Identification of an I-E^d-Restricted T-Cell Epitope of Escherichia coli Outer Membrane Protein F

Abstract

A predominant T-cell epitope of Escherichia coli outer membrane protein F (OmpF) that encompasses amino acids 295 to 314 was identified in H-2^d mice. BALB/c-derived T-cell hybridomas generated against this region were CD3⁺, CD4⁺, CD8⁻, and T-cell receptor αβ⁺ and secreted TH-1-associated cytokines (interleukin-2 [IL-2] and gamma interferon), but not a TH-2-associated cytokine (IL-4), when restimulated with peptide 295-314. Class II⁺ mouse lymphoma (A20) cells, but not class II(−) mouse mastocytoma (P815) cells, supported IL-2 secretion of hybridomas when substituted for syngeneic splenocytes as antigen-presenting cells (APCs). Antibodies specific for I-E^d blocked IL-2 secretion by hybridomas, but I-A^d-specific antiserum did not. When transfected L cells expressing I-A^d (AαAβ^d), I-E^d (EαEβ^d), or the hybrid molecule I-EαAβ^d were used as APCs, hybridomas recognized peptide only when presented by the I-E^d-transfected cells. When peptide 295-314 truncated at either the C or the N terminus of the sequence was used, the minimal epitope was determined. Critical residues were determined by using alanine-substituted peptide analogues. T-cell hybridomas were only stimulated by peptides that encompassed amino acids 295 to 303 (9-mer), and the core sequence required a minimum of three additional amino acids at either the amino or the carboxy terminus to induce IL-2 secretion. Critical residues were determined to be phenylalanine at position 295, threonine at position 300, and tyrosines at positions 301 and 302. This study is the first to identify a minimal T-cell epitope and major histocompatibility complex restriction element of the OmpF protein and confirms previous observations that there is considerable degeneracy in the length of peptides that can bind I-E^d and variability in the amino acid composition of the C and N termini of these peptides.

Antigen recognition by CD4⁺ T cells involves the uptake and proteolysis of native proteins by antigen-presenting cells, followed by the association and surface presentation of smaller peptide fragments bound to class II major histocompatibility complex (MHC) molecules. Interaction of this complex with clonotypic receptors on the surface of T cells results in a cascade of intracellular events collectively termed T-cell activation. Acid elution, followed by protein sequencing of naturally processed peptides from affinity-purified human and murine class II molecules (13, 14, 17, 29, 41, 47, 64, 65), has shown considerable variability in the length of bound peptides (11 to 17 amino acids in length) and that a single determinant may be presented as multiple peptide species, depending on the length of the amino- or carboxy-terminal flanking residues. Crystallographic analysis of the MHC class II molecule HLA-DR1 (9, 62) provided the structural explanation for the heterogeneity in length of class II-bound peptides since it was observed that peptides are bound in an extended conformation inside the binding groove, which is open at both ends, unlike the class I peptide-binding groove. Additionally, binding studies have revealed that peptide anchor residues (or their respective binding sites within the groove) are more degenerate in their specificity compared to the stringent binding of class I ligands. While allele-specific binding motifs for class II molecules have been more difficult to identify because of the open-ended structure of the peptide-binding groove and the resulting heterogeneity of bound peptides, several allele-specific motifs have emerged (http://syfpeithi.bmi-heidelberg.com/).

Porins are a family of heterotrimeric pore-forming molecules present in the outer membranes of gram-negative members of the family Enterobacteriaceae. Monomeric porin molecules associate to form stable trimeric hydrophilic transmembrane channels that allow passive diffusion of solutes, including antibiotics, across the lipid bilayer. The primary amino acid sequence of porins from several species of the family Enterobacteriaceae show a high degree of intra- and interspecies homology, and studies using polyclonal and monoclonal antibodies also demonstrate immunological relatedness (8, 59, 69). Porins from Escherichia coli (outer membrane protein F [OmpF], OmpC, and PhoE) are best characterized in terms of their functional, genetic, and immunochemical properties. The three-dimensional structure of E. coli OmpF has been determined by X-ray crystallographic analysis (15), allowing correlation of functional and immunologic properties with the structural properties of the molecule. The structure consists of 16 antiparallel β strands forming a barrel that is embedded in the membrane. The external segments of the barrel consist of loop structures, seven of which are surface exposed and one of which folds back inside the barrel. The trimeric complex is formed by hydrophobic interaction between side chains of amino acid residues that forms the external surfaces of adjacent barrels. There is strong interest in the enterobacterial porins for studies of antibiotic resistance, for their potential use in diagnostic assays, and, because of their antigenic and immunomodulating properties (1, 6, 23, 24, 29, 34, 57, 58), as potential vaccine candidates. The identification of species-specific (or genetically permissive) B- and T-cell epitopes has provided useful insights into the evaluation of porin molecules for this purpose.

