Introduction
Parkinson’s disease is the second most common human neurodegenerative disease and is characterised by the preferential loss of dopaminergic neurons in the substantia nigra. The cell and molecular events that give rise to this neuronal loss are not properly understood but a number of lines of evidence suggest that abnormalities in α-synuclein are central to the disease process. First, mutations within the gene encoding α-synuclein and increased α-synuclein gene dosage involving duplication and triplication events cause familial, dominantly inherited forms of the disease [
1,
12,
41,
45,
49,
66,
74,
92]. Second, α-synuclein is the major protein constituent of Lewy bodies and Lewy neurites which are hallmark pathologies of Parkinson’s disease [
76,
80]. Finally, overexpression of wild-type and familial mutant α-synuclein can induce aspects of disease in transgenic mice [
6,
7,
18].
Despite this evidence, the mechanisms by which altered α-synuclein metabolism might cause disease are not properly understood. α-Synuclein is a 140 amino acid protein of unclear function that is enriched in synapses and perinuclear regions of neurons. Within neurons, α-synuclein localizes to cytosolic and membrane compartments including synaptic vesicles, mitochondria and the endoplasmic reticulum (ER) (see reviews [
32,
33,
91]). Recent evidence suggests that its membrane localization involves targeting to lipid rafts (also known as detergent resistant membranes; DRM) [
31].
Overexpression of wild-type or familial mutant α-synuclein has been shown to damage a number of physiological processes. These include Ca
2+ homeostasis [
11,
38], lipid metabolism [
9,
31], mitochondria including mitochondrial transport and biogenesis [
22,
32,
60,
90,
91], the ER [
59] and autophagy [
88]. Indeed, the difficulty in deciphering α-synuclein toxicity is linking these apparently diverse pathological changes to a common disease pathway.
One area of cell physiology that impacts upon all of these features involves signalling between ER and mitochondria. Approximately 5–20% of the mitochondrial surface is closely apposed (10–30 nm distances) to ER membranes; these specialized regions of ER are termed mitochondria-associated ER membranes (MAM). A large body of evidence demonstrates that mitochondria communicate directly with ER through MAM to regulate a number of fundamental cellular processes. These include Ca
2+ homeostasis, lipid metabolism, mitochondrial ATP production, mitochondrial transport and biogenesis, ER stress and the unfolded protein response (UPR) and autophagy (see reviews [
46,
48,
63,
70,
82,
83]). Any α-synuclein-induced damage to ER–mitochondria associations thus represents a plausible route for explaining many features of Parkinson’s disease.
The mechanisms by which regions of ER are recruited to mitochondria are not fully known but electron microscopy (EM) studies reveal the presence of structures that appear to tether the two organelles [
16]. Recently, the integral ER protein VAPB was shown to bind to the outer mitochondrial membrane protein PTPIP51 to form at least some of these tethers [
20,
26,
40,
78,
79]. Thus, modulating VAPB and/or PTPIP51 expression induces appropriate changes in ER–mitochondria contacts and Ca
2+ exchange between the two organelles which is a physiological readout of ER–mitochondria associations [
20,
78]. Here, we show that α-synuclein binds to VAPB, disrupts the VAPB-PTPIP51 interaction and perturbs ER–mitochondria associations. We also confirm that a proportion of α-synuclein is present in MAM. Moreover, we demonstrate that α-synuclein-induced loosening of ER–mitochondria contacts affects Ca
2+ exchange between the two organelles. As such, our findings reveal a new molecular mechanism to link α-synuclein and Parkinson’s disease.
Discussion
Despite the wealth of data linking α-synuclein to Parkinson’s disease, the targets for α-synuclein toxicity are not fully understood. Here, we show that α-synuclein perturbs ER–mitochondria associations and that this involves disruption to the VAPB-PTPIP51 tethering proteins. Importantly, this damage is also seen in iPS cell-derived dopaminergic neurons from Parkinson’s disease patients harbouring triplication of the α-synuclein locus. Although we used different α-synuclein triplication and control iPS cell clones in these latter studies, it will be important to confirm in future studies that other genetic differences between the clones do not contribute to this phenotype. This could be achieved by analyses of genetically corrected mutant clones such that the genetic backgrounds are identical. Using a range of assays including immunoprecipitation, cellular GST pull-down, proximity ligation and in vitro binding of recombinant proteins, we also show that α-synuclein is a direct binding partner for VAPB. Interestingly, earlier mass spectrometry proteomic analyses also suggested that α-synuclein was complexed with VAPB although these studies did not discriminate between direct and indirect binding nor provide confirmatory data [
58].
