Supplementary Information for SAR by kinetics for drug discovery in protein misfolding diseases Sean Chia*, Johnny Habchi*, Thomas C. T. Michaels*, Samuel I. A. Cohen, Sara Linse, Christopher M. Dobson, Tuomas P. J. Knowles, Michele Vendruscolo Michele Vendruscolo Email:
[email protected] This PDF file includes: Supplementary text Figs. S1 to S8 Tables S1 to S5 References for SI reference citations
www.pnas.org/cgi/doi/10.1073/pnas.1807884115
1
Materials and Methods Preparation of Aβ42 peptides The
recombinant
Aβ(M1-42)
peptide
(MDAEFRHDSGY
EVHHQKLVFF
AEDVGSNKGA IIGLMVGGVV IA), here called Aβ42, was expressed in the E. coli BL21 Gold (DE3) strain (Stratagene, CA, U.S.A.) and purified as described previously with slight modifications (1). Briefly, the purification procedure involved sonication of E. coli cells, dissolution of inclusion bodies in 8 M urea, and ion exchange in batch mode on diethylaminoethyl cellulose resin followed by lyophylisation. The lyophilised fractions were further purified using Superdex 75 HR 26/60 column (GE Healthcare, Buckinghamshire, U.K.) and eluates were analysed using SDS-PAGE for the presence of the desired peptide product. The fractions containing the recombinant peptide were combined, frozen using liquid nitrogen, and lyophilised again. Compounds A-C, and I were purchased from ArkPharm (Illinois, USA); Compounds R and S were purchased from Molport (Riga, Latvia); and all other compounds were obtained from Sigma-Aldrich (Germany). All chemicals were of the highest purity available. Preparation of Aβ42 samples for kinetic experiments Solutions of monomeric Ab42 were prepared by dissolving the lyophilized Ab42 peptide in 6 M guanidinium hydrocholoride (GuHCl). Monomeric forms were purified from potential oligomeric species and salt using a Superdex 75 10 ⁄ 300 GL column (GE Healthcare) at a flowrate of 0.5 mL/min, and were eluted in 20 mM sodium phosphate buffer, pH 8 supplemented with 200 µM EDTA and 0.02% NaN3. The centre of the peak was collected and the peptide concentration was determined from the absorbance of the integrated peak area using e280 = 1490 l mol-1 cm-1. The obtained monomer was diluted with buffer to the desired concentration and supplemented with 20 μM thioflavin T (ThT) from a 2 mM stock. All samples were prepared in low binding Eppendorf tubes on ice using careful pipetting to avoid introduction of air bubbles. Each sample was then pipetted into multiple wells of a 96well half-area, low-binding, clear bottom and PEG coating plate (Corning 3881), 80 µL per well, in the absence and the presence of different molar-equivalents of small molecules.
2
ThT fluorescence assay Assays were initiated by placing the 96-well plate at 37 ºC under quiescent conditions in a plate reader (Fluostar Omega, Fluostar Optima or Fluostar Galaxy, BMGLabtech, Offenburg, Germany). The ThT fluorescence was measured through the bottom of the plate using a 440 nm excitation filter and a 480 nm emission filter. The ThT fluorescence was followed for three repeats of each sample. Data sets for bexarotene, compounds D-H, and compound J were taken from literature (2) for analysis and the creation of the KIA fingerprints. Theoretical analysis The time evolution of the total fibril mass concentration, M(t), is described by the following integrated rate law (3, 4) : !(#) !(%)
*+ ,-+ /0 + ,-+ .
= 1 − )*
35 4 /0 64 /3 *1 ,-+ .
