JOURNAL OF BACTERIOLOGYVol. 87, No. 1, pp. 136-144 January, 1964Copyright © 1964 by the American Society for Microbiology

Printed in U.S.A.

HEMOLYSIS OF RABBIT ERYTHROCYTES BY PURIFIEDSTAPHYLOCOCCAL ALPHA-TOXIN

II. EFFECTS OF INHIBITORS ON THE HEMOLYTIC SEQUENCE

Louis Z. COOPER, MORTON A. MADOFF, AND Louis WEINSTEIN

Infectious Disease Service, Pratt Clinic, New England Center Hospital, and Department of Medicine,Tufts University School of Medicine, Boston, Massachusetts

Received for publication 29 August 1963

ABSTRACT

COOPER, Louis Z. (New England Center Hos-pital, Boston, Mass.), MORTON A. MADOFF, ANDLouis WEINSTEIN. Hemolysis of rabbit erythro-cytes by purified staphylococcal alpha-toxin. II.Effect of inhibitors on the hemolytic sequence. J.Bacteriol. 87:136-144. 1964.-Study of the timecourse of hemolysis of rabbit erythrocytes bypurified staphylococcal alpha-lysin revealed thatthe specific toxin-red cell reaction occurs duringthe prelytic period. This reaction could be pre-vented or decreased by alpha-lysin antitoxinadded early, but not by antitoxin added at theend of the prelytic phase or at any time thereafter.In contrast, hemolysis is suppressed temporarilyby sucrose and permanently by polyethyleneglycol, even when these are added during theperiod of rapid release of hemoglobin. Whensucrose is present together with alpha-lysin andred cells only during the prelytic period, and whenthe cells are then washed and resuspended inphosphate-buffered saline, their subsequenthemolysis is not altered by the presence of thesugar. This is not so when antitoxin is employed.When erythrocytes are laked by a measured excessof alpha-lysin, only a portion of the originalhemolytic activity can be recovered. Repeatedexposure of lysin to red cells produces a loss ofactivity represented by a linear function whenlogs of residual activity are plotted sequentially.Once alpha-lysin has reacted with red cells, itdoes not appear to be available for attachment toother erythrocytes.

Kinetic studies of hemolysis by purifiedstaphylococcal alpha-lysin demonstrated thathemolytic activity is directly related to the dura-tion of incubation and the concentration oftoxin, and is inversely proportional to the logconcentration of erythrocytes. Analysis of thetime course of hemolysis revealed a sigmoidcurve characterized by a prelytic lag phase anda period of rapid, linear release of hemoglobin

tailing off asymptotically. The duration of lag;was inversely proportional to alpha-lysin con-centration; the rate of hemoglobin release wasdirectly proportional to the quantity of toxinand red cells. The optimal rate of hemolysisoccurred between 34 and 42 C. The lysin wasunstable to heat and dilution (Cooper, Madoff,and Weinstein, 1964). The shape of the curve ofthe time course of hemolysis observed in thesestudies resembled that of certain other bacterialtoxins (Herbert and Todd, 1941; Bernheimer,1947).Bernheimer (1947) demonstrated and defined

several stages in the hemolytic process by study-ing the effects of specific antibody and sucrose onthe time course of lysis of rabbit erythrocytesexposed to the hemolysin of Clostridium septicum."Split titration" studies of staphylococcalalpha-lysin were carried out by Lominski andArbuthnott (1962). They retitrated a partiallypurified toxin preparation as many as five times,and noted no decrease in activity even when sixsamples of erythrocytes were used. The resultsof their investigation confirm those of Forssman(1939), but not those of Levine (1938), whoreported irreversible adsorption of the Freund-lich isothermic type. The purpose of the presentstudy was to define the specific steps in the reac-tion between staphylococcal alpha-toxin andrabbit erythrocytes by use of specific antiserumand osmotic stabilizers, and to investigatewhether purified alpha-toxin is used up in thisprocess.

