Distinct thermodynamic signatures of oligomer generation in the aggregation of the amyloid-β peptide

Abstract

Mapping free-energy landscapes has proved to be a powerful tool for studying reaction mechanisms. Many complex biomolecular assembly processes, however, have remained challenging to access using this approach, including the aggregation of peptides and proteins into amyloid fibrils implicated in a range of disorders. Here, we generalize the strategy used to probe free-energy landscapes in protein folding to determine the activation energies and entropies that characterize each of the molecular steps in the aggregation of the amyloid-β peptide (Aβ42), which is associated with Alzheimer's disease. Our results reveal that interactions between monomeric Aβ42 and amyloid fibrils during fibril-dependent secondary nucleation fundamentally reverse the thermodynamic signature of this process relative to primary nucleation, even though both processes generate aggregates from soluble peptides. By mapping the energetic and entropic contributions along the reaction trajectories, we show that the catalytic efficiency of Aβ42 fibril surfaces results from the enthalpic stabilization of adsorbing peptides in conformations amenable to nucleation, resulting in a dramatic lowering of the activation energy for nucleation.

Figure 1. Kinetics of Aβ42 aggregation from purely monomeric peptide at different temperatures and initial monomer concentrations.

Normalized experimental reaction profiles, monitored by ThT fluorescence, for Aβ42 aggregation from purely monomeric peptide for different initial concentrations of monomeric peptide and at different temperatures (a) 26°C, (b) 29°C, (c) 31°C, (d) 33°C, (e-f) 36°C, (g) 40°C and (h) 45°C in 20 mM sodium phosphate, 0.2 mM EDTA, 0.02% sodium azide, pH 8.0, with 6 µM ThT. The initial concentrations of monomers were 5.0 µM (purple), 4.5 µM (red), 4.0 µM (pink), 3.5 µM (orange), 3.0 µM (yellow), 2.5 µM (green), 2.2 µM (cyan) and 1.9 µM (blue). Note the different scales on the time axes; Supplementary Fig. 1 shows the data with the same scale for each panel. The data were recorded in two sets of four temperatures with measurements at 36°C included in both sets to act as a reference condition. The solid lines are global fits at each temperature using the analytical integrated rate law for Aβ42 aggregation. Two combinations of the microscopic rate constants, k₊k_n and k₊k₂, are used to globally fit the entire data set in each panel, in terms of rate constants for elongation (k₊), primary nucleation (k_n) and secondary nucleation (k₂). The rate parameters determined at each temperature from the global fitting are plotted in Fig. 3a,b.

Figure 2. Kinetics of pre-seeded Aβ42 aggregation at different temperatures and initial monomer concentrations.

Normalized experimental reaction profiles, monitored by ThT fluorescence, for Aβ42 aggregation in the presence of pre-formed fibrils for different initial concentrations of monomeric peptide and at different temperatures (a) 29°C, (b) 33°C, (c) 36°C and (d) 40°C. The initial concentrations of monomers were 5.0 µM (yellow), 4.0 µM (pink) and 3.0 µM (purple). The solid lines are global fits at each temperature using the analytical integrated rate law for Aβ42 aggregation. The data in each panel is fitted globally with a single parameter, k₊/L(0), where L(0) is the average length of the pre-formed fibrils. The rate parameters determined at each temperature from the global fitting are plotted in Fig. 3c.

Figure 3. Arrhenius behavior of the microscopic rate constants for Aβ42 aggregation.

Arrhenius plots showing the temperature dependence of the rate parameters determined from the analytical fitting in Figs. 1, 2: (a) for k₊, (b) for the combined rate parameter k₊k_n, (c) for the combined rate parameter k₊k₂, in terms of rate constants for elongation (k₊), primary nucleation (k_n) and secondary nucleation (k₂). (d) shows how the data from (a-c), which is plotted as the faded data points and lines, are combined to give the temperature dependencies of the individual rate constants. Interestingly, while k₊ and k_n increase at higher temperatures, k₂ has a weak temperature dependence with the opposite trend. The error bars for the rate parameters represent standard errors from the global fits shown in Figs. 1-2 and for the temperatures indicate the ranges of fluctuations recorded during each experiment.

Figure 4. Activation energies of fibril elongation, primary nucleation and secondary nucleation in Aβ42 amyloid formation.

(a) The free energies of activation, and the enthalpic and entropic contributions, determined from our measurements. The activation energies for elongation and primary nucleation and primary nucleation consist of enthalpic barriers and favorable entropies of activation, whereas the enthalpic and entropic contributions to the free energy are reversed in sign for secondary nucleation. The values and standard errors shown were calculated by fitting the data shown in Fig. 3 with ΔH^‡ϴ = −R ∂(log k)/∂(1/T) and ΔG^‡ϴ = −RT log(k/A), which were combined to give TΔS^‡ϴ = ΔH^‡ϴ −ΔG^‡ϴ. The entropic term is shown at T = 298K. The values are given per mole of reaction at a standard state of 1M. (b) Schematic showing the different temperature dependencies of the two nucleation processes that both generate aggregates from monomeric peptides. The structures used in the visual representations of each process are adapted from Refs. and .

Figure 5. Mapping the energy landscape for secondary nucleation.

(a) SPR measurements of the adsorption of monomers onto fibrils as a function of concentration and temperature. The data are fitted to the Langmuir isotherm to determine K_D fitted at each temperature. (b) The variation in the surface coverage α(T) = 1/K_D(T), the rate constant of the subsequent nucleation reaction k_r(T), and the overall rate of secondary nucleation k₂(T) = α(T)²k_r(T). The corresponding values of K_D are K_D(20°C) = 11μM, K_D(30°C) = 26μM, K_D(40°C) = 64μM. (c) The energy landscape for secondary nucleation assuming a standard state of 1M, showing the trajectory from starting materials, through adsorption and reaction, to products. Since our experiments are typically at micromolar concentrations, the same landscape is shown assuming a standard state of 1μM in Supplementary Fig. 6, where ΔG_ads > 0. Note that, according to Kramers theory of diffusive reactions, the activation parameters determined from the temperature dependence of the rate constants correspond to the highest free energy barrier measured relative to the reactants.