Vol. 55, No. 7INFECTION AND IMMUNITY, JUlY 1987, p. 1641-16460019-9567/87/071641-06$02.00/0Copyright C) 1987, American Society for Microbiology

Purification, Characterization, and Toxicity of theSulfhydryl-Activated Hemolysin Listeriolysin 0 from

Listeria monocytogenesCHRISTIANE GEOFFROY,' JEAN-LOUIS GAILLARD,2 JOSEPH E. ALOUF,' AND PATRICK BERCHE2*

Unite des Antigenes Bacteriens (Centre National de la Recherche Scientifique, UA040557), Institut Pasteur, 75724 ParisCedex 15,1 and Faculte de Medecine Necker-Enfants Malades, Laboratoire de Microbiologie,

75730 Paris Cedex 15,2 France

Received 2 February 1987/Accepted 6 April 1987

We purified and characterized an extracellular hemolysin produced by Listeria monocytogenes. Hemolysinproduction was greatly enhanced by growing bacteria in resin (Chelex)-treated medium. This hemolysin wasseparated as a homogeneous protein of 60,000 daltons by using thiol-disulfide exchange affinity chromatogra-phy. This protein was a sulflydryl-activated toxin, termed listeriolysin 0, which shared the classical propertiesof other bacterial sulfhydryl-activated toxins: (i) inhibition by very low amounts of cholesterol; (ii) activationby reducing agents and suppression of the lytic activity by oxidation; (iii) antigenic cross-reactivity withstreptolysin 0. However, listeriolysin 0 differed remarkably from the other sulfhydryl-activated toxins in thatits cytolytic activity towards erythrocytes from various animal species was maximum at low pH (-5.5) and wasundetectable at pH 7.0. This suggests that the lytic activity of the toxin in host tissues might be better expressedin the acidic microenvironment, including macrophage phagosomes where bacteria presumably replicate.Listeriolysin 0 was lethal to mice (50% lethal dose of ca. 0.8 ,ug) and induced a rapid inflammatory reactionwhen injected intradermally. These results favor the view that listeriolysin 0 might play a major role duringintracellular replication of L. monocytogenes, ultimately promoting death of infected macrophages.

Listeria monocytogenes (sensu stricto) is a facultativeintracellular bacterium responsible for severe infections ofhumans and animals. Virtually all strains isolated fromnatural infections produce a zone of hemolysis on blood agarmedium (18, 35, 36, 39). However, the nature of the factor(s)responsible for this phenomenon remains controversial. Af-ter the demonstration of a soluble extracellular hemolysin byHarvey and Faber (20), a number of workers attempted toisolate it (17, 22-24, 26, 30, 38). The hemolysin (listeriolysin)was apparently a heat-labile antigenic protein, the lyticactivity of which was enhanced by reducing agents and wassuppressed by oxidation (22, 30) and by exposure to choles-terol (26) or to antistreptolysin 0 (anti-SLO) (22). Theseproperties and the cardiotoxic effects of the hemolytic ma-terial (27, 37), which were apparently similar to those elicitedby SLO (19), suggested that listeriolysin might belong to thegroup of the sulfhydryl (SH)-activated bacterial toxins (2,40), the prototype of which is SLO (1). However, otherreports were not in agreement with this view. In the firstplace, the hemolysin appeared to have a molecular weight(Mr) of 170,000 (23), whereas SLO and related purified toxinshave molecular weights in the range of 60,000 (2). Otherstudies indicated an Mr of 10,000 for a material activatableby reducing agents (38). On the other hand, a lipolyticmaterial of 52,000 M, was associated with the 170,000-Mrhemolysin (23). A hemolytic lecithinase was also described(24). Recently, a partially purified SLO-like hemolysin of Mr55,000 to 60,000 released by Listeria ivanovii has beenreported (32). Rabbit antisera against this hemolysin havebeen used for the detection of an Mr 60,000 hemolysin, called

* Corresponding author.

a-listeriolysin, in L. monocytogenes (32). Another unrelatedhemolysin(s) not detectable by these antisera was termed3-listeriolysin (32).The pathogenic role of the hemolytic material during

infection by L. monocytogenes is poorly understood. Thismaterial is cardiotoxic, lethal to mice (27, 37), and cytolyticfor many eucaryotic cells, including macrophages (5, 25, 42).It also lyses isolated organelles such as lysosomes (25).These findings suggested that the hemolysin might be impli-cated in the mechanisms of the intracellular growth of L.monocytogenes in macrophages, as inferred by electronmicroscopic study of infected mouse spleen (5). This viewhas been questioned (6), but received recent support from awork showing a close relationship between virulence andhemolysin secretion (15). A nonhemolytic L. monocyto-genes mutant obtained by transposon insertion was found tohave lost the capacity to grow in host tissues of infectedmice, and virulence was restored in its hemolytic revertantstrain (15). The present study was undertaken to furthercharacterize and to purify to homogeneity the putativeSH-activated listeriolysin as a step toward the elucidation ofthe mechanisms of virulence of L. monocytogenes. Sincetoxin production is restricted in high iron concentration (13),bacteria were grown in a medium treated with Chelex, aprocedure which has been described to remove free iron(31). This treatment allowed us to obtain high levels ofhemolysin. The toxin was then separated from the culturesupernatant as a homogeneous protein of Mr 60,000 by usingthiol-disulfide exchange affinity chromatography, as previ-ously used for the purification of three SH-activated toxins,SLO (34), alveolysin (16), and perfringolysin 0 (J. E. Aloufand C. Geoffroy, manuscript in preparation). The purifiedhemolysin from L. monocytogenes displayed the classical

1641

on February 21, 2020 by guest

http://iai.asm.org/

Dow

nloaded from

1642 GEOFFROY ET AL.

properties of the SH-activated toxins and induced stronginflammatory reactions.