MATERIALS AND METHODS

Purified staphylococcal alpha-hemolysin wasprepared by the method of Madoff and Wein-stein (1962). Rabbit erythrocytes were collecteddaily from the central artery of the ear, placedin Alsever's solution, washed twice in 0.155 M

136

on June 26, 2018 by guesthttp://jb.asm

.org/D

ownloaded from

VOL. 87, 1964 EFFECT OF INHIBITORS ON HEMOLYSIS BY ALPHA-LYSIN

NaCl, and suspended in phosphate-bufferedsaline (PBS) containing 0.145 M NaCl and 0.01M phosphate (pH 6.9).The methods for measuring alpha-lysin ac-

tivity, degree of hemoglobin release, 50%hemolysis, and the procedure for studying thetime course of erythrocyte lysis, as well as thehemolytic unit (HU), were described previously(Cooper et al., 1964).Bovine serum albumin (bovine albumin

powder, Cohn Fraction V, C-grade, Calbiochem)was added to PBS in a concentration of 1.0g/100 ml in some phases of these studies tostabilize dilute purified alpha-lysin. A 1.5 Msolution of sucrose was prepared in PBS. Poly-ethylene glycol (PEG) was made by addingone part of Carbowax 4000 (Union CarbideChemical Co., New York, N.Y.) to seven partsof PEG (molecular weight 300). This materialwas solid at room temperature; after meltingin a boiling-water bath, it remained liquid at37C.

Alpha-lysin antiserum (horse) was obtainedfrom the Wellcome Research Laboratories,Beckenham, England, and contained 900 unitsof antialpha-hemolysin per ml. This antitoxinwas heated to 56 C for 30 min prior to use, andwas appropriately diluted in PBS.

RESULTS

Effect on the hemolytic reaction of removal ofalpha-lysin during prelysis. Purified alpha-lysin(titer 300 HU) was diluted 1:80 in PBS, and20-ml portions were placed in each of two flasks(A and B); 20 ml of PBS were added to a thirdflask (E), and the three flasks were warmed to37 C in a reciprocal shaker bath. To each flaskwere added 20 ml of a prewarmed 4% rabbiterythrocyte suspension (RRBC). Incubationwas allowed to proceed for exactly 3 min. Thesuspensions were poured into chilled centrifugecups, and were centrifuged immediately at1,000 X g at 4 C for 2min.The supernatant fluid from flask A was pi-

petted quantitatively into a fresh flask (C),and 0.8 ml of packed RRBC was added. Simul-taneously, the alpha-lysin-exposed "button"of RRBC from flask A was resuspended to 40ml in PBS. The "button" from flask B wasresuspended in its own alpha-lysin-containingsupernatant fluid and transferred to flask B'."Button" E was also resuspended in its own PBS

60.

50- 0FLASKA'e50-',@ o-FLASK B' CONTROL40- t5/. *-FLASK D CONTROL30 * ,' A-FLASK C

20-

10-

2 4 6 8 10 12 14 16 IS 20 22 24 26 28MINUTES

FIG. 1. Effect of removing erythrocytes fromalpha-lysin after $ min of incubation at 37 C. FlaskA' contained erythrocytes resuspended in PBS. FlaskB' contained erythrocytes resuspended in originalalpha-lysin-containing solution. Flask C containeduntreated erythrocytes to which alpha-lysin fromflask A was added. Flask D contained erythrocytesto which alpha-lysin was added at the onset of thesecond incubation period.

and transferred to flask E'. At this time, anadditional control (D), consisting of 20 ml ofdiluted alpha-lysin and 20 ml of the 4% RRBCsuspension, was prepared.

All flasks (A', B', C, D, E') were promptlyreturned to the 37-C shaker bath, and 2.0-mlsamples were removed from each flask at specificintervals for determination of the per cent hemo-lysis. The results of these studies (Fig. 1) indi-cated that: (i) RRBC exposed to alpha-lysinduring only a portion of the prelytic lag period,then resuspended in buffered saline (A'), stillproceeded to hemolyze, although to a lesserextent than did cells which remained in PBScontaining alpha-lysin (B'); and (ii) that thesupernatant fluid from toxin incubated withRRBC for 3 min of the lag period containedsufficient residual alpha-lysin to hemolyze freshred cells (C), although more slowly and to alesser extent than did a control (D), which con-tained hemolysin not previously exposed to redcells. It was not possible to conclude from thisstudy whether the decreased hemolytic activitynoted when fresh erythrocytes were added toalpha-lysin previously exposed to RRBC repre-sented irreversible utilization of toxin in the firstred-cell exposure or simple inactivation of dilutealpha-lysin by heat.

