The chemical basis for interfacial activation of monomeric phospholipases A2. Autocatalytic derivatization of the enzyme by acyl transfer from substrate

A basic monomeric phospholipase A2 from the venom of the American water moccasin, Agkistrodon piscivorus piscivorus, undergoes Ca2+-dependent, autocatalytic acylation during the course of hydrolysis of both model and natural phospholipid substrates. Acylation occurs at 2 lysine residues, Lys-7 and Lys-10, in the NH2-terminal alpha-helical segment of the enzyme, and when both positions are fully derivatized, the stable bisacylphospholipase A2 becomes a dimer in solution. The acylated enzyme is fully activated toward monomolecular layers of lecithins. Similar studies applied to the monomeric phospholipases A2 from porcine pancreas and from the venom of Agkistrodon contortrix contortrix also showed irreversible activation of the enzymes by substrate with the same kinetic consequences and formation of dimers. Acylation thus enables these enzymes to overcome the lag period observed under such conditions with native monomeric phospholipases, a phenomenon referred to as interfacial activation. Activation of the enzyme by acylation potentiates the phospholipase for interfacial recognition via formation of a dimeric enzyme. The naturally occurring phospholipase A2 dimer from Crotalus atrox venom displays no lag in the hydrolysis of lecithin monolayers nor does it undergo substrate level acylation. These facts support our proposal that dimerization concomitant with acylation is responsible for the large rate enhancements seen in the hydrolysis of aggregated phospholipids by monomeric phospholipases. Our findings demonstrate for the first time a chemical mechanism for interfacial activation of and interfacial recognition by phospholipases A2.     

Properties of phospholipase A2 from the venom of the large hornet Vespa orientalis

Some properties (catalytic and hemolytic activity, pH and temperature optima, stability, substrate specificity, effects of detergents and metal ions, N-terminal sequence, chemical modification of histidine in the enzyme active center, etc.) of phospholipase A2 from hornet (Vespa orientalis) venom were studied. It was shown that phospholipase A2 from hornet venom differs essentially from other enzymes of this species in terms of stability, catalytic properties and structural features. The active center of the enzyme contains an essential histidine residue, similar to other phospholipases A2 from various sources. Unlike other known forms of phospholipase A2, the enzyme under study exerts a pronounced hemolytic action. The hemolysis is inhibited by Ca2+ at concentrations capable of inducing the activation of the hydrolytic activity of the enzyme.     

Trypsin-activated complex of human factor B with cobra venom factor (CVF), cleaving C3 and C5 and generating a lytic factor for unsensitized guinea pig erythrocytes. I. Generation of the activated complex.

A complex formed between cobra venom factor (CVF) and isolated human factor B (B) was found to be converted by trypsin to a stable enzyme, CVF-B which cleaved the third component (C3) and the fifth component (C5) of human complement. The formation of CVF-B by trypsin required divalent cations, whereas the formation of the lytic factor from human serum occurred even in the presence of EDTA. CVF-B purified by gel filtration could initiate the hemolysis of unsensitized guinea pig erythrocytes when incubated with human complement components C5 to C9 in 0.01 M EDTA buffer. C3 was not required for the lysis of guinea pig erythrocytes initiated by CVF-B because of the beta1C precipitation line formed between human serum and anti-beta1C antibody did not inhibit the hemolysis by CVF-B in agarose gel. Treatment of beta1C and beta1F globulins in whole human serum with CVF-B in the presence of 0.01 M EDTA converted them to components with higher mobilities on immunoelectrophoresis.     

Kinetics of the hydrolysis of mixed micelles of dipalmitoyllecithin with triton X-100 catalyzed by a phospholipase A2 from the venom of Agkistrodon halys blomhoffii

