Edited by Pablo G. Debenedetti, Princeton University, Princeton, NJ, and approved December 20, 2018 (received for review November 26, 2018)
Phase diagrams of atomic systems are calculated routinely by computer simulations, but such calculations are absent for even the simplest peptides. Previous simulations are mainly nonequilibrium and focus on the assembly of peptides from the monomeric state to the aggregated state. To obtain accurate equilibrium solubilities, it is necessary to simulate many assembly and disassembly events of fibrillar aggregates, which is notoriously difficult as it requires breaking many hydrogen bonds. We overcome these challenges and calculate the equilibrium phase diagram of amyloid β protein (16–22) (Aβ16–22), the archetypal amyloid former, using a realistic protein model. Importantly, our prediction of Aβ16–22 solubility over temperatures from 277 to 330 K agrees well with experimental measurements.
The aggregation of monomeric amyloid β protein (Aβ) peptide into oligomers and amyloid fibrils in the mammalian brain is associated with Alzheimer’s disease. Insight into the thermodynamic stability of the Aβ peptide in different polymeric states is fundamental to defining and predicting the aggregation process. Experimental determination of Aβ thermodynamic behavior is challenging due to the transient nature of Aβ oligomers and the low peptide solubility. Furthermore, quantitative calculation of a thermodynamic phase diagram for a specific peptide requires extremely long computational times. Here, using a coarse-grained protein model, molecular dynamics (MD) simulations are performed to determine an equilibrium concentration and temperature phase diagram for the amyloidogenic peptide fragment Aβ16–22. Our results reveal that the only thermodynamically stable phases are the solution phase and the macroscopic fibrillar phase, and that there also exists a hierarchy of metastable phases. The boundary line between the solution phase and fibril phase is found by calculating the temperature-dependent solubility of a macroscopic Aβ16–22 fibril consisting of an infinite number of β-sheet layers. This in silico determination of an equilibrium (solubility) phase diagram for a real amyloid-forming peptide, Aβ16–22, over the temperature range of 277–330 K agrees well with fibrillation experiments and transmission electron microscopy (TEM) measurements of the fibril morphologies formed. This in silico approach of predicting peptide solubility is also potentially useful for optimizing biopharmaceutical production and manufacturing nanofiber scaffolds for tissue engineering.
Author contributions: Y.W., S.J.B., S.A., and C.K.H. designed research; Y.W. and S.J.B. performed research; Y.W., S.J.B., and S.A. analyzed data; and Y.W., S.E.R., A.J.W., S.A., and C.K.H. wrote the paper.
The authors declare no conflict of interest.
This article is a PNAS Direct Submission.
This article contains supporting information online at www.pnas.org/lookup/suppl/doi:10.1073/pnas.1819592116/-/DCSupplemental.
Published under the PNAS license.
