Identification of Cross-Neutralizing Epitopes on Toxic Shock Syndrome Toxin-1 and Staphylococcal Enterotoxin B by LIWINA TIN YAN PANG B.Sc, The University of British Columbia, 1995 A THESIS SUBMITTED IN PARTIAL FULFILMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF SCIENCE in THE FACULTY OF GRADUATE STUDIES (Department of Experimental Medicine) We accept this thesis as conforming to the required standard THE UNIVERSITY OF BRITISH COLUMBIA April 1999 ® Liwina Tin Yan Pang, 1999 In presenting this thesis in partial fulfilment of the requirements for an advanced degree at the University of British Columbia, I agree that the Library shall make it freely available for reference and study. I further agree that permission for extensive copying of this thesis for scholarly purposes may be granted by the head of my department or by his or her representatives. It is understood that copying or publication of this thesis for financial gain shall not be allowed without my written permission. Department The University of British Columbia Vancouver, Canada DE-6 (2/88) Abstract Toxic Shock Syndrome (TSS) is primarily caused by Toxic Shock Syndrome Toxin-1 (TSST-1), Staphylococcal Enterotoxin A (SEA), Staphylococcal Enterotoxin B (SEB), and Staphylococcal Enterotoxin C (SEC). These toxins belong to a family of pyrogenic toxin superantigens (PTAgs) produced by Staphylococcus aureus. These PTAgs share similar immunobiological properties, including the induction of massive release of cytokines and stimulation of T cell proliferation in a VP-specific manner. The crystal structures of most PTAgs are now known. They share a similar basic structure even though their primary sequences are different. Despite the structural and immunobiological similarities, no cross-reactivity between TSST-1 and other PTAgs has been demonstrated in serological assays and neutralization assays. Recently, our laboratory has developed a neutralizing murine anti-TSST-1 monoclonal antibody (MAb5) which demonstrated significant cross-reactivity with SEB in ELISA. Since previous studies have shown that MAb5 can neutralize TSST-l-induced superantigenic and lethal activities both in vitro and in vivo, the present study was undertaken to evaluate possible cross-neutralizing epitopes on TSST-1 and SEB using MAb5. Our experimental approach in addressing this question included: (1) assessment of the ability of MAb5 to cross-neutralize SEB-induced superantigenic activities in vitro and in vivo; (2) determination of the location of the cross-reactive epitopes on TSST-1 and SEB by epitope peptide mapping; (3) evaluation of the importance of the cross-reactive epitope as a functional domain by the use of synthetic peptides. MAb5 was found to partially inhibit SEB-induced T cell mitogenesis (63%) and TNFa secretion (70%) in human blood mononuclear cells (HPBMC) in a dose dependent ii manner, while a control anti-TSST-1 monoclonal antibody, MAb6, had no effect. In order to locate the cross-reactive epitopes on TSST-1 and SEB, binding of MAb5 to a set of overlapping decapeptides homologous to TSST-1 (1-194) and SEB (1-239) were determined. MAb5 recognized a linear sequence of 5 amino acids ( " Y S P A F 5 6 ) on TSST-1, as well as a linear sequence of 4 amino acids ( 8 5 F G A N 8 8 ) on SEB. These epitopes are structurally similar but are topologically located in different areas of the respective toxins. Furthermore, a 10 mer peptide corresponding to SEB (83-92), 8 3 D W G A N Y Y Y Q 9 2 , was found to inhibit SEB-induced T cell mitogenesis (90%) and T N F a secretion (90%) from H P B M C in vitro, while an irrelevant, scrambled peptide had no effect. These data suggest that MAb5 recognizes structurally similar motifs of SEB and TSST-1, and contains neutralizing activity against these related PTAgs. iii TABLE OF CONTENTS Page Abstract ii Table of contents iv List of Tables viii List of Figures ix List of Abbreviations xi Acknowledgement xiii Chapter 1. Literature Review 1 1.1. Introduction 2 A. Historical Overview 2 B. Hallmark of Superantigens 3 C. Classes of Superantigens 5 I. Virally Encoded Superantigens 5 II. Bacterial Superantigens 6 1.2. Pyrogenic Toxin Superantigens 7 A. Shared Immunobiological Properties 10 I. Pyrogenicity 10 TI. Enhancement of Lethal Endotoxic Shock 10 m. Superantigenicity 11 B. Unique Properties 11 I. Toxic Shock Syndrome Toxin-1 11 iv n . Staphylococcal Enterotoxins 14 HI. Streptococcal Pyrogenic Exotoxins 15 1.3. Structural Characteristics of PTSAgs 15 1.4. M H C Class II Molecules - The Receptor for Superantigen 19 1.5. M H C Class II Binding Sites of PTSAgs 22 A. PTSAgs Bind M H C class H Molecule Using Common Structural 22 Strategies B . SEA, SEC2, SED, and SEE : The Role of Zinc 29 1.6. TCR Binding Sites of PTSAgs 30 1.7. Superantigens and Human Diseases 3 5 A. Biolological Responses to Superantigens 35 B. PTSAgs and Disease Association 36 I. Toxic Shock Syndrome 36 n . Streptococcal Toxic Shock Syndrome 38 HI. Other Diseases 39 C. Potential Therapeutic Applications of Superantigens 41 Chapter 2. Materials and Methods 43 2.1. Purification of TSST-1 and SEs 44 2.2. Silver Staining and Western Blot 46 2.3. Quantitation of TSST-1 and SEs by Immunoassay 47 2.4. Preparation of Monoclonal Antibodies 49 2.5. Characterization of Monoclonal Antibodies 49 v 2.6. Purification of Human peripheral blood mononuclear cells 50 2.7. Mitogenicity assays 50 2.8. Cytokine assays 51 2.9. Animal studies 52 2.10. Synthesis of support-coupled peptides 53 2.11. ELISA Testing 54 2.12. Inhibition of SEB-induced mitogenesis by synthetic peptides 55 2.13. Inhibition of SEB-induced TNFa production by synthetic peptides 56 2.14. Statistical Analysis 56 Chapter 3. Cross-Neutralization of Toxic Shock Syndrome Toxin-1 and 58 Staphylococcal Enterotoxin B by Anti-TSST-1 Monoclonal Antibodies. 3.1. Introduction 59 3.2. Results 63 3.3. Discussion 72 Chapter 4. Identification of Cross-Neutralizing Epitopes on Toxic Shock 79 Syndrome Toxin-1 and Staphylococcal Enterotoxin B by Epitope Peptide Mapping and Synthetic Peptide Inhibition Studies. 4.1. Introduction 80 4.2. Results 82 4.3. Discussion 91 vi Chapter 5. Summary o f Results and Conclus ions R E F E R E N C E S vii List of Tables Table! Biochemical Properties of TSST-1, Staphylococcal 8 Enterotoxins and Streptococcal Pyrogenic Exotoxins Table U . Biological Properties of Pyrogenic Toxin Superantigens 9 Table DT. VP Specificities of Staphylococcal Pyrogenic Toxin Superantigens 12 Table IV. The TSST-1:DR1 and SEB:DR1 interface 28 Table V . Contacts between the TCR VB 8.2 chain and Staphylococcal 31 Superantigens Table VI. Characterization of Anti-TSST-1 MAb5: reactivity in E L I S A and 64 Western Blot. Table V B . Neutralization studies of MAb5 on SEB-induced lethality in a D - 71 galactosamine sensitized murine model of lethal shock. viii List of Figures Figure 1. Hypothetical structure for the complex of class I I M H C , TCR, 4 and superantigens as proposed by Marrack and Kappler, 1990. Figure 2. Three-dimensional structure ribbon diagrams of PTSAgs. 16 Figure 3. Ribbon diagram of TSST-1. 17 Figure 4. Common structural strategies used by SEA, SEB, and TSST-1 23 for binding to M H C class II a subunits. Figure 5. Structures of complexes between HLA-DR1 and the superantigens 26 SEB and TSST-1. Figure 6. Ribbon diagram of TSST-1 shown in back-side view indicating 34 mutations which lead to alteration of biological activities Figure 7. Dose response of SEB-induced T N F a secretion from human 65 P B M C . Figure 8. Effect of anti-TSST-1 Mab5 on TSST-l-induced T N F a secretion 66 from human P B M C . Figure 9. Dose response of SEB-induced mitogenesis of human P B M C . 68 Figure 10. Effect of anti-TSST-1 Mab5 on TSST-l-induced human P B M C 69 mitogenesis. Figure 11. Dose Response of SEB-induced lethality in D-galactosamine 70 sensitized murine model. Figure 12. Effect of anti-TSST-1 MAb5 on TSST-l-induced mitogenesis of 74 human P B M C . Figure 13. Effect of anti-TSST-1 MAb5 on TSST-l-induced T N F a secretion 75 from human P B M C . Figure 14. Conserved regions of primary structure among sequences in the 77 PTSAgs family. Figure 15. Scans of TSST-1 (46-70) decapeptides with (a) anti-TSST-1 MAb5; 83 (b) isotypic control, IgGl . IX Figure 16. Window/ window + 1 analysis on TSST-1 epitope T(47-56). 84 Figure 17. Scans of SEB (1-239) decapeptides with (a) anti-TSST-1 MAb5; 86 (b) isotypic control, IgGl . Figure 18. Window/ window + 1 analysis on SEB epitope (83-92). 