stimulated the PA formation and hemolysis induced by alpha-toxin, and GDP[ S] ... hemolysis is due to activation of phospholipid metabolism systems through ...
INFECTION AND IMMUNITY, Sept. 1996, p. 3930–3933 0019-9567/96/$04.0010 Copyright q 1996, American Society for Microbiology
Vol. 64, No. 9
Phospholipid Metabolism Induced by Clostridium perfringens Alpha-Toxin Elicits a Hot-Cold Type of Hemolysis in Rabbit Erythrocytes SADAYUKI OCHI, KATSUSHI HASHIMOTO, MASAHIRO NAGAHAMA, AND JUN SAKURAI* Department of Microbiology, Faculty of Pharmaceutical Sciences, Tokushima Bunri University, Tokushima 770, Japan Received 18 March 1996/Returned for modification 23 May 1996/Accepted 21 June 1996
GTP and AlF42 significantly stimulated the late phosphatidic acid (PA) formation induced by Clostridium perfringens alpha-toxin in rabbit erythrocyte lysates. Pertussis toxin blocked the PA production. AlF42 markedly enhanced phosphatidylethanol production induced by alpha-toxin in the presence of ethanol. GTP[gS] stimulated the PA formation and hemolysis induced by alpha-toxin, and GDP[bS] inhibited them. An H-to-G mutation at position 126 (H126G) induced the PA formation and hemolysis in a Co21 concentration-dependent manner. H148G induced neither the PA formation nor hemolysis. These results suggest that the toxin-induced hemolysis is due to activation of phospholipid metabolism systems through GTP-binding protein. branes induced by the toxin, we investigated the relationship between the PA formation and hemolysis induced by the toxin in rabbit erythrocytes. Purification of wild-type and variant alpha-toxins was performed as described in detail previously (11). Erythrocyte membranes were prepared from rabbit erythrocytes (6 3 1011 cells per ml) suspended in 0.02 M Tris-HCl buffer (pH 7.5) (TB) as described previously (15). Erythrocyte lysates were prepared by lysing washed erythrocytes as follows. Erythrocytes (6 3 1011 cells per ml) were lysed in an equal volume of ice-cold TB. The resultant lysed cells were termed the lysate. Saponin-treated erythrocytes were prepared according to the previously described method (15) except that b-escin was used instead of saponin (gypsophila). Hemolytic activity of wild-type and variant alpha-toxin was determined in TB containing 0.9% NaCl and 0.3 mM Ca21, and the concentration of phosphatidylethanol in membranes was determined as described previously (15). The PA concentration was determined, and assay of ADP-ribosylation was performed as described previously (16). Protein concentration was determined by the method of Lowry et al. (8), by using bovine serum albumin as a standard. All mean values are shown with their calculated standard errors. The Student’s t test was used to determine the significance of differences between controls and experimental groups; a P value of 0.05 or less was considered statistically significant. The effect of GTP-binding protein on the late PA formation induced by the toxin was investigated. When erythrocyte lysates were incubated with the toxin (5 ng) in the presence of GTP at 378C for 20 min, GTP at a range of 10 to 250 mM stimulated the toxin-induced [g-32P]PA formation (the late PA formation) in the presence of [g-32P]ATP in a dose-dependent manner. When membranes were used under these conditions, GTP only slightly stimulated the PA formation. AlF42 at concentrations of 0.25 to 0.75 mM also stimulated the late PA formation (data not shown). We have reported that the late PA formation is dependent on activation of endogenous PLD in erythrocytes, as shown in Fig. 1 (15). PLD is known to catalyze a transphosphatidylation reaction in which the phosphatidyl moiety of the phospholipid substrate is transferred to appropriate nucleophiles such as ethanol (4). Therefore, to determine whether AlF42 activates a PLD pathway in lysates, PA
