Biological activities of the peptides of staphylococcal enterotoxin C formed by limited tryptic hydrolysis.

Abstract : Staphylococcal entertoxin C1 is converted to a doubly cleaved molecule by trypsin digestion with one of the scissions internal to the disulfide loop and one external to it. The larger, disulfide-containing polypeptide (Mr = 22,000) exhibited excellent binding to antiserum to the intact enterotoxin. The residual amino terminal fragment (Mr = 6,500) also bound to this antibody but only weakly. Only the carboxyl terminal carboxamidomethylated moiety of the 22,000 Mr polypeptide (Mr = 19,000) combined with anti-enterotoxin C1. Both the 22,000 Mr and 6,500 Mr polypeptides could partially inhibit the binding of entertoxin C1 to its antibody in a competitive system. It is suggested that enterotoxin C1 possesses three major antigenic determinants, two on Cam 19,000 and one on the 6,500 Mr fragment.


Staphylococcal enterotoxin
Cl is converted to a doubly cleaved molecule by trypsin digestion with one of the scissions internal to the disulfide loop and one external to it (Spero, L., Griffin, B. Y., Middlebrook, J. L., and Metzger, J. F. (1976) J. Biol. Chem. 251,55804588).
The larger, disulfide-containing polypeptide (3& = 22,000) exhibited excellent binding to antiserum to the intact enterotoxin. The residual NH2terminal fragment (Mr = 6,500) also bound to this antibody, but only weakly. Only the COOH-terminal carboxamidomethylated moiety of the i?l, = 22,000 polypeptide (&f, = 19,000) combined with anti-enterotoxin Cl. Both the iP2, = 22,000 and M, = 6,500 polypeptides could partially inhibit the binding of enterotoxin C1 to its antibody in a competitive system. It is suggested that enterotoxin Cl possesses three major antigenic determinants, two on carboxamidomethyl1M, = 19,000 and one on the M, = 6,500 fragment.
A significant degree of refolding to a native-like conformation is indicated for the M, = 22,000 and the carboxamidomethyl iV& = 19,000 materials by (a) the strong binding of these polypeptides to anti-enterotoxin C1, (b) the strong binding of enterotoxin Cl to antibody to the M, = 22,000 polypeptide, and (c) their circular dichroic spectra in the far ultraviolet.
The M, = 6,500 polypeptide exhibited mitogenic activity, but not emetic activity. Conversely, the 32, = 22,000 polypeptide was able to induce diarrhea in rhesus monkeys, but was not mitogenic, suggesting that the active sites for these activities are widely separated on the enterotoxin molecule.
Two of the enterotoxins elaborated by certain strains of Staphylococcus aureus, enterotoxin B and enterotoxin C1, have been shown to be susceptible to limited specific enzymic digestion by trypsin (1, 2). The primary cleavage in both instances occurs at a site on the polypeptide chain interior to the disulfide loop. A secondary cleavage also occurs which for enterotoxin C1 goes rapidly to completion. The doubly cleaved molecule, enterotoxin Cl-Tz, may be represented in schematic linear form, where the indicated molecular weights were obtained from amino acid analysis: * This work was presented in part at the 67th Annual Meeting of the American Society of Biological Chemists, San Francisco, Calif., June 1976. The views of the authors do not purport to reflect the position of the Department of the Army or the Department of Defense. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "aduertisement" in accordance with 18 USC. Both enterotoxin B-T and enterotoxin Cl-T2 retain all the biological activities of the parent toxins: binding to antibody, mitogenicity, and the induction of emesis and diarrhea. Moreover, enterotoxin Cl-T2 behaves as a single particle with essentially unaltered conformation.
The extraordinary lability to limited tryptic hydrolysis does not appear to play a role in enterotoxicity.
Although the site of nicking is in the disulfide loop and thus bears a structural similarity to that occurring with several of the bacterial exotoxins (3-5), a parallel biological activation does not take place (1). We have suggested (2) that the kind of lability exhibited by these enterotoxins is associated with p turn structures at the protein surface. The labile Lys-Thr bond in enterotoxin B was found by the Chou and Fasman procedure (6) for the prediction of secondary structure to be part of such a region. It was noted too that a naturally nicked bond in concanavalin A was placed by x-ray crystallography between the 2nd and 3rd residues of an exposed p turn (7). A similar situation with human &microglobulin has come to our attention (8). A Tyr-Ser bond at positions 10 and 11 is unusually susceptible to cleavage by chymotrypsin.
Application of the Chou and Fasman method (6) indicates that this serine residue is the first member of a tetrapeptide with a high probability for forming a p turn. These specific cleavages can provide a means of obtaining large well defined polypeptide fragments and we report here on the association of the polypeptides of enterotoxin C1 with the various biological activities of the parent enterotoxin.
It is demonstrated that both major tryptic polypeptides of enterotoxin Cl-T2 possess antigenic determinants, but that mitogenic activity is restricted to one fragment and emetic activity is restricted to the other.

