Fisheries Science 62(4), 538-542 (1996)

Hemolysin from Edwardsiella tarda Strain ET16 Isolated from Eel Anguilla japonica Identified as a Hole-forming Toxin

Jau-Der Chent and Shiang-Long Huang

Department of Aquaculture, National Taiwan Ocean University, Keelung, Taiwan 20224,

Republic of China

(Received October 6, 1995)

Edwardsiella tarda strain ET16 isolated from the liver of diseased eel Anguilla japonica could

secret a hemolysin across cell membrane rather than cell-associated hemolysins that are reported by

most known E. tarda strains. Hemolysin exported to the culture medium was partially purified by the

application of Sephacryl S-100 HR and DEAE-Sephacel. The molecular weight of the functional

hemolysin protein was estimated to be approximately 34 kDa by the recovery of hemolytic activity in

situ after SDS-PAGE. The action of the hemolysin on tilapia erythrocytes was scrutinized under SEM.

E. tarda hemolysin caused holes between 0.5-2.5ƒÊm in diameter to be formed on the membrane of the

tilapia erythrocytes.

Key words: Edwardsiella tarda, hemolysin, hole-forming toxin

Edwardsiella tarda, a gram negative bacterium, is a common pathogen which is frequently isolated from infected cultured fishes.1,2) Histopathological studies has been made by Japanese groups on the examination of the natural infected tilapia3) and diseased farm-reared eels and it revealed that the infection by E. tarda could be classified into suppurative interstitial nephritis forma) and suppurative hepatitis form.5) Pathological activities such as adhesion or invasion were also studied but failed to conclude.6) Those E. tarda strains were found to produce hemolysins but do not produce Escherichia coli-like enterotoxin, such as proteolytic enzymes and phospholipases. However, two dermatonecrotic exotoxic substances were found.7) In addition, the fate of E. tarda after intramuscular injection of eels were also studied8) The multiplication of bacteria was observed on all organs examined in eels which were injected with a lethal dose of the virulent E. tarda strain. This finding led to conclude that the virulent E. tarda strain has the ability to resist the clearance function of the eel.

Pathogenic properties of different E. tarda strains from clinical and environmental sources were also investigated.9,10) Those E. tarda strains were invasive in HEp-2 cell monolayers, produced a cell-associated hemolysin and siderophores. The ability of E. tarda strains to penetrate and replicate in clutured epithelial cells were dependent upon cell-associated hemolysin. This cytolysin/hemolysin was responsible for the toxic effects observed in HEp-2 cells during the infection-replication process of edwardsiellae.11) Thus, the hemolysin appears to play a role in the release of internalized and replicated bacteria from infected cells.

E. tarda which isolated from diseased eels in Taiwan pro

duced the same symptoms of necrosis on liver and kidney

as reported by the Japanese group.12) Since hemolysis com

monly occurs and is accompanied by serious hemorrhagic

symptoms, the hemolysin generated by this species ap

pears to be responsible for its virulence. However, the

hemolysin of E. tarda strains which have been reported all

show as cell-associated (not filterable) hemolysin.6,9-11,13)

Here, we report that a virulent E. tarda ET16 strain which

exported hemolysin into the culture medium could be par

tially purified by Sephacryl S-100 HR and DEAE-Sephacel

chromatography. The mass of the functional hemolysin

protein was estimated to be approximately 34 kDa by the

recovery of hemolytic activity in situ after SDS-PAGE. In

addition, we show that E. tarda hemolysin functions as a

hole-forming toxin and that its action can be directly ob

served by scanning electron microscopy (SEM). Holes be

tween 0.5-2.5ƒÊm in diameter were formed on the mem

brane of tilapia erythrocytes.

Materials and Methods

Preparation of Extracellular ProductsCells of E. tarda strain ET16 (provided by Dr. H. Y.

