INFECTION AND IMMUNITY, May 1971, p. 678-687 Vol. 3, No. 5Copyright @ 1971 American Society for Microbiology Printed in U.S.A.

Properties of the Hemolytic Activities ofEscherichia coli

EVERETT C. SHORT, JR., AND HAROLD J. KURTZ

College of Veterinary Medicine, University of Minnesota, St. Paul, Minnesota 55101

Received for publication 16 February 1971

Some properties of the cell-free and cell-associated hemolysins of Escherichia coliwere studied. Several strains of E. coli that were isolated from intestines of pigswith edema disease produce large quantities of cell-free hemolysin when grownin the presence of an extract of meat. The component of meat that stimulatesproduction of cell-free hemolysin is not extracted by lipid solvents and is not dialyz-able. The cell-free hemolysin is an acidic substance that occurs in two forms. It isinactivated by trypsin but not by lecithinase, lysozyme, ribonuclease, or deoxyribonu-clease, shows optimum activity between pH 7 and 8, and requires calcium ion foractivity. It does not appear to be an enzyme. The kinetics of the lytic reaction aremost consistent with the hypothesis that one molecule of cell-free hemolysin issufficient to lyse one erythrocyte and that it is inactivated in the lytic reaction. Thecell-free hemolysin does not sufficiently damage the cell during the prelytic periodto cause lysis after the hemolysin-calcium-erythrocyte complex has been disrupted.The cell-associated hemolysin was not separated from the cell by autolysis, freezing,sonic treatment, or treatment with trypsin or lysozyme. It appears to be closely as-sociated with the metabolic status of the cell. Organisms that are highly hemolyticunder usual conditions of assay immediately lose most of their hemolytic capability inthe presence of sodium cyanide, streptomycin, nalidixic acid, and rifampin.

Certain strains of Escherichia coli are capableof hemolyzing red blood cells. Lovell and Rees(6) first reported the isolation of a filterablehemolysin from young cultures of E. coli grownin a cooked meat medium. Later Smith (9)described two hemolysins produced by strains ofE. coli, one (alpha) that could be obtained freefrom bacterial cells and another (beta) that couldnot. According to Smith, who tested a largenumber of hemolytic E. coli, alpha hemolysinwas not produced on ordinary laboratory mediabut only on cooked meat medium. However,Snyder and Koch (10) reported that both filter-able and nonfilterable hemolysin were producedboth in defined media and in heart infusionmedium. Recently, a third hemolysin (gamma)has been described in mutants that are resistantto nalidixic acid (12).The purpose of this paper is to report the

production of hemolysins by strains of E. coliisolated from the intestine of post-weanling pigswith edema disease and to describe some prop-erties of the cell-free and cell-associated hemoly-sins from a typical organism of this group. Weare concerned with the hemolysins of E. colibecause hemolysin production has been associ-ated with strains that produce enteric diseases in

some species of animals and extra-intestinaldiseases in man. The role of these substances inthe genesis of disease is a problem of practicalimportance. The problem is not likely to besolved until they can be purified and studied inanimals and cell preparations.

MATERIALS AND METHODSOrganisms. The organisms used in this work were he-

molytic strains of E. coli that were isolated from theintestines of pigs with edema disease. Petri plates con-taining sheep erythrocyte agar medium were inoculatedwith intestinal contents and incubated aerobically at37 C overnight. Hemolytic colonies were purified bysubculturing on MacConkey agar. Typical lactose-fermenting colonies were also submitted to the indole,citrate utilization, hydrogen sulfide production, andcarbohydrate fermentation tests. The bacteria weremaintained on agar slants. Cultures were routinelyincubated in a shaking water bath at about 120 cyclesper min at 37 C. The concentration of bacteria in liquidculture was determined by measuring the turbidity at600 nm. The turbidity of the culture was related to theconcentration of viable cells by plate counts. Allportions of this work after the initial tests of the con-ditions required for hemolysin production were con-ducted with a single strain of E. coli, serotype 0141:85ac:NM (Minnesota stock culture 4331).

Media. Alkaline meat extract broth was prepared

678

on February 20, 2020 by guest

http://iai.asm.org/

Dow

nloaded from

HEMOLYTIC ACTIVITIES OF E. COLI

as described by Smith (9). Other media used in thiswork were those described by Snyder and Koch (10)and by Fraser and Jerrel (4). Nutrient Broth was aproduct of Difco.

Preparation of alpha-hemolysin. A culture of E.coli was grown in alkaline meat broth to late logphase at which point both beta and alpha hemoly-sins are present at nearly maximum levels. The culturewas chilled in ice, and the cells were removed bycentrifugation at 25,000 X g for 30 min. The centri-fuge was kept at 5 C. The supernatant was filteredthrough a sterile membrane filter with 45-,Am pores(Millipore Corp., Bedford, Mass.). The filtrate,which was the crude alpha hemolysin, was stored at4 C at pH 7.5. The hemolysin is stable for severalweeks at 4 C. The concentration of hemolysin in thispreparation was usually between 50 and 80 units perml.

Erythrocyte suspensions. Blood was collected fromnormal sheep and mixed with Alsever's solution.The cells were used within a week after bleeding.Before use in hemolysin assays, the cells were washedthree times with 0.01 M tris(hydroxymethyl)amino-methane (Tris)-chloride buffer, pH 7.5, containing0.135 M sodium chloride (Tris-buffered saline).After washing, the cells were suspended in the samebuffer at 2% concentration. The 2%,Cc cell suspensionwas always used in assays on the same day it wasprepared.

