Date of Award

2021

Document Type

Restricted Thesis

Degree Name

Bachelor of Arts

Department

Chemistry

First Advisor

K. Aurelia Ball

Abstract

ArkA12 is a proline rich intrinsically disordered protein (IDP) region within a larger, more structured, non-tyrosine kinase (ArkA). The disordered region binds to the AbpSH3 domain due to the domain’s high affinity towards PxxP motifs within the substrate. SH3 domains are numerous in the human genome and are largely responsible for cytoskeleton regulation, signal transduction, and gene expression. A binding mechanism for ArkA12 and AbpSH3 has only recently been elucidated via atomistic computer simulations: depicting ArkA12 in a predominately polyproline II helical structure in the fully bound state, with all five prolines in the trans conformation and no isomerization. The structure of the proline side chain allows peptide bonds to sample between the cis and trans states more readily than the other common amino acids. It is difficult to capture the peptide bond isomerization utilizing classical experimental and computational methods due to the characteristic rapid conformational switching in IDPs, as well as the proline isomerization rates occurring on the order of seconds, making it improbable to detect with the current technology. Multiple proteins have evolved to use proline isomerization as a “switch” for processes requiring quick turnaround times such as gene regulation.

Current experimental circular dichroism data of ArkA12 suggests that the in vitro ensemble may be composed, in part, of structures that contain prolines in cis; reinforcing the need to represent this kind of sampling in the computational data to ensure that the computational ensemble is complete, increasing the likelihood of the data accurately complimenting experimental results. There is currently no data in the literature that has observed proline peptide bond isomerization in proteins of lengths larger than ~5 residues, even with enhanced methods. Computational methods that were able to capture the cis conformation used programs that relied heavily on statistical probability calculations which do not subject conformational ensembles to a primarily time-dependent simulation trajectory: complicating the ease and feasibility of comparing the data to experiments, as well as increasing the likelihood of observing wider deviations from the experiments. The same studies only used peptides containing a single proline or peptides with more than one proline with only one proline captured isomerizing. The largest hinderance to sufficient isomerization sampling is the time that the proline peptide bond requires to flip; this means that the free energy barrier between the cis and trans states may be programmed to be too large in the common molecular dynamics forcefield, thus highlighting the need to counteract the default forcefield effects to accelerate sampling.

The current work uses the gaussian accelerated molecular dynamics algorithm along with a modified peptide torsional angle barrier in the Amberff14SB force field to artificially lower the potential energy landscape to promote faster and more diverse structure sampling. Two microseconds of independent simulation data showed simultaneous cis-trans isomerization for all five prolines in ArkA12 starting from an extended structure. Calculating the theoretical circular dichroism spectrum of the computational ensemble and comparing it to the experimental spectrum can possibly be used to decide whether cis sampling occurs for ArkA12 in vitro, as well as confirm whether the sampling method yields a closer fit to the experiments compared to non-modified sampling methods. The results suggest that this gaussian accelerated-lower barrier computational method can be used to sample proline cis-trans isomerization for similar proline-rich IDPs at the nanosecond timescale without implicitly facilitating certain trajectory processes that may compromise the data’s relevancy to the experimental results.

Creative Commons License

Creative Commons Attribution 4.0 International License
This work is licensed under a Creative Commons Attribution 4.0 International License.

Share

COinS