Chung, National Taiwan University) were isolated from

the liver of diseased eel, Anguilla japonica , and the characteristics of this strain is the same as reported previously.12)

Cells had been grown in TSB (Difco) at 37•Ž to a density

of 4•~108 CFU/ml, were chilled on ice for 15 min and sub

sequently centrifuged at 8100•~g , 4•Ž for 10 min. The crude supernatants were immediately filtered by passing

them through a 0.2-ƒÊm filter membrane (Nalgene) to re

move large particles. From these, the crude culture media

containing extracellular products (ECP) were poured into

a stirred ultrafiltration cell (Amicon) equipped with a

YM10 membrane (MW 10000). Gas (nitrogen) pressure

was applied directly to the cell. Solutes greater than the

molecular weight cutoff of the YM10 membrane were

retained in the cell. The volume was duly concentrated as

•õ To whom correspondence should be addressed: Dr. Jau-Der Chen Department of Aquaculture National Taiwan Ocean University 2, Pei-Ning

Rd. Keelung, Taiwan 20224 R.O.C.

Edwardsiella tarda hemolysin 539

desired.

Hemolytic Activity AssayThe ƒÀ-hemolytic phenotype of strain ET16 has been de

fined by the appearance of a clear lytic zone on the surface

of TSB blood agar plates. Since the determination of

hemolytic activity might be affected by the blood used as a

substrate, samples of tilapia and sheep blood were com

pared by assaying under the same ondditions. Determina

tion oof hemolytic units (HU) of the crude ECP was carried

out by using sheep blood according to Buckley and

Howard.14) Tilapia blood was also used as a substrate. In

this case, the fish were immersed in ice-water for 15 min,

and a 2.5-ml syringe with a 25 gauge needle, pre-rinsed

with 10mM EDTA, was used to take the blood from the

lateral line of the tail region from the cold-anesthetized tila

pia. The erythrocyte numbers for sheep and tilapia blood were estimated with a hemacytometer; they were 7•~107

cells/ml and 2•~107 cells/ml, respectively. Crude ECP

from strain ET16 was concentrated 150-fold using a stirred

ultrafiltration cell (Amicon). The concentrated ECP (0.1

ml) was mixed with an equal volume of PBS buffer and

serially diluted. Diluted blood (1.6%) of sheep or tilapia

was then added. The hemolytic assay was carried out by in

cubation at 37•Ž for 2 h.

Effect of Temperature on Stability of the HemolysinThe hemolytic activity of the crude ECP of strain ET16

was assayed at 37•Ž, with 256 HU routinely sampled. To

determine the effect of temperature on the stability of the

hemolysin, cells of strain ET16 grown to a density of

4•~108 CFU/ml were pelleted at 15000•~g, 4•Ž for 10

min. The crude ECP were immediately filtered by passing

through a 0.2-ƒÊm filter membrane. They were divided into

aliquots, which were incubated in water baths at different

temperatures (4•Ž, 37•Ž, 50•Ž, 55•Ž and 60•Ž) for 30

min. Subsequently, 0.1ml of each treated ECP sample

was analyzed for hemolytic activity at 37•Ž for 2 h.

Chromatography and Recovery of Hemolytic Activity in Situ after SDS-PA GE Chromatography was carried out at 4•Ž. For gel filtra

tion, a C26/100 column (2.6•~100cm, Pharmacia) was

packed with Sephacryl S-100 HR and equilibrated with

PBS buffer (0.1 M sodium phosphate, pH 7.4; 0.1 M NaCl)

at a flow rate of 0.7ml/min. Two ml (22 mg) of the crude

concentrated culture media, which contained hemolysin,

was then loaded onto the bed of Sephacryl S-100 HR for

separation according to size. The collected 80 fractions (6

MI/tube) were immediately used in the hemolytic assay to

determine the position of the hemolysin. The eluents pos

sessing the highest titer of hemolytic activity were subse

quently loaded on the bed of DEAE-Sephacel in a C16/40

column (1.6•~40cm, Pharmacia). They were eluted by a

0.1mM to 4 M NaCl gradient at a flow rate of 0.7ml/min.

Again, 80 fractions (3ml/tube) were collected and immedi

ately used to determine the position of the hemolysin.

