Production of a unique cytotoxin by Klebsiella oxytoca

Microbial Pathogenesis 1989 ; 7 : 203-211

Production of a unique cytotoxin by Klebsiella oxytoca Junzaburo Minami,' Akinobu Okabe,' Junji Shiode 2 and Hideo Hayashi'* 'Department of Microbiology, Kagawa Medical School, 1750-1 Ikenobe, Miki-cho, Kita-gun, Kagawa 761-07, Japan and 2Department of Bacteriology, Okayama University Medical School, Shikata-cho, Okayama 700, Japan

(Received May 2, 1989; accepted in revised form June 20, 1989)

Minami, J . (Dept of Microbiology, Kagawa Medical School, 1750-1 lkenobe, Miki-cho, Kitagun, Kagawa 761-07, Japan), A . Okabe, J . Shiode and H . Hayashi . Production of a unique cytotoxin by Klebsiella oxytoca . Microbial Pathogenesis 1989; 7 : 203-211 . Certain strains of Klebsiella oxytoca isolated from patients with hemorrhagic enterocolitis produced a unique cytotoxin . The cytotoxin induced rounding of tissue culture cells, such as HEp-2, Vero, CHO and HeLa cells . The induced morphologic changes were indistinguishable between cell types . Seventy to 80% of the rounded cells died in 48 h incubation . The cytotoxin was purified 1000-fold from culture supernatant by Sephadex G-25 and Bio-Gel P-2 gel filtration followed by reversed-phase high-performance liquid chromatography . The molecular weight of the purified cytotoxin was estimated to be <651 by mass spectrometry . The minimum concentrations of the purified cytotoxin required to cause 50% of rounding of cells were 0 .6 pg/ml for HEp-2, 0.8 pg/ml for Vero, .8 0 pg/ml for CHO and 1 .4 pg/ml for HeLa cells. The type strain of K. oxytoca, ATCC 13182, did not produce the cytotoxin and only the clinically isolated strains did, suggesting that the cytotoxin may play a role in the pathogenesis of the organisms . Key words : cytotox

'e

a; hemorrhagic enterocolitis .

Introduction There have been many reported cases of acute enterocolitis associated with the administration of antimicrobial agents . Etiological studies of such enterocolitis suggested that there may be some involvement of bacterial infection in conjunction with the therapeutic administration of antimicrobial agents . For example, pseudomembranous colitis, which is induced by the administration of clindamycin or lincomycin, is caused by abnormal proliferation of Clostridium difficile in the human colon ." In some cases of acute hemorrhagic enterocolitis, ampicillin or other penicillin derivatives have been suspected to be the inducible agents" and the association of a Klebsiella oxytoca infection has been speculated .' Chida et a1. 5 reported that, in 11 out of 11 cases studied, K. oxytoca was isolated on an average of 10' bacterial cells per g of faeces from the patients with hemorrhagic diarrhea after receiving penicillin derivatives. Although K. oxytoca was suspected as a causative organism of the enterocolitis, its etiological role has not yet been well elucidated . In this report, we showed that clinically isolated strains of K. oxytoca from patients with hemorrhagic enterocolitis produced a unique cytotoxin in the culture supernatant, whereas the type strains of K .

Author to whom correspondence should be addressed . 0882-4010/89/090203+09 $03 .00/0

° 1989 Academic Press Limited



2 04

J . Minami et al.

oxytoca, ATCC 13182, did not . The cytotoxin was purified 1000-fold and some properties of the cytotoxin were examined .

