ACTIVATION PHENOMENON OF CLOSTRIDIUM BOTULINUM TYPE E TOXIN

ACTIVATION PHENOMENON OF CLOSTRIDIUM BOTULINUM TYPE E TOXIN JULIA GERWING, CLAUDE E. DOLMAN, AND DAVID A. ARNOTT Department of Bacteriology and Immunology, The University of British Columbia, Vancouver, British Columbia, Canada Received for publication February 17, 1962 ABSTRACT GERWING, JULIA (The University of British Columbia, Vancouver, Canada), CLAUDE E. DOLMAN, AND DAVID A. ARNorr. Activation phenomenon of Clostridium botulinum type E toxin. J. Bacteriol. 84:302-306. 1962-Highly purified preparations of both nonactivated and trypsin-activated type E botulinus toxins have been analyzed in an ultracentrifuge. Untreated botulinus type E toxin was found to have a sedimentation constant (S2o,w) of 5.6 Svedberg units, whereas trypsin-activated toxins would not form a boundary under identical conditions. These and other considerations indicate that the mechanism of tryptic activation involves a fragmentation process whereby more toxic sites become exposed. The data to be presented here support the first of these alternatives, for studies of the sedimentation rates of both trypsin-activated and nonactivated type E toxins indicate that the activated toxia molecules are considerably smaller than those of the original toxin. MATERIALS AND METHODS The strains, materials, and methods used in this work were as previously described (Gerwing, Dolman, and Arnott, 1961). Ultracentrifugation 0.6 0.5 Downloaded from http://jb.asm.org/ The observations that certain proteolytic enzymes are capable of activating toxins of Clostridium botulinum type E (Dolman, 1953, 1957; Sakaguchi and Tohyama, 1955a, b; Duff, Wright, and Yarinsky, 1956) have remained undisputed, but no clear evidence about the mechanism involved has yet come to light. At least four plausible hypotheses present themselves for consideration. One of these assumes that "protoxin" or "precursor" molecules of relatively large size can be fragmented by certain proteolytic enzymes in such fashion as to release additional toxic activity. Another possibility is that the toxin molecules "unfold" as a result of the enzymatic cleavage of specific amino-acid bonds, thus exposing toxic sites. A third hypothesis visualizes a detachable group on a protoxin molecule, whose removal by enzyme action unmasks the toxic potentialities of the residuum. Fourthly, the enzyme might act only indirectly upon the toxin molecule, by causing some other constituent of toxic filtrates to yield a potentiating factor. z 0 0 U) conm 0.4 0.3 0.2 0.1 VOLUME IN ML. - O -O.10M BUFFER > FIG. 1. Second elution of type E toxin through DEAE cellulose. Material used represents nine pooled preparations of toxin purified by previous elution. Broken line: 280-m,u absorption; solid line: 260-m,u absorption. 302 on April 30, 2018 by guest

VOL. 84, 1962 ACTIVATION OF TYPE E BOTULINUS TOXIN 303 1.8 was carried out in a Beckman/Spinco model E analytical centrifuge. RESULTS 1.86 ' jwe have shown previously that highly purified preparations of both nonactivated and trypsinactivated type E toxins can be produced by 1.4 ethanol precipitation of the toxins followed by elution through diethylaminoethyl (DEAE) cellulose columns (Gerwing et al., 1961). In the experiments reported here, to obtain a further z1. degree of purification, the toxic peak samples '' TOXIN obtained from nine primary fractionations were a. i pooled, dialyzed against running water for 24 hr, n to. A and dried by lyophilization. The dried preparai \ tions were resuspended in distilled water and eluted through DEAE cellulose with sodium 0.8 acetate buffer (ph 6.5) in ascending concentration. The results of this procedure, carried out on o.e 1'.Sg '.\vnonactivated and trypsin-activated toxins, are 0.6 3 V > summarized in Fig. 1 and 2, respectively. In each case, the further purification permitted the separation of another component, which was not 0.4 '\ toxic. [The potencies of the toxic fractions were 0. only slightly higher than those reported previously by us (Gerwing et al., 1961), i.e., approximately 6 X 105 and 1 X 107 MLD per mg of N2 0.2 for nonactivated and activated toxin, respectively. This relatively small improvement in potency, despite the elimination of a nontoxic component, can be explained in terms of the 200 400 demonstrable loss of toxicity suffered by these VOLUME IN ML highly purified preparations in the course of < 0-0.20 M BUFFER. lyophilization.] Once again, the toxic peak FIG. 2. Second elution of trypsin-activated type E samples were dialyzed against running water for toxin through DFAE cellulose. Material used repre- 24 hr and dried by lyophilization. These preparasents nine pooled preparations of toxin purified by tions were resuspended in physiological saline at a previous elution. Broken line: 280-m, absorption; concentration of 2% protein and used for ultrasolid line: 222 m,u absorption. centrifuge analysis. - ; ~ ~~~~~~~~~~~~~~~ --------- ----- FIG. 3. Schlieren diagram showing ultracentrifuge behavior of type E toxin. Pictures taken 8 min apart at 70 angle. Material represents 1.75% protein. Speed of centrifugation: 259,700 X g; temperature: 12 C.

