A dynamic structural rearrangement in the phylogenetically conserved helix 27 of Escherichia coli 16S rRNA has been proposed to directly affect the accuracy of translational decoding by switching between “accurate” and “error-prone” conformations. To examine the function of helix 27 in eukaryotes, random and site-specific mutations in helix 27 of the yeast Saccharomyces cerevisiae 18S rRNA have been characterized. Mutations at positions of yeast 18S rRNA corresponding to E. coli 886 (rdn8), 888 (rdn6), and 912 (rdn4) increased translational accuracy in vivo and in vitro, and caused a reduction in tRNA binding to the A-site of mutant ribosomes. The double rdn4rdn6 mutation separated the killing and stop-codon readthrough effects of the aminoglycoside antibiotic, paromomycin, implicating a direct involvement of yeast helix 27 in accurate recognition of codons by tRNA or release factor eRF1. Although our data in yeast does not support a conformational switch model analogous to that proposed for helix 27 of E. coli 16S rRNA, it strongly suggests a functional conservation of this region in tRNA selection.

We describe a new computational technique to predict conformationally switching elements in proteins from their amino acid sequences. The method, called ASP (Ambivalent Structure Predictor), analyzes results from a secondary structure prediction algorithm to identify regions of conformational ambivalence. ASP identifies ambivalent regions in 16 test protein sequences for which function involves substantial backbone rearrangements. In the test set, all sites previously described as conformational switches are correctly predicted to be structurally ambivalent regions. No such regions are predicted in three negative control protein sequences. ASP may be useful as a guide for experimental studies on protein function and motion in the absence of detailed three-dimensional structural data.