Effect of the addition of antitoxin during variousphases of the hemolytic reaction. To each of fiverows of test tubes were added 1-ml samples of

137

on June 26, 2018 by guesthttp://jb.asm

.org/D

ownloaded from

COOPER, MADOFF, AND WEINSTEIN J. BACTERIOL.

purified alpha-lysin (titer 320 HU) diluted 1:100in PBS. After being prewarmed to 37 C in theshaker bath, 0.75-ml portions of a 4% suspensionof RRBC, also heated to 37 C, were added toeach tube at time 0. At appropriate intervalsthereafter, a tube from each row was removed,and the degree of hemolysis was determined as

described previously (Cooper et al., 1964). A fewseconds after 1 min, 0.25 ml of horse-serum anti-toxin, diluted 1 :1,000 in PBS, was added to eachremaining tube in the first row. The same quan-tity of antitoxin was added to each remainingtube in the second, third, and fourth rows after2, 3, and 5 min, respectively. The fifth row servedas a control.A hemolysis curve of the expected shape with

a 6-min prelytic lag phase, followed by a linearperiod of rapid lysis until maximal hemolysiswas approached, was observed in the control(Fig. 2). In sharp contrast, the tubes to whichantitoxin was added after 1 min of incubationshowed little hemolysis; those in which antibodyhad been put after 2 min, and later, exhibitedan increasing degree of lysis until the rate andextent of hemolysis in the control row was ap-proached ("5-min row"). Similar studies demon-strated that antitoxin added at time 0 completelyinhibited the reaction. When it was added at or

after the onset of the period of rapid lysis, it didnot interfere with hemolysis. These data indicatethat the hemolytic action of alpha-lysin occurs

and is essentially completed during the lag periodbefore hemoglobin is released from cells.

80.

70-

0 60

50.

vg 40-

t 30

20O

101

x-CONTROL*-ANTITOXIN F I MINUTE

A-ANTITOXIN p 2 MINUTES_ ,,II _' 3 is°-

is DB 5Iap, - ,

0~

4.-

,,,,°'Y1 0

a a Ae~~. . I I a-I..t tt 1o 15 MINUTESLA. o

FIG. 2. Effect of addition of alpha-lysin anti-serum during various phases of the hemolytic reac-

tion.

90.

80-

70.

c,, 60-%

50.

x 40-

30-

20-

10-

*-SUCROSE ADDED I MIN. 0-A - SUCROSE ADDED P 3 MIN.A " 5"o " "P7'6j*-CONTROL

^ ,D O;--- *.I *'

5.

10 MINUTES 15

FIG. 3. Effect of addition of sucrose during vari-ous phases of the hemolytic reaction.

Effects of sucrose or polyethylene glycol on hemo-lysis. A rate study was carried out as describedabove. Sucrose (1.5 M, 0.25 ml) was added toeach tube remaining in the first row after 1 minof incubation; similar quantities were added toeach remaining tube in the second, third, andfourth rows after 3, 5, and 7.5 min, respectively.In contrast to the previous study and others to bedescribed, the tubes were kept in an ice bathuntil the last ones were removed from the 37-Cshaker bath. They were all centrifuged simul-taneously at 4 C, and the per cent hemolysis wasdetermined (Fig. 3).The addition of sucrose caused prompt delay in

release of hemoglobin from erythrocytes treatedwith alpha-lysin. The dip in the curves afteraddition of the sucrose suggested that lysis con-

tinued in the ice bath prior to centrifugation.Data from studies in which 0.25-ml portions

of PEG were substituted for sucrose are presentedin Fig. 4. After 9 min of incubation, PEG wasadded to one row, and sucrose simultaneously toanother, without discernible difference in effecton hemolysis. In this experiment, all tubes were

chilled in an ice bath for only 15 sec prior tocentrifugation at 4 C; the dips after addition ofPEG or sucrose were much more shallow thanthose illustrated in Fig. 3.

This study demonstrated that PEG was as

effective as sucrose. In fact, it completely pre-vented release of hemoglobin throughout theentire period of observation.

Comparison of the curves obtained when tubeswere held in an ice bath until all were removedfrom the 37-C bath and then centrifuged at 4 C,with those observed when they were plunged intoan ice bath for just 15 sec and then centrifugedin the cold immediately, indicates that signifi-

138

on June 26, 2018 by guesthttp://jb.asm

.org/D

ownloaded from

VOL. 87, 1964 EFFECT OF INHIBITORS ON HEMOLYSIS BY ALPHA-LYSIN

X70 gtt__°n_°_O740i60

~50 0

40 Ii730

20

10 i2 42 4 6 89s 11 15 2t a

I MINUTES

A

FIG. 4. Comparison of the effects of addition ofsucrose or polyethylene glycol (PEG) during theperiod of rapid hemolysis.

cant hemolysis occurs while tubes are in the icebath if incubation at 37 C has been of sufficientduration. This accounts for the dips in the hemo-globin-release curves after addition of sucrose orPEG. Since the rate of release of hemoglobin is sorapid during the linear period of hemolysis, eventhe 15- to 20-sec delay at the onset of centrifuga-tion allowed for measurable hemolysis (Fig. 5).

