Differing processing requirements of four recombinant antigens containing a single defined T-cell epitope for presentation by major histocompatibility complex class II.

A set of predictive rules governing the likelihood of generating a particular peptide-major histocompatibility complex (MHC) class II complex from an intact antigen has not been fully elucidated. We investigated the influence of positional and structural constraints in the region of the epitope by designing a set of recombinant antigens that each contained the well-characterized T-cell epitope moth cytochrome c (MCC) (88-103), which is specifically recognized by the monoclonal antibody (mAb) D4 when complexed with H-2Ek. Our model antigens contained MCC(88-103) either peripherally, at or towards the C-terminus, or internally. Their abilities to bind directly to soluble H-2Ek, and the extent of D4 epitope formation from them by antigen processing-competent and -incompetent cell lines, were determined. Here we report that three of these four antigens yielded MCC(88-103)/H-2Ek complexes independently of the conventional MHC class II antigen-processing and presentation pathway, and in each case the epitope was carried peripherally; two bound directly as intact proteins, probably as a result of spatial separation of the epitope from the major globular domain, and one was processed to peptide by a cell-surface protease. One protein, which carried the epitope inserted into an internal loop, acted as a conventional processing-dependent MCC(88-103) delivery vehicle. Thus, this epitope has different presentation requirements depending on its context. These antigens constitute a panel whose framework could be modified to further define predictive rules for antigen processing for presentation through the different MHC class II complex-generating pathways.

Determination of the relationship between T cell responsiveness and the number of MHC-peptide complexes using specific monoclonal antibodies.

We describe the generation of three mAbs that recognize the complex of the class II MHC molecule IEk bound to a peptide derived from the carboxyl terminus of moth cytochrome c (residues 95-103). Reactivities of these mAbs are sensitive to single alterations in the sequence of both helices of the MHC molecule and to the bound peptide. The epitopes of these reagents are distinct but overlap substantially. One of these mAbs specifically blocks lymphokine release by T cells responsive to this complex but not others. We have used another to examine how the number of complexes on an APC is related to its ability to stimulate T cells. We find that 200-400 complexes per cell are necessary and sufficient to induce a degree of stimulation, whereas maximum stimulation is achieved only if more than 5000 complexes are present. The analysis indicates that T cell activation is a stochastic process.

A T cell receptor-specific blockade of positive selection.

To investigate the influence of endogenous peptides on the developmental processes that occur during thymocyte selection, we have used monoclonal antibodies that preferentially recognize the major histocompatibility complex (MHC) molecule I-Ek when it is bound to the moth cytochrome c peptide (88-103). One of these antibodies (G35) specifically blocks the positive selection of transgenic thymocytes expressing a T cell receptor that is reactive to this peptide- MHC complex. Furthermore, G35 does not block the differentiation of transgenic T cells bearing receptors for a different I-Ek-peptide complex. This antibody recognizes a subset of endogenous I-Ek-peptide complexes found on a significant fraction of thymic antigen-presenting cells, including cortical and medullary epithelial cells. The sensitivity of G35 to minor alterations in peptide sequence suggests that the thymic peptide-MHC complexes that mediate the positive selection of a particular class II MHC-restricted thymocyte are structurally related to the complexes that can activate it in the periphery.

Mechanisms of T cell epitope immunodominance analyzed in murine listeriosis.

We demonstrate conclusively that the bacterial exotoxin listeriolysin O (LLO) is a target Ag for eliciting CD4+ T cell responses following infection with Listeria monocytogenes. The minimal I-Ek-restricted immunodominant CD4+ T cell epitope was identified as peptide 215-226 (p215-226). Most LLO-specific T cell hybridomas recognized p203-226, p208-226, p215-226, and p215-234, although each exhibited a characteristic pattern of preferential reactivity. One hybridoma (IIIC5) reacted to p203-226 but not to p208-226 or any other LLO peptide tested. With APCs from B10 congenic mice and cells transfected with either I-Ak or I-Ek, IIIC5 recognized p203-216 with I-Ak, while a different hybridoma (IB5) recognized p215-226 with I-Ek. Competitive binding studies demonstrated that of 15 LLO peptides tested, only p203-226, p215-226, and p215-234 had high affinity for I-Ek, while p203-226 could also bind to isolated I-Ak. Of nine LLO peptides tested, only p215-234 bound multiple class II MHC alleles. These findings suggest that the immunodominance of p203-226 may be due in part to the presence of multiple T cell epitopes with I-Ek- and I-Ak-binding capability. Many of the rules of immunodominance observed with model Ags are also operative in our murine model of bacterial infectious disease. Furthermore, a novel mechanism of immunodominance based on newly defined structural features of MHC molecules is implicated. This information is crucial for rational vaccine development.

