Nonstop mRNAs generate a ground state of mitochondrial gene expression noise

Abstract

A stop codon within the mRNA facilitates coordinated termination of protein synthesis, releasing the nascent polypeptide from the ribosome. This essential step in gene expression is impeded with transcripts lacking a stop codon, generating nonstop ribosome complexes. Here, we use deep sequencing to investigate sources of nonstop mRNAs generated from the human mitochondrial genome. We identify diverse types of nonstop mRNAs on mitochondrial ribosomes that are resistant to translation termination by canonical release factors. Failure to resolve these aberrations by the mitochondrial release factor in rescue (MTRFR) imparts a negative regulatory effect on protein synthesis that is associated with human disease. Our findings reveal a source of underlying noise in mitochondrial gene expression and the importance of responsive ribosome quality control mechanisms for cell fitness and human health.

A kinetic dichotomy between mitochondrial and nuclear gene expression drives OXPHOS biogenesis

“…To validate this, we analyzed mitoribosome profiling data to determine whether ribosome-protected fragments (RPF) contain uncleaved mt-mRNA (Fig 4C-F) 5 . We found several transcripts with unprocessed 3’-ends in the ribosomes (Fig 4D, F and Fig S4C, Table S6), consistent with the observation that mRNAs lacking stop codons are nevertheless commonly translated by mitoribosomes 35 . However, only one transcript, COX3 , had unprocessed 5’ ends in the ribosome (Fig 4C, E, and S4B), in stark contrast to the nanopore direct RNA-seq, in which we found a large fraction of unprocessed transcripts also for ND1 and COX1 (Fig 2).…”

“…About 20% of COX1 transcripts were unprocessed at the 5’-end in whole cell lysate, while less than 1% of transcripts were unprocessed in the mitoribosome (Figures 4C, and S4A, Table S3). Interestingly, an equally large proportion of COX1 transcripts ended prematurely before the 3’-end, and this population was not strongly depleted in the mitoribosome, consistent with the observation that mRNAs lacking stop codons are nevertheless commonly translated by mitoribosomes 42,43 (Figure 4C). For ND1 and ND2, we found a smaller depletion of 5’-end unprocessed transcripts in the ribosome (Figure S4A).…”

“…Mitochondrial mRNAs, except ND6 , have short, 40-60 nt long, poly(A) tails 40 compared to nuclear-encoded mRNA (110 nt long) 25,41 . For most mt-transcripts, the poly(A) tail is required to complete the stop codon 40 , but some reports suggest that it is dispensable 42,43 . We wondered if poly(A) tailing kinetics were under less stringent time requirements and would occur more slowly than processing.…”

“…To validate this, we analyzed mitoribosome profiling data to determine whether S3), which is consistent with the fact that mRNAs lacking stop codons are nevertheless commonly translated by mitoribosomes 35 . However, only one transcript, COX3, had unprocessed 5 prime ends in the ribosome (Fig 4C , E, and S4B), in stark contrast to the nanopore direct RNA-seq where we found a large fraction of unprocessed transcripts (Fig 2).…”

“…Indicating that for ND1 and COX1, 5'-end processing might be a pre-requisite of ribosome association. To validate this, we analyzed mitoribosome profiling data to determine whether ribosome-protected fragments (RPF) contain S6), consistent with the observation that mRNAs lacking stop codons are nevertheless commonly translated by mitoribosomes 35 . However, only one transcript, COX3, had unprocessed 5' ends in the ribosome (Fig 4C , E, and S4B), in stark contrast to the nanopore direct RNA-seq, in which we found a large fraction of unprocessed transcripts also for ND1 and COX1 (Fig 2).…”