Selenocysteine codons decrease polysome association on endogenous selenoprotein mRNAs

Abstract

Selenocysteine incorporation has been reported to be inefficient in all systems studied, including Escherichia coli, baculovirus-insect cell systems, rabbit reticulocyte in vitro translation systems, transiently transfected mammalian cells, and intact animals. Nonetheless, full-length selenoproteins containing up to 17 selenocysteine residues are produced in animals, indicating that the efficiency observed in manipulated systems might not accurately reflect the true efficiency of this process in nature. To begin to address this apparent discrepancy, we have examined the polysome profiles of endogenously expressed selenoprotein mRNAs in a mammalian cell line, and compared them with nonselenoprotein mRNAs. We report that three selenoprotein mRNAs, type 1 deiodinase, glutathione peroxidase and selenoprotein P, are under-loaded with ribosomes, based on their predicted open reading frame sizes. The average numbers of ribosomes per mRNA correspond to the sizes predicted by termination at the UGA selenocysteine codons. Appropriate loading on the type 1 deiodinase mRNA is seen following substitution of a cysteine codon for the selenocysteine codon, indicating that the UGA codon confers a translational penalty on the mRNA. Surprisingly, ribosomal loading is also increased by the expression of eukaryotic release factors eRF1 and eRF3. These results suggest that the presence of a selenocysteine codon confers a translational penalty on selenoprotein mRNAs, and that increased levels of release factors may alter the kinetics of termination.