|Home | About | Journals | Submit | Contact Us | Français|
The Campylobacter genus consists of a number of important human and animal pathogens. Although the 16S rRNA gene has been used extensively for detection and identification of Campylobacter species, there is currently limited information on the 23S rRNA gene and the internal transcribed spacer (ITS) region that lies between the 16S and 23S rRNA genes. We examined the potential of the 23S rRNA gene and the ITS region to be used in species differentiation and delineation of systematic relationships for 30 taxa within the Campylobacter genus. The ITS region produced the highest mean pairwise percentage difference (35.94%) compared to the 16S (5.34%) and 23S (7.29%) rRNA genes. The discriminatory power for each region was further validated using Simpson's index of diversity (D value). The D values were 0.968, 0.995, and 0.766 for the ITS region and the 23S and 16S rRNA genes, respectively. A closer examination of the ITS region revealed that Campylobacter concisus, Campylobacter showae, and Campylobacter fetus subsp. fetus harbored tRNA configurations not previously reported for other members of the Campylobacter genus. We also observed the presence of strain-dependent intervening sequences in the 23S rRNA genes. Neighbor-joining trees using the ITS region revealed that Campylobacter jejuni and Campylobacter coli strains clustered in subgroups, which was not observed in trees derived from the 16S or 23S rRNA gene. Of the three regions examined, the ITS region is by far the most cost-effective region for the differentiation and delineation of systematic relationships within the Campylobacter genus.
Members of the Campylobacter genus are Gram-negative, nutritionally fastidious, microaerophilic organisms that are spiral, curved, or rod shaped and inhabit the gastrointestinal tracts of humans and animals. The first Campylobacter species to be isolated was Campylobacter fetus (Vibrio fetus), isolated from the uterine mucus of a sheep in 1906 (48). Since then, the Campylobacter genus has grown to comprise 17 formally named species. The most recognized species within the Campylobacter genus is Campylobacter jejuni, a gastrointestinal pathogen and a leading bacterial cause of acute diarrhea and gastroenteritis, accounting for 400 million cases in adults and children worldwide each year (2, 21, 44). Campylobacter coli and a number of other non-jejuni Campylobacter species are also considered to be important human and animal pathogens (1, 3, 14, 40, 56).
The 16S rRNA gene has been utilized extensively for rapid detection and identification of Campylobacter species (32, 36, 37, 39). This is largely due to the fact that the 16S rRNA gene is of considerable length (~1,500 bp), and it is ubiquitous in members of the Campylobacter genus and almost all bacteria (9, 55). The fact that certain regions of the 16S rRNA gene are highly conserved, and that any changes in the sequence are therefore likely to be an accurate measure of time, makes it a useful molecular marker for the study of phylogenetic relationships (28). However, this high sequence similarity observed between members of the Campylobacter genus also makes it difficult to differentiate between species such as C. jejuni and C. coli on the basis of the 16S rRNA gene (23, 43). In addition, this problem is further compounded by the remarkably similar phenotypic, host, and ecologic characteristics which many Campylobacter species share.
Currently, there is relatively little sequence data available on the 23S rRNA gene and the internal transcribed spacer (ITS) region that lies between the 16S and 23S rRNA genes for members of the Campylobacter genus. Although the potential of using the 23S rRNA gene and the ITS region for species differentiation and systematics has not been extensively investigated in the majority of Campylobacter species, previous studies have examined the ability of these regions to differentiate a small number of thermotolerant Campylobacter species, including C. jejuni, C. coli, Campylobacter lari, and Campylobacter upsaliensis (16, 18, 29). While the 23S rRNA genes in a number of strains of C. jejuni and C. coli share a similar sequence identity (41), previous studies have reported the presence of strain-specific intervening sequences (IVS) within the 23S rRNA gene of C. jejuni, C. coli, C. fetus, and C. upsaliensis (26, 53). Such an observation suggests that the 23S rRNA gene may be useful for differentiation at the strain level. While the presence of IVS elements does not appear to relate to the pathogenicity in C. jejuni (31), IVS are known to lead to rRNA fragmentation and have been hypothesized to be remnants of transposable elements (15, 35, 51, 53).
