International Journal of Systematic and Evolutionary Microbiology (2002), 52, 1391–1395 DOI: 10.1099/ijs.0.02021-0

Phylogenetic analysis of Prevotella nigrescens, NOTE Prevotella intermedia and Porphyromonas gingivalis clinical strains reveals a clear species clustering

Institute of Veterinary Peter Kuhnert,1 Joachim Frey,1 Niklaus P. Lang2 and Lisa Mayfield2 Bacteriology1 and School of Dental Medicine2 , University of Bern, Switzerland Author for correspondence: Peter Kuhnert. Tel: j41 31 6312485. Fax: j41 31 6312634. e-mail: peter.kuhnert!vbi.unibe.ch

Prevotella nigrescens, Prevotella intermedia and Porphyromonas gingivalis are oral pathogens from the family Bacteroidaceae, regularly isolated from cases of gingivitis and periodontitis. In this study, the phylogenetic variability of these three bacterial species was investigated by means of 16S rRNA (rrs) gene sequence comparisons of a set of epidemiologically and geographically diverse isolates. For each of the three species, the rrs gene sequences of 11 clinical isolates as well as the corresponding type strains was determined. Comparison of all rrs sequences obtained with those of closely related species revealed a clear clustering of species, with only a little intraspecies variability but a clear difference in the rrs gene with respect to the next related taxon. The results indicate that the three species form stable, homogeneous genetic groups, which favours an rrs-based species identification of these oral pathogens. This is especially useful given the 7% sequence divergence between Prevotella intermedia and Prevotella nigrescens, since phenotypic distinction between the two Prevotella species is inconsistent or involves techniques not applicable in routine identification.

Keywords: phylogeny, variation, 16S rRNA, oral pathogens

The periodontal pocket provides a unique habitat for parison, these periodontal pathogens are found rela- over 500 species of micro-organisms. These micro- tively infrequently and at low proportions within organisms are organized in a biofilm such that specific plaque samples from periodontally healthy individ- groups of bacteria are associated with one another in uals. Further evidence of the role of these pathogens complexes (Socransky et al., 1988). It is evident that a in destructive periodontal disease is provided by limited number of species of the oral microflora are studies demonstrating that the treatment aimed at associated with periodontal disease and may be identi- suppression of these species results in clinical im- fied in high proportions at sites with periodontal provement registered as reduced bleeding, reduced inflammation and destruction (Slots et al., 1986; Dzink pocket depth and gains in probing attachment level et al., 1988). (Slots et al., 1985; van Winkelhoff et al., 1988; Loos et al., 1988; Renvert et al., 1990). Microbial monitoring Prevotella intermedia and Porphyromonas gingivalis in dental practice may be an adjunct to diagnosis, are frequently isolated from subgingival plaque of choice of therapy, treatment evaluation, and risk individuals with periodontal disease. Prevotella nigres- evaluation and prognosis (Greenstein, 1988; Listgar- cens has been found to occur less frequently in ten, 1988; Maiden et al., 1990). Thus, a simple, reliable periodontal sites compared with Prevotella intermedia and reproducible method for identification of the (Dahlen et al., 1990; Haffajee et al., 1992). In com- bacterial species is desirable.

...... The three periodontal pathogens investigated in this Published online ahead of print on 28 January 2002 as DOI 10.1099/ study are all from the family Bacteroidaceae. Por- ijs.0.02021-0. phyromonas gingivalis (formerly Bacteroides gingivalis) The GenBank accession numbers for the 16S rRNA sequences from the has been recognized for some time as a single species family Bacteroidaceae are listed in Table 1. and can be relatively easily identified by phenotypic

02021 # 2002 IUMS Printed in Great Britain 1391 P. Kuhnert and others

Table 1. List of strains used for rrs gene sequence comparison

Strain* Original name Geographical origin† Clinical origin‡ Accession no.

