Major histocompatibility complex class II-associated peptides determine the binding of the superantigen toxic shock syndrome toxin-1

2894 A. von Bonin et al. Eur. J . Immunol. 1995.25: 2894-2898

Arne von Bonin', Svenja Ehrlich', Georg Malcherek' and Bernhard Fleischer'

Bernhard-Nocht-Institut for Tropical Medicine, Hamburg, Germany Department of Neurology, University of Tubingen, Tubingen, Germany

Major histocompatibility complex class 11- associated peptides determine the binding of the superantigen toxic shock syndrome toxin-1

Superantigens bind to major histocompatibility complex (MHC) class I1 proteins and interact with variable parts of the Tcell antigen receptor (TCR) P-chain. Cross-linking the TCR with MHC class I1 molecules on the antigen-presenting cell by the superantigen leads to Tcell activation that plays an essential role in pathogenesis. Recent crystallographic data have resolved the structure of the complexes between HLA-DR1 and staphylococcal enterotoxin B (SEB) and toxic shock syndrome toxin-1 (TSST-l), respectively. For TSST-1, these studies have revealed possible contact sites between the superantigen and the HLA- DR1-bound peptide. Here, we show that TSST-1 binding is dependent on the MHC-11-associated peptides by employing variants of T2 mutant cells deficient in loading of peptides to MHC class I1 molecules as superantigen-presenting cells. On HLA-DR3-transfected T2 cells, presentation of TSST-1, but not SEB, was dependent on HLA-DR3-associated peptides. Thus, although these superantigens can be recognized in the context of multiple MHC class11 alleles and isotypes, they clearly bind to specific subsets of MHC molecules displaying appropriate peptides.

1 Introduction

T cell recognition of superantigen [ 1-31 is distinguished from recognition of conventional peptide antigen by two properties: T cell recognition of superantigen-MHC complexes is not restricted by classical MHC elements, and the T cell specificity for superantigen is determined by the VP element, essentially independent of other components of the Tcell receptor [4, 51. These two characteristics have been incorporated into a model in which the superantigen activates the T cell by directly cross-linking the MHC class I1 molecule on the antigen-presenting cell (APC) and the TCR VP element on the T cell [6,7]. Moreover, in contrast to processed protein-derived antigens, bacterial superantigens bind as intact molecules to MHC class11 molecules outside the peptide-binding groove [8, 91, form- ing a complex recognizable by Tcell receptors in a VP-specific manner [4]. Recent crystallographic data of complexes formed between TSST-l/HLA-DRl [lo] and SEBIHLA-DR1 [ l l ] molecules revealed that, although sharing an identical binding site and displaying a very simi- lar three-dimensional structure, both superantigens differ in the binding to the HLA-DR1-molecule: SEB exclus- ively contacts amino acids located in the HLA-DR1 a-chain whereas TSST-1 also contacts the @-chain and residues of the HLA-DR1-associated peptide [lo, 111. Binding studies employing mutated HLA-DR1 molecules also sug- gested a selective binding of TSST-1 and SEB to different subsets of MHC-molecules [ 121.

[I 147281

Correspondence: Arne von Bonin, Bernhard-Nocht-Institut fur Tropenmedizin, Bernhard-Nocht-Str. 74, D-20359 Hamburg, Ger- many (Fax.: +49 40-31182386)

Abbreviations: TSST-1: Toxic shock syndrome toxin-1 SEB: Sta- phylococcal enterotoxin B ApoB: Apolipoprotein B alAR a-1 Anti-trypsin

Key words: Superantigen /Toxic shock syndrome toxin-1 / Major histocompatibility complex class IT / Peptide

To address the role of MHC-associated peptides in superantigen binding, we made use of the MHC-deficient cell line T2 [13]. T2.DR-transfectants are able to present exo- genous peptides, but are unable to process native antigen due to a large deletion in the MHC-encoding gene locus [ 141. As a consequence, peptides associated with HLA-DR molecules on T2.DR transfectants are derived to a large extent (> 50 YO) from the invariant chain (CLIP, class 11- associated invariant chain peptides) [15, 161. The T2-DR- specific processing defect can be overcome by supertrans- fecting T2.DR-cells with genes encoding the HLA-DMA and HLA-DMB-chains, resulting in an almost normal MHC class I1 peptide repertoire on T2.DR/DM transfectants [17-191.

