Linear prediction enhancement of 2D heteronuclear correlated spectra of proteins.

Linear prediction has been used to extrapolate the t1 domain of natural abundance 1H-13C correlated two-dimensional (2D) FIDs of insulin. The FIDs were obtained by two different heteronuclear correlation experiments, one that utilizes heteronuclear multiple-quantum coherence during t1, and one that utilizes 13C single-quantum coherence. It is shown that the enhancement of the resolution and sensitivity in the F1 dimension of the Fourier transform spectrum that results from the linear prediction extrapolation allows the t1 domain to be confined to a relatively short time period where the signal intensity is at maximum. In particular, it is found that the enhancement thus obtained is sufficiently good to allow an observation of the difference between the F1 line widths in the single-quantum and double-quantum coherence spectra.