Blocking the FKBP12 induced dendrimeric burst in aberrant aggregation of α-synuclein by using the ElteN378 synthetic inhibitor

Abstract α-Synuclein (α-syn), a disordered cytoplasmatic protein, plays a fundamental role in the pathogenesis of Parkinson’s disease (PD). Here, we have shown, using photophysical measurements, that addition of FKBP12 to α-syn solutions, dramatically accelerates protein aggregation, leading to an explosion of dendritic structures revealed by fluorescence and phase-contrast microscopy. We have further demonstrated that this aberrant α-syn aggregation can be blocked using a recently discovered non-immunosuppressive synthetic inhibitor of FKBP12, ElteN378. The role of FKBP12 and of ElteN378 in the α-syn aggregation mechanism has been elucidated using molecular dynamics simulations based on an effective coarse-grained model. The reported data not only reveal a new potent synthetic drug as a candidate for early stage treatment of α-syn dependent neurodegenerations but also pave the way to a deeper understanding of the mechanism of action of FKBP12 on α-syn oligomeric aggregation, a topic which is still controversial.

SDS electrophoresis. The concentration of FKBP12 in Tris buffer was determined by UV absorbance at 280 nm using 280 = 9970 M −1 . ElteN378 was synthesized as reported in Ref. 1 In Figure S1, we show the chemical structure of the chiral ElteN378 (on the right) along with the structure of the ElteN378 bound to a FKBP12 protein (on the left). 1 The design of ElteN378 is such that the two carbonyl oxygen atoms (highlighted in the chemical structure of Figure S1) can make two H-bonds with the hydrogen atom of the hydroxy group in Tyr82 (PDB atom label HH) and the amide hydrogen atom NH of Ile56 of the FKBP12 protein, while the two non polar phenyl moiety interact via π − π stacking. Thioflavine T (ThT) was purchased from Sigma-Aldrich Chemie GmbH (Schnelldorf, Germany). Tris 10 mM/HCl buffer was prepared with Trizma®base, Trizma®hydrochloride and NaCl purchased from Aldrich and used as received. All solutions were prepared with deionized water (resistivity =18 MΩ cm, pH = 5.6 at 20 • C) obtained from a Milli-RO coupled with a Milli-Q set up (Millipore, Italy). Except for the moderate agitation, which is known to greatly accelerate the diffusionlimited aggregration process, these experimental conditions were chosen in order to observe aggregation phenomena using physiological protein concentration 2-4 and pH levels. In this regard, it should be mentioned that in many biochemical in vitro studies of α-syn aggregation, high levels of α-syn monomer (40-200 µM) are commonly used and the process is often seeded with polyamines or fibril fragments. With such experimental setup, aggregation occurs very quickly, producing mature fibrils in just 20-40 h. The significance of these kind of studies for the aggregation kinetics of α-syn in physiological conditions (and in PD etiology in general) is unclear.

Photo-physical investigation of aggregation kinetics
Fluorescence spectra were recorded on a LS50B spectrofluorimeter (Perkin-Elmer). Excitation and emission slits were set to 10 nm. QS cell with 0.3 cm optical path length from Hellma (Hellma GmbH & Co. KG, Muellheim, Germany) were used for fluorescence experiments, whereas 1 cm OP cuvettes were used for absorption measurements. Cuvettes were cleaned with piranha solution and carefully rinsed with water and ethanol. The cuvettes were dried by nitrogen flushing prior to each measurement. All fluorescence measurements were run at 20 • C unless otherwise stated. The emission spectra of the corresponding blank solutions were always recorded separately and subtracted. Emission spectra of the samples containing the ThT probe were obtained exciting in the 300-450 nm range. Emissiondependent excitation spectra were acquired for emission wavelength in the range 430-500 nm. Reported data represent the average of at least three independent experiments.

Imaging of mature fibrillar aggregates
Images of 90-days old samples were taken with a Diaphot 300 Inverted microscope (Nikon) equipped with a LWD condenser capable of phase contrast, differential interference and bright field microscopy. The instrument was provided also with an epi-fluorescence attachment module (TMD-EF, Nikon) with a 100W Epi Fluorescence HBO mercury arc lamp house and power supply, an Epi-Fluorescence beam delivery system with selectable filters cassettes installed. Epi-fluorescence images were taken with different filters combination to selectively isolate the 450 nm emitting species and the 480 n emitting ThT species. In the images reported in the paper we used the BV combination (IF 400-450 excitation filter, with main excitation wavelength 436 nm, plus a DM455 dichroic mirror). Both 20x and 40x objectives were used to explore the homogeneity of the samples, scale bars are reported in the images accordingly. Aliquots withdrawn from the samples after 90 days of incubation where imaged immediately after withdrawal with and without addition of ThT to the samples. We applied 5 to 10 µl of this sample onto a microscope glass slide covered with a cover-slip at a distance of ca. 15 µm (as measured by z-scanning the samples with a Leica Confocal Laser Scanning Microscope). Samples were prepared following procedures already described 5 that includes proper sealing to avoid solvent evaporation either with enamel or wax (recipe from Nikon Imaging Center) to prevent evaporation. Similar results were also obtained using

