Unformatted Attachment Preview
Small molecule inhibits α-synuclein aggregation,
disrupts amyloid fibrils, and prevents degeneration
of dopaminergic neurons
Jordi Pujolsa,b,1, Samuel Peña-Díaza,b,1, Diana F. Lázaroc,d,e, Francesca Peccatif,g, Francisca Pinheiroa,b, Danilo Gonzálezf,
Anita Carijaa,b, Susanna Navarroa,b, María Conde-Giménezh,i, Jesús Garcíaj, Salvador Guardiolaj, Ernest Giraltj,k,
Xavier Salvatellaj,l, Javier Sanchoh,i, Mariona Sodupef,l, Tiago Fleming Outeiroc,d,e,m,n, Esther Dalfób,o,p,
and Salvador Venturaa,b,l,2
Institut de Biotecnologia i Biomedicina, Universitat Autònoma de Barcelona, 08193 Bellaterra, Spain; bDepartament de Bioquímica i Biologia Molecular,
Universitat Autònoma de Barcelona, 08193 Bellaterra, Spain; cDepartment of Experimental Neurodegeneration, University Medical Center Göttingen,
37073 Göttingen, Germany; dCenter for Biostructural Imaging of Neurodegeneration, University Medical Center Göttingen, 37073 Göttingen, Germany;
Center for Nanoscale Microscopy and Molecular Physiology of the Brain, University Medical Center Göttingen, 37073 Göttingen, Germany; fDepartament
de Química, Universitat Autònoma de Barcelona, 08193 Bellaterra, Spain; gLaboratoire de Chimie Théorique, Sorbonne Universités, CNRS, F-75005 Paris,
France; hDepartment of Biochemistry and Molecular and Cell Biology, University of Zaragoza, 50018 Zaragoza, Spain; iInstitute for Biocomputation and Physics
of Complex Systems (BIFI), University of Zaragoza, 50018 Zaragoza, Spain; jInstitute for Research in Biomedicine (IRB Barcelona), The Barcelona Institute of
Science and Technology (BIST), 08028 Barcelona, Spain; kDepartment of Inorganic and Organic Chemistry, University of Barcelona, 08028 Spain; lInstitució
Catalana de Recerca i Estudis Avançats (ICREA), 08010 Barcelona, Spain; mMax Planck Institute for Experimental Medicine, 37075 Göttingen, Germany; nInstitute
of Neuroscience, The Medical School, Newcastle University, Newcastle upon Tyne NE2 4HH, United Kingdom; oFaculty of Medicine, University of Vic-Central
University of Catalonia (UVic-UCC), 08500 Vic, Spain; and pInstitut de Neurociències, Universitat Autònoma de Barcelona, 08193 Bellaterra, Spain
Parkinson’s disease (PD) is characterized by a progressive loss of
dopaminergic neurons, a process that current therapeutic approaches
cannot prevent. In PD, the typical pathological hallmark is the accumulation of intracellular protein inclusions, known as Lewy bodies
and Lewy neurites, which are mainly composed of α-synuclein. Here,
we exploited a high-throughput screening methodology to identify a
small molecule (SynuClean-D) able to inhibit α-synuclein aggregation.
