of vicinal thiol groups alters the voltage response of the mPTP, such that oxidation tunes the voltage sensor to open the pore at relatively higher ΔΨs while reduction shifts the gating potential to lower values.¹⁵ However, the identity of the proteinaceous voltage sensor remains unknown.

Sulfide quinone oxidoreductase (SQOR) attracted our attention because it: (1) is located within the IMM, (2) binds and reduces CoQ during its catalysis of hydrogen sulfide (H₂S), and (3) contains redox active vicinal thiol groups at cysteine 379 and 201.^{16, 17} Interestingly, prior studies have also demonstrated that H₂S can inhibit calcium-induced mPTP opening.^{18, 19} Thus, we hypothesize that SQOR is a novel regulator of the mPTP and may be the voltage sensor. In this work, we aimed to define the voltage threshold for low conductance mPTP opening and quantify expression and activity of SQOR in cardiomyocyte mitochondria from newborn Fragile X syndrome (FXS) mouse pups (a model system of CoQ excess and relatively decreased mPTP open probability).¹³ We further utilized pharmacological inhibition and genetic silencing of SQOR to determine the effect on calcium-induced opening and voltage response of the mPTP in adult murine cardiac mitochondria. Here we demonstrate that increased SQOR expression and activity is associated with decreased mPTP open probability and a shift in the threshold for opening to lower ΔΨs. Furthermore, inhibition or loss of SQOR expression increased open probability of the pore in vitro and in the isolated-perfused heart, likely by interfering with voltage sensing. Thus, we propose a novel role for SQOR as a contributor to voltage sensing by mPTP and may be a component of the voltage sensing apparatus that modulates its gating potential.

2 MATERIALS AND METHODS

2.1 Animals

The care of animals in this study was in accordance with NIH and Columbia University Medical Center Institutional Animal Care and Use Committee guidelines and conformed to provisions of the Animal Welfare Act (NIH/DHHS) and the Association for Assessment and Accreditation of Laboratory Animal Care (AAALAC). Fmr1 KO (FVB.129P2-Pde6b⁺ Tyr^c-ch Fmr1^tm1Cgr/J), FVB controls (FVB.129P2-Pde6b⁺ Tyr^c-ch/AntJ), and C57BL6/J mouse strains were purchased from Jackson Laboratory (Bar Harbor, ME). Sqor KO mice (C57BL6/J-SqordE6) were generated by the Genetically Modified Mouse Model Shared Resource (Columbia University, NY). Fmr1 KO and FVB control mice were bred to yield pups that were postnatal day 10 to model a timepoint in human infancy, while C57Bl6/J were studied at 6-8 weeks of age to model for a timepoint in young adulthood. Sqor KO along with wild-type littermate control mice were studied at 4–6 weeks of age.

2.2 Sqor KO mouse generation

Sqor KO mice were generated using CRISPR/Cas9 technology to delete the entire Sqor exon 6 (Figure S1A,B). The CRISPOR tool was used to design four sgRNAs flanking the Sqor exon 6 (Figure S1B). The four sgRNAs were combined with the Cas9 protein to form ribonucleoprotein complex and injected into the pronuclei of fertilized C57BL/6J eggs. A male founder with the deletion of the Sqor exon 6 (Sqor^dE6) was identified and bred with female wild-type C57BL/6J to transmit the Sqor^dE6 allele and obtain Sqor heterozygous mice. The genotyping was performed with standard PCR and gel electrophoresis (Figure S1C,D). Heterozygous mice were back-crossed to the C57BL/6J background for more than six generations. To generate Sqor KO mice, heterozygous male and female mice were bred together. Genotyping was confirmed by PCR. Sqor^dE6 was inherited according to the predicted Mendelian ratio and Sqor KO mice were phenotypically similar to wild-type littermates before weaning. Similar to the previously described Sqor^ΔN/ΔN mutation (mice that lack mitochondrial SQOR and inappropriately express cytoplasmic SQOR), Sqor KO animals ceased to grow and gradually became emaciated after weaning and developed ataxia (Figure S1E,F).²⁰ All Sqor mutants died by 7 weeks of age (Figure S1G). Heterozygotes appeared phenotypically normal and were viable and fertile.

2.3 Body weight measurement

Mouse body weight was determined with an analytical balance (Mettler Toledo PB303-S, Columbus OH).

