Phospholipase A2 activity in lysates of mast cells such as rat mastocytoma RBL-2H3 cells and mouse bone marrow-derived IL-3-dependent mast cells (BMMC) was measured using phosphatidylcholine (PC), phosphatidylethanolamine (PE), or phosphatidylserine (PS) as a substrate. Both types of cells exhibited phospholipase A2 activity with a similar pH profile; the optimum pH observed with PS as a substrate was 5.5-7.4, whereas that with PE or PC was 8.0-9.0. PE and PC bearing an arachidonate at the sn-2 position were cleaved more efficiently by PE, PC-hydrolyzing phospholipase A2 than phospholipids with a linoleate. A monoclonal antibody raised against rabbit platelet 85-kDa cytosolic phospholipase A2 absorbed the PE, PC-hydrolyzing activity. PS-hydrolyzing activity was purified from RBL-2H3 cells and BMMC by sequential heparin-Sepharose, butyl-Toyo-pearl, and reverse-phase HPLC. On reverse-phase HPLC, the PS-hydrolyzing activity of RBL cells was separated into two peaks, A and B. The peak B activity was inhibited by the anti-rat 14-kDa group II phospholipase A2 antibody, while the peak A activity was not. The partially purified peak A activity hydrolyzed PS about 10-fold more efficiently than PE at optimum pH of 5.5-7.4. No appreciable hydrolysis was observed with PC or phosphatidylinositol (PI). Thus, mast cells may express at least three distinct phospholipases A2; 14-kDa group II phospholipase A2, 85-kDa cytosolic arachidonate preferential phospholipase A2, and a novel phospholipase A2 that shows high substrate specificity for PS.