Staphylococcal delta toxin stimulates endogenous phospholipase A2 activity and prostaglandin synthesis in fibroblasts

467

Biochimica et Biophysics Acta, 663 (1981) 467-479 @ ElsevierlNorth-Holland

Biomedical

Press

BBA 57742

STAPHYLOCOCCAL DELTA TOXIN STIMULATES ENDOGENOUS PHOSPHOLIPASE A2 ACTIVITY AND PROSTAGLANDIN SYNTHESIS IN FIBROBLASTS

JON P. DURKIN

* and W.T. SHIER

**

Cell Biology Laboratory, the Salk Institute San Diego, CA 92138 (U.S.A.)

(Received

July 28th,

for Biological

Studies,

Post Office

Box

85800,

1980)

Key words: Delta toxin; (Staphylococcus, Mouse

Cy tolysis, Phospholipase fibroblast)

AZ

; Prostaglandin

synthesis;

Summary Delta toxin, one of at least four toxins produced by pathogenic strains of the skin bacterium Staphylococcus aureus, is an amphipathic polypeptide possessing hemolytic and cytolytic activity. Delta toxin stimulates high levels of phospholipase A2 activity in 3T3 mouse fibroblasts with concomitant synthesis and release of prostaglandins. Alpha toxin, another hemolytic toxin produced by strains of S. aureus, did not stimulate phospholipase A2 or prostaglandin release in these cells. Analysis of the release of lactate dehydrogenase and fl-galactosidase (cytoplasmic and lysosomal marker enzymes, respectively) from delta-toxin-treated cells indicated that cytolytic concentrations of the toxin damage the cell-surface membrane more extensively than lysosomal membranes. During a 30 min exposure, delta toxin stimulated 3T3 cells to hydrolyze up to 32% of the lipids biosynthetically labeled by incorporation of [ 3H]arachidonic acid. A relatively high percentage of the free arachidonic acid formed in delta-toxin-treated 3T3 cells was converted to prostaglandins (up to 41.3% and 8.3% converted to chromatographically identifiable prostaglandins Ez and Fzo,, respectively, in 30 min), with optimal conversion occurring at sublytic toxin concentrations. The degree of activation of phospholipase AZ in 3T3 cells by a range of concentrations of delta toxin correlates with cytotoxicity assessed by failure to exclude trypan blue dye. Analysis of the calcium depen-

* Present address: Animal and Cell Physiology Group, Division of Biological Sciences. National Research Council of Canada ** To whom all correspondence should be sent at (present address): College of Pharmacy, University of Minnesota, Minneapolis, MN 55455. U.S.A.

468

dency of the toxin-activated phospholipase AZ was consistent with a cell-surface, Ca”dependent enzyme. The phospholipase AZ exhibits a degree of specificity for substrate lipids containing polyunsaturated fatty acid residues which can serve as precursors for prostaglandin formation. Enzymatic activity was not inhibited by diisopropylfluorophosphate (5 mM), N-ethylmaleimide (5 mM) or p-bromophenacylbromide (0.1 mM). Delta toxin did not activate detectatle phospholipase AZ in subcellular preparations containing plasma membrane.

Introduction Pathogenic strains of the common skin bacterium Staphylococcus aureus produce and excrete at least four toxins possessing hemolytic and cytolytic activity [l]. Staphylococcal alpha toxin exhibits preferential hemolysis of rabbit erythrocytes [ 21 by a mechanism which remains poorly understood [ 31. In contrast, the hemolytic activity of beta toxin is preferentially directed towards sheep erythrocytes [ 41. Wiseman and Caird [ 51, among others, have demonstrated that beta toxin is a phospholipase C, with a high specificity for sphingomyelins. Delta toxin, the subject of this report, is unique among staphylococcal hemolysins by virtue of its broad range of cellular targets, although it does exhibit a degree of specificity for human erythrocytes [ 1,6]. Preparations of delta toxin have been shown to disrupt (i) cells in tissue culture [6], (ii) bacterial protoplasts and spheroplasts [7], (iii) lysosomes [8], and (iv) mitochondria [ 91. Purified delta toxin is a protein of about 103 000 daltons which dissociates into 21000- and 5000daIton subunits under denaturing conditions [lo]. Better perceived as an aggregate than a subunit protein, delta toxin separates into several components upon isoelectric focusing with the majority of the hemolytic activity residing in a band isoelectric at approx. pH 9.5 [ll]. Bernheimer and associates [12] have recently investigated the surface behavior of delta toxin and have concluded, on the basis of its conformational flexibility and high surface activity, that the protein possesses the structural characteristics of an amphipathic protein; i.e a molecule which behaves both as a protein and a lipid. Several types of study have been reported which provide information about the mechanism of delta toxin-induced cytolysis. The lytic mechanism does not appear to involve enzyme activities associated with the toxin [ 11,121. Its toxic activity is readily inhibited by normal sera, proteolytic enzymes, and a variety of phospholipids [ll]. Sepharose-coupled toxin retains partial lytic activity, suggesting that an interaction with cell surface membranes may suffice in initiating human erythrocyte lysis [ 131. In addition to its cytolytic and hemolytic activities, delta toxin induces rat and rabbit hepatic mitochondrial lysis with concomitant inhibition of respiration and phosphorylation through a mechanism retaining electron transport integrity only in the presence of exogenous cytochrome c [ 91. The detergents Triton X-100 and deoxycholate cause similar perturbations in intact mitochondria, and consequently it has been suggested that delta toxin-induced mitochondrial lysis proceeds by a detergent-like mechanism [ 91. However, we have shown that delta toxin-induced cytolysis of