The pH dependence of kinetic parameters for the hydrolysis of mixed micelles of 1,2-dipalmitoyl-sn-glycero-3-phosphorylcholine (diC16PC) with Triton X-100, catalyzed by the intact and the N-terminal alpha-NH2-modified phospholipases A2 (PLA2s) of Agkistrodon halys blomhoffii, was studied at 25 degrees C and ionic strength 0.1 in the presence of saturating amounts of Ca2+. The pH dependence of the kinetic parameters for the hydrolysis of monodispersed diC6PC, catalyzed by the modified enzyme, was also studied under the same conditions, and the data were compared with the previous results for the intact enzyme [Teshima, K. et al. (1986) J. Biochem. 100, 1655-1662]. The pK values of the catalytic group, His 48, and Tyr 52 were found to shift from 5.55 to 7.00 and from 10.50 to 11.50, respectively, on binding of the micellar substrates to the enzyme. On the other hand, no participation of these ionizable groups was observed for the binding of the monodispersed substrate. On the basis of the present finding and the X-ray crystallographic studies on bovine pancreatic PLA2 [Dijkstra, B.W. et. al. (1981) J. Mol. Biol. 147, 97-123] and on a PLA2 of Crotalus atrox venom [Brunie, S. et al. (1985) J. Biol. Chem. 260, 9742-9749], the hydrogen-bonding of Tyr 73, which is involved in the lipid-water interface recognition site, to His 48 and Tyr 52 in the active center was strongly suggested to be important for the hydrolysis of micellar substrates.     

Increase in osmotic fragility of bovine erythrocytes induced by bacterial phospholipases C

Bovine erythrocytes were treated with each of three bacterial phospholipases C; phosphatidylcholine-hydrolyzing phospholipase C (PCase) of Clostridium perfringens, sphingomyelinase C (SMase) of Bacillus cereus and phosphatidylinositol-specific phospholipase C (PIase) of Bacillus thuringiensis. An increase in osmotic fragility was detected by means of a coil planet centrifugation (CPC) apparatus (Biomedical Systems Co., Tokyo) after the treatment with these enzymes. The peak of hemolysis normally observed in the untreated erythrocytes at the range between 50 and 100 mOsM shifted to 160 to 200 mOsM with the progress of sphingomyelin hydrolysis by phospholipase C of C. perfringens. Sphingomyelinase C of B. cereus showed two different effects on bovine erythrocytes: In the absence of divalent cations or in the presence of Ca2+ alone, the peak of hemolysis shifted to the region from 130 to 160 mOsM, without appreciable hydrolysis of sphingomyelin, while in the presence of Mg2+ or Mg2+ plus Ca2+, the peak of hemolysis further shifted to the region from 160 to 200 mOsM with the hydrolysis of sphingomyelin. Abrupt shift in osmotic fragility to a much higher region around 250 mOsM was produced by treatment with increasing amounts of phosphatidylinositol-specific phospholipase C. In this case, a significant amount of acetylcholinesterase was released from the erythrocyte membrane without hot or hot-cold hemolysis. The mechanism of alteration of osmotic fragility of bovine erythrocytes by treatment with phospholipases C seems to differ from case to case, depending upon the specific action of each enzyme toward the membrane phospholipids.     

Differences in the pattern of attack of acidic,neutral, and basic phospholipases A2 of A. halys blomhofii on human erythrocyte membranes: Problems in interpretation of phospholipid location

Purified acidic (pI 4.9), neutral (pI 6.9), and basic (pI 8.7) phospholipase A₂ from Agkistrodon halys blomhofii showed characteristically different patterns of hemolysis and phospholipid hydrolysis of intact human erthyrocytes. Acidic and neutral enzymes were nonlytic in the early periods of incubations with intact erythrocytes whereas the basic enzyme caused immediate hemolysis (5–8%). Under nonlytic conditions acidic and neutral enzymes hydrolyzed only phosphatidyl choline (PC) (20 and 50%, respectively), whereas basic enzyme hydrolyzed not only PC (60%) but nearly 15% of the phosphatidylethanolamine (PE). Both PC and PE were hydrolyzed significantly when the three phospholipases A₂ were incubated individually with erythrocyte lysate or hypotonic ghosts (sealed or unsealed). The order of substrate preference for acidic and neutral enzymes was always PC > PE. On the contrary basic enzyme exhibited the property of substrate specificity reversal. It hydrolyzed PC faster than PE when the membranes were sealed whereas PE hydrolysis was faster than PC hydrolysis in unsealed membranes. Interestingly only the basic enzyme showed activity in the absence of Ca²⁺ and in the presence of 0.5 mm EDTA. Phospholipase C (Bacillus cereus or Clostridium perfringens) did not show the property of substrate specificity reversal although their ability to hydrolyze PC and PE was different. In general this study demonstrates the unique activity patterns of three physically different pure phospholipases A₂ on human erythrocyte membranes which could be of value in selectively modifying membrane phospholipids. In addition it also throws an important light on the fact that results obtained with phospholipases should be interpreted with caution particularly as regards the localization of phospholipids in membranes.     

Organization of phospholipids in human red cell membranes as detected by the action of various purified phospholipases.