EXPERIMENTAL PROCEDURES
Materials-Staphylococcal enterotoxin C, and C,-Ta were prepared as previously described (2). The M, = 22,000 and M, = 6,500 polypeptides of enterotoxin Cl-T2 were separated on a column of Sepharose 6B (Pharmacia) in 6 M guanidine hydrochloride (Schwarz/Mann, ultrapure grade) (2). For some testing of biological activity, it was necessary to purify these two polypeptides further. The M, = 6,500 polypeptide was rechromatographed under the same conditions.
The single resultant fraction was dialyzed free of guanidine and passed through an affinity column of rabbit antibody to enterotoxin C, bound to Sepharose 4B (9). This was to remove any residual trace of intact enteritoxin Cl-T2 re-formed by complementation of contaminating M, = 22,000 polypeptide with the M, = 6,500 fragment. The M, = 22,000 polypeptide, although appearing as a single peak in the gel filtration used for its isolation, was significantly contaminated Biological Activity of Peptides of Enterotoxin C with enterotoxin Cl-T,. Rechromatographing this fraction three times, each time selecting the latter one-half to two-thirds of the peak, served to reduce the contamination to levels acceptable for assay.
After reduction with P-mercaptoethanol and alkylation with iodoacetamide, the M, = 22,000 polypeptide was separated into its constituent peptides, Cam' M, = 4,000 and Cam M, = 19,000 by chromatography on Sepharose 6B in 6 M guanidine hydrochloride. These reactions were carried out as previously described (2) except that the medium was 6 M with respect to guanidine hydrochloride. Preparation of Antisera-Anti-enterotoxm C, was prepared by intracutaneous injection of the enterotoxin without adjuvant in New Zealand white rabbits.' A regimen based on that developed by Silverman (10) was employed.
Only those sera giving identical Ouchterlony titers were pooled. Rabbit antiserum for the M, = 22,000 polypeptide was produced by intramuscular inoculation of three 100.,ag doses of polypeptide at weekly intervals in 10% rabbit serum albumin.
Most of the serum used in this study was from a bleeding of a single rabbit made 1 week after a second course of immunization administered 3 months after the first series of injections.
Labeling of Enterotoxin C, and Z'ryptic Peptides-All labeling was carried out by the gaseous diffusion method of Gruber and Wright (11) with rz31. Enterotoxin C, was labeled in phosphate-buffered saline. The peptides were dissolved in 6 M guanidine hydrochloride during labeling and unbound radioisotope was removed by dialysis against 6 M guanidine hydrochloride.

Antigen-binding Capacity rind Competiti~~e
Binding Assc~y-These techniques are based on the ability of protein A-containing strains of S. aureus to react specifically and with high affinity with the Fc portion of IgG (12). In the antigen-binding capacity assay, samples of labeled antigen (25 to 100 ng in 500 ~1) were added to 500-~1 volumes of 2-fold, serially diluted antiserum in phosphate-buffered saline containing 0.5% bovine serum albumin.