87 Figure 19. Effect of synthetic peptide SEB(83-92) on SEB-induced mitogenesis 89 of human PBMC. Figure 20. Effect of synthetic peptide SEB(83-92) on SEB-induced T N F a 90 secretion from human P B M C . Figure 21 Structural motif of (a) T S S T - 1 5 2 Y S P A F 5 6 , (b) SEB M V F G A N 8 8 , 92 (c) SEB 8 5 F G A N Y . x List of Abbreviations APC Antigen Presenting Cell B S A Bovine serum albumin BHI Brain Heart Infusion Broth °C Degree Celsius E D T A Ethylene diaminetetra acetic acid E L I S A Enzyme-linked Immunosorbent Assay FBS Fetal bovine serum h Hour 125j Iodine-125 Ig Immunoglobulin TL-1 Interleukin-1 IL-2 Interleukin-2 TL-4 Interleukin-4 IL-6 Interleukin-6 IL-7 Interleukin-7 IFN-y Interferon-gamma kb kilobase kd kilodalton K d Dissociation constant LPS Lipopolysaccharide M A b Monoclonal antibody M A M M . arthriditis mitogen min Minute MHCII Major Histocompatibility Complex Class II Antigen M L S Minor Lymphocytes Stimulating ng/ml nanogram per millitre n M nanomolar O D . Optical density P B M C Peripheral blood mononuclear cells PBS Phosphate buffered saline pg/ml picogram per millitre P i Isoelectric point p M picomolar PTSAgs pyrogenic toxin superantigens r.m.s. root mean square SAgs superantigens SE Staphylococcal enterotoxin SEA Staphylococcal enterotoxin A SEB Staphylococcal enterotoxin B SEC Staphylococcal enterotoxin C SED Staphylococcal enterotoxin D SEE Staphylococcal enterotoxin E SEF Staphylococcal enterotoxin F xi SEH Staphylococcal enterotoxin H SDS-PAGE Sodium dodecyl sulfate polyacrylamide gel electrophoresis S. aureus Staphylococcus aureus S. pyrogenes Streptococcus pyrogenes SPEA Streptococcal pyrogenic A SPEB Streptococcal pyrogenic B SPEC Streptococcal pyrogenic C TCR T cell receptor TNF Tumor necrosis factor TSS Toxic shock syndrome TSST-1 Toxic shock syndrome Toxin-1 V Volts vp Beta chain of variable region of T cell antigen receptor xii Acknowledgement I would like to thank Dr. Anthony Chow for his encouragement, advice and guidance throughout my studies. I would also like to acknowledge Dr. Winnie Kum, who aided me considerably over the course of research, especially in the in vitro assays, toxin purification, and animal model studies. I gratefully thank Anita Kwok for her advice in our lengthy discussions and her continual support and encouragement. I would like to express my gratitude to Dr. Natalie Strynadka and Mr. Steven Ness at the Department of Biochemistry for their knowledgeable contributions in the epitope mapping studies. Their time and effort is much appreciated. Special thanks are extended to Dr. Hermann Ziltener and Dr. Lawrence Mcintosh for serving as members of my advisory committee; and Ryan Hung for helping me with the computer programs. Last but not least, I would also like to thank my good friend, Soo Sen Lee, who shared with me the joy and tears during my studies; my fiancee, Alex Chan, who put up with my frustrations during those difficult times; my sister, Loretta Pang, who gave me her full support by being my blood donor; and my parents, who encouraged me throughout the entire studies. Without their continual support and encouragement, this thesis would never be accomplished. xiii Chapter 1 Literature Review 1 1.1. Introduction A . Historical Overview The term "superantigen" was first introduced by White et al. in 1989 to describe a group of bacterial toxins which induced massive proliferation of T cells in a VT3 specific manner (White et al., 1989). This was quickly followed by the insight that minor lymphocytes stimulating (MLS) antigens, a class of murine endogenous antigens which were known since the early 1970s, had similar functional properties (Kappler et al., 1988 and MacDonald et al., 1988). M L S antigens were discovered by Festenstein in 1973 because of their enormous T cell stimulatory capacity in mixed lymphocyte reactions with cells from major histocompatibility complex (MHC)-identical strains of mice (Festenstein, 1973). Since the genes encoding M L S antigens were mapped outside the M H C loci, they were named "minor" lymphocyte stimulating antigens to differentiate from M H C antigens. However, the in vivo consequences of expression were not realized until 1988, when two groups of investigators reported the absence of VP8.1 and Vf56 bearing T cell subsets from the peripheral lymphoid tissues in mice expressing M L S antigens (Kappler et al., 1988; MacDonald et al., 1988). Clearly, the M L S antigens were recognized by the T cell receptor (TCR), and the major determinant of M L S reactivity appeared to be the expression of a particular VP gene segment. This property distinguished M L S antigens from conventional peptide antigens, in which their recognition by T cells require all the variable elements of TCR (i.e. V a , Ja, VP, DP, and jp regions). Furthermore, M L S antigens were recognized as self-antigens since M L S reactive T cell subsets are deleted from the repertoire during thymic maturation as part of the self-tolerance process (Kappler et al., 1988). 2 Staphylococcal enterotoxins (SEs), toxic shock syndrome toxin-1 (TSST-1) and streptococcal pyrogenic exotoxins (SPEs), which were known for years as the major cause of food poisoning and toxic shock syndrome in human, were found to have immunobiological properties with striking similarities to M L S antigens. In fact, as early as 1970, Peavy et al. have reported the ability of staphylococcal enterotoxin A (SEA) and staphylococcal enterotoxin B (SEB) to induce T cell mitogenesis of mouse spleen cells and human peripheral lymphocytes (Peavy et al., 1970). The significance of these T cell responses and the superantigenic nature of these toxins were not, however, appreciated immediately. This breakthrough in understanding was dependent on the development of monoclonal antibodies (MAbs) that are specific for T cells expressing a particular VP domain. With such MAbs, these toxins were found to stimulate massive proliferation of T cells expressing a particular set of VP gene segment and injection of SEs into neonatal mice produced particular T cell VP subset deletion as in M L S (Janeway, 1989 and White et al., 1989). As a result, the connection between M L S antigens and these bacterial toxins was established and they were all designated as "superantigens". B. Hallmark of Superantigens Superantigens differ from conventional antigens in several ways: (1) Superantigens do not require processing by antigen presenting cells (APCs). In contrast, conventional antigens are internalized and processed by APCs before being loaded into the peptide binding groove formed by the a- and P- chains of class n M H C molecule. (Dellabona et al., 1990; Fleischer and Schrezenmeir, 1988; Fleischer, 1991) (Fig. 1); (2) Superantigens bind directly to relatively invariant regions on M H C class n molecules outside the conventional peptide-binding groove, and thus presentation to T cells is not class H 3 Tcell Antigen Presenting Cell Figure 1. Hypothetical structure for the complex of M H C class TI molecule, TCR, and superantigens as proposed by Marrack and Kappler, 1990. Figure taken from Herman et al., 1991. 4 restricted, violating the rule of M H C restriction seen in all other peptide-antigen-specific T cell responses (Dellabona et al., 1990; Kotzin et al., 1993); (3) The recognition of superantigen/MHC class II complex by TCR depends entirely on the variable domain of the TCRP chain (VT3), while a conventional peptide antigen requires a very specific interaction with the third hypervariable region of the TCR (Marrack and Kappler, 1990). This latter region of the TCR consists of the joining elements of both V J a and VDJP; thus, the responding frequency to a conventional antigen is usually about 1 in 10,000. On the other hand, since the relative number of VP genes is limited, a given superantigen is capable of interacting with as many as 5 to 30% of the entire T cell repertoire (Drake and Kotzin, 1992). C. Classes of Superantigens I. Virally Encoded Superantigens Superantigens are a diverse collection of molecules that share the ability to activate specific T cell subsets. Two groups of superantigens have been described, namely viral and bacterial superantigens. At the beginning of the 1990s, the exciting discovery was made that murine M L S antigens are encoded by the murine mammary tumor virus (MMTV) proviral D N A that had been integrated into the germline (Frankel et al., 1991 and Dyson et al., 