α-Synuclein-mediated inhibition of ATF6 processing into COPII vesicles disrupts UPR signaling in Parkinson's disease

Highlights

• α-Synuclein inhibits efficient signaling by the unfolded protein response (UPR).

• α-Synuclein reduces ATF6 signaling through direct physical association.

• α-Synuclein inhibits ATF6 incorporation into COPII vesicles

• Aberrant UPR signaling is a salient feature of Parkinson's disease (PD).

• Study provides a novel cellular mechanism of how ER stress and the UPR contribute to neurodegeneration in PD.

Abstract

The unfolded protein response (UPR) monitors the folding environment within the endoplasmic reticulum (ER). Accumulation of misfolded proteins within the ER activates the UPR resulting in the execution of adaptive or non-adaptive signaling pathways. α-Synuclein (α-syn) whose accumulation and aggregation define the pathobiology of Parkinson's disease (PD) has been shown to inhibit ER–Golgi transit of COPII vesicles. ATF6, a protective branch of the UPR, is processed via COPII mediated ER–Golgi transit following its activation via ER stress. Using cellular PD models together with biochemical reconstitution assays, we showed that α-syn inhibited processing of ATF6 directly through physical interactions and indirectly through restricted incorporation into COPII vesicles. Impaired ATF6 signaling was accompanied by decreased ER-associated degradation (ERAD) function and increased pro-apoptotic signaling. The mechanism by which α-syn inhibits ATF6 signaling expands our understanding of the role ER stress and the UPR play in neurodegenerative diseases such as PD.

Introduction

Parkinson's disease (PD) is characterized by the abnormal accumulation and aggregation of α-synuclein, which leads to the selective loss of dopaminergic neurons (DA-neurons) (Braak et al., 2003). Accumulation of α-synuclein (α-syn^WT) in idiopathic PD or expression of the missense point mutants (E46K, A30P, A53T) in familial PD is strongly associated with age of disease onset and severity (Narhi et al., 1999). The mechanism of how α-syn disturbs cellular function leading to PD pathology (e.g., loss of DA-neurons, Lewy bodies) remains a topic of ongoing investigation. α-Syn has been shown to localize to and accumulate within various cellular compartments and organelles such as the nucleus, mitochondria, and ER. This pattern of widespread localization by α-syn is surprising given the protein's lack of a nuclear localization signal, mitochondrial import signal, signal sequence, or an ER-retention signal such as KDEL. (Devi et al., 2008, Maroteaux et al., 1988). Of particular importance to this present study are recent findings examining human PD tissue and PD animal models, which showed α-syn accumulation within the ER lumen (Colla et al., 2012a, Colla et al., 2012b, Oaks et al., 2013).

Previous studies have demonstrated that α-syn over-expression strongly inhibited ER–Golgi transport of COPII vesicles, thereby disrupting normal transit and processing of proteins within the secretory pathway. α-Syn-mediated inhibition on secretory function has been verified using a variety of secretory substrates and reporters such as the human Dopamine Transporter (hDAT) and the GFP-tagged viral glycoprotein, VSVG (Oaks et al., 2013, Thayanidhi et al., 2010). Several studies have found that α-syn^WT or the A53T (α-syn^A53T) variant (herein collectively referred to as α-syn) inhibited SNARE and RAB function leading to declines in ER–Golgi transit of COPII vesicle while over-expression of several secretory accessory proteins (Ykt6, Ypt1p, Sec22b, RAB1, RAB3a, and RAB8A) reduced α-syn dependent toxicities in cellular PD models (Cooper et al., 2006, Gitler et al., 2008, Sancenon et al., 2012, Thayanidhi et al., 2010). Chemical screens have led to the discovery of small molecules, which increased ER–Golgi transport, thereby decreasing cytotoxicity associated with α-syn over-expression (Su et al., 2011). These studies collectively support the detrimental effects that α-syn has on proper secretory function and provide motivation for the discovery and biochemical elucidation of how such effects may be causal in the premature cell death of α-syn over-expressing neurons such as those found in the SN of PD patients.

