Can NMR-HetCA\nbe a Reliable Prediction Tool for the\nDirect Identification of Bioactive Substances in Complex Mixtures?

Conventional isolation methods in natural products chemistry\nare\ntime-consuming and costly and often result in the isolation of moderately\nactive compounds or the detection of already known natural products\n(NPs). A fast and cost-effective way to identify bioactive metabolites\nin plant extracts prior to isolation has been developed based on the\nnuclear magnetic resonance (NMR)-heterocovariance approach (NMR-HetCA).\nIn order to evaluate in depth the application of this chemometrics-based\ndrug discovery methodology, simple mixtures of 10 standard NPs simulating\na fast centrifugal partition chromatography (FCPC) fractionation (artificial\nfractions, ArtFrcts), as well as a more complex mixture of 59 natural\nstandard substances simulating a crude plant extract (artificial extract,\nArtExtr), were prepared. FCPC was employed for the fractionation of\nthe ArtExtr, while the inhibitory activity of all fractions against\nDPPH was evaluated, and their chemical profile was recorded using\nNMR spectroscopy. Spectral information was processed in the MATLAB\nenvironment, and statistical approaches, including HetCA and statistical\ntotal correlation spectroscopy (STOCSY), were applied to identify\nbioactive compounds. Total heterocovariance plots (pseudospectra)\nfacilitated the detection of highly correlated metabolites and led\nto the direct identification of 52.6% of the active compounds. The\nsuccess in identifying the ArtExtr bioactive substances increased\nto 63.2% when spectral alignment was implemented. HetCA incorporates\nchromatographic (fractionation), spectroscopic (NMR profiling), and\nbioactivity results along with advanced chemometrics and could be\nestablished as a method of choice for the rapid and effective identification\nof bioactive NPs in plant extracts prior to isolation.