Rodin and Ohno proposed that antisense homology between aminoacyl-tRNA synthetase (aaRS) class I and II-defining catalytic peptides implied their origin as sense/antisense proteins from opposite strands of one primordial gene. Structural homology between class I and II aaRSs, respectively, and two contemporary sense/antisense proteins (1), showed that tertiary scaffolds organizing class I and II catalytic peptides also can be coded sense/antisense. Further verification of this unusual hypothesis requires the resolution of an important stumbling block: contemporary class I aaRS signature sequences are at least 80 residues further apart than those of class II aaRS with which they have putative sense/antisense homology. A 75-residue segment connecting the first and second halves of the Rossmann dinucleotide binding fold contributes only indirectly to the structure of the catalytic site of Tryptophanyl-tRNA synthetase, the smallest class I enzyme, and provides no catalytically important residues. Bioinformatic analysis of the class I aaRS superfamily reveal few, if any, highly conserved residues in this insert and show that it underlies virtually all structural adaptive radiation in the other nine class I aaRS. Protein design considerations suggest that the insert can be replaced by a single peptide bond. Deleting this segment is a natural solution to the conundrum posed by the mismatch of aaRS class signature peptide intervals. Demonstration of catalytic activity in this fragment would therefore verify an important prediction of the Rodin-Ohno hypothesis.

We will also discuss two additional implications of this hypothesis for early protein evolution:

1. One aaRS from each class sufficed to initiate the protein evolutionary ?big bang? using error-prone, binary patterns drawn from the two major subclasses of amino acids, Ia and IIa, which now include, respectively, (LIVMRC) and (AGSTHP).
2. The contemporary proteome may have evolved largely from the two aaRS progenitors. Class I aaRS share a Rossmann fold and so belong to the largest known superfamily, which contributed ∼50 distinct enzymes to the Last Universal Common Ancestor (LUCA) (2), and were plausibly progenitors of this superfamily. There is less evidence, so far, for a corresponding superfamily including class II aaRS, which should be apparent if the two classes arose simultaneously from sense/antisense coding by the same gene.

Acknowledgement

Supported by NIGMS.

References and Footnotes

1. Carter, C. W., Jr., Duax, W. L. Mol. Cell 10, 705-708 (2002).
2. Aravind, L., et al., Curr. Op. Str. Biol. 12, 392-399 (2002).