Conserved Functional Dynamics in the Cell Signaling Protein Pin1: Investigations by Bioinformatics and NMR

In the past few years, several functional site prediction methods have been introduced that aim to predict the biologically relevant amino acids in proteins based purely off of a multiple sequence alignment. These methods utilize a variety of approaches including conservation, divergence, and perturbation to identify the functionally important amino acid positions (primary sequence or 3D structure) within the protein sequence. Yet, the question remains if these methods accurately predict the functionally important residues within a protein found experimentally. This dissertation investigates the capability of the bioinformatic functional site prediction methods. Residues known to participate in conformational dynamics associated with substrate binding, as observed experimentally by Nuclear Magnetic Resonance (NMR) spectroscopy, are compared to those residues highlighted by the prediction methods.

The comparison of results given above will provide insight into the protein evolution-dynamics-function relationship. A test of how well the current prediction methods perform at selecting the functionally relevant residues is presented herein.

Specifically, because dynamics are encoded in the protein sequence and are required for function, then a mutation of a functionally important position will alter both the internal dynamics and its activity. The time scales for such motions can be on the order of slow (μs-ms) to fast (ps-ns) motions. Nuclear Magnetic Resonance (NMR) is a non-invasive technique which probes such dynamics relevant for function.

Overlap exists between the bioinformatic prediction methods and the experimental data for many of the residues used in the model system, human Pin1. Many of the residues identified experimentally by NMR to have conformational dynamics due to substrate binding are the same residues predicted to be functionally relevant from the functional site prediction methods. For a select few amino acid positions, the agreement between the two approaches is not perfect. What follows in this thesis is a discussion on two mutations of Pin1. For the one mutation, M130A, agreement exists between the functional site prediction methods and experimental intrinsic dynamics. In the other, I28A, the agreement between the two methods fails.