With overlapping synthetic polypeptides spanning the entire length of the E. coli porin, the OmpF sequence, a dominant T-cell epitope encompassing amino acids 295 to 314 was identified in a genetically permissive manner in H-2^d, H-2^k, and H-2^b mice (69). Cytokine quantitation revealed that proliferating T cells were polarized toward a TH-1 T-helper-type response. Since TH-1 cells exert a protective role against facultatively intracellular pathogens by activation of antimicrobial effector functions and direct cytotoxic activity against infected macrophages (48), better characterization of this response was warranted. T-cell hybridomas were produced with T lymphocytes from BALB/c (H-2^d) mice immunized with native OmpF. T-cell hybridomas specific for the immunodominant sequence, OmpF amino acids 295 to 314, were identified and used to determine H-2 restriction of this epitope and the minimal amino acid sequence required for hybridoma stimulation. By replacement of each amino acid in the minimal peptide with an alanine residue, those amino acids critical to T-cell activation were identified.

MATERIALS AND METHODS

Porin purification.

The CM6 strain of E. coli B/r, which produces OmpF but not OmpC (5), was the source of OmpF. Bacteria were grown to mid-log phase in nutrient media at 37°C. Porin was extracted by the method of Nurminen (42), followed by size exclusion chromatography on Sephacryl S-200 in the presence of 1% sodium dodecyl sulfate to reduce lipopolysaccharide contamination. Levels of residual lipopolysaccharide per microgram of porin did not exceed picogram amounts, as detected with the Limulus amoebocyte lysate assay (E-Toxate kit; Sigma Chemical Company, St. Louis, Mo.). The purity of the preparation was assessed by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (36) with 10 to 20% linear gradient gels. Total protein levels were determined with a bicinchoninic acid assay kit (Pierce Chemical Company, Rockford, Ill.).

Synthetic peptides.

Synthetic peptides were produced by standard 9-fluorenylmethoxycarbonyl polyamide solid-phase synthesis with an Applied Biosystems (Foster City, Calif.) 430A peptide synthesizer. Peptides were synthesized on p-hydroxymethylphenoxymethyl polystyrene resins with 2-(1H-benzotriazol-1-yl)-1,1,3,3-tetramethyluronium hexafluorophosphate as the coupling agent. Peptides were deprotected and cleaved from the resin in 92 to 95% trifluoroacetic acid-H₂O containing the appropriate scavenger chemicals. Peptides were purified by reverse-phase (C₈) high-pressure liquid chromatography with an acetonitrile gradient of 0 to 70% and 0.1% trifluoroacetic acid as the mobile phase. Acidic peptides were purified on a reverse-phase polymeric (300-A-polystyrenedivinylbenzene) column (Vydac, Hesperia, Calif.) with 5 mM ammonium acetate, pH 8.5, as the mobile phase. Molecular weights of selected peptides were confirmed by electrospray mass spectrometry.

Cell lines.

P815 DBA/2 mastocytoma cells (class II negative), A20 BALB/c B-lymphoblastoid cells (class II positive), L cells, and the T-cell receptor (TCR)-negative T-cell fusion partner BW5147.G.1.4 (αβ⁻) were purchased from the American Type Culture Collection (Manassas, Va.). Cells were maintained Dulbecco's modified Eagle medium (DMEM) supplemented with 10% heat-inactivated fetal calf serum, 2 mM l-glutamine, 100 U of penicillin per ml, 100 μg of streptomycin sulfate per ml, and 5 × 10⁻⁵ M 2-mercaptoethanol (DMEM complete). Transfected L-cell lines RT 1.1.12, which does not express MHC class II; RT 2.3.3H-D6, expressing the murine class II molecule I-A^d; RT10.3B-c1, expressing the I-E^d molecule; and RT 7.7H 14.3, expressing the mixed-isotype I-EαAβ^d molecule, were kindly provided by R. N. Germain, National Institute of Allergy and Infectious Diseases, National Institutes of Health, Bethesda, Md. Transfected cells were maintained in complete DMEM supplemented with hypoxanthine (15 μg/ml)-aminopterin (0.2 μg/ml)-thymidine (5 μg/ml) and were periodically tested for surface class II expression by staining with class II-specific antibodies with a Coulter Epics Elite flow cytometer.

