REVIEW article

Front. Mol. Neurosci., 09 December 2021

Sec. Brain Disease Mechanisms

Volume 14 - 2021 | https://doi.org/10.3389/fnmol.2021.797833

Mitochondrial Function and Parkinson’s Disease: From the Perspective of the Electron Transport Chain

  • 1. Department of Neurology, National Taiwan University Hospital, Taipei, Taiwan

  • 2. Division of Neurology, Department of Internal Medicine, Lo-Hsu Medical Foundation, Lotung Poh-Ai Hospital, Yilan County, Taiwan

  • 3. College of Medicine, National Taiwan University, Taipei, Taiwan

  • 4. National Institute of Infectious Diseases and Vaccinology, National Health Research Institutes, Miaoli County, Taiwan

  • 5. Graduate Institute of Physiology, National Taiwan University, Taipei, Taiwan

  • 6. Department of Medical Research, National Taiwan University Hospital, Taipei, Taiwan

Abstract

Parkinson’s disease (PD) is known as a mitochondrial disease. Some even regarded it specifically as a disorder of the complex I of the electron transport chain (ETC). The ETC is fundamental for mitochondrial energy production which is essential for neuronal health. In the past two decades, more than 20 PD-associated genes have been identified. Some are directly involved in mitochondrial functions, such as PRKN, PINK1, and DJ-1. While other PD-associate genes, such as LRRK2, SNCA, and GBA1, regulate lysosomal functions, lipid metabolism, or protein aggregation, some have been shown to indirectly affect the electron transport chain. The recent identification of CHCHD2 and UQCRC1 that are critical for functions of complex IV and complex III, respectively, provide direct evidence that PD is more than just a complex I disorder. Like UQCRC1 in preventing cytochrome c from release, functions of ETC proteins beyond oxidative phosphorylation might also contribute to the pathogenesis of PD.

Introduction

Parkinson’s disease (PD) is the second most common neurodegenerative disease worldwide with 6.1 million patients globally in 2016 (), and up to 12.9 million people estimated to be affected by 2040 (). PD is clinically characterized by slowly progressive, levodopa-responsive bradykinesia with either rigidity or resting tremor (). Dopaminergic neuronal loss and Lewy body formation in the substantia nigra pars compacta are the hallmarks of most PD pathology.

Etiologies of PD are complex. Most individuals with PD are idiopathic. Only 10% of them showed a clear Mendelian inheritance (), and environmental factors such as pesticides, heavy metals, illicit substances, and diets also contribute (). In 1983 four cases of levodopa-responsive parkinsonism were described after intravenous injection of synthetic heroin containing 1-Methyl-4-phenyl-1,2,3,6-tetrahydropyridine (MPTP) (). The neurotoxicity of MPTP results from its oxidized metabolite, 1-methyl-4-phenylpyridine (MPP+) (; ), which damages oxidative phosphorylation (OXPHOS) by inhibiting the electron transport chain (ETC) of mitochondria (). The keystone discovery initiated an era of mitochondrial pathology in PD. Nowadays, a plethora of evidence from electrophysiological and anatomical perspectives has shown that mitochondrial health is essential in the integrity of dopaminergic neurons, especially those in SNc (; ). Therefore, to disentangle the complex pathophysiology of PD, mitochondria is undoubtedly one of the key players that should not be missed.

In this review, we will briefly introduce current knowledge about ETC and OXPHOS and their relationships with PD. We will then summarize some of the well-known and newly identified PD-associated genes and their direct or indirect influences on ETC. Finally, we discuss the OXPHOS-independent functions of ETC proteins and their possible implications in PD pathogenesis.

Electron Transport Chain and Parkinson’s Disease

Mitochondria are the powerhouses of eukaryotic cells. This double membrane-bound organelle generates most adenosine triphosphate (ATP) through OXPHOS, processed by ETC embedded in the inner mitochondrial membrane (IMM). ETC is composed of transmembrane complexes I (cI) to V (cV) and two electron carriers, the ubiquinone (i.e., CoQ) and the cytochrome c (cyt c). For ATP production, electrons are transferred from NADH and FADH2 to oxygen via the transport chain, coupled with the generation of a proton gradient across IMM (Zhao et al., 2019; Figure 1).

FIGURE 1

Complex I (cI) is an NADH-ubiquinone oxidoreductase that pumps four protons into the intermembrane space (IMS) upon each NADH oxidation. cI consists of three modules: the N, Q, and P modules. NADH generated from the tricarboxylic acid cycle delivers its electron to the N module. The electron was then passed to the Q module where the CoQ is reduced to ubiquinol (CoQH2) and induces conformational changes of the P module to allow proton translocation (). Similarly, cII also transfers electrons to CoQ, but by dehydrogenizing succinate to fumarate, which is a part of the tricarboxylic acid cycle. cII is composed of two subunits: the enzymatic subunit (SDHA-SDHB) in the matrix and the anchoring subunit (SDHC-SDHD) across the IMM. The enzymatic subunit exploits two electrons from succinate and passes them through a series of FeS clusters to the anchoring subunit where CoQ is reduced to CoQH2. Different from cI, no proton is translocated for the reactions that occur in cII (). Electrons from both cI and cII are carried by CoQH2 to enter the Q cycle processed by cIII. The catalytic activity of cIII depends on three redox-active subunits: cytochrome b (MT-CYB), the Rieske iron-sulfur [Fe2-S2] protein (UQCRFS1), and cytochrome c1 (CYC1) (Yang and Trumpower, 1986). Electrons donated by CoQH2 were transduced to cyt c via these three subunits sequentially. Each Q cycle generates two reduced cyt c and transports two protons (). The reduced cyt c then shuttles electrons to cIV, namely the cyt c oxidase (COX), whose catalytic core includes subunits 1, 2, and 3. Electrons are exploited from reduced cyt c at subunit 2 and reduce molecular oxygen to water at subunit 1. One proton is simultaneously translocated to IMS via subunit 3 upon each cyt c oxidization (Wilson and Prochaska, 1990). The oxidized cyt c then returns to its pool. In addition to working as an electron carrier, the oxidized cyt c also serves as a scavenger of reactive oxygen species (ROS) (Wang et al., 2003), a cardiolipin peroxidase (), and an apoptosis activator once released to the cytosol (). The functions of cyt c beyond OXPHOS are discussed in the latter part of this review.

