How Did tRNA Come to Play a Role in Translation? Covalent linkage of a basic amino acid to a 3'-terminal tRNA-like genomic tag might have improved the efficiency or specificity of replication in an RNA world, perhaps by permitting the negatively charged RNA replicase to bind more tightly to the negatively charged RNA genome. Alternatively, if aminoacylation interfered with replication, charging could have limited the number of genomes in the replicative pool or prevented free genomic tags from competing for the replicase. The aminoacylation activity could have arisen very early, even as a variant of the replicase itself. Aminoacylation chemically resembles RNA polymerization, and a variant replicase could have evolved to catalyze aminoacylation, just as a group I ribozyme, which naturally catalyzes phosphoester bond transfer, can be redesigned to catalyze reactions at a carbon center (Piccirilli et al., 1992). The specificity with which modern group I ribozymes bind L-arginine leaves little doubt that an aminoacyl-tRNA synthetase made of RNA could charge tRNA with considerable specificity (Connell et al., 1993). In any case, as we discuss in greater detail elsewhere (Maizels and Weiner, 1987), replication would have provided the driving force for the first two steps in the evolution of protein synthesis.
The apparent diversity in size and quaternary structure of modern aminoacyl-tRNA synthetases has long been puzzling. All these enzymes perform the same two-step reaction using an enzyme-bound aminoacyl-adenylate intermediate, and one might therefore have expected that all would be descended from a single ancestral protein. This mystery was refined but not clarified by the realization that modern aminoacyl-tRNA synthetases can be divided into two structurally and functionally distinct classes: synthetases with the classical Rossman nucleotide-binding fold charge the 2' hydroxyl of tRNA, and synthetases with a seven-stranded antiparallel b-sheet generally charge the 3' hydroxyl (Cusack et al., 1990; Eriani et al., 1990; Ruff et al., 1991; Cavarelli et al., 1993). However, if aminoacyl-tRNA synthetases first arose in an RNA world, as we suggested (Weiner and Maizels, 1987), and were then transformed by stepwise replacement of RNA with protein as envisioned by White (White, 1982), even a single RNA enzyme could give rise to multiple protein enzymes because there is unlikely to be a unique path for replacement of RNA by protein (Weiner and Maizels, 1987; Benner et al., 1989).
The Top Half of tRNA Is Ancient. The tRNA-like structures in early genomes may have consisted simply of a coaxial stack of the TΦC arm on the CCA acceptor stem. We base this suggestion on two different kinds of evidence (Figure 5 and Table 1). First, the top half of modern