Preparation of red cell fractions. Red cells werelysed at 4 C in 10 mm Tris-chloride buffer (pH 7.5)at a lysate-packed cell volume ratio of 50:1. A frac-tion (lysed-cell suspension) of 3 ml was removed anddiluted with 9 ml of the buffer containing 10 mmTris-chloride, 20 mm calcium chloride, and 170 mMsodium chloride. The remainder of the preparationwas centrifuged at 10,000 X g for 15 min. The super-natant was removed, and 3 ml was diluted as above(cell-free lysate). The precipitate was resuspended inTris-buffered saline, resedimented, and resuspendedin 17 ml of Tris-buffered saline containing 20 mmcalcium chloride (ghosts).

Measurement of hemolytic activity. Hemolyticactivity was measured on the basis of the amount ofhemoglobin released in a solution containing hemol-ysin or hemolytic bacteria, erythrocytes, Tris-chloride buffer, calcium chloride, and sodium chlo-ride. The incubation mixture (2.0 ml total volume)contained 1 sheep erythrocytes, 0.01 M Tris-chloridebuffer, pH 7.5, 0.02 M calcium chloride, 0.115 MNaCl, and 0.01 to 0.1 units of hemolysin. The hemol-ysin was diluted for assay in 0.01 M Tris-chloridebuffer, pH 7.5, containing 0.135 M sodium chloride.The assay vessels were incubated at 40 C for 60 min.At the end of the incubation period, the cells weresedimented by centrifugation at 1,000 X g for 10min. The amount of hemoglobin in the supernatantsolution was determined by measuring its absorbanceat 540 nm in a model DB spectrophotometer (Beck-man Instruments, Inc., Fullerton, Calif.). The ex-tinction coefficient of hemoglobin monomer at 540nm was taken as 14.7 X 103M-1 cm7l. One unit ofactivity is defined as the amount that causes the re-lease of 1 /Amole of hemoglobin monomer in 1 hr

under the conditions used for assay. Control vesselswithout hemolysin were treated in the same way asthe assay vessels in each experiment. The amount ofhemoglobin released was shown to be approximatelyproportional to the amount of hemolysin in the in-cubation vessel up to about 0.1 units of hemolysin.

Chromatography. Sephadex G-200 and Sepharose6B were products of Pharmacia. The columns ofSephadex G-200 and Sepharose 6B were 1.5 cm indiameter and 136 cm long. The eluant was Tris-buffered saline, and the elution rate was 6 ml per hr.

Diethylaminoethyl (DEAE) cellulose was Cellex-D,a product of Bio-Rad Laboratories. Samples weredialyzed against 10 mm phosphate buffer, pH 6,before chromatography on DEAE cellulose. A column(9 mm in diameter and 150 mm long) was prepared.Elution was carried out with a linear gradient ofsodium chloride generated by two cylindrical vesselsplaced in series. Both vessels contained 10 mm phos-phate buffer, pH 6, and the vessel away from thecolumn also contained 400 mm sodium chloride.

Other materials. The dialysis tubing used in theseexperiments was boiled for 15 min in 10% sodiumcarbonate, then in 10 mm ethylenediaminetetraaceticacid (EDTA), pH 10, and finally in distilled water.The tubing was washed with distilled water aftereach step, and it was stored in 10 mm EDTA, pH7.5. Casein hydrolysate was a product of NutritionalBiochemicals Corp., Cleveland, Ohio. Enzymes werefrom Worthington Biochemical Corp., Freehold,N.J. The water was distilled, passed through a mixedbed demineralizer and an organic removal cartridge(Barnstead), and redistilled from an all-glass still.

RESULTSProduction of hemolysins. The production of

alpha- and beta-hemolytic activities was examinedwith 12 different isolates of hemolytic E. coli.The whole cultures of all of the organisms werehemolytic when grown on any of the media used.The maximum concentration of beta hemolysin incultures was usually between 100 and 500 unitsper ml. Culture supernatants from which thecells had been removed by centrifugation con-tained only very low levels of alpha hemolysin(generally less than 1 unit per ml) when cellswere grown on any of the media except the alka-line meat extract broth. When cells were grownin alkaline meat extract broth, alpha hemolysinwas produced in large quantities (between 20 and50 units per ml). A typical pattern of cell growthand hemolysin production in alkaline meat ex-tract medium is shown in Fig. 1.

Conditions that affect production of alphahemolysin. The dialyzability of the substancepresent in meat broth that stimulates productionof alpha hemolysin was tested. Meat extractmedium (100 ml) was placed in a dialysis bag,and the dialysis bag was placed in 100 ml ofFraser's medium. The meat broth was allowed todialyze overnight at 4 C. The dialyzed meat

VOL. 3, 1971 679

on February 20, 2020 by guest

http://iai.asm.org/

Dow

nloaded from

SHORT AND KURTZ

500

eE

a

>%11

4.

ID

E:4

400-

3001-

801k601-

401-

20

I-

A

A -0/

#/ / A

A* _ *I~ A tI, I I I I

0 2 3 4 5Time ( hrs)

6 7

10I

E0

109 °ex

0

o -1

1O 8 E

z

._i

107 .

