Abstract
Higher proportions (91%) of 168 corynebacterial isolates were positively identified by partial rpoB gene determination than by that based on 16S rRNA gene sequences. This method is thus a simple, molecular-analysis-based method for identification of corynebacteria, but it should be used in conjunction with other tests for definitive identification.
The genus Corynebacterium comprises more than 60 species, most of which have been isolated from humans or animals. Since certain Corynebacterium species are likely nonpathogenic, accurate identification of any clinical isolate of a Corynebacterium sp. is important (3). However, this may not be easy if conventional methods are used as they may require using such fastidious procedures as chromatography. Furthermore, commercially available biochemical test systems (e.g., API strips [Biomerieux, Marcy l'Etoile, France]) may not incorporate all of the tests necessary for the identification of every Corynebacterium species (2). The introduction of molecular methods, including 16S rRNA gene sequence analysis, has paved the way for much tighter circumscription of the genus and more reliable identification of Corynebacterium species (8, 10). However, since the 16S rRNA gene sequences of corynebacteria show very little polymorphism (6), accurate molecular identification is only possible by sequencing the complete 16S rRNA gene (approximately 1,500 bp). Previously, we designed universal primers for rpoB amplification and sequencing that made it possible to amplify and sequence 434- to 452-bp fragments (6). We demonstrated that this partial sequence was polymorphic enough for the accurate identification of all of the Corynebacterium species tested and that it yielded high bootstrap values in the phylogenetic trees that had been constructed by using this partial gene sequence. However, in that work only reference strains were studied. In the present study, we compared 16S rRNA gene and partial rpoB gene sequencing for identification of 168 Corynebacterium isolates recovered from various clinical specimens over a 5-year period ending in 2003.
Only the corynebacteria isolated from selected samples were included. Corynebacteria isolated from sterile sites or observed within polymorphonuclear neutrophils on Gram staining of the samples were all included. In the case of samples taken from nonsterile sites of patients suspected of having osteoarticular infections, isolates were submitted for identification if recovered at least three times in samples taken at three different times and if polymorphonuclear neutrophils were observed on Gram staining. The following species that were not described at the time of our previous work (6) or that are not valid species represented by reference strains were also studied: Corynebacterium nigricans (CCUG 48176T [the superscript T means that it is the reference strain]) (11), “Corynebacterium pseudogenitalium” (CCUG 28787), Corynebacterium tuberculostearicum (CCUG 45418T), CDC group G1 (CCUG 18436), CDC group G2 (CCUG 18437), and CDC group F1 (CCUG 32360).
The methods used for the amplification and sequencing of the partial rpoB gene sequence (with primers C2700F and C3130R) and the 16S rRNA gene, multiple-sequence alignments, calculation of percent similarities, and construction of phylogenetic trees were as described before (5, 6).
In our previous study (6), the type strain of each species was examined and we proposed that two Corynebacterium isolates belonged to the same species if they showed ≥95.9% similarity. However, on the basis of the results obtained in the present study, a cutoff of ≥95% appears to be more appropriate as it is in agreement with data obtained from a range of similarities observed among different groups of isolates (Table 1) and the phylogenetic tree in Fig. 1. The basis for determination of cutoffs has been detailed elsewhere (7). For example, this cutoff value enabled us to group 20 C. jeikeium and 13 C. aurimucosum isolates. This group allocation is supported by a phylogenetic tree that incorporated all of the isolates in the group (data not shown). Taking this as a cutoff value, it was possible to identify 153 (91%) of the 168 isolates tested. A significantly higher proportion of strains was positively identified when the identification system was based on the partial rpoB gene sequence compared to that based on the 16S rRNA gene sequence (136 [81%] of 168; P = 0.01 by chi-square test).
TABLE 1.
