Mechanisms by which intracellular calcium induces susceptibility to secretory phospholipase A2 in human erythrocytes.

Exposure of human erythrocytes to the calcium ionophore ionomycin rendered them susceptible to the action of secretory phospholipase A(2) (sPLA(2)). Analysis of erythrocyte phospholipid metabolism by thin-layer chromatography revealed significant hydrolysis of both phosphatidylcholine and phosphatidylethanolamine during incubation with ionomycin and sPLA(2). Several possible mechanisms for the effect of ionomycin were considered. Involvement of intracellular phospholipases A(2) was excluded since inhibitors of these enzymes had no effect. Assessment of membrane oxidation by cis-parinaric acid fluorescence and comparison to the oxidants diamide and phenylhydrazine revealed that oxidation does not participate in the effect of ionomycin. Incubation with ionomycin caused classical physical changes to the erythrocyte membrane such as morphological alterations (spherocytosis), translocation of aminophospholipids to the outer leaflet of the membrane, and release of microvesicles. Experiments with phenylhydrazine, KCl, quinine, merocyanine 540, the calpain inhibitor E-64d, and the scramblase inhibitor R5421 revealed that neither phospholipid translocation nor vesicle release was required to induce susceptibility. Results from fluorescence spectroscopy and two-photon excitation scanning microscopy using the membrane probe laurdan argued that susceptibility to sPLA(2) is a consequence of increased order of membrane lipids.

Exposure of human erythrocytes to the calcium ionophore ionomycin rendered them susceptible to the action of secretory phospholipase A 2 (sPLA 2 ). Analysis of erythrocyte phospholipid metabolism by thin-layer chromatography revealed significant hydrolysis of both phosphatidylcholine and phosphatidylethanolamine during incubation with ionomycin and sPLA 2 . Several possible mechanisms for the effect of ionomycin were considered. Involvement of intracellular phospholipases A 2 was excluded since inhibitors of these enzymes had no effect. Assessment of membrane oxidation by cis-parinaric acid fluorescence and comparison to the oxidants diamide and phenylhydrazine revealed that oxidation does not participate in the effect of ionomycin. Incubation with ionomycin caused classical physical changes to the erythrocyte membrane such as morphological alterations (spherocytosis), translocation of aminophospholipids to the outer leaflet of the membrane, and release of microvesicles. Experiments with phenylhydrazine, KCl, quinine, merocyanine 540, the calpain inhibitor E-64d, and the scramblase inhibitor R5421 revealed that neither phospholipid translocation nor vesicle release was required to induce susceptibility. Results from fluorescence spectroscopy and two-photon excitation scanning microscopy using the membrane probe laurdan argued that susceptibility to sPLA 2 is a consequence of increased order of membrane lipids.
Under normal conditions, healthy cell membranes resist catalysis by secretory phospholipase A 2 (sPLA 2 ) 1 (1)(2)(3)(4). However, they may become susceptible under circumstances that cause alteration of membrane physical properties (1)(2)(3)(4). Previous studies using artificial membranes demonstrated that alterations that increase susceptibility generally increase the ani-onic charge of the outer leaflet, increase bilayer curvature, and/or decrease interactions among neighboring phospholipids (5)(6)(7)(8)(9). In some cases, enhanced susceptibility of artificial membranes depends on an increase in the order of the phospholipids (8, 10 -14). These changes increase susceptibility by augmenting the binding of sPLA 2 and/or by improving access of membrane phospholipids to the active site of the enzyme (5-12, 15, 16).
It is not known whether the properties that induce susceptibility to sPLA 2 in artificial membranes also contribute to the vulnerability of biological membranes to attack by the enzyme. In order to address this issue, we manipulated various properties of erythrocyte membranes by preparing different types of ghosts as explained in the accompanying particle (17). We found that the factors that determined the degree of susceptibility were increased exposure of phosphatidylserine, an anionic phospholipid, and increased membrane order. These interpretations agreed with those from studies of susceptibility using artificial membranes (5)(6)(7)(8)(9)(10)(11)(12)(13)(14)(15)(16). The next question, then, is whether these same factors are important in the hydrolysis of intact cells by sPLA 2 under conditions at which they have become susceptible such as in the presence of specific hormones, after treatment with certain toxins, during apoptosis, or following cellular trauma (2)(3)(4)18).
One feature common among many of the conditions that render cells susceptible to sPLA 2 is the elevation of intracellular calcium (1)(2)(3)(4). We have used human erythrocytes as an experimental model to determine whether phosphatidylserine exposure and/or an increase in the order of membrane phospholipids are relevant factors in the induction of catalysis by sPLA 2 when intracellular calcium is increased. In addition, we examined other hypotheses that have been proposed to explain the ability of certain agents to render cell membranes susceptible to sPLA 2 : 1) prior activation of intracellular phospholipase(s) A 2 (19,20), 2) release of microvesicles from the plasma membrane (1), 3) oxidation of membrane phospholipids (21).