Porphyromonas gingivalis Type strain ATCC 33277T ATCC Gingival sulcus AF414809 1-PGI OMGS 945 Sweden Periodontitis AF414810 2-PGI OMGS 2101 Kenya Periodontitis AF414811 3-PGI PFG 83 Indonesia  AF414812 40-PGI OMGS 716 Sweden  AF414813 42-PGI OMGS 789 Sweden  AF414814 25-PGI P02-2 Switzerland Periodontitis AF414815 26-PGI P06-2 Switzerland Periodontitis AF414816 27-PGI P06-3 Switzerland Periodontitis AF414817 28-PGI P10-1 Switzerland Periodontitis AF414818 29-PGI P11-6rough Switzerland Periodontitis AF414819 30-PGI P12-1 Switzerland Periodontitis AF414820 Prevotella intermedia Type strain ATCC 25611T ATCC Empyema AF414821 5-PIN MH 7 UK  AF414822 6-PIN MH 16 UK  AF414823 32-PIN HG 1674 The Netherlands  AF414824 35-PIN OMGS 874 Sweden Periodontitis AF414825 36-PIN OMGS 1277 Sweden Periodontitis AF414826 10-PIN P02-1 Switzerland Periodontitis AF414827 15-PIN P05-6 Switzerland Periodontitis AF414828 16-PIN P05-7 Switzerland Periodontitis AF414829 22-PIN P09-14 Switzerland Periodontitis AF414830 23-PIN P09-15 Switzerland Periodontitis AF414831 24-PIN P09-16 Switzerland Periodontitis AF414832 Prevotella nigrescens Type strain ATCC 33563T UK Gingivitis AF414833 7-PNIG MHS 1-18 Denmark Periodontitis AF414834 8-PNIG MH 20 UK  AF414835 9-PNIG MH 11 UK  AF414836 31-PNIG HG 1672 The Netherlands  AF414837 37-PNIG OMGS 865 Sweden  AF414838 11-PNIG P03-2 Switzerland Periodontitis AF414839 13-PNIG P04-4 Switzerland Periodontitis AF414840 17-PNIG P06-9 Switzerland Periodontitis AF414841 18-PNIG P07-10 Switzerland Periodontitis AF414842 20-PNIG P08-12 Switzerland Periodontitis AF414843 21-PNIG P08-13 Switzerland Periodontitis AF414844

* Strain acronyms: PGI for Porphyromonas gingivalis, PIN for Prevotella intermedia, PNIG for Prevotella nigrescens. † ATCC, American Type Culture Collection, Manassas, VA, USA. ‡ The strains were isolated from human individuals; , not determined. means (Krieg & Holt, 1984). Prevotella intermedia is groups representing the two species (Frandsen et al., an anaerobic, black-pigmented, Gram-negative rod 1995). and was recently split into two distinct species – Prevotella intermedia and Prevotella nigrescens –on Several methods have since been utilized to distinguish the basis of biochemical and DNA reassociation Prevotella intermedia from Prevotella nigrescens in the studies (Shah & Gharbia, 1992). The two species were clinical laboratory, with varying reliability. Phenotypic later confirmed by a polyphasic approach including identification and discrimination of Prevotella nigres- multilocus enzyme electrophoresis, ribotyping, and cens and Prevotella intermedia based on lipase pro- SDS-PAGE profiles. These different approaches car- duction, as originally described by Shah & Gharbia ried out with a representative sample number of (1992), proved to be problematic and was not repro- isolates produced clustering in the same two distinct ducible in different laboratories (Frandsen et al., 1995).

1392 International Journal of Systematic and Evolutionary Microbiology 52 Phylogeny of Prevotella and Porphyromonas spp.

Table 2. PCR and sequencing primers used for rrs amplification and sequence determination ...... The optimized primer set was adapted from Kuhnert et al. (1996).

Name Sequence Position*

16SUNI-L AGAGTTTGATCATGGCTCAG 8–27 16SUNI-R GTGTGACGGGCGGTGTGTAC 1410–1391 16SRNAI-S CTACGGGAGGCAGCAGTGGGG 341–361 16SRNA1-S CTACGGGAGGCAGCAGTGAGG 341–361 16SRNAII-S GTGTAGCGGTGAAATGCGTAG 682–702 16SRNA2-S GTGTAGGGGTAAAATCCGTAG 682–702 16SRNAIV-S GGTTAAGTCCCGCAACGAGCGC 1087–1108 16SRNA4-S GCTTAAGTGCCATAACGAGCGC 1087–1108 16SRNAV-S CCCCACTGCTGCCTCCCGTAG 361–341 16SRNAVI-S CTACGCATTTCACCGCTACAC 702–682 16SRNA6-S CTACGGATTTTACCCCTACAC 702–682 16SRNAVIII-S GCGCTCGTTGCGGGACTTAACC 1108–1087 16SRNA8-S GCGCTCGTTATGGCACTTAAGC 1108–1087

* Relative to the Escherichia coli sequence (J01859).

