Biomolecular Transduction of Mechanical Signals: Simulation Study of Force-Modulated Protein Activity

Mechanical signals have been shown to regulate various physiological behaviors, including growth, differentiation, apoptosis, motility and gene expression. The details of the processes by which these mechanical cues result in cellular change, or mechanotransduction, are largely unknown. Abnormal mechanotransduction can arise through changes in cell mechanics, variations in extracellular matrix structure or deregulation of the molecular mechanisms by which cells sense and respond to mechanical cues. This has been implicated in pathological conditions, such as atherosclerosis, asthma and carcinogenesis. One proposed mechanism for mechanotransduction is that a force-driven conformational change results in altered protein activity, such as force-modulated binding affinity. In this framework, key protein molecules act as mechanosensors and mechanotransducers; the combined effect of these functions is to convert mechanical signals into chemical ones, thereby connecting to cellular biochemistry and downstream effectors. We are taking an integrated approach to investigate key features of molecular events related to mechanotransduction, with a focus on biophysical modeling of the mechanisms of force-regulated binding properties. This work includes an analysis of simple mechanical problems that provide insights applicable to more complex systems, as well as simulation studies on model problems and full-scale proteins of mechanical interest.
Methods

Computer simulations provide atomistic information to aid in the design and analysis of experimental studies. Connecting atom-level simulations with macroscopic observables, specifically force-dependent binding affinity in the case of mechanotransduction, is essential to interpret experiments. Current simulation methods allow the estimation of free energy changes in the context of chemical perturbations. In order to expand existing techniques and apply them to mechanotransduction, we built extensions to incorporate an applied force as the perturbing factor. Method development is coupled with careful validation on theoretical systems for which the free energy can be computed independently. We designed potential energy landscapes based on experimental and computational characterizations of those of proteins, including overall shape and roughness, representative energy barrier heights and well spacing.

In parallel, we have chosen a protein system relevant to mechanotransduction for biological applications. We are focusing the efforts on focal adhesion targeting (FAT) domain of focal adhesion kinase, binding to a paxillin peptide. We model the protein in its bound and unbound states from the available crystal structure.

Our designed complex theoretical systems allowed us to identify key factors influencing accuracy, in particular the extent of phase space exploration, leading to the need for extensive sampling methods. Computations are performed with techniques identical to protein simulations, and important parameters such as equilibration and production times are being optimized. Thus, the use of model systems has permitted us to test and refine the computational methodology. This is particularly important to achieve converged simulations and accurate results.

Applications of the free energy method to protein systems will provide detailed understanding of the molecular mechanisms involved in force-modulated protein activity. By probing an array of force application schemes, varying both force intensity and directionality, we expect to gain much insight into atomistic events leading to cellular-level changes. Progress in this area could lead to applications in biomedicine, where molecules mediating mechanotransduction can represent targets for therapeutic intervention.