![[Feedback]](/icons/banner/feedbackACT.gif){kind=icon}
![[For Subscribers]](/icons/banner/subscriberACT.gif){kind=icon}
![[Archive]](/icons/banner/archiveACT.gif){kind=icon}
![[Search]](/icons/banner/searchACT.gif){kind=icon}
![[Contents]](/icons/banner/tocACT.gif){kind=icon}
The Biotechnology Centre of Oslo and School of Pharmacy, University of Oslo, Oslo, Norway
Received 17 January 2005/ Accepted 12 April 2005
 |
ABSTRACT |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
subunit of DNA polymerase III. Another striking finding is that the observed 3' splice site of B.c.I4 occurs 56 bp after the predicted end of the intron. This apparently unusual splicing mechanism may be related to structural irregularities in the 3' terminus. Finally, we also show that the intergenic introns of B. cereus ATCC 10987 are transcribed with their upstream genes and do splice in vivo. |
INTRODUCTION |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
20 years ago (33), they were first found in bacteria 10 years later (15) and lately have been identified in archaea of the Methanosarcina genus (11, 55). About 25% of the completely sequenced microbial genomes, covering a diverse range of bacterial phyla, contain one or more introns (either full length or fragmented). A compilation of bacterial group II introns showed that these elements are often inserted in intergenic regions, and when located inside genes, they are rarely found within highly conserved or housekeeping genes (10). Usually, bacterial group II introns are located on mobile DNA elements such as plasmids, insertion elements, transposons, or pathogenicity islands, which could account for their spread among bacteria (4, 10).
Group II introns typically consist of a catalytic RNA containing an internal protein-encoding open reading frame (ORF), although many ORF-less introns exist in eukaryotic organelles. Functional introns are able to excise out of RNA transcripts (self-splicing), and insert (reverse splice) into identical intronless DNA sites (process called homing) or into novel (ectopic) genomic locations but at a much lower frequency (retrotransposition). The homing process is highly site specific and occurs at target regions spanning 30 bp around the insertion point (18, 51). Selection of splice sites is determined by base pairings between three motifs in the intron RNA (exon-binding sequences EBS1, EBS2, and EBS3 or 
) and the complementary sequences in the flanking exons (intron-binding sequences IBS1, IBS2, and IBS3 or ', respectively), and these pairings are required for both splicing and reverse splicing (insertion). The splicing reactions are intrinsically catalyzed by the RNA part, but the intronic protein participates in vivo in both the splicing and insertion events (see references 5 and 55 for reviews). Because of the similar in vitro splicing mechanism, including formation of a lariat structure, group II introns are thought to be the ancestors of the nuclear spliceosomal introns of eukaryotes. The secondary structure of the group II RNA is made up of six domains linked by tertiary interactions, where domain V is the presumed catalytic core and domain VI contains a bulged A that is the branching point of the lariat (34, 35, 42, 54). The structure is used to divide the group II introns into subclasses (54). The intron-encoded protein (IEP), located in domain IV, is a multifunctional protein that can have three functional domains: a reverse transcriptase (RT) domain for synthesis of DNA strand upon insertion, a maturase (X) domain involved in splicing, and an endonuclease (En or Zn) domain for target DNA cleavage, although the latter region is lacking in most IEPs. In between the X and En domains is a DNA-binding (D) region (3, 5, 50, 57).
For many published bacterial genome sequences, the IEPs are annotated, but the complete intron elements are not defined, and it is thus not known whether or not the neighboring genes are interrupted by the introns unless further analysis is done (10). In this study, we have characterized the group II introns present in the genomes of two strains of Bacillus cereus (ATCC 14579 and ATCC 10987), an opportunistic pathogen causing food poisoning (27, 44). B. cereus is a member of the B. cereus group of bacteria that includes, among others, B. anthracis, the mammalian pathogen causing anthrax (45). These organisms are genetically very closely related, and many phenotypic and pathogenic differences are due to the presence of plasmids. To date, the only group II introns known from B. cereus group bacteria are the two introns in B. anthracis that have been defined by sequence analysis, named B.a.I1 and B.a.I2 (10). They are both located on the pXO1 virulence plasmid (40). A recent experimental study has revealed that the B.a.I2 intron has unusual structural features allowing alternative splicing, while B.a.I1 splices normally (47). Here, we present a comparative sequence analysis of group II introns in B. cereus and B. anthracis, including secondary structure predictions, and an experimental analysis by RT-PCR of the in vivo splicing abilities of the B. cereus introns.
