Given a little coaxing, monomers of α-synuclein huddle into non-toxic, soluble multimers, according to recent findings. As described September 27 in Scientific Reports, arachidonic acid, the most abundant fatty acid in the brain, provides the incentive. Researchers led by David Klenerman at the University of Cambridge in England reported these multimers sport α-helices and dissociate without hassle. Unlike fatty acid-deprived counterparts that form β-sheet-rich fibrils, the multimers appeared benign to neurons. The researchers proposed that in the cramped quarters of the membrane-rich synapse, fatty acids tip the balance from harmful, rigid fibrils of α-synuclein toward the friendlier, soluble species.

“Collectively, these intriguing data describe the in vitro reconstitution of metastable α-helical α-synuclein multimers that have similar properties to those we have described in experiments on neurons, erythrocytes, and other cells,” wrote Dennis Selkoe of Brigham and Women’s Hospital in Boston in a comment to Alzforum (see full comment below).

Like other amyloidogenic proteins, α-synuclein is known for forming fibrils. These form Lewy body inclusions found in people with Parkinson’s disease and other synucleinopathies. However, researchers have long tried to discern what configuration the native soluble protein takes. Over decades a consensus gradually built that the soluble α-synuclein was a monomer lacking secondary structure. In 2011, two studies turned this view upside down. Researchers in Selkoe’s lab, and in the labs of Quyen Hoang of Indiana University School of Medicine in Indianapolis and Thomas Pochapsky, Dagmar Ringe, and Gregory Petsko at Brandeis University in Waltham, Massachusetts, claimed that native α-synuclein predominantly existed as a soluble tetramer with α-helical structure. Further, they proposed that toxic fibrils are only formed by the less common disordered monomers (see Aug 2011 news; Wang et al., 2011). 

Other investigators could not corroborate these findings. A collaboration among six different research groups—including Hilal Lashuel’s at École Polytechnique Fédérale de Lausanne, Switzerland, and Eliezer Masliah’s, then at the University of California, San Diego—maintained that a disordered monomer is the normal physiological form (Feb 2012 news). Still others had a slightly different take. In two separate studies, researchers in Tom Südhof’s lab at Stanford University and Subhojit Roy’s lab at the University of California, San Diego reported that in healthy neurons, α-synuclein resides in synaptic compartments as a multimer that comingles with membranes and helps cluster synaptic vesicles (see Oct 2014 news). What’s more, preventing α-synuclein’s association with membranes promoted its aggregation into insoluble, neurotoxic forms, claimed Südhof and colleagues (see Burré et al., 2015Apr 2015 conference coverage). 

Fat-Free Fibrils. Without ARA (top), a mix of α-synuclein monomers, oligomers, and fibrils form (all in left panel). Centrifugation removes fibrils (supernatant, middle; pellet, right). With ARA, only oligomers form (bottom at two magnifications). [Image courtesy of Iljina et al., Scientific Reports, 2016.]

Now the new evidence. Given α-synuclein’s physiological role within membrane-rich synaptic compartments, first author Marija Iljina and colleagues wondered how α-synuclein would behave in the presence of fatty acids, which are found in cell membranes and can be enzymatically released from the lipid bilayer (see Rossetto et al., 2006). Because arachidonic acid (ARA) is the most abundant fatty acid in the gray matter of the brain, Iljina decided to test it on α-synuclein first. The researchers mixed 35μM of fluorescently labeled, full-length α-synuclein with 1μM ARA, and monitored protein aggregation via fluorescence resonance energy transfer. In the presence of ARA, α-synuclein rapidly oligomerized. An apparent tetramer was the most predominant species, but larger species, ranging up to 20-mers, appeared after a few hours. Without ARA, few multimers formed, and only after vigorous shaking of the sample. Shaking breaks apart growing fibrils to increase the number of nucleation “seeds” for oligomerization. 

To investigate the conformation of α-synuclein in the samples, the researchers used circular dichroism spectroscopy, which detects different secondary structures. This revealed that α-synuclein transformed from a random coil into an α-helical shape shortly after the addition of ARA, and remained in that state for at least 24 hours. In contrast, β-sheets formed in samples lacking the fatty acid. A closer look at the samples under the electron microscope revealed that α-synuclein mixed with the fatty acid formed large oligomers of various shapes and sizes, while α-synuclein without ARA consisted of a mixture of monomers, smaller oligomers, and insoluble fibrils that could be recovered by centrifugation. Iljina speculated that oligomers formed in the absence of the fatty acid were precursors to fibrils.