EXPERIMENTAL PROCEDURES
Materials-Erythrocytes were obtained from healthy individuals undergoing routine physicals at Brigham Young University McDonald Health Center. The samples were stored overnight at 4°C in EDTA vacutainers from which patient identification was removed. Control experiments comparing fresh blood with samples stored overnight demonstrated that the storage conditions did not influence the results. Erythrocytes were isolated by centrifugation, and resuspended to the original hematocrit in MBSS (NaCl ϭ 134 mM, KCl ϭ 6.2 mM, CaCl 2 ϭ 1.6 mM, MgCl 2 ϭ 1.2 mM, Hepes ϭ 18.0 mM, and glucose ϭ 13.6 mM, pH 7.4, 37°C).
Snake venom sPLA 2 (monomeric aspartate 49 (AppD49) from the venom of Agkistrodon piscivorus piscivorus) was isolated according to published procedures and was used in all experiments except those shown in Fig. 13 (22). Human group V and IIa sPLA 2 were provided generously by Wonhwa Cho (University of Illinois, Chicago, IL) and Michael Gelb (University of Washington, Seattle, WA). The final concentrations of sPLA 2 used in experiments were 1 g/ml for AppD49 and human group V and 2 g/ml for human group IIa.
ADIFAB, laurdan, merocyanine 540, and cis-parinaric acid were obtained from Molecular Probes (Eugene, OR). Ionomycin and E-64d were procured from Calbiochem (La Jolla, CA), and phenylhydrazine, diamide, and quinine were obtained from Sigma. The scramblase inhibitor, R5421, was a kind gift from Jeffrey T. Billheimer at Dupont Merck Pharmaceutical Co. (Wilmington, DE). DAPA, factor Va, factor Xa, prothrombin, and thrombin were acquired from Hematologic Technologies, Inc. (Essex Junction, VT). Pharmacological agents were dissolved in the appropriate solvents (Me 2 SO or ethanol). Control experiments demonstrated that these solvents did not have effects on the experimental data at the concentrations used.
Phospholipid Extraction and Thin Layer Chromatography-Washed erythrocytes (30 l) were suspended in MBSS to a final volume of 1 ml (about 1.5 ϫ 10 8 cells/ml) and incubated in the presence or absence of 0.3 M ionomycin with or without AppD49 sPLA 2 for 20 min at 37°C. Cells were then separated by centrifugation (6500 rpm for 60 s) in a microcentrifuge (about 3000 ϫ g) and pellets were frozen in liquid nitrogen to quench the reaction. Samples were quickly thawed, and lipids were extracted with chloroform and methanol by the method of Bligh and Dyer (23). In brief, 100 l of chilled MBSS was added to suspend the pellet followed by 125 l of chloroform and 250 l of methanol. After vortexing the tubes for 10 s, 125 l of water were added. The samples were then vortexed and centrifuged for 30 s at 3000 ϫ g. After removing half of the water layer, the protein layer was carefully removed with a pipette tip. The remainder of the water layer was then discarded and the residual organic layer was dried under a nitrogen stream to ϳ10% of the original volume. The sample was then spotted onto a silica gel thin-layer chromatography plate. Phospholipids and lysophospholipids were separated by thin-layer chromatography in 6.5:2.5:1 (v/v) chloroform:methanol:acetic acid. Lipids were stained by iodine vapor. Spots were identified by comparison to standards. The resulting phosphatidylcholine and phosphatidylethanolamine spots on the silica gel were analyzed by both phosphate assay according to the method of Bartlett (24) and by densitometry. For densitometric measurements, samples were photographed with a digital camera using a Coomassie Blue filter under direct light and the digital image was quantified using standard digitizing computer software.
Fluorescence Spectroscopy-Washed erythrocytes were suspended in 2 ml of MBSS in a fluorometer sample cell to a final density of about 3-4 ϫ 10 6 cells/ml. Measurements with fluorescent probes were obtained at 37°C using a Fluoromax (Spex Industries) photon-counting spectrofluorometer. Sample homogeneity was maintained by continuous gentle stirring with a magnetic stir bar. Simultaneous assessment of fluorescence intensity at multiple excitation and emission wavelengths was obtained by rapid sluing of monochromator mirrors using control software provided with the instrument. Band pass was set at 4.25 nm for all experiments.
Hydrolysis by sPLA 2 -Release of fatty acids from cells was assayed with an acrylodan-labeled fatty acid-binding protein (ADIFAB) (65 nM final, excitation ϭ 390 nm, emission ϭ 432 and 505 nm; Refs. 3 and 25). The results were quantified by calculation of the generalized polarization (GP) as described (3,26). The values of GP as a function of time were fit to a double exponential equation by nonlinear regression. The amount of hydrolysis at 100 s following sPLA 2 addition was then calculated using parameter values from the nonlinear regression results.
