Comparison of the computational NMR chemical shifts of choline with the experimental data

Abstract

One of the main biological markers of the presence of cancer in living patients is an over-expression of total choline (tCho), which is the sum of free choline and its derivatives. 1H Magnetic Resonance Spectroscopy, or H-MRS, enables the quantification of tCho via its proton spectra, and thus has the potential to be a diagnostic tool for the presence of cancer and an accurate early indicator of the response of cancer to treatment. However, it remains difficult to quantify individual choline derivatives, since they share a large structural similarity ((CH3)3-N+-CH2-CH2-O-), of which the strongest signal detectable by MRS is that of the choline "head group": the three methyl groups bonded to the nitrogen. This work used ACENet, a high performance computing system, to attempt to model the NMR parameters of choline derivatives, with the focus of this report being free choline. Optimized structures were determined using Density Functional Theory and the B3LYP electron correlation functional. The Polarizable Continuum Model was used to evaluate solvent effects. The Gauge-Invariant Atomic Orbital method was found to be the superior method for calculating the NMR parameters of cholines.