MATERIALS AND METHODS

Bacterial strain and culture media. The hemolysin waspurified from the hemolytic L. monocytogenes strain EGDserovar 1/2a, originally from the Trudeau Institute. Thebacteria were grown in brain heart infusion (BHI) broth(Diagnostics Pasteur, Marnes la Coquette, France) or in aChelex-treated medium prepared as follows. A 10-fold-concentrated solution was first prepared: proteose peptoneno. 3 (Difco Laboratories, Detroit, Mich.), 200 g; yeastextract (Difco), 50 g; Na2HPO4- 12H20, 83 g; KH2PO4, 7 g;quartz-distilled water, 1,000 ml. This solution was stirred(100 rpm) for 8 h at room temperature (Incubator Shaker,model G25; New Brunswick Scientific Co., Edison, N.J.).Chelex 100 beads (100 to 200 mesh, sodium form; Bio-RadLaboratories, Richmond, Calif.) were added to obtain a 2%final suspension. The resin was then removed by filtrationthrough a 0.45-jim-pore size Nalgene filter (Nalgene Co.,Rochester, N.Y.). The Chelex-treated concentrate was fur-ther diluted in quartz-distilled water (1:10), adjusted to pH7.5, and autoclaved at 115°C for 20 min. Sterile glucose (2%final) and sodium bicarbonate (0.25% final) were addedbefore inoculation.Hemolysin purification. Ten milliliters of an overnight

bacterial culture in BHI broth was grown in 500 ml ofChelex-treated medium for 18 h at 37°C and then used toinoculate 27 liters of the same medium, dispensed in 5-literflasks. After a 24-h incubation at 37°C, the bacteria wereremoved by centrifugation at 10,000 x g for 20 min. Thesupernatant fluid was concentrated by ultrafiltration at 4°C inan Amicon DC-2 apparatus (Lexington, Mass.), usingAmicon HlP30 hollow fibers, which eliminate material ofbelow Mr 30,000. The crude concentrate, adjusted to pH 6.0,was then passed through two tandem columns of thiopropyl-Sepharose 6B (Pharmacia, Uppsala, Sweden) prepared asfollows. Four grams of freeze-dried thiopropyl-Sepharose6B, containing ca. 35 ,umol of available SH per ml of gel, wasused for each column. The wet powder was allowed to swellin phosphate-buffered saline (PBS) (pH 6.0) for 2 h. Theslurry was poured into a Pharmacia column (2.5 by 30 cm)and equilibrated by washing with 100 ml of buffer. The firstcolumn was previously reduced with 20 mM cysteine, allow-ing reduction of hemolysin. The resulting effluent was passeddirectly through the second column, retaining the hemolysinand other reduced materials, which were then eluted with 5mM dithiothreitol in PBS (pH 7.5). The eluate was concen-trated by ultrafiltration in a stirred cell equipped with anAmicon PM-30 membrane and gel filtered on a SephacrylS-200 (Pharmacia) column (2.5 by 100 cm). The elutedhemolytic effluent was concentrated as before and passedthrough a Bio-Gel P-100 (Bio-Rad) column (2.5 by 100 cm),and the appropriate effluent was finally gel filtered on aFractogel HW-50 (Merck Laboratories, Darmstadt, FederalRepublic of Germany) column (2.5 by 100 cm). All columnswere equilibrated and eluted with PBS (pH 6.0) containing10% glycerol.Hemolsyin assay. The hemolysin assay is based on the

estimation of the hemolytic activity of the toxin (activatedwith 20 mM cysteine) towards erythrocytes from humansand various animal species (sheep, horse, rabbit), as previ-ously described (4). Measurement was made of the opticalabsorbance at 541 nm of hemoglobin released from erythro-cytes (6 x 108 cells per ml) incubated (at 37°C for 45 min)

with 1 ml of appropriately diluted toxin in PBS (pH 6.0)containing 0.1% bovine albumin (Sigma Chemical Co., St.Louis, Mo.). One hemolytic unit (HU) is the amount of toxinneeded to release half the hemoglobin (50% lysis) of theerythrocytes. It is estimated graphically by plotting percentlysis versus toxin volume on a log-probit graph (4).

SDS-polyacrylamide gel electrophoresis. Proteins were an-alyzed by sodium dodecyl sulfate(SDS)-polyacrylamide gelelectrophoresis as described (28). Samples of 25 to 50 jxlwere boiled for 90 s in 2% (wt/vol) SDS-5% (voUvol)2-mercaptoethanol-10% (vol/vol) glycerol-0.002% (wt/vol)bromophenol blue in 0.1 M Tris hydrochloride buffer (pH6.8). Electrophoresis was performed in a linear gradient of7.5 to 25% (wt/vol) acrylamide at 5 mA for 15 h. The proteinswere visualized by staining with Coomassie brilliant blue orsilver nitrate.