E

FIG. 1. Productioni of hemolysins in alkaline meatextract medium. Concentration of bacterial cells, *;alp/ia hemolysin, A; and beta hemolysin, A. Samplesfrom the whole culture were diluted with Tris-bufferedsaline for assay of beta hemolysin.

broth and the Fraser's medium (diffusate) weretransferred to Erlenmeyer flasks, inoculated withhemolytic E. coli, and incubated on the shakerbath. Cell-free hemolysin was produced in nor-mal quantities (40 units per ml) in the dialyzedmeat broth culture, but less than 1 unit per mlwas produced in Fraser's medium (diffusate). Asimilar experiment was carried out with uncookedmeat broth, and the results were the same aswith the cooked meat broth.

Lipids were extracted from muscle tissue bysequential extraction with acetone (2X), ether(3 x), boiling ethanol (1 X), and chloroform-methanol, 3:1 (3 X). Approximately 4 ml ofsolvent was used per gram of tissue in each ex-traction. The first two solvents were used toextract cholesterol and phospholipids. Both ex-tracted triglycerides. The last two solvents wereused as general extractants to remove as muchlipid as possible from the meat. No detectablelipid was obtained in the final extraction withchloroform-methanol. Each lipid fraction waspooled, evaporated to dryness, and tested for itsability to stimulate hemolysin production. Eachfraction was added to a separate flask containing50 ml of Fraser's medium which was then inocu-lated with hemolytic E. coli. The cultures weretested periodically for alpha hemolysin through-out the growth period. None of the lipid fractionsstimulated production of appreciable alphahemolysin. The meat residue from which the fathad been extracted was used to prepare meatextract broth in the usual way. This broth stimu-lated production of alpha hemolysin to about the

0

U 20 40 60 80 1OU 12O 140Elution volume (ml)

FIG. 2. Gel filtration of alpha hemolysin. -, Hemo-lytic activity;., A280. The supernatant solution ob-tained by centrifugation oJ a culture of hemolytic E.coli was the alpha hemolysin preparation.

same extent as meat extract broth prepared bythe usual procedures.

Physical and chemical properties of alphahemolysin. A sample of alpha hemolysin (3 ml)containing 27 units of activity was chromat-ographed on a column of Sephadex G-200.Fractions of 2.97 ml were collected. The ab-sorbance at 280 nm of each fraction wasmeasured, and samples were tested forhemolytic activity. The hemolytic activity wasrecovered in two major peaks (Fig. 2). Approxi-mately 80% of the hemolytic activity was re-covered. Although the peaks are not well re-solved, they consistently appeared as separatepeaks in different experiments. Because all ofthe hemolytic activity was eluted from SephadexG-100 in the void volume as a single sharp peakof activity, the molecular weight of the smallerhemolytic species is probably in excess of 3 X105.A sample of alpha hemolysin (4 ml) was also

passed through a column of Sepharose 6B.In this experiment, the hemolytic activity was

680 INFEC IMMUN .

on February 20, 2020 by guest

http://iai.asm.org/

Dow

nloaded from

HEMOLYTIC ACTIVITIES OF E. COLI

100

2- 80

I-

0

o 6C

0

E40

-240-

cr 20

, e

.4._

0 20 40 60

Concentration of calcium chloride (mM)

3 4 5 6 7 8 FIG. 4. Activity ofalpha hemolysin as a function of

pH calciwn ion concentration in the assay mixture.

FIG. 3. Solubility ofalpha hemolysin at different pHvalues.

also recovered in two major peaks (Fig. 2).The hemolysins represented by the two peaks ofactivity were eluted together from DEAE cellu-lose. They also behaved the same way underdifferent conditions of assay.

After the hemolysin had been subjected to gelfiltration and chromatography on DEAE cellu-lose, it was subjected to hydrolysis by lecithinase,lysozyme, deoxyribonuclease, ribonuclease, andtrypsin. Of the enzymes tested, only trypsininactivated the hemolysin. Other enzymes werewithout affect even at very high levels.Tubes of hemolysin were placed in an ice bath

and adjusted to pH values between 3 and 8 byadding hydrochloric acid or sodium hydroxide.Visible precipitate was present only at pH 4, 4.5,and 5.0. The tubes of hemolysin were centri-fuged at 20,000 X g for 15 min. The supernatantsolutions were collected, adjusted to the samefinal volume, and assayed by the standard pro-cedure at pH 7.5. The results are shown in Fig.3. Residual activity was recovered from theredissolved precipitates.

Conditions for hemolysis by alpha hemolysin.The effect of varying the calcium ion concen-tration on hemolytic activity is illustrated inFig. 4. Changing the erythrocyte concentrationin the assay vessel did not affect the magnitudeof the requirement for calcium. The alphahemolysin is inactive when incubated in thepresence of 10 mm EDTA. The alpha hemolysinis active over a broad range of pH, and theoptimum range for assay is between pH 7 and 8(Fig. 5). Buffers used were ammonium acetate,pH 5; histidine chloride, pH 6 and 7; Tris-

.1

.Ll

E

._-4*10

pH

FIG. 5. Activity ofalpha hemolysin as a function ofthe pH of the assay mixture.

