PINK1 Loss-of-Function Mutations Affect Mitochondrial Complex I Activity via NdufA10 Ubiquinone Uncoupling

See allHide authors and affiliations

Science  11 Apr 2014:
Vol. 344, Issue 6180, pp. 203-207
DOI: 10.1126/science.1249161

In the PINK1

Pathogenic mutations in the kinase PINK1 are causally related to Parkinson's disease (PD). One hypothesis proposes that PINK1 regulates mitophagy—the clearance of dysfunctional mitochondria. A second hypothesis suggests that PINK1 has a direct effect on mitochondrial complex I, affecting the maintenance of the electron transport chain (ETC) resulting in decreased mitochondrial membrane potential and dysfunctional mitochondria. In support of the second hypothesis, Morais et al. (p. 203, published online 20 March) observed a complex I deficit in fibroblasts and neurons derived from induced pluripotent stem cells from PINK1 patients before any mitophagy was induced. The phosphoproteome of complex I in liver and brain from mice deficient for Pink1, compared to wild-type animals, revealed that Ser250 in complex I subunit NdufA10 was differentially phosphorylated. Ser250 is critically involved in the reduction of ubiquinone by complex I, explaining why Pink1 knockout mice, flies, and patient cell lines show decreased mitochondrial membrane potential. Synaptic defects in pink1 null mutant Drosophila could be rescued using phosphomimetic NdufA10.


Under resting conditions, Pink1 knockout cells and cells derived from patients with PINK1 mutations display a loss of mitochondrial complex I reductive activity, causing a decrease in the mitochondrial membrane potential. Analyzing the phosphoproteome of complex I in liver and brain from Pink1−/− mice, we found specific loss of phosphorylation of serine-250 in complex I subunit NdufA10. Phosphorylation of serine-250 was needed for ubiquinone reduction by complex I. Phosphomimetic NdufA10 reversed Pink1 deficits in mouse knockout cells and rescued mitochondrial depolarization and synaptic transmission defects in pinkB9-null mutant Drosophila. Complex I deficits and adenosine triphosphate synthesis were also rescued in cells derived from PINK1 patients. Thus, this evolutionary conserved pathway may contribute to the pathogenic cascade that eventually leads to Parkinson’s disease in patients with PINK1 mutations.

Mutations in PINK1, a mitochondrial targeted Ser/Thr kinase, cause a monogenic form of Parkinson’s disease (PD) (1, 2). Loss of PINK1 function mutations interfere with Parkin-mediated carbonyl cyanide m-chlorophenyl hydrazone (CCCP)–induced mitophagy (35) and mitochondrial fusion and fission defects (4). However, an early and invariant phenotype of PINK1 loss of function in different species is an enzymatic defect in mitochondrial complex I and a decrease in mitochondrial membrane potential (Δψm) (68). In contrast to effects of PINK1 on toxin-induced mitophagy like CCCP (4, 5, 9), these complex I defects are observed in cell culture and Drosophila neurons under resting conditions with normal-appearing mitochondria (6, 8). Because Pink1−/− mice display only subtle, and somewhat controversial, phenotypes of altered mitochondrial morphology (6, 8, 10, 11), it remains unresolved to what extent decreased mitophagy, or, alternatively, primary complex I deficiency, or both, are involved in those defects (6). In pink1 and parkin Drosophila models (1214), phenotypes are more pronounced. Thorax muscle degeneration and flight deficits can be rescued by expression of the fission-promoting gene drp1 or by ablating the fusion-promoting gene opa1 (15, 16), linking these molecules to the role of PINK1 and Parkin in fusion and fission defects. Intriguingly, other pink1-related phenotypes, such as defective neurotransmitter release, adenosine triphosphate (ATP) depletion, and loss of Δψm, cannot be rescued efficiently in Drosophila neurons by fission gene Drp1 (17, 18) but can be rescued by genes restoring the proton motive force (19) or by NDi1, a yeast rotenone-insensitive reduced form of nicotinamide adenine dinucleotide (NADH)–quinone oxidoreductase (17). This suggests that two parallel molecular pathways are affected by pink1 deficiency in flies, and both could be relevant to our understanding of the role of PINK1 in PD.