As the final step of OXPHOS, the proton gradient established by cI, cIII, and cIV drives cV to generate ATP. cV consists of two functional domains: F1 and F0. Protons passing through the F0 domain across the IMM to the matrix release the energy provided by the proton electrochemical gradient. The F1 domain at the matrix then uses the energy to charge cells by phosphorylating ADP to ATP (). Each ATP synthesis requires the translocation of 3 or 4 protons (; ).

Dysfunction of ETC has been associated with various neurological diseases from infantile-onset neurodevelopmental regression (such as Leigh syndrome) to adult-onset optic neuropathy or peripheral neuropathy (such as Leber hereditary optic neuropathy (LHON) in cI deficiency, and neurogenic muscle weakness, ataxia, and retinitis pigmentosa (NARP) in cV deficiency) (). Neurodegeneration diseases, especially PD, are also associated with ETC dysfunction. PD has been branded as a “cI disease” because of the discovery of MPP+ (the metabolites of MPTP) which is a cI inhibitor (; ). Rotenone, another commonly reported PD risk factor, is also a cI inhibitor (). In post-mortem studies of PD patients, cI functional deficiency is found in substantia nigra or frontal cortex (; ). Similar cI deficiency is also found in non-brain tissues including platelets, skeletal muscles, or fibroblasts, indicating it is a systematic phenomenon (; Wiedemann et al., 1999). Also, reports of cIII and cIV function in PD have been inconsistent (; Wiedemann et al., 1999). Impairment of ETC causes excessive ROS, disruption of the electrochemical potential of the proton gradient, insufficient ATP production, and even cell death. Although it is widely observed that ETC functions are compromised in PD, genetic mutations of ETC proteins (either nuclear-encoded or mitochondrial-encoded) are rarely linked to PD. The discrepancy might suggest that the ETC dysfunction is secondary to other mechanisms that closely regulate mitochondria. In other words, ETC represents a hub where different PD-causing patho-mechanisms converge.

PD-Associated Genes and the ETC

To date, over 20 PD-related genes have been discovered. Although these familial PD cases are not common, and some may even have atypical features such as early onset (<50 years of age at onset) or early cognitive or psychiatric involvement, they provide great insights into the pathogenesis of PD. Many of the PD-causing genes affecting quality control systems of mitochondria, lysosomal regulation, or lipid/protein homeostasis, thus directly or indirectly influence ETC functions. In the following section, we will discuss the best-known genes and some newly identified ones with a focus on their influences on ETC (Figure 2 and Tables 1, 2).

FIGURE 2

TABLE 1

GeneMechanismComplex IComplex IIComplex IIIComplex IVComplex VReferences
PRKN and PINK1Protein turnover+++++Vincow et al., 2013
Translation derepression+++
PINK1Phosphorylation+
PARK7Chaperone+++; ;
PLA2G6Ceramide metabolism++
GBAUnknown, probably regulating mitophagy, GlcCer metabolism, and NAD + production+++;
LRRK2Unknown, probably regulating mitophagy+++
SNCAMutant monomers or oligomers++; ;
Fibril-induced Lewy body formation+++
CHCHD2Chaperone+;
Transcription factor+
UQCRC1cIII core subunits+;

PD-associated genes and their influences on each ETC complex.

“ + ” indicates evidence of direct or indirect interactions but the exact involved subunits are still unclear.

TABLE 2

ComplexSubunits*FunctionsAssociated PD genes
Complex INDUFB8An accessory subunit of the P modulePRKN, PINK1
NDUFS2A core subunit of the Q modulePRKN, PINK1
NDUFB11An accessory subunit of the P modulePRKN, PINK1
NDUFA7An assembly subunitPRKN, PINK1
NDUFV2A core subunit of N modulePRKN, PINK1
NDUFA5An accessory subunit of Q modulePRKN, PINK1
NDUFA9An accessory subunit of Q modulePRKN, PINK1
NDUFA13An accessory subunit of P modulePRKN, PINK1
NDUFA8An accessory subunit of P modulePRKN, PINK1
NDUFB10An accessory subunit of P modulePRKN, PINK1
NDUFS3A core subunit of the Q modulePRKN, PINK1
NDUFS1A core subunit of the N modulePRKN, PINK1, PARK7
NDUFA10An accessory subunit of P modulePINK1
ND1An core subunit of P modulePARK7
Complex IISDHBAn enzymatic subunitPRKN, PINK1
Complex IIIUQCRBMaintaining Cytochrome b stability with UQCRQPRKN, PINK1
UQCRC2A core subunit of cIII; in situ processing of UQCRFS1 (probable)PRKN, PINK1
UQCRC1A core subunit of cIII; in situ processing of UQCRFS1 (probable)-
UQCRFS1A core subunit of cIII, responsible for transducing electrons from Cytochrome b to Cytochrome c-1UQCRC1
CYC1To reduce cytochrome cUQCRC1
UQCRQMaintaining Cytochrome b stability with UQCRBUQCRC1
UQCR10UnclearUQCRC1
UQCR11UnclearUQCRC1
Complex IVCOX5AResponse to hypoxia and regulate NO productionPRKN, PINK1
COX4I1 and-2Modulation of oxygen affinityPRKN, PINK1, CHCHD2
COX2To oxidize cytochrome cPRKN, PINK1
NDUFA4CIV maintenancePARK7
Complex VATP5L, 5L2A subunit of the F0 domainPRKN, PINK1
ATP5H, ATP5HL1A subunit of the F0 domainPRKN, PINK1
ATP5F1A subunit of the F0 domainPRKN, PINK1
ATP5POA subunit of the F0 domainPRKN, PINK1
ATP5F1BA subunit of the F1 domainPARK7, SNCA