The usefulness of separation of whole-cell proteins by of the rrs gene and to evaluate published PCR assays, SDS-PAGE was also questioned (Premaraj et al., we carried out a study on the rrs gene sequences of 1999), and the ultimate comparison of genomic DNA clinical isolates of the three species Porphyromonas is time-consuming. Therefore, to distinguish the two gingivalis, Prevotella intermedia and Prevotella nigres- species, several methods are normally used in parallel. cens in comparison to their corresponding type A simple and robust identification method, however, is strains. For that purpose, we selected a representative needed for the clinical laboratory. sample of different geographically and epidemiologi- cally unrelated strains. They are listed in Table 1 and Paster et al. (1994) did a comprehensive study to included strains from individual patients, with perio- determine the phylogenetic relationships within the dontal disease, from Sweden, the UK, Denmark, The family Bacteroidaceae, using sequence analysis of the Netherlands, Kenya, Indonesia and Switzerland. In 16S rRNA. They found Porphyromonas gingivalis in addition, the type strains of the three species were the Porphyromonas cluster, and Prevotella intermedia included as the references with which the clinical and Prevotella nigrescens closely related but clearly isolates have to be compared. Clinical strains were separated within the Prevotella cluster, showing a clear previously identified by classical methods described difference between the two genera as well as between above, which included various phenotypic as well as the two species Prevotella intermedia and Prevotella genotypic approaches, but excluded rrs sequencing. nigrescens. This study was performed only with the The rrs gene sequence in our study was determined by type strains of these three species (in the case of a modification of a described method (Kuhnert et al., Prevotella intermedia, one additional isolate was in- 1996), using the PCR and sequencing primers listed in cluded) and hence does not give any clues as to the Table 2. Amplification of 1360 bp from the rrs gene intraspecies variability of the three pathogens. This was done directly using 2 µl culture and primers variability was shown to be high for several different 16SUNI-L and 16SUNI-R in a PCR. The reaction was reasons, including technical (e.g. sequencing mistakes) carried out in a 50 µl volume containing 5 µl10 Taq and biological reasons (e.g. different rrs copies) (Clay- i DNA polymerase buffer (Roche Molecular Biochemi- ton et al., 1995). Certain phenotypically defined cals), 20 pmol each primer, 1 mM dNTP (Roche bacterial species also can consist of more than one Molecular Biochemicals) and 2 5UTaq DNA poly- genotypic entity. This can be a particular problem in n merase (Roche Molecular Biochemicals). The PCR PCR detection and in identification methods based on conditions for the PE9600 thermal cycler (Applied rrs genes. For oral pathogens, several such PCR assays, Biosystems) equipped with MicroAmp tubes (Applied mainly based on the rrs gene, have been described Biosystems) were as follows: 3 min at 94 C followed (Conrads et al., 1996; Ashimoto et al., 1996; Premaraj m by 35 cycles at 94 C for 30 s, 54 C for 30 s and 72 C et al., 1999; Baumgartner et al., 1999). Some of them, m m m for 60 s. A final extension step of 7 min at 72 Cwas however, have been shown to be of low sensitivity, m included. After purification of the resulting amplifi- since the PCR methods were based on sequence data cation product, using the High Pure PCR Purification from single strains. Kit (Roche Molecular Biochemicals), the 1n36 kb PCR To obtain information on the intraspecies variability fragment was sequenced in both directions using http://ijs.sgmjournals.org 1393 P. Kuhnert and others

Prevotella

...... Fig. 1. 16S rRNA (rrs) gene sequence-based phylogenetic tree of strains analysed in this study and related GenBank entries from the family Bacteroidaceae. Clinical strains as well as type strains of Prevotella intermedia, Bacteroides Prevotella nigrescens and Porphyromonas gingivalis were included. All isolates of the three species form individual clusters with P. macacae (formerly salivosa) (L26103) their corresponding type strains. The acces- sion numbers of the sequences used are given in parentheses. The accession numbers of Porphyromonas sequences from this study are listed in Table 1. Actinobacillus actinomycetemcomitans was used as an outgroup. Strain acronyms: PIN for Prevotella intermedia; PNIG for Prevotella nigrescens; PGI for Porphyromonas gingivalis. The scale bar represents 5% sequence divergence.