First, we confirmed the pathological relevance of the previously reported Δψm defects in Pink1−/− mice and Drosophila using human fibroblasts and two induced pluripotent stem (iPS) cell lines derived from PD patients with PINK1 mutations. Fibroblasts contained homozygous p.Q456X nonsense (L2122) or p.V170G missense (L1703) mutations (20), and iPS cells were derived from two PD patients with c.1366C>T; p.Q456X nonsense (L2122 and L2124) mutations. Integrity of the mitochondrial-targeted red fluorescent protein–labeled mitochondrial network was qualitatively and quantitatively (fragmented versus elongated) not different between control (L2134 and L2132) and patient (L1703 and L2122) fibroblasts (fig. S1, A and B). Δψm was significantly decreased in the patient fibroblasts as assessed by the electrochemical potentiometric dye tetramethyl rhodamine ethyl ester (TMRE) (fig. S1, C and D). Overall ATP content in these PINK1 mutant fibroblasts was also decreased when compared with age-matched controls (fig. S1E) (20). In neuronal differentiated iPS cells (L2124 and L2122) (21), Δψm and ATP content (fig. S1, F to H) were lowered compared with controls (L2134 and L2135), confirming that clinical mutations in the context of human cells and human neurons display similar deficits as cell lines derived from Pink1-null mice and flies.

Cells display a specific deficit in the enzymatic activity of complex I (6). Therefore, we immunocaptured complex I from isolated mouse mitochondria (Fig. 1, A and B) and obtained independent phosphoproteomes from three brain and three liver preparations, covering 40 out of the 46 complex I subunits (table S1). Nine previously unknown phosphosites were identified (table S2), but they were present both in Pink1-deficient and wild-type tissue. Only Ser250 in complex I subunit NdufA10 (Fig. 1C and table S2) was lacking in knock-out material. In contrast, the unphosphorylated peptide was identified in three out of the six knock-out samples (fig. S2). The Ser250 is conserved from Drosophila to human, suggesting that it is functionally important (Fig. 1C). Consistent with this notion, stable transfection of phosphomimetic NdufA10S250D or wild-type NdufA10wt, but not phosphorylation-deficient NdufA10S250A, rescued the complex I activity defect in NdufA10-deficient HeLa cells (fig. S3, A and B).

Fig. 1 Lack of phosphorylation of the complex I subunit NdufA10 at Ser250.

(A and B) Schematic overview of the analysis of complex I phosphoproteome. Mitochondrial-enriched fractions from Pink1+/+ and Pink1−/− mouse brain and liver were treated with 1% dodecyl maltoside, and the solubilized mitochondrial protein complexes were immunocaptured using 20-kD complex I subunit antibody. The immunocaptured complex I was further analyzed on SDS–polyacrylamide gel electrophoresis, followed by colloidal Coomassie staining (B). The visualized protein bands were excised from the gel, digested with trypsin, and fractionated by cation exchange chromatography to enrich for phosphopeptides. Phosphopeptides were further analyzed by liquid chromatography–mass spectrometry on a Nano-LC-LTQ-Orbitrap-MS. (C) Alignment of human, mouse, and Drosophila NdufA10 revealed that the identified PINK1-dependent phosphoserine (in magenta) is conserved across species. (D) Pink1+/+ and Pink1−/− MEFs were stably transduced with 3xFLAG-tagged wild-type NdufA10wt, phosphorylation-deficient NdufA10S250A, and phosphomimetic NdufA10S250D. Expression levels were analyzed by immunoblotting of mitochondria-enriched fractions using antibodies to FLAG and Hsp60 (loading control). (E, F, and G) Phosphomimetic NdufA10S250D restores the mitochondrial membrane potential and ATP levels in Pink1−/− MEFs. Cells were stably transduced with NdufA10 constructs as indicated and loaded with 10 nM TMRE (E), and quantification of TMRE intensity (F) over mitochondrial regions of interest was performed using ImageJ software. ATP levels were measured in cell lysates (G). Statistical analysis: Student’s t test; **, P < 0.01; ns, not significant; mean ± SD; n = 100. Scale bar, 10 μm. (H) Live imaging of Δψm reveals that phosphomimetic NdufA10S250D restores the mitochondrial membrane potential. Pink1+/+, Pink1−/−, and Pink1−/− expressing the phosphomimetic NdufA10 mutants’ fibroblast cells were loaded with 10 nM TMRM in the presence of 2 μg/ml cyclosporine H. Sequential images of TMRM fluorescence were acquired every 60 s over a 40-min time course. TMRM fluorescence over mitochondrial regions of interest was measured. When indicated (arrows), 2 μg/ml oligomycin (an ATP synthase inhibitor) and 2 μM carbonyl cyanide 4-trifluoromethoxyphenylhydrazone (an uncoupler) were added. Mean ± SD; n = 5 independent experiments. (I) Phosphomimetic NdufA10 mutant protects Pink1-devoid cells from H202-induced apoptosis. Pink1+/+, Pink1−/−, and Pink1−/− fibroblasts expressing the phosphomimetic NdufA10 mutants were treated with increasing concentrations of H2O2, and after 4 hours apoptosis was determined using the luminescent cleaved Caspase-Glo 3/7 assay. Mean ± SD; n = 3 independent experiments.

