The Amyloid Phenomenon and Its Links with Human Disease

Abstract

The ability of normally soluble proteins to convert into amyloid fibrils is now recognized to be a generic phenomenon. The overall cross-β architecture of the core elements of such structures is closely similar for different amino acid sequences, as this architecture is dominated by interactions associated with the common polypeptide main chain. In contrast, the multiplicity of complex and intricate structures of the functional states of proteins is dictated by specific interactions involving the variable side chains, the sequence of which is unique to a given protein. Nevertheless, the side chains dictate important aspects of the amyloid structure, including the regions of the sequence that form the core elements of the fibrils and the kinetics and mechanism of the conversion process. The formation of the amyloid state of proteins is of particular importance in the context of a range of medical disorders that include Alzheimer's and Parkinson's diseases and type 2 diabetes. These disorders are becoming increasingly common in the modern world, primarily as a consequence of increasing life spans and changing lifestyles, and now affect some 500 million people worldwide. This review describes recent progress in our understanding of the molecular origins of these conditions and discusses emerging ideas for new and rational therapeutic strategies by which to combat their onset and progression.

Figure 1.

Predicted costs of AD in the United States (see also Alzheimer’s Association 2015). (From Alzheimer’s Disease Facts and Figures 2015, Alzheimer’s Association 2015; reprinted, with permission, from Elsevier © 2015.)

Figure 2.

Images of amyloid fibrils formed from the SH3 domain of PI3 kinase. The fibrils were formed in vitro by incubation of the protein (whose native structure is represented in Fig. 3) at low pH (Guijarro et al. 1998). Although the SH3 domain has no links with disease, the fibrils (which have diameters of just a few nanometers) have all the characteristics of the fibrillar aggregates isolated from the tissue of patients suffering from amyloid diseases such as Alzheimer’s and Parkinson’s (Chiti and Dobson 2006). This observation gave rise to the concept that the amyloid structure is a common or “generic” form of protein architecture, albeit one that is generally observed in nature only in pathological states (Dobson 1999).

Figure 3.

Comparison of examples of native and amyloid structures of protein molecules. On the left are ribbon diagrams of the native structures of three small proteins: an SH3 domain (top), myoglobin (bottom), and acylphosphatase (middle). The native structures differ in their topologies and contents of α-helices and β-sheets resulting from the dominance of side-chain interactions within their highly evolved sequences. On the right is a molecular model of an amyloid fibril. (Image kindly provided by Helen Saibil, Birkbeck College, London, from data reported in Jiménez et al. 1999.) The fibril was produced from the SH3 domain whose native structure is shown on the left and consists of four “protofilaments” that twist around one another to form a hollow tube with a diameter of ∼6 nm. The β-strands (flat arrows) are oriented perpendicularly to the fibril axis and are linked together by hydrogen bonds involving main-chain amide and carbonyl groups, many of which are intermolecular, to form a continuous structure in each protofilament. The protofilaments are held together by much weaker interactions involving primarily side-chain contacts. As the main chain is common to all polypeptides, the core protofilament structures of fibrils from different sequences have common features, differing only in detail as a result of differences in the nondominant effects of side-chain packing. The arrow indicates that when the native states of globular proteins are destabilized, they tend to convert into the generic amyloid structure, as described in the text. (Reprinted from Dobson 2008.)

Figure 4.

Structure of an amyloid fibril at atomic resolution. The structure shown is one of several polymorphs of the amyloid fibrils that are formed from an 11-residue fragment of transthyretin. The combination of cryo-electron microscopy imaging (A) with solid-state NMR analysis has enabled the determination of an atomic-level structure (B). A more detailed view (C) shows the hierarchical organization of the amyloid fibril in which the three filaments that form the mature fibril illustrated here are in turn formed by pairs of cross-β protofilaments, which are each composed of pairs of β-sheets. The fibril surfaces are shown as electron density maps, and the constituent β-sheets are shown in a ribbon representation; oxygen, carbon, and nitrogen atoms are shown in red, gray, and blue, respectively. (From Knowles et al. 2014; reprinted, with permission, from The American Association for the Advancement of Science © 2009 and Fitzpatrick et al. 2013 with permission from National Academy of Sciences © 2013.)