Results Cytotoxicity of the culture supernatant of K. oxytoca OK-1 The culture supernatant of K. oxytoca OK-1, a clinically isolated strain, was examined on monolayered HEp-2 cells for the cytotoxicity and the result is shown in Fig . 1(b) . The effects were in sequential changes, first causing rounding of cells, then detaching of the cells from culture plate and eventual death of the cells, as were observed under phase-contrast microscopy . Those changes were detectable within 24 h and reached a maximum in 48 h incubation . The mortality of the rounded cells measured by trypan blue exclusion was 70 to 80% after 48 h incubation . The culture supernatant of K. oxytoca ATCC 13182, a type strain of K. oxytoca, in contrast, did not show such cytotoxic effects on the tissue culture cells . Therefore the cytotoxic effects of the culture supernatant suspected to be due to a specific substance produced by K. oxytoca OK-1 . Relationship between cytotoxin production and bacterial growth phase Production of the cytotoxin by K. oxytoca OK-1 was analyzed in relation to bacterial growth phase (Fig . 2) . Cytotoxic activity was detectable in the culture supernatant of K. oxytoca OK-1 at the late exponential to early stationary growth phase . The cytotoxic activity was maximum at the early stationary growth phase, and then decreased along with cell cultivation . It was no longer detectable in the culture supernatant of 48 h culture . The cytotoxin tends to be produced at a limited growth phase and seems to be inactivated or degraded during further cultivation . Other clinically isolated strains, K. oxytoca KA-1 and KA-2, showed the same properties as K. oxytoca OK-1 both in the kinetics of production of cytotoxin and its cytotoxicity on HEp-2 cells (data not shown) . The cytotoxic activities of culture supernatants obtained from 12 h cultures of K. oxytoca KA-1 and KA-2 were both 240 CD 50/ml . As shown in Fig . 2, K. oxytoca ATCC 13182 did not produce cytotoxin at any growth phase . Purification of the cytotoxin The cytotoxin produced by K . oxytoca OK-1 was purified by gel filtration and reversed-phase high -performance liquid chromatography (RP-HPLC) . The purification procedures are summarized in Table 1 . The all purification steps were performed at 4°C except RP-HPLC step which was carried out at room temperature . The culture supernatant of K. oxytoca OK-1 was subjected to gel filtration on Sephadex G-25 (Fig . 3) and then on Bio-Gel P-2 (Fig . 4) . By these procedures, the specific activity of the cytotoxin increased about 300-fold, with 24 .5% recovery . The elution profile of gel filtration in Fig . 4 was striking, in that the cytotoxin was eluted after NaN 3 which was used as a molecular marker . Final purification was achieved by RP-HPLC of the Bio-Gel P-2 fractions . Figure 5 shows the elution profile of RP-HPLC . The elution was carried out by using a gradient of acetonitrile in H 2 O . The major peak eluted by 18% (v/v) acetonitrile had cytotoxic activity which was measured after lyphilization of the effluent . The gradient of acetonitrile was sharply increased from 25 to 100% during elution of the terminal 15 fractions, because cytotoxic activity was not detected in these fractions . The specific activity of the cytotoxin increased 1000-fold and the recovery was 8 .3% of the original culture supernatant . Bacterial medium (Trypto-Soya broth) and the culture supernatant obtained from 12 h culture of K. oxytoca ATCC 13182, as well as the culture supernatant of

205

Fig . 1 . Phase-contrast photograph of HEp-2 cells incubated with or without cytotoxin of K. oxytoca OK1 . (a) HEp-2 cells control cultured for 48 h in tissue culture medium (MEM containing 5% of heatinactivated fetal bovine serum) . (b) HEp-2 cells cultured for 48 h in tissue culture medium containing 20 pl of a culture supernatant of K. oxytoca OK-1 . (c) HEp-2 cells cultured for 48 h in tissue culture medium containing 1 pg of H PLC-purified cytotoxin . Culture supernatant was obtained from the 12 h culture of K . oxytoca OK-1 (9 .9x10 9 cells/ml) . HEp-2 cells were cultured in Lab-Tek Tissue Culture Chamber/Slide (Miles Scientific, Naperville) in 800 µl of tissue culture medium . HEp-2 cells precultured in MEM containing 10% of heat-inactivated fetal bovine serum, and then the medium was replaced by fresh tissue culture medium with or without cytotoxin . Magnification is 100 diameters .



J . Minami et al.

2 06

500

E 0 U

400

T

0 O

300

X 0 0

200

100

LAA&M

IOTA

10

0

I

A[ APAu AIA I

20

30

40

50

1

1 1, 40

60

70

Incubation time (h) Fig . 2 . The relationship between cytotoxin production and bacterial growth phase . 20 h precultures of K. oxytoca ATCC 13182 and OK-1 in Trypto-Soya broth were inoculated into fresh Trypto-Soya broth at 1 % (v/v) inoculum size and then cultured at 37'C, 120 rpm . Samples were taken at the indicated times and analyzed for the cytotoxic effect on H Ep-2 cells . A, growth of the K. oxytoca ATCC 13182; o, growth of the K. oxytoca OK-1 ; A, cytotoxic activity of the culture supernatant of K. oxytoca ATCC 13182 ; •, cytotoxic activity of the culture supernatant of K . oxytoca OK-1 .

