Identification of a Collagenase-Inhibiting Flavonoid from Alchemilla vulgaris Using NMR-Based Metabolomics

This paper describes the use of ¹H NMR profiling and chemometrics in order to facilitate the selection of medicinal plants as potential sources of collagenase inhibitors. A total of 49 plants with reported ethnobotanical uses, such as the healing of wounds and burns, treatment of skin-related diseases, rheumatism, arthritis, and bone diseases, were initially chosen as potential candidates. The in vitro collagenase inhibitory activity of hydroalcoholic extracts of these plants was tested. Moreover, their phytochemical profiles were analyzed by ¹H NMR and combined with the inhibitory activity data by an orthogonal partial least squares model. The results showed a correlation between the bioactivity and the concentration of phenolics, including flavonoids, phenylpropanoids, and tannins, in the extracts. Considering the eventual false-positive effect on the bioactivity given by tannins, a tannin removal procedure was performed on the most active extracts. After this procedure, Alchemilla vulgaris was the most persistently active, proving to owe its activity to compounds other than tannins. Thus, this plant was selected as the most promising and further investigated through bioassay-guided fractionation, which resulted in the isolation of a flavonoid, quercetin-3-O-β-glucuronide, as confirmed by NMR and HRMS spectra. This compound showed not only a higher activity than other flavonoids with the same aglycone moiety, but was also higher than doxycycline (positive control), the only Federal Drug Administration-approved collagenase inhibitor. The approach employed in this study, namely the integration of metabolomics and bioactivity-guided fractionation, showed great potential as a tool for plant selection and identification of bioactive compounds in natural product research.