Role of water in Protein Aggregation and Amyloid Polymorphism

A variety of neurodegenerative diseases are associated with the formation of amyloid plaques. Our incomplete understanding of this process underscores the need to decipher the principles governing protein aggregation. Most experimental and simulation studies have been interpreted largely from the perspective of proteins: the role of solvent has been relatively overlooked. In this Account, we provide a perspective on how interactions with water affect folding landscapes of A$\beta$ monomers, A$\beta_{16-22}$ oligomer formation, and protofilament formation in a Sup35 peptide. Simulations show that the formation of aggregation-prone structures (N$^$) similar to the structure in the fibril requires overcoming high desolvation barrier. The mechanism of protofilament formation in a polar Sup35 peptide fragment illustrates that water dramatically slows down self-assembly. Release of water trapped in the pores as water wires creates protofilament with a dry interface. Similarly, one of the main driving force for addition of a solvated monomer to a preformed fibril is the entropy gain of released water. We conclude by postulating that two-step model for protein crystallization must also hold for higher order amyloid structure formation starting from N$^$. Multiple N$^*$ structures with varying water content results in a number of distinct water-laden polymorphic structures. In predominantly hydrophobic sequences, water accelerates fibril formation. In contrast, water-stabilized metastable intermediates dramatically slow down fibril growth rates in hydrophilic sequences.

💡 Research Summary

This Account presents a comprehensive perspective on how water governs the aggregation of amyloid‑forming peptides, focusing on three model systems: amyloid‑β (Aβ) monomers, the Aβ₁₆₋₂₂ fragment, and a polar peptide derived from the yeast prion Sup35. Using explicit‑solvent molecular dynamics simulations, the authors demonstrate that water is not a passive background but an active participant that shapes the free‑energy landscape, determines kinetic bottlenecks, and generates structural polymorphism.

For Aβ monomers, the simulations reveal a rugged landscape populated by many compact conformations. The aggregation‑prone “N*” states, which contain the D23‑K28 salt bridge and the V24‑GSN turn observed in fibrils, are separated from the lowest‑energy basins by a high desolvation barrier. Water molecules tightly coordinate the charged side chains; a water‑mediated salt bridge persists for several picoseconds, making the removal of water a prerequisite for the formation of the native‑like N* conformation. Chemical cross‑linking that pre‑forms the D23‑K28 bridge accelerates aggregation by ~10³‑fold, confirming that overcoming the water barrier is rate‑limiting.

In the Aβ₁₆₋₂₂ system, the central hydrophobic cluster LVFFA drives rapid expulsion of water from the inter‑peptide crevices. Early aggregates are disordered, “dry” droplets that quickly evolve into antiparallel β‑sheet oligomers. The growth proceeds via a dock‑lock mechanism: a solvated monomer first diffuses onto the oligomer surface (fast docking), then undergoes a slower conformational conversion to a β‑strand that locks into the existing nematic droplet. Importantly, water does not re‑enter the interior during this step; the driving force is the favorable hydrophobic side‑chain contacts and the formation of inter‑peptide salt bridges (K16‑E22). Thus, for highly hydrophobic sequences water expulsion is an early event and not the kinetic bottleneck.

Conversely, the Sup35 peptide is highly polar. When two β‑sheets approach, a one‑dimensional water wire forms along the interface, stabilized by hydrogen bonds to amide groups. This wire creates a long‑lived metastable state that dramatically slows assembly. Only after the trapped water is released does a dry interface emerge, yielding a helically twisted proto‑filament. Here water acts as a kinetic inhibitor, and the rate‑determining step is the dehydration of the pore.

The authors connect these findings to the thermodynamics of protein crystallization, where the release of structured hydration layers provides a large positive entropy gain (ΔS_water) that can outweigh the loss of translational and rotational entropy of the proteins. Analogous water‑release mechanisms have been documented for Tobacco Mosaic Virus assembly and for Aβ oligomerization, where the number of expelled water molecules (5–30 per oligomer) contributes significantly to the overall entropy increase that drives fibril growth.

A key conceptual advance is the proposal of a two‑step model for amyloid formation, mirroring the nucleation‑growth paradigm of crystallization. First, water is expelled from hydrophobic contacts, generating a “dry” nucleus (N*). Second, structural ordering aligns the peptides into a cross‑β architecture. Because multiple N* conformations can exist with varying water content, the model naturally explains the emergence of polymorphic fibrils: each water‑laden N* seeds a distinct polymorph, and the balance between water‑accelerated and water‑stabilized intermediates dictates the overall kinetics. In predominantly hydrophobic sequences water accelerates fibrillization, whereas in hydrophilic sequences water‑stabilized intermediates retard growth.

Overall, the paper underscores that water is a dual‑role agent—both a catalyst and a barrier—depending on sequence composition. By modulating water‑protein interactions (e.g., designing molecules that lower desolvation barriers or disrupt water wires), it may be possible to control amyloid aggregation pathways, offering new avenues for therapeutic intervention in neurodegenerative diseases.