Structural and functional role of threonine 112 in a superantigen Staphylococcus aureus enterotoxin B.

Bacterial superantigens are potent T-cell stimulatory protein molecules produced by Staphylococcus aureus and Streptococcus pyogenes. Their superantigenic activity can be attributed to their ability to cross-link major histocompatibility complex class II molecules with T-cell receptors (TCRs) to form a tri-molecular complex. Each superantigen is known to interact with a specific V(beta) element of TCR. Staphylococcal enterotoxin B (SEB, a superantigen), a primary cause of food poisoning, is also responsible for a significant percentage of non-menstrual associated toxic shock syndrome in patients with a variety of staphylococcal infections. Structural studies have elucidated a binding cavity on the toxin molecule essential for TCR binding. To understand the crucial residues involved in binding, mutagenesis analysis was performed. Our analysis suggest that mutation of a conserved residue Thr(112) to Ser (T112S) in the binding cavity induces a selective reduction in the affinity for binding one TCR V(beta) family and can be attributed to the structural differences in the native and mutant toxins. We present a detailed comparison of the mutant structure determined at 2.0 A with the previously reported native SEB and SEB-TCR V(beta) complex structures.

Bacterial superantigens are potent T-cell stimulatory protein molecules produced by Staphylococcus aureus and Streptococcus pyogenes. Their superantigenic activity can be attributed to their ability to cross-link major histocompatibility complex class II molecules with Tcell receptors (TCRs) to form a tri-molecular complex. Each superantigen is known to interact with a specific V ␤ element of TCR. Staphylococcal enterotoxin B (SEB, a superantigen), a primary cause of food poisoning, is also responsible for a significant percentage of nonmenstrual associated toxic shock syndrome in patients with a variety of staphylococcal infections. Structural studies have elucidated a binding cavity on the toxin molecule essential for TCR binding. To understand the crucial residues involved in binding, mutagenesis analysis was performed. Our analysis suggest that mutation of a conserved residue Thr 112 to Ser (T112S) in the binding cavity induces a selective reduction in the affinity for binding one TCR V ␤ family and can be attributed to the structural differences in the native and mutant toxins. We present a detailed comparison of the mutant structure determined at 2.0 Å with the previously reported native SEB and SEB-TCR V ␤ complex structures.
Staphylococcus aureus is one of the most prevalent causes of foodborne illness throughout the world (1,2). Illness occurs following ingestion of staphylococcal enterotoxins (SEs) 1 produced by S. aureus growing in contaminated food. Manifestations, which appear 4 -6 h after ingestion of contaminated food, include nausea, vomiting, diarrhea, abdominal cramps, and headache. Their production during systemic S. aureus infections has been often associated with fatal toxic shock syndrome (3,4). The relationship between the enterotoxic and superantigenic properties of SEs is not clearly understood, although there is some evidence suggesting that these biological responses may be mediated by distinct mechanisms involving different sites on the toxin (5). Based on antibody specificities, the SEs have been divided into eight distinct serological types, and S. aureus strains can express either none, one, or a combination of SEs. The serotypes defined to date are designated SEA, SEB, SEC1, SEC2, SEC3, SED, SEE, SEG, SEH, SEI, and SEJ (1, 2, 6 -8). Within the family, a high level of amino acid sequence homology is found. In addition to their enterotoxic properties, SEs have superantigenic properties (9).
The immunomodulatory effects of SEs, such as immunosuppression, enhancement of endotoxic shock, and induction of cytokine release, can be attributed to the fact that SEs are potent T-cell mitogens. The ability of SEs to induce proliferation of T-cells is somewhat similar to conventional antigen presentation by MHC class II molecules to T-cell receptors (TCRs) and hence known as microbial superantigens. However, unlike conventional antigens, the T-cell mitogenicity of SEs does not require antigen processing and lacks the normal specificity of the TCR for specific epitopes in response to conventional antigens. They are selective in that they do not induce the proliferation of all T-cells but selectively stimulate T-cells expressing certain V ␤ families of TCRs, and the T-cell response to SEs is not self-MHC class II-restricted (9 -11). Thus superantigens cross-link MHC class II molecules and the TCR via sites that are close to but distinct from those involved in conventional antigen presentation. This bypasses the normal TCR specificity for conventional T-cell epitopes, resulting in the stimulation of a substantial proportion of the total T-cell population and overproduction of cytokines.