304 GERWING, DOLMAN, AND ARNOTT J. BACTERIOL. FIG 4. Schlieren diagram showing ultracentrifuge behavior of trypsin-activated type E toxin. Pictures taken 16 min apart at 700 angle, after centrifuge had attained speed of 260,000 X g for 40 min. Material represents 1.75% protein. Temperature of centrifugation: 21C. The Schlieren diagrams of the ultracentrifuge behavior of the nonactivated toxin are shown in Fig. 3. This preparation contained two components, with sedimentation constants (S20o,) of 5.6 and 1.1 Svedberg units. To establish whether both of these components were toxic, a second ultracentrifuge run was made using a moving p)artition separation centerl)iece. This run wn-as made at 260,000 X g for 90 min. Separation of the two components was thus effected, and the results showed that only the heavier component (S20,w= 5.6) contained toxic activity. Trypsinactivated toxin was similarly analyzed, and gave Schlieren diagrams (Fig. 4) which suggested that the activated material was too small to form a moving boundary under the conditions of the experiment. DISCUSSION The ultracentrifuge analyses carried out on purified preparations of nonactivated and trypsinactivated botulinus type E toxins appear to throw some light on the tryptic-activation mechanism. In our hands, nonactivated toxin showed a sedimentation constant within the range of many proteins and formed a moving boundary easily discernible with the Schlieren oltical system. Under the same conditions, activated toxins displayed no boundary, but seemed to form a concentration gradient at the meniscus of the centerpiece. The trypsinactivated toxin prepared by Fiock, Yarinsky, and Duff (1961) showed the presence of two components in the ultracentrifuge, with sedimentation constants of 12.5 and 4.7 Svedberg units. As these authors did not allude to any final verification of toxicity, it seems justifiable to suggest that at least one of their components might have proved nontoxic. In any case, their preparations were not comparable to ours, since theirs were derived from whole cultures and subjected to relatively crude purification procedures. Our data suggest that type E botulinus toxin, in the course of activation by trypsin, is broken into smaller components, probably not of uniform size, which are too small to form a boundary in the ultracentrifuge under the stated conditions. (This suggestion is supported by our recent observation that toxic molecules will dialyze into distilled water from trypsin-activated toxins, whereas untreated toxins are not dialyzable). In other words, the process of tryptic activation probably entails some degree of molecular fragmentation, with consequent increase in the number of active groups. In some respects, the mechanism of types A and B toxin production may be analogous. According to Bonventre and Kempe (1960), in the course of cytolysis, endogenous proteolytic enzymes are released along with large protoxin molecules, and may cause partial degradation of the latter with consequent exposure of more toxic sites. Two deductions from the above concept seem warranted. First, bacterial or nonbacterial enzymes may carry molecular degradation of botulinus toxin (regardless of type) to a stage where toxicity declines and eventually disappears, thus accounting, in part at least, for the fickle stability of these toxins. Second, a preliminary enzyme activation of certain types of toxins might be extended by exposing the products, under appropriate conditions, to another proteolytic enzyme. A few experimental observations bearing on this latter possibility can be cited. Sakaguchi and Tohyama (1955a) isolated from botulogenic fish a feebly toxigenic type E strain of C. botulinum, and also a nontoxic, proteolytic organism of the Clostridium genus, the latter being regarded by them as a contaminant. When

