Fluorine NMR spectroscopy is a powerful tool for studying biomolecular structure, dynamics, and ligand binding, yet the origins of 19F chemical shifts are not well understood. Herein, we use electronic structure calculations to describe the changes in 19F chemical shifts of 2F- and 4F-histidine/(5-methyl)-imidazole upon acid titration. While the protonation of the 2F species results in a deshielded chemical shift, protonation of the 4F results in an opposite, shielded chemical shift. The deshielding of 2F-histidine/(5-methyl)-imidazole upon protonation can be rationalized by concomitant decreases in charge density on fluorine and a reduced dipole moment. These correlations do not hold for 4F-histidine/(5-methyl)-imidazole, however. Molecular orbital calculations reveal that for the 4F species, there are no lone pair electrons on the fluorine until protonation. Analysis of a series of 4F-imidazole analogues, all with delocalized fluorine electron density, indicates that the deshielding of 19F chemical shifts through substituent effects correlates with increased C-F bond polarity. In summary, the delocalization of fluorine electrons in the neutral 4F species, with gain of a lone pair upon protonation may help explain the difficulty in developing a predictive framework for fluorine chemical shifts. Ideas debated by chemists over 40 years ago, regarding fluorine's complex electronic effects, are shown to have relevance for understanding and predicting fluorine NMR spectra.

No abstract

The ability of electronic structure methods (11 density functionals, HF, and MP2 calculations; two basis sets and two solvation models) to accurately calculate the 19F chemical shifts of 31 structures of fluorinated amino acids and analogues with known experimental 19F NMR spectra has been evaluated. For this task, BH and HLYP, ωB97X, and Hartree-Fock with scaling factors (provided within) are most accurate. Additionally, the accuracy of methods to calculate relative changes in fluorine shielding across 23 sets of structural variants, such as zwitterionic amino acids vs. side chains only, was also determined. This latter criterion may be a better indicator of reliable methods for the ultimate goal of assigning and interpreting chemical shifts of fluorinated amino acids in proteins. It was found that MP2 and M062X calculations most accurately assess changes in shielding among analogues. These results serve as a guide for computational developments to calculate 19F chemical shifts in biomolecular environments.

The solvation layer surrounding a protein is clearly an intrinsic part of protein structure–dynamics–function, and our understanding of how the hydration dynamics influences protein function is emerging. We have recently reported simulations indicating a correlation between regional hydration dynamics and the structure of the solvation layer around different regions of the enzyme Candida antarctica lipase B, wherein the radial distribution function (RDF) was used to calculate the pairwise entropy, providing a link between dynamics (diffusion) and thermodynamics (excess entropy) known as Rosenfeld scaling. Regions with higher RDF values/peaks in the hydration layer (the first peak, within 6 Å of the protein surface) have faster diffusion in the hydration layer. The finding thus hinted at a handle for rapid evaluation of hydration dynamics at different regions on the protein surface in molecular dynamics simulations. Such an approach may move the analysis of hydration dynamics from a specialized venture to routine analysis, enabling an informatics approach to evaluate the role of hydration dynamics in biomolecular function. This paper first confirms that the correlation between regional diffusive dynamics and hydration layer structure (via water center of mass around protein side-chain atom RDF) is observed as a general relationship across a set of proteins. Second, it seeks to devise an approach for rapid analysis of hydration dynamics, determining the minimum amount of information and computational effort required to get a reliable value of hydration dynamics from structural data in MD simulations based on the protein–water RDF. A linear regression model using the integral of the hydration layer in the water–protein RDF was found to provide statistically equivalent apparent diffusion coefficients at the 95% confidence level for a set of 92 regions within five different proteins. In summary, RDF analysis of 10 ns of data after simulation convergence is sufficient to accurately map regions of fast and slow hydration dynamics around a protein surface. Additionally, it is anticipated that a quick look at protein–water RDFs, comparing peak heights, will be useful to provide a qualitative ranking of regions of faster and slower hydration dynamics at the protein surface for rapid analysis when investigating the role of solvent dynamics in protein function.

Ribonuclease A is the oldest model for studying enzymatic mechanisms, yet questions remain about proton transfer within the active site. Seminal work by Jackson et al. (Science, 1994) labeled Ribonuclease A with 4-fluorohistidine, concluding that active-site histidines act as general acids and bases. Calculations of 4-fluorohistidine indicate that the π-tautomer is predominant in all simulated environments (by ~17 kJ/mol), strongly suggesting that fluoro-labeled ribonuclease A functions with His119 in π-tautomer. The tautomeric form of His119 during proton transfer and tautomerism as a putative mechanistic step in wild-type RNase A remain open questions and should be considered in future mechanistic studies.