Within the 16S-ITS-23S operon, the ITS region in C. jejuni subsp. jejuni, C. coli, and C. lari has been reported to be highly variable in size and/or sequence composition and contains different types of sequences encoding tRNA molecules (8, 29). The diversity of the ITS region observed previously in these Campylobacter species suggests that this region may be useful for differentiation and identification of other nonthermotolerant Campylobacter species. Indeed, the ITS region has increasingly been used for differentiation between bacterial species or strains, including Escherichia coli strains (20), Mycobacterium spp. (47), cyanobacteria (4, 46), and acetic acid bacteria (50), which cannot be easily differentiated using the 16S rRNA gene. Furthermore, the versatility of the ITS region as a molecular tool has also been exploited in environmental microbiology in the study of marine microbial diversity (6, 7, 19, 49).
The aim of this study was therefore to examine the ability of the 16S rRNA gene, 23S rRNA gene, and ITS region to differentiate and delineate systematic relationships between 30 members of the Campylobacter genus.
The Campylobacter taxa used in this study were obtained from our own laboratory, the University of New South Wales Culture Collection, the American Type Culture Collection (ATCC), or the National Collection of Type Cultures (NCTC), as shown in Table Table1.1. Arcobacter butzleri RM4018, a species from a genus closely related to the Campylobacter genus, was used in this study for comparative analysis. All bacteria were grown on horse blood agar supplemented with 5% defibrinated horse blood for 2 to 4 days at 37°C under 5% CO2 microaerophilic conditions generated by the Campylobacter system BR0056A (Oxoid Limited, Hampshire, United Kingdom). Bacteria were harvested and washed once using phosphate-buffered saline (PBS) prior to DNA extraction. Additional near-complete 16S and 23S rRNA genes and ITS regions from a range of Campylobacter species, not available from our collection, were obtained from GenBank (http://www.ncbi.nlm.nih.gov/).
Bacterial DNA was extracted using the Qiagen Puregene core kit A (Qiagen, Hilden, Germany) according to the manufacturer's instructions. Campylobacter species DNA was subjected to PCR analysis using combinations of 18 PCR primers to amplify the 16S rRNA gene, ITS region, and 23S rRNA gene (Table (Table2).2). The target position of each primer is illustrated in Fig. Fig.1.1. One pair of primers was used in a single PCR. PCR analysis was performed in a 25-μl reaction mixture consisting of 10 pmol of each primer pair (Sigma-Aldrich, Sydney, Australia), 1× PCR buffer (Fisher Biotech, Subiaco, Australia), 200 nM each deoxynucleotide triphosphate (Fisher Biotech), 1.5 mM MgCl2 (Fisher Biotech), 0.7 U Taq polymerase (Fisher Biotech), and 20 ng DNA. The thermal cycling conditions for the PCR were as follows: 94°C for 5 min, 40 cycles of 94°C for 20 s, 52°C for 20 s, and 72°C for 2 min, followed by 72°C for 7 min. Five microliters of the PCR product was subjected to gel electrophoresis (1.5% agarose), stained with 1× GelRed nucleic acid gel staining solution (Biotium, Hayward, CA) for 10 min, and visualized under UV transillumination.
PCR products were sequenced using BigDye Terminator chemistry (Applied Biosystems, Foster City, CA) on an ABI 3730 capillary DNA sequencer (Applied Biosystems). Prior to the sequencing reaction, PCR products were purified using the QIAquick PCR purification kit (Qiagen) according to the manufacturer's instructions. Sequencing of both the 5′ and 3′ ends of the amplicons was performed, using 1 μl ABI Prism BigDye Terminator version 3.1 (Applied Biosystems), 10 pmol/μl of the required primer, 50 to 200 ng DNA, and Milli-Q water (Millipore, Bedford, MA) to make up the final volume of 20 μl. The sequencing program consisted of 96°C for 1 min and 30 cycles of 96°C for 10 s, 50°C for 5 s, and 60°C for 4 min.
