The Problems of Mitochondrial Intronic Reading Frames and Their Products

Two important points need to be mentioned here to fully appreciate the situation in 1980-1981.

First, the w intronic reading frame was not written using the universal genetic code. Instead, it contains five UGA "stop" codons, shown to encode tryptophan in mitochondria (Barell et al. 1979; Macino et al. 1979), and eight other codons subsequently shown to be non-universal (Bonitz et al. 1980). With these differences from the normal code, I had no hope of expressing an active protein in E. coli or the in vitro systems of the time.

Second, genetic experiments using novel respiratory deficient mitochondrial mutations had led to the idea that some of these mutations affected the expression of their gene in an unusual manner (Slonimski and Tzagoloff 1976). With the discovery of introns, it became rapidly clear that the two genes concerned (encoding apocytochrome b and subunit I of cytochrome oxidase, respectively) were mosaic and that the unusual mutations were in their introns (Slonimski et al. 1978; reviewed in Dujon 1979,1981) and prevented splicing (Church et al. 1979). Interestingly, some of them were complemented in trans by a wild-type intron, as if an intron product existed. In particular, the presence of a long open reading frame in the sequence of the second intron of the cytochrome b gene strongly suggested that the trans-active component necessary for proper gene expression, and altered in the mutants, was the intron translation product, inferring that the intron product would help the splicing of its own intron RNA. The putative intron-encoded protein, called a "matu-rase" (Jacq et al. 1980; Lazowska et al. 1980), was linked in-frame with its upstream exon, a situation now known to be common to many intron-encoded proteins but different from the w intron. In addition, the deduced sequence of the cytochrome b maturase showed very little resemblance with the potential product of the w intron, except for a few short C-terminal motifs.

Indeed, it took a few more years, and the sequencing of several additional introns, to recognize that many intron-encoded proteins share two typical, but short, motifs that were originally designated PI and P2 (Michel et al. 1982) or LAGLI-DADG (Hensgens et al. 1983) from their amino-acid consensus sequence. This motif, in reality, is part of the dodecapeptide strings that form helices located at the interface between two domains and contain the two conserved catalytic acidic residues characteristic of this large family of proteins, now commonly referred to as LAGLIDADG homing endonucleases (see Chevalier et al., this Vol.).

Despite the numerous publications that appeared in the years 1981-1984 indicating the generality of the maturase model in yeast and other fungi (for review, see Lambowitz and Perlman 1990), the model did not apply to the œ intron. The first reason was that RNA splicing of this intron did not require any mitochondrially translated protein. Indeed, the œ intron was the second intron shown to be self-splicing in vitro (Tabak et al. 1984), soon after the discovery of the catalytic activity of the Tetrahymena intron (see Sect. 4, below). The second reason was that the maturase model concerned RNA splicing but did not address the polarity of recombination in crosses, which was the hallmark of the œ intron. The discovery of I-Scel had to await yet another strange turn.

0 0

Post a comment