Clostridium perfringens alpha-toxin, which is a main agent of gas gangrene, is known to have hemolytic, lethal, dermonecrotic, and phospholipase C (PLC) activities (13). We have reported that the toxin induces contraction of isolated rat aorta (5) and ileum (14), and that the toxin-elicited contraction of the isolated aorta and ileum is related to activation of phosphatidylinositol turnover (5, 14). Especially the toxin-induced contraction of isolated aorta was found to be closely linked to stimulation of thromboxane A2 synthesis (6). Similar results have subsequently been reported with PLC produced by other microorganisms (7, 9). Alpha-toxin is known to cause hemolysis of various animal erythrocytes (13). We have reported that incubation of rabbit erythrocyte membranes with the toxin results in biphasic production of phosphatidic acid (PA) (15). Furthermore, as shown in Fig. 1, the rapid PA formation induced by the toxin is due to activation of endogenous PLC regulated by GTP-binding protein in erythrocyte membranes (15, 16), and the late PA formation is dependent on activation of endogenous phospholipase D (PLD) (15). Neomycin inhibits both the toxin-induced increase in turbidity of egg yolk suspensions and hemolysis of saponin-permeabilized rabbit erythrocytes but does not inhibit the hemolysis of intact erythrocytes (15). Moreover, the inhibitor blocks the toxin-induced PA formation (15). From these data, it was suggested that the toxin-induced hemolysis is related to the toxin-induced PA formation. Recently, we reported that the hemolytic activity of the toxin was not separated from the PLC activity of the toxin by replacement of four histidine residues, which are essential to the activities of the toxin, with neutral amino acids by sitedirected mutagenesis (11). Artificial membranes, liposomes, composed of choline-containing phospholipids which are hydrolyzed by the toxin were shown to be destroyed by the toxin (10). However, the Bacillus cereus PLC, which does not possess hemolytic activity, was reported to hydrolyze phosphatidylcholine in liposomes but not to destroy the liposomes (2). In the present work, to clarify the destruction of biological mem* Corresponding author. Mailing address: Department of Microbiology, Faculty of Pharmaceutical Sciences, Tokushima Bunri University, Yamashiro-cho, Tokushima 770, Japan. Phone: 0886-22-9611. Fax: 0886-55-3051. 3930
VOL. 64, 1996
NOTES
3931
FIG. 1. Model for PA formation pathways activated by alpha-toxin in rabbit erythrocytes.
and phosphatidylethanol were measured after incubation of lysates with the toxin in the presence or absence of AlF42. As shown in Table 1, the toxin or 0.75 mM AlF42 alone significantly stimulated PA and phosphatidylethanol production. The magnitude of the toxin-induced PA formation stimulated by AlF42 almost resembled that of the toxin-induced phosphatidylethanol production stimulated by the agent. The toxin stimulated at least three times more PA and phosphatidylethanol production in the presence of AlF42 than was stimulated by the toxin or AlF42 alone. Further, to clarify whether the late PA formation is regulated through GTP-binding protein, the effect of pertussis toxin, which ADP-ribosylates GTP-binding protein (3), on the late PA formation was investigated (data not shown). Pretreatment of erythrocyte lysates with pertussis
TABLE 1. Effect of AlF42 on PA and phosphatidylethanol production induced by alpha-toxin in rabbit erythrocyte lysates Treatment
PA (cpm)a
Phosphatidylethanol (dpm)b
None AlF42 Alpha-toxin Alpha-toxin 1 AlF42
741 6 131 9,890 6 1,080 5,190 6 1,450 24,800 6 3,470c
572 6 110 10,940 6 1,040 5,660 6 1,510 28,600 6 2,990c
a Erythrocyte lysates were incubated with alpha-toxin (5 ng of protein per ml) and 10 mCi of [g-32P]ATP per ml in the presence or absence of 0.75 mM AlF42 as described in the text. PA formation was determined as described in the text. Values represent means 6 standard errors for four to five experiments. b Erythrocyte lysates were incubated with alpha-toxin (5 ng of protein per ml) and 50 mCi of [3H]ethanol per ml in the presence or absence of 0.75 mM AlF42 as described in the text. Phosphatidylethanol formation was determined as described in the text. Values represent means 6 standard errors for four to five experiments. c P , 0.01, compared with the values induced by the toxin or AlF42 alone.