Analyses were performed using programs available from the Australian National Genomic Information Service (ANGIS) at the University of Sydney. PILEUP from the GCG package (12) and Multicomp (45) were used for multiple sequence alignment and comparison. The average pairwise percent difference was calculated with Multicomp using a method described by Nei and Miller (42). The ability of an individual region to discriminate isolates was further determined using Simpson's index of diversity (D value) performed as previously described (25) using the following formula:
where N is the total number of isolates analyzed, S is the total number of Campylobacter sequence types obtained, and nj is the number of isolates belonging to the jth type. Phylip was used to generate neighbor-joining trees and bootstrap values (17). Secondary structures for Campylobacter 23S rRNA sequences were predicted using the GeneBee program (5) available from Moscow State University (http://www.genebee.msu.su/services/rna2_reduced.html). The Aragorn program was used for the identification of different types of tRNA molecules, the number of bases, and % GC content within the ITS region (34).
The 16S rRNA gene, 23S rRNA gene, and ITS region of 20 members of the Campylobacter genus were generated by using PCR (60 sequences), and sequences from an additional 10 members of the Campylobacter genus were obtained from GenBank (30 sequences). Sequences of each region from 30 Campylobacter isolates were used to determine inter- and intraspecies sequence variability. The mean pairwise percentage difference between all Campylobacter species when using the ITS sequence was 35.94%, which was significantly higher than the 5.34% and 7.29% obtained when using 16S and 23S rRNA gene sequences, respectively (P < 0.0001; paired t test). The discriminatory power for each individual region was calculated using Simpson's index of diversity (D value) (25), where a D value of 1 denotes the greatest ability of a region in discriminating different isolates. The 23S rRNA gene had the highest D value (0.995), followed closely by the ITS region (D value = 0.968), while the 16S rRNA gene had the lowest D value (0.766).
Comparison of the sequences from the ITS region between species showed that Campylobacter curvus 525.92 and Campylobacter hominis ATCC BAA381 were the most variable (70.21% pairwise difference). The ITS region was able to differentiate between the two C. hominis strains (25.85% difference), which were found to be identical when using the 16S rRNA gene. However, the ITS region was identical in all of the four C. coli strains, and therefore it could not be used for differentiation of the strains within this species.
The highest pairwise percentage difference between strains of the same species was between C. hominis ATCC BAA-381 and C. hominis UNSWCD (25.85%). The intraspecies pairwise differences using the ITS region for other Campylobacter species were as follows: C. coli, 0%, four strains; Campylobacter concisus, 1.19 to 5.49%, four strains; C. jejuni (including subspecies), 0 to 12.84%, 10 strains; and Bacteroides ureolyticus, 0.7 to 7.69%, five strains. The ITS region produced the highest pairwise percentage difference between strains of C. jejuni, C. concisus, C. hominis, and B. ureolyticus compared to the 16S rRNA gene and the 23S rRNA gene.
Comparison of the 16S rRNA gene sequences between species showed that C. curvus 525.92 and B. ureolyticus UNSWE were the most variable (11.74% difference). The 16S rRNA gene sequence was unable to differentiate the majority of C. jejuni and C. coli strains from each other. For example, the 16S rRNA gene sequence of C. coli X7 and C. coli X10 was identical to four other strains of C. jejuni.
The highest pairwise percentage difference between strains of the same species was 1.43% (between C. coli ATCC 33559 and C. coli RM2228/C. coli X7/C. coli X10). The intraspecies differences using the 16S rRNA gene sequences for all other Campylobacter species were as follows: C. coli, 0 to 1.43%, four strains; C. concisus, 0.15 to 0.45%, four strains; C. hominis, 0%, two strains; C. jejuni (including subspecies), 0 to 0.23%, 10 strains; and B. ureolyticus, 0.08 to 0.0.38%, five strains.
Comparison of the 23S rRNA gene sequences between species showed that C. curvus 525.92 and C. jejuni subsp. doylei 269.97 were the most variable (13.82% difference). The 23S rRNA gene was unable to distinguish two strains of C. coli (strains X7 and X10) and two strains of C. jejuni (strains 100 and BABS091400).
