The mechanism of substrate recognition by promiscuous aminoacyl-tRNA synthetases

Faithful translation of the genetic information, imprinted in the mRNA, into a functional protein is a fundamental biological process. Aminoacyl-tRNA synthetases (aaRSs) are ancient enzymes that ensure fidelity of gene translation by coupling every proteinogenic amino acid with the cognate transfer RNA (tRNA). Whereas selection of the amino acid is dictated by the architecture of the aminoacylation site, recognition of tRNA requires specific interactions between the aaRS and several structural motifs in the tRNA. Because of these features it was presumed that aaRSs are highly specific and cannot act on tRNAs with different structures and anticodon sequences. This also served as a basis for the proposal that the genetic code is universal and that perhaps arose as the result of the “frozen accident”. However, in the yeast mitochondria the CUN box codes for threonine instead of the canonical leucine, whereas the opal stop codon is recoded into a signal for selenocysteine incorporation in all domains of life. The main goal of this doctoral thesis is to explain at the structural level how these genetic code reassignment events are promoted. These mechanisms are of particular interest since they alter the interpretation of the genetic code and thus facilitate its evolution. Our results show that the leucine-to-threonine recoding in the yeast mitochondria is promoted by a promiscuous ThrRS (MST1) and a non-canonical tRNA1Thr that harbors enlarged anticodon loop and anticodon sequence complementary to the CUN codon box. We show that MST1 employs distinct mechanisms for binding to tRNA1Thr and tRNA2Thr, that pre-transfer editing is utilized for selection against serine, and that unique properties of the anticodon-binding domain are perhaps important for binding to distinct isoacceptor tRNAs. On the other hand, recoding of the opal stop codon is facilitated by promiscuous SerRS that acts on both tRNASer and tRNASec. In contrast to MST1, human SerRS does not bind anticodon loops of either tRNA. Further, our results show that SerRS binds both tRNAs with similar binding affinities, yet utilizes tRNASer as a superior substrate presumably because of its canonical fold. Hence, MST1 and SerRS use completely different strategies for promoting recoding of particular codons.