chloride, pH 7 through 9; sodium glycinate,pH 9 and 10. The relationship between tempera-ture of incubation and hemolytic activity isshown in Fig. 6. At 60 C the erythrocytes turnbrown and lyse. When the concentration oferythrocytes in the assay mixture is increasedand the quantity of hemolysin is kept constant,the amount of hemoglobin released is pro-portional to the logarithm of the erythrocyteconcentration up to a concentration of 2 to4% red cells (Fig. 7). A concentration of 1% wasarbitrarily selected for use in routine assays.The effect of erythrocyte concentration on the

extent of hemolysis is shown in Table 1. Thedata used to prepare Fig. 7 were recalculatedto obtain the observed per cent hemolysis ateach concentration of erythrocytes. The Poissonformula (1) was used to obtain the calculated

I IoIo-0 0

l0lI,IO

0

0

0l/'

III

DO _ ___

o -!0

'0

0 I I

100-0-- ~

soo-

80~~~~~~~~0~~~~~

60 _

40

20 O

oI 5 6 7 8 9 10

I I IV I I . ---r-

VOL. 3, 1971 681

on February 20, 2020 by guest

http://iai.asm.org/

Dow

nloaded from

SHORT AND KURTZ

7

-0

I A I

20 30 40 50

TABLE 1. Variation of per cent hemolysis withconcentration of erythrocytes

Concn of erythrocytes Observed Calculated(%) hemolysis (%) hemolysis (%)

16 2 28 3 44 6 72 10 131 15 250.5 23 440.25 23 69

Temperature (IC) (10 to 100 ,M). Control vessels contained no6. Activity of alpha hemolysin as a.function of reducing agent. There was no detectable inhibi-

nperature of the assay medium. tion of hemolytic activity by any of the reducingagents. Assays were also carried out in the

, , | | Z l presence of cholesterol and lecithin (1 to 10mg/ml), and no inhibition was observed.

Direct observation of hemolysis. The course of* hemolysis of erythrocytes by alpha hemolysin

-.- ~10 was observed with a phase-contrast microscope0 in order to determine whether partial lysis of a

cell population represented complete lysis ofsome of the cells or leakage from all of the cells.For this purpose the assay was performed withsufficient hemolysin to release between 10 and30% of the hemoglobin in a suspension oferythrocytes. The preparation was observed

V/ _ through a phase-contrast microscope throughoutthe period of hemolysis. Shortly after the start ofincubation, spherocyte formation was observed.

J II | | E | The fraction of red cells that underwent sphering0.25 050 1.0 2.0 4.0 ao 16.0 increased as the incubation period continued.

Concentration of erythrocytes ( The lysing of spherocytes was always observed asan instantaneous bleaching. A large portion of

7. Activity of alpha hemolysin as a function of the red cell population appeared normal through-ncentration of erythrocytes in the assay mixture. out the incubation period.

Time course of hemolysis by alpha hemolysin.ent hemolysis at each concentration of The relationship between the rate of release ofrocytes. The assumptions were: (i) only hemoglobin and the time of incubation of rediolecule of hemolysin is required to lyse one cells with alpha hemolysin is shown for differentrocyte; (ii) the hemolysin is used up in the red cell concentrations in Fig. 8. Each vesselreaction; and (iii) the ratio of hemolysin contained 20 ml total volume of incubationules to erythrocytes in the vessel containing mixture. Cell-free hemolysin (0.8 units) waserythrocytes was sufficiently low that added to each incubation vessel after it was pre-lly none of the red cells sustained two or incubated at 40 C for 5 min. A control vessel,hits. Under these conditions 2% hemolysis without hemolysin, was prepared for each celld be produced at a multiplicity of 0.0202 concentration. The concentrations of the otherlysin molecules per red cell. substances in the incubation vessels were thetest the possibility that alpha hemolysin same as described for routine assays. The vesselsact by oxidation of some component of were incubated in a shaker bath, shaking just

111 surface and that the oxidation could be sufficiently fast to keep the erythrocytes inited by one of the common reducing agents, suspension. Fractions of 2.0 ml were withdrawns of the hemolysin were carried out in the from the assay and control vessels at differentice of mercaptoethanol (1 to 10 mM), times, transferred to prechilled tubes, and centri-bic acid (1 to 10 mM), and dithiothreitol fuged at 2,000 x g for 3 minat 4 C. The amount

100

:._'2 80U0

_. 60

E& 40

- 200)r

I-

I.

FIG.the ten

Ic

GD00

cGDGDI.

FIGthe co)

per cerythrone nerythrlyticmolec16%virtuamoreshoul4hemolTo

mightthe ccinhibiassayspreserascort

682 INFEC. IMMUN.

on February 20, 2020 by guest

http://iai.asm.org/

Dow

nloaded from

HEMOLYTIC ACTIVITIES OF E. COLI

0~~~~%

6~~~~~

7-

o

0 10 20 30 40Time (min)

FIG. 8. Time course of hemolysiscentrations of erythrocytes. The initof erythrocytes in the incubationt mi8%; *, 4%; 0, 2%;E, 1%; A, 0.5

of hemoglobin released was calcA540 of the supernatant solutionis released at an increasing rate 125 min. After about 35 min the rzhemoglobin decreases. Additionwere designed to explore the fbilities for the rapid decreasehemolysis just as the peak rate(i) The lysed cells or their conticalcium ion concentration belowby providing additional bindintinhibitor of hemolysin is release(cells. (iii) Only a fraction of the c

tible to the action of hemolysin. (isin is used up in the reaction.The following experiment waE

demonstrate whether or not addiat the end of the period of ma

activity would stimulate additiactivity. A vessel (30 ml) whicherythrocytes, 20 mm calcium chsodium chloride, and 10 mi Tr7.5), and 0.6 units of hemolysirand incubated at 40 C. Samples oftaken at 10-min intervals, andhemoglobin released was measchloride was added 50 min afterincubation to achieve a concentrfSampling and measurement (

release were continued at 10-minadditional 50 min. A control c

carried out which differed only inwas added at 50 min. The rathemoglobin followed the same pimin as shown for the 1 % cell c

