Abstract
Under adverse conditions, cells activate stress response pathways to ensure survival. Such responses involve the inhibition of global cap-dependent translation, thereby conserving metabolic energy, restricting the production of toxic proteins, and releasing translational machinery to be re-tasked to produce critical stress response proteins. Although necessary, this repression represents a block that essential transcripts must then escape through alternative methods of translation initiation, such as the use of an internal ribosome entry site (IRES). In viruses, IRESes are well documented to have distinct structures and to require a limited repertoire of cellular factors for their activity. Conversely, cellular IRESes are less well studied and understood. In general, cellular IRESes are found in the 5′ untranslated region (UTR) of mRNA and have been identified in many critical cellular stress response transcripts, highlighting the importance of this mechanism of translation initiation. Understanding the mechanisms that govern cellular IRESes has the potential for transformative insights into protective adaptation and the varied and diverse methods of eukaryotic translation initiation. We previously identified cellular IRESes in the murine insulin receptor (Insr) and insulin-like growth factor 1 receptor (Igf1r) transcripts. These transcripts are targets of the stress responsive transcription factor FOXO and are upregulated during low nutrient stress. We showed that these cellular IRESes are resistant to the inhibition of translation initiation factors eIF4E andviii
eIF4A. However, the specific mechanisms governing the IRES-mediated translation of these transcripts remain largely unknown. Here, by examining how alternative ternary complexes may drive Insr and Igf1r IRES-mediated translation, we found that eIF5B promotes the activity of these two cellular IRESes and that of the Hepatitis C Virus (HCV). Furthermore, we discovered that this stimulation occurs through a previously uncharacterized, non-canonical role of eIF5B that requires its highly charged and disordered amino terminus. We observed that the N-terminal domain of eIF5B drives self-association and puncta formation in cells, that eIF5B condensate formation is triggered by cellular stress, that eIF5B clustering is sufficient to promote IRES activity, and that Insr IRES containing RNA are present in these puncta. These data indicate that N-terminal domain mediated eIF5B self-association creates a local cellular environment that preferentially favors non-canonical, IRES-mediated translation. Moreover, we found that the Insr and Igf1r cellular IRESes are able to mediate translation in the cell independently of the scaffolding protein eIF4G1 and that eIF4G1 and eIF5B depletion prompt significant changes in global active translation. While a complete attenuation of translation is not achieved in either case, the effect of eIF4G1 degradation is much more dramatic than that of eIF5B. These findings, however, indicate the potential identification of transcripts that are either sensitive or insensitive to eIF4G1 or eIF5B depletion.