To assess the role of this site in the context of Pink1, we stably transfected Pink1+/+ and Pink1−/− mouse embryonic fibroblasts (MEFs) with NdufA10wt, NdufA10S250A, and NdufA10S250D (Fig. 1D). Although the defect in Δψm assessed by TMRE labeling was not rescued with NdufA10wt or NdufA10S250A, NdufA10S250D completely restored Δψm and ATP levels (Fig. 1, E to G). No effects on Δψm were observed in Pink1+/+ MEFs expressing the NdufA10 mutants (Fig. 1, E to G). Real-time imaging of Δψm with tetramethylrhodamine methyl ester (TMRM) further corroborated that NdufA10S250D restored this defect (Fig. 1H). Additionally, NdufA10S250D appeared to exert a protective effect toward previously reported susceptibility of Pink1−/− mutant cells to H2O2-induced apoptosis (Fig. 1I). Thus, a phosphomimetic mutation at Ser250 in NdufA10 is sufficient to restore the defect in Δψm in Pink1−/− mutant cells.

To study the physiological relevance of the PINK1-dependent phosphorylation event, we assessed to what extent Drosophila pink1 mutant phenotypes could be rescued by expression of wild-type (A10wt), phosphorylation-deficient (A10SA), or phosphomimetic (A10SD) NdufA10. We focused first on phenotypes dependent on complex I deficiency. Pink1B9-null mutant Drosophila failed to maintain neurotransmitter release at neuromuscular junctions (NMJ) during high-frequency stimulation (10 Hz) (fig. S4A) (6), a defect that was fully rescued when expressing A10SD but not with A10wt or A10SA (Fig. 2A). Basal neurotransmitter release was not affected under the conditions tested (Fig. 2B). This deficit is caused by a defect in reserve pool (RP) vesicle mobilization (6), which can be assessed using FM 1-43, a lipophilic dye that labels these vesicles. Deficient loading of the RP vesicles was rescued upon expression of A10SD in pink1B9 flies (Fig. 2C and fig. S4B) but not with A10wt or A10SA. This phenotype was ATP dependent and was caused by a loss of Δψm in the mitochondria at the NMJ as assessed with JC-1, a green fluorescent potentiometric dye that shifts to red fluorescence as a function of a normal negative Δψm (22). Synaptic mitochondria of pink1B9 mutants expressing A10SD showed red JC-1 aggregates comparable to control (Fig. 2D and fig. S4C), indicating restoration of the Δψm, whereas A10wt or A10SA displayed similar weak signals as the pink1B9 mutant. Also, ATP synthesis was restored upon expression of A10SD (Fig. 2F). Δψm was not disturbed in control pink1rev NMJs expressing NdufA10 mutants (fig. S5, A and B). Thus, the synaptic phenotype in pink1B9 flies was significantly rescued by A10SD. Pink1B9 flies also show severe muscle degeneration and flight defects (12, 13). These phenotypes were not rescued upon A10SD expression, suggesting that most likely they are linked to other PINK1 functions (fig. S5, C and D). Also, CCCP-induced Parkin recruitment was not restored in Pink1−/− cells expressing the phosphomimetic NdufA10S250D mutant (fig. S5E), possibly ruling out an effect of NdufA10 on the mitophagy pathway. We investigated more closely the mitochondria in the flight muscles by performing electron microscopy. A10SD improved cristae structural organization without rescuing overall mitochondria morphology (Fig. 2, E and G). Thus, the muscular degeneration in Drosophila appears to be mainly the result of mitochondrial fusion and fission defects caused by PINK1 deficiency (16); however, restoring complex I enzymatic activity aids cristae organization leading to improved bioenergetics but does not improve the mitochondria morphology or the muscular degeneration.

