Peptide ligand screening of α-synuclein aggregation modulators by in silico panning
© Abe et al; licensee BioMed Central Ltd. 2007
Received: 26 July 2007
Accepted: 16 November 2007
Published: 16 November 2007
α-Synuclein is a Parkinson's-disease-related protein. It forms aggregates in vivo, and these aggregates cause cell cytotoxicity. Aggregation inhibitors are expected to reduce α-synuclein cytotoxicity, and an aggregation accelerator has recently been reported to reduce α-synuclein cytotoxicity. Therefore, amyloid aggregation modulating ligands are expected to serve as therapeutic medicines.
We screened peptide ligands against α-synuclein by in silico panning, a method which we have proposed previously. In this study, we selected as the target a very hydrophobic region known as the amyloid-core-forming region. Since this region cannot be dissolved in water, it is difficult to carry out the in vitro screening of its peptide ligand. We carried out 6 rounds of in silico panning using a genetic algorithm and a docking simulation. After the in silico panning, we evaluated the top peptides screened in silico by in vitro assay. These peptides were capable of binding to α-synuclein.
We demonstrated that it is possible to screen α-synuclein-binding peptides by in silico panning. The screened peptides bind to α-synuclein, thus affecting the aggregation of α-synuclein.
Protein misfolding and aggregation are involved in many diseases, such as Alzheimer's disease, Parkinson's disease and Prion disease, and such proteins accumulate as inclusion bodies in the brain. Lewy bodies are inclusion bodies observed in Parkinson's-disease patients. The major component of the Lewy body is amyloid-like fibrils of α-synuclein . The familial mutants of α-synuclein A53T, A30P and E46K accelerate α-synuclein aggregation and/or fibrillation and cause autosomal-dominant Parkinson's disease [2–4]. These results strongly support the idea that α-synuclein is the pathogenic protein of Parkinson's disease. It is known that α-synuclein is one of the natively unfolded proteins which have little or no ordered secondary structure under physiological condition. However, changes in various environmental factors (e.g., pH, ion strength, agitation) induce the formation of α-synuclein aggregates and amyloid-like fibrils in vitro . Especially, the aggregates called "protofibrils," an intermediate in the fibrillogenesis process, have more cytotoxicity than the amyloid-like fibrils of most of the proteins which generate fibrils . Therefore, aggregation inhibitors are expected to serve as therapeutic medicines, and their effect against several amyloid-forming proteins has been reported [7–11]. However, inhibitors which do not inhibit protofibril formation but amyloid fibril formation, such as L-DOPA, enhance cytotoxicity . Recently, Bodner et al. have reported that aggregation accelerator compounds decrease α-synuclein cytotoxicity . Thus, aggregation accelerators are also expected to serve as therapeutic medicines.
To obtain an effective peptidic aggregation inhibitor, the peptide should bind to the region that plays an important role in amyloid fibril formation. The hydrophobic central region of α-synuclein called the "non-amyloid-β component of the amyloid plaque" (NAC) (residues 61–95) is the most important region in the formation of amyloid-like fibrils. In particular, some of the partial peptides in NAC, such as α-synuclein 68–78 (GAVVTGVTAVA), form amyloid-like fibrils by themselves . We have also reported that a double mutant (V70T and V71T) of α-synuclein does not exhibit aggregation and fibrillation activity. Thus, the binding of peptide ligands to this region should affect α-synuclein aggregation. However, these amyloid-core-forming regions are so hydrophobic that these peptides cannot be dissolved in water. Therefore, computational screening would be useful to screen ligands against such a target.
In related studies, ADAPT and ENPDA have been reported [23, 24]. ADAPT is a program for designing small molecules, and ENPDA is a program for designing peptide ligands that is similar in methodology to our work; both methods were effective in screening ligands. In this study, we carried out in silico panning in order to screen the peptides which bind to the α-synuclein's amyloidogenic region and affect α-synuclein fibril formation.
In silico panning
We selected an 11-mer peptide, α-synuclein 68–78 (GAVVTGVTAVA), as the target region. Although the three-dimensional structure of α-synuclein has been reported (1XQ8 ), this structure is in the Sodium Dodecyl Sulfate (SDS) micelle binding state and its major secondary structure is an α-helix. Under physiological condition, α-synuclein does not form a defined structure. Therefore, the unfolded state of the target 11-mer peptide is better suited as a docking target. Of course, docking simulations against unfolded proteins are challenging. However, amyloid fibrils generally have many β-sheet structures. Since the target region can aggregate by itself and mutants in this region lose their aggregation activity, this region is considered to form β-strands easily. In this study, we hypothesized that it would show a linear motif. We constructed the structure of the target 11-mer peptide using the MOE protein builder.