Antibodies.

Monoclonal antibodies specific for murine CD3ɛ, CD4, CD8a, TCR, murine class II molecules, and murine interleukin-2 (IL-2), IL-4, and gamma interferon (IFN-γ) were purchased from BD Biosciences Pharmingen (San Diego, Calif.).

Immunization of mice.

BALB/cJ mice were purchased from Jackson Laboratories (Bar Harbor, Maine), used at an age of 6 to 8 weeks, and maintained in an Association for Assessment and Accreditation of Laboratory Animal Care-accredited facility. Mice were immunized subcutaneously at the base of the tail with 100 μg of OmpF emulsified in complete Freund's adjuvant.

T-cell hybridomas.

Spleens and popliteal and inguinal lymph nodes were harvested from mice 8 to 10 days following immunization. Red blood cells were lysed with 0.17 M ammonium chloride, followed by depletion of B cells and macrophages on columns containing glass beads coated with anti-mouse immunoglobulin (R & D Systems, Minneapolis, Minn.). The postcolumn purity of the T-cell suspension was >85% as determined by staining with anti-Thy-1 antibody. T cells were activated in vitro for 2 days with syngeneic spleen cells and 10 μM peptide 295-314 in complete DMEM. Hybridomas specific for peptide 295-314 were produced by a standard fusion protocol (31) with polyethylene glycol 1500 in conjunction with the fusion partner BW5147. Following fusion, cells were distributed into 96-well flat-bottom culture plates. Hybridomas were selected by addition of DMEM supplemented with hypoxanthine-aminopterin-thymidine, expanded into 24-well plates, and tested for surface CD3 expression by flow cytometry. CD3⁺ hybridomas were tested for specificity by incubation of hybridoma cells with peptide (10 μM) and 4 × 10⁵ mitomycin C (Sigma)-treated syngeneic spleen cells in a total volume of 250 μl in 96-well plates. After 24 h, 100 μl of supernatant was removed and tested for IL-2 production. To determine the TH phenotype, some duplicate cultures were tested for IL-4 and IFN-γ production. Hybridomas were cloned with a Coulter Epics Elite flow cytometer-cell sorter equipped with an autoclone attachment. A stable clone (2C7/C8) was chosen for further evaluation on the basis of its high levels of IL-2 secretion in response to peptide and native porin trimer.

Assays for T-cell stimulation.

All assays were conducted with 96-well round-bottom culture plates incubated at 37°C in 5% CO₂ in air. For MHC restriction assays, 10⁴ hybridoma cells, 10⁵ antigen-presenting cells, and peptide were added to each well in a total volume of 250 μl. Antisera specific for class II molecules were included in the incubation mixture in some wells at a concentration of 5 μg in a 250-μl total volume. Minimal sequence recognized by hybridoma cells was determined with peptide homologues titrated over a range of concentrations and tested for the ability to stimulate IL-2 production. All experiments were repeated a minimum of three times, and the experimental error was <10%.

Cytokine quantitation.

Cytokine concentrations of culture supernatants were determined by capture enzyme-linked immunosorbent assay with paired monoclonal antibodies specific for the murine cytokine of interest (BD Biosciences Pharmingen). Dynatech (Chantilly, Va.) Immulon II plates were coated overnight with the capture antibody (5 μg/ml) suspended in 0.1 M carbonate-bicarbonate buffer, pH 9.6. Plates were washed three times with phosphate-buffered saline-0.5% Tween 20, and then 100 μl of culture supernatant was added. Plates were incubated overnight at 4°C and washed four times with phosphate-buffered saline-Tween 20, and then 100 μl of biotin-conjugated anti-murine cytokine (2 μg/ml) was added. Plates were incubated for 2 h and washed six times, and then 100 μl of a 1:1,000 dilution of peroxidase-streptavidin (Jackson Immunoresearch, West Grove, Pa.) was added. After 30 min, the plates were washed eight times and the color was developed with tetramethylbenzidine. A₄₅₀ was read with a Spectramax 250 enzyme-linked immunosorbent assay plate reader (Molecular Devices, Sunnyvale, Calif.). Standard curves were generated with serial dilutions of murine recombinant cytokines (R & D Systems).

RESULTS

T-cell hybridoma.