|
MATERIALS AND METHODS |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
|
Secondary structures of B. cereus intron RNAs (RT ORF removed) were computationally predicted by constrained folding using the MFOLD version 3.1 package (58) following the consensus structures of the different intron classes described previously (54) and available at the Group II intron database. That is, conserved and identifiable sequence motifs corresponding to a given consensus structure were forced during the folding computation. To resolve local regions where there was no obvious match to consensus motifs, B. cereus intron sequences were aligned using BLASTN (2) with the individual members of a given intron family in order to identify homologous elements. The structures of the B. anthracis B.a.I1 and B.a.I2 introns can be found in this database and in a report by Robart et al. (47).
Interaction between the last intron base (
') and a nucleotide () located between domains II and III is determinant for correct 3' splice site selection (28, 47). In the search for 3' intron boundaries, introns were, if necessary, extended downstream of domain VI based on the requirement for the presence of both -' and EBS3-IBS3 nucleotide pairings, even though the length of the 3' end could disagree with the consensus of the intron family. Indeed, Robart et al. (47) recently showed that both pairings are critical for the functionality of the B. anthracis intron B.a.I2 and that B.a.I2 actually contains an extra 3' nucleotide compared to the previously proposed sequence (10). Potential motifs involved in base pairing between the introns and their exons were identified based on their conserved location in the intron secondary structures and their complementarity to exon boundaries.
When comparing the genomic organization of intron loci, genes were considered as putative orthologues in two bacterial strains if they satisfied a reciprocal best-hit relationship based on TBLASTX (2) searches of the complete gene sets of the two organisms. Hits were judged as significant if the E value was lower than 10–5 and the match covered at least 70% of the length of both genes (intragenic introns not included). B. anthracis chromosomal data were taken from the complete chromosome sequence of B. anthracis Ames (45) (GenBank accession no. NC_003997). Potential rho-independent transcriptional terminators present in the loci were predicted using TRANSTERM (13), and only the terminators with a statistical confidence of 98% or more were selected, as recommended by the authors of the program.
B. cereus group II introns were named according to the nomenclature used previously (10), where B.c. stands for B. cereus and where I and F denote full-length intron and intron fragment, respectively. A letter in parentheses was appended to distinguish different copies of the same intron.
RNA isolation. B. cereus ATCC 10987 and B. cereus ATCC 14579 were grown on Luria Bertani (LB) agar plates at pH 7 and 30°C. An overnight culture (16 h) of LB medium was used to inoculate a 250-ml medium (pH 7, 30°C, and 240 rpm) to give an optical density of 0.05 at 600 nm. Samples (2 to 10 ml) for RNA isolation were taken out after 3, 6, and 8 h. These time points represent the early phase, the mid/late-exponential phase, and the beginning of stationary phase of the growth curve, respectively. Samples were put into an equal volume of ice-cold methanol before they were spun down and frozen to –80°C. Total RNA was extracted using the QIAGEN RNasy Midi kit (QIAGEN Gmbh, Hilden, Germany). DNase treatment was conducted as suggested in the QIAGEN protocol with RNase-Free DNase Set (QIAGEN Gmbh, Hilden, Germany) and also after incubation for 20 min at 37°C with DNase I (Amersham Biosciences, Little Chalfont, United Kingdom).
RT-PCR and PCR amplification. RT-PCR and PCR experiments were conducted to determine whether the group II introns of B. cereus ATCC 10987 and B. cereus ATCC 14579 were able to splice in vivo following a strategy basically identical to that reported previously by Roberts et al. (48; see Fig. 1). For each intron, primers specific for the 5' and 3' exon-intron junctions and the intronic ORF were designed using the Primer3 software available at the Whitehead Institute for Biomedical Research website (49) (http://frodo.wi.mit.edu/cgi-bin/primer3/primer3_www.cgi/), and the specificity of the primer sequences was further assessed by matching them against the complete genome sequences. The primers are listed in Table 2. Note that for the B. cereus ATCC 14579 intron, the reverse primer in the 3' exon had to be placed within the region overlapping a potential rho-independent transcription terminator because the terminator was located only eight bases downstream of the intron (see Fig. 3).
