Abstract
This work demonstrates the use of the GruPol database to predict the functional group dipole moments and polarizabilities of glucagon in the presence of NaCl, simulating an electric charge distribution on the protein's backbone. A new feature of the database allows for the inclusion of ions on the protein backbone, effectively simulating a protein salt and predicting the impact on electrical properties. Glucagon was selected as a proof-of-concept molecule due to its relatively small chain, which enabled benchmarking against quantum mechanical calculations. Firstly, we simulated 70 different ionic configurations, varying the number of Na(+) and Cl(-) ions from zero to four NaCl moieties. Additionally, we investigated the effects of solvation under two distinct conditions: one involving just the peptide and water, and the other also including NaCl at a concentration of approximately 4.2 mol L(-1). Regarding the ab initio results, GruPol showed good accuracy, with an angular direction error of around 10° and a 15% difference in the magnitude of the dipole moments. However, the error in polarizability values was higher, most likely due to the lack of an augmented basis set in the ab initio quantum calculations (M06-HF/cc-pVDZ). The database entries were generated using the same functional along with the aug-cc-pVDZ basis set. In solution, a high ionic concentration lowered the overall dipole moment, while the main components of polarizability increased.