Fig. 8. The addition of calcium da second round of hemolysis. Thi

,-r----i , of hemoglobin in both the control and the tubeto which calcium was added remained at lessthan 0.25 nmoles/min throughout the incubationperiod from 50 to 100 min.The possibility that lysed red cells or their

contents inhibit alpha hemolysin was tested inthe following experiment. A vessel (10 ml) wasprepared as described in the preceeding experi-

111*O0

ment except that it contained approximately 0.2unit of hemolysin. After 1 hr of incubationat 40 C, two samples of 2 ml each were taken,

<e and the amount of hemoglobin released wasmeasured. At the same time 0.12 unit of hemolysin

4 was added to the remaining 6 ml of incubation50 60 70 mixture. At the end of the second hour, two

samples of 2 ml were taken, and the amounts at various con- of hemoglobin released was measured. Duringial cotcentratsons the first hour of incubation 24 nmoles of hemo-%x;ture wer, :

. globin per ml of incubation mixture was re-leased. During the second hour an additional

ulated from the 24 nmoles of hemoglobin per ml of incubationl Hemoglobin mixture was released. In a control tube, tobetwee. n 10 and which no hemolysin was added at the end of the

ate of release of first hour of incubation, 25 nmoles of hemoglobinial experiments per ml was released by the end of the first hour,

and only 3 nmoles of hemoglobin per ml was.ollowing possi- released during the second hour of incubation.in the rate of Failure of the hemolysin to release additional@ was attained. hemoglobin during the second hour in this con-ents reduce the trol vessel indicates that it is no longer active.effective levels In addition to showing that lysed cells do notg sites. (ii) An inhibit hemolysin, this experiment also demon-d by lysing red strates that lysis does not stop because only aells are suscep- fraction of the cells is susceptible to the action ofv) The hemoly- alpha hemolysin. The second charge of hemolysin

s performed to was essentially as effective as the first. The directtion of calcium proportionality between the concentration of

hemolytic hemolysin and the amount of hemoglobin re-

,onalrked olyti leased under the usual conditions for assay pro-onal hemolytIc vides further evidence that there is not a popula-

Jcontained1% tion of cells that is markedly more susceptble toXride11 H hemolysis.is-chloride (pH The experiment shown in Table 2 was per-

f was prepared formed to examine the inhibitory capacity of thef2 mleach were red cell contents and the ghosts. Assay vesselsthe amount. of contained 1.0 ml of the added component, 0.1th tartCf the ml of hemolysin, and 1.0 ml of a 2% suspensionthe startof the of erythrocytes. One control vessel without hemo-ation of 30mo lysin was run with each added component. Theof hemoglobin vessels were incubated at 40 C. Each vessel wasexpeinerv fran removed from the water bath 60 min aftertxperiment was erythrocytes were added. The vessels were centri-e of release of fuged and the A540 of the supernatant solutionsattern out to 50 was measured. The concentrations of the lysed-oncentration in cell suspension and the cell-free lysate werelid not promote one-fourth the concentration represented in thee rate of release original vessel. This reduced the concentration of

VOL. 3,) 1971 683

on February 20, 2020 by guest

http://iai.asm.org/

Dow

nloaded from

SHORT AND KURTZ

TABLE 2. Effect of erythrocyte fractionzs onhemolysis of erythrocytes by alpha-hemolysin

Time Amt ofAdded component eryth- hemoglobinAddedcomponent rocytes liberated

added (nmoles)(min)

Lysed red cell suspension 0 20.230 9.560 4.0

Cell-free lysate 0 28.530 16.360 14.9

Ghosts 0 15.030 3.460 1.4

Tris-buffered saline with 0 25.920 mm CaCl2 30 15.0

60 11.6

these components in the assay vessels to the con-centration present in the usual assay vessel aftersufficient hemolysis has occurred to give an A540of 0.4. Thus, if lysis under usual assay conditionsreleases sufficient inhibitor to terminate hemolysis,hemolysis should have been inhibited in thesevessels. The cell-free lysate caused no inhibition.The lysed cell suspension produced a slight in-hibition of hemolysis which was enhanced bypreincubating the lysate with alpha hemolysinbefore adding the red cell suspension. The ob-served degree of inhibition may be accountedfor on the basis that the sites on the lysed cellsare able to bind hemolysin but fail to releasehemoglobin. This possibility is supported by theresults of the experiment in which red cell ghostswere added to assay vessels before and at thesame time as the red cells were added. In theseexperiments the concentration of ghosts was thesame as the concentration of red cells. When theghosts and the red cells were added to the assayvessel at the same time, 15 nmoles of hemoglobinwas liberated. In the control tube, withoutghosts, 25.9 nmoles was liberated. It can becalculated from Fig. 7 that in the vessel con-taining ghosts approximately 18 nmoles of hemo-globin should have been liberated if the ghostsand the erythrocytes had the same affinity foralpha hemolysin. Preincubation of the hemolysinwith ghosts for 30 min reduced the amount ofhemoglobin released to 3.4 nmoles upon subse-quent incubation with erythrocytes. It can becalculated from Fig. 7 and| 8 that under theseconditions approximately 6 nmoles should be

E0.6

L_.0.4

0

1/J

0

'0% 0

0%

0%0% 0

I

100

80-nC

E

6000

40_

20

o.~~~~~~~~~~~~~~~~~~~0

trI

0 10 20 30 40 50 60Time of addition of EDTA (min)

FIG. 9. Requirement for calcium during the hemo-lytic period. 0, Fraction of hemoglobin released in 60min that was not released in the time interval indicated.0, Relative rate ofrelease ofhemoglobin.

released if ghosts reacted with hemolysin at thesame rate as red cells.

