In our study 'Monitoring the 5′UTR landscape reveals isoform switches to drive translational efficiencies in cancer' we used Nanopore long-read sequencing and cap analysis of gene expression (CAGE-seq) to documents the 5' and 3' untranslated region (UTR) landscape of epidermal stem cells, wild-type keratinocytes and squamous cell carcinoma. To serve as a resource, we made the datasets publicly available in the UCSC genome browser under the following link:
These datasets will allow you to 1. check out the isoform usage of your favorite gene by browsing the Oxford Nanopore Technology (ONT) datasets (Nanopore long-read sequencing) in different cell types of the skin: epidermal stem cells (EpSC), wild-type keratinocytes (Keratinocytes) and squamous cell carcinoma (SCC);
2. assess the exact 5' end and heterogeneity of transcription start site usage by the CAGE-seq data mapping to the negative and positive strand of the reference genome (e.g. EpSC neg and EpSC pos);
3. analyse how the 5' transcription start site tunes the translational efficiency of the mRNA isoform by the CAGE-seq data in LP (low polysomes=inefficiently translated) and HP (high polysomes=efficiently translated);
4. identify the 5' transcription start site isoforms that are most efficiently translated (ratio HP vs LP).
A genome-wide analysis of our data showed that squamous cell carcinoma and epidermal stem cells — compared to wild-type keratinocytes — express numerous alternative mRNA isoforms, many of them encoding 5' terminal oligopyrimidine (TOP) mRNAs, e.g. ribosomal protein mRNAs. Focussing on squamous cell carcinoma, we find that a specific cohort of genes switch towards an increased use of 5'TOP and pyrimidine-rich translational element (PRTE) motifs to drive efficient and mTORC1-dependent translation of the 5'UTR isoforms with identical coding sequences. This heterogeneity in transcription start site usage of 5'TOP motif-containing mRNAs represents a simple but effective way to tune the translational efficiency of these mRNAs and their sensitivity to the mTORC1-dependent nutrient-sensing pathway without changing the protein's primary sequence. Notably, we found that 5'TOP motif-containing, but not the TOP-less, RPL21 transcript isoform expression strongly correlates with overall survival in human head and neck squamous cell carcinoma patients. These findings warrant isoform-specific analyses in human cancer datasets and highlight the role of translational control during cancer formation.