Simulation procedures
The α-syn monomer chain is represented in our coarse-grained (CG) approach by 15 beads, made up of three distinct part of five beads length each, i.e. a central hydrophobic part (corresponding to the NAC domain) and two terminal hydrophilic parts. As shown in Figure   S3, in the monomer, the CG beads have hence only two possible colors, namely hydrophilic (terminal segments 1-5 and 11-15) and hydrophobic (central segment 6-10). Consecutive beads are bound via a stretching harmonic potential with r 0 = 1.5 nm equilibrium distance and with a stiff force constant of 230 in units of RT × nm −2 . A harmonic bending potential with equilibrium angle α =120 • and force constant k = 230 in units of RT ×rad −2 is enforced between three consecutive beads, irrespective of their color. The equilibrium bead-bead distance and bending angle are chosen so that the 15 beads CG α-syn monomer, when in the fully extended state, has a length of the order of 20 nm. We use a solvent free model with re-normalized bead-bead non-bonded interactions so as to mimic, in a water environment, the aggregation of hydrophobic moieties and solubilization of the hydrophilic groups. To this end, a strongly repulsive atom-atom potential is assigned to the hydrophilic beads (1-5 and 11-15) while an attractive atom-atom potential is used to model the interactions of the hydrophobic NAC domain central beads. The potential functional form for the bead-bead non bonded interaction is of the Lennard-Jones type and standard Lorentz-Berthelot mixing rules applies for the hydrophilic-hydrophobic interactions. The Lennard-Jones well depth and σ parameters for hydrophobic-hydrophobic interactions are tuned so as to obtain, at a distance of 1.5 nm, approximately a gain of 10 RT units per bead (representing approximately 10 residues), corresponding to a reasonable mean value of 1 RT units gain for the aggregation of two hydrophobic residues in water. 6 Potential parameters are reported in the Figure S3 that exemplifies the CG model used for α-syn and α-syn/FKBP12 mixtures.

Fluorescence spectra of ThT 3 µM in Tris Buffer
ThT emission and excitation spectra in buffer solution (shown in the Figure S4   The dissociation constant for the FKBP12-α-syn complex in presence of the tight binding inhibitor ElteN378 is given by where C is the concentration of the equimolar FKBP12 α-syn mixture, C E is the ElteN378 nominal concentration such that C E < C, and [FS] is the concentration of the FKBP12-α-syn complex at equilibrium. In Eq. 1 we have tacitly assumed that: i) all added ElteN378 is instantly bound to FKBP12; 1 ii) the concentration of the species [F n S] with n > 1 are negligible when C=1 µM iii) the equilibrium in the reaction FKBP12 + αsyn ↔ FS is established at day zero.
We now define f = (C − C E )/C as the fraction of available (not inhibited) FKBP12 and r = K d /C as the ratio between the FKBP12-α-syn dissociation constant and the concentration of the 1:1 FKBP12/α-syn mixture. Solving Eq. 1 for φ ≡ [FS]/C, we obtain that the fraction of FKBP12 proline bound α-syn with respect to the initial concentration C is given by (2) φ in Eq. 2 depends hence on the reduced quantity r = K d /C. In Figure S7 we show fraction of FKBP12 bound α-syn as a function of the fraction of available (not ElteN378-inhibited) FKBP12, ranging from 0 (C Elte = C F ) to 1 (C Elte = 0, no inhibitor). Assuming r = 1 (i.e.  According to our aggregation model, the FKBP12 globular protein induces ramification of the supramolecular structure and is hence expected to be localized on a hydrophilic segment at the junction of the branches. The localization of FKBP12 in the aggregates is assessed from our configurational data using the Voronoi tessellation algorithm. 19 The final configurations are examined by computing in all samples the Voronoi volumes on the sub-ensemble of beads made by all the hydrophobic units and by all beads (mutated or not) in position 2 of an hydrophilic segment (see Figure S3). Voronoi volumes are evaluated in periodic boundary conditions using the minimum image convention and by closing the polyhedrons, when needed, by introducing 8 distant octahedral additional vertices placed at a distance of 10 4 nm from the central bead. In this manner, exposed beads have faces determined by the distant octahedral vertices, yielding extremely high Voronoi volumes. As discussed above, in the FKBP12-α-syn mixture, a fraction of these hydrophilic beads is replaced by hydrophobic ones. As shown in Figure S8, the Voronoi volume distribution in the final state of pure α-syn Figure S8: Voronoi volume distribution of the hydrophobic beads in pure α-syn. Volumes for exposed residues are in the range V = 10 4 : 10 12 nm.
peaks at about 20 nm 3 indicating a tight packing of the hydrophobic core of mature fibrils.
Based on the distribution of Figure S8, we define a bead to be buried (i.e. within the interior of the hydrophobic core of the supramolecular aggregates) when its Voronoi volume is less than 50 nm 3 . In Table S1 we show, for various values of the ratio φ = [FS]/C], the results for the fraction of buried FKBP12 beads with respect to the total number of FKBP12 beads compared to the fraction of buried (non mutated) hydrophilic beads in position 2. The FKBP12 beads are clearly much more buried, on the average, with respect to the "normal" hydrophilic beads in position 2, that are basically solvent exposed in the vast majority with c M2−syn buried 0. Most of the non buried FKBP12 beads exhibit extremely large Voronoi volume indicating that these units are in general solvent exposed. We also note that, with increasing  Table S1, we expect that 1/3 of α-syn monomers are FKBP12-bound and that in these complexes only about the 10% of the FKBP12 is buried in the supramolecular cores. Hence, based on the simulation results, we estimate that in the 1:1 FKBP12-α-syn 1 µM mixture the amount of buried (non washable) FKBP12 in mature aggregates is in the nanomolar range (20:40 nM).