Evidence of Native α-Synuclein Conformers in the Human Brain*

Background: α-Synuclein aggregation is associated with Parkinson disease. The native state of α-synuclein in the human brain is unclear. Results: Evidence consistent with the presence of conformationally diverse metastable multimers of α-synuclein in the human brain is presented. Conclusion: Multimeric species of α-synuclein are present within the human brain. Significance: Understanding the native state of α-synuclein in the brain is critical for determining the causes or propensity for aggregation. α-Synuclein aggregation is central to the pathogenesis of several brain disorders. However, the native conformations and functions of this protein in the human brain are not precisely known. The native state of α-synuclein was probed by gel filtration coupled with native gradient gel separation, an array of antibodies with non-overlapping epitopes, and mass spectrometry. The existence of metastable conformers and stable monomer was revealed in the human brain.

␣-Synuclein aggregation is central to the pathogenesis of several brain disorders. However, the native conformations and functions of this protein in the human brain are not precisely known. The native state of ␣-synuclein was probed by gel filtration coupled with native gradient gel separation, an array of antibodies with non-overlapping epitopes, and mass spectrometry. The existence of metastable conformers and stable monomer was revealed in the human brain.
␣-Synuclein is a soluble protein consisting of 140 amino acids with a predicted molecular mass of 14,460. 16 Da and an isoelectric point of 4.67. ␣-Synuclein is expressed in neurons throughout the central nervous system (1,2). The precise functions of ␣-synuclein remain uncertain, although the preferential localization to presynaptic nerve terminals and its interaction with phospholipids and proteins suggest regulatory functions associated with synaptic activity, dopamine metabolism, and lipid vesicle trafficking (3)(4)(5)(6)(7)(8)(9). The steady state levels of ␣-synuclein are regulated by protein degradation pathways, including lysosomal and proteasomal removal (10,11).
Extensive biochemical and biophysical studies have revealed a remarkably dynamic structural flexibility (12)(13)(14)(15)(16)(17)(18)(19). Unique features of the N-terminal peptide sequence (residues 10 -86) are imperfect 11-residue repeats that enable the protein to achieve amphipathic ␣-helical conformation upon interaction with negatively charged phospholipids (3). Residues 61-95 constitute the hydrophobic core of the protein, also known as the non-amyloid-␤ component of Alzheimer disease (NAC) region. This part of the protein allows ␣-synuclein to achieve ␤-sheet conformation that promotes self-aggregation, which under pathological conditions, coalesces to generate amyloid fibrils and pathological inclusions (20,21). The C terminus is rich in negatively charged amino acids and is critical for selfassembly as well as for interacting with other proteins and small molecules (22)(23)(24)(25). Therefore, it is reasonable to assume that various ␣-synuclein conformers (defined as conformationally diverse ␣-synuclein multimers) are formed physiologically in neurons upon different molecular interactions, potentially establishing several equilibria that dictate functional utilization and avoidance of amyloid fibril formation.
A recent debate has centered on the size and folded state of these putative native conformers of ␣-synuclein (26 -32). Evaluation of protein sizes under native states that rely on methods based on first principles such as sedimentation equilibrium ultracentrifugation and scanning transmission electron microscopy (STEM) 2 suggested that ␣-synuclein isolated from fresh human red blood cells or cultured neuroblastoma cells exhibited the properties of a natively folded tetramer rich in ␣-helical content (26). The STEM data, revealed the presence of several ␣-synuclein molecules of different sizes (ranging from monomer to octamers) (26), which was independently corroborated for ␣-synuclein isolated from bacterial systems but with care to preserve natural conformers (28).
These different native conformers have not been recognized in previous studies employing isolated protein from bacterial expression systems with or without heating, in part because the monomeric protein is disordered, and under some analytical methods, it can co-migrate with native multimeric conformers (12,29). ␣-Synuclein from brain tissue of mice, rats, and humans or expressed in mammalian cell lines has displayed similar mobility in native and denaturing gels due to the unfolded monomer produced in Escherichia coli (29).
To investigate the presence of native conformers of ␣-synuclein in the human brain, we employed two analytical methods: gradient native gels and gel filtration coupled with native gradient gel analysis. These methods are not based on first principles for the determination of the size of the protein, but they are advantageous in that they are suitable for complex protein mixtures. Moreover, we confirmed the identity of the ␣-synuclein species by mass spectrometry and employed different extraction methods and lipid removal approaches, as well as an array of antibodies mapped to different epitopes on the protein to explore native conformers.

