Interaction of the antifolate antibiotic trimethoprim with phosphatidylcholine membranes: a 13C and 31P nuclear magnetic resonance study

GR Painter, R Grunwald and B Roth

The Wellcome Research Laboratories, Research Triangle Park, North Carolina 27709.

The interaction of the bacterial dihydrofolate reductase inhibitor trimethoprim with small unilamellar 1,2-dimyristoyl-sn-glycero-3- phosphorylcholine vesicles was studied using 13C and 31P NMR spectroscopy. In an effort to determine whether trimethoprim passively permeates the vesicle membrane, an impermeant, anionic complex of the paramagnetic ion Dy3+ was added to the extravesicular compartment. Based on the downfield shift that the Dy3+ complex induces in the [2- 13C] resonance of trimethoprim in free solution, membrane permeation and movement of the drug into the intravesicular space can in principle be established from observation of the C2 chemical shift alone. In contrast to what is predicted by a two-compartment model separated by a semipermeable barrier, the presence of vesicles virtually reverses the effect of the shift reagent on the [2-13C] carbon resonance. These results suggest that the majority of the trimethoprim might be sequestered within the vesicle membrane. A saturable decrease in the spin-lattice relaxation time and a saturable increase in the line width at half-height of the [2-13C] resonance as a function of vesicle concentration indicated that trimethoprim does in fact bind to the phospholipid matrix of the membrane bilayer. The KD for the interaction calculated from the relaxation data was 9.7 +/- 0.3 X 10(-4) M at a pH of 7.01 and an ionic strength of 0.015 M. The chemical shift of the [2- 13C] resonance is unaffected by interaction with the electroneutral membrane, and the pKa increases by only 0.16 upon binding. These results point to an interfacial location for the pyrimidine moiety. Using the paramagnetic shift reagent Pr3+ and the 31P NMR signal from the phosphodiester groups of the membrane lipids, trimethoprim was shown to displace Pr3+ ions from binding sites on the outer membrane surface as would be expected if the polar pyrimidine ring were located at or near the membrane surface. The extent to which trimethoprim and trimethoprim derivatives modified in the 3'- and 4'-positions interact with the exo face of the membrane is strongly dependent on the type of substituent and whether it is in the 3'- or 4'-position. Van der Waals interactions between the 5-benzyl sidechain and the hydrophobic fatty acid region of the membrane interior appear to be necessary for the polar portion of trimethoprim to compete favorably for the membrane- binding site with the polyvalent Pr3+ ion.(ABSTRACT TRUNCATED AT 400 WORDS)

Volume 33, Issue 5, pp. 551-558, 05/01/1988
Copyright © 1988 by American Society for Pharmacology and Experimental Therapeutics