1991). Expression of endogenous superantigens during T-cell ontogeny in mice causes physical and functional deletion of T cells bearing VP elements specific for those antigens (MacDonald et al., 1988; Woodland et al., 1991; Pullen et al., 1988). However, to date, an analogous system has not been found in humans. While only the MMTV-encoded viral superantigen has been well defined, there are reports of association of superantigen activity with other viruses. The nucleocapsid of rabies virus has been shown to be a superantigen 5 in humans and also in mice (Lafon et al., 1992 and Lafon et al., 1994). In addition, superantigenic activity has been attributed to two human tumor viruses, cytomegalovirus (CMV) and Epstein-Barr virus (EBV) (Dobrescu et al., 1995; Sutkowski et al., 1996). Furthermore, a number of different groups have proposed that human immunodeficiency virus (HTV), the cause of human acquired immunodeficiency syndrome (ADDS), encodes a superantigen that may be involved in host CD4 + T cell depletion (Dadaglio et al., 1994; Akolkar et al., 1995). Tl. Bacterial Superantigens A large number of bacterial toxins made up the growing list of bacterial superantigens. Bacterial superantigens can be subdivided into pyrogenic toxin superantigens (PTSAgs) and those that are not pyrogenic toxins. The PTSAgs, including TSST-1, SEs, and group A SPEs, are probably the most well studied superantigens. PTSAgs are distinguished from other bacterial superantigens due to their shared three-dimensional structure and immunobiological properties, which will be discussed in detail later. Superantigens that do not belong to the PTSAgs family include staphylococcal exofoliative toxins (ExFT), Mycoplasma arthritidis mitogen (MAM), Yersinia enterocolitica superantigen (YES), Yersinia pseudotuberculosis mitogen (YPM), and streptococcal M protein (Schlievert, 1997; Stuart and Woodland, 1992; Abe et al., 1993; Tomai et al., 1990). 6 1.2. Pyrogenic Toxin Superantigens (PTSAgs) PTSAgs form a subfamily within the larger superantigen family that includes a large group of simple, nonglycosylated proteins, secreted by Staphylococcus aureus, group A streptococci, and certain non-group A streptococci. Well characterized members of this family include TSST-1 and SEs serotypes A , B , Cn, D, E , G, and H produced by S. aureus; SPEs serotypes A , B , C, and F (mitogenic factor), and streptococcal superantigen (SSA) produced by group A streptococci (Bohach et al., 1990; Schlievert et al., 1995). They are grouped under the same subfamily based upon their shared biochemical, immunobiological, and three-dimensional structural properties, as well as their involvement in related illnesses such as Toxic shock syndrome (TSS) and other immune-mediated diseases (Kotzin et al., 1993). Biochemically, these toxins are small polypeptides with a neutral to basic isoelectric point (Blomster-Hautamaa et al., 1986; Bergdoll, 1993; Schlievert et al., 1990). They are generally resistant to acid, heat, and protease (Bohach et al., 1990). The biochemical properties of the PTSAgs are summarized in table I. In addition to sharing biochemical properties, the PTSAgs share many immunobiological activities, such as the ability to induce fever, to enhance host susceptibility to endotoxin shock, and to stimulate T cell proliferation via a VP specific manner (Bohach et al., 1990). While all PTSAgs possess these common immunobiological properties, each subclass of toxins exhibits unique characteristics as well. The shared and unique biological activities of PTSAgs are summarized in Table II. 7 Table I. Biochemical properties of Toxic Shock Syndrome Toxin-1, Staphylococcal Enterotoxins, and Streptococcal Pyrogenic Toxins. Toxin Isoelectric Point Number of Amino Acids (mature protein) Molecular Weight (kDa) TSST-1 7.2 194 22 049 SEA 7.3 233 27 100 SEB 8.5 239 28 336 SEC1 8.5 239 27 531 SEC2 7.0 239 27 589 SEC3 8.0 239 27 563 SED 7.4 228 26 360 SEE 7.0 230 26 425 SEH 5.65 217 25 210 SPEA 5.0-5.5 221 25 787 SPEB 8.0-9.0 253 27 588 SPEC 6.7-7.0 208 24 354 Derived from Bergdoll, 1983; Schlievert et al., 1990; Hauser and Schlievert, 1990, and Ren et al., 1994. 8 Table H. Biological Properties of Pyrogen Toxin Superantigens Subfamilies of PTSAgs Property TSST-1 Staphylococcal Enterotoxin Streptococcal Pyrogenic Exotoxins Pyrogenicity + + + Enhancement of endotoxin Shock + + + Superantigenicity + + + Interference with liver clearance function + + + Emesis - + -Cardiotoxicity + - • -Ability to bind endothelial cells + ? + Source: Bohach et al., 1990; Leonard and Schlievert, 1992; Schwab et al., 1995; Lee et al.,1991. 9 A. Shared Immunobiological Properties I. Pyrogenicity The term "pyrogenic toxin superantigens" is used for this class of superantigens because they are among the most potent pyrogens known. They are capable of inducing fever when injected into experimental animals (Kim and Watson, 1970; Schlievert et al., 1982). It is believed that PTSAgs cause fever through inducing the release of human endogenous pyrogens, interleukin-1 (IL-1) and tumor necrosis factor (TNF), from macrophages, and by having direct effects on the hypothalamus fever response control center (Fast et al., 1989; Bohach et al., 1990; Schlievert et al., 1995). PTSAgs' distinctive fever response can easily be distinguished from other pyrogens based upon the time of maximal response. PTSAgs-induced fever is characterized by a fairly linear rise of body temperature with a peak at 4 hour after injection, in contrast to the biphasic fever response to endotoxins, another well known pyrogen, with peaks at 1 and 3 hour post-injection (Bohach et al., 1990). II. Enhancement of Lethal Endotoxin Shock Another important biological property shared by all of the PTSAgs is the ability to enhance host susceptibility to the lethal effect of gram-negative bacterial endotoxin by up to 100,000-fold (Bohach et al., 1990). The precise mechanism underlying the enhancement phenomenon remains to be eludicated but appears to involve PTSAgs interference with liver clearance function (Schlievert et al., 1995). This is supported by the observation of fatty acid replacement of the liver on autopsy in staphylococcal TSS patients (Larkin et al., 1982; Stone and Schlievert, 1987). Moreover, SPEs have been 10 previously shown to inhibit R N A synthesis in isolated liver cells and their nuclei in SPEs-treated rabbits (Schlievert et al., 1979). However, whether PTSAgs blockage of R N A synthesis is the direct effect amplifying susceptibility to endotoxins is unclear. III. Superantigenicity The most intensively studied biological property of the PTSAgs at present is their capacity to induce dramatic T cell proliferation, which is referred to as superantigenicity. PTSAgs are able to stimulate both CD4 + and C D 8 + T cells regardless of their antigenic specificity (Kotzin et al., 1993). Interestingly, different PTSAgs have different specificities for the variable regions of the VJ3 region of the TCR (Table HI). For example, TSST-1 is capable of preferentially stimulating the proliferation of human T-cell subset bearing V02, which accounts for approximately 10% of all T cells (Choi et al., 1990). In an acute TSS patient, as many as 60-70% of all circulating T cells may display T cell receptors containing VP2 segment (Choi et al., 1990). The non-antigen-specific T cell stimulation elicit massive release of cytokines, including interleukin (TL)-la, -2, -4, -6, -10, interferon (IFN)-y, tumor necrosis factor (TNF)-a and B, and subsequently lead to capillary leak and hypotension in the host (Miethke et al., 1993). It is postulated that the uncontrolled release of cytokines underly the pathogenesis of TSS (Kain et al., 1993). B. Unique Properties I. Toxic shock syndrome toxin-1 (TSST-1) TSST-1 was first identified by Schlievert et al. as staphylococcal pyrogenic exotoxin C (Schlievert et al., 1981) and by Bergdoll and his coworkers as enterotoxin F 11 Table HI. VP Specificities of Staphylococcal Pyrogenic Toxin Superantigens. Toxin VP Specificity Human Mouse SEA 1.1,5.3, 6.3,6.4,6.9, 7.3 7.4, 9.1, 18 1,3, 10, 11, 12, 17 SEB 3, 12, 14, 15, 17, 20 3,7,8.1,8.2, 8.3,17 SEC1 3, 12 3, 8.2, 8.3, 11, 17 SEC2 12, 13.1, 13.2, 14, 15, 17,20 3, 8.2, 10, 17 SEC3 5,12 3,7,8.1,8.2 SED 5,12 3,7,8.2, 8.3, 11, 17 SEE 5.1,6.1, 6.2, 6.3, 8, 18 11, 15, 17 TSST-1 2 3, 4, 15, 17 Source: Marrack and Kappler, 1990; Herman et al., 1991, and Uchiyama et al., 1994. 