1. The action of eight purified phospholipases on intact human erythrocytes has been investigated. Four enzymes, e.g. phospholipases A2 from pancreas and Crotalus adamanteus, phospholipase C from Bacillus cereus, and phospholipase D from cabbage produce neither haemolysis nor hydrolysis of phospholipids in intact cells. On the other hand, both phospholipases A2 from bee venom and Naja naja cause a non-haemolytic breakdown of more than 50% of the lecithin, while sphingomyelinase C from Staphylococcus aureus is able to produce a non-lytic degradation of more than 80% of the sphingomyelin. 2. Phospholipase C from Clostridium welchii appeared to be the only lipolytic enzyme tested, which produces haemolysis of human erythrocytes. Evidence is presented that the unique properties of the enzyme itself, rather than possible contaminations in the purified preparation, are responsible for the observed haemolytic effect. 3. With non-sealed ghosts, all phospholipases produce essentially complete breakdown of those phospholipids which can be considered as proper substrates for the enzymes involved. 4. Due to its absolute requirement for Ca2+, pancreatic phospholipase A2 can be trapped inside resealed ghosts in the presence of EDTA, without producing phospholipid breakdown during the resealing procedure. Subsequent addition of Ca2+ stimulates phospholipase A2 activity at the inside of the resealed cell, eventually leading to lysis. Before lysis occurs, however, 25% of the lecithin, half of the phosphatidylethanolamine and some 65% of the phosphatidylserine can be hydrolysed. This observation is explained in relation to an asymmetric phospholipid distribution in red cell membranes.     

Hemolytic assay for venom phospholipase A2

A rapid and sensitive spectrophotometric assay for venom phospholipase A₂ based on the hemolysis of guinea pig erythrocytes in the presence of decomplemented serum and cardiotoxin (direct lytic factor) is described. This assay is particularly useful for rapid multisample analyses, such as those used in monitoring chromatography fractions, and is specific for phospholipase A₂ in she presence of other potentially hemolytic venom components. The hemolytic mechanism is shown to be a combination of the action of lysophospholipids liberated from lipoproteins in the serum and the synergistic action of phospholipase A₂ and cardiotoxin on the erythrocyte membrane.     

日本蝮蛇蛇毒碱性磷脂酶A2同源物的分离及鉴定

We purified and characterizated a phospholipase A2 homologue from Agkistrodon blomhoffii ussurensis snake venom. We used Hitrap SP cation exchange and Superdex 75 columns chromatography to obtain a basic protein, used SDS-PAGE to analyse molecular mass, and IEF (Isoelectric focusing electrophoresis) IEF to identify isoelectric point. The molecular mass was 16 kDa, and the isoelectric point was 8.56. We detected its phospholipase A2 activity on egg yolk phospholipids, hemolytic activity on washed erythrocytes, and anticoagulant effect on pig platelet-rich plasma, as well as the N-terminal sequence with protein sequencer. The results showed that it had no phospholipase A2 activity and hemolytic activity, but had obvious anticoagulant effect on in witro. The N terminal sequence (21 amino acid residues) compared with other phospholipases A2 demonstrated that the protein was homogenous with BPLA2s from Agkistrodon halys Palls.     

The role of Asp-49 and other conserved amino acids in phospholipases A2 and their importance for enzymatic activity

The role of aspartic acid-49 (Asp-49) in the active site of porcine pancreatic phospholipase A2 was studied by recombinant DNA techniques: two mutant proteins were constructed containing either glutamic acid (Glu) or lysine (Lys) at position 49. Enzymatic characterization indicated that the presence of Asp-49 is essential for effective hydrolysis of phospholipids. Conversion of Asp-49 to either Glu or Lys strongly reduces the binding of Ca2+ ions, in particular for the lysine mutant, but the affinity for substrate analogues is hardly affected. Extensive purification of naturally occurring Lys-49 phospholipase A2 from the venom of Agkistrodon piscivorus piscivorus yielded a protein that was nearly inactive. Inhibition studies showed that this residual activity was due to a small amount of contaminating enzyme and that the Lys-49 homologue itself has no enzymatic activity. Our results indicate that Asp-49 is essential for the catalytic action of phospholipase A2. The importance of Asp-49 was further evaluated by comparison of the primary sequences of 53 phospholipases A2 and phospholipase homologues showing that substitutions at position 49 are accompanied by structural variations of otherwise conserved residues. The occurrence of several nonconserved substitutions appeared to be a general characteristic of nonactive phospholipase A2 homologues.     

