(378e) High-throughput, Comprehensive Analysis of Eukaryotic Translation Initiation Motifs by FACS-Seq

The probability that a scanning ribosome will stop at a start codon and initiate translation is a function of the sequence surrounding the AUG. Unlike bacterial translation initiation, which uses well defined Watson-Crick base paring between the ribosome and the ribosome binding site, eukaryotic translation initiation is not well understood. Marilyn Kozak’s consensus sequence provides helpful rules of thumb but is not quantitative. We have revisited this problem using a high-throughput, quantitative approach named FACS-seq which couples fluorescence activated cell sorting (FACS) with next-generation sequencing. Our mutational library consisted of random permutations of the 6 bases preceding and 2 bases following a start codon driving the expression of a fluorescence reporter. In agreement to Kozak’s results, we find the -3 position to be the largest determinant of translation initiation. Contrary to Kozak’s results, we find the +4 to be of little importance and instead find the two bases neighboring the -3 position to play a major role. A translation initiation prediction model incorporating the FACS-seq data was found to be in agreement with ribosomal footprinting data and was able to predict translation of upstream open reading frames. In principle, the FACS-seq method described here could be applied to any mutational library that can be paired with a fluorescence reporter.