ADP-ribosylation in Clostridium difficile toxin-treated cells is not related to cytopathogenicity of toxin B

Biochimica et Biophysica Acta, 1091 (1991) 51-54 © 1991 Elsevier Science Publishers B.V. (Biomedical Division)0167-4889/91/$03.50 ADONIS 0167488991000596

51

BBAMCR 12839

ADP-ribosylation in Clostridium difficile toxin-treated cells is not related to cytopathogenicity of toxin B Inger Florin and Monica Thelestam Department of Bacteriology, Karolinska Institutet, Stockholm (Sweden) (Received 30 January 1990)

Key words: ADP ribosylation; Toxin B; Stress; Cytopathogenicity; (C dtfficile); (Human fibroblasts)

ADP-ribosylation of a protein in human fibroblasts treated with partially purified CIostridium diffieile toxin B was previously reported. Here we show that the same protein was ADP-ribosylated also in human fibroblasts exposed to supernatant from a C. difficile strain producing neither toxin A nor toxin B. Furthermore, in Chinese hamster ovary and in Veto cells, showing toxin B-induced cytopathogenic effect, the protein was not significantly ADP-ribosylated. The results indicate that the ADP-ribo~?iation is unrelated to the cytopathogenic effect of toxin B. It appears to be caused by another unidentified factor from C. difficile, and the substrate may correspond to a protein modified endogenously in cells exposed to stressful situations. Cellular actin was not ADP-ribosylated by toxin B.

Introduction

Materials and Methods

Two toxins, toxin A and toxin B, from Clostridium difficile are involved in the etiology of antibiotic-associated colitis [1-3]. In toxin B-treated cultured cells, a characteristic cytopathogenic effect (CPE) appears after a latency period, during which the toxin is internalized via endocytosis [4] and processed in the lysosomes [5]. Some component of the micro filament system appears to be the intracellular target of toxin B [6-10], but the molecular mechanism of action is not clarified. We have reported ADP-ribosylation of a protein in human fibroblasts treated with partially purified toxin B [11], but were not able to conclude whether this ADP-ribosylation was caused by toxin B or by an hitherto unidentified factor produced by C. difficile. Since there is still no generally accepted purification method for toxin B, we have characterized the relationship between the ADP-ribosylation and the cytopathogenie effect of toxin B using other approaches, and found that these two effects are unrelated to each other.

Chemicals. Eagle's minimal essential medium, Ham's F-10 medium, calf serum and trypsin were obtained from Flow Laboratories, Irvine, U.K. Nicotinamide [U14C]adenine dinucleotide (NAD) (specific activity 260287 mCi/mmol) was purchased from Amersham International, Bucks., U.K. Cytochalasin B and cycloheximide were from Sigma, St. Louis, MO. Acrylamide (> 99.9% pure) was from Bio-Rad, Richmond, CA. Toxins. Toxin B was purified from dialysis cultures of a clinical isolate of C. difficile as previously described [12]. A strain of C. difficile producing neither toxin A nor toxin B (kindly provided by Bo Aronsson, National Bacteriological Laboratory, Stockholm, Sweden) was cultivated and the supernatant purified under identical conditions as for the toxin-producing strain. Diphtheria toxin (300 LF/mg) was kindly donated by P. Askel0f and P. Gillenius, National Bacteriological Laboratory, Stockholm, Sweden. Cultivation of cells. Human diploid embryonic lung fibroblasts (line MRC-5) were cultivated as previously described [13] in Eagle's minimal essential medium supplemented with 10% newborn calf serum, 5 mM Lglutamine, penicillin (100 U/ml) and streptomycin (100 /~g/ml). Chinese hamster ovary (CHO) cells (line K-I) and African green monkey kidney (Vero) cells were cultivated in Ham's F-10 medium supplemented with 10% fetal calf serum, 5 mM L-glutamine, penicillin (100 U/ml) and streptomycin (100 #g/ml) under otherwise

Abbreviations: LF, flocculating unit; CPE, cytopathogenic effect; TCDso, 50% tissue culture dose; DT, diphtheria toxin; EF-2, elongation factor 2; DNase, deoxyribonuclease. Correspondence: I. Florin, Department of Bacteriology, Karolinska institutet, S-10401 Stockholm, Sweden.