VOL. 84, 1962 ACTIVATION OF TYPE E BOTULINUS TOXIN 305 both cultures were grown together, toxin production was increased about 100-fold; but when the "contaminant" was grown in mixed culture with a type A strain of C. botulinum, the latter's toxigenic capacity was unaltered. The same authors (Sakaguchi and Tohyama, 1955b) later ascribed the enhancement of type E toxigenicity to a filterable proteinaselike substance, elaborated by the "contarninant" culture, and acting optimally at ph 5.3. In our view, this "contaminant" might have been a proteolytic ("TP") mutant of its nonproteolytic type E fellow-isolate (Dolman, 1957). Apparently, trypsin is more effective than the foregoing bacterial enzymes as a secondary activator of botulinus toxins. We have shown, for example, that the variable degree of activation undergone by type E toxin on exposure to the proteolytic "TP" mutant may be further enhanced by incubation with 0.1 % trypsin for 1 hr at 37 C. Again, tryptic activation ratios ranging from about 10:1 to 30:1 were obtained with different preparations of type B toxin, irrespective of the proteolytic or nonproteolytic character (Dolman et al., 1960) of their parent cultures. The molecular fragmentation hypothesis and its corollaries, as applied to type E toxin-activation processes, require further experimental verification and elucidation. In particular, a careful study of the effects of activation upon the antitoxin-binding power of the toxin might prove very revealing. Meanwhile, the results of centrifugal analysis reported above seem difficult to explain on any other basis. An alternative suggestion made in our previous paper (Gerwing et al., 1961), that trypsin acts here by cleavage of specific amino-acid bonds, with consequent molecular "unfolding," seems incompatible with findings which point to a substantial reduction in size of the activated toxin molecule. The "potentiation" hypothesis, i.e., that trypsin does not attack the toxin molecule directly but merely acts upon some other constituent in toxic filtrates to yield a potentiating factor, also can be ruled out on the ground that it entails no reduction in size of the toxin molecule. Moreover, as pointed out by Duff et al. (1956), type E toxin can still be activated with trypsin after it has been considerably purified and hence presumably freed from possible potentiators. Finally, Sakaguchi and Sakaguchi (1959) have postulated a masking group, detachable by trypsin from the type E protoxin or precursor molecule. According to them, in the activation process the toxin moiety of the precursor molecule remains intact; but a specific kind of molecular fragmentation occurs involving the separation of a group containing ribonucleic acid, and also a masking group. They ascribed further masking properties to the nucleic acid component itself. The validity of this hypothesis can be challenged on two grounds. The detachment of specific inhibitory fragments from a toxin precursor whose molecules appear to be homogeneous (Fig. 3) should yield activated toxin of reduced but still fairly even molecular size, an expectation which seems incompatible with our findings (Fig. 4). An even more serious objection to the above conception, as lpointed out previously (Gerwing et al., 1961), is our inability to detect nucleic acids in partially purified preparations, either before or after tryptic activation. We believe the discrepancies may be due to our preparations being derived from culture filtrates relatively free from bacterial cytoplasm, whereas the Sakaguchis used culture extracts derived by centrifugation, presumably containing substantial amounts of ribonucleic acid of cell-wall origin. ACKNOWLEDGMENTS This work was carried out in part with the support of a research grant fromn the Medical Research Council of Canada. The authors express appreciation fo, the cooperation received from M. J. Pratt, Federal Department of Agriculture Research Station, Vancouver, in connection with the ultracentrifuge analyses. LITERATURE CITED BONVENTRE, P. F., AND L. L. KEMPE. 1960. Physiology of toxin production by Clostridium botulinum types A and B. IV. Activation of the toxin. J. Bacteriol. 79:24-32. DOLMAN, C. E. 1953. Clostridium botulinum type E. Intern. Congr. Microbiol., Proc. 6th Congr., Rome 4:130-132. DOLMAN, C. E. 1957. Recent observations on type E botulism. Can. Public Health J. 48:187-198. DOLMAN, C. E., M. ToMSICH, C. C. R. CAMPBELL, AND W. B. LAING. 1960. Fish eggs as a cause of human botulism. J. Infectious Diseases 106: 5-19.

306 GERWING, DOLMAN, AND ARNOTT J. BACTERIOL. DUFF, J. T., G. G. WRIGHT, AND A. YARINSKY. 1956. Activation of Clostridium botulinum type E toxins by trypsin. J. Bacteriol. 72: 455-460. FIOCK, M. A., A. YARINSKY, AND J. T. DUFF. 1961. Studies on immunity to toxins of Clostridium botulinum. VII. Purification and detoxification of trypsin-activated type E toxin. J. Bacteriol. 82:66-71. GERWING, J., C. E. DOLMAN, AND D. A. ARNOTT. 1961. Purification and activation of Clostridium botulinum type E toxin. J. Bacteriol. 81: 819-822. SAKAGUCHI, G., AND S. SAKAGUCHI. 1959. Studies on toxin production of Clostridium botulinum type E. III. Characterization of toxin precursor. J. Bacteriol. 78:1-9. SAKAGUCHI, G., AND Y. TOHYAMA. 1955a. Studies on the toxin production of Clostridium botulinum type E. I. A strain of genus Clostridium having the action to promote type E botulinal toxin production in a mixed culture. Japan. J. Med. Sci. Biol. 8:247-253. SAKAGUCHI, G., AND Y. TOHYAMA. 1955b. Studies on the toxin production of Clostridium botulinum type E. II. The mode of action of the contaminant organisms to promote toxin production of type E organisms. Japan. J. Med. Sci. Biol. 8:255-262. Downloaded from http://jb.asm.org/ on April 30, 2018 by guest