Lightning strikes twice: intron-intein coincidence.

‘‘Multiple introns, and the prospect that these occur within several genes in the same metabolic pathway, suggest a regulatory role for splicing . . . ’’ (1). That statement was made more than a decade ago by our colleagues and ourselves in reference to the phage T4 group I introns, all three of which reside in genes involved in nucleotide metabolism. The intervening years were distinguished by a striking absence of any demonstration of a regulatory role for these introns. Furthermore, they seem not to confer any selective advantage to T4 phage, and, despite the discovery of more introns in genes of DNA metabolism, the question of ‘‘why exclusively in those genes?’’ has been all but forgotten. Now we are faced with a report in this issue of the Proceedings of an unrelated phage and a different type of splicing element, an intein; not only does this element coexist with a group I intron in the same gene but also the gene is one of nucleotide metabolism (2). With lightning apparently striking twice in the same place, we are again forced to confront our doubts about pure chance. What might account for the colocalization of two different types of intervening sequence in the same gene on the same metabolic pathway? The now familiar group I introns self-splice at the RNA level (3), whereas inteins, which are in-frame protein fusions, selfsplice at the protein level (ref. 4; Fig. 1). The first inteins were described just a few years ago, and immediately examples emerged in all three biological kingdoms, the archaea, bacteria, and eukarya (5–7). The recent identification of more than 50 inteins, mostly by sequence comparisons (refs. 4, 8, and 9; New England Biolabs Intein Database Intein Registry at http:yywww.neb.com; S. Pietrokovski at http:yywww.blocks. fhcrc.orgy;pietroyinteins) has shown them to be present in a wide range of organisms. However, until now, none had been found in bacteriophage. Until, that is, the above-mentioned report, which identifies an intein and a group I intron in the gene for a putative ribonucleotide reductase subunit (the bnrdE gene) in the Bacillus subtilis bacteriophage SPb (2). Lazarevic et al. (2) also identified a group I intron in the neighboring bnrdF gene, encoding a second putative subunit of SPb ribonucleotide reductase. The bnrdE and bnrdF introns are in the same general family as the three group I introns of phage T4 (1, 10) and the introns in Bacillus phages b22 (11) and SPO1 and three of its close relatives (12). Remarkably, all but one of the introns and the single intein so far identified in bacteriophage are located in genes involved in DNA metabolism (10–13). The SPO1 intron is in the DNA polymerase gene, whereas b22 and T4 have introns in their thymidylate synthase genes. The remaining T4 introns are in nrdB, which encodes a small subunit of ribonucleotide reductase, and sunY, renamed nrdD, encoding an anaerobic ribonucleotide reductase. Given the small number of group I introns so far discovered in bacteriophage genes and the apparently very low occurrence of inteins in these organisms, these findings are provocative indeed. Although self-splicing introns in bacteria and lower eukaryotes are not confined to genes of DNA metabolism, but rather are situated in a different spectrum of loci, including tRNA, rRNA, and energy metabolism genes (14), the genes playing host to inteins are once again heavily biased toward DNA metabolism (refs. 4, 8, and 9; New England Biolabs Intein Database Intein Registry at http:yywww.neb.com; S. Pietrokovski at http:yywww.blocks.fhcrc.orgy;pietroy inteins). Approximately 70% of known inteins are confined to such functions. They include DNA polymerases, helicases, gyrases, RecA recombinase, and, once again, ribonucleotide reductases, both aerobic and anaerobic. In addition to splicing at the RNA or protein levels, many group I introns and inteins are mobile genetic elements at the DNA level, by virtue of endonucleases encoded within them (Figs. 1 and 2). These endonucleases cleave intron-minus or intein-minus alleles of their cognate genes, initiating a unidirectional gene conversion event that results in copying of the intein or intron DNA (Fig. 2). There has been much discussion of the evolution of mobile group I introns and inteins. It has been argued that endonuclease genes are the ancestral mobile elements that