DNA-catalyzed reductive amination.

Deoxyribozymes are particular DNA sequences that have catalytic ability, analogous to protein enzymes as functional amino acid sequences. The discovery of natural RNA enzymes (ribozymes) spurred the search for artificial deoxyribozymes, which are identified by in vitro selection. Most deoxyribozyme-catalyzed reactions involve phosphodiester bond cleavage or ligation, although other reactions such as Diels–Alder reaction and thymine dimer photoreversion have also been reported. We have initiated a comprehensive effort to identify deoxyribozymes that catalyze reactions of amino acid side chains, initially focusing on small peptide substrates and with the longer-term goal of covalently modifying large proteins. Towards this goal, as part of a recent study we examined the ability of several new deoxyribozymes to modify tyrosine (or serine) side chains of tripeptide substrates that are not covalently tethered to the deoxyribozyme, although the tether was required to enable the selection process. Deoxyribozymes identified using tethered peptide substrates were also active with untethered peptide substrates, but with substantially lower rate and yield. Therefore, in new experiments, here we performed in vitro selection directly using untethered peptide substrates. This effort required a modified selection procedure that was intended to involve reductive amination as a step merely to “capture” catalytically active DNA sequences. Surprisingly, this selection process provided deoxyribozymes that operate entirely independently of the tripeptide substrate and instead catalyze a Ni-dependent reductive amination involving the N-amine of a guanosine nucleobase on an RNA substrate, which reacts with an oligonucleotide-dialdehyde substrate. Reductive amination is a key biochemical process, for example, for amino acid biosynthesis and catabolism using amino acid dehydrogenases (oxidases) and transaminases (aminotransferases). In vitro, enzymatic reductive amination is important in both laboratory-scale and industrial-scale organic synthesis. Reductive amination may also have been important in RNA World scenarios. Our unexpected discovery of DNAcatalyzed reductive amination suggests further exploration of the abilities of nucleic acid enzymes to catalyze this interesting and potentially useful class of reaction. For any in vitro selection strategy that does not involve physical compartmentalization or segregation of individual candidate sequences, the selection design must include a mechanism by which individual, functional sequences become separable from the vast excess of nonfunctional sequences. Importantly, this mechanism must operate in parallel fashion without requiring independent interrogation (i.e., screening) of each candidate sequence, which is impractical for the typical populations of > 10 sequences. Although many in vitro selections have relied upon binding of functional sequences to a solid support (bead), typically through biotin– streptavidin interactions, our efforts have generally relied upon polyacrylamide gel electrophoresis (PAGE) shift. A key advantage of the gel-shift approach is that functional sequences are separated largely on the basis of oligonucleotide length (size), which has increased or decreased substantially upon successful reaction. In contrast, bead-based methods are blind to the length of binding sequences and are therefore subject to artifacts that are avoided in the gelshift approach. Shown in Figure 1a is the particular selection strategy used in our recent report for identifying deoxyribozymes that covalently modify the tyrosine (or serine) side chain of an oligonucleotide-anchored, tethered tripeptide substrate. The resulting deoxyribozymes function by catalyzing Tyr (or Ser) side chain reaction with a 5’-triphosphorylated RNA substrate, leading to a Tyr-RNA (or Ser-RNA) nucleopeptide linkage. The tether is mandatory to enable the selection process; functional DNA sequences append the oligonucleotide anchor to themselves upon covalently modifying the anchored tripeptide, thereby leading to a gel shift that enables their separation. The new deoxyribozymes were subsequently evaluated with untethered substrates lacking the oligonucleotide anchor. In all cases greatly reduced rate and yield were observed, indicating that these deoxyribozymes depend strongly on the oligonucleotide anchor of the tripeptide substrate for optimal function. Omitting the oligonucleotide anchor and directly using an untethered peptide during the selection process would not induce a substantial gel shift for catalytically active DNA sequences, because the small peptide mass is insufficient in this regard. Therefore, seeking to retain gel shift as the physical basis of selection, here we designed the two-stage strategy shown in Figure 1b. First, an untethered peptide was used during the selection step, which—in accord with our recent report—included both Mg and Mn cations as available catalytic cofactors (50 mm HEPES, pH 7.5, 40 mm MgCl2, 20 mm MnCl2, 150 mm NaCl, 2 mm KCl, 37 8C, 14 h). Second, in the “capture” step, any deoxyribozymes that [*] O. Y. Wong, A. E. Mulcrone, Prof. S. K. Silverman Department of Chemistry University of Illinois at Urbana-Champaign 600 South Mathews Avenue, Urbana, IL 61801 (USA) E-mail: [email protected] Homepage: http://www.scs.illinois.edu/silverman/