Multiple lines of evidence support the hypothesis that the early evolution of life was dominated by RNA, which can both transfer information from generation to generation through replication directed by base-pairing, and carry out biochemical activities by folding into functional structures. To understand how life emerged from prebiotic chemistry we must therefore explain the steps that led to the emergence of the RNA world, and in particular, the synthesis of RNA. The generation of pools of highly pure ribonucleotides on the early Earth seems unlikely, but the presence of alternative nucleotides would support the assembly of nucleic acid polymers containing nonheritable backbone heterogeneity. We suggest that homogeneous monomers might not have been necessary if populations of heterogeneous nucleic acid molecules could evolve reproducible function. For such evolution to be possible, function would have to be maintained despite the repeated scrambling of backbone chemistry from generation to generation. We have tested this possibility in a simplified model system, by using a T7 RNA polymerase variant capable of transcribing nucleic acids that contain an approximately 1∶1 mixture of deoxy-and ribonucleotides. We readily isolated nucleotide-binding aptamers by utilizing an in vitro selection process that shuffles the order of deoxy-and ribonucleotides in each round. We describe two such RNA/DNA mosaic nucleic acid aptamers that specifically bind ATP and GTP, respectively. We conclude that nonheritable variations in nucleic acid backbone structure may not have posed an insurmountable barrier to the emergence of functionality in early nucleic acids.abiogenesis | genetic takeover | RNA progenitor | origin of life G iven that RNA is likely to have played a key role early in the evolution of life (1), understanding the origins of the RNAbased biosphere is a critical aspect of understanding the origin of life. The difficulties associated with the prebiotic synthesis of a macromolecule as complicated and fragile as RNA have stimulated interest in the hypothesis that RNA was preceded by simpler and/or more stable progenitor nucleic acids (2-4). The systematic exploration of nucleic acids that are structurally related to RNA has revealed that the formation of stable WatsonCrick base-paired, antiparallel duplex structures is compatible with a surprising degree of variation in the sugar-phosphate backbone (5). These studies imply that many distinct nucleic acids are in principle capable of mediating the inheritance of genetic information. On the other hand, recent advances in the prebiotic chemistry leading to the pyrimidine ribonucleotides (6) have revived the prospects of the RNA-first model, with the attendant advantage of avoiding a difficult genetic takeover. Both RNAfirst and RNA-late models assume that life began with a single, well-defined genetic polymer. Perhaps the greatest problem implicit in this assumption is the challenge of explaining the origin of pure nucleotide monomers on the early Earth. Indeed, the supp...