469

several cell lines proceeds in a cell cycle dependent manner [ 141. Cells in mitosis and early G1 exhibit substantial resistance to the lytic effects of delta toxin relative to cells in interphase. The observed cell cycle dependence of delta toxin cytolysis is inconsistent with a lytic mechanism based solely upon detergent action. This view is substantiated by our observation that both melittin (a potent cytolytic peptide from bee venom possessing a variety of properties in common with delta toxin) and lysophosphatidylcholine lyse mitotic and interphase 3T3 cells with equal efficiency [ 141. These observations have prompted an examination of other plausible mechanisms by which delta toxin-induced cytolysis could possibly proceed. We have shown that a variety of hemolytic and cytolytic toxins [15--171 induce prolonged activation of high levels of endogenous phospholipase AZ (phosphatide 2acylhydrolase, EC 3.1.1.4) activity in cultured cells with concomitant synthesis and release of prostaglandins into the culture medium. This enzymatic activity catalyses the hydrolysis of the cell’s own membrane phosphoglycerides to yield hydrolysis products with natural detergent properties, including lysophosphatidylcholine and free fatty acids. We have proposed [ 161 that under certain conditions this process may constitute the cytolytic mechanism for several of these toxins. In this study we report that staphylococcal delta toxin induces prolonged stimulation of phospholipase AZ activity in 3T3 fibroblast cultures, with concomitant synthesis and release of prostaglandins in high yield. Despite antibiotic therapy, staphylococcal infections remain a problem of considerable medical importance. The observation that a commonly produced staphylococcal toxin stimulates the production of prostaglandins suggests a potential mechanism for a variety of secondary effects of staphylococcal infections. In 3T3 cells, sufficient cellular lipid is hydrolyzed in the presence of delta toxin to suggest that the activation of cellular phospholipase may play an important role in the cytolytic mechanism of delta toxin. These considerations have prompted this investigation of the cytolytic mechanism of delta toxin and a partial characterization of the phospholipase AZ activated by this toxin in 3T3 mouse fibroblasts. Materials and Methods Chemicals

Delta toxin was prepared and purified by the procedure of Kreger et al. [ll] from a culture of S. aureus-Wood 46M, kindly supplied by Dr. A.W. Bernheimer. This material was a homogeneous protein prelsaration as indicated by the complete absence of detectable cysteine, arginine, histidine, tyrosine, and proline residues [ll] by amino acid analysis on a Beckman 120-B amino acid analyzer. A solution of this preparation at 18 I-(gprotein per ml is equivalent to 1 hemolytic unit per ml determined with 1% human erythrocytes by the method of Thelestam et al. [ 181. Authentic samples of alpha and delta toxins were the generous gifts of Dr. A.W. Bernheimer; identical results were obtained with delta toxin from this source and with that prepared in this laboratory. The delta toxin used in these studies contained no detectable phospholipase AZ activity when assayed [ 151 with [choline-methyl-‘4C]phosphatidylcholine

470

(New England Nuclear) as substrate under conditions capable of detecting hydrolysis of 1 pmol phosphatidylcholine per min. Prostaglandins Ez and Fza were the gifts of John Pike, of Upjohn. Ionophore A23187 was obtained from Eli Lilly, [5,6,8,9,11,12,14,15(n)-3H]arachidonic acid (135 Ci/mmol), [8,9,11, 12,14,15(n)-3H]icosa-8,11,14-trienoic acid (120 Ci/mmol) and [ 1-14C]linoleic acid (61 mCi/mmol) were obtained from Amersham; [1,2-‘4C]choline (45 pCi/ mmol) from ICN Pharmaceuticals; and [9,10(n)-3H]oleic acid (6.67 Ci/mmol) and [9,10(n)-3H]palmitic acid (498 mCi/mol) from New England Nuclear. Unless otherwise indicated, all other biochemicals were purchased from Sigma. Cell culture

