Computational Amide I Spectroscopy from the ground up : building and benchmarking new tools to study disordered peptide ensembles

Abstract

In the "form follows function" paradigm of structural biology, we seek to understand-and control-protein function based on our knowledge of protein structure. This process is often difficult for intrinsically disordered proteins and peptides (IDPs) which possess inherent structural disorder in their functional forms. A prototypical example is elastin, a disordered structural protein that provides elastic properties to skin, lungs, and other connective tissues. In such cases, the "form" of the protein must be thought of not as an individual structure, but as a heterogeneous ensemble of structures. The characterization of such ensembles is complicated both by the inherent disorder of the system and by the fact that many common experimental techniques function poorly when applied to IDPs. In this work, we present our recent progress in developing experimental and computational tools for characterizing IDP ensembles using Amide I (backbone carbonyl stretch) vibrational spectroscopy. In this approach, the infrared (IR) absorption frequencies of isotope-labeled amide bonds act as sensitive probes of local electrostatic environment and, ultimately, of local structure. By producing and characterizing experimentally a progression of increasingly complex model systems ranging from dipeptide fragments to isotope-labeled proteins, we develop an efficient and robust spectroscopic model capable of predicting Amide I vibrational frequencies from atomistic protein structures to within a few cm-¹ of error. We apply these methods to the analysis a family of short (eight-residue) elastin-like peptides (ELPs), fragments of the elastin protein, whose local structure is believed to be critical to elastin function. Using our empirically-parameterized frequency maps, we test and refine molecular dynamics ensembles by quantitative comparison against isotope-labeled experimental data. This combination of isotope-labeled IR data, high level spectroscopic modeling, and well-sampled molecular dynamics ensembles provides both local conformational insight into the molecular structures underlying elastic function and a point of departure for testing and refining force-field based structure prediction methods for intrinsically disordered systems.