SynuClean-D significantly reduces the in vitro aggregation of wildtype α-synuclein and the familiar A30P and H50Q variants in a substoichiometric molar ratio. This compound prevents fibril propagation
in protein-misfolding cyclic amplification assays and decreases the
number of α-synuclein inclusions in human neuroglioma cells. Computational analysis suggests that SynuClean-D can bind to cavities in
mature α-synuclein fibrils and, indeed, it displays a strong fibril disaggregation activity. The treatment with SynuClean-D of two PD Caenorhabditis elegans models, expressing α-synuclein either in muscle or in
dopaminergic neurons, significantly reduces the toxicity exerted by
α-synuclein. SynuClean-D–treated worms show decreased α-synuclein
aggregation in muscle and a concomitant motility recovery. More importantly, this compound is able to rescue dopaminergic neurons from
α-synuclein–induced degeneration. Overall, SynuClean-D appears to
be a promising molecule for therapeutic intervention in Parkinson’s
Parkinson’s disease α-synuclein protein aggregation
inhibition dopaminergic degeneration
Interfering with α-Syn aggregation has been envisioned as a
promising disease-modifying approach for the treatment of PD
(1). However, the disordered nature of α-Syn precludes the use of
structure-based drug design for the discovery of novel molecules
able to modulate α-Syn aggregation. Therefore, many efforts have
focused on the analysis of large collections of chemically diverse
molecules to identify lead compounds (10). Recently, we have
developed an accurate and robust high-throughput screening
methodology to identify α-Syn aggregation inhibitors (11). Here,
we describe the properties of SynuClean-D (SC-D), a small molecule identified with this approach (SI Appendix, Fig. S1). We first
performed a detailed in vitro biophysical characterization of the
inhibitory and disaggregation activities of SC-D and tested its
performance in human neural cells. Finally, we validated the effects in vivo in two well-established Caenorhabditis elegans models
of PD, which express α-Syn either in muscle cells or in dopaminergic
Parkinson’s disease is characterized by the accumulation of amyloid deposits in dopaminergic neurons, mainly composed of the
protein α-synuclein. The disordered nature of α-synuclein and its
complex aggregation reaction complicate the identification of
molecules able to prevent or revert the formation of these inclusions and the subsequent neurodegeneration. By exploiting a
recently developed high-throughput screening assay, we identified SynuClean-D, a small compound that inhibits α-synuclein
aggregation, disrupts mature amyloid fibrils, prevents fibril
propagation, and abolishes the degeneration of dopaminergic
neurons in an animal model of Parkinson’s disease.
Downloaded by guest on April 14, 2020
arkinson’s disease (PD) is the second most prevalent neurodegenerative disorder after Alzheimer’s disease (AD) and
is still incurable (1). PD is the most common synucleinopathy, a
group of neurodegenerative disorders that includes dementia with
Lewy bodies and multiple system atrophy (MSA), among others (2,
3). Although the synucleinopathies are multifactorial disorders, the
molecular events triggering the pathogenic breakthrough of the
disease converge to the abnormal aggregation of α-synuclein (α-Syn)
in dopaminergic neurons (4, 5). α-Syn aggregation also occurs in
oligodendrocytes in patients with MSA (6). α-Syn is an intrinsically
disordered protein, which is expressed at high levels in the brain.
α-Syn function is thought to be related to vesicle trafficking (7). This
wild-type protein is the main component of cytoplasmic Lewy bodies
(LB) and Lewy neurites (LN) in sporadic PD (8). In addition,
dominantly inherited mutations in α-Syn, as well as multiplications of
the gene encoding for α-Syn (SNCA), cause familial forms of PD (9).
Author contributions: S.V. designed research; J.P., S.P.-D., D.F.L., F. Peccati, F. Pinheiro,
D.G., A.C., S.N., M.C.-G., J.G., S.G., and E.D. performed research; J.P., S.P.-D., D.F.L., E.G., X.S.,
J.S., M.S., T.F.O., E.D., and S.V. analyzed data; and M.S., E.D., and S.V. wrote the paper.
Conflict of interest statement: J.P., S.P.-D., M.C.-G., J.S., E.D., and S.V. are inventors on a
patent application (PCT/EP2018/054540) related to the compound in this study.
This article is a PNAS Direct Submission.
Published under the PNAS license.
J.P. and S.P.-D. contributed equally to this work.
To whom correspondence should be addressed. Email: email@example.com.
This article contains supporting information online at www.pnas.org/lookup/suppl/doi:10.
Published online September 24, 2018.
PNAS | October 9, 2018 | vol. 115 | no. 41 | 10481–10486
Edited by Gregory A. Petsko, Weill Cornell Medical College, New York, NY, and approved August 16, 2018 (received for review April 3, 2018)
neurons. The inhibitor reduced α-Syn aggregation, improved
motility, and protected against neuronal degeneration.
