Benchmarking Implicit Solvent Folding Simulations of the Amyloid β(10−35) Fragment

Andrew Kent, Abhishek Jha, James E. Fitzgerald, Karl F. FreedThe Journal of Physical Chemistry B 2008 · DOI

TL;DR

This paper benchmarks an implicit solvent model for simulating the amyloid-beta (Aβ) peptide, a key player in Alzheimer's disease. By comparing simulations with experimental data and explicit solvent simulations, they demonstrate that their implicit solvent approach offers a significant speedup while maintaining accuracy, making it a promising tool for studying Aβ aggregation.

What Problem Does This Paper Address?

Alzheimer's disease is linked to the aggregation of amyloid-beta (Aβ) peptides into toxic oligomers and fibrils. Understanding how these peptides fold and aggregate is crucial for developing treatments. However, simulating the long timescales (microseconds and beyond) required for aggregation using traditional all-atom explicit solvent molecular dynamics (MD) is computationally prohibitive. This necessitates the development of faster, yet accurate, computational models. The Aβ(10-35) fragment is particularly interesting because its monomeric properties have been experimentally measured, providing a benchmark for validating new simulation methods.

Methods

The core of this study is the validation of an implicit solvent model for simulating the Aβ(10-35) peptide. This model treats the solvent as a continuous medium, reducing computational cost compared to explicitly modeling every water molecule. The authors employed Langevin dynamics (LD) simulations, which incorporate solvent friction and random forces.

To validate their implicit solvent model, they performed several key comparisons:

  • Implicit vs. Explicit Solvent: They compared their implicit solvent simulations with a new, computationally expensive (200x slower) all-atom explicit solvent MD simulation using the CHARMM force field and TIP3P water.
  • Force Field Dependence: They tested the influence of different force fields (CHARMM, OPLS-aal, and GS-AMBER94) on the simulation results.
  • Initial Structure Dependence: They investigated how starting simulations from different initial conformations (native NMR structure, fully stretched, and unfolded) affects the convergence and properties of the simulations.

They calculated various structural and dynamic properties, including self-diffusion constant, end-to-end distance, radius of gyration, RMSD from the NMR structure, NMR order parameter (S²), solvent accessible surface area (SASA), and residue-residue contacts.

Datasets

The primary "dataset" for validation was experimental data for the Aβ(10-35) monomer. This included experimental measurements of NMR order parameters and diffusion constants. Additionally, they generated a 10 ns explicit solvent MD trajectory using the CHARMM force field and TIP3P water as a benchmark for their implicit solvent simulations. They also utilized an NMR-derived structure of the Aβ(10-35) peptide as a starting point for some simulations.

Key Insights and Ideas

  • Implicit Solvent Accuracy: The implicit solvent model, when combined with the CHARMM force field, shows excellent agreement with explicit solvent simulations and experimental data for many properties. This suggests it can accurately capture key aspects of Aβ(10-35) behavior.
  • Speed vs. Accuracy Trade-off: The implicit solvent model provides a substantial speedup (up to 200x) compared to explicit solvent simulations, making it feasible to explore longer timescales relevant to aggregation.
  • Force Field Matters: The choice of force field significantly impacts simulation results, particularly for the NMR order parameter. This highlights the need for careful force field selection and potential refinement for accurate protein dynamics.
  • Peptide Flexibility is Key: Aβ(10-35) is highly flexible, adopting a dynamic strand-loop-strand motif. This flexibility, coupled with a solvent-exposed hydrophobic patch (LVFFA region), is crucial for its aggregation propensity.
  • Equilibration Takes Time: Even with the faster implicit solvent model, achieving equilibrium for the Aβ(10-35) peptide requires simulations of at least 30 ns, underscoring the importance of long simulation times for reliable results.

Key Results

  • Diffusion Constant: Implicit solvent simulations with the CHARMM force field yielded diffusion constants close to experimental values and explicit solvent simulations. The OPLS-aal force field also provided good agreement.
  • Structural Properties: Radius of gyration and end-to-end distance from implicit solvent simulations generally agreed well with experimental data and explicit simulations, especially after 30 ns of simulation.
  • NMR Order Parameter (S²): The Zhang-Brüschweiler method for calculating S² showed good agreement with experimental data across different force fields. The CHARMM force field consistently predicted slightly lower S² values but still showed excellent agreement with experimental data for most residues.
  • Solvent Accessible Surface Area (SASA): The hydrophobic LVFFA region remained largely exposed to the solvent across all simulations, consistent with its proposed role in aggregation.
  • Equilibration Time: Simulations starting from various initial structures (NMR, stretched, unfolded) generally required around 30 ns to reach equilibrium.

Limitations and Future Work

  • Force Field Refinement: The study highlights the significant variation in NMR order parameters across different force fields, suggesting a need for further refinement of these force fields for accurately describing flexible protein regions.
  • Experimental Conditions: The authors note that experimental S² measurements were performed at pH 5.7, while their simulations likely correspond to neutral pH, which could influence the comparison. They also mention that experimental S² values were obtained using a different method (Lipari-Szabo), which they found less suitable for flexible peptides.
  • Future Work: The authors express that their validated implicit solvent model is well-positioned for future simulations of Aβ aggregation mechanisms and dynamics, especially for larger systems and longer timescales. They also suggest that future simulations of multiple peptides will be essential to reveal the detailed aggregation process and the role of the central hydrophobic region.

Summary updated

Abstract

A pathogenetic feature of Alzhemier disease is the aggregation of monomeric beta-amyloid proteins (Abeta) to form oligomers. Usually these oligomers of long peptides aggregate on time scales of microseconds or longer, making computational studies using atomistic molecular dynamics models prohibitively expensive and making it essential to develop computational models that are cheaper and at the same time faithful to physical features of the process. We benchmark the ability of our implicit solvent model to describe equilibrium and dynamic properties of monomeric Abeta(10-35) using all-atom Langevin dynamics (LD) simulations, since Alphabeta(10-35) is the only fragment whose monomeric properties have been measured. The accuracy of the implicit solvent model is tested by comparing its predictions with experiment and with those from a new explicit water MD simulation, (performed using CHARMM and the TIP3P water model) which is approximately 200 times slower than the implicit water simulations. The dependence on force field is investigated by running multiple trajectories for Alphabeta(10-35) using the CHARMM, OPLS-aal, and GS-AMBER94 force fields, whereas the convergence to equilibrium is tested for each force field by beginning separate trajectories from the native NMR structure, a completely stretched structure, and from unfolded initial structures. The NMR order parameter, S2, is computed for each trajectory and is compared with experimental data to assess the best choice for treating aggregates of Alphabeta. The computed order parameters vary significantly with force field. Explicit and implicit solvent simulations using the CHARMM force fields display excellent agreement with each other and once again support the accuracy of the implicit solvent model. Alphabeta(10-35) exhibits great flexibility, consistent with experiment data for the monomer in solution, while maintaining a general strand-loop-strand motif with a solvent-exposed hydrophobic patch that is believed to be important for aggregation. Finally, equilibration of the peptide structure requires an implicit solvent LD simulation as long as 30 ns.
BibTeX
@article{kent2008benchmarking,
  author = {Andrew Kent and Abhishek Jha and James E. Fitzgerald and Karl F. Freed},
  title = {{Benchmarking Implicit Solvent Folding Simulations of the Amyloid β(10−35) Fragment}},
  journal = {The Journal of Physical Chemistry B},
  year = {2008},
  doi = {10.1021/jp077099h},
}
Figure