Protein protic and aprotic interactions systematically mapped via IR spectroscopy and polarizable molecular dynamics

Nitrile vibrational frequencies can readily report on the environmental electric field except in hydrogen bonding (H-bonding) environments, where factors other than the field contribute to frequency shifts; these environments are very commonly found in proteins, among other forms of condensed phase matter, so an alternative observable for unambiguous field determination is desirable. We previously showed that the nitrile's transition dipole moment (TDM) – and therefore peak area – experiences a new vibrational Stark effect that linearly correlates with the field in all environments, unlike frequencies. As a proof-of-concept, we demonstrated this relationship using nitrile probes in solvents and applied it to four variants of photoactive yellow protein (PYP) with nitrile-containing ortho-cyanophenylalanine residues introduced via amber suppression (JACS, 144, 7562-7567 (2022)). Eight further PYP mutants were constructed from one variant to modify the H-bonding/electric fields experienced by the nitrile probe. IR spectroscopy yielded four consistent IR lineshapes, interpreted as environmental phenotypes, which included mixtures of protic and aprotic environments. TDM-based electric fields were assessed, and high-resolution crystal structures were obtained to further interpretation of the spectroscopic data. The structures showed that either the mutated residue or newly accessible waters around the nitrile serve as the H-bond donor(s) and could explain the environmental phenotypes on a qualitative level. For a more quantitative assessment, we performed molecular dynamics (MD) on PYPs with polarizable AMOEBA force fields: a strong correlation is found between average electric fields from simulation and experimental TDM analysis, further establishing that the field-TDM relationship extends to complex protein environments. Also, characterizations of the nitrile environment (average structure, H-bond donor identities, etc.) correlated well with environmental phenotype, implying that polarizable force fields realistically model local protein interactions.