Figure 5.

Different types of aggregates. AFM images of aggregates formed during the conversion of α-synuclein from the soluble monomeric form into amyloid fibrils. The series of images shows the spectrum of approximately spherical oligomers prior to more numerous aggregates that include thin protofibrils and mature fibrils. (From Apetri et al. 2006; reprinted, with permission, from the Journal of Molecular Biology © 2006.)

Figure 6.

The effect of mutations in the sequence of the 42-residue human Alzheimer Aβ-peptide on neuronal dysfunction in transgenic fruit flies (Luheshi et al. 2007). (A) A climbing assay of flies expressing the wild-type sequence (left) and two mutational variants (right) predicted to reduce the peptide's aggregation propensity; the healthier the flies, the higher up the tube they can climb. (B) A similar experiment with flies expressing the Aβ-peptide containing the E22G “Arctic mutation” (left-hand tube). The two right-hand tubes are of peptides that contain additional mutations that decrease the propensity to form oligomeric prefibrillar aggregates. (Adapted from Dobson 2008.)

Figure 7.

A summary of the general classes of mechanisms that create protein aggregates. Primary nucleation pathways result in the formation of new aggregates from interactions solely between soluble monomers. Such aggregates can grow (and in the case of fibrillar species, elongate) by the addition of further monomeric species. But the number of aggregates can increase by additional “secondary” processes including monomer-independent events, such as fibril fragmentation, which generate new aggregates at a rate that depends only on the level of the aggregates already present in the solution, and monomer-dependent events such as surface-catalyzed secondary nucleation; this latter process creates new aggregates at a rate that depends on the concentrations of both monomeric protein and existing aggregates. The rate constants of the various processes are labeled. Further details are given in Knowles et al. (2009). (From Knowles et al. 2009; reprinted, with permission, from The American Association for the Advancement of Science © 2009.)

Figure 8.

Identification of the inhibition of secondary nucleation and evaluation of the consequences on the generation of toxic oligomers by means of kinetic analysis. (A–C) Kinetic reaction profiles for the aggregation of Aβ42 are shown from left (blue) to right (green) for reactions in the absence of the molecular chaperone called BRICHOS, and with 10%, 20%, 50%, 75%, and 400% Aβ42 monomer equivalents of BRICHOS. The effect of BRICHOS saturates at a stoichiometry of approximately one monomer equivalent, and the blue dashed line is the integrated rate law for Aβ42 aggregation through primary and secondary nucleation using previously determined rate constants. The dashed green lines show predictions for the resulting reaction profiles when each of (A) primary nucleation, (B) fibril elongation, and (C) secondary nucleation are inhibited by the chaperone. Note the characteristic differences in the change in the shape of the reaction profile in each case. The prediction for the case where the chaperone solely and entirely suppresses secondary nucleation is matched essentially perfectly in the presence of excess BRICHOS. The thin dotted lines in C are theoretical predictions for the intermediate BRICHOS concentrations using the association and dissociation rate constants determined from separate binding experiments. (D–F) Time evolution of the concentration of low-molecular-weight oligomeric species predicted by kinetic analysis. The blue line corresponds to the situation without the presence of BRICHOS, and the dashed green lines show predictions when each of (D) primary nucleation, (E) fibril elongation, and (F) secondary nucleation are suppressed by BRICHOS. The results reveal that this particular chaperone effectively inhibits the latter process, which has been shown to be that responsible for the proliferation of toxic oligomeric species in the aggregation process in the absence of the chaperone. (From Cohen et al. 2015; adapted, with permission, from the authors.)