toxin (10 mg) at 378C for 30 min resulted in almost total abolition of the late PA formation. The treatment with pertussis toxin in the presence of 10 mCi of [32P]NAD under the condition resulted in labeling of a protein band corresponding to approximately 41 kDa on sodium dodecyl sulfate-polyacrylamide gel electrophoresis. On the other hand, alpha-toxininduced PA formation in lysates treated with heated pertussis toxin was not different from that in untreated lysates, indicating that heated pertussis toxin is unable to block the PA formation. Heated pertussis toxin also did not ADP-ribosylate protein in lysates. It is likely that the late PA formation is connected with activation of GTP-binding protein in rabbit erythrocytes. The effects of GTP[gS] and GDP[bS] on the toxin-induced hemolysis and PA formation were investigated. As shown in Table 2, the toxin-induced hemolysis in the presence of GTP[gS] at concentrations of 0.5 and 5.0 mM was approximately 150 and 250%, respectively, of that induced by the toxin in the absence of GTP[gS]. The toxin-evoked PA formation in the presence of 0.5 and 5.0 mM GTP[gS] was approximately 130 and 260% of control, respectively. The data show that GTP[gS] dose dependently stimulated the toxin-induced hemolysis and PA formation. Incubation with GDP[bS] at concentrations of 0.02, 0.2, and 2 mM resulted in approximately 87, 68, and 15%, respectively, of the hemolysis induced by the toxin in the absence of GDP[bS]. The toxin-evoked PA formation in the presence of 0.2 and 2.0 mM GDP[bS] was approximately 65 and 22% of control, respectively. GDP[bS] dose dependently inhibited the toxin-induced hemolysis and PA formation. It therefore appears that GTP-binding protein regulates the toxin-induced hemolysis as well as the PA formation. We have reported that the H-to-G mutation at position 126 (H126G) lyses sheep
3932
NOTES
INFECT. IMMUN.
TABLE 2. Effect of GTP[gS] or GDP[bS] on hemolysis and PA production caused by alpha-toxin Nucleotide analog (mM)
GTP[gS] 0 0.5 1.0 5.0 GDP[bS] 0 20 200 2,000
PAb (cpm)
Alpha-toxin (ng/ml)
Hemolysisa (%)
3 3 3 3
20 6 3 30 6 2c 41 6 5c 50 6 6e
841 6 40 894 6 51 NDd 980 6 59c
1,790 6 173 2,330 6 172c NDd 4,650 6 410c
5 5 5 5
100 87 6 4c 68 6 3c 15 6 5e
871 6 57 884 6 79 788 6 53 743 6 40
8,510 6 1,270 7,650 6 907 5,610 6 129c 1,900 6 156e
Without toxin
With toxin
a Saponin-treated erythrocytes were incubated with alpha-toxin (3 or 5 ng of protein per ml) and various concentrations of guanine nucleotide analog as described in the text. Hemolytic activity was determined as described in the text. Values represent means 6 standard errors for four to five experiments. b Rabbit erythrocyte lysates were incubated with alpha-toxin (3 or 5 ng of protein per ml) and various concentrations of guanine nucleotide analog in the presence of 10 mCi of [g-32P]ATP per ml as described in the text. PA formation was determined as described in the text. Values represent means 6 standard errors for four to five experiments. c P , 0.05, compared with hemolysis or PA formation induced by 3 or 5 ng of the toxin alone. d Not determined. e P , 0.01, compared with hemolysis or PA formation induced by 3 or 5 ng of the toxin alone.