Canonical OXPHOS functions of the mitochondrial complex subunits affected by PD-associated genes in Table 1.

*All designated by human gene names.

PINK1 and PRKN: Key Players of Mitochondria Quality Control

Mutations in PRKN (previously known as PARK2, which encodes parkin) and PINK1 are the most common causes of early onset PD. PRKN mutants on 6q25.2-27 were found in 1998 (), and PINK1 mutants on 1p35-1p36 were identified in 2001–2002 (, , ). Mutations in the two genes have since been reported worldwide (). Early onset PD associated with either gene showed similar clinical features, including slow disease progression, good levodopa response, lower limb dystonia, early psychiatric symptoms, and a higher likelihood of motor complications than idiopathic PD (). Pathologically, Lewy bodies can be absent in patients with either PRKN or PINK1 mutations (; ). These similarities reflect their common mechanism in PD pathogenesis.

It is now known that parkin, along with PINK1, is critical for mitochondria quality control through various pathways including mitophagy, formation of mitochondrial-derived vesicles, mitochondrial fission, and facilitation of mitochondrial biogenesis (). When mitochondria are jeopardized, PINK1 is recruited to and stabilized on the OMM. The stabilized PINK1 then recruits and activates parkin, an E3 ligase that assembles ubiquitin chains on its substrates including OMM proteins which subsequently recruits ubiquitin-binding autophagy receptors and ultimately leads to autophagosome formation and lysosomal degradation. Other substrates of parkin include Mitofusin-1 and -2, whose proteolysis divert mitochondria from fusion into fission (; Ziviani et al., 2010); parkin interacting substrate (PARIS, also known as ZNF746), whose degradation stimulates mitochondrial biogenesis (; ); and also hnRNP F, a translation repressor that regulates the localized synthesis of nuclear-encoded mitochondrial proteins (including some ETC components) (). In addition to PINK1-dependent phosphorylation, other post-translational modifications of parkin, such as S-nitrosylation, are also critical for the E3 ligase activity and solubility of parkin and have been comprehensively reviewed elsewhere ().

The integrity of ETC is highly correlated with the PINK1-Parkin-mediated quality control mechanisms. Fibroblasts or leukocytes with parkin loss-of-function mutations showed impaired cI activities (; ), and human fibroblasts or mouse embryonic fibroblasts of PINK1 mutants also showed cI dysfunction (; ). The turnover rates of mitochondrial proteins could be an indicator of mitochondrial quality control. A study using fly heads showed that ETC mitochondrial proteins had a greater dependence on parkin or Pink1 than other non-ETC mitochondrial proteins. Nineteen of the ETC proteins had longer half-lives in parkin mutants than in autophagy-impaired Atg7 mutants, suggesting the turnover of these proteins may be achieved via an a selective manner. More than half belong to cI (10/19), while proteins from the other complexes are also affected (1 in cII, 2 in cIII, 3 in cIV, and 3 in cV) (Vincow et al., 2013). The predominance of cI proteins likely explains why cI dysfunction is invariably reported in PRKN or PINK1 mutants.

Activities of cI could also be regulated by PINK1 independent of parkin. An accessory subunit of cI, Ndufa10, is pivotal for the ubiquinone-reducing capacity of cI, and the phosphorylation of Ndufa10 requires PINK1 (). Overexpression of Ndufa10 or expression of its phosphomimetics could reverse the defects in pinkB9-null mutant Drosophila, but not parkin mutants (; ). Ndufa10 phosphorylation also rescued the ROS-induced apoptosis in PINK1 knockout mouse fibroblasts (), suggesting that PINK1 is anti-apoptotic. The anti-apoptotic effects of PINK1 are also attributed to the phosphorylation of the mitochondrial chaperone TRAP1 (also known as Hsp75) which prevents cyt c release from mitochondria (). In tumor cells, TRAP1, phosphorylated by ERK1/2 and Src, prohibited the generation of ROS by inhibiting cII and cIV, respectively (). Whether the phosphorylation of TRAP1 by PINK1 has any influence on ETC functions awaits further investigation.

DJ-1: A Prominent Chaperone in Mitochondria

DJ-1 (also known as PARK7) mutations cause autosomal recessive, early onset PD. It was initially reported in 2003 from two genetically isolated consanguineous families (), and has a frequency of about 0.4% in early onset PD, much lower than that of PRKN and PINK1 ().

