INFECTION AND IMMUNITY, Dec. 1973, p. 938-946Copyright © 1973 American Society for Microbiology

Vol. 8, No. 6Printed in U.S.A.

Effects of Staphylococcal Alpha-, Beta-, Delta-,and Gamma-Hemolysins on Human Diploid

Fibroblasts and HeLa Cells: Evaluation of a NewQuantitative Assay for Measuring Cell Damage

MONICA THELESTAM, ROLAND MOLLBY, AND TORKEL WADSTROM

Department ofApplied Microbiology, Department of Bacteriology, Karolinska Institutet, S-104 01 Stockholm60, and Department of Bacteriology, Statens Bakteriologiska Laboratorium, S-105 21 Stockholm, Sweden

Received for publication 28 June 1973

Human diploid embryonic lung fibroblasts and HeLa cells were cultivated inEagle minimal essential medium supplemented with 10% calf serum. Monolayercultures were labeled with 3H-uridine and treated with highly purified staphylo-coccal alpha-, beta-, delta-, or gamma-hemolysin. The release of solubleradioactive substances into the medium was used as an indicator of damage tothe cell membrane after treatment with each hemolysin. The assay methoddescribed is simple, sensitive, and rapid. It allows quantitative estimation ofchanges in membrane permeability to be detected before a morphologicaldamage is observed microscopically. Upon incubation for up to 30 min withhighly purified staphylococcal hemolysins, only delta-hemolysin caused releaseof a significant amount of tritiated substances from fibroblasts. Such leakageoccurred immediately after addition of delta-lysin and was independent oftemperature. With minor exceptions, this was similar to the release of isotopesafter treatment of the cells with the nonionic detergent Triton X-100. Treatmentof fibroblasts with combinations of two or three of these toxins gave neithera synergistic nor an antagonistic effect. Evidence is presented which indicatesthat delta-hemolysin is the only important fibroblast damaging activity in crudepreparations of extracellular proteins of four strains of S. aureus, whereas HeLacells are susceptible also to purified alpha-toxin.

Staphylococcus aureus produces at least fourdifferent cytolytic toxins (31, 41) called alpha-,beta-, gamma-, and delta-hemolysins, since theerythrocyte was the first type of cell used forassaying the cytolytic activity. The hemolyticassay is still the simplest and most convenientone. The four hemolysins can be differentiatedon the basis of the sensitivity of erythrocytesfrom different animal species to each. Alpha-toxin is most active on rabbit erythrocytes,beta-toxin is most active on sheep erythrocytes,and gamma-toxin is most active on rabbit andhuman erythrocytes, whereas delta-toxin lyseserythrocytes from many different animal spe-cies (31, 41). Beta-toxin is a sphingomyelinase C(27), but the mechanism of the cell membraneinteraction of the three other toxins is unknown(31). In addition, it is still not known what typesof cells in animal hosts are the important targetfor the lethal action of these four toxins.

This investigation evaluates the use of humandiploid fibroblasts and HeLa cells in tissueculture for studies on cytotoxic effects of highlypurified staphylococcal hemolysins. Subjectiveobservations of morphological changes are onlyqualitative and do not reveal subtle effects onthe cell membrane. This report describes asystem for quantitative estimation of cell mem-brane damage by measuring leakage of tritiatedsubstances from the cytoplasm. The cell culturesystem used has been well standardized in aseries of investigations (19-23).

Since the criteria of purity of staphylococcaltoxins were not defined in most biological stud-ies, it is quite possible that the cytotoxic effectsdescribed by several authors might have beencaused by contaminants possessing broadersubstrate specificity than the toxins in question(1, 3, 12, 31, 34).

9:38

on February 23, 2021 by guest

http://iai.asm.org/

Dow

nloaded from

CYTOTOXICITY OF STAPHYLOCOCCAL HEMOLYSINS

MATERIALS AND METHODS

Materials. Staphylococcal alpha-toxin (alpha-hemolysin) was purified by isoelectric focusing (39)and Sephadex chromatography (R. Mollby and T.Wadstr6m, unpublished data) from a substrain ofWood 46 which produces delta-toxin but not entero-toxin B, beta-toxin or gamma-toxin. The Wood 46strain was kindly supplied by A. W. Bernheimer.

Beta-toxin (beta-hemolysin) was purified fromstrain R 1 as recently described (40) and by Biogelchromatography (31).

Purified delta-toxin (delta-hemolysin) was kindlysupplied by A. Kreger. It was also purified in our

laboratory according to Kreger et al. (17) from strainNewman, kindly supplied by G. Wiseman. This strainmainly produces delta-toxin, but also low quantitiesof alpha-toxin.

Gamma-toxin (gamma-hemolysin) was producedfrom strain Smith SR (30) and purified by ionexchange chromatography, Sephadex chromatogra-phy, and isoelectric focusing (31). In addition, thisstrain also produces alpha-toxin.

