Fig. 2. Bi-directional PIC scanning contributes to start codon selection.

a Schematic of cross-linking between m⁷G cap and eIF4E. Tethering eIF4E to m⁷G cap will prevent start codon selection in the ‘threading’ model, while permitting recognition of start codon downstream of the ‘blind spot’ in the ‘slot-in’ model. b Translational effects of cross-linking between m⁷G cap and eIF4E using Fluc mRNAs with β-globin 5ʹ UTR. The top panel used m⁷s⁶G capped mRNAs followed by cross-linking under UV 360 nm. The bottom panel used redox reactions for cross-linking. In vitro translation was conducted in RRL and Fluc activities were monitored by luminometry. Error bars; mean ± SEM; n = 3 biological replicates. c Translational effects of cross-linking between m⁷G cap and eIF4E using Fluc mRNAs with ultra-short 5’UTR (2 nt). The top and bottom panels are similar to (b). Error bars; mean ± SEM; n = 3 biological replicates. d Selection of closely spaced AUG codons by toe-printing assays. Two AUG codons with identical flanking sequences but different length of the spacer were inserted into Fluc mRNAs with β-globin 5’UTR. Expected positions corresponding to the ribosome leading edge are highlighted. The results are representative of three biological replicates. e Selection of start codons flanking the IRES. The polio IRES was inserted between the uORF encoding SIINFEKL and GFP with 6 nt or 18 nt downstream of the uORF stop codon. The mRNA reporters were capped with non-functional ApppG. The flow cytometry scatterplots of HEK293-K^b cells transfected with mRNA reporters are representative of three biological replicates. f For the mRNA reporters as in (e), a stem loop with different stability was placed near the 5ʹ end (left panel) or an out-of-frame AUG was inserted downstream of the main AUG. The flow cytometry scatterplots of HEK293-K^b cells transfected with mRNA reporters are representative of three biological replicates. Source data are provided as a Source Data file.