Fig. 2 Restoration of synaptic defects in Drosophila pink1B9 null mutants by expressing phosphomimetic NdufA10.

(A) Relative excitatory junction potentials (EJPs) amplitudes measured in 2-mM Ca2+ during 10 min of 10-Hz stimulation in pink1B9-null mutants expressing wild-type (A10wt), phosphodeficient (A10SA), and phosphomimetic (A10SD) NdufA10 mutants. Inset represents an overlay of a raw data trace of EJPs recorded for 10 min at 10 Hz in 2-mM calcium of pink1B9+A10SD (black) and pink1B9+A10wt (gray). The deficit to maintain normal EJP amplitude during a 10-Hz stimulation train observed in pink1B9 mutant expressing A10wt is restored when phosphomimetic A10SD is present. Mean ± SEM; n = 4 larvae for A10wt, 7 for A10SA, and 8 for A10SD. (B) Traces of basal neurotransmitter release measured at 1 Hz in 2-mM Ca2+ in pink1B9-null mutants expressing NdufA10 mutants. The average EJP amplitudes recorded are pink1B9+A10wt, 56.4 ± 1.9 mV; pink1B9+A10SA, 52.4 ± 2.1 mV; and pink1B9+A10SD, 57.0 ± 3.3 mV. Basal neurotransmitter release is not affected in pink1B9 larvae expressing the NdufA10 mutants. (C) RP labeling at Drosophila larval NMJs in controls (pink1REV) and pink1 mutants (pink1B9). Both the exo/endo cycling pool (ECP) and RP were labeled with FM1-43; after depolarization, only ECP vesicles, but not RP vesicles, were unloaded. Synapses were imaged after this unloading procedure. Quantification of fluorescence intensity of loaded RP vesicles was normalized to loading intensity of controls. The loading defect in pink1B9 is restored upon expression of phosphomimetic NdufA10. (D) Quantification of mitochondrial membrane potential at NMJ boutons in controls (pink1REV) and pink1 mutants (pink1B9) using the ratiometric dye JC-1, where the red JC-1 fluorescence emission to green emission (in the same area) is compared. The mitochondrial membrane potential was restored upon expression of phosphomimetic NdufA10. (F) Analysis of ATP levels in these mutant flies where decreased ATP content is restored in mutant flies expressing the phosphomimetic A10SD. (E and G) Electron micrographs [(E), inset in right column] and corresponding quantification of mitochondria that have organized cristae independently of overall mitochondrial morphology (E) of adult fly muscle in controls (pink1REV) and pink1 mutants (pink1B9) showing that cristae organization is partially restored in mutant flies expressing A10SD. Statistical analysis: Student’s t test; ***P < 0.001; **P < 0.01; *P < 0.05; mean ± SD; n = 8 larvae. Scale bar, 4.5 μm.

NdufA10 is located in subunit Iγ of complex I in close vicinity to the ND1 and ND3 subunits (23, 24). We hypothesized that the identified phosphorylation site on NdufA10 could structurally influence the ubiquinone binding cavity. We performed enzymatic assays for complex I to assess the reduction of coenzyme Q1 (CoQ1) and decylubiquinone, a ubiquinone analog (Fig. 3A). Reduction of both electron acceptors, CoQ1 and decylubiquinone, was significantly affected in complex I prepared from Pink1−/− MEFs expressing NdufA10wt but was restored with NdufA10S250D (Fig. 3, B and C). The effect on complex I was specific for its ubiquinone reductase enzymatic function because another enzymatic assay that employs only the NADH-binding site of complex I, which is based on the reduction of the artificial substrate hexammineruthenium (HAR) (25) (Fig. 3A), was not affected in Pink1−/− cells or in Pink1−/−-expressing NdufA10 mutants (Fig. 3D). Thus, NdufA10 is required for the binding and/or reduction of the physiological complex I substrate ubiquinone.