We used T1 (HLA-DR7') and T2 cells, both transfected with the DR3 cells allele (Tl.DR3, T2.DR3) and Raji cells (HLA-DR3+, DRwlO'). Depending on the superantigen used to stimulate different T cell hybridomas, we observed distinct capacities of Tl.DR3 and T2.DR3 cell lines to induce IL-2 production. Addition of peptides naturally associated with HLA-DR3 to T2.DR3 cells strongly influ- enced the stimulation and binding of TSST-1, but not of SEB. Thus, MHC classII-bound peptides play an impor- tant role in creating the trimolecular complex consisting of MHC, superantigen and TCR.

2 Materials and methods

2.1 Cell lines and flow cytometry

All cells were cultured in Iscove's medium supplemented with 5 % FCS, 2 m ~ glutamine and 50pglml gentamycin. Tl.DR3, T2.DR3 and T2.DR3 cells supertransfected with HLA-DMA and HLA-DMB molecules, T7..DR3/DM+, were all kindly provided by P. Cresswell. T2.DR3/DM+ cells express the 16.23 mAb epitope (data not shown and [ 191) and express less HLA-DR3/CLIP+-complexes. For selection, HLA-DR3 transfectants were maintained in

0014-2980/95/1010-2894$10.00 + .25/0 0 VCH Verlagsgesellschaft mbH, D-69451 Weinheim, 1995

Eur. J. Immunol. 1995.25: 2894-2898

0.5mg/ml G-418 (Gibco). For FACS staining, 1 x 10' of the indicated cells were incubated with 50 y1 supernatant of the anti-DR-specific mAb L243 or the DR3-specific mAb 16.23 for 30min at 4°C. After extensive washing, an FITC-labeled rabbit anti-mouse antiserum (0.5 yghample, Dianova, Hamburg) was added for 30min at 4°C. Cells were analyzed in a FACScan (Becton Dickinson, Moun- tain View, CA).

2.2 Proliferation assays

FRN2.7 cells (5 X 1O4/we1l) and 2.5 x 104/well of the indicated APC were co-cultured overnight (as triplicates) with the indicated amount of TSST-1. After 18 h of culture, the IL-2 content in the supernatant was determined in a standard CTLL assay [20]. TSST-1 was purified from a transfected Staphylococcus aureus as described [21]. Low molecular weight material was obtained as described else- where [22] in a slightly modified procedure. Briefly, 1 x 10' Raji cells were lysed with 10ml 0.1 % trifluoroacidic acid (TFA, pH2.0). Cellular debris was removed by ultracent- rifugation at 100 000 x g for 30 min at 4 "C. The supernatant was applied to a Centricon 10 spin ultrafilter (Amicon, Witten, Germany). The filtrate was lyophilized and dis- solved in 500yl serum-free CG medium (Vitromex, Ger- many). APC (5 x 10') were cocultured for 3 h at 37 "C with 50 (1OO)yl of the extract in 1ml Iscove's medium, cells were washed and incubated with 1 yg/ml TSST-1. The peptides naturally associated with HLA-DR3 ApoB (apolipoprotein B, residues 2177-2193: ISNQLTLDSNT- KYFHK) and alAT (al-anti-trypsin, residues 149-164: VDTFLEDVKNLYHSEA) were incubated at 10 yg/ml with T2.DR3 cells ( 5 X 10"/ml) for 3 h at 37 "C prior to the addition of the TSST-1 and Tcells. For SEB modification, Tl.DR3 (open squares) or T2.DR3 cells (filled circles) (5 x 105/ml) were incubated with the indicated serial dilu- tions of recombinant SEB (Serva, Germany) for 18h at 37 "C.

3 Results and discussion

3.1 TSST-1 induces a strong IL-2 response by Tl.DR3

All cells expressed comparable amounts of HLA-DR on the cell surface, but only Raji and Tl.DR3 cells express a conformational epitope for the HLA-DR3-specific mAb 16.23 [23] (Fig. 1), implying that 16.23 does not recognize HLA-DRYCLIP complexes. Titration of TSST-1 on Raji and Tl.DR3 cells resulted in a strong stimulation of the VB2' Tcell hybridoma FRN2.7 [24], whereas T2.DR3 induced a less pronounced IL-2 production (Fig. 2). This cannot be explained by the co-expression of HLA-DR7 molecules on Tl.DR3 cells or HLA-DR10 molecules on Raji cells, since the EBV-transformed B cell line ALK, homozygous for HLA-DR3, stimulated FRN2.7 cells to produce IL-2 levels comparable to T1.DR3 and Raji cells (data not shown).

cells, but not T2.DR3 cells

3.2 Peptides naturally associated with HLA-DR3 enhance TSST-1-induced stimulation of T2.DR3 cells

To obtain peptides naturally associated with HLA-DR3 from the stimulating Raji cells, low molecular weight cellu-

MHC class I1 peptides determine the binding of TSST-1 2895

e, Tl.DR3 Raji

h n, T2.DR3 $, ~ T2

Figure I . HLA-DR expression on Raji, Tl.DR3,T2.DR3, and T2 cells. All HLA-DR3' cells express comparable amounts of HLA-DR-molecules (shaded peaks) on the cell surface, but differ in the expression of the 16.23-specific HLA-DR3 epitope. The diagrams show the mean fluorescence intensity (x-axis) versus the relative cell number (y-axis). For negative controls, cells were incubated without the first antibody (small arrow).

lar extracts enriched for HLA-DR-associated peptides were prepared. These peptide pools were added to T2.DR3 cells, resulting in a significant increase in the response of FRN2.7 cells, indicating a possible role of HLA-DR3-associated peptides for the binding specificity of TSST-1. The same extracts show no effect onHLA-DR- T2 cells (Fig. 3A). Moreover, when T2.DR3/DM cells were used as APC, again, a dramatic increase in the response to the TSST-1 stimulus was seen (Fig. 3B). To address directly the role of HLA-DR3-associated peptides in TSST-1 binding, we incubated T2.DR3 cells in the pres- ence of the peptides ApoB2877-92 or alAT149-162, which are naturally associated with HLA-DR3 [25] and exhibit a comparable high affinity for HLA-DR3 molecules [26]. Addition of the peptides to T2.DR3 cells resulted in a sub- stantial increase in the TSST-1-induced response (Fig. 3 C). Peptide addition to untransfected T2 cells had no effect (Fig. 3D), nor did the addition of ovalbumine-derived peptide OVA325-339, which is known to bind to the mouse I-Ad allele but not to HLA-DR3, (data not shown).

0 2 0.5 0.1

pg/ml TSST-1

Figure 2. After TSST-1 incubation, Tl.DR3 and T2.DR3 cells differ in their capacity to induce IL-2 secretion, by VP2+ Tcell hybridomas. Raji (A), Tl.DR3 (O),T2.DR3 (0) andT2 (+)were compared for their ability to induce a specific IL-2 response in FRN2.7 (VP2+) cells after modification with the indicated amount of purified TSST-1. Negative controls were FRN2.7 cells incubated with APC without the adition of TSST-1 (5 230 cpm). The data represented are typical for three independent experiments.

2896 A. von Bonin et al. Eur. J. Immunol. 1995.25: 2894-2898

TBDRJ T2 Raji --

TSST-1 (lwdml): + + + - + +

Rsji-Extract (pl): - SO l o o 100 - 100 -

8wo nl A 8000

2 0.5 0.1

pg/ml TSST-1

& TZDRYDM B

2 0.5 0.1

pg/ml TSST-1

B 2000 -

2 0.5 0.1

pglml TSST-1

3.3 Addition of peptides naturally associated with HLA- DR3 enhance the binding of TSST-1 on T2.DR3 cells

Since addition of peptides naturally associated with HLA- DR3, either derived from cellular extracts or given as purified synthetic peptides, induced a higher stimulation of Vg+ T cell hybridomas, the most likely explanation is that the TSST-1 binding to MHC-I1 complexes is facilitated. Therefore, we performed a surface staining with an TSST- I-specific antiserum on Tl.DR3 and T2.DR3 cells to detect surface-bound TSST-1. As shown in Fig.4 addition of TSST-1 to Tl.DR3 cells resulted in a strong specific stain-

Figure 4. HLA-DR3-specific peptides enhance the binding of TSST-1. Tl.DR3 cells (A) and T2.DR3 cells (B) were incubated with (+ TSST-1) or without (-TSST-1) 5 pg/ml recombinant TSST-1 for 2 h at 37°C. After washing, the cells were stained in a standard staining procedure (see Sect. 2.1) employing 30 p1 anti-TSST-1- specific antiserum (rabbit, 1 : 300). Addition of 10 pg/ml alAT was performed overnight at 37 "C. Prior to TSST-1 addition, surplus peptide was removed by washing. In the same experiment, MHC- I I T 2 cells exhibited no specific TSST-1 binding (data not shown). Depicted is the mean fluorescence intensity (x-axis) versus the relative cell number (y-axis).

Figure 3. HLA-DR3-associated peptides have a positive effect on TSST-1-induced stirnu- lation. (A) Low molecular weight extracts derived from Raji cells were preincubated with T2.DR3 cells (shaded bars) and T2 cells (black bars) prior to incubation with TSST-1. Unspe- cific background due to APC alone was < 300cpm. (B) T2.DR3/DM transfectants were compared with T2.DR3 cells and T1.DR3 for their ability to induce IL-2 in VP2' FRN2.7 T cell hybridomas after TSST-1 stimulation. Addition of the peptides naturally associated with HLA-DR3 ApoB (0) and a lAT (W) enhance the TSST-1-induced stimulation on T2.DR3 cells (C) (0) but not on untransfected T2 cells (D). After 18 h of coculture, the IL-2 content in supernatants was determined in a standard CTLL assay. In all experiments, the unspecific background of unmodified APC (without TSST-1) was below < 200cpm; Tl.DR3 cells as APC showed in the same test a proliferative response of 14 166 ? 1421 cpm (1 pg/ml TSST-1).

ing. In contrast to Tl.DR3 cells, T2.DR3 cells showed a less pronounced TSST-1 binding. Addition of the HLA- DR3-derived alAT peptide to T2.DR3 cells induced a reproducible significant increase in the surface staining of TSST-1. Note that a relatively high concentration of TSST- 1 was used in this experiment. In the proliferation assays, lower concentrations of TSST-1 were used, which were too low to be detectable on the cell surface. This small increase of peptide-induced MHC/TSST-1 complexes, however, was sufficient to stimulate a response of Vg2' Tcell hybridomas. These results are in line with findings from Valitutti et al. [27] and others showing that very low num- bers of functional MHC/peptide complexes are sufficient for an effective stimulation of T cells.

3.4 TSST-1, but not SEB, binding is dependent on HLA- DR3-associated peptides

It is known that TSST-1 and staphylococcal enterotoxin B (SEB) do not block each others' binding [28,29]. One rea- son for this lack of competition might be a differential binding of TSST-1 and SEB to MHC class I1 subsets displaying different peptides. As shown in Fig. 5, a different situation was observed when Tl.DR3 and T2.DR3 cells were pre-incubated with SEB. In contrast to TSST-1- induced stimulation, T2.DR3 cells were potent stimulator cells when SEB was used as a superantigen, whereas a significantly reduced stimulation was observed when Tl.DR3 cells were incubated with SEB. The addition of Raji cell- derived peptide extracts to SEB-modified T2.DR3 cells had no effect, and the addition of ApoB2877-92 and alAT149-162 peptides did not change the SEB-induced stimulation. These results again indicate that SEB binds in

Eur. J. Immunol. 1995.25: 2894-2898 MHC class I1 peptides determine the binding of TSST-1 2897

antigen-presenting cell. The interaction of these superantigens with only variable parts of the immune system's receptor molecules is probably the basis for the high poly- morphism of the staphylococcal enterotoxin superantigen family.

We thank f? Cresswell (Yale University, Section of Immunobiology, New Haven, CT) for giving us the n, T2. DR3 and TI. DR3 cell lines, and S. Frosch and A . Fuhrmann for helpful discussions and the critical reading of the manuscript.

Received August 9, 1995; accepted August 21, 199.5

1 0.1 0.01

pglml SEB

10000

4000

+ TZ.DR3 + Extr.

U TZ.DR3

+ TZ.DR3 + Extr. 0-1 I I

1 0.1 0.01

p'glml SEB

U TZ.DR3+ApoB

18000

16000

12000

10000 1 0.1 0.01

pg/ml SEB

Figure 5. Optimal SEB binding requires peptide MHC class I1 complexes that are distinct from peptide/MHC class I1 molecules that are optimal for TSST-1 binding. (A) Compared to Tl.DR3 cells, T2.DR3 cells incubated with SEB displayed a significantly higher IL-2 response in Vfi14' cells. The IL-2 content in the supernatant was determined as described in Fig. 2. Controls without the addition of SEB were 1555 k 154 cpm (Tl.DR3) and 1756 f 134 cpm (T2.DR3). (B) The addition of low molecular weight extracts from Raji cells (100 ~ 1 , compare also Fig. 3) does not improve the IL-2 response of T2.DR3 cells. (C) In contrast to the TSST-1- dependent stimulation, the peptides naturally associated with HLA-DR3 ApoB2177-2193 and alAT149-164 do not show a positive effect on SEB-induced stimulation. The depicted data are typical for three independent experiments. Controls without SEB were < 2200 cpm (B) and < 1600 cpm (C).

a different way to HLA-DR3 molecules than TSST-1. The binding of SEB might be peptide independent, or might depend on peptides expressed extremely well on T2.DR3 cells.

4 Conclusions

Binding of superantigens can be mediated by various mechanisms, e .g . staphylococcal enterotoxin A (SEA) binding is dependent on Zn2+ bridges and corresponding histidine residues in the MHC class I1 molecule [30]. Our studies reveal that two superantigens, both members of the polymorphic staphylococcal enterotoxin family, depend differentially on the peptides present in the binding grooves of th MHC class I1 molecule. This finding helps to explain why these two superantigens do not com- pete with each other and bind to different subsets of MHC molecules. Our data show that superantigens are not only specialized in using a diverse repertoire of variable parts of the T cell receptor, but also exploit the extremely high vari- ability of MHC/peptide complexes expressed on the

5 References

1 Marrack, P., Winslow, G. M., Choi, Y., Scherer, M., Pullen, A. , White, J . and Kappler, J. W., Irnrnunol. Rev. 1993.131: 79.

2 Fleischer, B., Chem. Immunol. 1992. 55: 1. 3 Janeway, C. A. Jr., Adv. Immunol. 1991. 50: 1. 4 White, J., Herman, A., Pullen, A. M., Kubo, R., Kappler, J.

and Marrack, P., Cell 1989. 56: 27. 5 Woodland, D. L. and Blackman, M. A, , Imrnunol. Today

1993. 14: 208. 6 Janeway, C. A. Jr., Yagi, C., Conrad, P. J., Katz, M. E., Jones,

B., Vroegop, S. and Buxser, S . , Imrnunol. Rev. 1989. 107: 61. 7 Janeway, C. A. Jr., Cell 1990. 63: 659. 8 Dellabona, P., Peccoud, J., Kappler, J., Marrack, P., Benoist,

C. and Mathis, D., Cell 1990. 62: 1115. 9 Herman, A. , Labrecque, N., Thibodeau, J., Marrack, P.,

Kappler, J . and SCkaly, R. P., Proc. Natl. Acad. Sci. USA 1991. 88: 9954.

10 Kim, J., Uran, R. G., Strominger, J. L. and Wiley, D. C., Sci- ence 1994. 266: 1870.

11 Jardetzky, T. S. , Brown, J . H., Gorga, J . C., Stern, L. C., Urban, R. G., Chi, Y., Stauffacher, C., Strominger, J. L. and Wiley, D. C., Nature 1994. 368: 711.

12 Thibodeau, J., Cloutier, I., Lavoie, P. M., Labrecque, N., Mourad, W., Jardetzky, T. and SCkaly, R. P., Science 1994.266: 1874.

13 Salter, R. D. , Howell, D. N. and Cresswell, P., Irnrnunogenet- ics 1985. 21: 235.

14 Riberdy, J. M. and Creswell, P., J . Irnrnunol. 1992.148: 2.586. 15 Sette, A., Ceman, S., Kubo, R. T., Sakaguch, K . , Appella, E.,

Hunt, D. F., Davis, T. A. , Michel, H. , Shabanowitz, J . , Rudersdorf, R., Grey, H. M. and DeMars, R., Science 1992. 258: 1801.

16 Riberdy, J. M., Newcomb, J. R., Surman, M. J., Barbosa, J . A. and Cresswell, P., Nature 1992. 360: 474.

17 Morris, P., Shaman, J., Attaya, M., Amaya, M., Goodman, S. , Bergman, C., Monaco, J. J. and Mellins, E., Nature 1994. 368: 551.

18 Fling, S. P., Arp, B. and Pious, D., Nature 1994. 368: 54. 19 Denzin, L. K., Robbins, N. F., Carboy-Newcomb, C. and

Cresswell, F?, Immunity 1994. I : 595. 20 Grabstein, K., Eisenmann, J., Modizuki, D., Schanebeck, P.,

Cordon, P., Hopp, T., March, C. and Gilles, S., J . Exp. Med. 1986. 163: 145.

21 Blanco, L., Choi, E . M., Connoly, K., Thompson, M. R. and Bonventre, F? F., Infect. Immun. 1990. 58: 3020.

22 Rotzschke, O., Falk, K., Wallny, H.-J., Faath, S. and Ram- mensee, H.-G., Science 1990. 249: 283.

23 Johnson, J. P., Meo, T., Riethmuller, G. , Schendel, D. J. and Wank, R., J . Exp. Med. 1982. 156: 104.

24 Romagne, F., Besnardeau, L. and Malissen, B., Eur. J . Irnmu- nol. 1992. 22: 2749.

25 Malcherek, G., Falk, K., Rotzschke, O., Rammensee, H.-G., Stevanovic, S . , Gnau, V., Jung, G. and Melms, A,, Int. Immu- nol. 1993. 5: 1229.

2898 A. von Bonin et al. Eur. J. Immunol. 1995.25: 2894-2898

26 Malcherek, G., Gnau, v., Stevanovic, S. , Rammensee, H. G. ,

27 Valitutti, S. , Muller, S. , Cella, M., Padovan, E. and Lanzavec-

28 Chintagumpala, M. M., Mollick, J. A. and Rich, R. R., J.

29 Scholl, P. R., Diez, A. and Geha, R. S . , J. lmmunol. 1989.

30 Fraser, J. D., Urban, R. G. , Strominger, J. L. and Robinson, Jung, G. and Melms, A., J. Immunol. 1994.153: 1141.

chia, A., Nature 1995. 375: 148.

Immunol. 1991. 147: 3876.

143: 2583.

H., Proc. Natl. Acad. Sci., USA 1992. 89: 5507.

Documents

Major histocompatibility complex class II-associated peptides determine the binding of the superantigen toxic shock syndrome toxin-1