12 (Bergdoll et al., 1981) independently in 1981 as the causative agent of staphylococcal TSS. When staphylococcal pyrogenic exotoxin C and enterotoxin F were later shown to be identical, the term "toxic shock syndrome toxin-1" (TSST-1) was adopted to describe the toxin (Bergdoll and Schlievert, 1984). TSST-1 is produced by nearly 100% of the S. aureus isolated from vaginal or cervical cultures in cases of menstruation-associated TSS, and approximately 50% of S. aureus isolates obtained from other body sites in nonmenstrual TSS cases (Bergdoll and Schlievert, 1984; Schlievert, 1986). The toxin is encoded by the gene tst present in the chromosome as a large heterologous insert (Blomster-Hautamaa et al., 1986). tst occurs in two major forms, tsfR, where H refers to human isolate of S. aureus and the gene encodes TSST-1, and tstO where O refers to ovine isolate of S. aureus and the gene encodes TSST-0 (Blomster-Hautamaa et al., 1986; Lee et al., 1992). TSST-1 and TSST-O are translated as proteins consisting of 234 amino acids of which the first 40 residues comprise signal peptides (Blomster-Hautamaa et al., 1986; Lee et al., 1992). The mature secreted toxins are single polypeptide chains with molecular weights of approximately 22,000 and isoelectric points of 7.2 and 8.5 for TSST-1 and TSST-O, respectively (Blomster-Hautamaa et al., 1986; Schlievert et al., 1995). These two proteins differ from each other by seven amino acids at position 19, 55, 57, 69, 80, 132, and 140 (Lee et al., 1992). The consequence of these differences is that TSST-1 is biologically active whereas TSST-O lacks significant biological activity (Lee et al., 1992). These toxins have no cysteines and thus no internal disulfide loop, contain a high percentage of hydrophobic amino acids, and is generally resistant to heat and proteolysis (Schlievert et al., 1995). TSST-1 shares all the biological properties as SEs except the ability to induce emesis as characterized by the SEs. 13 Furthermore, TSST-1 shares little sequence similarities with other members of the PTSAgs (Schlievert et al., 1995). II. Staphylococcal Enterotoxins The SEs form a subgroup of related toxins in the PTSAgs family. They are distinguishable from other members of the PTSAgs by their ability to induce emetic response when ingested (Bohach et al., 1990). This unique feature of the SEs has long been recognized since the 60s. Hence, they are the causative agents in staphylococcal food poisoning (SFP), a very common form of food-associated gastroenteritis in the United States and worldwide (Holmberg and Blake, 1984). In recent years, SEs have received more attention for causing TSS (Schlievert, 1986). It is now known that these toxins, particularly SEB and SEC, are responsible for 50% of nonmenstrual toxic shock syndrome (Murray et al., 1995). Barber is generally credited for providing the first evidence for enterotoxin production by Staphylococcus aureus in 1914 (Holmberg and Blake, 1984). To date, seven major antigenic forms of SEs, including SEA, SEB, SEC, SED, SEE, SEG, and SEH, have been identified based primarily on their antigenic heterogeneity (Schlievert, 1997). Based on minor differences in antigenicity, SEC can be further differentiated into three subtypes (SEC1, SEC2, and SEC3). Comparison of primary amino acid sequences indicates that SEs can be divided into 2 groups: group 1 contains SEA, SED, SEE, and SEG; group 2 consists of SEB and SEC subtypes (Betley et al., 1992). 14 UJ. Streptococcal Pyrogenic Exotoxins (SPEs) SPEs have been known to cause scarlet fever since the beginning of the century (Dick and Dick, 1924). For years, it was thought that the only property of these toxins was rash production. In the early 1960s, the discovery of their distinctive fever-producing ability led Watson to name these proteins streptococcal pyrogenic exotoxins (SPEs) A , B , C (Watson, 1959). Although the T lymphocyte proliferative effects of the toxins have been known since 1970s (Hanna and Watson, 1973), their superantigenicity was recognized only approximately 20 years later when investigators noted their specificity for the VP chain of the TCR (Marrack and Kappler, 1990). It is now known that these toxins are also responsible for streptococcal toxic shock syndrome (Cone et al., 1987 and Stevens et al., 1989). Unlike the SEs, the SPEs share relatively little primary sequence homology with each other or with other PTSAgs. SPEA is an exception, which shares 45-50% sequence similarity with SEB and the three SEC subtypes (Schlievet et al., 1995). 1.3. Structural Characteristics of PTSAgs In addition to their shared immunobiological activity, the PTSAgs also have many common structural features even though there is little primary sequence similarities among the PTSAgs (Fig. 2). Recently, the three-dimensional structures of SEA (Schad et al., 1995), SEB (Swaminathan et al., 1995), SEC2 (Papageorgiou et al., 1995), SED (Sundstrom et al., 1996), TSST-1 (Prasad et al., 1993; Acharya et al., 1994; Papageorgiou et al., 1996) and SPEC (Roussell et al., 1997) were determined. These PTSAgs share strikingly similar architecture as typified by the ribbon structure of TSST-1 shown in Fig. 3. The TSST-1 structure is the simplest of the family. Basically, it is a two-domain kidney 15 Three dimensional structure ribbon diagram of (A) toxic shock syndrome Toxin-1 (TSST-1); (B) Staphylococcal Enterotoxin A (SEA); (C) Staphylococcal Enterotoxin B (SEB); (D) Staphylococcal Enterotoxin C 3 ( S E C 3 ) , as determined by X-ray crystallographic analyses. Domain A and B of each toxins are shown on the left and right of each figure respectively. 16 Figure 3. Ribbon diagram of TSST-1. Helices are labeled A-D and are represented as coils. P strands are labeled 1-13 and are shown as arrows. From: Prasad et al., 1993. 17 shaped molecule containing a long, diagonal, solvent accessible a helix spanning the center of the molecule. The C-terminal domain (domain A), the larger of the two domains, is a so-called P-grasp motif composed of a long central a helix resting against a five-stranded P sheet (Schlievert et al., 1995; Overington, 1992). Amino acid residues in domain A of TSST-1 (Murray et al., 1996; Hurley et a l , 1995), SEA (Hudson et al. 1993), SEB (Jardetsky et al., 1994), SEC (Li et al., 1998), and SEE (Hudson et al., 1993) have been implicated in TCR binding. The N-terminal domain (domain B) is characterized by a "claw-like" barrel structure composed of 5 P strands. Crystallographic studies of the three-dimensional structure of PTSAgs complexed with M H C n molecules have revealed that specific regions of domain B are responsible for contacting the M H C class II molecules (Jardetsky et al., 1994; Kim et al., 1994). Key features of the TSST-1 molecule are the two grooves, referred to as the front- and backside grooves formed by the central a helix. The N-terminal a helix and its subsequent loop located on the opposite sides of the central a helix define the wall of these grooves. The larger backside groove is more exposed to the environment than the front side groove. Mutational analysis have indicated that residues within the back-side-groove, including amino acids in the central and N-terminal a helix, are important for T cell interaction (Murray et al., 1994; Deresiewicz et al., 1994). Despite the many overall structural similarities, each toxin seems to have its unique structural features. Close analysis of their structures reveals 4 features not found in the TSST-1 structure: (Schlievert et al., 1995 and Earhart et al., 1998): (1) an elongated N -terminus of approximately 20 amino acids preceding the amino-terminal a-helix; (2) a second a helix sits between domain A and domain B ; (3) a highly flexible "disulfide loop" at the top of domain B ; (4) an a-helix at the bottom of domain B . It is speculated that 18 TSST-1 contains the minimal structural requirements for superantigenicity. Additional structural features found in SEs may be associated with their unique sub-group specific biological activities. For example, the disulfide loop present in all SEs but missing in TSST-1 has been implicated in the emetic properties of the SEs, which is absent in TSST-1 (Schlievert et al., 1995). Site-directed mutagenesis studies of SEC1 suggested that local conformations near or within the disulfide loop appear to be important for emetic activity (Hovde et al., 1994). 