Interestingly, multimers formed with ARA appeared to be less stable than those formed without ARA—they were easier prey for the proteasome and proteinase K, and also dissociated more readily when diluted. When the fatty acid was removed, larger oligomers dissolved, leaving only smaller ones. These retained their α-helical structure. The researchers speculated this is because they still bound ARA. This could explain why small oligomers, such as the tetramers identified by Selkoe and Hoang, can be isolated from intact cells, Iljina said. Selkoe agreed that was possible.

The researchers found that α-helical, multimeric species also formed at physiological concentrations of α-synuclein (2μM) and ARA (10μM). However, the fatty acid had less effect on the oligomerization of α-synuclein harboring the PD-associated A30P, A53T, or E46K mutations.

These in vitro findings meshed nicely with those from the previous tetramer papers as well as those pointing to α-synuclein multimerization in the presence of synaptic vesicles. Iljina pointed out that the story could be different in the expanse of the cytosol, where fatty acids such as ARA are less abundant. There, it is unclear which species would predominate, although the survival of ARA-formed oligomers after washing suggests they could be present in the cytosol as well as in the synaptic compartments.

Finally, the researchers compared the effects of different α-synuclein oligomers on neurons and microglia. Those generated without ARA triggered the abundant release of reactive oxygen species from cortical neurons, and also caused cell death, while those generated with ARA did neither. The ARA-less oligomers also stimulated a microglial cell line to pump out more TNFα, an inflammatory cytokine.

Although data from Südhof’s and other labs suggested that α-synuclein multimers are primarily associated with membranes, other researchers have found the multimers (in particular, the tetramer) in the cytosol. How might the new data shed light on that discrepancy? ARA, like other membrane-associated fatty acids, can be released from the cell membrane. Hence, cytosolic tetramers may have ARA bound to them. Jacqueline Burré, who reported that α-synuclein multimers help cluster synaptic vesicles while in Südhof’s lab, speculated that α-synuclein, which is highly enriched on synaptic vesicles, may detach from them when arachidonic acid releases from synaptic phospholipids. “This remains to be tested,” she wrote. Burré now runs her own lab at Weill Cornell Medical College in New York.

Selkoe proposed a similar idea, suggesting that tetramers form on membranes but can then detach into the cytosol and can be recovered from there (see Dettmer et al., 2013). “Perhaps in the process of tetramerization and detachment from membranes, a bound lipid component stabilizes the helical multimer, akin to the role of ARA in the in vitro experiments of Iljina et al.,” he wrote.

Pochapsky welcomed further evidence of multimeric α-synuclein. “The concept of α-synuclein as spaghetti soup in the cytosol, in the presence of so many possible nucleating agents and chaperones (including, as we now see, ARA) really needs to give way to the reality that α-synuclein can assume different forms under different conditions,” he wrote (see full comment below).

The idea that α-synuclein exists as a multimer stabilized by fatty acids makes biological sense, commented Petsko, now of Weill Cornell in New York. “We have to find an explanation not only for the fact that a lot people get PD, but also for the fact that most people don’t,” he said. “It makes more sense that α-synuclein can adopt a structure, and that it does something.” Petsko, who recently reported that caspase-1 cleavage of α-synuclein promotes its toxic aggregation, pointed out that there are likely several fatty acids or other molecules capable of stabilizing α-synuclein multimers, as well as multiple pathways that might bungle their formation and cause disease, such as mutations and/or truncations (see Wang et al., 2016). Critics of the α-synuclein tetramer hypothesis did not respond in time for this article.

“The central lesson overall is that altered tetramer-to-monomer equilibria may represent the ultimate upstream event in the initiation of α-synuclein misfolding and human synucleinopathies, coming well before any ‘pathogenic spread’ of abnormal oligomers,” commented Selkoe. “Stabilizing endogenous tetramers/multimers thus presents an entirely new therapeutic approach for PD, dementia with Lewy bodies, multiple system atrophy, and other α-synuclein disorders.”

The strategy has precedent. Tafamidis, approved in the EU for treating familial amyloid polyneuropathy, stabilizes tetramers of transthyretin, preventing the protein from forming fibrils (see Aug 2011 news). 