From the eluents of fractions 9 and 10 which were collect

ed after DEAE-Sephacel chromatography, 2.5ml were

concentrated to 80ƒÊl in a centricon 10 concentrator (cutoff

MW 10000, Amicon). Concentrated ECP (50•~) of strain

ET16 was used as a control. Aliquots of the concentrated

samples were boiled in a SDS sample buffer for 3 minutes

and loaded into the wells of two parallel 10% SDS-PA

gels. One gel was stained with Coomassie brilliant blue R-

250 to examine the bands of polypeptides. The other gel,

in order to renature the trapped polypeptides within the

gel, was soaked in several changes of PBS to remove the

SDS.15) Meanwhile, a 1% agarose gel plate of the same size

as that of the renatured gel and containing 2% defibrinat

ed sheep blood was prepared. The renatured gel was gently

pressed onto this blood agarose gel plate, wrapped with a

plastic membrane and incubated at 37•Ž for 18 h.

Preparation of Hemolysin-Treated Erythrocytes for SEMA 0.2ml samples of tilapia blood was washed twice with

1 ml PBS buffer and then suspended in 1ml PBS buffer.

To this suspension was added 0.1ml of the eluents (256

HU) of fraction 9 which contained the hemolysin that had

been collected after DEAE-Sephacel chromatography.

The mixture was incubated at 37•Ž and after 2, 5 and 10

min, 10 ƒÊ1 of the reaction mixtures were dropped onto a

1.6% poly-L-lysine hydrobromide (Sigma, MW 150000

300000) coating surface. The reaction was stopped immedi

ately by adding the fixative (3.3% paraformaldehyde; 5%

glutaraldehyde). After I h of fixation, the samples were

washed with PBS buffer and fixed by 1% osmium

tetraoxide. The subsequent standard procedures consisted

of dehydration, CPD drying, gold coating and SEM

photography and were carried out by the EM center of

National Taiwan Ocean University.

Results

Species Variation of Red Blood Cell SensitivityResults are shown in Fig. 1. A hemolytic titer of 1024

HU was determined with the sheep blood as a substrate, whereas a titer of 16384 HU was obtained with tilapia blood as substrate. This result indicates that the blood of tilapia is more sensitive than that of sheep to attack by the hemolysin. In any case, high values of hemolytic activity could be measured by using blood from either source.

Effect of Temperature on Stability of the HemolysinResults are shown in Fig. 2. There is no difference in

hemolytic activity when the ECP are held at 4•Ž or 37•Ž.

Loss of hemolytic activity was observed at a temperature

of 55•Ž; the hemolytic activity of the ECP was reduced to

32 HU. Heating the ECP to 60•Ž caused the hemolytic ac

tivity to become negligible. However, the hemolytic activi

Fig. 1. Determination of hemolytic activity.

The numbers above the figure represent the serial order of dilution

of each sample in the first and third rows; add 10 for the second and fourth rows.

540 Chen and Huang

ty of the ECP could be retained at low temperatures.

When assays were done after 14 days storage at 4•Ž and

60 days storage at -20•Ž, in both cases, 128 HU were

found.

Characterization of the Functional HemolysinThe hemolytic activity of 80 eluents which were collect

ed after DEAE-Sephacel chromatography were determined. Only fractions 9 (256 HU), 10 (256 HU), 11 (64 HU), and 12 (32 HU) exhibited hemolytic activity. To determine the size of the hemolysin, the samples were concentrated and run in two parallel 10% polyacrylamide gels. As shown in Fig. 3, the ECP of strain ET16 gave a clear hemo

Fig. 2. Effect of temperature on hemolytic activities of the extracellular

products from strain ET16.

lytic band in the region of 34 kDa. Clear hemolytic bands were also observed at 34 kDa from the samples of fractions 9 and 10 which were collected after DEAE-Sephacel chromatography. These two eluents evidently contained functional E. tarda hemolysin and the MW of the hemolysin was estimated to be approximately 34 kDa.

Hole-forming activity of E. tarda hemolysinThe action of E. tarda hemolysin on the membrane of

tilapia erythrocytes was immediately to be seen (Fig. 4).

The erythrocytes were severely deformed after 2 minutes'

treatment with hemolysin. In contrast to the normal

erythrocytes (Fig. 4A), the membrane of the erythrocytes

was invaginated, probably due to leakage (Fig. 4B). Holes

between 0.5-2.5ƒÊm in diameter (Figs. 4B and 4C) were

found randomly on the surface membrane of the erythro

cytes. As seen in Fig. 4D, this hole-forming action of the

hemolysin evidently causes leakage of the contents from

the erythrocytes. After 5 min mixing with the hemolysin,

the tilapia erythrocytes were almost completely deformed.