52 identical conditions as the lung fibroblasts. The cells were free of mycoplasma infection as determined by Hoechst staining [14] and cultivation. The unit of toxin activity was the 50% tissue culture dose (TCDs0), i.e., the toxin dilution inducing within 20 h a characteristic actinomorphic cytopathogenic effect in 50~ of human fbroblasts exposed under standardized conditions [13]. The amount of protein corresponding to 1 TCDs0 varied from 0.1 to 1 ng in different preparations. Exposure of all three cell types to a dose of 1000 TCDs0 for 2 h resulted in a CPE in all exposed cells. No CPE developed upon exposure of the cells to supernatant from C difficile lacking toxins A and B, even after several days of exposure. Assay of ADP-ribosylation in intact cells. Confluent monolayers of cells in six-well plates were incubated at 37°C with 1-2 /~Ci [14C]adenine NAD in 1 ml cell culture medium per well. (NAD is taken up very slowly by intact cells [15]. A labeling of intraceUular NAD is nevertheless obtained also by extracellular hydrolysis to adenosine which is easily taken up and incorporated into NAD intracellularly [16].) After 1 h, toxin B (1000 TCDs0), purified culture supernatant from C. difficile lacking toxins A and B (same amount of protein as used for toxin treatment), diphtheria toxin (30 LF) or cycloheximide (100/~M) was added and the incubation continued for 2 h. Extracellular radioactivity was carefully removed by rinsing twice with Hanks' balanced salts solution. The cells were solubilized in 0.1t~ SDS, whereupon the cells from two wells (corresponding to 302-368/~g protein) were pooled, and the protein precipitated with trichloroacetic acid as previously described [13]. The amount of protein was determined by the method of Lowry et al. [17].

Sodium &,decylsulfate polyacrylamide gel electrophore. sis. Trichloroacetic acid-precipitated protein pellets were boiled in preparation buffer and analyzed by electrophotcsis on 7-20~ gradient gels according to Laemmli [18]. After staining g~th Coomassie brilliant blue R-250, the gels were preparetl for fluorography according to Chamberlain [19] and dried. Fuji X-ray RX films were exposed to the gels for 2-6 weeks at - 7 0 ° C before development. All experiments were performed at least twice, Results and Discussion When intact human fibroblasts, preincubated with [t4C]adenine labeled NAD, were exposed to partially purified toxin B until a CPE was developed in all cells, one cellular protein (approx. 75 kDa) was clearly labeled (Ref. 11, and Fig. 1, lane 3). In untreated fibroblasts the same protein was very weakly labeled (ReL 11, and Fig. 1, lane 2). When human fibroblasts were exposed to supernatant from C. difficile lacking toxins A and B, the same protein was labeled to at least the same extent as

1

2 " 3"~

5"

6

7

8

9

Fig. 1. ADP-ribosylation in toxin-treated cells. Autoradiograms of gel electrophoresed protein from MRC-5 cells (lanes 1-5) and CHO cells (lanes 6-9) preincubated with [14C]NAD for 1 h, and treated for 2 h as follows: supernatant from C difficile lacking toxins A and B (lane 1), partially purified toxin B (contaminated with factor D) (lanes 3, 8), diphtheria toxin (lanes 4, 7), cycioheximide (lanes 5, 6), no treatment (lanes 2, 9). Molecular mass markers (from top to bottom): phosphorylase b (94 kDa), albumin (67 kDa), ovalbumin (43 kDa), carbonic anhydrase (30 kDa), trypsin inhibitor (20 kDa) and ot-lactalbumin (14 kDa).

in toxin B-treated cells (Fig. 1, lane 1), although no CPE appeared. This result indicates that the ADP-ribosylation is unrelated to the CPE of toxin B. Both CHO and Vero cells are sensitive to toxin B [5,20]. When intact prelabeled CHO cells were exposed to partially purified toxin B until a CPE was developed in all cells, a protein with the same molecular mass eapprox. 75 kDa) (Fig. 1, lane 8) as in human fibroblasts was labeled. In contrast to the situation in human fibroblasts, however, the intensity of this label was not stronger than in the control cells not exposed to toxin B (Fig. 1, lane 9). The same labeling pattern was observed in Vero cells (data not shown). This result further strengthens the conclusion that the ADP-ribosylation is unrelated to the CPE of toxin B. Consistent with this conclusion, we also observed that cytochalasin B-induced microfilament disorganization [21] was not associated with ADP-ribosylation of any protein in human fibroblasts (data not shown). If the ADP-ribosylation had occurred as a consequence of the characteristic microfilament disorganization taking place after either cytochalasin B or toxin B treatment [6,9], an ADP-ribosylated protein would have been expected to appear also in cells exposed to cytochalasin B. The ADP-ribosylation observed in human fibroblasts could be caused by an unidentified factor from C. d(fficiie. The production of biologically active factors

