Comparative Kinetic Stabilities of Staphylococcal Enterotoxin Types A, B, and C,*

guanidine hydrochloride solution were only 4- to 5-fold larger than those previously reported for enterotoxin type B (Warren, J. R., Spero, L., and Metzger, J. F. (1974) Biochemistq 13, 1678-1683). In addition, the types B and C, enterotoxins unfolded at nearly identical rates in 6


Staphylococcal
enterotoxin types A and C, were observed by viscosimetry and near-ultraviolet difference spectroscopy to unfold at concentrations of aqueous guanidine hydrochloride greater than 1 M. Apparent rate constants of unfolding calculated from spectral curves differed markedly for the two enterotoxins.
Rate constants for the unfolding of enterotoxin A in 2 or 3 M guanidine hydrochloride solution were 30-to 40-fold larger in value than those observed for enterotoxin C,. In contrast, rate constants for the unfolding of enterotoxin C, in 4 or 5 M guanidine hydrochloride solution were only 4-to 5-fold larger than those previously reported for enterotoxin type B (Warren, J. R., Spero, L., and Metzger, J. F. (1974) Biochemistq 13, 1678-1683. In addition, the types B and C, enterotoxins unfolded at nearly identical rates in 6 M guanidine hydrochloride and 8 M urea solution. Enterotoxin A unfolded about 50-fold faster in 8 M urea than enterotoxin B and C,. Therefore, unfolding of enterotoxin A by guanidine hydrochloride or urea appears to have a considerably lower activation energy than unfolding of enterotoxin B or C, by the denaturants.
It is suggested that the observed differences in kinetic stability reflect a significant dissimilarity of the native structure of enterotoxin A to the native structures of enterotoxin B and C,. Enterotoxin A is known to demonstrate greater biological activity than enterotoxin B. Consequently, the dissimilarity of enterotoxin A structure indicated by the isothermal denaturation data is probably of functional importance. -To obtain precise information on possible differences in conformational stabilities of the three enterotoxin variants, detailed kinetic analysis of enterotoxin unfolding in guanidine hydrochloride or urea solution was accomplished. The unfolding rates of the enterotoxins in guanidine solution varied widely. For example, unfolding of enterotoxin B in 3 M guanidine hydrochloride as followed by change in hze7 was not complete until about 8000 min of elapsed time (Fig. 2). In sharp contrast, unfolding of enterotoxin A reached equilibrium within 50 min. Enterotoxin C, required approximately 3000 min for unfolding in 3 M guanidine.
Semilog plots were linear over 90 to 95% of the unfolding curves of enterotoxin A in 1.5, 2, and 3 M guanidine hydrochloride (Fig. 3) and of enterotoxin C, in 2, 3, 4, 5 and 6 M guanidine hydrochloride (Fig. 4). The apparent rate constants of unfolding were obtained from the slopes of the straight line semilog plots and are reported in Table I Table I. The rate constants of unfolding were determined at the same concentration of the three enterotoxins (1 x lo-" M), and thus the values are directly comparable. Such a comparison reveals that enterotoxin A unfolding was 40-fold faster in 2 or 3 M guanidine hydrochloride solution than enterotoxin C, unfolding. However, the unfolding of enterotoxin C, in 4 or 5 M guanidine hydrochloride was only 4-to &fold faster than the unfolding of enterotoxin B; the unfolding rates of these two toxins were comparable in 6 M guanidine.
Likewise, enterotoxin C, unfolded with an apparent rate constant of 2.1 x 10m4 min' in 8 M urea. This value is very close to that previously reported for enterotoxin B in 8 M urea (2.0 x lo-" min') (81, whereas the apparent rate constant of enterotoxin A unfolding in 8 M urea (118.1 x 10m4 min-'1 was 50-fold greater than the unfolding rate constants of enterotoxin B or C, in the urea solution. Therefore, a large difference in kinetic stability toward isothermal denaturation by guanidine hydrochloride or urea appears to exist between enterotoxin B (the most stable) and enterotoxin A (the least stable). Despite the difference in the equilibrium stabilities of enterotoxin B and enterotoxin C,, these two enterotoxin variants are similar in kinetic stability toward isothermal denaturation.