Requirement for calcium during the course ofhemolysis by alpha hemolysin. The possibilitythat calcium is required for the lytic action ofalpha hemolysin only during some initial phaseof cell damage which occurs early in the incuba-tion period was tested as shown in Fig. 9. Aseries of assay vessels, volume of 2 ml, was pre-pared as usual except that the concentration ofcalcium chloride was 10 mm. At zero time 1.0ml of 0.1 M EDTA, pH 7.5, was added to thefirst tube; and at the end of each 5-min intervalEDTA was added to an additional tube. Alltubes were incubated until the end of 60 min;then they were centrifuged and the amount ofhemoglobin liberated in each was measured.Addition of EDTA at any time during the in-cubation period prevents hemolysis during theremainder of the incubation period. Thus, itappears that calcium and presumably the hemo-lysin-calcium-erythrocyte (HCE) complex mustbe maintained up to or very near to the lyticevent. Since no hemolysis occurred in tubes towhich EDTA was added at 0, 5, and 10 min, it isapparent that EDTA effectively inhibited thelytic action of the hemolysin. Comparison ofthe rate curves in Fig. 8 and 9 indicates thatremoval of calcium ion terminates lysis as effec-tively as removal of the red cells. If lysis persistedfor a time after the addition of EDTA, the ratecurve in Fig. 9 would be displaced to the left.

Attempts to dissociate beta-hemolytic activity

INFEC. IMMUN.684

on February 20, 2020 by guest

http://iai.asm.org/

Dow

nloaded from

HEMOLYTIC ACTIVITIES OF E. COLI

from the cells. The possibility was consideredthat beta- and alpha-hemolytic activity are causedby the same molcular species but differ in thatone has been dissociated from the cell surface,whereas the other remains bound. If this werethe case, then it might be possible to release thebeta hemolysin activity in the process of disrupt-ing the cell wall. To test this possibility, cellswere grown to late log phase. Assays were per-formed to quantitate hemolytic activity. Thecells were then disrupted by sonic treatment,freeze thawing, autolysis, and digestion by trypsinand lysozyme. Cell-free hemolysin was not ob-tained in the cell-free 15,000 X g supernatantsolutions of any of these preparations. Further-more, the cell-associated activity declined as thenumber of viable cells declined.The results of the above experiments suggested

the possibility that the beta-hemolytic activityis an unstable product that is liberated by theviable cell. It was also considered reasonablethat such a product might be relatively small.To test the possibility that the beta-hemolyticactivity was a diffusible metabolite, a growingculture (1 X 108 cells/ml) was placed in a dialysisbag and immersed in 25 ml of a 2% suspensionof erythrocytes in Tris-buffered saline containing20 mm calcium chloride. The preparation wasincubated at 40 C. Samples of 2 ml of the erythro-cyte suspension were taken at 30 and 60 min,and the degree of hemolysis was determined.Under these conditions there was no apparentlysis of red cells. When the procedure was modi-fied so that the erythrocyte suspension wasinside the dialysis bag and the culture was out-side, hemolysis still did not occur.

Inhibition of beta-hemolytic activity. Experi-ments were performed to assess whether agentsthat interfere with cellular metabolism affectbeta-hemolytic activity. The agents used werepotassium cyanide, 40 and 400 LM; streptomycin,200 jug/ml; rifampin, 100 ,ug/ml, and nalidixicacid, 100 ,ug/ml. The assay vessels were preparedin the usual way except that the diluted bacteriaand the interfering agent were added to theassay vessel just before the red cells.

Cells were tested in the early, middle, and lateperiods of exponential growth. All of the anti-biotic agents completely inhibited beta-hemolyticactivity. Potassium cyanide at the lower levelcaused 80 to 90% inhibition of beta-hemolyticactivity; at the higher level it caused completeinhibition. The cell-free supernatants from thecells used in the above experiments were alsoassayed in the presence of the same interferingagents. The agents did not impair the alphahemolysin. Twelve different strains of hemolytic