The highest percentage of pairwise difference between strains of the same species was observed between C. jejuni INN-73-83 094400 and C. jejuni subsp. doylei 269.97 (1.55% difference). Examination of the 23S rRNA gene showed that the intraspecies differences for the remaining Campylobacter species were as follows: C. coli, 0 to 0.47%, four strains; C. concisus, 0.12 to 2.51%, four strains; C. hominis, 0.51%, two strains; C. jejuni (including subspecies), 0 to 1.55%, 10 strains; and B. ureolyticus, 0.04 to 0.28%, five strains.
The ITS region is a region which lies between the 16S rRNA gene and the 23S rRNA gene. The ITS regions of 30 members of the Campylobacter genus were highly variable in length and % GC content. The average length of the Campylobacter ITS region was 880 bp. The longest ITS region was 1,646 bp, which was observed in C. hominis ATCC BAA-381 (Table (Table3).3). The shortest ITS region was 591 bp in length from C. concisus UNSWCD. In addition, all C. concisus strains harbored ITS regions that were relatively short in length (591 to 698 bp) but higher in % GC contents (32.8 to 34.5%) in comparison with other Campylobacter species. The average % GC content of the Campylobacter ITS region was 29.3%. C. concisus 13826 had the highest % GC content in the ITS region at 34.5%, while C. hominis ATCC BAA-381 had the lowest % GC content at 22.7%. The variability in length and % GC content was also evident between strains of a specific Campylobacter species. The lengths of the ITS region for 10 C. jejuni strains ranged from 714 bp (C. jejuni toxigenic strain Mexico INN-73-83 094400) to 828 bp (C. jejuni subsp. doylei 269.97), and the % GC content ranged from 27.5% (C. jejuni subsp. jejuni 81116) to 29.4% (C. jejuni toxigenic strain Mexico INN-73-83 094400).
All Campylobacter species examined in this study had a 16S-ITS-23S operon configuration except for C. hominis and C. upsaliensis. The 16S and 23S rRNA genes of these two species were not arranged in an operon. The 16S-ITS-23S operon was determined by PCR amplification, which generates a visible product using primers that target a region spanning the 16S-ITS or ITS-23S region. No PCR amplicon was observed when the ITS region of C. hominis and C. upsaliensis was amplified by PCR analysis using primers 1494R(RC) or RC1494M (forward primers which anneal to the 3′ end of the 16S rRNA gene) and M83 or M83M (reverse primers which anneal to the 5′ end of the 23S rRNA gene), whereas PCR amplicons were visible for other Campylobacter species tested in this study. Examination of whole genomes from fully sequenced C. hominis ATCC BAA-381 (GenBank accession no. CP000776) and C. upsaliensis RM3195 (GenBank accession no. AAFJ00000000) confirmed that these species do not have the 16S-ITS-23S operon. The region identified to be immediately downstream of the 16S rRNA gene and upstream of the next open reading frame was used as the ITS region for C. hominis and C. upsaliensis, given that in this region tRNA coding genes were identified in a configuration similar to that of other Campylobacter species.
The following three tRNA configurations were observed in 30 Campylobacter taxa: (i) tRNAAla(TGC) and tRNAIle(GAT), (ii) tRNAIle(GAT) and tRNAAla(TGC), and (iii) tRNAAla(TGC) (Table (Table3).3). C. fetus subsp. fetus 82-40 was the only Campylobacter species with tRNA configuration 3, a configuration with a single tRNA-coding gene. The tRNA-coding gene for tRNAAla(TGC) was identified at position 145 to 220 within the ITS region of C. fetus subsp. fetus 82-40, but no tRNAIle(GAT) was identified. All Campylobacter species except for C. concisus and Campylobacter showae had tRNA configuration 1, characterized by the presence of tRNAAla(TGC) followed by tRNAIle(GAT) (the most prevalent configuration). C. concisus, C. showae, and A. butzleri had tRNA configuration 2, which contained a reversed order of the tRNA-coding genes compared to configuration 1.