Kinetics of T-cell receptor binding to peptide/I-Ek complexes: correlation of the dissociation rate with T-cell responsiveness.

Recognition by T-cell antigen receptors (TCRs) of processed peptides bound to major histocompatibility complex (MHC) molecules is required for the initiation of most T-lymphocyte responses. Despite the availability of soluble forms of TCRs and MHC heterodimers, this interaction has proven difficult to study directly due to the very low affinity. We report here on the kinetics of TCR binding to peptide/MHC complexes in a cell-free system using surface plasmon resonance. The apparent association rates for the interactions of related peptide/MHC complexes to one such TCR are relatively slow (900-3000 M-1.s-1) and dissociation rates are very fast (0.3-0.06 s-1) with t1/2 of 2-12 s at 25 degrees C. The calculated affinity of the engineered soluble molecules compares well with previously reported competition data for native TCRs or competition data reported here for native peptide/MHC complexes, indicating that these soluble heterodimers bind in the same manner as the original molecules expressed on cells. We also find that the peptide variants which give weaker T-cell stimulatory responses have similar affinities but distinctly faster dissociation rates compared with the original peptide (when loaded onto the MHC molecule) and that this later property may be responsible for their lower activity. This has implications for both downstream signaling events and models of TCR-peptide antagonists.

Use of global amino acid replacements to define the requirements for MHC binding and T cell recognition of moth cytochrome c (93-103).

Substitution with all naturally occurring L-amino acids at each of 11 residues of the IEk-restricted month cytochrome c (93-103) epitope has allowed us to analyze the requirements for MHC binding and T cell recognition to a level of definition not previously possible. Substitutions at only three positions systematically affect MHC binding and three others appear to be the major TCR contacts. Interestingly, changing residues involved in MHC binding can ablate T cell recognition without altering MHC association. Additionally, residue identity at two positions that do not appear critical for MHC binding, nor to be involved in specific T cell contact, nonetheless dramatically affect T cell responses. This suggests that peptides differing only slightly in sequence can have significantly altered conformations within the class II MHC binding groove. We have also developed a simple scoring program that uses the binding data to quantitate how well a given peptide fits the MCC motif. All strongly immunogenic IEk-restricted epitopes score highly (> or = 0.70, where 1.0 is perfect concordance), and only 3% of all potential nonameric peptides in the two main protein sequence databases have scores greater than 0.70. This indicates that the global amino acid replacement approach using a single peptide is an efficient means of deriving binding motifs for a given class II MHC molecule, and should aid in the identification of novel T cell epitopes.

Formation of functional peptide complexes of class II major histocompatibility complex proteins from subunits produced in Escherichia coli.

Class II major histocompatibility complex molecules play a major role in the immune response by binding peptide fragments of exogenous antigens and displaying them on the surfaces of antigen-presenting cells, where they can be recognized by T cells. To facilitate structural and functional studies of these molecules, we have produced truncated alpha and beta chains of the murine class II molecule I-Ek in Escherichia coli (Ec-I-Ek) and have developed conditions to fold them in the presence of specific peptides with yields of complex approaching 2%. Reconstitution is specific since only unlabeled peptide known to bind I-Ek compete with biotinylated peptide, as assessed by ELISA. Complexes of the refolded heterodimer (Ec-I-Ek) with either of two different peptide antigens remain associated during nonreducing SDS/PAGE. Immobilized Ec-I-Ek-peptide complexes stimulate lymphokine production by three T-cell clones in an antigen-specific manner with a dose-response relation comparable to previously described soluble I-Ek molecules produced in CHO cells. These results demonstrate that folding of Ek alpha and Ek beta polypeptides does not require any other protein to produce the biologically relevant conformation and that carbohydrate modification of this class II molecule is not necessary for alpha beta T-cell recognition.

pH affects both the mechanism and the specificity of peptide binding to a class II major histocompatibility complex molecule.

