(505a) Modeling Entropy Changes upon Peptide Binding

Peptides are promising recognition elements for many biosensing applications. Small in size and easy to produce through biological expression pathways and chemical synthesis, peptides have many favorable biosensing features. Peptide’s bind to other molecules through a thermodynamic process governed by changes in Gibbs free energy, which in turn is determined from two other thermodynamic properties: changes in enthalpy and entropy. A current challenge associated with molecular mechanics force fields is the significantly limited manner in which they predict entropy changes, resulting in flawed predictions of Gibbs free energy changes. Accurately predicting both thermodynamic properties that determine changes in Gibbs free energy can be helpful for consistent and successful computational peptide design. This work describes the development of a model that rapidly predicts entropy changes upon peptide binding to guide peptide design overall.

In this study, we focused on collecting and analyzing a data set of 161 protein-peptide complexes with experimentally determined structures and experimentally measured binding properties. Each complex is composed of a peptide between 5 and 20 amino acids in length bound to a single-domain protein. The systems are entirely composed of the twenty common amino acids, with no post-translational modifications or non-canonical amino acids. Prior literature had reported for each complex Isothermal Titration Calorimetry experiments sufficient to allow for the calculation of entropy changes upon peptide binding: changes in enthalpy, temperature, and either changes in Gibbs free energy or dissociation constants. Upon identification of this data set, we performed Molecular Dynamics simulations on the unbound peptides and bound complexes to observe the conformational changes the peptides undergo when bound. This presentation will discuss our findings from the collected protein-peptide complex data set, specifically the features that are shown to be most important in characterizing entropy change upon peptide binding. With the lessons learned from both experimental (ITC) and computational (MD simulations) experiments, we are designing peptide recognition elements for biosensors as part of a larger project on designing in-line, continuous sensors for biomanufacturing applications.