Errors in Measuring and Interpreting Values of Coupling Constants J from PE.COSY Experiments

Abstract

The accuracy of the measured coupling constants J in a PE.COSY experiment is discussed in terms of the linewidth effects and, more importantly, the effect of transverse cross relaxation, which modulates the chemical shift, phase, and linewidth of the multiplet components. The 36° PE.COSY spectra were calculated as a function of the correlation time for a scalar-coupled four-spin system by solving the complete Liouville equation. The interproton distances and J values of the spin system were chosen to model accurately the H1′-H2′ R-H2′ S-H3′ atoms of a typical B-form DNA sugar ring. The results show that cross-relaxation effects modulate the measured coupling constants J H1′-H2′ R and J H1′-H2′ S . Furthermore, the linewidth effect, which may cause unequal cancellation in the PE.COSY cross-peak pattern, also compromises the accuracy of the J (H1′-H2′ R ) measurement, but not that of the J H1′-H2′ S measurement. Conditions for equal cancellation in E.COSY or PE.COSY cross-peak patterns are discussed. The trend in the calculated apparent values of coupling constants for H1′-H2′ R and H1′-H2′ S with increasing correlation time shows that, if the effects of transverse cross relaxation and/or unequal cancellation are ignored in analyzing DNA sugar conformation, either an unrealistically exaggerated N-S dynamic equilibrium or an erroneous static sugar conformation will be inferred using Karplus-type relationships calibrated on small molecules.