Effects of antiserum, sucrose, and PEG onhemolysis when added, then removed, during theprelytic lag period. This experiment was designedto examine the effect on erythrocytes of the addi-tion of sucrose, PEG, or antiserum at the onsetof incubation or early in the lag phase, and theeffect of removal of these substances by washingand resuspending the cells in buffered saline.

Four flasks (A, B, C, and D), containing 20 mlof purified alpha-lysin (titer 160 HU) diluted1:50, were placed in a shaker bath (37 C) for 1min. To each flask were then added 15 ml of aprewarmed 4% suspension of rabbit erythro-cytes. After 1 min, 5 ml of prewarmed PBS,PEG, sucrose, and antiserum were added toflasks A, B, C, and D, respectively. The contentsof each were poured 1 min later into chilled cupsand immediately centrifuged at 4 C for 2 min.The supernatant fluids were discarded and re-placed with cold PBS. The red-cell "buttons"were resuspended and then promptly centrifugedfor 2 min. The "buttons" were again suspendedin PBS, transferred to fresh flasks A, B, C, andD, and incubated at 37 C. At timed intervals,2-ml samples were removed from each flask, andper cent hemolysis was determined.

The results of this study indicated that sucrosedid not alter subsequent hemolysis significantlywhen it was present only during the prelyticphase (Fig. 6). Exposure to antiserum duringthe same period completely prevented the hemo-lytic reaction, however. Both toxin-treated anduntreated erythrocytes hemolyzed spontane-ously when washed and resuspended in PBS toremove PEG. This was thought to be due tomechanical injury to the cell membrane causedby rapid shrinkage and expansion. It preventedobservation of the effect of PEG on the prelyticphase. The addition of PEG, sucrose, or anti-serum to the appropriate flasks simultaneouslywith the erythrocytes, rather than 1 min laterin the early prelytic period, did not alter theresults of this study.

Effects on hemolytic activity of exposure of alpha-lysin to rabbit erythrocytes. Although the experi-ments described above demonstrated that suffi-cient alpha-lysin remained in solutions exposedbriefly to rabbit red cells to hemolyze additional

80-

70-

X 60.50

6 40-

30.20.10.

0

/f' -STAT SPIN

21-DELAYED/ t CENTRIFUGING

3 6 9 12 l5 18 21 MINUTES

FIG. 5. Effect on the release of hemoglobin, fromerythrocytes treated with alpha-lysin, of immediatecentrifugation at 4 C versus holding all tubes in anice bath until the last is ready for centrifugation.

80- *- BUFFER ONLY

70 o0- SUCROSEP 60 A -- ANTITOXIN 0

50-

30-

FIG. 6. Effects on hemolysis of alpha-lysin anti-serum and sucrose when added, then removed, duringthe prelytic period.

139

on June 26, 2018 by guesthttp://jb.asm

.org/D

ownloaded from

COOPER, MADOFF, AND WEINSTEIN

erythrocytes, they did not provide data of aquantitative nature. In this study, alpha-lysinactivity was determined in the usual fashion(Cooper et al., 1964). Hemolytic activity (HU)was measured after 30 min of incubation at 37 Cand considered "original activity." A 1-ml sampleof the supernatant fluid from the second tube,which was always completely hemolyzed, wasretitered with a fresh 2% suspension of rabbiterythrocytes; this process was repeated threetimes.Because Forssman (1939) suggested that a

finite time period is required before alpha-lysinis "released" from hemolyzed red cells, the "sec-ond tube" in each titration was allowed to incu-bate 15, 35, and 45 min, respectively, in each ofthree separate experiments. The results weresubstantially identical; therefore, data from onerepresentative study are presented in Table 1 and

TABLE 1. Results of "split titrations" of alpha-lysin

Dilution ofTitration Material titered original Activity Control

no. alpha-lysin (HU) (HU)in first tube

1 Original 1:5 4,1752 Second tube 1:20 1,850 3,680

of no. 13 Second tube 1:80 830 3,200

of no. 24 Second tube 1:320 400 2,614

of no. 3

4400-

4000

3600-

3200 -

2800 -

2400 -

2000

1600 -

1200.

800S

400.