MATERIALS AND METHODS
Extraction of ␣-Synuclein from Human Brain-Human brain tissue was homogenized in 10 volumes of homogenization * This work was supported, in whole or in part, by National Institutes of Health buffer (150 mM NaCl, 100 mM HEPES, pH 7.4, 10% glycerol, and 0.1% n-octyl-␤-glucopyranoside) with protease inhibitor mixture (Sigma) added. The tissue was homogenized with the aid of a mechanical homogenizer and left on ice for 10 min prior to centrifugation at 14,000 ϫ g for 10 min. The supernatant was retained, and total protein was measured using a BCA protein assay kit (Thermo). To avoid the potential for the detergent to promote ␣-synuclein folding, other extraction buffers included: homogenization buffer without 0.1% n-octyl-␤-glucopyranoside; 50 mM phosphate buffer, pH 7.0; PBS, pH 7.1; 50 mM ammonium acetate, pH 7.4; or 25 mM HEPES/150 mM NaCl, pH 7.25. Native conformation of ␣-synuclein was disrupted by heating, unless otherwise stated, at 55°C for 10 min.
Gel Filtration-Tissue homogenates were freshly prepared, and 1 mg of native or heat-denatured protein was injected onto a Superdex 200 (10/300 GL) column (GE Healthcare) connected to an Agilent 1100 series HPLC. Protein was eluted with 25 mM HEPES, 150 mM NaCl, pH 7.25, mobile phase, and a total of 48 fractions were collected. The fractions corresponding to a hydrodynamic radius greater than 40 Å or less than 30 Å were pooled into four samples each, whereas the fractions that corresponded to a size of 32.3-37.5 Å remained unpooled. All samples were concentrated using Amicon 3-kDa molecular mass cutoff filters (Millipore) prior to native gradient gel electrophoresis. The column was calibrated using a mixture of globular proteins including blue dextran, thyroglobulin, ferritin, catalase, aldolase, albumin, ovalbumin, chymotrypsin, and ribonuclease A.
Sucrose Gradient Ultracentrifugation-For sedimentation, 4 -20% sucrose gradients totaling 5 ml in 10 mM Tris, pH 7.5, were prepared in Beckman Ultra Clear (13 ϫ 51 mm) tubes. 200 l of protein standard or human brain homogenate (1 mg) was layered in buffer and centrifuged at 40,000 rpm for 16 h at 4°C using a Sorvall ultracentrifuge equipped with a Beckman SW 55 Ti swinging bucket rotor. Fractions were collected in 400-l increments, concentrated using 3-kDa cutoff filters (Amicon), and analyzed on standard SDS or native gradient gels. Gels containing the standard proteins were stained with colloidal blue, whereas the native gels consisting of the brain protein homogenate were transferred and blotted for ␣-synuclein (Syn211, 1:1000). Standard proteins were as follows: catalase (11.3 S), aldolase (7.3 S), albumin (4.6 S), ovalbumin (3.5 S), chymotrypsin (2.6 S), and ribonuclease A (2.0 S).
Statistical Analysis-All experiments were performed with an n ϭ 3-5. Data analyzed using GraphPad Prism software (version 5.02) and presented as mean Ϯ S.E. One-way analysis of variance with Tukey's post hoc test was used to determine whether groups differed significantly from control levels; significance was set at p Ͻ 0.05.

RESULTS
To characterize the native conformations of ␣-synuclein, postmortem human brain extracts were separated by gel filtration followed by native gel electrophoresis (Fig. 1A). A corresponding heat-denatured brain extract was analyzed in parallel (Fig. 1B). The brains were homogenized in simple salt buffer and in the absence of detergents or kosmotropic agents. ␣-Synuclein was eluted from gel filtration columns into four main fractions corresponding to Stokes radii of 32.3-37.5 Å for both the native and the heat-denatured brain extracts. The gel filtration fractions were then separated on 4 -16% gradient clear native gels, which allow separation of oligomeric states of proteins with a pI less than 5.4 (33). The heat-denatured human brain ␣-synuclein revealed a single band with an apparent molecular mass of 53 kDa corresponding to the recombinant unfolded monomer. Furthermore sucrose gradient ultracentrifugation determined an average sedimentation coefficient value of 1.0 S for the major heat-insensitive ␣-synuclein species in human brain (Fig. 1D). Additional heat-sensitive species of ␣-synuclein were observed in the natively prepared homogenate corresponding to average values of 1.4 S, 2.9 S, and 3.8 S (Fig. 1C).
In the natively extracted brains, additional immunoreactive bands with apparent higher molecular masses of 57, 62, and 70 kDa were resolved. To confirm that these immunoreactive bands are composed of ␣-synuclein, we employed mass spectroscopy. In-gel trypsin digestion of the individual protein bands followed by mass spectroscopy identified five peptides originating from the N terminus of human ␣-synuclein (table in Fig. 2B). The absence of lysine and arginine residues in the ␣-synuclein C terminus prohibited the identification of trypsin-cleaved peptides from this region. Nonetheless, the five peptides we recovered provided 52% sequence coverage and were identified in all of the immunoreactive bands.