|
|
|
5 µg of total RNA as a template according to the supplier's protocol. A negative control was conducted without addition of reverse transcriptase. A portion of the cDNA and negative control was then amplified by PCR using both the forward and reverse primers at a concentration of 0.4 µM, each deoxynucleoside triphosphate at a concentration of 0.8 mM, and 1 U of Dynazyme (Finnzymes Oy, Espoo, Finland). PCR was primarily run with one denaturation step at 94°C for 5 min, followed by 30 to 38 cycles of a 1-min denaturation step at 94°C, a 50-s annealing step at 58°C, and a 50-s extension step at 72°C, followed by a final extension step of 7 min at 72°C. For the primer pair covering a transcriptional terminator in B. cereus ATCC 14579 the temperature for the annealing step was set to 56.5°C. Products were analyzed by electrophoresis on a 0.8% agarose gel to determine their lengths. PCR fragments of ligated exons were sequenced to ascertain the splicing boundaries. Sequencing was performed as described previously (22). Furthermore, we designed three additional primer pairs for the B.c.I2(a) intron and its insertion locus in B. cereus ATCC 10987, as follows. (i) A second primer pair was designed for the 3' intron-exon junction. The forward primer was the one given in Table 2, while the reverse primer was located at position 2 in an unannotated ORF (ORFX) predicted to be encoded within the 3' exon BCE3693 in the opposite DNA strand (see Fig. 2). The sequence and annealing temperature of the new reverse primer were 5'TGGATGCGATATTTGTAATCATTC3' and 60.1°C, respectively. The expected PCR product length would be 765 bp. (ii) A primer pair linking ORFX to the BCE3694 ORF predicted downstream of BCE3693 was designed. The forward primer (5'GCAATGTACCGTTCCTTTACG3', 59.5°C) was located in ORFX, while the reverse primer (5'CCAACAGCATCCGTAACTATTTC3', 59.9°C) was located in BCE3694, which would make a PCR product of 794 bp. (iii) A primer pair specific for the 5' end of B.c.I2(a)'s host gene was designed. The left (5'ACGAATTTCAACCTGCACCT3', 59.6°C) and right (5'TGTGCCGTTCCATTAACTGT3', 59.1°C) primers were located at positions 284 and 935 within the BCE3690 ORF, respectively, giving a PCR product of 652 bp. In all three cases, the RT-PCR and PCR analyses were done as described above.
|
|
|
RESULTS |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
The most striking finding relates to B.c.I2(a) in the B. cereus ATCC 10987 chromosome. This intron is positioned in the opposite orientation relative to the host gene (split into ORFs BCE3690 and BCE3693) (Fig. 2). The function of the protein encoded by BCE3690+BCE3693 is unknown. A search of the Interpro database (http://www.ebi.ac.uk/interpro/) for conserved motifs in the translated sequence of BCE3690+BCE3693, which is 2,520 amino acids long, returned significant hits matching the bacterial adhesion domain (Interpro identification no. IPR008966) and a domain of unknown function, DUF11 (identification no. IPR001434). The sequence is repetitive and made up of 19 copies of the adhesion domain and 18 copies of DUF11. The adhesion domain (120 amino acids) is found in adhesins, cell surface proteins involved in attachment to host cells during pathogenesis (6). BCE3690+BCE3693 shares weak and partial amino acid sequence similarity (36 to 56%) to a range of long proteins (>1,500 residues) from diverse bacteria, including some of the longest proteins from B. anthracis Ames (BA1618 and BA3601) and B. cereus ATCC 14579 (BC1592, BC2639, BC3546-7, and BC4927), which are annotated as either hypothetical proteins, repeat domain proteins, or cell surface/cell adhesion proteins.
Another remarkable feature of the B. cereus ATCC 10987 chromosome is the presence (in the same orientation) of the B.c.I3 intron within the gene encoding the subunit of the gram-positive specific DNA polymerase III (also known as polC [BCE3856+BCE3859]). The orthologous proteins in B. anthracis (BA3955) and B. cereus ATCC 14579 (BC3816) are 99% identical to the joint translated sequence of BCE3856+BCE3859 in B. cereus ATCC 10987. In fact, the protein is well conserved among gram-positive organisms like Bacillus subtilis, Bacillus halodurans, Listeria species, Enterococcus faecalis, Staphylococcus aureus, and Streptococcus species (amino acid identity of 50 to 74%).
