Upstream Open Reading Frames (uORFs) have emerged as major cis-acting elements involved in regulating mRNA turnover and translation (Wethmar, 2014; Ingolia 2014). We and others have found that genetic variation in putative uORFs has contributed to variation in translation in yeast (McManus et al., 2014), mice (Hou et al., 2014), and humans (Cenik et al., 2015). Furthermore, mutations in uORFs are associated with human disease (Calvo et al., 2009). Functional uORFs can be challenging to identify, as they often reside in alternative regions of mRNA transcript leaders, and have limited sequence conservation. Curiously, recent work has uncovered a large number of uORFs initiating with non-AUG start codons, with functions implicated in development and stress responses (Ingolia et al., 2009; Brar et al., 2011). While the catalog of such noncanonical uORFs continues to increase, little is known about their functions and evolution.
To study the evolution of uORFs, we used targeted RNA-sequencing to map transcript start and end sites in all five Saccharomyces sensu stricto yeast species. Comparison of UTR lengths in these species revealed that genes involved in translation and glycolysis have highly conserved transcription start and end sites, many that have remained entirely stable over 15 million years. Many genes contain conserved alternative transcription start and end sites, suggesting transcriptional regulation of UTR sequences. As expected from prior work, “AUG” triplets are depleted from 5’ UTRs in each species. Surprisingly, we find that “UUG” is the most enriched triplet in 5’ UTRs, suggesting selective pressure to maintain uORFs initiating with this codon. To investigate the importance of UUG enrichment, we searched 5’ UTRs for functional uORFs using ribosome profiling data from S. cerevisiae, S. paradoxus, and S. uvarum using a novel machine learning algorithm (uORF-seeker). This analysis identified both conserved and species-specific AUG- and UUG-uORFs in hundreds of orthologous 5’ UTRs. Interestingly, conserved UUG uORFs are highly enriched in genes involved in regulating cellular growth during stress. Our results show that UUG- and AUG-uORFs have similar rates of conservation and implicate UUG-uORFs in repressing cellular growth genes in response to stress.