Clinical utility of the ATP hydrolysis assay for the diagnosis of complex V deficiency in cultured skin fibroblasts

Abstract

Complex V catalyzes the formation of ATP from ADP and P_i through the dissipation of the proton gradient generated during the process of oxidative phosphorylation. Most complex V genetic disorders are caused by missense mutations in the mtDNA-encoded subunits, a (MT-ATP6) and A6L (MT-ATP8). Nuclear DNA-encoded gene mutations are increasingly recognized as causes of complex V defects and exhibited both autosomal recessive and autosomal dominant inheritance. Most identified variants are novel and confirmation by functional assays is important especially for variants demonstrating autosomal dominant inheritance. A kinetic spectrophotometric assay of the Complex V enzymatic hydrolysis activity has been reported. Here we report the clinical utility of this assay for the diagnosis of complex V deficiency after optimization and validation for the diagnosis of isolated complex V disorders due to both nuclear and mitochondrial DNA encoded variants and also for use in combined respiratory chain deficiencies. This assay was able to identify all nuclear DNA-encoded complex V deficiencies, whereas a decrease in complex V activity was observed in some patient cell lines with combined deficiencies. However, this assay only identified 50% of the mitochondrial DNA-encoded complex V deficiencies due to pathogenic variants in MT-ATP6/8. In conclusion, the enzymatic assay of complex V has best clinical utility for nuclear DNA-encoded complex V defects.

Introduction

Complex V catalyzes the formation of ATP from ADP and P_i through the dissipation of the proton gradient generated across the inner mitochondrial membrane established during the process of cellular respiration. The enzymatic activity is reversible allowing for hydrolysis of ATP to ADP and inorganic phosphate (Pi). Complex V consists of 16 structural subunits of which two are mitochondrial DNA-encoded (MT-ATP6 and MT-ATP8) and the remaining are nuclear DNA-encoded [1–3] (Figure 1). It consists of two functional domains designated F1 which protrudes into the matrix space and F0 which resides in the inner mitochondrial membrane. The enzymatic activity of the complex resides in the F1 domain in the catalytic lobular head formed by three repeating α and ß subunits (genes ATP5F1A and ATP5F1B). The F1 and F0 domains are connected by a central stalk composed of the γ, δ, and ε subunits to the c-ring located in the F0 domain within the inner mitochondrial membrane. The central arm interacts with the c-ring causing it to rotate as the protons pass through the pore located in the a-protein. Attached to the side of the catalytic lobular head is the OSCP protein (gene ATP5PO) and the lateral arm. The lateral arm extends into the inner mitochondrial membrane where it is attached to the a and A6L proteins, which are mitochondrial DNA (mtDNA) encoded.

Figure 1: Complex V hydrolysis assay.

(A) Complex V schematic showing the locations of the variant subunits analyzed in this study. (B) ATP hydrolysis enzyme assay in which ATP is converted to ADP and P_i by ATP synthase that is coupled via pyruvate kinase and lactate dehydrogenase to the oxidation of NADH to NAD⁺. (C) Spectrophotometric tracings illustrating the total ATPase activity and the oligomycin insensitive ATPase activity. The oligomycin-sensitive ATPase activity is calculated from subtracting the oligomycin-insensitive ATPase activity from the total ATPase activity.

Most complex V genetic disorders are caused by missense mutations in the mtDNA-encoded subunits, a and A6L (MT-ATP6 and MT-ATP8, respectively) [1–2,4,5]. These disorders can present with either Leigh disease or with Neuropathy, Ataxia, and Retinitis Pigmentosa (NARP) syndrome. Nuclear DNA-encoded gene mutations are now frequently recognized as causes of complex V defects and can be both autosomal recessive inheritance and autosomal dominant inheritance. Many of these variants are novel and particularly for autosomal dominant inheritance, confirmation by functional assays is important. Clinical presentations vary and can include developmental delays, cognitive dysfunction, epilepsy, motor problems such as spasticity and dystonia, cardiomyopathy, cataracts, Leigh syndrome, lactic acidosis, elevated 3-methylglutaconic acid, and hyperammonemia.