E. coli that were isolated in this laboratory gaveessentially the same results in this experiment.

DISCUSSIONCooked or uncooked meat broth and broth

prepared from meat homogenate after extractionof lipid contained a nondialyzable substancethat promoted the production of alpha hemolysinby certain strains of hemolytic E. coli. Lovell andRees (6) and Smith (9) have previously reportedthat cooked meat broth was required for theproduction of alpha hemolysin. Snyder andKoch using a strain of E. coli type 06 found thatdefined medium was as satisfactory as meat brothfor the production of alpha hemolysin (10).Snyder kindly supplied us with a culture of hisstrain of E. coli, and our results with his cellswere in agreement with his findings. Thus, itappears that different strains of E. coli havedifferent requirements for the production ofalpha hemolysin. However, all of the strains ofhemolytic E. coli that we have isolated from pigshave produced much higher levels of alphahemolysin when grown in meat medium than indefined medium. The role of some component inthe meat extract medium in elevating cell-freehemolytic activity is not known, but severalpossibilities come to mind. (i) It stabilizes amolecule that is produced and liberated by theE. coli. (ii) It is converted to the hemolytic sub-stance by the action of the E. coli. (iii) It stimu-lates the liberation or production of the hemolyticsubstance by E. coli.There are at least two molecular species of

alpha hemolysin. The larger species is excludedby Sepharose 6B and the smaller by SephadexG-100. The possibility that the smaller species is afragment or subunit of the larger is supportedby the observations that the hemolytic propertiesof the two are similar and the two forms are notseparated by chromatography on DEAE cellulose.Cell-free hemolysin appears to contain a proteinor polypeptide component which is essential forhemolytic activity as indicated by the inactivationof the hemolytic activity by trypsin. Snyder andZwadyk (11) recently reported that the alphahemolysin of another strain of E. coli, type 06,was inactivated by trypsin and chymotrypsin.Minimum solubility near pH 4 indicates thatalpha hemolysin is an acidic substance.The alpha hemolysin requires calcium ions

for activity. The activity increases as the concen-tration of calcium ion increases until the con-centration is approximately 10 mm. Snyder andZwadyk (11) found that 5 mm calcium ion causedmaximum activity of the alpha hemolysin pro-duced by E. coli, type 06. This difference in

VOL. 3, 1971 685

on February 20, 2020 by guest

http://iai.asm.org/

Dow

nloaded from

SHORT AND KURTZ

requirement for calcium ion and the differencein the requirement for meat extract medium forproduction of alpha hemolysin indicate that thehemolysins of various strains of E. coli are notidentical.When the concentration of alpha hemolysin is

held constant and the red cell concentration isincreased, the amount of hemoglobin that isreleased is increased (Fig. 7 and 8). This resultappears to be due to the formation of an in-creased number of productive hemolysin erythro-cyte interactions at the higher cell concentrations.The decreased number of productive interactionsat lower erythrocyte concentrations is probablydue to an increased frequency of multiple hits onsingle erythrocytes. This interpretation raises aquestion as to the number of hits by alphahemolysin that are required to lyse one red cell.The data presented in Table 1 support thehypothesis that only one molecule of hemolysinis required to lyse one erythrocyte. The valuescalculated with the Poisson formula approximatethe experimental values over a 10-fold range oferythrocyte concentrations and then diverge at thelowest concentrations. The reason for the di-vergence is not clear but it is possible that in themost dilute erythrocyte suspensions (i) the con-centration of available reacting sites is so lowthat the reaction proceeds very slowly; (ii) alarger portion of the hemolysin is inactivatedbefore it can interact with erythrocytes; and(iii) erythrocytes are less easily lysed than in theconcentrated suspensions. If more than onemolecule of hemolysin were required to lyse anerythrocyte, then the calculated per cent lysiswould increase even more rapidly than shownin Table 1.

Preliminary evidence indicates that E. colialpha hemolysin probably does not damagecells by oxidation of sulfhydryl groups. Thisconclusion is strengthened by the observationthat dissociation of the HCE complex terminateslysis immediately. At least a portion of cellsdamaged by oxidation should bfe sufficientlydamaged at any time after the prelytic periodthat lysis would continue for several minutesafter dissociation of the HCE complex.We conclude, from direct observation of cell

lysis, that the release of hemoglobin is an allor none phenomenon for each cell. The sameconclusion has been drawn for the lysis oferythrocytes by staphylococcal alpha toxin (2).The time course of hemolysis of erythrocytes

by E. coli alpha hemolysin is roughly similar tothat of staphylococcal alpha toxin. A prelyticlag period is followed by a period of rapid re-lease of hemoglobin which in turn is followedby a diminishing rate of release of hemoglobin.

When the time course curves that are generatedby the two hemolysins are inspected more closely,several important differences become apparent.The prelytic lag period has been reported to bedirectly proportional to the reciprocal of theconcentration of staphylococcal alpha toxin (2).No direct and consistent relationship between thereciprocal of the concentration of hemolysin andlag period was detected with E. coli hemolysin.Cooper et al. (2) and Marucci (7) indicated thatstaphylococcal alpha toxin causes a period oflinear release of hemoglobin which tails offasymptotically. We did not observe such a periodof linear release of hemoglobin by E. coli alphahemolysin. For this reason we have not beenable to relate the rate of hemolysis during theperiod of constant rate directly to the concen-tration of hemolysin as has been done withstaphylococcal alpha toxin.The decrease in the rate of hemolysis after