Prothrombinase Assay-Exposure of phosphatidylserine in the outer leaflet of the bilayer was detected by an increase in the fluorescence intensity of dansylarginine-N- Microvesicle Release-The release of vesicles from the plasma membrane was monitored simultaneously with other fluorescence observations by recording the intensity of scattered light (excitation ϭ 500 nm, emission ϭ 510 nm; Ref. 4). For simultaneous measurements with the prothrombinase assay (see above), excitation and emission wavelengths were 600 and 610 nm, respectively.
Oxidation-Oxidation of membrane phospholipids was monitored by use of the fluorescent probe, cis-parinaric acid (1.12 M final, excitation ϭ 303 nm, emission ϭ 416 nm; Ref. 28). Measurements of light scattering for microvesicle release and ADIFAB fluorescence were made simultaneously. The data were corrected for time-dependent light scattering artifacts caused by microvesicle release.
Scanning Electron Microscopy-Erythrocytes were prepared by a modification of Schneider's method (29). Briefly, the preparations were washed in 0.1 M phosphate buffer at pH 7.4. 9.5 ml of 1ϫ10 8 cells/ml were incubated in a jar having a 5.5-cm diameter, in the presence or absence of 0.3 M ionomycin, and allowed to settle onto cover glasses, previously coated with poly-L-lysine, at 4°C overnight. Samples were then fixed in 2% glutaraldehyde for 2.5 h. Following fixation, the cells were washed six times in sodium cacodylate buffer (pH 7.3), fixed in 2% osmium tetroxide for 2 h at 23°C, and washed six times in sodium cacodylate buffer. Samples were dehydrated through a graded series of ethanol solutions (10,30,50,70,95, and 100%) for 10 min each then washed three times in acetone. The slides were then subjected to critical point drying, using carbon dioxide. Finally, samples were sputter coated with gold for 2 min. Images were obtained on a JEOL JSM 840A scanning electron microscope.
Membrane Fluidity-Membrane order was assessed using laurdan GP (26). Laurdan (2.5 M final) was added to samples of erythrocytes prepared as described above for fluorescence spectroscopy. Fluorescence emission was then monitored as a function of time at dual wavelengths (excitation ϭ 350 nm, emission ϭ 435 and 500 nm) for at least 5 min to establish the baseline. Various agents followed by ionomycin or control solvent were added as described in Fig. 9. Changes in laurdan GP were then assessed by calculating the difference in the slope of laurdan GP before and after addition of ionomycin or control solvent under each experimental condition.
Two-photon Microscopy-The two-photon excitation images were collected on an Axiovert 35 inverted microscope (Zeiss, Thornwood, NY), with a Zeiss 20X LD-Achroplan (0.4 N.A., air) using a titanium-sapphire laser excitation source (Coherent, Palo Alto, CA) tuned to 770 nm and pumped by a frequency-doubled Nd:vanadate laser (Coherent, Palo Alto, CA) as described previously (30). The laser was guided by a galvanometer-driven x-y scanner (Cambridge Technology, Watertown, MA) to achieve beam scanning in both x and y directions. A frequency synthesizer (Hewlett-Packard, Santa Clara, CA) controlled the scanning rate of 9 s to acquire a 256 ϫ 256-pixel frame that covered approximately a 60 ϫ 60-m region. Dual images were collected simultaneously using a beam-splitter, two emission short-pass filters (centered at about 450 and 500 nm), and two detectors for calculation of GP (26).
Samples were incubated with or without the agents indicated in Fig.  11 as described above for hydrolysis experiments. Laurdan (250 nM) was added to the samples 250 s after the addition of the last agent in the experiment. Samples were incubated with stirring for an additional 50 s, and a 0.5-ml aliquot was then transferred to 1 ml of fresh MBSS in a heated microscopy sample dish (36°C). Cells were allowed an additional 5 min to settle, and images were then obtained.
In some cases (e.g. Fig. 10), cells were incubated prior to the onset of the experiment with 5 M laurdan for 1 h at 36°C, and excess laurdan was removed by centrifugation. Cells were suspended in 2 ml of fresh MBSS and transferred to the microscopy dish. After allowing the cells to settle, baseline images were obtained. Ionomycin was then added directly to the sample, and additional images were acquired. Finally, sPLA 2 was added, and the time course of changes in laurdan fluorescence was monitored by repeated acquisition of images of the same field.
Statistical Analysis-In all figures that contain summaries from multiple replicates, the data are expressed as the mean Ϯ S.E. Each replicate sample included in the data represents data from a separate blood donor. Large comparisons of hydrolysis or light scattering data among many groups sharing some of the same blood samples (e.g. Figs. 2, 7, 9, and 12) were accomplished by one-way analysis of variance followed by Dunnett's post-test for multiple comparisons. Since the number of samples per group was unbalanced, it was not possible to consider sample pairing in the analyses of variance. This increased the possibility of missing real differences that would only be identified by paired comparisons of samples matched by blood donor. Accordingly, the various treatment groups were also compared with the group treated with ionomycin alone using Student's paired t test (two-tailed) for those samples that were matched by blood donor. Although the results of this secondary analysis agreed with those of the analysis of variance and post-test in most instances, there was one example in which the results were significant only when the analysis was confined to paired samples. In this case, the level of significance was very high (p ϭ 0.004), and the data were therefore interpreted as being statistically significant (see legend to Fig. 7).