Since we have succeeded in screening peptide ligands by using a tetramer peptide in the previous study, we selected a tetramer peptide in this study as well. To carry out effective screening using a genetic algorithm, we restricted the type of amino acids composing the peptides. In this study, since the target region was very hydrophobic, we selected the hydrophobic amino acids and the polar amino acids that tended to form hydrogen bonds (glycine, alanine, valine, serine, threonine, proline, glutamine, asparagine, phenylalanine and tyrosine), and the total diversity is 104 = 10,000. However, even a docking simulation of all 10,000 sequences would be difficult because a peptide docking simulation requires more time than the docking simulation of a small compound. We therefore applied a genetic algorithm to the peptide in silico screening.
Ranking of the superior peptides after in silico panning
Binding analysis by surface plasmon resonance
The affinities of the screened peptides were determined by surface plasmon resonance measurement. The screened peptides were immobilized on a CM5 sensor chip by amine coupling. In this study, we evaluated QSTQ, GSQQ, SQTQ and AQTQ which were the top peptides, and PTYF, which was an inferior peptide. The N-terminal glutamine could have formed pyroglutamine, and the N-terminal proline did not have a primary amine. Thus, we added an alanine residue to the N-terminals of these peptides to immobilize them on the CM5 sensor chip and evaluated them. The dissociation constants were determined using a Scatchard plot. We were able to confirm the peptide binding to α-synuclein, and the KD value of QSTQ was determined to be 19 μM. AQTQ also bound to α-synuclein with similar affinity of QSTQ. GSQQ and SQTQ have lower affinity than QSTQ and AQTQ. In the case of PTYF, its KD was more than 10 times that of QSTQ. Since docking simulation still has problems inaccuracy, we have to evaluate several peptides to find good ligands. In this study, we can obtain the high affinity peptide by assaying the top 4 peptides.
Effect of synthetic peptide ligands on α-synuclein aggregation
Peptide docking simulation takes more time than the docking simulation of small molecules because of the structural flexibility of peptides. Therefore, it is difficult to evaluate the peptide library comprehensively by means of a docking simulation. Yoshimori et al. have proposed a stochastic approach to reduce the number of evaluated peptides . Alternatively, we applied a GA to reduce the number of evaluated peptides. We call this method "in silico panning," and have previously reported the screening of a glucose dehydrogenase noncompetitive inhibitor. For the recognition of the target protein, peptides need to have functional amino acids, and the amino-acid sequence is more important than the position of each amino acid. In some cases, the combination of independent important amino acids results in loss of binding affinity. The GA can search the sequence space by optimizing the amino-acid sequence. The GA comprises three steps: selection, crossover and mutation. Through the operation of sequence crossover, it is possible to avoid falling into the local minimum. GAs are used in several fields, especially in informational science, and the results demonstrate the efficiency of GA-based screening. In this study, we evaluated 150 peptides in silico. This represents 1.5% of the total library, and we succeeded in obtaining the peptide that affects α-synuclein aggregation and fibrillation. Therefore, our approach was sufficiently effective for peptide ligand screening. At present, there are many docking simulation softwares, and some of them can perform the calculations more rapidly. Thus, in silico panning has the potential to be applied it to a larger library.
There are some reports of peptides that affect α-synuclein aggregation [7, 27, 28]. These peptides are homologous peptides. Amyloid-forming proteins, including α-synuclein, can form self-aggregates. Therefore, homologous peptides should bind to the proteins, and these peptides have the potential to act as protein aggregation inhibitors. It is known that peptides partially homologous to amyloid β bind to amyloid β and the modified homologous peptide inhibits amyloid β aggregation [9, 10]. Concerning α-synuclein, El-Agnaf et al. have reported that peptides partially homologous in the NAC region, especially the α-synuclein 60–80 region, bind to α-synuclein across its full length, and that these peptides, modified to dissolve in water, inhibit α-synuclein aggregation. Therefore, if the peptides binding these regions can be screened, they could be used to inhibit α-synuclein aggregation.
