The first, indirect clue to the functional role of the a intron-encoded protein came from its evolutionary conservation in other yeast species. In 1981,1 had started to screen a large collection of yeast isolates (65) representing numerous (48) distinct species (their phylogeny was unclear at this time) for the possible presence of homologues to the a intron of S. cerevisiae. Surprisingly, all the Kluyveromyces species responded positively whereas only few Sac-charomyces did, and none of the other yeasts. Alain Jacquier, who started his Ph.D. work in my lab at this time, sequenced the corresponding piece of mitochondrial DNA in Kluyveromyces thermoloterans and discovered that, not only was an intron inserted within the mitochondrial rRNA gene at the exact same site as the a intron in S. cerevisiae, but there was also a long open reading frame at the same position (Jacquier and Dujon 1983). We now know that K. thermotolerans and S. cerevisiae are farther apart in terms of molecular evolution than man and fishes (Dujon et al. 2004, and unpubl. data). However, even without this notion, the evolutionary conservation of the a intronic reading frame in K. thermotolerans played a considerable role in the discovery of the first homing endonuclease, helping me to design the artificial gene (see Sect. 6, below).
The second clue for the direct action of the a intron-encoded protein in the propagation of its own intron in crosses came from an observation made several years earlier by Julius Subik in Bratislava (Subik et al. 1977). I was very intrigued by his results because he had isolated a mn mutant from a strain that I showed clearly possessed the a intron, whereas all previous mn mutants were missing it. The sequence once again was immediately informative (Jacquier and Dujon 1985). The intron and flanking exons were normal, but a single-base deletion was obvious in the middle of the intron reading frame, generating a -1 frameshift mutation leading to a stop codon immediately downstream. This was better than I could have ever expected. I called this mutation (for deficient) to differentiate it from the mn mutations which were single nucleotide substitutions in the intron-less gene next to the site of insertion of the intron (see Sect. 2, above). The md mutation was the proof that the translation product of the a intron was necessary for the "mobility" of this intron.
Together with polymorphic variations in exons used as molecular markers, this mutant also allowed us to show that the insertion of the intron in the in-tron-less gene was determining the co-conversion of the flanking exon sequences with an efficiency decreasing with distances. The ancient "polarity" of recombination was explained by a bidirectional repair mechanism initiated at the site of the intron insertion. After the discovery of I-Scel and its exceptional cleavage specificity (see Sect. 7, below), I used the word homing to designate the "transposition" of mobile introns to cognate sites, hence the term homing endonucleases now widely used.
The last clue about the mechanism responsible for the propagation of the a intron in crosses was the demonstration that a double-strand break is transiently formed in mitochondrial DNA during crosses between and yeast strains. The idea of a break was not novel (see Sect. 2, above). However, with the analogy to the double-strand break generated by the HO endo-nuclease to initiate mating-type switching in yeast (Strathern et al. 1982), we again insisted on the similarity to the a intron propagation (Dujon and Jacquier 1983).
In these years, the competition between my laboratory and that of Ronald Butow was fierce. Using properly synchronized mating populations with very sensitive molecular methods, they were able to demonstrate the transient appearance of a double-strand break at the intron insertion site of the intron-less gene in crosses with a wild-type strain (Zinn and Butow 1985). Meanwhile, my laboratory had also observed the break using less sophisticated techniques (Dujon et al. 1985). The break was specific to the wild-type DNA after crosses with a wild-type strain but was absent from mn DNA. Using an imaginative combination of mutagenesis of the a intron during its transfer to the intron-less gene, Butow's group was able to mutate the intronic reading frame, creating mutant strains that had lost the "polarity" of recombination (Macreadie et al. 1985). This was not different from the md mutation that nature had spontaneously created in Subik's strain (Jacquier and Dujon 1985). However, with their intron mutants, they showed that the double-strand break was not formed at the site.
How to demonstrate the exact role played by the a intron-encoded protein in the formation of the DNA break? By the end of 1983, before any of these results were published, I had speculated that, for a number of logical reasons, the intronic protein itself, and not cellular cofactors, had to be the active endo-nuclease. Consequently, I had decided to engineer the intronic reading frame to express it in E. coli.
Was this article helpful?