Stock cultures of Swiss 3T3-4a mouse fibroblasts were obtained from Dr. M. Vogt (this Institute) and maintained in Dulbecco’s modified Eagle’s medium, supplemented with 10% calf serum (Irvine Scientific Sales Co., Irvine, CA) in a 12% CO2 atmosphere with 0.05% trypsin used to effect transfer. To prepare test cultures, cells grown to confluence were subcultured at 1.2 - 10’ cells/dish in 2 ml of Dulbecco’s modified Eagle’s medium containing 0.2% calf serum in 3.4 cm (Falcon) plastic dishes for 24 h. Cultures established in this manner were used throughout the course of the study. Phospholipase

assay

The culture medium was changed to 2 ml of fresh medium containing 10% calf serum and 1 &!i of [3H]arachidonic acid or other labeled fatty acids. A labeling period of 24 h (more than one generation time) was routinely used in order to minimize the presence of pools of varying specific activity. Under these labeling conditions the percentage of incorporated [ 3H]arachidonic acid found in the following cellular lipids were: phosphatidylcholine, 48.5 ? 4.5; phosphatidylethanolamine, 5.8 f 0.4; phosphatidylserine, 4.5 + 0 .l ; sphingomyelin plus phosphatidylinositol, 10.9 * 1.4; lysophosphoglycerides, 5.2 f 0.9; free fatty acids, 2.3 f 0.7; triacylglycerols, 4.0 f 0; and other lipids, 18.8 + 1.6. These values were obtained by extracting the lipids from washed cells on plastic dishes by a modified Folch procedure [15], fractionating by thin-layer chromatography on silica gel using the solvent system diisobutylketone/acetic acid/saline (80 : 40 : 6), v/v and determining the relative amount of radioactivity co-migrating with authentic standards visualized with iodine vapor. Under the labeling conditions used in this study with [3H]arachidonic acid, [3H]palmitic acid, [3H]icosaS,11,14-trienoic acid and [ 14C]linoleic acid, an average of lOO%, 68.8%, 69.8%, 98.2% and 85.1%, respectively, of the incorporated radioactivity was shown to be in the C-2 position of cellular phosphoglycerides extracted from cells and treated with purified phospholipase AZ (Cro talus adaman teus venom, Worthington) as described [ 15,191. To assay phospholipase activation the biosynthetically labeled cells were washed on the dish three times with 2 ml aliquots of medium and incubated for 30 min (unless otherwise indicated) in 0.6 ml Dulbecco’s modified Eagle’s medium containing delta toxin. The radioactive lipids released into the medium were extracted and analyzed as described [ 151 and the results presented as the percentage of the total incorporated label released by cells into the medium during the labeling period. All treatments were performed in triplicate. Verifi-

471

cation that the results obtained were not an artifact of tracer analysis was obtained by culturing 3T3 cells in the same manner in the absence of [3H]arachidonic acid, treating with delta toxin, extracting and chromatographing cell and medium lipids as described above without addition of authentic standards. The ratio of lysophosphatidylcholine phosphorus to phosphatidylcholine phosphorus was determined in eluates from chromatograms using the procedure of Ames [20]. Cy to toxicity Cells were cultured and treated in a manner identical to that described for phosphohpase assays except that no radiolabel was added to the culture medium. The cytolytic effects of delta toxin were assessed by determining the percentage of the cells able to exclude 0.013% trypan blue dye following a 30 min incubation with the indicated concentrations of delta toxin. Assay for fi-galactosidase and lactate dehydrogenase activity Following the treatment of 3T3 cells (unlabeled) with delta toxin for 30 min, the medium (600 ~1) was removed and centrifuged to remove any nonadherent cells, and aliquots (250 ~1) were tested for fl-galactosidase [21] and lactate dehydrogenase [22] activities. The total cellular content of these activities was determined by scraping toxin treated and control untreated cells from dishes into 2.0 ml HzO, combining with the original medium and assaying aliquots for each enzymatic activity. Delta toxin (up to 100 pg/ml) did not effect the total recoverable activity (i.e., activity present in lysed cells plus medium) for either enzyme. The presence of (6%) Triton X-100 in the Pgalactosidase assay medium did not alter the results obtained. All treatments were performed in triplicate. Results Characteristics of delta toxin-induced cytolysis in 3T3 fibroblasts Swiss 3T3 fibroblasts were highly sensitive (LD5,, a 13 pg/ml) to the lytic effects of staphylococcal delta toxin, as indicated by failure to exclude trypan blue dye (Fig. 1). Within lo-15 min after the addition of delta toxin (30 pg/ ml) to a cell population, profound changes in cellular morphology occurred (Fig. 2). Microscopic examination of the cells revealed that cytoplasmic protrusions (blebs) developed at multiple sites on the cell surface and thereafter expanded in a rapid and continuous manner. Within 30 min after the addition of toxin, these protrusions ruptured, releasing intracellular contents into the external medium. At this time the cells were found to stain strongly with 0.013% trypan blue dye and to be incapable of further growth upon the removal of residual toxin by washings [ 141. The effects of delta toxin on the release of cellular lactate dehydrogenase and flgalactosidase activities into the external medium are shown in Fig. 3. Lactate dehydrogenase and Pgalactosidase activities are highly compartmentalized in the cytoplasm and lysosomes, respectively, and have proved to be useful markers for damage to cell surface [22] and lysosomal [23] membranes, respectively. The incubation of cell populations with delta toxin for 30 min