A kinetic spectrophotometric enzyme linked ATP hydrolysis assay for use in the diagnosis of isolated complex V disorders has been reported [6–8]. The use of oligomycin, a specific inhibitor of complex V, allows the isolation of the hydrolytic activity from other ATPase within the cell. Therefore, the oligomycin-sensitive ATPase activity is calculated from subtracting the oligomycin-insensitive ATPase activity from the total ATPase activity. In this study, we analyze the clinical utility of this assay to diagnose both nuclear DNA- and mitochondrial DNA-encoded variants of Complex V disorders. It was further expanded to include disorders of intramitochondrial protein synthesis due to disorders of mitochondrial DNA maintenance, transcription, RNA processing and translation on the mitoribosomes, in which the identification of reduced complex V activity can facilitate the recognition of a combined defect.

Methods

Complex V activity was determined by a kinetic spectrophotometric assay that measures the oligomycin-sensitive ATP synthase hydrolytic activity as described with some modifications [6,7]. In this assay, ATP is converted to ADP and P_i by ATP synthase that is coupled via pyruvate kinase and lactate dehydrogenase to the oxidation of NADH to NAD⁺ (Fig 1B). Fibroblasts were grown in Dulbecco minimal essential medium alpha modification (α-MEM) supplemented with 10%Hyclone fetal clone III serum replacement, 1X non-essential amino acids and 1X penicillin and streptomycin antibiotic/ amphotericin antimycotic solution. All cell lines used in this analysis had previously been tested for mycoplasma (Lonza MycoAlert^™ Mycoplasma Detection Assay Kit) and were found to be free of mycoplasma contamination. Cells were harvested at 80% confluency from one T150 cm² flask following treatment with trypsin, washed three times with 1X Dulbecco phosphate buffered saline (DPBS), and centrifuged at 1000 x g for 10 min after each wash to pellet cells. Cells were assayed immediately after harvesting.

The cell pellet was resuspended in 1000 μL MegaFB buffer (250 mM sucrose, 2 mM HEPES, 0.1mM EGTA, pH 7.4) and transferred to a glass homogenizer. The fibroblasts were homogenized using a Teflon pestle and approximately 30 strokes. The resulting homogenate was transferred to a microfuge tube and centrifuged at 1000 x g for 5 min. The supernatant was reserved; the pellet was resuspended in 800 μL MegaFB buffer and homogenized and centrifuged as previously described. The supernatants were combined and centrifuged at 10,000 x g for 10 min to pellet the mitochondria. The resulting pellet was then resuspended in 100 μL MegaFB buffer and used directly for the complex V ATP hydrolysis enzyme assay. Protein was assayed using the Bio-Rad protein assay (Hercules, CA).

For the ATPase reaction, 50 μL 10X assay buffer (400 mmol/L Tris-HCl, pH 8.0, 50 mmol/L MgCl₂, 100 mmol/L KCL, 40 U/mL lactate dehydrogenase, 40 U/mL pyruvate kinase, 2 mmol/L NADH, 20 mmol/L phosphoenolpyruvate (PEP), 5 mmol/L ATP, 1% bovine serum albumin (BSA), 10 μmol/L antimycin A, 3 mM diadenosine-pentaphosphate (Ap5A), and 30 μmol/L carbonyl cyanide 4-(trifluoromethoxy)phenylhydrazone (FCCP), mitochondrial homogenate equivalent to 5 μg protein, and water pre-warmed at 30°C to complete the assay volume to 500 μL were combined. After mixing, the assay reaction mixture was then sonified for 10 sec using a Branson Sonifier 550 at 20% output and immediately placed in a Cary 4000 UV/VIS spectrophotometer to measure the ATP hydrolysis activity. The oxidation of NADH was followed at 340 nm for 5 min at 37°C, after which oligomycin was added and the assay was followed for an additional 5 min to determine the oligomycin-insensitive ATPase activity. The oligomycin-sensitive ATPase activity is calculated from subtracting the oligomycin-insensitive ATPase activity from the total ATPase activity and reflects the complex V hydrolytic activity. Suppliers and catalogue numbers for both tissue culture and the complex V hydrolysis assay are provided in Suppl Table 1.