While there are many reports of aggregation inhibitors, there are some reports of aggregation accelerator peptides [7, 27, 28]. These are peptides homologous to the NAC region. Surprisingly, El-Agnaf et al. have reported that a peptide designed to inhibit α-synuclein aggregation actually accelerated the aggregation. This peptide lacks one amino acid from the α-synuclein aggregation inhibitor peptide sequence. The protein fibrillation process is assumed to comprise three steps: conformational change, the formation of an amyloid core through self-aggregation, and the elongation of amyloid fibrils. The amyloid-core-forming region is highly hydrophobic. Thus, this region should be isolated to some extent from the solvent. This region may become exposed to the solvent upon the conformational change resulting from the addition of certain factors, such as ligands, which might induce the aggregation of these regions.
The ligand with strong binding activity may inhibit aggregation, which is followed by conformational change. Since amyloid fibrils have a very stable conformation, it is thought that the binding ability of self-aggregation is very strong. Weakly bound ligands would not sufficiently inhibit this aggregation followed by conformational change. Therefore, peptides lacking some of the amino acids of the inhibitor peptide sequence would accelerate α-synuclein aggregation and induce fibrillation. In this study, the screened peptides may have bound to the hydrophobic core region, but the binding ability was not high. These peptides might induce conformational changes by binding to the hydrophobic region, and accelerate the α-synuclein aggregation and subsequent fibrillation.
We demonstrated that in silico panning can be used to screen α-synuclein-binding peptides. Although α-synuclein does not have a defined structure, we were able to screen the peptides by hypothesizing that the target region forms a β-sheet structure. The screened peptides showed low dissociation constants (for example, KD = 19 μM) and affected α-synuclein aggregation and fibrillation. These peptides accelerated α-synuclein aggregation and fibrillation, and may be able to decrease α-synuclein cytotoxicity by decreasing the protofibril amount.
Since our in silico panning approach has a great potential and is applicable to a larger library, it may be useful for the ligand screening of certain disease-related proteins.
Construction of the initial generation and target
We designed the peptides of the initial generation randomly, without bias. Since the target peptide, α-synuclein 68–78, is very hydrophobic, we expected that the peptide ligands would bind to the target peptide through hydrophobic interaction and hydrogen bonding. To avoid an excessively large sequence space, we selected 10 amino acids: glycine, alanine, valine, serine, threonine, proline, glutamine, asparagine, phenylalanine and tyrosine. The molecular operating environment (MOE) from the Chemical Computing Group, Inc. was used for the model building. After the three-dimensional construction of these peptides, in order to optimize these peptides' local conformations, these structures were subjected to an MMFF94 energy minimization protocol until the root mean square of the conjugate gradient reached <0.05 kcal mol-1 Å-1. The target peptide, α-synuclein 68–78, was constructed in the same way.
The docking simulation described here was carried out using the MOE-Dock default parameters based on the simulated annealing protocol. We selected the MMFF94s force field. The MOE suite utilized here includes an implemented version of the Generalized Born/Surface Area contact function that models the electrostatic contribution to the free energy of the solvation in a continuous solvent model. A docking box was constructed around α-synuclein 68–78. We carried out the MOE-Dock 30 run in our case.
We implemented a simple genetic algorithm to produce the next generation. After the docking simulation, the peptides were ranked by their calculated docking energy. The more or less superior peptide sequences were duplicated, and the sequences were recombined at a 1.0 crossover rate. Finally, single-amino-acid mutations were introduced into the peptides at a mutation rate of 5 to 20%. Population sizes of each generation are selected between 20 and 25 according to number of superior peptide.
Expression and purification of α-synuclein
Human wild-type α-synuclein expressed in Escherichia coli BL21(DE3) cells transformed the pET28a/human α-synuclein WT plasmid. Following induction with 1 mM isopropyl-1-thio-β-D-galactopyranoside, the α-synuclein was purified by anion exchange chromatography . The bacterial cells were harvested by centrifugation at 5,000 g for 10 min and lysed in 500 mM NaCl, 20 mM Tris-HCl, pH 8.0, and 1 mM EDTA. The lysate was then boiled for 20 min, and centrifuged for 10 min at 20,000 g. The supernatant was dialyzed against 20 mM Tris-HCl, pH 8.0, loaded onto a ResourseQ 6-ml column (Amersham Biosciences) equilibrated in the same buffer, and eluted with a 0 ~1 M NaCl gradient. The α-synuclein-containing fraction was dialyzed against 10 mM Tris-HCl, pH 7.4.
