Aqueous leaf extracts of Boldoa purpurascens are widely used because of their diuretic, natriuretic, antiurolithiatic, anti-inflammatory and antihypertensive properties. The major component of the extract is the flavonoid 4′,5-dihydroxy-6,7-methylenedioxyflavonol-3-O-α-L-rhamnopyranosyl-(1→2)-β-D–xylopyranoside, but it is not known if this compound is responsible for the biological activity. The objective of this work is to develop effective in silico tools that allow predicting the possible activity of the flavonoid aglycone as an inhibitor of metalloproteases that regulate renal fluid excretion. First, a mathematical ligand-based classification model was developed, using an artificial intelligence and machine learning technique of support vector machines to find the relationship between chemical structure and biological activity. This showed good fit of the statistical parameters with an accuracy greater than 90%, offering a priori information of the flavonoid activity. Subsequently, the flavonoid aglycone was docked to the active site of the enzymes thermolysin (PDB: 6YMS), angiotensin-converting enzyme (PDB: 6TT4) and neprilysin (PDB: 6SUK) using the Extra Precision glide method (Glide-XP), showing conformations with binding energies lower than −5 Kcal/mol. In this study, possible interactions were determined at the catalytic site, where the coordination of negatively charged pharmacophoric groups with the zinc atom of these enzymes is observed. Finally, a preliminary in vivo evaluation was carried out using a diuresis-natriuresis model with sodium quantification in urine which revealed good activity profiles. These results are in correspondence with the ethnopharmacological use of the plant as a diuretic-natriuretic and for the treatment of hypertension.
Read full abstract