erythrocytes and hydrolyzes phosphatidylcholine and sphingomyelin in the presence of divalent cations such as Co21, and that the mutant H148G indicates no activities in the presence or absence of these cations (11). To determine whether the toxin-induced PA formation is linked to hemolysis in rabbit erythrocytes, we measured hemolysis and PA formation induced by H126G or H148G (Table 3). When H126G was incubated with rabbit erythrocytes and lysates in the presence of 10 mM Co21, the variant toxin dose dependently induced hemolysis and PA formation. In addition, when rabbit erythrocytes and lysates were treated with 1.0 mg of the variant toxin per ml in the presence of Co21, the levels of toxin-induced hemolysis and PA formation were dependent on Co21 concentration. However, H148G showed neither PA formation nor hemolysis in the presence of 100 mM Co21. It therefore appears that the occurrence of hemolysis is in parallel with that of the late PA formation. We have reported that the late PA formation is inhibited by addition of primary alcohol such as ethanol and that the toxin stimulates phosphatidylethanol formation in the presence of ethanol, showing that the late PA formation is mediated through a PLD pathway (15). GTP enhanced the late PA formation in erythrocyte lysates, but did so only slightly in membranes, suggesting that the late PA formation is activated by the combination of GTP and a factor such as GTP-binding protein in the rabbit erythrocyte cytosol. AlF42 and GTP[gS], which are known to activate GTP-binding protein (4), activated the late PA formation, and AlF42 also stimulated the toxin-induced phosphatidylethanol formation in erythrocyte lysates. Moreover, pretreatment of erythrocyte lysates with pertussis toxin resulted in complete loss of the late PA formation. Therefore, endogenous PLD seems to be regulated by GTPbinding protein in rabbit erythrocytes (Fig. 1). Recently, PLD has been observed in mammalian cells and tissues (4). PA formation induced by various agonists has been reported to be produced from phospholipids such as phosphatidylcholine via
activation of PLD rather than a PLC-diacylglyceride kinase pathway (4). Agwu et al. (1) and Olson et al. (12) report that PLD activation is mediated through activation of GTP-binding protein. These reports support the possibility that the toxin activates the enzyme through activation of GTP-binding protein. On the other hand, we have reported that the biphasic formation of PA induced by the toxin is first mediated by an endogenous PLC pathway and then by a PLD-phosphatidate phosphatase pathway (15), and that the rapid PA formation is elicited by endogenous PLC which is regulated by GTP-binding protein (Fig. 1) (16). Therefore, it is possible that the occurrence of the late PA formation is the result of the activation of the rapid PA formation through GTP-binding protein, suggesting that the stimulation of the late PA formation by GTP and AlF42 may be a secondary event to the activation of the rapid PA formation by these agents. However, the possibility that the elevation of the late PA formation is due to the activation of PLC through GTP-binding protein cannot be excluded. Accordingly, it is concluded that the late PA formation induced by the toxin is at least indirectly regulated by GTP-binding protein. GTP[gS] dose dependently stimulated the late formation of PA and hemolysis induced by the toxin, and GDP[bS], a GTPbinding protein inhibitor, simultaneously inhibited both processes, indicating that the late PA formation and hemolysis induced by the toxin are regulated by GTP-binding protein. These observations suggest that the hemolysis is closely associated with the pathway of the PA formation, although it is not clear whether the late PA formation is essential to the process of hemolysis. We have reported that H126G, which is unable to bind to sheep erythrocytes in TB containing 0.9% NaCl, binds to erythrocytes and lyses the cells in the buffer with divalent cations such as Zn21 and Co21 (11). PLC activity of H126G is also observed in the presence of Co21, and the H148G mutant, which is able to bind to sheep erythrocytes, indicates no activity of the toxin (11). In the present work, the late PA formation and hemolysis induced by H126G were dependent on the concentration of Co21 added, and H148G elicited no activity of
TABLE 3. Effect of Co21 on hemolysis and PA formation induced by H126G or H148G % Hemolysis with indicated concn of CoCl2 (mM)a
Variant toxin (mg/ml)
H148G 0 0.5 1.0 H126G 0 0.5 1.0
PA (cpm) with indicated concn of CoCl2 (mM)b
0
10
100
0
10
100
0 0 0
0 0 0
0 0 0
759 6 90 860 6 74 822 6 86
763 6 69 890 6 90 891 6 95
698 6 77 908 6 84 932 6 92
0 0 0
0 38 6 3c 56 6 4c
0 NDd 78 6 8c
699 6 48 783 6 96 791 6 95
898 6 61 2,798 6 346c 3,642 6 655c
721 6 46 NDd 4,600 6 870c
a Rabbit erythrocytes were incubated with the variant toxins (0.5 and 1.0 mg of protein per ml) in the presence of 10 or 100 mM CoCl2 at 378C for 20 min as described in the text. Hemolytic activity was determined as described in the text. Values represent means 6 standard errors for four to five experiments. b Rabbit erythrocyte lysates were incubated with the variant toxins (0.5 and 1.0 mg of protein per ml) and 10 mCi of [g-32P]ATP per ml in the presence of 10 or 100 mM CoCl2 as described in the text. PA formation was determined as described in the text. Values represent means 6 standard errors for four to five experiments. c P , 0.01, compared with the value induced by the variant toxin in the absence of Co21. d Not determined.