DISCUSSION
Unfolding of enterotoxin B by 8 M urea followed by reduction and alkylation of the protein's single intramolecular disulfide bond and then extensive dialysis against neutral salt solution results in a folded enterotoxin B derivative lacking the kinetic stability toward isothermal denaturation observed for untreated native enterotoxin B (12). This interference with stable refolding of enterotoxin B by "bulky" carboxamidomethyl or carboxymethyl groups covalently positioned at half-cystines 92 and 112 indicates that the 92-112 disulfide bond is located in a region of toxin structure critical for native conformation. Carboxamidomethyl enterotoxin B unfolded in 2 M guanidine hydrochloride solution with an apparent rate constant of 44.1 x 10m4 min' (12). The unfolding behavior of the carboxamidomethyl derivative of enterotoxin B (12) is thus very similar to that reported in this paper for native enterotoxin A (Fig. 1, Table I). The important question /3 sheet at residues 89 to 94 and a p sheet at 111 to 118 (14). The primary structure of enterotoxin A in the immediate vicinity of its single disulfide bond is now available (15) and  can be applied to this region. The 6 amino acid residues contiguous to half-cystine 112 at positions 113 to 118 of enterotoxin B have been conserved in enterotoxin A (Met-Tyr-Gly-Gly-Val-Thr).
The p breaker Glu-119 of enterotoxin B is substituted for in enterotoxin A by the /3 former leucine; also, the p former threonine at position 111 of enterotoxin B is separated from the COOH-terminal side half-cystine of enterotoxin A by the weak /3 former alanine. Thus, the COOH-terminal side half-cystine of enterotoxin A closely resembles half-cystine 112 of enterotoxin B by being incorporated into a peptide sequence having a high probability for p sheet formation. Likewise, the 5 amino acid residues just proximal to the NH,-terminal side half-cystine of enterotoxin A (Tyr-Thr-Gly-Tyr-Gin) are p formers, with the exception of glycine which is a p indifferent residue. Including the threonine immediately distal to the NH,-terminal side halfcystine of enterotoxin A, the calculated value for the p sheet conformational parameter of this stretch of amino acid residues is 1.20 (a value ~1.05 indicates a high probability of p sheet formation for a given sequence of 5 or more amino acid residues (13)). This value is close to our previously reported value of 1.28 for /3 sheet 89 to 94 of enterotoxin B. It appears, therefore, that the disulfide bonds of both enterotoxin types A and B serve to covalently lock two /3 sheets in a stable, antiparallel configuration.
Consequently there are no obvious differences apparent for probable secondary structures in the immediate vicinity of the intramolecular disulfide bonds which might account for the difference in kinetic stability. However, Chou-Fasman analysis (13) of the enterotoxin B sequence also indicates a high probability for the incorporation of the 92-112 disulfide bond in a large core of repeating antiparallel p sheet from residues 81 to 148 (14). It is possible that deletion of strong /3 formers or substitution by amino acid residues with @ breaker properties (or both) more distal from the disulfide than those presently analyzed might prohibit formation of a stable repeating antiparallel p sheet around the intramolecular disulfide bond of enterotoxin A. The specific differences in amino acid sequence which could result in an absence of stable secondary structures in enterotoxin A can be detemined only when the complete sequence of this variant becomes known.
Unfolding kinetics in guanidine hydrochloride solution have been previously utilized in this laboratory to quantitate the effect of peptide bond "nicking" and disultide bond reduction on the conformation of enterotoxin B (121, and to demonstrate complementation between the 1 to 97 and 98 to 239 peptide chains of enterotoxin B (16). The present report demonstrates that the staphylococcal enterotoxins, although expressing qualitatively identical toxic and mitogenic activities (1, 41, vary greatly in kinetic stability. Although the near-and farultraviolet circular dichroism spectra of enterotoxin types B and C, are essentially identical (14, 17), a measurable difference in the kinetic stability of these two variants was found (Table I). Also, other studies have shown that trypsin cleavage of two peptide bonds in enterotoxin C, induced only minor arises as to whether or not native conformations vicinal to changes in the near-and far-ultraviolet circular dichroic the disulfide bond could explain the marked variation in spectra, but doubled the protein's unfolding rate in 4 M kinetic stabilities of the enterotoxin variants (Table I). Anal-guanidine (17). Thus, variation in kinetic stability appears to ysis of the primary amino acid sequence of enterotoxin B by approach antigenic reactivity (1) as a sensitive indicator of the technique of Chou and Fasman (13) suggests that the structural differences between natural and derivatized var-92-112 disulfide bond forms a covalent cross-link between a iants of enterotoxin.

Kinetic Stabi&
Comment can be made on the degree of structural homology between the enterotoxin variants. Significant quantitative differences in the biological activity of enterotoxin types A and B have been described. Enterotoxin A has been shown to demonstrate greater immunosuppressive activity than enterotoxin B. A dose of 0.1 wglml of enterotoxin A was found sufficient to suppress the in vitro humoral antibody responses of C57BL/6 spleen cells to sheep erythrocytes by greater than 90%; comparable inhibition was not obtained by enterotoxin B until toxin concentrations of 3 to 6 pg/ml (2). Also, the intravenous 50% effective dose of the A variant for enterotoxicity in rhesus monkeys is 0.03 pg/kg (0.017 to 0.065, 95% confidence limits) (lo), compared to a 50% effective dose of 0.10 pg/kg (0.05 to 0.20, 95% confidence limits) for the B variant (9). The quantitative difference in biological activity of enterotoxins A and B corresponds to the large difference in kinetic stability for the two enterotoxins (Table I). Rate constants for the unfolding of enterotoxin types B and C, are much closer in value than those reported for enterotoxin A (Table I). Biological data statistically equivalent to data reported for enterotoxins A and B are not available for enterotoxin C,. However, enterotoxin types B and C, closely resemble each other by the presence of peptide bonds highly susceptible to trypsin hydrolysis (18). Such labile peptide bonds are absent in the enterotoxin A molecule (18). Also, Spero and his colleagues (17) have recently reported that the NH,-terminal 6500-dalton fragment and its adjacent 4000dalton fragment obtained from enterotoxin C, by double cleavage with trypsin are very similar in amino acid composition to the equivalent peptide segments at positions 1 to 54 and 55 to 97 of enterotoxin B, respectively. Available biological and chemical data are compatible, therefore, with the kinetic data of Table I in suggesting a high degree of structural homology between enterotoxin types B and C,, and significantly less structural homology of these two enterotoxins with enterotoxin A. Since enterotoxin A is apparently a more active toxic and immunosuppressive agent than enterotoxin B (2, 9, lo), it will be important to specifically identify those r, Shambaunh.