Anticodon loop modifications of E. coli tRNAArgICG modulate structural flexibility and restrict decoding of the rare CGA codon

Abstract

Six codons are decoded by tRNAArg in E. coli, and three (CGU, CGC and CGA) are read by tRNAArg1,2isoacceptors. The anticodon stem and loop (ASL) domains of these isoacceptors differ only in the identity of the residue at position 32 as either a 2-thiocytosine (s2C32) or unmodified cytosine for tRNAArg1 and tRNAArg2 respectively. These isoacceptors also contain other naturally-occurring modifications at positions 34 (inosine, I) and 37 (2-methyladenosine, m2A37). To investigate the roles of these modifications in proper folding of the ASL, six ASL constructs, differing in their array of modifications, were analyzed by biophysical spectroscopic methods as well as functional binding assays. Although thermal denaturation and circular dichroism spectroscopy showed that the modifications contribute competing thermodynamic and base stacking properties indicating significant differences in structural dynamics, native gel electrophoresis and NMR spectroscopy clearly show that the equilibrium solution conformations of the ASLs are nearly identical, but do not possess the canonical U-turn structure needed for codon binding. Ribosome filter binding assays show that both s2C32 and m2A37 restrict binding to the rare CGA codon. Inosine seems to be the only modification required for cognate codon binding in the ribosomal A-site; therefore, ribosome binding experiments are being performed with the near cognate codons UGU, UGC and UGA to determine if the modifications inhibit misreading. Taken together, the results clearly indicate that chemical modifications modulate the flexibility of the loop; however, it remains unclear how they function in codon recognition. This research is supported by the National Science Foundation.