Crude culture supernatants of these strains are

referred to as crude alpha-, beta-, delta-, and gamma-toxins, respectively. The purified preparations were

devoid of protease, lipase, staphylokinase, hyaluron-ate lyase, enterotoxin B, leucocidin, nuclease, andglucosidase activities (29, 40). The purity of thesepreparations was also controlled by analytical acryl-amide electrophoresis and immunoelectrophoresis(40). 3H-uridine (specific activity, 20 Ci/mmol) was

purchased from the Radiochemical Centre, Amer-sham, Buckinghamshire, England; Aquasol T M Uni-versal L.S.C. cocktail was obtained from New Eng-land Nuclear Chemicals GmbH, Frankfurt, W. Ger-many. Polystyrene multi-dish disposable trays (FB-6-TC, 3 cm), from Linbro Chemical Co., Inc., NewHaven, Conn., were used as culture vessels.The components of Eagle minimal essential me-

dium (7) and Hank balanced salt solution as well as

calf serum were obtained from Flow LaboratoriesLtd., Irvine, Scotland. Soy bean lecithin (grade III-S)was purchased from Sigma Chemical Co., St. Louis,Mo. and antiserum against alpha-toxin (CPP68/36)was from Wellcome Research Laboratories, LangleyCourt, Beckenham, England. Triton X-100 (technicalgrade) was bought from Rohm and Haas Co., Phila-delphia, Pa. All chemicals used were of analyticalgrade unless otherwise stated.Methods: cultivation and maintenance of cell

cultures. Diploid fibroblasts were isolated fromhuman embryonic lung tissue and cultivated in Rouxbottles. The cultivation procedure has recently beendescribed (20-22). Cultures between the 5th and the10th passages after 15 to 30 cell divisions were used forthe toxicity tests. In addition, two cloned HeLa lineswere used. Cells were seeded at a density of 90,000cells per culture (13,000 cells/cm2) in Linbro trays andincubated with 3 ml of Eagle minimal essentialmedium supplemented with 10% calf serum, 4 mMglutamine, 1 mM sodium pyruvate, and penicillin(100 IU/ml) and streptomycin (100 Ag/ml). The cellswere incubated at 37 C in a humid atmosphere

containing 5% CO2. The medium was changed after 5to 7 days, and the cultures were used when completemonolayers had developed after another 2 to 7 days.

Toxicity testing. Monolayers were incubated for 2h at 37 C with medium containing 1 UCi of 3H-uridineper ml. After further incubation for 2 h with freshmedium lacking 3H-uridine, the monolayers werewashed with Hanks balanced salt solution three timesto remove extracellular radioactivity before toxin wasapplied. The toxins were diluted in Eagle mediumwithout calf serum. One milliliter of toxin solution perculture was added, and the cultures were incubatedfor 10 and 30 min. Control cultures were incubatedwith Eagle medium alone. The medium was thenremoved and centrifuged (1,000 x g for 10 min at4 C). One-tenth milliliter of the supernatant wastransferred to a scintillation vial containing 10 ml ofAquasol. Samples were counted in a Nuclear Chicagoliquid scintillator for 1 min. All tests were performedin duplicate or triplicate. The spontaneous release ofisotope in the control cultures never exceeded 3 to 6%and was deducted in the data presented.To test the effects on cells in suspension, monolay-

ers were removed from culture bottles by trypsiniza-tion (0.25% trypsin) after usual labeling and washingof the cells with Hanks balanced salt solution. Thetrypsinization itself caused certain cell damage with arelease of about 12% of the maximal release (seebelow). The cells were centrifuged at 1,000 x g for 10min at 4 C, resuspended in fresh 3H-free medium,counted, and distributed in test tubes (0.7 x 10'cells/tube). Samples of toxin were diluted, 1 ml wasthen added to each tube, and the cell suspensionswere thoroughly mixed and incubated at 37 C.

Standardization of toxic effects. To compareresults between experiments, maximal release of solu-ble intracellular radioactive substances was measuredby incubating six control cultures in each experimentwith 0.06 M sodium borate buffer (pH 7.8) at 37 C for15 min, followed by scraping with a rubber policemanto lyse the fibroblasts, a procedure which leaves thenuclei intact (25). Because sodium borate buffer doesnot lyse the plasma membrane of HeLa cells, thenonionic detergent Triton X-100 was used instead. Ata concentration of 0.25% (vol/vol), this detergentcauses lysis of the cytoplasmic membrane, leaving thenuclei intact (M. Thelestam, unpublished data).

All experimental results were then expressed aspercentage of the maximal release.

Standard deviation of the maximal release whenmeasured in six parallel samples was 4 3 to 7%. Thisclosely paralleled the standard deviation in number ofcells per culture, which was also determined for sixcultures in each experiment. The standard deviationwas considered acceptable in this biological assaysystem.Measurement of hemolytic activity. The toxin

was diluted in tris(hydroxymethyl)aminometh-ane-hydrochloride-buffered saline (pH 7.0) supple-mented with 1 mM MgCl2, and 1% (vol/vol) washederythrocytes in the same buffer were added (30, 39,40). The mixture was incubated at 37 C for 1 h,hemolytic titers were read after 2 h at 4 C, and thedilution of hemolysin which gave a 50% hemolysis was

VOL. 8, 1973 939

on February 23, 2021 by guest

http://iai.asm.org/

Dow

nloaded from

THELESTAM, MOLLBY, AND WADSTROM

determined. The inverted value of this dilution indi-cated the number of hemolytic units (HU) in theundiluted toxin. Alpha-, delta-, and gamma-toxinwere tested on rabbit and human erythrocytes. Beta-toxin was tested on sheep and human blood cells.

