Interplay of tRNA-like structures from plant viral RNAs with partners of the translation and replication machineries.

The surprise was great in the early 1970s when it was shown that valine could be covalently attached to the 3' terminus of the genomic RNA from turnip yellow mosaic virus (TYMV) by valyl-tRNA synthetase (1) and, soon after, that tyrosyland histidyl-tRNA synthetases can aminoacylate the RNAs from brome mosaic virus (BMV) and tobacco mosaic virus (2, 3). The surprise was even greater when it was realized that the sequences and secondary foldings of the anticipated "tRNA" domains deviated markedly from those of canonical tRNAs (reviewed in refs. 4-6; Fig. 1) and that tRNA mimicry was linked with the mandatory presence of pseudoknots, the new type of RNA fold discovered in the tRNA-like domain of TYMV (8) and now found in many other RNAs (9, 10). It was soon conjectured about possible physiological roles of these tRNA-like structures (e.g., refs. 4, 11, and 12). When it was proven that they do not participate in protein synthesis, an easy explanation was to consider tRNA-like structures as insignificant remnants of evolutionary processes, but a more positive viewpoint was to consider them as possible actors during the life cycles of the viruses. On the other hand, the peculiar structural features of tRNA-like structures makes them attractive natural tRNA variants useful for investigating the specificity rules underlying recognition of tRNAs by aminoacyltRNA synthetases (6, 13).