The possible first tetrad recruited by the genetic code (top panel) (reviewed in [26, 27]). Consistent with our model, these four amino acids show dual complementarity not only in the legitimate Gly(GCC) – Ala(GGC) and Asp(GUC) – Val(GAC) pairings but also in the Gly(GCC) – Val(GAC) and Ala(GGC) – Asp(GUC) pairings, with G-U and even A*C mismatch at the central position . Accordingly, the Ala-Gly pair could generate the new Val-Asp pair via C→U and G→A transitions (ibid.) The Ala→Val expansion of the code is accompanied by a change of tRNA recognition from the major (blue) to the minor (yellow) groove side, whereas the complementary Gly→Asp leaves the mode (blue) unchanged. This asymmetry could be associated with a risk of pleiotropic tRNA mis-aminoacylations by ribozymes. The schematic below (under the tetrad) (adapted from ) demonstrates this risk for Ala→Val and Gly→Asp expansions with legitimate (solid arrows) and 'wobbling' (dotted arrows) recognitions of anticodons by r-aaRSs. The Ala(GCC) → Val(GUC) expansion is prone to multiple mis-aminoacylations of the old tRNAAla by the new r-ValRS (red dotted arrow). In contrast, the complementary Gly(GGC) → Asp(GAC) expansion has no such disadvantage (green dotted arrow). To escape potential pleiotropic complications a different mode of tRNA recognition is required that would safely distinguish the r-ValRS from the already established r-AlaRS. This is precisely what happened: AspRS preserved the (same as GlyRS) recognition from the major groove (blue), whereas ValRS adopted the recognition from the minor groove (yellow).