RESULTS

Effects of crude and purified toxins. Table 1shows that crude alpha-, beta-, and delta-toxinscaused a concentration-dependent release ofradioactive nucleotides from fibroblasts afterincubation for 10 min, whereas crude gamma-toxin did not. When the purified toxins were

tested, the results were quite different (Table2). Purified alpha- and beta-toxins did notrelease radioactive substances from the fibro-blasts. On the other hand, the effect of delta-toxin did not diminish upon purification. Itseems probable that the activities found withcrude alpha- and beta-toxins were due to con-

taminating delta-toxin.All four crude toxins caused release of radio-

active compounds from HeLa cells upon incu-bation for 30 min at 37 C (Table 3). Of thepurified preparations, both alpha- and delta-toxins caused isotope release, whereas purifiedbeta- and gamma-toxins were without effect on

these cells.

TABLE 1. Effect on fibroblasts of crude alpha-, beta-,delta-, and gamma-toxins in different concentrationsa

HU/mlPercentage

Toxin Sheep Rabbit Human of maximalred blood red blood red blood release

cells cells cells

Alpha 3.2 0.1 2232 1 7664 2 84

Beta 3 x 102 0.1 293 x 103 1 906x 103 2 91

Delta 0.1 0.1 201 1 58

2 2 92

Gamma 1.6 0.1 1

16 1 2

32 2 2

a 3H-uridine-labeled diploid fibroblasts were incu-bated for 10 min at 37 C with crude toxins diluted infresh medium at concentrations indicated. The mem-

brane damage was measured as the release of radioac-tive substances into the medium after toxin treat-ment. The observed values are expressed as percent-age of maximal release obtained by treating the cellswith sodium borate buffer.

TABLE 2. Effect on fibroblasts of purified alpha-,beta-, delta-, and gamma-toxinsa

HU/mlPercentage

Toxin Sheep Rabbit Human of maximalred blood red blood red blood release

cells cells cells

Alpha 100 0.2 8Beta > 105 < 0.1 0Delta 1 1 1 93Gamma 12 6 9

a 3H-uridine-labeled diploid fibroblasts were incu-bated for 30 min at 37 C with purified toxins dilutedin fresh medium at concentrations indicated. Theresults are expressed as in Table 1.

TABLE 3. Effect on HeLa cells of crude and purifiedalpha-, beta-, delta-, and gamma-toxinsa

HU/ml Percentage

Toxin Sheep Rabbit Human ofxmred blood red blood red blood maximal

cells cells cells release

CrudeAlpha 4 1 66Beta 600 1 30Delta 0.5 0.5 84Gamma 26 3.2 56

Purif'iedAlpha 8 0.1 73Beta > 105 < 0.2 0Delta 0.25 0.5 80Gamma 6.4 3.2 11

a 3H-uridine-labeled HeLa cells were incubated for30 min at 37 C with crude and purif'ied toxins dilutedin fresh medium at concentrations indicated. Mem-brane damage was measured as the release of radioac-tive substances into the medium after toxin treat-ment. The observed values are expressed as percent-age of maximal release obtained by treating the cellswith Triton X-100 (0.25%, vol/vol).

The effect of crude alpha-toxin on HeLa cellsmay still mainly have been due to contaminat-ing delta-toxin. In terms of hemolytic units(rabbit erythrocytes), a higher amount of puri-fied than crude alpha-toxin was required toobtain the same amount of isotope release. Theeffect of gamma-toxin was probably due tocontaminating alpha-toxin.

In Fig. 1, where the membrane-damagingeffect on fibroblasts of crude alpha-, beta-,delta-, gamma-toxins is plotted against thehemolytic activity upon human erythrocytes, itis seen that, except for gamma-toxin, the mem-brane effect was almost exclusively related tothe number of HU per milliliter. Becauseneither alpha- nor beta-hemolysin acts uponhuman blood cells in these concentrations (30),

940 INFECT. IMMUNITY

on February 23, 2021 by guest

http://iai.asm.org/

Dow

nloaded from

CYTOTOXICITY OF STAPHYLOCOCCAL HEMOLYSINS

._Ex

mE

%O-0

C

0.1 1 2Hemolytic units per ml (human RBC)

FMG. 1. Effect of crude alpha, beta, delta, andgamma toxins in different concentrations on diploidfibroblasts after incubation for 10 min at 37 C. Therelease of radioactive compounds into the medium,expressed as percentage of the maximal release, isplotted against the hemolytic units on human eryth-rocytes. For further details see Table 1. Alpha-toxin0, beta-toxin, 0, delta-toxin, 0, gamma-toxin, x.

the membrane effect was probably due to thecontaminating delta-toxin.Crude gamma-toxin also contains alpha-tox-

in, but neither of these two toxins evidentlydamaged the diploid fibroblast membrane aftera 30-min incubation period.