*1 ,-+
2
7 894 #
(Eq. S1)
To capture the complete assembly process, only two particular combinations of the rate constants define most of the macroscopic behaviours. These are related to the rate of formation of new aggregates through primary pathways : = ;2=, => ?(0)>A and
through
secondary
pathways
B = ;2=, =C ?(0)>5 ,D ,
where
the
initial
concentration of soluble monomers is denoted by m(0), nc and n2 describe the dependencies of the primary and secondary pathways on the monomer concentration, and kn, k+ and k2 are the rate constants of the primary nucleation, elongation and secondary nucleation, respectively. Inhibitors can interfere with the aggregation process by inhibiting one or more of the individual microscopic reactions. We can identify the microscopic events that are inhibited by the chemical compounds by applying the above equation to describe the macroscopic aggregation profiles shown in Figs. S1 and S6, and comparing the set of microscopic rate constants k+k2 and k+kn required to describe the time evolution of the fibril formation in the absence and presence of small molecules. Since surface-catalysed secondary nucleation is primarily responsible for the generation of Ab42 oligomers (3), a molecule that inhibits primary nucleation leaves
3
relatively unaffected the overall number of oligomers produced during the aggregation process. Instead, since primary nucleation is the initial event leading to aggregation, the presence of this small molecule retards the formation of the critical amount of fibrils needed for surface-catalysed secondary nucleation to dominate, and thus effectively reduces the flux toward oligomers. By contrast, a molecule that inhibits secondary nucleation redirects free monomers towards elongation, and thus reduces the total overall number of oligomers on top of decreasing the flux toward oligomers (5). We observed that the molecules affect the aggregation process of Ab42 to different extents (Figs. S1 and S6), and thus inhibit the primary nucleation and secondary nucleation pathways to different degrees (Figs. S2 and S8). Using the rate constants (kn, k2 or k+) in the presence of the molecules, we can estimate the reactive flux toward oligomers (r(t)) as (1, 6, 7) : E(F) = => ?(F)>A + =C ?(F)>5 H(F)
(Eq. S2)
From the plots in Figs. 2A and 3A, we can predict the time of which the generation of oligomers reaches a peak, as well as the total number of oligomers generated over time (time integral of r(t)). In the presence of the molecules, as a result of the inhibition in both primary and secondary nucleation, the time taken for the oligomers to reach a peak is delayed, which consequently translates to a delay in the rate of oligomerization. The total number of oligomers generated over time is also reduced in the presence of the small molecules. The evolution of the increase in peak time for the flux toward oligomers, the reduction in the total number of oligomers formed, and the increase in half-time as a function of small molecules concentration x (Figs. S2, D
S4 and S8), can be described by a functional dependence of the form I = KD,9LM
N
to
give estimates of the OIC50PT, OIC25T, and KIC50M values respectively. This formula emerges from an explicit mathematical treatment of the kinetic equations of protein aggregation in the presence of small molecules in the limit of fast binding of the compound to the various aggregate species (8).
4
Generation of a 3D model of the pharmacophore 3D conformations of the positive compounds were first generated using CORINA (9). These conformations (in a mol2 format) were used as input with default options, and the algorithm for PharmaGist was subsequently employed to generate the pharmacophore model (10). The features of the pharmacophore calculated was matched by all the molecules that were used as the inputs. The pharmacophore was subsequently visualised using ROCS (11, 12). Calculation of physico-chemical properties of the small molecules Constitutional descriptors and molecule properties were calculated using E-Dragon 1.0 software (13). The apolarity of the components of the molecules was expressed through the Ghose-Crippen octanol-water coefficient, and the number of rotatable bonds was expressed through the number of single bonds, not in a ring, bound to a nonterminal heavy atom. The average length of the linker was calculated by summing average bond lengths obtained from literature (14).
5
Figure S1. Effect of bexarotene and its derivatives on Ab42 aggregation. Kinetic profiles of the aggregation of a 2 µM Ab42 solution in the presence of either 1% DMSO (black) or 2 µM (green), 6 µM (orange), or 10 µM (red) of (a) bexarotene, (b) compound A, (c) compound B, (d) compound C, (e) compound D, (f) compound E, (g) compound F, (h) compound G, (i) compound H, (j) compound I, and (k) compound J. The continuous lines represent the integrated rate laws for Ab42 aggregation fitted to the experimental data. Data from (a), (e)-(i), and (k) are obtained from literature (2).
6
Figure S2. Evolution of the apparent reaction rate constants and half-times for bexarotene and its derivatives. (a) Evolution of normalised t1/2 of the aggregation kinetics of 2 µM Ab42 in the presence of increasing concentrations of different small molecules (represented in different colours), as derived from Fig. S1. (b,c) Dependence of the apparent reaction rate constants (kapp) of (b) primary pathways (k+kn), and (c) secondary pathways (k+k2), as derived from Fig. S1, is shown for increasing concentrations of different small molecules (represented in different colours).
