Compilation and analysis of group II intron insertions in bacterial genomes: evidence for retroelement behavior

Abstract

Group II introns are novel genetic elements that have properties of both catalytic RNAs and retroelements. Initially identified in organellar genomes of plants and lower eukaryotes, group II introns are now being discovered in increasing numbers in bacterial genomes. Few of the newly sequenced bacterial introns are correctly identified or annotated by those who sequenced them. Here we have compiled and thoroughly analyzed group II introns and their fragments in bacterial DNA sequences reported to GenBank. Intron distribution in bacterial genomes differs markedly from the distribution in organellar genomes. Bacterial introns are not inserted into conserved genes, are often inserted outside of genes altogether and are frequently fragmented, suggesting a high rate of intron gain and loss. Some introns have multiple natural homing sites while others insert after transcriptional terminators. All bacterial group II introns identified to date encode reverse transcriptase open reading frames and are either active retroelements or derivatives of retroelements. Together, these observations suggest that group II introns in bacteria behave primarily as retroelements rather than as introns, and that the strategy for group II intron survival in bacteria is fundamentally different from intron survival in organelles.

Figure 1

Examples of group II intron insertions in bacterial genomes. Orange color denotes group II intron RNA domains (stem–loops) and RT ORF (outlined box). Blue boxes represent ORFs that are interrupted by the intron; the purple box indicates an IS element interrupting an intron. Arrows show gene orientations. GenBank accession numbers with correct intron boundaries are on the left, and numbering in the diagram is in kilobases according to the GenBank entries. Introns with similar insertions are shown on the right. (A) The intron interrupts an ORF. (B) The intron has apparently inserted between ORFs. (C) The intron has inserted after a putative teminator structure. (D) It is unclear whether or not the intron is inserted into an ORF. (E) The intron has inserted into an ORF but in the wrong orientation. (F) The intron is truncated by an IS element. (G) The intron is internally deleted.

Figure 2

Multiple natural insertion sites for group II introns. For each intron, the flanking sequence is shown for 50 bp upstream and 50 bp downstream, with the intron sequence abbreviated gtgcg---ac (or a variation). GenBank accession numbers are on the right, with redundant entries in parentheses. When the same intron is inserted into multiple sites, a consensus sequence is shown beneath the set. Color shadings denote introns (orange), the experimentally determined boundaries of the L.l.I1 homing site (purple), 5S sequence (green) and the terminator stem–loops (blue). T residues following the terminator stem–loop sequences are in upper case. Abbreviations for introns and intron fragments are according to Tables 1–3. (A) Alignment of the flanking sequences of the L.l.I1 intron. The L.l.I1 intron copies are ≥99% identical. (B) Alignment of the flanking sequences of S.f.I1, E.c.I4 and Y.p.F1. Intron sequences are >93% identical. (C) Alignment of flanking sequences of introns of bacterial class C. All introns are inserted after potential terminator sequences (blue shading and upper-case T residues). (D) Example RNA secondary structure of the terminator for the B.h.I1 intron.

Figure 3

Model for the spread of the E.c.I4 intron among three species and three homing sites. All introns in the family are >93% identical. See text for description.