DNA and RNA can be equally efficient catalysts for carbon-carbon bond formation.

Catalysis by nucleic acids was merely a theoretical possibility until the discovery of catalytic RNAs (ribozymes) in the early 1980s.1 Although a variety of natural ribozymes have since been identified,2 analogous catalytic DNAs (deoxyribozymes) have not been found in nature. In the laboratory, many artificial ribozymes and deoxyribozymes have been identified through in vitro selection by starting with pools of random sequences.3 The repertoire of artificial ribozymes discovered in this fashion encompasses many chemical reactions including phosphodiester cleavage and ligation,4 RNA polymerization,5 redox reactions,6 carbon-carbon bond formation (Diels-Alder reaction),7,8 and many others.9 Because DNA catalysts were identified later than RNA10 and because natural ribozymes provide a strong motivation to study artificial RNA catalysts, the catalytic abilities of DNA have not been examined as thoroughly as those for RNA.11 An early speculation was that the lack of 2′-hydroxyl groups in DNA would likely impair its catalytic efficiency relative to RNA,12 providing a specific concern about the functional range of DNA as a catalyst. The available data for the most commonly studied DNA-catalyzed reaction, RNA cleavage,13 suggest that RNA and DNA should be equally competent, although in both cases the highest theoretically possible rate enhancements have likely not been achieved.14 In this report we investigated DNA catalysis of the Diels-Alder reaction, anticipating that the results would allow a clear comparison between the catalytic efficiencies of artificial ribozymes and deoxyribozymes for this important carbon-carbon bond-forming reaction. We identified deoxyribozymes that can catalyze the Diels-Alder reaction as efficiently as the reported ribozymes, providing evidence that DNA can be as catalytically efficient as RNA for C-C bond formation.15 We began by considering the Diels-Alder ribozyme that was identified by Jäschke and co-workers using in vitro selection.8 This ribozyme catalyzes the bimolecular Diels-Alder reaction between suitably functionalized anthracene and maleimide substrates with multiple turnover. The structural basis for catalysis has been elucidated through X-ray crystallography,16 and the scope of substrate tolerance has been explored.17 One minimal form of this Diels-Alder ribozyme, 39M49, has 49 nucleotides.8 With 39M49 in mind, we arranged two parallel deoxyribozyme selection experiments. In the first selection experiment, designated “DAR” for “Diels-Alder Random”, we used an entirely random 40nucleotide (N40) sequence pool. In the second selection experiment, designated “DAB” for “Diels-Alder Biased”, we used a biased (i.e., partially randomized) pool that was derived from the 39M49 ribozyme sequence but synthesized as DNA, with 36 nucleotides of the sequence partially randomized. Each of these 36 nucleotides had 70% probability of having the original nucleotide identity and 10% probability each of having the other three possible identities. On this basis, the mean number of nucleotide differences between the 39M49 ribozyme and an arbitrary DNA sequence from the biased pool is ca. 11 nucleotides (30% of 36 ) 10.8), although a wide range of mutations per sequence is statistically represented. Separately using the DAR and DAB pools, we performed in vitro selection as illustrated in Figure 1.18 The selection process was initiated by primer extension on a DNA template using Taq polymerase and a DNA oligonucleotide primer with anthracene attached at the 5′-end via a hexaethylene glycol (HEG) tether. Each selection round consisted of three iterated steps: (1) Incubation with DTME (dithiobismaleimidoethane) to allow the Diels-Alder reaction to proceed [the key selection step]; (2) Treatment with a 5′-thiol-DNA to attack the unreacted maleimide moiety of DTME, followed by PAGE separation of the extended DNA strands; and (3) PCR amplification to regenerate the anthracene-tethered DNA pool, now enriched in catalytically active sequences. In each round, the key selection step used incubation conditions of 100 μM DTME in 50 mM Tris, pH 7.5, 200 mM Na+, and 100 mM K+ with 20 mM each Mg2+ and Ca2+ along with 5 μM each Mn2+, Co2+, Cu2+, and Zn2+ (all as Clsalts) at 30 °C for 1 h; these were the same ion concentrations used during the original ribozyme selection.8 After 10 selection rounds, robust activities of 49% (DAR) and 33% (DAB) were observed.18 After two additional rounds using only a 5-min incubation during the Diels-Alder selection step (leading to DAR and DAB activities of 18% and 13%), both round 12 pools were cloned, and individual deoxyribozyme clones were screened for catalytic activity. Many active sequences were found in both selection pools.18 One particular clone, DAB22, also showed catalytic activity when tested in trans using the anthracene-HEG small-molecule substrate that was not covalently tethered to DNA (i.e., Anthr-HEG). DAB22 had 13 mutations relative to the parent 39M49 sequence, but its mfold-predicted secondary structure revealed no apparent relationship to the parent ribozyme.18 Therefore, although DAB22 originated from the biased pool, it is a new catalytic sequence.19 The enantioselectivity of DAB22 was not Figure 1. Strategy for in vitro selection of deoxyribozymes that catalyze the Diels-Alder reaction. Published on Web 02/14/2008