All of the identified tRNAAla(TGC) molecules consisted of 76 bases, and all of the tRNAIle(GAT) molecules consisted of 77 bases (Table (Table3).3). The % GC content in all tRNAAla(TGC) was 60.5%, which was highly conserved. In contrast, the % GC content in tRNAIle(GAT) was more variable, with all Campylobacter species having a GC content of 53.2% except C. hominis ATCC BAA-381 and C. hominis UNSWCD strains, which had a GC content of 54.5%. A. butzleri RM4018 had a tRNAIle(GAT) GC content of 55.8%, which was greater than all of the Campylobacter species examined in this study.
Sequences of the 16S rRNA gene, 23S rRNA gene, and ITS region were aligned to generate neighbor-joining (NJ) trees. A. butzleri RM4018 (accession no. CP000361) was used as an outgroup. In the NJ tree derived from the 16S rRNA gene sequences, there were four major clusters (Fig. (Fig.2A).2A). Cluster I consisted of very closely related C. coli and C. jejuni strains, C. lari, C. fetus subsp. fetus, and C. upsaliensis. Thermotolerant Campylobacter species which have the ability to grow at 42°C were found only in cluster I (C. jejuni, C. coli, C. lari, and C. upsaliensis). C. fetus subsp. fetus, which is not considered a thermotolerant Campylobacter species, also clustered in cluster I. However, some strains of C. fetus subsp. fetus have been reported to grow at 42°C (13, 24). Cluster II contained all B. ureolyticus strains only. Cluster III included both C. hominis strains. Cluster IV comprised of C. concisus, C. showae, and C. curvus.
In the NJ tree derived from the 23S rRNA gene sequences, members of the Campylobacter genus were divided into four major clusters (Fig. (Fig.2B),2B), similar to the tree derived from the 16S rRNA gene data. The first cluster contained C. jejuni, C. coli, C. upsaliensis, and C. lari. All thermotolerant Campylobacter species were grouped in this cluster in a fashion similar to that observed in the 16S rRNA gene. The second cluster consisted of C. concisus, C. showae, and C. curvus. Cluster III consisted of C. fetus subsp. fetus and two C. hominis strains. The clustering of C. fetus subsp. fetus differed in the 23S rRNA gene tree compared with the 16S rRNA gene tree in which C. fetus subsp. fetus clustered with C. jejuni, C. coli, C. lari, and C. upsaliensis. B. ureolyticus strains formed a distinct cluster (cluster IV), similar to that of the tree derived from the 16S rRNA gene.
For the NJ gene tree based on Campylobacter sequences derived from the ITS region, a markedly different topology was observed, characterized by the presence of three clusters (Fig. (Fig.2C).2C). All B. ureolyticus strains formed a discrete cluster (cluster I), consistent with observations in the trees derived from the 16S and 23S rRNA genes. Cluster II included all strains of C. jejuni, C. coli, and C. lari, which was congruent with the clustering of thermotolerant Campylobacter species observed when using the 16S and 23S rRNA genes. A closer inspection of the tree revealed that the ITS region was able to further divide C. jejuni and C. coli strains into subclusters, with the exception of one strain of C. jejuni (INN-73-83 094400) which subclustered with the C. coli strains. Cluster III contained C. concisus and C. showae. However, C. curvus did not cluster with C. concisus and C. showae, as was observed using the 16S and 23S rRNA gene data. Interestingly, cluster III was more closely related to A. butzleri than to the remaining Campylobacter taxa. The relatedness of C. concisus and C. showae to A. butzleri was also reflected by their unique tRNA configuration (configuration 2, 16S-tRNAIle(GAT)-tRNAAla(TGC)-23S), unlike most other Campylobacter species (configuration 1, 16S-tRNAAla(TGC)-tRNAIle(GAT)-23S), as highlighted earlier. C. fetus subsp. fetus, C. curvus, C. hominis, and C. upsaliensis were very divergent and did not cluster with other Campylobacter species.