2.4 Mitochondrial isolation

Mouse heart mitochondria were isolated as previously described.¹³ Cardiac ventricles were harvested and homogenized in ice-cold isolation buffer (225 mM mannitol, 75 mM sucrose, 1 mM EGTA, 5 mM HEPES-KOH (pH 7.2), and 1 mg mL⁻¹ of fatty-acid-free bovine serum albumin (BSA)). The homogenate was spun at 1100 g for 2 min at 4°C. Supernatant (0.8 mL) was removed and then layered on 0.7 mL 15% Percoll® gradient and centrifuged at 18 500g for 10 min at 4°C. The mitochondria fraction was collected and resuspended in 0.7 mL of washing buffer (250 mM sucrose, 5 mM HEPES-KOH (pH 7.2), 0.1 mM EGTA, and 1 mg mL⁻¹ of BSA). The suspension was centrifuged at 10 000g for 5 min at 4°C. The mitochondrial pellet was resuspended in washing buffer and mitochondrial protein concentrations were determined using the Pierce bicinchoninic acid (BCA) assay kit (ThermoFisher Scientific, Waltham, MA).

2.5 Quantification of ΔΨ during proton leak respiration

The voltage threshold for mPTP opening was determined during proton leak respiration as previously described.²¹ Oxygen consumption and ΔΨ were simultaneously measured using a Clark-type electrode (Oxytherm, Hansatech, UK) and a tetraphenylphosphonium (TPP⁺) ion sensitive electrode (World Precision Instruments, Sarasota, FL). Isolated cardiac ventricle mitochondria (0.1 mg) were added to 1 mL of respiration buffer (200 mM sucrose, 25 mM KCl, 2 mM K₂HPO₄, 5 mM HEPES-KOH (pH 7.2), 5 mM MgCl₂, 0.2 mg mL⁻¹ BSA) containing 80 ng mL⁻¹ nigericin (to collapse ΔpH), 5 μM rotenone and 5 mM succinate at 37°C. Proton leak respiration was induced with oligomycin (2.5 μg mL⁻¹). Sensitivity or insensitivity to CsA (1 μM) was used to determine open probability of the mPTP over a range of ΔΨs. Sensitivity to CsA (open mPTP) was defined as a concomitant decline in oxygen consumption with a rise in ΔΨ (closure of a CsA-sensitive leak channel; i.e., the mPTP). ΔΨ was determined based on TPP⁺ calibration curves using the Nernst equation.

For experiments involving pharmacological SQOR inhibition: 120 μM (N,N,N′,N′-tetramethyl-p-phenylenediamine) TMPD and 480 μM ascorbate were used to reduce cytochrome c directly (in lieu of succinate) given that SQOR inhibitors also inhibit Complex III. Antimycin A (4.5 μM) was used to inhibit SQOR.^{22, 23} CsA sensitivity was assessed as described above over a range of ΔΨs. ΔΨ was determined based on TPP⁺ calibration curves using the Nernst equation. Of note, auto-oxidation of TMPD is inherent to its use. Though we did not directly measure TMPD auto-oxidation, experiments were performed under identical conditions with the assumption that the degree of TMPD auto-oxidation would be similar between experiments.

2.6 Western blotting

20 μg samples of cardiac mitochondrial protein were subjected to SDS-acrylamide gel electrophoresis and immunoblotting. SQOR expression was assessed using a primary polyclonal rabbit anti-SQOR antibody (Sigma, hpa017079). Mitochondrial protein loading was assessed with a primary monoclonal antibody to mouse voltage-dependent anion channel [(VDAC); Abcam, (ab15895)] and cytosolic protein loading was assessed with a primary monoclonal antibody to mouse actin (ThermoFisher Scientific MA5-15739). Appropriate secondary antibodies were utilized, signal was detected with enhanced chemiluminescence (ECL; Amersham Pharmacia Biotech, Piscataway, New Jersey, USA), and density was measured using scanning densitometry. SQOR density was normalized to VDAC.

2.7 SQOR enzyme activity

Crude mitochondrial extracts were prepared as previously described.¹³ Isolated mitochondria underwent five freeze–thaw cycles. Steady-state SQOR glutathione reductase activity was determined spectrophotometrically by measuring the rate of oxidation of decylubiquinone (DB) at 278 nm at 25°C.²⁴ Assays were conducted in 1 mL of 20 mM HEPES buffer (pH 7.2) with 1 mM ETDA, 2 mM KCN, 4 μM rotenone, 5 mM malonate, 0.025% dodecyl maltoside, 100 μM DB, 25 mM glutathione (GSH). 100 μg of isolated mitochondria was added to initiate the reaction. To determine specific activity, 20 μg of Antimycin A (AA) was added to the buffer in a parallel assay to inhibit SQOR. Specific activity was calculated using 12.7 mM⁻¹ cm⁻¹ as the extinction coefficient of DB at 278 nm.