472

J 0t

y-y+/ 205

, I

2

STAPHYLOCOCCAL

5 DELTA

IO TOXIN

50

20

100

(yg/ml)

Fig. 1. Effct of delta toxin concentration on viability of 3T3 fibroblasts in culture. Following a 30 min incubation with delta toxin, the viability of the cell population was assessed by its ability to exclude trypan blue dye.

resulted in the release into the medium of the majority (68%) of the lactate dehydrogenase activity present in the cells. In contrast, a much lower percentage of the total cellular content of fl-galactosidase activity was found in the medium following delta toxin treatment (11%). These results suggest a greater degree of toxin-induced damage to the cell-surface membrane than to lysosoma1 membranes. Stimulation of endogenous phospholipase A2 activity by delta toxin Staphylococcal delta toxin stimulated the release of radioactive arachidonic acid and prostaglandins from cultured 3T3 fibroblasts biosynthetically labeled with [ 3H]arachidonic acid (Fig. 4). In contrast, staphylococcal alpha toxin did not stimulate the release of detectable phospholipase AZ products during a 1 h incubation of [3H]arachidonic acid-labeled 3T3 cells with a range (0.01 pg/ml to 50 pg/ml) of toxin concentrations. We have previously suggested that the activation of endogenous phospholipase AZ for prolonged periods can be a lethal cellular event [15,16] and may constitute a widespread toxic mechanism. In the case of delta toxin, about 15% deacylation of the [3H]arachidonate labeled lipids of the cell was achieved within 30 min at toxin concentrations capable of 50% destruction of the cell population in the same time period. This degree of phospholipid hydrolysis is sufficient to account for toxin-induced cytolysis, particularly if the plasma membrane is hydrolyzed more extensively than other membranes. Prostaglandin production in cultured cells appears to be dependent on the generation of free arachidonic acid in the cell [ 241. As shown in Fig. 4b, substantial percentages of labeled arachidonic acid were released in the form of prostaglandin E, and to a lesser extent as F-type prostaglandins when cells were treated with lytic and sublytic concentrations of delta toxin. The optimal conversion of free arachidonic acid produced by the delta taxi-stimulated phospholipase AZ to prostaglandin Ez occurred at 3.5 pg/ml (41.3% conversion). More efficient conversion of free arachidonic acid to prostaglandin Ez was found to occur at sublytic delta toxin concentrations.

473

474 40 A

4

2

STAPHYLOCOCCAL

DELTA

TOXIN

Fig. 3. Release of cellular P-galactosidase dase (0). lactate dehydrogenase (0).

( pg

/ml

0 l-?dl!s! 0 02 05 I STAPHYLOCOCCAL

)

and lactate

dehydrogenase

Fig. 4. Effect of delta toxin concentration on the release able as (A) total C3Hlarachidonic acid-derived free fatty F-type prostaglandins (A).

activities

of radioactivity acids (m), and

2 5 DELTA

IO 20 50 TOXIN (fig/ml)

by delta toxin.