Percent similarities observed in the 16S rRNA and partial rpoB gene sequences with those of the closest relatives of the 168 clinical isolates studied
| Isolatea | No. of isolates tested | Closest relative (highest 16S rRNA similarity, %) | Closest relative (highest rpoB similarity, %) |
|---|---|---|---|
| 42 | C. amycolatum (99) | C. amycolatum (97-98) | |
| 20 | C. jeikeium (99) | C. jeikeium (95-97) | |
| 25850 | 2 | C. jeikeium (99) | C. jeikeium (93) |
| 59614 | 2 | C. jeikeium (99) | C. jeikeium (91) |
| 13 | C. aurimucosum (99-100) | C. aurimucosum (95-99) | |
| 1 | C. aurimucosum (99) | C. singulare (96) | |
| 5 | C. pseudodiphtheriticum (99-100) | C. pseudodiphtheriticum (98-99) | |
| 5 | C. propinquum/peudodiphtheriticum (99) | C. propinquum 97 | |
| 110393 | 1 | C. propinquum/peudodiphtheriticum (99) | C. propinquum (91) |
| 5 | C. accolens (98) | C. accolens (88-89) | |
| CDC G1/G2 (99-100) | CDC G1/G2 (97-100) | ||
| 3 | C. tuberculostearicum (99-100) | C. tuberculostearicum (98) | |
| 5 | C. urealyticum (98-99) | C. urealyticum (98-99) | |
| 3 | C. simulans (99) | C. simulans (97) | |
| 32 | C. striatum (99) | C. striatum (96-97) | |
| 4 | C. striatum/xerosis (99) | C. striatum (97) | |
| 4 | C. coylae (98-99) | C. coylae (95-98) | |
| 4 | C. mucifaciens (97) | C. afermentans (98-99) | |
| 1 | C. coyleae (98) | C. afermentans (96) | |
| 1 | C. coyleae/mucifaciens (97) | C. afermentans (96) | |
| 3301750 | 1 | C. mucifaciens (98) | C. afermentans/C. mucifaciens (93) |
| 4 | C. glucuronolyticum (98-99) | C. glucuronolyticu (99) | |
| 31595 | 1 | C. glucuronolyticum (93) | C. glucuronolyticum/seminale (89) |
| 1 | C. xerosis (98) | C. freneyi (98) | |
| 2300500 | 1 | C. xerosis/argentoratense (94) | C. ulcerans (89) |
| 89349 | 1 | C. kroppenstedtii (99) | C. kroppenstedtii (94) |
| 1 | C. glutamicum (99) | C. glutamicum (99) | |
| 1 | C. diphtheriae (99) | C. diphtheriae (98) | |
| 1 | C. durum (99) | C. durum (98) | |
| 1 | C. riegelii (99) | C. riegelii (98) | |
| 1 | Turicella otitidis (99) | Turicella otitidis (99) | |
| 110960 | 1 | C. phocae (96) | C. aurimucosum (88) |
The 10 clinical isolates whose identification is not clear-cut are indicated on the left.
FIG. 1.
Dendrogram representing phylogenetic relationships of Corynebacterium type species and not-well-classified isolates by the neighbor-joining method. The tree was derived by alignment of concatenated 16S rRNA and rpoB genes sequences. The sizes of the 16S rRNA and rpoB gene sequences aligned ranged from 1224 to 1248 and 425 to 443 bp, respectively. The support of each branch, as determined from 1,000 bootstrap samples, is indicated by the value at each node (in percent). Only bootstrap values of ≥70% are given. Mycobacterium smegmatis was taken as the outgroup.
We constructed two separate phylogenetic trees by using 16S rRNA and partial rpoB gene sequences, respectively (data not shown). For well-identified groups of species, the grouping was almost identical. This prompted us to construct yet another tree by using the sequences of both the partial rpoB gene and the 16S rRNA gene (9a). The tree so constructed gave high bootstrap values (Fig. 1). The 97% similarity value for the partial rpoB gene between the C. aurimucosum and C. nigricans type species and the position of these bacteria in the phylogenetic tree confirmed the notion that these species are synonymous (1).
