Group II self-splicing introns in bacteria

Abstract

Like nuclear premessenger introns, group II self-splicing introns are excised from primary transcripts as branched molecules, containing a 2'-5' phosphodiester bond. For this reason, it is widely believed that the ribozyme (catalytic RNA) core of group II introns, or some evolutionarily related molecule, gave rise to the RNA components of the spliceosomal splicing machinery of the eukaryotic nucleus. One difficulty with this hypothesis has been the restricted distribution of group II introns. Unlike group I self-splicing introns, which interrupt not only organelle primary transcripts, but also some bacterial and nuclear genes, group II introns seemed to be confined to mitochondrial and chloroplast genomes (reviewed in ref. 6). We now report the discovery of group II introns both in cyanobacteria (the ancestors of chloroplasts) and the gamma subdivision of purple bacteria, or proteobacteria, whose alpha subdivision probably gave rise to mitochondria. At least one of these introns actually self-splices in vitro.

The dispersal of five group II introns among natural populations of Escherichia coli

“…Full length copies of E.c.I1 and E.c.I3 with flanks were PCR amplified from genomic DNA of ECOR43 and ECOR52, respectively, with Pfu DNA polymerase+ Amplification of E.c.I1 was with the primers 59-CAGGGAAAGATCACGGTG and 59-CTCTACTTTCCAGGCTTATCTGTAATCGGG, with an annealing temperature of 48 8C+ Amplification of E.c.I3 was with the primers 59-CGGACCGGAAAGGTATCCATCCACAAAATC and 59-GAGGTCAGATCAACTTTCCAGGGCAACAG with an annealing temperature of 55 8C+ Products were cloned and both strands were sequenced+ Sequence data are deposited in GenBank (Accession Nos+ AF512502 and AF512503)+ Both sequences are completely identical to partial sequences reported in Ferat et al+ (1994), and the E.c.I3 sequence is 98+4% identical to an intron copy in Shigella (GenBank Accession No+ AF348706), except that the Shigella intron is truncated by 34 bp at its 59 end+…”

“…Five ECOR strains were previously reported to contain E.c.I1 (ECOR17, 22, 31, 43, and 67;Ferat et al+, 1994)+ Based on southern hybridizations (not shown), each strain contains one full-length intron in the expected ISEc1 homing site+ In addition, each strain probably contains at least one fragment, as suggested by weak hybridizations of unexpected size (Table 1)+ Interestingly, the intron insertion site is 15 bp from the 59 end of ISEc1, suggesting an abbreviated homing site extending to only Ϫ15 rather than the typical Ϫ25+ To investigate the boundary of the homing site, inverse PCR was used to obtain sequence of the upstream flank+ Two different 59 sequences were obtained+ ECOR17, 22, and 67 have an upstream sequence "a," which has no match in GenBank, and ECOR31 and ECOR43 have the sequence "b," which matches RhsE sequence (Fig+ 2A)+ Two more fragments of E.c.I1 were found in GenBank entries for ECOR11 and ECOR45, each with different upstream sequences (Fig+ 2A)+ Together, the four upstream sequences suggest no homing site conservation beyond Ϫ15, although it remains possible that the ISEc1 transposed after intron insertion+ All 72 ECOR strains were screened by southern blots with a full-length E.c.I1 probe, but because the E.c.I1 and E.c.I2 probes cross-hybridized, they will be discussed together below+…”

“…RmInt1 of Sinorhizobium meliloti is the only other group II intron to be investigated within a population (MartinezAbarca et al+, 1998;Muñoz et al+, 2001)+ Like E. coli, RmInt1 is variable in copy number, with 0-11 copies in different strains; however, unlike E. coli introns, RmInt1 is quite prevalent, with most S. meliloti strains containing at least one full-length RmInt1 copy+ Interestingly, many intron-containing S. meliloti strains contain host IS elements unfilled with intron, as deduced from hybridizations (Martinez-Abarca et al+, 1998)+ Although the hybridizations were not determined to represent unmutated homing sites, the combined observations in S. meliloti and E. coli suggest that it may be common for intron-containing bacteria to contain unfilled homing sites+ Evidence for independent mobility of the introns Because all five E. coli introns are located in mobile DNAs, it has been speculated that their dispersal may rely on host element mobility rather than independent mobility (Ferat et al+, 1994)+ Several observations in this study argue for independent mobility, the clearest being E.c.I4 composition in the ECOR38/39/40/41 grouping, in which each strain has a different intron composition+ Importantly, there are three distinguishable homing sites: an IS911 site, a IS629 site in ECOR9,38,39,, and a polymorphic IS629 site in ECOR40+ Therefore, E.c.I4 has inserted independently into new homing sites at least twice within the ECOR38/39/40/41 grouping+ The E.c.I4 composition in ECOR9 also supports independent mobility, because the intron is located in both IS629 and IS911 sites, whereas no related ECOR strains contain the intron+ Similarly, E.c.I1 appears to have independently inserted into four different ISEc1 homing sites (Fig+ 2A), although it is also possible that the ISEc1 element transposed after intron insertion+ For E.c.I2, E.c.I3, and E.c.I5, there is no information about exon polymorphisms, and therefore no direct evidence for independent mobility; however, their distribution suggests that they may also be independently mobile at some frequency+…”