Purified SH-activated toxins and anti-SLO serum. SLO andalveolysin were obtained by a procedure previously de-scribed (16, 34). Perfringolysin 0 was prepared similarly(unpublished data). Pneumolysin was kindly provided byMary K. Johnson (Tulane University, New Orleans, La.).Hyperimmune horse anti-SLO (serum no. 525), previouslydescribed (3), and nonimmune horse serum (GIBCO, Pais-ley, Scotland) were used in immunoprecipitation and neu-tralization experiments.

Hemolytic activity after incubation with sterols. Steroidswere dissolved in double-distilled absolute ethanol at aconcentration of 1 mg of cholesterol per ml, or 10 mglml forepicholesterol and dehydroepiandrosterone (Sigma). One-milliliter samples of toxin at a concentration of 30 HU/ml inPBS (pH 6.0) supplemented with 0.1% ovalbumin (Sigma)were incubated for 30 min at 22°C with 10 ,ul of variousdilutions of the tested sterols in ethanol. The controlsincluded 1 ml of toxin and 10 pl of ethanol. The hemolyticactivity of the mixture was then measured, and the concen-tration of sterol was determined as that inhibiting the hemo-lytic activity of 1 HU of hemolysin.

Hemolytic activity after incubation with SH-group reagents.Reagents used were HgCl2, iodoacetic acid, N-ethylmal-eimide, iodoacetamide, p-chloromercuribenzoate, tosyl-lysine chloromethyl ketone, and tosylphenylalaninechloromethyl ketone, provided by Sigma. All reagents weredissolved (10 mM) in PBS (pH 6.0), except for the last two,which were dissolved (100 mM) in absolute ethanol, and thenincubated for 30 min at 22°C with toxin samples. Thehemolytic activity was determined with the appropriatecontrols as described above for sterols.Mouse toxicity tests. Specific-pathogen-free ICR female

Swiss mice (Charles River, St. Aubin les Elboeuf, France) 6to 8 weeks old were used. Mice were injected intravenously(i.v.) or intraperitoneally with 0.5 ml of PBS (20 mMcysteine) at various pHs, containing increasing amounts ofhemolysin. The 50% lethal dose (LD50) was determined bythe probit method. Neutralization experiments were carriedout with pure toxin preincubated in vitro with cholesterol(0.2, wt/wt; 30 min at 37°C), with 1 mM of N-ethylmaleimide(3 h at 37°C), or with horse anti-SLO (no. 525) or nonimmunehorse serum (appropriately diluted in PBS, pH 6.8, for 30min at 37°C and injected i.v. in a volume of 0.25 ml). Toestimate the inflammatory potential of the hemolysin, vari-ous amounts of the toxin (diluted in PBS, pH 6.8) wereinjected into the right hind footpad of mice in a volume of 50pul. Footpads were measured 30 min later with dial calipers(Schnelltaster, Hessen, Federal Republic of Germany) thatare capable of measuring 0.05-mm increments in thickness.Other methods. Agarose double-immunodiffusion assays

INFECT. IMMUN.

on February 21, 2020 by guest

http://iai.asm.org/

Dow

nloaded from

PURIFICATION OF SH-ACTIVATED L. MONOCYTOGENES TOXIN

TABLE 1. Purification of SH-activated toxin from L. inonlo(YtogenleTotal Total activitv Sp act Recovery

Purification step Vol (ml) protein 1(60Ht( prtein Rce(mg) H U) (HU/mg of pi-otein) (%)

1. Culture supernatant fluid 27,000 7.290 43.00 5.925 100.02. Concentrated supernatant 650 2,145 42.00 21.666 98.03. Thiopropyl-Sepharose 6B 16 56 4.80 85.714 11.4

chromatography4. Sephacryl S-200 gel filtration 20 33 2.80 227.692 5.85. Bio-Gel P-100 gel filtration 4.5 3 1.35 437.782 3.16. Fractogel HW-50 gel filtration 8.0 0.60 0.60 1.000.000 1.3

were performed on glass slides in 1% agarose in PBS (pH6.8). Wells of 4 mm were cut in the agar gel plates and filledwith appropriate antigens and sera. Proteins were assayedaccording to the method of Bradford (9) with Bio-Radreagents. Histological studies were performed on 1-,um-thicksections of the footpad fixed in Bouin solution, embedded inparaffin, and stained with hematoxylin and eosin.