Adsorption of sphingomyelinase of Bacillus cereus onto erythrocyte membranes

Sphingomyelinase of Bacillus cereus proved to be specifically adsorbed onto mammalian erythrocyte membranes in the presence of either Ca2+ or Ca2+ plus Mg2+ in the order of sphingomyelin content; i.e., sheep, bovine greater than porcine greater than rat erythrocytes. No appreciable adsorption was observed in the presence of Mg2+ alone nor in the absence of divalent metal ions. The enzyme adsorption onto bovine erythrocytes was dependent upon the incubation temperature. By shifting the temperature from 37 to 0 degrees C, sphingomyelinase once adsorbed onto the surface of bovine erythrocytes was released into the supernatant. Ca2+ proved to be an essential factor for the enzyme adsorption: The addition of 1 mM Ca2+ enhanced the adsorptive process, but inhibited sphingomyelin hydrolysis and hot or hot-cold hemolysis of erythrocytes, while the addition of 1 mM Ca2+ plus 1 mM Mg2+ enhanced sphingomyelin breakdown and hemolysis as well as the enzyme adsorption. However, when the amount of sphingomyelin fell off to 0.2-0.7 nmol/ml or less by the action of sphingomyelinase, the enzyme once adsorbed was completely released from the surface of erythrocytes. The result indicates that the major binding site for sphingomyelinase is sphingomyelin. In the presence of 1 mM Mg2+ alone, the enzymatic hydrolysis of sphingomyelin and hemolysis proceeded whereas the enzyme adsorption was not encountered during 60 min incubation at 37 degrees C. The change in the molar ratio of Ca2+ to Mg2+ affected the enzyme adsorption and sphingomyelin breakdown; the higher Ca2+ enhanced the adsorption whereas the higher Mg2+ stimulated sphingomyelin hydrolysis.     

Membrane alterations in cellular aging: Susceptibility of phospholipids in density (age)-separated human erythrocytes to phospholipase A2

Human erythrocytes were separated into four density (age) groups representing the top 10% (young), bottom 10% (old), and two middle fractions of 40% each (intermediary ages). When these erythrocytes of different age groups were treated with the low levels of a purified basic phospholipase A₂ from Agkistrodon halys blomhofii, under conditions where little or no hemolysis occurred, the optimum extent of phosphatidylcholine (PC) hydrolysis in all age groups was the same, but interestingly, the rate of its hydrolysis was two to three times faster in the older cells compared to younger erythrocytes. On the other hand, hydrolysis of phosphatidylethanolamine (PE) of younger erythrocytes by the phospholipase A₂ was negligible under the particular experimental conditions. However, in erythrocytes of older age groups, both the rate and extent of PE hydrolysis by the enzyme increased in a distinctive fashion. Concomitant with the above pattern of PC and PE hydrolysis, the shape changes in the erythrocytes also were different; whereas all older erythrocytes became echinocytic only two-thirds of the younger erythrocytes showed a similar shape change. These observations firmly establish that during in vivo aging of normal erythrocytes in circulation significant changes in the structural organization of membrane phospholipids take place. Importance of this phenomenon in membrane phospholipid asymmetry studies and in the elimination of senescent cells also is discussed.     

Studies on an activator of the (Ca2+ plus Mg2+)-ATPase of human erythrocyte membranes.

1. An activator of the (Ca2+ plus Mg2+)-stimulated ATPase present in the human erythrocytes (membrane) has been isolated in soluble form from hemolysates of these cells. Partial purification has been achieved through use of carboxymethyl-Sephadex chromatography. The resulting activator fraction contained no hemoglobin and only 0.3% of the total adenylate kinase activity of the cell. 2. Whereas the activator was released from erythrocytes subjected to hemolysis in 20 miosM buffer at pH 7.6 or at pH 5.8, only the membranes prepared at pH 7.6 were affected by it. 2. Whereas the activator was released from erythrocytes subjected to hemolysis in 20 miosM buffer at pH 7.6 or at pH 5.8, only the membranes prepared at pH 7.6 were affected by it. 3. When (Ca2+ plus Mg2+)-ATPase activity was measured by 32Pi release from (gamma-32P)ATP, freeze-thawed erythrocytes, as well as membranes prepared at pH 5.8 and at pH 7.6, expressed lower values than noted by assay for total Pi release. When ADP instead of ATP was used as substrate, significant amount of Pi were released by these erythrocyte preparations. Further study revealed (a) production of ATP and AMP from ADP with membranes and hemolysate alone, and (b) exchange of the gamma-and B-position phosphate on (gama-32P)ATP in the presence of membranes plus hemolysates. These observations established the presence of adenylate kinase activity in the (membrane-free) hemolysates and in membranes. It further supports the conclusion that Pi release from ADP by human erythrocytes (freeze-thawed) and by their isolated membranes is due to formation of ATP by adenylate kinase and hydrolysis of this generated ATP by (Ca2+ plus Mg2+)-ATPase. 4. The following points were also established: (a) absence of an ADPase in human erythrocytes; (b) the (Ca2+ plus Mg2+)-ATPase activator enhanced cleavage only of the gama-position of ATP and (c) the (Ca2+ plus Mg2+)-ATPase activator is neither adenylate kinase nor hemoglobin.     

