B-epitopes: when complexity matters
Specificity of antibody that will be used for research or diagnostics is of primary importance. Lately, most of antibodies are developed against a selected peptide epitope from the sequence of a protein. This raises the question: can the specificity of antibody prepared against a given epitope be predicted in advance, at the stage when the epitope has been selected?
We had recently an opportunity to perform an analysis for specificity of several antibodies developed by a third party. These were all prepared against peptide epitopes (9-12-mers) and purified on a peptide column from hyperimmune serum. In other words, each of them was as good as antibody against this epitope can be.
We have analysed the epitopes used for these peptides (and the entire sequences of the proteins they originated from) using EpiQuest-C, our software that defines the complexity (as we view it) profile of a protein primary sequence. Each possible 9-mer from the entire sequences and from peptide epitopes was assigned its calculated complexity index, CGI. As some peptides were longer than 9-mer, we have determined the highest and lowest value of their CGI.
Interestingly, all peptides could be divided into two groups according to their highest CGI value, Tier 1 and 2: the first was above the Mean value, the second -below the mean value (Fig.1). The selected peptide epitopes were not on average different from the whole population of 9-mers that could be selected from the given sequences.
Fig.1. CGI of 25 epitopes from 25 individual proteins were compared with all possible 9-mers within the whole sequences of the proteins. Mean CGI (blue line) of each group of 9-mer peptides were not different. The peptides used for epitopes could be separated into 2 groups according to their CGI (above and below Mean)
What was interesting, is that the reactivity of the antibodies in western blot was quite different between the Tier 1 and 2 antibodies with respect to specificity and background. Thus, practically all Tier 1 antibodies were excellent in blot (10 out of 12, 83%, with 2 having some non-specific reaction). In contrast, in Tier 2 from 13 antibodies 3 did not give any specific reaction in blot at all, 6 demonstrated additional reactivity to some other proteins or were excessively sticky to the membrane, giving the background. Only 4 antibodies (30%) showed good, clean reaction as did the majority of Abs from Tier 1 (Fig 2)
Fig.2. Shown are the highest and the lowest GCI for 9-mers within each epitope, as the epitopes may be up to 15 aa long (orange boxes). Blots for every affinity purified antibody raised against every epitope is shown below for each epitope used. (+) or (-): flawless antibodies or antibodies showing some non-specific cross-reactivity or background. Black box indicated antibodies that gave no signal that could be considered as specific.
This data is more illustrative than conclusive due to the small size of selection, although if to divide the antibodies roughly into 3 groups (no-background at all, slight background and only non-specific) and compare the CGI of epitopes between the groups, it is possible to claim with certain confidence (Fig. 3, p<0.05) that high GCI of epitope assures antibody specificity in a solid phase assay (like W. Blot)
Fig.3. The 25 peptide epitopes used for making specific antibodies were divided into 3 groups according to the respective antibody reactivity in the western blot assay: w/o specific signal, with some noise, good. The CGI, a complexity index, as defined by EpiQuest-C is significantly different for the "good" antibodies group from the other two groups.