A T-cell hybridoma specific for immunodominant determinant OmpF amino acids 295 to 314 was derived from BALB/c mice immunized with native porin emulsified in colonization factor antigen. Flow cytometric analysis of lymphocyte surface antigen expression confirmed that the cells were CD3⁺, CD4⁺, CD8⁻, and TCRαβ⁺ (Fig. 1). Cytokine quantitation of culture supernatant from hybridoma cells incubated with OmpF amino acids 295 to 314 revealed high concentrations of IL-2 (18.56 ± 0.67 ng/ml) and IFN-γ (12.38 ± 0.75 ng/ml) and no measurable IL-4 (0.08 ± 0.003 pg/ml), characteristic of a TH-1-type helper cell response. Peak concentrations of culture supernatant IL-2 were obtained after 18 to 24 h of culture, whereas a 72- to 96-h incubation period was required for maximal levels of IFN-γ. For this reason, measurements of culture supernatant IL-2 were used for determination of MHC restriction and identification of the minimal epitope and critical peptide residues for hybridoma stimulation.

FIG. 1.

Surface lymphocyte antigen expression of hybridoma T cells as assessed by two-color flow cytometry. Panels: A, isotype controls; B, 97.17% of cells are CD3ɛ⁺ and TCRαβ⁺; C, 97.85% of cells are CD4⁺ and CD8a⁻; D, 97.17% of cells are CD3ɛ⁺ and TCRγδ⁻. PE, phycoerythrin; FITC, fluorescein isothiocyanate.

MHC restriction.

MHC restriction of the 20-mer peptide consisting of OmpF amino acids 295 to 314 (amino acids with no signal peptide) and the minimal epitope FEVGATYYFNKN was initially determined by replacing syngeneic spleen cells with several cultured cell lines for use as antigen-presenting cells. MHC class II⁺ B-lymphoblastoid cells derived from BALB/c mice (A20 cells) were able to support T-cell hybridoma activation, but the class II⁻ cell line P815 (DBA/2 mouse mastocytoma) was not (Fig. 2). Specificity of restriction was determined by using transfected L cells as antigen-presenting cells (Fig. 3) or by the inclusion of specific anti-MHC class II monoclonal antibodies in the T-cell proliferation assay (Fig. 4). Transfected L cells expressing I-EαEβ^d were able to support T-cell activation, but L cells expressing I-AαAβ^d were not. L cells transfected with the hybrid molecule I-EαAβ^d and L-cell transfectants that do not express class II molecules did not support T-cell activation. To further corroborate MHC restriction, class II-specific monoclonal antibodies specific for I-E^d,k,r,p (clone 14-4-4S) and I-E^d/I-A^d (clone 29-G) were able to completely inhibit IL-2 production by the T-cell hybridoma, while the anti-I-A^d antibody (clone 39-10-8) had no effect. Antibodies specified for I-E^d completely inhibited the T-cell responses to the full-length 20-mer and to the minimal epitope.

FIG. 2.

Class II MHC restriction. A20 cells (class II⁺) were able to present peptide to T-cell hybridomas and supported IL-2 production. When P815 cells (class II⁻) were used as antigen-presenting cells, no stimulation of hybridomas was observed.

FIG. 3.

Class II MHC restriction with transfected L cells for antigen presentation. Hybridomas recognized antigen only when presented by L cells transfected with and expressing surface I-E^d (EαEβ). Hybridomas were unable to recognize antigen presented by transfected L cells expressing I-A^d (AαAβ), L cells expressing the mixed-isotype molecule I-EαAβ^d, or transfected L cells that do not express MHC class II molecules.

FIG. 4.

Inhibition of IL-2 production by MHC class II-specific monoclonal antibodies. Inclusion of class II-specific monoclonal antibodies in the hybridoma stimulation assay demonstrated that antibodies that bind the I-E^d molecule (clones 14-4-4S [specific for I-E^d,k,r,p] and 29G [specific for I-E^d/I-A^d]) were able to completely inhibit T-cell activation, as measured by IL-2 secretion. There was no measurable effect on IL-2 levels when antibody 39-10-8 (specific for I-A^d) was included in the T-cell stimulation assay. The stimulatory capacity of the minimal determinant (peptide 295-306) was inhibited to a degree similar to that of parent peptide 295-314.

Identification of minimal epitope.