Discussion

Aeromonas hydrophila, which releases a cytolytic toxin

called aerolysin,16-22) is also a common pathogen and is

often isolated from infected farmed-fishes in aquacul

ture.23-26) Aerolysin's mechanism of hole formation has been

studied in some detail.271 It has been shown to bind to

erythrocytes using glycophorin as a receptor. It then ag

gregates, forming 3 nm holes in the membrane. In this

report, we show by direct observation that larger holes,

0.5-2.5ƒÊm in diameter, were formed by the action of E.

tarda hemolysin on the membrane of tilapia erythrocytes.

The mechanism of this hole formation remains unknown.

By SDS-PAGE, the mixture is boiled in SDS sample buffer for 3min before loading on to the gel in order to denature proteins. Despite of the hemolytic activity is heat-labile as shown in Fig. 2, the mixture after a short ex

posure by heat still maintain a low level of hemolytic activity and it is enough to form a hemolytic band in situ after SDS-PAGE. The hemolytic band was found on SDS

- PAGE to have a mass of 34 kDa, but since additional stained polypeptide bands were seen (see Fig. 3), other protein functions cannot be excluded. The possibility of other protein functions besides hole-forming, such as directly

Fig. 3. SDS-PAGE and the recovery of hemolytic activity in situ in the renatured gel.

Edwardsielta tarda hemolysin 541

Fig. 4. Formation of hole(s) on the membrane of tilapia erythrocytes by the activity of E. tarda hemolysin.

(A) Tilapia erythrocytes were not exposed to the hemolysin. (B) and (C) Samples mixing with hemolysin for 2 min were stopped by the fixative.

Holes between 0.5-2.5 ƒÊm in diameter are shown. (D) The contents of an erythrocyte are seen in the act of leaking outwards.

degrading cell membrane components2s,29) or altering the membrane lipids as a result of membrane rupture by osmotic pressure30) will require further investigation.

Acknowledgments We thank Miss. G. D. Huang and Mr. J. G. Chang

for their help in SEM preparation. This work was supported by National

Science Council , Republic of China grant NSC 84-2321-B-019-021 and NSC 85-2321-B-019-030.

References

1) J. D. Chen and G. H. Kou: Studies on bacterial distribution in

pond-cultured eels. COA Fish., 10, 25-42 (1987).2) C. Y. Liu and M. C. Chien: Studies on pathogenesis of edwardsiello

sis in experimentally infected eels. Fish Dis. Res., 7, 68-78 (1986).3) S. S. Kubota, N. Kaige, T. Miyazaki, and T. Miyashita: Histopatho

logical studies on edwardsiellosis of tilapia-I. Bull. Fac. Fish. Mie

Univ., 9, 155-165 (1981).4) T. Miyazaki and S. Egusa: Histopathological studies of edwardsiel

losis of the Japanese eel (Anguilla japonica)-I. suppurative intersti

tial nephritis form. Fish Pathol., 11(1), 33-43 (1976).

5) T. Miyazaki and S. Egusa: Histopathological studies of edwardsiel

losis of the Japanese eel (Anguilla japonica)-‡U. suppurative hepati

tis form. Fish Pathol., 11(2), 67-75 (1976).

6) M. Ullah and T. Arai: Pathological activities of naturally occurring

strains of Edwardsiella tarda. Fish Pathol., 18(2), 65-70 (1983).

7) M. Ullah and T. Arai: Exotoxic substances produced by Edwardsiel

la tarda. Fish Pathol., 18(2), 71-75 (1983).

8) R. Kusuda and S. Ishihara: The fate of Edwardsiella tarda bacteria

after intramuscular injection of eels, Anguilla japonica. Nippon Sui

son Gakkaishi., 47, 475-479 (1981).

9) J. M. Janda and S. L. Abbott: Expression of an iron-regulated

hemolysin by Edwardsiella tarda. FEMS Microbiol. L., 111, 275

280(1993).

10) J. M. Janda, S. L. Abbott, S. Kroske-Bystrom, W. K. W. Cheung,

C. Powers, R. P. Kokka, and K. Tamura: Pathogenic properties of

Edwardsiella species. J. Clin. Microbiol., 29, 1997-2001 (1991).