RESULTS
Solubilization of Peptides-Obtaining the peptides in a stable, soluble state from the concentrated guanidine hydrochloride solutions in which they were isolated was a significant problem, especially for the M,. = 22,000 fragment. Dialysis against aqueous buffers over a wide pH range resulted in precipitation of most of the polypeptide and an unstable solution. It was soluble in both anionic (sodium dodecyl sulfate) and cationic (cetylpyridinium chloride) detergent solutions but serologic activity could not be demonstrated in the presence of either detergent. Successful solubilization was achieved by dilution into concentrated solutions of bovine serum albumin. The general procedure consisted of the dropwise addition of a 5 to lo-mg/ml solution of the polypeptide in 6 M guanidine hydrochloride (obtained by dry concentration with Aquacide (Calbiochem)) to a vigorously stirred 10% buffered solution of the albumin. Concentrations up to 1 mg/ml were readily obtained and t,he solutions were stable for several days. Except where noted, all biological measurements were carried out on peptides prepared in this manner.
Serologic Properties of the Tryptic Peptides-The binding capacities of the radiolabeled major peptides of enterotoxin C,-T2, the M, = 6,500 fragment, and the M, = 22,000 fragment, are compared in Fig. 1 with that for intact '?-enterotoxin Cr. The data have been normalized to equal amounts of labeled antigen; this is based on the assumption that over the range of antigen employed, the ratio of antibody to antigen at the endpoint is independent of antigen level. It is readily apparent that the M, = 22,000 polypeptide bound very well to the antibody to the whole enterotoxin; on a weight basis, the affinity was nearly one-half that of enterotoxin Cl, and essentially all of the label was precipitable.
The binding of the M, = 6,500 polypeptide was significant but considerably weaker.
Competitive inhibition of these two polypeptide fragments with the native enterotoxin is presented in Fig. 2. Both materials competed successfully with the native toxin for its homologous antibody. As anticipated, the M, = 22,000 polypeptide was the more potent inhibitor, but a plateau of reaction was not reached with either fragment at its highest available concentration.
If it is assumed that each fragment contained a unique determinant, some antibodies for intact toxin would be incapable of combining with the fragment under assay and total inhibition of the native enterotoxinantiserum reaction could not be achieved.
Neither of these polypeptide preparations contained M, = 28,000 material by polyacrylamide gel electrophoresis. However, as noted earlier, the detectable limit of the method is slightly better than 0.1%. This does not present a problem for the M, = 22,000 polypeptide because significant inhibition was observed at concentrations where contamination levels with whole enterotoxin of 1% would be necessary to produce the observed effect. Moreover, if the inhibitory response were indeed due to contamination, the inhibition curve would have been identical with that seen with enterotoxin Cl, but displaced to the right (cf Ref. 13, Fig. 6). With the M, = 6,500 polypeptide, we must rely on the methods employed in its isolation and purification; after the second chromatographic separation, M, = 28,000 material was not detected on analytic gels and subsequent purification with the antibody affinity column was demonstrated to be capable of removing enterotoxin contaminants several orders of magnitude greater than would be present.
These difficulties are obviated in the antigen-binding capacity assay where trace contamination cannot contribute significantly to the percentage of labeled antigen bound to the antibody. Binding data with antisera to enterotoxin C, and to the M, = 22,000 polypeptide are shown in Table I for intact enterotoxin and its four tryptic peptides. In these calculations, it was assumed that all the antigen bound at the 50% endpoint was in the form Ag,Ab (15), so that for the bivalent antibody, the molar ratio of antigen to antibody was 4. Endpoints were estimated from log-log plots of the volume of antibody against the percentage af antigen bound. These graphs were linear from about 20 to 80% bound and had nearly identical slopes. This latter property facilitated the endpoint estimation when an extrapolation from the data was required. These molar ratios emphasize the contrast between the excellent binding of the M, = 22,000 polypeptide and the comparatively weak binding of the M, = 6,500 polypeptide to anti-enterotoxin C,. Lack of binding of Cam M, = 4,000 to this antiserum indicated the absence of a native determinant on this portion of the M, = 22,000 fragment. In contrast, Cam M, = 19,000 bound very well to the antiserum, requiring only 6 times as much antibody as the whole enterotoxin.
It was, however, less efficacious than the M, = 22,000 polypeptide. This may be attributed to either the absence of a determinant or to inferior restoration of native structure in the somewhat smaller polypeptide. The lack of binding by Cam M, = 4,000 leads us to favor the latter explanation.
The excellent binding of the M, = 22,000 polypeptide to anti-enterotoxin C, is indicative of a high degree of refolding of this fragment to a native conformation. This contention is supported by the surprisingly strong binding of the whole enterotoxin to antiserum raised against the M,. = 22,000 polypeptide. Enterotoxin C, is a very stable, compact protein and the antibody sites with which it combines reflect immunoglobulin biosynthesis induced by determinants intrinsic to the intact native enterotoxin, i.e. these regions must also be present in the immunogen, the M, = 22,000 polypeptide. It is well established that antibodies elicited by immunization with denatured protein either fail to react, or do not react extensively, with the native protein (16). Conversely, it is also clear that complete refolding does not exist in the M, = 22,000 polypeptide. This was demonstrated by 1) the binding of Cam M, = 4,000 for the antibody to the M, = 22,000 polypeptide despite its complete failure to react with anti-enterotoxin C, and 2) by the better binding of Cam M, = 19,000 to the anti-M, = 22,000 polypeptide than to anti-enterotoxin C,.