DJ-1 is ubiquitously expressed () and serves a dual role as an oxidative sensor and also an antioxidant protein. Immunostaining of oxidized DJ-1 (oxDJ-1) revealed reactivity in SNc, striatum, and inferior olivary nucleus in postmortem brain of elders with Lewy body pathology. Co-localization of oxDJ-1 and α-synuclein (α-Syn) was also reported (). The cysteine residue (C106) in its active site could be sequentially oxidized into the sulfenylated form (-SOH), sulfinated form (-SO2H), and sulfonic form (-SO3H) (). The sulfinated form is the active form that not only prevents α-Syn fibrillation (Zhou et al., 2006) but also plays protective roles in promoting mitochondria fission and cell viability to rotenone (). These results suggested that DJ-1 is an oxidative stress-dependent chaperone.

Since mitochondria are the major site of intracellular ROS production, it is intuitive that DJ-1 plays an important role in mitochondria physiology. The subcellular and submitochondrial fractionation of mouse brains shows that abundant endogenous DJ-1 is localized in mitochondria (Zhang et al., 2005). Depletion of DJ-1 significantly reduced the oxygen consumption rate, ATP production (), and cI activity (). Proteomics assays revealed improper cI assembly in DJ-1-deficient neuronal cells related to loss of CI-75kD (encoded by Ndufs1) (). Direct interactions of DJ-1 with ND1 and with NDUFA4 were also reported (). Of note, NDUFA4 is a subunit of cIV, rather than within cI as initially considered (). Whether cIV activities are influenced by DJ-1 remains unknown.

Recently, showed DJ-1 binds to the F1FO ATP synthase β subunit. ATPase β subunit is crucial for maintaining the mitochondrial membrane potential through inhibition of proton leakage from the pores formed by the c-subunit of cV, thereby enhancing the ATP synthesis (). The ATPase β subunit was decreased in DJ-1 knockout mice, thus the ATP production, mitochondrial membrane potential, and neurite outgrowth of dopaminergic neurons were all compromised (). The interactions of DJ-1 and ATP synthase β subunit again demonstrate its role as a mitochondrial chaperone.

PLA2G6: ETC Dysfunction and Lipid Imbalance

PLA2G6-associated neurodegeneration (PLAN) consists of a series of rare diseases: infantile neuroaxonal dystrophy (INAD), atypical neuroaxonal dystrophy (ANAD), dystonia-parkinsonism (DP), and autosomal recessive early onset parkinsonism (AREP). Among them, DP and AREP are both characterized by adult-onset as well as levodopa-responsiveness, and are not reported until 2009 and 2011, respectively (; ). Though previously under-reported, recent studies showed that PLA2G6 might contribute to early onset PD as frequently as 0.54–1.3%, more common than DJ-1 (; Zhao et al., 2020).

PLA2G6 encodes a group VIA calcium-independent phospholipase A2β enzyme (iPLA2β), which is responsible for the selective hydrolysis of glycerophospholipids to generate fatty acids and lysophospholipids (). In PLA2G6-knockout mice, granules containing collapsed mitochondria were observed in axons (). In iPLA2-VIA (the homolog of PLA2G6) null Drosophila, aberrant mitochondria with reduced respiratory chain activities of cI and cII or ATP production were demonstrated (; ). These changes, however, are unlikely associated with changes in cardiolipin composition (), which is the main component of mitochondria inner membrane. Lin et al. showed that none of the glycerophospholipids significantly changed in iPLA2-VIA null Drosophila. Rather, they showed a decrease in ceramide phosphoethanolamines, and an increase in ceramides, dihydroceramides, and other sphingolipid intermediates (). Some ceramides are associated with compromised respiratory chain activities, increased ROS production, impaired mitochondria membrane potential, mitophagy, and apoptosis (). By using liquid chromatography-tandem mass spectrometry (LC-MS/MS), it was revealed that subunit 2 of cIV might have allosteric interactions with C6-ceramide (). Reduced cIV activities were found in the liver of ceramide synthase 2 deficient mice where very long acyl chain (C22-C24) ceramides were barely detectable while C16-ceramide was accumulated (Zigdon et al., 2013). In cells that have excessive levels of dihydroceramide and dihydrosphingolipids due to the lack of dihydroceramide desaturase 1, the activities of cI and cIV (especially the latter) were inhibited (). In these animal models, it is still not clear whether cIV activity is affected by PLA2G6/iPLA2-VIA and how the imbalance of lipids influences cI and cII.

GBA1: Lysosomal Dysfunction Linking to ETC Impairment

GBA1 is not a traditional PARK-designated gene. Its biallelic mutations cause Gaucher’s disease, which is a lysosomal storage disease. However, increasing reports have identified GBA1 variants as common risk factors of PD (; ). The combined odds ratio of any GBA1 mutant in PD patients versus controls was as high as 5.43 (95% CI, 3.89 to 7.57) according to an international multi-center study, with the most common variants being N370S and L444P (). PD patients with GBA1 variants have a more aggressive motor deterioration and accelerated course of dementia than other PD patients ().

GBA1 encodes β-glucocerebrosidase (GCase), a lysosomal enzyme hydrolyzing glucosylceramide (GlcCer) into glucose and ceramide. The enzyme activity of GCase decreases in mutation carriers and elder individuals (; ). In contrast to the PLA2G6 mutants that cause accumulation of ceramide, GCase deficiency results in the accumulation of GlcCer, as well as alterations of glucosylsphingosine (GlcSph), sphingosine (Sph), sphingosine-1-phosphate (S1P) in different brain regions (). GCase deficiency impaired lysosomal recycling () and lysosomes enriched with GlcCer accelerates and stabilizes soluble α-Syn oligomers, which eventually become amyloid fibrils (). GCase deficiency also causes significant mitochondrial morphological changes, decreased oxygen consumption rate, and reduced respiratory chain complex activities (; ).