A considerable amount of structural data is now available describing the molecular architecture for members of the SE family and has identified the MHC class II and TCR recognition sites on the toxin molecules (for recent reviews see Refs. [12][13][14]. Among the SEs, SEB is the most potent toxin, and our study has focused on the role of critical amino acid residues at the TCR-binding cavity (see Refs. [15][16][17]. It has been suggested that the residues that surround the T-cell receptor-binding cavity on the surface of SEBs are most important to TCR binding; Asn 23 , Tyr 175 , and Asn 179 have been shown to play a key role in TCR binding (18).
Based on the native structure of the toxin, we have constructed several point mutants and have studied the effect of each mutation on TCR specificity. One such mutation, Thr 112 3 Ser (T112S), a conserved residue in all SEs and located in the TCR binding cavity, induced altered T-cell specificity, and we describe the differences in molecular interactions through the elucidation of the crystal structure of the mutant at 2.0-Å resolution. We also address the molecular basis of the interaction of SEB with T-cells bearing different TCR V ␤ chains.

EXPERIMENTAL PROCEDURES
Chemicals and Enzymes-Unless otherwise stated, all chemicals and enzymes were purchased from Sigma or Roche Molecular Biochemicals.
Site-directed Mutagenesis by "Overlap" PCR-The wild type seb gene was amplified using PCR with S. aureus strain S6 template DNA and the oligonucleotide primers 5Ј-GGGAATTCATGTATAAGAGATTATT-TATTTC-3Ј and 5Ј-GGGAAGCTTGGCAACAAGGGGTTAATGCTA-3Ј. For SEBT112S, two DNA fragments that flanked the site of the mutation were first amplified using PCR with S. aureus strain S6 template DNA. The amplified DNA fragments were purified and used in a subsequent round of PCR to generate a DNA fragment that encoded the mutated seb gene. The amplified wild type and mutated seb genes were treated with the Klenow fragment of DNA polymerase I and ligated with SmaI-digested pBluescript. The cloned DNA was sequenced to confirm the authenticity of the genes. For preliminary studies the cloned DNA was excised with EcoRI and HindIII and cloned into similarly digested plasmid pKK223-3 (Amersham Biosciences). The recombinant plasmids were used to transform Escherichia coli strain XL-1 Blue (Stratagene, The Netherlands), and the recombinant bacteria were selected on L-agar containing 100 g/ml ampicillin (19). For production of purified protein for T-cell mitogenicity studies and for crystallization, each gene was subcloned into plasmid pMalc2 (New England Biolabs, Hitchin, UK) to generate a gene encoding a maltosebinding protein-SEB fusion protein. These recombinant plasmids were used to transform E. coli strain XL1 Blue, which was cultured as described above.
Expression and Purification of SEB and T112S Mutant-For initial studies, E. coli XL-1 Blue cells containing the appropriate recombinant pKK223-3 plasmid were cultured in 10 liters of L-broth supplemented with 100 g/ml ampicillin and 25 g/ml tetracycline at 30°C. When the A 600 nm of the culture reached 0.4 units, isopropyl-␤-D-thiogalactopyranoside was added to a final concentration of 0.5 mM. After a further incubation of 4 h, the cells were collected by centrifugation (10,000 ϫ g, 10 min, 4°C) and resuspended in 30 mM Tris-HCl buffer (pH 8.0) containing 1 mg/ml lysozyme, 1 mM EDTA, and 20% (w/v) sucrose. This was incubated for 30 min at 0°C and subsequently centrifuged (10,000 ϫ g, 10 min, 4°C); the supernatant was removed, and (NH 4 ) 2 SO 4 was added to 80% (561 g/liter) saturation. Precipitated material was collected by centrifugation (20,000 ϫ g, 10 min, 4°C) and dissolved in 20 ml of 5 mM sodium phosphate buffer (pH 7.0). After dialysis against 3ϫ 1 liter of the same buffer, 2.5 ml of pH 3-10 ampholytes (Bio-Rad) were added, and the proteins were separated by preparative isoelectric focusing using a Rotophor system (Bio-Rad). Fractions containing SEB were isolated and dialyzed against 5 mM sodium phosphate buffer (pH 7.0).