Comparative sequence analysis of B. cereus and B. anthracis group II introns and IEPs. Alignment and comparison of the complete nucleotide sequences of the B. anthracis and B. cereus group II introns revealed that in B. cereus ATCC 10987, B.c.I2(a) (chromosome) and B.c.I2(b) (pBc10987plasmid) are identical and are thus copies of the same intron, B.c.I2, which also shares 89% overall sequence identity to B.a.I2 from the B. anthracis pXO1 plasmid. More surprisingly, B.c.I1(b) is identical to B.c.I1(a) of B. cereus ATCC 14579 (both chromosomal), while B.c.I1(c) (pBc10987plasmid) differs from them by only a single (nonsynonymous) nucleotide difference. B.c.I1(a), B.c.I1(b), and B.c.I1(c) also represent copies of the same element, B.c.I1. These observations suggest that there have been recent movements of introns within and possibly between B. cereus genomes. Furthermore, two subsets of introns could be defined based on sequence homology: B.c.I1 is more closely related to B.a.I1, while B.c.I2, B.c.I3, B.c.I4, and B.c.I5 are more similar to B.a.I2. Overall, the nucleotide sequence identity ranges from 46 to 89% between introns of the same subset (59 to 95% similarity between the IEPs) and from 38 to 45% between elements of the two groups (53 to 62% similarity between the IEPs). The IEP of B.c.I1 seems to lack some elements of the endonuclease (En) domain. The En domain typically consists of two motifs containing conserved histidine residues (consensus E-X1-HH and H-X3-H), characteristic of endonucleases of the H-N-H family. These motifs alternate with motifs containing of one or two pairs of cysteine residues forming a zinc-finger-like region (17, 50). Although a pair of cysteines is present, no obvious H-N-H pattern can be identified (see Fig. S2 in the supplemental material). A sequence search of the Interpro database returned a significant hit to the H-N-H endonuclease domain (Interpro identification no. IPR002711) for all IEPs, except that of B.c.I1. Interestingly, the three B.c.I1 copies are not inserted within genes, as opposed to the seven other B. cereus and B. anthracis introns (Table 1).
Comparative structure analysis of B. cereus and B. anthracis group II intron RNAs. All introns start with the common 5' splice site GUGYG motif. Predictions of the secondary structure of the catalytic RNA (ORF removed) indicated that all B. cereus group II introns could be folded into one of two known structural subtypes within class IIB (Table 1 and Fig. 3). Interestingly, B.c.I1, which is more similar to B.a.I1 than to B.a.I2, folded into the same structural class as B.a.I1, namely, the B1 class (according to the nomenclature of Toor et al. [54]), and the B.a.I2-like introns B.c.I2, B.c.I3, B.c.I4, and B.c.I5 folded into the same class as B.a.I2, i.e., the B2-like class. These observations are in agreement with the coevolution rule, which predicts that group II intron RNAs coevolve with their corresponding ORFs (54).
Potential motifs involved in base pairing between the introns and their exons are shown in Fig. 4. Candidates for EBS3-IBS3 and -' pairings (where IBS3 is the first base of the 3' exon and ' is the last base of the intron) could be found near the base of the domain VI stem for all B. cereus introns except B.c.I4 (Fig. 3 and 4). In a situation similar to that of B.a.I2, the requirements for proper base pairings necessitate B.c.I2 to end four nucleotides after domain VI as opposed to the usual three nucleotides for B.c.I1, B.c.I3, B.c.I5, and all other characterized group IIB introns. Furthermore, like B.a.I2, B.c.I2 also contains structural irregularities such as a C-C mispair at the base of the catalytic domain V stem and an extra stem-loop region at the base of domain I (Fig. 3). Finally, the 3' terminus of B.c.I4 is unusual; ending the intron three or four bases after domain VI would create no complementary EBS3-IBS3 pair (A-C or A-G). The first canonical EBS3-IBS3 pairing (A-U) would require the intron to be extended to eight bases after domain VI, but in this case, there would be no typical -' pair.