2.8 Calcium loading capacity

Calcium uptake and release were measured using cardiac mitochondria (50–100 μg) in 1 mL of respiration buffer (200 mM sucrose, 25 mM KCl, 2 mM K₂HPO₄, 5 mM HEPES-KOH (pH 7.2), 5 mM MgCl₂, 0.2 mg mL⁻¹ BSA) containing 5 μM rotenone, oligomycin (2.5 μg mL⁻¹), and 10 mM succinate at 37°C.^{12, 25} Calcium concentration was determined using an ion-sensitive selective electrode (World Precision Instruments, Sarasota, FL) and calculated based on calibration using sequential additions of CaCl₂.^{12, 25}

For experiments involving pharmacological SQOR inhibition: 120 μM TMPD +480 μM ascorbate was used to reduce cytochrome c directly (in lieu of succinate) given that SQOR inhibitors also inhibit Complex III. 25 μM Phenylarsine oxide (PAO) and 200 μM atractyloside (ATR) were utilized to open the mPTP, and 1 μM CsA and 100 μM ADP were added to inhibit the pore. 4.5 μM AA and 10 μM myxothiazol (MYX) were added to inhibit SQOR and 10 μM 2-methoxy-antimycin A (methoxyAA) served as a control inhibitor since it is an antimycin A analogue that does not inhibit electron transport.^{22, 23, 26}

2.9 Isolated-perfused mouse heart preparation

4- to 6-week-old mice were heparinized (10 kU kg⁻¹ ip) and anesthetized with intraperitoneal pentobarbital (70 mg kg⁻¹). The heart was rapidly excised and the aorta was cannulated (22-guage cannula) as previously described.²⁷ Retrograde perfusion was initiated at constant flow (5 mL/min) with a modified Krebs–Henseleit buffer containing (mmol/L) NaCl 120, KCl 4.7, MgSO₄ 1.2, KH₂PO₄ 1.2, CaCl₂ 1.25, NaHCO₃ 25, and glucose 11. Non-recirculating buffer was maintained at pH 7.4 equilibrated with 95% O₂–5% CO₂ at 37°C. Hearts were allowed to stabilize for 20 min. Exclusion criteria were enforced as previously described.

2.10 Real-time confocal imaging

Calcein-acetoxymethyl (AM) ester (Sigma-Aldrich) and tetramethylrhodamine ethyl ester (TMRE) (Sigma-Aldrich) were prepared in DMSO as 4 and 1 mM stock solutions, respectively. Cobalt chloride (CoCl₂) (Sigma-Aldrich) and 2,3-butanedione monoxime (BDM) (Sigma-Aldrich) were prepared in modified Krebs–Henseleit buffer. Following the stabilization period, isolated-perfused hearts were loaded with calcein-AM (5 μM) and TMRE (20 μM) followed by a 12-min washout.²⁸ Hearts were placed under a confocal microscope (Nikon Eclipse) and temperature was maintained at ~37°C. All imaging was performed using a 1.3–numerical aperture (NA) 40× oil immersion objective lens to obtain 1024 × 1024 pixel images for cardiomyocyte analysis. Hearts were simultaneously perfused with BDM (10 mM) (to prevent contraction-induced movement) and CoCl₂ (2 mM) to quench calcein fluorescence. Ventricular cardiomyocytes were easily identified and images were obtained 15 and 60 min after quenching. Cardiomyocytes in 1–3 imaged fields per mouse were identified and calcein and TMRE fluorescence were quantified using Fiji image analysis software within each region of interest. Calcein and TMRE fluorescence were expressed in arbitrary units and wildtype values were arbitrarily set to equal 1.

2.11 Sulfide measurement

Sulfide levels were measured in cardiac ventricle tissue homogenate or in isolated mitochondria using a H₂S colorimetric assay kit (Elabscience Biotechnology, Houston, TX) following the manufacturer's directions.

2.12 Quantification and statistical analysis

Data were analyzed and presented as mean ± standard error of the mean (SEM) or standard deviation (SD) as indicated. The sample size for each experiment is indicated in each figure legends. Pairwise assessment was made using a two-tailed, Student's t-test. One-way ANOVA with Tukey's post hoc test was utilized to compare three or more groups considering only one variable. Significance was set at p < .05; ****p < .0001, ***p < .001, **p < .01, *p < .05. Statistical analysis was performed using GraphPad Prism 9 software (GraphPad Software, La Jolla, CA).

