摘要：Single-molecule discrimination among amino acids is crucial to the realization of next-generation protein sequencing. Owing to the heterogeneous charge and subtle volume difference of underivatized amino acids, it remains a challenge for single-molecule techniques to recognize each of them. Here, we report the direct detection of twenty proteinogenic amino acids using a copper(II)-functionalized MspA nanopore. The binding sites for copper(II) ion are constructed by introducing histidine mutation (N91H) to M2MspA protein. With copper ion binding to histidine residues, amino acids can reversibly coordinate the copper-histidine complex, generating well-defined current signals. Using this strategy, all twenty amino acids can be detected. Assisted by a machine learning algorithm, we can identify 100% of signals with 70.2% accuracy or 60% of signals with 93.4% accuracy in the validation set. In successively addition experiment, each amino acid in a mixture of 10 amino acids can be identified precisely. Furthermore, we use carboxypeptidase A1 to partly release the C-terminal amino acids of peptides with different lengths (9, 10 and 22 residues). The hydrolysates of peptides can be identified and distinguished. These results demonstrate the feasibility of this system for amino acids detection and peptide identification, shedding new lights on the development of single-molecule protein sequencing.