When data sets were fully matched by blood sample for all treatment groups (i.e. Figs. 3, 4, and 13), they were analyzed in two steps. First, results within each treatment group were normalized to the value of an appropriate internal standard matched by blood sample (see figure legends for details). Second, the normalized values were tested for treatment effects using Student's t test (two-tailed) with the value of 1.0 as the null hypothesis. Since multiple (two to three) treatment groups were compared with the same standard in these cases, a correction was made to the critical value of p accepted as indicating statistical significance (traditionally 0.05) according to the formula p ϭ 1-0.95 1/n , where n ϭ the number of comparisons. For n ϭ 2, the critical value of p ϭ 0.025; for n ϭ 3, the critical value of p ϭ 0.017. Fig. 1, the extent of fatty acid release from erythrocyte membranes in the presence of sPLA 2 was greatly enhanced by a 10-min prior incubation of the cells with ionomycin. The average response among multiple samples is displayed in Fig. 2. Control experiments in which Ca 2ϩ was replaced by EGTA in the extracellular medium revealed that this and the other results described below for ionomycin were due to Ca 2ϩ entry into the cell rather than direct effects of the ionophore. Experiments in which the time of incubation with ionomycin was varied revealed that the effect developed after a latency of about 100 s and reached a maximum within about 300 s (not shown). In contrast to results obtained with lymphocytes (4), the hydrolysis was not sufficient to consume the cells, and little or no hemolysis was observed at the end of the reaction.

Effect of Ionomycin-As shown in
Comparative Hydrolysis of Phosphatidylcholine and Phosphatidylethanolamine-In order to assess the relative amount of hydrolysis of the two major glycerophospholipid species, experiments were conducted at much higher cell densities (about 40-fold higher) than in the experiment of Fig. 1. Under such conditions, the time course of onset of the effect of ionomycin was much slower such that the lag time for achieving maximum hydrolysis rates was about 20 min. The release of microvesicles in the presence of ionomycin described below was also proportionally slower. Thin-layer chromatography experiments investigating hydrolysis of erythrocytes in the presence of sPLA 2 or ionomycin alone revealed no significant hydrolysis of either phosphatidylcholine or phosphatidylethanolamine after a 20-min incubation. In contrast, both substrates were hydrolyzed significantly when erythrocytes were incubated with sPLA 2 and ionomycin together (Fig. 3).
Intracellular Phospholipase A 2 -Erythrocytes were incubated with ionomycin in the presence or absence of MAFP (19) or BEL (31), inhibitors of intracellular phospholipases A 2 . As shown in Fig. 4, the amount of hydrolysis after 100 s with sPLA 2 was not statistically different in cells treated with MAFP and ionomycin compared with cells treated with only ionomycin. Similar results were obtained with BEL as the inhibitor (data not shown). When MAFP was incubated with erythrocytes at low cell density in the spectrofluorometer (i.e. as in Figs. 1 and 2), it caused nonspecific perturbation of the cell membrane that directly rendered the cells susceptible to sPLA 2 . Consequently, these experiments were conducted at higher cell densities and longer incubation times similar to the experiments described above for thin-layer chromatography (i.e. Fig. 3). The remainder of the experiments described below were completed at low cell density in the spectrofluorometer as in Figs. 1 and 2. The less specific inhibitor, AACOCF 3 , that had Asterisks represent values that differed significantly from ionomycin alone (**: p Ͻ 0.01, by analysis of variance as explained under "Experimental Procedures"). High KCl, quinine, or R5421 alone had no effect (not shown). previously been shown to inhibit an intracellular phospholipase A 2 activity in erythrocytes was also tested (32). Like MAFP and BEL, it did not reduce the susceptibility induced by ionomycin (data not shown). Therefore, activation of intracellular phospholipases A 2 appeared unnecessary for Ca 2ϩ -induced susceptibility to sPLA 2 in erythrocytes.
Phospholipid Translocation-Ionomycin stimulates transbilayer migration of membrane phospholipids (33). Fig. 5 demonstrates the time course of phosphatidylserine exposure (increase in DAPA fluorescence intensity) upon the addition of ionomycin. Prior treatment of cells for 10 min with R5421, an inhibitor of scramblase activity (34), caused a substantial decrease in the exposure of phosphatidylserine stimulated by ionomycin. Inhibition of phospholipid translocation by R5421 did not alter the susceptibility to sPLA 2 in the presence of ionomycin (Fig. 2).