For larger scale production of SEB and T112S, the recombinant plasmids were transferred to E. coli X1776. These recombinant bacteria were each cultured in tryptone soya broth (Oxoid, Basingstoke, UK) supplemented with the following to the final concentrations indicated: 20 mM Tris-HCl (pH 7.5), 5 mM magnesium chloride, 0.5% (w/v) glucose, 0.01% (w/v) diaminopimelic acid, 0.005% (w/v) thymidine, and 50 g/ml ampicillin. Cultures (400 ml) were incubated with shaking (180 rpm) at 37°C. When the A 600 nm of the culture reached 0.7 units, isopropyl-␤-D-thiogalactopyranoside was added to a final concentration of 0.5 mM. Cultures were incubated for a further 3.5 h, and cells were harvested by centrifugation at 12,000 ϫ g at 4°C for 10 min. The cells were resuspended in 30 ml of column buffer as follows: 20 mM Tris-HCl (pH 7.4), containing 200 mM NaCl, 1 mM EDTA, 0.06 mM phenylmethylsulfonyl fluoride, and 1 mM dithiothreitol. Lysozyme (5 ml of 10 mg/ml solution) was added, and the suspension was incubated at room temperature for 30 min. Aliquots were sonicated on ice 3 times with 20-s intervals. The crude cell extract was centrifuged at 12,000 ϫ g for 30 min at 4°C, and the supernatant was decanted and filtered using a 0.22-m disposable filter. The supernatant was applied to a 40-ml amylose resin column (New England Biolabs), which had previously been equilibrated with column buffer. The column was washed with 400 ml of column buffer. Elution buffer (column buffer containing 10 mM maltose) (40 ml) was added to the drained resin, and the column was rolled for 1 h at room temperature. The eluate was collected as 1-ml fractions, which were analyzed at 280 nm and by SDS-PAGE. Fractions containing an optimum concentration of protein were pooled and cleaved with 1.6 ml (1600 units) of factor Xa (Amersham Biosciences) in column buffer containing 1 mM CaCl 2 and 0.05% (w/v) SDS. Cleavage was allowed to continue overnight at room temperature, and the cleaved product was dialyzed against 25 mM Tris-HCl (pH 8.0) containing 20 mM NaCl overnight at 4°C. The sample was filtered using a 0.22-m filter, before loading onto a 16/10 Mono Q anion exchange column (Amersham Biosciences). Cleaved SEB did not bind to the column, and was collected as fractions, whereas the maltose-binding protein and factor Xa bound to the column. The fractions of SEB were pooled and concentrated by stirred cell ultrafiltration (Millipore UK Ltd., Watford, Hertfordshire, UK).
The purified proteins were quantified using a Coomassie Blue dye binding assay (Bio-Rad) or using a BCA assay (Perbio Science UK Limited, Cheshire, UK) and also quantified by enzyme-linked immunosorbent assay using anti-SEB monoclonal antibodies. The SEB preparations were analyzed by SDS-PAGE (10 -15% gradient gels; Amersham Biosciences) followed by Coomassie Blue staining. The gels were scanned using a Bio-Rad Gel Document 1000.