Inhibition of alpha- and delta-toxins. Ta-bles 4 and 5 give further proof for the hypothesisthat contaminating delta-toxin of the crudepreparations is responsible for the membraneeffect. Addition of 50 jAg of lecithin per ml to thecrude toxins diminished their membrane-damaging effect considerably. Lecithin specifi-cally inhibits delta-toxin (4, 14). Heating ofcrude alpha- or beta-toxin could not abolish itsmembrane-damaging effect (Table 5). This also

indicates that the membrane damage was notcaused by the heat-labile alpha- or beta-toxinactivities in these preparations, but rather bythe heat-stable delta-toxin (17). Furthermore,when anti-alpha serum was mixed with crudealpha-toxin, the membrane damage on fibro-blasts was not reduced. However, anti-alphaserum completely abolished the membrane ef-fect of purified alpha-toxin on HeLa cells.Prolonged incubation with pure alpha-,

beta-, and gamma-toxins. Table 6 shows howthese three purified toxins affected fibroblastswith prolonged incubation for up to 22 h. Beta-and gamma-toxins did not cause any significantrelease after 22 h of incubation, whereas alpha-toxin produced an increasing release of radioac-tive substances. Prolonged incubation was notfurther studied because released material wasevidently reincorporated due to the cell metabo-lism.

TABLE 5. Effect of heating on the fibroblastmembrane-damaging activity of crude alpha-, beta-,

and delta-toxinsa

HU/ml

Sheep Rabbit Human PercentageToxin red red red of maximal

blood blood blood releasecells cells cells

Alpha (control) 32 0.5 33Heated alpha 1 0.5 35Beta (control) 103 0.5 51Heated beta 103 0.5 70Delta (control) 1 1 93Heated delta 1 1 90

a 3H-uridine-labeled diploid fibroblasts were incu-bated for 30 min at 37 C with heated (60 C, 10 min)and unheated crude alpha-, beta-, and delta-toxinsdiluted in Eagle medium. The results are expressed asin Table 1.

TABLE 4. Inhibition of the membrane-damaging activity of crude alpha-, beta-, and delta-toxins by soy beanlecithina

HU/ml Percentage of maximal release

Toxin Sheep red blood Rabbit red blood Human red Fibroblasts HeLa cellscells cells blood cells Fbolss HL el

Alpha (control) 32 0.5 33 63Alpha + lecithin 32 0.05 5 34

Beta (control) 103 0.5 51 31Beta + lecithin 103 0.05 10 13

Delta (control) 1 1 93 NDhDelta + lecithin 0.05 0.05 8 ND

a 3H-uridine-labeled cells were incubated for 30 min at 37 C with crude toxin or mixtures of crude toxin andsoy bean lecithin suspension (50 Ag/ml) diluted in Eagle medium. The results are expressed as in Table 1.

& ND, not done.

VOL. 8, 1973 941

on February 23, 2021 by guest

http://iai.asm.org/

Dow

nloaded from

THELESTAM, MOLLBY, AND WADSTROM

TABLE 6. Effect on fibroblasts of purified alpha-,beta-, and gamma-toxins with prolonged incubation

timea

Alpha- Beta- Gamma-Incubation Control toxin toxin toxintime (h) (10 HU), (100 HU)c (6 HU),

0.5 9,700 23,186 10,243 19,6711 12,071 37,743 12,271 18,9572 15,871 53,586 14,286 21,3433 19,329 73,286 18,786 18,2007 20,314 67,214 17,143 18,557

22 33,786 90,071 42,771 41,357

a 3H-uridine-labeled diploid fibroblasts were incu-bated at 37 C for different time periods with purifiedtoxins diluted in fresh medium at concentrationsindicated. As control, fibroblast cultures were incu-bated in fresh medium without added toxin. Thenumber of counts per minute released per 106 cells ispresented.

b On rabbit erythrocytes.c On sheep erythrocytes.

Effect of beta-toxin in different systems.Beta-toxin shows a typical "hot-cold" effect onsheep erythrocytes; i.e., the cells are not lysedafter incubation at 37 C but require a furtherincubation at 4 C. A similar experiment wasperformed with fibroblasts. After incubation for30 min at 37 C, the cells were stored at 4 C for90 and 180 min. However, not even the highestconcentration of purified toxin applied (105HU/ml) caused any significant release of radio-active nucleotides.To make sure that the resistance of fibro-

blasts was not due to the arrangement of thecells in a dense monolayer, the effect of beta-toxin was also tested on fibroblasts in a suspen-sion. Treatment with purified beta-toxin gaveno additional release of radioactivity as com-pared with the control.Kinetic studies on the effect of delta-toxin.

Purified delta-toxin seems to have an instan-taneous effect on the fibroblast membrane (Fig.2). Within 2 min after incubation with 0.5 to 1HU (human erythrocytes) at 37 C, 70 to 80%of the maximal release of nucleotides occurred.No obvious visible cell damage was observedmicroscopically with these concentrations until75 to 90% of the maximal release of radioactivecompounds had occurred. The cells then be-came granulated and decreased in size butwere not detached from the plastic surface ex-cept in cultures to which higher concentrationsof toxin (1 to 2 HU) were added. With lowerconcentrations, the release never exceeded 50to 60% of the maximal release. In these cases,morphological changes were not observed untilafter a 30-min incubation period (Fig. 2, lowerarrow).