RESULTS

Purification of the SH-activated hemolysin. Bacteria weregrown for 18 h at 37°C in Chelex-treated medium, whichallowed a remarkably greater hemolysin production. Thehemolytic titer was about 1,500 HU/ml in the chelatedmedium (109 bacteria per ml), as compared to 64 HU/ml inBHI (1010 bacteria per ml). A summary of the purificationsteps, all carried out at 4°C, is presented in Table 1. Culturesupernatant (27 liters) was concentrated to 650 ml by ultra-filtration. The concentrate was applied to thiopropyl-Sepharose 6B columns, and the bound toxin was then eluted,concentrated to 16 ml, and applied to a Sephacryl S-200column. The pooled effluents that contained the peak ofhemolytic activity were sequentially passed through Bio-GelP-100 and Fractogel HW-50 columns, and the eluate wasconcentrated to 8 ml. This fraction exhibited a specificactivity of ca. 106 HU/mg and appeared as a single polypep-tide chain of Mr 60,000, visualized as one sharp band bySDS-polyacrylamide gel electrophoresis after staining with

1 2 3 4 5 6 7

_ 67K

_ui' 43K

Coomassie brilliant blue (Fig. 1) or silver nitrate (notshown). Increasing the load of pure toxin up to 5 ,ug ofprotein did not reveal any additional band.

Characteristics of the purified hemolysin. The purifiedhemolysin displayed the usual properties of the SH-activatedcytolysins. The hemolytic activity of the toxin was inhibitedby cholesterol. In the presence of ovalbumin as proteinprotectant, 1 HU was inhibited by ca. 0.2 ng of cholesterol.Epicholesterol was a very weak inhibitor (15 ,ug/HU), al-though it did not inhibit alveolysin or SLO under the sameconditions. Dehydroepiandrosterone, which has the samesteroid nucleus as cholesterol but lacks the aliphatic sidechain, was inactive even at 100 ,ug/HU, as previously foundfor other SH-activated toxins (33). The hemolytic activity ofthe toxin was totally inhibited by 1 mM HgCl2 or p-chloromercuribenzoate. Inhibition by the mercurials wasreversed by 2 mM dithiothreitol or cysteine. lodoacetic acidat 2 mM, 2 mM iodoacetamide, 1 mM tosyllysine chloro-methyl ketone, and 1 mM tosylphenylalanine chloromethylketone were not inhibitory. Under the same conditions,alveolysin activity was inhibited by 50 and 15% with thelatter two reagents, respectively. The toxin was antigenicallyrelated to SLO, since the pure toxin gave a single immuno-precipitate line by gel diffusion versus horse anti-SLO (Fig.2), but did not immunoprecipitate with nonirnmune horseserum at all (not shown).

Effect of pH on hemolytic activity. The hemolytic activityof the purified hemolysin from L. monocytogenes as afunction of pH was compared with those of SLO, perfringo-lysin 0, pneumolysin, and alveolysin. The results observedon sheep erythrocytes are illustrated in Fig. 3. The activityof listerial hemolysin had the lowest optimum pH (pH 5.5),as compared to those of pneumolysin (pH 6.0), perfringo-lysin 0 and alveolysin (pH 6.5), and SLO (pH 7.0). Incontrast to the other four hemolysins, the listerial hemolysinexhibited a narrow pH range of activity, and hemolysis wasnot observed at pH 7.0. The hemolytic activity was fully

- 30K

20K

4 1 4.4 K

FIG. 1. SDS-polyacrylamide gel electrophoresis of Table 1 frac-tions (30 ,ul, 2,000 to 9,000 HU). Lanes: 1, culture supernatant (8,ug); 2, step 2 (10 p.g); 3, step 3 (100 ,ug); 4, step 4 (50 ,ug); 5, step 5(20 ,ug); 6, step 6 (2.2 ,ug); 7, molecular weight markers. Gels werestained with Coomassie brilliant blue. Purified hemolysin (lane 6)migrates as a single band of M, ca. 60,000.

FIG. 2. Immunological cross-reactivity between L. monocyto-genes hemolysin and SLO. Upper wells: right, purified hemolysin(10 ,ug); left, purified SLO (10 p.g). Bottom well: 20 ,ul of undilutedequine anti-SLO no. 525. No immunoprecipitation was observedwith nonimmune horse serum used as controls (not shown).

VOL. 55, 1987 1643

Im4w qlmmw jumm

"......

on February 21, 2020 by guest

http://iai.asm.org/

Dow

nloaded from

1644 GEOFFROY ET AL.

100

;> 90

>. 80

p 70.

60

t;50O4 4002 30

= 20

teo1

CHOLESTEROL FOOTPAD SWELLING ( 0.1 mm)0 2 4 6 8 10 12 14 16 18 20 22

lI g +

0.1 gg0.1 ±g +

0.01 igg0.01 ig +

0.001 gg0.001 gg +

4.5 5 5.5 6 6.5

pH

FIG. 3. Effect of pH on the lytic activity on sheep erythrocytesof purified hemolysin from L. monocytogenes (s), as compared withother SH-activated toxins: pneumolysin (x), alveolysin (LI),perfringolysin 0 (0), and SLO (0). A 10-HU sample of pure toxinwas incubated with 6 x 108 sheep erythrocytes per ml.