The synergism of cardiotoxin and phospholipase A2 in hemolysis

The synergistic effect of exogenous cobra phospholipase A₂ on the hemolysis rate of guinea pig erythrocytes by highly purified snake venom cardiotoxins was investigated. In the presence of phospholipase A₂ the reaction was not only faster and had a lower activation energy but followed a sigmoidal instead of a linear time course. Similar results were obtained using porcine pancreatic phospholipase A₂. Significantly, addition of even a trace of cobra phospholipase A₂ (approx. 0.1%, w/w) was sufficient to bring about the full synergistic effect, emphasizing the stringent purity requirements for any meaningful investigation of cardiotoxin's own action. The possibility that the action of cardiotoxin on its own may involve the stimulation of an endogenous phospholipase is discussed in the light of the results obtained with exogenous cobra enzyme.     

Purification and properties of lysophospholipase from the venom of the giant hornet Vespa orientalis

Using biospecific chromatography on polylysocephamide, a toxic phospholipase possessing a presynaptic effect on neuromuscular preparations was isolated from the venom of the giant hornet Vespa orientalis. The enzyme was shown to possess a high hydrolytic activity towards 1-acyllysophosphatidylcholine within a narrow pH range (pH optimum 7.5). The enzyme activity was suppressed by detergents of various chemical composition. Lysophospholipase caused an intensive hemolysis of washed human erythrocytes. The catalytic and hemolytic functions of the enzyme were sensitive to metal ions, however, in a different degree. Ca2+ and Mn2+ activated, while Cu2+ and Zn2+ inhibited the enzyme. Mg2+ and Sr2+ had no effect on the enzyme activity.     

Kinetics of erythrocyte lysis by snake venom cardiotoxins

Hemolysis of guinea pig erythrocytes by snake venom cardiotoxins was investigated with a semi-automatic method based on light-scattering changes of erythrocyte suspensions at 700 nm which are directly related to hemoglobin release. Small amounts of phospholipase-free cardiotoxin (<100 μg) could be conveniently and rapidly assayed with the high reproducibility in a recording spectrophotometer, and reliable kinetic data were accumulated.Cardiotoxins from two different genera (Hemachatus haemachates and Naja mossambica mossambica) displayed virtually identical hemolytic properties. Hemolysis increased linearly with time, in contrast with a sigmoidal pattern when phospholipase was present as an impurity. Low concentrations of Ca²⁺ (<1 mM) stimulated cardiotoxin action. A limiting plateau rate of hemolysis reached during concentration dependence experiments in which the level of either cardiotoxin or of erythrocytes was varied, suggested that the interaction of cardiotoxin with erythrocyte membranes is a saturation phenomenon only at a high ratio of cardiotoxin: erythrocytes. No hemolysis was observed with an homologous neurotoxin of S-methylated cardiotoxin, providing evidence for specificity. The linear Arrhenius plots obtained for the temperature dependence of cardiotoxin-induced hemolysis strengthened the conclusion that its action involves more than a detergent-like effect on membrane phospholipids.     