K. oxytoca OK-1, were concentrated 10-fold by lyophilization and then subjected to gel filtration on Sephadex G-25 . Cytotoxic activity was not detected in either of the Sephadex G-25 fractionations . Characteristics of the purified cytotoxin The effect of the purified cytotoxin on HEp-2 cells is shown in Fig . 1(c) . It was the same as shown in culture supernatant in terms of sequential morphologic changes . The cytotoxin also affected Vero, CHO and HeLa cells in the same manner . The minimum concentrations required to cause 50% of rounding of cells were 0 .6 tg/ml Table 1

Summary of purification of K. oxytoca cytotoxin

Step Culture supernatant' 100 ml Sephadex G-25 Bio-Gel P-2 RP-HPLC

Dry weight (mg)

1 CD50` (pg)

Purification (-fold)

Recovery (%)

2400 (780) b

150

100 .0

14 .6 1 .8 (1 .9) n 0 .2

1 .4 0 .46

107 326

65 .2 24 .5

0 .15

1000

8 .3

' Culture supernatant obtained from the 12 h culture of K . oxytoca OK1 in Trypto-Soya broth at 37°C . The bacterial concentration of the culture was 7 .0 x 109 /ml . ' Protein assayed by the method of Lowry et al .,' with bovine serum albumin as a standard . `Cytotoxic dose determined as described under Materials and methods . CD 50 was expressed as the dry weight of sample required to cause 50% of rounding of cells .



K. oxytoca cytotoxin

207 4 .0

E C 3 .0 0 W N m C 2 .0 a0 N Q I •0

50

100 150 200 Fraction number (3 ml/tube)

250

Fig . 3 . Sephadex G-25 gel filtration chromatography of cytotoxin . 10x concentrated culture supernatant of K. oxytoca OK-1 after 12 h incubation was chromatographed . The effluent was monitored at 280 nm . Fractions marked with the heavy bar had cytotoxic activity .

for HEp-2, 0 .8 µg/ml for Vero, 0 .8 ug/ml for CHO and 1 .4 yg/ml for HeLa cells . The morphologic changes induced by the purified cytotoxin were indistinguishable from one cell type to another (data not shown) . The molecular weight of the cytotoxin could not be measured by gel filtration . Mass spectrometry was used to determine the molecular weight . Figure 6 shows the mass spectrum of the cytotoxin purified by RP-HPLC . It showed molecular ion peak at m/z 651 . It cannot be concluded that the molecular weight of the cytotoxin was 651, because the purity of cytotoxin was not 100% . However, the mass spectrum indicated that the molecular weight of the cytotoxin was less than 651 . This result was consistent with the fact that the cytotoxin passed through Spectra/Por 6 (molecular weight cutoff of 1000) dialysis membrane tubing (Spectrum Medical Industries, Los Angeles) .

3 .0 CE 13

a bcd I III 2 .0

50

100 150 200 Fraction number (3 ml/tube)

Fig . 4 . Bio-Gel P-2 gel filtration chromatography of cytotoxin . Cytotoxin-containing fractions from the Sephadex G-25 gel filtration were pooled, lyophilized, and dissolved in 3 .0 ml water and loaded onto BioGel p-2 column (2 .2 by 75 cm) . The effluent was monitored at 280 nm . Fractions marked with the heavy bar had cytotoxic activity . Arrows in figure indicate the positions of molecular markers elution : a, bovine serum albumin ; b, vitamin Bt2; c, glutathione, disulfide ; d, glutathione, reduced ; e, NaN3 . The elution of molecular markers was monitored at 230 nm for glutathione, reduced and at 280 nm for the others .



a J . Minami et al.