chromatographicdy (B) prostaglandio

no

@&lactosi-

E2

identify(o) and

The rate of release of arachidonic acidderived fatty acids from delta toxin treated 3T3 cells is shown in Fig. 5. Following an initial lag period (approx. 5 min), rapid release of free fatty acids is observed. Since delta toxin exists in solution in a highly aggregated state [lo], the initial lag observed in the release of phospholipase AZ products may reflect a slow dissociation of toxin into active subunits upon exposure to cellular components. Partial characterization of the delta toxin-induced phospholipase A2 Virtually all phospholipase AZ preparations exhibit a dependence on Ca2’ at concentrations of 1 mM or more [25]. All well characterized preparations have been reported to be activated by Ca2+ except the lysosomal phospholipase A with a low pH optimum, which has been reported to be inhibited by Ca2’ and activated by EDTA [ 261. The Ca2’dependence of the phospholipase A2 activated in 3T3 cells by delta toxin was investigated by determining the extent of toxin-induced stimulation of the enzyme under the following sets of conditions: (i) in medium (free Ca2+, 1.8 mM); (ii) in medium supplemented with (10 mM) EDTA, a chelating agent which effectively removes free Ca” and Mg2+; and (iii) in medium supplemented with (10 mM) EDTA and (5 @I) A23187, an ionophore which transports divalent cations across cell membranes. In 1969 Borle [27] reported that HeLa cells biosynthetically labeled with 45Ca2’ contain two exchangeable pools of Ca2+, one which exchanges rapidly (half-time about 1.5 min) and one which exchanges more slowly (half-time about 30 min). These pools were interpreted as representing cell surface and

475

/

0 TIME (mm1

FL

IO

STAPHYLOCOCCAL

DELTA

TOXIN

20

50 (ug/ml)

Fig. 5. Rate of release of radiolabeled [3Hlarachidonic acid-derived free fatty acids from 3T3 fibroblasts treated with delta toxin at 0 n&n1 (*I 4 &ml (A), 20 n&n1 (0). and 40 &ml (a). Fig. 6. Effect of Ca%, EDTA and divalent cation ionophore A23187 on the release of radiolabeled free fatty acids from biosynthetically labeled 3T3 cells. Assays were conducted in medium (1.8 mM Ca’M) (0); medium containing 10 mM EDTA (a); and medium containing 10 mM EDTA and 5 PM ionophore A23187 (A).

cytoplasmic Ca”, respectively [27]. Similar results have been obtained with 3T3 cells under the conditions used in this study [15,19] and the addition of the divalent cation ionophore A23187 was found to cause the rapid exchange of both pools (half-time, 1.35 min). Consequently, a phospholipase AZ with a cell surface Ca*‘-binding site would be expected to be inhibited within minutes in medium containing (10 mM) EDTA in the presence or absence of (5 PM) A23187. In contrast, a cytoplasmic Ca*+dependent phospholipase A2 should be inhibited by EDTA much more effectively in the presence of the ionophore during a 30 min assay. A lysosomal phospholipase A (low pH optimum) should be stimulated in the presence of EDTA and ionophore, as has been observed with the direct lytic factor from the venom of Hemachatus hemachatus [ 151. As shown in Fig. 6, the phospholipase A2 stimulated by delta toxin in the presence of Ca*+ was inhibited under condition8 in which free Ca*+ was removed from the medium by chelation. Such results are consistent with the activation of a Ca*‘dependent enzyme with a cell surface Ca*+-binding site. The low levels of phospholipase A activity observed in the presence of EDTA (with or without A23187) and cytolytic concentrations of toxin may reflect secondary lysosomal activation following substantial toxin-induced damage to the cell surface membrane. Toxin-induced permeability of the cell to Ca*+ has been proposed as a general mechanism by which low levels of cytotoxins cause cell lysis over extended time periods [ 281. In 3T3 cells, basal phospholipase AZ activity was found not to be significantly affected during a 30 min treatment with ionophore A23187 (5 fl) (data not shown). Thus, activation of an intracellular phospholipase by a delta toxin-induced Ca*+ flux appears to be an unlikely mechanism during the first 30 min of treatment with the toxin. The substrate specificity of delta toxin-stimulated phospholipase AZ was investigated. Release of radiolabel from 3T3 cells during toxin treatment was analyzed using cell populations biosynthetically labeled with various fatty acids (Fig. 7). A much higher percentage of incorporated radioactivity was released

476

O-

0

15

3 STAPHYLOCOCCAL

6

15 DELTA

TOXIN

30 (,,g,,,,

60 )

Fig. 7. Substrate specificity of delta toxin stimulated endogenous phospholipase A2 activity stimulated in 3T3 fibroblasta by delta toxin. Release of radiolabeled fatty acids from cells biosynthetically labeled with [3H]arachidonic acid (0). [3H]oleic acid (0). [3H]palmitic acid (o), [3Hlicosa-8,11,14-triienoic acid (a) and [14Cllinoleic acid (a) was determined in the presence of a range of delta toxin concentrations.