11) J. M. Janda, S. L. Abbott, and L. S. Oshiro: Penetration and repli

cation of Edwardsielta spp. in HEp-2 cells. Infect. Immun., 59, 154

542 Chen and Huang

161 (1991).

12) S. C. Kuo, H. Y. Chung, and G. H. Kou: Edwardsiella anguillimortiferum isolated from edwardsiellosis of cultured eel (Anguilla

japonica). JCRR Fish., 29, 1-6 (1977).13) J. J. Watson and F. H. White: Hemolysins of Edwardsiella tarda.

Can J. Comp. Med., 43, 78-83 (1979).14) J. T. Buckley and S. P. Howard: Aerolysin from Aeromonas

hydrophila. Method Enzymol., 165, 193-199 (1988).15) A. L. Rosenthal and S. A. Lacks: Nuclease detection in SDS-poly

acrylamide gel electrophoresis. Anal. Biochem., 80, 76-90 (1977).16) T. Chakraborty, B. Huhle, H. Bergbauer, and W. Goebel: Cloning,

expression, and mapping of the Aeromonas hydrophila aerolysin

gene determinant in Escherichia coli K-12. J. Bacteriol., 167, 368-374(1986).

17) J. S. G. Dooley, R. Lallier, D. H. Shaw , and T. J. Trust: Electrophoretic and immunochemical analyses of the lipopolysaccharides from various strains of Aeromonas hydrophila. J. Bacteriol. , 164, 263-269 (1985).

18) J. S. G. Dooley and T. J. Trust: Sruface protein composition of

Aeromonas hydrophila strains virulent for fish: Identification of a surface array protein. J. Bacteriol., 170, 499-506 (1988).

19) S. P. Howard and J. T. Buckley: Protein export by a gram-negative

bacterium: Production of aerolysin by Aeromonas hydrophila. J. Bacteriol., 161, 1118-1124 (1985a).

20) S. P. Howard and J. T. Buckley: Activation of the hole-forming aer

olysin by extracellular processing. J. Bacteriol., 163, 336-340

(1985b).21) S. P. Howard and J. T. Buckley: Molecular cloning and expression

in Escherichia coli of the structural gene for the hemolytic toxin aer

olysin from Aeromonas hydrophila. Mol. Gen. Genet., 204, 289

295 (1986).

22) S. P. Howard, W. J. Garland, M. H. Green, and J. T. Buckley:

Nucleotide sequence of the gene for the hole-forming toxin aeroly

sin of Aeromonas hydrophila. J. Bacteriol., 169, 2869-2871 (1987).

23) S. Egusa: On the motile Aeromonads. Fish Pathol., 2,36-49 (1967).

24) G. H. Kou: Studies on the pathogenecity of the fish pathogen, Aero

monas liquefaciens-‡V. Variation of pathogenecity caused by inocu

lation into fish. J. Fish. Soc. Taiwan, 2, 16-19 (1973).

25) H. Wakabayashi, K. Kanai, T. C. Hsu, and S. Egusa: Pathogenic

activities of Aeromonas hydrophila biovar hydrophila (Chester)

Popoff and Veron, 1976 to fishes. Fish Pathol., 15, 319-325 (1981).

26) H. Y. Chung and G. H. Kou: Studies on the change in number of

cells of fish pathogen, Aeromonas hydrophila inoculated intramus

cularly in eels. J. Fish. Soc. Taiwan, 3, 15-20 (1974).

27) S. P. Howard and J. T. Buckley: Membrane glycoprotein receptor

and hole-forming properties of a cytolytic protein toxin. Biochemis

try, 21, 1662-1667 (1982).

28) A. W. Bernheimer and B. Rudy: Interactions between membranes

and cytolytic peptides. Biochim. Biophys. Acta., 864, 123-141

(1986).

29) F. Hugo, J. Reichwein, M. Arvand, S. Kramer, and S. Bhakdi: Use

of a monoclonal antibody to determine the mode of transmem

brane pore formation by streptolysin O. Infect. Immun., 54, 641

645(1986).

30) J. E. Alouf: Streptococcal toxins (streptolysin O, streptolysin S,

erythrogenic toxin) Pharmacol. Ther., 11, 661-717 (1980).