The effect of different concentrations of puri-fied delta-hemolysin on fibroblasts was testedat 4, 13, and 37 C for 30 min. The effect ofdelta-toxin at 4 C was almost the same as at37 C (Table 7). The cytolytic effect of delta-toxin was thus shown to be very rapid, with nonoticeable lag phase, and largely independent ofthe temperature. A minimal critical concentra-tion was not observed with delta-toxin, as in thecase of the detergent Triton X-100 (Fig. 3).Combination of toxins. Purified alpha-, be-

ta-, and delta-toxins were mixed in the combi-nations shown in Table 8 and incubated withdiploid fibroblasts at 37 C. The observed releaseof radioactive material seemed -o be relatedonly to the presence of delta-toxin. Purifiedalpha- and beta-toxins together did not showany membrane-damaging effect. These resultssupport the proposition that the activities of the

X 100

-rE

.E

so

-0

u

=

2 4 6 10 30Minutes

FIG. 2. Effect on diploid fibroblasts of purifieddelta-toxin in different concentrations in relation toincubation time. The release of radioactive com-pounds into the medium is expressed as percentage ofthe maximal release obtained by treating the cellswith sodium borate buffer. Incubation was performedat 37 C. The following concentrations of purifieddelta-toxin were used: I HU (human erythrocytes)(0); 0.5 HU (O); 0.3 HU (U); 0.2 HU (0); 0.1 HU(A). Arrows indicate when morphological changes, asseen in the light microscope, first appeared. Thesechanges were considered typical for delta-toxin. In-serted figure shows the corresponding effect of thesurface active agent Triton X-100 (0.05 %, vol/vol).

TABLE 7. Effect on fibroblasts ofpurified delta-toxinat different temperaturesa

Concn (HU/ml, Percentage of maximal release athuman red blood

cells) 4 C 13 C 37 C

0.25 63 62 630.5 64 78 701.0 60 80 81

a Incubation time was 30 min; for further details seeTable 1.

942 INFECT. IMMUNITY

I

on February 23, 2021 by guest

http://iai.asm.org/

Dow

nloaded from

CYTOTOXICITY OF STAPHYLOCOCCAL HEMOLYSINS

I s

I-t

100 TrltonX-100

so XOM Q0per cent Wv/O) Triton

a I A II.

0.1 02 0.3 O 0.5Hwmolytic uits per ml (human RBC)

FIG. 3. Dose-response curve of the effect on diploidfibroblasts of purified delta-toxin and Triton X-100(inserted). Incubation was at 37 C for 30 min; forfurther details, see Fig. 2. Arrows indicate whenmorphological changes, as seen in the light micro-scdpe, first appeared. These changes were consideredtypical for delta-toxin.

TABLE 8. Effect on fibroblasts of combinations ofpurified toxinsa

HUPercentage

Toxin Sheep Rabbit Human ofred red red maximal

blood blood blood releasecells cells cells

Alpha 100 0Beta > 105 0Delta 0.25 41Delta 0.5 53Alpha + beta > 10 100 2Alpha + delta 100 0.25 51Alpha + delta 100 0.5 66Beta + delta > 105 0.25 56Beta + delta > 105 0.5 62Alpha + beta + > 105 100 0.5 61

delta

a Incubation time was 30 min; for further details seeTable 1.

crude toxins shown in Table 1 were caused bycontaminating delta-toxin.

DISCUSSIONTissue culture techniques have been used for

many years to study the effects of bacterialprotein toxins (37). Most of the earlier studiesdealt with the effects of diphtheria toxin (18) onmammalian cells cultivated by differentmethods (5, 6), whereas only a few studies havebeen devoted to other protein toxins, such asenterotoxins (35), neurotoxins (28), and a fewmembrane active toxins, such as streptolysins 0

and S and staphylococcal alpha-toxin (37). Adetailed study of human diploid fibroblasts byHayflick and Moorhead (11) stimulated re-search with these cells. A simplified standard-ized cultivation technique has been developed(20-23), and strictly defined conditions havebeen used for biological testing of antibiotics(19) and carcinogens (24; M. Thelestam and J.Litwin, unpublished data).The method described in this paper offers

several advantages over conventional methodsfor measuring cytotoxic effects, such as observa-tion of morphological changes under the micro-scope, determination of total cell protein as ameasure of cell growth (8), measure of acidproduction by color change of an indicator inthe medium, or measure of uptake of vitalstains such as trypan blue or neutral red (37,43). These procedures are all less suitable forquantitative measurement of cell damage.However, release of enzymes, such as aldolaseand beta-glucuronidase, from cells damaged bystaphylococcal delta-toxin has been studied,and a quantitative assay should be possible bythis method (9).A method similar to the one described in this

paper was developed for the study of the cyto-toxic effect of alpha-toxin by labeling cells with35S-methionine (26). Recently, Hallander andBengtsson (10) labeled cells with 32P to measurecell damage with crude and partially purifiedstaphylococcal toxins. Release of 5'Cr fromlabeled target cells has been used in immuno-logical investigations (42).The main advantages of a method based on

the release of radioactive metabolites from thecells are: (i) high sensitivity; (ii) quantificationof the cell damaging effect; (iii) suitability forstudies of cytoplasmic membrane damage with-out visible morphological changes; and (iv) highreproducibility and simplicity for studies ofmany cultures simultaneously.The well-standardized conditions for the out-

growth of fibroblasts to monolayers yieldingapproximately 0.7 x 106 cells per culture makethis method highly reproducible. In contrast,HeLa cells are less well suited for this type ofassay, since they do not grow to completemonolayers but tend to grow more irregularlydue to lack of contact inhibition. However, thisand other heteroploid cell lines have mostlybeen used for cytotoxicity studies, since they areeasy to grow and readily available (37).

