Computational Analysis of the Interactions between the S100B Extracellular Chaperone and Its Amyloid β Peptide Client

Abstract

S100B is an astrocytic extracellular Ca²⁺-binding protein implicated in Alzheimer's disease, whose role as a holdase-type chaperone delaying Aβ₄₂ aggregation and toxicity was recently uncovered. Here, we employ computational biology approaches to dissect the structural details and dynamics of the interaction between S100B and Aβ₄₂. Driven by previous structural data, we used the Aβ_25-35 segment, which recapitulates key aspects of S100B activity, as a starting guide for the analysis. We used Haddock to establish a preferred binding mode, which was studied with the full length Aβ using long (1 μs) molecular dynamics (MD) simulations to investigate the structural dynamics and obtain representative interaction complexes. From the analysis, Aβ-Lys28 emerged as a key candidate for stabilizing interactions with the S100B binding cleft, in particular involving a triad composed of Met79, Thr82 and Glu86. Binding constant calculations concluded that coulombic interactions, presumably implicating the Lys28(Aβ)/Glu86(S100B) pair, are very relevant for the holdase-type chaperone activity. To confirm this experimentally, we examined the inhibitory effect of S100B over Aβ aggregation at high ionic strength. In agreement with the computational predictions, we observed that electrostatic perturbation of the Aβ-S100B interaction decreases anti-aggregation activity. Altogether, these findings unveil features relevant in the definition of selectivity of the S100B chaperone, with implications in Alzheimer's disease.

Keywords: amyloids; chaperones; docking; molecular dynamics; protein aggregation; protein folding diseases; protein interactions.

Figure 1

Molecular representations of Aβ42, Aβ_25–35, S100B, and the main experimental data which guided the molecular docking procedure. (a) Aβ₄₂ peptide cartoon structure (Protein Data Bank, PDB entry 1Z0Q) obtained from NMR in a 30/70 ratio solution of hexafluoro-2-propanol/water [22] (Lys28 and Ile31 residues are depicted in sticks to highlight their spatial proximity) and the NMR-based experimental evidence showing Lys28 and Ile31 interacting with S100B (redrawn from NMR interaction data from [8]). (b) Haddock solution used in this study with the Aβ_25–35 structure located in the S100B (grey surface) binding cavity. (c) Full-length Aβ₄₂ rebuilt from the Haddock solution in α-helical conformation. N-ter: N-terminus.

Figure 2

Structural equilibration properties, the hydrophobicity/hydrophilicity (SAS^hydro) index, and the final conformations for the long molecular docking (MD) simulations. (a) Helical content of S100B, Aβ_25–35 and Aβ_N-ter during production MD runs. (b) Interfacial area between Aβ_25–35/Aβ_N-ter and S100B. (c) SAS^hydro indexes for the Aβ_25–35 region at the interface. (d) Average SAS^hydro index values for the hydrophobic and hydrophilic residues in the Aβ_25–35 segment. A floating window of 100 ns was used in the time series to reduce the local fluctuations. The final interfacial Solvent Acessible Surface Area (SASA) values were obtained by averaging the two interfacial areas, one mapped on the protein and another mapped on the peptide surface. The hydrophobic and hydrophilic residues in the SAS^hydro index calculations were separated based on their sign (positive for hydrophilic and negative for hydrophobic residues). The average SAS^hydro index values were obtained from the equilibrated segments (the last 450 ns). The error bars were calculated from the standard error of the mean between replicates. A representative structure of the Aβ₄₂ peptide is shown in the cartoon with the most relevant residues, in terms of SAS^hydro index values, represented with spheres.

Figure 3

Mapping the interactions of Lys28 from Aβ42 with S100B residues. (a) Structural representation of the Aβ₄₂:S100B complex highlighting the important role of Lys28 (yellow) and distances to nearby residues Met79 (pink, 0.74 nm), Thr82 (green, 0.37 nm) and Glu (blue, 0.58 nm). (b) Probability density distributions of minimum distances between Lys28 (the terminal amino group) and the hot spot partner residues on the S100B surface.

Figure 4

Estimation of the binding energy for the Aβ:S100B complex and experimental evidence supporting the relevance of coulombic interactions for Aβ:S100B complex stabilization. (a) Structural representation of the Aβ:S100B complex (depicted as light grey cartoon with Aβ_25–35 in yellow, on the protein grey surface) highlighting the contributions of Lys28 and Ile31 residues (marked with sticks). (b) Molecular mechanics Poisson–Boltzmann surface area (MM-PBSA) binding energy calculated for the Aβ_25–35:S100B complex as the sum of all energetic terms involved over the equilibrated segments of the simulations and with the error bars calculated from the standard error of the mean between replicates. (c) Aggregation of 5 µM Aβ₄₂ at 37 °C with or without 25 µM Ca²⁺-S100B under low (left) and high (right) ionic strength conditions. (d) Binding impairment between S100B and monomeric Aβ₄₂ at high ionic strength accounts for the partial depletion of S100B inhibitory activity, for example, over the mechanism of primary nucleation, which is exclusively dependent on monomeric Aβ₄₂ concentration (top). Half-time (t_1/2) and lag time (t_lag) values of Aβ₄₂ aggregation in all tested conditions (error bars represent standard deviation, n = 3) (bottom).