international Journal of Systematic Bacteriology (1 999), 49, 1695-1 700 Printed in Great Britain Phylogenetic evidence for reclassification of Calymmatobacterium granulomatis as Klebsiella granulomatis comb. nov. Jenny 5. Carter,’l2 Francis J. B~wden,~Ivan Ba~tian,~Garry M. Myers,’ K. S. Sriprakash’ and David J. Kemp’ Author for correspondence : David J. Kemp. Tel : + 6 18 8922 84 12. Fax : + 6 18 8927 5 187 e-mail : [email protected] 1 Menzies School of Health By sequencing a total of 2089 bp of the 16s rRNA and phoE genes it was Research, Darwin, demonstratedthat Calymmatobacterium grandomatis (the causative Austra Iia organism of donovanosis) shows a high level of identity with Klebsiella * Centre for Indigenous species pathogenic to humans (Klebsiellapneumoniae, Klebsiella Natural and Cultural Resource Management, rhinoscleromatis). It is proposed that C. grandomatis should be reclassified as Faculty of Aboriginal and Klebsiella granulomatis comb. nov. An emended description of the genus Torres Strait Islander Klebsiella is given. Studies, Northern Territory University, Darwin, Australia 3 Institute of Medical and Keywords : Calymmatobacteriurn, Klebsiella, sequence data, phylogenetic inferences Veterinary Science, Adelaide, Australia 4 AIDS/STD Unit, Royal Darwin Hospital, Darwin, Australia Calymmatobacterium granulomatis is the presumed ganism (Richens, 1991) have prevented further char- causative agent of donovanosis, an important cause of acterization of this relationship. genital ulceration that occurs in small endemic foci in all continents except Europe and Antarctica. The name Non-cultivable pathogenic eubacteria have been C. granulomatis was originally given to the pleo- identified by PCR using primers targeting conserved morphic bacterium cultured from donovanosis lesions genes (Fredricks & Relman, 1996). We have shown by Aragiio & Vianna (1913). Although these early through sequencing a 334 bp region of the phosphate cultures are highly dubious and were probably not of porin (phoE) gene that C. granulomatis has a high the organism itself (Richens, 1985), the name C. degree of molecular identity with other Klebsiella granulomatis has retained precedence over others. species in this region (Bastian & Bowden, 1996). We present here an analysis of the almost complete 16s C. granulomatis is known to be an encapsulated, non- rRNA and phoE sequences for C. granulomatis, K. motile, facultatively anaerobic, Gram-negative bac- rh inoscleromat is and Klebsiella pneumoniae. terium (Chandra & Jain, 199 1 ; Davis, 1970; Davis & At least two punch biopsies or swabs were obtained Collins, 1969; Dodson et al., 1974; Kuberski et al., from the lesions of eight patients with clinical dono- 1980). C. granulomatis has been associated with the vanosis. One punch biopsy was fixed and examined for genus Klebsiella because of the above characteristics Donovan bodies by the slow Giemsa technique (Sehgal and common antigenicity (Maddocks et al., 1975; & Jain, 1987) whilst DNA was extracted from the Packer & Goldberg, 1950). Furthermore, two members second punch biopsy (seven samples) (Mullenbach et of the genus Klebsiella that produce clinical disease al., 1989) or swab (one sample) (Beige et al., 1995). and pathologic changes very similar to those of C. DNA was also obtained from type cultures of K. granulomatis are Klebsiella rhinoscleromatis (Levine & pneumoniae [NCTC 9633T (National Collection of Hoyt, 1947; Shaw & Martin, 1961; Welsh et al., 1963), Type Cultures)] and K. rhinoscleromatis (NCTC which is implicated in a granulomatous disease of the 5046T). nose, and Klebsiella ozaenae (Richens, 1985) which is implicated in chronic atrophic rhinitis. Richens (1985) Primer sets used in this study include (5’-ACCTA- goes so far as to place C. granulomatis in the genus CCGCAACACCGACTTCTTCGG-3’ and 5’-TGA- Klebsiella, although difficulties in culturing the or- TCAGAACTGGTAGGTCAT-3’, 604 bp phoE prod- 01071 0 1999lUMS 1695 J. S. Carter and others Table 7. Strain designations and accession numbers for sequences obtained from the GenBank and EMBL databases, and for sequences novel to this study ...... .... Abbreviations : NCTC, National Collection of Type Cultures; ATCC, American Type Culture Collection; DSM, DMSZ - Deutsche Sammlung fiir Mikroorganismen und Zellkuturen ; JCM, Japanese Collection of Microorganisms ; PAH, Princess Alexandra Hospital, Brisbane, Australia; T, type Culture. Organism 16s rRNA phoE Strain Accession Strain Accession Calymmatobacteriurn granulomatis * AF009 17 17 * AF00923 17 Klebsiella rhinoscleromatis NCTC 5046T AF009 1697 NCTC 5046T AF0092307 Klebsiella pneumoniae DSM 30104T X87276 NCTC 9633T AF064793T Klebsiella oxytoca JCM 1665 AB004754 K26 X68022 Klebsiella ornithinolytica JCM 7251 AB004756 Klebsiella plan ticola DSM 3069T X932 15 Enterobacter cloacae l A04378 Enterobacter aerogenes JCM 1235 AB004750 Citrobacter freundii ATCC 29935 M5929 1 1 X6802 1 Esc her ic h ia co li ATCC 43895 283205 K12 V003 16 Serratia marcescens ATCC 13880 M59 160 UOC-5 1 U8 1967 Salmonella typhi Stlll US8545 1 x74595 Yersinia enterocolitica PAHll472 X68674 * Sequences obtained from clinical specimens. t Sequences novel to this study. 