We have compared the contribution of electrostatic forces in the binding of antigenic peptides to the class II MHC molecule, IEk, at weakly acidic (pH 5.4) and neutral (pH 7.5) pH values. The binding of specific moth cytochrome c (MCC) and hemoglobin (Hb) peptides to IEk is very sensitive to ionic strength at pH 7.5 but not at pH 5.4, indicating that the mechanism of peptide binding is pH-dependent. Substitution of the C-terminal Lys in MCC for an Ala residue selectively destroyed peptide binding at neutral pH and increased the dissociation rate at least 30-fold, implicating this residue in the pH-dependent electrostatic interaction. The presence of a C-terminal Lys in many of the peptides that are restricted to IEk suggests that this electrostatic interaction is widely used to bind peptides to this MHC molecule. We also probed the electrostatic environment of the peptide binding groove adjacent to the N-terminus of the bound peptide by rapid-diffusion fluorescence energy transfer using a terbium-labeled MCC peptide. In this region of the peptide binding groove, more negative charge is present at pH 7.5 than at pH 5.4. These findings indicate the importance of MHC carboxylates to the mechanism and specificity of peptide binding. The biological importance of having two distinct mechanisms of peptide binding at different pH may be that it acts to broaden the spectrum of antigenic peptides that can be presented to T-cells.

Two-dimensional nuclear magnetic resonance analysis of a labeled peptide bound to a class II major histocompatibility complex molecule.

The formation of peptide/major histocompatibility complex (MHC) complexes and their subsequent recognition by T cells is a pivotal event in the initiation of an immune response. While X-ray crystal structures are now available for class I MHC/peptide complexes, little detailed structural information is known about the class II MHC equivalent, and there are no solution structure data for either. A 16 amino acid residue moth cytochrome c peptide (residues 88 to 103) was 13C-labeled for two-dimensional isotope-edited NMR analysis. The peptide was labeled either selectively in the methyl groups of alanine residues or uniformly at every carbon position, and bound to unlabeled soluble mouse I-Ek class II MHC molecules. Although alpha-helical in the native cytochrome c protein and with no uniform structure in solution, the peptide is bound to the I-Ek molecule with the alpha-carbon atoms of the 11 C-terminal residues held in the binding groove. This indicates that the class II MHC peptide binding site is somewhat larger than that of class I MHC molecules (> or = 11 amino acid residues versus 8 to 10 amino acid residues), consistent with recent data on eluted peptides. Despite the large size of the complex (approximately 70 kDa), nuclear Overhauser effects are clearly detectable between peptide side-chains and the MHC molecule. Indications of the buried or exposed nature of particular side-chains within the bound peptide are derived from the NMR data and these are used together with information from previous biological studies to propose a crude model of the interaction of the peptide with the groove of the MHC molecule. We find no evidence for a conformational change in the peptide/MHC complex in the spectra at pH 5.0 versus pH 7.0, despite a 40-fold faster on-rate for the peptide at the lower pH value.

pH dependence and exchange of high and low responder peptides binding to a class II MHC molecule.

We have compared the binding kinetics of two antigenic peptides to a soluble class II MHC molecule. One of the peptides provokes a strong T cell response and the other a much weaker one. Both show greatly increased (approximately 40-fold) association rates at pH 5 in comparison to neutral pH, consistent with the low pH environment of late endosomes being most conducive to class II MHC--peptide binding. Interestingly, the weak peptide has a much faster off-rate that is significantly increased at pH 5 and it can be entirely replaced in an exchange reaction by the stronger one. This suggests that one characteristic of immunodominant peptides is that of nearly irreversible binding, such that they will be strongly selected for in the course of class II MHC transit and recycling through endosomal compartments. Modelling the parameters of this peptide exchange also suggests that a large fraction of the GPI-chimeric MHC molecules used in this study are 'empty' with respect to endogenous peptides, or else occupied with extremely weak ones, consistent with their inability to load processed peptides intracellularly.

