all: there is no reason a priori why the mature 5' ends of modern tRNAs could not be generated directly by transcription, yet there is only one instance known in which this is so (Lee et al., 1989).

If early tRNAs functioned as telomeres, it would also explain the surprising fact that all cells contain tRNA nucleotidyltransferase, an activity that can regenerate the 3'-terminal CCA sequence of tRNAs. This activity can be viewed as a telomerase, responsible for maintaining the integrity of the terminal CCA. Interestingly, although eubacteria and archaea encode the 3' terminal CCA of tRNA, eukaryotes do not, and they must rely instead on tRNA nucleotidyltransferase to produce mature tRNAs (Palmer et al., 1992). The primitive state is not yet known, but as tRNA nucleotidyltransferase is present in all cells, it seems most plausible that genomically encoded CCA was devised later to circumvent a slow or inefficient step in tRNA processing. This may also explain why the telomerase function of tRNA nucleotidyltransferase can in certain cases be augmented by the replicase itself. A number of modern RNA and DNA polymerases, including Qß RNA replicase (Blumenthal and Carmichael, 1979) and Taq DNA polymerase will add an untemplated A to the 3' end of a newly synthesized molecule (for example, see Tse and Forget, 1990). Polyadenylation of polymerase II transcripts, which often occurs following a CA dinucleotide (Wigley et al., 1990; Raabe et al., 1993), may be viewed as an exaggerated example of this.

Viruses May Be Clues to Early Evolution. The Qß RNA genome is replicated by an enzyme composed of four subunits. Only subunit II is encoded by the phage genome; the other three subunits—ribosomal protein S1 and elongation factors Tu and Ts (Blumenthal and Carmichael, 1979)—are components of the translation apparatus. If Qß is viewed as no more than a cellular parasite, then this simple bacteriophage appears to have stolen elements of the protein synthesis machinery for its replication. However, if tRNAs arose early to function in replication, it would be natural for factors that recognized tRNA to accompany this molecule as it took on a new role in translation. The presence of translation factors in an RNA replicase may therefore be viewed as evidence that these two processes have long been intimately connected.

In addition, the notion that some modern viruses still reveal their ancient origins allows us to see viruses in a new way. Viruses usually command our attention as vectors of disease, and the extraordinary genomic diversity of modern viruses is rarely appreciated by nonvirologists. There are double-stranded viruses, single-stranded viruses, circular viruses and linear viruses, RNA viruses, DNA viruses, viruses



The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement