Epitope designates an area of an antigen that interacts with a specific antibody. It has generally replaced the former name of antigenic determinant. By symmetry, the term paratope has been proposed for the antibody, instead of antibody combining site (also designated as antigen binding site), but is much less widely used. An extension of this terminology is to be found as idiotopes and allotopes that designate epitopes that define idiotypic and allotypic (see Allotype, alloantigen) specificities, respectively.
Epitopes are defined functionally by the interaction of the antigen with a particular antibody, so it does not represent a preformed antigenic structure in itself; instead, it is defined by the immune system. This means that identification of epitopes will depend upon the nature and the number of B-cell clones that will respond to the administration of antigen. It is therefore very likely that two individuals will have a different pattern of recognition of the same antigen, although large overlaps are the rule. A consequence of this is that the best tool to define an epitope is a monoclonal antibody, which constitutes a reliable and standardized reagent. The obvious requisite for a region of the antigen to be recognized as an epitope is its accessibility—that is, that it be in contact with the solvent, which is favored by the presence of hydrophilic and/or charged amino acids.
Defining the structural basis of an epitope on natural antigens, such as proteins, is an almost impossible task, because B cells, and hence antibodies, recognize epitopes directly as native structures of the antigen. This implies that most epitopes have a complex three-dimensional structural organization that is generally lost upon cleavage of the antigen. Limited proteolysis of a protein may retain some regions that still contain native epitopes, but it is quite obvious that isolation of a "pure" epitope is most often not feasible. Interesting attempts to isolate epitopes from natural proteins have nevertheless been made with some success. This is the case of myoglobin and lysozyme, for which a loop of 20 amino acids clamped by a disulfide bond has long been a prototype of an "isolated" epitope. Because of their repetitive structural organization, polysaccharides are somewhat better tools to derive epitopes. This was the case for dextran, a polymer of glucose that allowed Kabat to demonstrate that various oligosaccharides were competing with the binding of dextran to its antibodies and indicating that they could be considered epitopes of this antigen. In fact, because the optimal inhibition was obtained by a hexasaccharide, it was concluded that the size of the anti-dextran combining site was that required to accommodate a molecule having the size of this inhibitor. Most of our knowledge of the antigenicity of proteins has been gained thus far with synthetic polyamino acids that constitute simplified models of proteins, because they can have a rather monotonous organization; for example, three or four amino acids might be linked covalently to a backbone of poly-L-lysine, like those extensively studied by the group of Sela in Israel. These models favored, however, sequential epitopes, as opposed to the conformational ones that are commonly encountered in natural protein antigens. The use of haptens was of course only a first approach to define an antigenic determinant, and they gave information of interest regarding the exquisite specificity of antibody recognition.
Monoclonal antibodies (mAbs) have proven of great value in epitope mapping the surface of an antigen. Mapping is based on inhibition of one mAb by others. When complete and reciprocal inhibition is observed between two mAbs, it is concluded that both antibodies recognize the same epitope. Although not a definitive proof, it remains an acceptable conclusion. Partial inhibition is suggestive that the two mAbs recognize overlapping epitopes. Consequently, a sufficiently large collection of mAbs permits exhaustive mapping of the epitopes, which may cover the entire surface of the antigen. One must realize, however, two important points: (1) Some regions are more frequently recognized than others; these are the immunodominant epitopes; and (2) the number of mAbs that can bind simultaneously to a given antigen molecule is necessarily limited by the steric hindrance imposed by the surface of the antibody combining site that interacts with the epitope. Having said that, it is quite obvious that the overall number of epitopes that can be detected on one antigen molecule is directly related to its surface area.
By extension, the notion of T-cell epitopes has been defined. From a physiological point of view, it is somewhat more complicated in that it is linked directly to the mechanism of antigen presentation and thus requires the interaction of the antigenic structure with two partners, the multiple histocompatibility complex (MHC) molecules and the T-cell receptor (TCR) itself. Because antigen processing and presentation have been defined for only proteins thus far, such T-cell epitopes are peptides. In a sense, the structural basis for a T-cell epitope is simpler and more directly accessible than in the case of B cells, because these peptides are short and are recognized as a sequential determinant. Strictly speaking, however, the term epitope is reserved to that part of the peptide that interacts with the TCR, whereas regions that interact with the MHC molecule, which are of decisive importance, because it conditions immunogenicity, are described sometimes as the "agretope" Extensive analysis by site-directed mutagenesis has clearly demonstrated which key residues contributed the epitope or the agretope. Because of the differences in the structures of the cavity that bind the peptides in various MHC molecules, the size of the peptide is limited to 9 amino acid residues for those that bind to class I MHC molecules, whereas it may be somewhat larger for those that are presented by class II. A very large number of such small overlapping peptides may be derived from one protein antigen. The problem of recognition by the MHC molecules, which are in very limited number for any given individual (a few discrete molecules), is therefore very different from the huge repertoire expected from Ig and TCR molecules. What happens is that MHC molecules will "choose" only a very few peptides from one antigen, those that fit with the minimum requirements that drive the interaction with the agretope. This is why every individual will use a different way to respond to the same antigen. Using a simplified model of antigen, such as a monotonous synthetic polypeptide, the chance is high that the molecule will be immunogenic only in those individuals with the appropriate MHC molecules. This was the case of the strains of mice injected with synthetic polypeptides that did or did not respond, initiating the first approach to the understanding of the genetic control of the immune response and starting the elucidation of the physiological role of MHC molecules.
Both the B-cell and T-cell epitopes raise directly the problem of the repertoire of the immune system, a central notion that is directly related to the number of potential antibodies and TCR that a given organism can potentially make (see Repertoire).