In this assay system, the cells are incubatedwith the toxin dilution in fresh culture mediumwithout serum. This is important, since delta-toxin is partially inhibitied by serum and sincesera from different animal species might inter-fere with the cytotoxic assay as reported for

VOL. 8, 1973 943

on February 23, 2021 by guest

http://iai.asm.org/

Dow

nloaded from

THELESTAM, MOLLBY, AND WADSTROM

staphylococcal enterotoxin B (35). Addition ofserum is unnecessary in our system because ofthe short incubation time. Furthermore, thecells have not been in contact with proteolyticenzymes for at least 7 days when used in thisassay system. This is advantageous since pro-teases may alter the membrane architecture,e.g., so as to resemble malignant cells in thebehavior towards plant lectins (32, 38) or so asto respond differently to staphylococcal entero-toxin B (36).Crude alpha- and beta-toxins caused a rapid

release of nucleotides, which was probablycaused by contaminating delta-toxin. Thishypothesis is supported by the following facts:(i) highly purified alpha-toxin was without ef-fect; (ii) the release was closely correlatedto the hemolytic activity on human erythro-cytes; (iii) the release caused by the crudetoxins was abolished by preincubation withlecithin; (iv) the release was not affected byheating at 60 C for 10 min; and (v) the releasecaused by crude alpha-toxin was not affected bypreincubation with antiserum against alpha-toxin.On the other hand, it is important to notice

that highly purified alpha-toxin caused a cer-tain membrane damage on fibroblasts uponincubation for longer periods of time. However,this effect progressed slowly, causing only a

slight swelling of the cells after 3 h and nofurther changes in morphology during the fol-lowing 22-h period. It is thus probable thatneither alpha-, beta-, nor gamma-toxin is ableto interact with the human fibroblast mem-brane on short-time incubation, and only alpha-toxin interacts when incubated for longer pe-riods of time. Moreover, purified alpha-toxincaused a release of nucleotides from HeLa cellsupon incubation for 30 min, whereas purifiedbeta- and gamma-toxins did not.

Addition of a preparation of partially purifiedalpha-toxin to human amnion, rabbit kidney, orHEP-2 cells was reported to cause morphologi-cal changes both in the cytoplasm and nucleus(2, 15). Release of 35S-methionine from rabbitkidney and Ehrlich ascites cells revealed thatthe action was rapid and reached a maximumafter about 5 min (26). This was very similar tothe action of delta-toxin on fibroblasts (or HeLacells) as shown in Fig. 2. More recently, Hal-lander and Bengtsson (10) showed that partiallypurified alpha-toxin was toxic for calf kidneycells but not for human or monkey kidney cellsas measured by release of 32P from the cells. Onthe other hand, delta-toxin had a more pro-nounced effect on the human cells, but with amuch slower release of the isotope than shownin Fig. 2.

It is not evident in the above mentioned orany of the other studies on cytotoxicity ofstaphylococcal alpha-toxin, recently reviewedby Jeljaszewicz (12), how pure these prepara-tions were, and they might in several cases havebeen contaminated with delta-toxin and entero-toxin B. It is thus probable that the substratespecificity of alpha-toxin is much more limitedthan previously believed. Another explanationof the reported cytotoxic effects of alpha-toxincould be that the membrane structures of differ-ent mammalian cells are different, as indicatedby the effect of alpha-toxin on HeLa cells in thisstudy. However, since purified alpha-toxin isoften contaminated with delta-toxin, the firstexplanation seems more probable (4, 29, 33).Korbecki and Jeljaszewicz (16) showed that

crude and partially purified beta-toxin (sphin-gomyelinase C) was cytotoxic for KB cells. Thiswas also confirmed by Wiseman (43) on KBcells and cell lines of different origins.Wadstrom and Mollby (41) showed that purifiedbeta-toxin was cytotoxic for HeLa cells, humanembryonic muscle fibroblasts, and humanthrombocytes, whereas corresponding concen-trations of beta-toxin were not toxic to humanembryonic lung fibroblasts or HeLa cells in thisstudy. This again points to the fact that differ-ent mammalian cells might show quite differentsensitivities to these toxins. However, also inthe case of beta-toxin, varying quantities ofcontaminating delta-toxin might have causedsome of the conflicting results reported earlier.The low specific hemolytic activity of delta-

toxin, the kinetics of its hemolytic action, itswide spectrum of action on different artificialand biological membranes, and its inhibition byvarious phospholipids, as well as the kineticsfound in this study (rapid and temperature-independent release), indicate that it is a sur-factant polypeptide (13, 17, 33), as first pointedout by Bernheimer (3), and that it is probablyunique among bacterial proteins.