1 2 3 4 1 2 3 4-TITRATION #

FIG. 7. Effects on hemolytic activity of repeatedre-exposure of alpha-lysin to erythrocytes, "splittitrations. "

Fig. 7. The studies were performed in duplicate;average activity is presented in the table. Alpha-lysin, serially diluted and incubated at 37 C, towhich no erythrocytes were added constitutedthe control. A striking loss of activity occurredwith each retitration; this was greatly in excessof that observed with the control toxin not previ-ously exposed to red cells. When plotted semi-logarithmically, the decrease in activity waslinear (Fig. 7). Activity loss by control toxinwas variable.Because of the possibility that loss of activity

was caused by inhibition of alpha-lysin by theproducts of hemolysis which accumulated witheach retitration, the "split titration" method wasmodified as follows. A portion of the 2% erythro-cyte suspension was mechanically hemolyzed byrepeated freeze-thawing. The red-cell ghosts wereremoved by centrifugation at 34,000 X g for 30min at 4 C. This solution, which contained allthe soluble products of mechanical laking oferythrocytes, was added to the control dilutionsof alpha-lysin in the same manner in whichwhole cells were added to the experimental group.Decreases in alpha-lysin activity paralleled thoseobserved when toxin lysed intact erythrocytes(Table 2, Fig. 8).The results of these studies indicated that a

loss of alpha-lysin activity occurred with eachexposure to erythrocytes; this phenomenon didnot appear to depend only on the toxin actinghemolytically; it was also influenced by thepresence of soluble materials produced duringmechanical disruption of erythrocytes.An attempt was made to determine whether

alpha-lysin, exposed to a sample of rabbit redcells, would hemolyze other erythrocytes addedlate in the prelytic lag period.To each of three flasks (A, B, and C), each

containing 20 ml of a 3% suspension of pre-warmed RRBC, were added 20 ml of dilute,purified alpha-lysin heated to 37 C. Bufferedsaline (20 ml) added to 3% red cells served as acontrol (D). The reaction mixtures were incu-bated in a 37-C shaker bath for 2 min, quicklypoured into chilled cups, and centrifuged for 2min at 4 C. The hemoglobin-free supernatantfluids were discarded and replaced with coldbuffered saline. The red-cell buttons were sus-pended in the buffer and then recentrifuged for 2min at 4 C. The supernatant fluids from thiswash were again discarded and replaced with 20ml of buffered saline, the cell buttons were re-

140 J . BACTERIOL .

on June 26, 2018 by guesthttp://jb.asm

.org/D

ownloaded from

VOL. 87, 1964 EFFECT OF INHIBITORS ON HEMOLYSIS BY ALPHA-LYSIN

suspended, and the contents of the cups werepoured into fresh flasks, A', B', C', and D',respectively. To flask A' was added an additional20 ml of buffered saline; to B', an additional 20ml of 3% RRBC; to C', 20 ml of a 6% RRBCsuspension; and, to D', 20 ml of buffer. The flaskswere returned to the 37-C shaker bath, and 2.0-mlportions were removed at 1-min intervals. Theoptical density of the supernatant fluid was deter-mined, and per cent hemolysis was calculated.The essentially identical hemoglobin-release

curves demonstrated that the additional erythro-

TABLE 2. Results of "split titrations" of staphylococ-cal alpha-lysin with addition of hemoglobin

supernatant fluid to control dilutions

Dilution ofTitration Material titered original Activity Control

no. Maeilttrdalpha-lysin (HU) (HU)in first tube

1 Original 1:5 31,0002 Second tube 1:20 11,500 13,900

of no. 13 Second tube 1:80 5,500 5,860

of no. 24 Second tube 1:320 2,100 2,700

of no. 35 Second tube 1:1280 1,050 1,700

of no. 4

100.000

10.000

1`-

I->

1.000

o

- EXPERIMENTAL RETITEREDA .1 a

0 CONTROL CONTAININGHgb SUPERNATE

0

0

1. 2. 3. 4. 5. 6.

TITRATION NO.

FIG. 8. Effects on hemolytic activity of "splittitrations" as in Fig. 7. In this study, the "control"received hemoglobin-containing supernatant fluid.

% HEMOLYSIS100/50/33'/4

70/35/23 /,

50/25/16%.

20/ 10/ 6/,* A A

2 4 6 8 10 12 14 I16 I 20 22 24 26 42 64MINUTES

FIG. 9. Effect on subsequent hemolysis of addinguntreated erythrocytes to erythrocytes exposed to,then removed from, alpha-lysin.

cytes did not lead to increased hemolysis (Fig. 9).Because it was shown previously (Cooper et al.,1964) that increase in red-cell concentrationincreases the rate of hemolysis even in verydilute toxin, the results of this experiment indi-cate that, once alpha-lysin has reacted witherythrocytes, it is no longer available to reactwith others. It is used up in the hemolytic reac-tion or so tightly bound to the red-cell leafletthat it is not recoverable under any of the con-ditions examined in this study.