Microvesicle Release-Upon introduction of ionomycin, erythrocytes released microvesicles after a short lag period as expected (1). The release of these vesicles was conveniently monitored in real time concurrently with assessment of susceptibility by measuring the amount of light scattered by the sample at 500 nm. As shown in Fig. 6, the intensity of scattered light increased about 100 s after addition of ionomycin and rose until reaching a plateau about 500 s later. We verified this interpretation by direct observation of the samples by scanning electron microscopy. As shown in Fig. 6C, treatment with ionomycin caused the erythrocytes to assume a diminished size and rounded shape (spherocyte) and to extrude small pieces of its membrane as microvesicles.
To test whether the microvesicle release was required for susceptibility to occur, we prevented the release by adding either high KCl (89 mM) or a Ca ϩ2 -activated K ϩ channel blocker (quinine) to the extracellular medium (35,36). Cells incubated in the high KCl buffer demonstrated a decrease, while cells treated with quinine showed a complete inhibition, of the amount of microvesicles present after treatment with ionomycin (Fig. 7). In the case of quinine, a decrease in light scattering intensity was observed presumably because of the reduction in cell size due to the shape transition (37). There was no significant effect of either treatment on the ability of ionomycin to induce susceptibility to sPLA 2 (Fig. 2).
Microvesicle release was also inhibited by use of an erythrocyte (type ) calpain inhibitor, E-64d (38). One of the effects of elevated intracellular Ca 2ϩ in erythrocytes and platelets is activation of calpain, an intracellular cytoskeletal protease (38). The resulting cytoskeletal damage appears to be involved in the process of microvesicle release (39). As expected, E-64d reduced significantly the level of microvesicle release observed in the presence of ionomycin (Fig. 7). However, like the other inhibitors of microvesicle release, it did not reduce ionomycinstimulated susceptibility to sPLA 2 (Fig. 2). Higher concentrations (up to 140 M), while able to block microvesicle release completely, also did not lower the level of hydrolysis by sPLA 2 (not shown). Therefore, it appeared that microvesicle release was not necessary for the cells to become susceptible to sPLA 2 .
Oxidation-In contrast to the positive controls, diamide and phenylhydrazine, ionomycin did not cause oxidation (i.e. reduction of cis-parinaric acid intensity) of erythrocyte membranes

FIG. 3. Hydrolysis of phosphatidylethanolamine (PE) and phosphatidylcholine (PC) in the presence of ionomycin and/or sPLA 2 .
Erythrocytes were incubated with or without 0.3 M ionomycin for 5 min; then sPLA 2 was added to half of the samples and the incubation continued for 20 min. Aliquots were removed, and phospholipids were extracted and separated by thin-layer chromatography as explained under "Experimental Procedures." The data were normalized to the amount of phospholipid observed under control conditions (no ionomycin or sPLA 2 ). Significance was assessed by Student's t test with adjustment for multiple comparisons as explained under "Experimental Procedures" (*, p ϭ 0.013 for "PC both," and p ϭ 0.0016 for "PE both," n ϭ 3).

FIG. 4. Inhibition of intracellular phospholipase A 2 does not alter the ability of ionomycin to induce susceptibility.
Erythrocytes were incubated for 10 min in the presence or absence of 20 M MAFP under the conditions described for the thin-layer chromatography experiments (i.e. Fig. 3). Ionomycin (0.3 M) was then added and the incubation continued for 45 min. A 50-l aliquot of each sample was then transferred to a spectrofluorometer sample cell (2 ml final volume), and susceptibility was then assessed using ADIFAB as described under "Experimental Procedures." Significance was assessed by normalizing the data for each group to that obtained with ionomycin alone and evaluated by Student's t test as described under "Experimental Procedures" (***, p Ͻ 0.0001). (Fig. 8). Erythrocytes treated with diamide alone for 10 min did not become susceptible to sPLA 2 . Likewise, diamide did not alter microvesicle release (not shown) nor the amount of hydrolysis observed when ionomycin was present (Fig. 2). A second oxidizing agent, phenylhydrazine, also did not cause the cells to become susceptible during a 10-min incubation (Fig. 2). Interestingly, in contrast to diamide, phenylhydrazine significantly impaired the effect of ionomycin on susceptibility (Fig.  2). In addition, phenylhydrazine caused an increase in the intensity of scattered light reminiscent of the effect of ionomycin to induce microvesicle release (Fig. 7).
Membrane Order-Fluidity of the membrane was assessed by fluorescence spectroscopy. Cells were labeled with laurdan, and the effects of various agents on GP values were determined. In general, an increase in the value of GP corresponds to an increase in membrane order (26). As shown in Fig. 9, ionomycin treatment caused a reproducible elevation of the value of GP. This effect was blocked by EGTA demonstrating that it required Ca 2ϩ and was not simply a direct effect of intercalation of ionophore into the membrane. Incubation of the cells in high KCl, E-64d, or R5421 had no significant effect on the response to ionomycin. Since phenylhydrazine treatment inhibited the ability of ionomycin to induce susceptibility, we also considered its effect on membrane order. In contrast to the other agents tested, phenylhydrazine did cause a significant decrease in the GP value. Control observations revealed that the effect of phenylhydrazine on laurdan GP was caused, at least in part, by time-dependent changes in the optical density of phenylhydrazine. We therefore repeated some of the experiments of Fig. 9 using two-photon microscopy to detect laurdan GP under conditions at which laurdan fluorescence arising directly from the membrane could be distinguished from indirect optical effects of the experimental agents. As shown in Fig. 10A, untreated cells displayed a non-uniform distribution of laurdan GP values.