Peptide synthesis and purification
All peptides were synthesized using 9-fluorenylmethyoxycarbonyl (Fmoc) chemistry and a solid phase system using the peptide synthesizer PSSM8 (Shimadzu Co., Ltd., Kyoto, Japan). The synthesis resin, TentaGel S PHB-coupled Fmoc-amino acid, was purchased from HiPep Laboratories. Deprotection/cleavage from the resin was performed through treatment with a 90:5:5 (v/v/v) trifluoroacetic acid (TFA)/water/triisopropylsilane mixture for 2 h. Since the glutamine at the N-terminal could undergo a side-reaction, we conducted the acetylation of the N-terminal amine. The synthesized peptides were purified by reverse-phase liquid chromatography using a C18 ODS-column (TOSOH). The mass weights were confirmed by ESI-MS (JEOL).
Surface plasmon resonance
All peptides were immobilized onto the CM5 sensor chip by amine coupling. The CM5 sensor chip (Biacore) was activated by an equimolar mix of NHS (N-hydroxysuccinimide) and EDC (N-ethyl-N'-(dimethylaminopropyl) carbodiimide), coupled with 100 μg/mL of the synthesized peptide in pH-6.0 sodium acetate buffer, and then blocked with ethanolamine. For reference, only activation and protection was carried out, in the same way as mentioned above. The final coupling level of each peptide was approximately 100 RU. Each of the α-synuclein solutions was injected, and at different concentrations a signal increase was observed. The sensor surface was regenerated using 0.5% SDS. The dissociation constants were determined using a Scatchard plot and an equilibrium point.
Fibril formation and analysis
All the protein samples were concentrated to approximately 4.0 mg/ml by means of Amicon Ultra-15 filters (Millipore) in 10 mM Tris-HCl, pH 7.4, and centrifuged at 150,000 g for 60 min to remove the insoluble aggregates. The peptides were precipitated in the same buffer and mixed with α-synuclein. The α-synuclein and peptide mixtures were incubated in 0.5 ml of the same buffer with 0.02% NaN3 at 37°C by shaking in 1.5 ml tubes. The protein concentrations were kept at approximately 2.0 mg/ml. Fibril formation was monitored by thioflavin T (TfT) fluorescence: 10 μl aliquots were sampled from the incubated samples and added to 1.0 ml of 25 μM TfT in 10 mM Tris buffer, pH 7.4. Fluorescence emission was measured at 485 nm at an excitation wavelength of 450 nm on a FP6500 spectrofluorometer (JASCO). The Rayleigh scattering at 330 nm was used to monitor the total aggregation.
We thank Dr. Chikashi Nakamura and Ms. Nao Kaneko of the Research Institute for Cell Engineering, the National Institute of Advanced Industrial Science and Technology, for the peptide synthesizer and their valuable advice on peptide synthesis.
- Spillantini MG, Crowther RA, Jakes R, Hasegawa M, Goedert M: alpha-Synuclein in filamentous inclusions of Lewy bodies from Parkinson's disease and dementia with lewy bodies. Proc Natl Acad Sci U S A 1998, 95(11):6469–6473. 10.1073/pnas.95.11.6469PubMed CentralView ArticlePubMedGoogle Scholar
- Conway KA, Harper JD, Lansbury PT: Accelerated in vitro fibril formation by a mutant alpha-synuclein linked to early-onset Parkinson disease. Nat Med 1998, 4(11):1318–1320. 10.1038/3311View ArticlePubMedGoogle Scholar
- Conway KA, Lee SJ, Rochet JC, Ding TT, Williamson RE, Lansbury PT Jr.: Acceleration of oligomerization, not fibrillization, is a shared property of both alpha-synuclein mutations linked to early-onset Parkinson's disease: implications for pathogenesis and therapy. Proc Natl Acad Sci U S A 2000, 97(2):571–576. 10.1073/pnas.97.2.571PubMed CentralView ArticlePubMedGoogle Scholar