Binding to MHC Class II Molecules-A flow cytometric assay was employed to detect binding of SEB and T112S mutant to Raji (a human lymphoblastoid cell line that expresses MHC class II) and RJ2.2.5 (an MHC class II-deficient mutant). Briefly, cells were suspended in Hanks' balanced salt solution with HEPES (3 mM), 2% (v/v) fetal bovine serum, and 0.1% (w/v) sodium azide. 5 ϫ 10 5 cells were incubated for 1 h at 37°C with 10 Ϫ7 M SEB or T112S mutant, and surface-bound toxins were detected by immunofluorescence following labeling with polyclonal rabbit anti-SEB (diluted 1:100) for 30 min at room temperature, followed by FITC sheep anti-rabbit IgG (Southern Biotechnology, Birmingham, AL) for 30 min at room temperature. Cells were washed between steps. 10,000 cells were acquired on a Becton Dickinson FAC-Scan, and data were analyzed using Becton Dickinson Lysis II software. Results were plotted as histograms of green fluorescence (Fl-1), and MHC class II binding was detected as a shift to the right of mean fluorescence intensity on Raji cells in the presence of bound toxins, compared with the MHC class II-negative mutant. Note that a small proportion of RJ2.2.5 cells bound the sheep anti-rabbit Ig nonspecifically in the presence or absence of toxins, which does not otherwise obscure the result.
V ␤ Typing-Direct immunofluorescence was used to determine the TCR V ␤ families expressed by human T-cells activated by SEB or the mutant by flow cytometry. Mononuclear cells were separated from the peripheral blood of 10 volunteers by one-step density centrifugation. Cells were cultured for 3 days with SEB (5 ϫ 10 Ϫ7 M), the mutant T112S (5 ϫ 10 Ϫ7 M), or the polyclonal mitogen phytohemagglutinin (PHA), 10 g/ml. T-cell blasts were prepared for flow cytometry in phenol red-free Hanks' balanced salt solution supplemented with 5 mM HEPES buffer, 2% fetal bovine serum, and 1% sodium azide. Cells (5 ϫ 10 5 ) were incubated for 30 min at 5°C in the dark with direct FITC-conjugated mAb specific for families of human T-cell TCR V ␤ (a kit comprising antibodies to V␤ 3.1, 5.1, 5.2/5.3, 5.3, 6.7, 8, 12, and 13 and V␣ 2, 12.1, Serotec, Oxford, UK). Cells were also incubated with a direct FITC-conjugated mAb specific for CD3 (PharMingen, Oxford, UK) to determine the proportion of T-cells after stimulation. Direct FITC-conjugated isotype control, mAbs were also used to determine the level of background labeling of T-cells after stimulation, which was subtracted from the test data. After washing, data from 10,000 cells were acquired for each sample and analyzed using a FACScan flow cytometer with CellQuest software (Becton Dickinson, Oxford, UK). The results are expressed as the percent of T-cells expressing a particular TCR V ␤ chain.
Protein Crystallization-Crystals of T112S mutant were grown at 16°C using the hanging drop vapor diffusion method. Samples (2 l) of the reservoir solution (0.8 ml) containing 30% (w/v) PEG 4,000, 50 mM sodium citrate buffer (pH 4.6), and 0.1 M ammonium acetate were mixed with an equal volume of the protein stock solution on siliconized coverslips. The crystals belong to orthorhombic space group P2 1 2 1 2 1 with two molecules in the asymmetric unit. The crystallization condition and packing of molecules in the mutant structure are different from those previously reported for the native SEB toxin (16,20).
X-ray Data Collection-Diffraction data were collected at 100 K using crystallization reservoir solution as cryoprotectant. The first data set was collected to 2.0 Å at the synchrotron radiation source (Electra), Trieste, Italy, using an MAR 345 image plate. Seventy nine images were collected with an oscillation range of 1.5°per image. A second data set (55 images) was collected at Daresbury, UK, station PX 9.5 to 1.8 Å with an oscillation range of 1.5°using an MAR CCD detector. Data processing, scaling, and merging of the two data sets were carried out using the HKL suite (21). However, data between 2.0 and 1.8-Å reso-lution shell were rather weak, and hence only up to 2.0 Å was used. The final R merge was 11% with an overall completeness of 92.7%. The data processing statistics are given in Table I.