Citrate synthase activity was then assayed in a subset of controls and patients as detailed previously [9] using complex V homogenates. Furthermore, the effect of freezing cell pellets for up to 24 hours before processing on the hydrolytic activity of complex V was examined to assess enzyme stability, and the sonication step was optimized.

Cells

Cells were studied deidentified under an IRB-approved study protocol (COMIRB# 18-1828). All cell lines have known pathogenic variants. Thirty-four cell lines with genetically confirmed primary mitochondrial disease were used for the evaluation of this assay (Suppl Table2) This cohort included six cell lines with isolated complex V deficiency due to mutations in the nuclear DNA-encoded genes. The pathogenicity of the variants of the nuclear defects had been documented in previous publications [10–12]. There were ten cell lines with isolated complex V deficiency due to mutations in the mitochondrial DNA encoded gene MT-ATP6. Fifteen cell lines with defects in mitochondrial translation were similarly investigated, and as a control three cell lines with isolated deficiencies of complexes I, II, and IV.

Statistical presentation

Normality of distribution was first evaluated using Shapiro-Wilk and Kolmogorov-Smirnov statistics. Data were presented as mean and standard deviation if normally distributed, and as median and interquartiles if not normally distributed. A log transformation was used for specific distributions to better adjust to the normal distribution. Comparisons between two populations were done by Student t-test or by Mann-Whitney-U statistic if normally distributed or not, respectively. Relationships between 2 variables were calculated as the Pearson correlation coefficient for normally distributed populations or the Spearman rank correlation for non-normally distributed populations.

Results

Validation

The enzyme activity displayed linearity from 5 to 50 μg protein (R² = 0.98 and 0.99 for two different controls, respectively). For sonication, the mean oligomycin-sensitive enzyme activity using 20% amplitude setting (N=30) was 38% higher than using 10% amplitude setting (N=10). Similarly, the % oligomycin inhibition was 10.3% higher using 20% amplitude setting compared to using 10% amplitude setting, indicating greater specificity of the assay for complex V (Suppl Table 2). Enzyme stability was affected by freezing, resulting in a reduction of 18 to 60% less activity in a normal control cell line repeated twice and four patient cell lines (Suppl Table 4). The final assay had good accuracy with the within run precision for three normal control cell lines showed a COV or 13%, 12.6% and 8.6% and the between run precision showed a COV of 20.8%, 21.3% and 14.2%.

The control range for the complex V hydrolysis assay was established from 30 fibroblast cell lines as 205.85 to 440.56 nmol.min⁻¹.mg protein⁻¹ (Table 1), which corresponds to 65% to 140% of the mean of controls, and did not significantly differ from the normal distribution. In all controls, the oligomycin-sensitive ATPase activity was greater than 70% of total activity. The natural log (ln) of the activity of the controls was also normally distributed and provided the best discrimination between affected and control fibroblast cell lines. Using this measure, the Z-scores of the normal fibroblast population ranged between −1.89 and +1.66 SD, and patient values could be expressed as Z-scores [13]. The ratio of complex V enzyme activity over citrate synthase enzyme activity introduced greater variability and provided less discrimination between controls and the target disease population and was not further employed (Suppl Tables 5 and 6).