Ilgina and colleagues plan to look at how other fatty acids affect the structure of α-synuclein, and proposed that the acids could work as therapeutic treatment. However, the solution may not be straightforward. A previous study reported that the configuration of α-synuclein in the presence of another brain fatty acid—docosahexanoic acid (DHA)—depended on concentration. A little DHA stabilized soluble multimers, while too much triggered formation of fibrils (see De Franceschi et al., 2011).—Jessica Shugart

Comments

  1. α-Synuclein Conformation: The Felicity of Helicity

    Over the last 2½ decades, hundreds of papers have been written about the conformational states of α-synuclein, most all of them analyzing the recombinant protein in vitro, with few attempted analyses of physiological αSyn in cells. The universal consensus until five years ago was that αSyn exists biologically as a "natively unfolded" monomer, and this continues to be the broadly held assumption in the field (Fauvet et al., 2012). As a result of this strong conviction, new information suggesting that αSyn can also exist in vivo in α-helically folded multimers, principally sizing as tetramers, has gained limited acceptance. But a growing number of labs has provided evidence for the existence of the latter forms in living cells and sometimes after non-denaturing purification (Bartels et al., 2011; Wang et al., 2011; Dettmer et al., 2013; Westphal and Chandra, 2013). One telling example is the purification of αSyn from human brain under native conditions: a five-step protocol retained some α-helical multimers until the final step of full purification, which then yielded just the 14 κDa monomer (Luth et al., 2015). This result was interpreted as suggesting the loss of an unknown “stabilizing co-factor (e.g., a small lipid)” that keeps tetramers/multimers together within cells and until they are fully purified. Supportive of this interpretation was evidence that helical multimers in cells are immediately depolymerized to monomers upon cell lysis, unless the cells are lysed in a highly concentrated solution that retains ‘molecular crowding’ (see Fig. 5D in Dettmer et al., 2013).

    This elegant new study provides, among other insights, an explanation for the aforementioned findings. The authors used single molecule FRET to show that incubation of the recombinant unfolded monomer with arachidonic acid (ARA) allowed the formation of α-helical multimers whose apparent size distribution peaked at 4 monomer units and whose conformation was α-helical by CD. The ARA-mediated multimers were distinct from oligomers formed by incubating αSyn alone; the authors showed previously and also here that the latter process leads to β-sheet-rich oligomers and fibrils that are neurotoxic. In contrast, the ARA-induced helical multimers resisted fibril formation and were not neurotoxic, and they produced much less pro-inflammatory activation of microglia than the β-sheet-rich oligomers (Fig. 4e,f). The initial experiments were done at supraphysiological concentrations of the two components, but they were repeated at 2 μM aSyn and 10 μM ARA, yielding smaller numbers of α-helical multimers whose apparent size again distributed around a peak of 4 monomer units (Fig. 5b, d). While the latter concentrations are in a physiological range for brain tissue, one needs to confirm that micelles of ARA were not present at 10 μM. However, the authors used 100 κDa spin filters to separate the multimers (which should be tetramers or greater to be retained) from excess free ARA that flowed through, and CD confirmed that soluble α-helical multimers were still present in the retentate (Fig. 5d).

    Collectively, these intriguing data describe the in vitro reconstitution of metastable helical α-synuclein multimers that have closely similar properties to those we have described in experiments on neurons, erythrocytes, and other cells (Bartels et al., 2011; Dettmer et al., 2013), as the authors pointed out. The physiological nature of such assemblies is supported by the findings of Iljina et al. and by our description of highly similar tetramers of β- and γ-synuclein, which are not implicated in PD pathogenesis (Dettmer et al., 2013). In accord, all familial PD missense mutations have been shown to decrease the tetramer:monomer ratio in intact neural cells, and amplifying the E46K mutation by placing it into the two adjacent KTKEGV repeat motifs abrogates multimers, increases free monomers, and leads to neurotoxicity and round cytoplasmic inclusions (Dettmer et al., 2015). Thus, specific residues in the imperfect KTKEGV repeat motifs of aSyn help provide a structural basis for tetramer/multimer formation (Dettmer et al., 2015).