The identities of several isolates appearing to be closely related with C. jeikeium (four isolates), C. propinquum (one isolate), C. afermentans/C. mucifaciens (one isolate), C kroppenstedtii (one isolate), and C. accolens (five isolates), having ≤94% similarity in the partial rpoB sequences, were not clear-cut. The five isolates related to C. accolens may be considered as having been identified correctly as they grouped with either CDC group G1 or G2. Phylogenetic analysis showed that four isolates having 91 to 93% similarity with C. jeikeium (Table 1) stood well apart from the major cluster of C. jeikeium. This species was first delineated on the basis of similarities in DNA sequences and protein electrophoresis patterns (4). The heterogeneity of C. jeikeium was further confirmed by DNA-DNA hybridization and antibiotic susceptibility patterns (9). Since our collection of strains included only one of C. kroppenstedtii, it was difficult to define whether isolate 89349 was indeed a genotype or a genospecies. This is because it fell short (94% similar to the type strain) of the cutoff value we have chosen, as described above, for the definition of a genospecies. Isolate 110393, which belonged to the C. propinquum/pseudodiphtheriticum group, was found to be an independent taxon. With only 91% partial rpoB gene similarity to the type strain of the species, it may well represent a new genospecies. Actually, isolate 3301750 belongs to the C. afermentans/C. mucifaciens group, and its separation from C. afermentans and C. mucifaciens on the basis of concatenated rpoB and 16S rRNA genes is supported by a bootstrap value of 88% (Fig. 1). However, these data obtained by rpoB analysis are not congruent with those obtained from analysis of the 16S rRNA gene (data not shown). However, caution should be exercised in interpreting such discrepancies, as the phylogenetic tree based on 16S rRNA gene sequences, because of low bootstrap values, has questionable value in this group. Extensive polyphasic study, along with complete rpoB gene sequencing, may prove worthwhile in the further characterization and delineation of isolates 89349, 110393, and 3301750. The partial rpoB genes of isolates 31595, 2300500, and 110960 were, respectively, 89, 89, and 88% similar to those of the closest species. These data, along with those obtained from the phylogenetic tree (Fig. 1), indicate that these isolates most likely represent new Corynebacterium genospecies.
Finally, C. aurimucosum, a newly described species (12), along with such well-known species as C. amycolatum, C. jeikeium, and C. striatum, is not an infrequent isolate in clinical practice. This study confirms that partial rpoB sequencing is a simple and efficient mean for identification of corynebacteriain clinical practice. The small size of the sequenced fragment renders it more convenient than 16S rRNA gene sequencing. However, in some ambiguous cases it should be used in conjunction with other tests for definitive identification.
Nucleotide sequence accession numbers.
The partial rpoB sequences of C. nigricans, C. pseudogenitalium, C. tuberculostearicum, CDC group G1, CDC group G2, and CDC group F1 included in this study (corresponding to positions 2713 to 3012 of the complete C. diphtheriae type strain rpoB gene) have been deposited in the GenBank database under accession numbers AY781343, AY581868, AY581869, AY581870, AY581871, and AY581872. The sequences of isolates 25850, 59614, 110393, 3301750, 89349, 31595, 2300500, and 110960 were deposited under accession no. AY581873, AY581874, AY581875, AY581876, AY581877, AY581878, AY581879, and AY581880, respectively, for the rpoB gene and under accession no. AY581881, AY581882, AY581883, AY581884, AY581885, AY581886, AY581887, and AY581888, respectively, for the 16S rRNA gene.
Acknowledgments
We thank Enevold Falsen for constructive criticism.
