Progress toward automated metabolic profiling of human serum: Comparison of CPMG and gradient-filtered NMR analytical methods

Abstract

The investigation of drug delivery and metabolism requires the analysis of molecules in complicated biological matrices such as human serum. In NMR-based metabonomic analysis, T2 relaxation editing with a CPMG filter is commonly used to suppress background signals from proteins and other endogenous components. Radio frequency pulse imperfections and incomplete irradiation across the spectral bandwidth can cause phase and baseline distortions in CPMG spectra. These distortions are exacerbated by water suppression techniques. Baseline correction methods included in commercially available data processing software packages may be incapable of producing artifact-free spectra. To increase the analytical precision of metabolic profiling, one NMR spectroscopist may be responsible for manually phasing and baseline correcting hundreds of spectra individually to remove operator-dependent variations, significantly reducing throughput. For metabonomic analysis of human serum, it was observed that the application of a pulsed field gradient filter produced 1H NMR spectra well suited to automatic phasing routines. Superior baseline characteristics, an increased tolerance to radio frequency pulse imperfections, and improved water suppression were achieved. A concomitant reduction in signal intensity compared with the CPMG method was easily recovered by increasing the number of scans. Principal component analysis (PCA) of spectra, acquired under a variety of experimental conditions, revealed the improved reproducibility and robustness of1H NMR pulsed field gradient-filtered metabonomic analyses of serum compared to the CPMG method.

Department

Chemistry

Publication Date

9-1-2005

Journal Title

Journal of Pharmaceutical and Biomedical Analysis

Publisher

Elsevier

Digital Object Identifier (DOI)

10.1016/j.jpba.2004.09.060

Document Type

Article

Rights

Copyright © 2005, Elsevier

Share

COinS