A method for probing the mutational landscape of amyloid structure

Abstract

Motivation: Proteins of all kinds can self-assemble into highly ordered β-sheet aggregates known as amyloid fibrils, important both biologically and clinically. However, the specific molecular structure of a fibril can vary dramatically depending on sequence and environmental conditions, and mutations can drastically alter amyloid function and pathogenicity. Experimental structure determination has proven extremely difficult with only a handful of NMR-based models proposed, suggesting a need for computational methods.

Results: We present AmyloidMutants, a statistical mechanics approach for de novo prediction and analysis of wild-type and mutant amyloid structures. Based on the premise of protein mutational landscapes, AmyloidMutants energetically quantifies the effects of sequence mutation on fibril conformation and stability. Tested on non-mutant, full-length amyloid structures with known chemical shift data, AmyloidMutants offers roughly 2-fold improvement in prediction accuracy over existing tools. Moreover, AmyloidMutants is the only method to predict complete super-secondary structures, enabling accurate discrimination of topologically dissimilar amyloid conformations that correspond to the same sequence locations. Applied to mutant prediction, AmyloidMutants identifies a global conformational switch between Aβ and its highly-toxic 'Iowa' mutant in agreement with a recent experimental model based on partial chemical shift data. Predictions on mutant, yeast-toxic strains of HET-s suggest similar alternate folds. When applied to HET-s and a HET-s mutant with core asparagines replaced by glutamines (both highly amyloidogenic chemically similar residues abundant in many amyloids), AmyloidMutants surprisingly predicts a greatly reduced capacity of the glutamine mutant to form amyloid. We confirm this finding by conducting mutagenesis experiments.

Availability: Our tool is publically available on the web at http://amyloid.csail.mit.edu/.

Contact: lindquist_admin@wi.mit.edu; bab@csail.mit.edu.

Fig. 1.

Amyloid fibril schemas used for analysis. Amyloid fibril schemas, diagrammed from side and top perspectives. Red indicates a single fibril peptide flanked by two gray adjacent peptides along the fibril axis. (a) Schema 𝒫, a 2-sheet β-solenoid with unrestricted number of rungs per peptide and parallel intra- and interchain interactions. (b) Schema 𝒜, identical to 𝒫 except with antiparallel interchain interactions. (c) Schema 𝒮, a serpentine cross-β structure with unrestricted number of packed intrachain β-sheets. All β-strand hydrogen bonds formed interchain.

Fig. 2.

β-strand ‘kinks’ extend schemas in Figure 1 to allow AmyloidMutants to model sharp β-sheet turns like those found in β-solenoids. Kink represents a deviation in the standard β-sheet in/out residue sidechain orientation.

Fig. 3.

AmyloidMutants per-residue β-strand assignments indicate amyloid core regions, comparable with existing per-residue amyloidogenicity predictors. AmyloidMutants predictions (green) outperform those tools available for testing when using their default settings and thresholds. BETASCAN (gold) (Bryan et al., 2009), ZYGGREGATOR (blue) (Tartaglia and Vendruscolo et al., 2008), TANGO (cyan) (Fernandez-Escamilla et al., 2004), PASTA (red) (Trovato et al., 2007) and Waltz (purple) (Maurer-Stroh et al., 2010), when compared against experimental structure models supported by NMR, H/D-exchange and mutational analysis (black) (Luca et al., 2007; Lührs et al., 2005; Mukrasch et al., 2009; Vilar et al., 2007; Wasmer et al., 2008, 2010). Note, the BETASCAN, ZYGGREGATOR, TANGO and PASTA tools most closely match our tool's ability to predict full-length per-residue amyloidogenicity, whereas Waltz aims to predict short hots pots that could specifically adopt a steric zipper.

Fig. 4.

AmyloidMutants structure predictions match experimentally observed β-strand interactions of Aβ_1-42 (a), HET-s (b), amylin (c), α-synuclein (d) and tau (e). (a) Diagram depicts Aβ_1-42 β-strand in gray, residues in blue (with in/out orientation) and β-sheet/β-sheet packing as one β-strand above another, packed residues facing center. Predicted structure (green arrows) mirrors NMR structure (Lührs et al., 2005) (black arrows), including most packing orientations. Predicted kink occurs because schema does not account for known D23/K28 salt bridge. (b) Similar depiction of HET-s prediction (top, green arrows) compared with NMR model (Wasmer et al., 2008) (bottom, black arrows) shows near identical match, including residue orientations and kink location. (c) Top two amylin predictions (solid, striped green arrows) align well to NMR model (Luca et al., 2007) (black arrows). Predictions differ only by their inclusion of Phe23 (*) within β-sheet, a residue experimentally shown to form non-β-sheet interpeptide interactions not considered by schema. (d) Top two α-synuclein predictions (i,ii) agree very well with H/D exchange data (iii,iv) and NMR model (v) (Heise et al., 2005; Vilar et al., 2007). (e) Tau predictions identify 7/8 β-regions observed experimentally (Mukrasch et al., 2009). The highest AmyloidMutants scores (red boxes) specifically identify regions 274–279 and 305–310, positions believed crucial to fibril nucleation (von Bergan et al., 2000).

Fig. 5.

HET-s/4N→Q is defective for amyloid assembly. Purified proteins were filtered through a non-binding membrane either before or after incubation for 24 h in a physiological buffer. Protein aggregates that formed during the incubation are retained on the surface of the membrane, as visualized by Ponceau-S staining.