The role of aspartic acid-49 in the active site of phospholipase A2. A site-specific mutagenesis study of porcine pancreatic phospholipase A2 and the rationale of the enzymatic activity of [lysine49]phospholipase A2 from Agkistrodon piscivorus piscivorus' venom

In order to probe the role of Asp-49 in the active site of porcine pancreatic phospholipase A2 two mutant proteins were constructed containing either Glu or Lys at position 49. Their enzymatic activities and their affinities for substrate and for Ca2+ ions were examined in comparison with the native enzyme. Enzymatic characterization indicated that the presence of Asp-49 is essential for effective hydrolysis of phospholipids. Conversion of Asp-49 to either Glu or Lys strongly reduces the binding of Ca2+ ions in particular for the lysine mutant but the affinity for substrate analogues is hardly affected. Extensive purification of [Lys49]phospholipase A2 from the venom of Agkistrodon piscivorus piscivorus yielded a protein which was 4000 times less active than the basic [Asp49]phospholipase A2 from this venom. Inhibition studies with p-bromophenacyl bromide showed that this residual activity was due to a small amount of contaminating enzyme and that the Lys-49 homologue itself is inactive. The results obtained both with the porcine pancreatic phospholipase A2 mutants and with the native venom enzymes show that Asp-49 is essential for the catalytic action of phospholipase A2.     

Activation of the fifth and sixth component of the complement system: similarities between C5b6 and C(56)a with respect to lytic enhancement by cell-bound C3b or A2C, and species preferences of target cell

Brief shift of purified C5 and C6 at 0 degrees C to pH 6.4, followed by immediate neutralization, results in the generation of a factor, designated C(56)a, that lyses erythrocytes together with C7, C8, and C9. We compared C(56)a and C5b6 generated by an alternative-pathway convertase, with regard to their action on different target cells. We found tht C(56)a is similar to C5b6 in the following properties: 1) Together with C7, C(56)a forms a stable intermediate on either sheep or guinea pig erythrocytes. 2) Membrane-bound C3b, or A2C incorporated in the membrane, enhances lysis by C(56)a-9, as well as lysis by C5b6-9. We also found that the lysis of EC(56)a7 or EC5b67 intermediates by C8 and C9 depends on the species of the erythrocytes and the species of C8 and C9. Thus, lysis of sheep erythrocytes is more efficient with guinea pig C8 and C9 than with human C8 and C9. In the case of guinea pig erythrocytes, this relationship is reversed, i.e., these cells lyse more efficiently when human C8 and C9 are used. Enhancement of lysis by membrane-bound C3b or A2C does not abrogate this species incompatibility pattern.     

The conjugation of tyramine with sulphate by liver and intestine of different animals.

Tyramine was conjugated with sulphate by extracts of monkey intestine and livers of monkey, rat, mouse, guinea pig and man. The activity measured in monkey intestine was almost three times that of monkey liver. Labelled tyramine sulphate synthesized from [14C] tyramine, [3H] tyramine or Na235SO4, on acid hydrolysis, released its radioactive precursor. Liver extracts of monkey, rat, mouse and guinea pig synthesized respectively 145,66,21 and 6 pmol of [14C] tyramine sulphate/min per mg of protein. Except with the monkey, intestine exhibited very low activity. trans-2-Phenylcyclopropylamine, a monoamine oxidase inhibitor, was added as a routine to the enzyme preparation, as its omission resulted in the production of p-hydroxyphenylacetic acid in appreciable amounts. This oxidative deamination of tyramine, however, did not decrease the sulpho-conjugation of tyramine. The low Km (9.1 muM) of sulphotransferase for tyramine is probably responsible.     

Exploring the action and specificity of cobra venom phospholipase A2 toward human erythrocytes, ghost membranes, and lipid mixtures

We have investigated the action and substrate specificity of phospholipase A2 (EC 3.1.1.4) purified from cobra venom (Naja naja naja) toward intact and Triton-solubilized human erythrocytes, toward ghost membranes, and toward extracted ghost lipids in mixed micelles with Triton X-100. We have found that: (i) phospholipids in the outer surface of intact erythrocytes are extremely poor substrates for the phospholipase, (ii) phospholipids in ghost erythrocyte membranes and in Triton-solubilized erythrocytes are suitable substrates for the enzyme, (iii) in these latter systems which contain a mixture of lipids, phosphatidylethanolamine is preferentially hydrolyzed, whereas in model studies on individual phospholipid species in mixed micelles with Triton, phosphatidylcholine is the preferred substrate of the enzyme, and (iv) the preferential hydrolysis of phosphatidylethanolamine is also observed for extracted ghost lipid mixtures in mixed micelles. These results demonstrate a dependence of phospholipase A2 activity on the ghosting procedure and a dependence of substrate specificity on the presence of other lipids. The relevance of these findings to the interpretation of membrane lipid asymmetry studies utilizing phospholipases is considered in detail.