|
RT-PCR and PCR amplification of spliced and unspliced products. For all B. cereus ATCC 10987 group II introns except B.c.I2(a) (discussed below), RT-PCR and PCR amplification using primers specific for the junctions between the introns and their 5' and 3' flanking exons (Table 2) generated products of the expected sizes (data not shown). This demonstrates that unspliced RNA was present in the total RNA sample. In agreement with this, PCR products of expected lengths were also obtained when primers designed for the intron-encoded ORFs were used (data not shown). For the intragenic B.c.I2(b), B.c.I3, B.c.I4, and B.c.I5 introns, detection of the intron-exon junctions reveals that there was transcription of their host genes. In the case of the intergenic B.c.I1(b) and B.c.I1(c) elements, for which the primer for the 5' exon was positioned within the gene upstream of the intron, amplification of the PCR products indicates that these introns were actually transcribed together with their upstream genes. Using a primer pair consisting of the primers for the 5' and 3' exons (Table 2), short products having a size corresponding to that expected for fragments lacking the intron could be amplified for all six introns (Fig. 5A). Sequencing of the fragments confirmed that they were made up of ligated exons and, with the exception of B.c.I4, confirmed the locations of the intron boundaries predicted by the comparative sequence and structure analyses described above (Fig. 3 and 4) and that coding exons were in frame. These results show that spliced RNA was present and that the introns, including the intergenic ones, are functional and able to splice in vivo under the growth conditions applied and at the time point(s) tested. The results also suggest that the intron-associated genes annotated as hypothetical ORFs (Table 1) are real genes.
|
-' (A-U) and EBS3-IBS3 (A-U) pairings could occur (Fig. 3 and 4).
In the case of B.c.I2(a), which is positioned in the opposite orientation relative to its host gene (BCE3690+BCE3693 [Fig. 2 and see Fig. S1 in the supplemental material]), no PCR product was detected using any of the primer pairs designed for the intronic ORF or the intron-exon junctions when cDNA synthesis was initiated using the reverse primers by RT-PCR. Using a primer pair designed for the 5' end of the host gene (in BCE3690), a PCR product of an expected size of 650 bp was obtained, indicating that the host gene of B.c.I2(a) was expressed (data not shown). The size of the whole transcript would be 10.4 kb, B.c.I2(a) being positioned between 6.3 and 9.2 kb, so the absence of detection of the intron could be due to RNA degradation at the 3' end, RNA folding, or premature transcription termination. In any case, B.c.I2(a) would not be able to splice out of the transcript because of its reverse orientation relative to the gene. Surprisingly, RT-PCR using the forward primer instead of the reverse primer and subsequent PCR revealed that this intron was part of an RNA molecule transcribed from the DNA strand opposite to the host gene. Both 5' and 3' intron-exon junctions were detected, as well as the presence of a spliced molecule (Fig. 5A). All products had the expected sizes, and sequencing of the spliced fragment confirmed the positions of the predicted intron-exon junctions.
As can be seen in Fig. 2, the genes downstream of B.c.I2(a)'s 3' exon (BCE3694 and BCE3695) are positioned in the same orientation as B.c.I2(a). Furthermore, two additional hypothetical ORFs (ORFX, 411 bp, genomic coordinates 3447633 to 3447223, and ORFY, 189 bp, genomic coordinates 3447837 to 3447649) oriented the same way could be predicted within the 3' exon using Glimmer2 (12). A PCR product of an expected size of 760 bp was obtained when RT-PCR was started using the forward primer located in the intron and a reverse primer located at position 2 in ORFX, indicating that the whole ORFX was transcribed with the intron (Fig. 5B). In contrast, no transcript containing parts of ORFX and BCE3694 could be detected when primers positioned in these two ORFs were used, as expected due to the presence of a predicted rho-independent transcription terminator (Fig. 2 and 5B).
With respect to the B. cereus ATCC 14579 B.c.I1(a) intron, expression of the intron ORF was detected, and a PCR product of expected length was also obtained for the 5' intron-exon junction, suggesting that a transcript spanning the upstream gene and the intron was present (data not shown). However, no fragment was obtained for the 3' intron-exon boundary, and in agreement with this, no spliced RNA could be detected.
