We compared the efficiency of two commonly used cellular major histocompatibility complex (MHC) class I peptide-binding assays to identify a cytotoxic T lymphocyte (CTL) epitope-containing peptide among length variants derived from the human papilloma virus type 16 (HPV 16) oncoprotein E7. Although both assays identified the same sequence (E7 49-57) as the most efficient Db-binding peptide, the efficiency by which they did so differed markedly. In a peptide competition cytotoxicity (PCC) assay, based on inhibition of CTL lysis by competition for binding to MHC class-I molecules between a known CTL epitope-containing peptide and peptide of interest, E7 49-57 bound 45-fold more efficiently to Db than the second Db-binding peptide in line. In the widely used RMA-S MHC class I peptide-binding assay, based on peptide-induced stabilization of 'empty' MHC class-I molecules at the surface of antigen-processing defective RMA-S cells, this difference was only 3 fold. Similar differences were observed when other Db-restricted CTL clones and CTL epitope-containing peptides were used in the PCC assay. The same phenomenon was observed when peptide binding affinities for H-2Kb were analyzed in both assays. We conclude that the PCC assay discriminates more efficiently between high- and low-affinity MHC class I binding peptides than the RMA-S assay. This observation is ascribed to the fact that peptide-MHC class I dissociation is an important parameter in the PCC but not the RMA-S assay.