CaltechTHESIS
  A Caltech Library Service

First-Principles-Based Simulations for G Protein-Coupled Receptor Activation and for Large-Scale Nonadiabatic Electron Dynamics

Citation

Dong, Sijia S. (2017) First-Principles-Based Simulations for G Protein-Coupled Receptor Activation and for Large-Scale Nonadiabatic Electron Dynamics. Dissertation (Ph.D.), California Institute of Technology. doi:10.7907/Z98C9T8D. http://resolver.caltech.edu/CaltechTHESIS:01032017-003321706

Abstract

This thesis focuses on simulating large molecular systems within and beyond the Born-Oppenheimer framework from first principles. Two approaches have been developed for very different but important applications.

The first one is a hybrid method based on classical force fields that predicts the high-energy ensemble of three-dimensional structures of a class of proteins critical in human physiology: the G protein-coupled receptors (GPCRs). GPCRs' functions rely on their activation marked by a series of conformational changes related to binding of certain ligands, but the short of experimental structures has hampered the study of their activation mechanism and drug discovery. Our method, combining homology modeling, hierarchical sampling, and nanosecond-scale molecular dynamics, is one of the very few computational methods that can predict their active-state conformations and is one of the most computationally inexpensive. It enables the conformational landscape and the first quantitative energy landscape of GPCR activation to be efficiently mapped out.

This method, named ActiveGEnSeMBLE, allows the inactive- and active-state conformations of GPCRs without an experimental structure to be systematically predicted. We have validated the method with one of the most well-studied GPCRs, human β2 adrenergic receptor (hβ2AR), and applied the method on a GPCR without an experimental structure, human somatostatin receptor 5 (hSSTR5). Insights on GPCR activation as well as structure prediction methods are discussed.

The second one is a semiclassical approach for large-scale nonadiabatic dynamics of condensed systems in extreme conditions, termed Gaussian Hartree Approximated Quantum Mechanics (GHA-QM). Many nonadiabatic processes related to important applications (e.g. renewable energy) happen in large systems, but existing excited state dynamics methods are too computationally demanding for their long timescale simulations. GHA-QM is based on the electron force field (eFF) framework where we model electrons as Gaussian wavepackets and nuclei as classical point charges, and obtain a simplified solution to the time-dependent Schrödinger equation as the equation of motion. We employ a force field philosophy approximating the total energy as a sum of electronic kinetic energies, electrostatic energies and a Pauli correction, which corrects for the lack of explicit antisymmetry in the wavefunctions. New designs of the Pauli potential and preliminary results on hydrogen systems are discussed. With the new development, we hope to improve the accuracy and range of applications of eFF to simulate the nonadiabatic dynamics of hundreds of thousands of electrons on nanosecond timescale.

Item Type:Thesis (Dissertation (Ph.D.))
Subject Keywords:molecular dynamics; G protein-coupled receptor; somatostatin receptor; nonadiabatic dynamics; semiclassical dynamics; force field
Degree Grantor:California Institute of Technology
Division:Chemistry and Chemical Engineering
Major Option:Chemistry
Minor Option:Computational Science and Engineering
Thesis Availability:Mixed availability, specified at file level
Research Advisor(s):
  • Goddard, William A., III
Group:Materials and Process Simulation Center
Thesis Committee:
  • Heath, James R. (chair)
  • Gray, Harry B.
  • Marcus, Rudolph A.
  • Goddard, William A., III
Defense Date:20 December 2016
Non-Caltech Author Email:sijiadong.nature (AT) gmail.com
Funders:
Funding AgencyGrant Number
Materials and Process Simulation CenterUNSPECIFIED
NSFEFRI-1332411
NIHR01AI040567
DOEDE-SC0014607
Record Number:CaltechTHESIS:01032017-003321706
Persistent URL:http://resolver.caltech.edu/CaltechTHESIS:01032017-003321706
DOI:10.7907/Z98C9T8D
Related URLs:
URLURL TypeDescription
http://dx.doi.org/10.1016/j.bpj.2016.04.028DOIArticle adapted for Chapters 2 and 3
http://dx.doi.org/10.1002/cmdc.201500023DOIArticle adapted for Chapters 3 and 4
ORCID:
AuthorORCID
Dong, Sijia S.0000-0001-8182-6522
Default Usage Policy:No commercial reproduction, distribution, display or performance rights in this work are provided.
ID Code:10001
Collection:CaltechTHESIS
Deposited By: Sijia Dong
Deposited On:12 Jan 2017 01:01
Last Modified:26 Jan 2017 18:22

Thesis Files

[img] PDF (Complete Thesis) - Final Version
Restricted to Caltech community only until 9 January 2018.
See Usage Policy.

19Mb
[img]
Preview
PDF (Front Matter and Chapters 1-4) - Final Version
See Usage Policy.

19Mb
[img] PDF (Chapter 5) - Final Version
Restricted to Caltech community only until 9 January 2018.
See Usage Policy.

600Kb
[img] PDF (Appendix D) - Final Version
Restricted to Caltech community only until 9 January 2018.
See Usage Policy.

138Kb
[img] Archive (ZIP) (Supporting PDB files for Chapters 2-4) - Supplemental Material
See Usage Policy.

1029Kb

Repository Staff Only: item control page