53 separate from toxin A and B have been reported [22-25]. In fact, an ADP-ribosyltransferase not associated with toxin A or toxin B and having actin as a substrate has been detected in one C. difficile strain (out of 15 tested) [24]. An enterotoxin separate from toxin A, designated as fraction C, was isolated from both cytotoxic and non-cytotoxic strains of C. difficile [25], but whether it possesses ADP-ribosyltransferase activity has not been reported. Since the relationships between all these factors are not yet clear, we will designate the activity described here as factor D. The 75 kDa-protein ADP-ribosylated in human fibroblasts treated with factor D-contaminated toxin B (Fig. 1, lane 3) comigrates with a protein strongly labeled in human fibroblasts treated with diphtheria toxin (DT) (Ref. 11, and Fig. 1, lane 4), which was used as positive control to ensure that the cells were able to take up the radioactive label and perform ADP-ribosylations. The protein was initially assumed to be elongation factor 2 (EF-2; 95 kDa), the sub~trate for DT [26], since the molecular mass was in the right range and since we were not aware of any other substrate for DT. It was confirmed by two-dimensional gel electrophoresis (according to O'Farrell [27]) that the protein labeled in factor D-treated cells was the same as that labeled after DT-treatment. This finding was further substantiated by the observation that the labeling intensity of the DT-induced band was decreased by exposing the cells to factor D before addition of DT and labeled NAD (data not shown). Nevertheless, the protein did not react with anti-EF-2 antibodies (kindly provided by Lars Nilsson and Odd Nygfird, Department of Cell Biology, University of Stockholm, Sweden) in immunoprecipitation tests or Western blots (data not shown). Later on, in cell types other than human fibroblasts, a second ADP-ribosylated protein of a slightly higher molecular mass was observed (Fig. 1, lane 7) after DT-treatment. Presumably, this protein is identical with EF-2. Indeed, this band was detected also in DT-treated human fibroblasts when the films were exposed to the gels for several months (data not shown). The 75 kDa-protein initially reported to be ADPribosylated in cells treated with partially purified toxin B (and now known to have a slightly lower molecular weight than EF-2) has been suggested [24] to correspond to a protein of 78-89 kDa (in different gel systems) referred to as Sp83 [28], which is the primary acceptor protein for ADP-ribose in various types of cell exposed to stressful situations. Ledford et al. [29] also identified an acceptor protein (80 kDa) for ADP-ribose, which they referred to as P80. The extent of ADP-ribosylation of these proteins is increased after, e.g., inhibition of protein synthesis. Indeed, in both human fibroblasts and CHO cells treated with cycloheximide (Fig. 1, lanes 5 and 6, respectively), at a concentration which inhibits protein synthesis by more than 90% [29],

a strongly ADP-ribosylated protein comigrated with the protein labeled in factor D-treated cells (Fig. 1, lane 3). This finding is consistent with the suggestion that the labeled band corresponds to a protein, which acts as an acceptor protein for ADP-ribose at a stressful situation induced by factor D. In conclusion, factor D seems to be an hitherto unidentified factor produced by C. difficile. In human fibroblasts it causes ADP-ribosylation probably of a stress-inducible protein, without causing any visible cytopathogenic effect. In CHO and Veto cells, however, no ADP-ribosylation is detected. Since several clostridial species [30-32] have been shown recently to ADP-ribosylate cellular actin, and since toxin B disorganizes the microfilament system [6-10], a similar activity in C. difficile toxin B has been searched for. Popoff et al. [31] found no labeling of cellular actin after microinjection of toxin B into X. laevis oocytes loaded with [32p]NAD. However, with this approach it might not be possible to conclude whether the toxin was truly active in the oocytes, as it requires lysosomal activation upon its endocytosis in order to intoxicate at least cultured mammalian cells [5]. Nevertheless, neither in intact human fibroblasts nor in CHO cells clearly intoxicated by toxin B did we detect any ADP-ribosylation of cellular actin (43 kDa) (Fig. 1, lanes 3 and 8, respectively). Since we might have failed to detect ADP-ribosylated actin in whole cell homogenates, we also separated the actin by chromatography of fibroblast homogenates on DNase columns [33,34], but the actin fractions contaired no radioactive label at all (data not shown). In summary, two major conclusions concerning toxin B can be drawn from this work: (i) The previously reported ADP-ribosylation of a cellular protein in human fibroblasts treated with partially purified toxin B is not related to the cytopathogenic effect of toxin B and thus not caused by toxin B. The reported ADP-ribosylation is probably an endogenous modification due to a stressful situation induced by a contaminating unidentified factor from C. difficile. (ii) Toxin B does not ADP-ribosylate actin.

Acknowledgements The accurate and reliable technical assistance of Lena Norenius is gratefully acknowledged. We thank Michael Gill, Tufts University, Boston, U.S.A., for making us aware of stress-inducible ADP-ribosylation. This investigation was supported by the Swedish Medical Research Council (grant No. 16X-05969).