Screening of various B. cereus group strains for the B.c.I2(a) intron and its host gene.
Given the unusual orientation of B.c.I2(a) in B. cereus ATCC 10987 and the fact that the host gene encodes a putative cell surface protein, 92 B. cereus group strains were screened for the presence of the intron and the gene. For 28 of the 92 strains tested, PCR products of an expected size of 450 bp were obtained using the primer pair designed for the 5' exon-intron junction of B.c.I2(a) (Table 3). This means that these strains encode a gene homologous to the BCE3690+BCE3693 gene of B. cereus ATCC 10987 and that the gene is interrupted by an intron that is inserted in the reverse orientation. For five other strains that were negative for the intron, a fragment of 300 bp was obtained using the primers located in the 5' and 3' exons, which is indicative of the presence of an uninterrupted homologue to BCE3690+BCE3693, as is the case in B. anthracis (Fig. 2). All the remaining 59 strains were negative for both primer sets, suggesting that they, like B. cereus ATCC 14579, lack the corresponding gene (Table 3). Testing of a few of the strains by PCR using the primer sets designed for the 5' end of BCE3690 and the 3' intron-exon junction also produced negative results. It is, however, possible that some isolates carry a divergent homologue for which the primers, based on the B. cereus ATCC 10987 genome sequence, were not well suited. Interestingly, 26 of the 33 strains carrying a homologue to BCE3690+BCE3693 are patient isolates. In particular, a group of 15 closely related strains isolated from patients affected by periodontitis were all positive (AH 812, AH 813, AH 816, AH 817, AH 818, AH 819, AH 820, AH 823, AH 824, AH 825, AH 826, AH 827, AH 828, AH 829, and AH 831).
|
DISCUSSION |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
One of the most striking findings in this analysis is the presence, in the chromosome of B. cereus ATCC 10987, of the B.c.I2(a) intron located on the reverse strand of a gene encoding a putative cell wall or cell surface protein (host gene BCE3690+BCE3693 [Fig. 2 and see Fig. S1 in the supplemental material]). Such an arrangement implies that the complementary sequence of the intron would be part of the mRNA transcript of the host gene, and therefore, the intron would not be able to fold into the proper RNA structure required for splicing and would not be recognized by an IEP, thereby disrupting the host gene. To our knowledge, the introns inserted in the opposite strand of the gene encoding a DNA methyltransferase in Xylella fastidiosa (10) and of a gene encoding a hypothetical protein in Thermosynechococcus elongatus (37, 38) are the only two published examples of such a gene disruption. Interestingly, in a PCR screen of 92 B. cereus group strains, 28 were shown to harbor an intron similar to B.c.I2(a) inserted in the complementary strand of a gene homologous to BCE3690+BCE3693 (Table 3). Even more remarkable is the fact that B.c.I2(a) is transcribed on the reverse strand relative to the host gene and splices out of the transcript in vivo in B. cereus ATCC 10987 (Fig. 5A). The antisense RNA seems to be the product of the transcription of one (or two) internal hypothetical ORF (ORFX and possibly ORFY) that we predicted to be located in the opposite strand within the 3' exon (Fig. 2 and 5B). Corresponding ORFs could be predicted in other closely related B. cereus group genomes; however, they do not share any significant homology to known proteins, and no insight about their function could be inferred from sequence analysis. Antisense ORFs have been reported for viruses (e.g., see references 7 and 26).
It is intriguing that among the B. cereus/B. thuringiensis strains examined here, 75% of those carrying a BCE3690+BCE3693-like gene are from clinical sources (Table 3). Most of them, the strains isolated from periodontitis cases in particular, are phylogenetically very close to each other and to the pathogen B. anthracis, according to multilocus enzyme electrophoresis analysis (20, 21) (Helgason, unpublished). The function of the BCE3690+BCE3693 gene is unknown, but the corresponding protein exhibits characteristics of cell surface proteins. Proteins exposed at the cell surface might potentially play a role in interaction of the bacterium with host cells and thus pathogenicity. The apparently biased distribution of the gene towards patient isolates and B. anthracis could suggest so. Interestingly, among 14 BCE3690+BCE3693-positive B. cereus/B. thuringiensis strains that were tested for the presence of an S layer, 12 were shown to have an S layer on the cell surface, as do B. anthracis isolates, while 6 out of 7 BCE3690+BCE3693-negative strains do not seem to possess an S layer (36). It is not clear what could be the possible implication(s) of the B.c.I2(a)-mediated disruption. Translation of the mRNA up to the intron insertion point would make a sequence containing 16 of the 19 adhesion-like domains representing 84% of the size of the full-length protein, so a truncated protein might still be functional. The production of a smaller protein may have no phenotypic consequences if the purpose of the protein is only to be embedded in an extracellular matrix. This may also explain why the gene disruption could have been maintained in many B. cereus group strains.