    Together, the emerging information supports a model in which unfolded αSyn monomers, after their synthesis on the ribosome, contact the outsides of vesicles (e.g., synaptic vesicles) and fold into α-helical monomers, as was first found to occur in vitro long ago (Davidson et al., 1998). We hypothesize that the membrane-folded monomers can assemble into an energetically favored tetramer (and perhaps multimers of tetramers) that can detach from the membrane into the aqueous cytosol, where they are principally recovered (Dettmer et al., 2013).  Perhaps in the process of tetramerization and detachment from membranes, a bound lipid component stabilizes the helical multimer, akin to the role of ARA in the in vitro experiments of Iljina et al.

    Regarding function, there is evidence that αSyn multimers in living neurons can occur on the surface of synaptic vesicles, clustering them and attenuating their diffusion to the plasma membrane for exocytosis (Wang et al., 2014). Tetramer-abrogating mutations can prevent this, suggesting that the dynamic equilibrium between helical tetramers and unfolded monomers helps regulate this function. Any genetic or environmental factors that shift the equilibrium toward excess free monomers could both alter this function and predispose to the formation of β-sheet-rich oligomers that would exert cytotoxicity over time.

    Relevant to this debate about αSyn native state, Theillet et al. (Theillet et al., 2016) recently used solution-phase in-cell NMR to report that unfolded, bacterial αSyn, when electroporated into mammalian cells, showed no signs of undergoing membrane interactions, oligomerization, aggregation, targeted degradation, or interaction with other cytosolic proteins. They interpreted this lack of change as indicating that “αS is disordered in mammalian cells.” An explanation for the apparent discrepancy between the various studies mentioned above and Theillet et al. arises from the methods chosen to address which forms of αSyn exist in intact cells. Complexes larger than monomers and membrane-bound forms of αS are difficult to detect via solution NMR due to their lower tumbling rates and significant line broadening, which together cause markedly decreased NMR signal/noise ratios. Thus, it is problematic to use this method to attempt to prove their non-existence. By the authors’ estimate, up to 20 percent of their electroporated αSyn could form multimers that they would not detect by NMR, as also pointed out in an accompanying commentary (Alderson and Bax, 2016). When we followed the electroporation protocol of Theillet et al. but assessed the assembly state of the introduced αSyn by intact-cell crosslinking, we observed αSyn multimers similar to those we have described for endogenous αSyn. We believe that acute electroporation of large amounts of αSyn and analysis by in-cell solution NMR may not be a quantitatively sensitive way to resolve dynamic metastable interactions of αSyn in vivo.

    In summary, intact-cell crosslinking (Bartels et al., 2011; Dettmer et al., 2013), fluorescence protein complementation (Wang et al., 2014; Dettmer et al., 2015), and FRET (Burré et al., 2014) are three independent methods that have been used to observe the existence of αSyn tetramers/multimers in cells. This line of research has now been substantially enhanced by the impressive in vitro reconstitution of α-helical multimers by Iljina et al., providing an exciting advance by which to study their assembly, disassembly and biological properties in much more detail. The central lesson overall is that altered tetramer-to-monomer equilibria may represent the ultimate upstream event in the initiation of αSyn misfolding and human synucleinopathies, coming well before any "pathogenic spread" of abnormal oligomers. Stabilizing endogenous tetramers/multimers thus presents an entirely new therapeutic approach for PD, dementia with Lewy bodies, multiple system atrophy, and other αSyn disorders.

    References:

    . Parkinson's disease: Disorder in the court. Nature. 2016 Feb 4;530(7588):38-9. Epub 2016 Jan 25 PubMed.

    . α-Synuclein occurs physiologically as a helically folded tetramer that resists aggregation. Nature. 2011 Aug 14;477(7362):107-10. PubMed.

    . α-Synuclein assembles into higher-order multimers upon membrane binding to promote SNARE complex formation. Proc Natl Acad Sci U S A. 2014 Oct 7;111(40):E4274-83. Epub 2014 Sep 22 PubMed.

    . Stabilization of alpha-synuclein secondary structure upon binding to synthetic membranes. J Biol Chem. 1998 Apr 17;273(16):9443-9. PubMed.

    . In vivo cross-linking reveals principally oligomeric forms of α-synuclein and β-synuclein in neurons and non-neural cells. J Biol Chem. 2013 Mar 1;288(9):6371-85. Epub 2013 Jan 14 PubMed.