Table 1:

Reference Range and Control Data

Controls n=30	OMY-Sensitive ATPase Activity	% OMY inhibition	% of control	Z-score of Ln(activity)
MEAN	315.61	82.82	100%	0.00
MIN	205.85	72.50	65%	−1.89
MAX	440.56	89.73	140%	1.66
P5	210.99	76.46	67%	−1.78
P95	432.10	87.16	137%	1.57
SD	66.43	3.72
COV	21.05%	4.49%

The reference range of the Complex V enzyme activity expressed in nmol.min⁻¹.mg protein⁻¹ from 30 controls is provided with the oligomycin-sensitive activity, the relative percentage of oligomycin sensitive portion of the total activity. The resulted are also expressed as the percentage of the mean of the controls and as the Z-score of the natural log transformed control data. Abbreviations used: OMY: oligomycin; COV: coefficient of variance; P5: 5^th percentile; P95: 95^th percentile; SD standard deviation.

Patient cell lines: clinical utility

All nuclear encoded complex V deficiencies (N=6) showed a marked decrease in complex V activity (14% to 42% of mean of controls, or Z-score of −9.18 SD to −3.98 SD). The three autosomal dominant ATP5F1A cases had higher residual activity (33-42% of mean of controls, or Z-score of −5.05 to −3.98) compared to the autosomal recessive cases (14-27% of mean of controls, Z-score of −9.18 to −6.01) (Table 2). Defects were similarly identified on replicate analysis.

Table 2.

Oligomycin-sensitive ATPase activity in fibroblast cell lines with isolated complex V deficiencies

Nuclear DNA Complex V Deficiencies
Cell Line	Phenotype	OMY-Sensitive ATPase Activity	% OMY inhibition	% of control	Z-score of ln(activity)
5059	TMEM70	85.22	60.17	27%	−6.01
5059	TMEM70	59.46	64.06	19%	−7.69
1123	ATP5F1A	131.58	81.82	42%	−3.98
1179	ATP5F1A	116.75	72.48	37%	−4.54
1345	ATP5F1A	104.64	66.36	33%	−5.05
1018	ATP5PD	73.74	63.00	23%	−6.69
1100	ATP5PO	43.26	44.17	14%	−9.18
1100	ATP5PO	53.75	57.73	17%	−8.16
Mitochondrial DNA Complex V Deficiencies
5268	MT-ATP6: m.9176C>T	240.14	80.35	76%	−1.17
5268	MT-ATP6: m.9176C>T	151.43	70.89	48%	−3.33
1019	MT-ATP6: m.9032T>C	170.68	75.03	54%	−2.77
1062	MT-ATP6: m.8969G>A	187.05	76.21	59%	−2.34
1380	MT-ATP6: m.8579C>T	226.82	85.29	72%	−1.44
5344	MT-ATP6: m.9035T>C	333.68	83.42	106%	0.36
5357	MT-ATP6: m.8993T>C	383.17	85.80	121%	1.01
7004	MT-ATP6: m.8993T>C	302.36	86.34	96%	−0.10
1099	MT-ATP6: m.8993T>C	166.44	75.32	53%	−2.88
1087	MT-ATP6: m.8993T>G	275.83	87.00	87%	−0.53
5258	MT-ATP6: m.8993T>G	206.42	75.87	65%	−1.88
5258	MT-ATP6: m.8993T>G	209.75	75.32	66%	−1.80

The results of the Complex V enzyme activity expressed in nmol.min⁻¹.mg protein⁻¹ from 15 patients with known genetic defects in complex V is provided with the oligomycin-sensitive activity, the relative percentage of oligomycin sensitive portion of the total activity, the percentage of the mean of the controls, and as the Z-score of the natural log transformed control data. Abbreviations used: OMY: oligomycin, Complex V activity expressed in nmol/min.mg protein

Cell lines with isolated complex V deficiency due to pathogenic variants in the mitochondrial DNA encoded MT-ATP6 gene had borderline activities. The activities observed were between − 1.5 and −3.5 SD in 50% of cases, with only 4 cases below the normal range (Table 2). Moreover, this decrease did not always replicate consistently in this diagnostic group (e.g., m.9176C>T showed activities of 76% and 48% of mean controls on repeat analysis).

