Search

Human papillomavirus (HPV) is the most common viral infection of the reproductive tract, which can increase the risk of cervical cancer. Development of HPV-specific vaccines that can be recognized and presented by human leukocyte antigen (HLA) system is considered as an important approach to prevention and treatment of cervical cancer. In this study, a combination of bioinformatics analysis and peptide binding assay was described to identify potential cytotoxic T lymphocyte epitopes from the HPV genome. In the procedure, a structure-based quantitative structure–activity relationship (SB-QSAR) was used to analyze the intermolecular interactions between HLA proteins and peptide epitopes, and the obtained nonbonded descriptors were then correlated linearly and nonlinearly with HLA–peptide affinities by using partial least squares regression and support vector machine, respectively. Several SB-QSAR models were established and evaluated rigorously, from which the best predictor was applied to virtual screening against thousands of nonapeptide segments generated from the HPV-16 genome. With this protocol a number of peptide ligands were inferred to have high affinity for a variety of frequently observed HLA-A alleles, including 0101, 0201, 0202, 0203, 0206 and 0301, from which several promising candidates were selected and tested in vitro to determine their binding strength toward HLA proteins. Consequently, several peptides were identified as potent HLA binders with half-maximal binding BL50 at micromolar level. The molecular mechanism underlying the high-affinity binding of identified peptide ligands to HLA proteins was also investigated in detail.