1.4. M H C class H molecules - The Receptor for Superantigens Convincing evidence has been provided that staphylococcal superantigens (SSAgs) bind specifically to class IT M H C molecules, and that this interaction is important for T cell stimulation (Fleischer and Schrezenmeier, 1988; Fischer et al., 1989; Mollick et al., 1989; Scholl et al., 1989). Although binding to the M H C class II molecules is a shared property among SSAgs, a significant degree of heterogenity is observed in regard to: (1) affinity of PTSAg binding to M H C class II; (2) requirement of metal atoms; (3) repertoire of compatible M H C class II molecules; (4) molecular regions of M H C class II molecule and PTSAgs involved in binding. The dissociation constants for the binding of staphylococcal superantigens to 5 S human M H C class TI molecules were estimated to range from 10" - 1(T M (Chintagumpala et al., 1991). The binding affinities of TSST-1, SEC1 and SEB for H L A -D R are lower than that of SEA but substantially higher than those of SED or SEE (Mollick et al., 1991 and Chintagumpala et al., 1991). A correlation is noted between the binding affinities of these toxins to M H C class II molecules and the toxins' potencies of T cell 19 proliferation. The affinity of the staphylococcal superantigens for M H C class II molecules is usually inversely correlated with the amount of toxins required for efficient T cell stimulation; i.e. toxins which required the lowest concentrations to elicit T cell responses displayed the highest affinity for HLA-DR1 (Mollick et al., 1991). This is true for SEA, SEB, SEC1 and SED but not for SEE and TSST-1 (Mollick et al., 1991; Chintagumpala et al., 1991). SEA, SEC2, SED, and SEE, but not SEB and TSST-1, have been found to possess a zinc binding site (Fraser et al., 1992; Schad et al., 1995; Sundstrom et al., 1996; Papageorgiou et al., 1995). Earlier biochemical studies demonstrated that binding of radiolabeled SEA to M H C class II molecules was completely inhibited by E D T A , indicating that divalent cations were important for this interaction (Fraser et al., 1992). Zinc ions, but not magnesium ions, calcium ions, manganese ions, ferrite ions, and copper ions, were able to restore completely the binding of SEA to H L A - D R (Fraser et al., 1992). Moreover, zinc ions bound directly to SEA but not to H L A - D R , suggesting that they are required to stabilize the class II binding domain on SEA (Fraser et al., 1992). The ability of various human and murine M H C class II molecules to present PTSAgs has been well documented (Scholl et al., 1989; Mollick et al., 1991; Scholl et al., 1990; Mollick et al., 1989). Although the presentation of PTSAgs is generally described as not being restricted to particular M H C class II proteins, the efficacy of presentation of PTSAgs to T cell varies when different alleles and isotypes of M H C class n molecules are used (Herrmann et al., 1989; Scholl et al., 1990). In general, H L A - D R is the most effective human H L A molecule for PTSAg presentation (Herman et al., 1990). However, alleles of H L A - D R were shown to have heterogeneity in binding to particular PTSAgs. SEA was 20 shown to bind to most H L A - D R molecules except for the DRw53 molecules (Herman et al., 1990). Yet, DRw53 displays an intact affinity for SEB or SEC (Herman et al., 1991; Karp and Long; 1992; Herrmann et al., 1989). Herman et al determined that there was a hierarchy of SEs presentation to human M H C class H molecules with H L A - D R > H L A -D Q > H L A - D P (Herman et al., 1990). However, SEC seems to be an exception to this rule. SEC2 binds predominantly to H L A - D Q and SEC 3 to H L A - D R and H L A - D Q (Herrmann et al., 1989). Similarly, studies of the murine analogue of H L A - D R , I-E, demonstrated consistently better presentation of toxins than the murine analogue of H L A -DQ, I-A (Yagi et al., 1990). However, I-A alleles were shown to bind TSST-1 more efficiently than I-E molecules (Scholl et al., 1990). Alleles of murine class n molecule I-A were also shown to have different capacity to bind TSST-1. Specific binding of TSST-1 to I-A d and I -A k was observed but transfected cells expressing I -A b did not bind this toxin (Scholl et al., 1990). Competition studies suggest that multiple overlapping or independent M H C class H binding sites exist for superantigens (Scholl et al., 1989; Chintagumpala et al., 1991; See et al., 1990). SEA competes with SEB and TSST-1 for binding to H L A - D R but a mixture of TSST-1 and SEB does not display S E A binding to H L A - D R , suggesting that SEA has multiple binding sites, one of which overlaps with SEB and TSST-1, and one that is unique (Fraser, 1989; See et al., 1990). Although the class II binding sites of TSST-1 and SEB (as well as SEA) overlap, the toxins do not compete with each other because they bind different subsets of DR1 molecules (Scholl et al., 1989). This data suggests that there are at least two binding sites for staphylococcal superantigens on class II molecules, one 21 common binding region shared by these toxins and an additional unique binding region for SEA and its related toxins (Chintagumpala et al., 1991). 1.5. MHC class II Binding Sites of PTSAgs A. PTSAgs Bind MHC Class H Molecule Using Common Structural Strategies The most direct information on the interaction of superantigens with M H C class II molecules has been provided by the elucidation of the three-dimensional crystal structures of the SEB.HLA-DR1 and TSST-1 :HLA-DR1 complex (Jardetzky et al., 1994; K i m et al., 1994). Together with superimposition of the crystal structures of PTSAgs, a similar binding mode involving the N-terminal domain of the toxins and the a-chain of the DR1 molecules was noted (Ulrich et al., 1995; Ulrich et al.,1998). The M H C class II binding sites on PTSAgs consists of principally two structurally conserved region which are located in domain B and are essential for the molecular recognition of the a chain of H L A -D R (Fig. 4) (Ulrich et al., 1995; Ulrich et al., 1998). The first region is the hydrophobic binding loop, centered around a leucine residue (SEA L48; SEB L45; TSST-1 L30), in most of the PTSAgs (Ulrich et al., 1998). The alignment of primary sequences of PTSAgs reveals a conservation of this leucine residue among PTSAgs except SPEC, suggesting that it may serve as an important contact for PTSAg: M H C class II binding (Acharya et al., 1994; Swaminathan et al., 1992; Schad et al., 1995; Ulrich et al., 1998). The crystal structure of SEB:DR1 reveals that Phe44, Leu45, and Phe47 point into a hydrophobic pocket of the DR1 molecule (Jardetzky et al., 1994). The importance of these residues in M H C class IT binding had been demonstrated by mutagenesis studies (Kappler et al., 1992). 22 Figure 4. Common structural strategies used by SEA (a), SEB (b), and TSST-1 (c) for binding to M H C class II a subunits. C represents carboxyl-terminus; N represents amino terminus; Z n 2 + in (a) represents bound to zinc ion. Adapted from Ulrich et al., 1998. 23 In TSST-1, the corresponding region makes similar contacts to DR1 as observed in SEB:DR1 complex (Kim et al., 1994). Additional evidence for the importance of this hydophobic loop between p strands 1 and 2 of TSST-1 (residues 27-32) in complex formation comes from mutational studies performed by our laboratory and others. Site-directed and random mutagenesis studies demonstrated that TSST-1 G31R mutant and TSST-1 G31S/S32P double mutants were unable to bind an HLA-DR1 molecule or stimulate human peripheral blood lymphocytes (Hurley et al., 1995; Kum et al., 1996). Although the SEA:DR1 crystal complex is not yet available, the modeled DR1-S E A complex also presents similar interaction in the analogous region (Ulrich et al., 1995). Site-directed mutagenesis indicated that changes of residue Phe 47 in S E A to either alanine or serine resulted in a six to seven fold decrease in M H C class II binding (Abrahmsen et al., 1995; Hudson et al., 1995). Moreover, alanine substitution of His 50 in S E A reduced M H C class n binding by threefold (Abrahmsen et al., 1995). In addition, Phe-47-Gly, Leu-48-Gly, Phe-47-Ala, and Phe-47-Ser mutation drastically reduced T cell proliferation (Harris et al., 1993; Harris and Betley, 1995, Abrahmsen et al., 1995; Hudson et a l , 1995). The SEA residues Phe 47, Leu 48, and His 50 are equivalent to residues Phe 44, Leu 45, and Phe 47 on SEB, which are all important in the interface between SEB and HLA-DR1 (Jardetzky et al., 1994). Furthermore, Thibodeau et al have recently