DIsCUSSION

Menk (1932) was the first to point out thatthe hemolysis produced by C. septicum toxinconsisted of two phases: (i) combination of lysinwith erythrocytes, and (ii) lysis of cells. Bern-heimer (1947) studied the effects of sucrose andantitoxin on the time course of hemolysis. Thepresent experiments are essentially similar tothose of Bernheimer except that staphylococcalalpha-lysin was examined. This does not neces-sarily imply that these toxins are the same, orthat their mechanisms of action are identical.Neither staphylococcal alpha-lysin nor C. septi-cum hemolysin, which most closely resemblesalpha-lysin in the kinetics of the hemolytic reac-tion, have demonstrable lecithinase activity, acommon property of numerous clostridial toxins(McFarlane, 1955).The present studies demonstrate that alpha-

lysin exerts its effect in the prelytic period. Theearlier in this phase that antitoxin is added, thegreater is its effect in decreasing subsequenthemolysis; when added sufficiently late in theprelytic period, it is niot inhibitory. The shapesof a family of curves produced by addition of

141

g :s:.-.A.

0 A 6% RBC ADDED

9 BUFFER ONLY

A 3 % ROC ADDED

on June 26, 2018 by guesthttp://jb.asm

.org/D

ownloaded from

COOPER, MADOFF, AND WEINSTEIN

antitoxin after increasingly longer incubation inthe prelytic period were found to be similar tothose produced by increasing concentrations ofalpha-lysin (Cooper et al., 1964).However, when sucrose and alpha-lysin were

present only during the prelytic period, and whenboth were then removed, subsequent hemolysiswas not significantly altered. The continuedpresence of 0.3 M sucrose markedly delayedlysis, and addition of this compound, even duringthe period of rapid release of hemoglobin, ab-ruptly flattened the hemolysis curve, an effectquite distinct from that observed with antitoxin.PEG which, like sucrose, does not penetrate thered-cell membrane (Doebbler and Rinfret, 1962)also prevented further release of hemoglobinwhen it was added to toxin-erythrocyte mixtures.However, red cells exposed to this agent hemo-lyzed spontaneously when they were resuspendedin buffer free of PEG. This effect may be relatedto the "mechanical trauma" produced by therapid shrinking and subsequent swelling causedby the drastic osmotic changes. These results arecompatible with those presented in the previouspaper (Cooper et al., 1964), and confirm theobservation that the prelytic phase is integralto the hemolytic reaction, because duration ofthe lag period was inversely related to toxinconcentration and varied with changes in incuba-tion temperature; increase in temperature short-ened the lag until thermal inactivation becameprominent above 42 C.Hemolysis resulting from the action of such

varied agents as lipid solvents (Wilbrandt, 1941),X-radiation (Ting and Zirkle, 1943), C. septicumhemolysin (Bernheimer, 1947), antibody andcomplement (Green, Barrow, and Goldberg,1959), and sulfhydryl inhibitors (Jacob andJandl, 1962) occurs via a common pathwaycharacterized by increased permeability of thecell to cations, and swelling just prior to lysis-"colloid osmotic swelling" (Wilbrandt, 1941;Davson and Danielli, 1939; Ponder, 1948).Potassium ion release was demonstrated earlyin the prelytic period (Madoff, Cooper, and Wein-stein, 1964) and cell-swelling just prior to "ghost-ing" was observed with phase microscopy con-firmed by photomicrography (Cooper and Mitus,unpublished data).Forssman (1939) suggested that the action of

alpha-lysin is enzymatic because irreversibleadsorption is not necessary for hemolysis. On the

other hand, Levine (1938) contended that theadsorption of this toxin was not reversible, butconformed to the Freundlich adsorption iso-therm. Lominski and Arbuthnott (1962) recentlypresented evidence of total recovery of alpha-lysin after hemolysis. By use of a "split titra-tion" technique, they exposed alpha-lysin to sixconsecutive portions of red cells vwithout loss ofhemolytic activity. The present study, in whichhighly purified toxin, stabilized in bovine serumalbumin, was used, did not confirm these results.Activity was found to decrease in a linear fashionwhen plotted semilogarithmically. A similar lossof activity in controls exposed to the supernatantfluid of a suspension of mechanically lysed eryth-rocytes suggested inhibition by an unidentifiedsubstance. Decreases in activity in control titra-tions due to heating and dilution were much lessthan in those in which alpha-lysin mas exposedto erythrocytes. The difference between our dataand those of Lominski and Arbuthnott (1962)cannot be explained at present.