    . Parkinson-causing α-synuclein missense mutations shift native tetramers to monomers as a mechanism for disease initiation. Nat Commun. 2015 Jun 16;6:7314. PubMed.

    . KTKEGV repeat motifs are key mediators of normal α-synuclein tetramerization: Their mutation causes excess monomers and neurotoxicity. Proc Natl Acad Sci U S A. 2015 Aug 4;112(31):9596-601. Epub 2015 Jul 7 PubMed.

    . α-Synuclein in central nervous system and from erythrocytes, mammalian cells, and Escherichia coli exists predominantly as disordered monomer. J Biol Chem. 2012 May 4;287(19):15345-64. Epub 2012 Feb 7 PubMed.

    . Purification of α-synuclein from human brain reveals an instability of endogenous multimers as the protein approaches purity. Biochemistry. 2015 Jan 20;54(2):279-92. Epub 2014 Dec 23 PubMed.

    . Structural disorder of monomeric α-synuclein persists in mammalian cells. Nature. 2016 Feb 4;530(7588):45-50. Epub 2016 Jan 25 PubMed.

    . α-synuclein multimers cluster synaptic vesicles and attenuate recycling. Curr Biol. 2014 Oct 6;24(19):2319-26. Epub 2014 Sep 25 PubMed.

    . A soluble α-synuclein construct forms a dynamic tetramer. Proc Natl Acad Sci U S A. 2011 Oct 25;108(43):17797-802. Epub 2011 Oct 17 PubMed.

    . Monomeric synucleins generate membrane curvature. J Biol Chem. 2013 Jan 18;288(3):1829-40. Epub 2012 Nov 26 PubMed.

  2. The formation of helix-rich α-synuclein (αSyn) multimers in the presence of arachidonic acid (ARA, a common fatty acid found in brain tissue) described by Iljina et al. in their Sci. Rep. article is both welcome further evidence for such species, and, frankly, a little embarrassing for those of us who have previously published experimental evidence for their existence. The authors performed some fairly simple yet elegant experiments showing conclusively that helix-rich, non-toxic, and easily dissociated αSyn multimers form with much different characteristics than the αSyn oligomers formed in the absence of ARA. The first thing that crossed my mind upon reading this article was “Why didn’t I think of that?” The second was “Why are we still arguing about this?” It is admittedly difficult to drop a long-cherished viewpoint, but the concept of αSyn as spaghetti soup in the cytosol, in the presence of so many possible nucleating agents and chaperones (including, as we now see, ARA) really needs to give way to the reality that αSyn can assume different forms under different conditions. Clinging to an all-or-nothing view of αSyn behavior is blinding us to multiple possibilities for PD prophylaxis/treatment.

    One other important aspect of this work to note is that apparently ARA needs to be above the critical micelle concentration in order for the multimers to form, but upon dilution of ARA, they remained relatively stable (there was some increase in oligomer formation, and some loss of material, but the helicity of the remaining multimers was maintained). This indicates (as we have previously suggested) that the multimers are kinetic, rather than thermodynamic, products, and likely the result of interaction between cell membrane components and bound αSyn. Furthermore, it appears that the lipid is a component of the multimers, not just a catalyst for their formation. In other words, the method of introducing αSyn into an experimental system will affect what form of αSyn is observed.    

  3. The paper is interesting and agrees with our previous studies where we reported multimerization of α-helical α-synuclein in the presence of liposomes and brain membranes that mediate SNARE-complex formation, and the protection of membrane-bound α-synuclein from aggregation and thus cellular toxicity (Burré et al., 2014; Burré et al., 2015). Overall, it has been becoming clear in the last couple of years that there are more species of α-synuclein in a neuron than previously anticipated, which exist in various equilibria. Along this line, it would be helpful for the α-synuclein field to distinguish these various α-synuclein species by using separate terms for pathologically relevant species that form β-sheets (e.g., oligomers) versus physiologically functional species that are α-helical and do not aggregate (e.g., we have previously introduced the term “multimers”). In this vein, it is confusing that the authors use the term “aggregate formation” in presence of ARA in Fig. 1a/b/d, but discuss at the same time that α-synuclein in presence of ARA is α-helical and is at least partially protected from fibril formation (Fig. 1c).