3 RESULTS

3.1 Decreased mPTP open probability and altered voltage sensing in newborn Fmr1 KO cardiomyocyte mitochondria

Fmr1 mutant newborn mouse cardiomyocyte mitochondria have excess CoQ, relatively decreased mPTP open probability, and less proton leak than wildtype controls.¹³ However, inducible high-conductance mPTP opening has never been assessed in this model system, and it is unknown if voltage sensing is impaired in the context of low-conductance pore opening. Thus, we first determined the threshold for mPTP opening in cardiac mitochondria by quantifying the calcium loading capacity in newborn Fmr1 mutant and FVB control cardiac mitochondria (Figure 1A). The amount of calcium required to trigger mPTP opening was markedly and significantly increased in Fmr1 KO cardiac mitochondria, confirming reduced sensitivity to calcium-induced opening mutants (Figure 1B,C).

FIGURE 1

Altered mPTP and SQOR characteristics in Fmr1 KO cardiomyocyte mitochondria. (A) Schematic of the calcium loading capacity assay is depicted. The amount of calcium (Ca²⁺) required to induce mPTP opening following uptake is quantified in isolated mitochondria during leak respiration. (B) Representative traces of calcium uptake and release in Fmr1 KO and FVB control mitochondria are depicted. Numbers indicate calcium loading capacity. (C) Graphical representation of calcium loading capacity. n = 3–4 per group. Values are expressed as means ± SEM. p values were calculated by Student's t-test. **p < .01. (D) Representative tracings of CsA insensitivity at a relatively high ΔΨ and CsA sensitivity at a relatively low ΔΨ are depicted. O₂ consumption (above) and ΔΨ (below) were simultaneously recorded in isolated mitochondria during succinate-dependent leak respiration. CsA-sensitive mitochondria (red) exhibited a decline in O₂ consumption rate (red numbers, measured in nmol mL⁻¹ min⁻¹ mg mitochondrial protein⁻¹) with a concomitant increase and stabilization of ΔΨ indicating a CsA-mediated blockade of proton leak. CsA failed to block leak in insensitive mitochondria (black). Arrows indicate timing of the addition of mitochondria (mito), oligomycin (oligo), and CsA to the respiratory buffer. (E) CsA sensitivity was determined over a range of ΔΨs in Fmr1 KO and FVB newborn isolated heart mitochondria. Data were fit to variable slope stimulus–response curves and the voltage at which there was 50% CsA sensitivity (EV₅₀) were compared with confidence intervals set to 95%. ****p < .0001. (F) Representative immunoblot of SQOR in Fmr1 KO and FVB newborn isolated heart mitochondria. Voltage-dependent anion channel (VDAC) was used as a loading control. (G) Graphical representation of normalized cardiomyocyte SQOR densities is shown. The mean values from FVB controls were arbitrarily set to 1. n = 11–13 per group. Values are expressed as means ± SEM, p values were calculated using Student's t-test. ****p < .0001. (H) SQOR specific activity was determined spectrophotometrically in Fmr1 KO and FVB newborn isolated heart mitochondria. n = 8 per group. Values are expressed as means ± SEM, p values were calculated using Student's t-test. *p < .05.

Next, we assessed open probability of the mPTP in Fmr1 KOs and FVB controls during leak respiration (low-conductance) over a range of ΔΨs. We simultaneously measured oxygen consumption and ΔΨ in isolated mitochondria using a Clark-type electrode and an ion-selective TPP⁺ electrode, respectively, and determined open/closed state of the mPTP by assessing CsA sensitivity/insensitivity in oligomycin-treated mitochondria. Sensitivity to CsA and closure of the mPTP was defined as a decline in respiration with a concomitant rise in ΔΨ (i.e., CsA-induced block of proton leak) (Figure 1D). The CsA sensitivity-voltage curve for Fmr1 KOs was left shifted with a significantly lower EV₅₀ compared to FVB controls, indicating CsA insensitivity (i.e., decreased open mPTP probability) at ΔΨ voltages where FVB leak respiration was sensitive to CsA (Figure 1E). Thus, the gating potential of the mPTP was shifted to lower ΔΨs in Fmr1 KO mitochondria.