restored by lowering the pH from 7.0 to 5.5. Similar resultswere observed with human, horse, and rabbit erythrocytes.Mouse toxicity. The LD50 of purified hemolysin was deter-

mined by i.v. injection into mice. The LD50 was estimated atabout 0.8 p.g per mouse (Table 2). Mice died with convul-sions and opisthotonos. When toxin was administered inPBS at pH 6.8, 6.0, or 5.5, the mice died within 1 to 2 min,as opposed to control mice, which received the same volume(0.5 ml) of PBS at various pH levels but did not display anyclinical symptom. At pH 7.2, the mice died after a delay of 30to 60 min, but the LD50 was the same. When toxin wasinjected by the intraperitoneal route, animals died severalhours later, and the LD50 was assessed to be about 1.7 ,ug permouse. No mortality was observed from injecting intra-dermally up to 5 p.g of toxin.The hemolysin lost all its lethal potential when treated

with 1 mM N-ethylmaleimide (3 h at 37°C) with cholesterol(0.2, wt/wt; 30 min at 37°C), or with heat (1 h at 60°C) (Table2). In addition, a lethal dose of toxin (1 ,ug) was neutralizedby incubation (30 min at 37°C) with equine anti-SLO before

TABLE 2. Toxicity of hemolysin in mice

Hemolysin Treatment before Mortality"(t1g)' injection

5.0 10/102.5 10/101.2 10/100.6 0/105.0 Cholesterol' 0/105.0 N-Ethylmaleimide" 0/105.0 Heating' 0/101.0 Anti-SLO (1:1)' 0/51.0 Anti-SLO (1:10)t 5/51.0 Normal serum (1:1)' 5/51.0 Normal serum (1:10)Y 5/5

Hemolysin was injected intravenously in a volume of 0.5 ml of PBS (pH6.8).

" Mortality in 6- to 8-week-old Swiss mice: nuniber of dead mice/total ofinjected mice; except in the group receiving anti-SLO 1:10. in which micesurvived only for 5 to 6 h. all mice died within minuites after i.v. injection.

Incubation with cholesterol (0.2. wt/wt; 30) min at 37C)."Incubation with 1 mM N-ethylmaleimide (3 h at 37°C).' For 1 h at 60°C.'Incubation (30 min at 37C) with equine anti-SLO no. 525 or nonimmuine

control serum (undiluted or diluted in PBS. pH 6.8).

FIG. 4. Inflammatory reaction in mice, elicited by purifiedhemolysin from L. monocytogenes diluted in PBS (pH 6.8) and

-,'-~C injected into the right-hind footpad. Solid bars, Hemolysin alone;7.5 8 8.5 open bars, same doses of hemolysin previously incubated with

cholesterol (0.2, wt/wt, 30 min at 37°C).

i.v. injection (Table 2). In contrast to control mice receivingthe same amount of toxin preincubated with nonimmunehorse serum, which died within minutes, the animals in-jected with neutralized toxin survived either undefinitely(undiluted anti-SLO) or only for 5 to 6 h (anti-SLO 1:10 inPBS, pH 6.8).The toxin diluted in PBS (pH 6.8) was also injected

intradermally into mouse footpads. It induced a rapid inflam-matory response that peaked in 0.5 h and slowly faded overthe next 24 h. Histological examination showed an exudatemostly constituted of polymorphonuclear cells. The dosedependence of this inflammatory reaction is shown in Fig. 4.One microgram of hemolysin induced an inflammatory swell-ing of about 20 U (1 U = 0.1 mm) of thickness. Even minutedoses (0.01 ,ug) were capable of inducing a significant reac-tion (-6 U). When the toxin was previously incubated withcholesterol, the inflammatory reaction was almost com-pletely abrogated.

DISCUSSIONAlthough it has generally been thought that L. monocyto-

genes produced an SH-activated hemolysin, this hemolysinhad not previously been available in a sufficiently hemogene-ous form to assess its nature unambiguously. The failure ofprevious attempts to purify it is very likely due to severalfactors: (i) the low levels of listeriolysin released in conven-tional culture media (20 to 200 HU/ml), which also vary fromstrain to strain (32); (ii) inappropriate purification methods;(iii) the presence of more than one hemolysin in culture fluids(23, 24, 32, 42). These difficulties were overcome in thepresent work, which for the first time allowed the purifica-tion of the hemolysin to homogeneity and its clear-cutcharacterization as an SH-activated toxin. A high level ofextracellular toxin (1,500 HU/ml) was obtained in a resin(Chelex)-treated broth, as opposed to the amount ofhemolysin produced in BHI broth (64 HU/rnl). This con-firmed previous observations indicating that toxin produc-tion is optimum at low iron levels and declines as a functionof increasing iron concentration (13). Similar iron control isknown for diphtheria toxin, Pseudomonas aeruginosaexotoxin A, and other toxins (21). The purification of thehemolysin was initiated with thiol-disulfide exchange col-umns, which selectively retain SH-containing proteins. An-other crucial technical improvement which allowed us toobtain homogeneous hemolysin was the final gel filtrationstep through Fractogel HW-50, a new hydrophilic matrix.The yield of toxin purification was very low, due to impor-tant losses at stage 3 of the purification (Table 1). Animprovement of the purification procedure is currently inprogress.