Another surprising finding concerns the B.c.I4 intron of B. cereus ATCC 10987, for which the spliced RNA product was shorter than expected (Fig. 5A). The actual start of the 3' exon was 56 nucleotides after the base of domain VI, as opposed to 3 or 4 nucleotides in all other characterized group IIB introns (Fig. 3 and 4). B.c.I4 is located on the pBc10987plasmid within the gene (BCEA0033+BCEA0036) orthologous to the pXO1-70 ORF (BXA0099) of the B. anthracis pXO1 plasmid, which encodes a hypothetical protein containing a DNA primase domain. Homologous sequences were also found in B. cereus ATCC 43881 and B. thuringiensis ATCC 33679 (41), while the extra 56 bp at the end of B.c.I4 do not share significant homology to any sequence in the public databases. Interestingly, translation of the observed spliced sequence would produce a protein of exactly the same size as BXA0099/pXO1-70, while the spliced sequence generated by an intron ending 3 or 4 bases after domain VI would contain premature stop codons. The observed splicing event therefore avoids gene disruption. It seems that B.c.I4 has an unusual splicing mechanism, and further laboratory experiments will be required to elucidate this observation. It will be particularly interesting to determine whether the extra segment is part of the intron RNA structure.
Among the other three intragenic group II introns of B. cereus ATCC 10987 [B.c.I2(b), B.c.I3, and B.c.I5] that are all inserted in the same orientation as their host genes and splice normally in vivo (Fig. 5A), B.c.I3 is a noticeable case. B.c.I3 is interesting because it is inserted within the chromosomal gene encoding the subunit of the gram-positive specific DNA polymerase III (polC [BCE3856+BCE3859]), while insertions of group II introns into housekeeping or conserved loci are quite infrequent in bacterial genomes. Only four instances have been reported (1, 10, 14, 31, 32). B.c.I3 constitutes the first case of a bacterial group II intron entirely inserted in a gene coding for a conserved protein involved in the essential DNA replication process (an intron possibly inserted in a replicative DNA helicase in the Onion yellows phytoplasma genome is also reported at the Group II intron database). Splicing of B.c.I2(b), B.c.I3, and B.c.I5 puts the neighboring ORFs in frame in one sequence of a size identical to that of the orthologous noninterrupted ORF in B. anthracis, implying that the resulting protein is not affected. However, if the splicing efficiency of an intron is very low, a large fraction of the mRNA molecules is nonfunctional, which could have an important effect on the level of protein expression.
The B. anthracis intron B.a.I2 on the pXO1 plasmid is very unusual in that it uses an alternative 3' splice site four nucleotides downstream of the main splice site (47). Splicing at the alternative site occurs at a frequency of only 4% in vivo, but this sets the flanking ORFs in frame, while splicing at the main site keeps them out of frame (47). This mechanism seems to be related to irregularities in the B.a.I2 RNA secondary structure permitting more flexibility and is thought to have evolved for regulating gene expression of the downstream ORF (47). B.c.I2 of B. cereus ATCC 10987 shares some of B.a.I2's structural peculiarities (Fig. 3). Furthermore, -' and EBS3-IBS3 pairings require that B.c.I2 ends with a T and that the 3' exon starts with a T, and therefore, the TTTT sequence present at the 3' intron-exon junction of the B.c.I2(b) copy on the pBc10987plasmid could produce two potential alternative 3' splice sites (Fig. 4). This raises the question about the possibility of alternative splicing of B.c.I2(b) and the TTTT motif providing a regulatory mechanism for expression of the host gene (BCEA0121). We did not detect alternatively spliced RNA products for B.c.I2(b) in our RT-PCR and PCR assays, and in contrast to B.a.I2, splicing of B.c.I2(b) at the expected 3' site produces in-frame exons (Fig. 5A), while the use of the two potential alternative sites would generate a stop codon within the first 12 bases of the ligated transcript. This would basically disrupt BCEA0121, which encodes a conserved hypothetical protein, and it may therefore be unlikely that alternative splicing occurs in vivo, unless it is important for the bacterium to suppress expression of the gene under particular environmental or physiological conditions, and thus, this would provide a putative regulatory mechanism. Alternative splicing may be unlikely for the other intragenic introns B.c.I2(a), B.c.I3, and B.c.I5 because there is no potential alternative splice sites in the near proximity (up to 15 bases) of the 3' boundary (Fig. 4), or they do not exhibit the structural irregularities seen in B.a.I2 (Fig. 3).
