The lack of even a marginal similarity between the two aminoacyl-tRNA synthetase (aaRS) classes suggests their independent origins (Eriani et al., 1990; Nagel and Doolittle, 1991). Yet, this independence is a puzzle inconsistent with the common origin of transfer RNAs, the coevolutionary theory of the genetic code (Wong, 1975, 1981) and other associated data and ideas. We present here the results of antiparallel 'class I versus class II' comparisons of aaRSs within their signature sequences. The two main HIGH- and KMSKS-containing motifs of class I appeared to be complementary to the class II motifs 2 and 1, respectively. The above sequence complementarity along with the mirror-image between crystal structures of complexes formed by the opposite aaRSs and their cognate tRNAs (Ruff et al., 1991), and the generally mirror ('head-to-tail') mapping of the basic functional sites in the sequences of aaRSs from the opposite two classes led us to conclude that these two synthetases emerged synchronously as complementary strands of the same primordial nucleic acid. This conclusion, combined with the hypothesis of tRNA concerted origin (Rodin et al., 1993a,b), may explain many intriguing features of aaRSs and favor the elucidation of the origin of the genetic code.

If the table of the genetic code is rearranged to put complementary codons face-to-face, it becomes apparent that the code displays latent mirror symmetry with respect to two sterically different modes of tRNA recognition. These modes involve distinct classes of aminoacyl-tRNA synthetases (aaRSs I and II) with recognition from the minor or major groove sides of the acceptor stem, respectively. We analyze the anticodon pairs complementary to the face-toface codon couplets. Taking into account the invariant nucleotides on either side (5 0 and 3 0 ), we consider the risk of anticodon confusion and subsequent erroneous aminoacylation in the ancestral coding system. This logic leads to the conclusion that ribozymic precursors of tRNA synthetases had the same two complementary modes of tRNA aminoacylation. This surprising case of molecular mimicry (1) shows a key potential selective advantage arising from the partitioning of aaRSs into two classes, (2) is consistent with the hypothesis that the two aaRS classes were originally encoded by the complementary strands of the same primordial gene and (3) provides a 'missing link' between the classic genetic code, embodied in the anticodon, and the second, or RNA operational, code that is embodied mostly in the acceptor stem and is directly responsible for proper tRNA aminoacylation.

Gene duplication is commonly regarded as the main evolutionary path toward the gain of a new function. However, even with gene duplication, there is a loss-versus-gain dilemma: most newly born duplicates degrade to pseudogenes, since degenerative mutations are much more frequent than advantageous ones. Thus, something additional seems to be needed to shift the loss versus gain equilibrium toward functional divergence. We suggest that epigenetic silencing of duplicates might play this role in evolution. This study began when we noticed in a previous publication (Lynch M, Conery JS [2000] Science 291:1151-1155) that the frequency of functional young gene duplicates is higher in organisms that have cytosine methylation (H. sapiens, M. musculus, and A. thaliana) than in organisms that do not have methylated genomes (S. cerevisiae, D. melanogaster, and C. elegans). We find that genome data analysis confirms the likelihood of much more efficient functional divergence of gene duplicates in mammals and plants than in yeast, nematode, and fly. We have also extended the classic model of gene duplication, in which newly duplicated genes have exactly the same expression pattern, to the case when they are epigenetically silenced in a tissue- and/or developmental stage-complementary manner. This exposes each of the duplicates to negative selection, thus protecting from "pseudogenization." Our analysis indicates that this kind of silencing (i) enhances evolution of duplicated genes to new functions, particularly in small populations, (ii) is quite consistent with the subfunctionalization model when degenerative but complementary mutations affect different subfunctions of the gene, and (iii) furthermore, may actually cooperate with the DDC (duplication-degeneration-complementation) process.

A total of 1268 available (excluding mitochondrial) tRNA sequences was used to reconstruct the common consensus image of their acceptor domains. Its structure appeared as a 11-bp-long double-stranded palindrome with complementary triplets in the center, each flanked by the 3'-ACCD and NGGU-5' motifs on each strand (D, base determinator). The palindrome readily extends up to the modern tRNA-like cloverleaf passing through an intermediate hairpin having in the center the single-stranded triplet, in supplement to its double-stranded precursor. The latter might represent an original anticodon-codon pair mapped at 1-2-3 positions of the present-day tRNA acceptors. This conclusion is supported by the striking correlation: in pairs of consensus tRNAs with complementary anticodons, their bases at the 2nd position of the acceptor stem were also complementary. Accordingly, inverse complementarity was also evident.at the 71st position of the acceptor stem. With a single exception (tRNAPhe-tRNAGlu pair), the parallelism is especially impressive for the pairs of tRNAs recognized by aminoacyl-tRNA synthetases (aaRS) from the opposite classes.The above complementarity still doubly presented at the key central position of real single-stranded anticodons and their hypothetical double-stranded precursors is consistent with our previous data pointing to the double-strand use of ancient RNAs in the origin of the main actors in translation-tRNAs with complementary anticodons and the two classes of aaRS.In the L-folded three dimensional structure of present-day tRNAs, their anticodon and 3'-terminal amino acid attachment site are separated by a distance of "70 A so that tRNA is unable to be correctly aminoacylated by itself. The specific aminoacylation is provided enzymatically by 20 different species of aminoacyl-tRNA synthetases (aaRS), one for each amino acid and the cognate set of isoacceptor tRNAs.On the one hand, to avoid the "chicken or egg" circular argument, direct interaction between anticodon and amino acid had to be hypothesized at 3' half of proto-tRNAs (1-3). In addition, the genetic code could have already existed in the prebiotic RNA world, possibly by direct aminoacylation of anticodons helped by RNA synthetase-like activity inherent in group