The mechanism of how a lysosomal enzyme such as GCase influences ETC is elusive. In gba knockout mice inhibited activities of cI and cII-III, but not cIV, were observed. Reduced co-localization of mitochondrial marker with LC3 and parkin recruitment suggested impaired mitophagy which might explain the compromised ETC activities (). However, the mitophagy theory unlikely explains the discrepancy between cIV and other complexes (). CoQ, which transfers electrons from cI and cII to cIII, has been shown beneficial for mitochondrial function in fibroblasts from patients of Gaucher disease (), but definite evidence showing deficits of CoQ in GCase depleted cells is lacking.

The ETC dysfunction can also result from alterations of mitochondrial metabolism. In iPSC neurons derived from PD patients carrying GBA1 variants, cI perishment with decreased expression of nicotinamide mononucleotide adenylyltransferases2 (NMNAT2) was found. Supplementation of NAD+ precursor nicotinamide riboside rescued mitochondrial defects and autophagy. The lipidomic analysis of mitochondria showed no accumulation of GlcCer in mitochondria except for C16-GlcCer (). Where do the C16-GlcCer exactly locate and whether it regulates ETC functions are still unanswered.

LRRK2: A Master Kinase Regulating the Mitochondrial Quality Control System

Mutations in LRRK2, which encodes the leucine-rich repeat kinase 2 (LRRK2), are the most prevalent causes of autosomal dominant PD worldwide. LRRK2-associated PD (LRRK2-PD) are characterized by late-onset (>60 years), with clinical features and treatment response resembling idiopathic PD (; ; Zhao et al., 2020). G2019S is the most common variant in Europe and North America, accounting for 4% of familial PD and 1% of sporadic PD (), while in Asia this variant is rare but prominent allelic heterogeneity was observed (, ). One of the main features of LRRK2-PD is its incomplete penetrance which varies among each variant and even different ethnicities (). A polygenic risk score has been developed to predict the vulnerability of G2019S carriers, and genes involving the vacuolar functions, lysosomal functions, and endocytic pathways were included (). Recent studies suggested that mitochondrial DNA damage is also a potential biomarker in LRRK2-PD (). Reduced cI activities and increased mitochondria DNA copy number were observed in fibroblasts derived from G2019S PD patients than from carriers (), implying mitochondria in association with the incomplete penetrance of the G2019S variant.

LRRK2 is an enzyme with both kinase and GTPase activities (). The kinase activity of LRRK2 is responsible for phosphorylation of a subset of Rab small GTPase (), which are important for exocytosis of synaptic vesicles, and endolysosomal and Golgi apparatus protein sorting (). Most PD-associated LRRK2 mutations represent gain-of-function alleles: the G2019S and I2020T in the kinase domain cause hyper-phosphorylation of Rabs (West et al., 2005; ), and R1441C/G/H in the Roc-COR domain increase kinase activities through disruption of the GTP hydrolysis activity (Weiss, 2008). Hyperphosphorylation of the Rabs disrupts their interaction with GTP/GDP exchange factors (GEFs), and Rab GDP dissociation inhibitors (GDIs), resulting in their inactivation and membrane-cytosol redistribution ().

Accumulating evidence has shown the impairment of mitochondria in LRRK2 mutants. Fibroblasts from LRRK2-PD patients exhibited defects including reduced mitochondria membrane potential, decreased ATP production, and decreased activities of complexes I, II, and IV than healthy controls. Elongation of mitochondria and increased interconnectivity of mitochondria were also observed in LRRK2-PD fibroblasts (). LRRK2 is recruited to the mitochondrial outer membrane by mitochondria-anchored Rab29 (also called Rab7L1) (). RAB29 is also a risk gene of PD identified in PARK16 loci from several GWASs (), and in addition to mitochondria Rab29 is also distributed at trans-Golgi network, lysosomes, and autophagic vesicles to regulate membrane trafficking, lysosome homeostasis, and axonal transport of autophagosomes with LRRK2 (; ; ). When gain-of-function LRRK2 mutants are anchored to mitochondria, it phosphorylates and traps Rab10 nearby (), causing decreased interaction of Rab10 with OPTN (optineurin), which is an autophagy receptor, therefore impairing mitophagy (Wauters et al., 2020). LRRK2 also regulates mitophagy through interaction with Miro, which anchors mitochondria to microtubules in mitochondria axonal transport (). In mitophagy, PINK1/parkin-dependent phosphorylated Miro is targeted to proteasome degradation, thus “quarantining” damaged mitochondria from further transport (Wang et al., 2011). Upon CCCP treatment, Miro interacting with wild-type LRRK2 was degraded with time, while in the LRRK2 G2019S mutant, Miro showed decreased interaction with the mutant as well as resistance to degradation, delaying mitochondrial arrest and clearance in iPSC-derived neurons (). In addition to mitophagy, LRRK2 is also involved in other mitochondrial quality control pathways like fusion and fission and its cytoskeleton dynamics and trafficking (). Although evidence showing direct regulation of ETC by LRRK2 is limited, these quality control systems mentioned earlier do have profound impacts on ETCs. A recent in vitro study showed that PINK1-dependent phosphorylation of Ser111 of Rab8A antagonistically regulates the phosphorylation of Thr72 of Rab8A by LRRK2 (), suggesting that there might be even closer crosstalk between the two master kinase.