3.2 SQOR expression and activity are increased in newborn Fmr1 KO mouse cardiomyocyte mitochondria

Given that Fmr1 KOs demonstrate decreased open mPTP probability, a shift in gating potential, and harbor excess CoQ, we next explored the role of SQOR. We first quantified SQOR levels in isolated mitochondria via immunoblot analysis and then measured SQOR enzyme kinetic activity via spectrophotometry in Fmr1 mutants and FVB controls. Steady-state levels and specific activity of SQOR were significantly increased in Fmr1 KO mitochondria compared to controls (Figure 1F–H). Thus, SQOR expression and catalytic activity were increased in Fmr1 mutant mitochondria. Of note, we also found that H₂S levels were increased in Fmr1 KO cardiac ventricle tissue homogenates compared to FVB controls (Figure S3A). So, it is possible that the increase in SQOR expression and activity in Fmr1 mutants could result from the increase in H₂S.

3.3 SQOR inhibition decreases the threshold for high conductance mPTP opening and alters mPTP voltage sensing during low conductance opening

Given the association between increased SQOR expression and activity and decreased mPTP open probability in Fmr1 KO cardiomyocytes, we next assessed the effect of SQOR inhibition in wild-type adult murine cardiomyocyte mitochondria. First, we quantified the threshold for high conductance mPTP opening. Calcium loading capacity was determined in the presence of the SQOR inhibitors, antimycin A (AA) and myxothiazol (MYX), along with the control inhibitor, methoxy-AA. Because AA and MYX also inhibit Complex III of the electron transport chain, we utilized TMPD and ascorbate to directly reduce cytochrome c in these experiments. Pore activators (ATR and PAO) and inhibitors (desensitizers) (CsA and ADP) were also employed. As expected, ATR and PAO significantly decreased calcium loading capacity compared to non-treated controls while CsA and ADP significantly increased it (Figure 2A,B). AA and MYX significantly decreased calcium loading capacity versus non-treated controls while methoxyAA had no significant effect (Figure 2A,B). The profound and significant decrease in calcium loading capacity induced by AA was abrogated independently by both CsA and ADP (Figure 2A,B). Thus, SQOR inhibition decreased the threshold for high conductance mPTP opening and could be overcome. Importantly, pharmacologic inhibition of SQOR did not significantly alter H₂S levels in isolated mitochondria (Figure S3C).

FIGURE 2

SQOR inhibition sensitizes the mPTP to calcium-induced opening and disrupts voltage gating of the mPTP in wildtype adult mouse cardiac mitochondria. (A) Representative traces of the calcium loading capacity assay in isolated mitochondria are depicted. Effect of mPTP activators (PAO and ATR), mPTP inhibitors (CsA and ADP), and SQOR inhibitors (AA and myxothiazol) was tested. MethoxyAA and non-treated mitochondria served as negative controls. Numbers indicate calcium loading capacity. (B) Graphical depiction of calcium loading capacity is shown. Values are means ± SEM from 17 (no treatment), 2 (PAO), 4 (ATR), 4 (CsA), 3 (ADP), 5 (AA), 6 (MYX), 2 (methoxy-AA), 4 (AA+CsA), 3 (AA+ADP) biological replicates. *p < .05, **p < .01, and ****p < .0001 by one-way ANOVA. (C) Representative tracings demonstrating CsA sensitivity in AA-treated mitochondria and CsA insensitivity in untreated mitochondria are shown. O₂ consumption (above) and ΔΨ (below) were simultaneously recorded in isolated mitochondria during TMPD/ascorbate-induced leak respiration. CsA was added at similar ΔΨs to specifically inhibit proton leak via the mPTP. AA-treated mitochondria exhibited a decline in O₂ consumption rate (red numbers, measured in nmol mL⁻¹ min⁻¹ mg mitochondrial protein⁻¹) and a concomitant increase and stabilization of ΔΨ after CsA addition. Arrows indicate the addition of mitochondria (mito), AA, and CsA. (D) CsA sensitivity was determined over a range of ΔΨs in AA-treated and untreated mitochondria. Data were fit to variable slope stimulus–response curves and the voltage at which there was 50% CsA sensitivity (EV₅₀) was determined. AA-treated mitochondria was CsA sensitive at 100% of voltages examined.