Disorders of the intramitochondrial protein synthesis can have combined deficiencies of respiratory chain complexes that contain mitochondrial DNA-encoded subunits including as complex V. Indeed, in a series of 15 such cell lines, deficient Complex V ATPase activity was found in 33% of cases (Table 3). Most decreases were mild, Z-score between −2.7 and −2 SD (55-62% of mean of controls), and the degree of decrease in complex V activity did not correlate with the degree of decrease in other enzyme complexes. Interestingly, in contrast, a very pronounced decrease was observed in two patients with pathogenic variants in the gene C12orf65 (Z-scores of −5.22 SD and −6.43 SD, 25% and 32% of mean of controls). As expected, the activity of complex V was normal in the cell lines with isolated deficiencies of complexes I, II, and IV.

Table 3.

Oligomycin-sensitive ATPase activity in fibroblast cell lines with combined OXPHOS and isolated deficiencies

Combined OXPHOS Deficiencies
Cell Line	Phenotype	OMY Sensitive ATPase Activity	% OMY inhibition	% of control	Z-score ln(activity)
• aminoacylation translational defects
1352	NARS2	180.05	79.96	57%	−2.52
7100	CARS2	310.84	88.92	98%	0.03
1355	TARS2	194.42	82.16	62%	−2.16
1351	FARS2	349.80	86.03	111%	0.58
5364	WARS2	366.65	83.27	116%	0.80
• mitochondrial RNA processing
5363	MTO1	375.47	90.53	119%	0.91
• mitochondrial DNA maintenance
5237	POLG1	362.95	87.58	115%	0.76
• mitochondrial DNA
5362	MT-TK	382.46	88.23	121%	1.00
1199	MT-TY	286.60	88.14	91%	−0.35
• other translational defects
1343	TUFM	174.06	77.30	55%	−2.68
1110	C12orf65	101.00	66.01	32%	−5.22
1376	C12orf65	77.87	66.67	25%	−6.43
1217	MRPS22	276.73	83.23	88%	−0.51
5360	MRPS18C	409.04	89.75	130%	1.31
1311	MTFMT	326.56	85.05	103%	0.26
Isolated Complex Deficiency
5232	SURF1	393.26	85.77	125%	1.13
7009	SDHA	270.99	77.23	86%	−0.61
1309	NDUFAF6	353.97	87.15	112%	0.64

The results of the Complex V enzyme activity expressed in nmol.min⁻¹.mg protein⁻¹ from patients with known genetic defects in genes other than complex V subunits are provided with the oligomycin-sensitive activity, the relative percentage of oligomycin sensitive portion of the total activity, the percentage of the mean of the controls, and as the Z-score of the natural log transformed control data. OMY: oligomycin, Complex V activity expressed in nmol/min.mg protein

Discussion

This assay demonstrated excellent clinical utility in the identification of patients with nuclear DNA-encoded complex V gene defects and proved to be a powerful tool for the validation of variants of unknown significance in nuclear genes including those defects inherited as autosomal dominant. However, it had limited clinical utility to evaluate MT-ATP6 variants. There are variable reports in the literature about the functionality in MT-ATP6 defects. Some identified more consistent low complex V enzyme activity in post-600 g supernatants of muscle homogenate [7,8], with less reduction in fibroblasts [7]. Other reports summarized a much more variable reduction in ATP hydrolysis assay with normal results in 19/22 m.8993T>G and 6/6 m.8993T>C, and 1/1 m.9101T>C, but decreased ATP hydrolysis results in one patient each with m.8969G>A, m.8618ins, m.9035T>C, m.9134A>G, and m.9185T>C [4].