REFERENCES
- 1.Daneshvar, M. I., D. G. Hollis, W. S. Weyant, J. G. Jordan, J. P. MacGregor, R. E. Morey, A. M. Whitney, D. J. Brenner, A. G. Steigerwalt, L. O. Helsel, P. M. Raney, J. B. Patel, P. N. Levett, and J. M. Brown. 2004. Identification of some charcoal-black-pigmented CDC fermentative coryneform group 4 isolates as Rothia dentocariosa and some as Corynebacterium aurimucosum: proposal of Rothia dentocariosa emend. Georg and Brown 1967, Corynebacterium aurimucosum emend. Yassin et al. 2002, and Corynebacterium nigricans Shukla et al. 2003 pro synon. Corynebacterium aurimucosum. J. Clin. Microbiol. 42:4189-4198. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Funke, G., and K. A. Bernard. 2003. Coryneform gram-positive rods, p. 472-501. In P. R. Murray, E. J. Baron, J. H. Jorgensen, M. A. Pfaller, and R. H. Yolken (ed.), Manual of clinical microbiology. ASM Press, Washington, D.C.
- 3.Funke, G., A. von Graevenitz, J. E. Clarridge III, and K. A. Bernard. 1997. Clinical microbiology of coryneform bacteria. Clin. Microbiol. Rev. 10:125-159. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Jackman, P. J. H., D. G. Pitcher, S. Pelczynska, and P. Borman. 1987. Classification of corynebacteria associated with endocarditis (group JK) as Corynebacterium jeikeium sp. nov. Syst. Appl. Microbiol. 9:83-90. [Google Scholar]
- 5.Khamis, A., P. Colson, D. Raoult, and B. La Scola. 2003. Usefulness of rpoB gene sequencing for identification of Afipia and Bosea species including a strategy for the choice of discriminative partial sequences. Appl. Environ. Microbiol. 69:6740-6749. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Khamis, A., D. Raoult, and B. La Scola. 2004. rpoB gene sequencing for identification of Corynebacterium species. J. Clin. Microbiol. 42:3925-3931. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.La Scola, B., Z. Zeaiter, A. Khamis, and D. Raoult. 2003. Gene sequence based criteria for species definition in bacteriology: the Bartonella paradigm. Trends Microbiol. 11:318-321. [DOI] [PubMed] [Google Scholar]
- 8.Pascual, C., P. A. Lawson, J. A. E. Farrow, M. N. Gimenez, and M. D. Collins. 1995. Phylogenetic analysis of the genus Corynebacterium based on the 16S rRNA gene sequences. Int. J. Syst. Bacteriol. 45:724-728. [DOI] [PubMed] [Google Scholar]
- 9.Riegel, P., D. de Briel, G. Prevost, F. Jehl, and H. Monteil. 1994. Genomic diversity among Corynebacterium jeikeium strains and comparison with biochemical characteristics and antimicrobial susceptibilities. J. Clin. Microbiol. 32:1860-1865. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9a.Rokas, A., B. L. Williams, N. King, and S. B. Caroll. 2003. Genome scale approaches to resolving incongruence in molecular phylogenies. Nature 425:798-804. [DOI] [PubMed] [Google Scholar]
- 10.Ruimy, R., P. Riegel, P. Boiron, H. Montiel, and R. Christen. 1995. Phylogeny of the genus Corynebacterium deduced from analyses of small-subunit ribosomal DNA sequences. Int. J. Syst. Bacteriol. 45:740-746. [DOI] [PubMed] [Google Scholar]
- 11.Shukla, S. K., K. A. Bernard, M. Harney, D. N. Frank, K. D. Reed. 2003. Corynebacterium nigricans sp. nov.: proposed name for a black-pigmented Corynebacterium species recovered from the human female urogenital tract. J. Clin. Microbiol. 41:4353-4358. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Yassin, A. F., U. Steiner, and W. Ludwig. 2002. Corynebacterium aurimucosum sp. nov. and emended description of Corynebacterium minutissimum Collins and Jones (1983). Int. J. Syst. Evol. Microbiol. 52:1001-1005. [DOI] [PubMed] [Google Scholar]