    We and others have reported that α-helical multimers of α-synuclein play a role in clustering of synaptic vesicles in the presynaptic terminal (Diao et al., 2013; Wang et al., 2014). The physiological role of ARA-induced multimers of α-synuclein remains to be determined, in particular because α-synuclein is highly enriched on synaptic vesicles and may not encounter much cytosolic ARA under normal physiological conditions. Possibly, α-synuclein multimers may come off synaptic vesicles upon releasing ARA from synaptic vesicle phospholipids, but this remains to be tested.

    A few things I would have loved to see or am surprised about:

    • It would have been interesting to see whether α-synuclein results in clustering of ARA-containing micelles, similar to the reported clustering of synaptic vesicles and synaptic vesicle mimics; it is curious that the oligomer FRET signals in presence of ARA take so long to form (Fig. 3c);
    • The data for A30P is surprising (Fig. 5f), as A30P has been reported by numerous groups to bind to lipid membranes/micelles substantially less than α-synuclein wild-type, A53T, or E46K;
    • We have previously determined the molecular configuration of α-synuclein multimers on membranes using multiple FRET pairs (Burré et al., 2014). It would be interesting to see if ARA induces a configuration similar to that.

    References:

    . α-Synuclein assembles into higher-order multimers upon membrane binding to promote SNARE complex formation. Proc Natl Acad Sci U S A. 2014 Oct 7;111(40):E4274-83. Epub 2014 Sep 22 PubMed.

    . Definition of a molecular pathway mediating α-synuclein neurotoxicity. J Neurosci. 2015 Apr 1;35(13):5221-32. PubMed.

    . Native α-synuclein induces clustering of synaptic-vesicle mimics via binding to phospholipids and synaptobrevin-2/VAMP2. Elife. 2013;2:e00592. PubMed.

    . α-synuclein multimers cluster synaptic vesicles and attenuate recycling. Curr Biol. 2014 Oct 6;24(19):2319-26. Epub 2014 Sep 25 PubMed.

    . α-Synuclein assembles into higher-order multimers upon membrane binding to promote SNARE complex formation. Proc Natl Acad Sci U S A. 2014 Oct 7;111(40):E4274-83. Epub 2014 Sep 22 PubMed.

Make a Comment

To make a comment you must login or register.

References

News Citations

  1. An α-Synuclein Twist—Native Protein a Helical Tetramer
  2. Synuclein—Researchers Out of Sync on Structure
  3. Synuclein Oligomers: Is EnSNAREing Synaptic Vesicles Their True Calling?
  4. Form and Function: What Makes α-Synuclein Toxic?
  5. Amyloid-Blocking Drug Poised for Approval for Rare Disease

Paper Citations

  1. . A soluble α-synuclein construct forms a dynamic tetramer. Proc Natl Acad Sci U S A. 2011 Oct 25;108(43):17797-802. Epub 2011 Oct 17 PubMed.
  2. . Definition of a molecular pathway mediating α-synuclein neurotoxicity. J Neurosci. 2015 Apr 1;35(13):5221-32. PubMed.
  3. . Presynaptic enzymatic neurotoxins. J Neurochem. 2006 Jun;97(6):1534-45. PubMed.
  4. . In vivo cross-linking reveals principally oligomeric forms of α-synuclein and β-synuclein in neurons and non-neural cells. J Biol Chem. 2013 Mar 1;288(9):6371-85. Epub 2013 Jan 14 PubMed.
  5. . Caspase-1 causes truncation and aggregation of the Parkinson's disease-associated protein α-synuclein. Proc Natl Acad Sci U S A. 2016 Aug 23;113(34):9587-92. Epub 2016 Aug 1 PubMed.
  6. . Structural and morphological characterization of aggregated species of α-synuclein induced by docosahexaenoic acid. J Biol Chem. 2011 Jun 24;286(25):22262-74. PubMed.

Further Reading

Papers

  1. . New insights into cellular α-synuclein homeostasis in health and disease. Curr Opin Neurobiol. 2016 Feb;36:15-22. Epub 2015 Aug 15 PubMed.
  2. . Parkinson-causing α-synuclein missense mutations shift native tetramers to monomers as a mechanism for disease initiation. Nat Commun. 2015 Jun 16;6:7314. PubMed.

Primary Papers

  1. . Arachidonic acid mediates the formation of abundant alpha-helical multimers of alpha-synuclein. Sci Rep. 2016 Sep 27;6:33928. PubMed.