INFECT. IMMUN.

on February 21, 2020 by guest

http://iai.asm.org/

Dow

nloaded from

PURIFICATION OF SH-ACTIVATED L. MONOCYTOGENES TOXIN

The purified toxin displayed a specific activity of the sameorder (106 HU/mg) as that reported for other purified SH-activated toxins (3). Its molecular weight was also similar(ca. 60,000). The toxin was lethal to mice at a level (LD50 Of0.8 ,ug) comparable to that reported for SLO (0.2 ,ug) (1). Itdisplayed the classical properties of the SH-activated toxins:(i) it is inhibited by very low amounts of cholesterol; (ii) it isactivated by reducing agents and its toxicity is suppressed byreagents which modify SH groups; (iii) it is antigenicallyrelated to SLO, as assessed by immunoprecipitation andneutralization tests with horse anti-SLO. On the basis ofthese similarities, we propose that this hemolysin be termedlisteriolysin 0, according to the nomenclature recommendedby Bernheimer when other cytolysin(s) than the SH-activated one are produced by a microorganism, as is thecase for Streptococcus sp. and Clostridum sp. (7). Whethera-listeriolysin (32) and listeriolysin 0 are identical remains tobe clarified in spite of apparent common properties asregards molecular weight and cross-reactivity with SLO.Moreover, by using immunoblotting with a rabbit immuneserum raised against highly purified listeriolysin 0, wedetected this toxin in culture supernatants of bacteria, grownunder adequate conditions as described in this work, from allL. monocytogenes strains currently tested and obtainedfrom human clinical isolates (manuscript in preparation).

Several of our findings concerning the properties and theconditions of production of the SH-activated toxin might beof critical importance in terms of pathogenicity of L. mono-cytogenes. The purified toxin was cytolytic at very lowdoses: 1 HU corresponds to 1 ng, equivalent to ca. 1010molecules. It can be calculated that only 30 to 40 moleculesof listeriolysin 0 is needed to lyse a single erythrocyte. Thisnumber is similar to that found with SLO (1). Likewise,Bhakdi et al. (8) estimated that the presence of approxi-mately 100 SLO monomers per cell is required to generateone functional lesion on a cell. These results can be com-pared to the amounts of listeriolysin 0 produced in vitroafter 18-h incubation at 37°C. It can be calculated thatlisteriolysin 0 production reached about 15,000 moleculesper bacterial cell in Chelex-treated medium (109 bacteria perml produce 1,500 HU/ml). Even in inappropriate medium forhemolysin production (BHI broth), the amount of listeri-olysin 0 still reached about 64 molecules per bacterial cell(1010 bacteria per ml produce 64 HU/ml). Thus it is quitereasonable to suppose that the amount of listeriolysin 0produced by one or several bacterial cells inside phagosomesis sufficient to damage macrophage membranes. In thisconnection, it is known that the amount of free iron in hosttissues is very low, 10-18 M (12), due to the high associationconstant of the host iron-binding proteins (12), suggestingthat a significant increase of toxin production could occur inthe cellular microenvironment during infection (13).The most interesting finding concerns the optimum pH of

listeriolysin 0 activity (Fig. 3), which is significantly lowerthan that of all other SH-activated toxins purified so far. Thecytolytic activity was maximum at pH -5.5 and undetect-able at pH 7.0, whereas the other toxins were mostly activeat pH 2 6.5 to 7.0. This finding is apparently at variance withthe observation of hemolysis zones on blood agar preparedat neutral pH. It should be kept in mind that growing bacteriaacidify the surrounding medium, thus favoring binding oftoxin molecules onto cell membrane cholesterol. This sug-gests that listeriolysin 0 would display better activity in theacidic environment in vivo, including in phagosomes wherebacteria presumably replicate. In this connection, it is tempt-ing to relate this finding to the fact that L. monocytogenes is

the only intracellular facultative microorganism among thebacterial pathogens secreting SHI-activated toxins.

Altogether, these results support the model of intracellulargrowth of L. monocytogenes proposed by Armstrong andSword in 1966 (5, 41) and substantiated by a recent report(15). The sequence of events following bacterial invasion ofthe host tissues during infection starts with a very rapidphagocytosis of invasive bacteria (29) in a iron-deprivedenvironment (12), which stimulates hemolysin secretion(13). Internalization of bacteria inside acidic phagosomes ofresident macrophages would favor local accumulation oflisteriolysin 0 in the cellular microenvironment and subse-quent irreversible binding to the membrane cholesterol. Thisleads to disruption of intracellular membranes and toxinexport in the macrophage cytoplasm, rendering the cytoplas-mic Fe3" available for bacterial growth through a nonspecificmechanism of iron acquisition (14) and ultimately stimulatingrapid bacterial replication (41). Moreover, it might be spec-ulated that the local production of listeriolysin 0 associatedwith bacterial multiplication in host tissues is also implicatedin the formation of mutiple discrete infectious foci, whichappear in the spleen and liver during the initial phase ofinfection and are mainly populated with polymorphonuclearcells (29). This is supported by the observation that thecytolytic toxin can induce an inflammatory reaction even atvery low doses (0.01 ,ug) when inoculated in the footpad(Fig. 4). On the view of recent reports on other SH-dependent toxins (10, 11), listeriolysin 0 might similarlyinitiate the generation of leukotrienes from granulocytes andmacrophages upon binding on their cytoplasmic membranein infectious foci, and thereby potentiate inflammatory reac-tion during the early phase of infection. Finally, since thereis evidence that the process of intracellular bacterial growthis a prerequisite for in vivo induction of T-cell-mediatedimmunity (6a), we are currently studying the role of listeri-olysin 0 during the initiation of acquired immunity to L.monocytogenes.