SNCA: A Vicious Cycle of Proteinopathy and Mitochondriopathy

SNCA is the first identified gene associated with PD in 1997 (). While α-synuclein (α-Syn), which is encoded by SNCA, is one of the fundamental components in the Lewy body, patients with SNCA mutants (A30P, E46K, H50Q, G51D, A53E, and A53T) or multiplications are really rare, characterized by autosomal dominant inheritance, widely distributed age of onset (from 19 to 81 years), and more rapid cognitive impairment (). Physiologically α-Syn is soluble in the cytosol, mainly located at presynaptic terminals to regulate the genesis, maintenance, and release of presynaptic vesicles (; ; ). The α-synucleinopathy results from the imbalance among synthesis, aggregation, and clearance of α-Syn. As a result, aggregation of α-Syn takes place in the cytosol or around membranes. The conformations of the aggregates include oligomers, protofibrils, and fibrils (). Accelerated oligomerization is the common trait among different disease mutants (), and a higher propensity of fibrilization might correlate to the earlier onset age in A53T than A30P or E46K (; ). Post-translational modifications like phosphorylation at Ser129 and truncation at the carboxyl terminal are also pathogenic (; ), while Ser87 is protective ().

Both oligomers and fibrils (especially the amyloid fibrils) are toxic to mitochondria (). Overexpression of α-Syn, cytosolic acidification, and Ser129 phosphorylated pathogenic α-Syn could all induce translocation of α-Syn to mitochondria (; ; Wang et al., 2019). By using a split-GFP tool, the translocated α-Syn was found distributed at the OMM and IMS but not in the matrix. Mutants (A30P and A53T) were more likely translocated to the IMS than the wild-type, and more α-Syn would be translocated to IMS upon oxidative stress or cI inhibition (). In turn, α-Syn inhibits cI activities in either monomer or oligomerized form (; ; ). Oligomeric α-Syn also selectively oxidized the ATP synthase β subunit and caused mitochondrial lipid peroxidation and increased mitochondrial permeability transition pore opening (). These suggest a vicious feed-forward loop between α-Syn aggregation and mitochondrial dysfunction.

Synergistic with oligomers, α-Syn fibrils impairs mitochondria by sequestering mitochondria and other organelles into Lewy bodies. added preformed fibrils (PFF) of α-Syn exogenously to the cultured neurons and found that at day 14 inclusions composed of loose fibrils start to interact with organelles, and at day 21 high enrichment of proteins from the mitochondria (including complexes I, II, and V), the endoplasmic reticulum, and the Golgi were detected in the Lewy body-like inclusions, corresponding to the reduction of the maximum OXPHOS capacity and maximum electron transport capacity at day 21. The important discovery links the two competitive but not conflicting theories of proteinopathy and mitochondriopathy in PD pathogenesis.

UQCRC1 and CHCHD2: Two Recently Found PD-Causing Genes Implicating Apoptotic Neuronal Death

The association between CHCHD2 and autosomal dominant late-onset PD was discovered in four independent families through genome-wide linkage analysis (). CHCHD2 is a member of the coiled-coil-helix-coiled-coil-helix (CHCH) domain-containing protein family known to participate in mitochondrial respiration, redox regulation, membrane ultrastructure, and dynamics in the IMS (). CHCHD2 binds with cIV and regulates its activity. Under hypoxia, CHCHD2 activates the transcription of itself and COX4I2, which encodes cIV subunit-4 isoform 2 (COX IV-2) (). The findings are compatible with the clinical data showing that fibroblasts derived from a patient with homozygous A71P mutant in CHCHD2 show reduced cIV activity in addition to cI deficiency (). CHCHD2 is also important in maintaining the cristae structure and preventing apoptosis. Cristae are the home of ETC complexes and changes in the structure would alter the distribution of mitochondrial complexes, thus influencing the efficiency of OXPHOS (; Wilkens et al., 2013). In the flight muscles of CHCHD2 null Drosophila, the mitochondrial cristae structure was disorganized (). Cristae integrity is regulated by OPA1 and the MICOS complex (). In cells carrying PD-linked CHCHD2 mutations, the expression level of MICOS components, such as Mitofilin, MINOS1, CHCHD3, CHCHD6, were all reduced (Zhou et al., 2019). In Drosophila, Chchd2 knockout increases the degradation of OPA1 by peptidase YME1L (). Similarly, OPA1 degradation is also observed in CHCHD2 and CHCHD10 double-knockout mice (). Mass spectrometry analysis reveals that CHCHD2 interacts with MICS1 (), which is involved in cristae organization and cyt c stabilization, thereby prohibiting cells from apoptosis ().

UQCRC1 is a newly identified PD-causing gene that is clinically characterized by autosomal dominant late-onset parkinsonism with polyneuropathy (, ). UQCRC1 encodes the mitochondrial ubiquinol-cyt c reductase core protein 1 (UQCRC1), which is a core subunit of cIII. UQCRC1 complexes with UQCRC2 and interacts with UQCRFS1, CYC1, and other cIII subunits to regulate cIII activity (; ). UQCRC1 and UQCRC2 are homologs to the mitochondrial-processing peptidase subunits and are predicted to process UQCRFS1 in situ during cIII assembly (). However, in vivo experimental validation is required.