Structure Determination and Refinement-The structure of T112S mutant was determined by molecular replacement with the program AMoRe (22) using the 1.5-Å native SEB (monomer) structure (16) as a starting model. Clear solutions were found for the two molecules in the asymmetric unit. This resultant structure was subjected to rigid body refinement followed by the calculation of an average electron density map. The structure was refined by simulated annealing using the maximum likelihood target as implemented in the program CNS (23) with non-crystallographic symmetry restraints. The progress of refinement was followed by monitoring both R free and R cryst (23) values. The refinement was continued using cycles of simulated annealing and B factor refinement. Electron density maps (F o Ϫ F c and 2F o Ϫ F c ) were calculated after each cycle of refinement and visualized using the program O (24). Water molecules were added to the model manually based on the density seen in the electron density maps and also by using the water pick protocol with the program CNS. In the final rounds of refinement the non-crystallographic restraints were gradually released, and the two molecules were checked and refined individually. The final model has a crystallographic R factor (R cryst ) of 22.7% for all data from 40.0 to 2.0-Å resolution and an R free of 24.6% for 5% of the data omitted (Table I).

RESULTS AND DISCUSSION
Purification of SEB Wild type and T112S Mutant-The TCRbinding site on SEB has been identified previously from the three-dimensional structure of the complex as a shallow cavity on the surface of the enterotoxin (17,25). Mutational analysis has confirmed the role in T-cell binding of many of the residues forming this cavity (17,26). Thr 112 is one of the TCR-binding site "rim" residues proposed by Swaminathan et al. (20), and this residue is conserved in SEA, SEB, SEC1, SEC2, SEC3, SED, and SEE (25). However, the role of this residue has not yet been determined experimentally. Hence we have investigated the properties of a T112S mutant. For the production of recombinant wild type SEB or T112S mutant, the encoding genes were expressed as maltose-binding protein fusions in E.
coli. The proteins were purified using affinity chromatography and cleaved with factor Xa to release SEB or T112S mutant. Based on scanning densitometry of SDS-PAGE gels, the purified proteins were judged to be at least 95% pure, and the yield obtained was ϳ8 mg of wild type SEB per liter of culture and 10 mg of T112S mutant per liter of culture.
Binding to MHC Class II Molecule-The TCR and MHC class II-binding sites are adjacent on the surface of SEB (17,27). However, residue Thr 112 is located on the region of the TCRbinding cleft which is distal to the MHC class II-binding site. This suggested that unless gross conformational changes had occurred in the T112S mutant protein, binding to MHC class II would be unaffected by this mutation. The ability of SEB and T112S mutant to bind the Raji human lymphoblastoid cell line in a flow cytometric assay was used as a measure of relative affinity of binding to MHC class II, as used previously (28). SEB and T112S mutant bound to Raji cells with equal fluorescence intensity (Fig. 1, b and c). The assay was considered to be specific as the MHC class II-negative mutant RJ2.2.5 failed to bind SEB or T112S mutant (Fig. 1, e and f), and the rabbit anti-SEB and FITC sheep anti-rabbit Ig failed to bind either cell type in the absence of SEB or T112S (Fig. 1, a and c). The results confirm that the T112S mutation has no effect on MHC class II binding and suggest that the structure of the MHC class II binding region was unaffected by this amino acid substitution. Other workers (26) have also shown that amino acid substitutions at other positions within the TCR-binding site of SEB do not influence the ability to bind MHC class II. Like these workers, we are therefore able to interpret changes in T-cell activation by the T112S mutant protein as indicative of changes in the TCR-binding site rather than changes in the ability to bind MHC class II.