5 pmol of the PCR and sequencing primers in eight identification and detection of Prevotella nigrescens, individual sequencing reactions. To apply a general Prevotella intermedia and Porphyromonas gingivalis. sequencing strategy taking into account sequence vari- This is especially useful for the direct detection and ability in the primer binding sites, the following primer identification of these pathogens in clinical material. pairs were used in the same reaction: 16SRNAI-S and 16SRNA1-S, 16SRNAII-S and 16SRNA2-S, Sequence variation within the species Prevotella inter- 16SRNAIV-S and 16SRNA4-S, 16SRNAVI-S and media was low, showing 18 variable positions among 16SRNA6-S, and 16SRNAVIII-S and 16SRNA8-S. the 1365 nucleotides. Most of these positions revealed The sequencing reaction was done using the dRhoda- ambiguous nucleotides, indicating the presence of more than one copy of the rrs gene. The type strain mine Terminator Sequencing Kit (Applied Biosystems) T and run on an ABI310 Genetic Analyser (Applied (ATCC 25611 ) showed the most ambiguous positions (n l 13). The type strain for Prevotella nigrescens Biosystems). Sequences were aligned and edited using T the software  (Gene Codes). The final (ATCC 33563 ) showed less-ambiguous positions proof-read sequences of the strains analysed in this (n l 6). The variability was also less than in Prevotella study and selected GenBank entries of the most related intermedia, a maximum of 13 out of 1363 nucleotides species of the family Bacteroidaceae were used for a being variable among the different strains, most of distance matrix calculated using the neighbour-joining them showing ambiguous nucleotides. Interestingly, in method (; GCG Package), and a tree was built Prevotella nigrescens we found in several strains, using the Jukes–Cantor algorithm in  (Kumar including the type strain, an insertion\deletion of one et al., 1993). or even two T at position 195 in one of the copies of the rrs genes. This insertion\deletion was not found with The resulting tree shows a clear clustering of all 11 strains 8-PNIG, 31-PNIG and 37-PNIG but was found strains of each of the three species around the rrs gene in the other isolates of Prevotella nigrescens. In the sequence of the corresponding type strain (Fig. 1). The sequence chromatograms this resulted in overlapping three species are clearly separated from their nearest peaks after position 195, as described by others relatives on the basis of the rrs gene. This is especially (Reischl et al., 1998). This difficulty was nevertheless noteworthy for the two species Prevotella nigrescens resolved by the sequence analysis of the opposite and Prevotella intermedia. The high correlation be- strand. The interspecies variability of rrs gene se- tween classical identification and rrs sequence analysis quences between Prevotella nigrescens and Prevotella strongly suggests that genotypic identification by rrs intermedia is more than 7%, allowing a clear-cut gene sequencing is suitable for the differentiation, distinction of the two species based on the rrs sequence.