Emetic Activity of the Tryptic Peptides of Enterotoxin
Cl-All the animals used for test of emetic activity of the tryptic polypeptides had no antibody titer to enterotoxin B or C, by hemagglutination assay. No animals died and all appeared to be completely normal within 24 h after inoculation.
The M, = 6,500 polypeptide produced no emesis or diarrhea in monkeys injected intravenously with doses up to 10 pg/kg, the equivalent to 300 median effective (ED,,,) doses of the intact enterotoxin.
The results of assay of the M, = 22,000 polypeptide in rhesus monkeys are shown in Table II. This prep-TABLE  1 Binding of enterotoxin C, and trypticpeptides derived from it with ant&eta The antigens were labeled and binding capacity was determined as described under 'Zxperimental Procedures." Fifty per cent endpoints were estimated from log-log plots of the volume of antibody added against the percentage of antigen bound. Antibody levels and molar ratios were calculated assuming that all antigen bound at the endpoint was in the form of Ap.>Ab.
Molar ratio of total antibody to labeled antigen at 505 endl&nt Antigen Anti-&f, = 22,000 Anti-enterotoxin C, Biological Activity of Peptides of Enterotoxin C aration of the polypeptide contained 0.25% contamination with M, = 28,000 material which represents less than onefourth of an EDm of enterotoxin at the higher dose of the M, = 22,000 polypeptide at which diarrhea occurred. A positive response to this level of enterotoxin C1 has not been observed.
A positive result at the same level was obtained with another preparation of the M, = 22,000 polypeptide in cynomolgus monkeys. It must be noted that no emesis was seen with this large polypeptide and the diarrhea was always less severe than with intact enterotoxin. We concluded that the active site for emesis and diarrhea is located within that portion of the molecule isolated on the M, = 22,000 polypeptide. The greatly reduced activity may be due to any of several factors operating singly or in combination, e.g. faster metabolic turn- over, impaired conformation, or weaker binding to the putative receptor.

Mitogenic
Activity of the Tryptic Peptides of Enterotoxin Cl-The M, = 6,500 polypeptide demonstrated a low level of mitogenic activity. A typical response is shown in Fig. 3. In these experiments, the M, = 6,500 polypeptide was prepared by removal of the guanidine in which it was isolated by dialysis against phosphate-buffered saline. No mitogenic activity was found for the M, = 22,000 polypeptide under any of several solubilixed conditions. Circular Dichroisrn of the Peptides--In Fig. 4 are presented the circular dichroic spectra of the four tryptic peptides in the far ultraviolet. The spectrum of the native enterotoxin (2) in that region is included for comparison. The curves for the M* = 6,500 polypeptide and for Cam M, = 4,000 are typical of random coil conformation. Both Cam M, = 19,000 and the M, = 22,000 fragments, however, show significant similarity to that of the intact enterotoxin. The spectrum of enterotoxin Ci is very much like that of enterotoxin B which has been reported by analysis of its circular dichroism and by the Chou and Fasman procedure (6) for the prediction of secondary structure to contain about 30% P-pleated sheet and about 10% a helix (17). A striking feature of this structure is extensive grouping of anti-parallel P-pleated sheet around the disulflde loop. Much of this would be retained in both of these large polypeptides.

DISCUSSION
The assays for mitogenic and emetic activity of the two major polypeptides of enterotoxin G-T2 indicate that these activities are associated with widely separated regions of the polypeptide chain. Mitogenesis was induced by the NH*-terminal fragment, and diarrhea by the remaining portion of the molecule. We earlier suggested (14) that a residual mitogenic activity of the closely related enterotoxin B,3 detoxified by treatment with formaldehyde, implied that the mitogenic and emetic sites of this enterotoxin were not identical. The present studies point to a similar conclusion and provide a general localization of the sites responsible for these two activities on enterotoxin Cl.
Assay The very good binding of the M, = 22,000 polypeptide and the much weaker binding of the M, = 6,500 polypeptide with anti-enterotoxin C1 is conveniently interpreted in terms of the concept of Sachs et al. (18) that the efficiency of binding of a combining fragment is a function of the equilibrium between native and disordered forms. In this light, the smaller polypeptide is largely in random conformation in solution, while the M, = 22,000 polypeptide has an equilibrium greatly in favor of the native structure. CD spectra in the far ultraviolet, which indicated a lack of organized structure for the M, = 6,500 material but were suggestive of a secondary structure for the M, = 22,000 polypeptide similar to that of the intact enterotoxin, are consistent with this view. It is also supported biologically by the observation that the native enterotoxin binds extremely well to antibody to the M, = 22,000 fragment.
It may be noted that a significant degree of native structure on the M, = 22,000 polypeptide also implies that the primary nucleation site for folding is in this part of the amino acid sequence. Further, it is contrary to the idea that virtually an entire sequence is required for folding to a native conformation from a disordered state (19). Wetlaufer and co-workers (20) have obtained a native-like structure from the peptide 13-105 of hen egg lysozyme.