Using gene sequences from the 16S and 23S rRNA genes and the ITS region, we generated an NJ tree using all three regions (combined tree). Five major clusters were generated (Fig. (Fig.2D).2D). Cluster I consisted of C. jejuni strains only. Cluster II consisted of all C. coli strains and C. jejuni INN-73-83 094400. The combined tree was the only tree that was able to clearly group C. jejuni and C. coli strains into distinct clusters with the exception of C. jejuni INN-73-83 094400. Clusters III and IV were comprised of C. hominis and B. ureolyticus, respectively, showing grouping congruency with trees derived from three individual regions. C. concisus, C. showae, and C. curvus formed one group (cluster V), which is comparable to the clustering using individual regions. C. fetus subsp. fetus, C. lari, and C. upsaliensis were very divergent and did not cluster with other Campylobacter species. Thermotolerant Campylobacter species did not form one distinct cluster but were closely related to each other, with C. jejuni and C. coli taxa positioned in clusters I and II, respectively, and C. lari and C. upsaliensis located in close proximity to clusters I and II.
The 16S and 23S rRNA gene sequences were aligned against their corresponding genes in E. coli strain K-12 (accession no. NC000913). Intervening sequences (IVS) were present in the 16S rRNA gene of one of the 30 Campylobacter taxa (3%). An IVS of approximately 200 bp in size was found in C. curvus 525.92. The IVS was located at position 220 with respect to the 16S rRNA gene in E. coli K-12. The presence of IVS was also observed in the 23S rRNA gene sequences of 12 of the 30 Campylobacter taxa (40%). While there was only one IVS identified in the 23S rRNA gene of each taxon, the IVS differed in length, position, and type (Table (Table4).4). Not all of the strains within a specific species had identical IVS lengths, positions, and sequences. For example, three types of IVS were identified in C. jejuni.
There were six different IVS types, ranging from 37 bp (C. jejuni UNSW091300 and C. jejuni subsp. jejuni 81116) to 240 bp (C. curvus 525.92). The most frequently detected IVS was 143 bp in length and in position 1023 (with respect to E. coli K-12), and this type of IVS was found in five Campylobacter taxa, including three C. coli strains (RM2228, X7, and X10) and two C. jejuni strains (INN-73-83 094400 and RM1221). Two Campylobacter taxa (C. jejuni subsp. doylei 269.97 and C. upsaliensis RM3195) harbored a 172-bp IVS at position 1024. C. jejuni UNSW091300 and C. jejuni subsp. jejuni 81116 harbored a 37-bp IVS at position 1204. C. curvus 525.92, C. fetus subsp. fetus 82-40, and C. hominis UNSWCD all harbored a unique type of IVS with variable lengths and positions.
The presence of IVS in the Campylobacter genus appeared to be strain dependent. Of the four strains of C. coli examined, three contained IVS of the same sequence at the same location. C. coli strain ATCC 33559 was the only C. coli strain without an IVS in its 23S rRNA gene. Furthermore, IVS were present in 5 of 10 strains of C. jejuni, including three IVS of different sequence composition at two distinct positions. The five C. jejuni strains that did not contain an IVS were C. jejuni 100, 84-25, BABS 091400, NCTC 11168, and RP0001. C. hominis strain UNSWCD also contained an IVS, whereas none was found in C. hominis ATCC BAA381. The remaining 11 strains which did not have IVS included all five B. ureolyticus strains (UNSWC, UNSWE, UNSWJ, UNSWM, UNSWR), all four C. concisus strains (13826, ATCC 51561, ATCC 51562, UNSWCD), C. lari, and C. showae. C. curvus 525.92 was the only Campylobacter species examined to have contained IVS elements in both the 16S and 23S rRNA gene.
The secondary structure of the 23S rRNA molecules harboring an IVS from all Campylobacter taxa was predicted. As shown in Fig. Fig.3A3A using C. curvus 525.92 as a representative, the IVS was located at one extremity and was not a part of the central structure of the predicted 23S rRNA molecule. Removal of the IVS from the sequence of the 23S rRNA gene resulted in no major change to the secondary structure of the 23S rRNA molecule (Fig. (Fig.3B).3B). The predicted secondary structures of two fragmented 23S rRNA molecules obtained after excision of the IVS are shown in Fig. 3C and D.