In a previous study (69), peptide 295-314 was identified as an immunodominant epitope. To more precisely define the CD4⁺ response to this epitope, T-cell hybridomas were stimulated with a set of peptides that were sequentially truncated by one amino acid at either the amino or the carboxy terminus. The sequences of these peptides and the IL-2 responses of hybridoma T cells to these peptides are shown in Fig. 5. The IL-2 response to the 12-mer peptides VNYFEVGATYYF and FEVGATYYFNKN indicated that these are the minimal epitopes required for stimulation, since reduction of the sequence by a single amino acid at either the amino or the carboxy terminus resulted in complete loss of hybridoma activity. Additionally, 12-mer peptides extended at the amino or carboxy terminus were able to stimulate hybridomas. Closer examination of these sequences revealed that each contains the core nine-amino-acid sequence FEVGATYYF, but a three-amino-acid extension on either the carboxy or the amino terminus is required.

FIG. 5.

Identification of the minimal sequence recognized by T-cell hybridoma. The boundaries of the minimal sequence required for stimulation of hybridoma cells, peptide 295-314, were sequentially reduced in size by one amino acid from the amino and carboxy termini, and these smaller peptides were tested for the ability to stimulate IL-2 secretion. The minimal sequence recognized encompassed the core amino acid sequence FEVGATYYF (residues 295 to 303); however, a minimum of three additional amino acids is required at either the amino or carboxy end of the core sequence.

Determination of critical residues.

To identify amino acids that are critical for hybridoma stimulation, analogues of the minimal epitope FEVGATYYFNKN were synthesized by sequential replacement of each amino acid with an alanine residue. The conservative substitution of glycine for alanine was made at position 299, where alanine appeared in the native sequence. When used as antigens for the stimulation of hybridomas, several peptides with substitutions in the core sequence were unable to support T-cell activation (Fig. 6). Alanine substitutions at positions F₂₉₅, T₃₀₀, Y₃₀₁, and Y₃₀₂ resulted in complete loss of stimulation. IL-2 secretion was reduced somewhat when V₂₉₇ was replaced with alanine. Replacement of the amino acids at positions E₂₉₆, G₂₉₈, A₂₉₉, and F₃₀₃ and alanine substitutions outside of the core sequence had little or no effect on the T-cell response.

FIG. 6.

Identification of critical amino acid residues. Alanine-substituted analogues of the minimal peptide FEVGATYYFNKN were used to stimulate T-cell hybridoma cells. Alanine substitutions for F₂₉₅, T₃₀₀, Y₃₀₁, and Y₃₀₂ eliminated activation of the T cells.

DISCUSSION

Effective T-cell help is necessary for the induction of B- and T-cell-dependent immunity to facultatively intracellular pathogens, some of which are gram-negative members of the family Enterobacteriaceae. A structural and antigenic feature common to these organisms is the family of porin molecules. Previous studies to investigate T-cell recognition of E. coli OmpF involved the use of overlapping synthetic polypeptides encompassing the entire OmpF sequence in proliferation assays with T cells isolated from several strains of inbred mice (69). A dominant TH-1-like, class II-restricted T-cell epitope was identified in mice immunized with native E. coli OmpF that encompasses amino acids 295 to 314 of the monomeric protein. This epitope was recognized in a genetically unrestricted manner since it was able to stimulate T-cell proliferation in several strains of sensitized mice (H-2^d,k,b). In examining the three-dimensional structure of OmpF (15), it was observed that the amino acid 295 to 314 T-cell epitope comprises a portion of the transmembrane region of the protein. The transmembrane regions of E. coli OmpF show limited sequence heterogeneity compared to OmpF sequences from other members of the family Enterobacteriaceae and compared to other E. coli porin molecules (OmpC, PhoE) (3, 56). The comparison of known porin sequences from E. coli (11), Salmonella enterica serovar Typhimurium (39), S. enterica serovar Typhi (45), and Shigella flexneri 2a (67) reveals a sequence identity of the majority of the amino acids in the amino acid 295 to 314 segment, with conservative substitutions in the remaining residues. Since conserved segments of the protein may serve as potential candidates for immunodiagnosis or peptide-based vaccines, epitopes within this sequence region of OmpF were selected for further studies to determine the specificity of the T-cell response with synthetic polypeptides. In this study, the T-cell hybridomas generated against the amino acid 295 to 314 sequence were CD3⁺, CD4⁺, CD8⁻, and TCRαβ⁺ and exhibited a TH-1-like phenotype, secreting high levels of IL-2 and IFN-γ but no IL-4 when peptide stimulated. One hybridoma was chosen for further characterization on the basis of its stability in culture and its high level of IL-2 secretion in response to peptide stimulation and to the native OmpF trimer. The minimal epitope was identified as a 12-mer composed of a core sequence of nine amino acids, FEVGATYYF (residues 295 to 303). Additionally, a minimum of three additional amino acids are required at either the carboxy or the amino terminus to induce T-cell stimulation. MHC restriction of peptide recognition was also studied in inhibition experiments with monoclonal antibodies specific for monomorphic determinants of the MHC class II molecules and by using transfected L cells expressing specific class II molecules as antigen-presenting cells. It was determined that presentation of this epitope is I-E^d restricted. Since L cells expressing the hybrid molecule I-EαAβ^d were not able to stimulate IL-2 secretion, it was concluded that a residue(s) present on the Eβ chain is essential for hybridoma activation, serving as a contact residue(s) for either peptide or TCR. L cells expressing the hybrid molecule were included in these studies, since I-EαAβ^d represents a small proportion of the surface class II presented on H-2^d B cells (61), and peptides restricted by this molecule have been identified (19, 40, 50).