20 8 1 . 00

100 II II II I I 1 I I I I I

0 .75 E C 0 K) N

I I I

0 . 50

50

I I I

I

I

/I

0

75

25

1 I 1

0 0 .25

l

1 I i I 0 30 60



90

C 0

m

I

N Q

0 4)

120

Fraction number (I ml/tube)

Fig . 5 . RP-HPLC of cytotoxin . Lyophilized fraction from Bio-Gel P-2 gel filtration step was dissolved in 150µl of H20 and 75 µl of the dissolved sample was loaded onto C-18 column (7 .8 by 300 mm) equilibrated in H20 . A linear gradient (---) of acetonitrile used was : 0 to 25% in 30 min, 25 to 100% in 5 min, 100 to 0% in 5 min . The flow rate was 3 .0 ml/min, and one-ml fractions were collected. The effluent was monitored at 230 nm . Fractions marked with the heavy bar had cytotoxic activity . Cytotoxic activity was measured after lyophilization of the effluent .

100

149

-10

8 0 T C 4, C 4)

50

5

18

0 a)

429

Gr 355 71 223

IiIM .

0

100

200

300

{'

400 m/z

SU3 651 .+,,,, . . .,,,,~ 0

500

600

700

800

Fig . 6 . Mass spectrum of HPLC-purified cytotoxin . The mass spectrometry conditions are described in the Materials and methods section . The arrow in figure indicates the position (m/z 300) at which the relative intensity was scaled up 10 times .



K. oxytoca cytotoxin

209

Discussion Cytotoxin production by enteric pathogens has been increasingly investigated in recent years . Shigella sp .," Escherichia coli,` ° Clostridium difficile,' 1 .12 Campylobacter jejuni 13 and Salmonella sp .11 18 produce cytotoxins . These cytotoxins are mostly proteins and have high molecular weights more than 10 kDa . Especially Shiga toxin and Shiga-like toxins (SILT-1, SLT-11) are well characterized . Shiga toxin inhibits protein synthesis in eukaryotic cells, but it is effective only on HeLa and Vero cells . On the other hand, K. oxytoca cytotoxin seems to be a very low molecular weight substance and less specific for cell types in the cytotoxicity than Shiga toxin . It affected four unrelated mammalian cell lines similarly and induced rounding of the cells . By phase-contrast microscopy, these changes were indistinguishable between cell types . Similar morphologic changes in mammalian cells were observed by C . difficile cytotoxin, which induced rounding of CHO, HeLa and Y1 cells ." Although the morphologic changes induced by both cytotoxins are similar to each other, the mechanism for the cytotoxic activity may be different . The molecular weight differed from one another . K. oxytoca cytotoxin, which is a low molecular weight substance and causes the rounding followed by death in tissue-cultured mammalian cells, seems to be a unique cytotoxin differing from cytotoxins reported in members of Enterobacteriacea . K. oxytoca cytotoxin was purified 1000-fold from culture supernatant, and caused rounding of CHO cells at >,0 .8 pg/ml . C. difficile toxin A and B caused rounding of CHO cells at >0 .5 pg/ml and >0 .2 ng/ml, respectively .12 Therefore the specific activity of K . oxytoca cytotoxin is similar to that of C . difficile toxin A . The role of the cytotoxin in the pathogenesis of K . oxytoca or in correlation with hemorrhagic enterocolitis has not been clarified as yet . But the fact that K . oxytoca cytotoxin was produced only by clinically isolated strains and not by the type strain (ATCC 13182) suggests that cytotoxin production may play a role in the pathogenesis . It is probable that the cytotoxin plays a role in the local damage to intestinal mucosa to cause hemorrhagic enterocolitis . The full elucidation of the mechanism for cytotoxic activity and its role in pathogenesis remains to be determined . Materials and methods Organisms. Three clinically isolated strains of K. oxytoca (OK-1, KA-1 and KA-2) and the type strain K. oxytoca, ATCC 13182 (American Type Culture Collection, Rockville), were used in this study . K. oxytoca OK-1, KA-1, and KA-2 were isolated from patients with hemorrhagic diarrhea after administration of penicillin derivatives for common cold ; K. oxytoca OK-1, from a 38-year-old patient taken amoxillin (1 g/day) for 3 days, K . oxytoca KA-1 from a 21-year-old patient taken amoxillin (1 g/day) for 6 days, and K . oxytoca KA-2 from a 52-year-old patient taken ampicillin (1 g/day) for 3 days . Purification of cytotoxin . K. oxytoca OK-1 was precultured in Trypto-Soya broth (17 .0 g of peptone, 3 .0 g of Soya peptone, 5 .0 g of NaCl, 2 .5 g of dextrose, and 2 .5 g K2HPO4 in one liter, pH 7 .3, Nissui Pharmaceutical Co ., Tokyo) at 37°C for 20 h with shaking at 120 rpm . The precultured cells were inoculated into Trypto-Soya broth at 1% (v/v) and cultured at 37°C for 12 h with shaking at 120 rpm . The final pH of the cultured medium was 7 .5 to 8 .2 . The pH of the cultured medium was adjusted to 7 .4 with HCI and filtrated through 0 .2 pm membrane filter and used as the culture supernatant . The culture supernatant (100 ml) was lyophilized and then dissolved in 10 ml of H20 and fractionated on a 2 .2 by 85 cm Sephadex G-25 column (Pharmacia, Uppsala) equilibrated in 0 .02 M ammonium acetate (pH 7 .0) . The column effluent was monitored at 280 nm, and cytotoxin-containing fractions were identified by their cytotoxic effect on HEp-2 cells . Fractions containing cytotoxin were pooled, lyophilized, dissolved in 3 .0 ml of H20 and then loaded onto 2 .2 by 75 cm Bio-Gel P-2 column (200-400 mesh, Bio-Rad Laboratories, Richmond) equilibrated in 0 .02 M ammonium acetate (pH 7 .0) . Cytotoxincontaining fractions eluted from Bio-Gel P-2 column were pooled, lyophilized, dissolved in 150 pl of H20 and further purified by reversed-phase high-performance liquid chromatography