Higher values were concentrated along the rims of the diskocytes. The addition of ionomycin increased the GP value of these peripheral regions and expanded their size (Figs. 10B and 11A). In agreement with the measurements displayed in Fig. 9, phenylhydrazine blocked the response to ionomycin (Fig. 11B).
Two-photon images of erythrocyte ghosts also revealed a non-uniform distribution of laurdan GP (17). Hydrolysis by sPLA 2 appeared related to the presence of regions of high GP since these regions expanded and became more ordered following sPLA 2 addition (17). As shown in Fig. 10, the same phenomenon was observed with intact erythrocytes treated with ionomycin. After addition of sPLA 2 , regions of low fluidity (yellow to red color) expanded systematically and became more ordered (Fig. 10, C and D).
The possibility that the susceptibility of erythrocytes to sPLA 2 was dependent on membrane order was further investigated using merocyanine 540. Merocyanine 540 binds to the outer leaflet of erythrocyte membranes and induces the shape transition from diskocytes to spherocytes without release of microvesicles or flip-flop of membrane lipids (40,41). Addition of merocyanine 540 to erythrocytes also caused a significant increase in laurdan GP (0.17 Ϯ 0.009 GP units, mean Ϯ S.E., n ϭ 3, p Ͻ 0.003 by Student's one-sample t test) comparable to that produced by ionomycin (e.g. Fig. 11B). Likewise, the agent rendered the membranes susceptible to sPLA 2 (Fig. 12, control and ionomycin data from Fig. 2 included for comparison).
Human sPLA 2 - Fig. 13 displays repetition of key experiments using human group V sPLA 2 instead of the snake venom enzyme. The extent of hydrolysis was about half of that observed with the AppD49 enzyme as reported previously (4). Nevertheless, the fundamental trends observed with ionomycin and phenylhydrazine were similar for the human sPLA 2 compared with the venom enzyme (compare Figs. 2 and 13). Experiments were also repeated with human group IIa sPLA 2 . In this case, however, the activity was very low, and quantitative interpretation of the data was not feasible.  Fig. 6 were repeated in the presence or absence of the agents listed. The ordinate indicates the change in average intensity of scattered light from that measured immediately prior to ionomycin (or phenylhydrazine or control diluent) addition to that observed 10 min later. Asterisks represent values that differed significantly from ionomycin alone (*, p Ͻ 0.05; **, p Ͻ 0.01, by analysis of variance; ***, p ϭ 0.004 by paired t test; the details of both analyses are explained under "Experimental Procedures"). High KCl and quinine alone were indistinguishable from control. The numbers of replicates were: Control, 37; Ionomycin, 52; High KCl ϩ ionomycin, 12; Quinine ϩ ionomycin, 6; E-64d ϩ ionomycin, 3; Phenylhydrazine, 4; Phenylhydrazine ϩ ionomycin, 16. See Fig. 2 for the details of the concentrations and incubation times. Concentrations and incubation times were the same as described in the legend to Fig. 2 with the exception that cells were incubated 5 min with phenylhydrazine prior to ionomycin addition. In the EGTA experiment, 1 mM EGTA replaced the 1.6 mM Ca 2ϩ in the MBSS.

DISCUSSION
The ability of intracellular Ca 2ϩ to govern the susceptibility of cell membranes to sPLA 2 has been observed in a number of cell types. For example, S49 lymphoma cells normally resist the action of sPLA 2 until treated with agents that elevate intracellular Ca 2ϩ levels such as ionophore, lysolecithin, or the plant toxin thionin (3,4). A similar phenomenon has been observed in other cells such as HL-60, MOLT-4, Raji, erythrocytes, and platelets (1,4,42).
While it is clear that this phenomenon is general, at least among blood cells, the mechanisms involved are much less established. Nevertheless, a few hypotheses have been proposed based on a variety of observations: 1) prior activation of intracellular phospholipase(s) A 2 (19,20,43,44); 2) release of microvesicles from the plasma membrane (1); 3) oxidation of membrane phospholipids (21); 4) transbilayer migration of phosphatidylserine and phosphatidylethanolamine (2,9,16,45); and 5) changes to other microscopic physical properties of the membrane (17). Based on the data shown in Figs. 2 and 6 -8, the first three hypotheses were excluded as explanations for the susceptibility to sPLA 2 observed in the presence of ionomycin. As discussed below, the results of this study combined with those of the accompanying article (17) contend that alterations to specific physical properties related to membrane fluidity are responsible for susceptibility to the enzyme. Importantly, these results validate the assumption that information obtained from studies of artificial bilayers relates to biological membranes.