Fig. 3 Expression of phosphomimetic NdufA10 rescues ubiquinone-reducing capacity of complex I.

(A) Schematic representation of complex I. Complex I is composed of 45 different subunits that assemble into a structure of ~1 MD. Electrons that arise from the oxidation of NADH are transferred to a noncovalently bound flavin mononucleotide and subsequentially passed through a series of iron-sulfur clusters (Fe-S), finally reaching the acceptor ubiquinone (Q), that is reduced to ubiquinol (QH2). NdufA10 is located in subunit Iγ within the membrane arm domain of complex I, in close vicinity to the predicted ubiquinone binding pocket. (B, C, and D) Analysis of enzymatic function of complex I. Spectrophotometric assays were performed to measure complex I (NADH:ubiquinone oxidoreductase) and citrate synthase activities on mitochondria homogenates from Pink1+/+ and Pink1−/− MEFs rescued with NdufA10 mutants as indicated. In (B) and (C), NADH:ubiquinone reduction (rotenone sensitive) with electron acceptor CoQ1 or decylubiquinone was used, respectively, and in (D), NADH:HAR reduction was measured. Values were normalized to citrate synthase activity. (E and F) Complex I enzymatic activity is restored when PINK1 is localized to the mitochondria. Pink1−/− MEFs were stably transduced with truncated PINK1 mutants that alter the localization of PINK1, and complex I enzymatic assays were performed. Statistical analysis: Student’s t test; ***P < 0.001; **P < 0.01; *P < 0.05; ns, not significant; mean ± SD; n = 3 independent experiments.

To determine whether PINK1 needs to be targeted to mitochondria to facilitate NdufA10 activity, we introduced PINK1 truncated mutants into the Pink1−/− MEFs (Fig. 3E). Amino acids between position 77 and 112 are sufficient to target PINK1 to the mitochondria (26). PINK1-ΔN77 was indeed targeted to the mitochondria, whereas the ΔN113 and ΔN153 were mainly present in nonmitochondrial fractions (Fig. 3E). Only ΔN77 restored complex I enzymatic activity (Fig. 3F). This and the association of PINK1 with the inner mitochondrial membrane, as shown with proteinase K protection assays and the extraction with sodium carbonate (fig. S6 and fig. S7, A and B), suggests that PINK1 needs to be targeted to the mitochondria to activate NdufA10, although more complex scenarios cannot be ruled out.

We finally investigated whether Pink1−/− cells expressing human wild-type or PINK1 containing PD-causing mutations could be rescued with NdufA10S250D (fig. S8A). NdufA10S250D was able to restore fully the decylubiquinone reduction reaction in complex I from cells expressing PINK1 clinical mutants G309D and W437X or the artificial kinase inactive (KD) PINK1 mutant (Fig. 4A). Furthermore, Δψm was fully restored upon NdufA10S250D expression in the cells expressing the PD-causing mutations (Fig. 4B).

Fig. 4 Restoration of mitochondrial membrane potential deficits caused by PINK1 PD-causing mutations.

(A) Respiratory chain measurements performed on mitochondria homogenates from Pink1−/− MEFs rescued with human PINK1 wild-type (wt) or PD-causing mutants or artificial kinase dead mutant (KD). The cells were stably transduced with NdufA10 phosphomimetic mutants and analyzed by spectrophotometric assays of complex I (NADH:ubiquinone oxidoreductase, rotenone sensitive) and citrate synthase enzyme activities. Values were normalized to citrate synthase activity. The enzymatic activity of complex I is rescued in the presence of NdufA10S250D mutant. (B) Quantification of mitochondrial membrane potential in the same cell lines. Cells were loaded with 10 nM TMRE, and quantification of TMRE intensity over mitochondrial regions of interest was performed using ImageJ software. The mitochondrial membrane potential is restored when phosphomimetic NdufA10S250D is coexpressed. Statistical analysis: Student’s t test; **, P < 0.01; *, P < 0.05; ns, not significant; mean ± SD; for (A), n = 3 independent experiments; for (B), n = 70 images. Scale bar, 10 μm. (C) Quantification of mitochondrial membrane potential in control (L2134 and L2132) and PINK1 patient (L1703 and L2122) fibroblasts electroporated with GFP-tagged NdufA10 mutants loaded with 10 nM TMRE. TMRE fluorescence over mitochondrial regions of interest was quantified using ImageJ software; only cells that were GFP positive were analyzed. Statistical analysis: Student’s t test; **P < 0.01; *P < 0.05; ns, not significant; mean ± SD; n = 80 images. Scale bar, 10 μm.

