Mitochondrial Translation Initiation and Translational Activation

We have used mutational and revertant analysis to study the elements of the 54nucleotide COX2 5’untranslated leader involved in translation initiation in yeast mitochondria and in activation by the COX2 translational activator, Pet1 1 lp. We generated a collection of mutants with substitutions spanning the entire COX2 5’-UTL by in vitro mutagenesis followed by mitochondrial transformation and gene replacement. The phenotypes of these mutants delimit a 31-nucleotide segment, from -16 to -46, that contains several short sequence elements necessary for COX2 5’-UTL function in translation. The sequences from -16 to -47 were shown to be partially sufficient to promote translation in a foreign context. Analysis of revertants of both the series of linker-scanning alleles and two short deletion/ insertion alleles has refined the positions of several possible functional elements of the COX2 5”untranslated leader, including a putative RNA stem-loop structure that functionally interacts with Petlllp and an octanucleotide sequence present in all S. cerevisiae mitochondrial mRNA 5’-UTLs that is a potential rRNA binding site. T HE principal function of the mitochondrial genetic system is the production of several subunits of respiratory enzyme complexes located in the mitochondrial inner membrane. Most components of the machinery responsible for the expression of mitochondrially encoded genes are distinct from the components responsible for nuclear gene expression. In the yeast Saccharomyces cerevisiae, over 100 nuclear genes function in translation of the seven respiratory protein subunits encoded in mitochondrial DNA (mtDNA; reviewed in FOX 1996). One unusual feature of the yeast mitochondrial translation system is that translation of each of the mitochondrial mRNAs studied so far requires at least one gene-specific translational activator in addition to the general translation machinery (reviewed in DIECKMANN and STAPLES 1994; FOX 1996). Translation of COX2 mRNA (encoding cytochrome c oxidase subunit 11) is specifically dependent on the PET111 nuclear gene product (POUTRE and FOX 1987; MULERO and F o x 1993b), while translation of COX3 mRNA is specifically dependent on the nuclear gene products of PET54, PET122 and PET494 (COSTANZO and F o x 1986; C o s TANZO et al. 1986; FOX et al. 1988a). The activator proteins appear to tether mRNAs encoding hydrophobic proteins to the mitochondrial inner membrane (MCMULLIN and FOX 1993) and mediate ribosomemRNA interactions (MCMULLIN et al. 1990; €€UTTER et al. 1991), promoting cotranslational membrane insertion. Activator proteins may play additional roles in miCorresponding author: Thomas D. Fox, Section of Genetics and Development, 101 Biotechnology Building, Cornell University, Ithaca, NY 148532703, E-mail: [email protected] Genetics 147: 87-100 (September, 1997) tochondrial translation initiation, possibly positioning ribosomes on the mRNh (reviewed in FOX 1996). The 5”untranslated leaders (5”UTLs) of the COX2, COX3 and COB mRNAs contain the sites of action for the mRNA-specific translational activators that act on each of these genes (RODEL and F o x 1987; COSTANZO and F o x 1988; MULERO and FOX 1993b). The 5‘-UTLs of the eight major mitochondrial mRNAs vary greatly in sequence features, complexity, and length from the 54nucleotide COX2 leader to the 954nucleotide COB leader (reviewed in GRIVELL 1989). Outside of the common feature of A/U richness, shared sequence or structural features in S. cereuisiaemitochondrial5’-UTLs have not so far been discerned. The COX2 and COX3 5‘UTLs have been demonstrated to be sufficient to direct the translation of other structural genes, including noncognate mitochondrially encoded genes (MULERO and F o x 1993b), and the synthetic ARGP reporter gene (STEELE et al. 1996; H. M. DUNSTAN and T. D. FOX, un-