- Greenbaum EA, Graves CL, Mishizen-Eberz AJ, Lupoli MA, Lynch DR, Englander SW, Axelsen PH, Giasson BI: The E46K mutation in alpha-synuclein increases amyloid fibril formation. J Biol Chem 2005, 280(9):7800–7807. 10.1074/jbc.M411638200View ArticlePubMedGoogle Scholar
- Hashimoto M, Hsu LJ, Sisk A, Xia Y, Takeda A, Sundsmo M, Masliah E: Human recombinant NACP/alpha-synuclein is aggregated and fibrillated in vitro: relevance for Lewy body disease. Brain Res 1998, 799(2):301–306. 10.1016/S0006-8993(98)00514-9View ArticlePubMedGoogle Scholar
- Bucciantini M, Giannoni E, Chiti F, Baroni F, Formigli L, Zurdo J, Taddei N, Ramponi G, Dobson CM, Stefani M: Inherent toxicity of aggregates implies a common mechanism for protein misfolding diseases. Nature 2002, 416(6880):507–511. 10.1038/416507aView ArticlePubMedGoogle Scholar
- El-Agnaf OM, Paleologou KE, Greer B, Abogrein AM, King JE, Salem SA, Fullwood NJ, Benson FE, Hewitt R, Ford KJ, Martin FL, Harriott P, Cookson MR, Allsop D: A strategy for designing inhibitors of alpha-synuclein aggregation and toxicity as a novel treatment for Parkinson's disease and related disorders. Faseb J 2004, 18(11):1315–1317.PubMedGoogle Scholar
- Horiuchi M, Baron GS, Xiong LW, Caughey B: Inhibition of interactions and interconversions of prion protein isoforms by peptide fragments from the C-terminal folded domain. J Biol Chem 2001, 276(18):15489–15497. 10.1074/jbc.M100288200View ArticlePubMedGoogle Scholar
- Soto C, Sigurdsson EM, Morelli L, Kumar RA, Castano EM, Frangione B: Beta-sheet breaker peptides inhibit fibrillogenesis in a rat brain model of amyloidosis: implications for Alzheimer's therapy. Nat Med 1998, 4(7):822–826. 10.1038/nm0798-822View ArticlePubMedGoogle Scholar
- Tjernberg LO, Naslund J, Lindqvist F, Johansson J, Karlstrom AR, Thyberg J, Terenius L, Nordstedt C: Arrest of beta-amyloid fibril formation by a pentapeptide ligand. J Biol Chem 1996, 271(15):8545–8548. 10.1074/jbc.271.15.8545View ArticlePubMedGoogle Scholar
- Nagai Y, Tucker T, Ren H, Kenan DJ, Henderson BS, Keene JD, Strittmatter WJ, Burke JR: Inhibition of polyglutamine protein aggregation and cell death by novel peptides identified by phage display screening. J Biol Chem 2000, 275(14):10437–10442. 10.1074/jbc.275.14.10437View ArticlePubMedGoogle Scholar
- Conway KA, Rochet JC, Bieganski RM, Lansbury PT Jr.: Kinetic stabilization of the alpha-synuclein protofibril by a dopamine-alpha-synuclein adduct. Science 2001, 294(5545):1346–1349. 10.1126/science.1063522View ArticlePubMedGoogle Scholar
- Bodner RA, Outeiro TF, Altmann S, Maxwell MM, Cho SH, Hyman BT, McLean PJ, Young AB, Housman DE, Kazantsev AG: Pharmacological promotion of inclusion formation: a therapeutic approach for Huntington's and Parkinson's diseases. Proc Natl Acad Sci U S A 2006, 103(11):4246–4251. 10.1073/pnas.0511256103PubMed CentralView ArticlePubMedGoogle Scholar
- El-Agnaf OM, Irvine GB: Aggregation and neurotoxicity of alpha-synuclein and related peptides. Biochem Soc Trans 2002, 30(4):559–565. 10.1042/BST0300559View ArticlePubMedGoogle Scholar
- Ikebukuro K, Okumura Y, Sumikura K, Karube I: A novel method of screening thrombin-inhibiting DNA aptamers using an evolution-mimicking algorithm. Nucleic Acids Res 2005, 33(12):e108. 10.1093/nar/gni108PubMed CentralView ArticlePubMedGoogle Scholar