Mapping T-cell receptor-peptide contacts by variant peptide immunization of single-chain transgenics.

To test models of T-cell recognition, mice transgenic for T-cell receptor alpha or beta chain have been immunized with variant peptides that force changes in the resulting T-cell response. In particular, charge substitutions on the peptide often elicit reciprocal charges in the junctional (CDR3) sequences of T-cell receptor V alpha or V beta chains, indicating direct T-cell receptor-peptide contact, and allowing derivation of a topology for the T-cell receptor-MHC interaction. At one position on the peptide, variants transformed a homogeneous V beta response into a very heterogeneous one.

Molecular components of T-cell recognition.

We review recent data that increase our understanding of the ternary complex of the T cell receptor (TCR), antigenic peptides, and molecules of the major histocompatibility complex (MHC). Studies using synthetic peptide analogs for T-cell antigens have identified peptide residues that appear to interact with the MHC molecule and/or the TCR. The logical extension of these studies, using a complete replacement set of peptide analogues for a model peptide antigen, has more precisely defined the biochemical character of putative MHC and TCR contact residues, and indicated that the TCR is highly sensitive to subtle changes in peptide conformation. Insight into the binding site for peptide on the TCR has recently come from variant peptide immunization of TCR single-chain transgenic mice. These experiments indicate that residues encoded by the V(D)J junctions of both TCR chains contact peptide directly. TCR-MHC contacts have also been studied, using in vitro-mutagenized MHC molecules, particularly those altered at residues predicted to point "up," toward the TCR. These studies reveal that TCR-MHC contacts appear to be quite flexible, and vary between even closely related TCRs. A measure of the affinity of TCR for peptide/MHC complexes has come from competition experiments using soluble MHC complexed with specific peptides. This affinity, with a KD of 5 x 10(-5) M, is several orders of magnitude lower than that of most antibodies for their protein antigens and suggests that the sequence of events leading to T-cell activation begins with antigen-independent adhesion.

Low affinity interaction of peptide-MHC complexes with T cell receptors.

The interaction of antigen-specific T cell receptors (TCRs) with their ligands, peptides bound to molecules of the major histocompatibility complex (MHC), is central to most immune responses, yet little is known about its chemical characteristics. The binding to T cells of a labeled monoclonal antibody to the TCR was inhibited by soluble class II MHC heterodimers complexed to different peptides. Inhibition was both peptide- and TCR-specific and of low affinity, with a KD = 4 x 10(-5) to 6 x 10(-5) M, orders of magnitude weaker than comparable antibody-antigen interactions. This finding is consistent with the scanning nature of T cell recognition and suggests that antigen-independent adhesion precedes TCR engagement.

Expression of a class II major histocompatibility complex (MHC) heterodimer in a lipid-linked form with enhanced peptide/soluble MHC complex formation at low pH.

A murine class II major histocompatibility complex (MHC) heterodimer, Ek, expressed as a glycan-phosphatidyl inositol-anchored chimera on Chinese Hamster Ovary cells, can present peptides, but not processed antigen to T cells. This chimeric MHC requires a 100-times higher peptide concentration to achieve a two- to four-times lower level of T cell stimulation. Cleavage with phosphatidylinositol-specific phospholipase C and purification result in large quantities of heterodimer in a water-soluble form. Plates coated with this material and then incubated with peptide can efficiently stimulate the appropriate T cell hybridomas. This stimulation is significantly enhanced when peptides are preincubated with the plate-bound MHC molecules in a pH range (5.0-5.5) similar to that of late endosomes. More than half of the soluble Ek molecules can form a specific complex with cytochrome c peptides in this pH range. This suggests that class II MHC molecules undergo distinct conformational changes in endosomal compartments that render them more capable of forming functional complexes with peptide antigens, irrespective of other cell components.

Recognition of the PB1, neuraminidase, and matrix proteins of influenza virus A/NT/60/68 by cytotoxic T lymphocytes.