1394 International Journal of Systematic and Evolutionary Microbiology 52 Phylogeny of Prevotella and Porphyromonas spp.

Porphyromonas gingivalis showed no ambiguous pos- Dzink, J. L., Socransky, S. S. & Haffajee, A. D. (1988). The pre- itions in the type strain (ATCC 33277T) rrs sequence, dominant cultivable microbiota of active and inactive lesions of J Clin Periodontol 15 and only a small intraspecies variability of six mis- destructive periodontal diseases. , 316–323. Frandsen, E. V., Poulsen, K. & Kilian, M. (1995). Confirmation of the matches within 1363 bp resulting in less than 0n5% species Prevotella intermedia and Prevotella nigrescens. Int J Syst variability. Bacteriol 45, 429–435. Greenstein, G. (1988). Microbiologic assessments to enhance perio- On the basis of rrs gene sequence analysis, we conclude dontal diagnosis. J Periodontol 59, 508–515. that the three species investigated form homogeneous Haffajee, A. D., Socransky, S. S., Smith, C. & Dibart, S. (1992). The phylogenetic groups, clearly separating them from use of DNA probes to examine the distribution of subgingival species in their most closely related taxa (Fig. 1). This is a good subjects with different levels of periodontal destruction. J Clin Periodon- reason for using rrs-based identification techniques, tol 19, 84–91. either by means of a specific PCR or by rrs sequence Krieg, N. R. & Holt, J. G. (editors) (1984). Bergey’s Manual of determination, as an alternative to conventional identi- Systematic Bacteriology, 8th edn. Baltimore: Williams & Wilkins. fication methods – especially for the closely related Kuhnert, P., Capaul, S. E., Nicolet, J. & Frey, J. (1996). Phylogenetic positions of Clostridium chauvoei and Clostridium septicum based on Prevotella species. However, primers for these methods 16S rRNA gene sequences. Int J Syst Bacteriol 46, 1174–1176. must be developed from rrs gene sequences of several Kumar, S., Tamura, K. & Nei, M. (1993). : Molecular Evolution strains for each species as presented in the current Genetics Analysis, version 1.0. University Park, PA: The Pensylvania study. Promising signature sequences suitable for the State University. design of specific primers are in the region 160–210 for Listgarten, M. A. (1988). A rationale for monitoring the periodontal Prevotella intermedia as well as in the region 420–480 microbiota after periodontal treatment. J Periodontol 59, 439–444. for Prevotella nigrescens (relative to the rrs sequences Loos, B., Claffey, N. & Crigger, M. (1988). Effects of oral hygiene of the type strains). These regions show high inter- measures on clinical and microbiological parameters of periodontal species variability, yet they are well conserved within disease. J Clin Periodontol 15, 211–216. the corresponding species. Maiden, M. F., Carman, R. J., Curtis, M. A., Gillett, I. R., Griffiths, G. S., Sterne, J. A., Wilton, J. M. & Johnson, N. W. (1990). Detection of high-risk groups and individuals for periodontal diseases: laboratory markers based on the microbiological analysis of subgingival plaque. J Clin Periodontol 17, 1–13. Acknowledgements Paster, B. J., Dewhirst, F. E., Olsen, I. & Fraser, G. J. (1994). Phylogeny of Bacteroides, Prevotella, and Porphyromonas spp. and We are grateful to Ellen V. G. Frandsen, A. J. Van Winkel- related bacteria. J Bacteriol 176, 725–732. hoff and G. Dahlen for providing us with strains. We thank Premaraj, T., Kato, N., Fukui, K., Kato, H. & Watanabe, K. (1999). Regula Hirschi and Marianne Weibel for technical as- Use of PCR and sodium dodecyl sulfate-polyacrylamide gel electro- sistance. This work was supported by the Priority Pro- phoresis techniques for differentiation of Prevotella intermedia sensu gramme Biotechnology of the Swiss National Science Foun- stricto and Prevotella nigrescens. J Clin Microbiol 37, 1057–1061. dation (grant 5002-057817). Reischl, U., Feldmann, K., Naumann, L., Gaugler, B. J. M., Ninet, B., Hirschel, B. & Emler, S. (1998). 16S rRNA sequence diversity in Mycobacterium celatum strains caused by the presence of two different copies of 16S rRNA gene. J Clin Microbiol 36, 1761–1764. Renvert, S., Wikstrom, M., Dahlen, G., Slots, J. & Egelberg, J. References (1990). Effect of root debridement on the elimination of Actinobacillus Ashimoto, A., Chen, C., Bakker, I. & Slots, J. (1996). Polymerase actinomycetemcomitans and Bacteroides gingivalis from periodontal chain reaction detection of 8 putative periodontal pathogens in pockets. J Clin Periodontol 17, 345–350. subgingival plaque of gingivitis and advanced periodontitis lesions. Shah, H. N. & Gharbia, S. E. (1992). Biochemical and chemical studies Oral Microbiol Immunol 11, 266–273. on strains designated Prevotella intermedia and proposal of a new Prevotella nigrescens Int J Syst Bacteriol 42 Baumgartner, J. C., Bae, K. S., Xia, T., Whitt, J. & David, L. L. pigmented species, sp. nov. , 542–546. (1999). Sodium dodecyl sulfate-polyacrylamide gel electrophoresis and polymerase chain reaction for differentiation of Prevotella intermedia Slots, J., Emrich, L. J., Genco, R. J. & Rosling, B. G. (1985). and Prevotella nigrescens. J Endod 25, 324–328. Relationship between some subgingival bacteria and periodontal pocket depth and gain or loss of periodontal attachment after treatment of Clayton, R. A., Sutton, G., Hinkle, P. S., Jr, Bult, C. & Fields, C. adult periodontitis. J Clin Periodontol 12, 540–552. (1995). Intraspecific variation in small-subunit rRNA sequences in Slots, J., Bragd, L., Wikstrom, M. & Dahlen, G. (1986). The GenBank: why single sequences may not adequately represent pro- occurrence of Actinobacillus actinomycetemcomitans, Bacteroides gingi- karyotic taxa. Int J Syst Bacteriol 45, 595–599. valis and Bacteroides intermedius in destructive periodontal disease in Conrads, G., Mutters, R., Fischer, J., Brauner, A., Lutticken, R. & adults. J Clin Periodontol 13, 570–577. Lampert, F. (1996). PCR reaction and dot-blot hybridization to Socransky, S. S., Haffajee, A. D., Dzink, J. L. & Hillman, J. D. (1988). monitor the distribution of oral pathogens within plaque samples of Associations between microbial species in subgingival plaque samples. periodontally healthy individuals. J Periodontol 67, 994–1003. Oral Microbiol Immunol 3, 1–7. Dahlen, G., Wikstrom, M., Renvert, S., Gmur, R. & Guggenheim, B. van Winkelhoff, A. J., van der Velden, U. & de Graaff, J. (1988). (1990). Biochemical and serological characterization of Bacteroides Microbial succession in recolonizing deep periodontal pockets after a intermedius strains isolated from the deep periodontal pocket. J Clin single course of supra- and subgingival debridement. J Clin Periodontol Microbiol 28, 2269–2274. 15, 116–122.

http://ijs.sgmjournals.org 1395