Active self-splicing group I introns in 23S rRNA genes of hyperthermophilic bacteria, derived from introns in eukaryotic organelles

Abstract

Group I introns are common in the 23 rRNA genes of mitochondria and chloroplasts. Often, they encode "homing endonucleases," which target highly conserved gene sequences and drive interorganellar intron mobility, even across species and genus lines. Most bacterial 23S rRNA genes show these same endonuclease-sensitive target sequences. However, only two bacterial 23S rRNA genes are known to contain group I introns: that of Simkania negevensis [Everett, K. D., Kahane, S., Bush, R. M. & Friedman, M. G. (1999) J. Bacteriol. 181, 4734-4740], where the intron is not spliced and probably limits growth, and that of Coxiella burnetii [Seshadri, R., et al. (2003) Proc. Natl. Acad. Sci. USA 100, 5455-5460], where no direct evidence of splicing exists. Both bacteria are intracellular parasites and might have acquired introns from eukaryotic hosts. Here we provide direct evidence for splicing, and evolutionary evidence for mobility, of group I introns in the 23S rRNA genes of several free-living hyperthermophilic bacteria of the genus Thermotoga. These bacteria do not live closely with eukaryotes, but phylogenetic analyses suggest that their introns were also acquired from eukaryotic (probably algal) organelles. In vivo, their introns must be spliced at temperatures approaching 90 degrees C, making them the most thermostable natural ribozymes so far described. We demonstrate that at least some of these introns can also self-splice in vitro.

Fig. 4.

Minimum evolution tree of 16S rRNA sequences and intron sequences. (a) Based on logdet distances estimated from 16S rRNA from all Thermotogales strains included in the study, rooted by Aquifex aeolicus.(b) Based on Kimura 2P distances estimated from 16S rRNA from the strains most closely related to T. neapolitana NS-E. (c) Based on Kimura 2P distances estimated from the intron sequences from E. coli position 1931. All trees were estimated in paup* (19) by using 10 random stepwise additions and TBR branch swapping. Numbers on nodes indicate number of times the node was recovered in 100 bootstrap replicates. Presence of a group I intron in the 23S rRNA is indicated in a and b.

Fig. 1.

Folding of the two Thermotoga introns according to Cech et al. (37). (a) The intron from T. neapolitana NS-E, Tna.bL1931. (b) The intron from T. subterranea SL1, Tsu.bL1926. Sites where polymorphisms were observed among the additional strains that contained an intron similar to Tna.bL1931 are indicated by stars on the structure along with the mutation observed. (Inset) Insertion sites of the two introns relative to part of domain IV of the T. maritima MSB8 23S rRNA, numbered according to the E. coli 23S rRNA. The intron positions in domain IV are found in sequences that form the interface with the 30S ribosome subunit (5).

Fig. 2.

RT-PCR on RNA isolated from T. neapolitana NS-E. Lane 1, RNA treated with DNase before reverse transcriptase step; lane 2, RNA treated with RNase before reverse transcriptase step; lane 3, RNA; lane 4, no RNA in reverse transcriptase step; lane 5, T. neapolitana NS-E DNA; lane 6, negative control PCR step. A 972-bp fragment in addition to the 273-bp fragment could be amplified from large amounts of fresh RNA (data not shown), suggesting that this PCR product represents either small amounts of DNA contamination or low levels of unspliced rRNA. DNA-dependent PCR on the RNA (i.e., no reverse transcription step) gave only the 972-bp band (data not shown).

Fig. 3.

(a) RT-PCR on in vitro splice products obtained by incubating transcripts of the Tna.bL1931 intron (lanes 1–14) and Tsu.bL1926 intron (lanes 15–24) with 273 bp of flanking exon sequence at various temperatures. Lanes 1 and 15, incubated with splicing buffer on ice; lanes 2 and 16, 30°C; lanes 3 and 17, 40°C; lanes 4 and 18, 50°C; lanes 5 and 19, 60°C; lanes 6 and 20, 70°C; lanes 7 and 21, 80°C; lanes 8 and 22, 90°C; lanes 9 and 23, 100°C; lanes 10 and 24, 105°C. Incubation time was 5 min for the incubations at 30–90°C and 2 min for the incubations at 100°C and 105°C. Lane 11, heating the Tna.bL1931 intron–exon transcript to 100°C followed by cooling on ice before addition of splice buffer; lanes 12–14, the Tna.bL1931 intron–exon transcript incubated at 100°C for 5 s (lane 12), 10 s (lane 13), and 30 s (lane 14). (b) Time-course analysis of the self-splicing reaction of Tna.bL1931. RNA was analyzed directly after incubation at 60°C, 90°C, and 100°C at different times. Lane 1, 0 min (incubated with splicing buffer and 0.05 M EDTA on ice); lanes 2–9, incubated at 60°C for 0.5 min (lane 2), 1 min (lane 3), 2 min (lane 4), 5 min (lane 5), 10 min (lane 6), 20 min (lane 7), 30 min (lane 8), 60 min (lane 9); lanes 10–13, incubated at 90°C for 0.5 min (lane 10), 1 min (lane 11), 2 min (lane 12), 5 min (lane 13); lanes 14–16, incubated at 100°C for 0.5 min (lane 14), 1 min (lane 15), and 2 min (lane 16). The full-length transcript is 1,238 nt, the religated exon is 539 nt, and the intron is 699 nt.

Fig. 5.

Maximum-likelihood tree of single-LAGLIDADG homing endonucleases carried by group I introns. The 23 first hits to single-LAGLIDADG homing endonucleases in blast-x searches, with the Thermotoga sequences as probes, was selected to build the phylogeny. Chloroplast and mitochondrial introns are indicated by a “c” or an “m,” respectively. Double-LAGLIDADG homing endonucleases were excluded because they have a higher rate of divergence (26). The tree was estimated by using a JTT+Γ model in phylip version 3.6 (17) with a substitution matrix provided by E. Tillier (personal communication), with 10 random additions of the sequences and global rearrangements. The α-parameter was estimated in puzzle version 4.0 (18). Values at nodes indicate number of times the node was recovered in 100 bootstrap replicates (bold numbers) or puzzle support (italic numbers). Insertion position in the 23S rRNA of the group I intron carrying the endonuclease is indicated where this was given in the GenBank entry or available at www.rna.icmb.utexas.edu (21). Specific insertion positions were not available for m. Chlorella vulgaris, m2. Chlorella vulgaris, m. Acanthamoeba castellanii, m2. Acanthamoeba castellanii, and m. Chaetosphaeridium globosum.