These results conform to the structure of the class II binding site in that, unlike that of class I, the class II binding groove is open at both ends, allowing for variability in peptide length. Longer peptides bind via a core region of nine amino acids, with flanking residues extending from either end of the open groove. Crystallographic analyses of human and mouse class II molecules (20-22, 46, 52, 62) have shown that the peptide-binding site consists of four polymorphic pockets that accommodate specific amino acid residues. A proposed binding motif of I-E^d has been identified through peptide-binding studies (4, 19, 35, 53, 54, 63) and by pool sequencing analysis of naturally processed ligands (51). The most important peptide anchor residue is referred to as P1 (or i), with additional anchors at relative positions P4 (i + 3), P6 (i + 5), and P9 (i + 8). At the P1 anchor, nonpolar amino acids are more prevalent, with aromatic amino acids (W, Y, F) being preferred. The P4 and P6 binding pockets are more degenerate in their amino acid preferences, but in the anchor P9, the basic residues Lys and Arg are preferred. With alanine-substituted analogues, it is apparent that residues in the core sequence FEVGATYYF either are critical to MHC binding or serve as TCR contact residues. It is likely that F₂₉₅ serves as the P1 anchor residue, but the completion of direct class II binding studies is required to reach this conclusion. Furthermore, we cannot exclude the possibility that substitution analogues of the minimal epitope interact with the I-E^d molecule in different configurations, as has been reported in previous peptide-class II binding studies (30, 35, 37, 70) or, less likely, in a different peptide-binding register (2).

Assays to determine critical residues with alanine-substituted analogues also demonstrate that substitution of the peptide flanking residues in our minimal epitope does not affect stimulation of hybridoma T cells. Despite this, several other investigators have demonstrated that peptide flanking residues may, in fact, play a more critical role in determining T-cell recognition of some peptide epitopes than was previously thought (2, 12, 49, 68). Additionally, peptide flanking residues can contribute to higher-affinity binding to class II molecules (7, 38) or may effect cellular mechanisms involved in antigen processing (12, 16, 25, 26). Determining the fine specificity of additional OmpF-specific T-cell hybridomas or the use of substitution-containing peptides into which amino acids other than alanine were introduced could reveal T-cell recognition patterns of the OmpF minimal peptide that may involve additional residues.

A well-defined potential vaccine candidate containing analogues of a limited number of constant T-cell epitopes that are present in these conserved regions of the porin molecule could have broad specificity against enterobacterial disease. Such a vaccine could have a high level of flexibility if conserved sequences were coupled to other protective B- or T-cell epitopes and if it had the capacity to generate protective T-cell responses in a genetically diverse population. A subset of promiscuous peptides that are recognized in the context of multiple murine or human class II alleles has been described (10, 13, 18, 19, 27, 28, 32, 33, 43, 44, 60). Degenerate binding of peptides to multiple alleles of murine I-E, and its human homolog HLA-DR, has been attributed to the presence of a monomorphic α chain. However, degenerate binding to highly polymorphic class II alleles has also been reported (10, 27, 41, 55, 66). The promiscuous recognition of conserved OmpF peptide epitopes and the ability of sensitized T cells to recognize heterologous porin peptides are under investigation. A high degree of promiscuity of T- and B-cell responses with native porin proteins has been previously demonstrated (69). Our studies have permitted a better understanding of the immune response to the OmpF protein, the identification of epitopes that may be relevant to induction of these responses, and the structural characteristics of this protein that influence class II-peptide recognition by T cells.