SynuClean-D Inhibits α-Syn Aggregation in Vitro. The formation of
α-Syn amyloid fibrils can be reproduced in vitro by incubating the
recombinant protein. However, fibril growth is very slow and highly
variable, complicating drug screening (12). We have implemented
a robust high-throughput kinetic assay to screen large chemical libraries in the search for α-Syn aggregation inhibitors (11). The assay
uses thioflavin-T (Th-T) as readout of amyloid formation, completing highly reproducible reactions in 30 h. Approximately 14,400
chemically diverse compounds of the HitFinder Collection from
Maybridge were screened with this approach. SC-D [2-hydroxy-5nitro-6-(3-nitrophenyl)-4-(trifluoromethyl)nicotinonitrile], a small
aromatic compound, was identified as one of the molecules of
potential interest (SI Appendix, Fig. S1). Many compounds with
promising pharmacological characteristics never become drugs
because they are rapidly metabolized in the liver and therefore
have low oral bioavailability. SC-D was metabolically stable in the
presence of human hepatic microsomes, with an intrinsic clearance of <5 μL·min−1·mg−1 (SI Appendix, Fig. S2).
Incubation of 70 μM α-Syn with 100 μM SC-D impacted α-Syn
aggregation, as monitored by Th-T fluorescence (Fig. 1A). The
analysis of the aggregation curves indicated that the autocatalytic
rate constant in the presence of the compound (ka 0.25 h−1) was
25% lower than in its absence (ka 0.33 h−1). SC-D increases t50 by
1.5 h and reduces by 53% the amount of Th-T–positive material
at the end of the reaction. By measuring light scattering, we
confirmed that the observed changes in Th-T fluorescence
reflected an effective decrease in the levels of α-Syn aggregates,
with a reduction of 48 and 58% in the scattering signal at the end
of the reaction in the presence of SC-D when exciting at 300 and
340 nm, respectively (Fig. 1B). Nanoparticle tracking analysis
indicated that the presence of SC-D increased the number of
particles of <100 nm and decreased the formation of large aggregates (150 to 500 nm) (SI Appendix, Fig. S3). Finally, transmission electron microscopy (TEM) images confirmed that samples
incubated with SC-D contained smaller and much fewer fibrils per
field than untreated samples (Fig. 1 C and D). The inhibitory activity of SC-D was dose-dependent and still statistically significant at
10 μM (1:7 compound:α-Syn ratio), where it reduces the final Th-T
signal by 34% (Fig. 1E).
We further investigated whether SC-D was active against the
aggregation of α-Syn variants associated with PD (1). SC-D was
able to reduce by 45 and 73% the amount of Th-T–positive aggregates at the end of the reaction for the H50Q and A30P α-Syn
familial variants, respectively (Fig. 1F).
The inhibitory activity of SC-D was also assessed using
protein-misfolding cyclic amplification (PMCA) (13). Conceptually based on the nucleation-dependent polymerization model
for prion replication, PMCA has been recently adapted to amplify α-Syn amyloid fibrils (14). The PMCA technique combines
cycles of incubation at 37 °C, to grow fibrils, and sonication, to
break fibrils into smaller seeds. In our conditions, a single cycle
of amplification was sufficient to generate amyloid-like protease
K (PK)-resistant α-Syn assemblies, but the highest levels of
protection were attained after four rounds (Fig. 1G). When the
same experiment was performed in the presence of SC-D, we
observed a substantial decrease in the amount of PK-resistant
material (Fig. 1H), indicating that the molecule was interfering
with α-Syn template seeding amyloid formation.
