The secondary structure of the complete mRNAs and its variants were predicted using the RNAfold [46 (link)] and mFOLD servers [47 (link)]. We used the default server conditions and simulated the 30S-bound structural layout by incorporating unpaired constrains from the SD to the start of the coding sequence (S1 File). To model the in vivo secondary structure, unpaired nucleotides resulting from chemical probing were set as unbound constrains [38 (link)] (S1 File).
Three-dimensional structures of the mRNAs were built using the RNAcomposer server, and the resulting pdb file was visualized in Pymol and MOE software [48 (link)] (Chemical Computing Group, Montreal, Canada). Structural alignments and RMSD between two structures were calculated using the MOE software (Chemical Computing Group, Montreal, Canada). For the structural modelling of the 30S IC and pre-IC, the predicted SR IIIb and TIR until +20 from the GUG were aligned with the bound mRNA of PDBs 5lmp and 5lmv, respectively [42 (link)]. mRNA structural alignments and visualization of the resulting models were performed with Pymol software (Schrödinger, New York, New York). Alignment and secondary structure prediction of the mTufA SETI of E. coli, Shigella dysenteriae, and Salmonella enterica were performed using the PETfold web server (S1 File) [49 (link)].
Free full text: Click here