In this paper, we also provide, for the first time, experimental evidence of splicing of intergenic bacterial group II introns. B.c.I1(b) and B.c.I1(c) of B. cereus ATCC 10987 are transcribed with their upstream genes and splice out in vivo (Fig. 5A). The transcripts likely end at a rho-independent terminator predicted to be present before the downstream gene (see Fig. S1 in the supplemental material). B.c.I1(a) of B. cereus ATCC 14579 seems to be transcribed with its upstream gene as well; however, no conclusive results regarding its splicing ability could be obtained. Although B.c.I1(a) should intrinsically be able to splice, since it is identical to the functional B.c.I1(b) and B.c.I1(c) introns, it may not actually splice in vivo due to the very short 3' exon. It could be that the stem-loop structure formed by the transcriptional terminator located only eight bases downstream of the 3' intron boundary interferes with the binding of the IEP to unspliced RNA. To our knowledge, no experimental study involving such a short exon has been reported to date. The observed splicing of B.c.I1(b) and B.c.I1(c) raises a question: why would intergenic introns keep the ability to splice? As suggested previously (10), the splicing ability would be maintained because splicing and reverse splicing are required for mobility and insertion into new DNA target sites. Movement of introns could provide a regulatory or adaptive repertoire to the bacterium, for example, under stressful conditions, as shown for other mobile elements (e.g., see references 39 and 56). However, the IEP of the B. cereus intergenic introns seems to be missing functional parts of the En domain which is involved in the mobility process, in particular, the H-N-H motif (see Fig. S2 in the supplemental material). More than half of the bacterial group II introns compiled to date lack the En domain (9, 10), but they may still be able to move to new sites, as demonstrated experimentally for the Sinorhizobium meliloti S.me.I1 (RmInt1) intron (29). The fact that B.c.I1 is present in three different locations across two bacterial strains indicates that mobility events have occurred (Fig. 1 and 4).
Group II introns were found in additional B. cereus group strains for which genomic data are available (see Table S1 in the supplemental material). Group I introns, which are not related to group II introns and have a different structure and splicing mechanism (see reference 8 for a review) are also present in B. cereus group bacteria (see Table S2 in the supplemental material). To date, only the B. anthracis group I intron inserted within the recA gene, involved in DNA recombination and repair, has been shown to splice in vivo (30).
Comparative and detailed sequence and structure analysis can help unravel and correct genome annotations. This was exemplified by the finding of intragenic introns showing that the upstream and downstream ORFs were actually part of the same gene. Moreover, in the original genome
reports (27, 44, 45), a small conserved hypothetical ORF (120 nt on average) was predicted to be located in the 5' region of the catalytic RNA part of all seven intragenic introns of B. cereus and B. anthracis analyzed in this study (see Fig. S1 in the supplemental material). These are clearly overpredictions generated by the gene-finding program Glimmer2 (12) in this case and are certainly not true genes.
|
ACKNOWLEDGMENTS |
|---|
This work was supported by grants from the Norwegian Research Council through the Strategic University Programme (SUP) and the Functional Genomics (FUGE-CAMST) platform.
|
FOOTNOTES |
|---|
Supplemental material for this article may be found at http://jb.asm.org/.
|
REFERENCES |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
DNA digested with BstEII (New England Biolabs, Beverly, MA).