“…Group II introns are retroelements found in the genomes of bacteria, mitochondria, and chloroplasts (Bonen & Vogel, 2001;Belfort et al+, 2002)+ Intron mobility occurs by site-specific insertion into defined target sites (retrohoming), or, at much lower frequencies, insertion into noncognate sites (retrotransposition)+ Group II intron retroelements consist of a self-splicing intron RNA structure, which encodes a multifunctional protein that facilitates intron splicing and intron mobility+ The intron-encoded protein is a reverse transcriptase (RT) containing the seven domains common to all RTs, but also contains a domain X that aids in splicing of the intron in vivo, and a Zn domain, which provides a nuclease activity used in mobility+ The mobility mechanism is well established and requires catalytic activities of both the self-splicing intron RNA and the intronencoded protein+ First, the intron reverse splices into the sense strand of the double-stranded DNA target site, and then the Zn domain cleaves the antisense strand and the RT reverse transcribes the intron using the cleaved DNA as a primer (Bonen & Vogel, 2001;Belfort et al+, 2002)+ Group II intron retroelements were discovered relatively recently in bacterial genomes+ The first identification was achieved through PCR screening using degenerate primers (Ferat & Michel, 1993), which identified introns in a cyanobacterium (Calothrix ) and a proteobacterium (Azotobacter )+ Because these bacteria are relatives of the ancestors of eukaryotic organelles, the finding suggested that mobile introns might have spread from bacteria to organelles+ By now, due to the many bacterial genomes sequenced, it is clear that group II introns are widely dispersed throughout the eubacterial kingdom (Martinez-Abarca & Toro, 2000b;Dai & Zimmerly, 2002)+ Bacterial group II introns differ in several respects from organellar introns+ Unlike organellar introns, bacterial introns are not located in conserved genes, but are associated with mobile DNAs+ The bacterial introns are often inserted outside of genes, and one subclass of introns inserts exclusively after transcriptional terminators rather than into ORFs+ In contrast to organellar introns, all known bacterial group II introns encode RT ORFs and are either active retroelements or inactive derivatives+ Finally, over half of bacterial group II intron sequences are fragments, suggesting that the introns survive by constantly inserting into new locations rather than assuming a relatively stable position in a conserved gene, as occurs in organelles+ Together these observations led us to suggest that group II introns in bacteria are adapted to function mainly as retroelements (Dai & Zimmerly, 2002)+ Group II introns were discovered in Escherichia coli in 1994, and were named IntA, IntB, IntC, and IntD (Ferat et al+, 1994)+ Dot blot screening of strains of the ECOR collection indicated that only 18 of the 72 ECOR strains contain an intron+ Full sequence was initially reported only for IntB, with partial sequences reported for IntA, IntC, and IntD (Ferat et al+, 1994)+ Subsequently IntD was sequenced independently in a Shigella pathogenicity island (Rajakumar et al+, 1997), and in E. coli plasmid pB171, where it is associated with a second homing site …”

“…Full sequences of E.c.I1 and E.c.I3 were obtained by PCR amplification of the introns based on predicted flanking exon sequences and host strains (Ferat et al+, 1994), followed by cloning and sequencing (see Material and Methods)+ Based on the completed sequences, all five introns can be described+ E.c.I1, E.c.I2, and E.c.I3 are related to each other and belong to the previously defined subclass "bacterial class D+" (Intron subclasses are based on phylogenetic groupings of the intron-encoded ORFs, but each subclass also has a distinct intron secondary structure; Toor et al+, 2001;Zimmerly et al+, 2001+) Typical of bacterial class D, the introns are IIB-like in RNA structure (Fig+ 1A,B), and the ORFs lack a Zn domain+ Although the Zn domain is required for mobility of Lactococcus Ll+ltrB (Cousineau et al+, 1998), many bacterial introns do not contain Zn domains and at least one of these is efficiently mobile nevertheless (Martinez-Abarca & Toro, 2000a)+ E.c.I1 and E.c.I2 are 67% identical in DNA sequence, whereas E.c.I3 is much less related, with 44% identity to E.c.I2 over its ORF+ E.c.I1 sequence from ECOR43 has a stop codon in RT domain 6 and a frame shift in domain X, suggesting loss of mobility and splicing functions+ E.c.I1 and E.c.I2 are inserted into an ISEc1 element, which is itself contained within an Recombination hot spot (Rhs) element (Ferat et al+, 1994)+ ISEc1 was previously called an H-repeat, but has been renamed ISEc1 because of its resemblance to IS elements (Mahillon & Chandler, 1998)+ The Rhs element is an ;6-kb DNA found in five copies in the sequenced K-12 genome+ There are at least eight varieties of Rhs elements (A-H), whose core sequences are .70% identical, but which differ in organization and lengths of spacer elements (Zhao et al+, 1993;Bachellier et al+, 1996;Wang et al+, 1998)+ At the 39 terminus of most Rhs elements is an ISEc1 copy, which is flanked by 11 bp inverted repeats+ E.c.I1 is inserted 4 bp after the upstream inverted repeat, whereas E.c.I2 is inserted near the 39 end of the ISEc1 ORF+ E.c.I3 is also located in an IS element, IS679+ E.c.I4 belongs to the subclass "bacterial class A," and its only other relative is the essentially identical intron in the closely related bacterium Shigella (99+4% identity)+ A possible secondary structure is shown in Figure 1C+ According to information in the databases, the intron has two homing sites in IS629 and IS911 elements, which share only 60% identity (Fig+ 2D)+ (We use the term "homing site" to denote the insertion site for all five introns in this study+ Although homing has not been experimentally demonstrated, it is suggested because the introns are repeatedly found within the same flanking sequences+)…”