A primary function of ER–mitochondria associations is to deliver Ca
2+ from ER stores to mitochondria. This delivery involves release of Ca
2+ from IP3 receptors in MAM and uptake via the mitochondrial voltage dependent anion channel (VDAC) [
46,
48,
63,
70,
82]. Mitochondria require Ca
2+ to efficiently produce ATP since several dehydrogenases in the tricarboxylic acid cycle are Ca
2+ regulated [
30]. Thus, disruption to the VAPB-PTPIP51 tethers perturbs IP3 receptor-mediated ER–mitochondria Ca
2+ exchange and mitochondrial ATP production [
20,
78,
79]. Consistent with these findings, we show that disruption of the VAPB-PTPIP51 tethers by α-synuclein is accompanied by reductions in IP3 receptor-mediated Ca
2+ delivery to mitochondria. We also show that there are reductions in mitochondrial ATP production in α-synuclein expressing cells which is in line with reduced mitochondrial Ca
2+ levels. However, α-synuclein has been shown to damage a number of mitochondrial proteins that could impact upon ATP production. These include VDAC, mitochondrial ATP synthase and TOM20 [
22,
56,
69]. Thus, there may be several routes including the VAPB-PTPIP51 tethers by which α-synuclein could deleteriously affect mitochondrial function and ATP production, and the precise contributions of each of these are not as yet clear.
Neurons are particularly dependent upon correct Ca
2+ signalling since it is involved in depolarization and synaptic activity [
10]. Neurons also consume large amounts of energy [
50]. Thus changes to Ca
2+ signalling and mitochondrial ATP production are strongly implicated in Parkinson’s disease and other neurodegenerative diseases [
2,
10,
14,
62,
65]. Indeed, elegant molecular modelling studies have shown that even relatively small reductions in mitochondrial ATP production can be sufficient to induce many salient features of neurodegenerative diseases [
51]. The α-synuclein induced disruptions to ER–mitochondria Ca
2+ exchange and mitochondrial ATP production that we describe here are thus likely to be major drivers of disease.
Changes in mitochondrial morphology have been associated with α-synuclein [
8] and there is evidence of mitochondrial “rounding up” and clustering in our α-synuclein expressing cells. Whether such morphological alterations are linked to the changes in ER–mitochondria contacts that we describe are not clear. We did not detect any gross changes to ER morphology. However, several recent studies have shown that ER–mitochondria contact sites regulate mitochondrial biogenesis, division and DNA synthesis [
25,
47,
53]. One possibility is that the effects of α-synuclein on mitochondrial morphology are linked to its function at ER–mitochondria contact sites.
In immunoprecipitation assays, we found that whilst both wild-type and mutant α-synuclein bound to VAPB, the mutants bound slightly stronger. However, in the functional assays involving ER–mitochondria Ca2+ exchange and mitochondrial ATP production we did not detect robust differences between wild-type and mutant α-synuclein. This may be due to the sensitivity of these assays in detecting changes induced by relatively small alterations in binding of wild-type and mutant α-synuclein to VAPB. However, our findings are in line with human disease phenotypes. Triplication of the α-synuclein gene leading to increased expression of wild-type α-synuclein is pathogenic but the familial mutants are not associated with such overexpression. The increased binding of mutant α-synuclein to VAPB, therefore, provides a possible explanation for the similar pathogenic effects of wild-type and mutant α-synuclein involving the VAPB-PTPIP51 tethers. Thus, triplication generates increased α-synuclein which binds to VAPB to disrupt the VAPB-PTPIP51 tethers. Mutant α-synuclein expressed at normal levels binds slightly stronger to VAPB to equally disrupt the VAPB-PTPIP51 tethers.