into the medium from cells labeled with [3H]arachidonic acid or C3H]icosa-8, 11,14-trienoic acid than from cells labeled with other fatty acids. Since both of these unsaturated fatty acids are known prostaglandin precursors, these results are consistent with delta toxin stimulating a phospholipase AZ with a degree of specificity for fatty acid residues which can serve as prostaglandin precursors. The ability of delta toxin to induce stimulation of phospholipase AZ activity and prostaglandin release in [3H]arachidonic acid-labeled 3T3 cells was tested in the presence of three potential phospholipase AZ inhibitors. paruBromphenacylbromide (0.1 mM) and N-ethylmalemide (5 mM), which alkylate histidine [29] and sulhydryl [20] groups, respectively, have been shown to inhibit different phospholipase AZ activities stimulated in 3T3 cells by other agents [ 17,131. Serum-stimulated prostaglandin release from a transformed Balb/3T3 cell line is inhibited by agents which derivatize active-site serine residues [32], a class of agents which includes diisopropylfluorophosphate [ 301. However, phospholipase AZ activity and prostaglandin release stimulated in 3T3 cells by delta toxin was found to be unaltered by the presence of any of these reagents with or without pretreatment by the agent. Microsomes prepared from 3T3 cells biosynthetically labeled with [ 1,2-14C]choline [15] were assayed for delta toxin-activatable phospholipase AZ activity. Although these preparations contained plasma membrane, as indicated by Fi’nucleotidase and adenylate cyclase activities, no detectable increase in phosphospholipase activity was observed when microsomes were assayed [ 151 with delta toxin concentrations optimal for activation of phospholipase AZ in intact cells (data not shown). In fact, (40 pg/ml) delta toxin was found to inhibit the phospholipase AZ activity stimulated in these preparations by (10 mM) CaClz. Discussion It is apparent from the results presented in Figs. 1 and 4 that the delta toxin concentrations required for activation of phospholipase AZ and for cytotoxicity

correlate; i.e. both activities occur within the same delta toxin concentration range (5-50 I,cg/ml)and exhibit half maximal effects at the same concentration (10 pg/ml) of toxin. Delta toxin cytotoxicity was observed to increase dramatically when more than 10% of the cellular [3H]arachidonic acid-labeled lipids had been hydrolyzed. The observation that delta toxin caused more damage to the cell-surface membrane than to lysosomal membranes (Fig. 3) argues against the activation of lysosomal enzymes being a primary and critical event in delta toxin cytolysis. However, the possibility of secondary involvement of lysosomal enzymes cannot be eliminated, particularly in light of the finding that significant activation of phospholipase A by delta toxin occurred in the presence of the Ca’+chelating agent EDTA (Fig. 6). The Ca2+dependency of the delta toxin&imulated phospholipase A2 (Fig. 6) suggests a predominantly cell-surface Ca”-binding site and, hence, a cellsurface phospholipase A2. These results are consistent with a model in which activation of a cell surface phospholipase A2 is the primary cytolytic event in the mechanism of delta toxin cytolysis, since the cell surface appears to be both the site of phospholipase A2 activation and the site of major toxin-induced membrane damage. Additional evidence that phospholipase A2 activity plays a role in delta toxin cytolysis is provided by our observation that mitotic cells, resistant to the effects of delta toxin, have substantially reduced levels of activatable phospholipase A2 activity relative to cells in interphase [ 141. This reduction in cytotoxicity and in phospholipase A2 response to delta toxin in mitotic cells could represent one or more of the following: (i) reduced association of the toxin with mitotic cells; (ii) reduced efficiency of the phospholipase activation mechanism by cell-associated toxin; (iii) reduced responsiveness of the phospholipase A2 molecules in mitotic cells to stimulation; or (iv) fewer phospholipase A2 molecules in mitotic cells. If either of the latter two possibilities were true, the phospholipase A2 activated by delta toxin could play a significant cell-cycledependent regulatory role either by controlling prostaglandin synthesis or by altering membrane lipid properties. Permeabilization of the cell-surface membrane to Ca2+ has been proposed as a general cytolytic mechanism [ 281 by which low levels of cytotoxins induce a slow rate of death in cultured cells (approx. 6 h). Under the conditions used in this study (i.e., the assessment of cytotoxicity after a 30 min treatment) permeabilization of the cell-surface membrane (Fig. 3) to Ca2+ cannot be responsible for cytotoxicity, because effective concentrations of the divalent cation ionophore A23187 are not cytolytic within 30 min (data not shown). We have suggested [ 15,161 that for some toxins at higher concentrations, activation of cellular phospholipase A2 may be the principal cytolytic mechanism responsible for cell death occurring in an intermediate time period (less than 30 min). At high concentrations of surface-active toxins, such as melittin, direct surfactant effects appear responsible for rapid (within 1 min) cytolysis. Delta toxin exhibits substantially less surface activity than melittin [ 12,331 and therefore, very high concentrations of toxin would probably be required to effect cytolysis by surfactant action alone. Under the conditions used in this study, direct surfactant effects are unlikely to account for delta toxin-induced cytolysis. This conclusion is supported by our observation that mitotic and early G1 cells are resistant to the effects of delta toxin at concentrations nor-