Additional evidence indicating the irreversiblenature of the toxin-cell reaction was the failureto demonstrate an increase in hemolysis whenfresh erythrocytes were added to cells washedfree of excess toxin.

Inability to recover alpha-lysin after hemolysisdoes not rule out the possibility that its actionis enzymatic. Previous studies (Cooper et al.,1964) indicated that there are many more siteson the red-cell leaflet at which the toxin canreact than are necessary for hemolysis. Leci-thinase activity was demonstrated in clostridialhemolysins (McFarlane, 1955), and Soule,Marinetti, and Morgan (1959) showed cleavageof bonds between sphingomyelin and protein,and degradation of the former in chicken erythro-cytes hemolyzed by mumps virus. However, therelationship between these effects and the mecha-nism of the hemolysis produced by this agenthas not been established. Studies of possiblephospholipase activity in purified alpha-lysinare in process in this laboratory. Magnusson,Doery, and Gulasekharam (1962) and Doeryet al. (1963) reported various phospholipaseactivities in culture filtrates of staphylococciproducing alpha-beta toxin.On the basis of the data presented in this paper,

the following scheme is proposed as representingthe steps involved in the lysis of erythrocytesexposed to alpha-toxin:

142 J. BACTERIOL.

on June 26, 2018 by guesthttp://jb.asm

.org/D

ownloaded from

VOL. 87, 1964 EFFECT OF INHIBITORS ON HEMOLYSIS BY ALPHA-LYSIN

Requires:

1. Alpha-lysin

Membrane integritycompromised:

2. Heat 1. K+ loss3. Time 2. Swelling of cell

Erythrocyte Altered erythrocyte

1. Inhibited by antitoxin2. Delayed by low temperature

Hemoglobin loss

Altered erythrocyte Erythrocyteghost

1. Inhibited by osmotic stabilizers2. Not inhibited by antitoxin3. Inhibited little by low temperature

In phase I, which corresponds to the prelyticperiod, alpha-lysin is adsorbed to the erythro-cyte, and, over a period of time, which is prob-ably the rate-limiting step, it reacts with a struc-tural component of the red-cell membrane causingloss of membrane integrity and "leakage" ofparticles of small molecular weight. This reactionis probably enzymatic in nature. It is quite sensi-tive to incubation temperature and the concen-

trations of the reactants, and is inhibited byspecific antibody. Phase microscopy demon-strated that the duration of this stage is variablefrom cell to cell; that is, prelytic swelling, whichoccurs at the end of this phase, begins at differ-ent times in individual erythrocytes. Althoughthe time interval before visible swelling may bemany minutes when dilute toxin is used, the"life" of the "swollen cell" is only a few seconds.Phase II, in which hemoglobin leaks from the

swollen erythrocyte, is also brief. It does notrequire the presence of alpha-lysin, is not in-hibited by antitoxin, and is little slowed at 4 C.This reaction can be delayed dramatically byraising the osmolarity of the suspending isotonicsolution with such materials as sucrose and PEG.This phase appears to be the final common

pathway for a wide variety of hemolytic agents,and is probably not enzymatic, but simplyleakage through a badly damaged plasma mem-

brane (colloid osmotic swelling).Strikingly similar changes were observed when

cells in tissue culture were exposed to purifiedalpha-lysin, with leakage of substances of lowmolecular weight preceding swelling of the cellsand release of cell proteins (Madoff, Artenstein,and Weinstein, 1963). The lethal action of thistoxin is inhibited by incubating it with partiallypurified phospholipids of neural origin (Northand Doery, 1961; Madoff and Dain, in prepara-tion). It is probable that the widespread action

of alpha-lysin represents a primary action oncell membranes.

ACKNOWLEDGMENTS

This investigation was supported in whole byU.S. Public Health Service research grantAI-02564 from the National Institute of Allergyand Infectious Diseases.

L. Z. Cooper was a Post-doctoral ResearchFellow and M. A. Madoff was a Research CareerDevelopment Awardee, National Institute ofAllergy and Infectious Diseases.We thank Madeline Zizza and John J. Mc-

Kenna for invaluable technical assistance.

LITERATURE CITED

BERNHEIMER, A. W. 1947. Comparative kineticsof hemolysis induced by bacterial and otherhemolysins. J. Gen. Physiol. 30:337-353.

