Interfacial Recognition by Bee Venom Phospholipase A 2 : Insights into Nonelectrostatic Molecular Determinants by Charge Reversal Mutagenesis † Farideh Ghomashchi, ‡ Ying Lin, ‡ Mark S. Hixon, ‡ Bao-Zhu Yu, § Robert Annand, ‡ Mahendra K. Jain,* ,§ and Michael H. Gelb* ,‡ Departments of Chemistry and Biochemistry, UniVersity of Washington, Seattle, Washington 98195-1700, and Department of Chemistry and Biochemistry, UniVersity of Delaware, Newark, Delaware 19716 ReceiVed October 13, 1997; ReVised Manuscript ReceiVed December 4, 1997 ABSTRACT: The basis for tight binding of bee venom phospholipase A 2 (bvPLA2) to anionic versus zwitterionic phospholipid interfaces is explored by charge reversal mutagenesis of basic residues (lysines/ arginines to glutamates) on the putative membrane binding surface. Single-site mutants and, surprisingly, multisite mutants (2-5 of the 6 basic residues mutated) are fully functional on anionic vesicles. Mutants bind tightly to anionic vesicles, and active-site substrate and Ca 2+ binding are not impaired. Multisite mutants undergo intervesicle exchange slightly faster than wild type, especially in the presence of salt. It is estimated that electrostatic contribution to interfacial binding is modest, perhaps 2-3 kcal/mol of the estimated 15 kcal/mol. Elution properties of bvPLA2 from HPLC columns containing solid phases of tightly packed monolayers of phosphocholine amphiphiles suggest that ionic effects provide a modest portion of the interfacial binding energy and that this contribution decreases as the number of cationic residues mutated is increased. These results are consistent with the observation that Gila monster venom PLA2 (Pa2), which is homologous to bvPLA2, has high activity on anionic vesicles despite the fact that it has only a single basic residue on its putative interfacial recognition face. Results with bvPLA2 mutants show that manoalogue and 12-epi-scalaradial inactivate bvPLA2 by modification of K94. Also, deletion of the large -loop (residues 99-118) is without consequence for interfacial binding and catalysis of bvPLA2. All together, the preferential binding of bvPLA2 to anionic vesicles versus phosphatidylcholine vesicles is mainly due to factors other than electrostatics. Therefore hydrogen-bonding and hydrophobic interactions must provide a major portion of the interfacial binding energy, and this is consistent with recent spectroscopic studies. The hydrolysis of the sn-2 ester of glycerophospholipids by phospholipase A 2 (PLA2) 1 necessarily occurs at the lipid- water interface because naturally occurring phospholipids have virtually no solubility in water as monomers. All PLA2s characterized to date (4-6) are water-soluble en- zymes that must bind to organized interfaces for catalytic turnover (7-11). Fourteen-kilodalton PLA2s are secreted from a number of cell types and are found in animal venoms, pancreatic digestive fluids, and mammalian cells of the inflammatory system. As supported by X-ray structures (12, 13), these PLA2s require Ca 2+ as an essential active-site cofactor to promote binding of a single substrate molecule in the active site and for the chemical step of lipolysis but not for binding to the interface (4, 14, 15). There seems to be a general consensus that the binding of secreted PLA2s to membranes has significant electrostatic and hydrophobic components (16-18). These enzymes bind several orders of magnitude more tightly to anionic vesicles and mixed-micelles compared to neutral interfaces (19-21). In fact, at low ionic strength, the binding of most, if not all, 14-kDa PLA2s to anionic vesicles is virtually irreversible such that the bound enzyme hydrolyzes all of the phospho- lipids in the outer layer of vesicles without desorption of enzyme into the aqueous phase (9, 22, 23); this highly processive interfacial catalysis is termed scooting. Tight binding of 14-kDa PLA2s to anionic interfaces makes sense when one considers that biological interfaces have anionic surface potential due to the presence of anionic phospholipids and bile salts. Increasing ionic strength apparently leads to weaker enzyme-membrane binding such that the enzyme now hops from vesicle to vesicle (24). High-resolution crystal structures are available for several 14-kDa PLA2s (reviewed in ref 18), and all of these † This work was supported by Grant HL36235 and Research Career Development Award GM562 to M.H.G., Grant GM29703 to M.K.J., and postdoctoral fellowship GM15464 to R.R.A., all from the National Institutes of Health. * To whom correspondence should be addressed. ‡ University of Washington. § University of Delaware. 1 Abbreviations: bvPLA2, 16-kDa phospholipase A2 from the venom of honeybees (Apis mellifera); dansyl-DHPE, N-dansyl-1,2-dihexadecyl- sn-glycero-3-phosphoethanolamine; Δ99-118, mutant bee venom phospholipase A 2 in which residues 99-118 have been deleted; diC6- thioPC, 1,2-bishexanoylthio-sn-glycero-3-phosphocholine; diC6thioPM, 1,2-bishexanoylthio-sn-glycero-3-phosphomethanol; DMPC and DMPM, 1,2-dimyristoyl-sn-glycero-3-phosphocholine and -phosphomethanol; DTPC and DTPM, 1,2-ditetradecyl-sn-glycero-3-phosphocholine and -phosphomethanol; E5, K14E/R23E/K85E/K94E/K133E mutant of bee venom phospholipase A 2; EGTA, ethylene glycol bis(-aminoethyl ether)-N,N,N′,N′-tetraacetic acid; i-face, interfacial recognition face of phospholipase A2; Pa2, Pa2 isozyme of Gila monster venom (Helo- derma suspectum) phospholipase A2; PLA2, 14-18 kDa secreted phospholipase A2; WT, wild-type bee venom phospholipase A2. 6697 Biochemistry 1998, 37, 6697-6710 S0006-2960(97)02525-7 CCC: $15.00 © 1998 American Chemical Society Published on Web 04/23/1998