418

mally cytolytic for cells in interphase [14]. In fact, in cells which have been examined, mitotic cells are more susceptible to detergent-induced cytolysis than are cells in interphase [ 34,351. In addition, lysophosphatidylcholine and melittin (both of which exhibit high surface activty) were found to lyse mitotic and interphase cells at equal effectiveness over a wide range of concentrations [ 141, Thus, the observed cell cycle dependency of delta toxin-induced lysis of cultured cells is not compatible with a cytolytic mechanism proceeding by detergent action alone. Thelestam and Mollby [36,37] have observed that delta toxin induced timeand temperature-independent release of small labeled molecules from cultured human fibroblasts before the induction of detectable morphological changes associated with cytolysis. With larger labeled molecules, leakage from the cells was induced by delta toxin in a slow, time-dependent manner by low concentrations of toxin, or rapidly by high concentrations of toxin [36]. These results are consistent with the delta toxin inducing non-enzymatic membrane changes, which do not lead to rapid cell lysis, as well as time-dependent effects which do. Activation of cellular phospholipase AZ is a timedependent cytolytic process consistent with these observations. The observation that delta toxin stimulates cellular production of free arachidonic acid and other prostaglandin precursors suggests plausible mechanisms for secondary effects of staphylococcal infections at various sites in the body. Free arachidonic acid introduced into virtually any tissue, either added from exogenous sources or released from tissue phospholipids, is rapidly converted into a variety of metabolites as a result of the action of fatty acid cyclooxygenase or lipoxygenase [ 381. The initial products of these enzymes are converted to various prostaglandins and other metabolites, the nature of which depends on the tissue in question. Many of these metabolites have potent biological activities [39] and appear to play roles in processes such as the mediation of inflammation and the regulation of blood clotting, smooth muscle contraction, chemotaxis and maintenance of blood pressure. An example of a plausible secondary effect of S. aureus infection due to the production of prostaglandins in infected tissues is the induction of miscarriages by vaginal infections [ 401. Prostaglandin F,, , a known abortifacient [ 411, could be produced in the vagina in response to delta toxin production and contribute with other factors in promoting premature parturition. Similarly, delta toxin-induced prostaglandin synthesis could play a role in mediating the inflammatory response associated with localized infections and the alterations in blood pressure during S. aureus-induced septic shock. Subsequent to the preparation of this manuscript it was reported [42] that delta toxin induces arachidonic acid release and prostaglandin synthesis in lOTi cells. Sublytic concentrations of delta toxin were also found to inhibit the binding of epidermal growth factor to cell-surface receptors in HeLa and rat embryo cells. The authors indicate that delta toxin shares several effects in cell culture with tumor promoters, such as 12-O-tetradecanoyl-phorbol-13-acetate. These findings suggest that delta toxin may prove to be a useful probe in studies of membrane structure and function.