VOL. 64, 1996
NOTES
wild-type toxin in the presence or absence of the cation. These observations strongly support the contention that the late PA formation is linked to hemolysis induced by the toxin. Thus, it is concluded that the toxin-induced hemolysis is linked to a PA formation pathway activated through PLC and/or PLD which is regulated by GTP-binding protein. We thank Keiko Kobayashi, Minako Okamoto, Masami Hiramatsu, Aki Hiraoka, and Kaori Matsuo for competent technical assistance. The research was supported in part by a grant from the Ministry of Education, Science, and Culture of Japan. REFERENCES 1. Agwu, D. E., L. C. McPhail, M. C. Chabot, L. W. Daniel, R. L. Wykle, and C. E. McCall. 1989. Choline-linked phosphoglycerides: a source of phosphatidic acid and diglycerides in stimulated neutrophils. J. Biol. Chem. 264: 1405–1413. 2. Bhamidipati, S. P., and J. A. Hamilton. 1993. NMR studies of phospholipase C hydrolysis of phosphatidylcholine in model membranes. J. Biol. Chem. 268:2431–2434. 3. Black, W. J., J. J. Munoz, M. G. Peacock, P. A. Schad, J. L. Cowell, J. J. Burchall, M. Lim, A. Kent, L. Steinman, and S. Falkow. 1988. ADP-ribosyltransferase activity of pertussis toxin and immunomodulation by Bordetella pertussis. Science 240:656–659. 4. Exton, J. H. 1994. Phosphatidylcholine breakdown and signal transduction. Biochim. Biophys. Acta 1212:26–42. 5. Fujii, Y., S. Nomura, Y. Oshita, and J. Sakurai. 1986. Excitatory effect of Clostridium perfringens alpha toxin on the rat isolated aorta. Br. J. Pharmacol. 88:531–539. 6. Fujii, Y., and J. Sakurai. 1989. Contraction of the rat isolated aorta caused
Editor: A. O’Brien
7.
8. 9.
10.
11.
12.
13. 14. 15.
16.
3933
by Clostridium perfringens alpha toxin (phospholipase C): evidence for the involvement of arachidonic acid metabolism. Br. J. Pharmacol. 97:119–124. Levine, L., D. M. Xiao, and C. Little. 1988. Increased arachidonic acid metabolites from cells in culture after treatment with the phosphatidylcholine-hydrolyzing phospholipase C from Bacillus cereus. Prostaglandins 34: 633–642. Lowry, O. H., N. J. Rosebrough, A. L. Farr, and R. J. Randall. 1951. Protein measurement with the Folin phenol reagent. J. Biol. Chem. 193:265–275. Meyers, D. J., and R. S. Berk. 1990. Characterization of phospholipase C from Pseudomonas aeruginosa as a potent inflammatory agent. Infect. Immun. 58:659–666. Nagahama, M., K. Michiue, and J. Sakurai. 1996. Membrane-damaging action of Clostridium perfringens alpha-toxin on phospholipid liposomes. Biochim. Biophys. Acta 1280:120–126. Nagahama, M., Y. Okagawa, T. Nakayama, E. Nishioka, and J. Sakurai. 1995. Site-directed mutagenesis of histidine residues in Clostridium perfringens alpha-toxin. J. Bacteriol. 177:1179–1185. Olson, S. C., E. P. Bowman, and J. D. Lambeth. 1991. Phospholipase D activation in a cell-free system from human neutrophils by phorbol 12myristate 13-acetate and guanosine 59-O-(3-thiotriphosphate): activation is calcium dependent and requires protein factors in both the plasma membrane and cytosol. J. Biol. Chem. 266:17236–17242. Sakurai, J. 1995. Toxins of Clostridium perfringens. Rev. Med. Microbiol. 6:175–185. Sakurai, J., Y. Fujii, and M. Shirotani. 1990. Contraction induced by Clostridium perfringens alpha toxin in the isolated rat ileum. Toxicon 28:411–418. Sakurai, J., S. Ochi, and H. Tanaka. 1993. Evidence for coupling of Clostridium perfringens alpha-toxin-induced hemolysis to stimulated phosphatidic acid formation in rabbit erythrocytes. Infect. Immun. 61:3711–3718. Sakurai, J., S. Ochi, and H. Tanaka. 1994. Regulation of Clostridium perfringens alpha-toxin-activated phospholipase C in rabbit erythrocyte membranes. Infect. Immun. 62:717–721.