SynuClean-D Disrupts Preformed α-Syn Fibrils. The progress of
α-Syn PMCA reactions can also be monitored by using the Th-T
signal as the readout for fibril assembly (15). Consistent with PK
resistance analysis, Th-T fluorescence of α-Syn increased significantly after four cycles of PMCA (Fig. 2A). Surprisingly, in the
presence of SC-D, the Th-T signal not only did not increase, but
Downloaded by guest on April 14, 2020
Fig. 1. Effect of SynuClean-D on the aggregation of α-Syn in vitro. (A) α-Syn aggregation kinetics in the absence (black) and presence (blue) of SC-D followed
by Th-T–derived fluorescence. (B) Light-scattering signal at 300 and 340 nm, both in the absence (white) and presence (blue) of SC-D. (C and D) Representative
TEM images in the absence (C) and presence (D) of SC-D. (E) Inhibition of α-Syn aggregation in the presence of different concentrations of SC-D. (F) H50Q and
A30P α-Syn variant aggregation in the absence (white) and presence (blue) of SC-D. (G and H) Bis/Tris gels of PMCA samples in the absence (G) and presence
(H) of SC-D, both analyzed after PK digestion. Soluble α-Syn and PMCA steps 4 and 5 are shown. Th-T fluorescence is plotted as normalized means. Final points
were obtained at 48 h. Error bars are represented as SE of mean values; **P < 0.01 and ***P < 0.001.
10482 | www.pnas.org/cgi/doi/10.1073/pnas.1804198115
Pujols et al.
We did not detect any perturbations in chemical shifts or peak
intensities with respect to the original α-Syn spectrum in the
presence of 100 μM concentration of the molecule (SI Appendix,
Fig. S4), indicating that SC-D does not bind α-Syn monomers.
Induced-fit docking simulations of α-Syn–SC-D revealed four
major poses for its interaction with α-Syn fibrils (16): two internal,
with SC-D fully inserted in the fibril (poses 1 and 2), and two
external, with SC-D partially exposed (poses 3 and 4) (SI Appendix, Fig. S5). In the internal poses, the ligand is sandwiched
between two parallel β-sheets of the Greek-key motif and interacts
with the side chains of ALA53, VAL55, THR59, GLU61, THR72,
and GLY73. The only difference between pose 1 and 2 lies in the
orientation of the compound in the binding pocket. PELE (17)
interaction energies are stronger for the internal poses, where
SC-D binds essentially through dispersion interactions into a
solvent-excluded cavity, than for external ones, where SC-D inserts into a surface groove of the fibril. In light of these calculations, we predict that SC-D binds into the core of α-Syn fibrils.
MM/GBSA calculations (SI Appendix, Table S1) (18, 19) show
that internal binding pose 1 (Fig. 3) exhibits the largest binding
energy with the fibril, the computed ΔGbind being −18.4 ± 4.1 kcal·
mol−1. The main contribution comes from the van der Waals term,
representing roughly 80% of the interaction. This is not surprising
given the nature of SC-D, a planar aromatic molecule. Plots of the
reduced density gradient versus the density (SI Appendix, Fig.
S6A) provide information on the nature of the noncovalent interactions in the system (20). Peaks in the negative and positive
regions of the x axis are indicative of attractive and repulsive interactions, respectively. The region around zero corresponds to
the weakest noncovalent van der Waals contacts. Though weak,
these interactions are present in large number and involve the
whole body of the molecule, being the largest contribution to the
binding energy. Their spatial extension is shown in SI Appendix,
Fig. S6B. For pose 1, the noncovalent interaction plot shows that
Fig. 2. Disaggregational capacity of SynuClean-D. (A) Th-T fluorescence of
the different PMCA passes of both treated (blue) and untreated (black)
samples with SC-D. (B) Aggregation kinetics of α-Syn after the addition of SCD at different time points. (C and D) Th-T–derived fluorescence (C) and lightscattering (D) assays before and after the addition of SC-D to preformed
α-Syn fibrils. (E and F) Representative TEM images in the absence (E) and
presence (F) of SC-D. Th-T fluorescence is plotted as normalized means. Error
bars are represented as SE of mean values; ***P < 0.001.
Downloaded by guest on April 14, 2020
began to decrease after the third cycle. This suggested that SC-D
might disrupt newly formed amyloid fibrils.