The results reported with erythrocyte ghosts in the accompanying paper (17) suggest that exposure of phosphatidylserine can promote susceptibility, although multiple regression analysis revealed that it was a less important contributor than membrane properties assessed by laurdan. What the experiments with ghosts were unable to determine was whether phosphatidylserine exposure was required or instead simply ancillary or even redundant for making the membrane susceptible to the enzyme. The logic of the two studies was to identify first in the ghosts possible candidates for the relevant membrane changes and then ask whether those changes applied to Ca 2ϩ ionophore treatment of intact erythrocytes. In the case of phosphatidylserine exposure, the appropriate conclusion is that such may promote susceptibility, but it is not an absolute requirement during ionomycin treatment of erythrocytes. This assertion is based on two results. First, R5421 treatment inhibited the exposure of phosphatidylserine substantially (Fig.  5) but did not affect the amount of hydrolysis catalyzed by sPLA 2 in the presence of ionomycin (Fig. 2). Second, merocyanine 540, which does not induce translocation of phosphatidylserine (41), was able to cause susceptibility. These observations corroborate results obtained with S49 cells in which it was shown that susceptibility to sPLA 2 during apoptosis precedes significant exposure of phosphatidylserine (46).
Comparison of the laurdan results shown in Figs. 9 -11 with the susceptibility data (Fig. 2) suggests that changes in membrane order could be largely responsible for the induction of susceptibility by ionomycin. The agreement with the results from erythrocyte ghosts described in the accompanying article (17) is strong. First, membrane order was found to be the major predictor of susceptibility in ghosts when the various factors were considered together in multiple regression analysis. Second, comparison of the levels of susceptibility and change in GP induced by ionomycin (Figs. 2 and 11) with those reported for the ghosts demonstrates that the similarity is quantitative as well as qualitative. In addition, the data obtained with merocyanine 540 suggest that the relationship between membrane order and susceptibility is a general phenomenon rather than depending on influx of calcium (Fig. 12). Importantly, these results support the concept that principles learned from biophysical studies with artificial membranes apply to biological systems (5)(6)(7)(8)(9)(10)(11)(12)(13)(14)(15)(16). Attempts have been made previously with cultured cells to determine whether changes in membrane order detectable with fluorescent probes might explain the action of Ca 2ϩ to render them vulnerable to sPLA 2 (3,4). The results from those studies were inconclusive probably because of the diversity of membranes accessible to the probes. These studies with erythrocytes have the experimental and interpretive advantage of avoiding complications due to intracellular membranes.
It is likely that regional increases in the order of membrane lipids increases susceptibility both by enhancing the binding of the enzyme as well as creating membrane defects that facilitate migration of phospholipids into the active site of sPLA 2 as discussed (17). As with the ghosts, the microscopy images supported the idea that hydrolysis was focused at such regions of reduced fluidity (Fig. 10). How an elevation in intracellular Ca 2ϩ concentration would cause this increase in membrane order is not clear. One likely possibility is that Ca 2ϩ ions entering the cell bind to phospholipids, especially phosphatidylserine and phosphatidylinositols, on the inner leaflet of the membrane. This binding would cause the lipids to become more ordered on the inner leaflet. Increased order on the inner face would then be likely to enhance the ordering of lipids on the outer face leaflet since the physical properties of phospholipids across membranes are coupled. Typical biochemical effects of Ca 2ϩ such as involvement of calmodulin and kinases appear not to be involved based on results with S49 lymphoma cells (4).
The ability of phenylhydrazine to impede the effects of ionomycin on susceptibility and membrane order was unexpected. The mechanism of this inhibition is not yet clear. Phenylhydrazine has been reported to cause a variety of effects on erythrocytes such as proteolysis, hemolysis, formation of Heinz bodies, and alterations to phospholipid distribution and dynamics (47)(48)(49)(50)(51)(52)(53). However, these effects of phenylhydrazine are unlikely to be relevant to ionomycin-induced susceptibility and membrane order shown in Figs. 2, 9, and 11 since they were observed only after prolonged incubation with the agent for a period exceeding 1 h. In contrast, our results occurred immediately. We considered possible direct effects of phenylhydrazine on sPLA 2 by monitoring the consequence of phenylhydrazine incubation on the ability of the enzyme to hydrolyze artificial liposomes. In this case, no inhibition by phenylhydrazine was observed. It is also unlikely that these results reflected a direct effect of phenylhydrazine on ionomycin. Repetition of the experiments in Fig. 5 in the presence of phenylhydrazine demonstrated that the agent did not alter the ability of ionomycin to induce translocation of phosphatidylserine to the membrane exterior (not shown). Also, a different oxidizing agent, diamide, did not interfere with the responses to ionomycin (Fig. 2).