ACKNOWLEDGMENTS

We gratefully acknowledge Michael Gill (Tufts University, Bos-ton, Mass.) for reviewing the manuscript and P. Sansonetti forhelpful discussion. We extend our grateful appreciation to I.Razafimanantsoa for her invaluable assistance in supervising thecare for experimental animals throughout this study, and LydiaOussadi for typing the manuscript.

This work was supported by a grant (DES 3009) from theDirection de l'Enseignement Superieur (Universite Paris V).

LITERATURE CITED1. Alouf, J. E. 1980. Streptococcal toxins (streptolysin 0, strep-

tolysin S, erythrogenic toxins). Pharmacol. Ther. 11:661-717.2. Alouf, J. E., and C. Geoffroy. 1984. Structure activity relation-

ships in sulfhydryl-activated toxins, p. 165-171. In J. E. Alouf,F. J. Fehrenbach, J. H. Freer, and J. Jeljaszewicz (ed.), Bac-terial protein toxins. Academic Press, London.

3. Alouf, J. E., C. Geoffroy, A. M. Gilles, and P. Falmagne. 1986.Structural relatedness between five bacterial sulfhydryl-activat-ed toxins: streptolysin 0, perfringolysin 0, alveolysin, pneumo-lysin and thuringiolysin, p. 49-50. In P. Falmagne, J. E. Alouf,F. J. Fehrenbach, J. Jeljaszewicz, and M. Thelestam (ed.),Bacterial proteins: toxins. Gustav Fischer-Verlag, Stuttgart.

4. Alouf, J. E., M. Viette, R. Corvazier, and M. Raynaud. 1965.Preparation et proprietes de serums de chevaux antistreptoly-sine 0. Ann. Inst. Pasteur 108:476-500.

5. Armstrong, B. A., and C. P. Sword. 1966. Electron microscopyof Listeria monocytogenes-infected mouse spleen. J. Bacteriol.91:1346-1355.

6. Barza, M. 1985. Listeriosis and milk. N. Engl. J. Med.

VOL. 55, 1987 1645

on February 21, 2020 by guest

http://iai.asm.org/

Dow

nloaded from

1646 GEOFFROY ET AL.

312:438-440.6a.Berche, P., J. L. Gaillard, and P. Sansonetti. 1987. Intracellular

growth of Listeria monocytogenes as a prerequisite for in vivoinduction of T cell-mediated immunity. J. Immunol. 138:2266-2271.

7. Bernheimer, A. W. 1976. Sulfhydryl-activated toxins, p. 85-97.In A. W. Bernheimer (ed.), Mechanisms in bacterial toxinology.John Wiley and Sons, Inc., New York.

8. Bhakdi, S., M. Roth, Z. Sziegoleit, and J. Tranum-Jensen. 1984.Isolation and identification of two hemolytic forms of strepto-lysin 0. Infect. Immun. 46:394-400.

9. Bradford, M. 1976. A rapid and sensitive method for thequantitation of microgram quantities of protein, utilizing theprincipe of protein-dye binding. Anal. Biochem. 72:248-257.

10. Bremm, K. D., H. J. Brom, J. E. Alouf, W. Konig, B. Spur, A.Crea, and W. Peters. 1984. Generation of leukotrienes fromhuman granulocytes by alveolysin from Bacillus alvei. Infect.Immun. 44:188-193.

11. Bremm, K. D., W. Konig, P. Pfeiffer, I. Rauschen, K. Theobald,M. Thelestam, and J. E. Alouf. 1985. Effect of thiol-activatedtoxins (streptolysin 0, alveolysin, and theta toxin) on thegeneration of leukotrienes and leukotriene-inducing and -me-tabolizing enzymes from human polymorphonuclear granulo-cytes. Infect. Immun. 50:844-851.

12. Bullen, J. J. 1981. The significahce of iron in infection. Rev.Infect. Dis. 3:1127-1138.

13. Cowart, R. E., and B. G. Foster. 1981. The role of iron in theproduction of hemolysin by Listeria monocytogenes. Curr.Microbiol. 6:287-290.

14. Cowart, R. E., and B. G. Foster. 1985. Differential effects of ironon the growth of Listeria monocytogenes: minimum require-ments and mechanism of acquisition. J. Infect. Dis. 151:721-730.

15. Gaillard, J. L., P. Berche, and P. Sansonetti. 1986. Transposonmutagenesis as a tool to study the role of hemolysin in thevirulence of Listeria monocytogenes. Infect. Immun. 52:50-55.

16. Geoffroy, C., and J. E. Alouf. 1983. Selective purification bythiol-disulfide interchange chromatography of alveolysin, a sulf-hydryl-activated toxin of Bacillus alvei. J. Biol. Chem.258:9968-9972.

17. Girard, K. F., A. J. Sbarra, and W. A. Bardawil. 1963. Serologyof Listeria monocytogenes. I. Characteristics of the solublehemolysin. J. Bacteriol. 85:349-355.

18. Groves, R. D., and H. J. Welshimer. 1977. Separation ofpathogenic from apathogenic Listeria monocytogenes by threein vitro reactions. J. Clin. Microbiol. 5:559-563.