These findings were further confirmed by analyzing fresh homogenates of human brain without pre-fractionation by gel filtration and by using different ␣-synuclein antibodies with non-overlapping epitopes (Fig. 2, A and B) (34). Recombinant protein purified from boiled E. coli extracts and fresh human RBC extracts were used as controls. Recombinant ␣-synuclein migrates to the same apparent molecular mass on clear native gels as the heat-inactivated brain extracts (Fig. 2B), whereas the RBC extracts contain ␣-synuclein conformers with closely similar molecular masses as the brain extracts (Fig. 2B). Three different ␣-synuclein antibodies recognized the ␣-synuclein con-formers, indicating that the conformers contain accessible epitopes spanning the entire ␣-synuclein sequence.
The native conformers of ␣-synuclein are sensitive to heat as both the gel filtration fractionated and the unfractionated human brain extracts showed a single band under clear native gel electrophoresis upon heating (Figs. 1B and 2B). Therefore, we performed a melting point thermostability analysis of the natively extracted human brain fraction (Fig. 2C). The data revealed that ␣-synuclein conformers were stable at temperatures up to 50°C for 10 min, with 70% of human brain-derived FIGURE 1. Conformers of ␣-synuclein in the human brain. A, gel filtration analysis coupled with native gradient gel electrophoresis and blotting with anti-␣-synuclein antibody Syn211 of human brain extracts reveals natively folded conformers (arrowheads). Arrow delineates monomer migration. B, parallel analysis by gel filtration and native gradient gel electrophoresis of heat-denatured brain extracts. Recomb., recombinant protein. C, sucrose gradient ultracentrifugation coupled with native gradient gel electrophoresis of native human brain extracts. D, parallel heat-denatured brain extract. The apparent molecular masses (kDa), stokes radius (Å), and sedimentation coefficient (s 20,w ) of the ␣-synuclein bands were calculated using globular protein standards and a second order polynomial fit. B, Western blots of native (ctrl), heat-denatured (heat), or lipidex treated brain extracts probed using ␣-synuclein antibodies SNL4 (1:1000), Syn121 (1:1000), or Syn211 (1:1000), recombinant protein and fresh RBC lysate used as reference controls for monomer and conformers, respectively. Inset, bands that were analyzed by LC-MS/MS with corresponding peptides that were mapped to human ␣-synuclein. C, quantification of the heat stability of native ␣-synuclein conformers. Data are presented as mean Ϯ S.E., n ϭ 3, ***p Ͻ 0.001. Data presented are representative of three different human brains from subjects without clinical and histological evidence of neurodegeneration or brain disease. ␣-synuclein existing as conformers, whereas the remaining 30% existed as a heat-stable monomer. At temperatures 55°C and higher, a significant downshift in the ratio of ␣-synuclein conformer to monomer occurred, with ␣-synuclein being almost completely monomer above 70°C. These data suggest that the brain ␣-synuclein conformers may represent metastable states distinct from the heat-stable unfolded ␣-synuclein monomer.
Previous studies have documented that inclusion of negatively charged phospholipids, detergents, or molecules that will promote intramolecular hydrogen bonding could significantly increase the ␣-helical content of ␣-synuclein (3,16). Inclusion or exclusion of detergents during the preparation of human brain homogenates followed by analysis on 4 -16% clear native gels showed the presence of the same ␣-synuclein conformers and monomer. Although the inclusion of detergents (n-octyl-␤-glucopyranoside, Triton X-100) improved the extraction efficiency, the sizes and relative proportion of the ␣-synuclein conformers and monomer were not altered, indicating that the ␣-synuclein conformers are not an artifact of detergents and exist in brain tissue under native conditions. Moreover, treatment of brain homogenates with lipidex to remove proteinbound lipids did not significantly decrease the intensity of the ␣-synuclein conformers (Fig. 1D), indicating that the association of ␣-synuclein with phospholipids does not account for the increase in apparent molecular mass on clear native gels of the various ␣-synuclein conformers. Taken together, these data suggest that ␣-synuclein conformers are natively present in the human brain.

DISCUSSION
␣-Synuclein is a highly conserved, abundantly expressed protein, which in part is localized in presynaptic terminals in the CNS (1)(2)(3)(4)(5)(6)(7)(8)(9). For reasons that are not well understood, the protein assembles to highly organized amyloid fibrils and non-amyloid amorphous aggregates that constitute neuronal inclusions in several neurodegenerative disorders, including Parkinson disease (20,21). The precise biological function, the molecular and biochemical triggers of aggregation, and the aggregated forms (oligomers, fibrils) that cause neurotoxicity remain unclear. Recently, the native conformation and assembly state of ␣-synuclein in healthy cells and tissues have come under considerable debate (26 -32). Resolving these issues is critical from the standpoint of therapeutic interventions that could keep ␣-synuclein in its physiological state.