Two other studies have investigated the effect of α-synuclein on ER–mitochondria associations but the findings are inconsistent [
11,
31]. In particular, one reports that wild-type and mutant α-synuclein reduce whereas the other reports that wild-type increases contacts [
11,
31]. The reasons for these different findings are not clear but both studies included confocal microscopy experiments to quantify ER–mitochondria associations. ER–mitochondria contacts are defined as involving 10–30 nm distances which is at least an order of magnitude beyond the resolution of the confocal microscope; appropriate microscopy methods, therefore, need to be used when quantifying contacts [
16,
37,
55,
63]. Here, we utilized EM, proximity ligation assays and super resolution SIM methods to quantify ER–mitochondria contacts and all revealed that expression of wild-type and mutant α-synuclein decrease ER–mitochondria contacts. Such methods afford better resolution for properly quantifying ER–mitochondria associations. We also found that α-synuclein disrupts binding between the ER–mitochondria tethering proteins VAPB and PTPIP51. Finally, we show that α-synuclein expression perturbs Ca
2+ uptake by mitochondria following IP3 receptor-mediated release from ER stores which is a physiological readout of ER–mitochondria associations. Together, these findings using different but complementary methods and approaches demonstrate that overexpression of α-synuclein disrupts ER–mitochondria contacts.
Recently, Tar DNA-binding protein-43 (TDP-43) and Fused in Sarcoma (FUS), two proteins intimately linked to fronto-temporal dementia and related amyotrophic lateral sclerosis (FTD/ALS) have also been shown to disrupt ER–mitochondria associations [
78,
79]. As is the case with α-synuclein, these effects of TDP-43 and FUS involve breaking of the VAPB-PTPIP51 tethers. However, for TDP-43 and FUS, this breaking involves activation of GSK-3β; GSK-3β is a regulator of the VAPB–PTPIP51 interaction and so controls ER–mitochondria associations and Ca
2+ exchange [
78,
79]. GSK-3β has been linked to α-synuclein and Parkinson’s disease [
27,
54,
57] but we found no evidence that either wild-type or mutant α-synuclein expression caused activation of GSK-3β. Rather, we found that α-synuclein bound directly to VAPB. Thus, there appears to be different routes by which neurodegenerative disease insults can impact upon ER–mitochondria tethering via the VAPB–PTPIP51 interaction, some involving activation of GSK-3β and some such as we describe here for α-synuclein, involving binding to the tethering proteins.
α-Synuclein toxicity has been linked to a number of seemingly diverse pathological features in Parkinson’s disease. These include damage to mitochondria, the ER, axonal transport, autophagy, Ca
2+ homeostasis and lipid metabolism [
9,
11,
13,
22,
31,
32,
38,
59,
60,
68,
71,
84,
88,
90,
91]. Indeed, the difficulty in deciphering α-synuclein toxicity is linking these different pathological changes to a common disease pathway. However, all of these physiological processes are regulated by signalling between ER and mitochondria at MAM [
46,
48,
63,
70,
82,
83]. Thus, our demonstration that α-synuclein binds to VAPB, disrupts the VAPB-PTPIP51 tethers, ER–mitochondria contacts and signalling, represents a plausible route whereby α-synuclein may damage such a variety of cellular functions. These molecular findings facilitate a proper dissection of the role of ER–mitochondria signaling in α-synuclein linked Parkinson’s disease.
Acknowledgements
This work was supported by grants from Parkinson’s UK, Alzheimer’s Research UK, the UK Medical Research Council and The Wellcome Trust. We thank Salman Azhar (Geriatric Research, Education and Clinical Center, Veterans Affairs Palo Alto Health Care System, Palo Alto USA) and Michel Goedert (MRC Laboratory of Molecular Biology, Cambridge UK) for plasmids, and Henry Houlden (UCL Institute of Neurology, London UK) and Katrina Gwinn (National Institute of Neurological Diseases and Stroke, Bethesda USA) for assistance with the generation of iPS cells. We also thank staff within the Centre for Ultrastructural Imaging and the Nikon Imaging Centre at King’s College for assistance with microscopy. Finally, we thank John Hardy (UCL) for his support. The research leading to these results has received funding from the Innovative Medicines Initiative Joint Undertaking under Grant Agreement No. 115439, resources of which are composed of financial contribution from the European Union’s Seventh Framework Programme (FP7/2007-2013) and EFPIA companies in kind contribution.