19. Halbert, S. P. 1970. Streptolysin 0, p. 69-98. In T. C. Montie,S. Kadis, and S. J. Ajil (ed.), Microbial toxins, vol. 3. AcademicPress, Inc., New York.

20. Harvey, P. C., and J. E. Faber. 1941. Some biochemicalreactions of the Listerella group. J. Bacteriol. 41:45-46.

21. Holmes, R. K., and L. M. Russel 1983. Role of iron in regulatingsynthesis of bacterial toxins, p. 359-362. In D. Schlessinger(ed.), Microbiology-1983. American Society for Microbiology,Washington, D.C.

22. Jenkins, E. M., A. N. Njoku-Obi, and E. W. Adams. 1964.Purification of the soluble hemolysins of Listeria monocyto-genes. J. Bacteriol. 88:418-424.

23. Jenkins, E. M., and B. B. Watson. 1971. Extracellular antigensfrom Listeria monocytogenes. I. Purification and resolution of

hemolytic and lipolytic antigens from culture filtrates of Listeriamonocytogenes. Infect. Immun. 3:589-594.

24. Khan, M. A., A. Seaman, and M. Woodbine. 1973. Listeriamonocytogenes haemolysin: lecithinase. Zentralbl. Bakteriol.Parasitenkd. Infektionskr. Hyg. Abt. 1 Orig. Reihe A 225:66-79.

25. Kingdon, G. C., and C. P. Sword. 1970. Effects of Listeriamonocytogenes hemolysin on phagocytic cells and lysosomes.Infect. Immun. 1:356-362.

26. Kingdon, G. C., and C. P. Sword. 1970. Biochemical andimmunological effects of Listeria monocytogenes hemolysin.Infect. Immun. 1:363-372.

27. Kingdon, G. C., and C. P. Sword. 1970. Cardiotoxic and lethaleffects of Listeria monocytogenes hemolysin. Infect. Immun.1:373-379.

28. Laemmli, U. K. 1970. Cleavage of structural proteins during theassembly of the head of bacteriophage T4. Nature (London)227:680-685.

29. Mackaness, G. B. 1962. Cellular resistance to infection. J. Exp.Med. 116:381-406.

30. Njoku-Obi, A. N., E. M. Jenkins, J. C. Njoku-Obi, J. Adams,and V. Covington. 1963. Production and nature of Listeriamonocytogenes hemolysins. J. Bacteriol. 86:1-8.

31. O'Brien, A. D., and G. D. Laveck. 1982. Immunochemical andcytotoxic activities of Shigella dysenteriae 1 (Shiga) and Shiga-like toxins. Infect. Immun. 35:1151-1154.

32. Parrisius, J., S. Bhakdi, M. Roth, J. Tranum-Jensen, W. Goebel,and H. P. R. Seeliger. 1986. Production of listeriolysin by beta-hemolytic strains of Listeria monocytogenes. Infect. Immun.51:314-319.

33. Prigent, D., and J. E. Alouf. 1976. Interaction of streptolysin 0with sterols. Biochim. Biophys. Acta 443:288-300.

34. Prigent, D., C. Geoffroy, and J. E. Alouf. 1978. Purification de lastreptolysine 0 par chromatographie covalente sur gel de thiol-agarose. C.R. Acad. Sci Paris 287:951-954.

35. Rocourt, J., F. Grimont, P. A. D. Grimont, and H. P. R.Seeliger. 1982. DNA relatedness among serovars of Listeriamonocytogenes sensu lato. Curr. Microbiol. 7:383-388.

36. Rocourt, J., A. Schrettenbrunner, and H. P. R. Seeliger. 1983.Diffdrenciation biochimique des groupes gdnomiques de Listeriamonocytogenes (sensu lato). Ann. Microbiol. (Inst. Pasteur)134:65-71.

37. Siddique, I. H., and G. W. Cooper. 1969. Hemodynamic effectof listerial hemolysin in dogs. Am. J. Physiol. 216:1399-1403.

38. Siddique, I. H., I. F. Lin, and R. A. Chung. 1974. Purificationand characterization of hemolysin produced by Listeria mono-cytogenes. Am. J. Vet. Res. 35:289-296.

39. Skalka, B., J. Smola, and K. Elischerova. 1982. Routine test forin vitro differentiation of pathogenic and apathogenic Listeriamonocytogenes. J. Clin. Microbiol. 15:503-507.

40. Smyth, C. J., and J. L. Duncan. 1978. Thiol-activated (oxygen-labile) cytolysins, p. 129-183. In J. Jeljaszewicz and T. Wad-strom (ed.), Bacterial toxins and cell membranes. AcademicPress, Inc., New York.

41. Sword, C. P. 1966. Mechanisms of pathogenesis in Listeriamonocytogenes. I. Influence of iron. J. Bacteriol. 92:536-542.

42. Watson, B. B., and J. C. Lavizzo. 1973. Extracellular antigensfrom Listeria monocytogenes. II. Cytotoxicity of hemolytic andlipolytic antigens of Listeria for cultured mouse macrophages.Infect. Immun. 7:753-758.

INFECT. IMMtJN.

on February 21, 2020 by guest

http://iai.asm.org/

Dow

nloaded from