T-cell V ␤ Distribution on T-cells Activated by SEB and T112S Mutant-Staphylococcal enterotoxins are known to activate human T-cells bearing a range of TCR V ␤ chains, but the precise spectrum of activity differs between the individual enterotoxins. To investigate whether the Thr 112 -Ser substitution had modulated the ability of the protein to activate T-cells, preliminary studies were undertaken in which human peripheral blood T-cells were stimulated with SEB or T112S, and V␤-specific mRNA levels were measured using semi-quantitative PCR. These experiments revealed a reduced expression of TCR V␤3 mRNA by T-cells activated by T112S compared with those activated by native SEB. 2 To confirm this preliminary result, we undertook a flow cytometric analysis by measuring TCR V ␤ family expression by T-cells following activation with SEB, T112S mutant, or PHA.
Peripheral blood mononuclear cells isolated from 10 human volunteers were cultured for 3 days in the presence of SEB, T112S mutant, or PHA, and the expansion of T-cells expressing different TCR V ␤ was determined by flow cytometry. SEB is known to activate selectively human V ␤ 3, V ␤ 12, V ␤ 14, V ␤ 15, V ␤ 17, and V ␤ 20 T-cells, whereas PHA activates all V ␤ families in the proportions present in peripheral blood (29). The results indicated that both SEB and T112S mutant selectively induced T-cells bearing V ␤ 3.1 and 12 compared with PHA as detected by the range of antibodies used (Fig. 2). As expected, SEB did not induce T-cells bearing V ␤ 5, 6.7, 8, or 13. The T112S mutant also failed to activate these T-cells (Fig. 2). SEB and the T112S mutant appeared to activate T-cells bearing V ␤ 12 to an equal degree (Fig. 2). However, in all of the individuals tested, the proportion of V ␤ 3.1 T-cells induced by T112S mutant was reduced compared with SEB. Indeed, the proportion of V ␤ 3.1 T-cells activated by T112S mutant was in the range 24 -86% below that for SEB, suggesting a substantial reduction in the 2 J. D. Hayball, personal communication.

FIG. 2. TCR V ␤ families expressed by human toxin-stimulated peripheral blood T-cells. Peripheral bood T-cells from 10 volunteers
were activated with SEB, T112S mutant, or PHA and labeled with FITC-conjugated anti-V ␤ -specific antibodies. Results from unstimulated peripheral blood T-cells from one volunteer is also shown (peripheral blood lymphocytes).  affinity of T112S mutant for the V ␤ 3.1 family of TCR compared with native SEB. As expected, PHA, SEB, and T112S mutant induced the same proportion of V␣ 2 and V␣ 12.1 T-cells, as V␣ TCR families are not known to be selectively expanded by superantigens. These findings indicated that the affinity of the enterotoxin toward the V ␤ 3.1 TCR had been reduced by the introduction of the Thr 112 to Ser substitution, without altering the response of V ␤ 12 T-cells.
Structure of T112S Mutant-The structure of T112S mutant was determined at 2.0-Å resolution (Fig. 3). The final model contained 3802 non-hydrogen protein atoms and 173 water molecules (Table I). The root mean square deviation between monomer pairs (C ␣ atoms) for the mutant is 0.41 Å, and between the mutant and native SEB structure (C ␣ atoms) is 0.51 Å. The regions that deviate most between the native and mutant structures include the disulfide loop, the first 20 residues at the N terminus, and residues 41-62 (Fig. 4). Part of the disulfide loop and the N-terminal residues are better ordered in the mutant structure, which could be due to the fact that Thr 112 to Ser data were collected under cryogenic conditions, whereas the native structure was at room temperature. Exclusion of these regions improves the root mean square deviation between the two structures to 0.3 Å for Thr 112 to Ser to native, and 0.4 Å for Thr 112 to Ser monomer pairs. The Ramachandran plot for both molecules shows 87.1% of the residues in the most favorable regions and no residues in the disallowed regions. Residues 98 -107 have not been modeled due to poor density, and residue 238 (in both molecules) was modeled as alanine due to insufficient density. Molecule 1 will be used below.
