Chemical replication of nucleic acids

Regiospecificity of chemical RNA copying

During the nonenzymatic copying of an RNA template by primer extension, either the 2′ or the 3′ hydroxyl of the primer can attack the phosphate of the incoming activated monomer, generating either a 2′-5′- or a 3′-5′-linkage. The resulting complementary strand will therefore contain a mixture of 2′-5′- and 3′-5′-linkages. Such backbone heterogeneity was thought to disrupt the folding—and hence the function—of RNA molecules. Our studies challenged this view and showed that even in the presence of substantial backbone heterogeneity, RNA aptamers and ribozymes can retain their molecular recognition and catalytic functions. Our high-resolution crystal structures rationalized this observation at the atomic level: RNA duplexes can buffer local structural changes caused by 2′-5′-linkages, resulting in a minimally altered global structure. Furthermore, 2′-5′-linkages reduce the energetic barriers of the ribose pseudorotation cycle, allowing RNA molecules to sample a wider range of conformations.

Fidelity

In order for protocells to have evolved in a Darwinian manner, the process of RNA replication must have been accurate enough to allow for the transmission of genetic information. This requires an error rate less than the reciprocal of the effective genome size, or about 1-2% for an RNA genome encoding 1 or 2 small ribozymes. Unfortunately, chemical copying using the four standard nucleobases is error-prone, and a major source of error is the G:U mismatch. Our recent studies showed that this problem can be alleviated by using s²U, a modified nucleotide that stabilizes canonical A:U pairs and destabilizes G:U wobble pairs. This single atom substitution improves both the kinetics and the fidelity of chemical copying of mixed-sequence RNA templates. We are actively exploring the potential roles of such noncanonical nucleotides during the infancy of the RNA world.

Reaction mechanism

A mechanistic understanding of nonenzymatic RNA copying requires detailed knowledge of the steps along the reaction pathway, including the binding of the monomers to the template and the ensuing chemical reaction that forms new phosphodiester bonds. Using NMR techniques, we have measured the thermodynamic association constants of nucleotide monophosphates to the template and probed conformations of both free and bound activated monomers. The results showed that RNA but not DNA templates induce conformational changes of the activated monomers. The changes in the conformation of the monomers increase their reactivity, and this explains why RNA templates are superior to DNA templates for nonenzymatic primer extension. Recently, we have obtained transition state models of the phosphoroimidazolide hydrolysis reaction by combining quantum mechanical calculations with experimental kinetic isotope effects and linear free energy relationships. The experimental observations are best interpreted by a calculated S_N2-like concerted transition structure with extensive scissile bond fission. We are now extending this line of research to understand how metal ions catalyze the chemical step of the RNA copying reaction.

Activated helper oligos

It has long been possible to copy C-rich templates by primer extension with 2-methylimidazole (2MeIm)-activated nucleotides. In contrast, the copying of mixed sequence templates containing all four nucleotides has been essentially impossible. We have recently found that by using an activated trimer helper oligo, e.g. with 2-MeIm on the 5′-phosphate, there is a dramatic acceleration of the reaction between the primer and the adjacent monomer. By using a set of activated helper trinucleotides, along with all four activated monomers, we can now copy short mixed sequence templates containing all four nucleotides, in good yield.

Alternative genetic polymers

In addition to RNA, many alternative genetic polymers have been proposed and some may have played a role in the early evolution of life. Our lab has characterized an extensive array of such genetic polymers, including threose nucleic acids, glycerol nucleic acids, as well as ribonucleotides with 2′-amino or 3′-amino modifications. Both 2′-amino-modified and 3′-amino-modified nucleotides are more reactive than their hydroxyl counterparts, enabling rapid and efficient copying of all four nucleobases on homopolymeric RNA and DNA templates. Future investigations will focus on copying mixed templates with all four nucleobases, within in fatty acid vesicles.