479

Acknowledgments We gratefully acknowledge gifts of samples of staphylococcal toxins and S. aureus Wood 46M from Dr. A.W. Bernheimer and the technical assistance of Mr. J.T. Trotter. This research was supported in part by a fellowship from the National Research Council of Canada (to J.P.D.); a grant from the Cystic Fibrosis Foundation; Grant Number CA20636, awarded by the National Cancer Institute, D.H.E.W., and Grant Number PCM80-11784, awarded by the National Science Foundation. References 1 Jeljaszewicz. J. (1912) in The Staphylococci (Cohen, J.O.. ed.), pp. 249-280. Wiley Interscience, New York 2 Bernheimer, A.W. (1965) Ann. N.Y. Acad. Sci. 128.112-123 3 Cassidy, P. and Harshman, S. (1916) Biochemistry 15.2348-2355 4 Smith, M.L. and Price, S.A. (1938) J. Pathol. Bacterial. 41. 361-311 13,369-316 5 Wiseman. G.M. and Caird. J.D. (1961) Can. J. Microbial. 6 Gladstone, G.P. and Yosbida, A. (1961) Brit. J. Expt. Fathol. 48, 11-19 7 Bemheimer. A.W.. Avigad, L.S. and Grushoff. P. (1968) J. Bacterial. 96.481491 8 Bernheimer. A.W. and Swartz, L.L. (1964) J. Bacterial. 81.1100-1104 9 Rahal. J.J. (1912) J. Inf. Diseases 126.96-103 10 Kantor. H.S., Temples, B. and Shaw. W.V. (1912) Arch. Biochem. Biophys. 151.142-156 11 Kreger, A.S.. Kim, KS., Zaboretsky, F. and Bemheimer. A.W. (1911) Infect. Immun. 3.449465 G.. Basu. M.K.. Buckelew. A.R. and Bemheimer, A.W. (19’71) Biochim. Biophys. Acta 465, 12 Colacicco, 318390 13 Lee, S.H. and Haque, R. (1916) Biochem. Biophys. Res. Commun. 68.1116-1118 14 Durkin, J.P. and Shier, W.T. (1980) Biochem. Biophys. Res. Commun. 94.980-981 15 Shier, W.T. (1919) Proc. Nati. Acad. Sci. U.S.A. 16.195-199 16 Shier, W.T. (1919) Toxicon 11, S166 11 Shier, W.T. and Trotter, J.T. (1980) Biochim. Biophys. Acta 619, 235-246 18 Thelestam. M.. M5iIby. R. and Wadstrom. T. (1913) Infect. Immun. 8.938-946 19 Shier, W.T. (1980) Proc. Nati. Acad. Sci. U.S.A. 77.131-141 20 Ames, B.N. (1966) Methods Enzymol. 8.115-118 21 Ho. M.W., Seek. J.. Schmidt, D, Veath. M.L., Johnsson. W.. Brady, R.O. and O’Brien, J.S. (1972) Am. J. Hum. Genet. 24.3145 22 Bergmeyer, H.V., Bern& E. and Hess, B. (1965) in Methods of Enzymatic Analysis (Bergmeyer, H.V.. ed.). pp. 136-143. Academic Press, New York 23 Vaes. G. (1966) Methods Enzymol. 8, 509-514 24 Kunze. h. and Vogt, W. (1911) Ann. N.Y. Acad. Sci. 180.123-125 25 van den Bosch, H. (1914) Annu. Rev. Biochem. 43, 243-211 G.L., Boshouwers. F.M.G. and van Deenen. L.L.M. (1969) J. Lipid Res. 10, 26 Waite, M.. Scherphof. 411420 27 Borle, A.B. (1969) J. Gen. Phys. 53.51-69 28 Schanne, F.A.X.. Kane, A.B.. Young, E.E. and Farber, J.L. (1919) Science 206.100-102 13.1446-1454 29 Volwerk, J.J., Pieterson, W.A. and de Haas, G.H. (1914) Biochemistry A. (1910) in Advances in Protein Chemistry 30 Spande, T.F., Witkop, B.. Degani, Y. and Patchornik, (Anfinsen, C.B.. Edsall, J.T. and Richards, F.M.. eds.). Vol. 24. PP. 91-308. Academic Press, New York 31 Shier, W.T.. Durkin, J.P. and Trotter, J.T. (1919) Fed. Proc. 38,406 32 Pang. S.S., Hong, S.L. and Levine, L. (1911) J. Biol. Chem. 252. 1408-1413 G. and Weissman. G. (1969) J. Biol. Chem. 244.3515-3582 33 Sessa. G.. Freer, J.H., Colacicco, 34 Graham, J.M.. Sumner. M.C.B.. Curtis, D.H. and Psstemak, C.A. (1913) Nature 246,291-295 35 Pasternak. C.A.. Sumner, M.C.B. and Collin. R.C.L.S. (1914) in Celi Cycle Controls (PadiBa. G.M.. Cameron, I.L. and Zimmerman, A., eds.), PP. 111-124. Academic Press, New York 36 Thelestam, M. and Miillby, R. (1915) Infect. Immun. 11.640-648 31 Thelestam. M. and M6Bby. R. (1915) Infect. Immun. 12, 225-232 in Phospholipases and Prostaglandins (Gaili, C., GaBi, G. and Porcellati. G.. eds.). 38 Vogt, W. (1918) Vol. 3, pp. 8945. Raven Press, New York 39 Bartmann. W. (1915) Angew. Chem. Int. Edit. 14.331344 40 Naeye, R.L. and Peters. E.C. (1918) Pediatrics 61.111-111 Plenum (RamweB, P.W.. ed.). Vol. 2, pp. 115-203. 41 Persaud. T.V.N. (1914) in The Prostaglandins Press, New York 42 Umezawa, K., Weinstein, I.B. and Shaw, W.V. (1980) Biochem. Biophys. Res. Commun. 94, 625429