References 1 Borriello. S.P. (ed.) (1984) Antibiotic-associated diarrhoea and colitis. The role of Clostridium difficile in gastrointestinal disorders. Martinus Nijhoff. The Hague.

54 2 Lyerly, D.M., Krivan, H.C. and Wilkins, T.D. (1988) Clin. Microbiol. Rev. 1, 1-18. 3 Rolfe, R.D. and Finegold, S.M. (eds.) (1988) Clostridium difficile: Its role in intestinal disease, Academic Press, San Diego. 4 Florin, I. and Thelestam, M. (1983) Biochim. Biophys. Acta 763, 383-392. 5 Florin, I. and Thelestam, M. (1986) Microb. Pathog. 1, 373-385. 6 Thelestam, M. and BriSnneg~ird, M. (1980) Scand. J. Infect. Dis. Suppl. 22, 16-29. 7 Wedei, N., Toselli, P., Pothoulakis, C., Fails, B., Oliver, P., Fransblau, C. and LaMont, T. (1983) Exp. Cell Res. 148, 413-422. 8 Pothoulakis, C., Barone, LM., Ely, R., Faris, B., Clark, M.E., Franzblau, C. and LaMont, J.T. (1986) J. Biol. Chem. 261, 13161321. 9 Mitchell, M.J., Laughon, B.E. and Lin, S. (1987) Infect. Immun. 55, 1610-1615. 10 Ottlinger, M,E. and Lin, S. (1988) Exp. Cell Res. 174, 21=-229. 11 Florin, I. and Thelestam, M. (1986) Biochem. Biophys. Res. Commun. 139, 64-70. 12 Caspar, M., Florin, !, and Thelestam, M. (1987) J. Cell. Physiol. 132, 168-172, 13 Florin, I. and Thelestam, M. (1981) Infect. lmmun. 33, 67-74. 14 Chen, T.R, (1977) Exp. Cell Res. 104, 255-262. 15 Nolde, S. and Hilz, H. (1972) Physiol. Chem. 353, 505-513. 16 Colyer, R.A., Burdette, K.E. and Kidwell, W.R. (1973) Biochem. Biophys, Res. Commun. 53, 960-966. 17 Lowry, O.H,, Rosebrough, N.J., Farr, A.L. and Randall, R.J. (1951) J. Biol. Chem. 193, 265-275. 18 Laemmli, U.K. (1970) Nature 227, 680-685.

19 20 21 22 23

24 25 26 27 28 29 30 31 32 33 34

Chamberlain, J.P. (1979) Anal. Biochem. 98, 132-135. Bowman, R.A. and Riley, T.V. (1986) FEMS Lett. 34, 31-35. Brett, J.G. and Godman, G.C. (1986) Tissue Cell 18, 175-199. Banno, Y., Kobayashi, T., Kono, H., Watanabe, K., Ueno, K. and Nozawa, Y. (1981) Biochem. Intern. 2, 629-635. Justus, P.G., Martin, J.L., Goldberg, D.A., Taylor, N.S., Bartlett, J.G., Alexander, R.W. and Mathias, J.R. (1982) Gastroenterology 83, 836-843. Popoff, M.R., Rubin, E.J., Gill, D.M. and Boquet, P. (1988) Infect. Immun. 56, 2299-2306. Giuliano, M., Piemonte, F. and Gianfrilli, P.M. (1988) FEMS Microbiol. Lett. 50, 191-194. Honjo, T., Nishizuka, Y., Kato, I. and Hayaishi, O. (1971) J. Cell Biol. 246, 4251-4260. O'Farrell, P.H. (1975) J. Biol. Chem. 250, 4007-4021. Carisson, L. and Lazarides, E. (1983) Proc. Natl. Acad. Sci. USA 80, 4664-4668. Ledford, B.L. and Jacobs, D.F. (1986) Eur. J. Biochem. 161, 661-667. Aktories, K., B~irmann, M., Ohishi, l., Tsuyama, S., Jakobs, K.H. and Habermann, E. (1986) Nature 322, 390-392. Popoff, M.R. and Boquet, P. (1988) Biochem. Biophys. Res. Commun. 152, 1361-1368. Schering, B., B!irmann, M., Chatwal, G.S., Geipel, U. and Aktoties, K. (1988) Eur. J. Biochem. 171,225-229. Blikstad, I., Markey, F., Carlsson, L., Persson, T. and Lindberg, U. (1978) Cell 15, 935-943. Aktories, K., Ankenbauer, T., Schering, B. and Jakobs, K.H. (1986) Era. J. Biochem. 161, 155-162.