Planta Med 2008; 74 - PG23
DOI: 10.1055/s-0028-1084776

In silico discovery of natural chemosensitizers from Eriobotrya japonica

PH Pfisterer 1, JM Rollinger 1, L Schyschka 2, AM Vollmar 2, H Stuppner 1
  • 1Institute of Pharmacy/Pharmacognosy, Center for Molecular Biosciences Innsbruck, Leopold-Franzens University, Innrain 52c, 6020 Innsbruck, Austria
  • 2Department of Pharmacy, Center of Drug Research, Pharmaceutical Biology, Ludwig-Maximilians Univesity, Butenandtstr. 5–13 Building B, 81377 Munich, Germany

Triterpene acids and polyphenols of the medicinal plant loquat, Eriobotrya japonica Lindl. (Rosaceae), are described in literature showing cytotoxic and antitumor-promoting effects [1,2]. Their molecular mechanism has not been elucidated yet. This prompted us to investigate the potential interactions of E. japonica-constituents on the X-linked inhibitor of apoptosis protein (XIAP). Targeting the baculovirus IAP repeat (BIR) 3 domain of XIAP represents a promising strategy for the discovery of chemosensitizers [3]. In this study we used a previously generated and validated pharmacophore model on the XIAP BIR3 domain [4] as search query for the virtual screening of a small 3D multiconformational database. It consists of 116 molecules of known secondary metabolites from the leaves of E. japonica. The putative hits (26.7%) of the virtual screening procedure mainly belong to the phytochemical classes of acylated and non acylated triterpene acids. In order to evaluate the reliability of the pharmacophore model, virtual hits and further non predicted constituents were isolated by using different chromatographic methods. Their structures were confirmed by LC-MS and 1D- and 2D-NMR. In the Nicoletti test [5] 3-O-p-coumaroyl tormentic acid was identified as the most promising compound. The combination of sub-optimal concentrations of the chemotherapeutic etoposide with this virtual hit strongly induced cell death in S-Jurkat and XIAP overexpressing Jurkat cells. The combination of etoposide with compounds, which have not been virtually predicted, showed no chemosensitizing effects at all. These findings underline the high quality of the pharmacophore model and the efficiency of the in silico strategy.

References: 1. Banno, N. et al. (2005) Biol. Pharm. Bull. 28:1995–99. 2. Ito, H. et al. (2000) Chem. Pharm. Bull. 48:687–693. 3. Schimmer, AD. (2006) 13:179–188. 4. Bliem, CB. et al. (2006) Planta Med 72: P084. 5. Nicoletti, I. et al. (1991)J. Immunol. Methods 139:271–279.