Molecular basis of human nuclear and mitochondrial tRNA 3’-processing

Abstract

Eukaryotic transfer RNA (tRNA) precursors undergo sequential processing steps to become mature tRNAs. In humans, ELAC2 carries out 3 prime-end processing of both nucleus-encoded (nu-tRNAs) and mitochondria-encoded tRNAs (mt-tRNAs). ELAC2 is self-sufficient for processing of nu-tRNAs, but requires TRMT10C and SDR5C1 to process most mt-tRNAs. Here, we show that TRMT10C-SDR5C1 specifically facilitate processing of structurally degenerate mt-tRNAs lacking the canonical elbow. Structures of ELAC2 in complex with TRMT10C, SDR5C1 and two divergent mt-tRNA substrates reveal two distinct mechanisms of pre-tRNA recognition. While canonical nu-tRNAs and mt-tRNAs are recognized by direct ELAC2-RNA interactions, processing of non-canonical mt-tRNAs depends on protein-protein interactions between ELAC2 and TRMT10C. These results provide the molecular basis for tRNA 3 prime-processing in both the nucleus and mitochondria and explain the organelle-specific requirement for additional factors. Moreover, they suggest that TRMT10C-SDR5C1 evolved as a mitochondrial tRNA maturation platform to compensate for the structural erosion of mt-tRNAs in bilaterian animals.