It is important to use highly purified toxinsfor biological testings in order to be able tostudy the spectrum on different cell types fromdifferent animal hosts. This is quite obviouswhen, for example, the different effects of crudeor semipurified alpha-toxin on muscle cells,fibroblasts, thrombocytes, and other cells arecompared with the findings reported in thisinvestigation. It also means that most of theprevious effects attributed to alpha-toxinshould be re-investigated with purified toxindevoid of delta-toxin and other contaminants(12, 31, 33). Also, the contradictory reports onthe toxicity of beta-toxin might be due to thepresence of a contaminant in purified toxinpreparations.

944 INFECT. IMMUNITY

on February 23, 2021 by guest

http://iai.asm.org/

Dow

nloaded from

CYTOTOXICITY OF STAPHYLOCOCCAL HEMOLYSINS

It is important to point out that in spite of thefact that alpha-, beta-, and gamma-toxins may

have a more limited toxic effect on cell culturesthan previously believed, these toxins are stilllethal for different animals in 1- to 100-.ugquantities. The possible role of specific targetcells or organs for these different lethal toxins isstill completely unknown.

ACKNOWLEDGMENTSWe are very grateful to J. Litwin and C. J. Smyth for

inspiring and helpful ideas. We also thank A. W. Bernheimer,J. P. Arbuthnott. and J. J. Rahal for stimulating criticism.The skillful technical assistance of G. Blomquist. B. Forn-stedt, M. Kjellgren, I. Kuhn, and L. Norenius is herebygratefully acknowledged.

This work was supported by the Swedish Board forTechnical Development (grant no. 69-577/UT441) and theSwedish Medical Research Council (grant no. 16X-2562).

LITERATURE CITED1. Arbuthnott, J. P. 1970. Staphylococcal alpha-toxin, p.

189-236. In S. J. Ail, S. Kadis, and T. C. Montie (ed.),Microbial toxins. vol. 3. Academic Press Inc., NewYork.

2. Artenstein, M. S., M. A. Madoff, and L. Weinstein. 1963.Studies on the biological activity of purified staphylo-coccal alpha-toxin. Yale J. Biol. Med. 35:373-389.

3. Bernheimer, A. W. 1970. Cytolytic toxins of bacteria, p.

183-212. In S. J. Ajl, S. Kadis. and T. C. Montie (ed.),Microbial toxins. vol. 1. Academic Press Inc., NewYork.

4. Bernheimer. A. W., L. S. Avigad, and P. Grushoff. 1968.Lytic effects of staphylococcal alpha-toxin and delta-hemolysin. J. Bacteriol. 96:487-491.

5. Carrel, A. 192:3. A method for the physiological study oftissues in vitro. J. Exp. Med. 38:407-418.

6. Dulbecco, R. 1952. Production of plaques in monolayertissue cultures by single particles of an animal virus.Proc. Nat. Acad. Sci. U.S.A. 38:747-752.

7. Eagle, H. 1959. Amino acid metabolism in mammaliancell cultures. Science 130:432-437.

8. Gabliks, J. 1972. Protective action of 2, 2-bis(para-chlorophenyl)-1, 1, 1-trichloroethane against intoxica-tion of cells by staphylococcal enterotoxin B. Infect.Immunity 6:364-369.

9. Gladstone, G. P., and A. Yoshida. 1967. The cytopathicaction of purified staphylococcal delta-hemolysin. Brit.J. Exp. Pathol. 48:11-19.

10. Hallander. H. O., and S. Bengtsson. 1967. Studies on thecell toxicity and species specificity of purified staphylo-coccal toxins. Acta. Pathol. Microbiol. Scand.70:107-119.

11. Hayflick, L., and P. S. Moorhead. 1961. The serialcultivation of human diploid strains. Exp. Cell Res.25:585-621.

12. Jeljaszewicz, J. 1973. Hemolysins, alpha, beta and delta.In J. 0. Cohen (ed.), The staphylococci. John Wileyand Sons, Inc., New York.

13. Kantor, H. S., B. Temples, and W. V. Shaw. 1972.Staphylococcal delta-hemolysin: purification and char-acterization. Arch. Biochem. Biophys. 151:142-156.

14. Kapral, F. A. 1972. Inhibition of Staphylococcus aureus

delta hemolysin by phospholipids. Proc. Soc. Exp.Biol. Med. 141:519-521.

15. Korbecki, M., and J. Jeljaszewicz. 1964. Action of staphy-lococcal alpha-hemolysin on KB-cells. Zentrabl. Bak-teriol. Parasitenk. Infektionskr. Hyg. I Abt. Orig.192:430-433.

16. Korbecki, M., and J. Jeljaszewicz. 1965. Action ofstaphylococcal toxins in cell culture. J. Infect. Dis.115:205-213.

17. Kreger, A. S., K.-S. Kim, F. Zaboretzky, and A. W.Bernheimer. 1971. Purification and properties of staph-ylococcal delta hemolysin. Infect. Immunity 3:449-465.

18. Levaditi, C., S, Mutermilch. 1913. Action de la toxinediphterique sur la survie des cellules in vitro. C. R.Soc. Biol. 74:379-382.

