Protein dynamical distribution as ensemble in their phase space is directly linked to allostery. Some triggering events, such as binding process or chemical modification, could cause significant change of protein conformation as well as dynamics distribution. These led to two complementary allosteric models: conformation-driven and dynamics-driven allostery. Mutations of certain amino acid residues could alter distribution of protein conformation or dynamics, and regulate its allosteric function. This suggests that overall space for conformational and dynamical distribution potentially accessible for certain protein is much larger than those of specific structure, such as wild type. Certain perturbation including mutation could cause protein to dwell in additional (referred as hidden) conformational space or dynamical states. Therefore, directing proteins into these hidden states could serve as a novel approach in protein engineering. We propose to apply rigid residue scan method recently developed in our lab and machine learning models for data mining to systematically explore hidden conformational space and dynamical states of proteins with quantitative evaluations based on individual residue assessment. Our long-term goal is to exploit hidden conformational spaces and dynamical states of proteins and develop precise regulations of protein allostery and function. Two parallel and complementary specific aims are proposed to test our hypothesis that many proteins have hidden conformational space and dynamical states inaccessible to the wild type but could be accessed through certain perturbation, and that additional space and states lead to precise allostery regulation or even new protein functions.
In Aim 1, we propose to identify residues guiding conformational driven allostery to reach certain hidden conformational space for conformation driven allosteric circadian clock proteins. We will apply rigid residue scan method to efficiently sample hidden conformational space, and develop dimension deduction and Markov state model to theoretically describe protein conformational space. We will also apply machine learning and feature selection methods to develop classification model to quantitatively measure the conformation driven allostery.
In Aim 2, we propose to identify residues of dynamics driven allosteric protein as keys to reach hidden dynamical states for dynamics driven allosteric circadian clock proteins. Similarly, we will apply rigid residue scan method to efficiently sample hidden dynamical space, and apply machine learning and feature selection methods to quantitatively measure the dynamics driven allostery. Mutants will be designed for the identified key residues in both aims. These mutants will be subjected to further molecular simulation for validation purpose. Only those computationally validated mutants will be subject to experimental test. We expect that the proposed research will lead to a novel theoretical model of protein allostery and provide computational tools to quantitatively predict the impacts on protein allostery from specific residue, which could benefit experimental study of the circadian clock proteins to design mutants as novel optogenetic tools.
s Protein allostery regulates protein functional activities by perturbation at a distant site. Allosteric modulation has become a novel approach to develop optogenetic tools to control biological processes. Here we propose to exploit hidden conformational spaces and dynamical states of proteins and develop precise regulations of protein allostery and function. Our efforts will lead to a novel theoretical model of protein allostery and provide computational tools to quantitatively predict the impacts on protein allostery from specific residue. This could greatly benefit experimental study of the circadian clock proteins to design mutants with new allosteric states as optogenetic tools.