shown that mutations in the a 1 domain of H L A - D R affected the binding of SEA (Thibodeau et al., 1997). The second structural binding motif is a polar pocket (SEA Y92, Y108, D70; SEB Y89, Y I 15, E67), derived from three P-strand elements located at the upper part of domain B , that accommodates Lys39 of D R l a chain (Ulrich et al., 1995; Ulrich et al., 1998). Site-directed mutagenesis studies showed that the substitution of alanine for SEB Y89 or SEA 24 Y92 eliminated the hydrogen bond with Lys39 of D R l a chain and resulted in 100 fold reduction in DR1 binding (Ulrich et al., 1995). In addition, the importance of aLys39 of D R l a chain in M H C class II binding had also been demonstrated by mutagenesis studies. By mutation of aLys39 to alanine, the binding of both TSST-1 and SEB are abolished (Jardetzky et al., 1994). Although Lys39 of DR1 a chain hydrogen bonds to TSST-1 residues Pro50 and Ser53 at the TSST-1:DR1 interface(Kim et al., 1994), TSST-1 is the only PTS Ag which lacks this particular motif, as represented by the positioning of residues T69,185, and 146 in Fig. 4c (Ulrich et al., 1998). The polar pocket combined with other surface elements, such as the disulfide loop, stabilize the PTSAg:HLA-DR complexes (Ulrich et al., 1998). Although the disulfide loops differ in structure between SEA and SEB, both SEA A97 and SEB Y94 located within the loop apparently bind in a similar manner to D R a chain (Ulrich et al., 1995). Replacement of these residues with alanine disrupted M H C class II binding (Ulrich et al., 1995). Additional contact region involving the a-helix of DR1 P-chain and TSST-1 residues 73-80 was identified in the TSST-1 :DR1 interface (Kim et al., 1994). These TSST-1 residues form part of a mobile loop to optimize TSST-1 contacts with both the D R i p chain (residue 55-63) and the peptide antigen (Papageorgiou et al., 1996) (Fig. 5b, 5d). Thus, TSST-1 binding is likely to be peptide-dependent (Hsu and Huber, 1995). In contrast, SEB interacts entirely with the external surface of the a 1 domain of HLA-DR1 and no contacts with the DR1 P chain and peptide was observed in the complex (Fig. 5a, 5c). Comparison of the TSST-1 :DR1 complex with the SEB:DR1 complex reveals that 11 of the 17 DR1 residues in the TSST-1 :DR1 interface are common to the SEB:DR1 25 Figure 5. Structures of complexes between HLA-DR1 and the superantigens SEB and TSST-1. (a) Crystal structure of SEB-DR1 complex; (b) Crystal structure of TSST-l-DR-1 complex; (c) Schematic structure of the SEB-DR1 complex, showing the exclusive binding of SEB to the a chain of DR1, primarily off one edge of the peptide binding groove; (d) Schematic structure of the TSST-1-DR1 structure, showing how TSST-1, in contrast to SEB, binds to both a and P chains of DR1 and contacts part of the bound peptide. Taken from Hsu et al., 1995. 26 interface, and 13 of the 19 residues on TSST-1 that contact DR1 molecule are homologous to residues on SEB that contact DR1, demonstrating that there is overlap between the M H C class II binding sites for TSST-1 and SEB (Table IV) (Kim et al., 1994). Despite the similarity and the apparent overlap of their respective MHC-binding sites, TSST-1 and SEB do not compete with each other for binding to H L A - D R (Scholl et al., 1989). It was proposed that the affinity of TSST-1 and SEB for M H C class II molecules might be influenced by the nature of the peptide contained in the conventional peptide-binding groove (Kim et al., 1994; Thibodeau et al., 1994). Hence, only a portion of M H C class II molecules carrying the appropriate peptides could bind the toxins and present them to T cells (Wen et al., 1996). Another possible explanation would be the existence of a second binding site on DR1 for SEB, or TSST-1. Although the SEB:DR1 complex reveal only one M H C class II binding site on SEB, several groups of investigators have independently implicated that residues outside the binding interface defined in the crystal complex are involved in class f l binding. Using random mutagenesis, Kappler and his associates had identified mutants at positions 14 and 17 of SEB that caused a 100-fold reduction in H L A -D R binding (Kappler et al., 1992). Moreover, peptides containing residues 1-33 and 9-20 were shown to compete with SEB for binding to H L A - D R , implicating that these regions are involved in M H C class n binding (Soos and Johnson, 1994; KoMLSar et al., 1994). These implicated residues are located on the outside edge and back of the large P-grasp domain, which somewhat overlaps with the zinc coordinated M H C class II binding site. However, zinc atom is absent in the SEB crystal structure (Swaminathan et al., 1992). 27 Table TV. The TSST-1 :DR1 interface. Underlined residue numbers are also found in the DR1: SEB interface. TSST-1 residue SEB residue Location DR1 contact residues D27 D42 PI 100 i I i i i SEB 1.09 >100 SEA > 35 > 100 a Toxin was coated at 0.1 ug. b Titre was designated as the lowest concentration that gave an A410 of 3 times background. c IgGl is the isotypic control. d Relative intensities of the signals in Western Blot are indicated: - , no reaction; +++++, very strong reaction. 64 ~ 1500-, 1 • •—I 1 1 1 1 1 1 1 1 1 1 1 1 1 1 o 10 ^ O-"IO10IO-9IO-8IO-7IO-6IO-5IO-4IO-3IO-2IO1 10° 10' 10 2 [SEB] ( nM) Figure 7. Dose response of SEB-induced TNFa secretion of human PBMC. 3X105 cells were incubated with increasing concentration of SEB for 18 h at 37°C. Supernatants were then harvested and the concentration of TNFa in the supernatant was measured by TNFa assay. Results represent the mean ± SEM for triplicate measurement in 4 different donors. 65 Figure 8. Effect of anti-TSST-1 MAb5 on SEB-induced TNFa secretion from Human PBMC. Results represent the mean + SEM for triplicate measurement in 3 donors. * represents statistically significant inhibition (p < 0.05; 1-way ANOVA). 66 Inhib i t ion of S E B - i n d u c e d Mitogenesis by M A b 5 . Dose response study of SEB-induced human PBMC mitogenesis was carried out and lpM SEB, within the linear range of the dose response curve, was used in the inhibition study (Fig. 9). Serial dilutions of MAb5 were tested for their ability to neutralize SEB-induced mitogenesis of human PBMC from 4 different donors (Fig. 10). MAb5 was able to inhibit SEB-induced mitogenesis significantly (P < 0.05) in a dose dependent manner. Again, MAb6 showed no effect. A n i m a l studies. The in vivo effect of MAb5 was examined for its ability to neutralize SEB-induced lethality in the D-galactosamine sensitized BALB/c mouse model of lethal shock (Table VTI). A predetermined dose of SEB LD90 was injected i.p. with or without MAbs one hour after D-galactosamine sensitization (Fig. 11). 15 animals per group were studied and death after 72 hours was the end point. At the end of the 72 hour period, 73% mortality (11 of 15) was observed in mice challenged with SEB. This was not significantly different from the mortality of 87% (13 of 15) in the presence of MAb5, or the mortality of 80% (12 of 15) in the presence of MAb6. Both MAb5 and MAb6 were unable to neutralize the lethal effect of SEB in D-galactosamine sensitized mice at the concentrations studied (P > 0.05, 1-tailed, Fisher's exact test). 67 [SEB] ( n M ) Figure 9. Dose response of SEB-induced mitogenesis of human PBMC. Results represent the mean ± SEM for triplicate measurement in 4 donors. 68 ~ 4000CH Q. O 30000-o e-o o 0) 1 200004 = 100004 I I F ^ I i—i rTfl o: s s m LU 2 5 CO CD in e oo + m m CO X X I J2 CM + m ui co m .a O CN + CO HI to o o CO + CD LU CO m .a + 00 LU CO m .a co + CO LU CO o o CO + CO UJ CO C O + CQ Ul CO CO + m UJ CO Figure 10. Effect of anti-TSST-1 MAb5 on SEB-induced mitogenesis of human PBMC. Results represent the mean ± SEM for triplicate measurements in 4 donors. * represents statistically significant inhibition (p < 0.05; 1-way ANOVA). 69 % Lethality %Survival 0 2 4 6 8 10 12 14 16 18 20 22 Dose of SEB (u,g/mouse) Figure 11. Dose response of SEB-induced lethality in D-galactosamine sensitized murine model. 