Иммунные исследования
Открытый доступ
ISSN: 1745-7580
ISSN: 1745-7580
Jing Sun, Tianlei Xu, Shuning Wang, Guoqing Li, Di Wu, Zhiwei Cao
Background
As an essential step of adaptive immune response, the recognition between antigen and antibody triggers a series of self-protection mechanisms. Therefore, the prediction of antibody-binding sites (B-cell epitope) for protein antigens is an important field in immunology research. The performance of current prediction methods is far from satisfying, especially for conformational epitope prediction. Here a multi-perspective analysis was carried on with a comprehensive B-cell conformational epitope dataset, which contains 161 immunoglobulin complex structures collected from PDB, corresponding to 166 unique computationally defined epitopes. These conformational epitopes were described with parameters from different perspectives, including characteristics of epitope itself, comparison to non-epitope surface areas, and interaction pattern with antibody.
Results
According to the analysis results, B-cell conformational epitopes were relatively constant both in the number of composing residues and the accessible surface area. Though composed of spatially clustering residues, there were sequentially linear segments exist in these epitopes. Besides, statistical differences were found between epitope and non-epitope surface residues with parameters in residual and structural levels. Compared to non-epitope surface residues, epitope ones were more accessible. Amino acid enrichment and preference for specific types of residue-pair set on epitope areas have also been observed. Several amino acid properties from AAindex have been proven to distinguish epitope residues from non-epitope surface ones. Additionally, epitope residues tended to be less conservative under the environmental pressure. Measured by topological parameters, epitope residues were surrounded with fewer residues but in a more compact way. The occurrences of residue-pair sets between epitope and paratope also showed some patterns.
Conclusions
Results indicate that, certain rules do exist in conformational epitopes in terms of size and sequential continuity. Statistical differences have been found between epitope and non-epitope surface residues in residual and structural levels. Such differences indicate the existence of distinctiveness for conformation epitopes. On the other hand, there was no accordant estimation for higher or lower values derived from any parameter for epitope residues compared with non-epitope surface residues. This observation further confirms the complicacy of characteristics for conformational epitope. Under such circumstance, it will be a more effective and accurate approach to combine several parameters to predict the conformation epitope. Finding conformational epitopes and analysing their properties is an important step to identify internal formation mechanism of conformational epitopes and this study will help future development of new prediction tools.