Next, we determined the effect of SQOR inhibition on low conductance open probability of the mPTP during leak respiration over a range of ΔΨs. We simultaneously measured oxygen consumption and ΔΨ in isolated cardiac mitochondria using a Clark-type electrode and an ion selective TPP⁺ electrode, respectively, and determined open/closed state of the mPTP by assessing CsA sensitivity/insensitivity during oligomycin-induced respiration. Mitochondria were energized with TMPD and ascorbate to bypass Complex III. We compared the effect of AA-treatment versus no treatment. As expected, non-treated cardiac mitochondria demonstrated CsA sensitivity over a range of ΔΨs beginning at ~223 mV (Figure 2C,D). However, AA-treated mitochondria were sensitive to CsA at all ΔΨs tested (Figure 2C,D). Thus, SQOR inhibition opened the mPTP and altered its voltage dependence during low conductance opening. As with the calcium loading capacity assays, pharmacologic inhibition of SQOR did not significantly change H₂S levels in the isolated mitochondria (Figure S3C).

3.4 Genetic silencing of SQOR increases high-conductance mPTP open probability in isolated mitochondria and alters mPTP voltage sensing in the isolated-perfused heart

Given the non-specific nature of SQOR pharmacological inhibition, we next aimed to assess the specific effect of genetic silencing. Global Sqor KO mice were generated using the CRISPR/Cas9 approach and genotype was confirmed using standard PCR, immunoblot analysis, and assessment of SQOR activity (Figure S1). Sqor KOs were bred and tested along with littermate wild-type controls. We first assessed the threshold for high-conductance mPTP opening in cardiomyocyte mitochondria in each strain using the calcium loading capacity assay. Consistent with pharmacological inhibition, Sqor KOs demonstrated significantly lower calcium loading capacity compared with wild-type controls (Figure 3A,B). Thus, genetic silencing of SQOR increased high-conductance open probability of the mPTP. Notably, H₂S levels measured in Sqor KO cardiac ventricle tissue homogenates were not significantly different from wild-type controls.

FIGURE 3

Increased mPTP sensitivity to calcium-induced opening and altered voltage gating in Sqor KO cardiomyocyte mitochondria. (A) Representative traces of the calcium loading capacity assay in isolated mitochondria obtained from Sqor KO mice and wild-type littermate controls are depicted. Numbers indicate calcium loading capacity. (B) Graphical depiction of calcium loading capacity is shown. Values are means ± SEM from 7–8 biological replicates. p values were calculated by Student's t test. ***p < .001. (C–E) Real-time confocal imaging of the isolated perfused mouse heart was used to measure ΔΨ and determine the open probability of the mPTP. Following stabilization, hearts were treated with calcein-AM and TMRE for 30 min. Calcein was subsequently quenched with cobalt chloride. (C) Calcein (green) and TMRE (red) were easily seen and colocalized within actively respiring mitochondria of Sqor KO and wild-type littermate control ventricular cardiomyocytes immediately following cobalt quenching. Scale bar is 50 μm. (D) Representative images of cardiomyocyte calcein and TMRE in Sqor KOs and wild-type littermate controls obtained at 25× 60 min after cobalt quenching. Scale bar is 50 μm. (E) Graphical depiction of myocardial TMRE and calcein fluorescence quantified 60 min post-cobalt. Values are means ± SEM. Control values were arbitrarily set to equal 1. Data represent 3–4 biological replicates per group, cardiomyocytes quantified in section ranged from 24 to 61. p values were calculated using Student's t-test. ****p < .0001.

Interestingly, CsA had no significant effect on the calcium loading capacity in Sqor mutants (Figure S2A,B). The lack of CsA sensitivity in Sqor KOs was also seen during leak respiration, thus, preventing us from determining voltage dependence of low conductance opening of the mPTP (Figure S2C). Although not tested, this phenomenon may relate to changes in cyclophilin D expression given the evidence for a potential SQOR-cyclophilin D interaction (CypD KO mouse hearts demonstrate significant SQOR upregulation).²⁹ Thus, as an alternative, we opted to simultaneously assess ΔΨ and open probability of the mPTP in situ.

To achieve this, we performed real-time confocal imaging of isolated-perfused Sqor KO and wild-type mouse hearts to image calcein and TMRE fluorescence within cardiomyocytes over time. Calcein-AM readily gains access to all cellular compartments, including mitochondria, where it is converted to fluorescent calcein by esterases in living tissue. Cobalt chloride is then used to quench non-mitochondrial calcein fluorescence to enable visualization of mitochondrial entrapped calcein. However, mPTP opening allows cobalt to enter the mitochondrial matrix to abolish intramitochondrial calcein fluorescence. Thus, quenching of mitochondrial fluorescence is an indicator of an open mPTP. Actively respiring mitochondria were simultaneously labeled with TMRE for colocalization and for real-time dynamic quantification of the ΔΨ.