Extensive biochemical studies using protein isolated from bacterial expression systems and protein overexpressed in rodents indicated that ␣-synuclein is a monomer with little secondary structure (12,29). ␣-Synuclein purified from bacterial expression system without heating had an apparent molecular mass of 20 Ϯ 3 kDa measured on SDS containing acrylamide gels, a sedimentation coefficient in sucrose gradient of s 20,w ϭ 1.7 S, and a Stokes radius derived from gel filtration of 34 Å (12). The same Stokes radius was also reported for the human protein expressed in the mouse brain (35). Additional measurements of ␣-synuclein hydrodynamic properties by small angle x-ray scattering indicated that the protein is more compact than that of a random coil (13). The relative compactness of ␣-synuclein was also apparent by NMR studies, indicating that long range interactions allow the formation of ␣-synuclein ensembles that may resist aggregation (14,15,19). The hydrodynamic properties and apparent molecular mass of ␣-synuclein on native gels indicated that the unfolded monomer resembles a globular protein of 57-58 kDa. These observations prompted investigators to re-examine the size of ␣-synuclein using methods such as sedimentation equilibrium analytical centrifugation and STEM that rely on first principles and are suitable for the precise determination of protein size under the native state. When these methods were employed, ␣-synuclein isolated from human red blood cells and human neuroblastoma cells exhibited properties of a natively folded tetramer (26,27). A sedimentation equilibrium value of 4.78 S and a molecular mass of 57,753 Da were determined for ␣-synuclein isolated from human red blood cells. Unbiased counting of the roughly spherical ␣-synuclein molecules revealed a distribution of differently sized particles ranging from 10 to 175 kDa, with a peak at ϳ55 kDa (26). These observations as well as NMR data from ␣-synuclein purified from bacteria under conditions that will preserve ␣-helical content indicated that natively, ␣-synuclein exists in part as a tetramer rich in ␣-helical content (26 -28). This conclusion is not entirely surprising given the recognized structural plasticity of ␣-synuclein that allows the protein to achieve ensembles of different molecular size in solution and within cells.
Because the native state of ␣-synuclein in the human brain is unknown, we employed methodologies suitable for complex protein mixtures to explore the native size of ␣-synuclein in minimally processed human brain tissue. The data reveal the presence of distinct, thermolabile conformers of ␣-synuclein with Stokes radii ranging from 32.3-37.5 Å, sedimentation coefficients ranging from 1.4 S to 3.8 S, and apparent molecular masses of 53, 57, 62, and 70 kDa in gradient native gels. The 53-kDa species corresponds to the unstructured monomer. Heating of the brain extracts above 55°C collapses the higher molecular mass ␣-synuclein conformers into the 53-kDa species.
The human brain ␣-synuclein conformers were recognized by different antibodies with non-overlapping epitopes and were each confirmed by in-gel digestion and mass spectrometric detection of the resulting peptides. The use of 4 -16% gradient native gels was important in separating native conformers from the monomer because separation was not observed when 4 -12% gradient native gels were used, consistent with previously published data (29). However, improved immunodetection was achieved when human brain was first fractionated by gel filtration followed by native gel electrophoresis. Consistent with previous observations, the native conformers collapse to a single species under SDS-PAGE separation (29). The native ␣-synuclein conformers are distinct from pathological oligomeric intermediates detected in the brains of transgenic mice expressing human ␣-synuclein as well as oligomers stabilized by oxidized catechols or oxidative cross-linking (35). We have previously documented that the pathologically relevant oligomers are both SDS-stable and heat-stable and have a Stokes radius of 45-53 Å or larger, which differs from the native conformers that are sensitive to SDS and heat and have smaller Stokes radii (35).
Collectively, the data indicate that metastable ␣-synuclein conformers are present in the human brain. What accounts for these species and why is this important? Although methods are not currently available to characterize these conformers in complex protein mixtures, the emerging data from numerous laboratories have clearly established a conformational plasticity for ␣-synuclein. Many factors such as ionic and non-ionic kosmotropes in the human brain can promote intramolecular hydrogen bonding and partially stabilize these conformers. Negatively charged phospholipids also promote N-terminal ␣-helical content. Other factors may include post-translational modifications such as N-terminal acetylation or protein-protein interactions, which are primarily facilitated by the flexible C terminus. Although potential functional roles for the vesicular-bound ␣-synuclein are emerging, the function of the soluble cytosolic ␣-synuclein conformers is still unclear. Strategies that preserve or stabilize natural ␣-synuclein conformers may provide therapeutic approaches for human brain diseases characterized by prominent aggregation of ␣-synuclein into ␤-sheet rich assemblies that are pathogenic.