19. Litwin, J. 1970. The effect of commercial and pure

gentamicin on the growth of human diploid lungfibroblasts. Acta Pathol. Microbiol. Scand.78B:273-276.

20. Litwin, J. 1970. Standardization of human diploid cellcultivation. Appl. Microbiol. 20:899-906.

21. Litwin. J. 1971. Standardization of human diploid fibro-blast cultivation: centrifugation procedure. Appl. Mi-crobiol. 21:575-579.

22. Litwin, J. 1971. Standardization of human diploid fibro-blast cultivation: trypsinization procedure. Appl. Mi-crobiol. 21:169-174.

23. Litwin, J. 1972. Human diploid cell response to variationsin relative amino acid concentrations in Eagle medium.Exp. Cell Res. 72:566-568.

24. Litwin, J., and M. Thelestam. 1971. Use of diploid lungfibroblasts from human embryos for biological testing.Environ. Mutagen. Soc. News 4:37-38.

25. Litwin, J., and M. Thelestam. 1972. The lysis of humanfibroblasts with borate buffer. Acta Pathol. Microbiol.Scand. 80B:115-116.

26. Madoff, M. A., M. S. Artenstein, and L. Weinstein. 1963.Studies on the biologic activity of purified staphylococ-cal alpha-toxin. II. The effect of alpha-toxin on Ehrlichascites carcinoma cells. Yale J. Biol. Med. 35:382-389.

27. Maheswaran. S., and R. K. Lindorfer. 1967. Staphylococ-cal beta-hemolysin. II. Phospholipase C activity ofpurified beta hemolysin. J. Bacteriol. 94:1313-1319.

28. Mesrobeanu, L., and I. Mesrobeanu. 1971. Salmonellatyphimurium and Escherichia coli neurotoxins, p.

301-336. In S. Kadis, T. C. Montie, and S. J. Ajl (ed.).Microbial toxins, vol. 2A. Academic Press Inc., NewYork.

29. Mollby, R.. C. E. Nord, and T. Wadstr6m. 1973. Biologi-cal activities contaminating preparations of phospholi-pase C alpha-toxin from Clostridium perfringens. Toxi-con 11:139-147.

30. M6llby, R., and T. Wadstrom. 1971. Separation ofgamma hemolysin from Staphylococcus aureus Smith5R. Infect. Immunity 3:633-635.

31. M6llby. R., and T. Wadstrom. 1973. Purification ofstaphylococcal beta-, gamma- and delta-hemolysins, p.298-313. In J. Jeljaszewicz (ed.), Topics in microbiol-ogy and immunology, vol. 11. Karger Press, Basel.

32. Nicolson, G. L. 1972. Topography of membrane con-

canavalin A sites modified by proteolysis. Nature N.Biol. 239:193-196.

33. Rahal, J. J. 1972. Comparative effects of purified staphv-lococcal alpha- and delta-toxins on mitochondrial me-

tabolism. J. Infect. Dis. 126:96-103.34. Raskova, H., and K. Masek. 1971. Staphylococcal toxins,

p. 3-38. In H. Raskova (ed.), International ency-clopedia of pharmacology and therapeutics, Sect. 71.Pergammon Press, Oxford.

35. Schaeffer, W. I. 1970. Interaction of staphylococcal enter-otoxin B with cell cultures. Infect. Immunity1:455-458.

36. Schaeffer, W. I., J. Gabliks. and R. Calitis. 1967. Interfer-ence by trypsin in the interaction of staphylococcalenterotoxin B and cell cultures of human embryonicintestine. J. Bacteriol. 93:1489-1492.

37. Solotorovsky, N., and W. Johnson. 1970. Tissue cultureand bacterial protein toxins, p. 277-327. In S. J. Ajl. S.

Kadis, and T. C. Montie (ed.), Microbial toxins, vol. 1.

Academic Press Inc., New.York.38. Tarro, G. 1973. Appearance in trypsinized normal cells of

reactivity with antibody presumably specific for malig-

VOL. 8, 1973 945

on February 23, 2021 by guest

http://iai.asm.org/

Dow

nloaded from

THELESTAM, MOLLBY, AND WADSTROM

nant cells. Proc. Nat. Acad. Sci. U.S.A. 70:325-327.39. Wadstrom, T. 1968. Studies on extracellular proteins

from Staphylococcus aureus. IV. Purification of alpha-toxin by isoelectric focusing. Biochim. Biophys. Acta168:228-242.

40. Wadstr6m, T., and R. Mollby. 1971. Studies on extracel-lular proteins from Staphylococcus aureus. VI. Produc-tion and purification of beta-hemolysin in large scale.Biochim. Biophys. Acta 242:288-307.

INFECT. IMMUNITY

41. Wadstr8m, T., and R. Mollby. 1972. Some biologicalproperties of purified staphylococcal hemolysins. Toxi-con 10:511-519.

42. Wigzell, H. 1965. Quantitative titrations of mouse H-2antibodies using CrP'-labelled target cells. Transplan-tation 3:423-431.

43. Wiseman, G. M. 1968. The nature of staphylococcalbeta-hemolysin. II. Effect on mammalian cells. Can. J.Microbiol. 14:179-181.

946

on February 23, 2021 by guest

http://iai.asm.org/

Dow

nloaded from