Modifications Modulate Anticodon Loop Dynamics and Codon Recognition of E. coli tRNAArg1,2

Three of six arginine codons are read by two tRNAArg isoacceptors in Escherichia coli. The anticodon stem and loop of these isoacceptors (ASLArg1,2) differs only in that the position 32 cytidine of tRNAArg1 is posttranscriptionally modified to 2-thiocytidine (s2C32). The tRNAArg1,2 are also modified at positions 34 (inosine, I34) and 37 (2-methyladenosine, m2A37). To investigate the roles of modifications in the structure and function, we analyzed six ASLArg1,2 constructs differing in their array of modifications by spectroscopy and codon binding assays. Thermal denaturation and circular dichroism spectroscopy indicated that modifications contribute thermodynamic and base stacking properties, resulting in more order but less stability. NMR-derived structures of the ASLArg1,2 showed that the solution structures of the ASLs were nearly identical. Surprisingly, none possessed the U-turn conformation required for effective codon binding on the ribosome. Yet, all ASLArg1,2 constructs efficiently bound the cognate CGU codon. Three ASLs with I34 were able to decode CGC, whereas only the singly modified ASLArg1,2ICG with I34 was able to decode CGA. The dissociation constants for all codon bindings were physiologically relevant (0.4–1.4 μM). However, with the introduction of s2C32 or m2A37 to ASLArg1,2ICG, the maximum amount of ASL bound to CGU and CGC was significantly reduced. These results suggest that, by allowing loop flexibility, the modifications modulate the conformation of the ASLArg1,2, which takes one structure free in solution and two others when bound to the cognate arginyl-tRNA synthetase or to codons on the ribosome where modifications reduce or restrict binding to specific codons.