210

J . Minami et alt

on a C-18 column (TSK-gel ODS-120 T, 7 .8 by 300 mm, Toyo Soda Manufacturing Co ., Tokyo) . Sample was loaded onto the column equilibrated in H 2 0 . A linear gradient of acetonitrile was used for elution . 1 ml fractions were collected . The fractions eluted by 0 to 20% acetonitrile were directly lyophilized . The fractions eluted by 20 to 100% acetonitrile were concentrated by centrifugal concentrator (Toyo Scientific Industrial Co ., Tokyo) at room temperature before lyophilization .

Assay for cytotoxic activity. HEp-2 cells were cultured in Eagle minimum essential medium with Earl salt, penicillin (100 U/ml) and streptomycin (100 pg/ml) (MEM), supplemented with 10% heat-inactivated fetal bovine serum at 37°C in a 5% CO 2 atmosphere . Freshly trypsinized cells were cultured in 96-well microculture plates for 24 h to be monolayered . The old medium in each of the wells was replaced by 200 yl of freshly prepared MEM containing 5% heatinactivated fetal bovine serum and 50 pl of serial dilution of fractions containing cytotoxin with PBS (Dulbecco's phosphate buffered saline without Mgt' and Ca 21, pH 7 .4) . Culture supernatant was diluted at least two-fold with PBS for cytotoxic assay . The plates were further incubated for 48 h at 37°C in a 5% CO2 atmosphere . After the incubation, the plates were washed three times in PBS with shaking by Tray-mixer (Fujirebio, Tokyo) . By the washing, affected or killed cells were detached from the bottom surface of the wells . The remaining intact cells were dissolved in 18 pl of 1 M NaOH solution and the protein concentration of the solution in each well was measured by the method of Lowry et a1. 9 To each well, 180 pl of the solution containing 2% Na 2 CO 3, 0.01% CuSO 4 and 0 .02% sodium potassium tartrate was added . After 10 min incubation at room temperature with gentle shaking, 18 pl of Phenol reagent was added and the plate was kept at room temperature for 1 h with shaking . The absorbance was measured at 750 nm in a Titertek Multiskan spectrophotometer (Flow Laboratories, McLean) . The cytotoxic dose required to reduce the absorbance by 50% of control (CD 50) was determined . Mass spectrometry . Mass spectrometry was carried out with Hitachi M-80 mass spectrometer (Hitachi, Tokyo) . Ionization energy was 20 eV and ion source temperature was 200°C .

