Specific Sequence Stretches Drive Aggregation
The specific physico-chemical properties of amino acids and their combination in linear patterns along the primary structure play a major role in the potential of a given polypeptide to aggregate and, thus, define the intrinsic determinants of protein aggregation. However, it has been shown that not all the protein sequence is equally important for the ability of a protein to aggregate—but, instead, there are short sequence fragments that promote and guide the formation of amyloid-like structures (Ventura et al. 2004; Ivanova et al. 2004). This principle defines the “amyloidogenic stretch” hypothesis, and such fragments are commonly referred to as “Hot Spots” or aggregation-prone regions (APRs). Consistent with the intrinsic determinants of protein aggregation, these segments are characterized by an enrichment in hydrophobic amino acids—both aliphatic (Val, Leu, Ile) and aromatic (Phe, Trp, Tyr) (Rousseau et al. 2006b).
Structural Determinants of Amyloid-like Aggregation
As already discussed, the X-ray diffraction patterns of amyloids, amyloid-like fibrils, and different kinds of apparently amorphous aggregates share a cross-p super-secondary level of structure. However, the failure of these different kinds of protein aggregates to attain a sufficiently regular 3-dimensional assembly (even in the case of the apparently macroscopically ordered amyloid-like fibrils) hampered for a long time the description of amyloid structure at atomic detail. Fortunately, advances in solid state Nuclear Magnetic Resonance (ssNMR) (Petkova et al. 2002; Ritter et al. 2005) and in the microcrystallization of short amyloidogenic peptides (Makin et al. 2005; Nelson et al. 2005; Rodriguez et al. 2015) have elucidated the fine molecular architecture of the amyloid-like fibrils formed by different proteins and by peptides thereof. Most of the solved structures confirm a cross-p core composed of two opposite p-sheets running perpendicular to the axis of the fibril; although fibrils formed by certain amyloidogenic proteins adopt a p-helix structure instead, where three p-strands are arranged facing each other on every turn (Tycko 2011; Eisenberg and Jucker 2012; Tycko and Wickner 2013). The molecular complementarity required for the assembly of each pair of facing strands in the cross-p conformation is particularly highlighted by the crystallographic structures of amyloidogenic peptides (Sawaya et al. 2007), which reveal how the docking of facing strands defines a “steric zipper” formed by the inward-pointing side chains. At the same time, the structures of these peptides reflect the array of posible arrangements that can be adopted by p-strands to build the cross-p structure. The atomic detail provided by these experimental structures provide an outstanding framework to rationalize the role of the determinants of protein aggregation we have introduced before. In first place, it allows for an understanding of how the small sequence stretches defining APRs can guide and promote the formation of amyloid-like structure, since only a small portion of the polypeptide is strictly required in order to contribute a p-strand for the establishment of the cross-p core of the fibril, while the rest of the molecule may well remain exposed to the solvent or even attached as either a partially or completely structured fragment (Sambashivan et al. 2005). Next, the high degree of molecular complementarity required to build the cross-p conformation explains how, while the formation of amyloid-like structures is thermodynamically driven by the backbone-mediated hydrogen bond network, its assembly is limited by the requirement of amino acidic combinations able to provide appropriate physico-chemical properties and shape complementarity. In addition, residues whose side chains are responsible for the contacts between opposite strands sustaining the solvent-protected “steric zipper” tend to have an apolar character. Because the geometry of the p-conformation results in contiguous amino acids pointing out in opposite directions, this implies residues that do not participate of the “steric zipper” would be located in the solvent-exposed face of the strand, explaining why a sequential pattern that alternates hydrophobic and hydrophilic amino acids is well accommodated by amyloid-like structures.