In this project, long-time collaborators Dill, Fennell and Vlachy propose experiments and theory to improve computational models of solvation in complex protein solutions. Explicit- and implicit-solvent models have grown increasingly powerful, but are not able to handle important challenges in modeling proteins in concentrated environments, or in high-salt environments, or in the complex formulations of biological drugs (monoclonal antibodies). Researchers cannot yet reliably compute protein aggregation, solubilities, solution viscosities, the formation of amyloid oligomers or fibrils, or Hofmeister effects, nor can they yet rationally design solution formulations that prevent the precipitation of biologicals, or design optimal conditions for protein crystallization.
The specific aims here are: (1) Statistical mechanics solution theory and systematic experiments on protein type, concentration and salt-series. To apply a new statistical mechanical approach (Wertheim theory for strongly-associating media applied to the KVD model of protein solutions) to treat complex multicomponent mixtures of proteins, with salts and excipients over a range of concentrations, and as a function of temperature. This is a bottom-up first-principles approach (based on a Hamiltonian of all the intermolecular interactions), not based on averaged-solvent such as implicit-solvent or DLVO approaches. There is a huge need for systematic studies of protein association, pairwise and multimeric, vs. salts and conditions. This will be done by Vlachy's group, which has long-standing expertise, and in conjunction with collaborators at Amgen. Extensive preliminary theory and experiment results are in hand. (2) SEA Water and i-PMF. Fast physical simplified models of solvation at the atomistic scale for solvation free energies and PMFs are in development. Recent proofs show SEA Water is as accurate as explicit-solvent and as fast as implicit-solvent. (3) Making our methods available. SEA water methods will be shared with the scientific community by incorporating them into standard molecular dynamic simulation packages like OpenMM and Amber. Additional solubility prediction tools will be offered on a website.
Proteins are the nano-machines of our body. Proteins called monoclonal antibodies are also a major source of drugs, called biologicals, made by Biotech companies. Proteins often associate with each other: sometimes in biologically favorable ways, as when they form complexes to perform biological functions, and sometimes in biolog- ically unfavorable ways, as when they form amyloid, the basis for Alzheimer's and Parkinson's diseases and when they precipitate and aggregate in the formulations of biological drugs. This project will use theory and experiments to improve our understanding of protein interactions and aggregation in a variety of complex environments, such as in cells and in biological formulations. This will help us better understand disease states of proteins and help us provide better drug formulations.
Showing the most recent 10 out of 74 publications
