An epitope is the portion of the surface of an antigen that binds to an antibody, or the peptide fragment of a protein antigen that binds to the T lymphocyte antigen receptor when presented by the cognate major histocompatibility protein. The best way to identify an antibody epitope is from a crystal structure of the antibody:antigen complex, where the contacts are evident. There are several servers that attempt to predict epitopes.
Antibody epitopes may also be called determinants, which is an historically earlier but equally good term. The term epitope implies that the determinant is on the surface of the antigen ("epi").
T Lymphocyte Epitopes
It is unfortunate that epitope has caught on as the term to describe the peptide fragments that T cells recognize, since these are not necessarily derived from the surfaces of protein antigens, but may be derived from portions that were buried in the folded protein. The terms cryptotope and unfoldon are almost never used, but are perhaps more descriptive.
Antibody epitopes can be made up of discontinuous portions of a protein antigen's sequence, or of a continuous portion. In contrast, T cell epitopes always represent a continuous fragment of the sequence of a protein antigen.
Antibody epitopes can occur on the surfaces of native folded proteins, or equally well on denatured conformations of proteins. Peptides are typically too short to have a well-defined fold, yet sometimes can simulate the epitope, binding to antibodies. T cell epitopes are always peptide fragments, and hence, represent a denatured (unfolded) form of the native protein.
Epitope Prediction Servers
Over a dozen servers that predict epitopes are available (Google Search for "epitope prediction server"). Only a few of these are listed below. Please help by adding more, with descriptions. It would also be very useful to have performance comparisons.
In the absence of the crystal structure of an antibody:antigen complex, a common way to identify the epitope recognized by a particular antibody is to display random peptides (for example, using phage display libraries), and then to identify the sequences of the peptides with the highest affinity for the antibody. These sequences can then be used to predict where the epitope lies on the native protein, taking into account that the epitope on the native protein may be discontinuous. In order to use this strategy, the 3D structure of the protein antigen must be known.
In alphabetical order:
ElliPro is a web-tool that implements a method for identifying continuous epitopes in the protein regions protruding from the protein's globular surface and, together with a residue clustering algorithm, the MODELLER program and the Jmol viewer, allows the prediction and visualization of antibody epitopes in a given protein sequence or structure. The methods are published.
EpiSearch requires as input the sequences of peptides that bind to the antibody in question, and an uploaded 3D model of the protein antigen. It offers an interactive 3D view of the epitopes, in Jmol, as well as a list of possible epitopes. No publication is cited.
The Epitopia Server predicts immunogenic regions in general. It will accept either a protein sequence, or a 3D protein structure. It "implements a machine learning scheme to rank individual amino acids in the protein, according to their potential of eliciting a humoral immune response". (Thus, it does not require a list of peptides that bind to an antibody of interest.) When a 3D model is submitted, it can be visualized in FirstGlance in Jmol colored by predicted immunogenicity. The methods are published.
The Pepitope Server predicts epitopes on the surface of a 3D protein antigen model, based on a list of peptides that bind to the antibody. The 3D model can be a PDB code or uploaded. Results can be visualized interactively in FirstGlance in Jmol. The algorithm is also available as a stand-alone program called PepSurf (with C++ source), licensed for non-commercial use by Tel-Aviv University. The methods are published.
References and Notes
- ↑ Ponomarenko J, Bui HH, Li W, Fusseder N, Bourne PE, Sette A, Peters B. ElliPro: a new structure-based tool for the prediction of antibody epitopes. BMC Bioinformatics. 2008 Dec 2;9:514. PMID:19055730 doi:10.1186/1471-2105-9-514
- ↑ Rubinstein ND, Mayrose I, Pupko T. A machine-learning approach for predicting B-cell epitopes. Mol Immunol. 2009 Feb;46(5):840-7. Epub 2008 Oct 22. PMID:18947876 doi:10.1016/j.molimm.2008.09.009
- ↑ Mayrose I, Penn O, Erez E, Rubinstein ND, Shlomi T, Freund NT, Bublil EM, Ruppin E, Sharan R, Gershoni JM, Martz E, Pupko T. Pepitope: epitope mapping from affinity-selected peptides. Bioinformatics. 2007 Dec 1;23(23):3244-6. Epub 2007 Oct 31. PMID:17977889 doi:10.1093/bioinformatics/btm493