To address the time window in which SC-D is active, we set up
aggregation reactions with a constant amount of SC-D at different time intervals. As presented in Fig. 2B, the effect of SC-D
on the final amount of amyloid structures was independent of
whether it was added at the beginning (4 h), in the middle (12 h),
or at the end (18 h) of the exponential phase, or even when the
reaction had already attained a plateau (24 h). These results
suggested again ability to disrupt/destabilize fibrils.
To confirm the fibril-disrupting activity of SC-D, 4-d mature α-Syn
fibrils were incubated in the absence or presence of the compound
for 24 h. Incubation with SC-D promoted a 43% reduction in Th-T
fluorescence emission (Fig. 2C). Moreover, light-scattering measurements indicated a reduction in the amount of detectable aggregates by 29 and 39% at 300 and 340 nm, respectively (Fig. 2D).
Consistently, TEM images illustrated how 4-d-incubated α-Syn
tended to form big fibrillary clusters (Fig. 2E), which became
completely disrupted in the presence of SC-D (Fig. 2F).
α-Syn Fibrils Can Accommodate SynuClean-D. To assess if SC-D can
bind monomeric and soluble α-Syn, the recombinant protein was
isotopically labeled and NMR 1H-15N-HSQC spectra of 70 μM
[15N]α-Syn were recorded in the absence and presence of SC-D.
Pujols et al.
Fig. 3. Characterization of SynuClean-D–fibril interaction. General view (A) and
zoom (B) of the most stable binding pose of SC-D on the α-Syn fibril model.
PNAS | October 9, 2018 | vol. 115 | no. 41 | 10483
besides van der Waals, an H-bond contact is responsible for the
binding of SC-D (SI Appendix, Fig. S6A).
SynuClean-D Inhibits the Formation of Intracellular α-Syn Aggregates
in Cultured Cells. We tested the potential toxicity of SC-D for
human neuroglioma (H4) and human neuroblastoma (SH-SY5Y)
cells. For both cell lines, the molecule was innocuous at concentrations as high as 50 μM (Fig. 4A and SI Appendix, Fig. S7). We
used a well-established cell model that enabled us to assess α-Syn
inclusion formation. H4 cells were transiently transfected with
C-terminally modified α-Syn (synT) and synphilin-1, which results
in the formation of LB-like inclusions, as we previously described
(9). The formation of α-Syn inclusions was assessed 24 h after
treatment by immunofluorescence (Fig. 4D). Upon treatment with
1 and 10 μM SC-D, we observed a significant increase in the
number of transfected cells devoid of α-Syn inclusions (SC-D,
1 μM: 42.4 ± 1.0%; SC-D, 10 μM: 49.5 ± 4.5%) relative to untreated samples (control: 28.7 ± 2.0%) (Fig. 4B). SC-D treatment
also promoted a significant decrease in the number of transfected
cells displaying more than five aggregates (SC-D, 1 μM: 35.5 ±
5.0%; SC-D, 10 μM: 32.5 ± 6.6%) relative to control cells (control: 49.6 ± 5.6%) (Fig. 4C).
SynuClean-D Inhibits α-Syn Aggregation in a C. elegans Model of PD.
Downloaded by guest on April 14, 2020
Next, we tested SC-D in a living system. We used a well-studied
nematode model of PD, the strain NL5901, in which human α-Syn
fused to the yellow fluorescent protein (YFP) is under control of
the muscular unc-54 promoter, transgene pkIs2386 [Punc-54::αSYN::YFP] (21). Muscle expression has been used successfully to
model protein-misfolding diseases and to identify modifier genes
without considering neuronal effects (21, 22). To determine the
effects of SC-D in α-Syn accumulations, animals at the fourth larval
stage (L4) (23) were incubated with and without the compound, to
analyze the inhibitor efficiency in aged worms at 9 d posthatching
Fig. 4. Inhibition of α-Syn aggregate formation in cultured cells. (A) Human
neuroglioma cell (H4) survival when incubated with different concentrations
of compound (blue) and without (white) the compound. (B and C) Reduction
of α-Syn inclusion formation in human cultured cells in the presence of
different concentrations of SC-D. (B) Percentage of transfected cells devoid
of α-Syn aggregates. (C) Percentage ...