Structural Alterations at the TCR-binding Site in the Mutant-Analysis of the structure of the T112S mutant reveals that many of the gross features of the TCR-binding site are preserved as in the native SEB structure (16). Based on crystallographic studies, it has been established that the main contact interface in the SEB-TCR V ␤ complex (17) involves SEB side-chain atoms and TCR V ␤ backbone atoms, mainly CDR1, CDR2, and HV4 regions ( Fig. 4 and Table II). In addition, mutagenesis and structural studies have identified regions corresponding to residues 20 -33 (␣2-helix), 55-61 (␤2-␤3 loop), 87-91 (␤4), 112 (␤5), 177 (␣4 -␤9 loop), and 210 -214 (␣5) as being crucial for TCR binding in SEB (15)(16)(17). In particular, SEB residues Asp 23 (located in the ␣2-helix), Val 26 (solventexposed and situated in the middle of the ␣2-helix), Asn 31 (solvent-accessible), and Val 33 (buried) are implicated in TCR binding. The ␤2-␤3 loop (residues 55-61) is flexible, and the positioning of this loop may affect binding to the TCR and the solvent accessibility of residues such as Asn 88 , Tyr 89 , and Thr 112 (Fig. 5). Thr 112 is buried, and as such, any effects on TCR recognition are likely to be indirect, for example by perturbation of the local environment such as its hydrogen bond to Tyr 89 that is implicated in TCR binding (16). The residue Phe 177 is located within the ␣4 -␤9 loop and makes several contacts with the TCR V ␤ backbone in the SEB complex (17). The additional region of interest in the site is the ␣5-helix and, in particular, residue Gln 210 . This SEB residue is involved in direct interaction with the TCR molecule through hydrogen bond formation with the V ␤ main chain (17).
The Mutation Site-In the native SEB structure, Thr 112 forms a hydrogen bond with Tyr 89 and makes van der Waals contact with Tyr 61 . The mutation of Thr 112 to Ser causes the loss of the hydrogen bond with Tyr 89 and van der Waals interaction with Tyr 61 . As a result, the flexible loop region 52-64 displays significant change in conformation, moving up to 20 Å away, and hence the Thr 112 -Tyr 61 interaction is no longer possible (ϳ12 Å, Fig. 5 and Table II). The movement of Tyr 61 has implications for T-cell recognition (16,30). Asn 60 is also located in this loop and forms part of the TCR-binding site (Fig.  5, top and middle). In the mutant structure, Asn 60 also moves some 14 Å from its original position. Leder et al. (31) measured the energetic contribution of individual residues to the TCRbinding site and found that Asn 60 makes a modest, yet significant contribution to the binding of SEB to TCR. It was also Comparison of the contacts between the TCR V ␤ chain and SEB. a refers to the first SEB-TCR V ␤ complex molecule in the asymmetric unit. b n refers to the second.
found in binding studies that mutation of Asn 60 resulted in less efficient stimulation of T-cells. Superposition of the T112S mutant structure with that of the SEB-TCR V ␤ complex (17) reveals that neither Asn 60 nor Tyr 61 can make their usual interactions with the TCR V ␤ chain (Fig. 5, middle). molecule (3.04 Å), forms a hydrogen bond with Asn 63 (2.86 Å), and interacts with the side chain of Val 64 . These interactions appear to stabilize the local environment of the mutation site in the absence of the ␤2-␤3 loop ( Fig. 5 and Table II).
The SEB-MHC class II-binding site as defined by Jardetzky et al. (27) appears to be intact in the T112S mutant structure. This site is formed by residues Phe 44 to Phe 47 , Glu 67 , Tyr 89 to Ser 96 , Tyr 115 , Asp 209 , and Ser 211 . The loss of the hydrogen bond between Thr 112 and Tyr 89 in the mutant structure seems to have no significant effect on the architecture of the MHC class II-binding site, because superposition of native SEB and the mutant structures reveals all the residues to be in equivalent positions (Fig. 4).