The unfolded protein response (UPR) serves as a sentinel of protein folding within the ER, maintaining both the quantity and quality of protein folding (Ron and Walter, 2011). Upon activation, the UPR orchestrates cellular changes in transcription, translation, and protein degradation. These changes are mediated through the three canonical sensors of the metazoan UPR; IRE1, ATF6, and PERK. Each sensor detects misfolded proteins via their ER luminal domains and signals through their respective cytoplasmic domains. Upon activation, full-length ATF6 (ATF6^FL) transits to the Golgi via COPII vesicular transport, where sequential cleavage by regulated intramembrane proteolysis (RIP) occurs, releasing the soluble cytosolic domain (aa1-373), a basic leucine zipper (bZip) transcription factor (ATF6^N) (Haze et al., 1999). Activated IRE1 proceeds to cleave the bound unspliced XBP1 mRNA (XBP1^U), to produce the mature transcription factor, XBP1^S (Yoshida et al., 2001). Cooperation between the ATF6 and IRE1-XBP1 signaling pathways occurs through the heterodimerization of the transcriptional domains inducing strong expression of genes necessary for maintaining ER folding homeostasis, including ER resident chaperones, ER associated degradation [ERAD] genes, and genes involved in regulation of the secretory pathway (Yamamoto et al., 2007, Yoshida et al., 2003). Increased induction of UPR-dependent genes is one component of the Integrated Stress Response (ISR), during which the individual branches of the UPR work in concert to promote cellular recovery and survival (Ron and Walter, 2007).

While there is accumulating evidence supporting involvement of the UPR in the pathobiology of PD, it remains unclear if UPR activation and signaling are protective, benign, or causative in PD (Bellucci et al., 2011, Gorbatyuk et al., 2012, Hoozemans and Scheper, 2012, Hoozemans et al., 2007, Nijholt et al., 2012, Valdés et al., 2014). Moreover, interactions between α-syn and components of the UPR remain unexplored. Analysis of PD tissue and PD models has focused mainly on observations of the PERK branch, with results showing increased expression of the PERK downstream targets, p-eIF2α, CHOP, and ATF4 (Bellucci et al., 2011, Colla et al., 2012a, Colla et al., 2012b, Gorbatyuk et al., 2012, Holtz and O'Malley, 2003, Hoozemans and Scheper, 2012, Hoozemans et al., 2007, Mutez et al., 2014, Nijholt et al., 2012). As previously mentioned, several studies have shown that α-syn inhibits COPII ER–Golgi transport, a pathway necessary for activation and signaling by ATF6. α-Syn has also been shown to form aggregates within the ER lumen suggesting that cytotoxicity may occur through a combination of α-syn localization to the ER and cytoplasm (Colla et al., 2012a, Colla et al., 2012b, Oaks et al., 2013). Other studies have provided evidence supporting the protective role of ATF6 in preventing degeneration of DA-neurons in chemically induced PD models and in the protection of cells from chronic ER stress (Egawa et al., 2011, Hashida et al., 2012, Wu et al., 2007). Therefore, diminished ATF6-processing through the effects of ER and cytoplasmic localized α-syn inhibition on COPII ER–Golgi transport would ultimately disrupt signal integration among the UPR branches through reduced cooperation between ATF6^N and XBP1^S. Collectively, these and other studies provide compelling evidence that accumulation of α-syn would have deleterious effects on the ability of the UPR to induce and maintain proper folding programs within the ER lumen while inhibiting the ability of the ER to efficiently remove misfolded proteins through ER-quality control measures such as ERAD.

In this report we demonstrate the ability of α-syn to disrupt UPR signaling through a combination of ER luminal and cytoplasmic insults. First, we show that GRP78, an ER resident chaperone, and general marker of ER stress levels is increased in DA-neurons of the SN from PD patient samples when compared to age-matched control tissue. Increased expression of GRP78 in PD is contradicted by the accompanying observation of decreased expression levels of ATF6^FL and decreased nuclear localization of the cytosolic signaling domain, ATF6^N. Second, we demonstrated that α-syn over-expression results in a decay of downstream UPR signaling pathways such as ERAD. Third, biophysical studies showed that α-syn formed a stable detergent-resistant complex with ATF6^FL. Lastly, biochemical reconstitution of COPII budding showed that transient over-expression of α-syn reduced ATF6^FL incorporation into COPII vesicles and caused deviations from the normal biophysical traits of such vesicles with observed increases in vesicle sizes and alterations to vesicular morphology. These data support the conclusions that α-syn over-expression and accumulation results in inhibition of ATF6 processing during periods of ER stress thus preventing execution of adaptive signaling programs, with the affected cell becoming increasingly incapable of adapting to ER stress, resulting in the premature death of selective neuronal populations.