Eukaryotic translation initiation is a complex process involving the ribosome, mRNA, Met-tRNAiMet and numerous eukaryotic initiation factors (eIFs). Decades of studies driven by those using the model eukaryote yeast Saccharomyces cerevisiae revealed that the key process is the formation of codon-anticodon base pairing in the small ribosome P-site. Stringent initiation is enabled by formation of the 48S ribosomal pre-initiation complex (PIC) strictly at the AUG start codon while excluding initiation at other sites. Intriguingly, however, many non-canonical start sites are utilized in some biological contexts and diseases such as cancer and neurodegenerative disorders. The list of non-AUG start sites within the human genome is far from being complete and how the use of these sites and hence protein production from these sites are regulated remains an open question. Thus, Aim 1 of this grant is to make such a list of non-AUG start sites through genome-wide translation profiling of well characterized cancer model systems, verify some of these sites and determine the mechanism driving the observed non-AUG translational regulation in cancer. The Aim 2 is to study the mechanistic role of 5MP and Met-tRNAiMet adenosine N6- threonylcarbamoylation (t6A) in controlling non-AUG translation. It will be tested if 5MP mutations found in many types of cancer can alter initiation accuracy, thereby affecting patients' prognosis. Our preliminary studies suggested that t6A located 3' of Met-tRNAiMet anticodon can discriminate specifically against GUG and UUG start codons, in contrast to eIF1 being more universal non-AUG discriminator. Combining molecular dynamics simulation methods, we will test if the recently discovered cyclic t6A serves the discriminating role and determine how the cooperation or competition between t6A and eIF1 promotes stringent initiation and leaky scanning crucial for translational regulation. The Aim 3 is to study yet a distinct mechanism of start codon selection that is exploited during the heat shock response at the translational level. This mechanism was discovered through translational profiling of yeast eIF3i mutant defective in its interaction with RNA-binding eIF3g subunit at a high temperature. The working hypothesis assumes that, at a high temperature, stable PIC formation at the AUG codon requires additional mRNA elements anchoring the initiating ribosome at its entry site. These hypothetical mRNA elements are located downstream of the start codon and include a novel eIF3g-binding motif 5'-GUCG-3' and a downstream stem- loop that potentially binds the entry site-side of the 40S, thereby stabilizing the PIC formation. This model will tested using yeast as a model system.