Calcein fluorescence was easily visualized within Sqor KO and wild-type littermate control ventricular cardiomyocytes and colocalized with TMRE within mitochondria (Figure 3C). TMRE fluorescence observed at 60-min post-cobalt quenching was not significantly different between Sqor KO cardiomyocytes and controls (Figure 3D,E). Thus, the ΔΨ generated by Sqor KO and control cardiomyocytes were similar at this timepoint. In contrast, calcein fluorescence in Sqor KOs was markedly and significantly decreased 60 min post-cobalt quenching compared to controls (Figure 3D,F). Thus, Sqor KO hearts demonstrated increased mPTP open probability relative to control hearts. Importantly, dramatic loss of calcein fluorescence with relative preservation of TMRE (on par with controls) indicated that mPTP opening was out of proportion to any loss of ΔΨ that might be expected as a consequence of the ischemia–reperfusion inherent to any Langendorff preparation. Hence, much of the mPTP opening detected in Sqor KO hearts was likely due to low conductance opening. The findings suggest that genetic silencing of SQOR increased the gating potential of the mPTP in the heart, causing the pore to open at higher ΔΨs.

4 DISCUSSION

Closure of the mPTP during development is thought to trigger cellular maturation. In developing cardiac mitochondria, the mPTP is initially open and then closes as the heart matures beyond the postnatal period.⁸ It is curious then that Fmr1 KO mouse cardiac mitochondria harbor a relatively closed mPTP in the newborn heart. Although premature closure of the mPTP in this murine model is not associated with any specific cardiac phenotype, Fmr1 mutant cardiomyocyte mitochondria are characterized by decreased proton leak along with excess CoQ levels. CoQ analogs are known to modulate the mPTP, yet the mechanisms by which they exert such biological activity are poorly understood. Thus, we exploited the Fmr1 KO model system in the current work in an attempt to elucidate novel regulatory mechanism(s) responsible for such aberrant mPTP characteristics.

We found that the Fmr1 KO cardiac mPTP demonstrated a marked desensitization to calcium and a significant shift in its gating potential to lower membrane potentials. Thus, altered voltage sensing was likely responsible for the decreased open probability of the mPTP in Fmr1 KO cardiomyocyte mitochondria. To assess the potential contribution of SQOR to voltage sensing, we confirmed that SQOR expression and activity were, indeed, significantly increased in Fmr1 KO cardiomyocyte mitochondria, suggesting a role in mPTP regulation. However, since we also detected elevated H₂S levels in Fmr1 KO cardiac ventricle tissue homogenates, it remains unclear whether inhibition of the mPTP in Fmr1 KOs is driven by H₂S independently or in combination with elevation in SQOR. The increase in H₂S in Fmr1 KOs could be due to upregulation of H₂S synthetic enzymes: cystathionine β-synthase (CBS) and 3-mercaptopyruvate sulfur transferase (3-MST) (given that each is a known target of the translational repressor, FMRP [genetically silenced in the Fmr1 KOs]).³⁰ Furthermore, it is possible that the elevation in SQOR level and activity in this strain results as a consequence of increased H₂S. Future work will need to determine the cause of H₂S elevation and assess whether the effects of H₂S on the pore are dependent on SQOR, since we did not directly assess the effect of H₂S on the mPTP in this work. Despite these uncertainties, we were able to demonstrate that both inhibition and genetic silencing of SQOR induced mPTP opening in wild-type mouse heart mitochondria in vitro and in situ. In these scenarios, loss of SQOR function did not alter H₂S levels; thus, the activation of the mPTP observed appeared to be independent of H₂S. Taken together, these findings provide evidence that SQOR contributes to the regulation of the mPTP.