7
Figure S3. Effects of bexarotene and its derivatives on Ab42 aggregation in the presence of 30% of preformed seed fibrils. Kinetic profiles of the aggregation of a 2 µM Ab42 solution in the presence of 30% of preformed seed fibrils, and either 1% DMSO (black) or 6 µM (orange), or 10 µM (red) of (a) bexarotene, (b) compound B, (c) compound D, (d) compound E, (e) compound F, (f) compound H, and (g) compound J. The continuous lines represent the integrated rate laws for Ab42 aggregation fitted to the experimental data.
8
Figure S4. Estimation of OIC50PT and OIC25T for bexarotene and its derivatives. Evolution of the increase in (a) OIC50PT, i.e. the time to reach the peak value of the generation of Ab42 oligomers, and in (b) OIC25T, i.e. the reduction in the total number of oligomers generated over time with increasing concentrations of different small molecules (represented in different colours), as derived from Fig. 2A.
9
Figure S5. Generation of a pharmacophore from the alignment of bexarotene and its positive derivatives. The pharmacophore consists of two main components: an apolar moiety and a polar moiety separated by about 6 Å.
10
Figure S6. Effect of rhodanine-based compounds on Ab42 aggregation. Kinetic profiles of the aggregation of a 2 µM Ab42 solution in the presence of either 1% DMSO (black) or 2 µM (green), 6 µM (orange), 10 µM (red), or 20 µM (blue) of (a) compound K, (b) compound L, (c) compound M, (d) compound N, (e) compound O, (f) compound P, (g) compound Q, (h) compound R, and (i) compound S. The continuous lines represent the integrated rate laws for Ab42 aggregation fitted to the experimental data.
11
Figure S7. Effect of rhodanine-based compounds on Ab42 aggregation in the presence of 30% of preformed seed fibrils. Kinetic profiles of the aggregation of a 2µM Ab42 solution in the presence of 30% of preformed seed fibrils, and either 1% DMSO (black) or 2 µM (green), 6 µM (orange), 10 µM (red), or 20 µM (blue) of (a) compound M, (b) compound N, (c) compound O, and (d) compound R. The continuous lines represent the integrated rate laws for Ab42 aggregation fitted to the experimental data.
12
Figure S8. Estimation of the KIC50M, OIC50PT, and OIC25T parameters for rhodanine-based compounds. (a) Evolution of normalised t1/2 of the aggregation kinetics of 2 µM Ab42 in the presence of increasing concentrations of different small molecules (represented in different colours), as derived from Fig. S6 (b,c) The dependence of the apparent reaction rate constants (kapp) of (b) primary pathways (k+kn), and (c) secondary pathways (k+k2), as derived from Fig. S6, is shown with increasing concentrations of different small molecules (represented in different colours). (d,e) Evolution of (d) the increase in time to reach the peak value of the generation of Ab42 oligomers, and (e) the reduction in the total number of oligomers generated over time with increasing concentrations of different small molecules (represented in different colours), as derived from Fig. 3A.
13
Attribute Inhibition of overall aggregation Reduction in total number of oligomers formed
Delay in the flux toward oligomers
Associated
Expression in the KIA
parameter
fingerprint
KIC50M
1/ KIC50M
OIC25T
1/ OIC25T
OIC50PT
1/ OIC50PT
Table S1: Summary of the three parameters in the KIA fingerprints: (1) OIC50PT, which is associated with a delay in the flux toward oligomer, (2) OIC25T, which is associated with a reduction in the total number of oligomers generated, and (3) the macroscopic parameter KIC50M, which describes the overall rate of aggregate formation.
14
Molecule
Linker
Average
Rotatable
Length
Bonds
(Å)
(around linker)
Delay in Hydrogen
Hydrogen
flux
Bond
Bond
toward
Acceptor
Donor
oligomers (1/OIC50PT)
Reduction in total
Inhibition of
number of
overall
oligomers
aggregation
formed
(1/KIC50M)
(1/OIC25T)
O
B
~2.96
2
ü
O
0.13
0.12
0.18
Bexarotene
~2.94
2
O
O
0.37
0.21
0.36
J
~4.40
2
O
O
0.36
0.3
0.4
~4.18
2
ü
ü
0.5
0.34
0.48
~3.03
2
O
ü
0.83
0.3
0.59
O
F N H
O
D
O
Table S2: Effects on Ab42 oligomer production associated with the physico-chemical properties of the linker group R2 (Fig. 2) in bexarotene and its derivatives. Values in green indicate an increase in potency with respect to bexarotene, while values in red indicate a reduction in potency.
