Mutational effects on protein folding stability and antigenicity: the case of streptococcal pyrogenic exotoxin A

Abstract

The influence of mutationally induced changes in protein folding on development of effective neutralizing antibodies during vaccination remains largely unexplored. In this study, we probed how mutational substitutions of streptococcal pyrogenic exotoxin A (SPEA), a model bacterial superantigen, affect native conformational stability and antigenicity. Stability changes for the toxin variants were determined using circular dichroism and fluorescence measurements, and scanning calorimetry. Self-association was assayed by dynamic light scattering. Inactivated SPEA proteins containing particular combinations of mutations elicited antibodies in HLA-DQ8 transgenic mice that neutralized SPEA superantigenicity in vitro, and protected animals from lethal toxin challenge. However, a highly destabilized cysteine-free mutant of SPEA did not provide effective immunity, nor did an irreversibly denatured version of an otherwise effective mutant protein. These results suggest that protein conformation plays a significant role in generating effective neutralizing antibodies to this toxin, and may be an important factor to consider in vaccine design.

Introduction

Vaccines against protein toxins can often be generated from mutationally inactivated toxins produced in E. coli. However, the impact of these mutations on conformational stability, as it relates to effective antigenicity, remains relatively unexplored. Consequently, this study focused on examining the impact of mutant residues on protein folding, and subsequent vaccine design, against streptococcal pyrogenic exotoxin A (SPEA) [1]. SPEA1 is a major virulence factor released by Streptococcus pyogenes and is associated with scarlet fever and severe invasive infections [1], [2], [3], [4].

SPEA belongs to a class of toxins that are collectively referred to as superantigens. These proteins possess the ability to simultaneously bind to the major histocompatibility complex (MHC), outside of the peptide antigen presentation groove, and the Vβ chain of T-cell receptors (TCR). The resulting unrestricted interactions between T cells and antigen-presenting cells lead to the activation of a large population of T cells, as well as the hyperproduction of inflammatory cytokines. Subsequent physiological consequences include rash, hypotension, and, in severe cases, multiple organ failure. The continuing public health threat posed by streptococcal infections and the emergence of antibiotic-resistant strains of this species of bacteria indicate the need for a protective vaccine [5]. In accordance with this goal, efforts to develop a vaccine against SPEA from a recombinant, mutationally inactivated toxin mandate a better understanding of the relationships between structure, function, and protective antigenicity for this protein.

The X-ray crystal structure of SPEA has been solved at 2.6-Å resolution [6], revealing N- and C-terminal subdomains, and a surface loop constrained by a disulfide bond between residues C87 and C98. In addition, another cysteine (located at position 90) is also inside this loop (Fig. 1). Modeling and mutational analyses [7], [8] have predicted that the disulfide loop is important for interactions with both TCR and MHC in the ternary SPEA–TCR–MHC complex. A recent cocrystal structure of the C90S mutant of SPEA with a mouse TCR β chain [9] confirmed the involvement of loop residues in binding to TCR.

Chromatographic separation of recombinantly produced SPEA protein isolated at pH 7.9 revealed the presence of both monomer and dimer species, leading Papageorgiou et al. [6] to speculate that residue C90 in one toxin monomer forms a disulfide bond to the C90 of another SPEA molecule. We confirm here using dynamic light scattering measurements on a C90A mutant protein at pH 7.0 that this residue does indeed form such a disulfide-linked dimer. The SPEA protein examined by Papageorgiou and colleagues was isolated at pH 5.7 to inhibit disulfide formation and crystallized in an asymmetric unit of a tetramer, which has raised speculation that this toxin may function as such in vivo.

A zinc-binding site on SPEA at the locus of residues D77 and H106 has also been identified and confirmed by several groups [6], [9], [10], [11]. Baker et al. [11] found that residues E33, D77, H106, and H110 ligate zinc ion with a K_D of 2.3 μM. The significance of zinc for physiological function of SPEA is not yet clear. However, Hartwig and Fleischer (1993) [12] found that a SPEA D77A mutant protein was unable to bind class II MHC on the surface of antigen-presenting cells, while Baker et al. [11] suggested that the zinc site is a potential second MHC-binding locus, as judged from molecular modeling.

Other studies have identified additional SPEA mutations and residues that are important for the activity of this toxin. The naturally occurring substitution V76I, known as the SPEA3 allele [13], has been shown to possess an eightfold higher affinity for HLA-DQ than SPEA1 [7], [8]—even though it is not located at the predicted MHC-binding surface. Residues N20 [5], [14] and L24 have been predicted to be at or near the TCR interaction surface, while residue L42 is believed to be located in the SPEA contact site for MHC class II molecules [7], [8]. Residue N20 is observed to contact the TCR β chain in the structure of Sundberg et al. (2002) [9].

The objective of this study was to determine the effects of previously studied mutations on SPEA stability and test for a correlation with changes in the effectiveness of neutralizing antibodies elicited during vaccination. Knowledge of the significance of protein folding vs antigenicity gained from studying model systems such as SPEA may be valuable in designing more robust vaccines against bacterial SAgs and other proteins.