To date, only three UQCRC1 human mutations have been identified and functionally validated: Y314S, I311L, and p.Ala25Glyfs*27. All three mutants show reduced maximal oxygen consumption rate, decreased ATP production, and increased ROS in SH-SY5Y cell lines. Drosophila and mice with the Y314S variant both showed degeneration of dopaminergic neurons and locomotor defects (), providing direct evidence for the involvement of cIII in PD pathogenesis.

In addition to regulating cIII activities, UQCRC1 is also anti-apoptotic. In cardiac cells, UQCRC1 prevents ischemia/reperfusion injury by activating the PI3K/Akt/GSK-3β pathway, upregulating the anti-apoptotic protein Bcl-2, and downregulating the pro-apoptotic protein Bax (Yi et al., 2020). In Drosophila, our recent data reveal that uqcr-c1 associates with cyt c in mitochondria to gate its release (). Previous reports have shown that complex I and II mediate signals for apoptosis (). Together, these findings suggest that the respiratory chain complexes are important regulators of apoptosis.

The Many Faces of ETC Proteins: Functions Beyond Bioenergetics and Possible Implications in Neurodegeneration

Abundant studies have focused on mitochondrial quality control, lysosomal functions, apoptosis, and α-Syn aggregation in PD pathogenesis and their influences on ETC functions. In contrast, whether ETC proteins have any direct involvement in these pathways other than bioenergetics remains unclear in neurodegenerative diseases.

The roles of ETC proteins in apoptosis have been extensively studied in cancer cell biology. Both cI and cII are implicated as apoptotic sensors via different mechanisms. In cI, cleavage of NDUFS1 by caspase-3 inhibits cI activities, causing ROS production and collapse of mitochondria membrane potential. The cleavage-resistant mutant NDUFS1 D255A decreases ROS formation and delays the loss of plasma membrane integrity (), suggesting cI inhibition as an accelerator of apoptosis. A similar mechanism is found in granzyme A-induced apoptosis, in which NDUFS3 was cleaved (). For cII, the acidification of matrix in response to mitochondrial stress causes the disintegration of cII by dissociating the anchoring subunit SDHC-SDHD from the enzyme subunit SDHA-SDHB. As the SDHA-SDHB subunit is still enzymatically active, it produces excessive ROS and causes apoptosis. cII, therefore, is a pH sensor of the matrix in programmed cell death.

The roles of complexes III and IV in apoptosis are less clear, but the electron carrier between the two complexes, cyt c, has been known as a key player by activating apoptosis protease activating factor-1 (Apaf-1) when released to the cytosol. Post-translational modifications, especially phosphorylation, of cyt c have regulatory roles in both electron transport and apoptosis (). Tissue-specific phosphorylation of cyt c at Ser47 and Tyr97 that are enriched in porcine brain tissues and insulin-treated porcine tissues, respectively (; ), are both anti-apoptotic not only by inhibiting caspase-3 activities () but also by reducing the cyt c-COX interactions to lower the COX reaction rate and ROS generation (; ; ). It is not known whether the post-translational modifications of cyt c alter any interactions with cIII. From the lessons of UQCRC1, we now know that cIII can keep cyt c from being released from mitochondria.

Emerging evidence disclosed that there are other moonlight functions of cyt c. Abundant cyt c was shown to translocate to the nucleus prior to caspase activation in the cytosol during the early phase of apoptosis (). The translocated cyt c hijacks histone chaperones such as suppressor of variegation, enhancer of zeste and trithorax (SET)/template-activating factor (TAF)-Iβ, inhibiting the nucleosome assembly/disassembly activity and DNA repairing (; ). DNA damage is common in many neurodegenerative disorders, including PD (). Cyt c has also been identified as one of the components of brainstem Lewy bodies (Wakabayashi et al., 2007) and might play as a trigger in α-Syn aggregation (). α-Syn fibrils rather than mutant oligomers or monomers specifically interact with cyt c according to an in vitro study (). Whether similar aggregations could be replicated in vivo and whether post-translational modifications of cyt c as mentioned earlier have any effects on these moonlight functions are still unknown.

Some evidence showed that ETC proteins are also involved in mitophagy or autophagy. For example, autophagy is inhibited by the cIII inhibitors Antimycin A or myxothiazol (). Another example is ECSIT (evolutionarily conserved signaling intermediate in Toll pathways), which is essential in cI formation (Vogel et al., 2007), found as a parkin substrate and co-localized with LC3B in CCCP-treated macrophages (). Upon loss of ECSIT, parkin still accumulates on the mitochondria, but the recruitment of LC3BII to mitochondria is compromised. This suggested that parkin-induced ECSIT ubiquitination is upstream of LC3BII recruitment for the initiation of mitophagy (), It is not clear whether ECSIT participates in receptor-mediated mitophagy like other mitochondrial outer membrane proteins such as FUNDC1, NIX, and BNIP3. Further investigation is needed to elucidate whether ECSIT mediates mitochondria clearance via a similar mechanism in the nervous system.

Discussion

In this review, we focus on the central role of ETC in PD pathogenesis. We have discussed that ATP production by mitochondria is directly or indirectly affected by reduced turnover and altered post-translational modifications of ETC proteins, impaired assembly or instability of mitochondrial complexes, abnormal protein aggregation, and dysregulated lipid metabolism. While mutations of ETC proteins are rarely linked as a direct cause of PD (except for UQCRC1), impaired ETC proteins may not only lead to excessive ROS production, but also cause impeded mitophagy, α-Syn aggregation, and apoptosis in PD pathogenesis.

