The discovery of bioactive natural
products remains a time-consuming
and challenging task. The ability to link high-confidence metabolite
annotations in crude extracts with activity would be highly beneficial
to the drug discovery process. To address this challenge, HPLC-based
activity profiling and advanced UHPLC-HRMS/MS metabolite profiling
for annotation were combined to leverage the information obtained
from both approaches on a crude extract scaled down to the submilligram
level. This strategy was applied to a subset of an extract library
screening aiming to identify natural products inhibiting oncogenic
signaling in melanoma. Advanced annotation and data organization enabled
the identification of compounds that were likely responsible for the
activity in the extracts. These compounds belonged to two different
natural product scaffolds, namely, brevipolides from a Hyptis
brevipes extract and methoxylated flavonoids identified in
three different extracts of Hyptis and Artemisia spp. Targeted isolation of these prioritized compounds led to five
brevipolides and seven methoxylated flavonoids. Brevipolide A (1) and 6-methoxytricin (9) were the most potent
compounds from each chemical class and displayed AKT activity inhibition
with an IC50 of 17.6 ± 1.6 and 4.9 ± 0.2 μM,
respectively.