Proton transfer in guanine–cytosine base pair analogues studied by NMR spectroscopy and PIMD simulations. (20th September 2018)

• Record Type: Journal Article
• Authors: Pohl, Radek; Socha, Ondřej; Slavíček, Petr; Šála, Michal; Hodgkinson, Paul; Dračínský, Martin
• Is Part Of: Faraday discussions. Volume 212(2019)
• Page Start: 331
• Page End: 344
• Publication Date: 2018-09-20
• Subjects: Chemistry -- Periodicals; Metallurgy -- Periodicals; Electrochemistry -- Periodicals; 540
• DOI: 10.1039/c8fd00070k
• Languages: English
• ISSNs: 1359-6640

Abstract

Abstract : NMR isotope shifts provide a unique possibility to study the nuclear quantum effects of hydrogen-bonded nucleobases. Abstract : It has been hypothesised that proton tunnelling between paired nucleobases significantly enhances the formation of rare tautomeric forms and hence leads to errors in DNA replication. Here, we study nuclear quantum effects (NQEs) using deuterium isotope-induced changes of nitrogen NMR chemical shifts in a model base pair consisting of two tautomers of isocytosine, which form hydrogen-bonded dimers in the same way as the guanine–cytosine base pair. Isotope effects in NMR are consequences of NQEs, because ro-vibrational averaging of different isotopologues gives rise to different magnetic shielding of the nuclei. The experimental deuterium-induced chemical shift changes are compared with those calculated by a combination of path integral molecular dynamics (PIMD) simulations with DFT calculations of nuclear shielding. These calculations can directly link the observable isotope-induced shifts with NQEs. A comparison of the deuterium-induced changes of 15 N chemical shifts with those predicted by PIMD simulations shows that inter-base proton transfer reactions do not take place in this system. We demonstrate, however, that NMR isotope shifts provide a unique possibility to study NQEs and to evaluate the accuracy of the computational methods used for modelling quantum effects in molecules. Calculations based on the PBE functional from the general-gradient-approximation family provided significantly worse predictions of deuterium isotope shifts than those with the hybrid B3LYP functional.