It might be argued that for the long time scale of PD progression, apoptosis is an acute cellular event that on the surface does not perfectly explain the development of PD, but when patients start to have detectable parkinsonism symptoms, there is usually at least 40–60% dopaminergic neuronal loss in the substantia nigra (), suggesting progressive accumulation of neuronal loss. At the organismal level, apoptosis of dopaminergic neurons occurs over time, responding to stress such as aggregated proteins or ETC dysfunction. At the cellular level, it takes time for a cell to reach the threshold that breaks the balance between pro-apoptotic and anti-apoptotic pathways. For example, ROS accumulation, alterations of mitochondrial outer membrane permeability, and the attenuation of antiapoptotic signals may fluctuate, and mutant cells with the propensity of releasing cyt c do not mean that it releases cyt c and activate caspases immediately. This concept of ‘apoptotic threshold’ has been well illustrated by a beautiful mathematical model (). In fact, apoptosis has been implicated in many neurodegenerative diseases, including PD (). Although it remains unclear whether loss of ETC function leads to protein aggregation in vivo, apoptosis represents an important pathogenic mechanism that warrants our attention.

In terms of mitochondrial dysfunction, the examples of UQCRC1 and CHCHD2 have suggested that PD is not just a cI disorder. Various genetic models of PD have now shown deficits in cII to cV. The conventional belief of PD as a cI disorder might be the result of publication bias or study limitations in earlier studies. It has been shown that even in cI, the significance of the deficiency could be influenced by the purity of the samples and the assays we use (). By using high-resolution quantitative fluorescence immunohistochemistry, Reeve et al. found that both cI and cIV were deficient in the remaining dopaminergic synapses of idiopathic PD (). By using imaging mass cytometry, a recent study revealed that there is even more widespread deficiency from complexes I to V in the postmortem human midbrain (), contrary to the findings of earlier studies that were based on the measurements of biochemical activities only.

Mitochondrial therapies have been proposed as a potential treatment of PD (); however, many attempts have failed. The anti-diabetic drugs glitazones which improves mitochondrial functions (such as activities of cI and cIV) and biogenesis (; ) did lower the incidence of PD in retrospective cohort studies (; ), but failed to show benefits in modifying disease progression of early PD (). The randomized clinical trial of high dose CoQ did not slow disease progression (). The limited bioavailability of CoQ has been considered as one of the main reasons for its failure (). Longer follow-up or earlier administration of the drugs at prodromal stages may be necessary to reveal the potential benefits. Idebenone, a synthetic short-chain analog of CoQ with improved lipophilicity and bioavailability (), having been approved as an orphan drug in treating LHON by European Medicines Agency (EMA) () and Duchenne muscular dystrophy by the United States Food and Drug Administration (FDA) (), is tested in two PD clinical trials: (1) Idebenone Treatment of Early Parkinson’s Disease symptoms (ITEP) (NCT03727295), and (2) A Study of Efficacy and Safety of Idebenone vs. Placebo in Prodromal Parkinson Disease (SEASEiPPD) (NCT04152655). The recruitment status of the former is due to end in May 2021, while the latter is still recruiting and is scheduled to complete by January 2023.

In conclusion, ETC dysfunction is a common feature in PD, secondary to various pathomechanisms. Increasing evidence has shown that cI is not the only affected ETC complex. Other complexes, such as cIII and cIV, are also involved. ETC dysfunction not only reduces ATP production but also induces apoptosis or impaired mitophagy that undermines neuronal viability in PD. So far there is still no silver bullet to prevent or to ameliorate PD progression. Non-etheless, new compounds or existing drugs await testing as potential therapeutic strategies for PD via improving ETC functions ().

Publisher’s Note

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.

Statements

Author contributions

J-LL, T-YL, and C-CC: conceptualization. J-LL and T-YL: writing – original draft. C-CC: funding acquisition and supervision. All authors writing – review and editing.

Funding

This work was supported by grants from the Ministry of Science and Technology of Taiwan (MOST) (108-2311-B-002-011-MY3 and 110-2314-B-002-202) to C-CC.

Acknowledgments

We thank the members of Chan Lab for critical comments on this manuscript.

Conflict of interest

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

References

Summary

Keywords

Parkinson’s disease, electron transport chain, mitochondria quality control, mitophagy, apoptosis

Citation

Li J-L, Lin T-Y, Chen P-L, Guo T-N, Huang S-Y, Chen C-H, Lin C-H and Chan C-C (2021) Mitochondrial Function and Parkinson’s Disease: From the Perspective of the Electron Transport Chain. Front. Mol. Neurosci. 14:797833. doi: 10.3389/fnmol.2021.797833

Received

19 October 2021

Accepted

18 November 2021

Published

09 December 2021

Volume

14 - 2021

Edited by

Chuang-Rung Chang, National Tsing Hua University, Taiwan

Reviewed by

Shigeru Yanagi, Gakushuin University, Japan; Kostas Tokatlidis, University of Glasgow, United Kingdom; Kuen-Jer Tsai, National Cheng Kung University, Taiwan

Updates

Copyright

*Correspondence: Chih-Chiang Chan,

This article was submitted to Brain Disease Mechanisms, a section of the journal Frontiers in Molecular Neuroscience

Disclaimer

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article or claim that may be made by its manufacturer is not guaranteed or endorsed by the publisher.

Outline

Figures

Cite article

Copy to clipboard


Export citation file


Share article

Article metrics