This distinction in the clinical utility of this assay is most likely due to the mechanistic nature of the defect in complex V and rooted in the biogenesis of this complex. Complex V first assembles the catalytic head structure through oligomerization of the three α-subunits and the three β-subunits helped by the assembly factors ATPAF1 and ATPAF2, respectively. These then assemble to the central stalk and the peripheral stalk, with the assembly of the c-ring aided by TMEM70. This V* subcomplex is stable and has catalytic activity, without needing the late assembling subunits such as MT-ATP6 [1,5,14]. As a result, pathogenic variants in subunits of the V* only variably affect the stability of this intermediate where the ATPase activity resides. The gene ATP5F1A encodes for a protein that is required for the stability of the F1 module and is involved in the early stages of complex V assembly. Both ATP5PD and ATP5PO genes encode proteins that are part of the peripheral stalk that is linked to the F1 subcomplex [5,15]. The gene TMEM70 for examples encodes for an assembly factor which is required in the early stages of the formation of complex V. Variants in this gene have been shown to result in complex V assembly defects.

Disorders of intramitochondrial protein synthesis cause a combined protein defect that can affect all the complexes that include mitochondrial DNA encoded subunits and theoretically should have a combined deficiency of the activities of all these complexes. Experience shows that this is rather variable, and often only includes deficient activity of a single complex. Deficient activity in complex V helps to identify that a combined deficiency is present rather than an isolated complex activity. Like other complexes, this is not consistently present, estimated at one third of cases, and like MT-ATP6 variants, the decrease was mild. It is not clear why the decrease in C19Orf65 is an exception with a much more marked decrease in the complex V activity, more in line with nuclear DNA-encoded complex V defects. A defect in complex V assembly and activity with near absent complex V holocomplex has been reported before [16–19], which reflects the impact on this assay. The Mtrfr protein is associated with the release of the nascent peptide chain from the mitoribosome and the quality control system for stalled nascent peptide chains [20]. The mechanism of this profound impact on complex V assembly is not immediately clear and appears to exceed that of reduced a-protein synthesis.

In contrast, the results of the variants in MT-ATP6 are variable. Decreased synthesis and thus abundance of the a-protein encoded by the MT-ATP6 gene may result in complex V instability resulting in reduced hydrolytic activity, but this is variable. Examples where this effect was evident was found in two aminoacyl synthetase disorders NARS2 and TARS2 and was profound in two patients with mutations in the C12orf65 (MTRFR) gene. Mutations that primarily affect proton translocation or the coupling of proton flow to the rotation of the rotor [4] would affect ATP synthesis or the transmembrane protein gradient and membrane potential, and not necessary the hydrolytic activity, as the stability of the enzyme is not affected. Both the a-protein and A6L protein are incorporated late in the assembly of this complex and form part of the proton channel that allows the passage of protons across the inner mitochondrial membrane. The dissipation of this proton gradient is then coupled to ATP synthesis. In this study, the hydrolytic activity was reduced in only some ATP6 cell lines. This assay therefore lacks the required sensitivity to consistently confirm variants of unknown significance in the ATP6 gene as non-pathogenic. Because the occurrence of variants in ATP8 are rare compared to ATP6 variants, this study did not confirm the clinical utility of this assay for the diagnosis of complex V deficiencies due to mutations in the ATP8 gene.

In conclusion, this assay is intended to primarily identify nuclear DNA-encoded defects in complex V and to functionally validate VUS in these genes including for autosomal dominant inheritance. For this role, this assay has excellent clinical utility. However, this assay has limited sensitivity to identify defects in MT-ATP6, and in some it may confirm decreased activity, However, it should not be used to exclude pathogenicity of a variant of unknown significance, which is a limitation of the utility of this assay. Evaluation of other functional assays of the a-protein and the impact of the variants identified in the MT-ATP6 gene to be used for validation of its pathogenicity are still under study, with no universally applicable assay yet identified [4]. The recently published assay in skeletal muscle for a similar complex V hydrolytic assay shows promising data for MT-ATP6 variants, but needs further validation [21].