25 min of incubation suggests that some com-ponent of the HCE complex is being used upor is being inactivated by some product of thelytic reaction. The decrease in the rate of re-lease of hemoglobin is greater than can be ac-counted for by the decrease in the number oferythrocytes. For example, the rate of hemolysisdecreased by 75% between 35 and 65 min in avessel containing 2% erythrocytes, and duringthe same time the concentration of erythrocytesfell by only 5%. Furthermore, the addition of redcells to an assay mixture at the end of an hourfollowed by another hour of incubation failedto stimulate a second round of lysis. It is note-worthy that experiments of this type have givenvariable results with staphylococcal alpha toxin.Lominski and Arbuthnott (5) and Marucci (8)were able to demonstrate the persistence ofstaphylococcal alpha toxin through severalcycles of cell lysis, but Cooper et al. (3) reportedthat they could not demonstrate persistence ofalpha toxin through more than one lytic cycle.Addition of excess calcium did not stimulatelysis by alpha hemolysin. Contents of erythrocytesdid not inhibit lysis. Ghosts caused no greaterinhibition of lysis than could be accounted foron the basis that ghosts have about the sameafinity as intact erythrocytes for hemolysin.Cell-free hemolysin was the only componentthat caused a second round of hemolysis whenadded at the end of the normal incubationperiod. These data suggest that the E. coli alphahemolysin is used up in the lytic reaction and isnot an enzyme.We have observed that in the ordinary assay

the amount of hemoglobin released is directlyproportional to the amount of alpha hemolysinadded to the assay vessel up to about 30 per cent

686 INFEC. IMMUN.

on February 20, 2020 by guest

http://iai.asm.org/

Dow

nloaded from

HEMOLYTIC ACTIVITIES OF E. COLI

lysis. This is consistent with the hypothesis thatthe hemolysin is not an enzyme and only onemolecule of hemolysin is required to lyse one

erythrocyte. The per cent of cells hemolyzed bydifferent amounts of hemolysin when the con-centration of erythrocytes is kept constant canbe calculated with the Poisson formula. When thecalculated per cent hemolysis is plotted againsthemolysin concentration, the theoretical curveapproximates a straight line up to about 30%lysis.When rabbit erythrocytes are exposed to

staphylococcal alpha toxin during only a portionof the prelytic lag period, lysis still occurs; andthe later in the prelytic lag period that the alphatoxin is inactivated, the greater the ultimateextent of lysis (3, 8). If the lysin is inactivated atthe end of the prelytic lag period, lysis is notinhibited at all (8). These observations indicatethat cells damaged by alpha toxin are doomedto eventual destruction even though the alphatoxin might be inactivated. In contrast, with E.coli alpha hemolysin lysis does not occur unlessthe HCE complex is maintained up to the lyticevent.The beta-hemolytic activity appears to be differ-

ent from alpha hemolysin. No method of celldisruption that was tried liberated beta-hemolyticactivity into the medium. Growth conditionsthat markedly depressed production of alphahemolysin had little if any effect on cell-as--sociated activity. Furthermore, Smith (9) hasfound that the beta-hemolytic activity was notneutralizable by alpha hemolysin antiseru..Smith (9) also reported that some E. coli isolatedfrom intestines of different species of animalsproduce alpha but not beta hemolysin and othersproduce beta but not alpha hemolysin. The

inhibition of beta-hemolytic activity by severalcompounds that inhibit nucleic acid synthesis,protein synthesis, or aerobic metabolism isinterpreted to mean that beta-hemolytic activityis dependent on the metabolic status of theorganism.

ACKNOWLEDGMENTSThis work was supported by the Hatch Fund, GAR Fund, and

NC-62 Regional Research Funds.We gratefully acknowledge the excellent technical assistance of

Betty McKay and Marina Cheung.

LITERATURE CITED1. Adams, M. H. 1959. In Bacteriophages, p. 470. Inters6iience

Publishers, Inc., New York.2. Cooper, L. Z., M. A. Madoff, and L. Weinstein. 1964. Hemo-

lysis of rabbit erythrocytes by purified staphylococcal alpha-toxin. 1. Kinetics of the lytic reaction. J. Bacteriol. 87:127-135.

3. Cooper, L. Z., M. A. Madoff, and L. Weinstein. 1964. Hemol-ysis of rabbit erythrocytes by purified staphylococcal alpha-toxin. II. Effects of inhibitors on the hemolytic se-

quence. J. Bacteriol. 87:136-144.4. Fraser, D., and E. A. Jerrell. 1953. The amino acid composi-

tion of T3 bacteriophage. J. Biol. Chem. 205:291-295.5. Lominski, I., and J. P. Arbuthnott. 1962. Some characteristics

of Staphylococcus alpha hemolysin. J. Pathol. Bacteriol.83:515-520.

6. Lovell, R., and T. A. Rees. 1960. A filterable haemolysin fromEscherichia coli. Nature (London) 188:755-756.

7. Marucci, A. A. 1963. Mechanism of action of staphylococcalalpha-hemolysin. I. Some factors influencing the measure-

ment of alpha-hemolysin. J. Bacteriol. 86:1182-1188.8. Marucci, A. A. 1963. Mechanism of action of staphylococcal

alpha-hemolysin. 11. Analysis of the kinetic curve and in-hibition by specific antibody. J. Bacteriol. 86:1189-1195.

9. Smith, H. W. 1963. The hemolysins of Escherichia coli. J.Pathol. Bacteriol. 85:197-211.

10. Snyder, I. S., and N. A. Koch. 1966. Production and charac-teristics of hemolysins of Escherichia coli. J. Bacteriol. 91:763-767.

11. Snyder, I. S., and P. Zwadyk. 1969. Some factors affecting theproduction and assay of Escherichia coli haemolysins. J.Gen. Microbiol. 55:139-143.

12. Walton, J. R., and D. H. Smnith. 1969. New hemolysin (u)produced by Escher iclhia coli. J. Bacteriol. 98:304-305.

VOL. 3, 1971 687

on February 20, 2020 by guest

http://iai.asm.org/

Dow

nloaded from