COOPER, L. Z., M. A. MADOFF, AND L. WEINSTEIN.1964. Hemolysis of rabbit erythrocytes bypurified staphylococcal alpha-toxin. I. Ki-netics of the lytic reaction. J. Bacteriol. 87:127-135.

DAVSON, H., AND J. F. DANIELLI. 1938. CXXXIV.Studies on the permeability of erythrocytes.V. Factors in cation permeability. Biochem.J. 32:991-1001.

DOEBBLER, G. F., AND A. P. RINFRET. 1962. Theinfluence of protective compounds and coolingand warming conditions on hemolysis oferythrocytes by freezing and thawing. Bio-chim. Biophys. Acta 58:449-458.

DOERY, H. M., B. J. MAGNUSSON, I. M. CHEYNE,AND J. GULASEKHARAM. 1963. A phospholipasein staphylococcal toxin which hydrolysessphingomyelin. Nature 198:1091-1092.

FORSSMAN, J. 1939. Studies on staphylococci.XIV. The fixation of staphylolysin by bloodcorpuscles and the hemolytic action of staph-ylolysin. Acta Pathol. Microbiol. Scand. 16:335-345.

GREEN, H., P. BARROW, AND B. GOLDBERG. 1959.Effect of antibody and complement on perme-ability control in ascites tumor cells anderythrocytes. J. Exptl. Med. 110:699-713.

HERBERT, D., AND E. W. TODD. 1941. Purificationand properties of a hemolysin produced bygroup A hemolytic streptococci (streptolysin0). Biochem. J. 35:1124-1139.

JACOB, H. S., AND J. H. JANDL. 1962. Effects ofsulfhydryl inhibition on red blood cells. I.Mechanism of hemolysis. J. Clin. Invest.41:779-792.

143

on June 26, 2018 by guesthttp://jb.asm

.org/D

ownloaded from

COOPER, MADOFF, AND WEINSTEIN

LEVINE, B. S. 1938. Studies of staphylococcaltoxin. The toxin-red cell reaction. J. Immunol.35:131-139.

LOMINSKI, I., AND J. P. ARBUTHNOTT. 1962. Somecharacteristics of staphylococcus alpha he-molysin. J. Pathol. Bacteriol. 83:515-520.

MADOFF, M. A., M. S. ARTENSTEIN, AND L. Wein-STEIN. 1963. Studies of the biologic activityof purified staphylococcal alpha-toxin: II.The effect of alpha-toxin on Ehrlich ascitescarcinoma cells. Yale J. Biol. Med. 35:382-389.

MADOFF, M. A., L. Z. COOPER, AND L. WEINSTEIN.1964. Hemolysis of rabbit erythrocytes bypurified staphylococcal alpha-toxin. III. Po-tassium release. J. Bacteriol. 87:145-149.

MADOFF, M. A., AND L. WEINSTEIN. 1962. Purifica-tion of staphylococcal alpha-hemolysin. J.Bacteriol. 83:914-918.

McFARLANE, M. G. 1955. On the biochemicalmechanisms of action of gas-gangrene toxins.In J. W. Howie and A. J. O'Hea [ed.], Mecha-nisms of microbial pathogenicity. CambridgeUniversity Press, London.

MAGNUSSON, B. J., H. M. DOERY, AND J. GULASEK-HARAM. 1962. Phospholipase activity ofstaphylococcal toxin. Nature 196:270-271.

MENK, W. 1932. Zur wirkungdes pararauschbrandhamolysins. Zentr. Bakteriol. Abt. I. Orig.123:55-60.

NORTH, E. A., AND H. M. DOERY. 1961. Inactiva-tion of staphylococcal, tetanus and diphtheriatoxins by ganglioside. Brit. J. Exptl. Pathol.42:23-29.

PONDER, E. 1948. Hemolysis and related phenom-ena. Grune and Stratton, New York.

SOULE, D. W., G. V. MARINETTI, AND H. R. MOR-GAN. 1959. Studies of the hemolysis of redblood cells by mumps virus. IV. Quantitativestudy of changes in red blood cell lipids andvirus lipids. J. Exptl. Med. 110:93-102.

TING, T. P., AND R. E. ZIRKLE. 1943. The natureand cause of hemolysis produced by X-rays.J. Cellular Comp. Physiol. 22:233-249.

WILBRANDT, W. 1941. Osmotische natur sogenanternicht-osmotischer haemclysin. Kolloidosmo-tische hamolyse. Arch. Ges. Physiol. 245:22-52.

144 J . BACTERIOL.

on June 26, 2018 by guesthttp://jb.asm

.org/D

ownloaded from