5: Strain not specified in GenBank report. uct) ; (5’-ATGAAAAAGAGTACTCTGGCA-3’and firming and extending the results obtained previously 5’-CAGACCGAAGTCGAACTGATACTG-3’, (Bastian & Bowden, 1996). 840 bp phoE product) ; (5’-CCTAACACATGCAA- GTCGA-3’ and 5‘-ATCCGACTTGACAGACCG-3’, In order to further investigate this relationship we 538 bp 16s rRNA product); (5’-CGGTCTGTCA- determined the sequence of a coding region spanning AGTCGGAT-3’ and 5’-TGTAGCACGTGTGTA- 95% of the phoE gene (1001/1059 bp) from K. GCC-3’, 653 bp 16s rRNA product). pneumoniae, K. rhinoscleromatis and C. granulomatis isolates, along with sequences for a 1088 bp region of One microlitre of template DNA was included in a the 16s rRNA from C. granulomatis and K. rhino- 25 pl PCR mixture that comprised the following scleromatis. The 16s rRNA sequence of K.pneumoniae solution: 10 mM Tris/HCl pH 8.0, 50mM KCl, [DSM 30104T(DSMZ, Deutsche Sammlung fur Mikro- 2.5 mM MgCl,, 0.01 YOgelatin, 0-2 pM dNTPs, 0.5 pM organismen und Zellkulturen)] was obtained from the primers and 0.5 Units Taq polymerase. Samples were GenBank database. These sequences were aligned with subjected to 30 cycles of denaturation (94 “C, 40 s), the sequences of other Klebsiella sp. and related annealing (55 “C, 40 s forphoE primers; 52 “C, 40 s for Enterobacteriaceae available from the GenBank/ 16s primers) and extension (72 “C, 60 s) in a thermal EMBL databases using the program CLUSTAL w cycler (Corbett Research, Sydney). All steps of PCR (Thompson et al., 1994). Due to the limited amount of template preparation, PCR mix and gel analysis were complete phoE sequence data available, Klebsiella conducted in three separate dedicated workrooms. planticola, Klebsiella ornithinolytica and Yersinia enter- Template negative controls were included in each ocolitica were only included in the 16s rRNA align- reaction. Products of the expected size were purified ment. For similar reasons different species of Entero- and sequenced in both directions using either the dye bacter have been compared for the 16s rRNA and terminator method (Applied Biosystems) or a manual phoE genes. T7 sequencing kit (Pharmacia-Biotech). Phylogenetic analyses was performed using the PHYLIP Initially, a 604 bp region of phoE was amplified and software package (Felsenstein, 1993). Programs em- sequenced from 4/4 patients with histologically con- ployed included SEQBOOT: to produce multiple data firmed donovanosis and from 4/4 patients with nega- sets by bootstrap resampling (100 iterations) ;DNADIST : tive histology for whom donovanosis remained the to compute distances between species for all data sets clinical diagnosis. The same Klebsiella-like sequence with the Jukes-Cantor algorithm; NEIGHBOR : to pro- was amplified from all eight patients thereby con- duce 1000 trees (100 data sets x 10 jumbled input 1696 lnterna tional Journal of Systematic Bacteriology 49 Phylogenetic analysis of C. granulornatis ~~ to this study are provided in Table 1 with the strain Yersinia enterocolitica designations. Phylogenetic trees, including the con- fidence values of branching are shown as cladograms Klebsiella pneumoniae in Fig. 1 for the 16s rRNA andphoE. It is evident from I tree topologies that C. has a close Calymmatobacterium granulomatis granulornatis phylogenetic relationship with members of the Klebsiella rhinoscleromatis Klebsiella genus known to be human pathogens (K. 71 rhinosclerornatis, K. pneurnoniae). Bootstrap analysis ,-- Serratia marcescens data confirmed that this association was highly sig- nificant for both the phoE and the 16s rRNA genes -1 -1 r Salmonellatyphi (bootstrap values 100 YO).Nucleotide similarities be- 54 tween the three isolates ranged from 98.8-99*8% for Escherichia coli the 16s rRNA and 99.7-99.8 YOfor phoE. IKlebsiella oxytoca PCR-based identification of non-cultivable microbial pathogens may be confounded by the incidental Enterobacler aerogenes amplification of colonizing non-pathogenic bacteria. Citrobacterfreundii Fredricks & Relman (1996) have proposed several '"4 criteria for establishing microbial disease causation by r Klebsiella planticola molecular methods. Their guidelines stipulate that : the sequence-based identification should be reproducible ; L Klebsiella ornithinolytica the nucleic acid sequence should be found at its highest concentration in diseased tissue but