70 Table VE. Neutralization studies of Mab5 on SEB-induced lethality in D-galactosamine sensitized murine model. Group Dead/Total Saline control 0/5 SEB alone 11/15 SEB + Mab5 13/15 SEB + Mab6 12/15 TSST-1 alone 4/5 TSST-1 +Mab5 0/5 71 DISCUSSION In order to evaluate whether shared antigenic determinants exist on TSST-1 and SEB, a murine anti-TSST-1 neutralizing monoclonal antibody, MAb5, was assessed for its ability to cross-react with a panel of staphylococcal superantigens. The ELISA results indicated that MAb5 was capable of cross-reacting with SEB, suggesting that TSST-1 and SEB share at least one cross-reactive epitope, which was recognized by our MAb5. However, it is likely that this cross-reactivity was of lower affinity since 35 fold more MAb5 was required to demonstrate reactivity to SEB by ELISA compared to TSST-1 in ELISA. It is interesting to note that this cross-reaction is not detectable in western blot. This might be expected if the recognized determinant is a discontinuous conformational epitope which consists of linearly distant residues brought together by folding of the polypeptide chain. The denaturing condition in SDS-PAGE might disrupt the specific conformation of the epitope required for recognition by MAb5, and thus no longer recognizable in western blot. However, solid phase ELISA (used in this study) has the same problem as western blot since coating of toxins on the ELISA plates might also alter the conformation of the molecules. It is well documented that solid phase ELISA preferentially detects MAbs that recognize sequential epitopes and may not be effective in detecting MAbs that recognize conformational epitope (Brennand et al., 1986; Friguet et al., 1984; Vaidya et al., 1985; Fieseretal, 1987). An alternative possibility is that the cross-reactive epitope is inaccessible to MAb5 in western blot, or that western blot is not sensitive enough to detect this low affinity cross-reaction. Even though the nature of the recognized SEB eptiope, whether continuous or conformational, is not distinguishable at 72 this point, our ELISA data strongly suggested that MAb5 recognizes cross-reactive epitope on SEB. Anti-TSST-1 MAb5 could neutralize the superantigenic activity, including mitogenicity and TNFa release, induced by both TSST-1 (Fig. 12 and Fig. 13) and SEB (Fig. 8 and Fig. 10). However, MAb5 was less effective at neutralizing SEB. Approximately, 1000 fold more MAb5 was required to significantly inhibit both SEB-induced mitogenesis and TNFa release than that needed for the neutralization of TSST-1. In addition, MAb5 can only partially neutralize SEB-induced activity. In contrast, MAb5 caused nearly complete neutralization of TSST-1. The ineffectiveness of MAb5 in completely neutralizing SEB-induced activity in vitro is expected and can be attributed to its lower affinity to SEB than to TSST-1 as reflected in the ELISA assay. However, the in vitro data strongly suggest that TSST-1 and SEB share cross-neutralizing epitopes. Previously, our laboratory has demonstrated the neutralizing effect of MAb5 on TSST-1-induced lethality (LD90) in D-galactosamine mice using a dose of TSST-1: MAb5 molar ratio of 1:5. In order to evaluate if MAb5 can also neutralize SEB-induced lethality in vivo, the same molar ratio of SEB (LD90) and MAb5 was used in this pilot study. It is not surprising to observe that MAb5 was not able to neutralize SEB-induced lethality in vivo in this study, since 1000 times more MAb5 was required to significantly cross-neutralize SEB-induced activity in vitro. Based on the in vitro data, it seems logical to predict that a much higher dose of MAb5 might be needed to demonstrate any cross-neutralizing effect in SEB-challenged mice, although the precise mechanism for SEB-induced toxicity and the protective mechanism of MAb5 in the murine model are not fully 73 300001 « 9r Figure 12. Effect of anti-TSST-1 MAb5 on TSST-1-induced mitogenesis of human PBMC. Results represent the mean ± S E M of triplicate measurement in 2 donors. * represents statistically significant inhibition ( p < 0.05; one-way ANOVA). 74 IOOO-I Figure 13. Effect of anti-TSST-1 MAb5 on TSST-l-induced TNFa secretion from human PBMC. Results represent the mean ± SEM of triplicate measurement in 2 donors. * represents statistically significant inhibition (p < 0.05; one-way ANOVA). 75 understood. Thus, there are insufficient data to conclude that MAb5 is not able to cross-neutralize SEB-induced lethality in vivo. A dose response experiment should be done to further evaluate the cross-neutralizing effect of MAb5 in vivo. The results obtained with our MAb studies showed that determinants on SEB and TSST-1 are located at or near a functionally important site on the toxins required for mitogenicity and cytokine production, two biological properties shared by all staphylococcal superantigens. Although it is generally known that TSST-1 exhibits minimal sequence homology with SEs, amino acid sequence alignment analysis reveal that the PTSAgs possess 4 uniformly conserved consensus sequences (Hoffmann et al., 1994). Fig. 14 illustrates the 4 regions with large degree of homology between the amino acid sequences of PTSAgs family. It is possible that these conserved regions or motifs adjacent to these regions are recognized by MAb5 and are important functional domains for mitogenicity, TNFa release and possibly other shared immunobiological properties of these PTSAgs. However, no conclusive evidence has been shown that all these conserved regions are responsible for superantigenic activities of all PTSAgs and the exact location of our recognized cross-reactive epitopes are not known. Alternatively, the motifs that direct all these common biological activities might be derived from the folding of PTSAgs, and these coriformational epitopes might be found in other parts of the molecules. Thus, definitive studies are necessary to determine the nature and location of these cross-reactive determinants. Realizing that PTSAgs elicit superantigenicity through binding to the major histocompatibility complex (MHC) class II molecules and interacting with the specific VJ3 region of the T cell receptor (TCR), it is presumed that MAb5 neutralized SEB-induced 76 Region 1 Region 3 Toxin Residue # Toxin Residue# SEA 79 K Y K G K K V D L Y G SEA 147 K K N V T V Q E L D L Q A R R Y L SEB 76 K Y K D K Y V D V F G SEB 152 K K K V T A Q E L D Y L T R H Y L SEO 76 K Y K D E V V D V Y G SEC1 151 K K S V T A Q E L D I K A R N F L SEC2 76 K Y K D E V V D V Y G SEC2 151 K K S V T A Q E L D I K A R N F L SEC3 76 K Y K D E V V D V Y G SEC3 151 K K S V T A Q E L D I K A R N F L SED 74 H F K S K N V D V Y P SED 142 K K N V T V Q E L D A Q A R R Y L SEE 76 K Y K G K K V D L Y G SEE 144 K K E V T V Q E L D L Q A R H Y L SPEA 70 L F K D K N V D I Y G SPEA 137 K K V T A Q E L D Y K V R K Y L SPEC 63 F K R D D H V D V F G SPEC 124 K D I V T F Q E I D F K I R K Y L TSST-1 56 F T K G E K V D L N T TSST-1 121 K K - - - - Q - L - I TSST-1 129 L D F E I R H Q L Reeion2 Retrion4 Toxin Residue # Toxin Residue# SEA 106 C Y G G V T L H D N N SEA 209 L L R I Y R D N K T I N S E SEB 113 C Y G G V T E H N G N SEB 213 Y L Y N D N K V D S K SEC1 110 C Y G G I T K H E G N SEC1 213 Y L Y N D N K T V D S K SEC2 110 C Y G G I T K H E G N SEC2 213 Y L Y N D N K T V D S K SEC3 110 C Y G G I T K H E G N SEC3 213 Y L Y N D N K T V D S K SED 101 C T Y G G V T P H E G N SED 204 Q L R I YS D N K T L S T E SEE 103 C Y G G V T L H D N N SEE 206 L L R I Y R D N K T I N S E SPEA 98 C I Y G G V T N H E G N SPEA 197 Y L I Y K D N E T L D S N SPEC 83 Y I Y G G I T P A Q N N SPEC 184 I F A K Y K D N R I IN K TSST-1 83 F Q I S G V T N T E K L TSST-1 179 P P IN I D E I K T I E A E Figure 14. Conserved regions of primary structure among sequences in the PTSAg family. Residue # refers to the primary sequence position of the first (N-terminal) conserved residue in each region. Taken from Hoffmann et al., 1994. 77 superantigenic activity by binding to the toxin and blocking these interactions (Hamad et al., 1994; Shimonkevitz et al., 1996). However, further studies are needed to deduce the precise mechanism involved in the cross-neutralizing activity exhibited by MAb5. The report published recently by Ulrich and his coworkers, while this manuscript was in preparation, demonstrated that antibodies induced by SEB mutant vaccines cross-reacted with heterologous PTSAgs, including TSST-1, SEA, and SEC1 (Ulrich et al., 1998). They further showed that this vaccine protected both mice and rhesus monkeys from lethal toxic shock when they were challenged with these toxins (Ulrich et al., 1998). Interestingly, these cross-reactive antibodies also had a low titre for TSST-1 and confer partial protection when the animals were challenged with TSST-1 (Ulrich et al., 1998). However, the precise locations of these cross-reactive antigenic epitopes were not known. Our observations in this study substantiate the fact that TSST-1 and SEB do share common cross-reactive epitopes that are important for superantigenic activities. 78 Chapter 4 Identification of Cross-Neutralizing Epitopes on Toxic Shock Syndrome Toxin-1 and Staphylococcal Enterotoxin B by Epitope Peptide Mapping and Synthetic Peptide Inhibition Studies. 79 \ I N T R O D U C T I O N Staphylococcal and streptococcal pyrogenic superantigens (PTSAgs) are exotoxins that are associated with a variety of immune-mediated diseases, including food poisoning, toxic shock syndrome (TSS), and other chronic immune disorders, due to their profound effects upon the immune systems (Kotzin et al., 1993; Schlievert et al., 1993). PTSAgs are characterized by their ability to bind major histocompatibility (MHC) class II receptors on antigen presenting cells (APCs) and T cell receptors (TCR) which express particular VP elements (Herman et al., 1989; Fields et al., 1996; Schlievert et al., 1997). These interactions lead to massive proliferation of T cells and excessive release of proinflammatory cytokines (Hackett wt al., 1993; Grossman et al., 1992). The latter superantigenic properties are thought to be responsible for the pathogenesis of TSS (Parsonnet, 1989). One of the ways to control PTSAgs-associated disease is vaccination. Formaldehyde inactivated toxoid vaccine had been developed in an attempt to prevent TSS. However, the vaccination approach to controlling SAg-associated disease is complicated by the diversity of PTSAgs produced by clinical isolates. PTSAgs can be divided into sequence homology groups consisting of (1) SEA, SED, and SEE; (2) SEB, SEC1, SEC2, SEC3, streptococcal superantigen (SSA), SPEA, and SPEC; (3) TSST-1 (Ulrich et al., 1995; Ulrich et al., 1998). It would not be practical to develop specific vaccines for each of these PTSAgs. Despite the considerable sequence divergence among these PTSAgs, they exhibit strikingly similar protein folds and structural architectures (Swaminathan et al., 1992; Schad et al., 1995, Sundstrom et al., 1996, Acharya et al., 1994; 80 Papageorgiou et al., 1995). Recently, Ulrich et al. demonstrated that cross-reactive antibodies against most PTSAgs could be induced by immunizing mouse with just one mutant PTSAg vaccine and thereby suggested the presence of shared neutralizing epitopes among these toxins (Ulrich et al., 1998). Yet, the precise locations of these antigenic determinants are not known. This information, however, is essential for developing a common, cross-reactive vaccine. Therefore, before an effective vaccine can be developed, it is important to have more information about critical recognition sites that can induce cross-reactive antibodies for neutralization against the whole panel of PTSAgs. To this end, much effort had been focused on the mapping of shared antigenic determinants of the SEs within the sequence homology groups, especially SEB and SEC1 (Spero and Morlock, 1979; Thompson et al., 1984, Turner et al., 1992). Little is known about shared antigenic determinant across different sequence homology groups. Previously, our laboratory had developed an anti-TSST-1 monoclonal antibody, MAb5, which was able to neutralize TSST-1, as well as SEB-induced superantigenic activities in vitro. In the current study, by mapping the neutralizing domains on TSST-1 and SEB recognized by MAb5, we were able to describe two topographically different epitopes in TSST-1 and SEB in which functional activities were located. 81 RESULTS Determination of a TSST-1 Epitope recognized by MAb5. In order to precisely identify the epitope on TSST-1 defined by MAb5, overlapping decapeptides spanning a small portion of TSST-1 sequence (T46-70) were made since this region has been previously characterized as containing epitopes) binding to MAb5 (Kum et al., unpublished data). The amino acid sequence of TSST-1 had been deduced from the nucleotide sequence of TSST-1 (Lee et al., 1992). The results obtained for the synthesized decapeptides when tested by ELISA with two different antibodies, anti-TSST-1 MAb5 and isotypic control IgGl, were shown in Fig. 15a and Fig. 15b. Using limited dilution of MAb5 (1 |ig/ml), scan (a) in Fig. 15 showed that MAb5 reacted strongly with one single epitope, 4 7FPSPYYSPAF 5 6. The isotypic control, IgGl, did not react with any of the peptides as shown in Fig. 15b, thereby demonstrating the specificity of the reaction with MAb5. In order to identify the minimum TSST-1 antigenic sequence in contact with MAbS, MAb5 was reacted with multiple sets of overlapping peptides, which differ only in length, spanning the TSST-1 sequence 4 7FPSPYYSPAF 5 6. This multiple length scanning or "window" approach, analogous to viewing the protein through multiple windows of different width, identifies the location of linear epitopes and define their boundaries or ends of minimal antibody-binding sequence (Geysen et al., 1987). It can be seen in Fig. 16. that none of the tri- or tetrapeptides in this region reacted with MAb5, but one pentapeptide, 5 2 YSPAF 5 6 did. Window/window+1 analysis applied to the penta- and 82 (a) 3-i (b) 2.5-. ri d w 2.0-Figure 15. Scan of TSST-1 (46-70) decapeptides with (a) anti-TSST-1 MAb5; (b) isotypic control IgGl. 83 Figure 16. Window/ window + 1 analysis on TSST-1 epitope T(47-56). 84 hexapeptide set demonstrates that the omission of the N-terminal Y51 from 5 1 YYSPAF 5 6 or C-terminal T57 of 5 2 YSPAFT 5 7 did not reduce the reactivity of pentapeptide 3 2 YSPAF 5 6 , suggesting that Y51 and T57 were not directly involved in antibody binding. However, analysis of the tetra- and pentapeptide set demonstrates that further omission of N-terminal Y52 or C-terminal F56 abolished the reactivity of tetrapeptide 5 3 SPAF 5 6 and 5 2 YSPA 5 5 . Taken altogether, these results suggest the presence of a single sequential epitope comprising the five residues 5 2 YSPAF 5 6 , represents the minimum epitope needed for recognition by MAb5. Determination of a SEB epitope recognized by M A b S . To reveal MAb5's cross-reactive epitope on SEB, the initial attempt was to synthesize a series of overlapping decapeptides corresponding to SEB (residues 62-118), since secondary structural alignment of TSST-1 and SEB revealed that TSST-1 (41-75) aligned with SEB (63-89) (Acharya et al., 1994). Using a limited concentration of MAbS, ELISA testing indicated that MAb5 also recognized one single peptide, 8 3 DVFGANYYYQ 9 2 with a good signal to background ratio (Fig. 17a). Again, the isotypic control IgGl did not react with any of the peptides, indicating the specificity of the reaction with MAb5 (Fig. 17b). In order to further characterize this SEB epitope, ^DWGANYYYQ 9 2 , peptides corresponding to the complete set of all the possible linear octa-, hepta-, hexa-, and pentapeptide homologous with the native sequence were synthesized. Significant binding of MAb5 to pentapeptides 8 4 V F G A N 8 8 and 8 5 FGANY 8 9 was still observed when the peptide target was sequentially reduced from octapeptides to pentapeptides, suggesting that the sequential epitope defined by MAb5 on SEB comprises fewer than 5 residues, likely within the tetrapeptide 8 5 F G A N 8 8 (Fig. 18). 85 (a) (b) Figure 17. Scans of SEB (1-239) decapeptides with (a) anti-TSST-1 MAb5; (b) isotypic control, IgGl. 86 5 mer 3.0-, 7 mer 3.0-. 2.5-2.0-Q 6 1-5-6 mer 3.0-, 2.5-2.0-a O 1.5-8 mer 3.0-n 2.5-2.0-Q O 15-B83-90 B84-91 B85-92 Figure 18. Window/window + 1 analysis on SEB epitope (83-92) 87 Peptide inhibition studies. In an attempt to relate the mapped epitopes with function, synthetic peptide corresponding to SEB (residues 83-92), ^DWGAhmTYTQ92, was tested for its ability to inhibit SEB-induced superantigenic activities. First of all, peptides were tested for stimulation of proliferation and TNFa production. Neither SEB peptide nor the scrambled control peptide showed any increase over control values. Next, these peptides were tested for their ability to inhibit SEB-induced superantigenic activities. SEB(83-92) significantly ( p < 0.05) inhibited both SEB-induced mitogenesis and TNFa production in a dose dependent manner while the scrambled peptide had no effect (Fig. 19 and Fig. 20). SEB(83-92) was effective at 1.5 mM for inhibition of both SEB-induced mitogenesis and TNFa production. 88 a o o a. L. o U ii a 40000-30000H S u 20000-S H 10000-I t^nini-nriHf^nr^i O M « (S (S g os os o\ as O i i t • ci m co 00 00 00 00 " CQ CQ C3 in m