Examination of available whole genomes of 10 Campylobacter isolates (C. coli RM2228, C. concisus 13826, C. curvus 525.92, C. fetus subsp. fetus 82-40, C. hominis ATCC BAA-381, C. jejuni NCTC 11168, C. jejuni RM1221, C. jejuni subsp. jejuni 81116, C. lari RM2100, and C. upsaliensis RM3195) showed that only three members of the Campylobacter genus harbored multiple rrn operons that consisted of variable ITS regions. Two sequence compositions of the ITS region were observed in C. concisus 13826, C. fetus subsp. fetus 82-40, and C. coli RM2228. To address whether the variability of the ITS regions affected the systematic relationship, the two versions of the ITS region from C. concisus 13826, C. fetus subsp. fetus 82-40, and C. coli RM2228 were included in an NJ tree generated from the ITS region. The two different ITS sequences of C. concisus 13826 and C. fetus subsp. fetus 82-40 remained in their respective clusters and were not in conflict (data not shown). In C. coli RM2228, both ITS sequences were clustered in the C. jejuni/C. coli cluster. One ITS sequence of C. coli RM2228 clustered with the C. coli subcluster within the C. jejuni/C. coli cluster, whereas the second ITS sequence clustered closer to the C. jejuni subcluster. No other C. coli strains examined in this study appear to have variable ITS sequences following sequencing of the 5′ and 3′ ends of the ITS region using the ITS primer pair RC1494M/1494R(RC) and M83/M83M. All Campylobacter species investigated possessed identical 16S and 23S rRNA genes despite multiple copies of rrn operons.
The Campylobacter genus consists of a number of important pathogens in human and veterinary medicine. In this study, we investigated the characteristics of the 16S rRNA gene, 23S rRNA gene, and ITS region for 30 taxa within the Campylobacter genus and their potential to be used in species differentiation and delineation of systematic relationships. Similar to the closely related members of the Helicobacter genus, IVS can be found in both the 16S and 23S rRNA genes (11, 35). In this study, we identified the presence of IVS elements in the 23S rRNA gene of 12 of the 30 taxa within the Campylobacter genus. We have shown for the first time that C. curvus and C. hominis also contain an IVS element in their 23S rRNA genes. In addition, IVS of various lengths within the 23S rRNA gene were identified in different strains of C. jejuni, C. coli, C. fetus, and C. upsaliensis, which is in agreement with previously reported studies (26, 31, 51). The strain-specific appearance of IVS within the 23S rRNA gene suggests that this gene would be useful for differentiation between strains that contain highly similar or identical 16S rRNA gene sequences. Our study revealed that of the Campylobacter species examined, none had more than one IVS within the 23S rRNA gene. In contrast, members of the phylogenetically related Helicobacter genus have been shown to contain one or more IVS within the 23S rRNA gene, in four distinct positions (11). Interestingly, the presence of IVS is known to lead to rRNA fragmentation induced by posttranscriptional excision, which mediates the removal of IVS from within the rRNA molecule of bacteria, including Campylobacter species and members of the Alphaproteobacteria (15, 31, 51, 53). Ribosomal RNAs without IVS remain intact after transcription. The evolutionary advantage and function of these IVS elements is, however, still unclear. Thus far, no relationship between C. jejuni strains containing an IVS and their pathogenicity has been observed (31). This suggests that IVS may not have an essential function and that they may be remnants of transposable elements, which have been previously inserted into and subsequently excised from the 23S rRNA genes (31, 35).
The ITS region found between the 16S and 23S rRNA genes has been previously investigated only in a small group of Campylobacter species consisting of C. jejuni, C. coli, and C. lari, but not in any other members of the Campylobacter genus. These studies have shown that all the Campylobacter species investigated contain a 5′-16S-tRNAAla-tRNAIle-23S-3′ tRNA configuration (8, 30). Examination of other members of the Campylobacter genus in our study revealed two other possible tRNA configurations. The most prevalent configuration was found to be the 5′-16S-tRNAAla-tRNAIle-23S-3′ configuration, which was observed in all Campylobacter species except C. concisus, C. showae, and C. fetus subsp. fetus. In C. concisus and C. showae, we observed an inverted tRNA configuration not previously reported in members of the Campylobacter genus (5′-16S-tRNAIle-tRNAAla-23S-3′). This inverted tRNA configuration has also been shown in Bacillus subtilis (38) and cyanobacteria (27). The second configuration not previously reported was found in C. fetus subsp. fetus, which had only one tRNA-coding gene (tRNAAla), a characteristic that is found primarily in Gram-positive bacteria (22).