15
Delay in flux Molecule
Apolar Group
Lipophilicity
toward oligomers (1/OIC50PT)
Reduction in total
Inhibition of
number of
overall
oligomers
aggregation
formed
(1/KIC50M)
(1/OIC25T)
C
2.32
N.A
N.A
N.A
E
4.63
0.22
0.24
0.22
0.37
0.21
0.36
1.1
0.5
1.1
4.75
Bexarotene
N
H
~5.27
Table S3: Effects on Ab42 oligomer production associated with the physico-chemical properties of the apolar group R3 (Fig. 2) in bexarotene and its derivatives. Values in green indicate an increase in potency with respect to bexarotene, while values in red indicate a reduction in potency.
16
Molecule
Polar Group
Polar Surface Area
Activity
86.49
Inhibits aggregation
103.72
Does not inhibit aggregation
O
M,N,O,R
NH S
S O
S
N
NH2
S
S
Table S4: Effects on the inhibition of Ab42 aggregation associated with the physicochemical properties of the polar group R1 (Fig. 3) in rhodanine-based compounds. Upon the substitution of a more polar group, molecule S does not inhibit significantly the aggregation of Ab42.
17
Delay in flux Molecule
Apolar Group
Lipophilicity
toward oligomers (1/OIC50PT)
Reduction in total
Inhibition of
number of
overall
oligomers
aggregation
formed
(1/KIC50M)
(1/OIC25T)
HO
K
1.56
N.A
N.A
N.A
O
2.22
0.12
0.08
0.12
N
3.14
0.59
0.36
0.56
0.91
0.48
0.91
O
3.35
R
Table S5: Effects on Ab42 oligomer production associated with the physico-chemical properties of the apolar group R3 (Fig. 3) in rhodanine-based compounds. Values in green indicate an increase in potency with respect to Molecule K.
18
SI References 1.
Habchi J, et al. (2016) An anti-cancer drug suppresses the primary nucleation reaction that initiates the formation of toxic Aβ aggregates associated with Alzheimer’s disease. Sci Adv 2(2):e1501244.
2.
Habchi J, et al. (2016) Systematic development of small molecules to inhibit specific microscopic steps of Aβ42 aggregation in Alzheimer’s disease. Proc Natl Acad Sci:201615613.
3.
Cohen SIA, et al. (2013) Proliferation of amyloid-β42 aggregates occurs through a secondary nucleation mechanism. Proc Natl Acad Sci U S A 110(24):9758–9763.
4.
Meisl G, et al. (2016) Molecular mechanisms of protein aggregation from global fitting of kinetic models. Nat Protoc 11(2):252–272.
5.
Arosio P, Vendruscolo M, Dobson CM, Knowles TPJ (2014) Chemical kinetics for drug discovery to combat protein aggregation diseases. Trends Pharmacol Sci 35(3):127–135.
6.
Cohen SIA, et al. (2015) A molecular chaperone breaks the catalytic cycle that generates toxic Aβ oligomers. Nat Struct Mol Biol 22(3):207–213.
7.
Michaels TCT, Lazell HW, Arosio P, Knowles TPJ (2015) Dynamics of protein aggregation and oligomer formation governed by secondary nucleation. J Chem Phys 143(5):054901.
8.
Arosio P, et al. (2016) Kinetic analysis reveals the diversity of microscopic mechanisms through which molecular chaperones suppress amyloid formation. Nat Commun 7(2):10948.
9.
Sadowski J, Gasteiger J, Klebe G (1994) Comparison of Automatic ThreeDimensional Model Builders Using 639 X-ray Structures. J Chem Inf Model 34(4):1000–1008.
10.
Schneidman-Duhovny D, Dror O, Inbar Y, Nussinov R, Wolfson HJ (2008) PharmaGist: a webserver for ligand-based pharmacophore detection. Nucleic Acids Res 36:223–228.
11.
ROCS 3.2.2.2: OpenEye Scientific Software, Santa Fe, NM.
12.
Hawkins PCD, Skillman a. G, Nicholls A (2007) Comparison of shapematching and docking as virtual screening tools. J Med Chem 50(1):74–82.
13.
Tetko I V., et al. (2005) Virtual computational chemistry laboratory - Design and
19
description. J Comput Aided Mol Des 19(6):453–463. 14.
Haynes WM (2014) CRC handbook of chemistry and physics (CRC press).
20