We have investigated the recognition of the PB1, neuraminidase, and matrix (M1) proteins of influenza virus A/NT/60/68 (H3N2 subtype) by secondary in vitro stimulated polyclonal cytotoxic T lymphocyte (CTL) populations. While these three proteins have different functions and cellular locations, they can all be recognized as target antigens. However, the immunogenicity of these proteins for CTLs is under strict genetic control. Thus, PB1 protein is recognized as a cross-reactive target antigen by CTLs raised in CBA (H-2k) but not BALB/c (H-2d) mice. CBA, but not BALB/c mice, also generate a low-level CTL response to the neuraminidase. This latter response was only detectable following in vivo priming of CBA mice with a recombinant vaccinia virus expressing neuraminidase (N2-VACC). The matrix protein, expressed from recombinant vaccinia virus M-VACC, was not recognized as an antigen by CTL generated from either CBA or BALB/c strains of mice. By contrast, human HLA-A2-restricted influenza virus-specific CTLs were shown to recognize this matrix protein as a target antigen. Endogenous expression of as little as 90 amino acids of the matrix protein was sufficient to render target cells susceptible to lysis by such CTLs.

A region of the influenza A/NT/60/68 PB2 protein containing an antigenic determinant recognized by murine H-2Dd restricted cytotoxic T lymphocytes.

We have used a recombinant vaccinia virus to investigate the recognition of the PB2 protein of influenza A/NT/60/68 (H3N2) by murine polyclonal CTL populations. PB2 is recognized as a major cross-reactive target antigen. Recognition of PB2 is under strict genetic control, since BALB/c (H-2d) but not CBA (H-2k) mice are responders. We also demonstrate, by use of cell lines transfected with individual genes encoding class I molecules of the H-2d haplotype, that recognition of PB2 occurs in conjunction with the H-2Dd but not the H-2Kd or H-2Ld molecules. In contrast, recognition of the nucleoprotein of A/PR/8/34 by BALB/c-derived polyclonal CTL is restricted via the H-2Kd molecule. By using three recombinant vaccinia viruses expressing deleted forms of the PB2 protein we show that at least one epitope of the PB2 protein resides within the amino-terminal 256 amino acids. This approach offers an effective method to map the regions of large proteins containing epitopes recognized by CTL.

Protection against lethal influenza with neuraminidase.

The role of the neuraminidase in eliciting protection against a lethal influenza A virus [A/Ck/Penn/1370/83 (H5N2)] infection was investigated in chickens. Isolated N2 neuraminidase administered in adjuvant did not prevent infection but did prevent systemic spread of virus and death of chickens. N2 expressed in a recombinant vaccinia virus protected chickens when administered in adjuvant but was less effective when allowed to replicate and produce pox on the chicken's comb. Chickens vaccinated with isolated N2 in adjuvant or with inactivated H5N2 influenza virus were protected from clinical signs and death after challenge with A/Ck/Penn/1370/83 influenza virus. However, these animals were completely susceptible and died of infection with a heterologous subtype (H7N7) of influenza virus. The role of the different antigenic determinants of the N2 NA was investigated in chickens by passive transfer of monoclonal antibodies. Antibodies to antigenic determinants rimming the enzyme active center reduced disease signs in approximately half of the birds but did not significantly reduce virus levels. Antibodies to one of the two independent antigenic determinants that are distant from the enzyme active center were most effective at reducing virus replication and disease signs. This is surprising because antigenic variants could not be selected in vitro with these antibodies and suggests that they may facilitate clearance of virus. Antibodies to the other determinant that is located distally to the enzyme site were ineffective at providing protection.

Nuclear location of all three influenza polymerase proteins and a nuclear signal in polymerase PB2.

In order to re-examine the sub-cellular location of the three influenza A/NT/60/68 polymerase proteins PB1, PB2 and PA in infected cells, specific antisera for each polymerase component have been prepared by immunizing rabbits with polymerase-beta-galactosidase fusion proteins synthesized in Escherichia coli. We show that polymerase PB1, PB2, and PA are predominantly associated with the nucleus of influenza-infected MDCK cells by immunocytochemical techniques. In the case of polymerase PB2 we investigate the possibility that nuclear accumulation is an intrinsic property of the PB2 protein. Using a vaccinia-PB2 recombinant virus, we show that PB2 accumulates intra-nuclearly in monkey CV-1 cells in the absence of any other influenza protein, suggesting it contains an intrinsic nuclear signal.