The differentiation of C. jejuni and C. coli continues to be a significant taxonomic problem, as both species share remarkably similar phenotypic, genotypic, host, and ecologic characteristics and are virtually indistinguishable using the 16S or 23S rRNA gene, as shown in this study. Currently, the most reliable method to distinguish C. jejuni and C. coli is believed to be the presence of the hippuricase gene in C. jejuni but absent in C. coli (36). Interestingly, delineation of the systematic relationships between members of the Campylobacter genus showed that when the 16S and 23S rRNA genes and the ITS region were combined, we were able to clearly differentiate strains of C. jejuni and C. coli, except for one C. jejuni strain (INN-73-83 094400) originally isolated from the diarrheic feces of a patient from Mexico. While our data question the designated identity of this C. jejuni strain, the results suggest the need to use multiple regions for the clearer separation of C. jejuni and C. coli. Where the ITS region alone was also able to differentiate between C. jejuni and C. coli strains, this was to a lesser degree than when using the 16S-ITS-23S operon. The differentiation of C. coli strains remains problematic as none of the three regions, including the ITS region, was able to differentiate all of the strains. The use of other genotypic methods such as pulsed-field gel electrophoresis (PFGE) and/or phenotypic methods is required to distinguish strains of C. coli (10).
Another major taxonomic problem in the Campylobacter genus is the taxonomic position of B. ureolyticus, which is currently ambiguous. Vandamme et al. (54) conducted a polyphasic taxonomic study of B. ureolyticus and found that this species has a quinone content, DNA base ratio, and phenotypic characteristics similar to those of other Campylobacter species, differing only in its fatty acid composition and ability to digest casein and gelatin (54).
On the basis of the 16S and 23S rRNA genes and the ITS region, our genotypic analyses of five B. ureolyticus strains indicated that they were most closely related to C. hominis. This bacterium is morphologically similar to B. ureolyticus, both of which are nonmotile, rod-shaped organisms that exhibit a swarming characteristic on horse blood agar, possibly mediated by the use of pili. In addition, on the basis of the 16S rRNA gene and the ITS region, we have shown that C. jejuni and C. coli were more closely related to B. ureolyticus than they were to C. concisus, C. curvus, and C. showae. Given this new evidence of a close systematic grouping of B. ureolyticus with other members of the Campylobacter genus on the basis of regions other than the 16S rRNA gene, there is an increasing prospect of classifying B. ureolyticus as a formal member of the Campylobacter genus.
Comparative sequence analyses of the 16S and 23S rRNA genes and the ITS region of members of the Campylobacter genus revealed that the most discriminatory region for species and strain differentiation was the ITS region. While the 23S rRNA gene has the highest D value, it has a relatively low mean pairwise difference. In contrast, the ITS region has a significantly higher mean pairwise difference and a D value similar to that of the 23S rRNA gene. Although it is clear that when the 16S and 23S rRNA genes and the ITS region were combined, a resultant NJ tree elicited taxonomic relationships of the highest resolution between members of the Campylobacter genus. However, one of the major limitations involved in examining all three regions is the time, effort, and cost required to amplify, sequence, and assemble a region of 5,500 nucleotides in length. Indeed, the entire process required 14 different PCR and sequencing primers. Of the three regions examined, the ITS region is by far the best region for bacterial identification, differentiation, and systematic analysis, as it is the shortest (~1,000 bp) and is highly discriminatory. In contrast to the four and eight different PCR and sequencing primers required to obtain the near-complete 16S rRNA gene (1,500 bp) and 23S rRNA gene (2,500 to 3,000 bp), only one pair of primers is required for the amplification and sequencing of the complete ITS region. In conclusion, the ITS region is the most cost-effective region for differentiation and delineation of systematic relationships for members of the Campylobacter genus.
Published ahead of print on 26 March 2010.