Modeling and predicting co-translational protein folding with chemical kinetic and molecular dynamic simulations

Open Access
- Author:
- Nissley, Daniel
- Graduate Program:
- Chemistry
- Degree:
- Doctor of Philosophy
- Document Type:
- Dissertation
- Date of Defense:
- December 06, 2018
- Committee Members:
- Edward P. O'Brien, Dissertation Advisor/Co-Advisor
Edward Patrick O'Brien Jr., Committee Chair/Co-Chair
Philip C Bevilacqua, Committee Member
William George Noid, Committee Member
Reka Z Albert, Outside Member - Keywords:
- co-translational folding
protein synthesis
protein folding
synonymous codon
non-equilibrium - Abstract:
- Proteins are linear polymers of amino acids that perform myriad functions within living cells. Most proteins must form a specific three-dimensional structure, often referred to as the native state, in order to perform their biological function. The process of reaching the native state is termed protein folding. It is often assumed that thermodynamics alone determines the specific conformation of the native state of a protein, meaning that its structure is determined by the amino-acid sequence alone. Within living cells, however, proteins are translated based on mRNA templates by the ribosome in a distinctly out-of-equilibrium fashion. The non-equilibrium nature of translation means that kinetics can trump thermodynamics in determining the conformation of a protein, and recent experiments indicate that seemingly small changes to the kinetics of translation can radically alter protein structure and function and even lead to disease. One way in which the speed of translation can be perturbed is through synonymous codon mutations, which change the rate of translation but not the primary sequence of the nascent protein that is produced. Understanding how the non-equilibrium nature of protein synthesis influences the likelihood that a given protein will correctly fold and function is therefore critical to understanding protein biogenesis. This thesis contains four theoretical and computational studies of co-translational protein folding and its influence on protein conformations after synthesis is complete. The state of the experimental and computational literature is summarized in Chapter 1. In Chapter 2 I describe a chemical kinetic model that is able to accurately predict experimental co-translational folding probabilities for the first time. This chemical kinetic model is used to make the novel prediction that some yeast proteins which typically fold post-translationally can be made to fold co-translationally by recoding their mRNA sequences to contain the slowest-translating synonymous codon at each codon position. This chemical kinetic model is general, in principle, for all proteins and all organisms. The fluorescent technique FRET has recently been used to monitor protein folding on the ribosome both in vitro and in vivo. One disadvantage of such fluorescent techniques is that they produce a single value per time point, providing minimal structural information. In Chapter 3, I present results from low-friction, coarse-grain Langevin dynamics simulations that elucidate the structural origins of FRET measurements on the ribosome. These simulations are in strong agreement with experimental time series and reveal the underlying co-translational folding trajectory at a spatial resolution of 3.8 Å. I also show that the alternative hypothesis that nascent chain compaction occurs due to collapse as is expected for a polymer in poor solvent is not consistent with the experimental data. Dimensional collapse, however, could not be ruled out. I therefore suggest alternative dye positions that could be used to differentiate domain folding and domain dimensional collapse in future experiments. Chapter 4 of this thesis presents the hypothesis that the pathogenesis of Huntington’s Disease is, at least in part, due to the dysfunction of a co-translational process. The genetic cause of Huntington’s Disease is the expansion of a poly-glutamine region in Exon 1 of the HTT gene. Individuals with 35 or more CAG codons (which encode the amino acid glutamine) in their HTT gene will suffer disease symptom onset within a typical human lifespan. The age of symptom onset decreases linearly as the number CAG codons increases beyond 35. Based on strong circumstantial experimental evidence and a simple kinetic model, I argue that the expansion of the CAG codon region in the transcript leads to an increase in the speed of translation at a key time during the synthesis of huntingtin protein that leads to its misprocessing and the onset of disease symptoms. I go on to propose experiments to test this hypothesis. In Chapter 5 I describe high-throughput simulations of the translation elongation, translation termination, and post-translational dynamics of a representative subset of the E. coli cytosolic proteome. I find that roughly one in four proteins is kinetically trapped for as long as 3 minutes after the completion of protein synthesis. Finally, in Chapter 6 I summarize the conclusions that can be drawn from, and the future directions related to, my work. One obvious future goal is the extension of my simulations of multi-domain E. coli proteins to the study of how domain interfaces influence protein folding. That is, do domains fold independently or do they rely upon one another? My E. coli proteome data set also includes diverse information about translation termination kinetics, and preliminary simulations reveal that electrostatic effects seem to largely determine its timescales. In summary, the results presented in this thesis advance understanding of co-translational protein folding and the influence translation kinetics can have on protein conformations.