Boreskov

PRE‐TRANSLATIONAL ORIGIN OF THE GENETIC CODERodin S.N. 1,2 , Eörs Szathmáry 2,3 and Rodin A.S. 2,41 Beckman Research Institute of the City of Hope, Duarte, CA 91010, USA2 Collegium Budapest, Szentháromság u. 2, H‐1014 Budapest, Hungary3 Parmenides Center for the Study of Thinking, Kirchplatz 1, D‐82049 Munich, Germany4 Human Genetics Center, SPH, University of Texas, Houston, TX 77225, USAThe major challenge to the understanding of the genetic code origins is the archetypal“key‐lock vs. frozen accident” dilemma (Crick, 1968). Recently we have re‐examined thedilemma (Rodin et al., 2009, 2011) in light of modular structures of tRNAs and aminoacyl‐tRNAsynthetases (aaRS) (actually bringing the code into action), and the updated library of aminoacid‐binding sites of RNA aptamers selected in vitro (Yarus et al., 2009; Yanus et al., 2010).The aa‐binding sites of arginine, isoleucine and tyrosine contain both their cognatetriplets, anticodons and codons. We observed that this puzzling error‐prone simultaneouspresence is associated with palindrome‐dinucleotides (Rodin et al., 2011). For example, onebaseshift to the left brings arginine codons CGN, with CG at 1‐2 positions, to the respectiveanticodons NCG, with CG at 2‐3 positions. Technically, the concomitant presence of codonsand anticodons is also expected in the reverse situation, with codons containing palindromedinucleotidesat their 2‐3 positions, and anticodons exhibiting them at 1‐2 positions. A closeranalysis reveals that, surprisingly, RNA binding sites for Arg, Ile and Tyr “prefer”, exactly asin the actual genetic code, anticodon(2‐3)/codon(1‐2) tetramers to their anticodon(1‐2)/codon(2‐3) counterparts, despite the seemingly perfect symmetry of the latter. However,since in vitro selection of aa‐specific RNA aptamers apparently had nothing to do withtranslation, this striking preference provides a new strong support for the notion of thegenetic code emerging before translation, in response to catalytic (and possibly other) needsof ancient RNA life.Consistent with the pre‐translational origin of the code are our updated phylogeneticstudy of tRNA genes (Rodin et al., 2009) and a new model of gradual (Fibonacci iterationlike)evolutionary growth of tRNAs – from a primordial coding triplet and 5’‐DCCA‐3’ (D is abase‐determinator) to the eventual 76 base‐long cloverleaf‐shaped molecule (Rodin et al.2011). The pre‐translational genetic code origin is also consistent with the partition of aaRSsin two structurally unrelated classes with sterically mirror modes of tRNA recognition (Erianiet al., 1990; Delarue, 2007). This partition seems to have protected proto‐tRNAs withcomplementary anticodons from otherwise very likely confusion, the only “exception” being112