The SEB Tyr 61 deletion mutant (SEB⌬61Y) described by Hayball and Lake (30) is able to bind MHC class II molecules as wild type but behaves as an altered ligand for a T-cell clone (AC20) that expresses the V ␤ 17 TCR. SEB⌬61Y possessed the ability to partially activate T-cells based on its capacity to induce TCR down-regulation and IL-2 receptor up-regulation but failed to elicit the secretion of IL-2, IL-3, IL-4, interferon ␥, or cell proliferation. Similarly, substitution of Tyr 61 for Ala also causes a change in its V ␤ usage profile (32).
Conclusions-Previous workers (15) have suggested that Thr 112 is directly involved in interactions with the TCR. Others (16) have suggested that this residue plays only an indirect role in the interaction of SEB with the TCR. Our results now allow the role of Thr 112 to be defined more precisely. From our data we can conclude that the mutation of Thr 112 to Ser did not ablate T-cell reactivity. In fact, the ability to bind MHC class II as well as to activate human V ␤ 12 T-cells was unaffected, and the only detectable effect was a selective reduction in the activation of V ␤ 3.1 T-cells. Our structural data suggest that Thr 112 affects TCR binding indirectly through maintaining the local environment of the TCR-binding site by retaining contacts with other residues in the site, namely Tyr 61 and Tyr 89 . Unlike Tyr 61 deletion mutants, the ability of the T112S mutant to stimulate some subsets of T-cells is maintained. Hayball and Lake (30) propose that residues Asn 60 and Tyr 61 confer specificity for human TCR V ␤ 17 and murine TCR V ␤ 7 and 8.1. However, the effect of the Thr 112 -Ser mutation on the human TCR V ␤ 17 T-cell response was not investigated in the present study.
The V ␤ profile does differ between the T112S mutant and the wild type toxin, but the mutation causes a less dramatic effect than deletion of Tyr 61 , suggesting that Thr 112 has a more subtle effect on the TCR-SEB interaction site. Structurally these altered specificities could be accounted for in several ways. First, it is important to note that residues 60 and 61 are located in the middle of a loop region. The conformation of this loop appears to depend on the stabilizing interactions with Thr 112 . Therefore, it is possible that for a proportion of those T112S molecules binding to the TCR, the loop and hence these residues are in a (near) correct orientation. In turn, this could be dependent on the conformational flexibility of particular V ␤ chains.
Second, contact with the TCR V ␤ chain in these positions could be preserved by a network of water molecules replacing Tyr 61 and replacement of Asn 60 by Lys 109 . In our model of T112S in complex with TCR V ␤ chain, the amide nitrogen atom of Lys 109 is in the same region of space as the amide nitrogen atom of Asn 60 and may therefore form a hydrogen bond with the TCR V ␤ chain. However, in the T112S mutant structure, Lys 109 is located at the non-crystallographic dimer interface, forming contacts with residues 17 and 18 of the second molecule. As such it is not possible to envisage a direct role for Lys 109 in TCR binding because of these crystal packing interactions in the mutant structure. However, we cannot definitely rule out the possibility of a role for this residue in solution.
A comparison between the structures of SEB and SEC2 indicated that amino acids Leu 20 , Glu 22 , Val 26 , Asn 31 , Ala 87 , Tyr 91 , and Glu 92 of SEB account for differences in V ␤ specificity between the two superantigens (16). Our data showing the selective reduction in activation of human V ␤ bearing T-cells suggests that Thr 112 is another residue that determines V ␤ specificity for SEB due to the effect of the T112S mutation on V ␤ 3.1 but not V ␤ 12 T-cell activation. Our results add significantly to construction of a complete picture of the topology of the interaction site between superantigens and the TCR.