The Group I Intron Encoded Homing Endonuclease IScel

I knew that I-Scel could not be similar to the bacterial restriction enzymes for two reasons. First, E. coli survived my plasmid expression experiments. Second, more importantly, during mitochondrial crosses, the novel intron copy was always inserted at the appropriate location, not elsewhere in the mitochondrial genome. Therefore, my enzyme had to be specific enough to cleave only one site in the mitochondrial genome. The parallel with HO was also suggestive (Kostriken and Heffron 1984). My priority was, therefore, to map the recognition site of I-Scel and estimate its recognition specificity. As this had to be done in vitro, not in vivo, purification of I-Scel was needed.

I engaged in this task using laborious classical protein purification methods, without any knowledge of the proper assay conditions that we determined much later (Monteilhet et al. 1990). None led to a completely pure fraction but some preparations were sufficiently depleted of contaminating nucleases for in vitro assays. Using these fractions, my first idea was to determine the mode of cleavage of I-Scel. This was done using terminally labeled DNA fragments, and the enzyme was shown to generate a 4bp staggered cut with 3' overhangs (Colleaux et al. 1988). On each DNA strand, hydrolysis occurs immediately upstream of the scissile phosphate groups, leaving 3' OH and 5' phosphate ends that can be religated directly with DNA ligase and ATP (Fig. 2). This mode of cleavage has now been shown as a general characteristic of the LAGLIDADG subclass of intron-encoded homing endonucleases (see Chevalier et al., this Vol.). It is also the mode of cleavage of most inteins. We determined much later (Perrin et al. 1993) that the double-strand cleavage activity of I-Scel results from two successive single-strand nicks during the same reaction, the bottom strand being cleaved first. It took 10 more years before this cleavage preference could be explained in atomic detail (Moure et al. 2003). In 1993, we had also determined that the turnover number of I-Scel reactions is very low, the enzyme being inhibited by one of the two cleaved products.

In 1987, determination of its recognition site was the next step to understanding the specificity of I-Scel. This was done using in vitro cleavage assays on mutant sites generated by oligonucleotides synthesized with one degenerate position at a time (Colleaux et al. 1988). The surprise was that the recognition site of I-Scel extended over 18 nucleotides, from position -7 to position + 11 relative to the intron insertion site (Fig. 2). This was unprecedented. Even if the specificity of recognition of I-Scel is less than 418 (because some nucleotide substitutions are tolerated), this was much higher than anything known before. Only HO had a similarly long recognition site.

With I-Scel, back in 1988, the route was opened to cleave entire genomes at a single, predetermined site (see Sect. 9, below). However, before going in to that, the important point for the time was that the recognition site of I-Scel was extending on both sides of the intron insertion site, explaining why the intron-plus DNA is not cleaved, and consequently how mitochondria escape self-destruction (Fig. 2). The mechanism of intron homing was now clear (Fig. lb), but there was no explanation of its very existence. A unique self-splicing group I intron of yeast mitochondria, conferring no phenotype to the cell, was encoding an extremely specific DNA endonuclease only for the sake of its own propagation.

0 0

Responses

  • eligio
    What is i_scel enzyme?
    5 years ago

Post a comment