- Noma T, Ikebukuro K: Aptamer selection based on inhibitory activity using an evolution-mimicking algorithm. Biochem Biophys Res Commun 2006, 347(1):226–231. 10.1016/j.bbrc.2006.06.083View ArticlePubMedGoogle Scholar
- Yohei Yokobayashi, Kazunori Ikebukuro, Scott McNiven, Karbe : Directed evolution of trypsin inhibiting peptides using a genetic algorithm. J Chem Soc Perkin Trans 1 1996, 20: 2435–2437. 10.1039/p19960002435View ArticleGoogle Scholar
- Zhang W, Loughran MG, Kanna S, Yano K, Ikebukuro K, Yokobayashi Y, Kuroda R, Karube I: Exploration of structural features of monomeric helical peptides designed with a genetic algorithm. Proteins 2003, 53(2):193–200. 10.1002/prot.10509View ArticlePubMedGoogle Scholar
- Li H, Li C, Gui C, Luo X, Chen K, Shen J, Wang X, Jiang H: GAsDock: a new approach for rapid flexible docking based on an improved multi-population genetic algorithm. Bioorg Med Chem Lett 2004, 14(18):4671–4676. 10.1016/j.bmcl.2004.06.091View ArticlePubMedGoogle Scholar
- Jasbir Singh, Mark A. Ator, Edward P. Jaeger, Martin P. Allen, David A. Whipple, James E. Soloweij, Swapan Chowdhary, Treasurywala AM: Application of Genetic Algorithms to Combinatorial Synthesis:A Computational Approach to Lead Identification and Lead Optimization. J Am Chem Soc 1996, 118: 1669–1676. 10.1021/ja953172iView ArticleGoogle Scholar
- Weber L, Wallbaum S, Broger C, Gubernator K: Optimization of the Biological Activity of Combinatorial Compound Libraries by a Genetic Algorithm. Angew Chem Int Ed Engl 1995, 34: 2280–2282. 10.1002/anie.199522801View ArticleGoogle Scholar
- Yagi Y, Terada K, Noma T, Ikebukuro K, Sode K: In silico panning for a non-competitive peptide inhibitor. BMC Bioinformatics 2007, 8: 11. 10.1186/1471-2105-8-11PubMed CentralView ArticlePubMedGoogle Scholar
- Pegg SC, Haresco JJ, Kuntz ID: A genetic algorithm for structure-based de novo design. J Comput Aided Mol Des 2001, 15(10):911–933. 10.1023/A:1014389729000View ArticlePubMedGoogle Scholar
- Belda I, Madurga S, Llora X, Martinell M, Tarrago T, Piqueras MG, Nicolas E, Giralt E: ENPDA: an evolutionary structure-based de novo peptide design algorithm. J Comput Aided Mol Des 2005, 19(8):585–601. 10.1007/s10822-005-9015-1View ArticlePubMedGoogle Scholar
- Ulmer TS, Bax A, Cole NB, Nussbaum RL: Structure and dynamics of micelle-bound human alpha-synuclein. J Biol Chem 2005, 280(10):9595–9603. 10.1074/jbc.M411805200View ArticlePubMedGoogle Scholar
- Yoshimori A, Takasawa R, Tanuma S: A novel method for evaluation and screening of caspase inhibitory peptides by the amino acid positional fitness score. BMC Pharmacol 2004, 4: 7. 10.1186/1471-2210-4-7PubMed CentralView ArticlePubMedGoogle Scholar
- Du HN, Li HT, Zhang F, Lin XJ, Shi JH, Shi YH, Ji LN, Hu J, Lin DH, Hu HY: Acceleration of alpha-synuclein aggregation by homologous peptides. FEBS Lett 2006, 580(15):3657–3664. 10.1016/j.febslet.2006.05.050View ArticlePubMedGoogle Scholar
- Kuroda Y, Maeda Y, Hanaoka H, Miyamoto K, Nakagawa T: Oligopeptide-mediated acceleration of amyloid fibril formation of amyloid beta(Abeta) and alpha-synuclein fragment peptide (NAC). J Pept Sci 2004, 10(1):8–17. 10.1002/psc.485View ArticlePubMedGoogle Scholar
- Kobayashi M, Kim J, Kobayashi N, Han S, Nakamura C, Ikebukuro K, Sode K: Pyrroloquinoline quinone (PQQ) prevents fibril formation of alpha-synuclein. Biochem Biophys Res Commun 2006, 349(3):1139–1144. 10.1016/j.bbrc.2006.08.144View ArticlePubMedGoogle Scholar
This article is published under license to BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.