We electroporated fibroblasts derived from PINK1-mutant patients and controls (fig. S1) with green fluorescent protein (GFP)–tagged forms of the NdufA10 mutants, and the Δψm was rescued as assessed by TMRE in the patient-derived fibroblasts expressing GFP and NdufA10S250D but not in those expressing NdufA10wt or NdufA10S250A (Fig. 4C). Thus, restoration of the pseudophosphorylation status of NdufA10 rescues complex I activity in cells harboring PD-causing mutations in PINK1.

Phosphorylation of Ser250 in NdufA10 regulates the ubiquinone reductase ability of complex I. Crystal structures from complex I reveal that subunits NdufS2 and NdufS7 are involved in electron donation to ubiquinone (27, 28) and that contacts between the peripheral arm and membrane domain of complex I are mediated by NdufS2, ND1, and ND3, leading to the formation of a cavity capable of harboring the large hydrophobic substrate ubiquinone (23, 29). NdufA10 is located close to ND1 and ND3; therefore, phosphorylation of this site could regulate the interaction of complex I with ubiquinone. Vitamin K2, an alternative electron carrier for ubiquinone, can rescue the pink1B9 mutant phenotype in Drosophila (19), in agreement with this hypothesis. Although the most parsimonious explanation for our observations is that PINK1 phosphorylates NdufA10, we have not been able to demonstrate such activity directly and therefore cannot rule out an indirect effect of PINK1 on this site. Nonetheless, our findings provide a molecular link between PINK1 dysfunction and activity of complex I. We conclude that PINK1 has dual functions in mitochondrial homeostasis. Under steady-state conditions, PINK1 is needed to maintain complex I and electron transport chain (ETC) activity. The clinical mutations create a latent situation where this phosphorylation is affected and ETC function becomes destabilized, as shown here under steady-state culture conditions. When additional stress is exerted—for example, in the presence of CCCP or other mitochondrial toxins—defects in mitophagy are also observed (30). Such a multiple-hit hypothesis for familial PD makes sense given the relatively late onset of the disease. Rescuing complex I activity by activating phosphorylation or inhibiting dephosphorylation of NdufA10 at residue Ser250 could thus potentially prevent or partially attenuate the disease.

Supplementary Materials

Materials and Methods

Figs. S1 to S8

Tables S1 and S2

References (3140)

References and Notes

  1. Acknowledgments: This work was supported by the Fund for Scientific Research Flanders (FWO); research fund KU Leuven; the Hercules Foundation, Federal Office for Scientific Affairs (IAP P7/16); a Methusalem grant of the Flemish Government, VIB, Agentschap voor Innovatie door Wetenschap en Technologie (IWT), the European Research Council (ERC StG and AdG to P.V. and B.D.S.); the Queen Elisabeth Foundation; the Hermann and Lilly Schilling Foundation (to C.K.); the Deutsche Forschungsgemeinschaft (to A.G. and to C.K.); the Fritz Thyssen Foundation (to A.G.); and the StemBANCC consortium (to C.K.). B.D.S. is the Arthur Bax and Anna Vanluffelen chair for Alzheimer’s disease. S.V. is supported by an FWO postdoctoral fellowship, and L.A. is supported by an IWT predoctoral fellowship. B.D.S., V.A.M., and P.V. are the inventors on a patent application comprising diagnostic assays measuring the phosphorylation status of NdufA10 and screening methods for compounds able to restore or increase the phosphorylation of NdufA10. B.D.S. is a paid consultant for the Alzheimer’s disease research programs at Janssen Pharmaceutica, Envivo, and Remynd NV. The data reported in this paper are tabulated in the main paper and in the supplementary materials.
View Abstract

Navigate This Article