SQOR is a flavoenzyme anchored to the inner mitochondrial membrane that catalyzes the first step in the mitochondrial sulfide oxidation pathway. Specifically, SQOR couples H₂S oxidation with CoQ reduction and likely glutathione oxidation, relying on two redox active cysteine residues (Cys201 and Cys379) within its active site.^{17, 24} Changes in the redox state of vicinal thiol groups have been proposed to alter voltage gating of the mPTP and tune the voltage sensor. However, the proteinaceous identity of the sensor is unknown. Although we did not specifically assess for changes in thiol group redox state as part of this work, our findings suggest that SQOR may contribute to the voltage sensing of the mPTP. Moreover, SQOR may be a key enzyme that links two established regulators of mPTP activity: CoQ and H₂S. The notion that SQOR contributes to pore regulation is supported by the fact that H₂S inhibits calcium-induced mPTP opening in a variety of tissues and that loss of SQOR expression sensitizes murine brain, heart, and liver to hypoxic injury in a manner highly reminiscent of pathological mPTP opening.^{18-20, 31-34} In addition, genetic silencing of SQOR causes severe abnormalities in growth and development and results in early death.²⁰ Future work, however, will need to determine if the effects of knocking out SQOR result from a failure of the mPTP to close during maturation.

Our identification of a novel role for SQOR carries significance. This is because the exact molecular identity of the pore remains ill-defined. Therefore, identifying a new proteinaceous component or potential regulator of the mPTP opens the door for further investigation and insight. Our findings may help to better resolve the pore structure, advance our knowledge of the regulatory mechanisms, and gain an understanding of how the mPTP becomes dysregulated in disease states. Importantly, SQOR may prove to be a valuable therapeutic drug target in the future, whereby enhancing its activity or expression could help to combat or prevent the myriad of diseases caused by pathological mPTP opening.

Although our data demonstrate that SQOR contributes to modulation of the mPTP by altering its voltage threshold for opening, the findings raise several unanswered questions. First, it is unknown how SQOR modulates mPTP gating potential. It is possible that the mechanism relates to utilization of a disulfide bridge in its redox center; however, more investigation will be necessary to resolve this. Future electrophysiological studies employing direct channel recordings will be required to further define the role of SQOR in mPTP voltage gating. In addition, further studies are needed to specifically define how loss of fragile X mental retardation protein (FMRP) changes H₂S production and increases SQOR levels and activity. Increased expression of SQOR could be a response driven by elevated H₂S in Fmr1 KOs, but there could be other explanations as well. In support of this concept, prior studies have shown that exogenous supplementation of CoQ can induce SQOR expression. Thus, it is possible that elevated CoQ levels within the newborn Fmr1 KO mouse heart also drive overexpression of SQOR. Furthermore, since SQOR binds CoQ, utilizing it as an electron acceptor, it is intriguing to consider that the previously described effects of CoQ on mPTP are mediated by SQOR. Moreover, because SQOR is a redox active protein, changes in SQOR activity could influence the redox status of CoQ and GSH. Oxidative stress does influence mPTP open probability, and while we did not measure the redox status of CoQ or GSH and its influence on the mPTP in this study, future work will need to address whether SQOR plays a role in integrating this signal as well. One further question to ponder is why the mPTP was insensitive to CsA in Sqor KO mitochondria. Somehow, lack of SQOR interfered with the ability of CsA to interact with its binding target, CypD, or altered the expression of CypD. Prior work found upregulation of SQOR in the CypD KO mouse heart. So, there may be some relationship between SQOR and CypD, and it is possible that genetic silencing of SQOR resulted in downregulation of CypD. Obviously, future work will need to explore these questions and hypotheses further.

In summary, we found decreased gating potential of the mPTP and increased expression and activity of SQOR in a model system of CoQ excess and relatively decreased mPTP open probability. We further found that pharmacological inhibition and genetic silencing of SQOR increased mPTP open probability in adult murine cardiac mitochondria, likely by interfering with voltage sensing. Thus, SQOR contributes to regulation of the mPTP and may be a component of the voltage sensing apparatus that modulates the gating potential of the mPTP.

AUTHOR CONTRIBUTIONS

K.G. generated the hypothesis, designed and conducted the experiments, analyzed the data, performed statistical analysis, interpreted the results, wrote the manuscript. A.W. conducted the experiments. E.J. generated the hypothesis, edited the manuscript. R.L. generated the hypothesis, designed the experiments, analyzed the data, performed statistical analysis, interpreted the results, and wrote the manuscript.

FUNDING INFORMATION

Supported by NIH/NIGMS T32GM008464 (KKG), Columbia University Irving Medical Center Target of Opportunity Provost award to the Department of Anesthesiology (KKG), Society of Pediatric Anesthesia Young Investigator Research Award (KKG), and NIH/NINDS R01NS112706 (EAJ and RJL).

DISCLOSURES

The authors have no conflicts of interest to declare.

Open Research

DATA AVAILABILITY STATEMENT

All relevant data within the paper and in Supporting information are available upon request from corresponding author.