Cyclic depsipeptide cyclo-[D-Hmp(1)-L-MeVal(2)-L-Phe(3)-L-MePhe(4)-L-Pro(5)-L-aIle+ ++(6)-L-MeVal(7)-L-Leu(8)-L-betaHOMeVal(9)], the antifungal antibiotic aureobasidin A (AbA), was reported to interfere with ATP-binding cassette (ABC) transporters in yeast and mammalian cells, particularly the MDR1 P-glycoprotein (Pgp), a transmembrane phospholipid flippase or "hydrophobic vacuum cleaner" that mediates multidrug resistance (MDR) of cancer cells. In a standardized assay that measures Pgp function by the Pgp-mediated efflux of the calcein-AM Pgp substrate and uses human lymphoblastoid MDR-CEM (VBL(100)) cells as highly resistant Pgp-expressing cells and the cyclic undecapeptide cyclosporin A (CsA) as a reference MDR-reversing agent (IC(50) of 3.4 microM), AbA was found to be a more active Pgp inhibitor (IC(50) of 2.3 microM). Out of seven natural analogues and 18 chemical derivatives of AbA, several were shown to display even more potent Pgp-inhibitory activity. The Pgp-inhibitory activity was increased about 2-fold by some minor modifications such as those found in the naturally occurring aureobasidins AbB ([D-Hiv(1)]-AbA), AbC ([Val(6)]-AbA), and AbD [gammaHOMeVal(9)]-AbA). The replacement of the [Phe(3)-MePhe(4)-Pro(5)] tripeptide by an 8-aminocaprylic acid or the N(7)()-desmethylation of MeVal(7) led to only a 3.3-fold decreased capacity to inhibit Pgp function, suggesting that the Pgp inhibitory potential of aureobasidins, though favored by the establishment of an antiparallel beta-sheet between the [D-Hmp(1)-L-MeVal(2)-L-Phe(3)] and [L-aIle(6)-L-MeVal(7)-L-Leu(8)-] tripeptides, does not critically depend on the occurrence of the [L-Phe(3)-L-MePhe(4)-L-Pro(5)-L-aIle(6)] type II' beta-turn secondary structure. In contrast, the most potent Pgp inhibitors were found among AbA analogues with [betaHO-MeVal(9)] residue alterations, with some data suggesting a negative impact of the [L-Leu(8)-L-betaHOMeVal(9)-D-Hmp(1)] gamma-turn secondary structure on Pgp inhibitory potential. The [2,3-dehydro-MeVal(9)]-AbA was the most potent Pgp inhibitory aureobasidin, being 13-fold more potent than AbA and 19-fold more potent (on a molar basis) than CsA. Finally, there was no correlation between the SAR for the human MDR1 Pgp inhibition and the SAR for Saccharomyces cerevisiae antifungal activity, which is mediated by an inositol phosphoceramide synthase activity.