Catalytic DNA sequences (deoxyribozymes, DNA enzymes, or DNAzymes) have been identified by in vitro selection for various catalytic activities. Expanding the limits of DNA catalysis is an important fundamental objective and may facilitate practical utility of catalysts that can be obtained from entirely unbiased (random) sequence populations. In this study, we show that DNA can catalyze Zn 2+ -dependent phosphomonoester hydrolysis of tyrosine and serine side chains (i.e., exhibit phosphatase activity). The best deoxyribozyme decreases the half-life for phosphoserine hydrolysis from as high as >10 10 y to <1 h. The phosphatase activity also occurs with nonpeptidic substrates but with reduced efficiency, indicating a preference for phosphopeptides. The newly identified deoxyribozymes can function with multiple turnover using free peptide substrates, have activity in the presence of human cell lysate or BSA, and catalyze dephosphorylation of a larger protein substrate, suggesting broader application of DNA catalysts as artificial phosphatases. D evelopment of catalysts is a major impetus for much of modern chemical research. Nature's biomolecular protein and RNA catalysts are responsible for a wide range of chemical reactions, and protein enzymes in particular can achieve large rate enhancements (1, 2). Although DNA catalysts are unknown in nature, in vitro selection [first pioneered for RNA (3)] is readily applied to identify catalytically active artificial DNA sequences (4-6). Importantly, DNA (and RNA) catalysts can be identified by starting with entirely random sequence pools, whereas directed evolution of proteins typically requires a known, catalytically active starting point (7,8). A growing range of chemical reactions has been shown to be catalyzed by DNA (4-6). For DNA phosphodiester hydrolysis, the uncatalyzed (spontaneous) half-life for P-O bond cleavage of ∼30 million y is reduced to as little as 0.5 min by a DNA catalyst (9, 10). However, in this reaction the DNA catalyst interacts with its DNA substrate by extensive Watson-Crick base pairing, and such an approach cannot be generalized to nonoligonucleotide substrates such as peptides and proteins. With the exception of the ribosome, the natural ribozymes catalyze RNA cleavage and ligation, and they generally have more modest rate enhancements (11-13), limited by the relatively high uncatalyzed half-lives of these reactions.DNA catalysts that covalently modify peptide and protein substrates are fundamentally interesting and likely have practical value, especially for biologically relevant chemical modifications.We have initiated studies into DNA-catalyzed modifications of amino acid side chains of peptide substrates, such as nucleopeptide linkage formation involving tyrosine and serine (14-16). A major challenge in such studies is to achieve catalysis even though the DNA catalyst cannot engage in any preprogrammable WatsonCrick binding interactions with the substrate. In this report, we show that DNA can catalyze Zn 2+ -dependent hydrolysis of tyrosine...