The data in this study also support the possibility that the shape transition from diskocytes to spherocytes induced by ionomycin (Fig. 6) is related to susceptibility. The agents in Fig.  2 that did not block hydrolysis by sPLA 2 after addition of ionomycin also did not alter the shape transition (based on visual inspection of the images used to generate Fig. 11). In FIG. 12. Effect of merocyanine 540 on susceptibility to sPLA 2 . Hydrolysis of erythrocyte phospholipids by sPLA 2 was assessed in the presence of 10 M merocyanine 540 (MC540) (10 min incubation prior to sPLA 2 addition) as described in Fig. 2. The "control" and "ionomycin" data of Fig. 2 are included for comparison. The amount of hydrolysis observed with merocyanine 540 was significantly different from the control (no merocyanine) by analysis of variance (*, p Ͻ 0.05, n ϭ 4).
FIG. 13. Ionomycin induces susceptibility to catalysis by human Group V sPLA 2 . The experiments shown in Fig. 1 were repeated with or without phenylhydrazine (0.5 mM). For statistical analysis, the results were normalized to the amount of hydrolysis observed in the "Ionomycin" treatment group. Significance was assessed by Student's t test as explained under "Experimental Procedures" (control, p ϭ 0.015; Phenylhydrazine ϩ ionomycin, p ϭ 0.002, n ϭ 4). contrast, phenylhydrazine inhibited both. Likewise, merocyanine 540 caused both the shape transition and increased susceptibility. Nevertheless, it is doubtful that the important factor is the actual shape of the erythrocyte per se since it was shown in the accompanying article (17) that the overall morphology of erythrocyte ghosts was unrelated to hydrolysis by sPLA 2 . Likewise, it is unlikely that the decreased cell volume resulting from K ϩ efflux during Ca 2ϩ uptake (35) could be the basis of enhanced susceptibility to sPLA 2 . This assertion is based on the observation that high KCl medium, sufficient to block the reduction in cell volume (36), failed to inhibit the vulnerability of the cells to attack by sPLA 2 . It is more likely that the molecular processes leading to the shape transition during ionomycin treatment also promote the alterations in membrane microscopic properties that result in enhanced hydrolysis by sPLA 2 .
Erythrocytes are a common model system used for studying plasma membrane structure and its relationship with membrane proteins, cytoskeleton, and a variety of pathologies. Changes that occur upon elevation of intracellular Ca 2ϩ are thought to be representative of similar changes that occur in other cells during processes such as platelet activation and apoptosis (36,54). Nevertheless, the intracellular Ca 2ϩ concentration required for these phenomena in erythrocytes is much higher than that achieved in other cells (1, 34 -36, 54, 55). Intracellular Ca 2ϩ in erythrocytes treated with ionophore equilibrates rapidly and completely with extracellular Ca 2ϩ which is high micromolar to millimolar in most investigations (1, 34 -36, 54, 55). This raises possible concerns regarding the relevance of observations in erythrocytes to physiological or pathological states (36). Such concerns are challenges one commonly faces when using a simplified system as a model. However, the benefit of obtaining information leading to testable hypotheses that can then be applied to more complex systems often exceeds the disadvantages.
The results from this study offer this benefit and also potentially relate to pathological conditions at which intracellular Ca 2ϩ levels are very high. First, as stated above, it has been difficult to identify in nucleated cells what changes in the plasma membrane were involved in the induction of susceptibility due to the diversity of membranes present in the cells. By using erythrocytes, we have identified a testable mechanism that can now be investigated in cultured cells using imaging technology such as the two-photon method. Second, many studies have suggested that membrane changes identified in erythrocytes during Ca 2ϩ loading such as PS exposure, microvesicle release, and diminished membrane fluidity may relate to alterations present in several erythrocyte pathologies: cell aging, secondary effects of hypertension, spherocytosis, thalassemias, and sickle cell disease (e.g. Refs. 36 and 56 -61). For example, sickle cells contain high levels of intracellular Ca 2ϩ (62,63) that may explain alterations to their membranes such as vesicle release (60), phosphatidylserine exposure (61), and possible enhanced susceptibility to sPLA 2 in the absence of ionophore (64). Furthermore, Ca 2ϩ -induced membrane changes in erythrocytes are analogous to events occurring in lymphocytes during apoptosis (54). Recent data has revealed that cells undergoing apoptosis also become vulnerable to sPLA 2 early in the apoptotic process (46). These observations suggest that one physiological role of sPLA 2 is to help clear aging, damaged, or dying cells in which intracellular Ca 2ϩ levels may become very high. Such could help explain participation of the enzyme in pathological conditions involving damaged cells such as ischemia, sepsis, and inflammatory disease. Based on the data of Fig. 13, it is more likely that the group V enzyme would be involved in these processes physiologically than the group IIa.
Possible relevance of these results to the more recently discovered group X enzyme has not yet been investigated.