We thank Mr S . Katayama for useful discussion . This work was supported by a Grant-in-Aid for Scientific Research (B62480156) from the Ministry of Education, Science and Culture of Japan .

References 1 . Bartlett JG, Chang TW, Gurwith M, Gorbach SL, Onderdonk AB . Antibiotic-associated pseudomembranous colitis due to toxin-producing Clostridia . N Engl J Med 1978 ; 298 : 531-4 . 2 . George RH, Symonds JM, Dimock F et al. Identification of Clostridium difficile as a cause of pseudomembranous colitis . Br Med J 1978 ; 1 : 695 . 3 . Larson HE, Price AB, Honour P, Borriello SP . Clostridium difficile and the aetiology of pseudomembranous colitis . Lancet 1978; is 1063-6 . 4 . Lusk RH, Fekety R, Silva J, Browne RA, Ringler DH, Abrams GD . Clindamycin-induced enterocolitis in hamsters . J Infect Dis 1978 ; 137 : 464-75 . 5 . Chida T, Nakaya R, Tsuji M et aL Intestinal microflora of patients with antibiotic-associated hemorrhagic colitis associated with Klebsiella oxytoca and Clostridium difficile enterotoxin . Kansenshogaku Zasshi 1986 ;60 :608-15 . 6 . Gould PC, Khawaja Fl, Rosenthal WS . Antibiotic-associated hemorrhagic colitis . Am J Gastroenterol 1982 ; 77 : 491-3 . 7 . Sakurai Y, Tsuchiya H, Ikegami F, Funatomi T, Takasu S, Uchikoshi T . Acute right-side hemorrhagic colitis associated with oral administration of ampicillin . Dig Dis Sc i 1979 ; 24 : 910-5 . 8 . Toffler RB, Pingoud EG, Burrell MI . Acute colitis related to penicillin and penicillin derivatives . Lancet 1978 ; ii : 707-9. 9 . Lowry OH, Rosebrough NJ, Farr AJ, Randall RJ . Protein measurement with the Folin phenol reagent . J Biol Chem 1951 ; 193: 265-75 . 10 . O'Brien AD, Holmes RK . Shiga and Shiga-like toxins . Microbiol Rev 1987; 51 : 206-20 . 11 . Donta ST, Shaffer SJ . Effects of Clostridium difficile toxin on tissue-cultured cells . J Infect Dis 1980 ; 141 :218-22 . 12 . Lima AAM, Lyerly DM, Wilkins TD, Innes DJ, Guerrant RL . Effects of Clostridium difficile toxins A and B in rabbit small and large intestine in vivo and on cultured cells in vitro . Infect Immun 1988; 56 : 582-8 . 13 . Guerrant RL, Wanke CA, Pennie RA, Barrett LJ, Lima AAM, O'Brien AD . Production of a unique cytotoxin by Campylobacterjejuni. Infect Immun 1987 ; 55 : 2526-30.



K. oxytoca cytotoxin 14 . 15 . 16 . 17 . 18 .

211

Ketyi I, Pacsa S, Emody L, Vertenyi A, Kocsis B, Kuch B . Shigella dysenteriae 1-like cytotoxic enterotoxins produced by Salmonella strains . Acta Microbiol Hung 1979 ; 26 : 217-23 . O'Brien AD, LaVeck GD, Thompson MR, Formal SB . Production of Shigella dysenteriae type 1-like cytotoxin by Escherichia coli. J Infect Dis 1982 ; 146 : 763-9 . Baloda SB, Faris A, Krovacek K, Wadstrom T . Cytotoxic enterotoxins and cytotoxic factors produced by S . enteritidis and S. typhimurium. Toxicon 1983 ; 21 : 785-90 . Koo FCW, Peterson JW . Cell-free extracts of Salmonella inhibit protein synthesis and cause cytotoxicity in eukaryotic cells . Toxicon 1983 ; 21 : 309-20. Ashkenazi S, Cleary TG, Murray BE, Wanger A, Pickering LK . Quantitative analysis and partial characterization of cytotoxin production by Salmonella strains . Infect Immun 1988; 56 : 3089-94 .